Raspberry‐Like Microspheres of Core–Shell Cr2O3@TiO2 Nanoparticles for CO2 Photoreduction

Abstract To promote the interaction of p–n semiconductors, raspberry‐like microspheres of core–shell Cr2O3@TiO2 nanoparticles have been fabricated through a five‐step process. Raman spectroscopy of products calcined at various temperatures reveal that the titania shell causes crystal distortion of the Cr2O3 core, without changing the microstructures of the fabricated core–shell microspheres. In situ and time‐resolved synchrotron‐based powder XRD reveals the formation of monoclinic TiO2 in the fourth step, but these monoclinic TiO2 nanocrystals undergo a phase transition when the applied calcination temperature is above 550 °C. As a result, TiO2(B), a magnéli phase of Ti4O7 and Cr2Ti6O15 compounds, resulting from inner doping between Cr2O3 and TiO2, is formed. The close interaction of Cr2O3 and TiO2 forms a p–n junction that decreases the recombination of photogenerated electron–hole pairs, leading to enhanced production of CH4 by photocatalytic reduction of CO2.


Introduction
In contrast to the use of fossil fuels and associated adverse global environmental effects, solar energy has the potential to provide our energy demands if it can be efficiently harvested and transformed. The photocatalytic reduction of CO 2 with H 2 O to valuable hydrocarbons, such as methane or methanol, is promising to reduce CO 2 emissions, as well as offering renewable energy alternatives. [1] However, the efficiency of CO 2 photoreduction is still very low to date. To improve the conversion efficiency, the development of nanomaterials with well-defined sizes, shapes, crystal phases, structure and composition are becoming increasingly important. [2] Nanomaterials have attracted enormous attention owing to their interesting properties and applications in diverse areas, such as photocatalysis, [3] nanoelectronics, [4] and integrated catalysis. [5] The preparation of such materials is currently regarded as among the most challenging areas in chemistry. [2] In particular, core-shell nanostructures with conducive and versatile compositions and structures are highly desirable in certain applications as they are task-specific nanomaterials with multifunctional capabilities. [2,6] Moreover, nanomaterials with synergetic properties between the core and the shell have become a very important class for emerging applications, such as enhanced optical devices, [7] tailored magnetic materials, [8] energy storage materials, [9] fuel cells, [10] dye-sensitized solar cells [11] and many important catalytic [12] and photocatalytic reactions. [3b, 13] Chromium(III) oxide (Cr 2 O 3 , eskolaite), which is a p-type semiconductor with a band gap of approximately 3.5 eV, has been recently proposed and used as a photocatalyst or co-catalyst for different photocatalytic reactions. [3a, 14] Maeda et al. proposed that the oxidation and reduction reactions in the photocatalytic overall water splitting process could be separated by using core-shell Rh/Cr 2 O 3 nanostructures. [3a] To overcome the rapid recombination of photogenerated electrons and holes, Hu et al. proposed the coupling between Cr 2 O 3 and WO 3 to induce the formation of p-n junctions. [14a] Chen et al. suggested that Cr 2 O 3 could act as the surface holes trapper within the Cr 2 O 3 -carbon nanotubes/TiO 2 nanocomposite to reduce the recombination rate of the photogenerated electron-hole pairs. [15] A recent study proposed that the presence of oxygen vacancies within Cr 2 O 3 could be used for methanol synthesis. [16] The fabrication of the core-shell structured Cr 2 O 3 :P@fibrousphosphorus hybrid composites revealed that the core-shell structure not only can reduce the recombination of the photogenerated electron-hole pairs, but also can enhance the optical properties of the composite. [3b] Doping of P into Cr 2 O 3 at the interface of the fibrous-phosphorus and Cr 2 O 3 composite resulted in the extension of the absorption tail into the near IR region, which can neither be observed in Cr 2 O 3 nor in fibrousphosphorus. Moreover, Cao et al. reported that the inner doping of Cr on TiO 2 thin films could significantly enhance the photo(electro)catalytic water splitting efficiency. [17] Recently, Zhao and co-workers proposed that the use of core-shell structures offers an excellent system for light/chemical CO 2 photoreduction. For instance, the wrapping of reduced graphene oxide onto the Pt-decorated thin sheet TiO 2 exhibited an apparent quantum efficiency of 1.93 % in the CO 2 photore-To promote the interaction of p-n semiconductors, raspberrylike microspheres of core-shell Cr 2 O 3 @TiO 2 nanoparticles have been fabricated through a five-step process. Raman spectroscopy of products calcined at various temperatures reveal that the titania shell causes crystal distortion of the Cr 2 O 3 core, without changing the microstructures of the fabricated coreshell microspheres. In situ and time-resolved synchrotronbased powder XRD reveals the formation of monoclinic TiO 2 in the fourth step, but these monoclinic TiO 2 nanocrystals undergo a phase transition when the applied calcination temperature is above 550 8C. As a result, TiO 2 (B), a magnØli phase of Ti 4 O 7 and Cr 2 Ti 6 O 15 compounds, resulting from inner doping between Cr 2 O 3 and TiO 2 , is formed. The close interaction of Cr 2 O 3 and TiO 2 forms a p-n junction that decreases the recombination of photogenerated electron-hole pairs, leading to enhanced production of CH 4 by photocatalytic reduction of CO 2 .
[a] Dr. J. Z. Y. Tan, Prof. M. M. Maroto-Valer duction reaction to CH 4 . [18] The group also proposed the fabrication of core-shell bimetallic (i.e., Au@Pd and Pt@Ru) nanoparticles decorated on TiO 2 to enhance the optical properties of TiO 2 , reduce the recombination of photogenerated charges and improve the CO 2 adsorption capability for CO 2 photoreduction. [19] In addition, the product selectivity of CO 2 photoreduction can be tuned through adjusting the Au/Pd molar ratio on TiO 2 . [19b] In this work, we aim at achieving p-n nanojunctions between Cr 2 O 3 and TiO 2 to enhance the interaction between the p-and n-type semiconductors, and consequently, increase CO 2 photoreduction efficiency. Unlike the typical core-shell structure, which consists of a layer of shell coated on a core particle, novel raspberry-like microspheres were constructed herein from many core-shell Cr 2 O 3 @TiO 2 nanoparticles. A thin layer of TiO 2 was coated on the Cr 2 O 3 nanoparticles of the microsphere, establishing the p-n nanojunction between Cr 2 O 3 and TiO 2 . The microstructure and elemental analysis of the nanocomposite were examined to characterize the morphology and distribution of TiO 2 on the Cr 2 O 3 . Using in situ and time-resolved synchrotron-based powder X-ray diffraction (PXRD) and Raman spectroscopy, the change of crystal phase of the fabricated Cr 2 O 3 @TiO 2 nanocomposite with increasing calcination temperature was observed. The photocatalytic efficiencies of the pristine Cr 2 O 3 and the fabricated Cr 2 O 3 @TiO 2 were evaluated for CO 2 photoreduction.

Results and Discussion
The laboratory and synchrotron-based PXRD patterns of the as-prepared Cr 2 O 3 revealed high crystallinity and the peaks matched well with Cr 2 O 3 in the database (eskolaite, JCPDS No.: 38-1479; Figure 1 A,C). The laboratory-based diffraction patterns showed that the Cr 2 O 3 phase in the mixture had no observable changes in peak positions after calcination at all temperatures (Figure 1 A). The absence of the TiO 2 peak was probably due to the low amount of TiO 2 within the sample. When the calcination temperature increased from 400 to 850 8C, all the peaks of Cr 2 O 3 remained (Figure 1 A). However, the overall peak intensity of the laboratory-based PXRD spectra decreased, indicating a possible phase transition triggered in Cr 2 O 3 during the calcination treatment at different temperatures (this is discussed further below; Figure 1 A, c-e). [20] Moreover, a number of new peaks between 24.5-33.38 were observed in the laboratory-based PXRD. To study the effect of calcination and to elucidate the new peaks observed in the laboratory-based PXRD, Raman spectroscopy was carried out (Figure 1 B) and in situ and time-resolved synchrotron-based PXRD were conducted (Figure 1 C).
Raman spectroscopy is a powerful technique to probe the crystal lattice vibrations in the study of nanomaterials. [21] When the calcination temperature was increased to 550 8C, new weak peaks centered at approximately 113.4 and 437.9 cm À1 , which were assigned to TiO 2 (B), appeared (Figure 1 B). [22] Moreover, the shoulder peak at approximately 417 cm À1 was very likely a result of magnØli Ti 4 O 7 . [23] The peaks centered at 302.6, 349.2, 523.6, 551.2 and 611.3 cm À1 can be assigned to crystalline Cr 2 O 3 (Figure 1 B). [24] The pattern of these peaks did not change with increasing calcination temperatures, but the intensities did. The Raman intensity of the Cr 2 O 3 @TiO 2 nanocomposites increased with calcination temperature compared with the pristine Cr 2 O 3 (inset of Figure 1 B). Raman intensity has been used to investigate alterations of crystal structure, evaluating the distortions of the crystal structure. [25] It is proposed here that the host lattice, which was the core Cr 2 O 3 , was distorted by the shell titania during the calcination treatment. With the increase in calcination temperature, the extent of the distortion was enhanced, resulting a stronger intensity of the Raman spectra. Moreover, owing to the reorientation of Cr 2 O 3 crystals, new peaks centered at 446.2, 713.5, 847.6, 906.7 and 927.6 cm À1 emerged. However, this crystal reorientation or distortion would cause the decrease in crystallinity of Cr 2 O 3 . Hence, the laboratory-based PXRD intensity of the TiO 2 /Cr 2 O 3 nanocomposites decreased with the calcination temperature (Figure 1 A).
Throughout the in situ PXRD experiment, the strong characteristic peaks of Cr 2 O 3 at 9.38, 12.78 and 13.78 were indexed to the (0 1 2), (1 0 4), and (11 0) lattice planes, respectively (Figure 1 C). In addition, there was a broad peak centered at approximately 3.48, which was later shifted to approximately 4.68 when the temperature was elevated from 80 to 550 8C, corresponding to the monoclinic TiO 2 (JCPDS No.: 65-6429), which was observed previously in a TiO 2 nanocomposite calcined below 550 8C. [26] The shift of this broad peak during the in situ calcination process could be due to the thermal contraction of TiO 2 . When the in situ calcination temperature was further increased to 700 8C, this broad peak disappeared because the monoclinic TiO 2 was unstable above 550 8C. Beyond 550 8C, several new peaks appeared and they were assigned to TiO 2 (B) ( After the coating of SiO 2 and titanium(IV) butoxide (TBT), the surface of the microspheres became smoother, losing the raspberry-like microstructure (see the Supporting Information, Figure S1 a, b). Then, the SiO 2 protective layer was removed by using an alkaline hydrothermal etching assisted crystallization method, which also crystallized the titania layer, although the crystallinity was very weak (Figure 1 C starting from room temperature). Excessive TBT coated on the SiO 2 layer was also removed together with SiO 2 (Figure S1 b). As a result, the Cr 2 O 3 @TiO 2 microspheres recovered the raspberry-like microstructure (Figure 2 d). No significant change in the overall spherical morphology was observed after calcination. After the removal of SiO 2 and calcination treatment at 700 8C, the Cr 2 O 3 @TiO 2 sample exhibited highly dispersed and homogeneous distribution of titania within each microsphere (Figure 2 eh). The TEM-EDX (energy-dispersive X-ray) line scan profile revealed that the thickness of Ti was about 10 nm, coating on Cr 2 O 3 nanoparticle (Figure 2 i), which was also observed by TEM ( Figure S2). The successive coating of titania on each of the Cr 2 O 3 nanoparticles was promoted by the porous structure of the Cr 2 O 3 core (Figure 2 j).
After the removal of SiO 2 , the close contact of the core and shell layers (Figure 2 k,l), which were the Cr 2 O 3 and TiO 2 , respectively, formed the oxide composite (i.e., Cr 2 Ti 6 O 15 ) as revealed by the in situ PXRD (Figure 1 C). This was further evident from the HRTEM, in which the lattice spacing of the Cr 2 Ti 6 O 15 (1 0 8) facet with 3.02 , was observed in between the TiO 2 (0 0 2) and Cr 2 O 3 (1 0 4) facets (Figure 2 m).
The surface chemistry of 700-Cr 2 O 3 @TiO 2 was elucidated by using X-ray photoelectron spectroscopy (XPS). High-resolution Cr 2p spectra (Figure 3 A) exhibited two major components  centered at 576.9 and 586.9 eV attributed to Cr 3 + species from Cr 2 O 3 . [28] Minor components positioned at 579.9 and 589.1 eV corresponded to the Cr 6 + from the chromium precursor. [29] The high-resolution Ti 2p (Figure 3 B) spectrum presented two major components at 458.6 and 464.4 eV, which are assignable to Ti 4 + . [9] The presence of Ti 3 + from Ti 4 O 7 was exhibited in the small peak centered at 457.7 eV. [9] The O 1s spectra were numerically fitted with three types of surface oxygen centered at 530.3, 530.5 and 532.6 eV, which were assigned to Ti-O-Ti, Ti-O-Cr and O-Cr, respectively (Figure 3 C). [9,28] The as-prepared Cr 2 O 3 sample exhibited two absorption peaks at approximately 400 and 535 nm ( Figure S3), originating from the 4 A 2g ! 4 T 1g and 4 A 2g ! 4 T 2g electronic transitions in Cr 2 O 3 , respectively. [30] The former is characteristic of the Cr 3 + ions of six-coordinate geometry and the latter implies octahedral symmetry. A dramatic decline in the absorption intensity, particularly for the trough centered at about 535 nm (Figure S3), in the fabricated Cr 2 O 3 @TiO 2 samples was observed when the calcination temperature was increased from 400 to 850 8C. This phenomenon was very likely due to the distortion of the octahedral symmetry in the Cr 2 O 3 by the titania moiety, as evident from the formation of Cr 2 Ti 6 O 15 (Figure 1 C), during the calcination process. [31] The distortion of the octahedral symmetry slightly decreased the band gap energy of the fabricated Cr 2 O 3 @TiO 2 nanocomposites, which was derived from the diffuse reflectance spectra (Figure 3 D), from 3.0 to 2.8 eV (Figure 3 E). Although this variation in electronic property has been reported in the literature previously, the mechanism is still unclear and further studies are needed. [31] The photoreduction of CO 2 was conducted for 8 h for the Cr 2 O 3 sample and the calcined Cr 2 O 3 @TiO 2 samples (Figure 3 F). The photoproducts were analyzed after each hour. In the control experiments, no methane was produced in the absence of photocatalyst, water and CO 2 under UV light irradiation. When the photocatalyst was loaded into the reactor in the dark (0 h), no product was obtained. When the reaction was run with He (i.e., presence of photocatalysts and water under UV light irradiation), no conversion was observed. However, in the presence of photocatalyst, CO 2 , H 2 O, and UV light, the production of CO in a trace amount and CH 4 was observed (Table S1). The photoproduction of CH 4 was genuine, which was evident from the negligible weight loss of 400-Cr 2 O 3 @TiO 2 in the thermogravimetric analysis (TGA) spectrum ( Figure S4). The maximum cumulative photoproduction of CH 4 from CO 2 by the pristine Cr 2 O 3 was approximately 82 mmol g catalyst h À1 (Figure 3 F). The photoproduction of CH 4 was significantly enhanced when the core-shell structured Cr 2 O 3 @TiO 2 sample was used. The maximum cumulative photoproduction of CH 4 increased with the calcination temperature of the Cr 2 O 3 @TiO 2 samples, from approximately 105 mmol g catalyst h À1 for sample 400-Cr 2 O 3 @TiO 2 to about 168 mmol g catalyst h À1 for sample 700-Cr 2 O 3 @TiO 2 (AQE = 0.296 %; Figure 3 F). Further increasing the calcination temperature to 850 8C had a detrimental effect on the photoreduction performance, leading to only about 63 mmol g catalyst h À1 of cumulative CH 4 production. This was probably because when the calcination temperature was increased to 850 8C, the inner doping between TiO 2 and Cr 2 O 3 was greatly enhanced, resulting in crystal distortion and reorientation in the Cr 2 O 3 by the titania moiety, as observed in the Raman studies. As a result, the overall crystallinity of the 850-Cr 2 O 3 @TiO 2 sample was decreased, leading to an inferior photocatalytic activity. When using the optimized sample, 700-Cr 2 O 3 @TiO 2 , the photoreduction of CO 2 was further tested up to 16 h to investigate the durability of the photocatalyst. The cumulative production of CH 4 was stabilized at 171.4 mmol g catalyst h À1 , indicating the photoproduction rate of CH 4 decreases with time ( Figure S5).
The increase in the CH 4 production of the 400-, 550-and 700-Cr 2 O 3 @TiO 2 samples compared with the pristine Cr 2 O 3 was due to the core-shell microstructure, which resulted in a close contact of the p-n junction formed between Cr 2 O 3 and TiO 2 , as seen in the TEM-EDX line mapping. In addition, the inner doping at the interface between TiO 2 and Cr 2 O 3 , which resulted in the formation of Cr 2 Ti 6 O 15 , was speculated to enhance the photocatalytic activity, which was also reported previously. [17] The heterojunction formed thus increased the charge separation efficiency as evident from the photoluminescence spectra (PL; Figure S6). Based on the characterization results, the overall photocatalytic mechanism was proposed. Upon irradiation with UV light, electrons in TiO 2 and Cr 2 O 3 were excited from their respective valence bands to their conduction bands. The formation of p-n junctions in the core-shell microstructure shortened the migration distance of the photogenerated electrons. As a result, the photogenerated electrons in Cr 2 O 3 migrated to the conduction band of TiO 2 ; whereas the photogenerated holes accumulated at the valence band of Cr 2 O 3 . [14b, 32] The electrons then photoreduced the adsorbed CO 2 to CH 4 . [32] Meanwhile, the accumulated photogenerated holes underwent photooxidation of H 2 O.

Conclusions
Raspberry-like structured Cr 2 O 3 @TiO 2 microspheres were synthesized by using a five-step process involving hydrothermal and calcination treatments. The pristine Cr 2 O 3 microspheres were made up from many Cr 2 O 3 nanoparticles, which were coated with a thin layer of TiO 2 shell on each of the Cr 2 O 3 nanoparticles, forming a novel core-shell nanostructure within a microsphere. Different calcination temperatures (i.e., 400-850 8C) were applied to the Cr 2 O 3 @TiO 2 samples after the removal of the SiO 2 layer. When the applied calcination temperature was above 550 8C, inner doping at the interface of Cr 2 O 3 and TiO 2 nanoparticles through the formation of Cr 2 Ti 6 O 15 was observed in the Raman spectra and synchrotron-based in situ PXRD. The p-n heterojunction between Cr 2 O 3 and TiO 2 at the nanoscale level and the newly formed Cr 2 Ti 6 O 15 phase were proposed to enhance the CO 2 photocatalytic reduction efficiency with a maximum cumulative product of approximately 168 mmol g catalyst À1 h À1 when the Cr 2 O 3 @TiO 2 sample was calcined at 700 8C.
All the reagents were used as received. Milli-Q water (resistivity higher than 18.2 MW cm À1 ) was collected from a Millipore academic purification system.

Method
Raspberry-like microspheres consisting of core-shell Cr 2 O 3 @TiO 2 nanoparticles were fabricated by following a five-step process (Scheme 1), as described below.

Synthesis of Cr 2 O 3 microspheres
The synthesis of Cr 2 O 3 was adopted from a previous study. [33] Briefly, K 2 Cr 2 O 7 (0.5 mmol), acrylamide (1.0 mmol) and H 2 O (2.1 mol) were mixed and stirred until a homogeneous orange solution was obtained. The solution was then transferred into a 100 mL Teflonlined stainless-steel autoclave and treated at 180 8C for 12 h. After the hydrothermal treatment, the obtained dark-green powder was washed consecutively in a series of sequential washing steps alternating ethanol and water for 3 times. After drying the powder in an oven at 75 8C, the product was calcined at 900 8C with ramping 10 8C min À1 and maintained at 900 8C for 1 h.

Synthesis of Cr 2 O 3 @SiO 2 microspheres
The synthesis of Cr 2 O 3 @SiO 2 microspheres was prepared according to a previous study. [8b] First, Cr 2 O 3 microspheres (0.15 g) were dispersed in ethanol (3 mL). Second, the mixture was added into a round-bottom flask containing ethanol (280 mL), Milli-Q water (70 mL) and concentrated NH 3 (5 mL, 28 wt %) in an ultrasound bath and treated ultrasonically for 30 min. After that, TEOS (4 mL) was added dropwise (flow rate: 0.4 mL min À1 ). Then, the solution was kept under continuous mechanical stirring for 12 h at room temperature. Finally, Cr 2 O 3 @SiO 2 microspheres were separated by using a centrifuge (ThermoFisher Sarvall Primo Benchtop Centrifuge, 4000 rpm for 15 min), followed by washing consecutively in a series of sequential washing steps alternating ethanol and water for 3 times.

Synthesis of Cr 2 O 3 @SiO 2 @TBT microspheres
The obtained product of Cr 2 O 3 @SiO 2 microspheres was redispersed in ethanol (200 mL) and mixed with concentrated NH 3 (0.9 mL, 28 wt %) under ultrasound for 30 min. After that, TBT (2 mL) was added dropwise (flow rate: 0.4 mL min À1 ). Then, the solution was kept under continuous mechanical stirring for 24 h at 45 8C. The resulting Cr 2 O 3 @SiO 2 @TBT microspheres were separated from the so-

Synthesis of Cr 2 O 3 @titania microspheres
The removal of SiO 2 was conducted by using an alkaline hydrothermal etching assisted crystallization method.
[8b] The Cr 2 O 3 @SiO 2 @TBT product obtained in the previous step was mixed with NaOH solution (20 mL, 1.0 m). The solution was then transferred to a 100 mL Teflon-lined stainless-steel autoclave. The autoclave was heated at 150 8C for 24 h, and then allowed to cool down to room temperature in air. The product obtained was then immersed in aqueous HCl (100 mL, 0.1 m) for 20 min, and subsequently washed with Milli-Q water until the pH value was close to 7, and then was dried at 75 8C in an oven overnight.

Characterization
The morphology of the synthesized Cr 2 O 3 and the Cr 2 O 3 @TiO 2 samples after calcination at different temperatures was examined by field emission scanning electron microscopy (FE-SEM, Quanta 200 F FEI), transmission electron microscopy (TEM) and high-resolution (HR) TEM (FEI Titan Themis 200) operated at 200 kV. The TEM was equipped with an energy-dispersive X-ray spectroscopy (EDX) detector. To investigate the structures of the nanospheres, samples were embedded in TAAB 812 resin and sliced into approximately 90 nm thick sections. The crystallinity of the synthesized products was assessed by powder X-ray diffraction (laboratory-based PXRD, Bruker Advanced Diffractometer) equipped with Cu Ka source (l = 1.5418 ). In situ and time-resolved synchrotron-based PXRD studies were carried out at the Australian Synchrotron at the Powder Diffraction beamline. The X-ray energy was 21 keV, and the wavelength (l = 0.590928 ) was calibrated by using a LaB 6 standard (NIST SRM 660b). The methodology was similar to previous in situ PXRD studies. [34] The precursor sample (dried Cr 2 O 3 @TBT) was loaded into a quartz glass capillary (1.0 mm OD and 0.1 mm wall thickness); silica glass wool plugs were placed before and after the sample section to prevent sample movement during oscillation and heating. Both ends of the capillary were open to air. The loaded capillary was placed at the X-ray beam center and heated (10 8C min À1 ) to the target temperature (700 8C) by a hot air blower under the capillary. The temperature was sensed by a K-type thermocouple about 2 mm beneath the capillary and was calibrated by using KNO 3 and quartz temperature standards. In situ PXRD patterns were collected during the calcination process by using a position-sensitive MYTHEN detector over the 2q range 1.5-81.58 with a time resolution of 2 min. Raman spectra were collected by using a Renishaw inVia Raman Microscope with 514 nm excitation source. X-ray photoelectron spectroscopy (XPS) was performed with a Scienta 300 XPS machine incorporated with a rotating Al Ka X-ray source operating at 13 kV, 333 mA (4.33 kW). Electron analysis was performed by using a 300 cm radius hemispherical analyzer and lens system. The electron counting system consists of a multichannel plate, phosphorescent screen and CCD camera. All multichannel detection counting is done by using proprietary Scienta software. The elements present were determined by a wide energy range survey scan (200 mV step, 20 ms dwell time, 150 eV pass energy and summed over three scans). The high-resolution scans were performed at a similar pass energy (150 eV) as in the survey scan, but a step size of 20 mV, a dwell time of 533 ms and summed over three scans. The instrument operated at a base pressure of 1 10 À9 mbar; the energy scale is calibrated by using the Au 4f, Ag 3d and Cu 2p emission lines, where the half width of the Au 4f 7 emission line is approximately 1.0 eV. All of the sample was mixed with a small amount of Ag powder to act as a binding energy reference. All data analysis and peak fittings were performed by using the CasaXPS software. Diffuse reflectance studies were carried out by using a UV/Vis spectrometer (PerkinElmer Lambda 900) equipped with an integrating sphere (150 mm). Sample weight loss was analyzed by using a TGA Q500 Thermogravimetric Analyzer from TA Instruments in the air (ramping temperature 10 8C min À1 up to 900 8C). Photoluminescence (PL) was performed with a Fluorescence Spectrometer (PerkinElmer LS 55) with a 300 nm excitation wavelength and a cut-off filter at 350 nm.

Photocatalytic testing
The prepared samples were tested for CO 2 photoreduction, as described previously. [35] Briefly, Cr 2 O 3 @TiO 2 powder (0.1 g) sample was first placed in a stainless-steel photoreactor with a quartz window and then the photoreactor was purged overnight by a stream of The apparent quantum efficiency (AQE) was measured under similar photocatalytic reaction conditions using the same UV lamp (75 mW cm À1 , 365 nm). The incident flux was determined by Laboratory Spectroradiometer (Apogee Instruments). The AQE values of CH 4 evolution for CO 2 photoreduction reaction were calculated according to the following equation: FCH 4 % ð Þ ¼ 8 Â number of evolved CH 4 molecules ð Þ number of incident photons Â 100 %