Spectroscopic investigation of the mechanism of photocatalysis.

Reaction mechanisms of various kinds of photocatalysts have been reviewed based on the recent reports, in which various spectroscopic techniques including luminol chemiluminescence photometry, fluorescence probe method, electron spin resonance (ESR), and nuclear magnetic resonance (NMR) spectroscopy were applied. The reaction mechanisms elucidated for bare and modified TiO2 were described individually. The modified visible light responsive TiO2 photocatalysts, i.e., Fe(III)-deposited metal-doped TiO2 and platinum complex-deposited TiO2, were studied by detecting paramagnetic species with ESR, •O2- (or H2O2) with chemiluminescence photometry, and OH radicals with a fluorescence probe method. For bare TiO2, the difference in the oxidation mechanism for the different crystalline form was investigated by the fluorescence probe method, while the adsorption and decomposition behaviors of several amino acids and peptides were investigated by 1H-NMR spectroscopy.

1. Introduction

TiO2 photocatalysts have been widely utilized for the oxidation of organic pollutants [1,2,3,4]. For further practical applications, the improvement in the photocatalytic efficiency and the extension of the effective wavelength to visible region are desired. To develop photocatalysts, understanding of the detailed photocatalytic mechanisms is prerequisite. Recently, the reaction mechanisms of TiO2 photocatalysis have been extensively reviewed [5] and the authors also reviewed the reports published up to 2011 from the view of the detection of active oxygen species [6]. In this manuscript, recent development in the reaction mechanism mainly reported by our group was reviewed. Main techniques used were ESR spectroscopy for the state of photoinduced electron and holes, fluorescence probe method for the formation of OH radical and NMR spectroscopy for the adsorption and decomposition of biological molecules in solution.

2. Spectroscopic Methods for Investigating Photocatalysis

2.1. ESR (Electron Spin Resonance) Spectroscopy

ESR spectroscopy is conventionally used to detect unpaired electrons. Photocatalytic reactions proceed by the two following reactions: reduction of reactants with photoexcited electrons and oxidation of reactants with holes. Therefore, it is important to examine the generation behavior of these active species. In TiO2 photocatalytic systems, two kinds of active species (photoexcited electron and hole) are generated on absorbing photons. Some of the electrons and holes are trapped at Ti and O atoms, to become Ti3+ and O−, respectively. Therefore, by detecting these trapped electrons and holes using ESR spectroscopy under light irradiation of different wavelengths, the generation behavior of excited species can be examined. Moreover, the electron transfer between photocatalysts and co-catalysts can be also examined because if the electron transfer occurs, the amount of the unpaired electron in the co-catalyst should change before and after light irradiation. Therefore, ESR spectroscopy is very useful to elucidate photocatalytic reaction mechanism.

2.2. Chemiluminescence Photometry

Reduced oxygen molecules such as superoxide radical (•O2−) and H2O2 can be detected by chemiluminescence with luminol (LH2, aminodiazabenzoquinone). The one electron oxidized state of luminol (•L−) reacts with •O2− to form the excited state of 3-aminophthalic acid to emit fluorescence in alkaline solution [7], where •L− is formed from LH2 by the oxidation with •O2− [8]. Since •O2− is rather stable in alkaline solution, after the irradiation on photocatalyst was stopped, luminol is injected to measure the amount of •O2− by the chemiluminescence intensity. The same chemiluminescence was obtained from H2O2 by the reaction with L that is two-electron oxidized state of LH2 [9]. To measure the amount of H2O2 in solution, after mixing luminol, hemoglobin was added to oxidize luminol, because L is rather unstable [8]. Luminol chemiluminescence method has some problems. It is available only in alkaline solution, and luminol emits light with SiO2 in the absence of •O2− and H2O2. Therefore, in the case of SiO2 deposited TiO2, instead of luminol, MCLA and lucigenin were employed for the detection of •O2− and H2O2, respectively, by means of chemiluminescence photometry [10].

2.3. Florescence Probe Method

Hydroxyl radical (•OH) has been recognized as a key active species in the oxidation mechanism in photocatalysis [9,11]. For the detection of •OH we employed coumarin. It reacts with •OH to produce 7-OH coumarin (umbelliferone) which emits strong fluorescence [12]. After the irradiation of a coumarin aqueous solution containing photocatalysts powder for a given time, the fluorescence intensity of the fluorescent products (umbelliferone) in the solution was measured. The •OH concentration could be calculated from the concentration of umbelliferone with the aid of data of radiation chemistry [12]. Since carboxyl group is known to adsorb on TiO2, the similar experiments were performed for 3-carboxylic acid derivative of coumarin (CCA, Figure 1), and ensured the reaction with •OH to form OH-CCA as illustrated in Figure 1 [13].

Figure 1

Probing reaction of OH radical with CCA (coumarin 3-carboxy acid) to form fluorescent molecule OH-CCA (7-hydroxy coumarin 3-carboxy acid).

2.4. NMR (Nuclear Magnetic Resonance) Spectroscopy

1H-NMR spectroscopy has been recognized as an effective technique to investigate the behaviors of the reactant molecules in the photocatalytic systems. The adsorption and the decomposition of biomolecules such as amino acids and peptides in the aqueous suspension of photocatalysts can be investigated with 1H-NMR spectroscopy with relatively feasible experimental procedures [14] as follows. Firstly, 1H-NMR of organic molecules dissolved in the solvent are measured. Then certain amount of the photocatalysts is added to the solution. From the initial decrease in the intensity of 1H-NMR peaks of the corresponding reactant molecules the amount of the adsorption can be estimated. Then, by measuring the decrease in the intensities of reactant molecules for various photoirradiation times, one could estimate the photodecomposition rates of the reactants [15].

3. Mechanism of Photocatalysis

3.1. Bare-TiO2 and Visible-Light Responsive TiO2 Photocatalysts

General scheme of photocatalysis applied for the oxidation of pollutant is shown in Figure 2. Light absorption in semiconductor corresponds to the formation of an electron (e−) in the conduction band (CB) and a hole (h+) in the valence band (VB). Usually e− reduces O2 in air to form •O2− and H2O2.

Figure 2

General reaction processes for the photocatalytic oxidation of organic molecules.

The photocatalytic oxidation of organic compounds is accelerated with oxygen [16]. The consumption of O2 at the oxidation site of the photocatalyst has been suggested from the experiment of electrochemical probe reactions at the surface of illuminated TiO2 photoelectrode [17]. Therefore, the generalized oxidation mechanism of organic molecules (RH) can be illustrated as shown in Figure 2. Organic reactants RH will degrade by losing one carbon atom by releasing CO2 through the intermediates like aldehyde R’CHO or carboxylate R’COO−. Although •OH has been often regarded to play an important role in the actual oxidation mechanism of photocatalytic reactions, •OH is not involved in the main oxidation process for organic compounds. In place of it, the surface trapped holes play the role of oxidation, which may be acknowledged as the surface adsorbed •OH in the de-protonated form as stated below.

For the extension of the practical applications of photocatalysts, the utilization of visible light has been intensively promoted. Figure 3 shows the energy levels of several representative visible-light responsive photocatalysts. Since the one-electron reduction potential of O2 is very close to that of the CB bottom of TiO2 and the energy level of VB has sufficient oxidation ability, the shift of the VB by doping N (or, C and S) anions has been attempted to absorb visible light (b). In this case, photogenerated holes at the donor level should have the oxidation ability similarly to that of bare TiO2 [9].

Figure 3

Classification of visible-responsive photocatalysts by the primary reaction mechanism. (a) Unmodified TiO2; (b) Nitrogen doped TiO2; (c) Fe(III) grafted TiO2; (d) Fe(III) grafted metal doped TiO2; (e) Platinum complex deposited TiO2. Adopted with permission from [18]. © 2013 American Chemical Society.

Recently, interfacial charge transfer (IFCT) type absorption originating from the excitation of VB electrons to deposited (or grafted) metal ions has been proposed (c). In this case, if the deposited compound has a catalytic ability of O2 reduction, the efficiency is expected to be increased [19,20]. Since the absorbance of IFCT is very small, to increase the absorption, the transferring of the excited electron to the graft level by doping of metal ions was proposed (d). Photocatalysts of photosensitization type were also proposed, in which the stable metal complex such as PtCl62− is deposited as a sensitizer (e). The deposited compound absorbs the visible light and transfers the excited electron to produce a cation radical D+, which can oxidize organic pollutant molecules. In this case, the enough oxidation power with good stability is required for the oxidized sensitizer D+ [21,22]. To confirm the suggested reaction mechanism, several spectroscopic methods have been applied to the detection of the paramagnetic species produced on the catalysts along with the primary products (•O2−, H2O2, •OH).

3.1.1. Fe(III) Grafted TiO2 Based Photocatalysts

Fe(III) grafted TiO2 (Fe/TiO2) showed the photocatalytic activity under visible light irradiation. The quantum efficiency of Fe/TiO2 prepared under optimized condition was reported to be 22% [20]. We examined the photocatalytic reaction mechanism of the Fe/TiO2 using ESR spectroscopy [23]. As shown in Figure 4A, under visible light irradiation, the ESR signal assigned to Fe3+ (g = 4.3) was decreased and the ESR signal assigned to trapped holes (g = 2.01) at the TiO2 host was observed. In the case of TiO2 without the grafting of Fe3+, the trapped hole signal was scarcely observed as compared to the Fe/TiO2 under visible light irradiation. This means that electrons at VB are directly transferred to the grafted Fe3+ rather than CB (Figure 4B). Using ESR spectroscopy, we could reveal for the first time that the direct electron transfer from the VB of TiO2 to the Fe3+ is the origin of the visible light response.

Figure 4

(A) ESR spectra of (a) Fe3+ and (b) holes for Fe/TiO2 and (c) holes for TiO2 before and after light irradiation with different wavelengths; (B) Photocatalytic reaction mechanism for Fe/TiO2 under visible light irradiation. Reprinted with permission from [23]. © 2012 American Chemical Society.

Photocatalytic reactions cannot proceed when the photogenerated electrons do not react with molecular oxygen, which is the only molecule to be reduced in ambient atmosphere even though the redox potential of the photogenerated hole is positive enough to decompose organic compounds. Therefore, it is important to confirm the reduction of O2 into •O2− by one-electron or H2O2 by two-electron reductions. Under visible light irradiation, for Fe/TiO2, the production of H2O2 was dominant as compared to •O2−. This means that the excited Fe2+ can reduce O2 to H2O2 through two-electron process. Since the electrons having a potential energy of +0.695 V (vs. SHE at pH = 0) can reduce O2 to H2O2 by two-electron process [24], the redox potential of the grafted Fe3+ to Fe2+ was equal to or less than +0.695 V (vs. SHE at pH = 0). Since photogenerated electrons were consumed by the reduction of O2 to H2O2, holes remained at valence band could decompose organic substances efficiently, resulting in the high performance.

Moreover, after the grafting of Fe3+ on the TiO2 doped with metal (M) ions such as Ru, Ir, Rh or Cr ions (Fe/M:TiO2), the photocatalytic activities were enhanced compared to the Fe/TiO2 as shown in Figure 5A. The visible light response was increased in the order Ir > Cr > Ru > Rh. In the case of the Fe/Ru:TiO2, based on the measurements by ESR spectroscopy, Ru ions were doped as tetravalent and played a role as an acceptor level. Then, the photoexcited Ru ions, by receiving electrons from the VB, immediately transfer electrons to the grafted Fe3+ under visible light irradiation. This indirect electron transfer from the VB to the Fe3+ via the doped Ru ions occurred in addition to the direct electron transfer from the VB to the Fe3+, leading to the enhancement of photocatalytic activity.

Figure 5

(A) Photocatalytic activities of the Fe/M:TiO2 against gaseous acetaldehyde under visible light irradiation of (a) λ = 470 nm and (b) λ = 625 nm; (B) The amount of (a) •O2− and (b) H2O2 generated on the M:TiO2 and the Fe/M:TiO2, respectively under visible light irradiation of λ = 625 nm; (C) Schematic energy level diagram for the Fe/M:TiO2. Reprinted with permission from [25]. © 2014 Elsevier B. V.

Furthermore, we examined the desirable character of doped metal ions for photocatalytic performance in detail. Figure 5B shows the amounts of •O2− and H2O2 generated on the M:TiO2 and the Fe/M:TiO2, respectively, under visible light irradiation. The electrons having a potential energy of +0.38 V (vs. SHE at pH = 0) can reduce O2 into •O2− under the experimental conditions (pH = 11.5) [24]. Therefore, by measuring the amount of •O2−, the redox potential of dopants can be relatively estimated. For the non-doped TiO2, the generation amount of •O2− was larger than that of the M:TiO2. For the non-doped TiO2, O2 was reduced to •O2− by an electron excited at conduction band from defect level under visible light irradiation. This result indicated that all kinds of dopant used played a role as acceptor, because if they play a role as donor, the generation amount of •O2− should increase due to electron excitations from dopants to CB. Among the M:TiO2, the generation amount of •O2− was decreased in the order of Rh > Cr > Ru > Ir. Since the order of the redox level of dopants should be consistent with that of the •O2− amount, their redox levels would be more negative in the order of Rh > Cr > Ru > Ir. Secondly, when electrons transfer to the grafted Fe3+, the excited Fe2+ can reduce O2 to H2O2 by a two-electron process as mentioned below. Therefore, we can determine the degree of electron transfer to the Fe3+ from the dopants by the measurement of amount of H2O2. The generation amount of H2O2 was decreased in the order of Ir > Ru > Cr > Rh (Figure 5B). This H2O2 generation tendency was opposite to that of •O2−. This means that through the dopant with more positive redox potential, electrons can transfer more easily to the Fe3+ as illustrated in Figure 5C. This is due to the small energy loss of electrons when the redox potential of dopant was close to that of the Fe3+. Therefore, we concluded that the high photocatalytic activity of the Fe/Ir:TiO2 under visible light irradiation was attributable to the acceptor level due to Ir4+ formed close to the redox potential of the grafted Fe3+ (Figure 5C).

In the case of TiO2 codoped with Rh and Sb ions, the efficiency of the indirect electron transfer to the Fe3+ was lowered compared to the TiO2 doped with Rh ions alone (Figure 6A). By codoping with Sb ions, Rh4+ was reduced to Rh3+ and the formed Rh3+ played a role as donor [26,27]. This indicated that the efficiency of the indirect electron transfer of Rh3+ → CB → Fe3+ was lower than that of VB → Rh4+ → Fe3+ (Figure 6B).

Figure 6

(A) The co-doped effect on photocatalytic activities for CO2 formation from gaseous acetaldehyde under visible light irradiation of λ = 470 nm. (a) Before and (b) after grafting of Fe3+; (B) Indirect electron transfer paths to the Fe3+ via (a) donor level and (b) acceptor level. Reprinted with permission from [25]. © 2014 Elsevier B. V.

The energy gap of the redox potential between conduction band and the grafted Fe3+ was larger than that between the doped Rh4+ and the grafted Fe3+ and therefore electrons photoexcited at CB could not effectively transfer to the Fe3+ because of the large energy loss. From these results, forming acceptor level closed to the redox potential of the grafted Fe3+ was important for high performance of Fe/TiO2 based photocatalysts under visible light irradiation [25]. Recently, a better energy level matching in Fe/M:TiO2 was achieved by employing Fe3+ as a doping metal ion [28].

3.1.2. Pt Chloride Deposited TiO2 Photocatalysts

Pt4+ chloride deposited TiO2 (PtCl/TiO2) also showed a photocatalytic activity under visible light irradiation and its quantum efficiency was 9.8% [20]. In the past, Kisch et al., reported a mechanistic hypothesis to explain PtCl/TiO2 activity [21,29]. The proposed hypothesis was that the PtCl undergoes homolytic Pt-Cl cleavage by absorbing of light, generating a Pt3+ intermediate and a chlorine atom, the Pt3+ injects an electron to the conduction band of TiO2, and then the Cl radical oxidizes organic compounds. However, it is not clear whether the Pt-Cl cleavage in the PtCl/TiO2 system would occur. In addition, there is no sufficient evidence to support the injection of electron from Pt3+ to the conduction band of TiO2. Therefore, we clarified the charge transfer between the PtCl and TiO2 under visible-light irradiation using ESR spectroscopy [30].

For a bare TiO2 without deposition of PtCl, under visible light irradiation, both ESR signals assigned to trapped electrons and holes were not observed (Figure 7A(a)). For the PtCl/TiO2, a signal assigned to Pt3+ was observed. This means that Pt4+ chloride complexes were charge-separated into Pt3+ and Cl radicals. Then in the TiO2 host, trapped electrons (g ≈ 1.98) were observed (Figure 7A(b)). These results proved that TiO2 could receive electrons from excited Pt3+ as well as the hypothetical mechanism. However, unlike the hypothetical mechanism, trapped hole signal (g = 2.01) was also observed. Based on the results, some electrons in the VB of TiO2 would be excited to the orbital of the Cl radicals similarly to the case of direct electron transfer from the VB of TiO2 to the grafted Fe3+ for the Fe/TiO2 photocatalysts. Since the redox potential (+3.0 V vs. SHE at pH = 0) of the VB of rutile TiO2 is more positive than that (2.47 V vs. SHE at pH = 0) of Cl/Cl− [31], the high photocatalytic activity of PtCl/TiO2 would be owing to the generation of holes in the TiO2 host.

Figure 7

(A) ESR spectra of electrons and holes for (a) TiO2 and (b) PtCl/TiO2; (B) (a) •O2− and (b) H2O2 generated on the PtCl/TiO2 under visible light irradiation. (C) Photocatalytic reaction mechanism of PtCl/TiO2. Reprinted with permission from [30]. © 2012 American Chemical Society.

Generation behaviors of •O2− and H2O2 under visible light irradiation were also examined for the PtCl/TiO2 as shown in Figure 6B. •O2− was predominantly generated compared to H2O2. This was opposite behavior to the Fe/TiO2 for which H2O2 was dominantly generated rather than •O2−. This means that photoexcited electrons have a higher potential energy than +0.38 V (vs. SHE at pH = 0) and therefore the reduction of O2 to •O2− was produced by the electrons photoexcited at CB of TiO2, which supported the ESR results [30].

A plausible reaction mechanism for the PtCl/TiO2 photocatalyst is illustrated in Figure 7C. Photoexcited Pt3+ generated by ligand- metal charge transfer in deposited PtCl complex by adsorption of visible light gives an electron to the TiO2 CB and then the electron is consumed by reduction of O2 into •O2−. Some of the photogenerated Cl radicals can decompose organic substances and the other receive electrons by photo-excitation from the valence band of TiO2, resulting the generation of holes in TiO2. The organic substances can be efficiently decomposed by the generated holes in TiO2 with strong oxidation ability.

3.1.3. Comparison of the Visible-Light Responsive TiO2 Photocatalysts

The reaction mechanisms of various modified TiO2 were investigated by detecting •OH quantitatively by means of a coumarin fluorescence probe method [18]. The photocatalysts investigated were nitrogen-doped, Fe(III)-grafted, Fe(III)-grafted Ru-dopedTiO2, and Pt-complex-deposited, whose diffuse reflectance spectra are shown in Figure 8A. On the irradiation with 470 nm light in the presence of coumarin, the concentration of umbelliferone was increased (Figure 8B). From the slope, the formation rate of •OH was calculated. Then, the •OH quantum yield was calculated with the absorbed light intensity which was evaluated from the absorption and irradiance spectra in Figure 8A. The quantum yield ranged from 10−5 for N-TiO2 to 4 × 10−4 for Fe/TiO2 [18]. In the presence of 0.14 mM H2O2, the •OH yield decreased for N-TiO2 while it increased for Fe/TiO2. The increase for Fe/TiO2 suggests that H2O2 is a reaction intermediate for producing •OH.

Figure 8

(A) Absorption spectra as the complement of the reflectance (1-R) and the irradiance spectra of LED used in the study; (B) Concentration of umbelliferone generated under 470-nm irradiation in aqueous coumarin solution was plotted as a function of the irradiation time. Reprinted with permission from [18]. © 2013 American Chemical Society.

The photocatalytic activity was evaluated by the rate of CO2 generation associated with acetaldehyde decomposition and then it was plotted in Figure 9A as a function of the •OH formation rate for each photocatalyst. The CO2 generation rates of the photocatalysts were positively correlated with those of the •OH formation. However, the formation rates of CO2 were extremely larger (103 times) than those of •OH. This finding indicates that the oxidation reaction predominantly takes place at the photocatalyst surface with the trapped holes. The good correlation in the figure suggests that •OH in the bulk solution is equilibrated with trapped holes (Equation (1)), but the equilibrium is significantly shifted to the surface trapped holes. •OH + Ti

Figure 9

(A) Relationship between the formation rates of CO2 and •OH under the irradiation of 470 nm LED. The CO2 formation rate is a measure of the photocatalytic reaction rate in the acetaldehyde decomposition in aqueous suspension system; (B) Schematic illustration of reaction mechanism of Fe(III)-grafted Ru-doped TiO2 photocatalyst based on the detection of •OH and CO2. Key: IFCT, interfacial charge transfer; h+tr, surface trapped hole. Reprinted with permission from [18]. © 2013 American Chemical Society.

The highest photocatalytic activity in the suspension system was obtained for Fe(III)-deposited Ru-doped TiO2 (Fe(III)/Ru:TiO2) whose reaction mechanism is shown in Figure 9B. On the basis of the ESR and chemiluminescence experiments mentioned above [23], the CB electrons are formed by two step excitation with visible light irradiation and O2 is reduced to H2O2. The grafted Fe3+ is reduced by Ru3+ or by IFCT and then the formed Fe2+ produces •OH from H2O2. The •OH produced in solution is adsorbed on the TiO2 surface to form trapped holes which could oxidize organic compounds, such as acetaldehyde, leading to CO2.

3.2. Photocatalysis with Bare TiO2

3.2.1. Reactivities of Rutile and Anatase Surfaces

The photogeneration of molecular oxygen at rutile TiO2 electrode is a famous historical reaction [32]. To investigate the oxidation mechanism, •OH formation was measured by employing three electrodes of rutile TiO2 (100), (110), and (001) [33]. Figure 10A shows the amount of the produced umbelliferone, which is normalized to the number of charges used in the reaction. For all electrodes, the photocurrent efficiency of •OH was less than 1%, while that of O2 was about 100%. This observation implies that the conventionally proposed mechanism to produce O2 via •OH formation is not a major mechanism in water oxidation at TiO2 surface. Figure 10B shows the plausible reaction steps in the formation of O2 and •OH through surface peroxo (Ti-O-O-Ti). By cleaving Ti-O bond in the peroxo, O2 is formed as shown in Figure 10B(a) [34]. When O-O bond is cleaved instead of Ti-O bond, •OH is formed as a byproduct. The •OH formation in Figure 10A increases in the order of (001) < (110) < (100), which can be explained by the strength of Ti-O bond deduced from the surface structure [33].

Figure 10

(A) The amount of produced umbelliferone normalized for the current charge is plotted against the coumarin concentration in solution for the rutile (100), (110), and (001) TiO2 electrodes; (B) Plausible reaction steps starting from peroxo to form (a) O2 and (b) •OH at the TiO2 surface. Reprinted with permission from [33]. © 2013 American Chemical Society.

Though rutile TiO2 shows high activity for O2 evolution, anatase TiO2 is known to have a higher activity in the photocatalytic oxidation of organic molecules [35]. The difference in the generation process of •OH between rutile and anatase was investigated by using CCA and coumarin [36]. Figure 11A shows the quantum yields of •OH generation measured with coumarin and CCA together with the adsorbed fraction of CCA. As shown in Figure 11A, anatase and anatase-contained TiO2 (ST-01, P25, and F1) generated •OH in the substantial yields. The quantum yield for OH-CCA was much larger than that for umbelliferone, indicating that •OH is formed at the TiO2 surface and diffused into bulk solution. Furthermore, this observation indicates that the active site is different from the adsorption site of –COO− group. Since H2O2 is produced in photocatalysis, the effect of H2O2 on the •OH generation was investigated. Figure 11B shows the effect of the addition of H2O2 on the formation rate for (a) OH-CCA and (b) umbelliferone. On the addition of H2O2, the •OH generation for pure anatase TiO2 decreased but increased for rutile and rutile-contained TiO2. This phenomenon has been reported previously for other several TiO2 powders [37]. The amount of •O2− was significantly increased with the addition of H2O2 [37]. Although the formation of •OH from H2O2 by CB electrons is commonly suggested in the •OH generation mechanism, the fact that the •O2− was significantly increased with H2O2 denied the one-electron reduction of H2O2.

Figure 11

(A) Quantum yields of •OH using the different probe molecules, coumarin (blue) and CCA (brown), and the fraction of adsorbed CCA (green) for four kinds of TiO2 powders; (B) Effect of the addition of 0.14 mM H2O2 on the formation rates of (a) OH-CCA and (b) umbelliferone. Reprinted with permission from [36]. © 2014 American Chemical Society.

Since the increase is remarkable for anatase than rutile TiO2, with the addition of H2O2, the generation of •OH at anatase surface was replaced by the oxidation of H2O2 to form •O2−, as illustrated in Figure 12A.

Figure 12

(A) Photocatalytic processes at the conduction band (C.B.) and the valence band (V.B.) of TiO2 with anatase and rutile crystalline types in the absence and the presence of H2O2. The thickness of arrows expresses the degree of the reaction rate; (B) Plausible mechanisms of •OH generation at anatase TiO2 (upper part) and rutile TiO2 (lower part). Reprinted with permission from [36]. © 2014 American Chemical Society.

For rutile TiO2 adsorbed H2O2 is equivalent to the surface peroxo, Ti-O-O-Ti and promotes the formation of •OH as stated above. The detailed generation mechanism of •OH on anatase and rutile TiO2 surfaces can be proposed as shown in Figure 12B. On the anatase surface, photogenerated valence band holes, h+, are trapped at the surface oxygen to form trapped holes (Ti−O•) that can be regarded as the adsorbed •OH in the deprotonated form (•O−) [18] then an •OH is released to the solution as represented by Equation (1). On the other hand, for rutile TiO2, since the crystalline structure is packed more tightly than that for anatase, the stability of the surface trapped holes may be different. By trapping of h+ predominantly near the trapped hole, Ti-peroxo is formed. As described above in Figure 10B(b), •OH radical is produced by h+ from H2O with Ti-peroxo, which plays the role of a catalyst. Thus, the increase of the •OH generation with H2O2 for rutile TiO2 can be explained.

3.2.2. Adsorption and Decomposition of Glycine Related Peptides

The application of photocatalysts to biological fields for their antibacterial effect and in medical treatments for diseases, including cancer, has been proceeding extensively [38,39]. It is believed that the active oxygen species generated on the photocatalysts such as H2O2, •OH, and singlet oxygen are involved in the attack to kill various kinds of virus and bacteria [40]. However, the mechanism underlying the photobiological activity is not yet well understood. Since the photocatalytic process is expected to occur at the interface between the photocatalysts and the liquid medium, the interface between protein molecules and inorganic materials has recently received much attention.

Proteins and peptides are composed of various kinds of amino acids. For a proper understanding of the adsorptive and photocatalytic interactions between the surface of the photocatalysts and proteins/peptides, fundamental knowledge on the adsorption and photocatalytic reactivity of individual constituent amino acids would be necessary.

TiO2 is widely used for practical applications as a photocatalyst. The surface of TiO2 is amphiphilic, which consists of hydrophobic and hydrophilic parts [41]. The hydrophilic parts involve two kinds of hydroxyl group, that is, the acidic bridged hydroxyl group and the basic terminal hydroxyl group. Both groups can be adsorptive and/or photocatalytic active sites, depending on the kinds of titanium dioxides which are characterized by different particle size, surface area, and crystal forms such as anatase, rutile and brookite. The photocatalyst with different characteristic surface shows different adsorbability and photocatalytic activity [40].

It was demonstrated that both hydrophilic and hydrophobic sites are adsorptive sites but that only hydrophobic sites are photocatalytically active for ST-01 TiO2 (100% anatase crystal form with a BET surface area of 320 m2·g−1 and a particle size of 9 nm; Ishihara Sangyo Ltd., Osaka, Japan) [15]. After the calcinations at 973 K hydrophilic parts of the surface of ST-01 can be eliminated and a highly hydrophobic surface (designated as HT-TiO2) is created without changing the crystal form [15]. By employing these characteristics, the adsorption and decomposition sites of the simplest amino acid glycine, whose adsorbability on the TiO2 surface is still controversial [42], and its homopeptides (Gly-Gly and Gly-Gly-Gly) were investigated by 1H-NMR spectroscopy [43]. For Gly-Gly and Gly-Gly-Gly the carboxylic group and the peptide bond were assigned as the adsorptive sites of the peptides on the surface of ST-01. The adsorption feature of Gly-Gly-Gly on TiO2 (ST-01) are illustrated in Figure 13; the peptide would adsorb by the C-terminal carboxyl group most probably with the terminal hydroxyl group at 5-coordinated Ti of TiO2 as is generally believed [42]. The photo decomposition took place by the weak adsorption of the peptide bonds on the surface of TiO2 (ST-01).

Figure 13

Schematic presentation of the adsorption of the peptide (Gly-Gly-Gly) on the hydrophilic and hydrophobic surface of TiO2 (ST-01). Reprinted with permission from [43] © 2014 American Chemical Society.

On the other hand, when a hydrophobic side chain Leu is incorporated, in addition to the peptide bonds and the carboxylic group, the adsorption of hydrophobic leucyl residue on the hydrophobic parts of TiO2 surface would take place. As shown in Figure 14, for HT-TiO2, the adsorption of the Leu- containing peptides increased with the increase of the number of the peptide bond that is, Leu < Leu-Gly, Gly-Leu < Leu-Gly-Gly (Figure 14C).

Figure 14

Equilibrium adsorption and rate of decomposition under the UV irradiation measured at 297 K for Leu, Leu-Gly, Gly-Leu, and Leu-Gly-Gly in the aqueous suspensions of TiO2; (A,B) for untreated TiO2 and (C,D) for HT-TiO2 (TiO2 calcined at 973 K). Reprinted with permission from [43] © 2014 American Chemical Society.

However, the decomposition rates are almost the same (Figure 14D). These facts suggest that both the peptide bond and leucyl side chain could adsorb on the hydrophobic surface of TiO2 but photocatalytic decomposition should take place through the adsorption of the leucyl side chain which would adsorb preferably on the photocatalytic active part of the hydrophobic TiO2 surface. Thus leucyl residue would adsorb preferably on the active site of the hydrophobic part of TiO2 instead of the peptide bonds and photocatalysis proceeds. The adsorption feature of Leu-Gly-Gly, on TiO2 (ST-01) are illustrated in Figure 15.

Figure 15

Schematic presentation of the plausible adsorption of the peptide (Leu-Gly-Gly) on the hydrophilic and hydrophobic surface of TiO2 (ST-01). Reprinted with permission from [43] © 2014 American Chemical Society.

However, as shown in Figure 14A,B it was found that Leu-Gly showed remarkably low adsorbability and decomposition rate as compared to Gly-Leu due to the specific conformation, in which the positively charged amino group and negatively charged carboxyl group interact strongly by electrostatic force [43]. Thus, when a peptide or proteins take a specific conformation, photocatalysis does not work effectively. For the effective use of TiO2 it would be necessary to acquire information on the surface conformation of the corresponding proteins/peptides to access the surface of the photocatalysts. By combining the information about the surface conformation of proteins/peptides and the active sites of TiO2 (hydrophobic or hydrophilic), we could design the TiO2 effective to diminish the specific virus, bacteria or environmental hazardous materials.

3.2.3. Glutathione and Related Amino Acids

With increased applications of TiO2 nanoparticles, the concerns about their potential human toxicity and their environmental impact have also increased. Although details of human biological responses to TiO2 exposure are still unavailable, numerous in vitro examinations concerning cellular responses induced by TiO2 have been reported [44,45,46].

Glutathione is a tri-peptide capable of diminishing active oxygen species in living cells. In spite of the importance of glutathione in defense against oxidative stress, its actual affects and the mechanism for the TiO2-induced cytotoxicity and genotoxicity have not been completely elucidated yet.

The photocatalytic decomposition of glutathione and related amino acids in TiO2 suspension was investigated with 1H NMR spectroscopy [47]. The results suggest, that as shown in Figure 16A, both glutathione in reduced (GSH) and oxidative forms (GSSG) are adsorbed on the TiO2 surface by carboxyl or amino groups but not by the thiol group (SH) of the side chain which plays a crucial role in the glutathione cycle (Scheme 1), to be degraded. This suggests that the function of glutathione cycle should be deteriorated in living cells by the adsorption. However, the decomposition rates are considerably slow as compared with those of the constituent amino acids (Glu, Cys and Gly) as shown in Figure 16B, possibly reflecting the self-defensive property against active oxygen species.

Figure 16

(A) Schematic presentation of the plausible adsorption of glutathione on the surface of TiO2 (ST-01); (B) (a) Adsorption and (b) decomposition rates of glutathione (GSH and GSSG) and the constituent amino acids (Glu, Cys, and Gly) in aqueous suspension of TiO2 (5 mg/0.4 mL D2O) under UV irradiation at 297 K. Reprinted with permission from [47] © 2012 American Chemical Society.

Scheme 1

Glutathione cycle.

4. Conclusions

TiO2 photocatalysts have been utilized for the oxidation of organic pollutants. For the development of further practical applications, the improvement of the activity with the aid of an understanding of the detailed mechanism(s) of action is a prerequisite. The primary process of photocatalysis reported in the literatures still have some confusion. To clarify the reaction mechanism, the proper and reliable detection of primary active species, such as trapped electrons, trapped holes, •O2− and •OH, in photocatalytic systems is required. By employing various spectroscopic techniques we have succeeded in elucidating some of the mechanisms of important photocatalytic reactions. Further investigations are proceeding in our laboratory.