Elsevier

Journal of Catalysis

Volume 269, Issue 1, 1 January 2010, Pages 33-43
Journal of Catalysis

Selective decomposition of formic acid on molybdenum carbide: A new reaction pathway

https://doi.org/10.1016/j.jcat.2009.10.012Get rights and content

Abstract

Selective decomposition of formic acid is important as a prototype to study selective bond cleavage of oxygenates. We demonstrate that carbon-modified Mo(1 1 0), C–Mo(1 1 0), is up to 15 times more selective for the dehydrogenation of formic acid than Mo(1 1 0). Reflection absorption infrared spectroscopy (RAIRS) indicates that carbidic carbon blocks active sites for C–O bond cleavage, decreasing the rate of dehydration. Steady-state reactive molecular beam scattering (RMBS) shows that dehydration is the dominant reaction pathway on clean Mo(1 1 0), while C–Mo(1 1 0) selectively promotes dehydrogenation. Kinetic analysis of RMBS data reveals that formic acid dehydrogenation on Mo(1 1 0) has an activation energy of 34.4 ± 3.3 kJ mol−1 while the C–Mo(1 1 0) surface promotes distinct pathways for dehydrogenation with an activation energy of only 12.8 ± 1.0 kJ mol−1. RAIRS spectra suggest the new pathways include the formation of monodentate formate, and at temperatures of 500 K and greater, direct activation of the C–H bond to form carboxyl, both of which decompose via a CO2δ- intermediate to evolve CO2 and H2.

Graphical abstract

Mo2C is up to 15 times more selective towards formic acid dehydrogenation than Mo. Suppressed C–O bond dissociation leading to the formation of monodentate formate and carboxyl is responsible.

  1. Download : Download high-res image (60KB)
  2. Download : Download full-size image

Introduction

Transition metal carbides (TMCs) have attracted significant interest following seminal research by Boudart and co-workers who reported platinum-like reactivity of tungsten carbide catalysts [1]. Since this discovery numerous research groups have investigated the reactivity of transition metal carbides with particular focus on Group VI TMCs, namely tungsten and molybdenum carbides [2], [3], [4]. Early transition metals have high binding energies for many molecules preventing facile desorption and repeated reaction. However, their catalytic behavior can be improved by tempering the binding energy of adsorbates through the addition of carbon and subsequent carbide formation. Many investigations have shown that TMCs can catalyze a number of reactions (many at rates matching or exceeding the best known Group VIII transition metals) involving hydrogen transfer (hydrogenation, isomerization, hydrodesulfurization) [5]. Molybdenum carbide (Mo2C) in particular has been shown to be active for hydrogenation and dehydrogenation reactions [6], and recent research has demonstrated that Mo2C is highly active for the low temperature water–gas shift reaction, in some cases performing better than commercial alumina supported Cu–Zn catalysts [7], [8], [9], [10], [11]. Further, TMCs have the added benefits of comparatively low cost (with respect to Pt-group metals), high thermal stability, mechanical durability, and greater tolerance to common catalyst poisons [2], [6]. In addition to the immediate economic benefits of using TMCs as replacements for precious metal catalysts, it is necessary to consider the sustainability of relying on such rare materials. The US Geological Survey estimates the relative abundance1 of Pd, Pt, Rh and Ir as 7 × 10−4, 7 × 10−4, 2 × 10−4, and 1 × 10−5 [12]. In sharp contrast Mo and W (two of the metals used for TMC catalysts) have a relative abundance1 of 2 and 1, making them ∼103–105 times more abundant than the Pt-group elements.

Significant demand for platinum and platinum-based alloy catalysts has been generated by the need for oxygen reduction and fuel oxidation catalysts for fuel cells. The ubiquitous proton exchange membrane fuel cell (PEMFC) suffers from well known disadvantages due to the use of hydrogen fuel (storage costs and safety) which has promoted interest in alternative liquid fuel cells. Direct Formic Acid Fuel Cells (DFAFCs) are a safe and convenient alternative for portable applications [13], [14], [15]. Still, DFAFCs utilize Pt and Pd-based catalysts; in fact, the most active reported catalysts are Pd–Pt alloys [15], [16], [17]. Electrochemical investigations have established that CO formed by HCOOH decomposition acts to poison active sites, gradually deceasing the activity of the catalyst [18]. Whereas removal of CO can be achieved by oxidation to CO2 by reaction with adsorbed hydroxyl groups present on the surface, in a manner somewhat analogous to the associative water–gas shift mechanism, this reaction appears to be slow and consequently in some cases is a rate-limiting step [18]. The greater activity of Pd-based electrocatalysts has been attributed to an increased rate for direct conversion of formic acid to CO2 which avoids the formation of CO [15], [16]. Therefore, a potential catalyst for formic acid decomposition should promote direct dehydrogenation of formic acid with high activity for CO removal by desorption or by oxidation with surface hydroxyls. The catalytic behavior of formic acid with Mo2C has not been investigated, although Mo2C and WC have already shown promise as replacements for platinum anodes in PEMFCs [19] and direct methanol fuel cells [20].

In many cases, especially for energy applications, the overall desired reaction is dehydrogenation [HCOOH  CO2 + H2]2 as opposed to dehydration [HCOOH  CO + H2O]. Thus, it is of primary interest to determine the elementary steps of the reaction pathway and how these steps may be influenced or controlled by careful modification of the catalyst. The surface chemistry of formic acid has been reviewed by Columbia and Thiel [21], although a number of investigations have been performed since. Briefly, on many clean and oxygen modified transition metal surfaces formic acid initially reacts to create formate and other surface species by either (a) unimolecular deprotonation [HCOOH  HCOO + H] or (b) bimolecular dehydration [2HCOOH  HCOO + H + HCO + OH]. After this first step, the formate species undergoes one of the following reactions; (c) unimolecular deprotonation [HCOO  CO2 + H], (d) reaction with a second formate molecule through the bimolecular “hot hydrogen” pathway [2HCOO  CO + CO2 + 2H + O], or (e) complete decomposition to atomic surface species. In all cases, the ultimate fate of the products from these reaction mechanisms is strongly influenced by the binding energy of these species to the surface, dictating whether they undergo direct desorption, recombinative desorption, or complete dissociation. In the case of Group VIII transition metals, ultra-high vacuum (UHV) studies show that formic acid adsorption on to the clean metal surfaces followed by temperature programmed desorption reacts both by dehydrogenation and dehydration to yield comparable amounts of CO2 and CO on Ni(1 0 0) [22], Ni(1 1 0) [23], [24], Ni(1 1 1) [25], Ru(0 0 0 1) [26], Pd(1 1 1) [27], [28] and Pt(1 1 0) [29]. On the other hand, only CO2 and H2 evolve from the reaction of formic acid dosed at low temperatures on Pt(1 1 1) [30], [31], while the Pt(1 0 0) surface is reported to be inert to formic acid [32]. Additionally, Dahlberg et al. investigated the steady-state reaction of formic acid on polycrystalline platinum foil by molecular beam and found that the reaction proceeded with ∼85% selectivity towards dehydrogenation [33].

Here, we investigate the reaction of formic acid on molybdenum and molybdenum carbide surfaces using two general approaches. First, formic acid decomposition was studied using reflection absorption infrared spectroscopy (RAIRS) and temperature programmed desorption (TPD) on a clean molybdenum surface, Mo(1 1 0), and a molybdenum carbide model catalyst, C–Mo(1 1 0) to determine the effects of carburization towards selective bond cleavage. Vibrational spectroscopy performed in this study indicates that the addition of carbon suppresses low temperature C–O bond dissociation on C–Mo(1 1 0) in comparison to Mo(1 1 0), creating less CO during the initial stages of formic acid decomposition. Second, formic acid decomposition was investigated utilizing reactive molecular beam scattering (RMBS) on the two model catalysts, Mo(1 1 0) and C–Mo(1 1 0). Analysis of the kinetics of this reaction at near steady-state conditions, in a manner comparable to the investigation of Dahlberg et al. on Pt [33], provides product selectivities and apparent activation energies for the direct dehydrogenation and dehydration reaction pathways for formic acid decomposition. Results from these “single-collision” experiments demonstrate that carburization of the Mo(1 1 0) surface improves selectivity towards direct gas phase dehydrogenation at steady-state by as much as ∼1500%. Kinetic data and vibrational spectra acquired at steady-state suggest that the increased selectivity is due to two individual effects: (1) carbidic carbon decreases the rate of dehydration without modifying the activation energy, which remains nearly constant at ∼16 kJ mol−1, therefore, it seems that carbon atoms block the active sites responsible for dehydration, and (2) formation of a carbidic layer on Mo(1 1 0) decreases the apparent activation energy for dehydrogenation from 34.4 ± 3.3 to 12.8 ± 1.0 kJ mol−1 suggesting a change in the reaction mechanism due to the presence of carbidic carbon. Vibrational spectroscopy indicates that carbon-modified Mo(1 1 0) generates new, distinct surface intermediates which are absent on unmodified Mo(1 1 0) including monodentate formate, anionic carbon dioxide, CO2δ-, and carboxyl, OCOH.

Section snippets

Experimental

Our experiments investigating the reaction of formic acid with molybdenum carbide were conducted employing an ultra-high vacuum (UHV) molecular beam surface scattering apparatus with a base pressure less than 1 × 10−10 Torr which has been previously described in detail [34]. Briefly, the apparatus contains an Auger electron spectrometer (AES, Physical Electronics 10–500), a quadrupole mass spectrometer (QMS, Extrel C-50), a Fourier transform infrared spectrometer (FTIR, Bruker Tensor 27) combined

Infrared spectroscopy of HCOOH and DCOOH

Infrared spectra of HCOOH and DCOOH adsorbed onto the Mo(1 1 0) surface are displayed in Fig. 1a and b while spectra on the C–Mo(1 1 0) surface are displayed in Fig. 2a and b. At a sample temperature of 77 K formic acid remains intact on both surfaces (i.e., Mo(1 1 0) and C–Mo(1 1 0)) and the vibrational spectra closely resemble each other and published spectra of crystalline HCOOH and HCOOH multilayers on Mo(1 1 0) [41], [42]. The distinct absorption peaks corresponding to fundamental vibrational modes

Conclusions

We have studied the decomposition of formic acid on the Mo(1 1 0) and C–Mo(1 1 0) surfaces with vibrational spectroscopy, temperature programmed desorption, and reactive molecular beam scattering. Carburization of the model catalyst had dramatic effects on the reaction pathway and product selectivity. The introduction of carbidic carbon on the Mo(1 1 0) surface was shown to suppress the production of formyl, while increasing the fraction of bridge-bonded formate. During TPD of formic acid, the C–Mo(1 1

Acknowledgments

We acknowledge the Defense Threat Reduction Agency (CBT070005974), the National Science Foundation (CTS-0553243) and the Welch Foundation (F-1436). This material is based upon work supported by, or in part by, the US Army Research Laboratory and the US Army Research Office under contract/Grant No. W911NF-09-1-0130. DWF acknowledges helpful conversations with Dr. Jinlong Gong.

References (74)

  • J.G. Chen et al.

    J. Mol. Catal. A

    (1998)
  • S.T. Oyama

    Catal. Today

    (1992)
  • E.V. Rebrov et al.

    Catal. Today

    (2007)
  • M. Nagai et al.

    J. Catal.

    (2006)
  • Y.W. Rhee et al.

    J. Power Sources

    (2003)
  • C. Rice et al.

    J. Power Sources

    (2002)
  • C. Rice et al.

    J. Power Sources

    (2003)
  • P. Waszczuk et al.

    Electrochem. Commun.

    (2002)
  • R. Larsen et al.

    J. Power Sources

    (2006)
  • N.M. Markovic et al.

    Surf. Sci. Rep.

    (2002)
  • M.R. Columbia et al.

    J. Electroanal. Chem.

    (1994)
  • J.B. Benziger et al.

    Surf. Sci.

    (1979)
  • J.L. Falconer et al.

    J. Catal.

    (1978)
  • J.L. Davis et al.

    Surf. Sci.

    (1991)
  • F.S. Thomas et al.

    Surf. Sci.

    (2004)
  • N.R. Avery

    Appl. Surf. Sci.

    (1982)
  • S.C. Dahlberg et al.

    J. Catal.

    (1975)
  • B. Frühberger et al.

    Surf. Sci.

    (1995)
  • H.H. Hwu et al.

    J. Catal.

    (2005)
  • J. Wang et al.

    Surf. Sci.

    (1997)
  • J.W. He et al.

    Surf. Sci.

    (1992)
  • Y. Mikawa et al.

    J. Mol. Spectrosc.

    (1967)
  • R.B. Barros et al.

    Surf. Sci.

    (2005)
  • H.H. Hwu et al.

    Surf. Sci.

    (2003)
  • F. Zaera et al.

    Surf. Sci.

    (1986)
  • M. Okada et al.

    Surf. Sci.

    (1997)
  • S. Haq et al.

    Surf. Sci.

    (1995)
  • N.R. Avery

    Appl. Surf. Sci.

    (1983)
  • J.G. Chen et al.

    Chem. Phys. Lett.

    (1991)
  • A.J. Jaworowski et al.

    Surf. Sci.

    (2001)
  • B. Frühberger et al.

    Surf. Sci.

    (1995)
  • J.G. Chen et al.

    Surf. Sci.

    (1996)
  • J.B. Benziger et al.

    J. Catal.

    (1978)
  • J.B. Benziger et al.

    J. Catal.

    (1979)
  • E.I. Ko et al.

    Appl. Surf. Sci.

    (1979)
  • R.J. Madix et al.

    J. Catal.

    (1973)
  • J. McCarty et al.

    J. Catal.

    (1975)
  • Cited by (56)

    • Unravelling the reaction mechanism of gas-phase formic acid decomposition on highly dispersed Mo<inf>2</inf>C nanoparticles supported on graphene flakes

      2020, Applied Catalysis B: Environmental
      Citation Excerpt :

      According to DFT-based calculations conducted over molybdenum carbide surfaces, the activation energy of unimolecular dehydration is actually higher than that of decarboxylation [33]. Moreover, reactive molecular beam studies over a Mo2C catalyst conducted by Flaherty, Berglund, and Mullins indicated a very high initial production rate of CO shortly after introducing the reactive beam to the sample [37]. This signal rapidly decreased but reached a sustained value while the CO2 signal increased continuously during the reaction period.

    • Non-precious molybdenum-based catalyst derived from biomass: CO-free hydrogen production from formic acid at low temperature

      2018, Energy Conversion and Management
      Citation Excerpt :

      In this catalyst, K should serve as the promoter for the HCOOH decomposition. The reaction pathway of formic acid decomposition on the molybdenum carbide without K was investigated in details by Flaherty et al. [23]. Meanwhile, the mechanism of the vapor-phase formic acid decomposition on the catalysts doped with K species was proposed by Jia et al. [50–52].

    View all citing articles on Scopus
    View full text