NOTE
331
remove, at least partially, surface aluminum. The isobutene
selectivity was slightly improved, showing that the sur-
face aluminum has a detrimental effect on the isobutene
selectivity.
From these results, it isobserved that the surface dealumi-
nated sample exhibitsa high isobutene selectivitycompared
to the fresh nontreated sample. These observations are in
line with those reported recently by Xu et al. (2) who have
attributed this effect to the removal of non-shape-selective
acid sites.
A very large improvement in isobutene selectivity is ob-
served on a hydrothermally dealuminated sample, either
treated with oxalic acid or untreated. Again these results
are in agreement with those of Xu et al. (2); these au-
thors have attributed these improved performances to the
fact that the “lowering of the number of acid sites de-
creases interactions between intermediates in adjacent acid
sites.”
In order to discriminate between the two effects of the
dealumination treatment (reduction of the density of acid
sites and reduction of the porosity of FER by extraframe-
work aluminum), we have tested a H-Cs-FER material.
The initial selectivity of this solid, containing nearly the
same number of protonic sites as the H-Li-FER, is very
high indicating that the major factor governing the selec-
tivity is the space available around the active site. It is now
likely, for hydrothermally dealuminated FER, that the im-
provement in isobutene selectivity is due to the presence
of aluminum debris in the cavities and pores of the FER,
decreasing the space available in the vicinity of the acid
sites, thus rendering the FER shape selective for the n–
butene skeletal isomerization. The ageing of the solid with
time on stream is probably at the origin of the same phe-
nomenon: coke deposits increasingly block the porosity,
causing an increase of the isobutene selectivity with time on
stream.
FIG. 1. Change in the isobutene selectivity as a function of the con-
version for different solids. The reactant is 1-butene. Selectivity is on a
carbon basis, with all linear butenes (1-butene, 2-cis, and 2-trans butene)
being considered as reactants. Isobutene selectivity is given by the num-
ber of moles of isobutene formed/number of moles of 1-butene trans-
formed into products different from 2-butenes. Conversion is defined
as (Cin ꢀ Cout/Cin) ꢃ 100, with Cin being the number of moles of 1-C4H8,
Cout = 61-C4H8 + 2-cis C4H8 + 2-trans C4H8).
that the fresh nontreated solid has a poor isobutene selec-
tivity. Treatment with oxalic acid improved the isobutene
selectivity, clearly indicating that the surface acid sites are
not selective as reported earlier (1).
Coming back to the surprising result reported earlier in-
dicating that FER zeolites are more selective than MTT (5),
it has to be recalled that the results were certainly obtained
on coked materials since the isobutene yields for FER were
quite high. Very probably, the reverse order of classifica-
tion would be obtained with uncoked zeolites. Work is in
progress to confirm these interpretations.
For the samples dealuminated via the hydrothermal tr-
eatment, the selectivity is improved, showing that the re-
moval of some framework aluminum and the consecutive
formation of extraframework aluminum either in the mi-
croporosity or in the mesoporosity is beneficial for the
isobutene selectivity.
Indeed, it could be speculated that it is the density of
protonic sites which is the major factor because the dealu-
mination causes a decrease in the number of acid sites. In
order to confirm (or to refute) this, we have exchanged a
fresh H-FER with a lithium solution (1 M solution) at 353 K
in order to have a H-Li-FER containing only 5% of the ini-
tial protons (see Table 1). This sample did indeed exhibit
a reduced activity but the selectivity is not much improved
(Fig. 1), clearly indicating that the density of acid sites is not
governing the isobutene selectivity.
ACKNOWLEDGMENT
This work has been funded by Total Raffinage et Distribution.
REFERENCES
1. Xu, W. Q., Yin, Y. G., Suib, S. L., Edwards, J. C., and O’Young, C. L.,
J. Phys. Chem. 99, 9943 (1995).
The hydrothermally dealuminated sample has been sub-
mitted to an additional oxalic acid treatment in order to
2. Xu, W. Q., Yin, Y. G., Suib, S. L., Edwards, J. C., and O’Young, C. L.,
J. Catal. 163, 232 (1996) [and references of the same authors included].