- Email: [email protected]
An unusual form of non-Arrhenius behaviour in ethyne hydrogenation over palladium catalysts
Catalysis, 55 (1989) L5-L8 Elsevier Science Publishers B.V., Amsterdam -
L5
Applied
Printed in The Netherlands
An Unusual Form of Non-Arrhenius Behaviour in Ethyne Hydrogenation over Palladium Catalysts RICHARD B. MOYES, DAVID W. WALKER, PETER B. WELLS and DAVID A. WHAN* School of Chemistry,
University
of Hull, Hull, HU6 7RX (U.K.)
and ELIZABETH A. IRVINE ICI Chemicals & Polymers Ltd, Research & Technology Middlesbrough, Cleveland, TS6 8JE (U.K.)
Department,
P.O. Box 90, Wilton,
(Received 25 September 1989)
The selective hydrogenation of the ethyne present in “ethene” streams from naphtha crackers is of considerable commercial importance, especially in the purification of ethene to be used for the manufacture of polythene where ethyne can destroy the polymerisation catalyst. The variation with temperature of the rate of metal catalysed hydrogenation of unsaturated hydrocarbons does not always obey the Arrhenius law. For example, the rate of benzene hydrogenation has been reported to exhibit a maximum at 417 K on palladium catalysts [l] and on nickel the slope of the Arrhenius plot has been reported to change in magnitude at about the same temperature [ 2 1. The rate of hydrogenation of ethene on nickel catalysts displays a maximum at temperatures ranging from 408 to 433 K [ 3,4]. The explanation of the maximum in rate has been variously interpreted in terms of either desorption of reactants at higher temperatures [5,6] or a change in reaction mechanism or rate limiting step at the inversion temperature [ 7,8 J. We have determined the rates of ethyne hydrogenation in a static batch reactor (volume 150 ml) as a function of temperature and ethyne pressure. The extent of reaction in a Pyrex vessel was monitored by means of a pressure transducer and samples were withdrawn as required for gas chromatographic analysis. Three catalysts were used, as follows: catalyst A, 0.04 wt.-% palladium on y-alumina (support surface area 90 m2 g-l, palladium area 173 m* gPd_l); catalyst B, 0.036 wt.-% palladium on a-alumina (support surface area 5.7 m2 g-l, palladium area 150 m2 gPd_‘); catalyst C, unsupported palladium powder prepared by reduction of palladium chloride in flowing hydrogen (palladium surface area 1.9 m2 gPd-‘). Before admission of the reactants, catalyst
0166-9834/89/$03.50
0 1989 Elsevier Science Publishers B.V.
L6
samples were reduced in 200 Torr of hydrogen at 473 K for 2 h and then pumped for 0.5 h. Plots showing the temperature dependence of the activity for the hydrogenation of ethyne on the palladium catalysts described are presented in Fig. 1. All display a marked and reproducible break in both activity and activation energy at temperatures around 363 to 368 K. Behaviour was reversible in that results were collected in random order over the whole temperature range. Similar behaviour, in the same temperature range, was observed over catalyst A when perdeuteroethyne reacted with deuterium. The initital rates of ethyne hydrogenation were determined as a function of ethyne pressure at various temperatures and the results obtained using catalyst B are illustrated in Fig. 2. At a given temperature, the initial rate at first increased with increasing ethyne pressure, then dropped sharply, and finally declined slowly. The pressure at which the discontinuity, and change in ethyne order from positive to slightly negative, are observed increases with increasing temperature of reaction. Similar behaviour was observed with catalysts A and C. The above observations may be correlated in a manner such as that presented diagrammatically in Fig. 3. In Fig. 3a, which is a plot of initial rate of hydrogenation as a function of initial ethyne pressure, temperature dependence would be represented by an ordinate such as the line PQ: appropriate transfer of the information from this ordinate to an Arrhenius plot would yield Fig. 3b. We are thus of the opinion that the discontinuities in the Arrhenius plots and in the initial rates of reaction as a function of initial ethyne pressure are manifestations of the same phenomenon. If the observed phenomenon were associated with the j? to a phase change in palladium hydride, the transition temperature on the Arrhenius plot would
2.2
2.6
3.0 lO'/T
3.4
2.2
2.6
50
2.2
2.6
3.0
103/T
Fig. 1. Variation of the initial rates of ethyne hydrogenation, r, with temperature. (a) Reaction over 0.152 g of catalyst A (initial reactant pressures: ethyne 50.2 Torr, hydrogen 104 Torr 1; (b ) reaction over 0.019 g of catalyst B (initial reactant pressures: ethyne 30.0 Torr, hydrogen 60.6 Torr ); (c) reaction over 0.006 g of catalyst C (initial reactant pressure: ethyne 29.9 Torr, hydrogen 60.8 Torr). Units of r=Torr min-‘.
L7
htlal
ethyne
pressure
ITorr
Fig. 2. Variation of the initial rates of ethyne hydrogenation, r, with initial ethyne pressure for reaction over 0.019 g of catalyst B. (0 ) 356 K, (a) 384 K, ( l ) 402 K. Initial hydrogen pressure: 60.4 Torr throughout. Units of r=mol hydrogen consumed (gPd) -i min-i.
Initial
ethyne
pressure
IO+
Fig. 3. A diagrammatic illustration of the proposed relationship between the discontinuities observed in the rate versus ethyne pressure plots and in the Arrhenius plots for the ethyne hydrogenation reaction [ (a) and (b) see text].
be expected to shift to a lower value for the reaction of perdeuteroethyne with deuterium [ 91. Such a shift was not observed. Nevertheless, the discontinuities are characteristic of a phase change and we suggest that these initial results may be interpreted on the basis of two different modes of packing of ethyne molecules on the palladium surface. At pressures below the discontinuity, the reaction order in ethyne is positive, indicating that ethyne is not particularly strongly adsorbed. On passing to the high-pressure side of the discontinuity, the reaction order in ethyne is slightly negative; this indicates that ethyne is strongly adsorbed, probably to an extent which reduces the hydrogen coverage of the surface and thus leads to a lower rate of hydrogenation. Activation energies determined for heterogeneously catalysed reactions are apparent values and can differ from true values because
L8
of heat of adsorption effects [lo]. In general, when reactants and products are weakly adsorbed the true and apparent activation energies differ by an amount equal to the sum of the heats of adsorption of the reactants. However, if the surface coverage of a reactant approaches that corresponding to a monolayer, the heat of adsorption of that molecule is not included in the measured apparent activation energy for the reaction. It would thus appear that, on passing through the observed pressure discontinuity region, ethyne adsorption changes from weak to strong and that the breaks in the Arrhenius plots are a direct consequence of this change, Any alteration in the strength of adsorption of ethyne, or in ethyne surface coverage, with ethyne pressure or temperature should have implications on both the selectivity of ethyne hydrogenation (ratio of ethene to ethane produced) and the yield of oligomers formed [ll]. These additional factors are under investigation, ACKNOWLEDGEMENT
We thank Imperial Chemical Industries plc and the Science and Engineering Research Council for the award of a CASE studentship to DWW.
REFERENCES 1 2 3 4 5 6 7 8 9 10 11
R.Z.C. van Meertenand J.W.E. Coenen, J. Catal., 46 (1977) 13. H. Kubica, J. Catal., 12 (1968) 223. P. Pareja, A. Amariglio and H. Amariglio, C.R. Acad. Sci., Ser. C., 276 (1973) 1505. K. Miyahara and H. Narumi, J. Res. Inst. Catal., Hokkaido University, 6 (1958) 250. H. zur Strassen, Z. Phys. Chem., 169 (1934) 81. G.H. Twiggand E.K. Rideal, Proc. Roy. Sot., A171 (1939) 55. T. Tucholski and E.K. Rideal, J. Chem. Sot., (1935) 1701. A. Farkas and L. Farkas, J. Am. Chem. Sot., 60 (1938) 22. F.A. Lewis, p. 118. The Palladium Hydrogen System, Academic Press, New York, 1967. G.C. Bond, p. 126. Catalysis by Metals, Academic Press, London, 1962. W.T. McGown, C. Kemball and D.A. Whan, J. Catal., 51 (1978) 173.
L5
Applied
Printed in The Netherlands
An Unusual Form of Non-Arrhenius Behaviour in Ethyne Hydrogenation over Palladium Catalysts RICHARD B. MOYES, DAVID W. WALKER, PETER B. WELLS and DAVID A. WHAN* School of Chemistry,
University
of Hull, Hull, HU6 7RX (U.K.)
and ELIZABETH A. IRVINE ICI Chemicals & Polymers Ltd, Research & Technology Middlesbrough, Cleveland, TS6 8JE (U.K.)
Department,
P.O. Box 90, Wilton,
(Received 25 September 1989)
The selective hydrogenation of the ethyne present in “ethene” streams from naphtha crackers is of considerable commercial importance, especially in the purification of ethene to be used for the manufacture of polythene where ethyne can destroy the polymerisation catalyst. The variation with temperature of the rate of metal catalysed hydrogenation of unsaturated hydrocarbons does not always obey the Arrhenius law. For example, the rate of benzene hydrogenation has been reported to exhibit a maximum at 417 K on palladium catalysts [l] and on nickel the slope of the Arrhenius plot has been reported to change in magnitude at about the same temperature [ 2 1. The rate of hydrogenation of ethene on nickel catalysts displays a maximum at temperatures ranging from 408 to 433 K [ 3,4]. The explanation of the maximum in rate has been variously interpreted in terms of either desorption of reactants at higher temperatures [5,6] or a change in reaction mechanism or rate limiting step at the inversion temperature [ 7,8 J. We have determined the rates of ethyne hydrogenation in a static batch reactor (volume 150 ml) as a function of temperature and ethyne pressure. The extent of reaction in a Pyrex vessel was monitored by means of a pressure transducer and samples were withdrawn as required for gas chromatographic analysis. Three catalysts were used, as follows: catalyst A, 0.04 wt.-% palladium on y-alumina (support surface area 90 m2 g-l, palladium area 173 m* gPd_l); catalyst B, 0.036 wt.-% palladium on a-alumina (support surface area 5.7 m2 g-l, palladium area 150 m2 gPd_‘); catalyst C, unsupported palladium powder prepared by reduction of palladium chloride in flowing hydrogen (palladium surface area 1.9 m2 gPd-‘). Before admission of the reactants, catalyst
0166-9834/89/$03.50
0 1989 Elsevier Science Publishers B.V.
L6
samples were reduced in 200 Torr of hydrogen at 473 K for 2 h and then pumped for 0.5 h. Plots showing the temperature dependence of the activity for the hydrogenation of ethyne on the palladium catalysts described are presented in Fig. 1. All display a marked and reproducible break in both activity and activation energy at temperatures around 363 to 368 K. Behaviour was reversible in that results were collected in random order over the whole temperature range. Similar behaviour, in the same temperature range, was observed over catalyst A when perdeuteroethyne reacted with deuterium. The initital rates of ethyne hydrogenation were determined as a function of ethyne pressure at various temperatures and the results obtained using catalyst B are illustrated in Fig. 2. At a given temperature, the initial rate at first increased with increasing ethyne pressure, then dropped sharply, and finally declined slowly. The pressure at which the discontinuity, and change in ethyne order from positive to slightly negative, are observed increases with increasing temperature of reaction. Similar behaviour was observed with catalysts A and C. The above observations may be correlated in a manner such as that presented diagrammatically in Fig. 3. In Fig. 3a, which is a plot of initial rate of hydrogenation as a function of initial ethyne pressure, temperature dependence would be represented by an ordinate such as the line PQ: appropriate transfer of the information from this ordinate to an Arrhenius plot would yield Fig. 3b. We are thus of the opinion that the discontinuities in the Arrhenius plots and in the initial rates of reaction as a function of initial ethyne pressure are manifestations of the same phenomenon. If the observed phenomenon were associated with the j? to a phase change in palladium hydride, the transition temperature on the Arrhenius plot would
2.2
2.6
3.0 lO'/T
3.4
2.2
2.6
50
2.2
2.6
3.0
103/T
Fig. 1. Variation of the initial rates of ethyne hydrogenation, r, with temperature. (a) Reaction over 0.152 g of catalyst A (initial reactant pressures: ethyne 50.2 Torr, hydrogen 104 Torr 1; (b ) reaction over 0.019 g of catalyst B (initial reactant pressures: ethyne 30.0 Torr, hydrogen 60.6 Torr ); (c) reaction over 0.006 g of catalyst C (initial reactant pressure: ethyne 29.9 Torr, hydrogen 60.8 Torr). Units of r=Torr min-‘.
L7
htlal
ethyne
pressure
ITorr
Fig. 2. Variation of the initial rates of ethyne hydrogenation, r, with initial ethyne pressure for reaction over 0.019 g of catalyst B. (0 ) 356 K, (a) 384 K, ( l ) 402 K. Initial hydrogen pressure: 60.4 Torr throughout. Units of r=mol hydrogen consumed (gPd) -i min-i.
Initial
ethyne
pressure
IO+
Fig. 3. A diagrammatic illustration of the proposed relationship between the discontinuities observed in the rate versus ethyne pressure plots and in the Arrhenius plots for the ethyne hydrogenation reaction [ (a) and (b) see text].
be expected to shift to a lower value for the reaction of perdeuteroethyne with deuterium [ 91. Such a shift was not observed. Nevertheless, the discontinuities are characteristic of a phase change and we suggest that these initial results may be interpreted on the basis of two different modes of packing of ethyne molecules on the palladium surface. At pressures below the discontinuity, the reaction order in ethyne is positive, indicating that ethyne is not particularly strongly adsorbed. On passing to the high-pressure side of the discontinuity, the reaction order in ethyne is slightly negative; this indicates that ethyne is strongly adsorbed, probably to an extent which reduces the hydrogen coverage of the surface and thus leads to a lower rate of hydrogenation. Activation energies determined for heterogeneously catalysed reactions are apparent values and can differ from true values because
L8
of heat of adsorption effects [lo]. In general, when reactants and products are weakly adsorbed the true and apparent activation energies differ by an amount equal to the sum of the heats of adsorption of the reactants. However, if the surface coverage of a reactant approaches that corresponding to a monolayer, the heat of adsorption of that molecule is not included in the measured apparent activation energy for the reaction. It would thus appear that, on passing through the observed pressure discontinuity region, ethyne adsorption changes from weak to strong and that the breaks in the Arrhenius plots are a direct consequence of this change, Any alteration in the strength of adsorption of ethyne, or in ethyne surface coverage, with ethyne pressure or temperature should have implications on both the selectivity of ethyne hydrogenation (ratio of ethene to ethane produced) and the yield of oligomers formed [ll]. These additional factors are under investigation, ACKNOWLEDGEMENT
We thank Imperial Chemical Industries plc and the Science and Engineering Research Council for the award of a CASE studentship to DWW.
REFERENCES 1 2 3 4 5 6 7 8 9 10 11
R.Z.C. van Meertenand J.W.E. Coenen, J. Catal., 46 (1977) 13. H. Kubica, J. Catal., 12 (1968) 223. P. Pareja, A. Amariglio and H. Amariglio, C.R. Acad. Sci., Ser. C., 276 (1973) 1505. K. Miyahara and H. Narumi, J. Res. Inst. Catal., Hokkaido University, 6 (1958) 250. H. zur Strassen, Z. Phys. Chem., 169 (1934) 81. G.H. Twiggand E.K. Rideal, Proc. Roy. Sot., A171 (1939) 55. T. Tucholski and E.K. Rideal, J. Chem. Sot., (1935) 1701. A. Farkas and L. Farkas, J. Am. Chem. Sot., 60 (1938) 22. F.A. Lewis, p. 118. The Palladium Hydrogen System, Academic Press, New York, 1967. G.C. Bond, p. 126. Catalysis by Metals, Academic Press, London, 1962. W.T. McGown, C. Kemball and D.A. Whan, J. Catal., 51 (1978) 173.