10 Application of Molecular Sieve Zeolites to Catalysis By H. F. LEACH Department of Chemistry University of Edinburgh In the past decade the subject of catalysis has been discussed in Annual Reports on three occasions.' In 1964 and 1968 aspects of catalytic reactions over metals were considered and in 1966 the topic was homogeneous catalysis by complexes of Group VII elements. The present report will review some of the work that has been published concerning the role that crystalline molecular sieve zeolites have played in the catalytic reactions of hydrocarbons. This area of research has undoubtedly been one of the major growth points in catalysis during the last decade. The application of zeolites as catalysts for various processes in the petroleum refining industry was discovered during the 1 9 5 9 4 2 period soon after the Union Carbide Corporation began to produce synthetic Linde molecular sieves on a large commercial scale.The upsurge in publication of papers since that period emphasizes the importance of these materials. At the 1st international conference on zeolites in London in 1967, Milton estimated2 that over 60% of installed catalytic cracking capacity in the United States employed zeolite-based catalysts and in 1970 Arey3 suggested that the figure had risen to approximately 90% for fluid catalytic cracking units. Weisz4 has compared the industrial impact of the zeolite catalyst to that of the Haber ammonia and Ziegler-Natta polymerization catalysts that produced Nobel prizes for catalytic science and Weisz quotes the annual consumption of zeolite-containing catalysts for petroleum cracking as > 100 OOO tom4 The catalytic application of zeolites is a topic that has been well served by review articlq in the past few year^.^-^ At the 2nd international conference on zeolites in Worcester U.S.A.in 1970 there were two authoritative invited papers by Venuto" and by Rabo and Poutsma." The former paper classified I G. C. Bond Ann. Reports 1964 61 99; 1966,63,27; Ann. Reports ( A ) 1968,65 121. R. M. Milton in 'Molecular Sieves' Society of Chemical Industry London 1968, p. 199. W. F. Arey jun. Oil and Gas J . 1970 68 64. P. B. Weisz Ann. Rev. Phys. Chem. 1970 21 175 Kh. M. Minachev V. I. Garanin and Ya. I. Isakov Russ. Chem. Rev. 1966 35 903. J. Turkevich Catalysis Rev.1968 1 1 . P. B. Venuto and P. S. Landis Adv. Catalysis 1968 18 259. F. G. Ciapetta Chimica e Industria 1969 51 1173. Kh. M. Minachev Kinetika i Kataliz 1970 11 413. l o P. B. Venuto Adv. Chern. Ser. 1971 No. 102 260. " J. A. Rabo and M. L. Poutsma Ado. Chem. Ser. 1971 No. 102 284 196 H . F. Leach many of the organic reactions that had been observed over zeolite catalysts, whereas the latter paper considered in more detail the structural and mechanistic aspects of zeolite catalysts with particular reference to the cracking of cumene and hexane. In the present Report some of the important arguments arising out of these articles will be discussed and attention drawn to a number of important papers that have appeared in the past two years. 1 Structure of Zeolites The catalytic properties of zeolites are very closely related to their unique crystal structures.These structures provide the following features (i) a three dimen-sional lattice containing uniform pores of molecular dimensions (ii) a high surface area which is only accessible to molecules able to diffuse through the porous network and (iii) excellent thermal stability at relatively high tempera-tures. In view of this correlation between structure and catalytic properties it is necessary to discuss briefly some of the more important structural details of zeolites. The origin of the events leading to the discovery of zeolite catalysts lies in inorganic chemistry and crystallography rather than in catalysis itself. Much of the early work on the structure of naturally-occurring zeolitic minerals and the synthesis of zeolite materials was done by Barrer and his colleagues.The articles by Barrer,” and also by Breck,I3 summarize much of the information concerning the classification and detailed structure of these materials. In order to analyse catalytic results it is essential to be aware of the zeolite framework structure and also to obtain as much information as possible concerning the location of cations and water molecules within the zeolite lattice. X Y and Faujasite.-The X and Y zeolites are the most widely used in catalytic studies (in various modified forms). Both are synthetic forms of an alumino-silicate framework which has the same topology as the naturally-occurring mineral faujasite. Turkevich6 has described in detail the structure of these materials.The primary structural unit in the faujasite structure is the so-called sodalite cage or P-cage which is a truncated octahedron containing 24 silicon (or aluminium) tetrahedra. It is composed of six four-membered rings and eight six-membered rings (rings of four or six oxygens respectively joined via silicon or aluminium) and has a free diameter in the cavity of approximately 0.66 nm. The pore opening of a four-membered ring is too small to be important in catalytic terms but that of a six-membered ring is 0.22 nm which is sufficient to permit the entry of some molecules. The three-dimensional open porous structure of the faujasite is derived by the joining of these sodalite cages in a tetrahedral arrangement through hexagonal prisms with each such prism being joined to a six-ring face of a sodalite cage.Thus a hexagonal prism will be composed of two parallel six-ring faces and six four-membered sides. ’’ R. M. Barrer Endeavour 1964 23 122. l 3 D. W. Breck J . Chem. Educ. 1964 41 678 Application of Molecular Sieve Zeolites to Catalysis 197 The joining of the sodalite cages in this manner produces the characteristic faujasite cage (often referred to as the a-cage or supercage) which is the major cavity in the zeolite structure. The cage is a 26-hedron made up of eighteen four-membered rings four six-membered rings and four twelve-membered rings. These latter rings are arranged as the four sides of a tetrahedron and afford the most important entry points into the cavity.The free diameter within the supercage is approximately 1.30nm but the diameter of the pore opening is 0 . 8 0 . 9 0 n m which is sufficient to permit entry of aromatic compounds and some branched-chain hydrocarbons. The capacity of the supercage for a variety of molecules has been listed6 and ranges from 28 molecules of water to 2.1 mole-cules of perfluorodimethylcyclohexane. MeierI4 classified the structural frameworks of faujasite and many of the other naturally-occurring zeolites in great detail and more recently’ stereograms of these structures have appeared. In the latter paper the skeletal framework drawings were based on the T-atoms (Si or Al) with the T-0-T bridges represented by straight lines. The drawings were constructed as stereopairs and the stereoscopic representation was arranged so that the selected viewing direc-tion clearly indicated the main channels in the particular zeolite class.Smith16 has tabulated the T-0 distances and T-0-T angles for a variety of hydrated and dehydrated ionic forms of the faujasite-type structure. The location of the cations in X and Y zeolites has been the subject of intense activity and speculation. In the mid-60’s three main types of site were generally recog-nized i.e. type S located in the hexagonal prisms between the sodalite units, type S, located in the open six-membered faces of the sodalite unit and type S,, located on the walls of the cavity. As single-crystal samples became available, and X-ray diffraction data more precise the number of defined cation sites increased and for a time there was a degree of confusion about the nomenclature.However the nomenclature adopted by Smith (illustrated in Figures 1 and 2 of reference 16) has now been adopted by the majority of workers in the field. It is as follows site I situated in the centre of the hexagonal prism ; site I’ on a triad axis displaced into the sodalite cage from the hexagonal face shared by the sodalite cage and the hexagonal prism ; site 11’ on a triad axis displaced into the sodalite cage from the open (unshared) six-membered ring of the sodalite unit ; site I1 and site 11* displaced (slightly and considerably respectively) from this ring into the supercage; site I11 displaced into the supercage from bridging four-membered rings ; and site V very nearly at the centre of the twelve-membered rings separating the supercages.These well-defined locations are the positions that the cations required to compensate for the excess negative charge arising out of the silicon-oxygen-aluminium skeleton will occupy. When first synthesized zeolites are usually in the sodium form and there will be an equivalent number of univalent sodium cations to ‘framework’ aluminium to establish electrical neutrality within the l 4 W. M. Meier in ‘Molecular Sieves’ Society of Chemical Industry London 1968 p. 10. l 5 W. M. Meier and D. H. Olson Ado. Chem. Ser. 1971 No. 101 155. l 6 J . V. Smith A&. Chem. Ser. 1971 No. 101 171 198 H . F. Leach material. It is a well-known feature of zeolites that they undergo ion exchange, with the sodium cations being replaced by a variety of other cationic species.As will be discussed later such cation exchange can have a profound effect upon the catalytic activity of the zeolite species. Rees17 has given a very comprehensive report on the phenomenon of ion exchange in zeolites both from the viewpoint of equilibrium and kinetic studies. Barrer and R e e ~ ’ ~ * ~ ~ and Sherry,20 have very thoroughly investigated the ion-exchange equilibria in X and Y zeolites. A recent article by Sherry” is another major contribution to this topic. The location of the exchanged cations will be of vital importance in catalytic studies. Table 1 in reference 17 gives an excellent summary of the position regarding the site occupancy of a variety of cation-exchanged faujasites both in the hydrated and dehydrated forms.The bulk of this data was acquired by X-ray diffraction techniques with the work of Bennett and Smith,22 and of Olson and colleagues23 being particularly prominent. A recent paperz4 has added to the knowledge of the location of the cations in highly-exchanged cerium-)< zeolites. It has been reported that in the hydrated sample the cerium cations were distributed between sites I1 and 1’ and on dehydration in air at 813 K cation migration to I and I’ sites took place with a site occupancy of 0.74 quoted for I’ after dehydration. It should be remembered that the X-ray diffraction techniques actually yield maps of electron density which are subjectively analyzed. With the complex situation present in cation-exchanged zeolites the quoted precision of the data regarding the cation distributions should always be treated with caution.However there can be little doubt in general concerning the different distributions that have been observed (dependent upon the nature of the cation) or the manner in which such distributions can be altered on dehydration. As previously mentioned both the X and Y synthetic zeolite forms are similar to that of faujasite. Their main difference lies in the relative Si to A1 content. In the X-type material the composition can be represented by the general formula Nag6[(A102)8,(Si02),06],264H20 with a Si A1 ratio of about 1.2,whereas the Y-type material has the general formula of Na56[(A102)56(Si02)136],264H20 with a Si:Al ratio of about 2.4 similar to that of naturally-occurring faujasite.A necessary consequence of the higher aluminium content of the X zeolites, relative to the Y-type is that they have an appreciably higher cation density. Mordenite Erionite and 0ffretite.-A considerable amount of work has been published concerning the catalytic properties of modified forms of mordenite. 1 7 18 1 9 2 0 2 1 2 2 23 2 4 L. V. C. Rees Ann. Reports(A) 1970 67 191. R. M. Barrer J. A. Davies and L. V. C. Rees J. Inorg. Nuclear Chem. 1968 30, 3333; l969,31 2599. R. M. Barrer L. V. C. Rees and M. Shamsuzzoha J. Inorg. Nuclear Chem. 1966 28, 629. H. S. Sherry J . Phys. Chem. 1966 70 1158. H. S. Sherry Adu. Chem. Ser. 1971 No. 101 350. J. M. Bennett and J. V. Smith Materials Res. Bull. 1969 4 343 and earlier papers.D. H. Olson J. Phys. Chem. 1970,74 2758; and earlier papers. F. D. Hunter and J. Scherzer J. Catalysis 1971 20 246 Application of Molecular Sieve Zeolites to Catalysis 199 Some recent interest has also been centred on synthetic zeolites related to the naturally-occurring erionite and offretite minerals. Consequently it is necessary to consider briefly these structures which are appreciably different from that of fauj asi te. The structure of the fully hydrated form of mordenite was originally determined by Meier.25 It is one of the most silica-rich zeolites known with a Si A1 ratio of ca. 5. The idealized unit cell composition corresponds to a general formula of Na,[(A102)8(Si02),,],24H20. The structure contains none of the cubic sym-metry of the faujasite structure and is in fact orthorhombic.A particular feature is the importance and predominance of five-membered rings and the mordenite framework can be constructed from a chain-like sequence of tetrahedra which can be cross-linked to other identical chains. Such chains run parallel to the c axis of the unit cell and enclose wide parallel near-elliptical channels (with free diameter 0.59 x 0.70nm) which are circumscribed by puckered twelve-membered rings. These main channels in the mordenite are linked by additional pockets which can be entered through an eight-membered ring. These orifices have a free diameter of about 0.28nm so they can accommodate a limited number of molecules (smaller than n-butane). There are however effectively no interconnecting channels in mordenite as these side channels have restrictions which essentially prohibit transfer of molecules from one main channel to another.The pore structure of mordenite does not therefore have the three-dimensional character of the X and Y zeolites. Another difference which can have important catalytic consequences is that with highly-siliceous zeolites such as mordenite all the exchangable cations can be directly replaced by H+ ions by treatment with dilute acid without destroying the lattice framework. Some of the cations have been located in the side pockets and others are considered to be in the main channels of the pore system. However the cation distribution in mordenite has not been determined in as detailed a manner as for the faujasite materials. Offretite and erionite are both members of the chabazite group of naturally-occurring zeolites and are hexagonal.There has been considerable confusion in the literature concerning their structure and they were originally considered identical. Bennett and Gard,26 however showed that the ‘c’ spacing in offretite was half that in erionite. Both species can be regarded as layer structures with the primary building units being hexagonal prisms and single hexagons but they differ in their stacking sequence. In erionite layer A consists of six-membered rings with layer B being hexagonal prisms. The next layer of six-membered rings layer C is rotated through 60” relative to layer A and the resultant stacking sequence is ABCBABCB. In offretite all the layers of six-membered rings are superimposed and therefore identical.Consequently the stacking sequence for offretite along the c direction is ABAB. A paper by Whyte et aL2’ illustrates 2 5 2 6 J. M. Bennett and J. A. Gard Nature 1967 214 1005. ” T. E. Whyte jun. E. L. Wu G. T. Kerr and P. B. Venuto J. Catalysis 1971 20 88. W. M. Meier 2. Krisr. 1961 115 439 200 H. I;. Leach clearly the manner in which large cavities are built up by such stacking arrange-ments in both erionite and offretite. A consequence of the difference in the structures is that offretite is considerably more open than erionite and contains channels with a free diameter of 0.63 nm. The absorptive capacity (a property closely-associated with catalytic activity) of offretite will thus be considerably higher than that of erionite although only a small degree of stacking disorder will considerably constrict the channel widths in offretite.Gard and Tait28 reported that there were two different types of cation site located on the axes parallel to c of the single six-membered rings and the large channels which were readily ion exchanged. Electron microscopy, together with X-ray diffraction result^,^ has provided further information con-cerning the cation distribution in unfaulted offretite and has further indicated the effect upon the adsorption properties of the size of the cation in the main channel. Synthetic preparations of both offretite and erionite have been reported and in several instances their catalytic properties examined. Some of the recent types of synthetic zeolites have been reported3' to have similar structures to those just discussed.Zeolite T has been shown to have a similar structure to offretite,28 and the structure of zeolite L3 is based on similar polyhedral cages to those found in erionite and offretite. Such new zeolitic species are likely to have useful and interesting catalytic properties but as yet only a very limited amount of published work has appeared on such aspects of these materials. 2 Catalytic Activity of Zeolites Undoubtedly the major application of zeolites in catalytic fields has been in the area of hydrocarbon cracking. Another growing use is in hydrocracking processes, which give the refiner added flexibility in oils which can be processed and products which can be made. However zeolites have also been used as catalysts in a great number of other very important hydrocarbon reactions.Venuto and Landis' have surveyed the very prolific research literature of the 1960-68 period and have clearly demonstrated the diversity of organic reactions studied. More recently V e n ~ t o ~ ~ has listed some 50 important organic reactions that workers in the Mobil laboratories have shown to be catalytically affected by zeolites to a greater or lesser degree. In the limited space available here attention will be focussed on some of the more important industrial reactions such as cracking alkylation dealkylation isomerization and oxidation with particular attention being paid to recent publications in these areas. Cracking Reactions.-Since the discovery3 that the Na-X zeolite exhibited an activity for hydrocarbon cracking similar to that of the most active silica-alumina J.A. Gard and J . M. Tait Adv. Chem. Ser. 1971 No. 101 230. 2 9 R. Aiello R. M. Barrer J . A. Davies and I. S. Kerr Trans. Faraday SOC. 1970,66 1610. 30 D . W. Breck Adv. Chem. Ser. 1971 No. 101 1 . 3 1 R. M. Barrer and H. Villiger 2. Krisf. 1969 128 352. 3 2 P. B. Venuto Chem. Tech. 1971 1 215. 3 3 P. B. Weisz and V. J. Frilette J . Phys. Chem. 1960 64 382 Application of Molecular Sieve Zeolites to Catalysis 20 1 catalyst there has been a vast amount of published work on this subject. The product pattern from Na-X was similar to that of the radical-controlled mech-anism and not to that obtained on the typical acidic catalyst. However when it was established that further enhancement of the cracking activity was exhibited by Ca-X and with a typical acidic cracking pattern of products a whole range of unusually acidic catalysts was rapidly developed.There is now very little doubt that in this particular type of reaction the operation of the zeolite catalyst is best explained in terms of the formation of carbonium ion intermediates at acid sites on the internal surface. The mechanism of hydrocarbon cracking is assumed to involve the initial formation of a carbonium ion (C,H,,- on the acid site after which each transformation event regenerates a carbonium entity. With alkane cracking this entity could be formed either from the addition to the hydrogen (proton) of a Bronsted acid site of some olefinic thermally-produced species or from the action of Lewis acid sites where the vacant electron pair orbitals would be satisfied by the abstraction of a hydride ion i.e.hydrogen atom plus electron from the alkane species. Even now it is experimentally difficult to differentiate between these two alternatives. From the large amount of work reported on the study of acidity in zeolite cracking catalysts the Bronsted acidity would appear to be the important controlling factor. However there are numerous examples where the importance of the Lewis type of acidity has also been demonstrated. Much of the early work relating to the carbonium ion activity has been summarized by Rabo and Poutsma." The work has indicated that (i) the locus of activity is on the intracrystalline zeolite surface as demonstrated by the molecular sieve phenomenon (ii) bi- and multi-valent cations induce high carboniogenic activity in X and Y zeolites (iii) the decationized Y obtained through heat treatment of NH,-exchanged Y had enhanced activity relative to the cation-exchanged Y zeolites (iu) comparable forms of Y are more active and in certain forms more stable than X-type zeolites and (v) loading of small amounts of noble metal by cation exchange greatly enhances the catalytic activity.Perhaps the most successful cracking catalysts have been those using the rare-earth faujasite-type exchanged zeolites. These materials have a particularly high thermal and hydrothermal stability are not subject to poisoning by metal impurities in the raw material and have a high selectivity.Their success can in part be attributed to the peculiarity in reaction selectivity they exhibit rather than to the capability of attaining higher reaction velocities ; the ratio of medium molecular weight (liquid) range to gaseous range hydrocarbon products is very favourable. In this connection there has recently been a marked increase in the industrial importance of rare-earth elements and it has indeed been reported3, that the major employment of rare-earth elements is now in cracking catalysts. Hirschler3 proposed that the source of carboniogenic activity in acidic zeolite catalysts was the hydroxyl protons and that the exchanged cations influenced 3 4 R. L. Koffler Proc. 7th Rare Earth Research Conference 1968 vol. 2 p. 697 3 5 A. E. Hirschler J .Catalysis 1963 2 428 202 H. F. Leach the geometry and acidity of such protonic sites. P l a n ~ k ~ ~ also suggested the role of the hydroxyl proton in the carboniogenic activity and postulated that hydrolysis of the cation produced the protons : Re3+ + H,O [ReOH]*+ + H + Consequent to this Ward3' showed that the ionic strength of the cation would control the amount of hydroxyl introduced either through hydrolysis during cation exchange or by ionization of water upon activation. Therefore if the catalytic activity is dependent upon hydroxyl concentration then it will parallel the cation strength there are various reports of such a correlation. Ward37 observed that the catalytic cracking of cumene by decationized Y was accom-panied by major changes in the i.r.spectrum of the catalyst particularly in the absorption bands associated with the hydroxy-groups. The literature concerning i.r. spectroscopic studies of zeolites has recently been reviewed,38 and this technique has been used extensively to examine the nature and location of hydroxy-groups. The acidity of cation-exchanged zeolites has also been examined using basic molecules such as ammonia pyridine and piperidine as probes. These molecules have the property that their interaction with Bronsted acid sites, with Lewis acid sites with cations and their hydrogen-bonding interactions give rise to different species detectable by i.r. spectroscopy. The series of papers by Ward on the nature of active sites on zeolites is an excellent example of the application of i.r.spectroscopy. In the latest of these39 it has been reported that the transition-metal-Y zeolites display only protonic acidity with no detectable Lewis acidity after calcination at 753 K. One of the standard test reactions for the characterization of cracking catalysts is the cracking (or dealkylation) of cumene and Rabo and Poutsma" have summarized many of the early papers concerned with this reaction. It is of the Friedel-Crafts type and is normally rationalized in terms of proton attack at an aromatic carbon atom with displacement of the side chain as a carbonium Apart from kinetic data for the cumene cracking differing product distribu-tions are also observed e.g. after calcination at 823 K La-Y was more active than Ca-Y and the C3 fraction over La-Y contained a higher proportion of propane.Eberly and Kimberlin4' have examined the cumene-cracking activity of a series of single-component rare-earth Y-type zeolites and correlated their findings with i.r. data. They noted an hydroxyl vibration in the range 347G3520 cm- ', characteristic of the rare earth and which increased linearly with ionic radius. From pyridine adsorption effects it was concluded that the Bronsted acidity also increased slightly with ionic radius although it was affected more by the calcina-tion conditions. Lower calcination temperatures produced greater acidity and 36 C. J. Planck in 'Proceedings ofthe 3rd International Congress on Catalysis,' Amsterdam, 3 7 J. W. Ward J. Catalysis 1968 10 34; 11 238 259. 3 8 J. W. Ward AdiI. Chem. Ser.1971 No. 101 380. 3 9 J. W. Ward J. Catalysis 1971 22 237. 40 P. E. Eberly jun. and C. N. Kimberlin jun. Adv. Chem. Ser. 1971 No. 102 374. 1964 p. 727 Application of’ Molecular Sieve Zeolites to Catalysis 203 also more-active cumene-cracking catalysts. However they stated that changes in Bronsted acidity could not explain all the nuances of catalytic activity observed. in a study of the cumene-cracking activity of hydrogen-, calcium- and lanthanum-exchanged X and Y zeolites reported that both the number and strength of acid sites in the zeolites increased with increasing extent of exchange. The catalytic activity could be correlated more readily with acid strength than with total acidity. Turkevich and Ono4 examined the cumene-cracking activity of a Y zeolite that had been partially exchanged with NH;, and also contained 2-3.5 % by weight of Pd.The catalyst was active up to a pretreatment temperature of 773 K but a sharp drop in activity was observed at greater pretreatment temperatures. From the effects of quinoline poisoning (titration) they concluded that the cumene-cracking activity was due entirely to Bronsted acid sites. In the same paper the cracking of 2,3-dimethylbutane was also examined and it appeared that a small number of Lewis acid sites were necessary to initiate this reaction. A study43 on the quinoline poisoning of cumene-cracking activity for Ce-Y, Ca-Y and NH,-Y zeolites has shown that a minimum dose of approximately one quinoline molecule per supercage was sufficient and that there was a cor-relation between quinoline adsorption and poisoning.The effects were considered to arise mainly from blockage of the supercages and to some extent the inter-pretation was at variance with the view of Turkevich that quinoline was a specific poison for Bronsted acid sites. Earlier studies had found4 that the amount of quinoline required to poison the cumene-cracking activity of an Ag-Y zeolite correlated reasonably well with the expected number of surface hydroxyl species. The cumene-cracking activity of a series of exchanged Y zeolites has been des-~ribed,,~ and the relative activity found was La-Y > NH,-Y > Ca-Y > Na-Y. The extent of conversion increased with increasing cation exchange and the major products below 773K were benzene and propylene. Also examined was the effect of altering the SiOz :A1,0 ratio.With the Ca-Y zeolite the cracking activity appeared to pass through a maximum as the SiO,:Al,O ratio was increased (the maximum occurring at about 4.6) whereas with the NH,-Y a steady increase was observed. Topchieva et a1.46-50 have also examined the role of cations in Y zeolites, and the SiO :A1203 ratio in respect of cumene-cracking activity. They observed the promoting effect of surface hydration and suggested that the active centres for cracking were surface hydroxy-groups where the H atom is protonized. Otouma et 4 1 H. Otouma Y . Arai and H. Ukihashi Bull. Chem. Soc. Japan 1969 42 2449. 42 J . Turkevich and Y. Ono Adu. Chem. Ser. 1971 No. 102 315. 43 M. S. Goldstein and T. R. Morgan J . Catalysis 1970 16 232.4 4 J. T. Richardson J. Catalysis 1967 9 182. 4 5 K. Tsutsumi and H. Takahashi Seisan-Kenkyu 1969,21,455 457. 46 K. V. Topchieva and Ho Chi Thanh Neftekhimiya 1970 10 525. 4 7 K. V. Topchieva and Ho Chi Thanh Doklady Akad. Nauk S.S.S.R. 1970,193 641. 4 8 K. V. Topchieva and Ho Chi Thanh Kinetika i Kataliz 1970 11 490. 4 9 K. V. Topchieva and E. N. Rosolovskaya Zhur.fiz. Khim. 1970 44 870. 5 0 K. V. Topchieva and Cho Shi Thuong Doklady Akad. Nauk S.S.S.R. 1971,198 141 204 H . F. Leach They rationalized the observation that the cation-exchanged zeolites were more active than the decationized forms by suggesting that the incorporation of multivalent cations would stabilize the hydroxy-groups against thermal deactiva-tion. The thermal stability of the active centres was higher for zeolites with large SiO :A1203 ratios and it was reported that dealumination increased the cracking activity.No simple direct relationship between aluminium content and catalytic activity was noted but maximum activity corresponded to the situation when about 50% of the aluminium had been removed from the zeolite lattice. The fact that La-Y was more active than Ca-Y was attributed to differences in the cation locations of the respective zeolites. It has been reported" that the Ca2+ ions exhibit a preference for the hidden type.1 sites in the hexagonal prisms, whereas the tervalent rare-earth ions have a tendency to occupy positions in the sodalite cage. The advantage of a high silica content in the zeolite is one of the prime reasons for the relatively recent upsurge in the catalytic applications of synthetic mordenite.Burbidge et a/.51 have indicated the growing role that mordenite-type zeolites are playing in the petroleum industry as catalysts (and as adsorbents). In general mordenite catalysts display a high initial activity but activity main-tenance is often relatively poor. By varying the synthesis conditions mordenites can be prepared with differing structural and adsorptive properties. Such mordenites have been classified52 as 'large-port' or 'small-port' depending on their ability to adsorb large molecules such as benzene and cyclohexane. Further varieties of mordenite with differing adsorptive properties can be produced by removing aluminium from the mordenite lattice structure by strong acid treat-ment e.g.with HCl.53 Eberly and K i m b e r l i ~ ~ ~ ~ have compared the cumene-cracking activity of an H-mordenite sample with a conventional Si02:Ai203 ratio of 12 with that of a highly aluminium-deficient mordenite (SO2 :A1203 ratio of 64). The activity ( A ) of both catalysts decreased with time at temperatures above 500 K according to the relationship A = at" where n was approximately -0.5. The aluminium-deficient mordenite was considerably more active and this was mainly attributed to the larger adsorption capacity and the greatly decreased resistance to adsorp-tive diffusion. A paper by Weiss and co-workersS5 reports further studies on cumene cracking over aluminium-deficient large-port mordenites. In this case the alumina content (on an anhydrous basis) was reduced in stages from 11.2% (corresponding to the conventional H-mordenite) to an ultimate value of 0.1%.In this latter sample the Si:A1 ratio was extremely high (of the order of 600) and it corres-ponded essentially to a silicic acid composition with a mordenite crystal-lattice. 5 1 B. W. Burbidge I. M. Keen and M. K. Eyles Adv. Chem. Ser. 1971 No. 102 400. 5 2 L. B. Sand in 'Molecular Sieves' Society of Chemical Industry London 1968 p. 71. 53 L. 1. Piguzova E. N. Prokof'eva M . M . Dubinin N. R. Bursian and Yu. A. Shavandin, Kinetika i Kataliz 1969 10 3 15. 5 4 P. E. Eberly jun. and C. N. Kimberlin jun. Ind. and Eng. Chern. (Product Res. and Development) 1970 9 3 3 5 . 5 5 H . S. Bierenbaum S. Chiramongkol and A. H. Weiss J .Catalysis 1971 23 61 Application of Molecular Sieve Zeolites to Catalysis 205 However it still contained sufficient Bronsted acid exchange sites to be an active cracking catalyst. The initial activity of this material was lower than that of the less-dealuminated samples but it had an appreciably lower rate of activity decline so that after a short time on-stream it became the most active catalyst. It was suggested that the essential nature of the catalytically active site i.e. the Bronsted acid site was not altered by the removal of such a drastic quantity of alumina. ’ The explanation of the improved lifetime of the low-alumina mordenite was that the low density of acid sites would reduce the rate of formation of higher molecular weight condensation products (capable of blocking the mordenite channel system).Further the more-open pore structure would facilitate the desorption of such heavy products. The removal of structural aluminium from Y zeolites also leads to improved thermal properties. An important and still somewhat controversial topic in this context is the production of so-called ultrastable fa~jasite.’~ The preparation involved specific thermal treatment of an Na-Y zeolite in which virtually all the sodium had been replaced by NH; by ion exchange. Ambs and Flank57 suggested that the thermal stability of synthetic faujasite was dependent only upon the level of sodium present and that there was no significant difference between decationized Y and ultrastable faujasite. However Kerr58 showed that the high thermal stability produced in a hydrogen-Y zeolite by heating at 973-1073 K (with the chemical water remaining in the environment of the zeo1ite)corresponded to a situation in which approximately 25 % of the aluminium was present in the cationic form.It was also shown that materials of improved thermal stability could be produced by removal of up to 50% of the framework aluminium from a Na-Y zeolite. In a detailed examination of calcination conditions it was dem~nstrated~~ that the geometry of the zeolite bed during calcination of the NH,-Y zeolite significantly affected the nature of the final product. The ultra-stable faujasites were only formed under conditions where the removal of ammonia and water from the bed was impeded. A detailed X-ray diffraction study6’ has indicated that the ultrastable Y faujasite has lost 15 framework aluminium atoms per unit cell and also a significant number of framework oxygen atoms.Jacobs and Uytterhoeven6 have reported an i.r. study of deepbed-calcined NH,-Y zeolites and describe two additional absorption bands at ca. 3700 cm- ’ and 3600 cm- ’ which appear to be characteristic of ultrastable faujasites. They have assigned these two bands to framework hydroxy-species created during the deepbed calcination procedure and have suggested that they could correspond to two distinct locations in the framework. 5 6 C. V. McDaniel and P. K. Maher in ‘Molecular Sieves’ Society of Chemical Industry, 5’ W. J. Ambs and W. H. Flank J . Catalysis 1969 14 118. 5 8 G. T. Kerr J . Phys. Chem. 1967,71 4155; 1968 72 2594.5 9 G . T. Kerr J . Carafysis 1969 15 200. 6 o P. K. Maher F. D. Hunter and J. Scherzer Adu. Chem. Ser. 1971 No. 101 266. 6 1 London 1968 p. 186. P. Jacobs and J. B. Uytterhoeven J . Catalysis 1971 22 193 206 H. I;. Leach Some of the patent literature emphasises the important industrial application of ultrastable faujasites. McDaniel and co-workers claim62 that after heating for two hours at 1173 K the surface area of stabilized faujasite-type zeolites was only reduced from 870 to 635 m2 g- '. Other patents that inclusion of an ultrastable component in the zeolite catalyst gives higher hydrocracking and hydrodenitrification activities with lower fouling rates and a lower starting temperature (for the same conversion) than with conventional zeolite catalysts.The material produced by dealumination of a Y zeolite (to a SiO :A1,0 molar ratio of 16.7) has been claimed6* to produce a material useful as a support for cracking catalysts. Beaumont et a1.69*70 have used the cracking of iso-octane (with isobutene as a major product) to examine the catalytic activity of X and Y zeolites exchanged with Ca2+ and La3+. They also examined the acidity of these materials using Hammett and arylmethanol indicators. Their results indicated that the number of acid sites was dependent upon the valency of the exchanged cation and that the cracking activity was affected by the nature of the carrier gas (considerably enhanced when hydrogen was used). They concluded that the acidic and catalytic properties were increased when ion exchange took place at the inner cation sites rather than at sites located near or in the supercage.In examining the effect of dealumination (by repeated extraction with edta) they found that the catalytic activity was apparently unaltered until more than 35% of the aluminium was removed from the structure and they observed three acidity sites of differing strengths. M o s c o u ~ ' ~ ~ has used chemical methods to investigate the acid sites on rare-earth-exchanged (RE) zeolites i.e. by LiAlH reaction and Karl Fischer titration. After heating at 473-573 K the RE-Y zeolites contained only one acidic hydroxy-group in the supercage for each rare-earth ion introduced. For a RE-Y zeolite with a Si02:A1203 molar ratio of 5.0 and with 75% of the sodium exchanged by rare-earth cations the density of acidic hydroxy-groups (Bronsted acid sites) was quoted as 7 x lo2' sites g-'.This value is some 70 times higher than that for amorphous silica-alumina catalysts. Ben Taarit et have also examined the acidic (and oxidizing) properties of rare-earth-exchanged Y zeolites. The i.r. spectrum of RE-Y after calcination at various temperatures suggested the presence of both Bronsted- and Lewis-type acidity. 6 2 U.S.P. 3 595 611/1971. 6 3 U.S.P. 3 535 227/1970. 6 4 U.S.P. 3 536 605/1970. 6 5 U.S.P. 3 536 606/1970. 6 6 U.S.P. 3 558 471/1971. " G.P. 2 000 026/1970. 6 8 G.P. 2 061 285/1971. 6 9 R. Beaumont and D. Barthomeuf Cornpt. rend. 1969,269 C 617; 1971 272 C 363. '' R. Beaumont D. Barthomeuf and Y. Trambouze Ado. Chem. Ser.1971 No. 102,327. " L. Moscou and M. Lakeman J . Curalysis 1970 16 173. 7 2 L. Moscou Adv. Chem. Ser. 1971 No. 102 337. 7 3 Y. Ben Tarrit M. Mathieu and C. Naccache Adv. Chem. Ser. 1971 No. 102 362 Application of Molecular Sieve Zeolites to Catalysis 207 Brief mention has been made of the fact that the cracking patterns observed over Na-X zeolites did not correspond to those observed over typical acidic catalysts. Rabo and Poutsma’ have reported an examination of the cracking of n-hexane and other paraffins over K-Y Na-Y and Na,K-exchanged L zeolites at 773 K. The observed product distributions over these alkali-exchanged zeolites could be rationalized in terms of a modified radical mechanism with double-bond-shift isomerization processes occurring as secondary reactions subsequent to rather than an inherent part of the cracking process.It seems clear that with these materials the catalytic activity cannot be associated with ionic-type mechanisms. Miale and we is^^^ have reported that the catalytic cracking activity (for n-hexane) of Na-X was increased by a factor of five to ten by contact with H,S or S followed by oxygen (air) or with SO alone. In each case the same sulphur-zeolite complex appeared to be formed the sulphur content corresponding to one sulphur atom for two of the mobile sodium atoms of the aluminosilicate lattice. The chemistry of the hydrocracking process has been reviewed7 with particular reference to the shape selectivity exhibited by zeolites. The use of the small-pore 5A zeolites as shape-selective catalysts was first described by Eng,76 and sub-sequently developed by several authors.Robson et have reported the use of synthetic erionite as a selective catalyst for hydrocracking. Tests on a c5&6 naphtha showed strong selectivity for converting n-paraffins to gaseous products, particularly propane. The selectivity decreased and other components of the naphtha feed were cracked as the temperature was raised. X-Ray and electron diffraction data indicated that the synthetic erionite used contained intergrowths of the related offretite structure (a relatively large-port zeolite). Such impurities were believed to be responsible for the manner in which the catalyst exhibited appreciable conversion of branched-chain paraffins which could not enter the erionite pore structure.Workers at the Mobil laboratories have r e p ~ r t e d ~ ~ * ~ * studies on a synthetic offretite containing tetramethylammonium ions (probably located in the large intracrystalline pores along the c axis of the offretite lattice). It is claimed7’ that these materials are good hydrocracking catalysts. Lattice-associated hydroxy-groups confirmed as protonic in nature by interaction with ammonia were generated from the tma cations. Dehydroxylation of the acid offretite at about 773 K generated electron-acceptor (Lewis acid) sites. It was noted that higher temperatures were required to convert the NH4-offretite to the acid form than with NH,-Y. Vacuum fragmentation of the bulky tma cations within the narrow offretite channels produced a complex product mixture which could be ra-tionalized however by superimposition of a variety of ‘classical’ reaction patterns.i4 J. N. Miale and P. B. Weisz J . Catalysis 1971 20 288. ’’ G . E. Langlois and R. F. Sullivan Adu. Chem. Ser. 1970 No. 97 38. 7 h U.S.P. 3 039 95311962. 7 7 H. E. Robson G. P. Hamner and W. F. Areyjun. Adu. Chem. SPY. 1971 No. 102.417. ’’ E. L. Wu T. E. Whyte. jun. and P. B. Venuto J . Caralysis 1971 21 384. 7 9 U.S.P. 3 578 398/1971 208 H. F. Leach It has also been shown" that thermal decomposition of a series of methyl-ammonium-cation-exchanged Y zeolites will generate protonic sites. A brief examination of the patent literature indicates that the application of zeolites as catalysts in cracking and hydrocracking processes is still generating a tremendous amount of industrial research.In the past two years some 40 patents additional to those already mentioned have appeared. These reveal the continued interest in the use of rare-earth-exchanged faujasite materials and also reflect the growing employment of zeolites of the mordenite and erionite type. Alkylation Dealkylation and Isomerization Reactions.-Venuto and co-w o r k e r ~ ~ ~ have investigated a wide variety of alkylation reactions over zeolite catalysts. They have concluded that such reactions generally proceed via carbonium-ion-type mechanisms and that they showed great similarity to the corresponding features commonly reported for electrophilic aromatic substitu-tion in the presence of strong protonic acids such as concentrated H2S0,, liquid HF or promoted Lewis acids.Normally ortho,para-orientation was observed and selectivity for attack on the reactive (nucleophilic) aromatic nucleus was exhibited in competitive situations. They based their conclusions about the nature of the reaction on an analysis of the structures of the alkyl-aromatic products the patterns of substrate reactivity and the pathways of side-reactions. Their results provided qualitative evidence for a Rideal-like mechanism where in an olefin alkylation the initial step would be the fast reversible non-competitive adsorption of the olefin on the catalyst acidic sites. Mays and Pickert' reported that multivalent-cation-exchanged and deca-tionized Y zeolites were excellent catalysts for the alkylation of aromatic hydro-carbons with C,-Cl2 olefins or alkyl halides.High reaction rates were obtained at low temperatures and pressures with high selectivities. The observed activity was comparable to that of aluminium chloride promoted by HC1 and consider-ably higher than that exhibited by amorphous silica-alumina or supported (or non-supported) mineral acids. The zeolites would alkylate with good selectivity and give high yields with species such as t.hiophen or phenolic ethers which would be readily degraded by mineral acids. Pickert et ~ 1 . ~ ~ observed that the alkylation activity of faujasite catalysts (for the benzene-propene alkylation reaction) was enhanced with increase in calcina-tion temperature. The condition for maximum activity corresponded to the situation where all the residual hydroxy-groups associated with catalytically-active sites had been removed.It was suggested that the carbonium ion inter-mediates were formed by polarization of the reactant hydrocarbons (by the electrostatic fields set up by the cations). However Rabo et reported that E. L. Wu G . H. Kuhl T. E. Whyte jun. and P. B. Venuto Adv. Chem. Ser. 1971, No. 101 490. 8 1 R. L. Mays and P. E. Pickert in 'Molecular Sieves' Society of Chemical Industry, London 1968 p. 112. 8 2 P. E. Pickert A. P. Bolton and M. A. Lanewala presented at the 59th American Institute of Chemical Engineers Meeting Columbus Ohio 1966. 8 3 J. A. Rabo C. L. Angell and V. Schomaker in 'Proceedings of the 4th International Congress on Catalysis' Moscow 1968 vol. 3 p. 966 Application of Molecular Sieve Zeolites to Catalysis 209 the activity of a La-Y zeolite (for the toluene-propene alkylation reaction) was not affected by changes in the calcination temperature.The La-Y showed high activity at room temperature whether activated at 823 K or near 973 K. It was postulated that oxygen-deficient lattice sites (Lewis acid sites) generated during the 973 K activation could have alkylation activity. More recently Morita and ~ o - w o r k e r s ~ ~ in a study of the alkylation activity (for benzene-ethylene alkyla-tion) of La-Y reported a decrease in activity with increase in catalyst-pretreatment temperature (over the range 448-548 K). They suggested the catalyst was being poisoned by adsorption of excess water on the active sites. Recent patents have claimed that the H-Y zeolite is a good catalyst for benzene alkylation with propane,8 and that Y zeolites containing rare-earth manganese, or aluminium ions are good alkylation catalysts for the preparation of Cs hydrocarbons from isobutane and but-1 -ene.86 Isakov et ~ l .~ ’ have reported that benzene alkylation (with propene) proceeds rapidly on Y zeolites exchanged with Ca2+ or Cd2+. An increase in activity with increase in extent of ion exchange was observed. An examinations8 of the alkylation activity of toluene with methanol for a series of Y zeolites has found the following relative order of activity: RE-Y > H-Y > bivalent forms > univalent forms. Selective formation of p-xylene in the product mixture was noted eg. in excess of 50% compared to the thermodynamic equilibrium quantity of about 22 ”/,.The authors suggested that depression of the secondary isomerization reaction of the initially-formed xylene in the zeolite supercages was responsible for such selectivity. Addition of HC1 to a Mn-Y zeolite promoted the p-xylene selectivity presumably suggesting a correlation with Bronsted acidity. There was also a correlation between the extent of cation exchange and the catalytic activity (and the p-xylene selectivity). Sidorenko and G a l i ~ h ~ ~ have also investigated the mechanism of toluene methylation over zeolite catalysts. For Na-alkaline-earth zeolites they noted that Y zeolites were more active than X zeolites and the activity decreased from magnesium to strontium. They postulated a mechanism involving the decomposi-tion of protonated methanol (the essential first step) to methylene radicals and H 3 0 + acid sites on the catalyst.The absence of either ethylbenzene or styrene was attributed to the specific geometry of adsorption of the toluene which prevented insertion of the methylene radical into the C-H bonds of the side chain. Kawakami et a/.’’ have stated that alkali-metal Y zeolites are active for methyl migration of anisole in the temperature range 598-723 K and they reported relative activities in the order Li-Y > Na-Y > K-Y > Rb-Y. The rates 8 4 Y. Morita M. Takayasu and H. Matsumoto Kogyo Kagaku Zasshi 1970 73 2540. 8 5 G.P. 1934426/1970. 86 G.P. 1931 425/1970. ‘I Y. I. Isakov N. V. Mirzabekova V. I. Bogomolov and Kh. M. Minachev Neftekhimiya, 1970 10 520.T. Yashima K. Yamazaki H. Aymad M. Katsuta and N. Hara J. Catalysis 1970, 16 273; 17 151. 8 9 Y. N. Sidorenko and P . N. Galich Ukrain khim. Zhur. 1970 36 1234. 90 S. Kawakami S. Takanashi and S. Fujii Kog-vo Kagaku Zasshi 1971,74 899 210 H . F. Leach of formation of phenol cresol and methylanisole together with the rate of con-version of anisole were all zero order with respect to anisole. The zeolites that catalyse alkylation processes will also act in general as catalysts for dealkylation and patent claims invariably link together alkylation and dealkylation activity. However it is evident7 that considerably higher temperatures are normally required for dealkylation. The cumene-dealkylation reaction discussed in the previous section is probably the most widely studied example of this type.A recent patent’’ refers to the use of a modified mordenite catalyst for the selective dealkylation of C + hydrocarbons with the selective nature of the catalyst being a function of the mordenite pore structure. A large number of papers have been published concerned with the catalytic activity of zeolites for the transalkylation and isomerization of alkyl-aromatic species. In the very thorough investigation of the relationship between catalytic activity and zeolite structural properties Ward3’ correlated his i.r. data with o-xylene isomerization activity. Attention has recently been focussed upon the activity of alkaline-earth Y zeolite^,'^,^^ where NH,-Y has been back-exchanged with Ca2+ and Mg2+. The changes in activity observed as the extent of exchange was altered have been correlated with accessible acid site concentra-tions and ascribed to the different polarizing effects of the two cations on the site strength.The observed increase in catalytic activity was matched by an increase in the Bronsted acid site concentration with the Mg-Y zeolites being more acidic than Ca-Y. The Bronsted acidity and the o-xylene isomerization activity was also reported to increase with increase in the SiO :A1203 molar ratio. Sidorenko and G a l i ~ h ’ ~ have also reported that zeolites containing alkaline-earth metals are more active for rn-xylene rearrangements than those containing alkali metals. They attributed their results to the greater acidity of water molecules associated with the polyvalent ions.proposed that the isomerization of diethyl-benzenes over a modified Y zeolite occurred via a transalkylation mechanism involving di-phenylethane-type intermediates. However such a mechanism cannot occur unless readily-extractable a-hydrogen atoms are available i.e. not if the alkyl group is t-butyl. Csicsery and Hi~kson’~ examined the isomerization of 2-ethyl-1-methylbenzene over a series of Y zeolites (in the temperature range 473-673 K). They found that two major independent reactions were taking place i.e. isomeriza-tion to 3-ethyl-1-methylbenzene and 4-ethyl-l-methylbenzene and transethyla-tion to toluene and diethylmethylbenzenes. For a given catalyst the isomeriza-tion transethylation ratio increased with increasing water content of the reaction mixture.They concluded that their results could be explained if the isomerization process was primarily catalysed by Bronsted acid sites whereas the transethyla-tion was primarily a Lewis acid-catalysed reaction ; or if the transethylation was Bolton et 91 Fr.P. 2 010 151/1970. y 2 J. W. Ward J . Catalysis 1970 17 3 5 5 . 9 3 94 Y. N. Sidorenko and P. N. Galich Ukrain. khim. Zhur. 1970 36 1120. y 5 A. P. Bolton M. A. Lanewala and P. E. Pickert J . Org. Chem. 1968 33 1513. y 6 S. M. Csicsery and D. A. Hickson J . Cafalysis 1970 19 386. R. C. Hansford and J. W. Ward Adu. Chem. Ser. 1971 No. 102 354 Application of Molecular Sieve Zeolites to Catalysis 21 1 catalysed by a Bronsted-Lewis site-pair and the isomerization was catalysed by a single-type Bronsted acid site.Csicsery' has examined the shape-selectivity exhibited by mordenite for transalkylation processes. Symmetrical trialkylbenzenes are normally the pre-dominant components of trialkylbenzene isomer mixtures at thermodynamic equilibrium. However although mordenite catalysts have adequate isomeriza-tional activity no symmetrical isomers are formed (in marked contrast to the product distribution observed over X and Y zeolites). There is thus shape-selective kinetic control where the special crystal structure of the catalyst prevents the formation of the thermodynamically favoured isomer. The effective channel diameters in acid msrdenite are between the minimum cross-sections of the wider symmetrical trialkylbenzenes and the other trialkylbenzene species (approximately 0.86 and 0.82 nm respectively).It was noted that the mordenite catalysts were deactivated much faster than the Y zeolites presumably because the mordenite pore system can be more easily blocked by strongly adsorbed molecules. The importance for transalkylation reactions of aromatic species of catalysts based on the mordenite structure is demonstrated by several patents9*-"' where either the acid form or nickel-containing mordenites are claimed to be active for such processes at attractively low temperatures. Japanese workerslo2 reported that H-mordenite was an effective catalyst for the disproportionation of toluene but that bivalent cation-exchanged synthetic mordenites exhibited little activity for this reaction. There have been a number of reports of the manner in which multivalent metal cation and decationized Y zeolites (usually containing small amounts of a noble metal) exhibit high activity and selectivity for the isomerization of C,-C6 n-paraffins.Penchev et a1.Io3 have compared the acidity and catalytic activity (for the isomerization of n-hexane and cyclohexane) of Ca-Y Mg-Y and a Y zeolite containing Pt. They noted that the acidity (and activity) increased with Ca2+ and Mg2+ content but that the Pt content did not apparently affect the zeolite acidity. Kubasov et / . I o 4 also examined cyclohexene isomerization over a series of Y zeolites and reported that the activity fell in the order L-Y > H-Y > Ca-Y. Maximum isomerization was noted just before the onset of cracking and the formation of cyclohexane (evidently by a parallel reaction on different active centres) also took place.They were not able to establish any correlation of the catalytic activity with a particular acidity type and sug-gested that the conversion occurred on both types of active centres by different 9 7 9 8 G.P. 1925 102/1970. 9 9 G.P. 1946 187/1970. l o o G.P. 2 000 491/1971. l o ' G.P. 2 006 902/1971. l o 3 V. Penchev V. Kanazirev and Khr. Minchev Cornpt. rend. Acad. bulg. Sci. 1969, l o 4 A. A. Kubasov A. N. Ratov K. V. Topchieva and L. M. Vishnevskaya Veslnik S. M. Csicsery J. Catalysis 1970 19 394; 1971 23 124. T. Yashima H. Moslehi and N. Hara Bull. Japan Petrol. Znst. 1970 12 106. 22 899. Moskou. Unit?. 1970 11 406 212 H . F. Leach mechanisms.A recent article"' describes the hydroisomerization process operated by Shell where C and C paraffins are converted in the vapour phase, over zeolites containing noble metal into highly-branched compounds with high octane numbers. Beecher and Voorhies'06 showed that a synthetic H-mordenite catalyst had a high n-hexane isomerization activity with or without dispersed noble metal. The effect of pressure on the rate constant for the isomerization process was consistent with a dual-site catalytic mechanism. The hydroisomerization of cyclohexane (to methylcyclopentane) has also been r e p ~ r t e d ' ~ ~ ' ~ ~ to provide data consistent with a dual-site mechanism over a series of H-mordenite catalysts containing Pd and with differing Si02 :A120 molar iatios. The activation energies for the isomerization were similar to those for large-port zeolites so it was concluded that there were no macropore diffusion limitations to the results.Cyclohexane isomerization has been examined,' O9 over Pt-Al,O,-mordenite catalysts and it was found that simultaneous isomerization and dehydrogenation occurred with significant diffusional effects. Ammonia-chemisorption experi-ments suggested that the mordenite component of the catalyst was some twelve times more acidic than the alumina and the isomerization activity was reported to be directly related to the catalyst acidity. Minachev et al.' ' have also examined the catalytic isomerization properties (for cyclohexane and n-pentane) of synthetic mordenites. They reported that H-mordenite was more active than bivalent and tervalent cation-exchanged forms and that the Na- Li- and K-exchanged mordenites had negligible iso-merization activity.They postulated that the mechanism over H-mordenite was different to that normally attributed to metal-zeolite isomerization catalysts. It was suggested that the carbonium ion species was formed by the splitting off of a hydride ion from the saturated molecule of the starting hydrocarbon and not via the attachment of a proton. Eberly et a1.l" have investigated the de-alumination of a series of acid mordenite samples (containing 0.5 % by weight Pd) with particular regard to the effect upon acidity and catalytic activity for n-pentane hydroisomerization ; SiO A120 ratios from 12-97 were prepared. 1.r. spectral changes together with ammonia-adsorption data indicated that the surface acidity was decreasing with increase in the Si02 :A120 ratio.As the catalytic activity also decreased it was concluded that the surface acidity was the dominant factor in the process. The isomerization of cyclopropane over Na-Y and NH,-Y has been cor-related112 with the Bronsted acidity of the zeolites. The nature of the deuteriated I o 5 H. W. Kouwenhoven and W. C. Van Zijill Langhout Chem. Eng. Progr. 1971,67,65. l o b R. Beecher and A . Voorhies Ind. and Eng. Chem. (Product Res. and Development), 1969 8 366. l o ' J . R. Hopper Diss. Abs. ( B ) 1970 30 5026. l o ' A. Voorhies and J. R. Hopper Adv. Chem. Ser. 1971 No. 102 410. I o 9 D. E. Allan Diss. Abs. ( B ) 1971 31 4652. ' l o Kh.Minachev V. Garanin T. Isakova V. Kharlamov and V. Bogomolov Ado. ' I 1 'I2 Z . M. George and H. W. Habgood J . Phys. Chem. 1970,74 1502. Chem. Ser. 1971 No. 102 441. P. Eberly C. N. Kimberlin and A. Voorhies J . Catalysis 1971 22,419 Application of Molecular Sieve Zeolites to Catalysis 21 3 propene species obtained when the isomerization was carried out over a corn-pletely deuteriated catalyst suggested that the cyclic C3H6D+ ion was able to equilibrate with the various isotopic forms before ring-opening occurred. Flockhart et al.' ' have considered the effect of calcination temperature upon the cyclopropane isomerization activity of a Na-Y zeolite partially exchanged with NH;. They also found a clear correlation between Bronsted acidity and activity but a second mechanism appeared to be operative at high calcination temperatures (ca.930 K) where the Bronsted acidity was low. Under such condi-tions the electron-donor power as measured by the formation of trinitro-benzene anion radicals from the adsorbed parent molecule was at a maximum. It was therefore postulated that the active site could be either a Lewis acid centre or possibly an electron-transfer site of the type reponsible for the redox activity of zeolites. The isomerization of n-butenes without skeletal rearrangement has received considerable attention as a test reaction for the characterization of zeolite catalysts. Dimitrov and Leach' l4 found that Na-X was relatively inactive (requiring temperatures above 475 K) but a marked increase in activity was observed on exchange with Cu2 + cations.The initial cis-trans but-2-ene product distribution over Na-X and the low-exchanged Cu-X was compatible with a radical-type mechanism. Cross et a!.' 15,116 have reported a range of initial product ratios for a series of cation-exchanged X zeolites. Results with Ce-X (and the majority of other zeolites examined) were indicative of a carbonium ion mechanism. Over Ni-X and to a lesser extent Zn-X however a radical-type mechanism was proposed. Evidence in support of this was obtained by following the isomerization in the presence of (i) deuterium (when an enhancement of rate and extensive exchange were observed) or (ii) deuterium oxide (when very little exchange occurred and the reaction rate was virtually unchanged). Tempere and co-workers"7 have reported cis:trans product ratios of ca.2.0 for the but-1-ene isomerization over a series of X and Y zeolites but over an A-type zeolite reported that trans-but-2-ene was the major product. They concluded that the active sites for the isomerization were hydroxy-groups localized in hexagonal sites (corresponding to an i.r. absorption band at 3600cm-') and that the activity was not directly dependent upon the electrostatic field of the exchanged cation. Hall et al.' l8 have examined the manner in which the activity and selectivity (for the but-1-ene isomerization) of an Na-Y zeolite were modified by exchanging small amounts of Ca2+ and by creation of a cation deficiency by hydrolysis. As the Ca2+ content (and the acidity) was increased cis-trans isomerization was enhanced relative to double-bond migration and it was also reported that the ' I 3 B.D. Flockhart L. McLoughlin and R. C. Pink Chem. Comm. 1970 818. Chr. Dimitrov and H. F. Leach J . Catalysis 1969 14 336. N. E. Cross C. Kemball and H. F. Leach Adu. Chem. Ser. 1971 No. 102 389. N. E. Cross C. Kemball and H. F. Leach J . Chem. SOC. ( A ) 1971 3315. ' I ! J. F. Tempere J. Kermarec and B. Imelik Bull. SOC. chim. France 1970 3808 4227. ''* W. K. Hall E. A. Lombardo and G. A. Sill J. Catalysis 1971 22 54 214 H. F. Leach catalytic activity increased with increasing cation deficiency (up to 0.94 % original Na+ extracted by hydrolysis). From data obtained using added water as a co-catalyst it was concluded that the reaction intermediate was the s-butyl carbonium ion.Furthermore it was postulated that a pure alkali-Y zeolite containing no cation deficiency no bivalent ions and therefore no decationized sites would have negligible catalytic activity for the isomerization of n-butenes. Oxidation Reactions.-Crystalline aluminosilicates show very little intrinsic catalytic activity for oxidation reactions and oxidation processes over zeolites invariably feature materials containing transition-metal ions e.g. Mn-Y has been reported7 to catalyse the oxidative dehydrogenation of ethylbenzene to styrene and the selective oxidation of benzyl alcohol to benzaldehyde in the temperature range 5 2 3 4 4 3 K. Fripiat and ~ o - w o r k e r s ~ ~ ~ - ~ ~ ~ reported the liquid-phase oxidation of several hydrocarbon species by X zeolites containing Co Mn, or Mo.In the oxidation of p-xylene over a mixed CwMn-X zeolite they observed high selectivity for terephthalic acid formation and over Mo-X propene was selectively oxidized to propene oxide. It was postulated that the high oxidation activity exhibited was related to the promotion of electron-unpairing in the transition-metal cations by the zeolite support. is concerned with catalytic oxidation of propene and ethylene over Y zeolites containing transition-metal cations. They noted that the oxidation of propene over Cu-Y in the presence of steam exhibited some selectivity with the preferential formation of acrolein. The relative order of oxidation activity was given as Pd z Pt > Cu > T1 > Ag > Mn > Ni > Co > Zn > V > Cr > Na.The reaction orders were all observed to be approximately 0.5 in oxygen over the various zeolites but to vary for propene. It was therefore concluded that the olefin adsorption was important in the oxidation process, and the activity sequence was correlated with a parameter expressing the tendency of the metal cation to form a dative n-bond. Iron-containing zeolites have also been reported'23 to exhibit good catalytic activity for propene oxidation. Van Sickle and P r e ~ t ' ~ ~ have examined the reaction of oxygen with cyclo-pentene but-1-ene and but-2-ene adsorbed on cobalt-exchanged A and X zeolites in the temperature range 298-363 K. They reported oxidation rates some 5OMOO times greater than in more conventional homogeneous oxidation systems.However there were a multiplicity of products many of which were tightly bound to the zeolites (especially the A-type zeolites). The products were also different in structure the principal products of homogeneous oxidations are hydroperoxides but the more prominent volatile products from the butene The series of papers by Mochida et J. Rouchaud L. Sondengam and J. J. Fripiat Bull. SOC. chim. France 1968 4387. I 2 O J. Rouchaud P. Mulkay and J. J. Fripiat Bull. SOC. chim. belges 1968 77 537. 1 2 ' J. Rouchaud and J. J. Fripiat Bull. SOC. chim. France 1969 78. I. Mochida S. Hayata A. Kato and T. Seiyama J . Catalysis 1969 15 314; 1970, 19,405; 1971 23 3 1 . 1 2 3 L. V. Skalkina I. K. Kolchin E. Y. Margolis N. F. Ermolenko S. A. Levina and L. N. Melashevich Izuest.Akad. Nauk S.S.S.R. Ser. khim. 1970 980. I z 4 D. E. Van Sickle and M. L. Prest J . Cata/ysis 1970 19 209 Application of Molecular Sieve Zeolites to Catalysis 21 5 oxidation over zeolites were methyl ethyl ketone crotonaldehyde and but-2-ene- 1-01. In an e.s.r. investigation of cationic oxidation sites in faujasite Richardson 12' suggested that an electron-transfer process took place at the Cu2+ ion in Cu-Y zeolites containing 2% by weight of Cu. Naccache and Ben Taarit'26 have further examined the oxidizing (and acidic) properties of Cu-Y using e.s.r. and i.r. spectroscopic techniques. They found that at low activation temperatures the zeolite exhibited Bronsted acidity and no true Lewis acid sites could be detected. However carbon monoxide treatment of the zeolites caused reduction (cupric to cuprous) and formed Lewis acid centres.They concluded that the oxidizing properties of the cupric Y zeolite could be attributed to the cupric ions whereas those of the reduced samples were due to the true Lewis acid sites. Miscellaneous Catalytic Studies-In this section a limited number of further points of catalytic significance which in themselves do not justify separate sections will be discussed. In the paper74 in which the cracking activity of Na-X was reported to be increased by the inclusion of sulphur it was also noted that the catalytic activity was increased by the inclusion of selenium or tellurium. However the catalyst then became a dehydrocyclization catalyst rather than a cracking catalyst. Further communications'27~'28 from the Mobil laboratories have reported a more detailed examination of this novel dehydrocyclization catalyst particularly the catalyst containing tellurium.The test reaction for this activity was the aromatization of n-hexane (to benzene). X-Ray examination indicated the presence of tellurium both within the sodalite cage and within the supercage. It was established that the presence of a significant concentration of cations in type I11 sites was essential for the catalytic activity and it was postulated that the catalytically-active entity was a tellurium ion situated in the zeolite supercage and co-ordinated to cations in type I1 and type I11 sites. It is generally considered that for the majority of situations catalysis over zeolites occurs within the zeolite porous structure.In shape-selective catalysis the situation corresponds to molecules of suitable dimensions continuously enter-ing and leaving the molecular sieve cavities. The situation of reactant selectivity corresponds to the case where only one of two classes of reactant molecules can pass through the pores; on the other hand with product selectivity only those products (formed within the porous structure) with suitable dimensions can diffuse out and appear as observed products. However there are clearly many instances where reactions which cannot occur within the zeolite have been reported, e.g. the formation of acetophenone tetramer over H-Y," and in such circum-stances the external surface must play some catalytic role. Differences in reac-tivity and selectivity of such surface sites could be expected as the environment J.T. Richardson J . Catalysis 1967 9 172. 126 C. Naccache and Y . Ben Taarit J . Catalysis 1971 22 171. 1 2 ' W. H. Lang R. J. Mikovsky and A. J. Silvestri J. Catalysis 1971 20 293. 2 8 R. J. Mikovsky A. J. Silvestri E. Dempsey and D . H. Olson J. Catalysis 1971,22,371 216 H . F. Leach of the reactant molecules would be totally different and no diffusion restrictions would be applicable. Such surface (external) reactions could reduce the selec-tivity of the zeolite catalyst and in fact enhanced selectivity has recently been claimed129 by reduction of the catalytic activity of the external surface. An erionite zeolite containing nickel when treated with copper acetate was reported to have negligible undesirable side-reactions (presumably on the external surface between molecules unable to enter the pore structure) e.g.isoparaffin formation from 2-methylpentane hydrocracking. A similar improvement in selectivity has been disclosed'30 by poisoning of the external sites with a large organic, phosphorus-containing compound such as tricresyl phosphate. Thomas and BarmbyI3' proposed that the primary cracking ofgas oil molecules on zeolite catalysts occurred on the external surface of the catalysts. They considered that the improved gasoline products obtained relative to those obtained over amorphous silica-alumina materials resulted from subsequent hydrogen-transfer reactions of the gasoline species within the zeolitic structure.However Weisz4 has suggested that the unusually high kdk ratio for the interior of the zeolite where k refers to hydrogen transfer and k to cracking is sufficient to account for the unusual product distributions of the zeolites and that participa-tion of the external surface is not substantial. It has already been noted that mordenite zeolites particularly the H form, are active alkylation catalysts e.g. for alkylation of benzene with propene. In this context the studies of Satterfield and co-workers concerning the diffusion of hydrocarbon species within mordenite systems are very relevant. The rates of diffusion of methane butane isobutane and perfluorobutane were observed' to exhibit a maximum with time in single crystals of Na-mordenite in the temperature range 298-373 K.A more detailed examination of the diffusion properties of benzene and cumene in H-mordenite' 3 3 indicated that counter-diffusion of these species was very small (relative to the situation in Y zeolites). Observed changes in the desorptive diffusion coefficient of cumene were attributed to the slow formation in the H-mordenite pores of large molecular species via cumene disproportionation which blocked the pores. It was concluded that the major portion of the internal area of H-mordenite was apparently unavailable for reaction of aromatic species at temperatures moderately above ambient. The suggestion was made that the alkylation-type reactions must occur either on the external surface or just within the pore mouth. Venuto" has also postu-lated pore-mouth catalysis to explain the relative isomerization and deuteriation rates of 2,3-dimethylbut-l-ene over a deuteriated Y zeolite.It was there suggested that the limited extent of deuteriation corresponded to a situation in which the majority of the intracrystalline OD groups were effectively inaccessible and that U.S.P. 3 554 900/1971. 130 U.S.P. 3 575 845/1971. C. L. Thomas and D. S. Barmby J . Catalysis 1968 12 341. 1 3 * C. N. Satterfield and W. G. Margetts Amer. Inst. Chem. Engineers J . 1971 17 295. 1 3 ' C. N. Satterfield J. R. Katzer and W. R. Vieth Ind. and Eng. Chem. (Fundamentals), 1971 10 478 Application of Molecular Sieve Zeolites to Catalysis 217 only those species near to the external surface were operative in the exchange (deuteriation) reaction.Although the majority of reactions catalysed by zeolites have been shown to occur as a result of the acidic nature of the catalysts mention has already been made of particular instances where radical-type reactions are observed and intermediates of a non-ionic character have been postulated. A paper by Venuto and L a n d i ~ ' ~ ~ in which the formation of stilbenes from reactions of benzyl-type mercaptans was examined gives an example where the presence of the zeolite alters the reaction very considerably (both in terms of increased activity and stilbene selectivity) but not as a direct consequence of the acidity of the zeolite. It has been suggested that in this particular instance the enhanced activity can possibly be explained in terms of the large surface area and a concentration (of reactant) effect rather than the more-specific catalyst interactions normally invoked to rationalize the catalytic activity of zeolites.The redox activity of zeolites was briefly mentioned in connection with the isomerization of cyclopropane and Roginskii et ~ 2 1 . ' ~ have further examined the redox activity of a series of transition-metal-exchanged Y zeolites. They reported that in all cases there was enhanced catalytic activity relative to Na-Y, for the oxidation of hydrogen carbon monoxide ethylene and ammonia. They concluded that the active centres were closely associated with the transition-metal cations. Minachevg has also discussed the redox catalytic activity of zeolites and noted that in such reactions zeolites with a lower SiO A1,03 ratio, and hence a higher cation density are generally more active.This is in marked contrast to reactions in which carbonium ions are involved. Although possibly not a redox reaction the high activity of Na-mordenite for the hydrogenation of benzene (in the absence of any conventional hydrogenating components such as Pt Pd or Ni) reported by Minachev et al.,'" appears to be a particularly important reaction in this context. The hydrogenation activity decreased sharply when Na+ was replaced by H' and could be generated by treatment of H-mordenite with sodium chloride so the direct participation of sodium cations in the benzene hydrogenation seems well established. The major role played by i.r. spectroscopy and X-ray diffraction techniques in the characterization of zeolite catalysts has already been indicated.It is instructive to briefly consider some of the other techniques that are now being employed to obtain information about zeolites. Calorimetric studies both heats of adsorption' 36 and heats of imrner~ion,'~~ have recently been reported where the data have been used (i) to measure the acidity of sites in X zeolites and (ii) to suggest that the electrostatic field is the source of the catalytically-active site for carbonium ion reactions of Ca-Y zeolites. In an examination of the nature of rare-earth-exchanged Y zeolites Bolton' 38 has combined i.r. studies with thermo-' 3 4 P. B. Venuto and P. s. Landis J . Catalysis 1971 21 330. 1 3 ' S . Z. Roginskii 0. V. Al'tshuler 0. M. Vinogradova V.A. Seleznev and I. L. 1 3 6 Y . Okamoto T. Imanaka and S. Teranishi Bull. Chem. SOC. Japan 1970,43 3353. 1 3 ' Tsitovskaya Doklady Akad. Nauk S . S . S . R . 1971 196 872. K. Tsutsumi and H. Takahashi J . Phys. Chern. 1970,74 2710. A. P. Bolton J. Catalysis 1971 22 9 218 H. F. Leach gravimetric analysis. The latter technique indicated that stoicheiometrically, one water molecule was associated with each rare-earth cation. TungI3' has proposed the concept of dynamic Bronsted acidity i.e. variation of acid strength with time to explain aspects of the activity of zeolites. This proposal arose out of his measurements of the dielectric response to temperature and electrical frequency changes which indicated considerable ionic movement over the zeolite surface.Dyer et ~ 2 1 . ' ~ ' have observed the mobility of Sr2+ and Ba2+ cations in X zeolites using radiochemical techniques and Gallezot and c o - ~ o r k e r s ' ~ have reported an X-ray study of the movement of copper cations in Y zeolites after ammonia or pyridine chemisorption. They observed some displacement of Cu2 + from I sites to .I' sites after ammonia treatment and a more substantial migration of Cu2+ into the zeolite supercages (from I and I' sites) subsequent to pyridine adsorption. Naccache and Ben Taarit'42 have interpreted changes in the e.s.r. spectrum of Cu-Y zeolites after treatment with water ammonia or pyridine in terms of a similar migration of the cupric ions. Mikkeiken et have also reported an e.s.r. (and optical) spectroscopic study of the localization of Cu2+ in Y zeolites.They concluded that the cations were initially adsorbed as octa-hedral aquo-complexes. On dehydration the symmetry decreased and after treatment at 773 K in uacuo the Cu2+ became stabilized in hexagonal sites of the sodalite cages. Leith and c o - ~ o r k e r s ' ~ ~ employed e.s.r. techniques in an investigation of the nature of Cu-X zeolites. The spectra suggested the presence of exchanged cations in more than one crystallographic environment. It was shown that the Cu2 + in one particular environment reacted preferentially with but-1-ene at temperatures where the Cu-X zeolites were active for n-butene isomerization. Following the early work of Stamires and T ~ r k e v i c h ' ~ ~ there have been an extensive number of reports on the application of e.s.r.techniques to the study of the generation of free-radical species on zeolites. Papers describing the formation of biphenyl cation radicals from benzene over an ammonium-mor-denite catalyst,'46 and the e.s.r. spectra arising from the adsorption of olefins on RE-Y zeolites'47 typify such studies. In an investigation of the properties of lanthanide-exchanged zeolites NeikamI4* used e.s.r. data to show that the high radigenic activity of cerium-exchanged zeolites (relative to other lanthanide-exchanged zeolites) resulted from the presence of Ce4+ which was formed during activation in the presence of oxygen. 139 S. E. Tung J . Caralysis 1970 17 24. 14' 1 4 2 C. Naccache and Y . Ben Taarit Chem. Phys. Letters 1971 11 1 1 . 1 4 3 I. Mikheiken V. A. Shvets and V. B. Kazanskii Kinetika i Kataliz 1970 11 747. 144 I. R. Leith C. Kemball and H. F. Leach Chem. Comm. 1971 407. 145 D. N . Stamires and J. Turkevich J . Amer. Chem. SOC. 1964 86 749. 146 Y . Kuita T. Sonoda and M. Sato J . Catalysis 1970 19 82. 14' G. Raseev J . Caralysis 1971 20 120. 14* W. C. Neikam J . Caralysis 1971 21 102. A. Dyer R. B. Gettins and R. P. Townsend J . Inorg. Nuclear Chem. 1970 32 2395. P. Gallezot Y. Ben Taarit and B. Imelik Compt. rend. 1971 272 C 261 Application of Molecular Sieve Zeolites to Catalysis 219 Stevenson'49 has indicated how a detailed examination of the lineshape of the 'H n.m.r. spectrum of an H-Y zeolite can be analysed to provide data concern-ing the precise location of protons in the zeolitic framework. Freude et a1.I5' have also analysed the n.m.r. lineshape of a decationized Y zeolite and concluded that the hydroxy-groups existed mainly in pairs with an inner proton-proton distance of 0.37nm. Such an analysis requires the modification of what is a rather sophisticated model in the first instance ; consequently the technique has not yet been widely employed. However such work is an indication of the sophisticated and detailed data concerning the structure of zeolite catalysts that one might expect to be available in the future. 149 R. L. Stevenson J . Catalysis 1971 21 113. D. Freude D. Mueller and H. Schmiedel Surface Sci. 1971 25 289