Zirconia-supported 12-tungstophosphoric acid as a solid catalyst for the synthesis of linear alkyl benzenes

Article abstract of DOI:10.1016/j.jcat.2004.09.024

The liquid-phase alkylation of benzene with 1-octene and 1-dodecene was investigated with zirconia-supported 12-tungstophosphoric acid (TPA) as catalysts. We prepared the catalysts, with different TPA loading (5-20 wt% calcined at 750°C) and calcination temperatures (15 wt% calcined from 650 to 850°C), by suspending hydrous zirconia in a methanol solution of TPA, followed by drying and calcination. These catalysts were characterized by X-ray diffraction, DTG-DTA, FTIR pyridine adsorption, NH₃-TPD, and ³¹P MAS NMR spectroscopy measurements. The catalyst with optimum TPA loading (15%) and calcination temperature (750°C) was prepared in different solvents and characterized by ³¹P MAS NMR spectroscopy. The XRD results indicate that TPA stabilizes the tetragonal phase of zirconia. The catalysts show both Bronsted and Lewis acidity, and 15% TPA on zirconia calcined at 750°C shows the highest acidity. ³¹P MAS NMR spectra show two types of phosphorous species: one is the Keggin unit and the other is the decomposition product of TPA. The relative amount of each depends on TPA loading, calcination temperature, and the solvent used for the catalyst preparation. Under reaction conditions of 84°C and a benzene/1-olefin molar ratio of 10 (time 1 h), the most active catalyst, 15% TPA, calcined at 750°C, gave more than 98% olefin conversion with selectivity for 2-phenyl octane (53.5%) and 2-phenyl dodecane (47%).

Full text of DOI:10.1016/j.jcat.2004.09.024

Journal of Catalysis 231 (2005) 1–10
www.elsevier.com/locate/jcat
Zirconia-supported 12-tungstophosphoric acid as a solid catalyst
for the synthesis of linear alkyl benzenes
a
b
a,∗
Biju M. Devassy , F. Lefebvre , S.B. Halligudi
a
Inorganic Chemistry and Catalysis Division, National Chemical Laboratory, Pune 411 008, India
b
Laboratoire de Chemie Organometallique de Surface, CNRS-CPE, Villeurbanne cedex, France
Received 10 July 2004; revised 22 September 2004; accepted 23 September 2004
Available online 28 December 2004
Abstract
The liquid-phase alkylation of benzene with 1-octene and 1-dodecene was investigated with zirconia-supported 12-tungstophosphoric
◦
acid (TPA) as catalysts. We prepared the catalysts, with different TPA loading (5–20 wt% calcined at 750 C) and calcination temperatures
◦
(
15 wt% calcined from 650 to 850 C), by suspending hydrous zirconia in a methanol solution of TPA, followed by drying and calcination.
These catalysts were characterized by X-ray diffraction, DTG-DTA, FTIR pyridine adsorption, NH -TPD, and P MAS NMR spectroscopy
measurements. The catalyst with optimum TPA loading (15%) and calcination temperature (750 C) was prepared in different solvents and
characterized by P MAS NMR spectroscopy. The XRD results indicate that TPA stabilizes the tetragonal phase of zirconia. The catalysts
show both Brønsted and Lewis acidity, and 15% TPA on zirconia calcined at 750 C shows the highest acidity. P MAS NMR spectra show
two types of phosphorous species: one is the Keggin unit and the other is the decomposition product of TPA. The relative amount of each
depends on TPA loading, calcination temperature, and the solvent used for the catalyst preparation. Under reaction conditions of 84 C and a
31
3
◦
31
◦
31
◦
◦
benzene/1-olefin molar ratio of 10 (time 1 h), the most active catalyst, 15% TPA, calcined at 750 C, gave more than 98% olefin conversion
with selectivity for 2-phenyl octane (53.5%) and 2-phenyl dodecane (47%).

2004 Elsevier Inc. All rights reserved.
Keywords: Zirconia; 12-Tungstophosphoric acid; Solvent effect; Linear alkyl benzene
1
. Introduction
the above reaction [2,3]. Various catalysts, such as zeolites,
clays, heteropoly acids (HPAs), sulfated zirconia, and im-
mobilized ionic liquids, were tested for this reaction [4–8].
The Detal process developed by UOP uses solid acid cata-
lysts for the alkylation of benzene with heavy olefins under
liquid-phase conditions [1].
HPAs are a unique class of materials that are active
both in redox and acid catalysis [9,10]. These are poly-
oxometalates made up of heteropoly anions with metal–
oxygen octahedra as the basic structural unit. The Keggin-
type HPAs are the most important in catalysis, 12-tungsto-
phosphoric acid (TPA) is the usual catalyst of choice because
of its high acidic strength, relatively high thermal stability,
and lower oxidation potential compared with molybdenum
HPAs. They are strong Brønsted acid catalysts, and their
acidity is stronger than that of conventional solid acids like
zeolites and mixed oxides.
The alkylation of aromatic hydrocarbons with olefins is
applied on a large scale in the chemical industry. Alkyla-
tion of benzene with C10–14 linear alkenes is used for the
synthesis of linear alkyl benzenes (LABs), which are the
primary raw material for the production of LAB sulfonates,
a surfactant detergent intermediate [1]. Traditionally this re-
action is catalyzed by a homogeneous Lewis acid such as
AlCl3 or a strong Brønsted acid such as HF, which are highly
toxic, generate a substantial amount of waste, and cause se-
vere corrosion problems. At present, considerable efforts are
being made to find efficient, sustainable, recyclable, and eco-
friendly solid acid catalysts that can successfully catalyze
*
Corresponding author.
E-mail address: halligudi@cata.ncl.res.in (S.B. Halligudi).
0021-9517/$ – see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2004.09.024

2
B.M. Devassy et al. / Journal of Catalysis 231 (2005) 1–10
HPAs can be used either directly as a bulk material or in
Zirconium hydroxide was prepared by hydrolysis of 0.5 M
zirconyl chloride solution by the dropwise addition of aque-
ous NH3 (10 M) to a final pH of 10. The precipitate was
filtered and washed with ammoniacal water (pH 8) until it
was determined to be free from chloride ions by the silver
nitrate test. The hydrous zirconia thus obtained was dried at
supported form. The supported form is preferable because
of its high surface area compared with the bulk material (5–
2
−1
8
m g ) and better accessibility of reactants to the active
sites. Acidic or neutral solids, such as silica, active car-
bon, and acidic ion-exchange resin, which interact weakly
with HPAs, have been reported to be suitable as HPA sup-
ports [11]. Serious problems associated with these types of
materials are their susceptibility to deactivation during or-
ganic reactions due to the formation of carbonaceous de-
posits (coke) on the catalyst surface. The thermal stability
of HPAs is not high enough for conventional regeneration
◦
120 C for 12 h, powdered well, and dried for another 12 h.
Each time, 4 ml of methanol per gram of solid support was
used, and the mixture was stirred in a rotary evaporator for
8–10 h. After stirring, the excess methanol was removed at
◦
ca. 50 C under vacuum. The resulting solid materials were
◦
dried at 120 C for 24 h and ground well. We prepared a
◦
by the burning of coke at 500–550 C, as routinely used
series of catalysts with different TPA loading by changing
the TPA concentration in methanol. The dried samples were
in the case of zeolites and aluminosilicates [12]. Thus the
preparation of an active and stable HPA in supported form
is essential to utilize fully the potential of these materials as
catalysts.
In recent years zirconia has attracted much attention as
both a catalyst and a catalyst support because of its high
thermal stability and the amphoteric character of its sur-
face hydroxyl groups [13,14]. Extensive studies have been
carried out on zirconia modified by isopolytungstate (WOx–
ZrO2), which acts as an efficient solid acid catalyst [15–17].
But relatively few works are reported on zirconia modified
by heteropolytungstates as a catalyst in acid-catalyzed reac-
tions [18–20].
◦
then calcined in air at 750 C. Samples with 15% TPA were
◦
calcined between 650 and 850 C to understand the role of
the calcination temperature on the properties and activity of
the catalysts. All samples were calcined in shallow quartz
boats placed inside a 3-cm-diameter quartz tube in a tube
◦
−1
furnace. The samples were heated at a rate of 5 C min to
the final temperature, held for 4 h under static conditions,
◦
−1
and cooled at a rate of 5 C min to room temperature. To
understand the role of the solvent used for the preparation,
the catalyst with optimum TPA loading (15%) was prepared
with different solvents. For comparison, we also prepared a
catalyst with 15% TPA simply by grinding a mixture of zir-
conium hydroxide and TPA (neat) for 10 min, followed by
drying and calcination as mentioned above. The catalysts are
represented by x TZ-t, where x represents weight percent-
age, T represents TPA, Z represents zirconia, and t denotes
Recently we have shown that zirconia can be used as a
suitable support for heteropolytungstate, for example, for
2-TPA [21]. The present study deals with the preparation,
1
characterization, and application of zirconia-supported TPA
in the synthesis of LAB by alkylation of benzene with higher
linear alkenes such as 1-octene and 1-dodecene. The catalyst
with optimum TPA loading and calcination temperature was
prepared in different solvents to study the role of the solvent
used for catalyst preparation in deciding its catalytic activ-
ity in the above reaction. The catalysts were characterized
by X-ray diffraction, DTG-DTA, FTIR pyridine adsorption,
NH3-TPD, and 31P MAS NMR spectroscopy.
◦
calcination temperature ( C).
2.3. Characterization
The specific surface areas of the catalysts were measured
by N2 physisorption at liquid nitrogen temperature with a
Quantachrome Nova-1200 surface area analyzer and stan-
dard multipoint BET analysis methods. Samples were dried
◦
at 300 C in a dynamic vacuum for 2 h before N2 physisorp-
tion measurements.
2
. Experimental
X-ray diffraction (XRD) measurements of the catalyst
powder were recorded with a Rigaku Geigerflex diffrac-
tometer equipped with Ni-filtered Cu-Kα radiation (λ =
2
.1. Chemicals
1
.5418 Å). The volume percentage of the tetragonal phase
Zirconyl chloride (ZrOCl2 · 8H2O) and ammonia (25%)
(Vt) of the calcined samples was estimated with the formula
proposed by Toraya et al. [22]:
were obtained from S.D. Fine Chemicals, Ltd. (Mumbai).
1
-Octene (98%), 1-dodecene (95%), 12-TPA (H3PW12O40 ·
Im(11-1) + Im(111)
xH2O), and all other solvents (99.9%) were purchased from
Aldrich. Benzene (99.7%) was obtained from E. Merck In-
dia, Ltd. (Mumbai). All of the chemicals were used as re-
ceived without further purification.
Xm =
Vm =
,
Im(11-1) + Im(111) + It(111)
1
.311Xm
and Vt = 1 − Vm,
1
+ 0.31Xm
where Im(hkl) is the integral intensity of the (hkl) reflec-
2
.2. Catalyst preparation
tions of the monoclinic phase and I (111) is the intensity of
t
the (111) reflection of the tetragonal phase.
We prepared the catalysts by suspending a known amount
of dried zirconia powder in a methanol solution of TPA.
Thermogravimetric and differential thermal analysis (TG-
DTA) measurements were performed on a Setaram TG-DTA

B.M. Devassy et al. / Journal of Catalysis 231 (2005) 1–10
3
◦
9
2 apparatus from room temperature to 1000 C in flowing
netic stirrer; the temperature was measured at the reaction
mixture. The catalyst, freshly activated at 500 C for 2 h,
−1
◦
dry air (ca. 50 ml min ), with α-Al2O3 as a reference. For
each experiment 25–30 mg of the sample was used, with
a heating rate of 10 C min . TGA curves are depicted as
first derivative DTG of the direct weight loss traces.
The nature of the acid sites (Brønsted and Lewis) of
the catalyst samples was characterized by in situ FTIR
spectroscopy with chemisorbed pyridine. For catalysts with
different TPA loading, we performed pyridine adsorption
was weighed in the reactor, and then benzene was added
according to the desired proportion. Finally, 1-alkene was
added to obtain the desired molar ratio of benzene to 1-al-
kene. For example, a typical reaction mixture consists of
21.86 g (280 mmol) of benzene and 3.14 g of 1-octene
(28 mmol), together with 0.125 g of catalyst. After 1 h, the
reaction was stopped and the catalyst was separated. The fil-
trate was analyzed with a Shimadzu 14B gas chromatograph
and a HP-5 capillary column (cross-linked 5% ME silicone,
30 m × 0.53 × 1.5 µm film thickness), coupled with FID.
Products were identified by GC-MS.
◦
−1
(
Nicolet model 60 SXB) studies by heating in situ a self-
supporting wafer (20 mg) of the sample from room tem-
◦
◦
−1
perature to 400 C at a rate of 5 C min under vacuum
−6
◦
(
10 mbar). The samples were kept at 400 C for 3 h, fol-
lowed by cooling to 100 C. Pyridine vapor (10 mm Hg)
◦
was introduced into the cell and allowed to equilibrate for
◦
4
5 min, followed by evacuation at 100 C for 30 min, and
3. Results and discussion
the IR spectrum was recorded. Then the temperature was
slowly increased and IR spectra were recorded at different
temperatures up to 400 C.
The FTIR (Shimadzu SSU 8000) pyridine adsorption
studies for the 15 TZ catalyst calcined at different temper-
atures were carried out in the diffuse reflectance infrared
Fourier transform (DRIFT) mode. A calcined powder sam-
ple in a sample holder was placed in a specially designed
cell. The samples were then heated in situ from room tem-
3
3
.1. Characterization of the catalysts
.1.1. Surface area
◦
◦
The pure zirconium hydroxide dried at 120 C showed a
2
−1
◦
surface area of 331.6 m g . After calcination at 750 C,
the surface area decreased to 16 m g . Addition of TPA to
the support results in an increase in the surface area, which
2
−1
2
−1
reaches maximum at ca. 53.2 m g for 15% TPA loading
Table 1). Since this increase is affected by the change in
◦
◦
−1
perature to 400 C at a rate of 5 C min in a flowing stream
(
−1
◦
(
40 ml min ) of pure N2. The samples were kept at 400 C
for 3 h and then cooled to 100 C. Pyridine vapor (20 µl)
was then introduced under N2 flow, and the IR spectra were
recorded at different temperatures up to 400 C. We attained
a resolution of 4 cm after averaging over 500 scans for all
of the IR spectra reported here.
weight of the catalyst by loading, the surface area per gram
of the support indicated that the loading of TPA gradually
increased the surface area. This can be explained by the fact
that the added TPA forms a surface overlayer that reduces
the surface diffusion of zirconia and inhibits sintering. It sta-
bilizes the tetragonal phase of zirconia, which leads to an
increase in surface area.
The nominal WO3 loading corresponding to each TPA
loading and calcination temperature is determined to calcu-
late the tungsten (W) surface density from the measured sur-
face area. The W surface densities, expressed as the number
◦
◦
−1
The total amount of acid sites present on the catalysts was
evaluated by temperature-programmed desorption (TPD) of
NH3. It was carried out after 0.1 g of the catalyst sample
◦
was dehydrated at 500 C in dry air for 1 h and purged with
◦
helium for 0.5 h. The temperature was decreased to 125 C
under a flow of helium, and then 0.5-ml NH3 pulses were
supplied to the samples until no further uptake of ammonia
was observed. NH3 was desorbed in He flow by an increase
−2
of W atoms per square nanometer of area (W atoms nm ),
were obtained with the equation
◦
◦
−1
in the temperature to 540 C at 10 C min , and NH3 de-
sorption was measured with a TCD detector.
Surface density of W
ꢀ
ꢁ
23
=
[WO3 loading (wt%)/100] × 6.023 × 10
31
P MAS NMR spectra (Bruker DSX-300 spectrometer)
ꢂ
/
231.8 (formula weight of WO3)
were recorded at 121.5 MHz with high power decoupling
with a Bruker 4-mm probe head. The spinning rate was
ꢃ
BET surface area (m g−1) × 1018
2
×
1
0 kHz, and the delay between two pulses was varied be-
tween 1 and 30 s to ensure that complete relaxation of the
and are presented in Table 1. These results show that an in-
crease in TPA loading results in an increase in the surface
density of W. The specific surface area of TZ catalyst also
depends on the calcination temperature. The surface density
of W increased with the calcination temperature because of
the concomitant decrease in the ZrO2 surface area (Table 1).
31
P nuclei occurred. The chemical shifts are given relative
to external 85% H3PO4.
2
.4. Catalyst testing
The liquid-phase alkylation reactions were carried out in
a 50-ml glass batch reactor (slurry reactor) with an anhy-
drous CaCl2 guard tube. The temperature was maintained
with a silicon oil bath equipped with a thermostat and a mag-
3.1.2. X-ray diffraction
The XRD pattern of the catalysts with different TPA load-
ing and calcined at 750 C (Fig. 1) shows that the presence
◦

4
B.M. Devassy et al. / Journal of Catalysis 231 (2005) 1–10
Table 1
Surface area, phase composition, FTIR pyridine adsorption and NH -TPD data of various catalysts
3
a
b
Sample
SBET
SBET
(m g of ZrO )
2
Surface density
t-ZrO
(vol%)
B/L
(I
Acidity
2
2
−1
2
−1
−2
−2
(NH nm )
3
(m g of sample)
(W nm
)
/I
1536 1442)
c
Z-750
16.0
40.2
46.3
53.2
52.3
105.4
78.0
53.2
40.7
30.4
16.0
42.3
51.5
62.5
65.4
124.0
91.6
62.5
47.8
35.7
0
5.6
21.9
75.7
84.2
94.6
95.4
93.8
84.2
79.3
58.9
n.e.
n.e.
5
1
1
2
1
1
1
1
1
TZ-750
3.18
5.53
7.23
9.75
3.64
4.93
7.23
9.46
12.66
0.32
0.64
1.06
0.27
0.59
0.92
1.50
1.20
n.e.
3.44
4.68
5.21
2.65
2.86
n.e.
5.21
n.e.
3.86
0 TZ-750
5 TZ-750
0 TZ-750
5 TZ-650
5 TZ-700
5 TZ-750
5 TZ-800
5 TZ-850
a
For catalysts with different TPA loading, pyridine adsorption was carried out under vacuum and for 15% catalyst with different calcination temperature, it
was carried out under N flow.
2
b
Acidity values obtained from NH -TPD.
3
Not evaluated.
c
Fig. 1. X-ray diffraction patterns of (a) ZrO , (b) 5, (c) 10, (d) 15, and (e) 20% catalysts.
2
of TPA strongly influences the crystallization of zirconium
hydroxide into zirconia. Pure zirconia calcined at 750 C is
literature [19,23]. It can also be seen that up to a 15% TPA
◦
◦
loading, for catalysts calcined at 750 C and for 15% cata-
◦
mainly monoclinic, with only a small amount of the tetrag-
onal phase (Table 1). For catalysts with low TPA loading
lyst, up to 750 C calcination, no XRD peaks that could be
attributed to the polyacid or to its decomposition products
are observed, indicating that TPA is highly dispersed on the
support. When the TPA loading is higher than 15%, or when
◦
and calcined at 750 C, the XRD pattern can be described as
the sum of the monoclinic and tetragonal phases of zirconia;
this latter phase becomes dominant for catalyst with 20%
TPA. The tetragonal content of zirconia at a fixed loading
depends on the calcination temperature; for 15% catalyst,
◦
the calcination temperature exceeds 750 C for 15% load-
◦
ing, new peaks appear in the region of 23–25 C, which is
characteristic of WO3 [24]. This indicates that (i) TPA de-
composes, at least partially, and (ii) tungsten oxide formed
by this decomposition is present as relatively large particles.
Thus, when the TPA loading exceeds monolayer coverage
(see below), the excess TPA that does not interact with zir-
conia decomposes and the XRD spectra show the presence
of bulk crystalline WO3.
◦
zirconia exists mainly in the tetragonal form at 650 C, and
the tetragonal content decreases with increasing calcination
temperature (Table 1).
Thus, the added TPA stabilizes the tetragonal phase of
zirconia, and such stabilization of tetragonal ZrO2 in the
presence of TPA and other oxides has been reported in the

B.M. Devassy et al. / Journal of Catalysis 231 (2005) 1–10
5
Fig. 2. DTG (A) and DTA (B) curves of (a) hydrous ZrO , (b) 5, (c) 10, (d) 15, (e) 20, (f) 30% catalysts, and (g) TPA hydrate.
2
3
.1.3. Thermal analysis (DTG-DTA)
between TPA’s terminal W=O oxygen atoms and surface
≡Zr+ species or to the formation of water by the reaction of
TPA protons with ≡ZrOH hydroxyl groups [19,26]. Along
The DTG and DTA spectra of all of the samples dried
◦
at 120 C along with TPA hydrate are shown in Fig. 2. The
DTG analysis of pure TPA hydrate (H3PW12O40 · xH2O)
shows three stages of weight loss (endothermic effects) [25].
The first weight loss occurred from room temperature to
◦
with the exothermic effect at 445 C, another exothermic
◦
peak appeared in the temperature range of 500–690 C for
samples with 5–20% TPA, and the peak position moved to
higher temperature with increasing TPA content. The sample
◦
1
25 C because of the loss of physisorbed water. The second
◦
◦
one from 130 to 305 C accounts for the loss of crystalliza-
tion water, and the third one in the range of 370–550 C
accounts for the loss of 1.5 molecules of H2O originating
from all acidic protons. The total mass loss of the sample
corresponded to 21 H2O molecules per Keggin unit (KU).
The DTA shows an exothermic peak at 607 C due to the
complete decomposition of the Keggin structure to form a
mixture of oxides followed by its crystallization.
The zirconium hydroxide showed two endothermic and
one exothermic peak. The first endothermic peak from room
temperature to 180 C was due to the loss of physisorbed
water, and the second peak from 182 to 335 C was ascribed
to the dehydration and dehydroxylation of amorphous zir-
conium hydroxide. The exothermic effect at 445 C, termed
glow exotherm, is responsible for the crystallization of zir-
conium hydroxide to metastable tetragonal ZrO2. The DTG
behaviors of all of the TZ samples were almost similar, but
differed in DTA behavior. The DTA of the samples shows a
broad exothermic peak from 273 to 360 C with a peak maxi-
mum at 305 C, assigned to the formation of Zr–O–W bonds
with 30% TPA showed a sharp exothermic peak at 730 C.
◦
This is due to the crystallization of tungstate species formed
by the decomposition of TPA.
3.1.4. FTIR pyridine adsorption
◦
Adsorption of pyridine as a base on the surface of solid
acids is one of the most frequently applied methods for
the characterization of surface acidity. The use of IR spec-
troscopy to detect adsorbed pyridine enables us to distin-
guish among different acid sites. FTIR pyridine adsorption
spectra of catalysts with different TPA loading are shown
in Fig. 3. The catalysts showed Brønsted (B) and Lewis (L)
◦
◦
−
1
−1
acidity at 1536 cm and at 1442 cm ; the B/L ratios cal-
culated from the IR absorbance intensities [27] are given
in Table 1. The B/L ratio shows that the relative Brønsted
acidity increases with TPA loading up to 15% and decreases
with further loading. For 15% catalyst, the Brønsted char-
◦
◦
◦
acter increases up to 750 C calcination, and above 750 C
an increase in Lewis acidic character is observed (Table 1).
Thus, the catalyst 15 TZ-750 shows the maximum relative
◦
◦

6
B.M. Devassy et al. / Journal of Catalysis 231 (2005) 1–10
Fig. 4. TPD ammonia profile of (a) 5, (b) 10, (c) 15, and (d) 20% catalysts.
Fig. 3. The IR spectra of pyridine adsorbed on (a) 5, (b) 10, (c) 15, and
◦
(d) 20% catalysts after in situ activation at 300 C.
When the coverage exceeds a monolayer, the polyanion is
not stabilized by zirconia and decomposes to its oxides.
Brønsted acidity, and this corresponds to a monolayer of
TPA on ZrO2.
.1.6. 31P MAS NMR
3
Since Brønsted acidity increases with calcination tem-
perature up to monolayer coverage, we propose that during
calcination, support dehydroxylates and undergoes crystal-
lization, and during this process the interaction of HPA with
This is one of the most useful characterization techniques
for studying the state of phosphorous in heteropoly acids.
The chemical shift depends upon the phosphorous environ-
ment and on factors such as hydration number, addenda
+
the support is partially weakened, and some of the H ions
31
metal ion, support, etc. [28–32]. The P MAS NMR spec-
that are involved in interaction with the support are freed and
act as Brønsted acid sites.
tra of the catalysts with 5–20% TPA and 15% catalyst cal-
◦
cined from 650 to 850 C show that the state of phospho-
rous in the catalyst depends on TPA loading and calcination
temperature (Fig. 5). For low loadings and at a calcination
3
.1.5. TPD of NH3
This adsorption–desorption technique permits the deter-
◦
temperature of 750 C, a broad signal above −20 ppm is
mination of the strength of acid sites present on the cata-
lyst surface, together with total acidity. The TPD profiles
of the catalysts with different TPA loading are shown in
Fig. 4, and the amounts of sorbed NH3 per square nanome-
ter of the catalysts with different TPA loading and of cata-
observed, attributed to phosphorous (P–OH) in the Keggin
unit [21,28,32], and at higher loadings and at higher calci-
nation temperatures, a new signal appears below −20 ppm,
attributed to phosphorous oxide (P–O–P) resulting from the
decomposition of the polyoxometalate [19]. This phospho-
rous oxide represents 20 and 45% of the total phosphorous
for 15 and 20 TZ-750 catalysts and 80% for the 15 TZ-850
catalyst. To the best of our knowledge, this is the highest
thermal stability reported so far for any of the heteropoly
acid catalysts.
lyst with 15% TPA calcined at different temperatures (650,
7
◦
50, and 850 C) are given in Table 1. All samples show a
broad TPD profile, revealing that the surface acid strength
is widely distributed. It is evident from the data that there
is an initial increase in the acidity up to 15% loading, and
thereafter the acidity decreases. For 15 TZ catalysts, calcined
at different temperatures, the amount of adsorbed ammonia
increases with calcination temperature and reaches a maxi-
The 31P MAS NMR spectra of catalysts with differ-
ent TPA loading and calcination temperatures show that
the Keggin unit starts decomposition at a TPA loading of
◦
◦
mum at 750 C.
ca. 15% at 750 C. The results from different characteriza-
If we take into account the other characterizations, these
results can be explained as follows: at lower TPA loadings,
the polyoxometalate retains its structure and acidity, whereas
for higher loadings it decomposes, at least partially, into its
oxides. Thus, the highest acidity corresponds to about one
monolayer of polyoxometalate and is evident from 31P MAS
NMR spectra. So we can describe the evolution of the cata-
tion techniques suggest that a geometric monolayer of TPA
on zirconia was attained for 15 TZ-750 catalyst.
The 15 TZ-750 catalyst was selected to study the role of
the solvent used for catalyst preparation. The NMR spec-
tra of the catalysts prepared in different solvents show two
signals, one above −20 ppm and one below −20 ppm (not
shown). A careful examination of the spectra shows that the
signal intensities depend upon the solvent used for the cat-
alyst preparation, confirming the vital role played by the
solvent in stabilizing the Keggin structure of HPA in the final
catalyst. We have observed a direct relation between P–OH
◦
lysts as follows: up to a TPA loading of ca. 15% at 750 C, or
◦
for a 15% catalyst up to a calcination temperature of 750 C
(
i.e., up to a monolayer), the heteropolyanion is well dis-
persed on the zirconia surface and it retains its integrity.

B.M. Devassy et al. / Journal of Catalysis 231 (2005) 1–10
7
(A)
(B)
31
Fig. 5. P MAS NMR spectra of (A) catalysts with different TPA loading (a) 5, (b) 10, (c) 15, and (d) 20% and (B) 15% catalyst calcined at different
◦
temperature (a) 650, (b) 750, and (d) 850 C.
Table 2
vents varies with the dielectric constant of the solvent. In
solvents with a high dielectric constant (e.g., solvents like
DMF and methanol), the anion–anion interaction seems to
be absent or negligible and the catalyst prepared with such
solvents retains a maximum amount of TPA in Keggin form,
that is, such catalysts show the highest P–OH intensity. In
the case of DMF as a solvent, DMF is protonated by TPA
and the protonated DMF interacts with the Keggin anion, as
in the case of tetraalkylammonium salts of HPA, explain-
ing the lack of anion–anion interaction for the acid form in
DMF [35,36]. TPA in CH3OH forms protonated monomers
P–OH intensity, dielectric constant, octene conversion and product selectiv-
ity of 15% catalyst prepared in different solvents (conditions: total weight =
◦
2
5 g; catalyst weight = 0.125 g; temperature = 84 C; benzene/1-octene
(molar ratio) = 10; time = 1 h)
a
Solvent
Dielectric P–OH
Octene
MOB
DOB
constant intensity conversion selectivity selectivity
(%)
(mol%)
(%)
(%)
Neat
Ether
–
32
33
36
40
44
46
52
65
80
82
12.1
14.2
14.3
24.0
27.0
27.6
32.8
43.4
53.4
55.5
100
100
100
0
0
0
0.4
0.6
0.8
0.6
1.0
4.5
4.6
2.21
4.30
6.20
6.02
7.60
20.60
79.70
32.60
36.70
1,4-Dioxane
Acetic acid
Ethyl acetate
THF
Acetone
Water
99.6
99.4
99.2
99.4
99.0
95.5
95.4
+ +
(
CH3OH2 ) and dimers ((CH3OH)2H ), and these ions in-
teract more with Keggin ions than TPA does directly with
the solvent [37]. TPA in other solvents under study shows
varying degrees of anion–anion interactions arising from the
formation of hydrogen-bonding networks in the solvent, re-
sulting in a solvent structure that holds the anion within
interaction proximity [35]. Hence we presume that, as the
dielectric constant of the solvent decreases, the ability of the
solvent to undergo structured solvent and anion–anion inter-
action increases.
The hydroxyl group of the zirconium hydroxide tends to
be either positively or negatively charged below or above the
IEP (pH 6–7) [38]. Above the IEP, the hydroxyl groups tend
to be negatively charged, so that it is relatively difficult for
the anions to be adsorbed on the hydroxide surface, because
of electrostatic repulsion. On the other hand, the adsorption
Methanol
DMF
a
Relative amount of TPA in Keggin form.
intensity and the dielectric constant of the solvent [33] used
for the catalyst preparation (Table 2). It is evident from Ta-
ble 2 that the P–OH intensity (i.e., the amount of TPA in
Keggin form) increases with the dielectric constant of the
solvent, except for water as the solvent.
From spectroscopic studies Rocchiccioli-Deltcheff et al.
showed the existence of anion–anion interaction in the acid
form of HPA in solid state and in solution [34,35]. The
present work shows that the interaction of HPA with sol-

8
B.M. Devassy et al. / Journal of Catalysis 231 (2005) 1–10
of the anions is favored below the IEP of the support. Hence,
the adsorption of tungstophosphate ions on the surface of
zirconium hydroxide is favorable when water is used as the
solvent, which is not applicable in the case of nonaqueous
solvents. The solvent water has the highest dielectric con-
stant (79.7), but the catalyst prepared in water shows a P–OH
intensity lower than that prepared in methanol. Here one
must consider the stability of TPA in water. The maximum
limit for the stability of TPA in water is 0.1 M [35]. However,
the concentration of TPA in our experimental condition cor-
responds to ∼ 0.01 M. Thus the low P–OH intensity of the
catalyst prepared in water is attributed to the low stability of
TPA in water at low concentration.
Table 3
Octene conversion and product selectivity over various catalysts (condi-
◦
tions: total weight = 25 g; catalyst weight = 0.125 g; temperature = 84 C;
benzene/1-octene (molar ratio) = 10; time = 1 h)
Sample
Octene
conversion
TOF
MOB
selectivity
(%)
DOB
−
3
(10 mol
selectivity
(%)
−1 −1
(mol%)
mol
s
)
W
5
1
1
TZ-750
0 TZ-750
5 TZ-750
1.3
3.7
20.8
50.0
33.6
23.1
8.1
32.5
22.5
5.0
100
100
0
0
14.9
53.4
47.6
41.1
8.7
34.6
24.0
5.4
95.5
97.9
98.7
100
97.8
99.0
100
4.5
2.1
1.3
0
2.2
1.0
0
20 TZ-750
25 TZ-750
15 TZ-650
15 TZ-700
15 TZ-800
15 TZ-850
The different behavior of the catalyst prepared in differ-
ent solvents can also originate from limitation of diffusion
of TPA into the pores of the support. The size of the Keggin
anion (12 Å) is on the order of the pore size of the support
To establish the relation between catalytic activity, TPA
loading, and calcination temperature, the turnover frequen-
cies (TOF, (molmolW s ) of different catalysts were calcu-
lated (Table 3). The catalyst 15 TZ-750 showed the highest
TOF, and the surface density of this catalyst is found to be
7.23 W nm (Table 1), which corresponds to monolayer
coverage of TPA on zirconia [21]. This clearly indicates that,
irrespective of loading and calcination temperature, catalytic
activity depends on TPA coverage, and the highest activity
corresponds to monolayer of TPA on zirconia.
(
diameter < 2 nm) [39]; the rate of diffusion is controlled by
−
1
−1
the anion size, and the large polyoxoanions should have a
lower diffusion rate. Since the HPA anions are very weakly
solvated in solvents, the solubility of heteropoly acids de-
pends on the solvation of cations [40]. Thus the effective
size of the Keggin unit can vary from solvent to solvent,
and hence the diffusivity of the polyanion ultimately results
in different dispersion of HPA on the support and hence
the amount of intact TPA present in the ZrO2 surface af-
ter calcination. These results are in good agreement with the
observations of Fournier et al., who suggest the use of DMF
as a solvent to achieve good dispersion during the prepa-
ration of supported heteropoly acid catalysts [35]. We have
used methanol as a solvent because of its higher volatility
and ease in handling.
−2
Therefore, the catalyst with optimum TPA loading (15%)
and calcination temperature (750 C) was taken to study
◦
the role of the solvent used for catalyst preparation. The
variation of conversion and selectivity for the 15% catalyst
prepared in different solvents showed that the conversion
increases with P–OH intensity (Table 2). The catalyst pre-
pared without solvent showed the lowest octene conversion
of 12.1%, and the catalyst prepared in DMF showed the
highest conversion of 55.5%. As the conversion increased,
the selectivity for monoalkylation decreased from 100% for
the catalyst prepared without solvent to 95.4% for the one
prepared in DMF. It has to be noted that there is not much
difference in octene conversion for the catalysts prepared in
methanol and DMF.
3
3
.2. Catalytic activity
.2.1. Alkylation of benzene
With these catalysts, the main reactions were alkene
double-bond shift isomerization and benzene alkylation.
Monoalkylbenzenes (MOBs) were the main reaction prod-
uct, whereas dialkylbenzenes (DOBs) and alkene dimers
(
DIMs) appeared in small amounts. The conversion was ex-
The 15 TZ-750 catalyst prepared with methanol as a sol-
vent was used to study the effect of temperature in the range
pressed as the percentage of alkene converted into products.
The effect of TPA loading on octene conversion and prod-
uct selectivity is shown in Table 3. The 5 TZ catalyst showed
◦
of 55–84 C. The results indicate that temperature has a dras-
◦
tic effect on the conversion of octene (Fig. 6). At 55 C,
1
5
.3% conversion, and conversion increased to a maximum of
3.4% at 15% loading. The selectivity for mono- and dialky-
the conversion was only 0.4%, and it increased to 6.2% at
◦
75 C. An increase of 47% octene conversion was observed
◦
lated products depended on TPA loading, and the catalyst
with 15% TPA showed 95.5% monoalkylation selectivity.
To study the effect of calcination temperature, we used
when the temperature was increased from 75 to 84 C (boil-
ing point of the reaction mixture).
We studied the effect of the benzene/octene molar ratio
(4 to 10) on conversion and product selectivity while keep-
ing the total weight of the reaction mixture constant under
otherwise similar conditions (Fig. 6). As the molar ratio was
increased from 4 to 10, the octene conversion increased from
29 to 53.4%, and the selectivity for dialkylation decreased
from 13.4 to 4.5%. The dimerization of octene was also ob-
served at low benzene/octene molar ratios, and the dimer
◦
5 TZ catalyst calcined between 650 and 850 C. The cata-
1
◦
lyst calcined at 650 C showed 8.7% octene conversion, and
conversion increased to 53.4% at a calcination temperature
of 750 C (Table 3). With an increase in calcination temper-
ature, selectivity for dialkylation increased up to 750 C and
then decreased. The high activity of 15 TZ-750 catalyst is
due to its high Brønsted acidity.
◦
◦

B.M. Devassy et al. / Journal of Catalysis 231 (2005) 1–10
9
◦
zene with 1-octene at 84 C (2 wt% catalyst, 1 h, and 10:1
molar reactants ratio). To study the recycling, after the first
cycle the catalyst was separated (without washing), dried in
◦
air at 80 C for 4 h, and reused with fresh reaction mixture.
After the first use, the catalyst retains only 30% of initial
conversion. Thermal analysis of the separated catalyst shows
a continuous weight loss (3.8%) in the temperature range
◦
of 230–580 C, attributed to adsorbed products and alkene
oligomers. This result suggests that the loss in catalytic ac-
tivity is due the blockage of active sites of the catalyst by
heavy aromatics and oligomerized octene [7]. The deacti-
vated catalyst can be partially regenerated by refluxing with
dichloromethane. The dried catalyst obtained after refluxing
with dichloromethane gave 60% of initial conversion. Af-
ter the first recycling, the catalyst was reused twice without
appreciable loss in activity, after regeneration. The regenera-
tion was achieved by catalyst separation followed by drying
Fig. 6. Effect of reaction conditions on octene conversion and product se-
lectivity. (A) Effect of temperature (conditions: total weight = 25 g; cata-
lyst weight = 0.125 g; benzene/1-octene (molar ratio) = 10; time = 1 h).
◦
◦
at 120 C for 4 h and calcination at 600 C for 4 h in the pres-
ence of air. After the second regeneration, the catalyst shows
more than 96% initial conversion.
(
B) Effect of molar ratio (conditions: total weight = 25 g; catalyst
◦
weight = 0.125 g; temperature = 84 C; time = 1 h). (C) Effect of
◦
catalyst weight (conditions: total weight = 25 g; temperature = 84 C;
benzene/1-octene (molar ratio) = 10; time = 1 h).
4
. Conclusions
content decreased with increasing amount of benzene; it is
absent at a molar ratio of 10.
Zirconia-supported 12-tungstophosphoric acid acts as an
efficient and stable solid acid catalyst. The activity of the
catalyst in the alkylation of benzene with 1-octene is found
to depend on TPA loading, calcination temperature, and sol-
vent used for the catalyst preparation. Solvents like methanol
and DMF are the best for catalyst preparation. The catalyst
with a monolayer of TPA on zirconia (15 TZ-750) prepared
in methanol showed the highest activity. Under the reaction
The effect of catalyst concentration on octene conversion
shows that a catalyst concentration of 0.5 wt% (of the total
mass of the reactants) gave an octene conversion of 53.4%,
which increased to 98.1% with 2 wt% catalyst, while the
MOB selectivity remained almost the same (Fig. 6). Thus,
◦
under reaction conditions of 84 C, 1 h, and a benzene/olefin
molar ratio of 10, alkylation of benzene with 1-octene gave
◦
conditions of 84 C and a benzene/1-olefin molar ratio of 10
9
8.1% octene conversion with 93.7% MOB selectivity (iso-
(
time 1 h), 15 TZ-750 catalyst gave more than 98% olefin
conversion, 53.5% selectivity for 2-phenyl octane, and 47%
selectivity for 2-phenyl dodecane.
mer distribution 57.2% 2-PO, 26% 3-PO, and 16.8% 4-PO)
and 6.3% DOB selectivity.
◦
Under the above reaction conditions (84 C, 1 h, and
benzene/olefin molar ratio of 10), the reaction of benzene
with 1-dodecene was also performed. as the desired product
monododecyl benzene (phenyl dodecane—PD) is the pre-
cursor to linear alkyl benzene sulfonate, which is the most
widely used surfactant in the detergent industry. The 2 wt%
catalyst concentration gave a dodecene conversion of 41%,
and the conversion increases to 99% with 5% catalyst. As the
conversion increased from 41 to 99%, the monododecyl ben-
zene (MDB) selectivity decreased from 98.8 to 95.3%, and
didodecyl benzene selectivity increased from 1.2 to 4.7%.
At a dodecene conversion of 99%, the MDB isomer distrib-
ution was 49.4% 2-PD, 17.3% 3-PD, 11% 4-PD, and 22.3%
Acknowledgments
The authors thank the reviewer for constructive sugges-
tions on an earlier version of this paper. This work was
carried under a DST-SERC-sponsored project. B.M.D. ac-
knowledges CSIR New Delhi for a research fellowship.
References
[
[
[
1] J.A. Kocal, B.V. Vora, T. Imai, Appl. Catal. A 221 (2001) 295.
2] A. Corma, Chem. Rev. 95 (1995) 559.
3] A. Corma, H. Garcia, Chem. Rev. 103 (2003) 4307.
5
+6 PD. Sulfated zirconia was shown to be more active [7],
and zeolites are more selective catalysts for 2-isomer, but at
low activity for this reaction [41].
[4] S. Sivasankar, A. Thangaraj, J. Catal.138 (1992) 386.
[
[
[
5] J.L.B. Tejero, A.M. Danvila, US Patent 5 146 026 (1992) to Petro-
quimica Espanola, S.A. Petresa.
6] R.T. Sebulsky, A.M. Henke, Ind. Eng. Chem. Process Res. Dev. 10
For supported heteropoly acid catalysts, it is important to
study recycling of the catalyst, since the limited thermal sta-
bility of heteropoly acids usually prevents regeneration of
the deactivated catalyst by thermal methods. The recyclabil-
ity of 15 TZ-750 catalyst was tested in the alkylation of ben-
(1971) 272.
7] J.H. Clark, G.L. Monks, D.J. Nightingale, P.M. Price, J.F. White, J. Ca-
tal. 193 (2000) 348.

10
B.M. Devassy et al. / Journal of Catalysis 231 (2005) 1–10
[8] C. DeCastro, E. Sauvage, M.H. Valkenberg, W.F. Hölderich, J. Ca-
[26] J.B. Moffat, Catalysis by Acids and Bases, Elsevier, Amsterdam, 1985,
tal. 196 (2000) 86.
p. 157.
[
9] C.L. Hill (Ed.), Chem. Rev. 98 (1998) 1.
[27] B.H. Davis, R.A. Keogh, S. Alerasool, D.J. Zalewski, D.E. Day, P.K.
Doolin, J. Catal. 183 (1999) 45.
[28] M. Misono, Chem. Commun. (2001) 1141.
[29] C.J. Dillon, J.H. Holles, R.J. Davis, J.A. Labinger, M.E. Davis, J. Ca-
tal. 218 (2003) 54.
[
[
10] T. Okuhara, N. Mizuno, M. Misono, Adv. Catal. 41 (1996) 113.
11] S. Damyanova, L.M. Gomez, M.A. Banares, J.L.G. Fierro, Chem.
Mater. 12 (2000) 501.
[
[
[
[
[
12] I.V. Kozhevnikov, Chem. Rev. 98 (1998) 171.
13] K. Tanabe, T. Yamaguchi, Catal. Today 20 (1994) 185.
14] G.K. Chuah, Catal. Today 49 (1999) 131.
15] D.G. Barton, S.L. Soled, E. Iglesia, Topics Catal. 6 (1998) 87.
16] S. Kuba, P. Lukinskas, R.K. Grasselli, B.C. Gates, H. Knözinger, J. Ca-
tal. 216 (2003) 353.
[30] A. Ghanbari-Siahkali, A. Philippou, J. Dwyer, M.W. Anderson, Appl.
Catal. A 192 (2000) 57.
[31] A. Molnar, T. Beregszaszi, A. Fudala, P. Lentz, J.B. Nagy, Z. Konya,
I. Kiricsi, J. Catal. 202 (2001) 379.
[32] S. Uchida, K. Inumaru, M. Misono, J. Phys. Chem. B 104 (2000) 8108.
[33] I.M. Smallwood, Handbook of Organic Solvent Properties, Arnold,
London NW1 3BH, 1996.
[34] C. Rocchiccioli-Deltcheff, M. Fournier, R. Franck, R. Thouvenot, In-
org. Chem. 22 (1983) 207.
[
17] F.D. Gregorio, V. Keller, J. Catal. 225 (2004) 45.
18] E. Lopez-Salinas, J.G. Hermandez-Cortez, M.A. Cortes-Jacome,
J. Navarrete, M.E. Llanos, A. Vazquez, H. Armendariz, T. Lopez,
Appl. Catal. A 175 (1998) 43.
[
[
19] E. Lopez-Salinas, J.G. Hermandez-Cortez, I. Schifter, E. Torres-
Garcia, J. Navarrete, A. Gutierrez-Carrillo, T. Lopez, P.P. Lottici,
D. Bersani, Appl. Catal. A 193 (2000) 215.
[35] M. Fournier, R. Thouvenot, C. Rocchiccioli-Deltcheff, J. Chem. Soc.
Faraday Trans. 87 (1991) 349.
[36] G. Zukowska, J.R. Stevens, K.R. Jeffrey, Electrochem. Acta 48 (2003)
[
[
20] S. Patel, N. Purohit, A. Patel, J. Mol. Catal. A 192 (2003) 195.
21] B.M. Devassy, S.B. Halligudi, S.G. Hegde, A.B. Halgeri, F. Lefebvre,
Chem. Commun. (2002) 1074.
2
157.
37] Y. Hirano, K. Inumaru, T. Okuhara, M. Misono, Chem. Lett. (1996)
111.
[
1
[22] H. Toraya, M. Yoshimura, S. Somiya, J. Am. Ceram. Soc. C 67 (1984)
[38] A.A. Parks, Chem. Rev. 65 (1965) 177.
119.
[39] M. Valigi, D. Gazzoli, G. Ferraris, E. Bemporad, Phys. Chem. Chem.
Phys. 5 (2003) 4974.
[23] A. Khodakov, J. Yang, S. Su, E. Iglesia, J. Catal. 177 (1998) 343.
[24] JCPDS-International Center for diffraction data, 1990, Card 43-1035.
[25] I.V. Kozhevnikov, Catalysts for Fine Chenical Synthesis, Catalysis by
polyoxometalates, vol. 2, Wiley, New York, 2002, p. 15.
[
[
40] L.C.W. Baker, D. Glick, Chem. Rev. 98 (1998) 3.
41] J.L.G. de Almeida, M. Dufaux, Y. Ben Tarrit, C. Naccache, J. Am. Oil
Chem. Soc. 71 (7) (1994) 675.