Elsevier

Chemical Engineering Science

Volume 55, Issue 18, 15 September 2000, Pages 3929-3940
Chemical Engineering Science

Adsorption-enhanced steam–methane reforming

https://doi.org/10.1016/S0009-2509(99)00597-7Get rights and content

Abstract

Experimental and theoretical studies of steam–methane reforming in the presence of a hydrotalcite-based CO2 adsorbent are presented. Attention is given to the analysis of the transient behaviour of a tubular (integral) reactor when an Ni-based catalyst is admixed with the adsorbent. Considerable enhancement of the methane conversion is experimentally demonstrated. Enhancement arises from the favourable shifts in the reaction equilibria of the reforming and water–gas shift reactions towards further CO2 production. As predicted, the potential for conversion enhancement is shown to increase under the conditions of a high reactor space time, high operating pressure, or a low steam-to-methane feed ratio, i.e. when reaction equilibrium limitations are important. A mathematical model, accounting for mass transfer limited adsorption kinetics, non-linear (Langmuirian) adsorption equilibria and a general reaction kinetic model, is shown to accurately predict the observed elution profiles from the reactor, and thus the degree of conversion enhancement.

Introduction

The advantages of coupling reaction systems with some form of in situ separation have been widely reported in the literature. Such hybrid configurations may substantially improve reactant conversion or product selectivity and, for reversible reactions, establish a more favourable reaction equilibrium than that which could be achieved under conventional reactor operation. Reaction enhancement may enable a lower temperature of operation, which in turn may alleviate the problems associated with catalyst fouling, high process energy requirements and poor energy integration within the plant environment. For gas-phase catalytic reactions, the separation can be based on adsorption, selective permeation through a membrane, or through simultaneous reaction of the targeted molecule (e.g. the reaction inhibitor) with a chemical acceptor.

A comprehensive review on membrane-based reaction systems has been given by Armor (1995). Advances have been made in the use of metallic membranes (often Group VIII metals which only small molecules like hydrogen can permeate) and, more recently, polymeric, ceramic and zeolitic membranes. The membranes may act as permselective barriers, or as an integral part of the catalytically active surface. Practical issues such as membrane pore blockage, thermal and mechanical stability, and the dilution caused by the need for sweep (i.e. permeate purge) gases, have limited the usefulness of the membrane reactor systems. Nevertheless, the benefits of the membrane systems have been demonstrated though a wide number of experimental reaction studies, examples of which include the dehydrogenation of ethane (Tsotsis, Champagnie, Vasileiadis, Zraka & Minet, 1992), cyclohexane (Sun & Khang, 1988), ethylbenzene (Wu, Gerdes, Pszczolowski, Bhave & Liu, 1990), and acetylene (Itoh, Xu & Sathe, 1993), CO production via the water–gas shift reaction (Uemiya, Sato, Ando & Kikuchi, 1991), and steam–methane reforming (Adris et al., 1994, Adris et al., 1997).

In comparison to the membrane reactors, a relatively small amount of work has been carried out on systems combining reaction with adsorption or chemical acceptor-based separation processes. Even so, such processes offer distinct advantages to the membrane-based systems in terms of the material tolerance to high temperatures and pressures, and the wide choice and availability of adsorbents for achieving the desired separations under reaction conditions. Furthermore, even through use of a purge gas for regeneration, the effective separation of the primary adsorbate from other non- or weakly adsorbing species can be achieved. Some examples of recent works employing chemical acceptors or adsorbents for reaction enhancement are now summarised.

Han and Harrison (1994) studied hydrogen production via the water–gas shift reaction using CaO as a CO2 acceptor in a tubular reactor. CO conversions were reported which exceeded that of the thermodynamic equilibrium conversion under the specified operating conditions. Brun-Tsekhovoi, Zadorin, Katsobashvili and Kourdyumov (1986) showed a very significant enhancement of CH4 conversion to H2 in a fluidised bed reactor containing Ni-based catalyst balls, and a specially treated form of dolomite as adsorbent. Typical industrial operating conditions were considered in this work, i.e. pressure levels of 103–104 kPa, and an operating temperature of 627°C. Goto, Tagawa and Oomiya (1993) studied the dehydrogenation of cyclohexane over a Pt–alumina catalyst and CaNi5 alloy as a hydrogen acceptor. The workers showed that at 150–190°C and ambient pressure, the overall conversion of cyclohexane to benzene could be exceeded by three-fold when compared to the catalyst-only case. Most recently, Carvill, Hufton and Sircar (1996) and Hufton, Mayorga and Sircar (1999) describe the general concept of the Sorption Enhanced Reaction Process (SERP), which utilises pressure and concentration swing adsorption principles for reaction enhancement; see also Vaporciyan and Kadlec (1989) and Alpay, Chatsiriwech and Kershenbaum (1995). CO production via the reverse water–gas shift reaction was specifically considered by Carvill et al. (1996), in which NaX zeolite was used as an adsorbent for water. The authors showed that at 250°C and 480 kPa, a CO2 conversion of 36% could be achieved; a conventional plug flow reactor required an operating temperature of 565°C for the same conversion. Furthermore, the process generated a high purity (+99% (v/v)) CO stream as product. Hufton et al. (1999) applied the SERP concept to H2 production via the steam–methane reforming (SMR) reactions. In specific, using a hydrotalcite-based CO2 adsorbent, and a commercial Ni-based catalyst, the authors showed that at 450°C and 480 kPa, +95% (v/v) H2 could be produced directly from reactor. The CH4 to H2 conversion was 82%, which could only be achieved at approximately 650°C with a conventional SMR reactor.

The present work also considers the sorption-enhanced steam–methane reforming (SE-SMR) process. The key reactions of the SMR process are given as:CH4+H2OCO+3H2,ΔH298=206kJ/mol,CH4+2H2OCO2+4H2,ΔH298=164.9kJ/mol,CO+H2OCO2+H2,ΔH298=−41kJ/mol.Reforming reactions (1) and (2) are strongly endothermic, so the forward reaction is favoured by high temperatures, while the water–gas shift reaction (3) is moderately exothermic and is therefore favoured by low temperatures. The reforming reactions will also be favoured at low pressures, whereas the water–gas shift reaction is largely unaffected by changes in pressure. In the presence of a selective CO2 adsorbent, the conversion of CH4 to CO2 though reaction (2) is favoured, as is the production of CO2 through CO intermediate. For a reaction which is not kinetically limited, the use of an adsorbent will thus enable a lower operating temperature for a desired conversion. However, on equilibration of the adsorbent, the separation effect is, of course, lost. This necessitates the periodic regeneration of the adsorbent and thus, for example, the pressure and concentration swing type operations mentioned above. In other words, sorption-enhanced reaction processes are inherently dynamic in operation. Adequate design and scale-up of such processes will thus require information on the kinetics of adsorption and desorption, as well as reaction kinetic models under transient conditions in the presence of an adsorbent.

Research work on the kinetics of the SMR process dates back to the 18th century (see Sabatier, 1922; Marek & Hahn, 1932), with the first extensive study by Akers and Camp in 1955. A good review of the work up until 1970 is given by Van Hook (1980), which covers kinetic studies over porous nickel catalysts and nickel foil over large temperature (260–1000°C) and pressure (100–5000 kPa) ranges. A considerable amount of work on the kinetic aspects of the SMR process has been carried out since 1970; see, for example, Schnell (1970),Ross and Steel (1973),Allen (1975),Phung Quach and Rouleau (1975),Munsted and Grabke (1981),De Deken, Deves and Froment (1982), and Xu & Froment, 1989a, Xu & Froment, 1989b. To date, the rate models proposed by Xu and Froment (1989a) are considered to be the most general in form, and have been extensively tested under typical industrial operating conditions (see, also, Elnashaie, Adris, Al-Ubaid & Soliman, 1990). However, like most previous work, the models are applicable to steady-state kinetics, and untested for forced-dynamic operation. It is interesting to note, however, that where some attention has been given to the transient kinetics, ideal surface conditions were maintained though vacuum operations; see, for example, Ross and Steel (1973). This, of course, limits the applicability of such models for process design applications.

In this work, attention is given to the analysis of SMR reaction kinetics under transient conditions depicting SERP-type operation, both in the presence and absence of a selective absorbent for CO2. Particular attention is given to the transient analysis of the Xu and Froment (1989a) kinetic model. The work then considers the influence of operating parameters on the degree of separation enhancement. In doing so, mathematical models of the process are verified, which can ultimately be used in the design, analysis or scale-up of pressure or concentration swing based adsorptive reactors.

Section snippets

Experimental

A commercial Ni-based catalyst (United Catalyst Inc.) containing 25–35% Ni, 25–35% NiO, 5–15% MgO and 15–25% sodium silicate, was used in this work. The original catalyst (1/8″ cylindrical pellets; BET area of 137.6 m2/g) was crushed and sieved into two particle size groups: 0.11–0.25 and 0.25–0.5 mm. The CO2 adsorbent consisted of an industrially supplied potassium promoted hydrotalcite, which was previously measured for its capacity and stability under wet gas conditions (see Ding & Alpay, 2000

Governing equations

A dynamic model accounting for non-isothermal, non-adiabatic, and non-isobaric operation, was developed to describe both the SMR and SE-SMR processes. For the SE-SMR process, the reactions and adsorption were assumed to take place on the surfaces of the catalyst and adsorbent, respectively. A Langmuir model was used to describe the adsorption equilibria of CO2, and a linear driving force (LDF) model for the intraparticle mass transfer of the adsorbent; further details of equilibria and kinetic

Results and discussion

Typical reactor effluent concentration profiles measured in the SMR and SE-SMR experiments under similar operating conditions (450°C, 445.7 kPa, H2O/CH4=6) are shown in Fig. 2(a) and (b), respectively. It can be seen that in the presence of CO2 adsorbent, the transient period (∼5 min) is much longer than that of the SMR run (∼1.5 min). However, since the experimental conditions are nearly identical, the steady-state concentrations of CH4, H2, and CO2 are approximately equal, indicating no strong

Conclusions

The steady-state kinetic model of Xu and Froment (1989a) for steam–methane reforming has been shown to be applicable to transient reactor operation, both in the presence and absence of adsorbent. In particular, a reactor model accounting for non-isothermal, non-adiabatic and non-isobaric operation, and mass transfer limited adsorption, was found to accurately predict the elution profiles of the reaction species over an admixture of Ni catalyst and hydrotalcite adsorbent. Kinetic and equilibrium

References (43)

  • E. Alpay et al.

    Combined reaction and separation in pressure swing processes

    Chemical Engineering Science

    (1995)
  • Armor, J. N. (1995). Membrane catalysis: What is it now, what needs to be done? Catalysis Today, 25,...
  • R.B. Bird et al.

    Transport phenomena

    (1960)
  • N.M. Bodrov et al.

    Kinetics of the reaction of methane with water vapour catalysed by nickel on a porous carrier

    Kinetika L Kataliz

    (1967)
  • N.M. Bodrov et al.

    Kinetics of the reaction of methane with steam on the surface of nickel at 400–600°C

    Kinetika L Kataliz

    (1968)
  • A.R. Brun-Tsekhovoi et al.

    The process of catalytic steam-reforming of hydrocarbons in the presence of carbon dioxide acceptor

  • B.T. Carvil et al.

    Sorption-enhanced reaction process

    A.I.Ch.E. Journal

    (1996)
  • J.C. De Deken et al.

    Steam reforming of natural gas

    Chemical Reaction Engineering ACS Symposium Series

    (1982)
  • A.P. De Wash et al.

    Heat transfer in packed bed

    Chemical Engineering Science

    (1972)
  • Ding, Y., Alpay, E. (2000). Equilibria and kinetics of high temperature CO2 adsorption on hydrotalcite adsorbent, in...
  • S. Ergun

    Fluid flow through packed columns

    Chemical Engineering Progress

    (1952)
  • Cited by (297)

    View all citing articles on Scopus
    View full text