Catalysis Volume 6

A Review of the Recent Literature Published up to mid–1982

By G. C. Bond, G. Webb

The Royal Society of Chemistry

Copyright © 1983 The Royal Society of Chemistry
All rights reserved.
ISBN: 978-0-85186-574-4

Contents

Chapter 1 Oscillatory Phenomena in Heterogeneous Catalysed Oxidation Reactions By D. Mukesh, M. Goodman, C. N. Kenney, and W. Morton, 1,
Chapter 2 Strong Metal–Support Interactions By G. C. Bond and R. Burch, 27,
Chapter 3 The Catalytic Hydrogenation of Organic Compounds – A Comparison between the Gas-phase, Liquid-phase, and Electrochemical Routes By M. D. Birkett, A. T. Kuhn, and G. C. Bond, 61,
Chapter 4 Structural Characterization of Surface Species and Surface Sites by Conventional Optical Spectroscopies By A. Zecchina, E. Garrone, and E. Guglielminotti, 90,
Chapter 5 Use of Radiotracers in the Study of Surface Catalysed Processes By G. F. Berndt, 144,
Chapter 6 Hydroformylation By B. A. Murrer and M. J. H. Russell, 169,
Chapter 7 Formation of Oxygenated Products from Synthesis Gas By E. K. Poets and V. Ponec, 196,

CHAPTER 1

Oscillatory Phenomena in Heterogeneous Catalysed Oxidation Reactions Oxidation Reactions

BY D. MUKESH, M. GOODMAN, C. N. KENNEY, AND W. MORTON

1 Introduction

The observation of oscillations in heterogeneous catalytic reactions is an indication of the complexity of catalyst kinetics and makes considerable demands on the theories of the rates of surface processes. In experimental studies the observed fluctuations may be in catalyst temperature, surface species concentrations, or most commonly because of its accessibility, in the time variation of the concentrations of reactants and products in contact with the catalyst. It is now clear that spontaneous oscillations are primarily due to non-linearities associated with the rates of surface reactions as influenced by adsorbed reactants and products, and the large number of experimental studies of the last decade have stimulated a considerable amount of theoretical kinetic modelling to attempt to account for the wide range of oscillatory behaviour observed.

Several homogeneous gas- and liquid-phase reactions are now also known to exhibit self oscillations and it is clear that many living organisms depend on coupled oscillatory reactions catalysed by enzymes to control biological functions. However, only heterogeneous oxidation reactions catalysed by noble metals are reviewed here. Experimental studies are first described, followed by a discussion of kinetic analyses which have been put forward to account for them. Particular attention is given to the most extensively studied system to date, the oxidation of CO over Pt catalysts.

2 CO Oxidation

Sheintuch and Schmitz have thoroughly reviewed oscillatory oxidation reactions up to 1977. A year later another review was published by Slinko and Slinko. Varghese et al. reported oscillations in the rate of CO production during the oxidation of CO over Pt supported on γ-[Al.sub.2][O.sub.3] catalyst between 100–150 °C. They also observed multiple peak limit cycles in gas-phase concentrations in the presence of hydrocarbon impurities. Oscillations have been observed by Plichta and Schmitz and by Sheintuch over Pt foil over a temperature range of 150–250 °C. The former authors also detected simultaneous oscillations in the catalyst temperature. The time period of oscillation observed by Sheintuch was of the order of 25 min. The oscillations were of single and multiple peak type. The feed gas concentration was varied between 0–10% CO and 10–20 0. Turner et al. observed rate oscillations on Pt wire over a wide range of temperature from 150–300°C. The oscillations were observed when the ratio of partial pressures of CO to 0 was in the range of 0.001 to 0.045. Dauchot and Van Cakenberghe have also observed oscillations on Pt wire. Gray et al. and Barkowski et al. have found oscillations in the production of CO2 when a feed of 5% CO in O2 was passed over polycrystalline Pt at 250°C. Hugo and Jakubith reported oscillations when a mixture of CO and air was passed over Pt gauze at 120°C. The time period of these oscillations was around 2min. Beusch et al., McCarthy et al., and Rathousky et al. have observed oscillations at 180 °C over a Pt catalyst supported on A10. Beusch et al. also found oscillations in catalyst temperature of the order of 2–3 °C. The oscillations observed by Rathousky et al. could be spikes or quasi-sinusoid al with a period of oscillations as high as 8h. Turner et al. found oscillations in C0 production under similar conditions even when the reaction was carried out over polycrystalline Pd or Ir wire. Sales et al. observed formation of an ‘oxide’ layer on the Pt wire during the reaction and they suggested this fact as the cause of the observed oscillations.

3 H2 Oxidation

The oxidation reaction of H2 has also been shown to exhibit relaxation and sinusoidal oscillations over noble metal catalysts for a wide range of temperatures. Thus Wicke et al. observed oscillations during the oxidation of H2 over 0.4% Pt on SiO2–Al2O3 support over a temperature range of 95–200 °C. The variation in the catalyst temperature was of the order of [+ or -] 20 °C. Rajagopalan and Luus observed oscillations over Pt wire in the presence of impurities. Beusch et al. observed similar oscillations when a mixture of 3.14% H2 in air was passed over Pt catalyst at 80 °C. Horak and Jiracekand Zuniga and Luus have also observed oscillations over Pt catalyst. The former authors observed that there was a very large difference between the catalyst and the gas temperature when oscillations occur. Kurtanjek et al. observed oscillations on Ni plate over a temperature range of 160–400 °C. Belyaev et al. found oscillations over Ni foil at 180°C in the presence of excess H. The period of oscillation varied between 6 and 120 s. Schmitz et al. observed multi-peak oscillations and chaotic behaviour when H2 oxidation was carried out on Ni catalyst in the presence of excess H2.

4 Hydrocarbon Oxidation

Vayenas et al. observed multiple peak oscillations during the oxidation of ethene over polycrystalline Pt film between 200–400 °C. Formation and disintegration of an oxide film on the catalyst was given as the reason for the observed periodic behaviour. Sheintuch and Luus recently observed oscillations during the oxidation of propene when 1% propene in 0 was passed over Pt wire over a temperature range of 175–228 °C. Krylov and his co-workers have also reported oscillations in cyclohexane oxidation on Zeolite NaX and in propene oxidation on the surface of CaO–MgO solid solutions.

5 NH3 Oxidation

Stephanopoulos et al, have reported oscillations in studies on the oxidation of NH3 over Pt wire or foils. In this case there were temperature fluctuations on the catalyst of the order of 20 °C. The feed consisted of 20&dnash;40% NHin air. The period of oscillation varied from 1–200s. The oxidation products consisted of NO, N2, and H2O.

6 Oxidation of CO/H2 or CO/Hydrocarbon Mixtures

Oscillations have been observed in this department when mixtures of gases were oxidized over supported Pt catalyst, although such behaviour was not found during the oxidation of CO alone. Multi-peak oscillations have been reported during the oxidation of mixtures of CO and but-1-ene above 150°C when the feed consisted of 2% CO, 3% O2, and 1% but-1-ene. The period of oscillation varied from 1.5 to 90 min. Goodman has observed sinusoidal oscillations during the oxidation of a mixture of CO and H2 or CO and trans-but-2-ene at similar conditions. Cutlip and Kenney have also observed relaxation-type oscillations during the oxidation of a mixture of CO and propene.

In contrast to other studies, oxidation carried out in this department on a Pt/γ-Al2O3 catalyst has not uncovered any oscillatory behaviour in the temperature range of 100–185 °C. Addition of a hydrocarbon like but-1-ene, but-2-ene, or propene induces sinusoidal or relaxation type oscillations at temperatures above 150°C. The experimental set-up used consists of a continuous recycle reactor system. The catalyst is packed in the cylindrical tubes. The gas flow rates are precisely measured with a bubble flow-meter. The reactor outlet is connected to a magnetic deflection mass spectrometer. An electronic peak select unit allows up to four mass numbers to be continuously monitored. The output data are connected to a PDP 11/45 computer for automatic and fast data logging. The data thus stored in the computer can be analysed later. The line diagram of the experimental set up is given in Figure 1.

Limit cycles observed when mixtures containing 2% CO, 3% O2, and 1% propene at 68 cm3 min-1 and 44 cm3 min-1 are shown in Figures 2 and 3 and mixtures containing 2% CO, 3% O2, and 1% but-1-ene at 50cm3 min-1 and 68.9 cm3min-1 are shown in Figures 4a and 4b. The but-1-ene system exhibits multi-peak relaxation oscillations at higher flow rates with very long time period of oscillation. A feature of this catalyst system is that the CO2 product is adsorbed on the A12O3 support and desorbs somewhat slowly relative to the other rate processes.

7 Forced Periodic Oscillations

A related oscillatory phenomenon is that in which the concentration of one or more reactants, fed to a flow reactor, is varied in time. Such forced periodic feed oscillations during oxidation reactions have now been studied by a number of authors. It is found that not only can conversion be increased but the selectivity of certain parallel reactions can be improved, which may be of value in industrial applications. Cutlip and Abdul-Karem and Jain have observed increased conversion during the oxidation of CO over both Pt and V2O5 catalysts. Hegedus et al. have also observed an improvement in conversion during the oxidation of CO and NO over Pt/γ-Al2O3 at 500°C. The feed was switched between NO and a mixture of CO and O2. Unni et al. observed a 30% improvement in selectivity during the oxidation of SO2 over V2O5. Renken et al. showed experimentally that improvement in selectivity could be achieved during the oxidation of ethene over Ag catalyst. Horn et al. and Bailey et al. have demonstrated theoretically that the selectivity where parallel reactions occur could be improved by varying feed composition in periodic manner.

Another advantage of forced periodic feed experiments, which has not been fully exploited so far, is that the technique could be used for kinetic model discrimination, a technique in which large deviations could be induced into calculated reponses between rival models under consideration. Hawkins has carried out experiments on oxidation of CO for discriminating between several Hougen and Watson rival models. Cutlip et al. have compared experimental forced periodic feed CO oxidation experimental transients with simulations using an elementary step model and compared theory with experiment in studies of the variation of the conversion as a function of time period of the forced oscillation.

8 Steady State and Dynamic Models

In considering kinetic models which can display oscillatory behaviour, it is useful to recall the Langmuir–Hinshelwood approach to a simple reaction such as the oxidation of CO, taking place in a closed system and consider the commonly adopted assumptions:

CO + 1/2O2 -> CO2

The adsorption of the gases on surface metal sites followed by reaction could be written as:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

The rate of adsorption of CO from the gas phase is

d (CO)/dt = -k1(CO)θv + k-1 (CO-S) (2)

whilst on the catalyst surface:

dθ CO/Dt = d(CO – S)/dt = k1(CO)θ v – k-1(CO – S) – k3(CO – S)(O – S) (3)

where θv is the number of vacant sites = 1 – θCO – θ0, and θCO is the surface coverage of CO. Similar expressions exist for O2 adsorption but allow for dissociative adsorption. The surface reaction rate term is usually written as first order in both adsorbed species, that is second order in surface concentrations. Thus:

d(CO2)/dt = k3θCO θ 0 – k-3(CO2) (4)

where θCO and θ0 are the surface coverages of CO and O2, respectively, on the catalyst. The following 5 assumptions are then usually made.

1. The adsorption–desorption steps are fast compared with surface reaction so the term in k3 can be neglected in equation (3).

2. A steady state is established between gas-phase and surface concentrations so the time derivatives in equations (2) and (3), etc. are effectively zero.

3. Surface concentrations such as θCO may be written in terms of k1/k-1 = KCO, the equilibrium adsorption constant, and gas-phase concentrations.

4. In this reacion the reactants compete for the same type of surface site.

5. CO2 is not adsorbed on the metal surface sites giving k-3 = 0.

These assumptions lead to an expression for the rate of formation of CO2 in terms of reactant gas-phase pressures.

d(CO2])/dt = k3θCO θ0

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (5)

Such equations have some success, albeit often qualitative, in describing the variation of reaction rate with CO and O2 partial pressures. In particular if [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII], then at low pCO the rate is proportional to pCO, and at sufficiently high pCO the rate falls due to high coverage of surface sites with CO, the rate becoming inversely dependent on CO pressure as first observed by Langmuir. It will also be noted that in this very simple formulation all rate constants are assumed to be independent of surface coverage.

In an open system, such as a well mixed flow reactor (CSTR), the convective transport of reactants and products must be allowed for; the analogue of equation (2) is:

Vd(CO)out/dt = vin(CO)in – vout (CO)out – k1(CO)outθ v + k-(CO-S) (2′)

Here V is the reactor gas-phase volume, v the volumetric flow rate to and leaving the reactor, and the concentration defined by the subscript ‘out’ applied throughout the reactor since it is well mixed.

If the set of assumptions (1–3) apply, then the steady state concentration of CO leaving a CSTR for a given (CO)in is obtained by solving the non-linear algebraic equation:

vin(CO)in – vout(CO) out = rx = 0 (5′)

where rx, the rate of CO oxidation, is given by (5′). We shall call equations (2), (3), and (2′) with finite derivative terms ‘elementary step’ equations and equations of the form (5) and (5′) ‘steady state’ equations.

Equation (5′) has a number of important and well known features which follow from inspection of Figure 5, where the O2 concentration is calculated from an analogous coupled mass balance equation for O2.

(a) For appropriate values of v, V, and (CO)in, the system can have one or two stable steady states, one of high conversion and one of low conversion. These stable states can be on either side of a third ‘unstable state’, that is unstable with respect to concentration perturbations in (CO).

(b) The flow reactor shows quite different features from those displayed by the corresponding batch reactor; for which the rate-concentration behaviour would have to be much more non-linear than (5) to show multiples states.

(c) Alterations in the relative positions of the rate-concentration curve (by altering the temperature) or the mass balance line (by changing the inlet concentration or flow rate) can induce transitions from one stable state to the other.

(d) A theoretical mechanism which might produce (non-linear) oscillations is the existence of an additional ‘slow’ variable which periodically alters the relative position of the rate envelope relative to the mass balance line so producing periodic transitions between high and low conversion states. This picture is conceptually relatively simple but unfortunately of limited utility in accounting for experimental observations.

In an attempt to describe what are clearly time dependent phenomena, recourse has been made to solving equations of the forms (2), (3), and (3′). The systematic study of the dynamics of a heterogeneous reacting system of n reactants unfortunately involves the solution of at least (2n + 1) coupled non-linear ordinary differential equations; n for gas-phase reactants, as in equation (2′), n for each surface concentration, as in equation (3), and at least one surface reaction equation like equation (4). The absence of a tractable theory of the stability and behaviour of sets of coupled non-linear differential equations, which are also ‘stiff’, has necessitated the use of simplifying assumptions wherever possible. A widely adopted approach is to reduce the dynamical variables in the problem preferably to two, and apply a standard linearized stability analysis to the equilibrium points of the differential equations.

9 Modelling of Oscillations in CO Oxidation

The oxidation of CO over Pt/Al2O3 catalyst can be represented with an elementary step model described in equation (1), together with an extra equation for the adsorption of CO2 on the support:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1a)

where S’ represents a vacant site on alumina. The simpler assumptions that can be made for describing the reaction in a CSTR are given below.

1. Neglect reaction between adsorbed O2 and gaseous CO (Eley–Rideal step).

2. Rate constants are independent of surface coverages.

3. Oxygen dissociatively adsorbs on the catalyst.

4. All gases compete for the same type of sites on the Pt.

The model equations describing the response of a well stirred reactor catalysing the above reaction system are:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (6)

The rate constants are in dimensionless form. The differential equations for the two gases (C1 and C2) and their corresponding surface concentrations (C4 and C5) can be represented in a simplified form as:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (7)

The modified rate constants α1, α 2, β1, β2, β3, β4, and ε are defined in the symbols table. Figure 6 shows the transient behaviour of the system defined in (2) for the set of rate constant values given in Table 1. The system exhibits sinusoidal type oscillation with a time period of 0.87 residence time. It is seen that excess CO is required to simulate these oscillations. However, it is found from our work and that of others that simple elementary step models alone cannot generate oscillations with excess O2 in the feed, which is a condition under which many experimentally observed oscillations occur.

(Continues…)Excerpted from Catalysis Volume 6 by G. C. Bond, G. Webb. Copyright © 1983 The Royal Society of Chemistry. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.