**2.2.1 Synthesis of TiO2**

342 Solar Radiation

The production of reactive species by a TiO2 photocatalyst is influenced by a series of factors, such as surface acidity and pH of the reaction medium, control of the kinetic of recombination of charge carriers, interfacial electron-transfer rate, optical absorption of the semicondutor, phase distribution, morphology, specific surface area and porosity (Hoffmann et al., 1995; Furube et al, 2001; Diebold, 2003; Carp et al., 2004; Kumar & Devi

The reactions (1) to (4) combined with other (Hoffmann et al., 1995; Machado et al., 2008; Kumar & Devi, 2011) give an approximate view of the chain reactions that compose a

Different semiconductors are able to trigger the heterogeneous photocatalytic processes.

TiO2 stands in front of others for its abundance, low toxicity, good chemical stability over a wide pH range, photosensitivity, photostability, insolubility in water, low cost, chemical inertness, biological and chemical inertness, and stability to corrosion and photocorrosion (Martin et al., 1995; Augugliaro et al., 2007). However, its band gap energy limits, in principle, its application in photocatalytic processes induced by solar radiation, since the radiation incident on the biosphere consists of approximately 5 % UV, 43 % visible and 52 %,

The introduction of changes in the crystalline structure of TiO2 through the introduction of dopant ions and/or modifying ions and associations between TiO2 and other semiconductor oxidesin order to expand the use of incident radiation, is particularly important if the aim is to use solar radiation in photocatalytic processes. The synthesis of new materials based on TiO2 has resulted in substantial progress towards the improvement of the photocatalytic activity of this semiconductor (Imhof & Pine, 1997; Cavalheiro et al.,2008; Eguchi et al., 2001; Agostiano et al., 2004; Machado et al., 2008; Zaleska et al., 2010; Batista, 2010; Machado et al.,

Titanium dioxide can be found in nature in the form of three different polymorphs: Anatase, Rutile and Brookite (Hanaor & Sorrell, 2011; Khataee et al., 2011; Kumar & Devi, 2011). Among these polymorphs, the thermodynamically more stable is the rutile, which can be obtained from the conversion of anatase, which in turn is the most photoactive polymorph

Technological applications of titanium oxide are quite large. In addition to the previously described, TiO2 has been used in filters to absorb ultraviolet radiation (sunscreens, for example), pigments, in chemical sensors for gases (Pichat et al., 2000), as constituents of

Strong light absorption and suitable redox potential are prerequisites for photocatalytic reactions. Growing interest has focused on doped TiO2 catalysts (Ohno et al., 2003; Luo et al., 2004; Li et al., 2005; Labat et al., 2008; Yang et al., 2008; Long et al., 2009; Zhang et al., 2010; Zaleska et al., 2010; Iwaszuk & Nolan, 2011; Long & English, 2011; Spadavecchia et al., 2011; Kumar & Devi, 2011;), however current achievements are still far from the ideal

ceramic materials for bone and dental implants (Chen et al., 2008), among others.

**2.2 Changes in the structure and surface of titanium dioxide** 

Other in addition to TiO2 are: CdS, ZnO, ZnS, and Fe2O3 (Nogueira & Jardim, 1998).

(2011).

2011b).

goal.

heterogeneous photocatalytic process.

harvesting infrared (Kumar & Devi, 2011).

(Hoffmann et al., 1995; Khataee et al., 2011).

We have performed the synthesis of titanium dioxide using different methodologies (Batista, 2010; Oliveira, 2011). A modification was introduced in the methodology of the synthesis by precipitation of TiO2 using titanium tetraisopropoxide as precursor suggested by Batista (Batista, 2010). It consists in making the whole process, since the solubilization of the precursor in 2-propanol, always under the action of ultrasound. The solid obtained was dried at 60 ° C and subjected to heat treatment at 400 ° C. This new photocatalyst has been adopted in our most recent studies since it has shown impressive photocatalytic activity in the mineralization of different organic substrates (Machado et al., 2011b). As a result, we have studied the introduction of modifications in order to enlarge it, especially expanding it to the visible.

After annealing, the semiconductor was highly crystalline, being only anatase with average crystallite size around 12 nm, estimated from the line width obtained for the peak of greatest intensity in XRD (**Fig.2**). For a semiconductor synthesized according to a similar methodology adopted by Batista, the minimum crystallite size obtained was equal to 22 nm (Batista, 2010).

From the curves of diffuse reflectance, the band gap of the synthesized TiO2 and TiO2 P25 Degussa were estimated. For this, we used Tauc´s method (Wood & Tauc, 1972). For the synthesized TiO2 was obtained a value equal to 3.18 eV while for TiO2 P25 Degussa the estimated band gap was equal to 3.20 eV, in agreement with the value described by many authors (Hoffmann et al., 1995; Machado et al., 2008; Batista, 2010). The earlier versions obtained by precipitation, reported by Batista in his DSc Thesis (Batista, 2010) showed no photocatalytic activity due to its proper degree of aggregation and in some cases limited surface area. Most likely, due to the significant aggregation observed in semiconductor synthesized by Batista (2010), the recombination of charge carriers was more favored at the expense of photocatalytic reactions. It is very likely that the introduction of ultrasound in the synthesis process resulted in significant increase in the dispersion of the particles formed during the formation of critical nuclei, resulting in the precipitation of particles with minimal or no aggregation. Morphological characterization of this new photocatalyst is ongoing.

Fig. 2. X-ray diffraction patterns of TiO2 synthesized by solubilization of titanium tetraisopropoxide in 2-propanol and subsequent hydrolysis and precipitation by slow addition of ultrapure water.

The mineralization of food dye trartrazine, C.I. 19140, mediated by this new photocatalyst is presented as an example. It was promoted at pH 3, using 100 mg/L of photocatalyst, in experiments on laboratory scale, using as radiation source a 400 W high pressure mercury vapor lamp. 4 L of the model effluent were used per experiment. Hydrogen peroxide (166 mg/L) was added as an extra font of radicals (Machado et al., 2003a). The results were compared to the obtained under the same conditions using TiO2 P25 Degussa as photocatalyst. Additionally, all photolysed samples underwent the following tests: pH monitoring, spectrophotometric measurements through the use of a UV/VIS dual beam Shimadzu UV-1650PC spectrophotometer. The aliquots collected in the experiments in the presence of the photocatalyst were filtered using Millipore filters (0.45 m of mean pore size) to remove suspended TiO2 before the measurements. The experimental setup is similar to that described in previous studies (Machado, 2003; Oliveira, 2012).

After 120 minutes of reaction, 52% of mineralization was reached with the use of the synthesized TiO2. For TiO2 P25-mediated degradation, the mineralization was 84% under

the synthesis process resulted in significant increase in the dispersion of the particles formed during the formation of critical nuclei, resulting in the precipitation of particles with minimal or no aggregation. Morphological characterization of this new photocatalyst is

20 40 60 80

**Bragg´s angle** 

The mineralization of food dye trartrazine, C.I. 19140, mediated by this new photocatalyst is presented as an example. It was promoted at pH 3, using 100 mg/L of photocatalyst, in experiments on laboratory scale, using as radiation source a 400 W high pressure mercury vapor lamp. 4 L of the model effluent were used per experiment. Hydrogen peroxide (166 mg/L) was added as an extra font of radicals (Machado et al., 2003a). The results were compared to the obtained under the same conditions using TiO2 P25 Degussa as photocatalyst. Additionally, all photolysed samples underwent the following tests: pH monitoring, spectrophotometric measurements through the use of a UV/VIS dual beam Shimadzu UV-1650PC spectrophotometer. The aliquots collected in the experiments in the presence of the photocatalyst were filtered using Millipore filters (0.45 m of mean pore size) to remove suspended TiO2 before the measurements. The experimental setup is similar

After 120 minutes of reaction, 52% of mineralization was reached with the use of the synthesized TiO2. For TiO2 P25-mediated degradation, the mineralization was 84% under

Fig. 2. X-ray diffraction patterns of TiO2 synthesized by solubilization of titanium tetraisopropoxide in 2-propanol and subsequent hydrolysis and precipitation by slow

to that described in previous studies (Machado, 2003; Oliveira, 2012).

ongoing.

**Intensity (a.u.)**

addition of ultrapure water.

the same conditionsIn the absence of H2O2, the levels of mineralization were respectively 24 and 38% for the synthesized TiO2 and TiO2 P25 **(Fig. 3).** The mineralization was estimated from measurements of dissolved organic carbon using a Shimadzu TOC-VCPH Total Organic Carbon Analyzer.

Fig. 3. Mineralization of tartrazine by heterogeneous photocatalysis using: (a) TiO2 synthesized in reaction in the absence of H2O2; (b) TiO2 P25 in reaction in the absence of H2O2; (c) TiO2 synthesized, in reaction in the presence of H2O2, and (d) TiO2 P25 reaction in the presence of H2O2.

The changes introduced during the solubilization and synthesis process itself should have been enough to guarantee a level of ordering of the particles formed. The final product after thermal treatment of the oxide formed proved to be 100% anatase.

## **2.2.2 Photocatalysts based on the association between a photosensitizing dye and a semiconductor oxide**

Electron transfer at the interface between a photoactive species and the semiconductor surface is a fundamental aspect for organic semiconductor devices (Grätzel, 2001; Ino et al., 2005). Certain photoactive compounds has proven to be able, when electronically excited, to inject electrons in the conduction band of semiconductors (Grätzel, 2001; Ino et al., 2005; Rehm et al., 1996; Nazeeruddin et al., 1993; Asbury et al., 2001; Krüger et al., 2001; Argazzi et al., 1998; Xargas et al., 2000; Tennakone et al., 1997; Sharma et al., 1991; Hao et al., 1998; Chen et al., 1997; Wu et al., 2000), increasing the performance of dye-sensitized solar cells. In particular, ultrafast charge separation led by electron injection from electronically excited photoactive molecules to the conduction band of a wide-gap metal oxide, and a good electronic coupling between dye molecules and surface of the substrate are key steps for improving the performance of these materials (Rehm et al., 1996; Nazeeruddin et al., 1993; Asbury et al., 2001). In the dye sensitization process, dye gets excited rather than the TiO2 particles to appropriate singlet and triplet states, being subsequently converted to cationic dye radicals after electron injection to TiO2 CB (Benko et al., 2002). The electrons injected to TiO2 CB react with the preadsorbed O2 to form oxidizing species (superoxide, hydroperoxyl and hydroxyl radicals) which combined to the species produced from photoexcited TiO2, induce oxidative reactions (Wu et al., 1998). Thus TiO2 plays an important role in electrontransfer mediation, even though TiO2 itself is not excited. A photodegradation mechanism of dyes under visible irradiation without TiO2 photoexcitation was recently presented by Kumar & Devi (2011). The formation of singlet oxygen has been reported in some cases (Stylidi et al., 2004).

The association between photosensitizing dyes and oxides semiconductors with photocatalytic activity constitutes a strategy for obtaining more efficient photocatalysts for a wider range of applications. These photosensitizing dyes, when excited by photons of lower energy, allow the injection of electrons from these species to the conduction band of the semiconductor increasing the concentration of charge carriers (Benkö et al., 2002; Sharma et al., 2006; Machado et al., 2008; Shang et al., 2011; Kumar & Devi, 2011). The electrons, in turn, can be transferred to reduce organic acceptors adsorbed on the catalyst surface (Machado et al.,2008). Thus, the photocatalyst composites containing a photosensitizing dye associated with the photoactive semiconductor have, in general, improved photocatalytic activity. The possibility of utilization of solar radiation, because they have the range of absorption expanded to the visible, makes it possible achieve important contributions in solving problems concerning effluent treatment (Machado et al.,2008). Machado and coworkers (2003b; 2008; 2011; Duarte et al., 2005) have studied composites prepared by the association between zinc phthalocyanine (ZnPc) and titanium dioxide, obtained by coating TiO2 particles using a solution of zinc phthalocyanine followed by controlled drying of the organic suspension (Machado et al., 2008). These materials have been intensively characterized (Machado et al., 2008; Batista et al., 2011). A decrease between 20 and 30% in the specific surface area (SSA) is verified for the composites when compared to the TiO2 P25 (Machado et al., 2008; Oliveira et al., 2011; Batista et al., 2011). This difference should be as a result of the incorporation of ZnPc aggregates on the surface of the semiconductor. The changes in the specific area caused by the incorporation of zinc phthalocyanine do not imply distortions in the crystal structure (Machado et al., 2008). Scanning tunneling microscopy of different metal phthalocyanines confirm that the above mentioned aggregates are adsorbed onto the semiconductor surface (Qiu et al., 2004).

For these composites, the surface sensitization by electron transfer via physisorbed ZnPc should compensate the decrease in surface area, increasing the efficiency of the photocatalytic process. It should be emphasized that the extended range of wavelengths shifted to the visible region of the electromagnetic spectrum, which is capable of positively influencing the electron transfer between the excited dye and the semiconductor conduction band tends to improve electron–hole separation (Machado et al., 2008; Carp et al., 2004;

particular, ultrafast charge separation led by electron injection from electronically excited photoactive molecules to the conduction band of a wide-gap metal oxide, and a good electronic coupling between dye molecules and surface of the substrate are key steps for improving the performance of these materials (Rehm et al., 1996; Nazeeruddin et al., 1993; Asbury et al., 2001). In the dye sensitization process, dye gets excited rather than the TiO2 particles to appropriate singlet and triplet states, being subsequently converted to cationic dye radicals after electron injection to TiO2 CB (Benko et al., 2002). The electrons injected to TiO2 CB react with the preadsorbed O2 to form oxidizing species (superoxide, hydroperoxyl and hydroxyl radicals) which combined to the species produced from photoexcited TiO2, induce oxidative reactions (Wu et al., 1998). Thus TiO2 plays an important role in electrontransfer mediation, even though TiO2 itself is not excited. A photodegradation mechanism of dyes under visible irradiation without TiO2 photoexcitation was recently presented by Kumar & Devi (2011). The formation of singlet oxygen has been reported in some cases

The association between photosensitizing dyes and oxides semiconductors with photocatalytic activity constitutes a strategy for obtaining more efficient photocatalysts for a wider range of applications. These photosensitizing dyes, when excited by photons of lower energy, allow the injection of electrons from these species to the conduction band of the semiconductor increasing the concentration of charge carriers (Benkö et al., 2002; Sharma et al., 2006; Machado et al., 2008; Shang et al., 2011; Kumar & Devi, 2011). The electrons, in turn, can be transferred to reduce organic acceptors adsorbed on the catalyst surface (Machado et al.,2008). Thus, the photocatalyst composites containing a photosensitizing dye associated with the photoactive semiconductor have, in general, improved photocatalytic activity. The possibility of utilization of solar radiation, because they have the range of absorption expanded to the visible, makes it possible achieve important contributions in solving problems concerning effluent treatment (Machado et al.,2008). Machado and coworkers (2003b; 2008; 2011; Duarte et al., 2005) have studied composites prepared by the association between zinc phthalocyanine (ZnPc) and titanium dioxide, obtained by coating TiO2 particles using a solution of zinc phthalocyanine followed by controlled drying of the organic suspension (Machado et al., 2008). These materials have been intensively characterized (Machado et al., 2008; Batista et al., 2011). A decrease between 20 and 30% in the specific surface area (SSA) is verified for the composites when compared to the TiO2 P25 (Machado et al., 2008; Oliveira et al., 2011; Batista et al., 2011). This difference should be as a result of the incorporation of ZnPc aggregates on the surface of the semiconductor. The changes in the specific area caused by the incorporation of zinc phthalocyanine do not imply distortions in the crystal structure (Machado et al., 2008). Scanning tunneling microscopy of different metal phthalocyanines confirm that the above mentioned aggregates are adsorbed

For these composites, the surface sensitization by electron transfer via physisorbed ZnPc should compensate the decrease in surface area, increasing the efficiency of the photocatalytic process. It should be emphasized that the extended range of wavelengths shifted to the visible region of the electromagnetic spectrum, which is capable of positively influencing the electron transfer between the excited dye and the semiconductor conduction band tends to improve electron–hole separation (Machado et al., 2008; Carp et al., 2004;

(Stylidi et al., 2004).

onto the semiconductor surface (Qiu et al., 2004).

Wang et al., 1997; Shourong et al., 1997; Zhang et al., 1997; Zhang et al., 1998). These composites have shown to be better photocatalysts for wastewater decontamination, mainly mediated by visible light, than pure TiO2 (Machado et al., 2003b; Duarte et al., 2005; Machado et al., 2008; França, 2011; Oliveira et al., 2012), performance that remains even when reused (Machado et al., 2008).

The zero point charge pH (pHZPC) was estimated for TiO2 P25 and a composite containing 1.6% m/m of ZnPc by zeta potential measurements, carried out in a disperse suspension using a Zetasizer Nano ZS90. The estimated value for the composite, pHZPC = 5.50, lower than the one for P25 (pHZPC = 6.25) suggests a differentiated behavior for the composite since its surface is negatively charged in a pH range in which P25 is still with the surface positively charged. The value measured for TiO2 P25 agrees with the reported in the literature (Hoffmann et al., 1995). The morphological characteristics of both samples were investigated by SEM, carried out in a Philips XL-30 microscope coupled to a eld emission gun and a EDX analytical setup. The micrographs show the occurrence of macro-aggregates in the composite and spherical particles around 25 nm in P25. The estimated concentration of ZnPc on P25 surface is around 1.6%, confirmed by EDX measurements (Batista et al., 2011). Also, the thickness of ZnPc coating, homogeneity, and aggregation on the TiO2 composite surface were evaluated by TEM using a Philips CM-120 microscope. The improvement of visible light absorption in TiO2/ZnPc and electronic surface properties of this composite (Machado et al., 2008) are responsible for an almost three times faster mineralization of Ponceau 4R (C.I. 16255), an azo dye employed in the food industry, when compared with the result obtained using only TiO2 P25, and still much higher than the presented by the other TiO2-based photocatalysts (Oliveira et al., 2012). This dye is classified as a carcinogen in some countries and is currently listed as a banned substance by U.S. Food and Drug Administration (FDA).

The highest photocatalytic activity of TiO2/ZnPc 1.6% seems to be the result of synergism between the photocatalytic characteristics inherent to TiO2 P25 with the redox properties and charge transport of ZnPc Frenkel's "J" aggregates on the semiconductor surface (Fidder et al., 1991; Kim et al., 2006; Machado et al., 2008; Machado et al., 2011a). The sensitization of TiO2 P25, induced by zinc phthalocyanine aggregates was effective in producing more active photocatalysts.

**Fig. 4** presents the diffuse reflectance spectra (DRS) of ZnPc, TiO2 and some of the studied TiO2/ZnPc composites.

Unlike what occurs with TiO2 (**Fig. 4a**), for composite materials obtained by the association between TiO2 and ZnPc there is a significant electronic absorption for wavelengths above 390 nm. Comparison between the graphs presented in As can be seen in **Fig. 4 (a to e),** the UV-Vis absorption spectrum (DRS) of these composites is not the result of an additive effect between the absorption spectra of the precursors. The absorption spectra of the composites are quite different from the typical absorption profiles of TiO2 (**Fig. 4a**) and pure ZnPc in the solid state (**Fig. 4f**) or even in very dilute liquid solutions (Miranda et al., 2002).

The absorption spectrum of these composites is characterized by an intense absorption band below 460 nm, and a large, intense and non structured absorption band above 475 nm. Both bands are most probably the result of superposition of electronic states of TiO2 and ZnPc aggregates.

Fig. 4. Diffuse reectance spectra (DRS) of TiO2 and TiO2/ZnPc composites, prepared with different percent in mass of ZnPc. TiO2 P25 (a) and composite containing: 1.0%of ZnPc (b); 2.5%of ZnPc (c); 5.0%of ZnPc (d); composite containing 2.5% of ZnPc, using TiO2 P25 as reference (e) and DRS of pure ZnPc (f). Barium sulphate was used as reference for (a) to (d) (Machado et al., 2008).

E**g**(Composites)=2.7 eV

300 400 500 600 700 800

**(a)**

Q Band

**(f)**

**(b)**

**(c)**

**(e)**

Q-Band

**(d)**

Wavelength, nm

300 400 500 600 700 800

Wavelengthnm

Fig. 4. Diffuse reectance spectra (DRS) of TiO2 and TiO2/ZnPc composites, prepared with different percent in mass of ZnPc. TiO2 P25 (a) and composite containing: 1.0%of ZnPc (b); 2.5%of ZnPc (c); 5.0%of ZnPc (d); composite containing 2.5% of ZnPc, using TiO2 P25 as reference (e) and DRS of pure ZnPc (f). Barium sulphate was used as reference for (a) to (d)

0.0

10

(Machado et al., 2008).

15

20

25

Soret (B) band

e**u**

F(R)

30

35

40

45

0.5

Soret

1.0

F(R)

E**g**(TiO2

)=3.1 eV

e (n,\* transition) **<sup>g</sup>**

1.5

In **Fig. 4e** the shape of the bands in the ultraviolet and visible portions of the electronic spectrum of the composite containing 2.5% m/m of ZnPc, obtained using TiO2 as reference, is very different from that observed for pure ZnPc in the solid state, **Fig. 4f**. In the visible, it presents a large and intense three peak band centered by a red shifted Q band, with maximum at 683 nm. The batochromic shift of the absorption maximum associated to Q band, suggests the occurrence of Frenkel's J aggregates of ZnPc (**Fig. 5**) in the composites (Köhler & Schmid, 1996; Eisfeld & Briggs, 2006; Chen, Z. et al., 2008), which agrees with results of a theoretical study employing methods of Density Functional Theory on the formation of aggregates of zinc phthalocyanine (Machado et al*.*, 2011a). The bathochromic shift of the absorption maximum of the Q band highlights the differentiated nature of these compounds against pure TiO2 and ZnPc. The Soret (B) band also presents a different shape compared to its equivalent in pure ZnPc in the solid state (**Fig. 4f**), and is red shifted. The spectrum of **Fig. 4e** is very similar to the absorption spectrum for a flash-evaporated ZnPc thin film deposited on a glass substrate (Senthilarasu et al., 2003), in which the two energy bands characteristic of phthalocyanines are evident, one in the region between 500 and 900 nm, with an absorption peak at 690 nm, related to the Q band, and the other, very intense, at 330 nm, attributed to Soret (B) band (Meissner & Rostalski, 2001), similar to that reported for the absorption spectrum for thin films of Magnesium Phthalocyanine (Mi et al., 2003). The unstructured band in the visible and the red shifted Q band of these composites can be attributed to the strong intermolecular interactions due to ZnPc aggregation (ZnPcagg), resulting in coupling effects of excitons on the allowed transitions, with significant effects on the mobility of charge carriers (Hoffmann, 2000).

Fig. 5. Representation of the molecular structure of a Frenkel's J aggregate of ZnPc formed by four grouped individual molecules, indicating the sharing the same ligand MO between the ZnPc 2 and 3, in the HOMO (Machado et al., 2011a).

**Fig. 4f** presents the diffuse reflectance spectrum of pure ZnPc. The intense absorption peak at 552 nm, is related to the Q band and is attributed to very intense \* transitions (Leznoff & Lever, 1990). The Soret band presents an absorption maximum at 301 nm. A low intensity and non structured absorption band with the absorption peak centered at 416 nm, is related to an n\* transition involving the e**u** azanitrogen lone pair orbital with the eg LUMO (Ricciardi et al., 2001). A set of three very small intensity low energy bands, above the Q band, can also be observed.

The Eg value for the TiO2/ZnPc composites, 2.7 eV, lower than the estimated for pure TiO2 (Hoffmann et al., 1995), has a value similar to the estimated for iron (II) phthalocyanine excitons (2.6 eV) in TiO2/FePc blends (Sharma et al., 2006) and other metal phthalocyanine associated to semiconductor oxides (Iliev et al., 2003). For ZnPc thin films, Senthilarasu et al. assigned an Eg of 1.97 eV (Senthilarasu et al., 2003) with a directly allowed optical transition, near the value estimated for the peak absorption Q-band (2.25 eV) of pure ZnPc in the solid state (**Fig. 4f**). The Eg for the composites might be related to the coupling between TiO2 and ZnPc electronic states and their positive implications. Similar to TiO2/FePc blends (Sharma et al., 2006) and ZnPc thin films (Ino et al., 2005; Senthilarasu et al., 2003), the photoexcitation of ZnPc aggregates should result in the formation of e- /ZnPc+ pairs, followed by electron transfer from ZnPc excitons to the conduction band of bulk TiO2, which explains at least in part the improved photocatalytic activity observed for some of the ZnPc/TiO2 composites (Machado et al., 2008; Oliveira et al., 2012). Sharma et al. reported charge separation after photo-excitation of TiO2/FePc composite film due to charge transfer from FePc to TiO2 resulting in FePc(h+) and TiO2(e-) (Sharma et al., 2006). Additionally, they reported that the charge transport and the current leakage through FePc films and the photo-generation are due to the efficient dissociation of exciton at the donor–acceptor interface of the bulk, and that the higher holes mobility in the organic material layer, combined with lower conductance leakage, leads to the more efficient collection of photogenerated carriers. Thus, the electronic coupling strength between donor and acceptor is one of the critical conditions to ensure the occurrence of such electron transfer (Ino et al., 2005; Rehm et al., 1996; Senthilarasu et al., 2003; Meissner & Rostalski, 2001).

The spectrum presented in **Fig. 4e** is very similar to the absorption spectrum for a flashevaporated ZnPc thin film deposited on a glass substrate (Senthilarasu et al., 2003), in which the two energy bands characteristic of phthalocyanines are evident, one in the region between 500 and 900 nm, with an absorption peak at 690 nm, related to the Q band, and the other, very intense, at 330 nm, attributed to Soret (B) band (Meissner & Rostalski, 2001), similar to that reported for the absorption spectrum for thin films of Magnesium Phthalocyanine (Mi et al., 2003).

#### **2.3 Solar photocatalysis using a compound parabolic concentrator (CPC) reactor**

#### **2.3.1 Design and construction of a CPC reactor**

The study of new technologies has now focused on decontamination methods feasible alternatives that are environmentally friendly, and allow its application in large scale, with easy operation and low cost.

The economic use of AOPs based on the use of solar radiation in the treatment of wastewater has been proposed for their low cost, especially in regions with high insolation

**Fig. 4f** presents the diffuse reflectance spectrum of pure ZnPc. The intense absorption peak at 552 nm, is related to the Q band and is attributed to very intense \* transitions (Leznoff & Lever, 1990). The Soret band presents an absorption maximum at 301 nm. A low intensity and non structured absorption band with the absorption peak centered at 416 nm, is related to an n\* transition involving the e**u** azanitrogen lone pair orbital with the eg LUMO (Ricciardi et al., 2001). A set of three very small intensity low energy bands, above

The Eg value for the TiO2/ZnPc composites, 2.7 eV, lower than the estimated for pure TiO2 (Hoffmann et al., 1995), has a value similar to the estimated for iron (II) phthalocyanine excitons (2.6 eV) in TiO2/FePc blends (Sharma et al., 2006) and other metal phthalocyanine associated to semiconductor oxides (Iliev et al., 2003). For ZnPc thin films, Senthilarasu et al. assigned an Eg of 1.97 eV (Senthilarasu et al., 2003) with a directly allowed optical transition, near the value estimated for the peak absorption Q-band (2.25 eV) of pure ZnPc in the solid state (**Fig. 4f**). The Eg for the composites might be related to the coupling between TiO2 and ZnPc electronic states and their positive implications. Similar to TiO2/FePc blends (Sharma et al., 2006) and ZnPc thin films (Ino et al., 2005; Senthilarasu et al., 2003), the

followed by electron transfer from ZnPc excitons to the conduction band of bulk TiO2, which explains at least in part the improved photocatalytic activity observed for some of the ZnPc/TiO2 composites (Machado et al., 2008; Oliveira et al., 2012). Sharma et al. reported charge separation after photo-excitation of TiO2/FePc composite film due to charge transfer from FePc to TiO2 resulting in FePc(h+) and TiO2(e-) (Sharma et al., 2006). Additionally, they reported that the charge transport and the current leakage through FePc films and the photo-generation are due to the efficient dissociation of exciton at the donor–acceptor interface of the bulk, and that the higher holes mobility in the organic material layer, combined with lower conductance leakage, leads to the more efficient collection of photogenerated carriers. Thus, the electronic coupling strength between donor and acceptor is one of the critical conditions to ensure the occurrence of such electron transfer (Ino et al., 2005;

The spectrum presented in **Fig. 4e** is very similar to the absorption spectrum for a flashevaporated ZnPc thin film deposited on a glass substrate (Senthilarasu et al., 2003), in which the two energy bands characteristic of phthalocyanines are evident, one in the region between 500 and 900 nm, with an absorption peak at 690 nm, related to the Q band, and the other, very intense, at 330 nm, attributed to Soret (B) band (Meissner & Rostalski, 2001), similar to that reported for the absorption spectrum for thin films of Magnesium

**2.3 Solar photocatalysis using a compound parabolic concentrator (CPC) reactor** 

The study of new technologies has now focused on decontamination methods feasible alternatives that are environmentally friendly, and allow its application in large scale, with

The economic use of AOPs based on the use of solar radiation in the treatment of wastewater has been proposed for their low cost, especially in regions with high insolation

/ZnPc+ pairs,

photoexcitation of ZnPc aggregates should result in the formation of e-

Rehm et al., 1996; Senthilarasu et al., 2003; Meissner & Rostalski, 2001).

the Q band, can also be observed.

Phthalocyanine (Mi et al., 2003).

easy operation and low cost.

**2.3.1 Design and construction of a CPC reactor** 

(Malato et al., 2002; Machado et al., 2003; Sattler et al., 2004a, 2004b; Machado et al.,2004; Palmisano et al., 2007; Machado et al., 2008; Torres et al., 2008; Li et al., 2009). Literature reports suggest that the reactors most suitable for application in solar photocatalysis are CPC type (Malato et al., 1997; Malato et al., 2002; Sattler et al., 2003a, 2003b; Machado et al.,2004; Duarte et al., 2005; Machado et al.,2008).

CPC reactors are static collectors of solar radiation with reflective surfaces in the form of involute positioned around cylindrical tubes, **Fig. 6**. Reflectors with this geometry allows the pock up of solar radiation, either by direct incidence, as the diffuse radiation, directing it to a glass tube through which circulates the effluent to be treated (Duarte et al., 2005).

Fig. 6. Representation in two angles (a and b) of a CPC reactor, detailing one of the reflectors in the form of involute (c), and pipes the fixed to the body of the reactor (d).

Our CPC reactor was designed to process up to 150 L of effluent, This reactor consists in a module with an aperture of about 1.62 m2, elevation angle adjusted to the latitude of Uberlândia, Brazil (19o S), ensuring a better use of incident radiation. The reflecting surface contains 10 borosilicate glass tubes (external diameter 32 mm, wall thickness of 1.4 mm, and length of 1500 mm), mounted in parallel and connected in series, each on double parabolic shaped inox steel reflector surfaces (Duarte et al., 2005). A centrifugal pump of 0.50 HP with rotor and housing made in inert material has been used to ensure a flow of 2 m3/h.

The flow of effluent in tubular reactors is usually turbulent, which may cause loss of efficiency in the capture of solar radiation. However, this difficulty can be minimized during the design of the reactor, and the use of balanced amounts of the catalyst, in the case of heterogeneous photocatalysis, so as to guarantee a uniform flow and a good dispersion of the photocatalyst in the effluent to be treated, minimizing possible effects of co-absorption of the incident radiation (Duarte et al., 2005). Non-uniform flows implies in non uniform residence times that can lower efficiency compared to the ideal conditions (Koca & Sahin, 2002). In the case of the heterogeneous processes with photocatalyst powder in suspension, sedimentation and depositing of the catalyst along the hydraulic circuit should be avoided and turbulent flow in the reactor needs be guaranteed. Reynolds's number varying between 10 000–50 000 ensures fully turbulent flow and avoids the settlement of the photocatalyst particles in the tubes (Malato et al., 2002). In our project, the Reynolds' number were defined as being Reglass = 34,855.4 and RePVC = 40,070.0, for glass and PVC, the materials where the effluent with the photocatalyst in suspension circulate.

Details of the project of a CPC reactor similar to the built in our laboratory are available in Duarte et al., 2005.

#### **2.3.2 Photocatalytic degradation of organic substrates using solar radiation**

#### **2.3.2.1 Degradation of organic matter present in a model-effluent simulating the wastewater produced by a pulp and paper industry, using TiO2 P25 and the composite TiO2/ZnPc 2.5% m/m**

The performance of the studied composites to degrade organic matter present in wastewaters, in reactions mediated by solar irradiation, and the possibility of reuse of such photocatalysts, was evaluated monitoring the consumption of the organic matter content during the treatment of three 50 L batches of a model effluent (an aqueous solutions containing 160 mg L-1 of a sodium salt of lignosulphonic acid (Sigma-Aldrich), possessing a mean molecular mass of 52,000 D. The reactions were done at pH 3, with the addition of hydrogen peroxide (30 mg L-1), used as additional source of reactive species (Machado et al., 2003a), and monitored by chemical oxygen demand (COD) analysis of aliquots of effluent samples collected at different accumulated doses of UV-A radiation (this option was due to operational limitations. However, the spectral pattern of the visible light does not change significantly during the execution of the experiments). To evaluate the observed (global) reaction kinetics, the temporal variations were substituted by the UVA accumulated dose, which warrants the reproducibility of these experiments under different latitude and weather conditions. The incident UV-A radiation was monitored using a Solar Light PMA-2100 radiometer. All reactions were stopped when the accumulated dose of UVA reached 900 kJ m-2 (Machado et al., 2008). This corresponds to about 3 hours of sunlight on a sunny day, or 5 to 6 hours during a cloudy day with moderate to high nebulosity in Uberlândia, MG, Brasil (Duarte et al., 2005).

The COD measurements considered the Environmental Protection Agency (EPA) recommended method (Jirka & Carter, 1975).

A same sample of the photocatalyst (100 mg per liter of effluent), containing initially 2.5% of ZnPc, was used to treat the three effluent batches. The treatment of each batch was performed using a CPC (Compound Parabolic Concentrator) reactor (Duarte et al., 2005).

The flow of effluent in tubular reactors is usually turbulent, which may cause loss of efficiency in the capture of solar radiation. However, this difficulty can be minimized during the design of the reactor, and the use of balanced amounts of the catalyst, in the case of heterogeneous photocatalysis, so as to guarantee a uniform flow and a good dispersion of the photocatalyst in the effluent to be treated, minimizing possible effects of co-absorption of the incident radiation (Duarte et al., 2005). Non-uniform flows implies in non uniform residence times that can lower efficiency compared to the ideal conditions (Koca & Sahin, 2002). In the case of the heterogeneous processes with photocatalyst powder in suspension, sedimentation and depositing of the catalyst along the hydraulic circuit should be avoided and turbulent flow in the reactor needs be guaranteed. Reynolds's number varying between 10 000–50 000 ensures fully turbulent flow and avoids the settlement of the photocatalyst particles in the tubes (Malato et al., 2002). In our project, the Reynolds' number were defined as being Reglass = 34,855.4 and RePVC = 40,070.0, for glass and PVC, the materials where the

Details of the project of a CPC reactor similar to the built in our laboratory are available in

The performance of the studied composites to degrade organic matter present in wastewaters, in reactions mediated by solar irradiation, and the possibility of reuse of such photocatalysts, was evaluated monitoring the consumption of the organic matter content during the treatment of three 50 L batches of a model effluent (an aqueous solutions containing 160 mg L-1 of a sodium salt of lignosulphonic acid (Sigma-Aldrich), possessing a mean molecular mass of 52,000 D. The reactions were done at pH 3, with the addition of hydrogen peroxide (30 mg L-1), used as additional source of reactive species (Machado et al., 2003a), and monitored by chemical oxygen demand (COD) analysis of aliquots of effluent samples collected at different accumulated doses of UV-A radiation (this option was due to operational limitations. However, the spectral pattern of the visible light does not change significantly during the execution of the experiments). To evaluate the observed (global) reaction kinetics, the temporal variations were substituted by the UVA accumulated dose, which warrants the reproducibility of these experiments under different latitude and weather conditions. The incident UV-A radiation was monitored using a Solar Light PMA-2100 radiometer. All reactions were stopped when the accumulated dose of UVA reached 900 kJ m-2 (Machado et al., 2008). This corresponds to about 3 hours of sunlight on a sunny day, or 5 to 6 hours during a cloudy day with moderate to high nebulosity in Uberlândia,

The COD measurements considered the Environmental Protection Agency (EPA)

A same sample of the photocatalyst (100 mg per liter of effluent), containing initially 2.5% of ZnPc, was used to treat the three effluent batches. The treatment of each batch was performed using a CPC (Compound Parabolic Concentrator) reactor (Duarte et al., 2005).

**2.3.2 Photocatalytic degradation of organic substrates using solar radiation 2.3.2.1 Degradation of organic matter present in a model-effluent simulating the wastewater produced by a pulp and paper industry, using TiO2 P25 and the composite** 

effluent with the photocatalyst in suspension circulate.

Duarte et al., 2005.

**TiO2/ZnPc 2.5% m/m** 

MG, Brasil (Duarte et al., 2005).

recommended method (Jirka & Carter, 1975).

As reference, an additional effluent batch was treated under similar conditions using pure TiO2 P25 as photocatalyst.

The degradation of the sodium salt of lignosulphonic acid (LSA) suggests higher photocatalytic efficiency for the TiO2/ZnPc composite. **Fig. 7** shows a more effective LSA degradation under the action of TiO2/ZnPc, which increases with reuse, with significant changes in the degradation profile due to the use of the recovered composite While under the action of TiO2 P25 was reached 60% degradation, under the same conditions, with the unused composite, the degradation reached 96%. For the composite in both the first and second reuse, the degradation of the LSA was about 90%. The change in profile suggests that other processes, less likely to occur before, became important for the overall reaction (Machado et al., 2008). The production of singlet oxygen by photosensitization from 3ZnPc\*, for example, is an event plausible if the level of aggregation of ZnPc is reduced. The formation of singlet oxygen has been reported in some cases (Stylidi et al., 2004).

On the other hand, the better hydration of the surface of the composite due to the increasing number of cycles of use, should favor reactions from the valence band.

Fig. 7. Degradation of the organic load present in 50 L of a model waste water containing LSA monitored in terms of relative chemical oxygen demand (COD/COD0), induced by: (1) TiO2 P25; (2) TiO2/ZnPc 2.5%; (3) TiO2/ZnPc 2.5% in the first recycling; (4) TiO2/ZnPc 2.5% in the second recycling.

Despite the fact that part of ZnPc adsorbed to the surface of TiO2 P25 may have been degraded during the photocatalytic process, surprisingly, the photocatalytic efficiency of the composite did not decrease when reused. Results suggest that the composite can be reused at least five times before making any significant loss of photocatalytic efficiency.
