Double-heterojunction structure of Sb_xSn_1-xO₂/TiO₂/CdSe for efficient decomposition of gaseous 2-propanol under visible-light irradiation

Abstract

Highly crystallized antimony-doped tin oxide (ATO; Sb_xSn_1-xO₂, x = 0.1) of ∼50 nm size was prepared by co-precipitation of SnCl₄·5H₂O and SbCl₃, followed by heat-treatment at 1000 °C. The prepared ATO nanoparticles of deep blue color revealed a profound light-absorption in the visible range. ATO/TiO₂ composites were prepared by covering the surface of ATO nanoparticles with TiO₂ using the sol–gel method. Under visible-light irradiation (λ ≥ 420 nm), the prepared ATO/TiO₂ showed a notable photocatalytic efficiency in decomposing gaseous 2-propanol (IP), which seemed to be caused by the hole-transfer mechanism between the valence bands (VB) of ATO and TiO₂, since the ATO's VB level is located lower than that of TiO₂. Subsequently, a double-heterojunction ATO/TiO₂/CdSe structure was prepared by loading CdSe quantum dots (QDs) onto the surface of the ATO/TiO₂, which dramatically enhanced the visible-light photocatalytic efficiency. In fact, the catalytic activity of ATO/TiO₂/CdSe in evolving CO₂ from IP, was ∼3 times that of ATO/TiO₂ and twice that of typical N-doped TiO₂. The unexpectedly high efficiency of ATO/TiO₂/CdSe seemingly is due to the unique band matching among these semiconductors. With sensitization of ATO and CdSe, not only the holes but also the electrons are generated in the VB and CB, respectively, of TiO₂ under visible-light irradiation.

---

1. Introduction

Photocatalytic decomposition of organic pollutants has attracted widespread attention over the last few decades. TiO₂ is known as the most efficient photocatalyst among the various semiconductors, due to its unique characteristics in band position, surface structure, and electron mobility, as well as its extended chemical stability and non-toxicity.^1–6 However, TiO₂, due to its wide band gap (E_g = 3.2 eV), can utilize only the photons in the UV region (λ < 380 nm), which limits its commercial applications in sunlight or indoors.^7–11 Hence, in order to be able to utilize the major portion of the solar spectrum, development of photocatalysts that are functional under visible-light is indispensable.

Thus far, several design strategies for visible-light photocatalysts have been pursued, among which are lowering of the TiO₂ conduction band (CB) level by doping transition metal ions,^12–17 and elevation of the valence band (VB) level by substituting anions such as N, C, S, or B for the oxygen sites in TiO₂.^18–22 Recently, utilization of charge transfer reaction by grafting metal ions onto the TiO₂ surface has been reported,^23–25 and it seems to be a noticeable strategy in designing visible-light photocatalysts. Significant progress has been made thus far, but for the purpose of practical applications, the visible-light photocatalytic efficiency must be further improved.

Another promising approach is the formation of a heterojunction structure between the sensitizer and the main photocatalyst. According to the relative energy band location there, two kinds of heterojunctions can be designed. First, as outlined in Scheme 1(a) (the Type-A heterojunction), the CB of sensitizer semiconductor is positioned higher than that of the main photocatalyst, in the present case TiO₂. The Type-A heterojunction can be formed by loading, for example, metal chalcogenide quantum dots (QDs) or inorganic dyes onto the TiO₂ surface. Under visible-light irradiation, the sensitizer-A is excited, and the electrons are then transported to the TiO₂ CB. These electrons can induce various reduction reactions and participate in decoloration reactions of organic dyes,^26–30 but CO₂ evolution resulting from complete oxidation of organic pollutants is difficult, due to the unavailability of the OH˙ radical on the TiO₂ surface. Serpone et al. reported that complete decomposition also is possible with CdS/TiO₂,³¹ but the efficiency is low, because the photocatalytic reaction occurs via the following complicated steps.

Scheme 1 (a) Schematic diagram of Type-A heterojunction. Under visible-light irradiation, sensitizer-A is excited, and the photo-excited electrons are injected into the CB of TiO₂. (b) Schematic diagram of Type-B heterojunction. Under visible-light irradiation, sensitizer-B is excited, and the induced holes are transported to the VB of TiO₂. (c) Schematic diagram of double-heterojunction. Under visible-light irradiation, both the sensitizer-A and B are excited. As a result, electrons and holes are generated in the TiO₂ CB and VB, respectively.

CdS + hν → CdS(e⁻_CB + h⁺_VB) (1)

CdS(e⁻_CB + h⁺_VB) + TiO₂ → CdS(h⁺_VB) + TiO₂(e⁻_CB) (2)

TiO₂(e⁻_CB) + O₂ → TiO₂ + O^·−₂ (3)

O^·−₂ + H⁺ → HO₂˙ (4)

HO₂˙ + Organic_(ads) → Intermediates → Products (5)

In the second heterojunction type (Type-B heterojunction), the VB of the sensitizer, as described in Scheme 1(b), is located lower than that of the main photocatalyst such as TiO₂. Under visible-light irradiation, the electrons in the VB of the sensitizer are excited to its CB. Thereby, the VB of the sensitizer is rendered partially vacant, and the electrons in the VB of TiO₂ can be transferred to that of the sensitizer. As a result, holes are generated in the VB of TiO₂, initiating in turn various oxidation reactions. Considering the powerful oxidative ability of the holes in the VB of TiO₂, efficient and complete oxidation of organic compounds is expected with this Type-B heterojunction structure under visible-light. However, only a few systems utilizing hole transfer between the sensitizer and TiO₂ have been reported, owing to the scarcity of appropriate sensitizers. Thus far, a few Type-B heterojunction systems, WO₃/TiO₂,³² FeTiO₃/TiO₂,³³ FeOOH/TiO₂,³⁴ Bi₂O₃/BiOCl,³⁵ and Bi₃O₄Cl/BiOCl³⁶ have been reported to be effective under visible-light irradiation. Notably, it was found that the FeTiO₃/TiO₂ heterojunction, formed by an FeTiO₃ nanodisk and TiO₂ nanoparticles, exhibited a photocatalytic efficiency comparable to that of typical N-doped TiO₂.^33,37

VB position of SnO₂ (+ 3.6 eV vs. NHE) has been reported to be considerably lower than that of TiO₂ (+ 2.9 eV vs. NHE), but its wide bandgap (E_g), 3.5 eV, does not allow utilization of visible-light.^31,38,39 In the present work, we doped Sb ions into SnO₂ to form antimony-doped tin oxide (ATO), which absorbs a considerable portion of visible-light with its deep-blue color. We found that ATO is an effective sensitizer for use with TiO₂ in forming the Type-B heterojunction structure with high visible-light photocatalytic activity.

For further enhancement of visible-light catalytic efficiency, we doubly combined the two different sensitizers with TiO₂. That is, ATO (sensitizer-B) was designed to be located in the core of the TiO₂ structure, whereas CdSe (sensitizer-A) was loaded onto the TiO₂ surface. According to Scheme 1(c), both sensitizers are excited by visible-light irradiation, and resultantly the electrons and holes are transported to the CB and VB, respectively, of the TiO₂. These electrons and holes generated on TiO₂ are expected to enhance photocatalytic oxidation reactions, since not only the holes but also the electrons are considered to be essential for the efficiency of those oxidation reactions. In the present work, the photocatalytic behavior of ATO/TiO₂/CdSe and the mechanistic roles of ATO and CdSe in enhancing its photocatalytic oxidation reactions were investigated.

2. Experimental section

2.1. Preparation of ATO

The ATO nanoparticles, used in forming the heterojunction with TiO₂, were synthesized by the co-precipitation and subsequent heat-treatment.^40–43 In a typical synthesis of 10 mol% Sb-doped ATO particles, 9 mmol of tin(IV) chloride pentahydrate (SnCl₄·5H₂O, Aldrich) and a stoichiometric amount of antimony(III) chloride (SbCl₃, Aldrich) were dissolved in 50 ml of anhydrous ethanol. With vigorous stirring, tetrabutyl ammonium hydroxide (TBAH, 40 wt% water solution, Aldrich) was added dropwise, until the pH of solution was adjusted to 7.5. In the course of the stirring, Sn and Sb species were gradually co-precipitated, and the collected precipitate was washed with de-ionized water several times. It was dried in air at 60 °C for 12 h, and then calcined at 1000 °C for 2 h.

2.2. Preparation of ATO/TiO₂ composites

A total of 3.67 g titanium isopropoxide (97%, Aldrich) was stabilized in a mixed solution of 40 ml ethanol, 1 ml concentrated nitric acid, and 1 ml water. Then the prepared Ti-precursor solution was gently stirred for 4 h. A stoichiometric amount of 10 mol% Sb-doped ATO particles was then added to this solution. In order to obtain 5/95 ATO/TiO₂ composite (in wt% ratio), 53 mg 10 mol% Sb-doped ATO particles pre-suspended in 10 ml ethanol was added and gently stirred overnight. The sample was then dried at 80 °C for 24 h, and subsequently heat-treated at 250 °C for 2 h. As a blank sample, the bare TiO₂ was prepared by the same procedure, but without adding ATO.

2.3. Preparation of cadmium selenide (CdSe) and its surface modification

Cadmium selenide (CdSe) quantum dots (QDs) were synthesized in noncoordinating solvent octadecene (ODE, Aldrich) with the addition of trioctylphosphine oxide (TOPO, Aldrich) as a capping agent.⁴⁴ Typically, 0.0154 g CdO, 0.339 g oleic acid (OA), 2.72 g ODE, and 0.928 g TOPO were mixed in a three-neck flask, and heated to 300 °C under an anhydrous N₂ flow. A solution consisting of 0.0190 g Se, 0.356 g trioctylphosphine (TOP, Aldrich) and 1.625 g ODE in a beaker was quickly added to the heated solution. The temperature of the mixture was then adjusted to 260 °C so as to induce the growth of CdSe nanocrystals. After 5 min, the reaction was quenched by adding 30 ml of cold toluene.

The TOPO capping the CdSe QD was exchanged for mercaptopropionic acid (MPA, Aldrich) following the procedure reported by Leschkies et al.,⁴⁵ in order to allow for CdSe QD attachment to the TiO₂ surface. According to the typical procedure, 0.2 mmol MPA was dissolved in 100 mL anhydrous methanol, the pH of which solution was adjusted to 11.4 by adding tetramethylammonium hydroxide (TMAH, Aldrich). 40 mg TOPO-capped CdSe was then added this solution. The mixed solution was heated to 63 °C under Ar atmosphere, and refluxed for ∼24 h in darkness. The formed MPA-capped CdSe QD was precipitated by adding a mixture of ethyl acetate and diethyl ether (1 : 1 in volume). The collected precipitate was washed several times with ethyl acetate to remove residual MPA or TOPO. The prepared CdSe is then readily soluble in protic solvents.

2.4. Preparation of ATO/TiO₂/CdSe composites

Into a 50 ml ethanol suspension containing 1.00 g 5/95 ATO/TiO₂ composite, stoichiometric amounts of the MPA-capped CdSe were dispersed, and the resultant solution was stirred at 60 °C for 6 h. In order to obtain, for example, 1 wt% CdSe-loaded ATO/TiO₂ composite, 10.1 mg MPA-capped CdSe was added. It is known that during the stirring the MPA-capped CdSe QD can bind to the surface of the ATO/TiO₂, due to the high binding affinity of the carboxylic group for TiO₂. The formed ATO/TiO₂/CdSe composites were then precipitated by centrifugation, and the collected particles were heated at 250 °C for 2 h.

2.5. Characterizations

X-Ray diffraction (XRD) patterns of the prepared photocatalytic samples were obtained using a Rigaku Multiflex diffractometer with monochromatic light-intensity Cu Kα radiation. XRD scanning was performed under ambient conditions over 2θ region of 20–80° at a rate of 2°/min (40 kV, 20 mA). UV-visible diffuse reflectance spectra were acquired by a Perkin-Elmer Lambda 40 spectrophotometer. BaSO₄ was used as the reflectance standard. Transmission electron microscopy (TEM) and energy dispersive X-ray spectroscopy (EDX)-mapping images were obtained by a JEOL JEM2100F operated at 200 kV. One milligram of the synthesized particles was dispersed in 50 mL of ethanol, and a drop of the suspension was then spread on a holey amorphous carbon film deposited on the copper grid.

2.6. Evaluation of photocatalytic activity

The visible-light photocatalytic efficiencies of the photocatalytic samples were estimated by monitoring the decomposition of gaseous 2-propanol (IP). An aqueous suspension containing 8.0 mg of the photocatalytic samples was spread on a 2.5 × 2.5 cm² Pyrex glass in film form and subsequently dried at room temperature. Preparatory to the photocatalytic measurement, the prepared films were irradiated by a 300 W Xe lamp for 3 h, in order to remove possible organic residues. The gas reactor system used to measure the photocatalytic activity has been described elsewhere.⁴⁶ The net volume of the gas-tight reactor was 200 mL, and the photocatalytic film was located at the center of the reactor. The entire area of the photocatalytic film (2.5 × 2.5 cm²) was irradiated by a 300 W Xe lamp through a UV cut-off filter (λ < 420 nm, Oriel) and a water filter to cut-off the IR spectrum. After evacuation of the reactor, 1.6 μL of the water-diluted IP (IP : H₂O = 1 : 9 in volume) was injected into the reactor. The initial concentration of gaseous IP in the reactor was maintained at 117 ppm in volume (ppmv). Thus the ultimate concentration of CO₂ evolved, with all of the IP decomposed, would be 351 ppmv, as shown in the following equation:

2 (CH₃)₂CHOH (g) + 9 O₂ (g) → 6 CO₂ (g) + 8 H₂O (g) (6)

The total pressure of the reactor was then increased to 750 Torr by the addition of oxygen gas. Under these conditions, the IP and H₂O remained in the vapor phase. After a certain duration of irradiation, 0.5 mL of the gas in the reactor was automatically collected and sent to a gas chromatograph (Agilent Technologies, Model 6890N) using an auto sampling valve system. For CO₂ detection, a methanizer was installed between the GC column outlet and the FID detector.

3. Results and discussion

ATO nanoparticles at different Sb-doping levels were prepared by co-precipitation and subsequent 2 h heat-treatment at 1000 °C. As the Sb-doping level was increased, the bluish color of the ATO particles grew bolder and deeper. This suggests that a higher doping level would induce more efficient utilization of visible light, but in fact it was found from the X-ray diffraction patterns shown in Fig. 1(a) that only a limited amount of Sb ion could be doped into the SnO₂ lattice. That is, the pure cassiterite SnO₂ phase (JCPDS, No. 77-0450) could be maintained only up to 10 mol% of Sb incorporation, whereas further incorporation led to the formation of impurity peaks, suggesting that extra Sb species were segregated to form the Sb₂O₄ cervantite phase (JCPDS, No. 77-0127). This is consistent with previous observations.^47–49 The average crystallite size of the 10 mol% Sb-doped ATO particles, as determined from the (110) and (101) peaks by application of the Scherrer equation, was ∼39 nm, strongly indicating the highly crystallized structure.

Fig. 1 (a) XRD patterns of SnO₂ and several ATO nanoparticles of different compositions. (b) XRD patterns of CdSe, TiO₂, ATO (Sb_xSn_1-xO₂, x = 0.1), 5/95 ATO/TiO₂, and ATO/TiO₂/CdSe (CdSe: 1 wt%) composites.

Fig. 2(a) shows the 10 mol% Sb-doped ATO powder exhibiting a deep blue color. This color is associated with ATO's internal energy bands formed between the CB and VB of SnO₂. In the ATO lattice, the oxidation state of the Sb ions substituting for Sn⁴⁺ will be 5+ and 3+. Hence, its chemical formula can be written Sn⁴⁺_0.9Sb⁵⁺_xSb³⁺_0.1-xO_1.95+x, where 0 < x < 0.1.⁵⁰ The presence of Sb⁵⁺ and Sb³⁺ ions produces additional energy bands originating from Sb 4d and 5p, below the SnO₂ CB. Thus the CB level is lowered considerably, whereas the VB level originating from O 2p is not appreciably affected. As a result, the bandgap of ATO is reported to be 2.55 eV or less,^51–53 which allows for visible light absorption.

Fig. 2 (a) Photographic image of ATO (Sb_xSn_1-xO₂, x = 0.1) powders. (b) TEM image of ATO nanoparticles. (c) High-resolution TEM image of a single ATO nanoparticle. (d) Magnification of dotted-rectangular area indicated in (c). (e) SAED pattern obtained from entire area of single ATO nanoparticle is shown in (c).

Fig. 2(b) and 2(c) show TEM images of 10 mol% Sb-doped ATO particles of ∼50 nm size. The particles are moderately monodispersed and mostly mutually separated, without significant agglomeration. The dotted rectangular area of the single ATO particle shown in Fig. 2(c) was further magnified (Fig. 2(d)). The TEM image reveals the fringe patterns uniformly formed over the entire area, suggesting that each ATO particle is a single crystal. The spacing of the fringe patterns was determined to be 3.35 Å, corresponding to d₁₁₀ of the cassiterite phase. Fig. 2(e) shows the selected area electron diffraction (SAED) pattern monitored over the entire area of Fig. 2(c)'s single particle. The SAED pattern, as is apparent, exhibited clear diffraction spots indexed to the tetragonal P4₂/mnm space group of the cassiterite phase, indicating that the prepared 10 mol% Sb-doped ATO nanoparticle was a single crystal without any impurity phases.

Fig. 3(a) shows the TEM image of the CdSe prepared in noncoordinating solvent ODE with the addition of TOPO.⁴³ The as-prepared CdSe nanoparticles were monodispersed and of 6–7 nm size. The high resolution TEM image in Fig. 3(b) shows a CdSe QD with uniform fringe patterns over its entire area. The fringe spacing of 3.51 Å was determined to be d₁₁₁ of the CdSe in the cubic F4̄3m space group. Fig. 1(b) shows the XRD patterns of the pure CdSe, TiO₂, ATO and ATO/TiO₂/CdSe composites. The observed (111), (220), and (311) XRD peaks confirm that the prepared CdSe QD is in the pure cubic phase (JCPDS, No. 88-2346). The sol–gel derived TiO₂, heat-treated at 250 °C, showed diffraction peaks at 25.31°, 37.79°, 48.05°, 55.07°, and 62.69°, corresponding to the (101), (004), (200), (211), and (204) peaks, respectively, of the anatase TiO₂ (JCPDS, No. 71-1166). The XRD patterns of the ATO/TiO₂/CdSe composites were a simple mixture of those of ATO, TiO₂ and CdSe, without other impurity peaks, suggesting that no chemical reaction had occurred among the ATO, TiO₂ and CdSe in the course of the heat-treatment at 250 °C.

Fig. 3 (a) TEM image of as-prepared CdSe QDs and (b) a high resolution image.

The TEM images in Fig. 4(a) show the 5/95 ATO/TiO₂ composites prepared by the sol–gel method. The ATO particles of ∼50 nm size are fully covered with TiO₂. Fig. 4(b) is a high-magnification image of the interface between TiO₂ and ATO. The sharp interface without any chasm suggests that the TiO₂ is tightly bound to the ATO particles.

Fig. 4 (a) TEM image of 5/95 ATO/TiO₂, and (b) high-magnification image of interface between ATO and TiO₂.

A double heterojunction structure of ATO/TiO₂/CdSe was formed by loading the CdSe QDs onto the surface of the 5/95 ATO/TiO₂ composite. Scheme 2(a) is a schematic diagram of the preparation of the structure. Fig. 5(a) and 5(b) shows TEM images of the 1 wt% CdSe-loaded ATO/TiO₂, clearly showing that the large ATO particles are located in the core and fully covered with less crystallized TiO₂. The high magnification image (Fig. 5(c)) for the dotted rectangular part of c in Fig. 5(a), exhibits that the small CdSe quantum dots are scattered on the surface of the TiO₂. Since the ATO particles were embedded in a large quantity of the sol–gel derived TiO₂, the CdSe QDs and ATO particles were not in contact each other. They were spatially separated by the TiO₂ layer.

Fig. 5 (a) TEM images of 1 wt% CdSe-loaded ATO/TiO₂ composite, and (b) and (c) are the high resolution images magnifying the dotted rectangular areas of b and c, respectively, shown in (a).

Scheme 2 (a) Formation scheme for ATO/TiO₂/CdSe double heterojunction system. (b) Energy band diagram indicating electron flows under visible-light irradiation. All of the potential levels versus NHE are indicated.

EDX-mapping images were obtained for the 1 wt% CdSe-loaded ATO/TiO₂ composite shown in the TEM image of Fig. 6(a). Fig. 6(b) shows an image of the Sn-mapping, clearly indicating that the dark particles embedded in the TiO₂ structure (Fig. 6(a)) are aggregated ATO. The Ti-mapping image shown in Fig. 6(c) was very close to the original TEM image. Fig. 6(d) shows an image of the Cd-mapping. The overall mapping signal is very weak, but the image is close to that of Ti, suggesting that the CdSe QDs are uniformly distributed over the entire surface of TiO₂ in ATO/TiO₂.

Fig. 6 (a) TEM image of ATO/TiO₂/CdSe (CdSe: 1 wt%) sample. EDX mapping images were obtained for the same area by monitoring the concentrations of (b) Sn, (c) Ti, and (d) Cd, respectively.

Fig. 7 shows the UV-visible absorption spectra for the TiO₂, CdSe, ATO, ATO/TiO₂ and ATO/TiO₂/CdSe composites. TiO₂, due to its wide band gap, showed a low absorption in the visible region, whereas ATO revealed a profound absorption over the entire visible region. As a result, the ATO/TiO₂ composites exhibited an appreciable absorption in the visible region. The characteristic absorption edge of CdSe QD, shown in the absorption spectra in Fig. 7, indicated that its band gap is 1.7 eV, which corresponds to the reported value of bulk CdSe.^54–56 As a result, the ATO/TiO₂/CdSe composite showed a significant absorption in the visible region, reflecting efficient utilization of visible light in the photocatalytic reactions.

Fig. 7 UV-visible absorption spectra of TiO₂, CdSe, ATO, 5/95 ATO/TiO₂, and ATO/TiO₂/CdSe (CdSe: 1 wt%) powders, taken at diffuse reflectance mode.

The photocatalytic activities of the CdSe/TiO₂, ATO/TiO₂ and ATO/TiO₂/CdSe composites in decomposing gaseous IP were evaluated under visible-light irradiation (λ ≥ 420 nm). The photocatalytic reaction proceeded through initial oxidation of IP to acetone and then ultimate mineralization to CO₂. Fig. 8(a) depicts the removal of the IP with several ATO/TiO₂ composites of different compositions under visible-light irradiation. The photocatalytic activities of several ATO/TiO₂ composites were remarkably higher than those of the end members such as the bare ATO nanoparticles or the blank TiO₂ sample. Among them, the 5/95 ATO/TiO₂ composite exhibited the highest catalytic activity. About 80% of IP was decomposed within 2 h of visible-light irradiation. When the ATO exceeded 5 wt%, however, the photocatalytic activity of the composites gradually decreased. The photocatalytic activity also was evaluated with respect to the evolution of CO₂ under visible-light irradiation. As shown in Fig. 8(b), the 5/95 ATO/TiO₂ composites showed the highest CO₂-evolving efficiency. The amount of CO₂ evolved in 2 h was 8.1 ppmv, whereas both the bare ATO and TiO₂ showed negligible evolution.

Fig. 8 Photocatalytic decomposition of gaseous IP with ATO, TiO₂, 1/99 CdSe/TiO₂ and several ATO/TiO₂ composites under visible-light irradiation. (a) Remnant IP percentage as function of irradiation time. (b) Amount of evolved CO₂ (ppmv).

For the purpose of comparison, several CdSe/TiO₂ samples were prepared by loading different amounts of CdSe QD onto the surface of TiO₂. Such are considered to be typical examples of the Type-A heterojunction structure, since the CB level of the CdSe is higher than that of TiO₂, as shown in Scheme 1(a). It was found that 1/99 CdSe/TiO₂ (wt% ratio) showed the optimum photocatalytic efficiency in decomposing IP and in evolving CO₂. As shown in Fig. 8(a) and (b), its efficiency was appreciably higher than that of the bare CdSe or TiO₂, but only ∼1/4 that of the 5/95 ATO/TiO₂ composites in evolving CO₂. The lower catalytic activity of CdSe/TiO₂ would be caused by the fact that the holes, the most active component in the photocatalytic oxidation reaction, cannot be generated in the VB of TiO₂. Under visible-light irradiation, the electrons of CdSe QDs are excited to its CB, and then transferred to the TiO₂ CB. These electrons are useful to the reduction reaction, but will not be used efficiently in the oxidation reactions such as IP decomposition.

For further enhancement of visible-light photocatalytic activity, we prepared a double heterojunction structure by loading CdSe QDs onto the surface of 5/95 ATO/TiO₂. Fig. 9(a) shows the photocatalytic activity of the CdSe-modified ATO/TiO₂ composites in decomposing gaseous IP under visible light irradiation. The composites demonstrated remarkably enhanced catalytic activity, compared with the bare 5/95 ATO/TiO₂. Among them, the 1.0 wt% CdSe loaded ATO/TiO₂ composite exhibited the highest photocatalytic activity: about 97% of 2-propanal was decomposed in 2 h of irradiation. The photocatalytic activity also was evaluated according to the amount of CO₂ evolved under visible-light irradiation, as shown in Fig. 9(b). The 1.0 wt% CdSe-loaded 5/95 ATO/TiO₂ composites demonstrated the highest efficiency in evolving CO₂: 23.2 ppmv after 2 h of irradiation, which is ∼3 times that of the bare 5/95 ATO/TiO₂ composite. Its efficiency, indeed, was ∼13 times and ∼2 times that of Degussa P25 and the typical N-doped TiO₂,^33,37 respectively.

Fig. 9 Photocatalytic decomposition of gaseous IP with ATO/TiO₂, several ATO/TiO₂/CdSe composites, and N-doped TiO₂ under visible-light irradiation. (a) Remnant IP percentage as function of irradiation time. (b) Amount of evolved CO₂ (ppmv). (c) Ultimate decomposition of IP with ATO/TiO₂/CdSe (CdSe: 1.0 wt%) under long-term irradiation.

Fig. 9(c) plots the trends of IP decomposition and CO₂ evolution with 1.0 wt% CdSe loaded ATO/TiO₂ composite under long-term visible-light irradiation. In 3 h, the IP, with an initial concentration of 117 ppm, was completely decomposed, while the acetone level was maximized at this time. Then, the acetone slowly disappeared, and in 70 h, the amount of evolved CO₂ attained the maximum level. The measured CO₂ level, 346 ppm, was close to the theoretical amount (351 ppm), suggesting that this photocatalytic system completely mineralized the IP without forming any intermediates.

Scheme 2b outlines the photocatalytic mechanism of the ATO/TiO₂/CdSe composite system. Under visible-light irradiation, the electrons in the VB of ATO and CdSe are excited to their CBs. Then, the electrons in the VB of TiO₂ can move to that of ATO, because the VB position of ATO lies below that of TiO₂. At the same time, the photo-excited electrons in the CB of CdSe can move to that of TiO₂, since the CB position of TiO₂ lies below that of CdSe. As a result, both electrons and holes can be generated in the CB and VB of TiO₂ under visible-light irradiation. In general, the photo-excited electron–hole pairs in bare TiO₂ can be recombined quickly. In this double heterojunction structure, however, it is expected that the e⁻–h⁺ pairs generated in the TiO₂ will have a relatively long lifetime, since the electrons and holes are space-charge-separated. In the ATO/TiO₂/CdSe structure, ATO nanoparticles are embedded in the core, and CdSe QDs are loaded onto the surface of the TiO₂ structure. Thus the electrons will be generated in the region neighboring the CdSe, whereas the holes will be formed near the ATO. Therefore, the space-charge-separated holes and electrons in TiO₂ will have more opportunity to participate in photocatalytic reactions.

It was demonstrated in the present work that the double-heterojunction ATO/TiO₂/CdSe composite, generating both electrons and holes on the surface of TiO₂, demonstrates remarkably higher efficiency in decomposing IP and evolving CO₂ than the ATO/TiO₂ system which generates holes only. This strongly suggests that not only the holes in the VB of TiO₂ but also the electrons in its CB play an important role in the complete decomposition of organic compounds. As indicated in eqn (3) and (4), HO₂˙ can be formed from the electrons in the CB of TiO₂. Therefore, both the HO˙ and HO₂˙ will be the key components in inducing the decomposition reactions.

On the other hand, in the case of the ATO/TiO₂/CdSe composite system under visible-light irradiation, the generated electrons in the CB of ATO and the holes in the VB of CdSe do not directly participate in oxidation reactions. In fact, it has been known that the holes in the VB of CdSe can be consumed by the adsorbed –OH group and/or induce photoanodic corrosion of CdSe,³¹ as shown in eqn (7). Then, what is the fate of the accumulated electrons in the CB of ATO? Considering that the CB level of ATO is considerably lower than the standard hydrogen potential, direct electron transfer to oxygen molecules, requiring −0.284 V (vs. NHE), will be difficult, as shown in eqn (8) . Thus we suppose that the electrons in the CB of ATO would be transported to oxygen species through the processes described in eqn (9)–(11). At present, we do not have concrete evidence for this. Presently mechanism studies are underway in order to fully understand the electron-transfer-passways in these novel heterojunction structures.

nh⁺ + nCdSe → nSe + nCd²⁺ (7)

O₂ + e⁻ → ˙O₂⁻, E₀ = −0.284 V (vs. NHE) (8)

O₂ + H⁺ + e⁻ → HO₂˙, E₀ = −0.046 V (9)

O₂ + 2H⁺ + 2e⁻ → H₂O₂, E₀ = +0.682 V (10)

O₂ + 4H⁺ + 4e⁻ → 2H₂O, E₀ = +1.23 V (11)

4. Conclusions

Deep blue-colored ATO (Sb_xSn_1-xO₂, x = 0.1) nanoparticles of ∼50 nm size were successfully prepared by co-precipitation and post-annealing at 1000 °C. The prepared ATO is a typical Type-B sensitizer, since it offers profound absorption in the visible range and its VB level is considerably lower than that of TiO₂. By the hole-transfer mechanism, the 5/95 ATO/TiO₂ composite exhibited notably high visible-light photocatalytic activity. The CO₂ evolved in 2 h was 8.1 ppmv, whereas both the bare ATO and TiO₂ showed negligible evolution. The double-heterojunction-structured 1% CdSe-loaded ATO/TiO₂ composite demonstrated remarkably enhanced the photocatalytic activity in mineralization of IP. The amount of CO₂ evolved was 23.2 ppmv after 2 h irradiation, which is ∼3 times and twice that of the bare ATO/TiO₂ and the typical N-doped TiO₂, respectively. The ultra-high visible-light photocatalytic activity of this novel double-heterojunction structure is due to the generation of both the electrons and holes generated in the CB and VB, respectively, of TiO₂ under visible-light irradiation. Hence, it is suggested that not only the holes in the VB but also the electrons in the CB play an important role in the complete decomposition of organic compounds, and, further, that both the HO˙ and HO₂˙ are key components in inducing such oxidation reactions.

Acknowledgements

This work has been supported by the Korean Center for Artificial Photosynthesis (KCAP) funded by the Ministry of Education, Science, and Technology (NRF-2011-C1AAA001-2011-0030278), the Ministry of Environment (Project No. 022-061-026), and the National Research Foundation of Korea (Project No. 2011-0002995).

References

1. T. Inoue, A. Fujishima, S. Konishi and K. Honda, Nature, 1979, 277, 637.
2. M. R. Hoffmann, S. T. Martin, W. Choi and D. W. Bahnemann, Chem. Rev., 1995, 95, 69.
3. E. S. Tang, J. H. Won, S. J. Hwang and J. H. Choy, Adv. Mater., 2006, 18, 3309.
4. Z. Ding, G. Q. Lu and P. F. Greenfield, J. Phys. Chem. B, 2000, 104, 4815.
5. Y. Ou, J. Lin, S. Fang and D. Liao, Catal. Commun., 2007, 8, 936.
6. Y. Huang, Z. Zheng, Z. Ai, L. Zhang, X. Fan and Z. Zou, J. Phys. Chem. B, 2006, 110, 19323.
7. C. Burda, Y. Lou, X. Chen, A. C. S. Samia, J. Stout and J. L. Gole, Nano Lett., 2003, 3, 1049.
8. V. Stengl, V. Houskova, S. Bakardjieva and N. Murafa, ACS Appl. Mater. Interfaces, 2010, 2, 575.
9. S. Klosek and D. Raftery, J. Phys. Chem. B, 2001, 105, 2815.
10. S. Sakthivel and H. Kisch, ChemPhysChem, 2003, 4, 487.
11. K. Y. Song, M. K. Park, Y. T. Kwon, H. W. Lee, W. J. Chung and W. I. Lee, Chem. Mater., 2001, 13, 2349.
12. M. Anpo and M. Takeuchi, J. Catal., 2003, 216, 505.
13. B. K. Vijayan, N. M. Dimitrijevic, J. Wu and K. A. Gray, J. Phys. Chem. C, 2010, 114, 21262.
14. J. C. Xu, M. Lu, X. Y. Guo and H. L. Li, J. Mol. Catal. A: Chem., 2005, 226, 123.
15. W. Choi, A. Termin and M. R. Hoffmann, J. Phys. Chem., 1994, 98, 13669.
16. P. Bouras, E. Stathatos and P. Lianos, Appl. Catal., B, 2007, 73, 51.
17. B. Peng, X. Meng, F. Tang, X. Ren, D. Chen and J. Ren, J. Phys. Chem. C, 2009, 113, 20240.
18. R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki and Y. Taga, Science, 2001, 293, 269.
19. F. Dong, H. Wang and Z. Wu, J. Phys. Chem. C, 2009, 113, 16717.
20. T. Tachikawa, S. Tojo, K. Kawai, M. Endo, M. Fujitsuka, T. Ohno, K. Nishijima, Z. Miyamoto and T. Majima, J. Phys. Chem. B, 2004, 108, 19299.
21. W. Zhao, W. Ma, C. Chen, J. Zhao and Z. Shuai, J. Am. Chem. Soc., 2004, 126, 4782.
22. J. C. Yu, W. Ho, J. Yu, H. Yip, P. K. Wong and J. Zhao, Environ. Sci. Technol., 2005, 39, 1175.
23. T. Morikawa, Y. Irokawa and T. Ohwaki, Appl. Catal., A, 2006, 314, 123.
24. T. Morikawa, T. Ohwaki, K. Suzuki, S. Moribe and S. T. Kubota, Appl. Catal., B, 2008, 83, 56.
25. H. Irie, K. Kamiya, T. Shibanuma, S. Miura, D. A. Tryk, T. Yokoyama and K. Hashimoto, J. Phys. Chem. C, 2009, 113, 10761.
26. D. Liu and P. V. Kamat, J. Electroanal. Chem., 1993, 347, 451.
27. L. Spanhel, H. Weller and A. Henglein, J. Am. Chem. Soc., 1987, 109, 6632.
28. X. Yu, Q. Wu, S. Jiang and Y. Guo, Mater. Charact., 2006, 57, 333.
29. J. C. Kim, J. K. Choi, Y. B. Lee, J. H. Hong, J. I. Lee, J. W. Yang, W. I. Lee and N. H. Hur, Chem. Commun., 2006, 5024.
30. Y. Bessekhouad, N. Chaoui, M. Trzpit, N. Ghazzal, D. Robert and J. V. Weber, J. Photochem. Photobiol., A, 2006, 183, 218.
31. N. Serpone, P. Maruthamuthu, P. Pichat, E. Pelizzetti and H. Hidaka, J. Photochem. Photobiol., A, 1995, 85, 247.
32. S. Y. Chai, Y. J. Kim and W. I. Lee, J. Electroceram., 2006, 17, 909.
33. Y. J. Kim, B. Gao, S. Y. Han, M. H. Jung, A. K. Chakraborty, T. G. Ko, C. M. Lee and W. I. Lee, J. Phys. Chem. C, 2009, 113, 19179.
34. S. B. Rawal, A. K. Chakraborty and W. I. Lee, Bull. Korean Chem. Soc., 2009, 30, 2613.
35. S. Y. Chai, Y. J. Kim, M. H. Jung, A. K. Chakraborty, D. W. Jung and W. I. Lee, J. Catal., 2009, 262, 144.
36. B. Gao, A. K. Chakraborty, J. M. Yang and W. I. Lee, Bull. Korean Chem. Soc., 2010, 31, 1941.
37. H. Irie, Y. Watanabe and K. Hashimoto, J. Phys. Chem. B, 2003, 107, 5483.
38. M. S. Wrighton, D. L. Morse, A. B. Ellis, D. S. Ginley and H. B. Abrahamson, J. Am. Chem. Soc., 1976, 98, 44.
39. Y. Xu and A. A. Schoonen, Am. Mineral., 2000, 85, 543.
40. C. A. Vincent and D. G. C. Weston, J. Electrochem. Soc., 1972, 119, 518.
41. B. Orel, U. Lavrencis-Stangar, Z. Crnjak-Orel, P. Bukovec and M. Kosec, J. Non-Cryst. Solids, 1994, 167, 272.
42. F. Bai, Mater. Lett., 2006, 60, 3126.
43. H. J. Jeon, Mater. Lett., 2005, 59, 1801.
44. Q. Dai, D. Li, H. Chen, S. Kan, H. Li, S. Gao, Y. Hou, B. Liu and G. Zou, J. Phys. Chem. B, 2006, 110, 16509.
45. K. S. Leschkies, R. Divakar, J. Basu, E. E. Pommer, J. E. Boercker, C. B. Carter, U. R. Kortshagen, D. J. Norris and E. S. Aydil, Nano Lett., 2007, 7, 1793.
46. Y. T. Kwon, K. Y. Song, W. I. Lee, G. J. Choi and Y. R. Do, J. Catal., 2000, 191, 192.
47. Y. M. Cross and D. R. Pyke, J. Catal., 1979, 58, 61.
48. T. Nutz, U. Z. Felde and M. Haasea, J. Chem. Phys., 1999, 110, 12142.
49. V. Dusastre and D. E. Williams, J. Phys. Chem. B, 1998, 102, 6733.
50. C. A. Vincent, J. Electrochem. Soc., 1972, 119, 515.
51. Z. Q. Li, Y. L. Yin, X. D. Liu, L. Y. Li, H. Liu and Q. G. Song, J. Appl. Phys., 2009, 106, 83701.
52. N. Naghavi, C. Marcel, L. Dupont, J.-B. Leriche and J.-M. Tarascon, Solid State Ionics, 2003, 156, 463.
53. A. Klein, C. Körber, A. Wachqu, F. Säuberlich, Y. Gassenbauer, S. P. Harvey, D. E. Proffit and T. O. Mason, Materials, 2010, 3, 4892.
54. P. V. Kamat, Chem. Rev., 1993, 93, 207–300.
55. C. Harris and P. V. Kamat, ACS Nano, 2009, 3, 682.
56. J. H. Bang and P. V. Kamat, ACS Nano, 2009, 3, 1467.

---