(343e) Kinetics of Catalyst Deactivation during Hydrogen and Ncm Production by Ccvd

The catalytic decomposition of hydrocarbons (e.g. methane) over transition metals (Ni, Fe and Co) catalysts, also called Catalytic Chemical Vapour Deposition, CCVD, has recently been receiving attention as an alternative route to the production of hydrogen and nanocarbonaceous materials (NCM) from natural gas [1-4]. Hydrogen is predicted to become a major source of energy in the future [5]. Hydrogen is a clean fuel that emits no CO2 when burned or used in H2-O2 fuel cells, can be stored as a liquid or gas, is distributed via pipelines, and has been described as a long-term replacement for natural gas [5]. Therefore, a growing demand is forecast in all sectors, including petroleum refining where the increasing need to process heavy and high-sulphur content crudes is accompanied with the lowering of hydrogen co-product in the catalytic reforming process. Steam reforming of methane and other hydrocarbon feedstocks has been the most widely used and usually the most economical technology for the production of hydrogen [5,6]. However, this route makes hydrogen an indirect source of CO2. In addition, the co-product of steam reforming, CO, must be removed by two subsequent steps: water-gas shift and methanation. The complete removal of CO is not economical and therefore, the hydrogen thus produced is not suitable for low-temperature fuel cells given that the catalyst is poisoned by CO [1,7] One of the advantages of catalytic decomposition of hydrocarbons it that is avoided direct formation of CO2 and, therefore subsequent steps for CO removal are not needed. In addition to Hydrogen production, CCVD produces nanocarbonaceous materials, namely carbon nanotubes, CNT's, and carbon nanofibres, CNF's. Infact, CNF's have been known for a long time as a nuisance that often appears during catalytic conversion of carbon containing gases. The recent outburst of interest in the NCM originates from their potential for unique applications as well as their chemical similarity to fullerenes and carbon nanotubes [8]. These materials have potential utilisation as gas (e.g. hydrogen) storage, polymer additives, and as catalyst supports. The accumulation of carbon in form of CNT's and CNF's allows the catalyst to maintain its activity in some cases for an extended period of time. However, catalyst deactivation usually occurs through the formation of encapsulating carbon on the nickel particles. The mechanism of carbon filament formation resulting from the decomposition of hydrocarbons on catalyst metal particles has been extensively studied in the past but few kinetic studies, including all the stages, have been reported In the present paper we report the results of characterisation and catalytic behaviour of coprecipitated Ni-Mg-Al and Co-Mg-Al catalysts during the reaction of methane and acetylene decomposition. A complete kinetic study has been made of the main operating variables, (temperature and gas composition) of these reactions. It is worth noting that most kinetic studies presented in the literature only consider the period of constant carbon formation rate. However, our experiments indicate that the rate of carbon formation is not constant and follows a quite complex pattern. For this reason, the evolution of the carbon formation rate over time, including catalyst deactivation, has been measured during the complete duration of each experiment. The influence has been studied of the operating temperature and feed composition on the carbon formation rate, methane conversion as well as hydrogen production.

Kinetic model of carbon growth. The kinetic model developed was developed to have the dependence of carbon content, and hydrogen production, with time of reaction and operating conditions. This model takes into account all stages of carbon formation, nucleation and filament growth and catalyst deactivation. The first step of the mechanism is the decomposition of methane on the metallic surface of the catalyst. As a consequence of this the carbon atoms leaved on the surface react with the metal forming a surface carbide. This carbide is unstable at operating conditions regenerating the metallic phase and introducing the carbon atoms inside the metallic particle [9,10]. The kinetics of metallic surface carburization can be described as: dCB/dt=(rD)*(CB0*a-CB) rD represents the rate of methane decomposition. As was shown by Snoeck et at. [10], this term depends on the operating conditions, and this dependence can be deduced from the mechanism of the methane decomposition. CB represents the carbon concentration in the metallic particle surface at the side in contact with gas phase. CB0 is represents the maximum carbon concentration attainable at the surface of metallic particles, in absence of catalyst deactivation. The catalyst activity, a, usually decreases as a consequence of encapsulating coke formation. Given that, at the operating conditions used, part of the encapsulating coke formed can be eliminated in situ by hydrogen present in the reaction, the deactivation rate is expressed in terms of a ?Deactivation Model with Residual Activity (DMRA)?[11]. One example of these DMRA is given by the following expression: -da/dt=kG*(a-aS)_d The term d is de deactivation kinetic order, kG is the global deactivation kinetic function and aS is the residual activity of the catalyst. Like rD, kG and aS are also dependent on the operating conditions during the reaction. This deactivation kinetic model can be changed for a mechanistic model in order to get a more fundamental approach of the reactions involved during deactivation process. After decomposition of the surface carbide, the atoms of carbon diffuse through metallic particles, reaching the particle side in contact with the support. When the concentration of carbon at the support side is higher than the solubility of carbon nanofibres, CF, begins the extrusion of filaments. The diffusion rate, and therefore the filaments production rate, can be expressed as: rC(t)=kC*(CB-CF) The term kC is the effective transport coefficient for carbon. Finally, the amount of carbon produced at a given time is calculated by integration of above equation: mC(t)=Integral(rC(t)*dt) This model was solved for different deactivation kinetics cases (including mechanistically derived equations) and then used to fit the curves of carbon content vs. time obtained working at different operating conditions. In all cases, the model predicts very well the experimental data and the dependence of the parameters model with respect to temperature, hydrogen and hydrocarbon partial pressures was obtained. The application of the model allows us to determinate the intrinsic kinetic parameters appearing in rD and kG. i.e. partial orders with respect hydrogen and hydrocarbon, apparent activation energies and pre-exponential factors. These results can also be used to discriminate what is the most appropriated mechanism involved in both, the main reaction (hydrocarbon decomposition) and the deactivation reaction (encapsulating coke formation).

Acknowledgments. The authors acknowledge financial support from DGI-MCYT, Madrid, Spain (Grants PPQ2001-2479 and CTQ 2004-03973/PPQ. 2005-2008).

References 1. N. Z. Muradov and T. N. Veziroglu; Int. J. Hydrogen Energy, 30 (2005) 225. 2. N. Z. Muradov; Int. J. Hydrogen Energy, 26 (2001) 1165. 3. T. Zhang and M.D. Amiridis; Appl. Catal. A, 167 (1998) 1161. 4. J.I. Villacampa, C. Royo, E. Romeo, J.A. Montoya, P. del Angel and A. Monzón; Appl. Catal. A, 252 (2003) 363?383. 5. J.N. Armor, Appl. Catal. A; 176 (1999) 159. 6. J.M. Abrardo, V. Khurana; Hydrocarbon Proc., 79 (1995) 43. 7. M. Steinberg and H.C. Cheng; Int. J. Hydrogen Energy, 14 (1989) 797. 8. K.P. de Jong and J. W. Geus, Catal. Rev.?Sci and Eng., 42 (2000) 481. 9. I. Alstrup, I., J. Catal. 109 (1988) 241. 10. J.-W. Snoeck, G.F. Froment and M. Fowles, J. Catal., 169, (1997) 240; ibid 169 (1997) 250.10 A. Borgna, E. Romeo and A. Monzón; Chem. Eng. J.; 94(2003)19.