Quick viewing(Text Mode)

Heterogenised Homogeneous Catalysts

Heterogenised Homogeneous Catalysts

Heterogenised Homogeneous Catalysts

RHODIUM CATALYSTS FOR

By Michael S. Scnrrell Instituttet for Kemiindustri, Technical University of Denmark, Lyngby, Denmark

Heterogeneous versions of homogeneous catalysts can often he produced and may 1we certain advantages in use particularly on a commercial scale. The catalytic chemistr-y of supported rhodium compounds for the carhonylation of methanol is described and the behaviour of heterogeneous and homogeneous operation compared in order to illustrate the relations between the two catulyst types. Particular attention is given to the factors which injuence the activity and the selectivity of the heterogenised form.

Recently a tremendous interest has arisen phase reactants. Our knowledge concerning concerning the behaviour of “heterogenised” these points is still fragmented, but sufficient forms of catalysts, such as transition metal work has now been reported to indicate that complexes, which were previously used in supported versions can exhibit activities and homogeneous media (I). The growth of selectivities comparable with those of their activity in this area reflects the importance homogeneous counterparts. An excellent which is being attached to the considerable illustration is provided by the use of rhodium potential offered by the new generation of catalysts in the carbonylation of methanol to catalysts. Homogeneous catalysts frequently . exhibit very high activity and selectivity but their use on a large scale is often complicated Carbonylation with Homogeneous by difficulties associated with their separation Catalysts from the reaction products and also with It is helpful to consider first the behaviour problems of corrosion. of rhodium catalysts in homogeneous media. These drawbacks can in principle be over- The components of the active material com- come by binding the catalyst to a solid prise a rhodium compound and a halogen material, thus using the active compound promoter, which is preferably iodine (6, 7). in a heterogenised form. Suitable support The latter may be supplied as methyl iodide, materials which have been used include in- hydrogen iodide, calcium iodide or iodine soluble polymers (2, 3,4) and silica (5). itself. Reaction with methanol is normally The extent to which the catalytic behaviour carried out (6, 7, 8) at 150 to 225OC and at a of such heterogeneous systems parallels that total pressure of I to IOO atmospheres. displayed by the unsupported complex and Selectivity for the production of acetic acid is the possible effects on reaction of the support typically about 99 per cent. itself are two obviously important factors Traces of dimethyl ether and possibly which must be considered in the first in- are produced, and this is stance. In addition, we must consider the particularly evident if the reaction is con- stability of the bound catalysts and examine ducted using methanol itself as solvent. their operation with both liquid and vapour Acetic acid may not be recovered in all cases

Platinum Metals Rev., 1977, 21, (3), 92-96 92 because of esterification to methyl acetate. RhCI,, Rh,O,, RhCI(CO)(PPh,), (Ph= The rate of carbonylation is found (7, 8) phenyl) and Rh(CO),CI,. Suitable solvents to be directly proportional to the concentra- are low molecular weight hydrocarbons such tions of rhodium and iodine, but is inde- as benzene, or one of the reactants, methanol, pendent of methanol concentration and acetic acid, methyl acetate or water. pressure. The suggested The catalysts have provided the basis for reaction sequence is depicted below: the development of new technologies for methanol carbonylation (7) and also for the CH,OH+HISCH,I +HaO (i) related alkene hydroformylation process (9). RhL, + CH,I+CH,Rh(I)L, Successful though the homogeneous cata- CH,Rh(I)L,+ COSCH,Rh(CO)(I)L, (iii) lysts have proved, it is clear that hetero- CH,Rh(CO)(I)L,~CH,CO-Rh(I)L, (iv) geneous analogues, if sufficiently active and selective, could offer considerable advantages CH:3CO-Rh(I)Lm+H20,C in solving both separation and corrosion RhL,+CH,COOH +HI (v! problems (10). Step (ii), the of methyl iodide to the rhodium complex is rate deter- Heterogenised Catalysts mining (7,8), with the remaining steps taking Several examples of methanol carbonyla- place much more rapidly. Although the tion by solid catalysts comprising in each case initial complex is denoted RhL,, ligands of one of a range of rhodium compounds bound different types may be present. Roth et a1 (7) to a support material have been reported have demonstrated that rates of reaction and recently. The carriers have included carbon, product distributions are very similar for a alumina, poly(styrene-divinylbenzene) and a variety of rhodium compounds, including type X molecular sieve zeolite.

Platinum Metals Rev., 1977, 21, (3) 93 The polymer supported catalyst (11) was for operation at 2ooOC (13). An active prepared by linking chlorocarbonylbis- material is produced by impregnation of the (triphenylphosphine)rhodium, IuICl(C0)- carbon with a solution of rhodium nitrate (PPhJ, to membranes or beads of poly- followed by decomposition of the latter at (styrene-divinylbenzene). The synthesis temperatures of 300°C and above. An in- method (3) is summarised in the Figure. crease in activity is seen on using tempera- Carbonylation was carried out using vapour tures in excess of 300°C and this appears to phase reactants with the membrane or by be associated with conversion of the nitrate suspending the beads in a liquid reactant to oxide. Decomposition in hydrogen rather mixture. Catalytic activity was associated than nitrogen resulted in a lower rate of with the presence of a Rh(1) complex. A reaction and from this evidence it might be steady conversion of rhodium from Rh(I) inferred that rhodium in the reduced state or to Rh(II1) during the reaction resulted in a as free metal is not required. The situation fall in activity of the membrane catalyst. The remains uncertain however, since hydrogen activity of the catalyst in liquid medium also treatment has a favourable effect on the declined with time but in this case loss of activity of catalysts obtained from rhodium rhodium from the support appeared to be trichloride. Selectivities for carbon supported responsible. A reduction in this loss might, catalysts of up to 99 per cent have been it was suggested, be obtained by using a reported (12). different polymer as support, or by increasing A number of organometallic complexes of the number of coordinating groups linking rhodium have been examined for carbonyla- the complex and support. Some dimethyl tion activity when supported on y-alumina ether was produced in side reactions with (IS). The carrier was pre-dried at 650°C and these catalysts. the complexes incorporated by impregnation The mechanism suggested by Roth et a1 of the oxide with a benzene solution of the (7) for homogeneous and supported required compound. The resulting solids by kinetic investigations (8) also affords an were active for conversion of methanol (120 explanation of the kinetics observed for the to zoo°C, vapour phase reactants, total heterogenised catalysts in both liquid and pressure I atmosphere), but selectivity for vapour phase operation. The presence of an acetate formation was in nearly every case iodine based promoter is again found to be rather low (less than 50 per cent) because of essential. One difference concerns the ether production, This was true of catalysts evidence that two adjacent rhodium centres based on, for example, RhCI(PPh,),, are involved in the oxidative addition of (codRhCI), and (codRhOCH,), (cod= methyl iodide for the heterogeneous catalyst cycloocta-1,5-diene). On the other hand rather than one as shown in the reaction those derived from RhCI(CO)(PPh,), had sequence for homogeneous reaction. about the same overall activity but were much Carbon has also been found to provide a more selective (approaching gg per cent). useful support (12, 13, 14). Here too a close As with the work using carbon the exact similarity in behaviour of heterogeneous and state of the rhodium complex in an active homogeneous catalysts is observed (14). catalyst remains obscure. Because different Rhodium trichloride may be used as a starting materials resulted in catalysts having starting material (14) but rhodium nitrate is different selectivities there is some reason to preferred (12, 13). The nitrate based cata- suppose that the molecular identity of the lysts are more active than those prepared complex is retained, at least to some extent. from such complexes as Rh(acac),, The rather large amounts of ether produced RhCI(CO)(PPh,), and RhCI(PPh3),, where by most of these catalysts may very well be acac is acetylacetonate, by factors of up to 10 due to the use of alumina as the support,

Platinum Metals Rev., 1977, 21, (3) 94 I I Comparison of Catalyst Activities for Methanol Carbonylation Rates of reaction at 250"C,with po=1 atmosphere and molar ratio methano1:rnethyl iodide=lO. All reactants in the vapour phase for heterogenised catalysts.

Rhodium content of Rate, expressed as Catalyst catalyst, in weight g methyl acetate/g Rh/h Reference per cent

RhCI(CO)(PPh,),-alumina 1.39 40 15 Rh (N 0,),-carbon 3 25 12 RhCI,-N aX 0.25 50 17 Rh CI ,-NaX 1.0 10 17 R hCI(CO)(P,),* 3.6 %** 11 RhCI, homogeneous catalysis 3 x103 8

*Pp=p-(polystyryl)diphenylphosphine ligand **Molar ratio methano1:methyl iodide typically>l5

particularly since the oxide was dried at high ing lower or higher rhodium contents than temperature. Such treatment results in high this displayed a smaller rate per unit weight activity for the dehydration of methanol to of metal. dimethyl ether (16). The reason for the low Like alumina, zeolites are expected to activity of the catalyst based on RhCl(C0)- exhibit a tendency to produce dimethyl ether (PPh,), for ether formation is not clear. from the elimination reactions of methanol or The use of a molecular sieve zeolite as a methyl iodide (IS), but unlike those based on support is of particular interest because of the the former oxide, the zeolite carbonylation influence which these materials have had in catalysts had reasonably high selectivities. the design of conventional heterogeneous Again, the reason may lie in the treatment of catalysts. the support before use. The zeolite was An active carbonylation catalyst has recently apparently not subjected to high temperature been prepared by the impregnation of a drying and, in any event, impregnation with sodium-X zeolite with rhodium trichloride the rhodium salt took place in aqueous solution (17). As with all the other catalysts solution. the presence of iodine is necessary. Selec- These observations naturally lead one to tivity for acetate production approached go suspect that support pretreatment procedure per cent, the main side product being may have a very great influence on the dimethyl ether. On decreasing the rhodium catalytic behaviour of the heterogenised content of the catalyst from I weight per cent systems, particularly when the carriers are to 0.05 weight per cent there resulted a steady potentially active materials such as alumina, increase in the rate of reaction per unit zeolites and silica. weight of rhodium, by a factor of about 20 at 25ooC, but selectivity for acetate formation Relative Activities of Heterogenised declined a little. This contrasts with the Catalysts situation found for rhodium nitrate on A comparison of the activities of various carbon where maximum effective use of the heterogenised catalysts for methanol car- rhodium was obtained at a rhodium level of bonylation, given in the Table, reveals that about z weight per cent (13). Catalysts hav- the rate per unit weight of rhodium is not

Platinum Metals Rev., 1977, 21, (3) 95 markedly dependent upon the choice of References support or rhodium compound used. Similar I J. C. Bailar, Jr., Card. Rev.-Sci. Engng., 1974, 1% 17 conclusions have been drawn before (11, 13). 2 Z. M. Michalska and D. E. Webster, Platinzim Me have however seen that the selectivity for Metals Rev., 1974, 18, (2), 65 acetate production is greatly influenced by 3 W. 0. Haag and D. D. Whitehurst, Proc. 5th Int. Congress on Catalysis, Miami Beach the nature of the starting materials. Further- 1972, publ. by North Holland Publ. Co., more, there appear to be differences in the Amsterdam, 1973, I, 465 4 C. U. Pittman, Jr., L. R. Smith and S. E. stability of the supported catalysts. Although Jacobson, “Catalysis, Heterogeneous and studies of the deactivation characteristics of Homogeneous”, Elsevier, Amsterdam, 1975, these materials have been extensive, the P* 393 not 5 R. D. Hancock, I. V. Howell, R. C. Pitkethly loss of activity with use seems to be especially and P. J. Robinson, “Catalysis, Heterogeneous noticeable with the polymer based systems. and Homogeneous”, Elsevier, Amsterdam, 19759 P. 361 However, the use of these materials as cata- 6 F. E. Paulik and J. F. Roth, Chem. Commun., lyst supports is still a very recent development 1968, (24, 1578 7 J. F. Roth, J. H. Craddock, A. Hershman and and significant improvements may well result F. E. Paulik, Chem. TechnoZ., Oct. 1971, 600 from further work. 8 J. Hjortkjaer and V. W. Jensen, Id.Engng. Insufficient data are available at present to Chem., Prod. Res. Dev., 1976,15, (I), 46 9 F. J. Smith, Platinum Metals Rev., 1975, 19, enable the calculation of activities on a rate (3), 93 per unit active site basis (turnover number). 10 R. G. Schultz, German Offen.I, 939,286; 1970 This is due to the difficulty, typically asso- 11 M. S. Jarrell and B. C. Gatcs, J. Catalysis, 1975%40, (21,255 ciated with heterogeneous catalysts, of know- 12 R. G. Schultz and P. D. Montgomery, J. ing just how many active centres are present Catalysis, 1969, 13, (I), 105 in a given quantity of material under working 13 R. G. Schultz and P. D. Montgomery, Am. Chem. SOC.,Div. Petrol. Chem., Preprints, conditions. However, if the heterogeneous 1972917, BI3 forms are compared with the homogeneous 14 K. K. Robinson, A. Hershman, J. H. Craddock and J. F. Roth,J. Catalysis, I972,27, (3), 389 catalyst (8) it can be seen in the Table that in 15 A. Krzywicki and G. Pannetier, Bull. Sot. order for the turnover numbers for the two Chim. Fr., 1975, (5-6), 1093 catalyst types to be equal, it would be neces- 16 H. Knozinger, Adv. Catalysis, 1976,25, 184 17 B. K. Nefedov, N. S. Sergeeva, T. V. Zueva, sary to assume that only about I per cent of E. M. Shutkina and Ya.T. Eidus, Izv. Akad. the total number of rhodium atoms in the Nauk S.S.S.R., Ser. Khim., 1976,(3), 582 heterogenised material is active. Such a 18 H. Noller and W. Kladnig, Catal. Rev.-Sci. Engng., 1976,~~149 figure does not appear unreasonable. 19 G. C. Bond, Platinum Metals Rev., 1975, Until active site concentrations are estab- 19, (41, 126 lished for these supported catalysts a more rigorous comparison of activities is not possible. Nevertheless the outstanding simi- larity between homogeneous and hetero- Acknowledgements geneous analogues for methanol carbonylation The paper “Forming Fibres from Basalt Rock”, which appeared in the January ‘977, *I> (‘1, has been demonstrated. This State of affairs issue of this Journal reviewed results of the is very encouraging for we are now one step project on basalt fibres conducted at the Materials Science and Engineering Department of Washing- nearer in achieving the often mentioned goal ton State University under a grant from the of basing future catalyst design on molecular Pacific Northwest Regional Commission with which the author was associated. Detailed results or enzymic (‘9). In view Of the are contained in the final report to the Corn- many examples of homogeneous catalysis by mission “Use of Basalt Rock for the Production of Mineral Fiber, Contract No. NR-3001 (1975)”~ complexes Of the it is by R. V. Subramanian, H. F. Austin, R. A. V. Raff, likelv that a considerable number of hetero- G. L. Sheldon, R. T. Dailey and D. Wullenwaber. Omission of specific -acknowledgements in genised systems be based On the Same the paper to these sources of data and finance materials. support is regretted.

Platinum Metals Rev., 1977, 21, (3) 96