Influence of the active phase loading in carbon supported molybdenum–cobalt catalysts for hydrodeoxygenation reactions

https://doi.org/10.1016/S1387-1811(02)00492-4Get rights and content

Abstract

The influence of the active phase loading on the physico-chemical properties and on the hydrodeoxygenation (HDO) activity of carbon supported catalysts was studied. Four molybdenum catalysts (Mo), with a MoO3 loading ranging from 6 to 15 wt.% and four cobalt promoted molybdenum catalysts (CoMo), with a total oxide content (MoO3+CoO) ranging from 7.2 to 18 wt.% (with constant Co/(Co+Mo) ratio), were prepared and characterized. Both series of samples exhibit a non-uniform distribution of the active phase between the inside and outside of the carbon particles. The dispersion decreases with increasing metal loading. Cobalt seems to be mainly impregnated on the external particle surface and to be responsible for a partial remobilization of molybdenum which migrates towards the external part of the particles where cobalt is impregnated. Bulk cobalt oxide or bulk molybdenum oxide were never detected. For MoO3 contents higher than 6 wt.%, micropore blocking takes place. For a MoO3 content lower than 6 wt.%, most of the molybdenum is located in the microporous structure. It is fairly well dispersed and relatively strongly bound to the support. The CoMo catalysts were tested in HDO reactions. Due to the decrease of the dispersion, the catalytic activity does not increase proportionally with the amount of active phase. The most important parameter for the reaction of ketonic and ester groups is the dispersion of the active phase.

Introduction

Carbon supported sulfided molybdenum-based catalysts seem attractive for HDS and hydrotreating, especially for slurry processes. The absence of acidity of the support makes them potentially useful for special hydrotreating processes [1], [2], especially in the case of oils obtained by pyrolysis of biomass.

Reports indicated that such catalysts were as active, and even more active, than similar ones supported on alumina and silica [3]. In particular, smaller amounts of molybdenum gave better catalysts with activated carbons than with silica or alumina supports. But the performances of carbon-, silica- and alumina-supported catalysts tended to the same limit when the amount of molybdenum was increased. These results are explained by the authors in terms of interactions between active phase and support, dispersion of molybdenum and an easy sulfidation process during activation. The high surface area (SA) of activated carbons allows to obtain a well dispersed and completely sulfided active phase at low metal loading, in spite of the weakness of the interactions between molybdenum and the carrier; sintering and agglomeration are favored with increasing metal content. With alumina, strong interactions occur with molybdenum oxide. The transformation of the oxide species to the sulfide phase is difficult at low metal loading, the fraction of sulfidable molybdenum increases at higher loading, and a good dispersion is maintained.

In a previous work [4] we have shown that molybdenum is not homogeneously distributed between the inside and outside of the carbon particles in catalysts supported on activated carbons prepared with 15 wt.% of MoO3 and 3 wt.% of CoO. Moreover, like on silica supports, the addition of cobalt seems to be responsible for a partial remobilization of molybdenum. This leads to an accumulation of molybdenum at the external surface of the carbon particles. But bulk MoO3 and CoOx oxides were never detected.

In spite of their better performances compared to silica- or alumina-supported catalysts at low active phase loading, those supported on carbon are still not sufficiently active for applications. To take full advantage of the positive characteristics of activated carbon, like weak acidity and low coking tendency, it is crucial to obtain high activity per metal atom, also at high active phase loading. To explore the potential of activated carbon as catalyst support, we have studied a series of four molybdenum catalysts (Mo) with a MoO3 loading ranging from 6 to 15 wt.% and four cobalt-promoted molybdenum catalysts (CoMo), with a total oxide content (MoO3+CoO) ranging from 7.2 to 18 wt.% (with constant Co/(Co+Mo) ratio). The textural and structural properties were characterized. A special objective was to better understand the interactions existing between the two metals when impregnated in different amounts and how these interactions affected the dispersion and the distribution. After sulfidation, the bimetallic catalysts were tested in hydrodeoxygenation (HDO) reactions. The context was the upgrading of the liquids produced from the pyrolysis of biomass [5].

Section snippets

Catalyst preparation

The support was a commercial activated carbon, BKK-100 (particle size diameter between 0.3 and 0.5 mm, specific SA 1200 m2 g−1, total pore volume (Tot PV) 0.9 cm3 g−1). This carbon is obtained by thermal treatment of bituminous coal. This explains the presence of inorganic impurities, mainly calcium, iron and magnesium (0.5, 0.8 and 0.5 wt.%, respectively). The catalysts were prepared following the incipient wetness method. Prior to impregnation, the activated carbon was dried under argon flow at

Textural characteristics

The results of the textural characterization of the activated carbon used as support, Mox and CoyMox samples are reported in Table 2. The same results are presented in Fig. 2. The SA (right-hand scale), the Tot PV, the micro PV (calculated with the Dubinin–Astakov method) and the difference between Tot PV and micro PV, namely meso PV (left-hand scale) have been plotted as a function of the amount of molybdenum oxide.

The textural characteristics of the mono- and bimetallic samples do not

Physico-chemical characterization

Considering first the case of molybdenum, its introduction in the support causes a decrease of porosity higher than expected. As there is no detectable change in the pore diameter, this effect must be explained by pore blocking. The dispersion, as measured by XPS, remains very good. The amount of molybdenum involved in pore blocking is therefore small, compared to that forming a molecular thickness layer. It is presumably limited to the formation of a thin “top” in the narrower mouths.

There is

Conclusions

XPS characterization has shown a non-uniform distribution of the active elements distributed between the inside and outside of the carbon particles. The dispersion decreases with increasing metal loading. Cobalt seems to be mainly impregnated on the external particle surface and to be responsible for a partial dissolution and migration of molybdenum to the external parts. The formation of bulk cobalt or molybdenum oxides was never detected. For a MoO3 content higher than 6 wt.%, micropore

Acknowledgements

The financial support of UE (project JOR3-CT95-0025) is gratefully acknowledged.

References (29)

  • F. Rodriguez-Reinoso

    Carbon

    (1998)
  • A. Centeno et al.

    J. Catal.

    (1995)
  • J.C. Duchet et al.

    J. Catal.

    (1983)
  • M. Ferrari et al.

    Carbon

    (2002)
  • F.P. Daly et al.

    Appl. Catal.

    (1984)
  • L.M. Gandia et al.

    J. Catal.

    (1994)
  • D.R. Penn

    J. Electron Spectrosc. Relat. Phenom.

    (1976)
  • J.H. Scofield

    J. Electron Spectrosc. Relat. Phenom.

    (1976)
  • E. Laurent et al.

    Appl. Catal. A

    (1994)
  • E. Furimsky

    Appl. Catal. A: Gen.

    (2000)
  • B.S. Gevert et al.

    Appl. Catal.

    (1987)
  • J. Sonnemans et al.

    J. Catal.

    (1973)
  • M. Ferrari et al.

    J. Catal.

    (2001)
  • P. Gajardo et al.

    J. Catal.

    (1980)
  • Cited by (41)

    • Selective breaking of C−O bonds in hydrodeoxygenation of 4-methylphenol over CoMoS/ZrO<inf>2</inf>

      2021, Ranliao Huaxue Xuebao/Journal of Fuel Chemistry and Technology
    View all citing articles on Scopus
    View full text