Morphological effects on the photocatalytic properties of SnO2 nanostructures

Journal of Alloys and Compounds 810 (2019) 151718

Contents lists available at ScienceDirect

Journal of Alloys and Compounds journal homepage: http://www.elsevier.com/locate/jalcom

Morphological effects on the photocatalytic properties of SnO2 nanostructures  wka a, Sumanta Sain b, Spyder-Ryder I. Sloman a, Olga Montes c, Arik Kar a, *, 1, Joanna Olszo  n Ferna ndez c, Swapan K. Pradhan d, Andrew E.H. Wheatley a, ** Asuncio a

Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge, CB2 1EW, UK Department of Materials Science, Indian Association for the Cultivation of Science, Jadavpur, Kolkata, 700 032, India Instituto de Ciencia de Materiales de Sevilla (CSIC et Univ. Sevilla), Avda. Am erico Vespucio Nr. 49, CIC Cartuja, 41092, Sevilla, Spain d Materials Science Division, Department of Physics, The University of Burdwan, Golapbag, Burdwan, West Bengal, 713 104, India b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 27 July 2018 Received in revised form 15 April 2019 Accepted 5 August 2019 Available online 5 August 2019

The photocatalytic properties of SnO2 nanocrystals are tuned by varying their morphology and microstructure. SnO2 nanoparticles and nanowedges have been synthesized using hydrothermal methods, while microwave irradiation techniques have given nanospheres. Detailed structural and chemical characterization of these different morphologies has been accomplished. The influence of SnO2 morphology on photocatalytic activity has been examined by monitoring the degradation of aqueous methylene blue dye. Results demonstrate that changing the morphology of the SnO2 modulates both surface area and levels of surface defects and that these alterations are reflected in the photocatalytic properties of the materials. The degradation of methylene blue dye (98%) in the presence of SnO2 nanoparticles under simulated solar irradiation is superior to previously reported photocatalyst performance and is comparable to that of standard TiO2 (Degussa P-25). The SnO2 nanoparticles perform better than both the nanowedges and nanospheres and this is attributed to the number of surface defects available to the high surface area material. They also reveal outstanding recyclability and stability. © 2019 Elsevier B.V. All rights reserved.

Keywords: Dye Morphology Photocatalysis Semiconductor SnO2 nanocrystals

1. Introduction The ability to morphologically-control the formation of inorganic nanocrystals is a target that has attracted considerable attention by virtue of the resulting ability to modulate the often unique properties available on the nanoscale [1]. As one of the most significant classes of inorganic nanocrystal, metal oxide semiconductors find applications in diverse areas of science and technology. This is particularly true of TiO2 and ZnO, for which the tailoring of physicochemical properties through the achievement of morphological control is well established [2]. This contrasts with the situation for another readily accessible metal oxide semiconductor, tin oxide (SnO2). This n-type, wide band gap (3.6 eV at 300 K) semiconductor [3] is employed in a range of applications

* Corresponding author. ** Corresponding author. E-mail address: [email protected] (A. Kar). 1 Present address: Indian Institute of Engineering Science and Technology, Shibpur, Howrah 711103, Kolkata, India. https://doi.org/10.1016/j.jallcom.2019.151718 0925-8388/© 2019 Elsevier B.V. All rights reserved.

including solar cells [4], catalytic support materials [5], transparent electrodes [6], and solid-state chemical sensors [7]. Applications interest derives from the material's high charge carrier density, chemical stability, outstanding electrical, optical, and electrochemical properties, and ability to present high levels of surface defects and oxygen vacancies. It has been established both that these defects and vacancies participate in the relaxation dynamics and recombination processes associated with photoinduced charge carriers [8] and that the size and shape of SnO2 nanocrystals significantly influences their formation [9]. Modulating physical properties through efficient morphological control over SnO2 nanocrystal synthesis therefore represents an area of major significance. The photocatalytic degradation of organic pollutants in water and air using metal oxide nanocrystals has garnered widespread interest in recent years [10]. Previous work has established that SnO2 nanocrystals can degrade various persistent organic pollutants, including detergents, dyes, pesticides and volatile organic compounds under simulated solar irradiation [11]. However, the rapid recombination rate of photogenerated electron-hole pairs in

2

A. Kar et al. / Journal of Alloys and Compounds 810 (2019) 151718

SnO2 nanocrystals has hitherto hindered the commercialization of this technology [12]. Attempts to overcome this drawback have generally focused on controlling the energy gap between the valence and conduction bands by tuning the morphology of the SnO2 nanocrystals [13]. However, it has also been suggested that differing SnO2 nanocrystal size and shape promises to influence photocatalytic properties through variation of the number of oxygen-related defects presented at the particle surface [14]. It is clear that preparing SnO2 nanocrystals with controllable size, shape, and morphology remains a noteworthy challenge [15], though recent advances in the construction of nanoparticles, -cubes and -rods [16], nanobelts [17], florets [18], and ultrathin nanosheets [16,19] have been reported. Nevertheless, to date, few reports exist on the connection between SnO2 nanocrystal size with photocatalytic activity [20]. Therefore, although the controllable synthetic targeting of overall SnO2 morphology is described and nanoparticles (NPs) [21], florets [22] and nanorods [23], have been adapted and shown to be active in degrading dyes and other organic pollutants, the literature at present contains very limited reports that compare and correlate the photocatalytic activity of SnO2 nanocrystals with morphology [23a] to explicitly explain trends in the former in terms of the latter at the atomistic level. Additionally, reports attempting to unambiguously relate photocatalytic activity to the level of surface defects and oxygen vacancies are very few. In this latter context, limited inroads have been made into improving the sensing and photocatalytic properties of SnO2 by divulging specific crystal facets [24] or of varying properties through controlled sizing or shaping [25]. In this work we investigate how photocatalytic properties of SnO2 can be varied by altering its morphology. NPs, nanowedges (NWs) and nanospheres (NSs) have been synthesized by using either hydrothermal or microwave irradiation methods. The crystalline phases inherent to these SnO2 nanocrystals have been evidenced and the population and type of surface defects and oxygen vacancies associated with each elucidated. Photocalalytic activity has been measured in the context of MB dye degradation and catalyst recovery and recyclability tests have been undertaken to establish photocatalyst stability. Finally, a photocatalytic mechanism has been proposed that seeks to correlate the extent of oxygen vacancies and surface defects available with photocatalytic properties.

2. Experimental section 2.1. Materials and preparation 2.1.1. General synthetic and analytical details Chemicals were obtained from Sigma-Aldrich (analytical grade reagents, HPLC grade solvents) and used without further purification. Aqueous solutions used HPLC grade water (Millipore).

2.1.3. Preparation of SnO2 nanowedges SnCl4,5H2O (0.526 g, 1.5 mmol) was dissolved in water (2.5 ml) and NaOH (0.6 g, 15.0 mmol) was dissolved in water (50 ml). The solutions were combined and stirred for 30 min. Cetyl trimethyl ammonium bromide (CTAB, 0.729 g, 2.0 mmol) was added and the resulting solution stirred for 1 h. The mixture was heated, isolated and washed as for the NP preparation above. 2.1.4. Preparation of SnO2 nanospheres [26] SnCl4,5H2O (0.351 g, 1.0 mmol) was dissolved in water (25 ml) and a solution of CTAB (0.365 g, 1.0 mmol) in water (5 ml) added. The mixture was stirred for 1 h. NaOH (0.8 g, 20.0 mmol) was dissolved in water (10 ml) and EtOH (10 ml) was added. This solution was added dropwise to the stirred SnCl4$5H2O solution to give a white precipitate that dissolved at pH 11e12. The solution was heated in a microwave reactor (CEM Discover, operating at 300 W) at 110
C for 5 min to give a white colloidal suspension. Isolation and washing was as for SnO2 NPs. 2.2. Characterization 2.2.1. Transmission electron microscopy (TEM) Samples were screened for size and morphology using either an FEI Tecnai 20 TEM operated at 200 kV, an FEI Talos 200S TEM operated at 200 kV or an FEI Tecnai G2 F30 TEM operated at 300 kV. Data were processed using Gatan Digital Micrograph 3.6.5. Sample preparation was by droplet coating of ethanolic suspensions on carbon-coated Cu grids (Agar Scientific, 300 mesh). 2.2.2. X-ray diffraction (XRD) XRD profiles were recorded using Ni-filtered CuKa radiation from a highly stabilized and automated PAN-analytical X-ray generator operated at 40 kV and 40 mA. The X-ray generator was coupled with a PW3071/60 bracket goniometer for sample mounting. Step-scan data (step size 0.02
2q, counting time 2 s/ step) were recorded for 20
e75
2q. 2.2.3. Rietveld analysis Rietveld analysis [27] has been employed to extract (micro) structural parameters from XRD patterns [28]. Iterative least squares refinement [29] assessed how full-width-at-halfmaximum (FWHM) values of all reflections varied according to the Caglioti relation [30].


1=2 FWHM ¼ Utan2 q þ Vtanq þ W

The integrated intensity of each peak was taken to depend on the (micro)structural parameters. The least square refinement procedure minimized the value of the weighted residue, Rwp:

"P

ðYio  Yic Þ2 P 2 i wi Yio

i wi

Rwp ¼ 2.1.2. Preparation of SnO2 nanoparticles SnCl4,5H2O (0.787 g, 2.2 mmol) was dissolved in water (15 ml) and the resulting solution stirred for 30 min. NaOH (0.8 g, 20.0 mmol) was dissolved in water (20 ml) and 20 ml of EtOH was added. This solution was added dropwise to the stirred SnCl4$5H2O solution to give a white precipitate which dissolved at pH 11e12. The clear solution was placed in a 40 ml Teflon-lined steel chamber that was then heated in a furnace at 200
C for 24 h. The chamber was air-cooled to room temperature before the contents were removed, centrifuged (8000 rpm for 5 min) and washed with water (2  25 ml) and absolute EtOH (3  25 ml). Subsequent drying at 60
C in air gave SnO2 NPs.

(1)

#1 2

(2)

Goodness-of-fit was monitored by comparing Rwp with the expected residue, Rexp:

" Rexp ¼

ðN  PÞ P 2 i wi Yio

#1 2

(3)

Here, wi is the statistical weight, Yio and Yic are observed and calculated XRD intensities, N is the weight and number of experimental observations and P the number of fitting parameters. To estimate coherently diffracting domain (crystallite) size and r.m.s. lattice strain, the XRD profile was fitted using a pseudo Voigt (pV)

A. Kar et al. / Journal of Alloys and Compounds 810 (2019) 151718

function. This is a combination of Cauchyian- and Gaussian-type functions that better models crystallite size and strain broadening [31]. MAUD (Version 2.33) [32] was used for the refinement procedure [33,25b].

3

3. Results and discussion 3.1. TEM analysis Hydrothermally prepared SnO2 NPs are shown representatively

2.2.4. Raman spectroscopy Data were collected at room temperature on a Thermo Scientific DXR Raman microscope using a helium-neon laser with an excitation wavelength of 532 nm (laser power 10 mW).

2.2.5. X-ray photoelectron spectroscopy (XPS) XPS measurements were obtained using a Thermo Scientific KAlpha system. Binding energy was calibrated internally with respect to the C1s line.

2.2.6. UV-diffuse reflectance spectroscopy (UV-DRS) Data were collected on a Varian Cary-50 UVeVis spectrophotometer with a Harrick Video-Barrelino diffuse reflectance probe.

2.2.7. Brunauer-Emmett-Teller (BET) surface area analysis BET surface areas were obtained by nitrogen adsorption in a Nova 2000 Quantachrome sorptiometer 3000 analyzer. Samples were degassed at 77 K before measurements were commenced. Surface areas were determined by a multipoint BET method using the adsorption data in the P/P0 range 0.0e1.0, where P and P0 are the equilibrium and saturation pressures of adsorbates at the temperature of adsorption.

2.3. Photocatalytic testing The degradation of aqueous methylene blue (MB) under simulated solar irradiation was studied without sacrificial reagents. All experiments were done in triplicate (n ¼ 3). Typically, 5.0 mg of catalyst was added to 25 ml of a 3.0  105 M aqueous dye solution (pH 7), and the mixture stirred in the dark for 30 min to allow dye adsorption to the catalyst surface [34]. A 3.0 ml aliquot was centrifuged and the absorption of the supernatent determined to give the dye concentration before photocatalysis (C0). The remaining solution was irradiated with a 300 W xenon lamp (Solar Simulator model LSO306, AM 1.5 G filter, 1 sun illumination at 100 mW cm2). Degradation of lmax was monitored by UVevis spectroscopy (PerkinElmer LAMBDA 265) to obtain the concentration (C) of dye as a function of time in subsequent aliquots. During irradiation, samples resided in an ice bath to prevent evaporation or thermal degradation. Loss of MB was measured and calculated according to: Degradation (%) ¼ (1  C/C0)  100

(4)

Hydroxyl radicals (OH) created during photocatalysis were estimated via a terephthalic acid (TA) probe. In a typical process, 5.0 mg catalyst was dispersed in an aqueous solution of 30 ml TA (5  104 M) and NaOH (2  103 M). The resulting suspension was exposed to simulated solar irradiation and at regular intervals, 3.0 ml of the suspension was collected and centrifuged and the maximum fluorescence emission intensity measured with an excitation wavelength of 315 nm. The method relies on 2hydroxyterephthalic acid (TAOH) fluorescing at 425 nm. An Edinburgh Instruments FLS980 photoluminescence spectrophotometer was utilized.

Fig. 1. (a) TEM image (scale bar 20 nm) and corresponding size distribution (inset) for SnO2 NPs, (b) a representative HRTEM image (scale bar 5 nm) highlighting the distance between the lattice planes and (c) the corresponding FFT pattern.

4

A. Kar et al. / Journal of Alloys and Compounds 810 (2019) 151718

in Fig. 1a, along with HRTEM data (Fig. 1b) that resolves the crystallographic planes and the corresponding fast Fourier transform (FFT) pattern (Fig. 1c). The formation of monocrystalline NPs of uniform shape and size is clearly demonstrated. Mean particle size and distribution is 11.2 ± 1.0 nm (Fig. 1a inset) and the distance between the lattice planes (0.33 nm; Fig. 1b) agrees with the (110) plane of tetragonal SnO2 [35]. The FFT pattern is also consistent with tetragonal SnO2. TEM analysis revealed that attempts at the inclusion of CTAB in the synthesis of SnO2 NPs gives a product in which NP formation is accompanied by that of nanowedges (NWs). The latter superficially dominate low magnification TEM images by virtue of their size (typically 200e400 nm in length; Fig. 2a and also ESI Fig. S1y). The NPs are ca. 5e20 nm in size, though an accurate mean size distribution is difficult to achieve because they are capable of existing as individual NPs (ESI Fig. S2) or of forming dense agglomerates (Fig. 2a and b and ESI Fig. S3). Moreover, a closer inspection of highresolution (HR) TEM images reveals that the narrow ends of the NWs are based upon regions of agglomerated NPs (Fig. 2c (and righthand expansions) and ESI Fig. S4). Typically, the NWs exhibit smooth edges and are ca. 400 nm long. HRTEM analysis of the edge of one such NW evidences monocrystallinity and reveals lattice fringes (0.26 nm separation) that index as the (101) plane of tetragonal SnO2 (Fig. 2c, and left-hand expansions). This appearance contrasts starkly with that of the tip of the NW. Taken together, these data lead us to the view that the monocrystalline NPs are the primary nanocrystallites that seed formation of the much larger NWs; a thesis that gains traction in the light of Rietveld analysis (below). In contrast to the nanomaterials prepared hydrothermally, TEM analysis in Fig. 3aeb establishes that SnO2 prepared in the presence of CTAB using microwave irradiation exists as essentially uniformly-sized and shaped nanospheres (NSs) with diameters approximating to 200e300 nm [26]. HRTEM data (Fig. 3c) reveals that these NSs express lattice planes consistent with tetragonal SnO2 [25,36]. However, in contrast to NW formation, that of the NSs is attributable to the agglomeration of very small (ca. 3e4 nm) SnO2 NPs. This correlates well with the results of Rietveld analysis (below). The FFT pattern (Fig. 3d) of the region highlighted in Fig. 3c points to highly crystalline tetragonal SnO2, with the diffraction

Fig. 3. (a, b) TEM images of SnO2 NSs (scale bars 50 nm), (c) representative HRTEM image of a NS revealing fine-structure (scale bar 5 nm) and (d) the FFT pattern of the circled region in (c).

spots indexing as the (110) and (211) planes.

3.2. XRD phase identification XRD patterns of each of the three types of SnO2 sample revealed by TEM (see above) are fitted by Rietveld refinement [27]. Results are shown in Fig. 4; each pattern is indexed as tetragonal SnO2 (ICSD database code 154960; space group P42/mnm; a ¼ 4.7311 Å, c ¼ 3.1815 Å), pointing to the formation of pristine SnO2. Fig. 4a shows observed (Io), calculated (Ic) and residual (Io-Ic) patterns for SnO2 NPs. Reflections in Fig. 4a are sharper than those for SnO2 NSs (Fig. 4c) indicating that the primary nanocrystallites in the NPs are,

Fig. 2. (a) Representative TEM image of SnO2 NWs coexisting with agglomerated NPs; (b) an isolated agglomerate of NPs; (c) A further example of NW formation with HRTEM analysis of the terminus revealing a high density of agglomerated NPs while that of the main body of the NW reveals monocrystalline appearance with lattice fringes (0.26 nm apart) indexing as the (101) plane of tetragonal SnO2.

A. Kar et al. / Journal of Alloys and Compounds 810 (2019) 151718

Fig. 4. Indexed and simulated XRD patterns for nanostructured SnO2: (a) NPs, (b) NWs and (c) NSs. Red dots and black lines represent observed (Io) and calculated (Ic) intensities, respectively. (Io-Ic) shows post-fitting residual signal. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

in fact, larger than those in the NSs. Rietveld analysis reveals that for the NPs, lattice parameters a and c are larger and marginally smaller, respectively, with respect to those of standard (bulk) SnO2 (Table 1). Primary nanocrystallite size and r.m.s. lattice strain for the NPs are found to be 12.91 nm and 5.27  103, respectively. The close match between primary nanocrystallite size and individual NP size observed by TEM establishes individual NPs to be monocrystalline. Rietveld analysis also suggests oxygen deficiencies in the SnO2 NP unit cell (Table 1). Fig. 4b and Table 1 show the XRD patterns and lattice parameters of the SnO2 NWs. The suggestion, based on TEM data, that ca. 5e20 nm NPs seed the production of gross NW structures when SnO2 synthesis is conducted in the presence of CTAB is reinforced by XRD analysis. This reveals a primary nanocrystallite size of 12.67 nm. At 7.67  103, r.m.s. lattice strain in the NWs is computed to be significantly greater than in the NP sample. This can be explained by considering NW formation to be attributable to the agglomeration of a large number of primary nanocrystallites, each displaying surface anionic, oxygen vacancies. Rietveld analysis estimates the level of oxygen deficiency to be comparable to that in the NP sample. These vacancies logically furnish the nanocrystallites with an overall cationic charge, making them capable of agglomerating through weak hydrogen-like polarization bonding. This coupling mechanism has recently been shown to yield large spherical nanoobjects [33]. The coupling of nanocrystallites with similar cationic charge is synonymous with the development of r.m.s strain at the inter-particle junction since the lattice strain inherent to the crystal lattices of the individual nanocrystallites can no longer propagate to the agglomerate surface for annihilation. As

5

a result, r.m.s lattice strain in the NWs is higher than that in the NPs, even though primary nanocrystallite size is comparable in either sample. XRD patterns for SnO2 NSs are shown in Fig. 4c. The FWHM of all reflections are larger than those of NPs and NWs, reinforcing the view from TEM data that NSs in fact comprise smaller primary nanocrystallites. In spite of the appearance of non-uniform broadening of individual peaks, all peaks can be viewed as having uniformly broadened on account of peak overlap. For example, FWHM values for the (211) and (301) peaks are larger compared to those of (110) and (101) on account of overlap with adjacent low intensity peaks (e.g. (220), (002), (310), (112) and (320)). Consistent with reduced primary nanocrystallite size, the values of both a and c lattice parameters of SnO2 NSs are relatively large. Primary nanocrystallite size and r.m.s. lattice strain are calculated to be 2.19 nm and 1.2  103, respectively. This relatively low value of r.m.s. lattice strain is characteristic of very small individual nanocrystallites where bulk defects are unlikely to be stabilized, instead being driven to the nanocrystallite surface [37]. The calculated value of crystallite size shows excellent agreement with TEM observations. Moreover, consistent with the values of both size and r.m.s. strain, the SnO2 NS unit cell demonstrates significantly reduced levels of oxygen vacancy as compared with NP and NW unit cells (Table 1). 3.3. Atomic modeling ESI Fig. S5 shows the modelled structure of bulk SnO2. It reveals regular SnO6 octahedra and atomic coordinates for Sn and O of (0, 0, 0) and (0.30660, 0.30660, 0), respectively, suggesting Sn atoms occupy corner positions in the unit cell. Fig. 5a shows calculated bond lengths in SnO2 NPs. Whereas SneO bonds in bulk SnO2 are all computed to be 2.05 Å with a non-bonding Sn,,,Sn distance of 3.18 Å, SnO2 NPs see all SneO bonds increase to 2.06 Å. Meanwhile, non-bonding Sn,,,Sn distances remain unchanged. Fig. 5b shows also changes in SnO2 NWs, with SneO bonds in the equatorial plane increasing from 2.05 Å to 2.10 Å, and axial SneO bonds decreasing

Fig. 5. Changes in SneO bond lengths with respect to bulk SnO2 (ICSD database code 154960) in SnO2 NPs (a) and NWs (b).

Table 1 (Micro)structural parameters for nanostructured SnO2. SnO2 (Tetragonal, Sp. Gr. P42/mnm)

NPs NWs NSs

Lattice parameters (Å) a

c

4.7526 4.7575 4.8058

3.1845 3.1811 3.2622

Particle size (nm)

12.91 12.67 02.19

r.m.s. lattice strain (  103)

5.27 7.67 1.20

O-site occupancy

0.5509 0.5295 0.9960

Fractional coordinates (O) x

y

z

0.2697 0.2967 0.306

0.3218 0.2956 0.305

0.0 0.0 0.0

6

A. Kar et al. / Journal of Alloys and Compounds 810 (2019) 151718

to 1.99 Å. Once again, Sn,,,Sn non-bonding distances remain unchanged. In contrast, the NSs show SneO and Sn,,,Sn bonds unaltered with respect to bulk SnO2. Changes to OeSneO bond angles between bulk and nanostructured SnO2 are shown in Fig. 6. Significant changes are observed in the equatorial plane of the SnO6 octahedron only for the NWs; 101.5
and 78.5
(NPs) and 98.5
and 81.5
(NWs) compare with 101.7
and 78.3
(bulk). The directional elongation/compression of SneO bonds in the NWs is attributable to the polarization coupling that takes place between adjacent NPs during NW formation. Evidently, this interaction between multiple NPs causes the spherical charge distribution between cation and anion at the NP surface to become ellipsoidal. 3.4. BET surface area analysis For heterogeneous photocatalysis, surface area plays a vital role [38,39]. In this work, BET surface areas and pore size distributions of each sample were determined by measuring nitrogen adsorption-desorption isotherms (Fig. 7). Both the adsorption and desorption branches of each isotherm formed a hysteresis loop characteristic of type IV behaviour. This suggests interconnected mesoporous networks with disordered and inhomogeneous size distributions resulting from the aggregation of primary nanocrystallites. Specific BET surface areas were found to be 97.6, 79.6 and 10.1 m2g1 for the NPs, NWs and NSs, respectively (Table 2). The higher surface area recorded for the NPs is rationally consistent with there being more surface defect sites available to capture the photoinduced holes and thus retard electron-hole recombination and enhance photocatalytic activity.

Fig. 7. Nitrogen adsorption-desorption isotherms at 77 K (P and P0 are the equilibrium and the saturation pressures of N2 at the temperature of adsorption) for different SnO2 morphologies.

Table 2 Nitrogen sorption porosimetry studiesa for nanostructured SnO2. Photocatalyst

SBET (m2g1)

Pore volume (cm3g1)

Mean pore size (nm)

NPs NWs NSs

97.6 79.6 10.1

0.08301 0.05711 0.03405

18.1 29.3 10.9

a Surface areas determined by the BET technique. Pore volume and mean pore diameters are obtained by applying Barrett-Joyner-Halenda theory to the adsorption branch [62].

3.5. Raman spectroscopy Fig. 8 illustrates the Raman spectra of the three SnO2 samples. Each reveal peaks at 472e486, 627e631 and 765-775 cm1 consistent with Eg, A1g and B2g vibrational modes, respectively [40]. These are all consistent with the tetragonal rutile structure [41]. Compared to previously reported bulk SnO2 [42], the Eg and A1g modes showed minor shifts attributable to size effects [43]. Meanwhile, an additional peak observed at 573-575 cm1 is due to surface defects [44]. This dominates the spectrum for the NP sample; consistent with the high proportion of surface area demonstrated by this sample and indicating a correspondingly high level of surface oxygen vacancies. However, a comparison of the relative strength of this band with respect to A1g in the NWs and NSs is highly informative. In the NWs (SBET ~82% that of the NPs) the defect peak is comparatively weak whereas in the NSs it is relatively strong in spite of their low SBET value (~10% that of the NPs). This lets us refine the model suggested by BET surface area analysis, which suggested a straightforward correlation between SBET and the number of surface defect sites available. Hence, while NPs

Fig. 8. Raman spectra of SnO2 NSs (a), NWs (b) and NPs (c) (Spectra are as acquired and not normalised).

Fig. 6. OeSneO bond angles in tetragonal SnO2 in (a) NPs and (b) NWs.

demonstrate significant numbers of surface defects by virtue of their high value of SBET, NSs appear to do so in spite of a relatively low SBET. In contrast, NWs exhibit a relatively high SBET but relatively low number of surface defects. Maintaining the view that surface defects encourage catalytic activity, this suggests that NPs will prove active. However, NSs may show an impressive catalytic activity per unit surface area, whilst that of NWs is likely to prove

A. Kar et al. / Journal of Alloys and Compounds 810 (2019) 151718

poor. 3.6. XPS analysis Fig. 9a and b shows XPS peaks characteristic of Sn3d and O1s levels in the SnO2 NPs [45]. Peaks at 486.3 and 494.7 eV correspond to Sn3d5/2 and Sn3d3/2, respectively [46], and are consistent with Sn(IV) (indexed Standard ESCA Spectra of the Elements and Line Energy Information, ɸCo., USA) [47]. An O1s peak is seen at 530.1 eV in Fig. 9b [45]. Corresponding data for SnO2 NWs and NSs are shown in Fig. 9c and d and Fig. 9e and f, respectively. In none of the samples (NPs, NWs, NSs) were any remnants of CTAB detected. The most noteworthy feature that was observed was the growth

7

through Fig. 9b, d and f of a high binding energy shoulder on the O1s line. This is most clearly apparent in the NS sample (Fig. 9f). We are not aware of any reports that unambiguously correlate the effect of oxygen vacancies on binding energies. However, current data suggest a dominant O1s line in each sample attributable to lattice oxide [48], with the NS sample most clearly including a shoulder of a type documented to be enhanced by atmospheric exposure [49], and consistent with OH [50] or CO2 3 [51] groups saturating surface vacancies [52]. The apparent contradiction between this data and the Rietveld analysis above can be explained if the NSs are viewed as exhibiting a large number of oxygen vacancies only at or about their exterior surface. The large number of oxygen vacancies being proposed at the NS surface can be viewed as being more than

Fig. 9. Selected XPS data for SnO2 NPs (binding energy spectra for (a) Sn3d; (b) O1s), NWs ((c) Sn3d; (d) O1s) and NSs ((e) Sn3d; (f) O1s).

8

A. Kar et al. / Journal of Alloys and Compounds 810 (2019) 151718

offset by the inherently low surface:volume ratio of individual nanospheres in comparison to that of the isolated particles in the NP sample. These arguments are consistent with both the Raman spectroscopic data (see above) and, below, the photoactivities of the samples. 3.7. UV-DRS UV-DRS was used to differentiate the optical absorptions of the different morphologies of SnO2. In each case, the absorption edge corresponds to electron transitions from the valence band to the conduction band and this edge can be used to calculate the optical band gap (Eg). The optical absorption coefficient, a, of a semiconductor close to the band edge is expressed by [53].



a ¼ k hy  Eg

n . hy

(5)

Where k is a constant and n depends on the nature of the transition. In this case, n ¼ ½. Eg can be estimated from the corresponding Tauc plot. Fig. S6a shows data for all three morphologies, with NSs and NWs being red shifted compared to NPs. Estimation of Eg from the corresponding Tauc plots (Fig. S6b) has given values of 3.95, 3.75 and 3.93 eV for NPs, NWs and NSs, respectively. 3.7.1. Photocatalytic activity tests Photocatalytic activity was evaluated for the treatment of aqueous MB with nanoscopic SnO2 under simulated solar irradiation. Fig. 10 illustrates representative changes in MB lmax ¼ 664 nm as a function of time for different SnO2 morphologies, with the catalytic efficiency maximized for use of the NP photocatalyst. Fig. 11a shows the decomposition (±SD) of MB in each test. Outstandingly, 98.0 (±0.7) % of MB was degraded in the presence of SnO2 NPs after 120 min, a photocatalytic performance significantly exceeding that reported recently in the literature [54]. This contrasts with 58.0 (±2.5) % and 33.0 (±3.7) % MB degradation for the use of SnO2 NWs and NSs, respectively. To quantitatively evaluate photocatalytic activity, reaction rate constants (k) were estimated assuming pseudo first-order kinetics and a low initial pollutant concentration [55]. Fig. 11b confirms first order kinetics, with k (±SD) ¼ 31.5  103 (±0.3  103), 7.0  103 (±0.2  103) and 3.2  103 (±0.1  103) min1 for NPs, NWs and NSs, respectively. On the face of it, this suggests a ca. 10-fold enhancement in activity for the SnO2 NPs relative to NSs. Reference experiments i) with SnO2 NPs but without light, and ii) with light in the absence of catalyst were performed. Neither showed significant degradation at lmax ¼ 664 nm (see ESI Fig. S7), though information on MB absorption by the NP catalyst (i.e. that with the highest SBET) in a single experimental cycle was presented. Consequently, 15 and 6% decreases were noted in lmax for reference experiments i and ii, respectively. These data compare with the 98.0 (±0.7) % drop in lmax for irradiated NPs described above. A more nuanced picture of the activity emerges, however, if the catalytic activity per unit surface area is considered (Table 3). NPs remain the most effective catalyst tested by virtue of their high surface area (which brings with it a large number of surface defects, see Raman spectroscopy (Fig. 8)), with an activity of 1.16  104 mol dm3 s1 m2 (see Table 3). However, the NSs perform comparably per unit surface area. This is consistent with the vision that, in spite of their modest surface area, they present a large population of surface defects (see again Raman spectroscopy (Fig. 8)). The detrimental effect of the NWs presenting a considerable surface area but, according to Raman data, a small number of surface defects can additionally be seen; they present both poor activity and poor activity per unit surface area.

Fig. 10. Changes to MB lmax ¼ 664 nm catalyzed by SnO2 NPs (a), NWs (b) and NSs (c).

A. Kar et al. / Journal of Alloys and Compounds 810 (2019) 151718

9

Fig. 11. (a) Plot of C/C0 (%) (where C0 and C are the concentrations of dye before and after irradiation, respectively) for MB as a function of irradiation time in the presence of SnO2 NPs (green), NWs (red) and NSs (blue) (mean ± SD, n ¼ 3). (b) Plot of ln (C0/C) as a function of irradiation time. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Table 3 Photocatalytic activity in the degradation of MBa by nanostructured SnO2. Photocatalyst

SBET of catalyst used (m2)b

k (  103 min1)

k (  101 s1)

Activity per unit surface area (  104 mol dm3 s1 m2)

NPs NWs NSs

0.488 0.398 0.051

31.5 7.0 3.2

18.9 4.2 1.9

1.16 0.32 1.11

a b

[MB] ¼ 3  105 M. Quantity of catalyst 0.005 g.

The photocatalytic activity of the SnO2 NPs has been compared with that of standard TiO2 Degussa P25 powder. Fig. S8a (see ESI) demonstrates that 5.0 mg of either catalyst induce comparable decreases in lmax for MB, with photocatalytic efficiencies of 98.5% (NPs) and 99.1% (P25) recorded in concurrent tests (Fig. S8b). To investigate the stability and reusability of the SnO2 NPs, cycling tests on the degradation of MB were commenced with the catalyst recovered by centrifugation between tests. As shown in Fig. S9a, after three cycles of MB photodegradation, the catalyst showed only a slight loss of efficiency; 98.4, 86.7, and 84.5% of MB was degraded in the first, second, and third cycles, respectively. Fig. S9b demonstrates the XRD patterns of photocatalytic NPs before and after three MB decomposition cycles, with the lack of observable changes suggesting that both crystalline phase and structure remain intact. Taken together, these data suggest that the difficulties associated with complete recovery of catalyst during centrifugation may be responsible for the modest loss in performance. The photocatalytic properties of SnO2 depend on its ability to produce electron-hole pairs [56] and of these to induce the creation of radicals [57], e.g. by the excited electrons forming superoxide (O 2 ) and the holes reacting to form hydroxyl radicals (OH) that can quench pollutants [58] (such as MB) according to:

þ 2SnO2 (e CB) þ O2þ 2H /2SnO2 þ H2O2  SnO2 (e CB) þ H2O2/SnO2 þ OH þ OH

MB þ O2 þ OH/intermediates/degraded products A terephthalic acid (TA) probe [59] was used to verify that the photocatalytic degradation of MB proceeded here through the photoinduced creation of OH. Upon excitation at 315 nm, the maximum intensity at 425 nm gradually increased as irradiation time (see ESI Fig. S10), evidencing OH photogeneration. The literature already recognises that O2 ions can escape from a host SnO2 lattice, leading to oxygen vacancies or defects (Vo) [60]. Moreover, these vacancies can trap electrons, leading to the formation of a Vo state [40]. On photoexcitation, positive holes and negative electrons are created in the valence (VB) and conduction (CB) bands of SnO2, respectively, enabling two types of relaxation  process. Firstly, a hole in the VB (hþ VB) and an electron in the CB (eCB) can radiatively recombine but secondly, a hole in the VB can be trapped by an oxygen vacancy [40] to inhibit charge-carrier recombination and improve photocatalytic activity [61]. In comparison to the other morphologies tested, the SnO2 NPs present the most surface defects and this is reflected in their superior performance.

þ SnO2 þ hv/SnO2 (e CB þ hVB) þ þ  SnO2 (hþ VB) þ H2O/SnO2 (hVB) þ H þ OH  SnO2 (hþ VB) þ OH /SnO2 þ OH  SnO2 (e CB) þ O2/O2 þ SnO2

4. Conclusions In summary, we have synthesized three distinctly different samples of nanostructured SnO2 in order to test the influence of morphology and agglomeration on photocatalytic properties. Surface defects and oxygen vacancies in each morphology were inspected by Raman spectroscopy and Rietveld analysis,

10

A. Kar et al. / Journal of Alloys and Compounds 810 (2019) 151718

respectively. Results correlated well with performance in the photocatalytic degradation of MB; monocrystalline SnO2 NPs proved most effectual, though NSs demonstrated a comparable activity per unit surface area on account of a high density of surface defect sites as evidenced by both Raman spectroscopy and XPS. The activity reported here is significantly greater than that previously reported for SnO2 NPs [54]. These data are attributed to high populations of surface defect sites in the NPs and NSs promoting the capture of photoinduced holes and the subsequent retardation of electron-hole recombination. Recycling tests confirm that the stability of the SnO2 NP photocatalysts reported here is significant, though it is necessary to consider the possible effect of incomplete catalyst recovery during centrifugation steps on long-term catalyst efficiency. The further development of these SnO2 NPs as potential photocatalysts is therefore contingent upon their immobilization. The development of novel porous matricies for this purpose is now being initiated.

[12] [13] [14] [15] [16] [17] [18] [19]

[20]

[21]

Acknowledgements A. K. acknowledges support from the Royal Society's Newton International Fellowship (NF130808) and Alumni follow-on funding schemes. S. S. acknowledges an RA-II Fellowships Grant from IACS, Kolkata. A. F. and O. M acknowledge the Spanish Ministry MCIU for financial support (CTQ2015-65918-R and RTI2018093871-B-I00, MCIU/AEI/FEDER, UE ) and the Laboratory for Nanoscopies and Spectroscopies LANE at the ICMS for TEM analysis. Thanks go also to Drs. Tim Jones and Jill Geddes of Schlumberger Gould Research and Mr. Joshua Mehta and Dr. Da Shi (Cambridge) for help with Raman, X-ray photoelectron and UV-DR spectroscopies, and also to Dr. Zlatko Saracevic of the Department of Chemical Engineering and Biotechnology (Cambridge) for help with BET surface area analysis. We thank the reviewers for constructive ideas on comparing catalyst activity per unit surface area. Data for this paper are available at the University of Cambridge data repository (see https://doi.org/10.17863/CAM.42669).

[22]

[23]

[24]

Appendix A. Supplementary data [25]

Supplementary data to this article can be found online at https://doi.org/10.1016/j.jallcom.2019.151718. References

[26] [27]

[1] (a) B.K.H. Yen, N.E. Stott, K.F. Jensen, M.G. Bawendi, Adv. Mater. 15 (2003) 1858e1862; (b) T. Hyeon, Chem. Commun. (2003) 927e934; (c) W.S. Seo, J.H. Shim, S.J. Oh, E.K. Lee, N.H. Hur, J.T. Park, J. Am. Chem. Soc. 127 (2005) 6188e6189. [2] (a) Y. Liu, Y. Jiao, Z. Zhang, F. Qu, A. Umar, X. Wu, ACS Appl. Mater. Interfaces 6 (2014) 2174e2184; (b) J.S. Chen, X.W. Lou, Small 9 (2013) 1877e1893. [3] A. Kar, A. Patra, Trans. Indian Ceram. Soc. 72 (2013) 89e99. [4] P.G. Harrison, M.J. Willet, Nature 332 (1988) 337e339. [5] W.W. Wang, Y.J. Zhu, L.X. Yang, Adv. Funct. Mater. 17 (2007) 59e64. [6] J. Zhu, Z. Lu, S.T. Aruna, D. Aurbach, A. Gedanken, Chem. Mater. 12 (2000) 2557e2566. [7] (a) Y. Wang, X. Jiang, Y. Xia, J. Am. Chem. Soc. 125 (2003) 16176e16177; (b) G. Xi, J. Ye, Inorg. Chem. 49 (2010) 2302e2309; (c) S. Sain, A. Kar, A. Patra, S.K. Pradhan, CrystEngComm 16 (2014) 1079e1090. [8] Z.R. Dai, Z.W. Pan, Z.L. Wang, J. Am. Chem. Soc. 124 (2002) 8673e8680. [9] A. Kar, S. Sain, S. Kundu, A. Bhattacharyya, S.K. Pradhan, A. Patra, ChemPhysChem 16 (2015) 1017e1025. [10] (a) M. Kong, Y. Li, X. Chen, T. Tian, P. Fang, F. Zheng, X. Zhao, J. Am. Chem. Soc. 133 (2011) 16414e16417; (b) J.L. Yang, S.J. An, W.I. Park, G.C. Yi, W. Choi, Adv. Mater. 16 (2004) 1661e1664; (c) S. Wu, H. Cao, S. Yin, X. Liu, X. Zhang, J. Phys. Chem. C 113 (2009) 17893e17898. [11] (a) S. Wu, H. Cao, S. Yin, X. Liu, X. Zhang, J. Phys. Chem. C 113 (2009) 17893e17898;

[28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38]

[39] [40] [41] [42] [43]

(b) S. Zhuang, X. Xu, Y. Pang, J. Hu, C. Yang, L. Tong, Y. Zhou, RSC Adv. 3 (2013) 20422e20428. A. Kar, A. Patra, J. Mater. Chem. C 2 (2014) 6706e6722. G. Wang, W. Lu, J.H. Li, J.Y. Choi, Y. Jeong, S.Y. Choi, J.B. Park, M.K. Ryu, K. Lee, Small 2 (2006) 1436e1439. A. Kar, S. Sain, S. Kundu, A. Bhattacharyya, S.K. Pradhan, A. Patra, ChemPhysChem 16 (2015) 1017e1025. H. Wang, A.L. Rogach, Chem. Mater. 26 (2014) 123e133. M. Wang, Y. Gao, L. Dai, C. Cao, X. Guo, J. Solid State Chem. 189 (2012) 49e56. P.H. Suman, A.A. Felix, H.L. Tuller, J.A. Varela, M.O. Orlandi, Sens. Actuators B 208 (2015) 122e127. T. Zima, I. Bataev, J. Solid State Chem. 243 (2016) 282e289. (a) T. Tao, L. He, J. Li, Y. Zhang, Mater. Lett. 138 (2015) 45e47; (b) Y.-C. Chang, J.-C. Lin, S.-Y. Chen, L.-Y. Hung, Y.-R. Lin, C.-Y. Chen, Mater. Res. Bull. 100 (2018) 429e433; (c) Y. Zhu, T. Cao, Z. Li, C. Chen, Q. Peng, D. Wang, Y. Li, Sci. China Mater. 61 (2018) 1527e1535. (a) D.R. Chu, J.H. Mo, Q. Peng, Y.P. Zhang, Y.G. Wei, Z.B. Zhuang, Y.D. Li, ChemCatChem 3 (2011) 371e377; (b) S. Wu, H. Cao, S. Yin, X. Liu, X. Zhang, J. Phys. Chem. C 113 (2009) 17893e17898. (a) K. Prakash, P.S. Kumar, S. Pandiaraj, K. Saravanakumar, S. Karuthapandian, J. Exp. Nanosci. 11 (2016) 1138e1155; (b) P.V. Viet, C.M. Thi, L.V. Hieu, J. Nanomater. (8) (2016) 4231046; (c) P. Li, Y. Lan, Q. Zhang, Z. Zhao, T. Pullerits, K. Zheng, Y. Zhou, J. Phys. Chem. C 120 (2016) 9253e9262; (d) S. Begum, M. Ahmaruzzaman, J. Photochem, B. Photobiol, Biology 184 (2018) 44e53. (a) N. Talebiann, F. Jafarinezhad, Ceram. Int. 39 (2013) 8311e8317; (b) J. Wang, H. Fan, H. Yu, Optik 127 (2016) 580e584; (c) X. Chen, D. Chu, L. Wang, W. Hu, H. Yang, J. Sun, S. Zhu, G. Wang, J. Tao, S. Zhang, J. Alloy. Comp. 729 (2017) 710e715; (d) X. Chen, D. Chu, L. Wang, W. Hu, H. Yang, J. Sun, S. Zhu, G. Wang, J. Tao, S. Zhang, J. Alloy. Comp. 768 (2018) 517e524. (a) M. Guan, X. Zhao, L. Duan, M. Cao, W. Guo, J. Liu, W. Zhang, J. Appl. Phys. 114 (2013) 114302; (b) S.A.H. Juybari, H.M. Moghaddam, Appl. Surf. Sci. 457 (2018) 179e186; (c) A. Sadeghzadeh-Attar, Solar Energy Mater. Sol. Cells 183 (2018) 16e24. (a) X. Han, M. Jin, S. Xie, Q. Kuang, Z. Jiang, Y. Jiang, Z. Xie, L. Zheng, Angew. Chem. Int. Ed. 48 (2009), 9180-9180; (b) X. Wang, X. Han, S. Xie, Q. Kuang, Y. Jiang, S. Zhang, X. Mu, G. Chen, Z. Xie, L. Zheng, Chem. Eur J. 18 (2012) 2283; (c) H. Wang, K. Dou, W.Y. Teoh, Y. Zhan, T.F. Hung, F. Zhang, J. Xu, R. Zhang, A.L. Rogach, Adv. Funct. Mater. 23 (2013) 4847e4853; (d) C. Wang, D. Cai, B. Liu, H. Li, D. Wang, Y. Liu, L. Wang, Y. Wang, Q. Li, T. Wang, J. Mater. Chem. A 2 (2014) 10623e10628; (e) G. Zhou, X. Wu, L. Liu, X. Zhu, X. Zhu, Y. Hao, P.K. Chu, Appl. Surf. Sci. 349 (2015) 798e804; (f) Z. Jin, M. Yang, S.-H. Chen, J.-H. Liu, Q.-X. Li, X.-J. Huang, Anal. Chem. 89 (2017) 2613e2621. (a) M. D'Arienzo, D. Cristofori, R. Scotti, F. Morazzoni, Chem. Mater. 25 (2013) 3675e3686; (b) A. Kar, S. Sain, S. Kundu, A. Bhattacharyya, S.K. Pradhan, A. Patra, ChemPhysChem 16 (2015) 1017e1025; (c) Y. Yang, Y. Wang, S. Yin, Appl. Surf. Sci. 420 (2017) 399e406. A. Kar, S. Sain, D. Rossouw, B.R. Knappett, S.K. Pradhan, A.E.H. Wheatley, J. Alloy. Comp. 698 (2017) 944e956. (a) H.M. Rietveld, Acta Crystallogr. 22 (1967) 151e152; (b) H.M. Rietveld, J. Appl. Crystallogr. 2 (1969) 65e71; (c) R.A. Young, D.B. Willes, J. Appl. Crystallogr. 15 (1982) 430e438. R. A. Young (ed.), The Rietveld Method, 1996, International Union of Crystallography, Oxford University Press, New York. L. Lutterotti, P. Scardi, J. Appl. Crystallogr. 23 (1990) 246e252. G. Caglioti, A. Paoletti, F.P. Ricci, Nucl. Instrum. Methods 3 (1958) 223e228. (a) R.A. Young, D.B. Wiles, J. Appl. Crystallogr. 15 (1982) 430e438; (b) L. Lutterotti, P. Scardi, P. Maistrelli, J. Appl. Crystallogr. 25 (1992) 459e462. (a) L. Lutterotti, The maud program. http://maud.radiographema.com/; (b) L. Lutterotti, Nucl. Instrum. Methods Phys. Res. B 268 (2010) 334e340. S. Sain, A. Kar, A. Patra, S.K. Pradhan, CrystEngComm 16 (2014) 1079e1090. Y. Xu, C.H. Langford, Langmuir 17 (2001) 897e902. A. Kar, S. Kundu, A. Patra, RSC Adv. 2 (2012) 10222e10230. A. Kar, S. Sain, D. Rossouw, B.R. Knappett, S.K. Pradhan, A.E.H. Wheatley, Nanoscale 8 (2016) 2727e2739. J. Schiøtz, K.W. Jacobsen, Science 301 (2003) 1357e1359. (a) X. Xu, J. Zhuang, X. Wang, J. Am. Chem. Soc. 130 (2008) 12527e12535; (b) M. D'Arienzo, L. Armelao, A. Cacciamani, C.M. Mari, S. Polizzi, R. Ruffo, R. Scotti, A. Testino, L. Wahba, F. Morazzoni, Chem. Mater. 22 (2010) 4083e4089. J. Yu, G. Wang, B. Cheng, M. Zhou, Appl. Catal., B 69 (2007) 171e180. A. Kar, S. Kundu, A. Patra, J. Phys. Chem. C 115 (2011) 118e124. J.X. Wang, D.F. Liu, W.Y. Zhou, G. Wang, S.S. Xie, Solid State Commun. 130 (2004) 89e94. A. Kar, S. Sain, S. Kundu, A. Bhattacharyya, S.K. Pradhan, A. Patra, ChemPhysChem 16 (2015) 1017e1025. C. Xu, G.D. Xu, Y. Liu, X. Zhao, G. Wang, Scr. Mater. 46 (2002) 789e794.

A. Kar et al. / Journal of Alloys and Compounds 810 (2019) 151718 [44] G. Xi, J. Ye, Inorg. Chem. 49 (2010) 2302e2309. [45] L.M. Cukrov, P.G. McCormick, K. Galatsis, W. Wlodarski, Sens. Actuators B 77 (2001) 491e495. [46] L. Cao, H. Wan, L. Huo, S. Hi, J. Colloid Interface Sci. 244 (2001) 97e101. [47] Y.-D. Wang, C.-L. Ma, X.-D. Sun, H.-D. Li, Nanotechnology 13 (2002) 565e569. [48] X.-B. Li, X.-W. Wang, Q. Shen, J. Zheng, W.-H. Liu, H. Zhao, F. Yang, H.-Q. Yang, ACS Appl. Mater. Interfaces 5 (2013) 3033e3041. [49] J.-C. Dupin, D. Gonbeau, P. Vinatierb, A. Levasseur, Phys. Chem. Chem. Phys. 2 (2000) 1319e1324. [50] D. Pradhan, M. Kumar, Y. Ando, K.T. Leung, ACS Appl. Mater. Interfaces 1 (2009) 789e796. [51] M.W. Roberts, R.S.C. Smart, Chem. Phys. Lett. 69 (1980) 234e239. [52] (a) M. Foley, C. Ton-That, M.R. Phillips, Appl. Phys. Lett. 93 (2008) 243104; (b) K. Ghosh, M. Kumar, H.F. Wang, T. Maruyama, Y. Ando, Langmuir 26 (2010) 5527e5533. [53] A. Kar, A. Patra, J. Phys. Chem. C 113 (2009) 4375e4380. [54] (a) S. Wu, H. Cao, S. Yin, X. Liu, X.J. Zhang, J. Phys. Chem. C 113 (2009)

[55] [56] [57] [58] [59] [60] [61] [62]

11

17893e17898; (b) X. Liu, L. Pan, T. Chen, J. Li, K. Yu, Z. Sun, C. Sun, Catal. Sci. Technol. 3 (2013) 1805e1809. Y. Feng, L. Li, M. Ge, C. Guo, J. Wang, L. Liu, ACS Appl. Mater. Interfaces 2 (2010) 3134e3140. H. Hidaka, J. Zhao, E. Pelizzetti, N. Serpone, J. Phys. Chem. 96 (1992) 2226e2230. J. Li, W. Ma, C. Chen, J. Zhao, H. Zhu, X. Gao, J. Mol. Catal. A Chem. 261 (2007) 131e138. Y. Nosaka, S. Komori, K. Yawata, T. Hirakawa, A.Y. Nosaka, Phys. Chem. Chem. Phys. 5 (2003) 4731e4735. S. Khanchandani, S. Kundu, A. Patra, A.K. Ganguli, J. Phys. Chem. C 117 (2013) 5558e5567. N. Chiodini, A. Paleari, D. Dimartino, G. Spinolo, Appl. Phys. Lett. 81 (2002) 1702e1704. Z.R. Dai, Z.W. Pan, Z.L. Wang, J. Am. Chem. Soc. 124 (2002) 8673e8680. E.P. Barrett, L. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373e380.