Alkylation of benzene catalysed by supported heteropolyacids

Article abstract of DOI:10.1016/S1381-1169(99)00078-3

A new kind of solid acid based on heteropolyacids has been proved to be a useful catalyst in alkylation of benzene with propylene to obtain cumene. The new catalysts are based on a heteropolyacid supported on zirconia treated with very low contents of sulphate anions. Two different heteropolyacids have been studied, both with Keggin structure, (H₃PMo₁₂O₄₀, H₃PW₁₂O₄₀), and both produced active catalysts. The zirconia was used either pure or doped with few percent of iron oxide. The catalysts have been characterised through XRD, ³¹P NMR and HREM. These catalysts are very active in alkylation reaction even at mild temperature. Their two outstanding features are the low n-propylbenzene production and the possibility to regenerate it at moderate temperature (35°C). The best catalyst is based on H₃PW₁₂O₄₀ supported on zirconia doped with iron and treated with sulphate ions.

Full text of DOI:10.1016/S1381-1169(99)00078-3

Journal of Molecular Catalysis A: Chemical 146 Ž1999. 37–44
www.elsevier.comrlocatermolcata
Alkylation of benzene catalysed by supported heteropolyacids
Alberto de Angelis, Stefano Amarilli, Donatella Berti, Luciano Montanari,
)
Carlo Perego
EniTecnologie, Õia F. Maritano 26, 20097 S. Donato Milanese, MI, Italy
Abstract
A new kind of solid acid based on heteropolyacids has been proved to be a useful catalyst in alkylation of benzene with
propylene to obtain cumene. The new catalysts are based on a heteropolyacid supported on zirconia treated with very low
contents of sulphate anions. Two different heteropolyacids have been studied, both with Keggin structure, ŽH PMo O ,
3
12 40
H PW O ., and both produced active catalysts. The zirconia was used either pure or doped with few percent of iron oxide.
3
12 40
3
1
The catalysts have been characterised through XRD, P NMR and HREM. These catalysts are very active in alkylation
reaction even at mild temperature. Their two outstanding features are the low n-propylbenzene production and the possibility
to regenerate it at moderate temperature Ž3508C.. The best catalyst is based on H PW O supported on zirconia doped
3
12 40
with iron and treated with sulphate ions. q 1999 Elsevier Science B.V. All rights reserved.
Keywords: Alkylation; Benzene; Heteropolyacids
1
. Introduction
phoric acid, supported on kieselguhr SKPA ,
Ž .
patented by Ipatieff w3x. Only few plants are
The alkylation of benzene with propylene
based on the Monsanto technology, which uses
gives cumene, a very important petrochemical
commodity used for the production of phenol
and acetone w1x. More than 90% of the world’s
phenol demand is derived from cumene. Hence,
the growth rate of cumene is closely associated
with that of phenol. Cumene demand is ex-
pected to grow about 3.5% a year over the next
aluminum trichloride
processes have problems of corrosion, waste
treatment AlCl3 and exhaust catalyst disposal
SKPA . This accounts for the efforts devoted
to the research of new more efficient, safe and
environment-friendly catalysts.
Different kinds of solid acids have been ex-
tensively tested for alkylation of aromatics such
as: heteropolyacids, pillared clays, silica alu-
mina and acidic zeolites. Acidic zeolites have
been thoroughly evaluated for benzene alkyla-
tion with propylene since 1965 w4x, but their use
AlCl3 as catalyst. Both
Ž .
Ž
.
Ž
.
5
years w2x. The cumene capacity in the world is
about 8 million tonsryear distributed over about
0 plants. The most widespread process is the
4
UOP. The catalyst for such a process is phos-
in commercial plants has been announced only
recently by Dow, Mobil, CDTech, UOP and
Enichem w5x. During the search for new cata-
)
Corresponding author. Fax: q39-2-5205-6364; E-mail:
cperego@enitecnologie.eni.it
1
381-1169r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
PII: S1381-1169Ž99.00078-3

3
8
A. de Angelis et al.rJournal of Molecular Catalysis A: Chemical 146 (1999) 37–44
lysts, which led to the development of the
heteropolyacid and the free inorganic acid. The
aqueous solution is made of 200 cc of 0.1 N
solution of sulphuric acid in which 20 g of the
heteropolyacid is dissolved. In such solution 20
g of the preformed oxide is suspended, stirred
for about 4 h then filtered and washed with
distilled water till neutrality of the mother wa-
ter. The sample is then dried under air flow at
temperature between 2008C and 3508C for a
period of 12 h. The choice of the temperature
depends on the thermal stability of the hetero-
polyacid w9,10x.
Enichem process based on zeolite beta w6,7x, we
performed a wide screening considering solid
acids other then zeolites, such as pillared clays,
mesoporous silica–alumina
polyacids.
MSA and hetero-
Ž .
In this paper, we shall describe a new kind of
catalyst based on a heteropolyacid supported on
zirconia treated with very low content of sul-
phate anion w8x. Such catalyst shows pretty high
activity and a very low formation of n-propyl-
benzene NPB . Furthermore, the catalyst can
Ž .
be easily reactivated at relatively low tempera-
ture without significant loss of activity or change
in its crystalline structure also after some reac-
tion cycles.
2
.4. Reaction apparatus
The catalysts have been tested in a fixed bed
reactor, at temperature of 100–1508C and pres-
sure of 4 MPa; under this condition the feed is
liquid. The reactor was charged with 3 g of
catalyst, crushed from 5 tonsrcm pelletised
wafers and sieved to 20–40 mesh particles. The
catalyst was then flushed in slowly flowing dry
nitrogen while heating up to 1508C for 3 h.
The benzenerpropylene molar ratio is 7 and
the reagents are fed separately with two differ-
2
. Experimental
2
H PW O
and H PMo O , H SO ,
3
12 40
3
12 40
2
4
Ž
98%
.
ZrOCl , Fe
Ž
NO
.
, all analytically pure
2
3 3
and from Fluka are used as received.
2
.1. Synthesis of ZrO2
The oxide is prepared by hydrolysing in
aqueous solution an opportune precursor such as
ent pumps. Total WHSV
velocity is in the range 1–20 h . Benzene was
Ž
weight hourly space
y
1
.
ZrOCl with ammonia till the pH value of the
suspension reaches 10, filtering, washing it
carefully with plenty of water, then drying it at
a C.Erba reagent, )99% pure. Propylene was
supplied by SIAD, )99.9% pure. All these
reagents were used as supplied.
2
2
508C for 24 h.
The analysis of gases is performed on a GC
.
Hewlett Packard 5890 connected on-line with
Ž
2
.2. Synthesis of ZrO doped with iron
the plant. Liquid reaction samples are taken by
2
cooling the effluents at 108C and analysed
The mixed oxide is prepared treating an
through GC
capillary column PONA
mm O.D., coated with a 0.5 mm thick film of
stationary phase . Compounds from C to C
have been detected.
Ž
Hewlett Packard 5890
.
using a
50 m long and 0.21
aqueous solution of ZrOCl2 and Fe
Ž
NO
.
Ž
3 3
Ž
0.5% weight as Fe with ammonia till the pH
.
value of the suspension reaches the value of 10,
filtering, washing it thoroughly with distilled
water and finally drying it at 2508C for 24 h.
.
3
15
2
.5. Catalyst characterisation
2
.3. Catalysts preparation
The catalyst is prepared by suspending the
preformed oxide in aqueous solution of the
The catalysts have been characterised through
HREM-EDS and XRD.

A. de Angelis et al.rJournal of Molecular Catalysis A: Chemical 146 (1999) 37–44
39
Fig. 1. Reaction scheme of alkylation of benzene with propylene.
3
1
2
.5.1. X-ray diffraction (XRD) analysis
The XRD data were collected using a Philips
2.5.3. P-MAS-NMR analysis
3
1
P-MAS-NMR spectra
Ž
121.44 MHz
.
were
equipment with monochromatic Cu Ka radia-
ls1.5418 A˚
tion . The spectra were taken at
ambient conditions in step-scanning mode, in
the range 58-2u-708 step 0.038, time 10
. Qualitative phase analysis was carried
measured at room temperature on a Bruker
AMX-300 equipped with a Double Bearing
MAS device. Two hundred transient responses
were collected with pulse duration of 3 ms
flip angle and with repetition time of 16 s.
Ž
.
Ž
Ž
308
srstep
.
.
out using the Siemens Diffrac AT package run
on a IBM PC 330 P-75.
Spinning rate was 5 kHz. Exponential multipli-
Crystal sizes were calculated from diffraction
line broadening applying the Scherrer equation.
2
.5.2. High-resolution transmission electron mi-
croscopy (HREM) and energy dispersiÕe spec-
troscopy (EDS) analysis
HREM experiments were done on a Jeol
JEM-3010 microscope max. acceleration volt-
Ž
˚
.
age: 300 kV, max. resolution: 1.6 A equipped
with a Link Isis EDS apparatus.
The catalyst powder was embedded in epoxy
resin and then microtomed with a diamond knife
˚
to obtain sections ;400%700 A thick. Images
s
Fig. 2. Catalytic activity of H PW O on ZrO –SO at differ-
3
12 40
2
4
were obtained at 300 kV.
ent WHSV ŽT s1008C, t.o.s. 15 h..

4
0
A. de Angelis et al.rJournal of Molecular Catalysis A: Chemical 146 (1999) 37–44
Table 1
Catalytic test results of H PW O rZrO –SO
s
3
12 40
2
4
Reaction temperature Ž8C.
Propylene molar conversion, %
AromaticsrC3 selectivity, %
NPB Žppmrcumene.
1
1
00
50
99
99
93
96
96
184
Influence of reaction temperature.
Ps4 MPa.
WHSVs1 h
t.o.s.s15 h.
y1
.
cation of 20 Hz was applied prior to Fourier
Transformation. 1% H PO in D O was used as
which can alkylate benzene or crack to lighter
products olefins and paraffins . But the most
Ž
.
3
4
2
an external reference.
unwanted impurity is NPB. In fact, it is not
possible to separate this component from the
cumene by distillation and than its formation
should be prevented. A thermodynamic equilib-
rium exists between cumene and NPB; the latter
is favoured at increasing temperatures.
3
. Cumene chemistry
The most important reactions involved in the
alkylation of benzene with propylene are shown
in Fig. 1. Cumene is obtained according to the
classical carbenium ion mechanism in which
propylene gives the propyl cation which attacks
a benzene molecule. Then, through transfer of a
proton, the propylbenzenium cation rearranges
to cumene. However, it can be subsequently
4. Results and discussion
4.1. Catalytic actiÕity and selectiÕity
Fig. 2 reports the catalytic activity for the
alkylated to give diisopropylbenzenes
Ž
DIPB
.
catalyst based on H PW O
Ž
HPA
.
on zirco-
nia treated with sulphated anion. The sulphur
content in this sample is very low 0.11% wt.
while the tungsten is 18.6% wt. . The results of
the test have been reported as propylene conver-
sion, selectivity to useful aromatics cumene
and DIPB and NPB measured as ppmrcu-
mene
3
12 40
Ž
ortho, meta and para . The m-DIPB is the most
.
favoured isomer according to the thermody-
namic distribution in the range 100–1508C.
Nevertheless, from the DIPB, it is possible to
recover cumene through the transalkylation re-
action carried out on an acid catalyst in pres-
ence of benzene. There are also several reac-
tions that lead to unwanted by-products. For
example propylene form C –C oligomers
Ž
.
,
Ž
.
Ž
.
Ž
.
.
This catalyst is very active. In fact the propy-
lene conversion is higher than 90% also when
6
9
Table 2
s
s
Catalytic test results of H PW O rZrO –SO vs. H PMo O rZrO –SO
3
12 40
2
4
3
12 40
2
4
Catalyst
Propylene molar conversion, %
AromaticsrC3 selectivity, %
NPB Žppmrcumene.
H PW O
99
71
93
95
96
50
3
12 40
H PMo12O40
3
Ts1008C.
Ps4 MPa.
WHSVs1 h
t.o.s.s15 h.
y1
.

A. de Angelis et al.rJournal of Molecular Catalysis A: Chemical 146 (1999) 37–44
41
Table 3
Catalytic test results of H PW O rZrO –SO
s
3
12 40
2
4
Catalyst
Support
ZrO –SO
4
Propylene molar conversion, %
AromaticsrC3 selectivity, %
NPB Žppmrcumene.
H PW O
99
99
93
92
96
-10
3
12 40
2
H PW O
ZrO –Fe O –SO
4
3
12 40
2
2
3
Influence of the dopant.
Ts1008C.
Ps4 MPa.
WHSVs1 h
t.o.s.s15 h.
y1
.
the WHSV reaches the value of 20 hy1
Fig. 2
.
doping sulphated zirconia with few percent of
iron oxide increases the strength of the acid
sites w12x. Compared to those of pure zirconia,
such modified acid sites probably increase the
relative stability of secondary carbenium ion
with respect to the primary carbenium ion. As
the latter is responsible for NPB formation w13x,
this could account for better selectivity of the
catalyst doped with iron.
Ž
.
The selectivity to useful aromatics lowers in-
creasing the WHSV; this is due to an increase
of oligomers production. The production of NPB
is much lower
Ž
96–97 ppmrcumene at WHSV
s1 than the one reported for SKPA
.
Ž
200 ppm
.
w6x. This lower value of NPB is due to the
milder temperature
Ž
1008C
.
required by
s
H PW O rZrO –SO with respect to the
3
12 40
2
4
SKPA 2008C .
Ž .
s
Also for H PW O rZrO –SO increasing
4.2. Catalyst deactiÕation and regeneration
3
12 40
2
4
the reaction temperature of 508C
Ž
from 100 to
the value of NPB is nearly twice from
6 to 184 ppmrcumene. ŽTable 1
. The selec-
1
9
508C
.
Ž
The catalyst based on H PW O on sul-
3
12 40
.
phated zirconia has been subjected to a life test,
feeding the propylene and benzene mixture
tivity to aromatics instead increases from 93 to
9
6% owing to the lower formation of oligomers
Ž
h
molar ratios7
.
for the first 18 h at WHSVs1
y
1
y1
at higher temperatures.
and then at WHSVs5 h . This catalyst
This catalytic system is still active using a
different metal in the Keggin structure: molyb-
denum, for example, can be substituted for
deactivates after more than 90 h when the
propylene conversion is 70% less Fig. 3
Ž
.
.
A common method to regenerate a catalyst is
tungsten in H PMe O
Ž
Table 2
.
. This change
to treat it at a temperature around 5508C under
3
12 40
implies a decrease of the acid strength of the
heteropolyacid w11x; as a consequence, there is a
great loss of propylene conversion while the
selectivity to aromatics is nearly the same. This
is an obvious consequence of the minor activity
of the catalyst.
Doping the zirconia with a low content of
iron oxide this catalytic system becomes much
more selective, lowering the content of NPB
Ž
Table 3
.
, the amount of which is less than the
10 ppm . Propylene con-
detectable threshold
Ž
.
version and the aromatics selectivity are nearly
the same as those obtained using pure zirconia
as support. It is well known from literature that
s
Fig. 3. Life test of H PW O on ZrO –SO
.
3
12 40
2
4

4
2
A. de Angelis et al.rJournal of Molecular Catalysis A: Chemical 146 (1999) 37–44
Table 4
Catalytic test results of H PW O rZrO –SO
s
3
12 40
2
4
Catalyst
Propylene molar conversion, %
AromaticsrC3 selectivity, %
NPB Žppmrcumene.
H PW O fresh
99
98
97
93
93
94
96
81
54
3
12 40
H PW O 1st cycle
3
12 40
H PW O 2nd cycle
3
12 40
Regeneration of the catalyst.
Ts1008C.
Ps4 MPa.
WHSVs1 h
t.o.s.s15 h..
y1
.
air flow. This results in the combustion of the
carbonaceous resides, responsible for deactiva-
tion. This method cannot be applied to this kind
of catalyst because it is well known from the
literature w9x that the thermal stability of hetero-
4.3. Chemical–physical characterisation
The feature that the catalyst shows no signifi-
cant differences after each reaction cycle is in
agreement with the fact that there are no signifi-
cant differences between the crystalline struc-
ture of the fresh catalyst and that of the one
regenerated after two reaction cycles.
polyacids, as free acids, does not exceed 4508C.
We have found that it is possible to perform the
regeneration of H PW O
supported on
3
12 40
s
ZrO –SO under air flow at temperature lower
From XRD analysis, the only crystalline
2
4
than 3508C for 2 h. Under these conditions there
is an almost negligible lowering of propylene
conversion and of selectivity to useful aromatics
after each reaction-regeneration cycle while
there is a continuous reduction of NPB forma-
phase present in the fresh catalyst Fig. 4 is the
Ž .
Keggin acid, H PW O . After one activity-re-
3
12 40
generation cycle Fig. 4 there are no significant
Ž .
differences in the spectra: no change in the
crystallinity of the Keggin acid is detected and
no crystallisation of the amorphous support oc-
curs. The conclusion is the same also after the
tion Table 4 .
Ž .
The possibility of regenerating this catalyst at
mild temperature could exist probably due to
the good redox properties of the heteropolyacids
which allow the combustion of carbonaceous
resides at lower temperature.
second activity-regeneration cycle
Ž
Fig. 4
.
.
Fig. 4. XRD spectrum of catalyst Žfresh, after 1st cycle, after 1st
regeneration, after 2nd regeneration..
s
Fig. 5. HREM Analysis of H PW O on ZrO –SO
.
3
12 40
2
4

A. de Angelis et al.rJournal of Molecular Catalysis A: Chemical 146 (1999) 37–44
43
3
1
Fig. 6. P-MAS-NMR spectra of Ža. fresh catalyst, Žb. catalyst after 1 cycle, Žc. catalyst after 1st regeneration, Žd. catalyst after 2nd cycle
and Že. after 2nd regeneration.
From HREM analysis of the H PW O
fresh catalysts, HPA is well dispersed on the
the NMR spectra, we found that the sharp signal
Keggin structure is 20% of all phosphorous
3
12 40
Ž
.
amorphous zirconia matrix and forms spongy
while the broad one is 80%.
aggregates
Ž
Fig. 5
.
whose size is between 100
The attribution of the broad signal to Keggin
ion strongly interacting with the support should
be more correct because the lacunary hetero-
polyanions are not very stable, while the cata-
lyst, as seen above, is stable under reaction
conditions.
and 500 nm. The aggregates are formed by
crystallites of 10–20 nm, made by hexagonal
Ž .
packing of about 1 nm units the Keggin unit .
After two activity-regeneration cycles the Keg-
gin structure of the heteropolyacid is preserved:
no changes are observed in the morphology and
the aggregates are still crystalline.
These data are fully confirmed by XRD and
P NMR. P-MAS-NMR is the most revealing
5. Conclusions
3
1
31
method for examining the state of phosphorous
The catalytic system based on a Keggin hete-
ropolyacid supported on zirconia treated with
sulphate anion shows a good activity in the
alkylation of benzene with propylene to cumene.
It is possible to have an active catalytic system
using different heteropolyacids with very low
content of sulphate anion, and as support zirco-
nia either pure or doped with few percent iron
oxide. Choosing a suitable composition of the
catalyst and reaction conditions, it is possible to
lower the formation of NPB up to almost negli-
gible values. Besides, this catalyst can be easily
3
1
heteropolycompound w14x. Fig. 6 shows the P-
MAS-NMR spectra of the catalyst as fresh
after 1 and 2 activity cycles, and after the
respective regeneration states
c, e. ŽFig. 6
a ,
Ž .
Ž
b
.
d
Ž .
Ž
.
.
All spectra are very similar and consist of a
sharp signal centred at y15.2 ppm, attributed to
the P of the Keggin structure w14x, and of a
much broader one centred at y12.8 ppm which
can be due to a lacunary heteropolyanion w14x or
to Keggin anion strongly interacting with the
ZrO support w15x. From the deconvolution of
2

4
4
A. de Angelis et al.rJournal of Molecular Catalysis A: Chemical 146 (1999) 37–44
w5x G.R. Meima, M.J.M van der Aalst, M.S.U Samson, J.M.
Garces, J.G. Lee, in: J. Weitkamp, B. L u¨ cke ŽEds.., Proceed-
ings of the DGMK-Conference, Catalysis on Solid Acids and
Bases, March 14–15, 1996, Berlin, Germany, 1996, p. 125.
w6x G. Bellussi, G. Pazzuconi, C. Perego, G. Girotti, G. Terzoni,
J. Catal. 157 Ž1995. 227.
regenerated at mild conditions, maintaining un-
changed its crystalline structure and its catalytic
properties.
w7x C. Perego, S. Amarilli, R. Millini, G. Bellussi, G. Girotti, G.
Acknowledgements
Terzoni, Microporous Materials 6 Ž1996. 395.
w8x A. de Angelis, S. Amarilli, C. Perego, G. Girotti, Italian
Patent MI95rA 001730 Ž1995..
We thank EniTecnologie for permission to
publish this paper. We also thank Mr. Signoroni
w9x L. Malaparade, in: P. Pascal, Noveau Trait e` de Chimie
Minerale, Vol. 14, Masson Editeurs, Paris, 1959, p. 902.
Secondo
Ž
EniTecnologie
.
for his helpful work.
w10x
C. Rocchiccioli-Deltcheff, A. Aouissi, M.M. Bettahar, S.
Launay, M. Fournier, J. Catal. 164 Ž1996. 16–27.
w11x T. Okuhara, N. Mizuno, M. Misono, Advanced in Catalysis,
Vol. 41, Academic Press, San Diego, CA, 1996, pp. 139–145.
References
w12x T.K. Cheung, J.L. D’itri, B.C. Gates, J. Catal. 151 Ž1995.
1
39–145.
w1x P.R. Pujado, J.R. Salazar, C.V. Berger, Hydrocarbon Process
5 Ž1976. 91.
w13x M.F. Bentham, G.J. Gajda, R.H. Jensen, H.A. Zinnen, in: J.
Weitkamp, B. L u¨ cke, Proceedings of the DGMK-Con-
ference, Catalysis on Solid Acids and Bases, Berlin, Ger-
many, March 14–15, 1996, 1996, p. 155.
5
w2x W.P. Kleinloh, Proc. 1997 World Petrochemical Conference
ŽCMAI., Houston, TX.
w3x V.N. Ipatieff, US Patent 2,382,318 Ž1945..
w4x Kr.M. Minachev, Ya.I. Isakov, V.I. Garanin, L.I. Piguzova,
V.I. Bogomolov, A.S. Vitukina, Neftekhimiya 5 Ž1965. 676.
w
4 I.V. Kozhevnikov, Catal. Lett. 30 1995 241–252.
x Ž .
1
w15x
T.-H. Chang, J. Chem. Soc. Faraday Trans. 91
Ž . Ž .
2 1995
375–379.