Barium, calcium and magnesium doped mesoporous ceria supported gold nanoparticle for benzyl alcohol oxidation using molecular O₂

Abstract

In the era of sustainable energy, catalysis using gold nanoparticles has drawn considerable attention from world researchers. Oxidation of benzyl alcohol by molecular O₂ is an atom efficient path to synthesize benzaldehyde. Nanocrystalline ceria has been proven as a useful support to disperse gold nanoparticles since last few years, however there are a few reports on mesoporous ceria supported gold nanoparticles. In this work a systematic investigation was carried out to improve the activity of Au/CeO₂ catalyst by incorporating Ba^2+, Ca²⁺ and Mg²⁺ cations into the ceria lattice through a sol–gel procedure. Both the doped ceria and ceria supported gold nanoparticles are characterized by BET S.A, XRD, TEM, SAXS, XPS, TPR, CO₂-TPD techniques. BET S.A measurements show the mesoporous oxides where H3 hysteresis loops are found. The decrease in the crystallite size of ceria after doping by metal cations is observed in the XRD measurement. The TEM and HRTEM characterization shows the nanocrystalline particle size around 30–50 nm and gold nanoparticles around 10–15 nm in size. Distribution in the particle size for doped ceria have been obtained using SAXS measurements where narrow distributions of ceria particles are found in the 10–20 nm range. The existence of oxide vacancies and the mixture of Ce³⁺/Ce⁴⁺ oxidation states are observed for doped ceria materials in the XPS investigation. The strong gold-support interaction was also evidenced by XPS characterization where oxidic gold was found on the doped ceria surface. Lowering of the reduction peak in ceria after gold nanoparticle deposition was observed from TPR investigation whereas the change in basic site distribution is observed from CO₂ TPD experiment, instigating new insights into the surface properties of the catalysts. The catalytic activities of the catalysts were determined for benzyl alcohol oxidation reactions using molecular O₂. The catalytic activity was in the order of Au/Ba–CeO₂ > Au/Ca–CeO₂ > Au/Mg–CeO₂ > Au/CeO₂. The synergistic effect of gold nanoparticles and dopant cations to the ceria was explained in this work.

---

Introduction

Research on the unusual properties of gold nanoparticles compared to bulk gold has been well nurtured in the past few decades. Pioneering reports on hydrochlorination of acetylene by Prof. Hutchings¹ and the catalytic action of gold nanoparticles for CO oxidation by Prof. Haruta² had continued to nurture several other explorations and exploitations of gold nanoparticles for green chemistry practices.^3–6 For different reactions in which gold nanoparticles display catalytic activity, selective oxidation of alcohols to carbonyl compounds has attracted special attention because of its importance in organic synthesis at both laboratory and industrial level.^7–10 In particular, selective liquid phase aerobic oxidation of alcohols to corresponding carbonyl compounds using solid catalysts under mild conditions is a shared desire from an economic and environmental perspectives.^11,12

Synthesis and development of CeO₂ nanomaterials^13,14 with enhanced thermal, redox and acid–base properties is of paramount interest in designing catalysts for specific catalytic reactions. For example, in three-way catalysis (TWC) for the treatment of automotive internal combustion of engine exhaust gases, one of the most important properties of ceria is oxygen storage capacity, due to the Ce⁴⁺/Ce³⁺ redox cycles between its lattice ions.¹⁵ Doping of lower valent cations to the ceria is also a common strategy for tuning reducibility and OSC of ceria for lower temperature applications.^16–19 Oxygen storage capacity (OSC) of ceria is related to its reducibility which in turn depends on the morphology of its crystallites.

There is an attraction to use doped ceria as a support for gold nanoparticles in the oxidation reactions because of the anticipated synergistic effect between gold nanoparticles and ceria via its lattice oxygen vacancy.^20,21 Though there are a number of reports for using doped ceria as a support for gold nanoparticles in the case of CO oxidation reactions there are however few reports to show the role of dopant cation on the activity of oxidation of benzyl alcohol. In fact work on non siliceous mesoporous oxides have instigated the material researcher to develop synthetic strategies for different porous oxides.^22,23 A recent report by Chowdhury et al., revealed that doping can influence surface acid–base properties of mesoporous ceria and its catalytic behaviour.²⁴ This aspect of doped ceria materials has received less attention. Thus in this contribution, we attempt to investigate the influence of alkaline (Ba²⁺, Ca²⁺, Mg²⁺) doping on morphology, redox and acid–base properties of the mesoporous ceria which can tune the activity of gold nanoparticles. We introduced Ba²⁺, Ca²⁺, Mg²⁺ (dopant to cerium molar ratio 4/100) to the ceria lattice by sol–gel procedure using triethanolamine–water mixture as a solvent. It is anticipated that incorporation of the cations into a ceria lattice structure will influence redox and acid–base properties of ceria in favour of synergistic interactions with gold nanoparticles towards enhanced activity for aerobic oxidation benzyl alcohol to benzaldehyde.

Experimental

Catalyst preparation

Preparation of Ca, Mg and Ba doped –CeO₂ oxide support. The support material i.e. alkaline earth metal doped cerium oxide was prepared by the non-hydrothermal sol–gel method described in our previous publication.²⁴ In brief, a solution of Ce(NO₃)₃·6H₂O was added to triethanolamine under stirring conditions at room temperature. Then a nitrate precursor of the dopants was added to the solution. Afterwards the solution was stirred for 10 minutes after which tetraethylammonium hydroxide (TEAOH) was added dropwise under constant stirring conditions leading to a gel mixture of composition TEA (triethanolamine) :Ce :H₂O : Ba/Mg/Ca : TEAOH 0.2 : 0.1 : 1.1 : 0.004 : 0.1. The gel was aged for 24 h under stirring conditions then dried at 110 °C for 24 h and calcined at 700 °C for 10 h with temperature ramp 1 °C min⁻¹.

Deposition of gold on alkali metal doped CeO₂ support. The gold nanoparticles were deposited on the support by the procedure given by Haruta et al.²⁵ In this typical procedure 150 ml deionized water was taken in a beaker and a required amount (2.33 wt% in solution) of HAuCl₄·3H₂O was added to it. Then the solution was heated to 70 °C from the initial temperature of 30 °C. After attaining the temperature, the pH was raised to 7.0 from the initial pH 2.87 by adding aqueous NaOH solution dropwise. Then the support material (0.75 g) was added and the whole content was stirred for one hour keeping the pH constant at 7.0. Afterwards the material was filtered off, washed with water and dried in vacuum for 12 h followed by calcinations at 300 °C for 4 hours.

Catalyst characterization

BET Surface area and porosity measurement. The nitrogen adsorption–desorption isotherm of the samples were measured at liquid nitrogen temperature with a Quantachrome Autosorb-1C-TPD at 77 K. Pre-treatment of the samples was done at 473 K for 3 h under high vacuum. The surface area was determined by Brunauer–Emmett–Teller (BET) equation. Pore size distributions were calculated using NLDFT (Non linear Density Functional Theory) model of cylindrical pore approximation.

X-ray diffraction (XRD). XRD patterns are recorded at room temperature on a D8 ADVANCE (BRUKER AXS, Germany) diffractometer using CuK_α radiation with parallel beam (Gobel Mirror). The catalysts are ground to a fine powder prior to measurement. The scans are recorded in the 2θ range between 10–75° using a step size of 0.02° and scan speed of 2s/step. Peaks are identified by search match technique using DIFFRAC^plus software (BRUKER AXS, Germany) with reference to the JCPDS database. The software TOPAS 3.0 from Bruker AXS (2005) is used for refinement of cerianite diffraction peak (111) located at 2θ = 28.68° to determine the average crystallite size, in which it is possible to make all of the corrections and refinement processes. The software package TOPAS 3.0 uses the fundamental parameter approach (FPA), and is therefore capable of estimating the instrumental influence. The Double-Voigt approach in TOPAS is used with both calculated (FPA) and measured instrument functions. Employing integral breadth β_i, Balzar et al.,²⁶ have developed a more generalized treatment (β_i = λ/L_Volcos θ; λ: wave-length of X-ray, L_Vol: Volume weighted crystallite size, θ: angle of diffraction angle) of domain-size broadening which is independent of the crystallite shape. This conception directly leads to a formula identical with the Scherrer equation except that the constant assumes a value of unity.

Small angle X-ray scattering (SAXS). Small-angle X-ray scattering (SAXS) measurements were performed using a laboratory based SAXS instrument with Cu K_α X-ray source. Variation of scattering intensity with wave vector transfer (q = 4π sin(θ)/λ was measured for the powder samples.

Transmission electron microscope (TEM). Samples were prepared by placing sample mixture drops directly on the Formvar polymer coated grids using a micropipette. The Au NPs present in the aqueous mixture were allowed to settle, and the extra solvent was subsequently removed by placing a TEM grid on a neat filter paper and dried at ambient temperature condition for half a day. The morphology, size, and shape distribution of the Au NPs were recorded using TECNAI FE12 TEM instrument operating at 120 kV. In each image, more than 150 particles were analyzed using SIS imaging software to create size distribution histogram. The diffraction patterns were recorded at selected areas to determine the crystal structure and phases of crystals at 660 mm camera length.

High resolution transmission electron microscope (HRTEM). The HR-TEM investigation was done on JEOL JEM 2100 microscope operating at 200 KV acceleration voltage using lacey carbon coated Cu grid of 300 mess size.

X-ray photoelectron spectroscopy (XPS). X-ray photoelectron spectroscopy (XPS) measurements were carried on KRATOS-AXIS 165 instrument equipped with dual aluminium–magnesium anodes using Mg K radiation (hm 1/4 1253.6 eV) operated at 5 kV and 15 mA with pass energy 80 eV and an increment of 0.1 eV. The samples were degassed out for several hours in an XPS chamber to minimize air contamination to the sample surface. To overcome the charging problem, a charge neutralizer of 2 eV was applied and the binding energy of C 1s core level (BE 1/4 284.6 eV) of adventitious hydrocarbon was used as standard. The XPS spectra were fitted using a nonlinear square method with the convolution of Lorentzian and Gaussian functions after a polynomial background was subtracted from the raw spectra.

Inductively coupled plasma-optical emission spectroscopy (ICP-OES). An inductively coupled plasma optical emission spectrometer (ICP-OES, IRIS Intrepid II XDL, Thermo Jarrel Ash) was used to determine the concentrations of gold ions in the aqueous solutions.

Temperature programmed reduction (TPR). TPR profiles of the samples are recorded with ChemiSorb 2720 (Micrometrics, USA) equipped with a TCD detector. The TPR profiles are obtained by reducing the catalyst samples by a gas mixture of 10% H₂ in Ar with a flow rate of 20 mL min⁻¹ while the temperature is increased from ambient temperature to 700 °C at a rate of 10 °C min⁻¹.²⁷ Hydrogen consumption in the profiles were evaluated by peak area of CuO TPR calibrations.

Temperature programmed desorption (TPD). The TPD was carried out by ChemiSorb 2720 (Micrometrics, USA) equipped with a TCD detector which is heated under He flow from room temperature to 350 °C for 30 minutes. The dried catalyst was cooled to 80 °C. CO₂-TPD was obtained using 5.2% CO₂ in He for the CO₂ uptake. The amount of CO₂ desorbed in the TPD peak area was evaluated by pulses of CO₂ in a flow of He.

Catalyst activity studies. Evaluation of the catalysts for aerobic oxidation of benzyl alcohol oxidation was carried out in a 50 ml two necked round bottom flask. A reaction batch consisted of 4 mmol benzyl alcohol, 4 ml dry toluene as solvent, and 0.1 ml dodecane as an internal standard. The mixture was homogenized by vigorous shaking followed by addition of 0.1 g catalyst. The reaction flask is kept in the oil bath at 120 °C, magnetically stirred at 260 rpm and oxygen bubbles allowed into the reaction mixture at 50 ml min⁻¹ flow rate (maintained by AALBORG mass flow controller). After 3 h run, the content of the flask was diluted with acetone and centrifuged to separate the catalyst. Analysis of the mixture was carried out using a Gas Chromatograph (CIC-India) equipped with SE-30 column at 200 °C oven temperature, injector and detector (FID) temperature at 225 °C.

Results and discussions

BET S.A and porosity measurement

The representative results obtained from BET S.A and porosity analysis for Ca–CeO₂ are depicted in Fig. 1 and Table 1. A close look at the isotherm reveals that there is much less adsorption in the low pressure region and the hysteresis loop is almost like an H3 isotherm. This may be attributed to the so called pore blocking or percolation effect.²⁸ The effect is expected to occur when the pore has access to the external surface only through a narrower neck as in an ink bottle pore.

Fig. 1 BET S.A and pore size distribution of Ca–CeO₂ (Ca/Ce = 4/100).

Table 1 BET S.A and porosity result

Catalyst	Surface area (m^2 g^−1)	Pore size (nm)	Pore volume (cc g^−1)
CeO2	4	—	—
CaCeO2	20	8.84	0.03
BaCeO2	10	3.06	0.01
MgCeO2	7	2.98	0.01

---

The wide body of an ink bottle pore is filled at the vapour pressure which corresponds to the delayed condensation and remains filled during desorption until the narrow neck empties first at a lower vapour pressure. Thus in a network of ink bottle pores evaporation of the capillary condensate is obstructed by the pore necks. As observed in the case of Ba–CeO₂, Mg–CeO₂ (ESI†) and Ca–CeO₂ the porous network is different in each case because of surface heterogeneity, particle size varies in one to another. For both the Ba–CeO₂ and Mg–CeO₂ surface area is less however for Ca–CeO₂ the surface area is 20 m² g⁻¹.

The presence of mesoporosity is confirmed from the pore size distribution where the NLDFT (Non linear density functional theory) model²⁹ is adopted to calculate the pore size assuming the cylindrical pores. The surface area of CeO₂ is very low whereas higher surface area, different porosity and pore volume in three different samples indicate the role of dopant alkaline cation during sol–gel synthesis as the gelation step is highly dependent on the basicity of the medium.

X-ray diffraction (XRD)

The XRD pattern of CeO₂ and Ba, Ca and Mg doped CeO₂ oxides are depicted in Fig. 2. The peaks of all the samples could be indexed as (111), (200), (220), (311), (222), (400) planes of cubic fluorite structure (Space Group: Fm3m, JCPDS 34-0394) of CeO₂. As depicted in Table 2 the particle sizes of doped CeO₂ are found to be lower for all the doped CeO₂. The crystallite size was lowest for Ca–CeO₂ compared to Mg–CeO₂ and Ba–CeO₂ as calculated from XRD analysis. The higher surface area and mesoporosity were observed for Ca–CeO₂ as described in the earlier section. The small angle X-ray scattering (SAXS), which is described in the later section, was done to verify the size distribution of the doped ceria. The lattice parameter of the BaCeO₂ was increased compared to MgCeO₂ which strongly emphasizes the incorporation of Ba to the CeO₂ as the size of Ba²⁺ is higher (1.49 Å) than Mg²⁺ (0.86 Å).

Fig. 2 XRD pattern of the (a) CeO₂, (b) Ba–CeO₂ (Ba/Ce = 4/100), (c) Mg–CeO₂ (Mg/Ce = 4/100), (d) Ca–CeO₂ (Ca/Ce = 4/100).

Table 2 Quantitative XRD analysis

Catalyst	Phase	d(111) (Å)	Crystallite size (Å)
MgCeO2	Cerianite-(Ce)	3.1086	286
BaCeO2	Cerianite-(Ce)	3.1153	205
CaCeO2	Cerianite-(Ce)	3.0983	112
CeO2	Cerianite-(Ce)	3.1106	547

---

The non existence of any peak corresponding to the individual oxide or hydroxides of Ba, Ca and Mg reveal the formation of nanocrystalline solid solution of the metal cations into the ceria lattice. Tsunekawa³⁰et al. investigated the lattice expansion of the CeO₂ nanoparticles, and they found that “the reduction of the valence induces an increase in the lattice constant due to the decrease in electrostatic forces”. However, Rietveld refinement of the XRD pattern will not be an ideal to investigate whether the dopant (Ba²⁺, Ca²⁺ and Mg²⁺) is dispersed on the ceria surface and its local chemical environment. Therefore XPS characterization was undertaken which is described in the later section.

TEM (transmission electron microscopy) and HRTEM (high resolution transmission electron microscopy)

The TEM images obtained from CeO₂, Ba–CeO₂, Mg–CeO₂ and Ca–CeO₂ materials are displayed in Fig. 3. The nanocrystalline particles are observed in 20–30 nm size range similarly to the results obtained from earlier XRD studies. In the present case the preparation procedure is a sol–gel process using triethanolamine–H₂O mixture which we have adopted instead of a precipitation method. In the precipitation method because of the strong basicity and higher charge, Ce⁴⁺ ions usually undergo strong hydration resulting in the formation of [Ce(H₂O)x(OH)y]^(4−y)+ complex, where (x + y) is the coordination number of Ce⁴⁺. Now in the preparation procedure adopted here there was no pinkish purple colour noticed, it is expected that Ce ⁴⁺ did not form any complex with H₂O and OH⁻ despite the fact that both H₂O and OH⁻ ions were present in the solution. As reported in the literature³¹ there might be a possibility of [Ce(H₂O)n(OH)y]^(4−y)+ and [Ce(TEA)₂(NO₃)](NO₃)₂ complex in the solution. Rapid deprotonation of [Ce(H₂O)n(OH)y^(4−y)+ has a significant effect on the morphology of the nanocrystalline CeO₂ powder due to the polymeric nature of [Ce(H₂O)n(OH)y]^(4−y)+ which might act as a nucleation centre.

Fig. 3 TEM image of (a) CeO₂, (b) Ca–CeO₂(Ca/Ce = 4/100), (c) Mg–CeO₂(Mg/Ce = 4/100), (d) Ba–CeO₂(Ba/Ce = 4/100).

Therefore the presence of highly basic cations might influence the complex formation of TEA and Ce⁴⁺ resulting in different morphologies as revealed in the TEM study. In fact during nanoparticle formation, nucleation and growth are the main cause towards its size distribution. As reported by Jiang et al.³² for the synthesis of ZnO nanoparticles, TEA hinders the anisotropic growth by adsorbing on its crystal plane and forming microspheres.

The TEM images and particle size distribution of gold nanoparticles for Au/CeO₂, Au/Ba–CeO₂, Au/Ca–CeO₂ and Au/Mg–CeO₂ are presented in Fig. 4. The Au/Ba–CeO₂ had smaller particles in the size range 3–4 nm compared to other gold deposited supports.

Fig. 4 TEM image and particle size distribution of (a) Au/CeO₂, (b) Au/Mg–CeO₂(Mg/Ce = 4/100), (c) Au/Ca–CeO₂(Ca/Ce = 4/100), (d) Au/Ba–CeO₂(Ba/Ce = 4/100).

The HRTEM image of Au/Ba–CeO₂ was shown in Fig. 5 where smaller gold nanoparticles in the 3–4 nm range are distinctly visible.

Fig. 5 HRTEM image of (a) and (b) Au/Ba–CeO₂ (Ba/Ce = 4/100).

Small angle X-ray scattering (SAXS) studies

The scattered intensities I(q) were recorded as a function of scattering vector q(= 4π sin(θ)/λ, where 2θ is the scattering angle, and λ is the X-ray wavelength). Fig. 6 shows the SAXS profiles of all the samples. To obtain the particle size distribution, SAXS data were analyzed in the light of a polydisperse spherical particle model. For such a case, I(q) may be approximated as
where, P(q,R) represents the form factor of a spherical particle of radius R, i.e.,
D(R) represents the particle size distribution, i.e., D(R)dR indicates the probability of having size R to R + dR.

Fig. 6 SAXS data from the Ceria samples.

In the present case, a standard log normal distribution was assumed. S(q,R_av) represents the inter-particle structure factor with R_av representing average R value. In the present case, a fractal like³³ interparticle structure factor was assumed for Ca–CeO₂. For other CeO₂, Ba–CeO₂, Mg–CeO₂, consideration of structure factor was not required during fitting of the model to the experimental data. The fitting of the above model to the experimental data are shown in the Fig. 7.

Fig. 7 Fitting of the model to the experimental SAXS data is compared (solid line represents the fit in each case).

The estimated size distributions are plotted in Fig. 8. It is found that except Mg–CeO₂ all the other samples have a narrow size distribution indicating the surface homogeneity in those cases.

Fig. 8 Particle size distribution as obtained from the fitting of the model curve to the experimental SAXS data.

From SAXS, the data specific surface area (Σ = S/V) was calculated using the formula results as shown in Table 3. The higher surface area of Ca–CeO₂ supports the results obtained from BET measurement studies.

Table 3 Quantitative SAXS analysis

Catalyst	Specific surface area
CeO2	5.8 × 10^8 m^−1
CaCeO2	8.2 × 10^8 m^−1
BaCeO2	7.0 × 10^8 m^−1
MgCeO2	5.6 × 10^8 m^−1

---

X-ray photoelectron spectroscopy (XPS)

XPS has remained as a pivotal technique to identify the oxidation state, chemical environment of the element present in the surface.^34,35 No peak due to residual Na and Cl are observed in the XPS survey (ESI†). The presence of Ba, Ca and Mg over the ceria surface is confirmed using XPS spectroscopy and the results are shown in Fig. 9.

Fig. 9 XPS spectra for Ba (3d), Ca (2p), Mg (2p) core level electrons.

The two peaks due to Ba 3d_5/2 and 3d_3/2 are observed at 779.9 eV and 795.3 eV respectively in the Ba–CeO₂ catalyst. The peak around 779.0 eV can be assigned to the Ba²⁺ species.³⁶ The Mg 2p core electron binding energy was around 50.397 eV in the case of Mg–CeO₂ which indicates the existence of all the Mg species in the Mg²⁺ state.³⁷ In the Ca–CeO₂ catalyst, the Ca 2p core electron shows binding energy around 344.9 eV and another one around 348.0 eV which is typical of the Ca²⁺ oxidation state.³⁸ To investigate the nature of the interaction between the metal center (Ba²⁺, Ca²⁺ and Mg²⁺) in the doped ceria before and after gold deposition, we need to analyze the Ce (3d), O (1s) and Au (4f) core electron binding energies as described in the following section.

The core level binding energy for Ce 3d_5/2 and Ce 3d_3/2 of the samples as well as quantitative analysis show some interesting features (ESI†). Cerium compounds are known to exhibit rather complex features due to hybridization with ligand orbitals and fractional occupancy of the valence 4f orbitals.²⁴ The peaks around 882.4 eV, 887 eV, 890 eV and 897.9 eV are denoted as V, V′, V′′, V′′′ where as peaks at 900.4 eV, 904.1 eV, 908.1 eV and 916.1 eV are denoted by U, U′,U′′ and U′′′. The peaks labelled U are due to 3d_3/2 spin-orbit states, and those labelled V are due to the corresponding 3d_5/2 states. The U′′′/V′′′ doublet is due to the primary photoemission from Ce(IV)O₂i.e. Ce 3d⁹ 4f⁰ O 2p⁶ Ce(IV). The U/V and U′′/V′′ doublets are shakedown features resulting from the transfer of one or two electrons from a filled O 2 p orbital to an empty Ce 4f orbital, i.e. Ce 3d⁹ 4f² O 2p⁴ and Ce 3d⁹ 4f¹ O 2p⁵ Ce(IV) in the final states. The U′/V′ doublet is due to photoemission from Ce cations, i.e. Ce 3d⁹ 4f¹ O 2p⁶ of Ce (III). The spin-orbit splitting of V′′/U′′ is 18.0 eV is in the same order reported in literature.³⁹ The presence of U′′/V′′ doublet peaks in the spectrum indicates that the sample contains some oxygen vacancies and is in a partially reduced state. The Ce 3d spectrum basically denotes a mixture of Ce³⁺/Ce⁴⁺ oxidation states giving rise to a myriad of peaks indicating that the surface of the sample is not fully oxidized.⁴⁰ It is also proposed that the concentration of oxygen vacancies in ceria is related to the particle sizes, and smaller nanoparticles possess a higher concentration of oxygen vacancies. As the hierarchically structured ceria fabricated in this work consist of nanosized crystals of ceria as building units which is evidenced from XRD, TEM and SAXS investigations in the earlier section, a high concentration of the oxygen vacancy would be expected. Indeed, recent theoretical calculations of unsupported CenO_2n−x(n) clusters by Loschen et al.⁴¹ showed the formation of Ce³⁺ ions on the corner and edge sites, the relative amount of which increases with decreasing particle size. Interestingly we found that Ca–CeO₂ had the smallest particle size from the XRD measurement which has a maximum Ce³⁺/Ce⁴⁺ ratio as observed from XPS quantitative analysis. Apart from that partial photoreduction of CeO₂ during XPS measurements is a well-known fact in the literature. The reduction is mainly due to the progressive elimination of surface hydroxyls and oxygen ions from the CeO₂ surface upon vacuum treatment.⁴⁰ The Ce 3d spectrum after gold deposition shows a shift in the binding energy of the peak due to Ce³⁺ for Ba–CeO₂, Ca–CeO₂, Mg–CeO₂ towards higher value compare to Au/CeO₂. This clearly indicates the strong gold-support interaction and also highlights synergistic behaviour of both gold and dopant cation. Therefore we need a look to the XPS pattern of Au species in the later section.

The representative O (1s) spectrum of CeO₂ and metal (Ca, Mg, Ba) doped CeO₂ have been investigated (ESI†). Though there is a possibility of several types of oxygen species in the doped ceria however the O 1s spectra consisted mainly of three peaks; a dominant one at energy 529.2 eV, another at 531.4 eV along with a peak at 533.5 eV. The one at 529.2 eV is the oxygen present in the CeO₂ lattice. The peak at 531.0 eV is due to Ce₂O₃ which is more covalent in nature. The weak third peak at a higher binding energy may be due to adsorbed oxygen species (from adsorbed water).⁴² There is a shift of O1s peak (529.0 eV) towards the lower energy side after Ca²⁺ and Mg²⁺ insertion however for Ba²⁺ the position remained almost unchanged. As doping of lower valent cations either to the CeO₂ lattice or Ce₂O₃ lattice generate excess negative charge shifting the O (1s) binding energy to the lower binding energy side. Interestingly for the BaCeO₂ sample the position of the 531.0 eV peak shifted to higher value which may due to a change in the covalent nature of the ceria–oxygen bond.

The O (1s) spectra obtained for Au deposited undoped and doped ceria (ESI†) show a large shift of O (1s) binding energy (both for Ce₂O₃ and CeO₂) towards higher values for Ca²⁺ and Mg²⁺ doped samples. Though for the Ba²⁺ doped sample there is a shift to the higher side but it is to a lesser extent compared to other counterparts. From the results it can be speculated that lattice oxygen is tightly bound for Au/Ca–CeO₂ and Au/Mg–CeO₂ but for Au/Ba–CeO₂ it is weakly bound. Also it is interesting to note that interactions between gold and doped ceria are synergistic in nature where the dopant cation is having a decisive role towards the binding energy shift. May be the gold nanoparticles reside on the oxide vacancies in the doped ceria resulting strong interaction between gold and ceria. It is well known that in the oxidation reaction, nucleophilic lattice oxygen can be inserted into the activated organic molecule leading to oxygenated product.

The XPS spectrum obtained for Au (4f) core electrons is depicted in Fig. 10. The major peaks are in 83.0–84.0 eV, 86.7–88.0 eV. The binding energies of the gold metal were reported as 4f level at 83.9 eV (Au 4f_7/2) and 87.7 eV (Au 4f_5/2). It is well documented in the literature that Au species present as metallic Au⁰, Au⁺ and Au³⁺ species depending on the size of gold nanoparticles and the nature of supports. As reported by Visco et al.⁴³ both Au⁰ and Au³⁺ species was found in Au/Fe₂O₃ catalyst whereas Soares et al.⁴⁴ reported that for Au/TiO₂ catalysts prepared by the deposition–precipitation method, there was only metallic gold present.

Fig. 10 Au 4f core level XPS spectra of (a) Au/ Ba–CeO₂ (Ba/Ce = 4/100), (b) Au/Mg–CeO₂ (Mg/Ce = 4/100), (c) Au/Ca–CeO₂ (Ca/Ce = 4/100), (d) Au/CeO₂.

When gold nanoparticles are deposited on Ba–CeO₂, Ca–CeO₂ and Mg–CeO₂ the Au 4f binding energy is shifted towards higher values indicating formation of more oxidic gold on the catalyst surface. The absence of a chlorine peak in the photoelectron spectra indicates the elimination of chlorine during washing. The presence of peakes due to oxidic gold must be attributed to Au–oxygen-containing species of the Au⁺ and Au³⁺ types. Oxidic gold species pull the electron from oxygen thereby increasing the core electron binding energy as evidenced from Fig. 10.

The unresolved issue is the oxidation state of the Au active site during the oxidation reaction. Some researchers claim the ionic Au as the active species.^45,46 One model proposed is an ensemble consisting of metallic Au and Au–OH species.^47,48 In contrast, other researcher proposals that the active site is metallic Au, which is based on the fact that the only detectable species is metallic Au on an active catalyst.

TPR (temperature programmed reduction)

Fig. 11 shows the TPR profiles of doped and undoped ceria supports. The profiles depict typical reduction peaks at about 550 °C and 800 °C. Peaks between the range at 0 ≤ x ≤ 600 °C are relevant to redox properties of ceria, peak temperatures and hydrogen consumptions of these peaks provides a measure of reducibility and OSC of supports,⁴⁹ hence our focus is on the peaks in this temperature for the discussion of the catalytic properties of the supports. We found that alkaline cation doping enhanced the OSC of ceria in order of increasing ionic radius (Fig. 12).

Fig. 11 H₂-TPR of (a) Au/CeO₂, (b) Au/Mg–CeO₂ (Mg/Ce = 4/100), (c) Au/Ca–CeO₂ (Ca/Ce = 4/100), (d) Au/Ba–CeO₂ (Ba/Ce = 4/100), (e) CeO₂, (f) Mg–CeO₂ (Mg/Ce = 4/100), (g) Ca–CeO₂ (Ca/Ce = 4/100), (h) Ba–CeO₂ (Ba/Ce = 4/100).

Fig. 12 Percentage in H₂ consumption of doped relative to undoped ceria.

Theoretical calculations by Kehoe et al., indicated the same order for reduction energies for alkaline cations doped ceria.⁵⁰Fig. 13 showed that ceria reducibility is virtually unaffected for Mg–CeO₂, Ca–CeO₂ but enhanced by barium doping. The reduction peaks are lowered by 400 °C (Fig. 11) when gold nanoparticles are deposited on the supports. It appears that the gold nanoparticles are located in the oxygen vacancies in the supports. This close proximity and hydrogen spill-over effect by gold nanoparticles account for lowering of the reduction temperature.

Fig. 13 H₂ consumption from TPR profiles of the catalysts (a) CeO₂, (b) Mg–CeO₂ (Mg/Ce = 4/100), (c) Ca–CeO₂ (Ca/Ce = 4/100), (d) Ba–CeO₂ (Ba/Ce = 4/100), (e) Au/CeO₂, (f) Au/Mg–CeO₂ (Mg/Ce = 4/100), (g) Au/Ca–CeO₂ (Ca/Ce = 4/100), (h) Au/Ba–CeO₂ (Ba/Ce = 4/100).

Fig. 14 shows CO₂ desorption profiles of the ceria supports and corresponding gold deposited catalysts. Desorption temperature and amount of CO₂ desorbed are measures of basic strength and site density, respectively. For comparative analysis, the profiles are demarcated into temperature ranges: 50–250 °C, 250–500 °C and 500–800 °C and are designated as weak, medium and strong basic sites, respectively. Although the basicity of alkaline cations increases with increasing order of ionic radius, basic site distribution of the supports and the catalysts did not depict the same trend (Fig. 15). Basicity of doped material depends on the nature of defects created and coordination behaviour of dopants. It is observed from XPS studies described in an earlier section, that there are different types of electronic interaction between dopant cation, cerium, oxygen and gold atoms.

Fig. 14 CO₂-TPD of (a) Au/CeO₂, (b) Au/Mg–CeO₂ (Mg/Ce = 4/100), (c) Au/Ca–CeO₂ (Ca/Ce = 4/100), (d) Au/Ba–CeO₂ (Ba/Ce = 4/100), (e) CeO₂, (f) Ca–CeO₂ (Ca/Ce = 4/100), (g) Ba–CeO₂ (Ba/Ce = 4/100), (h) Mg–CeO₂ (Mg/Ce = 4/100).

Fig. 15 CO₂ desorbed from TPD profiles of the catalysts (a) CeO₂, (b) Mg–CeO₂ (Mg/Ce = 4/100), (c) Ca–CeO₂ (Ca/Ce = 4/100), (d) Ba–CeO₂ (Ba/Ce = 4/100), (e) Au/CeO₂, (f) Au/Mg–CeO₂ (Mg/Ce = 4/100), (g) Au/Ca–CeO₂ (Ca/Ce = 4/100), (h) Au/Ba–CeO₂ (Ba/Ce = 4/100).

It had been shown that dopants in ceria prefer to assume coordination of their own oxides rather than that of cubic coordination of Ce⁴⁺.⁵⁰ In addition, the nature of defect also depends on the relative size of dopant and doped ions. Because of its smaller size, Mg²⁺ may be located in the lattice interstitials rather than substituting the lattice site of Ce⁴⁺.⁵¹ This may account for why Mg-doped supports fall out of expected trends based on ionic radius. It also interesting that the site distributions change dramatically after gold deposition on the supports. Densities of medium and strong sites in undoped, Mg and Ba doped ceria decrease after gold deposition. They represent preferential location of the deposited gold nanoparticles. However, gold deposition increases the basicity of Ca-doped ceria. Surface basicity of a metal oxide is related to the ionic character of it surface oxygen species. As observed from the quantitative XPS analysis (ESI†) of O (1s) core electron binding energy there is a change in the percentage of ionic contribution of Au/Ca–CeO₂ correlating to a higher basicity of the Au/Ca–CeO₂ catalyst.

Catalytic activity results

The catalyst activity for benzyl alcohol oxidation using molecular O₂ was measured for all of the samples. All the catalysts have high selectivity towards benzaldehyde as shown Table 4. The CeO₂ had very low activity which was more for Ca–CeO₂, Ba–CeO₂, Mg–CeO₂. However, it is very difficult to compare the activity of the above three catalysts. No leaching of dopants was observed for doped CeO₂ as evidenced from EDS measurement (ESI†). There is a remarkable increase in the activity of doped ceria after gold deposition. The activity of Au/M–CeO₂ (M = Ca, Ba, Mg) was higher compared to Au/CeO₂ instigating a close look for the synergistic effect of gold nanoparticles and dopant cation.

Table 4 Result of benzyl alcohol oxidation reaction

Reactant	Catalyst	Actual gold loading (% Au)^a	Conversion (%)	TOF^# (h^−1)	Selectivity (%)
a Determined from ICP-OES measurement. Reaction condition: Toluene solvent, 4 ml; benzyl alcohol, 4 mmol; catalyst, 100 mg; O2, 50 ml min^−1; reaction time, 3 h; dodecane used as an internal standard.
PhCH2OH	Mg–CeO2 (4 mol% doped)	—	7.73	—	>99
PhCH2OH	Ba–CeO2 (4 mol% doped)	—	9.62	—	>99
PhCH2OH	Ca–CeO2 (4 mol% doped)	—	7.62	—	>99
PhCH2OH	CeO2	—	2.99	—	>99
PhCH2OH	2.33 wt% of Au–CeO2	0.02	7.99	1049.58	>99
PhCH2OH	2.33 wt% of Au–MgCeO2	0.10	13.07	343.25	>99
PhCH2OH	2.33 wt% of Au–BaCeO2	0.03	36.80	3221.70	>99
PhCH2OH	2.33 wt% of Au–CaCeO2	0.32	25.00	205.13	>99

---

It is important to note that the charge transfer from supports to deposited gold clusters is crucial for enhanced catalytic activities. The activity of gold catalysts in the oxidation reactions is somewhat surprising, because dissociatively adsorbed oxygen is generally thought to be the active oxidant in related reactions on other noble metals.⁵² The origin of the catalytic activity in supported gold nanoparticles is still not well understood, with many questions remaining about the site of the reaction and the nature of the interaction between gold nanoparticles and the support on supported gold catalysts. It has been reported that there is a correlation between the stabilization of superoxide species and gold cationic sites for CO oxidation reaction.⁵³

The Au/Ba–CeO₂ catalyst is having the highest activity. The gold nanoparticles distributed over Ba–CeO₂ supports have smaller sizes compare to other supports as evident from TEM and HRTEM studies described earlier. From the XPS characterization it is found that oxygen species is more labile in the Au/Ba–CeO₂ catalyst. Also the TPR results infer that deposition of gold nanoparticles make Au/Ba–CeO₂ more reducible as lowering of the reduction temperature is observed. As reported in the literature the adsorption of molecular O₂ plays a dominant role for CO oxidation reaction⁵⁴ and propylene epoxidation reaction.⁵⁵ Therefore the gold nanoparticles acts as a centre for oxygen adsorption which spill over to the Ba–CeO₂ surface making it highly active for alcohol oxidation. The basic site distribution on the Ba–CeO₂ catalyst decreases after gold nanoparticle deposition, however similar observations were prominent for the Au/Mg–CeO₂ catalyst. The Au/Mg–CeO₂ catalyst has the lowest activity levels so the change of distribution of basic sites is responsible for the lower activity of the Au/Mg–CeO₂ catalyst and redox sites are responsible for the Au/Ba–CeO₂ catalyst. It has been shown that dehydrogenation of alcohols to aldehydes can take on metal oxides catalysts.⁵⁶ The conversion involved the initial alkoxy intermediate formation on weak Lewis acid–strong Brønsted base site pairs. But the reaction is limited by removal of hydrogen. Thus, catalytic performance CeO₂ for aerobic oxidation of benzyl alcohol can be attributed to dehydrogenation on acid–base site pairs followed by hydrogen removal as H₂O by weakly coordinated surface oxygen ions. The used surface oxygen ions are replenished by dissociative adsorption of molecular oxygen to complete the catalytic cycle.

Though gold nanoparticles promote the catalytic activity of Mg–CeO₂ compared to undoped Mg–CeO₂, however more decreases in the basic sites cause the lowering of activity compared to Au/BaCeO₂. The higher activities of Au/Ca–CeO₂ compared to Au/Mg–CeO₂ could not be explained from TPR and TPD results. A quantitative estimation of % of Ce³⁺/ Ce⁴⁺ shows that the population of Ce ³⁺ is more in the Au/Ca–CeO₂ samples compare to Au/Ba–CeO₂ and Au/Mg–CeO₂ samples. As observed from textural properties, the Ca–CeO₂ has a higher surface area and pore size compared to Ba–CeO₂ and Mg–CeO₂ and having a smaller crystallite size compared to other counterparts as evidenced from XRD and TEM studies. The different type of gelation during the growth of Ca doped ceria might induce different electronic properties as found in the XPS results. The larger population of Ce³⁺ ions (ESI†) indicates more oxide vacancies in the Au/Ca–CeO₂ catalyst compare to Au/Mg–CeO₂ might result easily in the diffusion of molecular oxygen during the reaction demonstrating higher catalytic activity. Since the presence of Au increases the activity of the doped CeO₂, Au nanoparticles most probably constitute the dominant active sites here (involved in the complete catalytic cycle) while the acid–base and redox sites on the CeO₂ play complementary roles.

Conclusions

In summary, undoped and alkaline metal (Ba, Ca, Mg) doped mesoporous ceria were prepared by modified sol–gel methods using a triethanolamine–water mixture as a solvent. The Ca–CeO₂ has a higher surface area, pore size as observed from BET S.A and porosity measurements studies. Non existence of any peak due to Ba, Ca and Mg containing compounds in the XRD pattern emphasize the formation of nanocrystalline ceria containing solid solutions. The TEM characterizations reveal the highly crystalline nanoparticles in the range of 20–30 nm. The gold particles are smaller in size when deposited over Ba–CeO₂ surfaces compared to CeO₂, Ca–CeO₂ and Mg–CeO₂ supports. The role of dopant cations to the chemical environment of the ceria lattice are highlighted from the XPS investigation. A large number of oxide vacancies along with a mixed valence Ce³⁺/Ce⁴⁺ species are observed in the XPS studies. The strong gold-support interaction resulting in the shift of Ce 3d core electron peaks are observed when both the gold nanoparticles and dopant cations are present. The TPR studies show that Ba–CeO₂ is more reducible compared to Mg–CeO₂ and Ca–CeO₂. In all cases the reduction temperature of doped ceria is lowered by gold deposition showing higher catalytic activity towards benzyl alcohol oxidation. The reducibility of the Au/Ba–CeO₂ catalyst was more compared to the Au/Ca–CeO₂ and Au/Mg–CeO₂ catalyst. Simultaneously, the lattice oxygen atoms are more labile in the case of Au/Ba–CeO₂ which is more active in the benzyl alcohol oxidation reaction. The alteration of basic sites due to gold deposition is observed from the CO₂ TPD studies. Though the basic site density is decreased for Au/Ba–CeO₂ and Au/Mg–CeO₂ catalysts but for Au/Ca–CeO₂ it is increased. The higher activity of the Au/Ca–CeO₂ catalyst compare to Au/Mg–CeO₂ is due to the large number of oxygen vacancies leading to the highest Ce³⁺/Ce⁴⁺ ratio as evidenced from XPS measurements.

Acknowledgements

BC would like to acknowledge the financial grant from DST, Govt of India under the Fast Track Young Scientist Scheme (project no SR/FTP/ETA-16/2007-08). CS would like to acknowledge UGC, Govt of India for providing research fellowship. Authors acknowledge Prof. Asim Bhaumik (IACS, Calcutta) for helping them to carry out BET S.A porosity measurement experiments.

References

1. G. J. Hutchings, J. Catal., 1985, 96, 292.
2. M. Haruta, T. Kobayashi, H. Sano and N. Yamada, Chem. Lett., 1987, 405.
3. D. T. Thompson, Gold Bulletin. (Berlin, Ger.), 1998, 31, 111.
4. C. B. Geoffrey and D. T. Thompson, Catal. Rev. Sci. Eng., 1999, 41, 319.
5. C. B. Geoffrey, Gold Bulletin. (Berlin, Ger.), 2001, 34, 117.
6. G. J. Hutchings, Gold Bulletin. (Berlin, Ger.), 2004, 37, 3.
7. T. Mallat and A. Baiker, Chem. Rev., 2004, 104, 3037.
8. T. Punniyamurthy, S. Velusamy and J. Iqbal, Chem. Rev., 2005, 105, 2329.
9. A. V. Sánchez, J. G. Á. Zárraga and J. Mex, Chem. Soc., 2007, 51, 213.
10. C. P. Vinod, K. Wilson and A. F. Lee, J. Chem. Technol. Biotechnol., 2011, 86, 161.
11. R. J. J. Jachuck, D. K. Selvaraj and R. S. Varma, Green Chem., 2006, 8, 29.
12. P. Maity, D. Mukesh, S. Bhaduri and G. K. Lahiri, J. Chem. Sci., 2009, 121, 377.
13. M. Epifani, E. Pellicer, J. Arbiol and J. R. Morante, Chem. Mater., 2009, 21, 862.
14. X. Liu, K. Zhou, L. Wang, B. Wang and Y. Li, J. Am. Chem. Soc., 2009, 131, 3140.
15. A. Trovarelli, Catal. Rev. Sci. Eng., 1996, 38, 439.
16. Z. Hu and H. Metiu, J. Phys. Chem. C, 2011, 115, 17898.
17. A. B. Kehoe, D. O. Scanlon and G. W. Watson, Chem. Mater., 2011, 23, 4464.
18. Z. Hu, B. Li, X. Y. Sun and H. Metiu, J. Phys. Chem. C, 2011, 115, 3065.
19. N. Qiu, J. Zhang, Z. Wu, T. Hu and P. Liu, Cryst. Growth Des., 2012, 12, 629.
20. A. Abad, P. Concepcióń, A. Corma and H. Garcia, Angew. Chem., Int. Ed., 2005, 44, 4066.
21. J. Guzman, S. Carrettin and A. Corma, J. Am. Chem. Soc., 2005, 127, 3286.
22. A. K. Patra, S. K. Das and A. Bhaumik, J. Mater. Chem., 2011, 21, 3925.
23. S. K. Das, M. K. Bhunia and A. Bhaumik, Dalton Trans., 2010, 39, 4382.
24. G. Postole, B. Chowdhury, B. Karmakar, K. Pinki, J. Banerji and A. Auroux, J. Catal., 2010, 269, 110.
25. B. Chowdhury, K. K. Bando, J. J. Bravo-Suárez, S. Tsubota and M. Haruta, J. Mol. Catal. A: Chem., 2012, 359, 21.
26. D. Balzar, N. Audebrand, M. R. Daymond, A. Fitch, A. Hewat, J. I. Langford, A. Le. Bail, D. Louër, O. Masson, C. N. McCowan, N. C. Popa, P. W. Stephens and B. H. Toby, J. Appl. Crystallogr., 2004, 37, 911.
27. S. Liang, F. G. Teng, Bulgan, R. Zong and Y. Zhu, J. Phys. Chem. C, 2008, 112, 5307.
28. G. Ertl, H. Knozinger, F. Schuth and J. Weitkmp, Hand book of heterogeneous catalyst, WILEY-VCH, Weinheim, 2nd edn, 2008, vol. 2, p. 728.
29. P. I. Ravikovitcha, G. L. Haller and A. V. Neimarka, Adv. Colloid Interface Sci., 1998, 203, 76.
30. S. Tsunekawa, K. Ishikawa, Z. Q. Li, Y. Kawazoe and A. Kasuya, Phys. Rev. Lett., 2000, 85, 3440.
31. R. K. Pati, I. C. Lee, K. J. Gaskell and S. H. Ehrman, Langmuir, 2009, 25, 67.
32. H. Jiang, J. Hu, F. Gu and C. Li, J. Phys. Chem. C, 2008, 112, 12138.
33. J. Texeira, J. Appl. Crystallogr., 1988, 21, 781.
34. B. M. Reddy, B. Chowdhury, I. Ganesh, E. P. Reddy, T. C. Rojas and A. Fernandez, J. Phys. Chem. B, 1998, 102, 10176.
35. B. M. Reddy, B. Chowdhury and P. G. Smirniotis, Appl. Catal., A, 2001, 219, 53.
36. U. A. Joshi, S. Yoon, S. Baik and J. S. Lee, J. Phys. Chem. B, 2006, 110, 12249.
37. J. S. Valente, E. Lima, J. A. Toledo-Antonio, M. A. Cortes-Jacome, L. Lartundo-Rojas, R. Montiel and J. Prince, J. Phys. Chem. C, 2010, 114, 2089.
38. M. Vallet-Regi, J. Pérez-Pariente, I. Izquierdo-Barba and J. A. Salinas, Chem. Mater., 2000, 12, 3770.
39. V. Shapovalov and H. Metiu, J. Catal., 2007, 245, 205.
40. A. K. Sinha and K. Suzuki, J. Phys. Chem. B, 2005, 109, 1708.
41. C. Loschen, S. T. Bromley, K. M. Neyman and F. Illas, J. Phys. Chem. C, 2007, 111, 10142.
42. B. M. Reddy, G. Thrimurthulu, L. Katta, Y. Yamada and S. E. Park, J. Phys. Chem. C, 2009, 113(36), 15882.
43. A. M. Visco, F. Neri, G. Neri, A. Donato, C. Milone and S. Galvagno, Phys. Chem. Chem. Phys., 1999, 1, 2869.
44. J. M. C. Soares, P. Morrall, A. Crossley, P. Harris and M. Bowker, J. Catal., 2003, 219, 17.
45. E. D. Park and J. S. Lee, J. Catal., 1999, 186, 1.
46. R. M. Finch, N. A. Hodge, G. J. Hutchings, A. Meagher, Q. A. Pankhurst, M. R. H. Siddiqui, F. E. Wagner and R. Whyman, Phys. Chem. Chem. Phys., 1999, 1, 485.
47. G. C. Bond and D. T. Thompson, Gold Bull. (Berlin, Ger.), 2000, 41, 33.
48. C. K. Costello, M. C. Kung, H. S. Oh, Y. Wang and H. H. Kung, Appl. Catal., A, 2002, 232, 159.
49. G. R. Rao and B. G. Mishra, Bulletin of the Catalysis Society of India, 2003, 2, 122.
50. A. B. Kehoe, D. O. Scanlon and G. W. Watson, Chem. Mater., 2011, 23, 4464.
51. F. Y. Wang, S. Chen, Q. Wang, S. Yu and S. Cheng, Catal. Today, 2004, 97, 189.
52. V. A. Bondzie, S. C. Parker and C. T. Campbell, J. Vac. Sci. Technol., A, 1999, 17, 1717.
53. J. Guzman, S. Carrettin and A. Corma, J. Am. Chem. Soc., 2005, 127, 3286.
54. M. Date, M. Okumura, S. Tsubota and M. Haruta, Angew. Chem., Int. Ed., 2004, 43, 2129.
55. B. Chowdhury, J. J. Bravo-Suárez, N. Mimura, S. Tsubota and M. Haruta, J. Phys. Chem. B, 2006, 110, 22995.
56. J. I. D. Cosimo, C. R. Apesteguia, M. J. L. Gines and E. Iglesia, J. Catal., 2000, 190, 261.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c2cy20523h

---