Synthesis of TiO2 nanoparticles using novel titanium oxalate complex towards visible light-driven photocatalytic reduction of CO2 to CH3OH

Applied Catalysis A: General 437–438 (2012) 28–35

Contents lists available at SciVerse ScienceDirect

Applied Catalysis A: General journal homepage: www.elsevier.com/locate/apcata

Synthesis of TiO2 nanoparticles using novel titanium oxalate complex towards visible light-driven photocatalytic reduction of CO2 to CH3 OH Quang Duc Truong 1 , Thi Hang Le, Jen-Yu Liu, Cheng-Chi Chung, Yong-Chien Ling ∗ Department of Chemistry, National Tsing Hua University, Hsinchu 30013, Taiwan

a r t i c l e

i n f o

Article history: Received 19 December 2011 Received in revised form 1 April 2012 Accepted 7 June 2012 Available online 15 June 2012 Keywords: TiO2 nanoparticles CO2 reduction CH3 OH Titanium oxalate complex Visible light response

a b s t r a c t TiO2 nanoparticles (NPs) with controlled crystalline structure and morphology were synthesized by a facile hydrothermal method using a novel titanium oxalate complex. The structure, morphology, and spectral properties of the synthesized TiO2 NPs were characterized by X-ray diffraction, Raman spectroscopy, scanning/transmission electron microscopy, and UV–vis diffuse reflectance spectroscopy. The titania phases of anatase, rutile, or brookite can be easily tuned by tailoring the solution pH during reaction. Highly ordered flower-like rutile could be obtained with oxalic acid additive. The synthesized TiO2 catalysts showed excellent visible light absorption and remarkable photocatalytic activity for CO2 reduction to CH3 OH under both UV–vis and visible light irradiation, mainly due to doped carbon and nitrogen. Bicrystalline anatase–brookite composite afforded maximum CH3 OH yield, attributed mainly to the unique electrical band structures and efficient charge transfer between the two crystalline phases. © 2012 Elsevier B.V. All rights reserved.

1. Introduction The foreseen shortage of fossil fuels is of global concern. The development of alternative and renewable energy sources has gained increased attention. Since the pioneering work of water splitting using a TiO2 electrode by Honda and Fujishima in early 1970s, the conversion of solar energy into chemical energy has been intensively studied [1]. Typical examples include transformation of solar energy into hydrogen via photocatalytic water splitting [2] and solar-driven photocatalytic conversion of CO2 into chemical fuels. The latter approach provides intriguing opportunity for economic utilization and simultaneous reduction of CO2 , which is beneficial to tackle global warming. Considerable efforts have been devoted to explore the photoreduction of CO2 into chemical fuel such as CO, CH4 , or CH3 OH [3–13]. A variety of catalysts such as metal oxides [3–6], InTaO4 [7], ZnGa2 O4 [8], Zn2 GeO4 [9,10], and Ti-based materials [11–13] have been explored for CO2 conversion in either liquid or gas phase. Among them, titanium oxide (TiO2 ) is one of the most promising candidates, presumably owing to its powerful oxidation properties, stability, efficiency, and long-term durability [13–23]. To improve the TiO2 application efficacy, current research has focused on the development of light-harvesting

∗ Corresponding author. Tel.: +886 3 5721484; fax: +886 3 5711082. E-mail address: [email protected] (Y.-C. Ling). 1 Current address: Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai 980 8577, Japan. 0926-860X/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.apcata.2012.06.009

strategy [18–20] and extended its photo-response to visible light region [13–17,21,22]. Indeed, a great deal of effort on investigating TiO2 doped with metal [13–20] or nonmetal elements [21,22] has been reported. By comparison, the effect of TiO2 crystalline structure, i.e. brookite phase, and morphology on the CO2 photoreduction has not been considered appreciably. Kakihana et al. recently proposed a novel route using water-soluble compounds formed by coordinating small organic molecules with metals, facilitating the control of nucleation and assembly to afford desired nanostructures by bottom–up approach [24–27]. A new series of water-soluble titanium complexes were developed as precursors for the selective synthesis and morphological control of TiO2 polymorphs [24–28]. The synthesized TiO2 were usually doped with nonmetal elements during the nucleation and crystallization stage, providing increased photoactivity under visible light irradiation [25]. We recently adopted this approach for the synthesis of FeTiO3 /TiO2 with increased photocatalytic performance [29]. Herein, we report the controlled synthesis of TiO2 NPs with various phases and morphologies for photocatalytic reduction of CO2 to CH3 OH. Particularly, we propose a hydrothermal method using a novel titanium oxalate complex to synthesize TiO2 catalysts with tunable crystalline structure and morphology. The oxalic acid ligand possess unique symmetry structure and rich coordination properties, assisting the control of nucleation and morphology to yield TiO2 catalysts containing anatase, rutile, or brookite phase. The size and morphology effect on TiO2 catalyst photocatalytic reduction of CO2 to CH3 OH was investigated in detail.

Q.D. Truong et al. / Applied Catalysis A: General 437–438 (2012) 28–35

2. Experimental 2.1. Synthesis of TiO2 NPs TiO2 was prepared by a hydrothermal method using titanium complex with oxalic acid as chelating agent [29]. Typically, metallic titanium powder (5 mmol, 99.4%, Alfa Aesar, America) was completely dissolved in a cold mixture of aqueous ammonia solution (10 ml, 28%, J.T. Baker, Germany) and hydrogen peroxide solution (40 ml, 30%, J.T. Baker, Germany). After removing the excess reagents by aging at 353 K, a yellowish gel-like peroxo-titanic acid intermediate was obtained and dissolved in 15 ml distilled water containing 2% H2 O2 to produce a peroxo-titanic acid solution with pH 5. In a typical synthesis, oxalic acid (7.5 mmol, 99.5%, Merk, Germany) was added to the peroxo-titanic acid solution. The solution color changed from yellow to red, suggesting the formation of titanium oxalate complex. Appropriate amounts of aqueous ammonia solution were subsequently added to the complex solution to adjust solution pH from 2.0 to 4.0, 6.0, 8.0, to 10.0. Additional amount of oxalic acid was added further lower the pH down to 1.5 (0.1 M) and 1.0 (0.2 M). The solutions were diluted to 20 ml (0.1 M) placed in a Teflon-lined stainless steel autoclave, heated at 473 K for 24 h, and finally cooled down to room temperature. The white/yellow solid powder was separated by centrifugation, washed with distilled water, and sonicated until a neutral pH was obtained. The obtained powder was dried at 353 K for 1 day to yield TiO2 catalysts.

above the reactor. A Pyrex glass was placed on top of the reactor to cut off light with wavelengths  < 300 nm. A UV cut-off filter consisted of 2 M NaNO2 solution was applied to cut off entire UV region ( < 400 nm) [32]. The photocatalytic reaction was continued up to 9 h and in every 3 h interval the reactor was allowed to cool down naturally for CH3 OH desorption from the catalyst. The sample solution was centrifuged and analyzed by GC-FID for CH3 OH with a 2 m Porapak Q column. The rate of CH3 OH evolution was defined as the total amount of evolved CH3 OH divided by the reaction time. The photocatalytic activity of the synthesized TiO2 NPs was compared with that of commercial P25 (anatase/rutile = 4/1) TiO2 catalyst. The photoreduction of CO2 by various catalysts has been extensively studied [3–23]. The following reaction was proposed for the selective production of CH3 OH: CO2 + 4H2 O → 2CH3 OH + 3O2

2.3. Photocatalytic reduction of CO2 The photocatalytic reduction of CO2 was carried out in an aqueous system under UV–vis or visible light irradiation according to our previous report [29,30]. In brief, 0.05 g catalyst was dispersed in 30 ml distilled water containing sodium bicarbonate (NaHCO3 , 0.08 M). The solution was prepared in a 30 cm × 15 cm × 5 cm reactor. The reaction system was initially tested to afford the optimal catalyst concentration of 1.65 g dm−3 , which was close to the value recommended by Ballari et al. [31]. Fixed catalyst concentration was used in all experiments, aiming to eliminate uncertainties associated with light reflection and reactor geometry. The solution was pursed by N2 to remove oxygen and stood in dark for 30 min until reaching adsorption–desorption equilibrium. The solution temperature was kept constant by a water bath. UV–vis and visible light irradiation was provided by a 500 W high-pressure Xe lamp placed

(1)

Prior to the photocatalytic reaction, the TiO2 catalyst was dispersed in distilled water and irradiated by UV light from a high-pressure mercury lamp for 5 h, aiming to remove any organic residues on TiO2 surface. We carried out control experiments in dark first and no CH3 OH was detected for all tested catalysts. Similarly, no CH3 OH was detected for control experiments under light irradiation but without catalyst. Preliminarily results from these control experiments indicate that both light irradiation and catalysts are indispensible for CO2 photoreduction to CH3 OH. The reaction quantum yield (QE) is estimated using CH3 OH yield noting that six electrons are required to reduce CO2 to CH3 OH [33]. The equation is as follows:

2.2. Characterization of TiO2 NPs The crystalline phase was characterized using powder X-ray diffraction (XRD; Rigaku D/MAX-IIB, 40 kV and 30 mA) with Cu ˚ The data were collected in 2– scanK␣ radiation ( = 1.5406 A). ning mode at a scan speed of 4◦ min−1 and a step-size of 0.02◦ . The morphology was examined using field-emission scanning electron microscopy (FE-SEM; Hitachi S-4700) and transmission electron microscopy (TEM; Jeol JEM-2010). N2 adsorption and desorption isotherms were measured at 77 K (Micromeritics ASAP 2010) to yield Brunauer–Emmett–Teller (BET) specific surface area. The phase composition was confirmed using Raman spectroscopy (Jasco R-3000) using 732.2 nm and 90 W laser for excitation. The incident beam was focused onto the catalyst sample using microscope 100× lenses to yield 2 ␮m spot. All samples were also subjected to diffuse reflectance spectroscopy (Shimadzu UV 2450) measurements. The samples were subjected to UV irradiation using a high-pressure mercury lamp for 5 h to remove organic residues on surface and placed onto a Si wafer before Time-of-flight secondary ion mass spectrometry (ToF-SIMS, ION-TOF IV) analysis using a pulsed 25 keV 69 Ga+ primary ion beam.

29

˚Methanol (%) = 100 × Mole of photon =

[6 × mole of CH3 OH yield] [mole of photon absorbed by catalyst]

[I × S] [NA × E]

(2) (3)

where I is light intensity (2.5 mW cm−2 and 0.12 mW cm−2 for UV and visible light, respectively); S is the irradiated area of the reactor (30 cm × 15 cm); E is the photon energy, (6.63 × 10−19 J at 300 nm and 4.97 × 10−19 J at 400 nm); NA is the Avogadro number (6.022 × 1023 mol−1 ). 3. Results and discussion It is well-known that precursor strongly affects the crystalline phase and morphology of the synthesized particles. In the case of titanium complex precursors, the ligand or complex structure and experimental condition crucially affect the crystalline phase and morphology of the synthesized nanostructures [24–27]. We prepared TiO2 catalysts with various phase and morphology by hydrothermal treatment of titanium oxalate complex. The effect of solution pH and oxalic acid additive on crystallite formation, size, shape, as well as the resulting photocatalytic activity of the synthesized TiO2 NPs were studied extensively. 3.1. Effect of solution pH Fig. 1 shows the XRD patterns of the synthesized TiO2 catalysts, indicating that all the samples consist of crystalline titanium oxide. The dominant crystalline phase was anatase when treating the complex with pH 2.0 solution (Fig. 1a) and changed to rutile as the solution pH increased to 4.0 and 6.0 (Fig. 1b and c). Further addition of ammonia to increase the solution pH to 8.0 and 10.0 resulted in the formation bicrystalline anatase–brookite composite (Fig. 1d and e). The brookite phase is evidenced from the presence of characteristic peak at 2 = 30.81◦ . Other brookite characteristic peaks at 2 = 25.34◦ and 2 = 25.69◦ were overlapped by those of anatase, yielding abnormally intensive peak at 2 = 25.32◦ .

30

Q.D. Truong et al. / Applied Catalysis A: General 437–438 (2012) 28–35

A,B

B

A

A

Intensity (arb. units)

A

A

A

(e)

(d) R

R

R

R

R

R

R

(c) brookiteanatase

Intensity (a.u)

R

(c)

(b) rutile

(b)

(a) anatase

(a) 20

25

30

35

40

45

50

55

60

65

70

75

80

2θ/ degree, CuKα

0

200

We used the method proposed by Mitsuhashi et al. to estimate the phase composition of anatase (A%) and brookite (B%) as follows: B% = 100

IB121 (IB120 + IA )

(4)

A% = 100 − B% where IA101 , IB120 and IB121 represents the XRD peak intensity of anatase (1 0 1), brookite (1 2 0) and (1 2 1), respectively [34]. The calculated phase composition is A = 69% and B = 31% with solution pH 8.0; A = 73% and B = 27% with solution pH 10.0. Similarly, the phase composition of anatase (A%) and rurile (B%) can also be estimated as follows: A% = 100

IA (IA + 1.265 IR )

R% = 100 ×

(5)

1.265IR (IA + 1.265 IR )

where IA and IR represent the XRD peak intensity of the anatase (1 0 1) and rutile (1 1 0) [35]. Table 1 lists the estimated phase composition of all catalysts and the corresponding crystallite size estimated using Scherrer equation. The crystalline structure was further confirmed using Raman spectroscopy. The Raman spectra shown in Fig. 2, indicating that anatase phase (curve a in Fig. 2) has characteristic bands at 395 cm−1 , 520 cm−1 and 630 cm−1 ; whereas rutile phase (curve b in Fig. 2) has characteristic bands at 445 cm−1 and 614 cm−1 . Generally, Raman spectroscopy is an efficient tool to clarify the presence of brookite phase [36]. The bicrystalline anatase–brookite

400

600

800

1000

1200

Raman shift / cm-1

Fig. 1. XRD patterns of the synthesized TiO2 catalysts with different solution pH: (a) 2.0, (b) 4.0, (c) 6.0, (d) 8.0, and (e) 10.0. A: anatase; B: brookite; R: rutile.

Fig. 2. Raman spectra of the synthesized TiO2 catalysts with different solution pH: (a) 2.0, (b) 6.0, and (c) 10.0.

composite shows additional Raman peaks at 247 cm−1 and 322 cm−1 from brookite crystallite besides the characteristic peaks of anatase crystallite [36,37], indicating that brookite crystallite has formed with solution pH 10.0. Fig. 3 shows the TEM images of the synthesized TiO2 catalysts, indicating size varied from 15 to 100 nm. The anatase catalyst obtained with solution pH 2.0 display spherical, square, or rod shape with diameter of 15 nm (Fig. 3a). The rutile catalyst obtained with solution pH 6.0 display 20–100 nm rod-like NPs (Fig. 3b) self-aggregated around a common center (the SEM image inset in Fig. 3b). The bicrystalline anatase–brookite composite obtained with solution pH 10.0 disperse uniformly and display spherical morphology with an average diameter of 10 nm (Fig. 3c). The enlarged TEM image (Fig. 3d) reveals the presence of rod-like NPs with an average diameter of 20 nm and length up to 50 nm, which is typical of brookite catalyst [24,38]. Fig. 3d also shows that the spherical anatase NPs are decorated or adhered strongly onto the surface of brookite rod-like NPs. Based on the estimated crystallite size in Table 1 and the TEM images in Fig. 3, we conclude that the anatase catalyst possesses spherical shape; whereas brookite catalyst possesses rod-like shape. It is plausible that simultaneous growth of anatase and brookite in our experimental condition has resulted in the formation of bicrystalline anatase–brookite composite. Similar growth process has been reported in hydrothermal reaction of titanium–lactate complex where the nucleation of anatase and brookite was controlled in situ by released OH− dosing [38].

Table 1 Chemical and physical properties of the synthesized TiO2 catalysts. pH

2 4 6 8 10 1

Crystallite sizea (nm)

Phase composition (%)

Energy band gap (eV)

A

B

R

A

B

R

97 100 6 69 73 0

0 0 0 31 27 0

3 0 94 0 0 100

12 11 – 9 9 –

– – – 25 25 –

– – 20 – – 33

3.10 – 3.00 – 2.85 3.00

a Estimated by the Scherrer equation: D(h k l) = (K )/(ˇ cos ) where K is the shape factor,  the wavelength of the Cu K␣ radiation, ˇ the full width at half-maximum (fwhm) of the (h k l) peak, and  the diffraction angle.

Q.D. Truong et al. / Applied Catalysis A: General 437–438 (2012) 28–35

31

Fig. 3. TEM images of the synthesized TiO2 catalysts with (a) anatase, (b) rutile, and (c, d) anatase–brookite phase. The inset in (b) showing a large-view SEM image of the corresponding sample.

Tomita et al. have reported direct effects from ligand on phase formation in the hydrothermal treatment of titanium complexes [24,25]. The ligand might coordinate to titanium in an octahedron environment which further assembles in skewed chains or linear chains polymerization, representing the initial structure of anatase or rutile crystallites [24,39]. We adopted a ligand/metal ratio of 1.5/1, aiming to afford titanium oxalate complex with structure similar to that reported by Uppal et al. [40]. Such complex structure is suitable for the polymerization of TiO6 octahedral unit in a skewed chain (see Supplementary material, Fig. S1), leading to the growth of anatase crystallites upon hydrothermal treatment. Interestingly, brookite appeared readily assembled in hydrothermal reaction with high solution pH. Up to date, the growth of brookite titania by using titanium–glycolate or titanium–picolinate complex with high solution pH has been successfully demonstrated [24,26]. The titanium–glycolate structurally resembled to the brookite unit and expectedly afforded high brookite crystallites. The titanium oxalate complex possessing specific octahedral arrangement [40] is similar to that of titanium–glycolate complex [24], facilitating the growth of brookite TiO2 polymorph with high solution pH as

demonstrated by Tomita et al. (see Supplementary material, Fig. S2). Furthermore, the solution pH has been reported playing an important role in the selective crystallization of anatase or rutile [38,41]. Hydroxide ions directly involves in the nucleation and growth of titania from a chelating complex [41]. At near neutral pH condition (4.0–6.0), the chelation of ligands to Ti is relatively stable, facilitating condensation along z-axis and forming mostly corner-shared bonds, which initiates the crystallization of rutile. At basic condition (pH 8.0–10.0), the chelation of ligands to Ti is fragile and readily cleaved due to the frequent nucleophilic attack to Ti by abundant hydroxide ions [41]. The hydrolysis and condensation thus occur primarily along edge-sharing bonds, resulting in the formation of anatase and brookite crystallites. 3.2. Effect of oxalic acid additive The carboxyl group of additive could strongly bind onto the specific crystal facets of titania surface and lead to various growth rates on different facets, resulting in the appearance of anisotropic nanostructures [27], which motivated us to examine the effect of

Q.D. Truong et al. / Applied Catalysis A: General 437–438 (2012) 28–35

oxalic acid additive on crystallite formation and morphology of the synthesized TiO2 catalysts. The original complex solution was added with different amounts of oxalic acid before hydrothermal treatment. The XRD patterns (Fig. 4) reveal that small amount of oxalic acid additive strongly affects the crystalline phase of the synthesized TiO2 catalyst. Rutile was the dominant phase (Fig. 4b and c) obtained with additive; whereas anatase was the dominant phase (Fig. 4a) without additive. As mentioned above, the crystallization of anatase from complex solution undergoes the polymerization of TiO6 octahedron into skewed chains. The presence of oxalic acid additives in solution prevents the formation of skewed chains through their chelation to TiO6 octahedron at bridging oxo group [42] (see Supplementary material, Fig. S3). The chelation would lead complex demeric in linear TiO6 octahedron chains, facilitating the growth of rutile crystallites. This is in good agreement with previous study that small amount of citric acid additive could promote the growth of rutile crystallites in hydrothermal treatment of amorphous titania [42]. Fig. 5 shows the TEM and SEM images of TiO2 catalysts obtained by hydrothermal treatment of titanium oxalate complex with oxalic acid additive. The TEM image (Fig. 5) reveals that the catalyst possessing multiple branches which assembled into anisotropic structure and afforded diversified shapes such as star and flower. The SEM image (Fig. 5b) indicates the presence of monodispersed and highly ordered flower-like NPs up to several hundred nanometers in diameter. The enlarged SEM image (Fig. 5c) further reveals that the catalyst consisted of several rod-shaped crystals interconnected through a common center. 3.3. UV–vis diffuse reflectance spectra and band gap energy The UV–vis diffuse reflectance spectra of the synthesized TiO2 catalysts and Degussa P25 are shown in Fig. 6A. The Dugussa P25 shows absorption in UV region (curve a) only; whereas the

R R R R Intensity (arb. units)

32

R

R

R

R

R

(c)

A

(b) A

A

A A

20

25

30

35

40

45

50

55

60

(a) 65

70

75

80

2θ/ degree, CuKα Fig. 4. XRD patterns of the synthesized TiO2 catalysts (a) without and (b) with 0.1 M (pH 1.5); (c) with 0.2 M (pH 1.0) oxalic acid additive, A: anatase; R: rutile.

synthesized TiO2 catalysts shows absorption (curves a–d) shift toward longer wavelength. The anatase catalyst (curve a), rutile catalyst (curve b), and flower-like rutile (curve d) show visible light absorption in the range of 400–600 nm; whereas the anatase–brookite composite (curve c) display strong absorption up to 700 nm. The absorption intensity of the bicrystalline anatase–brookite composite dominates that of anatase and rutile NPs. The absorption intensity in the visible-light region increases in the order anatase < rutile < anatase–brookite.

Fig. 5. (a) TEM image and (b, c) SEM images of the synthesized TiO2 catalysts with 0.2 M (pH 1.0) oxalic acid additive.

Q.D. Truong et al. / Applied Catalysis A: General 437–438 (2012) 28–35

Intensity

OH

33

O2

CNO

C2H CN

C3H

NO2

TiO2

NO3

(a) (b) 10 15 20 25

30 35 40 45 50 55 60 65

70 75 80 85

Mass / u Fig. 7. ToF-SIMS negative ion mass spectra of (a) synthesized anatase–brookite catalyst and (b) commercial P25.

the band gap energy decrease from 3.10 eV of anatase to 2.85 eV of anatase–brookite, agreeing well with the absorption shift to visible region observed in the anatase–brookite DRS spectrum (curve c in Fig. 6A). 3.4. Photocatalytic reduction of CO2 to CH3 OH The photocatalytic activity of the synthesized TiO2 catalysts was evaluated in terms of photoreduction of CO2 to CH3 OH under UV–vis ( > 300 nm) and visible ( > 400 nm) light irradiation. Generally, CH3 OH is formed by CO2 photoreduction on the photocatalyst upon UV or visible light irradiation. The photogenerated holes in the valence band (VB) can initiate the oxidation of water by the following half-reaction: Fig. 6. (A) UV–vis diffuse reflectance spectra of the synthesized TiO2 catalysts with (a) anatase, (b) rutile, (c) anatase–brookite, (d) flower-like rutile phase, and (e) P25. The inset in (A) showing the photograph of the corresponding samples. (B) Plot of the Kubelka–Munk function versus the energy of absorbed light.

The remarkable absorption efficiency of the synthesized TiO2 catalysts in visible light region was presumably attributed to the carbon and nitrogen doping. Tomita et al. reported that 1–4 wt% of carbon and 0.2–0.5 wt% of nitrogen were detected in titania nanopowders synthesized from titanium complexes by hydrothermal method. The strong binding of small organic molecules with Ti usually facilitated the carbon and/or nitrogen doping of synthesized particles during the nucleation and crystallization process [25]. We speculated that the synthesized TiO2 catalyst might also be doped by carbon and/or nitrogen. ToF-SIMS analysis of precleaned anatase–brookite bicrystaline catalyst was carried out. The resultant ToF-SIMS spectrum (Fig. 7a) show additional fragments of C2 H− (m/z 25), C3 H− (m/z 37), CN− (m/z 26), NO2 − (m/z 46), NO3 − (m/z 62) compared to that of Degussa P25 (Fig. 7b), indicating the presence of surface carbon and nitrogen. The strong absorption in 450–460 nm region by anatase–brookite crystallite (curve c in Fig. 6A) implied the presence of Ti N bonding in TiO2 structures. This absorption edge has been reported being related to the newly formed N1s orbit by doping nitrogen into the molecular structure [36]. Moreover, we also found that the visible-light absorption intensity of the synthesized TiO2 catalysts increased with higher solution pH which was prepared by adding larger amounts of aqueous ammonia solution. For instance, anatase–brookite catalyst prepared with solution pH 10.0 shows the most intensive visible light adsorption. We concluded that ammonia provides the nitrogen source doped into the synthesized NPs. To further study the doping effect on the optical property of TiO2 catalysts, we estimated the band gap energy by plotting the modified Kubelka–Munk function [F(R)E]1/2 calculated from the optical absorption spectrum vs. the energy of irradiation light (Fig. 6B). The respective band gap energy is listed in Table 1, indicating that

H2 O → 0.5O2 + 2H+ + 2e−

0 (Eox = 1.23 V vs. NHE)

(6)

The VB edge of TiO2 is about 2.7 V (vs. NHE) which is more positive than E0 (O2 /H2 O), facilitating subsequent photooxidative reaction process. The photogenerated electrons in the conduction band (CB) can initiate CO2 photoreduction by the following half-reaction: 0 CO2 + 6e− + 6H+ → CH3 OH + H2 O (Ered = 0.03 V vs. NHE)

(7)

Fig. 8 shows that the CB edge of TiO2 is −0.5 V (vs. NHE) that is lower than E0 (CO2 /CH3 OH), facilitating CO2 photoreduction to CH3 OH. Koˇcí et al. have successfully conducted UV light-induced photocatalytic reduction of CO2 on TiO2 catalyst in liquid phase. The yield of detectable products in decreased order of H2 > CH4 > CH3 OH  CO was found [43,44]. We focused on CH3 OH product in liquid phase since CH3 OH is regarded as a convenient liquid fuel and raw material for synthetic hydrocarbons and their products. Table 2 lists the photocatalytic performance in terms of CH3 OH yield and QE, as well as the specific BET surface area of the

Fig. 8. Schematic diagram illustrating the charge transfer between two crystalline phases.

34

Q.D. Truong et al. / Applied Catalysis A: General 437–438 (2012) 28–35

Table 2 Photocatalytic performance and BET specific areas of the TiO2 catalysts. pH

– 2 6 10 1

Catalyst

P25 Anatase Rutile Anatase-brookite Flower-like rutile

BET (m2 g−1 )

52.6 62.3 35.8 50.3 30.7

CH3 OH yield (␮mol g−1 h−1 )/QE (%) UV–vis irradiation

Visible light irradiation

0.176/0.00176 0.185/0.00185 0.300/0.00300 0.590/0.00590 0.320/0.00320

0.045/0.0067 0.141/0.0211 0.192/0.0288 0.478/0.0717 0.250/0.0375

synthesized TiO2 catalysts. Upon UV–vis light ( > 300 nm) irradiation, the anatase TiO2 afforded CH3 OH yield (bar b in Fig. 9A) comparable to that of P25 (bar b in Fig. 9A). The CH3 OH yield afforded by rutile TiO2 (bar c in Fig. 9A) is significantly higher than that by anatase. Above all, the anatase–brookite exhibited remarkable photoactivity with 0.590 ␮mol g−1 h−1 CH3 OH yield (bar d in Fig. 9A) that is 3.4 times higher than that by P25 or anatase catalyst. Upon visible light ( > 400 nm) irradiation, the photocatalytic activity of all synthesized TiO2 catalysts was significantly better than that of P25. Similar to UV light irradiation, anatase–brookite catalyst showed the highest improvement among all samples. The CH3 OH yield by anatase–brookite is 0.478 ␮mol g−1 h−1 , which is 2.5 and 2.3 times higher than those by anatase and rutile catalysts. There are few reports on photocatalytic reduction of CO2 under visible light irradiation. For instance, Ozcan et al. have reported visible light-driven photoproduction of CH4 using dye-sensitized Pt/TiO2 with a production rate of 0.26 ␮mol g−1 h−1 [23]. Varghese et al. applied N-doped TiO2 nanotube arrays for photoreduction of CO2 under the visible light irradiation with a total evolution rate (CO, CH4 and other hydrocarbons) of 0.33 ␮mol g−1 h−1 [21]. Zhang et al. recently reported an impressive photoreduction rate under visible light irradiation with a CO production rate of 2.4 ␮mol g−1 h−1 [22]. On comparison, the photoproduction rate obtained in this work (CH3 OH of 0.48 ␮mol g−1 h−1 ) is higher than that obtained using dye-sensitized or N-doped TiO2 [21,23]. The quantum

Fig. 9. (A) The CH3 OH yield in the photoreduction of CO2 by (a) commercial P25, and the synthesized TiO2 catalysts (b) anatase, (c) rutile, (d) anatase–brookite, (e) flower-like rutile. (B) The CH3 OH yield in the photocatalytic reduction of CO2 by anatase–brookite catalyst over a 9 h time course.

efficiency (QE) of the synthesized TiO2 catalyst was estimated (Table 2). The CH3 OH QE obtained under visible light irradiation ranges from 0.021 to 0.072%, which is comparable to the CH4 QE obtained under UV irradiation reported by Wu et al. [33]. In order to investigate the robust nature of the synthesized TiO2 catalysts, the photoreaction experiment was carried over long reaction time. Fig. 9B shows the CH3 OH yield by anatase–brookite catalyst as a function of irradiation time, indicating that CH3 OH amount increased linearly with light irradiation time up to several hours. The evolution rate was high at the early reaction stage and leveled off after 3 h reaction time, presumably due to the generation of large amounts of CH3 OH which became participating in competing for photogenerated holes on catalyst. The fact that the synthesized TiO2 catalysts, irrespective of the structural difference, all exhibit improved photocatalytic performance under both UV–vis and visible light irradiation suggests the possibility of carbon and nitrogen doping. We used organic ligand and ammonia aqueous solution during the synthesis which might cause the doping of carbon and nitrogen elements onto the synthesized TiO2 catalysts. The TiO2 doped with non-metal elements such as nitrogen or iodine have been demonstrated possessing improved photocatalytic activity for CO2 reduction [21,22], indicating the doping of non-metal elements on the synthesized TiO2 catalysts is an important factor for enhancing their catalytic efficiency. The rutile catalyst shows higher CH3 OH yield (bar c in Fig. 9A) than that of anatase catalyst (bar b in Fig. 9A) despite of their lower specific surface area. This enhancement may be attributed to the high crystalline purity and specific surface property of rutile. It is well-known that larger crystallite size might lead to lower density of crystalline defects; hence, it will increase the photocatalytic activity of rutile. In addition, as shown in Fig. 3b, the rutile catalyst displays a relatively smooth surface, indicating the presence of flat exposed facets. The rutile has strong tendency to grow along c-axis with the exposure of (1 1 0) facets [27]. Oxygen vacancies located on TiO2 (1 1 0) surface is considered as active sites capable of enhancing the absorption by and charge transfer to adsorbed molecules [45]. Highly adsorptive ability of rutile (1 1 0) has been reported in the adsorption of formate and acetate species [46] as well as water adsorption/dissociation [47]. The unique flat facets in rutile become the preferential sites for adsorption of carbon-containing molecules such as CO3 2− or HCO3 − [48]. More importantly, the presence of both {1 1 0} and {1 1 1} facets could lead the efficient separation of photogenerated charges among the crystal facets co-exposed. Consequently, upon light irradiation the photogenerated electrons can be transferred directly to the carbonate species, accelerating the photoreductive reaction. Furthermore, 6% anatase was found in the rutile TiO2 , facilitating charge carrier separation and consequently yielding improved photocatalytic activity [49]. Therefore, despite of its smaller specific surface area, the rutile catalyst still exhibit higher photocatalytic activity than that of the anatase catalyst. The remarkable photocatalytic performance of the anatase–brookite catalyst under both UV–vis and visible light irradiation may be benefit from the visible-light response and the junction effect between two crystalline phases. The anatase–brookite catalyst display an intensive visible light absorption (curve c in Fig. 6) and small band gap energy of 2.85 eV (Table 2), which also contribute to the high photocatalytic efficiency afforded by anatase–brookite catalyst than that by anatase and rutile catalysts. Furthermore, the band gap energy of brookite (3.25–3.29 eV) is positioned more positive than that of anatase (3.2 eV) and corresponding CB energy edge (−0.55 eV) is slightly lower than that of anatase (−0.5 eV) (Fig. 8). The band gap energy changed but it may only alter the VB energy rather than CB energy [36]. Therefore, the difference in CB energy between the anatase and brookite crystallites still remained. Upon light irradiation, the electron-hole pairs

Q.D. Truong et al. / Applied Catalysis A: General 437–438 (2012) 28–35

are formed by photo-excitation. Taking advantage of CB energy edge difference, electrons in brookite CB can readily transfer to anatase CB, leaving behind the holes in brookite VB. Consequently, electron-hole pairs are well-separated and abundant electrons on anatase surface are available for CO2 reduction. Additionally, anatase catalyst possessing large specific area also facilitated electrons transfer to the adsorbed carbonate species, attributing to increased CH3 OH yield. The junction effect between two crystallites has been reported responsible for remarkable photocatalytic activity in anatase–brookite composite [38,50]. For example, Ozawa et al. have reported a 5.4 times higher photocatalytic activity by anatase–brookite composite compared to that of anatase powder in the oxidation of acetic aldehyde [51]. Zhang et al. recently reported very impressive CO2 photoreduction performance by iodine-doped anatase–brookite composite under both UV and visible light irradiation [22]. Similar improvement in photocatalytic performance by anatase–brookite composite demonstrated in this work evidenced that junction effect between two crystallites is also crucial for CO2 photoreduction by TiO2 catalyst. 4. Conclusions TiO2 NPs with controlled crystalline structure and morphology were synthesized by hydrothermal treatment of a novel titanium oxalate complex. The desired phase of anatase, rutile, or brookite can be easily tuned by tailoring the solution pH. Highly ordered flower-like rutile catalyst was obtained using oxalic acid additive. The synthesized TiO2 NPs showed excellent visible light absorption and remarkable photocatalytic performance for CO2 reduction under both UV–vis and visible light irradiation, which is attributed to the doping of carbon and nitrogen onto the synthesized TiO2 catalyst. The maximum CH3 OH yield of 0.478 ␮mol g−1 h−1 was afforded by bicrystalline anatase–brookite catalyst under visible light irradiation, presumably due to its unique electrical band structures and efficient charge transfer between two crystallites. Acknowledgements We gratefully acknowledge the financial support from the National Science Council of Republic of China (NSC98-2113-M-007016-MY3 and NSC99-2627-M-007-011) and National Tsing Hua University. The constructive and valuable comments by the reviewers to improve the quality of the work are also acknowledged. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.apcata.2012.06.009. References [1] A. Fujishima, K. Honda, Nature 238 (1972) 37–38. [2] A. Kudo, Y. Miseki, Chem. Soc. Rev. 38 (2009) 253–278. [3] X.J. Feng, J.D. Sloppy, T.J. LaTemp, M. Paulose, S. Komarneni, N.Z. Bao, C.A. Grimes, J. Mater. Chem. 21 (2011) 13429–13433. [4] C.C. Lo, C.H. Hung, C.S. Yuan, J.F. Wu, Sol. Energy Mater. Sol. Cells 91 (2007) 1765–1774. [5] K. Teramura, T. Tanaka, H. Ishikawa, Y. Kohno, T. Funabiki, J. Phys. Chem. B 108 (2004) 346–354.

35

[6] K. Teramura, H. Tsuneoka, T. Shishido, T. Tanaka, Chem. Phys. Lett. 467 (2008) 191–194. [7] Z.Y. Wang, H.C. Chou, J.C.S. Wu, D.P. Tsai, G. Mul, Appl. Catal. A: Gen. 380 (2010) 172–177. [8] S.C. Yan, S.X. Ouyang, J. Cao, M. Yang, J.Y. Feng, X.X. Fan, L.J. Wan, Z.S. Li, J.H. Ye, Y. Zhou, Z.G. Zou, Angew. Chem. Int. Edit. 49 (2010) 6400–6404. [9] Q. Liu, Y. Zhou, J.H. Kou, X.Y. Chen, Z.P. Tian, J. Gao, S.C. Yan, Z. Zou, J. Am. Chem. Soc. 132 (2010) 14385–14387. [10] N. Zhang, S.X. Ouyang, P. Li, Y.J. Zhang, G.C. Xi, T. Kako, J.H. Ye, Chem. Commun. 47 (2011) 2041–2043. [11] T. Inoue, A. Fujishima, S. Konishi, K. Honda, Nature 277 (1979) 637–638. [12] M. Anpo, H. Yamashita, K. Ikeue, Y. Fujii, S.G. Zhang, Y. Ichihashi, D.R. Park, Y. Suzuki, K. Koyano, T. Tatsumi, Catal. Today 44 (1998) 327–332. [13] Y. Li, W.-N. Wang, Z. Zhan, M.-H. Woo, C.-Y. Wu, P. Biswas, Appl. Catal. B: Environ. 100 (2010) 386–392. [14] O. Ishitani, C. Inoue, Y. Suzuki, T. Ibusuki, J. Photochem. Photobiol. A: Chem. 72 (1993) 269–271. [15] S.H. Chien, M.C. Kuo, C.H. Lu, K.N. Lu, Catal. Today 97 (2004) 121–127. [16] T. Yui, A. Kan, C. Saitoh, K. Koike, T. Ibusuki, O. Ishitani, ACS Appl. Mater. Interfaces 3 (2011) 2594–2600. [17] C.C. Yang, Y.H. Yu, B. van der Linden, J.C.S. Wu, G. Mul, J. Am. Chem. Soc. 132 (2010) 8398–8406. [18] I.H. Tseng, W.C. Chang, J.C.S. Wu, Appl. Catal. B: Environ. 37 (2002) 37–48. [19] I.H. Tseng, J.C.S. Wu, H.Y. Chou, J. Catal. 221 (2004) 432–440. [20] J.C.S. Wu, H.M. Lin, C.L. Lai, Appl. Catal. A: Gen. 296 (2005) 194–200. [21] O.K. Varghese, M. Paulose, T.J. LaTempa, C.A. Grimes, Nano Lett. 9 (2009) 731–737. [22] Q. Zhang, Y. Li, E. Ackerman, M. Gajdardziska-Josifovska, H. Li, Appl. Catal. A: Gen. 400 (2011) 195–202. [23] O. Ozcan, F. Yukruk, E.U. Akkaya, D. Uner, Top. Catal. 44 (2007) 523–528. [24] K. Tomita, V. Petrykin, M. Kobayashi, M. Shiro, M. Yoshimura, M. Kakihana, Angew. Chem. Int. Edit. 45 (2006) 2378–2381. [25] K. Tomita, M. Kobayashi, V. Petrykin, S. Yin, T. Sato, M. Yoshimura, M. Kakihana, J. Mater. Sci. 43 (2008) 2217–2221. [26] M. Kobayashi, V. Petrykin, K. Tomita, M. Yoshimura, M. Kakihana, Chem. Mater. 19 (2007) 5373–5376. [27] M. Kobayashi, V. Petrykin, M. Kakihana, K. Tomita, J. Am. Ceram. Soc. 92 (2009) S21–S26. [28] Q.D. Truong, M. Kobayashi, H. Kato, M. Kakihana, J. Ceram. Soc. Jpn. 119 (2011) 513–516. [29] Q.D. Truong, J.Y. Liu, C.C. Chung, Y.C. Ling, Catal. Commun. 19 (2012) 85–89. [30] J.Y. Liu, B. Garg, Y.C. Ling, Green Chem. 13 (2011) 2029–2031. [31] M.M. Ballari, R. Brandi, O. Alfano, A. Cassano, Chem. Eng. J. 136 (2008) 50. [32] K. Maeda, T. Takata, M. Hara, N. Saito, Y. Inoue, H. Kobayashi, K. Domen, J. Am. Chem. Soc. 127 (2005) 8286–8287. [33] T. Nguyen, J.C.S. Wu, Appl. Catal. A 335 (2008) 112–120. [34] T. Mitsuhashi, M. Watanabe, Miner. J. 9 (1978) 236–240. [35] R.A. Spurr, H. Myers, Anal. Chem. 29 (1957) 760–762. [36] S. Yin, Y. Aita, M. Komatsu, J. Wang, Q. Tang, T. Sato, J. Mater. Chem. 15 (2005) 674–682. [37] G.A. Tompsett, G.A. Bowmaker, R.P. Cooney, J.B. Metson, K.A. Rodgers, J.M. Seakins, J. Raman Spectrosc. 26 (1995) 57–62. [38] T.A. Kandiel, A. Feldhoff, L. Robben, R. Dillert, D.W. Bahnemann, Chem. Mater. 22 (2010) 2050–2060. [39] Y. Zheng, E. Shi, Z. Chen, W. Li, X. Hu, J. Mater. Chem. 11 (2001) 1547–1551. [40] R. Uppal, C.D. Incarvito, K.V. Lakshmi, A.M. Valentine, Inorg. Chem. 45 (2006) 1795–1804. [41] N.M. Kinsinger, A. Wong, D. Li, F. Villalobos, D. Kisailus, Cryst. Growth Des. 10 (2010) 5254–5261. [42] H. Yin, Y. Wada, T. Kitamura, S. Kambe, S. Murasawa, H.H. Mori, T. Sakata, S. Yanagida, J. Mater. Chem. 11 (2001) 1694–1703. ˇ ´ J. Jirkovsky, ´ O. Solcová, Appl. [43] K. Koˇcí, L. Obalová, L. Matˇejová, D. Plachá, Z. Lacny, Catal. B: Environ. 89 (2009) 494–502. ˇ ´ D. Plachá, L. Capek, A. ˚ L. Obalová, S. Krejˇcíková, Z. Lacny, [44] K. Koˇcí, K. Matˇeju, ˇ Appl. Catal. B: Environ. 96 (2010) 239–244. Hospodková, O. Solcová, [45] T.L. Thompson, J.T. Yates, Chem. Rev. 106 (2006) 4428–4453. [46] F.P. Rotzinger, J.M. Kesselman-Truttmann, S.J. Hug, J. Phys. Chem. B 108 (2004) 5004–5017. [47] R. Schaub, R. Thostrup, N. Lopez, Phys. Rev. Lett. 87 (2001) 266104. [48] L.F. Liao, C.F. Lien, D.L. Shieh, M.T. Chen, J.L. Lin, J. Phys. Chem. B 106 (2002) 11240–11245. [49] T.A. Kandiel, R. Dillert, A. Feldhoff, D.W. Bahnemann, J. Phys. Chem. C 114 (2010) 4909–4915. [50] J.C. Yu, J. Yu, W. Ho, L. Zhang, Chem. Commun. 194 (2001) 2–1943. [51] T. Ozawa, M. Iwasaki, H. Tada, T.K. Akita, S.I. Tanaka, J. Colloid Interf. Sci. 281 (2005) 510–513.