Synthesis and Catalytic Properties of Titanium Containing Zeolites

Synthesis and Catalytic Properties of Titanium Containing Zeolites

P.J. Grobet et al. (Editors) /Innovation in Zeolite Materials Science © Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands 413 ...

670KB Sizes 0 Downloads 10 Views

P.J. Grobet et al. (Editors) /Innovation in Zeolite Materials Science © Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands


SYNTHESIS AND CATALYTIC PROPERTIES OF TITANIUM CONTAINING ZEOLITES B. NOTARI ENI Coordinamento Ricerca e Innovazione 20097 San Donato Milanese, MILANO, ITALY ABSTRACT The synthesis of titanium containing zeolites and their catalytic properties when Ilsed in oxidations of organic substrates with H are described. The 20 2 special cases of phenol to hydroquinone/catechol mixture and of propylene to propylene oxide are discussed. It is assumed that rr" enters the silicalite structure at random during synthesis: each Ti 1V is surrounded by si" and isolated from other Ti'v by long Si-O-Si-O sequences. It is proposed that isolated Ti"'have properties different from those of ri" having other Ti1vas near neighbours, in particular that isolated Ti lV have a reduced rate for H 20 2 decomposition; this, together with "restricted transition state selectivity" typical of medium pore zeolites gives rise to high yields of useful oxidation products. Analogies and differences with catalysts consisting of Ti0 deposited 2 on high surface area Si0 are discussed. 2 1. INTRODUCTION

The discovery of the new high silica zeolites of the ZSM family and their unique properties in catalytic

reactions has stimulated research in the field

of zeolites and zeolite - like materials. The effectiveness of organic bases in influencing the crystallization of precursors towards specific structures has offered a new tool which appeared promising also for the synthesis of materials in which a part of the Si

atoms of the framework is replaced by

elements different from Al'" . A research program was carried out at ENI centered around the synthesis of new metallosilicates. The idea of introducing foreign atoms in the crystal lattice of solids to change their catalytic properties is certainly not new: many catalysts are obtained by modifying in different ways the basic structure of a solid, with the consequence that important physical and chemical properties are changed, as well as the behaviour of these solids when used as catalysts. The nature, concentration and location of the foreign atoms introduced and the correlation with corresponding changes in catalytic properties is therefore of great interest for both theory and practice of catalysis.


Our first attempts were directed towards the introduction of


in the crystal

lattice of silicalite (ref. I, 2): convincing evidence that isomorphous substitution of


with s'" had occurred was obtained with the synthesis of

borosilicates haVing the framework structures of NU-I, B, 2SM-5 and 2SM-II high silica zeolites, named boralites. The regular decrease of the unit cell parameters with the boron content in the amount that could be calculated on the

si" and B"' is consistent si'" in the framework.

basis of the difference in radii between the hypotesis that B"' has substituted

only with

Recently we have been successfull in synthesizing a t.i taruim derivative of Silicalite-I (ref. 3,4,5): in this case the evidence for framework substitution of

sr'" with



was supplied by the regular increase in the unit cell

parameters with the ti t.anum content. The new material has been named tit.snnm silicalite-I or TS-L

I t has also been shown that Ti


is uniformly distributed

within the crystal of TS-I and that other t.Lt.anirm containing phases are below detection limits. From the difference in electronegativity between 0 (3.5) and Ti (1.32) it can be inferred that the bond has a marked ionic character, but the designation of ions for Ti'v and 0= seems excessive:we shall simply use the notation Ti'v and 0=. The preparation of a zeolite containing ti and silicon was claimed in the patent literature (ref. 6); it is doubtful whether a t.Ltanrim containing zeolite has really been produced. Attempts made in our laboratory to produce the claimed ti tanium containing zeolite according to the informations given in the patents have failed. Only amorphous compounds were obtained.

2. SYNTHESIS OF TS-I The synthesis of TS-I can be carried out with a variety of Si and Ti precursors in the presence of tetraalkylarnnonilnTI bases. A simple and very effective procedure employs

tetraethylorthosilicate, tetraethylortotitanate,

tetrapropylammonilnTI hydroxide and water, in proper ratio and all of high purity (ref. 3): the synthesis does not require mineralizing alkali metals; in fact these can prove detrimental, whi-le the absence of materials other than those essential for the synthesis simplifies very much the purification of the zeolite after the crystallization step. The hydrolysis of the precursors followed by hydrothermal crystallization carried out under autogenous pressure and stirring

415 in the temperature range 160°C .;. 180°C provides the crystalline material which can be separated from the mother liquors, washed and dried; careful calcination to prevent temperatures in excess of 550°C provides the material in the form of a powder which can be used to manufacture the catalyst in the form and dimensions required by the different specific applications.


TS-l acts as a very efficient catalyst in the oxidation of a large number of organic products with H as the oxidant. 202 Table 1 indicates the different reactions studied.



oxidations catalyzed by TS-l

Reactant s





+ "2°2


R<, C=C


@OH 0H

o ©J OH

+ "2°2

+ "2°2







+ "2°2

C-C" C=C










+"P +"P






Aromatic hydrocarbons are hydroxylated to phenols or substituted phenols, phenol itself is hydroxylated to catechol and hydroquinone, olefins are oxidized to the corresponding epoxides, primary alcohols to aldehydes and secondary alcohols to ketones. The interesting aspect of these oxidations is that they can be carried out with the diluted solution of hydrogen peroxide in water (40'~ no loss in selectivity which in almost all cases is higher than

wt H 8C1'~.

with 20 2) Even in the

presence of water propylene can be epoxidized to propylene oxide with selectivity of 98%.


The catalytic activity of TS-1 constitutes the basis for the development of new technologies in different fields, but particularly for the production of chemicals which are obtained through selective oxidation reactions. TS-1 catalyzed oxidations with H could compete with oxidation processes in 202 which alkyl-hydroperoxides are used as oxidants or could constitute an alternative to catalysts in oxidation processes which already use H . Two 20 2 examples will illustrate the case. 4.1 HYDROQUINONE AND CATECHOL

A field in which the catalytic properties of TS-1 have already found industrial application is that of hydrcqutnone and catechol production. In Table 2 the chemical reactions on which the different industrial processes are based for the production of hydroquinone and catechol are indicated. The aniline process has been used by Eastman Kodak in the USA for many years: it produces hydroquinone only: due to the large amount of by-products and the expensive waste treatments required, it may soon be abandoned. The p-diisopropyl-benzene process developed by Signal and Goodyear also produces only hydroquinone. In Europe the three different processes which have been developed by Rhone Poulenc (ref. 12), Brichima (ref. 13) and Enichem (ref. 7) all use H as 202 the oxidant, and all produce both hydroquinone and catechol. The total world production is 50. CXXJ Tons/y, 50';& of which by the H 02 process. No other 2 industrial processes are being operated: catechol has been produced for some time via hydrolysis of o-chlorophenol, but the only existing plant has been closed. A process based on acetylene and carbon monoxide has been brought to the stage of pilot plant by Lanza but not industrialized.





ani line Ea strran




p-dllsopropylbenzene Signal, Goodyear

c2J 0

2CH 3-CO-CH3

hydrogen peroxide

Rhone Poul enc, Brichima, Enichem


The main featurescf the H based processes are given in Table 3. 2 The Rhone Poulenc process uses strong acids as catalyst, like HCl0 or H 3P04; 4 . . the Br-Ichirna process uses a Fenton reagent made of Fe++ and Co++ ; the Eru.chem process uses TS-1 as catalyst. TABLE 3

H Phenol hydroxylation 202 Rhone Poulenc







(TS - 1)

Phenol conversion, %




yield H % 202 Phenol selectivity %










Tars Tars + diphenols



1.4 10

2.3 20

1 12

418 The Enichem process can operate at a mach higher conversion of phenol maintaining an acceptable level of tar fonnation and high selectivities on both H and phenol; the other processes have to limit conversion due to H 202 20 2 decomposition and/or excessive tar formation. High phenol conversions mean substantial reduction in the recycle of unreacted phenol, a very important factor in the overall economics of the process. The data of Table 3 and particularly the high yields on H of the TS-l 20 2 process with respect to the Fe++/Co++ process indicate that the reaction does not proceed through the radical decompoi'5ition of H typical of the Fe ++ /Co ++ 202 catalyzed reaction. The mach higher conversions of the TS-l process with respect to the HCl0

4 catalyzed process at comparable level of tars indicate that the reaction does not proceed through the intermediate formation of peroxonium ion H-O-9-H which is considered the oxidizing agent in the presence of strong acids.


possibility can also be excluded on the basis of the weak acidity of TS-1. In the Enichem process the catechol/hydroquinone ratio is equal to one: this means more hydroquinone with respect to the other H based processes. A 20 2 unique feature of the process is that the catalyst can be modified to change the relative proportions of the two products so that the production can be adapted to changes in market demand (ref. 14). 4.2 PROPYLENE OXIDE The second example selected is that of propylene oxide (PO) production. The chemical reactions on which the different industrial processes are based for the production of propylene oxide are indicated in Table 4 (ref. 15). On a total production of 3.000.000 Tons/y of PO, 1.760.000 Tons are produced according to the chlorohydrin process, 1.340.000 Tons according to the hydroperoxide processes. Tert-butyl-hydroperoxide is the most widely used. In the Daicel process, used in only one 12.000 Tons/y plant, peracetic acid obtained by acetaldehyde oxidation is used. (ref. 16). The H route developed by Ugine Kuhlman (ref. 17) and Enichem (ref. 9) has 202 not been industrialized. A version of the H route has been developed by 20 2 Bayer Degussa (ref. 18): propionic acid is reacted with H to form 20 2 perpropionic acid which is then used for epoxidation of propylene. Also this version has not been industrialized.


Chlorohydrin processes Dow, Basf, Bayer 2HOCI


CH3 -C~-CH2 OH


CH -CH-CH 3 I 12 OH CI Hydroperoxides processes Oxi r-ane , Arco, Shell R-0-0-H

CH -CH=CH 2 3


C -


cat -

/0\ CH -GH-CH 3 2





Hydrogen peroxide Ugine Kuhlman, Enichem cat

---+ The catalysts used in the epoxidation step are both homogeneous and heterogeneous. Shell (ref. 19) has developed the Ti0

on Si0 heterogeneous 2 2 catalyst which contains the same chemical elements as TS-1. I t consists of Ti0

on Si0 and is obtained by the reaction of a compound of Ti with high 2 2 surface area silica, followed by transformation of the Ti compound into the ~2";b

oxide. For TiC1 OH





the reactions can be described as follows:


/CI Ti, 0/ 0 , I



Si <, + '0/ \

TiC1 4



HO "PH 'Ti 0/ '0 I




T' 0/ 1'0 I I -rj(O/Si,\:,

"-Ti=O has been postulated to explain the The formation of the ti tanyl group ./" catalytic activity of TiO/Si0

catalyst in the epoxidation of olefins: its 2 reaction with alkylhydroperoxides would produce surface titaniun alkyl-


compounds which would then react with olefins to produce the

epoxides (ref. 20).








<, /OH Ti



The performances of the TiOz!Si0

catalyst in hydroperoxide epoxidation of 2 propylene are improved upon silylation with silylating agents like Cl-Si-(CH 3)3 (ref. 21): most likely the silylating agents react with the Si-OH groups present at the surface of the silica which are responsible for secondary reactions of the PO initially produced. Also for TS-1 in H epoxidation of propylene an improvement in selectivity 202 has been observed upon silylation (ref. 22). l-'lwe the secondary reaction is mainly hydration of PO to propylene glycol, water being the only by-product present. The results for H epoxidation with TS-1 and for EthylBenzeneHydroPeroxide 202 epoxidation with TiO/Si0 and the effect of silylation on the two catalysts 2 are given in Table 5. Table 5.

Epoxidation of propylene with H and EBHP 202

Oxidant Catalyst



H 202 H 202 Conversion



PO Selectivity



1.5 TiO/Si0 silylated 2 (ref. 20) TS - 1



TS - 1 silylated (ref.22)



EBHP Conversion


PO Selectivity






The selective epoxidation of olefins with H has been considered a major 202 challenge for homogeneous and heterogeneous catalysis: many different approaches have been tried. The results obtained with TS-1 demonstrate that the solution of the problem has been found. The drawbacks which have been considered to constitute


an obstacle to the industrial use of H can in fact be easily overcome: lower 20 2 rates of reaction which have been observed for polar solvent can be easily canpensated by longer reaction times, which in the case of heterogeneous catalytic processes simply require the use of higher catalyst/feed ratio; secondary hydration of epoxides to glycols is negligible on silylated TS-1. The major differences between H and alkylhydroperoxide processes are the 20 2 availabili ty of rEM materials and the production of by-products . Availability of H at a convenient price and guarantees concerning future supply must of 202 course be secured before any engajement in industrial plant is taken. The same is true for the hydrocarbons used to produce the hydroperoxides, ethylbenzene and iso-butane. For every Ton of PO, 2.5, Tons of methylphenylcarbinol or 3 Tons of tert-butylalcohol (TEA) are coproduced. The carbinol is generally dehydrated to styrene. TEA constitutes a special case. Until recently (ref. 23) it had found an outlet in the oxygenated octane boosters market, as a mixture with methanol (OXINOL), but more recently decreased prWes of oil and the preference of refiners for more effective non alcoholic octane boosters like MethylTertiaryButylEther (MTBE) have decreased interest in this use. AReO has recently announced that it will convert TEA into isobutene and will use this for MTBE production. The elimination of lead additives in gasoline has created an enormous market demand for MTBE, whose production has increased from almost nothing to 5 Million Tons/y in 1986 and is expected to grow to 10 Millions Tons/y in 1990. These figures are so large when compared to chemicals that no risk of having excess TEA exists: however, since the isobutene necessary for METE production can also be obtained, and is in fact obtained, through the simple dehydrogenation of isobutane, the market value of TEA (or isobutene) co-produced must be competitive with this al ternative process. However, even

though the risk of unsold by-product is

unlikely when tertiarybutylhydroperoxide is used, there is no doubt that larger investments are required for both the production of the hydroperoxide and the transformation of the by-product TBA into MTBE. It seems appropriate to say that by-products formation in certain cases can be accepted even in large industrial processes provided a large enough market is at hand, and the overall economics of the process are satisfied by a value of the by-product competitive with alternative processes ,


Ti tanium has a stable valence of 4 and in an oxidizing medium it is very likely that this valence is maintained. An examination of the chemistry of


compounds irrrnediately shows that Ti'"

has a strong tendency to assume a high coordination number: with oxygen, six groups in octahedral coordination form a stable and very frequently observed configuration, but to do this Ti


must have near neighbours capable of

increasing their coordination number to satisfy at the same time titanium valency of four and coordination of six. When bulky groups are linked to


tetrahedral coordination is also observed. Coordination of seven in a pentagonal pyramidal arrangement like in peroxo compounds and of eight like in Ti(N0 are also observed.


From the crystalline structure and the regular change in uni t cell parameters which are consistent with isomorphous substitution of

si" with

Ti lv it seems

justified to represent T8-1 as a silicalite in which few Ti lV have taken the place of 8i IV. The interpretation of the catalytic activi ty of T8-1 must take into consideration the role played by these few Ti 'v: in fact pure silicalite is totally inactive, and other phases containing Ti have not been identified. Due to the fact that T8-1 crystallizes from a homogeneous solution, it is reasonable to assume that the distribution of Ti 'v in the crystal lattice is at random':' since the 8i / Ti most Ti


ratio is in the range 40 + 90 in typical preparations,

must be isolated from each other by long sequences of -0-8i-0-8i-0- .

If Ti'v replaces a 8i'v it should be tetrahedrally coordinated by 0 -1

the presence of a band at 980 cm

: however,

closely corresponds to the band observed in

other titanium compounds containing the . ,"Ti = 0 group, whose streching / -1 frequency is 975 cm with bond distances of 1,66 + 1,79 A ; furthermore, 0

hydroxyl groups are present at the surface as shown by the increase in selectivity which is obtained upon silylation. Finally, near neighbour

positions ofTi 'v are occupied by

si'" which

in a field of

0- is stable only in tetrahedral coordination. A simple representation of the sites where substitution has occurred which takes into consideration the various pieces of experimental evidence could be (3)


Other more elaborated and detailed representations could be given, should the present model prove inadequate to interpret all experimental facts. Til>' in TS-1 maintains the strong affinity of soluble


salts for H and 202 gives rise to a strong yellow colour which can

in fact the addition of H 0 2 2 be attributed to the formation of surface titaniumperoxocanpounds which can be in the hydrated or dehydrated form OH

°<:>OH /Si, °


°"'-,OH 0"'-/ -0 OH ° \/ °/ Si , '0 /'Ti" 0 / Si, 0

° /Si, °








and which constitutes the actual oxidants. Work carried out on Mo(VI) and W(VI) peroxocanpounds (ref. 24) has demonstrated that peroxocanpounds can act as oxidants in stoichiometric epoxidations involving a nucleophilic attack of the substrate to the peroxidic oxygen: in the presence of excess H the peroxo compound is regenerated and this 20 2 accounts for the catalytic nature of the reaction. It seems reasonable to assume that a similar mechanism operates in the case of Ti(IV) peroxocanpounds. The relevance of isolated Ti'Y and the connection with catalytic performances appears to hold also for the TiO/Si0 selectivities are obtained when Ti0

catalyst. In fact high epoxide 2 is distributed on high surface area Si0

2 2 and its concentration is limited to 2% (ref. 20). It is very likely that at this low concentration Ti'Y are isolated and surrounded by

sr'" .



or Ti0 alone, or physical mixtures of Si0 and Ti0 or various metal 2 2 2 2 titanates exhibit no significant activity. Similarly, supporting Ti0 on 2 carriers different from Si0 like Al ' MgO or zr0 leads to catalysts 203 2 2 whose activity is lower or nil. One piece of evidence seems very convincing: when the Ti0

concentration on the catalyst is reduced from 4% to 0.4%, all 2 other conditions being equal, an increase in epoxide selectivity is obtained. The only effect that a reduction in the concentration of Ti0 increase in the degree of dispersion of each



rs " :

can have is an 2 chances for each rr" of

as near neighbours increase, as does the selectivity of the catalyst.

The correlation between the isolated Ti'v and selectivity of the catalyst in epoxidation could be due to the fact that on


having other


as near

neighbours, a mechanism proceeding through a bimolecular interaction of surface peroxo species could be operating





which would give rise to a high decomposition rate of H or hydroperoxides 202 to 02 . This mechanism could not operate on perfectly isolated T{v. Of course low decomposition of H


(or hydroperoxides as well) means greater

stabili ty of titanium peroxo compound whose reduction can only be carried out by the organic substrate with increased yields of usefUl oxidized products. When the different results between TS-l and Ti0


in the hydroxylation of

phenol are analyzed the existence of a "restricted transition state selectivity" must be assumed to explain the small amount of tars formed. 6. CONCWSIONS

The discovery of TS-l and of its catalytic properties in oxidation of organic compounds with H has opened the way to new technological possibilities and 202 has demonstrated that H can be a very selective oxidizing agent when proper 202 catalysts and conditions are used. This is of remarkable scientific interest and has al.ready proved of industrial interest in one case. When other industrial applications are considered, many factors must be taken into consideration which have a large Irrpact on the final econanic evaluation: availability and prices of raw materials and of H ' value of by-products, dimensions and perspectives 202 of the different by-products applications, integration with other productions, investments required, price trends of the different products in the future. The results could well be different from one case to another, and H 202 processes could prove superior in quite a number of them. REFERENCES

1. M. Taramasso, G. Pe.rego and B. Notari, in L.V.C. Rees (Ed) Proc. Fifth

Int. Conf. on Zeolites, Naples, 1980, London Heyden and Sons, p. 40


2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

M. Taramasso, G. Manara, V. Fattore and B.Notari, U.S. Pat. 4,656,016 (1987) M. Taramasso, G. Perego and B. Notari, U.S. Pat. 4,410,501 (1983) M. Taramasso, G. Manara, V. Fattore and B. Notari, U.S. Pat.4,666,692 (1987) G. Perego, G. Bellussi, C. Como, M. Taramasso, F. Buonomo, A. Esposito, in Y. Murakami, A. Iijima, J .W. Ward, (Eds.), Proc. Seventh Int. Conf. on Zeolites, Tokyo, 1986, Tonk Kodansha Amsterdam Elsevier, p. 129 D.A. Young, U.S. Pat. 3,329,481 (1967) A. Esposito, M. Taramasso, C. Neri and F. Buonomo, U.K. Pat. 2.116.974 C. Neri, A. Esposito, B. Anfossi and F. Buonomo, Eur. Pat. 100.119 (1984) C. Neri, B. Anfossi and F. Buonomo, Eur-, Pat. 100.118 (1984) F. Maspero and U. Romano, Pat. Pend. A. Esposito, C. Neri and F. Buonomo, U.S. Pat. 4,480,135 (1984) J. Varagnat, Ind. Eng. Chem. Prod. Res. Dev.Vol.15, N° 3, 1976, p.212 P. Maggioni and F. Minisci, La Chim. e l'Ind. 59, 4 (1977) p. 239 G. Bellussi, A. Giusti, A. Esposito and F. Buonomo, Eur. Pat. 202269-A Inf. Chimie 247, March 1984, p. 131 Hydr. Proc. Nov. 1981, p. 225 J. P. Shirmarm and S. Y. Delavarenne, Hydrogen Peroxide in Organic Chemistry, Ed. Documentation Industrielle, Paris, 1980, p. 45 Hydr. Proc. Nov. 1981, p. 223 Brit. Pat. 1.249.079 (1971) to Shell Oil R.A. Sheldon, J. Mol. Cat. 7, (1980), p. 107 H.P. Wulff, U.S. Pat. 3,923,843 (1975) to Shell Oil M.G. Clerici, U. Romano, Eur. Pat. 230949-A I. Pasquon in Catalysis today, Vol. 1, N° 3, 1987, p. 303 G. Amato, A. Arcoria, F.P. Ballistreri, G.A. Tomaselli, O. Bortolini, V. Conte, F. Di Furia, G. Modena and G. Valle, J. Mol. Cat. 37 (1986)p. 165