Structure and speciation of chromium ions in chromium doped Fe₂O₃ catalysts

Abstract

In this paper, we report a detailed characterisation of chromium doped iron oxide catalysts using a range of techniques to establish the nature of chromium species in the near surface and bulk of iron oxide, high-temperature shift (HTS) catalysts. In particular we have employed X-ray absorption spectroscopy Cr K-edge near edge and extended fine structure data for comparison with chemical and X-ray photoelectron spectroscopy. There was excellent agreement between the techniques in terms of identification and quantification of Cr⁶⁺ and Cr³⁺ species as a function of calcination temperatures between 100 and 500 °C.

---

1. Introduction

Iron and copper based catalysts are widely used for the water–gas shift reactions in which steam and carbon monoxide are reacted to produce hydrogen and carbon dioxide. The operating temperatures for this reaction are between 300 and 450 °C.^1–3 The majority of high-temperature shift catalysts contain small amounts of chromium oxide which acts as a stabiliser to prevent sintering and hence loss of surface area.^4,5 While such benefits are realised using chromium oxide (Cr₂O₃) doping, Cr⁶⁺ can also be detected in addition to Cr³⁺ depending on catalyst preparation conditions. Although a number of studies have addressed the issue of Cr⁶⁺ and how to avoid its formation, the need remains for improved quantification and the ability to distinguish how Cr³⁺ and Cr⁶⁺ are distributed between surface and bulk sites.^6–15

Therefore, it is our aim in this paper to demonstrate the application of a combination of techniques to understand the bulk and near-surface structure of the HTS catalysts after different pre-treatments. Principally, we compare X-ray absorption spectroscopy (XAS) both near edge (XANES) and extended fine structure (EXAFS), with X-ray powder diffraction (XRD), X-ray photoelectron spectroscopy (XPS) and quantitative chemistry.

A quantitative wet chemical method does exist to accurately measure levels of Cr⁶⁺ in a Cr⁶⁺/Cr³⁺ mixture. However, this approach cannot discriminate between surface and bulk material.

X-ray diffraction methods are ideal for characterising the bulk solids, if they are crystalline, but not amorphous or poorly crystalline components.

Furthermore, it is also difficult to distinguish the electronic state of chromium ions in the material using this technique. XAS is an atom-specific technique, widely used in catalytic science to obtain local structure data on crystalline and amorphous materials, but once again this technique provides bulk rather than near-surface information. XPS is the ideal technique to gain near-surface information, in that it can distinguish oxidation states of the metal ions but not structural details and it can provide some quantitative information. Therefore, we have used a combination of all three X-ray techniques to determine local structure and composition of chromium 6+ and 3+ ions in the HTS catalysts as a function of pre-treatment conditions.

2. Experimental part

2.1. Catalyst preparation

The iron/chromium oxide catalyst precursors used in this study were prepared using a co-precipitation method. Briefly, iron and chromium nitrate (at a Fe/Cr molar ratio of 9.8) was dissolved in distilled water and precipitated by addition of aqueous sodium carbonate. The brown precipitate was aged under agitation at 70 °C for 1 h and washed to remove sodium impurities before separating into five batches for calcination. The five batches were subjected to 100, 200, 300, 400, 500 °C respectively with a 5 hour dwell time at target temperature.

2.2. Catalyst characterization

Chemical extraction of the Cr⁶⁺ species was conducted by heating 3 g of the sample with 3 g of NaHCO₃ in 150 ml of distilled water at 150 °C for 1 hour. The Cr⁶⁺ content of the resulting filtrate was determined by iodometric titration.⁷

XRD data of HTS catalysts were collected using a Bruker AXS D8 diffractometer equipped with a 45 position sample changer and operating with Cu Kα radiation (nickel filtered) source. Room temperature scans of all the samples were performed from 10 to 130° 2θ with a step size of 0.02° using the θ/θ coupled scan mode. Data were processed using Bruker AXS Diffrac Plus, EVA V16 software and compared with the ICDD PDF files (PDF2, Release 2003) to identify the phases present in the catalysts.

The XAS spectra at Cr and Fe K-edges were measured at room temperature on the Dutch Belgium beam line (BM26A)¹⁶ at the ESRF using a water cooled Si(111) double crystal monochromator with the synchrotron ring operating in 16 bunch mode giving ∼60 mA electron current at 6 GeV energy. Higher harmonic rejection was performed using vertically focussing mirror that is installed after monochromator.¹⁶

Cr K-edge X-ray absorption spectra of all the samples were measured in the fluorescence mode because of low concentration of chromium species using Cr Kα fluorescence as a function of incident X-ray energy. XAS spectra were recorded employing a 9-element Ge detector. XAS data were acquired in step scans over a k range from 3 Å⁻¹ to 12 Å⁻¹ using k³ weighting for the counting time per data point increasing from 5 s at 3 Å⁻¹ to 10 s at 12 Å⁻¹. Fe K-edge data for HTS catalysts were recorded in transmission mode over a similar k-range. Three scans of each catalyst were recorded and averaged.

Data reduction and the curve fitting analysis of the extended X-ray absorption fine structure (EXAFS) data were performed in Athena¹⁷ and EXCURVE98¹⁸ respectively. Phase shifts and backscattering factors were calculated ab initio using Hedin–Lundqvist potentials. The EXAFS data were analysed using the single scattering wave approximation.¹⁹ Refinement was carried out with k³ weighting in the range 3.3 to 11.8 Å⁻¹. The non-structural parameter AFAC (the amplitude reduction factor due to many electron processes) was extracted from EXAFS data of Cr and Fe foils. During refinement of foils EXAFS the coordination numbers for the first shell were fixed to crystallographic values. The best fit to data yielded AFAC 0.7 and 0.9 for Fe and Cr respectively. It is noteworthy that the refined coordination numbers in EXAFS are generated with 10–20% uncertainties^20,21 and the individual errors obtained from the analysis are given in the tables. In addition, to determine the nature of chromium species and the amount of Cr⁶⁺ in the catalyst, principal component analysis (PCA) was performed on the near edge spectra (XANES) using model compounds, Cr₂O₃ and K₂Cr₂O₇, representing Cr³⁺ and Cr⁶⁺ respectively. Each XANES of the catalysts was fitted to data from the Cr³⁺ and Cr⁶⁺ model compounds using Athena software.¹⁷

XPS measurements were carried out using a Thermo Fischer demonstration Escalab 250 operated with monochromatic aluminium K_α radiation as a 650 μm spot 200 W power. Samples were dusted onto conductive tape mounted on a sample stub. Charge compensation was provided by an argon ion in-lens flood gun at 20 V combined with an in-chamber flood source at ‘zero energy’ settings. Energy scales were corrected to the carbon 1s signal maxima (285 eV). Quantification was carried out by peak area measurements using curve fitting after background subtraction and applying sensitivity factors derived by Schofield.²²

3. Results and discussion

XRD patterns recorded at room temperature (Fig. 1) reveal that the samples calcined at temperatures below 500 °C contains both amorphous and crystalline components, while the sample calcined at 500 °C appears to be highly crystalline. The crystalline component is identified as Fe₂O₃ (haematite) in all samples. There was no evidence of any significant X-ray peak shifts to confirm Fe/Cr/O solid solution formation or stoichiometric differences in any of these samples, as evidenced by the calculated lattice parameter data obtained through Rietveld analysis (TOPAS software). However, Cr³⁺ and Fe³⁺ ionic radii are very similar, both form α-alumina oxides structure and here Cr is very much the minority species. Furthermore, interatomic potential computational modelling has indicated ready solubility of Cr³⁺ in haematite.²³

Fig. 1 XRD patterns of HTS catalysts calcined at various temperatures.

The results of quantitative chemical analysis are shown in Table 1. The data indicate that total Cr⁶⁺ levels are low for material calcined at 100 °C. The level then increases markedly up to 400 °C, when over 50% of the Cr appears to be Cr⁶⁺, and subsequently diminishes after heating to 500 °C. This variation with temperature is discussed alongside data from other techniques below.

Table 1 Relative amounts of Cr⁶⁺ and Cr³⁺ in catalysts calcined at various temperatures. The experimental errors are less than 8%

Calc. temp. (°C)	Bulk analysis			Surface analysis
	Chemical	XANES	EXAFS	XPS
	Cr^6+/Cr^3+ atom ratio				Total Cr/Fe^b
a Too low to measure. b Bulk Cr/Fe formulation 0.11.
100	0.05	^a	^a	^a	0.08
200	0.32	0.39 ± 0.02	0.37 ± 0.01	0.48	0.07
300	0.83	1.16 ± 0.01	1.00 ± 0.01	1.01	0.09
400	1.18	1.48 ± 0.01	1.50 ± 0.01	1.46	0.11
500	0.21	0.22 ± 0.02	0.30 ± 0.01	0.82	0.13

---

The results of XPS analysis, which are derived from the spectra in Fig. 2 by curve fitting, are shown Table 1. The spectra clearly indicate the presence of both Cr³⁺ (binding energy 576.8 eV) and Cr⁶⁺ (binding energy 579.4 eV) although the relative amounts vary strongly with calcination temperature. The ratio of Cr to Fe data in Table 1 suggest that total Cr at the surface slightly increase as a function of calcination temperature. This observation would be in agreement with the electron microscopy analysis by Edwards et al.,²⁴ who measured composition variation from 6.3 ± 2.3 wt% at the centre of a catalyst grain to 10.7 ± 2.3 wt% at the edge. The Cr⁶⁺ concentration matches that measured by bulk chemical analysis with perhaps a slight surface excess up to 400 °C. After calcination at 500 °C the XPS data indicate a strong surface excess of Cr⁶⁺.

Fig. 2 XPS data of the chromium containing iron oxide samples heated at various temperatures. We show here only the Cr 2p spectra and they were normalised and corrected using carbon 1s binding energy.

X-ray absorption spectroscopy have been used extensively to characterise chromium containing catalysts.^11–14 Normalized Cr K-edge XANES spectra recorded in the fluorescence mode for the chromium containing iron oxide catalysts are shown in Fig. 3 along with two model compounds, Cr₂O₃ and K₂Cr₂O₇, representing Cr³⁺ and Cr⁶⁺ oxidation states respectively. The XANES spectra for the two model compounds containing Cr³⁺ and Cr⁶⁺ ions are distinctly different with the Cr₂O₃ data showing a pre-edge doublet (∼5994 eV) with very low intensity compared to K₂Cr₂O₇ which has an intense single pre-edge peak. This latter feature has been reported in several publications and identified as a 1s–3d transition. Although this quadrupole transition is not allowed according to the selection rules, due to mixing of p and d states coupled with local coordination symmetry and density of unoccupied states, the transition becomes allowed. This pre-edge feature has been exploited as a signature identifying tetrahedral coordination geometry in a variety of d^o transition metals states in catalytic materials.^25–29

Fig. 3 Cr K-edge XANES of HTS catalyst and the reference materials containing Cr⁶⁺ and Cr³⁺.

Comparison of the XANES for HTS-100 (dashed line) with that of the Cr₂O₃ model compound suggests that the chromium in the catalyst is similar to Cr₂O₃, in terms of local structure but it is important to remember it is not crystalline (Fig. 1) and Cr is only a minority dopant. The two small pre-edge features at 5993 eV and 5995 eV can be assigned to the 1s–3d(t_2g) and 1s–3d(e_g) electronic transitions respectively.^29–32 The 2 eV separation of the Cr³⁺ pre-edge features agrees with 2–3 eV octahedral crystal field splitting between the t_2g and e_g levels measured for several Cr³⁺compounds in local octahedral coordination.³⁰ However, the pre-edge XANES for the catalysts calcined above 100 °C are distinctly different to Cr₂O₃ in that they show the strong pre-edge feature indicating the presence of Cr⁶⁺ in widely varying amounts as a function of calcination temperature.

Quantification of the fluorescence XANES data was carried out using principal component analysis by taking Cr₂O₃ and K₂Cr₂O₇ as model compounds to obtain the best linear combination fit. A typical fit to the catalyst data (HTS-500) is depicted in Fig. 4. The variation of the ratio of Cr⁶⁺ to Cr³⁺ derived by this method (Table 1) shows a similar pattern to that displayed by the results of bulk chemical analysis, i.e. a minimal amount of Cr⁶⁺ at 100 °C rising to a maximum at 400 °C and then dropping away after 500 °C calcination, when Cr⁶⁺ becomes unstable and reduces to Cr³⁺. This pattern of behaviour differs in one respect from that observed by XPS, which is a surface sensitive technique, in that the latter indicates a considerable excess of Cr⁶⁺ at the surface after calcination at 500 °C compared to the bulk. This suggests that surface Cr³⁺ re-oxides during cooling to room temperature.

Fig. 4 Cr K-edge fluorescence XANES of HTS-500 catalyst (solid line) and the fitted profile (dashed line) obtained by a linear combination least square fit to the Cr₂O₃ and K₂Cr₂O₇ reference materials.

In order to investigate further the scope for XAS methods to derive information on the occurrence of Cr⁶⁺, a second approach has been examined, the quantification of bulk transmission EXAFS data.

Fe K-edge and Cr K-edge EXAFS analysis Fig. 5 and 6 show the best fit between the calculated and experimental Fe K-edge and Cr K-edge EXAFS data respectively. The EXAFS fitting results for both Fe K-edge and Cr K-edge are summarized in Table 2. All the samples, except for the one heated at 500 °C, showed a large amorphous component by XRD in addition to some haematite. In order to determine whether the amorphous component also has similar local environment to that of the crystalline part, we analysed in detail the Fe K-edge EXAFS data of all the samples. The structural parameters obtained from Fe K-edge EXAFS data for HTS catalysts revealed that the structure of HTS-100 catalysts is very close to the haematite structure α-Fe₂O₃.³³ The iron oxygen distances are found to be at R_Fe–O = 1.96 Å and R_Fe–O = 2.09 Å, and the iron–iron distances R_Fe–Fe = 2.89 Å, 3.05 Å, and 3.40 Å. These inter atomic distances are in very good agreement with the structure of haematite.³³ The FeO₆ octahedra in haematite share one face (R_Fe–Fe = 2.89 Å), three edges (R_Fe–Fe = 2.97 Å), three double corners (R_Fe–Fe = 3.37 Å) and six double corners R_Fe–Fe = 3.7 Å. The inter-atomic iron–oxygen distances of, R_Fe–O, and iron–iron, R_Fe–Fe, in HTS-200, -300, -400, -500 are similar to those in the crystal structure of haematite.

Fig. 5 Fe K-edge EXAFS spectra analysed in k-space (after background subtraction, k³-weighted) of HTS catalysts together with corresponding Fourier transforms. The magnitude and imaginary part of Fourier transforms along with best fits are shown. The solid line is experimental data and the dashed line is the best fit.

Fig. 6 Cr K-edge EXAFS spectra analyzed in k-space (after background subtraction, k³-weighted) of HTS catalysts and Fourier transform of corresponding EXAFS spectra. The magnitude and imaginary part of Fourier transforms along with best fits are shown. The solid line is experimental data and the dashed line is the best fit.

Table 2 Fe K-edge and Cr K-edge EXAFS results for the HTS catalysts and chromium containing model compounds, where N is coordination number; when refined N is reported with less than 20% uncertainty, R inter-atomic distance between atom pairs, σ²-Debey Waller factor and R_Fit-goodness of fit. Uncertainties in the last digit shown are given in parentheses

Catalyst	Fe K-edge					Cr K-edge
	Scatt.	N	R (Å)	σ ^2 (Å^2)	R Fit	Scatt.	N	R (Å)	σ ^2 (Å^2)	R Fit
HTS-100	O	3	1.96(1)	0.004(1)	16	O	6	1.98(1)	0.005(1)	25
	O	3	2.09(1)	0.012(3)		Fe(Cr)	1	2.99(2)	0.003(3)
	Fe	1	2.89(1)	0.004(1)		Fe(Cr)	3	3.11(9)	0.012(1)
	Fe	3	3.05(1)	0.007(1)		Fe(Cr)	3	3.43(4)	0.020(1)
	Fe	3	3.40(2)	0.022(6)
HTS-200	O	4	1.95(1)	0.005(1)	19	O	1	1.64(1)	0.003(1)	23
	O	2	2.14(1)	0.005(1)		O	4	1.98(1)	0.005(1)
	Fe	1	2.92(1)	0.005(2)		Fe(Cr)	1	2.97(2)	0.005(2)
	Fe	3	3.06(1)	0.010(3)		Fe(Cr)	3	3.13(2)	0.012(4)
	Fe	3	3.41(1)	0.016(3)		Fe(Cr)	3	3.42(3)	0.015(4)
HTS-300	O	2	1.89(1)	0.004(2)	28	O	2	1.64(1)	0.004(1)	22
	O	3	2.03(1)	0.007(3)		O	3	1.98(1)	0.005(1)
	Fe	1	2.89(2)	0.007(3)		Fe(Cr)	1	2.97(2)	0.010(5)
	Fe	3	3.03(1)	0.011(4)		Fe(Cr)	2	3.14(2)	0.014(7)
	Fe	3	3.38(2)	0.016(5)		Fe(Cr)	3	3.40(2)	0.016(3)
HTS-400	O	4	1.95(1)	0.005(1)	18	O	2	1.64(1)	0.004(1)	20
	O	2	2.15(1)	0.004(1)		O	2	1.98(1)	0.004(1)
	Fe	1	2.90(1)	0.005(2)		Fe(Cr)	1	3.03(1)	0.010(1)
	Fe	3	3.05(1)	0.009(3)		Fe(Cr)	2	3.38(1)	0.014(3)
	Fe	3	3.38(1)	0.017(4)
HTS-500	O	3	1.94(1)	0.004(2)	26	O	1	1.63(1)	0.008(4)	33
	O	3	2.09(1)	0.004(4)		O	5	1.99(1)	0.005(1)
	Fe	1	2.88(7)	0.004(9)		Fe(Cr)	3	2.96(1)	0.008(3)
	Fe	3	2.97(3)	0.004(4)		Fe(Cr)	1	3.12(2)	0.011(9)
	Fe	3	3.37(1)	0.005(1)		Fe(Cr)	3	3.40(2)	0.007(4)
	Fe	3	3.68(1)	0.005(2)		Fe(Cr)	3	3.71(2)	0.010(6)
K2Cr2O7						O	4	1.61(1)	0.008(1)	29
						Cr	1	3.14(1)	0.004(2)
Cr2O3						O	6	1.99(1)	0.005(1)	39
						Cr	1	2.65(2)	0.006(4)
						Cr	3	2.89(1)	0.005(1)
						Cr	3	3.42(1)	0.005(2)
						Cr	6	3.65(1)	0.008(2)

---

Before we analysed the Cr K-edge data of the catalysts, we tested the modelling procedure including calculated phase shift and amplitude factors by analysing in detail the two model compounds, K₂Cr₂O₇ and Cr₂O₃. In K₂Cr₂O₇ the chromium atom is surrounded by four tetrahedrally arranged oxygen atoms at a distance of R_Cr(VI_)–O = 1.61 Å. The chromium–chromium distance was found to be at R_Cr(VI_)–Cr = 3.14 Å. In Cr₂O₃ the chromium atom is surrounded by six octahedrally spaced oxygen atoms at R_Cr(III_)–O = 1.99 Å; the chromium–chromium distances are R_Cr(III_)–Cr = 2.65, 2.89, 3.42, 3.65 Å. The EXAFS fitting results for K₂Cr₂O₇ and Cr₂O₃ are consistent with these reported structural models derived from XRD.

Comparison of the Fourier transforms of HTS-100, -200, -300, and -400 (Fig. 7) clearly show that the magnitude of the Fourier transform peaks corresponding to the Cr–O distance at ∼2 Å decreases with calcination temperature. These changes due to thermal treatment correspond to a partial transformation in the local atom environment from the octahedral Cr³⁺ to tetrahedral Cr⁶⁺ coordination as detected by XANES and supported by chemical and XPS analysis. Furthermore, Fig. 7 confirms that this process is reversed on heating to 500 °C, i.e. the level of Cr⁶⁺ is significantly reduced.

Fig. 7 Fourier transforms of k³ weighted Cr K-edge EXAFS as a function of the atomic distance, R, for HTS catalysts calcined at various temperatures.

While the EXAFS data in Fig. 5 for the Fe K-edge from the four largely amorphous materials, HTS-100, -200, -300 and -400 look broadly similar and obviously different to that for crystalline haematite (HTS-500), the Cr K-edge data look very different from one another. These structural changes in the local environment around Cr reflect the dramatic differences in relative amounts of Cr⁶⁺ and Cr³⁺ detected by other techniques as discussed above. The results of detailed curve fitting for the Cr K-edge EXAFS data are shown in Table 2. It is important to note when viewing Table 2 that it is not possible to distinguish Fe and Cr as backscatters since they have similar X-ray scattering power.

For HTS-100 the nearest neighbour oxygen shell can be fitted with a single octahedral shell with a Cr–O distance of 1.98 Å which is similar to Cr₂O₃ and close to the average value from the Fe K-edge in HTS-100. For the next nearest neighbour cations three shells were fitted following a similar pattern to the Fe K-edge in HTS-100 rather than Cr₂O₃, i.e. cations at 2.99 Å (1Fe), 3.11 Å (3Fe), and 3.43 Å (3Fe). The XANES data discussed above has already confirmed the sole presence of Cr³⁺ and these two results would suggest that Cr³⁺ substitutes for Fe³⁺ in the amorphous matrix in agreement with predictions from computational modelling.²³

The Cr K-edge EXAFS fitting results for HTS-200, -300 and -400 indicate that it is necessary to have two Cr–O distances, one representing a short Cr–O distance of ∼1.64 Å (similar to K₂Cr₂O₇) and the other with Cr–O distance of ∼1.98 Å, similar to Cr₂O₃. The results obtained for the short Cr–O distance is consistent with the Cr⁶⁺–O tetrahedral distance in K₂Cr₂O₇ and the long Cr–O distance agrees with the octahedral distance found in Cr₂O₃. The variation in the relative number of O atoms in the two first shell sites reflects the changes in the relative amount of Cr⁶⁺ and Cr³⁺ as a function of calcination temperature. Indeed, the relative oxygen occupation in the two sites, tetrahedral and octahedral, can provide a quantitative estimate of the amount of the two Cr oxidation states as shown in Table 1. For the second shell of HTS-200 and -300 EXAFS fitting finds three cations atoms, one at a short distance of 2.97 Å and the other two/three at a longer distance of 3.13 Å plus three iron(chromium) atoms at 3.40 Å. This pattern is broadly similar to data for the Fe K-edge. While the next nearest neighbour cation shells are slightly different for HTS-200 and -300, the data for HTS-400 seems very different reflecting the large amount of Cr⁶⁺ compared to Cr³⁺ and the presence of tetrahedral geometry. However, the radii of the fitted cation shells match the average values derived for HTS-200 and -300, i.e. the overall structure is dominated by the Fe–O octahedral matrix.

HTS-500 has been already identified by XRD above as being crystalline haematite. It is not surprising therefore that EXAFS data for the Cr K-edge bears some resemblance to the Fe K-edge. However, a short Cr–O distance (1.63 Å) contribution to the EXAFS can be detected corresponding to the presence of Cr⁶⁺. XPS data discussed above indicates this oxidation state is located primarily at the surface.

4. Conclusions

Analysis of Cr doped Fe₂O₃ catalysts chemically and by XPS, XANES and EXAFS has revealed Cr³⁺ is the dominant species for the largely amorphous material calcined at 100 °C. However upon heating to 400 °C levels of Cr⁶⁺ increase markedly. There is excellent agreement between the four techniques in terms of identification and quantification of the relative levels of the two Cr species. On heating to 500 °C, when the material crystallises to form haematite, the Cr⁶⁺ also becomes unstable and reduces to Cr³⁺. XPS indicates that the residual Cr⁶⁺ after 500 °C calcination is located at the surface and is probably formed from re-oxidation of Cr³⁺ during a cooling to room temperature. However, total Cr/Fe ratio measured by XPS suggests that Cr³⁺ is doped into the haematite lattice. Importantly we have also shown that both principal component analysis of XANES data and modelling EXAFS data differentiating Cr⁶⁺ and Cr³⁺ oxygen distances provide good method of analysing mixed oxidation states.

Acknowledgements

We thank EPSRC and NWO for funding and ESRF for the provision of beam time. GS thanks the Royal Society, UK for an Industry Fellowship. VM thanks to Martin Martis for his help during the beamtime.

References

1. L. Lloyd, D. E. Ridler and M. V. Twigg, Catalyst Handbook, ed. Wolfe Publishing, 2nd edn, 1989.
2. C. Ratnasamy and J. P. Wagner, Catal. Rev.: Sci. Eng., 2009, 51, 325.
3. C. Rhodes, G. J. Hutchings and A. M. Ward, Catal. Today, 1995, 23, 43.
4. G. C. Chinchen, R. H. Logan and M. S. Spencer, Appl. Catal., 1984, 12, 89.
5. R. L. Keiski and T. Salmi, Appl. Catal., A, 1992, 87, 185.
6. F. Domka, A. Basinska and R. Fiedorow, Surf. Technol., 1983, 18, 275.
7. M. A. Edwards, D. M. Whittle, C. Rhodes, A. M. Ward, D. Rohan, M. D. Shannon, G. J. Hutchings and C. J. Kiely, Phys. Chem. Chem. Phys., 2002, 4, 3902.
8. J. Y. Lee, D.-W. Lee, K.-Y. Lee and Y. Wang, Catal. Today, 2009, 146, 260.
9. P. Kappen, J.-D. Grunwaldt, B. S. Hammershøi, L. Tröger and B. S. Clausen, J. Catal., 2001, 198, 56.
10. G. Doppler, A. X. Trautwein, H. M. Ziethen, E. Ambach, R. Lehnert, M. J. Sprague and U. Gonser, Appl. Catal., 1988, 40, 119.
11. E. Groppo, C. Lamberti, S. Bordiga, G. Spoto and A. Zecchina, Chem. Rev., 2005, 105, 115.
12. E. Groppo, C. Prestipino, F. Cesano, F. Bonino, S. Bordiga, C. Lamberti, P. C. Thüne, J. W. Niemantsverdriet and A. Zecchina, J. Catal., 2005, 230, 98.
13. P. C. Thune, R. Linke, W. J. H. van Gennip, A. M. de Jong and J. W. Niemantsverdriet, J. Phys. Chem. B, 2001, 105, 3073.
14. B. M. Weckhuysen, R. A. Schoonheydt, J.-M. Jehng, I. E. Wachs, S. J. Cho, R. Ryoo, S. Kijlstra and E. Poels, J. Chem. Soc., Faraday Trans., 1995, 91, 3245.
15. M. Scariot, M. S. P. Francisco, M. H. Jordão, D. Zanchet, M. A. Logli and V. P. Vicentini, Catal. Today, 2008, 133–135, 174.
16. S. Nikitenko, A. M. Beale, A. M. J. van der Eerden, S. D. M. Jacques, O. Leynaud, M. G. O'Brien, D. Detollenaere, R. Kaptein, B. M. Weckhuysen and W. Bras, J. Synchrotron Radiat., 2008, 15, 632.
17. B. Ravel and M. Newville, J. Synchrotron Radiat., 2005, 12, 537.
18. N. Binsted and S. S. Hasnain, J. Synchrotron Radiat., 1996, 3, 185.
19. S. J. Gurman, N. Binsted and I. Ross, J. Phys. C: Solid State Phys., 1984, 17, 143.
20. K. Teo Boon, EXAFS: Basic Principles and Data Analysis, 1986.
21. G. Bunker, Introduction to XAFS: A Practical Guide to X-ray Absorption Fine Structure Spectroscopy, Cambridge University Press, 2010.
22. J. H. Scofield, J. Electron Spectrosc. Relat. Phenom., 1976, 8, 129.
23. S. Benny, Doctoral thesis (EngD), University College London (UCL), 2010.
24. M. A. Edwards, D. M. Whittle, C. Rhodes, A. M. Ward, D. Rohan, M. D. Shannon, G. J. Hutchings and C. J. Kiely, Phys. Chem. Chem. Phys., 2002, 4, 3902.
25. F. Farges, G. E. Brown and J. J. Rehr, Geochim. Cosmochim. Acta, 1996, 60, 3023.
26. C. Kongmark, V. Martis, A. Rubbens, C. Pirovano, A. Lofberg, G. Sankar, E. Bordes-Richard, R.-N. Vannier and W. Van Beek, Chem. Commun., 2009, 4850.
27. G. Sankar, J. M. Thomas and C. R. A. Catlow, Top. Catal., 2000, 10, 255.
28. F. Farges, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 1.
29. M. L. Peterson, J. G. E. Brown, G. A. Parks and C. L. Stein, Geochim. Cosmochim. Acta, 1997, 61, 3399.
30. R. G. Burns, Mineralogical applications of crystal field theory, Cambridge Univ. Press, 2nd edn, 1993, 7.
31. M. L. Peterson, G. E. Brown and G. A. Parks, MRS Proc., 1996, 432, 75.
32. A. M. Beale, D. Grandjean, J. Kornatowski, P. Glatzel, F. M. F. de Groot and B. M. Weckhuysen, J. Phys. Chem. B, 2006, 110, 716.
33. R. L. Blake, R. E. Hessevick, T. Zoltai and L. W. Finger, Am. Mineral., 1966, 51, 123.

---