Gaining insight into the kinetics of partial oxidation of light hydrocarbons on Rh, through a multiscale methodology based on advanced experimental and modeling techniques

A. Beretta, A. Donazzi, G. Groppi, M. Maestri, E. Tronconi and P. Forzatti

DOI: 10.1039/9781849737203-00001

This chapter updates previous reviews on the catalytic partial oxidation of methane and light hydrocarbons over noble metals; specifically, it focuses on the development of experimental and modeling tools that in recent years allowed to measure with accuracy and formalize the kinetics of the surface process, thus setting the basis for the engineering of short contact time CPO reformers. Such advanced tools include special micro-reactor designs for the kinetic investigation under isothermal conditions, first-principle microkinetic schemes, techniques for the spatially resolved measurement of temperature and concentration profiles inside working adiabatic reactors, detailed reactor models accounting for the role of transport phenomena in structured catalysts as well as that of homogeneous reactions. These contributions pave a multi-scale path, that runs from the fundamentals of surface kinetics to the reactor optimal design.

1 Introduction

Small scale reformer designs are currently being developed for the distributed and small scale production of H2. Future and futuristic applications are: the fuelling of H2-vehicles, the use of cogeneration systems based on fuel cells for residential heat and power supply, the valorization of CH4-rich bio-gas, the on-board generation and utilization of H2-rich streams as co-fuels, reducing agents and/or feed of Auxiliary Power Unit systems based on SOFC.

Fuel reformers produce hydrogen or hydrogen/COX mixtures by exploiting the catalytic conversion of gaseous or liquid fuels in the presence of steam, CO2 and/or O2 or air. Among various processes and reactor solutions that have been proposed (externally heated reformers, exo-endo reformers, autothermal reformers), the catalytic partial oxidation (CPO) of hydrocarbon fuels in insulated short contact time reactors has several advantages: flexibility to scale up and down, fast light off, resistance to extinction after load changes, high H2 yields at millisecond contact times, and the possibility of fully autothermal operation after light off. This is important in systems of a smaller scale than conventional hydrogen or syngas plants, particularly if restrictions on weight and/or volume are tight.

The CPO process consists of the selective conversion of a hydrocarbon fuel and O2 into CO and H2 using metals (Ni, Pd, Pt, Rh but also Fe, Co and other transition metals), typically deposited in the form of a catalytic wash-coat over high-void fraction structures such as foams or honeycomb monoliths. Lanny Schmidt and his group at Minnesota University have pioneered the process and during the last twenty years largely demonstrated its potential over a wide range of fuels, from gaseous to liquid to solid ones.

In recent years the development of novel experimental techniques and theoretical methods based on DFT has made possible the advancement of fundamental studies of CH activation on metals, which have led to the development of reliable kinetic schemes that can be used to interpret the complex behavior of short contact time-CPO reformers and address their engineering. The surface kinetics of the partial oxidation of light hydrocarbons have been investigated to a lesser extent, but recent studies have shown important analogies between the paths of activation of hydrocarbon fuels of different nature.

The aim of this chapter is to exemplify such recent advances, by reviewing findings from our and other laboratories that, along a multi-scale methodology, run from the surface kinetics to the reactor optimal design. Other previous review papers, in particular the recent paper by Enger, Rune and Holmen, have given a broad view of mechanicistic findings (based also on the development of in-situ or operando characterization techniques) for various catalyst systems. In this work, the emphasis is on the development of experimental and modeling tools that have allowed to recognize and formalize the kinetics of the surface reaction, and also to identify the roles of surface chemistry, gas-phase chemistry, heat and mass transfer phenomena in short contact time CPO reformers.

Kinetic studies on Rh are mainly treated in this chapter, the reason being that experimental, theoretical and modeling studies have presently reached a maturity and a comprehension of the process drivers that is still missing in the case of other metals. However, investigations on other catalysts, mostly Pt, will be also commented because of their important methodological contributions. First (Section III), we illustrate the kinetic findings from isothermal reactor designs, wherein the surface chemistry was isolated and kinetically controlled data were collected on CH4, C2H6 and C3 H8 activation in the presence of O2 and steam. Then (Section IV) we present the microkinetic modeling of CH4 partial oxidation on Rh based on a priori estimations of the kinetic parameters; the model was extremely useful in identifying the prevalent reaction paths behind the observed kinetic dependences. Concerning then the more complex and representative scale of adiabatic, short contact time CPO-reformers, a state of the art technique for the experimental investigation is represented by the axially resolved sampling technique; this will be also presented (Section IV). It is an extremely powerful technique that provides detailed information on the axial evolution of the process stoichiometry and of the corresponding temperature evolution. However, a thorough interpretation of the data needs the means of reactor modeling, which in turn needs the best possible description of the transport phenomena characteristic of the specific reactor configuration (Section V).

Finally, the chapter presents (Section VI) some case-studies wherein the spatially resolved measurements of temperature and composition, in combination with predictive modeling, have driven important advances in the comprehension of the CPO process and set the basis for the rational design of improved CPO-reformers. We conclude by presenting recent results on the optimal design of CPO-reformers, aimed at improving the thermal management of the reactor and preserving the catalyst stability (Section VII).

The ending paragraph attempts then to identify issues that still remain open, including some controversial findings between mechanicistic studies (elsewhere treated) and kinetic studies.

2 The severity of the CPO process

Methane partial oxidation, with stoichiometry:

[FORMULA OMITTED] (1)

is a mildly exothermic reaction. However, several reactions may be catalyzed by metals in the presence of CH rich streams, including the total oxidation of CH (2), the steam reforming (3) and CO reforming (4) of CH4, the water gas shift reaction (5).

[FORMULA OMITTED] (2)

[FORMULA OMITTED] (3)

[FORMULA OMITTED] (4)

[FORMULA OMITTED] (5)

The presence of a direct route of partial oxidation has not been excluded by some authors in the literature. But, in practice, the catalytic conversion of CH4 in the presence of O2 or air involves simultaneously both strongly exothermic and endothermic reactions; sharp gradients of temperature and concentration will tend to develop within the catalyst volume, when the process is run under adiabatic conditions in insulated short contact time reactors.

Thus, although adiabatic short contact time reactors are extremely flexible and useful for demonstrating the process or testing a new catalytic formulation under representative conditions at the laboratory scale, they are not suitable tools for a rigorous kinetic investigation. The quantitative analysis of the observed performances needs the means of mathematical modeling, but the description of the flow pattern and the quantification of transfer phenomena, if we consider for instance the case of foams, may not be consolidated. It is also useful to consider that since the final temperature profile which establishes within an insulated structured catalyst is a result of the operating conditions, it inevitably varies at varying conditions; thus, in an autothermal reactor parameters cannot be varied one at a time as in the traditional kinetic studies.

The understanding of the surface kinetics calls for the development of dedicated lab-scale reactors, wherein the kinetic control of the heterogeneous reaction dominates over other chemical and physical phenomena and the catalyst temperature is uniform in the whole bed. Different approaches have been adopted in the literature and witness an effort of either eliminating any artefact, or to contain them within a measurable form.

3 Experimental tools for the kinetic investigation of the CPO of CH4 and light hydrocarbons under isothermal conditions: reactor designs

We first discuss two alternative designs, recently applied in the literature for studying the kinetics of CH4 activation on Rh and other metals: the differential ultra-diluted packed-bed reactor adopted by Iglesia and co-workers to eliminate any transport artifact, and the annular reactor proposed by our group to obtain integral data under reduced and well denned impact of mass transfer. Other reactor designs are also mentioned.

3.1 The gradient-less differential micro-reactor

The ideal reactor for the kinetic investigation is the one where the temperature and the concentration of species (reactants and products) at the active site have the same values as in the gas phase and any transport limitation is ruled out; along this principle, in recent years Iglesia's group has developed an "infinitely" diluted micro-reactor and applied it for the fundamental study of CH4 activation with O2, steam and CO2 over Pt, Rh and other metals.

Two levels of dilution are involved: intraparticle dilution and bed dilution. Both levels are pushed up to a point where turnover rates do not depend any longer on them. Preferred values of intra-particle dilution on Pt were found to be higher than 200, while optimal values of bed dilution were found to be higher than 2000.

The intra-particle dilution was obtained through a laborious process: a certain, measurable amount of catalyst was first mixed with a comparable amount of inert powders, then the mixed powders were pelletized, crushed, sieved to a suitable mesh size, and then mixed again with another definite amount of inert powders. This procedure needs to be repeated many, many times before achieving the desired diluent to catalyst ratio (at least 200, as mentioned above). The diluted catalyst particles were then mixed with particles of inert material in order to realize the desired bed dilution.

The preparation procedure is such that the final catalyst density referred to the reactor volume is extremely low, only chemical events become kinetically relevant and the reactor behaves as a gradient-less reactor. However, because of the low amount of catalyst which at the end of the procedure is present in the reactor, contamination becomes a major issue: any part of the reactor (quartz wool, thermocouple sheath, the internal walls) must be carefully cleaned and protected with specific coatings. Coherently, attention must be paid when following the preparation procedures. Finally, from a fundamental viewpoint, this high metal dilution can possibly lead to loss of activity due to metal/support effects (e.g. formations of alloys or solid solutions) as pointed out by Jones et al.

Concerning the field of application of this reactor design, it must be considered that the influence of bypass in the reacting flow can be neglected only by operating under differential regime. This is in fact the regime within which the kinetic investigations on CH4 activation with H2O and CO2 were studied by Wei and Iglesia over various metals. Similarly, differential regime data were collected to study the dependence of oxygen coverage on the kinetics of CH4 oxidation on Pt, Rh and Pd surfaces and ethane oxidation on Pt surfaces by Chin, Iglesia et al.

It was demonstrated that over Ni, Pt and Rh the turnover frequency of the forward reforming of CH4 with steam or CO2 (later identified as Regime 4 of CH4 activation) is independent from the nature and the amount of the co-reactant, while it is proportional to the partial pressure of CH4, in strict analogy with methane pyrolysis:

rH2O-ref = rCO2-ref = rpyr = k PCH4 (6)

In more recent studies, the infinitely diluted fixed bed reactor has been successfully applied to investigate the kinetics of CH4 oxidation and C2H6 oxidation on Pt surfaces; four distinct kinetic regimes have been identified at varying oxygen coverage. It has been shown that over oxygen saturated Pt surfaces (Regime 1) CH4 oxidation is kinetically limited only by CH4 activation and the turnover rate is proportional to CH4 partial pressure and independent from O2 partial pressure. At decreasing O2 /CH4 feed ratio, a dramatic enhancement of the turnover rate is observed (up to two orders of magnitude higher than in Regime 1) and an inverse dependence from Opartial pressure together with a quadratic dependence from the partial pressure of the fuel come into place (Regime 2). The experimental evidence on the positive effect of decreasing surface ooxygen coverag is supported by DFT calculations, according to which CH4 adsorption is strongly favored on O*-* site pairs, because of a concerted stabilization of (Pt-CH3---H-O *) transition state. At further decreasing the O2 surface coverage, O2 activation becomes the limiting reaction step and the turn over rate becomes proportional to Opartial pressure (Regime 3). The limiting case of O2 surface depletion is than represented by pyrolysis and reforming reactions (identified with Regime 4 as above mentioned). Regimes, turn over rates and kinetic dependences for methane and ethane oxidation on Pt are reported in Fig. 1.

3.2 The annular micro-reactor

The concept of annular reactor was originally introduced by McCarthy to study the catalytic combustion of methane; it was later applied and further improved in our laboratory to study the short contact time oxidative dehydrogenation of light alkenes over Pt as well as the kinetics of CH4 combustion over Pd.

Figure 2 reports a schematic representation. The reactor consists of a catalyst coated ceramic tube, coaxially inserted into a quartz tube giving rise to an annular duct through which the gas stream flows in laminar regime.

The main geometrical parameters of the reactor are listed in Table 1.

The reactor design has been optimized, with the main aim of performing experiments at very high space velocities without appreciable pressure drop, far from thermodynamic constraints, under a limited or even negligible effect of external and internal mass transfer limitations; these, given the simple geometry of the catalyst layer and the known flow field of the reactor, can be easily quantified by the use of well assessed literature correlations. The thermal behavior of the annular reactor has been quantitatively analyzed within a EUROKIN project, which has shown that the mechanism of heat dissipation consists of back-conduction of the heat of reaction from the catalyst layer through the alumina tube and radiation from the entire surface of the tube (the coated portion and the surface upstream) towards the oven wall. Such a mechanism is so efficient that quasi isothermal conditions are easily obtained when the reactor is operated with diluted feed streams. Instead, under concentrated feed streams axial temperature gradients develop, but their extent is largely moderated by the radiation towards the inner furnace surface. Also, it has been shown that the sliding thermocouple inside the ceramic tube provides a reliable estimate of the actual catalyst temperature.

The design specifically represented in Fig. 1 has been extensively applied to investigate the process of CH4 activation on Rh with O2 and/or CO2 and/or H2O. The wide investigated experimental field is summarized in Table 2. Tests with CH4/O2 mixtures (the CPO tests) were performed from 300 to 850 °C; the operation at high space velocity (varied from 8 x 105 to 4.5 x 106) results in a decreased sensitivity of the conversion curves on temperature, so that it enlarges significantly the T-window for the kinetic investigation.

Experiments were purposely performed both in differential and in integral regime, up to complete conversion of the reactants. As more clearly shown in the following, the production of synthesis gas on Rh occurs only consecutively to complete O2 conversion. Thus, pieces of information on the kinetics of H2 and CO production can be obtained only at high conversion of reactants.

A model of the annular reactor, accounting for gas-solid concentration gradients and intra-porous mass transfer limitations, was necessary to derive the kinetic dependences from the data.