JOURNAL OF CATALYSIS 93, 30-37 (1985)

The Role of Different Types of Acid Site in the Cracking of Alkanes

on Zeolite Catalysts
A. CORMA,*J J. PLANELLES,~ J. SANCHEZ-MARfN,t AND F. ToMAst
*Institute de Cat&is y Petroleoquhica, CSIC, Serrano 119, Madrid, 6 Spain and TDepartamento
Quimica-Fisica, Facultad de Ciencias Quimicas, Avd. Dr. Moliner, Burjasot, Valencia, Spain

Received October 17, 1983; revised June 3, 1984

The cracking of n-heptane on a rare-earth (RE)HY zeolite has been studied in the temperature
range 400-470°C. By measuring the initial selectivities with respect to the various reaction prod-
ucts, and carrying out molecular orbital calculations it has been found that cracking can initially
take place on Bronsted acid sites via protolytic cracking, or on Lewis acid sites via a classical p-
scission mechanism. The activation energy for protolytic cracking is lower than for p-scission both
for n-heptane and for the iso-heptanes. From energetic considerations, a scheme of cracking can be
suggested where the cracking proceeds by hydride transfer and p-scission mainly via branched
heptanes. 0 1985 Academic Press, Inc.

INTRODUCTION A pentacoordinated carbocation is

formed via protonation of the alkane by a
It is generally accepted that the catalytic strong Bronsted site (13).
cracking of alkanes takes place by a carbe- The initial carbenium ion is formed via
nium ion-type mechanism. A matter of dis- protonation of an olefin generated by ther-
cussion is the first step of the reaction, i.e., mal cracking or present as an impurity in
the formation of the carbenium ion. In the the feed (27-19).
case of cracking with superacids in the liq- The electric fields associated with the
uid phase, it has been shown that the initiat- cations can be responsible for the polariza-
ing step occurs by protonation of the alkane tion of the reactant molecules (20).
to give a carbonium ion (1-5). In the case of Several studies have produced partial
heterogeneous catalysis with solid super- proofs for one or another mechanism for
acids, Hattori et al. (6, 7) have proposed the initiation of the cracking of alkanes.
that an initial carbenium ion is formed by Thus, it has been observed that the OH
abstraction of a hydride ion from the alkane groups localized in the supercage of a HY
by a Lewis acid site of the catalyst. When zeolite are consumed during the catalytic
zeolites have been used as cracking cata- cracking of alkanes (21,22). Anufriev et al.
lysts several hypotheses for the initiation (23) have seen an increase in the rate of
step have been reported: cracking of alkanes when olefins were in-
A carbenium ion is formed via abstrac- troduced into the feed. This result agrees
tion of a hydride ion by a Lewis site (tri- with the observations of Aldridge et al.
coordinated aluminum and/or AlO+-type (22). These authors have shown that there
species) (8-12). is an induction period for the cracking of n-
A carbenium ion is formed via abstrac- hexane on a LaY zeolite, and that the in-
tion of a hydride ion by a strong Bronsted duction period is inversely proportional to
site (13-16). the reaction rate.
L6pez Agudo et al. (24) have studied the
r To whom correspondence should be addressed. cracking of n-heptane on a CrHNaY zeolite
30
0021-9517/85 $3.00
Copyright Q 1985 by Academic Press. Inc.
All rights of reproduction in any form reserved.
 ACID SITES IN CRACKING OF ALKANES 31

and obtained the initial selectivities with re-allowed mass balances to be made, and
spect to the various reaction products in the only experiments with mass balances of 100
absence of decay. Under these conditions + 5% were considered.
they observed that the paraffin/olefin (P/O) In a typical run, the reactor was initially
ratio is higher than one, while aromatics purged with N2 for 20 min. An amount of
and coke were found as secondary prod- 10.0 ml of n-heptane (6.88 g) was pumped
ucts. Therefore, a source of hydrogen is through the reactor at a constant rate, over
needed to explain their results. a given weight of catalyst (given cat/oil ra-
In the present work the mechanism for tio (P)), for a total time tf = 2.5 min. Addi-
the initiation and development of paraffin tional data up to tf = 50.0 min were ob-
cracking has been studied by investigating tained at a constant P value by changing the
the cracking of n-heptane on a (RE)HY ul- rate at which the same quantity of reactant
trastable zeolite and by carrying out molec- was pumped through the same packing.
ular orbital calculations to simulate the re- Various cat/oil curves at different values of
actions postulated on the basis of the P were obtained by repeating the process
experimental results. with different weights of catalyst in the re-
actor. At the end of each run, the reactor
EXPERIMENTAL
was purged with N2 and all liquid and gas-
Materials. The catalyst was prepared by eous products were analyzed by gas chro-
repeated exchange of a NaY zeolite (SK- matography on a SE-30 silicone on
40, WA1 = 2.25), obtained from Linde Chromosorb P column and a Porapak Q-
Company, with a solution containing La3+ silica gel column, respectively, using a HP-
and Ce3+ ions in a ratio of 3 to 1, and subse- 5840A chromatograph. In our conditions
quently with a solution containing NH: and with the gas mixtures analyzed, hydro-
ions. The Na+ form was stirred in a La3+ + gen could be detected in concentrations of
Ce3+ solution for 6 h, and then filtered and 100 ppm. The analysis of the coke depos-
washed. The filtrate was dried at 110°C for ited on the surface of the catalyst was car-
12 h and calcined at 550°C for 2 h. This ried out using air as the oxidant and mea-
procedure was repeated until 80% of the suring the amount of Hz0 and CO* formed.
original Na+ content was exchanged. Sub- MO Calculations. A semiempirical ap-
sequently, the remaining Na+ was ex- proach to the MO formalism has been
changed by NH: as above. Thereafter the adopted to deal with the size of the mole-
zeolite was pelletized, crushed, and cule studied. The MIND0/3 method (25)
steamed at 500°C for 3 h. has been selected because of its usefulness
The n-heptane used was a Carlo Erba for the study of various properties of or-
99.9% purity reactant. ganic ions, both classical and nonclassical,
Procedure. The experiments were car- in the gas phase, mainly those concerned
ried out in a fixed-bed tubular reactor. The with calculations of energies of reactions
reactant was charged at the top of the reac- involving fragmentations (26). MIND0/3
tor by means of a pump, feeding at a con- calculations have proved useful in a great
stant rate. To minimize the thermal effects variety of cases in giving the correct order
of the cracking reaction the catalyst was of magnitude of the energies of alternative
diluted to a constant volume using ground transition states and hence they can be
glass of the same mesh size as the catalyst. helpful in determining reaction mechanism
Condensed liquid products from the reac- (27). All the geometrical optimizations
tion were trapped in a pot, after passing were performed using the variable metric
through a condenser located below the re- method described by Murtagh and Sargent
actor, while the gaseous products were (28) and adapted by Rinaldi (29) for semi-
trapped in gas burettes. The closed system empirical energy calculations.
32 CORMA ET AL.

the large number of atoms needed to repre-

sent a significant fragment of the surface of
the catalyst. In any case, the energies in-
volved in intermolecular interactions are
much smaller than those involved in the
FIG. 1. Starting points adopted for the search of the
possible minima of protonated n-heptane. See text for breaking of chemical bonds. We think,
details. therefore, that these interaction energies
would only introduce small quantitative
modifications in our results.
The stationary points on the MIND0/3
energy hypersurface of the protonated n- RESULTS AND DISCUSSION
heptane have been studied. The n-heptane
A classical P-cracking mechanism taking
molecule was arranged in a symmetrical
place over an acid site gives a P/O (paraffin/
conformation in order to reduce the number
olefin) ratio of unity in the reaction prod-
of starting points to be considered. Three
ucts. In the case where the initiation of
main protonation schemes were adopted:
cracking occurs via olefins, that ratio would
(1) protonation in the middle of a C-C
also be unity. The behavior of the P/O ratio
bond, (2) protonation in the middle of a C-
may be observed experimentally by merely
H bond (A in Fig. l), and (3) protonation
looking at the values of the initial selectivi-
over a C atom to yield a pentacoordinated
ties with respect to the various reaction
structure. Two classes of starting point
products, since under these conditions the
were taken for protonation over the C-C
results are not disguised by consecutive re-
bond, viz. (i) placing the H+ just over the
actions such as, for instance, hydrogen
bond (0 in Fig. l), or (ii) placing it to form a
transfer. We have obtained here the selec-
bent-bridged structure (0 in Fig. 1). Simi-
tivity curves for the several groups of prod-
larly, two types of attack of a C atom were
ucts formed, in the absence of decay and
considered, viz. (i) over the H-C-H face
thermal cracking (Fig. 3). From these
(0 in Fig. l), or (ii) on top of the H-C-C
curves the values of the initial selectivities
face (0 in Fig. 1). The complexes with a
have been calculated at different levels of
linear C-H-H structure have been shown
conversion. In Fig. 3g one sees that the P/O
to be less stable than other structures (30)
ratio is significantly greater than unity for
and will not be considered here. All starting
low conversion levels (<2%), and it de-
points which have been considered are pre-
creases as the level of conversion in-
sented in Fig. 1.
creases. Obviously, the high P/O ratio
The two-dimensional cross section of the
found cannot be explained only in terms of
cracking hypersurface was obtained by in-
a cracking mechanism through Lewis acid
dependently scanning the distances dl (C-
sites, since some kind of hydrogen supplier
C) and d2 (C-H) (see Fig. 2), starting from
would be needed. Possible hydrogen do-
protonated n-heptane structures of the lin-
nors are cycloalkanes, while producing aro-
ear bridge type (C-H-C). The molecular
orbital study corresponds to the transfor-
mation of an isolated protonated hydrocar-
bon into different cracking products. A bet-
ter understanding of the process would be
obtained if the interaction energy between
the protonated hydrocarbon and catalyst A B C

could be calculated. However, MO calcula- FIG. 2. Minima found in the hypersurface corre-
tions on such an interacting model are not sponding to n-C7H&. Only that part of the n-heptane
feasible at the MIND0/3 level, because of skeleton related with protonation is depicted.
 ACID SITES IN CRACKING OF ALKANES 33

a strong molecular complex between Hz

4I (1 LI and a heptyl carbenium ion. The structures

IM2t&
0 4 12 20 28 36 0 4 12 20 28 36
B and C represent a hydrogen atom bridged
between two carbon atoms, forming a linear
C-H-C bridge or a bent-bridge structure.
Such structures are similar to those re-
ported for small carbonium ions on the ba-
$4I c 12
t
d sis of accurate ab initio calculations (31).
$61 / 81 The eight stationary energy structures have
i8w, , kr; been obtained and are summarized in Fig.
4.
0 4 12 20 28 36 0 4 12 20 28 36 The first possibility of interaction is the
attack of one proton from the catalyst on a
C-H bond. It produces species which have
been postulated during the isomerization
and cracking of alkanes in superacid media
(complex A) (1-5). It is reasonable to ex-
pect that such an intermediate complex will
0 4 12 20 28 36 0 4 12
Total cLm"erSlO" 1%)
20 28 36 lead to the formation of H2 leaving a carbe-
nium ion. Indeed, Olah et al. (I-5) have
reported the formation of molecular hydro-
gen during the cracking of alkanes in the
presence of superacids in the liquid phase.
In the case of zeolites, the formation of
hydrogen via the abstraction of a hydride
ion from the hydrocarbon by means of a
FIG. 3. Selectivity curves to the different groups of Bronsted acid site has been postulated by
products at 400°C: (a) Cz + C,; (b) C$- + C:-; (c) C3 + Minachev et al. (32) to explain the effect of
C,; (d) C:- + C:-; (e) aromatics; (f) coke; (g) P/O ratio H2 partial pressure on the rate of the isom-
of the reaction products at: (--) 400°C; (---) 430°C; erization of alkanes in the presence of a ze-
(.-.-.) 470°C; (0) 0.07, (0) 0.146, (A) 0.29 cat/oil (g/g).
olite. However, in our work, we have not

matics, and coke. However, under our

experimental conditions (Fig. 3) both
aromatics and coke appear as secondary
products, and therefore they cannot be con-
sidered responsible for the high P/O ratio
found in the cracked products at very low
levels of conversion. The other possible
sources of hydrogen are the OH groups of
the catalyst. If this were so, then it would
be necessary to postulate, for the first step
of the cracking reaction, an interaction be-
tween the n-heptane and the Bronsted acid
centers of the catalyst. We have simulated
that interaction, and three types of minima
have been found in the MIND0/3 surface of FIG. 4. Structures and energies in kcal mol-I asso-
protonated n-heptane, labeled as A, B, C ciated with the various minima in the hypersurface of
(see Fig. 2). The A type can be described as protonated n-heptane.
34 CORMA ET AL.

structures (see Fig. 4) correspond to linear

bridge C-like structures where the hydro-
gen atom is placed between two secondary
carbons. Other structures such as those la-
beled 1 and 2 (Fig. 4) show calculated
MIND0/3 energies that are similar to those
of the more stable bridged structures 6 and
7. However, Hiraoka and Kebarle (36-38)
have shown experimentally from the study
of the reaction surface of CjH$ that the C-
C breaking process starts from structures
type B or C, i.e., bridge-like structures,
while MIND0/3 calculations lead to the
same conclusion (39). Thus we have per-
formed our cracking surface study starting
from the more stable linear bridged struc-
ture. In Figs. 5 and 6 are shown the two-
dimensional cross sections of the energy
hypersurfaces corresponding to the struc-
tures labeled as 6 and 7 in Fig. 4. The two

FIG. 5. Two-dimensional cross section of the energy

surface for the cracking process of structure 7 (Fig. 4)
leading to C3 and Cd. Each curve from 1 to 7 corre-
sponds to a one-dimensional cross section associated
with a fixed value of d, in the range from 2.5 to 3.7 A,
with a scanning step of 0.2 A. Dashed lines indicate the
minimum energy reaction path. ET is the molecular
energy. AE = 17.5 kcal . mol-’ is the energy barrier for
cracking to CT + C,; A.& = 13.5 kcal . mol-r is the
energy barrier for cracking to C, + Ci.

seen hydrogen as a reaction product during

the cracking of n-heptane, and in those
cases where hydrogen has been reported as
a product (33,3#) the results were obscured
by thermal cracking (35). In any case a
mechanism involving Bronsted acid sites as
hydride ion abstracting agents will produce
a carbenium ion which, by cracking, will
yield a theoretical P/O ratio of unity, which
we did not observe in our experiments.
The other possibility for the interaction FIG. 6. Two-dimensional cross section for the crack-
of a Bronsted acid with an alkane molecule ing process of structure 6 (Fig. 4), leading to C2and Cs.
& is the molecular energy. AE, = 15 kcal . mol-‘, is
is the attack of a proton on a C-C bond of the energy barrier for cracking n-heptane into Cl +
the alkane to give structures of the type B C,; hE2 = 12.5 kcal . mol-I is the energy barrier for
and C bridge (see Fig. 2). The more stable cracking into Cz + Cl.
 ACID SITES IN CRACKING OF ALKANES 35

all mechanism can be expressed in a simpli-

fied way by the equations
ZH + Cj*CjH+ . . . Z-
CjH+ . . . z- ---, c, + c; . . . z-
(j = m + n) Protolytic cracking
c; . . . z-+cj-+c,+c,~. . . z-
Hydride transfer
c; . . . z- + c:- + c: . . . z-
(j = 1 + S) P-cleavage
Reoctlon coordmate

In the scheme ZH represents an acid zeo-

FIG. 7. Energetic profiles associated with the follow-
ing processes: (1) H- transfer from (CH&CH andlite and Cj is a paraffin, wherej is the num-
ber of carbon atoms; C, and CT- are alkanes
(CH,),C+ (Ref. (40)). (2) Protolytic cracking to give Cj
and olefines , respectively.
+ Cd. (3) /3-Cracking of i-heptane to give C-C+-C +
nC4 (Ref. (40)). (4) P-Cracking of n-heptane to give
It is clear that with the proposed mecha-
C+-C-C + nC4 (Ref. (40)). nism it is not necessary to invoke the for-
mation of H2 as a reaction product. Fur-
parameters dl and d2 have been taken as thermore, by looking at Fig. 7, one sees
independent. From those figures, one sees that the process which involves the lowest
that the energy barrier for the protolytic activation energy is the hydride transfer be-
cracking process is in the range of 12 to 18 tween an adsorbed carbenium ion and a
kcal . mol-‘. Moreover, when the bridged paraffin, followed by the protolytic crack-
complex is formed it is much easier to crack ing which requires a lower activation en-
than to interconvert from minima of the ergy than the P-cracking via carbenium
bridged species to those of complex ones, ions, either directly from n-heptane or
since the latter process shows an energetic through the iso forms.
barrier higher than either the protolytic These results indicate that from an en-
cracking or dehydrogenation (36-39). ergy point of view the protolytic cracking,
Therefore, in this second case a carbonium together with P-cracking through n-heptane
ion is formed that can crack to produce a isomers, should be the process responsible
paraffin, which desorbs into the gaseous for the first cracking events and the forma-
stream, and an adsorbed carbenium ion. tion of the surface covered with carbenium
Thus, a surface covered with carbenium ions. When such a surface is formed, the
ions is generated. The carbenium ions will reaction will proceed by P-cracking, mainly
keep the reaction going by creating new CT via the branched isomers of heptane.
carbenium ions through a hydride transfer It should be noted that Olah et al. (1-5)
reaction have reported the formation of H2 when su-
peracids are used as catalysts in homoge-
C&<7) + c7 + cl + c: nous liquid phase, where there are hydro-
gen atoms providing a very low electron
The heptyl carbenium ions produce, by p- density. On the other hand, Hattori et al.
scission, one olefin and a new adsorbed car- (6, 7) have reported that in the presence of
benium ion. This mechanism may explain heterogeneous superacids no evolution of
not only the high P/O ratio found at low Hz was observed. These authors inter-
conversion levels, but also its decrease (for preted that feature on the basis of the lower
still lower levels of conversion) when in- acidity of the Bronsted acid sites on hetero-
creasing the level of conversion. The over- geneous catalysts compared with homoge-
36 CORMA ET AL.

neous superacid catalysts. Hence it can be bond being formed will be weaker than in
considered that in the heterogeneous case the first case, i.e., the H2 structure will not
the Bronsted sites provide an electronic be clearly established in the case of zeolite
density considerably larger than in the ho- catalysts. If this is true, the formation of A-
mogeneous case. like structures cannot be considered to be
In the case of zeolites, the acid strength likely; we propose for this case, complexes
of the Bronsted sites is lower than that of of the type
superacids (42), and as indicated before, we
have not detected Hz in the products. Thus,
in the case of our zeolite catalyst the elec-
tronic charge transfer from the C-H bond
to the H-H bond, which is being formed,
will be less important than for the case of cat.
superacid catalysts. In other words, in the The strength of the acid sites in zeolites
case of zeolites the C-H bond remains would be able to produce B- or C-like struc-
stronger than for superacid homogeneous tures which can give protolytic cracking in
catalysts, and consequently the new H-H consecutive steps:

+ Alkane

In conclusion it can be said that cracking takes place through two parallel pathways:

PVX. t hyd. t
CARBONIUM/ION- ALK.+CARBENIUM/ION- ALK.+ CARBENlUM/lON
H+ Crack. Tranrf.
f
ALKANE

&scission
CARBENlUM/lON
h *
m
+

It is obvious that in the first stages of the decrease in the P/O ratio obtained at short
reaction the product distribution should be contact time and times on stream. Indeed,
dependent on the ratio carbonium ions/car- such behavior is observed in Fig. 3g.
benium ions or, similarly, on the Brgnsted/
Lewis site ratio.
Following the reaction scheme presented Note added in proof: While this paper was in press,
above, and taking into account the values Haag and Dessau (W. D. Haag, R. M. Dessau, “Pro-
ceedings 8th International Congress on Catalysis,
of the activation energies for the different Berlin, July 1984.” H-305) also proposed the possibil-
processes, it is obvious that an increase in ity of protolytic cracking of 3-Methylpentane on zeo-
the reaction temperature would produce a lite catalysts.
 ACID SITES IN CRACKING OF ALKANES 37

REFERENCES Congress on Catalysis, Amsterdam, 1964.” p.

1264. North Holland, Amsterdam, 1965.
1. Olah, G. A., .I. Amer. Chem. Sot. 94, 808 (1972). 21. Bolton, A. P., and Bujalski, L. R., J. Cutal. 23,
2. Olah, G. A., DeMember, J. R., and Shen, J., J. 331 (1971).
Amer. Chem. Sot. 95, 4952 (1973). 22. Aldridge, L. P., McLaughlin, J. R., and Pope, C.
3. Olah, G. A., Shen, S., and Schlosberg, R. H., J. G., J. Cut&. 30, 409 (1973).
Amer. Chem. Sot. 95, 4957 (1973). 23. Anufriev, D. M., Kuznetsov, P. N., and Ione, K.
4. Olah, G. A., Halpem, Y., Shen, J., and MO, Y. G., React. Kinet. Catal. Lett. 9, 297 (1978).
K., .I. Amer. Chem. Sot. 95, 4960 (1973). 24. Lopez Agudo, A., Asensio, A., and Corma, A., .I.
5. Olah, G. A., MO, Y. K., and Olah, J. A., J. Amer. Cutul. 69, 274 (1981).
Chem. Sot. 95, 439 (1973). 25. Bingham, R. C., Dewar, M. J. S., and Lo, D. M.,
6. Hattori, H. Takahaski, 0, Tagaki, M., and J. Amer. Chem. Sot. 97, 1294, 1302, 1311
Tanabe, K., J. Cutal. 68, 132 (1981). (1975).
Dewar, M. J. S., and Rzepa, H. S., J. Amer.
7. Takahashi, O., and Hattori, H., J. Catal. 68, 144 26.
(1981). Chem. Sot. 99, 7432 (1977).
27. Flanigan, M. C., Komornicki, A., and McIverin,
8. Tung, S. E., and McIninch, E., J. Cat&. 10, 166 J. W., in “Modem Theoretical Chemistry” (G. A.
(1968).
Segal, Ed.). Plenum, New York, 1977.
Nate, D. M., Ind. Eng. Chem. Prod. Res. Dev. 8, 28. Murtagh, B. A., and Sargent, R. W. H., Comput.
31 (1969). J. l3, 185 (1970).
Borodzinski, A., Corma, A., and Wojciechowski, 29. Rinaldi, D., Comput. Chem. 1, 101 (1977).
B. W., Cunud. J. Chem. Eng. 58, 219 (1980). 30. Corma, A., Sanchez, J., and Tomas, F., J. Molec.
11. Uytterhoeven, J. B., Christner, L. G., and Hall, Curd. 19, 9 (1983).
W. K., J. Phys. Chem. 69, 2117 (1965). 31. Raghavachari, K., Whiteside, R. A., Pople, J. A.,
12. Kerr, G. T., in “Molecular Sieves” (W. M. Meier and Schleyer, P. von R., J. Amer. Chem. Sot.
and J. B. Uytterhoeven, Eds.), Adv. Chem. Ser., 103, 5649 (1981).
Vol. 121, p. 219. Amer. Chem. Sot., Washington, 32. Minachev, K. H., Garanin, V., Isakova, T.,
D.C., 1973. Karlamov, V., and Bogomolov, V., Adv. Chem.
Greensfelder, B. S., Voge H. H., and Good, Cl. Ser. 102, 441 (1971).
M., Ind. Eng. Chem. 49, 742 (1957). 33. Good, G. M., Voge, H. H., and Greensfelder, B.
14. Janardhan, P. B., and Rajeswari, R., Ind. Eng. S., Znd. Eng. Chem. 39, 1032 (1949).
Chem. Prod. Res. Dev. 16, 1 (1977). 34. Miale, J. N., Chen, N. Y., and Weisz, P. B., J.
15. Poutsma, M. L., in “Zeolite Chemistry and Catal- Cutul. 6, 278 (1966).
ysis” (J. A. Rabo, Ed.), No. 171, p. 505. Amer. 35. Corma, A., and Wojciechowski, B. W., Prepr.
Chem. Sot., Washington, D.C., 1976. Amer. Pet. Div., August 1983.
36. Hiraoka, K., and Kebarle, P., J. Chem. Phys. 63,
16. Gates, B. C., Katzer, J. R., and Schuit, G. C. A.,
394 (1975).
“Chemistry of Catalytic Process,” p. 31. Mc-
37. Hiraoka, K., and Kebarle, P., Cunud. J. Chem.
Graw-Hill, New York, 1979.
53, 970 (1975).
Greensfelder, B. S., Voge, H. H., and Good, G. 38. Hiraoka, K., and Kebarle, P., J. Amer. Chem.
M., Ind. Eng. Chem. 41, 2572 (1949). Sot. 98, 6119 (1976).
18. Weisz, P. B., Annu. Rev. Phys. Chem. 21, 175 39. Planelles, J., Sanchez-Mar&, S., and Tomas, F.,
(1970). Theo&em. 17, 65 (1984).
19. Scherzer, J., and Ritter, R. E., Ind. Eng. Chem. 40. Planelles, J., Sanchez-Martn, J., Tomas, F., and
Prod. Res. Dev. 17, 219 (1978). Corma, A., J. Chem. Sot. Perkin II, in press.
20. Pickert, P. E., Rabo, J. A., Dempsey, E., and 41. Jacobs, P. A., “Carboniogenic Activity of Zeo-
Schomaker, V., “Proceedings, 3rd International lites.” Elsevier, Amsterdam, 1977.