Applied Catalysis A: General 193 (2000) 155171

Hydroisomerization of normal hexadecane with platinum-promoted

tungstate-modied zirconia catalysts
Shuguang Zhang, Yulong Zhang, John W. Tierney, Irving Wender

Department of Chemical and Petroleum Engineering, 1249 Benedum Hall, University of Pittsburgh, Pittsburgh, PA 15261, USA
Received 6 May 1999; received in revised form 3 September 1999; accepted 7 September 1999
Abstract
The present work examines the activity, selectivity and long-termstability of platinum-promoted tungstate-modied zirconia
(Pt/WO
3
/ZrO
2
) catalysts for the hydroisomerization of long-chain linear alkanes under relatively mild conditions, using
n-hexadecane as a model compound. The preparation of catalysts is described. A trickle-bed reactor is used to compare the
activities and selectivities of three Pt/WO
3
/ZrO
2
catalysts prepared by different methods and to investigate the effects of
tungsten loading and of reaction conditions for hydroisomerization of n-hexadecane. A run of 93.5 h was conducted using
the most highly active Pt/WO
3
/ZrO
2
catalyst which contains 0.5 wt.% well-dispersed Pt and 6.5 wt.% W. Reaction conditions
were manipulated and ve shutdown (feed stopped, reactor temperature lowered but H
2
ow rate maintained) and restart
operations were carried out. This catalyst showed high activity and stability. Considerable success has also been achieved in
converting n-hexadecane to isohexadecanes for 100 h at temperatures of about 220

Cand under H
2
pressure as lowas 160 psig
(1 psig =6.895 kPa). Our best results (highest i-C
16
yield) are a 79.1 wt.% n-C
16
conversion, 89.9 wt.% i-C
16
selectivity and
71.1 wt.% i-C
16
yield at 218

C, 160 psig H
2
, H
2
/n-C
16
mole ratio =2 and weight hour space velocity (WHSV) =1 h
1
. The
Pt/WO
3
/ZrO
2
catalyst is rugged and has properties which allow one to propose that it may attain commercial use after further
study. 2000 Elsevier Science B.V. All rights reserved.
Keywords: Tungstate-modied zirconia; Long-chain linear alkanes; Hydroisomerization; Isohexadecanes; Platinum
1. Introduction
The acid-catalyzed isomerization of alkanes is of
growing importance in determining the nature of
transportation fuels, including environmentally clean
high octane gasoline, high cetane diesel fuel, low
pour point jet fuel and lubricant base stocks. How-
ever, isomerization of long-chain alkanes generally
precedes cracking, leading to extensive undesired
cracking reactions; this tends to limit catalytic parafn
isomerization to C
4
-C
6
alkanes [1,2]. Alkanes above

Corresponding author.
E-mail address: wender@engrng.pitt.edu (I. Wender).
C
7
+ crack easily [3]; they have a large number of
methylene units that, in the presence of Lewis acids,
can form carbenium ions by abstraction of hydride
ions [4] or, in the presence of strong Bronsted acids,
can form carbonium ions [5]. It is highly desirable
to obtain environmentally clean catalysts that can be
used to isomerize long-chain parafns with minimum
cracking. Catalysts with a certain optimal balance of
metal and acid functions at suitable reaction condi-
tions are generally needed to suppress cracking in
order to achieve high isomerization selectivity for
long-chain parafns [68].
Efforts [912] have been made to avoid the use
of present commercial catalysts for the isomerization
0926-860X/00/$ see front matter 2000 Elsevier Science B.V. All rights reserved.
PII: S0926- 860X( 99) 00425- 1
156 S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171
of alkanes: HF is particularly dangerous while cat-
alysts containing halides such as AlCl
3
or sulfuric
acid are corrosive and pose signicant environmen-
tal challenges including disposal of wastes [2]. Re-
cently, there have been efforts by industrial [1315]
and academic researchers [1618] to obtain solid acid
catalysts which are environmentally more suitable for
alkane isomerization with minimal cracking. The ob-
jective is to combine high activity at low temperatures
and pressures with catalyst stability and regenerability.
Considerable interest has been focused on the use
of strong solid acids based on anion-modied zirco-
nium oxide catalysts [1620]. Sulfated zirconia was
found to be effective in catalyzing butane isomeriza-
tion at room temperature, a thermodynamically dif-
cult reaction that does not take place in 100% sulfuric
acid [21]. Farcasiu et al. have furnished evidence that
the high activity of sulfated zirconia for the isomer-
ization of parafns is associated with a redox process
which produces ion radicals leading to the formation
of carbenium ions [22,23]. A commercial process for
naphtha isomerization using sulfated zirconia has been
reported [24].
Hino and Arata [25] found that the substitution of
small amounts of WO
3
for SO
4
2
to give ZrO
2
/WO
3
resulted in a strong solid acid. The ZrO
2
/WO
3
is of
somewhat lower acid strength than ZrO
2
/SO
4
[2,4,26].
A number of researchers then found [2729] that,
as with ZrO
2
/SO
4
[30], addition of Pt to ZrO
2
/WO
3
was, in the presence of hydrogen, an active and se-
lective catalyst for n-alkane isomerization. Iglesia and
co-workers [2,31] found that Pt/WO
3
/ZrO
2
was more
selective than Pt/SO
4
/ZrO
2
in the hydroisomerization
of n-heptane, partly the result of increased hydride
transfer rates which limit the lifetime of adsorbed car-
benium ions. Iglesia and co-workers [1,2,31] proposed
that platinum particles dissociate molecular hydrogen
into hydrogen atoms which spill over to acid sites on
the catalyst surface to form protons and hydride ions.
The state of Pt on the surface of anion-modied zir-
conia is still in controversy [3237].
The hydroisomerization and/or hydrocracking of
n-hexadecane, FischerTropsch (FT) waxes and
polyolens in microautoclaves have been success-
fully carried out in this laboratory using Pt/SO
4
/ZrO
2
and Pt or Ni promoted WO
3
/ZrO
2
catalysts [3840].
Pt/WO
3
/ZrO
2
was more selective for n-hexadecane
hydroisomerization (300

C, 500 psig [cold] H

2
,
20 min) and more stable in the hydrocracking of
FT wax and polyolens (375

C, 1500 psig [cold]

H
2
, 25 min) than was Pt/SO
4
/ZrO
2
. Preliminary
results from XANES analyses on fresh and used
Pt/WO
3
/ZrO
2
catalysts indicated that platinum and
tungsten existed as zerovalent platinum and hexava-
lent tungsten, in both materials [38]. Hydroisomer-
ization of n-hexadecane using Pt/SO
4
/ZrO
2
catalysts
has also been studied by Davis et al. in a trickle bed
continuous reactor [41]. The results indicated that
high isomerization selectivity was difcult to achieve
even at low n-hexadecane conversion. There have
been few studies [14,42] of the promising catalyst,
Pt/WO
3
/ZrO
2
, for hydroisomerization of long-chain
linear parafns (>C
7
) in a continuous reactor system.
The objective of this study, carried out in a small
continuous trickle-bed reactor, is to examine the pos-
sibility of the use of Pt/WO
3
/ZrO
2
as a catalyst for
the isomerization, better termed hydroisomerization,
of n-alkanes above C
7
, using n-hexadecane as a model
compound. The activity, selectivity and long-term sta-
bility of this catalyst is investigated. A later paper will
apply this catalyst to the hydroisomerization of longer
n-alkanes in FT waxes and in various polymers, such
as polyethylene and polypropylene.
It should be pointed out that Chevron has introduced
a silicoaluminophosphate catalyst, SAPO-11, in their
Isodewaxing process [8]. Metal-promoted SAPO-11
is more selective than zeolites for long-chain parafn
isomerization [43]; however, it is generally used at
high pressure (1000 psig), high hydrogen to feed mole
ratios (30/1) and at about 340

C, a relatively high
reaction temperature. Isodewaxing seems to be aimed
mainly at producing lubricant oil, although transporta-
tion fuels can be produced from waxy feedstocks.
The Pt/WO
3
/ZrO
2
catalyst appears to be environ-
mentally benign, stable, robust and likely regenerable
and could be a viable alternative to Isodewaxing for
long-chain alkanes.
2. Experimental
2.1. Catalyst preparation and characterization
In their study of the isomerization of n-butane to
isobutane, Larsen and coworkers [28] compared two
S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171 157
Pt promoted tungstated zirconia catalysts which had
the same Pt and tungstate loadings. The rst cata-
lyst was prepared by adding Pt to WO
3
/ZrO
2
which
was already calcined at a high temperature (823

C);
it had a high Pt dispersion after another calcination at
500

C. The second one was prepared by calcination

at 823

C of a zirconia to which W and Pt had been

added by impregnation, resulting in a catalyst with a
low Pt dispersion. They found that the rst catalyst
increased selectivity toward hydrogenolysis products
in n-butane hydroisomerization while hydrogenoly-
sis was suppressed on the second catalyst. However,
Figoli et al. claimed that addition of Pt to a calcined
tungstated zirconia, followed by a second calcination,
generated a catalyst with better catalytic activity and
stability for n-hexane hydroisomerization than a cata-
lyst from a one step calcination (830

C, 3 h) of a Pt
and tungstate promoted zirconia, on which there were
less accessible Pt particles [44].
We studied the effects of two different prepa-
ration methods on the activity and selectivity of
Pt/WO
3
/ZrO
2
catalysts for n-hexadecane hydro-
isomerization. Two Pt (0.5 wt.%)/WO
3
/ZrO
2
cata-
lysts (I and II) were prepared. They had the same
composition, 0.5 wt.% Pt and 6.5 wt.% W. Both
started with zirconium hydroxide which had been
precipitated by addition of ammonium hydrox-
ide to zirconium chloride, but they differed in the
Fig. 1. Three Pt/WO
3
/ZrO
2
catalysts prepared by different methods.
following ways. There were two calcinations in the
preparation of I, the rst at 700

C after zirconium
hydroxide was modied by impregnation of ammo-
nium metatungstate ((NH
4
)
6
W
12
O
39
xH
2
O) solution
and the second at 500

C after the 700

C calcined
tungstated zirconia was impregnated with a platinum
salt (H
2
PtCl
6
6H
2
O); For II, there was only one cal-
cination at 700

C after co-impregnation of zirconium

hydroxide by the solution of tungsten and metal salts.
Based on our previous results, 700

C was chosen as
a calcination temperature to obtain WO
3
/ZrO
2
with
highest activity after Pt promotion. We used 500

C,
for the other calcination, as did Larsen et al. and
Figoli et al.
A third catalyst was prepared in a different way.
A zirconium hydroxide sample to which WO
3
had
already been added was obtained from Magne-
sium Elektron, Ltd. (MEL). In our laboratories, this
tungstated zirconia (6.7 wt.% W) sample was calcined
at 700

C and then was promoted by impregnation

with Pt, followed by another calcination at 500

C.
The resulting Pt (0.5 wt.%)/WO
3
/ZrO
2
catalyst was
labeled as PtWZ(MEL). The details of the prepara-
tion procedures for these three catalysts are shown in
Fig. 1.
Comparison of I and II (Fig. 2) showed that I was
more active for n-C
16
hydroisomerization. A series of
six catalysts with different W but with the same Pt
158 S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171
Fig. 2. Comparison of conversion and selectivity for hydroisomerization of n-C
16
using I and II [Reaction conditions: 300

C, 300 psig,
H
2
/n-C
16
mole ratio =2, 6th hour reaction results].
loadings were then prepared under the same conditions
used for I so as to study the effects of tungsten loading
and reaction variables.
BET specic surface areas (by nitrogen adsorp-
tion) and platinum dispersions (by carbon monox-
ide adsorption) were measured using a Micromerit-
ics ASAP 2010 instrument. X-ray diffraction exper-
iments were carried out on a Phillips XPert X-ray
diffractometer; a nickel-ltered Cu K radiation at
40 kV and 30 mA was used. Scanning electron mi-
croscopy (SEM) characterizations were carried out
using a Phillips SEM equipped with a XL30 Field
Emission Gun. Temperature-programmed desorption
of ammonia (NH
3
-TPD) experiments were conducted
using Altamira Instruments AMI-1. The catalyst was
pretreated in 30 cc/min He at 450

C for 1 h. Ammo-
nia was adsorbed at 100

C for 10 min and then He

(30 cc/min) was used to ush the catalyst for 1 h to
remove physisorbed ammonia. Then the temperature
was increased to 700

C at a rate of 10

C/min in a he-
lium ow (30 cc/min). TCD was used to monitor the
desorption process of ammonia.
2.2. Reactor system, operating procedure and
product analyses
Catalytic activity and selectivity tests were carried
out in a continuous trickle bed with a 0.305 in. i.d.
and 17 in. long stainless steel reactor. Reaction tem-
peratures were controlled by a computer and system
pressures by a back-pressure regulator. n-Hexadecane
(99 wt.% n-C
16
, from ICN Biomedical Inc.) was de-
livered from a feed tank into the reactor by a syringe
pump at a constant rate of 6.6 ml/h. Because of the
low ow rate, ow through the back-pressure regula-
tor was intermittent and there was a pressure surge of
10 psig every 2 min. Total gas ow rate (pure hydro-
gen or a mixture of hydrogen and helium) was kept
constant for catalyst tests. Input rates of hydrogen and
helium (make-up gas to keep gas ow rate constant
when studying pressure effects) were each controlled
by a separate mass ow meter.
The catalysts were crushed to 4060 mesh pellets
and placed into the center of the reactor after mixing
with an equal volume of quartz (5070 mesh). The
quartz had been washed with 10 wt.% hydrochloric
acid and double distilled deionized water and then cal-
cined at 700

C for 3 h before use. Blank runs showed

the pretreated quartz had no catalytic activity for n-C
16
conversion. It was also used as packing at each end of
the reactor. The length of catalyst bed was about 6.3 in.
when weight hour space velocity (WHSV) was 1 h
1
.
Since the ow rates of gas and liquid feed were main-
tained constant, variations in WHSV were made by
changing the amount of catalyst so that possible heat
and mass transfer or channeling problems would be
same for all runs. Before reactions, the catalysts were
activated at 450

C with 20 ml/min of air for 1 h. Gen-

erally, a steady state was reached on a fresh catalyst
S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171 159
within 5 h at 300

C on stream (TOS) but at longer

times at low reaction temperatures.
Two of the long runs included shutdown operations
and restarts. The shutdown procedure was as follows:
feed (n-C
16
) was halted but H
2
ow and pressure were
maintained, oven temperature was lowered to room
temperature and the H
2
ow decreased to 10 cc/min.
To restart a reaction, the H
2
ow was adjusted to
reach the desired ow rate and pressure, the oven tem-
perature was raised and n-hexadecane addition com-
menced. Catalysts experienced no deactivation during
a series of changes of reaction conditions (T, P and
H
2
/n-C
16
mole ratio) and ve shutdown and restart
operations or during a run of 100 h at about 220

C,
160 psig, WHSV=1 h
1
and H
2
/n-C
16
mole ratio =2.
In one case, the reaction was arbitrarily stopped and
the used but still active catalyst was calcined at 450

C
in 20 ml/min of air for 3 h to determine its activity af-
ter high temperature treatment.
Liquid products were collected in an ice-water
cooled vial for analysis using an HP-5980 GC. The
column was an HP-1 (Cross-linked Methyl Silicon
Gum, 25 m0.2 mm0.33 m). The GC oven tem-
perature increased from 40

C to 280

C at a rate of
5

C/min and then maintained for 10 min. The injector

temperature and the FID detector temperature were
320

C and 280

C, respectively. Helium was used as a

Fig. 3. NH
3
-TPD patterns of Catalysts I and II.
carrier gas at a rate of 50 cc/min. Cracking products,
normal hexadecane and isohexadecanes with only one
methyl group were identied using a GC (HP-5890
II)-MS (HP-5970). The column used in the GC-MS
was an HP-1 (Cross-linked Methyl Silicon Gum,
50 m0.2 mm0.33 m). The temperature program
for the GC-MS oven was the same as that for the GC
except that the rate was 3

C/min. Identication of C
16
isomers with two or more branches was achieved by
comparison with spectra found in the literature [7,45].
3. Results and discussion
n-Hexadecane conversion was dened as the dif-
ference between n-C
16
weight percentage in the feed
(100 wt.%) and that in the reaction products; the i-C
16
selectivity was calculated by dividing the i-C
16
weight
percentage in the products by n-C
16
conversion; the
i-C
16
yield was the weight percentage of branched C
16
isomers in the products. To obtain comparable n-C
16
conversions with different catalysts, activity and se-
lectivity tests were carried out at the same tempera-
ture, usually at 300

C. Reactions at lower tempera-

tures, from 210 to 250

C, using our most active cata-

lyst (I) were investigated and the inuence of reaction
pressure and H
2
/n-C
16
mole ratios were studied.
160 S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171
Fig. 4. Distribution of branched C
16
isomers versus n-C
16
conversion at various WHSV. Catalyst: I, 300

C, 300 psig, H
2
/n-C
16
mole
ratio =2, 6th hour results.
3.1. Comparison of catalysts I, II and PtWZ(MEL)
Catalysts I and II were tested at 300

C, 300 psig,
H
2
/n-C
16
mole ratio =2 and various WHSVs. Results
at steady states are shown in Fig. 2. n-Hexadecane
conversion decreased and i-C
16
selectivity increased
for both catalysts when the WHSV was increased,
but the n-C
16
conversion with catalyst II declined
more rapidly. At the same level of n-C
16
conversion,
90 wt.% for instance, the i-C
16
selectivity with I was
about 20 wt.% higher than that obtained with II. These
two catalysts had similar acidity, as shown by their
TPD patterns in Fig. 3. A possible explanation is
that the high temperature calcination (700

C), after
platinum salt impregnation used in the preparation of
II, resulted in a low degree of platinum dispersion.
These two catalysts have similar surface areas, about
67.5 m
2
/g, but carbon monoxide chemisorption exper-
Table 1
Surface properties of PtWZ(MEL), I and II
Catalyst I II PtWZ(MEL)
Pt (wt.%) 0.5 0.5 0.5
W (wt.%) 6.5 6.5 6.7
Surface area (m
2
/g) 67.5 67.6 49.2
Pore volume (cm
3
/g) 0.16 0.19 0.05
Pore diameter () 88.2 113.2 44.0
Platinum dispersion (%) 71.0 9.8 54.5
iments show the platinum dispersion of I to be 71.0%,
much higher than that of II (9.8%) (Table 1). Met-
als such as platinum [1], palladium or nickel [38]
can provide hydride ions by dissociation of hydrogen
molecules and enhance acidity by hydrogen spillover
[4648]. Lack of available hydride ions due to low
platinum dispersion may result in low activity or long
residence times for reaction intermediates and increase
the opportunity for cracking to occur.
The distribution of branched hexadecane isomers
obtained from the hydroisomerization of n-C
16
using
catalyst I varied at different n-C
16
conversions (Fig.
4). At high conversions, more normal hexadecanes
were converted to monobranched isomers by alkyl
(mainly methyl) shifts; monobranched i-C
16
isomer-
ized by further alkyl shifts so that the percentage
of multibranched isomers increased. The latter are
more easily cracked since the chances of -scission
increases when there are more tertiary carbon atoms
in isomers, but -scission was probably inhibited
by fast hydride ion transfer on this type of catalyst
[2,31]. Cracked products increased markedly as the
concentration of multibranched hexadecanes grew.
As intermediates between monobranched and multi-
branched isomers, the amount of dimethyltetrade-
canes remained fairly constant. A typical GC-MS
spectrum of hexadecane isomers is shown in Fig. 5.
A similar product distribution was obtained by Girgis
S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171 161
Fig. 5. A typical GC-MS spectrum of hexadecane isomers in the products of n-C
16
hydroisomerization.
and Tsao [7], who studied the hydroisomerization of
n-hexadecane using a Pt promoted proprietary zeolite.
They developed a reaction network for this catalyst.
In the network, monobranched isomers were proposed
to be the primary product, the multibranched isomers
(isomers with two or more branches) were the sec-
ondary products; cracked products were formed from
the multibranched isomers.The catalytic activities of
PtWZ(MEL) and I for n-C
16
hydroisomerization were
compared at 220

C, 300 psig, WHSV=1 h

1
and
H
2
/n-C
16
mole ratio =2. The results were essentially
the same (Table 2), although I had a higher surface
area (Table 1). It is interesting that these two catalysts
differed greatly in their XRD patterns (Fig. 6). The
peaks of tetragonal WO
3
were only observed in the
XRD pattern of I. A possible explanation is that the
tungsten oxide in PtWZ was so well dispersed that
the WO
3
clusters were too small to be identied by
XRD. If the tungsten oxide was well dispersed, the
high activity of PtWZ(MEL) would not be surprising
Table 2
Comparison of two catalysts having similar compositions
a
Catalyst n-C
16
conversion (wt.%) i-C
16
selectivity (wt.%) i-C
16
yield (wt.%)
PtWZ (MEL) 53.8 96.5 51.9
I 53.8 97.6 52.5
a
Reaction conditions: 220

C, 300 psig, H
2
/n-C
16
mole ratio =2/1, WHSV=1 h
1
. Results at steady state.
although its surface area was not as large as I. Almost
all zirconia in PtWZ(MEL) was in the tetragonal
phase, while there was much more monoclinic zirco-
nia in I. Both monoclinic and tetragonal zirconia had
activity for n-C
16
hydroisomerization. As determined
by SEM, PtWZ(MEL) consists entirely of particles of
about 2 m which are much larger than particles in I
(Fig. 7 and Fig. 10a).
3.2. Effect of tungsten loading on catalytic activity
BET surface areas, pore sizes and pore volumes of
six catalysts with different amounts of tungsten are
shown in Table 3. These catalysts were tested at the
same temperature (300

C), pressure (300 psig) and hy-

drogen to n-C
16
ratio (2/1 mol/mol) (Fig. 8). A rel-
atively high temperature was chosen so that steady
states could be reached within several hours (generally
5 h) and n-C
16
conversions over each catalyst were
162 S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171
Fig. 6. XRD proles of PtWZ(MEL) and I.
high enough for comparisons of activity and selectiv-
ity. Catalysts with 6.5 and 8 wt.% tungsten have high
hydroisomerization selectivities at high conversions.
Generally, about 8 wt.%tungsten loading is considered
as the amount which forms a monolayer coverage over
tungsten modied zirconia [31,49]. For Pt/WO
3
/ZrO
2
with different tungsten loadings, both the surface areas
and the relative amount of tetragonal zirconia shown
in Fig. 9 increased with increasing tungsten contents.
Fig. 7. SEM patterns of PtWZ(MEL).
The 15 wt.% W catalyst, which has the largest surface
area and the largest amount of tetragonal zirconia, was
not the most active catalyst for hydroisomerization of
n-hexadecane under these test conditions. Our opti-
mum tungsten loading range, 6.58 wt.%, is consistent
with that obtained in n-butane isomerization by Larsen
and Petkovic, who pointed out that inactive tungsten
oxide species could be formed above a 6 wt.% W limit
[29]. It has been suggested that a two layer coverage on
S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171 163
Table 3
Physical properties of Pt (0.5 wt.%)/WO
3
/ZrO
2
with different tung-
sten contents
Catalyst
a
A B C D E F
Tungsten content (wt.%) 3 4.5 6.5 8 10 15
Surface area (m
2
/g) 42.5 56.4 67.5 74.9 90.0 99.1
Pore diameter () 125.1 111.8 96.8 88.2 87.3 72.5
Pore volume (cm
3
/g) 0.14 0.16 0.16 0.16 0.20 0.18
a
Prepared with same procedure for I.
tungstated zirconia occurred at 16 wt.% W, generating
a maximum activity for n-pentane isomerization [49].
Parera et all found that the optimum tungsten loading
was 15 wt.% for n-butane isomerization [50]. The dif-
ference with tungsten loadings to achieve various ac-
tivities is likely the result of differences in preparation
methods, starting materials or test reactions.
At 50 wt.% isomerization selectivity, overall iso-
merization and cracking rates are equal. From Fig.
8, a 95 wt.% conversion of n-hexadecane with a
50 wt.% isomerization selectivity could be achieved
using the Pt/WO
3
/ZrO
2
(6.5 wt.% W) catalyst at
WHSV=8.5 h
1
(220

C, 300 psig, and H

2
/n-C
16
mole ratio =2). Above this conversion, crack-
ing will exceed isomerization so that more than
50 wt.% of converted n-hexadecane will form cracked
Fig. 8. Effect of tungsten content (315 wt.%) on n-C
16
conversion () and i-C
16
selectivity (- - - - ). [Catalysts prepared by same manner
as that for I; see Table 3 for details of A, B, C, D, E, F] reaction conditions:300

C, 300 psig, H
2
/n-C
16
mole ratio =2, 6th hour results.
products. Since the conversion at 50 wt.% selec-
tivity using catalyst I is high, it appears to be a
very good catalyst for the hydroisomerization of
n-hexadecane.
In the XRD patterns of Pt (0.5 wt.%)/WO
3
/ZrO
2
with different tungsten loadings (Fig. 9), the peak rep-
resenting tetragonal zirconia at 2 =30

was slightly
shifted toward a lower 2 value with increased tung-
sten loading. Peaks representing tetragonal tungsten
oxide (WO
3
) had the greatest area at 6.5 wt.% W and
then diminished at higher tungsten loadings, a likely
indication of strong interaction between tungstate
species and the zirconia support. The stabilization of
tetragonal zirconia by tungstate has been shown by
others [44,51]. Effects of variables during impregna-
tion, such as the pH values of the salt solutions and
the pore size of zirconium hydroxide, on catalytic
activity for butane isomerization have been studied
by Parera and coworkers [50].
The amount of platinum was too small to be iden-
tied by XRD. The SEM characterization shows that
there is little difference in texture between catalysts
containing 6.5 and 15 wt.% W (Fig. 10). No Pt or
WO
3
entities were observed in SEM characteriza-
tion, probably because their domain sizes were too
small.
164 S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171
Fig. 9. XRD patterns of Pt (0.5 wt.%)/WO
3
/ZrO
2
with different tungsten contents. Catalyst prepared using the method for I (see Fig. 1).
(a) 6.5 wt.% W; (b) 15 wt.% W.
TPD experiments using ammonia as the adsorbate
were carried out to obtain a measure of the acidities of
these Pt/WO
3
/ZrO
2
(4.515 wt.% W) catalysts. TPD
experiment was also carried out with an H-ZSM-5
Fig. 10. SEM patterns of Pt (0.5 wt.%)/WO
3
/ZrO
2
.
sample (SiO
2
/Al
2
O
3
=30, Na =0.02 mol%) for com-
parison (Fig. 11 ). The amounts of ammonia adsorbed
and the temperatures of ammonia desorption increased
with higher tungsten loadings. However, compared
S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171 165
Fig. 11. NH
3
-TPD patterns of HZSM-5 and Pt (0.5 wt.%)/WO
3
/ZrO
2
with different W loadings.
Fig. 12. Effect of operations of calcination, shutdown and restart on the activity of catalyst I for n-C
16
hydroisomerization.
166 S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171
Fig. 13. Effects of reaction variables: T, P and H
2
/n-C
16
mole ratio. Catalyst: I, WHSV=1 h
1
n-C
16
conversion () and i-C
16
selectivity
(- - - -).
with the HZSM-5 sample, Pt promoted tungstated zir-
conia has fewer acid sites and a weaker acid strength.
The acidity of these Pt (0.5 wt.%)/WO
3
/ZrO
2
catalysts
may not be the principal cause for high activity for
hydroisomerization.
3.3. Effect of high temperature treatment
A 72 h run started with fresh catalyst I (Fig. 12).
It took more than 20 h to reach a steady state at
230

C, 300 psig, WHSV=1 h

1
and H
2
/n-C
16
mole
ratio =2. The temperature was varied to obtain both
a high n-C
16
conversion and a high i-C
16
selectiv-
ity. After 34 h, the reaction was arbitrarily stopped
although the catalyst did not deactivate. The cata-
lyst was calcined in the reactor at 450

C, 3 h in a
20 cc/min air ow to see if the calcination would
affect its activity. When operated again at the same
conditions, the induction time was 9 h, shorter than
that with the fresh catalyst. Compared with the fresh
catalyst, n-C
16
conversion and i-C
16
selectivity with
the calcined catalyst were about 3 wt.% lower and
9 wt.% higher, respectively. Operations of shutdown
and restart had little effect on activity and selectiv-
ity. In the presence of hydrogen, no deactivation was
observed in our experiments.
3.4. Performance of catalysts under various
operations
The performance of catalyst I at different operating
conditions was investigated during a run of 93.5 h
TOS (Fig. 13). There were ve shutdown and restart
operations. Reaction variables, including T, P and
H
2
/n-C
16
mole ratios, were manipulated at different
levels. Steady states may not have been reached at
some points, but the results showed the response of
n-C
16
conversion and i-C
16
selectivity to changes
of reaction conditions and the rugged nature of this
catalyst.
S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171 167
Fig. 14. Effect of reaction temperature on n-C
16
conversion and i-C
16
selectivity. Reaction conditions: 300 psig, H
2
/n-C
16
mole ratio =2,
WHSV=1 h
1
.
Table 4
Comparison of two sets of reaction conditions
a
Reaction 300

C, WHSV=24 h
1
, 230

C, WHSV=1 h
1
,
conditions 300 psig, H
2
/n-C
16
mole 300 psig, H
2
/n-C
16
mole
ratio =2 ratio =2
n-C
16
conversion (wt.%) 82.8 85.9
i-C
16
selectivity (wt.%) 74.3 83.1
i-C
16
yield (wt.%) 61.5 71.4
a
Catalyst: I, steady-state results.
In the temperature range from 210 to 250

C,
lowering the temperature resulted in lower conver-
sion but higher selectivity (Fig. 14). An 85.9 wt.%
n-C
16
conversion with an 83.1 wt.% i-C
16
selectivity
was achieved at 230

C, WHSV=1 h
1
. Compara-
ble n-C
16
conversion (82.8 wt.%) was also obtained
at higher temperature and higher WHSV (300

C,
WHSV=24 h
1
), but the i-C
16
selectivity was only
74.3 wt.% (Table 4). This suggests that larger i-C
16
yields will be obtained at low temperatures and a
low WHSV, rather than at high temperatures and a
high WHSV. For a once-through process, operation at
230

C is desirable because of the high i-C

16
yield of
71.4 wt.%.
An increase of total reaction pressure but with con-
stant hydrogen partial pressure resulted in low conver-
sion with high selectivity (Fig. 15). The hydrogen to
n-C
16
mole ratio has an effect similar to that due to
total pressure (Table 5). At low H
2
/n-C
16
mole ratios
(low hydrogen partial pressures), high n-C
16
conver-
sions and low i-C
16
selectivity were obtained, indicat-
ing a negative reaction order with respect to hydrogen.
It is likely that H
2
and n-C
16
compete for adsorption
on the catalyst. A similar effect of hydrogen partial
pressure on conversion was observed by Larsen et all
in butane isomerization [28] and by Iglesia et al. in
heptane isomerization [31] using Pt/WO
3
/ZrO
2
. For
isomerizations of short chain parafns (<C
7
) using
168 S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171
Fig. 15. Effect of reaction pressure on n-C
16
conversion and i-C
16
selectivity. Reaction conditions: 230

C, H
2
/n-C
16
mole ratio =2,
WHSV=1 h
1
.
sulfated zirconia with or without Pt, both positive
and negative effects of H
2
pressure were reported
[19,52].
Catalyst I lost activity in the absence of hydrogen
at 180

C, but the activity could be restored by resum-

ing hydrogen addition. Essentially the same results
(85.9 wt.% n-C
16
conversion and 82.74 wt.% i-C
16
Table 5
Effect of H
2
/n-C
16
mole ratio on n-C
16
conversion and i-C
16
selectivity
T (
o
C) 230 200 180
TOS (h) 31 61.5 66.5 72.5 77.5 78.5 82.5
H
2
/n-C
16
mole ratio 2 1 1 0.5 0.5 0 0
n-C
16
conversion (wt.%) 85.9 96.5 28.9 36.1 10.1 21.0
a
6.1
b
i-C
16
selectivity (wt.%) 83.1 52.2 99.4 99.5 95.0 100.0
a
94.1
b
i-C
16
yield (wt.%) 71.4 50.4 28.7 36.0 9.6 21.0
a
5.7
b
i-C
16
distribution (wt.%)
Multibranched i-C
16
26.6 38.9 3.8 5.3 0 2.4 0
Dimethyltetradecanes 42.8 42.8 30.0 31.7 19.8 28.6 14.0
7- and 8-methylpentadecane 6.4 4.2 15.7 13.3 17.8 16.7 35.1
6-methylpentadecane 5.7 3.0 13.6 14.4
c
14.3
c
5-methylpentadecane 4.6 2.8 10.8 10.3 12.5 11.4 14.0
4-methylpentadecane 4.1 2.4 9.0 8.3 10.4 9.5 10.5
3-methylpentadecane 4.5 2.8 9.0 8.6 12.5 9.5 17.5
2-methylpentadecane 3.9 2.6 6.6 6.4 8.3 8.1 8.8
3-ethyltetradecane 0.8 0.4 1.4 1.7 0 0 0
a
The result was obtained at 1st hour after changing H
2
/n-C
16
ratio from 0.5 to 0.
b
The result was obtained at 5th hour after changing H
2
/n-C
16
ratio from 0.5 to 0.
c
The peak of 6-methylpentadecane was not separated from that of 7- and 8-methylpentadecane.
selectivity at TOS =31 h, 84.64 wt.% n-C
16
conver-
sion and 84.12 wt.% i-C
16
selectivity at TOS =93.5 h)
were obtained when reaction conditions were switched
back to 230

C, WHSV=l h
1
, 300 psig, H
2
/n-C
16
mole ratio =2 after shutting off hydrogen for 6 h. It
is likely that hydrogen cleans the catalyst surface and
restores hydroisomerization activity.
S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171 169
Fig. 16. XRD patterns of fresh, washed and dried, washed and calcined catalysts (Washing, drying and calcination were carried out on
Catalyst I used for 93.5 h).
At the end of the 93.5 h TOS run, the catalyst was
taken out of the reactor and washed with n-pentane.
Half of the used washed catalyst was dried at 60

C
in air for 1 h and was labeled as ; the other half was
calcined at 450

C for 3 h in air in an oven and was

labeled as . Characterizations using XRD showed
little difference among the fresh catalyst, and (Fig.
16). Catalyst contained 0.99 wt.% carbon.
Fig. 17. Performance of catalyst I at a low pressure (160 psig).
Catalyst I was operated at 160 psig for 100 h at
a temperature of about 220

C with no shutdowns,
restarts or high temperature calcinations during the
run (Fig. 17). No sign of deactivation was observed
at 100 h TOS, indicating a possibility of operation at
even lower pressure. At this point (218

C, 160 psig H
2
,
H
2
/n-C
16
mole ratio =2 and WHSV=1 h
1
), the best
results (highest i-C
16
yield) were a 79.1 wt.% n-C
16
170 S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171
conversion, 89.9 wt.% i-C
16
selectivity and 71.1 wt.%
i-C
16
yield.
Future work will study the selective hydrocracking
and hydroisomerization of high molecular weight ma-
terials, largely parafnic, such as FT waxes, natural
oils and polyolens using Pt promoted tungstated zir-
conia and similar catalysts in a continuous unit.
4. Conclusions
Calcination of a tungstate-modied zirconium hy-
droxide at 700

C followed by impregnation with

H
2
PtCl
6
6H
2
O solution and another calcination
at 500

C results in a highly active and selective

platinum-promoted tungstate-modied zirconia cat-
alyst (Pt/WO
3
/ZrO
2
) for the hydroisomerization of
n-hexadecane. The optimum range of tungsten load-
ing to achieve high isomerization selectivity at high
n-hexadecane conversion is between 6.5 and 8 wt.%.
Reaction conditions can be adjusted to obtain maxi-
mum i-C
16
yields or desirable branched isomers. This
catalyst is robust under proper operations, sustaining
changes of reaction conditions, shutdown and restart
and high temperature calcination. It retains a stable
activity and selectivity for n-C
16
hydroisomerization
when operated at a temperature of about 220

C and
a hydrogen pressure as low as 160 psig. A catalyst
prepared from an MEL tungstated zirconia has hy-
droisomerization activity similar to the best catalyst
prepared in our laboratories.
Acknowledgements
We thank the U.S. Department of Energy for
the nancial support of this work (Grant No.
DEFC2293PC93053). We are also grateful to
Prof. J. R. Blachere of the Materials Science and
Engineering Department, University of Pittsburgh for
his help on XRD and SEM analyses and discussions,
to Magnesium Elektron, Ltd. for providing catalyst
samples and to Magnesium Elektron, Inc. for carbon
analyses of a used catalyst.
References
[1] E. Iglesia, S.L. Soled, G.M. Kramer, J. Catal. 144 (1993) 238.
[2] D.G. Barton, S.L. Soled, G.D. Meitzner, G.A. Fuentes, E.
Iglesia, J. Catal. 181 (1999) 57.
[3] J. Weitkamp, Am. Chem. Soc. Symp. 20 (1975) 1.
[4] T.K. Cheung, B.C. Gates, Chemtech. Sep. (1997) 28.
[5] G.A. Olah, J. Am. Chem. Soc. 94 (1972) 808.
[6] J. Weitkamp, P.A. Jacobs, J.A. Martens, Appl. Catal. 8 (1983)
123.
[7] M.J. Girgis, Y.P. Tsao, Ind. Eng. Chem. Res. 35 (1996) 386.
[8] S.J. Miller, zeolites and related microporous materials: state
of the art: 1994, in: J. Weitkamp, H.G. Karge, H. Pfeifer, W.
Holderich (Eds.), Stud. Surf. Sci. and Catal., vol. 84, Elsevier,
Amsterdam, 1994, p. 2319.
[9] J. Kaspi, D.D. Montgomery, G.A. Olah, J. Org. Chem. 43
(1978) 3147.
[10] K. Tanabe, M. Misono, O. Yoshio, H. Hattori, New Solid
Acids and Bases, Kodansha, Tokyo, 1989.
[11] T. Yamaguchi, Appl. Catal. 61 (1990) 1.
[12] K. Arata, Appl. Catal. A 146 (1996) 3.
[13] C.D. Chang, J.G. Santiesteban, D.L. Stern, US Patent
5,345,026.
[14] S.L. Soled, W.E. Gates, E. Iglesia, US Patent 5,422,327.
[15] B.S. Umansky, M.V. Bhinde, C.Y. Hsu, C.S. Huang, US
Patent 5,629,257.
[16] K. Tanabe, M. Itoh, K. Morishige, H. Hattori, in: B. Delmon,
P.A. Jacobs, G. Poncelet (Eds.), Preparation of Catalysis,
Elsevier, Amsterdam, 1976, p. 65.
[17] M. Hino, K. Arata, J. Am. Chem. Soc. 101 (1979) 39.
[18] J.C. Yori, J.C. Luy, J.M. Parera, Appl. Catal. 46 (1989) 103.
[19] A. Sayari, Y. Yang, X. Song, J. Catal. 167 (1997) 346.
[20] S. Rezgui, B.C. Gates, Catal. Lett. 37 (1996) 5.
[21] K. Tanabe, H. Hattori, T. Yamaguchi, Critical Rev. Surf.
Chem. 1 (1) (1990) 1.
[22] D. Farcasiu, A. Ghenciu, J.Q. Li, J. Catal. 158 (1996) 116.
[23] A. Ghenciu, D. Farcasiu, J. Molec. Catal. A: Chem. 109
(1996) 273.
[24] C. Gosling, R. Rosin, P. Bullen, Petroleum Technol. Quarter.
(1997/1998) 55.
[25] M. Hino, K. Arata, J. Chem. Soc., Chem. Commun. (1988)
1259.
[26] M. Hino, K. Arata, J. Chem. Soc., Chem. Commun. (1980)
851.
[27] J.C. Yori, C.L. Pieck, J.M. Parera, Appl. Catal. 181 (1999) 5.
[28] G. Larsen, E. Lotero, S. Raghavan, R.D. Parra, C.A. Querini,
Appl. Catal. A 139 (1996) 201.
[29] G. Larsen, L.M. Petkovic, Appl. Catal. A 148 (1996) 155.
[30] T. Hosoi, T. Shimidzu, S. Itoh, S. Baba, H. Takaoka,
T. Imai, N. Yokoyama, Symposium on Preparation
and Characterization of Catalysts, Division of Petroleum
Chemistry, American Chemical Society, Los Angeles, 1988,
p. 562.
[31] E. Iglesia, D.G. Barton, S.L. Soled, S. Miseo, J.E.
Baumgartner, W.E. Gates, G.A. Fuentes, G.D. Meitzner, in:
J.W. Hightower, W.N. Delgass, E. Iglesia, A.T. Bell (Eds.),
11th Int. Congress on Catalysis 40 th Anniversary, Stud.
Surf. Sci. and Catal., vol. 101, Elsevier, Amsterdam, 1996,
p. 533.
[32] M. Hino, K. Arata, Catal. Lett. 30 (1995) 25.
[33] A. Sayari, A. Dicko, J. Catal. 145 (1994) 561.
[34] R. Comelli, Catal. Lett. 40 (1996) 67.
S. Zhang et al. / Applied Catalysis A: General 193 (2000) 155171 171
[35] J. Zhao, G.P. Huffman, Catal. Lett. 24 (1994) 385.
[36] K. Ebitani, H. Konno, T. Tanaka, H. Hattori, J. Catal. 135
(1992) 60.
[37] T. Shishido, T. Tanaka, H. Hattori, J. Catal. 172 (1997) 24.
[38] K.R. Venkatesh, J. Hu, W. Wang, G.D. Holder, J.W. Tierney,
I. Wender, Energy & Fuels 10 (1996) 1163.
[39] K.R. Venkatesh, Ph. D. Thesis, University of Pittsburgh, 1996.
[40] M.Y. Wen, I. Wender, J.W. Tierney, Energy & Fuels 4 (1990)
372.
[41] R.A. Keogh, D. Sparks, J. Hu, I. Wender, J.W. Tierney, W.
Wang, B. Davis, Energy & Fuels 8 (1994) 755.
[42] S.L. Soled, W.E. Gates, E. Iglesia, US Patent 5,648,589.
[43] R.J. Taylor, R. Pretty, Appl. Catal. A 119 (1994) 121138.
[44] S.R. Vaudagna, R.A. Comelli, N.S. Figoli, Appl. Catal. A
164 (1997) 265.
[45] E.N. Coppola, M.S. Thesis, University of Utah, 1987.
[46] K. Ebitani, J. Konishi, J. Hattori, J. Catal. 130 (1991) 257.
[47] K. Ebitani, J. Tsuji, H. Hattori, H.J. Kita, J. Catal. 135 (1992)
609.
[48] T. Shishido, H. Hattori, Appl. Catal. A 146 (1996) 157.
[49] J.G. Santiesteban, J.C. Vartuli, S. Han, R.D. Bastian, C.D.
Chang, J. Catal. 168 (1997) 431.
[50] J.C. Yori, C.R. Vera, J.M. Parera, Appl. Catal. A 163 (1997)
165.
[51] K. Arata, M. Hino, in: M.J. Phillips, M. Ternan (Eds.), Proc.
9th Int. Congress on Catalysis, Calgary, vol. 4, 1988, p. 1727.
[52] R.A. Comelli, Z.R. Finelli, S.R. Vaudagna, N.S. Figoli, Catal.
Lett. 45 (1997) 227.