Journal of Molecular Catalysis, 79 (1993) 253-264 Elsevier Science Publishers B.V., Amsterdam

253

MO11

Influence of support and metallic precursor on the hydrogenation of p-chloronitrobenzene over supported platinum catalysts

Bernard Cog*, Amina Tijani, Roger Dutartre and FranCois Fig&as Laboratoire de Chimie Organique Physique et Cindtique Chimique Appliqudes, UA 418 CNRS; ENSCM, 8 rue de l%cole Normale, 34053 Montpellier Ckdex, (France); tel. (+33)67144395, fax.

( + 33)67144353

(Received April 29,1992; accepted September 15,1992)

Abstract

Hydrogenation of p-chloronitrobenzene (CNB) has been studied, in methanol suspension, at 303 K and atmospheric pressure, over alumina, magnesia, titania and graphite supported plat- inum catalysts. The catalysts were prepared by using anionic, cationic or organometallic platinum precursors. The nature of the precursor does not modify markedly the catalytic properties of plat- inum for that reaction. However, the inhibiting effect ofp-chloroaniline (CAN), the main product formed on CNB hydrogenation is the highest on Pt/AI,Os prepared from the cationic precursor. For similar sizes of the Pt particles, the greatest influence on activity and selectivity is observed when titania is used as carrier. There is a tenfold increase of turnover frequency on Pt/TiOz compared to Pt/A1,03. At high CNB conversion ( > 98% ), the yield of CAN increases from 85.2% on Pt/Al,Os to 99.3% on Pt/TiOz reduced at high temperature (773 K). The improvement of CAN selectivity stems mainly from enhancement of the relative reactivity between hydrogenation of the nitro group and hydrogenolysis of the C-Cl bond of CAN. It is proposed that the migration of suboxide TiO, species (x < 2) on to the Pt particles, in a strong metal/support interaction state, is responsible for this behaviour. A schematic model of the reaction site is presented, in which the migrating TiO, adspecies on Pt activate the N=O bond which becomes highly susceptible to hy- drogen attack.

Key words: p-chloroaniline; p-chloronitrobenzene; hydrogenation; platinum; supported catalysts

Introduction

There is a growing interest to develop supported metallic catalysts for the selective hydrogenation of a specific function in polyfunctional organic mole- cules. To achieve this goal the properties of the metal catalyst can be tuned by modifying the size of the metal particles, by alloying or by initiating some kind of interaction between the metal particles and the carrier. For the hydrogen- ation of p-chloronitrobenzene (CNB ) we have studied the behaviour of some Pt catalysts in relation to the size of the platinum particles for monometallic Pt/A1203 catalysts [l], or with the effect of adding a second metal to Pt in PtM/A1203 catalysts (M = Sn, Pb, Ge, Al, Zn) [ 21. It was found that the se- lectivity to p-chloroaniline (CAN) is affected by either changing the size of Pt

*Author to whom correspondence should be addressed.

0304-5102/93/$06.00 0 1993 - Elsevier Science Publishers B.V. All rights reserved.

254 B. Coq et al./J. Mol. Catal. 79 (1993) 253-264

particles or by alloying Pt with a second metal. The highest yields in haloam- ines are obtained on large Pt particles [l] and with PtSn, PtZn and PtPb bimetallics [ 21. The reason for this behaviour is not a decrease of the reactivity of the C-Cl bond of CAN on large Pt particles or bimetallics, but a lower ad- sorption strength of CAN with respect to CNB. The weakening of the bond between Pt and CAN could stem from a higher electronic density at the Pt atom [ 1,2]. It is now well established that similar kinds of electronic modifi- cation can be induced by the nature of the support. Two good examples can be found in the hydrogenation of Q-unsaturated aldehydes over supported group VIII metals. The hydrogenation of crotonaldehyde to crotyl alcohol [ 31 is to- tally unselective over Pt/SiO,, but becomes selective over Pt/TiOa reduced at high temperature due to the so-called strong metal support interaction (SMSI) effect. The enhancement of selectivity to the unsaturated alcohol would come from improved activation of the C=O bond at the interface between Pt atoms and TiO, adspecies. According to the authors the electronic modification of Pt under the SMSI state would not play a prominent role. In contrast, the im- proved yield in cinnamyl alcohol from cinnamyl aldehyde over Pt/graphite (PtG) catalysts, compared to Pt/charcoal (PtC ), is thought to come from an electronic modification of Pt by graphite [ 41. An electron transfer from graph- ite to Pt would depress the interaction between the C=C double bond and the metallic surface. Moreover, in the same reaction, the nature of the metallic precursor used for preparing the catalyst has great relevance. Thus, it was found that platinum or ruthenium catalysts prepared from chloride salts are less selective to the unsaturated alcohols [ 3,5]. The selectivity in such cases is enhanced when the residual chloride ions are eliminated by reduction of the catalyst at high temperature albeit at the cost of activity. However, contrary behaviour was reported recently for the hydrogenation of cinnamaldehyde and crotonaldehyde over supported cobalt catalysts [ 6 1. The samples prepared from chloride precursors exhibited the highest selectivity to the unsaturated alcohols.

The hydrogenation of nitroaromatic compounds in terms of the effects of support and precursor is not well documented. Nevertheless, it has been re- ported that a basic support, like BaCO,, is more efficient than alumina in pre- venting dehalogenation [ 71.

The aim of the present study was to determine the influence of support and precursor on the hydrogenation of p-chloronitrobenzene over supported Pt catalysts. This work is a continuation of our previous studies on the effect of Pt particle size [ 11 and alloying [ 21 on the same reaction.

Experimental

Reactants High purity hydrogen (99.99%) was used for the catalytic experiments,

and ultra-high purity hydrogen (99.9995% ) was used for the adsorption mea-

B. Coq et al./J. Mol. Catal. 79 (1993) 253-264 255

surements. p-Chloronitrobenzene (Aldrich, purity > 99% ) and methanol, (SDS, > 99% ) were used without further purification. Platinum acetylaceton- ate (Pt (acac),, Strem), platinum tetramine dichloride (Pt ( NHB),C12, Strem) , hexachloroplatinic acid ( H2PtC16, Merck) were used as precursor for the cat- alysts. Pt (acac), was dissolved in anhydrous toluene (purity > 99.5% ), the other precursors in water. The carriers were y-alumina from RhGne-Poulenc (surface area 200 m2 gg ’ ) , TiO, from Degussa (surface area 120 m2 g- ’ ), MgO prepared by decomposition of magnesium ethylate and calcined at 573 K (sur- face area 25 m2 g-l ) and graphite HSAG300 from Lonza (surface area 300 m2 g- ’ ) functionalized by NaClO as described elsewhere [ 81.

Preparation of platinum catalysts The carriers were activated at 573 K overnight, then placed in contact

with Pt (acac) 2 in toluene solution. The Pt/A1203 samples from inorganic pre- cursors were prepared by ion exchange between H2PtC16 in acidic medium (pH 2), or Pt(NHB)&12 in basic medium (pH 9-10). After being in contact for several hours, the solution was either evaporated or filtered, dried at 373 K overnight, and then calcined at 573 K for 4 h (except for Pt/graphite). The final reduction step was performed under various conditions, as listed in Table 1.

Characterization The chemisorption of hydrogen was carried out in a conventional volu-

metric apparatus, at 298 K between 0 and 20 kPa pressure. The sample was first reactivated in situ under flowing hydrogen at 573 K overnight, then evac- uated to 1.2 x 10W4 Pa at the same temperature for 3 h. Hydrogen was then adsorbed. The linear part of the isotherm, usually between 10 and 20 kPa, was extrapolated to zero pressure to determine the hydrogen uptake. The size of

TABLE 1

Mean conditions for the preparation and characterization of the Pt/A1,03 catalysts

Catalyst

PtAlI PtAlXI PtAlXIII PtAlXIV PtMgO PtG PtTi (LTR) PtTi (HTR )

support

AM-h Al& 40, AW, MgO graphite TiOa TiO,

Precursor

Pt(acac)z 0.46 673 Pt(acac), 1.4 623 Pt(NH,),Cl, 0.30 623 H,PtCl, 0.37 623 Pt(acac), 0.26 623 Pt(acac), 0.20 573 Pt(acac)z 0.48 523 Pt(acac)z 0.48 773

wt.% Pt Reduction H/Pt Mean size temperature from TEM

(Xl (nm)

0.40 2.3 1.0 1.2 0.50 - 0.90 1.2 0.58 3.1 0.04 wide distribution 0.35 3.2 0.04 3.0

256 B. Coq et al./J. Mol. Catal. 79 (1993) 253-264

the metallic particles was checked by transmission electron microscopy using a JEOL 1OOCX apparatus. The agreement with chemisorption is good.

The chemical compositions of the solids were determined, after dissolu- tion, by elemental analysis at the Service Central d’Analyse (CNRS, Solaize, France ) .

The main characteristics of the catalysts are included in Table 1.

Catalytic tests Hydrogenation ofp-chloronitrobenzene (CNB ) was carried out in a 50 ml

four-necked flask connected to a cooler, a dropping funnel, a hydrogen gas line and a tube used to sample the reaction mixture. The catalyst was reactivated in situ under flowing hydrogen at 523 K overnight. After cooling to the reaction temperature ( T, - -303 K, unless otherwise specified), 25 ml of the reactant solution (CNB in methanol) was introduced through one arm of the flask. This solution had been previously purged with bubbling hydrogen. The reaction mixture was magnetically stirred at 700 rpm, and the reaction was carried out at atmospheric pressure under Hz flow.

Chemical analysis of the products was performed by gas chromatography (Carlo Erba, model 2450), equipped with a FID detector and a J & W capillary column (30 m x 0.32 mm i.d., DBl apolar phase). Reactants and products were identified by comparison with authentic samples, and by GC-MS coupling.

Results

Our aim was to compare the behaviour of Pt catalysts obtained from dif- ferent precursors or different supports, but having a similar size of Pt particles, since the conversion of CNB is size dependent [ 11. As far as the nature of the precursor is concerned, this goal was reached, and the Pt/Al,O, catalysts have Pt particle sizes of between 1 and 2 nm (Table 1) . It was previously reported that little influence of Pt dispersion occurs in this range [ 11. However, to ob- tain similar dispersion for Pt catalysts from the same precursor but on differ- ent supports was not as successful. PtAlI, PtMgO and PtTi(LTR) however, do exhibit comparable dispersions. Moreover, when Pt/Ti02 is reduced at high temperature (PtTi(HTR) ) the decrease in H, uptake is very likely to be due to the SMSI effect, since no clear decrease of Pt particle size occurred (Fig. 1). In contrast, the PtG sample exhibits a wide distribution of Pt particle size, and a reliable mean size was difficult to estimate, and may be meaningless.

The hydrogenation of halonitrobenzenes follows the formal reaction path- ways described elsewhere [ 2,9]. The only products appearing in detectable amounts during the reaction were chlorobenzene (CB ), aniline (AN), p-chlo- ronitrosobenzene (CNSB ), nitrobenzene (NB ), p-chlorophenylhydroxylam- ine (CPH) and azo- and azoxy-dichlorobenzenes. Actually, it seems that p- chloronitrosobenzene was formed mainly by decomposition of CPH in the GC

B. Cog et al./J. Mol. Catal. 79 (1993) 253-264 251

x

20 nm 1

particle size (wn)

Fig. 1. Distribution of Pt particle size and electron micrograph of Pt/Ti02 reduced at low tem- perature (a, PtTi ( LTR ) ) , and high temperature (b, PtTi (HTR ) ) .

258 B. Coq et al./J. Mol. Catal. 79 (1993j 253-264

apparatus as reported previously [ 2,101. Whatever the catalyst, CB, NB and azo-or azoxydichlorobenzenes were always detected in small amounts ( < 1% ).

Figure 2 shows the evolution with time of the composition of the reaction medium for typical Pt catalysts, which illustrate slightly different types of cat- alytic behaviour. The initial rates of CNB conversion were determined from these curves and reported in Table 2. They are expressed both in reaction rate and turnover frequency (TOF), or number of reactant molecules converted per second and per surface Pt atom.

The compositions of the reaction medium at CNB conversions greater than 98% are reported in Table 3 for all the catalysts. Better yields of CAN

100

25

0 50 1000 200 400

0 250 500 0 200 400 time IminI

Fig. 2. Hydrogenation of p-chloronitrobenzene over PtAlI (a), PtAlXIII (b), PtG (c) and PtTi(HTR) (d) samples as a function of time. T,=303 K; C,=O.52 mol 1-l. (0) p-chloronitro-

benzene; (0 ) p-chloroaniline; (0 ) aniline; (0 ) p-chloronitrosobenzene; traces ( < 1% ) of chlo-

robenzene, nitrobenzene, azo- and azoxy-dichlorobenzenes.

B. Coq et al./J. Mol. Catal. 79 (1993) 253-264 259

TABLE 2

Reaction rate (mol g-r s-‘) and specific activity per Pt surface atom (s-r) for the hydrogenation of p-chloronitrobenzene over supported Pt catalysts (T, = 303 K; CR = 0.52 mol 1-r )

Sample H/Pt Reaction rate Turnover frequency (mol gg’ s) X lo6 (s-l)

PtAlI 0.40 8.5 0.89 PtAlXI 1.0 12 0.17 PtAlXIII 0.50 1.6 0.21 PtAlXIV 0.90 2.0 0.12 PtMgO 0.58 15.5 2.0 PtG 0.04 2.2 4.5 PtTi(LTR) 0.35 19 2.2 PtTi (HTR) 0.04 7.0 7.1

TABLE 3

Yield of products for the hydrogenation ofp-chloronitrobenzene at high conversion ( > 98% ) over supported Pt catalysts (T,= 303 K; CR= 0.52 mol 1-r )

Sample Catalyst CNB Reaction Yield of products (mol% ) amount conversion time (mg) (%) (min) CB AN CNSB NB CAN Others”

PtAlI 100 99.4 95 - 13.2 0.07 - 85.2 0.86 PtAlXI 50 99.7 285 0.11 16.9 0.19 0.05 82.1 0.25 PtAlXIII 1000 99.8 1500 1.57 0.09 - 95.0 3.2 PtAlXIV 500 98.8 150 7.9 - 90.4 0.5 PtMgO 100 99.0 165 0.13 2.86 2.72 - 92.0 1.27 PtG 100 98.8 525 0.03 - 0.90 - 97.7 0.24 PtTi (LTR ) 200 98.5 150 - 0.57 - 97.7 0.26 PtTi (HTR) 100 99.7 370 0.03 0.35 - 99.3 -

“Azo- and azoxy-dichlorobenzenes.

were achieved with Pt/A1203 from inorganic precursors, whereas aniline for- mation was nearly totally suppressed with graphite or titania as supports. In particular, when Pt/TiOz is reduced at high temperature (SMSI state), chlo- roaniline in 99.3% yield was obtained, this improvement in selectivity was not offset by longer reaction times.

Discussion

In previous studies of the CNB hydrogenation over Pt/A1203 [ 11, and bimetallic PtM/Al,O, [ 21 catalysts obtained from organometallic precursors, we concluded that the kinetics of the reaction obeys a modified Langmuir- Hinshelwood model with competitive adsorption between CNB and hydrogen.

260 B. Coq et al/J. Mol. Catal. 79 (1993) 253-264

The competition of adsorption does not hold for high concentrations of CNB: on the Pt surface saturated by CNB, there are still Pt sites available for the chemisorption of the smaller hydrogen molecule. The following rate law (Eqn. (1) ) was then proposed which permits both the adsorption and rate constants to be determined:

(1)

where F is the reaction rate (mol gg ’ s-’ ), CR the concentration of CNB (mol l- ’ ) , PH the hydrogen pressure (atm ) , k the rate constant (mol g- ’ s- ’ ) ,A R the adsorption constant of CNB (1 mol-’ ), A, the adsorption constant of hy- drogen (atm- ’ ), j? a dimensionless number (p< 1) representing the fraction of Pt surface covered by CNB at saturation.

Actually, we found that the correcting factor j3 lies between 0.9 and 1 [ 1,2 1, thus the expression 1 &n ( 1 - p) //? is c ose to zero and eqn. (1) can be simpli- 1 tied to:

(2)

In agreement with this rate equation the sigmoid curves of Fig. 2 indicate the occurrence of three different kinetic regimes depending on the CNB con- centration: a negative order at high CNB concentration, then a zero order de- pendency and finally a positive order at the end of the CNB transformation. For PtTi (HTR) the concentration region where the negative order versus CNB applies is narrow. This could reflect competition between H, and CNB that is less than those with other catalysts. Nevertheless, the only sample for which the time history composition of the reaction medium differed was PtAlXIII; after a normal initiation of the reaction, the CNB conversion is slowed down sharply with the appearance of CAN. The logical conclusion which follows, would be strong inhibition of the reaction by CAN. But it is difficult to under- stand why such behaviour only occurred for that sample.

For the Pt/Al,O, catalysts the nature of the precursor little influences the initial specific activity expressed per Pt surface atom. Conversely, the impor- tant result of this study is the marked enhancement in TOF when platinum is supported on titania and reduced at high temperature. This unique behaviour is very likely to be due to the SMSI state for platinum. It has now been clearly established that TiO, species (2~ -c 2) can migrate over the metal surface under high temperature reducing conditions [ 11 and references therein]. It was pro- posed for Pt/TiOa [ 3,121, Ir/Nb205 [ 131 and Pt/Nb20, [ 141 reduced at high temperature, that the specific metallic sites at the borderline with these TiO, or NbO, adspecies are superactive for C=O hydrogenation. Hydrogen disso- ciates on the bare metal and sub-oxide species coordinate the oxygen atom of the C=O bond, which becomes more susceptible to hydrogen attack. A very similar interpretation was proposed for the enhancement of activity observed

B. Coq et al./J. Mol. Catal. 79 (1993) 253-264 261

for the hydrogenations of 2,4dinitrotoluene over PdFe/SiO, [ 151, p-chloro- nitrobenzene over PtSn, Zn/Al,O, [ 21, and nitrobenzene over PtSn/nylon [ 161. In these bimetallic formulations, ionic or electron deficient Fe or Sn species would be the active-centres for N=O activation, as described above for TiO, species.

We have reported previously a TOF of 4 s-l for 15 nm Pt particles sup- ported on alumina, under the same reaction conditions [ 11. Thus, the higher activity recorded over the PtG catalyst cannot be interpreted unambiguously as an influence of graphite on Pt. Indeed, the size of Pt particles ranges from 1 nm to 50 nm with a rough mean of cu. 10 nm.

The most interesting point deals with selectivity. In that reaction several undesired products can be formed like AN, CB, NB, azo- and/or azoxy-dichlo- robenzenes. The use of anionic or cationic precursors yields Pt/A1203 catalysts that are more selective to CAN than Pt/A1203 obtained from Pt (acac ),. The influence of precursor on catalysis and chemisorption is often ascribed to the presence of chlorine either during the preparation or in the final catalyst [ 3,17- 19]. Actually, PtAlXIII and PtAlXIV samples contain 0.1% and 2% Cl re- spectively, but PtAlXIII is the most selective to CAN and appears to be the most influenced by the nature of the precursor (Fig. 2b). Thus, the possible role of residual chlorine in the properties of the final catalyst is questionable. Rather, we would expect that during the nucleation of the Pt clusters, the nature of the precursor and the pH conditions influence the metal/support interaction inducing some specific particle morphologies, as proposed for Rh/ ALO [ 191.

The support modifies, more markedly than the metal precursor, the selec- tivity of the reaction. For the same reason as previously discussed for the over- all activity, the higher selectivity to CAN obtained on PtG is difficult to cor- relate directly to a support effect, owing to the wide distribution of Pt particle size. Moreover, the 0.1% Fe content of graphite could act as a promoting ad- ditive as reported for the hydrogenation of 2,4-dinitrotoluene over PdFe/SiO, [151.

In contrast, the very high CAN yield achieved on PtTi (HTR) is very likely due to the SMSI state and the migration of TiO, adspecies on Pt particles. The protection of CAN against dechlorination can be explained in terms of: (1) either the relative strength of adsorption between CNB and CAN increases when Pt enters the SMSI state. Hence, p-chloroaniline is easily desorbed from the Pt surface and dechlorination is suppressed; (2) or Pt in the SMSI state exhibits a lower activity for C-Cl bond hydrogenolysis.

In order to test these hypotheses two experiments were performed: first, the hydrodechlorination of CAN over PtAlXI and PtTi(HTR) and second, the hydrogenation of CNB in the presence of CAN on the same samples.

The effect of CAN concentration on the initial rate of CNB hydrogenation over PtAlXI and PtTi (HTR) catalysts is reported in Table 4. It appears that the inhibiting effect of CAN is slightly lower on PtTi (HTR) than on PtAlXI

262 B. Coq et al. jJ. Mol. Catal. 79 (1993) 253-264

TABLE 4

Effect ofp-chloroaniline concentration ( CCAN ) on the initial rate ofp-chloronitrobenzene hydro-

genation over Pt/Al,Oa and Pt/TiOz catalysts (‘I’,= 303 K; CR = 0.13 mol 1-i )

Sample Reaction rate (mol g-i s-l) X 10fi-

C CAN=0 CCAN = 1.3 mol 1-i

PtAlXI 18.8

PtTi (HTR) 3.3

0.70 0.33

TABLE 5

Reaction rate (mol g-i s-i) and specific activity (s-i) for the hydrodechlorination of p-chlo-

roaniline over Pt/Al,O, and Pt/Ti02 catalysts (Z’,= 303 K; CcAN = 0.13 mol 1-i )

Sample Reaction rate Turnover frequency

(mol g- 1 s-1) x 107 (s-1)

PtAlXI 39.6 0.056

PtTi (HTR ) 1.1 0.11

catalyst. A tenfold decrease of activity is observed on the latter, whereas it reaches a value of thirty on the former. In a first approximation we can con- clude that the ratio of the adsorption constants LCAN/;IR is reduced by a factor of three for Pt/TiOz in the SMSI state. In a previous study [2] we found that upon alloying Zn to Pt the ;1 c&AR ratio was reduced by a factor of ten. Thus, an increase in CAN selectivity from 82.1% to 97.2% occurred. In the present case on going from the PtAlXI to the PtTi (HTR) catalysts, the ;1 c&d n ratio decreases by a factor of three only, but the yield of CAN goes from 82.1% to 99.3%. Therefore, the decrease in the strength of CAN adsorption is not enough to account for this result. The rate of CAN hydrodechlorination over PtAlXI and PtTi(HTR) is reported in Table 5. A tenfold increase in reactivity be- tween CNB and CAN is apparent for PtTi (HTR) compared to PtAlXI. Very likely, this is the key factor responsible for the enhancement of CAN yield on PtTi (HTR). In this respect, the behaviour of PtTi (HTR) differs totally from that of PtZn/AlzOs [ 21 f or which the high CAN yield is mainly due to the lower CAN adsorption strength. As mentioned above, this high reactivity towards the N=O bond may be the characteristic feature of the mixed sites on the bor- derline between the bare platinum and the TiO, adspecies. If we take into account the specific behaviour of PtTi (HTR) we can tentatively propose the reaction site schematically shown in Fig. 3. The main feature of this site is the adsorption of CNB in part over TiO, species, which moreover accounts for both the low competition between hydrogen and CNB and the high TOF.

Furthermore, the high reactivity of TiO, species towards the N=O bond explains well the very low concentration ( < 1% ) of intermediates (CNSB and/

B. Cog et al./J. Mol. Catal. 79 (1993) 253-264 263

Pt surface /

Fig. 3. A schematic diagram of the reaction sites of SMSI-Pt/Ti02 for p-chloronitrobenzene

hydrogenation.

or CPH) during the reaction. This behaviour is similar to that reported for PtSn/Al,O, samples [ 21. After undergoing a redox cycling which gives rise to the SnO, adspecies, the concentration of the intermediates fell from 6% to 0.5%. In contrast, the reactivity of intermediates is the only point on which PtMgO differs from the PtAlI or the PtAlXI catalysts. Indeed, the concentra- tion of CPH or CNSB goes through a maximum value of 12% compared to 4% on Pt/A1203. That could be due to the basic character of magnesia.

The main conclusion is that the specificity of Pt/TiOz diminished at high temperature, SMSI state, for the selective hydrogenation ofp-chloronitroben- zene to p-chloroaniline. We propose that the suboxide TiO, species migrating on the Pt particles are responsible for promoting the Pt properties for this reaction. These adspecies polarize the N=O bond which becomes more suscep- tible to hydrogen attack. A yield of p-chloroaniline as high as 99.3% can be obtained with this dedicated catalyst.

Acknowledgements

The contribution of the Service Central d’Analyse du CNRS (Solaize, France) for chemical analyses is gratefully acknowledged.

References

B. Coq, A. Tijani and F. Figueras, J. Mol. Catal., 68 (1991) 331.

B. Coq, A. Tijani and F. Figueras, J. Mol. Catal., 71 (1991) 317.

M.A. Vannice and B. Sen, J. Catal., 115 (1988) 65. A. Giroir-Fendler, D. Richard and P. Gallezot, in M. Guisnet, J. Barrault, C. Bouchoule, D.

Duprez, C. Montassier and G. Perot (eds.), Heterogeneous Catalysis and Fine Chemicals, Poitiers, 1988, Elsevier, Amsterdam, 1988, p. 171.

A.A. Wismeijer, A.P.G. Kieboom and H. Van Bekkum, Appl. Catal., 25 (1986) 181.

Y. Nitta, Y. Hiramatsu and T. Imanaka, J. Catal., 126 (1990) 235.

264 B. Coq et al/J. Mol. Catal. 79 (1993) 253-264

8

9

10 11 12 13 14 15

16 17

18 K. Lu and B.J. Tatarchuk, J. Catal., 106 (1987) 166. 19 B. Coq, R. Dutartre, F. Figueras and T. Tazi, J. Catal., 122 (1990) 438.

P.N. Rylander, M. Kilroy and V. Coven, Engelhard Znd. Tech. Bull., 6 (1965) 11. D. Richard and P. Gallezot, in B. Delmon, P. Grange, P.A. Jacobs and G. Poncelet (eds.), Preparation of Catalysts ZV, Elsevier, Amsterdam, 1987, p. 71. W. Pascoe, in P.N. Rylander, H. Greenfield and R.L. Augustine (eds.), Catalysis of Organic Reactions, Marcel Dekker, New York, 1988, p. 121. G. Mestroni, G. Zassinovich, C. de1 Bianco and A. Camus, J. Mol. Catal., 18 (1983) 33. S.J. Tauster, Act. Chem. Res., 20 (1987) 389. B. Sen and M.A. Vannice, J. Catal., 113 (1988) 52. H. Yoshitake, K. Asakura and Y. Iwasawa, J. Chem. Sot. Faraday Trans. 1,85 (1989) 2021. H. Yoshitake and Y. Iwasawa, J. Catal., 125 (1990) 227. A. Benedetti, G. Fagherazzi, F. Pinna, G. Rampazzo, M. Selva and G. Strukul, Catal. Lett., 10 (1991) 215. S. Galvagno, A. Donato, G. Neri and R. Pietropaolo, J. Mol. Catal., 42 (1987) 379. H.F.J. van’t Blik, J.B.A.D. van Zon, T. Huizinga, J.C. Vis, D.C. Koningsberger and R. Prins, J. Am. Chem. Sot., 107 (1985) 3139.