Characterization of reduced α-alumina-supported nickel catalysts by spectroscopic and chemisorption measurements

https://doi.org/10.1016/j.apcata.2005.04.052Get rights and content

Abstract

Ni/α-Al2O3 catalysts, prepared from nickel acetylacetonate, with low nickel contents (0.4–3 wt.%), have been characterized after a reduction treatment at 400 and 550 °C by infrared spectroscopy of adsorbed CO (IR-CO), X-ray photoelectron spectroscopy (XPS), hydrogen chemisorption measurements, and temperature-programmed reduction (TPR). Relatively high dispersion and particle size smaller than 8 nm were established. It was observed that an increase of the metal loading resulted in an enhancement of the average particle size and a decrease of the dispersion, indicating that suppression of hydrogen adsorption encountered in supported nickel catalysts with low metal contents is not necessarily due to the amount of nickel as reported in the literature, but instead, can be explained in terms of change of the particle size.

Introduction

In the preparation of a supported metal catalyst, a surface compound is frequently formed in the first place. The ease of reduction of the surface compound(s), the particle size distribution and particle shape are affected, among the different treatments, by the nature of the support [1].

Nickel catalysts with less than 5% Ni have been little investigated in spite of their importance to academic studies and industrial applications. Indeed, catalysts with low Ni content allow obtaining an accurate estimation of metal-support interactions which may be used to improve the selectivity in hydrogenation reactions [2]. They also exhibit CO hydrogenation performances that differ from catalysts with high metal content [3], [4].

On the other hand, Ni/Al2O3 catalysts with metal loading lower than 3% are typical cases where the classical techniques for determining particle size and/or dispersion encounter experimental difficulties [5], [7]. Although H2 is normally considered as the best adsorbate for measuring an average crystallite size of nickel, neither this gas nor CO can be used when strong metal-support interactions (SMSI) effects are suspected [5], [6], [8]. For catalysts with less than 3% Ni, the adsorption stoichiometry H/Ni = 1 is not valid, as shown by the very low values in systems where a decrease of the metal dispersion with decreasing metal loading would hardly be expected [5]. X-ray diffraction line broadening or peak profile analysis is limited to crystalline particles which are not too small. In Ni/γ-Al2O3 catalysts, the predominant [1 1 1] reflection of Ni was completely obscured by a broad peak of the support, and the next most intense [2 0 0] reflection of Ni was detected only in samples with more than 15% Ni [6]. Ni/Al2O3 catalysts are not easily analysed by transmission electron microscopy (TEM) owing to insufficient contrast between nickel crystallites and pore structure of the support, particularly for Ni contents lower than 10% [6], [9]. Magnetic measurements inform on the average particle size only when the samples are exempt of remanent magnetism (absence of large nickel particles). If this is not the case, only estimates of the average diameter of the smaller particles and of the weight percentage of crystallites larger than ca. 15 nm may be obtained [8]. In addition, dispersion effects on the magnetic properties induce an error when the nickel precursor is partially reduced [10], and the fit with crystallite size and size distribution does not take into account the possibility of unusual distributions such as a bimodal particle size distribution [7].

Nickel catalysts were among the first ones to be investigated by infrared spectroscopy [11], [12] and studies relating the adsorption of infrared active molecules (e.g. CO) to the surface structure and chemical state of nickel constituted pioneering work in this area [11], [13], not only because nickel catalysts have been considered as the best methanation catalysts, but also because nickel has served as a model metal for theoretical approaches which describe the electronic behaviour and the chemisorption properties [14]. In particular, IR spectroscopy of adsorbed CO provides information on the composition and structure of surface compounds, the nature of the bond(s) between adsorbate and support surface, the existence of various types of surface compounds and active surface centres [15], and finally it also gives access to metal dispersion in supported catalysts [13], [16].

X-ray photoelectron spectroscopy (XPS) offers an interesting alternative approach to evaluate dispersion when usual techniques fail, since it can be applied to any catalytic system without the requirement of crystallinity (as in XRD), minimum particle size (as for TEM) or selective gas adsorption (as in chemisorption) [9]. However, this spectroscopy may be considerably influenced by the repartition of the catalyst components and, therefore, it is often necessary to combine XPS measurements with other characterization methods which allow to solve partially the problems mentioned above [17].

In this study, the main surface species and the principal changes occurring upon reduction have been examined for catalysts prepared with nickel acetylacetonate, Ni(acac)2, and was verified whether the final catalyst presents similar features as those mentioned above. For this purpose, catalysts with metal contents between 0.4 and 3% were reduced at 400 and 550 °C, and characterized by H2 chemisorption, IR-CO, XPS, and TPR techniques. These temperatures were selected on the basis of distinct TPR profiles [18]. Aspects related to the adsorption of Ni(acac)2, the thermal decomposition and catalytic properties in benzene hydrogenation have been previously addressed elsewhere [19], [20], [21], [22], [23].

Section snippets

Experimental

The materials and the preparation conditions of the nickel catalysts have been detailed elsewhere [18], [19], [20]. The support was an α-alumina (with BET specific surface area of 42 m2 g−1, and total pore volume of 0.21 cm3 g−1) obtained by calcination of γ-alumina from Rhone-Poulenc. Small amounts of transition phases (κ and θ) were identified by X-ray diffraction. Both AlVI and AlIV were found by 27Al MAS-NMR spectroscopy, with a AlVI/AlIV ratio of 82:18% [23]. After adsorption of designated

Infrared spectroscopy of adsorbed CO

Fig. 1A and B shows the IR spectra of CO adsorbed on Ni/Al2O3 catalysts reduced at 400 and 550 °C, respectively, for metal loadings between 0.4 and 1.6%. Table 1 indicates the frequencies (in wavenumbers) of the bands appearing in these figures and, in regard, values from the literature and the corresponding assignments [5], [13], [15], [16]. In general, two adsorption states of CO adsorbed on reduced Ni sites are observed in the region 2100–1900 cm−1: linearly bound mono- or multiple carbonyl(s)

IR spectroscopy of adsorbed CO

The CO stretching frequency is an excellent indicator for the way CO is bound to the metal. The precise frequency depends on the nature of the metal, its surface structure and the CO coverage. The type of bonding between CO and the supported metal is influenced, among other parameters, by metal dispersion, type of metal-support interaction, and resulting valency states.

It has been seen that at a reduction temperature of 550 °C, small amounts of Ni0 crystallites are formed, as inferred from the

Conclusion

Surface and dispersion measurements of Ni/Al2O3 catalysts prepared from Ni(acac)2, with less than 3% of metal have been carried out. The dispersion and particle size were determined by IR-CO, XPS, and H2 chemisorption techniques while TPR method was used to obtain information on the species present after a pre-reduction step and on the reduction degree. Qualitatively, the results obtained by IR-CO, XPS, and H2 chemisorption techniques presented similar trends: with increasing reduction

Acknowledgement

R.M. is indebted to Colciencias, Colombia, for financial support.

References (49)

  • C.H. Bartholomew et al.

    J. Catal.

    (1980)
  • D.G. Mustard et al.

    J. Catal.

    (1981)
  • M.V. Ciocco et al.

    Appl. Catal.

    (1988)
  • C. Hoang-Van et al.

    Appl. Catal.

    (1989)
  • G. Marcelin et al.

    J. Catal.

    (1985)
  • J.B. Peri

    J. Catal.

    (1984)
  • R. Molina et al.

    J. Catal.

    (1998)
  • R. Molina et al.

    Stud. Surf. Sci. Catal.

    (2000)
  • R. Molina et al.

    J. Catal.

    (2001)
  • S. Damyanova et al.

    J. Catal.

    (1996)
  • C. Hu et al.

    J. Mol. Catal.

    (1996)
  • C.H. Bartholomew et al.

    J. Catal.

    (1976)
  • S. Narayanan et al.

    Appl. Catal.

    (1995)
  • B. Mile et al.

    J. Catal.

    (1988)
  • J. Mérida-Robles et al.

    J. Catal.

    (1997)
  • A. Aguinaga et al.

    Appl. Catal.

    (1989)
  • P.H. Bolt et al.

    J. Catal.

    (1995)
  • E. Ruckenstein et al.

    J. Catal.

    (1984)
  • L. Kubelková et al.

    Appl. Catal.

    (1993)
  • J.C. Duchet et al.

    Catal. Today

    (1988)
  • S. Hoste et al.

    J. Electron Spectrosc. Relat. Phenom.

    (1979)
  • Y.J. Huang et al.

    Appl. Catal.

    (1988)
  • C.N. Satterfield

    Heterogeneous Catalysis in Industrial Practice

    (1991)
  • Y. Uemura et al.

    Sekiyu Gakkaishi

    (1986)
  • Cited by (0)

    View full text