Elsevier

Journal of Catalysis

Volume 287, March 2012, Pages 124-137
Journal of Catalysis

Understanding the stability of Co-supported catalysts during ethanol reforming as addressed by in situ temperature and spatial resolved XAFS analysis

https://doi.org/10.1016/j.jcat.2011.12.013Get rights and content

Abstract

Co catalysts, based on alumina supports modified with La2O3 and CeO2, have been explored in the steam reforming of ethanol (SRE). The addition of oxygen in the reactants, process called oxy-reforming of ethanol (ORE), and the effect of adding Pt as a promoter were also addressed. One of the main challenges of this system is to hinder carbon deposition that leads to catalyst deactivation. In this work, the stability against carbon deposition was correlated to the control of the Co2+/Co0 ratio, which was addressed by in situ temperature-resolved X-ray absorption near edge spectroscopy (XANES). Temperature- and spatial resolved XANES indicated that the nature of supports, the presence/absence of Pt promoter and the composition of oxidants (water and oxygen) in the feed stream determine the degree of reduction of Co under reaction conditions. The control of the Co2+/Co0 ratio can equilibrate the steps of ethanol activation and carbon oxidation, resulting in stable catalysts.

Graphical abstract

Temperature- and spatial resolved XANES indicated that the nature of supports and the composition of reactants determine the Co2+/Co0 ratio in reforming of ethanol, which control the activity and deposition of carbon.

  1. Download : Download high-res image (147KB)
  2. Download : Download full-size image

Highlights

► Effects of supports on stability of Pt–Co catalysts for ethanol reforming are explored. ► The nature of supports and the presence of Pt determine the degree of reduction of Co. ► The Co2+/Co0 ratio controls the activity and oxidation of the deposited carbon. ► Suppression of carbon deposition is found by changing O2/H2O ratio in the feed.

Introduction

Hydrogen is an important feedstock for fertilizer production and has a crucial role in petroleum refining to reduce impact of fuels on the environment. In the near future, it is expected that hydrogen will become an important energy carrier for use in fuel cells. The production of hydrogen from renewable sources such as ethanol has received special attention due to the environment appeal of this route. Hydrogen can be produced by steam reforming of ethanol (SRE) (Eq. (1)) and steam reforming of ethanol with addition of oxygen in reactants, called oxy-reforming of ethanol (ORE), both reactions catalyzed by metal surfaces.C2H5OH+H2O2CO2+4H2Density functional theory (DFT) studies on decomposition of ethanol over Pt(1 1 1) surfaces suggest that the abstraction of H from ethanol leads to the formation of intermediate OC–CHx and subsequent cleavage of C–C bond yielding CO and CHx species [1], [2]. These calculations are in agreement with experimental data obtained for SRE on Co- and Ni-supported catalysts [3], [4]. The experimental data indicate that the hydrogenation of CHx species to CH4 is favored at low temperatures [3], [4]. The formation of CH4, undesirable for maximizing the H2 production, can be avoided using transition metals with high reactivity to CHx species and low H hindrance, favoring the decomposition of CHx species to C and H2 via a pyrolytic mechanism [5]. On the other hand, this high reactivity of the metal implies in high activity for C–C bond cleavage and consequently C accumulates on metal surface. In this case, the presence of O species is crucial to reestablish the accessibility to these sites by C oxidation. The use of metals with low occupancy of the d orbital can contribute to both strong O interaction and high activity for C–C bond cleavage. However, if not kinetically equilibrated, these properties can result in deactivation due to the blockage of active sites by carbon deposits or due to the oxidation of active sites. Among the metals with low occupancy of the d orbital, Co [4], [6] and Rh [7], [8] have been described as active for reforming of ethanol.

The performance of catalysts is strongly influenced by the nature of support, which adsorbs ethanol and may lead to undesirable by-products. One critical case is the presence of acidic surfaces, which favors the ethanol dehydration forming ethylene and other by-products from the condensation of acetaldehyde [3], [9], [10]. The relative reaction rates on the support and on the metal directly influence the distribution of products. Thus, metals with higher activity are desirable. On the other hand, the increase in activity results in an increase in C on catalyst surface. In this case, a cooperative effect of O transfer from oxides to the metal–support interface will be desirable to promote oxidation of C and obtain active and stable catalysts. Considering that Co (i) has a high activity for the cleavage of C–H and C–C bonds and (ii) enables a higher free energy of adsorbed O species relative to other transition metals [5], these properties may confer potential active sites for obtaining kinetically equilibrated catalysts for reforming of ethanol.

Recently, attention has been attracted to hydrogen production from different reaction, such as SRE and ORE, using Co-based catalysts supported on various oxides, such as Al2O3, MgO, ZnO, SiO2, ZrO2, CeO2 and CeO2–ZrO2 [11], [12], [13], [14], [15], [16], [17], [18], [19], [20]. In spite of reportedly high activities and selectivities, Co-based catalysts still undergo significant deactivation, which is generally attributed to Co particle sintering, carbon deposition and oxidation of the metallic Co particles. Carbon deposition strongly depends on the reaction conditions, such as reaction temperature, water/ethanol molar ratio and oxygen/ethanol molar ratio. Oxygen addition to the feed enhances the gasification rate of the carbon deposits improving the catalyst stability. However, it may lead to oxidation of the metallic Co particles, which can result in activity loss in the reforming reactions [21], [22], [23], [24]. For example, Co/SiO2, Co–Rh/SiO2 and Co–Ru/SiO2 catalysts deactivated under ORE carried out between 623 and 673 K [25]. The decrease in ethanol conversion was accompanied by a decrease in hydrogen selectivity and an increase in acetaldehyde selectivity. The authors proposed that oxygen from the feed oxidized the surface of Co particles and the Co oxide favored the dehydrogenation of ethanol to acetaldehyde. The oxidation of Co particles was inhibited or decreased by the addition of a noble metal (e.g., Rh, Ru). The coexistence of metallic Co and CoOx phases was also observed in CeO2- and CeO2–ZrO2-supported Co catalysts [18], [20], [26], [27], [28], [29]. It is clear that the dependence on the degree of reduction of Co with the composition of reactants in the feed stream and with the type of support should not be neglected, and it has to be balanced with these requirements to decreasing the carbon deposition. It is important to stress that the changes in the oxidation state of Co have not been investigated under reaction conditions by using techniques such as XANES. The characterization of the nature of the active sites under reaction conditions is still a challenge, which is becoming less elusive with time.

In this work, we examine the phase changes of Co in Pt-promoted and unpromoted Co catalysts supported on Al2O3, La2O3/Al2O3, CeO2/Al2O3 and CeO2/MgAl2O4 in SRE and ORE reactions. Special effort was done to perform in situ characterization by XANES as a function of temperature and reaction conditions. The strong spatial dependence of the oxidation state of Co along the reactor was also addressed. The experiments aimed to contribute to the elucidation of issues concerning the Co-catalyzed ethanol reforming reactions such as (i) the sensitivity of stability to the amount of oxygen in the feed stream, (ii) the importance of support nature on catalyst stability and (iii) possible reasons of deactivation.

Section snippets

Catalyst preparation

Commercial γ-Al2O3 (Strem Chemicals, SBET = 200 m2 g−1) or MgAl2O4 synthesized by sol–gel method was used as support. The commercial γ-Al2O3 was calcined at 773 K (10 K min−1) during 3 h under flow of 30 mL min−1 of synthetic air in order to remove moisture and any impurities adsorbed.

Results

The supports Al, La–Al, Ce–Al and Ce–AM show surface areas (SBET) of 78, 74, 77 and 95 m2 g−1, respectively. The surface areas of the catalysts were similar to the ones measured for the supports except for Ce-containing catalysts, that is, Co/Ce–Al, Pt–Co/Ce–Al and Co/Ce–AM, which presented lower surface areas of 66, 65 and 85 m2 g−1, respectively (Table 1).

Fig. 1 presents the XRD patterns obtained for supports and unreduced catalysts. The catalysts show XRD peaks at 2θ equals to 36.8° and 44.8°

Discussion

Alumina-supported Co3O4 precursor is reduced by a two-step process (Co3O4  CoO  Co0). The CoO species formed mainly during the first step of reduction shows high interaction with the alumina surface, resulting in Co/Al catalyst with low degree of reduction by activation at temperatures lower than 973 K. The addition of Ce or La on the Al2O3 surface decreases the interaction of the CoO species with the support. For Co/La–Al sample, a fraction of Co2+ species shows stronger interaction with La–Al

Conclusions

Under SRE at low H2O/ethanol molar ratio (H2O/ethanol = 3) on alumina-supported Co catalysts, the deactivation of catalysts occurs mainly by carbon deposition. The rate of carbon deposition is connected to the density of Co0 sites, characterized by changes of H2 uptake. The supports influence the degrees of reduction of Co and have indirect influence on the control of carbon deposition.

Deposition of carbon decreases with addition of O2 in reactants in ORE. There are not simultaneous rapid

Acknowledgments

The authors gratefully acknowledge financial support from VALE and FAPESP (Fundação para o Amparo a Pesquisa do Estado de São Paulo-Brazil), FINEP (Financiadora de Estudos e Projetos) and the Brazilian Synchrotron Light Laboratory (LNLS), which is acknowledged for the use of its facilities and technical support in TEM (LME) and XAFS (DXAS and XAFS2 beamline) experiments.

References (57)

  • R. Alcalá et al.

    J. Catal.

    (2003)
  • J.W.C. Liberatori et al.

    Appl. Catal. A

    (2007)
  • R.U. Ribeiro et al.

    Appl. Catal. B

    (2009)
  • G. Jones et al.

    J. Catal.

    (2008)
  • H. Song et al.

    J. Catal.

    (2009)
  • A. Birot et al.

    Appl. Catal. B

    (2008)
  • A.M. da Silva et al.

    Appl. Catal. B

    (2011)
  • A. Machocki et al.

    Appl. Surf. Sci.

    (2010)
  • A.N. Fatsikostas et al.

    J. Catal.

    (2004)
  • J. Llorca et al.

    J. Catal.

    (2002)
  • D.R. Sahoo et al.

    Chem. Eng. J.

    (2007)
  • M.S. Batista et al.

    J. Power Sources

    (2004)
  • S.S.Y. Lin et al.

    Appl. Catal. A

    (2009)
  • S.M. de Lima et al.

    J. Catal.

    (2009)
  • S.S.Y. Lin et al.

    J. Catal.

    (2010)
  • F. Frusteri et al.

    Int. J. Hydrogen Energy

    (2006)
  • J. Kugai et al.

    J. Catal.

    (2006)
  • N. Laosiripojana et al.

    Appl. Catal. A

    (2007)
  • A.L.M. da Silva et al.

    Catal. Today

    (2011)
  • E.B. Pereira et al.

    J. Catal.

    (2008)
  • J. Llorca et al.

    Appl. Catal. A

    (2003)
  • V.A. de la Peña O’Shea et al.

    Catal. Today

    (2007)
  • G. Jacobs et al.

    Appl. Catal. A

    (2007)
  • G. Jacobs et al.

    Appl. Catal. A

    (2004)
  • C. Kwak et al.

    Appl. Catal. A

    (2005)
  • E. Baumgarten et al.

    J. Catal.

    (1985)
  • P.J.S. Prieto et al.

    J. Catal.

    (2010)
  • J. Xu et al.

    Surf. Sci.

    (1995)
  • Cited by (47)

    View all citing articles on Scopus
    View full text