Microcalorimetric investigation of mordenite and Y zeolites for 1-methylnaphthalene isomerisation

Article abstract of DOI:10.1039/a803931c

Adsorption microcalorimetry of pyridine has been used to determine the concentration and the distribution of strengths of the acid sites of Y zeolites and mordenite samples. 1-Methylnaphthalene was reacted over these samples at 623 K under atmospheric pressure in a continuous-flow fixed-bed reactor. For Y zeolites the relative extents of 1-methylnaphthalene isomerisation and disproportionation have been found to depend on the density of the acid sites. Coke growth favours isomerisation over disproportionation as time-on-stream increases. However, both disproportionation and coking are highly disfavoured in the channel system of mordenites, in spite of the relatively high density of the acid sites. The accumulation of strongly adsorbed feed and/or product molecules is responsible for the severe decay in activity of non-dealuminated mordenite.

Full text of DOI:10.1039/a803931c

View Article Online / Journal Homepage / Table of Contents for this issue
Microcalorimetric investigation of mordenite and Y zeolites for
1-methylnaphthalene isomerisation
I. Ferino,*¤ R. Monaci, E. Rombi and V. Solinas
Dipartimento di Scienze Chimiche, Universita di Cagliari, V ia Ospedale 72, 09124 Cagliari, Italy
Adsorption microcalorimetry of pyridine has been used to determine the concentration and the distribution of strengths of the
acid sites of Y zeolites and mordenite samples. 1-Methylnaphthalene was reacted over these samples at 623 K under atmospheric
pressure in a continuous-Ñow Ðxed-bed reactor. For Y zeolites the relative extents of 1-methylnaphthalene isomerisation and
disproportionation have been found to depend on the density of the acid sites. Coke growth favours isomerisation over dispro-
portionation as time-on-stream increases. However, both disproportionation and coking are highly disfavoured in the channel
system of mordenites, in spite of the relatively high density of the acid sites. The accumulation of strongly adsorbed feed and/or
product molecules is responsible for the severe decay in activity of non-dealuminated mordenite.
Microcalorimetry has gained importance as one of the most
reliable methods for the study of the surface acidity of solid
catalysts. Ammonia, pyridine and other basic molecules can
be conveniently used, according to speciÐc needs, to assess the
number and the strength of the acid sites. Surface acidity is a
key parameter in the catalysis of a wide range of reactions of
industrial interest. A potential route for upgrading low-cost,
largely available polynuclear aromatic feedstocks is the con-
at 823 K (heating rate 3.5 K min~1). To obtain the fully deca-
tionated HY sample ion-exchange was carried out under
reÑux at 368 K. HYd was prepared from HY by dealumina-
tion with SiCl at 573 K, according to the procedure reported
4
in ref. 8. Mordenite samples were commercial products from
Norton and ZEOCAT. The chemical composition of the cata-
lysts is reported in Table 1.
A Tian-Calvet heat Ñow equipment (Setaram) equipped
with a volumetric vacuum line was used for microcalorimetric
measurements. Each sample was pretreated overnight at 673
K under vacuum (10~3 Pa). Adsorption was carried out at
423 K, by admitting successive doses of pyridine and record-
ing the thermal e†ect; the run was stopped at a Ðnal equi-
librium pressure of 133.3 Pa. After overnight outgassing of the
sample at 423 K, a second adsorption was carried out, giving
an adsorption isotherm parallel to the Ðrst one. From the dif-
ference between the adsorption isotherms the amount of pyri-
dine irreversibly adsorbed was evaluated.
version
of
1-methylnaphthalene
(1-MN)
into
2-
methylnaphthalene (2-MN). The latter can be easily oxidised
to 2-methyl-1,4-naphthoquinone, used therapeutically, under
the name of Menadione, as a substitute for natural Vitamin K
in the treatment of bleeding disorders. A further potential use
of 2-MN is via its shape-selective methylation to 2,6-dimethyl-
naphthalene, used in the manufacture of Ðlms, Ðbres and
engineering plastics. Zeolites are strong candidates as catalysts
for the conversion of 1-MN to 2-MN: no pollution problems
originate from these acid solids, whose pore structure is able
to induce unique selectivity e†ects. 1-MN isomerisation to
2-MN was Ðrst reported to occur over magnesium-modiÐed
NaY zeolites.1 The behaviour of ), X, Y and ZSM-5 was then
investigated in our laboratory.2h5 All the zeolites were com-
pletely inactive in their fully Na-form; for H zeolites the
sequence of the initial reactant conversion was HY B H) A
HX A HZSM-5; selectivity to 2-MN ranged from ca. 70 to
100%; coking and activity decay were observed. Di†erences in
the nature of the carbonaceous deposits formed during 1-MN
conversion on zeolites of di†erent structural types have been
reported;6,7 besides coking, transalkylation was reported to
be the main undesired reaction.
The catalytic activity was tested in a Ðxed-bed Ñow micro-
reactor. The zeolite samples were activated in situ by Ñushing
with N for 2 h at 773 K before reaction. Liquid 1-MN was
2
fed by a syringe pump to the preheating section of the reactor,
where it was vaporised and mixed with a H stream
2
(H /hydrocarbon \30 mol/mol). The reaction was carried out
2
at 623 K under atmospheric pressure; contact time, W /F,
ranged between 0.23 ] 102 and 3.28 ] 102 g(cat) h [mol(1-
MN)]~1. The reactor effluent was collected in cold traps and
analysed by capillary gas chromatography (GC) (carrier gas:
H ; column: 100 m Petrocol DH; oven programme: 1 min at
2
383 K, heating at 1 K min~1 up to 493 K; detector: Ñame
In the present work microcalorimetry was used to deter-
mine the acid properties of Y zeolites and mordenite, di†ering
as to the extent of decationation and/or the Si/Al ratio; 1-MN
was reacted over these samples in a Ðxed-bed Ñow reactor and
their catalytic behaviour was interpreted by considering the
combined inÑuence of structure and acidity.
ionization).
The amount of coke was determined from the amount of
CO evolved when heating the catalyst at 773 K in an O
2
2
Ñow, after purging in N to desorb reactant and/or products
2
retained by the solid.
Table 1 Chemical composition of the zeolites
Experimental
sample
unit cell formula
NaY was a commercial sample from Union Carbide. The par-
tially decationated NaHY sample was prepared from NaY by
NaY
NaHY
HY
HYd
HMOR-5
HMOR-11
Na Al Si
O
51 51 141 384
Na
H
H
Al Si
O
ion-exchange with NH Cl aqueous solution, 0.5 M, at room
28 23 51 141 384
Al Si
O
O
4
51 51 141 384
temperature, followed by overnight calcination in Ñowing air
H
Al Si
30 30 162 384
H Al Si
O
8
8
40 96
O
H Al Si
4
4
44 96
¤ E-mail: ferino=vaxcal.unica.it
J. Chem. Soc., Faraday T rans., 1998, 94(17), 2647È2652
2647

View Article Online
kJ mol~1; reasonably, these sites are the Na` ions acting as
Lewis centres. Accordingly, the large peak in NaHY can be
associated with the reversible interaction of pyridine with the
weakly acidic Na` ions acting as Lewis centres. Accordingly,
the large peak in NaHY can be associated with the reversible
interaction of pyridine with the weakly acidic Na` ions
remaining after partial decationation, whereas the peak
around 110 kJ mol~1 indicates that a new set of sites of higher
strength is formed. Complete decationation leads to HY; as
expected, its spectrum no longer shows the low-energy peak,
whereas the peak at 110 kJ mol~1 appears markedly
enhanced; another peak is also visible around 132 kJ mol~1.
The present results for HY are quite similar to those pre-
viously obtained for other commercial HY samples of similar
Si/Al ratio9, whose acid-strength spectra for pyridine adsorp-
tion showed two sets of homogeneous sites with heats of
adsorption between 145 and 115 kJ mol~1; comparison of
FTIR and calorimetric data allowed us to conclude that these
sites were BrÔnsted in nature. The acid-strength distribution
for HYd is similar to that of the parent HY, the e†ect of the
mild dealumination being to decrease the number of acid sites.
Calorimetric results for mordenites are presented in Fig. 3.
The initial di†erential heats of adsorption are between 250
and 220 kJ mol~1 for HMOR-5 and HMOR-11, respectively.
Increasing coverage causes a sudden drop in the di†erential
heat to a plateau region, followed by a continuous decrease.
The amount of irreversibly adsorbed pyridine is 788 and 625
lmol g~1 for HMOR-5 and HMOR-11, respectively. The sites
Results and Discussion
Microcalorimetric characterisation
Calorimetric results for the Y samples are reported in Fig. 1,
where the di†erential heat of adsorption (Q ) is plotted vs.
diff
pyridine uptake (n). The initial values of Q are in the range
diff
240È195 kJ mol~1. All the samples show a stepwise decrease
in the di†erential heat with coverage. Non-speciÐc adsorption
could be responsible for the low di†erential heat observed at
the higher coverages. Calorimetric runs carried out on silica
samples (two commercial and one prepared by the solÈgel
technique) actually showed that pyridine is adsorbed to some
extent on pure silica, with di†erential heats not exceeding 60
kJ mol~1. Accordingly, pyridine uptake originating di†eren-
tial heats \60 kJ mol~1 were neglected. Some of the acid sites
interact reversibly with pyridine. They have di†erential heats
below 90È100 kJ mol~1 (depending on the particular sample).
The amount of the sites which adsorb pyridine irreversibly, n ,
i
is 88, 354, 746 and 566 lmol g~1 for NaY, NaHY, HY and HYd
respectively. The number of sites with Q [ 150 kJ mol~1
diff
is less than 50 lmol g~1 for any of the samples.
From the Q
distribution plots, [dn/dQ
vs. coverage curves of Fig. 1 the site energy
diff
vs. Q , were obtained. These
diff
diff
are presented in Fig. 2 and can be regarded as the acid-
strength spectra of the Y zeolites. The existence of di†erent
sets of sites, each set being homogeneous as to strength, is
revealed for each sample by the presence of peaks in its spec-
trum. Most of the sites of the parent NaY are homogeneous,
as indicated by the presence of a single, large peak around 84
with Q [ 150 kJ mol~1 do not exceed 65 lmol g~1. The
diff
acid-strength spectra are shown in Fig. 4. HMOR-5 shows a
large peak around 135 kJ mol~1 (with a shoulder at 140
kJ mol~1). The set of homogeneous sites observed on the
mildly dealuminated HMOR-11 has a strength similar to that
of HMOR-5.
Additional insight into the adsorption process can be
obtained by considering the time over which the heat is
evolved after each successive dose (thermokinetic parameter,
t ). This parameter is plotted vs. coverage in Fig. 5 for Y zeo-
O
lites. For all the samples except NaY, t Ðrst increases with
0
pyridine coverage, passes through a pronounced maximum
and then decreases. Comparison with Fig. 1 shows that the
maxima correspond to the end of the stepped portion of the
Q
vs. coverage plots, i.e. to the point beyond which
diff
reversible adsorption becomes established. The thermokinetic
parameter for mordenites is shown in Fig. 6. Comparison with
Fig. 3 shows that for these samples also the maxima occur at
the coverages for which irreversible adsorption is complete.
This indicates that surface di†usion is the mechanism by
which the sites are selectively titrated during sequential expo-
sure of the sample to pyridine.10 At low coverages, pyridine is
adsorbed on the strong sites, where it remains irreversibly
held, giving a low thermokinetic parameter. As coverage
increases, the number of strong sites still available decreases
and a competition between strong and weak sites takes place
Fig. 1 Di†erential heat of adsorption, Q , vs. pyridine uptake for Y
diff
zeolites. Open symbols refer to readsorption after evacuation.
Fig. 3 Di†erential heat of adsorption, Q , vs. pyridine uptake for
diff
Fig. 2 Site energy distribution plots for Y zeolites
mordenites. Open symbols refer to readsorption after evacuation.
2648
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Article Online
Fig. 4 Site energy distribution plots for mordenites
Fig. 6 Thermokinetic parameter, t , vs. pyridine uptake for morde-
0
nites
for the adsorption of pyridine molecules. Those eventually
adsorbed on the weak sites are able to migrate to the stronger
sites; the smaller the number of strong sites, the slower this
surface equilibration process (as the molecules must di†use
over greater distances) and hence the thermokinetic parameter
increases. After complete coverage of the stronger sites, only
the exchange of pyridine among weaker sites occurs. This is a
faster process resulting in a decrease in the thermokinetic
parameter. The short time for the heat evolution observed for
all the doses over NaY (Fig. 5) is in agreement with the weak
(Lewis) acidity of this sample.
Catalytic behaviour
1-MN was reacted over the whole group of catalysts at 623 K.
Fig. 7 shows the inÑuence of the contact time, W /F, on the
catalytic behaviour of HMOR-5, HMOR-11, NaHY and HYd
after 30 min on-stream, i.e. the time needed for the estab-
lishing of stationary Ñuodynamic conditions. 1-MN undergoes
selective conversion to 2-MN over both HMOR-5 (Fig. 7A)
and HMOR-11 (Fig. 7B); on the latter, trace amounts
of by-products (\1%) were revealed at W /F \ 1.36 ] 102
g(cat) h [mol(1-MN)]~1 and less than 3.5% of naphthalene
(N) was formed at 3.28 ] 102 g(cat) h [mol(1-MN)]~1. Ther-
modynamic equilibrium (39% 1-MN and 61% 2-MN) is prac-
tically attained over both mordenites, but it requires working
at the highest value of W /F in the case of HMOR-5, whereas
this occurs even at W /F \ 0.82 ] 102 g(cat) h [mol(1-MN)]~1
over HMOR-11.
The activity of HYd is remarkably high, as shown in Fig.
7C by the concentration of 1-MN in the effluent at the lower
W /F value. Even under these relatively mild conditions, para-
site reactions occur to some extent, as indicated by the moder-
ate presence of naphthalene and dimethylnaphthalenes. As
W /F is increased, a remarkable enhancement of N production
is observed, while 2-MN decreases; also DMNs seem to
decrease slightly. Over NaHY (Fig. 7D) DMNs are hardly
formed (less than 2% at the highest W /F value). 2-MN
increases up to a maximum and then smoothly decreases as
W /F is raised. N is present only in very small amounts at
W /F \ 0.82 ] 102 g(cat) h [mol(1-MN)]~1 beyond this point
its concentration increases with W /F, with a trend similar to
that for HYd.
NaY was found to be practically inactive at any W /F. The
initial (30 min on-stream) results for HY at W /F \ 0.82 ] 102
Fig. 5 Thermokinetic parameter, t , vs. pyridine uptake for Y zeo-
0
lites
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
2649

View Article Online
Fig. 7 InÑuence of the contact time, W /F, on the composition of the reactor effluent. (L), 1-MN; (K), 2-MN; ()), N; (|), DMNs. Results refer
to 30 min on stream.
g(cat) h [mol(1-MN)]~1 were: 27.3% 1-MN, 52.8% 2-MN,
9.8% N and 10.1% DMNs. This catalyst was initially rather
active, even at W /F \ 0.23 ] 102 g(cat) h [mol(1-MN)]~1:
45.8% 1-MN, 47.2% 2-MN, 3.5% N and 3.2% DMNs;
however, it showed a dramatic activity decay, becoming prac-
tically inactive (naphthalene conversion ca. 2%) after 3 h on-
stream.
The inÑuence of time-on-stream (TOS) on the catalytic
behaviour of HMOR-5, HMOR-11, HYd and NaHY has been
tested at W /F \ 3.28 ] 102 g(cat) h [mol(1-MN)]~1. Conver-
sion and selectivity data are plotted vs. TOS in Fig. 8 and 9
respectively. HMOR-5 undergoes severe deactivation, whereas
the other catalysts are still active after several hours on-stream
(Fig. 8). In the case of HYd a variation in selectivity occurs
along with conversion during deactivation (Fig. 9A). On the
fresh catalyst, naphthalene production is strongly favoured,
but it decreases dramatically with icreasing TOS, while 2-MN
selectivity simultaneously increases. DMNs are also formed;
after the early reaction period, their selectivity curve is practi-
cally superimposed on that for naphthalene. After several
hours on-stream the amount of by-products is very low,
2-MN being produced with a selectivity higher than 90%. The
selectivity of NaHY depends similarly on TOS (Fig. 9B) but in
this case the formation of DMNs is always low (\8%). For
both HMOR-5 and HMOR-11 the selectivity to the desired
product was practically complete at any TOS. Coking occurs
on the catalysts. The amount of coke deposited per g of cata-
Fig. 9 InÑuence of TOS on selectivity for HYd (A) and NaHY (B).
(L), 2-MN; (K), N; (|), DMNs. W /F as in Fig. 8.
Fig. 8 InÑuence of TOS on the conversion of 1-MN. (L), HMOR-5;
(K), HMOR-11; (|), NaHY; ()), HYd. Results refer to W /F \
3.28 ] 102 g(cat) h [mol(1-MN)]~1.
Fig. 10 InÑuence of TOS on coke formation for HMOR-5 ()),
HMOR-11 (|), NaHY (K) and HYd (L). W /F as in Fig. 8.
2650
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Article Online
lyst is shown as a function of TOS in Fig. 10. HMOR-5
appears to be “saturatedÏ by coke, after a few minutes of reac-
tion, and a smooth coke increase seems to occur on HMOR-
11. Catalyst “saturationÏ by coke occurs also for the two
faujasites, but it requires ca. 4 h on-stream for HYd and even
a few more hours for NaHY.
nism involving the formation of a diarylmethane intermediate
is generally invoked for the disproportionation of mono-
alkylaromatics.14 While it is generally accepted that dealkyla-
tion requires stronger sites than isomerisation,14 the question
whether the acid sites that catalyse intramolecular isomer-
isation are responsible also for disproportionation has been
debated.14h18 It has been suggested that disproportionation
requires stronger sites than isomerisation.15,17 Pyridine poi-
soning experiments18 support the view that both reactions
demand acid sites of equivalent strength in the case of o-
xylene and 1,2,4-trimethylbenzene transformation over HY.
Being a bimolecular reaction, disproportionation should be
strongly dependent on the density of the acid sites; owing to
the dimensions of the intermediate species, the structure of the
zeolite should also play a fundamental role. Because of the
several bimolecular reaction steps and the very bulky tran-
sition states involved, coking is very sensitive to both site
density and zeolite structure.19,20
The protonation of 1-MN should be the Ðrst step for both
its isomerisation and dealkylation. Though the latter could
occur before the protonated species has isomerised as well as
after, the concentration proÐles shown in Fig. 7D seem to
suggest that, over NaHY, dealkylation follows isomerisation.
The trend of naphthalene production for HYd (Fig. 7C) is
similar to that for NaHY (Fig. 7D), being shifted upwards for
the former because of the occurrence of disproportionation
superimposed on dealkylation. The concentration proÐles pre-
sented in Fig. 7C seem to suggest also that isomerisation and
disproportionation occur through parallel pathways, which is
consistent with di†erent initial steps. The presence of a
maximum in the 2-MN curve indicates the establishment of
consecutive reactions, likely dealkylation and perhaps (at the
intermediate W /F value) disproportionation. DMNs also
seem to undergo further (very limited) transformation at the
high W /F value.
Fig. 11 1-MN conversion as a function of the amount of pyridine
irreversibly adsorbed, n , for Y zeolites
i
The acid properties of the samples are expected to deter-
mine their catalytic behaviour. The observed inactivity of
NaY indicates that the weak sites, i.e. those adsorbing pyri-
dine reversibly, are not able to catalyse the transformation of
1-MN. 1-MN conversion depends on the population of the
acid sites strong enough to give irreversible interaction with
The occurrence of disproportionation (besides dealkylation
and isomerisation) over HYd and its absence over NaHY
might stem from the presence on the former and the lack on
the latter of some family of sites of particular strength. Calori-
metric data allow us to rule out such a possibility. For HYd,
pyridine (n ), as shown in Fig. 11 for Y zeolites. An interesting
i
di†erence between NaHY and HYd is the occurrence of
dealkylation practically without disproportionation over
NaHY, whereas both reactions are present as well as isomer-
isation in the case of HYd (compare the composition of the
reactor effluent in Fig. 7C and D). To correlate this di†erence
with the acid properties the mechanism of the various reac-
tions should be considered.
The isomerisation of mononuclear alkylaromatics occurs
through intramolecular transfer of the alkyl group.11 Repor-
tedly, m-xylene isomerisation can also proceed via a bimolecu-
lar mechanism involving disproportionation products as
intermediates,12,13 but this requires that the dispro-
portionation reaction accompanies isomerisation to a rela-
tively high extent. Accordingly, we can assume that 1-MN
isomerisation is monomolecular over our catalysts, where
isomerisation is, by far, the predominant reaction. A mecha-
30% of the active sites have Q [ 125 kJ mol~1, 62%
diff
between 125 and 100 kJ mol~1 and 8% \100 kJ mol~1. For
NaHY, 22, 50 and 28% of the active sites, respectively, are
found in the same regions of strength. The two samples do not
di†er substantially as to their acid-strength distribution, in the
sense that the fraction of the active sites populating each
region is signiÐcant in both cases; furthermore, dealkylation
(which is known to require stronger sites than
disproportionation14) occurs over NaHY; this would enable
disproportionation to occur if the strength was the relevant
parameter.
Assuming competition between disproportionation and
isomerisation on the sites with strength up to 125 kJ mol~1,
Scheme 1
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
2651

View Article Online
and competition of both with dealkylation on the sites strong-
er than 125 kJ mol~1, the concentrations of active sites able to
catalyse both isomerisation and disproportionation are 566
and 354 lmol g~1 for HYd and NaHY, respectively. Accord-
ingly, the occurrence of disproportionation over HYd is most
likely related to its higher density of active sites in comparison
with NaHY, where bimolecular disproportionation is dis-
favoured.
simply raising the temperature at 773 K in N Ñow until
neither reactants nor products were revealed in the exit
stream.
References
2
1
2
3
4
C. Dimitrov, Z. Popova and M. Tuyen, React. Kinet. Catal. L ett.,
1978, 1, 101.
V. Solinas, R. Monaci, B. Marongiu and L. Forni, Appl. Catal.,
1983, 5, 171.
V. Solinas, R. Monaci, B. Marongiu and L. Forni, Appl. Catal.,
1984, 9, 109.
V. Solinas, R. Monaci, B. Marongiu and L. Forni, in Catalyst
Deactivation, ed. B. Delmon and G. F. Froment, Elsevier,
Amsterdam, 1987, p. 493.
The route for coke formation is probably through the same
intermediate involved in disproportionation, as shown in
Scheme 1 (adapted from ref. 21 and 22). As expected, the
higher disproportionation activity of HYd compared to
NaHY is associated with a higher coking activity (Fig. 10).
Dealkylation is suddenly suppressed as TOS increases (Fig. 9)
because of poisoning of the stronger sites by coke; com-
petition between disproportionation and isomerisation then
occurs on the remaining sites. The size of the condensation
products shown in Scheme 1 is such that they can only emerge
with great difficulty from the supercages. Besides enhancing
their tendency towards further condensation leading to poly-
aromatic coke, the location of these molecules in the super-
cages slows down the di†usion of reactants and products
through the supercages. The formation of the cumbersome
intermediate of disproportionation is thus hindered; further-
more, the di†usion of the cumbersome dialkylated naphtha-
lenes would be slow compared to that of 2-MN. Thus,
shape-selectivity e†ects induced by coke formation enhance
isomerisation selectivity.
As expected, over HMOR-5 and HMOR-11, both dispro-
portionation and coking are highly disfavoured (Fig. 7A and
B and 10), as the channel system without large cavities in
these zeolites is not spacious enough to allow the formation of
the bulky reaction intermediates. The concentration of the
acid sites able to catalyse 1-MN transformation is 788
lmol g~1 for HMOR-5 and 625 lmol g~1 for HMOR-11.
In spite of this, HMOR-5 is less active than HMOR-11 (cf.
Fig. 7A with 7B). This is due to the slow di†usion of the cum-
bersome reactant and products within the monodimensional
channel system of HMOR-5, where hindering interaction with
the active sites present at a high density is originated. The
accumulation of strongly adsorbed feed and/or product mol-
ecules is responsible for the severe decay in activity of
HMOR-5 (Fig. 8). This is supported by the Ðnding that a
deactivated HMOR-5 sample (4 h on-stream, conversion
\10%) gave a renewal of activity (conversion [30%) after
5
6
7
8
9
L. Forni, V. Solinas and R. Monaci, I. & E.C. Res., 1987, 26,
1860.
M. M. Neuber, S. Ernst, H. Geerts, P. J. Grobet, P. A. Jacobs,
G. T. Kokotailo and J. Weitkamp, in ref. 4, p. 567.
M. Neuber, H. G. Karge and J. Weitkamp, Catal. T oday, 1988, 3,
11.
H. K. Beyer, I. M. Belenykaja, F. Hange, M. Tielen, P. J. Grobet
and P. A. Jacobs, J. Chem. Soc., Faraday T rans. 1, 1985, 81, 2889.
G. Colon, I. Ferino, E. Rombi, E. Selli, L. Forni, P. Magnoux
and M. Guisnet, Appl. Catal., A, 1998, 168, 81.
10 N. Cardona-Martinez and J. A. Dumesic, J. Catal., 1990, 125,
427.
11 M. L. Poutsma, in Zeolite Chemistry and Catalysis, ed. J. A.
Rabo, ACS Monograph 171, American Chemical Society, Wash-
ington, 1976, p. 437.
12 S. Morin, N. S. Gnep and M. Guisnet, J. Catal., 1996, 159, 296.
13 S. Morin, N. S. Gnep and M. Guisnet, Proceedings Int. Symp.
AcidÈBase Catalysis III, Rolduc, April 20È24, 1997, p. 43.
14 M. Guisnet and N. S. Gnep, in Zeolites: Science and T echnology,
ed. F. R. Ribeiro, A. E. Rodriguez, L. D. Rollmann and C. Nac-
cache, NATO ASI Series, Martinus Nijhof, the Hague, 1984, p.
571.
15 P. Ratnasamy, S. Sivasankar and S. Vishnoi, J. Catal., 1981, 69,
428.
16 P. Cartraud, A. Cointot, M. Dufour, N. S. Gnep, M. Guisnet, G.
Joly and J. Tejada, Appl. Catal., 1986, 21, 85.
17 J. A. Martens, J. Perez-Pariente, E. Sastre, A. Corma and P. A.
Jacobs, Appl. Catal., 1988, 45, 85.
18 G. Bourdillon, C. Gueguen and M. Guisnet, Appl. Catal., 1990,
61, 123.
19 S. M. Csicsery, Zeolites, 1984, 4, 208.
20 M. Guisnet and P. Magnoux, Appl. Catal., 1989, 54, 1.
21 J. R. Anderson, Q-N. Dong, Y-F. Chang and R. J. Western, J.
Catal., 1991, 127, 113.
22 P. Magnoux, C. Cana†, F. Machado and M. Guisnet, J. Catal.,
1992, 134, 286.
Paper 8/03931C; Received 26th May, 1998
2652
J. Chem. Soc., Faraday T rans., 1998, V ol. 94