Enhanced Catalytic Activity of Supported Gold Catalysts for Oxidation of Noxious Environmental Pollutant CO

Noble metal nanomaterials have attracted mounting research attention for applications in diverse fields of catalysis, biology, and nanotechnology. In the present study, we have undertaken a detailed investigation on synthesis, characterization, and catalytic activity studies for CO oxidation by nanogold catalysts supported over CeO 2 and CeO 2 -ZrO 2 (1: 1 mole ratio). The support systems were prepared by modified, simple precipitation technique and the Au supported samples were synthesized using deposition-precipitation with urea method. The physicochemical characterization was performed by XRD, ICP-AES, BET surface area, FT-IR, UV-Vis DRS, Raman Spectroscopy, TEM, and XPS techniques. Au/CeO 2 catalyst showed more than 80% CO conversions at 30 ∘ C, whereas Au/CeO 2 -ZrO 2 exhibited ∼ 100% CO conversion at that temperature. The catalytic performance of Au catalysts is highly dependent on the nature of the support.


Introduction
The curiosity in using gold as a catalyst component has increased during the last few years after the revolutionary work on Au catalysts by Haruta, reporting surprisingly high CO oxidation activity of gold nanoparticles [1][2][3][4][5][6][7][8][9][10]. When it is suitably employed, supported gold is capable of catalyzing a wide variety of reactions. Nevertheless, much skill is needed to achieve this. In recent times, the catalytic applications of gold have been explored extensively in several processes such as chemical processing, hydrotreatment, and purification of hydrogen for fuel cell applications and in reactions of environmental importance such as catalytic treatment of vehicle exhausts and removal of volatile organic compounds [3][4][5][6]. There is no denial to the fact that strict automotive regulations all over the world have necessitated the development of new catalysts to reduce automotive exhaust gases such as CO, NO , and hydrocarbons. Finding an efficient catalyst working at mild, environmentally benign conditions therefore remains a challenge.
Ceria (CeO 2 ) has been regarded as one of the most important components in many catalytic systems due to its remarkable redox properties and striking oxygen storage capacity [3,7]. It has been of wide interest for decades because of its important applications including three way catalysts for automotive emission control [7][8][9]. Under various redox conditions, the oxidation state of the cation may vary between +3 and +4. Its distinct defect chemistry and the ability to exchange lattice oxygen with the gas phase result in an oxide with unique catalytic properties [7][8][9][10]. Literature reveals that ceria and its mixed oxides are very good catalysts for CO oxidation reaction at high temperatures (above 350 ∘ C) [7][8][9][10][11][12][13]. In this view, there has been a tremendous growth in research for CO oxidation using Au catalysts supported over ceria that can be used for low temperature applications like abatement of environmental pollutants (e.g., CO, NO , and VOCs) and synthesis of chemicals via selective oxidation/hydrogenation [10][11][12][13]. Many active gold catalysts are composed of gold nanoparticles (Au-NPs) and a powder support that usually have simple metal-support interfaces. However, one problem generally encountered with these catalysts is the relatively weak metal-support interaction, which may lead to the sintering of gold nanoparticles at elevated temperatures. Therefore, several strategies have been implemented to construct gold 2 Indian Journal of Materials Science catalysts with diversified local structures and better catalytic performance [14][15][16].
It is noteworthy that CeO 2 itself has not been recognized as an excellent catalyst for low temperature CO oxidation [3]. However, keeping in consideration the ability of both ceria and Au in catalyzing CO oxidation reaction alone, there is always a good possibility of designing promising catalyst for the said reaction when ceria and Au are combined in effective ways. In view of this huge possibility, there is still need to explore systems having Au over ceria or ceria-based materials. Reports could be found for such systems in literature showing very good activity for the said reaction [4,17,18]. However, the methods of preparation of most the catalysts are based mainly on sol-gel technique. Here, we have tried to apply simpler synthesis method using coprecipitation technique without using any specific arrangement. Against this background, we have carried out a systematic investigation on synthesis of gold nanocatalysts supported over ceria and one of its most prominent mixed oxides, ceria-zirconia, which is also better known for superior thermal stability and easy reducibility [7]. The support systems were prepared by modified, simple precipitation technique and the Au supported samples were synthesised using deposition-precipitation with urea method. The physicochemical characterization was performed by XRD, ICP-AES, BET surface area, FT-IR spectroscopy, UV-Vis DRS, Raman Spectroscopy, TEM, and XPS techniques. Finally, the catalytic performance of the synthesised catalysts was evaluated for CO oxidation reaction.

Experimental
2.1. Materials. All the chemicals used in this study were of analytical grade and we used them without further purification. In all experiments, double distilled water was used.

Synthesis of CeO 2 (C) and CeO 2 -ZrO 2 (CZ).
Both ceria and ceria-zirconia (1 : 1 mole ratio based on oxides) solid solutions were prepared by adopting a coprecipitation method using ammonium cerium (IV) nitrate (Himedia) and zirconium oxychloride (Himedia), respectively. For preparation of ceria, required amount of the precursor was dissolved in double distilled water, whereas for the case of the mixed oxide required amounts of precursors were dissolved separately in double distilled water under mild stirring conditions and mixed together. Upon complete mixing, excess ammonium hydroxide was added dropwise until the precipitation was complete (pH = ∼8.5). The resulting slurry was filtered off and thoroughly washed with distilled water until it became free from anion impurities. The accumulated oxide and the mixed oxide pastes were carefully placed and covered in a clean ceramic crucible and allowed to dry overnight in a hood. The oxide and mixed oxide samples were then oven-dried at 120 ∘ C for 12 h. Finally, the samples were calcined at 500 ∘ C for 5 h in air atmosphere to remove water and any residual precursors remaining from the coprecipitation step. After cooling, the solid residues were ground using a ceramic mortar and pestle until fine powders were obtained. The rate of heating and cooling was always maintained at 10 ∘ C/min.

Synthesis of Au/CeO 2 (Au/C) and Au/CeO 2 -ZrO 2 (Au/CZ).
Gold was deposited over CeO 2 and CeO 2 -ZrO 2 to prepare 1 wt% of Au to the total amount of supports by employing a modified deposition-precipitation technique with urea (DPU). Typically, certain weighed amount of the support systems was dispersed in double distilled water under stirring condition and the temperature of the resulting suspension was increased up to 80 ∘ C. Then a solution of HAuCl 4 ⋅3H 2 O (Himedia) was added slowly followed by the addition of urea. The concentration of HAuCl 4 ⋅3H 2 O solution was corresponding to 1 wt% of Au and the concentration ratio of urea to total metal was maintained at 100. The resulting mixture was left for stirring for another 12 hours, covered by opaque aluminum foil to get rid of uncontrolled photoreduction. It was then centrifuged to gather the solid product and washed with distilled water until it became free from anion impurities. The collected solid was oven-dried at 120 ∘ C overnight, and finally it was calcined at 400 ∘ C for 12 hours.

Characterization of Catalysts.
The BET surface areas were determined by N 2 physisorption at liquid N 2 temperature on a Micromeritics Gemini 2360 instrument. The powder X-ray diffraction (XRD) patterns were recorded on a Rigaku Multiflex instrument using nickel-filtered CuK (0.15418 nm) radiation source and a scintillation counter detector. The intensity data were collected over a 2 range of 10-80 ∘ . The Raman spectra were collected on a DILORXY spectrometer equipped with a liquid-nitrogen-cooled CCD detector. Samples were excited with the emission line at 514 nm from an Ar + ion laser. UV-Vis diffuse reflectance spectra were recorded on a UV-visible spectrophotometer, Model U-4100 spectrophotometer (solid). Infrared spectra were taken in a FTIR spectrophotometer, Model Spectrum Two FT-IR Spectrometer (Perkin Elmer), by pelletizing the samples with KBr in the midinfrared region at an accelerating voltage of 200 V. Transmission electron microscopic (TEM) investigations were made on a JEM-2100 (JEOL) instrument equipped with a slow scan CCD camera. XPS studies were carried out on a V-G Microtech unit ESCA 3000 Spectrometer with two anodes, AlK (1486.6 eV) and MgK (1256.6 eV). Prior to analysis, the samples were dried and evacuated at high vacuum and then introduced into the analysis chamber. The recorded XPS spectra were charge corrected with respect to the C 1s peak at 284.6 eV. Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES) measurements were carried out (model: ARCOS M/s. Spectro, Germany) in order to estimate the amount of Au deposited on the support.

Catalytic Activity Measurements.
The catalytic activity of the synthesized catalysts was evaluated for the oxidation of CO at normal atmospheric pressure and temperatures in a fixed bed microreactor at a heating ramp of 5 ∘ C/min. About 80 mg catalyst sample (250-355 m sieve fraction) was placed in the reactor for evaluation. Temperature was measured directly at the catalyst bed, using a thermocouple placed in the hollow part of the reactor. The gases used (supplied by Assam Air Products) are helium (>99.9% purity), 10% CO in helium (CO purity, >99.9%), and 10% O 2 in helium (oxygen Table 1: BET surface area, pore size, crystallite size, and lattice parameter of C, CZ, Au/C, and Au/CZ samples.

Results and Discussions
The N 2 BET surface areas of various samples are presented in Table 1. The C and CZ samples calcined at 500 ∘ C bear rather high BET surface areas of 72 and 99 m 2 g −1 , respectively. Not surprisingly, the mixed oxide bears higher surface area than the pristine oxide. A gradual fall in surface area was noted when gold was deposited over those supports. This could be due to penetration of the Au-NPs into the pores of the support, thereby narrowing its pore diameter and blocking some of the micropores. A systematic XRD study was performed to understand the phase symmetry and structure of the prepared samples. The powder XRD patterns of the C, CZ, Au/C, and Au/CZ samples are presented in Figure 1.
All the patterns consist of prominent CeO 2 crystal phases. The characteristic peaks were observed at 2 = 28.6 ∘ , 32.9 ∘ , 47.5 ∘ , 56.7 ∘ , 59.3 ∘ , 69.5 ∘ , and 76.7 ∘ corresponding to the (111), (200), (220), (311), (222), (400), and (331) diffraction planes, respectively. It reveals the formation of face-centered cubic (fcc) fluorite type phase of CeO 2 . In the synthesized samples, no prominent peak for any other phase was found demonstrating the high purity of the synthesized product which in turn is in accordance with the literature [19,20]. Literature shows that, on doping metallic cation with radius smaller than Ce 4+ , the peak positions are slightly shifted towards the higher diffraction angles (2 ) [21]. In this case, dopant cation Zr 4+ bears ionic radius of 0.084 nm, whereas that of Ce 4+ is 0.097 nm. Therefore, the observed shift in 2 positions is the outcome of successful substitution of Ce 4+ ions with isovalent Zr 4+ . This also confirms the formation of CZ solid solutions [22,23]. This incident is quite reinforced by the fact that CZ has smaller crystallite size than pure CeO 2 (obtained from Scherrer equation) as can be seen from the values given in Table 1 [7,24]. The broad nature of the peaks distinctly indicated the nanocrystalline nature of the samples in general. In the case of the CZ sample, broad diffraction lines pertaining to cubic fluorite type phases with the composition Ce 0.6 Zr 0.4 O 2 (PDF-ICDD38-1439) and Ce 0.5 Zr 0.5 O 2 (PDF-ICDD38-1436) were identified. It was observed that the support systems (i.e., C and CZ) bear relatively smaller crystallite size than the final gold decorated catalysts (i.e., Au/C and Au/CZ). This indicates the smaller surface area of the gold supported samples, which is in agreement with the results obtained from the N 2 BET surface area analysis of the synthesized materials. It is evident from Figure 1 that, in addition to the diffraction peaks corresponding to bare CeO 2 , two more very weak peaks centered at 2 = 39.4 ∘ and 43.1 ∘ were also detected which may arise due to the (111) and (200) facets of Au [25,26]. However due to relatively low gold loading (only 1 wt%) diffraction peaks for gold are not well distinct which is also in accordance with the previous literature reports [25,26]. The actual gold contents in the gold supported samples were found to be 0.923 and 0.721, respectively, for Au/C and Au/CZ (in terms of % as determined by ICP-AES). The Raman spectra of C, CZ, Au/C, and Au/CZ samples are shown in Figure 2. As can be seen from the figure, for all samples, a common, sharp, distinct, and strong Raman band centered at around 460-470 cm −1 could be observed. This peak is due to the only Raman-active mode of the perfect cubic fluorite structure of ceria [7,27,28]. This is again ascribed to the symmetric breathing mode of the oxygen atoms surrounding Ce 4+ cation [27,28]. Moreover, the CZ sample shows one weak band at 598 cm −1 due to nondegenerate longitudinal optical (LO) mode of ceria, which arises due to relaxation of symmetry rules [29]. In particular, this band is correlated with the presence of the oxygen vacancies (O v ) and/or lattice defects (MO 8 -type complex defects). The defective structure is associated with relative amount of Ce 3+ ions, preferentially placed on the surface of the sample that could be enhanced when dopant enters the lattice of cerium oxide [17,30,31]. It is evident from Figure 2(b) that incorporation of Zr 4+ cation into the cubic fluorite lattice of ceria results in broadening of F 2g band and is blue shifted towards higher wave numbers. This phenomenon indicates the formation of CZ mixed oxide and the broadness of the spectrum could be due to the presence of smaller crystallite size of ceria which is also established from XRD and BET surface area investigations [32]. In addition, the weak band at around 315 cm −1 represents the displacement of oxygen atoms from their ideal fluorite lattice positions, which shows that CZ sample contains more number of oxygen vacancies [29,32]. Interestingly, no Raman lines due to ZrO 2 could be observed in line with XRD measurements. In the case of the gold supported samples, F 2g band has been shifted to a slightly lower wavenumber (∼460 cm −1 as shown in the inset of Figures 2(a) and 2(b)). This shifting of F 2g mode implies that the deposition of gold on CeO 2 lattice lowers the symmetry of the Ce-O bond [28,33]. The shift might indicate that the deposition of Au affects the support not only at the surface and it may be the result of the interaction with the chemicals involved, which in this case is urea. Moreover, it could be seen that F 2g peak in unsupported C and CZ samples is relatively very much stronger than that in case of gold containing samples. As the CZ systems contain more oxygen vacancies or defect sites, it is rational to expect better catalytic efficiency of Au/CZ sample [19,34].
The UV-Vis diffuse reflectance spectra for pristine ceria, doped ceria, and gold supported samples are shown in Figure 3. All the samples are found to be active in the UV region. Despite the difficulties to interpret the large bandwidth and specular reflectance usually observed in the DR spectra, this technique has extensive use to obtain information on surface coordination and different oxidation states of metal ions in ceria-based materials by measuring dd and f-d transitions and oxygen-metal ion charge transfer bands [35]. It is known from the previous literature that CeO 2 is fairly active in the near UV-Vis region due to its ntype semiconductivity nature [36]. UV absorption in CeO 2 is illustrated by a charge transfer transition between O 2p and Ce 4f, not by an inner transition between Ce 4f and Ce 3d bands [36]. As presented in Figure 3(a), the UV-Vis DR spectra of CeO 2 exhibit three major absorption bands at about ∼227, ∼312, and ∼331 nm. The absorption bands centered at ∼227 nm and ∼312 nm may be attributed to O 2− → Ce 3+ and O 2− → Ce 4+ charge transfer transitions, respectively, while the band at ∼331 nm may be likely to interband transition [37,38]. Incorporation of Zr 4+ into the CeO 2 lattice enables the material to absorb UV-Vis light relatively towards longer wavelength. It could be seen from Figure 3 that bare CZ sample shows absorption band at slightly higher wavelength with respect to pristine C. It is exposed that the light absorption efficacy of the Au-NPs varies with the nature of the supports, while the supports themselves exhibit little absorption of light with wavelengths longer than 400 nm [39]. Surface plasmon resonance (SPR) is a fascinating characteristic shown by noble metal nanoparticles which is primarily dependent on the size, shape, and hence the oscillation of free electrons of the NPs [40]. As an example, nanorods have been known to show two plasmon resonances, one due to the transverse oscillation of the electrons around 520 nm for gold which is independent of aspect ratio and other due to longitudinal plasmon resonance increases with larger aspect ratios [41,42]. It is obvious from Figure 3 that UV-Vis DR spectra of Au/C and Au/CZ samples show absorption bands at about 565 and 558 nm, respectively, which are the distinctive SPR absorption exhibited by the supported Au-NPs. In this case, existence of surface plasmon resonance is an indication of Au-NPs deposited over the C and CZ materials.
The FT-IR spectra of Au/C and Au/CZ samples are shown in Figure 4. The spectra were measured in the range from 400 cm −1 to 4000 cm −1 . The very small and broad absorption bands at ∼3855 and ∼3370 cm −1 were assigned to the stretching vibration of O-H functional group [18]. The peak at ∼2920 cm −1 may be attributed to the N-H stretching vibration of amine formed from the decomposition products of urea and the important peaks found at ∼2300 and ∼1380 cm −1 are due to the absorption of atmospheric CO 2 on the metallic cations [43,44]. Another small band observed at about ∼2850 cm −1 may be due to C-H bonds of organic compounds [3]. The band at ∼1615 cm −1 corresponds to O-H bending of molecularly physisorbed H 2 O [18]. The IR band appearing at ∼1250 and ∼1550 cm −1 is assumed to be due to the nitrogen oxide species [45]. The absorption band at ∼1450 and ∼1042 cm −1 could be ascribed to the presence of unwanted NO 3 − groups and other residues, respectively [18,46]. The bands sited in the range of 800-840 and ∼520 cm −1 may be due to Ce-O stretching frequency indicating the formation of ceria [47]. In addition to the peaks assigned above for both samples, the peak observed at 550 cm −1 could be recommended to Zr-O bond for the Au/CZ sample [48]. The HRTEM images of bare ceria, doped ceria, and supported Au-NPs are presented in Figure 5. The morphology and structure of the Au-NPs were resoluted from HRTEM micrographs. The formation of Au-NPs was identified and differentiated by the lattice spacing measurements. It could be seen that Au-NPs are well dispersed over the C and CZ supports and have narrow but almost uniform size distributions and the particles are nearly spherical in shape with crystalline nature as revealed from the SAED patterns. The HRTEM images of supported Au-NPs as exposed in and CZ is estimated to be in the range of 4 to 5 nm and the size of C and CZ samples is approximately 5 to 6 nm with fluorite cubic structure. These results are also in line with the crystallite sizes calculated by Scherrer equation from the XRD analysis. Furthermore, the corresponding selected area electron diffraction (SAED) patterns (shown as insets in Figure 5) clearly indicate the well crystallinity of the synthesized samples. According to previous literature, the cubic fluorite structure by and large exhibits only three electron diffraction rings which correspond to the interplanar spacings of ca. 3.1 (111), 2.7 (200), and 1.9 (220)Å [50,51]. This feature is quite prominent in the prepared materials which are once again in compliance with the formation of cubic fluorite pattern of the materials. XPS study was carried out in order to know the valence states of the elements contained in the synthesized catalysts. The XP spectra of Au 4f show a doublet corresponding to Au 4f 7/2 and Au 4f 5/2 states with a separation of ∼3.52 eV [52]. As could be observed from Figure 6, binding energies of Au 4f 7/2 state are ∼83.8 eV and ∼83.4 eV, respectively, in Au/C and Au/CZ catalysts. The observed difference is plausibly due to the size effect. Similarly, Au 4f 5/2 binding energies in Au/C and Au/CZ catalysts, respectively, are observed at ∼87.3 eV and ∼87.02 eV. The corresponding binding energy values of both Au 4f states clearly confirm the presence of metallic gold (Au 0 ) species. However, literature shows that Au species may present as either metallic gold (Au 0 ) or cationic gold (Au + 4f 7/2 at 84.5 eV and Au 3+ 4f 7/2 at 86.6 eV) species [52,53]. Thus, it can be concluded that Au species residing in the asprepared gold supported catalysts are only metallic in nature.
As per the literature, O 1s spectra of ceria-based materials show two binding energies corresponding to different types of oxygen ions [52]. As shown in Figure 7, the peak observed at relatively lower binding energy (∼530.0 eV) region stands for CeO 2 lattice oxygen. On the contrary, the very small and dull shoulder at ∼532 eV implies the presence of adsorbed carbonates and hydroxyl groups [52]. Intensity (a.u.) Binding energy (eV) Ce 3d XP spectra of the supported gold catalysts show two spin-orbit components, namely, 3d 3/2 and 3d 5/2 , which have been represented by the notations u and v, respectively, as shown in Figure 8. The corresponding peaks for Ce 3+ state are symbolized by u / (∼903.8 eV) and v / (∼885.4 eV) and those for the Ce 4+ state are represented by v (∼883.8 eV), u (∼902.5 eV), v // (∼889.4 eV), u // (908.2 eV), v /// (∼899.4 eV), and u /// (∼917.6 eV). Accordingly, it was seen that Ce exists together as Ce 3+ and Ce 4+ ions in both synthesized gold supported catalysts [49].
The Zr 3d XP spectrum of Au/CZ catalyst is shown in Figure 9. Two peaks at ∼182.7 and 184.6 eV are observed in the spectrum. The lower binding energy peak is assigned to Zr 3d 5/2 , whereas that at higher BE side is due to Zr 3d 3/2 . The  corresponding BE values for the two spin-orbit components in Zr 3d spectra indicate the presence of Zr 4+ ions [54].
The CO oxidation results of Au/C and Au/CZ samples as a function of reaction temperature are shown in Figure 10. It is well known that ceria and ceria-zirconia samples are very promising catalysts for CO oxidation at high temperatures (beyond 350 ∘ C). In the present case, our prime emphasis was to explore the catalysts for low temperature activity. Therefore, the reaction temperature range we have chosen is from room temperature to 90 ∘ C. Even at a lower temperature of 30 ∘ C, both samples exhibited excellent activity for CO conversion. As observed from the figure, the Au/CZ sample exhibited better CO conversion than Au/C. When Au/C gave more than 80% CO conversion, Au/CZ showed ∼100% CO conversion at that temperature. There was a slight increase in CO conversion rate when temperature was raised and the samples exhibited full CO conversion at 90 ∘ C. Interestingly, our synthesised catalysts have shown better catalytic activity than some recently reported results [4,17,52]. The observed more conversion by Au/CZ sample could be due to increased oxygen mobility in the defective fluorite structure of CZ sample generated by introduction of more Zr 4+ cations into the ceria core lattice compared to bare ceria. Thus, it could be assumed that the catalytic performance of Au catalysts is highly dependent on the nature of the support.

Conclusions
Nanosized ceria and ceria-zirconia solid solutions have been synthesized by a modified simple coprecipitation method and calcined at 500 ∘ C. The Au supported samples were successfully synthesised using deposition-precipitation with urea method. A comparative study for CO oxidation activity has been undertaken with gold supported over ceria and ceria-zirconia. The results of XRD and cell parameter values revealed that zirconium is incorporated into the ceria lattice and more defective sites are formed. Raman spectroscopic measurements suggested the presence of oxygen vacancies, lattice defects, and displacement of oxygen ions from their ideal lattice positions. The surface plasmon band obtained from UV-Vis DRS analysis established the successful deposition of gold in the synthesized catalysts. The nanometer dimension of the prepared catalysts was confirmed from both XRD and TEM analysis. Better CO oxidation activity was observed for the Au/CZ sample in comparison to that of Au/C sample. In summary, it can be concluded that gold supported nanosized ceria and ceria-zirconia solid solutions by simple deposition-coprecipitation method using urea are an effective and promising approach for making catalysts for low temperature CO oxidation.