R. Casanova¹, J. Mendialdua¹, A. Loaiza-Gil², A. Rodríguez¹ and F. Rueda¹.

1. Laboratorio de Física de Superficies,Dpto. de Física, Facultad de Ciencias.Universidad de Los Andes
  2. Laboratorio de Cinética y Catálisis,Dpto. de Química, Facultad de Ciencias.
Universidad de Los Andes, La Hechicera, Mérida 5101, VENEZUELA.
E-mail: rodrigoc@ciens.ula.ve

Abstract.
In this work, we present a surface characterization of iron phyllosilicate catalysts synthesized with the ammonia method performed by using X-ray excited Auger spectra (XAES) together with XPS data. The detailed study of the O KLL region shows that the oxygen atoms, in this catalyst, are bonded differently to silicon or phyllosilicate (Si₂O₅) than in SiO₂. This report shows that exist compounds that the mere consideration of the Auger parameter and their related is not sufficient for the complete identification, then it has been necessary the detailed study of the O KLL region Auger structure.

Keywords: Auger / XPS / catalyst / oxygen / silica

Resumen
En este trabajo presentamos los resultados de la caracterización superficial de un catalizador tipo filosilicato de hierro usando Espectroscopia de rayos X Auger (XAES) junto con datos de Espectroscopia de fotoelectrones de rayos X (XPS). El estudio detallado de la región O KLL muestra que los atomos de oxígeno en este catalizador están enlazados en forma diferente en el filosilicato (Si₂O₅) que en la silice (SiO₂). Este trabajo muestra que existen compuestos en los que la simple consideración del parámetro Auger y otros relacionados no es suficiente para la total identificación, haciendo necesario un estudio detallado de la estructura de la región Auger O KLL.

Palabras Clave: Auger / XPS / catalizador / oxígeno / sílica

1. Introduction

The frequent use of XPS in surface characterization is its proved ability to provide chemical environment information defined from the chemical shifts of the energy levels for a given atom. Unfortunately, the charge effect produced by X- ray irradiation is present in many materials giving rise to uncertainties in separating the shift due to charge effects and that due to the chemical environment. Wagner et al.[1] proposed the possibility of using the Auger parameter, which is independent of the charge referencing operation, for assigning chemical states thus allowing compound identification.
The KVV Auger transitions for the light elements involving valence levels have been used since they are highly sensitive to local environments and their variations can be used to identify chemical states. A detailed study of the oxygen Auger region in several compounds is very suitable to obtain information about the different chemical and/or structural oxygen environments, since the O 1s chemical shift in different oxides and other compounds is water molecule in the gas phase and the U(2p2p) and the U(2s2p) energies. Fuggle and coworkers[5] employed the O KLL Auger spectra in identification of surface chemistry.
In this work, we have used the oxygen KLL Auger region together with XPS data to extract information on the state and characteristics of an iron phyllosilicate catalyst[6] designated in this work SiFe10.

2.Experimental Section

General Remarks: The experimental measurements were carried out in a VSW spectrometer and its specifications have been reported earlier[7]. The spectrometer has attached a UHV sample treatment chamber with Argon ion etching facilities where samples can experience different sample treatments without exposing them to air. Vacuum in the spectrometer was in the high 10^-9 mbar range and its hemispherical analyzer was operated at 44 eV constant pass energy. Non-monochromatic AlKa radiation (1486.6 eV ) was used as the x-ray source with 300 watts constant power. The raw spectra were smoothed by using a Fourier transform routine after background subtraction according to Shirley’s method[8]. The 285.0 eV C 1s binding energy of adventitious carbon was used as internal binding energy reference for the majority of the samples, except for the case when the samples were treated with oxygen at 773 K in which better coherence is achieved if the O 1s at 530.3 eV or the Si 2p at 103.8 eV are used.
The samples studied in this work were prepared in the catalysis laboratory of Los Andes University according to the method described by Barbier and coworkers[ 9]. In short, 1.6 g of Fe(NO₃)₃ .9 H₂O( 100% purity, J.T. Baker) is added to 16 ml of distilled water at room temperature and stirred by a magnetic rod. The iron hydroxide become precipitated by the addition of some drops of ammonia solution (25% NH₃, Riedel de-Haën). A large excess of ammonia solution (16-20 ml) is added to dissolve the former precipitated. After one hour stirring, 2g. of silica aerosil 200 (specific surface area of 200m2/g, Degussa) are added to the solution. The system is dried in an oven at 353 K for 48 hours after other 2 hours stirring. The samples were reduced under hydrogen flow (10 ml/min) by slowly increasing the temperature up to 973 K at a heating rate of 1K/min.
Three samples has been studied in this work, two designated (1BT SiFe10) and (2BT SiFe10) which were not used in catalytic tests, and the other designated (AT SiFe10) which was used in catalytic tests. These samples were treated in the spectrometer chamber. 1BT SiFe10 sample was annealed, using a temperature ramp of 1 ºC/min, at the following temperatures: 134 ºC, 174 ºC, 217 ºC, 334 ºC, 589 ºC, and left at each temperature for 5 hours; then it was annealed in oxygen (7.10-5 mbar) at 500 ºC for 63 hours. These temperatures were chosen according to the information obtained from ATD data. AT SiFe10 sample was treated in H₂ (9.10-5 mbar) at 530 ºC for 55 hours, then it was annealed in oxygen (10^-4 mbar) at 500 ºC for 50 hours. 2BT SiFe10 sample was annealed at 134 ºC (left for 5 hours) then it was treated in H₂ (10^-4 mbar for 55 hours at 530 ªC.). Auger and XPS spectra of interest were taken for each sample once it was cooled down to ambient temperature after each thermal treatment.

3.Results and Discussion

In table 1 are listed the values for the modified Auger parameters a’ and b’, the energy difference between the KL₁L₂₃ and KL₂₃L₂₃ Auger transitions, the energy of the final state ion[3], the Auger parameter shifts with respect to H₂O(g), and the hole-hole correlation energies[4] for the KL₂₃L₂₃ and KL₁L₂₃ transitions.

Table 1. The values for the modified Auger parameters a‘ and b‘, the energy difference between the KL₁L₂₃ and KL₂₃ L₂₃ Auger transitions, the energy of the final state ion, the Auger parameter shift with respect to H₂O (g), and the hole-hole correlation energies for the KL₂₃L₂₃ and KL₁L₂₃ transitions.

j represents the charge effects
(*) The O1s level has been taken as energy reference

The great similarity that exists between the values in all the parameters reported for SiO₂ samples can be easily verified. BT SiFe10 sample annealed at 174ºC is the only one that differs slightly. Such similarities could indicate that the oxygen chemical environments in these samples are similar as well as the degree of covalence or ionicity in their bonding, indicating that the catalyst structure is like of silica supported iron oxide. However, our recent studies[6,10] reveal that this is not the case but rather a phyllosilicate type structure, Two important facts can be used to reject the Fe₂O₃/SiO₂ structure: one is that the quantitative N_o/N_Si ratio does not correspond to that of silica, and the other, the iron behavior in this catalyst is very different to that exhibited in the ULA catalyst[11] which is of Fe₂O₃/SiO₂ The samples SiFe10, present low iron content and consequently have low oxygen percentage bonded to iron; in addition to the fact that the catalytic active sites in this sample are related to iron, which is type. reflected in the changes experienced by the Fe 2p level during sample treatment, in contrast to the O1s and Si 2p levels which do not show changes; thus, it is not surprising that the oxygen KLL Auger region does not show important changes during sample treatment, and it should also have similarities in comparison to the oxygen Auger region of SiO₂.
In figure 1 the XPS spectra for the Fe 2p level of these samples are presented. In figure 2 the oxygen KLL Auger region are shown. We can clearly distinguish both similarities and differences between SiFe10 and SiO₂ samples.
The KL₁L₁ (¹S_o) transition is better defined in SiO₂ than in SiFe10 and the KL₁L₃ (³P_o) is more pronounced in SiO₂ than in SiFe10. The structure designated by E (see fig.2), assigned by Wagner et al[ 3] to SiO₂ , is more marked in the SiFe10 sample than in SiO₂. The KL₂L₂ transition at  499 eV is also more pronounced in SiFe10 than in SiO₂; this can not be attributed to a contribution coming from the oxygen bonded to iron since none of the important oxygen Auger transitions in Fe₂O₃ appears at this energy; particularly the most intense KL₂L₃ (¹D₂) transition at 512eV. The latter could be contributing on the high energy side of the KL₂L₃ transition of the oxygen bonded to silicon, but even this possibility is not clear, as the structure designated by G (see fig.2 ) is better defined in SiO₂ than in BT SiFe10, and it could be comparable to AT SiFe10 ( see figure2).

Fig. 1. Fe 2p XPS region of SiFe10 sample under the following conditions: a) before catalytic test (BT) and annealed at 174 º C, b) BT and annealed at 217 º C, c) BT and annealed at 334 ºC, d) BT and annealed at 589 ºC, e) BT and treated in oxygen ( 7.10-5 mbar) at 500 º C for 63 hours, f) after catalytic test (AT) in the as received condition, g) AT and treated in H2 (9.10-5mbar) for 55 hours at 530 º C, h) AT and treated in ¹S₀ (10^-4 mbar) for 50 hours, i) BT and treated in hydrogen (10^-4 mbar) for 55 hours at 530 ºC

Fig. 2. O KLL Auger spectra of: a) SiO₂ in the as received condition, b) calcined SiO₂. Fig. 2-c up to fig . 2-k represent the O KLL Auger spectra of sample SiFe10 in the same conditions as fig. 1-a up to fig.1-i. Letters A, B, C, D, on top of figure2-k correspond to KL₁L₁ , KL₁L₂ , KL₁L₃ , KL₂L₂ transitions while E, F, G correspond to KL₂L₃ transitions.

Likewise, it is shown that the oxygen bonded to iron KL₁L₂ transition could contribute to the intensity of the KL₁L₃ transition in SiFe10 samples, and however, that transition (3), in fig. 2-b, is more intense where Fe is absent
The strong KL₂L₂ (¹S₀) transition in SiFe10 samples appears to be typical of these materials. This transition does not change under sample treatment in contrast to the modifications observed in the Fe 2p spectral region, thus this transition should arise from oxygen atoms bonded differently to silicon than in SiO₂.This could be related to the lamellar structure in SiFe10 samples, where one of the oxygen atoms in the tetrahedron that contains a Si atom, is bonded to one silicon atom, while the other three oxygen atoms are bonded to two silicon atoms as in the SiO₂ structure. This could give rise to different bond lengths and a different degree of covalence and hybridization.

4. Conclusion
We can conclude, from our study, that the composition differences in oxygen-silicon compounds are necessarily reflected in significant differences of the values of the Auger parameters and their related.
It can also be concluded that exist cases where a detailed comparative study of the high resolution oxygen KLL Auger spectra helps compound identification specially in those oxygen and silicon compounds in which a consideration of the oxygen Auger parameters, and other related parameters, do not allow a clear identification

Acknowledgments
R.C., J.M., A.R., and F.R. thank CDCHT-ULA for funding some of our research on oxides. The authors wish to thank Dr. B. Fontal from the Department of Chemistry of Universidad de Los Andes-Venezuela for reading the manuscript. We also like to thank the technical assistance of Mr. J. Sarmiento.
A. Loaiza- Gil wish to thank CDCHT-ULA for the project C-942-99- 08-AA and FONACIT-Venezuela for the grant S1-2000000809.

5. References
1. C. D. Wagner, H. A. Six, W.T. Jansen and J. A. Taylor Applications of Surf. Sci., 9 (1981), 203[         [ Links ]STANDARDIZEDENDPARAG]
2. V. I. Nefedov, D. Gati, B. F. Dzhusinskii, N. P. Sergushin, Ya. V.S. Russian Journal of Inorganic Chemistry, 20(9), (1975) 1279[         [ Links ]STANDARDIZEDENDPARAG]
3. C. D. Wagner, D. E. Pasoja, H.F. Hillery, T. G. Kinisky, H.A.Six, W.T. Jansen and J. A. Taylor. J. Vac. Sci. Tecnol.,21(4), (1982) 933[         [ Links ]STANDARDIZEDENDPARAG]
4. G. Moretti J. Electron Spec. and Related Phenom.,59 (1992) 105-118[         [ Links ]STANDARDIZEDENDPARAG]
5. J.C.Fuggle, E. Umbach, R. Kakoscke, D. Menzel J. Electron Spec. and Related Phenom., 26 ( 1982) 111.         [ Links ]
6. A. Loaiza, P. Rodriguez, W. Velazquez D. Gómez, B. Fontal, M. Reyes, T. Suarez, Latin American Journal of Metal. And Mater, , 22(1), (2002) 47.         [ Links ]
7. J. Mendialdua, R. Casanova, F. Rueda, A. Rodríguez, J. Quiñones, L. Alarcón, E. Escalante, P. Hoffman. European J. of Inorg.Chem. Submitted[         [ Links ]STANDARDIZEDENDPARAG]
8. D. A. Shirley, Phys. Rev. B. 5 (1972) 4709[         [ Links ]STANDARDIZEDENDPARAG]
9. A. Babier, A. Hanif, J. A. Delmon, G.A. Martin, Applied Catalysis.,168 (1998) 333.         [ Links ]
10 . A. Loaiza, P. Rodríguez, R. Casanova, J. Mendialdua, A. Rodríguez, F. Rueda. Latin American Journal of Metal and Mater. In press[         [ Links ]STANDARDIZEDENDPARAG]
11. F. Rueda, J. Mendialdua, A. Rodríguez, R. Casanova, Y. Barbaux, L. Gengembre, L. Jalowiecki D. Bouqueniaux, J. Electron Spec. and Related Phenom.,70 (1995) 225        [ Links ]