**Section 1**

**Minerals and Glasses** 

10 Infrared Spectroscopy – Materials Science, Engineering and Technology

[8] Fowles, G.R. (1975).*Introduction to Modern Optics.* Second Edition. New York: Dover

[11] A.A. Michelson, Studies in Optics, University of Chicago, Press, Chicago (1962), 208 p [12] A.A Michelson and Morley, "on the Relative Motion of the Earth and the luminiferous

[13] Jean Baptiste Joseph Fourier, Oeuvres de Fourier, ( 1888); Idem Annals de Chimie et de Physique, 27, Paris, Annals of Chemistry and Physics, (1824) 236-281p [14] S. Tolansky, An Introduction to Interferometry, William Clowes and Sons Ltd.(1966),

[15] J. Anastassopoulou, E. Boukaki, C. Conti, P. Ferraris, E.Giorgini, C. Rubini, S. Sabbatini,

[18] Wikipedia, the free encyclopedia. *Infrared spectroscopy* http://en.wikipedia.org (July 28,

[19] Mount Holyoke College, South Hadley, Massachusetts. *Forensic applications of IR* 

[20] T. Theophanides, *Infrared and Raman Spectra of Biological Molecules*, NATO Advanced

[21] T. Theophanides, C. Sandorfy) *Spectroscopy of Biological Molecules*, NATO Advanced

[22] T. Theophanides *Fourier Transform Infrared Spectroscopy*, D. Reidel Publishing Co.

[23] T. Theophanides, *Inorganic Bioactivators*, NATO Advanced Study Institute, D. Reidel

[24] G. Vergoten and T. Theophanides, *Biomolecular Structure and Dynamics: Recent* 

[25] C. Conti, P. Ferraris, E. Giorgini, C. Rubini, S. Sabbatini, G. Tosi, J. Anastassopoulou, P.

[26] M. Petra, J. Anastassopoulou, T. Theologis & T. Theophanides, Synchrotron micro-FT-

[27] P. Kolovou and J. Anastassopoulou, "Synchrotron FT-IR spectroscopy of human bones.

[28] T. Theophanides, J. Anastassopoulou and N. Fotopoulos, *Fifth International Conference on* 

*experimental and Τheoretical Αdvances*, NATO Advanced Study Institute, Kluwer

Arapantoni, E. Boukaki, S FT-IR, T. Theophanides, C. Valavanis, FT-IR Microimaging Spectroscopy:Discrimination between healthy and neoplastic human

IR spectroscopic evaluation of normal paediatric human bone*, J. Mol Structure*, 78

The effect of aging". Brilliant Light in Life and Material Sciences, Eds. V. Tsakanov

*the Spectroscopy of Biological Molecules*, Kluwer Academic Publishers, Dodrecht,

Study Institute, D. Reidel Publishing Co. Dodrecht, 1978,372p.

Study Institute, D. Reidel Publishing Co. Dodrecht, 1984 , 646p

Ether" Am. J. of Science, 333-335(1887); F.Gires and P.Toumois," L'interféromètrie utilizable pour la compression lumineuse module en fréquence "Comptes Rendus

T. Theophanides, G. Tosi, Microimaging FT-IR spectroscopy on pathological breast

[10] Hecht, E. *Optics* .Fourth edition. San Francisco: Pearson Education Inc. (2002

de l'Académie des Sciences de Paris, 258, 6112-6115(1964)

tissues, *Vibrational Spectroscopy*, 51 (2009)270-275 [16] Melissa A. Page and W. Tandy Grubbs, J. Educ., 76(5), p.666 (1999) [17] Modern Spectroscopy, 2nd Edition, J.Michael Hollas,ISBN: 471-93076-8.

http://www.mtholyoke.edu (July 28, 2007

Publishing Co. Dodrecht, 1989,415p

Academic Publishers, The Netherlands, 1997, 327p

colon tissues , J. Mol Struc. 881 (2008) 46-51.

and H. Wiedemann, Springer, 2007 267-272p.

publications Inc., 336 p

253 p

2007).

Dodrecht, 1984.

(2005) 101

1991,409p

[9] Democritos, Avdera, Thrace, Greece, 460-370 BC

**1** 

*<sup>1</sup>Jožef Stefan Institute* 

*University of Ljubljana* 

*Slovenia* 

**Using Infrared Spectroscopy to** 

Tadej Rojac1, Primož Šegedin2 and Marija Kosec<sup>1</sup>

*<sup>2</sup>Faculty of Chemistry and Chemical Technology,* 

**Identify New Amorphous Phases –** 

**A Case Study of Carbonato Complex** 

**Formed by Mechanochemical Processing** 

Since the first laboratory experiments of M. Carey Lea and the original definition by F. W. Ostwald at the end of the 19th century, mechanochemistry, a field treating chemical changes induced in substances as a result of applied mechanical stress, has been evolved as an important area of chemistry from the viewpoint of both the fundamental research and applications (Takacs, 2004; Boldyrev & Tkačova, 2000). Whereas the fundamentals of mechanochemistry are still being extensively explored, the mechanical alloying, a powder metallurgy process involving ball milling of particles under high-energy impact conditions, met the commercial ground as early as in 1966 and was used to produce improved nickeland iron-based alloys for aerospace industry (Suryanarayana et al., 2001). In addition to metallurgy, the science and technology of mechanochemical processes are continuously developing within various other fields, including ceramics processing, processing of

Due to simplicity and technological reasons, the most common way to apply mechanical stress to a solid is via ball-particle collisions in a milling device. This is often referred to as the "high-energy milling" technique. What distinguish this method from the classical "wet ball-milling", used primarily for reducing particle size and/or mixing components, is that a powder or mixture of powders is typically milled in liquid-free conditions; under such circumstances, a larger amount of the kinetic energy of a moving ball inside a grinding bowl is transferred to the powder particles during collisions; this is also the origin of the term "high-energy" milling. Owing to the feasibility to conduct chemical reactions by high-energy milling, an often used term in the literature is

To carry out mechanochemical processes, various types of milling devices are used, including shaker, planetary, horizontal, attrition mill, etc. (Lu & Lai, 1998). One of the most

**1. Introduction** 

**1.1 Mechanochemistry and high-energy milling**

minerals, catalysis, pharmaceutics, and many others.

"mechanochemical synthesis".

## **Using Infrared Spectroscopy to Identify New Amorphous Phases – A Case Study of Carbonato Complex Formed by Mechanochemical Processing**

Tadej Rojac1, Primož Šegedin2 and Marija Kosec<sup>1</sup> *<sup>1</sup>Jožef Stefan Institute <sup>2</sup>Faculty of Chemistry and Chemical Technology, University of Ljubljana Slovenia* 

## **1. Introduction**

### **1.1 Mechanochemistry and high-energy milling**

Since the first laboratory experiments of M. Carey Lea and the original definition by F. W. Ostwald at the end of the 19th century, mechanochemistry, a field treating chemical changes induced in substances as a result of applied mechanical stress, has been evolved as an important area of chemistry from the viewpoint of both the fundamental research and applications (Takacs, 2004; Boldyrev & Tkačova, 2000). Whereas the fundamentals of mechanochemistry are still being extensively explored, the mechanical alloying, a powder metallurgy process involving ball milling of particles under high-energy impact conditions, met the commercial ground as early as in 1966 and was used to produce improved nickeland iron-based alloys for aerospace industry (Suryanarayana et al., 2001). In addition to metallurgy, the science and technology of mechanochemical processes are continuously developing within various other fields, including ceramics processing, processing of minerals, catalysis, pharmaceutics, and many others.

Due to simplicity and technological reasons, the most common way to apply mechanical stress to a solid is via ball-particle collisions in a milling device. This is often referred to as the "high-energy milling" technique. What distinguish this method from the classical "wet ball-milling", used primarily for reducing particle size and/or mixing components, is that a powder or mixture of powders is typically milled in liquid-free conditions; under such circumstances, a larger amount of the kinetic energy of a moving ball inside a grinding bowl is transferred to the powder particles during collisions; this is also the origin of the term "high-energy" milling. Owing to the feasibility to conduct chemical reactions by high-energy milling, an often used term in the literature is "mechanochemical synthesis".

To carry out mechanochemical processes, various types of milling devices are used, including shaker, planetary, horizontal, attrition mill, etc. (Lu & Lai, 1998). One of the most

Using Infrared Spectroscopy to Identify New Amorphous Phases –

a)

c)

polymorphic transitions

chemical reactions

decrease of crystallite size to nanometer range

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 15

b)

Fig. 1. a) Laboratory-scale planetary mill Fritsch Pulverisette 4, b) schematic representation of the movement of milling balls in a planetary mill (from Suryanarayana, 2001) and c) characteristic phenomena taking place in the solids as a result of high-energy collisions.

**High-energy milling** 

amorphization

plastic and elastic deformation

creation of structural defects

**1.2 Mechanochemical synthesis of complex ceramic oxides and underlying reaction** 

Mechanochemical synthesis (or high-energy milling assisted synthesis) has been found particularly useful for the synthesis of ceramic oxides with complex chemical composition, ranging from ferroelectric, magnetic and multiferroic oxides to oxides exhibiting semiconducting and catalytic properties. For an overview of the research activity in this

Whereas, in general, extensive literature data can be found on the mechanochemical synthesis of complex oxides, only limited studies are devoted to the understanding of mechanochemical reaction mechanisms. Primarily driven by the need to enrich our fundamental knowledge of mechanochemistry, the studies of reaction mechanisms have also been found to be essential in order to efficiently design a mechanochemical process, which includes the selection of milling parameters, milling regime, etc. (Rojac et al., 2010).

field the reader should consult Kong et al. (2008) and Sopicka-Lizer (2010).

**mechanisms** 

used, in particular for research purposes, is the planetary ball mill (Fig. 1a). A schematic view of the ball motion inside a grinding bowl of a planetary mill is illustrated in Fig. 1b. This characteristic ball motion results from two types of rotations: i) rotation of the grinding bowl around its center and ii) rotation of the supporting disc to which the bowls are attached; the two rotational senses are opposite (see Fig. 1b). In such a rotational geometry, the forces acting on the milling balls result into a periodical ball movement, illustrated by arrows in Fig. 1b, during which, when certain conditions are met, the balls are detached from the bowl's internal surface, colliding onto the powder particles on the opposite side. Even if simplified, the mathematical model derived from such an idealized ball movement agreed well with the experimental measurements of power consumption during milling (Burgio et al., 1991; Iasonna & Magini, 1996). In addition, this periodical movement was confirmed by numerical simulations (Watanabe et al., 1995a) and high-speed video camera recordings (Le Brun et al., 1993).

The high energy released during ball-powder collisions leads to various phenomena in the solid; this includes creation of a large amount of defects in the crystal structure, amorphization or complete loss of long-range structural periodicity, plastic and elastic deformation of particles, decrease of particle size down to the nanometer scale, increase of specific surface area of the powders, polymorphic transitions and even chemical reactions (Fig. 1c). Such changes result in distinct powder properties. The so-called mechanochemical reactions, which take place directly during the milling process without any external supply of thermal energy, make the method particularly interesting and distinguished from other conventional synthesis methods, which are typically based upon thermally driven reactions.

Due to their complexity, understanding mechanochemical reactions and the underlying mechanisms is a difficult task. In addition to local heating, provided by the high-energy collisions, modelling of the high-energy milling process revealed a large increase of pressure at the contact area between two colliding milling balls, which can reach levels of up to several GPa. It should be noted that both temperature and pressure rise are realized in tenths of microseconds, an estimated duration of a collision, illustrating the nonequilibrium nature of the mechanochemical process (Maurice & Courtney, 1990). Actually, during high-energy collisions the powder particles are subjected to a combination of hydrostatic and shear stress components, which further complicate the overall picture, even in apparently simple cases, such as polymorphic phase transitions. It was shown, for example, that conventional thermodynamic phase diagrams cannot be applied for polymorphic phase transitions realized during high-energy milling (Lin & Nadiv, 1979). In fact, the classical hydrostatic-pressure–temperature (*p*-*T*) phase diagram, e.g., in the case of a polymorphic transition between litharge and massicot forms of PbO, is considerably altered by introducing the shear component into the calculations; a twophase field region appears in the phase diagram, suggesting co-existence of the two polymorphs, rather than a sharp transition line characteristic for the conventional PbO *p*-*T* diagram. This might explain the often observed co-existence of two polymorphic modifications upon prolonged milling when "steady-state" milling conditions are reached (Lin & Nadiv, 1979; Iguchi & Senna, 1985). The influence of shear stress and local temperature rise on more complex mechanochemical reactions are still subject of intensive discussions.

used, in particular for research purposes, is the planetary ball mill (Fig. 1a). A schematic view of the ball motion inside a grinding bowl of a planetary mill is illustrated in Fig. 1b. This characteristic ball motion results from two types of rotations: i) rotation of the grinding bowl around its center and ii) rotation of the supporting disc to which the bowls are attached; the two rotational senses are opposite (see Fig. 1b). In such a rotational geometry, the forces acting on the milling balls result into a periodical ball movement, illustrated by arrows in Fig. 1b, during which, when certain conditions are met, the balls are detached from the bowl's internal surface, colliding onto the powder particles on the opposite side. Even if simplified, the mathematical model derived from such an idealized ball movement agreed well with the experimental measurements of power consumption during milling (Burgio et al., 1991; Iasonna & Magini, 1996). In addition, this periodical movement was confirmed by numerical simulations (Watanabe et al., 1995a) and high-speed video camera

The high energy released during ball-powder collisions leads to various phenomena in the solid; this includes creation of a large amount of defects in the crystal structure, amorphization or complete loss of long-range structural periodicity, plastic and elastic deformation of particles, decrease of particle size down to the nanometer scale, increase of specific surface area of the powders, polymorphic transitions and even chemical reactions (Fig. 1c). Such changes result in distinct powder properties. The so-called mechanochemical reactions, which take place directly during the milling process without any external supply of thermal energy, make the method particularly interesting and distinguished from other conventional synthesis methods, which are typically based upon

Due to their complexity, understanding mechanochemical reactions and the underlying mechanisms is a difficult task. In addition to local heating, provided by the high-energy collisions, modelling of the high-energy milling process revealed a large increase of pressure at the contact area between two colliding milling balls, which can reach levels of up to several GPa. It should be noted that both temperature and pressure rise are realized in tenths of microseconds, an estimated duration of a collision, illustrating the nonequilibrium nature of the mechanochemical process (Maurice & Courtney, 1990). Actually, during high-energy collisions the powder particles are subjected to a combination of hydrostatic and shear stress components, which further complicate the overall picture, even in apparently simple cases, such as polymorphic phase transitions. It was shown, for example, that conventional thermodynamic phase diagrams cannot be applied for polymorphic phase transitions realized during high-energy milling (Lin & Nadiv, 1979). In fact, the classical hydrostatic-pressure–temperature (*p*-*T*) phase diagram, e.g., in the case of a polymorphic transition between litharge and massicot forms of PbO, is considerably altered by introducing the shear component into the calculations; a twophase field region appears in the phase diagram, suggesting co-existence of the two polymorphs, rather than a sharp transition line characteristic for the conventional PbO *p*-*T* diagram. This might explain the often observed co-existence of two polymorphic modifications upon prolonged milling when "steady-state" milling conditions are reached (Lin & Nadiv, 1979; Iguchi & Senna, 1985). The influence of shear stress and local temperature rise on more complex mechanochemical reactions are still subject of intensive

recordings (Le Brun et al., 1993).

thermally driven reactions.

discussions.

Fig. 1. a) Laboratory-scale planetary mill Fritsch Pulverisette 4, b) schematic representation of the movement of milling balls in a planetary mill (from Suryanarayana, 2001) and c) characteristic phenomena taking place in the solids as a result of high-energy collisions.

#### **1.2 Mechanochemical synthesis of complex ceramic oxides and underlying reaction mechanisms**

Mechanochemical synthesis (or high-energy milling assisted synthesis) has been found particularly useful for the synthesis of ceramic oxides with complex chemical composition, ranging from ferroelectric, magnetic and multiferroic oxides to oxides exhibiting semiconducting and catalytic properties. For an overview of the research activity in this field the reader should consult Kong et al. (2008) and Sopicka-Lizer (2010).

Whereas, in general, extensive literature data can be found on the mechanochemical synthesis of complex oxides, only limited studies are devoted to the understanding of mechanochemical reaction mechanisms. Primarily driven by the need to enrich our fundamental knowledge of mechanochemistry, the studies of reaction mechanisms have also been found to be essential in order to efficiently design a mechanochemical process, which includes the selection of milling parameters, milling regime, etc. (Rojac et al., 2010).

Using Infrared Spectroscopy to Identify New Amorphous Phases –

closely the transitional amorphous phase.

mechanisms in which carbonate ions (CO3

a reduced piezoelectric response (Rojac et al., 2010).

Avvakumov et al. (2001).

comprising CO3

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 17

90-hours separately milled Nb2O5 (Fig. 2a, Nb2O5 90 h), revealed a much larger degree of amophization of Nb2O5 when co-milled with K2CO3; note the considerably weaker Nb2O5 peaks and higher XRD background in the case of the mixture as compared to separately milled Nb2O5. This suggests that the amorphization of Nb2O5 is not a consequence of the high-energy impacts only, but has its origin in the mechanochemical interaction with the carbonate. It should be emphasized that this is not an isolated case; examples involving transitional amorphous phases can also be found during mechanical alloying of mixture of metals (El-Eskandarany et al., 1997). Finally, a nucleation-and-growth mechanism from amorphous phase was recently proposed as a general concept to explain the mechanochemical synthesis of a variety of complex oxides, such as Pb(Zr0.52Ti0.48)O3, Pb(Mg1/3Nb2/3)O3, Pb(Zn1/3Nb2/3)O3, etc. (Wang et al., 2000a, 2000b; Kuscer et al., 2006). In order to understand mechanochemical reactions, it is thus indispensable to analyze more

It is clear from the above considerations that the most often used and widely reported XRD analysis becomes insufficient to provide detailed information about amorphous phases. The benefits of in-depth studies of mechanochemical reaction mechanisms by selection of appropriate analytical tools, able to provide data on a short-range (local) structural scale, such as nuclear magnetic resonance (NMR), X-ray photoelectron spectroscopy (XPS), electron paramagnetic resonance (EPR) spectroscopy, infrared spectroscopy (IR), Raman spectroscopy, etc., were demonstrated by the pioneering work of Senna, Watanabe and coworkers (Watanabe et al., 1996, 1997; Senna, 1997). In those cases, the synthesis of selected complex oxide systems have been studied from starting mixtures comprising typically hydroxide and oxide compounds; extensive data on these studies can be found in

Mechanochemical processing has recently provided important improvements in the synthesis of ceramic materials in the family of alkaline niobates tantalates, a rich group of materials exhibiting wide applicability; this includes KTaO3 and (K,Na,Li)(Nb,Ta)O3 (KNLNT), which are considered as promising materials for dielectric (microwave) and piezoelectric applications, respectively (Glinsek et al., 2011; Tchernychova et al., 2011; Rojac et al., 2008a, 2010). Since alkali carbonates are the most frequently used as starting alkali compounds, it naturally became of interest to understand in more details the mechanochemical reaction

important practical consequences. For example, in the case of the synthesis of the complex KNLNT solid solution, it was demonstrated that the identification of the reaction mechanism during mechanochemical processing is a key step leading to highly homogeneous KNLNT ceramics with excellent piezoelectric response. After identifying an intermediate amorphous carbonato complex, to which the present chapter is particularly devoted, it was found that a homogeneous KNLNT can only be obtained by providing the formation of this complex during the high-energy milling step. In other words, milling conditions that did not lead to the formation of the carbonato complex, e.g., milling in the "friction" mode instead of the "friction+impact" mode, resulted into considerable Ta-inhomogeneities and, consequently, to

In this chapter we present an overview of the studies of reaction mechanisms in systems

various analytical methods, including quantitative XRD analysis, thermal analysis and

2– ions. The chapter aims primarily at showing the importance of combining

2–) are involved. The results of these studies carry

One of the main difficulties in analyzing the complex mechanisms of mechanochemical reactions is the identification of amorphous phases, which are metastable and appear often transitional with respect to the course of the reaction. To illustrate an example, we present in Fig. 2 the mechanochemical synthesis of KNbO3 from a powder mixture of K2CO3 and Nb2O5 (Rojac et al., 2009). In the first 90 hours of milling, the initial crystalline K2CO3 and Nb2O5 (Fig. 2a, 0 h) are transformed into an amorphous phase, characterized by two broad "humps" centred at around 29° and 54° 2-theta (Fig. 2a, 90 h). The formation of the amorphous phase was confirmed by transmission electron microscopy (TEM), i.e., an amorphous matrix was observed with embedded nanocrystalline particles of Nb2O5 (Fig. 2b), which is consistent with the X-ray diffraction (XRD) pattern (Fig. 2a, 90 h). Further milling from 90 to 350 hours resulted in the crystallization from the amorphous phase; this is evident from the appearance of new peaks after 150 and 350 hours of milling, which were assigned to various potassium niobate phases with different K/Nb molar ratio (Fig. 2a, 150 and 350 h). Therefore, the amorphous phase represents a transitional phase of the reaction. In addition, comparison of the 90-hours milled K2CO3–Nb2O5 mixture (Fig. 2a, 90 h) with the

Fig. 2. a) XRD patterns of K2CO3–Nb2O5 powder mixture after high-energy milling for 20, 90, 150 and 350 hours. The non-milled mixture is denoted as "0 h". The pattern of the 90-hoursseparately-milled Nb2O5 is added for comparison. In order to prevent adsorption of water during XRD measurements, a polymeric foil was used to cover the non-milled powder mixture. b) TEM image of the K2CO3–Nb2O5 powder mixture after high-energy milling for 90 hours. Notations: K2CO3 (○, PDF 71-1466), Nb2O5 (●, PDF 30-0873), KNbO3 (▲, PDF 71- 0946), K6Nb10.88O30 (□ , PDF 87-1856), K8Nb18O49 (◊, PDF 31-1065), polymeric foil (F); "h" denotes milling hours (from Rojac et al., 2009).

One of the main difficulties in analyzing the complex mechanisms of mechanochemical reactions is the identification of amorphous phases, which are metastable and appear often transitional with respect to the course of the reaction. To illustrate an example, we present in Fig. 2 the mechanochemical synthesis of KNbO3 from a powder mixture of K2CO3 and Nb2O5 (Rojac et al., 2009). In the first 90 hours of milling, the initial crystalline K2CO3 and Nb2O5 (Fig. 2a, 0 h) are transformed into an amorphous phase, characterized by two broad "humps" centred at around 29° and 54° 2-theta (Fig. 2a, 90 h). The formation of the amorphous phase was confirmed by transmission electron microscopy (TEM), i.e., an amorphous matrix was observed with embedded nanocrystalline particles of Nb2O5 (Fig. 2b), which is consistent with the X-ray diffraction (XRD) pattern (Fig. 2a, 90 h). Further milling from 90 to 350 hours resulted in the crystallization from the amorphous phase; this is evident from the appearance of new peaks after 150 and 350 hours of milling, which were assigned to various potassium niobate phases with different K/Nb molar ratio (Fig. 2a, 150 and 350 h). Therefore, the amorphous phase represents a transitional phase of the reaction. In addition, comparison of the 90-hours milled K2CO3–Nb2O5 mixture (Fig. 2a, 90 h) with the

Fig. 2. a) XRD patterns of K2CO3–Nb2O5 powder mixture after high-energy milling for 20, 90, 150 and 350 hours. The non-milled mixture is denoted as "0 h". The pattern of the 90-hoursseparately-milled Nb2O5 is added for comparison. In order to prevent adsorption of water during XRD measurements, a polymeric foil was used to cover the non-milled powder mixture. b) TEM image of the K2CO3–Nb2O5 powder mixture after high-energy milling for 90 hours. Notations: K2CO3 (○, PDF 71-1466), Nb2O5 (●, PDF 30-0873), KNbO3 (▲, PDF 71- 0946), K6Nb10.88O30 (□ , PDF 87-1856), K8Nb18O49 (◊, PDF 31-1065), polymeric foil (F); "h"

b)

**nanocrystalline Nb2O5**

**amorphous phase** 

denotes milling hours (from Rojac et al., 2009).

a)

90-hours separately milled Nb2O5 (Fig. 2a, Nb2O5 90 h), revealed a much larger degree of amophization of Nb2O5 when co-milled with K2CO3; note the considerably weaker Nb2O5 peaks and higher XRD background in the case of the mixture as compared to separately milled Nb2O5. This suggests that the amorphization of Nb2O5 is not a consequence of the high-energy impacts only, but has its origin in the mechanochemical interaction with the carbonate. It should be emphasized that this is not an isolated case; examples involving transitional amorphous phases can also be found during mechanical alloying of mixture of metals (El-Eskandarany et al., 1997). Finally, a nucleation-and-growth mechanism from amorphous phase was recently proposed as a general concept to explain the mechanochemical synthesis of a variety of complex oxides, such as Pb(Zr0.52Ti0.48)O3, Pb(Mg1/3Nb2/3)O3, Pb(Zn1/3Nb2/3)O3, etc. (Wang et al., 2000a, 2000b; Kuscer et al., 2006). In order to understand mechanochemical reactions, it is thus indispensable to analyze more closely the transitional amorphous phase.

It is clear from the above considerations that the most often used and widely reported XRD analysis becomes insufficient to provide detailed information about amorphous phases. The benefits of in-depth studies of mechanochemical reaction mechanisms by selection of appropriate analytical tools, able to provide data on a short-range (local) structural scale, such as nuclear magnetic resonance (NMR), X-ray photoelectron spectroscopy (XPS), electron paramagnetic resonance (EPR) spectroscopy, infrared spectroscopy (IR), Raman spectroscopy, etc., were demonstrated by the pioneering work of Senna, Watanabe and coworkers (Watanabe et al., 1996, 1997; Senna, 1997). In those cases, the synthesis of selected complex oxide systems have been studied from starting mixtures comprising typically hydroxide and oxide compounds; extensive data on these studies can be found in Avvakumov et al. (2001).

Mechanochemical processing has recently provided important improvements in the synthesis of ceramic materials in the family of alkaline niobates tantalates, a rich group of materials exhibiting wide applicability; this includes KTaO3 and (K,Na,Li)(Nb,Ta)O3 (KNLNT), which are considered as promising materials for dielectric (microwave) and piezoelectric applications, respectively (Glinsek et al., 2011; Tchernychova et al., 2011; Rojac et al., 2008a, 2010). Since alkali carbonates are the most frequently used as starting alkali compounds, it naturally became of interest to understand in more details the mechanochemical reaction mechanisms in which carbonate ions (CO3 2–) are involved. The results of these studies carry important practical consequences. For example, in the case of the synthesis of the complex KNLNT solid solution, it was demonstrated that the identification of the reaction mechanism during mechanochemical processing is a key step leading to highly homogeneous KNLNT ceramics with excellent piezoelectric response. After identifying an intermediate amorphous carbonato complex, to which the present chapter is particularly devoted, it was found that a homogeneous KNLNT can only be obtained by providing the formation of this complex during the high-energy milling step. In other words, milling conditions that did not lead to the formation of the carbonato complex, e.g., milling in the "friction" mode instead of the "friction+impact" mode, resulted into considerable Ta-inhomogeneities and, consequently, to a reduced piezoelectric response (Rojac et al., 2010).

In this chapter we present an overview of the studies of reaction mechanisms in systems comprising CO3 2– ions. The chapter aims primarily at showing the importance of combining various analytical methods, including quantitative XRD analysis, thermal analysis and

Using Infrared Spectroscopy to Identify New Amorphous Phases –

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 19

Fig. 3. XRD patterns of Na2CO3–Nb2O5 powder mixture after high-energy milling for 5, 40, 160 and 400 hours. The non-milled mixture is denoted as "0 h". Notations: Na2CO3 (∆, PDF 19-1130), Nb2O5 (○, PDF 30-0873) and NaNbO3 (●, PDF 33-1270); "h" denotes milling hours

of both Na2CO3 and Nb2O5 decrease with milling time (Fig. 4a). While Nb2O5 persists in the mixture up to 280 hours (Fig. 4a, closed rectangular), Na2CO3 is no longer detected after 20 hours of milling (Fig. 4a, open rectangular). The amount of the XRD-amorphous phase rapidly increases in the initial part of the reaction, reaching a maximum of 91% after 110 hours of milling, after which it decreases with further milling. Note the constant amount of the XRD-amorphous phase after reaching 600 hours of milling. The formation of NaNbO3 follows a sigmoidal trend: at the beginning of the reaction the formation rate is low, after which it increases and slows down again in the final part of the reaction (Fig. 4b, open circles). Similarly like the XRD-amorphous phase, no differences in the amount of NaNbO3 are observed with milling from 600 to 700 hours, suggesting a constant NaNbO3-to-

From the quantitative analysis, shown in Fig. 4, an important observation can be derived by looking more closely at the initial stage of the reaction. An enlarged view of this part of the reaction is shown as inset in Fig. 4b. Here, we can see that in the initial 20 hours of milling, during which no NaNbO3 was detected, a large amount, i.e., 73%, of the amorphous phase was formed. Only subsequently, i.e., after 40 hours of milling, NaNbO3 was firstly detected.

(from Rojac et al., 2008b).

amorphous-phase mass ratio upon prolonged milling.

infrared spectroscopy, to obtain an overall picture of a complex reaction mechanism, such as the one encountered during mechanochemical processing. The first part of the chapter is devoted to the synthesis of NaNbO3 from a mixture of Na2CO3 and Nb2O5. After demonstrating the feasibility of synthesizing NaNbO3 directly by high-energy milling, we show systematically how a mechanism can be revealed by a built-up of data from various analytical methods. The focus is to gain insight into the amorphous phase, which represents a transitional phase in the synthesis of NaNbO3. In the second part of the chapter we will extent the studies to other systems based on sodium carbonate, i.e., Na2CO3–M2O5 (M = V, Nb, Ta). The transition-metal oxides were selected through the 5th group of the periodic table to allow systematic comparisons and propose potentially a general reaction mechanism.

#### **2. Mechanochemical reaction mechanism in the Na2CO3–Nb2O5 system studied by a combination of quantitative X-ray diffraction, thermal and infrared spectroscopy analysis**

#### **2.1 Quantitative X-ray diffraction analysis**

The mechanochemical synthesis of NaNbO3 from a Na2CO3–Nb2O5 mixture was followed by XRD analysis. Fig. 3 shows the XRD patterns of the Na2CO3–Nb2O5 mixture after selected milling times. The pattern of the non-milled mixture (Fig. 3, 0 h), which is a homogenized mixture of Na2CO3 and Nb2O5 powders just before mechanochemical treatment, can be fully indexed with the initial monoclinic Na2CO3 and orthorhombic Nb2O5 (Fig. 3, 0 h). The first 5 hours of high-energy milling are characterized by broaden peaks of the two reagents together with reduced peak intensity (Fig. 3, 5 h). After 40 hours of milling Na2CO3 was not observed anymore in the mixture, whereas traces of the newly formed NaNbO3 were first detected (Fig. 3, 40 h). Further milling from 40 to 400 h leaded to a progressive disappearance of Nb2O5 from the mixture at the expense of the growing NaNbO3. Note the long milling time, i.e., 400 hours, needed to obtain the final NaNbO3 free of any reagents (Fig. 3, 400 h). The low rate of the reaction between Na2CO3 and Nb2O5 resulted from the mild milling conditions, which were applied intentionally in order to enable a careful analysis of the individual reaction stages. It should be noted, however, that more intensive milling, resulting into NaNbO3 after 32 hours of milling, did not change qualitatively the course of the reaction (for details see Rojac et al., 2008b). The results of the XRD analysis from Fig. 3 confirm the mechanochemical formation of NaNbO3 according to the following reaction:

$$\text{Na}\_2\text{CO}\_3 + \text{Nb}\_2\text{O}\_5 \xrightarrow{} 2\text{NaNb}\_2\text{O} + \text{CO}\_2\tag{1}$$

In order to obtain a more quantitative picture of the mechanochemical reaction, we performed a quantitative XRD phase analysis using the Rietveld refinement method. In addition to the amount of the crystalline phases, i.e., Na2CO3, Nb2O5 and NaNbO3, we determined also the contribution from the XRD background, which we denoted as "XRDamorphous" phase. This was done using an internal standard method; details of the method can be found in Kuscer et al. (2006) and Rojac et al. (2008b).

The results of the refinement analysis in terms of the amounts of Na2CO3, Nb2O5, NaNbO3 and XRD-amorphous phase as a function of milling time are shown in Fig. 4. The amounts

infrared spectroscopy, to obtain an overall picture of a complex reaction mechanism, such as the one encountered during mechanochemical processing. The first part of the chapter is devoted to the synthesis of NaNbO3 from a mixture of Na2CO3 and Nb2O5. After demonstrating the feasibility of synthesizing NaNbO3 directly by high-energy milling, we show systematically how a mechanism can be revealed by a built-up of data from various analytical methods. The focus is to gain insight into the amorphous phase, which represents a transitional phase in the synthesis of NaNbO3. In the second part of the chapter we will extent the studies to other systems based on sodium carbonate, i.e., Na2CO3–M2O5 (M = V, Nb, Ta). The transition-metal oxides were selected through the 5th group of the periodic table to allow systematic comparisons and propose potentially a general reaction

**2. Mechanochemical reaction mechanism in the Na2CO3–Nb2O5 system studied by a combination of quantitative X-ray diffraction, thermal and** 

The mechanochemical synthesis of NaNbO3 from a Na2CO3–Nb2O5 mixture was followed by XRD analysis. Fig. 3 shows the XRD patterns of the Na2CO3–Nb2O5 mixture after selected milling times. The pattern of the non-milled mixture (Fig. 3, 0 h), which is a homogenized mixture of Na2CO3 and Nb2O5 powders just before mechanochemical treatment, can be fully indexed with the initial monoclinic Na2CO3 and orthorhombic Nb2O5 (Fig. 3, 0 h). The first 5 hours of high-energy milling are characterized by broaden peaks of the two reagents together with reduced peak intensity (Fig. 3, 5 h). After 40 hours of milling Na2CO3 was not observed anymore in the mixture, whereas traces of the newly formed NaNbO3 were first detected (Fig. 3, 40 h). Further milling from 40 to 400 h leaded to a progressive disappearance of Nb2O5 from the mixture at the expense of the growing NaNbO3. Note the long milling time, i.e., 400 hours, needed to obtain the final NaNbO3 free of any reagents (Fig. 3, 400 h). The low rate of the reaction between Na2CO3 and Nb2O5 resulted from the mild milling conditions, which were applied intentionally in order to enable a careful analysis of the individual reaction stages. It should be noted, however, that more intensive milling, resulting into NaNbO3 after 32 hours of milling, did not change qualitatively the course of the reaction (for details see Rojac et al., 2008b). The results of the XRD analysis from Fig. 3 confirm the mechanochemical formation of NaNbO3 according to the following

 Na2CO3 + Nb2O5 2NaNbO3 + CO2 (1) In order to obtain a more quantitative picture of the mechanochemical reaction, we performed a quantitative XRD phase analysis using the Rietveld refinement method. In addition to the amount of the crystalline phases, i.e., Na2CO3, Nb2O5 and NaNbO3, we determined also the contribution from the XRD background, which we denoted as "XRDamorphous" phase. This was done using an internal standard method; details of the method

The results of the refinement analysis in terms of the amounts of Na2CO3, Nb2O5, NaNbO3 and XRD-amorphous phase as a function of milling time are shown in Fig. 4. The amounts

can be found in Kuscer et al. (2006) and Rojac et al. (2008b).

mechanism.

reaction:

**infrared spectroscopy analysis** 

**2.1 Quantitative X-ray diffraction analysis** 

Fig. 3. XRD patterns of Na2CO3–Nb2O5 powder mixture after high-energy milling for 5, 40, 160 and 400 hours. The non-milled mixture is denoted as "0 h". Notations: Na2CO3 (∆, PDF 19-1130), Nb2O5 (○, PDF 30-0873) and NaNbO3 (●, PDF 33-1270); "h" denotes milling hours (from Rojac et al., 2008b).

of both Na2CO3 and Nb2O5 decrease with milling time (Fig. 4a). While Nb2O5 persists in the mixture up to 280 hours (Fig. 4a, closed rectangular), Na2CO3 is no longer detected after 20 hours of milling (Fig. 4a, open rectangular). The amount of the XRD-amorphous phase rapidly increases in the initial part of the reaction, reaching a maximum of 91% after 110 hours of milling, after which it decreases with further milling. Note the constant amount of the XRD-amorphous phase after reaching 600 hours of milling. The formation of NaNbO3 follows a sigmoidal trend: at the beginning of the reaction the formation rate is low, after which it increases and slows down again in the final part of the reaction (Fig. 4b, open circles). Similarly like the XRD-amorphous phase, no differences in the amount of NaNbO3 are observed with milling from 600 to 700 hours, suggesting a constant NaNbO3-toamorphous-phase mass ratio upon prolonged milling.

From the quantitative analysis, shown in Fig. 4, an important observation can be derived by looking more closely at the initial stage of the reaction. An enlarged view of this part of the reaction is shown as inset in Fig. 4b. Here, we can see that in the initial 20 hours of milling, during which no NaNbO3 was detected, a large amount, i.e., 73%, of the amorphous phase was formed. Only subsequently, i.e., after 40 hours of milling, NaNbO3 was firstly detected.

Using Infrared Spectroscopy to Identify New Amorphous Phases –

hours (from Rojac et al., 2006).

34.8 35.8 36.8 37.8 38.8 2°)

∆ Na2CO3

<sup>∆</sup> <sup>∆</sup> <sup>∆</sup> <sup>∆</sup>

<sup>∆</sup> <sup>∆</sup>

a) Na2CO3 + Nb2O5

∆

**40 h** 

**20 h**

**5 h**

**1 h**

**0 h**

Rel. Int. (a.u.)

Rel. Int. (a.u.)

∆

∆

**2.2 Thermal analysis** 

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 21

b) Na2CO3 (separate milling)

Fig. 5. XRD patterns of a) Na2CO3–Nb2O5 mixture and b) Na2CO3 after high-energy milling for up to 40 hours. The pattern in a) shows a narrow 2-theta region, i.e., from 34.8 to 38.8°, to highlight the changes upon milling in the peaks corresponding to Na2CO3. Note that all the peaks on the patterns of non-milled and 40-hours-separately-milled Na2CO3 in b) are indexed with monoclinic Na2CO3. Notation: Na2CO3 (∆, PDF 19-1130); "h" denotes milling

20 25 30 35 40 45 50 55 60 65 70 2°)

**0 h** 

**40 h** 

**Na2CO3** 

interaction between Na2CO3 and Nb2O5, a question that arises at this point is whether this interaction resulted into the carbonate decomposition. This is also relevant with respect to the nature of the amorphous phase. Obviously, further information could be obtained by following the decomposition of the carbonate during milling. This can be done using thermogravimetric (TG) analysis; the results of TG coupled with differential thermal analysis (DTA) and evolved-gas analysis (EGA) are presented in the following section.

In order to explore the origin of the reaction-induced amorphization and/or possible decomposition of Na2CO3 (Fig. 5) we were further focused on the initial part of milling, i.e.,

Fig. 6 presents the thermogravimetric (TG), derivative thermogravimetric (DTG), differential thermal analysis (DTA) and evolved-gas analysis (EGA) curves of the Na2CO3–Nb2O5 powder mixture in the first 40 hours of high-energy milling. The nonmilled Na2CO3–Nb2O5 mixture looses mass in several steps in a broad temperature range from 400 °C to 800 °C (Fig. 6a and b, 0 h). The total mass loss of this mixture upon annealing to 900 °C amounts to 11.7%, which agrees well with the theoretical mass loss of 11.8%, calculated according to equation 1 for the complete decomposition of Na2CO3 in an

results are presented for the samples treated in the first 40 hours of milling.

Fig. 4. Fractions of crystalline phases (Na2CO3, Nb2O5 and NaNbO3) and XRD-amorphous phase, determined by Rietveld refinement analysis, as a function of milling time. a) Na2CO3 and Nb2O5, b) NaNbO3 and XRD-amorphous phase. The inset of b) shows an enlarged view of the curves in the initial 80 hours of milling. The lines are drawn as a guide for the eye (from Rojac et al., 2008b).

From this simple observation we can infer that NaNbO3 is not formed directly, like assumed by equation 1, but through an intermediate amorphous phase. The transitional nature of the amorphous phase is further confirmed by the maximum in its amount after 110 hours of milling. Moreover, literature data go in favour of our conclusions. In fact, based on studies of the kinetics, the sigmoidal trend, like that observed in the case of NaNbO3 (Fig. 4b, open circles), is characteristic for multistep mechanochemical processes, such as the amorphization of a mixture of metals, where the phase transformation requires two or more impacts on the same powder fraction. In contrast, continuously decelerating processes, described by asymptotic kinetics, are typical for the amorphization of single-phase compounds, such as intermetallics, where the structure is already altered after the first impact (Delogu & Cocco, 2000; Cocco et al., 2000; Delogu et al., 2004). Therefore, independently of the analysis on the XRD-amorphous phase, the sigmoidal-like trend in the formation of NaNbO3 (Fig. 4b, open circles) suggests that the niobate is formed via a transitional phase.

In addition to the XRD-amorphous phase, we shall look at the changes induced in the Na2CO3 in the initial part of milling. Fig. 5 compares the XRD patterns of the Na2CO3–Nb2O5 mixture in the first 40 hours of milling (Fig. 5a) with the XRD patterns of Na2CO3 (Fig. 5b), which was high-energy milled alone, without Nb2O5, with exactly the same milling conditions as the mixture. While the peaks of Na2CO3 when milled together with Nb2O5 completely disappeared after 20 hours of milling (see open triangles in Fig. 5a), this is clearly not the case even after 40 hours if Na2CO3 was milled alone (see Fig. 5b). The broader peaks of Na2CO3 after 40 hours of separate milling (Fig. 5b, 40 h) are most probably a consequence of reduced crystallite size and increase in microstrains due to creation of structural disorder. The disappearance of the original crystalline Na2CO3 from the mixture, suggesting amorphization, is therefore an effect triggered by the presence of Nb2O5 rather than a pure effect of the high-energy collisions. In relation to this mechanochemical

Fig. 5. XRD patterns of a) Na2CO3–Nb2O5 mixture and b) Na2CO3 after high-energy milling for up to 40 hours. The pattern in a) shows a narrow 2-theta region, i.e., from 34.8 to 38.8°, to highlight the changes upon milling in the peaks corresponding to Na2CO3. Note that all the peaks on the patterns of non-milled and 40-hours-separately-milled Na2CO3 in b) are indexed with monoclinic Na2CO3. Notation: Na2CO3 (∆, PDF 19-1130); "h" denotes milling hours (from Rojac et al., 2006).

interaction between Na2CO3 and Nb2O5, a question that arises at this point is whether this interaction resulted into the carbonate decomposition. This is also relevant with respect to the nature of the amorphous phase. Obviously, further information could be obtained by following the decomposition of the carbonate during milling. This can be done using thermogravimetric (TG) analysis; the results of TG coupled with differential thermal analysis (DTA) and evolved-gas analysis (EGA) are presented in the following section.

#### **2.2 Thermal analysis**

20 Infrared Spectroscopy – Materials Science, Engineering and Technology

Fig. 4. Fractions of crystalline phases (Na2CO3, Nb2O5 and NaNbO3) and XRD-amorphous phase, determined by Rietveld refinement analysis, as a function of milling time. a) Na2CO3 and Nb2O5, b) NaNbO3 and XRD-amorphous phase. The inset of b) shows an enlarged view of the curves in the initial 80 hours of milling. The lines are drawn as a guide for the eye

0

0 100 200 300 400 500 600 700 800

Mass fraction (%)

0 20 40 60 80

Milling time (h)

NaNbO3

Milling time (h)

20

40

60

Mass fraction (%)

a) XRD-am. ph.

80

b)

Mass fraction (%)

100

From this simple observation we can infer that NaNbO3 is not formed directly, like assumed by equation 1, but through an intermediate amorphous phase. The transitional nature of the amorphous phase is further confirmed by the maximum in its amount after 110 hours of milling. Moreover, literature data go in favour of our conclusions. In fact, based on studies of the kinetics, the sigmoidal trend, like that observed in the case of NaNbO3 (Fig. 4b, open circles), is characteristic for multistep mechanochemical processes, such as the amorphization of a mixture of metals, where the phase transformation requires two or more impacts on the same powder fraction. In contrast, continuously decelerating processes, described by asymptotic kinetics, are typical for the amorphization of single-phase compounds, such as intermetallics, where the structure is already altered after the first impact (Delogu & Cocco, 2000; Cocco et al., 2000; Delogu et al., 2004). Therefore, independently of the analysis on the XRD-amorphous phase, the sigmoidal-like trend in the formation of NaNbO3 (Fig. 4b, open circles) suggests that the niobate is formed via a

In addition to the XRD-amorphous phase, we shall look at the changes induced in the Na2CO3 in the initial part of milling. Fig. 5 compares the XRD patterns of the Na2CO3–Nb2O5 mixture in the first 40 hours of milling (Fig. 5a) with the XRD patterns of Na2CO3 (Fig. 5b), which was high-energy milled alone, without Nb2O5, with exactly the same milling conditions as the mixture. While the peaks of Na2CO3 when milled together with Nb2O5 completely disappeared after 20 hours of milling (see open triangles in Fig. 5a), this is clearly not the case even after 40 hours if Na2CO3 was milled alone (see Fig. 5b). The broader peaks of Na2CO3 after 40 hours of separate milling (Fig. 5b, 40 h) are most probably a consequence of reduced crystallite size and increase in microstrains due to creation of structural disorder. The disappearance of the original crystalline Na2CO3 from the mixture, suggesting amorphization, is therefore an effect triggered by the presence of Nb2O5 rather than a pure effect of the high-energy collisions. In relation to this mechanochemical

(from Rojac et al., 2008b).

0 50 100 150 200 250 300

Nb2O5 Na2CO3

Milling time (h)

0

20

40

60

Mass fraction (%)

80

100

transitional phase.

In order to explore the origin of the reaction-induced amorphization and/or possible decomposition of Na2CO3 (Fig. 5) we were further focused on the initial part of milling, i.e., results are presented for the samples treated in the first 40 hours of milling.

Fig. 6 presents the thermogravimetric (TG), derivative thermogravimetric (DTG), differential thermal analysis (DTA) and evolved-gas analysis (EGA) curves of the Na2CO3–Nb2O5 powder mixture in the first 40 hours of high-energy milling. The nonmilled Na2CO3–Nb2O5 mixture looses mass in several steps in a broad temperature range from 400 °C to 800 °C (Fig. 6a and b, 0 h). The total mass loss of this mixture upon annealing to 900 °C amounts to 11.7%, which agrees well with the theoretical mass loss of 11.8%, calculated according to equation 1 for the complete decomposition of Na2CO3 in an

Using Infrared Spectroscopy to Identify New Amorphous Phases –

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 23

DTG (a.u.)

Fig. 6. a) TG, b) DTG, c) DTA and d) EGA(H2O, CO2) curves of the Na2CO3–Nb2O5 powder mixture after high-energy milling for 1, 5, 20 and 40 hours. The non-milled mixture is denoted as "0 h". Since the main EGA(H2O) signal was observed in the temperature range 25–350 °C the data are plotted accordingly. "h" denotes milling hours (from Rojac et al.,

EGA (a.u.)

0 100 200 300 400 500 600 700 800 900

H2O CO2

0 100 200 300 400 500 600 700 800 900

T (°C)

0 h 1 h

5 h

20 h 40 h

0 h

1 h

5 h

20 h

40 h

T (°C)

According to DTA, the decomposition of the carbonate in the milled samples is accompanied by an exothermic heat effect (Fig. 6c, milled samples). This is seen from the sharp and intense exothermic peaks appearing in all the milled samples in the temperature

To summarize, the DTA and EGA(CO2) analyses on the milled samples (Fig. 6c and d, milled samples) suggest a rather defined carbonate decomposition occurring in a narrow temperature range, which is not typical for a physical mixture of Na2CO3 and Nb2O5 (compare 0 h with milled samples in Fig. 6c and 6d; see also Jenko, 2006); this indicates a change in the chemical state of the carbonate upon milling and formation of a new phase.

According to the mass loss related to the CO2 release, which can be separated from the loss of H2O by combining EGA and TG curves, we can calculate the amount of the residual

range where the CO2 is released, i.e., 400–500 °C (compare Fig. 6c with Fig. 6d).

0 h

1 h 5 h

20 h

c) DTA d) EGA (H2O, CO2)

0 h

1 h

a) TG b) DTG

5 h

40 h exo

0 100 200 300 400 500 600 700 800 900

0 100 200 300 400 500 600 700 800 900

T (°C)

20 h 40 h

T (°C)

2006).

DTA (a.u.)

84

88

92

TG (%)

96

100

equimolar mixture with Nb2O5. The carbonate decomposition is further confirmed by EGA, which shows a release of CO2 in the temperature range 400–800 °C (Fig. 6d, 0 h, full line). Note also that the DTG peaks (Fig. 6b, 0 h) coincide with the EGA(CO2) peaks (Fig. 6d, 0 h, full line), showing that the measured mass loss in this sample is indeed entirely related to the decomposition of Na2CO3, which is triggered by the reaction with Nb2O5, like represented by equation 1.

High-energy milling resulted into several changes in the thermal behaviour of the Na2CO3–Nb2O5 mixture. Firstly, by inspecting the TG curves, a mass loss appears in the milled samples in the temperature range 25–300 °C, which was not observed prior milling (Fig. 6a, compare milled samples with the non-milled). According to the DTA curves (Fig. 6c), these mass losses between room temperature and 300 °C are accompanied by endothermic heat effects, which first manifest as a sharp endothermic peak at around 100 °C (Fig. 6c, 1 h), progressively evolving with milling into a broader endothermic peak, which expands from 80 °C to 250 °C (see for example Fig. 6c, 40 h). According to EGA(H2O), the mass losses in this low temperature range correspond to the removal of H2O (Fig. 6d, milled samples, dashed lines). The amounts of H2O removed from the samples milled for 0, 1, 5, 20, 40 hours, as determined from the TG curves (Fig. 6a, milled samples, 25–300 °C), are 0%, 2.5%, 4.0%, 4.8% and 5.1%, respectively. This suggests gradual adsorption of H2O on the powder with increasing milling time; taking into account that the milling was performed in open air and also considering the hygroscopic nature of Na2CO3, the adsorption of H2O is not surprising. We note that the H2O removal from the samples milled for longer periods, i.e., 5, 20 and 40 hours, takes place at temperatures higher than 100 °C (Fig. 6d, dashed lines), which might suggest water chemisorption rather than physical adsorption.

In addition to water adsorption, high-energy milling induced considerable changes in the thermal decomposition of the carbonate. This is best seen by inspecting the DTG and EGA (CO2) curves of the milled samples (Fig. 6b and 6d, milled samples). Firstly, it should be noted that in the temperature range between 350 °C and 500 °C the DTG peaks of the milled mixtures (Fig. 6b, milled samples) coincide with those of EGA(CO2) (Fig. 6d, milled samples, full lines), which means that the mass loss in this temperature range is related to the CO2 removal, i.e., to the carbonate decomposition. For the sake of discussion, we consider in the following only the EGA(CO2) curves (Fig. 6d, full lines). In contrast to the carbonate decomposition in the non-milled mixture (Fig. 6d, 0 h, 400–800 °C), occurring in several steps and in a broad temperature range, which is characteristic for a physical mixture of Na2CO3 and Nb2O5 particles (Jenko, 2006), the mixture milled for only 1 hour releases CO2 in a much narrower temperature range, i.e., 400–500 °C (Fig. 6d, 1 h). We attribute this effect to the smaller particle size after 1 hour of milling, which is known to decrease considerably the decomposition temperature of Na2CO3 in the Na2CO3–Nb2O5 mixture due to reduced diffusion paths (Jenko, 2006). In comparison with the 1-hour milled sample, upon milling for 5 hours only small changes are observed in the shape of the EGA(CO2) peak (Fig. 6d, 5 h, 400–500 °C). After 20 hours of milling a new, weak EGA(CO2) peak appears at 370 °C (Fig. 6d, 20 h), suggesting two-step carbonate decomposition; this peak then shifts to 400 °C upon 40 hours of milling (Fig. 6d, 40 h). Note that after 40 hours of milling the intense EGA(CO2) peak at 420 °C becomes sharper in comparison with shorter milling times, i.e., 1, 5 and 20 hours, indicating a more uniform decomposition of the carbonate.

equimolar mixture with Nb2O5. The carbonate decomposition is further confirmed by EGA, which shows a release of CO2 in the temperature range 400–800 °C (Fig. 6d, 0 h, full line). Note also that the DTG peaks (Fig. 6b, 0 h) coincide with the EGA(CO2) peaks (Fig. 6d, 0 h, full line), showing that the measured mass loss in this sample is indeed entirely related to the decomposition of Na2CO3, which is triggered by the reaction with Nb2O5,

High-energy milling resulted into several changes in the thermal behaviour of the Na2CO3–Nb2O5 mixture. Firstly, by inspecting the TG curves, a mass loss appears in the milled samples in the temperature range 25–300 °C, which was not observed prior milling (Fig. 6a, compare milled samples with the non-milled). According to the DTA curves (Fig. 6c), these mass losses between room temperature and 300 °C are accompanied by endothermic heat effects, which first manifest as a sharp endothermic peak at around 100 °C (Fig. 6c, 1 h), progressively evolving with milling into a broader endothermic peak, which expands from 80 °C to 250 °C (see for example Fig. 6c, 40 h). According to EGA(H2O), the mass losses in this low temperature range correspond to the removal of H2O (Fig. 6d, milled samples, dashed lines). The amounts of H2O removed from the samples milled for 0, 1, 5, 20, 40 hours, as determined from the TG curves (Fig. 6a, milled samples, 25–300 °C), are 0%, 2.5%, 4.0%, 4.8% and 5.1%, respectively. This suggests gradual adsorption of H2O on the powder with increasing milling time; taking into account that the milling was performed in open air and also considering the hygroscopic nature of Na2CO3, the adsorption of H2O is not surprising. We note that the H2O removal from the samples milled for longer periods, i.e., 5, 20 and 40 hours, takes place at temperatures higher than 100 °C (Fig. 6d, dashed lines), which might suggest water

In addition to water adsorption, high-energy milling induced considerable changes in the thermal decomposition of the carbonate. This is best seen by inspecting the DTG and EGA (CO2) curves of the milled samples (Fig. 6b and 6d, milled samples). Firstly, it should be noted that in the temperature range between 350 °C and 500 °C the DTG peaks of the milled mixtures (Fig. 6b, milled samples) coincide with those of EGA(CO2) (Fig. 6d, milled samples, full lines), which means that the mass loss in this temperature range is related to the CO2 removal, i.e., to the carbonate decomposition. For the sake of discussion, we consider in the following only the EGA(CO2) curves (Fig. 6d, full lines). In contrast to the carbonate decomposition in the non-milled mixture (Fig. 6d, 0 h, 400–800 °C), occurring in several steps and in a broad temperature range, which is characteristic for a physical mixture of Na2CO3 and Nb2O5 particles (Jenko, 2006), the mixture milled for only 1 hour releases CO2 in a much narrower temperature range, i.e., 400–500 °C (Fig. 6d, 1 h). We attribute this effect to the smaller particle size after 1 hour of milling, which is known to decrease considerably the decomposition temperature of Na2CO3 in the Na2CO3–Nb2O5 mixture due to reduced diffusion paths (Jenko, 2006). In comparison with the 1-hour milled sample, upon milling for 5 hours only small changes are observed in the shape of the EGA(CO2) peak (Fig. 6d, 5 h, 400–500 °C). After 20 hours of milling a new, weak EGA(CO2) peak appears at 370 °C (Fig. 6d, 20 h), suggesting two-step carbonate decomposition; this peak then shifts to 400 °C upon 40 hours of milling (Fig. 6d, 40 h). Note that after 40 hours of milling the intense EGA(CO2) peak at 420 °C becomes sharper in comparison with shorter milling times, i.e., 1, 5 and 20 hours, indicating a more

like represented by equation 1.

chemisorption rather than physical adsorption.

uniform decomposition of the carbonate.

Fig. 6. a) TG, b) DTG, c) DTA and d) EGA(H2O, CO2) curves of the Na2CO3–Nb2O5 powder mixture after high-energy milling for 1, 5, 20 and 40 hours. The non-milled mixture is denoted as "0 h". Since the main EGA(H2O) signal was observed in the temperature range 25–350 °C the data are plotted accordingly. "h" denotes milling hours (from Rojac et al., 2006).

According to DTA, the decomposition of the carbonate in the milled samples is accompanied by an exothermic heat effect (Fig. 6c, milled samples). This is seen from the sharp and intense exothermic peaks appearing in all the milled samples in the temperature range where the CO2 is released, i.e., 400–500 °C (compare Fig. 6c with Fig. 6d).

To summarize, the DTA and EGA(CO2) analyses on the milled samples (Fig. 6c and d, milled samples) suggest a rather defined carbonate decomposition occurring in a narrow temperature range, which is not typical for a physical mixture of Na2CO3 and Nb2O5 (compare 0 h with milled samples in Fig. 6c and 6d; see also Jenko, 2006); this indicates a change in the chemical state of the carbonate upon milling and formation of a new phase.

According to the mass loss related to the CO2 release, which can be separated from the loss of H2O by combining EGA and TG curves, we can calculate the amount of the residual

Using Infrared Spectroscopy to Identify New Amorphous Phases –

complex IR spectrum (Buijs & Schutte, 1961; Brooker & Bates, 1971).

vibrational bands of the CO3

The CO3

of CO3

spectrum of Nb2O5 is not shown). The IR vibrations of the free CO3

(Venyaminov & Prendergast, 1997).

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 25

Gatehouse et al., 1958). This is consistent with the fact that Nb2O5, which is also a part of the mixture, did not show any IR bands in the two examined wavenumber regions (the IR

symmetrical C–O stretching vibration, denoted as 1, is IR-inactive, while the 2, 3 and <sup>4</sup> are IR-active. According to Harris & Salje (1992), and Table 1, the strongest band of the nonmilled sample at 1445 cm–1 (Fig. 7a, 0 h) belongs to the assymetrical C–O stretching vibration

2– ion possesses two stretching and two bending vibrational modes. The

2– (3), while the weak band at 1775 cm–1 can be assigned to the combinational band of the type 1+4. No bands are observed in the 950–1150 cm–1 region (Fig. 7a, 0 h), consistent with absence of the IR-inactive1 vibration. With the exception of some differences in the position, the bands of the non-milled mixture, which belong to Na2CO3, are consistent with vibrations characteristic for the free CO32– ion with *D*3h symmetry. This is in agreement with the literature data and was explained as being a consequence of the small effect of the crystal field of Na+ ions on the symmetry of the CO32– in the Na2CO3 structure. This is somewhat different, for example, in Li2CO3, where a stronger interaction between crystal lattice and CO32– ions leads to lowered CO32– symmetry and, consequently, to a more

**Type of vibration Notation Wavenumber (cm–1)** 

C–O symmetrical stretching 1 (A1') 1063 Out-of-plane CO32– bending 2 (A2'') 879 C–O asymmetrical stretching 3 (E') 1415 In-plane CO32– bending 4 (E') 680 Table 1. Fundamental IR vibrations of carbonate (CO32–) ion with *D*3h symmetry. 2, 3 and <sup>4</sup> are IR-active vibrations, while 1 is IR-inactive (Gatehouse et al., 1958; Nakamoto, 1997).

Upon milling the Na2CO3–Nb2O5 mixture, considerable changes can be observed in the IR spectra (Fig. 7a, milled samples). After 1 hour of milling a new weak band appears at 1650 cm–1. The position of this band coincides with one of the strongest HCO3– bands typical for alkaline hydrogencarbonates (Watters, 2005). This is in agreement with the simultaneous loss of H2O and CO2 upon annealing this sample (Fig. 6d, 1 h), which is characteristic for the hydrogencarbonate decomposition. Furthermore, we should not eliminate the possibility of having the in-plane bending vibration of H2O, which also appears near 1650 cm–1

By further milling from 1 hour to 40 hours related and simultaneous trends can be noted: i) the 3(CO32–) vibration shifts from 1445 cm–1 (Fig. 7a, 1 and 5 h) to 1455 cm–1 (Fig. 7a, 20 h) and decreases in intensity until it completely disappears after 40 h of milling, ii) the <sup>3</sup> vibration is gradually replaced by new absorption bands appearing at 1605, 1530 and 1345 cm–1 (Fig. 7a, 40 h), and iii) a new band arises during milling, located at 1055 cm–1, which belongs to the symmetrical C–O stretching vibration of the CO32– ions (1) (Fig. 7a, see region 950–1150 cm–1). We can conclude from these results that milling induced a splitting of 3 and activation of 1 vibrations, suggesting a change of the CO32– symmetry from the

2– ions, present in the initial Na2CO3 (Harris & Salje, 1992;

2– ion having *D*3h point group symmetry are listed in Table 1.

carbonate in the mixture, i.e., the amount of the carbonate that did not decompose during high-energy milling. The total CO2 loss from the sample milled for 40 hours is 9.6%, corresponding to 85.0% of residual carbonate. Therefore, in the first 40 hours of milling, a minor amount of the carbonate decomposed, whereas the major part, according to XRD analysis (Fig. 5a), became amorphous. As mentioned in the previous section, the Na2CO3 amorphization is stimulated by the mechanochemical interaction with Nb2O5. This observation, together with the characteristic changes in the decomposition of the carbonate upon milling (Fig. 6), indicates a formation of a new carbonate compound. As a next step, it seems reasonable to explore the symmetry of the CO3 2– ions, which was done using infrared spectroscopy.

#### **2.3 Infrared spectroscopy analysis**

The IR spectra of the Na2CO3–Nb2O5 mixture before and after milling for various periods are shown in Fig. 7a. The two separate graphs in Fig. 7a show two different wavenumber regions, i.e., 950–1150 cm–1 and 1280–1880 cm–1. The spectrum of the non-milled mixture is composed of a weak band at 1775 cm–1 and a strong one at 1445 cm–1; no bands are observed in the lower wavenumber region between 950 and 1150 cm–1 (Fig. 7a, 0 h). Based on the literature data, the spectrum of the non-milled mixture can be entirely indexed with

Fig. 7. FT-IR spectra of a) Na2CO3–Nb2O5 powder mixture after high-energy milling for 1, 5, 20 and 40 hours and (b) Na2CO3 subjected to separate high-energy milling for 40 hours. The non-milled powders are denoted as "0 h". Note that, in contrast to the Na2CO3–Nb2O5 mixture (a), no splitting of 3(CO32–) is observed in the case of the separately milled Na2CO3 (b). Notation: \* Nujol, for bands assignment refer to Table 1; "h" denotes milling hours. (from Rojac et al., 2006).

carbonate in the mixture, i.e., the amount of the carbonate that did not decompose during high-energy milling. The total CO2 loss from the sample milled for 40 hours is 9.6%, corresponding to 85.0% of residual carbonate. Therefore, in the first 40 hours of milling, a minor amount of the carbonate decomposed, whereas the major part, according to XRD analysis (Fig. 5a), became amorphous. As mentioned in the previous section, the Na2CO3 amorphization is stimulated by the mechanochemical interaction with Nb2O5. This observation, together with the characteristic changes in the decomposition of the carbonate upon milling (Fig. 6), indicates a formation of a new carbonate compound. As a next step, it seems reasonable to explore the symmetry of the CO32– ions, which was done using infrared

The IR spectra of the Na2CO3–Nb2O5 mixture before and after milling for various periods are shown in Fig. 7a. The two separate graphs in Fig. 7a show two different wavenumber regions, i.e., 950–1150 cm–1 and 1280–1880 cm–1. The spectrum of the non-milled mixture is composed of a weak band at 1775 cm–1 and a strong one at 1445 cm–1; no bands are observed in the lower wavenumber region between 950 and 1150 cm–1 (Fig. 7a, 0 h). Based on the literature data, the spectrum of the non-milled mixture can be entirely indexed with

Fig. 7. FT-IR spectra of a) Na2CO3–Nb2O5 powder mixture after high-energy milling for 1, 5, 20 and 40 hours and (b) Na2CO3 subjected to separate high-energy milling for 40 hours. The non-milled powders are denoted as "0 h". Note that, in contrast to the Na2CO3–Nb2O5 mixture (a), no splitting of 3(CO32–) is observed in the case of the separately milled Na2CO3 (b). Notation: \* Nujol, for bands assignment refer to Table 1; "h" denotes milling hours.

spectroscopy.

(from Rojac et al., 2006).

**2.3 Infrared spectroscopy analysis** 

vibrational bands of the CO3 2– ions, present in the initial Na2CO3 (Harris & Salje, 1992; Gatehouse et al., 1958). This is consistent with the fact that Nb2O5, which is also a part of the mixture, did not show any IR bands in the two examined wavenumber regions (the IR spectrum of Nb2O5 is not shown).

The IR vibrations of the free CO3 2– ion having *D*3h point group symmetry are listed in Table 1. The CO3 2– ion possesses two stretching and two bending vibrational modes. The symmetrical C–O stretching vibration, denoted as 1, is IR-inactive, while the 2, 3 and <sup>4</sup> are IR-active. According to Harris & Salje (1992), and Table 1, the strongest band of the nonmilled sample at 1445 cm–1 (Fig. 7a, 0 h) belongs to the assymetrical C–O stretching vibration of CO3 2– (3), while the weak band at 1775 cm–1 can be assigned to the combinational band of the type 1+4. No bands are observed in the 950–1150 cm–1 region (Fig. 7a, 0 h), consistent with absence of the IR-inactive1 vibration. With the exception of some differences in the position, the bands of the non-milled mixture, which belong to Na2CO3, are consistent with vibrations characteristic for the free CO32– ion with *D*3h symmetry. This is in agreement with the literature data and was explained as being a consequence of the small effect of the crystal field of Na+ ions on the symmetry of the CO32– in the Na2CO3 structure. This is somewhat different, for example, in Li2CO3, where a stronger interaction between crystal lattice and CO32– ions leads to lowered CO32– symmetry and, consequently, to a more complex IR spectrum (Buijs & Schutte, 1961; Brooker & Bates, 1971).


Table 1. Fundamental IR vibrations of carbonate (CO32–) ion with *D*3h symmetry. 2, 3 and <sup>4</sup> are IR-active vibrations, while 1 is IR-inactive (Gatehouse et al., 1958; Nakamoto, 1997).

Upon milling the Na2CO3–Nb2O5 mixture, considerable changes can be observed in the IR spectra (Fig. 7a, milled samples). After 1 hour of milling a new weak band appears at 1650 cm–1. The position of this band coincides with one of the strongest HCO3 – bands typical for alkaline hydrogencarbonates (Watters, 2005). This is in agreement with the simultaneous loss of H2O and CO2 upon annealing this sample (Fig. 6d, 1 h), which is characteristic for the hydrogencarbonate decomposition. Furthermore, we should not eliminate the possibility of having the in-plane bending vibration of H2O, which also appears near 1650 cm–1 (Venyaminov & Prendergast, 1997).

By further milling from 1 hour to 40 hours related and simultaneous trends can be noted: i) the 3(CO32–) vibration shifts from 1445 cm–1 (Fig. 7a, 1 and 5 h) to 1455 cm–1 (Fig. 7a, 20 h) and decreases in intensity until it completely disappears after 40 h of milling, ii) the <sup>3</sup> vibration is gradually replaced by new absorption bands appearing at 1605, 1530 and 1345 cm–1 (Fig. 7a, 40 h), and iii) a new band arises during milling, located at 1055 cm–1, which belongs to the symmetrical C–O stretching vibration of the CO3 2– ions (1) (Fig. 7a, see region 950–1150 cm–1). We can conclude from these results that milling induced a splitting of 3 and activation of 1 vibrations, suggesting a change of the CO32– symmetry from the

Using Infrared Spectroscopy to Identify New Amorphous Phases –

also apply for this case (Fujita et al., 1962; Nakamoto, 1997).

to the C–O bond in the free, non-coordinated, CO3

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 27

symmetry; this comes from the equivalence of the three C–O bonds, which is schematically illustrated in Fig. 9 (bottom-left quadrant). The equivalence of these three C–O bonds is lost upon coordination, so that typically the C–O bond coordinated to the metal cation becomes weaker, while the C–O bonds not involved in metal binding becomes stronger with respect

et al., 1962; Brintzinger & Hester, 1966). This in turn leads to lowered CO32– symmetry, e.g., from *D*3h to *C*<sup>2</sup>, and to the activation of the 1 vibration (Fig. 9, bottom-right quadrant). In the case of monodentate coordination, also the *C*s symmetry is possible and arises when the M–O–C bond is not collinear (Fig. 9, upper-right quadrant); same IR spectroscopic changes

Fig. 9. Schematic representation of non-coordinated and coordinated CO32– ion and the corresponding point group symmetry elements. The changes in the 1 and 3 IR vibrations of the CO32– ion upon coordination are also shown. For simplicity, only monodentate

planes perpendicular and parallel to the principal axis, respectively, *S*n – n-fold rotationreflection operation. The number preceding the symmetry operation symbol refers to number of such symmetry elements that the molecule possesses. For further details consult

In parallel with the 1 activation, also the splitting of the 3 vibration occurs upon

quadrant). Doubly degenerate vibrations occur only in molecules possessing an axis higher than twofold, which is the case of the *D*3h symmetry, having a three-fold rotational axis (see

2– ion, the 3 vibration is doubly degenerate (Fig. 9, bottom-left

coordination is presented. Notations: *I* – identity, *C*n – n-fold axis of rotation,

Nakamoto, 1997.

coordination. In the free CO3

2– ion (Fig. 9, upper-right quadrant) (Fujita

h, 

– mirror

original *D*3h. We shall come back to this point after examining the fundamental relation between symmetry and IR vibrations of the carbonate ion.

An extensive review on the IR spectroscopic identification of different species arising from the reactive adsorption of CO2 on metal oxide surfaces can be found in Busca & Lorenzelli, 1982. In principle, the carbonate ion is a highly versatile ligand, which gives rise not only to simple mono- or bidentate structures, but also to a number of more complicated bidentate bridged structures. Some examples of CO3 2– coordinated configurations are schematically illustrated in Fig. 8.

Fig. 8. Schematic view of free (non-coordinated) and various types of coordinated CO32– ions.

When the CO32– ion is bound, through one or more of its oxygens, to a metal cation (denoted as "M" in Fig. 8), its point group symmetry is lowered. It is well known from the literature that the lowering of the CO32– symmetry, resulting from the coordination of the carbonate ion in a carbonato complex, causes the following changes in the IR vibrational modes of the free carbonate ion (Gatehouse et al., 1958; Hester & Grossman, 1966; Brintzinger & Hester, 1966; Goldsmith & Ross, 1967; Jolivet et al., 1980; Busca & Lorenzelli, 1982; Nakamoto, 1997):


The most characteristic of the above IR spectroscopic changes upon CO3 2– coordination is the infrared activation of the 1, i.e., the symmetrical C–O stretching vibration. This vibration, as mentioned earlier, is IR-inactive for the free carbonate ion, but also for most alkali, alkaline-earth and heavy-metal carbonates; it appears as a weak band only in certain carbonates of the aragonite type (Gatehouse et al., 1958). To derive the relation between symmetry and IR vibrations, we shall first look at the details of the 1 vibration. According to the IR selection rule, which states that *the vibration is IR-active if the dipole moment is changed during vibration*, we can understand that there will be no net change in the dipole moment during symmetrical C–O stretching vibration (1) of the CO3 2– ion with *D*3h

original *D*3h. We shall come back to this point after examining the fundamental relation

An extensive review on the IR spectroscopic identification of different species arising from the reactive adsorption of CO2 on metal oxide surfaces can be found in Busca & Lorenzelli, 1982. In principle, the carbonate ion is a highly versatile ligand, which gives rise not only to simple mono- or bidentate structures, but also to a number of more complicated bidentate

Fig. 8. Schematic view of free (non-coordinated) and various types of coordinated CO32–

bidentate chelate bidentate bridged

**M M M**

free carbonate ion monodentate

When the CO32– ion is bound, through one or more of its oxygens, to a metal cation (denoted as "M" in Fig. 8), its point group symmetry is lowered. It is well known from the literature

ion in a carbonato complex, causes the following changes in the IR vibrational modes of the free carbonate ion (Gatehouse et al., 1958; Hester & Grossman, 1966; Brintzinger & Hester, 1966; Goldsmith & Ross, 1967; Jolivet et al., 1980; Busca & Lorenzelli, 1982; Nakamoto, 1997):

The most characteristic of the above IR spectroscopic changes upon CO32– coordination is the infrared activation of the 1, i.e., the symmetrical C–O stretching vibration. This vibration, as mentioned earlier, is IR-inactive for the free carbonate ion, but also for most alkali, alkaline-earth and heavy-metal carbonates; it appears as a weak band only in certain carbonates of the aragonite type (Gatehouse et al., 1958). To derive the relation between symmetry and IR vibrations, we shall first look at the details of the 1 vibration. According to the IR selection rule, which states that *the vibration is IR-active if the dipole moment is changed during vibration*, we can understand that there will be no net change in the dipole moment during symmetrical C–O stretching vibration (1) of the CO32– ion with *D*3h

2– symmetry, resulting from the coordination of the carbonate

**M M**

**M**

**M**

2– coordinated configurations are schematically

between symmetry and IR vibrations of the carbonate ion.

bridged structures. Some examples of CO3

illustrated in Fig. 8.

ions.

that the lowering of the CO3

2. Shift of 2 vibration 3. Splitting of 3 vibration 4. Splitting of 4 vibration

1. Activation of IR-inactive 1 vibration

symmetry; this comes from the equivalence of the three C–O bonds, which is schematically illustrated in Fig. 9 (bottom-left quadrant). The equivalence of these three C–O bonds is lost upon coordination, so that typically the C–O bond coordinated to the metal cation becomes weaker, while the C–O bonds not involved in metal binding becomes stronger with respect to the C–O bond in the free, non-coordinated, CO3 2– ion (Fig. 9, upper-right quadrant) (Fujita et al., 1962; Brintzinger & Hester, 1966). This in turn leads to lowered CO32– symmetry, e.g., from *D*3h to *C*<sup>2</sup>, and to the activation of the 1 vibration (Fig. 9, bottom-right quadrant). In the case of monodentate coordination, also the *C*s symmetry is possible and arises when the M–O–C bond is not collinear (Fig. 9, upper-right quadrant); same IR spectroscopic changes also apply for this case (Fujita et al., 1962; Nakamoto, 1997).

Fig. 9. Schematic representation of non-coordinated and coordinated CO32– ion and the corresponding point group symmetry elements. The changes in the 1 and 3 IR vibrations of the CO3 2– ion upon coordination are also shown. For simplicity, only monodentate coordination is presented. Notations: *I* – identity, *C*n – n-fold axis of rotation, h, – mirror planes perpendicular and parallel to the principal axis, respectively, *S*n – n-fold rotationreflection operation. The number preceding the symmetry operation symbol refers to number of such symmetry elements that the molecule possesses. For further details consult Nakamoto, 1997.

In parallel with the 1 activation, also the splitting of the 3 vibration occurs upon coordination. In the free CO32– ion, the 3 vibration is doubly degenerate (Fig. 9, bottom-left quadrant). Doubly degenerate vibrations occur only in molecules possessing an axis higher than twofold, which is the case of the *D*3h symmetry, having a three-fold rotational axis (see

Using Infrared Spectroscopy to Identify New Amorphous Phases –

observed during separate milling of Na2CO3 (see Fig. 5b and Fig. 7b).

lowering of the CO3

al., to be published).

as central cation.

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 29

induced by milling, the 3 band at 1445 cm–1 is still present after 40 hours of separate milling. We emphasize that for this separate Na2CO3 milling the same milling conditions and same milling time, i.e., 40 hours, were applied as for the Na2CO3–Nb2O5 mixture. Therefore, the

only be explained by the presence of Nb2O5 or, in other words, by the participation of Nb5+

The mechanochemical formation of the carbonato complex is further supported by XRD analysis. By comparing XRD and IR data, we find that the 3 splitting and 1 activation, which took place progressively from 5 to 40 hours of milling (Fig. 7a), coincide with the amorphization of Na2CO3 (Fig. 5a). For example, after 20 hours of milling, when the split <sup>3</sup> bands are resolved for the first time and intense 1 band appeared (Fig. 7a, 20 h), the XRD peaks of Na2CO3 could not be detected anymore (Fig. 5a, 20 h). From this comparison we can conclude that the amorphization of Na2CO3 is closely related to the formation of the complex. The conclusion seems reasonable if we consider that the formation of the complex requires a reconstruction, i.e., coordination, of the CO32– ions; such reconstruction can eventually ruin the original Na2CO3 structure over the long range, make it undetectable to X-ray diffraction. The relation between amorphization and coordination is further supported by the fact that neither the amorphization of Na2CO3 nor the 3 splitting were

We finally note that after 40 hours of milling the powder mixture contains 81% of XRDamorphous phase (inset of Fig. 4b). According to this large amount and based on the fact that we did not detect any new crystalline phase during 40 hours of milling we can conclude that the carbonato complex is amorphous or eventually nanocrystalline to an extent that is undetectable with X-ray diffraction methods. This example illustrates that enriched information on a local structural scale can only be achieved by appropriate selection of analytical tools. The amorphous carbonato complex has recently been confirmed using Raman and nuclear magnetic resonance (NMR) spectroscopies (Rojac et

Another important aspect to discuss is the possible role of water on of the formation of the complex. Jolivet et al. (1982) emphasized the influence of the water molecules on the <sup>3</sup> splitting, which can be significant, depending on whether they can interact via hydrogen bonding with the carbonate group. This was demonstrated through various examples of lanthanide carbonates, where the hydrated forms showed different 3 splitting with respect to their dehydrated analogues. As an example, the hydrated form of the Na2Cu(CO3)2 complex, that is Na2Cu(CO3)23H2O, showed larger 3 splitting, i.e., 203 cm–1, in comparison with its dehydrated form, i.e., 138 cm–1 (see also Table 2, first two examples).

In our case, the possible influence of water molecules on the carbonate ion should be considered. In fact, we showed in section 2.2 (Fig. 6) that an amount of water was introduced in the sample from the air during the milling. Therefore, we have to examine more carefully the possible influence of water adsorption on the 3 splitting. This was done by quenching the 40-hours-treated sample in air from different temperatures so that controlled amounts of water were released; the quenched samples were then analyzed using IR spectroscopy. The IR spectra together with the TG and EGA curves are shown in Fig. 10. The mass losses after quenching at 100 °C, 170 °C and 300 °C were 0.8 %, 3.1 % and 5.0 %,

2– symmetry and the corresponding coordination of the CO3

2– ions can

Fig. 9, upper-left quadrant; *C*3 – three-fold axis) (Nakamoto, 1997). The lowering of the symmetry of the carbonate ion from *D*3h to either *C*<sup>2</sup> or *C*s, which means loss of the equivalence of the three C–O bonds in the CO32– and, therefore, loss of the three-fold rotational axis, leads to the separation (splitting) of the doubly degenerate vibrations (Fig. 9, bottom-right quadrant).

With respect to the changes upon coordination in the other two vibrational modes, i.e., <sup>2</sup> and 4, it should be noted that the splitting of the 4 vibration has been studied to a lesser extent in comparison with the characteristic 3 splitting. In addition, the shift of the <sup>2</sup> vibration upon coordination is typically small and the values for complexes do not differ greatly in comparison with those of simple carbonates (Gatehouse et al., 1958).

Coming back to our case from Fig. 7a, we can interpret the 3 splitting and 1 activation of the CO3 2– vibrations during milling of the Na2CO3–Nb2O5 mixture as characteristic of lowered CO32– symmetry, which is related to the mechanochemical formation of a carbonato complex. For comparison, we compiled in Table 2 the data of a number of carbonato complexes, with metals such as Cu and Co, provided from the literature. According to the notation of the coordinated *C*<sup>2</sup> symmetry, the 3 vibration of the *D*3h symmetry is now split into two components, which are denoted as 1 and 4 (second and third column in Table 2); the activated 1 vibration becomes 2 (fourth column in Table 2). The regions in which the carbonato complex absorption bands appear are 1623-1500 cm–1, 1362-1265 cm–1 (3 splitting) and 1080-1026 cm–1 (1 activation) (Table 2). By comparing these data with the our case, we can observe that the 3 split bands at 1605, 1530 and 1345 cm–1, and the 1 activated band at 1055 cm–1 from Fig. 7a (40 h) fall entirely within the wavenumber regions of carbonato complexes from Table 2.


Table 2. Some copper and cobalt carbonato complexes and the corresponding IR absorption bands related to CO32– vibrations. 1, 2 and 4 correspond to vibrations of CO3 2– in the *C*<sup>2</sup> symmetry notation; according to this notation, the doubly degenerate 3 vibration of the free CO32– ion, which splits into two components, is denoted as 1 and 4, whereas the activated 1 vibration is denoted as 2 (data compiled from Gatehouse et al., 1958; Fujita et al., 1962; Jolivet et al., 1982; Healy & White, 1972).

It is important to note that, in contrast to the Na2CO3–Nb2O5 mixture, the 3 vibration did not split when Na2CO3 was milled alone, i.e., without Nb2O5. This is seen in Fig. 7b, which shows the IR spectra of Na2CO3 before and after separate milling. Except for the reduced intensity, which might be related to the decreased crystallite size and structural disordering

Fig. 9, upper-left quadrant; *C*3 – three-fold axis) (Nakamoto, 1997). The lowering of the symmetry of the carbonate ion from *D*3h to either *C*<sup>2</sup> or *C*s, which means loss of the equivalence of the three C–O bonds in the CO32– and, therefore, loss of the three-fold rotational axis, leads to the separation (splitting) of the doubly degenerate vibrations (Fig. 9,

With respect to the changes upon coordination in the other two vibrational modes, i.e., <sup>2</sup> and 4, it should be noted that the splitting of the 4 vibration has been studied to a lesser extent in comparison with the characteristic 3 splitting. In addition, the shift of the <sup>2</sup> vibration upon coordination is typically small and the values for complexes do not differ

Coming back to our case from Fig. 7a, we can interpret the 3 splitting and 1 activation of the CO32– vibrations during milling of the Na2CO3–Nb2O5 mixture as characteristic of lowered CO32– symmetry, which is related to the mechanochemical formation of a carbonato complex. For comparison, we compiled in Table 2 the data of a number of carbonato complexes, with metals such as Cu and Co, provided from the literature. According to the notation of the coordinated *C*<sup>2</sup> symmetry, the 3 vibration of the *D*3h symmetry is now split into two components, which are denoted as 1 and 4 (second and third column in Table 2); the activated 1 vibration becomes 2 (fourth column in Table 2). The regions in which the carbonato complex absorption bands appear are 1623-1500 cm–1, 1362-1265 cm–1 (3 splitting) and 1080-1026 cm–1 (1 activation) (Table 2). By comparing these data with the our case, we can observe that the 3 split bands at 1605, 1530 and 1345 cm–1, and the 1 activated band at 1055 cm–1 from Fig. 7a (40 h) fall entirely within the wavenumber regions of carbonato

**Carbonato complex 4(B2) (cm–1) 1(A1) (cm–1) 2(A1) (cm–1)**  Na2Cu(CO3)2 1500 1362 1058 Na2Cu(CO3)23H2O 1529 1326 1066, 1050 K3Co(CO3)33H2O 1527 1330 1080, 1037 KCo(NH3)2(CO3)2 1623, 1597 1265 1026 Co(NH3)6Co(CO3)3 1523 1285 1073, 1031 Co(NH3)4CO3Cl 1593 1265 1030 Co(NH3)4CO3ClO4 1602 1284 not reported

Table 2. Some copper and cobalt carbonato complexes and the corresponding IR absorption bands related to CO32– vibrations. 1, 2 and 4 correspond to vibrations of CO32– in the *C*<sup>2</sup> symmetry notation; according to this notation, the doubly degenerate 3 vibration of the free CO32– ion, which splits into two components, is denoted as 1 and 4, whereas the activated 1 vibration is denoted as 2 (data compiled from Gatehouse et al., 1958; Fujita et al., 1962;

It is important to note that, in contrast to the Na2CO3–Nb2O5 mixture, the 3 vibration did not split when Na2CO3 was milled alone, i.e., without Nb2O5. This is seen in Fig. 7b, which shows the IR spectra of Na2CO3 before and after separate milling. Except for the reduced intensity, which might be related to the decreased crystallite size and structural disordering

greatly in comparison with those of simple carbonates (Gatehouse et al., 1958).

bottom-right quadrant).

complexes from Table 2.

Jolivet et al., 1982; Healy & White, 1972).

induced by milling, the 3 band at 1445 cm–1 is still present after 40 hours of separate milling. We emphasize that for this separate Na2CO3 milling the same milling conditions and same milling time, i.e., 40 hours, were applied as for the Na2CO3–Nb2O5 mixture. Therefore, the lowering of the CO3 2– symmetry and the corresponding coordination of the CO3 2– ions can only be explained by the presence of Nb2O5 or, in other words, by the participation of Nb5+ as central cation.

The mechanochemical formation of the carbonato complex is further supported by XRD analysis. By comparing XRD and IR data, we find that the 3 splitting and 1 activation, which took place progressively from 5 to 40 hours of milling (Fig. 7a), coincide with the amorphization of Na2CO3 (Fig. 5a). For example, after 20 hours of milling, when the split <sup>3</sup> bands are resolved for the first time and intense 1 band appeared (Fig. 7a, 20 h), the XRD peaks of Na2CO3 could not be detected anymore (Fig. 5a, 20 h). From this comparison we can conclude that the amorphization of Na2CO3 is closely related to the formation of the complex. The conclusion seems reasonable if we consider that the formation of the complex requires a reconstruction, i.e., coordination, of the CO32– ions; such reconstruction can eventually ruin the original Na2CO3 structure over the long range, make it undetectable to X-ray diffraction. The relation between amorphization and coordination is further supported by the fact that neither the amorphization of Na2CO3 nor the 3 splitting were observed during separate milling of Na2CO3 (see Fig. 5b and Fig. 7b).

We finally note that after 40 hours of milling the powder mixture contains 81% of XRDamorphous phase (inset of Fig. 4b). According to this large amount and based on the fact that we did not detect any new crystalline phase during 40 hours of milling we can conclude that the carbonato complex is amorphous or eventually nanocrystalline to an extent that is undetectable with X-ray diffraction methods. This example illustrates that enriched information on a local structural scale can only be achieved by appropriate selection of analytical tools. The amorphous carbonato complex has recently been confirmed using Raman and nuclear magnetic resonance (NMR) spectroscopies (Rojac et al., to be published).

Another important aspect to discuss is the possible role of water on of the formation of the complex. Jolivet et al. (1982) emphasized the influence of the water molecules on the <sup>3</sup> splitting, which can be significant, depending on whether they can interact via hydrogen bonding with the carbonate group. This was demonstrated through various examples of lanthanide carbonates, where the hydrated forms showed different 3 splitting with respect to their dehydrated analogues. As an example, the hydrated form of the Na2Cu(CO3)2 complex, that is Na2Cu(CO3)23H2O, showed larger 3 splitting, i.e., 203 cm–1, in comparison with its dehydrated form, i.e., 138 cm–1 (see also Table 2, first two examples).

In our case, the possible influence of water molecules on the carbonate ion should be considered. In fact, we showed in section 2.2 (Fig. 6) that an amount of water was introduced in the sample from the air during the milling. Therefore, we have to examine more carefully the possible influence of water adsorption on the 3 splitting. This was done by quenching the 40-hours-treated sample in air from different temperatures so that controlled amounts of water were released; the quenched samples were then analyzed using IR spectroscopy. The IR spectra together with the TG and EGA curves are shown in Fig. 10. The mass losses after quenching at 100 °C, 170 °C and 300 °C were 0.8 %, 3.1 % and 5.0 %,

Using Infrared Spectroscopy to Identify New Amorphous Phases –

for further investigations.

NaVO3, NaTaO3 and NaNbO3.

**mixtures** 

**analysis** 

oxide.

2011).

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 31

humid-free conditions will affect the formation of the niobate. These questions will be left

In-depth study of reaction mechanism limited to one system is often insufficient if fundamental characteristics governing certain type of mechanochemical reaction are to be determined. Following the results from the previous section, in which we identified an amorphous carbonato complex as a transitional stage of the reaction between Na2CO3 and Nb2O5, it is the next step to find out i) whether this mechanism is general for mechanochemical reactions involving CO32– ions and ii) which parameters control the decomposition of the carbonato complex. The latter is particularly important as the decomposition of the complex is a necessary step for the formation of the final binary

In order to study systematically the mechanochemical interaction between CO32– ions and various metal cations, which could possibly lead to the formation of the carbonato complex, we explored the reactions involving Na2CO3, as one reaction counterpart, and various 5th group transition-metal oxides, including V2O5, Nb2O5 and Ta2O5. The aim of the study was to determine the influence of the transition-metal oxide on i) the mechanochemical decomposition of Na2CO3 and ii) the rate of formation of the target binary oxides, i.e.,

The mechanochemical formation of NaMO3 (M = V, Nb, Ta) from respective Na2CO3–M2O5 (M = V, Nb, Ta) mixtures was followed by quantitative X-ray diffraction phase analysis using Rietveld refinement method. The fractions of NaMO3 (M = V, Nb, Ta) as a function of milling time are shown in Fig. 11. The rate of formation of the final oxides follows the order NaVO3>NaTaO3>NaNbO3. Note that the vanadate was formed within 4 hours, while the

Fig. 11. Fraction of NaMO3 (M = V, Nb, Ta), determined by Rietveld refinement analysis, as a function of milling time. The lines are drawn as a guide for the eye (from Rojac et al.,

**3. Mechanochemical reaction rate in Na2CO3–M2O5 (M = V, Nb, Ta) powder** 

**3.1 Quantitative X-ray diffraction, infrared spectroscopy and thermogravimetric** 

Fig. 10. a) TG and EGA(H2O,CO2) curves of the 40-hours high-energy milled Na2CO3–Nb2O5 powder mixture. The dashed lines on the graphs represent the temperatures at which the 40 hours milled sample was air-quenched. b) FT-IR spectra of the 40-hours high-energy milled Na2CO3–Nb2O5 powder mixture air-quenched from various temperatures (from Rojac et al., 2006).

respectively. According to EGA, these mass losses correspond entirely to the removal of water (Fig. 10a, see dashed lines). From Fig. 10b we can see that with increasing amount of released H2O the band at 1605 cm–1 gradually decreases at the expense of the band at 1530 cm–1. Note that the intensity of the band at 1345 cm–1 decreases, too. The results confirm the influence of H2O on the splitting of the 3 vibration.

There are some cases of carbonato complexes, as pointed out by Jolivet et al. (1982), in which water molecules can even modify the coordination state of the carbonate ion. This is also the case of the Na2Cu(CO3)23H2O complex, which contains both bidentate chelate and bridged carbonate ions, whereas its dehydrated form is exclusively a bridged structured (see also Fig. 8). Concerning our carbonato complex, it would be interesting to get more information about the actual influence of H2O on the CO32– coordination. Since apparently the H2O has an active role in the mechanochemical formation of the complex, namely, it affects the <sup>3</sup> splitting, and taking into account that this complex represents an intermediate phase from which the NaNbO3 is formed, it would also be interesting to find out whether milling in

**Transmittance (a.u.)**

b)

40 h

100 °C

170 °C

300 °C

quenching IR analysis

Fig. 10. a) TG and EGA(H2O,CO2) curves of the 40-hours high-energy milled Na2CO3–Nb2O5 powder mixture. The dashed lines on the graphs represent the temperatures at which the 40 hours milled sample was air-quenched. b) FT-IR spectra of the 40-hours high-energy milled Na2CO3–Nb2O5 powder mixture air-quenched from various temperatures (from Rojac et al.,

0 0.4 0.8 1.2 1.6 2

H2O

**300 °C**  *m* = 5.0 % EGA (10-8A)

1880 1680 1480 1280 **Wavenumber (cm-1)**

1530

1605

respectively. According to EGA, these mass losses correspond entirely to the removal of water (Fig. 10a, see dashed lines). From Fig. 10b we can see that with increasing amount of released H2O the band at 1605 cm–1 gradually decreases at the expense of the band at 1530 cm–1. Note that the intensity of the band at 1345 cm–1 decreases, too. The results confirm the

There are some cases of carbonato complexes, as pointed out by Jolivet et al. (1982), in which water molecules can even modify the coordination state of the carbonate ion. This is also the case of the Na2Cu(CO3)23H2O complex, which contains both bidentate chelate and bridged carbonate ions, whereas its dehydrated form is exclusively a bridged structured (see also Fig. 8). Concerning our carbonato complex, it would be interesting to get more information about the actual influence of H2O on the CO32– coordination. Since apparently the H2O has an active role in the mechanochemical formation of the complex, namely, it affects the <sup>3</sup> splitting, and taking into account that this complex represents an intermediate phase from which the NaNbO3 is formed, it would also be interesting to find out whether milling in

influence of H2O on the splitting of the 3 vibration.

0 200 400 600 800

0 200 400 600 800

*m* = 3.1 %

T (°C)

T (°C)

CO2

2006).

84

0

**100 °C** 

*m* = 0.8 % **170 °C** 

1

2

EGA (10-8A)

3

4

88

92

TG (%)

96

100

a)

humid-free conditions will affect the formation of the niobate. These questions will be left for further investigations.

### **3. Mechanochemical reaction rate in Na2CO3–M2O5 (M = V, Nb, Ta) powder mixtures**

#### **3.1 Quantitative X-ray diffraction, infrared spectroscopy and thermogravimetric analysis**

In-depth study of reaction mechanism limited to one system is often insufficient if fundamental characteristics governing certain type of mechanochemical reaction are to be determined. Following the results from the previous section, in which we identified an amorphous carbonato complex as a transitional stage of the reaction between Na2CO3 and Nb2O5, it is the next step to find out i) whether this mechanism is general for mechanochemical reactions involving CO32– ions and ii) which parameters control the decomposition of the carbonato complex. The latter is particularly important as the decomposition of the complex is a necessary step for the formation of the final binary oxide.

In order to study systematically the mechanochemical interaction between CO32– ions and various metal cations, which could possibly lead to the formation of the carbonato complex, we explored the reactions involving Na2CO3, as one reaction counterpart, and various 5th group transition-metal oxides, including V2O5, Nb2O5 and Ta2O5. The aim of the study was to determine the influence of the transition-metal oxide on i) the mechanochemical decomposition of Na2CO3 and ii) the rate of formation of the target binary oxides, i.e., NaVO3, NaTaO3 and NaNbO3.

The mechanochemical formation of NaMO3 (M = V, Nb, Ta) from respective Na2CO3–M2O5 (M = V, Nb, Ta) mixtures was followed by quantitative X-ray diffraction phase analysis using Rietveld refinement method. The fractions of NaMO3 (M = V, Nb, Ta) as a function of milling time are shown in Fig. 11. The rate of formation of the final oxides follows the order NaVO3>NaTaO3>NaNbO3. Note that the vanadate was formed within 4 hours, while the

Fig. 11. Fraction of NaMO3 (M = V, Nb, Ta), determined by Rietveld refinement analysis, as a function of milling time. The lines are drawn as a guide for the eye (from Rojac et al., 2011).

Using Infrared Spectroscopy to Identify New Amorphous Phases –

including the following:


mixtures (from Rojac et al., 2011).



cm–1), followed by Ta2O5 (305 cm–1) and Nb2O5 (270 cm–1).

Na2CO3–V2O5 325 Na2CO3–Ta2O5 305 Na2CO3–Nb2O5 270

(Britzinger & Hester, 1966). The 3 splitting, which reflects the CO3

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 33

The results from Fig. 12 suggest a general reaction mechanism in mixtures involving CO32– ions; in fact, in addition to the systems presented in this chapter, the carbonato complex was identified in a number of other alkaline-carbonate–transition-metal oxide mixtures,

A closer inspection of Fig. 12 reveals several differences between the three reaction systems. First of all, the degree of the 3 splitting is different depending on the metal cation, i.e., V5+, Nb5+ or Ta5+, to which the CO32– coordinate. The maximum splitting of 3 from the spectra of the Na2CO3–M2O5 (M = V, Ta, Nb) mixtures after 4, 72 and 150 hours of milling, respectively (Fig. 12), is collected in Table 3. The maximum 3 splitting is largest in the case of V2O5 (325

**Mixture Max 3 splitting (cm–1)** 

Table 3. Maximum splitting of 3(CO32–) vibration in Na2CO3–M2O5 (M = V, Nb, Ta) powder

Nakamoto et al. (1957) were the first to propose the degree of 3 splitting (3) as a criterion to distinguish between mono- and bidentate coordination in carbonato complexes. Their results showed that some bidentate cobalt carbonato complexes have 3 splitting of about 300 cm–1, while monodentate complexes of analogous chemical composition exhibit about 80 cm–1 of 3. Calculations based on models of XO3 (X = C, N) groups coordinated to a metal cation confirmed the larger splitting in the case of bidentate coordination, as compared to the monodentate coordination (Britzinger & Hester, 1966; Hester & Grossman, 1966). A general relationship between the type of coordination and 3 splitting, which we updated according to the critical review by Busca & Lorenzelli (1982), is shown schematically in Fig. 13a. While monodentate configurations show splitting of around 100 cm–1 or lower, larger 3 splitting can be expected for bidentate chelate and bidentate bridged coordinations.

In addition to the type of coordination, other factors influence the degree of the 3 splitting. As explained in the previous section, the coordination of the CO32– ion causes a rearrangement of the C–O bonds, i.e., the C–O bond coordinated to the metal cation is typically weakened, while the others, non-coordinated, are strengthened. Calculations showed that, for a given type of coordination, this CO32– polarization is more pronounced if the polarizing power of the central cation is high as it can attracts electrons more strongly

2– polarization, should

tantalate and niobate required much longer milling times to be the only crystalline phase detected in the mixtures, i.e., 72 and 150 hours, respectively. The results show that the type of the transition-metal oxide plays an important role in the formation kinetics of NaMO3 (M = V, Nb, Ta).

In order to verify whether the amorphous carbonato complex appears as a transitional phase in the three examined reactions, we performed an IR spectroscopy analysis. The results are presented in Fig. 12. In all the systems, a common trend, characteristic for the lowering of the CO3 2– symmetry, is observed during milling: i) the 3(CO3 2–) vibration shifts gradually to higher wavenumbers and decreases in intensity until it disappears after certain milling time, ii) the 3 vibration is replaced by new bands in the region 1650–1250 cm–1, showing <sup>3</sup> splitting (see 4 h, 72 h, 150 h in Fig. 12 a, b and c, respectively) and iii) the 1 vibration is activated. Note that the 1 activation in the case of the Na2CO3–V2O5 mixture could not be ascertained due to overlapping with the band at 1025 cm–1, related to the stretching vibration of the double vanadyl V=O bonds of V2O5 (Fig. 12a). According to the relation between symmetry and IR vibrational spectroscopy of the CO32– ion, described in detail in the previous section, the formation of the carbonato complex is confirmed in all the examined systems.

We note that the milling conditions for the mechanochemical synthesis of NaNbO3 presented in the previous section were different from the ones that we applied for the study presented here. This is seen from the different kinetics of the formation of NaNbO3, i.e., by comparing the timescale of the NaNbO3 fraction-versus-time curves from Fig. 11 (closed rectangular) and Fig. 4b (open circles). Therefore, the mechanism of the mechanochemical interaction between Na2CO3 and Nb2O5, in terms of the transitional carbonato complex, is qualitatively unaffected by the milling intensity.

Fig. 12. FT-IR spectra of Na2CO3–M2O5 (M = V, Nb, Ta) powder mixtures after different milling times (from Rojac et al., 2011).

The results from Fig. 12 suggest a general reaction mechanism in mixtures involving CO3 2– ions; in fact, in addition to the systems presented in this chapter, the carbonato complex was identified in a number of other alkaline-carbonate–transition-metal oxide mixtures, including the following:


32 Infrared Spectroscopy – Materials Science, Engineering and Technology

tantalate and niobate required much longer milling times to be the only crystalline phase detected in the mixtures, i.e., 72 and 150 hours, respectively. The results show that the type of the transition-metal oxide plays an important role in the formation kinetics of NaMO3

In order to verify whether the amorphous carbonato complex appears as a transitional phase in the three examined reactions, we performed an IR spectroscopy analysis. The results are presented in Fig. 12. In all the systems, a common trend, characteristic for the lowering of the CO32– symmetry, is observed during milling: i) the 3(CO32–) vibration shifts gradually to higher wavenumbers and decreases in intensity until it disappears after certain milling time, ii) the 3 vibration is replaced by new bands in the region 1650–1250 cm–1, showing <sup>3</sup> splitting (see 4 h, 72 h, 150 h in Fig. 12 a, b and c, respectively) and iii) the 1 vibration is activated. Note that the 1 activation in the case of the Na2CO3–V2O5 mixture could not be ascertained due to overlapping with the band at 1025 cm–1, related to the stretching vibration of the double vanadyl V=O bonds of V2O5 (Fig. 12a). According to the relation between symmetry and IR vibrational spectroscopy of the CO32– ion, described in detail in the previous section, the formation of the carbonato complex is confirmed in all the

We note that the milling conditions for the mechanochemical synthesis of NaNbO3 presented in the previous section were different from the ones that we applied for the study presented here. This is seen from the different kinetics of the formation of NaNbO3, i.e., by comparing the timescale of the NaNbO3 fraction-versus-time curves from Fig. 11 (closed rectangular) and Fig. 4b (open circles). Therefore, the mechanism of the mechanochemical interaction between Na2CO3 and Nb2O5, in terms of the transitional carbonato complex, is

Fig. 12. FT-IR spectra of Na2CO3–M2O5 (M = V, Nb, Ta) powder mixtures after different

(M = V, Nb, Ta).

examined systems.

qualitatively unaffected by the milling intensity.

milling times (from Rojac et al., 2011).


A closer inspection of Fig. 12 reveals several differences between the three reaction systems. First of all, the degree of the 3 splitting is different depending on the metal cation, i.e., V5+, Nb5+ or Ta5+, to which the CO32– coordinate. The maximum splitting of 3 from the spectra of the Na2CO3–M2O5 (M = V, Ta, Nb) mixtures after 4, 72 and 150 hours of milling, respectively (Fig. 12), is collected in Table 3. The maximum 3 splitting is largest in the case of V2O5 (325 cm–1), followed by Ta2O5 (305 cm–1) and Nb2O5 (270 cm–1).


Table 3. Maximum splitting of 3(CO32–) vibration in Na2CO3–M2O5 (M = V, Nb, Ta) powder mixtures (from Rojac et al., 2011).

Nakamoto et al. (1957) were the first to propose the degree of 3 splitting (3) as a criterion to distinguish between mono- and bidentate coordination in carbonato complexes. Their results showed that some bidentate cobalt carbonato complexes have 3 splitting of about 300 cm–1, while monodentate complexes of analogous chemical composition exhibit about 80 cm–1 of 3. Calculations based on models of XO3 (X = C, N) groups coordinated to a metal cation confirmed the larger splitting in the case of bidentate coordination, as compared to the monodentate coordination (Britzinger & Hester, 1966; Hester & Grossman, 1966). A general relationship between the type of coordination and 3 splitting, which we updated according to the critical review by Busca & Lorenzelli (1982), is shown schematically in Fig. 13a. While monodentate configurations show splitting of around 100 cm–1 or lower, larger 3 splitting can be expected for bidentate chelate and bidentate bridged coordinations.

In addition to the type of coordination, other factors influence the degree of the 3 splitting. As explained in the previous section, the coordination of the CO32– ion causes a rearrangement of the C–O bonds, i.e., the C–O bond coordinated to the metal cation is typically weakened, while the others, non-coordinated, are strengthened. Calculations showed that, for a given type of coordination, this CO3 2– polarization is more pronounced if the polarizing power of the central cation is high as it can attracts electrons more strongly (Britzinger & Hester, 1966). The 3 splitting, which reflects the CO3 2– polarization, should

Using Infrared Spectroscopy to Identify New Amorphous Phases –

three examined systems.

formation (M = V, Nb, Ta), shown in Fig. 11.

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 35

the order Nb5+<Ta5+<V5+ (see Xz/CN values in Table 5 and refer to the next section for details). Since it is generally accepted that the acidity scales with the cation charge density, i.e., *e*/*r* (Avvakumov et al., 2001), our correlation, in principle, agrees with the one of Jolivet et al. (1982) from Fig. 13b. However, even if a correlation exists, it should be interpreted carefully since in addition to the polarizing ability of the central cation, we should not neglect other influences on the 3 splitting, such as, for example, incorporation of water molecules, as demonstrated in previous section (Fig. 10), which might differ between the

In addition to the 3 splitting, another difference between the three reactions that should be noted is the much faster formation of the complex in the case of V5+ as central cation in comparison with Nb5+ and Ta5+ (Fig. 12). This can be seen by comparing the characteristic <sup>3</sup> splitting among the three systems taking into consideration the point of the transition of the original 3 vibration into split bands as a criterion for the CO32– coordination (Fig. 12). Whereas in the case of V2O5 the 3 vibration almost completely disappeared after 4 hours of milling, giving rise to split bands (Fig. 12a), at least 16 hours were needed in the case of Nb2O5 and Ta2O5 (Fig. 12b and c). This is in agreement with the kinetics of NaMO3

In contrast to the Na2CO3–Ta2O5 and Na2CO3–Nb2O5 systems, where the IR absorption bands of the carbonato complex are still clearly resolved after 72 and 150 hours of milling (Fig. 12b and c), these bands completely disappeared after only 16 h of milling in the case of the Na2CO3–V2O5 mixture (Fig. 12a). A reasonable explanation for the absence of the IR bands related to the complex is its decomposition. If this is the case, it suggests that the

We can compare quantitatively the carbonate decomposition in the three studied systems by using thermogravimetric analysis. Similarly like explained in the previous section, by separating the mass loss related to the H2O release from the one that is due to the CO2 release, we can estimate the amount of the residual carbonate in the powder mixtures. The results of this analysis are shown in Fig. 14. As expected, in all the mixtures the carbonate fraction decreases with increasing milling time; this is associated with the mechanochemically driven carbonate decomposition. We can summarize the reaction as follows: after being formed (Fig. 12), the carbonato complex decomposes (Fig. 14), leading to the formation of the final NaMO3 oxides (Fig. 11). Note that, in terms of the reaction

transition-metal oxide plays a role in the decomposition of the carbonato complex.

timescale, these three stages are not clearly separated, instead, they are overlapped.

The results of TG analysis (Fig. 14) reveal a substantial difference in the decomposition rate of the carbonate between the three systems: the fastest is in the case of V2O5, followed by Ta2O5 and Nb2O5. Note also that the carbonate fraction reaches a plateau after prolonged milling, which we denote here as the "steady-state" milling condition. The amount of the carbonate in this "steady-state" condition depends strongly on the type of the transitionmetal oxides participating in the reaction. Whereas the carbonato complex decomposed nearly completely in the case of vanadium, i.e., only 0.5% of residual carbonate was determined in the "steady-state" milling condition, 29% and 39% remained in the mixture in the case of Ta and Nb, respectively (Fig. 14 and Table 4). This is in agreement with the IR spectra from Fig 12, i.e., in contrast to the cases with Nb and Ta, no IR bands related to the carbonato complex are observed after prolonged milling of the Na2CO3–V2O5 mixture (Fig.

therefore depend on the polarizing power of the central cation. This was indeed confirmed experimentally by Jolivet et al. (1982), which identified a linear increase of 3 splitting with the polarizing power of the central cation for numerous carbonato complexes having the same bidentate coordination (Fig. 13b). For those cases, the polarizing power of the cation was assumed to be proportional to *e*/*r*2, where *e* and *r* are cation charge and radius, respectively. Therefore, the 3 splitting criterion for distinguishing between different types of coordination (Fig. 13a) should only be applied if the polarizing power of the cation is taken into account (Fig. 13b). As pointed out by Busca & Lorenzelli (1982), low values of <sup>3</sup> splitting, e.g., 100 cm–1, do not unequivocally indicate the presence of monodentate structure, particularly in cases of metals having low polarizing power (se also Fig. 13b).

Fig. 13. a) Schematic view of the influence of the type of coordination on splitting of 3(CO32–) vibration (from Busca & Lorenzelli, 1982) and b) correlation between splitting of 3(CO32–) vibration and polarizing power (*e*/*r*2) of the central cation for bidentate carbonato complexes. "*e*" and "*r*" denote cation charge and radius, respectively (from Jolivet et al., 1982).

By comparing with the literature data and considering the relationship shown in Fig. 13a, the maximum 3 splitting in the three mixtures from Table 3, being larger than 100 cm–1, might suggest bidentate chelate and/or bridged coordination. The increasing 3 from the system with niobium, having the smallest 3 of 270 cm–1, to that with vanadium, with the largest 3 of 325 cm–1 (Table 3), correlates with the increasing cation acidity, appearing in

therefore depend on the polarizing power of the central cation. This was indeed confirmed experimentally by Jolivet et al. (1982), which identified a linear increase of 3 splitting with the polarizing power of the central cation for numerous carbonato complexes having the same bidentate coordination (Fig. 13b). For those cases, the polarizing power of the cation was assumed to be proportional to *e*/*r*2, where *e* and *r* are cation charge and radius, respectively. Therefore, the 3 splitting criterion for distinguishing between different types of coordination (Fig. 13a) should only be applied if the polarizing power of the cation is taken into account (Fig. 13b). As pointed out by Busca & Lorenzelli (1982), low values of <sup>3</sup> splitting, e.g., 100 cm–1, do not unequivocally indicate the presence of monodentate structure, particularly in cases of metals having low polarizing power (se also Fig. 13b).

monodentate

3 0 cm–1 100 cm–1 100 – 300 cm–1

**M M M M**

bidentate chelate

bidentate bridged

Fig. 13. a) Schematic view of the influence of the type of coordination on splitting of 3(CO32–) vibration (from Busca & Lorenzelli, 1982) and b) correlation between splitting of 3(CO32–) vibration and polarizing power (*e*/*r*2) of the central cation for bidentate carbonato complexes. "*e*" and "*r*" denote cation charge and radius, respectively (from Jolivet et al.,

By comparing with the literature data and considering the relationship shown in Fig. 13a, the maximum 3 splitting in the three mixtures from Table 3, being larger than 100 cm–1, might suggest bidentate chelate and/or bridged coordination. The increasing 3 from the system with niobium, having the smallest 3 of 270 cm–1, to that with vanadium, with the largest 3 of 325 cm–1 (Table 3), correlates with the increasing cation acidity, appearing in

1982).

a)

Free CO32–

b)

the order Nb5+<Ta5+<V5+ (see Xz/CN values in Table 5 and refer to the next section for details). Since it is generally accepted that the acidity scales with the cation charge density, i.e., *e*/*r* (Avvakumov et al., 2001), our correlation, in principle, agrees with the one of Jolivet et al. (1982) from Fig. 13b. However, even if a correlation exists, it should be interpreted carefully since in addition to the polarizing ability of the central cation, we should not neglect other influences on the 3 splitting, such as, for example, incorporation of water molecules, as demonstrated in previous section (Fig. 10), which might differ between the three examined systems.

In addition to the 3 splitting, another difference between the three reactions that should be noted is the much faster formation of the complex in the case of V5+ as central cation in comparison with Nb5+ and Ta5+ (Fig. 12). This can be seen by comparing the characteristic <sup>3</sup> splitting among the three systems taking into consideration the point of the transition of the original 3 vibration into split bands as a criterion for the CO3 2– coordination (Fig. 12). Whereas in the case of V2O5 the 3 vibration almost completely disappeared after 4 hours of milling, giving rise to split bands (Fig. 12a), at least 16 hours were needed in the case of Nb2O5 and Ta2O5 (Fig. 12b and c). This is in agreement with the kinetics of NaMO3 formation (M = V, Nb, Ta), shown in Fig. 11.

In contrast to the Na2CO3–Ta2O5 and Na2CO3–Nb2O5 systems, where the IR absorption bands of the carbonato complex are still clearly resolved after 72 and 150 hours of milling (Fig. 12b and c), these bands completely disappeared after only 16 h of milling in the case of the Na2CO3–V2O5 mixture (Fig. 12a). A reasonable explanation for the absence of the IR bands related to the complex is its decomposition. If this is the case, it suggests that the transition-metal oxide plays a role in the decomposition of the carbonato complex.

We can compare quantitatively the carbonate decomposition in the three studied systems by using thermogravimetric analysis. Similarly like explained in the previous section, by separating the mass loss related to the H2O release from the one that is due to the CO2 release, we can estimate the amount of the residual carbonate in the powder mixtures. The results of this analysis are shown in Fig. 14. As expected, in all the mixtures the carbonate fraction decreases with increasing milling time; this is associated with the mechanochemically driven carbonate decomposition. We can summarize the reaction as follows: after being formed (Fig. 12), the carbonato complex decomposes (Fig. 14), leading to the formation of the final NaMO3 oxides (Fig. 11). Note that, in terms of the reaction timescale, these three stages are not clearly separated, instead, they are overlapped.

The results of TG analysis (Fig. 14) reveal a substantial difference in the decomposition rate of the carbonate between the three systems: the fastest is in the case of V2O5, followed by Ta2O5 and Nb2O5. Note also that the carbonate fraction reaches a plateau after prolonged milling, which we denote here as the "steady-state" milling condition. The amount of the carbonate in this "steady-state" condition depends strongly on the type of the transitionmetal oxides participating in the reaction. Whereas the carbonato complex decomposed nearly completely in the case of vanadium, i.e., only 0.5% of residual carbonate was determined in the "steady-state" milling condition, 29% and 39% remained in the mixture in the case of Ta and Nb, respectively (Fig. 14 and Table 4). This is in agreement with the IR spectra from Fig 12, i.e., in contrast to the cases with Nb and Ta, no IR bands related to the carbonato complex are observed after prolonged milling of the Na2CO3–V2O5 mixture (Fig.

Using Infrared Spectroscopy to Identify New Amorphous Phases –

mechanochemical interaction (Liao & Senna, 1993).

complete the reaction (Avvakumov et al., 2001).

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 37

metal). For example, in the case of the M(OH)2–SiO2 (M = Ca, Mg) mixtures, they showed experimentally that a larger acid-base potential between Ca(OH)2 and SiO2 brought a faster

The acid-base reaction mechanism is not confined to the hydroxyl groups only. Thermodynamic calculations showed that a correlation exists between the Gibbs free energies of a variety of reactions between oxide compounds and the acid-base potential between the participating oxide reagents for two-component systems: the larger the potential, the more negative the value of the Gibbs energy and, thus, the faster and more

In order to fully consider the acid-base properties of oxide compounds, one should take into account that the acidity of a cation, incorporated into a certain oxide compound, depends on the oxidation state and the coordination number. For example, when the oxidation degree of the manganese ion increases by unity, the acidity increases by 2–3 times; same trend is observed when the coordination number of Si4+ decreases from 6 to 4. For correct comparisons, the influence of these parameters on the acid-base properties of cations should be taken into account. This can be done by introducing the electronegativity of a cation, divided by the coordination number, which defines the cation-ligand force per one bond in a coordination polyhedron; the larger the force, the larger the acidity of the cation, i.e., the stronger is the

To address the acid-base properties of the transition-metal cations, we adopted the electronegativity scale for cations derived by Zhang (1982). Table 5 shows the electronegativities Xz for V5+, Ta5+ and Nb5+. The ratio Xz/CN, where CN refers to the coordination number of the cation, taken as being indicative of the acidity of the cations in

> **Cation Xz Xz/CN**  V5+ 2.02 0.40 Ta5+ 1.88 0.29 Nb5+ 1.77 0.27

Table 5. Electronegativity values Xz and Xz/CN ratios for V5+, Nb5+ and Ta5+. Xz and CN denote cation electronegativity defined by Zhang and coordination number, respectively. The Xz/CN ratio is taken as a parameter proportional to cation acidity (from Rojac et al., 2011).

The order of the cation acidity, i.e., V5+>Ta5+>Nb5+ (Table 5) correlates with our experimental results; in fact, the reaction rate follows the same order, i.e., Na2CO3– V2O5>Na2CO3–Ta2O5>Na2CO3–Nb2O5 (see Fig. 11). This means that the higher is the acidity of the cation involved, the faster is the reaction, including the formation and decomposition of the carbonato complex (Fig. 12 and 14), and the crystallization of the final oxides (Fig. 11). The agreement between the reaction rate sequence and the cation acidity or acid-base potential, where Na2CO3 is taken as basic and transition-metal oxides as acidic compound, suggests that the mechanochemical reactions studied here suit the concept of an acid-base interaction mechanism. Similar correlations between the acid-base properties and the mechanochemical reaction rate can also be found in other systems comprising CaO, as one reagent, and Al2O3, SiO2, TiO2, V2O5 or WO3, as the other reagent (Avvakumov et al. 1994).

ability to attract electron pairs forming covalent bonds (Avvakumov et al., 2001).

their respective oxides, is the highest for V5+, followed by Ta5+ and Nb5+.

Fig. 14. Fraction of carbonate, determined by TG analysis, as a function of milling time in Na2CO3–M2O5 (M = V, Nb, Ta) powder mixtures. The lines are drawn as a guide for the eye (from Rojac et al., 2011).


Table 4. Fraction of residual carbonate in Na2CO3–M2O5 (M = V, Nb, Ta) powder mixtures in the "steady-state" milling conditions (see carbonate fraction after prolonged milling in Fig. 14) (from Rojac et al., 2011).

12a, see 16 and 48 hours). Finally, it is important to stress that the complex is amorphous and could not be analyzed using Rietveld analysis (Fig. 11); instead, we were able to follow its formation and decomposition using IR and TG analyses, respectively (Fig. 12 and 14).

From the presented results we can infer that a common mechanism, characterized by the formation of an intermediate amorphous carbonato complex, link the reactions between Na2CO3 and M2O5 (M = V, Nb, Ta); however, considerable differences exist in the rate of the formation and decomposition of this carbonato complex and, consequently, in the crystallization of the final binary compounds. The sequence of the rates of these reactions, i.e., Na2CO3–V2O5>Na2CO3–Ta2O5>Na2CO3–Nb2O5, can be interpreted by considering the acid-base properties of the reagents involved.

#### **3.2 Acid-base mechanochemical reaction mechanism**

In their extensive work on the mechanochemical reactions involving hydroxide–oxide mixtures, Senna and co-workers (Liao & Senna, 1992, 1993; Watanabe et al., 1995b, 1996, Avvakumov et al., 2001) showed that the mechanism in these mixtures is governed by an acid-base reaction between different hydroxyl groups on the solid surface. The driving force for these reactions is the acid-base potential, i.e., the difference in the acid-base properties between an acidic and basic surface –OH group, which is determined by the type of metal on which it is bound, and therefore, on the strength of the M–OH bond (M denotes the

Fig. 14. Fraction of carbonate, determined by TG analysis, as a function of milling time in Na2CO3–M2O5 (M = V, Nb, Ta) powder mixtures. The lines are drawn as a guide for the eye

Na2CO3–V2O5 0.5 Na2CO3–Ta2O5 29 Na2CO3–Nb2O5 39 Table 4. Fraction of residual carbonate in Na2CO3–M2O5 (M = V, Nb, Ta) powder mixtures in the "steady-state" milling conditions (see carbonate fraction after prolonged milling in Fig.

12a, see 16 and 48 hours). Finally, it is important to stress that the complex is amorphous and could not be analyzed using Rietveld analysis (Fig. 11); instead, we were able to follow its formation and decomposition using IR and TG analyses, respectively (Fig. 12 and 14).

From the presented results we can infer that a common mechanism, characterized by the formation of an intermediate amorphous carbonato complex, link the reactions between Na2CO3 and M2O5 (M = V, Nb, Ta); however, considerable differences exist in the rate of the formation and decomposition of this carbonato complex and, consequently, in the crystallization of the final binary compounds. The sequence of the rates of these reactions, i.e., Na2CO3–V2O5>Na2CO3–Ta2O5>Na2CO3–Nb2O5, can be interpreted by considering the

In their extensive work on the mechanochemical reactions involving hydroxide–oxide mixtures, Senna and co-workers (Liao & Senna, 1992, 1993; Watanabe et al., 1995b, 1996, Avvakumov et al., 2001) showed that the mechanism in these mixtures is governed by an acid-base reaction between different hydroxyl groups on the solid surface. The driving force for these reactions is the acid-base potential, i.e., the difference in the acid-base properties between an acidic and basic surface –OH group, which is determined by the type of metal on which it is bound, and therefore, on the strength of the M–OH bond (M denotes the

**Mixture Carbonate fraction (%)** 

(from Rojac et al., 2011).

14) (from Rojac et al., 2011).

acid-base properties of the reagents involved.

**3.2 Acid-base mechanochemical reaction mechanism** 

metal). For example, in the case of the M(OH)2–SiO2 (M = Ca, Mg) mixtures, they showed experimentally that a larger acid-base potential between Ca(OH)2 and SiO2 brought a faster mechanochemical interaction (Liao & Senna, 1993).

The acid-base reaction mechanism is not confined to the hydroxyl groups only. Thermodynamic calculations showed that a correlation exists between the Gibbs free energies of a variety of reactions between oxide compounds and the acid-base potential between the participating oxide reagents for two-component systems: the larger the potential, the more negative the value of the Gibbs energy and, thus, the faster and more complete the reaction (Avvakumov et al., 2001).

In order to fully consider the acid-base properties of oxide compounds, one should take into account that the acidity of a cation, incorporated into a certain oxide compound, depends on the oxidation state and the coordination number. For example, when the oxidation degree of the manganese ion increases by unity, the acidity increases by 2–3 times; same trend is observed when the coordination number of Si4+ decreases from 6 to 4. For correct comparisons, the influence of these parameters on the acid-base properties of cations should be taken into account. This can be done by introducing the electronegativity of a cation, divided by the coordination number, which defines the cation-ligand force per one bond in a coordination polyhedron; the larger the force, the larger the acidity of the cation, i.e., the stronger is the ability to attract electron pairs forming covalent bonds (Avvakumov et al., 2001).

To address the acid-base properties of the transition-metal cations, we adopted the electronegativity scale for cations derived by Zhang (1982). Table 5 shows the electronegativities Xz for V5+, Ta5+ and Nb5+. The ratio Xz/CN, where CN refers to the coordination number of the cation, taken as being indicative of the acidity of the cations in their respective oxides, is the highest for V5+, followed by Ta5+ and Nb5+.


Table 5. Electronegativity values Xz and Xz/CN ratios for V5+, Nb5+ and Ta5+. Xz and CN denote cation electronegativity defined by Zhang and coordination number, respectively. The Xz/CN ratio is taken as a parameter proportional to cation acidity (from Rojac et al., 2011).

The order of the cation acidity, i.e., V5+>Ta5+>Nb5+ (Table 5) correlates with our experimental results; in fact, the reaction rate follows the same order, i.e., Na2CO3– V2O5>Na2CO3–Ta2O5>Na2CO3–Nb2O5 (see Fig. 11). This means that the higher is the acidity of the cation involved, the faster is the reaction, including the formation and decomposition of the carbonato complex (Fig. 12 and 14), and the crystallization of the final oxides (Fig. 11). The agreement between the reaction rate sequence and the cation acidity or acid-base potential, where Na2CO3 is taken as basic and transition-metal oxides as acidic compound, suggests that the mechanochemical reactions studied here suit the concept of an acid-base interaction mechanism. Similar correlations between the acid-base properties and the mechanochemical reaction rate can also be found in other systems comprising CaO, as one reagent, and Al2O3, SiO2, TiO2, V2O5 or WO3, as the other reagent (Avvakumov et al. 1994).

Using Infrared Spectroscopy to Identify New Amorphous Phases –

physical mixtures of Na2CO3 and Nb2O5 powders.

coordination and formation of an amorphous carbonato complex.

the reagents involved, the faster the mechanochemical reaction.

Institute of Molecular Physics, Polish Academy of Sciences, Poznan, Poland.

quantitative XRD phase analysis.

**5. Acknowledgments** 

**6. References** 

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 39

the amorphous phase, i.e., of up of 91%, formed in the initial part of the reaction. The amount of this amorphous phase then decreased by subsequent milling, leading to the crystallization of the final NaNbO3. Only limited information, including mainly the identification of the transitional nature of the amorphous phase, was obtained using

The decomposition of the carbonate in the Na2CO3–Nb2O5 mixtures was analyzed using TG analysis coupled with DTA and EGA. By following the carbonate decomposition upon annealing the mixtures milled for various periods we were able to infer about the changes occurring in the carbonate during high-energy milling. Characteristic changes in the carbonate decomposition upon increasing the milling time suggested a formation of an intermediate carbonate compound with rather defined decomposition temperature occurring in a narrow temperature range; such decomposition was found to be atypical for

A more accurate identification of the amorphous phase was made possible using IR spectroscopy analysis. Characteristic IR vibrational changes during milling, including splitting of 3 and activation of 1 C–O stretching vibrations of the CO32– ion, suggested lowered CO32– symmetry, which was interpreted as being a consequence of CO32–

Expanding the study of the Na2CO3–Nb2O5 to other systems, we showed in the second part of the chapter that the mechanism involving the transitional amorphous carbonato complex is common for several alkaline-carbonate–transition-metal oxide mixtures, including Na2CO3–M2O5 (M = V, Nb, Ta). The sequence of the reaction rate in these three systems, including the formation of the complex, its decomposition and crystallization of the final NaMO3 (M = V, Nb, Ta), was interpreted by considering the acid-base reaction mechanism. The largest the acid-base potential, i.e., the difference between acidic and basic properties of

The work was supported by the Slovenian Research Agency within the framework of the research program "Electronic Ceramics, Nano, 2D and 3D Structures" (P2-0105) and postdoctoral project "Mechanochemical Synthesis of Complex Ceramic Oxides" (Z2-1195). For the financial support, additional acknowledgements go to Network of Excellence MIND, COST Actions 528 and 539, and bilateral project PROTEUS BI-FR/03-001. For valuable discussions on the topic we thank Barbara Malič and Janez Holc. For the help with various analytical methods, Jana Cilenšek, Bojan Kozlevčar, Edi Krajnc, Anton Meden, Andreja Benčan, Goran Dražič and Bojan Budič are sincerely acknowledged. A special thank is given for the help in the laboratory to Sebastjan Glinšek, Mojca Loncnar, Tanja Urh, Živa Trtnik and Silvo Drnovšek. Collaborations from abroad include Olivier Masson and René Guinebretière from SPCTS, University of Limoges, France, and Bozena Hilczer and Maria Polomska from the

Avvakumov, E. G.; Devyatkina, E. T. & Kosova, N. V. (1994). Mechanochamical Reactions of Hydrated Oxides. *Journal of Solid State Chemistry*, Vol.113, No.2, pp. 379–383.

We showed in the previous section that after reaching a specific milling time ("steady-state" milling condition), the fraction of the residual carbonate did not change any longer if further milling was applied (Fig. 14). These carbonate fractions correlate with the acidity of the cations as well (compare Table 4 with Table 5). Note that the small carbonate fraction in the Na2CO3– V2O5 system, i.e., 0.5 % (Table 4), is consistent with the much larger acidity of V5+ as compared to Ta5+ or Nb5+ (see XZ/CN values in Table 5). The results seem reasonable considering the relation that was found between the acid-base potential and the reaction Gibbs free energy; however, insufficient thermodynamic data for the systems presented here prevent us from making further steps in this direction. We note that these results carry practical consequences, i.e., they suggest that attempting to eliminate the carbonate from a mixture, characterized by a low acid-base potential, by intensifying the milling might not be successful. In fact, the residual carbonate fraction seems to be dependent on the acid-base potential, which is a characteristic of a system, rather than on the milling conditions (Rojac et al., 2008b).

Even if the rate of the three examined reactions apparently agrees with the acid-base reaction concept, we shall not neglect other parameters that could influence the course of the reaction, such as, e.g., adsorption of H2O during milling, which might differ from one system to another. In order to directly verify this possibility, we plot in Fig. 15 the reaction rate constant versus Xz/CN. The reaction rate constant was obtained by fitting the curves from Fig. 11 with a kinetic model proposed for mechanochemical transformations in binary mixtures (Cocco et al., 2000). A linear relationship would be expected if the reaction rate will be largely dominated by the cation acidity. While the sequence of the reaction rate constants, i.e., V>Ta>Nb, agrees with the acid-base reaction mechanism, the non-linear relationship between the rate constant and Xz/CN from Fig. 15 suggests that, in addition to the cation acidity, probably other factors influence the reaction rate. The origin of these additional influences will be left open for further studies.

Fig. 15. Reaction rate constant versus XZ/CN for the reactions between Na2CO3 and M2O5 (M = V, Nb, Ta) (from Rojac et al., 2011).

#### **4. Conclusions**

A systematic study of the reaction mechanism during high-energy milling of a Na2CO3– Nb2O5 mixture, presented in the first part of the chapter, revealed that the synthesis of NaNbO3 takes place through an intermediate amorphous stage. Quantitative phase analysis using XRD diffraction and Rietveld refinement method showed indeed a large amount of the amorphous phase, i.e., of up of 91%, formed in the initial part of the reaction. The amount of this amorphous phase then decreased by subsequent milling, leading to the crystallization of the final NaNbO3. Only limited information, including mainly the identification of the transitional nature of the amorphous phase, was obtained using quantitative XRD phase analysis.

The decomposition of the carbonate in the Na2CO3–Nb2O5 mixtures was analyzed using TG analysis coupled with DTA and EGA. By following the carbonate decomposition upon annealing the mixtures milled for various periods we were able to infer about the changes occurring in the carbonate during high-energy milling. Characteristic changes in the carbonate decomposition upon increasing the milling time suggested a formation of an intermediate carbonate compound with rather defined decomposition temperature occurring in a narrow temperature range; such decomposition was found to be atypical for physical mixtures of Na2CO3 and Nb2O5 powders.

A more accurate identification of the amorphous phase was made possible using IR spectroscopy analysis. Characteristic IR vibrational changes during milling, including splitting of 3 and activation of 1 C–O stretching vibrations of the CO32– ion, suggested lowered CO32– symmetry, which was interpreted as being a consequence of CO32– coordination and formation of an amorphous carbonato complex.

Expanding the study of the Na2CO3–Nb2O5 to other systems, we showed in the second part of the chapter that the mechanism involving the transitional amorphous carbonato complex is common for several alkaline-carbonate–transition-metal oxide mixtures, including Na2CO3–M2O5 (M = V, Nb, Ta). The sequence of the reaction rate in these three systems, including the formation of the complex, its decomposition and crystallization of the final NaMO3 (M = V, Nb, Ta), was interpreted by considering the acid-base reaction mechanism. The largest the acid-base potential, i.e., the difference between acidic and basic properties of the reagents involved, the faster the mechanochemical reaction.

## **5. Acknowledgments**

38 Infrared Spectroscopy – Materials Science, Engineering and Technology

We showed in the previous section that after reaching a specific milling time ("steady-state" milling condition), the fraction of the residual carbonate did not change any longer if further milling was applied (Fig. 14). These carbonate fractions correlate with the acidity of the cations as well (compare Table 4 with Table 5). Note that the small carbonate fraction in the Na2CO3– V2O5 system, i.e., 0.5 % (Table 4), is consistent with the much larger acidity of V5+ as compared to Ta5+ or Nb5+ (see XZ/CN values in Table 5). The results seem reasonable considering the relation that was found between the acid-base potential and the reaction Gibbs free energy; however, insufficient thermodynamic data for the systems presented here prevent us from making further steps in this direction. We note that these results carry practical consequences, i.e., they suggest that attempting to eliminate the carbonate from a mixture, characterized by a low acid-base potential, by intensifying the milling might not be successful. In fact, the residual carbonate fraction seems to be dependent on the acid-base potential, which is a

characteristic of a system, rather than on the milling conditions (Rojac et al., 2008b).

influences will be left open for further studies.

(M = V, Nb, Ta) (from Rojac et al., 2011).

**4. Conclusions** 

Even if the rate of the three examined reactions apparently agrees with the acid-base reaction concept, we shall not neglect other parameters that could influence the course of the reaction, such as, e.g., adsorption of H2O during milling, which might differ from one system to another. In order to directly verify this possibility, we plot in Fig. 15 the reaction rate constant versus Xz/CN. The reaction rate constant was obtained by fitting the curves from Fig. 11 with a kinetic model proposed for mechanochemical transformations in binary mixtures (Cocco et al., 2000). A linear relationship would be expected if the reaction rate will be largely dominated by the cation acidity. While the sequence of the reaction rate constants, i.e., V>Ta>Nb, agrees with the acid-base reaction mechanism, the non-linear relationship between the rate constant and Xz/CN from Fig. 15 suggests that, in addition to the cation acidity, probably other factors influence the reaction rate. The origin of these additional

Fig. 15. Reaction rate constant versus XZ/CN for the reactions between Na2CO3 and M2O5

A systematic study of the reaction mechanism during high-energy milling of a Na2CO3– Nb2O5 mixture, presented in the first part of the chapter, revealed that the synthesis of NaNbO3 takes place through an intermediate amorphous stage. Quantitative phase analysis using XRD diffraction and Rietveld refinement method showed indeed a large amount of The work was supported by the Slovenian Research Agency within the framework of the research program "Electronic Ceramics, Nano, 2D and 3D Structures" (P2-0105) and postdoctoral project "Mechanochemical Synthesis of Complex Ceramic Oxides" (Z2-1195). For the financial support, additional acknowledgements go to Network of Excellence MIND, COST Actions 528 and 539, and bilateral project PROTEUS BI-FR/03-001. For valuable discussions on the topic we thank Barbara Malič and Janez Holc. For the help with various analytical methods, Jana Cilenšek, Bojan Kozlevčar, Edi Krajnc, Anton Meden, Andreja Benčan, Goran Dražič and Bojan Budič are sincerely acknowledged. A special thank is given for the help in the laboratory to Sebastjan Glinšek, Mojca Loncnar, Tanja Urh, Živa Trtnik and Silvo Drnovšek. Collaborations from abroad include Olivier Masson and René Guinebretière from SPCTS, University of Limoges, France, and Bozena Hilczer and Maria Polomska from the Institute of Molecular Physics, Polish Academy of Sciences, Poznan, Poland.

## **6. References**

Avvakumov, E. G.; Devyatkina, E. T. & Kosova, N. V. (1994). Mechanochamical Reactions of Hydrated Oxides. *Journal of Solid State Chemistry*, Vol.113, No.2, pp. 379–383.

Using Infrared Spectroscopy to Identify New Amorphous Phases –

*Powder Technology*, Vol.43, No.2, pp. 155–162.

*Journal of Molecular Structure*, Vol.60, pp. 93–98.

*Materials Science*, Vol.53, No.2, pp. 207–322.

Vol.44, No.3, pp. 1109–1117.

Ljubljana, Slovenia.

*Acta*, Vol.210, pp. 89–102.

Boston, USA.

No.5, pp. 1559–1565.

*Ionics*, Vol.66, No.3-4, pp. 313–319.

*Science and Engineering*, Vol.39, No.2, pp. 193–209.

*Solid State Ionics*, Vol.177, No.33-34, pp. 2987–2995.

*American Ceramic Society*, Vol.91, No.11, pp. 3789–3791.

*Transactions A*, Vol.21, No.1, pp. 289–303.

A Case Study of Carbonato Complex Formed by Mechanochemical Processing 41

Iasonna, A. & Magini, M. (1996). Power Measurements during Mechanical Milling: An

Iguchi, Y. & Senna, M. (1985). Mechanochemical Polymorphic Transformation and Its

Jenko, D. (2006). *Synthesis of (K,Na)NbO3 Ceramics*, PhD Thesis, University of Ljubljana,

Jolivet, J. P.; Thomas, Y. & Taravel, B. (1980). Vibrational Study of Coordinated CO32– Ions.

Jolivet, J. P.; Thomas, Y. & Taravel, B. (1982). Infrared Spectra of Cerium and Thorium Pentacarbonate Complexes. *Journal of Molecular Structure*, Vol.79, pp. 403–408. Kong, L. B.; Zhang, T. S.; Ma, J. & Boey, F. (2008). Progress in Synthesis of Ferroelectric

Kuscer, D.; Holc, J. & Kosec, M. (2006). Mechano-Synthesis of Lead-Magnesium-Niobate Ceramics. *Journal of the American Ceramic Society*, Vol.89, No.10, pp. 3081–3088. Le Brun, P.; Froyen, L. & Delaey, L. (1993). The Modeling of the Mechanical Alloying

Observations. *Materials Science and Engineering A*, Vol. 161, No.1, pp. 75–82. Liao, J. & Senna, M. (1992). Enhanced Dehydration and Amorphization of Mg(OH)2 in the

Liao, J. & Senna, M. (1993). Mechanochemical Dehydration and Amorphization of

Lin, I. J. & Nadiv, S. (1979). Review of the Phase Transformation and Synthesis of Inorganic

Lu, L. & Lai, M. O. (1998). *Mechanical Alloying,* Kluwer Academic Publisher, Boston, USA. Maurice, D. R. & Courtney, T. H. (1990). The Physics of Mechanical Alloying. *Metallurgical* 

Nakamoto, K.; Fujita, J.; Tanaka, S. & Kobayashi, M. (1957). Infrared Spectra of Metallic

Rojac, T.; Kosec, M.; Šegedin, P.; Malič, B. & Holc, J. (2006). The Formation of a Carbonato

Rojac, T.; Benčan, A.; Uršič, H.; Malič, B. & Kosec, M. (2008a). Synthesis of a Li- and Ta-

Rojac, T.; Kosec, M.; Malič, B. & Holc, J. (2008b). The Mechanochemical Synthesis of NaNbO3

Experimental Way to Investigate the Energy Transfer Phenomena. *Acta Materialia*,

Stationary State between Aragonite and Calcite. I. Effects of Preliminary Annealing.

Ceramic Materials Via High-Energy Mechanochemical Technique. *Progress in* 

Process in a Planetary Ball Mill: Comparison between Theory and In-situ

Presence of Ultrafine SiO2 Under Mechanochemical Conditions. *Thermochimica* 

Hydroxides of Ca, Mg and Al on Grinding With and Without SiO2. *Solid State* 

Solids Obtained by Mechanical Treatment (Mechanochemical Reactions). *Materials* 

Complexes. IV. Comparison of the Infrared Spectra of Unidentate and Bidentate Complexes. *Journal of the American Chemical Society*, Vol.79, No.18, pp. 4904–4908. Nakamoto, K. (1997). *Infrared and Raman Spectra of Inorganic and Coordination Compounds. Part* 

*A: Theroy and Applications in Inorganic Chemistry,* Kluwer Academic Publisher,

Complex during the Mechanochemical Treatment of a Na2CO3–Nb2O5 Mixture.

Modified (K,Na)NbO3 Solid Solution by Mechanochemical Activation. *Journal of the* 

Using Different Ball-Impact Energies. *Journal of the American Ceramic Society*, Vol.91,


Avvakumov, E.; Senna, M. & Kosova, N. (2001). *Soft Mechanochemical Synthesis,* Kluwer

Boldyrev, B. B. & Tkacova, K. (2000). Mechanochemistry of Solids : Past, Present, and Prospects. *Journal of Materials Synthesis and Processing*, Vol.8, No.3-4, pp. 121–132. Brintzinger, H. & Hester, R. E. (1966). Vibrational Analysis of Some Oxyanion-Metal

Brooker, M. H. & Bates, J. B. (1971). Raman and Infrared Spectral Studies of Anhydrous Li2CO3 and Na2CO3. *The Journal of Chemical Physics*, Vol.54, No.11, pp. 4788–4796. Buijs, K. & Schutte, C. J. H. (1961). The Infra-red Spectra and Structures of Li2CO3 and Anhydrous Na2CO3. *Spectrochimica Acta B*, Vol.17, No.9-10, pp. 927–932. Burgio, N.; Iasonna, A; Magini, M.; Martelli, S. & Padella, F. (1991). Mechanical Alloying of

Busca, G. & Lorenzelli, V. (1982). Infrared Spectroscopic Identification of Species Arising

Cocco, G.; Delogu, F. & Schiffini, L. (2000). Toward a Quantitative Understanding of the

Delogu, F. & Cocco, G. (2000). Relating Single-Impact Events to Macrokinetic Features in

Delogu, F.; Mulas, G.; Schiffini L. & Cocco, G. (2004). Mechanical Work and Conversion

El-Eskandarany, M. S.; Akoi, K.; Sumiyama, K. & Suzuki, K. (1997). Cyclic Crystalline-

Fujita, J.; Martell, A. E. & Nakamoto, K. (1962). Infrared Spectra of Metal Chelate

Gatehouse, B. M.; Livingstone, S. E. & Nyholm, R. S. (1958). The Infrared Spectra of Some Simple and Complex Carbonates. *Journal of the Chemical Society*, pp. 3137–3142. Glinšek, S.; Malič, B.; Rojac, T.; Filipič, C.; Budič, B. & Kosec, M. (2011). KTaO3 Ceramics

Goldsmith, J. A. & Ross, S. D. (1967). Factors Affecting the Infra-Red Spectra of Planar

Haris, M. J. & Salje, E. K. H. (1992). The Incommensurate Phase of Sodium Carbonates: An

Healy, P. C. & White, A. H. (1972). Crystal Structure and Physical Properties of Anhydrous

Hester, R. E. & Grossman, W. E. L. (1966). Vibrational Analysis of Bidentate Nitrate and Carbonate Complexes. *Inorganic Chemistry*, Vol.5, No.8, pp. 1308–1312.

the Fe-Zr System: Correlation between Input Energy and End Products. *Il Nuovo* 

from Reactive Adsorption of Carbon Oxides on Metal Oxide Surfaces. *Materials* 

Mechanical Alloying Process. *Journal of Materials Synthesis and Processing*, Vol.8,

Mechanical Alloying Processes. *Journal of Materials Synthesis and Processing*, Vol.8,

Degree in Mechanically Induced Processes. *Materials Science and Engineering A*,

Amorphous Transformations of Mechanically Alloyed Co75Ti25. *Applied Physics* 

Compounds. VIII. Infrared Spectra of Co(III) Carbonato Complexes. *The Journal of* 

Prepared by the Mechanochemically Activated Solid-State Synthesis. *Journal of the* 

Anions with D3h Symmetry – IV. The Vibrational Spectra of Some Complex Carbonates in the Region 4000–400 cm–1. *Spectrochimica Acta A*, Vol.24, No.8, pp.

Infrared Absorption Study. *Journal of Physics: Condensed Matter*, Vol.4, No.18, pp.

Sodium Copper Carbonate. *Journal of the Chemical Society, Dalton Transactions*, pp.

Complexes. *Inorganic Chemistry*, Vol.5, No.6, pp. 980–985.

Academic Publisher, Boston, USA.

*Cimento*, Vol.13D, No.4, pp. 459–476.

*Chemistry*, Vol.7, No.1, pp. 89–126.

No.3-4, pp. 167–180.

No.5-6, pp. 271–277.

993–998.

4399–4408.

1913–1917.

Vol.382, No.1-2, pp. 280–287.

*Letters*, Vol.70, No.13, pp. 1679–1681.

*Chemical Physics*, Vol.36, No.2, pp. 339–345.

*American Ceramic Society*, Vol.94, No.5, pp. 1368–1373.


**2** 

*Brazil* 

**Application of Infrared Spectroscopy to** 

Suédina M.L. Silva, Carla R.C. Braga, Marcus V.L. Fook,

**Analysis of Chitosan/Clay Nanocomposites** 

Claudia M.O. Raposo, Laura H. Carvalho and Eduardo L. Canedo *Federal University of Campina Grande, Department of Materials Engineering* 

In recent years, polymer/clay nanocomposites have attracted considerable interest because they combine the structure and physical and chemical properties of inorganic and organic materials. Most work with polymer/clay nanocomposites has concentrated on synthetic polymers, including thermosets such as epoxy polymers, and thermoplastics, such as polyethylene, polypropylene, nylon and poly(ethylene terephthalate) (Pandey & Mishra, 2011). Comparatively little attention has been paid to natural polymer/clay nanocomposites. However, the opportunity to combine at nanometric level clays and natural polymers (biopolymers), such as chitosan, appears as an attractive way to modify some of the properties of this polysaccharide including its mechanical and thermal behavior, solubility and swelling properties, antimicrobial activity, bioadhesion, etc. (Han et al., 2010). Chitosan/clay nanocomposites are economically interesting because they are easy to prepare and involve inexpensive chemical reagents. Chitosan, obtained from chitin, is a relatively inexpensive material because chitin is the second most abundant polymer in nature, next to cellulose (Chang & Juang, 2004). In the same way, clays are abundant and low-cost natural materials. Although chitosan/clay nanocomposites are very attractive, they were not extensively investigated, with relatively small number of scientific publications. In addition, the successful preparation of the nanocomposites still encounters problems, mainly related to the proper dispersion of nano-llers within the polymer matrix. In this chapter, in addition to discussing the synthesis and characterisation by infrared spectroscopy of chitosan/clay nanocomposites, data of x-ray diffraction and mechanical

Chitosan is a naturally occurring linear polysaccharide, closely related to chitin, a polymer widely distributed in the animal kingdom. The discovery of chitosan is ascribed to Rouget in 1859 when he found that boiling chitin in potassium hydroxide rendered the polymer soluble in organic acids. In 1894 Hoppe-Seyler named this material chitosan. Only in 1950 was the structure of chitosan finally resolved (Dodane &Vilivalam, 1998, as cited in Dash et al., 2011). Chitin can be extracted from crustacean shells, insects, fungi, insects and other biological materials (Wan Ngah et al., 2011). The main commercial sources of chitin are the

**1. Introduction** 

properties are also considered.

**1.1 Chitosan** 


## **Application of Infrared Spectroscopy to Analysis of Chitosan/Clay Nanocomposites**

Suédina M.L. Silva, Carla R.C. Braga, Marcus V.L. Fook, Claudia M.O. Raposo, Laura H. Carvalho and Eduardo L. Canedo *Federal University of Campina Grande, Department of Materials Engineering Brazil* 

#### **1. Introduction**

42 Infrared Spectroscopy – Materials Science, Engineering and Technology

Rojac, T.; Kosec, M.; Polomska, M.; Hilczer, B.; Šegedin, P. & Benčan, A. (2009).

Rojac, T.; Benčan, A. & Kosec, M. (2010). Mechanism and Role of Mechanochemical

Rojac, T.; Trtnik, Ž. & Kosec, M. (2011). Mechanochemical Reactions in Na2CO3–M2O5 (M =

Senna, M. & Isobe, T. (1997). NMR and EPR Studies on the Charge Transfer and Formation

Sopicka-Lizer, M. (2010). *High-Energy Ball Milling: Mechanochemical Processing of* 

Suryanarayana, C. (2001). Mechanical Alloying and Milling. *Progress in Materials Science*,

Tacaks, L. (2004). M. Carey Lea, The First Mechanochemist. *Journal of Materials Science*,

Tchernychova, E.; Glinšek, S.; Malič, B. & Kosec, M. (2011). Combined Analytical

Wang, J.; Xue, J. M.; Wan, D. M. & Gan, B. K. (2000a). Mechanically Activating Nucleation

Wang, J.; Xue, J. M. & Wan, D. (2000b). How Different is Mechanical Activation from

Watanabe, R.; Hashimoto, H. & Lee, G. G. (1995a). Computer Simulation of Milling Ball

Watanabe, T.; Liao, J. & Senna, M. (1995b). Changes in the Basicity and Species on the

Watanabe, T.; Isobe, T. & Senna, M. (1996). Mechanisms of Incipient Chemical Reaction

Watanabe, T.; Isobe, T. & Senna, M. (1997). Mechanisms of Incipient Chemical Reaction

Watters, R. L. (2005). NIST Standard Reference Database N°69, United States, June 2005

Zhang, Y. (1982). Electronegativities of Elements in Valence States and Their Applications. 1.

Solutions. *Analytical Biochemistry*, Vol.248, No.2, pp. 234–245.

*Journal of Solid State Chemistry*, Vol.115, No.2, pp. 390–394.

*Journal of Solid State Chemistry*, Vol.130, No.2, pp. 248–289.

*Solid State Chemistry*, Vol.122, No.1, pp. 74–80.

Release (robert.watters@nist.gov).

Transmission Electron Microscopy Approach to Reliable Composition Evaluation of KTaO3. *Journal of the American Ceramic Society*, Vol.94, No.5, pp. 1611–1618. Venyaminov, S. Y. & Pendergast, F. G. (1997). Water (H2O and D2O) Molar Absorptivity in

the 1000–4000 cm–1 Range and Quantitative Infrared Spectroscopy of Aqueous

and Growth of Complex Perovskites. *Journal of Solid State Chemistry*, Vol.154, No.2,

Thermal Activation? A Case Study with PZN and PZN-Based Relaxors. *Solid State* 

Motion in Mechanical Alloying (Overview). *Materials Transactions JIM*, Vol.36,

Surface of Me(OH)2–SiO2 (Me=Ca, Mg, Sr) Mixtures Due to Mechanical Activation.

between Ca(OH)2 and SiO2 under Moderate Mechanical Stressing. I. A Solid State Acid-Base Reaction and Charge Transfer Due to Complex Formation. *Journal of* 

between Ca(OH)2 and SiO2 under Moderate Mechanical Stressing. III. Changes in the Short-Range Ordering throughout the Mechanical and Thermal Processes.

Electronegativities of Elements in Valence States. *Inorganic Chemistry*, Vol.21, No.11,

*Ceramic Society*, Vol.29, No.14, pp. 2999–3006.

*Ceramic Society*, Vol.93, No.6, pp. 1619–1625.

*Nanopowders,* Woodhead Publishing, Boston, USA.

*Solid State Ionics*, Vol.190, No.1, pp. 1–7.

103, No.1, pp. 387–392.

Vol.46, No.3-4, pp. 1–184.

pp. 321–328.

No.2, pp. 102–109.

pp. 3886–3889.

Vol.39, No.16-17, pp. 4987–4993.

*Ionics*, Vol.127, No.1-2, pp. 169–175.

Mechanochemical Reaction in the K2CO3–Nb2O5 System. *Journal of the European* 

Activation in the Synthesis of (K,Na,Li)(Nb,Ta)O3 Ceramics. *Journal of the American* 

V, Nb, Ta) Powder Mixtures: Influence of Transition-Metal Oxide on Reaction Rate.

of Complexes through Incompletely Coordinated States. *Solid State Ionics*, Vol.101-

In recent years, polymer/clay nanocomposites have attracted considerable interest because they combine the structure and physical and chemical properties of inorganic and organic materials. Most work with polymer/clay nanocomposites has concentrated on synthetic polymers, including thermosets such as epoxy polymers, and thermoplastics, such as polyethylene, polypropylene, nylon and poly(ethylene terephthalate) (Pandey & Mishra, 2011). Comparatively little attention has been paid to natural polymer/clay nanocomposites. However, the opportunity to combine at nanometric level clays and natural polymers (biopolymers), such as chitosan, appears as an attractive way to modify some of the properties of this polysaccharide including its mechanical and thermal behavior, solubility and swelling properties, antimicrobial activity, bioadhesion, etc. (Han et al., 2010). Chitosan/clay nanocomposites are economically interesting because they are easy to prepare and involve inexpensive chemical reagents. Chitosan, obtained from chitin, is a relatively inexpensive material because chitin is the second most abundant polymer in nature, next to cellulose (Chang & Juang, 2004). In the same way, clays are abundant and low-cost natural materials. Although chitosan/clay nanocomposites are very attractive, they were not extensively investigated, with relatively small number of scientific publications. In addition, the successful preparation of the nanocomposites still encounters problems, mainly related to the proper dispersion of nano-llers within the polymer matrix. In this chapter, in addition to discussing the synthesis and characterisation by infrared spectroscopy of chitosan/clay nanocomposites, data of x-ray diffraction and mechanical properties are also considered.

#### **1.1 Chitosan**

Chitosan is a naturally occurring linear polysaccharide, closely related to chitin, a polymer widely distributed in the animal kingdom. The discovery of chitosan is ascribed to Rouget in 1859 when he found that boiling chitin in potassium hydroxide rendered the polymer soluble in organic acids. In 1894 Hoppe-Seyler named this material chitosan. Only in 1950 was the structure of chitosan finally resolved (Dodane &Vilivalam, 1998, as cited in Dash et al., 2011). Chitin can be extracted from crustacean shells, insects, fungi, insects and other biological materials (Wan Ngah et al., 2011). The main commercial sources of chitin are the

Application of Infrared Spectroscopy to Analysis of Chitosan/Clay Nanocomposites 45

water soluble below pH about 6.5 (Krajewska, 2004; Lavorgna et al., 2010). In acid conditions, when the amino groups are protonated (Fig. 2), chitosan becomes a soluble polycation (Chivrac et al., 2009). The presence of amino groups make chitosan a cationic polyelectrolyte (pKa ≈ 6.5), one of the few found in nature. Soluble chitosan is heavely charged on the NH3+ groups, it adheres to negatively charged surfaces, aggregates with polyanionic compounds, and chelates heavy metal ions. These characteristics offer

Fig. 2. Schematic illustration of chitosan: (a) at low pH (less than about 6.5), chitosan's amine groups become protonated (polycation); (b) at higher pH (above about 6.5), chitosan's

Increasingly over the last decade chitosan-based materials have been examined and a number of potential products have been developed for areas such as wastewater treatment (removal of heavy metal ions, occulation/coagulation of dyes and proteins, membrane purication processes), the food industry (anticholesterol and fat binding, preservative, packaging material, animal feed additive), agriculture (seed and fertilizer coating, controlled agrochemical release), pulp and paper industry (surface treatment, photographic paper), cosmetic sand toiletries (moisturizer, body creams, bath lotion) (No et al., 2000). Owing to the unparalleled biological properties, the most exciting uses of chitosan-based materials are in the area of medicine and biotechnology. Medicine takes advange of its biocompatibility, biodegradability to harmless products, nontoxicity, physiological inertness, remarkable afnity to proteins, hemostatic, fungistatic, antitumoral and anticholesteremic properties; it may be in drug delivery vehicles, drug controlled release systems, articial cells, wound healing ointments and dressings, haemodialysis membranes, contact lenses, articial skin, surgical sutures and for tissue engineering. In biotechnology they may nd application as chromatographic matrices, membranes for membrane separations, and notably as enzyme/cell immobilization supports (Felt et al., 2000; Krajewska, 2004). The current

Even though a number of potential products have been developed using chitosan-based materials, the tensile properties of pristine chitosan lms are poor (due to its crystallinity). Thermal stability, hardness, gas barrier properties and bacteriostatic activity frequently are not good enough to meet the wide ranges of demanding applications. Thus, modication (chemical modication, blending and graft copolymerization) of chitosan has gained importance as means of tailoring the material to the desired properties. In this context, synthesis of nanocomposites with layered silicate loadings was proposed as a novel approach to modify some of the properties of chitosan, including mechanical and thermal behavior (Wang et al., 2005; Wu and Wu, 2006), solubility and swelling properties in acidic media (Pongjanyakul et al., 2005), antimicrobial activity (Han et al., 2010; Wang et al., 2006)

extraordinary potential in a broad spectrum of chitosan applications.

interest in medical applications of chitosan is easily understood.

aminegroups are deprontonated and reactive.

shell waste of shrimps, lobsters, krills, and crabs. Several millions tons of chitin are harvested annually in the world, making this biopolymer an inexpensive and readily available resource (Dash et al., 2011). Chitosan is found naturally only in certain fungi (Mucoraceae), but it is easily obtained by the thermochemical deacetylation of chitin in the presence of alkali (Darder et al., 2003). Several methods have been proposed, most of them involving the hydrolysis of the acetylated residue using sodium or potassium hydroxide solutions, as well as a mixture of anhydrous hydrazine and hydrazine sulfate. The conditions used for deacetylation determines the polymer molecular weight and the degree of deacetylation (DD) (Dash et al., 2011; Lavorgna et al., 2010).

Chitosan is a copolymer whose chemical structure is shown in Fig. 1. The numbers on the extreme left ring are conventionally assigned to the six carbons in the glucopyranose ring, from C-1 to C-6. Substitution at C-2 may be an acetamido or amino group. Chitosan contains more than 50% (commonly 70 to 90%) of acetamido residues on the C-2 of the structural unit, while amino groups predominate in chitin. The degree of deacetylation (DD) serves as a diagnostic to classify the biopolymer as chitin or chitosan (Dash et al., 2011; Rinaudo, 2006). Notice that DD + DA =1.

The DD is the key property that affects the physical and chemical properties of chitosan, such as solubility, chemical reactivity and biodegradability and, consequently their applications. A quick test to differentiate between chitin and chitosan is based on solubility and nitrogen content. Chitin is soluble in 5% lithium chloride/N,N-dimethylacetamide solvent [LiCI/DMAc] and insoluble in aqueous acetic acid while the opposite is true of chitosan. The nitrogen content in purified samples is less than 7% for chitin and more than 7% for chitosan (Dash et al., 2011; Rinaudo, 2006).

Fig. 1. Chemical structure of chitin and chitosan.

In the solid state, chitosan is a semicrystalline polymer. Its morphology has been investigated and many polymorphs are mentioned in the literature. Single crystals of chitosan were obtained using fully deacetylated chitin of low molecular weight. The dimensions the orthorhombic unit cell of the most common form were determined as *a* = 0,807 nm, *b* = 0,844 nm, *c* = 1,034 nm; the unit cell contains two antiparallel chitosan chains, but no water molecules (Dash et al., 2011).

The degree of acetylation (DA) and the crystallinity of chitin molecules affect the solubility in common solvents. Reducing the acetylation level in chitosan ensures the presence of free amino groups, which can be easily protonated in an acid environment, making chitosan

shell waste of shrimps, lobsters, krills, and crabs. Several millions tons of chitin are harvested annually in the world, making this biopolymer an inexpensive and readily available resource (Dash et al., 2011). Chitosan is found naturally only in certain fungi (Mucoraceae), but it is easily obtained by the thermochemical deacetylation of chitin in the presence of alkali (Darder et al., 2003). Several methods have been proposed, most of them involving the hydrolysis of the acetylated residue using sodium or potassium hydroxide solutions, as well as a mixture of anhydrous hydrazine and hydrazine sulfate. The conditions used for deacetylation determines the polymer molecular weight and the degree

Chitosan is a copolymer whose chemical structure is shown in Fig. 1. The numbers on the extreme left ring are conventionally assigned to the six carbons in the glucopyranose ring, from C-1 to C-6. Substitution at C-2 may be an acetamido or amino group. Chitosan contains more than 50% (commonly 70 to 90%) of acetamido residues on the C-2 of the structural unit, while amino groups predominate in chitin. The degree of deacetylation (DD) serves as a diagnostic to classify the biopolymer as chitin or chitosan (Dash et al., 2011;

The DD is the key property that affects the physical and chemical properties of chitosan, such as solubility, chemical reactivity and biodegradability and, consequently their applications. A quick test to differentiate between chitin and chitosan is based on solubility and nitrogen content. Chitin is soluble in 5% lithium chloride/N,N-dimethylacetamide solvent [LiCI/DMAc] and insoluble in aqueous acetic acid while the opposite is true of chitosan. The nitrogen content in purified samples is less than 7% for chitin and more than

In the solid state, chitosan is a semicrystalline polymer. Its morphology has been investigated and many polymorphs are mentioned in the literature. Single crystals of chitosan were obtained using fully deacetylated chitin of low molecular weight. The dimensions the orthorhombic unit cell of the most common form were determined as *a* = 0,807 nm, *b* = 0,844 nm, *c* = 1,034 nm; the unit cell contains two antiparallel chitosan chains,

The degree of acetylation (DA) and the crystallinity of chitin molecules affect the solubility in common solvents. Reducing the acetylation level in chitosan ensures the presence of free amino groups, which can be easily protonated in an acid environment, making chitosan

of deacetylation (DD) (Dash et al., 2011; Lavorgna et al., 2010).

Rinaudo, 2006). Notice that DD + DA =1.

7% for chitosan (Dash et al., 2011; Rinaudo, 2006).

Fig. 1. Chemical structure of chitin and chitosan.

but no water molecules (Dash et al., 2011).

water soluble below pH about 6.5 (Krajewska, 2004; Lavorgna et al., 2010). In acid conditions, when the amino groups are protonated (Fig. 2), chitosan becomes a soluble polycation (Chivrac et al., 2009). The presence of amino groups make chitosan a cationic polyelectrolyte (pKa ≈ 6.5), one of the few found in nature. Soluble chitosan is heavely charged on the NH3+ groups, it adheres to negatively charged surfaces, aggregates with polyanionic compounds, and chelates heavy metal ions. These characteristics offer extraordinary potential in a broad spectrum of chitosan applications.

Fig. 2. Schematic illustration of chitosan: (a) at low pH (less than about 6.5), chitosan's amine groups become protonated (polycation); (b) at higher pH (above about 6.5), chitosan's aminegroups are deprontonated and reactive.

Increasingly over the last decade chitosan-based materials have been examined and a number of potential products have been developed for areas such as wastewater treatment (removal of heavy metal ions, occulation/coagulation of dyes and proteins, membrane purication processes), the food industry (anticholesterol and fat binding, preservative, packaging material, animal feed additive), agriculture (seed and fertilizer coating, controlled agrochemical release), pulp and paper industry (surface treatment, photographic paper), cosmetic sand toiletries (moisturizer, body creams, bath lotion) (No et al., 2000). Owing to the unparalleled biological properties, the most exciting uses of chitosan-based materials are in the area of medicine and biotechnology. Medicine takes advange of its biocompatibility, biodegradability to harmless products, nontoxicity, physiological inertness, remarkable afnity to proteins, hemostatic, fungistatic, antitumoral and anticholesteremic properties; it may be in drug delivery vehicles, drug controlled release systems, articial cells, wound healing ointments and dressings, haemodialysis membranes, contact lenses, articial skin, surgical sutures and for tissue engineering. In biotechnology they may nd application as chromatographic matrices, membranes for membrane separations, and notably as enzyme/cell immobilization supports (Felt et al., 2000; Krajewska, 2004). The current interest in medical applications of chitosan is easily understood.

Even though a number of potential products have been developed using chitosan-based materials, the tensile properties of pristine chitosan lms are poor (due to its crystallinity). Thermal stability, hardness, gas barrier properties and bacteriostatic activity frequently are not good enough to meet the wide ranges of demanding applications. Thus, modication (chemical modication, blending and graft copolymerization) of chitosan has gained importance as means of tailoring the material to the desired properties. In this context, synthesis of nanocomposites with layered silicate loadings was proposed as a novel approach to modify some of the properties of chitosan, including mechanical and thermal behavior (Wang et al., 2005; Wu and Wu, 2006), solubility and swelling properties in acidic media (Pongjanyakul et al., 2005), antimicrobial activity (Han et al., 2010; Wang et al., 2006)

Application of Infrared Spectroscopy to Analysis of Chitosan/Clay Nanocomposites 47

mass of 720 g/mol. Isomorphic substitution of Al3+ in the octahedral sheets by Mg2+ (less commonly Fe2+, Mn2+ , and other) and, less frequently, of Si4+ by Al3+ in the tetrahedral sheet, results in a net negative charge on the crystaline layer, wich is compensated by the presence of cations, such as Na+, K+, Ca2+, or Mg2+, sorbed between the layers and surrounding the edges. An idealised montmorillonite has 0.67 units of negative charge per unit cell, in other words, it behaves as a weak acid. These loosely held cations do not belong to the crystal structure and can be readily exchanged by other cations, organic or inorganic. The cation exchange capacity (CEC) of montmorillonite ranges from 0.8 to 1.2 meq/g of air-dried clay, resulting in 0.6–0.9 exchangeable cations per unit cell. The layers organize themselves to form stacks with a regular gap between them, the interlayer space or gallery. The electrostatic and Van der Waals forces holding the layers together are relatively weak, and the interlayer distance varies depending on the radius of the cation present and its degree of hydration. In general, the smaller the cation and the lower its charge, the higher the clay swells in water or alcohols. For montmorillonite, the swelling capacity decreases depending on the cation chemical type according to the following trend: Li+ > Na+ > Ca2+ > Fe2+ > K+ (Powell et al., 1998; Tettenhorst et al., 1962, as cited in Chivrac et al., 2009). The distance between two platelets of the primary particle, called inter-layer spacing or d-spacing (*d*001), depends on the silicate type, on the type of the counter-cation, and on the hydration state. For instance, *d*001 = 0.96 nm for anhydrous sodium montmorillonite, but *d*001 = 1.2-1.4 nm in usual, partially hydrated conditions, as

Commercial montmorillonite is available as a powder of about 8 μm particle size, each particle containing about 3000 platelets. Montmorillonite exhibits enhanced gel strength, mucoadhesive capability to cross the gastrointestinal barrier and adsorb bacterial and metabolic toxins such as steroidal metabolites. Because of these advantages in biomedical applications, it is sometimes called a medical clay. Bentonite (named after Ford Benton, Wyoming) is rich in montmorillonite (usually more than 80%) (Utracki, 2004; Wei et al., 2009; Holzer et al., 2010; Li et al., 2010). Its color varies from white to yellow, to olive green, to brown. The names bentonite and montmorillonite are often used interchangeably. However, the terms represent materials with different degrees of purity. Bentonite is the ore that comprises montmorillonite, inessentials minerals and others impurities. Beyond quartz, kaolinite, and many other minerals often present in minute proportions (feldspar, calcite, dolomite, muscovite, chlorite, hematite, etc), organic matter is present in bentonites as intrinsic impurities composed predominantly of humic substances (Bolto et al., 2001). Since competitive reactions can take place between the organic matter present in the bentonite and the chitosan, the extent of intercalation and polymer/clay interactions can be affected. Purification capable

of removing of organic matter from bentonites before intercalation is fundamental.

As mentioned previously, because of the polycationic nature of chitosan in acidic media, the biopolymer may be intercalated in sodium montmorillonite through cation exchange and hydrogen-bonding processes, the resulting nanocomposites showing interesting structural

Chitosan/clay nanocomposites represent an innovative and promising class of materials. Potential biomedical applications of chitosan/clay nanocomposites include: the intercalation of cationic chitosan in the expandable aluminosilicate structure of the clay is expected to affect the binding of cationic drugs by anionic clay; the solubility of chitosan at the low pH of gastric uid may decrease the premature release of drugs in the gastric environment;

determined by x-ray diffraction techniques (Utracki, 2004).

and functional properties.

and bioadhesion (Pongjanyakul and Suksri, 2009). Chemical structure of chitosan containing multiple functional groups (hydroxyl, carbonyl, carboxyl, amine, amide) creates new possibilities for bonding chitosan to clays.

## **1.2 Clays**

Clays are fine-grained, sedimentary rocks originated from the hydrothermal weathering volcanic volcanic ashes in akaline lakes and seas. As such, clays are classified based on their stratigraphic position, location, and mineral content. Clays contain minerals of definite crystaline structure and elementary composition, some as main components, many as impurities, which usually include organic matter in the form of humic acids. Notwithstanding the fundamental difference between clay and clay mineral, both terms are sometimes used as indistinctly, especially in the frequent occasions in which the clay has a single principal mineral component; in this sense, the clay is considered as the impure mineral and the mineral as the purified clay (Utracki, 2004).

Clays are classified on the basis of their crystal structure and the amount and locations of elelectric charge (deficit or excess) per unit cell. Crystalline clays range from kaolins, which are relatively uniform in chemical composition, to smectites, which vary in their composition, cation exchange properties, and ability to expand. The most commonly employed smectite clay for the preparation of polymeric nanocomposites is bentonite, whose main mineral component is montmorillonite (Utracki, 2004).

Montmorillonite is the name given to clay found near Montmorillonin in France, whereit was identified by Knight in 1896 (Utracki, 2004). Montmorillonite is a 2:1 layered hydrated aluminosilicate, with a triple-sheet sandwich structure consisting of a central, hydrous alumina octahedral sheet, bonded to two silica tetrahedral sheets by shared oxygen ions (Fig. 3). The unit cell of this ideal structure has a composition [Al2(OH)2(Si2O5)2]2 with a molar

Fig. 3. Schematic of a montmorillonite, layered clay mineral with a triple-sheet sandwich structure consisting of a central, hydrous alumina octahedral sheet (O), bonded to two silica tetrahedral sheets (T) by shared oxygens.

and bioadhesion (Pongjanyakul and Suksri, 2009). Chemical structure of chitosan containing multiple functional groups (hydroxyl, carbonyl, carboxyl, amine, amide) creates new

Clays are fine-grained, sedimentary rocks originated from the hydrothermal weathering volcanic volcanic ashes in akaline lakes and seas. As such, clays are classified based on their stratigraphic position, location, and mineral content. Clays contain minerals of definite crystaline structure and elementary composition, some as main components, many as impurities, which usually include organic matter in the form of humic acids. Notwithstanding the fundamental difference between clay and clay mineral, both terms are sometimes used as indistinctly, especially in the frequent occasions in which the clay has a single principal mineral component; in this sense, the clay is considered as the impure

Clays are classified on the basis of their crystal structure and the amount and locations of elelectric charge (deficit or excess) per unit cell. Crystalline clays range from kaolins, which are relatively uniform in chemical composition, to smectites, which vary in their composition, cation exchange properties, and ability to expand. The most commonly employed smectite clay for the preparation of polymeric nanocomposites is bentonite,

Montmorillonite is the name given to clay found near Montmorillonin in France, whereit was identified by Knight in 1896 (Utracki, 2004). Montmorillonite is a 2:1 layered hydrated aluminosilicate, with a triple-sheet sandwich structure consisting of a central, hydrous alumina octahedral sheet, bonded to two silica tetrahedral sheets by shared oxygen ions (Fig. 3). The unit cell of this ideal structure has a composition [Al2(OH)2(Si2O5)2]2 with a molar

Fig. 3. Schematic of a montmorillonite, layered clay mineral with a triple-sheet sandwich structure consisting of a central, hydrous alumina octahedral sheet (O), bonded to two silica

tetrahedral sheets (T) by shared oxygens.

possibilities for bonding chitosan to clays.

mineral and the mineral as the purified clay (Utracki, 2004).

whose main mineral component is montmorillonite (Utracki, 2004).

**1.2 Clays** 

mass of 720 g/mol. Isomorphic substitution of Al3+ in the octahedral sheets by Mg2+ (less commonly Fe2+, Mn2+ , and other) and, less frequently, of Si4+ by Al3+ in the tetrahedral sheet, results in a net negative charge on the crystaline layer, wich is compensated by the presence of cations, such as Na+, K+, Ca2+, or Mg2+, sorbed between the layers and surrounding the edges. An idealised montmorillonite has 0.67 units of negative charge per unit cell, in other words, it behaves as a weak acid. These loosely held cations do not belong to the crystal structure and can be readily exchanged by other cations, organic or inorganic. The cation exchange capacity (CEC) of montmorillonite ranges from 0.8 to 1.2 meq/g of air-dried clay, resulting in 0.6–0.9 exchangeable cations per unit cell. The layers organize themselves to form stacks with a regular gap between them, the interlayer space or gallery. The electrostatic and Van der Waals forces holding the layers together are relatively weak, and the interlayer distance varies depending on the radius of the cation present and its degree of hydration. In general, the smaller the cation and the lower its charge, the higher the clay swells in water or alcohols. For montmorillonite, the swelling capacity decreases depending on the cation chemical type according to the following trend: Li+ > Na+ > Ca2+ > Fe2+ > K+ (Powell et al., 1998; Tettenhorst et al., 1962, as cited in Chivrac et al., 2009). The distance between two platelets of the primary particle, called inter-layer spacing or d-spacing (*d*001), depends on the silicate type, on the type of the counter-cation, and on the hydration state. For instance, *d*001 = 0.96 nm for anhydrous sodium montmorillonite, but *d*001 = 1.2-1.4 nm in usual, partially hydrated conditions, as determined by x-ray diffraction techniques (Utracki, 2004).

Commercial montmorillonite is available as a powder of about 8 μm particle size, each particle containing about 3000 platelets. Montmorillonite exhibits enhanced gel strength, mucoadhesive capability to cross the gastrointestinal barrier and adsorb bacterial and metabolic toxins such as steroidal metabolites. Because of these advantages in biomedical applications, it is sometimes called a medical clay. Bentonite (named after Ford Benton, Wyoming) is rich in montmorillonite (usually more than 80%) (Utracki, 2004; Wei et al., 2009; Holzer et al., 2010; Li et al., 2010). Its color varies from white to yellow, to olive green, to brown. The names bentonite and montmorillonite are often used interchangeably. However, the terms represent materials with different degrees of purity. Bentonite is the ore that comprises montmorillonite, inessentials minerals and others impurities. Beyond quartz, kaolinite, and many other minerals often present in minute proportions (feldspar, calcite, dolomite, muscovite, chlorite, hematite, etc), organic matter is present in bentonites as intrinsic impurities composed predominantly of humic substances (Bolto et al., 2001). Since competitive reactions can take place between the organic matter present in the bentonite and the chitosan, the extent of intercalation and polymer/clay interactions can be affected. Purification capable of removing of organic matter from bentonites before intercalation is fundamental.

As mentioned previously, because of the polycationic nature of chitosan in acidic media, the biopolymer may be intercalated in sodium montmorillonite through cation exchange and hydrogen-bonding processes, the resulting nanocomposites showing interesting structural and functional properties.

Chitosan/clay nanocomposites represent an innovative and promising class of materials. Potential biomedical applications of chitosan/clay nanocomposites include: the intercalation of cationic chitosan in the expandable aluminosilicate structure of the clay is expected to affect the binding of cationic drugs by anionic clay; the solubility of chitosan at the low pH of gastric uid may decrease the premature release of drugs in the gastric environment;

Application of Infrared Spectroscopy to Analysis of Chitosan/Clay Nanocomposites 49

dried films were soaked with an aqueous solution of 1 M NaOH for 30 min to remove residual acetic acid, followed by rinsing with distilled water to neutralize, and then dried at

Chitosan/clay films were prepared by a casting/solvent evaporation technique. Firstly, 1% chitosan solutions were adjusted to pH = 4.9 by addition a 1M sodium hydroxide solution to form the NH3+ groups in the chitosan structure. Given that the primary amine group in the structure of the chitosan has a pKa = 6.3, 95% of the groups amine will be protonated at the final pH = 5. of the chitosan/clay mixture (Darder et al., 2005). After, the chitosan solution was slowly added to a 1 wt% clay suspension followed by stirring at 53 2°C for 4 h to obtain the films with chitosan/clay mass ratios of 1:1, 5:1 and 10:1. This chitosan/clay solution was cast into Petri dishes and dried at 50°C for 20 h to evaporate the solvent and form the films. Following the same procedure used for chitosan films, the dried films were soaked into an aqueous solution of 1 M NaOH for 30 min to remove residual acetic acid, followed by rinsing in distilled water to neutral and then dried at room temperature. The chitosan/purified sodium bentonite and chitosan/sodium montmorillonite films prepared from chitosan/clay mass ratio of the 1:1, 5:1 and 10:1 were denoted CS1:BNT1; CS5:BNT1;

Although the clay dispersion process is usually followed by x-ray diffraction and transmission electron microscopy, infrared spectroscopic techniques may shed light into the complex chemical and physical interactions involved, helping scientists and technologists to understand the mechanisms of nanocomposite formation, and leading to better products and production methods in the laboratory and the industrial plant. Furthermore, infrared spectroscopic is relatively rapid, is a common instrument found in most research laboratories, sample purity is not as critical and the method can be used with insoluble samples. This gives infrared spectroscopic methods an advantage over other methods,

Fourier transform infrared spectra of the chitosan films and the chitosan/clay films were collected using a Spectrum 400 Perkin Elmer operating in the range of 400-4000 cm−1 at a

XRD patterns were obtained using a Shimadzu XRD-6000 diffractometer with CuK<sup>α</sup> radiation (λ = 0.154 nm, 40 kV, 30 mA) at room temperature. XRD scans were performed on sodium montmorillonite and purified sodium bentonite, chitosan films and chitosan/clay films with a 2 range between 1.5º and 12.0º, at a scanning rate of 1º/min and a scanning step of 0.02º. The basal spacing (*d*001) value of the layered silicates and the chitosan/layer

Mechanical properties of chitosan films and chitosan/clay films were measured following ASTM D882 standard procedures. The films were cut in rectangular strips (80 10 mm) and the thickness of each sample was measured at three different locations and averaged. The tensile strength (TS), elastic modulus (EM) and elongation at break (E) of the samples were determined using a universal testing machine (EMIC, model DL1000) fitted with a load cell

room temperature. The chitosan films were coded CS.

CS10/BNT1 and CS1:MMT1; CS5:MMT1; CS10:MMT1, respectively.

which require elaborate and time-consuming sample preparation.

silicate films were computed using Bragg's law.

**2.3 Preparation of the chitosan/clay films** 

**2.4 Characterization** 

resolution of 4 cm1.

cationic chitosan may result in the efcient transport of negatively charged drugs; the presence of reactive amine groups on chitosan may provide ligand attachment sites for targeted drug delivery; etc. The limited solubility of a chitosan/clay nanocomposite drug carriers at gastric pH offers signicant advantages for colon-specic delivery of drugs that may destroyed in the acidic gastric environment or by the presence of gastric digestive enzymes. Furthermore, the mucoadhesive properties of chitosan may enhance the bioavailability of drugs in the gastrointestional tract.

Many actual applications of chitosan/clay nanocomposites are reported in the literature. Darder et al., 2005 prepared chitosan/montmorillonite nanocomposites and used them in potentiometric sensors for anion detection. Gecol et al., 2006 investigated the removal of tungsten from water using chitosan coated montmorillonite biosorbents. Chang and Juang, 2004 studied the adsorption of tannic acid, humic acid, and dyes from water using chitosan/activated clay composites. An and Dultz, 2007 reported the adsorption of tannic acid on chitosan–montmorillonite as well Pongjanyakul et al., 2005; Wang et al., 2005; Wu and Wu, 2006; Günister et al., 2007; Khunawattanakul et al., 2008; Pongjanyakul & Suksri, 2009. Darder et al. , 2005 synthesized functional chitosan/MMT nanocomposites, successfully used in the development of bulk modified electrodes. Wang et al., 2005 reported the effect of acetic acid residue and MMT loading in the nanocomposites.

However, there are few reports on chitosan/bentonite nanocomposites (Yang & Chen, 2007; Zhang et al., 2009; Wan Ngah et al., 2010). The physical properties and biological response of chitosan strongly depend on the starting materials and nanocomposite preparation conditions. In the present study chitosan/clay nanocomposites were prepared using two kinds of clay and different chitosan/clay ratios, to evaluate how these variables affect the dispersion of clay particles into the chitosan matrix. The samples obtained were characterized by infrared spectroscopy, x-ray diffraction, and mechanical (tensile) properties.

## **2. Experimental**

### **2.1 Materials**

Chitosan was supplied by Polymar (Fortaleza, CE, Brazil) and used without purification. The chitosan was obtained by deacetylation of chitin from crab shells, with a degree of deacetylation of 86.7%. Sodium bentonite (Argel 35) was provided by Bentonit União Nordeste (Campina Grande, PB, Brazil). The clay, coded BNT, was purified according to procedure reported elsewhere (Araujo et al., 2007); the cation exchange capacity (CEC) of the purified bentonite was 0.92 meq/g (Leite et al., 2010). Sodium montmorillonite (Cloisite Na+), coded MMT, with a CEC of 0.90 meq/g was supplied by Southern Clay Products (Gonzalez, TX, USA). Both of the clays, purified sodium bentonite (BNT) and sodium montmorillonite (MMT), were screened to 200 mesh size before mixed with chitosan.

#### **2.2 Preparation of chitosan films**

Chitosan solutions were prepared by dissolving chitosan in a 1% aqueous acetic acid solution at a concentration of 1 wt% under continuous stirring at 45°C for 2 h followed by vacuum filtering to remove the insoluble residue. This solution was cast into Petri dishes (radius 12 cm) and dried at 50°C for 20 h to evaporate the solvent and form the films. The dried films were soaked with an aqueous solution of 1 M NaOH for 30 min to remove residual acetic acid, followed by rinsing with distilled water to neutralize, and then dried at room temperature. The chitosan films were coded CS.

## **2.3 Preparation of the chitosan/clay films**

Chitosan/clay films were prepared by a casting/solvent evaporation technique. Firstly, 1% chitosan solutions were adjusted to pH = 4.9 by addition a 1M sodium hydroxide solution to form the NH3+ groups in the chitosan structure. Given that the primary amine group in the structure of the chitosan has a pKa = 6.3, 95% of the groups amine will be protonated at the final pH = 5. of the chitosan/clay mixture (Darder et al., 2005). After, the chitosan solution was slowly added to a 1 wt% clay suspension followed by stirring at 53 2°C for 4 h to obtain the films with chitosan/clay mass ratios of 1:1, 5:1 and 10:1. This chitosan/clay solution was cast into Petri dishes and dried at 50°C for 20 h to evaporate the solvent and form the films. Following the same procedure used for chitosan films, the dried films were soaked into an aqueous solution of 1 M NaOH for 30 min to remove residual acetic acid, followed by rinsing in distilled water to neutral and then dried at room temperature. The chitosan/purified sodium bentonite and chitosan/sodium montmorillonite films prepared from chitosan/clay mass ratio of the 1:1, 5:1 and 10:1 were denoted CS1:BNT1; CS5:BNT1; CS10/BNT1 and CS1:MMT1; CS5:MMT1; CS10:MMT1, respectively.

#### **2.4 Characterization**

48 Infrared Spectroscopy – Materials Science, Engineering and Technology

cationic chitosan may result in the efcient transport of negatively charged drugs; the presence of reactive amine groups on chitosan may provide ligand attachment sites for targeted drug delivery; etc. The limited solubility of a chitosan/clay nanocomposite drug carriers at gastric pH offers signicant advantages for colon-specic delivery of drugs that may destroyed in the acidic gastric environment or by the presence of gastric digestive enzymes. Furthermore, the mucoadhesive properties of chitosan may enhance the

Many actual applications of chitosan/clay nanocomposites are reported in the literature. Darder et al., 2005 prepared chitosan/montmorillonite nanocomposites and used them in potentiometric sensors for anion detection. Gecol et al., 2006 investigated the removal of tungsten from water using chitosan coated montmorillonite biosorbents. Chang and Juang, 2004 studied the adsorption of tannic acid, humic acid, and dyes from water using chitosan/activated clay composites. An and Dultz, 2007 reported the adsorption of tannic acid on chitosan–montmorillonite as well Pongjanyakul et al., 2005; Wang et al., 2005; Wu and Wu, 2006; Günister et al., 2007; Khunawattanakul et al., 2008; Pongjanyakul & Suksri, 2009. Darder et al. , 2005 synthesized functional chitosan/MMT nanocomposites, successfully used in the development of bulk modified electrodes. Wang et al., 2005

reported the effect of acetic acid residue and MMT loading in the nanocomposites.

However, there are few reports on chitosan/bentonite nanocomposites (Yang & Chen, 2007; Zhang et al., 2009; Wan Ngah et al., 2010). The physical properties and biological response of chitosan strongly depend on the starting materials and nanocomposite preparation conditions. In the present study chitosan/clay nanocomposites were prepared using two kinds of clay and different chitosan/clay ratios, to evaluate how these variables affect the dispersion of clay particles into the chitosan matrix. The samples obtained were characterized by infrared spectroscopy, x-ray diffraction, and mechanical (tensile)

Chitosan was supplied by Polymar (Fortaleza, CE, Brazil) and used without purification. The chitosan was obtained by deacetylation of chitin from crab shells, with a degree of deacetylation of 86.7%. Sodium bentonite (Argel 35) was provided by Bentonit União Nordeste (Campina Grande, PB, Brazil). The clay, coded BNT, was purified according to procedure reported elsewhere (Araujo et al., 2007); the cation exchange capacity (CEC) of the purified bentonite was 0.92 meq/g (Leite et al., 2010). Sodium montmorillonite (Cloisite Na+), coded MMT, with a CEC of 0.90 meq/g was supplied by Southern Clay Products (Gonzalez, TX, USA). Both of the clays, purified sodium bentonite (BNT) and sodium

montmorillonite (MMT), were screened to 200 mesh size before mixed with chitosan.

Chitosan solutions were prepared by dissolving chitosan in a 1% aqueous acetic acid solution at a concentration of 1 wt% under continuous stirring at 45°C for 2 h followed by vacuum filtering to remove the insoluble residue. This solution was cast into Petri dishes (radius 12 cm) and dried at 50°C for 20 h to evaporate the solvent and form the films. The

bioavailability of drugs in the gastrointestional tract.

properties.

**2. Experimental 2.1 Materials** 

**2.2 Preparation of chitosan films** 

Although the clay dispersion process is usually followed by x-ray diffraction and transmission electron microscopy, infrared spectroscopic techniques may shed light into the complex chemical and physical interactions involved, helping scientists and technologists to understand the mechanisms of nanocomposite formation, and leading to better products and production methods in the laboratory and the industrial plant. Furthermore, infrared spectroscopic is relatively rapid, is a common instrument found in most research laboratories, sample purity is not as critical and the method can be used with insoluble samples. This gives infrared spectroscopic methods an advantage over other methods, which require elaborate and time-consuming sample preparation.

Fourier transform infrared spectra of the chitosan films and the chitosan/clay films were collected using a Spectrum 400 Perkin Elmer operating in the range of 400-4000 cm−1 at a resolution of 4 cm1.

XRD patterns were obtained using a Shimadzu XRD-6000 diffractometer with CuK<sup>α</sup> radiation (λ = 0.154 nm, 40 kV, 30 mA) at room temperature. XRD scans were performed on sodium montmorillonite and purified sodium bentonite, chitosan films and chitosan/clay films with a 2 range between 1.5º and 12.0º, at a scanning rate of 1º/min and a scanning step of 0.02º. The basal spacing (*d*001) value of the layered silicates and the chitosan/layer silicate films were computed using Bragg's law.

Mechanical properties of chitosan films and chitosan/clay films were measured following ASTM D882 standard procedures. The films were cut in rectangular strips (80 10 mm) and the thickness of each sample was measured at three different locations and averaged. The tensile strength (TS), elastic modulus (EM) and elongation at break (E) of the samples were determined using a universal testing machine (EMIC, model DL1000) fitted with a load cell

Application of Infrared Spectroscopy to Analysis of Chitosan/Clay Nanocomposites 51

chitosan/MMT and chitosan/BNT films with 1:1, 5:1 and 10:1 chitosan/clay ratios, respectively (CS1:MMT1; CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1;

In the clay spectra (MMT and BNT), the characteristic absorption band at ~ 3622 cm-1 [OH] is assigned to the stretching vibration of AlOH and SiOH; at ~ 3416 cm-1 [OH] to the stretching vibration of H2O; at ~ 1628 cm-1 [HOH] to the bending vibration of H2O; at ~1118 cm-1 and at ~ 980 cm-1[Si-O] to the stretching vibration of SiO; at~913 cm-1 [Al-Al-OH] to the bending vibration of AlAlOH; at~882 cm-1 [Al-Fe-OH] to the bending vibration of AlFeOH; and at~841 cm-1 [Al-Mg-OH] to the bending vibration of AlMgOH (Awad et al., 2004; Bora et

In order to fully characterize the starting materials, a spectrum of pure chitosan was also recorded. The main bands appearing in that spectrum were due to stretching vibrations of OH groups in the range from 3750 cm-1 to 3000 cm-1, which are overlapped to the stretching vibration of N-H; and C–H bond in –CH2 (1 = 2920 cm−1) and –CH3 (2 = 2875 cm−1)groups, respectively. Bending vibrations of methylene and methyl groups were also visible at = 1375 cm−1 and = 1426 cm−1, respectively (Mano et al., 2003). Absorption in the range of 1680–1480 cm−1 was related to the vibrations of carbonyl bonds (C=O) of the amide group CONHR (secondary amide, 1 = 1645 cm−1) and to the vibrations of protonated amine group

 , 2 = 1574 cm−1) (Marchessault et al., 2006). Absorption in the range from 1160 cm−1 to 1000 cm−1 has been attributed to vibrations of CO group (Xu et al, 2005). The band located near = 1150 cm−1 is related to asymmetric vibrations of CO in the oxygen bridge resulting from deacetylation of chitosan. The bands near 1080–1025 cm−1 are attributed to CO of the ring COH, COC and CH2OH. The small peak at ~890 cm-1 corresponds to wagging of the saccharide structure of chitosan (Darder et al., 2003; Paluszkiewicz et al., 2011; Yuan et al., 2010).The assigned characteristic FTIR absorption bands of clay (MMT and BNT) and

groups interact electrostatically with the negatively charged

groups that do not interact electrostatically with the clay

band towards a lower frequency is observed in all

band also increases for higher amounts of

vibration

al., 2000; Leite et al., 2010; Madejová, 2003; Xu et al., 2009).

chitosan film (CS) derived from Fig. 4 are summarized in Table 1.

FTIR was also used to study the polymer/clay interaction, since a shift in the *NH*<sup>3</sup>

the chitosan/clay films (CS1:MMT1; CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1; CS10:BNT1) as show in Fig. 5 (spectra of Fig. 4 in the 1800–1400 cm-1 wavenumber range) and Table 2. Nevertheless, this shift is higher for chitosan/clay films with the lowest amounts of chitosan (CS1:MMT1; CS5:MMT1 and CS1:BNT1; CS5:BNT1), while the chitosan/clay films with the highest amounts of biopolymer (CS10:MMT1 and CS10:BNT1) show a frequency value that trends to that observed in the films of pure chitosan (CS). This

intercalated chitosan (CS10:MMT1 and CS10:BNT1) (Fig.5). The secondary amide band (1) at 1645 cm-1 of chitosan is overlapped with the HOH bending vibration band at 1628 cm-1 of the water molecules associated to the chitosan/clay films, which are present as in the starting clay, as expected for a biopolymer with high water retention capability (Darder et al., 2003; Darder et al., 2005; Han et al., 2010; Paluszkiewicz et al., 2011; Tan et al., 2007; Wang & Wang, 2007). Comparing the spectra of chitosan/MMT with the spectra of chitosan/BNT we can observe that the interaction of the chitosan with both clays (MMT and

CS10:BNT1).

( *NH*<sup>3</sup> 

may be expected when – *NH*<sup>3</sup>

fact may be related to the – *NH*<sup>3</sup>

BNT) is similar.

substrate (Fig.6). Besides, the intensity of the *NH*<sup>3</sup>

sites of the clay. In fact, a shift of the *NH*<sup>3</sup>

of 50 N, with initial gauge separation of 50 mm and a stretching speed of 5 mm/min. Reported results were the average of five independent measurements.

#### **3. Results and discussion**

#### **3.1 Infrared spectroscopy (FTIR)**

Fig. 4 shows FTIR spectra in the 4000–400 cm1 wave number range for sodium montmorillonite (MMT), purified sodium bentonite (BNT), chitosan film (CS),

Fig. 4. FTIR spectra in the 4000–400 cm1 wave number range for sodium montmorillonite (MMT), purified sodium bentonite (BNT), chitosan film (CS), chitosan/MMT and chitosan/BNT films with 1:1, 5:1 and 10:1 chitosan/clay ratios, respectively (CS1:MMT1; CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1; CS10:BNT1).

of 50 N, with initial gauge separation of 50 mm and a stretching speed of 5 mm/min.

Fig. 4 shows FTIR spectra in the 4000–400 cm1 wave number range for sodium montmorillonite (MMT), purified sodium bentonite (BNT), chitosan film (CS),

882

CS10:BNT1

CS5:BNT1

CS1:BNT1

CS10:MMT1

CS5:MMT1

CS1:MMT1

CS

BNT

MMT

1118

980

1574

2920 1645 2875

<sup>1628</sup> <sup>3416</sup>

882

913

4000 3500 3000 2500 2000 1500 1000 500

Wavenumber (cm-1

Fig. 4. FTIR spectra in the 4000–400 cm1 wave number range for sodium montmorillonite

(MMT), purified sodium bentonite (BNT), chitosan film (CS), chitosan/MMT and chitosan/BNT films with 1:1, 5:1 and 10:1 chitosan/clay ratios, respectively (CS1:MMT1;

CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1; CS10:BNT1).

)

Reported results were the average of five independent measurements.

**3. Results and discussion** 

**3.1 Infrared spectroscopy (FTIR)** 

Absorbance (a.u.)

3622

chitosan/MMT and chitosan/BNT films with 1:1, 5:1 and 10:1 chitosan/clay ratios, respectively (CS1:MMT1; CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1; CS10:BNT1).

In the clay spectra (MMT and BNT), the characteristic absorption band at ~ 3622 cm-1 [OH] is assigned to the stretching vibration of AlOH and SiOH; at ~ 3416 cm-1 [OH] to the stretching vibration of H2O; at ~ 1628 cm-1 [HOH] to the bending vibration of H2O; at ~1118 cm-1 and at ~ 980 cm-1[Si-O] to the stretching vibration of SiO; at~913 cm-1 [Al-Al-OH] to the bending vibration of AlAlOH; at~882 cm-1 [Al-Fe-OH] to the bending vibration of AlFeOH; and at~841 cm-1 [Al-Mg-OH] to the bending vibration of AlMgOH (Awad et al., 2004; Bora et al., 2000; Leite et al., 2010; Madejová, 2003; Xu et al., 2009).

In order to fully characterize the starting materials, a spectrum of pure chitosan was also recorded. The main bands appearing in that spectrum were due to stretching vibrations of OH groups in the range from 3750 cm-1 to 3000 cm-1, which are overlapped to the stretching vibration of N-H; and C–H bond in –CH2 (1 = 2920 cm−1) and –CH3 (2 = 2875 cm−1)groups, respectively. Bending vibrations of methylene and methyl groups were also visible at = 1375 cm−1 and = 1426 cm−1, respectively (Mano et al., 2003). Absorption in the range of 1680–1480 cm−1 was related to the vibrations of carbonyl bonds (C=O) of the amide group CONHR (secondary amide, 1 = 1645 cm−1) and to the vibrations of protonated amine group ( *NH*<sup>3</sup> , 2 = 1574 cm−1) (Marchessault et al., 2006). Absorption in the range from 1160 cm−1 to 1000 cm−1 has been attributed to vibrations of CO group (Xu et al, 2005). The band located near = 1150 cm−1 is related to asymmetric vibrations of CO in the oxygen bridge resulting from deacetylation of chitosan. The bands near 1080–1025 cm−1 are attributed to CO of the ring COH, COC and CH2OH. The small peak at ~890 cm-1 corresponds to wagging of the saccharide structure of chitosan (Darder et al., 2003; Paluszkiewicz et al., 2011; Yuan et al., 2010).The assigned characteristic FTIR absorption bands of clay (MMT and BNT) and chitosan film (CS) derived from Fig. 4 are summarized in Table 1.

FTIR was also used to study the polymer/clay interaction, since a shift in the *NH*<sup>3</sup> vibration may be expected when – *NH*<sup>3</sup> groups interact electrostatically with the negatively charged sites of the clay. In fact, a shift of the *NH*<sup>3</sup> band towards a lower frequency is observed in all the chitosan/clay films (CS1:MMT1; CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1; CS10:BNT1) as show in Fig. 5 (spectra of Fig. 4 in the 1800–1400 cm-1 wavenumber range) and Table 2. Nevertheless, this shift is higher for chitosan/clay films with the lowest amounts of chitosan (CS1:MMT1; CS5:MMT1 and CS1:BNT1; CS5:BNT1), while the chitosan/clay films with the highest amounts of biopolymer (CS10:MMT1 and CS10:BNT1) show a frequency value that trends to that observed in the films of pure chitosan (CS). This fact may be related to the – *NH*<sup>3</sup> groups that do not interact electrostatically with the clay substrate (Fig.6). Besides, the intensity of the *NH*<sup>3</sup> band also increases for higher amounts of intercalated chitosan (CS10:MMT1 and CS10:BNT1) (Fig.5). The secondary amide band (1) at 1645 cm-1 of chitosan is overlapped with the HOH bending vibration band at 1628 cm-1 of the water molecules associated to the chitosan/clay films, which are present as in the starting clay, as expected for a biopolymer with high water retention capability (Darder et al., 2003; Darder et al., 2005; Han et al., 2010; Paluszkiewicz et al., 2011; Tan et al., 2007; Wang & Wang, 2007). Comparing the spectra of chitosan/MMT with the spectra of chitosan/BNT we can observe that the interaction of the chitosan with both clays (MMT and BNT) is similar.

Application of Infrared Spectroscopy to Analysis of Chitosan/Clay Nanocomposites 53

1574 cm-1 1645

NH3

CO

1800 1700 1600 1500 1400

Wavenumber (cm-1)

Fig. 5. IR spectra of Fig. 4 in the 1800–1400 cm-1 wavenumber range of chitosan film (CS), chitosan/MMT and chitosan/BNT films prepared from 1:1, 5:1 and 10:1 chitosan–clay ratios

(CS1:MMT1; CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1; CS10:BNT1).

1570 cm-1

CS10:BNT1

CS5:BNT1

CS1:BNT1

CS10:MMT1

CS5:MMT1

CS1:MMT1

CS

1558 cm-1

1557 cm-1

1570 cm-1

Absorbance (a.u.)

1558 cm-1

1555 cm-1


\* = stretching vibration; S = symmetric stretching vibration;

aS = asymmetric stretching vibration; = wagging.

Table 1. Assignment of FTIR spectra of clays and chitosan derived from Fig. 4.


\* stretching band of secondary amide (-C=O).

Table 2. Frequency values of vibrational bands corresponding to the water molecules associated with the clay (MMT and BNT) and with the protonated amine group in the chitosan chain.

**Sample IR band (cm-1) Description\***  Clay (MMT and BNT) 3622 (O-H) for Al-OH and Si-OH

Chitosan film (CS) 3750-3000 (O-H) overlapped to the S(N-H) 2920 as(C-H) 2875 s(C-H)

> 1645 (-C=O) secondaryamide 1574 (-C=O)protonated amine

1313 s(-CH3) tertiary amide

**Sample HOH (cm-1)** *NH*<sup>3</sup>

1261 (C-O-H)

890 (C-H)

Table 1. Assignment of FTIR spectra of clays and chitosan derived from Fig. 4.

Clay (MMT and BNT) 1628\* -

Chitosan film (CS) 1645 1574

Chitosan/MMT (CS1:MMT1) 1638 1555 Chitosan/MMT (CS5:MMT1) 1640 1558 Chitosan/MMT (CS10:MMT1) 1645 1570

Chitosan/BNT (CS1:BNT1) 1640 1557 Chitosan/BNT (CS5:BNT1) 1641 1558 Chitosan/BNT (CS10:BNT1) 1643 1570

Table 2. Frequency values of vibrational bands corresponding to the water molecules associated with the clay (MMT and BNT) and with the protonated amine group in the

= stretching vibration; S = symmetric stretching vibration;

aS = asymmetric stretching vibration; = wagging.

stretching band of secondary amide (-C=O).

\*

\*

chitosan chain.

1426, 1375 (C-H)

1150, 1065, 1024 as(C-O-C) ands(C-O-C)

 **(cm-1)** 

(O-H) for H-O-H (HOH) for H-O-H 1118 and 980 (Si-O) out of plane (AlAlOH) (AlFeOH) (AlMgOH)

Fig. 5. IR spectra of Fig. 4 in the 1800–1400 cm-1 wavenumber range of chitosan film (CS), chitosan/MMT and chitosan/BNT films prepared from 1:1, 5:1 and 10:1 chitosan–clay ratios (CS1:MMT1; CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1; CS10:BNT1).

Application of Infrared Spectroscopy to Analysis of Chitosan/Clay Nanocomposites 55

0 2 4 6 8 10 12

MMT

CS1:MMT1 (1)

CS1:MMT1 (2)

CS1:MMT1 (3)

BNT

CS1:BNT1 (1)

CS1:BNT1 (2)

CS1:BNT1 (3)

0 2 4 6 8 10 12

0 2 4 6 8 10 12

0 2 4 6 8 10 12

0 2 4 6 8 10 12

0 2 4 6 8 10 12

0 2 4 6 8 10 12

0 2 4 6 8 10 12

Fig. 7. XRD pattern of sodium montmorillonite (MMT), purified sodium bentonite (BNT), chitosan/MMT and chitosan/BNT films prepared from 1:1 chitosan/clay ratios in triplicate [CS1:MMT1 (1), CS1:MMT1 (2), CS1:MMT1 (3) and CS1:BNT1 (1), CS1:BNT1 (2), CS1:BNT1

CS1:MMT1 (3)]. The shift of the basal reflection of MMT to lower angle indicates the formation of an intercalated nanostructure, while the peak broadening and intensity decreases most likely indicate the disordered intercalated or exfoliated structure (Utracki, 2004; Wang et al., 2005). Similar behavior was observed for CS/BNT (Fig. 7), i.e. the basal plane of BNT at 2θ = 6.3° disappears, substituted by a new weakened broad peak at around 2θ = 3.0° - 3.9° [CS1:BNT1 (1), CS1:BNT1 (2) and CS1:BNT1 (3)]. It is suggested that the MMT

2(°)

Intensity (a.u.)

d001=3,15

d001=2,82

d001= 1,40

d001=2,22

d001=2,24

and the BNT form intercalated and flocculated structures.

d001= 1,50

d001 =2,34

(3)].

Fig. 6. Schematic illustration of the intercalation of chitosan layers into the clay inter-layer spacing for films (a) with the lowest amounts of chitosan (CS1:MMT1 and CS1:BNT1) and (b) with the highest amounts of biopolymer (CS10:MMT1, CS5:MMT1 and CS10:BNT1, CS5:BNT1).

#### **3.2 X ray diffraction analysis (XRD)**

XRD is the principal method that has been used to examine the distribution/dispersion of the clay platelet stacks in the polymer matrix (Utracki, 2004). Depending on the relative distribution/dispersion of the stacks, three types of nanocomposites can be described: *intercalated nanocomposites*, where polymer chains are interleaved with silicate layers, resulting in a well ordered mutilayer morphology built up with alternating polymer and inorganic sheets; *flocculated nanocomposites*, where intercalated clay layers are sometimes bonded by hydroxylated edge-edge interactions, and *exfoliated/delaminated nanocomposites*, where individual clay layers are completely and homogenously dispersed in the polymer matrix (Wang et al., 2005).

FTIR data indicate that chitosan was intercalated into the MMT and BNT interlayers. However, to confirm the FTIR results, the MMT and BNT clays, as well as, chitosan/MMT and chitosan/BNT films prepared from 1:1, 5:1 and 10:1 chitosan/clay ratios, respectively (CS1:MMT1; CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1; CS10:BNT1) were analyzed by XRD and the results are shown in Figs. 7-9.

The XRD patterns of the MMT (Fig. 7) shows a reflection peak at about 2θ = 5.9°, corresponding to a basal spacing (d001) of 1.50 nm. After incorporating MMT within CS, with CS/MMT 1:1 ratio, the basal plane of MMT at 2θ = 5.9° disappears, substituted by a new weakened broad peak at around 2θ = 2.8° - 3.7° [CS1:MMT1 (1), CS1:MMT1 (2) and

Fig. 6. Schematic illustration of the intercalation of chitosan layers into the clay inter-layer spacing for films (a) with the lowest amounts of chitosan (CS1:MMT1 and CS1:BNT1) and (b) with the highest amounts of biopolymer (CS10:MMT1, CS5:MMT1 and CS10:BNT1,

XRD is the principal method that has been used to examine the distribution/dispersion of the clay platelet stacks in the polymer matrix (Utracki, 2004). Depending on the relative distribution/dispersion of the stacks, three types of nanocomposites can be described: *intercalated nanocomposites*, where polymer chains are interleaved with silicate layers, resulting in a well ordered mutilayer morphology built up with alternating polymer and inorganic sheets; *flocculated nanocomposites*, where intercalated clay layers are sometimes bonded by hydroxylated edge-edge interactions, and *exfoliated/delaminated nanocomposites*, where individual clay layers are completely and homogenously dispersed in the polymer

FTIR data indicate that chitosan was intercalated into the MMT and BNT interlayers. However, to confirm the FTIR results, the MMT and BNT clays, as well as, chitosan/MMT and chitosan/BNT films prepared from 1:1, 5:1 and 10:1 chitosan/clay ratios, respectively (CS1:MMT1; CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1; CS10:BNT1) were

The XRD patterns of the MMT (Fig. 7) shows a reflection peak at about 2θ = 5.9°, corresponding to a basal spacing (d001) of 1.50 nm. After incorporating MMT within CS, with CS/MMT 1:1 ratio, the basal plane of MMT at 2θ = 5.9° disappears, substituted by a new weakened broad peak at around 2θ = 2.8° - 3.7° [CS1:MMT1 (1), CS1:MMT1 (2) and

CS5:BNT1).

**3.2 X ray diffraction analysis (XRD)** 

analyzed by XRD and the results are shown in Figs. 7-9.

matrix (Wang et al., 2005).

Fig. 7. XRD pattern of sodium montmorillonite (MMT), purified sodium bentonite (BNT), chitosan/MMT and chitosan/BNT films prepared from 1:1 chitosan/clay ratios in triplicate [CS1:MMT1 (1), CS1:MMT1 (2), CS1:MMT1 (3) and CS1:BNT1 (1), CS1:BNT1 (2), CS1:BNT1 (3)].

CS1:MMT1 (3)]. The shift of the basal reflection of MMT to lower angle indicates the formation of an intercalated nanostructure, while the peak broadening and intensity decreases most likely indicate the disordered intercalated or exfoliated structure (Utracki, 2004; Wang et al., 2005). Similar behavior was observed for CS/BNT (Fig. 7), i.e. the basal plane of BNT at 2θ = 6.3° disappears, substituted by a new weakened broad peak at around 2θ = 3.0° - 3.9° [CS1:BNT1 (1), CS1:BNT1 (2) and CS1:BNT1 (3)]. It is suggested that the MMT and the BNT form intercalated and flocculated structures.

Application of Infrared Spectroscopy to Analysis of Chitosan/Clay Nanocomposites 57

Fig. 9 shows he XRD patterns of the MMT, BNT, chitosan/MMT and chitosan/BNT films prepared from 10:1 chitosan/clay ratios in triplicate [CS10:MMT1 (1), CS10:MMT1 (2), CS10:MMT1 (3) and CS10:BNT1 (1), CS10:BNT1 (2), CS10:BNT1 (3)]. With increasing CS content, the 2θ of (001) peak becomes lower and it is not possible to calculate the interlayer distance for each nanocomposite in the broad peaks, indicating that the MMT and the BNT forms intercalated and exfoliated structures. In all probability, exfoliated/delaminated

0 2 4 6 8 10 12

d001=1,50

MMT

CS10:MMT1 (1)

CS10:MMT1 (2)

CS10:MMT1 (3)

BNT

CS10:BNT1 (1)

CS10:BNT1 (2)

CS10:BNT1 (3)

0 2 4 6 8 10 12

0 2 4 6 8 10 12

0 2 4 6 8 10 12

d001 = 1,40

0 2 4 6 8 10 12

0 2 4 6 8 10 12

0 2 4 6 8 10 12

0 2 4 6 8 10 12

Fig. 9. XRD pattern of MMT, BNT, chitosan/MMT and chitosan/BNT films prepared from 10:1 chitosan/clay ratios in triplicate [CS10:MMT1 (1), CS10:MMT1 (2), CS10:MMT1 (3) and

2(°)

structures were obtained in this case.

Intensity (a.u.)

CS10:BNT1 (1), CS10:BNT1 (2), CS10:BNT1 (3)].

Fig. 8 shows he XRD patterns of the MMT, BNT, chitosan/MMT and chitosan/BNT films prepared from 5:1 chitosan/clay ratios in triplicate [CS5:MMT1 (1), CS5:MMT1 (2), CS5:MMT1 (3) and CS5:BNT1 (1), CS5:BNT1 (2), CS5:BNT1 (3)]. After incorporating MMT within CS, with CS/MMT 5:1 ratio, the basal plane of MMT at 2θ = 5.9° disappears, substituted by a new weakened broad peak at around 2θ = 2.0° - 3.8° [CS5:MMT1 (1), CS5:MMT1 (2) and CS5:MMT1 (3)]. In this case the 2θ values were smaller than the values observed for chitosan/MMT prepared from 1:1 ratio (Fig. 7), indicating that exfoliated/delaminated nanocomposites were be obtained. In the same way exfoliated/delaminated nanocomposites are probably obtained for Chitosan/BNT [CS5:BNT1 (1), CS5:BNT1 (2) and CS5:BNT1 (3)].

Fig. 8. XRD pattern of MMT, BNT, chitosan/MMT and chitosan/BNT films prepared from 5:1 chitosan/clay ratios in triplicate [CS5:MMT1 (1), CS5:MMT1 (2), CS5:MMT1 (3) and CS5:BNT1 (1), CS5:BNT1 (2), CS5:BNT1 (3)].

Fig. 8 shows he XRD patterns of the MMT, BNT, chitosan/MMT and chitosan/BNT films prepared from 5:1 chitosan/clay ratios in triplicate [CS5:MMT1 (1), CS5:MMT1 (2), CS5:MMT1 (3) and CS5:BNT1 (1), CS5:BNT1 (2), CS5:BNT1 (3)]. After incorporating MMT within CS, with CS/MMT 5:1 ratio, the basal plane of MMT at 2θ = 5.9° disappears, substituted by a new weakened broad peak at around 2θ = 2.0° - 3.8° [CS5:MMT1 (1), CS5:MMT1 (2) and CS5:MMT1 (3)]. In this case the 2θ values were smaller than the values observed for chitosan/MMT prepared from 1:1 ratio (Fig. 7), indicating that exfoliated/delaminated nanocomposites were be obtained. In the same way exfoliated/delaminated nanocomposites are probably obtained for Chitosan/BNT

0 2 4 6 8 10 12

d001=1,50

MMT

CS5:MMT1 (1)

CS5:MMT1 (2)

CS5:MMT1 (3)

BNT

CS5:BNT1 (1)

CS5:BNT1 (3)

0 2 4 6 8 10 12

0 2 4 6 8 10 12

2 4 6 8 10 12

d001=1,40

CS5:BNT1 (2) d001=1,89

d001=2,28

0 2 4 6 8 10 12

0 2 4 6 8 10 12

2 4 6 8 10 12

2 4 6 8 10 12

Fig. 8. XRD pattern of MMT, BNT, chitosan/MMT and chitosan/BNT films prepared from 5:1 chitosan/clay ratios in triplicate [CS5:MMT1 (1), CS5:MMT1 (2), CS5:MMT1 (3) and

2(°)

[CS5:BNT1 (1), CS5:BNT1 (2) and CS5:BNT1 (3)].

Intensity (a.u.)

CS5:BNT1 (1), CS5:BNT1 (2), CS5:BNT1 (3)].

Fig. 9 shows he XRD patterns of the MMT, BNT, chitosan/MMT and chitosan/BNT films prepared from 10:1 chitosan/clay ratios in triplicate [CS10:MMT1 (1), CS10:MMT1 (2), CS10:MMT1 (3) and CS10:BNT1 (1), CS10:BNT1 (2), CS10:BNT1 (3)]. With increasing CS content, the 2θ of (001) peak becomes lower and it is not possible to calculate the interlayer distance for each nanocomposite in the broad peaks, indicating that the MMT and the BNT forms intercalated and exfoliated structures. In all probability, exfoliated/delaminated structures were obtained in this case.

Fig. 9. XRD pattern of MMT, BNT, chitosan/MMT and chitosan/BNT films prepared from 10:1 chitosan/clay ratios in triplicate [CS10:MMT1 (1), CS10:MMT1 (2), CS10:MMT1 (3) and CS10:BNT1 (1), CS10:BNT1 (2), CS10:BNT1 (3)].

Application of Infrared Spectroscopy to Analysis of Chitosan/Clay Nanocomposites 59

The intercalation of the cationic biopolymer chitosan into layered silicate clays (montmorillonite and bentonite) through a cation exchange process results in nanocomposites with interesting structural and functional properties. The clay reduces the film-forming capability of chitosan leading to compact, robust, and handy threedimensional nanocomposites. The techniques employed in the characterization of the nanocomposites, infrared spectroscopy, x-ray diffraction, and mechanical properties in tension, confirm the high affinity between the clay substrate and the biopolymer, as well as the special arrangement of chitosan as a bilayer when the biopolymer amount exceeds the cation exchange capacity of the clay. The intercalation of the first layer of chitosan takes place mainly by electrostatic interactions between positive ammonium groups in the chitosan chain and negative sites in the clay. In contrast, hydrogen bonds between amino and hydroxyl groups of chitosan and the clay substrate are established in the adsorption of

This research was financially supported by CAPES, CNPq and INAMI (Brazil). We thank Prof. Heber Carlos Ferreira (Department of Materials Engineering, Federal University of Campina Grande) for graciously proving sodium montmorillonite (Cloisite Na+) samples.

An JH., Dultz S. (2007). Adsorption of tannic acid on chitosan–montmoril-lonite as a

Araujo, P. E. R.; Araújo, S. S.; Raposo, C. M. O.; Silva, S. M. L. (2007). 23th Polymer

Awad, W. H.; Gilman, J. W.; Nyden, M.; Harris, R. H.; Sutto, T. E.; Callahan, J.; Trulove, P.

Bolto, B.; Dixon, D.; Eldridge, R.; King, S. (2001). Cationic polymer and clay or metal oxide

Bora, M.; Ganguli, J. N.; Dutta, D. K. (2000). Thermal and spectroscopic studies on the

Chivrac, F.; Pollet, E.; Avérous, L. (2009). Progress in nano biocomposites based on

Darder, M.; Colilla, M.; Ruiz-hitzky, E. (2003). Biopolymer-Clay Nanocomposites Based on

Darder, M.; Colilla, M.; Ruiz-Hitzky, E. (2005). Chitosan clay nanocomposites: application as

Electrochemical sensors. *Applied Clay Science*. Vol.28, pp.199-208.

Processing. Society Annual Meeting. Vol. 23, pp.1.

function of pH and surface charge properties. *Applied Clay Science*. Vol. 36, pp. 256–

C.; Delong, H. C.; Fox, D. M. (2004). Thermal degradation studies of alkylimidazolium salts and their application in nanocompósitos. *Thermochimica Acta*.

combinations for natural organic matter removal. *Water Research.*Vol. 35, p. 2669-

decomposition of [Ni{di(2-aminoethyl)amine}2]- and [Ni(2,2":6",2"-terpyridine)2]- Montmorillonite intercalated composites *Thermochimica Acta*. Vol. 346, p.169-175. Chang, M. Y.; Juang, R. S. (2004). Adsorption of tannic acid, humic acid and dyes from water

using the composite of chitosan and activated clay. Journal of Colloid and Interface

polysaccharides and nanoclays. *Materials Science and Engineering R*. Vol. 67, pp.

Chitosan Intercalated in Montmorillonite. *ChemicalMaterials*. vol. 15, pp. 3774–3780.

the second layer.

**6. References** 

64.

2676.

1–17.

Vol. 409, p. 3-11.

Science. Vol. 278, pp.18–25.

**5. Acknowledgments** 

In summary, the morphology of the nanocomposites was affected by chitosan/clay rations. On the base of XRD patterns, it is suggested that the MMT and the BNT forms intercalated and exfoliated structures at higher CS content (CS10:MMT1, CS5:MMT1 and CS10:BNT1, CS5:BNT1), while decreasing the CS content (CS1:MMT1 and CS1:BNT1), clay layers (MMT and BNT) form intercalated and flocculated structures. According Wang et al., 2005, the formation of flocculated structure in CS/clay nanocomposites can be due to the hydroxylated edge-edge interactions of the clay layers. Since one chitosan unit possesses one amino and two hydroxyl functional groups, these groups can form hydrogen bonds with the clay hydroxyl edge groups, which leads to the strong interactions between matrix and clay layers (Fig.6a) and corroborate FTIR results. This strong interaction is believed to be the main driving force for the assembly of MMT and BNT in the CS matrix to form flocculated structures.

### **3.3 Mechanical properties**

Tensile properties of the chitosan film (CS) and chitosan/clay films prepared from 1:1, 5:1 and 10:1 chitosan/clay ratios, respectively (CS1:MMT1; CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1; CS10:BNT1) are collected in Table 3. The tensile strength (TS) and elastic modulus (EM) of chitosan films increase by the formation of nanocomposites, particularly for chitosan/clay prepared from 5:1 rations. The increase in the TS and EM of such nanocomposite films can be attributed to the high rigidity and aspect ratio of the nanoclay as well as the high affinity between the biopolymer and the clay. On the other hand, the chitosan/clay nanocomposites have shown significant decrease in elongation at break (EB). This reduction can be attributed to the restricted mobility of macromolecular chains.


(TS = tensile strength, EM = elastic modulus (EM), EB = elongation at break)

Table 3. Tensile properties of chitosan and chitosan/clay films.

#### **4. Conclusions**

In this study chitosan/clay nanocomposites were successfully prepared by the solution intercalation process. It was found that clay dispersion is affected by the kind of clay and the chitosan/clay ratio. Since the nanocomposites prepared with purified bentonite (BNT) showed similar behavior to that prepared with montmorillonite, less expensive bentonite may be employed in the preparation of chitosan/clay nanocomposites.

The intercalation of the cationic biopolymer chitosan into layered silicate clays (montmorillonite and bentonite) through a cation exchange process results in nanocomposites with interesting structural and functional properties. The clay reduces the film-forming capability of chitosan leading to compact, robust, and handy threedimensional nanocomposites. The techniques employed in the characterization of the nanocomposites, infrared spectroscopy, x-ray diffraction, and mechanical properties in tension, confirm the high affinity between the clay substrate and the biopolymer, as well as the special arrangement of chitosan as a bilayer when the biopolymer amount exceeds the cation exchange capacity of the clay. The intercalation of the first layer of chitosan takes place mainly by electrostatic interactions between positive ammonium groups in the chitosan chain and negative sites in the clay. In contrast, hydrogen bonds between amino and hydroxyl groups of chitosan and the clay substrate are established in the adsorption of the second layer.

#### **5. Acknowledgments**

This research was financially supported by CAPES, CNPq and INAMI (Brazil). We thank Prof. Heber Carlos Ferreira (Department of Materials Engineering, Federal University of Campina Grande) for graciously proving sodium montmorillonite (Cloisite Na+) samples.

#### **6. References**

58 Infrared Spectroscopy – Materials Science, Engineering and Technology

In summary, the morphology of the nanocomposites was affected by chitosan/clay rations. On the base of XRD patterns, it is suggested that the MMT and the BNT forms intercalated and exfoliated structures at higher CS content (CS10:MMT1, CS5:MMT1 and CS10:BNT1, CS5:BNT1), while decreasing the CS content (CS1:MMT1 and CS1:BNT1), clay layers (MMT and BNT) form intercalated and flocculated structures. According Wang et al., 2005, the formation of flocculated structure in CS/clay nanocomposites can be due to the hydroxylated edge-edge interactions of the clay layers. Since one chitosan unit possesses one amino and two hydroxyl functional groups, these groups can form hydrogen bonds with the clay hydroxyl edge groups, which leads to the strong interactions between matrix and clay layers (Fig.6a) and corroborate FTIR results. This strong interaction is believed to be the main driving force

Tensile properties of the chitosan film (CS) and chitosan/clay films prepared from 1:1, 5:1 and 10:1 chitosan/clay ratios, respectively (CS1:MMT1; CS5:MMT1; CS10:MMT1 and CS1:BNT1; CS5:BNT1; CS10:BNT1) are collected in Table 3. The tensile strength (TS) and elastic modulus (EM) of chitosan films increase by the formation of nanocomposites, particularly for chitosan/clay prepared from 5:1 rations. The increase in the TS and EM of such nanocomposite films can be attributed to the high rigidity and aspect ratio of the nanoclay as well as the high affinity between the biopolymer and the clay. On the other hand, the chitosan/clay nanocomposites have shown significant decrease in elongation at break (EB).

for the assembly of MMT and BNT in the CS matrix to form flocculated structures.

This reduction can be attributed to the restricted mobility of macromolecular chains.

Chitosan film (CS) 44.5 + 4.5 1774 + 63 7.7 + 0.5

Chitosan/MMT (CS1:MMT1) 84.9 + 3.7 5214 + 112 3.3 + 0.5 Chitosan/MMT (CS5:MMT1) 79.1 + 1.1 4449+ 329 4.6 + 0.7 Chitosan/MMT (CS10:MMT1) 68.5 + 1.4 3536 + 180 4.6 + 0.8

Chitosan/BNT (CS1:BNT1) 49.6 + 4.9 4075 + 73 2.4 + 0.9 Chitosan/BNT (CS5:BNT1) 62.1 + 4.5 3106 + 50 6.8 + 0.8 Chitosan/BNT (CS10:BNT1) 40.4 + 1.8 2421 + 87 5.9 + 0.8

(TS = tensile strength, EM = elastic modulus (EM), EB = elongation at break)

may be employed in the preparation of chitosan/clay nanocomposites.

Table 3. Tensile properties of chitosan and chitosan/clay films.

In this study chitosan/clay nanocomposites were successfully prepared by the solution intercalation process. It was found that clay dispersion is affected by the kind of clay and the chitosan/clay ratio. Since the nanocomposites prepared with purified bentonite (BNT) showed similar behavior to that prepared with montmorillonite, less expensive bentonite

**Sample TS (MPa)** EM (MPa) EB (%)

**3.3 Mechanical properties** 

**4. Conclusions** 


Application of Infrared Spectroscopy to Analysis of Chitosan/Clay Nanocomposites 61

Pandey, S. Mishra S. B., (2011). Organic–inorganic hybrid of chitosan/organoclay

Paluszkiewicz, C.; Stodolak, E.; Hasik, M.; Blazewicz, M. (2011). FT-IR study of

Pongjanyakul, T.; Priprem, A.; Puttipipatkhachorn, S. (2005). Investigation of novelalginate-

Pongjanyakul, T.; Suksri, H. (2009). Alginate –magnesium aluminum silicate lms forbuccal delivery of nicotine. *Colloids and Surfaces B: Biointerfaces.* Vol. 74, pp. 103–113. Rinaudo, M. (2006). Chitin and chitosan: Properties and applications. *Progress in Polymer* 

Tan, W.; Zhang, Y.; Szeto, Y.; Liao, L. (2007). A novel method to prepare

Utracki, L. A. (2004). Basic Elements of Polymeric Nanocomposite Technology, In: *Clay-*

Wan Ngah, W. S.; Ariff, N. F. M.; Hanaah, M. A. K. M. (2010).Preparation, characterization,

Wan Ngah, W. S.; Teong, L. C.; Hanaah, M. A. K. M. (2011). Adsorption of dyes and heavy

Wang, L.; Wang, A. (2007). Adsorption characteristics of Congo Red onto the

Wang, S. F.; Shen, L.; Tong, Y. J.; Chen, L.; Phang, I. Y.; Lim, P. Q. (2005). Biopolymer

Wang, X.; Du, Y.; Yang, J.; Wang, X.; Shi, X.; Hu Y. (2006). Preparation, characterization and

Wei, J. M.; Zhu, R. L.; Zhu, J. X.; Ge, F.; Yuan, P.; He, H. P. (2009). Simultaneous sorption of

Wu, T. M.; Wu, C. Y. (2006). Biodegradable poly (lactic acid)/chitosan-modied

Xu, X.; Ding, Y.; Qian, Z.; Wang, F.; Wen, B.; Zhou, H.; Zhang, S.; Yang, M. (2009).

*Science*. Vol. 361, pp. 509–520.

*Science*.Vol. 31, pp. 603-632.

England.

pp.225–236.

1446–1456.

pp. 979–985.

pp. 6738-6744.

*Materials*. Vol. 166, pp.195–199.

*Stability*. Vol.94, pp.113-123.

*Journal Control. Release*. Vol. 107, pp. 343– 356.

oligomeric cations. *Composites Science and Technology*.

*Polymer Degradation and Stability*. Vol. 90, pp.123-131.

Degradation and Stability. Vol. 91, pp. 2198–2204.

Vol. 79, pp. 784–788.

bionanocomposites for hexavalent chromium uptake, *Journal of Colloid and Interface* 

montmorillonite–chitosan nanocomposite materials, *Spectroc*hi*mica Acta Part A*.

magnesium aluminum silicate microcomposite lms for modi ed-releasetablets.

chitosan/montmorillonite nanocompósitos in the presence of hydroxyl-aluminum

*Containing Polymeric Nanocomposites*, pp. 73-96, Rapra Technology Limited,

and environmental application of crosslinked chitosan-coated bentonite for tartrazine adsorption from aqeous solutions. *Water, Air and Soil Pollution*. Vol. 206,

metal ions by chitosan composites: A review. *Carbohydrate Polymers*. Vol. 83, pp.

chitosan/montmorillonite nanocomposite. *Journal of Hazardous Materials*. Vol. 147,

chitosan/montmorillonite nanocomposites: Preparation and characterization.

antimicrobial activity of chitosan/layered silicate nanocompósitos. *Polymer*. Vol. 47,

crystal violet and 2-naphthol to bentonite with different CECs. *Journal of Hazardous* 

montmorillonite nanocomposites: preparation and characterization. Polymer

Degradation of poly(ethylene terephthalate)/clay nanocomposites during melt extrusion: Effect of clay catalysis and chain extension. *Polymer Degradation and* 


Dash, M.; Chiellini, F.; Ottenbrite, R.M.; Chiellini E. (2011). Chitosan—A versatile semi-

E. S. Costa; H. S. Mansur. (2008). Preparação e caracterização de blendas de quitosana/poli

Felt, O.; Carrel, A.; Baehni, P.; Buri, P.; Gurny, R. (2000) Chitosan as tear substitute: a

Gecol, H.; Miakatsindila, P.; Ergican, E.; Hiibel, S. R. (2006). Biopolymer coated clay particles for the adsorption of tungsten from water. *Desalination*. Vol. 197, pp. 165–78. Günister, E., Pestreli, D., Ünlü, C.H.; Güngör, N. (2007). Synthesis and characterization of chitosan–MMT biocomposite systems. *Carbohydrate Polymers*. Vol. 67, pp.358-365. Han, Y.; Lee, S.; Choi, K. H. (2010). Preparation andcharacterizationofchitosan–

Holzer, L.; Münch, B.; Rizzi, M.; Wepf, R.; Marschall, P.; Graule, T. (2010). 3D-microstructure

Khunawattanakul , W.; Puttipipatkhachorn , S.; Rades , T.; Pongjanyakul, T. (2008).

Krajewska, B. (2004). Application of chitin and chitosan-based materials for enzyme immobilizations: a review. *Enzyme and Microbial Technology*. Vol. 35, pp. 126-134. Lavorgna, M.; Piscitelli, F.; Mangiacapra, P.; Buonocore, G. (2010). Study of the combined

Leite, I. F.; Soares, A. P. S.; Carvalho, L.H.; Malta, O. M. L.; Raposo, C. M. O.; Silva, S. M. S.

Li, Q.; Yue, Q. Y.; Sun, H. J.; Su, Y.; Gao, B. Y. (2010). A comparative study on the properties,

Madejová, J. (2003). FTIR techniques in clay mineral studies. *Vibrational Spectroscopy*. Vol.31,

Mano, J.F.; Koniarova, D.; Reis, R.L. (2003).*Journal of Materials Science: Materials in Medicine*.

Marchessault, R.H.; Ravenelle, F.; Zhu, X.X. (2006). Polysaccharides for drug delivery and

No, H. K.; Lee, K. S.; Meyers, S. P. (2000). Correlation between physicochemical

characteristics and binding capacities of chitosan products. *Journal Food Science*. Vol.

engenharia de tecido. *Química Nova*. Vol.31, pp.1460–1466.

*and Therapeutics*. Vol. 16, pp. 261–270.

*Carbohydrate Polymers*. Vol. 82, pp. 291–298.

*Thermal Analysis and Calorimetry*. Vol. 100, pp.563.

pharmaceutical applications, *American Chemical Society*.

*Solids*. Vol. 71, pp. 464–467.

Vol. 47, pp. 330–342.

Vol.351, pp. 227–235.

1611.

pp.1-10.

Vol. 14, pp.127–135.

65, pp. 1134–1137.

pp. 981–1014.

synthetic polymer in biomedical applications. *Progress in Polymer Science*, Vol. 36,

(álcool vinílico) reticuladas quimicamente com glutaraldeído para aplicação em

wetting agent endowed with antimicrobial efficacy. *Journal of Ocular Pharmacology* 

claynanocompositeswith antimicrobial activity. *Journal of Physics and Chemistry of* 

analysis of hydrated bentonite with cryo-stabilized pore water. *Applied Clay Science*.

Chitosan–magnesium aluminum silicate composite dispersions: characterization of rheology, occulate size and zeta potential. *International Journal of Pharmaceutics*.

effect of both clay and glycerol plasticizer on the properties of chitosan lms.

(2010). Characterization of pristine and purified organobentonites. *Journal of* 

mechanism and process designs for the adsorption of non-ionic or anionicdyes onto cationic-polymer/bentonite. *Journal of Environmental Management*. Vol. 91, pp.1601–


**3** 

*Romania* 

**Structural and Optical Behavior of Vanadate-**

*1Department of Physics and Chemistry, Technical University of Cluj-Napoca, Cluj-Napoca* 

Tellurium oxide based glasses are of scientific and technological interest due to their unique properties such as chemical durability, electrical conductivity, transmission capability, high

Tellurate glasses have recently gained wide attention because of their potential as hosts of rare earth elements for the development of fibres and lasers covering all the main telecommunication bands and promising materials for optical switching devices [4, 5]. Recently, tellurate glasses doped with heavy metal oxides or rare earth oxides have received great scientific interest because these oxides can change the optical and physical properties

 Vanadium tellurate glasses showed better mechanical and electrical properties due to the V2O5 incorporated into the tellurate glass matrix. In the case of V2O5 contents below 20mol%, the three-dimensional tellurate network is partially broken by the formation of [TeO3] trigonal pyramidal units, which in turn reduce the glass rigidity and. When the V2O5 concentration is above 20mol%, the glass structure changes from the continuous tellurate

Due to the large atomic mass and high polarizability of the Pb+2 ions, heavy metal oxide glasses with PbO possess high refractive index, wide infrared transmittance, and hence they are considered to be promising glass hosts for photonic devices [7, 8]. The special significance of PbO is that it contributes to form stable glasses over a wide range of

Rare-earth ions doped glasses have been prepared and characterized to understand their commercial applications as glass lasers and also in the production of wide variety of other

The luminescence spectral properties of rare earth ions such as Eu+3 (4f6) and Tb+3 (4f 8), Sm+3 (4f5) and Dy+3 (4f9) in the heavy-metal borate glasses have shown interesting and encouraging results [10, 11]. Eu+3 and Tb+3 ions have shown prominent emissions (red and green) in the visible wavelength region, while Sm+3 and Dy+3 show strong absorption bands

refractive indices, high dielectric constant and low melting points [1-3].

concentrations due to its dual role as glass modifier and glass former.

**1. Introduction** 

of the tellurate glasses [5].

types of optical components [9].

network to the continuous vanadate network [6].

**Tellurate Glasses Containing PbO or Sm2O3**

*3National Institute for R&D of Isotopic and Molecular Technologies, Cluj-Napoca* 

*2Faculty of Physics, Babes-Bolyai University of Cluj-Napoca, Cluj-Napoca* 

E. Culea1, S. Rada1, M. Culea2 and M. Rada3


## **Structural and Optical Behavior of Vanadate-Tellurate Glasses Containing PbO or Sm2O3**

E. Culea1, S. Rada1, M. Culea2 and M. Rada3 *1Department of Physics and Chemistry, Technical University of Cluj-Napoca, Cluj-Napoca 2Faculty of Physics, Babes-Bolyai University of Cluj-Napoca, Cluj-Napoca 3National Institute for R&D of Isotopic and Molecular Technologies, Cluj-Napoca Romania* 

#### **1. Introduction**

62 Infrared Spectroscopy – Materials Science, Engineering and Technology

Xu, Y.; Kim, K.; Hanna, M.; Nag, D. (2005). Industrial Crops and Products. Vol. 21, pp. 185–

Yang, Y. Q.; Chen, H. J. (2007). Study on the intercalation organic bentonite and its adsorption. Journal of Xinyang Normal University. Vol. 20, pp. 338–340. Yuan, Q.; Shah, J.; Hein, S.; Misra, R.D.K. (2010). Controlled and extended drug release

Zhang, A. C.; Sun, L. S.; Xiang, J.; Hu, S.; Fu, P; Su, S. (2009). Removal of elemental mercury

*Fuel Chemistry and Technology*. Vol. 37, pp. 489–495.

behavior of chitosan-based nanoparticle carrier. *Acta Biomaterialia*. Vol. 6, pp. 1140–

from coal combustion flue gas by bentonite–chitosan and their modifier. *Journal of* 

192.

1148.

Tellurium oxide based glasses are of scientific and technological interest due to their unique properties such as chemical durability, electrical conductivity, transmission capability, high refractive indices, high dielectric constant and low melting points [1-3].

Tellurate glasses have recently gained wide attention because of their potential as hosts of rare earth elements for the development of fibres and lasers covering all the main telecommunication bands and promising materials for optical switching devices [4, 5]. Recently, tellurate glasses doped with heavy metal oxides or rare earth oxides have received great scientific interest because these oxides can change the optical and physical properties of the tellurate glasses [5].

 Vanadium tellurate glasses showed better mechanical and electrical properties due to the V2O5 incorporated into the tellurate glass matrix. In the case of V2O5 contents below 20mol%, the three-dimensional tellurate network is partially broken by the formation of [TeO3] trigonal pyramidal units, which in turn reduce the glass rigidity and. When the V2O5 concentration is above 20mol%, the glass structure changes from the continuous tellurate network to the continuous vanadate network [6].

Due to the large atomic mass and high polarizability of the Pb+2 ions, heavy metal oxide glasses with PbO possess high refractive index, wide infrared transmittance, and hence they are considered to be promising glass hosts for photonic devices [7, 8]. The special significance of PbO is that it contributes to form stable glasses over a wide range of concentrations due to its dual role as glass modifier and glass former.

Rare-earth ions doped glasses have been prepared and characterized to understand their commercial applications as glass lasers and also in the production of wide variety of other types of optical components [9].

The luminescence spectral properties of rare earth ions such as Eu+3 (4f6) and Tb+3 (4f 8), Sm+3 (4f5) and Dy+3 (4f9) in the heavy-metal borate glasses have shown interesting and encouraging results [10, 11]. Eu+3 and Tb+3 ions have shown prominent emissions (red and green) in the visible wavelength region, while Sm+3 and Dy+3 show strong absorption bands

Structural and Optical Behavior of Vanadate-Tellurate Glasses Containing PbO or Sm2O3 65

**10 20 30 40 50 60**

**2 theta [degree]**

Fig. 1. X-ray diffraction patterns for xPbO (or xSm2O3)·(100-x)[3TeO2·2V2O5] samples where

samples with 0≤x≤50mol% PbO while in the samples with x≥40mol% Sm2O3 the presence of

The absorption bands located around 460cm-1, 610-680 and 720 to 780cm-1 are assigned to the bending mode of Te-O-Te or O-Te-O linkages, the stretching mode of [TeO4] trigonal pyramids with bridging oxygen and the stretching mode of [TeO3] trigonal pyramids with

The IR spectrum of the pure crystalline and amorphous V2O5 is characterized by the intense band in the 1000-1020cm-1 range which is related to vibrations of isolated V=O vanadyl groups in [VO5] trigonal bipyramids. The band located at 950-970cm-1 was attributed to the

The examination of the FTIR spectra of the xMaOb(100-x)[3TeO22V2O5] where MaOb = PbO or Sm2O3 glasses and glass ceramics shows some changes in the characteristic bands corresponding to the structural units of the glass network (Fig. 2). These modifications can

**x=50% Sm2**

**x=40% Sm2**

**x=0-30% Sm2**

**x=0-50% PbO**

**O3**

**O3**

**O3**

**SmVO4**

the SmVO4 crystalline phase was detected.

non-bridging oxygen, respectively [14-18].

**intensity [a.u.]**

0≤x≤50mol%.

**3.1 FTIR spectroscopy** 

[VO4] units [19-21].

be summarized as follows: 1. x =10mol% MaOb (1)

in the NIR range (800–2200 nm) and intense emission bands in the visible region (550–730 nm) [11].

The present work deals with the role of lead and samarium ions in the short-range structural order of the vanadate-tellurate glass network. Lead and samarium-activated vanadatetellurate glasses have been investigated using infrared spectroscopy and ultraviolet-visible spectroscopy. The main goal is to obtain information about the influence of the radii and concentration of lead and samarium ions on the TeO4/TeO3 and VO5/VO4 conversion in vanadate-tellurate glasses and especially to illuminate aspects of the vanadate glass network using DFT calculations.

### **2. Experimental procedure**

Glasses were prepared by mixing and melting of appropriate amounts of lead (IV) oxide, tellurium oxide (IV), lead (II) oxide or samarium (III) oxide of high purity (99,99%, Aldrich Chemical Co.). Reagents were melted at 8750C for 10minutes and quenched by pouring the melts on stainless steel plates.

The samples were analyzed by means of X-ray diffraction using a XRD-6000 Shimadzu diffractometer, with a monochromator of graphite for the Cu-Kα radiation (λ=1.54Å) at room temperature.

The FT-IR absorption spectra of the glasses in the 370-1100cm-1 spectral range were obtained with a JASCO FTIR 6200 spectrometer using the standard KBr pellet disc technique. The spectra were carried out with a standard resolution of 2cm-1.

UV-Visible absorption spectra measurements in the wavelength range of 250-1050nm were performed at room temperature using a Perkin-Elmer Lambda 45 UV/VIS spectrometer equipped with an integrating sphere. These measurements were made on glass powder dispersed in KBr pellets. The optical absorption coefficient, α, was calculated from the absorbance, A, using the equation:

$$\mathbf{a} = 2.303 \text{ A/d}$$

where d is the thickness of the sample.

The starting structures have been built using the graphical interface of Spartan'04 [12] and preoptimized by molecular mechanics. Optimizations were continued at DFT level (B3LYP/CEP-4G/ECP) using the Gaussian'03 package of programs [13].

It should be noticed that only the broken bonds at the model boundary were terminated by hydrogen atoms. The positions of boundary atoms were frozen during the calculation and the coordinates of internal atoms were optimized in order to model the active fragment flexibility and its incorporation into the bulk.

#### **3. Results and discussion**

The vitreous or/and crystalline nature of the xPbO·(100-x)(3TeO2·2V2O5) and xSm2O3·(100 x)(3TeO2·2V2O5) samples with various contents of lead or samarium oxide (0≤x≤50mol%) was tested by X-ray diffraction. The X-ray diffraction patterns of the studied samples are shown in Fig. 1. The X-ray diffraction patterns did not reveal the crystalline phases in the

Fig. 1. X-ray diffraction patterns for xPbO (or xSm2O3)·(100-x)[3TeO2·2V2O5] samples where 0≤x≤50mol%.

samples with 0≤x≤50mol% PbO while in the samples with x≥40mol% Sm2O3 the presence of the SmVO4 crystalline phase was detected.

#### **3.1 FTIR spectroscopy**

64 Infrared Spectroscopy – Materials Science, Engineering and Technology

in the NIR range (800–2200 nm) and intense emission bands in the visible region (550–730

The present work deals with the role of lead and samarium ions in the short-range structural order of the vanadate-tellurate glass network. Lead and samarium-activated vanadatetellurate glasses have been investigated using infrared spectroscopy and ultraviolet-visible spectroscopy. The main goal is to obtain information about the influence of the radii and concentration of lead and samarium ions on the TeO4/TeO3 and VO5/VO4 conversion in vanadate-tellurate glasses and especially to illuminate aspects of the vanadate glass network

Glasses were prepared by mixing and melting of appropriate amounts of lead (IV) oxide, tellurium oxide (IV), lead (II) oxide or samarium (III) oxide of high purity (99,99%, Aldrich Chemical Co.). Reagents were melted at 8750C for 10minutes and quenched by pouring the

The samples were analyzed by means of X-ray diffraction using a XRD-6000 Shimadzu diffractometer, with a monochromator of graphite for the Cu-Kα radiation (λ=1.54Å) at

The FT-IR absorption spectra of the glasses in the 370-1100cm-1 spectral range were obtained with a JASCO FTIR 6200 spectrometer using the standard KBr pellet disc technique. The

UV-Visible absorption spectra measurements in the wavelength range of 250-1050nm were performed at room temperature using a Perkin-Elmer Lambda 45 UV/VIS spectrometer equipped with an integrating sphere. These measurements were made on glass powder dispersed in KBr pellets. The optical absorption coefficient, α, was calculated from the

α = 2.303 A/d

The starting structures have been built using the graphical interface of Spartan'04 [12] and preoptimized by molecular mechanics. Optimizations were continued at DFT level

It should be noticed that only the broken bonds at the model boundary were terminated by hydrogen atoms. The positions of boundary atoms were frozen during the calculation and the coordinates of internal atoms were optimized in order to model the active fragment

The vitreous or/and crystalline nature of the xPbO·(100-x)(3TeO2·2V2O5) and xSm2O3·(100 x)(3TeO2·2V2O5) samples with various contents of lead or samarium oxide (0≤x≤50mol%) was tested by X-ray diffraction. The X-ray diffraction patterns of the studied samples are shown in Fig. 1. The X-ray diffraction patterns did not reveal the crystalline phases in the

(B3LYP/CEP-4G/ECP) using the Gaussian'03 package of programs [13].

spectra were carried out with a standard resolution of 2cm-1.

nm) [11].

using DFT calculations.

**2. Experimental procedure** 

melts on stainless steel plates.

absorbance, A, using the equation:

where d is the thickness of the sample.

flexibility and its incorporation into the bulk.

**3. Results and discussion** 

room temperature.

The absorption bands located around 460cm-1, 610-680 and 720 to 780cm-1 are assigned to the bending mode of Te-O-Te or O-Te-O linkages, the stretching mode of [TeO4] trigonal pyramids with bridging oxygen and the stretching mode of [TeO3] trigonal pyramids with non-bridging oxygen, respectively [14-18].

The IR spectrum of the pure crystalline and amorphous V2O5 is characterized by the intense band in the 1000-1020cm-1 range which is related to vibrations of isolated V=O vanadyl groups in [VO5] trigonal bipyramids. The band located at 950-970cm-1 was attributed to the [VO4] units [19-21].

The examination of the FTIR spectra of the xMaOb(100-x)[3TeO22V2O5] where MaOb = PbO or Sm2O3 glasses and glass ceramics shows some changes in the characteristic bands corresponding to the structural units of the glass network (Fig. 2). These modifications can be summarized as follows:

1. x =10mol% MaOb (1)

Structural and Optical Behavior of Vanadate-Tellurate Glasses Containing PbO or Sm2O3 67

18]. The [TeO3] structural units are expected to participate in the depolymerization of the

The intensity of the band corresponding to the [TeO3] units (located at about 750cm-1) decreases and a new band located at about 800cm-1 appears with the adding of the Sm2O3 content. In detail, the band situated at about 800cm-1 is due to the asymmetric stretching of

An increasing trend was observed in the strength of the bands centered at ~1020cm-1. The feature of the band located at about 875cm-1 comes up in intensity. This effect is more pronounced when adding of samarium ions in the matrix network. This band is attributed

The gradual addition of the samarium (III) oxide leads not only to a simple incorporation of these ions in the host glass matrix but also generates changes of the basic structural units of the glass matrix. Structural changes reveal that the samarium ions causes a change from the continuous vanadate-tellurate network to a continuous samarium-vanadate-tellurate network interconnected through Sm-O-V and Te-O-Sm bridges. Then, the surplus of nonbridging oxygens is be converted to bridging ones leading to the decrease of the

By increasing the Sm2O3 content up to 40mol%, the evolution of the structure can be explained considering the higher capacity of migration of the samarium ions inside the glass network and the formation of the SmVO4 crystalline phase, in agreement with XRD data. The accumulation of oxygen atoms in the glass network can be supported by the formation

On the other hand, the lead oxide generates the rapid deformation of the Te-O-Te linkages yielding the formation of [TeO3] structural units. Further, the excess of oxygen can be accommodated in the host matrix by conversion of some [VO4] structural units into [VO5]

The broader band centered at ~ 670-850cm-1 can be attributed to the Pb-O bonds vibrations from the [PbO4] and [PbO3] structural units. The absorption band centered at about 470cm−<sup>1</sup> may be correlated with the Pb-O stretching vibration in [PbO4] structural units [24-26]. The increase in the intensity of the bands situated between 650 and 850cm-1 show that the excess of oxygen in the glass network can be supported by the increase of [PbOn] structural units

In brief, the variations observed in the FTIR spectra suggest a gradual inclusion of the lead ions in the host vitreous matrix with increasing of the PbO content up to 50mol%, while progressive adding of samarium oxide determines the increase in the intensity of the bands due to the ortho- and pyrovanadate structural units. The mechanisms of incorporation of the lead and samarium ions in the host matrices can be summarized as

glass network because they create more bonding defects and non-bonding oxygens.

to the vibrations of the V-O bonds from the pyrovanadate structural units.

VO4-3 entity from orthovanadate species [22, 23].

2. 10 ≤x ≤30 mol% MaOb (2)

connectivity of the network.

of ortho- and pyro-vanadate structural units.

3. x ≥ 40mol% MaOb (3)

structural units.

(with n=3 and 4).

following:

Fig. 2. FTIR spectra of xPbO (or xSm2O3)·(100-x)[3TeO2·2V2O5] samples where 0≤x≤50mol%.

The incorporation of network modifier lead ions into the vanadate-tellurate glasses enhances the breaking of axial Te-O-Te linkages in the [TeO4] trigonal bypiramidal structural units. As a consequence, three-coordination tellurium is formed and accumulated. The band centered at about 750cm-1 indicates the presence of the [TeO3] structural units [17, 18]. The [TeO3] structural units are expected to participate in the depolymerization of the glass network because they create more bonding defects and non-bonding oxygens.

The intensity of the band corresponding to the [TeO3] units (located at about 750cm-1) decreases and a new band located at about 800cm-1 appears with the adding of the Sm2O3 content. In detail, the band situated at about 800cm-1 is due to the asymmetric stretching of VO4 -3 entity from orthovanadate species [22, 23].

2. 10 ≤x ≤30 mol% MaOb (2)

66 Infrared Spectroscopy – Materials Science, Engineering and Technology

**x=50%**

**x=40%**

**absorbance [a.u.]**

**x=30%**

**x=10%**

**x=20%**

**absorbance [a.u.]**

**absorbance [a.u.]**

**400 500 600 700 800 900 1000 1100 wavenumber [cm-1**

**400 500 600 700 800 900 1000 1100**

**400 500 600 700 800 900 1000 1100**

**]**

**wavenumber [cm-1**

Fig. 2. FTIR spectra of xPbO (or xSm2O3)·(100-x)[3TeO2·2V2O5] samples where 0≤x≤50mol%.

The incorporation of network modifier lead ions into the vanadate-tellurate glasses enhances the breaking of axial Te-O-Te linkages in the [TeO4] trigonal bypiramidal structural units. As a consequence, three-coordination tellurium is formed and accumulated. The band centered at about 750cm-1 indicates the presence of the [TeO3] structural units [17,

400 500 600 700 800 900 1000 1100

400 500 600 700 800 900 1000 1100

**]**

**Sm**

**Pb**

An increasing trend was observed in the strength of the bands centered at ~1020cm-1. The feature of the band located at about 875cm-1 comes up in intensity. This effect is more pronounced when adding of samarium ions in the matrix network. This band is attributed to the vibrations of the V-O bonds from the pyrovanadate structural units.

The gradual addition of the samarium (III) oxide leads not only to a simple incorporation of these ions in the host glass matrix but also generates changes of the basic structural units of the glass matrix. Structural changes reveal that the samarium ions causes a change from the continuous vanadate-tellurate network to a continuous samarium-vanadate-tellurate network interconnected through Sm-O-V and Te-O-Sm bridges. Then, the surplus of nonbridging oxygens is be converted to bridging ones leading to the decrease of the connectivity of the network.

$$3. \quad \text{x} \gtrsim 40 \text{mol} \% \text{ M}\_{\text{a}} \text{O}\_{\text{b}} \text{(3)}$$

By increasing the Sm2O3 content up to 40mol%, the evolution of the structure can be explained considering the higher capacity of migration of the samarium ions inside the glass network and the formation of the SmVO4 crystalline phase, in agreement with XRD data. The accumulation of oxygen atoms in the glass network can be supported by the formation of ortho- and pyro-vanadate structural units.

On the other hand, the lead oxide generates the rapid deformation of the Te-O-Te linkages yielding the formation of [TeO3] structural units. Further, the excess of oxygen can be accommodated in the host matrix by conversion of some [VO4] structural units into [VO5] structural units.

The broader band centered at ~ 670-850cm-1 can be attributed to the Pb-O bonds vibrations from the [PbO4] and [PbO3] structural units. The absorption band centered at about 470cm−<sup>1</sup> may be correlated with the Pb-O stretching vibration in [PbO4] structural units [24-26]. The increase in the intensity of the bands situated between 650 and 850cm-1 show that the excess of oxygen in the glass network can be supported by the increase of [PbOn] structural units (with n=3 and 4).

In brief, the variations observed in the FTIR spectra suggest a gradual inclusion of the lead ions in the host vitreous matrix with increasing of the PbO content up to 50mol%, while progressive adding of samarium oxide determines the increase in the intensity of the bands due to the ortho- and pyrovanadate structural units. The mechanisms of incorporation of the lead and samarium ions in the host matrices can be summarized as following:

Structural and Optical Behavior of Vanadate-Tellurate Glasses Containing PbO or Sm2O3 69

**400 600 800 1000**

**400 600 800 1000**

400 600 800 1000

400 600 800 1000

**wavelength [nm]**

Fig. 3. UV-VIS absorption spectra of xPbO (or xSm2O3)·(100-x)[3TeO2·2V2O5] samples where

orthovanadate structural units for the samples with Sm2O3 content and the increase of the

The measurements of optical absorption and the absorption edge are important especially in connection with the theory of electronic structure of amorphous materials. The energy gap,

0≤x≤50mol% in function of lead (II) or samarium (III) oxide content.

number of [PbO3] structural units in the samples with PbO.

400 600 800 1000

**wavelength [nm]**

**absorbance [a.u.]**

**absorbance [a.u.]**

**absorbance [a.u.]**

**x=40%**

**x=30%**

**x=20%**

**Pb Sm**

**x=10%**

**x=50%**


#### **3.2 UV-VIS spectroscopy**

Optical absorption in solids occurs by various mechanisms, in all of which the phonon energy will be absorbed by either the lattice or by electrons where the transferred energy is covered. The lattice (or phonon) absorption will give information about atomic vibrations involved and this absorption of radiation normally occurs in the infrared region of the spectrum. Optical absorption is a useful method for investigating optically induced transitions and getting information about the energy gap of non-crystalline materials and the band structure. The principle of this technique is that a photon with energy greater than the band gap energy will be absorbed [27].

One of the most important concerns in rare earth doped glasses is to define the dopant environment. Hypersensitive transitions are observed in the spectra of all rare earth ions having more than one f electrons. Hypersensitive transitions of rare earth ions manifest an anomalous sensitivity of line strength to the character of the dopant environment [28, 29].

The measured UV-VIS absorption spectra of the lead and samarium-vanadate-tellurate glasses are shown in Fig. 3. The spectra show that the maxima of the absorption are located in the UV region for all investigated glasses containing PbO or Sm2O3.

The Pb+2 ions absorb strongly in the ultraviolet (310nm) and yield broad emission bands in the ultraviolet and blue spectral area [30]. The Sm+3 ions have five electrons in the f shell. The absorption bands due to the electron jump from the 6H5/2 ground state to the 6P5/2 (365nm), 6P7/2 (375nm), 6P3/2 (400nm), 4K11/2 (415nm), 4F15/2 (460nm) and 4F13/2 (475nm) excited states were observed [31].

The stronger transitions in the UV region can be due to the presence of the Te=O bonds from the [TeO3] structural units, the Pb=O bonds from [PbO3] structural units and the V=O bonds from [VO4] structural units which allow n-\* transitions. The intensity of these bands slightly increases and shifts towards higher wavelengths with increasing the concentration of PbO and Sm2O3. This may be due to the increase of the number of the V=O bonds from

i. For the samples with lead oxide, the Pb+2 ions can occupy a position in the chain itself and their influence on the V=O bonds is limited. Since the V=O mode position from about 1020cm-1 is preserved it can be concluded that the V=O bond is not directly influenced and the coordination number and symmetry of the [VO4] and [VO5] structural units do not change significantly. The increase in the intensity of the bands situated at 650 and 850cm-1 show that the PbO acts as a network former with a

ii. By increasing the samarium oxide content up to 30mol%, Sm+3 ions located between the vanadate chains may affect the isolated V=O bonds yielding to the depolimerization of the vanadate network in shorter and isolated chains formed of ortho- and pyrovanadate structural units. As a result they are markedly elongated and the vibrations frequency shifts toward lower wavenumbers. The increase of the content of samarium ions produces a strong depolymerization of the network leading to formation of SmVO4 crystalline phase, in agreement with the XRD data. The combined XRD and IR spectroscopy data show that the Sm2O3 acts as a network modifier with a strong effect

Optical absorption in solids occurs by various mechanisms, in all of which the phonon energy will be absorbed by either the lattice or by electrons where the transferred energy is covered. The lattice (or phonon) absorption will give information about atomic vibrations involved and this absorption of radiation normally occurs in the infrared region of the spectrum. Optical absorption is a useful method for investigating optically induced transitions and getting information about the energy gap of non-crystalline materials and the band structure. The principle of this technique is that a photon with energy greater than

One of the most important concerns in rare earth doped glasses is to define the dopant environment. Hypersensitive transitions are observed in the spectra of all rare earth ions having more than one f electrons. Hypersensitive transitions of rare earth ions manifest an anomalous sensitivity of line strength to the character of the dopant environment [28, 29].

The measured UV-VIS absorption spectra of the lead and samarium-vanadate-tellurate glasses are shown in Fig. 3. The spectra show that the maxima of the absorption are located

The Pb+2 ions absorb strongly in the ultraviolet (310nm) and yield broad emission bands in the ultraviolet and blue spectral area [30]. The Sm+3 ions have five electrons in the f shell. The absorption bands due to the electron jump from the 6H5/2 ground state to the 6P5/2 (365nm), 6P7/2 (375nm), 6P3/2 (400nm), 4K11/2 (415nm), 4F15/2 (460nm) and 4F13/2 (475nm)

The stronger transitions in the UV region can be due to the presence of the Te=O bonds from the [TeO3] structural units, the Pb=O bonds from [PbO3] structural units and the V=O bonds from [VO4] structural units which allow n-\* transitions. The intensity of these bands slightly increases and shifts towards higher wavelengths with increasing the concentration of PbO and Sm2O3. This may be due to the increase of the number of the V=O bonds from

in the UV region for all investigated glasses containing PbO or Sm2O3.

moderate effect on the vanadate-tellurate network.

about vanadate network.

the band gap energy will be absorbed [27].

excited states were observed [31].

**3.2 UV-VIS spectroscopy** 

Fig. 3. UV-VIS absorption spectra of xPbO (or xSm2O3)·(100-x)[3TeO2·2V2O5] samples where 0≤x≤50mol% in function of lead (II) or samarium (III) oxide content.

orthovanadate structural units for the samples with Sm2O3 content and the increase of the number of [PbO3] structural units in the samples with PbO.

The measurements of optical absorption and the absorption edge are important especially in connection with the theory of electronic structure of amorphous materials. The energy gap,

Structural and Optical Behavior of Vanadate-Tellurate Glasses Containing PbO or Sm2O3 71

**0**

**0**

**2**

**(h)**

Fig. 5. Plots of (αhν)2 versus hν for xSm2O3·(100-x)[3TeO2·2V2O5] samples where 0≤x≤50mol%.

The increase of the band gap may occur due to variation in non-bridging oxygen ion concentrations. In metal oxides, the valence band maximum mainly consists of 2p orbital of the oxygen atom and the conduction band minimum mainly consists of ns orbital of the metal atom. The non-bridging oxygen ions contribute to the valence band maximum. The non-bridging orbitals have higher energies than bonding orbitals. When a metal-oxygen bond is broken, the bond energy is released. The increase in concentration of the nonbridging oxygen ions results in the shift of the valence band maximum to higher energies and the reduction of the band gap. Thus, the enlarging of band gap energy due to increase in the PbO or Sm2O3 content suggests that non-bridging oxygen ion concentration decreases with increasing the PbO or Sm2O3 content that expands the band gap energy. In the glasses doped with Sm2O3, the non-bridging oxygen ions concentration decreases due to the formation of orthovanadate structural units. In the glasses doped with PbO, the nonbridging oxygen ions concentration decreases also because the lead atoms act as network formers and the accommodation with the excess of oxygen ions is possible by the increase of

The existence and variation of optical energy gap may be also explained by invoking the occurrence of local cross linking within the amorphous phase of the matrix network, in such

the polymerization degree of the network by Pb-O-Te and Pb-O-V linkages.

a way as to increase the degree of ordering in these parts.

**2 [cm-1·eV]2**

**4**

**6 Eg =2.0eV**

**(h)**

**2 [cm-1·eV]2**

**2**

**(h)**

**2 [cm-1·eV]2**

**4**

**Eg =2.09eV**

**x=50% PbO**

**x=40% PbO**

**Eg =2.08eV**

**x=30% PbO**

**6**

**1.0 1.5 2.0 2.5 3.0 3.5**

**1.0 1.5 2.0 2.5 3.0 3.5**

**1.0 1.5 2.0 2.5 3.0 3.5**

**1.0 1.5 2.0 2.5 3.0 3.5**

**1.0 1.5 2.0 2.5 3.0 3.5**

**1.0 1.5 2.0 2.5 3.0 3.5**

**h [eV]**

**0**

**0**

**0**

**20**

**(h)**

**2 [cm-1·eV]2**

**40**

**60**

**20**

**(h)**

**2 [cm-1·eV]2**

**40**

**Eg =1.85eV**

**x=10% PbO**

**x=0% PbO**

**Eg =1.84eV**

**20**

**(h)**

**2 [cm-1·eV]2**

**40**

**Eg =1.86eV**

**x=20% PbO**

Eg, is an important feature of semiconductors which determines their applications in optoelectronics [32]. Observations of the variation of Eg with increase in the modifier content can be attributed to the changes in the bonding that takes place in the glass

The nature of the optical transition involved in the network can be determined on the basis of the dependence of absorption coefficient (α) on phonon energy (hυ). The total absorption could be due to the optical transition which is fitted to the relation:

$$\mathbf{a} \text{ hu} = \mathbf{a}\_0 \text{ (ho-Eg)}^n$$

where Eg is the optical energy gap between the bottom of the conduction band and the top of the valence band at the same value of wavenumber, α0 is a constant related to the extent of the band tailing and the exponent n is an index which can have any values between ½ and 2 depending on the nature of the interband electronic transitions.

Extrapolating the linear portion of the graph (αhν)2 → 0 to hν axis, the optical band gaps, Eg are determined with increasing PbO and Sm2O3 content (Figs. 4 and 5). The optical band gap increases gradually from 1.84eV to 2.09eV and 2.21eV, respectively, by adding of PbO and Sm2O3. In either case the values are systematically increasing with the increase of x. It is to be noted that the curves are characterized by the presence of an exponential decay tail at low energy. These results indicate the presence of a well defined *π→π*\* transition associated with the formation of conjugated electronic structure [33].

Fig. 4. Plots of (αhν)2 versus hν for xPbO·(100-x)[3TeO2·2V2O5] glasses where 0≤x≤50mol%

Eg, is an important feature of semiconductors which determines their applications in optoelectronics [32]. Observations of the variation of Eg with increase in the modifier content

The nature of the optical transition involved in the network can be determined on the basis of the dependence of absorption coefficient (α) on phonon energy (hυ). The total absorption

α hυ = α0 (hυ-Eg)n where Eg is the optical energy gap between the bottom of the conduction band and the top of the valence band at the same value of wavenumber, α0 is a constant related to the extent of the band tailing and the exponent n is an index which can have any values between ½

Extrapolating the linear portion of the graph (αhν)2 → 0 to hν axis, the optical band gaps, Eg are determined with increasing PbO and Sm2O3 content (Figs. 4 and 5). The optical band gap increases gradually from 1.84eV to 2.09eV and 2.21eV, respectively, by adding of PbO and Sm2O3. In either case the values are systematically increasing with the increase of x. It is to be noted that the curves are characterized by the presence of an exponential decay tail at low energy. These results indicate the presence of a well defined *π→π*\* transition associated

**x=50% Sm2**

**x=30% Sm2**

**=1.90eV**

**O3**

**O3**

**0**

**0**

**(h)**

**2 [cm-1·eV]2**

**(h)**

Fig. 4. Plots of (αhν)2 versus hν for xPbO·(100-x)[3TeO2·2V2O5] glasses where 0≤x≤50mol%

**0**

**5**

**<sup>10</sup> Eg**

**2 [cm-1·eV]2**

**5 Eg =2.12eV x=40% Sm2**

**2**

**(h)**

**2 [cm-1·eV]2**

**4**

**1.0 1.5 2.0 2.5 3.0 3.5**

**1.0 1.5 2.0 2.5 3.0 3.5**

**hv [eV]**

**1.0 1.5 2.0 2.5 3.0 3.5**

can be attributed to the changes in the bonding that takes place in the glass

could be due to the optical transition which is fitted to the relation:

and 2 depending on the nature of the interband electronic transitions.

with the formation of conjugated electronic structure [33].

**<sup>20</sup> Eg <sup>E</sup> =2.21eV <sup>g</sup>**

**<sup>O</sup> x=20% Sm2**

**O3 <sup>3</sup>**

**1.0 1.5 2.0 2.5 3.0 3.5**

**1.0 1.5 2.0 2.5 3.0 3.5**

**1.0 1.5 2.0 2.5 3.0 3.5**

**hv [eV]**

**0**

**0**

**10**

**(h)**

**(h)**

**2 [cm-1·eV]2**

**2 [cm-1·eV]2**

**(h)**

**2 [cm-1·eV]2**

**20**

**Eg =1.85eV**

**Eg =1.84eV**

**x=0% Sm2**

**O3**

**30**

**10**

**=1.87eV**

**x=10% Sm2**

**O3**

Fig. 5. Plots of (αhν)2 versus hν for xSm2O3·(100-x)[3TeO2·2V2O5] samples where 0≤x≤50mol%.

The increase of the band gap may occur due to variation in non-bridging oxygen ion concentrations. In metal oxides, the valence band maximum mainly consists of 2p orbital of the oxygen atom and the conduction band minimum mainly consists of ns orbital of the metal atom. The non-bridging oxygen ions contribute to the valence band maximum. The non-bridging orbitals have higher energies than bonding orbitals. When a metal-oxygen bond is broken, the bond energy is released. The increase in concentration of the nonbridging oxygen ions results in the shift of the valence band maximum to higher energies and the reduction of the band gap. Thus, the enlarging of band gap energy due to increase in the PbO or Sm2O3 content suggests that non-bridging oxygen ion concentration decreases with increasing the PbO or Sm2O3 content that expands the band gap energy. In the glasses doped with Sm2O3, the non-bridging oxygen ions concentration decreases due to the formation of orthovanadate structural units. In the glasses doped with PbO, the nonbridging oxygen ions concentration decreases also because the lead atoms act as network formers and the accommodation with the excess of oxygen ions is possible by the increase of the polymerization degree of the network by Pb-O-Te and Pb-O-V linkages.

The existence and variation of optical energy gap may be also explained by invoking the occurrence of local cross linking within the amorphous phase of the matrix network, in such a way as to increase the degree of ordering in these parts.

Structural and Optical Behavior of Vanadate-Tellurate Glasses Containing PbO or Sm2O3 73

glass network. This effect seems more pronounced in doping of the network with the Sm+3 ions. These results are in agreement with XRD data which indicate the higher affinity of the samarium ions to attract structural units with negative charge yielding the formation

In this section, the purpose of the present paper was to continue the investigation of the structure of the vanadate-tellurate network and especially to illuminate structural aspects of the vanadate network using quantum-chemical calculations because the coordination state of vanadium atoms is not well understood. Figure 7 shows the optimized structure

of the SmVO4 crystalline phase for samples with x>30mol%.

Fig. 7. Optimized structure of the model for binary 3TeO2·2V2O5 glassy.

tetrahedrons are easy distorted around the vanadium center.

Analyzing the structural changes resulted from the geometry optimization of our model, we found that the vanadium ions are distributed into two crystallographic sites: the [VO4] tetrahedral and [VO5] square pyramidal units. The vanadium tetrahedrons are very regular with vanadium-oxygen distances ranging from 1.57 to 1.82Å and O-V-O angles ranging from 1040 to 1100 (with a mean value (109.50) very close to the ideal value (109.280) corresponding to the tetrahedral geometry). In our model, the V-O interatomic bond distances are ranging from 1.60 to 1.65Å, 1.75 to 1.80Å (the average V-O distance is 1.72Å) and O-V-O angles values are ranging from 101 to 1120. This result show that the [VO4]

proposed to the 3TeO22V2O5 glass network.

**3.3 DFT calculations** 

Refractive index is one of the most important properties in optical glasses. A large number of researchers have carried out investigations to ascertain the relation between refractive index and glass composition. It is generally recognized that the refractive index, n, of many common glasses can be varied by changing the base glass composition [34].

The observed decrease in the refractive index of the studied glasses accompanying to the addition of PbO or Sm2O3 content presented in the Fig. 6 can be considered as an indication of a decrease in number of non-bridging oxygen ions.

Fig. 6. The relationship between the optic gap, Eg, and refractive index, n, and the and the PbO or Sm2O3 contents (the line is only a guide for the eye).

In brief, we can conclude that the optical band gap increases with increasing the PbO or Sm2O3 contentof the glass. Since the basic structural units of the vanadate-tellurate glasses are known to be the [TeO3] and [VO4] structural units and the internal vibrations of these molecular units take part in the transitions. In this work, the increase of the optical band gap, Eg, to larger energies with increasing the PbO or Sm2O3 content is probably related to the progressive decrease in the concentration of non-bridging oxygen. This decrease in turn gives rise to a possible decrease in the bridging Te-O-Te and V-O-V linkages. The shift is attributed to structural changes which are the the result of the different (interstitial or substitutional) site occupations of the Pb+2 or Sm+3 ions which are added to the vanadatetellurate matrix and modify the network.

We assume that as the cation concentration increases, the Te-O-Te and V-O-V linkages develop bonds with Pb+2 or Sm+3 ions, which in turn leads to the gradual breakdown of the glass network. This effect seems more pronounced in doping of the network with the Sm+3 ions. These results are in agreement with XRD data which indicate the higher affinity of the samarium ions to attract structural units with negative charge yielding the formation of the SmVO4 crystalline phase for samples with x>30mol%.

## **3.3 DFT calculations**

72 Infrared Spectroscopy – Materials Science, Engineering and Technology

Refractive index is one of the most important properties in optical glasses. A large number of researchers have carried out investigations to ascertain the relation between refractive index and glass composition. It is generally recognized that the refractive index, n, of many

The observed decrease in the refractive index of the studied glasses accompanying to the addition of PbO or Sm2O3 content presented in the Fig. 6 can be considered as an indication

**0 10 20 30 40 50**

**x [moli%]**

In brief, we can conclude that the optical band gap increases with increasing the PbO or Sm2O3 contentof the glass. Since the basic structural units of the vanadate-tellurate glasses are known to be the [TeO3] and [VO4] structural units and the internal vibrations of these molecular units take part in the transitions. In this work, the increase of the optical band gap, Eg, to larger energies with increasing the PbO or Sm2O3 content is probably related to the progressive decrease in the concentration of non-bridging oxygen. This decrease in turn gives rise to a possible decrease in the bridging Te-O-Te and V-O-V linkages. The shift is attributed to structural changes which are the the result of the different (interstitial or substitutional) site occupations of the Pb+2 or Sm+3 ions which are added to the vanadate-

We assume that as the cation concentration increases, the Te-O-Te and V-O-V linkages develop bonds with Pb+2 or Sm+3 ions, which in turn leads to the gradual breakdown of the

Fig. 6. The relationship between the optic gap, Eg, and refractive index, n, and the and the

PbO or Sm2O3 contents (the line is only a guide for the eye).

2.0

2.4

**n**

2.8

**PbO Sm2 O3**

common glasses can be varied by changing the base glass composition [34].

of a decrease in number of non-bridging oxygen ions.

**2.0**

**E g**

tellurate matrix and modify the network.

**2.4**

**E**

**g [eV]**

**2.8**

**n**

In this section, the purpose of the present paper was to continue the investigation of the structure of the vanadate-tellurate network and especially to illuminate structural aspects of the vanadate network using quantum-chemical calculations because the coordination state of vanadium atoms is not well understood. Figure 7 shows the optimized structure proposed to the 3TeO22V2O5 glass network.

Fig. 7. Optimized structure of the model for binary 3TeO2·2V2O5 glassy.

Analyzing the structural changes resulted from the geometry optimization of our model, we found that the vanadium ions are distributed into two crystallographic sites: the [VO4] tetrahedral and [VO5] square pyramidal units. The vanadium tetrahedrons are very regular with vanadium-oxygen distances ranging from 1.57 to 1.82Å and O-V-O angles ranging from 1040 to 1100 (with a mean value (109.50) very close to the ideal value (109.280) corresponding to the tetrahedral geometry). In our model, the V-O interatomic bond distances are ranging from 1.60 to 1.65Å, 1.75 to 1.80Å (the average V-O distance is 1.72Å) and O-V-O angles values are ranging from 101 to 1120. This result show that the [VO4] tetrahedrons are easy distorted around the vanadium center.

Structural and Optical Behavior of Vanadate-Tellurate Glasses Containing PbO or Sm2O3 75

[6] S. Jayaseelan, P. Muralidharan, M. Venkateswarlu, N. Satyanarayana, Mater. Chem.

[9] W. A. Pisarski, T. Goryczka, B. Wodecka-Dus, M. Plonska, J. Pisarska, Mater. Sci. Eng.

[12] Spartan'04, Wavefunction Inc., 18401 Von Karman Avenue, Suite 370 Irvine, CA 92612. [13] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J.

A. Pople, Gaussian 03, Revision A.1, Gaussian, Inc., Pittsburgh PA, 2003.

[17] S. Rada, E. Culea, M. Culea, Borate-Tellurate Glasses: An Alternative of Immobilization of the Hazardous Wastes, Nova Science Publishers INC., New York, 2010.

[14] S. Rada, M. Culea, E. Culea, J. Phys. Chem. A 112(44) (2008) 11251. [15] S. Rada, M. Neumann, E. Culea, Solid State Ionics 181 (2010) 1164. [16] S. Rada, E. Culea, M. Rada, Mater. Chem. Phys. 128(3) (2011) 464.

[19] M. Rada, V. Maties, S. Rada, E. Culea, J. Non-Cryst. Solids 356 (2010) 1267.

[23] K. Gatterer, G. Pucker, H. P. Fritzer, Phys. Chem. Glasses 38 (1997) 293. [24] S. Rada, M. Culea, E. Culea, J. Non-Cryst. Solids 354(52-54) (2008) 5491. [25] S. Rada, M. Culea, M. Neumann, E. Culea, Chem. Phys. Letters 460 (2008) 196. [26] M. Rada, V. Maties, M. Culea, S. Rada, E. Culea, Spectrochim. Acta A 75 (2010) 507.

[18] S. Rada, E. Culea, J. Molec. Struct. 929 (2009) 141.

[30] D. Ehrt, J. Non-Cryst. Solids 348 (2004) 22.

[34] R. El-Mallawany, J. Appl. Phys. 72 (1992) 1774.

[20] S. Rada, R. Chelcea, E. Culea, J. Molec. Model. 17 (2011) 165. [21] S. Rada, E. Culea, M. Culea, J. Mater. Sci. 43(19) (2008) 6480. [22] K. V. Ramesh, D. L. Sastry, J. Non-Cryst. Solids 352 (2006) 5421.

[27] M. T. Abd El-Ati, A. A. Higazy, J. Mater. Sci. 35 (2000) 6175. [28] S. N. Misra, S. O. Sommerer, Appl. Spectrosc. Rev. 26 (1991) 151. [29] V. K. Tikhomirov, M. Naftaly, A. Jha, J. Appl. Phys. 86 (1999) 351.

[31] A. M. Nassar, N. A. Ghoneim, J. Non-Cryst. Solids 46 (1981) 181. [32] T. Aoki, Y. Hatanaka, D.C. Look, Appl. Phys. Lett. 76 (2000) 3257.

[33] W. R. Salaneck, C. R. Wu, J. L. Bredas, J. J. Ritsko, Chem. Phys. Lett. 127 (1986) 88.

A. Montgomery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, and J.

[7] G. A. Kumar, A. Martinez, A. Mejia, C. G. Eden, J. Alloys Compd. 365 (2004) 117. [8] J. Yang, S. Dai, Y. Zhou, L. Wen, L. Hu, Z. H. Jiang, J. Appl. Phys. 93 (2003) 977.

[10] G. Lakshminarayana, S. Buddhudu, Mater. Chem. Phys. 102 (2007) 181. [11] Thulasiramudu, S. Buddhudu, Spectrochim Acta A 67 (2007) 802.

Phys. 87 (2004) 370.

122 (2005) 94.

The [VO5] square pyramidal units are considerably distorted around the vanadium center and the V-O bond distances are ranging from 1.65 to 2.30Å. Such a behavior was reported for the two-dimensional layered vanadate compounds [35, 36]. This shows that there is instability in the nonequivalent V-O bonds in the polyhedron. In essence, this is due to the displacement of the vanadium atom from the centre of the polyhedron, whose asymmetry strongly depends on the manner of connection with the surrounding polyhedron. This deformation will be expressed more clearly in the formation of the vitreous matrix.

This structural model shows a very complex behavior of the vanadium atoms and their stabilization can be achieved by the formation of orthovanadate structural units or by the intercalation of [PbO3] structural units in the immediate vicinity of these units.

#### **4. Conclusions**

The X-ray diffraction patterns reveal the SmVO4 crystalline phase in the samples with x>30mol% Sm2O3 indicating that the samarium ions have an pronounced affinity towards the vanadate structural units. By adding of Sm2O3 content in the host matrix, the FTIR spectra suggests that the glass network modification has taken place mainly in the vanadate part whereas by adding of PbO, the network is transformed from a vanadate-tellurate network into a continuous lead-vanadate-tellurate network by Te-O-Pb and V-O-Pb linkages.

The UV-VIS absorption spectra of the studied samples reveal the additional absorptions in the 250-1050nm range due to the generation of *n→π*\* transitions and the presence of the transition or rare earth metallic ions. By increasing the metal oxide content up to 50mol%, the optical band gap energy increases. This suggests a decrease of the non-bridging oxygens due to the formation of orthovanadate (for adding Sm2O3) and [PbO3] (for adding PbO) structural units, respectively. The band gap energy was changed due to structural modifications of the network.

Our DFT investigations show that the penta-coordinated vanadium atoms show a unique influence on the structural properties of the glasses.

### **5. Acknowledgments**

The financial support of the Ministry of Education and Research of Romania-National University Research Council (CNCSIS, PN II-IDEI 183/2008, contract number 476/2009) is gratefully acknowledged by the authors.

#### **6. References**


The [VO5] square pyramidal units are considerably distorted around the vanadium center and the V-O bond distances are ranging from 1.65 to 2.30Å. Such a behavior was reported for the two-dimensional layered vanadate compounds [35, 36]. This shows that there is instability in the nonequivalent V-O bonds in the polyhedron. In essence, this is due to the displacement of the vanadium atom from the centre of the polyhedron, whose asymmetry strongly depends on the manner of connection with the surrounding polyhedron. This

This structural model shows a very complex behavior of the vanadium atoms and their stabilization can be achieved by the formation of orthovanadate structural units or by the

The X-ray diffraction patterns reveal the SmVO4 crystalline phase in the samples with x>30mol% Sm2O3 indicating that the samarium ions have an pronounced affinity towards the vanadate structural units. By adding of Sm2O3 content in the host matrix, the FTIR spectra suggests that the glass network modification has taken place mainly in the vanadate part whereas by adding of PbO, the network is transformed from a vanadate-tellurate network into a continuous lead-vanadate-tellurate network by Te-O-Pb and V-O-Pb

The UV-VIS absorption spectra of the studied samples reveal the additional absorptions in the 250-1050nm range due to the generation of *n→π*\* transitions and the presence of the transition or rare earth metallic ions. By increasing the metal oxide content up to 50mol%, the optical band gap energy increases. This suggests a decrease of the non-bridging oxygens due to the formation of orthovanadate (for adding Sm2O3) and [PbO3] (for adding PbO) structural units, respectively. The band gap energy was changed due to structural

Our DFT investigations show that the penta-coordinated vanadium atoms show a unique

The financial support of the Ministry of Education and Research of Romania-National University Research Council (CNCSIS, PN II-IDEI 183/2008, contract number 476/2009) is

[2] H. Nasu, O. Matsusita, K. Kamiya, H. Kobayashi, K. Kubodera, J. Non-Cryst. Solids

[4] G. Nunziconti, S. Bemeschi, M. Bettinelli, M. Brei, B. Chen, S. Pelli, A. Speghini, G. C.

deformation will be expressed more clearly in the formation of the vitreous matrix.

intercalation of [PbO3] structural units in the immediate vicinity of these units.

**4. Conclusions** 

linkages.

modifications of the network.

**5. Acknowledgments** 

**6. References** 

influence on the structural properties of the glasses.

[1] S. Tanaba, K. Hirao, N. Soga, J. Non-Cryst. Solids 122 (1990) 79.

Righini, J. Non-Cryst. Solids 345&346 (2004) 343.

[5] M. Ganguli, M. Bhat Harish, K. J. Rao, Phys. Chem. Glasses 40 (1999) 297.

gratefully acknowledged by the authors.

[3] B. Eraiah, Bull. Mater. Sci. 29(4) (2006) 375.

124 (1990) 275.


**Water in Rocks and Minerals – Species,** 

Jun-ichi Fukuda *Tohoku University* 

*Japan* 

**Distributions, and Temperature Dependences** 

Water is ubiquitously distributed in the interior of the earth, as various forms in rocks and minerals: In rocks, aggregates of minerals, fluid water in which molecular H2O is clustered, is trapped at intergranular regions and as fluid inclusions (e.g., Hiraga et al., 2001). In minerals, water is located as the form of –OH in their crystal structures as impurities. Surprisingly, such water species are distributed over five times in the earth's interior than ocean water (e.g., Jacobsen & Van der Lee, 2006), and play very important roles on earth dynamics such as deformation and reactions of minerals and rocks (e.g., Thompson & Rubie, 1985; Dysthe & Wogelius, 2006). In the filed of earth sciences therefore, people are trying to measure properties of water in rocks and minerals such as species, contents,

Infrared (IR) spectroscopy is a powerful tool to quantitatively measure these properties of water in rocks and minerals (See Aines & Rossman, 1984; Keppler & Smyth, 2006 for IR spectra of various rocks and minerals). In this chapter, for an advanced IR spectroscopic measurement, I introduce in-situ high temperature IR spectroscopy to investigate above matters. First, I use chalcedonic quartz, which contains fluid water at intergranular regions and –OH in quartz crystal structures. Next, I use beryl, a typical cyclosilicate which contains isolated (not clustered) H2O molecules in open cavities of the crystal structure. Changes of the states of water in chalcedonic quartz and beryl by temperature changes and dehydration will be discussed. Finally, I perform two-dimensional IR mappings for naturally deformed rocks to investigate water distribution in polymineralic mixtures, and discuss possible water

Transmitted IR spectra for rocks and minerals are generally measured by making thin sections of samples with thicknesses of from 20 to 200 μm, which depend on concentrations and absorption coefficients based on Beer-Lambert law. A Fourier Transform IR microspectrometer totally used in this study is equipped with a silicon carbide (globar) IR source and a Ge-coated KBr beamsplitter. IR light through a sample is measured using a

In-situ high temperature IR spectra were measured for a sample on a heating stage which was inserted into the IR path. The sample was heated at 100 °C/minute to desired

**1. Introduction** 

distribution, thermal behaviour, migration rate, etc.

transportation during rock deformation.

mercury-cadmium-telluride detector.

**2. Methods** 


## **Water in Rocks and Minerals – Species, Distributions, and Temperature Dependences**

Jun-ichi Fukuda *Tohoku University Japan* 

## **1. Introduction**

76 Infrared Spectroscopy – Materials Science, Engineering and Technology

[35] Sun, E. Wang, D. Xiao, H. An, L. Xu, J. Molec. Struct. 840 (2007) 53.

Water is ubiquitously distributed in the interior of the earth, as various forms in rocks and minerals: In rocks, aggregates of minerals, fluid water in which molecular H2O is clustered, is trapped at intergranular regions and as fluid inclusions (e.g., Hiraga et al., 2001). In minerals, water is located as the form of –OH in their crystal structures as impurities. Surprisingly, such water species are distributed over five times in the earth's interior than ocean water (e.g., Jacobsen & Van der Lee, 2006), and play very important roles on earth dynamics such as deformation and reactions of minerals and rocks (e.g., Thompson & Rubie, 1985; Dysthe & Wogelius, 2006). In the filed of earth sciences therefore, people are trying to measure properties of water in rocks and minerals such as species, contents, distribution, thermal behaviour, migration rate, etc.

Infrared (IR) spectroscopy is a powerful tool to quantitatively measure these properties of water in rocks and minerals (See Aines & Rossman, 1984; Keppler & Smyth, 2006 for IR spectra of various rocks and minerals). In this chapter, for an advanced IR spectroscopic measurement, I introduce in-situ high temperature IR spectroscopy to investigate above matters. First, I use chalcedonic quartz, which contains fluid water at intergranular regions and –OH in quartz crystal structures. Next, I use beryl, a typical cyclosilicate which contains isolated (not clustered) H2O molecules in open cavities of the crystal structure. Changes of the states of water in chalcedonic quartz and beryl by temperature changes and dehydration will be discussed. Finally, I perform two-dimensional IR mappings for naturally deformed rocks to investigate water distribution in polymineralic mixtures, and discuss possible water transportation during rock deformation.

## **2. Methods**

Transmitted IR spectra for rocks and minerals are generally measured by making thin sections of samples with thicknesses of from 20 to 200 μm, which depend on concentrations and absorption coefficients based on Beer-Lambert law. A Fourier Transform IR microspectrometer totally used in this study is equipped with a silicon carbide (globar) IR source and a Ge-coated KBr beamsplitter. IR light through a sample is measured using a mercury-cadmium-telluride detector.

In-situ high temperature IR spectra were measured for a sample on a heating stage which was inserted into the IR path. The sample was heated at 100 °C/minute to desired

Water in Rocks and Minerals – Species, Distributions, and Temperature Dependences 79

High temperature IR spectra were measured for the sample set on the heating stage (Fig. 2). With increasing temperature up to 400 °C, the broad band due to fluid water dominantly and asymmetrically shifts to high wavenumbers. Contrary to this, the band due to Si-OH slightly shifts to high wavenumbers. After quenching to RT from high temperatures, these water bands do not change from those before heating, indicating that vibrational states of fluid water and Si-OH are changed at high temperatures without dehydration. The

Fig. 2. In-situ high temperature IR spectra of a chalcedonic quartz (black lines) in the water stretching region (Replotted from Fukuda et al., 2009c). The spectrum at RT after heating at 400 °C is shown as gray line on the top, showing no significant change from the spectrum

Since vibrational energy of OH stretching of Si-OH is structurally limited within quartz crystal structures, the band is sharp even at high temperatures. The deviation of the wavenumber from free -OH stretching (around 3650 cm–1; summarized in Libowitzky, 1999) can be explained by the work of hydrogen bond in Si-OH…O-Si in quartz crystal structures, which weaken OH vibrational energy. With increasing temperature, the hydrogen bond distance is extended due to thermal expansion of quartz crystal structure (e.g., Kihara, 2001). Resultantly, the band due to Si-OH slightly shifts to high wavenumber and the band height

In fluid water, H2O is clustered and networked by various hydrogen bond strengths (e.g., Brubach et al., 2005). Therefore, fluid water shows the broad band. With increasing temperatures, the average coordination numbers of a H2O molecule to adjacent H2O molecules at confined intergranular regions of chalcedonic quartz are reduced due to increases of vibrational energies without dehydration. The average coordination number of a single H2O molecule in fluid water is 2–3 molecules at RT (Brubach et al., 2005), and 1–2 above supercritical temperature (Nakahara et al., 2001). This leads to significant shifts of wavenumbers to higher. Also, band heights of fluid water are decreased with increasing

is not so decreased (3599 cm–1 at 400 °C; Fukuda & Nakashima, 2008).

following is discussion for changes in the vibrational states of water.

**3.2 Water vibrations at high temperatures** 

before heating.

temperature, and spectra were collected at about 1 minute. Mapping measurements were carried out using an auto XY-stage under atomospheric condition.

#### **3. Water in rocks: As an example of chalcedonic quartz**

Fluid water and –OH are trapped in chalcedonic quartz, an aggregate of microcrystalline quartz (SiO2) grains (sometimes called as chalcedony or agate which has different transparency and is treated as a gem). Fluid water in chalcedonic quartz is dominantly trapped at intergranular regions such as grain boundaries and triple junctions of grains. IR spectra of chalcedonic quartz have been measured at room temperature (RT) (Frondel, 1982; Graetsch et al., 1985). In this section, I measure in-situ high temperature IR spectra for chalcedonic quartz as a representative material that abundantly contains fluid water and structurally-trapped –OH.

#### **3.1 Typical IR spectrum at RT and water species**

Figure 1 is a typical IR spectrum at RT for a thin section (ca. 100 μm thickness) of chalcedonic quartz. The band due to fluid water shows an asymmetric broad band ranging from 2750 to 3800 cm–1 with a shoulder around 3260 cm–1: this band feature is the same with that of simple fluid water which exists everywhere around us (e.g., Eisenberg & Kauzman, 1969). –OH in chalcedonic quartz is mainly trapped as Si-OH by breaking the network of SiO2 bonds (Kronenberg & Wolf, 1990), and the OH stretching band is sharp (3585 cm–1 at RT). The bending mode of fluid water, which should be seen at around 1600 cm–1, is hindered by many sharp Si-O stretching bands. Combination modes of the stretching and bending modes of fluid water and Si-OH are clearly seen for a thicker sample (1 mm in Fig. 1), and they are detected at 5200 cm–1 and 4500 cm–1 respectively. Contents of fluid water and Si-OH can be calculated from these bands' heights using their molar absorption coefficients of 0.761 and 1.141 L mol–1 cm–1, respectively (Scholze, 1960; Graetsch et al., 1985); in this case 0.32 wt % H2O and 0.28 wt % Si-OH, respectively (Fukuda et al., 2009a).

Fig. 1. Typical IR spectrum of chalcedonic quartz at RT. The sample thickness of 100 μm for 4000–1000 cm−1 and 1 mm for 5500–4000 cm−1 (combination modes of the stretching and bending modes of fluid water and Si-OH).

#### **3.2 Water vibrations at high temperatures**

78 Infrared Spectroscopy – Materials Science, Engineering and Technology

temperature, and spectra were collected at about 1 minute. Mapping measurements were

Fluid water and –OH are trapped in chalcedonic quartz, an aggregate of microcrystalline quartz (SiO2) grains (sometimes called as chalcedony or agate which has different transparency and is treated as a gem). Fluid water in chalcedonic quartz is dominantly trapped at intergranular regions such as grain boundaries and triple junctions of grains. IR spectra of chalcedonic quartz have been measured at room temperature (RT) (Frondel, 1982; Graetsch et al., 1985). In this section, I measure in-situ high temperature IR spectra for chalcedonic quartz as a representative material that abundantly contains fluid water and

Figure 1 is a typical IR spectrum at RT for a thin section (ca. 100 μm thickness) of chalcedonic quartz. The band due to fluid water shows an asymmetric broad band ranging from 2750 to 3800 cm–1 with a shoulder around 3260 cm–1: this band feature is the same with that of simple fluid water which exists everywhere around us (e.g., Eisenberg & Kauzman, 1969). –OH in chalcedonic quartz is mainly trapped as Si-OH by breaking the network of SiO2 bonds (Kronenberg & Wolf, 1990), and the OH stretching band is sharp (3585 cm–1 at RT). The bending mode of fluid water, which should be seen at around 1600 cm–1, is hindered by many sharp Si-O stretching bands. Combination modes of the stretching and bending modes of fluid water and Si-OH are clearly seen for a thicker sample (1 mm in Fig. 1), and they are detected at 5200 cm–1 and 4500 cm–1 respectively. Contents of fluid water and Si-OH can be calculated from these bands' heights using their molar absorption coefficients of 0.761 and 1.141 L mol–1 cm–1, respectively (Scholze, 1960; Graetsch et al., 1985); in this case 0.32 wt % H2O and 0.28 wt % Si-OH, respectively (Fukuda et al., 2009a).

Fig. 1. Typical IR spectrum of chalcedonic quartz at RT. The sample thickness of 100 μm for 4000–1000 cm−1 and 1 mm for 5500–4000 cm−1 (combination modes of the stretching and

carried out using an auto XY-stage under atomospheric condition.

**3. Water in rocks: As an example of chalcedonic quartz** 

**3.1 Typical IR spectrum at RT and water species** 

bending modes of fluid water and Si-OH).

structurally-trapped –OH.

High temperature IR spectra were measured for the sample set on the heating stage (Fig. 2). With increasing temperature up to 400 °C, the broad band due to fluid water dominantly and asymmetrically shifts to high wavenumbers. Contrary to this, the band due to Si-OH slightly shifts to high wavenumbers. After quenching to RT from high temperatures, these water bands do not change from those before heating, indicating that vibrational states of fluid water and Si-OH are changed at high temperatures without dehydration. The following is discussion for changes in the vibrational states of water.

Fig. 2. In-situ high temperature IR spectra of a chalcedonic quartz (black lines) in the water stretching region (Replotted from Fukuda et al., 2009c). The spectrum at RT after heating at 400 °C is shown as gray line on the top, showing no significant change from the spectrum before heating.

Since vibrational energy of OH stretching of Si-OH is structurally limited within quartz crystal structures, the band is sharp even at high temperatures. The deviation of the wavenumber from free -OH stretching (around 3650 cm–1; summarized in Libowitzky, 1999) can be explained by the work of hydrogen bond in Si-OH…O-Si in quartz crystal structures, which weaken OH vibrational energy. With increasing temperature, the hydrogen bond distance is extended due to thermal expansion of quartz crystal structure (e.g., Kihara, 2001). Resultantly, the band due to Si-OH slightly shifts to high wavenumber and the band height is not so decreased (3599 cm–1 at 400 °C; Fukuda & Nakashima, 2008).

In fluid water, H2O is clustered and networked by various hydrogen bond strengths (e.g., Brubach et al., 2005). Therefore, fluid water shows the broad band. With increasing temperatures, the average coordination numbers of a H2O molecule to adjacent H2O molecules at confined intergranular regions of chalcedonic quartz are reduced due to increases of vibrational energies without dehydration. The average coordination number of a single H2O molecule in fluid water is 2–3 molecules at RT (Brubach et al., 2005), and 1–2 above supercritical temperature (Nakahara et al., 2001). This leads to significant shifts of wavenumbers to higher. Also, band heights of fluid water are decreased with increasing

Water in Rocks and Minerals – Species, Distributions, and Temperature Dependences 81

the split of the band at 3585 cm–1 at RT) (Yamagishi et al., 1997). The appearance of the band at 3730 cm–1 at 0 minute heating at 500 °C is due to slight dehydration during heating from RT to 500 °C at 100 °C/minute. The wavenumbers of these new –OH bands at high temperature are slightly different from those at RT, presumably due to thermal expansions

In addition to fluid water in rocks and –OH in mineral crystal structures as described above, isolated (not clustered) H2O molecules are incorporated in open cavities of crystal structures, and they are sometimes coupled with cations. H2O in open cavities has well been studied for beryl, a typical cyclosilicate (after Wood & Nassau, 1967). Ideal chemical formula of beryl is Be3Al2Si6O18, and six-membered SiO4 rings are stacked along the crystallographic *c*-axis and make a pipe-like cavity called a channel (Fig. 4) (Gibbs et al., 1968). Isolated H2O is trapped in the channel, and forms two kinds of orientations, depending on whether it coordinates to a cation (called type II) or not (called type I) (after Wood & Nassau, 1967). Such cations are trapped in the channels to compensate the electrical charge balances caused by Be2+-Li+ and Si4+-Al3+ substitutions and lacks of Be2+ in the crystal structure of beryl. The cations in the channels are assumed to be mainly Na+, and some other alkali cations may be incorporated (Hawthorne and Černý, 1977; Aurisicchio et al., 1988; Artioli et al., 1993; Andersson, 2006). In this section, I introduce polarized IR spectra of beryl, and discuss changes of the states of type

Fig. 4. Crystal structure of beryl. The (001)-plane (i.e., viewed down from the *c*-axis) and the channel section. Positions of type I/II H2O, a cation, and CO2 are also shown. Modified after

The chemical composition of the natural beryl sample used in this study was analyzed by Xray wavelength dispersive spectroscopy for major atomic contents, inductivity coupled plasma-atomic emission spectroscopy for Be content, and atomic absorption spectroscopy for Li and Rb contents (Table 1). The type I/II H2O contents were determined from intensities of IR bands due to the asymmetric stretching of type I and the symmetric stretching of type II in a polarized IR spectrum at RT (See the spectrum in the next section), using their molar absorption coefficients of 206 L mol–1 cm–1 and 256 L mol–1 cm–1,

of crystals and changes of hydrogen bond distances at high temperature.

**4. H2O molecules in minerals: As an example of beryl** 

I/II H2O in the channels by temperature changes and dehydration.

Fukuda & Shinoda (2011).

**4.1 Chemical composition of the sample** 

temperatures. For example, the maximum band height at 400 °C is approximately 50 % of that at RT. This is also because of decreases of average numbers of H2O in areas that IR light captures (i.e., density; Schwarzer et al., 2005).

#### **3.3 Dehydration behaviour**

When the sample is kept at high temperatures, dehydration occurs. High temperature IR spectra were continuously measured to monitor dehydration. Figure 3 shows dehydration behaviour measured at 500 °C (Fig. 3a) and 400 °C (Fig. 3b). Both of the experiments were performed by heating during 500 minutes in total. Broad bands around 3800-3000 cm–1 in both spectra decrease with keeping at high temperatures, and the wavenumbers of the broad bands are not changed during heating. IR spectra at RT after heating (gray spectra in Fig. 3) also show decreases of fluid water. This indicates that fluid water was dehydrated through intergranular regions which are fast paths for mass transfers (See Ingrin et al., 1995; Okumura and Nakashima, 2004; Fukuda et al., 2009c for estimation of water diffusivity). Over 50 % of the band areas are decreased during the heating in 84 minutes at 500 °C. Band areas of 80 % are decreased in 250 minutes, and the features of the spectra are not changed after that. The RT temperature spectrum after 500 minutes heating at 500 °C also shows significant reduction of board band due to fluid water. This remained band due to fluid water may reflect fluid inclusions, which is tightly trapped at open spaces in crystal structures. On the other hand, 60 % of the band areas are preserved after heating in 500 minutes at 400 °C, and the RT spectrum after 500 minutes heating at 400 °C still shows a strong signal of fluid water.

Fig. 3. Dehydration behaviour at 500 °C (left) and 400 °C (right) in the water stretching region (after Fukuda et al., 2009c). The spectra at RT after heatings are also shown as gray lines. The integrated heating times are shown at the right of each spectrum.

Since contents of fluid water are reduced by dehydration, different degrees of hydrogen bonds of Si-OH at intergranular regions (surface silanol) to fluid water are formed. This leads to the appearances of several –OH bands (3660 and 3730 cm–1 at 500 °C, and possibly the split of the band at 3585 cm–1 at RT) (Yamagishi et al., 1997). The appearance of the band at 3730 cm–1 at 0 minute heating at 500 °C is due to slight dehydration during heating from RT to 500 °C at 100 °C/minute. The wavenumbers of these new –OH bands at high temperature are slightly different from those at RT, presumably due to thermal expansions of crystals and changes of hydrogen bond distances at high temperature.

## **4. H2O molecules in minerals: As an example of beryl**

80 Infrared Spectroscopy – Materials Science, Engineering and Technology

temperatures. For example, the maximum band height at 400 °C is approximately 50 % of that at RT. This is also because of decreases of average numbers of H2O in areas that IR light

When the sample is kept at high temperatures, dehydration occurs. High temperature IR spectra were continuously measured to monitor dehydration. Figure 3 shows dehydration behaviour measured at 500 °C (Fig. 3a) and 400 °C (Fig. 3b). Both of the experiments were performed by heating during 500 minutes in total. Broad bands around 3800-3000 cm–1 in both spectra decrease with keeping at high temperatures, and the wavenumbers of the broad bands are not changed during heating. IR spectra at RT after heating (gray spectra in Fig. 3) also show decreases of fluid water. This indicates that fluid water was dehydrated through intergranular regions which are fast paths for mass transfers (See Ingrin et al., 1995; Okumura and Nakashima, 2004; Fukuda et al., 2009c for estimation of water diffusivity). Over 50 % of the band areas are decreased during the heating in 84 minutes at 500 °C. Band areas of 80 % are decreased in 250 minutes, and the features of the spectra are not changed after that. The RT temperature spectrum after 500 minutes heating at 500 °C also shows significant reduction of board band due to fluid water. This remained band due to fluid water may reflect fluid inclusions, which is tightly trapped at open spaces in crystal structures. On the other hand, 60 % of the band areas are preserved after heating in 500 minutes at 400 °C, and the RT spectrum after 500 minutes heating at 400 °C still shows a

Fig. 3. Dehydration behaviour at 500 °C (left) and 400 °C (right) in the water stretching region (after Fukuda et al., 2009c). The spectra at RT after heatings are also shown as gray

Since contents of fluid water are reduced by dehydration, different degrees of hydrogen bonds of Si-OH at intergranular regions (surface silanol) to fluid water are formed. This leads to the appearances of several –OH bands (3660 and 3730 cm–1 at 500 °C, and possibly

lines. The integrated heating times are shown at the right of each spectrum.

captures (i.e., density; Schwarzer et al., 2005).

**3.3 Dehydration behaviour** 

strong signal of fluid water.

In addition to fluid water in rocks and –OH in mineral crystal structures as described above, isolated (not clustered) H2O molecules are incorporated in open cavities of crystal structures, and they are sometimes coupled with cations. H2O in open cavities has well been studied for beryl, a typical cyclosilicate (after Wood & Nassau, 1967). Ideal chemical formula of beryl is Be3Al2Si6O18, and six-membered SiO4 rings are stacked along the crystallographic *c*-axis and make a pipe-like cavity called a channel (Fig. 4) (Gibbs et al., 1968). Isolated H2O is trapped in the channel, and forms two kinds of orientations, depending on whether it coordinates to a cation (called type II) or not (called type I) (after Wood & Nassau, 1967). Such cations are trapped in the channels to compensate the electrical charge balances caused by Be2+-Li+ and Si4+-Al3+ substitutions and lacks of Be2+ in the crystal structure of beryl. The cations in the channels are assumed to be mainly Na+, and some other alkali cations may be incorporated (Hawthorne and Černý, 1977; Aurisicchio et al., 1988; Artioli et al., 1993; Andersson, 2006). In this section, I introduce polarized IR spectra of beryl, and discuss changes of the states of type I/II H2O in the channels by temperature changes and dehydration.

Fig. 4. Crystal structure of beryl. The (001)-plane (i.e., viewed down from the *c*-axis) and the channel section. Positions of type I/II H2O, a cation, and CO2 are also shown. Modified after Fukuda & Shinoda (2011).

#### **4.1 Chemical composition of the sample**

The chemical composition of the natural beryl sample used in this study was analyzed by Xray wavelength dispersive spectroscopy for major atomic contents, inductivity coupled plasma-atomic emission spectroscopy for Be content, and atomic absorption spectroscopy for Li and Rb contents (Table 1). The type I/II H2O contents were determined from intensities of IR bands due to the asymmetric stretching of type I and the symmetric stretching of type II in a polarized IR spectrum at RT (See the spectrum in the next section), using their molar absorption coefficients of 206 L mol–1 cm–1 and 256 L mol–1 cm–1,

Water in Rocks and Minerals – Species, Distributions, and Temperature Dependences 83

Fig. 5. Polarized IR spectra for natural beryl under different polarized conditions at RT. (a) From **E**//*c*-axis to **E***c*-axis in the (100)-section (the sample thickness of 20 μm). The angle of the *c*-axis (i.e., the direction of the channels) respective to **E** is shown on the left of each spectrum. (b) Under **E***c*-axis in the (001)-section (the sample thickness of 120 μm).

somewhat shows three bands at 1640, 1600, and 1546 cm–1 (e.g., Wood & Nassau 1967; Charoy et al., 1996; Łodziński et al., 2005), and I refer these three bands to the ν2-I related bands. The asymmetric stretching mode of CO2 molecules is detected at 2360 cm–1 under **E***c*-axis. These H2O and CO2 bands are not changed at any angles of **E** to the sample in the (001)-plane, corresponding to that these molecules are isotropically distributed due to the hexagonal symmetry of beryl. There are other unassigned bands; for example the sharp band at 3594 cm–1 can be seen. This band has been argued and might be due to Na+-OH in

the channels.

respectively (Goldman et al., 1977). The CO2 content was also calculated for the band at 2360 cm–1 from 800 L mol–1 cm–1 in Della Ventura et al. (2009). In the chemical composition, the Li and Na contents are relatively high in addition to the Si, Al and Be contents of major atoms. Be content (2.893 in 18 oxygen), which is lower than the ideal composition of beryl (Be3Al2Si6O18), must be replaced by Li, and Na must be incorporated as Na+ in the channels to compensate the electrical charge balance. However, Li+ may be also incorporated in the channels, since the Li content is not completely explained by Be2+-Li+ substitution (e.g., Hawthorne and Černý, 1977).


Table 1. Chemical composition of the beryl sample used in this study.

#### **4.2 Typical polarized IR spectra of beryl at RT and types of H2O**

Polarized IR spectra were measured by inserting a wire grid IR polarizer to IR light through the sample. Electric vector of IR light, **E** to the *c*-axis (i.e., the direction of the arraignments of the channels) were gradually changed and spectra were obtained (Fig. 5). Fundamental vibrations of type I/II H2O (asymmetric stretching; ν3, symmetric stretching; ν1, and bending modes; ν2) can be detected under different polarized conditions, which correspond to the orientations of type I/II H2O in the channels (Fig. 4) and IR active orientations of their vibrational modes: In a sample section of the (100)-plane (i.e., the section including the alignment of channels), the bands due to the ν3 mode of type I (referred to as ν3-I hereafter), ν1-II, and ν2-II are dominantly detected under **E**//*c*-axis (Fig. 5a). Their wavenumbers are 3698, 3597, and 1628 cm–1, respectively.

Under **E***c*-axis in the (100)-section (bottom spectra in Fig. 5a) or in the (001)-section (Fig. 5b), the bands due to ν3-II (3661 cm–1), ν1-I (3605 cm–1), and ν2-I are dominant. The ν2-I

respectively (Goldman et al., 1977). The CO2 content was also calculated for the band at 2360 cm–1 from 800 L mol–1 cm–1 in Della Ventura et al. (2009). In the chemical composition, the Li and Na contents are relatively high in addition to the Si, Al and Be contents of major atoms. Be content (2.893 in 18 oxygen), which is lower than the ideal composition of beryl (Be3Al2Si6O18), must be replaced by Li, and Na must be incorporated as Na+ in the channels to compensate the electrical charge balance. However, Li+ may be also incorporated in the channels, since the Li content is not completely explained by Be2+-Li+ substitution (e.g.,

Table 1. Chemical composition of the beryl sample used in this study.

**4.2 Typical polarized IR spectra of beryl at RT and types of H2O** 

3698, 3597, and 1628 cm–1, respectively.

Polarized IR spectra were measured by inserting a wire grid IR polarizer to IR light through the sample. Electric vector of IR light, **E** to the *c*-axis (i.e., the direction of the arraignments of the channels) were gradually changed and spectra were obtained (Fig. 5). Fundamental vibrations of type I/II H2O (asymmetric stretching; ν3, symmetric stretching; ν1, and bending modes; ν2) can be detected under different polarized conditions, which correspond to the orientations of type I/II H2O in the channels (Fig. 4) and IR active orientations of their vibrational modes: In a sample section of the (100)-plane (i.e., the section including the alignment of channels), the bands due to the ν3 mode of type I (referred to as ν3-I hereafter), ν1-II, and ν2-II are dominantly detected under **E**//*c*-axis (Fig. 5a). Their wavenumbers are

Under **E***c*-axis in the (100)-section (bottom spectra in Fig. 5a) or in the (001)-section (Fig. 5b), the bands due to ν3-II (3661 cm–1), ν1-I (3605 cm–1), and ν2-I are dominant. The ν2-I

Hawthorne and Černý, 1977).

Fig. 5. Polarized IR spectra for natural beryl under different polarized conditions at RT. (a) From **E**//*c*-axis to **E***c*-axis in the (100)-section (the sample thickness of 20 μm). The angle of the *c*-axis (i.e., the direction of the channels) respective to **E** is shown on the left of each spectrum. (b) Under **E***c*-axis in the (001)-section (the sample thickness of 120 μm).

somewhat shows three bands at 1640, 1600, and 1546 cm–1 (e.g., Wood & Nassau 1967; Charoy et al., 1996; Łodziński et al., 2005), and I refer these three bands to the ν2-I related bands. The asymmetric stretching mode of CO2 molecules is detected at 2360 cm–1 under **E***c*-axis. These H2O and CO2 bands are not changed at any angles of **E** to the sample in the (001)-plane, corresponding to that these molecules are isotropically distributed due to the hexagonal symmetry of beryl. There are other unassigned bands; for example the sharp band at 3594 cm–1 can be seen. This band has been argued and might be due to Na+-OH in the channels.

Water in Rocks and Minerals – Species, Distributions, and Temperature Dependences 85

beryl under **E**//*c*-axis (Fig. 6a) and **E***c*-axis (Fig. 6b) in the water stretching and bending regions. Spectral changes are different for each vibrational mode of type I/II H2O: Under **E**//*c*-axis at RT, the ν3-I and ν1-II bands are clearly seen at 3698 and 3597 cm–1, respectively. The ν3-I band is rapidly decreased in its height with increasing temperature; for example 60 % of the band height decreases at 200 °C, compared with that at RT (Fig. 7). The rapid decreases of the type I band are also seen for the ν1-I band and the ν2-I related bands under **E***c*-axis (Fig. 6b). Contrary to the case for the type I bands, only 20 % of the ν1-II and ν2-II bands are decreased at 200 °C. Alternatively, wavenumber shifts dominantly occur for these bands. The wavenumber of the former (3597 cm–1 at RT) and latter bands (1628 cm–1 at RT) linearly shift to lower and higher, with increasing temperature (Fig. 7). The changes of the ν3-II band (3661 cm–1 at RT) under **E***c*-axis are difficult to monitor because of overlapping with other bands at high temperatures. Since these changes are reversible upon heating and cooling, they are not due to dehydration but changes of the states of type I/II H2O in the

Fig. 7. Changes of band heights (left) and wavenumbers (right) of the ν3-I, ν1-II, and ν2-II bands with increasing temperature (Modified after Fukuda & Shinoda, 2011). Values are

These changes in band heights and wavenumbers of type I/II H2O are interpreted mainly due to the presence (type I) or absence of cations (type II) in the beryl channels (Fig. 4). Since type I H2O is not coupled with a cation, its position in the channels is easy lost with increasing temperature, resulting the rapid decreases of band heights. On the other hand, since the position of type II H2O is fixed by a cation (mainly Na+), the decreases in band heights with increasing temperature do not significantly occur, compared with those for type I bands. Alternatively, the wavenumber shifts occur. The inverse wavenumber shifts of type II bands in stretching and bending modes would be due to modifications in vibrational constants in the thermally-expanded beryl channels (Fukuda et al., 2009b), as similar to the

deviations of wavenumbers from ideal H2O molecule in RT spectra (Section 4.2).

channels. Discussion is as follows.

determined from the spectra in Fig. 6a.

Ideally, the ν3, ν1, and ν2 modes of isolated free H2O are detected at 3756, 3657, and 1595 cm–1, respectively at RT (Eisenberg & Kauzman, 1969). That the fundamental vibrations of type I/II H2O in beryl are deviated from those for ideal value, is due to interaction of type I/II H2O with channel oxygens and cations (Fig. 5). The wavenumbers of the stretching modes of type I/II H2O at RT are lower than those of free H2O, while those of the bending modes are higher. Falk (1984) experimentally demonstrated reverse correlations of band shifts between stretching and bending modes. The lower wavenumbers of the stretching modes than those of isolated water molecules are due to weak hydrogen bonds between type I/II and channel oxygens, similarly to the case for chalcedonic quartz. The reverse wavenumber shifts from ideal H2O between stretching and bending modes is mainly explained by changes in H-H repulsion constants in a simple spring model (Fukuda & Shinoda, 2008).

#### **4.3 Water vibrations at high temperatures**

Significant dehydration does not occur during short time heating from RT to 800 °C (temperature raise of 100 °C/minute and 1 minute for the measurements at each temperature; see Section 2). Figure 6 shows high temperature behaviour of type I/II H2O in

Fig. 6. High temperature behaviour of beryl from RT to 800 °C for the samples in Fig. 5 (Replotted from Fukuda & Shinoda, 2011). (a) under **E**//*c*-axis. (b) under **E***c*-axis. The RT spectra after heating at 800 °C are shown as gray lines, showing no significant dehyderation occured during heating.

Ideally, the ν3, ν1, and ν2 modes of isolated free H2O are detected at 3756, 3657, and 1595 cm–1, respectively at RT (Eisenberg & Kauzman, 1969). That the fundamental vibrations of type I/II H2O in beryl are deviated from those for ideal value, is due to interaction of type I/II H2O with channel oxygens and cations (Fig. 5). The wavenumbers of the stretching modes of type I/II H2O at RT are lower than those of free H2O, while those of the bending modes are higher. Falk (1984) experimentally demonstrated reverse correlations of band shifts between stretching and bending modes. The lower wavenumbers of the stretching modes than those of isolated water molecules are due to weak hydrogen bonds between type I/II and channel oxygens, similarly to the case for chalcedonic quartz. The reverse wavenumber shifts from ideal H2O between stretching and bending modes is mainly explained by changes in H-H

Significant dehydration does not occur during short time heating from RT to 800 °C (temperature raise of 100 °C/minute and 1 minute for the measurements at each temperature; see Section 2). Figure 6 shows high temperature behaviour of type I/II H2O in

Fig. 6. High temperature behaviour of beryl from RT to 800 °C for the samples in Fig. 5 (Replotted from Fukuda & Shinoda, 2011). (a) under **E**//*c*-axis. (b) under **E***c*-axis. The RT spectra after heating at 800 °C are shown as gray lines, showing no significant dehyderation

repulsion constants in a simple spring model (Fukuda & Shinoda, 2008).

**4.3 Water vibrations at high temperatures** 

occured during heating.

beryl under **E**//*c*-axis (Fig. 6a) and **E***c*-axis (Fig. 6b) in the water stretching and bending regions. Spectral changes are different for each vibrational mode of type I/II H2O: Under **E**//*c*-axis at RT, the ν3-I and ν1-II bands are clearly seen at 3698 and 3597 cm–1, respectively. The ν3-I band is rapidly decreased in its height with increasing temperature; for example 60 % of the band height decreases at 200 °C, compared with that at RT (Fig. 7). The rapid decreases of the type I band are also seen for the ν1-I band and the ν2-I related bands under **E***c*-axis (Fig. 6b). Contrary to the case for the type I bands, only 20 % of the ν1-II and ν2-II bands are decreased at 200 °C. Alternatively, wavenumber shifts dominantly occur for these bands. The wavenumber of the former (3597 cm–1 at RT) and latter bands (1628 cm–1 at RT) linearly shift to lower and higher, with increasing temperature (Fig. 7). The changes of the ν3-II band (3661 cm–1 at RT) under **E***c*-axis are difficult to monitor because of overlapping with other bands at high temperatures. Since these changes are reversible upon heating and cooling, they are not due to dehydration but changes of the states of type I/II H2O in the channels. Discussion is as follows.

Fig. 7. Changes of band heights (left) and wavenumbers (right) of the ν3-I, ν1-II, and ν2-II bands with increasing temperature (Modified after Fukuda & Shinoda, 2011). Values are determined from the spectra in Fig. 6a.

These changes in band heights and wavenumbers of type I/II H2O are interpreted mainly due to the presence (type I) or absence of cations (type II) in the beryl channels (Fig. 4). Since type I H2O is not coupled with a cation, its position in the channels is easy lost with increasing temperature, resulting the rapid decreases of band heights. On the other hand, since the position of type II H2O is fixed by a cation (mainly Na+), the decreases in band heights with increasing temperature do not significantly occur, compared with those for type I bands. Alternatively, the wavenumber shifts occur. The inverse wavenumber shifts of type II bands in stretching and bending modes would be due to modifications in vibrational constants in the thermally-expanded beryl channels (Fukuda et al., 2009b), as similar to the deviations of wavenumbers from ideal H2O molecule in RT spectra (Section 4.2).

Water in Rocks and Minerals – Species, Distributions, and Temperature Dependences 87

A cation is coordinated by one or two type II H2O due to a spatial restriction of the channel (Fig. 4). Therefore, the dominant band at 3597 (ν1-II) and 1628 cm–1 (ν2-II) before heating would be mainly due to doubly-coordinated type II to a cation, mainly Na+. Since type II H2O dehydrates by heating, singly-coordinated type II are created. The wavenumbers of singlyand doubly-coordinated H2O have been calculated for free H2O molecules. According to a numerical approach for water vibrations by Bauschlicher et al. (1991), the wavenumbers of H2O-Na+-H2O is higher in its stretching modes and lower in the bending modes than those of Na+-H2O in approximately 10 cm–1. This is consistent with wavenumber shifts in beryl due to

Another possibility for the wavenumber shifts of the ν1-II and ν2-II bands is the presence of Li+ in the channels. According to the calculation in Lee et al. (2004) for free H2O molecules, the wavenumbers of Li+-H2O is higher in its stretching modes and lower in the bending modes than those of Na+-H2O. Also, binding energy of Li+ to H2O is higher than that of Na+ to H2O, which indicates the stable stability of Li+-H2O during dehydration. If Li+ is trapped

A sharp and unassigned band is seen at 3594 cm–1 under **E***c*-axis. This band is also more stable than that for type I bands. The wavenumber of this band is different from any vibrational modes of type I/II H2O. Judging from the thermal stability and the wavenumber, this band may be related to Na+-OH in beryl, as its presence has been argued

Rocks are deformed at shear zones in the interior of the earth. Rocks, which underwent brittle and plastic deformation at shear zones, are called as cataclasites and mylonites, respectively. Brittle deformation of continental crusts (mainly granitoids) is dominated from the ground to 10-20 km depth. Plastic deformation of rocks is dominated below that with increasing temperature and pressure. Another important factor that significantly contributes to plastic deformation of rocks is water. Water contents in ppm order dramatically promote plastic deformation of minerals, as confirmed by deformation experiments (e.g., Griggs, 1967; Jaoul et al., 1984; Post & Tullis, 1998; Dimanov et al., 1999). Also, water contributes to solution-precipitation which sometimes involves reactions among minerals (especially, feldspar and mica in granitoids) (e.g., summarized in Thompson & Rubie, 1985; Dysthe & Wogelius, 2006). Then, solution-precipitation creep may also contribute to the strength of the crusts (Wintsch & Yi, 2002; Kenis et al., 2005). Thus, water contents and distribution as

In this section, I use IR spectroscopy to map two-dimensional water distributions as well as to consider its species in deformed granites. I especially focus on water distributions associated with solution-precipitation process of feldspar, and consider possible transport

Deformed granites were collected from outcrops in an inner shear zone of the Ryoke Metamorphic Belt in the Kishiwada district, Osaka Prefecture, SW Japan, and believed to be

the formation of singly-coordinated type II at 3587 (ν1-II) and 1638 cm–1 (ν2-II).

in the beryl channels, it can cause the wavenumber shifts observed in this study.

**5. IR mapping measurements for deformed rocks** 

well as its species are important for rock deformation.

in Andersson (2006).

mechanisms of water.

**5.1 Samples and analyses** 

#### **4.4 Dehydration behaviour**

The beryl sample was heated on the heating stage at 850 °C where dehydration is enhanced (Fukuda & Shinoda, 2008). Polarized IR spectra are shown only at RT quenched from 850 °C (Fig. 8), since in-situ high temperature IR spectra at 850 °C show broadened water bands (Fig. 6) and changes of each bands are not clearly monitored. Under **E**//*c*-axis, the ν3-I band at 3698 cm–1 disappeared at heating of 12 hours, without any wavenumber changes (Fig. 8a). This trend is same with the band at 3605 cm–1 (ν1-I) and three ν2-I-related bands under **E***c*axis (Fig. 8b). Some bands remain in the water bending region under **E***c*-axis, and they would be due to structural vibrations of beryl. The dehydration behaviour of type II bands are different with that of type I: Since the position of type II H2O is fixed by a cation, its dehydration is obviously slower than type I H2O. Under **E**//*c*-axis, new bands develop at 3587 and 1638 cm–1 with decreasing of the initial bands at 3597 (ν1-II) and 1628 cm–1 (ν2-II). The band at 3661 cm–1 (ν3-II) under **E***c*-axis also shows the wavenumber shifts to the lower with decreasing its intensity. These bands are stable after 24 hours heating. The appearances of these bands are explained as follows.

Fig. 8. IR spectra at RT quenched from heating at 850 °C, showing dehydration behaviour (Replotted from Fukuda & Shinoda, 2011). (a) under **E**//*c*-axis. (b) under **E***c*-axis. Heating times at 850 °C are shown at the right of each spectrum.

The beryl sample was heated on the heating stage at 850 °C where dehydration is enhanced (Fukuda & Shinoda, 2008). Polarized IR spectra are shown only at RT quenched from 850 °C (Fig. 8), since in-situ high temperature IR spectra at 850 °C show broadened water bands (Fig. 6) and changes of each bands are not clearly monitored. Under **E**//*c*-axis, the ν3-I band at 3698 cm–1 disappeared at heating of 12 hours, without any wavenumber changes (Fig. 8a). This trend is same with the band at 3605 cm–1 (ν1-I) and three ν2-I-related bands under **E***c*axis (Fig. 8b). Some bands remain in the water bending region under **E***c*-axis, and they would be due to structural vibrations of beryl. The dehydration behaviour of type II bands are different with that of type I: Since the position of type II H2O is fixed by a cation, its dehydration is obviously slower than type I H2O. Under **E**//*c*-axis, new bands develop at 3587 and 1638 cm–1 with decreasing of the initial bands at 3597 (ν1-II) and 1628 cm–1 (ν2-II). The band at 3661 cm–1 (ν3-II) under **E***c*-axis also shows the wavenumber shifts to the lower with decreasing its intensity. These bands are stable after 24 hours heating. The appearances

Fig. 8. IR spectra at RT quenched from heating at 850 °C, showing dehydration behaviour (Replotted from Fukuda & Shinoda, 2011). (a) under **E**//*c*-axis. (b) under **E***c*-axis. Heating

times at 850 °C are shown at the right of each spectrum.

**4.4 Dehydration behaviour** 

of these bands are explained as follows.

A cation is coordinated by one or two type II H2O due to a spatial restriction of the channel (Fig. 4). Therefore, the dominant band at 3597 (ν1-II) and 1628 cm–1 (ν2-II) before heating would be mainly due to doubly-coordinated type II to a cation, mainly Na+. Since type II H2O dehydrates by heating, singly-coordinated type II are created. The wavenumbers of singlyand doubly-coordinated H2O have been calculated for free H2O molecules. According to a numerical approach for water vibrations by Bauschlicher et al. (1991), the wavenumbers of H2O-Na+-H2O is higher in its stretching modes and lower in the bending modes than those of Na+-H2O in approximately 10 cm–1. This is consistent with wavenumber shifts in beryl due to the formation of singly-coordinated type II at 3587 (ν1-II) and 1638 cm–1 (ν2-II).

Another possibility for the wavenumber shifts of the ν1-II and ν2-II bands is the presence of Li+ in the channels. According to the calculation in Lee et al. (2004) for free H2O molecules, the wavenumbers of Li+-H2O is higher in its stretching modes and lower in the bending modes than those of Na+-H2O. Also, binding energy of Li+ to H2O is higher than that of Na+ to H2O, which indicates the stable stability of Li+-H2O during dehydration. If Li+ is trapped in the beryl channels, it can cause the wavenumber shifts observed in this study.

A sharp and unassigned band is seen at 3594 cm–1 under **E***c*-axis. This band is also more stable than that for type I bands. The wavenumber of this band is different from any vibrational modes of type I/II H2O. Judging from the thermal stability and the wavenumber, this band may be related to Na+-OH in beryl, as its presence has been argued in Andersson (2006).

## **5. IR mapping measurements for deformed rocks**

Rocks are deformed at shear zones in the interior of the earth. Rocks, which underwent brittle and plastic deformation at shear zones, are called as cataclasites and mylonites, respectively. Brittle deformation of continental crusts (mainly granitoids) is dominated from the ground to 10-20 km depth. Plastic deformation of rocks is dominated below that with increasing temperature and pressure. Another important factor that significantly contributes to plastic deformation of rocks is water. Water contents in ppm order dramatically promote plastic deformation of minerals, as confirmed by deformation experiments (e.g., Griggs, 1967; Jaoul et al., 1984; Post & Tullis, 1998; Dimanov et al., 1999). Also, water contributes to solution-precipitation which sometimes involves reactions among minerals (especially, feldspar and mica in granitoids) (e.g., summarized in Thompson & Rubie, 1985; Dysthe & Wogelius, 2006). Then, solution-precipitation creep may also contribute to the strength of the crusts (Wintsch & Yi, 2002; Kenis et al., 2005). Thus, water contents and distribution as well as its species are important for rock deformation.

In this section, I use IR spectroscopy to map two-dimensional water distributions as well as to consider its species in deformed granites. I especially focus on water distributions associated with solution-precipitation process of feldspar, and consider possible transport mechanisms of water.

#### **5.1 Samples and analyses**

Deformed granites were collected from outcrops in an inner shear zone of the Ryoke Metamorphic Belt in the Kishiwada district, Osaka Prefecture, SW Japan, and believed to be

Water in Rocks and Minerals – Species, Distributions, and Temperature Dependences 89

Fig. 9. (a) Optical microscopic image including the IR mapped area (bold square). (b) SEM-BSE image of the IR mapped area shown in (a): light gray; K-feldspar, medium gray; plagioclase, dark gray; quartz. (c) Water distribution mapped by integral absorbance of water stretching bands in the 3800–2750 cm–1 range of the IR spectra. The color contours from black to red approximately correspond to water contents from low to high. The boundaries of minerals are shown as dotted lines. Arrows with letters show the locations used to illustrate the various selected IR spectra in (d). Pl; plagioclase, Qtz; quartz.

the bending vibrations of fluid water. Other bands in the range of 2500–1500 cm–1 are due to the structural vibrations of plagioclase. Six sharp bands between 2000 and 1500 cm–1 for quartz are due to the structural vibrations of quartz, which are same with the spectra for chalcedonic quartz (Fig. 1). Recognition of the bending vibrations of fluid water in quartz is difficult because of these sharp structural bands. Water concentration in this plagioclase porphyroclast ranges from 200 to 700 ppm, with an average of 450 ppm, consistent with values reported in the literature (Hofmeister & Rossman, 1985; Beran, 1987; Johnson & Rossman, 2003). The heterogeneity of water distribution in plagioclase does not directly correspond to textures under the optical microscope and BSE. The amount of water in the quartz is much lower than that in the plagioclase, ranging from 80 to 300 ppm, with an

average of 130 ppm.

deformed at ~500 °C (Takagi, 1988; Imon et al., 2002; 2004). Sample thin sections of ~50 μm were at first observed under a polarized optical microscope and a back-scattered electron (BSE) image which reflects compositional differences in a scanning electron microscope (SEM). After IR mapping measurements, thin sections were again polished to suitable thickness (~20 μm) to observe detailed microstructures under the optical microscope.

IR mapping measurements were carried out along ~1000 μm traverses with 30 μm spatial resolution (aperture size) in steps of 30 μm. See Section 2 for the instrument of IR spectroscopy. The integral absorbances of the water stretching bands in the range 3800–2750 cm–1 are displayed as a color-contoured image for a measured sample area. Color-contoured images can be used as a qualitative representation of the distribution of water, since absorption coefficients tend to increase linearly with decreasing wavenumbers (Paterson, 1982; Libowitzky & Rossman, 1997). When the absolute water contents of the minerals are to be determined, Beer–Lambert law is applied, using the absorption coefficients for each mineral. For K-feldspar and plagioclase, I used the absorption coefficients of integral water stretching bands reported by Johnson & Rossman (2003) (15.3 ppm–1 cm–2), and for quartz, those reported by Kats (1962). Kats (1962) reported following relation between water content in quartz and integral absorbance of water stretching bands; C(H/106 Si)=0.812 x Aint/d, where Aint is the integral absorbance and d is the sample thickness in cm. Then, I converted H/106 Si value to a ppm H2O unit (1 ppm = 6.67 H/106 Si), as also adopted in Gleason & DeSisto (2008). To distinguish Si-OH and H2O contents separately, their combination bands of the stretching and bending modes should be used, as shown in Section 3. However, it is difficult for these samples, since sample thicknesses are thin for texture observations under the optical microscope, and 1 mm thickness is needed to measure the combination bands.

## **5.2 Typical water distribution in deformed granite**

At first, I introduce water distribution in the granite mylonite with typical microtexture (Fig. 9) (See Passchier & Trouw, 2005 for many textures of deformed rocks). As can be seen under the polarized optical microscope (Fig. 9a), quartz is recrystallized by subgrain rotation, and plastically deformed by dislocation creep. Quartz grains are elongated with the aspect ratio of ca. 3:1 and the long axis is ca. 250 μm. Feldspar is a relatively hard mineral in this deformation condition, is not plastically deformed, and behaves as rigid body sometimes with fracturing (i.e., brittle deformation). Such relatively hard minerals are called as porphyroclasts. Under the BSE image, rims of plagioclase are replaced by K-feldspar, which would be due to solution-precipitation with or without reaction called myrmekitization (e.g., Simpson & Wintsch, 1989). Myrmekitization is the following reaction; K-feldspar + Na+ + Ca2+ = plagioclase + quartz + K+, where cations are included in circulating fluid water. Water contents in these replaced K-feldspar are difficult to determine in this region because of the limitation of its distribution, and discussed for other regions later.

The IR spectra for both plagioclase and quartz show broad bands at 3800–2750 cm–1, which is due to the stretching vibration of fluid water (Fig. 9d) (See Section 3). Fluid water must be trapped as fluid inclusions within both minerals, since spatial resolution of 30 μm (aperture size) covers intracrystalline regions, rather than intergranular regions. The IR spectra for the plagioclase porphyroclast also exhibit sharp bands at 3625 and 3700 cm–1, which are due to the stretching vibrations of structural hydroxyl in plagioclase (e.g., Hofmeister & Rossman, 1985; Beran, 1987; Johnson & Rossman, 2003). The band at 1620 cm–1 in plagioclase is due to

deformed at ~500 °C (Takagi, 1988; Imon et al., 2002; 2004). Sample thin sections of ~50 μm were at first observed under a polarized optical microscope and a back-scattered electron (BSE) image which reflects compositional differences in a scanning electron microscope (SEM). After IR mapping measurements, thin sections were again polished to suitable

IR mapping measurements were carried out along ~1000 μm traverses with 30 μm spatial resolution (aperture size) in steps of 30 μm. See Section 2 for the instrument of IR spectroscopy. The integral absorbances of the water stretching bands in the range 3800–2750 cm–1 are displayed as a color-contoured image for a measured sample area. Color-contoured images can be used as a qualitative representation of the distribution of water, since absorption coefficients tend to increase linearly with decreasing wavenumbers (Paterson, 1982; Libowitzky & Rossman, 1997). When the absolute water contents of the minerals are to be determined, Beer–Lambert law is applied, using the absorption coefficients for each mineral. For K-feldspar and plagioclase, I used the absorption coefficients of integral water stretching bands reported by Johnson & Rossman (2003) (15.3 ppm–1 cm–2), and for quartz, those reported by Kats (1962). Kats (1962) reported following relation between water content in quartz and integral absorbance of water stretching bands; C(H/106 Si)=0.812 x Aint/d, where Aint is the integral absorbance and d is the sample thickness in cm. Then, I converted H/106 Si value to a ppm H2O unit (1 ppm = 6.67 H/106 Si), as also adopted in Gleason & DeSisto (2008). To distinguish Si-OH and H2O contents separately, their combination bands of the stretching and bending modes should be used, as shown in Section 3. However, it is difficult for these samples, since sample thicknesses are thin for texture observations under the optical microscope, and 1 mm thickness is needed to measure the combination bands.

At first, I introduce water distribution in the granite mylonite with typical microtexture (Fig. 9) (See Passchier & Trouw, 2005 for many textures of deformed rocks). As can be seen under the polarized optical microscope (Fig. 9a), quartz is recrystallized by subgrain rotation, and plastically deformed by dislocation creep. Quartz grains are elongated with the aspect ratio of ca. 3:1 and the long axis is ca. 250 μm. Feldspar is a relatively hard mineral in this deformation condition, is not plastically deformed, and behaves as rigid body sometimes with fracturing (i.e., brittle deformation). Such relatively hard minerals are called as porphyroclasts. Under the BSE image, rims of plagioclase are replaced by K-feldspar, which would be due to solution-precipitation with or without reaction called myrmekitization (e.g., Simpson & Wintsch, 1989). Myrmekitization is the following reaction; K-feldspar + Na+ + Ca2+ = plagioclase + quartz + K+, where cations are included in circulating fluid water. Water contents in these replaced K-feldspar are difficult to determine in this region because

The IR spectra for both plagioclase and quartz show broad bands at 3800–2750 cm–1, which is due to the stretching vibration of fluid water (Fig. 9d) (See Section 3). Fluid water must be trapped as fluid inclusions within both minerals, since spatial resolution of 30 μm (aperture size) covers intracrystalline regions, rather than intergranular regions. The IR spectra for the plagioclase porphyroclast also exhibit sharp bands at 3625 and 3700 cm–1, which are due to the stretching vibrations of structural hydroxyl in plagioclase (e.g., Hofmeister & Rossman, 1985; Beran, 1987; Johnson & Rossman, 2003). The band at 1620 cm–1 in plagioclase is due to

of the limitation of its distribution, and discussed for other regions later.

thickness (~20 μm) to observe detailed microstructures under the optical microscope.

**5.2 Typical water distribution in deformed granite** 

Fig. 9. (a) Optical microscopic image including the IR mapped area (bold square). (b) SEM-BSE image of the IR mapped area shown in (a): light gray; K-feldspar, medium gray; plagioclase, dark gray; quartz. (c) Water distribution mapped by integral absorbance of water stretching bands in the 3800–2750 cm–1 range of the IR spectra. The color contours from black to red approximately correspond to water contents from low to high. The boundaries of minerals are shown as dotted lines. Arrows with letters show the locations used to illustrate the various selected IR spectra in (d). Pl; plagioclase, Qtz; quartz.

the bending vibrations of fluid water. Other bands in the range of 2500–1500 cm–1 are due to the structural vibrations of plagioclase. Six sharp bands between 2000 and 1500 cm–1 for quartz are due to the structural vibrations of quartz, which are same with the spectra for chalcedonic quartz (Fig. 1). Recognition of the bending vibrations of fluid water in quartz is difficult because of these sharp structural bands. Water concentration in this plagioclase porphyroclast ranges from 200 to 700 ppm, with an average of 450 ppm, consistent with values reported in the literature (Hofmeister & Rossman, 1985; Beran, 1987; Johnson & Rossman, 2003). The heterogeneity of water distribution in plagioclase does not directly correspond to textures under the optical microscope and BSE. The amount of water in the quartz is much lower than that in the plagioclase, ranging from 80 to 300 ppm, with an average of 130 ppm.

Water in Rocks and Minerals – Species, Distributions, and Temperature Dependences 91

Figure 11 shows water distribution in an area that dominantly includes fine-grained plagioclase. The fine-grained plagioclase develops around plagioclase porphyroclasts, and some of it is closely associated with K-feldspar under the BSE image (Fig. 11b). Quartz in this region can be identified from its characteristic structural vibrations in the IR spectra (Fig. 11d), indicating that myrmekitization (K-feldspar + Na+ + Ca2+ = plagioclase + quartz + K+; see Section 5.2) occured during rocks deformation. Water contents in the area where fine-grained plagioclase grains are associated with small amounts of K-feldspar and quartz are roughly 2–4 times lower than those within plagioclase porphyroclasts, as inferred from the color contrasts in the IR mapping image (Fig. 11c). However, it is not possible to measure the absolute water contents in this area because of a mixture of plagioclase, K-

Fig. 11. (a) Optical microscopic image including the IR mapped area (bold square). (b) SEM-

BSE image of the mapped area shown in (a): light gray; K-feldspar, medium gray; plagioclase, dark gray; quartz. (c) Water distribution mapped by integral absorbance of water stretching bands in the 3800–2750 cm–1 range of the IR spectra. The boundaries of minerals are shown as dotted lines. Arrows with letters show the locations used to illustrate the various selected IR spectra in (d). Pl; plagioclase, FGP; fine-grained plagioclase, Qtz; quartz. IR spectra show quartz vibrational bands in fine-grained plagioclase. Absolute water contents for fine-grained plagioclase are not determined due to the mixtures of quartz and

K-feldspar.

feldspar, and quartz; consequently, the absorption coefficients are not clear.

#### **5.3 Water distribution around feldspar fine grains and possible water transportation**

Water distribution was measured for an area where fine-grained K-feldspar develops around K-feldspar and plagioclase porphyroclasts (Fig. 10a). The BSE image shows that finegrained K-feldspar regions, which are constructed by ~20 μm grains, contain patchy distribution of plagioclase (Fig. 10b). This indicates that solution-precipitation of K-feldspar, which may be accompanied with myrmekitization, occurred for the development of fine grains. The IR-mapped image shows that water contents in these regions are 220 ppm H2O in average; low and homogeneously distributed, although fluid water must be participated in the solution-precipitation process. The features of water stretching bands of fine-grained K-feldspar and K-feldspar porphyroclasts do not show structural –OH bands, differently from plagioclase (Fig. 10d); only broad bands can be seen at 3800–2750 cm–1. Water contents in K-feldspar and plagioclase porphyroclasts are 200-1150 ppm; heterogeneously distributed compared with those in fine-grained K-feldspar regions.

Fig. 10. (a) Optical microscopic image including the IR mapped area (bold square). (b) SEM-BSE image of the mapped area shown in (a): light gray; K-feldspar, dark gray; quartz. Finegrained K-feldspar regions contain patchy-distributed plagioclase, indicating solutionprecipitation occurred for the developments of these regions. (c) Water distribution mapped by integral absorbance of water stretching bands in the 3800–2750 cm–1 range of the IR spectra. The boundaries of minerals are shown as dotted lines. Arrows with letters show the locations used to illustrate the various selected IR spectra in (d). Pl; plagioclase, Kfs; Kfeldspar, FGK; fine-grained K-feldspar, Qtz; quartz.

**5.3 Water distribution around feldspar fine grains and possible water transportation**  Water distribution was measured for an area where fine-grained K-feldspar develops around K-feldspar and plagioclase porphyroclasts (Fig. 10a). The BSE image shows that finegrained K-feldspar regions, which are constructed by ~20 μm grains, contain patchy distribution of plagioclase (Fig. 10b). This indicates that solution-precipitation of K-feldspar, which may be accompanied with myrmekitization, occurred for the development of fine grains. The IR-mapped image shows that water contents in these regions are 220 ppm H2O in average; low and homogeneously distributed, although fluid water must be participated in the solution-precipitation process. The features of water stretching bands of fine-grained K-feldspar and K-feldspar porphyroclasts do not show structural –OH bands, differently from plagioclase (Fig. 10d); only broad bands can be seen at 3800–2750 cm–1. Water contents in K-feldspar and plagioclase porphyroclasts are 200-1150 ppm; heterogeneously distributed

Fig. 10. (a) Optical microscopic image including the IR mapped area (bold square). (b) SEM-BSE image of the mapped area shown in (a): light gray; K-feldspar, dark gray; quartz. Finegrained K-feldspar regions contain patchy-distributed plagioclase, indicating solutionprecipitation occurred for the developments of these regions. (c) Water distribution mapped by integral absorbance of water stretching bands in the 3800–2750 cm–1 range of the IR spectra. The boundaries of minerals are shown as dotted lines. Arrows with letters show the locations used to illustrate the various selected IR spectra in (d). Pl; plagioclase, Kfs; K-

compared with those in fine-grained K-feldspar regions.

feldspar, FGK; fine-grained K-feldspar, Qtz; quartz.

Figure 11 shows water distribution in an area that dominantly includes fine-grained plagioclase. The fine-grained plagioclase develops around plagioclase porphyroclasts, and some of it is closely associated with K-feldspar under the BSE image (Fig. 11b). Quartz in this region can be identified from its characteristic structural vibrations in the IR spectra (Fig. 11d), indicating that myrmekitization (K-feldspar + Na+ + Ca2+ = plagioclase + quartz + K+; see Section 5.2) occured during rocks deformation. Water contents in the area where fine-grained plagioclase grains are associated with small amounts of K-feldspar and quartz are roughly 2–4 times lower than those within plagioclase porphyroclasts, as inferred from the color contrasts in the IR mapping image (Fig. 11c). However, it is not possible to measure the absolute water contents in this area because of a mixture of plagioclase, Kfeldspar, and quartz; consequently, the absorption coefficients are not clear.

Fig. 11. (a) Optical microscopic image including the IR mapped area (bold square). (b) SEM-BSE image of the mapped area shown in (a): light gray; K-feldspar, medium gray; plagioclase, dark gray; quartz. (c) Water distribution mapped by integral absorbance of water stretching bands in the 3800–2750 cm–1 range of the IR spectra. The boundaries of minerals are shown as dotted lines. Arrows with letters show the locations used to illustrate the various selected IR spectra in (d). Pl; plagioclase, FGP; fine-grained plagioclase, Qtz; quartz. IR spectra show quartz vibrational bands in fine-grained plagioclase. Absolute water contents for fine-grained plagioclase are not determined due to the mixtures of quartz and K-feldspar.

Water in Rocks and Minerals – Species, Distributions, and Temperature Dependences 93

I thank K. Shinoda, J. Muto, T. Okudaira, and T. Hirono for their helpful comments on the manuscript. K. Shinoda is especially thanked for the advice on IR measurements and supports on sample preparation of beryl. T. Okudaira is also thanked for the supports for sample collection of the deformed rocks. This work was financially supported by a Grant-in-Aid for Scientific Research (212327) and (233694) by the Japan Society for the Promotion of

Aines, R.D. & Rossman, G.R. (1984). Water in minerals? A peak in infrared. *Journal of* 

Andersson, L.O. (2006). The positionof H+, Li+ and Na+ impurities in beryl. *Physics and* 

Artioli, G., Rinaldi, R., Ståhl, K. & Zanazzi, P.F. (1993). Structure refinements of beryl by

Aurisicchio, C., Fioravanti, G., Grubessi, O. & Zanazzi, P.F. (1988). Reappraisal of the crystal

Bauschlicher, C.W., Langhoff, S.R., Partridge, H., Rice, J.E. & Komornicki, A. (1991). A theoretical study of Na(H2O)+n (n = 1-4). *Journal of Chemical Physics*, 95, 5142-5148. Beran, A., (1987). OH groups in nominally anhydrous framework structures: an infrared

Brubach, J.B., Mermet, A., Filabozzi, A., Gerschel, A. & Roy, A. (2005). Signatures of the

Charoy, B., de Donato, P., Barres, O. & Pinto-Coelho, C. (1996). Channel occupancy in an

Della Ventura, G., Bellatreccia, F., Cesare, B., Harley, S. & Piccinini, M. (2009). FTIR

Dimanov, A., Dresen, G., Xiao, X., & Wirth, R. (1999). Grain boundary diffusion creep of

Dysthe, D.K. & Wogelius, R.A. (2006). Confined fluids in the Earth's crust – Properties and

Eisenberg, D. & Kauzman, W. (1969). *The Structure and Properties of Water,* Oxford University

Falk, M. (1984). The frequency of the H–O–H bending fundamental in solids and liquids.

Fitz Gerald, J.D. & Stünitz (1993). Deformation of granitoids at low metamorphic grade. I: reactions and grain size reduction. *Tectonophysics*, Vol. 221, pp. 269-297 Frondel, C. (1982). Structural hydroxyl in chalcedony (Type B quartz). *American Mineralogist*,

chemistry of beryl. *American Mineralogist,* Vol. 73, pp. 826-837

single-crystal neutron and X-ray diffraction. *American Mineralogist,* Vol. 78, pp. 762-

spectroscopic investigation of Danburite and Labradorite. *Physics and Chemistry of* 

hydrogen bonding in the infrared bands of water. *Journal of Chemical Physics*, Vol.

alkali-poor beryl from Serra Branca (Goias, Brazil): spectroscopic characterization.

microspectroscopy and SIMS study of water-poor cordierite from El Hoyazo, Spain: application to mineral and melt devolatilization. *Lithos*, Vol. 113, pp. 498-506

synthetic anorthite aggregates: the effect of water. *Journal of Geophysical Research*,

*Geophysical Research,* Vol. 89, pp. 4059-4071

*Chemistry of Minerals*, Vol. 33, pp. 403-416

*American Mineralogist*, Vol. 81, pp. 395-403

processes. *Chemical Geology*, Vol. 230, pp. 175-181

*Spectrochimica Acta*, Vol. A40, pp. 43-48

*Minerals*, Vol. 14, pp. 441-445

122, Article No. 184509

Vol. 104, pp. 10483-10497

Vol. 67. pp. 1248-1257

Press, Oxford

**7. Acknowledgment** 

Science for Young Scientists.

**8. References** 

768

In general, fluid water can be trapped at intergranular regions (grain boundaries) up to a few thousand ppm H2O, as reported for quartz aggregates whose each grain size is a few tens of micrometers to a few hundred micrometers (e.g., Nakashima et al., 1995; Muto et al., 2004; O'kane et al., 2005). In this study, intergranular regions must also be covered in the measurements for fine-grained K-feldspar- and plagioclase-dominant regions. The solution-precipitation process that produces fined-grained K-feldspar and plagioclase are subsequently and/or simultaneously enhanced by the ability of fluid water along intergranular regions to carry ions in solution, especially in situations where intergranular diffusion was promoted by an increase in surface area. (e.g., Simpson & Wintch, 1989; Fitz Gerald & Stünitz, 1993; Tsurumi et al., 2003). However, contrary to the previous knowledge on water contents at intergranular regions, water contents in fine-grained Kfeldspar- and plagioclase-dominant regions are low and homogeneous (av. 250 ppm). Therefore, it can be inferred that fluid water may not abundantly be trapped in newlycreated grains, and rather released during and/or after the solution-precipitation process. The release of fluid water and dissolved ions, which contributed to the process, was a result of the concentration gradients formed in the context of numerous newly-created intergranular regions within the mix of fine-grained feldspar and quartz. As a consequence, the entire process results in a positive feedback to promote the solutionprecipitation process.

## **6. Conclusions**

I investigated high temperature behaviour of water in rocks and minerals. Chalcedonic quartz was used as a representative rock which contains abundant fluid water at intergranular regions and –OH in quartz crystal structures. The average coordination numbers of water molecules were degreased with increasing temperature, which causes shifts of stretching vibrations to higher wavenumbers. Dehydration of fluid water in the chalcedonic quartz was monitored by keeping at high temperatures. Fluid water was rapidly dehydrated through intergranular regions at 500 °C, and new hydroxyl bands appeared with dehydration of fluid water.

States of water molecules, which are not clustered like fluid water, were investigated for beryl, typical cyclosilicate, using high temperature polarized IR spectroscopy. The beryl channels, open cavities in the crystal structures, contain two types of water molecules which freely exist or coordinate to cations from up and/or below them. The former type of water easily looses its specific position, resulting the rapid degreases of its IR band heights without dehydration, and shows rapid dehydration at 850 °C. The latter type of water shows significant changes in its wavenumbers with increasing temperatures. There are slight modifications in the wavenumbers during dehydration due to changes of coordination to cations during dehydration.

Distribution of fluid water was measured for deformed granites. K-feldspar and plagioclase fine grains were formed around porphyroclasts by solution-precipitation process. Water contents in fine-grained K-feldspar- and plagioclase-dominant regions show low and homogeneous distribution of fluid water, while water distributions in host porphyroclasts were heterogeneous. This indicates that fluid water, which was involved in the solution-precipitation process, was released during and/or after the solutionprecipitation process.

#### **7. Acknowledgment**

92 Infrared Spectroscopy – Materials Science, Engineering and Technology

In general, fluid water can be trapped at intergranular regions (grain boundaries) up to a few thousand ppm H2O, as reported for quartz aggregates whose each grain size is a few tens of micrometers to a few hundred micrometers (e.g., Nakashima et al., 1995; Muto et al., 2004; O'kane et al., 2005). In this study, intergranular regions must also be covered in the measurements for fine-grained K-feldspar- and plagioclase-dominant regions. The solution-precipitation process that produces fined-grained K-feldspar and plagioclase are subsequently and/or simultaneously enhanced by the ability of fluid water along intergranular regions to carry ions in solution, especially in situations where intergranular diffusion was promoted by an increase in surface area. (e.g., Simpson & Wintch, 1989; Fitz Gerald & Stünitz, 1993; Tsurumi et al., 2003). However, contrary to the previous knowledge on water contents at intergranular regions, water contents in fine-grained Kfeldspar- and plagioclase-dominant regions are low and homogeneous (av. 250 ppm). Therefore, it can be inferred that fluid water may not abundantly be trapped in newlycreated grains, and rather released during and/or after the solution-precipitation process. The release of fluid water and dissolved ions, which contributed to the process, was a result of the concentration gradients formed in the context of numerous newly-created intergranular regions within the mix of fine-grained feldspar and quartz. As a consequence, the entire process results in a positive feedback to promote the solution-

I investigated high temperature behaviour of water in rocks and minerals. Chalcedonic quartz was used as a representative rock which contains abundant fluid water at intergranular regions and –OH in quartz crystal structures. The average coordination numbers of water molecules were degreased with increasing temperature, which causes shifts of stretching vibrations to higher wavenumbers. Dehydration of fluid water in the chalcedonic quartz was monitored by keeping at high temperatures. Fluid water was rapidly dehydrated through intergranular regions at 500 °C, and new hydroxyl bands

States of water molecules, which are not clustered like fluid water, were investigated for beryl, typical cyclosilicate, using high temperature polarized IR spectroscopy. The beryl channels, open cavities in the crystal structures, contain two types of water molecules which freely exist or coordinate to cations from up and/or below them. The former type of water easily looses its specific position, resulting the rapid degreases of its IR band heights without dehydration, and shows rapid dehydration at 850 °C. The latter type of water shows significant changes in its wavenumbers with increasing temperatures. There are slight modifications in the wavenumbers during dehydration due to changes of

Distribution of fluid water was measured for deformed granites. K-feldspar and plagioclase fine grains were formed around porphyroclasts by solution-precipitation process. Water contents in fine-grained K-feldspar- and plagioclase-dominant regions show low and homogeneous distribution of fluid water, while water distributions in host porphyroclasts were heterogeneous. This indicates that fluid water, which was involved in the solution-precipitation process, was released during and/or after the solution-

precipitation process.

appeared with dehydration of fluid water.

coordination to cations during dehydration.

precipitation process.

**6. Conclusions** 

I thank K. Shinoda, J. Muto, T. Okudaira, and T. Hirono for their helpful comments on the manuscript. K. Shinoda is especially thanked for the advice on IR measurements and supports on sample preparation of beryl. T. Okudaira is also thanked for the supports for sample collection of the deformed rocks. This work was financially supported by a Grant-in-Aid for Scientific Research (212327) and (233694) by the Japan Society for the Promotion of Science for Young Scientists.

#### **8. References**


Water in Rocks and Minerals – Species, Distributions, and Temperature Dependences 95

Ingrin, J., Hercule, S. & Charton, T. (1995). Diffusion of hydrogen in diopside: results of dehydration experiments. *Journal of Geophysical Research*, Vol. 100, pp. 15489-15499 Jacobsen, S.D. & Van der Lee, S. (2006). *Earth's Deep Water Cycle*, Geophysical Monograph

Jaoul, O. (1984). Sodium weakening of Heavitree quartzite: preliminary results. *Journal of* 

Johnson, E. & Rossman, G.R. (2003). The concentration and speciation of hydrogen in

Kats, A. (1962). Hydrogen in alpha quartz. *Philips Research Report*, Vol. 17, pp. 1-31, 133–195,

Kenis, I., Urai, J.L., Van der Zee, W., Hilgers, C. & Sintubin, M. (2005). Rheology of fine-

numerical analysis. *Earth and Planetary Science Letters*, Vol. 233, pp. 351-360 Keppler, H. & Smyth, J.R. (2006). *Water in Nominally Anhydrous Minerals,* Reviews in Mineralogy and Geochemistry, Vol.62, The Mineralogical Society of America Kihara, K. (2001). Moleclar dynamics interpretation of structural changes in quartz, *Physics* 

Kronenberg, A.K. & Wolf, G.H. (1990). Fourier transform infrared spectroscopy

Lee, H.M., Tarakeshwar, P., Park, J., Kołaski, M.R., Yoon, Y.J., Yi, H.B., Kim, W.Y. & Kim,

Łodziński, M., Sitarz, M., Stec, K., Kozanecki, M., Fojud, Z. & Jurga, S. (2005). ICP, IR,

Muto, J., Nagahama, H., & Hashimoto, T. (2004). Microinfrared reflection spectroscopic

Nakahara, M., Matubayasi, N., Wakai, C. & Tsujino, Y. (2001). Structure and dynamics of water: from ambient to supercritical. *Journal of Molecular Liquids*, Vol. 90, pp. 75-83 Nakashima, S., Matayoshi, H., Yuko, T., Michibayashi, K., Masuda, T., Kuroki, N.,

O'kane, A., Onasch, C.M. & Farver, J.R. (2007). The role of fluids in low temperature, fault-

Okumura, S. & Nakashima, S. (2004). Water diffusivity in rhyolitic glasses as determined by in situ IR spectroscopy. *Physics and Chemistry of Minerals*, Vol. 31, pp. 183-189 Passchier, C.W. & Trouw, R.A.J. (2005). *Microtectonics (2nd Ed)*. Springer-Verlag, Heidelberg

feldspars using FTIR and 1H MAS NMR spectroscopy. *American Mineralogist*, Vol.

grained siliciclastic rocks in the middle crust−evidence from structural and

determinations of intragranular water content in quartz-bearing rocks: implications for hydrolytic weakening in the laboratory and within the earth, *Techtonophysics,* 

K.S. (2004). Insights into the structures, energetics, and vibrations of monovalent cation-(water)1−6 clusters. *Journal of Physical Chemistry A*, Vol. 108, pp. 2949-2958 Libowitzky, E. & Rossman, G.R. (1997). An IR absorption calibration for water in minerals.

Raman, NMR investigations of beryls from pegmatites of the Sudety Mts. *Journal of* 

mapping: application to the detection of hydrogen-related species in natural

Yamagishi, H., Ito, Y. & Nakamura, A. (1995). Infrared microspectroscopy analysis of water distribution in deformed and metamorphosed rocks. *Tectonophysics*, Vol.

related deformation of quartz arenite. *Journal of Structural Geology*, Vol. 29, pp. 819-

Series, Vol. 168, American Geophysical Union, Washington, D.C.

*Geophysical Research*, Vol. 89, pp. 4271-4280

*and Chemistry of Minerals*, Vol. 28, pp. 365-376

*American Mineralogist,* Vol. 82, pp. 1111-1115

*Molecular Structure*, Vol. 744, pp. 1005-1015

quartz. *Journal of Microscopy-Oxford*, Vol. 216, pp. 222-228

88, pp. 901-911

Vol. 172, pp. 255-271

245, pp. 263-276

836

201-279


Fukuda, J. & Shinoda, K. (2008). Coordination of water molecules with Na+ cations in a beryl

Fukuda, J. & Nakashima, S. (2008). Water at high temperatures in a microcrystalline silica

Fukuda, J., Peach, C.J., Spiers, C.J. & Nakashima, S. (2009a). Electrical impedance

Fukuda, J., Shinoda, K., Nakashima, S., Miyoshi, N. & Aikawa, N. (2009b). Polarized

Fukuda, J., Yokoyama, T. & Kirino, Y. (2009c). Characterization of the states and diffusivity

Fukuda, J. & Shinoda, K. (2011). Water molecules in beryl and cordierite: high-temperature

Gibbs, G.V., Breck, D.W. & Meagher, E.P. (1968). Structural refinement of hydrous and

Gleason, G.C. & DeSisto, S. (2008). A natural example of crystal-plastic deformation

Goldman, D.S. & Rossman, G.R. (1977). Channel constituents in cordierite. *American* 

Graetsch, H., Flörke, O.W. & Miehe, G. (1985). The nature of water in chalcedony and opal-C from Brazilian agate geodes. *Physics and Chemistry of Minerals*, Vol. 12, pp. 300-306 Grigss, D.T. (1967). Hydrolytic weakening of quartz and other silicates. *Geophysical Journal of* 

Hawthorne, F.C. & Černý, P. (1977). The alkali-metal positions in Cs-Li beryl. *Canadian* 

Hiraga, T., Nishikawa, O., Nagase, T. & Akizuki, M. (2001). Morphology of intergranular

microscopy. *Contributions to Mineralogy and Petrology*, Vol. 141, pp. 613-622 Hofmeister, A.M. & Rossman, G.R. (1985). A model for the irradiative colaration of smoky

Imon, R., Okudaira, T. & Fujimoto, A. (2002). Dissolution and precipitation processes in

pores and wetting angles in pelitic schists studied by transmission electron

feldspar and the inhibiting influence of water. *Physics and Chemistry of Minerals*,

deformed amphibolites: an example from the ductile shear zone of the Ryoke metamorphic belt, SW Japan. *Journal of Metamorphic Geology*, Vol. 20, pp. 297-308 Imon, R., Okudaira, T. & Kanagawa, K. (2004). Development of shape- and lattice-preferred

orientations of amphibole grains during initial cataclastic deformation and subsequent deformation by dissolution-precipitation creep in amphibolites from the Ryoke metamorphic belt, SW Japan. *Journal of Structural Geology*, Vol. 26, pp.

channels in beryl. *American Mineralogist*, Vol. 94, pp. 981-985

spectroscopy. *Mineralogical Magazine*, Vol. 73, pp. 825-835

*Minerals*, Vol. 35, pp. 347-357

*Petrological Sciences*, Vol. 104, pp. 176-181

*Chemistry of Minerals*, Vol. 38, pp. 469-481

*Mineralogist*, Vol. 62, pp. 1144-1157

*Mineralogist,* Vol. 15, pp. 414-421

*the Royal Astronomical Society,* Vol. 14, pp. 19-32

Vol. 1, pp. 275-285

Vol. 12, pp. 324-332

793-805

112-115

channel as determined by polarized IR spectroscopy. *Physics and Chemistry of* 

(chalcedony) by in-situ infrared spectroscopy: physicochemical states and dehydration behavior. *Journal of Mineralogical and Petrological Sciences*, Vol. 103, pp.

measurement of hydrous microcrystalline quartz. *Journal of Mineralogical and* 

infrared spectroscopic study of diffusion of water molecules along structure

of intergranular water in a chalcedonic quartz by high temperature in-situ infrared

vibrational behavior, dehydration, and coordination to cations. *Physics and* 

anhydrous synthetic beryl, Al2(Be3Si6)O18 and emerald, Al1.9Cr0.1(Be3Si6)O18. *Lithos*,

enhancing the incorporation of water into quartz. *Tectonophysics*, Vol. 446, pp. 16-30


**5** 

*1France 2Croatia 3Germany* 

**Attenuated Total Reflection –** 

**Infrared Spectroscopy Applied** 

**Electrolyte Solution Interfaces:** 

*1Chimie ParisTech - LECIME - CNRS UMR 7575, Paris* 

**to the Study of Mineral – Aqueous** 

*2Laboratory of Physical Chemistry, Department of Chemistry,* 

*Faculty of Science, University of Zagreb, Zagreb 3Karlsruhe Institute of Technology (KIT), Institute for* 

*Nuclear Waste Disposal (INE), Karlsruhe* 

**A General Overview and a Case Study** 

Grégory Lefèvre1, Tajana Preočanin2 and Johannes Lützenkirchen3

The present chapter gives an overview of the application of Attenuated total reflection – Infrared spectroscopy (ATR-IR) to the environmentally important mineral – aqueous electrolyte interface. At these interfaces the important adsorption processes occur that limit the availability of potentially toxic solutes. These retention processes may retard for example the migration of solutes in aquifer systems or even immobilize them on the aquifer material, which is usually a natural mineral. Selected solutes may also via a preliminary adsorption process, which weakens bonds, enhance both dissolution kinetics and the equilibrium

In the context of retardation (oxy)(hydr)oxide minerals are of major importance. At the surface of these minerals surface functional groups exist that are able to bind metal ions and organic ligands as well as they may promote the formation of so-called ternary surface complexes involving both metal ions and some ligand. To be able to quantify these retention phenomena in porous media (such as aquifers or soils) a physical model of solvent movement is coupled to a (chemical) adsorption model (usually some variant of the surface complexation approach). The intent in the chemical part of the model is to invoke as much understanding of the adsorption process as possible. Thus it turns out to be important whether an adsorption process results in monodentate or multidentate surface complexes. This can have profound consequences in the use of a surface complexation model under different conditions (Kulik et al., 2010; Kallay et al., 2011). Evaluating a surface complexation

**1. Introduction** 

solubility of a given mineral.


## **Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study**

Grégory Lefèvre1, Tajana Preočanin2 and Johannes Lützenkirchen3 *1Chimie ParisTech - LECIME - CNRS UMR 7575, Paris 2Laboratory of Physical Chemistry, Department of Chemistry, Faculty of Science, University of Zagreb, Zagreb 3Karlsruhe Institute of Technology (KIT), Institute for Nuclear Waste Disposal (INE), Karlsruhe 1France 2Croatia 3Germany* 

#### **1. Introduction**

96 Infrared Spectroscopy – Materials Science, Engineering and Technology

Paterson, M.S. (1982). The determination of hydroxyl by infrared absorption in quartz, silicate glasses and similar materials. *Bulletin de Minéralogie,* Vol. 105, pp. 20-29 Post, A. & Tullis, J. (1998). The rate of water penetration in experimentally deformed

Schwarzer, D. (2005). Energy relaxation versus spectral diffusion of the OH-stretching

Simpson, C. & Wintsch, R.P. (1989). Evidence for deformation-induced K-feldspar replacement by myrmekite. *Journal of Metamorphic Geology*, Vol. 7, pp. 261-275 Takagi, H., Mizutani, T. & Hirooka, K. (1988). Deformation of quartz in an inner shear zone

Thompson, A.B. & Rubie, D.C. (1985). *Metamorphic Reactions, Kinetics, Textures and* 

Tsurumi, J., Hosonuma, H. & Kanagawa, K. (2003). Strain localization due to a positive

Wintsch, R.P. & Yi, R. (2002). Dissolution and replacement creep: a significant deformation

Wood, D.L. & Nassau, K. (1967). Infrared spectra of foreign molecules in beryl. *Journal of* 

Yamagishi, H., Nakashima, S. & Ito, Y. (1997). High temperature infrared spectra of hydrous microcrystalline quartz, *Physics and Chemistry of Minerals*, Vol. 24, pp. 66-74

137

abstract).

*Physics*, Vol. 123, Article No. 161105

*Deformation,* Springer, New York

*Chemical Physics,* Vol. 47, pp. 2220-2228

*Geology*, Vol. 25, pp. 557-574

quartzite: implications for hydrolytic weakening. *Tectonophysics*, Vol. 295, pp. 117-

vibration of HOD in liquid-to-supercritical deuterated water. *Journal of Chemical* 

of the Ryoke belt – an example in the Kishiwada area, Osaka Prefecture. *Journal of the Geological Society of Japan,* Vol. 94, pp. 869-886 (in Japanese with English

feedback of deformation and myrmekite-forming reaction in granite and aplite mylonites along the Hatagawa Shear Zone of NE Japan. *Journal of Structural* 

mechanism in mid-crustal rocks. *Journal of Structural Geology*, Vol. 24, pp. 1179-1193

The present chapter gives an overview of the application of Attenuated total reflection – Infrared spectroscopy (ATR-IR) to the environmentally important mineral – aqueous electrolyte interface. At these interfaces the important adsorption processes occur that limit the availability of potentially toxic solutes. These retention processes may retard for example the migration of solutes in aquifer systems or even immobilize them on the aquifer material, which is usually a natural mineral. Selected solutes may also via a preliminary adsorption process, which weakens bonds, enhance both dissolution kinetics and the equilibrium solubility of a given mineral.

In the context of retardation (oxy)(hydr)oxide minerals are of major importance. At the surface of these minerals surface functional groups exist that are able to bind metal ions and organic ligands as well as they may promote the formation of so-called ternary surface complexes involving both metal ions and some ligand. To be able to quantify these retention phenomena in porous media (such as aquifers or soils) a physical model of solvent movement is coupled to a (chemical) adsorption model (usually some variant of the surface complexation approach). The intent in the chemical part of the model is to invoke as much understanding of the adsorption process as possible. Thus it turns out to be important whether an adsorption process results in monodentate or multidentate surface complexes. This can have profound consequences in the use of a surface complexation model under different conditions (Kulik et al., 2010; Kallay et al., 2011). Evaluating a surface complexation

Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of

E0

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 99

Z

n1 n2

<sup>E</sup> sample

crystal

**Reflected radiation**

1/2 <sup>1</sup> 2 2 d sin n p 21 <sup>2</sup> (2)

2 n (3)

d F n 1 F n (4) p v par v water =× +− × ( )

θ

Fig. 1. Schematic diagram of the attenuated total reflection of the infrared beam in a

falls to half its value at the interface (Z = 0.69 / γ). Another definition of the depth of penetration (dp) is given by Z = 1 / γ, i.e. a decay of the electric field of 63 %. Moreover, this value is lower than the actual depth sampled (dS), which is about three times dp (decay of the electric field of 95%) (Mirabella, 1993; Tickanen et al., 1991). Equation (1) can be used to obtain the value of dp in a homogeneous solution, but the determination of the penetration across oxy-hydroxide films is more complex. The depth of penetration, dp, is expressed as

> ( ) <sup>λ</sup> <sup>−</sup> = θ− π

= θ <sup>−</sup> πν

the particle material and the aqueous solution (Hug and Sulzberger, 1994):

3×dp), indicating that the deposited layer should be thinner than this value.

p 21 1 <sup>10000</sup> d sin n

In studies on the adsorption of ions onto layers of particles deposited on ATR crystals, it is important that the whole layer be probed. Otherwise sorption which takes place in the top of the layer (i.e. further away from the crystal) does not significantly contribute to the observed signal. To take into account the presence of a layer of particles (pores filled with solution) formula (1) can be used with a volume-weighted average of the refractive index of

where Fv is the volume fraction of solid and npar the refractive index of the pure solid. A volume fraction between 0.30 and 0.40 was estimated for TiO2 (npar = 2.6), leading to a maximum dp of 2.6 µm at 1100 cm-1. Thus, the actual depth sampled would be *ca.* 7 µm (dS =

( )<sup>−</sup>

1/2 2 2

**Incident radiation (wavelength** λ**)**

monoreflection ATR accessory.

or, with ν, the wavenumbers (cm-1):

(Coates, 1993)):

model based on macroscopic adsorption data alone usually is not unambiguous. Consequently, it is required to study the adsorption process at the molecular level. Various spectroscopic approaches have been used to resolve the adsorption mechanism, one being ATR-IR.

We give an introduction to the approach and an overview of its possible applicability (and in this context its use in contributing to the understanding of the acid-base chemistry of (oxy)(hydr)oxide mineral surfaces, the adsorption of anions and cations like the uranyl-ion, and the formation of ternary surface complexes can be mentioned in general). Our contribution focuses on a review on the interaction of small organic molecules with oxidic surfaces and we highlight previous studies and point to some controversary issues in selected studies that continue to exist despite extensive research. Obviously such studies relate to other vibrational spectroscopies like Raman or sum frequency generation vibrational spectroscopies.

Finally we discuss results from an experimental study on the mineral gibbsite (Al(OH)3) in the presence of 5-sulfosalicylic acid (5-SSA). We show how ideally such a study should be designed, starting from the study of the gibbsite-electrolyte solution system (i.e. in absence of 5-SSA) and that of 5-SSA in aqueous solution (i.e. in the absence of gibbsite). Furthermore, we show that it is necessary to study in aqueous solution the interaction of 5-SSA with dissolved aluminium, since the pH – dependent solubility of gibbsite will ultimately cause the appearance of aluminium ions in solution. The system involving gibbsite and 5-SSA is discussed in more detail. We relate the data to calculations of the species distribution for the solution systems, which indicate the dominant aqueous species thus facilitating the assignment of bands.

#### **2. Review of use of ATR in studies about adsorption of selected small organic molecules**

#### **2.1 Principles of ATR**

The Attenuated Total Reflection effect is based on the existence of an evanescent wave in a medium of lower index of refraction in contact with an optically denser medium in which the infrared beam is sent. This evanescent field decays exponentially in the less dense medium according to equation (1).

$$\mathbf{E} = \mathbf{E}\_0 \exp\left[ -\frac{2\pi}{\lambda\_1} \left( \sin^2 \theta - \mathbf{n}\_{21}^2 \right)^{1/2} Z \right] \tag{1}$$

where λ1 = λ / n1 is the wavelength of the radiation in the denser medium, λ the wavelength in free space, θ the angle of incidence with respect to the normal. The parameter n21 is defined as the ratio of the refractive indices, i.e. n21 = n2 / n1, where n1 and n2 are respectively, the refractive indices of the optically denser and less dense media, and Z is the distance from the surface (Mirabella, 1993) (see Fig. 1).

From the ATR element, the infrared beam probes only the first few micrometers of the sample medium. From equation (1), different parameters can be defined to characterize the depth of penetration. A first definition was the depth at which the electric field amplitude

Fig. 1. Schematic diagram of the attenuated total reflection of the infrared beam in a monoreflection ATR accessory.

falls to half its value at the interface (Z = 0.69 / γ). Another definition of the depth of penetration (dp) is given by Z = 1 / γ, i.e. a decay of the electric field of 63 %. Moreover, this value is lower than the actual depth sampled (dS), which is about three times dp (decay of the electric field of 95%) (Mirabella, 1993; Tickanen et al., 1991). Equation (1) can be used to obtain the value of dp in a homogeneous solution, but the determination of the penetration across oxy-hydroxide films is more complex. The depth of penetration, dp, is expressed as (Coates, 1993)):

$$\mathbf{d}\_{\mathbf{p}} = \frac{\lambda\_1}{2\pi} \left(\sin^2\theta - \mathbf{n}\_{21}^2\right)^{-1/2} \tag{2}$$

or, with ν, the wavenumbers (cm-1):

98 Infrared Spectroscopy – Materials Science, Engineering and Technology

model based on macroscopic adsorption data alone usually is not unambiguous. Consequently, it is required to study the adsorption process at the molecular level. Various spectroscopic approaches have been used to resolve the adsorption mechanism, one being

We give an introduction to the approach and an overview of its possible applicability (and in this context its use in contributing to the understanding of the acid-base chemistry of (oxy)(hydr)oxide mineral surfaces, the adsorption of anions and cations like the uranyl-ion, and the formation of ternary surface complexes can be mentioned in general). Our contribution focuses on a review on the interaction of small organic molecules with oxidic surfaces and we highlight previous studies and point to some controversary issues in selected studies that continue to exist despite extensive research. Obviously such studies relate to other vibrational spectroscopies like Raman or sum frequency generation

Finally we discuss results from an experimental study on the mineral gibbsite (Al(OH)3) in the presence of 5-sulfosalicylic acid (5-SSA). We show how ideally such a study should be designed, starting from the study of the gibbsite-electrolyte solution system (i.e. in absence of 5-SSA) and that of 5-SSA in aqueous solution (i.e. in the absence of gibbsite). Furthermore, we show that it is necessary to study in aqueous solution the interaction of 5-SSA with dissolved aluminium, since the pH – dependent solubility of gibbsite will ultimately cause the appearance of aluminium ions in solution. The system involving gibbsite and 5-SSA is discussed in more detail. We relate the data to calculations of the species distribution for the solution systems, which indicate the dominant aqueous species thus facilitating the

**2. Review of use of ATR in studies about adsorption of selected small** 

The Attenuated Total Reflection effect is based on the existence of an evanescent wave in a medium of lower index of refraction in contact with an optically denser medium in which the infrared beam is sent. This evanescent field decays exponentially in the less dense

( ) <sup>π</sup> = − θ− <sup>λ</sup>

where λ1 = λ / n1 is the wavelength of the radiation in the denser medium, λ the wavelength in free space, θ the angle of incidence with respect to the normal. The parameter n21 is defined as the ratio of the refractive indices, i.e. n21 = n2 / n1, where n1 and n2 are respectively, the refractive indices of the optically denser and less dense media, and Z is the

From the ATR element, the infrared beam probes only the first few micrometers of the sample medium. From equation (1), different parameters can be defined to characterize the depth of penetration. A first definition was the depth at which the electric field amplitude

0 21 1

1/2 2 2

<sup>2</sup> E E exp sin n Z (1)

ATR-IR.

vibrational spectroscopies.

assignment of bands.

**organic molecules 2.1 Principles of ATR** 

medium according to equation (1).

distance from the surface (Mirabella, 1993) (see Fig. 1).

$$\text{pd}\_{\text{p}} = \frac{10000}{2\pi \text{v} \text{n}\_1} \left(\text{sin}^2 \Theta - \text{n}\_{21}^2\right)^{-1/2} \tag{3}$$

In studies on the adsorption of ions onto layers of particles deposited on ATR crystals, it is important that the whole layer be probed. Otherwise sorption which takes place in the top of the layer (i.e. further away from the crystal) does not significantly contribute to the observed signal. To take into account the presence of a layer of particles (pores filled with solution) formula (1) can be used with a volume-weighted average of the refractive index of the particle material and the aqueous solution (Hug and Sulzberger, 1994):

$$\mathbf{d}\_{\rm p} = \mathbf{F}\_{\rm v} \times \mathbf{n}\_{\rm par} + (1 - \mathbf{F}\_{\rm v}) \times \mathbf{n}\_{\rm water} \tag{4}$$

where Fv is the volume fraction of solid and npar the refractive index of the pure solid. A volume fraction between 0.30 and 0.40 was estimated for TiO2 (npar = 2.6), leading to a maximum dp of 2.6 µm at 1100 cm-1. Thus, the actual depth sampled would be *ca.* 7 µm (dS = 3×dp), indicating that the deposited layer should be thinner than this value.

Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of

Variety of solids to best

Quantitative evaluation of

drawback.

up.

adsorbing species.

**2.2.2 Limitations in the wavenumber range** 

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 101

Ease of preparation ++ + -

studied ++ + - Optical quality (sensitivity) - + ++

spectra - + ++ Representativity/suspension ++ + - Flow cell - + ++

the confinement of the solutions are ignored. The last advantage of using a film is the possibility to perform experiments with a flow cell. This set-up allows recording spectra, while varying the composition of the solution, e.g. by modifying pH, or the concentration of

Once the procedure to prepare the interface has been chosen, the wavenumber range covered by the measurement is another important experimental aspect. Indeed, a number of interferences may occur between bands of adsorbed species and the experimental set-

The main limitation may arise from the **ATR element** itself. Each material has a transmission threshold, which may be located at a high wavenumber, such as silicium. Other materials with a low transmission threshold may be too reactive towards solutions. Thus, ZnSe can be attacked by acid or zinc-complexing species. A usual choice made by ATR-elements-dealers is an element made in ZnSe, but covered by a thin layer of diamond to increase its chemical resistance. This possibility exists only for small ATR crystals, allowing only few reflections of the infrared beam. To increase sensitivity, large ATR crystals are used. For example 40 mm × 10 mm crystals with a thickness of around 1 mm

Besides the above limitations due to the ATR element, two gases present in the ambient **atmosphere** lead to absorption bands in IR spectra: carbon dioxide, and water. The main bands (Fig. 2) consist in a doublet at 2361 and 2339 cm-1 (CO2), and numerous narrow peaks in the range 2000 – 1300 (H2O bending) and 4000 – 3400 (H2O stretching). Generally, the band of CO2 does not interfere with bands of adsorbates, but H2O bending can interfere with adsorbed organic molecules. Several methods exist to solve this problem. In fact, the presence of CO2 and water in the atmosphere of the spectrometer is not the actual problem since it is taken into account in the background spectrum. It is rather the evolution of their concentrations (or partial pressures) during the subsequent spectra collection that leads to the presence of bands, which varies with time. The less concentrated these gases are, the

allow dozens of reflections. Such crystals usually consist of a pure material.

Table 1. Summary of characteristics of the three methods of preparation of the solid/solution interface probed by ATR-IR: (++) strong advantage, (+) advantage, (-)

Method Paste Dried layer Film growth or

crystal only

## **2.2 Experimental**

Using an accessory allowing to record infrared spectra in ATR mode is the first requirement to get *in situ* signals of the solid/solution interface. However, the way to prepare this interface and even the choice of the accessory is not straightforward.

## **2.2.1 Protocols to produce a suitable solid-liquid interface**

The first step in ATR-related studies involves the formation of a suitable solid-liquid interface. To obtain a metal oxide / solution interface which can be probed by ATR, several methods have been described in literature.

The first one, described in the pioneering work by Tejedor-Tejedor and collaborators (Tejedor-Tejedor and Anderson, 1986; Tejedor-Tejedor and Anderson, 1990; Tickanen et al., 1991) consisted in a cylindrical internal reflection cell (a rod-shaped crystal of ZnSe) dipped in a suspension of 100 g/L goethite. This method is now less frequently used, to the advantage of horizontal ATR crystals. Using such instrumentation, Hug and Sulzberger (1994) have developed a method which has become standard. The approach consists in coating the ATR crystal by colloidal particles to form a film. As a typical protocol, a mixture of solid and ethanol is spread over the ATR crystal, then dried using a nitrogen flux. After drying, the layer is rinsed with water or with an electrolyte solution. More details are given in articles by Hug (1997) or Peak et al. (1999).

In another method, the equilibrium of the system solid/solution is reached by a classical batch experiment, using diluted suspensions of the solid. Then the suspensions are centrifuged to obtain a higher mass/volume ratio, for example 100-1000 g/L, or even a paste. The sample is then spread on the ATR crystal using a spatula (Villalobos and Leckie, 2001).

A final possibility is to use the surface of the crystal as the sample itself. Either the surface of the crystal is used as received, as ZnSe on which sodium dodecyl sulfate (Gao and Chorover, 2010) or Ge on which heptyl xanthate (Larsson et al., 2004) formed a monolayer, or the surface was chemically modified and is different from the bulk. Thus, Asay and Kim (2005) studied the adsorption of water molecules on the native layer of silica present on a silicium ATR crystal, or Wang et al. (2006) studied the adsorption of hexane and ethylbenzene from the vapor phase on a layer of zeolite grown directly on the surface of a silicium ATR crystal. Frederiksson and Holmgren (2008) have formed a PbS film on a ZnS ATR crystal by a chemical bath deposition process in order to study the adsorption of heptyl xanthate. In these latter studies, the system is very close to a film obtained by drying of a suspension, but the optical properties are expected to be better. Couzis and Gulari (1993) have deposited 600 Å of alumina by sputtering on a ZnSe crystal.

The advantages and drawbacks of the three methods to prepare the solid/solution interface discussed above are listed in table 1. As of today, the most common method is to prepare a dry layer, even though it is simpler to use a paste. However, using a paste has a major drawback since the contact between particles and the ATR crystal is not optimal, the sensitivity is low and depends on the suspension structure (which in general is pHdependent). On the other hand, using the results obtained with a dry layer to interpret macroscopic data obtained in well-dispersed suspensions can be tricky, since effects due to


Table 1. Summary of characteristics of the three methods of preparation of the solid/solution interface probed by ATR-IR: (++) strong advantage, (+) advantage, (-) drawback.

Representativity/suspension ++ + - Flow cell - + ++

the confinement of the solutions are ignored. The last advantage of using a film is the possibility to perform experiments with a flow cell. This set-up allows recording spectra, while varying the composition of the solution, e.g. by modifying pH, or the concentration of adsorbing species.

#### **2.2.2 Limitations in the wavenumber range**

100 Infrared Spectroscopy – Materials Science, Engineering and Technology

Using an accessory allowing to record infrared spectra in ATR mode is the first requirement to get *in situ* signals of the solid/solution interface. However, the way to prepare this

The first step in ATR-related studies involves the formation of a suitable solid-liquid interface. To obtain a metal oxide / solution interface which can be probed by ATR, several

The first one, described in the pioneering work by Tejedor-Tejedor and collaborators (Tejedor-Tejedor and Anderson, 1986; Tejedor-Tejedor and Anderson, 1990; Tickanen et al., 1991) consisted in a cylindrical internal reflection cell (a rod-shaped crystal of ZnSe) dipped in a suspension of 100 g/L goethite. This method is now less frequently used, to the advantage of horizontal ATR crystals. Using such instrumentation, Hug and Sulzberger (1994) have developed a method which has become standard. The approach consists in coating the ATR crystal by colloidal particles to form a film. As a typical protocol, a mixture of solid and ethanol is spread over the ATR crystal, then dried using a nitrogen flux. After drying, the layer is rinsed with water or with an electrolyte solution. More details are given

In another method, the equilibrium of the system solid/solution is reached by a classical batch experiment, using diluted suspensions of the solid. Then the suspensions are centrifuged to obtain a higher mass/volume ratio, for example 100-1000 g/L, or even a paste. The sample is then spread on the ATR crystal using a spatula (Villalobos and Leckie,

A final possibility is to use the surface of the crystal as the sample itself. Either the surface of the crystal is used as received, as ZnSe on which sodium dodecyl sulfate (Gao and Chorover, 2010) or Ge on which heptyl xanthate (Larsson et al., 2004) formed a monolayer, or the surface was chemically modified and is different from the bulk. Thus, Asay and Kim (2005) studied the adsorption of water molecules on the native layer of silica present on a silicium ATR crystal, or Wang et al. (2006) studied the adsorption of hexane and ethylbenzene from the vapor phase on a layer of zeolite grown directly on the surface of a silicium ATR crystal. Frederiksson and Holmgren (2008) have formed a PbS film on a ZnS ATR crystal by a chemical bath deposition process in order to study the adsorption of heptyl xanthate. In these latter studies, the system is very close to a film obtained by drying of a suspension, but the optical properties are expected to be better. Couzis and Gulari (1993)

The advantages and drawbacks of the three methods to prepare the solid/solution interface discussed above are listed in table 1. As of today, the most common method is to prepare a dry layer, even though it is simpler to use a paste. However, using a paste has a major drawback since the contact between particles and the ATR crystal is not optimal, the sensitivity is low and depends on the suspension structure (which in general is pHdependent). On the other hand, using the results obtained with a dry layer to interpret macroscopic data obtained in well-dispersed suspensions can be tricky, since effects due to

have deposited 600 Å of alumina by sputtering on a ZnSe crystal.

interface and even the choice of the accessory is not straightforward.

**2.2.1 Protocols to produce a suitable solid-liquid interface** 

methods have been described in literature.

in articles by Hug (1997) or Peak et al. (1999).

**2.2 Experimental** 

2001).

Once the procedure to prepare the interface has been chosen, the wavenumber range covered by the measurement is another important experimental aspect. Indeed, a number of interferences may occur between bands of adsorbed species and the experimental setup.

The main limitation may arise from the **ATR element** itself. Each material has a transmission threshold, which may be located at a high wavenumber, such as silicium. Other materials with a low transmission threshold may be too reactive towards solutions. Thus, ZnSe can be attacked by acid or zinc-complexing species. A usual choice made by ATR-elements-dealers is an element made in ZnSe, but covered by a thin layer of diamond to increase its chemical resistance. This possibility exists only for small ATR crystals, allowing only few reflections of the infrared beam. To increase sensitivity, large ATR crystals are used. For example 40 mm × 10 mm crystals with a thickness of around 1 mm allow dozens of reflections. Such crystals usually consist of a pure material.

Besides the above limitations due to the ATR element, two gases present in the ambient **atmosphere** lead to absorption bands in IR spectra: carbon dioxide, and water. The main bands (Fig. 2) consist in a doublet at 2361 and 2339 cm-1 (CO2), and numerous narrow peaks in the range 2000 – 1300 (H2O bending) and 4000 – 3400 (H2O stretching). Generally, the band of CO2 does not interfere with bands of adsorbates, but H2O bending can interfere with adsorbed organic molecules. Several methods exist to solve this problem. In fact, the presence of CO2 and water in the atmosphere of the spectrometer is not the actual problem since it is taken into account in the background spectrum. It is rather the evolution of their concentrations (or partial pressures) during the subsequent spectra collection that leads to the presence of bands, which varies with time. The less concentrated these gases are, the

Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of

Absorbance

(Venyaminov and Prendergast, 1997).

from \*Lide (1998), \*\* Dean (1999)

**2.3.1 Monoacids: Formic, acetic, benzoic, lauric** 

discussed in detail in the remainder of the section.

Acid pKa R ΔCOO (cm-1)

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 103

4000 3600 3200 2800 1200 1000 800 Wavenumbers (cm-1)

Fig. 3. Spectra of solids as dried layer on an ATR element: silica (), gibbsite (), goethite ()

below 890 cm-1 (Lefèvre et al., 2006). To be able to record spectra at lower wavenumbers, heavy water (D2O) can be used because the absorption bands are shifted by a factor of ca. 1.4 to lower wavenumbers. Thus, a good signal can be obtained for bands located between 850 and 950 cm-1 (Lefèvre et al., 2008) using the same 25-reflection crystal. It can be useful to avoid interferences with bands around 1650 cm-1 since D2O bending is located at 1209 cm-1

**2.3 Review of adsorption of carboxylic acids onto metal (hydr)oxides by ATR-IR** 

Formic 3.75 \* –H 230 192 (TiO2) Acetic 4.76 \* –CH3 137 90 (TiO2) Benzoic 4.19 \* –C6H5 154 109 (TiO2)

Lauric 4.90 \*\* –CH2–(CH2)9–CH3 136 185 (alumina)

Table 2. Characteristics of carboxylic acid (R-C(O)OH). ΔCOO=νas(COO)- νs(COO)

A number of monoacids are discussed in the context of this review. Table 2 gives some information on the monoacids both in solution and at the interface. The systems are

in solution

ΔCOO (cm-1) adsorbed

117 (Ta2O5) 122 (goethite) 141 (ZrO2)

Fig. 2. Spectrum of the atmosphere showing the contributions as discussed in the text.

lower are the bands, since the signal comes from the fluctuation of the partial pressures. Thus, some spectrometers are evacuated to enhance sensitivity and stability. Other spectrometers are purged with inert gas or with compressed dried air. Another possible solution to the problem consists in the use of spectrometers which are sealed and equipped with desiccant powder. In all cases, if bands of atmospheric compounds remain, they can be tentatively removed by subtracting the atmosphere spectra.

Since the studies generally consist in probing the species adsorbed on a **solid deposited** on the ATR crystal, it is important to take into account bands from the solid itself. For metal oxides, the absorption bands are generally located at low wavenumbers, which does not cause interferences with adsorbed species. Exceptions exist with light metals as SiO2 (around 1060 cm-1). For metal hydroxides, stretching of M-OH can lead to the presence of bands above 800 cm-1 as is the case with goethite (900 and 800 cm-1) or gibbsite (around 1000 cm-1).

Ideally, if the layer formed by particles is stable, the signal coming from the solid can be subtracted from the final spectra, and the presence of these bands does not hamper the detection and interpretation of bands from adsorbed species. However, in practice subtraction is often difficult due to the evolution of the signal of the solid with time or solution composition. Phenomena such as re-entrainment of particles by flowing solution, or swelling/shrinkage due to the change in surface potential can explain this problem.

Water is the most common **solvent** in environmental studies and its absorption bands can be a problem too. Stretching of H2O occurs around 3000-3600 cm-1 and interferes with stretching of surface hydroxyl groups. Bending takes place at 1643 cm-1 (Venyaminov and Prendergast, 1997), close to the stretching of C=O groups (see below). This can complicate the accurate measurement of νC=O maxima. Finally, water absorption is very strong below ca. 900 cm-1, and this can prevent the measurement of any bands in the lowest wavenumber range. In fact the actual threshold appears to depend on the number of reflections in the ATR system. For a monoreflection accessory, a measurement can be made down to 650 cm-1 without large absorption of H2O, while for a 25-reflection crystal, the signal becomes noisy

4000 3500 3000 2500 2000 1500 1000

Wavenumbers (cm-1)

lower are the bands, since the signal comes from the fluctuation of the partial pressures. Thus, some spectrometers are evacuated to enhance sensitivity and stability. Other spectrometers are purged with inert gas or with compressed dried air. Another possible solution to the problem consists in the use of spectrometers which are sealed and equipped with desiccant powder. In all cases, if bands of atmospheric compounds remain, they can be

Since the studies generally consist in probing the species adsorbed on a **solid deposited** on the ATR crystal, it is important to take into account bands from the solid itself. For metal oxides, the absorption bands are generally located at low wavenumbers, which does not cause interferences with adsorbed species. Exceptions exist with light metals as SiO2 (around 1060 cm-1). For metal hydroxides, stretching of M-OH can lead to the presence of bands above 800 cm-1 as is the case with goethite (900 and 800 cm-1) or gibbsite (around 1000 cm-1). Ideally, if the layer formed by particles is stable, the signal coming from the solid can be subtracted from the final spectra, and the presence of these bands does not hamper the detection and interpretation of bands from adsorbed species. However, in practice subtraction is often difficult due to the evolution of the signal of the solid with time or solution composition. Phenomena such as re-entrainment of particles by flowing solution, or

swelling/shrinkage due to the change in surface potential can explain this problem.

Water is the most common **solvent** in environmental studies and its absorption bands can be a problem too. Stretching of H2O occurs around 3000-3600 cm-1 and interferes with stretching of surface hydroxyl groups. Bending takes place at 1643 cm-1 (Venyaminov and Prendergast, 1997), close to the stretching of C=O groups (see below). This can complicate the accurate measurement of νC=O maxima. Finally, water absorption is very strong below ca. 900 cm-1, and this can prevent the measurement of any bands in the lowest wavenumber range. In fact the actual threshold appears to depend on the number of reflections in the ATR system. For a monoreflection accessory, a measurement can be made down to 650 cm-1 without large absorption of H2O, while for a 25-reflection crystal, the signal becomes noisy

Fig. 2. Spectrum of the atmosphere showing the contributions as discussed in the text.

tentatively removed by subtracting the atmosphere spectra.

Absorbance

Fig. 3. Spectra of solids as dried layer on an ATR element: silica (), gibbsite (), goethite ()

below 890 cm-1 (Lefèvre et al., 2006). To be able to record spectra at lower wavenumbers, heavy water (D2O) can be used because the absorption bands are shifted by a factor of ca. 1.4 to lower wavenumbers. Thus, a good signal can be obtained for bands located between 850 and 950 cm-1 (Lefèvre et al., 2008) using the same 25-reflection crystal. It can be useful to avoid interferences with bands around 1650 cm-1 since D2O bending is located at 1209 cm-1 (Venyaminov and Prendergast, 1997).

## **2.3 Review of adsorption of carboxylic acids onto metal (hydr)oxides by ATR-IR**

### **2.3.1 Monoacids: Formic, acetic, benzoic, lauric**

A number of monoacids are discussed in the context of this review. Table 2 gives some information on the monoacids both in solution and at the interface. The systems are discussed in detail in the remainder of the section.


from \*Lide (1998), \*\* Dean (1999)

Table 2. Characteristics of carboxylic acid (R-C(O)OH). ΔCOO=νas(COO)- νs(COO)

Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of

due to the low specific surface area of the minerals used in the study.

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 105

conclude that chemisorption is below the detection limit of the spectroscopy. This might be

Adsorption of **benzoic** acid on minerals was studied by several authors on quartz, albite, illite, kaolinite and montmorillonite (Kubicki et al., 1999), goethite (Tejedor-Tejedor et al., 1990), TiO2 (Tunesi and Anderson, 1992; Dobson and McQuillan, 1999), as well as Al2O3, ZrO2 and Ta2O5 (Dobson and McQuillan, 1999). Aqueous benzoate is characterized by bands at 1542 cm-1 (νas(COO)), 1388 cm-1 (νs(COO)) and 1593 cm-1 (νC=C). At pH < pKa, spectra are characterized by bands at 1705 cm-1 (νC=O), 1319 cm-1 (νCOH), 1279 cm-1 (δCOH), and bands associated with C=C and C-H vibrations (1603, 1494, 1452, 1178, 1073 and 1026 cm-1). Benzoic acid adsorbed on quartz displays bands of the aqueous species with two new peaks (at 1604 and 1569 cm-1). The lower frequency was found by calculation to correspond to a monodentate complex, and the higher one to an outer-sphere complex. On albite at pH 3, no peaks above 1700 cm-1 were observed, indicating that the C=O group is absent from the surface complex even in the pH range where the acid species predominates over the benzoate anion. This result is a direct evidence of the formation of a bidentate complex, stable over a wide range of pH. On goethite at pD 3.9 (Tejedor-Tejedor et al., 1990) and on TiO2 at pH 3.6 (Tunesi and Anderson, 1992), the νC=O mode is also absent. Another interesting point is that the asymmetric / symmetric carboxylate group stretching ratio decreases when benzoate interacts with Fe(III), which can be explained by the increase of coplanarity between the benzene ring and the νas(COO). These observations are consistent

with the formation of a bidentate complex. c.f. Fig. 5 (Tejedor-Tejedor et al., 1990).

(A)

Fig. 5. Proposed surface complexes of benzoate on (A) TiO2 and (B) goethite.

1992).

Ti

O

O

On TiO2, the difference between νas(COO) et νs(COO) for the adsorbed species is lower by 45 cm-1 compared to the corresponding difference for the solute species. It is believed that a lower value is indicative of a bidentate complex, and that such a large value indicates a chelate structure with a single centre (Fig. 5) (Tunesi and Anderson, 1992). On goethite, the difference was lower by 32 cm-1, consistent with a bridging complex (Tunesi and Anderson,

O

(B)

Fe

Fe

The bands pertaining to νas(COO) and νs(COO) modes of **laurate** in solution are located at 1547 and 1411 cm-1, respectively. Between 2850 and 3000 cm-1, several bands are reported corresponding to hydrocarbon stretching. Laurate anions were adsorbed onto alumina, which had been deposited on the ATR element by a sputtering technique the thickness of the film being 600 Å (Couzis and Gulari, 1993). The recorded spectra depended on contact time and pH. At pH 8, up to 20 minutes after initiation of the solid-liquid contact, the observed peaks mainly corresponded to the solute species and the authors inferred the presence of an outer-sphere surface complex, since the surface is positively charged at this

O

Several surface complexes can be formed with monoacids, such as monodentate, or bidentates (Fig. 4). Monodentate surface complexes can be distinguished from bidentates based on the occurrence or not of the free C=O group band, with a stretching frequency at about 1700 cm-1.

Fig. 4. Surface complexes between a monoacid and a metal oxide: (a) monodentate, (b) mononuclear bidentate and (c) binuclear bidentate

The ATR-FTIR spectrum of 1M of **formate** ion is characterized by bands located at 1350, 1383 and 1580 cm-1, assigned to νs(COO), δ(HCO) and νas(COO), respectively (Rotzinger et al., 2004). Spectra of formate adsorbed on TiO2 at pH 5.0, up to 30 mM display the presence of bands of formate ions and a new peak at 1540 cm-1, assigned to νas(COO) of species interacting with the surface. Spectra in D2O confirmed this assignment since only a small shift (8 cm-1) of this band was observed, which precludes the vibration of a protonated/deuterated species. A decrease of pH from 9 to 3 leads to the decrease of the peak area. A series of experiments where the adsorption of formic acid as a gas has been studied has shown the presence of bands of formic acid, formate, and a peak at ca. 1540 cm-1. In support of this, molecular calculations have been performed for the three hypothetical surface complexes (Fig. 4), leading to calculated frequencies. Calculations on the stability of the surface complexes were found to support the binuclear bidentate coordination.

Sorption of **acetate** ions has been studied by ATR on rutile (Rotzinger et al., 2004) and several other minerals (Kubicki et al., 1999). In solution, the acetate ion is characterized by bands at 1348-1349, 1415-1422 and 1552-1555 cm-1 (Rotzinger et al., 2004; Kubicki et al., 1999) assigned to δ(CH3), νs(COO), and νas(COO), respectively. Acetic acid is characterized by bands at 1279-1283, 1370-1371, 1392-1397(δCH3), 1642-1650 and 1711-1717 (νC=O) (Rotzinger et al., 2004; Kubicki et al., 1999). Spectra of adsorbed species have been recorded at pH 5.0 (ca. 1:1 mixtures of the acetate ion and acetic acid in solution since pH is close to pKa), and at total acetate concentrations up to 25 mM on TiO2 (Rotzinger et al., 2004), and at pH 3 and 6 in the presence of 2 M acetate on quartz, albite, illite, kaolinite and montmorillonite (Kubicki et al., 1999). On TiO2, bands of acetate are present with a new band at 1512 cm-1 assigned to νas(COO) shifted due to the adsorption. The absence of a band at ca. 1700 cm-1 indicates that the C=O group is not present in the surface species. On several minerals (Kubicki et al., 1999), spectra recorded at pH 3 and pH 6 are similar to spectra of solution species. With acetic acid adsorbed on quartz, two bands are seen around 1720 cm-1 (at 1709 and 1732), suggesting two different bonding environments. For the other minerals, the authors

Several surface complexes can be formed with monoacids, such as monodentate, or bidentates (Fig. 4). Monodentate surface complexes can be distinguished from bidentates based on the occurrence or not of the free C=O group band, with a stretching frequency at

M

(a) (b) (c)

The ATR-FTIR spectrum of 1M of **formate** ion is characterized by bands located at 1350, 1383 and 1580 cm-1, assigned to νs(COO), δ(HCO) and νas(COO), respectively (Rotzinger et al., 2004). Spectra of formate adsorbed on TiO2 at pH 5.0, up to 30 mM display the presence of bands of formate ions and a new peak at 1540 cm-1, assigned to νas(COO) of species interacting with the surface. Spectra in D2O confirmed this assignment since only a small shift (8 cm-1) of this band was observed, which precludes the vibration of a protonated/deuterated species. A decrease of pH from 9 to 3 leads to the decrease of the peak area. A series of experiments where the adsorption of formic acid as a gas has been studied has shown the presence of bands of formic acid, formate, and a peak at ca. 1540 cm-1. In support of this, molecular calculations have been performed for the three hypothetical surface complexes (Fig. 4), leading to calculated frequencies. Calculations on the stability of the surface complexes were found to support the binuclear bidentate

Sorption of **acetate** ions has been studied by ATR on rutile (Rotzinger et al., 2004) and several other minerals (Kubicki et al., 1999). In solution, the acetate ion is characterized by bands at 1348-1349, 1415-1422 and 1552-1555 cm-1 (Rotzinger et al., 2004; Kubicki et al., 1999) assigned to δ(CH3), νs(COO), and νas(COO), respectively. Acetic acid is characterized by bands at 1279-1283, 1370-1371, 1392-1397(δCH3), 1642-1650 and 1711-1717 (νC=O) (Rotzinger et al., 2004; Kubicki et al., 1999). Spectra of adsorbed species have been recorded at pH 5.0 (ca. 1:1 mixtures of the acetate ion and acetic acid in solution since pH is close to pKa), and at total acetate concentrations up to 25 mM on TiO2 (Rotzinger et al., 2004), and at pH 3 and 6 in the presence of 2 M acetate on quartz, albite, illite, kaolinite and montmorillonite (Kubicki et al., 1999). On TiO2, bands of acetate are present with a new band at 1512 cm-1 assigned to νas(COO) shifted due to the adsorption. The absence of a band at ca. 1700 cm-1 indicates that the C=O group is not present in the surface species. On several minerals (Kubicki et al., 1999), spectra recorded at pH 3 and pH 6 are similar to spectra of solution species. With acetic acid adsorbed on quartz, two bands are seen around 1720 cm-1 (at 1709 and 1732), suggesting two different bonding environments. For the other minerals, the authors

Fig. 4. Surface complexes between a monoacid and a metal oxide: (a) monodentate, (b)

M

O O

M

R

O O

R

about 1700 cm-1.

coordination.

M

O

O

R

mononuclear bidentate and (c) binuclear bidentate

conclude that chemisorption is below the detection limit of the spectroscopy. This might be due to the low specific surface area of the minerals used in the study.

Adsorption of **benzoic** acid on minerals was studied by several authors on quartz, albite, illite, kaolinite and montmorillonite (Kubicki et al., 1999), goethite (Tejedor-Tejedor et al., 1990), TiO2 (Tunesi and Anderson, 1992; Dobson and McQuillan, 1999), as well as Al2O3, ZrO2 and Ta2O5 (Dobson and McQuillan, 1999). Aqueous benzoate is characterized by bands at 1542 cm-1 (νas(COO)), 1388 cm-1 (νs(COO)) and 1593 cm-1 (νC=C). At pH < pKa, spectra are characterized by bands at 1705 cm-1 (νC=O), 1319 cm-1 (νCOH), 1279 cm-1 (δCOH), and bands associated with C=C and C-H vibrations (1603, 1494, 1452, 1178, 1073 and 1026 cm-1). Benzoic acid adsorbed on quartz displays bands of the aqueous species with two new peaks (at 1604 and 1569 cm-1). The lower frequency was found by calculation to correspond to a monodentate complex, and the higher one to an outer-sphere complex. On albite at pH 3, no peaks above 1700 cm-1 were observed, indicating that the C=O group is absent from the surface complex even in the pH range where the acid species predominates over the benzoate anion. This result is a direct evidence of the formation of a bidentate complex, stable over a wide range of pH. On goethite at pD 3.9 (Tejedor-Tejedor et al., 1990) and on TiO2 at pH 3.6 (Tunesi and Anderson, 1992), the νC=O mode is also absent. Another interesting point is that the asymmetric / symmetric carboxylate group stretching ratio decreases when benzoate interacts with Fe(III), which can be explained by the increase of coplanarity between the benzene ring and the νas(COO). These observations are consistent with the formation of a bidentate complex. c.f. Fig. 5 (Tejedor-Tejedor et al., 1990).

On TiO2, the difference between νas(COO) et νs(COO) for the adsorbed species is lower by 45 cm-1 compared to the corresponding difference for the solute species. It is believed that a lower value is indicative of a bidentate complex, and that such a large value indicates a chelate structure with a single centre (Fig. 5) (Tunesi and Anderson, 1992). On goethite, the difference was lower by 32 cm-1, consistent with a bridging complex (Tunesi and Anderson, 1992).

Fig. 5. Proposed surface complexes of benzoate on (A) TiO2 and (B) goethite.

The bands pertaining to νas(COO) and νs(COO) modes of **laurate** in solution are located at 1547 and 1411 cm-1, respectively. Between 2850 and 3000 cm-1, several bands are reported corresponding to hydrocarbon stretching. Laurate anions were adsorbed onto alumina, which had been deposited on the ATR element by a sputtering technique the thickness of the film being 600 Å (Couzis and Gulari, 1993). The recorded spectra depended on contact time and pH. At pH 8, up to 20 minutes after initiation of the solid-liquid contact, the observed peaks mainly corresponded to the solute species and the authors inferred the presence of an outer-sphere surface complex, since the surface is positively charged at this

Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of

and (COO-

from a bridging structure.

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 107

**Oxalic** acid is the simplest polyacid molecule (COOH)2 and its adsorption is the most common subject of study by ATR-IR on oxy-hydroxides of aluminum (Axe and Persson, 2001; Johnson et al., 2004; Rosenqvist et al., 2003; Yoon et al., 2004; Dobson and McQuillan, 1999), iron (Borda et al., 2003; Duckworth and Martin, 2001; Persson and Axe, 2001), chromium (Degenhardt and McQuillan, 1999; Garcia Rodenas et al., 1997), titanium (Hug and Sulzberger, 1994; Weisz et al., 2001; Weisz et al., 2002; Dobson and McQuillan, 1999), silicon (Kubicki et al., 1999), tantalum (Dobson and McQuillan, 1999) and zirconium (Dobson and McQuillan, 1999). The spectra of species in solution, i.e. (COOH)2, HOOCCOO-

at 1307 and 1571 cm-1, respectively, which are assigned to νas(COO) and νs(COO) modes. The spectrum of the oxalic acid species in solution is dominated by C=O stretching at 1735 cm-1 and C-OH stretching at 1227 cm-1. These features of oxalate and oxalic acid are consistent with theoretical frequency calculations (Axe and Persson, 2001). The spectra of hydrogen oxalate shows three peaks assigned to C=O stretching (1725 cm-1), νas(COO) (1620

Sorption of oxalate on boehmite was studied as a function of oxalate concentration and pH (Axe and Persson, 2001). Two different complexes were identified: an outer-sphere complex characterized by a spectrum similar to that of dissolved oxalate (two bands at 1577 and 1308 cm-1), and an inner-sphere complex. The assignment of this latter was based on the comparison of the spectra of the boehmite surface after sorption of oxalate (characterized by strong bands at 1722, 1702, 1413, 1288 cm-1) with the spectra of dissolved [Al(Ox)(H2O)4]+ (1725, 1706, 1412, 1281 cm-1). The very close resemblance suggests a mononuclear fivemembered chelate geometry. The possibility of a symmetric bridging coordination to two equivalent Al(III) ions was ruled out by Raman spectra of the surface species. Indeed, the comparison of Raman spectra of [Al(Ox)(H2O)4]+ with theoretical frequency calculations have indicated that the intensity of Raman bands can be used to distinguish a ring chelate

cm-1), and C-OH stretching (1240 cm-1) (Degenhardt and McQuillan, 1999).

Al

Fig. 6. Ring chelate of oxalate on alumina (Axe and Persson, 2001)

O

O

O

O

This interpretation has been supported by a study of oxalate sorption on corundum modelled by the CD-MUSIC model involving ATR-IR spectroscopy (Johnson et al., 2004). A mononuclear bidentate complex was found up to 14 µmol/m2, whereupon oxalate additionally adsorbed as an outer-sphere complex. Sorption of oxalate has also been studied on boehmite and corundum by Yoon et al. (2004) The peaks assigned to the inner-sphere complex in previous works (near 1286, 1418, 1700 and 1720 cm-1) were claimed to arise from the presence of several species. Evidence for this phenomenon comes from the observation that peaks at 1286 and 1418 cm-1 are shifted to 1297 and 1408 cm-1 as the oxalate surface coverage increases. The authors finally postulated the existence of two species: species "A" at 1286 and 1418 cm-1, and species "B" at 1297 and 1408 cm-1, respectively, which were

)2 were reported in several studies. The oxalate ion is characterized by two peaks

pH. For longer times of exposure, a new band appeared at 1597, along with the increase of the band at 1412 cm-1. This new band was assigned to νas(COO) of the adsorbed species. The difference between νas(COO) and νs(COO) for the surface species was higher than the value obtained for the solute species. This behaviour, contrary to that observed for carboxylic acids with a shorter alkyl chain (Table 3) has been interpreted as a different, i.e. monodentate, surface coordination. From the evolution of the spectra recorded in the hydrocarbon stretching range (2750 – 3000 cm-1), a chain-chain interaction is inferred after adsorption of laurate for short contact times, suggesting the association of the aliphatic chains at low surface coverage. For longer contact times, corresponding to a higher surface coverage, the chain-chain interactions become negligible.

### **2.3.2 Saturated and unsaturated diacids**

This section in a similar way as the previous one summarizes a number of studies on the adsorption saturated and unsaturated diacids to (oxy)(hydr)oxide minerals. The chemical speciation (in terms of the number of species in solution) becomes more complex for these compounds, which concomitantly enhances the possibilities of the diacids to form surface complexes of different stoichiometries in terms of bonding and proton balances. The diacids addressed are summarized in tables 3 (saturated diacids) and 4 (unsaturated diacids). The remainder of the section discusses in some detail published findings from ATR-FTIR spectroscopy.


Table 3. Characteristics of dicarbocylic acids (HO(O)C–R–C(O)OH).


Table 4. Characteristics of unsaturated dicarbocylic acids.

pH. For longer times of exposure, a new band appeared at 1597, along with the increase of the band at 1412 cm-1. This new band was assigned to νas(COO) of the adsorbed species. The difference between νas(COO) and νs(COO) for the surface species was higher than the value obtained for the solute species. This behaviour, contrary to that observed for carboxylic acids with a shorter alkyl chain (Table 3) has been interpreted as a different, i.e. monodentate, surface coordination. From the evolution of the spectra recorded in the hydrocarbon stretching range (2750 – 3000 cm-1), a chain-chain interaction is inferred after adsorption of laurate for short contact times, suggesting the association of the aliphatic chains at low surface coverage. For longer contact times, corresponding to a higher surface

This section in a similar way as the previous one summarizes a number of studies on the adsorption saturated and unsaturated diacids to (oxy)(hydr)oxide minerals. The chemical speciation (in terms of the number of species in solution) becomes more complex for these compounds, which concomitantly enhances the possibilities of the diacids to form surface complexes of different stoichiometries in terms of bonding and proton balances. The diacids addressed are summarized in tables 3 (saturated diacids) and 4 (unsaturated diacids). The remainder of the section discusses in some detail published findings from ATR-FTIR

Formula

OH

OH

O

O

O

O

O

OH

O

OH

coverage, the chain-chain interactions become negligible.

Acid pKa1, pKa2 Lide (1998) R Oxalic 1.23, 4.19 N.A. Malonic 2.83, 5.69 –CH2– Succinic 4.16, 5.61 –(CH2)2– Glutaric 4.31, 5.41 –(CH2)3– Adipic 4.43, 5.51 –(CH2)4– Table 3. Characteristics of dicarbocylic acids (HO(O)C–R–C(O)OH).

fom Lide (1998)

*o*-Phtalic 2.89, 5.51 OH

Table 4. Characteristics of unsaturated dicarbocylic acids.

Maleic 1.83, 6.07 OH

**2.3.2 Saturated and unsaturated diacids** 

Acid pKa1, pKa2

*trans*-Fumaric 3.03, 4.44

spectroscopy.

**Oxalic** acid is the simplest polyacid molecule (COOH)2 and its adsorption is the most common subject of study by ATR-IR on oxy-hydroxides of aluminum (Axe and Persson, 2001; Johnson et al., 2004; Rosenqvist et al., 2003; Yoon et al., 2004; Dobson and McQuillan, 1999), iron (Borda et al., 2003; Duckworth and Martin, 2001; Persson and Axe, 2001), chromium (Degenhardt and McQuillan, 1999; Garcia Rodenas et al., 1997), titanium (Hug and Sulzberger, 1994; Weisz et al., 2001; Weisz et al., 2002; Dobson and McQuillan, 1999), silicon (Kubicki et al., 1999), tantalum (Dobson and McQuillan, 1999) and zirconium (Dobson and McQuillan, 1999). The spectra of species in solution, i.e. (COOH)2, HOOCCOOand (COO- )2 were reported in several studies. The oxalate ion is characterized by two peaks at 1307 and 1571 cm-1, respectively, which are assigned to νas(COO) and νs(COO) modes. The spectrum of the oxalic acid species in solution is dominated by C=O stretching at 1735 cm-1 and C-OH stretching at 1227 cm-1. These features of oxalate and oxalic acid are consistent with theoretical frequency calculations (Axe and Persson, 2001). The spectra of hydrogen oxalate shows three peaks assigned to C=O stretching (1725 cm-1), νas(COO) (1620 cm-1), and C-OH stretching (1240 cm-1) (Degenhardt and McQuillan, 1999).

Sorption of oxalate on boehmite was studied as a function of oxalate concentration and pH (Axe and Persson, 2001). Two different complexes were identified: an outer-sphere complex characterized by a spectrum similar to that of dissolved oxalate (two bands at 1577 and 1308 cm-1), and an inner-sphere complex. The assignment of this latter was based on the comparison of the spectra of the boehmite surface after sorption of oxalate (characterized by strong bands at 1722, 1702, 1413, 1288 cm-1) with the spectra of dissolved [Al(Ox)(H2O)4]+ (1725, 1706, 1412, 1281 cm-1). The very close resemblance suggests a mononuclear fivemembered chelate geometry. The possibility of a symmetric bridging coordination to two equivalent Al(III) ions was ruled out by Raman spectra of the surface species. Indeed, the comparison of Raman spectra of [Al(Ox)(H2O)4]+ with theoretical frequency calculations have indicated that the intensity of Raman bands can be used to distinguish a ring chelate from a bridging structure.

Fig. 6. Ring chelate of oxalate on alumina (Axe and Persson, 2001)

This interpretation has been supported by a study of oxalate sorption on corundum modelled by the CD-MUSIC model involving ATR-IR spectroscopy (Johnson et al., 2004). A mononuclear bidentate complex was found up to 14 µmol/m2, whereupon oxalate additionally adsorbed as an outer-sphere complex. Sorption of oxalate has also been studied on boehmite and corundum by Yoon et al. (2004) The peaks assigned to the inner-sphere complex in previous works (near 1286, 1418, 1700 and 1720 cm-1) were claimed to arise from the presence of several species. Evidence for this phenomenon comes from the observation that peaks at 1286 and 1418 cm-1 are shifted to 1297 and 1408 cm-1 as the oxalate surface coverage increases. The authors finally postulated the existence of two species: species "A" at 1286 and 1418 cm-1, and species "B" at 1297 and 1408 cm-1, respectively, which were

Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of

complex in their studies, even if they have not mentioned this possibility.

absent in malonate and glutarate surface species.

Duckworth and Martin (2001) and Dobson and McQuillan (1999) have studied the effect of longer carbon chains on the adsorption of dicarboxylic acids. Dobson and McQuillan (1999) found surface structures similar to malonate (bridging bidentate via a loop) consistent with greater molecular flexibility. On the contrary, Duckworth and Martin (2001) have found either a bridging bidentate via a loop for malonate and glutarate, or a monodentate complex for succinate and adipate. This interpretation is supported by the behaviour of these ions in the dissolution of hematite: oxalate, malonate and glutarate promote the dissolution whereas succinate and adipate show less of an effect. However, the spectral evidence for this difference in the geometry of complexes is the presence of a peak at 1550-1540 cm-1, *i.e.* at a similar location as (νas(CO2)) of the solution species, whereas this stretching mode was

Three unsaturated dicarboxylic acids have been studied by ATR: fumaric (Dobson and McQuillan, 1999; Rosenqvist et al., 2003), maleic (Dobson and McQuillan, 1999; Borda et al., 2003; Rosenqvist et al., 2003, Johnson et al., 2004), (respectively trans- and cis- butendioic acids), and phtalic (Boily et al., 2000; Dobson and McQuillan, 1999; Hwang et al., 2007; Klug and Forsling, 1999; Kubicki et al., 1999; Nordin et al., 1997; Rosenqvist et al., 2003; Tunesi and Anderson, 1992) acid. On gibbsite, the spectra of these three species after sorption are

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 109

ATR-FTIR studies of the sorption of other saturated diacids HO(O)C-(CH2)*n*-C(O)OH with *n*=1 to 4 have been reported, as malonic (Dobson and McQuillan, 1999; Dolamic and Bürgi, 2006; Duckworth and Martin, 2001; Rosenqvist et al., 2003), succinic (Dobson and McQuillan, 1999; Duckworth and Martin, 2001), glutaric (Duckworth and Martin, 2001) and adipic (Dobson and McQuillan, 1999; Duckworth and Martin, 2001) acids. In solution, the dicarboxylate ions are characterized by νas(CO2) around 1560-1550 cm-1 and νs(CO2) around 1410-1350 cm-1 (Dobson and McQuillan, 1999). The value of νas(CO2) is close for all values of *n* (and equal to frequency in oxalate), but νs(CO2) increases with *n*, from ca. 1360 cm-1 (1310 cm-1 in oxalate) to 1400 cm-1. Dicarboxylic acids are characterized by ν(C=O) at 1718 cm-1 (malonic) and ν(C-O) at 1328 cm-1 (malonic). The –CH2– bending frequencies are located at 1410-1440 cm-1 and 1300-1250 cm-1. Malonate adsorbed on hematite at pH 5 is characterized by peaks at 1260 (δ(-CH2-)), 1349 (νs(CO2)), 1439 (δ(-CH2-)) and 1631 (ν(C=O)) cm-1. νs(CO2) remains at 1349 cm-1 without shift. In comparison with solution spectra, the intensity of – CH2- bending is enhanced, and the CO2 asymmetric stretching is replaced by ν(C=O) (70 cm-1 higher). Such an assignment is consistent with a single-bonded surface complex, a structure similar to adsorbed oxalate. The –CH2- bending band enhancement indicates a change of the dipole moment, which may be an indication of a strained surface structure with increased bond angles (Dobson and McQuillan, 1999). On other metallic oxides, as TiO2 (Dobson and McQuillan, 1999; Dolamic and Bürgi, 2006), ZrO2 (Dobson and McQuillan, 1999), Al2O3 (Dobson and McQuillan, 1999; Rosenqvist et al., 2003) and Ta2O5 (Dobson and McQuillan, 1999), the spectra of sorbed malonate species are similar: δ(-CH2-) at 1430-1450 and 1270- 1280 cm-1, ν(C=O) at 1580-1600 cm-1, and unshifted νs(CO2) at 1360-1380 cm-1. On gibbsite (Rosenqvist et al., 2003), the study of the evolution of the spectra with pH has shown two independent species: an inner-sphere complex (corresponding to a peak at 1438 cm-1) and an outer-sphere complex (corresponding to the unshifted νs(CO2)). Since the other authors (Dobson and McQuillan, 1999; Dolamic and Bürgi, 2006; Duckworth and Martin, 2001) have not studied the effect of pH, it is not possible to rule out the presence of an outer-sphere

assigned to an inner-sphere surface complex on boehmite and to dissolved oxalate coordinated to aqueous Al(III). This assumption is supported by oxalate promoted dissolution of the aluminum oxy-hydroxide arising from the complexation reactions of dissolved Al(III) cations. Quantum calculations of infrared vibrational frequencies of possible surface complexes were carried out on aluminium oxide clusters (Al18O12 and Al14O22) including monodentate, bidentates with 4- and 5-membered ring, and bridging bidentates. They showed that the bidentate 5-membered ring most closely matched the experimental observations (within 15 cm-1), while the simulation results for the other models showed deviations between 17 and 102 cm-1.

On hematite (Duckworth and Martin, 2001), the spectra of sorbed oxalate are similar to the above discussed surface complex on an aluminum oxy-hydroxide and consequently a 5 member bidentate complex was proposed. The effect of pH on the sorption of oxalate on goethite has also been studied (Persson and Axe, 2001). An outer-sphere surface complex and a 5-member ring inner-sphere surface complex were inferred from spectra of the goethite/oxalate system and the aqueous Fe(III)-oxalate complex. At low pH, the presence of outer-sphere surface complexes (COOH)2 was ruled out because of the absence of a band corresponding to these species in aqueous solutions (around 1735 and 1233 cm-1).

Oxalate sorption on chromium oxide (Degenhardt and McQuillan, 1999) is characterized by bands at 1708, 1680, 1405 and 1269 cm-1, which was interpreted as a side-on surface complex (both carboxylic groups interact with the surface), but without distinguishing between a 5 or a 7-member ring. An additional weakly bound oxalate ion was detected (bands at 1620- 1580 and around 1306 cm-1). The absence of absorption at 1725 cm-1 (corresponding to C=O group) eliminates the singly protonated oxalate species. However, an upward shift of νas(COO) is observed (from 1571 cm-1 in Ox2-), suggesting hydrogen bonding with the surface. On chromium oxide (Garcia Rodenas et al., 1997), spectra were recorded after exposure to 0.1 M oxalate solution at pH 3.6 followed by washing with pure water. The resulting spectra were compared to the spectra of Cr(Ox)3 3- species in solution. As in the work by Degenhardt and McQuillan (1999), an inner-sphere surface complex was inferred from the bands at 1710, 1680, 1410 and 1260 cm-1. The remaining shoulder at 1620 cm-1 and the peak at 1310 cm-1 were attributed to uncoordinated oxalate ions. Since the solid has been washed after contacting with oxalate solution, the solute species are expected to be removed, and these data could be reinterpreted as a species involving hydrogen bonding.

Oxalate sorption onto TiO2 was amongst the first *in situ* adsorption studies involving ATR-IR spectroscopy at the solid/solution interface (Hug and Sulzberger, 1994). Hug and Sulzberger (1994) have focused their study on the measurement of adsorbed oxalate to plot an isotherm curve. The isotherm at constant pH (3) was fitted by three Langmuir components, correlated with the three possible solute species (H2Ox/HOx-/Ox2-). Weisz et al. (2001, 2002) have used the same protocol, and have measured three Langmuir stability constants. Dobson and McQuillan (1999) have recorded the spectra of Na2[TiO(Ox)2]2. 3H2O(S), where the oxalate ion forms a µ2-oxo bridged Ti dimeric complex. Its spectrum is close to those obtained with oxalate adsorbed onto TiO2 and the comparison would lead to the interpretation of the spectroscopic results in terms of a bidentate-bridging surface complex. However, this interpretation disagrees with observations by Scott et al. (1973) who have shown on oxalato-Co(III) complexes that bidentate-chelating and bidentatebridging oxalato ligands are characterized by nearly identical spectra.

assigned to an inner-sphere surface complex on boehmite and to dissolved oxalate coordinated to aqueous Al(III). This assumption is supported by oxalate promoted dissolution of the aluminum oxy-hydroxide arising from the complexation reactions of dissolved Al(III) cations. Quantum calculations of infrared vibrational frequencies of possible surface complexes were carried out on aluminium oxide clusters (Al18O12 and Al14O22) including monodentate, bidentates with 4- and 5-membered ring, and bridging bidentates. They showed that the bidentate 5-membered ring most closely matched the experimental observations (within 15 cm-1), while the simulation results for the other models

On hematite (Duckworth and Martin, 2001), the spectra of sorbed oxalate are similar to the above discussed surface complex on an aluminum oxy-hydroxide and consequently a 5 member bidentate complex was proposed. The effect of pH on the sorption of oxalate on goethite has also been studied (Persson and Axe, 2001). An outer-sphere surface complex and a 5-member ring inner-sphere surface complex were inferred from spectra of the goethite/oxalate system and the aqueous Fe(III)-oxalate complex. At low pH, the presence of outer-sphere surface complexes (COOH)2 was ruled out because of the absence of a band

Oxalate sorption on chromium oxide (Degenhardt and McQuillan, 1999) is characterized by bands at 1708, 1680, 1405 and 1269 cm-1, which was interpreted as a side-on surface complex (both carboxylic groups interact with the surface), but without distinguishing between a 5 or a 7-member ring. An additional weakly bound oxalate ion was detected (bands at 1620- 1580 and around 1306 cm-1). The absence of absorption at 1725 cm-1 (corresponding to C=O group) eliminates the singly protonated oxalate species. However, an upward shift of νas(COO) is observed (from 1571 cm-1 in Ox2-), suggesting hydrogen bonding with the surface. On chromium oxide (Garcia Rodenas et al., 1997), spectra were recorded after exposure to 0.1 M oxalate solution at pH 3.6 followed by washing with pure water. The resulting spectra were compared to the spectra of Cr(Ox)33- species in solution. As in the work by Degenhardt and McQuillan (1999), an inner-sphere surface complex was inferred from the bands at 1710, 1680, 1410 and 1260 cm-1. The remaining shoulder at 1620 cm-1 and the peak at 1310 cm-1 were attributed to uncoordinated oxalate ions. Since the solid has been washed after contacting with oxalate solution, the solute species are expected to be removed, and these data could be reinterpreted as a species involving hydrogen bonding. Oxalate sorption onto TiO2 was amongst the first *in situ* adsorption studies involving ATR-IR spectroscopy at the solid/solution interface (Hug and Sulzberger, 1994). Hug and Sulzberger (1994) have focused their study on the measurement of adsorbed oxalate to plot an isotherm curve. The isotherm at constant pH (3) was fitted by three Langmuir components, correlated with the three possible solute species (H2Ox/HOx-/Ox2-). Weisz et al. (2001, 2002) have used the same protocol, and have measured three Langmuir stability constants. Dobson and McQuillan (1999) have recorded the spectra of

3H2O(S), where the oxalate ion forms a µ2-oxo bridged Ti dimeric complex. Its

spectrum is close to those obtained with oxalate adsorbed onto TiO2 and the comparison would lead to the interpretation of the spectroscopic results in terms of a bidentate-bridging surface complex. However, this interpretation disagrees with observations by Scott et al. (1973) who have shown on oxalato-Co(III) complexes that bidentate-chelating and bidentate-

bridging oxalato ligands are characterized by nearly identical spectra.

corresponding to these species in aqueous solutions (around 1735 and 1233 cm-1).

showed deviations between 17 and 102 cm-1.

Na2[TiO(Ox)2]2.

ATR-FTIR studies of the sorption of other saturated diacids HO(O)C-(CH2)*n*-C(O)OH with *n*=1 to 4 have been reported, as malonic (Dobson and McQuillan, 1999; Dolamic and Bürgi, 2006; Duckworth and Martin, 2001; Rosenqvist et al., 2003), succinic (Dobson and McQuillan, 1999; Duckworth and Martin, 2001), glutaric (Duckworth and Martin, 2001) and adipic (Dobson and McQuillan, 1999; Duckworth and Martin, 2001) acids. In solution, the dicarboxylate ions are characterized by νas(CO2) around 1560-1550 cm-1 and νs(CO2) around 1410-1350 cm-1 (Dobson and McQuillan, 1999). The value of νas(CO2) is close for all values of *n* (and equal to frequency in oxalate), but νs(CO2) increases with *n*, from ca. 1360 cm-1 (1310 cm-1 in oxalate) to 1400 cm-1. Dicarboxylic acids are characterized by ν(C=O) at 1718 cm-1 (malonic) and ν(C-O) at 1328 cm-1 (malonic). The –CH2– bending frequencies are located at 1410-1440 cm-1 and 1300-1250 cm-1. Malonate adsorbed on hematite at pH 5 is characterized by peaks at 1260 (δ(-CH2-)), 1349 (νs(CO2)), 1439 (δ(-CH2-)) and 1631 (ν(C=O)) cm-1. νs(CO2) remains at 1349 cm-1 without shift. In comparison with solution spectra, the intensity of – CH2- bending is enhanced, and the CO2 asymmetric stretching is replaced by ν(C=O) (70 cm-1 higher). Such an assignment is consistent with a single-bonded surface complex, a structure similar to adsorbed oxalate. The –CH2- bending band enhancement indicates a change of the dipole moment, which may be an indication of a strained surface structure with increased bond angles (Dobson and McQuillan, 1999). On other metallic oxides, as TiO2 (Dobson and McQuillan, 1999; Dolamic and Bürgi, 2006), ZrO2 (Dobson and McQuillan, 1999), Al2O3 (Dobson and McQuillan, 1999; Rosenqvist et al., 2003) and Ta2O5 (Dobson and McQuillan, 1999), the spectra of sorbed malonate species are similar: δ(-CH2-) at 1430-1450 and 1270- 1280 cm-1, ν(C=O) at 1580-1600 cm-1, and unshifted νs(CO2) at 1360-1380 cm-1. On gibbsite (Rosenqvist et al., 2003), the study of the evolution of the spectra with pH has shown two independent species: an inner-sphere complex (corresponding to a peak at 1438 cm-1) and an outer-sphere complex (corresponding to the unshifted νs(CO2)). Since the other authors (Dobson and McQuillan, 1999; Dolamic and Bürgi, 2006; Duckworth and Martin, 2001) have not studied the effect of pH, it is not possible to rule out the presence of an outer-sphere complex in their studies, even if they have not mentioned this possibility.

Duckworth and Martin (2001) and Dobson and McQuillan (1999) have studied the effect of longer carbon chains on the adsorption of dicarboxylic acids. Dobson and McQuillan (1999) found surface structures similar to malonate (bridging bidentate via a loop) consistent with greater molecular flexibility. On the contrary, Duckworth and Martin (2001) have found either a bridging bidentate via a loop for malonate and glutarate, or a monodentate complex for succinate and adipate. This interpretation is supported by the behaviour of these ions in the dissolution of hematite: oxalate, malonate and glutarate promote the dissolution whereas succinate and adipate show less of an effect. However, the spectral evidence for this difference in the geometry of complexes is the presence of a peak at 1550-1540 cm-1, *i.e.* at a similar location as (νas(CO2)) of the solution species, whereas this stretching mode was absent in malonate and glutarate surface species.

Three unsaturated dicarboxylic acids have been studied by ATR: fumaric (Dobson and McQuillan, 1999; Rosenqvist et al., 2003), maleic (Dobson and McQuillan, 1999; Borda et al., 2003; Rosenqvist et al., 2003, Johnson et al., 2004), (respectively trans- and cis- butendioic acids), and phtalic (Boily et al., 2000; Dobson and McQuillan, 1999; Hwang et al., 2007; Klug and Forsling, 1999; Kubicki et al., 1999; Nordin et al., 1997; Rosenqvist et al., 2003; Tunesi and Anderson, 1992) acid. On gibbsite, the spectra of these three species after sorption are

Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of

the platelets of about 200 nm and a thickness of about 10 nm.

Aldrich), AlCl3.6H2O (Merck p.a.), standard buffers (pH = 3, 5, 7, 9 and 11).

OH

SO3-

Fig. 8. Equilibrium constants for the deprotonation of 5-SSA (Panak, 1996).

SSA-gibbsite system.

**3.2 Experimental 3.2.1 Chemicals** 

depends on pH (Fig. 8).

OH

COOH

predominantly bound in Al-complexes up to pH 11.

SO3H

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 111

charging studies mentioned above to lay a foundation to the study of the ternary Cm-

Adsorption of 5-SSA to related minerals (i.e. alumina) has been previously studied previously by Jiang et al. (2002) by both a set of electrokinetic data and batch adsorption. Furthermore they reported IR data which they interpreted in terms of 5-SSA forming a

The gibbsite particles were synthesized by the following procedure: 1 mol.dm–3 aluminum chloride solution was titrated with 4 mol.dm–3 NaOH until pH reached a value of about 4.6. Dialysis was carried out at 70 °C during four months, with initially one change of water per day. Subsequently, water was changed two to three times a week. The gibbsite was stored as a suspension at a concentration of 41.9 g.dm–3. The radius of the particles was determined by several methods, including AFM and field flow fractionation yielding an average width of

The suspension and solution were prepared with de-ionized water (conductivity ~ 18.2 MΩ cm). All solutions and suspensions were prepared in plastic containers. The following chemical reagents were used: NaCl (*p.a*., Merck), HCl (0.1 mol.dm–3, titrival, Merck), NaOH (0.1 mol.dm–3, titrival, Merck), 5-sulfosalicylic acid (5-Sulfosalicylic acid.2H2O, Sigma

Speciation of 5-SSA in solution results in four species. The occurrence of these species

OH

COO-

O-

COO-

SO3-

SO3-

COOH

pKa1 = 1.52 pKa2 = 2.32 pKa3 = 11.30

In presence of aluminium ions, a number of Al(5-SSA)x species (with x=1 to 3) can be formed. The speciation of 5-SSA and Al as a function of pH for 5×10-2 mol.dm-3 of each component is plotted in Fig. 9. The relevant Al-species are shown in the upper part, indicating that Al is preferably bound to 5-SSA up to pH 8. Above pH 8, the tetra-hydroxo species of aluminium dominates the speciation under the given conditions. The lower part of Figure 9 shows the distribution of 5-SSA for the equimolar solution. 5-SSA is

bidentate surface complex involving the carboxylate group and the phenol group.

Fig. 7. Proposed structures of diacids adsorbed on hematite (Duckworth and Martin, 2001)

very similar to spectra of the carboxylate ion in solution. Consequently, Rosenqvist et al. (2003) have concluded that an outer-sphere complex was formed. On Al2O3, the same behaviour was shown for maleate and fumarate (Dobson and McQuillan, 1999). On corundum, high resolution spectra of the adsorbed maleate ion were recorded (Johnson et al., 2004), and only small differences between bands of dissolved and adsorbed species have been detected, supporting outer-sphere complexation. For phtalate, inner-sphere and outersphere complexes were found on ferric oxy-hydroxides (Boily et al., 2000; Hwang et al., 2007).

## **3. Adsorption of 5-sulfosalicylic acid onto gibbsite**

#### **3.1 Introduction**

The system 5-sulfosalicylic acid (5-SSA)/gibbsite is part of a broader study involving Cm adsorption onto gibbsite (Huittinen et al., 2009). Interaction of dissolved Cm with 5-SSA (Panak, 1996) and acid-base equilibria of gibbsite (Adekola et al., 2011) have been studied separately. Gibbsite as used in this study has a platelike morphology. Most of its surface relates to the basal plane. A very complex issue is the interfacial behaviour of the bare gibbsite. While the basal plane from a conventional point of view is considered quite inert, there are a number of indications that rather show that the basal plane can be quite reactive (Rosenquist et al., 2002; Gan and Franks, 2006). The present view is that a number of effects can be of importance, such as the precise conditions for the preparation of the gibbsite particles. This has been further studied by mimicking the basal plane by individual single crystals of sapphire, which are structurally very similar to the ideal basal plane of gibbsite (Lützenkirchen et al., 2010). Here we present some results on the interaction of 5-SSA with the gibbsite particles used in the Cm adsorption and basic charging studies mentioned above to lay a foundation to the study of the ternary Cm-SSA-gibbsite system.

Adsorption of 5-SSA to related minerals (i.e. alumina) has been previously studied previously by Jiang et al. (2002) by both a set of electrokinetic data and batch adsorption. Furthermore they reported IR data which they interpreted in terms of 5-SSA forming a bidentate surface complex involving the carboxylate group and the phenol group.

#### **3.2 Experimental**

110 Infrared Spectroscopy – Materials Science, Engineering and Technology

Fe

Fe

Fig. 7. Proposed structures of diacids adsorbed on hematite (Duckworth and Martin, 2001)

very similar to spectra of the carboxylate ion in solution. Consequently, Rosenqvist et al. (2003) have concluded that an outer-sphere complex was formed. On Al2O3, the same behaviour was shown for maleate and fumarate (Dobson and McQuillan, 1999). On corundum, high resolution spectra of the adsorbed maleate ion were recorded (Johnson et al., 2004), and only small differences between bands of dissolved and adsorbed species have been detected, supporting outer-sphere complexation. For phtalate, inner-sphere and outersphere complexes were found on ferric oxy-hydroxides (Boily et al., 2000; Hwang et al.,

The system 5-sulfosalicylic acid (5-SSA)/gibbsite is part of a broader study involving Cm adsorption onto gibbsite (Huittinen et al., 2009). Interaction of dissolved Cm with 5-SSA (Panak, 1996) and acid-base equilibria of gibbsite (Adekola et al., 2011) have been studied separately. Gibbsite as used in this study has a platelike morphology. Most of its surface relates to the basal plane. A very complex issue is the interfacial behaviour of the bare gibbsite. While the basal plane from a conventional point of view is considered quite inert, there are a number of indications that rather show that the basal plane can be quite reactive (Rosenquist et al., 2002; Gan and Franks, 2006). The present view is that a number of effects can be of importance, such as the precise conditions for the preparation of the gibbsite particles. This has been further studied by mimicking the basal plane by individual single crystals of sapphire, which are structurally very similar to the ideal basal plane of gibbsite (Lützenkirchen et al., 2010). Here we present some results on the interaction of 5-SSA with the gibbsite particles used in the Cm adsorption and basic

O

O

OH2+

O

O

OH2+

O

O

O

O

Fe

Fe

2007).

**3.1 Introduction** 

O

O

O

O

O

**3. Adsorption of 5-sulfosalicylic acid onto gibbsite** 

O

O

O

#### **3.2.1 Chemicals**

The gibbsite particles were synthesized by the following procedure: 1 mol.dm–3 aluminum chloride solution was titrated with 4 mol.dm–3 NaOH until pH reached a value of about 4.6. Dialysis was carried out at 70 °C during four months, with initially one change of water per day. Subsequently, water was changed two to three times a week. The gibbsite was stored as a suspension at a concentration of 41.9 g.dm–3. The radius of the particles was determined by several methods, including AFM and field flow fractionation yielding an average width of the platelets of about 200 nm and a thickness of about 10 nm.

The suspension and solution were prepared with de-ionized water (conductivity ~ 18.2 MΩ cm). All solutions and suspensions were prepared in plastic containers. The following chemical reagents were used: NaCl (*p.a*., Merck), HCl (0.1 mol.dm–3, titrival, Merck), NaOH (0.1 mol.dm–3, titrival, Merck), 5-sulfosalicylic acid (5-Sulfosalicylic acid.2H2O, Sigma Aldrich), AlCl3.6H2O (Merck p.a.), standard buffers (pH = 3, 5, 7, 9 and 11).

Speciation of 5-SSA in solution results in four species. The occurrence of these species depends on pH (Fig. 8).

Fig. 8. Equilibrium constants for the deprotonation of 5-SSA (Panak, 1996).

In presence of aluminium ions, a number of Al(5-SSA)x species (with x=1 to 3) can be formed. The speciation of 5-SSA and Al as a function of pH for 5×10-2 mol.dm-3 of each component is plotted in Fig. 9. The relevant Al-species are shown in the upper part, indicating that Al is preferably bound to 5-SSA up to pH 8. Above pH 8, the tetra-hydroxo species of aluminium dominates the speciation under the given conditions. The lower part of Figure 9 shows the distribution of 5-SSA for the equimolar solution. 5-SSA is predominantly bound in Al-complexes up to pH 11.

Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of


(squares). Temperature was 25 ºC.

**3.2.3 Spectroscopy** 

suspension.

**3.3 Results and discussion** 

**3.3.1 Spectral characterization of 5-SSA in solution** 

0

10

ξ / mV

20

30

40

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 113

1 2 3 4 5 6 7 8 9 10 11 12 13

pH

Fig. 10. Zeta-potentials of gibbsite particles in the absence (□,○) and in the presence (■,●) of 5-SSA. Ionic strength was controlled by NaCl: 10–1 mol.dm–3 (circles) and 10–2 mol.dm–3

For both ionic strengths studied, the electrokinetic data show (i) a strong shift of the isoelectric point and (ii) negative zeta-potentials over a wide pH range, indicating adsorption of 5-SSA and transfer of negative charges to the gibbsite surface. The speciation diagram (Fig. 9) actually shows that the speciation of 5-SSA is dominated both in the

The IR-ATR spectra of gibbsite in absence and presence of 5-SSA, as well as spectra of the aqueous solutions of the 5-SSA and 5-SSA/Al3+ were obtained using a Bruker spectrometer (IFS 55). A ZnSe crystal (multibounce) was used and for each measurement 1024 scans were recorded with a resolution of 4 cm–1. All measurements were made under dry argon atmosphere. The effect of pH, which has an influence on gibbsite surface charge as well on the speciation of 5-SSA in solution was examined. For spectra of the gibbsite layer in contact with solution of 5-SSA, the gibbsite layer was prepared by drying an aliquot of a

Solution spectra of 5-SSA, with and without aluminium ions have been recorded by ATR-IR. For pure 5-SSA solution at pH 2, 5 and 12, the spectra are shown in Fig. 11. At pH 2, the spectrum consists in several bands which have been assigned following previous works by Varghese et al. (2007) and Jiang et al. (2002) as shown in Table 5. From pH 2 to 5, some differences can be seen: bands around 1200 cm-1 decrease, bands at 1270 and 1306 cm-1

absence and the presence of Al by negatively charged aqueous species.

Fig. 9. Calculated speciation for an equimolar solution (5×10-2 M) of Al(III) (top) and 5-SSA (bottom). Uncomplexed species are in thin lines, and complex Al:5-SSA are in bold lines. Only species whose concentration are higher than 5% are represented.

#### **3.2.2 Electrokinetics**

The electrokinetic (zeta) potential of gibbsite particles was measured after adsorption of 5- SSA on gibbsite surfaces by means of a ZetaPals *(Brookhaven Instruments)*. The mass concentration of gibbsite particles was 0.1 g.dm-3 and 5-SSA concentration was 10–3 mol.dm-3. The experiments were performed at two different ionic strength values (*Ic* = 10-1 mol.dm-3 and 10-2 mol.dm-3). The results are shown in Figure 10 together with zeta potential of gibbsite particles in the absence of 5-SSA (Adekola et al., 2011).

Fig. 10. Zeta-potentials of gibbsite particles in the absence (□,○) and in the presence (■,●) of 5-SSA. Ionic strength was controlled by NaCl: 10–1 mol.dm–3 (circles) and 10–2 mol.dm–3 (squares). Temperature was 25 ºC.

For both ionic strengths studied, the electrokinetic data show (i) a strong shift of the isoelectric point and (ii) negative zeta-potentials over a wide pH range, indicating adsorption of 5-SSA and transfer of negative charges to the gibbsite surface. The speciation diagram (Fig. 9) actually shows that the speciation of 5-SSA is dominated both in the absence and the presence of Al by negatively charged aqueous species.

#### **3.2.3 Spectroscopy**

112 Infrared Spectroscopy – Materials Science, Engineering and Technology

**- 1:1**

2 4 6 8 10 12 pH

Fig. 9. Calculated speciation for an equimolar solution (5×10-2 M) of Al(III) (top) and 5-SSA (bottom). Uncomplexed species are in thin lines, and complex Al:5-SSA are in bold lines.

The electrokinetic (zeta) potential of gibbsite particles was measured after adsorption of 5- SSA on gibbsite surfaces by means of a ZetaPals *(Brookhaven Instruments)*. The mass concentration of gibbsite particles was 0.1 g.dm-3 and 5-SSA concentration was 10–3 mol.dm-3. The experiments were performed at two different ionic strength values (*Ic* = 10-1 mol.dm-3 and 10-2 mol.dm-3). The results are shown in Figure 10 together with zeta potential of

**1:3**

**Al(OH)4**

**1:3**

**SSA3-**

**1:2**

**Al 13**

**1:2**

**HSSA <sup>H</sup>** <sup>2</sup>**- <sup>2</sup>**

Only species whose concentration are higher than 5% are represented.

gibbsite particles in the absence of 5-SSA (Adekola et al., 2011).

0

0.01

**3.2.2 Electrokinetics** 

0.02

0.03

0.04

**1:1**

**Al3+**

**SSA-**

0.05

00.01

0.02

0.03

0.04

0.05

0.06

Concentration (M)

The IR-ATR spectra of gibbsite in absence and presence of 5-SSA, as well as spectra of the aqueous solutions of the 5-SSA and 5-SSA/Al3+ were obtained using a Bruker spectrometer (IFS 55). A ZnSe crystal (multibounce) was used and for each measurement 1024 scans were recorded with a resolution of 4 cm–1. All measurements were made under dry argon atmosphere. The effect of pH, which has an influence on gibbsite surface charge as well on the speciation of 5-SSA in solution was examined. For spectra of the gibbsite layer in contact with solution of 5-SSA, the gibbsite layer was prepared by drying an aliquot of a suspension.

#### **3.3 Results and discussion**

#### **3.3.1 Spectral characterization of 5-SSA in solution**

Solution spectra of 5-SSA, with and without aluminium ions have been recorded by ATR-IR. For pure 5-SSA solution at pH 2, 5 and 12, the spectra are shown in Fig. 11. At pH 2, the spectrum consists in several bands which have been assigned following previous works by Varghese et al. (2007) and Jiang et al. (2002) as shown in Table 5. From pH 2 to 5, some differences can be seen: bands around 1200 cm-1 decrease, bands at 1270 and 1306 cm-1

Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of

but a consequence on vibration of other groups would be possible.

**onto gibbsite** 

measurements.

pH 5 and no Al-SSA complex at pH 12.

1270 cm-1 indicates the presence of Ph-O-

calculation indicates the favourable formation of Al(OH)4-

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 115

The main impact would be on the vibration band of hydroxyl group which is deprotonated,

**3.3.2 Spectral characterization of 5-SSA in presence of aluminium ions or adsorbed** 

Spectra of solutions of 5-SSA in the absence and presence of aluminium ions have been recorded. The results are shown as dotted and grey lines in Figure 12. The main difference is the presence of bands at 1330 and 1270 cm-1 at pH 2 and pH 5. At pH 12, the pure 5-SSA presents the same band too, yet a difference between the spectra is rather observed in the shift of bands at 1308, 1428, 1470 cm-1 to 1326, 1431, 1480 cm-1 respectively. Speciation calculation suggests predominance of Al-SSA at pH 2, a mixture of Al-SSA and Al-SSA2 at

From the study of the effect of pH on pure SSA, it was concluded that the band visible at

presence of Ph-OH is absent in spectra at pH 5 in presence of aluminium ion, or less intense at pH 2. At both pH values, the spectra would indicate that the presence of aluminium promotes the deprotonation of the phenol group, which leads to a bond between 5-SSA and the aluminium ion. The band corresponding to carboxyl groups at 1370 cm-1 is not shifted in presence of aluminium. However, a hypothetical carboxyl-Al group could perturb also the intramolecular hydrogen bond, as suggested for salicylate-iron interaction, which makes difficult to use this method to evaluate the interaction with carboxyl groups. We may conclude that the affinity of carboxyl groups for aluminium ion is well known, and an interaction is expected, despite any direct experimental evidence. At pH 12, the speciation

complexes. However, the spectra evolves in presence of aluminium ions, mainly by the shift of bands at 1425 and 1468 cm-1 towards values close to the spectra of pure SSA at pH 2. This

As the last step, the spectra of 5-SSA in the presence of gibbsite have been recorded. The results are shown as black lines in Figure 12. In the 1500-1000 cm-1 range, no difference between spectra of pure 5-SSA solution and gibbsite-SSA was observed at pH 2 and 12. At pH 2, this behaviour is different from that of aluminium in solution. This might be consistent with an extent of 5-SSA adsorption that is too low to be detectable by the ATR

A comparison of the spectroscopic results to the zeta-potential measurements shown in Fig. 10 suggests the following picture. The differences in zeta potential of gibbsite particles with and without 5-SSA being present are substantial over the complete pH range investigated, except for pH 12. At pH 12, no adsorption is expected according to the electrokinetic results, but the spectra of 5-SSA in the presence of gibbsite do show a slight shift of 1425 cm-1 and 1468 cm-1 bands, similar to that observed with aluminium in solution. At pH 5, the spectra in presence of gibbsite are similar to that of 5-SSA in presence of aluminium ions, mainly due to the presence of a peak near 1270 cm-1. The effect of 5-SSA on zeta potential is strong, so it can be concluded that a complex between 5-SSA and aluminium atom occurs at the gibbsite surface at pH 5, similar to what has been observed with aluminium ion. In such a geometry (Fig. 13), the negative charge observed in the electrokinetic data would be consistent with

result remains unexplained but suggests an interaction between both species.

, while a band at 1294 cm-1 wich indicates the

and the absence of any Al-SSA

Fig. 11. Spectra of 5-SSA at pH 2 (black line), pH 5 (grey line), pH 12 (dotted line)


\* Varghese et al. (2007),

\*\* Jiang et al. (2002), sh: shoulder, c:carboxyl

Table 5. Assignments of bands in spectra of 5-SSA in solution at different pH values.

appear, while bands at 1370 cm-1, 1430 cm-1 increase. Around 1200 cm-1, bands corresponding to vibration of the sulfate group are expected, so this evolution would be in agreement with the change in the molecule from –SO3H to –SO3 - at pH 1.5. From pH 5 to 12, the shift of several bands is visible, the most important is 1292 cm-1 to 1306 cm-1. According to Jiang et al. (2002), this band corresponds to stretching of Ph-OH, which is deprotonated at pH > 11.3. Thus, the difference between these spectra would correspond to the change of speciation from pH 5, where only HSSA2- is present to pH 12, dominated by 83% of SSA3-.

pH 2 pH 5 pH 12

1218sh 1201sh 1198sh

1500 1400 1300 1200 1100 1000 Wavenumbers (cm-1)

Fig. 11. Spectra of 5-SSA at pH 2 (black line), pH 5 (grey line), pH 12 (dotted line)

Assignments from literature pH 2 pH 5 pH 12 νPh \* 1480 1479 1468 νPh,δCOHc \* 1439 1431 1425 νsCOc \*\* 1371 1373 1379 νasSO2 \* 1346-1330 1346-1330 1306 νPh-OH \*\* 1292 1294 1269

δCH \* 1159 1159 1151 νC-COOH \* 1082 1080 1084 νS-OH \* 1034 1030 1028

Table 5. Assignments of bands in spectra of 5-SSA in solution at different pH values.

appear, while bands at 1370 cm-1, 1430 cm-1 increase. Around 1200 cm-1, bands corresponding to vibration of the sulfate group are expected, so this evolution would be in agreement with the change in the molecule from –SO3H to –SO3- at pH 1.5. From pH 5 to 12, the shift of several bands is visible, the most important is 1292 cm-1 to 1306 cm-1. According to Jiang et al. (2002), this band corresponds to stretching of Ph-OH, which is deprotonated at pH > 11.3. Thus, the difference between these spectra would correspond to the change of speciation from pH 5, where only HSSA2- is present to pH 12, dominated by 83% of SSA3-.

wavenumbers (cm-1)

**1270**

**1430 1370 1306**

Absorbance

νsSO2 \* 1242sh

\*\* Jiang et al. (2002), sh: shoulder, c:carboxyl

\* Varghese et al. (2007),

νC-OH, δSOH \* 1180 1175

The main impact would be on the vibration band of hydroxyl group which is deprotonated, but a consequence on vibration of other groups would be possible.

#### **3.3.2 Spectral characterization of 5-SSA in presence of aluminium ions or adsorbed onto gibbsite**

Spectra of solutions of 5-SSA in the absence and presence of aluminium ions have been recorded. The results are shown as dotted and grey lines in Figure 12. The main difference is the presence of bands at 1330 and 1270 cm-1 at pH 2 and pH 5. At pH 12, the pure 5-SSA presents the same band too, yet a difference between the spectra is rather observed in the shift of bands at 1308, 1428, 1470 cm-1 to 1326, 1431, 1480 cm-1 respectively. Speciation calculation suggests predominance of Al-SSA at pH 2, a mixture of Al-SSA and Al-SSA2 at pH 5 and no Al-SSA complex at pH 12.

From the study of the effect of pH on pure SSA, it was concluded that the band visible at 1270 cm-1 indicates the presence of Ph-O- , while a band at 1294 cm-1 wich indicates the presence of Ph-OH is absent in spectra at pH 5 in presence of aluminium ion, or less intense at pH 2. At both pH values, the spectra would indicate that the presence of aluminium promotes the deprotonation of the phenol group, which leads to a bond between 5-SSA and the aluminium ion. The band corresponding to carboxyl groups at 1370 cm-1 is not shifted in presence of aluminium. However, a hypothetical carboxyl-Al group could perturb also the intramolecular hydrogen bond, as suggested for salicylate-iron interaction, which makes difficult to use this method to evaluate the interaction with carboxyl groups. We may conclude that the affinity of carboxyl groups for aluminium ion is well known, and an interaction is expected, despite any direct experimental evidence. At pH 12, the speciation calculation indicates the favourable formation of Al(OH)4 and the absence of any Al-SSA complexes. However, the spectra evolves in presence of aluminium ions, mainly by the shift of bands at 1425 and 1468 cm-1 towards values close to the spectra of pure SSA at pH 2. This result remains unexplained but suggests an interaction between both species.

As the last step, the spectra of 5-SSA in the presence of gibbsite have been recorded. The results are shown as black lines in Figure 12. In the 1500-1000 cm-1 range, no difference between spectra of pure 5-SSA solution and gibbsite-SSA was observed at pH 2 and 12. At pH 2, this behaviour is different from that of aluminium in solution. This might be consistent with an extent of 5-SSA adsorption that is too low to be detectable by the ATR measurements.

A comparison of the spectroscopic results to the zeta-potential measurements shown in Fig. 10 suggests the following picture. The differences in zeta potential of gibbsite particles with and without 5-SSA being present are substantial over the complete pH range investigated, except for pH 12. At pH 12, no adsorption is expected according to the electrokinetic results, but the spectra of 5-SSA in the presence of gibbsite do show a slight shift of 1425 cm-1 and 1468 cm-1 bands, similar to that observed with aluminium in solution. At pH 5, the spectra in presence of gibbsite are similar to that of 5-SSA in presence of aluminium ions, mainly due to the presence of a peak near 1270 cm-1. The effect of 5-SSA on zeta potential is strong, so it can be concluded that a complex between 5-SSA and aluminium atom occurs at the gibbsite surface at pH 5, similar to what has been observed with aluminium ion. In such a geometry (Fig. 13), the negative charge observed in the electrokinetic data would be consistent with

Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of

in surface complexes.

higher pH.

highlighted here.

**4. Conclusions** 

adsorbed carboxylic acids.

negative charge of the gibbsite in the presence of 5-SSA.

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 117

<sup>O</sup> <sup>O</sup> <sup>O</sup>

Al

SO3-

Fig. 13. Structure of the Al-SSA complex. Al is either an ion in solution, or bound to surface

the free sulfonate group, deprotonated in the whole pH range. At very low pH, the zetapotential turns positive even in the presence of 5-SSA, which can be explained by the speciation of the organic molecule itself, which would on average be less negative than at

At pH 12, there might still be adsorption of 5-SSA, since the spectra of 5-SSA/gibbsite system is close to those of 5-SSA/Al(III). However, it would be difficult to detect by the electrokinetic method, since the gibbsite itself turns negative at pH > 11.3. The fact that the solution spectra of 5-SSA are different with or without Al(III) species indicates the possibility that there is some interaction. Again, the importance of spectroscopic studies in pinpointing interactions which are not detectable by macroscopic methods can be

In the first part of this chapter, the principles and experimental protocols of the use of ATR-IR to probe solid/solution interfaces have been described. A review of the literature has shown that ATR-IR can be useful to get information about the surface speciation of

The second part was devoted to the comprehensive study of the adsorption of 5-SSA onto gibbsite platelets, and it was suggested that this phenomenon occurs via a dominant tridentate surface complex involving the phenolic and carboxylic groups. The ring structure of this surface complex had been previously postulated by Jiang et al. (2002) based on a limited set of spectra. The fact that the sulphate group is not involved in co-ordination to the surface, and therefore oriented towards the solution side of the interface, agrees with the

The solution study strongly indicates that the present speciation schemes for the 5-SSAdissolved aluminium system are incomplete. The spectra show differences between the absence and presence of Al(III) in solution at high pH, where within the current speciation scheme no complexes are expected. At high pH in the 5-SSA-gibbsite-system, the spectroscopic data also suggest interaction, which is undetectable by the electrokinetic

method, since the gibbsite surface has an overall negative charge in that pH-range.

Fig. 12. Spectra of SSA at (A) pH 2, (B) pH 5, (C) pH 12, as a pure solution (dotted line), a solution in presence of aluminium ions (grey line), a solution after the deposition of a gibbsite layer (black line).

Fig. 13. Structure of the Al-SSA complex. Al is either an ion in solution, or bound to surface in surface complexes.

the free sulfonate group, deprotonated in the whole pH range. At very low pH, the zetapotential turns positive even in the presence of 5-SSA, which can be explained by the speciation of the organic molecule itself, which would on average be less negative than at higher pH.

At pH 12, there might still be adsorption of 5-SSA, since the spectra of 5-SSA/gibbsite system is close to those of 5-SSA/Al(III). However, it would be difficult to detect by the electrokinetic method, since the gibbsite itself turns negative at pH > 11.3. The fact that the solution spectra of 5-SSA are different with or without Al(III) species indicates the possibility that there is some interaction. Again, the importance of spectroscopic studies in pinpointing interactions which are not detectable by macroscopic methods can be highlighted here.

## **4. Conclusions**

116 Infrared Spectroscopy – Materials Science, Engineering and Technology

5-SSA with gibbsite 5-SSA with Al3+ 5-SSA

Absorbance

Absorbance

Absorbance

gibbsite layer (black line).

(A)

(B)

(C)

**1470**

**1428 1308**

1500 1400 1300 1200 1100 1000 Wavenumbers (cm-1)

> 5-SSA with Al3+ 5-SSA

5-SSA with gibbsite

1500 1400 1300 1200 1100 1000 Wavenumbers (cm-1)

> 5-SSA 5-SSA with Al3+ 5-SSA with gibbsite

1500 1400 1300 1200 1100 1000 Wavenumbers (cm-1)

Fig. 12. Spectra of SSA at (A) pH 2, (B) pH 5, (C) pH 12, as a pure solution (dotted line), a solution in presence of aluminium ions (grey line), a solution after the deposition of a

**1330 1270**

**1270**

**1330**

In the first part of this chapter, the principles and experimental protocols of the use of ATR-IR to probe solid/solution interfaces have been described. A review of the literature has shown that ATR-IR can be useful to get information about the surface speciation of adsorbed carboxylic acids.

The second part was devoted to the comprehensive study of the adsorption of 5-SSA onto gibbsite platelets, and it was suggested that this phenomenon occurs via a dominant tridentate surface complex involving the phenolic and carboxylic groups. The ring structure of this surface complex had been previously postulated by Jiang et al. (2002) based on a limited set of spectra. The fact that the sulphate group is not involved in co-ordination to the surface, and therefore oriented towards the solution side of the interface, agrees with the negative charge of the gibbsite in the presence of 5-SSA.

The solution study strongly indicates that the present speciation schemes for the 5-SSAdissolved aluminium system are incomplete. The spectra show differences between the absence and presence of Al(III) in solution at high pH, where within the current speciation scheme no complexes are expected. At high pH in the 5-SSA-gibbsite-system, the spectroscopic data also suggest interaction, which is undetectable by the electrokinetic method, since the gibbsite surface has an overall negative charge in that pH-range.

Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of

*Miner. Engineer.* Vol. 21, 1000-1004, 0892-6875.

*Colloid Interface Sci.* Vol. 348, pp. 167-176, 0021-9797.

4301, 0016-7037.

pp. 6087-6092.

0021-9797.

9797.

7757.

0016-7037.

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 119

Duckworth, O.W. , & Martin, S.T. (2001), Surface complexation and dissolution of hematite

Fredriksson, A., & Holmgren, A. (2008). An in situ ATR-FTIR investigation of adsorption

Gan, Y., & Franks, G.V. (2006). Charging Behavior of the Gibbsite Basal (001) Surface in

Gao, X., & Chorover, J. (2010). Adsorption of sodium dodecyl sulfate (SDS) at ZnSe and α-

Garcia Rodenas, L. A., Iglesias, A. M., Weisz, A. D., Morando, P. J., & Blesa, M. A. (1997).

Hwang, Y. S., Liu, J, Lenhart, J. J. , & Hadad, C. M. (2007). Surface complexes of phthalic

Hug, S.J., & Sulzberger, B. (1994). In situ Fourier transform infrared spectroscopic evidence

Hug, S. H. (1997). In situ Fourier transform infrared measurements of sulfate adsorption on

Jiang, L. , Gao, L. , & Liu, Y. (2002). Adsorption of salicylic acid, 5-sulfosalicylic acid and

Johnson, S. B. , Yoon, T. H. , Slowey, A. J. , & Brown, G. E. (2004). Adsorption of organic

Johnson, S. B. , Yoon, Kocar, B.D. , & Brown, G.E. (2004). Adsorption of organic matter at

Kallay, N., Preočanin, T., Kovačević, D., Lützenkirchen, J., & Villalobos, M. (2011),

Klug, O., & Forsling, W. (1999). A spectroscopic study of phthalate adsorption on γ-

Kubicki, J.D., Schroeter, L.M., Itoh, M.J., Nguyen, B.N., & Apitz, S.E. (1999). Attenuated

adsorption of oxalate, *Langmuir* Vol. 20, 11480-11492, 0743-7463.

for dissolution processes, *Langmuir* Vol. 20, 4996-5006 , 0743-7463.

aluminum oxide, *Langmuir* Vol. 15, 6961-6968, 0743-7463.

*Acta* Vol. 84, 1, pp. 1-10, 0011-1643.

oxides by oxalic acid, *Inorg. Chem*. Vol. 36, 6423-6430, 0020-1669.

aqueous phase, *Langmuir* Vol. 10, pp. 3587-3597, 0743-7463.

by C1-C6 dicarboxylic acids at pH = 5.0, *Geochimi. Cosmochim. Acta* Vol. 65, 4289-

and orientation of heptyl xanthate at the lead sulphide/aqueous solution interface.

NaCl Solution Investigated by AFM Colloidal Probe Technique. *Langmuir* Vol. 22,

Fe2O3 surfaces: Combining infrared spectroscopy and batch uptake studies. *J.* 

Surface complexation description of the dissolution of chromium(III) hydrous

acid at the hematite/water interface, *J. Colloid Interface Sci.* Vol. 307, 124–134,

for the formation of several different surface complexes of oxalate on TiO2 in the

hematite in aqueous solutions. *J. Colloid Interface Sci.* Vol. 188, pp. 415-422, 0021-

Tiron at the alumina–water interface, *Colloids Surf. A.* Vol. 211, 165-172, 0927-

matter at mineral/water interfaces: 3. Implications of surface dissolution for

mineral/water interfaces. 2. outer-sphere adsorption of maleate and implications

Thermodynamics of the Reactions at Solid/Liquid Interfaces, *Croat. Chem.* 

total reflectance Fourier-transform infrared spectroscopy of carboxylic acids adsorbed onto mineral surfaces, *Geochim. Cosmochim. Acta* Vol. 63, 2709-2725,

A general conclusion of our experimental study would be that comprehensive studies of solid-solution interfaces via ATR-IR may contribute to solving the structure of surface complexes. Furthermore, such studies may help finding previously unidentified species in solution speciation schemes. Overall the interplay of surface complexation and solution complexation in this system indicates that the adsorption process may be very complex and that multi-method approaches are best suited to gain deeper understanding.

#### **5. References**


A general conclusion of our experimental study would be that comprehensive studies of solid-solution interfaces via ATR-IR may contribute to solving the structure of surface complexes. Furthermore, such studies may help finding previously unidentified species in solution speciation schemes. Overall the interplay of surface complexation and solution complexation in this system indicates that the adsorption process may be very complex and

Adekola, F., Fédoroff, M., Geckeis, H., Kupcik, T., Lefèvre, G., Lützenkirchen, J., Plaschke,

Asay, D. B. , & Kim, S.H. (2005). Evolution of the Adsorbed Water Layer Structure on

Axe, K., & Persson, P. (2001). Time-dependent surface speciation of oxalate at the water-

Boily, J.-F., Persson, P., & Sjöberg, S. (2000). Benzenecarboxylate surface complexation at

Borda, M. J. , Strongin, D. R., & Schoonen, M. A. (2003). A novel vertical attenuated total

Coates, J.P. (1993). The Industrial Applications of Infared Internal Reflection Spectroscopy*,* 

Couzist, A., & Gulari E. (1993). Adsorption of sodium laurate from its aqueous solution onto

Degenhardt, J., & McQuillan, A.J. (1999) Mechanism of oxalate ion adsorption on chromium

Dobson, K. D., & McQuillan, A. J. (1999). In situ infrared spectroscopic analysis of the

aqueous solutions, *Spectrochim. Acta A* Vol. 55, 1395-1405, 0584-8539. Dolamic, I., & Bürgi, T. (2006). Photoassisted decomposition of malonic acid on TiO2 studied

spectroscopy, *Spectrochim. Acta A* Vol. 59, 1103-1106, 0584-8539.

Dean, J.A. (1999) Lange's Handbook of Chemistry (15th ed.), McGraw-Hill.

M., Preocanin, T., Rabung, T., & Schild, D. (2011). Characterization of acid-base properties of two gibbsite samples in the context of literature results, *J. Colloid* 

Silicon Oxide at Room Temperature, *J. Phys. Chem. B* Vol. 109, 16760-16763, 1089-

boehmite (γ-AlOOH) interface: implications for dissolution, *Geochim. Cosmochim.* 

the goethite (α-FeOOH)/water interface: II. Linking IR spectroscopic observations to mechanistic surface complexation models for phthalate, trimellitate, and pyromellitate, *Geochim. Cosmochim. Acta* Vol. 64, No. 20, pp.

reectance photochemical ow-through reaction cell for Fourier transform infrared

*In: Internal Reflection Spectroscopy,* Dekker, New York, Mirabella, F. M. (Ed.), pp.53-

an alumina surface. A dynamic study of the surface-surfactant interaction using attenuated total reflection Fourier transform infrared spectroscopy, *Langmuir* Vol.

oxide-hydroxide from pH dependance and time evolution of ATR-IR spectra, *Chem.* 

adsorption of aliphatic carboxylic acids to TiO2, ZrO2, Al2O3, and Ta2O5 from

by in situ attenuated total reflection infrared spectroscopy, *J. Phys. Chem. B* Vol. 110,

that multi-method approaches are best suited to gain deeper understanding.

*Interface Sci.* Vol. 354, 306-317, 0021-9797.

*Acta* Vol. 65, 4481-4492, 0016-7037.

3453-3470, 0016-7037.

9, 3414-3421, 0743-7463.

14898-14904, 1089-5647.

*Phys. Lett.* Vol. 31, 179-184, 0009-2614.

**5. References** 

5647.

96.


Attenuated Total Reflection – Infrared Spectroscopy Applied to the Study of

*Chem*. Vol. 12, 655-663, 0020-1669.

602-611, 0743-7463.

*Langmuir* Vol. 2, pp. 203-210, 0743-7463.

*Langmuir* Vol. 6, 979-987, 0743-7463.

5647.

7463.

9797.

0920-5861.

495, 0743-7463.

Mineral – Aqueous Electrolyte Solution Interfaces: A General Overview and a Case Study 121

Scott, K.L., Wieghardt, K., & Sykes, A.G. (1973). Mu-oxalato-cobalt(III) complexes, *Inorg.* 

Tejedor -Tejedor, M.I., & Anderson, M.A. (1986). "In situ" attenuated total reflection Fourier

Tejedor -Tejedor, M.I., & Anderson, M.A. (1990). Protonation of phosphate on the surface of

Tejedor-Tejedor, M.I., Yost, E.C., & Anderson, M.A. (1990), Characterization of benzoic and

Tickanen, L.D., Tejedor-Tejedor, M.I., & Anderson, M.A. (1991). Quantitative

Tunesi, S., & Anderson, M.A. (1992). Surface effects in photochemistry: an in situ cylindrical

Varghese, H.T., Panicker, C. Y., & Philip, D. (2007). IR, Raman and SERS spectra of 5 sulphosalicylic acid dihydrate, *J. Raman Spectr.* Vol. 38, 309-315, 1097-4555. Venyaminov, S.Y., & Prendergast, F.G. (1997). Water (H2O and D2O) molar absorptivity in

Villalobos, M., & Leckie, J. O. (2001). Surface complexation modeling and FTIR study of

Wang, Z., Grahn, M., Larsson, M.L., Holmgren, A., Sterte, J., & Hedlund, J. (2006). Zeolite coated ATR crystal probes, *Sensors Actuators B*. Vol. 115, 685-690, 0925-4005. Weisz, A.D., Regazzoni, A.E., & Blesa, M.A. (2001). ATR-FTIR study of the stability trends of

Weisz, A.D., Garcia Rodenas, L., Morando, P.J., Regazzoni, A.E., & Blesa, M.A. (2002).

Yoon, T. H., Johnson, S. B., Musgrave, C. B., & Brown, G. E (2004). Adsorption of organic

immersed in water, *Solid State Ionics* Vol. 143, 125-130, 0167-2738.

solutions. *Anal. Biochem.* Vol. 248, 234-245, 0003-2697.

solution onto the TiO2 rutile (110) surface, *J. Phys. Chem. B* Vol. 108, 5004-5017, 1089-

transform infrared studies at the goethite (a-FeOOH)-aqueous solution interface.

goethite as studied by CIR-FTIR and electrophoretic mobility. *Langmuir* Vol. 6, pp.

phenolic complexes at the goethite/aqueous solution interface using cylindrical internal reflection Fourier transform infrared spectroscopy. Part 1. Methodology.

characterization of aqueous suspensions using attenuated total reflection Fourier transform infrared spectroscopy: influence of internal reflection element-particle interactions on spectral absorbance values. *Langmuir* Vol. 7, pp. 451-456, 0743-

internal reflection-Fourier transform infrared investigation of the effect of ring substituents on chemisorption onto TiO2 ceramic membranes, *Langmuir* Vol. 8, 487-

the 1000–4000 cm-1 range and quantitative infrared spectroscopy of aqueous

carbonate adsorption to goethite. *J. Colloid Interface Sci.* Vol. 235, pp. 15-32, 0021-

carboxylate complexes formed on the surface of titanium dioxide particles

FTIR study of the adorption of single pollutants and mixtures of pollutants onto titanium dioxide in water: oxalic and salycylic acids, *Catal. Today* Vol. 76, 103-112,

matter at mineral/water interfaces: I. ATR-FTIR spectroscopic and quantum chemical study of oxalate adsorbed at boehmite/water and corundum/water interfaces, *Geochim. Cosmochim. Acta* Vol. 68, No. 22, pp. 4505-4518, 0016-7037.


Kulik, D. A., Luetzenkirchen, J., & Payne, T. (2010*).* Consistent treatment of 'denticity' in

Larsson, M.L., Fredriksson, A., & Holmgren, A. (2004). Direct observation of a self-

polarized FTIR study. *J. Colloid Interface Sci.* Vol.273, 345–349, 0021-9797. Lefèvre, G., Noinville, S., & Fédoroff, M. (2006). Study of uranyl sorption onto hematite by

Lefèvre, G., Kneppers, J., & Fédoroff, M. (2008). Sorption of uranyl ions on titanium oxide studied by ATR-IR spectroscopy, *J. Colloid Interface Sci*. 327, 15-20, 0021-9797. Lide, D.R. (1998). Handbook of Chemistry and Physics 79th.; Lide, D.R., Ed.; CRC Press:

Huittinen, N., Rabung, Th., Lützenkirchen, J., Mitchell, S.C., Bickmore, B.R., Lehto, J., &

batch and TRLFS study, *J. Colloid Interface Sci.* Vol. 332, 158-164, 0021-9797. Lützenkirchen, J., Kupcik, T., Fuss, M., Walther, C., Sarpola, A., & Sundman, O. (2010).

Mirabella, F.M. (1993). Principles, Theory, and Practice of Internal Reflection Spectroscopy*,* 

Nordin, J. , Persson, P., Laiti, E., & Sjöberg, S. (1997). Adsorption of o-phthalate at the water-

Panak, P. (1996). Untersuchung von intramolekularen Energietransferprozessen in Cm(III)-

Peak, D., Ford, R.G., & Sparks, D.L. (1999). An in-situ FTIR-ATR investigation of sulfate

Persson, P., & Axe, K. (2005). Adsorption of oxalate and malonate at the water-goethite

Rosenqvist, J., Persson, P., & Sjöberg, S. (2002). Protonation and Charging of Nanosized

Rosenqvist, J., Axe, K., Sjöberg, S., & Persson P. (2003). Adsorption of dicarboxylates on

Rotzinger, F.P., Kesselman-Truttmann, J.M., Hug, S.J., Shklover, V. , & Grätzel, M. (2004).

0016-7037.

Boca Raton.

5411, 0169-4332.

13, 4085-4093, 0743-7463.

*Acta* Vol. 69, No. 3, pp. 541–552, 0016-7037.

*Colloids Surf. A* Vol. 220, 91-104, 0927-7757.

52.

p.

9797.

4604.

296, 608-613, 0021-9797.

surface complexation models. *Geochim. Cosmochim Acta* Vol. 74, 12 , pp. A544-A544,

assembled monolayer of heptyl xanthate at the germanium/water interface: a

in situ attenuated total reflection - infrared spectroscopy, *J. Colloid Interface Sci.* Vol.

Geckeis, H. (2009). Sorption of Cm(III) and Gd(III) onto gibbsite, alpha-Al(OH)3: A

Adsorption of Al(13)-Keggin clusters to sapphire c-plane single crystals: Kinetic observations by streaming current measurements, *Appl. Surf. Sci.* Vol. 256, 5406-

*In: Internal Reflection Spectroscopy,* Dekker, New York, Mirabella, F. M. (Ed.), pp.17-

boehmite (γ-AlOOH) interface: evidence for two coordination modes, *Langmuir* Vol.

und Tb(III) Komplexen mit organischen Liganden mit Hilfe der zeitaufgelösten Laserfluoreszencspektroskopie. Ph.D. Thesis, Technische Universität München, 236

bonding mechanisms on goethite. *J. Colloid Interface Sci*.Vol. 218, pp. 289-299, 0021-

interface: molecular surface speciation from IR spectroscopy, *Geochim. Cosmochim.* 

Gibbsite (α-Al(OH)3) Particles in Aqueous Suspension. *Langmuir* Vol. *18, pp.* 4598-

nano-sized gibbsite particles effects of ligand structure on bonding mechanisms,

Structure and vibrational spectrum of formate and acetate adsorbed from aqueous

solution onto the TiO2 rutile (110) surface, *J. Phys. Chem. B* Vol. 108, 5004-5017, 1089- 5647.


**6** 

*Latvia* 

**Research of Calcium Phosphates Using** 

Liga Berzina-Cimdina and Natalija Borodajenko

*Riga Technical University,* 

*Institute of General Chemical Engineering* 

**Fourier Transform Infrared Spectroscopy** 

In the biomaterial research field, nowadays a great attention is driven onto calcium phosphates synthesis and obtaining of ceramics that can be used in orthopedics and dentistry, in the form of coatings, granules, porous or solid blocks, as well as in the form of various composite materials. The most frequently studied, clinically tested and used synthetic materials based on calcium phosphate (CaP) are hydroxyapatite [HAp - Ca10(PO4)6(OH)2], β-tricalcium phosphate [β-TCP - Ca3(PO4)2] and biphasic HAp/β-TCP mixture. CaP ceramic demonstrates high biocompatibility and bioactivity while it contacts bone cells, builds a direct chemical connection between bone tissues and ceramic implant. As practice shows, purchased materials, most often commercial CaP materials, not always have properties and qualities defined by the manufacturer. Frequently, the manufacturer's information about the offered product is not complete or precise, by it troubling usage of raw CaP material for development of implants. The most common imperfections of CaP materials are – unpredictable properties after the high temperature treatment (composition and clarity of crystal phases, chemical composition, thermal stability, etc.) which become clear only after the high temperature treatment of the ready implant material has occurred. For several years, Riga Biomaterial Innovation and Development Centre (RBIDC) of Riga Technical University perform a wide range of

Properties of bioceramic implants obtained from various commercial and laboratory synthesized calcium phosphate precursors are different, since behavior of those precursors is different within the thermal treatment processes, which are a significant stage of obtaining

CaP synthesis methods and their technological parameters can significantly impact stoichiometry of the synthesis product, its grade of crystallization, particle size, bioceramic phase composition, thermal stability, microstructure and mechanical properties. The important technologic parameters that impact properties of calcium phosphate synthesis product and then also of bioceramic, are temperature of synthesis, pH of synthesis environment, reagent type and concentration, as well as selection of raw materials, their purity and quality. All of the above mentioned also brings a significant impact on the tissue

property studies of various commercial CaP raw materials.

response of these bioceramic implants.

**1. Introduction** 

ceramics.

Yost, E.C., Tejedor-Tejedor, M.I., & Anderson, M.A. (1990). In situ CIR-FTIR characterization of salicylate complexes at the goethite/aqueous solution interface, *Environ. Sci. Technol.* Vol. 24, pp. 822-828, 0013-936X.

## **Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy**

Liga Berzina-Cimdina and Natalija Borodajenko *Riga Technical University, Institute of General Chemical Engineering Latvia* 

### **1. Introduction**

122 Infrared Spectroscopy – Materials Science, Engineering and Technology

Yost, E.C., Tejedor-Tejedor, M.I., & Anderson, M.A. (1990). In situ CIR-FTIR characterization

*Technol.* Vol. 24, pp. 822-828, 0013-936X.

of salicylate complexes at the goethite/aqueous solution interface, *Environ. Sci.* 

In the biomaterial research field, nowadays a great attention is driven onto calcium phosphates synthesis and obtaining of ceramics that can be used in orthopedics and dentistry, in the form of coatings, granules, porous or solid blocks, as well as in the form of various composite materials. The most frequently studied, clinically tested and used synthetic materials based on calcium phosphate (CaP) are hydroxyapatite [HAp - Ca10(PO4)6(OH)2], β-tricalcium phosphate [β-TCP - Ca3(PO4)2] and biphasic HAp/β-TCP mixture. CaP ceramic demonstrates high biocompatibility and bioactivity while it contacts bone cells, builds a direct chemical connection between bone tissues and ceramic implant. As practice shows, purchased materials, most often commercial CaP materials, not always have properties and qualities defined by the manufacturer. Frequently, the manufacturer's information about the offered product is not complete or precise, by it troubling usage of raw CaP material for development of implants. The most common imperfections of CaP materials are – unpredictable properties after the high temperature treatment (composition and clarity of crystal phases, chemical composition, thermal stability, etc.) which become clear only after the high temperature treatment of the ready implant material has occurred. For several years, Riga Biomaterial Innovation and Development Centre (RBIDC) of Riga Technical University perform a wide range of property studies of various commercial CaP raw materials.

Properties of bioceramic implants obtained from various commercial and laboratory synthesized calcium phosphate precursors are different, since behavior of those precursors is different within the thermal treatment processes, which are a significant stage of obtaining ceramics.

CaP synthesis methods and their technological parameters can significantly impact stoichiometry of the synthesis product, its grade of crystallization, particle size, bioceramic phase composition, thermal stability, microstructure and mechanical properties. The important technologic parameters that impact properties of calcium phosphate synthesis product and then also of bioceramic, are temperature of synthesis, pH of synthesis environment, reagent type and concentration, as well as selection of raw materials, their purity and quality. All of the above mentioned also brings a significant impact on the tissue response of these bioceramic implants.

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 125

determined by their Ca/P molar ratio. From the point of view of chemistry, they are formed by three main elements: calcium, phosphorus and oxygen. Many calcium phosphates also contain hydrogen in an acidic phosphate anion (for example, HPO42-), hydroxyl groups (for example, Ca10(PO4)6(OH)2) or in a form of bonded water (for example, CaHPO4·2H2O). Majority of compounds of this class are poorly soluble in the water and non-soluble in alkaline solutions, but all of them easily dissolve in acids. Chemically pure calcium orthophosphates are white crystals with an average hardness, while natural materials are always of some other color which depends on the type and amount of impurities. Biologic calcium phosphates are main mineral components in the calcified tissues of the vertebrates

Main components of the natural bone tissues are calcium phosphates which, along with the other elements (Na, K, F, and Cl) form ~ 70% of the bone tissue mass. Also, bone tissues contain water (10% of mass) and collagen along with the other organic materials in small amounts. In living system CaP are found in the form of crystalline hydroxyapatite (HAp)

As bone substitution materials, calcium orthophosphates are researched for more than 80 years. The most significant characteristics of calcium phosphates are their bioresorbtion and bioactivity. They are non-toxic and biocompatible. Bioactivity shows as an ability to create a physical chemical bond between an implant and a bone. This process is called

Depending on the calcium/phosphorus (Ca/P) molar ratio and solubility of the compound, it is possible to obtain numerous calcium phosphates of different composition. Molar Ca/P ratio and solubility are connected with the pH of the solution. Majority of materials of this class are resorbable and dissolve when inserted in a physical environment. Calcium phosphates that are most frequently used in the biomaterial field are demonstrated in Table

For biomedical application, the following calcium phosphates are most frequently used: HAp (Ca/P = 1.67) and β-TCP (Ca/P=1.5), as well as biphasic calcium phosphate which

Hydroxyapatite Ca10(PO4)6(OH)2 is dominating and the most significant mineral phase in the solid tissues of the vertebrates. It consists of the same ions that form mineral part of teeth

A biological HAp usually has a calcium deficient; it is always substituted with a carbonate.

(A-type substitution (CO3)2- ↔ 2OH-) and (2) necessity after charge compensation, PO43-

provoke characteristic changes in the lattice parameters, crystallinity, crystal symmetry, thermal stability, morphology, and solubility, physical, chemical and biological

The most characteristic chemical groups in the FTIR spectrum of synthesized HAp are PO43-,

with CO32-

2- (B-type substitution). Substitution groups may

Two types of carbonate substitution are possible: (1) direct substitution of OH-

OH-, CO32-, as well as HPO42- that characterize non-stoichiometric HAp.

(Dorozhkin, 2009a).

and bones.

and in the amorphous calcium phosphate (ACP) form.

ostheointegration (Dorozhkin, 2009b).

1 (Dorozhkin, 2009c; El Kady, 2009; Shi, 2006).

substituting a tetrahedral group with CO3

characteristics (Shi, 2006).

mainly consists of HAp and β-TCP mixture in various ratios.

**2.2 FTIR absorption bands of the synthesized HAp** 

Fourier transform infrared spectroscopy (FTIR) is one of the methods which, systematically monitoring variations of structural characteristic groups and vibrations bonds, can provide an indirect evaluation of the synthesized Ca/P implant materials from TCP up to HAp and bioceramics, obtained from these materials.

FTIR spectroscopy has numerous advantages when used for chemical analysis of CaP products. First of all, an obtained spectrogram provides useful information about location of peaks, their intensity, width and shape in the required wave number range. Secondly, FTIR is also a very sensitive technique for determining phase composition. In the third place, FTIR is a comparatively quick and easy everyday approach.

During recent years, many authors' attention is turned onto synthesis of CaP and research of structure of the synthesized products depending on their technological parameters, with various methods, including an X-ray diffraction (XRD) and FTIR methods. However, the data of the literary sources is often incomplete or sometimes even contradictory. Studying various literary sources and analyzing the taken spectra of laboratory synthesized and commercial CaP products was aimed onto creating summary IR spectrum tables for the characteristic calcium phosphate chemical groups absorption bands. HAp stoichiometry is very important if the material has undergone a high temperature treatment.

A minor misbalance of synthesis product in a stoichiometric ratio (standard molar ratio of Ca/P is 1.67) during high temperature treatment can lead to composition of -, -TCP, or other phases. Thermally treating the stoichiometric calcium phosphates, it is possible to obtain stable phases at temperatures up to 1300 oC. One of the main non-stoichiometry reasons is inclusion of impurities, often substitutions of Ca2+ or interpenetration of other ions in the crystal lattice. In total, biological calcium phosphates are defined as calcium hydroxyapatites with deficient of calcium, Ca10-x(PO4)6-x(HPO4)x(OH)2-x (0<x<2), including the substituting atoms or groups, as, for example, Mg2+, Na+, K+, Sr2+, or Ba2+ substitute Ca2+, CO32-, H2PO4-, HPO42-; SO42- substitute PO43-; F- , Cl- , CO32- , PO43- substitute OH-.

The main target of our work was to perform a FTIR spectroscopy analysis of the CaP products synthesized in RBIDC laboratory and make summarizing conclusions about chemical groups of calcium phosphates, their variations under impact of synthesis parameters and further thermal treatment, as well as creating summary tables for:


## **2. Calcium phosphates and FTIR absorption bands of their chemical groups**

#### **2.1 Calcium phosphates**

Calcium phosphates as chemical compounds arise interest of the numerous fields of science, like geology, chemistry, biology and medicine. Many forms of calcium phosphates are

Fourier transform infrared spectroscopy (FTIR) is one of the methods which, systematically monitoring variations of structural characteristic groups and vibrations bonds, can provide an indirect evaluation of the synthesized Ca/P implant materials from TCP up to HAp and

FTIR spectroscopy has numerous advantages when used for chemical analysis of CaP products. First of all, an obtained spectrogram provides useful information about location of peaks, their intensity, width and shape in the required wave number range. Secondly, FTIR is also a very sensitive technique for determining phase composition. In the third place,

During recent years, many authors' attention is turned onto synthesis of CaP and research of structure of the synthesized products depending on their technological parameters, with various methods, including an X-ray diffraction (XRD) and FTIR methods. However, the data of the literary sources is often incomplete or sometimes even contradictory. Studying various literary sources and analyzing the taken spectra of laboratory synthesized and commercial CaP products was aimed onto creating summary IR spectrum tables for the characteristic calcium phosphate chemical groups absorption bands. HAp stoichiometry is very important if the material has undergone a high

A minor misbalance of synthesis product in a stoichiometric ratio (standard molar ratio of Ca/P is 1.67) during high temperature treatment can lead to composition of -, -TCP, or other phases. Thermally treating the stoichiometric calcium phosphates, it is possible to obtain stable phases at temperatures up to 1300 oC. One of the main non-stoichiometry reasons is inclusion of impurities, often substitutions of Ca2+ or interpenetration of other ions in the crystal lattice. In total, biological calcium phosphates are defined as calcium hydroxyapatites with deficient of calcium, Ca10-x(PO4)6-x(HPO4)x(OH)2-x (0<x<2), including the substituting atoms or groups, as, for example, Mg2+, Na+, K+, Sr2+, or Ba2+ substitute Ca2+, CO32-, H2PO4-, HPO42-; SO42- substitute PO43-; F- , Cl- , CO32- , PO43- substitute OH-.

The main target of our work was to perform a FTIR spectroscopy analysis of the CaP products synthesized in RBIDC laboratory and make summarizing conclusions about chemical groups of calcium phosphates, their variations under impact of synthesis

 CaP powders synthesized in the laboratory with a chemical solution precipitation method with different synthesis parameters (temperature, final suspension pH, maturation time) as-synthesized and then thermally treated at various temperatures

CaO containing materials of various origins (marble, eggshells, land snail shells) and

**2. Calcium phosphates and FTIR absorption bands of their chemical groups** 

Calcium phosphates as chemical compounds arise interest of the numerous fields of science, like geology, chemistry, biology and medicine. Many forms of calcium phosphates are

parameters and further thermal treatment, as well as creating summary tables for:

bioceramics, obtained from these materials.

temperature treatment.

from 200oC up to 1400oC; Commercial CaP products;

**2.1 Calcium phosphates** 

FTIR spectra of obtained products;

FTIR is a comparatively quick and easy everyday approach.

determined by their Ca/P molar ratio. From the point of view of chemistry, they are formed by three main elements: calcium, phosphorus and oxygen. Many calcium phosphates also contain hydrogen in an acidic phosphate anion (for example, HPO42-), hydroxyl groups (for example, Ca10(PO4)6(OH)2) or in a form of bonded water (for example, CaHPO4·2H2O). Majority of compounds of this class are poorly soluble in the water and non-soluble in alkaline solutions, but all of them easily dissolve in acids. Chemically pure calcium orthophosphates are white crystals with an average hardness, while natural materials are always of some other color which depends on the type and amount of impurities. Biologic calcium phosphates are main mineral components in the calcified tissues of the vertebrates (Dorozhkin, 2009a).

Main components of the natural bone tissues are calcium phosphates which, along with the other elements (Na, K, F, and Cl) form ~ 70% of the bone tissue mass. Also, bone tissues contain water (10% of mass) and collagen along with the other organic materials in small amounts. In living system CaP are found in the form of crystalline hydroxyapatite (HAp) and in the amorphous calcium phosphate (ACP) form.

As bone substitution materials, calcium orthophosphates are researched for more than 80 years. The most significant characteristics of calcium phosphates are their bioresorbtion and bioactivity. They are non-toxic and biocompatible. Bioactivity shows as an ability to create a physical chemical bond between an implant and a bone. This process is called ostheointegration (Dorozhkin, 2009b).

Depending on the calcium/phosphorus (Ca/P) molar ratio and solubility of the compound, it is possible to obtain numerous calcium phosphates of different composition. Molar Ca/P ratio and solubility are connected with the pH of the solution. Majority of materials of this class are resorbable and dissolve when inserted in a physical environment. Calcium phosphates that are most frequently used in the biomaterial field are demonstrated in Table 1 (Dorozhkin, 2009c; El Kady, 2009; Shi, 2006).

For biomedical application, the following calcium phosphates are most frequently used: HAp (Ca/P = 1.67) and β-TCP (Ca/P=1.5), as well as biphasic calcium phosphate which mainly consists of HAp and β-TCP mixture in various ratios.

#### **2.2 FTIR absorption bands of the synthesized HAp**

Hydroxyapatite Ca10(PO4)6(OH)2 is dominating and the most significant mineral phase in the solid tissues of the vertebrates. It consists of the same ions that form mineral part of teeth and bones.

A biological HAp usually has a calcium deficient; it is always substituted with a carbonate. Two types of carbonate substitution are possible: (1) direct substitution of OH with CO3 2- (A-type substitution (CO3)2- ↔ 2OH-) and (2) necessity after charge compensation, PO4 3 substituting a tetrahedral group with CO32- (B-type substitution). Substitution groups may provoke characteristic changes in the lattice parameters, crystallinity, crystal symmetry, thermal stability, morphology, and solubility, physical, chemical and biological characteristics (Shi, 2006).

The most characteristic chemical groups in the FTIR spectrum of synthesized HAp are PO43-, OH- , CO3 2-, as well as HPO42- that characterize non-stoichiometric HAp.

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 127

Substitutes phosphate ion, B-type HAp is formed (Meejoo, et al., 2006)

ions prove presence of HAp

Under influence of thermal treatment, absorption band becomes narrower

Characterizes HAp with deficient of calcium. (Raynaud, et al., 2002); Refers to non-stoichiometric HAp (Kwon, et al.,2003);

4 (Destainville, et al., 2003); bending mode (Han J-K., et al., 2006)

3 (Destainville, et al., 2003); bending mode (Han J-K., et al., 2006);

Synthesis residue that disappears during the calcifying process (Destainville, et al., 2003)

OH-

al., 2002); 2 (Destainville, et al., 2003);

al., 2002) 1 (Destainville, et al., 2003);

**groups Absorption bands, (cm-1) Description** 

873; 1450; 1640 (Meejoo, et al., 2006) 1650 (Raynaud, et al., 2002); 870 and 880; 1460 and 1530 (Ratner, 2004)

3500 (Meejoo, et al., 2006) 630 and 3540 (Destainville, et al., 2003), (Raynaud, et al., 2002); 3570 and 3420 (Han J-K., et al., 2006); 1650 (Raynaud, et al., 2002)

875 (Destainville, et al., 2003), (Raynaud, et al., 2002); 880 (Kwon, et al.,2003)

460 (Destainville, et al., 2003); (Raynaud, et

560 - 600 (Destainville, et al., 2003), (Raynaud, et al., 2002), (Mobasherpour & Heshajin, 2007); 602 un 555 (Han J-K., et al., 2006)

960 (Destainville, et al., 2003), (Raynaud, et

1020 -1120 (Destainville, et al., 2003), (Raynaud, et al., 2002); 1040 (Han J-K., et al., 2006); 1000 - 1100 (Mobasherpour & Heshajin, 2007);

(Raynaud, et al., 2002)

Table 2. FTIR absorption bands of synthesized HAp chemical groups.

**2.3 FTIR absorption bands of thermally treated calcium phosphates** 

As a biomaterial, HAp is mainly used in its ceramic form that was obtained by sintering the powder at 1000 – 1350 °C or as a coating on the implant surface. During the process of thermal decomposition of HAp, sintering of ceramic or obtaining the coating, physical, chemical, mechanical and, most important, biomedical properties may be negatively affected. Thus, HAp and other CaP materials should be thoroughly studied while thermally

NO3- 820 and 1380 (Destainville, et al., 2003);

water 2600 – 3600 (Meejoo, et al., 2006)

**Chemical** 

CO32-

OH-

Adsorbed

HPO42-

PO43-

treated.


Table 1. Calcium phosphates used in the biomaterial field.

Fig. 1. A typical FTIR spectrum of hydroxyapatite (Ratner, 2004)

PO43- group forms intensive IR absorption bands at 560 and 600 cm-1 and at 1000 – 1100 cm-1. Adsorbed water band is relatively wide, from 3600 to 2600 cm-1, with an explicit peak at 3570 cm-1, a weaker peak is formed at 630 cm-1. CO32- group forms weak peaks between 870 and 880 cm-1 and more intensive peaks between 1460 and 1530 cm-1. Absorption bands of chemical bonds of the synthesized HAp spectrum are summarized in Table 2.

ACP CaxHy(PO4)z .

DCPD CaHPO4 .

Octacalcium phosphate OCP Ca8(HPO4)2(PO4)4 . 5H2O 1.33 β-tricalcium phosphate β-TCP Ca3(PO4)2 1.50 α-tricalcium phosphate α-TCP Ca3(PO4)2 1.50

Hydroxyapatite HAp Ca10(PO4)6(OH)2 1.67 Tetra calcium phosphate TTCP (TetCP) Ca4(PO4)2O 2.00 β-Ca pyrophosphate CPP Ca2P2O7 <1.5 Oxyapatite OAp Ca10(PO4)6O 1.67

Table 1. Calcium phosphates used in the biomaterial field.

Fig. 1. A typical FTIR spectrum of hydroxyapatite (Ratner, 2004)

PO43- group forms intensive IR absorption bands at 560 and 600 cm-1 and at 1000 – 1100 cm-1. Adsorbed water band is relatively wide, from 3600 to 2600 cm-1, with an explicit peak at 3570 cm-1, a weaker peak is formed at 630 cm-1. CO32- group forms weak peaks between 870 and 880 cm-1 and more intensive peaks between 1460 and 1530 cm-1. Absorption bands of

chemical bonds of the synthesized HAp spectrum are summarized in Table 2.

Amorphous calcium

Dicalcium phosphate

Dicalcium phosphate

Hydroxyapatite with calcium deficient

phosphate

anhydride

dehydrate

**Name Abbreviation Chemical formula Ca/P** 

CDHA Ca10-x(HPO4)x(PO4)6-x(OH)2-x 0 ≤ x ≤ 1

DCPA CaHPO4 1.00

nH2O 1.2-2.2

1.5-1.67

2H2O 1.00


Table 2. FTIR absorption bands of synthesized HAp chemical groups.

#### **2.3 FTIR absorption bands of thermally treated calcium phosphates**

As a biomaterial, HAp is mainly used in its ceramic form that was obtained by sintering the powder at 1000 – 1350 °C or as a coating on the implant surface. During the process of thermal decomposition of HAp, sintering of ceramic or obtaining the coating, physical, chemical, mechanical and, most important, biomedical properties may be negatively affected. Thus, HAp and other CaP materials should be thoroughly studied while thermally treated.

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 129

In Tables 3-5, the data obtained from literary sources about the FTIR absorption bands of

**bands, (cm-1) Description** 

Synthesis residue is taken away, but the initial TCP remains unchanged (Destainville, et al., 2003);

Agglutination begins (Destainville, et al., 2003);

Maximum speed of compaction, comparing with HAp, β-TCP sintering occurs at a lower temperature. (Destainville, et al., 2003);

Disappeared; spectrum is similar β-TCP (Kwon, et al., 2003)

Lack of P2O74- proves that there is no CPP phase and spectrum is similar to pure β-TCP (Destainville, et al., 2003);

**Absorption** 

630 (Destainville, et al., 2003);

727 and 1200 (Destainville, et al., 2003)

Table 4. FTIR absorption bands of thermally treated TCP chemical groups.

be considered that CO2 is discharged from the sample between 450-950°C.

1200 β-TCPTCP (Destainville, et al., 2003);

Losing the adsorbed water do not impact lattice parameters. The water adsorbed on the surface discharges under temperature of less than 250°C, when the moisture is discharged from pores up to 500°C. With temperature rising, wide water bands at 3540 cm-1 become narrower and gradually disappear, but the sharp narrow peaks at 630 and 3570 cm-1 refer to

on the synthesis condition, a carbonate containing apatite is often obtained. Then, it should

Thermal stability is characterized by the decomposition temperature of HAp sample. The decomposition occurs when a critical dehydration point is achieved. In the temperatures less than the critical point, crystal structure of HAp remains unchanged in spite of the stage of dehydration. Achieving the critical point, a complete and irreversible dehydroxillation occurs, which results damage of HAp structure, decomposing onto tricalcium phosphate (β-TCP under 1200 °C and α-TCP in higher temperatures) and

At 900oC, β-TCP shoulders begin to show up at 947, 974 and 1120 cm-1, but during heating at higher temperature as, for example, 1200oC, β-TCP phase becomes more visible and PO43 peaks shift from 603 and 565 cm-1 to 601 and 571 cm-1, also from 1094 and 1032 to 1090 and

groups, which is characteristic to structure of HAp. Depending

thermally treated CaP chemical groups, is summarized.

**Chemical groups and phases** 

TCP β-TCP (Destainville, et al., 2003);

650 TCP;

OH-

P2O74-

variations of structural OH-

tetracalcium phosphate (TTCP).

1046 cm-1.

**Temperature, oC** 

750

950 - 1000


Table 3. FTIR absorption bands of thermally treated CaP chemical bonds.

During the thermal treatment, behavior of CaP is affected by various factors, like, atmosphere of sintering, ratio of Ca/P, method and conditions of powder synthesis, type and amount of impurities, sample size, particle size, etc.

During thermal treating HAp undergoes the following processes:



<sup>β</sup>- TCP 947, 974 and 1120(Meejoo,

3500 (Meejoo, et al., 2006) 630 and 3540 (Destainville, et al., 2003);

et al., 2006)

603 and 565; 1094 and 1032 (Meejoo, et al., 2006)

601 and 571; 1090 and 1046 (Meejoo, et al., 2006)

551; 585; 597; 613; 984; 1025; 1055 (Han J-K., et al., 2006)

Table 3. FTIR absorption bands of thermally treated CaP chemical bonds.

and amount of impurities, sample size, particle size, etc.

HAp decomposition with formation of other phases.

dehydration (separation of adsorbed water);

During thermal treating HAp undergoes the following processes:

800 CO32- 1450 (Meejoo, et al., 2006) Disappears (Meejoo, et al., 2006) 900 OH- Disappears (Meejoo, et al., 2006)

<sup>1200</sup>β- TCP Can see the characteristic peaks

1200 - 1400 β- TCP transforms onto -TCP

During the thermal treatment, behavior of CaP is affected by various factors, like, atmosphere of sintering, ratio of Ca/P, method and conditions of powder synthesis, type

dehydroxylation (separation of structured water), forming oxy-hydroxyapatite (OHAp)

**Absorption bands, (cm-1) Description** 

2- 1450 (Meejoo, et al., 2006) Intensity decreases;

Molecules of adsorbed water disappear (Mobasherpour & Heshajin, 2007);

Synthesis impurities disappear (Raynaud, et al., 2002);

Adsorbed water band becomes narrower (Meejoo, et al., 2006); Refers to variations of OH- (Destainville, et al., 2003);

β-TCP shoulders begin to show up (Meejoo, et al., 2006);

Shifts position at 1200 0C (Meejoo, et al., 2006);

better (Ratner, 2004);

Indicates that under influence of temperature, phosphates decompose and β-TCP shoulders become wider (Meejoo, et al., 2006).

(Mobasherpour & Heshajin, 2007);

**Temperature, oC** 

250 H2O

600 NO3

700 CO3

PO4

PO43-

1400 -TCP

and oxyapatite (OAp);

**Chemical groups and phases** 

> H2O, OH-

> > 3-


In Tables 3-5, the data obtained from literary sources about the FTIR absorption bands of thermally treated CaP chemical groups, is summarized.

Table 4. FTIR absorption bands of thermally treated TCP chemical groups.

Losing the adsorbed water do not impact lattice parameters. The water adsorbed on the surface discharges under temperature of less than 250°C, when the moisture is discharged from pores up to 500°C. With temperature rising, wide water bands at 3540 cm-1 become narrower and gradually disappear, but the sharp narrow peaks at 630 and 3570 cm-1 refer to variations of structural OH groups, which is characteristic to structure of HAp. Depending on the synthesis condition, a carbonate containing apatite is often obtained. Then, it should be considered that CO2 is discharged from the sample between 450-950°C.

Thermal stability is characterized by the decomposition temperature of HAp sample. The decomposition occurs when a critical dehydration point is achieved. In the temperatures less than the critical point, crystal structure of HAp remains unchanged in spite of the stage of dehydration. Achieving the critical point, a complete and irreversible dehydroxillation occurs, which results damage of HAp structure, decomposing onto tricalcium phosphate (β-TCP under 1200 °C and α-TCP in higher temperatures) and tetracalcium phosphate (TTCP).

At 900oC, β-TCP shoulders begin to show up at 947, 974 and 1120 cm-1, but during heating at higher temperature as, for example, 1200oC, β-TCP phase becomes more visible and PO4 3 peaks shift from 603 and 565 cm-1 to 601 and 571 cm-1, also from 1094 and 1032 to 1090 and 1046 cm-1.

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 131

In order to control speed of biodegradation, a biphasic calcium phosphate (BCP) bioceramic is developed, containing both HAp and TCP. By variation HAp/TCP ratio, it is possible to control bioactivity and biodegradation of implant. Since β-TCP is more soluble and HAp allows a biological precipitation of apatites, solubility of BCP depends on the ratio of HAp/TCP. Osteoconductivity among BCP, HAp and TCP does not

BCP is formed, if 1.500 < Ca/P < 1.667, which refers to a hydroxyapatite with calcium deficient, its chemical formula is Ca10-x(PO4)6-x (HPO4)x(OH)2-x (0< x <2). Ca/P ratio of synthesis sedimentary is not directly connected with the initial Ca/P ratio. At the constant pH, molar ratio of calcium phosphates may be varied by changing the temperature during

In order to achieve higher assay of the obtained material, its thermal stability and predictability of other properties, the calcium phosphates were synthesized in a laboratory using a wet chemical precipitation method from CaO (or calcium hydroxide Ca(OH)2 ) of various origins (commercial, marble, eggshells, land snails shells) as precursors and

Main advantages of this method are a simple synthesis process, a relatively quick obtaining of end product, possibility to obtain large quantities of end product, relatively cheap raw materials and the only by-product it gives is water. It is also important that calcium phosphates with nanometric crystal size can be obtained at a low process temperature (from

During synthesis, technological parameters like final pH of calcium phosphate suspension and synthesis temperature (T,°C) were changed. Final suspension pH was stabilized in the range of 5-11, using solution of acid. Synthesis temperature was changed in the range from room temperature (21°C) up to 70°C, the following parameters were controlled: acid solution adding speed (ml/min), stirring speed (rpm), synthesis temperature (T,°C), final suspension pH, stabilization time (h), maturity time (τ, h), drying temperature and time

For further obtaining HAp or biphasic bioceramic synthesized under impact of various parameters (final pH and synthesis temperature), powder is thermally treated. Samples are heated in different environments (air, vacuum and water vapor), variating thermal

In order to determine raw materials, structure of synthesized and heated powder, phase composition and functional groups, two important methods are used, complementing each other: Fourier transform infrared spectroscopy (FTIR) and X-ray difractometry (XRD). XRD method is widely used for apatite characterization, for it provides data about the crystal structure of material and its phase composition, however, it is not convenient to determine

significantly differ.

the synthesis process.

**3. Materials and methods** 

orthophosphoric acid H3PO4.

room to water boiling temperature).

(T,°C; h), calcifying temperature and time (T,°C; h).

**3.2 Analysis methods and sample preparation** 

treatment temperature in range 200-1400°C and processing time.

**3.1 Synthesis of calcium phosphates** 


Table 5. Analysis of FTIR absorption bands of thermally treated biphasic calcium phosphate (HAp/β-TCP) chemical groups

For stoichiometric HAp, HPO42- group is not detected, even though it can appear from the synthesis impurities (NO3-, NH4+).

Various studies show that in the result of HAp (OAp) decomposition, apart from TCP and TTCP, also other calcium compounds may form, like calcium pyrophosphate (CPP, β-Ca2P2O7) and calcium oxide (CaO).

Apatitic TCP (ap-TCP) Ca9(HPO4)(PO4)5(OH) is a calcium orthophosphate which, during thermal treatment at temperature higher than 750oC, transforms onto β-tricalcium phosphate Ca3(PO4)2.

During the synthesis of TCP, the most important controllable parameters are temperature and pH. According to the literary sources, pH is almost neutral or slightly acidic, and is synthesized at lower temperatures.

A pure stoichiometric β-TCP with a molar ratio Ca/P=1.500, is formed in the result of temperature treatment. If Ca/P > 1.500, HAp is formed as the second phase. When Ca/P ratio is changed for 1%, HAp is formed for 10 wt%. If Ca/P < 1.500, then DCPA is formed, this is proven by presence of calcium pyrophosphate Ca2P2O7.

In order to control speed of biodegradation, a biphasic calcium phosphate (BCP) bioceramic is developed, containing both HAp and TCP. By variation HAp/TCP ratio, it is possible to control bioactivity and biodegradation of implant. Since β-TCP is more soluble and HAp allows a biological precipitation of apatites, solubility of BCP depends on the ratio of HAp/TCP. Osteoconductivity among BCP, HAp and TCP does not significantly differ.

BCP is formed, if 1.500 < Ca/P < 1.667, which refers to a hydroxyapatite with calcium deficient, its chemical formula is Ca10-x(PO4)6-x (HPO4)x(OH)2-x (0< x <2). Ca/P ratio of synthesis sedimentary is not directly connected with the initial Ca/P ratio. At the constant pH, molar ratio of calcium phosphates may be varied by changing the temperature during the synthesis process.

## **3. Materials and methods**

130 Infrared Spectroscopy – Materials Science, Engineering and Technology

350 HPO42- Begins showing up (Raynaud,

**bands, (cm-1) Description** 

et al., 2002);

As a result of condensation, P2O7 is formed (Raynaud, et al., 2002);

Begins showing up if 1.5<Ca/P<1.677 (Raynaud, et al., 2002);

> Disappears (Raynaud, et al., 2002)

Proves presence of HAp (Destainville, et al., 2003);

Forming of pyrophosphate groups (Meejoo, et al., 2006);

> HAp with deficient of calcium, by decomposing, forms HAp and β - TCP(Raynaud, et al., 2002);

Disappears above 1000 oC, if 1.5<Ca/P<1.677 (Raynaud, et al., 2002);

**Absorption** 

et al., 2002);

820 and 1380 (Raynaud, et al., 2002);

630 and 3540 (Destainville, et al., 2003);

et al., 2006);

et al., 2002);

Table 5. Analysis of FTIR absorption bands of thermally treated biphasic calcium phosphate

For stoichiometric HAp, HPO42- group is not detected, even though it can appear from the

Various studies show that in the result of HAp (OAp) decomposition, apart from TCP and TTCP, also other calcium compounds may form, like calcium pyrophosphate (CPP, β-

Apatitic TCP (ap-TCP) Ca9(HPO4)(PO4)5(OH) is a calcium orthophosphate which, during thermal treatment at temperature higher than 750oC, transforms onto β-tricalcium

During the synthesis of TCP, the most important controllable parameters are temperature and pH. According to the literary sources, pH is almost neutral or slightly acidic, and is

A pure stoichiometric β-TCP with a molar ratio Ca/P=1.500, is formed in the result of temperature treatment. If Ca/P > 1.500, HAp is formed as the second phase. When Ca/P ratio is changed for 1%, HAp is formed for 10 wt%. If Ca/P < 1.500, then DCPA is formed,

4- 720 (Raynaud, et al., 2002);

**Temperature, oC** 

400 P2O7

600 NO3-

750 OH-

(HAp/β-TCP) chemical groups

Ca2P2O7) and calcium oxide (CaO).

synthesized at lower temperatures.

synthesis impurities (NO3-

phosphate Ca3(PO4)2.

700-900

**Chemical groups** 

350-720 HPO42- 875 (Raynaud,

800 P2O74- 715 (Meejoo,

1000 P2O74- 720 (Raynaud,

, NH4+).

this is proven by presence of calcium pyrophosphate Ca2P2O7.

## **3.1 Synthesis of calcium phosphates**

In order to achieve higher assay of the obtained material, its thermal stability and predictability of other properties, the calcium phosphates were synthesized in a laboratory using a wet chemical precipitation method from CaO (or calcium hydroxide Ca(OH)2 ) of various origins (commercial, marble, eggshells, land snails shells) as precursors and orthophosphoric acid H3PO4.

Main advantages of this method are a simple synthesis process, a relatively quick obtaining of end product, possibility to obtain large quantities of end product, relatively cheap raw materials and the only by-product it gives is water. It is also important that calcium phosphates with nanometric crystal size can be obtained at a low process temperature (from room to water boiling temperature).

During synthesis, technological parameters like final pH of calcium phosphate suspension and synthesis temperature (T,°C) were changed. Final suspension pH was stabilized in the range of 5-11, using solution of acid. Synthesis temperature was changed in the range from room temperature (21°C) up to 70°C, the following parameters were controlled: acid solution adding speed (ml/min), stirring speed (rpm), synthesis temperature (T,°C), final suspension pH, stabilization time (h), maturity time (τ, h), drying temperature and time (T,°C; h), calcifying temperature and time (T,°C; h).

For further obtaining HAp or biphasic bioceramic synthesized under impact of various parameters (final pH and synthesis temperature), powder is thermally treated. Samples are heated in different environments (air, vacuum and water vapor), variating thermal treatment temperature in range 200-1400°C and processing time.

## **3.2 Analysis methods and sample preparation**

In order to determine raw materials, structure of synthesized and heated powder, phase composition and functional groups, two important methods are used, complementing each other: Fourier transform infrared spectroscopy (FTIR) and X-ray difractometry (XRD). XRD method is widely used for apatite characterization, for it provides data about the crystal structure of material and its phase composition, however, it is not convenient to determine

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 133

While preparing bioceramic samples from various commercial materials available on the market, we have come across hardly predictable properties of the end product, like crystallinity degree, phase composition and, following, bioactivity and mechanical characteristics. One of the disadvantages while purchasing commercial calcium phosphate powders or commercial calcium phosphate ceramic materials is insufficient information about synthesis conditions of these calcium phosphates, raw materials and in which

It is significant to know if the purchased powder is thermally treated, and in which temperature range this thermal treatment was performed. Exactly temperature, at which the sample was obtained and processed, is one of the conditions that influences outcome of

groups are observed, but band shape, width and intensity are different. Differences in spectra are also observed at CO32- and HPO42-groups location and intensity. On Fig. 3, there are three spectra from different commercial HAp compared: "Fluka"(F), "Riedel-de Haën®"

and PO43-

For example, in spectra of several overviewed commercial hydroxyapatites, OH-

Fig. 3. FTIR spectra of commercial HAp products (1 - S-A; 2 - R-dH; 3 - Fluka)

Fig. 2. FTIR spectrum of the KBr pellet

proportions these materials are taken.

ceramic and phase composition.

**4.1 Commercial HAp products description** 

**4. Results and discussion** 

amount of [OH] or [CO3] groups in hydroxyapatite. FTIR method, in many cases is more sensitive than XRD when determining presence of new phases. Using FTIR, CaP can be characterized, considering three spectrum parameters:


During the research process, *X"Pert PRO* X-ray diffractometer has been used (*PANalitical*, the Netherlands). Samples were measured in a spinning mode, in the 2θ angle range from 20-90º, with a scanning step 0,0334o, with a CuKα radiation. Ratio of HAp/TCP is determined using a XRD semi-quantitative method after calibration line.

Calcium phosphate spectra are measured and analyzed with a FTIR spectrometer "Varian 800" of Scimitar Series, with a wave length range from 400 – 4000 cm-1, with precision of 4 cm-1 and RESOLUTION software.

Samples are prepared by mixing powder with KBr and pressing the pellet. This method of sample preparation has some complications and requires a certain experience, in order to obtain a good quality spectrum in everyday work routine. Special factors should be considered in order to perform invariable sample analysis by using a KBr method, and these include pellet thickness, particle dispersion, ensuring vacuum state during the pressing, pressure influence, ion exchange, etc.

In a prepared KBr pellet, there should be material concentration of 2-10% from the total weight. For preparing a 300g KBr pellet, from 1 to 5mg of the sample is required, and the pellet size will be 13 mm.

Powder grain size should be ~150 μm. The analyzed sample is crashed in a powder and thoroughly mixed with the KBr powder. A powder mixer "Pulverisette 23" was used for crashing and mixing of the powder. Prepared powder was located in a specific SPECAC (d = 13 mm) mould, and the required pressure was achieved by applying a uniaxial press (required pressure is ~5. 103 kg/cm2, pressing time 1 min).

KBr attracts water molecules from the environment and they create wide water bands in the spectrum, so they are hard to or even impossible to analyze. KBr powder should be of the highest assay, Riedel-de-Haen KBr (Lot 51520) brand was used with an assay in the range of 99.5-100.5%. Usually, absorption bands of the main impurities in the KBr are: OH- groups and H2O molecules (3500 cm-1 and 1630 cm-1), NO2 (1390 cm-1), SO42- (1160-1140 cm-1). Spectrum of the KBr used in our research is demonstrated on the Fig. 2. Considering that the powder is hygroscopic, KBr powder is dried at 105oC and kept in special hermetic containers. Prepared pellets with the analyzed material are dried once again for 24 h. The obtained pellets should be transparent and equally colored. Weak bands connected with water can also be compensated by using a KBr pellet of the same thickness, but not containing the analyzed material, for background spectrum measuring.

Fig. 2. FTIR spectrum of the KBr pellet

## **4. Results and discussion**

132 Infrared Spectroscopy – Materials Science, Engineering and Technology

amount of [OH] or [CO3] groups in hydroxyapatite. FTIR method, in many cases is more sensitive than XRD when determining presence of new phases. Using FTIR, CaP can be

 Location of absorption maximum indicates material composition, even slight variations of the composition influence energy of material bondings and, as follows, frequency of

 Considering the absorption maximum of [OH] vibrations, presence of HAp and its thermal stability can be determined, as well as hydroxyl group concentration in the

During the research process, *X"Pert PRO* X-ray diffractometer has been used (*PANalitical*, the Netherlands). Samples were measured in a spinning mode, in the 2θ angle range from 20-90º, with a scanning step 0,0334o, with a CuKα radiation. Ratio of HAp/TCP is

Calcium phosphate spectra are measured and analyzed with a FTIR spectrometer "Varian 800" of Scimitar Series, with a wave length range from 400 – 4000 cm-1, with precision of 4

Samples are prepared by mixing powder with KBr and pressing the pellet. This method of sample preparation has some complications and requires a certain experience, in order to obtain a good quality spectrum in everyday work routine. Special factors should be considered in order to perform invariable sample analysis by using a KBr method, and these include pellet thickness, particle dispersion, ensuring vacuum state during the pressing,

In a prepared KBr pellet, there should be material concentration of 2-10% from the total weight. For preparing a 300g KBr pellet, from 1 to 5mg of the sample is required, and the

Powder grain size should be ~150 μm. The analyzed sample is crashed in a powder and thoroughly mixed with the KBr powder. A powder mixer "Pulverisette 23" was used for crashing and mixing of the powder. Prepared powder was located in a specific SPECAC (d = 13 mm) mould, and the required pressure was achieved by applying a uniaxial press

KBr attracts water molecules from the environment and they create wide water bands in the spectrum, so they are hard to or even impossible to analyze. KBr powder should be of the highest assay, Riedel-de-Haen KBr (Lot 51520) brand was used with an assay in the range of 99.5-100.5%. Usually, absorption bands of the main impurities in the KBr are: OH- groups and H2O molecules (3500 cm-1 and 1630 cm-1), NO2 (1390 cm-1), SO42- (1160-1140 cm-1). Spectrum of the KBr used in our research is demonstrated on the Fig. 2. Considering that the powder is hygroscopic, KBr powder is dried at 105oC and kept in special hermetic containers. Prepared pellets with the analyzed material are dried once again for 24 h. The obtained pellets should be transparent and equally colored. Weak bands connected with water can also be compensated by using a KBr pellet of the same thickness, but not

103 kg/cm2, pressing time 1 min).

containing the analyzed material, for background spectrum measuring.

Peak width shows degree of the atoms' order in the apatite elementary cell.

determined using a XRD semi-quantitative method after calibration line.

characterized, considering three spectrum parameters:

variations;

sample.

cm-1 and RESOLUTION software.

pressure influence, ion exchange, etc.

pellet size will be 13 mm.

(required pressure is ~5.

#### **4.1 Commercial HAp products description**

While preparing bioceramic samples from various commercial materials available on the market, we have come across hardly predictable properties of the end product, like crystallinity degree, phase composition and, following, bioactivity and mechanical characteristics. One of the disadvantages while purchasing commercial calcium phosphate powders or commercial calcium phosphate ceramic materials is insufficient information about synthesis conditions of these calcium phosphates, raw materials and in which proportions these materials are taken.

It is significant to know if the purchased powder is thermally treated, and in which temperature range this thermal treatment was performed. Exactly temperature, at which the sample was obtained and processed, is one of the conditions that influences outcome of ceramic and phase composition.

For example, in spectra of several overviewed commercial hydroxyapatites, OH and PO4 3 groups are observed, but band shape, width and intensity are different. Differences in spectra are also observed at CO32- and HPO42-groups location and intensity. On Fig. 3, there are three spectra from different commercial HAp compared: "Fluka"(F), "Riedel-de Haën®"

Fig. 3. FTIR spectra of commercial HAp products (1 - S-A; 2 - R-dH; 3 - Fluka)

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 135

In order to control reaction process and possible appearance of by-products, control samples were taken every 5 minutes and analyzed by measuring their spectra. After taking a sample from reactor with a plastic dropper, it was inserted in a glass bottle, hermetically sealed and frozen by putting it in the mixture of dry ice and acetone. After full freezing, the temperature is supported by storing the bottle in the dry ice. Before inserting in the cryogenic drying device, bottles are covered with a perforated plastic film. From the spectra of samples prepared this way (Fig.5), synthesis with a final pH=9,3 and synthesis temperature 45oC, it can be seen that the synthesized material is formed with an apatitic HAp structure with a slight Ca

reducing along with reaction approaching its end. No by-products were detected, during the

 peaks at 3571 cm-1 and 631 cm-1 which is characteristic for HAp phase. Such control is very important for scaling the synthesis, relatively increasing amount of

Fig. 5. FTIR spectra of the samples from the synthesis series with final pH=9,3 at temperature 45oC, depending on the reaction occurrence time (from 1 to18).

**shells) and its influence on the CaP product properties** 

the obtained bioceramics.

**4.3 Selection of CaO containing materials of various origins (marble, eggshells, snail** 

Scientific literature contains very few information about research of how raw materials' (for example, CaO, Ca(OH)2) quality (chemical and physical properties) impacts properties of

Before starting synthesis of CaP products, selection of CaO containing raw material and complex research were performed. For "Ca" precursors, two commercial available synthetic

2- group, amount of which is constantly

peak disappearing at 3642 cm-1 and

deficient, which is proven by presence of the CO3

synthesis and amount of obtained CaP.

forming OH-

reaction, CaO has reacted fully which is proven by an OH-

(R-dH) and "Sigma Aldrich" (S-A). IR spectrum of commercial S-A product has a lower CO32- and HPO42- intensity that could mean a higher assay degree than the other materials have. These spectra also demonstrate that the products have a low crystallinity degree and they were not thermally treated.

Characteristic chemical groups of the commercial synthesized products are summarized in table 6.


Table 6. Characteristic chemical groups of commercial HAp FTIR absorption bands

In the FTIR spectra of commercial β-TCP products, PO4 3- groups are observed, which is characteristic to β-TCP (Fig. 4). In HAp IR spectrum, OH group peaks are observed (at 630 and 3570 cm-1), but there are no such in the IR spectrum of commercial β-TCP, which means that there is no HAp phase in this β-TCP product. In addition, at 725 cm-1, presence of P2O7 4 group can be observed, which is characteristic to calcium pyrophosphate phase.

Fig. 4. FTIR spectra of commercial product (Fluka β-TCP)

#### **4.2 Phase composition control during synthesis reaction**

As above mentioned, CaP material synthesis was performed in our laboratory, variating synthesis and further sample heating parameters, searching for their optimal combination depending on the required properties of the obtained material.

(R-dH) and "Sigma Aldrich" (S-A). IR spectrum of commercial S-A product has a lower CO32- and HPO42- intensity that could mean a higher assay degree than the other materials have. These spectra also demonstrate that the products have a low crystallinity degree and

Characteristic chemical groups of the commercial synthesized products are summarized in

1382; 1413; 1457; 1634; 1997

470; 553 - 610; 964; 1000 - 1150

1382; 1417; 2457; 1639; 1990; 2359

471; 561; 601; 605; 964; 1013 - 1120

**HAp Fluka R-dH S-A** 

**H2O adsorbed** 3100 - 3600 3000 -3600 3200 - 3600 **OH-** 635; 3568 3568 630; 3569;

**HPO42-** ; **CO32-** 891; 875 870 874

Table 6. Characteristic chemical groups of commercial HAp FTIR absorption bands

group can be observed, which is characteristic to calcium pyrophosphate phase.

In the FTIR spectra of commercial β-TCP products, PO43- groups are observed, which is characteristic to β-TCP (Fig. 4). In HAp IR spectrum, OH- group peaks are observed (at 630 and 3570 cm-1), but there are no such in the IR spectrum of commercial β-TCP, which means that there is no HAp phase in this β-TCP product. In addition, at 725 cm-1, presence of P2O74-

As above mentioned, CaP material synthesis was performed in our laboratory, variating synthesis and further sample heating parameters, searching for their optimal combination

**groups Absorption bands, cm-1**

they were not thermally treated.

**CO32-** 1386; 1411; 1635; 1997;

**PO43-** 470; 553 - 600; 964;

(2359 C≡C)

1000 - 1156

Fig. 4. FTIR spectra of commercial product (Fluka β-TCP)

**4.2 Phase composition control during synthesis reaction** 

depending on the required properties of the obtained material.

table 6.

**Commercial** 

**Chemical** 

In order to control reaction process and possible appearance of by-products, control samples were taken every 5 minutes and analyzed by measuring their spectra. After taking a sample from reactor with a plastic dropper, it was inserted in a glass bottle, hermetically sealed and frozen by putting it in the mixture of dry ice and acetone. After full freezing, the temperature is supported by storing the bottle in the dry ice. Before inserting in the cryogenic drying device, bottles are covered with a perforated plastic film. From the spectra of samples prepared this way (Fig.5), synthesis with a final pH=9,3 and synthesis temperature 45oC, it can be seen that the synthesized material is formed with an apatitic HAp structure with a slight Ca deficient, which is proven by presence of the CO3 2- group, amount of which is constantly reducing along with reaction approaching its end. No by-products were detected, during the reaction, CaO has reacted fully which is proven by an OH peak disappearing at 3642 cm-1 and forming OH peaks at 3571 cm-1 and 631 cm-1 which is characteristic for HAp phase.

Such control is very important for scaling the synthesis, relatively increasing amount of synthesis and amount of obtained CaP.

Fig. 5. FTIR spectra of the samples from the synthesis series with final pH=9,3 at temperature 45oC, depending on the reaction occurrence time (from 1 to18).

#### **4.3 Selection of CaO containing materials of various origins (marble, eggshells, snail shells) and its influence on the CaP product properties**

Scientific literature contains very few information about research of how raw materials' (for example, CaO, Ca(OH)2) quality (chemical and physical properties) impacts properties of the obtained bioceramics.

Before starting synthesis of CaP products, selection of CaO containing raw material and complex research were performed. For "Ca" precursors, two commercial available synthetic

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 137

Fig. 7. FTIR spectra of biogenic and commercial CaO after heating at 1000ºC for 1 h (F – "Fluka", Gl – land snail shells, Ol – egg shells and R – "Riedel-de Haën®")

Fig. 8. FTIR spectra of HAp products as-synthesized after drying at 105ºC for ~ 20 h from various CaO (F – "Fluka", Gl - land snail shells, Ol – egg shells and R – "Riedel-de Haën®").

CaO powders from "Riedel-de Haën®" (CaOR) and "Fluka" (CaOF) and two materials of biogenic origin, widespread in the nature – egg shells and land snail (*Arianta arbustorum)* shells, were chosen. These materials were selected as raw materials for obtaining CaO and further usage in the synthesis process of CaP.

In composition of commercial CaO, presence of Ca(OH)2 phase, small amount of MgO phase, as well as small amount of polymorphous CaCO3 modification – calcite phase (Fig. 6), is detected. Presence of CaCO3 is undesirable in CaO, for during the process of suspension obtaining, Ca(OH)2 creates an error in preparation of precise amount of the reagent and, as follows, product reproduction, also preventing obtaining a homogenous Ca(OH)2 suspension which is required for further synthesis of CaP. After heating biogenic and commercial CaO at 1000ºC, X-ray diagrams of all the synthesis materials demonstrated a CaO crystal phase with a small amount of MgO phase, along with Ca(OH)2 that formed in the result of CaO contact with air moisture; considering that CaO is hygroscopic (Siva Rama Krishna, et al., 2007).

Fig. 6. Commercial CaO X-ray diffractograms before calcifying

FTIR spectra were also taken for the thermally treated CaO samples, and their compositions were similar (Fig. 7.). In all the IR spectra, an explicit absorption peak is visible at 3642 cm-1, that indicates stretching variations of Ca(OH)2 (Ji, et al., 2009) and [OH] groups (Siva Rama Krishna, et al., 2007). Presence of CaO is also proven by a wide intensive absorption band of [Ca-O] group, which is centered at ~ 400 cm-1 (Ji, et al., 2009). Absorption peaks at 874 cm-1, 1080 cm-1 [113], as well as at 1420 cm-1 prove presence of the [CO3] groups in the samples which, therefore, shows a slight carboxilation of Ca(OH)2 from CO2 of the atmosphere. Absorption band from 3430-3550 cm-1 proves presence of adsorbed water molecules in the samples.

FTIR spectra of the synthesized powders demonstrate absorption bands of the chemical functional groups, characteristic to HAp phase (Fig. 8.). Number of [**CO3**] groups in those is different, but after thermal treatment at 1100ºC for 1 h, the spectra become very similar (Fig. 9).

CaO powders from "Riedel-de Haën®" (CaOR) and "Fluka" (CaOF) and two materials of biogenic origin, widespread in the nature – egg shells and land snail (*Arianta arbustorum)* shells, were chosen. These materials were selected as raw materials for obtaining CaO and

In composition of commercial CaO, presence of Ca(OH)2 phase, small amount of MgO phase, as well as small amount of polymorphous CaCO3 modification – calcite phase (Fig. 6), is detected. Presence of CaCO3 is undesirable in CaO, for during the process of suspension obtaining, Ca(OH)2 creates an error in preparation of precise amount of the reagent and, as follows, product reproduction, also preventing obtaining a homogenous Ca(OH)2 suspension which is required for further synthesis of CaP. After heating biogenic and commercial CaO at 1000ºC, X-ray diagrams of all the synthesis materials demonstrated a CaO crystal phase with a small amount of MgO phase, along with Ca(OH)2 that formed in the result of CaO contact with air moisture; considering that CaO is hygroscopic (Siva Rama

further usage in the synthesis process of CaP.

Fig. 6. Commercial CaO X-ray diffractograms before calcifying

FTIR spectra were also taken for the thermally treated CaO samples, and their compositions were similar (Fig. 7.). In all the IR spectra, an explicit absorption peak is visible at 3642 cm-1, that indicates stretching variations of Ca(OH)2 (Ji, et al., 2009) and [OH] groups (Siva Rama Krishna, et al., 2007). Presence of CaO is also proven by a wide intensive absorption band of [Ca-O] group, which is centered at ~ 400 cm-1 (Ji, et al., 2009). Absorption peaks at 874 cm-1, 1080 cm-1 [113], as well as at 1420 cm-1 prove presence of the [CO3] groups in the samples which, therefore, shows a slight carboxilation of Ca(OH)2 from CO2 of the atmosphere. Absorption band from 3430-3550 cm-1 proves presence of adsorbed water molecules in the

FTIR spectra of the synthesized powders demonstrate absorption bands of the chemical functional groups, characteristic to HAp phase (Fig. 8.). Number of [**CO3**] groups in those is different, but after thermal treatment at 1100ºC for 1 h, the spectra become very similar (Fig. 9).

Krishna, et al., 2007).

samples.

Fig. 7. FTIR spectra of biogenic and commercial CaO after heating at 1000ºC for 1 h (F – "Fluka", Gl – land snail shells, Ol – egg shells and R – "Riedel-de Haën®")

Fig. 8. FTIR spectra of HAp products as-synthesized after drying at 105ºC for ~ 20 h from various CaO (F – "Fluka", Gl - land snail shells, Ol – egg shells and R – "Riedel-de Haën®").

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 139

Thermal behavior data of laboratory synthesed calcium phosphates with various synthesis parameters were summarized. Final reaction pH was variated in the range from 5.0 up to

1 5,0 room TCP+CaHPO4 2 5,1 70 HAp/TCP (60/40) 3 5,3 45 HAp/TCP (35/65)

4 5,9 room TCP

6 9,3 45 HAp 7 10,7 45 HAp/CaO Table 7. Variable parameters of some calcium phosphate synthesis and calcium phosphate

It can be ascertained that, in spite of variable parameters of synthesis, all the spectra of samples and XRD diffraction diagrams, are similar, all the functional groups correspond with non-stoichiometric apatitic HAp structure with a low crystallinity degree (Fig. 10). The only slight differences that can be observed between spectra, are intensities of CO32- group bands, however, no significant differences in the number of OH- groups, depending on the

Fig. 10. FTIR spectrum of as-synthesized and drying at 105ºC (20 h) CaP products of 1-7

products phase composition after thermal treatment at 1100ºC for 1 h

synthesis parameters at the non-calcified samples, are detected (Table 8).

5 7 45 HAp/TCP (80/20)

Phase composition by XRD after thermal treatment at 1100oC (1h)

10.7 and synthesis temperature was chosen as room (22oC), 45oC and 70oC (Table 7).

synthesis Synthesis temperature, oC

Final pH of

Nr. of synthesis

synthesis

Fig. 9. FTIR spectra of HAp bioceramics obtained from various CaO after thermal treatment at 1100ºC for 1 h.

Using CaO containing materials of various origins, including the commercial CaO available on the market that seemingly correspond with the assay specification, it is still impossible to obtain a completely reproductable HAp bioceramic materials. XRD and FTIR of the HAp products synthesized and heated at 1100ºC, produce similar pictures, and post-synthesis morphology, phase composition and molecular structure of these bioceramics are identical. However, analyzing it with FE-SEM micrographs, it can be observed that obtained HAp products demonstrate different microstructures.

A homogenous, fine-grained microstructure with a small grain size – about 150-200 nm, is observed at the HAp bioceramic from commercial reagents, and non-homogenic grainy structure with irregular grain size in the range from 200 nm up to 1 mcm is observed at the ceramic which was synthesized using Ca of natural origin.

Therefore, the samples from commercial reagents demonstrated color changing, from white in the synthesized and just dried powder to average aquamarine, at ceramic samples heated at 1000 °C, up to light blue color at 1300 °C. It can be explained by oxidizing of the manganese impurities from Mn2+ up to Mn5+ and substitution of hydroxyapatite (PO43-) group with (MnO43-) (Ślósarczyk, et al., 2010). Color change can be also explained by other microelements or defects in the crystal lattice, but, in this case, FTIR and XRD analysis cannot give a precise answer to this.

#### **4.4 Thermal behavior of calcium phosphates depending on synthesis parameters**

Sample structure, phase composition and thermal stability after heating depend on synthesis parameters, especially from final pH and synthesis temperature. Also, connection between pH value and temperature (with temperature increasing, pH decreases); so, by combining and analyzing those parameters, both pure HAp and TCP materials with a good thermal stability are obtained and biphasic materials with various Ca/P ratios and various percentages of phase composition.

Fig. 9. FTIR spectra of HAp bioceramics obtained from various CaO after thermal treatment

Using CaO containing materials of various origins, including the commercial CaO available on the market that seemingly correspond with the assay specification, it is still impossible to obtain a completely reproductable HAp bioceramic materials. XRD and FTIR of the HAp products synthesized and heated at 1100ºC, produce similar pictures, and post-synthesis morphology, phase composition and molecular structure of these bioceramics are identical. However, analyzing it with FE-SEM micrographs, it can be observed that obtained HAp

A homogenous, fine-grained microstructure with a small grain size – about 150-200 nm, is observed at the HAp bioceramic from commercial reagents, and non-homogenic grainy structure with irregular grain size in the range from 200 nm up to 1 mcm is observed at the

Therefore, the samples from commercial reagents demonstrated color changing, from white in the synthesized and just dried powder to average aquamarine, at ceramic samples heated at 1000 °C, up to light blue color at 1300 °C. It can be explained by oxidizing of the manganese impurities from Mn2+ up to Mn5+ and substitution of hydroxyapatite (PO43-) group with (MnO43-) (Ślósarczyk, et al., 2010). Color change can be also explained by other microelements or defects in the crystal lattice, but, in this case, FTIR and XRD analysis

**4.4 Thermal behavior of calcium phosphates depending on synthesis parameters** 

Sample structure, phase composition and thermal stability after heating depend on synthesis parameters, especially from final pH and synthesis temperature. Also, connection between pH value and temperature (with temperature increasing, pH decreases); so, by combining and analyzing those parameters, both pure HAp and TCP materials with a good thermal stability are obtained and biphasic materials with various Ca/P ratios and various

at 1100ºC for 1 h.

products demonstrate different microstructures.

cannot give a precise answer to this.

percentages of phase composition.

ceramic which was synthesized using Ca of natural origin.

Thermal behavior data of laboratory synthesed calcium phosphates with various synthesis parameters were summarized. Final reaction pH was variated in the range from 5.0 up to 10.7 and synthesis temperature was chosen as room (22oC), 45oC and 70oC (Table 7).


Table 7. Variable parameters of some calcium phosphate synthesis and calcium phosphate products phase composition after thermal treatment at 1100ºC for 1 h

It can be ascertained that, in spite of variable parameters of synthesis, all the spectra of samples and XRD diffraction diagrams, are similar, all the functional groups correspond with non-stoichiometric apatitic HAp structure with a low crystallinity degree (Fig. 10). The only slight differences that can be observed between spectra, are intensities of CO3 2- group bands, however, no significant differences in the number of OH- groups, depending on the synthesis parameters at the non-calcified samples, are detected (Table 8).

Fig. 10. FTIR spectrum of as-synthesized and drying at 105ºC (20 h) CaP products of 1-7 synthesis

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 141

These locations of absorption bands of the functional groups explicitly indicate forming of a typical HAp structure in the synthesized samples (Nilen & Richter, 2008; Siva Rama

Weak absorption bands at 2365 cm-1 and 2344 cm-1 appeared due to attraction of CO2 from

Presence of [CO3] bands can be identified with explicitly visible absorption bands within the range between 1600-1400 cm-1 and at 875 cm-1, which are observed in the spectra of synthesized samples. Absorption bands of weak intensity, centered at 1418 cm-1 and 1458 cm-1 correspond to symmetrical and asymmetrical stretching modes of the [CO3]ν3 groups (C-O). Absorption band at ~875 cm-1 can prove presence of [CO3]ν2 stretching mode (at ~872 cm-1), intensity of which is approximately 1/5 share of [CO3]ν3, or presence of [HPO4] group absorption maximum (Siva Rama Krishna, et al., 2007; Landi, et al., 2000). Considering that the [HPO4] group band partially covers [CO3]ν2, it is complicated to detect, of which group is this band. Presence of this absorption band ascertains solution of atmosphere CO2 in the

Combination of absorption bands – absorption bands of [CO3]ν3 groups at 1418 cm-1 and 1458 cm-1, as well as at 875 cm-1, proves substitution of "B-type" [PO4] groups with [CO3] groups in the HAp crystal lattice (Barinov, et al., 2006; Siva Rama Krishna, et al., 2007).

Absorption maximum of [CO3] group at 875 cm-1 can also prove that "AB-type" ([PO4] and [OH] groups) substitution in the structure of HAp, as well as a weak absorption band at 3571 cm-1 in the synthesized HAp samples can mean the "AB-type" substitution (Barinov, et

A wide absorption band within the range from ~3600 cm-1 up to 3100 cm-1 points on ν3 and ν1 with H2O molecules bonded with hydrogen for stretching modes and an absorption band at 1629 cm-1 is referable onto deformation mode ν2 of H2O molecules (Siva Rama Krishna, et al., 2007), that proves presence of physically adsorbed water in the synthesized samples.

Processing the samples in the temperature range between 200 ºC to 1400 ºC in the air atmosphere, a similar sample behavior, even up to characteristic bioceramic sintering temperatures, is observed in all the syntheses. Dehydratation of the samples occurs up to 500oC, and at 600oC, the spectra demonstrate that adsorbed water band disappears from the spectrum, adsorbed and capillary water is eliminated from CaP. Within the temperature range between 500oC and 800oC, amount of CO32- groups in the samples also reduces, and bands of CO32- group fully disappear at 900 oC. FTIR spectra (Fig. 12) and XRD diffractograms (Fig. 13) of the thermally treated at 1100ºC samples considerably differ, comparing with the samples that just have been synthesized. A restructurization of functional [PO4] groups have occurred, and sample phase composition is considerably different from the combination of the initial synthesis parameters. It can be concluded that

A pure, stoichiometric and stable HAp in a wide temperature range is obtained using the following synthesis parameters: final pH=9,3 and synthesis temperature Ts=45oC (synthesis 6). Sample thermal treatment is performed in the air atmosphere, in the temperature range 200oC – 1400oC for 1 h. HAp spectra at various heating temperatures are demonstrated on

Krishna, et al., 2007; Kothapalli, et al., 2004; Landi, et al., 2000; Lioua, et al., 2004).

suspension, if synthesis occurs in the alkaline environment.

phases with a high crystallinity degree were formed.

the atmosphere.

al., 2006).

the Fig. 14 and 15.

Fig. 11. XRD patterns of as-synthesized CaP products from 4, 5, 6, 7 synthesis.


Table 8. Absorption bands of as-synthesized and drying at 105ºC (20 h) CaP products

Absorption bands of chemical functional groups characteristic to HAp phase can be defined as follows:


Fig. 11. XRD patterns of as-synthesized CaP products from 4, 5, 6, 7 synthesis.

**Chemical groups Absorption bands, cm-1**

**HPO42-** 875 (identifies HAp with deficient of calcium and nonstoichiometric structure)

Table 8. Absorption bands of as-synthesized and drying at 105ºC (20 h) CaP products

Absorption bands of chemical functional groups characteristic to HAp phase can be defined

 Absorption bands at 3570 cm-1 and 631 cm-1 are referable to structural [OH] groups (O-H) stretching and libration modes at the HAp crystallite surface or at the crystallites. Presence of [PO4] groups, characteristic to tetrahedral apatite structure, is proven by absorption bands at 472 cm-1 , which is characteristic to [PO4]ν2 group (ν2 O-P-O) bending variations; double band at 570 cm-1 and 602 cm-1 with a high resolution is referable to asymmetric and symmetric deformation modes of [PO4]ν4 group (ν4 O-P-O); absorption band at 963 cm-1 corresponds to a symmetric stretching mode; intensive absorption band in the range of 1040-1090 cm-1 corresponds to a band characteristic to [PO4]ν3 groups (ν3 P-O) at 1040 cm-1 and 1090 cm-1 asymmetrical stretching mode, which, as explicit maximums, can be observed after thermal

An absorption band of weak intensity within the range between 1950-

2100 cm-1 is connected with combinations of [PO4]ν3, ν1 modes.

**PO43-** 472; 570; 602; 963; 1000 - 1140;

**CO32-** 875; 1418; 1458; 1632 and 1650; 1994

**H2O adsorbed** 3100 - 3600; **OH-** 631; 3570;

treatment of the samples;

as follows:

These locations of absorption bands of the functional groups explicitly indicate forming of a typical HAp structure in the synthesized samples (Nilen & Richter, 2008; Siva Rama Krishna, et al., 2007; Kothapalli, et al., 2004; Landi, et al., 2000; Lioua, et al., 2004).

Weak absorption bands at 2365 cm-1 and 2344 cm-1 appeared due to attraction of CO2 from the atmosphere.

Presence of [CO3] bands can be identified with explicitly visible absorption bands within the range between 1600-1400 cm-1 and at 875 cm-1, which are observed in the spectra of synthesized samples. Absorption bands of weak intensity, centered at 1418 cm-1 and 1458 cm-1 correspond to symmetrical and asymmetrical stretching modes of the [CO3]ν3 groups (C-O). Absorption band at ~875 cm-1 can prove presence of [CO3]ν2 stretching mode (at ~872 cm-1), intensity of which is approximately 1/5 share of [CO3]ν3, or presence of [HPO4] group absorption maximum (Siva Rama Krishna, et al., 2007; Landi, et al., 2000). Considering that the [HPO4] group band partially covers [CO3]ν2, it is complicated to detect, of which group is this band. Presence of this absorption band ascertains solution of atmosphere CO2 in the suspension, if synthesis occurs in the alkaline environment.

Combination of absorption bands – absorption bands of [CO3]ν3 groups at 1418 cm-1 and 1458 cm-1, as well as at 875 cm-1, proves substitution of "B-type" [PO4] groups with [CO3] groups in the HAp crystal lattice (Barinov, et al., 2006; Siva Rama Krishna, et al., 2007).

Absorption maximum of [CO3] group at 875 cm-1 can also prove that "AB-type" ([PO4] and [OH] groups) substitution in the structure of HAp, as well as a weak absorption band at 3571 cm-1 in the synthesized HAp samples can mean the "AB-type" substitution (Barinov, et al., 2006).

A wide absorption band within the range from ~3600 cm-1 up to 3100 cm-1 points on ν3 and ν1 with H2O molecules bonded with hydrogen for stretching modes and an absorption band at 1629 cm-1 is referable onto deformation mode ν2 of H2O molecules (Siva Rama Krishna, et al., 2007), that proves presence of physically adsorbed water in the synthesized samples.

Processing the samples in the temperature range between 200 ºC to 1400 ºC in the air atmosphere, a similar sample behavior, even up to characteristic bioceramic sintering temperatures, is observed in all the syntheses. Dehydratation of the samples occurs up to 500oC, and at 600oC, the spectra demonstrate that adsorbed water band disappears from the spectrum, adsorbed and capillary water is eliminated from CaP. Within the temperature range between 500oC and 800oC, amount of CO32- groups in the samples also reduces, and bands of CO32- group fully disappear at 900 oC. FTIR spectra (Fig. 12) and XRD diffractograms (Fig. 13) of the thermally treated at 1100ºC samples considerably differ, comparing with the samples that just have been synthesized. A restructurization of functional [PO4] groups have occurred, and sample phase composition is considerably different from the combination of the initial synthesis parameters. It can be concluded that phases with a high crystallinity degree were formed.

A pure, stoichiometric and stable HAp in a wide temperature range is obtained using the following synthesis parameters: final pH=9,3 and synthesis temperature Ts=45oC (synthesis 6). Sample thermal treatment is performed in the air atmosphere, in the temperature range 200oC – 1400oC for 1 h. HAp spectra at various heating temperatures are demonstrated on the Fig. 14 and 15.

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 143

heating and then cooling a HAp sample in the air atmosphere, a complete HAp decomposition cannot be achieved, for during cooling [OH] groups get back in the structure

Fig. 14. FTIR spectra of HAp (synthesis 6) depending of thermal treatment temperature in

Fig. 15. FTIR spectra of HAp (synthesis 6) depending of thermal treatment temperature in

A pure β-tricalcium phosphate phase Ca3(PO4)2, obtained using the following synthesis parameters: final pH=5,8 and T= 22 oC (synthesis 4), it is synthesized at low temperatures

OH- group bands completely disappear at 700oC with a wave length 3569 cm-1 and 631 cm-1, which is characteristical for HAp. It can also be observed that at 900oC, bands that

range 200 – 1000oC for 1 h

range 1100 – 1400oC

and pH is acidic.

from the air and therefore a partial reversible α-TCP and TTCPtransform into HAp.

Fig. 12. FTIR spectra of calcium phosphates thermally treated at 1100oC for 1 h, depending on synthesis parameters (1, 2, 4, 6 and 7 synthesis)

Fig. 13. X-ray diffraction patterns of different calcium phosphate of calcium phosphates thermally treated at 1100oC for 1 h with different phase compositions for synthesis 6 – pure HAp, 4 – pure -TCP and 2 – biphasic mixture of HAp/-TCP

As can be seen after FTIR analysis which is often more sensitive than XRD, HAp begins to decompose when a sample is heated for 1 h at more than 1200°C, forming TCP shoulders at 948 cm-1, 975 cm-1 and poorly intensive band at 432 cm-1, which proves dehydroxilation of HAp and forming of α-TCP or OHAp phases. Therefore, XRD analysis shows that the decomposition occurs only at 1400°C, and is resulted with mixture of α-tricalcium phosphate (α-TCP, Ca3(PO4)2), tetracalcium phosphate (TTCP, Ca4(PO4)2O) and HAp. By

Fig. 12. FTIR spectra of calcium phosphates thermally treated at 1100oC for 1 h, depending

Fig. 13. X-ray diffraction patterns of different calcium phosphate of calcium phosphates thermally treated at 1100oC for 1 h with different phase compositions for synthesis 6 – pure

As can be seen after FTIR analysis which is often more sensitive than XRD, HAp begins to decompose when a sample is heated for 1 h at more than 1200°C, forming TCP shoulders at 948 cm-1, 975 cm-1 and poorly intensive band at 432 cm-1, which proves dehydroxilation of HAp and forming of α-TCP or OHAp phases. Therefore, XRD analysis shows that the decomposition occurs only at 1400°C, and is resulted with mixture of α-tricalcium phosphate (α-TCP, Ca3(PO4)2), tetracalcium phosphate (TTCP, Ca4(PO4)2O) and HAp. By

HAp, 4 – pure -TCP and 2 – biphasic mixture of HAp/-TCP

on synthesis parameters (1, 2, 4, 6 and 7 synthesis)

heating and then cooling a HAp sample in the air atmosphere, a complete HAp decomposition cannot be achieved, for during cooling [OH] groups get back in the structure from the air and therefore a partial reversible α-TCP and TTCPtransform into HAp.

Fig. 14. FTIR spectra of HAp (synthesis 6) depending of thermal treatment temperature in range 200 – 1000oC for 1 h

Fig. 15. FTIR spectra of HAp (synthesis 6) depending of thermal treatment temperature in range 1100 – 1400oC

A pure β-tricalcium phosphate phase Ca3(PO4)2, obtained using the following synthesis parameters: final pH=5,8 and T= 22 oC (synthesis 4), it is synthesized at low temperatures and pH is acidic.

OH- group bands completely disappear at 700oC with a wave length 3569 cm-1 and 631 cm-1, which is characteristical for HAp. It can also be observed that at 900oC, bands that

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 145

In the synthesis with final pH=5,00 and T=22 oC (synthesis 1), a complicated phase composition is formed from β-TCP, CaHPO4 and TTCP phases, and it remains stable up to 1200 oC. CaHPO4 phase can be recognized after a HPO42- group band at 897 cm-1, which is

In the synthesis with a final pH=10,74 and T=45oC (synthesis 7), a HAp structure is formed but at 1000oC, and additional OH- peak appears at 3644 cm-1, which means that there is a unreacted CaO left. The reaction did not occur completely and this synthesis, as well as the

> 875; 1384; 1418; 1457; 1636

> > 3572; 634

471; 556-604; 963

> 946; 975

Table 9. FTIR absorption bands (cm-1) of thermally treated laboratory synthesized HAp and

Dehydroxilation of HAp under temperature influence is proven by reduction of [OH] absorption maximums intensity. Absorption band of [OH] bonding at 632 cm-1 is very

800 900 1000 1100 1200 1300 1400


<sup>632</sup>3570 3570 3570

1045; 1094;

472; 568; 602; 961

944; 970; 1121 1037; 1045; 1090

472; 568 601; 961

943; 970; 1120

<sup>3595</sup>- - - -

1046; 1091

472; 569; 602; 962

944; 970; 1121

3572;

1046; 1091

472; 570; 602; 963

944; 971 1127

<sup>1212</sup>725;1210 - - - -

875; 1385; 1419; 1457

3572; 633

1091

473; 554-601; 963

946; 975; 1127

HPO42- 875 875 875 - - - -


visible in the spectra already at 200 oC.

T, oC

OH- 3572;

Chemical groups and phases

> CO3 2-

PO43-

PO43- HAp

PO43- -TCP

above mentioned synthesis 1, cannot be used in practice.

875; 1418; 1456; 1466; 1636

H2O adsorbed 3100 - 3600 3250-3600 3330 -

634

473; 550 - 640; 963

> 946; 975

**4.5 Thermal stability of HAp in various environments** 

P2O74- 725; 1211 725;


PO43- 1000 -1120 1006 - 1120 1046;

correspond to a CO3 2- functional group, disappear. At 700oC, a β-TCP phase begins to form, which is shown by characteristic shoulders which become more sharply explicit with increasing of temperature. Until 1100oC, also poorly intensive pyrophosphate (CPP) group bands at 727 and 1212 cm-1, that disappear at highertemperatures.

Since the sample does not contain CPP and HAp any longer, a pure TCP is obtained in the result, and it is stable up to 1400 oC (Fig. 16 and 17). A β-TCP phase can be identified in the spectrum by appearance of characteristic bands at 947 cm-1 and 975 cm-1.

Fig. 16. FTIR spectra of β-TCP (synthesis 4) depending of thermal treatment temperature in range 700 – 1000oC for 1 h

Fig. 17. FTIR spectra of β-TCP (synthesis 4) depending of thermal treatment temperature in range 1100 – 1400oC for 1 h

In the CaP syntheses with final pH=7, T=45oC (synthesis 5), final pH=5,1 and T=70oC (synthesis 2), final pH=5,3 and T=45oC (synthesis 3), spectra demonstrate that a biphasic mixture with various HAp/TCP proportions, respectively 80/20, 60/40 and 35/65, measured at 1100 oC with an XRD semi-quantitative method, is formed in the synthesized samples.

which is shown by characteristic shoulders which become more sharply explicit with increasing of temperature. Until 1100oC, also poorly intensive pyrophosphate (CPP) group

Since the sample does not contain CPP and HAp any longer, a pure TCP is obtained in the result, and it is stable up to 1400 oC (Fig. 16 and 17). A β-TCP phase can be identified in the

Fig. 16. FTIR spectra of β-TCP (synthesis 4) depending of thermal treatment temperature in

Fig. 17. FTIR spectra of β-TCP (synthesis 4) depending of thermal treatment temperature in

In the CaP syntheses with final pH=7, T=45oC (synthesis 5), final pH=5,1 and T=70oC (synthesis 2), final pH=5,3 and T=45oC (synthesis 3), spectra demonstrate that a biphasic mixture with various HAp/TCP proportions, respectively 80/20, 60/40 and 35/65, measured at 1100 oC with an XRD semi-quantitative method, is formed in the synthesized

bands at 727 and 1212 cm-1, that disappear at highertemperatures.

spectrum by appearance of characteristic bands at 947 cm-1 and 975 cm-1.

2- functional group, disappear. At 700oC, a β-TCP phase begins to form,

correspond to a CO3

range 700 – 1000oC for 1 h

range 1100 – 1400oC for 1 h

samples.

In the synthesis with final pH=5,00 and T=22 oC (synthesis 1), a complicated phase composition is formed from β-TCP, CaHPO4 and TTCP phases, and it remains stable up to 1200 oC. CaHPO4 phase can be recognized after a HPO4 2- group band at 897 cm-1, which is visible in the spectra already at 200 oC.

In the synthesis with a final pH=10,74 and T=45oC (synthesis 7), a HAp structure is formed but at 1000oC, and additional OH- peak appears at 3644 cm-1, which means that there is a unreacted CaO left. The reaction did not occur completely and this synthesis, as well as the above mentioned synthesis 1, cannot be used in practice.


Table 9. FTIR absorption bands (cm-1) of thermally treated laboratory synthesized HAp and -TCP (synthesis 6 and 4) chemical groups.

## **4.5 Thermal stability of HAp in various environments**

Dehydroxilation of HAp under temperature influence is proven by reduction of [OH] absorption maximums intensity. Absorption band of [OH] bonding at 632 cm-1 is very

Research of Calcium Phosphates Using Fourier Transform Infrared Spectroscopy 147

of numerous [PO4] groups vibrations occurs, so it is not possible to distinguish, which

600 1 HAp HAp + OAp/TCP 800 1 HAp HAp + OAp/TCP 900 1 HAp HAp + OAp/TCP 1000 1 HAp + α-TCP HAp + OAp/TCP/TTCP 1100 1 HAp + α-TCP + TTCP HAp + TCP + TTCP 1200 1 HAp + α-TCP + TTCP TCP + TTCP 1300 1 α-TCP + TTCP TCP + TTCP

Table 11. Phase composition after thermal treatment of HAp in the vacuum oven.

According to XRD data, supplying water vapor during heating of hydroxyapatite does not change phase composition, however slight changes in the sample structure are detected with FTIR analysis. After thermal treatment in the water vapor, absorption intensity of [OH] group increases in ratio to [PO4], comparing with a sample heated in the air, which could

This summarizing work can help to evaluate both synthesed CaP and structure, phase composition and properties of the CaP bioceramic products. FTIR spectrometry along with XRD is one of the most important, quickest and most available methods for studying CaP materials and in many cases more sensitive than XRD in order to detect forming of new phases. Using these methods, it was possible to detect synthesis parameters in order to obtain a pure and thermally stable HAp and -TCP, as well as biphasic mixtures with controlable ratio. From the results of above mentioned studies it can be concluded that a thorough selection of environment is required for processing HAp powder and it is also required to monitor behavior of the material during the heating, in order to obtain the desirable product. By variating conditions of thermal treatment, it is possible to improve structure of the synthesized HAp, for example, eliminate carbonate groups included in the structure during synthesis, increase number of [OH] groups, as well as slow (by thermal treatment with water vapor presence) or quicken (by thermal treatment in vacuum) HAp

Barinov S.M., Rau J.V., Cesaro S.N. (2006). Carbonate release from carbonated

hydroxyapatite in the wide temperature range// *J. Mater. Sci. Mater. Med.*, Vol.17.,

**Phase composition after XRD after FTIR** 

absorption bands belong to TCP, and which to TTCP phase).

**h** 

improve HAp properties and make this phase more stable.

**Temperature, °C Processing time,** 

**5. Conclusions** 

decomposition.

**6. References** 

pp. 597.–604

sensitive to temperature changes and at higher heating temperatures (over 1300°C) is only shown as shoulder of [PO4] absorption band at 601 cm-1, when [OH] bonding at 3572 cm-1 is more stable in the higher temperatures (Fig. 18).

Fig. 18. FTIR spectra of HAp heated in the air at 1200, 1300, 1450°C (1h)


Table 10. Phase composition after HAp thermal treatment in the air

Stability of hydroxyapatite phase depends on the partial pressure of the water in the atmosphere, so, in an environment with no presence of water, by addition of sufficient amount of energy, HAp will turn into more stable calcium phosphates in the waterless environment. After thermally treating the sample in vacuum, there is an absorption band detected at 948 cm-1 in the FTIR spectrum already at 600 °C. It appears in the result of [PO4] group fluctuations and usually points onto TCP phase, however the literary sources mention that the absorption band at this wave length could also be characteristic to oxyapatite phase. In the result of thermal treatment, in vacuum at 1300 °C, HAp phase has completely decomposed and absorption bands, characteristic to TCP and TTCP phases, are visible in the FTIR spectrum (overlapping


of numerous [PO4] groups vibrations occurs, so it is not possible to distinguish, which absorption bands belong to TCP, and which to TTCP phase).

Table 11. Phase composition after thermal treatment of HAp in the vacuum oven.

According to XRD data, supplying water vapor during heating of hydroxyapatite does not change phase composition, however slight changes in the sample structure are detected with FTIR analysis. After thermal treatment in the water vapor, absorption intensity of [OH] group increases in ratio to [PO4], comparing with a sample heated in the air, which could improve HAp properties and make this phase more stable.

## **5. Conclusions**

146 Infrared Spectroscopy – Materials Science, Engineering and Technology

sensitive to temperature changes and at higher heating temperatures (over 1300°C) is only shown as shoulder of [PO4] absorption band at 601 cm-1, when [OH] bonding at 3572 cm-1 is

> **Phase composition After XRD after FTIR**

1 HAp HAp + TCP 12 HAp HAp + TCP

3 HAp + α-TCP + TTCP HAp + TCP +TTCP

1 HAp + α-TCP + TTCP HAp + TCP +TTCP 3 HAp + α-TCP + TTCP HAp + TCP +TTCP 6 HAp + α-TCP + TTCP HAp + TCP +TTCP

more stable in the higher temperatures (Fig. 18).

**Heating time, h** 

**Temperature, °C** 

1300

1450

Fig. 18. FTIR spectra of HAp heated in the air at 1200, 1300, 1450°C (1h)

Table 10. Phase composition after HAp thermal treatment in the air

1000 8 HAp HAp 1100 1 HAp HAp 1200 1 HAp HAp + TCP

1400 1 HAp + α-TCP + TTCP HAp + TCP

1500 1 HAp + α-TCP + TTCP HAp + TCP +TTCP

Stability of hydroxyapatite phase depends on the partial pressure of the water in the atmosphere, so, in an environment with no presence of water, by addition of sufficient amount of energy, HAp will turn into more stable calcium phosphates in the waterless environment. After thermally treating the sample in vacuum, there is an absorption band detected at 948 cm-1 in the FTIR spectrum already at 600 °C. It appears in the result of [PO4] group fluctuations and usually points onto TCP phase, however the literary sources mention that the absorption band at this wave length could also be characteristic to oxyapatite phase. In the result of thermal treatment, in vacuum at 1300 °C, HAp phase has completely decomposed and absorption bands, characteristic to TCP and TTCP phases, are visible in the FTIR spectrum (overlapping This summarizing work can help to evaluate both synthesed CaP and structure, phase composition and properties of the CaP bioceramic products. FTIR spectrometry along with XRD is one of the most important, quickest and most available methods for studying CaP materials and in many cases more sensitive than XRD in order to detect forming of new phases. Using these methods, it was possible to detect synthesis parameters in order to obtain a pure and thermally stable HAp and -TCP, as well as biphasic mixtures with controlable ratio. From the results of above mentioned studies it can be concluded that a thorough selection of environment is required for processing HAp powder and it is also required to monitor behavior of the material during the heating, in order to obtain the desirable product. By variating conditions of thermal treatment, it is possible to improve structure of the synthesized HAp, for example, eliminate carbonate groups included in the structure during synthesis, increase number of [OH] groups, as well as slow (by thermal treatment with water vapor presence) or quicken (by thermal treatment in vacuum) HAp decomposition.

## **6. References**

Barinov S.M., Rau J.V., Cesaro S.N. (2006). Carbonate release from carbonated hydroxyapatite in the wide temperature range// *J. Mater. Sci. Mater. Med.*, Vol.17., pp. 597.–604

**7** 

*Russia* 

**FTIR Spectroscopy of Adsorbed Probe** 

**Properties of Supported Pt (Pd) Catalysts** 

Supported metal catalysts are important for many fields of applied chemistry, including chemical synthesis, petrochemistry, environmental technology, and energy generation/storage. For prediction of catalyst performance in a chosen reaction and optimization of its functions, it is necessary to know the composition of the surface active sites and have methods for estimating their amount and strength. One of the most available and well-developed methods for studying the composition and structure of the surface functional groups of supported metal catalysts is vibrational spectroscopy, in particular

Although Fourier transform infrared (FTIR) spectroscopy is widely employed for characterization of the catalyst surface (Paukshtis, 1992; Ryczkowski, 2001), it is still unclear whether the regularities obtained under conditions of spectral pretreatments and measurements (evacuation, temperature) can be used for interpreting and predicting the surface properties during adsorption of a precursor or in a catalytic reaction. Thus, aim of the present work is not only to demonstrate the possibilities of FTIR spectroscopy of adsorbed molecules for investigation of the surface functional groups in the chosen catalytic systems, but also to compare FTIR spectroscopy data with the data obtained for supported metal catalysts by other physicochemical methods and with the catalyst properties in model and commercially important reactions. Main emphasis will be made on quantitative determination of various surface groups and elucidation of the effect of their ratio on the

The study was performed with model and commercially important supports and catalysts: gamma alumina, which is among the most popular supports in the synthesis of supported metal catalysts for oil refining, petrochemistry, and gas emissions neutralization; supported platinum and palladium catalysts containing sulfated zirconia (Pt/SZ, Pd/SZ) or alumina-

**1. Introduction** 

with the use of adsorbed probe molecules.

acid-base, adsorption and catalytic properties of the surface.

**Molecules for Analyzing the Surface** 

Alexander V. Lavrenov1 and Vladimir A. Likholobov1,2

Olga B. Belskaya1,2, Irina G. Danilova3,

*1Institute of Hydrocarbons Processing SB RAS* 

*2Omsk State Technical University 3Boreskov Institute of Catalysis SB RAS* 

Maxim O. Kazakov1, Roman M. Mironenko1,


## **FTIR Spectroscopy of Adsorbed Probe Molecules for Analyzing the Surface Properties of Supported Pt (Pd) Catalysts**

Olga B. Belskaya1,2, Irina G. Danilova3, Maxim O. Kazakov1, Roman M. Mironenko1, Alexander V. Lavrenov1 and Vladimir A. Likholobov1,2 *1Institute of Hydrocarbons Processing SB RAS 2Omsk State Technical University 3Boreskov Institute of Catalysis SB RAS Russia* 

## **1. Introduction**

148 Infrared Spectroscopy – Materials Science, Engineering and Technology

Destainville A., Champion E., Bernache-Assollante D., et al. (2003). Synthesis,

Dorozhkin S.V. (2009). Calcium Orthophosphates in Nature, Biology and Medicine.

Dorozhkin S.V. (2009). Calcium orthophosphate-based biocomposites and hybrid

Dorozhkin S.V. (2009). Calcium Orthophosphate Cements and Concretes*. Materials*, 2: pp.

El Kady A.M., K.R.M., El Bassyouni G.T. (2009). Fabrication, characterization and bioactivity evaluation of calcium pyrophosphate/polymeric biocomposites*. Ceram. Int*. Ji G., Zhu H., Jiang X., et. al. (2009). Mechanical Strenght of Epoxy Resin Composites

Han J-K., Song H-Y., et al. (2006). Synthesis of height purity nano-sized hydroxyapatite

Kothapalli C., Wei M., Vasiliev A., et. al. (2004). Influence of temperature and concentration

Kwon S-H., Jun Y-K., Hong S-H., el al. (2003). Synthesis and dissolution behaviour of - TCP

Landi E., Tampieri A., Celotti G., et. al. (2000). Densification behaviour and mechanisms of synthetic hydroxyapatites. *J. Eur. Ceram. Soc.,* Vol. 20, pp. 2377–2387 Lioua S.-C., Chena S.-Y., Lee H.-Y., et. al. (2004). Structural characterization of nano-sized calcium deficient apatite powders. *Biomaterials,* Vol. 25, pp. 189-196. Meejoo S., Maneeprakorn W., Winotai P. (2006). Phase and thermal stability of

Mobasherpour I., Heshajin M. (2007). Synthesis of nanocrystalline hydroxyapatite by using precipitation method. *Journal of Alloys and Compounds*, N 430, pp. 330 – 333 Nilen R.W.N., Richter P.W. (2008). The thermal stability of hydroxyapatite in biphasic calcium phosphate ceramics. *J. Mater. Sci. Mater. Med.,* Vol. 19(4), pp. 1693–1702 Ratner B., Hoffman A., Schoen F. et. al. (2004). *Biomaterials Scienc. An Introduction to Materials* 

Raynaud S., Champion E., Bernache-Assollant D. Et al. (2002). Calcium phosphate apatite

Siva Rama Krishna D., Siddharthan A., Seshadri S. K., et. al. (2007). A novel route for

Ślósarczyk A., Paszkiewicz Z., Zima A. (2010). The effect of phosphate source on the sintering of carbonate substituted hydroxyapatite. *Ceram Int.,* Vol. 36, pp. 577-582

with variable Ca/P atomic ratio I. Synthesis, characterisation and thermal stability

synthesis of nanocrystalline hydroxyapatite from eggshell waste. *J. Mater. Sci. -* 

*in Medicine*, Second Edition // Academic Press, pp. 851

Shi D., (2006). *Introduction to biomaterials. World Scientific Publishing*, p. 253.

of powders*. Biomaterials* No 23, pp. 1065–1072

*Mater. Med.,* Vol.18., pp. 1735-1743

Reinforced by Calcined Pearl Shell Powders*. J. Appl. Polym. Sci*., Vol.114, pp. 3168.-

powder by microwave-hydrothermal method. *Materials Chemistry and Physics*, No.

on the sintering behavior and mechanical properties of hydroxyapatite// Acta

and HA/-TCP composite powders. *Journal of European Ceramic Society*, No. 23, pp.

nanocrystalline hydroxyapatite prepared via microwave heating. *Thermochimica* 

*Materials Chemistry and Physics,* No. 80, pp. 269 – 277

biomaterials. *J. Mater. Sci.,* 44(9): pp. 2343-2387

*Materials*, 2: pp. 399-498

221-291

3176

99, pp 235 – 239

1039–1045

Mater, Vol.52, pp. 5655-5663

*Acta,* No. 447, pp. 115–120

characterization and thermal behaviour of apatite tricalcium phosphate //

Supported metal catalysts are important for many fields of applied chemistry, including chemical synthesis, petrochemistry, environmental technology, and energy generation/storage. For prediction of catalyst performance in a chosen reaction and optimization of its functions, it is necessary to know the composition of the surface active sites and have methods for estimating their amount and strength. One of the most available and well-developed methods for studying the composition and structure of the surface functional groups of supported metal catalysts is vibrational spectroscopy, in particular with the use of adsorbed probe molecules.

Although Fourier transform infrared (FTIR) spectroscopy is widely employed for characterization of the catalyst surface (Paukshtis, 1992; Ryczkowski, 2001), it is still unclear whether the regularities obtained under conditions of spectral pretreatments and measurements (evacuation, temperature) can be used for interpreting and predicting the surface properties during adsorption of a precursor or in a catalytic reaction. Thus, aim of the present work is not only to demonstrate the possibilities of FTIR spectroscopy of adsorbed molecules for investigation of the surface functional groups in the chosen catalytic systems, but also to compare FTIR spectroscopy data with the data obtained for supported metal catalysts by other physicochemical methods and with the catalyst properties in model and commercially important reactions. Main emphasis will be made on quantitative determination of various surface groups and elucidation of the effect of their ratio on the acid-base, adsorption and catalytic properties of the surface.

The study was performed with model and commercially important supports and catalysts: gamma alumina, which is among the most popular supports in the synthesis of supported metal catalysts for oil refining, petrochemistry, and gas emissions neutralization; supported platinum and palladium catalysts containing sulfated zirconia (Pt/SZ, Pd/SZ) or alumina-

FTIR Spectroscopy of Adsorbed Probe Molecules for

a CO molecules and two cation sites simultaneously.

catalyst sample weight (g) in 1 cm2 cross-section of the light flux.

**2.2 Experimental** 

in the integral form:

and the band contour, respectively.

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 151

valence and coordination states of the cations, that is, on their abilities to accept σ-donation (increasing CO) and to donate -orbitals of the CO (decreasing CO, as for carbonyl complexes) (Davydov, 2003; Hadjiivanov & Vayssilov, 2002; Little, 1966). Carbonyls involving -donation can have different structures – linear or bridged – and the number of metal atoms bonded to the CO molecules can also be different. Complexes involving cations only be linear (terminal) because in an M-O-M situation the distance between cations is to great to form a bond between

The FTIR spectra were measured on a Shimadzu FTIR-8300 spectrometer over a range of 700-6000 cm–1 with a resolution of 4 cm–1 and 100 scans for signal accumulation. Before spectra recording, powder samples were pressed into thin self-supporting wafers (8-30 mg/cm2) and activated in a special IR cell under chosen conditions and further in vacuum (р < 10–3 mbar). FTIR spectra are presented in the optical density units referred to a

Quantitative measurements in FTIR spectroscopy are based on the empirical Beer–Lambert– Bouguer law interrelating the intensity of light absorption and the concentration of a substance being analyzed. For FTIR spectroscopy of adsorbed molecules, this law is applied

where A is the integral absorbance (cm–1), T0 and T are the transmittance along the base line,

[ /] , *A S N mol g*

where A0 is the integral absorption coefficient (integral intensity of absorption band (a.b.) for 1 mol of the adsorbate per 1 cm2 cross-section of the light flux), р is the weight of a

The concentration of surface OH groups in -Al2O3 was determined from the integral intensities of absorption bands νОH in the region of 3650-3800 cm–1 using the integral

In the present study, acidic properties of the samples were examined by FTIR spectroscopy using CO adsorption at -196 °C and CO pressure 0.1-10 mbar. An increase in the νСО band frequency of adsorbed CO relative to the value of free CO molecules (2143 cm–1) is caused by the formation of complexes with Lewis or Brønsted acid sites. Complexes with Lewis acid sites are characterized by the bands with a frequency above 2175 cm–1, whereas the frequency range from 2150 through 2175 cm–1 is typical of CO complexes with OH groups. The concentration of Lewis acid sites was measured by the integral intensity of CO band in the range of 2170-2245 cm–1. For alumina and compositions with prevailing fraction of

0

*p A*

The concentration of active sites on the catalyst surface was estimated by the formula

sample wafer (g), and S is the surface area of a sample wafer (cm2).

absorption coefficient A0 = 5.3 cm/mol (Baumgarten et al., 1989).

<sup>0</sup> log( ), *A T T dv <sup>v</sup>* (1)

(2)

promoted SZ (Pt/SZA), which are suitable for low-temperature isomerization of n-alkanes and hydroisomerization of benzene-containing fractions of gasoline.

### **2. Vibrational spectroscopy of adsorbed probe molecules for investigation of supported catalysts – Estimation of the strength and concentration of various surface sites**

### **2.1 FTIR spectra of adsorbed probe molecules**

FTIR spectroscopy of adsorbed probe molecules is one of the most available and welldeveloped methods for studying the composition and structure of the surface functional groups of supported metal catalysts. As the vibrational spectrum reflects both the properties of the molecule as a whole and the characteristic features of separate chemical bonds, FTIR spectroscopy offers the fullest possible information on the perturbation experienced by a molecule on contact with the solid surface, and often determines the structure of adsorption complexes and of surface compounds. Examination of supported metal catalysts deals with two types of surfaces strongly differing in their properties: surface of a support and surface of a metal-containing particle. Various species can reside on the support surface: hydroxyl groups of different nature; Lewis acid sites (coordinatively unsaturated surface cations); base sites (bridging oxygen atoms or oxygen atoms of OH groups); structures formed by impurity anions that remain after the synthesis (sulfate, nitrate and ammonia groups) or form upon contacting with air (carbonate-carboxylate structures).

Various spectroscopic probe molecules are widely used for characterization of Lewis and Brønsted acid sites on the surfaces of oxide catalysts. Among such probes are strong bases: amines, ammonia and pyridine, and weak bases: carbon oxide, carbon dioxide and hydrogen (Knözinger, 1976a; Kubelková et al., 1989; Kustov, 1997; Morterra & Magnacca, 1996; Paukshtis, 1992). Being a weaker base than ammonia, pyridine interacts with the sites widely varying in acidity. However, within each type of Lewis acid site, which is determined with pyridine as a probe molecule, there are distinctions in acidity that cannot be revealed with the use of strong bases. In this connection, very advantageous is the adsorption of weak bases like CO. The application of such probe molecules as CO or pyridine makes it possible to estimate both the concentration and the acid strength of OH groups and Lewis acid sites in zeolites, oxide and other systems (Knözinger, 1976b; Paukshtis, 1992). Concentration of the surface groups accessible for identification by FTIR spectroscopy is above 0.1 mol/g.

In the case of base surface sites, the concentration and strength can be characterized with deuterochloroform (Paukshtis, 1992). Surface of a metal-containing particle may consist of metal atoms with various oxidation states or different charge states caused by the metalsupport interaction. Metal cations and atoms on the surface can be detected only from changes in the spectra of adsorbed molecules, since vibrations of the metal-oxygen bonds on the surface belong to the same spectral region as lattice vibrations and thus are not observed in the measurable spectra, whereas vibrational frequencies of the metal-metal bonds are beyond the measuring range of conventional FTIR spectrometers. Surface atoms and nanoparticles of metals and metal ions are usually identified by the method of spectroscopic probe molecules such as CO (Little, 1966; Sheppard & Nguyen, 1978). Examination of the nature of the binding in Меn+CO complexes suggests that the frequency of adsorbed CO (CO) should depend on the valence and coordination states of the cations, that is, on their abilities to accept σ-donation (increasing CO) and to donate -orbitals of the CO (decreasing CO, as for carbonyl complexes) (Davydov, 2003; Hadjiivanov & Vayssilov, 2002; Little, 1966). Carbonyls involving -donation can have different structures – linear or bridged – and the number of metal atoms bonded to the CO molecules can also be different. Complexes involving cations only be linear (terminal) because in an M-O-M situation the distance between cations is to great to form a bond between a CO molecules and two cation sites simultaneously.

## **2.2 Experimental**

150 Infrared Spectroscopy – Materials Science, Engineering and Technology

promoted SZ (Pt/SZA), which are suitable for low-temperature isomerization of n-alkanes

**2. Vibrational spectroscopy of adsorbed probe molecules for investigation of** 

FTIR spectroscopy of adsorbed probe molecules is one of the most available and welldeveloped methods for studying the composition and structure of the surface functional groups of supported metal catalysts. As the vibrational spectrum reflects both the properties of the molecule as a whole and the characteristic features of separate chemical bonds, FTIR spectroscopy offers the fullest possible information on the perturbation experienced by a molecule on contact with the solid surface, and often determines the structure of adsorption complexes and of surface compounds. Examination of supported metal catalysts deals with two types of surfaces strongly differing in their properties: surface of a support and surface of a metal-containing particle. Various species can reside on the support surface: hydroxyl groups of different nature; Lewis acid sites (coordinatively unsaturated surface cations); base sites (bridging oxygen atoms or oxygen atoms of OH groups); structures formed by impurity anions that remain after the synthesis (sulfate, nitrate and ammonia groups) or

Various spectroscopic probe molecules are widely used for characterization of Lewis and Brønsted acid sites on the surfaces of oxide catalysts. Among such probes are strong bases: amines, ammonia and pyridine, and weak bases: carbon oxide, carbon dioxide and hydrogen (Knözinger, 1976a; Kubelková et al., 1989; Kustov, 1997; Morterra & Magnacca, 1996; Paukshtis, 1992). Being a weaker base than ammonia, pyridine interacts with the sites widely varying in acidity. However, within each type of Lewis acid site, which is determined with pyridine as a probe molecule, there are distinctions in acidity that cannot be revealed with the use of strong bases. In this connection, very advantageous is the adsorption of weak bases like CO. The application of such probe molecules as CO or pyridine makes it possible to estimate both the concentration and the acid strength of OH groups and Lewis acid sites in zeolites, oxide and other systems (Knözinger, 1976b; Paukshtis, 1992). Concentration of the surface groups accessible for identification by FTIR

In the case of base surface sites, the concentration and strength can be characterized with deuterochloroform (Paukshtis, 1992). Surface of a metal-containing particle may consist of metal atoms with various oxidation states or different charge states caused by the metalsupport interaction. Metal cations and atoms on the surface can be detected only from changes in the spectra of adsorbed molecules, since vibrations of the metal-oxygen bonds on the surface belong to the same spectral region as lattice vibrations and thus are not observed in the measurable spectra, whereas vibrational frequencies of the metal-metal bonds are beyond the measuring range of conventional FTIR spectrometers. Surface atoms and nanoparticles of metals and metal ions are usually identified by the method of spectroscopic probe molecules such as CO (Little, 1966; Sheppard & Nguyen, 1978). Examination of the nature of the binding in Меn+CO complexes suggests that the frequency of adsorbed CO (CO) should depend on the

**supported catalysts – Estimation of the strength and concentration of** 

and hydroisomerization of benzene-containing fractions of gasoline.

form upon contacting with air (carbonate-carboxylate structures).

**2.1 FTIR spectra of adsorbed probe molecules** 

**various surface sites** 

spectroscopy is above 0.1 mol/g.

The FTIR spectra were measured on a Shimadzu FTIR-8300 spectrometer over a range of 700-6000 cm–1 with a resolution of 4 cm–1 and 100 scans for signal accumulation. Before spectra recording, powder samples were pressed into thin self-supporting wafers (8-30 mg/cm2) and activated in a special IR cell under chosen conditions and further in vacuum (р < 10–3 mbar). FTIR spectra are presented in the optical density units referred to a catalyst sample weight (g) in 1 cm2 cross-section of the light flux.

Quantitative measurements in FTIR spectroscopy are based on the empirical Beer–Lambert– Bouguer law interrelating the intensity of light absorption and the concentration of a substance being analyzed. For FTIR spectroscopy of adsorbed molecules, this law is applied in the integral form:

$$A \approx \int \log(T\_0/T)\_v dv \,\tag{1}$$

where A is the integral absorbance (cm–1), T0 and T are the transmittance along the base line, and the band contour, respectively.

The concentration of active sites on the catalyst surface was estimated by the formula

$$\text{N} \left[ \mu \text{mol} \mid \text{g} \right] = \frac{A \cdot S}{p \cdot A\_0} \text{.} \tag{2}$$

where A0 is the integral absorption coefficient (integral intensity of absorption band (a.b.) for 1 mol of the adsorbate per 1 cm2 cross-section of the light flux), р is the weight of a sample wafer (g), and S is the surface area of a sample wafer (cm2).

The concentration of surface OH groups in -Al2O3 was determined from the integral intensities of absorption bands νОH in the region of 3650-3800 cm–1 using the integral absorption coefficient A0 = 5.3 cm/mol (Baumgarten et al., 1989).

In the present study, acidic properties of the samples were examined by FTIR spectroscopy using CO adsorption at -196 °C and CO pressure 0.1-10 mbar. An increase in the νСО band frequency of adsorbed CO relative to the value of free CO molecules (2143 cm–1) is caused by the formation of complexes with Lewis or Brønsted acid sites. Complexes with Lewis acid sites are characterized by the bands with a frequency above 2175 cm–1, whereas the frequency range from 2150 through 2175 cm–1 is typical of CO complexes with OH groups.

The concentration of Lewis acid sites was measured by the integral intensity of CO band in the range of 2170-2245 cm–1. For alumina and compositions with prevailing fraction of

FTIR Spectroscopy of Adsorbed Probe Molecules for

attributed to hydrogen-bonded OH groups (Paukshtis, 1992).

pentacoordinated aluminum atom.

crystallites.

**synthesis of Pt/Al2O3 catalysts** 

at low pH of the solution:

respectively

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 153

(Egorov, 1961; Peri, 1965; Tsyganenko & Filimonov, 1973; Zamora & Córdoba, 1978; Knözinger & Ratnasamy, 1978). These models are based on the spinel structure of transitional alumina modifications. Recent attempts to develop advanced models of the surface structure or refine the existing models were made by Tsyganenko and Mardilovich (Tsyganenko & Mardilovich, 1996) as well as Liu and Truitt (Liu & Truitt, 1997). Such advanced models admit the existence of fragments on the alumina surface, which comprise

At present, 7 absorption bands characterizing the isolated OH groups are commonly distinguished in FTIR spectrum of -Аl2O3. Low-frequency bands are assigned to the bridging OH groups located between aluminum atoms with different coordination: 3665-3675 (AlV(OH)AlIV), 3685-3690 cm–1 (AlVI(OH)AlIV), 3700-3710 cm–1 (AlVI(OH)AlV), and 3730-3740 (AlVI(OH)AlVI)1. High-frequency bands correspond to the terminal OH groups bound to one aluminum atom with different coordination: 3745-3758 (AlVIOH), 3765-3776 (AlVOH), and 3785-3792 (AlIVOH) cm–1. In addition, there is a broad a.b. at 3600 cm–1

The surface Lewis acidity is formed by electron-acceptor sites represented by coordinatively unsaturated aluminum cations on the Аl2O3 surface. The use of CO as a probe molecule makes it possible to estimate both the strength and the amount of Lewis acid sites, which is essential when alumina is employed as a catalyst or catalyst support. СО is adsorbed on the γ-Al2O3 surface to form three types of surface complexes (Della Gatta et al., 1976; Zaki & Knözinger, 1987; Zecchina et al., 1987). FTIR spectra of adsorbed CO show the following a.b. corresponding to stretching vibrations of CO molecule: 2180-2205 cm–1 (weak Lewis acid sites), 2205-2220 cm–1 (medium strength Lewis acid sites), and 2220-2245 cm–1 (strong Lewis acid sites). The high-frequency bands are related with two types of Lewis acid sites including AlIV ions located in configurations with crystallographic defects. The lowfrequency bands correspond to AlVI ions in the regular defects of low-index faces of

**3.1.1 FTIR spectroscopy for determining the sites of precursor anchoring during** 

In the synthesis of supported platinum catalysts, chloride complexes of platinum (IV) are commonly used as precursors. Their sorption on the alumina surface occurs from aqueous solutions and implies the involvement of OH groups of the support surface. Therewith, two main mechanisms of the interaction between metal complex and support are considered, implementation of each mechanism depending both on the chemical composition of a complex (degree of hydrolysis) and the ratio of various OH groups (Belskaya et al., 2008, 2011; Bourikas, 2006; Lycourghiotis, 2009). The first mechanism consists in the formation of outer sphere complexes; it implies electrostatic interaction between chloroplatinate and alumina surface, which is protonated and positively charged

Al—OH + H+ + [PtCl6]2- ↔ Al—ОН2+—[PtCl6]2- ↔ Al—[PtCl6]2– + H2O (7)

1AlIV, AlV and AlVI are aluminum atoms in tetrahedral, pentahedral and octahedral coordination,

alumina (Al2O3 and SO42–-ZrO2-Al2O3), the following *A*0 values (cm/mol) were used: 1.25 (2245-2220 cm–1), 1.0 (2200 cm–1), 0.9 (2190 and 2178-2180 cm–1); for zirconia and sulfated zirconia, *A*0 was equal to 0.8 cm/mol (Paukshtis, 1992). A value of the upward νСО frequency shift determines the strength of Lewis acid sites, as it is related to the heat of complex formation by the following formula (Paukshtis, 1992):

$$Q\_{CO} = 10.5 + 0.5 \cdot (\nu\_{CO} - 2143) \tag{3}$$

The concentration of Brønsted acid sites in the modified aluminum oxides was measured by the integral intensity of CO band in the range of 2170-2175 cm–1, *A*0 = 2.6 cm/mol. The concentration of Brønsted acid sites in sulfated samples was determined from the integral intensity of absorption band due to the pyridinium ion with a maximum at 1544 cm–1 (*A*0 = 3.5 cm/μmol) (Paukshtis, 1992). Adsorption of carbon monoxide on Brønsted acid sites at -196 °C results in the shift of OH bands to the lower frequency region due to perturbation of OH stretch by hydrogen bonding with CO molecule. The higher the shift of OH stretching vibration, the stronger the acidity of this OH group (Maache et al., 1993; Paze et al., 1997).

Basic properties of the samples were studied by FTIR spectroscopy using CDCl3 adsorption at 20 °C. At the formation of H bonds, deuterochloroform behaves as a typical acid. A decrease in the frequency of νСD band of CDCl3 adsorbed relative to the value of physisorbed molecules (2265 cm–1) is caused by the formation of complexes with base sites. The strength of base site was determined by the band shift of the CD stretching vibrations that occurred under CDCl3 adsorption. The strength can be recalculated into the proton affinity (PA) scale using the formula (Paukshtis, 1992):

$$0.\log(\Delta \nu\_{\rm CD}) = 0.0066PA - 4.36\tag{4}$$

The concentration of base sites was measured by the integral intensity of CD band in the range of 2190-2255 cm–1. The integral absorption coefficient was calculated from the correlation equations (Paukshtis, 1992):

$$A\_0 = 0.375 + 0.0158 \Delta \nu\_{\rm CD} \text{ for } \Delta \nu\_{\rm CD} > 13 \text{ cm}^{-1} \tag{5}$$

$$A\_0 = 0.125 + 0.0034 \Delta \nu\_{\rm CD} \text{ for } \Delta \nu\_{\rm CD} < 13 \text{ cm}^{-1} \tag{6}$$

#### **3. Investigation of supports and catalysts by FTIR spectroscopy**

#### **3.1 Gamma alumina. The role of studying the surface functional groups for understanding the processes of adsorption and catalysis**

Owing to its unique acid-base and structural properties, aluminum oxide, first of all γ-Al2O3, remains the most popular catalyst and catalyst support. The analysis of catalytic reactions usually deals with Lewis acid and base sites of Al2O3. However, in the catalyst synthesis, adsorption properties of the surface during its interaction with aqueous solutions strongly determine the composition of surface hydroxyl cover of alumina. It should be noted that modern concepts of the surface structure of aluminum oxides, which were developed in recent 50 years, are based mainly on the vibrational spectroscopy data. Various structural models of the aluminum oxide surface were suggested to explain the experimental data

alumina (Al2O3 and SO42–-ZrO2-Al2O3), the following *A*0 values (cm/mol) were used: 1.25 (2245-2220 cm–1), 1.0 (2200 cm–1), 0.9 (2190 and 2178-2180 cm–1); for zirconia and sulfated zirconia, *A*0 was equal to 0.8 cm/mol (Paukshtis, 1992). A value of the upward νСО frequency shift determines the strength of Lewis acid sites, as it is related to the heat of

> 10.5 0.5 ( 2143) *QCO*

The concentration of Brønsted acid sites in the modified aluminum oxides was measured by the integral intensity of CO band in the range of 2170-2175 cm–1, *A*0 = 2.6 cm/mol. The concentration of Brønsted acid sites in sulfated samples was determined from the integral intensity of absorption band due to the pyridinium ion with a maximum at 1544 cm–1 (*A*0 = 3.5 cm/μmol) (Paukshtis, 1992). Adsorption of carbon monoxide on Brønsted acid sites at -196 °C results in the shift of OH bands to the lower frequency region due to perturbation of OH stretch by hydrogen bonding with CO molecule. The higher the shift of OH stretching vibration, the stronger the acidity of this OH group (Maache et al., 1993; Paze et al., 1997).

Basic properties of the samples were studied by FTIR spectroscopy using CDCl3 adsorption at 20 °C. At the formation of H bonds, deuterochloroform behaves as a typical acid. A decrease in the frequency of νСD band of CDCl3 adsorbed relative to the value of physisorbed molecules (2265 cm–1) is caused by the formation of complexes with base sites. The strength of base site was determined by the band shift of the CD stretching vibrations that occurred under CDCl3 adsorption. The strength can be recalculated into the proton

log( ) 0.0066 4.36

The concentration of base sites was measured by the integral intensity of CD band in the range of 2190-2255 cm–1. The integral absorption coefficient was calculated from the

Owing to its unique acid-base and structural properties, aluminum oxide, first of all γ-Al2O3, remains the most popular catalyst and catalyst support. The analysis of catalytic reactions usually deals with Lewis acid and base sites of Al2O3. However, in the catalyst synthesis, adsorption properties of the surface during its interaction with aqueous solutions strongly determine the composition of surface hydroxyl cover of alumina. It should be noted that modern concepts of the surface structure of aluminum oxides, which were developed in recent 50 years, are based mainly on the vibrational spectroscopy data. Various structural models of the aluminum oxide surface were suggested to explain the experimental data

 *CD* for <sup>1</sup> 13 

 *CD* for <sup>1</sup> 13 

<sup>0</sup> 0.375 0.0158 *A*

<sup>0</sup> 0.125 0.0034 *A*

**understanding the processes of adsorption and catalysis** 

**3. Investigation of supports and catalysts by FTIR spectroscopy 3.1 Gamma alumina. The role of studying the surface functional groups for** 

*CO* (3)

*CD PA* (4)

*CD cm* (5)

*CD cm* (6)

complex formation by the following formula (Paukshtis, 1992):

affinity (PA) scale using the formula (Paukshtis, 1992):

correlation equations (Paukshtis, 1992):

(Egorov, 1961; Peri, 1965; Tsyganenko & Filimonov, 1973; Zamora & Córdoba, 1978; Knözinger & Ratnasamy, 1978). These models are based on the spinel structure of transitional alumina modifications. Recent attempts to develop advanced models of the surface structure or refine the existing models were made by Tsyganenko and Mardilovich (Tsyganenko & Mardilovich, 1996) as well as Liu and Truitt (Liu & Truitt, 1997). Such advanced models admit the existence of fragments on the alumina surface, which comprise pentacoordinated aluminum atom.

At present, 7 absorption bands characterizing the isolated OH groups are commonly distinguished in FTIR spectrum of -Аl2O3. Low-frequency bands are assigned to the bridging OH groups located between aluminum atoms with different coordination: 3665-3675 (AlV(OH)AlIV), 3685-3690 cm–1 (AlVI(OH)AlIV), 3700-3710 cm–1 (AlVI(OH)AlV), and 3730-3740 (AlVI(OH)AlVI)1. High-frequency bands correspond to the terminal OH groups bound to one aluminum atom with different coordination: 3745-3758 (AlVIOH), 3765-3776 (AlVOH), and 3785-3792 (AlIVOH) cm–1. In addition, there is a broad a.b. at 3600 cm–1 attributed to hydrogen-bonded OH groups (Paukshtis, 1992).

The surface Lewis acidity is formed by electron-acceptor sites represented by coordinatively unsaturated aluminum cations on the Аl2O3 surface. The use of CO as a probe molecule makes it possible to estimate both the strength and the amount of Lewis acid sites, which is essential when alumina is employed as a catalyst or catalyst support. СО is adsorbed on the γ-Al2O3 surface to form three types of surface complexes (Della Gatta et al., 1976; Zaki & Knözinger, 1987; Zecchina et al., 1987). FTIR spectra of adsorbed CO show the following a.b. corresponding to stretching vibrations of CO molecule: 2180-2205 cm–1 (weak Lewis acid sites), 2205-2220 cm–1 (medium strength Lewis acid sites), and 2220-2245 cm–1 (strong Lewis acid sites). The high-frequency bands are related with two types of Lewis acid sites including AlIV ions located in configurations with crystallographic defects. The lowfrequency bands correspond to AlVI ions in the regular defects of low-index faces of crystallites.

#### **3.1.1 FTIR spectroscopy for determining the sites of precursor anchoring during synthesis of Pt/Al2O3 catalysts**

In the synthesis of supported platinum catalysts, chloride complexes of platinum (IV) are commonly used as precursors. Their sorption on the alumina surface occurs from aqueous solutions and implies the involvement of OH groups of the support surface. Therewith, two main mechanisms of the interaction between metal complex and support are considered, implementation of each mechanism depending both on the chemical composition of a complex (degree of hydrolysis) and the ratio of various OH groups (Belskaya et al., 2008, 2011; Bourikas, 2006; Lycourghiotis, 2009). The first mechanism consists in the formation of outer sphere complexes; it implies electrostatic interaction between chloroplatinate and alumina surface, which is protonated and positively charged at low pH of the solution:

$$\sim \text{Al}-\text{OH} + \text{H}^\* + [\text{PtCl}\_6]^{2-} \leftrightarrow \sim \text{Al}-\text{OH} \cdot \text{2}^\* - [\text{PtCl}\_6]^{2-} \leftrightarrow \sim \text{Al}-[\text{PtCl}\_6]^{2-} + \text{H}\_2\text{O} \tag{7}$$

 1AlIV, AlV and AlVI are aluminum atoms in tetrahedral, pentahedral and octahedral coordination, respectively

FTIR Spectroscopy of Adsorbed Probe Molecules for

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 155

experimental data presented in Fig. 1 and Table 1 shows that adsorption of platinum complexes followed by anchoring of oxide platinum species forming on the -Al2O3 surface decreases the intensity only of high-frequency bands. As the amount of supported platinum species increases (platinum content of 0.5 and 1.0 wt%), the concentration of all types of terminal groups (AlVOH, AlVIOH, AlIVOH) decreases; so does the concentration of bridging OH groups AlVI(OH)AlVI bonded to octahedral aluminum. Exactly these types of OH groups seem to be involved in anchoring the anionic complexes of platinum (IV). Having more strong base properties, they are capable of interacting with chloroplatinate by mechanism (8), acting as the attacking ligand. However, the bridging groups with a.b. 3665-3710 cm–1 (AlVI(OH)AlIV,

AlV(OH)AlIV, AlVI(OH)AlV) are virtually not involved in platinum anchoring (Table 1).

complexes via electrostatic interaction (Kwak et al., 2009; Mei et al., 2010).


adsorbed CO

**alumina** 

According to analysis of the spectra of adsorbed CO, in the region of CO stretching vibrations all the samples have a.b. at 2245 and 2238 cm–1 characterizing CO complexes with strong Lewis acid sites, absorption bands at 2220 and 2205 cm–1 characterizing CO complexes with medium strength Lewis acid sites, and absorption bands at 2189-2191 cm–1 characterizing CO complex with weak Lewis acid sites. Deposition of platinum raises the concentration of nearly all types of Lewis acid sites, which is related with introduction of Cl– ion of the complex. Along with this, a substantial decrease in the concentration of weak Lewis acid sites is observed (Table 2). These electron-deficient sites may also take part in the anchoring of anionic platinum

Type of Lewis acid site Super strong Strong Medium I Medium II Weak

Samples Concentration, mol/g

Table 2. Types and concentrations of Lewis acid sites according to FTIR spectroscopy of

mechanism and strength of the metal complex-support interaction.

Thus, the analysis of FTIR spectra of alumina provides data on the nature and amount of various surface sites; moreover, it allows identification of the sites where active component precursor is anchored during catalyst synthesis, and makes it possible to hypothesize about

**3.1.2 Novel approaches to varying the composition of surface functional groups of** 

Variation of the acid-base properties of alumina surface is commonly performed by chemical modifying via the introduction of additional anions (halogens, sulfates, phosphates) or cations (alkaline or alkaline earth metals) (Bocanegra et al., 2006; Ghosh & Kydd, 1985; Lisboa et al., 2005; López Cordero et al., 1989; Marceau et al., 1996; Requies et

CO, cm-1 2245 2238 2220 2205 2189-2191 QCO, kJ/mol 61.5 58 48.5 41.5 34


1.1 4.2 5 24 470

The formation of inner sphere complexes is accompanied by a deeper interaction of a complex with the oxide surface via ligand exchange with the surface OH groups:

$$\sim \text{Al}-\text{OH} + [\text{PtCl}\_6] \\ \text{2} \cdot \leftrightarrow \sim \text{Al}-[\text{(OH)}-\text{PtCl}\_5] \cdot + \text{Cl}\text{-}\tag{8}$$

FTIR spectroscopy is widely used to investigate state of the surface at different stages of catalyst synthesis. When studying the precursor-support interaction, this method allows identification of OH groups involved in chemisorption of the metal complex. Analysis of

Fig. 1. FTIR spectra of surface hydroxyl groups of -Al2O3 (a), 0.5% Pt/-Al2O3 (b) and 1% Pt/-Al2O3 (c). The samples were calcined and outgassed at 500 °C


Table 1. Types and concentrations of hydroxyl groups in calcined alumina and Pt/alumina samples as determined by FTIR spectroscopy data

The formation of inner sphere complexes is accompanied by a deeper interaction of a

FTIR spectroscopy is widely used to investigate state of the surface at different stages of catalyst synthesis. When studying the precursor-support interaction, this method allows identification of OH groups involved in chemisorption of the metal complex. Analysis of

**3500 3600 3700 3800**

Fig. 1. FTIR spectra of surface hydroxyl groups of -Al2O3 (a), 0.5% Pt/-Al2O3 (b) and 1%

(3775 cm-1) 35 29 16 33

(3758 cm-1) 34 27 20 28

(3730-3740 cm-1) 107 96 78 89

(3705-3710 cm-1) 50 54 52 52

(3665-3670 cm-1) 36 35 34 43 OH 336 325 275 313 Table 1. Types and concentrations of hydroxyl groups in calcined alumina and Pt/alumina

(3690 cm-1) 62 75 67 62

**Wavenumber, cm-1**

Al—OH + [PtCl6]2- ↔ Al—[(ОН)—PtCl5]– +Cl– (8)

**3730**

**3758**


12 9 8 6

**3775**

**3792**

**(c) (b)**

> Hydrothermal treatment 180 °C, 3 h

**(a)**

complex with the oxide surface via ligand exchange with the surface OH groups:

*10 a.u.*

Pt/-Al2O3 (c). The samples were calcined and outgassed at 500 °C

**IR absorbance**

Sample

Concentration, mol/g

samples as determined by FTIR spectroscopy data

Type of OH group

AlIVOH (3790-3795 cm-1)

AlVOH

AlVIOH

AlVI(OH)AlVI

AlV(OH)AlVI

AlVI(OH)AlIV

AlV(OH)AlIV

experimental data presented in Fig. 1 and Table 1 shows that adsorption of platinum complexes followed by anchoring of oxide platinum species forming on the -Al2O3 surface decreases the intensity only of high-frequency bands. As the amount of supported platinum species increases (platinum content of 0.5 and 1.0 wt%), the concentration of all types of terminal groups (AlVOH, AlVIOH, AlIVOH) decreases; so does the concentration of bridging OH groups AlVI(OH)AlVI bonded to octahedral aluminum. Exactly these types of OH groups seem to be involved in anchoring the anionic complexes of platinum (IV). Having more strong base properties, they are capable of interacting with chloroplatinate by mechanism (8), acting as the attacking ligand. However, the bridging groups with a.b. 3665-3710 cm–1 (AlVI(OH)AlIV, AlV(OH)AlIV, AlVI(OH)AlV) are virtually not involved in platinum anchoring (Table 1).

According to analysis of the spectra of adsorbed CO, in the region of CO stretching vibrations all the samples have a.b. at 2245 and 2238 cm–1 characterizing CO complexes with strong Lewis acid sites, absorption bands at 2220 and 2205 cm–1 characterizing CO complexes with medium strength Lewis acid sites, and absorption bands at 2189-2191 cm–1 characterizing CO complex with weak Lewis acid sites. Deposition of platinum raises the concentration of nearly all types of Lewis acid sites, which is related with introduction of Cl– ion of the complex. Along with this, a substantial decrease in the concentration of weak Lewis acid sites is observed (Table 2). These electron-deficient sites may also take part in the anchoring of anionic platinum complexes via electrostatic interaction (Kwak et al., 2009; Mei et al., 2010).


Table 2. Types and concentrations of Lewis acid sites according to FTIR spectroscopy of adsorbed CO

Thus, the analysis of FTIR spectra of alumina provides data on the nature and amount of various surface sites; moreover, it allows identification of the sites where active component precursor is anchored during catalyst synthesis, and makes it possible to hypothesize about mechanism and strength of the metal complex-support interaction.

#### **3.1.2 Novel approaches to varying the composition of surface functional groups of alumina**

Variation of the acid-base properties of alumina surface is commonly performed by chemical modifying via the introduction of additional anions (halogens, sulfates, phosphates) or cations (alkaline or alkaline earth metals) (Bocanegra et al., 2006; Ghosh & Kydd, 1985; Lisboa et al., 2005; López Cordero et al., 1989; Marceau et al., 1996; Requies et

FTIR Spectroscopy of Adsorbed Probe Molecules for

concentration of other types of Lewis acid sites.

*10 a.u.*

**(b)**

components. Inset: portion of spectrum (b) magnified 10 times

**(a)**

**IR absorbance**

the FTIR spectrum of adsorbed CO

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 157

However, the formation of aluminum oxide compounds having their own surface OH groups (Al2O3/γ-Al2O3) resulted in substantial changes in the adsorption and acidic properties of the γ-Al2O3 surface. Thus, investigation of the sorption of chloride complexes of platinum (IV) on the modified γ-Al2O3 showed a 1.5-fold increase in the sorption capacity and an increased strength of the metal complex-support interaction (Mironenko et al., 2009). According to FTIR spectroscopy of adsorbed CO (Fig. 3, Table 3), the anchoring of aluminum oxide compounds on the surface of initial γ-Al2O3 support decreased the concentration of weak Lewis acid sites with a.b. СО = 2191 cm–1 (regular defects of the alumina surface including the octahedral aluminum ion) without changes in the

**2130 2160 2190 2220**

**2179**

**2162**

**Wavenumber, cm-1**

Fig. 3. FTIR spectra of CO adsorbed at -196 °C and a CO pressure of 4 mbar: (a) γ-Al2O3, (b) 3%Al2O3/γ-Al2O3. The dashed line shows the deconvolution of spectrum (b) into its

γ-Al2O3 155 360 8 3 526 3%Al2O3/γ-Al2O3 155 330 7 2 494 Table 3. Concentrations of Lewis acid sites characterized by different absorption bands in

A decrease in the surface acidity revealed by FTIR spectroscopy of adsorbed CO is in good agreement with the results of catalytic testing (Fig. 4). Alumina samples before and after modifying were compared in a model reaction of 1-hexene double-bond isomerization, which is sensitive to amount and strength of Lewis acid sites. Although the reaction conditions radically differ from conditions of spectral measurements, the observed decrease in 1-hexene conversion can be predicted and interpreted using FTIR spectroscopy data.

Sample Lewis acid site concentration, mol/g

**2207**

2179 cm–1 2191 cm–1 2207 cm–1 2225 cm–1 Σ

**2225**

 **x 10**

**2191**

al., 2006; Rombi et al., 2003; Scokart et al., 1979; Wang et al., 1994). These methods complicate the process of catalyst synthesis and can lead to non-reproducible results.

This Section presents some unconventional approaches to alumina modifying for controlling the state of its surface functional cover. One of approaches consists in altering the relative content of hydroxyl groups and Lewis acid sites on the γ-Al2O3 surface without changes in the chemical composition of support (Mironenko et al., 2009, 2011). For this purpose, two techniques are employed: chemisorption of aluminum oxalate complexes followed by their thermal decomposition, and hydrothermal treatment of γ-Al2O3. Besides, there is an approach leading to considerable enhancement of acidic properties of the surface. Such effect is provided by γ-Al2O3 promotion with silica. The formation of SiO2 takes place in the pore space of alumina during thermal decomposition of preliminarily introduced siliconcontaining precursor.

#### **3.1.2.1 Modifying the functional cover of the γ-Al2O3 surface using aluminum oxalate complexes**

The proposed method of modifying implies the chemisorption (in distinction to conventional methods of incipient wetness impregnation) of anionic aluminum oxalate complexes [Al(С2О4)2(H2O)2]– and [Al(С2О4)3]3– on the -Al2O3 surface. It is essential that this approach excludes both the formation of a bulk alumina phase in the porous space after decomposition of supported complexes, and considerable changes in the texture parameters. Analysis of FTIR spectra of the modified alumina surface hydroxyl cover revealed (Fig. 2) that chemisorption of the oxalate complexes and subsequent formation of aluminum oxide compounds supported on -Al2O3 (calcination at 550 °C) decreased mainly the intensity of two a.b. at 3670 and 3775 cm–1. These bands characterize the bridging and terminal OH groups bound to pentacoordinated aluminum atom. Probably these are exactly the groups that are involved in anchoring of aluminum oxalate complexes.

Fig. 2. FTIR spectra of surface hydroxyl groups of -Al2O3 (a) and 3%Al2O3/γ-Al2O3 (b). The samples were calcined and outgassed at 500 °C

al., 2006; Rombi et al., 2003; Scokart et al., 1979; Wang et al., 1994). These methods

This Section presents some unconventional approaches to alumina modifying for controlling the state of its surface functional cover. One of approaches consists in altering the relative content of hydroxyl groups and Lewis acid sites on the γ-Al2O3 surface without changes in the chemical composition of support (Mironenko et al., 2009, 2011). For this purpose, two techniques are employed: chemisorption of aluminum oxalate complexes followed by their thermal decomposition, and hydrothermal treatment of γ-Al2O3. Besides, there is an approach leading to considerable enhancement of acidic properties of the surface. Such effect is provided by γ-Al2O3 promotion with silica. The formation of SiO2 takes place in the pore space of alumina during thermal decomposition of preliminarily introduced silicon-

complicate the process of catalyst synthesis and can lead to non-reproducible results.

**3.1.2.1 Modifying the functional cover of the γ-Al2O3 surface using aluminum oxalate** 

that are involved in anchoring of aluminum oxalate complexes.

*10 a.u.*

**IR absorbance**

samples were calcined and outgassed at 500 °C

The proposed method of modifying implies the chemisorption (in distinction to conventional methods of incipient wetness impregnation) of anionic aluminum oxalate complexes [Al(С2О4)2(H2O)2]– and [Al(С2О4)3]3– on the -Al2O3 surface. It is essential that this approach excludes both the formation of a bulk alumina phase in the porous space after decomposition of supported complexes, and considerable changes in the texture parameters. Analysis of FTIR spectra of the modified alumina surface hydroxyl cover revealed (Fig. 2) that chemisorption of the oxalate complexes and subsequent formation of aluminum oxide compounds supported on -Al2O3 (calcination at 550 °C) decreased mainly the intensity of two a.b. at 3670 and 3775 cm–1. These bands characterize the bridging and terminal OH groups bound to pentacoordinated aluminum atom. Probably these are exactly the groups

**3500 3600 3700 3800**

Fig. 2. FTIR spectra of surface hydroxyl groups of -Al2O3 (a) and 3%Al2O3/γ-Al2O3 (b). The

**Wavenumber, cm-1**

**3710**

**3690**

**3670**

**3730**

**3758**

**3775**

**3792**

**(b) (a)**

containing precursor.

**complexes** 

However, the formation of aluminum oxide compounds having their own surface OH groups (Al2O3/γ-Al2O3) resulted in substantial changes in the adsorption and acidic properties of the γ-Al2O3 surface. Thus, investigation of the sorption of chloride complexes of platinum (IV) on the modified γ-Al2O3 showed a 1.5-fold increase in the sorption capacity and an increased strength of the metal complex-support interaction (Mironenko et al., 2009).

According to FTIR spectroscopy of adsorbed CO (Fig. 3, Table 3), the anchoring of aluminum oxide compounds on the surface of initial γ-Al2O3 support decreased the concentration of weak Lewis acid sites with a.b. СО = 2191 cm–1 (regular defects of the alumina surface including the octahedral aluminum ion) without changes in the concentration of other types of Lewis acid sites.

Fig. 3. FTIR spectra of CO adsorbed at -196 °C and a CO pressure of 4 mbar: (a) γ-Al2O3, (b) 3%Al2O3/γ-Al2O3. The dashed line shows the deconvolution of spectrum (b) into its components. Inset: portion of spectrum (b) magnified 10 times


Table 3. Concentrations of Lewis acid sites characterized by different absorption bands in the FTIR spectrum of adsorbed CO

A decrease in the surface acidity revealed by FTIR spectroscopy of adsorbed CO is in good agreement with the results of catalytic testing (Fig. 4). Alumina samples before and after modifying were compared in a model reaction of 1-hexene double-bond isomerization, which is sensitive to amount and strength of Lewis acid sites. Although the reaction conditions radically differ from conditions of spectral measurements, the observed decrease in 1-hexene conversion can be predicted and interpreted using FTIR spectroscopy data.

FTIR Spectroscopy of Adsorbed Probe Molecules for

*10 a.u.*

 **0.05 mmol Pt/g**

**Adsorption**

treatment at 550 °С

**IR absorbance**

at 500 °C

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 159

**3500 3600 3700 3800**

treatment for 3 h at 150 (b), 180 (c) and 200 °C (d). The samples were calcined and outgassed

**0 5 10 15 20**

Fig. 6. Isotherms of H2[PtCl6] adsorption from aqueous solutions on γ-Al2O3 (a) and γ-Al2O3 after hydrothermal treatment for 3 h at 150 (b), 180 (c) and 200 °C (d) followed by heat

Analysis of the FTIR spectra of adsorbed CO shows (Table 2) that introduction of the hydrothermal treatment step increases the concentration of all Lewis acid site types: weak (2180 and 2190 cm–1), medium strength (2205 cm–1), and strong (2225 cm–1). The obtained result is important for application of this modifying method in the synthesis of catalytic

compositions for the reactions requiring the presence of acid sites.

**CPt, mmol/l**

Fig. 5. FTIR spectra of hydroxyl cover of γ-Al2O3 (a) and γ-Al2O3 after hydrothermal

**Wavenumber, cm-1**

**0.22**

**(b)**

**(a)**

**0.24**

**(d) (c) (b)**

**(a)**

**3792**

**0.10**

**0.12**

**(c) (d)**

**3775**

**3758**

**3730**

**3710**

**3690**

**3670**

Fig. 4. Temperature dependence of the 1-hexene conversion: (a) γ-Al2O3, (b) 3%Al2O3/γ-Al2O3. Reaction conditions: atmospheric pressure, T = 90-110 °C, He : C6H12 molar ratio 3.3

#### **3.1.2.2 Modifying the functional cover of the γ-Al2O3 surface at hydrothermal treatment**

Hydrothermal treatment of γ-Al2O3 is commonly used for alteration of the porous structure parameters (Chertov et al., 1982). Our study demonstrated that this technique is efficient for controlling the state of the oxide surface. Hydrothermal treatment of γ-Al2O3 was carried out in a temperature range of 50-200 °C with the treatment time varying from 0.5 to 12 h. This produced a hydroxide phase of boehmite AlO(OH) on the γ-Al2O3 surface, which amount can be readily controlled by the treatment conditions. After hydrothermal treatment, the samples were calcined at 550 °C to reduce the oxide phase.

The FTIR spectroscopic examination revealed the hydrothermal treatment effect on the concentration and ratio of functional groups on the -Аl2O3 surface. Figure 5 shows FTIR spectra of the surface hydroxyl cover of initial γ-Al2O3 and γ-Al2O3 subjected to hydrothermal treatment at various temperatures with subsequent calcination at 550 °C. The quantitative analysis of FTIR spectroscopy data (Fig. 5 and Table 1) showed changes in the relative content of different surface OH groups of γ-Al2O3 with elevation of hydrothermal treatment temperature. This modifying technique was found to increase the fraction of lowfrequency bridging hydroxyl groups (ОН = 3710-3670 cm–1) from 50 to 70% of all OH groups and decrease the content of terminal hydroxyl groups (ОН = 3790-3760 cm–1) and especially the bridging group AlVI(OH)AlVI (ОН = 3730 cm–1). A decrease in the content of basic OH groups necessary for the anchoring of chloride platinum complexes decreased the adsorptivity of support with respect to [PtCl6]2–. The adsorption isotherms of H2[PtCl6] on the support pretreated at different temperatures (Fig. 6) demonstrate that the difference in adsorptivity can be quite high (more than a twofold), and calcination restoring the oxide phase cannot restore the relative content of functional groups of the surface and its adsorption properties in aqueous solutions. The presented experimental data illustrate that conclusions on the state of the surface obtained by FTIR spectroscopy can reflect and explain the processes occurring at the solid-liquid interface, i.e. under real conditions of the catalyst synthesis.

<sup>+</sup> <sup>H</sup>

*2 %*

**1-Hexene conversion**

synthesis.

**5.1 5.9**

**90 100 110**

**С**

**Temperature, 0**

Fig. 4. Temperature dependence of the 1-hexene conversion: (a) γ-Al2O3, (b) 3%Al2O3/γ-Al2O3.

**3.1.2.2 Modifying the functional cover of the γ-Al2O3 surface at hydrothermal treatment**  Hydrothermal treatment of γ-Al2O3 is commonly used for alteration of the porous structure parameters (Chertov et al., 1982). Our study demonstrated that this technique is efficient for controlling the state of the oxide surface. Hydrothermal treatment of γ-Al2O3 was carried out in a temperature range of 50-200 °C with the treatment time varying from 0.5 to 12 h. This produced a hydroxide phase of boehmite AlO(OH) on the γ-Al2O3 surface, which amount can be readily controlled by the treatment conditions. After hydrothermal

The FTIR spectroscopic examination revealed the hydrothermal treatment effect on the concentration and ratio of functional groups on the -Аl2O3 surface. Figure 5 shows FTIR spectra of the surface hydroxyl cover of initial γ-Al2O3 and γ-Al2O3 subjected to hydrothermal treatment at various temperatures with subsequent calcination at 550 °C. The quantitative analysis of FTIR spectroscopy data (Fig. 5 and Table 1) showed changes in the relative content of different surface OH groups of γ-Al2O3 with elevation of hydrothermal treatment temperature. This modifying technique was found to increase the fraction of lowfrequency bridging hydroxyl groups (ОН = 3710-3670 cm–1) from 50 to 70% of all OH groups and decrease the content of terminal hydroxyl groups (ОН = 3790-3760 cm–1) and especially the bridging group AlVI(OH)AlVI (ОН = 3730 cm–1). A decrease in the content of basic OH groups necessary for the anchoring of chloride platinum complexes decreased the adsorptivity of support with respect to [PtCl6]2–. The adsorption isotherms of H2[PtCl6] on the support pretreated at different temperatures (Fig. 6) demonstrate that the difference in adsorptivity can be quite high (more than a twofold), and calcination restoring the oxide phase cannot restore the relative content of functional groups of the surface and its adsorption properties in aqueous solutions. The presented experimental data illustrate that conclusions on the state of the surface obtained by FTIR spectroscopy can reflect and explain the processes occurring at the solid-liquid interface, i.e. under real conditions of the catalyst

Reaction conditions: atmospheric pressure, T = 90-110 °C, He : C6H12 molar ratio 3.3

treatment, the samples were calcined at 550 °C to reduce the oxide phase.

**6.6**

**9.4**

H H H

**6.9**

 **(a) (b)**

**11.3**

Fig. 5. FTIR spectra of hydroxyl cover of γ-Al2O3 (a) and γ-Al2O3 after hydrothermal treatment for 3 h at 150 (b), 180 (c) and 200 °C (d). The samples were calcined and outgassed at 500 °C

Fig. 6. Isotherms of H2[PtCl6] adsorption from aqueous solutions on γ-Al2O3 (a) and γ-Al2O3 after hydrothermal treatment for 3 h at 150 (b), 180 (c) and 200 °C (d) followed by heat treatment at 550 °С

Analysis of the FTIR spectra of adsorbed CO shows (Table 2) that introduction of the hydrothermal treatment step increases the concentration of all Lewis acid site types: weak (2180 and 2190 cm–1), medium strength (2205 cm–1), and strong (2225 cm–1). The obtained result is important for application of this modifying method in the synthesis of catalytic compositions for the reactions requiring the presence of acid sites.

FTIR Spectroscopy of Adsorbed Probe Molecules for

Si(OH) groups.

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 161

Meanwhile, in the case of modified support (Fig. 7(b)), there is an intense peak in the spectrum even at a two times lower platinum content corresponding to chloride complex [PtCl6]2–. Hence, anchoring of the complex does not produce noticeable changes in its chemical composition; moreover, there is a considerable decrease in the contribution of coordination binding with the surface involving OH groups of the support. Electrostatic interaction of a metal complex with the modified support with respect to equation (7) seems to prevail here. In addition, temperature-programmed reduction followed by chemisorption of H2 and CO probe molecules was used to confirm that a decrease in the bond strength between precursor and support decreases the reduction temperature of adsorbed platinum species and diminishes the dispersion of supported particles by a factor of more than 3.

**3.1.2.3 The effect of -Аl2O3 modifying with silica on acid-base properties of the surface**  Due to its high thermal stability, alumina modified with silica is a promising support for exhaust neutralization catalysts. Acid-base properties of the developed composition can be optimized by means of FTIR spectroscopy of adsorbed probe molecules. Composition of the support was varied by changing the silica concentration from 1.5 to 10 wt%. The modifying SiO2 compound was formed in the pore space of -Аl2O3 support during thermal

Figure 8 shows FTIR spectra in the region of OH group stretching vibrations of the aluminum oxide samples modified with silica in comparison with the spectrum of initial -Аl2O3. The spectra of all modified oxides show a decrease in intensity of absorption bands corresponding to OH groups of the bridging and terminal types as compared to initial -Аl2O3. The additional absorption bands appear at 3740-3745 cm–1, their intensity growing with silicon content of the sample. The band at 3745 cm–1 can be assigned to the terminal

**3500 3600 3700 3800**

**Wavenumber, cm-1** Fig. 8. FTIR spectra of hydroxyl groups for -Al2O3 calcined at 550 °C (a) and of 1.5% SiO2/ -Аl2O3 (b), 5% SiO2/-Аl2O3 (c), 7% SiO2/-Аl2O3 (d), 10% SiO2/-Аl2O3 (e) calcined at 450

**(c)**

**(b) (a)**

**3792**

**3775**

**3755**

**(d)**

**3740-3745**

**3730**

**3685**

**3670**

**3700**

**(e)**

decomposition of preliminarily introduced tetraethoxysilane Si(OC2H5)4.

*10 a.u.*

**IR absorbance**

°C. The samples were outgassed at 450 °C

Thus, the FTIR spectroscopy study demonstrated that the main effect of the proposed method for modifying of the alumina surface consists in changing the ratio of different type sites capable of anchoring the active component precursor, in particular, in diminishing the fraction of more basic OH groups capable of coordination binding of the complexes (the formation of inner sphere complexes). The conclusions based on FTIR spectroscopy data and concerning changes in the surface state able to affect the mechanisms of precursor-support interaction and strength of such interaction were supported by independent 195Pt NMR study. 195Pt MAS NMR can be used to acquire data on the composition of adsorbed complexes and their interaction with the surface (Shelimov et al., 1999, 2000). Among advantages of this method for investigation of platinum complexes is a wide overall range of chemical shift (ca. 15000 ppm). This allows a relatively simple identification of Pt (IV) complexes with different structure from their 195Pt chemical shift, which is very sensitive to the ligand environment. Thus, substitution of a Cl– ligand in [PtCl6]2– by Н2О or ОН– produces chemical shifts by 500 and 660 ppm, respectively.

However, in the study of complexes adsorbed on the support surface, the 195Pt NMR signals are observed only if octahedral symmetry of the complexes is retained or slightly distorted during the adsorption. Coordination anchoring of a complex on alumina, when one or several chloride ligands of [PtCl6]2– are substituted by hydroxyl groups of the support (see equation (8)), is accompanied by a substantial decrease in intensity and broadening of the peaks; sometimes NMR signals are not detected. Such situation is observed at the adsorption of complexes on the unmodified -Al2O3. At a maximum possible concentration of a metal complex for the chemisorption (platinum content of 4.5 wt%), the spectrum has only a broad peak with low intensity in the region characterizing [PtCl6]2– (Fig. 7(a)).

Fig. 7. 195Pt MAS NMR spectra of 4.5%Pt/-Al2O3 (unmodified support) (a) and 2.0%Pt/Al2O3 with modified support (hydrothermal treatment at 180 °C, 3 h) (b)

This fact agrees with the diffuse reflectance electron spectroscopy and EXAFS examination of adsorbed complexes (Belskaya et al., 2008, 2011) showing that platinum on the -Al2O3 surface is mainly a component of the hydrolyzed coordinatively anchored complexes.

Thus, the FTIR spectroscopy study demonstrated that the main effect of the proposed method for modifying of the alumina surface consists in changing the ratio of different type sites capable of anchoring the active component precursor, in particular, in diminishing the fraction of more basic OH groups capable of coordination binding of the complexes (the formation of inner sphere complexes). The conclusions based on FTIR spectroscopy data and concerning changes in the surface state able to affect the mechanisms of precursor-support interaction and strength of such interaction were supported by independent 195Pt NMR study. 195Pt MAS NMR can be used to acquire data on the composition of adsorbed complexes and their interaction with the surface (Shelimov et al., 1999, 2000). Among advantages of this method for investigation of platinum complexes is a wide overall range of chemical shift (ca. 15000 ppm). This allows a relatively simple identification of Pt (IV) complexes with different structure from their 195Pt chemical shift, which is very sensitive to the ligand environment. Thus, substitution of a Cl– ligand in [PtCl6]2– by Н2О or

However, in the study of complexes adsorbed on the support surface, the 195Pt NMR signals are observed only if octahedral symmetry of the complexes is retained or slightly distorted during the adsorption. Coordination anchoring of a complex on alumina, when one or several chloride ligands of [PtCl6]2– are substituted by hydroxyl groups of the support (see equation (8)), is accompanied by a substantial decrease in intensity and broadening of the peaks; sometimes NMR signals are not detected. Such situation is observed at the adsorption of complexes on the unmodified -Al2O3. At a maximum possible concentration of a metal complex for the chemisorption (platinum content of 4.5 wt%), the spectrum has

only a broad peak with low intensity in the region characterizing [PtCl6]2– (Fig. 7(a)).

**700 600 500 400 300 200 100 0**

**, ppm**

This fact agrees with the diffuse reflectance electron spectroscopy and EXAFS examination of adsorbed complexes (Belskaya et al., 2008, 2011) showing that platinum on the -Al2O3 surface is mainly a component of the hydrolyzed coordinatively anchored complexes.

Fig. 7. 195Pt MAS NMR spectra of 4.5%Pt/-Al2O3 (unmodified support) (a) and 2.0%Pt/Al2O3 with modified support (hydrothermal treatment at 180 °C, 3 h) (b)

**(b)**

**(a)**

**40 ppm**

ОН– produces chemical shifts by 500 and 660 ppm, respectively.

Meanwhile, in the case of modified support (Fig. 7(b)), there is an intense peak in the spectrum even at a two times lower platinum content corresponding to chloride complex [PtCl6]2–. Hence, anchoring of the complex does not produce noticeable changes in its chemical composition; moreover, there is a considerable decrease in the contribution of coordination binding with the surface involving OH groups of the support. Electrostatic interaction of a metal complex with the modified support with respect to equation (7) seems to prevail here. In addition, temperature-programmed reduction followed by chemisorption of H2 and CO probe molecules was used to confirm that a decrease in the bond strength between precursor and support decreases the reduction temperature of adsorbed platinum species and diminishes the dispersion of supported particles by a factor of more than 3.

## **3.1.2.3 The effect of -Аl2O3 modifying with silica on acid-base properties of the surface**

Due to its high thermal stability, alumina modified with silica is a promising support for exhaust neutralization catalysts. Acid-base properties of the developed composition can be optimized by means of FTIR spectroscopy of adsorbed probe molecules. Composition of the support was varied by changing the silica concentration from 1.5 to 10 wt%. The modifying SiO2 compound was formed in the pore space of -Аl2O3 support during thermal decomposition of preliminarily introduced tetraethoxysilane Si(OC2H5)4.

Figure 8 shows FTIR spectra in the region of OH group stretching vibrations of the aluminum oxide samples modified with silica in comparison with the spectrum of initial -Аl2O3. The spectra of all modified oxides show a decrease in intensity of absorption bands corresponding to OH groups of the bridging and terminal types as compared to initial -Аl2O3. The additional absorption bands appear at 3740-3745 cm–1, their intensity growing with silicon content of the sample. The band at 3745 cm–1 can be assigned to the terminal Si(OH) groups.

Fig. 8. FTIR spectra of hydroxyl groups for -Al2O3 calcined at 550 °C (a) and of 1.5% SiO2/ -Аl2O3 (b), 5% SiO2/-Аl2O3 (c), 7% SiO2/-Аl2O3 (d), 10% SiO2/-Аl2O3 (e) calcined at 450 °C. The samples were outgassed at 450 °C

FTIR Spectroscopy of Adsorbed Probe Molecules for

adsorbed CO

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 163

CO, cm–1 2240 2230 2220 2205 2190 QCO, kJ/mol 59.5 54 48.5 42.5 33


**2250**

**2235**

**(d)**

**(e)**

**(c) (b)**

**(a)**

Type of Lewis acid site Super strong Strong Medium I Medium II Weak

Table 5. Types and concentrations of Lewis acid sites according to FTIR spectroscopy of

**2217**

**2193**

**2150 2200 2250**

Fig. 9. FTIR spectra of adsorbed deuterochloroform on -Аl2O3 (a) and aluminum oxides

in Table 6. Strong base sites of alumina are characterized by a.b. with CD 2193 and 2217 cm–1, which correspond to the calculated PA values of 942 and 915 kJ/mol. The band at 2235 cm–1 corresponds to medium-strength base sites, whereas the band at 2250 cm–1 is attributed to weak base sites. Paukshtis (Paukshtis, 1992) hypothesized that the strong and mediumstrength base sites are bridging oxygen atoms (Al–O–Al), whereas the weak sites are oxygen atoms of OH groups (Al–OH). The introduction of 1.5-5% SiO2 results in disappearance of super strong base sites and an abrupt decrease in the concentration of other types of base sites, which can be related with sequential blocking of the surface by silica. The introduction of 7-10% SiO2 increases the concentration of strong and medium strength base sites, which evidences the formation of a new surface phase, probably aluminosilicate one. This phase blocks only partially the surface of initial alumina. All the samples modified with silica have high-frequency a.b. of adsorbed deuterochloroform at 2255-2258 cm–1 characterizing the

**Wavenumber, cm-1**

Samples Concentration, mol/g

 *5 a.u.*

**IR absorbance**

modified with 1.5% (b), 5% (c), 7% (d), and 10% (e) of SiO2

sites which basicity is close to that of silica gel OH groups.

Adsorption of CO on -Аl2O3 at -196 °C results in the shift of OH bands with OH 3700 and 3775 cm–1 to the lower frequency region due to perturbation of OH stretch by hydrogen bonding. Aluminum oxides have no Brønsted acidity according to the minor shift of OH stretching vibration (OH/СО) by 120-130 cm–1 (Maache et al., 1993; Paze et al., 1997). A large shift of OH stretching vibrations for the silicon-containing samples with OH/СО equal to 240-290 cm–1 indicates the formation of strong Brønsted acid sites on the sample surface. As demonstrated by Crépeau et al. (Crépeau et al., 2006), high acidity for amorphous silicaalumina can be shown by free silanol groups located nearby an Al atom. Acidity of OH groups at 3740 cm–1 of the 10% SiO2/Al2O3 sample approaches the acidity of bridged Si(OH)Al groups in zeolites with the band at 3610 cm–1 according to the value of the low frequency shift of OH vibrations with adsorbed CO (OH/СО = 300 cm–1).

For all tested systems, the spectrum of adsorbed CO has a.b. with CO = 2158-2163 cm–1 attributed to CO hydrogen-bonded with hydroxyl groups, and a.b. with CO = 2130-2135 cm–1 related with absorption of physisorbed CO. Besides, for all silicon-containing systems, there is a.b. with CO 2170 cm–1, which was assigned to CO hydrogen-bonded with Brønsted acid sites. Concentrations of these sites estimated from the intensity of a.b. with CO 2170 cm–1 are listed in Table 4. One may see that concentration of Brønsted sites increases with the silica content in the sample.


Table 4. Surface concentrations of Brønsted acid sites and specific surface area of silica promoted aluminum oxides

By using low-temperature CO adsorption, the concentration and strength of coordinatively unsaturated surface sites of the synthesized oxide systems were also estimated. The absorption bands and their intensities observed in the spectrum of -Аl2O3 calcined at 550 °C (Table 5) are close to those reported in the previous Sections. Minor distinctions are related with the preparation procedures and texture characteristics of -Аl2O3. The types of Lewis acid sites identified by CO adsorption on the surface of silicon-containing systems are close to the types of Lewis acid sites for -Al2O3. The concentration of super strong Lewis acid sites (QCO = 59.5 kJ/mol) decreases, whereas the concentration of strong Lewis acid sites (QCO = 54 kJ/mol) and medium strength Lewis acid sites (QCO = 42.5 kJ/mol) increases. The type of Lewis acid sites with CO 2230 cm–1 is typical of aluminosilicate structures and can be assigned to aluminum in a defect octahedral coordination, which is bonded to silicon atom in the second coordination sphere (Paukshtis, 1992).

Four types of base sites were identified on the alumina surface using deuterochloroform as a probe molecule. The spectra are shown in Fig. 9; site strengths and concentrations are listed

Adsorption of CO on -Аl2O3 at -196 °C results in the shift of OH bands with OH 3700 and 3775 cm–1 to the lower frequency region due to perturbation of OH stretch by hydrogen bonding. Aluminum oxides have no Brønsted acidity according to the minor shift of OH stretching vibration (OH/СО) by 120-130 cm–1 (Maache et al., 1993; Paze et al., 1997). A large shift of OH stretching vibrations for the silicon-containing samples with OH/СО equal to 240-290 cm–1 indicates the formation of strong Brønsted acid sites on the sample surface. As demonstrated by Crépeau et al. (Crépeau et al., 2006), high acidity for amorphous silicaalumina can be shown by free silanol groups located nearby an Al atom. Acidity of OH groups at 3740 cm–1 of the 10% SiO2/Al2O3 sample approaches the acidity of bridged Si(OH)Al groups in zeolites with the band at 3610 cm–1 according to the value of the low

For all tested systems, the spectrum of adsorbed CO has a.b. with CO = 2158-2163 cm–1 attributed to CO hydrogen-bonded with hydroxyl groups, and a.b. with CO = 2130-2135 cm–1 related with absorption of physisorbed CO. Besides, for all silicon-containing systems, there is a.b. with CO 2170 cm–1, which was assigned to CO hydrogen-bonded with Brønsted acid sites. Concentrations of these sites estimated from the intensity of a.b. with CO 2170 cm–1 are listed in Table 4. One may see that concentration of Brønsted sites increases with the silica

OH, cm–1 Concentration of Brønsted

acid sites, mol/g

frequency shift of OH vibrations with adsorbed CO (OH/СО = 300 cm–1).

m2/g


Table 4. Surface concentrations of Brønsted acid sites and specific surface area of silica

By using low-temperature CO adsorption, the concentration and strength of coordinatively unsaturated surface sites of the synthesized oxide systems were also estimated. The absorption bands and their intensities observed in the spectrum of -Аl2O3 calcined at 550 °C (Table 5) are close to those reported in the previous Sections. Minor distinctions are related with the preparation procedures and texture characteristics of -Аl2O3. The types of Lewis acid sites identified by CO adsorption on the surface of silicon-containing systems are close to the types of Lewis acid sites for -Al2O3. The concentration of super strong Lewis acid sites (QCO = 59.5 kJ/mol) decreases, whereas the concentration of strong Lewis acid sites (QCO = 54 kJ/mol) and medium strength Lewis acid sites (QCO = 42.5 kJ/mol) increases. The type of Lewis acid sites with CO 2230 cm–1 is typical of aluminosilicate structures and can be assigned to aluminum in a defect octahedral coordination, which is bonded to silicon

Four types of base sites were identified on the alumina surface using deuterochloroform as a probe molecule. The spectra are shown in Fig. 9; site strengths and concentrations are listed

Sample Specific surface area,

atom in the second coordination sphere (Paukshtis, 1992).

content in the sample.

promoted aluminum oxides


Table 5. Types and concentrations of Lewis acid sites according to FTIR spectroscopy of adsorbed CO

Fig. 9. FTIR spectra of adsorbed deuterochloroform on -Аl2O3 (a) and aluminum oxides modified with 1.5% (b), 5% (c), 7% (d), and 10% (e) of SiO2

in Table 6. Strong base sites of alumina are characterized by a.b. with CD 2193 and 2217 cm–1, which correspond to the calculated PA values of 942 and 915 kJ/mol. The band at 2235 cm–1 corresponds to medium-strength base sites, whereas the band at 2250 cm–1 is attributed to weak base sites. Paukshtis (Paukshtis, 1992) hypothesized that the strong and mediumstrength base sites are bridging oxygen atoms (Al–O–Al), whereas the weak sites are oxygen atoms of OH groups (Al–OH). The introduction of 1.5-5% SiO2 results in disappearance of super strong base sites and an abrupt decrease in the concentration of other types of base sites, which can be related with sequential blocking of the surface by silica. The introduction of 7-10% SiO2 increases the concentration of strong and medium strength base sites, which evidences the formation of a new surface phase, probably aluminosilicate one. This phase blocks only partially the surface of initial alumina. All the samples modified with silica have high-frequency a.b. of adsorbed deuterochloroform at 2255-2258 cm–1 characterizing the sites which basicity is close to that of silica gel OH groups.

FTIR Spectroscopy of Adsorbed Probe Molecules for

**IR absorbance**

incorporation.

*1 a.u.*

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 165

**1935**

**2030**

**2095**

**2090**

**2108**

**2125**

**2150**

**2170**

**2198**

**(e)**

**(d) (c) (b)**

**(a)**

**1800 1900 2000 2100 2200**

sharp band at 2090 cm−1 and the broad band around 1935 cm−1 are ascribed to terminal and bridge-coordinated CO on Pd0, respectively (Sheppard & Nguyen, 1978). The bands at 2030, 2125 and 2150 cm−1 can be attributed, respectively, to bridged CO complex with Pd+ and linear CO complexes with Pd+ and Pd2+ isolated ions; the band at 2170 cm−1 can be assigned to CO linearly adsorbed on Pd2+ in PdO species (Hadjiivanov & Vayssilov, 2002). Besides, the band at 2196–2198 cm−1 is present in all spectra and may be assigned to CO adsorption on Zr4+ ions (Morterra et al., 1993). The highest intensity of this a.b. is observed for the sample reduced in H2 at 300 °C. This fact confirms studies concerning the necessity of hightemperature reduction of SZ to obtain the greatest Lewis acidity of the catalyst after metal

According to FTIR spectroscopy data, in oxidized Pd/SZ sample a part of palladium is presented as Pd0 (high intensities of a.b. at 1935 and 2090 cm−1). Supposedly, the formation of metallic palladium can be caused by evacuation at high temperatures during the pretreatment in IR cell. This assumption was confirmed in a special experiment by means of UV-vis spectroscopy (Belskaya et al., 2010). Pd/SZ catalyst after the oxidation is characterized by a.b. at 20500, 34000, 39500 and 46000 cm−1. The a.b. at 20500 and 39500 cm−<sup>1</sup> can be attributed, respectively, to d–d transition and ligand-to-metal charge transfer of Pd2+ ions in D4h oxygen environment (Rakai et al., 1992). Evacuation at 300 °C decreases the concentration of Pd2+ ions in PdO (a decrease in the intensity of a.b. at 20500 cm−1 was observed). This experiment clearly demonstrates that a possible effect of pretreatment conditions on the state of catalyst surface in IR spectroscopy study (in this case, the effect of

In the FTIR spectra of CO adsorbed on Pd/SZ samples that were reduced at 150–200 °C (Fig. 10(b), (c)), the a.b. at 1930 and 2090–2095 cm−1 attributed to metallic palladium dominate. Bands at 2125–2170 cm−1, assigned to CO adsorbed on oxidized Pd ions, almost

Fig. 10. FTIR spectra of CO adsorbed on Pd/SZ (25 °C; 10 mbar) after oxidation in air at 400 °C (a), after reduction in hydrogen at 150 °C (b), 200 °C (c), 300 °C (d) and 350 °C (e).

All spectra are background subtracted. Spectra were offset for clarity

evacuation at elevated temperatures) should be taken into account.

**Wavenumber, cm-1**


Table 6. Types and concentrations of base sites according to FTIR spectroscopy of adsorbed CDCl3. \* The concentration may be overrated due to close proximity of the band of adsorbed and physisorbed deuterochloroform

Thus, Section 3.1 demonstrated the conventional approaches employed in FTIR spectroscopy for investigation of functional groups on the alumina surface. Original methods for modifying of this most popular support were reported. Applicability of FTIR spectroscopy for estimating the effect of modifying on the surface acid-base properties and optimizing the composition of surface sites was shown. Thus, FTIR spectroscopy in combination with other methods can explain changes in adsorption and catalytic characteristics and predict the behavior of oxide surface under real conditions of catalyst synthesis and testing.

## **3.2 State of palladium in Pd/SZ catalysts**

Sulfated zirconia (SZ) promoted with noble metals is a very effective catalyst for isomerization of alkanes due to its high activity at low temperatures and high selectivity of isomers formation (Song & Sayari, 1996). Information on the state of metal in isomerization catalyst is quite topical. For example, metallic platinum or palladium improve the dehydrogenating capacity of the catalyst, which affects the formation of isoalkanes and enhance the production of atomic hydrogen which is necessary for the removal of coke precursors (Vera et al., 2002, 2003). It should be noted that state of the metal is determined to a great extent by the conditions of oxidative and reductive treatment of catalysts before the reaction. Thus, the challenge is to find the optimal pretreatment temperatures allowing the formation of acid sites and retaining the metallic function.

A convenient method for solving this problem is FTIR spectroscopy of adsorbed CO molecules. In our earlier work (Belskaya et al., 2010), this method was used for studying the state of supported palladium particles in Pd/SZ under different conditions of catalyst pretreatment. In the experiment, the catalyst treatment in various gas media (air, hydrogen) and at different temperatures (100-400 °C) was performed directly in a spectrometer cell. CO adsorption was carried out over a pressure range of 0.1 to 10 mbar at room temperature.

IR spectra of CO adsorbed on the surface of Pd/SZ pretreated under different conditions are shown in Fig. 10. The spectrum of CO adsorbed on the sample that was activated in air shows several a.b. located at 1935, 2030, 2090, 2125, 2150, 2170 and 2198 cm−1 (Fig. 10(a)). The

Type of base site Super strong Strong Medium Weak I Weak II CD, cm–1 72 48-43 30-23 15 10-7

1.5% SiO2/Al2O3 - 38 74 - 570\* 5% SiO2/Al2O3 - - 21 - 560\* 7% SiO2/Al2O3 - 26 56 - 730\* 10% SiO2/Al2O3 - 80 124 - 700\* Table 6. Types and concentrations of base sites according to FTIR spectroscopy of adsorbed CDCl3. \* The concentration may be overrated due to close proximity of the band of adsorbed

Thus, Section 3.1 demonstrated the conventional approaches employed in FTIR spectroscopy for investigation of functional groups on the alumina surface. Original methods for modifying of this most popular support were reported. Applicability of FTIR spectroscopy for estimating the effect of modifying on the surface acid-base properties and optimizing the composition of surface sites was shown. Thus, FTIR spectroscopy in combination with other methods can explain changes in adsorption and catalytic characteristics and predict the behavior of oxide surface under real conditions of catalyst

Sulfated zirconia (SZ) promoted with noble metals is a very effective catalyst for isomerization of alkanes due to its high activity at low temperatures and high selectivity of isomers formation (Song & Sayari, 1996). Information on the state of metal in isomerization catalyst is quite topical. For example, metallic platinum or palladium improve the dehydrogenating capacity of the catalyst, which affects the formation of isoalkanes and enhance the production of atomic hydrogen which is necessary for the removal of coke precursors (Vera et al., 2002, 2003). It should be noted that state of the metal is determined to a great extent by the conditions of oxidative and reductive treatment of catalysts before the reaction. Thus, the challenge is to find the optimal pretreatment temperatures allowing the

A convenient method for solving this problem is FTIR spectroscopy of adsorbed CO molecules. In our earlier work (Belskaya et al., 2010), this method was used for studying the state of supported palladium particles in Pd/SZ under different conditions of catalyst pretreatment. In the experiment, the catalyst treatment in various gas media (air, hydrogen) and at different temperatures (100-400 °C) was performed directly in a spectrometer cell. CO adsorption was carried out over a pressure range of 0.1 to 10 mbar at room temperature.

IR spectra of CO adsorbed on the surface of Pd/SZ pretreated under different conditions are shown in Fig. 10. The spectrum of CO adsorbed on the sample that was activated in air shows several a.b. located at 1935, 2030, 2090, 2125, 2150, 2170 and 2198 cm−1 (Fig. 10(a)). The

РА, kJ/mol 942 915-908 885-857 839 Samples Concentration, mol/g


and physisorbed deuterochloroform

**3.2 State of palladium in Pd/SZ catalysts** 

formation of acid sites and retaining the metallic function.

synthesis and testing.

Fig. 10. FTIR spectra of CO adsorbed on Pd/SZ (25 °C; 10 mbar) after oxidation in air at 400 °C (a), after reduction in hydrogen at 150 °C (b), 200 °C (c), 300 °C (d) and 350 °C (e). All spectra are background subtracted. Spectra were offset for clarity

sharp band at 2090 cm−1 and the broad band around 1935 cm−1 are ascribed to terminal and bridge-coordinated CO on Pd0, respectively (Sheppard & Nguyen, 1978). The bands at 2030, 2125 and 2150 cm−1 can be attributed, respectively, to bridged CO complex with Pd+ and linear CO complexes with Pd+ and Pd2+ isolated ions; the band at 2170 cm−1 can be assigned to CO linearly adsorbed on Pd2+ in PdO species (Hadjiivanov & Vayssilov, 2002). Besides, the band at 2196–2198 cm−1 is present in all spectra and may be assigned to CO adsorption on Zr4+ ions (Morterra et al., 1993). The highest intensity of this a.b. is observed for the sample reduced in H2 at 300 °C. This fact confirms studies concerning the necessity of hightemperature reduction of SZ to obtain the greatest Lewis acidity of the catalyst after metal incorporation.

According to FTIR spectroscopy data, in oxidized Pd/SZ sample a part of palladium is presented as Pd0 (high intensities of a.b. at 1935 and 2090 cm−1). Supposedly, the formation of metallic palladium can be caused by evacuation at high temperatures during the pretreatment in IR cell. This assumption was confirmed in a special experiment by means of UV-vis spectroscopy (Belskaya et al., 2010). Pd/SZ catalyst after the oxidation is characterized by a.b. at 20500, 34000, 39500 and 46000 cm−1. The a.b. at 20500 and 39500 cm−<sup>1</sup> can be attributed, respectively, to d–d transition and ligand-to-metal charge transfer of Pd2+ ions in D4h oxygen environment (Rakai et al., 1992). Evacuation at 300 °C decreases the concentration of Pd2+ ions in PdO (a decrease in the intensity of a.b. at 20500 cm−1 was observed). This experiment clearly demonstrates that a possible effect of pretreatment conditions on the state of catalyst surface in IR spectroscopy study (in this case, the effect of evacuation at elevated temperatures) should be taken into account.

In the FTIR spectra of CO adsorbed on Pd/SZ samples that were reduced at 150–200 °C (Fig. 10(b), (c)), the a.b. at 1930 and 2090–2095 cm−1 attributed to metallic palladium dominate. Bands at 2125–2170 cm−1, assigned to CO adsorbed on oxidized Pd ions, almost

FTIR Spectroscopy of Adsorbed Probe Molecules for

without their chemical interaction with palladium.

**2–-ZrO2**

after reduction or not.

during catalyst synthesis.

**3.3 Alumina promoted Pt/SO4**

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 167

small metal particles. Bulk metal palladium is known to have Eb of Pd3d5/2 at 335.2 eV (Brun et al., 1999; Otto et al., 1992). In our case, a small increase of Eb is likely to originate from the size effect (Mason, 1983). The formation of bulk palladium sulfide was not observed because the corresponding value of Eb of Pd3d5/2, which is about 337.2 eV (Chaplin et al., 2007), was not detected. Thus, changes of the metal surface after high temperature reduction in the H2 atmosphere, which are demonstrated by the FTIR spectra, have no significant effect on the electronic state of metal particles: according to XPS data, Pd remains in the metallic state. H2S resulting from sulfate reduction blocks the active metal surface, most likely without formation of PdS in large amounts. However, it cannot be unambiguously concluded from Pd3d and S2p core-level spectra whether the sulfide film forms on the palladium surface

For assessment of the state of palladium, we also used a model reaction which is commonly employed to test the metallic function of catalysts – low-temperature (50-90 °C) hydrogenation of benzene. There was a clear effect of pretreatment conditions on the conversion of benzene to cyclohexane. Pd/SZ catalyst reduced at the temperature corresponding to metal formation (according to TPR data) demonstrated the highest conversion. In the case of reduction temperature above 120 °C, before the catalytic test H2S was detected in the exhaust gases, and poisoning of the metal function was observed. Thus, in Pd/SZ catalysts, after the reduction treatment at 300 °C, hydrogenation activity of palladium is strongly inhibited. However, the constant activation energy and complete recovery of hydrogenation activity under mild regeneration conditions (Belskaya et al., 2010) indicate that the metal surface is only blocked by sulfate decomposition products

Thus, FTIR spectroscopy of adsorbed CO used to examine the state of supported palladium in Pd/SZ catalysts provided data that agree well with the data obtained by independent methods – XPS and a model reaction for testing the metal function. Analysis of changes in the state of surface revealed by FTIR spectroscopy can be useful for explaining the adsorption and catalytic properties as well as for optimizing the conditions of thermal stages

The approaches for controlling the SO42–-ZrO2 acidity are of great practical importance, as they can change the catalyst activity and selectivity in various acid-catalyzed reactions (Hua et al., 2000; Lavrenov et al., 2007; Zalewski et al., 1999). In our works (Kazakov et al., 2010, 2011, 2012), we optimized the acidic and hydrogenation properties of bifunctional Pt/SO42–- ZrO2 catalyst for the one-step hydroisomerization of benzene-containing fractions, which is intended for elimination of benzene in gasoline while minimizing the octane loss. The introduction of alumina into the catalyst was suggested as the main modifying procedure. IR spectroscopy allowed us to elucidate the effect of catalyst composition on the properties of surface functional groups, to reveal the role of alumina in the formation of metal and acid sites, and provided a detailed characterization of the surface properties of optimal hydroisomerization catalyst. The work was performed with catalysts Pt/SO42–-ZrO2 (4.5 wt% SO42–), Pt/SO42–-ZrO2-Al2O3, (3.1 wt% SO42– and 67.8 wt% Al2O3) and Pt/Al2O3. Samples were denoted as Pt/SZ, Pt/SZA and Pt/A, respectively. Platinum concentration was 0.3 wt%. We studied also the supports used for the catalyst synthesis (SZ, SZA and A,

vanish. The differences between spectra of adsorbed CO on Pd/SZ samples reduced at 200 and 300–350 °C are dramatic (Fig. 10(c)–(e)). An increase in the reduction temperature suppresses the bridge coordinated CO IR bands, decreases the intensity and slightly shifts the stretching frequency of linearly adsorbed CO to higher wavenumbers. Such shift occurs when CO is chemisorbed on the sulfur-saturated Pd surface (Guerra, 1969; Jorgensen & Madix, 1985). So, appearance of the band at 2105–2108 cm−1 can be attributed to linear CO complexes with palladium in an electron-deficient state with high S-coverage. According to (Jorgensen & Madix, 1985; Ivanov & Kustov, 1998) a possible reason for the decrease in concentration of bridging CO is the partial covering of Pd surface by sulfur species. Thus, we suppose that low CO chemisorption capacity of Pd0 atoms in Pd/SZ samples reduced at high temperatures is due to partial sulfur coating of the metal surface. Therewith, new a.b. at 2135 and 2160 cm−1 are observed in the spectra of CO adsorbed on Pd/SZ samples that were reduced at 300–350 °C. These bands appear on admission of 0.1 mbar CO and grow in intensity with increasing CO pressure (Belskaya et al., 2010); they are attributed to CO complexes with oxidized Pd+ and Pd2+ species, probably in Pdn+–S2- sites (Vazquez-Zavala et al., 1994).

Fig. 11. Zr3p and Pd3d core-level spectra of the samples after oxidation in air at 400 °C (a) and after reduction in hydrogen at 300 °C (b). For clear identification, Pd3d spectra are multiplied by 3

IR spectroscopic data on the state of supported palladium in Pd/SZ after oxidation and reduction were compared with X-ray photoelectron spectroscopy (XPS) data for the same samples. Figure 11 shows the X-ray photoelectron spectra in the spectral region of Zr3p and Pd3d core-level lines. The difference curves between experimental spectrum and the envelope of the fit are presented under each spectrum. After calcination in air at 400 °C (Fig. 11(a)), palladium spectrum can be described by one doublet line with binding energy (Eb) of Pd3d5/2 336.5 eV. Such value of Eb is close to that for the oxidized palladium species in palladium oxide PdO (Brun et al., 1999; Pillo et al., 1997). After the action of H2 (Fig. 11(b)), Pd3d spectrum gives a peak with Eb (Pd3d5/2) 335.6 eV, which is assigned to

vanish. The differences between spectra of adsorbed CO on Pd/SZ samples reduced at 200 and 300–350 °C are dramatic (Fig. 10(c)–(e)). An increase in the reduction temperature suppresses the bridge coordinated CO IR bands, decreases the intensity and slightly shifts the stretching frequency of linearly adsorbed CO to higher wavenumbers. Such shift occurs when CO is chemisorbed on the sulfur-saturated Pd surface (Guerra, 1969; Jorgensen & Madix, 1985). So, appearance of the band at 2105–2108 cm−1 can be attributed to linear CO complexes with palladium in an electron-deficient state with high S-coverage. According to (Jorgensen & Madix, 1985; Ivanov & Kustov, 1998) a possible reason for the decrease in concentration of bridging CO is the partial covering of Pd surface by sulfur species. Thus, we suppose that low CO chemisorption capacity of Pd0 atoms in Pd/SZ samples reduced at high temperatures is due to partial sulfur coating of the metal surface. Therewith, new a.b. at 2135 and 2160 cm−1 are observed in the spectra of CO adsorbed on Pd/SZ samples that were reduced at 300–350 °C. These bands appear on admission of 0.1 mbar CO and grow in intensity with increasing CO pressure (Belskaya et al., 2010); they are attributed to CO complexes with oxidized Pd+ and Pd2+ species, probably in Pdn+–S2- sites (Vazquez-Zavala

Fig. 11. Zr3p and Pd3d core-level spectra of the samples after oxidation in air at 400 °C (a) and after reduction in hydrogen at 300 °C (b). For clear identification, Pd3d spectra are

IR spectroscopic data on the state of supported palladium in Pd/SZ after oxidation and reduction were compared with X-ray photoelectron spectroscopy (XPS) data for the same samples. Figure 11 shows the X-ray photoelectron spectra in the spectral region of Zr3p and Pd3d core-level lines. The difference curves between experimental spectrum and the envelope of the fit are presented under each spectrum. After calcination in air at 400 °C (Fig. 11(a)), palladium spectrum can be described by one doublet line with binding energy (Eb) of Pd3d5/2 336.5 eV. Such value of Eb is close to that for the oxidized palladium species in palladium oxide PdO (Brun et al., 1999; Pillo et al., 1997). After the action of H2 (Fig. 11(b)), Pd3d spectrum gives a peak with Eb (Pd3d5/2) 335.6 eV, which is assigned to

et al., 1994).

multiplied by 3

small metal particles. Bulk metal palladium is known to have Eb of Pd3d5/2 at 335.2 eV (Brun et al., 1999; Otto et al., 1992). In our case, a small increase of Eb is likely to originate from the size effect (Mason, 1983). The formation of bulk palladium sulfide was not observed because the corresponding value of Eb of Pd3d5/2, which is about 337.2 eV (Chaplin et al., 2007), was not detected. Thus, changes of the metal surface after high temperature reduction in the H2 atmosphere, which are demonstrated by the FTIR spectra, have no significant effect on the electronic state of metal particles: according to XPS data, Pd remains in the metallic state. H2S resulting from sulfate reduction blocks the active metal surface, most likely without formation of PdS in large amounts. However, it cannot be unambiguously concluded from Pd3d and S2p core-level spectra whether the sulfide film forms on the palladium surface after reduction or not.

For assessment of the state of palladium, we also used a model reaction which is commonly employed to test the metallic function of catalysts – low-temperature (50-90 °C) hydrogenation of benzene. There was a clear effect of pretreatment conditions on the conversion of benzene to cyclohexane. Pd/SZ catalyst reduced at the temperature corresponding to metal formation (according to TPR data) demonstrated the highest conversion. In the case of reduction temperature above 120 °C, before the catalytic test H2S was detected in the exhaust gases, and poisoning of the metal function was observed. Thus, in Pd/SZ catalysts, after the reduction treatment at 300 °C, hydrogenation activity of palladium is strongly inhibited. However, the constant activation energy and complete recovery of hydrogenation activity under mild regeneration conditions (Belskaya et al., 2010) indicate that the metal surface is only blocked by sulfate decomposition products without their chemical interaction with palladium.

Thus, FTIR spectroscopy of adsorbed CO used to examine the state of supported palladium in Pd/SZ catalysts provided data that agree well with the data obtained by independent methods – XPS and a model reaction for testing the metal function. Analysis of changes in the state of surface revealed by FTIR spectroscopy can be useful for explaining the adsorption and catalytic properties as well as for optimizing the conditions of thermal stages during catalyst synthesis.

#### **3.3 Alumina promoted Pt/SO4 2–-ZrO2**

The approaches for controlling the SO42–-ZrO2 acidity are of great practical importance, as they can change the catalyst activity and selectivity in various acid-catalyzed reactions (Hua et al., 2000; Lavrenov et al., 2007; Zalewski et al., 1999). In our works (Kazakov et al., 2010, 2011, 2012), we optimized the acidic and hydrogenation properties of bifunctional Pt/SO42–- ZrO2 catalyst for the one-step hydroisomerization of benzene-containing fractions, which is intended for elimination of benzene in gasoline while minimizing the octane loss. The introduction of alumina into the catalyst was suggested as the main modifying procedure. IR spectroscopy allowed us to elucidate the effect of catalyst composition on the properties of surface functional groups, to reveal the role of alumina in the formation of metal and acid sites, and provided a detailed characterization of the surface properties of optimal hydroisomerization catalyst. The work was performed with catalysts Pt/SO4 2–-ZrO2 (4.5 wt% SO4 2–), Pt/SO42–-ZrO2-Al2O3, (3.1 wt% SO4 2– and 67.8 wt% Al2O3) and Pt/Al2O3. Samples were denoted as Pt/SZ, Pt/SZA and Pt/A, respectively. Platinum concentration was 0.3 wt%. We studied also the supports used for the catalyst synthesis (SZ, SZA and A,

FTIR Spectroscopy of Adsorbed Probe Molecules for

is observed at 210 and 225 °C for Pt/SZA and Pt/A.

with frequencies for samples Pt/A and Pt/SZ.

**IR absorbance**

then evacuated at 500 °C

that observed for Pt/A.

 *1 a.u.*

**1830**

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 169

According to TPR data, platinum reduction on SZ surface (sample Pt/SZ) in the absence of chemisorption interaction starts from 90 °C. However, in the samples with alumina this process shifts toward higher temperatures, and a maximum rate of hydrogen consumption

The state of supported platinum in finished catalysts after reduction in hydrogen was investigated by FTIR spectroscopy of adsorbed CO (Fig. 13). The band with *ν*CO 2200-2208 cm-1, which is present in all the spectra, corresponds to CO complexes with Lewis acid sites of the catalysts (Morterra et al., 1993). The spectrum of Pt/A sample shows a.b. with *ν*CO 2065 cm–1 corresponding to stretching vibrations of CO linearly adsorbed on Pt0, and a broad band at 1830 cm–1 characterizing the bridging CO species on Pt0 (Apesteguia et al., 1984; Kooh et al., 1991). After CO adsorption on Pt/SZ, there appear bands at 2100 and 2150 cm–1, which have close intensities and correspond to linear CO complexes with Pt0 and Ptδ+, respectively (Grau et al., 2004; Morterra et al., 1997). In the case of Pt/SZA, the frequencies of CO (a.b. 2085 cm–1) adsorbed on metal platinum particles are intermediate in comparison

**1800 1900 2000 2100 2200**

**2100**

**2085**

**2065**

**2150**

**(a) (b) (c)**

**2200**

**Wavenumber, cm-1**

In comparison with Pt/SZ, the band corresponding to CO – Pt0 complexes for samples Pt/SZA and Pt/SZ is shifted toward higher frequencies by 20 and 35 cm-1, respectively. Such upward shift of *ν*CO was also observed for Pd/SZ samples (Section 3.2) and can be related to the presence of sulfur species on the metal surface (Apesteguia et al., 1984, 1987; J.R. Chang & S.L. Chang, 1998). These sulfur species are formed both at the stage of oxidative treatment and during the reduction; they can poison the metal partially or completely (Dicko et al., 1994; Iglesia et al., 1993). As a result, Pt/SZ demonstrates very poor hydrogenation activity and does not chemisorb hydrogen (Table 7). Pt/SZA sample has an enhanced hydrogenation activity in comparison with Pt/SZ; nevertheless, it is lower than

Fig. 13. FTIR spectra of CO (25 °C; 10 mbar) adsorbed on the catalysts: Pt/SZ (a), Pt/SZA (b), Pt/A (c). Prior to recording, the samples were reduced in hydrogen flow at 300 °C and

respectively). The preparation procedure for supports and catalysts is reported in (Kazakov et al., 2010, 2011).

#### **3.3.1 The effect of alumina introduction on the state of supported platinum**

The formation of platinum sites and the state of metal in a finished catalyst strongly depend on the interaction of precursor with the surface groups of support. As it was shown earlier, the composition and amount of hydroxyl groups on the support surface play a significant role in the platinum compounds anchoring from a solution of H2[PtCl6]. Infrared spectra of the OH stretching region for SZ, SZA and A supports are shown in Fig. 12. The spectrum of sample SZ is represented by a.b. at 3651 cm–1, which corresponds to bridging OH groups with acidic properties (Kustov et al., 1994; Manoilova et al., 2007). The FTIR spectrum of SZA sample has a.b. at 3772 and 3789 cm–1 assigned to terminal OH groups, and a.b. 3677 and 3728 cm–1 corresponding to bridging OH groups (Knözinger & Ratnasamy, 1978). The presence of these types of OH groups is typical for the γ-Al2O3 surface (sample A in Fig. 12).

Fig. 12. FTIR spectra of the catalyst supports in the OH stretching region: SZ (a), SZA (b), A (c). Spectra were offset for clarity. Prior to recording, the samples were evacuated at 400 °C

Significant distinctions in the hydroxyl cover suggest different mechanisms of the interaction between metal complex and support. Indeed, the presence of only the hydroxyl groups with acidic properties on the SZ surface explains the absence of chemisorption anchoring of the anionic [PtCl6]2- complex. After the introduction of 67.8 wt% alumina into sulfated zirconia sites for chloroplatinate ions sorption appear on the surface of mixed SO42–-ZrO2-Al2O3 support. However, the concentration of bridging AlVI(OH)AlVI groups (3728 cm–1) that are most active in chloroplatinate anchoring, and terminal groups (3772 and 3789 cm–1) on SZA support is lower than their concentration on the alumina surface, which causes a smaller fraction of complexes anchored by chemisorption (81% for SZA and 100% for A).

The reduction temperature of platinum species anchored on SZA support, which characterizes the strength of precursor – support interaction, also has a medium value.

respectively). The preparation procedure for supports and catalysts is reported in (Kazakov

The formation of platinum sites and the state of metal in a finished catalyst strongly depend on the interaction of precursor with the surface groups of support. As it was shown earlier, the composition and amount of hydroxyl groups on the support surface play a significant role in the platinum compounds anchoring from a solution of H2[PtCl6]. Infrared spectra of the OH stretching region for SZ, SZA and A supports are shown in Fig. 12. The spectrum of sample SZ is represented by a.b. at 3651 cm–1, which corresponds to bridging OH groups with acidic properties (Kustov et al., 1994; Manoilova et al., 2007). The FTIR spectrum of SZA sample has a.b. at 3772 and 3789 cm–1 assigned to terminal OH groups, and a.b. 3677 and 3728 cm–1 corresponding to bridging OH groups (Knözinger & Ratnasamy, 1978). The presence of these types of OH groups is typical for the γ-Al2O3

**3200 3300 3400 3500 3600 3700 3800**

Fig. 12. FTIR spectra of the catalyst supports in the OH stretching region: SZ (a), SZA (b), A (c). Spectra were offset for clarity. Prior to recording, the samples were evacuated at 400 °C

Significant distinctions in the hydroxyl cover suggest different mechanisms of the interaction between metal complex and support. Indeed, the presence of only the hydroxyl groups with acidic properties on the SZ surface explains the absence of chemisorption anchoring of the anionic [PtCl6]2- complex. After the introduction of 67.8 wt% alumina into sulfated zirconia sites for chloroplatinate ions sorption appear on the surface of mixed SO42–-ZrO2-Al2O3 support. However, the concentration of bridging AlVI(OH)AlVI groups (3728 cm–1) that are most active in chloroplatinate anchoring, and terminal groups (3772 and 3789 cm–1) on SZA support is lower than their concentration on the alumina surface, which causes a smaller fraction of complexes anchored by

The reduction temperature of platinum species anchored on SZA support, which characterizes the strength of precursor – support interaction, also has a medium value.

**Wavenumber, cm-1**

**(a)**

**(b)**

**(c)**

**3677**

**3651**

**3677**

**3767**

**3789**

**3789**

**3772**

**3728**

**3728**

**3.3.1 The effect of alumina introduction on the state of supported platinum** 

et al., 2010, 2011).

surface (sample A in Fig. 12).

**IR absorbance**

chemisorption (81% for SZA and 100% for A).

*5 a.u.*

According to TPR data, platinum reduction on SZ surface (sample Pt/SZ) in the absence of chemisorption interaction starts from 90 °C. However, in the samples with alumina this process shifts toward higher temperatures, and a maximum rate of hydrogen consumption is observed at 210 and 225 °C for Pt/SZA and Pt/A.

The state of supported platinum in finished catalysts after reduction in hydrogen was investigated by FTIR spectroscopy of adsorbed CO (Fig. 13). The band with *ν*CO 2200-2208 cm-1, which is present in all the spectra, corresponds to CO complexes with Lewis acid sites of the catalysts (Morterra et al., 1993). The spectrum of Pt/A sample shows a.b. with *ν*CO 2065 cm–1 corresponding to stretching vibrations of CO linearly adsorbed on Pt0, and a broad band at 1830 cm–1 characterizing the bridging CO species on Pt0 (Apesteguia et al., 1984; Kooh et al., 1991). After CO adsorption on Pt/SZ, there appear bands at 2100 and 2150 cm–1, which have close intensities and correspond to linear CO complexes with Pt0 and Ptδ+, respectively (Grau et al., 2004; Morterra et al., 1997). In the case of Pt/SZA, the frequencies of CO (a.b. 2085 cm–1) adsorbed on metal platinum particles are intermediate in comparison with frequencies for samples Pt/A and Pt/SZ.

Fig. 13. FTIR spectra of CO (25 °C; 10 mbar) adsorbed on the catalysts: Pt/SZ (a), Pt/SZA (b), Pt/A (c). Prior to recording, the samples were reduced in hydrogen flow at 300 °C and then evacuated at 500 °C

In comparison with Pt/SZ, the band corresponding to CO – Pt0 complexes for samples Pt/SZA and Pt/SZ is shifted toward higher frequencies by 20 and 35 cm-1, respectively. Such upward shift of *ν*CO was also observed for Pd/SZ samples (Section 3.2) and can be related to the presence of sulfur species on the metal surface (Apesteguia et al., 1984, 1987; J.R. Chang & S.L. Chang, 1998). These sulfur species are formed both at the stage of oxidative treatment and during the reduction; they can poison the metal partially or completely (Dicko et al., 1994; Iglesia et al., 1993). As a result, Pt/SZ demonstrates very poor hydrogenation activity and does not chemisorb hydrogen (Table 7). Pt/SZA sample has an enhanced hydrogenation activity in comparison with Pt/SZ; nevertheless, it is lower than that observed for Pt/A.

FTIR Spectroscopy of Adsorbed Probe Molecules for

Lewis acid sites, µmol/g

> 15 45 250

> > 29

hydrogen flow at 300 °C and then evacuated at 500 °C

X CH, % MCP yield,

%

conversion; CH – cyclohexane; MCP – methylcyclopentane

Catalyst νCO,

Pt/SZ

Pt/SZA

(Table 10).

conversion

Catalyst

Pt/A <sup>2204</sup>

cm–1

2208 2202 2192

2192

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 171

Lewis acid sites, µmol/g

Table 8. Acidic properties of Pt/SZ, Pt/SZA and Pt/A catalysts according to FTIR spectroscopy of adsorbed CO and pyridine. Prior to recording, samples were reduced in

Results of FTIR spectroscopic study are in good agreement with the model acid-catalyzed reactions of n-heptane and cyclohexane isomerization. The introduction of alumina into Pt/SZ decreases the total catalyst activity in n-C7H16 isomerization, which shows up as increase of the temperature of 50% n-heptane conversion from 112 to 266 °C (Table 9). For isomerization of cyclohexane to methylcyclopentane, higher operating temperatures are thermodynamically more favorable (Tsai et al., 2011). As a result, Pt/SZA catalyst is more efficient for cyclohexane isomerization due to higher selectivity at higher temperatures

Catalyst X n-C7, % t, °C Iso-C7 yield, % Selectivity to iso-C7, %

Selectivity to

Pt/SZ 70.0 57.1 81.7 91.8 26.2 29.6 Pt/SZA 4.4 4.4 99.3 74.4 69.6 93.7 Pt/A - - - 0.2 0.0 - Table 10. Isomerization of cyclohexane over Pt/SZ, Pt/SZA and Pt/A catalysts. Reaction conditions: 1.5 MPa, weight hourly space velocity 4.0 h–1, H2 : C6H12 molar ratio 5. X –

200 °C 275 °C

MCP, % X CH, % MCP yield,

%

Selectivity to MCP, %

Pt/SZ 50.0 112 43.3 86.5 Pt/SZA 50.0 266 47.1 94.2 Pt/A 12.9 300 6.3 47.0 Table 9. Isomerization of n-heptane over Pt/SZ, Pt/SZA and Pt/A catalysts. Reaction conditions: 1.5 MPa, weight hourly space velocity 4.0 h–1, H2 : n-C7H16 molar ratio 5. X –

Total Lewis acid sites, µmol/g

medium weak µmol/g

115 530 645 32

60 250 310 9

<sup>350</sup>29 350 379 0

Total Brønsted acid sites,


Table 7. Hydrogen chemisorption and benzene hydrogenation over Pt/SZ, Pt/SZA and Pt/A catalysts. Benzene hydrogenation conditions: 200 °C, 0.1 MPa, weight hourly space velocity 4.0 h–1, H2 : C6H6 molar ratio 8

The increasing accessibility of platinum sites for adsorption and catalytic reaction in a series Pt/SZ < Pt/SZA < Pt/A can be attributed both to a decrease in the content of sulfur compounds in the catalyst upon dilution of sulfated zirconia with alumina, and to a higher resistance to poisoning of more disperse supported platinum crystallites produced by chemisorption anchoring of a precursor (J.R. Chang et al., 1997).

#### **3.3.2 The effect of alumina introduction on the acidic properties of Pt/SO4 2–-ZrO2**

The strength and concentration of acid sites of catalysts Pt/SZ, Pt/SZA and Pt/A reduced at 300 °C were estimated from FTIR spectra of adsorbed CO and pyridine molecules. After CO adsorption, spectra of all the samples had the a.b. 2180-2208 cm–1 corresponding to CO complexes with Lewis acid sites of different strength, a.b. 2168-2171 cm–1 assigned to CO complexes with Brønsted acid sites, a.b. 2160-2162 cm–1 characterizing CO complexes with hydroxyl groups having weak acidic properties, and a.b. 2133-2148 cm–1 corresponding to the adsorption of physisorbed CO molecules. In all cases, only the medium strength (a.b. with *ν*CO 2204-2208 cm–1 for Pt/SZ, 2202-2208 cm–1 for Pt/SZA, and 2204 cm–1 for Pt/A) and weak Lewis acid sites (a.b. with *ν*CO 2194, 2188, 2180 cm–1 for Pt/SZ, and 2192 cm–1 for Pt/SZA and Pt/A) were detected. After the pyridine adsorption, we observed a.b. corresponding to complexes of pyridine molecules with Brønsted acid sites (1544 cm–1) and Lewis acid sites (1445 cm–1).

Data on the concentration of Lewis acid sites (calculated from the integral intensities of adsorbed CO a.b.) and Brønsted acid sites (calculated from the integral intensities of adsorbed pyridine a.b.) for the tested catalysts are listed in Table 8. The Pt/SZ catalyst has the highest content both of Lewis and Brønsted acid sites. For sample Pt/A, Brønsted acid sites able to protonate pyridine were not observed. Pt/SZA has intermediate position with respect to its acidic properties. The Brønsted acid sites content in this sample is 3.5 times lower as compared to Pt/SZ, which virtually corresponds to a decrease of ZrO2 amount in its composition (29.1 against 95.5 wt%, respectively). The total amount of Lewis acid sites in comparison with Pt/SZ decreases twofold. However, the ratio of medium strength and weak Lewis acid sites for Pt/SZA sample corresponds to the ratio revealed for Pt/SZ. Thus, the introduction of alumina into Pt/SZ system decreases the amount of Lewis and Brønsted acid sites, which is related to the effect of its dilution with a component having a lower intrinsic acidity. The observed nonadditive change in the concentration of acid sites, in particular Lewis acid sites, may be caused by interaction of the system components, which was noted earlier (Kazakov et al., 2010).

Pt/SZ 0.00 1.8 Pt/SZA 0.44 40.7 Pt/A 0.85 97.1

Table 7. Hydrogen chemisorption and benzene hydrogenation over Pt/SZ, Pt/SZA and Pt/A catalysts. Benzene hydrogenation conditions: 200 °C, 0.1 MPa, weight hourly space

The increasing accessibility of platinum sites for adsorption and catalytic reaction in a series Pt/SZ < Pt/SZA < Pt/A can be attributed both to a decrease in the content of sulfur compounds in the catalyst upon dilution of sulfated zirconia with alumina, and to a higher resistance to poisoning of more disperse supported platinum crystallites produced by

The strength and concentration of acid sites of catalysts Pt/SZ, Pt/SZA and Pt/A reduced at 300 °C were estimated from FTIR spectra of adsorbed CO and pyridine molecules. After CO adsorption, spectra of all the samples had the a.b. 2180-2208 cm–1 corresponding to CO complexes with Lewis acid sites of different strength, a.b. 2168-2171 cm–1 assigned to CO complexes with Brønsted acid sites, a.b. 2160-2162 cm–1 characterizing CO complexes with hydroxyl groups having weak acidic properties, and a.b. 2133-2148 cm–1 corresponding to the adsorption of physisorbed CO molecules. In all cases, only the medium strength (a.b. with *ν*CO 2204-2208 cm–1 for Pt/SZ, 2202-2208 cm–1 for Pt/SZA, and 2204 cm–1 for Pt/A) and weak Lewis acid sites (a.b. with *ν*CO 2194, 2188, 2180 cm–1 for Pt/SZ, and 2192 cm–1 for Pt/SZA and Pt/A) were detected. After the pyridine adsorption, we observed a.b. corresponding to complexes of pyridine molecules with Brønsted acid sites (1544 cm–1) and

Data on the concentration of Lewis acid sites (calculated from the integral intensities of adsorbed CO a.b.) and Brønsted acid sites (calculated from the integral intensities of adsorbed pyridine a.b.) for the tested catalysts are listed in Table 8. The Pt/SZ catalyst has the highest content both of Lewis and Brønsted acid sites. For sample Pt/A, Brønsted acid sites able to protonate pyridine were not observed. Pt/SZA has intermediate position with respect to its acidic properties. The Brønsted acid sites content in this sample is 3.5 times lower as compared to Pt/SZ, which virtually corresponds to a decrease of ZrO2 amount in its composition (29.1 against 95.5 wt%, respectively). The total amount of Lewis acid sites in comparison with Pt/SZ decreases twofold. However, the ratio of medium strength and weak Lewis acid sites for Pt/SZA sample corresponds to the ratio revealed for Pt/SZ. Thus, the introduction of alumina into Pt/SZ system decreases the amount of Lewis and Brønsted acid sites, which is related to the effect of its dilution with a component having a lower intrinsic acidity. The observed nonadditive change in the concentration of acid sites, in particular Lewis acid sites, may be caused by interaction of the system components, which

**2–-ZrO2**

velocity 4.0 h–1, H2 : C6H6 molar ratio 8

Lewis acid sites (1445 cm–1).

was noted earlier (Kazakov et al., 2010).

chemisorption anchoring of a precursor (J.R. Chang et al., 1997).

**3.3.2 The effect of alumina introduction on the acidic properties of Pt/SO4**

Catalyst H/Pt Benzene conversion, %


Table 8. Acidic properties of Pt/SZ, Pt/SZA and Pt/A catalysts according to FTIR spectroscopy of adsorbed CO and pyridine. Prior to recording, samples were reduced in hydrogen flow at 300 °C and then evacuated at 500 °C

Results of FTIR spectroscopic study are in good agreement with the model acid-catalyzed reactions of n-heptane and cyclohexane isomerization. The introduction of alumina into Pt/SZ decreases the total catalyst activity in n-C7H16 isomerization, which shows up as increase of the temperature of 50% n-heptane conversion from 112 to 266 °C (Table 9). For isomerization of cyclohexane to methylcyclopentane, higher operating temperatures are thermodynamically more favorable (Tsai et al., 2011). As a result, Pt/SZA catalyst is more efficient for cyclohexane isomerization due to higher selectivity at higher temperatures (Table 10).


Table 9. Isomerization of n-heptane over Pt/SZ, Pt/SZA and Pt/A catalysts. Reaction conditions: 1.5 MPa, weight hourly space velocity 4.0 h–1, H2 : n-C7H16 molar ratio 5. X – conversion


Table 10. Isomerization of cyclohexane over Pt/SZ, Pt/SZA and Pt/A catalysts. Reaction conditions: 1.5 MPa, weight hourly space velocity 4.0 h–1, H2 : C6H12 molar ratio 5. X – conversion; CH – cyclohexane; MCP – methylcyclopentane

FTIR Spectroscopy of Adsorbed Probe Molecules for

pp. 720-728, ISSN 0023-1584

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 173

Baumgarten, E., Wagner, R., & Lentes-Wagner, C. (1989). Quantitative Determination of

Belskaya, O.B., Danilova, I.G., Kazakov, M.O., Gulyaeva, T.I., Kibis, L.S., Boronin, A.I.,

Bocanegra, S.A., Castro, A.A., Guerrero-Ruíz, A., Scelza, O.A., & de Miguel, S.R. (2006).

Bourikas, K., Kordulis, C., & Lycourghiotis, A. (2006). The Role of the Liquid-Solid Interface

Chang, J.R., Chang, S.L., & Lin, T.B. (1997). γ-Alumina-Supported Pt Catalysts for Aromatics

*Journal of Catalysis*, Vol.169, No.1, (July 1997), pp. 338-346, ISSN 0021-9517 Chang, J.R., & Chang, S.L. (1998). Catalytic Properties of γ-Alumina-Supported Pt Catalysts

Chaplin, B.P., Shapley, J.R., & Werth, C.J. (2007). Regeneration of Sulfur-Fouled Bimetallic

Chertov, V.M., Zelentsov, V.I., & Lyashkevich, B.N. (1982). Production of Finely Divided

Crépeau, G., Montouillout, V., Vimont, A., Mariey, L., Cseri, T., & Maugé F. (2006). Nature,

Davydov, A. (2003). *Molecular Spectroscopy of Oxide Catalyst Surfaces*, John Wiley & Sons Ltd,

*Engineering*, Vol.48, No.4, (December 2006), pp.363-444, ISSN 0161-4940 Brun, M., Berthet, A., & Bertolini, J.C. (1999). XPS, AES and Auger Parameter of Pd and

*Nanocomposites*, Vol.1, No.2, pp. 99-133, ISSN 1949-4823

(May 2006), pp. 161-166, ISSN 1385-8947

1999), pp. 55-60, ISSN 0368-2048

pp. 5491-5497, ISSN 0013-936X

15172-15185, ISSN 1520-6106

ISBN 0-471-98731-X, Chichester, England

9517

ISSN 0021-888X

Hydroxyl Groups on Alumina by IR Spectroscopy. *Fresenius Zeitschrift für Analytische Chemie*, Vol.334, No.3, (January 1989), pp. 246-251, ISSN 0016-1152 Bel'skaya, O.B., Karymova, R.Kh., Kochubey, D.I., & Duplyakin, V.K. (2008). Genesis of the

Active-Component Precursor in the Synthesis of Pt/Al2O3 Catalysts: I. Transformation of the [PtCl6]2- Complex in the Interaction between Chloroplatinic Acid and the γ-Al2O3 Surface. *Kinetics and Catalysis*, Vol.49, No.5, (September 2008),

Lavrenov, A.V., & Likholobov, V.A. (2010). Investigation of Active Metal Species Formation in Pd-Promoted Sulfated Zirconia Isomerization Catalyst. *Applied Catalysis A: General*, Vol.387, No.1-2, (October 2010), pp. 5-12, ISSN 0926-860X Belskaya, O.B., Duplyakin, V.K., & Likholobov, V.A. (2011). Molecular Design of Precursor

in the Synthesis of Catalytic Nanocomposite System Pt-Al2O3. *Smart* 

Characteristics of the Metallic Phase of Pt/Al2O3 and Na-Doped Pt/Al2O3 Catalysts for Light Paraffins Dehydrogenation. *Chemical Engineering Journal*, Vol.118, No.3,

in the Preparation of Supported Catalysts. *Catalysis Reviews – Science and* 

PdO. *Journal of Electron Spectroscopy and Related Phenomena*, Vol.104, No.1-3, (July

Reduction: A Structural Investigation of Sulfur Poisoning Catalyst Deactivation.

for Tetralin Hydrogenation: Effects of Sulfur-Poisoning and Hydrogen Reactivation. *Journal of Catalysis*, Vol.176, No.1, (May 1998), pp. 42-51, ISSN 0021-

Pd-Based Catalysts. *Environmental Science & Technology*, Vol.41, No.15, (June 2007),

Boehmite Powder. *Journal of Applied Chemistry of the USSR*, Vol.55, pp. 2120-2122,

Structure and Strength of the Acidic Sites of Amorphous Silica Alumina: An IR and NMR Study. *Journal of Physical Chemistry B*, Vol.110, No.31, (August 2006), pp.

Thus, FTIR spectroscopy applied to investigation of Pt/SZA system proved to be a highly informative method, which allowed us to elucidate the role of alumina both in the formation of platinum sites and in the catalyst behavior in the acid-catalyzed reactions. Although state of the surface under conditions of FTIR spectroscopic examination strongly differ from its state upon contacting with aqueous solutions of metal complexes or in catalytic reactions, FTIR spectroscopy data on the state of supported platinum as well as on the nature and strength of acid sites can be used to optimize the composition of bifunctional catalyst Pt/SZA designed for hydroisomerization of benzene-containing fractions.

## **4. Conclusion**

The possibilities of FTIR spectroscopy, in particular with the use of adsorbed СО, pyridine or deuterochloroform probe molecules, for investigation of some model and industrially important supports and catalysts were demonstrated. The effect of chemical composition of a support (Al2O3, Al2O3-SiO2, SO4 2–-ZrO2, SO42–-ZrO2-Al2O3) and modification technique on the concentration and ratio of different types of OH groups and coordinatively unsaturated surface sites was shown.

Concentrations of the surface sites on supports before and after anchoring of the active metal component were compared to demonstrate a relation between composition of the functional surface groups, adsorption capacity of the support and strength of the interaction between metal complex precursor and support, and to identify the sites involved in anchoring of the active component. The impact of support nature and composition, conditions of oxidation and reduction treatments on the metal-support interaction and ratio of oxidized and reduced forms of supported metal (platinum or palladium) was revealed.

FTIR spectroscopy data for the examined catalytic systems were compared with the data of XPS, diffuse reflectance electron spectroscopy, H2 and CO chemisorption for determination of supported metal dispersion, and temperature-programmed reduction as well as with the results of testing in the following catalytic reactions: double-bond isomerization of 1-hexene, hydrogenation of benzene and isomerization of n-heptane and cyclohexane.

## **5. Acknowledgment**

This study was supported by the Russian Foundation for Basic Research, grant no. 09-03- 01013.

## **6. References**


Thus, FTIR spectroscopy applied to investigation of Pt/SZA system proved to be a highly informative method, which allowed us to elucidate the role of alumina both in the formation of platinum sites and in the catalyst behavior in the acid-catalyzed reactions. Although state of the surface under conditions of FTIR spectroscopic examination strongly differ from its state upon contacting with aqueous solutions of metal complexes or in catalytic reactions, FTIR spectroscopy data on the state of supported platinum as well as on the nature and strength of acid sites can be used to optimize the composition of bifunctional catalyst

The possibilities of FTIR spectroscopy, in particular with the use of adsorbed СО, pyridine or deuterochloroform probe molecules, for investigation of some model and industrially important supports and catalysts were demonstrated. The effect of chemical composition of a support (Al2O3, Al2O3-SiO2, SO42–-ZrO2, SO42–-ZrO2-Al2O3) and modification technique on the concentration and ratio of different types of OH groups and coordinatively unsaturated

Concentrations of the surface sites on supports before and after anchoring of the active metal component were compared to demonstrate a relation between composition of the functional surface groups, adsorption capacity of the support and strength of the interaction between metal complex precursor and support, and to identify the sites involved in anchoring of the active component. The impact of support nature and composition, conditions of oxidation and reduction treatments on the metal-support interaction and ratio of oxidized and reduced forms of supported metal (platinum or palladium) was revealed. FTIR spectroscopy data for the examined catalytic systems were compared with the data of XPS, diffuse reflectance electron spectroscopy, H2 and CO chemisorption for determination of supported metal dispersion, and temperature-programmed reduction as well as with the results of testing in the following catalytic reactions: double-bond isomerization of 1-hexene,

This study was supported by the Russian Foundation for Basic Research, grant no. 09-03-

Apesteguia, C.R., Brema, C.E., Garetto, T.F., Borgna, A., & Parera, J.M. (1984). Sulfurization

Apesteguia, C.R., Garetto, T.F., & Borgna, A. (1987). On the Sulfur-Aided Metal-Support

of Pt/Al2O3-Cl Catalysts: VI. Sulfur-Platinum Interaction Studied by Infrared Spectroscopy. *Journal of Catalysis*, Vol.89, No.1, (September 1984), pp. 52-59, ISSN

Interaction in Pt/Al2O3-Cl Catalysts. *Journal of Catalysis*, Vol.106, No.1, (July 1987),

Pt/SZA designed for hydroisomerization of benzene-containing fractions.

hydrogenation of benzene and isomerization of n-heptane and cyclohexane.

**4. Conclusion** 

surface sites was shown.

**5. Acknowledgment** 

0021-9517

pp. 73-84, ISSN 0021-9517

**6. References** 

01013.


FTIR Spectroscopy of Adsorbed Probe Molecules for

America

9517

1070-3632

7863

London, England

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 175

Kazakov, M.O., Lavrenov, A.V., Belskaya, O.B., Danilova, I.G., Arbuzov, A.B., Gulyaeva,

Knözinger, H. (1976a). Hydrogen Bonds in Systems of Adsorbed Molecules, In: *The* 

Knözinger, H. (1976b). Specific Poisoning and Characterization of Catalytically Active Oxide

Knözinger, H. & Ratnasamy, P. (1978). Catalytic Aluminas: Surface Models and

Kooh, A.B., Han, W.-J., Lee, R.G., & Hicks, R.F. (1991). Effect of Catalyst Structure and

Kubelková, L., Beran, S., & Lercher, J.A. (1989). Determination of Proton Affinity of Zeolites

Kustov, L.M., Kazansky, V.B., Figueras, F., & Tichit, D. (1994). Investigation of the Acidic

Kustov, L.M. (1997). New Trends in IR-Spectroscopic Characterization of Acid and Basic

Kwak, J.H., Hu, J., Mei, D., Yi, C.-W., Kim, D.H., Peden, C.H.F., Allard, L.F., & Szanyi, J.

Lavrenov, A.V., Basova, I.A., Kazakov, M.O., Finevich, V.P., Belskaya, O.B., Buluchevskii,

Lisboa, J. da S., Santos, D.C.R.M., Passos, F.B., & Noronha, F.B. (2005). Influence of the

Little, L.H. (1966). *Infrared Spectra of Adsorbed Species*, Academic Press, Inc, ISBN 0124521509,

Liu, X., & Truitt, R.E. (1997). DRFT-IR Studies of the Surface of γ-Alumina. *Journal of the* 

*Zeolites*, Vol.9, No.6, (November 1989), pp. 539-543, ISSN 0144-2449

(January-February 2012), Accepted, In Press, ISSN 0023-1584

0720403154, Amsterdam, The Netherlands

No.1, (January 1978), pp. 31-70, ISSN 0161-4940

(November 1994), pp. 143-149, ISSN 0021-9517

No.1, (March 2005), pp. 15-21, ISSN 0920-5861

1997), pp. 131-144, ISSN 1022-5528

pp. 1670-1673, ISSN 0036-8075

T.I., Drozdov, V.A., Duplyakin, V.K. (2012). Hydroisomerization of Benzene-Containing Gasoline Fractions on a Pt/SO42--ZrO2-Al2O3 Catalyst: III. Hydrogenating Properties of the Catalyst. *Kinetics and Catalysis*, Vol.53, No.1,

*Hydrogen Bond: Resent Developments in Theory and Experiments*, P. Schuster, G. Zundel, & C. Sandorfy, (Ed.), pp. 1263-1364, North-Holland Pub. Co., ISBN

Surfaces, In: *Advances in Catalysis*, D.D. Eley, H. Pines, & P.B. Weisz, (Ed.), pp. 184- 271, Academic Press, Inc, ISBN 0-12-007825-2, New York, The United States of

Characterization of Surface Sites. *Catalysis Reviews – Science and Engineering*, Vol.17,

Carbon Deposition on Heptane Oxidation over Supported Platinum and Palladium. *Journal of Catalysis*, Vol.130, No.2, (August 1991), pp. 374-391, ISSN 0021-

and Zeolite-Like Solids by Low-Temperature Adsorption of Carbon Monoxide.

Properties of ZrO2 Modified by SO42- Anions. *Journal of Catalysis*, Vol.150, No.1,

Sites in Zeolites and Oxide Catalysts. *Topics in Catalysis*, Vol.4, Nos.1-2, (November

(2009). Coordinatively Unsaturated Al3+ Centers as Binding Sites for Active Catalyst Phases of Platinum on γ-Al2O3. *Science*, Vol.325, No.5948 (September 2009),

E.A., & Duplyakin, V.K. (2007). Catalysts on the Basis of Anion-Modified Metal Oxides for Production of Ecologically Pure Components of Motor Fuels. *Russian Journal of General Chemistry*, Vol.77, No.12, (December 2007), pp. 2272-2283, ISSN

Addition of Promoters to Steam Reforming Catalysts. *Catalysis Today*, Vol.101,

*American Chemical Society*, Vol.119, No.41, (October 1997), pp. 9856-9860, ISSN 0002-


Della Gatta, G., Fubini, B., Ghiotti, G., & Morterra, C. (1976). The Chemisorption of Carbon

Dicko, A., Song, X.M., Adnot, A., & Sayari, A. (1994). Characterization of Platinum on

Ghosh, A.K., & Kydd, R.A. (1985). Fluorine-Promoted Catalysts. *Catalysis Reviews – Science and Engineering*, Vol.27, No.4, (December 1985), pp. 539-589, ISSN 0161-4940 Grau, J.M., Yori, J.C., Vera, C.R., Lovey, F.C., Condo, A.M., & Parera, J.M. (2004). Crystal

Guerra, C.R. (1969). Infrared Spectroscopic Studies of CO Adsorption by Metals. The Effect

Hadjiivanov, K.I., & Vayssilov, G.N. (2002). Characterization of Oxide Surfaces and Zeolites

Iglesia, E., Soled, S.L., & Kramer, G.M. (1993). Isomerization of Alkanes on Sulfated

Jorgensen, S.W., & Madix, R.J. (1985). Steric and Electronic Effects of Sulfur on CO Adsorbed

Kazakov, M.O., Lavrenov, A.V., Mikhailova, M.S., Allert, N.A., Gulyaeva, T.I., Muromtsev,

Kazakov, M.O., Lavrenov, A.V., Danilova, I.G., Belskaya, O.B., & Duplyakin, V.K. (2011).

*Catalysis*, Vol.144, No.1, (November 1993), pp. 238-253, ISSN 0021-9517 Ivanov, A.V., & Kustov, L.M. (1998). Investigation of the State of Palladium in the

*Bulletin*, Vol.47, No.1, (January 1998), pp. 55-59, ISSN 1066-5285

*Catalysis*, Vol.150, No.2, (December 1994), pp. 254-261, ISSN 0021-9517 Egorov, M.M. (1961). The Nature of the Surface of Catalytically Active Aluminum Oxide.

Nos.1-6, (September-October 1961), pp. 697-700, ISSN 0271-5007

(June 1976), pp. 90-98, ISSN 0021-9517

2004), pp. 141-152, ISSN 0926-860X

York, The United States of America

0021-9517

6028

ISSN 0023-1584

(February 1969), pp. 229-234, ISSN 0021-9797

No.4, (July 2011), pp. 573-578, ISSN 0023-1584

Hua, W., Xia, Y., Yue, Y., & Gao, Z. (2000). Promoting Effect of Al on SO4

Monoxide on Various Transition Aluminas. *Journal of Catalysis*, Vol.43, Nos.1-3,

Sulfated Zirconia Catalysts by Temperature Programmed Reduction. *Journal of* 

*Proceedings of the Academy of Sciences of the USSR, Physical Chemistry Section*, Vol.140,

Phase Dependent Metal-Support Interactions in Pt/SO42--ZrO2 Catalysts for Hydroconversion of n-Alkanes. *Applied Catalysis A: General*, Vol.265, No.2, (July

of Other Gases in Adsorption. *Journal of Colloid and Interface Science*, Vol.29, No.2,

by Carbon Monoxide as an IR Probe Molecule, In: *Advances in Catalysis*, B.C. Gates, H. Knözinger, (Ed.), pp. 307-511, Academic Press, Inc, ISBN 0-12-007847-3, New

Fe) Catalysts. *Journal of Catalysis*, Vol.196, No.1, (November 2000), pp. 104-114, ISSN

Zirconia: Promotion by Pt and by Adamantyl Hydride Transfer Species. *Journal of* 

Pd/SO4/ZrO2 System by Diffuse-Reflectance IR Spectroscopy. *Russian Chemical* 

on Pd(100). *Surface Science*, Vol.163, No.1, (November 1985), pp. 19-38, ISSN 0039-

I.V., Drozdov, V.A., & Duplyakin, V.K. (2010). Hydroisomerization of Benzene-Containing Gasoline Fractions on a Pt/SO42--ZrO2-Al2O3 Catalyst: I. Effect of Chemical Composition on the Phase State and Texture Characteristics of SO42-- ZrO2-Al2O3 Supports. *Kinetics and Catalysis*, Vol.51, No.3, (June 2010), pp. 438-443,

Hydroisomerization of Benzene-Containing Gasoline Fractions on a Pt/SO42--ZrO2- Al2O3 Catalyst: II. Effect of Chemical Composition on Acidic and Hydrogenating and the Occurrence of Model Isomerization Reactions. *Kinetics and Catalysis*, Vol.52,

2-/MxOy (M=Zr, Ti,


FTIR Spectroscopy of Adsorbed Probe Molecules for

Analyzing the Surface Properties of Supported Pt (Pd) Catalysts 177

Paukshtis, E.A., & Yurchenko, E.N. (1983). Study of the Acid-Base Properties of

Paukshtis, E.A. (1992). *Infrared Spectroscopy of Heterogeneous Acid-Base Catalysis* (in Russian),

Paze, C., Bordiga, S., Lamberti, C., Salvalaggio, M., Zecchina, A., & Bellussi, G. (1997). Acidic

Peri, J.B. (1965). A Model for the Surface of γ-Alumina. *Journal of Physical Chemistry*, Vol.69,

Pillo, T., Zimmermann, R., Steiner, P., & Hufner, S. (1997). The Electronic Structure of PdO

Requies, J., Cabrero, M.A., Barrio, V.L., Cambra, J.F., Güemez, M.B., Arias, P.L., La Parola,

Rombi, E., Cutrufello, M.G., Solinas, V., De Rossi, S., Ferraris, G., & Pistone, A. (2003).

Ryczkowski, J. (2001). IR Spectroscopy in Catalysis. *Catalysis Today*, Vol.68, No.4, (July 2001),

Scokart, P.O., Selim, S.A., Damon, J.P., & Rouxhet, P.G. (1979). The Chemistry and Surface

Shelimov, B.N., Lambert, J.-F., Che, M., & Didillon, B. (1999). Initial Steps of the Alumina-

Shelimov, B.N., Lambert, J.-F., Che, M., & Didillon, B. (2000). Molecular-Level Studies of

Sheppard, N., & Nguyen, T.T. (1978). The Vibrational Spectra of Carbon Monoxide

*Today*, Vol.116, No.3, (August 2006), pp. 304-312, ISSN 0920-5861

Vol.251, No.2, (September 2003), pp. 255-266, ISSN 0926-860X

No.2, (June 1979), pp. 209-222, ISSN 0021-9797

Vol.52, No.3, (March 1983), pp. 242-258, ISSN 0036-021X

No.24, (June 1997), pp. 4740-4751, ISSN 1520-6106

No.1, (January 1965), pp. 220-230, ISSN 0022-3654

Nauka, ISBN 5-02-029281-8, Novosibirsk

1992), pp. 869-875, ISSN 1144-0546

pp. 263-381, ISSN 0920-5861

1999), pp. 462-478, ISSN 0021-9517

ISSN 1381-1169

United States of America

Heterogeneous Catalysts by Infrared Spectroscopy. *Russian Chemical Reviews*,

Properties of H−β Zeolite as Probed by Bases with Proton Affinity in the 118-204 kcal mol-1 Range: A FTIR Investigation. *Journal of Physical Chemistry B*, Vol.101,

Found by Photoemission (UPS and XPS) and Inverse Photoemission (BIS). *Journal of Physics: Condensed Matter*, Vol.9, No.19, (May 1997), pp. 3987-3999, ISSN 0953-8984 Rakai, A., Tessier, D., & Bozon-Verduraz, F. (1992). Palladium-Alumina Catalysts – a Diffuse

Reflectance Study. *New Journal of Chemistry*, Vol.16, No.8-9, (August-September

V., Peña, M.A., & Fierro, J.L.G. (2006). Nickel/Alumina Catalysts Modified by Basic Oxides for the Production of Synthesis Gas by Methane Partial Oxidation. *Catalysis* 

Effects of Potassium Addition on the Acidity and Reducibility of Chromia/Alumina Dehydrogenation Catalysts. *Applied Catalysis A: General*,

Chemistry of Fluorinated Alumina. *Journal of Colloid and Interface Science*, Vol.70,

Supported Platinum Catalyst Preparation: a Molecular Study by 195Pt NMR, UV-Visible, EXAFS and Raman Spectroscopy. *Journal of Catalysis*, Vol.185, No.2, (July

Transition Metal – Support Interactions During the First Steps of Catalysts Preparation: Platinum Speciation in the Hexachloroplatinate/Alumina System*. Journal of Molecular Catalysis A: Chemical*, Vol.158, No.1, (September 2000), pp. 91-99,

Chemisorbed on the Surface of Metal Catalysts – a Suggested Scheme of Interpretation, In: *Advances in Infrared and Raman Spectroscopy*, R.J.H. Clark, R.E. Hester, (Ed.), pp. 67-147, Heyden & Son Inc, ISBN 0-85501-185-8, Philadelphia, The


López Cordero, R., Gil Llambías, F.J., Palacios, J.M., Fierro, J.L.G., & López Agudo, A. (1989).

Lycourghiotis, A. (2009). Interfacial Chemistry, In: *Synthesis of Solid Catalysts*, K.P. de Jong,

Maache, M., Janin, A., Lavalley, J.C., Joly, J.F., & Benazzi E. (1993). Acidity of Zeolites Beta

Manoilova, O.V., Olindo, R., Otero Areán, C., & Lercher, J.A. (2007). Variable Temperature

Mason, M.G. (1983). Electronic Structure of Supported Small Metal Clusters. *Physical Review* 

Mei, D., Kwak, J.H., Hu, J., Cho, S.J., Szanyi, J., Allard, L.F., & Peden, C.H.F. (2010). Unique

Mironenko, R.M., Belskaya, O.B., & Likholobov, V.A. (2009). Investigation of the Interaction

Mironenko, R.M., Belskaya, O.B., Danilova, I.G., Talsi, V.P., & Likholobov, V.A. (2011).

Morterra, C., Cerrato, G., Emanuel, C., & Bolis, V. (1993). On the Surface Acidity of Some

Morterra, C., & Magnacca, G. (1996). A Case Study: Surface Chemistry and Surface Structure

Morterra, C., Cerrato, G., Di Ciero, S., Signoretto, M., Pinna, F., & Strukul, G. (1997).

Otto, K., Haack, L.P., & de Vries, J.E. (1992). Identification of Two Types of Oxidized

*Environmental*, Vol.1, No.1, (February 1992), pp. 1-12, ISSN 0926-3373

*B*, Vol.27, No.2, (January 1983), pp. 748-762, ISSN 0163-1829

*Catalysis*, Vol.56, No.2, (August 1989), pp. 197-206, ISSN 0166-9834

Weinheim, Germany

426, ISSN 0144-2449

(May 1996), pp. 415-419, ISSN 0920-5861

Petersburg, Russia, June 15-19, 2009

0023-1584

5861

ISSN 0021-9517

349-367, ISSN 0021-9517

(September 2010), pp. 2688-2691, ISSN 1948-7185

Surface Changes of Alumina Induced by Phosphoric Acid Impregnation. *Applied* 

(Ed.), pp. 13-31, WILEY-VCH Verlag GmbH & Co. KGaA, ISBN 978-3-527-32040-0,

Dealuminated by Acid Leaching: An *FT*i.r. Study Using Different Probe Molecules (Pyridine, Carbon Monoxide). *Zeolites*, Vol.13, No.6, (July-August 1993), pp. 419-

FTIR Study on the Surface Acidity of Variously Treated Sulfated Zirconias. *Catalysis Communications*, Vol.8, No.6, (June 2007), pp. 865-870, ISSN 1566-7367 Marceau, E., Che, M., Saint-Just, J., & Tatibouёt, J.M. (1996). Influence of Chlorine Ions in

Pt/Al2O3 Catalysts for Methane Total Oxidation. *Catalysis Today*, Vol.29, Nos.1-4,

Role of Anchoring Penta-Coordinated Al3+ Sites in the Sintering of γ-Al2O3- Supported Pt Catalysts. *Journal of Physical Chemistry Letters*, Vol.1, No.18,

of Chloride Complexes of Platinum (IV) with Aluminum Oxide in a Structural Modification of the Surface (in Russian), *Proceedings of 24th International Chugaev Conference on Coordination Chemistry*, pp. 116-117, ISBN 5-85263-026-8, Saint

Modifying the Functional Cover of the γ-Al2O3 Surface Using Organic Salts of Aluminum. *Kinetics and Catalysis*, Vol.52, No.4, (August 2011), pp. 629-636, ISSN

Sulfate-Doped ZrO2 Catalysts. *Journal of Catalysis*, Vol.142, No.2, (August 1993), pp.

of Catalytic Aluminas, as Studied by Vibrational Spectroscopy of Adsorbed Species. *Catalysis Today*, Vol.27, Nos.3-4, (February 1996), pp. 497-532, ISSN 0920-

Platinum-Promoted and Unpromoted Sulfated Zirconia Catalysts Prepared by a One-Step Aerogel Procedure: 1. Physico-Chemical and Morphological Characterization. *Journal of Catalysis*, Vol.165, No.2, (January 1997), pp. 172-183,

Palladium on γ-Alumina by X-ray Photoelectron Spectroscopy. *Applied Catalysis B:* 


**8** 

**Hydrothermal Treatment of Hokkaido Peat –** 

There have been great changes in attitude toward the use of peat as an energy source since World War II (WEC, 2001). In Japan, peatland covers over 2500 km2 and accounts for a total energy resource of approximately 1.99 GJ.1010 (Spedding, 1988). Peatland in Japan is widely distributed throughout Hokkaido, which is the northernmost area of the country's four main islands. Although peatland also exists in other regions, its distribution is extremely localized. Peatland is distributed over an area of approximately 2000 km2 in Hokkaido (Noto, 1991), which is equivalent to approximately 6% of the flat area on this island. Peatland is also widespread in the northeastern part of Sapporo, which is the largest city in Hokkaido. Peatland in Japan is often lacustrine peat, which is formed when lakes and marshes become filled with dead plants from their surrounding areas and are then transformed into land. This type of peat is characterized by the spongy formation of plant fiber. In the peatland of Hokkaido, peat usually accumulates to a thickness of three to five meters on the ground surface, while the soft clay layer underlying is the peat is often over 20

meters thick. In some areas, a sand layer exists between the peat and the clay layers.

by Bergius in 1913, who termed the method hydrothermal carbonization.

One approach to study artificial coalification process is dewatering and conversion by hydrothermal treatment. Hydrothermal treatment of peat has been studied recently by the authors (Mursito et al., 2010; Mursito et al., 2010). In this method, raw peat is directly transformed without pretreatment or drying, which leads to greatly reduced costs. However, few studies have been conducted to evaluate the hydrothermal treatment of raw peat by means of all coalification process. Despite this lack of study, experiments imitating coalification by subjecting materials to heating with high pressure water were reported first

**1. Introduction** 

**An Application of FTIR and 13C NMR** 

Anggoro Tri Mursito1 and Tsuyoshi Hirajima2

*1Research Centre for Geotechnology, Indonesian Institute of Sciences (LIPI), Jl. Sangkuriang Komplek LIPI, Bandung 2Department of Earth Resources Engineering, Faculty of Engineering, Kyushu University,* 

*Motooka, Nishiku, Fukuoka* 

*1Indonesia 2Japan* 

**Spectroscopy on Examining of Artificial Coalification Process and Development** 


## **Hydrothermal Treatment of Hokkaido Peat – An Application of FTIR and 13C NMR Spectroscopy on Examining of Artificial Coalification Process and Development**

Anggoro Tri Mursito1 and Tsuyoshi Hirajima2

*1Research Centre for Geotechnology, Indonesian Institute of Sciences (LIPI), Jl. Sangkuriang Komplek LIPI, Bandung 2Department of Earth Resources Engineering, Faculty of Engineering, Kyushu University, Motooka, Nishiku, Fukuoka 1Indonesia 2Japan* 

#### **1. Introduction**

178 Infrared Spectroscopy – Materials Science, Engineering and Technology

Song, X., & Sayari, A. (1996). Sulfated Zirconia-Based Strong Solid-Acid Catalysts: Recent

Tsai, K.-Y., Wang, I., & Tsai, T.-C. (2011). Zeolite Supported Platinum Catalysts for Benzene

Tsyganenko, A.A., & Filimonov, V.N. (1973). Infrared Spectra of Surface Hydroxyl Groups

Tsyganenko, A.A., & Mardilovich, P.P. (1996). Structure of Alumina Surfaces. *Journal of the* 

Vazquez-Zavala, A., Fuentes, S., & Pedraza, F. (1994). The Influence of Sulfidation on the

Vera, C.R., Pieck, C.L., Shimizu, K., Yori, J.C., & Parera, J.M. (2002). Pt/SO42--ZrO2 Catalysts

Vera, C.R., Yori, J.C., Pieck, C.L., Irusta, S., & Parera, J.M. (2003). Opposite Activation

216, Elsevier, Inc, ISBN 978-0-444-98655-9, Amsterdam, The Netherlands Zaki, M.I., & Knözinger, H. (1987). Carbon Monoxide – a Low Temperature Infrared Probe

Zalewski, D.J., Alerasool, S., & Doolin, P.K. (1999). Characterization of Catalytically Active

Zamora, M., & Córdoba, A. (1978). A Study of Surface Hydroxyl Groups on γ-Alumina.

Zecchina, A., Escalona Platero, E., & Otero Areán, C. (1987). Low Temperature CO

*Catalysis*, Vol.107, No.1, (September 1987), pp. 244-247, ISSN 0021-9517

pp. 329-412, ISSN 0161-4940

4852, ISSN 0956-5000

4332

0254-0584

0920-5861

3654

2011), pp. 73-78, ISSN 0920-5861

(December 1973), pp. 579-589, ISSN 0022-2860

No.1-2, (June 2002), pp. 169-180, ISSN 0926-860X

Progress. *Catalysis Reviews - Science and Engineering*, Vol.38, No.3, (August 1996),

Hydrogenation and Naphthene Isomerization. *Catalysis Today*, Vol.166, No.1, (May

and Crystalline Structure of Oxides. *Journal of Molecular Structure*, Vol.19,

*Chemical Society, Faraday Transactions*, Vol.92, No.23, (December 1996), pp. 4843-

Crystalline Structure of Palladium, Rhodium and Ruthenium Catalysts Supported on Silica. *Applied Surface Science*, Vol.78, No.2, (June 1994), pp. 211-218, ISSN 0169-

Prepared from Pt Organometallic Compounds. *Applied Catalysis A: General*, Vol.232,

Conditions of Acid and Metal Functions of Pt/SO42--ZrO2 Catalysts. *Applied Catalysis A: General*, Vol.240, No.1-2, (February 2003), pp. 161-176, ISSN 0926-860X Wang, H., Tan, S., & Zhi, F. (1994). IR Characterization of Base Heterogeneity of Solid

Catalysts, In: *Acid-Base Catalysis II*, H. Hattori, M. Misono, & Y. Ono, (Ed.), pp. 213-

for the Characterization of Hydroxyl Group Properties on Metal Oxide Surfaces. *Materials Chemistry and Physics*, Vol.17, Nos.1-2, (April-May 1987), pp. 201-215, ISSN

Sulfated Zirconia. *Catalysis Today*, Vol.53, No.3, (November 1999), pp. 419-432, ISSN

*Journal of Physical Chemistry*, Vol.82, No.5, (March 1978), pp. 584-588, ISSN 0022-

Adsorption on Alum-Derived Active Alumina: An Infrared Investigation. *Journal of* 

There have been great changes in attitude toward the use of peat as an energy source since World War II (WEC, 2001). In Japan, peatland covers over 2500 km2 and accounts for a total energy resource of approximately 1.99 GJ.1010 (Spedding, 1988). Peatland in Japan is widely distributed throughout Hokkaido, which is the northernmost area of the country's four main islands. Although peatland also exists in other regions, its distribution is extremely localized. Peatland is distributed over an area of approximately 2000 km2 in Hokkaido (Noto, 1991), which is equivalent to approximately 6% of the flat area on this island. Peatland is also widespread in the northeastern part of Sapporo, which is the largest city in Hokkaido. Peatland in Japan is often lacustrine peat, which is formed when lakes and marshes become filled with dead plants from their surrounding areas and are then transformed into land. This type of peat is characterized by the spongy formation of plant fiber. In the peatland of Hokkaido, peat usually accumulates to a thickness of three to five meters on the ground surface, while the soft clay layer underlying is the peat is often over 20 meters thick. In some areas, a sand layer exists between the peat and the clay layers.

One approach to study artificial coalification process is dewatering and conversion by hydrothermal treatment. Hydrothermal treatment of peat has been studied recently by the authors (Mursito et al., 2010; Mursito et al., 2010). In this method, raw peat is directly transformed without pretreatment or drying, which leads to greatly reduced costs. However, few studies have been conducted to evaluate the hydrothermal treatment of raw peat by means of all coalification process. Despite this lack of study, experiments imitating coalification by subjecting materials to heating with high pressure water were reported first by Bergius in 1913, who termed the method hydrothermal carbonization.

Hydrothermal Treatment of Hokkaido Peat - An Application of FTIR and


12.3 67.5 32.5 12.8

57.6 5.8 2.2 33.8 0.7

77.7

20508

16727

*d.b* = dry basis; *a.r* = as received basis; *d.a.f* = dry ash free basis; *diff.* = differences

86.9

13.3 68.1 31.9 4.4

54.7 5.7 1.1 38.0 0.5


21,527

17307

Stirrer

Fig. 1. Schematic figures of hydrothermal batch type reactors.

Sealed & Screw

Vent

Gasometer

Takahashi peat moss and hydrothermally upgraded peat.

Proximate analysis

Ultimate analysis *(wt%) (d.a.f)*

Yield of solid products (Y) *(wt%) (d.b)*  Calorific value (CV) *(kJ.kg-1) (d.b)*

Effective calorific value

Gas Collector

N2

(ECV) *(kJ.kg-1)*

*(wt%) Moisture (a.r) Equilibrium Moisture* (X) *(a.r) Volatile Matter (d.a.f) Fixed Carbon (d.a.f)* 

*Ash (d.b)*

*C H N O (diff.) S* 

13C NMR Spectroscopy on Examining of Artificial Coalification Process and Development 181


5.6 56.4 43.6 8.3

67.0 5.5 1.7 25.1 0.6

65.3

25008

22436

150oC 200oC 250oC 270oC 300oC 330oC 350oC 380oC





2.4 37.2 62.8 8.2

79.1 5.4 1.7 13.3 0.5

48.3

30047

28233

3.0 44.0 56.0 7.5

75.2 5.4 1.7 17.2 0.5

50.4

29296

27264

100 V

3.2 45.4 54.6 7.4

74.3 5.5 1.7 17.9 0.5

53.9

28299

26242

Temperature Control

Vessel

Heater Elements

4.8 48.4 51.6 6.0

72.2 5.5 1.6 20.0 0.6

61.2

27985

25457


5.3 52.8 47.2 7.1

68.0 5.4 1.6 24.3 0.6

65.2

25836

23268

Properties Raw Treated temperatures (oC)


12.8 65.0 35.0 9.1

59.6 5.7 1.7 32.3 0.6

77.6

22427

18263

Table 1. Proximate and ultimate analysis, yield of solid products and calorific value of

Rotary Stirrer

H2O out H2O in

and Porapak Q columns. The column temperature was set at 60° C and argon was applied as the carrier gas at a rate of 30 mL/min. The results were recorded using a Shimadzu C-R8A Chromatopac data processor. The results of GC analysis are discussed elsewhere. The solid

The conversion of cold climate peat into liquid fuel has been studied and conducted in Germany, Canada, Sweden, Finland, Iceland and Israel (Björnbom et al., 1986). Although there has been much less interest in peat liquefaction than coal liquefaction, a large number of batch autoclave studies have evaluated the use of cold climate peat as the raw material for the formation of liquid fuel. The conversion of peat to liquid fuel in Sweden produced an organic product similar to very heavy oil after raw peat was treated with CO under high pressures and temperatures (Björnbom et al., 1981). In Canada, the conversion of peat to gas and liquid employed CO and/or H2 and water (Cavalier & Chornet, 1977).

Hydrothermal treatment of Hokkaido cold climate peat has also been investigated. However, it is still necessary to evaluate the liquid and gas products formed during the process to facilitate its energy and chemical utilization. The aim of this chapter is to characterize and determine the effectiveness of hydrothermal treatment for upgrading and dewatering processes on the solid products of Hokkaido cold climate peat as well as to determine its artificial coalification process by applying of FTIR and 13C NMR spectroscopy. A fundamental study of the effects of processed temperature on the products of hydrothermal treatment of cold climate peat is also described in this Chapter.

## **2. Experimental**

### **2.1 Materials**

Raw cold climate peat samples were obtained from peatland areas owned by the Takahashi Peat Moss Company, Hokkaido, Japan. The site is located in a peat mining area that consists of about 40 ha that already contained an open and systematic drainage system. The peat mining method used at the site is the cut and block and dry method, and the peat moss products are primarily used for agriculture and gardening. The peat in the study area is approximately 5–10 m thick and the water level is about 50 cm. Prior to World War II, about 1200 ha of peatland in this area were owned by the Japan Oil-Petroleum Company, which converted the harvested peat into oil. The typical properties of the peat from this mining site are shown in Table 1.

#### **2.2 Apparatus and experimental procedure**

All experiments were conducted in a 0.5 L batch-type reactor (Taiatsu Techno MA22) that was equipped with an automatic temperature controller and had a maximum pressure of 30 MPa and a maximum temperature of 400°C (Fig. 1) (Mursito et al., 2010). The raw peat samples were introduced to the reactor without any pretreatment except for milling. The amount of the raw peat added to the reactor was 300 g, which corresponded to 40 g of moisture-free peat. The reactor was pressurized with N2 to 2.0 MPa at ambient temperature, after which the raw peat was agitated at 200 rpm while the reaction temperature was automatically adjusted from 150°C to 380°C at an average heating rate of 6.6°C/min. Under supercritical conditions (380°C), the charge was 230 g and the initial pressure was 0.1 MPa. After the desired reaction time of 30 min, the reactor was cooled immediately.

After cooling, the gas products were released through a gasometer (Shinagawa DC-1) and their volume was determined by collection into a gas micro syringe (ITO MS-GANX00). The evolved gas composition was then determined by gas chromatography (GC) using a GC equipped with a thermal conductivity detector (Shimadzu GC-4C) using Molecular Sieve 5A


Hydrothermal Treatment of Hokkaido Peat - An Application of FTIR and 13C NMR Spectroscopy on Examining of Artificial Coalification Process and Development 181

The conversion of cold climate peat into liquid fuel has been studied and conducted in Germany, Canada, Sweden, Finland, Iceland and Israel (Björnbom et al., 1986). Although there has been much less interest in peat liquefaction than coal liquefaction, a large number of batch autoclave studies have evaluated the use of cold climate peat as the raw material for the formation of liquid fuel. The conversion of peat to liquid fuel in Sweden produced an organic product similar to very heavy oil after raw peat was treated with CO under high pressures and temperatures (Björnbom et al., 1981). In Canada, the conversion of peat to gas

Hydrothermal treatment of Hokkaido cold climate peat has also been investigated. However, it is still necessary to evaluate the liquid and gas products formed during the process to facilitate its energy and chemical utilization. The aim of this chapter is to characterize and determine the effectiveness of hydrothermal treatment for upgrading and dewatering processes on the solid products of Hokkaido cold climate peat as well as to determine its artificial coalification process by applying of FTIR and 13C NMR spectroscopy. A fundamental study of the effects of processed temperature on the products of

Raw cold climate peat samples were obtained from peatland areas owned by the Takahashi Peat Moss Company, Hokkaido, Japan. The site is located in a peat mining area that consists of about 40 ha that already contained an open and systematic drainage system. The peat mining method used at the site is the cut and block and dry method, and the peat moss products are primarily used for agriculture and gardening. The peat in the study area is approximately 5–10 m thick and the water level is about 50 cm. Prior to World War II, about 1200 ha of peatland in this area were owned by the Japan Oil-Petroleum Company, which converted the harvested peat into oil. The typical properties of the peat from this mining site

All experiments were conducted in a 0.5 L batch-type reactor (Taiatsu Techno MA22) that was equipped with an automatic temperature controller and had a maximum pressure of 30 MPa and a maximum temperature of 400°C (Fig. 1) (Mursito et al., 2010). The raw peat samples were introduced to the reactor without any pretreatment except for milling. The amount of the raw peat added to the reactor was 300 g, which corresponded to 40 g of moisture-free peat. The reactor was pressurized with N2 to 2.0 MPa at ambient temperature, after which the raw peat was agitated at 200 rpm while the reaction temperature was automatically adjusted from 150°C to 380°C at an average heating rate of 6.6°C/min. Under supercritical conditions (380°C), the charge was 230 g and the initial pressure was 0.1 MPa.

After cooling, the gas products were released through a gasometer (Shinagawa DC-1) and their volume was determined by collection into a gas micro syringe (ITO MS-GANX00). The evolved gas composition was then determined by gas chromatography (GC) using a GC equipped with a thermal conductivity detector (Shimadzu GC-4C) using Molecular Sieve 5A

After the desired reaction time of 30 min, the reactor was cooled immediately.

and liquid employed CO and/or H2 and water (Cavalier & Chornet, 1977).

hydrothermal treatment of cold climate peat is also described in this Chapter.

**2. Experimental** 

are shown in Table 1.

**2.2 Apparatus and experimental procedure** 

**2.1 Materials** 

*d.b* = dry basis; *a.r* = as received basis; *d.a.f* = dry ash free basis; *diff.* = differences

Table 1. Proximate and ultimate analysis, yield of solid products and calorific value of Takahashi peat moss and hydrothermally upgraded peat.

Fig. 1. Schematic figures of hydrothermal batch type reactors.

and Porapak Q columns. The column temperature was set at 60° C and argon was applied as the carrier gas at a rate of 30 mL/min. The results were recorded using a Shimadzu C-R8A Chromatopac data processor. The results of GC analysis are discussed elsewhere. The solid

Hydrothermal Treatment of Hokkaido Peat - An Application of FTIR and

samples were heated at 100 °C for 3 hours to obtain the cellulose.

**3.1 Sequential extraction of raw peat and solid products** 

substances in Takahashi moss peat are described in Table 2.

Humic substances, insoluble fractions and lignin

Extracted peat bitumen (benzene/ethanol soluble)

*Humic and fulvic acids (HA and FA) Humin (Hm), insoluble fractions and lignin* 

Water soluble compounds

Carbohydrates *Hemicelluloses Cellulose*

separated by filtration and centrifugation.

**3. Results and discussion** 

13C NMR Spectroscopy on Examining of Artificial Coalification Process and Development 183

8 hours with a mixture of benzene and ethanol (4:1 vol./vol.). The bitumen-free peat was then extracted in ultra-pure water at 100° C for 5 hours to obtain the water-soluble materials. Next, the samples were dried at room temperature, after which the residual peat was sequentially extracted with 2% HCl at 100 °C for 5 hours to obtain the hemicelluloses. The samples were then incubated in 72% H2SO4 at room temperature for 4 hours, after which the concentration of H2SO4 was adjusted to 4% by the addition of ultra-pure water and the

The residual peat was then correlated with the humic substances, lignin and other insoluble compounds. To accomplish this, the humic substances were removed by acid and alkali extraction. Briefly, the air-dried peat was washed twice with 0.01 M HCl and then treated with NaOH at pH 13.5 for 48 hours, which gave a supernatant fraction (humic and fulvic acids), and a fraction that contained the humin and other insolubles. All fractions were

The liquid content of the raw peat was 86.9 wt.%, which corresponded to the moisture content, and the products increased from 87.3 wt.% to 89.4 wt.% as the temperature increased. As the temperature increased, the gas products content increased from 1.6 wt.% to 4.8 wt.%. The increase in the liquid and gas products in response to increasing temperature suggest that dewatering and decomposition occurred during the process. Takahashi peat that contained most of the organic constituents of the original plant materials were least decomposed and peatification occurred shortly. The amounts and concentration of plant constituents, peat bitumen (benzene/ethanol soluble) and humic

Table 2. Humic substances, plant constituents and bitumen of Takahashi moss peat.

Extracted peat bitumen can be formed during natural decomposition and accounted for 9.0 wt.% of the raw peat, suggesting that the extracted materials correspond to the wax like hydrophobic formations. The hemicellulose and cellulose content was 30.7 wt.% and 40.7 wt.%, respectively. The high levels of carbohydrate indicate that the plant constituents still remained and were decomposed during peatification. The materials from which Takahashi moss peat is formed consist of cattails and reed grass (in Japanese, gama and ashi, which differ greatly from the raw materials that lead to the formation of tropical peat. This difference may explain why Takahashi peat differs from Pontianak peat. The water soluble

*(wt.%) dry-base*

0.9 4.2

30.7 40.7 9.0 14.5

and liquid phases were then collected from the reactor and separated by filtration (ADVANTEC 5C) using a water aspirator. The total moisture content of the filtered solid products was then determined using a moisture content analyzer (Sartorius MA 150).

## **2.3 Analysis**

The liquid product was filtered through a sheet of Advantec 0.45 µm pore size membrane filter prior to analysis. After dilution by a factor of 1000 using ultra-pure water, the total organic carbon (TOC) content and total inorganic carbon (TIC) content of the liquid product was determined using a Shimadzu TOC-5000A VCSH TOC analyzer. The organic compounds in the liquid product were identified and quantified by GC-MS (Agilent 6890N and JEOL Jms-Q1000GC (A)) using a J&W Scientific methyl silicon capillary column measuring 0.32 mm x 60 m. The split ratio was 99 and the column temperature was maintained at 40° C for 3 minutes, followed by an increase of 15 °C/min. to 250° C, which was maintained for 10 minutes. The composition of the sugar compounds in the liquid products was determined by highperformance liquid chromatography (HPLC) using a JASCO RI-2031 refractive index detector and Shodex KS-811, with 2 mM HClO4 applied at 0.7 mL/min. as the eluent. The results of liquid products content analysis are discussed elsewhere.

The elemental composition of the raw peat and solid product was determined using an elemental analyzer (Yanaco CHN Corder MT-5 and MT-6). Additionally, proximate analysis (based on JIS M 8812) total sulfur analysis (based on JIS M 8819) and calorific analysis (based on JIS M 8814) were conducted separately. The gross calorific value (CV) was measured using the bomb calorimetric method and the effective calorific value (ECV) of the sample at a constant pressure was determined based on JIS M 8814, which is followed by ISO 1928. The equilibrium moisture content of the dried solid product was further analyzed while maintaining their moisture contents according to JIS M 8811. Briefly, an aliquot of the sample was placed inside a desiccator containing saturated salt solution and then measured rapidly using a moisture content analyzer (Sartorius MA 150).

The primary components and the chemical structure of the raw peat and the solid product were further analyzed by Fourier transform infrared spectroscopy (FTIR) (JASCO 670 Plus) using the Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) technique and the JASCO IR Mentor Pro 6.5 software for spectral analysis. The cross polarization/magic angle spinning (CP/MAS) 13C NMR spectrum of raw peat and the solid product was measured using a solid state spectrometer (JEOL CMX-300). The measurement conditions were as follows: spinning speed in excess of 12 kHz, contact time of 2 ms, pulse repetition time of 7 s and scan number of 10,000. Chemical shifts are in ppm referenced to hexamethylbenzene. The curve fitting analysis of the spectrum was conducted using the Grams/AI 32 Ver. 8.0 software (Galactic Industries Corp., USA).

#### **2.4 Sequential extraction of peat bitumen, plant constituents and humic substances**

After drying at room temperature, the raw peat was milled and sieved through an 80-mesh screen, after which the plant constituents (hemicelluloses, cellulose and lignin), peat bitumen, humic substances (humic acid (HA), fulvic acid (FA) and humin (Hm)) and other insoluble contents were fractionated using methods that have been previously described. Briefly, the peat bitumen (benzene/ethanol-soluble) was extracted in a Soxhlet apparatus for 8 hours with a mixture of benzene and ethanol (4:1 vol./vol.). The bitumen-free peat was then extracted in ultra-pure water at 100° C for 5 hours to obtain the water-soluble materials. Next, the samples were dried at room temperature, after which the residual peat was sequentially extracted with 2% HCl at 100 °C for 5 hours to obtain the hemicelluloses. The samples were then incubated in 72% H2SO4 at room temperature for 4 hours, after which the concentration of H2SO4 was adjusted to 4% by the addition of ultra-pure water and the samples were heated at 100 °C for 3 hours to obtain the cellulose.

The residual peat was then correlated with the humic substances, lignin and other insoluble compounds. To accomplish this, the humic substances were removed by acid and alkali extraction. Briefly, the air-dried peat was washed twice with 0.01 M HCl and then treated with NaOH at pH 13.5 for 48 hours, which gave a supernatant fraction (humic and fulvic acids), and a fraction that contained the humin and other insolubles. All fractions were separated by filtration and centrifugation.

## **3. Results and discussion**

182 Infrared Spectroscopy – Materials Science, Engineering and Technology

and liquid phases were then collected from the reactor and separated by filtration (ADVANTEC 5C) using a water aspirator. The total moisture content of the filtered solid products was then determined using a moisture content analyzer (Sartorius MA 150).

The liquid product was filtered through a sheet of Advantec 0.45 µm pore size membrane filter prior to analysis. After dilution by a factor of 1000 using ultra-pure water, the total organic carbon (TOC) content and total inorganic carbon (TIC) content of the liquid product was determined using a Shimadzu TOC-5000A VCSH TOC analyzer. The organic compounds in the liquid product were identified and quantified by GC-MS (Agilent 6890N and JEOL Jms-Q1000GC (A)) using a J&W Scientific methyl silicon capillary column measuring 0.32 mm x 60 m. The split ratio was 99 and the column temperature was maintained at 40° C for 3 minutes, followed by an increase of 15 °C/min. to 250° C, which was maintained for 10 minutes. The composition of the sugar compounds in the liquid products was determined by highperformance liquid chromatography (HPLC) using a JASCO RI-2031 refractive index detector and Shodex KS-811, with 2 mM HClO4 applied at 0.7 mL/min. as the eluent. The results of

The elemental composition of the raw peat and solid product was determined using an elemental analyzer (Yanaco CHN Corder MT-5 and MT-6). Additionally, proximate analysis (based on JIS M 8812) total sulfur analysis (based on JIS M 8819) and calorific analysis (based on JIS M 8814) were conducted separately. The gross calorific value (CV) was measured using the bomb calorimetric method and the effective calorific value (ECV) of the sample at a constant pressure was determined based on JIS M 8814, which is followed by ISO 1928. The equilibrium moisture content of the dried solid product was further analyzed while maintaining their moisture contents according to JIS M 8811. Briefly, an aliquot of the sample was placed inside a desiccator containing saturated salt solution and then measured

The primary components and the chemical structure of the raw peat and the solid product were further analyzed by Fourier transform infrared spectroscopy (FTIR) (JASCO 670 Plus) using the Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) technique and the JASCO IR Mentor Pro 6.5 software for spectral analysis. The cross polarization/magic angle spinning (CP/MAS) 13C NMR spectrum of raw peat and the solid product was measured using a solid state spectrometer (JEOL CMX-300). The measurement conditions were as follows: spinning speed in excess of 12 kHz, contact time of 2 ms, pulse repetition time of 7 s and scan number of 10,000. Chemical shifts are in ppm referenced to hexamethylbenzene. The curve fitting analysis of the spectrum was conducted using the

**2.4 Sequential extraction of peat bitumen, plant constituents and humic substances**  After drying at room temperature, the raw peat was milled and sieved through an 80-mesh screen, after which the plant constituents (hemicelluloses, cellulose and lignin), peat bitumen, humic substances (humic acid (HA), fulvic acid (FA) and humin (Hm)) and other insoluble contents were fractionated using methods that have been previously described. Briefly, the peat bitumen (benzene/ethanol-soluble) was extracted in a Soxhlet apparatus for

liquid products content analysis are discussed elsewhere.

rapidly using a moisture content analyzer (Sartorius MA 150).

Grams/AI 32 Ver. 8.0 software (Galactic Industries Corp., USA).

**2.3 Analysis** 

## **3.1 Sequential extraction of raw peat and solid products**

The liquid content of the raw peat was 86.9 wt.%, which corresponded to the moisture content, and the products increased from 87.3 wt.% to 89.4 wt.% as the temperature increased. As the temperature increased, the gas products content increased from 1.6 wt.% to 4.8 wt.%. The increase in the liquid and gas products in response to increasing temperature suggest that dewatering and decomposition occurred during the process. Takahashi peat that contained most of the organic constituents of the original plant materials were least decomposed and peatification occurred shortly. The amounts and concentration of plant constituents, peat bitumen (benzene/ethanol soluble) and humic substances in Takahashi moss peat are described in Table 2.


Table 2. Humic substances, plant constituents and bitumen of Takahashi moss peat.

Extracted peat bitumen can be formed during natural decomposition and accounted for 9.0 wt.% of the raw peat, suggesting that the extracted materials correspond to the wax like hydrophobic formations. The hemicellulose and cellulose content was 30.7 wt.% and 40.7 wt.%, respectively. The high levels of carbohydrate indicate that the plant constituents still remained and were decomposed during peatification. The materials from which Takahashi moss peat is formed consist of cattails and reed grass (in Japanese, gama and ashi, which differ greatly from the raw materials that lead to the formation of tropical peat. This difference may explain why Takahashi peat differs from Pontianak peat. The water soluble

Hydrothermal Treatment of Hokkaido Peat - An Application of FTIR and

was calculated using the equation as described on (JIS M 8814 ) :

content in mass percentage for which the ECV is desired.

maintained at high humidity.

increased.

13C NMR Spectroscopy on Examining of Artificial Coalification Process and Development 185

solid product yield was 77.7 wt.% at the lower temperature of 150°C, while the minimum yield was 48.3 wt.% at the supercritical temperature (380°C). The decrease in the yield of the solid product was due to the extensive thermal decomposition of the raw material into liquid and gaseous products. Dewatering also occurred during the decomposition reaction. Indeed, the minimum moisture content of the filtered solid products was 20.9 wt.% when the reaction was conducted at 380°C, while the maximum moisture content of 47.9 wt.% was obtained at 150°C. Moreover, the equilibrium moisture content of the solid products obtained at 380°C was 2.4 wt.%, while it was 13.3 wt.% at 150°C. As the temperature increased, the equilibrium moisture content of the solid products decreased when the products were maintained at a constant humidity (77–79%), indicating that the product may be hydrophobic when subjected to high temperatures. In general, the use of higher temperatures for the hydrothermal treatment of raw peat improves the dewaterability of the solid product; therefore, solid products produced at higher temperatures have better resistance against moisture adsorption when they are

Lower equilibrium moisture contents resulted in the solid products having higher calorific values. In addition, the effective calorific value of peat fuel decreased as the level of moisture increased. The gross calorific value (CV) determined by the bomb calorimetric method and the effective calorific value (ECV) of the samples are shown in Table 1. The ECV

where ECV is the effective calorific value, which is the net calorific value at a constant pressure of the equilibrium moisture content sample in kJ/kg, CV is the gross calorific value at a constant volume of the dry sample in kJ/kg, H, O and N are the hydrogen, oxygen and nitrogen contents of the dry sample in mass percentage, respectively, X is the moisture

In the present study, CV increased from 20,508 kJ/kg in the 150°C product to 30,047 kJ/kg in the 380°C product, while the ECV of the 150°C product was 16,727 kJ/kg and that of the 380°C product was 28,233 kJ/kg. However, the yield decreased as the temperature

In addition, the fixed carbon content increased from 31.9 wt.% to 62.8 wt.%, and the volatile matter decreased from 68.1 wt.% to 37.2 wt.% as the temperature increased. The ash content of the raw peat and solid products also increased from 4.4 wt.% and 8.2 wt.%. All of the solid products had lower levels of volatile material than the raw peat. Moreover, the chemical variations in the C, H, N and O contents of the solid products following the hydrothermal reaction of peat at different temperatures were also very interesting. Hydrothermal treatment decomposed the raw peat, which resulted in the oxygen content decreasing from 38.0 wt.% to 13.3 wt.% as the temperature increased. These findings suggest that oxygen loss corresponds to dewatering and the decreased yield of solid product. There was also a significant correlation between oxygen loss and the calorific value. The extensive removal of oxygen–rich compounds from the raw peat resulted in a solid product with a low oxygen content and a high calorific value. Additionally, the carbon content of the solid products increased from 54.7 wt.% to 79.1 wt.%. Moreover, the hydrogen content decreased slightly while the nitrogen content increased slightly in response to increased temperature. The sulfur content was relatively stable (0.5 wt.% to 0.7 wt.%), regardless of treatment.

ECV = [CV – 212.2 x H – 0.8 x (O + N)] x (1 – 0.01 X) – 24.43 X (1)

content of the raw peat was 14.5 wt.%. The total humic substances, lignin and other insoluble fractions comprised 5.1 wt.% of the raw peat, of which HA and FA comprised 0.9 wt.%.

Figure 2 shows the effect of the processing temperature on the contents of peat bitumen, water soluble compounds and carbohydrates in the solid products. All other compounds consisted of char and other insoluble materials. The peat bitumen content in the solid products ranged from 4.7 to 26.0 wt.%, which suggests that peat bitumen formed in solid products formed in response to hydrothermal treatment. The peat bitumen decreased at low temperatures (150°C to 250°C), then increased slightly at 250°C and continued to increase with increasing temperature. Under supercritical conditions, the peat bitumen decreased slightly when compared to the peat bitumen content at 350°C, which may have been due to the intensive decomposition of organics in raw peat during the formation of gaseous and liquid products as a result of the supercritical reaction. The water soluble content accounted for 14.5 wt.% of the raw peat and 6.2 wt.% of the product formed at 380°C, indicating that this fraction decreased with increasing temperature. Both carbohydrates (hemicellulose and cellulose) decreased with increasing temperature and were no longer present in products formed at 300°C and above. Hydrothermal dewatering was highly affected by the decomposition of carbohydrates at 150°C to 270°C, suggesting that organics containing polysaccharides will be obtained in the liquid products.

Fig. 2. The effect of temperature on the concentration of peat bitumen (benzene/ethanol soluble), water soluble compounds and carbohydrates in solid products

#### **3.2 Properties of solid products**

Table 1 show the effects of temperature on the yield and moisture contents of solid products, respectively. The yield decreased as the temperature increased. Specifically, the maximum

content of the raw peat was 14.5 wt.%. The total humic substances, lignin and other insoluble fractions comprised 5.1 wt.% of the raw peat, of which HA and FA comprised 0.9 wt.%.

Figure 2 shows the effect of the processing temperature on the contents of peat bitumen, water soluble compounds and carbohydrates in the solid products. All other compounds consisted of char and other insoluble materials. The peat bitumen content in the solid products ranged from 4.7 to 26.0 wt.%, which suggests that peat bitumen formed in solid products formed in response to hydrothermal treatment. The peat bitumen decreased at low temperatures (150°C to 250°C), then increased slightly at 250°C and continued to increase with increasing temperature. Under supercritical conditions, the peat bitumen decreased slightly when compared to the peat bitumen content at 350°C, which may have been due to the intensive decomposition of organics in raw peat during the formation of gaseous and liquid products as a result of the supercritical reaction. The water soluble content accounted for 14.5 wt.% of the raw peat and 6.2 wt.% of the product formed at 380°C, indicating that this fraction decreased with increasing temperature. Both carbohydrates (hemicellulose and cellulose) decreased with increasing temperature and were no longer present in products formed at 300°C and above. Hydrothermal dewatering was highly affected by the decomposition of carbohydrates at 150°C to 270°C, suggesting that organics containing

Temperature (oC)

Benzene/Ethanol soluble

Water Soluble Hemicellulose Cellulose

50 100 150 200 250 300 350

Fig. 2. The effect of temperature on the concentration of peat bitumen (benzene/ethanol

Table 1 show the effects of temperature on the yield and moisture contents of solid products, respectively. The yield decreased as the temperature increased. Specifically, the maximum

soluble), water soluble compounds and carbohydrates in solid products

polysaccharides will be obtained in the liquid products.

Content (wt.%) d.b

0

~~

Raw

**3.2 Properties of solid products** 

10

20

30

40

50

solid product yield was 77.7 wt.% at the lower temperature of 150°C, while the minimum yield was 48.3 wt.% at the supercritical temperature (380°C). The decrease in the yield of the solid product was due to the extensive thermal decomposition of the raw material into liquid and gaseous products. Dewatering also occurred during the decomposition reaction. Indeed, the minimum moisture content of the filtered solid products was 20.9 wt.% when the reaction was conducted at 380°C, while the maximum moisture content of 47.9 wt.% was obtained at 150°C. Moreover, the equilibrium moisture content of the solid products obtained at 380°C was 2.4 wt.%, while it was 13.3 wt.% at 150°C. As the temperature increased, the equilibrium moisture content of the solid products decreased when the products were maintained at a constant humidity (77–79%), indicating that the product may be hydrophobic when subjected to high temperatures. In general, the use of higher temperatures for the hydrothermal treatment of raw peat improves the dewaterability of the solid product; therefore, solid products produced at higher temperatures have better resistance against moisture adsorption when they are maintained at high humidity.

Lower equilibrium moisture contents resulted in the solid products having higher calorific values. In addition, the effective calorific value of peat fuel decreased as the level of moisture increased. The gross calorific value (CV) determined by the bomb calorimetric method and the effective calorific value (ECV) of the samples are shown in Table 1. The ECV was calculated using the equation as described on (JIS M 8814 ) :

$$\text{ECV} = \left[\text{CV} - 212.2 \times \text{H} - 0.8 \times (\text{O} + \text{N})\right] \times (1 - 0.01 \text{ X}) - 24.43 \text{ X} \tag{1}$$

where ECV is the effective calorific value, which is the net calorific value at a constant pressure of the equilibrium moisture content sample in kJ/kg, CV is the gross calorific value at a constant volume of the dry sample in kJ/kg, H, O and N are the hydrogen, oxygen and nitrogen contents of the dry sample in mass percentage, respectively, X is the moisture content in mass percentage for which the ECV is desired.

In the present study, CV increased from 20,508 kJ/kg in the 150°C product to 30,047 kJ/kg in the 380°C product, while the ECV of the 150°C product was 16,727 kJ/kg and that of the 380°C product was 28,233 kJ/kg. However, the yield decreased as the temperature increased.

In addition, the fixed carbon content increased from 31.9 wt.% to 62.8 wt.%, and the volatile matter decreased from 68.1 wt.% to 37.2 wt.% as the temperature increased. The ash content of the raw peat and solid products also increased from 4.4 wt.% and 8.2 wt.%. All of the solid products had lower levels of volatile material than the raw peat. Moreover, the chemical variations in the C, H, N and O contents of the solid products following the hydrothermal reaction of peat at different temperatures were also very interesting. Hydrothermal treatment decomposed the raw peat, which resulted in the oxygen content decreasing from 38.0 wt.% to 13.3 wt.% as the temperature increased. These findings suggest that oxygen loss corresponds to dewatering and the decreased yield of solid product. There was also a significant correlation between oxygen loss and the calorific value. The extensive removal of oxygen–rich compounds from the raw peat resulted in a solid product with a low oxygen content and a high calorific value. Additionally, the carbon content of the solid products increased from 54.7 wt.% to 79.1 wt.%. Moreover, the hydrogen content decreased slightly while the nitrogen content increased slightly in response to increased temperature. The sulfur content was relatively stable (0.5 wt.% to 0.7 wt.%), regardless of treatment.

Hydrothermal Treatment of Hokkaido Peat - An Application of FTIR and

**3.3 Coalification properties of solid products** 

products.

**Atomic H/C Ratio**

products.

**0.2** 

**0.4**

**0.6**

**0.8**

**Bituminous**

**Anthracite**

**1.0**

**1.2**

**1.4**

**1.6**

13C NMR Spectroscopy on Examining of Artificial Coalification Process and Development 187

The carbon, hydrogen and oxygen contents of solid, liquid and gas products calculated based on analysis of the product are shown in Fig. 3. Conversion of the carbon in raw peat into solid, liquid and gas products was relatively slower than conversion of the hydrogen and oxygen. The oxygen and hydrogen content in the liquid product increased rapidly because the decomposition of peat increased suddenly, as shown in the yield of the solid

Figure 4 shows a plot of the coalification band of the cold climate peat and hydrothermally upgraded and coalified solid products. As the temperature of the hydrothermal treatment increased, the atomic H/C and O/C ratio of the solid products decreased. These results indicate that hydrothermally upgraded solid products produced at 250°C and 380°C had similar atomic H/C and O/C ratios following coalification between lignite and subbituminous coals. Heat and pressure causes a disruption of the colloidal nature of peat during hydrothermal treatment (Cavalier & Chornet, 1977; Lau et al., 1987), which results in the solid products having a low equilibrium moisture content. Extensive losses of oxygen also led to decreases in the equilibrium moisture content of the solid products. Moreover, oxygen from the peat could be removed by reduction (loss of oxygen) and dehydration reactions. Dehydration followed decarboxylation, while reduction followed by dehydrogenation (Kalkreuth & Chornet, 1982; Van Krevelen, 1950) of the solid products began at the same temperature (150°C). Hydrothermal dewatering causes dehydration, reduction and decarboxylation of the product to liquid and gas; therefore, decarboxylation

**Atomic O/C Ratio**

**Decarboxylation**

**<sup>300</sup>oC**

**Lignite <sup>270</sup><sup>o</sup> <sup>330</sup> <sup>C</sup> oC**

**<sup>200</sup><sup>o</sup> <sup>250</sup> <sup>C</sup> oC**

**Dehydration**

**Demethanation**

**<sup>150</sup>oC**

**Dehydrogenation Peat** 

 **0.02 0.1 0.2 0.3 0.4** 

Fig. 4. Coalification band representing raw moss peat and hydrothermally coalified solid

**Sub-Bituminous**

**Hydrogenation**

**Reduction Oxidation**

**<sup>350</sup><sup>o</sup> <sup>380</sup> <sup>C</sup> oC**

Fig. 3. The effect of temperature on the conversion of C, H and O of raw peat into solid, liquid and gas products.

The carbon, hydrogen and oxygen contents of solid, liquid and gas products calculated based on analysis of the product are shown in Fig. 3. Conversion of the carbon in raw peat into solid, liquid and gas products was relatively slower than conversion of the hydrogen and oxygen. The oxygen and hydrogen content in the liquid product increased rapidly because the decomposition of peat increased suddenly, as shown in the yield of the solid products.

### **3.3 Coalification properties of solid products**

186 Infrared Spectroscopy – Materials Science, Engineering and Technology

Solid product Liquid product Gas product

Temperature (oC)

50 100 150 200 250 300 350

Fig. 3. The effect of temperature on the conversion of C, H and O of raw peat into solid,

Carbon Content (wt.%) d.a.f

liquid and gas products.

Oxygen Content (wt.%) d.a.f

Hydrogen Content (wt.%) d.a.f

0

Raw

~~

20

40

60

80

100

0

~~

20

40

60

80

100

0

~~

20

40

60

80

100

Figure 4 shows a plot of the coalification band of the cold climate peat and hydrothermally upgraded and coalified solid products. As the temperature of the hydrothermal treatment increased, the atomic H/C and O/C ratio of the solid products decreased. These results indicate that hydrothermally upgraded solid products produced at 250°C and 380°C had similar atomic H/C and O/C ratios following coalification between lignite and subbituminous coals. Heat and pressure causes a disruption of the colloidal nature of peat during hydrothermal treatment (Cavalier & Chornet, 1977; Lau et al., 1987), which results in the solid products having a low equilibrium moisture content. Extensive losses of oxygen also led to decreases in the equilibrium moisture content of the solid products. Moreover, oxygen from the peat could be removed by reduction (loss of oxygen) and dehydration reactions. Dehydration followed decarboxylation, while reduction followed by dehydrogenation (Kalkreuth & Chornet, 1982; Van Krevelen, 1950) of the solid products began at the same temperature (150°C). Hydrothermal dewatering causes dehydration, reduction and decarboxylation of the product to liquid and gas; therefore, decarboxylation

Fig. 4. Coalification band representing raw moss peat and hydrothermally coalified solid products.

Hydrothermal Treatment of Hokkaido Peat - An Application of FTIR and

the other carbon functional groups.

**3.5 13C NMR results of solid products** 

13C NMR Spectroscopy on Examining of Artificial Coalification Process and Development 189

Mentor Pro 6.5 software and several publications (Kalkreuth & Chornet, 1982; Van Krevelen, 1950; Orem et al., 1996; Painter et al., 1981; Ibarra & Juan, 1985; Ibarra et al., 1996; Xuguang, 2005). Examination in a range of 3500–3300 cm-1 zone revealed a progressive lowering in relative peak intensity in –OH stretching mode of the solid products at 380oC. This peak is somewhat diminished in relative intensity, probably due to the dewatering of raw peat during hydrothermal treatment and the loss of hydroxyl-functionalized carbohydrates (Kalkreuth & Chornet, 1982). The spectrum in a range of 3000–2800 cm-1 showed existence of –CHx stretching mode in an aliphatic carbon. In addition, significant changes can also be observed in a range of 1800–1100 cm-1. The carbonyl C=O stretching vibration mode of carboxylic acid at 1770 cm-1 was observed initially but the signal was almost completely disappeared with treatment at 350–380°C. The relative intensity of the ketone carboxyl band C=O groups was clearly observed at 1680 cm-1 and shifted slightly at higher temperatures. A peak at 1470–1511 cm-1 assigned to the stretching vibration mode of C=C in aromatic ring carbons (Kalkreuth & Chornet, 1982), was gradually sharpened in relative intensity with increase in temperature. Peak assigned to the bending vibration mode of C–O–R in ethers were observed at 1080 cm-1, and with increasing temperature, were no longer present at 270°C. Peak assigned to the bending vibration mode of –C–H in aromatics was also observed at 900–700 cm-1 during the hydrothermal process at all temperatures. Peak assigned to aromatic nuclei CH at 880 cm-1 tended to sharpen in its relative intensity with temperature. Dehydration and decarboxylation were also affected, with lowering in relative intensity of OH stretch bonding and carboxyl groups being observed in response to treatment. The relative peak intensity of the aromatic ring carbons was sharply observed due to the thermal decomposition that occurred during treatment, and this greatly affected

The 13C NMR spectra of the carbon-functional groups of raw peat and hydrothermally treated solid products produced at different treated temperature are shown in Fig. 6. Determination and assignment of the peak area distribution of carbon-functional groups was based on several publications (Orem et al., 1996; Hammond et al., 1985; Freitas et al., 1999; Yoshida et al., 1987; Yoshida et al., 2002). The strongest peak in the 13C NMR spectrum was at 74–76 ppm, which corresponds to methoxyl carbons (OCH3) may have been related to the presence of carbohydrate carbons (i.e., hemicelluloses, cellulose). These were confirmed on Section 3.1 that the major components of Hokkaido peat are hemicelluloses and cellulose for about 71.4 wt.%. Secondly, the peaks at 30–32 ppm in raw peat containing aliphatic carbons (CHx (CH2 and CH3)), which was likely due to the occurrence of humic acids and related substances (i.e., humic substances). Peat contains the most important organic fraction in nature, humic substances, which are composed of humic acid (HA), fulfic acid (FA) and humin (Hm) (Cavalier & Chornet, 1977). Lignin, cellulose and hemicellulose decreased as the humification of peat increased. In addition, the peak areas of the aliphatic carbons (CHx) decreased progressively as the temperature increased. The spectrum of raw peat contained a peak area at 56–59 ppm and 64–65 ppm, which may have been due to ether carbons in lignin and cellulose in the 13C NMR spectrum. These results suggest that most organic constituents in the original plant material were least biodegraded, decomposed during peatification. These peaks decreased with increasing temperature and the ether carbon peak was nearly completely gone at 330°C. The area at 74–76 ppm, which

by hydrothermal treatment of raw tropical peat can produce organic soluble materials that contain carboxylic groups in the wastewater as well as gaseous products. Moreover, disruption of colloidal forms of peat by hydrothermal treatment can lead to extensive dehydration and possibly increase the number of organic soluble materials in wastewater.

#### **3.4 FTIR results of solid products**

Figure 5 shows the FTIR spectra of the raw peat and solid products. Assignments of the peaks in each spectrum of the main functional groups were conducted using the JASCO IR

Fig. 5. FTIR spectra of raw peat and solid products produced at all processing temperatures.

Mentor Pro 6.5 software and several publications (Kalkreuth & Chornet, 1982; Van Krevelen, 1950; Orem et al., 1996; Painter et al., 1981; Ibarra & Juan, 1985; Ibarra et al., 1996; Xuguang, 2005). Examination in a range of 3500–3300 cm-1 zone revealed a progressive lowering in relative peak intensity in –OH stretching mode of the solid products at 380oC. This peak is somewhat diminished in relative intensity, probably due to the dewatering of raw peat during hydrothermal treatment and the loss of hydroxyl-functionalized carbohydrates (Kalkreuth & Chornet, 1982). The spectrum in a range of 3000–2800 cm-1 showed existence of –CHx stretching mode in an aliphatic carbon. In addition, significant changes can also be observed in a range of 1800–1100 cm-1. The carbonyl C=O stretching vibration mode of carboxylic acid at 1770 cm-1 was observed initially but the signal was almost completely disappeared with treatment at 350–380°C. The relative intensity of the ketone carboxyl band C=O groups was clearly observed at 1680 cm-1 and shifted slightly at higher temperatures. A peak at 1470–1511 cm-1 assigned to the stretching vibration mode of C=C in aromatic ring carbons (Kalkreuth & Chornet, 1982), was gradually sharpened in relative intensity with increase in temperature. Peak assigned to the bending vibration mode of C–O–R in ethers were observed at 1080 cm-1, and with increasing temperature, were no longer present at 270°C. Peak assigned to the bending vibration mode of –C–H in aromatics was also observed at 900–700 cm-1 during the hydrothermal process at all temperatures. Peak assigned to aromatic nuclei CH at 880 cm-1 tended to sharpen in its relative intensity with temperature. Dehydration and decarboxylation were also affected, with lowering in relative intensity of OH stretch bonding and carboxyl groups being observed in response to treatment. The relative peak intensity of the aromatic ring carbons was sharply observed due to the thermal decomposition that occurred during treatment, and this greatly affected the other carbon functional groups.

## **3.5 13C NMR results of solid products**

188 Infrared Spectroscopy – Materials Science, Engineering and Technology

by hydrothermal treatment of raw tropical peat can produce organic soluble materials that contain carboxylic groups in the wastewater as well as gaseous products. Moreover, disruption of colloidal forms of peat by hydrothermal treatment can lead to extensive dehydration and possibly increase the number of organic soluble materials in wastewater.

Figure 5 shows the FTIR spectra of the raw peat and solid products. Assignments of the peaks in each spectrum of the main functional groups were conducted using the JASCO IR

1770

1680

1560

1470

1080

1170

1210

880

Wavenumber (cm-1) 4000 3500 3000 2500 2000 1500 1000 500

Fig. 5. FTIR spectra of raw peat and solid products produced at all processing temperatures.

**3.4 FTIR results of solid products** 

3450

2970

2900

Raw

150oC

200oC

Abs. (a.u.)

270oC

250oC

300oC

330oC

350oC

380oC

The 13C NMR spectra of the carbon-functional groups of raw peat and hydrothermally treated solid products produced at different treated temperature are shown in Fig. 6. Determination and assignment of the peak area distribution of carbon-functional groups was based on several publications (Orem et al., 1996; Hammond et al., 1985; Freitas et al., 1999; Yoshida et al., 1987; Yoshida et al., 2002). The strongest peak in the 13C NMR spectrum was at 74–76 ppm, which corresponds to methoxyl carbons (OCH3) may have been related to the presence of carbohydrate carbons (i.e., hemicelluloses, cellulose). These were confirmed on Section 3.1 that the major components of Hokkaido peat are hemicelluloses and cellulose for about 71.4 wt.%. Secondly, the peaks at 30–32 ppm in raw peat containing aliphatic carbons (CHx (CH2 and CH3)), which was likely due to the occurrence of humic acids and related substances (i.e., humic substances). Peat contains the most important organic fraction in nature, humic substances, which are composed of humic acid (HA), fulfic acid (FA) and humin (Hm) (Cavalier & Chornet, 1977). Lignin, cellulose and hemicellulose decreased as the humification of peat increased. In addition, the peak areas of the aliphatic carbons (CHx) decreased progressively as the temperature increased. The spectrum of raw peat contained a peak area at 56–59 ppm and 64–65 ppm, which may have been due to ether carbons in lignin and cellulose in the 13C NMR spectrum. These results suggest that most organic constituents in the original plant material were least biodegraded, decomposed during peatification. These peaks decreased with increasing temperature and the ether carbon peak was nearly completely gone at 330°C. The area at 74–76 ppm, which

Chemical Shift (ppm)

1: C=O; 2: COOH; 3: Ar–O; 4: Ar–C; 5: Ar–H; 6, 7, 8, 9 and 10: methoxyl carbons OCH3; 11: aliphatic carbons CHx (CH2 and CH3 )

Fig. 6. 13C NMR spectra of raw peat and solid products produced at all processing temperatures.

Hydrothermal Treatment of Hokkaido Peat - An Application of FTIR and

peat.

**4. Conclusion** 

**5. Acknowledgment** 

**6. References** 

13C NMR Spectroscopy on Examining of Artificial Coalification Process and Development 191

corresponds to methoxyl carbons (OCH3) may have been related to the presence of carbohydrate carbons. This peak area decreased with increasing temperature, eventually decreasing to almost undetectable limits. These findings indicate that carbohydrate carbons were decomposed easily by hydrothermal treatment. The peak representative of aromatic carbons bound to the hydrogen (Ar–H), aromatic non-oxygenated carbon (Ar–C) and aromatic oxygenated carbon (Ar–O) were observed at the area of 100–106 ppm, 127–130 ppm and 151–155 ppm respectively. Furthermore, increasing the temperature of the hydrothermal treatment resulted in an increase in relative area intensity of the aromatic carbons. Finally, the peak area of carboxyl carbons (COOH) in the region of 171–180 ppm and carbonyl carbon (C=O) peaks at 195–200 ppm were observed in the spectrum of raw

In this Chapter, the effectiveness of hydrothermal upgrading and dewatering of Hokkaido cold climate peat was evaluated at temperatures ranging from 150°C to 380°C, a maximum final pressure of 25.1 MPa and a residence time of 30 minutes. The hydrothermally dewatered peat fuel product had a significantly higher ECV than raw peat, with the raw peat having an ECV of 17,307 kJ/kg and the products having ECV values ranging from 16,727 kJ/kg to 28,233 kJ/kg. Hydrothermal dewatering may also have impacted the extensive dehydration process by causing a significant loss in the oxygen content. Additionally, the carbon content of the solid products increased from 54.7 wt.% to 79.1 wt.% as the temperature increased. The hydrothermally upgraded peat fuel also had an equilibrium moisture content that ranged from 2.4 wt.% to 13.3 wt.%. A significant loss of oxygen could result in the formation of solid products with low equilibrium moisture.

An application of FTIR and 13C NMR spectroscopy on hydrothermal coalification could determine the decomposition of organic compounds in peat at different treated temperature. Increasing the temperature of the hydrothermal treatment resulted in an increase in relative area intensity of the aromatic carbons bound to the hydrogen (Ar–H), aromatic nonoxygenated carbon (Ar–C) and aromatic oxygenated carbon (Ar–O). These mean and also

Financial support was provided by a Grant-in-Aid for Science Research (No. 18206092 and No. 21246135) from the Japan Society for the Promotion of Science (JSPS), the Global-Centre of Excellence in Novel Carbon Resource Sciences, Kyushu University and the New Energy

Björnbom, E.; Olsson, B. & Karlsson, O. (1986). Thermochemical refining of raw peat prior to

Björnbom, P.; Granath, L.; Kannel, A.; Karlsson, G.; Lindstrijm, L. & Björnbom, EP. (1981).

Cavalier, JC. & Chornet, E. (1977). Conversion of peat with carbon monoxide and water.

correspond to the increasing of aromaticity as well as coalification degree.

and Industrial Technology Development Organization (NEDO).

Liquefaction of Swedish peats. *Fuel*, Vol.60, pp. 7–13.

liquefaction. *Fuel*, Vol.65, pp. 1051–1056.

*Fuel*, Vol.56, pp. 57–64.

corresponds to methoxyl carbons (OCH3) may have been related to the presence of carbohydrate carbons. This peak area decreased with increasing temperature, eventually decreasing to almost undetectable limits. These findings indicate that carbohydrate carbons were decomposed easily by hydrothermal treatment. The peak representative of aromatic carbons bound to the hydrogen (Ar–H), aromatic non-oxygenated carbon (Ar–C) and aromatic oxygenated carbon (Ar–O) were observed at the area of 100–106 ppm, 127–130 ppm and 151–155 ppm respectively. Furthermore, increasing the temperature of the hydrothermal treatment resulted in an increase in relative area intensity of the aromatic carbons. Finally, the peak area of carboxyl carbons (COOH) in the region of 171–180 ppm and carbonyl carbon (C=O) peaks at 195–200 ppm were observed in the spectrum of raw peat.

## **4. Conclusion**

190 Infrared Spectroscopy – Materials Science, Engineering and Technology

1 2 3 456 78910 11

Raw

150<sup>o</sup> C

200o C

250<sup>o</sup> C

270<sup>o</sup> C

300o C

330<sup>o</sup> C

350<sup>o</sup> C

380o C

300 200 100 0 -100

Chemical Shift (ppm)

1: C=O; 2: COOH; 3: Ar–O; 4: Ar–C; 5: Ar–H; 6, 7, 8, 9 and 10: methoxyl carbons OCH3; 11: aliphatic

Fig. 6. 13C NMR spectra of raw peat and solid products produced at all processing

a.u

carbons CHx (CH2 and CH3 )

temperatures.

In this Chapter, the effectiveness of hydrothermal upgrading and dewatering of Hokkaido cold climate peat was evaluated at temperatures ranging from 150°C to 380°C, a maximum final pressure of 25.1 MPa and a residence time of 30 minutes. The hydrothermally dewatered peat fuel product had a significantly higher ECV than raw peat, with the raw peat having an ECV of 17,307 kJ/kg and the products having ECV values ranging from 16,727 kJ/kg to 28,233 kJ/kg. Hydrothermal dewatering may also have impacted the extensive dehydration process by causing a significant loss in the oxygen content. Additionally, the carbon content of the solid products increased from 54.7 wt.% to 79.1 wt.% as the temperature increased. The hydrothermally upgraded peat fuel also had an equilibrium moisture content that ranged from 2.4 wt.% to 13.3 wt.%. A significant loss of oxygen could result in the formation of solid products with low equilibrium moisture.

An application of FTIR and 13C NMR spectroscopy on hydrothermal coalification could determine the decomposition of organic compounds in peat at different treated temperature. Increasing the temperature of the hydrothermal treatment resulted in an increase in relative area intensity of the aromatic carbons bound to the hydrogen (Ar–H), aromatic nonoxygenated carbon (Ar–C) and aromatic oxygenated carbon (Ar–O). These mean and also correspond to the increasing of aromaticity as well as coalification degree.

## **5. Acknowledgment**

Financial support was provided by a Grant-in-Aid for Science Research (No. 18206092 and No. 21246135) from the Japan Society for the Promotion of Science (JSPS), the Global-Centre of Excellence in Novel Carbon Resource Sciences, Kyushu University and the New Energy and Industrial Technology Development Organization (NEDO).

## **6. References**

Björnbom, E.; Olsson, B. & Karlsson, O. (1986). Thermochemical refining of raw peat prior to liquefaction. *Fuel*, Vol.65, pp. 1051–1056.


**Section 2** 

**Polymers and Biopolymers** 

