**Meet the editor**

Adrián M.T. Silva, PhD, has expertise in the preparation, characterization, and application of nano- and macro-structured carbon-based materials (namely, activated carbons, carbon xerogels, carbon foams and fibers, carbon nanotubes, and graphene and its derivatives, among others). His most recent research interests are in the field of metal-free carbon catalysts and carbon-based compos-

ites for separation and reaction engineering, including water treatment as well as magnetic nanostructures for biomedical applications.

Sónia A.C. Carabineiro, PhD, has research interests in the synthesis and characterization of carbon materials (namely, activated carbons, carbon nanotubes, carbon xerogels, and microdiamonds and nanodiamonds, among others) to be employed as catalysts, templates, and supports for nanoparticles and for heterogenization of homogenous metal complexes, to be used for gas pol-

lution abatement and synthesis of high-added value products.

## Contents

## **Preface XI**


Chapter 6 **Flexible Low-Voltage Carbon Nanotube Heaters and their Applications 123** Seyram Gbordzoe, Rachit Malik, Noe Alvarez, Robert Wolf and Vesselin Shanov

Chapter 7 **Recent Trends of Reinforcement of Cement with Carbon Nanotubes and Fibers 137**

Oxana V. Kharissova, Leticia M. Torres Martínez and Boris I. Kharisov


## Preface

Chapter 7 **Recent Trends of Reinforcement of Cement with Carbon**

**Section 3 Nanofluids and Case Studies on Carbon Nanostructures 175**

Chapter 8 **Polycrystalline Diamond Thin Films for Advanced**

Chapter 9 **Nanofluids Based on Carbon Nanostructures 177**

Chapter 10 **Carbon Nanostructure‐Based Scale Sensors Using Inkjet Printing and Casting Techniques 199**

Hammad Younes, Amal Al Ghaferi and Irfan Saadat

Chapter 12 **Fully Configurable Electromagnetic Wave Absorbers by Using**

**Carbon Nanostructures 225**

**Coatings Reinforced with Well-Dispersed Graphene 211**

Davide Micheli, Roberto Pastore, Antonio Vricella and Mario

Chapter 11 **Corrosion Resistance and Tribological Properties of Epoxy**

Oxana V. Kharissova, Leticia M. Torres Martínez and Boris I. Kharisov

Hammad Younes, Amal Al Ghaferi, Irfan Saadat and Haiping Hong

**Nanotubes and Fibers 137**

**Applications 161** Irena Kratochvilova

**VI** Contents

Liu Shuan

Marchetti

Carbon materials are outstanding multipurpose resources that can show several forms. Car‐ bon atoms have the ability to bond in different manners that can assume distinct dimension‐ al arrangements from which diverse nanostructures are obtained. Graphene and its derivatives, carbon nanotubes, and diamonds are among the most recently investigated car‐ bon materials. These carbon materials are very versatile and can be easily tuned to obtain samples with different properties that can fit several various applications. This book aims to cover the topic of "Advances in Carbon Nanostructures" and includes 12 book chapters. The first nine chapters consist of significant reviews dealing with diverse carbon-based configu‐ rations and applications, while the last three chapters show some case studies in the field. These chapters are organized in three sections: (i) *Graphene and derivatives*, (ii) *Carbon nano‐ tubes and diamonds,* and (iii) *Nanofluids and case studies on carbon nanostructures*.

The first section deals with *Graphene and derivatives*. **Chapter 1** describes the preparation, properties, and applications of magnetic graphene-based nanocomposites, including materi‐ als based on elemental metals (such as Fe, Co, and Ni), magnetic nanoclusters, various mor‐ phological forms of iron oxides, ferrites, 3D graphene aerogels on Fe3O4 nanoclusters, singlemolecule magnets, and organometallic and polycomponent composites, among others. **Chapter 2** includes the latest advances on graphene-based membranes for gas separation and water purification, aiming to understand details involved in the molecular transport through these membranes and discuss the recent challenges, as well as the expected devel‐ opments in this topic. **Chapter 3** focuses on novel devices covering electric, acoustic, photon‐ ic, magnetic, and mechanical domains that can be obtained with graphene and introduces the laser-scribing technology, which enables wafer scale production of graphene devices. **Chapter 4** deals with light emission from graphene, a subject not yet fully explored, but with the potential to open the door toward obtaining atomically thin, flexible, and transparent light sources and graphene-based on-chip interconnects. **Chapter 5**, the last of the first sec‐ tion, discusses the work carried out on electrospinning of graphene oxide-based nanofiber materials, focusing on the combination of graphene oxide with polymeric materials, as well as their properties and applications.

The second section gathers the chapters dealing with *Carbon nanotubes and diamonds*. **Chap‐ ter 6** reports the design and fabrication of low-voltage carbon nanotube heaters and illus‐ trates their potential applications by the manufacturing and testing of heatable gloves and via deicing experiments. **Chapter 7**summarizes recent accomplishments in the reinforce‐ ment of cement using carbon nanotubes and fibers, with emphasis on several effects upon dispersion of carbon allotropes in concrete-water media along with their possible health im‐ pact. **Chapter 8** reviews the technological achievements in the synthesis of diamond thin films of the last decade, which led to the use of the remarkable diamond properties and to the development of a wide range of applications in various fields of engineering.

Finally, the third section covers the *nanofluids and case studies on carbon nanostructures*. **Chap‐ ter 9** reviews recent studies dealing with the use of both graphene and carbon nanotubes to improve the thermal conductivity of nanofluids (i.e., liquid suspensions of nanoparticles, which have higher thermal conductivity than the base fluids), highlighting the effect of us‐ ing magnetic fields to enhance their thermal conductivity. The last three chapters focus on specific case studies. In particular, **Chapter 10** deals with the fabrication and characteriza‐ tion of scale sensors using casting and inkjet printing of carbon nanotubes. **Chapter 11** re‐ ports the protective mechanism, wear resistance, and antifriction properties of graphenebased epoxy coatings used in marine environment. **Chapter 12** refers to fully configurable electromagnetic wave absorbers employing carbon nanostructures, a new method for the full design and configuration of layered carbon-based nanocomposite materials able to per‐ form the quasi-perfect imitation of any type of electromagnetic reflection coefficient target profile.

In summary, these book chapters present several fundamental concepts, recent advances, and future challenges in the synthesis and technological application of three target groups of nanostructured carbon materials. For this reason, this book is expected to stimulate the worldwide community working in this field, and the editors trust that readers will find this book interesting and helpful. The editors are extremely grateful to all authors, affiliated in four different continents, for their willingness to contribute to this book as well as for meet‐ ing the deadlines on time. An effort has been done to avoid overlapping of content between different chapters, but some differences in writing and style result from the large informa‐ tion of the many different contributions received. The editors also thank their main funding agencies (FCT—Fundação para a Ciência e Tecnologia through IF/01501/2013 and IF/ 01381/2013/CP1160/CT0007 with financing from the European Social Fund and the Human Potential Operational Programme and project POCI-01-0145-FEDER-006984 through FED‐ ER, COMPETE2020, and POCI) as well as the exceptional assistance of the InTech staff members, especially Ms. Romina Rovan, Ms. Ana Simčić, and Ms. Petra Nenadic, for their patience and unflagging support in all stages of the editorial process.

#### **Adrián M.T. Silva and Sónia A.C. Carabineiro**

Principal Associate Researchers Laboratory of Separation and Reaction Engineering—Laboratory of Catalysis and Materials (LSRE-LCM) Faculdade de Engenharia, Universidade do Porto Porto, Portugal

## **Graphene and Derivatives**

films of the last decade, which led to the use of the remarkable diamond properties and to

Finally, the third section covers the *nanofluids and case studies on carbon nanostructures*. **Chap‐ ter 9** reviews recent studies dealing with the use of both graphene and carbon nanotubes to improve the thermal conductivity of nanofluids (i.e., liquid suspensions of nanoparticles, which have higher thermal conductivity than the base fluids), highlighting the effect of us‐ ing magnetic fields to enhance their thermal conductivity. The last three chapters focus on specific case studies. In particular, **Chapter 10** deals with the fabrication and characteriza‐ tion of scale sensors using casting and inkjet printing of carbon nanotubes. **Chapter 11** re‐ ports the protective mechanism, wear resistance, and antifriction properties of graphenebased epoxy coatings used in marine environment. **Chapter 12** refers to fully configurable electromagnetic wave absorbers employing carbon nanostructures, a new method for the full design and configuration of layered carbon-based nanocomposite materials able to per‐ form the quasi-perfect imitation of any type of electromagnetic reflection coefficient target

In summary, these book chapters present several fundamental concepts, recent advances, and future challenges in the synthesis and technological application of three target groups of nanostructured carbon materials. For this reason, this book is expected to stimulate the worldwide community working in this field, and the editors trust that readers will find this book interesting and helpful. The editors are extremely grateful to all authors, affiliated in four different continents, for their willingness to contribute to this book as well as for meet‐ ing the deadlines on time. An effort has been done to avoid overlapping of content between different chapters, but some differences in writing and style result from the large informa‐ tion of the many different contributions received. The editors also thank their main funding agencies (FCT—Fundação para a Ciência e Tecnologia through IF/01501/2013 and IF/ 01381/2013/CP1160/CT0007 with financing from the European Social Fund and the Human Potential Operational Programme and project POCI-01-0145-FEDER-006984 through FED‐ ER, COMPETE2020, and POCI) as well as the exceptional assistance of the InTech staff members, especially Ms. Romina Rovan, Ms. Ana Simčić, and Ms. Petra Nenadic, for their

Laboratory of Separation and Reaction Engineering—Laboratory of Catalysis and Materials

**Adrián M.T. Silva and Sónia A.C. Carabineiro**

Faculdade de Engenharia, Universidade do Porto

Principal Associate Researchers

(LSRE-LCM)

Porto, Portugal

patience and unflagging support in all stages of the editorial process.

the development of a wide range of applications in various fields of engineering.

profile.

VIII Preface

## **Magnetic-Graphene-Based Nanocomposites and Respective Applications**

Oxana Vasilievna Kharissova, Beatriz Ortega García, Boris Ildusovich Kharisov and Ubaldo Ortiz Méndez

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64319

#### **Abstract**

Preparation, properties, and applications of magnetic-graphene-based nanocompo‐ sites are reviewed. Graphene magnetic nanocomposites include those on the basis of elemental metals (Fe, Co, Ni), magnetic nanoclusters, various morphological forms of iron oxides (Fe2O3, Fe3O4), ferrites MFe2O4, 3D graphene aerogels@hierarchical Fe3O4 nanoclusters, single-molecule magnets like TbPc2 (Pc: phthalocyanine), other organo‐ metallic-containing composites (benzene-metal-graphene), as well as polycomponent nanocomposites such as Ag/Fe3O4/G (G: graphene), Fe3O4/CdS/G, or FePc/Fe3O4/GO (GO: graphene oxide), among others. Their available synthesis methods consist commonly of hydrothermal and solvothermal techniques, sol-gel autocombustion, sonoelectrochemical polymerization, thermal expansion and thermal reduction, microwave-assisted technique, and covalent bonding chemical methods. Their current and potential applications are distinct devices, in particular for colorimetric detection of glucose, construction materials, analytical, sensor and biosensor applications, environmental remediation, compounds with antibacterial properties, catalysis and photocatalysis, biological imaging, oil absorption, etc.

**Keywords:** environmental applications, ferrites, grapheme, hybrids, iron oxides, lithi‐ um batteries, magnetic composites, metals

## **1. Introduction**

Graphene, which consists in a single layer of carbon atoms, has been intensively studied in the last decade due to its extraordinary structural and electronic properties, leading to numerous current and potential applications in electronic devices. Growing, doping, or adsorption of metal

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

nanoparticles and other inorganic/organic species on its surface can change, improve, and manipulate graphene magnetic and electronic properties. If these particles possess magnetic properties, the whole nanocomposite can have them too, opening new opportunities and applications for the formed hybrids.

Currently, the information on magnetic graphene and graphene oxide nanocomposites has not yet been extensively reviewed despite a number of original reports. At the same time, their importance is undoubtful due to numerous current and potential applications of graphene itself, and magnetic nanoparticles in a variety of fields of chemistry, physics, medicine and biology, in the frontiers with nanotechnology. This topic is rarely presented as short sections in some reviews [1–4], dedicated to graphene, in contrast to carbon nanotubes, where a concise review was recently published [5]. So, the main purpose of this chapter is to highlight magnetic graphene composites, and review their main synthesis methods, properties, and current and potential applications.

## **2. Metal-graphene composites**

A certain number of magnetic metals are present in the Periodic Table. Graphene composites have been reported only with some typical metals with paramagnetic properties, such as iron, cobalt, or nickel. Among theoretical studies of metal interactions with graphene surface, quantum Monte Carlo methods, studying the consequences of placing a magnetic adatom adjacent to a vacancy in a graphene sheet [6, 7], and description of electronic and magnetic properties of the graphene-ferromagnet interfaces [8] are particularly relevant. In addition, the plane-wave density functional theory (DFT) was used to study the properties of nanostructures of the type C60-M-G (M: Ti, Cr, Mn, Fe, or Ni; G: graphene) [9]. In this work, in particular, it was found that the high-spin C60-Cr-G nanostructure is more stable than its low spin analog. Structures containing distinct metals can be different. For instance, on the graphene surface, C60-Ti stands symmetrically upright, while for other metals, C60-M fragments are bent and nonequal in each composite (**Figure 1**). Authors suggested, based on calculations, that the C60- M fragment is flexible, respecting the bending motion. Importantly, the attachment of C60 to an M-G surface is more energetically favored than the decoration of graphene with C60-metal complexes. DFT studies were also carried out for Co-G composites (cobalt-vacancy defect [10] or single Co atom incorporated with divacancy in armchair graphene nanoribbon [11]). In this case, metal dopation was found to introduce magnetic properties into the formed composite.

Several reports were dedicated to *cobalt-graphene* nanocomposites. Studying single Co atoms on graphene on Pt(111) and their magnetic properties by scanning tunneling microscopy spinexcitation spectroscopy, upon hydrogen adsorption, three differently hydrogenated species were identified (**Figure 2**) [12], whose magnetic properties are very different from those of cobalt element. It was suggested that, due to the interaction of Co-G orbitals, large magnetic anisotropy stems from strong ligand field effects that take place. A series of reports studied the *intercalation* of cobalt between graphene and Ir(111) via distinct techniques. Thus, cobalt layers intercalated in between graphene and Ir(111) were characterized by a combination of X-ray magnetic circular dichroism and photoemission [13]. In particular, an induced magnetic moment in the graphene layer, antiparallelly oriented to that of cobalt, was established. Magnetic ordering was found to exist beyond monolayer intercalation. In a related report [14], cobalt-intercalated graphene on Ir(111) was studied by DFT calculations and spin-polarized scanning tunneling microscopy. It was revealed that the strong bonding between cobalt and graphene leads to a high corrugation within the Moiré pattern, which arises due to the lattice mismatch between the graphene and the Co on Ir(111).

nanoparticles and other inorganic/organic species on its surface can change, improve, and manipulate graphene magnetic and electronic properties. If these particles possess magnetic properties, the whole nanocomposite can have them too, opening new opportunities and

Currently, the information on magnetic graphene and graphene oxide nanocomposites has not yet been extensively reviewed despite a number of original reports. At the same time, their importance is undoubtful due to numerous current and potential applications of graphene itself, and magnetic nanoparticles in a variety of fields of chemistry, physics, medicine and biology, in the frontiers with nanotechnology. This topic is rarely presented as short sections in some reviews [1–4], dedicated to graphene, in contrast to carbon nanotubes, where a concise review was recently published [5]. So, the main purpose of this chapter is to highlight magnetic graphene composites, and review their main synthesis methods, properties, and current and

A certain number of magnetic metals are present in the Periodic Table. Graphene composites have been reported only with some typical metals with paramagnetic properties, such as iron, cobalt, or nickel. Among theoretical studies of metal interactions with graphene surface, quantum Monte Carlo methods, studying the consequences of placing a magnetic adatom adjacent to a vacancy in a graphene sheet [6, 7], and description of electronic and magnetic properties of the graphene-ferromagnet interfaces [8] are particularly relevant. In addition, the plane-wave density functional theory (DFT) was used to study the properties of nanostructures of the type C60-M-G (M: Ti, Cr, Mn, Fe, or Ni; G: graphene) [9]. In this work, in particular, it was found that the high-spin C60-Cr-G nanostructure is more stable than its low spin analog. Structures containing distinct metals can be different. For instance, on the graphene surface, C60-Ti stands symmetrically upright, while for other metals, C60-M fragments are bent and nonequal in each composite (**Figure 1**). Authors suggested, based on calculations, that the C60- M fragment is flexible, respecting the bending motion. Importantly, the attachment of C60 to an M-G surface is more energetically favored than the decoration of graphene with C60-metal complexes. DFT studies were also carried out for Co-G composites (cobalt-vacancy defect [10] or single Co atom incorporated with divacancy in armchair graphene nanoribbon [11]). In this case, metal dopation was found to introduce magnetic properties into the formed composite.

Several reports were dedicated to *cobalt-graphene* nanocomposites. Studying single Co atoms on graphene on Pt(111) and their magnetic properties by scanning tunneling microscopy spinexcitation spectroscopy, upon hydrogen adsorption, three differently hydrogenated species were identified (**Figure 2**) [12], whose magnetic properties are very different from those of cobalt element. It was suggested that, due to the interaction of Co-G orbitals, large magnetic anisotropy stems from strong ligand field effects that take place. A series of reports studied the *intercalation* of cobalt between graphene and Ir(111) via distinct techniques. Thus, cobalt layers intercalated in between graphene and Ir(111) were characterized by a combination of

applications for the formed hybrids.

4 Advances in Carbon Nanostructures

**2. Metal-graphene composites**

potential applications.

**Figure 1.** C60-M-G nanostructures. Reproduced with permission form the *American Chemical Society*.

**Figure 2.** Top and side view of atomic structures of CoH*<sup>n</sup>* (*n* = 1, 2, and 3) complexes. Reproduced with permission from the *APS Physics*.

In addition, nanometer-sizes films of cobalt were intercalated at the interface "graphene/ Ir(111)" and were characterized by Auger electron spectroscopy and spin-polarized low-

energy electron microscopy [15]. In resulting composites, graphene top layer was found to promote perpendicular magnetic anisotropy in the Co film. It was also revealed that the magnetic anisotropy energy is significantly larger for the graphene/Co interface than for the free metallic surface. Regarding metal intercalation, contrarily to cobalt, Fe-based intercala‐ tionofGOledtothe formationofreducedGO/Fe3Cmagnetichybrids [16].Ironcarbideparticles were encapsulated in a graphite cage, protecting them from agglomeration.

Among other available theoretical studies on Co/G nanocomposites, the single Co layer added on graphene showed ferromagnetic ordering, with perpendicular alignment to the graphene sheet [17], according to the data obtained by relativistic density functional theory, at the level of generalized gradient approximation. These properties are promising for magnetism in 2D systems despite experimental difficulties in obtaining regular structures of single layer of this metal on graphene. In addition, the oxidation state of cobalt, together with possible oxidation of metal in real conditions, is also important. Thus, XPS spectral data for graphene/Co composites, obtained from CoCl2. 6H2O diluted in ethyl alcohol and highly split graphite, showed two sets of 2*p*3/2,1/2 lines belonging to partly oxidized and metallic Co atoms [18]. The formation of this protective oxide layer prevents metallic cobalt from deep oxidation.

**Figure 3.** Scheme of synthesis of magnetic cobalt–graphene (MCG) nanocomposite. Reproduced with permission from the *Royal Society of Chemistry*.

The synthesis methods for cobalt-graphene nanocomposites frequently include high-temper‐ ature decomposition of precursors. Thus, a magnetic cobalt-G nanocomposite (**Figure 3**) was prepared by carbonizing a self-assembly of a cobalt-based metal organic framework, ZIF-67, and GO [19]. This composite based on cobalt and reduced graphene oxide (RGO) was used as a catalyst for the activation of peroxymonosulfate (PMS) in the process of decolorizing Acid Yellow dye in water. Resulting regeneration efficiency remained at 97.6% over 50 cycles, showing its effective and stable catalytic activity. The proposed mechanism of MCG activating PMS to generate sulfate radicals is shown in **Figure 4**. Another example is a combination of autocombustion and sol-gel methods, which led to Co-G nanocomposite, prepared from GO, Co(NO3)2, and citric acid as precursors [20]. A sol-gel was first prepared and then underwent autocombustion in Ar atmosphere, at 300°C, due to the action of the produced reducing agents H2 and CH4. In the formed nanocomposite, Co nanoparticles (with a diameter of about 10 nm) are homogeneously distributed on graphene surface. The same technique can be applied for loading Ni, Cu, Ag, and Bi on graphene surface. Pyrolisis was also used, for instance, in the case of cobalt phthalocyanine (CoPc), for the purpose of fabricating organic metal/graphene composites [21]. CoPc is capable of dispersing cobalt and its oxide onto graphene sheets due to π-interactions between them in the conditions of pyrolysis or oxidation. The process of the fabrication of Co/GC and Co3O4/GC is shown in **Figure 5**. The Co3O4/graphene nanocomposites showed remarkable lithium storage performance, including good rate capability, good cycle performance, and highly reversible capacity. It was suggested that CuO, Fe2O3 and other metal or metal-oxide-based graphene composites can also be prepared in this way.

energy electron microscopy [15]. In resulting composites, graphene top layer was found to promote perpendicular magnetic anisotropy in the Co film. It was also revealed that the magnetic anisotropy energy is significantly larger for the graphene/Co interface than for the free metallic surface. Regarding metal intercalation, contrarily to cobalt, Fe-based intercala‐ tionofGOledtothe formationofreducedGO/Fe3Cmagnetichybrids [16].Ironcarbideparticles

Among other available theoretical studies on Co/G nanocomposites, the single Co layer added on graphene showed ferromagnetic ordering, with perpendicular alignment to the graphene sheet [17], according to the data obtained by relativistic density functional theory, at the level of generalized gradient approximation. These properties are promising for magnetism in 2D systems despite experimental difficulties in obtaining regular structures of single layer of this metal on graphene. In addition, the oxidation state of cobalt, together with possible oxidation of metal in real conditions, is also important. Thus, XPS spectral data for graphene/Co composites, obtained from CoCl2. 6H2O diluted in ethyl alcohol and highly split graphite, showed two sets of 2*p*3/2,1/2 lines belonging to partly oxidized and metallic Co atoms [18]. The

formation of this protective oxide layer prevents metallic cobalt from deep oxidation.

**Figure 3.** Scheme of synthesis of magnetic cobalt–graphene (MCG) nanocomposite. Reproduced with permission from

The synthesis methods for cobalt-graphene nanocomposites frequently include high-temper‐ ature decomposition of precursors. Thus, a magnetic cobalt-G nanocomposite (**Figure 3**) was prepared by carbonizing a self-assembly of a cobalt-based metal organic framework, ZIF-67, and GO [19]. This composite based on cobalt and reduced graphene oxide (RGO) was used as a catalyst for the activation of peroxymonosulfate (PMS) in the process of decolorizing Acid Yellow dye in water. Resulting regeneration efficiency remained at 97.6% over 50 cycles, showing its effective and stable catalytic activity. The proposed mechanism of MCG activating PMS to generate sulfate radicals is shown in **Figure 4**. Another example is a combination of autocombustion and sol-gel methods, which led to Co-G nanocomposite, prepared from GO, Co(NO3)2, and citric acid as precursors [20]. A sol-gel was first prepared and then underwent autocombustion in Ar atmosphere, at 300°C, due to the action of the produced reducing agents H2 and CH4. In the formed nanocomposite, Co nanoparticles (with a diameter of about 10 nm)

the *Royal Society of Chemistry*.

6 Advances in Carbon Nanostructures

were encapsulated in a graphite cage, protecting them from agglomeration.

**Figure 4.** Proposed mechanism for PMS activation generating sulfate radicals. Reproduced with permission from the *Royal Society of Chemistry*.

**Figure 5.** Fabrication of Co/GC and Co3O4/GC: (1) dispersion of graphite oxide and CoPc in water by ultrasonication and subsequently chemical reduction of graphite oxide, (2) pyrolysis at 800°C, and (3) oxidation at 400°C. Reproduced with permission from the *John Wiley & Sons*.

The decolorization of organic dyes, such as the Acid Yellow as mentioned above, is one of the classical applications of graphene composites of cobalt and other magnetic metals. Cobalt nanoparticles (about 30 nm size), anchored on graphene sheets, were tested for heterogeneous oxidation of a dyeing pollutant, Orange II, with peroxymonosulfate (PMS) in aqueous solutions [22]. In comparison to pure cobalt, the incorporation of Co nanoparticles into graphene sheets resulted in much higher catalytic activity for Orange II degradation. The proposed reaction mechanism is described by reactions (1)–(7). This work shows the impor‐ tance of using other metal nanoparticles rather than iron for the decontamination of organic pollutants.

Co0 → Co2+ + 2e<sup>−</sup> *(1)* ≡Co<sup>0</sup> + O2 + 2H2O → ≡ Co2+ + 4OH<sup>−</sup> *(2)* ≡Co<sup>0</sup> + 2HSO5 − → ≡ Co2+ + 2SO4 −**.** + 2OH<sup>−</sup> *(3)* ≡Co2+ + HSO5 − → ≡ Co3+ + SO4 −**.** + OH<sup>−</sup> *(4)* ≡Co3+ + HSO5 − → ≡ Co2+ + SO5 −**.** + H+ *(5)* ≡Co<sup>0</sup> + 2 ≡ Co3+ → 3 ≡ Co2+ *(6)* SO4 −**. +** Orange II → many steps → CO2 + H2O *(7)*

**Figure 6.** (a) Optical image of the device after graphene growth: an Ni stripe is coated with CVD graphene and an Al2O3/Co electrode is then deposited on lithographed squares in UV resist. (b) Cross-sectional scheme of the junction. Reproduced with permission from the *American Chemical Society*.

In the case of *nickel*, its graphene or GO composites are not so widespread compared to cobalt or iron. Nickel itself is used as a support for graphene fabrication, for instance, by the CVD technique of CH4, leading to low-defect synthesis of bilayer graphene on evaporated poly‐ crystalline nickel films [23]. The CVD method also allows to obtain nickel coated with a few layers of graphene, resulting in a stable surface upon air exposure [24]. This way, graphenepassivated (against oxidation) ferromagnetic electrodes can be fabricated, which are suitable for spin devices (namely spin-polarized oxidation-resistant electrodes as shown in **Figure 6**).

The decolorization of organic dyes, such as the Acid Yellow as mentioned above, is one of the classical applications of graphene composites of cobalt and other magnetic metals. Cobalt nanoparticles (about 30 nm size), anchored on graphene sheets, were tested for heterogeneous oxidation of a dyeing pollutant, Orange II, with peroxymonosulfate (PMS) in aqueous solutions [22]. In comparison to pure cobalt, the incorporation of Co nanoparticles into graphene sheets resulted in much higher catalytic activity for Orange II degradation. The proposed reaction mechanism is described by reactions (1)–(7). This work shows the impor‐ tance of using other metal nanoparticles rather than iron for the decontamination of organic

**Figure 6.** (a) Optical image of the device after graphene growth: an Ni stripe is coated with CVD graphene and an Al2O3/Co electrode is then deposited on lithographed squares in UV resist. (b) Cross-sectional scheme of the junction.

pollutants.

→ Co2+ + 2e<sup>−</sup> *(1)*

8 Advances in Carbon Nanostructures

−

−

−

+ 2 ≡ Co3+ → 3 ≡ Co2+ *(6)*

+ 2HSO5

≡Co2+ + HSO5

≡Co3+ + HSO5

+ O2 + 2H2O → ≡ Co2+ + 4OH<sup>−</sup> *(2)*

→ ≡ Co2+ + 2SO4

→ ≡ Co3+ + SO4

→ ≡ Co2+ + SO5

−**.**

−**.**

−**.** + H+ *(5)*

 **+** Orange II → many steps → CO2 + H2O *(7)*

Reproduced with permission from the *American Chemical Society*.

+ 2OH<sup>−</sup> *(3)*

+ OH<sup>−</sup> *(4)*

Co0

≡Co<sup>0</sup>

≡Co<sup>0</sup>

≡Co<sup>0</sup>

SO4 −**.** In addition to CVD, graphene or GO hybrids with nickel have also been obtained by other methods, for example, grown on a graphene Moire on Rh(111) at 150 K [25]. These processes of graphene decoration with nickel nanoparticles can include simultaneous reduction of GO in the process of synthesizing GO/Ni composites [26], in particular simultaneous reduction of GO and nickel(II) ions by the one-step far-infrared-assisted method [27]. Also, nickel nano‐ particles were loaded by a graphene nanosheet by easily scalable and reproducible direct electrochemical deposition [28]. The graphene sheet decorated with magnetic nickel nanopar‐ ticles yielded a composite with soft magnetic and conductive properties, which efficiently promoted microwave absorbability (and was better than graphene alone).

*Iron*/graphene (or iron/GO) hybrids are very well studied and, together with other ironcontaining nanostructures below (iron oxides and their combinations with iron, as Fe@Fe2O3), are undoubtedly the center of magnetic metal/graphene field. The hybrids can be obtained by several methods, both classic chemical and "greener" routes. Thus, *focused solar radiation*, as a green chemical route, was applied for the fabrication of 3D metal/metal oxide nanoparticles dispersed on 2D ultrathin graphene by a simultaneous reduction and exfoliation process [29]. This procedure allowed the insertion of nanoparticles between ultrathin graphene layers acting as "spacers" between them. These metal/metal oxide nanoparticle dispersed graphene composites might have potential applications in environmental fields, conversion devices, energy storage sensing, and heterogeneous catalysis.

*Microwave irradiation* was used for the synthesis (**Figure 7**) of combined hierarchical 3D graphene/carbon nanotube/iron nanostructures (G-CNT-Fe) [30]. The formed composite consists of vertically aligned CNTs, which are grown in graphene sheets along with shorter branches of CNTs stemming out from both the vertically aligned CNTs and the graphene sheets. Zero-dimensional (0D) functional iron oxide nanoparticles decorated within this 3D hierarchical nanostructure (both on 1D nanotubes and 2D graphene sheets) provide outstand‐ ing lithium storage characteristics. Iron moieties were found to be present in mixed valence states of FeOOH and Fe2O3. In addition, iron nanoparticles can be first deposited onto a graphene/Cu substrate by *vacuum deposition* and then the hydrogenation was carried out at 1 atm of gaseous H2 and under liquid N2 [31]. Before the experiment, the stabilization of hydrogenated Fe nanoparticles on this support was predicted by DFT calculations, meanwhile the existence of Fe hydride is considered as nonreachable. Hydrogen was found to be released from hydrogenated Fe nanoparticles. In addition, a *thermal reduction* method was used for the synthesis of Fe/Fe3O4/G nanocomposite electrode material, using iron oxalate and exfoliated graphene oxide as precursors [32]. Its application, along with nickel, for rechargeable Ni/Fe alkaline batteries (discharge and charge capacities of 280 mAh/g) in hybrid electric vehicles can be applied in a large-scale production.

**Figure 7.** Two-step synthesis of hierarchical G-CNT-Fe 3D nanostructure and its application to anode material in lithi‐ um-ion batteries. (a) Step 1: microwave reaction for carbon nanotube growth. Step 2: iron nanoparticle decoration and fast explosive growth of smaller carbon nanotubes on original carbon nanotubes and graphene. (b) Schematic diagram for charge and discharge in G-CNT-Fe 3D anode material. Reproduced with permission from the *American Chemical Society*.

Sodium borohydride has been frequently used as classic reductive agent for the reduction of iron ions and/or GO. Thus, nano zero-valent iron (nZVI) -decorated graphene sheets were prepared via sodium borohydride reduction of GO and applied for Cr(VI) removal [33]. In a related report [34], nonhazardous superparamagnetic nanocomposites, consisting of iron nanoparticles (5 nm) and graphene, were synthesized from GO through intermediate forma‐ tion of Fe3+/GO complexes and their further reduction with NaBH4. Methyl blue solution, a dye often present in wastewater of dyeing industry, can easily be decolorized using this nanocomposite.

Iron/G (or GO), as well as iron oxide/G (or GO) hybrids, can also have applications in the catalysis area. A novel type of oxygen reduction reaction (ORR) electrocatalyst on the basis of few-walled (2–3 walls) CNT-G complexes was reported [35]. Abundant defects on the outer walls of the CNTs can be produced via partial unzipping of the outer CNTs walls and the formation of large quantities of graphene sheets, connected with the intact inner walls of the CNTs. These graphene sheets make easier the formation of catalytic sites for ORR on annealing in NH3. Nitrogen doping and Fe impurities are responsible for higher ORR activity in the CNT-G complexes. It was established that Fe atoms are often close to N atoms and are frequently situated along edges of defective graphene sheets. Elimination of Fe by purification leads to a considerable decrease in ORR activity. Indeed, both nitrogen and iron atomic species are important to the high ORR electrocatalytic activity. Another example is a cost-effective synthesis of nitrogen-doped graphene (NG; **Figure 8**) carried out by using cyanamide as a nitrogen source and graphene oxide as a precursor; iron nanoparticles were incorporated into NG using FeCl3 as precursor [36]. This composite was used as a model for the elucidation of the influence of nonnoble metals on the electrocatalytic performance. The NG supported with iron nanoparticles (5 wt%) showed high current density (8.20 mA cm2 ) in an alkaline solution and an excellent methanol crossover effect, high stability in distinct media, high surface area, among other advantages, in comparison with Pt and NG-based catalysts, thus allowing platinum replacement.

alkaline batteries (discharge and charge capacities of 280 mAh/g) in hybrid electric vehicles

**Figure 7.** Two-step synthesis of hierarchical G-CNT-Fe 3D nanostructure and its application to anode material in lithi‐ um-ion batteries. (a) Step 1: microwave reaction for carbon nanotube growth. Step 2: iron nanoparticle decoration and fast explosive growth of smaller carbon nanotubes on original carbon nanotubes and graphene. (b) Schematic diagram for charge and discharge in G-CNT-Fe 3D anode material. Reproduced with permission from the *American Chemical*

Sodium borohydride has been frequently used as classic reductive agent for the reduction of iron ions and/or GO. Thus, nano zero-valent iron (nZVI) -decorated graphene sheets were prepared via sodium borohydride reduction of GO and applied for Cr(VI) removal [33]. In a related report [34], nonhazardous superparamagnetic nanocomposites, consisting of iron nanoparticles (5 nm) and graphene, were synthesized from GO through intermediate forma‐ tion of Fe3+/GO complexes and their further reduction with NaBH4. Methyl blue solution, a dye often present in wastewater of dyeing industry, can easily be decolorized using this

Iron/G (or GO), as well as iron oxide/G (or GO) hybrids, can also have applications in the catalysis area. A novel type of oxygen reduction reaction (ORR) electrocatalyst on the basis of few-walled (2–3 walls) CNT-G complexes was reported [35]. Abundant defects on the outer walls of the CNTs can be produced via partial unzipping of the outer CNTs walls and the formation of large quantities of graphene sheets, connected with the intact inner walls of the CNTs. These graphene sheets make easier the formation of catalytic sites for ORR on annealing in NH3. Nitrogen doping and Fe impurities are responsible for higher ORR activity in the CNT-G complexes. It was established that Fe atoms are often close to N atoms and are frequently

can be applied in a large-scale production.

10 Advances in Carbon Nanostructures

*Society*.

nanocomposite.

**Figure 8.** (a) The content of three types of nitrogen in NG. (b) Schematic representation of NG. Reproduced with per‐ mission from the *American Chemical Society*.

Environmental problems of contamination with heavy metals and other pollutants (As, Cr) can be partially solved using iron/graphene composites, in particular those containing iron in both metallic and oxidized form. A one-pot thermal decomposition method was used for the preparation of graphene nanoplatelet composites decorated with core-shell Fe-Fe2O3 nano‐ particles [37]. These nanocomposites could be separated from the liquid-phase mixture with the aid of a permanent magnet. Efficient and effective adsorption of arsenic(III) from the polluted water was observed for this material (nearly complete As(III) removal within 1 ppb) and attributed to the increased adsorption sites existing in the presence of magnetic nanopar‐ ticles. Magnetic graphene nanocomposites (MGNCs), consisting of a core@double-shell structure of the nanoparticles with crystalline Fe as the core, iron oxide as the inner shell, and amorphous Si-S-O compound as the outer shell, were prepared by a thermodecomposition process (**Figure 9**) [38]. These composites were highly stable even in 1 M HCl aqueous acid and showed a fast and highly efficient removal of Cr(VI) from wastewater after 5 min (**Figure 10**), in contrast to several other materials (like carbon or waste biomass), the use of which require several hours or days and are not able to achieve 100% removal of Cr(VI).

**Figure 9.** Schematic illustration of the formation of the MGNCs. Reproduced with permission from the *American Chem‐ ical Society*.

**Figure 10.** Schematic adsorption mechanisms on graphene and MGNCs. Reproduced with permission from the *Ameri‐ can Chemical Society*.

Another application is related to polymers: graphene nanosheets, consisting of iron core and iron oxide shell nanoparticles, G/Fe@Fe2O3, were used as nanofillers for the fabrication of magnetic epoxy resin polymer nanocomposites [39]. The G/Fe@Fe2O3 was found to favor char formation from the epoxy resin; in addition, the tensile strength of polymer nanocomposite with 1.0 wt% Gr/Fe@Fe2O3 was found to be 58% higher than that of the pure epoxy and it was attributed to the high stiffness of graphene. It was suggested that the porous char layer with Gr/Fe@Fe2O3 may indicate the existence of Fe@Fe2O3, which helps the formation of gas, during the decomposition process of epoxy resin.

## **3. Fe2O3-graphene hybrids**

Iron can be present in graphene composites in the elemental form (as shown above) or as Fe core/oxide shell nanoparticles, iron oxides, among others. The ratio of iron nanoforms in *different oxidation states* depends, in particular, on O-containing groups present in the graphene surface, use of reductants, and other conditions. Magnetic iron-containing nanoparticles were loaded on the GO sheets due to the abundant oxygen-containing functionalities present in these carbon materials (hydroxyl, epoxy, and carboxyl functional groups), and their growth mechanism was studied [40]. Most of these functional groups were eliminated and the magnetic nanoparticles were partially converted to iron during thermal treatment under reducing conditions. Metal nanoparticles changed the GO lattice structure and intrinsic functionalities; this effect depended on the amount of iron precursor.

The *effect of pH* is also very important and it was studied for several noncovalent magnetic GObased materials, prepared using Fe2O3 microparticles, nanoparticles, and magnetic surfactants [41]. pH adjustment was used to effectively charge repulsion or attraction between Fe2O3 particles and the GO sheets (**Figure 11**). Each material caused coflocculation of GO at acidic pH, leading to materials that could be captured using an external magnetic field. The adsorp‐ tion of GO at low pH was explained by attractive electrical double-layer forces between the GO and Fe2O3 or surfactants. On the contrary, at higher values of pH, the dispersions are stable due to alike-charge repulsions. An intriguing effect was found with Fe2O3 nanoparticles: low concentrations resulted in the flocculation of GO and higher concentrations caused restabili‐ zation, which are explained by an effective overcharging of the GO surfaces. These systems were found to remove a model nanomaterial, gold nanoparticles, from water.

**Figure 9.** Schematic illustration of the formation of the MGNCs. Reproduced with permission from the *American Chem‐*

**Figure 10.** Schematic adsorption mechanisms on graphene and MGNCs. Reproduced with permission from the *Ameri‐*

Another application is related to polymers: graphene nanosheets, consisting of iron core and iron oxide shell nanoparticles, G/Fe@Fe2O3, were used as nanofillers for the fabrication of magnetic epoxy resin polymer nanocomposites [39]. The G/Fe@Fe2O3 was found to favor char formation from the epoxy resin; in addition, the tensile strength of polymer nanocomposite with 1.0 wt% Gr/Fe@Fe2O3 was found to be 58% higher than that of the pure epoxy and it was attributed to the high stiffness of graphene. It was suggested that the porous char layer with Gr/Fe@Fe2O3 may indicate the existence of Fe@Fe2O3, which helps the formation of gas, during

Iron can be present in graphene composites in the elemental form (as shown above) or as Fe core/oxide shell nanoparticles, iron oxides, among others. The ratio of iron nanoforms in *different oxidation states* depends, in particular, on O-containing groups present in the graphene surface, use of reductants, and other conditions. Magnetic iron-containing nanoparticles were loaded on the GO sheets due to the abundant oxygen-containing functionalities present in these carbon materials (hydroxyl, epoxy, and carboxyl functional groups), and their growth mechanism was studied [40]. Most of these functional groups were eliminated and the magnetic nanoparticles were partially converted to iron during thermal treatment under reducing conditions. Metal nanoparticles changed the GO lattice structure and intrinsic

functionalities; this effect depended on the amount of iron precursor.

*ical Society*.

12 Advances in Carbon Nanostructures

*can Chemical Society*.

the decomposition process of epoxy resin.

**3. Fe2O3-graphene hybrids**

**Figure 11.** (a) Conceptual scheme of the experiment: pH adjustment is used to affect charge attraction or repulsion be‐ tween the GO sheets and Fe2O3 particles. A magnetic field can be used to separate the Fe2O3 from solution or disper‐ sion. (b) Zeta potentials of graphene oxide and Fe2O3, demonstrating the pH ranges at which electrical double-layer attraction or repulsion would be expected. Reproduced with permission from the *American Chemical Society*.

The main synthesis methods for iron(III) oxide/graphene nanohybrids are, in general, similar to those used for metals described above or Fe3O4 that are explained below. Thus, iron oxide nanoparticles encased by permeable carbon layers of few-layer graphene were synthesized by high-pressure pyrolysis of ferrocene with pristine graphene [42]. The ferrocene precursor provides both carbon and iron, leading to the carbon-coated iron oxide, while the graphene works as a high-surface-area anchor, to obtain small iron oxide nanoparticles. This material was used to improve the electrochemical performance of iron-oxide-based electrodes on Liion batteries. Similarly, an iron-oleate precursor (**Figure 12**) was used for the preparation of an iron-oxide/graphene nanocomposite via a solventless thermal decomposition method [43]. Highly monodisperse γ-Fe2O3 nanoparticles were found to be in close contact with graphene. This nanomaterial can serve as a potentially valuable candidate anode material for high-rate Li-ion batteries.

**Figure 12.** Schematic representation of the direct preparation of iron-oxide/graphene nanocomposites by the solvent‐ less thermal decomposition method. Reproduced with permission from the *Royal Society of Chemistry*.

**Figure 13.** Scheme of the formation of the rGO/Fe2O3 composite. Reproduced with permission from the *American Chem‐ ical Society*.

In addition to high-temperature destructive methods, sol-gel and coprecipitation techniques are also common. Thus, reduced graphene oxide (rGO) tethered with maghemite (γ-Fe2O3) was prepared by a sol-gel process without a reducing agent in which sodium dodecylbenze‐ nesulfonate (NaDDBS) was added into the suspension for the prevention of undesirable formation of an iron oxide 3D network [44]. These composites were applied as anodes for half lithium-ion cells, exhibiting improved cycle life, reversible capacity, and good rate capability. Two-step synthesis (**Figure 13**), consisting of homogeneous precipitation and subsequent microwave-assisted reduction of the GO with hydrazine, led to reduced graphene oxide (rGO) platelets decorated with Fe2O3 nanoparticles, uniformly distributed on the surface of platelets [45]. The total specific capacity of rGO/Fe2O3 was determined to be higher than the sum of pure rGO and nanoparticular Fe2O3. In addition, polypyrrole (PPy) was reinforced with rGO and Fe2O3 to reach electrochemical stability and enhancement [46]. This ternary nanocomposite film was fabricated using a one-pot chronoamperometry approach.

The main synthesis methods for iron(III) oxide/graphene nanohybrids are, in general, similar to those used for metals described above or Fe3O4 that are explained below. Thus, iron oxide nanoparticles encased by permeable carbon layers of few-layer graphene were synthesized by high-pressure pyrolysis of ferrocene with pristine graphene [42]. The ferrocene precursor provides both carbon and iron, leading to the carbon-coated iron oxide, while the graphene works as a high-surface-area anchor, to obtain small iron oxide nanoparticles. This material was used to improve the electrochemical performance of iron-oxide-based electrodes on Liion batteries. Similarly, an iron-oleate precursor (**Figure 12**) was used for the preparation of an iron-oxide/graphene nanocomposite via a solventless thermal decomposition method [43]. Highly monodisperse γ-Fe2O3 nanoparticles were found to be in close contact with graphene. This nanomaterial can serve as a potentially valuable candidate anode material for high-rate

**Figure 12.** Schematic representation of the direct preparation of iron-oxide/graphene nanocomposites by the solvent‐

**Figure 13.** Scheme of the formation of the rGO/Fe2O3 composite. Reproduced with permission from the *American Chem‐*

In addition to high-temperature destructive methods, sol-gel and coprecipitation techniques are also common. Thus, reduced graphene oxide (rGO) tethered with maghemite (γ-Fe2O3) was prepared by a sol-gel process without a reducing agent in which sodium dodecylbenze‐ nesulfonate (NaDDBS) was added into the suspension for the prevention of undesirable

less thermal decomposition method. Reproduced with permission from the *Royal Society of Chemistry*.

Li-ion batteries.

14 Advances in Carbon Nanostructures

*ical Society*.

In addition to their use in batteries, as referred above, a few other uses are known for Fe2O3/G hybrids. Thus, functional nanocomposite-based selective separation of microcystin-LR (toxin belonging to the family of microcystins produced by cyanobacteria and known to be the most toxic of this group) from contaminated water was achieved (**Figure 14**), applying cyclodextrinfunctionalized magnetic composites consisting in porous silica and colloidal graphene [47]. In this material, the magnetic component offers easier separation of microcystin-LR from water and the cyclodextrin constituent offers host-guest interaction with microcystin-LR. Among all experimented cyclodextrins, γ-cyclodextrin was found to offer the best performance. The studied functional nanomaterials can be used for the development of advanced water purifi‐ cation systems. Catalytic applications are also reported, for example, for nanosized Fe2O3 composites on carbon matrix, modified with nitrogen-doped graphene (**Figure 15**), that are excellent catalysts for the chemoselective hydrogenation of nitroarenes to anilines with good yields (reaction (8)) [48].

**Figure 14.** Synthesis strategies for cyclodextrin functionalized magnetic graphene composite (G-Fe2O3-CD) and cyclo‐ dextrin functionalized magnetic mesoporous silica (MMS-CD). Reproduced with permission from the *American Chemi‐ cal Society*.

**Figure 15.** Nanoscaled Fe2O3 particles surrounded by N-doped graphene layers. Reproduced with permission from the *American Chemical Society*.

## **4. Fe3O4-graphene composites**

#### **4.1. Fe3O4/G aerogels**

Among a large variety of mixed-valent-iron-oxide/graphene (or GO) composites, considerable attention is paid to *aerogels* in contrast to the above-described magnetic graphene hybrids. Graphene and other aerogels and routes for their synthesis are well known [49]. The intro‐ duction of a magnetic component could lead to a larger variety of unusual properties and potential applications, namely, the easy removal of pollutants, such as crude oil. These nanocomposites can have a simple composition ( i.e., Fe3O4/G) or contain an additional component like a polymer. Thus, 3D N-doped graphene aerogel (N-GA)-supported Fe3O4 nanoparticles (Fe3O4/N-GAs; **Figure 16**) are known as efficient cathode catalysts for the oxygen reduction reaction (ORR) [50]. These hybrids were prepared via a combined hydrothermal self-assembly, freeze-drying, and thermal treatment process (**Figure 17**). The products showed an excellent electrocatalytic activity for the ORR in alkaline electrolytes, including a lower ring current, higher current density, higher electron transfer number (∼4), lower H2O2 yield, and better durability.

3D graphene aerogels containing Fe3O4 nanoparticles (Fe3O4/GA (**Figure 18**), the lightest magnetic elastomer ever reported with density about 5.8 mg**.** cm3 ) were hydrothermally prepared by self-assembly of graphene, simultaneously decorated with Fe3O4 nanoparticles [51]. They can be used to monitor the degree of compression/stretching of the material due to up to 52% reversible magnetic-field-induced strain and strain-dependent electrical resistance. Among more complex aerogels, hydrophobic graphene aerogel/Fe3O4/polystyrene composites (having extremely low density of 0.005 g cm−3, which corresponds to a volume porosity of 99.7%), with reticulated graphene structure, were solvothermally produced (**Figure 19**) [52]. Porous Fe3O4 nanoparticles were found to appear as partial substitutes for ethylenediamineassisted cross-linking and interconnections between graphene plates. These composites were applied for crude oil remediation, allowing intake capacity as much as 40 times its own mass, after 10 water-oil separation cycles.

**Figure 15.** Nanoscaled Fe2O3 particles surrounded by N-doped graphene layers. Reproduced with permission from the

Among a large variety of mixed-valent-iron-oxide/graphene (or GO) composites, considerable attention is paid to *aerogels* in contrast to the above-described magnetic graphene hybrids. Graphene and other aerogels and routes for their synthesis are well known [49]. The intro‐ duction of a magnetic component could lead to a larger variety of unusual properties and potential applications, namely, the easy removal of pollutants, such as crude oil. These nanocomposites can have a simple composition ( i.e., Fe3O4/G) or contain an additional component like a polymer. Thus, 3D N-doped graphene aerogel (N-GA)-supported Fe3O4 nanoparticles (Fe3O4/N-GAs; **Figure 16**) are known as efficient cathode catalysts for the oxygen reduction reaction (ORR) [50]. These hybrids were prepared via a combined hydrothermal self-assembly, freeze-drying, and thermal treatment process (**Figure 17**). The products showed an excellent electrocatalytic activity for the ORR in alkaline electrolytes, including a lower ring current, higher current density, higher electron transfer number (∼4), lower H2O2 yield, and

3D graphene aerogels containing Fe3O4 nanoparticles (Fe3O4/GA (**Figure 18**), the lightest

prepared by self-assembly of graphene, simultaneously decorated with Fe3O4 nanoparticles [51]. They can be used to monitor the degree of compression/stretching of the material due to up to 52% reversible magnetic-field-induced strain and strain-dependent electrical resistance.

cm3

) were hydrothermally

magnetic elastomer ever reported with density about 5.8 mg**.**

(8)

*American Chemical Society*.

16 Advances in Carbon Nanostructures

**4.1. Fe3O4/G aerogels**

better durability.

**4. Fe3O4-graphene composites**

**Figure 16.** Structure and morphology of Fe3O4/N-GAs catalysts. (a) XRD pattern and (b−d) typical SEM images of Fe3O4/N-Gas, revealing the 3D macroporous structure and uniform distribution of Fe3O4 nanoparticles in the graphene aerogels. The red rings in (d) indicate Fe3O4 nanoparticles encapsulated in thin graphene layers. Representative (e) TEM and (f) HRTEM images of Fe3O4/N-GAs showing an Fe3O4 nanoparticle wrapped by graphene layers. Repro‐ duced with permission from the *American Chemical Society*.

**Figure 17.** Fabrication process for the 3D Fe3O4/N-GAs catalyst. (a) Stable suspension of GO, iron ions, and polypyrrole (PPy) dispersed in a vial. (b) Fe- and PPy-supporting graphene hybrid hydrogel prepared by hydrothermal self-assem‐ bly and floating on water in a vial, and its ideal assembled model. (c) Monolithic Fe3O4/N-GAs hybrid aerogel obtained after freeze-drying and thermal treatment. Reproduced with permission from the *American Chemical Society*.

**Figure 18.** TEM and SEM images of the GA and Fe3O4/GA. (a) TEM image of GO and the diffraction pattern of a single flake (inset). (b) TEM image of Fe3O4 nanoparticle-decorated graphene sheets. (c, d) SEM images of microporous struc‐ tures of GA and Fe3O4/GA. (e, f) SEM images of cross-linking patterns of GA and Fe3O4/GA. Reproduced with permis‐ sion from the *American Chemical Society*.

**Figure 19.** Schematic illustration of the fabrication process of the graphene aerogel/Fe3O4/polystyrene composites CPFA and its cyclic utilization for oil removal in water. Reproduced with permission from the *American Chemical Soci‐ ety*.

### **4.2. Bicomponent Fe3O4/G hybrids**

**Figure 18.** TEM and SEM images of the GA and Fe3O4/GA. (a) TEM image of GO and the diffraction pattern of a single flake (inset). (b) TEM image of Fe3O4 nanoparticle-decorated graphene sheets. (c, d) SEM images of microporous struc‐ tures of GA and Fe3O4/GA. (e, f) SEM images of cross-linking patterns of GA and Fe3O4/GA. Reproduced with permis‐

**Figure 19.** Schematic illustration of the fabrication process of the graphene aerogel/Fe3O4/polystyrene composites CPFA and its cyclic utilization for oil removal in water. Reproduced with permission from the *American Chemical Soci‐*

sion from the *American Chemical Society*.

18 Advances in Carbon Nanostructures

*ety*.

Regarding the nonaerogel type of graphene-Fe3O4 hybrids, the hydrothermal [53] and solvothermal synthesis [54] have also been widely used although other methods are also frequently used, such as the atomic layer deposition, applied not only for Fe3O4/graphene but also for Ni/graphene composites [55]. Magnetic graphene foam with porous and hierarchical structures, on the basis of magnetite nanoparticles, was solvothermally obtained by gaseous reduction in a hydrothermal system and used for the adsorption of oil and organic solvents, thus serving for the cleanup of oil spills [56]. Fe3O4 nanoparticles on graphene foam possessed different morphologies, nanosheet arrays, or cubic structures, while controlling the reduction degree of graphene oxide under mild conditions. Distinct ratios "iron oxide/graphene" are important for several applications. Thus, G/Fe3O4 nanocomposites with different ratios of Fe3O4 to GO (mFe3O4: mG = 0.1:1, 0.2:1, 0.4:1, 0.6:1, 0.8:1 and 1:1) were prepared by solvothermal method and used for the removal of methylene blue dye from aqueous solutions [57]. The following morphologies were observed: uniform spherical homogeneously distributed Fe3O4 nanoparticles, with no agglomeration over the graphene sheets, and a uniform sheet-like shape of prepared graphene. Increasing the Fe3O4 nanoparticles on the surface of the graphene sheet was found to decrease the adsorption capacity, while the magnetization increased.

**Figure 20.** (a) A schematic representation of the preparation route of Fe3O4/rGO via a redox reaction between GO and Fe2+. Photos show a water/NH4OH (pH = 9) solution of Fe3O4/rGO (b) before and (c, left panel) after the redox reaction with Fe2+, and (c, right panel) with an applied magnet. Reproduced with permission from the *Royal Society of Chemistry*.

Several methods, considered as "greener", are sometimes reported for iron oxide/graphene composites. Thus, the "green" oxidation of Fe2+ cations in FeCl2 or FeSO4 by graphene oxide led to an *in situ* deposition of Fe3O4 nanoparticles onto the self-reduced graphene oxide (rGO) sheets (**Figure 20**) [58]. Strongly supraparamagnetic with highly chemical reactivity, electri‐ cal conductivity, good solubility, and excellent processability G@Fe3O4 nanocomposites (with an average diameter of Fe3O4 nanoparticles of 1.2–6.3 nm; coverage density of Fe3O4 nano‐ particles on graphene nanosheets of 5.3–57.9%) were prepared by a one-step "green" procedure (**Figure 21**) [59]. In addition, an approach for the deposition of iron oxide nanoparticles with *selective narrow size distribution* (0.5–7 and 1–3 μm), supported on differ‐ ent sizes of graphene oxide by coprecipitation, using Fe2+ and Fe3+ aqueous salt solutions and NH3, is described in reaction (9) [60]. The reduction of mitochondrial activity using these materials is size dependent, but the chemical functionalization of GO and Fe3O4 is a way to enhance the biocompatibility, making the system independent of the size distribution of GO.

**Figure 21.** Schematic illustration of the formation of graphene@Fe3O4. The inset shows TEM images of graphene@ Fe3O4 at reaction times of 0 (A), 1 (B), 5 (C), and 60 min (D). Scale bars are 20 nm. Reproduced with permission from the *American Chemical Society*.

(NH4)2Fe(SO4)2⋅6H2O + 2NH4Fe(SO4)2⋅12H2O + 8NH3⋅H2O → Fe3O4⋅4H2O + 6(NH4)2SO4 + 14H2O (9)

#### **4.3. Multicomponent Fe3O4/G hybrids**

More complex Fe3O4/G- and organic(inorganic)-containing systems are also common. Thus, coating a layer of mesoporous *silica* materials on each side of magnetic graphene, in the conditions of CTAB-assisted sol-gel process, with further calcination, led to obtaining magnetic graphene double-sided mesoporous nanocomposites (G/SiO2) with high surface area (168 cm2 /g) and large pore volume (0.2 cm3 /g) (**Figure 22**) [61]. The formed materials were applied to size-selective and specific enrichment and identification of peptides (peptidomics analysis) in human urine samples, protein digest solutions, and standard peptide mixtures. Another example is G@mSiO2-C18 materials (with a surface area of 315 cm2**.** g−1 and a uniform pore size of 3.3 nm) with extended plate-like morphology, prepared by coating mesoporous silica layers onto graphene via surfactant-mediated cocondensation sol-gel process, and applied as magnetic solid-phase adsorbents to the selective enrichment of phthalates in water [62].

materials is size dependent, but the chemical functionalization of GO and Fe3O4 is a way to enhance the biocompatibility, making the system independent of the size distribution of GO.

**Figure 21.** Schematic illustration of the formation of graphene@Fe3O4. The inset shows TEM images of graphene@ Fe3O4 at reaction times of 0 (A), 1 (B), 5 (C), and 60 min (D). Scale bars are 20 nm. Reproduced with permission from

(NH4)2Fe(SO4)2⋅6H2O + 2NH4Fe(SO4)2⋅12H2O + 8NH3⋅H2O → Fe3O4⋅4H2O + 6(NH4)2SO4 +

More complex Fe3O4/G- and organic(inorganic)-containing systems are also common. Thus, coating a layer of mesoporous *silica* materials on each side of magnetic graphene, in the conditions of CTAB-assisted sol-gel process, with further calcination, led to obtaining magnetic graphene double-sided mesoporous nanocomposites (G/SiO2) with high surface area (168

to size-selective and specific enrichment and identification of peptides (peptidomics analysis) in human urine samples, protein digest solutions, and standard peptide mixtures. Another

of 3.3 nm) with extended plate-like morphology, prepared by coating mesoporous silica layers

/g) (**Figure 22**) [61]. The formed materials were applied

g−1 and a uniform pore size

the *American Chemical Society*.

20 Advances in Carbon Nanostructures

**4.3. Multicomponent Fe3O4/G hybrids**

/g) and large pore volume (0.2 cm3

example is G@mSiO2-C18 materials (with a surface area of 315 cm2**.**

14H2O (9)

cm2

**Figure 22.** The workflow of synthesis of double-sided magnetic graphene/mSiO2. Reproduced with permission from the *John Wiley & Sons*.

A few reports are dedicated to *chitosan-containing* magnetic materials with mainly biomedical applications. Thus, an iron oxide/graphene oxide/chitosan (Fe3O4/GO/CS) composite was obtained by a solution mixing-evaporation method [63]. Among other data, with the incorpo‐ ration of 0.5 wt% Fe3O4 and 1 wt% GO, the tensile strength and Young's modulus of the composite significantly improved by about 28% and 74%, respectively, compared with chitosan. In addition, it was established from TGA that Fe3O4/GO/CS is less thermally stable than GO/CS composites, and graphite is more thermally stable than GO. In a related report [64], magnetic Fe3O4 nanoparticles were introduced into a water-dispersible and biocompatible chitosan-functionalized graphene, fabricated by a one-step ball milling of carboxylic chitosan and graphite. It could be an excellent catalyst for electrochemical biosensors, in particular for glucose detection, due to the presence of nitrogen (from chitosan) at the surface of graphene.

Also, superparamagnetic fluorescent Fe3O4/SiO2/G-CdTe QD s/chitosan nanocomposites (Fe3O4/SiO2/graphene-CdTe QDs/chitosan nanocomposites (FGQCs ), with a spherical diameter of 467 nm; QDs: quantum dots) were prepared and studied for improving the drug loading content [65]. In addition, a magnetic composite bioadsorbent on the basis of magnetic chitosan and graphene oxide (MCGO) was prepared [66, 67]. Its adsorption capacity for methyl blue was found to be about 90% of the initial saturation adsorption capacity after being used four times. The adsorption of methyl blue on MCGO strongly depends on ionic strength and pH, showing an ion exchange mechanism.

In the case of *polymers*, integrated hybrid structures (stable for more than 1 year) consisting of exfoliated expanded graphite (EG flakes, both naked and functionalized with branched polyethylenimine (PEI)) and Fe3O4 nanocrystals were fabricated by an *ex situ* process by the integration of iron oxide nanoparticles, coated with meso-2,3-dimercaptosuccinic acid (DMSA) or poly(acrylic acid) (PAA), onto the exfoliated EG flakes under the support of *N*-hydroxy‐ succinnimide (NHS) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) [68]. Such materials can have antibacterial properties. For example, *in situ* growth of silver nanoparticles onto the polyethylenimine (PEI)-mediated magnetic reduced GO resulted in a bactericidal material, Ag@rGO-Fe3O4-PEI composite (**Figure 23**) [69]. The material provides a very high killing rate of 99.9% for *E. coli* bacteria under a 0.5 min near-infrared (NIR) laser irradiation.

**Figure 23.** Schematic illustration of the preparation process of magnetic Ag@rGO-Fe3O4-PEI composites. Reproduced with permission from the *IOP Publishing*.

## **4.4. Applications of Fe3O4/G hybrids**

Solvothermally prepared G/Fe3O4 (**Figure 24**) is able to effectively remove both bacteriophage and bacteria in water [70]. Indeed, it is capable of eliminating a wide range of pathogens including not only bacteriophages, but also various bacteria such as *Shigella*, *Salmonella*, *E. Faecium*, *E. coli*, *E. faecalis*, and *S. aureus*, with removal efficiencies up to 94%. Graphene oxide, particularly as magnetic Fe3O4/GO particles, was used as an adsorbent for wastewater treatment (**Figure 25**) [71]. A variety of other Fe3O4/G (or GO) hybrid applications are shown in **Table 1**.

**Figure 24.** Synthesis route of G/Fe3O4. Reproduced with permission from the *American Chemical Society*.

Magnetic-Graphene-Based Nanocomposites and Respective Applications http://dx.doi.org/10.5772/64319 23

materials can have antibacterial properties. For example, *in situ* growth of silver nanoparticles onto the polyethylenimine (PEI)-mediated magnetic reduced GO resulted in a bactericidal material, Ag@rGO-Fe3O4-PEI composite (**Figure 23**) [69]. The material provides a very high killing rate of 99.9% for *E. coli* bacteria under a 0.5 min near-infrared (NIR) laser irradiation.

**Figure 23.** Schematic illustration of the preparation process of magnetic Ag@rGO-Fe3O4-PEI composites. Reproduced

Solvothermally prepared G/Fe3O4 (**Figure 24**) is able to effectively remove both bacteriophage and bacteria in water [70]. Indeed, it is capable of eliminating a wide range of pathogens including not only bacteriophages, but also various bacteria such as *Shigella*, *Salmonella*, *E. Faecium*, *E. coli*, *E. faecalis*, and *S. aureus*, with removal efficiencies up to 94%. Graphene oxide, particularly as magnetic Fe3O4/GO particles, was used as an adsorbent for wastewater treatment (**Figure 25**) [71]. A variety of other Fe3O4/G (or GO) hybrid applications are shown

**Figure 24.** Synthesis route of G/Fe3O4. Reproduced with permission from the *American Chemical Society*.

with permission from the *IOP Publishing*.

22 Advances in Carbon Nanostructures

in **Table 1**.

**4.4. Applications of Fe3O4/G hybrids**

**Figure 25.** Illustrative scheme of the dyeing process and the proposed treatment of its effluents; RB5 dye (top) and functionalized GO adsorbent (bottom). Reproduced with permission from the *John Wiley & Sons*.



**Table 1.** Main applications of Fe3O4/G composites.

## **5. Ferrite-graphene and related composites**

Simple or complex graphene or GO composites are known for a few ferrites, mainly those of cobalt and nickel, and usually used for environmental purposes or in batteries. NiFe2O4/G nanocomposites were prepared by mixing graphene and NiFe2O4 nanoparticles (obtained via a polyacrylamide gel method) into ethanol and further thermal drying at 60°C [82]. This composite exhibited significantly enhanced photocatalytic activity for the degradation of methylene blue in the conditions of irradiation of simulated sunlight, while NiFe2O4 nanopar‐ ticles were inert. NOH radicals were suggested as being the most active species causing dye degradation. NiFe2O4/GO nanohybrids based on NiFe2O4 and reduced GO consisted in nanosized NiFe2O4 crystals (5–10 nm in size) confined by few-layered rGO sheets [83]. This material seems a promising anode constituent for Li-ion batteries. The ternary nitrogen-doped graphene/nickel ferrite/polyaniline nanocomposite was hydrothermally prepared by a twostep approach process using urea, GO, Ni(NO3)2⋅6H2O, Fe(NO3)3⋅9H2O, and PANI as precur‐ sors, and applied in electrodes for supercapacitors [84]. It was confirmed that the introduction of N-heteroatom greatly improved the electrode specific capacitance.

**Composition Description Reference**

naphthylamine with different polarity. MWCNTs/FeO Fe2O3 were also produced in these synthesis as a byproduct, but graphene composite

[77]

[78]

[79]

[81]

nanomaterials. The adsorption capacity was shown to be naphthalene <

Removal of trace levels of six types of polybrominated diphenyl ethers in water treatment. The nanomaterials can be reused at least 20 times

Selective capture and convenient magnetic separation of target phosphopeptides, showing superior activity in comparison with

High-capacity lithium-ion battery anodes. The composites were prepared by electrostatic self-assembly between positively charged Fe3O4-APTMS [Fe3O4 grafted with (3-aminopropyl)trimethoxysilane (APTMS)] and negatively charged GO sheets and in an acidic solution (pH = 2) followed by *in situ* chemical reduction. An excellent rate

capability as well as much enhanced cycling stability.

Simple or complex graphene or GO composites are known for a few ferrites, mainly those of cobalt and nickel, and usually used for environmental purposes or in batteries. NiFe2O4/G nanocomposites were prepared by mixing graphene and NiFe2O4 nanoparticles (obtained via a polyacrylamide gel method) into ethanol and further thermal drying at 60°C [82]. This composite exhibited significantly enhanced photocatalytic activity for the degradation of methylene blue in the conditions of irradiation of simulated sunlight, while NiFe2O4 nanopar‐ ticles were inert. NOH radicals were suggested as being the most active species causing dye degradation. NiFe2O4/GO nanohybrids based on NiFe2O4 and reduced GO consisted in nanosized NiFe2O4 crystals (5–10 nm in size) confined by few-layered rGO sheets [83]. This material seems a promising anode constituent for Li-ion batteries. The ternary nitrogen-doped graphene/nickel ferrite/polyaniline nanocomposite was hydrothermally prepared by a twostep approach process using urea, GO, Ni(NO3)2⋅6H2O, Fe(NO3)3⋅9H2O, and PANI as precur‐

Use in lithium battery anodes. [80]

has the highest adsorption capacity among carbon-based

GO/Fe3O4 Super adsorbent to remove naphthalene, 1-naphthol, and 1-

Magnetic graphene composite absorbent (Fe3O4@PDDA/ GO@DNA) based on poly cationic

24 Advances in Carbon Nanostructures

A porous graphene composite affinity material, containing graphene scaffolds, Fe3O4 nanoparticles, and fully covered porous titania nanostructures

3D graphene foam-supported

The Fe3O4/G (graphene sheets) composite disks of μm dimensions

**Table 1.** Main applications of Fe3O4/G composites.

**5. Ferrite-graphene and related composites**

Fe3O4

core-shell Fe3O4@PDDA (poly(diallyldimethyl ammonium chloride) (PDDA)) and GO@DNA 1-naphthol < 1-naphthylamine.

commercial TiO2 affinity materials.

for remediation purposes.

In the case of cobalt, its graphene composites have been synthesized mainly by hydro/ solvothermally although an ultrasonic method was also reported [85]. A CoFe2O4/G nano‐ composite was hydrothermally prepared and was evaluated in the photocatalytic degradation of methylene blue dye under visible irradiation [86]. Its photocatalytic activity depends on the surface area and the structural and optical properties of samples, and behaves better than pure cobalt ferrite. The formation mechanism of hydrothermally synthesized CoFe2O4-FGS (FGS: functionalized graphene sheets) is described in detail by Li et al. [87]. In addition, porous CoFe2O4/rGO nanoclusters with different concentrations of graphene were solvothermally prepared (**Figure 26**) [88] and their electrochemical properties were evaluated using polyvi‐ nylidene fluoride and Na-alginate as binder materials to improve the anode performance of Li-ion batteries. The resulting CoFe2O4/rGO (20%) nanocomposites with sodium alginate binders exhibited promising electrochemical performance, such as excellent recycling behav‐ ior, good rate capability, and high reversible capacity.

**Figure 26.** Synthesis of the CoFe2O4/rGO composite. Reproduced with permission from the *Royal Society of Chemistry*.

Other graphene- (or GO)-ferrite hybrids are rare (for example, zinc-ferrite-reduced graphene oxide [89]), and their synthesis, properties, and applications are similar to those of nickel and cobalt. Thus, in a solvothermally fabricated uniform copper ferrite nanoparticle-attached graphene nanosheet (CuFe2O4/G), CuFe2O4 nanoparticles (with a diameter of about 100 nm) were shown to be densely and compactly deposited on graphene nanosheets [90]. This nanomaterial exhibited properties of a high-performance supercapacitor material. A related hydrothermal route was used for the formation of CuFe2O4 nanoparticles into the GO sheets, and the obtained material, with GO content varying from 10 to 30 wt%, was used as stable and good reproducible enzyme-free glucose sensor [91]. MnFe2O4 nanoparticle (MFNP)-decorated GO nanocomposites (MGONCs) were prepared through a miniemulsion and solvent evapo‐ ration process [92]. The loading of magnetic nanocrystals depends on the ratio of GO/magnetic nanoparticles. In a related report [93], MnFe2O4 nanoparticles were deposited on GO by thermal decomposition of manganese(II) acetylacetonate and iron(III) acetylacetonate precur‐ sors in triethylene glycol. The resulting GO/MnFe2O4 nanohybrids showed imperceptible *in vivo* toxicity, negligible hemolytic activity, and very low cytotoxicity, being effective T2 contrast agents.

In addition, barium hexaferrite nanoparticles, fabricated [94] via the citrate sol-gel combustion method, in a reaction medium consisting of various forms of graphene nanosheets (such as reduced graphite oxide, expanded graphite oxide, and expanded graphite) were used to prepare nanocomposites as microwave absorbing material for radars [95]. Finally, BiFeO3 nanoparticles that are important materials in photocatalysis [96] were prepared via a polya‐ crylamide gel route and transformed into a BiFeO3/G nanocomposite, by mixing with graphene in ethanol and further thermal drying at 60°C [97]. Methyl orange (MO) was found to be efficiently degraded by this material under simulated sunlight irradiation. This photocatalytic performance can be mainly ascribed to the efficient transfer of photogenerated electrons from BiFeO3 to graphene (**Figure 27**). A similar reduced graphene rGO/BiFeO3 nanocomposite, prepared by self-assembly, was also reported [98].

**Figure 27.** Schematic illustration of the photocatalytic mechanism of BiFeO3-graphene nanocomposite toward the deg‐ radation of MO. Reproduced with permission from the *Hindawi Publishing Corporation.*

An interesting example of a magnetic molecule and its composite with graphene is known for lanthanides. A device made of a graphene nanoconstruction, decorated with TbPc2 magnetic

molecules (Pc: phthalocyananine), was able to electrically detect the magnetization reversal of the molecules in proximity with graphene (**Figure 28**) [99, 100].

**Figure 28.** (a) Schematic representation of the TbPc2 single-molecule magnet (SMM). (b, c) Schematic view of the de‐ vice, showing in (b) the molecule attached to graphene and in (c) the nanoconstriction contacted by source (S) and drain (D) electrodes. The magnetic moments of the TbPc2 SMMs (hexyl and pyrenyl groups here omitted for clarity) on top of the constriction add another degree of freedom to tune the conductivity of the device. (d) False-colored SEM image of the device presented in the text. SiO2 substrate and etched graphene are colored in purple, while graphene conductive regions are colored in green. Source and drain electrodes are indicated. Reproduced with permission from the *American Chemical Society*.

## **6. Conclusions and outlook**

were shown to be densely and compactly deposited on graphene nanosheets [90]. This nanomaterial exhibited properties of a high-performance supercapacitor material. A related hydrothermal route was used for the formation of CuFe2O4 nanoparticles into the GO sheets, and the obtained material, with GO content varying from 10 to 30 wt%, was used as stable and good reproducible enzyme-free glucose sensor [91]. MnFe2O4 nanoparticle (MFNP)-decorated GO nanocomposites (MGONCs) were prepared through a miniemulsion and solvent evapo‐ ration process [92]. The loading of magnetic nanocrystals depends on the ratio of GO/magnetic nanoparticles. In a related report [93], MnFe2O4 nanoparticles were deposited on GO by thermal decomposition of manganese(II) acetylacetonate and iron(III) acetylacetonate precur‐ sors in triethylene glycol. The resulting GO/MnFe2O4 nanohybrids showed imperceptible *in vivo* toxicity, negligible hemolytic activity, and very low cytotoxicity, being effective T2

In addition, barium hexaferrite nanoparticles, fabricated [94] via the citrate sol-gel combustion method, in a reaction medium consisting of various forms of graphene nanosheets (such as reduced graphite oxide, expanded graphite oxide, and expanded graphite) were used to prepare nanocomposites as microwave absorbing material for radars [95]. Finally, BiFeO3 nanoparticles that are important materials in photocatalysis [96] were prepared via a polya‐ crylamide gel route and transformed into a BiFeO3/G nanocomposite, by mixing with graphene in ethanol and further thermal drying at 60°C [97]. Methyl orange (MO) was found to be efficiently degraded by this material under simulated sunlight irradiation. This photocatalytic performance can be mainly ascribed to the efficient transfer of photogenerated electrons from BiFeO3 to graphene (**Figure 27**). A similar reduced graphene rGO/BiFeO3 nanocomposite,

**Figure 27.** Schematic illustration of the photocatalytic mechanism of BiFeO3-graphene nanocomposite toward the deg‐

An interesting example of a magnetic molecule and its composite with graphene is known for lanthanides. A device made of a graphene nanoconstruction, decorated with TbPc2 magnetic

radation of MO. Reproduced with permission from the *Hindawi Publishing Corporation.*

contrast agents.

26 Advances in Carbon Nanostructures

prepared by self-assembly, was also reported [98].

The field of graphene and graphene oxide composites containing magnetic components can be considered as an important subject of the modern nanotechnology. Similarly to the filled or decorated carbon nanotubes, for which endohedral and exohedral functionalization with magnetic nanoparticles is well known, the graphene and graphene oxide magnetic nanocom‐ posites can also be obtained via distinct methods and have a variety of applications. Graphene magnetic nanocomposites are currently known for those on the basis of elemental metals (Fe, Co, and Ni), magnetic nanoclusters, iron oxides (Fe2O3, Fe3O4) in distinct morphological forms, ferrites MFe2O4, 3D graphene aerogels@hierarchical Fe3O4 nanoclusters, single-molecule magnets such as TbPc2 (Pc: phthalocyanine), other organometallic-containing composites (benzene-metal-graphene), as well as polycomponent nanocomposites such as Ag/Fe3O4/G (G: graphene), Fe3O4/CdS/G, or FePc/Fe3O4/GO (GO :graphene oxide), etc.

Magnetic graphene and graphene oxide hybrids can be prepared by a variety of methods. Hydrothermal and solvothermal techniques are the most common. Other routes are also available, such as sol-gel autocombustion, sonoelectrochemical polymerization, thermal expansion and thermal reduction, microwave-assisted technique, as well as covalent bonding chemical methods. The formed composites have or might have several current and/or potential applications, such as devices (supercapacitors or material for anodes of lithium batteries), construction (graphene oxide-ferrofluid-cement nanocomposites), analytical, sensor, and biosensor applications (graphene-based magnetic solid-phase extraction for the determination of five chloroacetanilide herbicides (alachlor, acetochlor, metolachlor, butachlor, and pretila‐ chlor), nitrite sensor, composites with peroxidase-like activity for colorimetric detection of glucose, etc.), environmental remediation (degradation and removal of water pollutants, such as chlorophenols, adsorbent for pesticide extraction, As and heavy metals (UVI)), compounds with antibacterial properties, and, of course, as magnetic nanocatalysts or nanophotocatalysts which can be simply magnetically removed from reaction media [101, 102].

As described above, magnetic graphene and graphene oxide composites can have applications, in particular, in the removal, decoloration, sequestration and adsorption of dyes [103], antibiotics and other organic contaminants, heavy metal separation (Hg, Cd, Cr, Cu, and As), and other environmental remediation aspects [104]. Other areas include biological imaging and further bioapplications [105], magnetic separation, electromagnetic materials and coat‐ ings, conducting polymer nanocomposites, aligned substrate for nanodevices, and spintronic devices with perpendicular magnetic anisotropy. 3D graphene aerogels can have applications in electrode materials, supercapacitors, oil absorption, and gas sensors due to their electrical conductivity, mechanical strength, and high porosity.

## **Author details**

Oxana Vasilievna Kharissova1 , Beatriz Ortega García1 , Boris Ildusovich Kharisov2\* and Ubaldo Ortiz Méndez3

\*Address all correspondence to: bkhariss@hotmail.com

1 Department of Physico-Mathematics, Autonomous University of Nuevo León, Monterrey, Mexico

2 Department of Chemistry, Autonomous University of Nuevo León, Monterrey, Mexico

3 FIME (Department of Mechanical & Electrical Engineering), Autonomous University of Nuevo León, Monterrey, Mexico

## **References**

(benzene-metal-graphene), as well as polycomponent nanocomposites such as Ag/Fe3O4/G (G:

Magnetic graphene and graphene oxide hybrids can be prepared by a variety of methods. Hydrothermal and solvothermal techniques are the most common. Other routes are also available, such as sol-gel autocombustion, sonoelectrochemical polymerization, thermal expansion and thermal reduction, microwave-assisted technique, as well as covalent bonding chemical methods. The formed composites have or might have several current and/or potential applications, such as devices (supercapacitors or material for anodes of lithium batteries), construction (graphene oxide-ferrofluid-cement nanocomposites), analytical, sensor, and biosensor applications (graphene-based magnetic solid-phase extraction for the determination of five chloroacetanilide herbicides (alachlor, acetochlor, metolachlor, butachlor, and pretila‐ chlor), nitrite sensor, composites with peroxidase-like activity for colorimetric detection of glucose, etc.), environmental remediation (degradation and removal of water pollutants, such as chlorophenols, adsorbent for pesticide extraction, As and heavy metals (UVI)), compounds with antibacterial properties, and, of course, as magnetic nanocatalysts or nanophotocatalysts

As described above, magnetic graphene and graphene oxide composites can have applications, in particular, in the removal, decoloration, sequestration and adsorption of dyes [103], antibiotics and other organic contaminants, heavy metal separation (Hg, Cd, Cr, Cu, and As), and other environmental remediation aspects [104]. Other areas include biological imaging and further bioapplications [105], magnetic separation, electromagnetic materials and coat‐ ings, conducting polymer nanocomposites, aligned substrate for nanodevices, and spintronic devices with perpendicular magnetic anisotropy. 3D graphene aerogels can have applications in electrode materials, supercapacitors, oil absorption, and gas sensors due to their electrical

, Beatriz Ortega García1

1 Department of Physico-Mathematics, Autonomous University of Nuevo León, Monterrey,

2 Department of Chemistry, Autonomous University of Nuevo León, Monterrey, Mexico

3 FIME (Department of Mechanical & Electrical Engineering), Autonomous University of

, Boris Ildusovich Kharisov2\* and

graphene), Fe3O4/CdS/G, or FePc/Fe3O4/GO (GO :graphene oxide), etc.

which can be simply magnetically removed from reaction media [101, 102].

conductivity, mechanical strength, and high porosity.

\*Address all correspondence to: bkhariss@hotmail.com

**Author details**

Mexico

Ubaldo Ortiz Méndez3

28 Advances in Carbon Nanostructures

Oxana Vasilievna Kharissova1

Nuevo León, Monterrey, Mexico


[26] Mural, P.K.S.; Pawar, S.P.; Jayanthi, S.; Madras, G.; Sood, A.K.; Bose, S. Engineering nanostructures by decorating magnetic nanoparticles onto graphene oxide sheets to shield electromagnetic radiations. ACS Appl. Mater. Interf., 2015, 7, 16266−16278.

[14] Decker, R.; Brede, J.; Atodiresei, N.; Caciuc, V.; Blugel, S.; Wiesendanger, R. Atomicscale magnetism of cobalt-intercalated graphene. Phys. Rev. B, 2013, 87, 041403(R).

[15] Rougemaille, N.; N'Diaye, A.T.; Coraux, J.; Vo-Van, C.; Fruchart, O.; Schmid, A.K. Perpendicular magnetic anisotropy of cobalt films intercalated under graphene. Appl.

[16] Vermisoglou, E.C.; Devlin, E.; Giannakopoulou, T.; Romanos, G.; Boukos, N.; Psycha‐ ris, V.; Lei, C.; Lekakou, C.; Petridis, D.; Trapalis, C. Reduced graphene oxide/iron carbide nanocomposites for magnetic and supercapacitor applications. J. Alloys

[17] Afshar, M.; Doosti, H. Magnetic properties of cobalt single layer added on graphene: a density functional theory study. Mod. Phys. Lett. B, 2015, 29(2),

[18] Zhidkov, I.S.; Skorikov, N.A.; Korolev, A.V.; Kukharenko, A.I.; Kurmaev, E.Z.; Fedorov, V.E.; Cholakh, S.O. Electronic structure and magnetic properties of graphene/

[19] Lin, K.-Y.A.; Hsu, F.-K.; Lee, W.-D. Magnetic cobalt–graphene nanocomposite derived from self-assembly of MOFs with graphene oxide as an activator for peroxymonosul‐

[20] Zhang, L.; Huang, Y.; Zhang, Y.; Ma, Y.; Chen, Y. Sol–gel autocombustion synthesis of graphene/cobalt magnetic nanocomposites. J. Nanosci. Nanotechnol., 2013, 13, 1129–

[21] Yang, S.; Cui, G.; Pang, S.; Cao, Q.; Kolb, U.; Feng, X.; Maier, J.; Mullen, K. Fabrication of cobalt and cobalt oxide/graphene composites: towards high-performance anode

[22] Yao, Y.; Xu, C., Qin, J., Wei, F.; Rao, M.; Wang, S. Synthesis of magnetic cobalt nano‐ particles anchored on graphene nanosheets and catalytic decomposition of Orange II.

[23] Umair, A.; Raza, H. Controlled synthesis of bilayer graphene on nickel. Nanoscale Res.

[24] Dlubak, B.; Martin, M.-B.; Weatherup, R.S., Yang, H.; Deranlot, C.; Blume, R.; Schloegl, R.; Fert, A.; Anane, A.; Hofmann, S.; Seneor, P.; Robertson, J. Graphene-passivated nickel as an oxidation-resistant electrode for spintronics. ACS Nano, 2012, 6(12), 10930–

[25] Sicot, M.; Bouvron, S.; Zander, O.; Rudiger, U.; Dedkov, Vu.S.; Fonin, M. Nucleation and growth of nickel nanoclusters on graphene Moire on Rh(111). Appl. Phys. Lett.,

materials for lithium ion batteries. ChemSusChem, 2010, 3, 236–239.

Phys. Lett., 2012, 101, 142403.

Compounds, 2014, 590, 102–109.

Co composites. Carbon, 2015, 91, 298–303.

fate. J. Mater. Chem. A, 2015, 3, 9480–9490.

Ind. Eng. Chem. Res. 2013, 52, 17341−17350.

1450262, 9 pp.

30 Advances in Carbon Nanostructures

1131.

Lett., 2012, 7, 437.

2010, 96, 093115.

10934.


[51] Xu, X.; Li, H.; Zhang, Q.; Hu, H.; Zhao, Z.; Li, J.; Li, J.; Qiao, Yu.; Gogotsi, Yu. Selfsensing, ultralight, and conductive 3D graphene/iron oxide aerogel elastomer deform‐ able in a magnetic field. ACS Nano, 2015, 9(4), 3969–3977.

[39] Zhang, X.; Alloul, O.; He, Q.; Zhu, J.; Joseph Verde, M.; Li, Y.; Wei, S.; Guo, Z. Strength‐ ened magnetic epoxy nanocomposites with protruding nanoparticles on the graphene

[40] Wang, Y.; He, Q.; Qu, H.; Zhang, X.; Guo, J.; Zhu, J.; Zhao, G.; Colorado, H.A.; Yu, J. et al. Magnetic graphene oxide nanocomposites: nanoparticles growth mechanism and

[41] McCoy, T.M.; Brown, P.; Eastoe, J.; Tabor, R.F. Noncovalent magnetic control and reversible recovery of graphene oxide using iron oxide and magnetic surfactants. ACS

[42] Sun, Z.; Madej, E.; Wiktor, C. et al. One-pot synthesis of carbon-coated nanostructured iron oxide on few-layer graphene for lithium-ion batteries. Chem. Eur. J., 2015, 21,

[43] Jang, B.; B. Chae, O.; Park, S.-K.; Ha, J.; Oh, S.-M.; Na, H.B.; Piao, Y. Solventless synthesis of an iron-oxide/graphene nanocomposite and its application as an anode in high-rate

[44] Tae Kim, I.; Magasinski, A.; Jacob, K.; Yushin, G.; Tannenbaum, R. Synthesis and electrochemical performance of reduced graphene oxide/maghemite composite anode

[45] Zhu, X.; Zhu, Y.; Murali, S.; Stoller, M.D.; Ruoff, R.S. Nanostructured reduced graphene oxide/Fe2O3 composite as a high-performance anode material for lithium ion batteries.

[46] Eeu, Y.C.; Lim, H.N.; Lim, Y.S.; Zakarya, S.A.; Huang, N.M. Electrodeposition of polypyrrole/reduced graphene oxide/iron oxide nanocomposite as supercapacitor

[47] Sinh, A.; Jan, N.R. Separation of microcystin-LR by cyclodextrin-functionalized magnetic composite of colloidal graphene and porous silica. ACS Appl. Mater. Interf.,

[48] Jagadeesh, R.V.; Natte, K.; Junge, H.; Beller, M. Nitrogen-doped graphene-activated iron-oxide-based nanocatalysts for selective transfer hydrogenation of nitroarenes.

[49] Jung, S.M.; Jung, H.Y.; Dresselhaus, M.S.; Jung, J.J.; Kong, J. A facile route for 3D aerogels from nanostructured 1D and 2D materials. Sci. Rep., 2012, 2, 849.

[50] Wu, Z.-S.; Yang, S.; Sun, Y.; Parvez, K.; Feng, X.; Müllen, K. 3D nitrogen-doped graphene aerogel-supported Fe3O4 nanoparticles as efficient electrocatalysts for the

oxygen reduction reaction. J. Am. Chem. Soc., 2012, 134, 9082−9085.

nanosheets. Polymer, 2013, 54, 3594–3604.

Appl. Mater. Interf., 2015, 7, 2124−2133.

16154–16161.

32 Advances in Carbon Nanostructures

property analysis. J. Mater. Chem. C, 2014, 2, 9478–9488.

Li-ion batteries. J. Mater. Chem. A, 2013, 1, 15442–15446.

electrode material. J. Nanomater., 2013, 2013, 653890, 6 pp.

for lithium ion batteries. Carbon, 2013, 52, 56–64.

ACS Nano, 2011, 5(4), 3333–3338.

ACS Catal., 2015, 5, 1526−1529.

2015, 7, 9911−9919.


freestanding membranes and their temperature dependent electronic transport properties. Carbon, 2012, 50, 1338–1345.

[77] Yang, X.; Li, J.; Wen, T.; Ren, X., Huang, Y.; Wang, X. Adsorption of naphthalene and its derivatives on magnetic graphene composites and the mechanism investigation. Colloids Surf. A: Physicochem. Eng. Aspects, 2013, 422, 118–125.

[63] Yadav, M.; Rhee, K.Y.; Park, J.S.; Hui, D. Mechanical properties of Fe3O4/GO/chitosan

[64] Zhang, W.; Li, X.; Zou, R.; Wu, H.; Shi, H.; Yu, S.; Liu, Y. Multifunctional glucose biosensors from Fe3O4 nanoparticles modified chitosan/graphene nanocomposites. Sci.

[65] Ou, Y.; Wang, F.; Huang, Y.; Li, D.; Jiang, Y.; Qin, Q.-H.; Stachurski, Z.H.; Tricoli, A.; Zhang, T. Fabrication and cyto-compatibility of Fe3O4/SiO2/graphene–CdTeQDs/CS nanocomposites for drug delivery. Colloids Surf. B: Biointerf., 2014, 117, 466–472. [66] Fan, L.; Luo, C.; Li, X.; Lu, F.; Qiu, H.; Sun, M. Fabrication of novel magnetic chitosan grafted with graphene oxide to enhance adsorption properties for methyl blue. J.

[67] Fan, L.; Luo, C.; Sun, M.; Li, X.; Lu, F.; Qiu, H. Preparation of novel magnetic chitosan/ graphene oxide composite as effective adsorbents toward methylene blue. Bioresource

[68] Jovici, N.; Calatayud, M.P.; Sanz, B.; Montone, A.; Goya, G.F. Ex situ integration of iron oxide nanoparticles onto exfoliated expanded graphite flakes in aqueous suspension.

[69] Wang, N.; Hu, B.; Chen, M.-L.; Wang, J.-H. Polyethylenimine mediated silver nano‐ particle-decorated magnetic graphene as a promising photothermal antibacterial agent.

[70] Zhan, S.; Zhu, D.; Ma, S.; Yu, W.; Jia, Y.; Li, Y.; Yu, H.; Shen, Z. Highly efficient removal of pathogenic bacteria with magnetic graphene composite. ACS Appl. Mater. Interf.,

[71] Kyzas, G.Z.; Deliyanni, E.A.; Matis, K.A. Graphene oxide and its application as an adsorbent for wastewater treatment. J. Chem. Technol. Biotechnol., 2014, 89, 196–205.

[72] Ye, Y.; Yin, D.; Wang, B.; Zhang, Q. Synthesis of three-dimensional Fe3O4/graphene aerogels for the removal of arsenic ions from water. J. Nanomater., 2015, 864864, 6 pp.

[73] Mishra, A.K.; Ramaprabhu, S. Ultrahigh arsenic sorption using iron oxide-graphene nanocomposite supercapacitor assembly. J. Appl. Phys., 2012, 112, 104315.

[74] Hu, X.-j.; Liu, Y.-g.; Wang, H.; Zeng, G.-m.; Hua, X.; Guo, Y.-m.; Li, T.-t.; Chen, A.-w.; Jiang, L.-h.; Guo, F.-y. Adsorption of copper by magnetic grapheneoxide-supported cyclodextrin: Effects of pH, ionicstrength, background electrolytes, and citric acid.

[75] Wang, Y.; Liang, S.; Chen, B.; Guo, F.; Yu, S.; Tang, Y. Synergistic removal of Pb(II), Cd(II) and humic acid by Fe3O4@mesoporous silica-graphene oxide composites. PLOS

[76] Narayanan, T.N.; Liu, Z.; Lakshmy, P.R.; Gao, W.; Nagaoka, Y.; Sakthi Kumar, D.; Lou, J.; Vajtai, R.; Ajayan, P.M. Synthesis of reduced graphene oxide–Fe3O4 multifunctional

composites. Composites: Part B, 2014, 66, 89–96.

Hazardous Mater., 2012, 215–216, 272–279.

J. Serb. Chem. Soc., 2014, 79(9), 1155–1167.

Nanotechnology, 2015, 26, 195703, 8 pp.

Chem. Chem. Eng. Res. Design, 2015, 93, 675–683.

Technol., 2012, 114, 703–706.

2015, 7, 4290−4298.

ONE, 2013, 8(6), e65634.

Rep., 2015, 5, 11129.

34 Advances in Carbon Nanostructures


composites with alginate binders for Li-ion battery applications. New J. Chem., 2014, 38, 3654–3661.


[102] Lim, C.W.; Lee, I.S. Magnetically recyclable nanocatalyst systems for the organic reactions. Nano Today, 2010, 5, 412–434.

composites with alginate binders for Li-ion battery applications. New J. Chem., 2014,

[89] Fei, P.; Zhong, M.; Lei, Z.; Su, B. One-pot solvothermal synthesized enhanced magnetic zinc ferrite–reduced graphene oxide composite material as adsorbent for methylene

[90] Zhang, W.; Quan, B.; Lee, C.; et al. One-step facile solvothermal synthesis of copper ferrite−graphene composite as a high-performance supercapacitor material. ACS Appl.

[91] Shahnavaz, Z.; Woi, P.W.; Aliasn, Y. A hydrothermally prepared reduced graphene oxide-supported copper ferrite hybrid for glucose sensing. Ceram. Int., 2015, 41, 12710–

[92] Peng, E.; Shi Guang Choo, E.; Chandrasekharan, P.; et al. Synthesis of manganese ferrite/graphene oxide nanocomposites for biomedical applications. Small, 2012, 8(23),

[93] Yang, Y.; Shi, H.; Wang, Y.; et al. Graphene oxide/manganese ferrite nanohybrids for magnetic resonance imaging, photothermal therapy and drug delivery. J. Biomater.

[94] Panwar, R.; Puthucheri, S.; Singh, D.; Agarwala, V. Design of ferrite–graphene-based thin broadband radar wave absorber for stealth application. IEEE Trans. Magnet., 2015,

[95] Durmus, Z.; Durmus, A.; Kavas, H. Synthesis and characterization of structural and magnetic properties of graphene/hard ferrite nanocomposites as microwave-absorbing

[96] Gao, T.; Chen, Z.; Huang, Q.; Niu, F.; Huang, X.; Qin, L.; Huang, Y. A review: prepa‐ ration of bismuth ferrite nanoparticles and its applications in visible-light induced

[97] Dai, J.F.; Xian, T.; Di, L.J.; Yang, H. Preparation of BiFeO3-graphene nanocomposites and their enhanced photocatalytic activities. J. Nanomater., 2013, 642897, 5 pp.

[98] Li, T.; Shen, J.; Li, N.; Ye, M. Hydrothermal preparation, characterization and enhanced properties of reduced graphene-BiFeO3 nanocomposite. Mater. Lett., 2013, 91, 42–44.

[99] Candini, A.; Klyatskaya, S.; Ruben, M.; Wernsdorfer, w.; Affronte, M. Graphene spintronic devices with molecular nanomagnets. Nano Lett., 2011, 11, 2634–2639.

[100] Lopes, M.; Candini, A.; Urdampilleta, M.; et al. Surface-enhanced Raman signal for terbium single-molecule magnets grafted on graphene. ACS Nano, 2010, 4(12), 7531–

[101] Govan, J.; Gun'ko, Yu.K. Recent advances in the application of magnetic nanoparticles as a support for homogeneous catalysts. Nanomaterials, 2014, 4, 222–241.

Appl. 2016 Jan;30(6):810–22. http://www.ncbi.nlm.nih.gov/pubmed/26296777

38, 3654–3661.

36 Advances in Carbon Nanostructures

12716.

7537.

3620–3630.

51(1), 2802804, 4 pp.

material. J. Mater. Sci, 2015, 50, 1201–1213.

photocatalyses. Rev. Adv. Mater. Sci., 2015, 40, 97–109.

blue removal. Mater. Lett., 2013, 108, 72–74.

Mater. Interf., 2015, 7, 2404−2414.


## **Next-Generation Graphene-Based Membranes for Gas Separation and Water Purifications**

Quan Xu and Wenwen Zhang

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64396

#### **Abstract**

Advanced membrane systems are regarded as an important portion of controllable separation processes, such as gas separation and water purification. The ideal materials should have good permeability for selected particle sizes, high stiffness to withstand high pressures applied, large surface area and micro- or nanopore structures for excellent selectivity. Recently, graphene with oxygen-containing functional groups and graphene oxide (GO) nanosheets, obtained via chemical oxidation of graphite, achieved tremendous properties that include excellent mechanical strength, large relative surface area, unique honeycomb lattice two-dimensional structure as well as narrow pore distribution, offering platform to be used as advanced, ultrathin membrane for a wide variety of purification process with high efficiency. In this review chapter, the potential application of such advanced materials for gas separation and water purification process is discussed. The fabrication and modification process and innovation of such advanced two-dimensional functional structure for purification and separation process are introduced. This review chapter will offer opportunity to understand details involved in gas and/or water molecular transport through thin, laminar graphene oxide and derived structures, as well as up to now progress in the field.

**Keywords:** graphene, graphene oxide, membrane, purification, separation

## **1. Introduction**

Separation, or separation technique, is regarded as an important process that converts a mixture of substances into distinct pure mixtures/fractions. In industry, separation is an important indispensable process for nature gas separation, hydrogen production, seawater desalination, etc. [1–3]. Traditional separation technologies, including distillation [4, 5],

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

adsorption [1], absorption [6, 7], etc., have been developed for decades. With the restriction of total energy resources on Earth and the continuous global population growth, the demand for high efficacy technologies to solve urgent environmental and energy problems urgently needs the formulation of advanced separation techniques. In comparison to traditional technolo‐ gies, membrane separation technology has been developed and became a well-established separation technology over the past decennia, with the benefit of overall low energy usage, flexible arranged process conductions, safety, materials sustainability and being economical and environmental friendly [8, 9]. Membranes work as a selective separation wall which target substances that can pass through and block others [8, 10, 11]. Two important factors are used to evaluate the performance of membranes: flux and selectivity [12]. Flux is defined as the volumetric flow rate, and selectivity is the status of molecules being passed through. The traditional membranes generally made from polymers, zeolite and silicon have been success‐ fully applied in industry [13]. A summary of membrane porous sizes *versus* application classification is exhibited in **Figure 1**.

**Figure 1.** Summary of membrane classification according to their pore sizes.

Recently, carbon-based nanomaterials have attracted intensive attention from scientists, due to their abundant reserves, easy accessibility, remarkable mechanical properties, biocompati‐ bility, as well as environmental friendly properties [14, 15]. Initially, carbon nanotubes (CNTs) were believed to be excellent candidates for membrane materials, due to their uniquely hollow structure with one open-end, and extremely high mechanical performance [16, 17]. However, the fabrication of vertically aligned CNT arrays with high density, large scale quantity and low defects is still challenging [18]. Thus, the generation of CNT-based separation membrane, for large-scale practical applications, remains a theoretical possibility. Another kind of carbonbased nanomaterials with great mechanical properties is carbon diamond. They contain amorphous sp3 -hybridized carbon atoms and are also considered as candidate materials for separation technology. Considered the large energy consumption and high cost, it is not economical and practical to use carbon diamonds for large-scale industrial applications [19]. Therefore, the development of more applicable and economically suited carbon-based nanomaterials for fabricating separation membranes is an urgent need.

adsorption [1], absorption [6, 7], etc., have been developed for decades. With the restriction of total energy resources on Earth and the continuous global population growth, the demand for high efficacy technologies to solve urgent environmental and energy problems urgently needs the formulation of advanced separation techniques. In comparison to traditional technolo‐ gies, membrane separation technology has been developed and became a well-established separation technology over the past decennia, with the benefit of overall low energy usage, flexible arranged process conductions, safety, materials sustainability and being economical and environmental friendly [8, 9]. Membranes work as a selective separation wall which target substances that can pass through and block others [8, 10, 11]. Two important factors are used to evaluate the performance of membranes: flux and selectivity [12]. Flux is defined as the volumetric flow rate, and selectivity is the status of molecules being passed through. The traditional membranes generally made from polymers, zeolite and silicon have been success‐ fully applied in industry [13]. A summary of membrane porous sizes *versus* application

classification is exhibited in **Figure 1**.

40 Advances in Carbon Nanostructures

amorphous sp3

**Figure 1.** Summary of membrane classification according to their pore sizes.

Recently, carbon-based nanomaterials have attracted intensive attention from scientists, due to their abundant reserves, easy accessibility, remarkable mechanical properties, biocompati‐ bility, as well as environmental friendly properties [14, 15]. Initially, carbon nanotubes (CNTs) were believed to be excellent candidates for membrane materials, due to their uniquely hollow structure with one open-end, and extremely high mechanical performance [16, 17]. However, the fabrication of vertically aligned CNT arrays with high density, large scale quantity and low defects is still challenging [18]. Thus, the generation of CNT-based separation membrane, for large-scale practical applications, remains a theoretical possibility. Another kind of carbonbased nanomaterials with great mechanical properties is carbon diamond. They contain

separation technology. Considered the large energy consumption and high cost, it is not economical and practical to use carbon diamonds for large-scale industrial applications [19].


The two-dimensional carbon, graphene, possessing a single-atom-thick sheet of sp2 -hybridized carbon atoms arrayed in a honeycomb pattern, opened a new door for researchers and scholars to fabricate the next generation of membranes with outstanding separation capacities [20, 21]. Graphene and its derivative, graphene oxide (GO), with very thin, stiff and large surface area and micro-/nanopore structure layers, are considered as the ideal candidates for membrane materials, due to their high durability, withstanding high applied working pressures and excellent selectivity [22–25]. Graphene is stripped out of graphite materials made of only carbon atoms, with a kind of two-dimensional crystal structures having just one-atom-thick sheet. The concept, graphene-based materials, as well as the preparation, has been explored for over a century. In the early middle of the nineteenth century, Brodie et al. [26] prepared GO and distinguished graphite as "lamellar" and "amorphous". In 1918, Haenni et al. [27] described in detail the properties of GO paper. Then Ruess et al. published the first transmis‐ sion electron microscopy (TEM) images of a few layers of graphene (between 3 and 10 layers) imaged in 1948. Hummers [28] provided a redox method to prepare GO with outstanding properties. Tsotsis et al. [29] reported that graphene membranes can be used for gas mixture separation. In 2004, Andre Geim, a physicist from the University of Manchester, and his copartner Konstantin Novoselov, successfully isolated single-layer graphene from a sheet of graphite, providing evidence that two-dimensional materials can exist alone in nature, and they shared the 2010 Nobel Prize of Physics for this work [30]. Nowadays, graphene and derived materials have attracted worldwide attention, and the global market for graphene is reported to have reached \$9 million in 2014 [22, 31, 32].

The first experimental use of graphene as membrane materials was performed by Bunch et al. [33], who demonstrated that graphene is impermeable to all standard gas atoms including helium. With high crystal quality of carbon atoms, the electron density of graphene aromatic rings can repel most atoms and molecules [34]. Scientists further suggest that nanoporous graphene (NPG), with monatomic layers, high mechanical strength and strong chemical stability, may have a great potential for nature-/biogas separation and seawater distillation [23, 24, 35, 36]. Moreover, its derived material, graphene oxide (GO), was identified as potential filtration/membrane, due to the controllable pore distribution, applicability under highpressure conditions, simple preparation procedure and the possibility for easy large-scale production. Many attempts were conducted for gas mixture separation and water or liquid molecules purification [36, 37].

In this chapter, the recent applications and research progresses of graphene and GO on separation membrane for gas separation and water purification are discussed. The chapter is divided into three sections. In the first section, the applications of graphene and GO on gas separation will be introduced. The second deals with water filtration. Finally, a summary and outlook of graphene-based separation membranes will be stated, including the recent chal‐ lenges and the expected further developments in the future.

### **2. Graphene and GO for gaseous selectivity separation**

Comparing with traditional membrane materials, graphene and GO membranes show remarkable properties, such as unique stiffness, hydrophobic characteristics and are easy scale up in various applications [38–41]. The monolayer graphene is known to resist to very small gaseous permeation, including He (helium) [33], as the electron-clouds on the graphene hexagonal rings could hamper the atoms and molecules to allow the transit of small molecules. Graphene is long being considered as an ideal gas barrier material. Although graphene prevents gas to pass through, nanoporous graphene (NPG), due to its high surface porosity, can efficiently allow gas flow to selectively pass through, thus becoming a potential membrane material candidate. The analysis of NPG as membrane materials starts with the simulation of this amazing material. In 2008, Sint et al. [42] designed functionalized nanopores in graphene monolayers and showed that they could provide highly selective passage of hydrated ions, using molecular dynamics (MD) simulations. In the same year, Jiang et al. [43] studied NPG and investigated the permeability and selectiv‐ ity of NPG and demonstrated the remarkable potential of graphene as a membrane for the gas separation of H2 and CH4. Their results indicated that H2 permeance could be up to 1 mol/ m2 s1 Pa1 , with a selectivity of amazing 108 for H2/CH4, which was up to 107 improvement in the value obtained with silica membranes (**Figure 2**). Du et al. [44] designed a series of NPG for separating N2 and H2 and found that there were different mechanisms for H2 and N2 to

**Figure 2.** (a) Creating a nitrogen-doped pore within one graphene sheet; (b) graphene pores with mixed nitrogen (green) and hydrogen (blue) terminations and (c) only hydrogen-terminated pores. (d) Equivalent pore size from elec‐ tron density calculations for nitrogen- and hydrogen-terminated pores; (e) pure hydrogen-terminated pores. Adapted with permission from [43].

permeate through the NPG membrane. The hydrogen flux was linear in relation to the pore size of NPG, while nitrogen flux was not. This revealed that the mechanisms of hydrogen and nitrogen permeation through the porous graphene are different. Trinh et al. [45] calculated bonding energy between gas mixtures including molecules of CO2 and H2 and graphite by performing MD simulation, and then analyzed selectivity and self-diffusion of CO2 and H2 on graphite membrane layers at different temperatures. **Figure 3** shows a series of typical snapshots, at different temperatures, of mixtures of CO2 and H2 on graphite surfaces, and the number of particles, as shown by Trinh et al. [45]. These authors also reported that the binding energy of CO2 on graphite surface was three times larger than that of H2, and the separation of CO2 over H2 was five times larger at lower temperature, compared to the value obtained at higher temperature. Other mechanisms, such as the pore translocation-limited, surface adsorption and rearrangement reaction, have been in-depth explored and discussed in the literature [46–49].

**2. Graphene and GO for gaseous selectivity separation**

m2 s1 Pa1

42 Advances in Carbon Nanostructures

with permission from [43].

Comparing with traditional membrane materials, graphene and GO membranes show remarkable properties, such as unique stiffness, hydrophobic characteristics and are easy scale up in various applications [38–41]. The monolayer graphene is known to resist to very small gaseous permeation, including He (helium) [33], as the electron-clouds on the graphene hexagonal rings could hamper the atoms and molecules to allow the transit of small molecules. Graphene is long being considered as an ideal gas barrier material. Although graphene prevents gas to pass through, nanoporous graphene (NPG), due to its high surface porosity, can efficiently allow gas flow to selectively pass through, thus becoming a potential membrane material candidate. The analysis of NPG as membrane materials starts with the simulation of this amazing material. In 2008, Sint et al. [42] designed functionalized nanopores in graphene monolayers and showed that they could provide highly selective passage of hydrated ions, using molecular dynamics (MD) simulations. In the same year, Jiang et al. [43] studied NPG and investigated the permeability and selectiv‐ ity of NPG and demonstrated the remarkable potential of graphene as a membrane for the gas separation of H2 and CH4. Their results indicated that H2 permeance could be up to 1 mol/

, with a selectivity of amazing 108 for H2/CH4, which was up to 107 improvement in

the value obtained with silica membranes (**Figure 2**). Du et al. [44] designed a series of NPG for separating N2 and H2 and found that there were different mechanisms for H2 and N2 to

**Figure 2.** (a) Creating a nitrogen-doped pore within one graphene sheet; (b) graphene pores with mixed nitrogen (green) and hydrogen (blue) terminations and (c) only hydrogen-terminated pores. (d) Equivalent pore size from elec‐ tron density calculations for nitrogen- and hydrogen-terminated pores; (e) pure hydrogen-terminated pores. Adapted

**Figure 3.** Typical snapshots of molecules absorbed on graphite surface at different temperatures (T) and total numbers of particles (N). The green, red and white represent C, O (oxygen) and H atoms, respectively [45].

Taking theoretical results into consideration, graphene and GO are promising materials to become the next-generation separation membranes. Scientists start to develop different strategies for synthesizing NPG. Most common methods include electron beam [50], ultra‐ violet-induced oxidative etching [51], electrochemical exfoliation [52] and creation of porous two-dimensional sheets by chemical synthesis [53] or 3D structures by sandwiching/tether‐ ing graphene with carbon nanotubes [54]. Fischbein and Drndić [50] introduced the use of ebeam irradiation to create nanopores on graphene sheets, allowing the porous size on the graphene to be controlled. Russo et al. prepared graphene nanopores with just a 3 Å radii, equivalent to 10 atoms removed, which is shown in **Figure 4** [55]. Ultraviolet light possesses energy to break certain chemical bonds and etching holes on graphene sheets. With this method, Bunch et al. [51] successfully prepared micrometer-sized graphene membranes with pores, which allowed smaller molecules to pass through but not larger molecules. Ambrosi and Pumera [52] tested different electrolytes in the method of electrochemically exfoliation, and suggested that graphene was a great candidate for electrochemical applications, such as biosensing and energy storage. O'Hern et al. [56] fabricated graphene composite mem‐ branes by transferring large area GO sheets onto polymer substrates. Strategies on how to precisely control the thickness and the chemical functionalization groups of graphene sheets [57] were also introduced. Recently, some novel methods are emerged to prepare graphene membranes. For example, Raghavan et al. [58] used plasma to prepare graphene mem‐ branes on the substrates including 6H-SiC etched by CF4- and Cl4-based plasmas and the results showed that the etching process could selectively etch silicon to produce carbon-rich surface layers on the silicon carbide substrate. When annealed, these carbon-rich layers formed graphene films with halogen- and oxygen-based defects. Reserbat-Plantey et al. [59] intro‐ duced an on-chip nanoelectromechanical system (NEMS) for preparing defect-controllable graphene which was believed to be the process allowing preparation of graphene at room temperature. Zhong et al. [60] reported that graphene sheets modified with the appropriate pore size of 4.06 Å could efficiently separate N2 and CO2 by molecular dynamic (MD) simulation and found that the graphene sheets behaved like a mesh that blocked CO2 permeance but allowed N2 to pass through. Besides that, Schrier [61] synthesized a porous graphene structure to separate He, Ne and CH4, and observed that the porous graphene could separate He from alkanes and noble gases. Hauser and Schwerdtfeger [62] discovered that nanopores functionalized graphene membrane with nitrogen-doped porous had the ability to efficiently separate <sup>3</sup> He from <sup>4</sup> He. Zhang et al. [63] designed covalently cross-linked freestanding few-layered GO films using a layer-by-layer deposition technique, in combination with photochemical cross-links, and achieved a controllable and adjustable graphene membrane with good stability that lasted up to 9 days of use. In 2016, Lozada-Hidalgo et al. [64] used graphene membranes to filtrate hydrogen ion isotopes, like molecular sieves, and found that deuteron permeation through these membranes is much slower than proton permeation. This work implies that graphene has the potential to be used as nuclear waste treatment. With the developed technique, larger high quality and uniform graphene sheets can be transferred to various substrates. Graphene sheets, the ultimate membrane platform, will have wider usages in gas separation technology.

Taking theoretical results into consideration, graphene and GO are promising materials to become the next-generation separation membranes. Scientists start to develop different strategies for synthesizing NPG. Most common methods include electron beam [50], ultra‐ violet-induced oxidative etching [51], electrochemical exfoliation [52] and creation of porous two-dimensional sheets by chemical synthesis [53] or 3D structures by sandwiching/tether‐ ing graphene with carbon nanotubes [54]. Fischbein and Drndić [50] introduced the use of ebeam irradiation to create nanopores on graphene sheets, allowing the porous size on the graphene to be controlled. Russo et al. prepared graphene nanopores with just a 3 Å radii, equivalent to 10 atoms removed, which is shown in **Figure 4** [55]. Ultraviolet light possesses energy to break certain chemical bonds and etching holes on graphene sheets. With this method, Bunch et al. [51] successfully prepared micrometer-sized graphene membranes with pores, which allowed smaller molecules to pass through but not larger molecules. Ambrosi and Pumera [52] tested different electrolytes in the method of electrochemically exfoliation, and suggested that graphene was a great candidate for electrochemical applications, such as biosensing and energy storage. O'Hern et al. [56] fabricated graphene composite mem‐ branes by transferring large area GO sheets onto polymer substrates. Strategies on how to precisely control the thickness and the chemical functionalization groups of graphene sheets [57] were also introduced. Recently, some novel methods are emerged to prepare graphene membranes. For example, Raghavan et al. [58] used plasma to prepare graphene mem‐ branes on the substrates including 6H-SiC etched by CF4- and Cl4-based plasmas and the results showed that the etching process could selectively etch silicon to produce carbon-rich surface layers on the silicon carbide substrate. When annealed, these carbon-rich layers formed graphene films with halogen- and oxygen-based defects. Reserbat-Plantey et al. [59] intro‐ duced an on-chip nanoelectromechanical system (NEMS) for preparing defect-controllable graphene which was believed to be the process allowing preparation of graphene at room temperature. Zhong et al. [60] reported that graphene sheets modified with the appropriate pore size of 4.06 Å could efficiently separate N2 and CO2 by molecular dynamic (MD) simulation and found that the graphene sheets behaved like a mesh that blocked CO2 permeance but allowed N2 to pass through. Besides that, Schrier [61] synthesized a porous graphene structure to separate He, Ne and CH4, and observed that the porous graphene could separate He from alkanes and noble gases. Hauser and Schwerdtfeger [62] discovered that nanopores functionalized graphene membrane with nitrogen-doped porous had the ability to

efficiently separate <sup>3</sup>

44 Advances in Carbon Nanostructures

He from <sup>4</sup>

will have wider usages in gas separation technology.

He. Zhang et al. [63] designed covalently cross-linked free-

standing few-layered GO films using a layer-by-layer deposition technique, in combination with photochemical cross-links, and achieved a controllable and adjustable graphene membrane with good stability that lasted up to 9 days of use. In 2016, Lozada-Hidalgo et al. [64] used graphene membranes to filtrate hydrogen ion isotopes, like molecular sieves, and found that deuteron permeation through these membranes is much slower than proton permeation. This work implies that graphene has the potential to be used as nuclear waste treatment. With the developed technique, larger high quality and uniform graphene sheets can be transferred to various substrates. Graphene sheets, the ultimate membrane platform,

**Figure 4.** Preparation of graphene nanopores by e-beam. Starting with pristine, one-atom-thick graphene (a), a mona‐ tomic site is created by bombardment with 3 keV Ar (Argon)+ ions, which remove a few carbon atoms in a single bom‐ bardment (b). After the single-bombardment, a small but stable defect-location is left in the layer surface (c). Then using a uniform e-beam enlargement of the defect-location is achieved by removing the edge atoms (d). Finally, the pore is opened until reaching the preferred size (e) [55].

Compared to graphene, GO membrane materials have a higher fracture energy (350 kJ m−3) and tensile strength (42 ± 2 GPa) [65]. GO can be bent over 90° and form a simple curve (**Figure 5a**) [65]. This stiff-yet-flexible property makes GO an ideal candidate as membrane material. Various routes have been reported for preparing graphene and its oxide derivative with good mechanical properties for different applications. Pierleoni et al. [66] used soluble graphene derivatives to fabricate gas separation membrane coatings on the surface of polymers and found that the use of 2D materials as surface coatings, instead of bulk additives, overcame common issues related to dispersion of graphene into a polymer matrix, and gave a remarkable advantage in preserving the mechanical properties of the bulk polymer. Kim et al. [67] successfully loaded a large area of graphene and GO on the polymer substrate and then measured the transport and selectivity of O2-N2 and CO2-N2. They considered that the defects and the wrinkles on the graphene and GO surfaces (**Figure 5b**) conferred the selectivity and the controllability to the membranes. Bai et al. [68] invented a new graphene nanostructure named graphene nanomesh (**Figure 5c**). They used block copolymer lithography to prepare NPG with neck widths as low as 5 nm and such materials exhibited great membrane perform‐ ance. Yang et al. [69] noticed that the wrinkles on the graphene could lead to buckling behavior, which would significantly influence graphene membrane performance. Li et al. [70] reported ultrathin GO membranes with thickness approaching 1.8 nm, loaded on porous anodic aluminum oxide (AAO). The selectivity of the membranes could reach as high as 3400 and 900 for H2/CO2 and H2/N2, respectively. With successful fabrications of gas flow channels and pore sizes, one can optimize the performances of graphene and GO membranes. Other parameters, including temperature and pressure of the system, can affect the effusing rate, selectivity and flux of graphene and GO membranes. Functional groups inserted at the edge of the pores or between the interlayers can also significantly enhance the selectivity and flux of the graphenebased membranes.

**Figure 5.** (a) GO materials can be folded over 90° [65]. (b) Atomic force microscopy (AFM) image of graphene mem‐ brane surface showing the defects and the wrinkles [67]. (c) Scanning electron microscope (SEM) image of graphene nanomesh [68].

**Figure 6.** (a) The SEM image of the SiNx membrane with 49 pores showed. (b) Comparison of hydrogen-carbon dioxide gas mixture separation performances between the graphene membranes and other membranes: GO, poly (1-trimethyl‐ silyl-1-propyne) (PMSP), polyetherimide (PEI), carbon molecular sieve (CMS), zeolite, silica, metal-organic framework (MOF) and SiC [71].

Traditional membranes, including silica-, polymer- and zeolite-based, follow the well-known law that selectivity is determined by the product of diffusivity and solubility [13]. However, this rule is no longer valid for graphene- and GO-based membranes. For atom-thick graphene, the selectivity is determined by the pore size of GO. The thickness of graphene or GO is also no longer linearly related to the diffusion rates. Although the mechanism is still under intensive study, NPG has shown fluxes of gas in far excess, compared to that of traditional membranes. For example, Celebi et al. [71] fabricated highly efficient mass transfer porous bilayer NPG onto a SiNx frame punctured with matrix pores with 4 μm in diameter (**Figure 6a**), which physically perforated double-layer graphene. The results (**Figure 6b**) showed that the perme‐ ation of the NPG membranes is much higher than that of other traditional membranes, including polymer, silicon and composite materials.

## **3. Graphene and GO for water purification**

which would significantly influence graphene membrane performance. Li et al. [70] reported ultrathin GO membranes with thickness approaching 1.8 nm, loaded on porous anodic aluminum oxide (AAO). The selectivity of the membranes could reach as high as 3400 and 900 for H2/CO2 and H2/N2, respectively. With successful fabrications of gas flow channels and pore sizes, one can optimize the performances of graphene and GO membranes. Other parameters, including temperature and pressure of the system, can affect the effusing rate, selectivity and flux of graphene and GO membranes. Functional groups inserted at the edge of the pores or between the interlayers can also significantly enhance the selectivity and flux of the graphene-

**Figure 5.** (a) GO materials can be folded over 90° [65]. (b) Atomic force microscopy (AFM) image of graphene mem‐ brane surface showing the defects and the wrinkles [67]. (c) Scanning electron microscope (SEM) image of graphene

**Figure 6.** (a) The SEM image of the SiNx membrane with 49 pores showed. (b) Comparison of hydrogen-carbon dioxide gas mixture separation performances between the graphene membranes and other membranes: GO, poly (1-trimethyl‐ silyl-1-propyne) (PMSP), polyetherimide (PEI), carbon molecular sieve (CMS), zeolite, silica, metal-organic framework

based membranes.

46 Advances in Carbon Nanostructures

nanomesh [68].

(MOF) and SiC [71].

Oceans contain about 97% of the world's water. Nowadays the cost of reverse osmosis (RO) decreased to 1.8 KWh/m3 , which is only 1/3 of the cost as was in 1990 [9]. Thermal desalination, such as multiple-effect distillation and multistage flash, needs much more energy to be used, compared to RO. Thus, the development of desalination can only become a sustainable option for water supply if a dramatically new technology is developed.

Graphene, with its extraordinary properties and environmentally friendly nature, has attracted the scientists' interest due to its potential applications in wastewater treatment and water desalination. As a single layer of graphene is impermeable to gases and liquids, NPG is considered as a potential membrane material for molecular sieving or water filtration. Advantages of NPG over RO membranes include negligible thickness and high mechanical strength, indicating faster water flux, with low pressure requirements. The first simulation work on water flux across graphene membrane was conducted by Suk et al. [72]. As shown in **Figure 7**, these authors compared the water flux performances of NPG and CNTs. Results indicated that NPG had higher water flux which did not significantly vary with the thickness of NPG. Then, Cohen-Tanugi and Grossman [73] showed that nanometer-scale pores in singlelayer freestanding graphene could effectively filter NaCl salt from water through MD simu‐ lation results. They demonstrated that NPG membrane had the ability to prevent salt passage, critically depending on the pore diameter and chemical functional groups bonded to the edges of graphene pores. Nair et al. [74] indicated anomalous water permeation to low friction between the water monolayer and pristine graphene regions. Based on that assumption, a capillary-driven flow mechanism was established for graphene-based membranes. Boukhva‐ lov et al. [75] analyzed the hybrid systems consisting of water and GO and believed formation of hexagonal ice bilayer in between the flakes, as well as melting transition of ice at the edges of flakes, were crucial to understand the water permeation across GO. Ma et al. [76] discovered the fast diffusion of water nanodroplets on graphene and opened a prospect for achieving fast and controllable motion of adsorbates across graphene and GO surface materials.

**Figure 7.** (a) Graphene membrane with nanopore diameter: Davg = 0.75 nm (left) and Davg = 2.75 nm (right). (b) Simula‐ tion setup. Cyan color represents carbon atoms; red and white colors represent oxygen and hydrogen atoms of a water molecule, respectively. Two water reservoirs are attached to each side of the porous graphene membrane. Ly=Lx=4 nm when the pore diameter is 0.75 nm, and Ly = Lx = 6 nm when the pore diameter is 2.75 nm. In the shaded region (Δz =1 nm), external forces are applied on water molecules to create a pressure drop across the membrane. (c) Water flux through a (10, 0) CNT with a diameter of 0.78 nm and through a graphene nanopore with a diameter of 0.75 nm. In both cases, a single-file water structure is observed. (d) Water flux through a (20, 20) CNT with a diameter of 2.71 nm and through a graphene nanopore with a diameter of 2.75 nm. A single-file water structure is not observed in this case [72].

In order to improve water permeability, small sheet size, high density of interedge spaces and wide nanochannels are required in graphene and GO membrane fabrication. Wei et al. [77] enhanced water flow in graphene membranes into the unique porous microstructures. These authors pointed out that the side-pinning effects caused by H-bonds between water molecules and oxidized regions could help fast water transportation in graphene channels, and that this, along the porous microstructure, were the principal factors responsible for the enhanced water flow. Other studies also indicated that intercalation of different-sized crosslinkers can tune the interlayer channel size [78]. Furthermore, Wei et al. [79] investigated interlayer flow and demonstrated that the breakdown of flow enhancement could be de‐ rived from GO chemical functionalization as well as relaxation of nanoconfinement. Also, both simulation and experiment were performed to evaluate the characteristics of mem‐ brane structure and operation parameter affect water flux in graphene and GO-based membranes. Cohen-Tanugi and Grossman [73] found out that the water flux through monolayer graphene had a linear relationship with the pore area by computational meth‐ ods. They also proposed that hydrophilic pores could increase water permeation, due to Hbonding between water molecules and pore edges. Hu and Mi [80] fabricated a series of novel GO membranes with different numbers of layers and tested their water purification performance. These authors observed that there was no evident correlation between water flux and the number of layers of graphene. Sheath and Majumder [81] prepared graphene membranes by evaporation-induced capillary-force self-assembly and found out that membranes had higher space rejection properties, while simultaneously enhancing water flux, compared to analogous materials produced by the traditional vacuum filtration techni‐ que. Han et al. [82] synthesized ultrathin graphene nanofiltration membranes (uGNMs) on

nanoporous microstructure substrates, using chemically converted graphene. The uGNMs showed high retention (>99%) for organic dyes and moderate retention (≈20–60%) for ion salts. The physical sieving and electrostatic interaction were considered as two important factors dominating the rejection process (**Figure 8**).

**Figure 7.** (a) Graphene membrane with nanopore diameter: Davg = 0.75 nm (left) and Davg = 2.75 nm (right). (b) Simula‐ tion setup. Cyan color represents carbon atoms; red and white colors represent oxygen and hydrogen atoms of a water molecule, respectively. Two water reservoirs are attached to each side of the porous graphene membrane. Ly=Lx=4 nm when the pore diameter is 0.75 nm, and Ly = Lx = 6 nm when the pore diameter is 2.75 nm. In the shaded region (Δz =1 nm), external forces are applied on water molecules to create a pressure drop across the membrane. (c) Water flux through a (10, 0) CNT with a diameter of 0.78 nm and through a graphene nanopore with a diameter of 0.75 nm. In both cases, a single-file water structure is observed. (d) Water flux through a (20, 20) CNT with a diameter of 2.71 nm and through a graphene nanopore with a diameter of 2.75 nm. A single-file water structure is not observed in this case

In order to improve water permeability, small sheet size, high density of interedge spaces and wide nanochannels are required in graphene and GO membrane fabrication. Wei et al. [77] enhanced water flow in graphene membranes into the unique porous microstructures. These authors pointed out that the side-pinning effects caused by H-bonds between water molecules and oxidized regions could help fast water transportation in graphene channels, and that this, along the porous microstructure, were the principal factors responsible for the enhanced water flow. Other studies also indicated that intercalation of different-sized crosslinkers can tune the interlayer channel size [78]. Furthermore, Wei et al. [79] investigated interlayer flow and demonstrated that the breakdown of flow enhancement could be de‐ rived from GO chemical functionalization as well as relaxation of nanoconfinement. Also, both simulation and experiment were performed to evaluate the characteristics of mem‐ brane structure and operation parameter affect water flux in graphene and GO-based membranes. Cohen-Tanugi and Grossman [73] found out that the water flux through monolayer graphene had a linear relationship with the pore area by computational meth‐ ods. They also proposed that hydrophilic pores could increase water permeation, due to Hbonding between water molecules and pore edges. Hu and Mi [80] fabricated a series of novel GO membranes with different numbers of layers and tested their water purification performance. These authors observed that there was no evident correlation between water flux and the number of layers of graphene. Sheath and Majumder [81] prepared graphene membranes by evaporation-induced capillary-force self-assembly and found out that membranes had higher space rejection properties, while simultaneously enhancing water flux, compared to analogous materials produced by the traditional vacuum filtration techni‐ que. Han et al. [82] synthesized ultrathin graphene nanofiltration membranes (uGNMs) on

[72].

48 Advances in Carbon Nanostructures

**Figure 8.** (a) Digital photo of an uGNM coated on an AAO disk (left) and a twisted uGNM coated on a PVDF mem‐ brane (right). (b) Schematic representation of a brGO: graphene sheet with a certain amount of holes; most of the oxi‐ dized groups are located on the edges and the periphery of the holes. Note that the real graphene sheets extend further than depicted. (c) Schematic view of possible permeation route: water molecules go through the nanochannels of the uGNMs and holes of the graphene sheets and finally reach the pores of supporting membranes. The blank squares present the holes on the graphene sheets (black line). The edges of the brGO and the periphery of the holes are nega‐ tively charged [82].

In order to further optimize the performances of NPG or GO for water treatment, the mecha‐ nism involved needs to be intensively studied. There are two main primary factors concerning the selectivity of the materials used in this process: size exclusion and interactions with functional groups, including chemical and electrostatic ones. Using classical molecular dynamics, Cohen-Tanugi and Grossman [73] concluded that nanopores in single-layer freestanding graphene could effectively filter NaCl salt from water. Small pores, low pressure and hydrophobic pores refuse salts more efficiently, due to direct size exclusion, while the larger effective volume of ions and the lack of H-bonding can lead to a higher energy barrier to ions passage. Prince et al. [83] devised an ultrawetting graphene-based membrane by incorporating amine and carboxyl functionalities onto graphene and then covalently attaching it to a polymer matrix to fabricate a water filtration membrane. They concluded that the graphene increased the water permeability of the membrane by 126% compared with stable membrane selectivity. Hu and Mi [84] prepared a kind of water separation membrane using GO nanosheets and this membrane displayed remarkable rejection of monovalent and divalent salts and moderate or high rejection of organic dyes. Yin et al. [85] prepared a thin-film nanocomposite (TFN) membrane containing GO sheets to enhance polyamide (PA) mem‐ branes for water purification application. These GO sheets were well dispersed in the polya‐ mide (PA) layer and their incorporation improved membrane performances. Authors suggested that interlayer spacing of GO may serve as water channels and hence contribute to the water permeability enhancement. Independently, Han et al. [82] obtained similar results, showing that the observed rejection organic dyes could be assigned to size, electrostatic interaction, and the rejection of ions is based on Donnan's exclusion (or Donnan effect, a name for the behavior of charged particles near a semipermeable membrane that sometimes fail to distribute evenly across the two sides of the membrane). Additionally, Huang et al. [20] investigated different factors including salt concentration, pH and pressure effects on the separation performance of GO membrane toward small molecules and demonstrated how these factors affect such process. Sun et al. [86] confirmed the selection ion penetration and water purification properties of GO membranes. They reported that sodium salts could quickly pass through GO membranes, while heavy-metal salts moved slower (**Figure 9**). Copper salts were absolutely unable to permeate through, as well as organic contaminants. Their followup work [87] revealed, by theoretical calculations, that the coordinative interactions play an important role on the selectivity of the transmembrane transport of alkali and alkaline earth cations. Joshi et al. [88] designed an experiment (**Figure 10a**) to measure the permeation properties of micrometer-thick GO laminates. The normalized permeation rates as per 1 M feed solution (**Figure 10b**) show that ultrafast permeation is only possible for ions smaller than a specific size. Size exclusion affected the selectivity as well as capillary forces could induce the anomalous fast transportation. Their follow-up work [89] indicated that graphene had a large possibility to be used as membranes for desalination and confirmed that 100% salt rejection could be achieved for commonly used ions by utilizing single-layer nonporous graphene. Liu et al. [90] studied the pressure-driven water transport inside the nanochannels of graphene membranes. They found that an ultrafast water flux was observed inside the channels, which increased linearly with the driving pressure, but decreased as the interaction strength between the water molecules and the bilayer increased.

**Figure 9.** Schematic diagram of the penetration processes of different ions through GO membranes [86].

membrane selectivity. Hu and Mi [84] prepared a kind of water separation membrane using GO nanosheets and this membrane displayed remarkable rejection of monovalent and divalent salts and moderate or high rejection of organic dyes. Yin et al. [85] prepared a thin-film nanocomposite (TFN) membrane containing GO sheets to enhance polyamide (PA) mem‐ branes for water purification application. These GO sheets were well dispersed in the polya‐ mide (PA) layer and their incorporation improved membrane performances. Authors suggested that interlayer spacing of GO may serve as water channels and hence contribute to the water permeability enhancement. Independently, Han et al. [82] obtained similar results, showing that the observed rejection organic dyes could be assigned to size, electrostatic interaction, and the rejection of ions is based on Donnan's exclusion (or Donnan effect, a name for the behavior of charged particles near a semipermeable membrane that sometimes fail to distribute evenly across the two sides of the membrane). Additionally, Huang et al. [20] investigated different factors including salt concentration, pH and pressure effects on the separation performance of GO membrane toward small molecules and demonstrated how these factors affect such process. Sun et al. [86] confirmed the selection ion penetration and water purification properties of GO membranes. They reported that sodium salts could quickly pass through GO membranes, while heavy-metal salts moved slower (**Figure 9**). Copper salts were absolutely unable to permeate through, as well as organic contaminants. Their followup work [87] revealed, by theoretical calculations, that the coordinative interactions play an important role on the selectivity of the transmembrane transport of alkali and alkaline earth cations. Joshi et al. [88] designed an experiment (**Figure 10a**) to measure the permeation properties of micrometer-thick GO laminates. The normalized permeation rates as per 1 M feed solution (**Figure 10b**) show that ultrafast permeation is only possible for ions smaller than a specific size. Size exclusion affected the selectivity as well as capillary forces could induce the anomalous fast transportation. Their follow-up work [89] indicated that graphene had a large possibility to be used as membranes for desalination and confirmed that 100% salt rejection could be achieved for commonly used ions by utilizing single-layer nonporous graphene. Liu et al. [90] studied the pressure-driven water transport inside the nanochannels of graphene membranes. They found that an ultrafast water flux was observed inside the channels, which increased linearly with the driving pressure, but decreased as the interaction

50 Advances in Carbon Nanostructures

strength between the water molecules and the bilayer increased.

**Figure 9.** Schematic diagram of the penetration processes of different ions through GO membranes [86].

**Figure 10.** (a) Experimental setup. A U-shaped tube is partitioned by GO membrane into two compartments that are referred to as feed and permeate. (b) Sieving through atomic-scale mesh. The small species permeate with approxi‐ mately the same speed, whereas large metal ions and organic molecules exhibit no detectable permeation [88].

Apart from water dissipation and purification, graphene and GO can also be used on other water treatment applications, such as recovery of acids from iron-based electrolytes. Sun et al. [91] reported that nano-GO (**Figure 11**) membranes could completely block Fe3+ when the ion concentration reached a certain extent (0.01 mol L−1), whereas the H+ permeated through unimpeded. This implies that the use of GO membranes can be applied to the recovery of acids from an FeCl3 solution. On the other hand, Huang et al. [21] prepared a GO membrane supported on a ceramic hollow fiber for separating dimethyl carbonate–water mixtures through a pervaporation process. They considered that the separation process followed a sorption-diffusion mechanism. Tang et al. [92] also utilized a pressurized ultrafiltration for dehydration of ethanol using GO membranes. However, the results obtained from binarycomponent feed tests were lower than the ideal water/ethanol selectivity, found by calculation. This may be attributed to the effects of intermolecular H-bonding between water molecules and the functional groups of GO nanosheets that may enlarge the interlayer spacing and allow more ethanol transport through the GO membrane. Recently, Lin et al. [93] devised a way to fast produce graphene in a CVD process and then grow it on Ni, Cu and a Cu-Ni alloy with proper composition. This work provided a possibility to produce large-scale graphene single crystals, which paved the route towards reaching practical applications of graphene, such as in wafer-scale electronic and optoelectronic devices. Wu et al. [94] used a similar approach to fast grow inch-sized single-crystalline graphene by letting a single nucleus interact with a monolayer at a fast rate. They synthesized a roughly 1.5-inch-large graphene one-atom-layer in 2.5 h and suggested that this method may prove effective for the synthesis of wafer-scale single-crystalline monolayers of other two-dimensional materials.

**Figure 11.** Schematic diagrams of the structure of nanocapillary networks in nano- and micro-GO membranes. It can speculated that when the lateral dimensions of the GO sheets used to form the membranes were reduced from micro‐ meter (left) to nanometer (right), the amount of nanocapillaries formed within the membranes should significantly in‐ crease, resulting in the enhancement of ion penetrations through GO membranes [91].

In summary, graphene and GO membranes are believed to be the next-generation separation materials for applications in water purification due to their significant intrinsic mechanical strength [95], high chemical stability [57], high antibacterial activity [96] and exquisite antifouling properties [65], as mentioned above. The narrow pore size and nanochannels of GO membranes can be tuned by induction and modification of surface functional groups. For water separation and purification, an ideal pore size distribution should be achieved as this is beneficial to precisely sieving ions through the size exclusion mechanism.

## **4. Conclusions and perspective**

Graphene and its oxide derivative, having two-dimensional one-atomic-layer, ultrafast permeation, remarkable mechanical properties and outstanding energy-efficiency, are promising candidates for the next-generation separation membranes for precise and selec‐ tive gas separation (**Figure 12**). Although graphene and GO have been reported as being extraordinary in both selectivity and flux when separating multiple gas mixtures and water purifications, there are still several challenges that need to be solved in order to achieve industrial applications. Firstly, the pores on the surface of graphene and GO sheets can be blocked by defects, impurities, polymer, molecular residues and grain boundaries. They will affect the performance of the membranes, namely the flux rate [97]. Therefore, the mixture should be carefully purified before entering into the graphene membrane systems. The operating conditions, such as pressure, temperature and flux rate, will affect membrane selectivity. These factors must be considered before using graphene- and GO-based mem‐ brane materials. Secondly, the selectivity for H2/H2S, CO2/CH4 or mixtures of three or more gases on gas separation applications of graphene and GO membranes still needs to be fully characterized. A reasonably analysis of the parameters that influence the process of gra‐ phene and GO is needed for predicting the mechanism of membrane selectivity. Those information are of great interest for industries, such as nature and shale gas companies. Although graphene is considered as the world's strongest material with Young's modulus up to 1 TPa, local defects, cracks and wrinkling of graphene and GO will affect the excellent property of graphene sheets [98]. The local bond breaking and bond rotation at the crack tip

of the graphene could lead to dramatic losses of mechanical properties of corresponding products, which may cause the membranes to be less robust and more fragile, and even generate unexpected large pores during constant usage. To solve this issue, a good solution might be to use a reliable porous substrate as support structure for graphene membranes. The connection and adhesion properties between membranes and support structures thus need to be fully revealed. Then, for the membrane itself, pore size is the most critical parameter in controlling the performance of graphene-based membranes; graphene and GO can enhance the selectivity of separating mixtures by adjusting its layer-to-layer distance and/or insert‐ ing functional groups on the surfaces and interlayer-faces of graphene sheet. The last challenge is the overall cost for the industry demands. Large-scale graphene and GO membrane fabrication techniques need to be developed, and related standard quality control tests need to be created to meet the industry requirements. Although the high production cost of GO so far impeded the development of large-scale applications, the coming advances on technolo‐ gy will certainly decrease the manufacturing prices of graphene and GO, and more methods will be developed to generate defect-free, larger area graphene, with controllable pore structure/size and interlayer distance. It is believed that this promising field will attract increasing interest from both academia and industry in the future.

**Figure 11.** Schematic diagrams of the structure of nanocapillary networks in nano- and micro-GO membranes. It can speculated that when the lateral dimensions of the GO sheets used to form the membranes were reduced from micro‐ meter (left) to nanometer (right), the amount of nanocapillaries formed within the membranes should significantly in‐

In summary, graphene and GO membranes are believed to be the next-generation separation materials for applications in water purification due to their significant intrinsic mechanical strength [95], high chemical stability [57], high antibacterial activity [96] and exquisite antifouling properties [65], as mentioned above. The narrow pore size and nanochannels of GO membranes can be tuned by induction and modification of surface functional groups. For water separation and purification, an ideal pore size distribution should be achieved as this is

Graphene and its oxide derivative, having two-dimensional one-atomic-layer, ultrafast permeation, remarkable mechanical properties and outstanding energy-efficiency, are promising candidates for the next-generation separation membranes for precise and selec‐ tive gas separation (**Figure 12**). Although graphene and GO have been reported as being extraordinary in both selectivity and flux when separating multiple gas mixtures and water purifications, there are still several challenges that need to be solved in order to achieve industrial applications. Firstly, the pores on the surface of graphene and GO sheets can be blocked by defects, impurities, polymer, molecular residues and grain boundaries. They will affect the performance of the membranes, namely the flux rate [97]. Therefore, the mixture should be carefully purified before entering into the graphene membrane systems. The operating conditions, such as pressure, temperature and flux rate, will affect membrane selectivity. These factors must be considered before using graphene- and GO-based mem‐ brane materials. Secondly, the selectivity for H2/H2S, CO2/CH4 or mixtures of three or more gases on gas separation applications of graphene and GO membranes still needs to be fully characterized. A reasonably analysis of the parameters that influence the process of gra‐ phene and GO is needed for predicting the mechanism of membrane selectivity. Those information are of great interest for industries, such as nature and shale gas companies. Although graphene is considered as the world's strongest material with Young's modulus up to 1 TPa, local defects, cracks and wrinkling of graphene and GO will affect the excellent property of graphene sheets [98]. The local bond breaking and bond rotation at the crack tip

crease, resulting in the enhancement of ion penetrations through GO membranes [91].

beneficial to precisely sieving ions through the size exclusion mechanism.

**4. Conclusions and perspective**

52 Advances in Carbon Nanostructures

Water purification is also a promising topic worth exploring, in order to achieve the funda‐ mental understanding that might lead to the specialized industrial application of grapheneand GO-based membranes. The first challenge is the precise design of user-controlled GO membranes, including the pattern of oxidized regions, interlayer space, pore size, modification of pore chemistry and the number of graphene layers. Further developments on the parameters that determine the small molecule permeation and separation performance are also necessary. Meanwhile, the effect of the interlayer space and surface defect of graphene and GO on fluid flux is being carefully studied. With those information, the mechanism affecting the selectivity of NPG and GO may be solved. The contradiction between flux and selectivity is considered to determine the optimal membrane fabrication parameters. Modification of the surface of GO membranes with different atoms and functional groups, to adjust water affinity, can achieve higher flux, without sacrificing selectivity of the membrane system. The second topic that needs to be considered is the state of GO in aqueous conditions. The swelling needs to be regulated in order to achieve the ideal selectivity. For industrial processes, the operating pressure is an important factor to both separation system and membrane materials. Operation in a large range of pressures will be highly beneficial. Finally, the last issue is the cost. For some special industrial applications, such as producing high purity acid from high-concentration FeCl3 solutions, pretreatment or posttreatment will increase energy consumption and total investment. Researchers are also suggested to minimize the gap between theoretical and experimental results on water/ion transport mechanisms across GO nanosheets in water and to solve impurity issue on the graphene- and GO-based membranes, including adsorption of metals and organic dyes [99], ionic clogging inside the capillaries, as well as coordination occupation on the surface of oxidized regions. These effects still need investigation and optimization. The key to successfully fabricate adequate membrane systems is to balance the antimicrobial efficiency and mechanical strength of the freestanding membranes.

**Figure 12.** The outlook of the research on graphene and GO membranes.

## **Author details**

Quan Xu1\* and Wenwen Zhang2

\*Address all correspondence to: xuquan@cup.edu.cn

1 State Key Laboratory of Heavy Oil Processing, China University of Petroleum, Beijing, China

2 College of Textile, North Carolina State University, North Carolina, USA

## **References**


**Figure 12.** The outlook of the research on graphene and GO membranes.

\*Address all correspondence to: xuquan@cup.edu.cn

1 State Key Laboratory of Heavy Oil Processing, China University of Petroleum, Beijing, China

[1] Geankoplis C. Transport Processes and Separation Process Principles (Includes Unit Operations), Fourth Edition. Upper Saddle River, NJ, USA: Prentice Hall Press; 2003.

[2] Baker RW. Research needs in the membrane separation industry: looking back, looking

2 College of Textile, North Carolina State University, North Carolina, USA

forward. Journal of Membrane Science 2010;362:134–136.

**Author details**

54 Advances in Carbon Nanostructures

**References**

Quan Xu1\* and Wenwen Zhang2


[34] Leenaerts O, Partoens B, Peeters FM. Graphene: a perfect nanoballoon. Applied Physics Letters. 2008;93(19):193107.

[19] Karan S, Samitsu S, Peng X, Kurashima K, Ichinose I. Ultrafast viscous permeation of organic solvents through diamond-like carbon nanosheets. Science. 2012;335(6067):

[20] Huang H, Ying Y, Peng X. Graphene oxide nanosheet: an emerging star material for novel separation membranes. Journal of Materials Chemistry A. 2014;2(34):13772–

[21] Huang K, Liu G, Lou Y, Dong Z, Shen J, Jin W. A graphene oxide membrane with highly selective molecular separation of aqueous organic solution. Angewandte Chemie

[22] Novoselov KS, Geim AK, Morozov SV, Jiang D, Zhang Y, Dubonos SV, et al. Electric field effect in atomically thin carbon films. Science. 2004;306(5696):666–669.

[23] Tsetseris L, Pantelides ST. Graphene: an impermeable or selectively permeable

[24] Gai J-G, Gong X-L, Wang W-W, Zhang X, Kang W-L. An ultrafast water transport forward osmosis membrane: porous graphene. Journal of Materials Chemistry A.

[25] Liu H, Dai S, Jiang D-E. Permeance of H2 through porous graphene from molecular

[26] Brodie BC. On the atomic weight of graphite. Philosophical Transactions of the Royal

[27] Kohlschütter V, Haenni P. Zur Kenntnis des Graphitischen Kohlenstoffs und der Graphitsäure. Zeitschrift für Anorganische und Allgemeine Chemie. 1919;105(1):121–

[28] Hummers Jr WS, Offeman RE. Preparation of graphitic oxide. Journal of the American

[29] Sedigh MG, Onstot WJ, Xu L, Peng WL, Tsotsis TT, Sahimi M. Experiments and simulation of transport and separation of gas mixtures in carbon molecular sieve

membranes. The Journal of Physical Chemistry. A. 1998;102(44):8580–8589.

[31] Geim AK. Graphene: status and prospects. Science. 2009;324(5934):1530–1534.

functionalizing graphene. MRS Bulletin. 2012;37(12):1187–1194.

[30] Geim AK, Novoselov KS. The rise of graphene. Nature Materials. 2007;6(3):183–191.

[32] Pantelides ST, Puzyrev Y, Tsetseris L, Wang B. Defects and doping and their role in

[33] Bunch JS, Verbridge SS, Alden JS, van der Zande AM, Parpia JM, Craighead HG, et al. Impermeable atomic membranes from graphene sheets. Nano Letters. 2008;8(8):2458–

International Edition. 2014;53(27):6929–6932.

Society of London 1859;149:249–259.

Chemical Society. 1958;80(6):1339–1340.

membrane for atomic species? Carbon 2014;67:58–63.

dynamics. Solid State Communications. 2013;175–176:101–105.

444–447.

56 Advances in Carbon Nanostructures

13782.

2014;2(11):4023.

144.

2462.


[63] Huang T, An Q, Luan X, Zhang Q, Zhang Y. Free-standing few-layered graphene oxide films: selective, steady and lasting permeation of organic molecules with adjustable speeds. Nanoscale. 2016;8:2003–2010.

[49] Zhou S, Bongiorno A. Density functional theory modeling of multilayer "epitaxial"

[50] Fischbein MD, Drndić M. Electron beam nanosculpting of suspended graphene sheets.

[51] Koenig SP, Wang L, Pellegrino J, Bunch JS. Selective molecular sieving through porous

[52] Ambrosi A, Pumera M. Electrochemically exfoliated graphene and graphene oxide for energy storage and electrochemistry applications. Chemistry – A European Journal.

[53] Kuhn P, Forget A, Su D, Thomas A, Antonietti M. From microporous regular frame‐ works to mesoporous materials with ultrahigh surface area: dynamic reorganization of porous polymer networks. Journal of American Chemical Society. 2008;130(40):

[54] Niu J, Li M, Choi W, Dai L, Xia Z. Growth of junctions in 3D carbon nanotube-graphene nanostructures: a quantum mechanical molecular dynamic study. Carbon 2014;67:627–

[55] Russo CJ, Golovchenko JA. Atom-by-atom nucleation and growth of graphene nanopores. Proceedings of the National Academy of Sciences. 2012;109(16):5953–5957.

[56] O'Hern SC, Stewart CA, Boutilier MSH, Idrobo J-C, Bhaviripudi S, Das SK, et al. Selective molecular transport through intrinsic defects in a single layer of CVD

[57] Dreyer DR, Park S, Bielawski CW, Ruoff RS. The chemistry of graphene oxide. Chemical

[58] Raghavan S, Denig TJ, Nelson TC, Chaudhari S, Stinespring CD. Characterization of graphene-on-insulator films formed using plasma based surface chemistry. Carbon

[59] Reserbat-Plantey A, Schädler KG, Gaudreau L, Navickaite G, Güttinger J, Chang D, et al. Electromechanical control of nitrogen-vacancy defect emission using graphene

[60] Wang Y, Yang Q, Li J, Yang J, Zhong C. Exploration of nanoporous graphene mem‐ branes for the separation of N2 from CO2: a multi-scale computational study. Physical

[61] Schrier J. Helium separation using porous graphene membranes. The Journal of

[62] Hauser AW, Schwerdtfeger P. Nanoporous graphene membranes for efficient 3He/4He

separation. The Journal of Physical Chemistry Letters. 2012;3(2):209–213.

graphene oxide. Accounts of Chemical Research. 2014;47(11):3331–3339.

Applied Physics Letters. 2008;93(11):113107–1.

graphene. ACS Nano. 2012;6(11):10130–10138.

NEMS. Nature Communications 2016;7:10218.

Physical Chemistry Letters. 2010;1(15):2284–2287.

Society Reviews. 2010;39(1):228–240.

Chemistry Chemical Physics. 2015.

2016;22(1):153–159.

13333–13337.

58 Advances in Carbon Nanostructures

2016;99:212–221.

634.

graphene. Nature Nanotechnology. 2012;7(11):728–732.


[92] Tang YP, Paul DR, Chung TS. Free-standing graphene oxide thin films assembled by a pressurized ultrafiltration method for dehydration of ethanol. Journal of Membrane Science 2014;458:199–208.

[77] Wei N, Peng X, Xu Z. Understanding water permeation in graphene oxide membranes.

[78] Mi B. Graphene oxide membranes for ionic and molecular sieving. Science.

[79] Wei N, Peng X, Xu Z. Breakdown of fast water transport in graphene oxides. Physical

[80] Hu M, Mi B. Layer-by-layer assembly of graphene oxide membranes via electrostatic

[81] Sheath P, Majumder M. Flux accentuation and improved rejection in graphene-based filtration membranes produced by capillary-force-assisted self-assembly. Philosophi‐

[82] Han Y, Xu Z, Gao C. Ultrathin graphene nanofiltration membrane for water purifica‐

[83] Prince JA, Bhuvana S, Anbharasi V, Ayyanar N, Boodhoo KVK, Singh G. Ultra-wetting graphene-based membrane. Journal of Membrane Science 2016;500:76–85.

[84] Hu M, Mi B. Enabling graphene oxide nanosheets as water separation membranes.

[85] Yin J, Zhu G, Deng B. Graphene oxide (GO) enhanced polyamide (PA) thin-film nanocomposite (TFN) membrane for water purification. Desalination 2016;379:93–101.

[86] Sun P, Zhu M, Wang K, Zhong M, Wei J, Wu D, et al. Selective ion penetration of

[87] Sun P, Zheng F, Zhu M, Song Z, Wang K, Zhong M, et al. Selective trans-membrane transport of alkali and alkaline earth cations through graphene oxide membranes based

[88] Joshi RK, Carbone P, Wang FC, Kravets VG, Su Y, Grigorieva IV, et al. Precise and ultrafast molecular sieving through graphene oxide membranes. Science.

[89] You Y, Sahajwalla V, Yoshimura M, Joshi RK. Graphene and graphene oxide for

[90] Liu B, Wu R, Baimova JA, Wu H, Law AW-K, Dmitriev SV, et al. Molecular dynamics study of pressure-driven water transport through graphene bilayers. Physical Chem‐

[91] Sun P, Wang K, Wei J, Zhong M, Wu D, Zhu H. Effective recovery of acids from ironbased electrolytes using graphene oxide membrane filters. Journal of Materials

ACS Applied Materials & Interfaces. 2014;6(8):5877–5883.

interaction. Journal of Membrane Science 2014;469:80–87.

cal Transactions of the Royal Society A. 2016;374(2060):20150028.

tion. Advanced Functional Materials. 2013;23(29):3693–3700.

Environmental Science & Technology. 2013;47(8):3715–3723.

graphene oxide membranes. ACS Nano. 2013;7(1):428–437.

on cation−π interactions. ACS Nano. 2014;8(1):850–859.

desalination. Nanoscale. 2016;8(1):117–119.

istry Chemical Physics. 2016;18(3):1886–1896.

2014;343(6172):740–742.

60 Advances in Carbon Nanostructures

Review E. 2014;89(1).

2014;343(6172):752–754.

Chemistry A. 2014;2(21):7734.


## **Laser-Scribing Technology for Wafer-Scale Graphene Devices**

He Tian, Mohammad Ali Mohammad, Wen-Tian Mi, Yi Yang and Tian-Ling Ren

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64228

#### **Abstract**

Graphene has attracted a lot of attention due to its amazing properties. A huge number of novel devices, covering the electric, acoustic, photonic, magnetic and mechanical domains, can be developed with graphene. Its ultrahigh mobility can enable ultra-fast transistors or photodetectors. However, the natural zero bandgap of graphene, with insufficient on/off ratio, limits its practical applications. In this chapter, we introduce laser-scribing technology that enables wafer-scale production of graphene devices. Moreover, such laser-scribed graphene (LSG) is, infact, semi-reduced graphene oxide with a finite bandgap, which is suitable for practical applications. We show five kinds of representative LSG devices and their integration. These devices are a resistive memory, an earphone, a strain sensor, a pressure sensor and a light-emitting device. These LSG devices are high-performance, flexible and low cost, which demonstrates the practical nature of laser-scribed graphene-based materials. Finally, an outlook is presented regarding how laser scribing, a serial patterning method, may lead to similar developments in various other serial lithography techniques, such as ion beam lithography.

**Keywords:** graphene, graphene oxide, two-dimensional device, laser-scribing technol‐ ogy, laser reduction, nanoelectronics

## **1. Introduction**

Graphene is a well-known two-dimensional material that has attracted wide attention due to its unique properties [1], which have not been observed in three-dimensional (bulk) materi‐ als. Graphene's mobility at room temperature can be as high as 200,000 cm2 V−1 s−1 [2], which

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

can be regarded as within the realm of ballistic transport and is much higher than the mobility of silicon. Moreover, graphene can absorb light in a wide range, even up to the far infrared [3], which is much broader than the capability of a conventional photodetector. Another powerful property is that its Young's modulus is 1 TPa [4], which makes it one of the strongest materials ever known. Graphene's thermal conductivity is also amazingly high at 5000 W/mK [5], which is also a record. Based on graphene's superior properties, a huge number of novel devices can be made using graphene, covering the electric, acoustic, photonic, magnetic and mechanical domains. However, graphene is not perfect. Its natural zero bandgap, with an insufficient on/off ratio, limits its practical applications. Especially, in logic circuits, the transistor needs to be turned off with a low current, which can reduce its power. This cannot be done in a graphene-based transistor, since the on/off ratio is only less than 10. Moreover, the current cannot be saturated under a high drain bias, which is also a drawback.

**Figure 1.** GO film obtained via the vacuum filtration method. (a) GO film on filter paper. (b) GO film on glass. (c) GO film on polyethylene terephthalate (PET). Reprinted by permission from Macmillan Publishers Ltd: Nature Nanotech‐ nology (3:270), copyright (2008).[8].

A number of researchers tried to artificially introduce a bandgap in graphene. There are mainly three ways to do this. The first two ways consist of using a bilayer graphene and a graphene nanoribbon. The bandgap of bilayer graphene can only be opened around 30 meV, which is still not large enough. The graphene nanoribbon can provide sufficient on/off ratio, when the width of the nanoribbon is reduced down to 10 nm. However, the fabrication of graphene nanoribbon down to 10 nm is very challenging. Since these two methods cannot be scaled up, we refrain from discussing the details here. In this chapter, we mainly focus on the third method—using reduced graphene oxide (rGO). The bandgap of rGO can be tuned from 0∼1.9 eV by controlling the quantity or level of reduction from graphene oxide (GO). Large scale graphene films have been prepared by chemical methods at low cost. A graphene oxide (GO) dispersion could be obtained by the Hummer's method [6], and the graphene oxide film can be prepared through various methods [7], that is spin-coating, spray-coating and vacuum filtration. Ultimately, the graphene film could be obtained by reduction. Some traditional methods include thermal and chemical reduction. Since there would be some oxygen func‐ tional groups remaining in the film after reduction, the width of the bandgap depends on how much oxygen will remain in the film. This can be tuned by controlling, for example, the reduction temperature. Large scale GO films can be prepared by a vacuum filtration method. This technique was firstly reported by Eda et al. (**Figure 1**), which revealed the application potential of rGO in transparent electrodes and transistors [8]. The next step is how to control the shape of rGO.

can be regarded as within the realm of ballistic transport and is much higher than the mobility of silicon. Moreover, graphene can absorb light in a wide range, even up to the far infrared [3], which is much broader than the capability of a conventional photodetector. Another powerful property is that its Young's modulus is 1 TPa [4], which makes it one of the strongest materials ever known. Graphene's thermal conductivity is also amazingly high at 5000 W/mK [5], which is also a record. Based on graphene's superior properties, a huge number of novel devices can be made using graphene, covering the electric, acoustic, photonic, magnetic and mechanical domains. However, graphene is not perfect. Its natural zero bandgap, with an insufficient on/off ratio, limits its practical applications. Especially, in logic circuits, the transistor needs to be turned off with a low current, which can reduce its power. This cannot be done in a graphene-based transistor, since the on/off ratio is only less than 10. Moreover,

the current cannot be saturated under a high drain bias, which is also a drawback.

**Figure 1.** GO film obtained via the vacuum filtration method. (a) GO film on filter paper. (b) GO film on glass. (c) GO film on polyethylene terephthalate (PET). Reprinted by permission from Macmillan Publishers Ltd: Nature Nanotech‐

A number of researchers tried to artificially introduce a bandgap in graphene. There are mainly three ways to do this. The first two ways consist of using a bilayer graphene and a graphene nanoribbon. The bandgap of bilayer graphene can only be opened around 30 meV, which is still not large enough. The graphene nanoribbon can provide sufficient on/off ratio, when the width of the nanoribbon is reduced down to 10 nm. However, the fabrication of graphene nanoribbon down to 10 nm is very challenging. Since these two methods cannot be scaled up, we refrain from discussing the details here. In this chapter, we mainly focus on the third method—using reduced graphene oxide (rGO). The bandgap of rGO can be tuned from 0∼1.9 eV by controlling the quantity or level of reduction from graphene oxide (GO). Large scale graphene films have been prepared by chemical methods at low cost. A graphene oxide (GO) dispersion could be obtained by the Hummer's method [6], and the graphene oxide film can be prepared through various methods [7], that is spin-coating, spray-coating and vacuum filtration. Ultimately, the graphene film could be obtained by reduction. Some traditional methods include thermal and chemical reduction. Since there would be some oxygen func‐ tional groups remaining in the film after reduction, the width of the bandgap depends on how much oxygen will remain in the film. This can be tuned by controlling, for example, the reduction temperature. Large scale GO films can be prepared by a vacuum filtration method. This technique was firstly reported by Eda et al. (**Figure 1**), which revealed the application

nology (3:270), copyright (2008).[8].

64 Advances in Carbon Nanostructures

Although large scale graphene films could be obtained by various methods, the laser-scribing procedure [9, 10] was developed to achieve a precise pattern of graphene, which had never been attempted before. The graphene pattern can be simply and cost-effectively obtained by a single-step laser scribing, even on flexible substrates.

The laser-scribing technology is a method in which a laser is used to convert the graphene oxide into reduced graphene oxide. With the use of the LightScribe DVD drive, specific patterns can be obtained. The laser (wavelength 788 nm, power 5 mW) in the LightScribe DVD drive can reduce the GO to rGO (**Figure 2a**). The laser can be controlled by a software, such as the Nero StartSmart software, and thus the input image can be printed onto the GO as rGO. An in-plane transistor array design is shown in **Figure 2b**, which could be revealed on the graphene film in ∼25 min by laser scribing. The golden film in **Figure 2c** is graphene oxide and the black part is multilayer graphene. Complex graphics could also be converted into graphene patterns, such as the Tsinghua university logo, shown in **Figure 2d** and **e**. And a coloured picture could be converted into a greyscale graphene pattern (**Figure 2f** and **g**), in which the greyscale directly corresponds to the degree of reduction.

**Figure 2.** Demos of laser-scribed graphene. (a) Laser-scribing platform. (b) Layout for wafer-scale in-plane transistors. (c) Printed wafer-scale in-plane transistors. (d) An original image of the Tsinghua University logo. (e) The reproduced Tsinghua logo using LSG technology. (f) An original colourful image. (g) The printed LSG pattern in grey scale. Re‐ printed by permission from Macmillan Publishers Ltd: Scientific Reports (4:3598), copyright (2014) [16].

## **2. Characterization of laser-scribed graphene**

In this section, we will demonstrate how to produce laser-scribed graphene(LSG) and dis‐ cuss its characterization. The spot size of the laser, which is 20 μm in this work, determines the precision of the laser-scribed graphene. Post-exposure, it could be observed under an optical microscope that the laser-scribed graphene consists of long micro-ribbons (**Figure 3**). **Figure 3a** and **b** are the optical images of the laser-scribed graphene micro-ribbons under low magnification. Under visible-light irradiation, the GO is white and the rGO is black.

Increasing the magnification, a white line in the centre of the rGO micro-ribbon could be observed. Remarkably, different focus planes of GO and rGO, which are clearly shown in **Figure 3c** and **d**, suggest that the laser-scribed graphene micro-ribbon has a 3D structure. The total length of the rGO micro-ribbon can be very long (up to cm scale). If roll-to-roll technology is integrated with such laser-scribing technology, the length of rGO micro-rib‐ bons can be even longer up to the meter scale. The most striking feature of the laser-scribing technology is that the rGO patterns can be obtained in a single step, which is time-efficient and low cost.

**Figure 3.** The optical image of a laser-scribed graphene micro-ribbon. (a) An optical image with focus on the GO sur‐ face. (b) The same optical image with focus on the rGO surface. (c) The zoomed-in image showing: (a) the GO surface and (d) the rGO surface, with z-direction profile.

**Figure 4.** A false coloured SEM image of a laser-scribed graphene micro-ribbon. The red part shows the GO surface, and the yellow part shows the LSG foam-like structure. Reprinted from Ref. [14] (CC BY 4.0).

A falsely coloured SEM image (**Figure 4**) of laser-scribed graphene reveals the 3D structure more clearly. The red part in the image represents the GO, and the yellow part represents the rGO. The contrast of light and shade in the image of the rGO ribbon indicates the valley in the centre and the elevation on the two sides, which resemble the foam-like structure.

Increasing the magnification, a white line in the centre of the rGO micro-ribbon could be observed. Remarkably, different focus planes of GO and rGO, which are clearly shown in **Figure 3c** and **d**, suggest that the laser-scribed graphene micro-ribbon has a 3D structure. The total length of the rGO micro-ribbon can be very long (up to cm scale). If roll-to-roll technology is integrated with such laser-scribing technology, the length of rGO micro-rib‐ bons can be even longer up to the meter scale. The most striking feature of the laser-scribing technology is that the rGO patterns can be obtained in a single step, which is time-efficient

**Figure 3.** The optical image of a laser-scribed graphene micro-ribbon. (a) An optical image with focus on the GO sur‐ face. (b) The same optical image with focus on the rGO surface. (c) The zoomed-in image showing: (a) the GO surface

**Figure 4.** A false coloured SEM image of a laser-scribed graphene micro-ribbon. The red part shows the GO surface,

A falsely coloured SEM image (**Figure 4**) of laser-scribed graphene reveals the 3D structure more clearly. The red part in the image represents the GO, and the yellow part represents the

and the yellow part shows the LSG foam-like structure. Reprinted from Ref. [14] (CC BY 4.0).

and low cost.

66 Advances in Carbon Nanostructures

and (d) the rGO surface, with z-direction profile.

The structural change of graphene before and after the laser scribing is shown in **Figure 5**. The thickness of the laser-scribed graphene is 10 μm, which is 10 times larger than that of the original GO film. It could be observed, in the magnified image, that the rGO is comprised of loosely stacked graphene layers and the GO film is made of dense layers. In the case of thermal reduction, a totally opposite behaviour is observed. The thickness of the graphene film decreases after thermal reduction, which results from the loss of the oxygen functional groups. However, the high-energy laser beam rapidly converts the oxygen functional groups into a gas containing oxygen, which finally leads to the loose structure of rGO and an increase of the total thickness.

**Figure 5.** The lateral view of the laser-scribed graphene 3D structure. (a) A SEM profile showing the LSG and GO. (b) Zoomed-in image showing the loosely stacked graphene layers. (c) Zoomed-in image showing the dense GO film. Re‐ printed by permission from Macmillan Publishers Ltd: Scientific Reports (4:3598), copyright (2014) [16].

**Figure 6.** TEM images of (a) laser-scribed graphene and (b) GO (insets in both panels show the diffraction patterns from the atomic structure). Reprinted by permission from Macmillan Publishers Ltd: Scientific Reports (4:3598), copy‐ right (2014) [16].

The diffraction pattern of the LSG (**Figure 6a**) confirms the hexagonal cellular lattice of graphene. On the contrary, the absence of a crystal lattice in the diffraction pattern of the stacked GO film (**Figure 6b**) indicates the broken C-C covalent bond, which results from the lattice defects bonding with oxygen functional groups.

Raman characterization is a quite efficient method to identify the properties of graphene. There are mainly three peaks present in graphene's Raman spectrum: the D-band (1350 cm−1), G-band (1580 cm−1) and 2D-band (2700 cm−1). The D-band indicates the defects of the carbon lattice. If the D-band is strong, defects are present. The G-band position is sensitive to the doping level of graphene. If there is a right shift of the G-band, it means that the graphene is doped. The 2D-band/G-band ratio shows the layer number information. If the ratio is larger than 1, the film under inspection should be monolayer graphene. If it is around 1, then this is a strong indication of bilayer graphene. When the layer thickness number is larger than 3, the ratio is lower than 1. As shown in **Figure 7**, the change of the lattice structure could also be revealed in the Raman spectrum. The decrease of the D peak intensity suggests that part of the C–C bonds were repaired after the laser scribing. However, since there are still some defects left in the laser-scribed graphene, the D peak remains obvious. Besides, the left shift of the G peak indicates that the doping level of graphene decreased after the reduction, resulting from the removal of oxygen. And the appearance of the obvious 2D peak after the laser scribing indicates that the stacked multilayer graphene is generated after reduction.

**Figure 7.** Raman spectra of GO and rGO.

X-ray photoelectron spectroscopy (XPS) is commonly used to identify the functional groups of various materials. Through the binding energy, the chemical formula and the electronic state of the elements can be revealed. The component of the laser-scribed graphene is shown in the XPS spectra (**Figure 8**). As shown in **Figure 8a**, the upper red spectrum is from the LSG and the lower black spectrum is from the GO. Comparing the LSG spectrum with the initial GO film spectrum, the oxygen peak obviously decreases for LSG. This is a direct evidence that the laser can reduce the oxygen from GO. The C1s peaks are analysed carefully to identify the functional groups on the carbon lattice. After fitting the sub-peaks, the GO film contains the C–C, disorder, carbonyl and π–π bonds. By comparing the LSG, it shows that the peaks of the C–C sp2 bond and π-π bond are enhanced after laser reduction.

**Figure 8.** XPS spectra of GO and rGO. (a) The whole spectra of GO and rGO with the C1s and O1s peaks. (b) C1s peak of rGO. (c) C1s peak of GO.

## **3. LSG devices and their application**

The diffraction pattern of the LSG (**Figure 6a**) confirms the hexagonal cellular lattice of graphene. On the contrary, the absence of a crystal lattice in the diffraction pattern of the stacked GO film (**Figure 6b**) indicates the broken C-C covalent bond, which results from the

Raman characterization is a quite efficient method to identify the properties of graphene. There are mainly three peaks present in graphene's Raman spectrum: the D-band (1350 cm−1), G-band (1580 cm−1) and 2D-band (2700 cm−1). The D-band indicates the defects of the carbon lattice. If the D-band is strong, defects are present. The G-band position is sensitive to the doping level of graphene. If there is a right shift of the G-band, it means that the graphene is doped. The 2D-band/G-band ratio shows the layer number information. If the ratio is larger than 1, the film under inspection should be monolayer graphene. If it is around 1, then this is a strong indication of bilayer graphene. When the layer thickness number is larger than 3, the ratio is lower than 1. As shown in **Figure 7**, the change of the lattice structure could also be revealed in the Raman spectrum. The decrease of the D peak intensity suggests that part of the C–C bonds were repaired after the laser scribing. However, since there are still some defects left in the laser-scribed graphene, the D peak remains obvious. Besides, the left shift of the G peak indicates that the doping level of graphene decreased after the reduction, resulting from the removal of oxygen. And the appearance of the obvious 2D peak after the laser scribing indicates

X-ray photoelectron spectroscopy (XPS) is commonly used to identify the functional groups of various materials. Through the binding energy, the chemical formula and the electronic state of the elements can be revealed. The component of the laser-scribed graphene is shown in the XPS spectra (**Figure 8**). As shown in **Figure 8a**, the upper red spectrum is from the LSG and

lattice defects bonding with oxygen functional groups.

68 Advances in Carbon Nanostructures

that the stacked multilayer graphene is generated after reduction.

**Figure 7.** Raman spectra of GO and rGO.

In this section, five different devices are presented, namely a memory device, an earphone, a strain sensor, a pressure sensor and a light emitter; all fabricated using laser-scribed graphene patterning technology. Each device will be presented in brief, while the readers are referred to the relevant citations to explore further details. Finally, multiple types of devices are patterned on the same surface in a single step, in close proximity, demonstrating the tremen‐ dously useful integration potential of this technique.

#### **3.1. Graphene resistive random access memory**

Using laser-scribing technology, a flexible graphene resistive random access memory or resistive memory (RRAM) [11] with a Fin-FET-like structure (**Figure 9a**) can be developed, which has great potential to increase the robustness of the device.

**Figure 9.** Flexible graphene RRAM device structure and fabrication process. (a) Device structure. (b) Fabrication proc‐ ess. Reprinted with permission from Nano Letters 14:3214. Copyright (2014) American Chemical Society [11].

As shown in **Figure 9b**, the GO dispersion was first spin-coated on the flexible polyethylene terephthalate (PET) substrates, following which the GO could be converted into rGO at specific locations. Then, the HfOx and the silver top electrodes were deposited and patterned consec‐ utively. For further details of the fabrication process and the testing method, please refer to Ref. [11].

**Figure 10.** Measurement results of the flexible graphene RRAM. (a) Forming process. (b) Switching performance in the first, 50th and 100th cycles. (c) Switching curves under different compliance currents. (d) High- and low-resistancestate distribution. (e) High- and low-resistance-state under different cycle numbers. (f) Retention testing up to 104 s for both high- and low-resistance-state. Reprinted with permission from Nano Letters 14:3214. Copyright (2014) American Chemical Society [11].

The excellent performance of the flexible graphene RRAM is clearly shown in **Figure 10**, including the forming-free feature (**Figure 10a**), repeatability (**Figure 10b**), multi-valued feature (**Figure 10c**), uniformity (**Figure 10d**), feasible memory window (**Figure 10e**) and stability (**Figure 10f**).

**Figure 11.** The working principle of the LSG RRAM. (a) The on-resistance *versus* the temperature. (b) The energy dia‐ gram for the filament formation. (c,d,e) The dynamic process in the filament region under increasing temperatures. Reprinted with permission from Nano Letters 14:3214. Copyright (2014) American Chemical Society [11].

The working principle of the graphene RRAM could be revealed by the temperature charac‐ teristics analysis (**Figure 11**). When the temperature is raised, the increase of the resistance indicates that the conductive filaments are made of silver. As the temperature is continually raised, some silver filaments transform into silver ions, leading to the increase of the resistance. As the silver ions are also conductive, if they increase, the resistance falls again.

This working principle may be confirmed by electrical measurements. The threshold voltage is only 0.5 V when the top electrode is made of silver. But for the device with Pt top electrodes, the threshold voltage is 2 V, which suggests that the soft breakdown voltage of the HfOx is 2 V. Therefore, it is obvious that the HfOx could not be broken down under 0.5 V, to form the oxygen filament, which confirms the claim that the filament is made of silver.

#### **3.2. Graphene earphone**

s for

**3.1. Graphene resistive random access memory**

70 Advances in Carbon Nanostructures

Ref. [11].

Chemical Society [11].

stability (**Figure 10f**).

which has great potential to increase the robustness of the device.

Using laser-scribing technology, a flexible graphene resistive random access memory or resistive memory (RRAM) [11] with a Fin-FET-like structure (**Figure 9a**) can be developed,

**Figure 9.** Flexible graphene RRAM device structure and fabrication process. (a) Device structure. (b) Fabrication proc‐ ess. Reprinted with permission from Nano Letters 14:3214. Copyright (2014) American Chemical Society [11].

As shown in **Figure 9b**, the GO dispersion was first spin-coated on the flexible polyethylene terephthalate (PET) substrates, following which the GO could be converted into rGO at specific locations. Then, the HfOx and the silver top electrodes were deposited and patterned consec‐ utively. For further details of the fabrication process and the testing method, please refer to

**Figure 10.** Measurement results of the flexible graphene RRAM. (a) Forming process. (b) Switching performance in the first, 50th and 100th cycles. (c) Switching curves under different compliance currents. (d) High- and low-resistancestate distribution. (e) High- and low-resistance-state under different cycle numbers. (f) Retention testing up to 104

both high- and low-resistance-state. Reprinted with permission from Nano Letters 14:3214. Copyright (2014) American

The excellent performance of the flexible graphene RRAM is clearly shown in **Figure 10**, including the forming-free feature (**Figure 10a**), repeatability (**Figure 10b**), multi-valued feature (**Figure 10c**), uniformity (**Figure 10d**), feasible memory window (**Figure 10e**) and

Flexible graphene earphones[12] can also be fabricated based on the laser-scribing technology. Compared with traditional earphones, the graphene earphone has the characteristics of flexibility and ultra-low thickness.

**Figure 12.** Graphene earphone. (a) Fabrication process of the graphene earphone using laser-scribing technology. (b) Wafer-scale graphene earphone after the laser scribing (inset showing the image of a single device). (c) SEM image showing the LSG profile. (d) I–V electrical properties before and after the laser scribing. Reprinted with permission from ACS Nano 8:5883. Copyright (2014) American Chemical Society [12].

Using the laser-scribing technology, the graphene sound source array could be fabricated on a PET substrate (**Figure 12a** and **b**). The I–V curve indicates the good conductivity of the graphene device. The sound source was wired out and packaged into a commercial earphone case (**Figure 13**). Further details about the fabrication and testing can be found in Ref. [12].

**Figure 13.** Packaged graphene earphone. (a) Graphene earphone in hand. (b) Graphene earphone after packaging. (c) The 3D assembled structure of the graphene earphone. (d) A pair of graphene earphones. Reprinted with permission from ACS Nano 8:5883. Copyright (2014) American Chemical Society [12].

The acoustic properties of the graphene earphone are tested on the measurement system, as shown in **Figure 14a**. The test results indicate that the sound pressure level of the frequency spectrum is flatter than that of commercial earphones, especially in the ultrasonic band. Moreover, the frequency spectrum of the graphene earphone covers not only the audible band (from 20 Hz to 20 kHz for human being) but also the ultrasonic band (from 20 to 50 kHz), which is the audibility zone for some animals. This implies that the graphene earphone could be a useful tool for interspecies communication, which is also demonstrated in **Figure 15**. The 35 kHz acoustic signal is conveyed to a dog via graphene earphones, and with the aid of some prior training, the dog stood up as soon as the signal was delivered (**Figure 15**).

**Figure 14.** Performance testing of the graphene earphone. (a) Acoustic test platform for the graphene earphone. (b) Acoustic spectrum of the graphene earphone compared with a commercial one. Reprinted with permission from ACS Nano 8:5883. Copyright (2014) American Chemical Society [12].

**Figure 15.** The graphene earphone as a tool for interspecies communication. (a) A pair of graphene earphones. (b) A dog wearing a graphene earphone. (c) Sound waves at 35 kHz applied to the dog signals it to stand up. Reprinted with permission from ACS Nano 8:5883. Copyright (2014) American Chemical Society [12].

#### **3.3. Graphene strain sensor**

**Figure 12.** Graphene earphone. (a) Fabrication process of the graphene earphone using laser-scribing technology. (b) Wafer-scale graphene earphone after the laser scribing (inset showing the image of a single device). (c) SEM image showing the LSG profile. (d) I–V electrical properties before and after the laser scribing. Reprinted with permission

Using the laser-scribing technology, the graphene sound source array could be fabricated on a PET substrate (**Figure 12a** and **b**). The I–V curve indicates the good conductivity of the graphene device. The sound source was wired out and packaged into a commercial earphone case (**Figure 13**). Further details about the fabrication and testing can be found in Ref. [12].

**Figure 13.** Packaged graphene earphone. (a) Graphene earphone in hand. (b) Graphene earphone after packaging. (c) The 3D assembled structure of the graphene earphone. (d) A pair of graphene earphones. Reprinted with permission

The acoustic properties of the graphene earphone are tested on the measurement system, as shown in **Figure 14a**. The test results indicate that the sound pressure level of the frequency spectrum is flatter than that of commercial earphones, especially in the ultrasonic band. Moreover, the frequency spectrum of the graphene earphone covers not only the audible band (from 20 Hz to 20 kHz for human being) but also the ultrasonic band (from 20 to 50 kHz), which is the audibility zone for some animals. This implies that the graphene earphone could be a

from ACS Nano 8:5883. Copyright (2014) American Chemical Society [12].

72 Advances in Carbon Nanostructures

from ACS Nano 8:5883. Copyright (2014) American Chemical Society [12].

Strain sensors are widely used in various application areas, but most of them are based on rigid substrates, which limit the sensitivity of the sensor. Using laser-scribing technology, a flexible graphene strain sensor [13] is achieved with greatly enhanced sensitivity. The sensi‐ tivity (GR) can be expressed by the following equation:

$$\text{GF} = (\Delta R / R) / \text{\AA} \tag{1}$$

Where Δ*R/R* is the relative resistance change, and *ε* is the applied strain.

The pattern of the graphene strain sensor array could be fabricated on the GO film via laserscribing technology on a flexible PET substrate with low cost and high speed (**Figure 16**). The laser-scribed graphene is made of a loose stack of graphene layers, which could greatly enhance the sensitivity of the graphene strain sensor.

**Figure 16.** Graphene strain sensor. (a) SEM image of the GO surface. (b) LSG surface showing the direction of the laser movement. (c) Wafer-scale strain sensor with different shapes. (d) One strain sensor image. (e) Strain sensor showing good flexibility. (f) Cross-section image of the LSG film. (g) Array of graphene strain sensors. (h) TEM image showing the single-layer graphene. Reprinted from Ref. [13] with permission of The Royal Society of Chemistry.

The test results of the graphene strain sensor are shown in **Figure 17**. The sheet resistance was recorded while stretching the sensor. The sensitivity calculated from the data is 9.49, and **Figure 17f** indicates the rapid response of the graphene strain sensor. Commercial strain sensors based on metal have a gauge factor of 2∼5. Our LSG strain sensor has a much higher sensitivity as compared to commercially available strain sensors, which enables it to have a wide variety of applications. The working principle can be explained as follows: LSG contains a lot of graphene sheets. The conductivity of LSG depends on the overlapping area of the graphene sheets. When the LSG is stretched, the overlapping area of graphene sheets decreas‐ es, which increases the resistance. Therefore, by measuring the resistance change, one can know the strain applied to the LSG.

**Figure 17.** Test results of the graphene strain sensor. (a) Graphene micro-ribbon sensor in its original state. (b) Gra‐ phene micro-ribbon sensor in its stretched state. (c) Optical image of the graphene micro-ribbon. (d) Resistance *versus* length. (e) Relative resistance change *versus* strain. (f) Dynamic response to instantaneous force. Reproduced from Ref. [13] with permission of The Royal Society of Chemistry.

#### **3.4. Graphene pressure sensor**

flexible graphene strain sensor [13] is achieved with greatly enhanced sensitivity. The sensi‐

The pattern of the graphene strain sensor array could be fabricated on the GO film via laserscribing technology on a flexible PET substrate with low cost and high speed (**Figure 16**). The laser-scribed graphene is made of a loose stack of graphene layers, which could greatly enhance

**Figure 16.** Graphene strain sensor. (a) SEM image of the GO surface. (b) LSG surface showing the direction of the laser movement. (c) Wafer-scale strain sensor with different shapes. (d) One strain sensor image. (e) Strain sensor showing good flexibility. (f) Cross-section image of the LSG film. (g) Array of graphene strain sensors. (h) TEM image showing

The test results of the graphene strain sensor are shown in **Figure 17**. The sheet resistance was recorded while stretching the sensor. The sensitivity calculated from the data is 9.49, and **Figure 17f** indicates the rapid response of the graphene strain sensor. Commercial strain

the single-layer graphene. Reprinted from Ref. [13] with permission of The Royal Society of Chemistry.

e

(1)

GF ( / ) / = D*R R*

Where Δ*R/R* is the relative resistance change, and *ε* is the applied strain.

tivity (GR) can be expressed by the following equation:

74 Advances in Carbon Nanostructures

the sensitivity of the graphene strain sensor.

Graphene pressure sensors can also be obtained by laser-scribing technology [14]. Conven‐ tional pressure sensors only have good sensitivity at pressure ranges lower than 5 Pa. How‐ ever, the sensitivity will obviously decrease when the pressure is larger than 5 Pa. It is thus necessary to develop pressure sensors with high sensitivity over a wide range of pressures.

As shown in **Figure 18**, the core device structure is a stacking over of the two LSG layers with the patterns oriented perpendicularly to one another. The upper layer is oriented in the *y*direction, and the lower layer is oriented in the *x*-direction. The profile of the LSG has a "V" shape with a height of 10.7 μm and a width of 19.8 μm. The fabrication process is illustrated in **Figure 19**, and the complete process details are provided in Ref. [14].

**Figure 18.** Device structure of LSG pressure sensor (the inset shows the real device in hand). Reprinted from Ref. [14] (CC BY 4.0).

**Figure 19.** Fabrication process for the LSG pressure sensor. Reprinted from Ref. [14] (CC BY 4.0).

The working principle of the LSG pressure sensor can be described as follows (**Figure 20**): when a force is applied on top, the inter layer distance of the LSG is reduced, which can change the resistance of the LSG. Since LSG has a porous structure, the force can not only increase the contact area but also increase the density of the LSG. Both these effects can increase the current pathway. When the pressure is released, the LSG foam releases and the interlayer distance recovers, which can lower the current to its initial level.

As shown in **Figure 18**, the core device structure is a stacking over of the two LSG layers with the patterns oriented perpendicularly to one another. The upper layer is oriented in the *y*direction, and the lower layer is oriented in the *x*-direction. The profile of the LSG has a "V" shape with a height of 10.7 μm and a width of 19.8 μm. The fabrication process is illustrated

**Figure 18.** Device structure of LSG pressure sensor (the inset shows the real device in hand). Reprinted from Ref. [14]

**Figure 19.** Fabrication process for the LSG pressure sensor. Reprinted from Ref. [14] (CC BY 4.0).

The working principle of the LSG pressure sensor can be described as follows (**Figure 20**): when a force is applied on top, the inter layer distance of the LSG is reduced, which can change the resistance of the LSG. Since LSG has a porous structure, the force can not only increase the contact area but also increase the density of the LSG. Both these effects can increase the current

(CC BY 4.0).

76 Advances in Carbon Nanostructures

in **Figure 19**, and the complete process details are provided in Ref. [14].

**Figure 20.** The working principle of the LSG pressure sensor. Reprinted from Ref. [14] (CC BY 4.0).

In order to test the LSG pressure sensor, we have set up a force testing system. This system contains a motor, a pressure sensor and a resistance analyser. In this system, a static force up to 113 kPa can be applied and the resistance analyser can acquire the resistance from the LSG pressure sensor simultaneously. As shown in **Figure 21**, if the pressure sensing range is up to 50 kPa, the sensitivity can be as high as 0.96 kPa−1. This indicates that the contact area and density obviously change in this pressure range. When the pressure is larger than 50 kPa, the pressure sensor starts to saturate with a sensitivity of 0.005 kPa−1. This demonstrates that we have successfully obtained a high sensitivity pressure sensor using laser-scribing technology.

**Figure 21.** Testing results of the LSG pressure sensor. Reprinted from Ref. [14] (CC BY 4.0).

#### **3.5. Graphene light-emitting device**

In conventional light-emitting devices, the emission wavelength cannot be changed after the devices are fabricated. This is because the bandgap of the material cannot be modified. Previously, there was no report on light-emitting devices with tunable wavelength in a wide range. Here, we obtained a wavelength tunable graphene light-emitting device based on laserscribing technology [15].

**Figure 22.** Device structure of the graphene light-emitting device. The semi-rGO layer functions as the light-emitting layer. Reprinted from Ref. [15] (CC BY 4.0).

As shown in **Figure 22**, a graphene in-plane transistor can be fabricated, based on laser-scribing technology. An in-plane transistor means that the gate, source and drain are all in the same plane. The gate electrical field can be applied to channel through the in-plane GO dielectric. All the components, namely the channel, source, drain and gate can be fabricated in a single laser-scribing step. After the device fabrication, a large current is applied to the channel (mA), which can burn the upper rGO away and expose the interface layer consisting of semi-rGO. A striking feature is that the whole fabrication process does not need vacuum or high tempera‐ ture.

**Figure 23.** Flexible graphene light-emitting device with red light emission. Reprinted from Ref. [15] (CC BY 4.0).

When the device is driven by a voltage of 10 V, there is a noticeable red light emission (**Figure 23**). Moreover, as it is directly fabricated on a PET substrate, the whole device is flexible. **Figure 24** shows the light emission spectra under different gate voltages. Each curve shows a single Lorentzian shape. When the gate voltage is zero, the light wavelength is around 690 nm. When the gate voltage is 50 V, the wavelength shifts to 470 nm. This indicates that increasing the gate voltage can shift the wavelength to a lower value. Moreover, the blue light emission efficiency is much lower than the red light efficiency from the intensity shown in the spectrum. Under different gate voltages, the selected semi-rGO is excited to emit light. The light emission mechanism can be explained by the Poole–Frenkel effect. Under a strong electrical field, the electrons can be excited by impact ionization effect and recombined with the holes. The extra energy can be released by emission of photons.

**Figure 24.** Different light emission spectra, obtained by tuning the gate voltage. Reprinted from Ref. [15] (CC BY 4.0).

#### **3.6. Wafer-scale graphene devices integration**

**3.5. Graphene light-emitting device**

scribing technology [15].

78 Advances in Carbon Nanostructures

layer. Reprinted from Ref. [15] (CC BY 4.0).

ture.

In conventional light-emitting devices, the emission wavelength cannot be changed after the devices are fabricated. This is because the bandgap of the material cannot be modified. Previously, there was no report on light-emitting devices with tunable wavelength in a wide range. Here, we obtained a wavelength tunable graphene light-emitting device based on laser-

**Figure 22.** Device structure of the graphene light-emitting device. The semi-rGO layer functions as the light-emitting

As shown in **Figure 22**, a graphene in-plane transistor can be fabricated, based on laser-scribing technology. An in-plane transistor means that the gate, source and drain are all in the same plane. The gate electrical field can be applied to channel through the in-plane GO dielectric. All the components, namely the channel, source, drain and gate can be fabricated in a single laser-scribing step. After the device fabrication, a large current is applied to the channel (mA), which can burn the upper rGO away and expose the interface layer consisting of semi-rGO. A striking feature is that the whole fabrication process does not need vacuum or high tempera‐

**Figure 23.** Flexible graphene light-emitting device with red light emission. Reprinted from Ref. [15] (CC BY 4.0).

Since a variety of graphene devices can be easily patterned onto the GO film via the laserscribing technology, it is possible to integrate multifunctional graphene devices [16] over an entire wafer (**Figure 25**). As shown in **Figure 25a**, a graphene transistor, photodetector and loudspeaker were integrated on the same chip. Furthermore, the substrates could even be flexible (**Figure 25b**). The patterning process of a whole wafer was finished in 25 min. This development brings our field one-step closer to the concept of heterogeneous integration, and it would be useful to investigate this technique and platform in further detail.

**Figure 25.** Wafer-scale graphene devices integration. (a) Multifunctional graphene devices containing transistors, pho‐ todetectors and loudspeakers. (b) Flexible graphene devices. (c) Wafer-scale graphene devices. Reprinted by permis‐ sion from Macmillan Publishers Ltd: Scientific Reports (4:3598), copyright (2014) [16].

## **4. Outlook: ion scribing for graphene oxide reduction**

The outstanding success of laser scribing for rGO-based device fabrication naturally leads one to imagine whether other serial writing techniques can be used as well. Besides laser lithog‐ raphy, other serial writing techniques include electron beam lithography, ion beam lithogra‐ phy, proton beam lithography and X-ray (or synchrotron) lithography. A review of literature reveals that reduction of GO has already been briefly investigated with electron beam lithography [17] and ion beam lithography [18]. Although the conversion efficiency of electron beam lithography for graphene oxide reduction is very low, and hence impractical, ion beam lithography is very promising. Lobo et al. [18] have investigated the morphology, composition and conductivity changes and confirmed the reduction of GO by Gallium ions. We have also begun investigations using ion beam lithography for GO reduction and device fabrication, and we observed that the conductivity of 300 nm GO films on SiO2/Si conducts up to 1.8 × 10−10 A (0.18 pA) for a Ga+ ion dose of 750 μC/cm2 , as shown in **Figure 26** [19].

**Figure 26.** GO reduction by Ga+ ion beams showing (a) the exposed and unexposed areas, and (b) the increase in con‐ ductivity at a dose of 750 μC/cm2 , as compared with unexposed GO.

#### **5. Summary**

In this chapter, the fabrication of wafer-scale graphene devices by one-step laser-scribing technology is demonstrated. Five kinds of novel graphene devices were developed, including a memory, an earphone, a strain sensor, a pressure sensor and a light-emitting device. The graphene resistive memory has a Fin-like structure with forming-free behaviour, stable switching, reasonable reliability and potential for 2-bit storage. The graphene earphone enables wide-band sound generation from 100 Hz to 50 kHz, which can be used for both humans and animals. The strain sensor based on graphene micro-ribbons has a gauge factor up to 9.49. The sensitivity of the graphene pressure sensor is as high as 0.96 kPa−1 in a wide pressure range (0∼50 kPa). A graphene light-emitting device is also developed with a tuneable emission wavelength. These results demonstrate that laser-scribing technology can be used as a platform to develop novel graphene devices. As a serial lithography technique, at the forefront of controlled graphene oxide reduction, laser scribing serves as a model for similar developments in other patterning techniques, such as ion beam lithography.

## **Author details**

**4. Outlook: ion scribing for graphene oxide reduction**

ion dose of 750 μC/cm2

, as compared with unexposed GO.

developments in other patterning techniques, such as ion beam lithography.

In this chapter, the fabrication of wafer-scale graphene devices by one-step laser-scribing technology is demonstrated. Five kinds of novel graphene devices were developed, including a memory, an earphone, a strain sensor, a pressure sensor and a light-emitting device. The graphene resistive memory has a Fin-like structure with forming-free behaviour, stable switching, reasonable reliability and potential for 2-bit storage. The graphene earphone enables wide-band sound generation from 100 Hz to 50 kHz, which can be used for both humans and animals. The strain sensor based on graphene micro-ribbons has a gauge factor up to 9.49. The sensitivity of the graphene pressure sensor is as high as 0.96 kPa−1 in a wide pressure range (0∼50 kPa). A graphene light-emitting device is also developed with a tuneable emission wavelength. These results demonstrate that laser-scribing technology can be used as a platform to develop novel graphene devices. As a serial lithography technique, at the forefront of controlled graphene oxide reduction, laser scribing serves as a model for similar

(0.18 pA) for a Ga+

80 Advances in Carbon Nanostructures

**Figure 26.** GO reduction by Ga+

ductivity at a dose of 750 μC/cm2

**5. Summary**

The outstanding success of laser scribing for rGO-based device fabrication naturally leads one to imagine whether other serial writing techniques can be used as well. Besides laser lithog‐ raphy, other serial writing techniques include electron beam lithography, ion beam lithogra‐ phy, proton beam lithography and X-ray (or synchrotron) lithography. A review of literature reveals that reduction of GO has already been briefly investigated with electron beam lithography [17] and ion beam lithography [18]. Although the conversion efficiency of electron beam lithography for graphene oxide reduction is very low, and hence impractical, ion beam lithography is very promising. Lobo et al. [18] have investigated the morphology, composition and conductivity changes and confirmed the reduction of GO by Gallium ions. We have also begun investigations using ion beam lithography for GO reduction and device fabrication, and we observed that the conductivity of 300 nm GO films on SiO2/Si conducts up to 1.8 × 10−10 A

, as shown in **Figure 26** [19].

ion beams showing (a) the exposed and unexposed areas, and (b) the increase in con‐

He Tian1,2, Mohammad Ali Mohammad1,3, Wen-Tian Mi1 , Yi Yang1 and Tian-Ling Ren1\*

\*Address all correspondence to: RenTL@tsinghua.edu.cn

1 Institute of Microelectronics and Tsinghua National Laboratory for Information Science and Technology (TNList), Tsinghua University, Beijing, China

2 Ming Hsieh Department of Electrical Engineering, University of Southern California, Los Angeles, CA, United States

3 School of Chemical and Materials Engineering (SCME), National University of Sciences and Technology (NUST), Sector H-12, Islamabad, Pakistan

## **References**


## **Chapter 4**

## **Light Emission from Graphene**

Young Duck Kim and Myung‐Ho Bae

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64051

#### **Abstract**

[9] El-Kady MF, Strong V, Dubin S, Kaner RB. Laser scribing of high-performance and flexible graphene-based electrochemical capacitors. Science. 2012;335(6074):1326-1330.

[10] El-Kady MF, Kaner RB. Scalable fabrication of high-power graphene micro-super capacitors for flexible and on-chip energy storage. Nature Communications. 2013;12(4):

[11] Tian H, Chen HY, Ren TL, Li C, Xue QT, Mohammad MA, Wu C, Yang Y, Wong HS. Cost-effective, transfer-free, flexible resistive random access memory using laserscribed reduced graphene oxide patterning technology. Nano Letters. 2014;14(6):3214–

[12] Tian H, Li C, Mohammad MA, Cui YL, Mi WT, Yang Y, Xie D, Ren TL. Graphene earphones: entertainment for both humans and animals. ACS Nano. 2014;8(6):

[13] Tian H, Shu Y, Cui YL, Mi WT, Yang Y, Xie D, Ren TL. Scalable fabrication of highperformance and flexible graphene strain sensors. Nanoscale. 2014;6(2):699-705. DOI:

[14] Tian H, Shu Y, Wang XF, Mohammad MA, Bie Z, Xie QY, Li C, Mi WT, Yang Y, Ren TL. A graphene-based resistive pressure sensor with record-high sensitivity in a wide

[15] Wang X, Tian H, Mohammad MA, Li C, Wu C, Yang Y, Ren TL. A spectrally tunable all-graphene-based flexible field-effect light-emitting device. Nature Communications.

[16] Tian H, Yang Y, Xie D, Cui YL, Mi WT, Zhang Y, Ren TL. Wafer-scale integration of graphene-based electronic, optoelectronic and electroacoustic devices. Scientific

[17] Chen L, Xu Z, Li J, Min C, Liu L, Song X, Chen G, Meng X. Reduction and disorder in graphene oxide induced by electron-beam irradiation, Materials Letters. 2011;65:1229.

[18] Lobo DE, Fu J, Gengenbach T, Majumder M. Localized deoxygenation and direct patterning of graphene oxide films by focused ion beams. Langmuir. 2012;28:14815.

[19] Mohammad MA, Mi WT, Li YX, Tian H, Yang Y, Ren TL. Manuscript in preparation

pressure range. Scientific Reports. 2015;5:8603. DOI: 10.1038/srep08603

DOI: 10.1126/science.1216744

82 Advances in Carbon Nanostructures

1475. DOI: 10.1038/ncomms2446

3219. DOI: 10.1021/nl5005916

10.1039/C3NR04521H

5883-5890. DOI: 10.1021/nn5009353

2015;6:7767. DOI: 10.1038/ncomms8767

DOI: 10.1016/j.matlet.2011.01.063

DOI: 10.1021/la303369m

(2016).

Reports. 2014;4:3598. DOI: 10.1038/srep03598

Graphene has attracted great interest due to its superior and unique physical proper‐ ties. The strong light‐matter interaction with ultrafast and broadband photoresponse of graphene has allowed for the development of state‐of‐the art optical components, such as photodetectors and optical modulators, making this material very promising for ultrafast optical communications. However, so far, light emission from graphene has remained elusive, although this would open the door towards obtaining atomically thin, flexible and transparent light sources and graphene‐based on‐chip interconnects. In this chapter, we review experimental results and techniques of the electrically driven light emission from graphene in the infrared and visible spectrum range.

**Keywords:** Graphene, light emission, power dissipation, high electric field, Phonon temperature

## **1. Introduction**

In order to obtain bright thermal radiation from gapless graphene, the radiative electron–hole recombination process is not enough due to the ultrafast energy relaxation that occurs through electron–electron ( ∼10 fs) and electron–phonon ( ∼1 ps) interactions. On the other hand, graphene's superior mechanical strength, high‐temperature stability and non‐equilibrium electron temperature, compared to lattice temperature, may enable the efficient thermal radiation under high electric field. However, the thermal radiation from electrically biased graphene is significantly influenced by the substrate. Especially, dominant vertical heat dissipation through underlying substrate, and extrinsic scattering effects, such as charged impurities and the surface polar optical phonon, limit the thermal radiation efficiency of graphene.

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Freely suspended structure enables the dramatic suppression of undesirable vertical heat dissipation and extrinsic scattering effects by substrate, and much more promising to yield efficient and brighter thermal radiation from graphene in the infrared to visible range. Furthermore, the emitted light from suspended graphene interacts with the reflected light from the separated substrate surface, resulting in interference effects that can be used to tune the radiation spectrum of this material. Electro‐thermal engineering will allow obtaining ad‐ vanced graphene light emitters for atomically thin, flexible and transparent lighting and ultrafast optical interconnects. In this chapter, we will provide an overview of the thermal radiation from electrically biased graphenes in the infrared and visible spectrum range.

### **2. High electric field transport in graphene on substrate**

Graphene's superior electronic [1], mechanical [2] and thermal [3] properties can be used for high mobility field‐effect transistors (FETs) [4] and mechanical resonators [5]. In the case of electronic devices, as shown in **Figure 1a**, that depicts a graphene FET, the applied electric field applied by the source-drain voltage generates a Joule self‐heating effect, which leads to a change of its electronic transport behaviour or device failure, in some cases [6]. For carbon nanotubes (CNTs), the observed electronic transport characteristics have been explained by heat transports along the CNT channels and between CNTs and the substrates [7]. There have been reports on the heat dissipation from graphene devise in a diffusive regime, where the chemical potential changes along the graphene channel. A direct method to study the heat dissipation is to accurately measure the temperature during device operation (see Sections 2.1– 2.3). The temperature measurement of graphene on a substrate biased by a source–drain voltage has been performed in four ways: (1) shift of Raman peaks [8–11]; (2) thermal scanning imaging [11]; (3) thermal radiation [9, 10, 12] and (4) ratio of the Stoke/anti‐Stoke intensities in Raman spectroscopy [9, 10, 13]. Methods (1) and (2) are sensitive to the acoustic phonon (AP) temperature (*Tap*), while methods (3) and (4) are linked to the electron and optical phonon (OP) temperatures (*Te* and *Top*), respectively.

**Figure 1.** (a) Graphene field‐effect transistor. S, D and G are source, drain and gate, respectively. *L* and *W* are length and width of graphene channel, respectively, and *tox* is thickness of SiO2 layer [14]. (b) Thermal model for a graphene on SiO2/Si substrate. *Rox* and *RSi* are thermal resistance of the oxide and silicon substrate, respectively. *Ksi* is thermal conductivity of Si substrate.

The Fourier's law for thermal conductance is:

Freely suspended structure enables the dramatic suppression of undesirable vertical heat dissipation and extrinsic scattering effects by substrate, and much more promising to yield efficient and brighter thermal radiation from graphene in the infrared to visible range. Furthermore, the emitted light from suspended graphene interacts with the reflected light from the separated substrate surface, resulting in interference effects that can be used to tune the radiation spectrum of this material. Electro‐thermal engineering will allow obtaining ad‐ vanced graphene light emitters for atomically thin, flexible and transparent lighting and ultrafast optical interconnects. In this chapter, we will provide an overview of the thermal radiation from electrically biased graphenes in the infrared and visible spectrum range.

Graphene's superior electronic [1], mechanical [2] and thermal [3] properties can be used for high mobility field‐effect transistors (FETs) [4] and mechanical resonators [5]. In the case of electronic devices, as shown in **Figure 1a**, that depicts a graphene FET, the applied electric field applied by the source-drain voltage generates a Joule self‐heating effect, which leads to a change of its electronic transport behaviour or device failure, in some cases [6]. For carbon nanotubes (CNTs), the observed electronic transport characteristics have been explained by heat transports along the CNT channels and between CNTs and the substrates [7]. There have been reports on the heat dissipation from graphene devise in a diffusive regime, where the chemical potential changes along the graphene channel. A direct method to study the heat dissipation is to accurately measure the temperature during device operation (see Sections 2.1– 2.3). The temperature measurement of graphene on a substrate biased by a source–drain voltage has been performed in four ways: (1) shift of Raman peaks [8–11]; (2) thermal scanning imaging [11]; (3) thermal radiation [9, 10, 12] and (4) ratio of the Stoke/anti‐Stoke intensities in Raman spectroscopy [9, 10, 13]. Methods (1) and (2) are sensitive to the acoustic phonon (AP) temperature (*Tap*), while methods (3) and (4) are linked to the electron and optical phonon (OP)

**Figure 1.** (a) Graphene field‐effect transistor. S, D and G are source, drain and gate, respectively. *L* and *W* are length and width of graphene channel, respectively, and *tox* is thickness of SiO2 layer [14]. (b) Thermal model for a graphene on SiO2/Si substrate. *Rox* and *RSi* are thermal resistance of the oxide and silicon substrate, respectively. *Ksi* is thermal

**2. High electric field transport in graphene on substrate**

temperatures (*Te* and *Top*), respectively.

84 Advances in Carbon Nanostructures

conductivity of Si substrate.

$$
\dot{\underline{Q}} = G\_{\text{th}} \Delta T,\tag{1}
$$

where *Q*˙ is the amount of heat transferred per unit time, Δ*T* is the temperature difference between two points of interest, and *Gth* is the thermal conductance. Here, the temperature corresponds to *Tap* since the acoustic phonons are the main contributors to heat conductions. **Figure 1b** shows how to apply the Fourier's law to graphene devices on a substrate, where *Rox* and *RSi* are the thermal resistances of the SiO2 layer and Si substrate, respectively. The spread‐ ing thermal resistance of the Si substrate is given by *RSi* ∼ 1/[2*ksi*(*LW*) 1/2] and *Rox* ∼ *tox*/(*koxLW*), where *kox*= 1.3 Wm-1K-1 and *ksi*= 50 Wm-1k-1 are the thermal conductivities of the SiO2 layer and Si substrate at room temperature, respectively [12]. *tox* is the thickness of SiO2 layer, and *L* (*W*) is the length (width) of the graphene channel. Eq. (1) can also be expressed as *ΔT* =*Rth P*, where *Rth* is the total thermal resistance, and *P* is the electrical power.

**Figure 2** shows a summary of temperature increase (ΔT) of graphene devices on substrates, as a function of electrical power density, determined by various temperature measurement techniques. The closed, opened and half‐closed scattered points show the acoustic, *Tap*, *Top* and *Te* and electron temperatures, respectively.

**Figure 2.** averaged temperature increase (ΔT) as a function of applied power per unit area of reported devices. De‐ tailed information of each device is described in **Table 1**.

The detailed measurement techniques can be found in the legend of **Figure 2**. For each sample, one can obtain the *Rth* of the system, through the slope data plotted in **Figure 2**. **Table 1** and **Figure 3** show the *Rth* comparison between the experiments and calculations, based on the relation: *Rth* ≈*Rox* + *RSi* (see **Figures 1b** and **3**).


**Table 1.** Summary of reported the thermal resistance of SiO2 layer (*Rox*), and Si substrate (*Rsi*) with calculated (*Rcal*) and experimentally measured (*Rexp*) total thermal resistance, where all samples are placed on SiO2/Si substrates.

**Figure 3.** Average calculated (*Rcal*) and measured (*Rexp*) total thermal resistance of each reference, where Rcal and Rexp were adopted from **Table 1** for corresponding references.

In order to treat the heat transfer from graphene on a substrate into the surrounding environ‐ ment, as shown in **Figure 1b**, let us first consider a one‐dimensional heat equation along the graphene channel [12]:

$$\frac{d^2T}{d\mathbf{x}^2} + p = \frac{\mathbf{g}}{k\_gA}(T - T\_0),\tag{2}$$

where *g* is the thermal conductance of the substrate per unit length, *kg* is the thermal conduc‐ tivity of graphene, *A* is the cross‐section of graphene, *p* = *I Fx*, *I* is the current and *Fx* is the electric field at *x*. For simplicity, if we consider an infinitely long graphene channel without bias current, Eq. (2) has a solution, *Θ* =*Θ*(0)*e* <sup>−</sup>*x*/*<sup>L</sup> <sup>H</sup>* , where *Θ*(*x*)=*<sup>T</sup>* (*x*)−*T*0 and *<sup>L</sup> <sup>H</sup>* <sup>=</sup> *kgA* / *<sup>g</sup>*. The *LH* is the lateral temperature diffusion (healing) length from a heat source through the graphene. The heat in a graphene device on a substrate is dissipated into a heat sink (*T* = *T0*) mainly in two means: a direct way through the substrate underneath graphene or indirect way via metal contacts. In a graphene device where *L* is comparable to *LH*, the heat dissipation through the metal contacts becomes important. In the opposite case of L >> LH, the heat flow through the substrate is the dominant heat dissipation path. *g* is approximately given by 1/[*L*(*Rox*+*RSi*)] (also see **Figure 1b**). For a monolayer graphene device on 300 nm thick, SiO2/Si substrate with *L* =  25 μm, *W* = 6 μm and graphene thickness *tg* = 0.35 nm, we get *<sup>L</sup> <sup>H</sup>* <sup>≈</sup> *toxtg kg kox* ≈0.2 μm [12]. Here, we used the thermal conductivity of a few μm size graphene on a substrate as *kg* ∼ 600 Wm-1K-1 [15]. Thus, in general, graphene devices of few micrometres in length directly dissipate their heat through the substrate underneath graphene, and the temperature profile of the graphene reflects the heat generation profile.

#### **2.1. Electron temperature of graphene**

**References** *L* **[μm]** *W* **[μm]** *tox* **[nm]** *Rox* **[K/W]** *RSi* **[K/W]** *Rcal* **[K/W]** *Rexp* **[K/W]** [8] 2.65 1.45 300 60,000 5100 65,100 73,500 [12] 28 6 300 1370 770 2140 2100 [11] 7 4 300 8240 1900 10,130 8300 [10] 4.15 1.45 300 38,350 4070 42,430 68,100 [9] 3.6 1.6 280 37,400 4160 41,560 69,800

**Table 1.** Summary of reported the thermal resistance of SiO2 layer (*Rox*), and Si substrate (*Rsi*) with calculated (*Rcal*) and

**Figure 3.** Average calculated (*Rcal*) and measured (*Rexp*) total thermal resistance of each reference, where Rcal and Rexp

In order to treat the heat transfer from graphene on a substrate into the surrounding environ‐ ment, as shown in **Figure 1b**, let us first consider a one‐dimensional heat equation along the

> <sup>2</sup> <sup>0</sup> ( ), *g dT g p TT dx k A*

where *g* is the thermal conductance of the substrate per unit length, *kg* is the thermal conduc‐ tivity of graphene, *A* is the cross‐section of graphene, *p* = *I Fx*, *I* is the current and *Fx* is the electric field at *x*. For simplicity, if we consider an infinitely long graphene channel without bias

is the lateral temperature diffusion (healing) length from a heat source through the graphene. The heat in a graphene device on a substrate is dissipated into a heat sink (*T* = *T0*) mainly in two means: a direct way through the substrate underneath graphene or indirect way via metal contacts. In a graphene device where *L* is comparable to *LH*, the heat dissipation through the metal contacts becomes important. In the opposite case of L >> LH, the heat flow through the substrate is the dominant heat dissipation path. *g* is approximately given by 1/[*L*(*Rox*+*RSi*)] (also see **Figure 1b**). For a monolayer graphene device on 300 nm thick, SiO2/Si substrate with *L* = 

+= - (2)

<sup>−</sup>*x*/*<sup>L</sup> <sup>H</sup>* , where *Θ*(*x*)=*<sup>T</sup>* (*x*)−*T*0 and *<sup>L</sup> <sup>H</sup>* <sup>=</sup> *kgA* / *<sup>g</sup>*. The *LH*

2

were adopted from **Table 1** for corresponding references.

current, Eq. (2) has a solution, *Θ* =*Θ*(0)*e*

graphene channel [12]:

86 Advances in Carbon Nanostructures

experimentally measured (*Rexp*) total thermal resistance, where all samples are placed on SiO2/Si substrates.

Thermal radiation can be observed for graphene devices on a substrate under high electric field, which is attributed to the electrically driven spontaneous emission from hot electrons of graphene [9, 10]. The thermalized charge carriers under electric field emit electromagnetic radiation, according to Planck's law, as a grey body:

$$I\left(E, T\_\circ\right) = \varepsilon\left(E\right)\frac{2E^3}{h^2C^2}\left[\exp\left(\frac{E}{k\_BT\_\circ}\right) - 1\right]^{-1},\tag{3}$$

where *I(E,Te)* is the spectral radiance, *ε*(*E*) is the emissivity of graphene, *E* is the energy of the emitted photon, *h* is Planck's constant, *c* is the speed of light, and *kB* is the Boltzmann constant.

**Figure 4.** (a) Thermal radiation spectrum from graphene on substrate under applied electrical power per unit area (symbols). The radiation spectra are well fitted to Planck's law (solid line) and electron temperatures are estimated. (b) Anti‐stoke and (c) Stokes Raman G mode as function of applied electric field and optical phonon temperature are indi‐ cated. (d) Electron and optical phonon temperature as function of the dissipated electrical power. It indicates the equi‐ librium of electron and optical phonon temperature. (e) Raman G mode frequency (symbol) features as function of optical phonon temperature. Solid line is the theoretically predicted temperature dependences under phonon equilibri‐ um. Dashed line is linearly extrapolation (from Ref. [9]).

The *Te*, extracted from its thermal radiation spectrum in the range of 1.3–2.6 eV, was of 1000– 2000 K [9], as shown in **Figure 4a**. From the extracted *Te*, we can assume that the strong electron‐ electron interaction rapidly thermalizes electrons and holes, and attains hot charge carriers by the applied electric field. Furthermore, the experimentally measured emissivity *ε*(*E*) of graphene of =1.6 ± 0.8 % is consistent with the measured absorptivity *α* =2.3 % for monolayer graphene [9, 10]. This means that, for graphene under high electric field, Kirchhoff's law is valid in the measured energy range.

#### **2.2. Optical phonon temperature of graphene**

Raman spectroscopy is a well‐known technique for the characterization of a number of layers, doping level, mechanical strain and temperature of graphene. In particular, for graphene devices under high electric field, Raman spectroscopy accurately provides the phonon temperature and population. In typical Raman spectroscopy, spontaneous Raman effects are due to the inelastic light scattering, with optical phonon modes of the material being deter‐ mined by measurement. The resulting inelastic scattered of excited photon emits or absorbs optical phonons. The Stoke process involves the creation of extra optical phonons (*n*+1), and the anti‐Stoke process is the annihilation of optical phonons (*n*).

In graphene, the phonon population for zone centre optical phonons can be estimated from the G peak signal intensity of the Raman spectra, which is proportional to phonon populations with effective temperature, and follows from the Bose‐Einstein distribution. Raman signal intensity ratio of the anti‐Stoke (*IAS*) and Stokes (*IS*) process is as follows:

$$\frac{I\_{\rm AS}}{I\_{\rm S}} \approx \exp(-\frac{\hbar o\_G}{k\_B T\_{op}}),\tag{4}$$

where ℏ*ωG* is the energy of G phonon. Experimentally, *Top* is extracted from the ratio *IAS*/*IS* as shown in **Figure 4b** and **c**. Due to the high optical phonon energy of graphene, anti‐Stoke Raman signals can be observed only above 440 K [9, 10, 13].

#### **2.3. Acoustic phonon temperature of graphene**

The Raman spectrum of graphene also exhibits a downshift of the Raman peak position with increasing temperature. This is due to the anharmonic coupling effect and anharmonic temperature of secondary phonons into G mode phonon decay and a thermal expansion effect [16]. Empirically, the down shift of the G peak can be measured as a linear function of *Tap* as:

$$
\alpha\_G = \alpha\_G^0 + \mathcal{Y}T\_{ap},
\tag{5}
$$

where *ω<sup>G</sup>* <sup>0</sup> is the G peak frequency at zero temperature, and *χ* is the coefficient for temperature correction term of -0.016 cm-1 K-1 [17], at moderate temperature.

#### **2.4. Non‐equilibrium phonon population in graphene**

electron interaction rapidly thermalizes electrons and holes, and attains hot charge carriers by the applied electric field. Furthermore, the experimentally measured emissivity *ε*(*E*) of graphene of =1.6 ± 0.8 % is consistent with the measured absorptivity *α* =2.3 % for monolayer graphene [9, 10]. This means that, for graphene under high electric field, Kirchhoff's law is

Raman spectroscopy is a well‐known technique for the characterization of a number of layers, doping level, mechanical strain and temperature of graphene. In particular, for graphene devices under high electric field, Raman spectroscopy accurately provides the phonon temperature and population. In typical Raman spectroscopy, spontaneous Raman effects are due to the inelastic light scattering, with optical phonon modes of the material being deter‐ mined by measurement. The resulting inelastic scattered of excited photon emits or absorbs optical phonons. The Stoke process involves the creation of extra optical phonons (*n*+1), and

In graphene, the phonon population for zone centre optical phonons can be estimated from the G peak signal intensity of the Raman spectra, which is proportional to phonon populations with effective temperature, and follows from the Bose‐Einstein distribution. Raman signal

> exp( ), *AS <sup>G</sup> S B op*

» - <sup>h</sup>

w

where ℏ*ωG* is the energy of G phonon. Experimentally, *Top* is extracted from the ratio *IAS*/*IS* as shown in **Figure 4b** and **c**. Due to the high optical phonon energy of graphene, anti‐Stoke

The Raman spectrum of graphene also exhibits a downshift of the Raman peak position with increasing temperature. This is due to the anharmonic coupling effect and anharmonic temperature of secondary phonons into G mode phonon decay and a thermal expansion effect [16]. Empirically, the down shift of the G peak can be measured as a linear function of *Tap* as:

<sup>0</sup> ,

<sup>0</sup> is the G peak frequency at zero temperature, and *χ* is the coefficient for temperature

*G G ap* = + *T* (5)

wwc

(4)

*I kT*

valid in the measured energy range.

88 Advances in Carbon Nanostructures

**2.2. Optical phonon temperature of graphene**

the anti‐Stoke process is the annihilation of optical phonons (*n*).

intensity ratio of the anti‐Stoke (*IAS*) and Stokes (*IS*) process is as follows:

*I*

Raman signals can be observed only above 440 K [9, 10, 13].

correction term of -0.016 cm-1 K-1 [17], at moderate temperature.

**2.3. Acoustic phonon temperature of graphene**

where *ω<sup>G</sup>*

As discussed in Sections 2.1–2.3, accurate measurements of graphene temperature provide a deep understanding of the heat dissipation mechanism and electron‐phonon interaction in biased graphene devices. Direct comparison of *Te* and *Top* temperatures of electrically biased graphene on a substrate shows that the *Te* and *Top* are in equilibrium up to 2000 K, as shown in **Figure 4d**.

In the case of CNTs, non‐equilibrium phonon distributions have been measured when electrically biased, as high‐energy optical phonons have larger populations than low‐energy phonons, such as radial‐breathing mode and acoustic phonons [7, 18].

Similarly to what was found with CNTs, non‐equilibrium phonon populations in electrically biased graphene on a substrate have been observed. **Table 1** and **Figure 3** [9, 10] showed that the elevated *Top* could be estimated by the ratio of Stoke/anti‐Stoke intensity of G phonons and *Te* measured by the thermal emission. For both cases, the experimentally obtained thermal resistance, *Rexp* is much larger than the predicted thermal resistance, *Rcal*, which could be due to the optical phonons and electrons not being in equilibrium with the acoustic phonon. This is consistent with the observation of the temperature deviation between *Top* and the anharmonic temperature of secondary phonons, for which G phonons decay for *Top* > 670 K [9]. From the device case in Ref. [10], the anharmonic temperature extracted from the 2D band showed similar temperatures to the *Top* and *Te* up to *Top* ∼ 720 K. In the reports, however, we should note that the temperature calibration for the 2D band was only obtained for 83 K < *T* < 373 K, while the measurements was performed at 300 K < *T* < 700 K, thus temperature correction could be needed for the high temperature region.

Experimental results show that strong electron‐optical phonon coupling enables the equili‐ brium of the electron and optical phonon temperature, whereas an energy relaxation bottle‐ neck to low‐energy phonons results in the non‐equilibrium phonon population. Furthermore, strong coupling with surface polar optical phonons and substrate acoustic phonons provides an extra electronic energy relaxation path, resulting in the effective cooling of acoustic phonons of graphene, under high electric field. Therefore, non‐equilibrium optical phonon temperature (*Top*) is given by:

$$T\_{ap} = T\_{ap} + \alpha (T\_{ap} - T\_0),\tag{6}$$

where *To* is the ambient temperature environment, and *α* is the non‐equilibrium coefficient. The *α* is defined by [7]:

$$
\alpha = \frac{R\_{op}}{R\_{op}},
\tag{7}
$$

where *Rop* and *Rap* are the thermal resistances related to optical phonon decay into acoustic phonons, and the heat conduction of acoustic phonons into the heat sink, respectively. Thus, the *Tap* and *Top* are given by *Tap* =*T*<sup>0</sup> + *PRap* and *Top* =*Tap* + *PRop*, respectively. For the devices from reference [9, 10], *Rexp* was ∼70,000 K/W while *Rcal* was ∼40,000 K/W (see **Table 1** and **Figure 3**). Since *Rcal* corresponds to *Rap* and *Rexp* ∼ *Rap* + *Rop*, we get *Rop* ∼ 30,000 K/W, resulting in *α* ∼ 0.75.

## **3. Infrared light emission from graphene on substrate**

Graphene has unique electronics and phonon states under high electric field, such as decou‐ pled electrons and low‐energy phonon temperature [9, 13]. These make graphene an efficient material as thermal radiation source, determined only by *Te*. Furthermore, gapless graphene exhibits high performance for optoelectronics such as photodetectors, optical modulator and plasmonic devices in the infrared range. Developing a high‐speed and efficient infrared light source from graphene has a great potential for optical communications, as well as medical applications.

Previous studies of infrared light emission from graphene were accomplished by graphene FETs on a substrate under high electric field [9, 10, 12]. Due to the self‐heating effect of graphene under electrical bias, only a small fraction (<10-6) of energy is converted into light and most of the energy is dissipated into the substrate and metal contact. The extracted *Te* of graphene light sources was reported to be in the range of 700–2000 K [9, 10].

In graphene FETs, charge carrier density profiles are determined by applied electric field and gate voltage. According to the self‐consistent electrical‐thermal model of charge and heat

**Figure 5.** (a) Infrared mapping of temperature profiles along the graphene,showing hotspot formation. The hot spot moves from source to drain, marking the location of minimum charge density and maximum electric field, following the device electrostatics (from Ref. [14]). (b) Imaged temperature map of belayed graphene field‐effect transistor (from Ref. [12])).

transport in graphene, the minimum charge carrier density spatial location is the cross‐point of the Fermi level across the charge neutrality point. The hot spot position of the infrared image directly reveals the spatial location of the charge neutrality point along the graphene channel, as shown in **Figure 5a** and **b**. In some case, the stationary hot spot of infrared light emission, even under variation of gate voltages and source‐drain bias, is attributed to the defects of graphene and trapped charges in the substrate [10, 12, 14].

Spontaneous thermal radiation is governed by the Fermi's golden rule, and the transition rate is determined by the optical density of the environment. This was shown in Ref. [19], through an optical micro cavity controlled graphene infrared light source, with emission around 925 nm,. Due to the strong light‐matter interaction of atomically thin graphene and the optical confinement effect, micro‐cavity structures enable the control of 20‐fold enhancement of photocurrent generation, which is spectrally selective near infrared thermal radiation. The thermal radiation confinement effect inhibits spontaneous thermal radiation wavelengths larger than the resonant wavelength of the optical cavity. The optical cavity‐induced suppres‐ sion of the radiative thermal radiation also leads to the self‐heating effect in graphene, which is affected by non‐radiative heat transfer through the substrate, metal electrodes and radiative heat transfer.

## **4. High electric field transport in suspended graphene**

the *Tap* and *Top* are given by *Tap* =*T*<sup>0</sup> + *PRap* and *Top* =*Tap* + *PRop*, respectively. For the devices from reference [9, 10], *Rexp* was ∼70,000 K/W while *Rcal* was ∼40,000 K/W (see **Table 1** and **Figure 3**). Since *Rcal* corresponds to *Rap* and *Rexp* ∼ *Rap* + *Rop*, we get *Rop* ∼ 30,000 K/W, resulting

Graphene has unique electronics and phonon states under high electric field, such as decou‐ pled electrons and low‐energy phonon temperature [9, 13]. These make graphene an efficient material as thermal radiation source, determined only by *Te*. Furthermore, gapless graphene exhibits high performance for optoelectronics such as photodetectors, optical modulator and plasmonic devices in the infrared range. Developing a high‐speed and efficient infrared light source from graphene has a great potential for optical communications, as well as medical

Previous studies of infrared light emission from graphene were accomplished by graphene FETs on a substrate under high electric field [9, 10, 12]. Due to the self‐heating effect of graphene under electrical bias, only a small fraction (<10-6) of energy is converted into light and most of the energy is dissipated into the substrate and metal contact. The extracted *Te* of graphene light

In graphene FETs, charge carrier density profiles are determined by applied electric field and gate voltage. According to the self‐consistent electrical‐thermal model of charge and heat

**Figure 5.** (a) Infrared mapping of temperature profiles along the graphene,showing hotspot formation. The hot spot moves from source to drain, marking the location of minimum charge density and maximum electric field, following the device electrostatics (from Ref. [14]). (b) Imaged temperature map of belayed graphene field‐effect transistor (from

**3. Infrared light emission from graphene on substrate**

sources was reported to be in the range of 700–2000 K [9, 10].

in *α* ∼ 0.75.

90 Advances in Carbon Nanostructures

applications.

Ref. [12])).

As discussed in the previous section, high electric field transport of graphene on a substrate is affected by the substrate and environmental conditions. Limited thermal radiation efficiency and electron temperature are due to the dominant heat dissipation to the substrate and extrinsic scattering effects, such as electron‐hole puddle, trapped charges [20] and surface polar phonon [21]. According to some authors [22, 23], suspended graphene shows ultrahigh mobility and fractional quantum Hall effect at low temperature, with low electric field. Therefore, in order to observe the intrinsic properties of graphene under high electric field, suspended structures are essential.

In suspended graphene, the self‐heating effect plays a role due to significantly reduced heat dissipation through the substrate. Thermally induced charge carriers, saturation velocity, charge carrier mobility and thermal conductivity of graphene are significantly affected by the applied electric field and temperature.

In the clean limit of suspended graphene, under high electric field as shown in **Figure 6a**–**c**, the total charge carrier density of graphene is dominated by thermally induced charge carriers (*nth*), represented as:

$$m\_{th} = \left(\frac{\pi}{6}\right) \left(\frac{k\_B T}{\hbar \nu\_F}\right)^2,\tag{8}$$

**Figure 6.** (a) Suspended graphene device, where colour scale indicates the temperature under high electric field. (b) SEM image of suspended graphene grown by CVD. (c) Calculated total charge carrier density as function of gate volt‐ age with increasing temperature. In suspended graphene, thermally generated charge carriers are dominant at high temperature and independent of gate voltage. (d) Simulated *I/W* versus *F* with variation of low‐field mobility‐based electro‐thermal self‐consistent simulation (from Ref. [24]).

where *vF* is the Fermi velocity of graphene. The intrinsic carrier density of graphene is *ni* =4(*nth* <sup>2</sup> <sup>+</sup> (*npd* / <sup>2</sup>)<sup>2</sup>), where *npd* is electron‐hole puddle‐induced carriers. The carrier density induced by electrostatic gate *Vg* is *ng* =*Cox*(*Vdr* −*Vg*) /*e*, where *Vdr* is the charge neutral voltage of graphene. In total, the electron (hole) carrier density (see **Figure 6c**) is as follows:

$$m\_e \left( n\_h \right) = \frac{1}{2} \left( \pm n\_g + \sqrt{n\_g^2 + n\_i} \right). \tag{9}$$

The current density *J* in the suspended graphene, with an applied electric field, is presented by the continuity equation:

$$J = e(n\_{\alpha} + n\_{h\iota})\nu\_{d\iota},\tag{10}$$

where *e* is the electron charge, *nex* (*nhx*) is the electron (hole) density, *vd* is the drift velocity along the graphene channel. The drift velocity is expressed by:

$$\mathbf{v}\_{d\boldsymbol{\alpha}} = \frac{\mu\_{\rm x} F\_{\rm x}}{\left[1 + \left(\mu\_{\rm x} F\_{\rm x} / \boldsymbol{\nu}\_{\rm sat}\right)^{\eta}\right]^{1/\eta}},\tag{11}$$

where *μx* is the temperature dependent charge carrier mobility, *Fx* is the electric field along the graphene channel, and *vsat* is a the saturation velocity determined by the intrinsic graphene optical phonon (160 meV) scattering [24, 25].

For a self‐consistent electrical‐thermal calculation of suspended graphene, the heat diffusion equation for the temperature profile (*Tx*) along the graphene channel takes the from:

$$\frac{d^2T\_x}{d\mathbf{x}^2} + \frac{1}{k\_x Wt} \frac{dP\_x}{d\mathbf{x}} - \frac{2\mathbf{g}\_x}{k\_x t} \left(T\_x - T\_0\right) = \mathbf{0},\tag{12}$$

where *Px (= I(Vx - IRc)*) is the local dissipated electrical power, *kx* is the temperature dependent thermal conductivity, *W* is width, *t* is thickness, *gx* is the thermal conductance per unit area (*gx*∼0 in vacuum), and *T0* is ambient temperature. Assuming that the temperature‐dependent charge carrier mobility is considered, (*Te*) =*μ*0(*T*<sup>0</sup> / *Te*)*<sup>β</sup>*, where *μ*0 is the ambient temperature mobility, *Te* is the electron temperature, and *β* varies from 1.5 to 2.5, depending on the kind of sample. Furthermore, thermal conductivity is assumed to be *k*(*Tap*) =*k*0(*T*<sup>0</sup> / *Tap*)*<sup>γ</sup>*, where *k0* is the ambient temperature thermal conductivity, *Tap* is the acoustic phonon temperature of gra‐ phene, and *γ* varies from 1.7 to 1.9 depending on the kind of sample.

**Figure 6.** (a) Suspended graphene device, where colour scale indicates the temperature under high electric field. (b) SEM image of suspended graphene grown by CVD. (c) Calculated total charge carrier density as function of gate volt‐ age with increasing temperature. In suspended graphene, thermally generated charge carriers are dominant at high temperature and independent of gate voltage. (d) Simulated *I/W* versus *F* with variation of low‐field mobility‐based

where *vF* is the Fermi velocity of graphene. The intrinsic carrier density of graphene is

induced by electrostatic gate *Vg* is *ng* =*Cox*(*Vdr* −*Vg*) /*e*, where *Vdr* is the charge neutral voltage

The current density *J* in the suspended graphene, with an applied electric field, is presented

where *e* is the electron charge, *nex* (*nhx*) is the electron (hole) density, *vd* is the drift velocity along

1/ , [1 ( / ) ] *x x*

m

h h

*<sup>F</sup> <sup>v</sup> F v*

m

*x x sat*

of graphene. In total, the electron (hole) carrier density (see **Figure 6c**) is as follows:

<sup>2</sup> <sup>+</sup> (*npd* / <sup>2</sup>)<sup>2</sup>), where *npd* is electron‐hole puddle‐induced carriers. The carrier density

( ) ( ) <sup>1</sup> <sup>2</sup> . <sup>2</sup> *eh g g i nn n n n* = ±+ + (9)

( ) , *ex hx dx J en n v* = + (10)

<sup>=</sup> <sup>+</sup> (11)

electro‐thermal self‐consistent simulation (from Ref. [24]).

the graphene channel. The drift velocity is expressed by:

*dx*

*ni* =4(*nth*

by the continuity equation:

92 Advances in Carbon Nanostructures

Using a self‐consistent electrical‐thermal calculation in suspended graphene, as shown in **Figure 6d**, clean samples show the saturation behaviour under high electric field. This implies that high charge carrier mobility of suspended graphene devices have a larger temperature dependence (*β* ∼2.5) due to the strong electron‐intrinsic graphene optical phonon interactions. However, disordered suspended graphene shows superliner behaviour and less dependence on temperature (*β* ∼ 1.5). This is due to dominant hopping transport with increase of thermally generated charge carriers under high electric field [24].

In contrast with suspended CNTs, suspended graphene does not show a negative differential conductance, due to the linear increase of density of states in 2D, whereas CNTs have strong 1D phonon scattering, resulting in a negative differential conductance [7].

Along the temperature profile of suspended graphene, drift velocity and charge carrier density are affected by current density. Theoretically, saturation velocity of graphene was predicted as *vsat* ∼2.7 ×10<sup>7</sup> *cm* / *s*, which is limited by 160 meV intrinsic graphene optical phonons [26]. Experimentally, saturation velocity before breakdown is measured as *vsat* ∼1.7 ×10<sup>7</sup> *cm* / *s* [24]. Of course, in a suspended graphene structure, flexural phonons [27] cannot be ignored, as they may be one of reasons for variation in saturation velocity.

Suspended graphene is well known as a material having the highest thermal conductivity at room temperature (*k*0∼ 2500 *W* / *mK*). However, thermal conductivity of suspended gra‐ phene displays an abrupt decrease (∼*T‐1.7*) at high temperature [24]. This is attributed to the flexural phonons modes which enable strong second‐order three‐phonon scattering [28] (∼*T‐2*) and first‐order Umklapp phonon–phonon scattering (∼*T*). The estimated average thermal conductivity of electrically biased suspended graphene at 1000 K is of ∼310 *W* / *mK*, for both mechanically exfoliated and chemical vapour deposited (CVD) grown graphene.

## **5. Bright visible light emission from suspended graphene**

Graphene has a great potential as an efficient thermal radiation light source due to its superior mechanical strength, high‐temperature stability, high current density and non‐equilibrium phonon modes. However, previous reports regarding the thermal radiation of supported graphene [9, 10, 12, 19] have been limited to the near infrared light emission range, and extremely small fractions of electrical power are converted into the photons (∼10-6). The main reason for the limited performance of graphene light emitters supported on a substrate is the dominant heat dissipation through the substrate, and significant hot carriers cooling by dominant extrinsic scattering origins, such as charged impurities, trapped charges and surface polar phonons. In the case of freely suspended graphene, these issues can be ignored. There‐ fore, suspended structures are promising for the observation of bright, efficient light emission radiation approaching the intrinsic properties.

There are several methods that can be used for the synthesis of suspended graphene structures. One is wet‐etching of a sacrificial layer after metallization onto the graphene. The other procedure is the transfer of patterned graphene onto pre‐fabricated trench substrate. In order to increase the fabrication yield and prevent the collapse of graphene, a critical point dryer process is important. Suspended graphene structures are fabricated using mechanically

**Figure 7.** (a) SEM image of suspended monolayer graphene devices. (b) Schematic of electrically biased suspended graphene and light emission from the centre of graphene. Micrographs of bright visible light emission from a suspend‐ ed mechanically exfoliated graphene: (c) few‐layer graphene under *Vsd* = 2.90 V, (d) multi‐layer graphene under *Vsd* =  7.90 V, and (e) monolayer graphene *Vsd* = 2.58 V. (f) Optical image of remarkably bright visible light emission from a suspended few‐layer graphene, which is visible even to the naked eye (from Ref. [25])

exfoliated graphene or large‐scale graphene grown by the CVD method. **Figure 7a** shows the obtained suspended graphene devices [25].

thermal conductivity of electrically biased suspended graphene at 1000 K is of ∼310 *W* / *mK*, for both mechanically exfoliated and chemical vapour deposited (CVD) grown graphene.

Graphene has a great potential as an efficient thermal radiation light source due to its superior mechanical strength, high‐temperature stability, high current density and non‐equilibrium phonon modes. However, previous reports regarding the thermal radiation of supported graphene [9, 10, 12, 19] have been limited to the near infrared light emission range, and extremely small fractions of electrical power are converted into the photons (∼10-6). The main reason for the limited performance of graphene light emitters supported on a substrate is the dominant heat dissipation through the substrate, and significant hot carriers cooling by dominant extrinsic scattering origins, such as charged impurities, trapped charges and surface polar phonons. In the case of freely suspended graphene, these issues can be ignored. There‐ fore, suspended structures are promising for the observation of bright, efficient light emission

There are several methods that can be used for the synthesis of suspended graphene structures. One is wet‐etching of a sacrificial layer after metallization onto the graphene. The other procedure is the transfer of patterned graphene onto pre‐fabricated trench substrate. In order to increase the fabrication yield and prevent the collapse of graphene, a critical point dryer process is important. Suspended graphene structures are fabricated using mechanically

**Figure 7.** (a) SEM image of suspended monolayer graphene devices. (b) Schematic of electrically biased suspended graphene and light emission from the centre of graphene. Micrographs of bright visible light emission from a suspend‐ ed mechanically exfoliated graphene: (c) few‐layer graphene under *Vsd* = 2.90 V, (d) multi‐layer graphene under *Vsd* =  7.90 V, and (e) monolayer graphene *Vsd* = 2.58 V. (f) Optical image of remarkably bright visible light emission from a

suspended few‐layer graphene, which is visible even to the naked eye (from Ref. [25])

**5. Bright visible light emission from suspended graphene**

radiation approaching the intrinsic properties.

94 Advances in Carbon Nanostructures

In order to observe light emission from suspended graphene under electrically bias, an experimental set‐up, similar to that shown in **Figure 7b**, is used under vacuum (∼10-5 Torr), at room temperature. In addition, to get clean graphene channel and reliable contact resistance, a slow current‐induced annealing process [29] is essential. Suspended graphene exhibits bright visible light at the centre of the graphene channel once the applied bias voltage exceeds the critical field (0.5 ∼ 1.0 V / μm), as shown in **Figure 7c**–**e**. The intensity of visible light is remarkably intense and can even be observed by the naked eye, without additional magnifi‐ cation, as shown in **Figure 7f**.

Y.D. Kim et al. [25] also demonstrated multiple visible light emission from a large‐scale suspended graphene array, which was fabricated from large‐scale CVD graphene. The synthesis of a large‐scale, robust and bright visible light emitter array from CVD graphene will allow obtaining graphene‐based flexible and transparent lighting and display modules. Furthermore, complementary metal‐oxide‐semiconductor (CMOS) compatibility of graphene technology will enable a hybrid graphene‐photonic platform for ultrafast optical communi‐ cations.

The radiation spectrum from electrically biased suspended graphene over a trench of depth (*D* ∼ 1000 nm) exhibits multiple emission peaks in the visible range, as shown in **Figure 8a**. These multiple emission peaks are quite interesting, since graphene is a gapless material, and the radiation spectrum was predicted to follow the featureless Planck's law of a grey body (Eq. (3)). The multiple peaks in the visible range are strongly affected by the trench depth rather than by the number of layers of graphene and electronic band structure.

**Figure 8.** (a) Spectra of visible light emitted from electrically biased suspended graphene exhibiting multiple strong emission peaks. Thermal radiation with interference effect (solid line) fit well to the experimental data (symbol). Esti‐ mated electron temperatures of suspended graphene are indicated as function of applied source‐drain bias. Inset: emission peak energies as a function of source‐drain bias and applied electric field. (b) Interference effect between re‐ flected (dashed arrow) and thermal radiation originated directly from graphene suspended over a trench. (c) Estimat‐ ed thermal conductivity and (d) optical phonon temperature of suspended graphene based on self‐consistent electrical and thermal transport, where it is assumed that *Top* = *Te*. (e) Measured temperature of graphene (symbol) and calculat‐ ed temperature (dashed line) as function of applied electric field (from Ref. [25]).

These multiple emission peaks in the visible range and significant modulation by trench depth can be understood by interference effects between the light emitted directly from the sus‐ pended graphene and light reflected from the substrate as shown in **Figure 8b**. When we neglect the light reflection and absorption by the graphene, the interference effect on thermal radiation is given by:

$$I\left(\left.o,D\right> = I\_0\left(o\right)\right| \frac{1+\left|r\left(o\right)\right|}{2} + Re\left[r\left(o\right)\exp\left(\frac{i2oD}{c}\right)\right],\tag{13}$$

where *I*0(*ω*)<sup>∼</sup> *<sup>ω</sup>* <sup>3</sup> / (exp(ℏ*<sup>ω</sup>* / *kBTe*)−1)) is the thermal radiation intensity of suspended graphene, *r*(*ω*) is the reflection coefficient of Si substrate, *ω* is the photon frequency, is the emissivity of graphene, *D* is the trench depth, *kB* is the Boltzmann constant, *Te* is the electron temperature of graphene, and *c* is the speed of light [25]. A strong interference effect enables the selectively enhancement of thermal radiation from graphene at a certain wavelength and can increase the radiation efficiency through the engineering of interference effects.

Based on thermal radiation with interference effect (Eq. 13), electron temperature of suspended graphene was extracted, as shown in **Figure 8a** (solid line) and approaches 2800 K. The light emission peak in the visible range rapidly increases with the applied electric field above a critical field (∼0.4 *V* / *μm*). This is attributed to the accumulation of hot electrons and optical phonons in suspended graphene, above the critical field, which is related to the activation of intrinsic graphene optical phonons under an electric field. In suspended graphene, low activation to generate intrinsic optical phonons in an electric field are achieved by the reduction of energy loss suffered by electric field‐induced hot electrons with extrinsic scattering sources, and prevent the direct cooling of hot electron and phonons through heat transfer to substrate. Current saturation in suspended graphene is observed and is known to be a signature of strong electron scattering by intrinsic optical phonons of graphene itself.

Raman spectroscopy provides an accurate measurement of acoustic phonons, optical phonon population and the temperature under applied electric field, as described in Sections 2.2 and 2.3. However, bright thermal radiation from electrically biased suspended graphene becomes significantly stronger than the Raman signal above ∼1500 K. In order to estimate the temper‐ ature profiles and thermal conductivity, self‐consistent numerical simulation of electrical and thermal transport is used, as described in Section 4. From numerical simulation based on the electrical transport data, thermal conductivity at the centre of the suspended graphene channel decreases from *k* ∼ 2700 *W* / *mK* (*Tap ∼* 300 K) to *k* ∼ 65 *W* / *mK* (*Tap ∼* 1800 K), as shown in **Figure 8c**. An abrupt decrease of thermal conductivity of suspended graphene also suppresses lateral heat dissipation along the graphene channel, meaning that hot electrons are confined and spatially localized at the centre of the graphene. As shown in **Figure 8d**, the maximum temperature is located at the centre of the graphene channel, which is consistent with the brightest light emission position as shown in **Figure 7**.

Effectively localization of hot electron in suspended graphene by unique electronic and thermal properties enables bright visible light emission. Based on the Stefan–Boltzmann law from measured electron temperature, as shown in **Figure 8e**, thermal radiation efficiency was estimated as ∼ 4.45 ×10−<sup>3</sup> , which is a 1000‐fold enhancement compared to the graphene devices on substrates (∼10-6). Furthermore, engineering of radiation spectrum enhancement in the visible range provides the further enhancement of radiation efficiency.

An electrically driven graphene light emitter has great advantages, such as being atomically thin, broadband emission, radiation spectrum tunability, surface and self‐emission all with a simple structure. Furthermore, ultrafast and broadband electrical and optical response of graphene has enabled the development of a nanoscale ultrafast light emitter. We expect to obtain atomically thin, flexible and transparent light sources and on‐chip optical interconnects for communications based on graphene.

## **Acknowledgements**

These multiple emission peaks in the visible range and significant modulation by trench depth can be understood by interference effects between the light emitted directly from the sus‐ pended graphene and light reflected from the substrate as shown in **Figure 8b**. When we neglect the light reflection and absorption by the graphene, the interference effect on thermal

> ( ) ( ) <sup>0</sup> ( ) <sup>1</sup> <sup>2</sup> , ( )( exp , <sup>2</sup> *<sup>r</sup> i D I DI Re r <sup>c</sup>* w

where *I*0(*ω*)<sup>∼</sup> *<sup>ω</sup>* <sup>3</sup> / (exp(ℏ*<sup>ω</sup>* / *kBTe*)−1)) is the thermal radiation intensity of suspended graphene, *r*(*ω*) is the reflection coefficient of Si substrate, *ω* is the photon frequency, is the emissivity of graphene, *D* is the trench depth, *kB* is the Boltzmann constant, *Te* is the electron temperature of graphene, and *c* is the speed of light [25]. A strong interference effect enables the selectively enhancement of thermal radiation from graphene at a certain wavelength and can increase the

Based on thermal radiation with interference effect (Eq. 13), electron temperature of suspended graphene was extracted, as shown in **Figure 8a** (solid line) and approaches 2800 K. The light emission peak in the visible range rapidly increases with the applied electric field above a critical field (∼0.4 *V* / *μm*). This is attributed to the accumulation of hot electrons and optical phonons in suspended graphene, above the critical field, which is related to the activation of intrinsic graphene optical phonons under an electric field. In suspended graphene, low activation to generate intrinsic optical phonons in an electric field are achieved by the reduction of energy loss suffered by electric field‐induced hot electrons with extrinsic scattering sources, and prevent the direct cooling of hot electron and phonons through heat transfer to substrate. Current saturation in suspended graphene is observed and is known to be a signature of strong

Raman spectroscopy provides an accurate measurement of acoustic phonons, optical phonon population and the temperature under applied electric field, as described in Sections 2.2 and 2.3. However, bright thermal radiation from electrically biased suspended graphene becomes significantly stronger than the Raman signal above ∼1500 K. In order to estimate the temper‐ ature profiles and thermal conductivity, self‐consistent numerical simulation of electrical and thermal transport is used, as described in Section 4. From numerical simulation based on the electrical transport data, thermal conductivity at the centre of the suspended graphene channel decreases from *k* ∼ 2700 *W* / *mK* (*Tap ∼* 300 K) to *k* ∼ 65 *W* / *mK* (*Tap ∼* 1800 K), as shown in **Figure 8c**. An abrupt decrease of thermal conductivity of suspended graphene also suppresses lateral heat dissipation along the graphene channel, meaning that hot electrons are confined and spatially localized at the centre of the graphene. As shown in **Figure 8d**, the maximum temperature is located at the centre of the graphene channel, which is consistent with the

 w

ww

radiation efficiency through the engineering of interference effects.

electron scattering by intrinsic optical phonons of graphene itself.

brightest light emission position as shown in **Figure 7**.

w

<sup>+</sup> é ù æ ö = + ê ú ç ÷ ë û è ø (13)

radiation is given by:

96 Advances in Carbon Nanostructures

The authors thank J. Hone, Y.D. Park and G. Arefe. YDK was supported by Grants from ONR (N00014‐13‐1‐0662 and N00014‐13‐1‐0464) and DE-SC0012592. MB was supported by Grants from the National Research Foundation of Korea (NRF‐2012‐M3C1A1‐048861, NRF‐ 2015R1A2A1A10056103) funded by the Korean government.

## **Author details**

Young Duck Kim1\* and Myung‐Ho Bae2\*

\*Address all correspondence to: yk2629@columbia.edu and mhbae@kriss.re.kr

1 Department of Mechanical Engineering, Columbia University, New York, NY, USA

2 Korea Research Institute of Standards and Science, Daejeon, Republic of Korea

## **References**


[15] Jae Hun Seol, Insun Jo, Arden L Moore, Lucas Lindsay, Zachary H Aitken, Michael T Pettes, Xuesong Li, Zhen Yao, Rui Huang, David Broido, Natalio Mingo, Rodney S Ruoff, and Li Shi. Two‐dimensional phonon transport in supported graphene. Science. 2010;328(5975):213–216. doi:10.1126/science.1184014

[3] Alexander A Balandin, Suchismita Ghosh, Wenzhong Bao, Irene Calizo, Desalegne Teweldebrhan, Feng Miao, and Chun Ning Lau. Superior thermal conductivity of

[4] Y M Lin, C Dimitrakopoulos, K A Jenkins, D B Farmer, H Y Chiu, A Grill, and Ph Avouris. 100‐GHz transistors from wafer‐scale epitaxial graphene. Science.

[5] Changyao Chen, Sunwoo Lee, Vikram V Deshpande, Gwan‐Hyoung Lee, Michael Lekas, Kenneth Shepard, and James Hone. Graphene mechanical oscillators with tunable frequency. Nat Nanotechnol. 2013;8(12):923–927. doi:10.1038/NNANO.

[6] Eric Pop. Energy dissipation and transport in nanoscale devices. Nano Res. 2010;3(3):

[7] Eric Pop, David Mann, Jien Cao, Qian Wang, Kenneth Goodson, and Hongjie Dai. Negative differential conductance and hot phonons in suspended nanotube molecular

[8] Marcus Freitag, Mathias Steiner, Yves Martin, Vasili Perebeinos, Zhihong Chen, James C Tsang, and Phaedon Avouris. Energy dissipation in graphene field‐effect transistors.

[9] Stéphane Berciaud, Melinda Y Han, Kin Fai Mak, Louis E Brus, Philip Kim, and Tony F Heinz. Electron and optical phonon temperatures in electrically biased graphene.

[10] Marcus Freitag, Hsin‐Ying Chiu, Mathias Steiner, Vasili Perebeinos, and Phaedon Avouris. Thermal infrared emission from biased graphene. Nat Nanotechnol. 2010;5(7):

[11] Insun Jo, I‐Kai Hsu, Yong J Lee, Mir Mohammad Sadeghi, Seyoung Kim, Stephen Cronin, Emanuel Tutuc, Sanjay K Banerjee, Zhen Yao, and Li Shi. Low‐frequency acoustic phonon temperature distribution in electrically biased graphene. Nano Lett.

[12] Myung‐Ho Bae, Zhun‐Yong Ong, David Estrada, and Eric Pop. Imaging, simulation, and electrostatic control of power dissipation in graphene devices. Nano Lett.

[13] Dong‐Hun Chae, Benjamin Krauss, Klaus von Klitzing, and Jurgen H Smet. Hot phonons in an electrically biased graphene constriction. Nano Lett. 2010;10(2):466–471.

[14] Myung‐Ho Bae, Sharnali Islam, Vincent E Dorgan, and Eric Pop. Scaling of high‐field transport and localized heating in graphene transistors. ACS Nano. 2011;5(10):7936–

Phys Rev Lett. 2010;104(22):227401. doi:10.1103/PhysRevLett.104.227401

wires. Phys Rev Lett. 2005;95(15):155505. doi:10.1103/PhysRevLett.95.155505

single‐layer graphene. Nano Lett. 2008;8(3):902–907. doi:10.1021/nl0731872

2010;327(5966):662–662. doi:10.1126/science.1184289

Nano Lett. 2009;9(5):1883–1888. doi:10.1021/nl803883h

147–169. doi:10.1007/s12274‐010‐1019‐z

497–501. doi:10.1038/nnano.2010.90

2011;11(1):85–90. doi:10.1021/nl102858c

doi:10.1021/nl903167f

7944. doi:10.1021/nn202239y

2010;10(12):4787–4793. doi:10.1021/nl1011596

2013.232

98 Advances in Carbon Nanostructures


visible light emission from graphene. Nat Nanotechnol. 2015;10(8):676–681. doi: 10.1038/NNANO.2015.118


## **Electrospun Graphene Oxide-Based Nanofibres**

Izzati Fatimah Wahab, Saiful Izwan Abd Razak, Nor Syuhada Azmi, Farah Nuruljannah Dahli, Abdul Halim Mohd Yusof and Nadirul Hasraf Mat Nayan

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64055

#### **Abstract**

visible light emission from graphene. Nat Nanotechnol. 2015;10(8):676–681. doi:

[26] Tian Fang, Aniruddha Konar, Huili Xing, and Debdeep Jena. High‐field transport in two‐dimensional graphene. Phys Rev B. 2011;84(12):125450. doi:10.1103/PhysRevB.

[27] Eduardo V Castro, H Ochoa, M I Katsnelson, R V Gorbachev, D C Elias, K S Novoselov, A K Geim, and F Guinea. Limits on charge carrier mobility in suspended graphene due to flexural phonons. Phys Rev Lett. 2010;105(26):266601. doi:10.1103/PhysRevLett.

[28] Denis L Nika, Artur S Askerov, and Alexander A Balandin. Anomalous size depend‐ ence of the thermal conductivity of graphene ribbons. Nano Lett. 2012;12(6):3238–3244.

[29] J Moser, A Barreiro, and A Bachtold. Current‐induced cleaning of graphene. Appl Phys

10.1038/NNANO.2015.118

84.125450

100 Advances in Carbon Nanostructures

105.266601

doi:10.1021/nl301230g

Lett. 2007;91(16):1653513. doi:10.1063/1.2789673

Graphene and graphene oxide have been the forefront of research in the field of carbon nanostructures. In this chapter, the authors attempt to review and discuss works on electrospinning of graphene oxide-based nanofibre materials, mainly focusing on their combination with polymeric materials. Explanation and insights towards their method of preparation, properties and applications will be discussed. The conclusion section will give some closing remarks regarding the future direction of graphene-based nanofibres.

**Keywords:** graphene oxide, electrospinning, nanofibres, nanocomposites, nanomate‐ rials, polymer matrices

## **1. Introduction**

The discovery of graphene around this decade changed the focus direction of material science into its unique and magnificent characteristics. Researchers throughout the world are compet‐ ing in finding ways to synthesis and explore the possible applications of graphene-based materials. Graphene oxide (GO) is an intermediate material that has better hydrophilicity and easier modification due to the presence of oxygen functionalities, compared to graphene. A broad range of applications have been reported over the years, including tissue engineering, sensor, power storage, coating and photonics. Electrospinning is one of the techniques that make use of GO materials in nanofibrous composites. This state-of-the art technique has been very

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

popular in facile 1-D nano-scaled fibres production and gives materials their excellent proper‐ ties in relation to high specific surface area.

## **2. Graphene oxide**

The novel nanostructure of graphene has been showing promising characteristics not only in fundamental science but also in variety of applications. Its remarkable chemical and physical properties urged into many attempts of synthesizing graphene in mass production. However, it fails to give adequate productivity for large-scale use [1]. Graphene is not easy to be handled when it comes to incorporating it with other materials and to transform it into other form. Current development in this field recognizes graphene oxide (GO) as the best alternative to alleviate this problem, by having a more solvent friendly structure [2]. The solubility of GO in water and other solvents is likely in favour to the large commercialization of this novel material.

### **2.1. Structure**

GO is basically a graphene sheet with oxygen functional groups, either epoxy, carboxyl or hydroxyl. It can be easily exfoliated from graphite through oxidation and sonication processes. Since graphene is a one-atomic layer of graphite, the structure of GO also consists in exfoliated layers of graphite oxide. Simplistically, the structural process is shown in **Figure 1**. The carbon compounds of **Figure 1** are organized in a honey-comb structure. One sheet of graphene contains six-carbon-atom hexagons, bonded by strong π and σ bonds.

The synthesis of GO involves oxidation of graphite into graphite oxide and exfoliation of graphite oxide into graphene oxide (**Figure 1**). There are several routes that can be utilized to oxidize graphite. The Brodie method, Brodie–Staudenmaier method and Hummers method are the earliest routes developed, but they lead to the release of toxic gases like NOx and ClOx [3]. Later, a typical modified Hummers method was introduced, and it has been used in many reported works in literature [4–6]. In general, graphite–graphene oxide conversion process can be summarized down into three steps [7]. First, the conversion of graphite into graphite intercalation compound (H2SO4–GIC), then, as a second step, H2SO4–GIC is converted into pristine graphite oxide (PGO). Finally, the conversion of PGO into GO takes place by the reaction of PGO with water.

The structure of GO on its own has become a research area interest among material scientists. This is due to the complexity of its non-stoichiometric composition [8]. The most recent and acceptable structure is constructed by Lerf and Klinowski. The Lerf–Klinowski model ignores the lattice-based model and focuses on a nonstoichiometric-amorphous alternative. It contains unoxidized benzene rings and aliphatic six-membered rings with epoxy, hydroxyl, carbonyl and/or carboxyl groups on the graphene sheet or at the graphene edges. Each oxygen atom is covalently bonded to two carbon atoms, forming an epoxy ring. Reduced graphene oxide (rGO) structure contains both sp2 and sp3 hybridized carbons and a high carbon/oxygen atoms ratio (C:O). sp2 carbons bind to three other atoms, sharing a double bond with one of them, while a sp3 carbon binds to a four other atoms.

A study on atomic and electronic structure of GO using annular dark field (ADF) imaging and electron energy loss spectroscopy (EELS) was conducted by Mkhoyan et al. in 2009 [9]. They found out that oxygen atoms randomly attach to graphene sites and convert sp2 carbon bonds into sp3 bonds. By ab initio calculated structure, two carbon atoms bonded to an oxygen atom are pulled above graphene plane. This configuration is fairly stable with −3.12 eV absorption energy. The bond length between these two carbon atoms expands to 1.514 Å from graphene 1.407 Å [9]. This is close to the diamond sp3 bond length of 1.54 Å. The sp<sup>3</sup> hybridization and π and σ bonds of graphene oxide contribute the most to its unique physical and chemical properties. The extent of polar groups on the surface and edges of the graphene oxide depends on the amount of the oxidizing agent and the time and temperature of the oxidation reaction.

#### **2.2. Properties**

popular in facile 1-D nano-scaled fibres production and gives materials their excellent proper‐

The novel nanostructure of graphene has been showing promising characteristics not only in fundamental science but also in variety of applications. Its remarkable chemical and physical properties urged into many attempts of synthesizing graphene in mass production. However, it fails to give adequate productivity for large-scale use [1]. Graphene is not easy to be handled when it comes to incorporating it with other materials and to transform it into other form. Current development in this field recognizes graphene oxide (GO) as the best alternative to alleviate this problem, by having a more solvent friendly structure [2]. The solubility of GO in water and other solvents is likely in favour to the large commercialization of this novel

GO is basically a graphene sheet with oxygen functional groups, either epoxy, carboxyl or hydroxyl. It can be easily exfoliated from graphite through oxidation and sonication processes. Since graphene is a one-atomic layer of graphite, the structure of GO also consists in exfoliated layers of graphite oxide. Simplistically, the structural process is shown in **Figure 1**. The carbon compounds of **Figure 1** are organized in a honey-comb structure. One sheet of graphene

The synthesis of GO involves oxidation of graphite into graphite oxide and exfoliation of graphite oxide into graphene oxide (**Figure 1**). There are several routes that can be utilized to oxidize graphite. The Brodie method, Brodie–Staudenmaier method and Hummers method are the earliest routes developed, but they lead to the release of toxic gases like NOx and ClOx [3]. Later, a typical modified Hummers method was introduced, and it has been used in many

contains six-carbon-atom hexagons, bonded by strong π and σ bonds.

ties in relation to high specific surface area.

**2. Graphene oxide**

102 Advances in Carbon Nanostructures

material.

**2.1. Structure**

**Figure 1.** Structural process of graphene oxide.

Spotted and closely studied only since 2004, graphene has already caught wide attention in research history after one by one of its interesting properties were reported. Being a single atomic layer of graphite, graphene is very thin and light. Yet, it is the strongest material ever discovered with intrinsic strength of 42 Nm−1 [10], so robust that it is 200 times stronger than steel and can outperform the hardness of diamond. It is transparent, but it can absorb light [11]. Having Young's modulus to approximately 1 TPa [10], it is elastic and pliable but, at the same time, is impermeable to all gases and liquids, except water [12]. It was also reported that graphene possesses superior thermal conductivity in the range of ~(4.84 ± 0.44) × 10<sup>3</sup> to (5.30 ± 0.48) × 10<sup>3</sup> W/mK [13]. These values suggest that it can far exceed carbon nanotubes in con‐ ducting heat. Another important character of graphene is its exceptional electrical properties for having high intrinsic mobility of 2 × 105 cm2 v−1s−1, compared to the highest known inorganic semiconductor (carbon nanotubes, with ~1 × 105 cm2 v−1s−1) [14, 15].

Despite of being the building block of graphite layers, the synthesis of graphene is not a straight-forward exfoliation process. The method that discovered graphene sheets in 2004, using cellophane tape, does not appear to be a readily scalable process. Graphite exists in large quantities and is not expensive, yet producing sufficient graphene sheets is still a huge challenge [3]. Even the use of chemical vapour deposition (CVD) to deposit graphene on arbitrary substrates is not possible at low temperatures and is incompatible for mass produc‐ tion processes. Another concern is that graphene is a difficult material to be incorporated into various polymer matrices. Although graphene has unique properties, due to the strong π and σ bonds, these also contribute to its weak reactivity and low solubility in common organic solvents.

In order to prepare graphene–polymer composites or graphene-based hybrid materials, stable dispersion of graphene with other matrices must be obtained. This is when GO becomes useful. The sp3 hybridization of GO contributes to a larger amount of free bonds, therefore increasing reactivity [16]. When graphite is oxidized, the functional groups make its hydrophilic and improve its exfoliation in aqueous media or organic solutions. The surface functionalization and solvents are the critical factors in determining interaction with functional groups of polymer or other matrixes. Several studies were done to determine the behaviour of GO towards a wide variety of solvents. Paredes et al. [17] studied the dispersibility of GO in acetone, methanol, ethanol, 1-propanol, ethylene glycol, dimethyl sulfoxide (DMSO), *N*,*N*dimethylformamide (DMF), *N*-methyl-2-pyrrolidone (NMP), pyridine, tetrahydrofuran (THF), dichloromethane (DCM), o-xylene and n-hexane. They have successfully identified some type of solvents (DMF, NMP, THF and ethylene glycol) that can maintain a stable GO solubility. Later, another similar study compared GO and rGO solubility in eighteen different solvents. That investigation suggests that polarity of solvents is not the only factor for good dispersibility [2]. NMP and ethylene glycol, besides water, were double-proved to be the best solvents for long-term stability.

The covalent character of C–O bonds in GO disrupts the sp2 conjugation of pure graphene and gives GO its insulating properties. Electrical conductivity can be recovered by restoring the π–network. GO, due to the presence of abundant oxygen, can be easily modified. Control of oxidation can provide tunable electronic properties, including the possibility of accessing zeroband gap graphene via bonds. GO, which contains a C/O content ratio of 2/1, is non-conductive (less than 1 S/m). GO can be partially reduced to conductive graphene-like sheets by removing the oxygen-containing groups. Restoration of the π–network and conjugated structure can be obtained by reducing GO to 6/1 of C/O atomic ratio. GO reduction is a process that converts sp3 to sp2 carbon, and thus recovers the important electrical properties. Some researchers make use of the oxygenated functionalities to bind with functional groups of conducting polymers, such as polyaniline, to improve the conductivity.

## **3. Electrospinning**

At nanoscale, materials exist in 0, 1, 2, or 3 dimensions. One dimensional nanofibres can be synthesized by several techniques such as phase separation, self-assembly, template synthesis and electrospinning. Among those techniques, electrospinning has been recognized as a facile process in fabricating polymeric fibres in nano-size [18]. It promptly gained popularity since its establishment, due to its potential implementation in various fields such as tissue engi‐ neering, chemical and biological sensing, energy generation and textiles.

### **3.1. Basics**

using cellophane tape, does not appear to be a readily scalable process. Graphite exists in large quantities and is not expensive, yet producing sufficient graphene sheets is still a huge challenge [3]. Even the use of chemical vapour deposition (CVD) to deposit graphene on arbitrary substrates is not possible at low temperatures and is incompatible for mass produc‐ tion processes. Another concern is that graphene is a difficult material to be incorporated into various polymer matrices. Although graphene has unique properties, due to the strong π and σ bonds, these also contribute to its weak reactivity and low solubility in common organic

In order to prepare graphene–polymer composites or graphene-based hybrid materials, stable dispersion of graphene with other matrices must be obtained. This is when GO becomes useful.

gives GO its insulating properties. Electrical conductivity can be recovered by restoring the π–network. GO, due to the presence of abundant oxygen, can be easily modified. Control of oxidation can provide tunable electronic properties, including the possibility of accessing zeroband gap graphene via bonds. GO, which contains a C/O content ratio of 2/1, is non-conductive (less than 1 S/m). GO can be partially reduced to conductive graphene-like sheets by removing the oxygen-containing groups. Restoration of the π–network and conjugated structure can be obtained by reducing GO to 6/1 of C/O atomic ratio. GO reduction is a process that converts

use of the oxygenated functionalities to bind with functional groups of conducting polymers,

At nanoscale, materials exist in 0, 1, 2, or 3 dimensions. One dimensional nanofibres can be synthesized by several techniques such as phase separation, self-assembly, template synthesis and electrospinning. Among those techniques, electrospinning has been recognized as a facile

carbon, and thus recovers the important electrical properties. Some researchers make

 hybridization of GO contributes to a larger amount of free bonds, therefore increasing reactivity [16]. When graphite is oxidized, the functional groups make its hydrophilic and improve its exfoliation in aqueous media or organic solutions. The surface functionalization and solvents are the critical factors in determining interaction with functional groups of polymer or other matrixes. Several studies were done to determine the behaviour of GO towards a wide variety of solvents. Paredes et al. [17] studied the dispersibility of GO in acetone, methanol, ethanol, 1-propanol, ethylene glycol, dimethyl sulfoxide (DMSO), *N*,*N*dimethylformamide (DMF), *N*-methyl-2-pyrrolidone (NMP), pyridine, tetrahydrofuran (THF), dichloromethane (DCM), o-xylene and n-hexane. They have successfully identified some type of solvents (DMF, NMP, THF and ethylene glycol) that can maintain a stable GO solubility. Later, another similar study compared GO and rGO solubility in eighteen different solvents. That investigation suggests that polarity of solvents is not the only factor for good dispersibility [2]. NMP and ethylene glycol, besides water, were double-proved to be the best

conjugation of pure graphene and

solvents.

104 Advances in Carbon Nanostructures

The sp3

sp3 to sp2

**3. Electrospinning**

solvents for long-term stability.

The covalent character of C–O bonds in GO disrupts the sp2

such as polyaniline, to improve the conductivity.

The first observation on electrospinning dates back from the study of the electrospraying method used by Rayleigh (1897) and Zeleny (1914) [19]. Formhals, however, is the one who patented the electrostatic force experimental setup of polymer filaments production. In 1934, Formhals published a patent in which describing the production of textile yarn from cellulose acetate [20]. Since then, further research was carried out in order to assemble fibre processing using electrostatic force. Later, the study of electrically driven jets, by Taylor in 1969, became a strong foundation in understanding the electrospinning process. Only around 1994, the electrospinning term was used. Along with its popularity and rapid establishment in many fields, electrospinning has undergone many modifications and upgrading, in order to maxi‐ mize its production and effectiveness in fabricating versatile nanofibres. There are several advanced techniques and setups of electrospinning instruments used favourably according to the required characterizations. Coaxial electrospinning, multilayer and mixed electrospinning, forced air-assisted electrospinning and air-gap electrospinning are a number of innovative processing techniques to enhance the function of electrospun fibres [21].

Basically, the typical basic components of any electrospinning setup are the high-voltage power supply, the spinneret (normally syringe with blunt-tip needle) and the collector (normally grounded). This basic experimental setup is shown in **Figure 2**. The polymer melt or solution placed into the syringe is held by retort stand (for vertical setups) or with a syringe pump (for vertical and horizontal setups). The tip of the syringe needle is connected to the positive electrode of the voltage supply. From here, the polymer liquid is pumped out of the syringe through the needle and brought to the grounded collector beneath the syringe or placed horizontally.

**Figure 2.** Electrospinning basic experimental setup.

**Figure 3.** Description of electrospinning principle.

This unique approach for producing continuous fibres from polymer melt or solution, works by the principle of surface tension acting on the polymer drops. By the time, electricity is supplied onto the polymer, the fluid droplet will gain electrostatic charges, being positive charged. Due to the interaction between two forces, a Taylor cone is formed. The increase in high-voltage supply then eventually achieves a certain critical point, at which the electrostatic repulsion can overcome the surface tension holding the polymer droplet. It results in a liquid jet ejection targeting the collector. **Figure 3** shows a clearer depiction of the principle behind electrospinning technique.

The polymer jets onto grounded collector from the spinneret can be divided into two segments. There are streaming jet and whipping jet (**Figure 4**). The streaming jet can be seen with appropriate lighting, while the whipping jet is invisible to the naked eye. The flow of the streaming jet may give a wise initial prediction on the morphology of collected nanofibres. Leach et al., in 2011, indicated that streams behaving inconsistently, non-uniformly, in short

**Figure 4.** Streaming and whipping jets.

and oscillating, might produce fibres with poor alignment, beading, wavy pattern and splattering [22]. Therefore, researchers engaging in electrospinning studies need to carefully study various parameters (which will be further explained in the next section) before the optimized nanofibre formation procedure can be achieved.

### **3.2. Working parameters**

This unique approach for producing continuous fibres from polymer melt or solution, works by the principle of surface tension acting on the polymer drops. By the time, electricity is supplied onto the polymer, the fluid droplet will gain electrostatic charges, being positive charged. Due to the interaction between two forces, a Taylor cone is formed. The increase in high-voltage supply then eventually achieves a certain critical point, at which the electrostatic repulsion can overcome the surface tension holding the polymer droplet. It results in a liquid jet ejection targeting the collector. **Figure 3** shows a clearer depiction of the principle behind

The polymer jets onto grounded collector from the spinneret can be divided into two segments. There are streaming jet and whipping jet (**Figure 4**). The streaming jet can be seen with appropriate lighting, while the whipping jet is invisible to the naked eye. The flow of the streaming jet may give a wise initial prediction on the morphology of collected nanofibres. Leach et al., in 2011, indicated that streams behaving inconsistently, non-uniformly, in short

electrospinning technique.

106 Advances in Carbon Nanostructures

**Figure 4.** Streaming and whipping jets.

**Figure 3.** Description of electrospinning principle.

Electrospinning deals with many working parameters in order to fabricate the required morphology and properties in the as-spun nanofibres. Other than the electrospinning exper‐ imental setup, like having different kind of collectors or spinnerets, electrospun nanofibres are also mainly affected by working parameters. The factors to be adjusted are the properties of polymer solution or melt, operational properties and ambient parameters [18, 22]. The electrospinning solution and operational properties are among the major parameters that influence electrospun nanofibres traits. Solution viscosity, surface tension and conductivity, along with electric field strength, distance between spinneret tip with collector and also the flow rate, are thoroughly investigated by researchers to maintain optimized fibre formation. Meanwhile, most researchers prefer to keep ambient parameters such as solution temperature, humidity and surrounding air velocity as constants throughout their studies. **Table 1** displays some factors taken into account for tuning the electrospun fibre morphology. Note that whenever finer or smaller diameter fibres are spun, they tend to get beading. Therefore, it is important for researchers to recognize their requirements and determine the optimized conditions for electrospun fibre formation. There are also some parameters that need minimum requirements to be executed in order to induce electrospinnability. The applied voltage and tip-to-collector distance, for example, are some of those parameters, having certain minimum


**Table 1.** Effect of parameters on fibre morphology [19, 23].

values which differ for every electrospinning material. Other parameters that rule electro‐ spinning process are the molecular weight of the material used, types of collector and of solvent used, and ambient parameters [18, 23].

## **4. Electrospinning of GOs with polymers**

The establishment of GO incorporation into polymeric electrospun nanofibres was only initiated around this decade. Since 2010, at least 34 literature works were published involving about 14 types of synthetic and natural polymers. There are some distinct reasons to add GO and its derivatives into electrospun nanofibres. The major purpose is to improve the mechan‐ ical and electrical properties of the electrospun material. Others include the morphological enhancement to meet certain diameters or porosity of the spun fibres. **Table 2** summarizes the reported polymer/GO combinations.


**Table 2.** Reported literatures on electrospun GO.

A fundamental study by Park et al. in 2013 described the effect of GO distribution and its location in 1D polymer composites. By controlling the distribution and location of the embed‐ ded nanomaterials in 1D nanoscale composites, the macroscopic properties can be enhanced and thus enable many practical applications. In their study, three types of polymers and two different solvents were used. The type of polymers and solvents are two processing conditions to manipulate the distribution and location of the GOs. Homogeneous colloidal dispersions of GO in the polymer solutions were successfully electrospun into polymer nanofibres, without the formation of beads or aggregates [24]. The location of GO was influenced by phase separation during the electrospinning process, which was in turn dependent on the compati‐ bility between the two components and the vapour pressure of the solvents. The thermal and mechanical properties of polymeric composite nanofibres were effectively improved by adding GO as nanofiller.

values which differ for every electrospinning material. Other parameters that rule electro‐ spinning process are the molecular weight of the material used, types of collector and of solvent

The establishment of GO incorporation into polymeric electrospun nanofibres was only initiated around this decade. Since 2010, at least 34 literature works were published involving about 14 types of synthetic and natural polymers. There are some distinct reasons to add GO and its derivatives into electrospun nanofibres. The major purpose is to improve the mechan‐ ical and electrical properties of the electrospun material. Others include the morphological enhancement to meet certain diameters or porosity of the spun fibres. **Table 2** summarizes the

**Graphene oxide Polymer References** Graphene oxide Poly (vinyl acetate) (PVAc) [56] Graphene oxide Polycaprolactone (PCL) [42, 56] Functionalized graphene oxide Poly (lactic acid) (PLA) [26, 41] Graphene oxide poly(lactic-co-glycolic acid) (PLGA) and collagen (Col) [40] Graphene oxide poly(lactic-co-glycolic acid) (PLGA) [32, 40] Graphene oxide PLA and polyurethane (PU) [38] Graphene oxide Polyurethane (PU) [38, 53, 55] Graphene oxide Polyethylene oxide (PEO)/chitosan (CS) [44]

Graphene oxide Poly (vinyl alcohol) (PVA) [31, 35, 44, 48] Reduced graphene oxide Poly (vinyl alcohol) (PVA) [24, 27] Graphene oxide PVA and chitosan CS [33, 47] Graphene oxide Nylon 6-6 [52] Reduced graphene oxide Nylon-6 [57] Graphene oxide Nylon-6 [34] Reduced graphene oxide Poly(carbonate urethane) (PCU) [30] Reduced graphene oxide polyvinyl pyrrolidone (PVP) [25] Reduced graphene oxide polyacrylonitrile (PAN) [25, 29, 36] Functionalized graphene oxide polyacrylonitrile (PAN) [28] Graphene oxide Polyacrylonitrile (PAN) [28, 50] Reduced graphene oxide Polyvinyl Butyral (PVB) [46] Graphene oxide Polyacrylic acid (PAA) [54]

used, and ambient parameters [18, 23].

108 Advances in Carbon Nanostructures

reported polymer/GO combinations.

**Table 2.** Reported literatures on electrospun GO.

**4. Electrospinning of GOs with polymers**

GOs have a curved morphological structure and form powder materials after exfoliation from graphite oxides. This powder structure can be used in paints, adhesives, coatings and lubri‐ cants, where GOs act as multifunctional additives. However, for applications where structured thickness is a limitation, the levelling of the graphene sheets in a plane is required. Thus, an effort to study the levelling of graphene sheets through electrospinning was recently carried out [25]. Those authors used PAN and PVP as examples of polar polymers for nanofibre matrices. GO was reduced using different agents, but later it was confirmed that the curved features of GO sheets cannot be prevented using arGO coating strategy. Nonetheless, the strategy of using electrospinning to extend single GO sheets seems effective. Levelled gra‐ phene sheets were successfully developed and showed good conductivity in the order of 103 Sm-1.

The effect of GO loading into PCL was investigated in presence of different oxidation levels [26]. GO dispersibility, wettability, polarity and charge relaxation time, which assess the fibre diameter, increase with oxidation level. Surface tension of PCL solution is invariable in the presence of GO, resulted in diameter variations of the spun fibres. The diameter of PCL nanofibres decreases with GO loading, due to the reduction of solution viscosity and increase in the conductivity. Another morphological study attempted to blend rGO into PVA [27]. The addition of a nanofiller seems to increase the viscosity and conductivity, but the nanofibre exhibited smaller diameters than those without nanofillers. Some GO particles were also observed to be protruded from the composite nanofibres under transmission electron micro‐ scope (TEM).

The manufacturing of commercial high-performance carbon nanofibre (CNF) has been carried out using PAN as precursor, due to its high carbon yield and superior fibre strength. It includes a typical process of oxidative stabilization and carbonization steps. Neverthe‐ less, most continuous CNFs produced to date have relatively poor graphitic structure, which is important for higher elastic moduli and thermal and electrical conductivities. This can be solved by increasing the carbonization temperature, because high temperature graphitiza‐ tion significantly improves the size and alignment of the graphitic crystallites in PANbased CNFs. This high-temperature condition is, however, unlikely since fibre processing at such conditions requires a specialized expensive equipment. Therefore, Papkov and co-

workers, in 2013, carefully made an effort to maximize the orientation of the polymer chains inside the polymer precursor fibre by incorporating oriented inclusions that had high surface area and could strongly interact with polymer chains [28]. GO was used since it can both organize the polymer chains and simultaneously serve as a 'templating' agent for the formation of graphitic crystallites during carbonization. Their close examination revealed the presence of crumpled GO particles inside the PAN matrix. The irregular shape of the radially crumpled sheets closely matches the popular model of the high-strength carbon fibre structure, which consists of irregular, longitudinally wrinkled turbostratic graphitic stacks with preferred axial orientation. It is very useful to represent a natural template to achieve similar high-strength CNFs structure in economic manufacturing protocol. In addition, the crumpled graphene oxide particles may provide multiple nucleation sites for graphitic crystal growth during carbonization, especially because the carbon atoms along the crumpled folds are more reactive.

The paramount role of graphene-based embedded in polymer matrix via electrospinning is to enhance the mechanical properties of the polymeric nanofibres. Wang et al., in 2015, examined PAN/GO composite nanofibres with lateral force microscopy and force-distance curves, in order to assess the surface properties like friction force and elasticity. The experiment results indicated that, with increasing GO concentrations, both the surface friction force and adhesive force increased [29]. A mechanical characterization of high-performance GO incorporated into an aligned fibro-porous PCU membrane was made, under static and dynamic conditions [30]. The electrospun membrane indicated that 55% of the tensile strength increased, a 127% rise in toughness, and the achievement of maximum strength reinforcement efficiency was reported at 1.5 wt.% GO loading. An even smaller loading of GO (0.02%) in PVA diluted aqueous solution was able to produce elevation of tensile strength of electrospun nanofibres by 42 times [31].

Another study on significant improvements in thermomechanical and surface chemical properties of poly(D, L-lactic-co-glycolic acid) (PLGA) upon addition of GO nanosheets was published in 2011 [32]. A significant enhancement of storage moduli and glass transition temperature of the PLGA/GO (2 wt.%) was presumably reported to be caused by the en‐ hanced chemical bonding between the GO oxygenated functional groups and the polymer hydroxyl groups. This nanocomposite fibre was more hydrophilic than the pristine PLGA fibre. Biocompatibility tests showed that the enhanced surface chemical properties of the PLGA/GO (2 wt.%) effectively improved neuronal cell proliferation and viability, indicating enhanced biocompatibility of PLGA nanocomposite nanofibres by the addition of 2-D GO nanofillers. Novel electrospun PVA embedded with glycine-modifed GO (GMGO/PVA) and PVA with nanogold deposited GO (AuG/PVA) was fabricated and its biocompatibility was critically analysed [33]. GMGO was synthesized by reducing GO with a biocompatible amino acid, glycine, for conductivity improvement. Then, AuG was modified from GMGO, by mixing with hydrogen tetrachloroaurate (HAuCl4). Cellular toxicity tests of these nanofibres showed almost no cytotoxicity to L929 cells, most likely due to nontoxicity of PVA nanofibres and the biocompatibility of modified GO.

There were also some experiments carried out with produce electrospun GO/polymer nanofibres for morphological modification, in order to meet certain requirements for applications. A spider-wave-like nano-net was fabricated from GO blend with nylon-6 which holds polyelectrolytic and polymorphic characteristics [34]. The amide (CO-NH) group of nylon-6 molecule can easily interact with GO by means of hydrogen bond formation or donoracceptor complexes. The strategy was to incorporate foreign materials into nylon-6 in order to change the morphology and crystallization of nanofibres. The pristine nylon-6 electro‐ spun mat appears well defined with no spider-wave-like nano-net. As GO was added into the polymeric solution, a double-layered morphology (bimodal fibre diameter distributed mat) was noticeable. Further increase of GO in the polymeric solution not only decreased the spider-wave-like structure but also decreased the spinnability and formed a film-like structure.

workers, in 2013, carefully made an effort to maximize the orientation of the polymer chains inside the polymer precursor fibre by incorporating oriented inclusions that had high surface area and could strongly interact with polymer chains [28]. GO was used since it can both organize the polymer chains and simultaneously serve as a 'templating' agent for the formation of graphitic crystallites during carbonization. Their close examination revealed the presence of crumpled GO particles inside the PAN matrix. The irregular shape of the radially crumpled sheets closely matches the popular model of the high-strength carbon fibre structure, which consists of irregular, longitudinally wrinkled turbostratic graphitic stacks with preferred axial orientation. It is very useful to represent a natural template to achieve similar high-strength CNFs structure in economic manufacturing protocol. In addition, the crumpled graphene oxide particles may provide multiple nucleation sites for graphitic crystal growth during carbonization, especially because the carbon atoms along the crumpled folds

The paramount role of graphene-based embedded in polymer matrix via electrospinning is to enhance the mechanical properties of the polymeric nanofibres. Wang et al., in 2015, examined PAN/GO composite nanofibres with lateral force microscopy and force-distance curves, in order to assess the surface properties like friction force and elasticity. The experiment results indicated that, with increasing GO concentrations, both the surface friction force and adhesive force increased [29]. A mechanical characterization of high-performance GO incorporated into an aligned fibro-porous PCU membrane was made, under static and dynamic conditions [30]. The electrospun membrane indicated that 55% of the tensile strength increased, a 127% rise in toughness, and the achievement of maximum strength reinforcement efficiency was reported at 1.5 wt.% GO loading. An even smaller loading of GO (0.02%) in PVA diluted aqueous solution was able to produce elevation of tensile strength of electrospun nanofibres by 42 times

Another study on significant improvements in thermomechanical and surface chemical properties of poly(D, L-lactic-co-glycolic acid) (PLGA) upon addition of GO nanosheets was published in 2011 [32]. A significant enhancement of storage moduli and glass transition temperature of the PLGA/GO (2 wt.%) was presumably reported to be caused by the en‐ hanced chemical bonding between the GO oxygenated functional groups and the polymer hydroxyl groups. This nanocomposite fibre was more hydrophilic than the pristine PLGA fibre. Biocompatibility tests showed that the enhanced surface chemical properties of the PLGA/GO (2 wt.%) effectively improved neuronal cell proliferation and viability, indicating enhanced biocompatibility of PLGA nanocomposite nanofibres by the addition of 2-D GO nanofillers. Novel electrospun PVA embedded with glycine-modifed GO (GMGO/PVA) and PVA with nanogold deposited GO (AuG/PVA) was fabricated and its biocompatibility was critically analysed [33]. GMGO was synthesized by reducing GO with a biocompatible amino acid, glycine, for conductivity improvement. Then, AuG was modified from GMGO, by mixing with hydrogen tetrachloroaurate (HAuCl4). Cellular toxicity tests of these nanofibres showed almost no cytotoxicity to L929 cells, most likely due to nontoxicity of PVA nanofibres and the

are more reactive.

110 Advances in Carbon Nanostructures

[31].

biocompatibility of modified GO.

The widespread use of electrospinning technique in tissue engineering application is no longer new. Many researchers have been exploring this subject. GO/polymer composites also found their way through this area, since they showed potential biocompatibility and noncytotoxici‐ ty characteristics. Scaffolds play an important role in tissue engineering as they not only provide support for cell attachment, but they also allow cell growth into tissue until the tissue is able to sustain itself. There are several types of GO/polymer composites been developed for tissue engineering scaffolds, such as PVA/GO [35], PAN/rGO [36], PVA/CS/GO [37], PLA/PU/ GO [38], PLA/hydroxyapatite (HA)/GO [39], GO/PLGA/Col [40], GO-g-PEG/PLA [41] and PCL/GO [42]. The polymeric scaffolds attain better thermal, mechanical and electrical properties, due to the intrinsic properties of GO. Cells adhesion and proliferation onto these nanofibrous scaffolds have been tested morphologically and quantitatively. Cells like mouse osteoblastic cells (MC3T3-E1), adipose-derived stem cells (ADSCs), skeletal myoblasts (C2C12), mouse marrow mesenchymal stem cells (MSCs) and low-differentiated rat pheo‐ chromocytoma cells (PC12-L) were cultured onto the samples. They confirmed that a small amount of GO did not restrain the proliferation and viability of cells, which demonstrated the appreciable cell affinity of GO. Cells are likely to spread on the composite scaffolds and some even enhance the cell growth by having similarrate of cell proliferation to that of tissue culture plates.

**Tissue engineering** is very important for regeneration and replacement of damaged tissues in the human body. Scaffolds for musculoskeletal and skin tissues have caught the attention of researchers in the last years, aiming to help many patients with injuries. On the other hand, tissue engineering for organs or tissues without regenerative abilities is an extensive area that needs serious exploration. Vascular tissue engineering is one of its branches, and it aims to fabricate functional vascular grafts to be used in vivo, in order to replace blood vessels and help their regeneration. Current commercial synthetic grafts used by surgeons are woven poly (ethylene terephthalate) (Dacron) and extended polytetrafluoroethylene (ePTFE). Although the aim was to substitute large diameter blood vessels, they are not suitable for small diameter grafts due to thrombosis and intimal hyperplasia risks. A thermoplastic polyur‐ ethane (TPU)/GO scaffold was fabricated using grounded rotation mandrel as electrospin‐ ning collector to obtain fibrous small diameter tubular scaffolds [43]. The tensile strength

(Young's modulus) and scaffolds hydrophilicity increased as GO amount increased. The suture retention strength and burst pressure of tubular TPU/GO scaffolds containing 0.5 wt. %. GO exceeded the mechanical requirements of human blood vessels. In addition, results of platelet adhesion tests indicated that a low loading of GO is unlikely to cause thrombosis on the scaffolds. Endothelial cells could form a layer on the inner surface of the tubular scaf‐ folds which resembles the native blood vessel structure, indicating that the TPU/GO electro‐ spun scaffolds, with low GO loading level, have the potential to be used in small diameter vascular graft applications.

Another impressive function of scaffolds that needs to be expanded is the ability to **control drug release** onto specific tissues or organs. Atargeteddrugdelivery system that can efficiently transport drugs is a way for a scaffold to improve the efficacy of the drug and to reduce the systemic toxicity. Electrospun nanofibres are very important materials due to the high surface area and high porosity of the nanofibres, which are able to provide higher drug encapsula‐ tion efficiency and better stability, compared to other systems. An electrospun scaffold for anticancer drug delivery was tested using a polyethylene oxide (PEO)/chitosan (CS)/GO nanocomposite [44]. In the nanofibre carrier system, GOs act as nanocarriers which control doxorubicin (DOX) release, an anthracycline antibiotic used in cancer chemotherapy intrave‐ nous administration. DOX and GO surface can make a strong bond through π–π stacking interactions as a non-covalent type of functionalization that may provide controlled release of drug. GO also contains epoxy and carboxylic functional groups that can react with the amine groups of CS. The π–π stacking interaction between DOX and GO with the fine pores of nanofibrous scaffolds allowed high drug loadings (98%).A strong pH dependence was shown by the DOX release result at pH 5.3 and 7.4. At pH 5.3, faster drug release can be seen which may resulted from the unstability of hydrogen-bonding interaction between GO and DOX.

Along with the development of tissue engineering field, the **fabrication of scaffolds** is also advancing and succeeding in providing a mimic environment to the real body system. Currently, combining biomaterial with living cells for tissue scaffolds is a critical approach. Luo et al., in 2015, have reported a fabrication of GO-doped ply (lactic-co-glycolic acid) (PLGA) nanofibrous scaffold mixed with human MSCs for osteogenic differentiation [45]. The porous structure and the diameter of produced fibres are similarto the topological structure of natural extracellular matrices (ECM), and therefore facilitating cell attachment and proliferation. GO plays two roles in this scaffold. First, it enhances the hydrophilic performance and proteinand inducer-adsorption ability of the nanofibres. Second, it accelerates the human MSCs adhesion, proliferation and differentiation towards osteoblasts.

Graphene-based materials have extraordinary electronic transport properties, strong mechan‐ ical strength, excellent electrocatalytic activities, high surface area, ease of functionalization, and thus have contributed to ion detector and **biosensing** applications. A sensor for Cu (II) detector based on rGO/polyvinyl butyral (PVB) nanofibres was fabricated by Ding and coworkers in 2015. Electrospun GO/PVB on glassy carbon electrode (GCE) was reduced through an electrochemical method, to form reduced graphene oxide (RGO)/PVB nanofibres [46]. Cu (II) is one of the heavy metals that contaminate water sources through casual discharge from mining, machinery manufacturing and metal smelting. Although copper is an essential

element in human body, the excessive intake of copper might results in Wilson's disease and Menke's syndrome [46]. The rGO/PVB nanofibres modified glassy carbon electrode (GCE) were used for Cu (II) detection by differential pulse anodic stripping voltammetry (DPASV) [46]. The as-fabricated sensor based on rGO/PVB nanofibres showed good analytical perform‐ ance with the linear range of 0.06−22 μM, a low detection limit of 4.1 nM (S/N=3), high sensitivity of 103.51 μA **∙**μM−1 cm−2, good selectivity and excellent reproducibility (RSD = 0.49%) [46].

(Young's modulus) and scaffolds hydrophilicity increased as GO amount increased. The suture retention strength and burst pressure of tubular TPU/GO scaffolds containing 0.5 wt. %. GO exceeded the mechanical requirements of human blood vessels. In addition, results of platelet adhesion tests indicated that a low loading of GO is unlikely to cause thrombosis on the scaffolds. Endothelial cells could form a layer on the inner surface of the tubular scaf‐ folds which resembles the native blood vessel structure, indicating that the TPU/GO electro‐ spun scaffolds, with low GO loading level, have the potential to be used in small diameter

Another impressive function of scaffolds that needs to be expanded is the ability to **control drug release** onto specific tissues or organs. Atargeteddrugdelivery system that can efficiently transport drugs is a way for a scaffold to improve the efficacy of the drug and to reduce the systemic toxicity. Electrospun nanofibres are very important materials due to the high surface area and high porosity of the nanofibres, which are able to provide higher drug encapsula‐ tion efficiency and better stability, compared to other systems. An electrospun scaffold for anticancer drug delivery was tested using a polyethylene oxide (PEO)/chitosan (CS)/GO nanocomposite [44]. In the nanofibre carrier system, GOs act as nanocarriers which control doxorubicin (DOX) release, an anthracycline antibiotic used in cancer chemotherapy intrave‐ nous administration. DOX and GO surface can make a strong bond through π–π stacking interactions as a non-covalent type of functionalization that may provide controlled release of drug. GO also contains epoxy and carboxylic functional groups that can react with the amine groups of CS. The π–π stacking interaction between DOX and GO with the fine pores of nanofibrous scaffolds allowed high drug loadings (98%).A strong pH dependence was shown by the DOX release result at pH 5.3 and 7.4. At pH 5.3, faster drug release can be seen which may resulted from the unstability of hydrogen-bonding interaction between GO and DOX.

Along with the development of tissue engineering field, the **fabrication of scaffolds** is also advancing and succeeding in providing a mimic environment to the real body system. Currently, combining biomaterial with living cells for tissue scaffolds is a critical approach. Luo et al., in 2015, have reported a fabrication of GO-doped ply (lactic-co-glycolic acid) (PLGA) nanofibrous scaffold mixed with human MSCs for osteogenic differentiation [45]. The porous structure and the diameter of produced fibres are similarto the topological structure of natural extracellular matrices (ECM), and therefore facilitating cell attachment and proliferation. GO plays two roles in this scaffold. First, it enhances the hydrophilic performance and proteinand inducer-adsorption ability of the nanofibres. Second, it accelerates the human MSCs

Graphene-based materials have extraordinary electronic transport properties, strong mechan‐ ical strength, excellent electrocatalytic activities, high surface area, ease of functionalization, and thus have contributed to ion detector and **biosensing** applications. A sensor for Cu (II) detector based on rGO/polyvinyl butyral (PVB) nanofibres was fabricated by Ding and coworkers in 2015. Electrospun GO/PVB on glassy carbon electrode (GCE) was reduced through an electrochemical method, to form reduced graphene oxide (RGO)/PVB nanofibres [46]. Cu (II) is one of the heavy metals that contaminate water sources through casual discharge from mining, machinery manufacturing and metal smelting. Although copper is an essential

adhesion, proliferation and differentiation towards osteoblasts.

vascular graft applications.

112 Advances in Carbon Nanostructures

A novel platform for enhanced biosensing of electrospun PVA/CS/GO was also engineered [47]. Authors used glucose oxidase (GOD) as a model enzyme. Nanofibres produced from electrospinning are an ideal biosensing platform because ofthe large surface area they provide, leading to high enzyme loading capacity and low hindrance for mass transfer. Electrospin‐ ning is also a simple, durable and adjustable prove to provide stability and reproducibility of biosensors. GOs can bind enzyme molecules well, through electrostatic interactions, and thus avoiding the leaching of enzymes from modified electrodes. Nanomaterials are needed in nanofibre platform to bind the enzyme, since accessibility of the substrate to enzyme will be inhibited when the enzyme is confined inside the nanofibres. The fabricated biosensing platform displayed optimal oxidation and reduction capabilities towards electro-active species, even when compared to the bare Pt electrode, due to the synergy effects of PVA/ chitosan nanofibres and GO nanosheets [47].

**Power generation** is one of the essential needs of human daily life. The increasing demand for green and rechargeable energy sources to power up many electronic devices urged into a variety of power storage system alternatives. Knowing the superior properties of graphene and high specific surface area of CNFs, a number of attempts have been made to incorporate these materials into the fabrication of electrodes for Li-batteries and supercapacitors. Current freestanding anode for Li-ion batteries of graphite has limited theoretical capacity (372 mA h g−1). Xu et al., in 2014, proposed the use of Si, a nontoxic and abundant element, as the anode, due to its highest ever-known theoretical capacity up to 4200 mA h g−1 [48]. However, Si has its limitations, as it can undergo up to 280% volumetric expansion upon lithiation, which leads to electrode pulverization and loss of electrical contacts, resulting in severe capacity fading. The solution to this problem is to convert bulk Si into particles of low dimensions and to uniformly disperse Si in an inactive and conductive matrix, like GO. Co-axial nozzle was used to electrospun Si/CNF/GOanode, in orderto avoidagglomerationduring the spinning process. The resulting electrode achieved a high rate capacity of 567 mA h g−1 at 1 mA h g−1 current density. It obtained excellent cyclic performance with a discharge capacity of 872 mA h g−1after 50 cycles and 91% capacity retention. Another approach is the synthesis of core-shell struc‐ tured nanofibres that consist of rGO/PAN (shell) and ZnO/carrier polymer (core) [49]. The current use of ZnO has actually gained a good theoretical capacity of 978 mA h g−1. Unfortu‐ nately, it also carries a huge volume expansion for approximately 228%, which causes damage of mechanical integrity. Considering the same solutions, Shilpa et al., in 2015, encapsulated ZnO nanoparticles in the hollow core of glassy carbon–reduced graphene oxide (C–rGO) where the void spaces should provide a buffer zone, accommodating the volume changes associated with lithiation/delithiation of the metal oxide and prevent attrition of nanoparti‐ cles by detachment and fragmentation. The outcome of this study gained capacitance of 815 mA h g−1 at a current density of 50 mA g−1 with a capacity retention of almost 80% after 100 cycles [49]. A freestanding electrode eliminates the use of current collector, which is an inactive material that only adds to the battery mass and volume.

Another CNF composite in form of fibre mats was prepared by electrospinning blend of GO and PAN in *N*,*N*-dimethylformamide (DMF) for supercapacitor electrodes [50]. Porous carbon material of CNFs was further improved by graphene-based electrical conductivity, through carbonization, to obtain CNF/graphene composites. GO concentration was the major factor responsible for the morphology and pore structures, where the optimized-specific surface area and conductivity were achieved at 5 wt.%. The electrode material (5 wt.% GO) exhibited enhanced electrochemical performance with a highest specific capacitance of 146.62 F g−1, and highest energy density of 14.83–19.44 W h kg−1 in the power density range of 400– 20,000 W kg−1. Over the last decades, there was another upcoming power storage system that has been successfully applied in numerous fields. **Latent heat storage** using phase change material (PCM) makes use of the solid–liquid phase change of a material in order to store thermal energy. Fatty acid esters showed potential significant energy storage ability, as organic PCMs, with satisfactory thermal properties, thermal conductivity and thermal reliability. Nevertheless, the use of organic PCMs may undergo energy leakage. The synthesis of PCM with a supporting material is crucial to overcome the leakage problem and im‐ prove its thermal conductivity. Ke et al., in 2014, report on the preparation of methyl stearate (MES) fatty acid ester with PAN/GO by electrospinning [51]. GO was selected as the heat transfer/diffusion promoter and also as supporting material in the phase change composite. The results suggest that by adding GO, beneficial effects on mechanical properties, thermal stability and thermal energy storage/release performance were gained.

A work has been published dealing with the use of nylon 6-6 excellent mechanical resistance, with functionalized GO as a **corrosion protection** composite coating [52]. 2% GO loading in the electrospun nylon 6-6 possessed coating capacitance in the order of 10−7 F/cm2 , compared to 10−4 F/cm2 for coating prepared by deposit. This value suggests charge storage capacity for the coating condition. Thus, the efficiency of deposited coating is better than electrospun coating. The well-known physicochemical properties of PU have been widely used in coating technology. However, its low hydrophilicity and mechanical properties are the main set‐ backs for its application. Meanwhile, GO-polymer composite fibres emerged as the subject of enormous interest because polymer nanocomposites, with graphene materials as fillers, have shown dramatic improvements in such needed properties. Electropinning of in situ synthe‐ sized PU/GO has been conducted for surface coating of metallic stents [53]. This study showed that the incorporation of small amounts of GO sheets through PU nanofibres could enhance the stability of coated layer on the surface of nonvascular stent. The grafted PU chains on the surface of GO can facilitate the beneficial stress transfer from the PU matrix to the GO sheets. Moreover, proper incorporation of GO sheets into the polymer matrix can improve the hydrophilicity of the materials, without any adverse biological effect. These improved properties of PU mat, induced by GO, make it a potential candidate for surface coating of medical devices.

In order to fabricate **water-swellable rubber** (WSR), a multi-scale hybrid nanofibre with enhanced elasticity, resilience, toughness and water swelling ability was electrospun [54]. WSR can be produced by chemical and physical methods. The chemical method is more complicat‐ ed and more environmentally unfriendly than the physical method. In the physical method, common hydrophobic rubbers are mixed with various water-absorbent materials, known as superabsorbent polymers (SAPs) (hydrogel form). Hydrophilic super water-absorbent resin does not disperse well in hydrophobic rubber, expecting easy break off of physical-synthe‐ sized WSR. Eventually, swelling ability will lose, strength is remarkably decreased, and water is polluted. The addition of hydrophilic fibres can fix then as they can act as internal water channels in hydrophobic materials for superabsorbent nanocomposites. These fibres can help to transfer waterfrom the surface of the rubber matrix to the hydrophilic SAP particles, as well as between encapsulated SAP particles in the rubber matrix, enhancing the water swelling ability of WSR composites. Fabrication of PAA/ethylene glycol (EG)/hyperbranched bis-MPA polyester-64-hydoxyl (HB)/GO showed synergistic effects on water swelling ability com‐ pared to without HB and GO. The electrospun fibre mats were added into conventionalrubber by simple physical mixing, and the rubber composites showed considerable enhancement of water swelling ability, mainly due to the water channels built.

Functional **shape memory** composite nanofibres with GO have also been fabricated with PU shape memory [55]. Shape memory PU/GO (SMPU/GO) has better shape memory effect (SE) and lower thermal shrinkage, compared to SMPU nanofibre mats. When the loading amount of GO increased to 4 wt.%, the thermal shrinkage ratio of the composite nanofibrous mats reached as low as 4.7 ± 0.3%, while the average fixation ratio and recovery ratio were as high as 92.1 and 96.5%, respectively. The study indicates that GO is a desirable reinforcing filler for preparing shape memory nanofibres with improved properties. Electrospun polymer with GO and modified GO fillers have been extensively investigated for a wide range of applications. Apart from all applications mentioned, electrospinning of polymer/GO also stands out as an optical element in fibre lasers [56] and photocatalysts [57], showing their important role in the area of photonics.

## **5. Conclusions**

cles by detachment and fragmentation. The outcome of this study gained capacitance of 815 mA h g−1 at a current density of 50 mA g−1 with a capacity retention of almost 80% after 100 cycles [49]. A freestanding electrode eliminates the use of current collector, which is an inactive

Another CNF composite in form of fibre mats was prepared by electrospinning blend of GO and PAN in *N*,*N*-dimethylformamide (DMF) for supercapacitor electrodes [50]. Porous carbon material of CNFs was further improved by graphene-based electrical conductivity, through carbonization, to obtain CNF/graphene composites. GO concentration was the major factor responsible for the morphology and pore structures, where the optimized-specific surface area and conductivity were achieved at 5 wt.%. The electrode material (5 wt.% GO) exhibited enhanced electrochemical performance with a highest specific capacitance of 146.62 F g−1, and highest energy density of 14.83–19.44 W h kg−1 in the power density range of 400– 20,000 W kg−1. Over the last decades, there was another upcoming power storage system that has been successfully applied in numerous fields. **Latent heat storage** using phase change material (PCM) makes use of the solid–liquid phase change of a material in order to store thermal energy. Fatty acid esters showed potential significant energy storage ability, as organic PCMs, with satisfactory thermal properties, thermal conductivity and thermal reliability. Nevertheless, the use of organic PCMs may undergo energy leakage. The synthesis of PCM with a supporting material is crucial to overcome the leakage problem and im‐ prove its thermal conductivity. Ke et al., in 2014, report on the preparation of methyl stearate (MES) fatty acid ester with PAN/GO by electrospinning [51]. GO was selected as the heat transfer/diffusion promoter and also as supporting material in the phase change composite. The results suggest that by adding GO, beneficial effects on mechanical properties, thermal

material that only adds to the battery mass and volume.

114 Advances in Carbon Nanostructures

stability and thermal energy storage/release performance were gained.

medical devices.

A work has been published dealing with the use of nylon 6-6 excellent mechanical resistance, with functionalized GO as a **corrosion protection** composite coating [52]. 2% GO loading in

to 10−4 F/cm2 for coating prepared by deposit. This value suggests charge storage capacity for the coating condition. Thus, the efficiency of deposited coating is better than electrospun coating. The well-known physicochemical properties of PU have been widely used in coating technology. However, its low hydrophilicity and mechanical properties are the main set‐ backs for its application. Meanwhile, GO-polymer composite fibres emerged as the subject of enormous interest because polymer nanocomposites, with graphene materials as fillers, have shown dramatic improvements in such needed properties. Electropinning of in situ synthe‐ sized PU/GO has been conducted for surface coating of metallic stents [53]. This study showed that the incorporation of small amounts of GO sheets through PU nanofibres could enhance the stability of coated layer on the surface of nonvascular stent. The grafted PU chains on the surface of GO can facilitate the beneficial stress transfer from the PU matrix to the GO sheets. Moreover, proper incorporation of GO sheets into the polymer matrix can improve the hydrophilicity of the materials, without any adverse biological effect. These improved properties of PU mat, induced by GO, make it a potential candidate for surface coating of

, compared

the electrospun nylon 6-6 possessed coating capacitance in the order of 10−7 F/cm2

Graphene oxide and its derivatives show comparable properties as alternatives to graphene. Their outstanding performance can easily be manipulated into a wide variety of applica‐ tions. An important highlight is that several researchers have shown that the biocompatibili‐ ty of produced composites is still preserved, with no significant cytotoxicity appearing at low GO loading. Utilizing electrospinning to fabricate polymer/GO composites is a wise choice to obtain high specific surface area fibres at the nanoscale. It displays a promising advantage in achieving optimum properties out of the electrospun composites. Nonetheless, further research towards commercializing polymer/GO electrospun fibres in applications and improvement on electrospinning machines needs to be done, so that one can make use of their outstanding advantages to a larger extent.

## **Author details**

Izzati Fatimah Wahab1,2, Saiful Izwan Abd Razak1,2\*, Nor Syuhada Azmi2 , Farah Nuruljannah Dahli3 , Abdul Halim Mohd Yusof3 and Nadirul Hasraf Mat Nayan4

\*Address all correspondence to: saifulizwan@utm.my

1 IJN-UTM Cardiovascular Engineering Centre, Universiti Teknologi Malaysia, Skudai, Johor, Malaysia

2 Faculty of Biosciences and Medical Engineering, Universiti Teknologi Malaysia, Skudai, Johor, Malaysia

3 Faculty of Chemical and Energy Engineering, Universiti Teknologi Malaysia, Skudai, Johor, Malaysia

4 Faculty of Engineering Technology, Universiti Tun Hussein Onn Malaysia, Batu Pahat, Johor, Malaysia

## **References**


[8] D. R. Dreyer, S. Park, C. W. Bielawski, and R. S. Ruoff, The chemistry of graphene oxide, *Chem. Soc. Rev*., vol. 39, no. 1, pp. 228–240, 2010.

**Author details**

116 Advances in Carbon Nanostructures

Malaysia

Johor, Malaysia

Johor, Malaysia

Johor, Malaysia

**References**

112, 2014.

161, 2014.

2, pp. 101–113, 2008.

no. 3, pp. 3060–3068, 2014.

Farah Nuruljannah Dahli3

Izzati Fatimah Wahab1,2, Saiful Izwan Abd Razak1,2\*, Nor Syuhada Azmi2

\*Address all correspondence to: saifulizwan@utm.my

, Abdul Halim Mohd Yusof3

1 IJN-UTM Cardiovascular Engineering Centre, Universiti Teknologi Malaysia, Skudai, Johor,

2 Faculty of Biosciences and Medical Engineering, Universiti Teknologi Malaysia, Skudai,

3 Faculty of Chemical and Energy Engineering, Universiti Teknologi Malaysia, Skudai,

4 Faculty of Engineering Technology, Universiti Tun Hussein Onn Malaysia, Batu Pahat,

[1] D. Li, M. B. Mueller, S. Gilje, R. B. Kaner, and G. G. Wallace, Processable aqueous dispersions of graphene nanosheets, Nat. Nanotechnol., vol. 3, no. 2, pp. 101–105, 2008.

[2] D. Konios, M. M. Stylianakis, E. Stratakis, and E. Kymakis, Dispersion behaviour of graphene oxide and reduced graphene oxide, *J. Colloid Interface Sci*., vol. 430, pp. 108–

[3] J. Li, X. Zeng, T. Ren, and E. van der Heide, The preparation of graphene oxide and its derivatives and their application in bio-tribological systems, *Lubricants*, vol. 2, pp. 137–

[4] G. L. Li, G. Liu, M. Li, D. Wan, K. G. Neoh, and E. T. Kang, Organo- and waterdispersible graphene oxide-polymer nanosheets for organic electronic memory and

[5] T. Kuilla, S. Bhadra, D. H. Yao, N. H. Kim, S. Bose, and J. H. Lee, Recent advances in graphenebasedpolymercomposites,*Prog.Polym.Sci*.,vol.35,no.11,pp.1350–1375,2010.

[6] A. Fazal and K. S. Fancey, Performance enhancement of nylon kevlar fiber compo‐ sites through viscoelastically generated pre-stress, *Polym. Polym. Compos*., vol. 16, no.

[7] A. M. Dimiev, J. M. Tour, Mechanism of graphene oxide formation, *ACS Nano*, vol. 8,

gold nanocomposites, *J. Phys. Chem. C*, vol. 114, no. 29, pp. 12742–12748, 2010.

,

and Nadirul Hasraf Mat Nayan4


[34] H. R. Pant, C. H. Park, L. D. Tijing, A. Amarjargal, D.-H. Lee, and C. S. Kim, Bimodal fiber diameter distributed graphene oxide/nylon-6 composite nanofibrous mats via electrospinning, *Colloids Surfaces A Physicochem. Eng. Asp*., vol. 407, pp. 121–125, 2012.

[22] M. K. Leach, Z.-Q. Feng, S. J. Tuck, and J. M. Corey, Electrospinning fundamentals: optimizing solution and apparatus parameters, *J. Vis. Exp*., no. 47, p. e2494, 2011.

[23] Z. M. Huang, Y. Z. Zhang, M. Kotaki, and S. Ramakrishna, A review on polymer nanofibers by electrospinning and their applications in nanocomposites, *Compos. Sci.*

[24] H. S. Park, B. G. Choi, W. H. Hong, and S.-Y. Jang, Controlled assembly of graphene oxide nanosheets within one-dimensional polymer nanostructure, *J. Colloid Interface*

[25] Y. Wang, J. Tang, S. Xie, J. Liu, Z. Xin, X. Liu, and L. A. Belfiore, Leveling graphene sheet through electrospinning and its conductivity, *RSC Adv*., vol. 5, no. 52, pp. 42174–

[26] S. Ramazani and M. Karimi, Electrospinning of Poly(e-caprolactone) solutions containing graphene oxide: effects of graphene oxide content and oxidation level,

[27] C. L. Huang, S. Y. Peng, Y. J. Wang, W. C. Chen, and J. H. Lin, Morphology of Poly(vinyl alcohol)/Graphene Nanosheets Composite Nanofibers, *Appl. Mech. Mater*., vol. 365–366,

[28] D. Papkov, A. Goponenko, O. C. Compton, Z. An, A. Moravsky, X. Z. Li, S. T. Nguyen, and Y. A. Dzenis, Improved graphitic structure of continuous carbon nanofibers via

[29] Q. Wang, Y. Du, Q. Feng, F. Huang, K. Lu, J. Liu, and Q. Wei, Nanostructures and surface nanomechanical properties of polyacrylonitrile/graphene oxide composite nanofibers by electrospinning, *J. Appl. Polym. Sci*., vol. 128, no. 2, pp. 1152–1157, 2013.

[30] S. Thampi, V. Muthuvijayan, and R. Parameswaran, Mechanical characterization of high-performance graphene oxide incorporated aligned fibroporous poly(carbonate urethane) membrane for potential biomedical applications, *J. Appl. Polym. Sci*., vol. 132,

[31] C. Wang, Y. Li, G. Ding, X. Xie, and M. Jiang, Preparation and characterization of graphene oxide/poly(vinyl alcohol) composite nanofibers via electrospinning, *J. Appl.*

[32] O. J. Yoon, C. Y. Jung, I. Y. Sohn, H. J. Kim, B. Hong, M. S. Jhon, and N. E. Lee, Nanocomposite nanofibers of poly(D, L-lactic-co-glycolic acid) and graphene oxide nanosheets, *Compos. Part A Appl. Sci. Manuf*., vol. 42, no. 12, pp. 1978–1984, 2011.

[33] Y.-H. Yu, C.-C. Chan, Y.-C. Lai, Y.-Y. Lin, Y.-C. Huang, W.-F. Chi, C.-W. Kuo, H.-M. Lin, and P.-C. Chen, Biocompatible electrospinning poly(vinyl alcohol) nanofibres embedded with graphene-based derivatives with enhanced conductivity, mechanical

strength and thermal stability, *RSC Adv*., vol. 4, no. 99, pp. 56373–56384, 2014.

graphene oxide templating, *Adv. Funct. Mater*., vol. 23, pp. 5763–5770, 2013.

*Technol*., vol. 63, no. 15, pp. 2223–2253, 2003.

*Polym. Compos*., vol. 37, no. 1, pp. 131–140, 2014.

*Polym. Sci*., vol. 127, no. 4, pp. 3026–3032, 2013.

*Sci*., vol. 406, pp. 24–29, 2013.

42177, 2015.

118 Advances in Carbon Nanostructures

pp. 1025–1029, 2013.

no. 16, p. n/a–n/a, 2015.


**Carbon Nanotubes and Diamonds**

[46] R. Ding, Z. Luo, X. Ma, X. Fan, L. Xue, X. Lin, and S. Chen, High sensitive sensor fabricated by reduced graphene oxide/polyvinyl butyral nanofibers for detecting Cu

[47] X. Su, J. Ren, X. Meng, X. Ren, and F. Tang, A novel platform for enhanced biosens‐ ing based on the synergy effects of electrospun polymer nanofibers and graphene

[48] Z. L. Xu, B. Zhang, and J. K. Kim, Electrospun carbon nanofiber anodes containing monodispersed Si nanoparticles and graphene oxide with exceptional high rate

[49] Shilpa, B. M. Basavaraja, S. B. Majumder, and A. Sharma, "Lectrospun hollow glassy carbon—reduced graphene oxide nano fibers with encapsulated ZnO nanoparticles: a free standing anode for Li-ion, *J. Mater. Chem. A Mater. energy Sustain*., vol. 3, pp. 1–8,

[50] B.-H. Kim and K. S. Yang, Structure and electrochemical properties of electrospun carbon fiber composites containing graphene, *J. Ind. Eng. Chem*., vol. 20, no. 5, pp. 3474–

[51] H. Ke, Z. Pang, Y. Xu, X. Chen, J. Fu, Y. Cai, F. Huang, and Q. Wei, Graphene oxide improved thermal and mechanical properties of electrospun methyl stearate/polya‐ crylonitrile form-stable phase change composite nanofibers, *J Therm Anal Calorim*, vol.

[52] C. Menchaca-Campos, C. García-Pérez, I. Castañeda, M. A. García-Sánchez, R. Guardián, and J. Uruchurtu, Nylon/graphene oxide electrospun composite coating, *Int.*

[53] H. R. Pant, P. Pokharel, M. K. Joshi, S. Adhikari, H. J. Kim, C. H. Park, and C. S. Kim, Processing and characterization of electrospun graphene oxide/polyurethane compo‐

[54] J. Zhao, N. Dehbari, W. Han, L. Huang, and Y. Tang, Electrospun multi-scale hybrid nano fiber/net with enhanced water swelling ability in rubber composites, *Mater. Des.*,

[55] L. Tan, L. Gan, J. Hu, Y. Zhu, and J. Han, Functional shape memory composite nanofibers with graphene oxide filler, *Compos. Part A Appl. Sci. Manuf*., vol. 76, pp. 115–

[56] Q. Bao, H. Zhang, J. Yang, S. Wang, D. Y. Tang, R. Jose, S. Ramakrishna, C. T. Lim, and K. P. Loh, Graphene-polymer nanofiber membrane for ultrafast photonics, *Adv. Funct.*

[57] H. R. Pant, B. Pant, P. Pokharel, H. J. Kim, L. D. Tijing, C. H. Park, D. S. Lee, H. Y. Kim, and C. S. Kim, Photocatalytic Ti02-RGO/nylon-6 spider-wave-like nano-nets via electrospinning and hydrothermal treatment, *J. Memb. Sci*., vol. 429, pp. 225–234, 2013.

site nanofibers for stent coating, *Chem. Eng. J*., vol. 270, pp. 336–342, 2015.

(II) in water, *Int. J. Anal. Chem*., vol. 2015.

oxides., *Analyst*, vol. 138, no. 5, pp. 1459–66, 2013.

capacities, *Nano Energy*, vol. 6, pp. 27–35, 2014.

2015.

120 Advances in Carbon Nanostructures

3479, 2014.

117, pp. 109–122, 2014.

vol. 86, pp. 14–21, 2015.

123, 2015.

*J. Polym. Sci*., vol. 2013, pp. 1–9, 2013.

*Mater*., vol. 20, no. 5, pp. 782–791, 2010.

## **Flexible Low-Voltage Carbon Nanotube Heaters and their Applications**

Seyram Gbordzoe, Rachit Malik, Noe Alvarez, Robert Wolf and Vesselin Shanov

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64054

#### **Abstract**

Carbon nanotube heaters recently gained more attention due to their efficiency and relative ease of fabrication. In this chapter, we report on the design and fabrication of low-voltage carbon nanotube (CNT) heaters and their potential applications. CNT sheets drawn from CNT arrays have been used to make the heaters. The sheet resistance of the CNT sheet is dependent on the number of layers accumulated during their formation, and it ranges from 3.57 kΩ/sq. for a 1-layer sheet to 6.03 Ω/sq. for a 300-layer sheet. The fabricated and studied CNT heaters revealed fast heating and cooling rate. Potential applications of these heating devices have been illustrated by manufactur‐ ing and testing heatable gloves and via deicing experiments using low-voltage CNT heaters.

**Keywords:** CNT sheet, CNT heater, heated gloves

#### **1. Introduction**

Lightweight and low-power-consuming heaters that can be easily integrated in different materials and devices are needed for many industries including aerospace and medicine. The cold seasons usually bring the risk for several health problems, such as hypothermia, frostbite, flu, and even heart attacks. These adverse conditions make it difficult for people to conduct their daily routines effectively and safely during cold weather, due to the lack of proper gear to protect them. Some of the main attributes of a heating material, which can be safely incorporated into apparel, are lightweight, flexibility, fast heating rate, and energy

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

efficiency. CNT sheets meet most of these qualities and thus are potential candidates as heating materials.

Carbon nanotubes have been the subject of immense interest since their official discovery in 1991 [1]. They attracted much attention due to their unique electrical, structural, mechanical and thermal conductivities, and their multiple potential applications [2–5]. It has been shown that CNT arrays can be processed into yarns and sheets, which have excellent properties [6– 11]. One of the newer applications of CNT sheets is their use as a heating material [12–17]. The resistivity of CNT films can be controlled by the number of layers within the sheet or by the reagents used for their densification [17, 18]. CNTs can be incorporated into polymers to form composites for heating applications [16, 19]. However, CNT/polymer composites have a limited range of performance temperatures due to possible decomposition of the polymers [20]. Powdered CNTs dispersed in solvents have been used to coat cotton fabric and create CNT/textile heaters [21]. CNT-based heaters have also found applications as efficient and durable electrodes for Faradaic water splitting [22]. CNT sheet heaters have been fabricated as transparent film heaters, due to their high transmittance and good heating capabilities [12, 15–17]. CNT heaters have also demonstrated to be more efficient than commercial heating materials such as nichrome [18, 23]. Additionally, CNT-based heaters revealed rapid heating capabilities [23], which make them ideal candidates for applications in aerospace and heated clothing.

In this study, we fabricated CNT sheets from CNT arrays produced by the chemical vapor deposition (CVD) process. These sheets were densified by solvents and used for the manu‐ facture of CNT heaters. The goal of this work is to study how the heater's design (number of layers in the CNT sheets, their post-processing) influences its performance, including the voltage required to generate heat and the rapid heating response. Finally, some possible applications of the CNT heater are provided in this study.

## **2. Experimental methods**

CNT sheet fabrication started with the synthesis of vertically aligned, spinnable CNTs, typically about 400 μm in length. Thin films of Fe and Co were used as a catalyst (1.2 nm), which were sputtered on 4-inch Si wafers with 5-nm Al2O3 as a buffer sublayer. The ob‐ tained surface-engineered substrates were placed in a deposition reactor ET 3000 made by chemical vapor deposition (CVD) Equipment Corporation for the growth of the CNT arrays. Details of the growth process have been published elsewhere [24].

Upon completion of the CNT growth, detached, drawable CNT arrays were transferred onto an Si wafer or glass. CNT sheets with the desired number of layers were then assembled by continuously accumulating CNT ribbon onto a drum covered with a Teflon film as shown in **Figure 1**.

The resulting CNT sheets were densified, layer by layer, using acetone, as they were collect‐ ed onto the drum. This solvent was used because a previous work showed that it can signifi‐ cantly improve the electro-thermal behavior of CNT sheets [18]. Each revolution of the Teflon drum added one CNT layer of 50–70 nm thick ribbon onto the sheet. CNT sheets with various numbers of layers (1, 10, 100, 200 and 300) were produced by this technique.

**Figure 1.** CNT ribbons being collected on a Teflon-covered drum from a CNT array.

efficiency. CNT sheets meet most of these qualities and thus are potential candidates as

Carbon nanotubes have been the subject of immense interest since their official discovery in 1991 [1]. They attracted much attention due to their unique electrical, structural, mechanical and thermal conductivities, and their multiple potential applications [2–5]. It has been shown that CNT arrays can be processed into yarns and sheets, which have excellent properties [6– 11]. One of the newer applications of CNT sheets is their use as a heating material [12–17]. The resistivity of CNT films can be controlled by the number of layers within the sheet or by the reagents used for their densification [17, 18]. CNTs can be incorporated into polymers to form composites for heating applications [16, 19]. However, CNT/polymer composites have a limited range of performance temperatures due to possible decomposition of the polymers [20]. Powdered CNTs dispersed in solvents have been used to coat cotton fabric and create CNT/textile heaters [21]. CNT-based heaters have also found applications as efficient and durable electrodes for Faradaic water splitting [22]. CNT sheet heaters have been fabricated as transparent film heaters, due to their high transmittance and good heating capabilities [12, 15–17]. CNT heaters have also demonstrated to be more efficient than commercial heating materials such as nichrome [18, 23]. Additionally, CNT-based heaters revealed rapid heating capabilities [23], which make them ideal candidates for applications in aerospace and heated

In this study, we fabricated CNT sheets from CNT arrays produced by the chemical vapor deposition (CVD) process. These sheets were densified by solvents and used for the manu‐ facture of CNT heaters. The goal of this work is to study how the heater's design (number of layers in the CNT sheets, their post-processing) influences its performance, including the voltage required to generate heat and the rapid heating response. Finally, some possible

CNT sheet fabrication started with the synthesis of vertically aligned, spinnable CNTs, typically about 400 μm in length. Thin films of Fe and Co were used as a catalyst (1.2 nm), which were sputtered on 4-inch Si wafers with 5-nm Al2O3 as a buffer sublayer. The ob‐ tained surface-engineered substrates were placed in a deposition reactor ET 3000 made by chemical vapor deposition (CVD) Equipment Corporation for the growth of the CNT arrays.

Upon completion of the CNT growth, detached, drawable CNT arrays were transferred onto an Si wafer or glass. CNT sheets with the desired number of layers were then assembled by continuously accumulating CNT ribbon onto a drum covered with a Teflon film as shown in

The resulting CNT sheets were densified, layer by layer, using acetone, as they were collect‐ ed onto the drum. This solvent was used because a previous work showed that it can signifi‐

applications of the CNT heater are provided in this study.

Details of the growth process have been published elsewhere [24].

**2. Experimental methods**

heating materials.

124 Advances in Carbon Nanostructures

clothing.

**Figure 1**.

CNT sheets were transferred onto a polyethylene terephthalate (PET) tape and then cut with a laser machine (Micro Machining Systems by Oxford Lasers, wavelength of 532 nm). PET tape was used as a supporting substrate, because this material is thermally stable up to 200°C and makes handling the CNT heaters easier. Samples were cut into 10 mm × 60 mm dimensions with a laser power of 5% and at a speed of 1 mm/s. The laser machine was also used to cut copperfoils into 20 mm × 5 mm strips at a laser power of 40% and speed of 1 mm/s. The copper strips were attached at both ends of the CNT sheet with silver paste as shown in **Figure 2**, to form the CNT heater.

**Figure 2.** CNT sheet assembled between two polyethylene terephthalate (PET) tapes with Cu electrodes attached at both ends to form a CNT heater.

A DC power source (Hewlett Packard E 3612A) was used to supply power to the CNT heaters. The surface temperatures of the CNT heaters were measured using an infrared (IR) camera (FLIR T640). The average temperature was determined by measuring the surface tempera‐ ture at three different locations across the sheet. The effective area of the CNT sheet actually heated up was 10 mm × 50 mm for all samples. The voltage was stepped until pre-deter‐ mined heater temperatures were reached, and the corresponding current was measured. Measurements were taken at room temperature and pressure, after the heaters were allowed to reach an equilibrium temperature.

The sheetresistance ofthe CNTsheets was measuredusing a Jandel 4-probe instrument (Model RM3000). Prior to the measurement of the sheet resistance, samples were placed on a flat substrate, and then the sheet resistance was measured in the direction parallel and perpen‐ dicular to the alignment of the CNT sheet.

## **3. Results and discussion**

#### **3.1. Effect of the number of layers on sheet resistance**

The sheet resistance of samples was measured in both parallel and perpendicular to the drawing directions, using four probe techniques, as shown in **Figure 3**. This method in‐ volves passing current through the two outer probes, while the resulting voltage drop is measured by the two inner probes. The measured sheet resistance is presented in **Table 1**.

**Figure 3.** Scheme of the four-probe measurement applied to the CNT sheets in (a) parallel and (b) perpendicular to the drawing direction.


**Table 1.** Resistance of CNT sheet in parallel and perpendicular to the drawing direction.

A DC power source (Hewlett Packard E 3612A) was used to supply power to the CNT heaters. The surface temperatures of the CNT heaters were measured using an infrared (IR) camera (FLIR T640). The average temperature was determined by measuring the surface tempera‐ ture at three different locations across the sheet. The effective area of the CNT sheet actually heated up was 10 mm × 50 mm for all samples. The voltage was stepped until pre-deter‐ mined heater temperatures were reached, and the corresponding current was measured. Measurements were taken at room temperature and pressure, after the heaters were allowed

The sheetresistance ofthe CNTsheets was measuredusing a Jandel 4-probe instrument (Model RM3000). Prior to the measurement of the sheet resistance, samples were placed on a flat substrate, and then the sheet resistance was measured in the direction parallel and perpen‐

The sheet resistance of samples was measured in both parallel and perpendicular to the drawing directions, using four probe techniques, as shown in **Figure 3**. This method in‐ volves passing current through the two outer probes, while the resulting voltage drop is measured by the two inner probes. The measured sheet resistance is presented in **Table 1**.

**Figure 3.** Scheme of the four-probe measurement applied to the CNT sheets in (a) parallel and (b) perpendicular to the

to reach an equilibrium temperature.

126 Advances in Carbon Nanostructures

dicular to the alignment of the CNT sheet.

**3.1. Effect of the number of layers on sheet resistance**

**3. Results and discussion**

drawing direction.

As the number of layers increased, the sheet resistance decreased accordingly. A growth in the number of layers leads to an increase in the density of CNTs, as shown by SEM images in **Figure 4**. The increase in density leads to a reduction in the electrical resistance, because the tube-to-tube contacts were increased, allowing for more electrons to be transferred through the sample.

**Figure 4.** SEM images of (a) 1-, (b) 10- and (c) 100-layer-densified CNT sheets (high magnification in inset) showing an increase in density with increasing number of layers.

The parallel sheet resistance is lower than the perpendicular sheet resistance for 10, 100, 200 and 300 layers. This is because there is less resistance when electrons travel along the tube

length, compared to the lateral tube-to-tube interactions, which take place in the perpendicu‐ lar case. This is, however, not the case for a 1-layer sheet. As shown in **Figure 4a**, the proba‐ bility of the 4 probes hitting the same bundle of sheets, to measure the resistance, is lower in the perpendicular direction compared to the parallel direction. The anisotropy of the sheets is, however, lower than what has been in reported in some cases. Inoue et al. reported an anisotropy of 7.3 between the sheet resistances in the parallel direction compared to that in the perpendicular direction [25].

#### **3.2. Electrical properties of CNT heaters**

The CNT sheets were connected to a DC power source, and the corresponding temperature was measured using an IR camera in ambient environment, as shown in **Figure 5**. The heat generated across the sheet is as a result of the Joule heating phenomenon. Joule heating is a phenomenon that occurs when electrical current is passed through a material with an electrical resistance. The resistance inherent to the material leads to a conversion of electrical energy to thermal energy. This is caused by the collision of the moving electrons with the atoms that are the main material's constituents. The quantity of heat (*Q*) that is yielded from the applied electrical energy is proportional to the square of the current (*I*) multiplied by the resistance (*R*) over a period of time (*t*) as per Eq. (1).

$$Q = I^2 R t \tag{1}$$

**Figure 5.** Schematic illustration of the temperature acquisition setup using an infrared (IR) camera.

However, we usually use another parameter, known as power (*P*), to describe how much electrical energy is converted into thermal energy. This is simply the quantity of heat (*Q*) divided by the time (*t*) as per Eq. (1), which leads to Eq. (2).

Flexible Low-Voltage Carbon Nanotube Heaters and their Applications http://dx.doi.org/10.5772/64054 129

$$P = I^2 R = \frac{V^2}{R} \tag{2}$$

The power (*P*) generated by the CNT sheet when connected to a voltage source is also directly proportional to the voltage (*V*) across the sheet and the current (*I*) passing through the sheet (Eq. (2)). Different CNT heaters were connected to a DC power source, and the voltage was changed to correspond to temperatures 30, 50, 70, 90 and 110°C. The resultant current and voltage at each temperature were noted and used to calculate the corresponding power. The effect of the number of layers on the power needed to reach a particular temperature is illustrated in **Figure 6**. It can be seen that, as the number of layers increased, the power needed to generate a particular temperature also increased. This can be explained using Eq. (2), which shows that the resistance is inversely proportional to the power. An increase in the number of layers leads to a decrease in resistance, and this in turn leads to an increase in power needed to attain a particular temperature. Therefore, sheets with high resistance need a lower power to attain the same temperature as sheets with lower resistances.

length, compared to the lateral tube-to-tube interactions, which take place in the perpendicu‐ lar case. This is, however, not the case for a 1-layer sheet. As shown in **Figure 4a**, the proba‐ bility of the 4 probes hitting the same bundle of sheets, to measure the resistance, is lower in the perpendicular direction compared to the parallel direction. The anisotropy of the sheets is, however, lower than what has been in reported in some cases. Inoue et al. reported an anisotropy of 7.3 between the sheet resistances in the parallel direction compared to that in the

The CNT sheets were connected to a DC power source, and the corresponding temperature was measured using an IR camera in ambient environment, as shown in **Figure 5**. The heat generated across the sheet is as a result of the Joule heating phenomenon. Joule heating is a phenomenon that occurs when electrical current is passed through a material with an electrical resistance. The resistance inherent to the material leads to a conversion of electrical energy to thermal energy. This is caused by the collision of the moving electrons with the atoms that are the main material's constituents. The quantity of heat (*Q*) that is yielded from the applied electrical energy is proportional to the square of the current (*I*) multiplied by the resistance (*R*)

**Figure 5.** Schematic illustration of the temperature acquisition setup using an infrared (IR) camera.

divided by the time (*t*) as per Eq. (1), which leads to Eq. (2).

However, we usually use another parameter, known as power (*P*), to describe how much electrical energy is converted into thermal energy. This is simply the quantity of heat (*Q*)

<sup>2</sup> *Q I Rt* = (1)

perpendicular direction [25].

128 Advances in Carbon Nanostructures

**3.2. Electrical properties of CNT heaters**

over a period of time (*t*) as per Eq. (1).

**Figure 6.** Power versus temperature for different number of layers within the CNT sheet.

Equation 2 also shows that the higherthe resistance of the sheet, the greaterthe voltage needed to attain a specific temperature. This trend is clearly seen in **Figure 7**. The lower number layer sheets (1 and 10) needed very high voltages to attain the same temperature as the high number layer sheets (100, 200 and 300). For practical applications, the heater might be limited by the available low voltage supply. In such a case, CNT sheets with less number of layers would not be a feasible option, and higher number of layer sheets would be preferred.

**Figure 7.** Voltage versus temperature for different number of layers within the CNT sheet.

The electrical resistivity of the CNT sheets was also investigated in this study. The resistivity (*ρ*) is related to the sheet resistance (*R*s) and thickness (*t*) of the sheet by Eq. (3).

$$
\mathcal{P} = \frac{R\_S}{t} \tag{3}
$$

The resistivity of the CNT sheet can therefore be calculated if the thickness of the CNT sheet is known. Cross-sectional SEM images were taken on a selected number of samples to determine their thicknesses. The sheets were cut perpendicular to the drawing direction with a laser powered 0.5%, using a cutting speed of 0.1 mm/s.

The lower power and reduced speed of the laser were important to ensure a clean cut of the sheet and avoid pullouts of the CNTs disturbing the thickness uniformity. In order to increase the conductivity of the CNT sheets for SEM imaging, the samples were mounted on a vertical stub and coated with a thin layer of Au/Pd by a sputtering machine (Denton Vacuum Desk IV) using a sputtering time of 25 s. SEM images of the different sheet thicknesses are dis‐ played in **Figure 8**. The SEM images show the CNT samples with uniform thickness across the sheet. The average thickness of 100, 200 and 300 layer sheets was 6.25, 13.32 and 20.31 μm, respectively. The calculated resistivity in the parallel direction using Eq. (2) for 100, 200 and 300 layer sheets was 1.08 × 10−2 Ω cm, 1.32 × 10−2 Ω cm, and 1.23 × 10−2 Ω cm, respectively. The resistivity of the sheets is comparable to the data reported in the literature where MWCNT films have been studied revealing a resistivity of 20 × 10−3 Ω cm [26]. Our resistivity is higher than those sheets composed of single-wall carbon nanotubes (SWCNTs), reported to be in the range of 10−4 to 10−5 Ω cm [4]. The high resistivity in our work is due to severalfactors, including the packing density of the nanotubes and the nature of the MWCNTs, which are less conduc‐ tive than SWCNTs.

Flexible Low-Voltage Carbon Nanotube Heaters and their Applications http://dx.doi.org/10.5772/64054 131

**Figure 8.** Cross-sectional SEM images showing the thickness of: (a) and (b) 100-layer CNT sheet; (c) and (d) 200-layer CNT sheet; and (e) and (f) 300-layer CNT sheet.

It has been shown in the literature that an improved packing density (by stretching and pressing) leads to improvement in electrical and mechanical properties [7, 27]. In the future, we intend to use these two approaches in order to increase the density of our CNT sheets and decrease their resistivity.

#### **3.3. Temperature profile of CNT sheets**

**Figure 7.** Voltage versus temperature for different number of layers within the CNT sheet.

a laser powered 0.5%, using a cutting speed of 0.1 mm/s.

tive than SWCNTs.

130 Advances in Carbon Nanostructures

(*ρ*) is related to the sheet resistance (*R*s) and thickness (*t*) of the sheet by Eq. (3).

The electrical resistivity of the CNT sheets was also investigated in this study. The resistivity

The resistivity of the CNT sheet can therefore be calculated if the thickness of the CNT sheet is known. Cross-sectional SEM images were taken on a selected number of samples to determine their thicknesses. The sheets were cut perpendicular to the drawing direction with

The lower power and reduced speed of the laser were important to ensure a clean cut of the sheet and avoid pullouts of the CNTs disturbing the thickness uniformity. In order to increase the conductivity of the CNT sheets for SEM imaging, the samples were mounted on a vertical stub and coated with a thin layer of Au/Pd by a sputtering machine (Denton Vacuum Desk IV) using a sputtering time of 25 s. SEM images of the different sheet thicknesses are dis‐ played in **Figure 8**. The SEM images show the CNT samples with uniform thickness across the sheet. The average thickness of 100, 200 and 300 layer sheets was 6.25, 13.32 and 20.31 μm, respectively. The calculated resistivity in the parallel direction using Eq. (2) for 100, 200 and 300 layer sheets was 1.08 × 10−2 Ω cm, 1.32 × 10−2 Ω cm, and 1.23 × 10−2 Ω cm, respectively. The resistivity of the sheets is comparable to the data reported in the literature where MWCNT films have been studied revealing a resistivity of 20 × 10−3 Ω cm [26]. Our resistivity is higher than those sheets composed of single-wall carbon nanotubes (SWCNTs), reported to be in the range of 10−4 to 10−5 Ω cm [4]. The high resistivity in our work is due to severalfactors, including the packing density of the nanotubes and the nature of the MWCNTs, which are less conduc‐

(3)

The heat generated by the CNT heater, when connected to a power source, is lost into the ambient environment during heating. The total heat is dissipated into the air and substrate through three mechanisms: conduction, convection and radiation [12].

The heating profiles with time of CNT sheets at selected low voltages (1.5, 3 and 4.5 V) are presented in **Figure 9**. The obtained data mimic possible low-voltage applications of the heaters. Sheets with 1 and 10 layers have not been included because they show negligible increase in temperature at the applied voltages.

The observed temperature evolution is divided into 3 stages: heating, saturation, and cooling. Cooling happens after the voltage source is turned off. During the saturation stage, the heat gained by the applied electric power is almost equal to the heat lost by radiation and convec‐ tion. The average time for the heater to reach the temperature saturation is 20 s. However, the average cooling time depends on the final temperature of the heater and, therefore, varies, as shown in **Figure 9**. This fast heating time is a desirable characteristic of a heater that will be incorporated into clothing or textiles, to keep the body warm during cold seasons. This quality

is also beneficial for deicing applications. We present some feasible applications of the CNT heaters created and studied in this work.

**Figure 9.** Heating profile as a function of time for 100-, 200- and 300-layer CNT sheets at (a) 1.5 V, (b) 3 V, and (c) 4.5 V.

## **4. Applications of CNT heaters**

#### **4.1. Heatable glove**

CNT heaters are easy to design and manufacture, which makes them attractive for a variety of applications. Their lightweight, mechanical flexibility and low voltage make the CNT heaters appealing candidates for apparel, to protect operators performing in cold weather. We have fabricated CNTheaters and embeddedthem into gloves, based on the approach described in this work. The heaters were made using 200 and 150 layer sheets. Such gloves are light‐ weight since the CNT heaters do not add significant mass (total change of 1.21 g) and can also be operated with either a 9-V battery or a rechargeable Li-ion (2700 mAh, 3.7 V) battery, as shown in **Figure 10**. This type of application has advantages because of the ease of fabrica‐ tion of the CNT heaters and the low voltage needed to power them.

#### **4.2. Deicing applications**

The use of CNT heaters for deicing applications was also studied. This approach was similar to that described by Janas et al. [23]. A 300-layer CNT heater was attached to the right wing of a model airplane. The test model was inserted into a freezer with a continuous voltage (3V) supply attached to the heater. Droplets of water were then sprinkled on both wings of the model airplane, shown in **Figure 11a**, and it was left in the freezer. The thermal image before ice formation is shown in **Figure 11b**, where the right wing temperature is above 0°C. After

an hour of cooling, the wing without the CNT heater had visible ice formation on its surface, while the wing with the functioning heater was dry as shown in **Figure 11c**. The correspond‐ ing infrared image, after ice formation, is shown in **11d**, where the entire model airplane is frozen except for the right wing. This experiment shows that CNT heaters can be used in deicing applications to prevent the formation of ice on airplanes, which is a major issue in the aerospace industry. The heater can be incorporated into the surface of the fuselage, which is made of polymer composite for most of the military and some of the advanced commercial airplanes.

**Figure 10.** CNT-heated gloves connected to (a) 9-V battery and (b) rechargeable lithium ion battery. The color scale represents temperature in °C obtained by an IR camera.

**Figure 11.** (a) Droplets of water on the wings of a model airplane with related (b) IR thermal image; (c) ice formation on the left wing of the model airplane after an hour of freezing and related (d) IR thermal image.

## **5. Conclusions**

is also beneficial for deicing applications. We present some feasible applications of the CNT

**Figure 9.** Heating profile as a function of time for 100-, 200- and 300-layer CNT sheets at (a) 1.5 V, (b) 3 V, and (c) 4.5 V.

CNT heaters are easy to design and manufacture, which makes them attractive for a variety of applications. Their lightweight, mechanical flexibility and low voltage make the CNT heaters appealing candidates for apparel, to protect operators performing in cold weather. We have fabricated CNTheaters and embeddedthem into gloves, based on the approach described in this work. The heaters were made using 200 and 150 layer sheets. Such gloves are light‐ weight since the CNT heaters do not add significant mass (total change of 1.21 g) and can also be operated with either a 9-V battery or a rechargeable Li-ion (2700 mAh, 3.7 V) battery, as shown in **Figure 10**. This type of application has advantages because of the ease of fabrica‐

The use of CNT heaters for deicing applications was also studied. This approach was similar to that described by Janas et al. [23]. A 300-layer CNT heater was attached to the right wing of a model airplane. The test model was inserted into a freezer with a continuous voltage (3V) supply attached to the heater. Droplets of water were then sprinkled on both wings of the model airplane, shown in **Figure 11a**, and it was left in the freezer. The thermal image before ice formation is shown in **Figure 11b**, where the right wing temperature is above 0°C. After

tion of the CNT heaters and the low voltage needed to power them.

heaters created and studied in this work.

132 Advances in Carbon Nanostructures

**4. Applications of CNT heaters**

**4.1. Heatable glove**

**4.2. Deicing applications**

A simple and easy way for fabricating flexible and low-voltage CNT heaters was described. The resistance of the CNT sheets with different numbers of layers was studied. It was observed that as the number of layers increased, the sheet resistance decreased, showing that this

property is dependent on the thickness of the sheet. The heating and cooling rates of these heaters were also studied. The obtained results suggested that the fast heating rate of the heaters makes them attractive for manufacturing body protection clothing used in cold weather. An example ofthe latter was demonstrated in this work by incorporating CNT heaters into a glove. Maximum glove temperatures of 39°C and 31°C were attained using a 9-V battery and Li-ion battery, respectively. Another application demonstrated that the CNT heater can find application in deicing solid surfaces exposed to freezing weather, such as airplanes. The CNT sheet heaters can be easily incorporated in polymer composites recently used as materials of choice for making advanced cars and fuselages.

## **Acknowledgements**

This work was funded by the National Science Foundation (NSF) through the following grants: CMMI-0727250, SNM-1120382, ERC-0812348, and DURIP-ONR. The partial support by the National Institute for Occupational Safety and Health Pilot Research Project Training Program of the UC Education and Research Center Grant #T42/OH008432-10 is also acknowledged. The authors are grateful for the financial support from the UC Office of Research and the Mathewson Renewable Energy Research Fund. Seyram Gbordzoe is thankful for the gradu‐ ate assistantship provided to him by the College of Engineering and Applied Science (CEAS) at UC and the Cincinnati Engineering Enhanced Mathematics and Science (CEEMS) pro‐ gram (NSF, DRL-1102990).

## **Author details**

Seyram Gbordzoe1 , Rachit Malik1 , Noe Alvarez2 , Robert Wolf3 and Vesselin Shanov1,2\*

\*Address all correspondence to: vesselin.shanov@uc.edu

1 Department of Mechanical and Materials Science Engineering, University of Cincinnati, Cincinnati, OH, USA

2 Department of Biomedical, Chemical and Environmental Engineering, University of Cin‐ cinnati, Cincinnati, OH, USA

3 Department of Electrical Engineering, University of Cincinnati, Cincinnati, OH, USA

## **References**

[1] Iijima, S., *Helical microtubules of graphitic carbon*. Nature, 1991. 354(6348): pp. 56–58.

[2] Jung , D., M. Han, L.J. Overzet, and G.S. Lee, *Effect of hydrogen pretreatment on the spincapability of a multiwalled carbon nanotube forest*. Journal of Vacuum Science & Technol‐ ogy B, 2013. 31(6): p. 06FI02.

property is dependent on the thickness of the sheet. The heating and cooling rates of these heaters were also studied. The obtained results suggested that the fast heating rate of the heaters makes them attractive for manufacturing body protection clothing used in cold weather. An example ofthe latter was demonstrated in this work by incorporating CNT heaters into a glove. Maximum glove temperatures of 39°C and 31°C were attained using a 9-V battery and Li-ion battery, respectively. Another application demonstrated that the CNT heater can find application in deicing solid surfaces exposed to freezing weather, such as airplanes. The CNT sheet heaters can be easily incorporated in polymer composites recently used as materials

This work was funded by the National Science Foundation (NSF) through the following grants: CMMI-0727250, SNM-1120382, ERC-0812348, and DURIP-ONR. The partial support by the National Institute for Occupational Safety and Health Pilot Research Project Training Program of the UC Education and Research Center Grant #T42/OH008432-10 is also acknowledged. The authors are grateful for the financial support from the UC Office of Research and the Mathewson Renewable Energy Research Fund. Seyram Gbordzoe is thankful for the gradu‐ ate assistantship provided to him by the College of Engineering and Applied Science (CEAS) at UC and the Cincinnati Engineering Enhanced Mathematics and Science (CEEMS) pro‐

, Noe Alvarez2

1 Department of Mechanical and Materials Science Engineering, University of Cincinnati,

3 Department of Electrical Engineering, University of Cincinnati, Cincinnati, OH, USA

2 Department of Biomedical, Chemical and Environmental Engineering, University of Cin‐

[1] Iijima, S., *Helical microtubules of graphitic carbon*. Nature, 1991. 354(6348): pp. 56–58.

, Robert Wolf3

and Vesselin Shanov1,2\*

of choice for making advanced cars and fuselages.

, Rachit Malik1

\*Address all correspondence to: vesselin.shanov@uc.edu

**Acknowledgements**

134 Advances in Carbon Nanostructures

gram (NSF, DRL-1102990).

**Author details**

Seyram Gbordzoe1

Cincinnati, OH, USA

**References**

cinnati, Cincinnati, OH, USA


## **Recent Trends of Reinforcement of Cement with Carbon Nanotubes and Fibers**

Oxana V. Kharissova, Leticia M. Torres Martínez and Boris I. Kharisov

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/62292

#### **Abstract**

[16] Kim, D., L. Zhu, D.-J. Jeong, K. Chun, Y.-Y. Bang, S.-R. Kim, J.-H. Kim, and S.-K. Oh, *Transparent flexible heater based on hybrid of carbon nanotubes and silver nanowires*. Carbon,

[17] Jung, D., D. Kim, K.H. Lee, L.J. Overzet, and G.S. Lee, *Transparent film heaters using multi-walled carbon nanotube sheets*. Sensors and Actuators A: Physical, 2013. 199: pp.

[18] Dawid Janas and Krzysztof K. Koziol, "Improved Performance of Ultra-Fast Carbon Nanotube Film Heaters," Jounal of Automation and Control Engineering, Vol. 2, No.

[19] Kunmo, C., K. Dongouk, S. Yoonchul, L. Sangeui, M. Changyoul, and P. Sunghoon, *Electrical and thermal properties of carbon-nanotube composite for flexible electric heating-unit*

[20] Lee, E. and Y.G. Jeong, *Preparation and characterization of sulfonated poly(1,3,4-oxadia‐ zole)/multi-walled carbon nanotube composite films with high performance in electric heating*

[21] Ilanchezhiyan, P., A.S. Zakirov, G.M. Kumar, S.U. Yuldashev, H.D. Cho, T.W. Kang, and A.T. Mamadalimov, *Highly efficient CNT functionalized cotton fabrics for flexible/*

[22] Janas, D., S. Kreft, S. Boncel, and K.K. Koziol, *Durability and surface chemistry of horizontally aligned CNT films as electrodes upon electrolysis of acidic aqueous solution*.

[23] Janas, D. and K.K. Koziol, *Rapid electrothermal response of high-temperature carbon*

[24] Alvarez, N.T., P. Miller, M. Haase, N. Kienzle, L. Zhang, M.J. Schulz, and V. Shanov, *Carbon nanotube assembly at near-industrial natural-fiber spinning rates*. Carbon, 2015. 86:

[25] Inoue, Y., Y. Suzuki, Y. Minami, J. Muramatsu, Y. Shimamura, K. Suzuki, A. Ghemes, M. Okada, S. Sakakibara, and H. Mimura, *Anisotropic carbon nanotube papers fabricated*

[26] Bacsa, W., A. Chatelain, T. Gerfin, R. Humphrey-Baker, L. Forro, and D. Ugarte, *Aligned carbon nanotube films: production and optical and electronic properties*. Science, 1995.

[27] Cheng, Q., J. Bao, J. Park, Z. Liang, C. Zhang, and B. Wang, *High mechanical perform‐ ance composite conductor: multi‐walled carbon nanotube sheet/bismaleimide nanocomposites*.

*from multiwalled carbon nanotube webs*. Carbon, 2011. 49(7): pp. 2437–2443.

Advanced Functional Materials, 2009. 19(20): pp. 3219–3225.

*and thermal stability*. Composites Part B: Engineering, 2015. 77: pp. 162–168.

*wearable heating applications*. RSC Advances, 2015. 5(14): pp. 10697–10702.

2, pp. 150–153 June, 2014. doi: 10.12720/joace.2.2.150–153.

Journal of Materials Science, 2014. 49(20): pp. 7231–7243.

*nanotube film heaters*. Carbon, 2013. 59: pp. 457–463.

*applications*. Electron Device Letters, IEEE, 2013. 34(5): pp. 668–670.

2013. 63: pp. 530–536.

176–180.

136 Advances in Carbon Nanostructures

pp. 350–357.

268(5212): pp. 845–847.

Recent achievements in the area of formation of carbon nanotubes (CNTs), nanocompo‐ sites, with cement are reviewed. The peculiarities of dispersion of CNTs in cementi‐ tious matrices are discussed, paying major attention to the CNT diameter, length and length-to-diameter ratio, concentration, functionalization, annealing, combination with other nanomaterials, and water-cement ratio. Several effects upon dispersion of carbon allotropes in concrete-water media are emphasized. It is also pointed out that the health impact should also be considered in further experiments on construction materials reinforced with CNTs.

**Keywords:** Carbon fibers, Carbon nanotubes, Cement, Composite, Dispersion, Rein‐ forcement

### **1. Introduction**

Concrete, steel, and asphalt coatings are conventional materials used in large scale and produced in huge quantities worldwide. In case of concrete (a mixture weak in tension and strong in compression), more than 2 tons per person are produced annually. Quantitative‐ ly, more than 11 billion metric tons are consumed every year all over the world. The cement industry is responsible for approximately 5–8% of all anthropogenic emissions of carbon dioxide worldwide [1]. Portland cement (containing, in dry phase, 63% calcium oxide, 20% silica, 6% alumina, and 3% iron (III) oxide, and small amounts of other substances) is a product with great, but not completely explored potential, despite intensive studies in the

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

last century. In the last century, when the nanotechnology era began, at the end of the 1980s, novel nanomaterials started to be used as additives in numerous applications, in particular for reinforcement of construction materials, and certain attention was given to cement composites. Their complex structure, when studied at the nanolevel, will definitely lead to new generations of highly durable concretes with "smart" properties. In order to improve concrete properties, a series of nanomaterials can be applied, such as nanoengineered polymers, superplasticizers, high strength fibers, or silicon dioxide (**Figure 1**).

**Figure 1.** Particle size versus specific surface area scale in to concrete materials [2]. Reproduced with permission.

Carbon nanotubes (CNTs), well known for having extraordinary properties, are considered as major candidates for diverse applications in nanotechnology. CNTs have been incorporated into materials for cement production, with excellent results, upon low quantities of added material. Thus, reinforcing cement with low additions of multiwalled carbon nanotubes (MWCNTs), ranging from 0.05 to 0.5 wt%, can represent a remarkable enhancement in the mechanical properties of cement. It is expected that CNTs, when added to concrete, will increase compression strength beyond 200 MPa, thus allowing the construction of mile-high skyscrapers. Nowadays, CNTs and carbon nanofibers (CNF) are two of the most prospective advanced materials for application in cement-based products, for the construction industry, due to their excellent material properties. According to the reported results, CNTs can reduce the occurrence of cracks, decrease porosity, and improve mechanical properties, thus extend‐ ing the cement durability. However, CNTs are insoluble in organic solvents and water, so surfactants should be normally added upon classic combination of ultrasonication and vigorous agitation, in order to disperse CNTs. There are other well-known less common specific methods for CNT "dissolution" in liquid media [3, 4]; some of which can be adjust‐ ed to concrete fabrication.

In this chapter, we discuss recent achievements of cement reinforcement with carbon allo‐ tropes. Taking into account a recently published excellent comprehensive review [5] in this area and other related publications [6, 7], we refer the particular aspects and peculiarities of dispersion of CNTs in cement mixtures.

## **2. Peculiarities of CNTs and CNFs dispersion in concrete**

last century. In the last century, when the nanotechnology era began, at the end of the 1980s, novel nanomaterials started to be used as additives in numerous applications, in particular for reinforcement of construction materials, and certain attention was given to cement composites. Their complex structure, when studied at the nanolevel, will definitely lead to new generations of highly durable concretes with "smart" properties. In order to improve concrete properties, a series of nanomaterials can be applied, such as nanoengineered

polymers, superplasticizers, high strength fibers, or silicon dioxide (**Figure 1**).

138 Advances in Carbon Nanostructures

**Figure 1.** Particle size versus specific surface area scale in to concrete materials [2]. Reproduced with permission.

Carbon nanotubes (CNTs), well known for having extraordinary properties, are considered as major candidates for diverse applications in nanotechnology. CNTs have been incorporated into materials for cement production, with excellent results, upon low quantities of added material. Thus, reinforcing cement with low additions of multiwalled carbon nanotubes (MWCNTs), ranging from 0.05 to 0.5 wt%, can represent a remarkable enhancement in the mechanical properties of cement. It is expected that CNTs, when added to concrete, will increase compression strength beyond 200 MPa, thus allowing the construction of mile-high skyscrapers. Nowadays, CNTs and carbon nanofibers (CNF) are two of the most prospective advanced materials for application in cement-based products, for the construction industry, due to their excellent material properties. According to the reported results, CNTs can reduce the occurrence of cracks, decrease porosity, and improve mechanical properties, thus extend‐ ing the cement durability. However, CNTs are insoluble in organic solvents and water, so surfactants should be normally added upon classic combination of ultrasonication and There are at least three main ways to add CNTs uniformly to concrete: addition to cement, to water, and as admixture. In the majority ofreports, ultrasonic dispersion techniques have been widely used to uniformly disperse CNTs [8]. Ultrasonication can be applied alone after adding CNTs to the concrete mass, or together with surfactants. Generally, pristine CNTs can be directly mixed with cement, but their hydrophobicity limits solubilization in aqueous media. So, frequently the CNTs are functionalized with HNO*3*/H*2*SO*<sup>4</sup>* mixture, forming -OH and - COOH functionalized CNTs, before being added to cement [9]. Alternatively, carboxylic functionalization can be done upon ultrasonication and polyacrylic acid polymer addition [10]. The use of superplasticizers (see below), magnetic stirring [11], CNT growth onto cement particles [12], and dry mixing with cement [13] are other options to disperse CNTs and CNFs in concrete, frequently accompanied by ultrasonication [14].

Generally, two main strategies are applied to provide an interfacial bonding between cementitious matrix and CNTs: (a) formation of hydration products of the cement matrix and (b) formation of covalent bonds on CNT surface using functional groups. **Figure 2** shows a scheme of interaction of the C-S-H (this means calcium-silicate-hydrate) cement product and a -COOH (carboxylic group) on the CNT surface[15]. These effects of *surface functionalization* of MWCNTs with -COOH groups on the strength and structure of Portland cement compo‐ sites have been deeply studied [16]. Thus, grafting of functional groups on the surface of the nanotubes allowed acceleration of cement hydration. The authors established that the use of carboxylated nanotubes contributed to early strength development. The MWCNT-rein‐ forced composites are characterized by high content in calcium silicate hydrates and very dense structure. The maximal compressive strength of 64 MPa (20% increase overthe reference material) was observed for the composite with 0.13% of MWCNTs (by the weight of cement). The addition of carboxylated MWCNTs, in the very low amount of 0.05%, provides a 30% increase of 1-day compressive strength of developed composites.

In addition to the CNT functionalization above, *surfactants* can be applied to decrease aggregation of CNTs, by modification of their surface, since CNTs tend to aggregate togeth‐ er in most solvents, due to van der Waals forces, and form nanotube clusters and bundles. CNTs need to be completely dispersed in concrete in order to explore their outstanding physical properties. Application of classic surfactants, such as sodium dodecyl sulfate (SDS) and Triton X-100, has been observed in a series ofreports. The effect of surfactants on pressure-

**Figure 2.** Reaction scheme between carboxylated nanotubes and cement hydration products (Ca(OH)2 and C-S-H).

sensitivity of CNT-filled cement mortar composites has been studied using sodium dodecyl sulfate and sodium dodecylbenzene sulfonate (NaDDBS) as surfactants to disperse MWNTs-COOH (OD < 8 nm, ID 2–5 nm, length 10–30 μm) in cement mortar (**Figure 3**) [17]. The authors revealed that, in comparison with SDS-based composites, the composites with NaDDBS show a more stable and sensitive response of electrical resistance to external force. In this respect, NADDBS is a more efficient surfactant, in comparison with SDS, for dispersion of MWCNTs in cement.

Other nanotechnology classic surfactants are less common to use for CNT solubilization in concrete. Stable homogeneous suspensions ofMWCNTs were prepared using gum arabic (GA) as dispersant (with an optimum GA concentration of 0.45 g/L1 ) and were incorporated into Portland cement paste [18]. The authors revealed that the compressive strength, as well as the flexural strength of the Portland cement composite, can be considerably improved by adding the treated carbon nanotubes. At 0.08 wt% MWCNT concentration, the flexural strength reached an increase in 43.38%. MWCNTs act as bridges and networks across cracks and voids, which transfer the load in case of tension, and the interface bond strength between the

**Figure 3.** SEM images of CNT-filled cement mortar composites. (A) With NaDDBS and 1 wt% of MWNTs. (B) With SDS and 1 wt% of MWNTs. Reproduced with permission from [17].

nanotubes and matrix is very strong. The interfacial interactions between surface-modified nanotubes and hydrations (such as C-S-H and calcium hydroxide) of cement produce a high bonding strength, and increase the load-transfer efficiency from cement matrix to the rein‐ forcement.

Among superplasticizers for concrete applications, we note polycarboxylate ether solution ("super plasticizer"), high-molecular-weight polyelectrolyte (i.e., poly (sodium 4-styrenesul‐ fonate) or PSS) [19]. In a related report [20], the dispersity of MWCNTs in water and in cement paste, as well as its effect on the compressive strength of composition with MWCNTs and cement, was comparatively investigated using ultrasonic dispersion and three kinds of surface active agents: polycarboxylate superplasticizer, polyvinylpyrrolidone (PVP), and washing powder (WP). Agglomeration of MWCNTs takes place easily in cement paste and in water. The reinforce ability is poorly exhibited, even if the ultrasound is applied, in case of using, as a surface-active agent, a single polycarboxylate superplasticizer. WP and PVP are able to create MWCNT dispersions in cement paste and also in water under ultrasonic treatment; in addition, PVPenhances the compressive strength, whileWPdecreases it, as more air bubbles are brought inside the cement paste. In addition, several peculiarities exist on the dispersing carbon nanoparticles, in particular CNFs, in cement paste [21]. When a superplasticizer is used as surfactant, CNFs can be uniformly dispersed in water by ultrasonic processing. However, a uniform distribution of CNFs in cement paste will not be obtained by mixing water-super‐ plasticizer-CNF dispersion with cement (used superplasticizer = high-range polycarboxylatebased water-reducing agent). To reach a better distribution in paste, either functionalized and highly dispersible CNFs should be used or the CNFs should initially be implanted or grown on cement particles. In addition, large cement particles prevent a uniform distribution, particularly when fibers are very small or are used in high doses. However, breaking cement grains into smaller particles is impractical and problematic. Fine-grain cement is very reactive and consumes a lot of water at a high rate. As a result, producing paste with typical water-

sensitivity of CNT-filled cement mortar composites has been studied using sodium dodecyl sulfate and sodium dodecylbenzene sulfonate (NaDDBS) as surfactants to disperse MWNTs-COOH (OD < 8 nm, ID 2–5 nm, length 10–30 μm) in cement mortar (**Figure 3**) [17]. The authors revealed that, in comparison with SDS-based composites, the composites with NaDDBS show a more stable and sensitive response of electrical resistance to external force. In this respect, NADDBS is a more efficient surfactant, in comparison with SDS, for dispersion of MWCNTs

**Figure 2.** Reaction scheme between carboxylated nanotubes and cement hydration products (Ca(OH)2 and C-S-H).

Other nanotechnology classic surfactants are less common to use for CNT solubilization in concrete. Stable homogeneous suspensions ofMWCNTs were prepared using gum arabic (GA)

Portland cement paste [18]. The authors revealed that the compressive strength, as well as the flexural strength of the Portland cement composite, can be considerably improved by adding the treated carbon nanotubes. At 0.08 wt% MWCNT concentration, the flexural strength reached an increase in 43.38%. MWCNTs act as bridges and networks across cracks and voids, which transfer the load in case of tension, and the interface bond strength between the

) and were incorporated into

as dispersant (with an optimum GA concentration of 0.45 g/L1

in cement.

140 Advances in Carbon Nanostructures

cement ratios such as 0.40 or 0.45 is very difficult or even impossible. Therefore, the authors consider it important to use fresh cement with a minimal amount of large grains and clumps for making CNF-CNT reinforced composites.

## **3. Improvement of concrete properties by addition of CNTs and CNFs**

It is well-known nowadays that mechanical and various other properties of concrete, indicat‐ ed below, are considerably improved by addition of CNTs. Thus, the inclusion of MWCNTs (outer diameter > 50nm, length 10-20 mm) in the cement mix improves both the tensile fracture characteristics and compressive strength when not mixed with a surfactant compound [22]. The improvement in the mechanical properties in specimens with the addition of CNTs was observed more clearly with increasing curing age. The fracture mechanics testresults indicated that the fracture properties of microconcrete and mortar are increased through proper dispersion of very low amount of MWCNTs (CNTs/c = 0.005). The vibration damping capacities of cement-based matrices with additions of MWCNTs (external diameter 20-40 nm, length 5-15 μm) were investigated with a free vibration testing method in an elastic system [23]. Adding small amounts of MWCNTs, several positive effects on the critical damping ratio (ζ) were observed for the cement matrix. This magnitude is increased in the MWCNTcement composites because of frictions among MWCNT matrix and multiple intertubes.

As an example of improvement of other cement properties, MWCNT (external diameter 20-40 nm, length 5-15 μm) reinforced cement composites (MWFRCs) were prepared with surfac‐ tant dispersion, ultrasonic treatment, and subsequently high-speed shear mixing processes [24]. Several following parameters can be considerably improved by adding MWCNTs: critical crack mouth opening displacement (*δC*, up to 119.4%), flexural strength (*σs*, up to 54.8%) of the cured nanocomposite, and the stress-intensity factor (KIC, up to 56.4%). The authors ex‐ plained these values mainly by "the superior pulling-out effect of dispersed and tough MWCNT fiber upon the notched cracks". The *σs*, *KIC*, and *δ<sup>C</sup>* values can be balanced by incorporation of MWCNTS treated with acids. Additional short carbon fibers provoked a negative effect, meanwhile nanophase carbon black improved the fracture toughness.

There are controversial results for cement paste with admixed CNT of up to 500 μm in length, showing an increase or decrease in flexural or compressive strength. A small increase in fracture energy and tensile strength of a CNT-reinforced cement paste with 50 × 50 μm was reported [25]. CNT clustering was proved to be the crucial factor for an increase in fracture energy and for an improvement in tensile strength. Data from **Table 1** demonstrate that even CNTs shorterthan 10 μm can be effective in providing added strength. In addition, the general transport properties (i.e., water sorptivity, water permeability, and gas permeability) of carbon-nanotube/cement composites were investigated [26]. Carboxyl-functionalized MWCNTs (outside diameter < 8 nm, inside diameter 2-5 nm, -COOH content 3.86 wt%, length 10-30 μm) were dispersed into cement mortar to fabricate the CNT-reinforced compo‐ sites, by applying ultrasonic energy, in combination with the use of surfactants (sodium dodecylbenzene sulfonate, SDS, and NaDDBS). Several coefficients of cement mortar (water


permeability coefficient, water sorptivity coefficient, and gas permeability coefficient) can be decreased by adding even small quantities of MWCNT-COOH (0.2%), thus increasing the durability of composites.

cement ratios such as 0.40 or 0.45 is very difficult or even impossible. Therefore, the authors consider it important to use fresh cement with a minimal amount of large grains and clumps

**3. Improvement of concrete properties by addition of CNTs and CNFs**

It is well-known nowadays that mechanical and various other properties of concrete, indicat‐ ed below, are considerably improved by addition of CNTs. Thus, the inclusion of MWCNTs (outer diameter > 50nm, length 10-20 mm) in the cement mix improves both the tensile fracture characteristics and compressive strength when not mixed with a surfactant compound [22]. The improvement in the mechanical properties in specimens with the addition of CNTs was observed more clearly with increasing curing age. The fracture mechanics testresults indicated that the fracture properties of microconcrete and mortar are increased through proper dispersion of very low amount of MWCNTs (CNTs/c = 0.005). The vibration damping capacities of cement-based matrices with additions of MWCNTs (external diameter 20-40 nm, length 5-15 μm) were investigated with a free vibration testing method in an elastic system [23]. Adding small amounts of MWCNTs, several positive effects on the critical damping ratio (ζ) were observed for the cement matrix. This magnitude is increased in the MWCNTcement composites because of frictions among MWCNT matrix and multiple intertubes.

As an example of improvement of other cement properties, MWCNT (external diameter 20-40 nm, length 5-15 μm) reinforced cement composites (MWFRCs) were prepared with surfac‐ tant dispersion, ultrasonic treatment, and subsequently high-speed shear mixing processes [24]. Several following parameters can be considerably improved by adding MWCNTs: critical crack mouth opening displacement (*δC*, up to 119.4%), flexural strength (*σs*, up to 54.8%) of the cured nanocomposite, and the stress-intensity factor (KIC, up to 56.4%). The authors ex‐ plained these values mainly by "the superior pulling-out effect of dispersed and tough MWCNT fiber upon the notched cracks". The *σs*, *KIC*, and *δ<sup>C</sup>* values can be balanced by incorporation of MWCNTS treated with acids. Additional short carbon fibers provoked a negative effect, meanwhile nanophase carbon black improved the fracture toughness.

There are controversial results for cement paste with admixed CNT of up to 500 μm in length, showing an increase or decrease in flexural or compressive strength. A small increase in fracture energy and tensile strength of a CNT-reinforced cement paste with 50 × 50 μm was reported [25]. CNT clustering was proved to be the crucial factor for an increase in fracture energy and for an improvement in tensile strength. Data from **Table 1** demonstrate that even CNTs shorterthan 10 μm can be effective in providing added strength. In addition, the general transport properties (i.e., water sorptivity, water permeability, and gas permeability) of carbon-nanotube/cement composites were investigated [26]. Carboxyl-functionalized MWCNTs (outside diameter < 8 nm, inside diameter 2-5 nm, -COOH content 3.86 wt%, length 10-30 μm) were dispersed into cement mortar to fabricate the CNT-reinforced compo‐ sites, by applying ultrasonic energy, in combination with the use of surfactants (sodium dodecylbenzene sulfonate, SDS, and NaDDBS). Several coefficients of cement mortar (water

for making CNF-CNT reinforced composites.

142 Advances in Carbon Nanostructures

**Table 1.** Strength gain when CNTs are added to cement (P = paste, M = mortar). Reproduced with permission from Ref. [25].

In cement composites, carbon fibers (CFs) can be applied alone or in combination with carbon nanotubes. The behavior of reinforced cement mortar composites round bars with MWCNTs (0.5 wt%; diameter 10-30 nm, length 1-2 μm, surface area 350 m2 /g) and CFs (2.25 wt%; length of fiber 5 mm, fiber thickness 0.3 mm), ultrasonically dispersed in the cement matrix, was elucidated [27]. Composite round bars were tested under direct tension, in order to evaluate their mechanical properties, such as ultimate load, deflection criteria, and stressstrain behavior. It was shown that the load carrying capacity of composite bars, under direct tension, is substantially higher than the plain controlled bar. Resulted plain cement bars with 2.25% carbon fiberreinforcement showed 38% increase in the ultimate load, whereas plain cement bars with 2.25% carbon fibers and 0.5% MWCNT showed 54% increase in the load carrying capacity. The presence of both carbon fibers and MWCNTs mainly contributed to the improvement of tensile load carrying capacity. Higher failure strain capacity of 44% was observed in bars with randomly distributed CF and MWCNTs as compared with plain bars.

Allotropes and hybrids of carbon can be obtained *in situ* on clinker particles as supports. Thus, the formation of a hybrid material (carbon clinker/nanofiber, **Figure 4**) was achieved by direct interaction of carbon nanofibers on clinker supporting particles (conditions: 550℃, fluidized bed reactor, gaseous C*2*H*2*, and CO*2*) [28]. This interaction resulted in an excellent CNFs/clinker matrix dispersion and a strong final composite. At the 0.4% CNFs concentration, the enhance‐ ment of the mechanical properties of the mortar resulted in an increase of more than 2.5-fold in the compressive strength. In a similar manner, carbon nanotubes were grown *in situ* by chemical vapor deposition on a Portland cement clinker, in order to produce a nanocompo‐ site material using clinker as support and catalyst [29]. Iron sources such as iron ore, steel mill scale, and red mud were used as additional transition metal catalysts to increase the carbon nanotube content. Thus, the clinker contains tetracalcium aluminoferrite (C4AF phase) that can be used as a catalyst for CNT growth. In a CVD process using ethylene as carbon source, and pure PC clinker both as a support and catalyst of CVD process, a CNT yield of 4.03%, in mass of clinker-CNT nanocomposite, was obtained.

**Figure 4.** TEM image of the CNFs produced on the surface of clinker particles. Reproduced with permission.

## **4. Effects observed upon concrete reinforcement with carbon nanotubes**

#### **4.1. Effects of CNT functionalization, annealing, concentration, and water-cement ratio**

–OH and –COOH functionalized CNTs have distinct physical properties and are more hydrophilic in comparison with pristine CNTs, so their effects on concrete properties can be different. The influence of functionalized MWCNT–OH and MWCNT–COOH on the impact resistance and compressive and flexural strengths of high–performance mortars (HPM) was studied [30]. The results of tests on reinforced high-performance mortar containing 0.1 wt% functionalized MWCNTs showed that the impact resistance, compressive strength, and flexural strength were 1400, 25.58, and 2% higher than those of HPM without MWCNTs. A more detailed comparison of distinct CNTs in cement matrix composites, prepared by addition of 0.5 wt% CNTs to plain cement paste, "was carried out [31]". Three different kinds of MWCNTs (**Table 2** and **Figure 5**) were used and compared: as-grown, annealed (high temperature (2100-2600℃); annealing effects on MWCNTs are described in [32, 33]), and carboxyl functionalized MWCNTs. The authors stated that high temperature annealing treatments remove lattice defects from the walls of CNTs, hence improving their mechanical

and pure PC clinker both as a support and catalyst of CVD process, a CNT yield of 4.03%, in

**Figure 4.** TEM image of the CNFs produced on the surface of clinker particles. Reproduced with permission.

**4. Effects observed upon concrete reinforcement with carbon nanotubes**

**4.1. Effects of CNT functionalization, annealing, concentration, and water-cement ratio**

–OH and –COOH functionalized CNTs have distinct physical properties and are more hydrophilic in comparison with pristine CNTs, so their effects on concrete properties can be different. The influence of functionalized MWCNT–OH and MWCNT–COOH on the impact resistance and compressive and flexural strengths of high–performance mortars (HPM) was studied [30]. The results of tests on reinforced high-performance mortar containing 0.1 wt% functionalized MWCNTs showed that the impact resistance, compressive strength, and flexural strength were 1400, 25.58, and 2% higher than those of HPM without MWCNTs. A more detailed comparison of distinct CNTs in cement matrix composites, prepared by addition of 0.5 wt% CNTs to plain cement paste, "was carried out [31]". Three different kinds of MWCNTs (**Table 2** and **Figure 5**) were used and compared: as-grown, annealed (high temperature (2100-2600℃); annealing effects on MWCNTs are described in [32, 33]), and carboxyl functionalized MWCNTs. The authors stated that high temperature annealing treatments remove lattice defects from the walls of CNTs, hence improving their mechanical

mass of clinker-CNT nanocomposite, was obtained.

144 Advances in Carbon Nanostructures

**Figure 5.** Electron microscopy analysis showing p-CNTs (a), a-CNTs (b), and f-CNTs (c). TEM images are reported on the left, while SEM images are on the right. Reproduced with permission.

strength. On the other hand, acid oxidative treatments increase chemical reactivity of pris‐ tine material, consequently chemical bonds between the reinforcement and the cement matrix are supposed to enhance the mechanical strength. Flexural and compressive tests showed a deterioration of the mechanical properties with functionalized MWCNTs, while a significant improvement is observed with both as-grown and annealed MWCNTs.


\*p-CNTs: pristine MWCNTs; a-CNTs: annealed MWCNTs; f-CNTs: functionalized MWCNTs.

**Table 2.** Characteristics of the three different MWCNTs\* dispersed in the cement. Reproduced with permission.

The piezoresistive property of the CNT/cement composite (**Figure 6**) shows an acidic treat‐ ment on CNTs (for their functionalization) and was used to explore its feasibility as an embedded stress sensor for civil structures, such as roadways, levees, and bridges [34]. It was shown that the electrical resistance of the CNT/cement composite changes with the compres‐ sive stress level, indicating the potential of using the CNT/cement composite as a stress sensor for civil structures (**Table 3**). It was also established that the dispersion-assistant surfactants could block the contacts among carbon nanotubes, thus impairing the piezoresistive re‐ sponse of the composite, while a higher CNT doping level could improve the sensitivity of the composite stress response. In addition, the effect of MWCNTs on strength characteristics and durability of concrete was investigated [35] by adding MWCNTs in different quantities (diameter 20-40 nm; length 1-10 μm; 0.015, 0.03, and 0.045%) with surfactants (super plasti‐ cizers-polycarboxylate 8H, 0.25% by weight of cement). Results showed an increase in compressive and splitting-tensile strengths of the samples with increasing MWCNT concen‐

**Figure 6.** Illustration of the CNT/cement fabrication process based on the acid treatment of CNTs. Reproduced with permission.

tration (**Table 4**). By increasing the percentage of MWCNTs in the concrete, the water absorption is reduced to a greater extent, which helps in improving the concrete durability and water resistance.


**Table 3.** Comparison of electrical resistance changes of CNT/cement composites with different CNT doping levels and under different compressive loads. Reproduced with permission.


**Table 4.** Compressive and split tensile strength. Reproduced with permission.

**Property p-CNTs a-CNTs f-CNTs** Deposition technique CVD CVD CVD Average diameter, nm 40–80 40–80 10–20 Average length, μm 400–1000 200–400 0.1–10 Carbon purity, wt% >92 >99 >95 Metal oxide, impurity, wt% <6 <1 <5 -COOH functionalization, wt% 0 0 <4

**Table 2.** Characteristics of the three different MWCNTs\* dispersed in the cement. Reproduced with permission.

The piezoresistive property of the CNT/cement composite (**Figure 6**) shows an acidic treat‐ ment on CNTs (for their functionalization) and was used to explore its feasibility as an embedded stress sensor for civil structures, such as roadways, levees, and bridges [34]. It was shown that the electrical resistance of the CNT/cement composite changes with the compres‐ sive stress level, indicating the potential of using the CNT/cement composite as a stress sensor for civil structures (**Table 3**). It was also established that the dispersion-assistant surfactants could block the contacts among carbon nanotubes, thus impairing the piezoresistive re‐ sponse of the composite, while a higher CNT doping level could improve the sensitivity of the composite stress response. In addition, the effect of MWCNTs on strength characteristics and durability of concrete was investigated [35] by adding MWCNTs in different quantities (diameter 20-40 nm; length 1-10 μm; 0.015, 0.03, and 0.045%) with surfactants (super plasti‐ cizers-polycarboxylate 8H, 0.25% by weight of cement). Results showed an increase in compressive and splitting-tensile strengths of the samples with increasing MWCNT concen‐

**Figure 6.** Illustration of the CNT/cement fabrication process based on the acid treatment of CNTs. Reproduced with

permission.

\*p-CNTs: pristine MWCNTs; a-CNTs: annealed MWCNTs; f-CNTs: functionalized MWCNTs.

146 Advances in Carbon Nanostructures

Additionally, two mechanical properties (28-day compressive strength and flexural strength) of CNTs (diameter <1 nm up to 50 nm and length from 1 μm to 1 cm) and CNFs (diameter of 70-200 nm and length of 50-200 μm) cement composites were investigated [36]. Composites with 0.1 and 0.2% of CNTs and CNFs, and water/cement ratios between 0.35 and 0.5, were prepared under sonication. Both CNT and CNF composites demonstrated significant in‐ crease in compressive strengths, when compared to plain mortar control samples (maxi‐ mum 154% for CNT and 217% for CNFs samples). Water/cement ratios in the range of 0.35-0.4 were found to produce the higher strengths, together with a 0.1% dosage rate for the CNTs/ CNFs. It seems that the CNTs are better dispersed in the cement matrix than the CNFs, because a correlation between the flow test results and the compressive strengths was detected for the CNT samples. One other example that CNT concentration should not exceed certain levels was described for the heavy-weight oil well cements [37]. When only 1 wt% of CNTs was added to the cement slurry, the yield point and plastic viscosity increased by 8 and 5 times, respec‐ tively, while the free water and fluid loss of cement slurry were reduced by 85 and 70%, respectively. In addition, compressive strength of cement stone increased by 73.8%. However, the increase in the additive concentration leads to reduced compressive strength and Young's modulus, because of unsuitable dispersion of nanoparticles in the cement stone matrix; thus, an optimum level of CNTs should be used.

#### **4.2. Length-to-diameter (l/d) aspect ratio effect**

In addition to the effects of CNT length described above (**Table 1**), the intriguing effects of different concentrations of long (diameter <8 nm, length 10-30 μm) MWCNTs-high length/ diameter aspect ratios of 1250–3750-and short MWCNTs (diameter 9.5 nm, length 1.5 μm) aspect ratio of about 157-in cement paste were studied for 7, 14, and 28 days [38]. Both short and long MWCNTs were produced by the catalytic chemical vapor deposition (CCVD) process. **Figure 7** shows the relationships between the length-to-diameter aspect ratio and the surface-area-to-volume (SA/V) ratio for different lengths of SWCNTs, MWCNTs, carbon nanofibers, and carbon microfibers (CMFs). The authors noticed that only CNTs can provide a very high surface-area-to-volume (SA/V) ratio, which is one of the most important and desired elements in fiber-reinforced composite systems, in order to obtain the best and the most efficient materials. SWCNTs are the only materials that have a SA/V ratio that exceeds 2.0 nm-1, especially when considering the ultra-long ones that have aspect ratios that can reach several millions. A higher SA/V ratio means a larger contact area between the fibers and the surrounding matrix, hence higher interaction with the matrix and more efficient reinforcing.

**Figure 7.** Length-to-diameter (l/d) aspect ratio effect on the surface-area-to-volume (SA/V) ratio for different lengths of SWCNTs, MWCNTs, CNFs, and CMFs. Reproduced with permission.

It was established that nanocomposites with low concentration of long MWCNTs yield comparable mechanical performance to the nanocomposites with higher concentration of short MWCNTs. **Figure 8** shows that the Young's modulus increases as the long CNTs' concentra‐ tion decreases, whereas the Young's modulus increases as the short CNTs' concentration increases. Also, the authors noticed that low concentrations of long CNTs can lead to a much higher increase in the Young's modulus, as compared to higher concentrations of short CNTs. In addition, an example of the microcrack bridging by the MWCNTs, within the cement paste, is shown in **Figure 9**, clearly indicating that many CNTs are bridging the microcrack. Both

modulus, because of unsuitable dispersion of nanoparticles in the cement stone matrix; thus,

In addition to the effects of CNT length described above (**Table 1**), the intriguing effects of different concentrations of long (diameter <8 nm, length 10-30 μm) MWCNTs-high length/ diameter aspect ratios of 1250–3750-and short MWCNTs (diameter 9.5 nm, length 1.5 μm) aspect ratio of about 157-in cement paste were studied for 7, 14, and 28 days [38]. Both short and long MWCNTs were produced by the catalytic chemical vapor deposition (CCVD) process. **Figure 7** shows the relationships between the length-to-diameter aspect ratio and the surface-area-to-volume (SA/V) ratio for different lengths of SWCNTs, MWCNTs, carbon nanofibers, and carbon microfibers (CMFs). The authors noticed that only CNTs can provide a very high surface-area-to-volume (SA/V) ratio, which is one of the most important and desired elements in fiber-reinforced composite systems, in order to obtain the best and the most efficient materials. SWCNTs are the only materials that have a SA/V ratio that exceeds 2.0 nm-1, especially when considering the ultra-long ones that have aspect ratios that can reach several millions. A higher SA/V ratio means a larger contact area between the fibers and the surrounding matrix, hence higher interaction with the matrix and more efficient reinforcing.

**Figure 7.** Length-to-diameter (l/d) aspect ratio effect on the surface-area-to-volume (SA/V) ratio for different lengths of

SWCNTs, MWCNTs, CNFs, and CMFs. Reproduced with permission.

an optimum level of CNTs should be used.

148 Advances in Carbon Nanostructures

**4.2. Length-to-diameter (l/d) aspect ratio effect**

**Figure 8.** Variation of the normalized Young's modulus with the CNT concentration for two different aspect ratios; long and short. Reproduced with permission.

**Figure 9.** SEM image showing the microcrack bridging and breakage of the MWCNTs within the cement paste compo‐ site ((C–S–H) means calcium-silicate-hydrate). Reproduced with permission.

pull-out and breakage of CNTs can also be observed from images. In fact, the CNT breakage implies a good bonding between the CNT surfaces and the surrounding cement paste. However, it was noted by authors that the presence of CNTs affects the chemical reaction of the hydrated cement.

#### **4.3. Effect of combination of CNTs with silica nanoparticles**

Nanosilica (NS) is applied as an additive to improve the cement properties, giving a 15–20% increase of compressive strength [39]. Its combination with CNTs could lead to better results. Thus, the behavior of hardened cement paste reinforced with MWCNTs (0.75% by weight of cement; diameter 10–30 nm, length 1–2 mm) and NS (0.5%; particle size 10–20 nm) was investigated (**Figure 10**) after both MWCNTs and NS being dispersed using a ultrasonic energy method [40]. MWCNTs were used to reinforce the cement paste at the nanoscale, while NS was used to increase strength by forming an additional CSH gel. It was shown that the addition of MWCNTs and NS almost doubled the flexural strength in comparison with PC (the control beam).

**Figure 10.** SEM image showing hardened cement paste modified with nanosilica and MWCNTs. Reproduced with permission.

## **5. Other specific studies of CNT/cement composites**

## **5.1. Computational studies**

pull-out and breakage of CNTs can also be observed from images. In fact, the CNT breakage implies a good bonding between the CNT surfaces and the surrounding cement paste. However, it was noted by authors that the presence of CNTs affects the chemical reaction of

Nanosilica (NS) is applied as an additive to improve the cement properties, giving a 15–20% increase of compressive strength [39]. Its combination with CNTs could lead to better results. Thus, the behavior of hardened cement paste reinforced with MWCNTs (0.75% by weight of cement; diameter 10–30 nm, length 1–2 mm) and NS (0.5%; particle size 10–20 nm) was investigated (**Figure 10**) after both MWCNTs and NS being dispersed using a ultrasonic energy method [40]. MWCNTs were used to reinforce the cement paste at the nanoscale, while NS was used to increase strength by forming an additional CSH gel. It was shown that the addition of MWCNTs and NS almost doubled the flexural strength in comparison with PC (the control

**Figure 10.** SEM image showing hardened cement paste modified with nanosilica and MWCNTs. Reproduced with

**4.3. Effect of combination of CNTs with silica nanoparticles**

the hydrated cement.

150 Advances in Carbon Nanostructures

beam).

permission.

Nanoscale and macroscale models were used to investigate the interfacial bond behavior of an individual CNT embedded in a cement matrix and the flexural response of a CNT/cement composite beam, respectively; also finite element (FE) models were applied to study the improvement in tensile strength and ductility of CNT-reinforced cement [41]. Industrial grade MWCNTs with a purity of 90 wt%, and a concentration of 0.25% by the total weight of cement paste, were ultrasonically dispersed in multiple steps before being added to cement (**Figure 11**). The FE models were successful in capturing the degradation in the CNT/cement bond strength and its effect on the flexural strength, ductility, and toughness of the compo‐ site. A comparison between the experimental and numerical behavior of the composite beams suggests an effective average shear strength value of 6.5 MPa at the interface of CNTs and cement matrix. Authors noted that, with a shear strength of up to 20.0 MPa, flexural strength, ductility, and toughness all increased with the interfacial bonding between CNTs and cement. This implies the significance of surface treatment of CNTs, which is commonly used to strengthen the bonding between CNTs and their surrounding matrix. As predicted by the FE analysis, increasing the shear strength from 6.5 to 20.0 MPa would increase the composite's flexural strength, ductility, and toughness by 141, 259, and 1976%, respectively. These results were significant, considering thatthe enhancement was providedby fibers alone, andhighlight the great potential of surface-treated CNTs in reinforcing cement.

```
Figure 11. Experimental preparation for three-point bending tests of composite beams. Reproduced with permission.
```
### **5.2. TEM approach**

A colloidal technique for transmission electron microscopy (TEM) of graphitic nanorein‐ forced cementitious composites was developed [42]. SWCNTs and MWCNTs (90 and 95% by mass with bulk densities of 0.14 and 0.27 g/cm3 at 20℃, respectively; the inner diameters of the tubes were specified as 0.8–1.6 and 2–5 nm with average tube lengths of 5–30 and 10–30 μm for SWNTs and MWNTs, respectively) were functionalized using an acid etching technique to obtain stable aqueous suspensions to incorporate in the mix design of a cement paste. The functionalized nanoreinforcement and binding characteristics were observed at the nano‐ scale using high-resolution TEM imaging. Functionalized CNTs were found to be well distributed and preferentially associated with the cementitious matrix.

#### **5.3. Study of electrical conductivity**

A model for the electrical conductivity of heterogeneous systems based on cement and CNTs was developed [43]. The electrical properties of heterogeneous variances were shown to be dependent of such key parameters, such as the degree of aggregation of conductive particles and the electrical conductivity of a single unit. When the threshold concentration of electri‐ cal conductivity of CNT and graphite dispersions equals to 0.15, the electrical conductivity of CNT's systems were found to be 5–6 times higher than graphite. The authors emphasized that the CNTs have anomalous properties in comparison with graphite electrical dependencies due to the size and shape of the particles. It was also suggested that the small size effect of CNTs on the primary aggregation and the elongated shape of the particles affects the concentra‐ tion at which a significant increase in electrical conductivity takes place.

#### **5.4. Ultrasound tests**

Tests with three different ratios (0.2, 0.4, and 0.6%, relatively high in comparison with those in reports described above) of carbon nanotubes (OD 5–60 nm, a length estimated from 5 to 30 μm) in concrete were carried out with the ultrasound technique. The measurement of the propagation delay of the signal through the specimens of cementitious materials with and without CNTs to reach the dynamic modulus of elasticity was carried out [44]. It was found that the mixture of 0.4% of CNTs was the mortar that showed the best performance, in comparison to the reference sample, reaching an increase of about 25% in the dynamic modulus of elasticity and a higher rate of structural compaction. The composites with addition of 0.2 and 0.6% of CNT did not show significant results for the same characteristics.

**Figure 12.** Cement foam concrete structure: (a) without nanotubes, (b) with 0.05% nanotubes (pore walls), (c) without nanotubes (perforated), and (d) stabilized with addition of 0.05% of nanotubes. Reproduced with permission.

#### **5.5. Study of special foam-like cement**

CNTs (80–90% of carbon, with a diameter of up to 100 nm and a length of up to 20 μm), agglomerated to fiber-shaped agglomerates with a diameter of up to 30 μm and a length of

up to 10 mm, were synthesized using the method of stimulation of dehydropolycondensa‐ tion and carbonization of aromatic hydrocarbons in chemical active environment (melts of aluminum, copper, nickel and iron salts), and used as a high strength dispersed reinforce‐ ment forthe synthesis of nonautoclave concrete foam produced on the base of Portland cement (**Figure 12**) [45]. It was established that the use of CNTs (0.05% by mass), in the production of these concretes, allows to decrease its heat conductivity to 12–20% and increase its compres‐ sive strength to 70%.

## **6. Health impact**

scale using high-resolution TEM imaging. Functionalized CNTs were found to be well

A model for the electrical conductivity of heterogeneous systems based on cement and CNTs was developed [43]. The electrical properties of heterogeneous variances were shown to be dependent of such key parameters, such as the degree of aggregation of conductive particles and the electrical conductivity of a single unit. When the threshold concentration of electri‐ cal conductivity of CNT and graphite dispersions equals to 0.15, the electrical conductivity of CNT's systems were found to be 5–6 times higher than graphite. The authors emphasized that the CNTs have anomalous properties in comparison with graphite electrical dependencies due to the size and shape of the particles. It was also suggested that the small size effect of CNTs on the primary aggregation and the elongated shape of the particles affects the concentra‐

Tests with three different ratios (0.2, 0.4, and 0.6%, relatively high in comparison with those in reports described above) of carbon nanotubes (OD 5–60 nm, a length estimated from 5 to 30 μm) in concrete were carried out with the ultrasound technique. The measurement of the propagation delay of the signal through the specimens of cementitious materials with and without CNTs to reach the dynamic modulus of elasticity was carried out [44]. It was found that the mixture of 0.4% of CNTs was the mortar that showed the best performance, in comparison to the reference sample, reaching an increase of about 25% in the dynamic modulus of elasticity and a higher rate of structural compaction. The composites with addition of 0.2

**Figure 12.** Cement foam concrete structure: (a) without nanotubes, (b) with 0.05% nanotubes (pore walls), (c) without nanotubes (perforated), and (d) stabilized with addition of 0.05% of nanotubes. Reproduced with permission.

CNTs (80–90% of carbon, with a diameter of up to 100 nm and a length of up to 20 μm), agglomerated to fiber-shaped agglomerates with a diameter of up to 30 μm and a length of

distributed and preferentially associated with the cementitious matrix.

tion at which a significant increase in electrical conductivity takes place.

and 0.6% of CNT did not show significant results for the same characteristics.

**5.3. Study of electrical conductivity**

152 Advances in Carbon Nanostructures

**5.4. Ultrasound tests**

**5.5. Study of special foam-like cement**

Considerable experimental data related to CNT toxicity at the molecular, cellular, and whole animal levels have been published [46]. The literature indicates considerable variability and uncertainty regarding the health impacts, reactivity, ecological effects, and environmental destination and transport of CNTs. In this respect, their addition to form high-performance cements should be also evaluated from the point of view of toxicity of resulted concrete, i.e., how the addition of CNTs may affect the environmental profile of cement. A recent report [47] evaluated hypothetical high-performance cements based on CNT reinforcement with a life cycle assessment (LCA) in order to compare the environmental impact of these new develop‐ ments to traditional cements. It was established that the inclusion of CNTs increases consid‐ erably the environmental impact of cement production (**Table 5**). In addition, progress in research on these kinds of systems is largely hampered by the intrinsically hydrophobic nature of CNTs and their chemical incompatibility with cement hydrates. The authors proposed alternatives to CNTs as reinforcement for cements such as inorganic nanotubes or plastic nanofibers.


**Table 5.** Comparative impacts of standard and reinforced cements. Reproduced with permission.

## **7. Conclusions**

The addition of CNTs and CNFs into cement forming nanocomposites leads to a considera‐ ble improvement of mechanical characteristics of concrete, since CNTs can act as effective bridges to minimize and limit the propagation of microcracks through the matrix. However, in order to reach the best effects, CNTs need to be well dispersed within the matrix to provide good bonding between them and the surrounding hydrated cement matrix, this being the main factor controlling concrete microstructure of cement composites. Addition of even 0.05% CNTs results in lower density, increased compressive strength, lower thermal conductivity, lower average pore diameter, and more homogeneous pore wall structure. The central idea of each cement-CNT experiment is to determine the best way to disperse CNTs. This can be reached by ultrasonication, use of surfactants (generally sodium dodecylbenzenesulfonate (SDBS), sodium deoxycholate (NaDC), Triton X-100 (TX10), GA, and cetyl trimethyl ammonium bromide (CTAB), whose dispersion ability varies considerably depending on molecular structure), use of cement admixtures (polycarboxylate, alkylbenzene, sulfonic acid, among many others), and covalent functionalization (commonly acid treatment to form -COOH and -OH functionalized CNTs). Combination of various chemical methods, such as the combina‐ tion of surface functionalization with polymers, and other techniques, like an innovative method of fabricating cementitious nanocomposites through growth of CNTs onto the cement particles, also leads to good results. It has been finally established that the homogeneous CNT dispersion can be reachedin the bestform bydispersing them firstin water, followedby mixing of the aqueous dispersion with mortar paste.

A series of effects have been discovered upon dispersion of carbon allotropes in concrete, showing the importance of diameter, length, and length-to-diameter ratio of CNTs, their concentration, functionalization, annealing, combination of CNTs with other nanomaterials, and water-cement ratio. We note that studies on CNT/cement nanocomposites are currently a relatively hot topic in nanotechnology [48–51] leading to best combinations of well-known materials having improved properties.

## **Author details**

Oxana V. Kharissova, Leticia M. Torres Martínez and Boris I. Kharisov\*

\*Address all correspondence to: bkhariss@hotmail.com

1 Department of Physico-Mathematical Sciences, Autonomous University of Nuevo León, Monterrey, Mexico

2 Department of Civil Engineering, Autonomous University of Nuevo León, Monterrey, Mexico

3 Department of Chemical Sciences, Autonomous University of Nuevo León, Monterrey, Mexico

## **References**

**7. Conclusions**

154 Advances in Carbon Nanostructures

of the aqueous dispersion with mortar paste.

materials having improved properties.

**Author details**

Monterrey, Mexico

Mexico

Mexico

The addition of CNTs and CNFs into cement forming nanocomposites leads to a considera‐ ble improvement of mechanical characteristics of concrete, since CNTs can act as effective bridges to minimize and limit the propagation of microcracks through the matrix. However, in order to reach the best effects, CNTs need to be well dispersed within the matrix to provide good bonding between them and the surrounding hydrated cement matrix, this being the main factor controlling concrete microstructure of cement composites. Addition of even 0.05% CNTs results in lower density, increased compressive strength, lower thermal conductivity, lower average pore diameter, and more homogeneous pore wall structure. The central idea of each cement-CNT experiment is to determine the best way to disperse CNTs. This can be reached by ultrasonication, use of surfactants (generally sodium dodecylbenzenesulfonate (SDBS), sodium deoxycholate (NaDC), Triton X-100 (TX10), GA, and cetyl trimethyl ammonium bromide (CTAB), whose dispersion ability varies considerably depending on molecular structure), use of cement admixtures (polycarboxylate, alkylbenzene, sulfonic acid, among many others), and covalent functionalization (commonly acid treatment to form -COOH and -OH functionalized CNTs). Combination of various chemical methods, such as the combina‐ tion of surface functionalization with polymers, and other techniques, like an innovative method of fabricating cementitious nanocomposites through growth of CNTs onto the cement particles, also leads to good results. It has been finally established that the homogeneous CNT dispersion can be reachedin the bestform bydispersing them firstin water, followedby mixing

A series of effects have been discovered upon dispersion of carbon allotropes in concrete, showing the importance of diameter, length, and length-to-diameter ratio of CNTs, their concentration, functionalization, annealing, combination of CNTs with other nanomaterials, and water-cement ratio. We note that studies on CNT/cement nanocomposites are currently a relatively hot topic in nanotechnology [48–51] leading to best combinations of well-known

1 Department of Physico-Mathematical Sciences, Autonomous University of Nuevo León,

2 Department of Civil Engineering, Autonomous University of Nuevo León, Monterrey,

3 Department of Chemical Sciences, Autonomous University of Nuevo León, Monterrey,

Oxana V. Kharissova, Leticia M. Torres Martínez and Boris I. Kharisov\*

\*Address all correspondence to: bkhariss@hotmail.com


[26] Han, B.; Yang, Z.; Shi, X.; Yu, X. Transport properties of carbon-nanotube/cement composites. *Journal of Materials Engineering and Performance*, 2013, *22*(1), 184–189.

[14] Nochaiya, T.; Chaipanich, A. Behavior of multi-walled carbon nanotubes on the porosity and microstructure of cement-based materials. *Applied Surface Science*, 2011,

[15] Fraga, J.L.; del Campo, J.M.; García, J.A. Carbon nanotube - cement composites in the cement composites in the construction industry: 1952–2014. A state of the art review. In 2nd International Conference on Emerging Trends in Engineering and Technology

[16] Petrunin, S.; Vaganov, V.; Sobolev, K. The effect of functionalized carbon nanotubes on the performance of cement composites. NANOCON, October 16–18, 2013, Brno, Czech

[17] Han, B.; Yu, X. Effect of surfactants on pressure-sensitivity of CNT filled cement mortar

[18] Wang, B.; Han, Y.; Pan, B.; Zhang, T. Mechanical and morphological properties of highly dispersed carbon nanotubes reinforced cement based materials. *Journal of Wuhan*

[19] Kołtuńczy, E.; Nowicka, G. Effect of poly(sodium-4-styrenesulphonate) additives on properties of cement suspensions. In *Proceedings of International Scientific Conference "Surfactants and Dispersed Systems in Theory and Practice"*, Ed., K.A. Wilk, Palma Press,

[20] Li, C.; Zhu, H.; Wu, M.; Yan, M. Effect of dispersity of multi-walled carbon nano‐ tubes on compression strength of cement. In *Fourth International Conference on Digital*

[21] Yazdanbakhsh, A.; Grasley, Z.; Tyson, B.; Abu Al-Rub, R.K. Distribution of carbon nanofibers and nanotubes in cementitious composites. *Journal of the Transportation Research Board*, No. 2142, Transportation Research Board of the National Academies,

[22] Şener, S.; Çağlar, Y.; Belgin, Ç.M.; Şener, K.C. Modified Arcan tests for concrete with multi-walled carbon nanotubes. *Advanced Materials Letters*, 2014, 5(8), 429–434.

[23] Duan, Z.; Luo, J. Effect of multi-walled carbon nanotubes on the vibration-reduction behavior of cement. In *Proceedings of SPIE, International Conference on Smart Materials and Nanotechnology in Engineering*, Ed., S. Du, J. Leng, A.K. Asundi, 2007, *6423*, 64230R.

[24] Luo, J.L.; Duan, Z.; Xian, G.; Li, Q.; Zhao, T. Fabrication and fracture toughness properties of carbon nanotube-reinforced cement composite. *The European Physical*

[25] Šmilauer, V.; Hlavácek, P.; Padevet, P. Micromechanical analysis of cement paste with

*257*(6), 1941–1945.

156 Advances in Carbon Nanostructures

Republic, EU.

Wrocław, 2007.

*Manufacturing & Automation*, 2013.

Washington, DC, 2010, 89–95.

*Journal Applied Physics*, 2011, 53, 30402.

carbon nanotubes. *Acta Polytechnica*, 2012, 52(6), 22–28.

(ICETET'2014), May 30–31, 2014, London, UK.

composites. *Frontiers in Materials*, 2014, 1, art. 27, 1–5.

*University of Technology - Materials Science Editorial*, 2013, 82–87.


[50] Sakulich, A.R.; Li, V.C. Nanoscale characterization of engineered cementitious composites (ECC). *Cement and Concrete Research*, 2011, 41(2), 169–175.

[38] Abu Al-Rub, R.K.; Ashour, A.I.; Tyson, B.M. On the aspect ratio effect of multi-walled carbon nanotube reinforcements on the mechanical properties of cementitious

[39] Sobolev, K.; Flores, I.; Hermosillo, R.; Torres-Martínez, L.M. Nanomaterials and nanotechnology for high-performance cement composites. In Proceedings of ACI Session on "*Nanotechnology of Concrete: Recent Developments and Future Perspectives*",

[40] Hunashyal, A.; Banapurmath, N.; Jain, A.; Quadri, S.; Shettar, A. Experimental investigation on the effect of multiwalled carbon nanotubes and nano-SiO*<sup>2</sup>* addition on mechanical properties of hardened cement paste. *Advances in Materials*, 2014, *3*(5), 45–

[41] Chan, L.Y.; Andrawes, B. Finite element analysis of carbon nanotube/cement compo‐ site with degraded bond strength. *Computational Materials Science*, 2010, *47*, 994–1004.

[42] Aich, N.; Zohhadi, N.; Khan, I.A.; Matta, F.; Ziehl, P.; Saleh, N.B. Applied TEM approach for micro/nanostructural characterization of carbon nanotube reinforced cementitious composites. *Journal of Research Updates in Polymer Science*, 2012, *1*, 14–23.

[43] Lopanov, A.N.; Shukhov, V.G.; Fanina, E.A.; Kalchev, D.N. Electrical conductivity and aggregation of carbon nanotubes in a heterogeneous system based on cement. *Middle-*

[44] Fernandes de Morais, J.; Naked Haddad, A. Analysis by ultrasound of the behavior of carbon nanotubes on cementitious composites. Journal of Nanotechnology and

[45] Yakovlev, G.; Keriene, J.; Gailius, A.; Girniene, I. Cement based foam concrete rein‐ forced by carbon nanotubes. *Journal of Materials Science (Medžiagotyra)*, 2006, *12*(2), 147–

[46] Liu, Y.; Zhao, Y.; Sun, B.; Chen, C. Understanding the toxicity of carbon nanotubes.

[47] Matanza, A.; Vargas, G.; Leon, I.; Pousse, M.; Salmon, N.; Marieta, C. Life cycle analysis of standard and high-performance cements based on carbon nanotubes composites for

[48] Manzur, T.; Yazdani, N. Strength enhancement of cement mortar with carbon nano‐ tubes. Early results and potential. *Journal of the Transportation Research Board*, No. 2142, TransportationResearch Board ofthe National Academies, Washington, DC, 2010, 102–

[49] Ghasemzadeh, H.; Akbari Jalalabad, E. Computing the compressive strength of carbon nanotube/cement composite. *International Journal of Civil Engineering*, 2011, *9*(3), 223–

construction applications. In *World SB-4*, Barcelona, October 28–30, 2014.

*East Journal of Scientific Research*, 2013, *17*(8), 1194–1199.

*Accounts of Chemical Research*, 2013, *46*(3), 702–713.

Advanced Materials, 2014, 2(2), 89–98.

nanocomposites. *Construction and Building Materials*, 2012, *35*, 647–655.

November 7, 2006, Denver, USA, 91–118.

51.

158 Advances in Carbon Nanostructures

151.

108.

229.

[51] Ubertini, F.; Luigi Materazzi, A.; D'Alessandro, A.; Laflamme, S. Natural frequencies identification of a reinforced concrete beam using carbon nanotube cement-based sensors. *Engineering Structures*, 2014, *60*, 265–275.

## **Polycrystalline Diamond Thin Films for Advanced Applications**

Irena Kratochvilova

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64701

#### **Abstract**

The technological achievements in diamond thin film synthesis over the past decade subsequently led to the utilization of outstanding diamond properties and develop‐ ment of a wide range of applications in various fields of engineering. However, since most chemical vapour deposition (CVD) diamond films are polycrystalline, their characteris‐ tics strongly depend on their microstructure. As the number of possible applications for polycrystalline CVD diamond increases, there is constant development and enhance‐ ment of the film properties. Polycrystalline diamond in the form of thin films delivers further advantages over thicker polycrystalline layer, e.g., smoother surface, less deposition time and less light absorption. Furthermore, besides the relevant diamond properties, the suitability for applications also depends on various material parameters such as substrate nature, substrate dimensions, possibility of non-planar geometries, surface morphology, electrical conductivity, capability of device fabrication, electro‐ chemical properties and cost.

**Keywords:** chemical vapour deposition polycrystalline diamond films, boron-doped polycrystalline diamond film, zirconium alloys protection against corrosion, water splitting

## **1. Introduction**

Due to unique combination of mechanical, thermal and optical properties, manmade diamond is commonly used nowadays in a number of industrial applications. After decades of research and gaining general knowledge, new foreseen applications can be realized through improvements of diamond properties. Achievements in the preparation of synthetic

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

diamond, in the form of thin films, allow for further research aimed at the development of modern applications.

A method, allowing preparation of polycrystalline diamond (PCD) films (**Figure 1**) under lowpressure and low-temperature regimes, was coined with the development of chemical vapour deposition (CVD). CVD apparatus is shown in **Figure 2**. A polycrystalline diamond film is formed during heteroepitaxial growth when the CVD deposition is carried out on a nondiamond substrate. The growth is achieved using small diamond crystal seeds dispersed onto a substrate. Due to eventual coalescence of different crystallites grown on variously oriented seeds (nucleation sites), a continuous PCD film with a columnar structure is formed. Such columnar structure is primarily made of diamond grains (crystallites) held together by thin layers of amorphous and graphitic inclusions located within the so-called grain boundaries. Overall, the PCD film thickness, the size of diamond grains at the surface and its roughness are determined at the moment when growth process is terminated [1].

**Figure 1.** Scanning electron microscopy top view images of the surface of the diamond film.

The technological achievements in the synthesis of diamond thin films over the past decade subsequently led to the utilization of outstanding diamond properties and development of a wide range of applications in various fields of engineering. However, since most CVD diamond films are polycrystalline, their characteristics strongly depend on their microstruc‐ ture. As the number of possible applications for polycrystalline CVD diamond increases, there is constant development and enhancement of the films properties. Polycrystalline diamond in the form of thin films delivers further advantages over thicker polycrystalline layer, e.g., smoother surface, less deposition time and less light absorption. Furthermore, besides the relevant diamond properties, the suitability for applications also depends on various materi‐ al parameters such as substrate nature, substrate dimensions, possibility of non-planar geometries, surface morphology, electrical conductivity, capability of device fabrication, electrochemical properties (given by the sp3 /sp2 ratio) and cost. Due to benefits of CVD diamond films over single-crystalline form (low-cost, large area products), a variety of applications already available expanded beyond classical mechanical use of diamond. These include CVD polycrystalline diamond for optical, electronic and electrochemical applica‐ tions, CVD diamond grades suitable for thermal management as well as PCD films for a number of further advanced applications [1–6].

diamond, in the form of thin films, allow for further research aimed at the development of

A method, allowing preparation of polycrystalline diamond (PCD) films (**Figure 1**) under lowpressure and low-temperature regimes, was coined with the development of chemical vapour deposition (CVD). CVD apparatus is shown in **Figure 2**. A polycrystalline diamond film is formed during heteroepitaxial growth when the CVD deposition is carried out on a nondiamond substrate. The growth is achieved using small diamond crystal seeds dispersed onto a substrate. Due to eventual coalescence of different crystallites grown on variously oriented seeds (nucleation sites), a continuous PCD film with a columnar structure is formed. Such columnar structure is primarily made of diamond grains (crystallites) held together by thin layers of amorphous and graphitic inclusions located within the so-called grain boundaries. Overall, the PCD film thickness, the size of diamond grains at the surface and its roughness

are determined at the moment when growth process is terminated [1].

**Figure 1.** Scanning electron microscopy top view images of the surface of the diamond film.

electrochemical properties (given by the sp3

The technological achievements in the synthesis of diamond thin films over the past decade subsequently led to the utilization of outstanding diamond properties and development of a wide range of applications in various fields of engineering. However, since most CVD diamond films are polycrystalline, their characteristics strongly depend on their microstruc‐ ture. As the number of possible applications for polycrystalline CVD diamond increases, there is constant development and enhancement of the films properties. Polycrystalline diamond in the form of thin films delivers further advantages over thicker polycrystalline layer, e.g., smoother surface, less deposition time and less light absorption. Furthermore, besides the relevant diamond properties, the suitability for applications also depends on various materi‐ al parameters such as substrate nature, substrate dimensions, possibility of non-planar geometries, surface morphology, electrical conductivity, capability of device fabrication,

/sp2

ratio) and cost. Due to benefits of CVD

modern applications.

162 Advances in Carbon Nanostructures

In this chapter, new technological applications of polycrystalline diamond films are present‐ ed. First, we illustrate electrical applications of diamond films. The electrical properties of diamond films can be tuned by addition of dopants, resulting in p-type or n-type conductiv‐ ity. Additionally, the electrical characteristics can be modified through surface termination with hydrogen, which gives rise to p-type surface conductivity. The origin of this effect is in strong covalent bonding between C and H that leads to the formation of two-dimensional surface layer for holes, spanning up to several nanometres below the diamond surface. Conducting microcrystalline, nanocrystalline and ultrananocrystalline diamond films features certain advantages and actively used as electrode material in electrochemistry. That includes chemical stability in aggressive media, optical transparency and wide working potential window. Therefore, diamond film electrodes nowadays find widespread use in various electroanalytical applications, e.g., chemical sensing, in vivo monitoring of bioactive spices, water purification and detection of organic pollutants.

Photoelectrochemical (PEC) cells for water splitting can offer a potentially inexpensive route for molecular hydrogen-fuel generation. In the frame of such application, nanocrystalline diamond has a potential to be utilized in PEC systems as a protective and electrically active coating for water oxidation. A unique diamond deposition technology allowing low temper‐ ature growth (ca. 450°C) was employed to fabricate boron-doped nanocrystalline diamond films on silicon solar cells. The goal of the study was to identify the properties of diamond films suitable for PEC application and to reveal the influence of diamond deposition process on the quality of silicon solar cells. Nanocrystalline boron-doped diamond (BDD) films possess grain boundaries containing a small fraction of non-diamond carbon impurities [4].

In the study of Ashcheulov et al. [4, 5], the conductivity of BDD polycrystalline films that were synthesized by microwave plasma-enhanced chemical vapour deposition on quartz sub‐ strates was studied. Conductivity is a complex physical quantity and in the case of BDD films there are many parameters that influence charge transport: (i) size of the diamond grains—the conductance of BDD films containing bigger grains is higher than the conductance of BDD films containing smaller diamond grains (for the same B content), (ii) diamond grain boun‐ daries, (iii) setting of B defects and specific morphology of the layer changed by B addition and (iv) diamond grain surface layer can have specific setting and concentration of B defects. Boron-doped diamond is presently the subject of considerable interest as a promising electrode material [5]. This is a consequence of several technologically important properties that distinguish it from common electrode materials used in electrochemistry, such as gold or platinum. Polycrystalline BDD films possess physical properties similar to those of bulk diamond, which includes outstanding hardness, high hole mobility, high thermal conductiv‐ ity and excellent resistance to radiation damage. Also, there is little evidence of degradation of BDD electrochemical activity with time. Due to these unique characteristics, to date, BDD electrodes are employed in a vast array of electrochemical applications. Additionally, then doped with high concentration of boron atoms, BDD films possess metallic conductivity and can also be optimized for light transparency [4, 5].

Nanocrystalline diamond layers were examined as anticorrosion protection of zirconium alloys nuclear fuel rods, operated in all commercial light-water and heavy-water nuclear reactors. Various corrosion tests were primarily focused on determining the ability of diamond layer to withstand emergency situations in nuclear reactors. The deposition of a polycrystal‐ line diamond layer was carried out using linear antenna microwave plasma-enhanced chemical vapour deposition (LA-MW PECVD) apparatus (**Figure 2**) [1].

**Figure 2.** Linear antenna microwave plasma-enhanced chemical vapour deposition apparatus.

## **2. Doped nanocrystalline and epitaxial diamond layers for photovoltaic applications**

It was showed that for hydrogen fuel generation based on water electrolysis by means of a light-driven photoelectrochemical process, boron-doped diamond is considered very prom‐ ising material for electrodes [4]. The hydrogen produced using solar energy has been denomi‐ nated as solar fuel. In essence, the photoelectrochemical process mimics photosynthesis by splitting water into H2 and O2. Nevertheless, water splitting cannot be achieved by sunlight directly because the water molecule cannot be electronically excited by sunlight photons. PEC water splitting is described by semiconductor's absorption of light photons with energies greater than its band-gap energy. Once photoexcited electrons and holes are generated in the conduction and valence bands, respectively, they separate and migrate towards the surface sites where they reduce and oxidize adsorbed water to produce gaseous oxygen and hydro‐ gen. It has been demonstrated that various materials and composite structures can be utilized as photoelectrodes in PEC water-splitting processes. However, there are still obstacles to direct water splitting using PEC cells based on single absorber material: (i) lack of efficient light absorption due to wide band gap of semiconductors, (ii) efficiency reduction caused by overpotential and insufficient band alignments and (iii) corrosion of the semiconductor. Moreover, high cost of light-absorbing materials and co-catalysts (e.g., Pt, RuO2 and IrO2), which can split water with reasonable efficiencies prohibited their use on a large scale and over large areas required for light harvesting. Thereby, to raise the conversion efficiency, new lowcost and effective absorber materials and material compositions should be developed.

electrodes are employed in a vast array of electrochemical applications. Additionally, then doped with high concentration of boron atoms, BDD films possess metallic conductivity and

Nanocrystalline diamond layers were examined as anticorrosion protection of zirconium alloys nuclear fuel rods, operated in all commercial light-water and heavy-water nuclear reactors. Various corrosion tests were primarily focused on determining the ability of diamond layer to withstand emergency situations in nuclear reactors. The deposition of a polycrystal‐ line diamond layer was carried out using linear antenna microwave plasma-enhanced

chemical vapour deposition (LA-MW PECVD) apparatus (**Figure 2**) [1].

**Figure 2.** Linear antenna microwave plasma-enhanced chemical vapour deposition apparatus.

**applications**

**2. Doped nanocrystalline and epitaxial diamond layers for photovoltaic**

It was showed that for hydrogen fuel generation based on water electrolysis by means of a light-driven photoelectrochemical process, boron-doped diamond is considered very prom‐ ising material for electrodes [4]. The hydrogen produced using solar energy has been denomi‐ nated as solar fuel. In essence, the photoelectrochemical process mimics photosynthesis by splitting water into H2 and O2. Nevertheless, water splitting cannot be achieved by sunlight directly because the water molecule cannot be electronically excited by sunlight photons. PEC water splitting is described by semiconductor's absorption of light photons with energies greater than its band-gap energy. Once photoexcited electrons and holes are generated in the conduction and valence bands, respectively, they separate and migrate towards the surface sites where they reduce and oxidize adsorbed water to produce gaseous oxygen and hydro‐ gen. It has been demonstrated that various materials and composite structures can be utilized as photoelectrodes in PEC water-splitting processes. However, there are still obstacles to direct water splitting using PEC cells based on single absorber material: (i) lack of efficient light absorption due to wide band gap of semiconductors, (ii) efficiency reduction caused by

can also be optimized for light transparency [4, 5].

164 Advances in Carbon Nanostructures

Silicon (Si) is an appealing material for photochemical processes due to a large number of photons that can be absorbed and converted. Further, Si is one of the most abundant ele‐ ments in the Earth's crust. Thereto, Si is an attractive candidate for water-splitting platforms because of its prevalence in photovoltaic (PV) industries and predominance in electronics. Moreover, Si surpasses the majority of common semiconductor materials: the 1.1 eV band gap of Si allows effective light absorption from the UV to the near-infrared wavelengths region, thus resulting in a higher photocurrent and conversion efficiency.

Enabling Si as a photoelectrode for practical solar-fuel production from water mimics natural photosynthetic systems, where light absorption and catalytic reaction do not happen on the same material. Although, Si exhibits significant advantages over wide band-gap semiconduc‐ tors (TiO2, Fe2O3, WO3, etc.) utilized in PEC devices, it has seldom been used in photocatalyt‐ ic applications. The main reason is the significant corrosion of Si in non-acidic electrolytes; therefore, the realization of PEC device based on Si must account for its inherent corrosion. Recently, Si has been integrated as one of the main components of a novel concept of un‐ wired macroscopic PEC device based on silicon solar cell of a standard structure (n-Si, p-Si and p+ -Si). In such system, p+ side of Si solar cell is coated with corrosion resistant/conduc‐ tive indium tin oxide (ITO) layer and catalyst for triggering the oxygen evolution reaction (OER).This approach enables fabrication of Si-based PEC devices with solar-to-fuel (STF) conversion efficiencies exceeding 10%. However, the scarcity and declining availability of indium forces to look for alternative corrosion resistant and electrically conductive material solutions.

Boron-doped diamond is presently the subject of considerable interest as a promising electrode material [4, 5]. This is a consequence of several technologically important properties that distinguish it from common electrode materials used in electrochemistry, such as gold or platinum. Polycrystalline BDD films possess physical properties similar to those of bulk diamond, which includes outstanding hardness, high hole mobility, high thermal conductiv‐ ity and excellent resistance to radiation damage. Also, there is little evidence of degradation of BDD electrochemical activity with time. Due to these unique characteristics, to date BDD electrodes are employed in a vast array of electrochemical applications. Additionally, then doped with high concentration of boron atoms, BDD films possess metallic conductivity and can also be optimized for light transparency. It has been recently shown that a large poten‐ tial window of BDD can be tuned by the fabrication of sp2 -abundant polycrystalline BDD films. Also, it has been reported, that BDD electrodes can be successfully utilized for disinfection of wastewater. On the whole, the above-mentioned properties make BDD electrodes an ideal candidate for implementation in PV-assisted electrochemical systems for solar-fuel genera‐ tion by water decomposition and wastewater treatment.

Essentially, due to the favourable alignment of Si and BDD valence bands, a system based on Si solar cell and BDD electrode integrated in a PV-PEC device could be developed. Similarly to already reported PV-PEC systems with ITO layer, the rear side of the Si solar cell (p-type side) immersed into water has to be protected by electrically conductive and non-corrosive material, thus BDD layer acting as an anode has the capacity to render such demand.

Overall, this approach circumvents problems caused by the intrinsic properties of Si, such as corrosion in aqueous solution, minimizes the cost that is about one-third of the total PV system cost and could provide higher conversion efficiency as opposed to classical PEC systems.

**Figure 3.** Schematic representation of the water-splitting system based on Si solar cell equipped with boron-doped diamond electrode deposited on the p-side of the Si solar cell with closed circuit on the externally wired cathode.

A schematic representation of the concept of the light-driven water-splitting process realiz‐ ed by PV-assisted PEC cell equipped with BDD electrode is illustrated in **Figure 3**. Here, the Si solar cell is responsible forthe incident light absorption upon which photogenerated carriers are swept across the pn-junction. Due to a large band gap of highly boron-doped diamond films, set into contact with the p-side of Si, electrons are reflected and prohibited to move towards the BDD surface. Meanwhile, due to metallic doping of the BDD film, the Fermi level lies below the diamond valence band maximum that enables photogenerated hole carriers to pass freely through the BDD layer. Hence, holes are permitted to enter the water or electro‐ lyte solution. Thus, the BDD film is responsible for water oxidation process and PV Si/BDD device could operate as photoanode, whereas the n-side of Si could be protected by a trans‐ parent glass layer may be further connected by a wire to a secondary electrode (cathode) that can be kept in a separate compartment to assist the separation of hydrogen and oxygen gases. Additionally, various OER and hydrogen evolution reaction catalysts can be deposited on the BDD surface and on the cathode, to facilitate the overall water-splitting process. Finally, due to the fact that commercially available single-junction Si solar cell is able to deliver potential of only ~0.7 V, which is not sufficient for triggering the water-splitting reaction (1.23 V), a tandem or a triple-junction of the c-Si solar cells connected in series would be necessary.

Essentially, due to the favourable alignment of Si and BDD valence bands, a system based on Si solar cell and BDD electrode integrated in a PV-PEC device could be developed. Similarly to already reported PV-PEC systems with ITO layer, the rear side of the Si solar cell (p-type side) immersed into water has to be protected by electrically conductive and non-corrosive

Overall, this approach circumvents problems caused by the intrinsic properties of Si, such as corrosion in aqueous solution, minimizes the cost that is about one-third of the total PV system cost and could provide higher conversion efficiency as opposed to classical PEC systems.

**Figure 3.** Schematic representation of the water-splitting system based on Si solar cell equipped with boron-doped diamond electrode deposited on the p-side of the Si solar cell with closed circuit on the externally wired cathode.

A schematic representation of the concept of the light-driven water-splitting process realiz‐ ed by PV-assisted PEC cell equipped with BDD electrode is illustrated in **Figure 3**. Here, the Si solar cell is responsible forthe incident light absorption upon which photogenerated carriers are swept across the pn-junction. Due to a large band gap of highly boron-doped diamond films, set into contact with the p-side of Si, electrons are reflected and prohibited to move towards the BDD surface. Meanwhile, due to metallic doping of the BDD film, the Fermi level lies below the diamond valence band maximum that enables photogenerated hole carriers to pass freely through the BDD layer. Hence, holes are permitted to enter the water or electro‐ lyte solution. Thus, the BDD film is responsible for water oxidation process and PV Si/BDD device could operate as photoanode, whereas the n-side of Si could be protected by a trans‐ parent glass layer may be further connected by a wire to a secondary electrode (cathode) that can be kept in a separate compartment to assist the separation of hydrogen and oxygen gases. Additionally, various OER and hydrogen evolution reaction catalysts can be deposited on the BDD surface and on the cathode, to facilitate the overall water-splitting process. Finally, due

material, thus BDD layer acting as an anode has the capacity to render such demand.

166 Advances in Carbon Nanostructures

## **3. Conductivity of boron-doped polycrystalline diamond films: influence of specific boron defects**

Diamond, as a carbon-based material, seems to be an exceptional choice for many applica‐ tions, holding many promises due to its excellent properties, such as chemical inertness, wide electrochemical potential window and surface tunability by surface functionalization. Borondoped diamond is a p-type semiconductor with *E*g ~ 5.5 eV. Boron acceptor level in semicon‐ ducting diamond is about 0.36 eV from the valence band (EVB). Upon the metallic doping, The Fermi level (EF) shifts towards the valence band of diamond. Crystalline diamond has a strong and rigid isotropic structure due to its cubic crystal symmetry—sp3 -hybridized carbon atoms are bound by strong covalent bonds. Crystalline diamond is insulator with large band gap (more than 5 eV). Dopant atoms in the diamond matrices (usually boron or nitrogen) can strongly change electrical conductivity. The most widely used dopant is boron, conferring a p-type semiconducting character to diamond. The boron source can be a volatile compound introduced into the reactant gases, such as trimethyl borate or B2H6, or solid boron located near the substrate during the growth of the diamond film. The boron doping level is often in the range of 1018–1020 atoms/cm3 or a B/C ratio of about 105 to 103 . Boron-doped diamond has randomly oriented microcrystallites with facets and grain boundaries characteristic of a polycrystalline material. B in a diamond lattice acts as an electron acceptor providing the diamond p-type semiconducting properties. The acceptor level introduced by B is quite deep 0.37 eV from the top of the valence band. At higher doping levels, conduction occurs by nearest neighbour and variable range hopping of holes between ionised B sites. Further addition of B results in a layer with conductivity comparable to metallic materials.

In the study of Ashcheulov et al. [5], the conductivity of boron-doped polycrystalline films was studied. BDD films were prepared by microwave plasma-enhanced chemical vapour deposition. Diamond is prepared by the decomposition of a mixture of methane (or other carbon-containing species) at pressures from 0.01 to 100 mbar and substrate temperatures of 250 to 800°C. If the chemicalreaction takes place in an ionized plasma, the ionized gas requires much less energy (lower temperatures) to react. Ionization energy is generally provided by a high-frequency field.

Concerning BDD film conductivity, the following factors are important: (i) size of the diamond grains—BDD films containing bigger grains have bigger conductivity (for the same B content); (ii) grain boundaries with non-diamond carbon impurities; (iii) structures of boron defects and setting of the material; (iv) the surface of each diamond grain can have a higher concentration of B, allowing specific conduction. Resistance of boron-doped diamond layers as a function of boron concentration is shown in **Figure 4**.

**Figure 4.** Resistance of boron-doped diamond layers as a function of boron concentration. Inset: Relation between the real boron concentration and the B/C ratio in the gas phase as measured by NDP.

Using a combination of theoretical and experimental techniques (plane-wave density func‐ tional theory, NDP, resistivity and Hall effect measurements, AFM, Raman spectroscopy), B defect location, vibration states, conductivity, free hole concentration inside the grains and hole mobility were investigated in [5]. Based on the analysis of specific structural, bond, interactions, charge localization parameters and spins, an explanation of the complex influence of specific B defects on the system charge transport mechanisms was set out:


## **4. Protection of zirconium nuclear fuel cladding by polycrystalline diamond film of nanosize thickness**

A 300 nm thick polycrystalline diamond layer has been used for the protection of zirconium alloy nuclear fuel cladding against undesirable oxidation with no loss of chemical stability and preservation of its functionality [2, 3, 6].

Reaction of fuel claddings with hot water steam triggered explosions of hydrogen gas in nuclear power reactors on several occasions, including the recent Fukushima accident. The probability of such accidents can be drastically lowered if the surface of the fuel claddings is coated with a suitable protective layer. In this chapter, we demonstrate that the 300 nm polycrystalline diamond layer on the surface of a Zr cladding rod improves the chemical stability of the cladding material while preserving its functionality.

Zirconium alloys are used in all commercial light-water (PWR, BWR and VVER) and heavywater (CANDU) nuclear reactors. Nuclear fuel cladding (NFCs) are typically produced from zirconium alloys (with a common subgroup trademarked Zircaloy2), a material with very low absorption cross-section of thermal neutrons, high ductility, hardness and corrosion resist‐ ance [1–4]. Zircaloy2 serves faultlessly under standard operation conditions ofthe plant sealing the uranium oxide fuel pellets. Normal operating temperatures in the reactor is about 300°C. In the case of some accident at temperatures above 800°C the so-called high-temperature corrosion begins; in this case, the peeling of the oxide layer previously protecting the metal from oxidation occurs, which can result in mechanical failure of the system. It is strongly exothermic and high auto-catalytic reaction between zirconium and steam, during which the dissociation of steam molecules occurs and causes the formation of zirconium dioxide, hydrogen and releases a large amount of heat. The result of the reaction is not only the formation of hydrogen, which, as a combustible gas, is a serious risk in the case of a severe accident (like in recent Fukushima accident), but also the release of large quantities of heat, which further complicates the core cooling and strengthens the further course of the hightemperature oxidation of zirconium alloys. Finally, there is also the degradation of the fuel cladding, one of the protective barriers that may lead to its failure and subsequent leakage of highly radioactive fission products from nuclear fuel into the primary circuit. Many protec‐ tive materials have been applied on the Zr alloys surfaces, but without any significant success. Just by applying a homogenous polycrystalline diamond layer on the surface of Zircaloy2, the Zircaloy2 cladding was successfully protected against undesirable changes and processes in the nuclear reactor environment. The layer of polycrystalline diamond will also prevent the reaction between zirconium and water steam. The deposition of polycrystalline diamond layer was carried out using LA-MW PECVD apparatus. In our recent work we demonstrated success: the 300 nm composite polycrystalline diamond layer with high crystalline diamond content and low roughness can protect zirconium alloy's nuclear fuel cladding against undesirable oxidation and consolidate its chemical stability while preserving its functionali‐ ty. After ion beam irradiation (10 dpa, 3 MeV Fe2+), the diamond film still showed satisfacto‐ ry structural integrity with both sp3 and sp2 carbon phases— it was confirmed by Raman spectroscopy (**Figure 5**). The zirconium alloys under the carbon-based protective layer after hot steam oxidation test (1100°C, 30 min) differed from the original zirconium alloy's material composition only very slightly—the diamond coating increases the material resistance towards high-temperature oxidation by forming a protective composite structure. We also found that the presence of the PCD coating blocks hydrogen diffusion into the Zr surface and protects Zr material from degradation. By more closely monitoring the oxidation rate and phase change of PCD-coated Zircaloy2 fuel cell rods as a function of temperature, we found that even attemperatures above 850°C (Zr beta phase thatis more openedfor oxygen/hydrogen

**Figure 4.** Resistance of boron-doped diamond layers as a function of boron concentration. Inset: Relation between the

Using a combination of theoretical and experimental techniques (plane-wave density func‐ tional theory, NDP, resistivity and Hall effect measurements, AFM, Raman spectroscopy), B defect location, vibration states, conductivity, free hole concentration inside the grains and hole mobility were investigated in [5]. Based on the analysis of specific structural, bond, interactions, charge localization parameters and spins, an explanation of the complex influence

**•** When the concentration of B in the PCD film is low, hopping conduction using the excited level of B defects is predominant. When the concentration of boron atoms in the diamond lattice is high enough so the wave functions of B overlap and form an impurity band passing to the top of the valence band metal-like conductivity behaviour of the material is reached. **•** High boron concentration in diamond lattice can support complex defects (B dimers in deformed lattice). B dimers have high acceptor energy so electrons excitation from top of the valence band is very improbable. The impurity states do not merge with the valence band, and free carrier concentration is reduced. Moreover, due to compensated spins of 2s

and 2p electrons in B dimmers, the probability of charge hopping is also reduced.

**4. Protection of zirconium nuclear fuel cladding by polycrystalline**

A 300 nm thick polycrystalline diamond layer has been used for the protection of zirconium alloy nuclear fuel cladding against undesirable oxidation with no loss of chemical stability and

**diamond film of nanosize thickness**

preservation of its functionality [2, 3, 6].

real boron concentration and the B/C ratio in the gas phase as measured by NDP.

168 Advances in Carbon Nanostructures

of specific B defects on the system charge transport mechanisms was set out:

diffusion) the Zircaloy2 surface could be effectively protected against hot steam corrosion. PCD anticorrosion protection of Zr nuclear fuel rods can significantly (30%) prolong life of Zr material in nuclear reactors even above phase transition.

**Figure 5.** Comparison of Raman spectra of the PCD coating (black curve) and PCD after ion beam irradiation (red curve). Both sp2 and sp3 C (1332 cm1 in Raman spectra) phases were present in the PCD layer.

After hot steam oxidation (950°C), a high level of structural integrity of the PCD layer was observed. Both sp2 and sp3 C (1332 cm1 in Raman spectra) phases were present in the protec‐ tive PCD layer. A greater relative weight gain of oxidized unprotected Zircaloy2 compared to that of the oxidized PCD-coated Zircalloy2 samples was supported by a thicker oxide layer of unprotected samples measured by impedance spectroscopy. The PCD layer blocks the hydrogen diffusion into the Zircaloy2 surface thus protecting the material from degradation. Hot steam oxidation tests confirmed that PCD-coated Zircaloy2 surfaces were effectively protected against corrosion. PCD film surface coating can prolong the service life of Zr nuclear fuel claddings in nuclear reactors.

Even under standard operating conditions in nuclear reactors (400°C hot water vapour), the polycrystalline diamond layer retains its original properties and participates both in the removal of heat released during the reactor operating mode, and further protects the coated surface against undesirable chemical reactions and changes in the composition of the struc‐ ture related to the diffusion of hydrogen atoms from dissociated water molecules to zirconi‐ um alloy—as shown in **Figures 6** and **7**. The polycrystalline diamond layer further reduces especially adverse high-temperature chemical reactivity of the zirconium alloy surface, thereby also reducing high-temperature dissociation of water vapour molecules and subse‐ quent formation of zirconium dioxide and explosive hydrogen. In the case of temperatureinduced changes in zirconium alloy, the layer will benefit from the mixed nature of the protective carbon layer, which besides crystalline diamond grains with sp3 -hybridized carbon also contains flexible amorphous phase sp2 from hybridized carbon capable of good adapta‐ tion to volume changes/expansion of the metal substrate without compromising the protec‐ tive layer integrity (**Figure 8**).

diffusion) the Zircaloy2 surface could be effectively protected against hot steam corrosion. PCD anticorrosion protection of Zr nuclear fuel rods can significantly (30%) prolong life of Zr

**Figure 5.** Comparison of Raman spectra of the PCD coating (black curve) and PCD after ion beam irradiation (red

After hot steam oxidation (950°C), a high level of structural integrity of the PCD layer was

tive PCD layer. A greater relative weight gain of oxidized unprotected Zircaloy2 compared to that of the oxidized PCD-coated Zircalloy2 samples was supported by a thicker oxide layer of unprotected samples measured by impedance spectroscopy. The PCD layer blocks the hydrogen diffusion into the Zircaloy2 surface thus protecting the material from degradation. Hot steam oxidation tests confirmed that PCD-coated Zircaloy2 surfaces were effectively protected against corrosion. PCD film surface coating can prolong the service life of Zr nuclear

Even under standard operating conditions in nuclear reactors (400°C hot water vapour), the polycrystalline diamond layer retains its original properties and participates both in the removal of heat released during the reactor operating mode, and further protects the coated surface against undesirable chemical reactions and changes in the composition of the struc‐ ture related to the diffusion of hydrogen atoms from dissociated water molecules to zirconi‐ um alloy—as shown in **Figures 6** and **7**. The polycrystalline diamond layer further reduces especially adverse high-temperature chemical reactivity of the zirconium alloy surface, thereby also reducing high-temperature dissociation of water vapour molecules and subse‐ quent formation of zirconium dioxide and explosive hydrogen. In the case of temperatureinduced changes in zirconium alloy, the layer will benefit from the mixed nature of the

in Raman spectra) phases were present in the PCD layer.

in Raman spectra) phases were present in the protec‐

from hybridized carbon capable of good adapta‐


material in nuclear reactors even above phase transition.

curve). Both sp2

observed. Both sp2

and sp3

170 Advances in Carbon Nanostructures

fuel claddings in nuclear reactors.

also contains flexible amorphous phase sp2

C (1332 cm1

and sp3

C (1332 cm1

protective carbon layer, which besides crystalline diamond grains with sp3

**Figure 6.** Comparison of Raman spectra of the PCD coating (black curve) and after hot steam oxidation (blue curve). Hot steam oxidation causes sp2 transition of the layer.

**Figure 7.** The 488 nm Raman spectra of PCD covered Zr alloy sample after 6 days in 400°C hot vapour/reactor testing. Spectra were taken from various parts of the tube. Both sp2 and sp3 C (1332 cm1 in Raman spectra) phases were present in the PCD layer.

**Figure 8.** Left: A sample of PCD-coated zirconium fuel cladding; right: a lengthwise cut of a sample of PCD-coated zirconium fuel cladding, which after being exposed to water vapour at 400°C for 3 days was completely straightened. The PCD layer has withstood even this strain without compromising its integrity. The scratch in the right lower corner (right) was made using a diamond point pen.

## **5. Conclusions**

From carbon nanomaterials especially polycrystalline diamond films have attracted more and more interest due to their unique electrical, optical and mechanical properties, which make them widely used for different applications. Diamond survives high temperatures, does not degrade over time and (important for the nuclear fuel cladding) has perfect neutron crosssection properties. Moreover, polycrystalline diamond layers consisting of crystalline (sp3 ) and amorphous (sp2 ) carbon phases could have suitable thermal expansion. Polycrystalline diamond films can be used in a wide range of applications: high boron- and phosphorus-doped nanocrystalline diamond layers that can have a significant impact in the industrial use of diamond in electrochemical applications. Boron-doped nanocrystalline diamond-coated silicon solar cells can be used for direct photoelectrochemical water splitting with high effectivity. In such a system, boron-doped diamond electrodes are responsible for water decomposition. Diamond could also be considered as a prospective material for electrodes and organic solar cells. Diamond layers were used as anticorrosion protection for zirconium alloys applied in nuclear reactors. Polycrystalline diamond anticorrosion protection of Zr alloys can significantly prolong life ofZr material in nuclearreactors even attemperatures aboveZr phase transition temperatures. Zr alloy's oxidation and hydrogenation lead to limitation criteria in justifying fuelrod operability—rods containing even non-irradiated nuclearfuel must be fired due to Zr alloy surface destruction by oxidation/corrosion in reactor. Thus, the protection of zirconium fuel rods' surfaces against corrosion in nuclear reactors' environment can protect nuclear fuel to important extent.

## **Author details**

Irena Kratochvilova

Address all correspondence to: krat@fzu.cz

Institute of Physics, Academy of Sciences Czech Republic v.v.i, Prague, Czech Republic

## **References**

**Figure 8.** Left: A sample of PCD-coated zirconium fuel cladding; right: a lengthwise cut of a sample of PCD-coated zirconium fuel cladding, which after being exposed to water vapour at 400°C for 3 days was completely straightened. The PCD layer has withstood even this strain without compromising its integrity. The scratch in the right lower corner

From carbon nanomaterials especially polycrystalline diamond films have attracted more and more interest due to their unique electrical, optical and mechanical properties, which make them widely used for different applications. Diamond survives high temperatures, does not degrade over time and (important for the nuclear fuel cladding) has perfect neutron crosssection properties. Moreover, polycrystalline diamond layers consisting of crystalline (sp3

diamond films can be used in a wide range of applications: high boron- and phosphorus-doped nanocrystalline diamond layers that can have a significant impact in the industrial use of diamond in electrochemical applications. Boron-doped nanocrystalline diamond-coated silicon solar cells can be used for direct photoelectrochemical water splitting with high effectivity. In such a system, boron-doped diamond electrodes are responsible for water decomposition. Diamond could also be considered as a prospective material for electrodes and organic solar cells. Diamond layers were used as anticorrosion protection for zirconium alloys applied in nuclear reactors. Polycrystalline diamond anticorrosion protection of Zr alloys can significantly prolong life ofZr material in nuclearreactors even attemperatures aboveZr phase transition temperatures. Zr alloy's oxidation and hydrogenation lead to limitation criteria in justifying fuelrod operability—rods containing even non-irradiated nuclearfuel must be fired due to Zr alloy surface destruction by oxidation/corrosion in reactor. Thus, the protection of zirconium fuel rods' surfaces against corrosion in nuclear reactors' environment can protect

) carbon phases could have suitable thermal expansion. Polycrystalline

) and

(right) was made using a diamond point pen.

**5. Conclusions**

172 Advances in Carbon Nanostructures

amorphous (sp2

nuclear fuel to important extent.


**Nanofluids and Case Studies on Carbon Nanostructures**

## **Nanofluids Based on Carbon Nanostructures**

Hammad Younes, Amal Al Ghaferi, Irfan Saadat and Haiping Hong

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64553

#### **Abstract**

A nanofluid consists in a liquid suspension of nanometer-sized particles. These fluids may contain (or not) surface-active agents to aid in the suspension of the particles. Nanometer-sized particles have higher thermal conductivity than the base fluids. Oxides, metals, nitrides, and nonmetals, like carbon nanotubes, can be used as nanoparticles in nanofluids. Water, ethylene glycol, oils, and polymer solutions can be used as base fluids. In this chapter, we summarize the recent studies of using CNTs and graphene to improve the thermal conductivity of nanofluids. Moreover, we refer to the studies about the effect of using magnetic fields on enhancing the thermal conductivity of nanofluids. Too much discrepancy about thermal conductivity of nanofluids can be found in the literature. For carbon nanofluids, unfortunately, no significant improvements on thermal conductivi‐ ty are observed using low concentrations. Different improvement percentages have been reported. This variation in the thermal conductivity can be attributed to many factors, such as particle size temperature, pH, or zeta potential. We believe that more research efforts need to be made in order to, first, improve the thermal conductivity of nano‐ fluids and, second, assess the effect of the different parameters and conditions on the thermal conductivity of nanofluids.

**Keywords:** CNTs, graphene, nanofluids, thermal conductivity, alignment and mag‐ netic field

## **1. Introduction**

The enhancement of the efficiency of heat transfer fluids has been of great interest to scientists and engineers for the past two decades. Conventional heat transfer fluids, such as mineral oil, water, and ethylene glycol (EG), have poor heat transfer properties, compared to solids. In recent years, nanofluids have been investigated as potential heat transfer fluids [1–11].

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The conventional coolants/fluids used in radiators and engines have poor heat transfer properties and contain millimeter- or micrometer-sized particles to improve the heat transfer properties that can clog new cooling technologies that include microchannels. The use of fluids with better heat transfer can allow engines to run at optimal temperatures and allow the use of smaller and lighter radiators, pumps, and other components. Smaller radiators allow a smaller front end of large vehicles, such as trucks. Fluids with better heat transfer can also lead to the production of lighter vehicles with better fuel economy, since less pumping power is needed. Although many factors such as pH, particle size, and zeta potential can affect the thermal conductivity (TC) of nanofluids, they are not promising to improve thermal conduc‐ tivity values [12–16].

Wensel et al. [17] found that using low concentrations of the nanoparticles do not enhance the TC effectively. For example, using of 4–5 vol.% metal oxides increases the thermal conductivity by 10–20% only, which is not the significant increase that is needed.

## **2. Carbon-based nanomaterials**

In this chapter, two carbon-based nanomaterial nanofluids will be discussed: carbon nanotube (CNT) and graphene-based nanofluids and their thermal conductivity.

### **2.1. Carbon nanotubes**

Carbon nanotubes (CNTs) have attracted significant attention since they were discovered by Iijima in 1991 [18]. The sp2 carbon–carbon bond in the plane of the graphene lattice is the strongest among of all chemical bonds. CNTs can be obtained with different numbers of walls, including single-walled CNTs, double-walled CNTs, and multiwall CNTs. They have very good thermal conductivity along the length of the tubes. It has been shown that the thermal conductivity of CNTs is unusually high (e.g., 3000 W/mK) for MWNTs [19] and even higher for SWCNTs (6000 W/mK). The thermal conductivity of CNTs is more than seven times higher than that of copper, which is well known for having a good thermal conductivity of 385 W/Mk [19–22]. Because of their remarkable thermal, electrical, mechanical properties, and low density, carbon nanotubes are ideal materials for reinforcement in composites [23–27], sensors [28], and many other applications, such as thermal managements [29–31].

#### *2.1.1. Carbon nanotubes and carbon nanotubes metal oxide-based nanofluids*

Due to the high thermal conductivity of CNTs, researchers started to use them to prepare nanofluids in order to improve the heat transfer. CNTs tend to aggregate into groups due to the large surface energy (strong van der Waals attractions between individual tubes) [32]. Therefore, a variety of approaches has been introduced to decrease the nanotubes agglomeration, namely, the modification of their chemistry through noncovalent adsorption using surfactants [33–36], covalent (functionalization) by chemical modification [37–42], and metal coating (like Ni-coated SWNTs). **Table 1** summarizes the work carried out so far dealing with the use of CNTs in nanofluids. Xie et al. [52] investigated the functionalized carbon nanotubes and found out that the thermal conductivity enhancement increased with the increase in nanotubes loading. In addition, it was found that the thermal conductivity decreased with the increase in thermal conductivity of the base fluid. Chen et al. [46] studied the effect of MWNTs on the thermal conductivity enhancement of ethylene glycol-based nanofluid and found that the enhancement reached up to 17.5% at 0.01 volume fraction. Ding et al. [49] observed a significant enhancement of the convective heat transfer in comparison with pure water as the working fluid, which depends on the flow condition, CNT concentration, and the pH level (although the latter is small). Grag et al. [44] studied the effect of ultrasonication on the thermal conductivity of MWCNT fluids and found that the thermal conductivity was enhanced by 20% when the ultrasonication was used for 40 min in 1 wt.% MWCNT concentration, in a 130 W, 20 kHz ultrasonication environment.

The conventional coolants/fluids used in radiators and engines have poor heat transfer properties and contain millimeter- or micrometer-sized particles to improve the heat transfer properties that can clog new cooling technologies that include microchannels. The use of fluids with better heat transfer can allow engines to run at optimal temperatures and allow the use of smaller and lighter radiators, pumps, and other components. Smaller radiators allow a smaller front end of large vehicles, such as trucks. Fluids with better heat transfer can also lead to the production of lighter vehicles with better fuel economy, since less pumping power is needed. Although many factors such as pH, particle size, and zeta potential can affect the thermal conductivity (TC) of nanofluids, they are not promising to improve thermal conduc‐

Wensel et al. [17] found that using low concentrations of the nanoparticles do not enhance the TC effectively. For example, using of 4–5 vol.% metal oxides increases the thermal conductivity

In this chapter, two carbon-based nanomaterial nanofluids will be discussed: carbon nanotube

Carbon nanotubes (CNTs) have attracted significant attention since they were discovered by

strongest among of all chemical bonds. CNTs can be obtained with different numbers of walls, including single-walled CNTs, double-walled CNTs, and multiwall CNTs. They have very good thermal conductivity along the length of the tubes. It has been shown that the thermal conductivity of CNTs is unusually high (e.g., 3000 W/mK) for MWNTs [19] and even higher for SWCNTs (6000 W/mK). The thermal conductivity of CNTs is more than seven times higher than that of copper, which is well known for having a good thermal conductivity of 385 W/Mk [19–22]. Because of their remarkable thermal, electrical, mechanical properties, and low density, carbon nanotubes are ideal materials for reinforcement in composites [23–27], sensors

Due to the high thermal conductivity of CNTs, researchers started to use them to prepare nanofluids in order to improve the heat transfer. CNTs tend to aggregate into groups due to the large surface energy (strong van der Waals attractions between individual tubes) [32]. Therefore, a variety of approaches has been introduced to decrease the nanotubes agglomeration, namely, the modification of their chemistry through noncovalent adsorption using surfactants [33–36], covalent (functionalization) by chemical modification [37–42], and metal coating (like Ni-coated SWNTs). **Table 1** summarizes the work carried out so far dealing with the use of CNTs in nanofluids. Xie et al. [52] investigated the functionalized carbon

carbon–carbon bond in the plane of the graphene lattice is the

by 10–20% only, which is not the significant increase that is needed.

(CNT) and graphene-based nanofluids and their thermal conductivity.

[28], and many other applications, such as thermal managements [29–31].

*2.1.1. Carbon nanotubes and carbon nanotubes metal oxide-based nanofluids*

tivity values [12–16].

1782 Advances in Carbon Nanostructures

**2.1. Carbon nanotubes**

Iijima in 1991 [18]. The sp2

**2. Carbon-based nanomaterials**


**Table 1.** Current research of carbon nanomaterials-based nanofluids.

Phuoc et al. [43] studied the thermal conductivity, viscosity, and stability of nanofluids made of MWCNT dispersed in water. The nanotubes have an outside diameter of 20–30 nm, inside diameter of 5–10 nm, length of 10–30 mm, and average density of 2.1 g/cm3 . Different weight ratio of Chitosan surfactant, namely, 0.1, 0.2, and 0.5 wt.%, was used to disperse MWCNTs in water. It was found that Chitosan dispersed and stabilized MWNTs in water efficiently; as only 0.1 wt.% Chitosan surfactant could disperse and stabilize 3 wt.% MWCNTs in water. The thermal conductivity of nanofluids containing 0.5–3 wt.% MWCNTs was enhanced from 2.3 to 13%. However, the enhancement of the thermal conductivity was independent of viscosity of the base fluid, which shows that the particle velocity does not have a significant effect on the thermal conductivity. **Figure 1** shows the dispersion behavior of the MWCNTs in deionized (DI) water (DW).

**Figure 1.** Nanofluid prepared by dispersing MWCNTs in deionized water (A) shows that MWNTs deposited at the bottom of the container and (B) shows that MWNTs maintain good dispersion: From Ref. [43].

Garg et al. [44] reported that 1 wt.% MWCNT increased the thermal conductivity up to 20% at 35°C. Gum arabic was used as a surfactant to stabilize the nanofluids. The study revealed that the thermal conductivity of CNT nanofluids increased considerably after 24°C. Meng et al. [45] found a 25.4% enhancement of the thermal conductivity of CNT glycol nanofluids with 4.0% mass fraction at room temperature. **Figure 2** shows that the transmittances from 200 to

**Figure 2.** TEM image of HNO3-treated CNTs (A) and UV–Vis–NIR transmittance spectra (B) of the CNTs glycol nano‐ fluids. From ref. [45].

1500 nm, the values drop from 50% to about 10%, indicating that a small amount of CNTs results in significant absorption enhancement of light.

Phuoc et al. [43] studied the thermal conductivity, viscosity, and stability of nanofluids made of MWCNT dispersed in water. The nanotubes have an outside diameter of 20–30 nm, inside

ratio of Chitosan surfactant, namely, 0.1, 0.2, and 0.5 wt.%, was used to disperse MWCNTs in water. It was found that Chitosan dispersed and stabilized MWNTs in water efficiently; as only 0.1 wt.% Chitosan surfactant could disperse and stabilize 3 wt.% MWCNTs in water. The thermal conductivity of nanofluids containing 0.5–3 wt.% MWCNTs was enhanced from 2.3 to 13%. However, the enhancement of the thermal conductivity was independent of viscosity of the base fluid, which shows that the particle velocity does not have a significant effect on the thermal conductivity. **Figure 1** shows the dispersion behavior of the MWCNTs in deionized

**Figure 1.** Nanofluid prepared by dispersing MWCNTs in deionized water (A) shows that MWNTs deposited at the

Garg et al. [44] reported that 1 wt.% MWCNT increased the thermal conductivity up to 20% at 35°C. Gum arabic was used as a surfactant to stabilize the nanofluids. The study revealed that the thermal conductivity of CNT nanofluids increased considerably after 24°C. Meng et al. [45] found a 25.4% enhancement of the thermal conductivity of CNT glycol nanofluids with 4.0% mass fraction at room temperature. **Figure 2** shows that the transmittances from 200 to

**Figure 2.** TEM image of HNO3-treated CNTs (A) and UV–Vis–NIR transmittance spectra (B) of the CNTs glycol nano‐

bottom of the container and (B) shows that MWNTs maintain good dispersion: From Ref. [43].

. Different weight

diameter of 5–10 nm, length of 10–30 mm, and average density of 2.1 g/cm3

(DI) water (DW).

1804 Advances in Carbon Nanostructures

fluids. From ref. [45].

Chen et al. [46] reported that 1 vol.% of MWCNTs without surfactant, placed in ethylene glycolbased nanofluid, yielded an enhancement of 17.5%. A mechano-chemical reaction method was used to enhance the MWCNTs dispersibility for producing a CNT nanofluid. **Figure 3** shows the thermal conductivities of nanofluids as a function of nanotube loadings in two different fluids: deionized water and ethylene glycol. The enhancement using ethylene glycol as fluid and without surfactant at 1 vol.% of MWCNTs is higher than the enhancement using DW as fluid and at 1 vol.% of MWCNTs.

**Figure 3.** Thermal conductivities of nanofluids as a function of nanotube loadings. From ref. [47].

Chen and Xie [47] investigated nanofluids made of SWCNTs and DWCNTs in water-based fluids, respectively. Those CNTs were functionalized by a wet-mechano-chemical reaction method in order to obtain a stable dispersion. The study revealed that the thermal conductivity enhancement was 15.6, 14.2, and 12.1% for SWCNT, DWCNT, and MWCNT at a volume fraction of 0.2%, respectively.

Choi et al. [12] reported that the enhancement of thermal conductivity at a 1% volume fraction of MWCNTs in poly (α-olefin) (PAO) oil was 160%, compared to the base fluid at room temperature.

Assael et al. [48] found a 38% enhancement of the thermal conductivity in 0.6 vol.% MWNTs nanofluids in which water was a base fluid and SDS surfactant was used to stabilize the nanofluids. Ding et al. [49] confirmed that 1.0 vol.% of MWCNT dispersed in water, with SDBS surfactant, lead to an enhancement of 79% at 30°C. Shaikh et al. [50] prepared nanofluids by dispersing CNTs, exfoliated graphite (EXG), and heat-treated nanofibers (HTT) in PAO oil. The thermal conductivity for 1% CNTs, EXG, and HTT was enhanced by 161, 131, and 104 %, respectively. Baby and Sundara [51] reported on the synthesis of silver nanoparticle decorated MWCNT-graphene mixture and dispersed the (Ag/(MWNT-HEG)) composite in ethylene glycol without surfactant. They found that only 0.04 vol.% of the nanoparticles enhanced the thermal conductivity by ∼8% at 25°C.

#### **2.2. Graphene**

Graphene, the first ever stable two-dimensional (2D) honeycomb lattice of sp2 -bonded carbon atoms, has attracted the interest of many researchers and is very promising for a wide range of applications due to its spectacular electronic, mechanical, and optical properties. Since few layers of graphene were isolated for the first time from graphite, using a tape in 2004 [53], the graphene production rate has increased rapidly in order to synthesize large-area and highquality graphene films, compatible with various applications. Actually, the history of graphene dates back to the 1960s, when single-atom plane of graphite, termed as graphite layer, was isolated. For several decades, the existence of 2D atomic structures remained as an unaccepted concept. Now the term of graphene, the basic block of graphite, is widely used to describe the 2D monolayer of carbon atoms with a honeycomb structure. Graphene lattice consists of repetitive two identical carbon atoms, A and B, that together form one unit cell (**Figure 4**). The distance between each carbon atom and its neighbors is 1.42 A ´ [54]. **Figure 5** shows first Brillouin zone and band structure of graphene using tight binding approximation. *K* and *K′* are Dirac points, the transition between the valence and the conduction band.

**Figure 4.** Graphene lattice.

**Figure 5.** First Brillouin zone and band structure of graphene using tight binding approximation. *K* and *K*' are Dirac points, the transition between the valence and the conduction band. Reproduced from [54].

#### *2.2.1. Graphene and graphene metal oxide-based nanofluids*

**2.2. Graphene**

1826 Advances in Carbon Nanostructures

**Figure 4.** Graphene lattice.

Graphene, the first ever stable two-dimensional (2D) honeycomb lattice of sp2

distance between each carbon atom and its neighbors is 1.42 A

are Dirac points, the transition between the valence and the conduction band.

atoms, has attracted the interest of many researchers and is very promising for a wide range of applications due to its spectacular electronic, mechanical, and optical properties. Since few layers of graphene were isolated for the first time from graphite, using a tape in 2004 [53], the graphene production rate has increased rapidly in order to synthesize large-area and highquality graphene films, compatible with various applications. Actually, the history of graphene dates back to the 1960s, when single-atom plane of graphite, termed as graphite layer, was isolated. For several decades, the existence of 2D atomic structures remained as an unaccepted concept. Now the term of graphene, the basic block of graphite, is widely used to describe the 2D monolayer of carbon atoms with a honeycomb structure. Graphene lattice consists of repetitive two identical carbon atoms, A and B, that together form one unit cell (**Figure 4**). The

Brillouin zone and band structure of graphene using tight binding approximation. *K* and *K′*

**Figure 5.** First Brillouin zone and band structure of graphene using tight binding approximation. *K* and *K*' are Dirac

points, the transition between the valence and the conduction band. Reproduced from [54].


´ [54]. **Figure 5** shows first

The high thermal conductivity of graphene encouraged researchers to use it to prepare nanofluids with super high thermal conductivity. Yu et al. [55] made stable nanofluids by dispersing graphene oxide nanosheets in ethylene glycol and found out that the thermal conductivity of the nanofluids was higher than that of the ethylene-glycol base fluid. The thermal conductivity of the nanofluids was enhanced up to 61.0% at 5.0 vol.% of graphene nanosheets (GNs), and almost constant for 7 days, indicating that the hydrophilic surfaces allow graphene oxide nanosheets to have good compatibility with ethylene glycol. Yu et al. [56] used a facial technique to prepare ethylene glycol-based nanofluids containing graphene nanosheets. They found that the thermal conductivity of the graphene nanofluids increased by 86% at 5.0 vol.% graphene, showing that the extensive presence of saturated sp3 bonds and oxygen atoms made graphene oxide nonconducting, hindered the thermal transport, and promoted phonon scattering. Mehrali et al. [57] prepared stable nanofluids by mixing graphene nanoplatelets and water using a tip ultrasonicator. In this study, four concentrations were prepared: 0.025, 0.05, 0.075, and 0.1 wt.% for three different specific surface areas of 300, 500, and 750 m2 /g. It was found that the highest thermal conductivity was 27.64% for a concentration of 0.1 wt.%, with a specific surface area of 750 m2 /g. Aravind and Ramaprabhu [58] adopted a simple thermal treatment method to prepare graphene nanosheets (GNs)-based nanofluids. The procedure was carried out without surfactant and harsh chemical treatments, which reduced the alkaline pH of graphite oxide suspension in ethylene glycol (EG) and deionized (DI) water-based fluids. It was found that the thermal conductivity at 25°C and 0.14% volume fraction of GN in EG and DI water is enhanced by 6.5 and 13.6%, respectively. Ijam et al. [59] prepared a glycerol-water-based nanofluid containing graphene oxide nanosheets. The mixture was stable for up to 5 months. The thermal conductivity of the prepared nanofluid with different temperatures (25–45°C) and weight fractions (0.02–0.1 wt. %) revealed an enhancement of thermal conductivity of 4.5% at 25°C for a weight fraction of 0.02 and 11.7% for a weight fraction of 0.1 wt.% and 45°C, respectively. Kole et al. [60] studied the thermal and electrical conductivity of stable and well dispersed functionalized grapheneethylene glycol. High purity graphite powder was used to make Graphene nanosheets. Then the Hummers' method [61] was followed, and exfoliation and reduction by hydrogen gas were used to obtain hydrogen-exfoliated graphene (HEG), which was functionalized with acid. The material was dispersed in distilled water nanofluids with volume concentration between 0.041 and 0.395 vol.%. It was found that the thermal conductivity at 0.395 vol.% was enhanced by 15%, and the thermal conductivity of both the base fluid and the prepared nanofluid increased linearly with temperature. However, the thermal conductivity ratios (TC of the nanofluids over the TC of base fluids) were nearly independent of temperature. The study also showed that the electrical conductivity of the f-HEG nanofluids had a significant enhancement of 8620% at 0.395 vol.% loading of f-HEG, in a base fluid of 70:30 mixture of EG and distilled water. Hajjar et al. [62] studied the thermal conductivity of graphene oxide nanofluids, which comprised of graphene oxide obtained by the Hummers' method and water. They found that 0.25 wt.% of graphene oxide could enhance the thermal conductivity by 33.9% at 20°C and up to 47.5% at 40°C. Ghozatloo et al. [63] investigated the stability and thermal conductivity behavior of the nanofluids made of graphene- and water-based fluids. They also analyzed the influence of time and temperature on the effective thermal conductivities for different concentrations of graphene that was functionalized in mild conditions by potassium persulfate. The thermal conductivity for 0.05 wt.% of the functionalized graphene is enhanced around 14.1% at 25°C and 17% at 50°C compared with water. Liu et al. [64] studied the thermodynamic properties (including thermal conductivity, viscosity, specific heat, and density) of graphene-dispersed nanofluids for temperatures ranging from room temperature to around 200°C. The studied nanofluids were made out of graphene and the ionic liquid 1 hexyl-3-methylimidazolium tetra fluoroborate ((HMIM)BF4). The work showed that the thermal conductivity of the nanofluid containing 0.06 wt.% graphene increased by 15.2–22.9% at the tested temperature range (25–200°C). Viscosity, specific heat, and density decreased, respectively, by 4.6–13.1, 3, and 2.8%, compared to (HMIM)BF4.

## **3. Effect of the alignment of the carbon nanostructures on the thermal conductivity of nanofluids**

Many researchers have studied the effect of the alignment of iron oxide nanoparticles and carbon nanomaterials on the thermal, electrical, and mechanical properties of fluids and polymers. **Table 2** summarizes most work about the effect of the alignment of iron oxide nanoparticles and carbon nanomaterials on the thermal conductivity of nanofluids.



influence of time and temperature on the effective thermal conductivities for different concentrations of graphene that was functionalized in mild conditions by potassium persulfate. The thermal conductivity for 0.05 wt.% of the functionalized graphene is enhanced around 14.1% at 25°C and 17% at 50°C compared with water. Liu et al. [64] studied the thermodynamic properties (including thermal conductivity, viscosity, specific heat, and density) of graphene-dispersed nanofluids for temperatures ranging from room temperature to around 200°C. The studied nanofluids were made out of graphene and the ionic liquid 1 hexyl-3-methylimidazolium tetra fluoroborate ((HMIM)BF4). The work showed that the thermal conductivity of the nanofluid containing 0.06 wt.% graphene increased by 15.2–22.9% at the tested temperature range (25–200°C). Viscosity, specific heat, and density decreased,

**3. Effect of the alignment of the carbon nanostructures on the thermal**

nanoparticles and carbon nanomaterials on the thermal conductivity of nanofluids.

microscopy

microscopy

TEM, Optical microscopy

microscopy

**Concentration Characterization TC**

Many researchers have studied the effect of the alignment of iron oxide nanoparticles and carbon nanomaterials on the thermal, electrical, and mechanical properties of fluids and polymers. **Table 2** summarizes most work about the effect of the alignment of iron oxide

**improvement**

With MF

MF 0.62 kG applied for 30 min 123%

No MF

SEM 15% 52% Sodium

Ni-CNTs Water 0.01 wt.% NA NA 75% NA Stable

**Stabilizing agent Stability**

Stable

dodecylbenzenesulfonate

(NaDDBS)

10 % 50% NaDDBS Stable

NA 11% NA Stable

18% 81% NA Stable

NaDDBS Stable

respectively, by 4.6–13.1, 3, and 2.8%, compared to (HMIM)BF4.

**conductivity of nanofluids**

**fluid**

Water 0.01 wt.% CNTs, 0.02 wt.% Fe2O3

Water 0.02 wt.% Optical

Ni-CNTs PAO 0.05 wt.% TEM, Optical

SWNT, 0.017 wt. % Fe2O3

Fe2O3 Water 0.4 vol.% TEM, Optical

Water 0.017 wt.%

**Reference Material Base**

1848 Advances in Carbon Nanostructures

CNTs +Fe2O3

CNTs +Fe2O3

SWCNT, Fe2O3

Hong et al. [65]

Wensel et al. [17]

Wright et al. [66]

Horton et al. [67]

Hong et al. [68]

Younes et al. [69]

**Table 2.** Current research of the effect of the alignment of iron oxide nanoparticles and carbon nanomaterials on the thermal conductivity of nanofluids.

Hong et al. [65] studied the effect of the alignment on the thermal conductivity of nanofluids under an external magnetic field and obtained the thermal conductivity and microscopic images taken with and without magnetic field, shown in **Figures 6** and **7**.

**Figure 6.** Experimental setup for taking optical microscopy images using the optical microscope, Leica Z16 APO under application of a magnetic field. From Ref. [70].

**Figure 7.** Experimental setup for measuring TC under application of a magnetic field. From Ref. [70].

**Figure 8** shows that the nanotubes align very well in the magnetic field direction.

**Figure 8.** SEM picture of 0.01 wt.% carbon nanotube + 0.02 wt.% Fe2O3 with magnetic field. From Ref. [65].

The heat transfer of nanofluids containing carbon nanotubes and magnetic-field-sensitive nanoparticles of Fe2O3 was enhanced when a magnetic field was applied for some time. The enhancement was attributed to the formation of aligned chains of Fe2O3 under the applied magnetic field and to the fact that chains help to connect the nanotubes. Wensel and Wright [17, 66] found that the thermal conductivity of nanofluids containing 0.02 wt.% was enhanced by 10% under a magnetic field. That was attributed to the aggregation of metal oxide particles on the surface of nanotubes by electrostatic attraction, which forms the aggregation chain along the nanotubes. Horton et al. [67] investigated the thermal conductivity of nanofluids contain‐ ing magnetic-metal-coated carbon nanotubes (Ni-coated SWCNTs), which significantly enhanced by 60% under a magnetic field. **Figure 9** shows a color online microscope image of 0.05 wt.% Ni-coated SWCNT in DI water, before (A) and after the magnetic field (B) was applied (*H* = 0.62 kG).

**Figure 9.** Color online microscope image of 0.05 wt.% Ni-coated SWCNT in DI water. (A) Before magnetic field and (B) after magnetic field (*H* = 0.62 kG). From Ref. [66].

**Figure 7.** Experimental setup for measuring TC under application of a magnetic field. From Ref. [70].

18610 Advances in Carbon Nanostructures

**Figure 8** shows that the nanotubes align very well in the magnetic field direction.

**Figure 8.** SEM picture of 0.01 wt.% carbon nanotube + 0.02 wt.% Fe2O3 with magnetic field. From Ref. [65].

The heat transfer of nanofluids containing carbon nanotubes and magnetic-field-sensitive nanoparticles of Fe2O3 was enhanced when a magnetic field was applied for some time. The enhancement was attributed to the formation of aligned chains of Fe2O3 under the applied magnetic field and to the fact that chains help to connect the nanotubes. Wensel and Wright [17, 66] found that the thermal conductivity of nanofluids containing 0.02 wt.% was enhanced

**Figure 10** shows TEM images of carbon nanotubes. (A) As received, uncoated CNTs, the image shows that the wall of one CNT is clear and nothing adhere to it. (B) Ni-coated CNTs, the image shows that the wall of the CNTs has been modified with some Ni adhered to the surface of the CNTs and (C) shows that the wall of the Ni-coated CNTs after applying the magnetic field experiment. It can be seen that the density of the Ni coated has decreased due to the effect of sonication and the magnetic field experiment.

**Figure 10.** TEM images of carbon nanotubes. (A) As received, uncoated. (B) Ni-coated nanotube before the magnetic field experiment, and (C) after the experiment. From Ref [67].

Hong et al. [68] studied the thermal conductivity of nanofluids containing CNTs and magneticfield-sensitive nanoparticles of Fe2O3 with NaDDBS surfactant and found that an electrostatic attraction among the nanotubes, the surfactant, and the metal oxide caused aggregation. NaDDBS surfactant with a negative charge attracted Fe2O3, which had positive zeta potential charge. When a cationic surfactant, cetyltrimethyl ammonium bromide (CTAB), replaced the anionic surfactant NaDDBS, no alignment was found because both the surfactant and Fe2O3 have the same positive charge. **Figure 11** shows that the nanotubes align very well in the direction of the magnetic field, either in scale bar 100 μm (A) or 10 μm (B).

**Figure 11.** Electron SEM images of 0.017 wt.% SWCNT, 0.017 wt.% Fe2O3, and 0.17 wt.% NaDSSB with magnetic field. (A) Scale bar 100 m, and (B) scale bar 10 m. From Ref. [68].

Younes et al. [27, 70] studied the alignment of different functionalized SWCNTs using Fe2O3 nanoparticles under an external magnetic field. These authors found that, even in the absence of NaDDBS as surfactant, some functionalized SWNTs, i.e., SWNT-SO2OH and SWNT-COOH, could disperse well in water and showed a clear alignment under an external magnetic field. **Figure 12** shows microscopy image of 0.016 wt.% SWNT-COOH and 0.016 wt.% Fe2O3 in DI water. This microscope image of 0.016 wt.% SWNT-COOH and 0.016 wt.% γ-Fe2O3 in DI water was obtained using a high-speed microscope video system. Magnetic field (*H* = 0.14 kG) was applied with an internal reference of 30 μm. As shown in **Figure 12**, it is clearly seen that the randomly dispersed SWNT-COOH/γ-Fe2O3 forms large and long lines, indicating that SWNT-COOH/γ-Fe2O3 nanoparticle mixture aligns under the external magnetic field.

**Figure 12.** Microscopy image of 0.016 wt.% SWNT-COOH, 0.016 wt.% Fe2O3 in DI water. At 8 min, the internal refer‐ ence is 30 μm. From Ref. [70].

Younes et al. [30] studied the effect of the alignment of carbon nanofibers in water and epoxy under an external magnetic field on the thermal conductivity of nanofluids. They found that the magnetic field aligned the nanofibers, not only in water but also in epoxy, as clearly seen in **Figure 13**. Adding 0.02 wt.% of carbon nanofibers to water enhanced the thermal conduc‐ tivity by 10%. The addition of an external magnetic field enhanced the thermal conductivity around 20%.

attraction among the nanotubes, the surfactant, and the metal oxide caused aggregation. NaDDBS surfactant with a negative charge attracted Fe2O3, which had positive zeta potential charge. When a cationic surfactant, cetyltrimethyl ammonium bromide (CTAB), replaced the anionic surfactant NaDDBS, no alignment was found because both the surfactant and Fe2O3 have the same positive charge. **Figure 11** shows that the nanotubes align very well in the

**Figure 11.** Electron SEM images of 0.017 wt.% SWCNT, 0.017 wt.% Fe2O3, and 0.17 wt.% NaDSSB with magnetic field.

Younes et al. [27, 70] studied the alignment of different functionalized SWCNTs using Fe2O3 nanoparticles under an external magnetic field. These authors found that, even in the absence of NaDDBS as surfactant, some functionalized SWNTs, i.e., SWNT-SO2OH and SWNT-COOH, could disperse well in water and showed a clear alignment under an external magnetic field. **Figure 12** shows microscopy image of 0.016 wt.% SWNT-COOH and 0.016 wt.% Fe2O3 in DI water. This microscope image of 0.016 wt.% SWNT-COOH and 0.016 wt.% γ-Fe2O3 in DI water was obtained using a high-speed microscope video system. Magnetic field (*H* = 0.14 kG) was applied with an internal reference of 30 μm. As shown in **Figure 12**, it is clearly seen that the randomly dispersed SWNT-COOH/γ-Fe2O3 forms large and long lines, indicating that SWNT-

**Figure 12.** Microscopy image of 0.016 wt.% SWNT-COOH, 0.016 wt.% Fe2O3 in DI water. At 8 min, the internal refer‐

COOH/γ-Fe2O3 nanoparticle mixture aligns under the external magnetic field.

direction of the magnetic field, either in scale bar 100 μm (A) or 10 μm (B).

(A) Scale bar 100 m, and (B) scale bar 10 m. From Ref. [68].

18812 Advances in Carbon Nanostructures

ence is 30 μm. From Ref. [70].

**Figure 13.** Alignment of CNFs in epoxy. (A) Optical microscopy images for 0.002 wt.% CNFs, 0.002 wt.% Fe2O3, and 0.02 wt.% NaDDBS in epoxy after 48 h. (B) Digital camera image of 0.02 wt.% CNFs, 0.02 wt.% Fe2O3, and 0.2 wt.% NaDDBS in epoxy after 2 h. Magnetic field (*H* = 1.2 kG) was applied. From Ref. [30].

Christensen et al. [26] studied the effect of nonionic surfactant on the dispersion and alignment of carbon nanotubes in the presence of Fe2O3. Gum arabic and Triton X-100 were used as nonionic surfactants. Gum arabic did not allow good dispersion or alignment in aqueous fluid. Triton X-100 allowed dispersion and alignment of both Fe2O3 and SWNT in an aqueous base

**Figure 14.** (a) Thermal conductivity ratio *k*/*kf* and the corresponding percentage of enhancement as a function of ap‐ plied magnetic field for 6.3 vol.% of Fe3O4. (B)–(D) Micrographs of ferrofluid emulsion of 200 nm, (B) without, (C) with low, and (D) with high magnetic field. (E)–(G) Schematic alignment of particles under the above three cases. From Ref. [71].

fluid. Due to the partial negative charge of Triton X-100, alignment did occur, but the alignment was slower, when compared to surfactants or functional groups containing stronger charges.

Glover et al. [72] prepared a nanofluids that consisted of water and 0–0.5 wt.% of sulfonated carbon nanotubes. The electrical conductivity of the prepared nanofluids was measured. The study found that the electrical conductivity increased by 13 times when only 0.5 wt.% of functional sulfonated carbon nanotubes was used. Philip et al. [71] observed a dramatic enhancement of thermal conductivity in a nanofluid containing magnetic particles of Fe3O4 (of 6.7 nm diameter) under a magnetic field (see **Figure 14**). The maximum enhancement in thermal conductivity was 300% (*k*/*kf* = 4.0) for a particle loading of 6.3 vol.%.

## **4. Conclusions**

The main focus of this chapter is to summarize the recent studies using CNTs and graphene to improve the thermal conductivity of nanofluids. Moreover, this chapter summarizes the investigation about the effect of using magnetic fields on enhancing the thermal conductivity of nanofluids. Too much discrepancy about thermal conductivity of nanofluids can be found in the literature. For carbon nanofluids, unfortunately, no significant improvements on thermal conductivity are observed using low concentrations. The thermal conductivity increased by approximately 10–20% when 1 vol.% SWCNT was used. However, at these high concentra‐ tions, the viscosity is greatly increased and the nanofluid becomes "mud-like," which are the difficulties reported in the literature. This variation in the thermal conductivity can be attributed to many factors, such as particle size, temperature, pH, or zeta potential. Finally, we believe that more efforts need to be done in order to improve the thermal conductivity of nanofluids and study the effect of the different parameters in thermal conductivity. In addition, more investigation about the use of magnetic field to align magnetic sensitive nanoparticles is required.

## **Author details**

Hammad Younes1\*, Amal Al Ghaferi1 , Irfan Saadat2 and Haiping Hong3

\*Address all correspondence to: hyounes@masdar.ac.ae

1 Department of Mechanical and Material Engineering, Masdar Institute of Science and Technology, Abu Dhabi, United Arab Emirates

2 Department of Electrical Engineering and Computer Science, Masdar Institute of Science and Technology, Abu Dhabi, United Arab Emirates

3 Department of Material and Metallurgical Engineering, South Dakota School of Mines and Technology, Rapid City, SD, USA

## **References**

fluid. Due to the partial negative charge of Triton X-100, alignment did occur, but the alignment was slower, when compared to surfactants or functional groups containing stronger charges. Glover et al. [72] prepared a nanofluids that consisted of water and 0–0.5 wt.% of sulfonated carbon nanotubes. The electrical conductivity of the prepared nanofluids was measured. The study found that the electrical conductivity increased by 13 times when only 0.5 wt.% of functional sulfonated carbon nanotubes was used. Philip et al. [71] observed a dramatic enhancement of thermal conductivity in a nanofluid containing magnetic particles of Fe3O4 (of 6.7 nm diameter) under a magnetic field (see **Figure 14**). The maximum enhancement in

The main focus of this chapter is to summarize the recent studies using CNTs and graphene to improve the thermal conductivity of nanofluids. Moreover, this chapter summarizes the investigation about the effect of using magnetic fields on enhancing the thermal conductivity of nanofluids. Too much discrepancy about thermal conductivity of nanofluids can be found in the literature. For carbon nanofluids, unfortunately, no significant improvements on thermal conductivity are observed using low concentrations. The thermal conductivity increased by approximately 10–20% when 1 vol.% SWCNT was used. However, at these high concentra‐ tions, the viscosity is greatly increased and the nanofluid becomes "mud-like," which are the difficulties reported in the literature. This variation in the thermal conductivity can be attributed to many factors, such as particle size, temperature, pH, or zeta potential. Finally, we believe that more efforts need to be done in order to improve the thermal conductivity of nanofluids and study the effect of the different parameters in thermal conductivity. In addition, more investigation about the use of magnetic field to align magnetic sensitive nanoparticles is

, Irfan Saadat2

1 Department of Mechanical and Material Engineering, Masdar Institute of Science and

2 Department of Electrical Engineering and Computer Science, Masdar Institute of Science

3 Department of Material and Metallurgical Engineering, South Dakota School of Mines and

and Haiping Hong3

= 4.0) for a particle loading of 6.3 vol.%.

thermal conductivity was 300% (*k*/*kf*

**4. Conclusions**

19014 Advances in Carbon Nanostructures

required.

**Author details**

Hammad Younes1\*, Amal Al Ghaferi1

Technology, Rapid City, SD, USA

\*Address all correspondence to: hyounes@masdar.ac.ae

Technology, Abu Dhabi, United Arab Emirates

and Technology, Abu Dhabi, United Arab Emirates


[27] H. Younes, Carbon Nanomaterials and Metal Oxide Nanoparticles and Their Applica‐ tions Toward Grease and Nanofluids, South Dakota School of Mines and Technology, Rapid City2013.

[13] Keblinski P, Eastman JA, Cahill DG. Nanofluids for thermal transport. Materials Today.

[14] Wang X, Xu X, Choi SU. Thermal conductivity of nanoparticle-fluid mixture. Journal

[15] Wen D, Ding Y. Experimental investigation into the pool boiling heat transfer of aqueous based γ-alumina nanofluids. Journal of Nanoparticle Research. 2005;7(2–3):

[16] Anoop K, Sundararajan T, Das SK. Effect of particle size on the convective heat transfer in nanofluid in the developing region. International Journal of Heat and Mass Transfer.

[17] Wensel J, Wright B, Thomas D, Douglas W, Mannhalter B, Cross W, et al. Enhanced thermal conductivity by aggregation in heat transfer nanofluids containing metal oxide nanoparticles and carbon nanotubes. Applied Physics Letters. 2008;92(2):023110.

[19] Biercuk M, Llaguno MC, Radosavljevic M, Hyun J, Johnson AT, Fischer JE. Carbon nanotube composites for thermal management. Applied Physics Letters. 2002;80(15):

[20] Berber S, Kwon Y-K, Tománek D. Unusually high thermal conductivity of carbon

[21] Nan C-W, Liu G, Lin Y, Li M. Interface effect on thermal conductivity of carbon

[22] Spitalsky Z, Tasis D, Papagelis K, Galiotis C. Carbon nanotube–polymer composites: chemistry, processing, mechanical and electrical properties. Progress in Polymer

[23] Hongtao L, Hongmin J, Haiping H, Younes H. Tribological properties of carbon nanotube grease. Industrial Lubrication and Tribology. 2014;66(5):579–583.

[24] Abueidda DW, Dalaq AS, Al-Rub RKA, Younes HA. Finite element predictions of effective multifunctional properties of interpenetrating phase composites with novel triply periodic solid shell architectured reinforcements. International Journal of

[25] Abueidda DW, Al-Rub RKA, Dalaq AS, Younes HA, Al Ghaferi AA, Shah TK. Electrical conductivity of 3D periodic architectured interpenetrating phase composites with carbon nanostructured-epoxy reinforcements. Composites Science and Technology.

[26] Christensen G, Younes H, Hong H, Smith P. Effects of solvent hydrogen bonding, viscosity, and polarity on the dispersion and alignment of nanofluids containing

Fe2O3 nanoparticles. Journal of Applied Physics. 2015;118(21):214302.

nanotube composites. Applied Physics Letters. 2004;85(16):3549–3551.

[18] Iijima S. Helical microtubules of graphitic carbon. Nature. 1991;354(6348):56–58.

of Thermophysics and Heat Transfer. 1999;13(4):474–480.

nanotubes. Physical Review Letters. 2000;84(20):4613.

2005;8(6):36–44.

19216 Advances in Carbon Nanostructures

265–274.

2767–2769.

2009;52(9):2189–2195.

Science. 2010;35(3):357–401.

Mechanical Sciences. 2015;92:80–89.

2015;118:127–134.


[52] Xie H, Lee H, Youn W, Choi M. Nanofluids containing multiwalled carbon nanotubes and their enhanced thermal conductivities. Journal of Applied Physics. 2003;94(8):4967.

[40] Shen J, Huang W, Wu L, Hu Y, Ye M. The reinforcement role of different aminofunctionalized multi-walled carbon nanotubes in epoxy nanocomposites. Composites

[41] Jeong G-H. Surface functionalization of single-walled carbon nanotubes using metal nanoparticles. Transactions of Nonferrous Metals Society of China. 2009;19(4):1009–

[42] Zhang L, Kiny VU, Peng H, Zhu J, Lobo RF, Margrave JL, et al. Sidewall functionali‐ zation of single-walled carbon nanotubes with hydroxyl group-terminated moieties.

[43] Phuoc TX, Massoudi M, Chen R-H. Viscosity and thermal conductivity of nanofluids containing multi-walled carbon nanotubes stabilized by chitosan. International Journal

[44] Garg P, Alvarado JL, Marsh C, Carlson TA, Kessler DA, Annamalai K. An experimental study on the effect of ultrasonication on viscosity and heat transfer performance of multi-wall carbon nanotube-based aqueous nanofluids. International Journal of Heat

[45] Meng Z, Wu D, Wang L, Zhu H, Li Q. Carbon nanotube glycol nanofluids: photothermal properties, thermal conductivities and rheological behavior. Particuology.

[46] Chen L, Xie H, Li Y, Yu W. Nanofluids containing carbon nanotubes treated by

[47] Chen L, Xie H. Surfactant-free nanofluids containing double- and single-walled carbon nanotubes functionalized by a wet-mechanochemical reaction. Thermochimica Acta.

[48] Assael M, Metaxa I, Arvanitidis J, Christofilos D, Lioutas C. Thermal conductivity enhancement in aqueous suspensions of carbon multi-walled and double-walled nanotubes in the presence of two different dispersants. International Journal of

[49] Ding Y, Alias H, Wen D, Williams RA. Heat transfer of aqueous suspensions of carbon nanotubes (CNT nanofluids). International Journal of Heat and Mass Transfer.

[50] Shaikh S, Lafdi K, Ponnappan R. Thermal conductivity improvement in carbon nanoparticle doped PAO oil: An experimental study. Journal of Applied Physics.

[51] Theres Baby T, Sundara R. Synthesis of silver nanoparticle decorated multiwalled carbon nanotubes-graphene mixture and its heat transfer studies in nanofluid. AIP

mechanochemical reaction. Thermochimica Acta. 2008;477(1–2):21–24.

Science and Technology. 2007;67(15):3041–3050.

Chemistry of Materials. 2004;16(11):2055–2061.

and Mass Transfer. 2009;52(21–22):5090–5101.

of Thermal Sciences. 2011;50(1):12–18.

2012;10(5):614–618.

2010;497(1–2):67–71.

2006;49(1–2):240–250.

2007;101(6):064302.

Advances. 2013;3(1):012111.

Thermophysics. 2005;26(3):647–664.

1012.

19418 Advances in Carbon Nanostructures


## **Carbon Nanostructure‐Based Scale Sensors Using Inkjet Printing and Casting Techniques**

Hammad Younes, Amal Al Ghaferi and Irfan Saadat

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64196

#### **Abstract**

[65] Hong H, Wright B, Wensel J, Jin S, Ye XR, Roy W. Enhanced thermal conductivity by the magnetic field in heat transfer nanofluids containing carbon nanotube. Synthetic

[66] Wright B, Thomas D, Hong H, Groven L, Puszynski J, Duke E, et al. Magnetic field enhanced thermal conductivity in heat transfer nanofluids containing Ni coated single

[67] Horton M, Hong H, Li C, Shi B, Peterson G, Jin S. Magnetic alignment of Ni-coated single wall carbon nanotubes in heat transfer nanofluids. Journal of Applied Physics.

[68] Hong H, Luan X, Horton M, Li C, Peterson G. Alignment of carbon nanotubes com‐ prising magnetically sensitive metal oxides in heat transfer nanofluids. Thermochimica

[69] Younes H, Christensen G, Luan X, Hong H, Smith P. Effects of alignment, pH, surfac‐ tant, and solvent on heat transfer nanofluids containing Fe2O3 and CuO nanoparticles.

[70] Younes H, Christensen G, Hong H, Peterson G. Alignment of different functionalized single wall carbon nanotubes using Fe2O3 nanoparticles under external magnetic field.

[71] Philip J, Shima P, Raj B. Enhancement of thermal conductivity in magnetite based nanofluid due to chainlike structures. Applied Physics Letters. 2007;91(20):203108. [72] Glover B, Whites KW, Hong H, Mukherjee A, Billups WE. Effective electrical conduc‐ tivity of functional single-wall carbon nanotubes in aqueous fluids. Synthetic Metals.

wall carbon nanotubes. Applied Physics Letters. 2007;91(17):173116.

Metals. 2007;157(10):437–440.

2010;107(10):104320.

19620 Advances in Carbon Nanostructures

Acta. 2011;525(1):87–92.

2008;158(12):506–508.

Journal of Applied Physics. 2012;111(6):064308.

Journal of Nanofluids. 2013;2(1):4–10.

In this chapter, the fabrication and characterization of scale sensor using carbon nanotubes (CNTs) are discussed. Two different methods are used to prepare the carbon nanomaterials for the sensor fabrication: CNT casting and the CNT inkjet printing. In addition, the sensors are integrated into Kelvin architectures. The electrical resistance of the carbon nanomaterial films is measured with and without adding a drop of brine to the surface of the film. The films are characterized by scanning electron microscopy (SEM), atomic force microscopy (AFM), transmission electron microscopy (TEM), and energy dispersive X‐ray spectroscopy (EDS). Electrical resistance of the casted CNT films and five layers of CNT inkjet printing are found to be close to 40.0 kΩ and 1.00  kΩ, respectively. Adding one drop of brine solution on the surface of the casted CNT film and five layers of CNT inkjet printing changed the resistance by 50% and 75%, respectively. The resetting process is done for all sensors by soaking in deionized water (DI water) for some time, and the electrical resistance is measured and found to be close to the initial electrical resistance.

**Keywords:** CNTs, Sensors, inkjet printing, smear casting, scale and electrical resist‐ ance

## **1. Introduction**

Carbon nanotubes (CNTs) are an important new class of nano‐2‐D and nano‐1‐D materials that have numerous novel and useful properties. Their nanostructures exhibit many interesting and often unique properties, and hence, they can be used in many novel technological and indus‐ trial applications. Due to their unique electrical and electronic properties, CNTs are consid‐ ered for various next‐generation device applications, that is nanofluids [1–3], scanning probes

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

[4], nanocomposites [5–9], grease [10, 11], electron magnetic shielding interface [12–16], solar thermal absorbers [17] and sensors [18–22]. The electrical resistance of the carbon nanomateri‐ als changes when chemicals in the surrounding make covalent or non‐covalentinteraction with the carbon nanomaterials. This change makes carbon nanomaterials good candidates for chemical sensors.The absorbedmolecules act asdopants that cause shiftingontheFermi energy of the carbon nanomaterials. In addition, the bonds formed between absorbed chemicals and the carbon nanomaterials change the band structure of the tube [23].

Scale is the creation of inorganic soluble salts from aqueous brines during oil and gas production, which deposits at the internal surface of the pipelines under the supersaturating conditions [24]. Calcium carbonate (CaCO3) and barium sulfate (BaSO4) are the most insolu‐ ble types of inorganic scale formed during the extraction of oil. The deposition can be seen not only in the pipelines but also in water‐handling equipment, pumps, valves, and any other parts that interact with water. The deposition of the scale leads to unavoidable damage of the equipment parts. Therefore, suspension of oil operations becomes essential in order to replace the damaged parts. In the petroleum industry, such interruptions lead to extremely high costs [25].

There has been multiple studies and research conducted on scale monitoring, testing, and optimization of inhibitors [26–28]. In general, these studies focused on the development of using new sensor materials that enabled monitoring and offline testing [29–32].

This chapter discusses the sensitivity and selectivity of using carbon nanotubes through two different preparation methods: the smear casted and the inkjet printing, to create sensors that detect scale in oil and gas pipelines. Moreover, the design and the microfabrication method of the sensors are discussed.

## **2. Experimental**

Single‐wall carbon nanotubes (SWCNT) were purchased from Sigma–Aldrich, Inc., USA. The CNT inkjet "CNTRENE® 3015 A3‐R" used for this printing was purchased from Brewer Science, Inc., USA. DMP2831 Inkjet Printer was used to print the CNT inkjet. The chemical surfactant sodium dodecyl benzene sulfonate (NaDDBS) was purchased from Sigma–Aldrich. Epoxy resins and hardener were purchased from West System Inc, Zayed port, Abu Dhabi. Releasing agents, Frekote, were purchased from Logistics Company Limited, in Dubai, UAE. Sonication was performed using a Branson Digital Sonifier, model 450. Scanning electron microscopy (SEM) images were acquired using the backscattered electron detector on a Zeiss Supra40VP variable pressure system. The electrical conductivity was obtained using Elite 300 Semi‐automatic Probe Station & Keithley 4200‐SCS Parameter Analyzer. The microstruc‐ tures of the samples were probed by scanning electron FEI quanta 250 ESEM and the FEI Nova NanoSEM 650. Atomic force microscopy data were obtained using (Asylum MFP‐3D, AFM) deflection 1.5 V, scan speed 0.7 Hz, applied voltage 500 V, contact mode imaging.

Casted technique: The epoxy used was a pre‐polymer consisting of bisphenol A attached to an epoxide group. Sodium dodecylbenzenesulfonate (NaDDBS) surfactant was added to

epoxy with a weight ratio of four times carbon nanotubes and mixed manually for 5 min. The mixture was then sonicated until the surfactant was dissolved completely, and a clear solution was obtained. About 0.5% of carbon nanotubes were added to the mixture and sonicated again. The mixture was then mixed manually for 5 min with hardener by weightratio of 10:2.63. Glass plate was coated with a layer of releasing agent (Frekote), and the mixture was casted on the glass plate by smear casting technique. The sample was cured at room temperature for 24 hours.

Carbon nanotube inkjet printing: CNT inkjet was used as it is, but the printing recipe was optimized in order to have a uniform continuous printing. CNT inkjet was printed on glass/ silicon oxide substrates.

## **3. Results and discussion**

#### **3.1. CNT casting**

[4], nanocomposites [5–9], grease [10, 11], electron magnetic shielding interface [12–16], solar thermal absorbers [17] and sensors [18–22]. The electrical resistance of the carbon nanomateri‐ als changes when chemicals in the surrounding make covalent or non‐covalentinteraction with the carbon nanomaterials. This change makes carbon nanomaterials good candidates for chemical sensors.The absorbedmolecules act asdopants that cause shiftingontheFermi energy of the carbon nanomaterials. In addition, the bonds formed between absorbed chemicals and

Scale is the creation of inorganic soluble salts from aqueous brines during oil and gas production, which deposits at the internal surface of the pipelines under the supersaturating conditions [24]. Calcium carbonate (CaCO3) and barium sulfate (BaSO4) are the most insolu‐ ble types of inorganic scale formed during the extraction of oil. The deposition can be seen not only in the pipelines but also in water‐handling equipment, pumps, valves, and any other parts that interact with water. The deposition of the scale leads to unavoidable damage of the equipment parts. Therefore, suspension of oil operations becomes essential in order to replace the damaged parts. In the petroleum industry, such interruptions lead to extremely high

There has been multiple studies and research conducted on scale monitoring, testing, and optimization of inhibitors [26–28]. In general, these studies focused on the development of

This chapter discusses the sensitivity and selectivity of using carbon nanotubes through two different preparation methods: the smear casted and the inkjet printing, to create sensors that detect scale in oil and gas pipelines. Moreover, the design and the microfabrication method of

Single‐wall carbon nanotubes (SWCNT) were purchased from Sigma–Aldrich, Inc., USA. The CNT inkjet "CNTRENE® 3015 A3‐R" used for this printing was purchased from Brewer Science, Inc., USA. DMP2831 Inkjet Printer was used to print the CNT inkjet. The chemical surfactant sodium dodecyl benzene sulfonate (NaDDBS) was purchased from Sigma–Aldrich. Epoxy resins and hardener were purchased from West System Inc, Zayed port, Abu Dhabi. Releasing agents, Frekote, were purchased from Logistics Company Limited, in Dubai, UAE. Sonication was performed using a Branson Digital Sonifier, model 450. Scanning electron microscopy (SEM) images were acquired using the backscattered electron detector on a Zeiss Supra40VP variable pressure system. The electrical conductivity was obtained using Elite 300 Semi‐automatic Probe Station & Keithley 4200‐SCS Parameter Analyzer. The microstruc‐ tures of the samples were probed by scanning electron FEI quanta 250 ESEM and the FEI Nova NanoSEM 650. Atomic force microscopy data were obtained using (Asylum MFP‐3D, AFM)

deflection 1.5 V, scan speed 0.7 Hz, applied voltage 500 V, contact mode imaging.

Casted technique: The epoxy used was a pre‐polymer consisting of bisphenol A attached to an epoxide group. Sodium dodecylbenzenesulfonate (NaDDBS) surfactant was added to

using new sensor materials that enabled monitoring and offline testing [29–32].

the carbon nanomaterials change the band structure of the tube [23].

costs [25].

the sensors are discussed.

198 200Advances in Carbon Nanostructures

**2. Experimental**

#### *3.1.1. Preparation of carbon nanotube‐based scale sensor*

As it has been discussed earlier, two types of preparation techniques for the CNTs have been used: smear casted and inkjet printing. The smear casting CNT nanocomposites were prepared bydispersing 0.5 wt% of CNTs in epoxy polymer using ultrahigh tip sonicator. Kelvin structure was then fabricated as can be seen in **Figure 1**.

#### *3.1.2. Characterization of carbon nanotube‐based scale sensor*

**Figure 2** is a cross‐sectional SEM image of the nanocomposites and shows that CNTs dis‐ perse very well in the epoxy polymer and make network structure essential for electrical conductivity. The electrical resistance of the prepared CNTs was measured just after the preparation and defined as resistance of original samples. The electricalresistance for samples were measured again after adding brine and defined as resistance with brine.

**Figure 2.** Images of cross section of the 1.0 wt% of CNTs in epoxy.

**Figure 3** is an SEM image of the surface of smear casted CNTs. There are many white spots on the surface of the sample, so we selected one of the largest (marked with a red cross) to carry on the EDS, as shown in **Figure 4**. As it can be seen, the sample has sodium and chlorine, which indicates the presence of sodium chloride on the surface.

**Figure 3.** SEM image of the surface of smear casted CNTs showing a spot of NaCl.

**Figure 4.** EDS spectrum for casted CNTs on a glass substrate.

Conductive atomic force microscopy (C‐AFM) was carried out for sample A, which is original, and for sample B that has brine. **Figure 5** shows that sample A has more tubes than B, which means that C‐AFM of casted CNTs affirm that the material A is more conductive than B, and this matches the resistance curve depicted in **Figure 6**.

**Figure 5.** CAFM for current mapping of rough casted CNTs, A: original sample B: with brine sample.

**Figure 6.** Electrical resistance of 0.5 wt% of smear casted CNTs.

**Figure 2.** Images of cross section of the 1.0 wt% of CNTs in epoxy.

200 202Advances in Carbon Nanostructures

indicates the presence of sodium chloride on the surface.

**Figure 3.** SEM image of the surface of smear casted CNTs showing a spot of NaCl.

**Figure 4.** EDS spectrum for casted CNTs on a glass substrate.

**Figure 3** is an SEM image of the surface of smear casted CNTs. There are many white spots on the surface of the sample, so we selected one of the largest (marked with a red cross) to carry on the EDS, as shown in **Figure 4**. As it can be seen, the sample has sodium and chlorine, which

#### *3.1.3. Electrical resistance of carbon nanotube‐based scale sensor*

**Figure 6** shows the resistance of the casted sample with sweeping voltage from 0 to 1.0. The resistance of the material seems to be very consistent with the swept voltage. Once one drop of brine solution was added to the surface of the smear casted CNTs, the resistance in‐ creased from 4.0 E+4 to almost 9.5 E+4 Ω. Chemical doping can induce strong changes in conductance. Such modifications can be easily detected by electron current signals, and these properties make CNTs extremely small sensors that are very sensitive to the surrounding chemicals.

### **3.2. CNT inkjet printing**

### *3.2.1. Preparation of carbon nanotube‐based scale sensor*

The CNT inkjet "CNTRENE® 3015 A3‐R" used for this printing was purchased from Brewer Science, Inc., USA. DMP2831 Inkjet Printer was used to print the CNT inkjet. About 1×1 cm2 of CNT inkjet printing was printed on Si substrate. Ebeam evaporator was used to deposit 200 nm Al as metal contact using a shadow mask. The printing was carried out at different temperatures and a second curing cycle was applied to the printed samples to ensure a very staple printing.

#### *3.2.2. Characterization of carbon nanotube‐based scale sensor*

After optimizing the printing recipe, it was critical to study the effect of the CNT printed layers on the electrical resistance; hereafter the sensor performance was assessed. About 3, 4, 5, and 6 layers of CNT inkjet were printed on Si substrates. The samples were characterized by measuring the electrical resistance of 4‐point Kelvin‐like structures and taking SEM and AFM images. **Figure 7** shows that 3, 4, 5, and 6 layers of CNT have the following electrical resist‐ ance, respectively: 5.5 kΩ, 4.2 kΩ, 1 kΩ, and 0.6 kΩ. This led us to conclude that while increasing number of printing passes, the electrical resistance decreases because the addition of CNT passes generates more paths for the electrons to go through and thus increases the conductivity.

**Figure 7.** Electrical resistance as a function of voltage of multilayer CNT inkjet printing.

**Figures 8** and **9** show SEM images for CNT inkjet printing of one and five layers of printing. The SEM images show that the CNTs are making network structures. SEM images of the five layers of CNT printing (**Figure 8**) shows more CNTs and thicker network structure than that of one layer of CNT printing (**Figure 7**), as expected.

**Figure 8.** SEM images for one layer of CNT inkjet printed on Si substrate.

conductance. Such modifications can be easily detected by electron current signals, and these properties make CNTs extremely small sensors that are very sensitive to the surrounding

The CNT inkjet "CNTRENE® 3015 A3‐R" used for this printing was purchased from Brewer Science, Inc., USA. DMP2831 Inkjet Printer was used to print the CNT inkjet. About 1×1 cm2 of CNT inkjet printing was printed on Si substrate. Ebeam evaporator was used to deposit 200 nm Al as metal contact using a shadow mask. The printing was carried out at different temperatures and a second curing cycle was applied to the printed samples to ensure a very

After optimizing the printing recipe, it was critical to study the effect of the CNT printed layers on the electrical resistance; hereafter the sensor performance was assessed. About 3, 4, 5, and 6 layers of CNT inkjet were printed on Si substrates. The samples were characterized by measuring the electrical resistance of 4‐point Kelvin‐like structures and taking SEM and AFM images. **Figure 7** shows that 3, 4, 5, and 6 layers of CNT have the following electrical resist‐ ance, respectively: 5.5 kΩ, 4.2 kΩ, 1 kΩ, and 0.6 kΩ. This led us to conclude that while increasing number of printing passes, the electrical resistance decreases because the addition of CNT passes generates more paths for the electrons to go through and thus increases the

chemicals.

staple printing.

conductivity.

**3.2. CNT inkjet printing**

202 204Advances in Carbon Nanostructures

*3.2.1. Preparation of carbon nanotube‐based scale sensor*

*3.2.2. Characterization of carbon nanotube‐based scale sensor*

**Figure 7.** Electrical resistance as a function of voltage of multilayer CNT inkjet printing.

**Figures 8** and **9** show SEM images for CNT inkjet printing of one and five layers of printing. The SEM images show that the CNTs are making network structures. SEM images of the five

**Figure 9.** SEM images for five layers of CNT inkjet printed on Si substrate.

The ultimate goal ofthis study was to fabricate a sensorthat can be used in oil and gas pipelines, which is a very hostile environment. Therefore, the sensor should be stable enough to sustain consistent results within the harsh conditions found in the pipelines. The stability of the CNT inkjet printed on silicon substrate was tested by immersing it in a solvent for some time and then obtaining SEM images, to examine the effect of the solvent on the printing.

Images obtained using tapping‐mode AFM (**Figures 10** and **11**) indicate that exposure to water for 72 hours does not have pronounced effects on the topography of the sample, in the case of inkjet‐printed CNT films. The roughness of the inkjet‐printed CNT decreased as a result of the exposure (**Figure 8**) by 26.3%. The roughness analyses were carried out on images obtained using tapping‐mode AFM topography scans.

**Figure 10.** AFM images of five layers of CNT inkjet printed on silicon oxide substrate.

**Figure 11.** AFM images of five layers of CNT inkjet printed on Si substrate after 72 hours of soaking in DI water.

#### *3.2.3. The resetting process of carbon nanotubes‐based scale sensor*

**Figure 12** shows the ability of the resetting film (made of CNT inkjet printing) to reset the resistance values. The resetting process is the ability of the film to get back to its primitive condition after soaking it in DI water. The electrical resistance of the CNT 5 passes printed film is 1.0E+03 Ω; by dipping the film for one hour in DI water, the electrical resistance de‐ creased to 5.0E+02 Ω. This treatment is needed to remove the left over ink from the surface as mentioned earlier. This value is the resistance of the optimized original sample and is thus used as the reference for the resetting process. Adding one drop of brine solution in‐ creases the electrical resistance up to 4.2E+03 Ω and decreases the sweeping voltage to 3.6E +03 Ω. By immersing the film for one hour in DI water, for the resetting process, we are able to obtain an electrical resistance very close to the electrical resistance of the original opti‐ mized sample.

**Figure 12.** Ability to reset the resistance values for five layers of CNTs printed on silicon oxide substrate.

## **4. Conclusion**

**Figure 10.** AFM images of five layers of CNT inkjet printed on silicon oxide substrate.

204 206Advances in Carbon Nanostructures

*3.2.3. The resetting process of carbon nanotubes‐based scale sensor*

**Figure 11.** AFM images of five layers of CNT inkjet printed on Si substrate after 72 hours of soaking in DI water.

**Figure 12** shows the ability of the resetting film (made of CNT inkjet printing) to reset the resistance values. The resetting process is the ability of the film to get back to its primitive condition after soaking it in DI water. The electrical resistance of the CNT 5 passes printed film is 1.0E+03 Ω; by dipping the film for one hour in DI water, the electrical resistance de‐ creased to 5.0E+02 Ω. This treatment is needed to remove the left over ink from the surface as mentioned earlier. This value is the resistance of the optimized original sample and is thus used as the reference for the resetting process. Adding one drop of brine solution in‐ creases the electrical resistance up to 4.2E+03 Ω and decreases the sweeping voltage to 3.6E +03 Ω. By immersing the film for one hour in DI water, for the resetting process, we are able

In summary, this chapter discussed the chemical sensing properties of CNTs obtained through smear casting and CNT inkjet printing techniques have been studied and compared. It was shown the ability of these films to be used for sensing salts, which are the precursors for the scale deposition. This was assessed through the modulation of the change in electrical conductance in the presence of various media. The SEM images showed that both CNT smear casted and the CNT inkjet printing were dispersed very well among the matrix and made a network structure. The results of these studies indicated that: (a) in case of casting, adding brine solution on the surface of the CNT casted films increased the resistance from 4.0E+3 kΩ to almost 9.5E+3 kΩ; (b) for the inkjet printing, five printing layers of CNT films was critical to obtain a consistent electricalresistance. Moreover, washing the original CNT inkjet printing sample with water was a precondition step to stabilize the CNT inkjet printing and ultimate‐ ly get consistent electrical resistance. The resistance of the CNT inkjet printing increased from 0.50 kΩ to 1.50 kΩ once the brine solution was added to the surface of the sample. The CNT inkjet printing sample was found to be stable even after more than 48 hours of soaking in water.

## **Acknowledgements**

The authors would like to thank the entire team of the Masdar Institute Cleanroom and microscopy suite for their contribution and support toward this work. In addition, we would like to acknowledge the help of Dr. Florent Favaux, and Khalid Marbou for SEM and AFM images. The authors would like also to appreciate Petroleum Institute for their partial funding ofthis research on behalf oftheAbuDhabi NationalOil Company (ADNOC), AbuDhabi, UAE.

## **Author details**

Hammad Younes1\*, Amal Al Ghaferi1 and Irfan Saadat2

\*Address all correspondence to: hyounes@masdar.ac.ae; hasy193@yahoo.com

1 Department of Mechanical and Material Engineering, Masdar Institute of Science and Technology, Abu Dhabi, United Arab Emirates

2 Department of Electrical Engineering and Computer Science, Masdar Institute of Science and Technology, Abu Dhabi, United Arab Emirates

## **References**


polymer/CNTs nanocomposites. In MRS Proceedings 1660, mrsf13–1660. Cambridge University Press.

[9] Younes H. Carbon Nanomaterials and Metal Oxide Nanoparticles and Their Applications Toward Grease and Nanofluids. PhD Dissertation. 2014.

**Author details**

206 208Advances in Carbon Nanostructures

**References**

2014;3(1):33–7.

2014;246:402–8.

Hammad Younes1\*, Amal Al Ghaferi1

Technology, Abu Dhabi, United Arab Emirates

and Technology, Abu Dhabi, United Arab Emirates

Journal of Nanofluids. 2013;2(1):25–8.

Technology. 2005;16(11):2138–46.

Mechanical Sciences. 2015;92:80–9.

and Irfan Saadat2

2 Department of Electrical Engineering and Computer Science, Masdar Institute of Science

[1] Younes H, Christensen G, Liu M, Hong H, Yang Q, Lin Z. Alignment of carbon nanofibers in water and epoxy by external magnetic field. Journal of Nanofluids.

[2] Younes H, Christensen G, Li D, Hong H, Ghaferi AA. Thermal conductivity of

[3] Christensen G, Younes H, Hong H, Peterson G. Alignment of carbon nanotubes comprising magnetically sensitive metal oxides by nonionic chemical surfactants.

[4] Nguyen CV, Ye Q, Meyyappan M. Carbon nanotube tips for scanning probe micro‐ scopy: fabrication and high aspect ratio nanometrology. Measurement Science and

[5] Abueidda DW, Dalaq AS, Al‐Rub RKA, Younes HA. Finite element predictions of effective multifunctional properties of interpenetrating phase composites with novel triply periodic solid shell architectured reinforcements. International Journal of

[6] Neupane S, Khatiwada S, Jaye C, Fischer DA, Younes H, Hong H, et al. Single‐walled carbon nanotubes coated by Fe2O3 nanoparticles with enhanced magnetic properties.

[7] Zhu Z, Hu Y, Jiang H, Li C. A three‐dimensional ordered mesoporous carbon/carbon nanotubes nanocomposites for supercapacitors. Journal of Power Sources.

[8] Kalakonda P, Georgiev GY, Cabrera Y, Judith R, Iannacchione GS, Cebe P, editors. Thermal and electrical transport properties of sheared and un‐sheared thin‐film

ECS Journal of Solid State Science and Technology. 2014;3(8):M39–M44.

nanofluids: review. Journal of Nanofluids. 2015;4(2):107–32.

1 Department of Mechanical and Material Engineering, Masdar Institute of Science and

\*Address all correspondence to: hyounes@masdar.ac.ae; hasy193@yahoo.com


## **Corrosion Resistance and Tribological Properties of Epoxy Coatings Reinforced with Well-Dispersed Graphene**

Liu Shuan

[22] Gautam M, Jayatissa AH. Ammonia gas sensing behavior of graphene surface decorated with gold nanoparticles. Solid‐State Electronics. 2012;78:159–65.

[23] Peng S, O'Keeffe J, Wei C, Cho K, Kong J, Chen R, et al., editors. Carbon nanotube chemical and mechanical sensors. In Proceedings of the 3rd International Workshop on Structural Health Monitoring 502–506. CRC Press, Stanford, CA, September 12–14,

[24] Crabtree M, Eslinger D, Fletcher P, Miller M, Johnson A, King G. Fighting scale—

[25] Sofela SO, Younes H, Jelbuldina M, Saadat I, Al Ghaferi A, editors. Carbon nanomate‐ rials based TSVs for dual sensing and vertical interconnect application. In Intercon‐ nect Technology Conference and 2015 IEEE Materials for Advanced Metallization

[26] Emmons, D.H., Graham, G.C., Holt, S.P., and Jordan, M.M. "On-site, near-real-time monitoring of scale deposition", SPE Annual Technical Conference and Exhibition,

[27] Jordan MM, editor. Deployment of real‐time scale deposition monitoring equipment to optimize chemical treatment for scale control during stimulation dlowback. In: SPE International Oilfield Scale Conference. Society of Petroleum Engineers, SPE Projects,

[28] Al‐Matar H, Al‐Ashhab JK, Mokhtar M, Ridzauddin S, editors. Techniques used to monitor and remove strontium sulfate scale in UZ producing wells. In: Abu Dhabi International Petroleum Exhibition and Conference. Society of Petroleum Engineers, Abu Dhabi International Petroleum Exhibition and Conference, 5-8 November, Abu

[29] Srivastava RK, Srivastava S, Narayanan TN, Mahlotra BD, Vajtai R, Ajayan PM, et al. Functionalized multilayered graphene platform for urea sensor. ACS Nano. 2011;6(1):

[30] Borini S, White R, Wei D, Astley M, Haque S, Spigone E, et al. Ultrafast graphene oxide

[31] Lu G, Ocola LE, Chen J. Gas detection using low‐temperature reduced graphene oxide

[32] Gonçalves V, Brandão L, Mendes A. Development of porous polymer pressure sensors

incorporating graphene platelets. Polymer Testing. 2014;37:129–37.

Conference (IITC/MAM), 2015 IEEE International 289–292. IEEE France.

removal and prevention. Oilfield Review. 1999;11(3):30–45.

Houston, Oct. 1999, SPE 56776.

Facilities & Construction, 6(1) 2008.

Dhabi, UAE, ISBN 978-1-55563-173-4.

humidity sensors. ACS Nano. 2013;7(12):11166–73.

sheets. Applied Physics Letters. 2009;94(8):083111.

2001.

208 210Advances in Carbon Nanostructures

168–75.

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64097

#### **Abstract**

Seawater environment is the most harsh corrosion media, thus the deterioration of metal materials in marine environment is particularly serious. Organic coatings are good options for metal protection. Regarding the large impact of metal corrosion on marine engineering and the important significance of metal anticorrosion, this chapter deals with the research and development of the protective mechanism, wear resistance, and antifriction properties of the graphene-based epoxy coating used in marine environ‐ ment. A highly efficient physical graphene dispersant was obtained by organic synthesis, and the graphene powder was dispersed uniformly in organic solvent and then added into epoxy resin. Moreover, the structure-activity relationship of the graphene-based epoxy coating was systematically investigated, taking into account its different amounts and dispersing characteristics in seawater. Consequently, the corrosion behavior and protection mechanism of the graphene-based heavy anticorro‐ sion coating are evaluated and clarified. The obtained results are of fundamental importance for the field.

**Keywords:** grapheme, corrosion resistance, tribology, epoxy coating, dispersion

## **1. Introduction**

The inhabitants of modernized nations live in metal-based societies. Almost all common metals or steels tend to react with their environments to some extents and at different corrosion rates. Corrosion is defined as the deterioration of a material, usually a metal, that requires the presence

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

of an anode, a cathode, an electrolyte, and an electrical circuit [1]. Corrosion usually occurs as a natural phenomenon.

The sea occupies about three-quarters of the Earth and has several kinds of chemical ions. Seawater is a powerful natural corrosion environmental, which contains almost all of the elements present on the Earth. Some of these chemicals or ions (e.g., Cl<sup>−</sup> , SO4 2−) are corrosive towards metals. Corrosion damages considerably the marine steel infrastructures, such as ships, offshore oil installations, submarine pipelines, and port terminals, which not only causes serious economic losses but also threats the safety of human life and ecology [2, 3]. It has been estimated that the metal corrosion losses in China totaled one trillion and eight hundred billion in 2014, which accounted for one-third of the average gross national product (GNP). Marine corrosion accounts for the majority of metal losses. Thus, an effective protection technology is necessary to inhibit corrosion. Up to now, several methods have been used to protect metals against corrosion such as cathodic protection, protective coating, corrosion inhibitors, and addition of alloy elements. Among those procedures, coating is the most effective and economical way to inhibit metal corrosion, as it can block the contact of the corrosive medi‐ um with the metal [4].

#### **1.1. Anticorrosion and failure mechanism of organic coating**

Organic coatings can be found everywhere in industrialized nations, in order to protect metal structures and metal equipment. They can protect metals from corrosion in two ways. First, the coating film acts as a physical barrier to separate the metal substrate from the corrosive environment. Second, the organic coating can contain corrosion inhibitors, which decrease the corrosion rate of the metal.

However, organic coatings are not perfect barriers as they can be penetrated by corrosive media, such as oxygen, water molecule, and ions. When this happens, metal corrosion can occur underneath the organic coating/metal interface. Thus, the corrosive media accumulat‐ edbeneath the organic coating not only leads to metal corrosion but alsodecreases the adhesion of the coating to the metal. Moreover, the corrosion products formed on the metal surface can also destroy the coating/metal system. Generally, there are three methods often used to improve the protective performance of the organic coatings. First, fillers or pigments, such as graphene, nanoclay, ZnO, and MoS2, are added into the coating to extend the diffusion path of corrosive media through the coating. Second, corrosion inhibitors are added to inhibit metal corrosion reaction underneath the coating. Third, high density, high cross-linked, thicker, and self-healing coating are needed to reduce the coating permeability [5].

#### **1.2. Graphene based epoxy coating used for metal protection**

Graphene (G) is a two-dimensional material with a single layer of around 0.335 nm and a diameter ranging from a few to several hundred microns. It has attracted increasing atten‐ tion from academic and industrial fields due to their unique nanostructure, excellent physi‐ cal properties, large specific surface area, super hydrophobic properties, and good compatibility with polymer matrixes [6, 7]. It has been reported that G can exhibit extraordi‐ nary corrosion resistance and self-lubricant characteristics due to its unique flexible graphit‐ ic layers, super-hydrophobic characteristics, extremely high strength, and easy shear capability on its densely packed and atomically smooth surface [8]. However, G nanosheets are prone to aggregate due to their strong van der Waals forces and high specific surface areas, which is difficult in achieving an homogenous dispersion of G in organic coatings. Many works on chemical modification approach have been reported to improve dispersion of G in compo‐ site coatings [9–11]. For example, Chang et al. [12] reported on the preparation of polyimide grafted G via a thermal imidization reaction. The well-dispersed G/polyimide coatings provided advanced corrosion protection of cold-rolled steel, as compared to neat polyimide coating. Compared to chemical modification, simply dispersing G into a polymer matrix can be particularly advantageous due to less damage made to G, and the high efficiency and ease of the preparation method.

In this chapter, the fabrication of composite coatings (with epoxy coating as the organic matrix and uniform dispersed G as anticorrosive barrier) is reported. The corrosion resistance and tribological properties of epoxy coatings reinforced with well-dispersed G are evaluated in details.

## **2. Corrosion resistance and tribological properties of G/epoxy composite coating**

#### **2.1. G powder dispersion performance in organic solvent**

of an anode, a cathode, an electrolyte, and an electrical circuit [1]. Corrosion usually occurs as

The sea occupies about three-quarters of the Earth and has several kinds of chemical ions. Seawater is a powerful natural corrosion environmental, which contains almost all of the

towards metals. Corrosion damages considerably the marine steel infrastructures, such as ships, offshore oil installations, submarine pipelines, and port terminals, which not only causes serious economic losses but also threats the safety of human life and ecology [2, 3]. It has been estimated that the metal corrosion losses in China totaled one trillion and eight hundred billion in 2014, which accounted for one-third of the average gross national product (GNP). Marine corrosion accounts for the majority of metal losses. Thus, an effective protection technology is necessary to inhibit corrosion. Up to now, several methods have been used to protect metals against corrosion such as cathodic protection, protective coating, corrosion inhibitors, and addition of alloy elements. Among those procedures, coating is the most effective and economical way to inhibit metal corrosion, as it can block the contact of the corrosive medi‐

Organic coatings can be found everywhere in industrialized nations, in order to protect metal structures and metal equipment. They can protect metals from corrosion in two ways. First, the coating film acts as a physical barrier to separate the metal substrate from the corrosive environment. Second, the organic coating can contain corrosion inhibitors, which decrease the

However, organic coatings are not perfect barriers as they can be penetrated by corrosive media, such as oxygen, water molecule, and ions. When this happens, metal corrosion can occur underneath the organic coating/metal interface. Thus, the corrosive media accumulat‐ edbeneath the organic coating not only leads to metal corrosion but alsodecreases the adhesion of the coating to the metal. Moreover, the corrosion products formed on the metal surface can also destroy the coating/metal system. Generally, there are three methods often used to improve the protective performance of the organic coatings. First, fillers or pigments, such as graphene, nanoclay, ZnO, and MoS2, are added into the coating to extend the diffusion path of corrosive media through the coating. Second, corrosion inhibitors are added to inhibit metal corrosion reaction underneath the coating. Third, high density, high cross-linked, thicker, and

Graphene (G) is a two-dimensional material with a single layer of around 0.335 nm and a diameter ranging from a few to several hundred microns. It has attracted increasing atten‐ tion from academic and industrial fields due to their unique nanostructure, excellent physi‐ cal properties, large specific surface area, super hydrophobic properties, and good compatibility with polymer matrixes [6, 7]. It has been reported that G can exhibit extraordi‐

, SO4

2−) are corrosive

elements present on the Earth. Some of these chemicals or ions (e.g., Cl<sup>−</sup>

**1.1. Anticorrosion and failure mechanism of organic coating**

self-healing coating are needed to reduce the coating permeability [5].

**1.2. Graphene based epoxy coating used for metal protection**

a natural phenomenon.

210 212Advances in Carbon Nanostructures

um with the metal [4].

corrosion rate of the metal.

G was supplied by Ningbo Morsh Technology Company from China. Four different physi‐ cal G dispersants, namely dispersant-1 (it has a basic triphenylmethane structure), disper‐ sant-2 (it is a sodium polyacrylate), dispersant-3 (it is a sodium hexametaphosphate), dispersant-4 (it has a carboxylated oligoaniline structure), were synthesized and used to disperse G. **Figure 1** depicts the G powder dispersed in ethanol solution by the different G dispersants after 30 days. Without any dispersant, the G powder was aggregated and quickly precipitated at the bottom of the container (**Figure 1a**). Dispersants-1 can be adsorbed on the surface of G, however, the Dispersants-1/G system precipitated after a few hours (**Figure 1b**); for Dispersants-2 and Dispersants-3, G powder was swelled and can be dispersed, but G precipitated after 12 h (**Figure 1c** and **d**). Dispersant-4 was able to stably disperse G in ethanol after even 30 days (**Figure 1e**), only a few G slurry can be found at bottom of glass bottle. Interestingly, when the dispersed G was dried by spraying, it still could be re-dispersed in several solvents, such as water, tetrahydrofuran (THF), and ethanol (**Figure 2**). **Figure 3** shows the SEM images of dispersed G in ethanol solution. It can be seen that the layered G was completely dispersed at length of ~5 μm, and no obvious aggregates existed. **Figure 4** depicts an atomic force microscopy (AFM) image of dispersed G deposited on a Si-coated substrate. On the topography image (**Figure 4a**), the G layer can be seen as slightly brighter than the Sicoated substrate. From the line scan shown in **Figure 4b**, it is very easy to estimate the thickness of G layer (about 3–6 nm), indicating that the aggregated G was exfoliated to ultra-thin layers.

**Figure 1.** G powder dispersed in ethanol solution using different G dispersants after 30 days; the mass ratio of disper‐ sant to G powder was 1:5. (a) G+ethanol; (b) G+dispersant-1+ethanol; (c) G+dispersant-2+ethanol; (d) G+disper‐ sant-3+ethanol; (e) G+dispersant-4+ethanol.

**Figure 2.** Schematic diagram of G dispersion and re-dispersion in different solvents.

**Figure 3.** SEM images of dispersed G at low magnification (a) and high magnification (b).

Corrosion Resistance and Tribological Properties of Epoxy Coatings Reinforced with Well-Dispersed Graphene http://dx.doi.org/10.5772/64097 213 215

**Figure 4.** AFM images (a) of a thin G film deposited on a Si-coated substrate and height profile (b) along the black line (from A to A0).

#### **2.2. Preparation of a G/epoxy composite coating**

**Figure 1.** G powder dispersed in ethanol solution using different G dispersants after 30 days; the mass ratio of disper‐ sant to G powder was 1:5. (a) G+ethanol; (b) G+dispersant-1+ethanol; (c) G+dispersant-2+ethanol; (d) G+disper‐

**Figure 2.** Schematic diagram of G dispersion and re-dispersion in different solvents.

**Figure 3.** SEM images of dispersed G at low magnification (a) and high magnification (b).

(a) (b)

sant-3+ethanol; (e) G+dispersant-4+ethanol.

212 214Advances in Carbon Nanostructures

**Figure 5** depicts the preparation process of G-based epoxy coatings. The well-dispersed G was mixed in the epoxy resin and cured by a polyamide agent (650 polyamide curing agent, light yellow, Amine value: 420 mgKOH/g, viscosity: 18,000 mPa s/40°C). **Figure 6a** and **b** shows typical SEM micrographs of fracture surfaces (cross section) of an epoxy coating containing 0.5 wt% G. It is well known that the fracture surface of pure epoxy thermoset exhibits oriented "bamboo"-like fracture patterns starting from the cracks. In contrast, this G based epoxy coating showed less brittle-like fracture morphology. The existence of highly dis‐ persed G effectively disturbs the crack propagation due to its two dimensional nanostruc‐ ture. **Figure 7** depicts TEM images of G (0.5 wt%) in epoxy coatings. It is shown that G is randomly dispersed in the coatings, and the good dispersion of G can extend the diffusion path of corrosive media in the epoxy coating.

**Figure 5.** Preparation process of G/epoxy composite coatings.

**Figure 6.** SEM images of cross sections of a G/epoxy composite coating (0.5 wt% G content).

**Figure 7.** TEM images of a G/epoxy composite coating (0.5 wt% G content).

## **2.3. Corrosion resistance and mechanism of G/epoxy composite coating**

In order to evaluate the improvement obtained with the G/epoxy composite coating, the pure epoxy coating, the D4/epoxy coating, and the G/epoxy coating were prepared in a similar way to the described above for the G-D4/epoxy coating (D4 means dispersant 4, it has a carboxy‐ lated oligoaniline structure). Electrochemical measurements were performed on a CHI-660E electrochemical workstation in 3.5% NaCl solution atroom temperature. A three-electrode setup was used for EIS test, a saturated calomel electrode (SCE) was used as the reference electrode, a platinum plate of 2.5 cm2 as the counter electrode, and coating/Q235 steels as the working electrode. Before electrochemical impedance spectroscopy (EIS) measurement, the coating/Q235 steel specimen was initially kept at an open circuit potential (OCP) for 1 h until a stable state was attained. For EIS, the test frequency range was 105 to 10−2 Hz and the amplitude of the sinusoidal voltage signal was 20 mV. ZsimpWin 3.21 software was used to fit the EIS spectra.

**Figure 8** shows the open circuit potential (*E*OCP) variation of pure epoxy, G/epoxy, D4/epoxy, and G-D4/epoxy as a function of immersion time in 3.5% NaCl solution. The *E*OCP of all coatings decreased to different degrees after the fluctuation at initial time and then remained basical‐ ly unchanged after 16 days. Compared with pure epoxy, D4/epoxy and G-D4/epoxy exhibit a larger *E*OCP, whereas G/epoxy displays a *E*OCP similar to that of the pure epoxy. The positive shift of *E*OCP for 0.30 V for G-D4/epoxy and 0.10 V for D4/epoxy after 35-day immersion shows that the G-D4 hybrid and D4 have obvious corrosion inhibition properties, and a similar phenomenon was reported by other researchers [13]. Furthermore, the G-D4 hybrid exhibits better corrosion inhibition than D4, while commercial G without D4 functionalization has almost no corrosion inhibition.

**Figure 6.** SEM images of cross sections of a G/epoxy composite coating (0.5 wt% G content).

**Figure 7.** TEM images of a G/epoxy composite coating (0.5 wt% G content).

electrode, a platinum plate of 2.5 cm2

fit the EIS spectra.

214 216Advances in Carbon Nanostructures

**2.3. Corrosion resistance and mechanism of G/epoxy composite coating**

In order to evaluate the improvement obtained with the G/epoxy composite coating, the pure epoxy coating, the D4/epoxy coating, and the G/epoxy coating were prepared in a similar way to the described above for the G-D4/epoxy coating (D4 means dispersant 4, it has a carboxy‐ lated oligoaniline structure). Electrochemical measurements were performed on a CHI-660E electrochemical workstation in 3.5% NaCl solution atroom temperature. A three-electrode setup was used for EIS test, a saturated calomel electrode (SCE) was used as the reference

working electrode. Before electrochemical impedance spectroscopy (EIS) measurement, the coating/Q235 steel specimen was initially kept at an open circuit potential (OCP) for 1 h until a stable state was attained. For EIS, the test frequency range was 105 to 10−2 Hz and the amplitude of the sinusoidal voltage signal was 20 mV. ZsimpWin 3.21 software was used to

**Figure 8** shows the open circuit potential (*E*OCP) variation of pure epoxy, G/epoxy, D4/epoxy, and G-D4/epoxy as a function of immersion time in 3.5% NaCl solution. The *E*OCP of all coatings

as the counter electrode, and coating/Q235 steels as the

**Figure 8.** Time dependence of the open circuit potential (EOCP) of pure epoxy (a), G/epoxy (b), D4/epoxy (c) and G-D4/ epoxy (d) coated Q235 steel electrodes immersed in 3.5% NaCl solution.

EIS was further employed to investigate the corrosion mechanism and evaluate the degrada‐ tion processes of the G-D4/epoxy composites, as shown in **Figure 9**. The Bode phase plots show one time constant at the initial stage of immersion (stage I), and then two time constants appeared due to the penetration of corrosive media during the immersion process (stage II). Usually, the time constant at the high frequency range can be attributed to the coating layer, while the one time constant at lower frequencies corresponds to a corrosion process taking place at the metal/coating interface. The pure epoxy, G/epoxy, and D4/epoxy coatings exhibit two time constants in short term immersion (2 days, 1 day, and 4 days, respectively), while the G-D4/epoxy coating shows one time constant in a high frequency range for 28-day immer‐ sion. The impedance modulus at low frequency (|Z|0.01Hz) can represent the corrosion protection of coating/metal system, which is in inverse proportion to corrosion rate. The |Z| 0.01Hz of the G-D4/epoxy-coated Q235 electrode was 1.6 × 109 Ω cm2 after 16-day immersion (**Figure 9D**), which was far larger than that of pure epoxy (2.2 × 107 Ω cm2 ), G/epoxy (3.1 × 106 Ω cm<sup>2</sup> ), and D4/epoxy (5.5 × 107 Ω cm<sup>2</sup> )-coated electrodes after 12-day immersion (**Figure 9A**–**C**). These results indicate that the G-D4/epoxy coating provides better corrosion protec‐ tion for the Q235 steel than other coatings. The increase in impedance for G-D4/epoxy coating can be attributed to the hydrophobicity and barrier effects of the well dispersed graphene.

**Figure 9.** Bode plots of pure epoxy (A), D4/epoxy (B), G/epoxy (C), and G-D4/epoxy (D) coated Q235 steel electrodes immersed in 3.5% NaCl solution after different times.

Water uptake of the coating was calculated based on the changes of the coating capacitance due to the variation of the dielectric constant. The water diffusion coefficient *D* could be calculated by means of the simplified Fick's second law of diffusion [14] (see Eq. (**1**)):

Corrosion Resistance and Tribological Properties of Epoxy Coatings Reinforced with Well-Dispersed Graphene http://dx.doi.org/10.5772/64097 217 219

$$\frac{\log Q\_{\circ} - \log Q\_{0}}{\log Q\_{\circ} - \log Q\_{0}} = \frac{2}{L} \sqrt{\frac{D}{\pi}} \sqrt{t} \tag{1}$$

where *Q*0, *Q*c, and *Q*∞ are the coating capacitances at the beginning of the immersion time *t*0, the time *t*c, and the time in saturated water absorption state *t*∞, respectively. *D* is the diffu‐ sion coefficient, *L* is the coating thickness. The coating capacitances obtained from the fitted corrosion parameters, were fitted with equivalent circuits (**Figure 10**) by ZsimpWin 3.21 software. *D* can be calculated from the lg*Q*<sup>c</sup> − *t* 1/2 curves (**Figure 11**), yielding the water diffusion coefficients of these 4 coating (pure epoxy, G/epoxy, D4/epoxy, and G-D4/epoxy), calculated to be 2.25 × 10−12, 23.6 × 10−12, 17.1 × 10−12 s and 0.619 × 10−12 cm2 /s, respectively (**Figure 12**). The significant decrease of water diffusion coefficient for G-D4/epoxy coating confirmed that the water penetration into the epoxy coating was impeded by the highly dispersed G, while the G/epoxy coatings without D4 exhibited much higher water diffusion coefficient due to the aggregation of G.

**Figure 10.** The equivalent circuit (a) for stage I and (b) for stage II used to fit the EIS data.

**Figure 11.** lgQc − t1/2 curves of different composite coatings immersed in 3.5% NaCl solution.

**Figure 9.** Bode plots of pure epoxy (A), D4/epoxy (B), G/epoxy (C), and G-D4/epoxy (D) coated Q235 steel electrodes

Water uptake of the coating was calculated based on the changes of the coating capacitance due to the variation of the dielectric constant. The water diffusion coefficient *D* could be

calculated by means of the simplified Fick's second law of diffusion [14] (see Eq. (**1**)):

immersed in 3.5% NaCl solution after different times.

216 218Advances in Carbon Nanostructures

**Figure 12.** The water diffusion coefficient D of different composite coatings immersed in 3.5% NaCl solution.

#### **2.4. Tribological properties of graphene/epoxy composite coating**

The coefficient of friction (COF) was conducted on UMT-3 tribometer (UMT-3, Bruker-CETR, USA) with a ball-on-plate configuration using the following conditions: load of 3N, sliding velocity of 1 Hz, friction duration of 20 min, and wear track length of 5 mm. All the experi‐ ments were performed under dry condition and seawater lubricated condition. **Figure 13** shows the friction coefficient of the G/epoxy coatings as function of test time under both of dry sliding and seawater lubrication, respectively. It was clear that the COF of G/epoxy coatings initially increases rapidly andthen reaches stability with time indry conditions, butin seawater environment, the COF of G/epoxy coatings initially decrease and then reach a stable state. The COF values of neat epoxy coating, 0.25%G-epoxy coating, and 0.5%G-epoxy coating were 0.49,

**Figure 13.** COF of three G-epoxy coating in dry and seawater friction conditions. (a) Neat epoxy coating; (b) 0.25%Gepoxy coating; (c) 0.5%G-epoxy coating.

0.45, and 0.42 in dry conditions, respectively. The COF values of neat epoxy coating, 0.25%Gepoxy coating, and 0.5%G-epoxy coating were 0.12, 0.08, and 0.06 in seawater, respectively. These results suggest that G can decrease the COF values of epoxy coatings in both dry and seawater conditions.

The wear traces of the different coatings were evaluated by using a surface profiler (Alpha-Step IQ, KLA-Tencor, USA). The wear resistance of the coatings is represented by wear trace cross-sectional area on account of the uniform wear trace length. **Figure 14** shows the wear trace areas of G/epoxy coatings with different mass ratio of G measured under dry sliding and seawater lubrication. It can be seen that the wear trace areas of composite coatings de‐ creased with the increase of G content in G/epoxy coatings. In addition, the COF and wear traces area, under seawaterlubrication, were much lowerthan those under dry sliding, during the whole process. Three main factors can cause these phenomena: First, the lubricating effect of seawater can effectively reduce the direct contact of the composite coatings with the counterpart surface. Second, the seawater medium can act as coolant which significantly reduces the friction heat, avoiding thermal softening, and resulting in lower COF and wear traces area. Finally, the wear debris can be continuously removed by seawater, leading to smooth and clean worn surface.

**Figure 14.** Cross-sectional area of wear trace of three G-epoxy coatings in dry and seawater friction conditions.

## **3. Conclusions**

**Figure 12.** The water diffusion coefficient D of different composite coatings immersed in 3.5% NaCl solution.

The coefficient of friction (COF) was conducted on UMT-3 tribometer (UMT-3, Bruker-CETR, USA) with a ball-on-plate configuration using the following conditions: load of 3N, sliding velocity of 1 Hz, friction duration of 20 min, and wear track length of 5 mm. All the experi‐ ments were performed under dry condition and seawater lubricated condition. **Figure 13** shows the friction coefficient of the G/epoxy coatings as function of test time under both of dry sliding and seawater lubrication, respectively. It was clear that the COF of G/epoxy coatings initially increases rapidly andthen reaches stability with time indry conditions, butin seawater environment, the COF of G/epoxy coatings initially decrease and then reach a stable state. The COF values of neat epoxy coating, 0.25%G-epoxy coating, and 0.5%G-epoxy coating were 0.49,

**Figure 13.** COF of three G-epoxy coating in dry and seawater friction conditions. (a) Neat epoxy coating; (b) 0.25%G-

epoxy coating; (c) 0.5%G-epoxy coating.

218 220Advances in Carbon Nanostructures

**2.4. Tribological properties of graphene/epoxy composite coating**

**1.** Graphene (G) can be uniformly dispersed in several solvents.


## **Acknowledgments**

The authors wish to acknowledge the financial support of the National Science Foundation of China (No. 41506098), China Postdoctoral Science Foundation (No. 2015M580528), Zhejiang Province Preferential Postdoctoral Science Foundation (No. BSW1502160) and Open Fund Project of Key Laboratory of Marine Materials and Related Technologies (LMMT-KFKT-2014-008) in Chinese Academy of Sciences.

## **Author details**

Liu Shuan

Address all correspondence to: liushuan@nimte.ac.cn

Ningbo Institute of Materials Technologies and Engineering, Chinese Academy of Sciences, Ningbo, China

## **References**


[5] B Chen, J Ma, L Gu, et al. Anticorrosive properties of oligoaniline containing photocured coatings. International Journal of Electrochemical Science, 10 (2015) 9154–9166.

**2.** The well-dispersed G can block the coating pores and decrease the electrolyte diffusion towards the epoxy coating, the water diffusion coefficients of epoxy coating decrease with

**4.** Friction coefficient and wear rate of the epoxy coating decrease with the addition of G; the friction coefficient and wear rate of G/epoxy composite coating in seawater was lower

The authors wish to acknowledge the financial support of the National Science Foundation of China (No. 41506098), China Postdoctoral Science Foundation (No. 2015M580528), Zhejiang Province Preferential Postdoctoral Science Foundation (No. BSW1502160) and Open Fund Project of Key Laboratory of Marine Materials and Related Technologies (LMMT-

Ningbo Institute of Materials Technologies and Engineering, Chinese Academy of Sciences,

[1] E McCafferty. Introduction to Corrosion Science. Springer, Alexandria Va, USA ,2009.

[2] S Liu, H Y Sun, N Zhang, et al. The corrosion performance of galvanized steel in closed

ior of the galvanized steel in simulated rust layer solution. Corrosion Science, 65 (2012)

[4] L Gu, S Liu, H C Zhao, et al. Anticorrosive oligoaniline-containing electroactive

concentration on corrosion behav‐

rusty seawater. International Journal of Corrosion, 2013 (2013) 1–9.

siliceous hybrid materials. RSC Advance, 5 (2015) 56011–56019.

[3] S Liu, H Y Sun, L J Sun, et al. Effects of pH and Cl**<sup>−</sup>**

addition of 0.5 wt% G.

220 222Advances in Carbon Nanostructures

than in dry conditions.

**Acknowledgments**

**Author details**

Liu Shuan

Ningbo, China

**References**

520–527

**3.** G increases the corrosion resistance of epoxy coatings.

KFKT-2014-008) in Chinese Academy of Sciences.

Address all correspondence to: liushuan@nimte.ac.cn


## **Fully Configurable Electromagnetic Wave Absorbers by Using Carbon Nanostructures**

Davide Micheli, Roberto Pastore, Antonio Vricella and Mario Marchetti

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/64213

#### **Abstract**

The configurable electromagnetic wave absorber (CEMA) defines a new method for the full design of layered carbon-based nanocomposites able to quasi-perfectly reproduce any kind of EM reflection coefficient (RC) profile. The method involves three main factors: (a) nanofillers-like carbon nanotube (CNT), carbon nanofiber (CNF), graphene nanoplatelet (GNP), and polyaniline (PANI) in different concentration versus the matrix; (b) the dielectric parameters of the nanoreinforced materials in the microwave range 2–18 GHz; (c) a numerical technique based on particle swarm optimization (PSO) algorithm within the MATLAB code of the EM propagation engine. Output is the layering of the wave absorber, that is, number of layers and material/thickness of each layer and the reflection/transmission simulated profiles. The frequency selective behavior is due to the multilayered composition, thanks to the direct/reflected wave combination tuning at interfaces. The dielectric characterization of the employed nanocomposites is presented in details: these materials constitute the database for the optimization code running toward the multilayer optimal solution. A FEM analysis is further proposed to highlight the EM propagation within the material's bulk at different frequencies. The mathematical model of layered materials, the PSO objective function used for RC target fitting, and some results are reported in the text.

**Keywords:** carbon nanoparticles, particle swarm optimization, layered electromagnet‐ ic wave absorber, nanocomposite materials

## **1. Introduction**

In the last 50 years, many research activities have been focused on materials and structures able to reduce the electromagnetic reflection coefficient (RC) in certain frequency ranges.

© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Applications are mostly in military technology, with radar absorbing materials (RAM) and radar absorbing structure (RAS) [1–6], and in electromagnetic compatibility (EMC) for EM shielding and EM interference (EMI) suppression purposes [7–9], as well as for antenna testing in anechoic chambers [10, 11]. Nowadays, the increase of wireless communication systems demands the use of specific technologies or technical solutions to reduce the mutual telecom‐ munication (TLC) systems interference. In fact, in order to reduce the energy impact and the constraints imposed by cost savings, most different wireless systems are often co-located in the same place, that is, sharing the same basic infrastructures. The proximity of such TLC systems can give rise to mutual radio frequency (RF) interferences due to intermodulation products in the shared antenna systems and/or to low efficiency in filtering out the spurious RF components. In this complex context, the availability of a new kind of materials, able to be frequency selective, absorbing, and tunable, is particularly attractive to reduce mutual EMI issues in specific frequency ranges [12]. Another field where configurable EM wave absorb‐ ers appear useful is the metrological science and technology. For example, in remote sens‐ ing, operations are sometimes necessary to often verify the sensors capability to detect and recognize certain material signatures on the Earth surface [13, 14]. Such type of testing and better sensors and algorithms calibration could be easily tuned in situ by using standard metrological samples, able to reproduce the desired EM target signatures. Further applica‐ tions of tunable EM wave absorbers are conceivable in the security field. Nowadays, some studies are focused on the detection of explosives using appropriate radar signatures [15, 16]: the possibility to design metrological samples able to reproduce some particular explosive matter radar signatures could represent an interesting aid for the safety upgrade in strategic/ critical environment.

This chapter is composed of three main sections. The first describes the nanocomposite materials database obtained from the dielectric characterization, by means of vector network analyzer (VNA) measurements, the mathematical modeling of the layered absorber, and the objective function optimization, used by the PSO algorithm. The second reports the main results of simulations, related to several ideal RC targets addressed; the third provides a discussion ofthe results basedupon a FEMnumerical analysis by COMSOLMultiphysics code.

## **2. Materials and methods**

#### **2.1. Nanocomposite materials**

Several microwave reflection profiles, even having complicated shapes, have been devised to design absorbers with fitting EM responses. The proposed strategy to achieve the target behavior is to combine different components from the microwave materials database in a multilayer assembly. Such a soft approach allows to exploit a variety of dielectric properties provided by the available constituents in the materials database, as well as the characteris‐ tics of the EM field propagation, through the physical discontinuities occurring at the several layers interfaces. Polymeric composites filled with carbon nanoparticles provide an ideal platform in this framework, thanks to the possibility to explore a wide range of EM trans‐

Applications are mostly in military technology, with radar absorbing materials (RAM) and radar absorbing structure (RAS) [1–6], and in electromagnetic compatibility (EMC) for EM shielding and EM interference (EMI) suppression purposes [7–9], as well as for antenna testing in anechoic chambers [10, 11]. Nowadays, the increase of wireless communication systems demands the use of specific technologies or technical solutions to reduce the mutual telecom‐ munication (TLC) systems interference. In fact, in order to reduce the energy impact and the constraints imposed by cost savings, most different wireless systems are often co-located in the same place, that is, sharing the same basic infrastructures. The proximity of such TLC systems can give rise to mutual radio frequency (RF) interferences due to intermodulation products in the shared antenna systems and/or to low efficiency in filtering out the spurious RF components. In this complex context, the availability of a new kind of materials, able to be frequency selective, absorbing, and tunable, is particularly attractive to reduce mutual EMI issues in specific frequency ranges [12]. Another field where configurable EM wave absorb‐ ers appear useful is the metrological science and technology. For example, in remote sens‐ ing, operations are sometimes necessary to often verify the sensors capability to detect and recognize certain material signatures on the Earth surface [13, 14]. Such type of testing and better sensors and algorithms calibration could be easily tuned in situ by using standard metrological samples, able to reproduce the desired EM target signatures. Further applica‐ tions of tunable EM wave absorbers are conceivable in the security field. Nowadays, some studies are focused on the detection of explosives using appropriate radar signatures [15, 16]: the possibility to design metrological samples able to reproduce some particular explosive matter radar signatures could represent an interesting aid for the safety upgrade in strategic/

This chapter is composed of three main sections. The first describes the nanocomposite materials database obtained from the dielectric characterization, by means of vector network analyzer (VNA) measurements, the mathematical modeling of the layered absorber, and the objective function optimization, used by the PSO algorithm. The second reports the main results of simulations, related to several ideal RC targets addressed; the third provides a discussion ofthe results basedupon a FEMnumerical analysis by COMSOLMultiphysics code.

Several microwave reflection profiles, even having complicated shapes, have been devised to design absorbers with fitting EM responses. The proposed strategy to achieve the target behavior is to combine different components from the microwave materials database in a multilayer assembly. Such a soft approach allows to exploit a variety of dielectric properties provided by the available constituents in the materials database, as well as the characteris‐ tics of the EM field propagation, through the physical discontinuities occurring at the several layers interfaces. Polymeric composites filled with carbon nanoparticles provide an ideal platform in this framework, thanks to the possibility to explore a wide range of EM trans‐

critical environment.

224 226Advances in Carbon Nanostructures

**2. Materials and methods**

**2.1. Nanocomposite materials**

**Figure 1.** Sample holder used to build ring-shaped samples of nanocomposite having dimensions of 3 mm and 7 mm for inner and outer diameter respectively.

port properties, from the insulating matrix to rather conductive composites, just by tuning the nano-filler concentration within the basic polymer. A materials database of 23 carbon micropowder- and nano-powder-filled polymeric composites has been set up and exploited. Such materials were obtained using commercial low cost products, in a well-established time-saving manufacturing procedure, in view of possible forthcoming scaling-up developments. In particular, a low-viscosity epoxy resin (Prime™ 20 LV—Gurit) was employed as polymeric matrix, while industrial grade multi-walled carbon nanotubes (MWCNT, NC7000—Nanocyl), carbon nanofibers (CNF—Sigma–Aldrich), graphene nanoplatelets (GNP, C750—XG Science), and polyaniline emeraldine base powder (PANI—Sigma–Aldrich) were added to the matrix as reinforcing materials, in different concentrations (from 0 up to 3 wt.%). The composite samples were shaped by means of an ensemble of hollow cylindrical molds (**Figure 1**) with 3 mm of internal diameter and 7 mm of the external one, as required by the specific techni‐ que adopted for the microwave characterization. The latter is based on the transmission line by means of coaxial air-line 50 Ω probe and has been carried out by means of a vector network analyzer (VNA, Agilent PNA-L N5235), via coaxial airline method in the frequency range of 2–18 GHz. The output of the measurements is obtained by means of the samples microwave reflection/transmission parameters and the materials complex electrical permittivity; the other typical dielectric properties of interest (loss tangent, skin depth, intrinsic impedance, etc.) can be then easily retrieved. The magnetic permeability of all the materials has not be taken into account during the analysis, since it is always very close to 1 (as expected for non-magnetic materials), as confirmed by the experiments. In **Figure 1**, the sample holder used to build ring samples of different reinforced materials is shown. Each hole of the sample holder is able to hostthe poured composite afterthe mixing and sonication procedures; afterthe curing process, the sample holderis open and the samples can be easily extracted. In **Figure 2**, the 50 Ω coaxial airline used for the VNA scattering parameters [17–19] measurements and for the subse‐ quent computation of the relative dielectric parameters is shown: the samples are inserted around the central conductor of the coaxial airline.

**Figure 2.** Coaxial airline 50 Ω probe used for the dielectric characterization of nanocomposite materials and for the measurement of EM reflection and transmission properties of layered materials.

**Figure 3.** Loss tangent and intrinsic wave impedance of the tested nanocomposite materials in the database averaged over the frequency band 2–18 GHz.

**Figure 3** shows the loss tangent and the intrinsic wave impedance of the nanocomposite materials present in the database, averaged in the frequency band 2–18 GHz. It can be seen that the CNT- and CNF-filled composites show higher EM losses and lower wave impedan‐ ces, thus anticipating the employment of such materials in layers where a significant EM absorption capability is required; on the other hand, the behavior of PANI/GNP-filled

materials is similar to that of the naked resin, thus suggesting the use of such materials in impedance matching layers. The behavior of the electric permittivity of CNT-based nanocom‐ posites is reported in **Figures 4** and **5** (real and imaginary parts, respectively). In the plots, a remarkable variation of permittivity can be observed upon variation of the CNT weight concentration, according to the change of resistive losses due to the conductive filler inclu‐ sion: such trend is more evident at lower frequencies, due to skin-depth effects. In **Table 1**, the database of materials is listed by the code number, thereafter used. In **Table 1**, the dielectric complex permittivity averaged overthe microwave range investigated is also reported to allow the readers a quick idea about the materials properties. A detailed description of the materi‐ als manufacturing procedure, as well as the full database microwave characterization plotting and analysis, is available in a previously published work [20].

**Figure 4.** Real part of permittivity of CNT-based nanocomposites.

**Figure 2.** Coaxial airline 50 Ω probe used for the dielectric characterization of nanocomposite materials and for the

**Figure 3.** Loss tangent and intrinsic wave impedance of the tested nanocomposite materials in the database averaged

**Figure 3** shows the loss tangent and the intrinsic wave impedance of the nanocomposite materials present in the database, averaged in the frequency band 2–18 GHz. It can be seen that the CNT- and CNF-filled composites show higher EM losses and lower wave impedan‐ ces, thus anticipating the employment of such materials in layers where a significant EM absorption capability is required; on the other hand, the behavior of PANI/GNP-filled

measurement of EM reflection and transmission properties of layered materials.

over the frequency band 2–18 GHz.

226 228Advances in Carbon Nanostructures

**Figure 5.** Imaginary part of permittivity of CNT-based nanocomposites.


**Table 1.** Microwave composite database: materials code, composition, and complex permittivity averaged in the frequency range 2–18 GHz.

#### **2.2. Mathematical model and objective function optimization**

The multilayered EM absorber design is based on the mathematical modeling of an EM wave propagating through a sequence of slabs made of different materials. Since the goal is to reproduce a given microwave reflection profile, it is worth to remind that the reflection coefficient (*RC*) at an air-matter interface can be evaluated by the relationship:

$$RC\_i = \left| \frac{Z\_i - Z\_0}{Z\_i + Z\_0} \right| \tag{1}$$

which relates the free-space impedance (*Z*<sup>0</sup> ≈ 377 Ω) to the input impedance (*Zi* ) seen at the impinged surface for a given *i*th value of frequency. Such quantity depends on the dielectric properties of the material; for a multilayer structure, it is affected by multiple reflections occurring at the material interfaces and can be obtained by iterating Eq. (1) layer by layer through the formalism:

**Code Material Average complex permittivity**

**Table 1.** Microwave composite database: materials code, composition, and complex permittivity averaged in the

The multilayered EM absorber design is based on the mathematical modeling of an EM wave propagating through a sequence of slabs made of different materials. Since the goal is to reproduce a given microwave reflection profile, it is worth to remind that the reflection

> 0 0


*i*

*i Z Z RC Z Z*

coefficient (*RC*) at an air-matter interface can be evaluated by the relationship:

*i*

**2.2. Mathematical model and objective function optimization**

 Resin PRIME TM 2.88–j0.19 CNT 0.5 wt% 7.36–j3.27 CNT 1.0 wt% 9.98–j6.27 CNT 1.5 wt% 12.76–j10.14 CNT 2.0 wt% 14.14–j11.21 CNT 2.5 wt% 15.87–j13.66 CNF 0.5 wt% 3.74–j0.38 CNF 1.0 wt% 4.39–j0.53 CNF 1.5 wt% 4.92–j0.69 CNF 2.0 wt% 5.97–j1.02 CNF 2.5 wt% 7.57–j1.74 CNF 3.0 wt% 9.74–j2.99 PANI 0.5 wt% 3.20–j0.21 PANI 1.0 wt% 3.16–j0.26 PANI 1.5 wt% 3.31–j0.28 PANI 2.0 wt% 3.09–j0.26 PANI 2.5 wt% 3.15–j0.28 GNP 0.5 wt% 2.90–j0.16 GNP 1.0 wt% 3.17–j0.21 GNP 1.5 wt% 3.20–j0.20 GNP 2.0 wt% 3.27–j0.20 GNP 2.5 wt% 3.36–j0.24 GNP 3.0 wt% 3.44–j0.27

228 230Advances in Carbon Nanostructures

frequency range 2–18 GHz.

$$Z\_{\rm in} = \eta\_{\rm in} \frac{Z\_{\rm in-1} \cos(k\_{\rm in} t\_{\rm u}) + j \eta\_{\rm in} \sin(k\_{\rm in} t\_{\rm u})}{\eta\_{\rm in} \cos(k\_{\rm in} t\_{\rm u}) + j Z\_{\rm in-1} \sin(k\_{\rm in} t\_{\rm u})} \tag{2}$$

$$
\eta\_u = \sqrt{\frac{\mu\_0}{\varepsilon\_0}} \sqrt{\frac{\mu\_{ru}}{\varepsilon\_{ru}}} \tag{3}
$$

$$k\_{\boldsymbol{\mu}}^2 = \left(\boldsymbol{\Sigma}\boldsymbol{\pi}f\right)^2 \mu\_0 \boldsymbol{\varepsilon}\_0 \mu\_{\boldsymbol{m}} \boldsymbol{\varepsilon}\_{\boldsymbol{m}} \tag{4}$$

where each layer (*u*) is denoted by its proper thickness (*tu*) and by the EM intrinsic wave impedance (*ηiu*) of the constituent material at the *i*th value of frequency (*f*). The intrinsic wave impedance and the wave number (*kiu*) are linked to the relative electric permittivity and magnetic permeability of the material constituting the *u*th layer (*εru* and *μru*, respectively) and to the vacuum electric permittivity and magnetic permeability (*ε*0 and *μ*0, respectively).

The adopted particle swarm optimization (PSO) algorithm evolves iteratively by minimiz‐ ing the following objective function (*OF*):

$$OF = \sum\_{\beta \text{square}} \frac{\left| S\_{RC} - T\_{RC} \right|^2}{\left| S\_{RC} + T\_{RC} \right|^2} \tag{5}$$

where *TRC* is the target RC, and *S*RC is the simulated RC of a given multilayer—that is, a "particle" filling the space of solutions—obtained by combining Eqs. (1) and (2). Input of the design/optimization code is the target RC profile to fit and the electric permittivity of the materials in the database (the magnetic permeability is here neglected since the measure‐ ments of the dielectric parameters have shown that *μr* is unitary for all the materials under test). The output is the composition of the multilayered structure, that is, number of layers, layering sequence, and material/thickness of each layer. The PSO ranges within the space of solutions in order to minimize the *OF* defined in Eq.(4): the constraints are mainly imposed by the freedom degrees in exploring the database of available materials, in orderto select the most appropriate for each layer, as well as the related thickness, limiting to a maximum of ten-layer structures. Further details on the EM wave/multilayer interaction modeling with a full description of the developed PSO algorithm can be found in some recent publications [21–23].

#### **3. Results**

The results of CEMA solutions for several RC targets are listed in **Tables 2**–**5**, where the target parameter, the materials of each layer coded with a number referred to **Table 1**, the thick‐ ness of each layer and the total thickness of the multilayer are reported. All the solutions are represented as ten-layer sequences: of course, a zero thick layer means that the PSO algo‐ rithm optimized the final layered material without the necessity of using that particular layer. In some cases, the indication of the PSO parameter *α* is also provided [21–23], whose mean‐ ing is here briefly recalled: by this parameter, the OF optimization is weighted with respect to the multilayer total thickness (e.g., *α* = 1 means that only OF will be taken into account by PSO, without taking care of the thickness minimization). In the next **Figures 6**–**15**, the plot of the RC simulations is reported, comparing to the corresponding targets (red dashed lines). Almost all the simulations are based on RC target imitation by providing a PEC at the end of the layered material, except two. In particular, CEMA 02 has been designed providing free-space environment on both sides, thatis, withoutthe presence of PEC as the back-end ofthe structure. For CEMA 24, two different design scenarios have been taken into account: the first one represents the usual RC target imitation with a PEC-backed multilayer, while the second one explores the possibility to consider the TC as target parameter to be assigned to the PSO algorithm, thus requiring free space at both sides (here the other PSO parameter *γ* is intro‐ duced, to point out the weight in RC/TC addressing: *γ* = 0 indicates the RC optimization with no regard to TC, vice versa for *γ* = 1, intermediate values represent an optimization process that takes into account for TC and RC imitation simultaneously—cfr. CEMA02 plot). In this special case, the expected EM losses phenomena within the material increase with the frequency, thus progressively lowering down the TC values: such unwanted collateral effect is unavoidable since the higher the frequency the greater the skin-depth effect, which in turns increases the EM wave losses within thick materials.

**Figure 6.** Plot of CEMA 00.

Fully Configurable Electromagnetic Wave Absorbers by Using Carbon Nanostructures http://dx.doi.org/10.5772/64213 231 233


**3. Results**

230 232Advances in Carbon Nanostructures

The results of CEMA solutions for several RC targets are listed in **Tables 2**–**5**, where the target parameter, the materials of each layer coded with a number referred to **Table 1**, the thick‐ ness of each layer and the total thickness of the multilayer are reported. All the solutions are represented as ten-layer sequences: of course, a zero thick layer means that the PSO algo‐ rithm optimized the final layered material without the necessity of using that particular layer. In some cases, the indication of the PSO parameter *α* is also provided [21–23], whose mean‐ ing is here briefly recalled: by this parameter, the OF optimization is weighted with respect to the multilayer total thickness (e.g., *α* = 1 means that only OF will be taken into account by PSO, without taking care of the thickness minimization). In the next **Figures 6**–**15**, the plot of the RC simulations is reported, comparing to the corresponding targets (red dashed lines). Almost all the simulations are based on RC target imitation by providing a PEC at the end of the layered material, except two. In particular, CEMA 02 has been designed providing free-space environment on both sides, thatis, withoutthe presence of PEC as the back-end ofthe structure. For CEMA 24, two different design scenarios have been taken into account: the first one represents the usual RC target imitation with a PEC-backed multilayer, while the second one explores the possibility to consider the TC as target parameter to be assigned to the PSO algorithm, thus requiring free space at both sides (here the other PSO parameter *γ* is intro‐ duced, to point out the weight in RC/TC addressing: *γ* = 0 indicates the RC optimization with no regard to TC, vice versa for *γ* = 1, intermediate values represent an optimization process that takes into account for TC and RC imitation simultaneously—cfr. CEMA02 plot). In this special case, the expected EM losses phenomena within the material increase with the frequency, thus progressively lowering down the TC values: such unwanted collateral effect is unavoidable since the higher the frequency the greater the skin-depth effect, which in turns

increases the EM wave losses within thick materials.

**Figure 6.** Plot of CEMA 00.


**Table 2.** CEMA solutions by PSO for different target: layers sequence (materials coded as in **Table 1**), single layers, and total multilayer thickness.



**Target CEMA Materials Layers**

TRC 13 CEMA 13 (*α* = 0.99) 11 0.0603 48.0260

**Table 2.** CEMA solutions by PSO for different target: layers sequence (materials coded as in **Table 1**), single layers,

**Target CEMA Materials Layers thickness (mm) Total thickness (mm)**

TRC 15 CEMA 15 (*α* = 0.99) 1 8.5696 44.8402 0 9.0000 4.9133 8.7613 0.6217 0.5130 9.0000 3.4488 0.0125 PEC − TRC 15 CEMA 15 (*α* = 0.96) 21 0 30.0380 8.6290 0

and total multilayer thickness.

234Advances in Carbon Nanostructures

**thickness (mm)**

 0.4403 PEC −

 2.4181 9.0000 4.4037 0 0.6423 9.0000 9.0000 9.0000 4.5016 PEC −

**Total thickness (mm)**

> **Table 3.** CEMA solutions by PSO for different target: layers sequence (materials coded as in **Table 1**), single layers, and total multilayer thickness.




**Target CEMA Materials Layers**

234 236Advances in Carbon Nanostructures

TRC 17 CEMA 17 (*α* = 0.99) 18 6.7023 52.3048

TRC 17 CEMA 17 (*α* = 0.96) 18 5.7320 37.2708

TRC 20 CEMA 20 (*α* = 0.96) 18 6.8693 33.4005

**thickness (mm)**

8 4.4908

12 9.0000 10 0.5107 10 9.0000 12 1.1707 11 3.4329 6 9.0000 14 8.9974

11 0

PEC −

5 0

5 0

10 0 7 0

PEC −

2 0

10 4.4797 3 6.7434 14 3.1642 14 9.0000

7 1.5600

10 2.9788 4 9.0000

14 9.0000 14 9.0000 **Total thickness (mm)**

> **Table 4.** CEMA solutions by PSO for different target: layers sequence (materials coded as in **Table 1**), single layers, and total multilayer thickness.



**Table 5.** CEMA solutions by PSO for different target: layers sequence (materials coded as in **Table 1**), single layers, and total multilayer thickness.

The PSO algorithm shows an intriguingly effective mimic capability in almost all the cases, even for the most demanding (actually "not physical") targets conceived. Of course, the reliability degree is strictly linked to the target shape, that is, to the peaks number and depth as well as to the sharpness of the RC variations to be hunted. TRC 00 and TRC 02 present a simple filter-shaped profile: the corresponding solutions are able to broadly follow the target trend, even if the sharp (ideal) square-like behavior, with deep peaks up to 20 dB, does not allow a satisfactory imitation around the edges. In particular, CEMA 02 almost loses track of the target at the frequency range boundaries, probably due to the TC weighting in this free-space backed situation. Other regular square-like target profiles (multi-filter behavior) with RC oscilla‐

tions of about 15 dB confirm both potentialities and issues of the method. In particular, the very difficult task of TRC 23 imitation is actually impossible to tackle, even if the quasi-total reproduction of all the oscillating peaks by CEMA 23 represents a surprising result in this case. TRC 16 is much easier to be imitated due to the greater peaks width, even if some inaccura‐ cies are discovered for both the CEMA 16 solutions at lowest frequencies. The results ob‐ tained by CEMA 24 can be seen as intermediate case of the latter two. An excellent mimic effectiveness is provided by CEMAs 15, 17, and 20, which are able to closely follow the corresponding targets: in such cases, the algorithm is "aided" by the lower peaks depth and mainly by smoother RC variations in frequency. The difference in pursuing triangle-shaped *versus* square-shaped RC oscillations is well established by the two CEMA 13 solutions, which alternatively match and miss the exotic "stair-profile" of TRC 13. Forthe TRC 13, 15, 16, 17, and 20, two CEMA solutions have been proposed by slight variation of *α*. It can be noticed that at lower *α* (0.96) the algorithm finds solutions with thinnertotal thickness: that is reflected by the simulations, which shows a little bit more inaccuracy (black dashed lines) around the sharpest RC variations. Concerning the transmission targets (*γ* ≠ 0), difficulty to conceive appreciable solutions in such framework due to the abovementioned intrinsic skin-depth effects, evident in both the situations investigated, needs to be pointed out. In TRC 02, the parameter *γ* is set to 0.5 (i.e., the optimization of RC and TC is equally weighted) and a constant low transmis‐ sion rate is defined: besides the loss effects, the simultaneous research of a layering to follow the RC target leads to an intermediate solution, not suitable for the required transmission (otherwise a quasi-full absorbing structure would have been identified). As for the TTC 24 (*γ* = 1), a challenging square-shaped transmission profile is defined (by analogy with TRC 24, *γ* = 0). The quite fast target oscillations are only remotely hinted by the CEMA 24 solution, whose average TC substantially decreases at higher frequency; nevertheless, it is worth noticing that the corresponding reflection presents a peaks series roughly reversed, in respect to the transmission target.

**Figure 7.** Plot of CEMA 02.

**Target CEMA Materials Layers**

TRC 24 CEMA 24 only TC (*α* = 0.99, *γ* = 0) 20 8.5429 25.5742

TTC 24 CEMA 24 only RC (*α* = 0.99, *γ* = 1) 20 9.0000 48.9902

**Table 5.** CEMA solutions by PSO for different target: layers sequence (materials coded as in **Table 1**), single layers,

The PSO algorithm shows an intriguingly effective mimic capability in almost all the cases, even for the most demanding (actually "not physical") targets conceived. Of course, the reliability degree is strictly linked to the target shape, that is, to the peaks number and depth as well as to the sharpness of the RC variations to be hunted. TRC 00 and TRC 02 present a simple filter-shaped profile: the corresponding solutions are able to broadly follow the target trend, even if the sharp (ideal) square-like behavior, with deep peaks up to 20 dB, does not allow a satisfactory imitation around the edges. In particular, CEMA 02 almost loses track of the target at the frequency range boundaries, probably due to the TC weighting in this free-space backed situation. Other regular square-like target profiles (multi-filter behavior) with RC oscilla‐

and total multilayer thickness.

236 238Advances in Carbon Nanostructures

**thickness (mm)**

13 7.9931 PEC −

22 9.0000 8 9.0000 9 0.1918 4 1.8001 12 9.0000 13 0 13 9.0000 13 1.9983 12 0 PEC −

**Total thickness (mm)**

**Figure 8.** Plot of CEMA 13.

**Figure 9.** Plot of CEMA 15.

A promising level of confidence is suggested from the reported results in each phase of the work, from the database materials preliminary characterization to the mathematical model‐ ing of the EM field/matter interaction, up to the PSO algorithm design optimization. The capability of the layering approach lies in the frequency selective interaction between the different materials and the EM field. Such concept is well explained by a finite element method (FEM) simulation where CEMA13 (*α* = 0.96) has been deeply analyzed by studying the multilayer structure inserted in a 50 Ω coaxial airline of the same type used to measure the dielectric properties of the database materials. The FEM analysis shows how the EM field propagates through the several layers at two different frequencies, 2.5 and 14 GHz, where the reflection coefficient, respectively, results maximum and minimum. Commercial code COMSOL Multiphysics has been used for the FEM analysis [20, 24]. In **Figure 16**, the mesh of the multilayer within the coaxial waveguide is depicted: the blue part on the left is the empty section of the airline, while the other is the part filled with ten layers of nanocomposite materials. The airline is simulated using a perfect electric conductor for inner, outer, and back

**Figure 10.** Plot of CEMA 16.

**Figure 8.** Plot of CEMA 13.

238 240Advances in Carbon Nanostructures

**Figure 9.** Plot of CEMA 15.

A promising level of confidence is suggested from the reported results in each phase of the work, from the database materials preliminary characterization to the mathematical model‐ ing of the EM field/matter interaction, up to the PSO algorithm design optimization. The capability of the layering approach lies in the frequency selective interaction between the different materials and the EM field. Such concept is well explained by a finite element method (FEM) simulation where CEMA13 (*α* = 0.96) has been deeply analyzed by studying the multilayer structure inserted in a 50 Ω coaxial airline of the same type used to measure the dielectric properties of the database materials. The FEM analysis shows how the EM field propagates through the several layers at two different frequencies, 2.5 and 14 GHz, where the reflection coefficient, respectively, results maximum and minimum. Commercial code COMSOL Multiphysics has been used for the FEM analysis [20, 24]. In **Figure 16**, the mesh of the multilayer within the coaxial waveguide is depicted: the blue part on the left is the empty section of the airline, while the other is the part filled with ten layers of nanocomposite materials. The airline is simulated using a perfect electric conductor for inner, outer, and back

surfaces. The Port1 is placed at the beginning of the coaxial airline on the left side of the figure. The maximum element size of the mesh is 1/10 of the minimum EM wavelength (*λ* = 3 × 108 / *f*max ~ 1.6 cm), while the minimum element size of the mesh is in the order of 10−7 m: these dimensions enable an optimal growth of the mesh and guarantee the needed accuracy in the EManalysis. In the coaxial airline, Port 1 isusedto compute the scattering parameter S11, related to the reflection coefficient. In the numerical simulation, the power set at the port is 3 dBm (2 mW) as in the real measurement. All the parameters adopted to run the FEM simulation are chosen in such a way to reproduce the experimental conditions of the microwave characteri‐ zation. In particular, the real and imaginary part of the complex electrical permittivity measured over the 2–18 GHz frequency band to create the database of nanocomposite materials have been adopted: such values are interpolated by COMSOL code in order to get out a functional relationship between dielectric properties and frequency. The EM interfaces of the code cover the modeling of EM fields and waves in the frequency domain, formulat‐ ing and solving the differential form of Maxwell's equations together with the initial and boundary conditions.

**Figure 11.** Plot of CEMA 17.

**Figure 12.** Plot of CEMA 20.

**Figure 13.** Plot of CEMA 23.

**Figure 14.** Plot of CEMA 24, optimization of RC.

Fully Configurable Electromagnetic Wave Absorbers by Using Carbon Nanostructures http://dx.doi.org/10.5772/64213 241 243

**Figure 15.** Plot of CEMA 24, optimization of TC.

**Figure 12.** Plot of CEMA 20.

240 242Advances in Carbon Nanostructures

**Figure 13.** Plot of CEMA 23.

**Figure 14.** Plot of CEMA 24, optimization of RC.

**Figure 16.** Mesh of the CEMA 13 (*α* = 0.96), ten-layered structure in the coaxial airline setup.

In **Figures 17** and **18**, the electric field (V/m) is reported at 2.5 and 14 GHz respectively. In **Figures 19** and **20**, the current density (A/m2 ) arrow lines and magnetic field (A/m) circular lines for the ten layers of CEMA13 are shown (arrow lines are used to show the region where the highest EM losses occur along the structure). In **Figures 21** and **22**, the power loss density (W/m3 ) at 2.5 and 14 GHz, respectively, is shown. At 14 GHz, the multilayer design provides the impedance matching with free space, thus resulting in a higher level of EM absorption, as compared to the values of the electric field at 2.5 GHz. Alongside, the signifi‐ cant current density at 14 GHz is well visible in the first and second layers, whereas at 2.5 GHz the impedance mismatch causes a great reflection of the EM field at the first layer, thus determining a lower power loss within the deep bulk. In fact, the power loss density at 2.5 GHz is quite lower compared to the value obtained at 14 GHz, where geometrically power losses are mainly confined within the 5 first layers. In other words, at 2.5 GHz, the EM field propagates through all the layers due to the impedance mismatch end. Due to the higher value of the reflection coefficient, the majority of the power is reflected back at the first air-multi‐ layer interface. At 14 GHz, the multilayer is impedance matched, the reflection coefficient is quite lower and the absorbed power dramatically increases. In **Figure 23**, a comparison

between the RC Target, RC simulated with mathematical model and RC simulated by FEM analysis is shown. The discrepancies with FEM simulation are due to the mesh and the approximation of the EM field solution, which is more difficult at high frequencies.

**Figure 17.** Electric field (V/m) at 2.5 GHz.

**Figure 18.** Electric field (V/m) at 14 GHz.

**Figure 19.** Magnetic field (A/m) circular line and current density (A/m2 ); red arrow lines at 2.5 GHz.

**Figure 20.** Magnetic field (A/m) circular line and current density (A/m2 ); red arrow lines at 14 GHz.

Fully Configurable Electromagnetic Wave Absorbers by Using Carbon Nanostructures http://dx.doi.org/10.5772/64213 243 245

**Figure 21.** Power loss density (W/m3 ) at 2.5 GHz.

between the RC Target, RC simulated with mathematical model and RC simulated by FEM analysis is shown. The discrepancies with FEM simulation are due to the mesh and the

approximation of the EM field solution, which is more difficult at high frequencies.

**Figure 17.** Electric field (V/m) at 2.5 GHz.

242 244Advances in Carbon Nanostructures

**Figure 18.** Electric field (V/m) at 14 GHz.

**Figure 19.** Magnetic field (A/m) circular line and current density (A/m2

**Figure 20.** Magnetic field (A/m) circular line and current density (A/m2

); red arrow lines at 2.5 GHz.

); red arrow lines at 14 GHz.

**Figure 22.** Power loss density (W/m3 ) at 14 GHz.

**Figure 23.** Comparison of RC Target, RC simulated, and RC simulated with FEM.

#### **4. Conclusions**

In this chapter, an attempt to identify a novel approach to design and optimize EM reference materials for metrological applications was introduced. The results highlight the intriguing potentialities of the proposed strategy, since even hard microwave reflection target profiles

are successfully addressed by the numerical design optimization technique developed. It is clear that several upgrades can be introduced in order to better refine the process. The above mentioned examples highlight some root of inaccuracy, especially when quite sharp edges occur along the target profile. Of course, the harder it is to follow a given profile, the more complex the optimized structure will become out in terms of layers number and thinness. Although the intriguing capability of the PSO algorithm to find solutions fitting such chal‐ lenging targets, it has to be outlined that the corresponding CEMA solutions are composed by at least four layers (almost always six or more), with total thickness over 4 cm, in the most of the situations. Furthermore, quite often the layers that fulfill the crucial EM absorbing role have thickness well below 1mm, thus pointing out that an eventual production process could represent a big task in this framework. Beyond possible PSO algorithm enhancements (convergence parameters, particles population, iterations), the effectiveness of the proposed strategy in terms of mimic capability would be significantly improved by enriching the materials database. The possibility to draw on a "quasi-continuous" spectrum of dielectric properties should allow the PSO algorithm to find solutions for reproducing any RC profile even more faithfully. On the other hand, by proper constraints in terms of layers number and thickness, the increased availability of materials should make the algorithm able to identify multilayered combinations easy to be practically achieved. In this respect, the employment of composites reinforced with nanoparticles sounds as the right way forward. With respect to the conventional materials, in fact, the family of nano-filled composites allows the unique opportunity to finely adjust the EM absorption property just by tuning the weight percent‐ age ofthe nanofiller within the matrix. Atthis purpose, the lowerthe incremental step in weight percentage, the greater the possibility of fine tuning: thus, the greater the database popula‐ tion, the better the possibility to select appropriate materials for mimic purposes. The research development will thus be addressed to enrich the database population, in orderto obtain even more precise and feasible solutions for the widest variety of electromagnetic requirements.

## **Author details**

Davide Micheli\* , Roberto Pastore, Antonio Vricella and Mario Marchetti

\*Address all correspondence to: davide.micheli@uniroma1.it

Astronautic Electric and Energy Engineering Department, Sapienza University of Rome, Rome, Italy

## **References**

[1] Sheffield RG, The official F-19 stealth fighter handbook, Radnor, Pennsylvania: Pam Williams ed, 1989. ISBN 10: 0874552176: 1–184.

[2] Knott EF, Shaeffer J, Tuley M, Radar cross section, 2nd edn, SciTech Publishing Inc., 2004: 112–115.

are successfully addressed by the numerical design optimization technique developed. It is clear that several upgrades can be introduced in order to better refine the process. The above mentioned examples highlight some root of inaccuracy, especially when quite sharp edges occur along the target profile. Of course, the harder it is to follow a given profile, the more complex the optimized structure will become out in terms of layers number and thinness. Although the intriguing capability of the PSO algorithm to find solutions fitting such chal‐ lenging targets, it has to be outlined that the corresponding CEMA solutions are composed by at least four layers (almost always six or more), with total thickness over 4 cm, in the most of the situations. Furthermore, quite often the layers that fulfill the crucial EM absorbing role have thickness well below 1mm, thus pointing out that an eventual production process could represent a big task in this framework. Beyond possible PSO algorithm enhancements (convergence parameters, particles population, iterations), the effectiveness of the proposed strategy in terms of mimic capability would be significantly improved by enriching the materials database. The possibility to draw on a "quasi-continuous" spectrum of dielectric properties should allow the PSO algorithm to find solutions for reproducing any RC profile even more faithfully. On the other hand, by proper constraints in terms of layers number and thickness, the increased availability of materials should make the algorithm able to identify multilayered combinations easy to be practically achieved. In this respect, the employment of composites reinforced with nanoparticles sounds as the right way forward. With respect to the conventional materials, in fact, the family of nano-filled composites allows the unique opportunity to finely adjust the EM absorption property just by tuning the weight percent‐ age ofthe nanofiller within the matrix. Atthis purpose, the lowerthe incremental step in weight percentage, the greater the possibility of fine tuning: thus, the greater the database popula‐ tion, the better the possibility to select appropriate materials for mimic purposes. The research development will thus be addressed to enrich the database population, in orderto obtain even more precise and feasible solutions for the widest variety of electromagnetic requirements.

, Roberto Pastore, Antonio Vricella and Mario Marchetti

Astronautic Electric and Energy Engineering Department, Sapienza University of Rome,

[1] Sheffield RG, The official F-19 stealth fighter handbook, Radnor, Pennsylvania: Pam

\*Address all correspondence to: davide.micheli@uniroma1.it

Williams ed, 1989. ISBN 10: 0874552176: 1–184.

**Author details**

244 246Advances in Carbon Nanostructures

Davide Micheli\*

Rome, Italy

**References**


[14] Hoeben R, Troch PA. Assimilation of active microwave observation data for soil moisture profile estimation. Water Resour. Res. 2000; 36, 2805–2819. doi:

[15] Martinez-Lorenzo JA, Rappaport C, Sullivan R, Angell A. Standoff concealed explo‐ sives detection using millimeter-wave radar to sense surface shape anomalies. Antennas and Propagation Society International Symposium, 2008. AP-S 2008. IEEE.

[16] Robert J, Douglass J, Gorman D, Burns TJ. System and method for standoff detection of human carried explosives. Patent US 7800527 B2, PCT/US2005/036593, 2010. [17] Micheli D, Apollo C, Pastore R, Marchetti M. Modeling of microwave absorbing structure using winning particle optimization applied on electrically conductive nanostructured composite material. 19th ICEM 2010, Rome (IT) 2010; ISBN

[18] Micheli D, Pastore R, Marchetti M, Gradoni G, Moglie F, Mariani Primiani V. Modeling and measuring of microwave absorbing and shielding nanostructured materials. IEEE EMC Europe 2012, Rome (IT) 2012; ISBN 978-1-4673-0718-5: 1–5. doi:10.1109/

[19] Micheli D, Radar absorbing materials and microwave shielding structure design, LAP

[20] Micheli D, Vricella A, Pastore R, Marchetti M. Synthesis and electromagnetic charac‐ terization of frequency selective radar absorbing materials using carbon nanopow‐

[21] Micheli D, Pastore R, Gradoni G, Mariani Primiani V, Moglie F, Marchetti M. Reduction of satellite electromagnetic scattering by carbon nanostructured multilayers. Acta

[22] Micheli D, Pastore R, Apollo C, Marchetti M, Gradoni G, Mariani Primiani V, Moglie F. Broadband electromagnetic absorbers using carbon nanostructure-based compo‐ sites. IEEE Trans. Microwave Theory Tech. 2011; 59(10): 2633–2646. doi:10.1109/

[23] Micheli D, Apollo C, Pastore R, Barbera D, Bueno Morles R, Marchetti M, Gradoni G, Mariani Primiani V, Moglie F. Optimization of multilayer shields made of composite nanostructured materials. IEEE Trans. Electromagn. C 2012; 54(1): 60–69. doi:10.1109/

[24] Pryor RW. Multiphysics modeling using COMSOL: a first principles approach.

Lambert Academic Publishing 2012; ISBN 978-3-8465-5939-0: 334–446.

ders. Carbon 2014; 77: 756–774. doi:10.1016/j.carbon.2014.05.080

Astron 2013; 88: 61–73. doi:10.1016/j.actaastro.2013.03.003

978-1-4244-4174-7: 1–10. doi:10.1109/ICELMACH.2010.5607881

10.1029/2000WR900100

246 248Advances in Carbon Nanostructures

EMCEurope.2012.6396810

TMTT.2011.2160198

TEMC.2011.2171688

Sudbury: Jones & Bartlett Learning; 2009.

pp 1–4. doi:10.1109/APS.2008.4619051

*Edited by Adrian M.T. Silva and Sonia A.C. Carabineiro*

Carbon atoms have the amazing ability to bond in remarkable different manners that can assume distinct astonishing dimensional arrangements from which absolutely diverse and interesting nanostructured carbon materials are obtained. This book aims to cover the most recent advances in (i) Graphene and derivatives, including graphene-based magnetic composites, membranes, wafer devices, and nanofibers for several applications, as well as some particular properties, such as light emission from graphene; (ii) Carbon nanotubes heaters and fibers for reinforcement of cement and diamond-based thin films; and (iii) Nanofluids consisting of both graphene and carbon nanotubes, apart from reporting some important case studies dealing with carbon nanostructures and their use in sensors, coatings, or electromagnetic wave absorbers.

Photo by XStudio3D / Can Stock

Advances in Carbon Nanostructures

Advances in Carbon

Nanostructures

*Edited by Adrian M.T. Silva* 

*and Sonia A.C. Carabineiro*