Gold nanoparticles supported on TiO₂–Ni as catalysts for hydrogen purification via water–gas shift reaction

Abstract

Gold nanoparticles deposited on TiO₂–Ni prepared by sol–gel process catalyses the CO oxidation to hydrogen purification. Gold-catalysts were characterized by UV-Vis and Raman spectroscopies, X-ray diffraction, H₂-TPR, N₂ physisorption, HRTEM and STEM-HAADF microscopies. These catalysts were applied in the water–gas shift reaction at temperatures from 30 to 300 °C and this reaction was studied by DRIFTS to understand the catalytic surface phenomena. The best CO conversion was showed by doped Au/TiO₂–Ni(1) with regard to Au/TiO₂ sol–gel and Au/TiO₂–P25 catalysts. DRIFTS confirm the strong and favorable effect of doping nickel ions into the reducible TiO₂ framework. Nickel contents from 1 to 10% enhance the WGS reaction in contrast to the undoped catalyst. Ni-doped TiO₂ support was practically inert for WGS reaction. These gold catalysts present significant activity in the water–gas shift reaction that allows purification of hydrogen from industrial sources at low-temperature.

---

1 Introduction

Hydrogen is a potential fuel solution for satisfying many of our energy requirements because it is storable, clean, and environmentally friendly. At present, nearly 95% of the hydrogen supply is produced from the reforming of crude oil, coal, natural gas, and biomass. The reformed fuel contains 1–10% of CO, which degrades the performance of the Pt electrodes utilized in the fuel cell systems. This is the reason why the water–gas shift reaction (WGSR) plays a key role for getting clean hydrogen for both fuel cells and other industrial applications, (CO + H₂O → CO₂ + H₂).¹

On the other hand, the WGSR is an important step to remove carbon monoxide from various chemical processes such as hydrogen from steam reforming of hydrocarbons and ammonia production.² This reaction involves the reduction of H₂O molecule and the oxidation of CO to produce CO₂ and H₂, breaking an energetic barrier of 41.1 kJ mol⁻¹ (exothermic reaction).³ This method is one of the most used when high-purity hydrogen is required for the operation of fuel cells that use sensitive membranes. Recently, the interest of WGSR has been renewed due to the possibility to obtain pure hydrogen for the more efficient use of fuel-cells or in other reactions related to environmental protection and clean technologies.^4,5

Currently the use of Au-based catalysts has been of interest in WGSR thanks to the good catalytic properties of nanometric gold at low reaction temperatures in contrast to others metal catalysts and the inert properties of bulk gold. It has been demonstrated that gold nanoparticles (AuNPs) can have remarkable catalytic properties in this reaction.⁶ AuNPs supported on metallic oxides are active for certain reactions including low-temperature catalytic combustion, partial oxidation of hydrocarbons, and reduction of nitrogen oxides.⁷ It is generally agreed that the catalytic activity of gold catalysts depends on the size, shape and dispersion of the gold particles, but the nature of the support material, the preparation method, and the activation procedure have a crucial role.^7–11 The role of the TiO₂ as substrate is very important, because, most of the works report that AuNPs supported on reducible oxide supports are more active than gold on nonreducible supports as Al₂O₃.¹² Andreeva et al. were the first to use AuNPs supported on reducible support for the WGSR displaying promising results.^13,14 TiO₂ as a support of AuNPs has been vastly used with outstanding results. However to enhance the activity, stability and AuNPs dispersion, doping TiO₂ framework has been the principal strategy.¹⁵ To our knowledge, there are no reports of the use of nickel-loaded TiO₂ as a reducible support for WGSR. Particularly, reducible supports such as TiO₂, CeO₂ and Fe₂O₃ are the most suitable for CO oxidation at low temperatures; however, its role in the CO oxidation is still under discussion.¹⁶

On the other hand, there is evidence that nickel oxide can act as reducible support and has a good performance in many reactions such as catalytic hydrogenation, Wolff–Kishner and Clemmensen reduction, methanation by Fisher Tropsch, among others.^17–19 Also, Sreethawong et al.²⁰ reported synthesis of mesoporous TiO₂-supported NiO photocatalyst by single step sol–gel process for photocatalytic H₂-production from methanol aqueous solution. Jing et al.²¹ investigated the fabrication of Ni-doped mesoporous TiO₂ and its photocatalytic activity for hydrogen evolution in methanol aqueous solution. In 2009, Jang et al.²² reported the enhanced photocatalytic H₂-production efficiency of nickel-intercalated titanate nanotube from methanol aqueous solution.²³ However, to the best of our knowledge, there aren't reports on the water–gas shift reaction H₂ production over Au/TiO₂–Ni.

In the present work, we study for the first time the WGSR at low temperatures using AuNPs supported on the TiO₂ reducible support modified with different contents of nickel. The role of nickel in the TiO₂ host lattice was determined by XRD, N₂ physisorption, TPR, UV-Vis and Raman spectroscopies and HRTEM and STEM microscopies. Diffuse reflectance infrared Fourier transform spectroscopic (DRIFTS) characterization was carried out to directly obtain insights of AuNPs support interactions with CO in the WGSR evaluation.

2 Experimental

2.1 Sol–gel hydrolysis process

TiO₂–Ni sol–gel catalysts were prepared by a controlled hydrolysis sol–gel process using titanium(IV) isopropoxide (Sigma-Aldrich 97%) and nickel nitrate hexahydrated(II) (97%, Sigma-Aldrich) as precursors, ethanol and distilled water as solvents. According to the previously reported sol–gel synthesis an appropriated amount of nickel nitrate was used to obtain 0.1, 0.5, 1.0, 2.5, 5.0, and 10 wt% of Ni content, and then nickel precursor was dissolved in distilled water.^24,25 The TiO₂–Ni catalyst was prepared by adding separately and simultaneously, drop by drop, 75.6 mL of titanium isopropoxide, 22.9 mL ethanol–18 mL water. The nickel solution was added to a mixture of 22.9 mL ethanol–18 mL water solution contained in a 4-neck round bottom flask (1 L) equipped with magnetic stirrer and thermometer. The alkoxide–ethanol–water molar ratio was 1/3/8. Later on, the solution was vigorously stirred at 50 °C until all the reagents were added. Subsequently, the solution was gradually heated to 70 °C. The gelled product was aged for 48 h at 70 °C. The solvents and unreacted part of precursors were removed at 80 °C and dried overnight under vacuum at 100 °C. Finally, the materials were thermally treated at 500 °C for 4 h at a rate of 2 °C min⁻¹. The catalysts were identified as TiO₂–Ni(X) where X indicates Ni content (wt%).

2.2 Synthesis of gold catalysis by deposition–precipitation with urea

The deposition of the gold nanoparticles was performed by deposition–precipitation with urea (DP Urea).^26,27 The gold precursor, HAuCl₄ (4.2 × 10⁻³ M) and the urea (0.42 M) were dissolved in 49.29 mL of distilled water at the initial pH of 2.4. Then, two grams of TiO₂ or TiO₂–Ni supports were added to this solution under constant stirring; thereafter, the suspension temperature was increased to 80 °C and kept constant for 16 h. The urea decomposition led to a gradual rise in pH from 2.4 to 7.²⁶ The amount of gold in the solution corresponded to a calculated gold loading of 2 wt% on the supported catalyst.

After the deposition–precipitation procedure, all the samples were centrifuged, washed with water at 50 °C and then centrifuged four times and dried under vacuum for 2 h at 100 °C. The thermal treatments were performed in a U reactor with a fritted plate of 1.5 cm of diameter; the calcination under a flow of dry air (1 mL min⁻¹ mg_sample⁻¹) was performed at 300 °C, for 2 h. All the catalysts were stored at room temperature under vacuum in a desiccator away from light in order to prevent any alteration.^28,29 Catalysts were identified as Au/TiO₂–Ni(X), where Au content was constant (2 wt%) and X indicates Ni content (wt%).

2.3 Water–gas shift reaction test

The WGSR was carried out in a flow reactor at atmospheric pressure and increasing temperature range from 25 to 300 °C (light off test). 50 mg of dried catalyst was first treated in situ in a flow of 40 mL min⁻¹ of air with a heating rate of 2 °C min⁻¹ up to 300 °C followed by a temperature plateau of 2 h, in order to form the AuNPs. After activation treatment the catalyst was cooled to 25 °C under the same gas.

The reactant gas mixture (5% CO, 10% water vapor in He balance) was introduced with a total flow rate of 40 mL min⁻¹, the heating rate was 2 °C min⁻¹. The products were monitored with an on-line gas chromatograph Agilent Technologies 6890N equipped with a FID detector and a HP Plot Q column. H₂O produced during the WGSR was trapped into U glass before the FID detector.

2.4 Characterization techniques

The thermally treated catalysts were characterized with X-ray diffraction in order to identify crystalline phases using a Bruker Advance 8 Diffractometer with CuKα radiation (1.5404 Å), the crystalline phases were confirmed with Raman spectroscopy, using a Micro-Raman Renishaw spectrometer equipped with an argon laser (514 nm). Diffuse reflectance UV-visible spectra of the catalysts were obtained using a Cary 5000 (UV-VIS-NIR) spectrophotometer, a spectrum of teflon (from Aldrich) was used as reference. The specific surface area was determined by N₂ physisorption in a Quantachrome sorptometer apparatus.

The hydrogen temperature programmed reduction (H₂-TPR) study of the dried catalysts was performed in a RIG-150 unit under a flow of 10% H₂/Ar gas mixture (30 mL min⁻¹) and with a heating rate of 10 °C min⁻¹ from room temperature to 600 °C. H₂O produced during the reduction process was trapped before the TCD detector. After ex situ thermal treatment under the identical conditions as for the WGSR, the catalysts were examined by transmission electron microscopy (TEM) in a Tecnai FEI 300 microscope operated at 300 kV. The samples were suspended in isopropanol and then sonicated for 5 min. Finally, the samples were mounted in a Cu TEM grid. The particle size distribution histograms for the catalysts were established from the measurements of 300 particles. The average particle diameter (d_s) was calculated using the following formula: d_s = Σn_id_i/Σn_i, where n_i is the number of particles of diameter d_i.

CO adsorption was followed by DRIFTS spectroscopy to characterize the metallic surface. The experiments were carried out in a Nicolet 670FT-IR spectrophotometer equipped with a Praying Mantis for DRIFTS and a low/high temperature reaction chamber by Harrick. In each experiment, approximately 25 mg of dried catalyst was packed in the sample holder and pretreated in situ under air flow (30 mL min⁻¹, heating rate 2 °C min⁻¹) up to the chosen temperature followed by a plateau for 1 h. After the thermal treatment, the catalyst was cooled to room temperature under the same gas flow and then purged with N₂ before the introduction of 5% CO in N₂ (30 mL min⁻¹). A spectrum recorded under N₂ flow was used as reference, then several spectra were recorded under the CO flow until the band intensity was stable; afterwards, temperature was increased under CO flow, and the spectra were recorded at increasing temperatures. Finally the catalyst was cooled again until 100 °C and the surface was cleaned with N₂, in order to simulate WGSR water was introduced and CO, spectra were recorded from 100 to 300 °C. The turnover frequencies (TOF), number of molecules of CO converted per surface atom of gold particles and per second, were determined from the reaction rates obtained in the kinetic regime and the dispersion of gold (surface atoms/total atoms in the particle). The dispersion of gold was calculated with the assumption that gold particles were cuboctahedral with a hexagonal face in contact with the TiO₂ surface.

3 Results and discussion

3.1 Water–gas shift reaction test

Fig. 1 shows the CO conversion as a function of the reaction temperature, from 25 °C to 300 °C for catalysts thermally treated at 300 °C with air flow. For all the materials the CO conversion starts up to 100 °C. The TiO₂–Ni sol–gel support does not present any CO conversion in the range of temperatures studied, see Fig. 1b. According to Fig. 1a, the highest CO conversion during WGSR was observed for the Au/TiO₂–Ni(1) catalyst, Au/TiO₂–Ni(0.1) catalyst shows the lowest CO conversion of all the gold–nickel catalysts, Au/TiO₂–Ni(0.5) catalyst show higher CO conversion from 100 °C to 180 °C range, after the 180 °C the CO conversion began to decrease notably until reaching the 46.2 percent. Au/TiO₂–Ni(2.5) and Au/TiO₂–Ni(5) catalysts show almost the same CO conversion. The increase of the amount of nickel incorporation into TiO₂ (Au/TiO₂–Ni(10) catalyst) seems to inhibit the CO conversion. Loadings lower than 1% wt of Ni also seem to inhibit the CO conversion.

Fig. 1 CO conversion during WGSR with in situ thermal treatment with air flow at temperatures from 30 to 300 °C, (a) Au/TiO₂–Ni catalyst and (b) comparison of bare TiO₂ and TiO₂–Ni(1) support regard to Au/TiO₂–Ni(1).

In Fig. 1b the activity of the most active catalyst (Au/TiO₂–Ni(1)) is compared with the one of the gold catalysts supported on bare TiO₂ synthesized by sol–gel and TiO₂ Degussa P25. At 180 °C the nickel doped gold catalyst, Au/TiO₂–Ni(1) clearly shows an increase of CO conversion in contrast to bare catalysts. It is important to note that all the nickel modified TiO₂ AuNPs showed higher CO conversion than the AuNPs supported on bare TiO₂ catalysts (Table 1). According to specific surface area (SSA) data show in Table 1, the method of synthesis provides SSA values up to 72 m² g⁻¹ for the most of the catalysts, only the catalyst Au/TiO₂–Ni(10) which has a SSA of 48.5 m² g⁻¹ presents a negative effect. TiO₂–P25 bare catalyst also show low SSA (51 m² g⁻¹). In Table 1 a direct relationship between SSA and %CO conversion can be noticed, meaning, the more SSA, the greater catalytic activity; therefore, when the nickel content is 10 wt% there is a low SSA and low CO conversion, 60% at 300 °C. The suitable amount of nickel incorporation on the sol that allows the doping effect seems to be less that 2.5 wt%. It is shown that the nickel ions incorporation has an important surface effect with AuNPs for enhancing the CO conversion from 100 °C to 300 °C probably due to the good dispersion of AuNPs.

Table 1 Physical and catalytic properties of the Au/TiO₂–Ni catalyst

Catalyst	Crystallite size (nm)	Specific area (m^2 g^−1)	CO conversion^a (%)	Reaction rate^a (mol CO per mol Au per h) (10^−3)	SPR^b (nm)
a CO conversion and reaction rate were determined at 300 °C.b Wavelength which occurs SPR (Surface Plasmon Resonance) of AuNPs.
Au/P25	19.7	51.3	54.9	1.37	546
Au/TiO2	16.3	72.7	44.6	1.12	546
Au/TiO2–Ni 1	13.9	79.0	75.9	1.90	518
Au/TiO2–Ni 2.5	13.3	78.8	75.3	1.88	522
Au/TiO2–Ni 5	10.9	76.3	75.0	1.88	526
Au/TiO2–Ni 10	13.7	48.5	60.2	1.51	549

---

3.2 Effect of the catalyst activation conditions on the WGSR

In order to study the role of thermal treatment activation of Au/TiO₂–Ni catalysts in WGSR, the more active catalyst, Au/TiO₂–Ni(1), was treated at 200 and 400 °C under air flow, see Fig. 2. The highest catalytic activity was obtained after an activation treatment at 300 °C in air, the same catalyst was treated at 300 °C with hydrogen flow. Fig. 2 shows that the use of hydrogen gas as a reducing agent instead of air, reduces the CO conversion of Au/TiO₂–Ni(1) catalyst. Thus activation at 300 °C in air seems to be the optimal thermal activation conditions for this catalyst in the WGSR. Furthermore, gold loading was varied at 1 and 3 wt% on support TiO₂–Ni(1) and using the optimal thermal treatment conditions for the WGSR, showing minor CO conversion regarding the optimal loading of 2 wt%. The determination of the turnover frequencies of two gold–nickel catalysts was carried out in order to obtain insights of nickel synergic effects on surface phenomena. The TOF for the Au/TiO₂–Ni(1) catalyst was 0.0053 h⁻¹ while for Au/TiO₂–Ni(10) was of 0.0047 h⁻¹. The result suggests that by incorporating 1% Ni the surface area was improved and AuNPs dispersion over TiO₂–N surface resulted in more active sites for WGSR enhancement.

Fig. 2 CO conversion during WGSR of Au/TiO₂–Ni catalyst varying Au content with in situ thermal treatment conditions.

To understand this behavior in the WGSR and correlate with catalytic activities of these materials, their structural characterization was carried out.

XRD characterization allows the identification of the structure and crystalline phases of catalysts. The diffractogram patterns presented in Fig. 3 show that in catalyst with a low nickel content (1 and 2.5 wt%) only anatase crystalline phase is detected (JCPDS 04-002-2678) whereas for higher nickel contents (5 and 10 wt%), nickel titanate crystalline phase is observed (JCPDS 04-012-0745). It is worth to note that other peaks attributable to metallic nickel or nickel oxides are not observed, see Fig. 3.

Fig. 3 X-ray diffraction patterns of TiO₂–Ni catalysts and bare TiO₂.

According to JCPDS cards of anatase and nickel titanate, 04-002-2678 and 04-012-0745, respectively, anatase has a tetragonal structure and the nickel titanate is rhombohedral. This indicates that both phases can coexist.³⁰ The observed reflections of NiTiO₃ show that at amounts bigger than 2.5% wt, nickel ions started to segregate from the TiO₂ framework and crystallized as a NiTiO₃ phase. This could be the reason why Au/TiO₂–Ni(10) catalyst had low SSA. It has been reported that titanate oxides present low SSA.³¹ The crystallite size of TiO₂–anatase was estimated by Scherrer equation,³² the results are reported in Table 1. The more crystalline sample was the bare TiO₂ P25, because of that, this sample presents a bigger crystallite size (19.7 nm).

In Fig. 4a, the results of Raman spectroscopy of Au/TiO₂–Ni catalysts are shown. One can observe vibrational modes of anatase (E_g – 144 cm⁻¹, E_g – 197 cm⁻¹, B_1g – 400 cm⁻¹, A_1g + B_1g – 515 cm⁻¹ and E_g – 640 cm⁻¹)^33,34 and nickel titanate phases for Au/TiO₂–Ni catalyst with more than 2.5% of Ni loading,^35,36 see Fig. 4a.

Fig. 4 Vibrational modes of Raman spectra of (a) Au/TiO₂–Ni catalysts and bare TiO₂, (b) Raman shift of E_g vibrational mode of TiO₂–anatase of TiO₂–Ni samples, (c) Raman shift of E_g vibrational mode of TiO₂–anatase of Au/TiO₂–Ni samples.

According to spectra shown in Fig. 4b we can assume that a substitutional doping occurred, because it is observed a red-shift between Au/TiO₂ and Au/TiO₂–Ni(2.5) samples of 1.6 cm⁻¹ in the main vibrational mode of anatase E_g (144 cm⁻¹) that is associated to O–Ti–O bending type vibrations, Fig. 4b and c. This shift is significant, considering the spectral resolution of the equipment of 0.1 cm⁻¹. This shift suggests a shortening and/or rigidity of the bonding Ti–O of this vibrational mode due to increase of vacancies produced by surface oxygen, attributable to nickel doping agent incorporation, which reduces the O/Ti rate.^37–39 Furthermore only the AuNPs incorporation in TiO₂ sol–gel produces an overall red-shift of 1.8 cm⁻¹ in the same vibrational mode E_g anatase indicating that AuNPs are giving electronic density to TiO₂ support, see Fig. 4c.

According to UV-Vis spectra showed in Fig. 5, the surface plasmon resonance (SPR) of AuNPs can be seen on the bare materials Au/P25 and Au/TiO₂ around 545 nm, which is usually located at 520 nm, but depending on the support properties this band could be blue-shifted or red-shifted.⁴⁰ Analyzing the shift of the Au SPR band of TiO₂–Ni catalysts, according to the plasmonic bands of the bare catalysts, the shift may be associated with their catalytic activity. The catalyst with larger blue-shift with respect to the bare Au/TiO₂ from 518 to 526 nm has higher CO conversion: in agreement with the literature, a blue-shift means that the support is donating electronic density toward AuNPs.⁴¹ It means that most of AuNPs are in reduced state, Au⁰. Therefore, the materials that presents plasmonic band from 546 to 549 nm have lower CO conversion. This red-shift indicates that electronic density is transferred from the AuNP to the support.⁴¹ This means that AuNPs has been partially oxidized. Thereby catalytic activity of the catalysts is correlating with the SPR position in the UV-Vis spectra.⁴²

Fig. 5 UV-vis diffuse reflectance spectra of the Au/TiO₂–Ni catalysts.

In the case of materials with nickel content of 5 and 10 wt%, that contain nickel titanate crystalline phase an additional band appears at 450 nm, Fig. 5. This band is due to the crystal field splitting of the 3d⁸ band associated with Ni²⁺ ions which splits up into two sub-bands as charge transfer bands of Ni²⁺ → Ti⁴⁺.³¹ This result supports the assumption of the segregation on Ni²⁺ ion to the surface and the formation of NiTiO₃ ilmenite crystalline phase for Ni content starting approximately from 5 wt%.

3.3 TPR characterization

TPR profiles are shown in Fig. 6. This programmed reduction technique was used to elucidate the oxidation state of AuNPs and nickel support in the dried catalysts obtained after DPU. It has been previously shown that in Au/TiO₂ catalysts prepared by DP urea, gold is oxidized as Au³⁺ species derived of HAuCl₄ precursor.⁴³ It was observed that all materials present a reduction peak that starts at 89 °C and ends at 185 °C. This reduction peak is related with the reduction of Au³⁺ species to Au⁰.⁴⁴ Depending on the support, variations of the temperature of reduction of the peak are observed, considering 128 °C as reference.⁴⁴ The Au/TiO₂–Ni(1) catalyst shows the lowest temperature reduction peak with a maximum at 115 °C. This indicates that this catalyst needs a minor energy to the Au³⁺ species reduction. Thus stabilized Au/TiO₂–Ni acts as the better catalyst in WGSR, Fig. 1.

Fig. 6 TPR analysis of Au/TiO₂–Ni catalysts.

Reduction peaks at temperatures around 450 °C may be related to the reduction of Ni²⁺ to Ni⁰ and those at temperatures higher than 500 °C indicate the reduction of the support Ti⁴⁺ to Ti³⁺.^45,46

3.4 HAADF and HRTEM characterization

A representative STEM-HAADF image of Au/TiO₂–Ni(1) catalysts activated ex situ in air at 300 °C shows a homogeneous dispersion of AuNPs on the TiO₂–Ni support (Fig. 7a). This catalyst has shown the best catalytic activity, according to Fig. 1a.

Fig. 7 Au/TiO₂–Ni(1) catalyst, (a) STEM-HAADF image and (b) histogram of 300 AuNPs.

A homogeneous dispersion of AuNPs on the TiO₂–Ni support was observed. The particle size distribution histogram for this catalyst (Fig. 7b) showed that the average diameter of AuNPs was centered at 3.2 nm, which agrees with previous published works where AuNPs were prepared by DP method using urea as precipitant agent.^26,27,47,48

Fig. 8a shows HRTEM images of Au/TiO₂–Ni(1) catalyst activated ex situ in air at 300 °C. In these images it can be observed AuNPs (black spots) and crystallographic planes for the two crystalline phases present in the catalyst. The lattice distances were measured using the software Digital Micrograph. In Fig. 8a the lattice distance of 0.34 nm corresponds to the (101) plane of anatase TiO₂, whereas the lattice distance of 0.28 nm can be attributed to the nickel titanate crystalline phase corresponding to the (104) plane. Thus, at low nickel contents, nickel ions start to segregate and the growth of nickel titanate on surface of TiO₂ starts. For the case of the catalyst Au/TiO₂–Ni(10), a higher ratio of crystals showing interplanar distance corresponding to the (104) plane of nickel titanate was observed regarding to the (101) plane of anatase TiO₂ (Fig. 8b). This result is in agreement with crystalline phases identified in the characterization by XRD (Fig. 3).

Fig. 8 HRTEM images for catalysts (a) Au/TiO₂–Ni(1) and (b) Au/TiO₂–Ni(10).

3.5 Surface composition of AuNPs/Ni–TiO₂ catalyst determined by DRIFTS

DRIFTS spectroscopy characterization allows to analyze the metallic activity of the catalyst and may elucidate the interaction of gold species with CO adsorption under the WGSR conditions. The study was focused on the Au/TiO₂–Ni(1) catalyst because it was the most active one and bare Au/TiO₂ in order to elucidate the effect of nickel on the WGSR.

Fig. 9a shows the DRIFTS spectra of Au/TiO₂–Ni(1), that show the characteristic band attributed to the interaction between CO and the catalyst surface. The band at 2103 cm⁻¹ corresponds to the vibration mode of Au⁰–CO which gradually decreases in intensity and shifts toward 2118 cm⁻¹. Some authors have suggested that this shift is related to the π backdonation from Au to CO and the decrement in intensity has been attributed to the Au^δ+–CO species formation.^49,50 However as in Fig. 9 this band do not change in intensity as a function of the temperature, it is most surely due to molecular CO adsorbed on the TiO₂ surface.^51,52

Fig. 9 DRIFTS spectra using Au/TiO₂–Ni(1) catalyst (a) CO adsorption at ambient temperature, (b) CO adsorption with temperature increases from 25 to 300 °C and (c) simulation of WGR reaction with CO and H₂O flows with increase of temperature from 100 to 300 °C.

Moreover it is observed an important increase in intensity of a band that appears at 2074 cm⁻¹ (between 0 and 20 minutes) that is due to the constant reduction of Au⁰ to Au^δ− and to the fact that the intensity of absorption of CO on Au^δ− is higher than the one on Au⁰.^53,54

Another band is detectable at 2172 cm⁻¹. This band has been attributed to the electrostatic interaction Au³⁺–CO which has been reported to be very stable due to the synergism between δ and π compounds of the bonding Au–CO.⁴⁹ However this band is also present during CO adsorption onto the bare sol–gel support TiO₂–Ni(1), see Fig. 10, thus this band can be attributed to CO adsorption onto Ti⁴⁺ of TiO₂, particularly of the anatase structure, meaning a linear adsorption of CO onto Ti^δ+ cations where probably δ = 4.^55–58 On the other hand, a band at 2120 cm⁻¹ (present in catalysts with and without gold) can be attributable to gaseous CO or to framework vibration of the catalyst support when temperature increases.^51,52

Fig. 10 DRIFTS spectra using support TiO₂–Ni(1) catalyst during CO adsorption with temperature increase from 25 to 300 °C.

With the purpose to analyze CO adsorption–desorption onto active sites of the catalyst, an experiment under continuous flux of CO as a function of temperature from 25 to 300 °C was carried out. Fig. 9b shows DRIFTS spectra where the following issues can be noticed. (i) An increase of the intensity of the 2077 cm⁻¹ band (Au^δ−–CO) followed by a gradual shift toward 2054 cm⁻¹. At 300 °C, this band decreases and the largest shift of 23 cm⁻¹ occurs. This red shift can be attributed to the decrease of the CO dipole–dipole interaction because of the decreasing CO coverage of the gold particles as temperature increases.⁵⁹ (ii) At 1905 cm⁻¹ a new band appears which gradually increases with regard to temperature. According to the literature data this band is attributable to carbonyl bridges formed on cationic sites, therefore this band can be related to the bridged structure Ni²⁺–(CO)₂.^48,60 This band is not visible in Au/TiO₂ previously reported by one of us.⁴⁴ An increase of the intensity of this band as a function of the temperature of the catalyst is observed in Fig. 9b. This can be due to surface saturation of Au sites with CO and the requirement of CO to keep on surface searching other kind of bonding. Decrease of this band at 250 °C is due to the desorption process by the increase of temperature. According to these issues, we can confirm the participation of Ni ions in the adsorption of CO. A higher content of nickel participating in CO adsorption enhance the efficiency of WGSR for the hydrogen purification. In Fig. 10 it is observed that at temperatures higher than 250 °C a band begins to appear at 2076 cm⁻¹. This band has been assigned to the linear adsorption of CO onto metallic nickel.^61,62 This can suggest that nickel crystals growth in the TiO₂ surface at high reaction temperatures.

Finally, an experiment to simulate WGSR conditions into the DRIFTS cell was done. In that case the experiment started at 100 °C which is the temperature at which CO conversion starts, see Fig. 1. Before the introduction of reactant WGSR mixture the catalyst surface was cleaned with N₂. In Fig. 9c the spectra at 100, 125 and 150 °C are shown. A similar behavior in the bands increased intensity of 2074 and 1905 cm⁻¹ (Au^δ−–CO and Ni–(CO)₂), respectively, is observed. However, at 300 °C these bands decrease due to the CO desorption. Moreover the 2074 cm⁻¹ band has the largest red-shift at this temperature.

The reaction is visible in DRIFTS from 200 °C. This is notably due to increase of the band assigned to the generation of CO₂ at 2400 cm⁻¹ associated to H₂ production.

At these temperatures, the maximum CO conversion is due to a remarkable redox processes in WGSR as seen in Fig. 1 and 2. At 300 °C, CO₂ desorption in both nickel Ni–(CO)₂ and Au^δ−–CO is observed. According to Fig. 1 and DRIFTS experiments simulating WGSR, the incorporation of nickel ion in the framework of TiO₂ has an important consequence in the adsorption of CO species. The WGSR was enhanced at Ni content lower than 2.5% favoring adsorption of CO and desorption of CO₂ as a product. The structural characterization shows that segregation of nickel as ilmenite NiTiO₃ structure starts from 2.5% of Ni and becomes a mixed oxide TiO₂–NiTiO₃ at 10% wt Ni. Considering the ionic radii of nickel, titanium, and oxygen of 0.69, 0.96, and 0.22 Å, respectively, some Ni²⁺ ions could replace some Ti⁴⁺ ions or some superficial O²⁻ ions during the gelling step, the TiO₂ framework will be stressed, which may constrain the growth of TiO₂–anatase crystallites, thereby favoring the formation of NiTiO₃ ilmenite structure.⁶³ According to the cited ionic radii, it is most probable that both Ni²⁺ and O²⁻ ions stressed the framework and constrain the growth of TiO₂–anatase crystallites, as it was observed in XRD characterization. Then at 5% of nickel, probably nickel ions replace some Ti⁴⁺ and O²⁻ ions start to form nickel titanate instead of TiO₂. Therefore, the coexistence of anatase tetragonal structure and a nickel titanate rhombohedral one inhibits the CO conversion in WGSR.

4 Conclusions

Supported Au/TiO₂–Ni was tested for the first time as catalyst in the water–gas shift reaction, obtaining more active catalyst than the corresponding Au/TiO₂ (sol–gel and P25) ones. It is concluded that not only AuNPs plays an essential role during this reaction, but also there are synergetic effects between AuNPs and nickel species present on the surface of TiO₂ which improves the CO conversion in WGSR. TiO₂–Ni support is completely inactive for WGSR in the temperature range studied. The Ni loading to achieve the maximum enhancement in CO conversion was 1% wt. TPR and DRIFTS characterization of AuNPs/TiO₂–Ni catalysts give insights into the strong interaction of Au⁰ with TiO₂–Ni support that makes sensitive the sol–gel systems to the WGSR without changing AuNPs properties. TOF calculations confirm that some active sites slightly more active were present when the Ni loading was 1% wt.

AuNPs supported on Ni-doped TiO₂ represents an effective and stable catalyst for hydrogen purification at low temperature by WGSR.

Acknowledgements

M. H. R. thanks CONACYT for the Ph. D. fellowship. The use of the infrastructure of CCADET and LINAN are also gratefully acknowledged. We thank V. Maturano-Rojas, A. Sandoval-García, B. A. Rivera-Escoto for their technical assistance. We thank SEP CONACYT-CB-2011/169597 and CB-2009/130407 projects and DGAPA-PAPIIT (Project 103513) for financial support. We also thank Dr H. C. Rosu for comments on the text.

Notes and references

1. J. A. Rodríguez, J. Evans, J. Graciani, J.-B. Park, P. Liu, J. Hrbek and J. F. Sanz, J. Phys. Chem. C, 2009, 113, 7364.
2. V. Idakiev, T. Tabakova, K. Tenchev, Z. Y. Yuan, T. Z. Ren, A. Vantomme and B. L. Su, J. Mater. Sci., 2009, 44, 6637.
3. Q. Fu, H. Saltsburg and M. Flytzani-Stephanopoulos, Science, 2003, 301, 935.
4. F. Boccuzzi, A. Chiorino, M. Manzoli, D. Andreeva, T. Tabakova, L. Ilieva and V. Iadakiev, Catal. Today, 2002, 74, 169.
5. H. Sakurai, A. Ueda, T. Kobayashi and M. Haruta, Chem. Commun., 1997, 271.
6. M. C. Daniel and D. Astruc, Chem. Rev., 2004, 104, 293.
7. M. Haruta, Catal. Today, 1997, 36, 153.
8. G. C. Bond and D. T. Thomson, Catal. Rev. Sci. Eng., 1999, 41, 319.
9. J.-D. Grunwaldt, C. Kiener, C. Wögerbauer and A. J. Baiker, J. Catal., 1999, 181, 223.
10. J.-D. Grunwaldt and A. Baiker, J. Phys. Chem. B, 1999, 103, 1002.
11. M. A. Sanchez-Castillo, C. Couto, W. B. Kim and J. A. Dumesic, Angew. Chem., Int. Ed., 2004, 43, 1140.
12. Y. Borensztein, L. Delannoy, A. Djedidi, R. G. Barrera and C. Louis, J. Phys. Chem., 2010, 114, 9008.
13. D. Andreeva, V. Idakiev, T. Tabakova and A. Andreev, J. Catal., 1996, 158, 354.
14. J. Hua, K. Wei, Q. Zheng and X. Lin, Appl. Catal., A, 2004, 259, 121.
15. D. C. Grenoble and M. M. Estadt, J. Catal., 1981, 67, 90.
16. D. Andreeva, Gold Bull., 2002, 35, 82.
17. C. A. Brown and V. K. Ahuja, J. Org. Chem., 1973, 38, 2226.
18. C. H. Herr, F. C. Whitmore and R. W. Schiessler, J. Am. Chem. Soc., 1945, 67, 2061.
19. D. Papa, E. Schwenk and B. Whitman, J. Org. Chem., 1942, 7, 587.
20. T. Sreethawong, Y. Suzuki, S. Yoshikawa and C. S. K. Rajapakse, Photochem. Photobiol. Sci., 2005, 4, 857.
21. D. Jing, Y. J. Zhang and L. J. Guo, Chem. Phys. Lett., 2005, 415, 74.
22. J. S. Jang, S. H. Choi, D. H. Kim, J. W. Jang, K. S. Lee and J. S. Lee, J. Phys. Chem. C, 2009, 113, 8990.
23. J. Yu, Y. Hai and B. Cheng, J. Phys. Chem. C, 2011, 115, 4953.
24. V. Rodríguez-González, A. Moreno-Rodríguez, M. May, F. Tzompantzi and R. Gómez, J. Photochem. Photobiol., A, 2008, 193, 266.
25. V. Rodríguez-González, F. Paraguay-Delgado, X. García-Montelongo, L. M. Torres-Martínez and R. Gómez, J. Ceram. Process. Res., 2008, 9, 606.
26. R. Zanella, S. Giorgio, C. R. Henry and C. Louis, J. Phys. Chem. B, 2002, 106, 7634.
27. R. Zanella, L. Delannoy and C. Louis, Appl. Catal., A, 2005, 291, 62.
28. R. Zanella and C. Louis, Catal. Today, 2005, 107–108, 768.
29. Y. Wu, K.-Q. Sun, J. Yu and B.-Q. Xu, Phys. Chem. Chem. Phys., 2008, 10, 6399.
30. K. Kakegawa, J. Mohri, S. Shirasaki and K. Takahashi, J. Am. Ceram. Soc., 1982, 65, 515.
31. Y.-J. Lin, Y.-H. Chang, W.-D. Yang and B.-S. Tsai, J. Non-Cryst. Solids, 2006, 352, 789.
32. P. Scherrer, Göttinger Nachrichten Gesell., 1918, 2, 98.
33. T. Ohsaka, F. Izumi and Y. Fujiki, J. Raman Spectrosc., 1978, 7, 321.
34. G. Busca, G. Ramis, J. M. Gallardo-Amores and V. Sanchez Escribano, J. Chem. Soc., Faraday Trans., 1994, 90, 3181.
35. M. I. Baraton, G. Busca, M. C. Prieto, G. Ricchiardi and V. Sanchez Escribano, J. Solid State Chem., 1994, 112, 9.
36. K. P. Lopes, L. S. Cavalcante, A. Z. Simoes, J. A. Varela, E. Longo and E. R. Leite, J. Alloys Compd., 2009, 468, 327.
37. B. Xin, Z. Ren, P. Wang, J. Liu, L. Jing and H. Fu, Appl. Surf. Sci., 2007, 253, 4390.
38. J. C. Parker and R. W. Siegel, J. Mater. Res., 1990, 5, 1246.
39. J. Zhu, J. Ren, Y. Huo, Z. Bian and H. Li, J. Phys. Chem. C, 2007, 111, 18965.
40. S. Link and M. A. El-Sayed, J. Phys. Chem. B, 1999, 103, 4212.
41. C. Wang, C. Liu, X. Zheng, J. Chen and T. Shen, Colloids Surf., A, 1998, 131, 271.
42. M. A. Centeno, C. Portales, I. Carrizosa and J. A. Odriozola, Catal. Lett., 2005, 102, 289.
43. A. Sandoval, A. Aguilar, C. Louis, A. Traverse and R. Zanella, J. Catal., 2011, 281, 40.
44. A. Sandoval, A. Gómez-Cortés, R. Zanella, G. Díaz and J. M. Saniger, J. Mol. Catal. A: Chem., 2007, 278, 200.
45. S. D. Robertson, B. D. McNicol, J. H. de Baas, S. C. Kloet and J. W. Jenkins, J. Catal., 1975, 37, 424.
46. V. Idakiev, L. Ilieva, D. Andreeva, J. L. Blin, L. Gigot and B. L. Su, Appl. Catal., A, 2003, 243, 25.
47. L. Delannoy, R. L. Chantry, S. Casale, Z. Y. Li, Y. Borensztein and C. Louis, Phys. Chem. Chem. Phys., 2013, 15, 3473.
48. A. Sandoval, C. Louis and R. Zanella, Appl. Catal., B, 2013, 140–141, 363.
49. M. Mihaylov, H. Knözinger, K. Hadjiivanov and B. C. Gates, Chem. Ing. Tech., 2006, 79, 795.
50. M. Daté, H. Imai, S. Tsubota and M. Haruta, Catal. Today, 2007, 122, 222.
51. C. Vignatti, M. S. Avila, C. R. Apesteguía and T. F Garetto, Int. J. Hydrogen Energy, 2010, 35, 7302.
52. D. Liu, G.-H. Que, Z.-X. Wang and Z.-F. Yan, Catal. Today, 2001, 68, 155.
53. K. Chakarova, M. Mihaylov, S. Ivanova, M. A. Centeno and K. Hadjiivanov, J. Phys. Chem. C, 2011, 115, 21273.
54. F. Boccuzzi, A. Chiorino, M. Manzoli, D. Andreeva and T. Tabakova, J. Catal., 1999, 188, 176.
55. D. J. C. Yates, J. Phys. Chem., 1961, 65, 746.
56. S. Derrouiche, P. Gravejat and D. Bianchi, J. Am. Chem. Soc., 2004, 126, 13010.
57. C. Morterra, J. Chem. Soc., Faraday Trans. 1, 1988, 84, 1617.
58. C. Morterra, G. Ghiotti and E. Garrone, J. Chem. Soc., Faraday Trans. 1, 1980, 76, 2102.
59. E. Roze, P. Gravejat, E. Quinet, J. L. Rousset and D. Bianchi, J. Phys. Chem. C, 2009, 113, 1037.
60. K. I. Hadjiivanov and G. N. Vayssilov, Adv. Catal., 2002, 47, 307.
61. A. Bandara, S. Dobashi, J. Kubota, K. Onda, A. Wada, K. Domen, C. Hirose and S. S. Kano, Surf. Sci., 1997, 387, 312.
62. J. Miragliotta, R. S. Polizotti, P. Rabinowitz, S. D. Cameron and R. B. Hall, Appl. Phys. A: Solids Surf., 1990, 51, 221.
63. http://www.nist.gov/pml/data/comp.cfm.

---