The role of platinum in hexane isomerization over Pt/FeOy/WOx/ZrO2

Article abstract of DOI:10.1006/jcat.2001.3229

Hexane isomerization was studied in Pt-free and Pt-containing highly acidic iron oxide/tungsten oxide zirconia (FeO_y/WO_x/ZrO₂) catalysts. Hexane isomerization on WO_x/ZrO₂-based catalysts occurred mainly on monofunctional acid sites via intermolecular hydride transfer reactions between neutral molecules and carbenium ions. The highly acidic character of WO_x/ZrO₂ was such that hexane isomerization did not proceed via the classical dual functional mechanism. These results were consistent with those obtained in strong acid systems, and contrast those of zeolite-based catalysts, where the presence of metal is required to achieve high conversions. The presence of platinum led to a moderately higher level of deuterium exchange in the isohexane products, but a large increase in the deuteration of the normal hexane starting material. At ～ 20% conversion, nearly all of the unreacted normal hexane molecules contain at least one deuterium atom over Pt/FeO_y/WO_x/ZrO₂, while only 4-5% of the hexane was deuterated over Pt-free FeO_y/WO_x/ZrO₂. These results can be attributed to branched isomer exchange with deuterated Broonsted sites, supplemented by non-isomer-selective C-H bond activation by platinum.

Full text of DOI:10.1006/jcat.2001.3229

Journal of Catalysis 202, 25–33 (2001)
doi:10.1006/jcat.2001.3229, available online at http://www.idealibrary.com on
The Role of Platinum in Hexane Isomerization over Pt/FeOy/WOx /ZrO2
J. G. Santiesteban,1 D. C. Calabro,1 C. D. Chang, J. C. Vartuli, T. J. Fiebig, and R. D. Bastian2
ExxonMobil Research and Engineering, Corporate Strategic Research, Clinton Township, Rt. 22 East, Annandale, New Jersey 08801
Received August 31, 2000; revised March 13, 2001; accepted March 13, 2001
I. INTRODUCTION
Hexane isomerization was investigated in Pt-free and Pt-
Light paraffin (C5 and C6) isomerization is an increas-
ingly important process in the manufacture of high-octane
clean fuels. This process and the alkylation of isobutane
with olefins are the only refinery processes which produce
significant quantities of the ideal component of reformu-
lated gasoline: isoparaffins. Currently, there are two major
commercial paraffin isomerization catalysts, Pt on highly
chlorinated alumina, and Pt/mordenite. The high activity
of the chlorinated alumina-based catalyst allows operation
at low temperature (<125 C), which thermodynamically
favors formation of the high-octane branched paraffins.
These catalysts require the constant addition of a chloride-
containing compound for chlorine maintenance and are
very sensitive to moisture and sulfur. In addition, they are
highly corrosive and are subject to stringent environmental
control. Pt/mordenite does not have these drawbacks, but is
considerably less active and consequently requires higher
operating temperatures ( 260 C) where thermodynamic
constraints give lower yields of the branched paraffins.
Therefore, there is a need for a catalyst that combines
the advantages of the above mentioned catalysts, i.e., high
enough acidity to isomerize light paraffins at low tempera-
ture and operational and environmental friendliness.
Modified zirconia-based catalysts have been offered
as an alternative. Highly acidic, sulfated zirconia is very
active for n-butane isomerization at low temperatures (1).
Arata and Hino (2, 3) reported a method for preparing a
tungsten oxide/zirconia solid acid catalyst which involves
the impregnation of Zr(OH)4 with aqueous ammonium
metatungstate followed by calcination in air at 800–850 C.
When employed as a support for Pt, or Pt with a secondary
metal (e.g., Fe, Mn), very effective dual functional catalysts
were generated.
containing highly acidic iron oxide/tungsten oxide zirconia (FeOy/
WOx/ZrO2) catalysts. It has been found that hexane isomerization
on WOx/ZrO2 based catalysts takes place mainly on monofunc-
tional acid sites via intermolecular hydride transfer reactions
between neutral molecules and carbenium ions. The highly acidic
character of WOx/ZrO2 is such that hexane isomerization does
not proceed via the classical dual functional mechanism. These
results are consistent with those obtained in strong acid systems,
and contrast those of zeolite-based catalysts, where the presence of
metal is required to achieve high conversions. The relative rate of
hydrogen transfer and the thermodynamic limitation imposed on
the hexane isomers dictate the product distribution. The presence
of Pt is found to specifically facilitate the approach to equilibrium
of the 2,2-dimethylbutane isomer. This isomer is shown to form
from the most sterically-hindered carbenium ion intermediate. The
results suggest that Pt-catalyzed hydrogen activation generates
substantial amounts of surface atomic hydrogen which readily un-
dergoes hydrogen transfer with this hindered intermediate, which
would otherwise undergo very slow hydride transfer with neutral
molecules. Our results also suggest that Pt-catalyzed hydrogen acti-
vation may play a more classical role in keeping the surface concen-
tration of hexyl cations low, and consequently secondary reactions,
oligomerization/cracking, that give rise to cracked products, are
inhibited. The results of deuterium exchange measurements in D2
at low conversion, are best explained by invoking both acid- and
metal-facilitated deuteration pathways. The presence of Pt results
in a moderately higher level of deuterium exchange in the isohexane
products, but a large increase in the deuteration of the normal
hexane starting material. At 20% conversion, nearly all of the
unreacted normal hexane molecules contain at least one deuterium
atom over Pt/FeOy/WOx/ZrO2, while only 4–5% of the hexane
is deuterated over Pt-free FeOy/WOx/ZrO2. These results can be
attributed to branched isomer exchange with deuterated Brønsted
sites, supplemented bynon-isomer-selectiveC–H bond activation by
c
2001 Academic Press
Pt.
A more recent preparation method based on the co-
precipitation of the tungsten salt with Zr(OH)4 to form a
WOx /ZrO2 catalyst with twice the strong acid site density
of the materials prepared by impregnation, has been re-
ported (4). These novel solid acids have paraffin isomeriza-
tion activities approaching that of the chlorinated alumina-
based catalyst, and much higher than that of Pt/mordenite.
Coupled with their high thermal stability and excellent
1 To whom correspondence should be addressed. Fax: 908-730-3198
(Calabro); 908-730-3323 (Santiesteban). E-mail: david.c.calabro@
exxonmobil.com; jose.g.santiesteban@exxonmobil.com.
2 Current address: Air Products & Chemicals, Inc., 7201 Hamilton
Blvd., Allentown, PA 18015-1501.
25
0021-9517/01 $35.00
c
Copyright
2001 by Academic Press
All rights of reproduction in any form reserved.

26
SANTIESTEBAN ET AL.
regenerability, the WOx /ZrO2-based catalysts show pro-
mise as an environmentally friendly replacement for the (Fe) has not been clearly described in these systems. We
halogen-containing catalysts. have observed that Fe enhances the paraffin isomerization
It should be noted that the role of the secondary metal
Two general mechanisms have been proposed for paraf- activity of these catalysts. A detailed description of the role
fin isomerization depending on the acid strength of the cata- of iron is lacking; however, elucidation of this important
lyst: bifunctional (metal–acid) and monofunctional (both issue is outside the scope of this paper.
metal-only and acid-only). The former takes place on metal
catalysts supported on relatively low-acidity supports such
as amorphous silica–alumina and zeolites and includes
three consecutive steps (5); (i) paraffin dehydrogenation
on the metal; (ii) skeletal rearrangement of the olefin on
the acid site via a carbenium ion mechanism; and (iii)
rehydrogenation of the isomerized olefin on the metal.
Metals (6) and zeolites (4) are capable of isomerizing paraf-
fins by themselves, but at very low rates. On catalysts ex-
hibiting strong acidity, such as highly chlorinated alumina
and BF3/HF, paraffin isomerization can take place solely
on acid sites via a monofunctional mechanism featuring a
hydride-transfer chain reaction (7). In the case of the com-
mercial chlorinated alumina, the noble metal was proposed
to limit coke formation and acid site deactivation (8).
A monofunctional mechanism has also been proposed
for the high acidity, modified zirconia catalysts (9–11). In
this case, Pt has the important secondary function of hydro-
gen activation, which contributes to acid site creation and
suppressed coking. Alternatively, a monofunctional mech-
anism has recently been proposed for the less acidic Pt/H-
catalyst involving acid-catalyzed methyl shift and hydro-
gen transfer reactions, with the principal role of Pt being
to prevent the formation of large, local concentrations of
olefins (12).
In this paper, we examine the role of Pt in the isomeriza-
tion of n-hexane over novel, high-acidity FeO y/WOx /ZrO2
catalysts. The results indicate that the acid sites in these
catalysts are strong enough to perform the isomerization
reaction via a monofunctional (i.e., hydride transfer) mech-
anism similar to that observed in other strong-acid cata-
lysts. Hydride transfer reactions from neutral molecules to
carbenium ions are fast, allowing hexane isomers, with the
exception of 2,2-dimethylbutane, to reach equilibrium val-
ues at low temperatures ( 150 C). The addition of Pt to
the FeO y/WOx /ZrO2 catalyst is required to increase the
formation rate of 2,2-dimethylbutane. It is proposed that
Pt opens a new reaction pathway that facilitates the for-
mation of 2,2-dimethylbutane. This pathway eliminates the
high kinetic barrier associated with intermolecular hydride
transfer of the sterically hindered 3,3-dimethyl-2-butane
cation precursor. A similar Pt role has been suggested
for Pt/SO4/ZrO2 (10). The presence of Pt also maintains
a low surface concentration of hexyl cations, minimizing
oligomerization/cracking reactions that give rise to cracked
products. Finally, the ability of Pt to facilitate hydrogen acti-
vation is determined by measuring the extent of deuterium
uptake during hexane isomerization in D2.
II. EXPERIMENTAL
Catalyst Synthesis
The FeO y /WOx /ZrO2 catalyst was prepared by a method
similar to the co-precipitation procedure described else-
where (4). In brief, an aqueous solution containing ammo-
nium metatungstate and ammonium hydroxide was added
to a solution containing iron sulfate and zirconyl chloride.
The precipitate was filtered, washed, and dried overnight
at 85 C and calcined in dry air at 825 C for 3 h. At room
temperature, Pt was added via incipient wetness with
H2PtCl6 6H2O (from Aldrich) dissolved in distilled water.
The Pt-containing catalyst was air-dried at 120 C for 1 h,
then air-calcined at 300 C for 2 h. The elemental compo-
sition of the catalyst was 0.5 wt.% Pt/0.6 wt.% Fe/15 wt.%
W/ZrO2, with a BET surface area of 62 m2/g.
Catalytic Data Collection
The hexane isomerization reaction over both Pt-free and
Pt-containing FeO y/WOx /ZrO2 catalysts was carried out in
a down-flow, fixed-bed reactor. Typically, 4 cc ( 5.0 g) of
14/30-mesh catalyst was charged into the reactor. The cata-
lyst was positioned in the central region of the reactor be-
tween two layers of sand, and pretreated in flowing nitrogen
(GHSV = 2300 h 1) at 300 C for 1 h to ensure removal of
adsorbed water from the reactor and catalyst. After catalyst
pretreatment, the reactor was cooled to reaction tempera-
ture, then pressurized with hydrogen at 450 psig. Hexane
(Aldrich) was slowly fed into the reactor using a standard
high-pressure ISCO pump while maintaining a high hydro-
gen flowrate (>1500 GHSV); then hydrogen and n-hexane
flows were adjusted to give a 2H2/n-C6 mol ratio. After
passing through a back-pressure regulator downstream of
the reactor, the effluent was diluted with nitrogen, and the
combined stream was sent to on-line gas chromatography
sampling for analysis. Heating and insulating tapes were
wrapped around the stainless steel lines connecting the re-
actor exit and the gas chromatograph sampling valve to
maintain temperature around 160 C to avoid condensation
of the reaction products in the transfer line. A fused silica
capillary column (DB-1, 60 m) run at 28 C isothermal was
used to determine the reaction product composition.
Deuterium Uptake Measurements
The fixed bed reactor was initially lined out, as de-
scribed above, at 18–20% overall hexane conversion with a

THE ROLE OF Pt IN Pt/FeO y/WOx /ZrO2 CATALYSTS
27
FIG. 1. Mass ion fragmentation pattern of 3-methylpentane product formed in (a) H2 and (b) D2.
hydrogen co-feed. Once stable operation was achieved, the sufficient purging, liquid product sampling was repeated to
reactor effluent was redirected through a dry ice trap for determine if the product deuterium returned to its baseline
manual collection. After collecting hourly liquid product level in hydrogen. All liquid samples were analyzed by GC
samples for several hours, feed hydrogen was replaced with and GC/MS using HP5890 and HP5890/5971A instruments,
deuterium. Deuterium feed was maintained for 2–3 h to respectively.
completelypurge hydrogen from the reactor system, where-
Figure 1 compares typical fragmentation patterns for
upon several hourly liquid products were again collected. 3-methylpentane from liquid products collected in the
Finally deuterium was replaced with hydrogen and, after presence of H2 and D2. The incorporation of deuterium

28
SANTIESTEBAN ET AL.
TABLE 1
is evidenced by the conversion of distinct, relatively iso-
lated weight loss lines in H2, to envelopes of lines extending
to high m/z in D2. The deuterium content of each isomer,
except 2,2-dimethylbutane, was determined from the rel-
+
Hexane Isomerization over Pt-Free and Pt-Containing FeOy/
WOx/ZrO2 at 210 C, 2H2/n-C6 Molar Ratio, 450 psig, 2 LHSV (ml
n-C6 Feed per ml Catalyst per Hour). Data Obtained after Five
Days on Feed
ative abundance of the 86 m/z parent ion peak (C6H14
)
within the envelope of deuterated parent ions. Since 2,2-
dimethylbutane did not exhibit a parent ion peak, the in-
Catalyst
+
Prod. Dist. (wt.% )
Pt-free
Pt-containing
tense 71 m/z methyl-loss peak (C5H11 ) was used.
The intensity of the 86 m/z line indicates the amount
C1 + C2
0.03
1.3
9.3
0.6
1.2
0.9
+
of deuterium-free C6H14 present in a given isomer. Deu-
C3
terium incorporation could increase this ion mass to a po-
i-C4
+
tential maximum of 100 m/z (C6D14 ), with all intermediate
n-C4
i-C5
n-C5
2,2-DMB
2,3-DMB
2-MP
3-MP
n-C6
0.9
5.6
0.9
5.3
0.9
1.0
0.4
24.4
9.1
29.6
18.2
13.4
Tr
86.6
5.0
masses included in the deuterated envelope. Therefore, the
percentage of the 86 m/z intensity to the cumulative inten-
sity over the 86–100 m/z range was taken as the relative
+
9.0
concentration of C6H14
,
30.0
18.6
18.4
Tr
81.6
18.0
6.5
!
I86
%
=
100,
[1]
C7
+
P
86
14
n=0(I86+n
)
n -C6 conv., wt.%
C5 , wt.%
2,2-DMB/Total C6’s, %
2,2-DMB thermo. equil., %
25.8
25.9
where, for example, I86 equals the measured intensity at
86 m/z. As stated above, the 71 m/z peak was used for deter-
mining deuterium incorporation into 2,2-dimethylbutane.
In this case, the relative concentration of the undeuterated
25.9
+
C5H11 , % 71, was calculated from
ethane and methane. These cracked products are formed
by isomerization and ß-scission of large carbenium ion in-
termediates that may be formed from oligomerization of
surface hexyl cations. The ß-scission of carbenium ions fa-
vors the formation of stable tertiary carbenium ion frag-
ments that are precursors of branched alkanes, isobutane
and isopentane, and disfavors the formation of primary car-
benium ion precursors of methane and ethane (13). The
data obtained over the Pt-free catalyst also indicates that,
unlike the other C6 isomers, the yield of 2,2-dimethylbutane
is uniquely far away from its thermodynamic value.
!
I71
%
=
100,
[2]
P
71
11
n=0(I71+n
)
The incorporation of deuterium into the product would
decrease the C6H14 fragment concentration as reflected in
a decreased % 86 value.
+
III. RESULTS AND DISCUSSION
The hexane isomerization results obtained over the
Scheme 1 summarizes the various reaction pathways that
Pt-free and Pt-containing FeO y/WOx /ZrO2 catalysts are are proposed to give rise to the observed product distribu-
shown in Table 1. Although Pt plays a critical role in the gen- tion. The first step is the generation of a secondary n-hexyl
eration of olefinic intermediates in the classical dual func- cation, for example 3-hexyl carbenium (I). In addition to
tional mechanism, this is not nearly as important in strong direct protolysis, hydride abstraction, C–H oxidation, or
acids, such as WOx /ZrO2, as in relatively weak acids, such as protonation of trace olefins (14), (I) may also form step-
amorphous silica alumina and zeolites. Accordingly, the Pt- wise via minor acid cracking of hexane to generate carbe-
containing FeO y/WOx /ZrO2 catalyst exhibits only slightly nium ions (15), which then undergo rapid intermolecular
higher overall hexane conversion than the Pt-free catalyst. hydride transfer with hexane to give the secondary n-hexyl
Table 1 also indicates that under identical reactor condi- cation (I). Regardless of how they are formed, these sec-
tions, Pt/FeO y/WOx /ZrO2 favors the formation of much ondary linear cations rapidly isomerize and equilibrate with
more 2,2-dimethylbutane and much less cracked products the monobranched tertiary 2- and 3-methyl-pentyl cations
(C5 ), than its Pt-free analogue. The product distribution (II), which subsequently equilibrate with the di-branched,
obtained on the monofunctional Pt-free FeO y/WOx /ZrO2 tertiary 2,3-dimethyl-2-butyl cation (III). The tertiary 2,3-
catalyst can be rationalized in terms of a hydride transfer dimethyl-2-butyl cation (III), in turn, equilibrates with
+
mechanism along with ß-scission of large (C6 ) carbenium the secondary 3,3-dimethy-2-butyl cation (IV), with the
ion intermediates. The cracked product (C5 ) distribution equilibrium favoring the more stable tertiary ion. In ad-
obtained over this monofunctional catalyst consists primar- dition to these relatively rapid skeletal rearrangements,
ily of isobutane and isopentane with very small amounts of the surface hexyl cations may undergo two other reaction

THE ROLE OF Pt IN Pt/FeO y/WOx /ZrO2 CATALYSTS
29
SCHEME 1
pathways; (a) intermolecular hydride transfer to complete the 3,3-dimethyl-2-butyl cation with a paraffin molecule.
a surface turnover and form the corresponding alkane Thus, on the Pt-free FeO y/WOx /ZrO2 catalyst, rearrange-
product, or (b) oligomerization/cracking reactions that give ment of the secondary 3,3-dimethyl-2-butyl cation (IV) to
rise to C5 products. The relative rates of these two reac- form the more stable tertiary 2,3-dimethyl-2-butyl cation
tion pathways and the thermodynamic limitation imposed (III) is preferred relative to the sterically hindered inter-
on the hexane isomer concentrations, dictate the product molecular hydride transfer.
distribution. The observed thermodynamic yields of 2-
Overall, for the Pt-free catalyst, one may conclude that
methylpentane, 3-methylpentane, and 2,3-dimethylbutane thermodynamic limitations on the concentration of 2-
indicate that their hexyl cation precursors, surface species methylpentane, 3-methylpentane, and 2,3-dimethylbutane,
(II) and (III), respectively, are at equilibrium, and un- along with the high kinetic barrier involved in the formation
dergo rapid intermolecular hydride transfer. On the other of 2,2-dimethylbutane, give rise to a relatively large concen-
hand, the low yield of 2,2-dimethylbutane, as noted above, tration of surface hexyl cations. These cations may undergo
suggests that its formation proceeds over a high kinetic oligomerization/cracking reactions to give C5 products.
barrier.
Now, let us turn our attention to the bifunctional (metal/
With a t-butyl group alpha to the charged center, IV is acid) Pt/FeO y/WOx /ZrO2 catalyst for hexane isomeriza-
the most sterically protected carbenium ion in Scheme 1. tion. The results presented in Table 1 indicate that the
Therefore, formation of the bimolecular hydride transfer addition of Pt to the catalyst has a striking effect on the
transition state with a neutral molecule would be most hin- product distribution. Platinum decreases the formation of
dered. This is illustrated by comparing the relative reaction cracked products, C5 , and greatly enhances the formation
rates of SN2-type reactions of ß-substituted alkyl halides, of 2,2-dimethylbutane to give equilibrium concentrations
where neopentyl halides (as in IV) react about three or- of this high octane isomer. Therefore, while all of the other
ders of magnitude slower than isobutyl halides (16). It is C6 isomers shown in Scheme 1 readily form via a mono-
likely that the suppressed formation of 2,2-dimethylbutane functional acid mechanism, 2,2-dimethylbutane formation
is due to the hindered intermolecular hydride transfer of alone is only facile in the presence of a metal function.

30
SANTIESTEBAN ET AL.
SCHEME 2
As stated above, it has been proposed that the role approach of 2,2-dimethylbutane to equilibrium is due to
of Pt in these highly acidic, modified zirconia catalysts is the low surface concentration of its energetically unfavored
hydrogen activation (9–11). The Pt-catalyzed hydrogen secondary 3,3-dimethyl-2-butyl cation precursor relative to
activation, provides a ready source of H for the hydride the more stable tertiary hexyl cation precursors of the other
transfer reactions in Scheme 1 which yield the observed C6 hexane isomers, and the high kinetic barrier to intermolec-
isomers. Accordingly, it has been proposed that the greater ular hydrogen transfer with this cation. The surface con-
+
selectivity of Pt/WOx /ZrO2 for C7 alkane isomerization centration of the secondary 3,3-dimethyl-2-butyl cation is
(as opposed to cracking) relative to Pt/SO4/ZrO2 is due to expected to increase only after the tertiary hexyl cations
its greater effectiveness at generating and storing spillover are at thermodynamic equilibrium with their correspond-
(activated) hydrogen (17).
ing isomer products, consistent with recent findings that 2,2-
While this proposed Pt role would be expected to dimethylbutane is a secondary isomerization product (12).
enhance overall isomerization selectivity, it also pro-
vides an explanation for the specific enhancement in 2,2-
dimethylbutane formation for the Pt-containing catalyst.
Since 2,2-dimethylbutane is formed via the most sterically
hindered carbenium ion, its formation would benefit the
most by the increased presence of a hydrogen transfer
reagent substantially smaller than hexane. The generation
of highly reactive free hydride via the Pt catalyzed het-
erolytic cleavage of H2 to H+ and H , eliminates the rate-
limiting, high-energy intermolecular hydride transfer be-
tween the 3,3-dimethyl-2-butyl cation (IV) and a paraffin
molecule.
Another possible pathway to 2,2-dimethylbutane, which
avoids this high kinetic barrier, might involve the Pt-
catalyzed hydrogenation of surface olefinic intermediates
which are in equilibrium with the 3,3-dimethyl-2-butyl
cation (IV), as shown in Scheme 2.
The Effect of Pt on Gas Phase Deuterium Incorporation
During Hexane Isomerization
In order to further probe the effect of platinum, hexane
isomerization was conducted in both hydrogen and deu-
terium atmospheres over both Pt-free and Pt-containing
catalysts. It was anticipated that Pt-catalyzed hydrogen
activation would increase the extent of deuterium incorpo-
ration via easy deuteride (D: ) transfer with the carbenium
ions shown in Scheme 1.
Product samples were collected at 20% hexane conver-
sion during the initial unit lineout in H2, after switching to
The residence time or concentration of these surface
olefinic intermediates is expected to be rather small at
the low temperature and high hydrogen pressure used in
this study (12). Nonetheless, the ability of Pt to rapidly
hydrogenate trace olefins might shift the above equilib-
rium sufficiently to the right to substantially increase 2,2-
dimethylbutane production.
Even with the addition of Pt, the approach of 2,2-
dimethylbutane to equilibrium is the slowest among the
hexane isomers. Figure 2 summarizes the hexane isomer
yields obtained as a function of temperature. The data
clearly indicate that 2,2-dimethylbutane requires the high-
est temperature to reach its equilibrium concentration. 2-
Methylpentane, 3-methylpentane, and 2,3-dimethylbutane
easily achieve their equilibrium compositions at a tempera-
ture of about 302 F. The yield of cracked products (C5
)
increases significantly as 2,2-dimethylbutane approaches
FIG. 2. Temperature dependence of C6 isomer distribution obtained
equilibrium ( 30% at 350 F). As discussed above, the slow with Pt/FeO y/WOx /ZrO2.

THE ROLE OF Pt IN Pt/FeO y/WOx /ZrO2 CATALYSTS
31
TABLE 2
Deuterium Incorporation during Hexane Isomerization
Catalyst: FeO y/WOx /ZrO2
% C6H14
Pt/FeO y/WOx /ZrO2
% C6H14
Cofeed gas: H2 (init.)
D2
H2 (final)
% loss C6H14
H2 (init.)
D2
H2 (final)
% loss C6H14
Isomer
n-C6
2-MP
3-MP
2,3-DMB
2,2-DMB
93.2
93.8
93.9
94.1
94.4
88.9
15.7
14.7
6
93.3
93.5
93.9
93.9
94.7
4.6
83.3
84.3
93.6
77.5
93.3
93.4
93.6
92.7
94.8
2.8
92
97.0
97.7
99.4
99.6
99.9
2.11
0.58
0.33
0.12
91.5
92.6
90.4
92.7
21.2
D2 cofeed, and finally after switching back to H2 for the with Pt-free FeO y/WOx /ZrO2, this number only drops to
conclusion of the run, as described above. Using Eqs. [1] 75–90% for the isohexanes, but is very low (4.6% ) for n-
and [2], the GC/MS fragmentation intensity data obtained hexane. These results suggest that more than one mech-
on these samples was used to calculate an approximate per- anism for deuterium incorporation is operable in this
+
cent of each isomer present as un-deuterated C6H14 (%
or % 71). These results are shown in Table 2.
system. In addition to enhanced deuteride transfer via Pt-
catalyzed D2 activation, another pathway is needed to ex-
86
Two noteworthy observations in this data are that (1) plain the substantial degree of deuteration in the Pt-free
+
even in H2 none of the isomers are present as 100% C6H14
,
case.
The deuteration of isoparaffins in the absence of isomer-
according to this simple calculation, and (2) the calculated
values in H2 before and after D2 cofeed agree very well. ization or cracking has been observed with previously D2O-
The latter indicates that sufficient gas phase purging had exchanged zeolite and SO4/ZrO2 catalysts(19). The authors
taken place between liquid sampling to assure a steady- propose a carbenium ion mechanism with an olefinic inter-
state gas phase composition and that this is well reflected mediate which is deuteronated by a deuterium-exchanged
in the calculated % 86. The deviation from 100% for the acid site, AD, enhanced by proximate nucleophilic centers
+
C6H14 parent ion in H2 can be attributed to the presence on the solid acid surface (bifunctional Lewis base-assisted
of 13C (approx. 100 [6 1.11] = 93.34).
reference point, the percent loss in C6H+14 in D2 was cal- to an unfavorable surface configuration (electronic and/or
culated for each isomer as the percent decrease in % steric) with regard to Lewis base assistance (20). If we as-
Brønsted acidity) (14, 20). Interestingly, normal paraffins
in H2 as the nondeuterated are not deuterated over D–Y zeolite (21), possibly due
Taking the calculated %
86
86
between samples collected in hydrogen and deuterium sume gas phase D2 exchange with the Brønsted acid sites in
(Table 2 and Fig. 3). While this simple normalization cor-
rects for the presence of 13C, it does not consider errors in
the m/z = 86 intensity in the deuterated samples due to the
concomitant loss of H (or D) during ionization, i.e., m/z 1,
+
m/z 2, etc. This causes an overestimation of C6H14 (un-
derestimation of % loss) since multiple ions can contribute
to the m/z = 86 intensity. Methods are available which
correct for these contributions (18); however, we did not
employ them since the estimated error is small relative to
the magnitude of the effects we observed.
Figure 3 shows the degree of deuteration (as % loss
of C6H14) for all of the C6 isomers. The results show
that a higher percentage of all of the product isomer
molecules contain at least one deuterium atom when Pt
is present, consistent with Pt-catalyzed hydrogen activa-
tion. Nonetheless, with the exception of n-hexane, the de-
gree of deuteration decreases only slightly in the absence
of Pt. For Pt/FeO y/WOx /ZrO2, 95–100% of the product
contains at least one deuterium per molecule, whereas
FIG. 3. The effect of Pt on the extent of deuterium incorporation into
the product isomers with D2 cofeed.

32
SANTIESTEBAN ET AL.
TABLE 3
very fast relative to the hydrocarbon desorption rate (25).
If so, a considerable yield of multideuterated molecules
would be expected, consistent with the envelope of high
mass parent and fragment ion peaks observed in Fig. 1. In
the case of Pt/FeO y/WOx /ZrO2, the presence of spillover
hydrogen (17, 26) may supplement the chemisorbed hydro-
gen to maintain a large inventory of activated H (D) atoms
for ready exchange with the surface-bound hydrocarbon
species.
Hexane Isomerization Products Obtained During Deuterium
Exchange Experiments at Low Conversion
Catalyst
H2/D2
C
% C5
% 2,2-DMB % C6 conv.
FeO y /WOx /ZrO2
H2
D2
197
197
0.35
0.28
0.44
0.35
22
18.9
Pt/FeO y/WOx /ZrO2
H2
D2
165
165
0.04
0.06
0.4
0.25
20.8
15.9
It is important to note that the metal-assisted deuterium
exchange described above is indiscriminate with respect to
the source of the C–H bond. Assuming no significant differ-
ence in the chemisorption tendenciesamongthe C6 isomers,
this pathway would incorporate deuterium comparably for
all of the product isomers, as observed.
WOx /ZrO2, as observed in zeolites (22), this acid-catalyzed
deuterium exchange mechanism would incorporate deu-
terium into the isohexanes but not the unreacted n-hexane,
as observed in Fig. 3.
Overall, the results in Fig. 3 suggest the following. Pt-
catalyzed deuterium activation results in greater deuterium
incorporation in the presence of Pt than in its absence. Deu-
terium exchange via deuterated bronsted acid sites could
explain the still high degree of deuteration with the Pt-free
catalyst. Based on previous observations on D–Y zeolite
(21), this might also explain the conspicuous lack of ex-
change with n-hexane in the absence of Pt. Finally, non-
selective, metal-assisted, C–H bond activation provides an
explanation for the uniformly high degree of reactant and
product deuteration even at low hexane conversion.
The data in Fig. 3 were obtained at low (16–22% ) hex-
ane conversions. The product selectivities for these runs are
shown in Table 3. In contrast to the results obtained at much
higher conversions (Table 1), the addition of Pt has no effect
on the formation of 2,2-dimethylbutane, but is still seen to
reduce the amount of cracking. Thus, at 20% conversion,
Pt has a much reduced impact on product selectivity and
incorporates modestly more deuterium into the isomerized
products. Clearly the most pronounced effect Pt has at low
conversion is the dramatic increase in the deuterium con-
tent of the n-hexane.
The extensive deuteration of n-hexane by the Pt/FeO y/
WOx /ZrO2 catalyst could occur via two metal-assisted path-
ways. As stated above, Pt-catalyzed H2 (D2) activation
should enhance the rates of all of the hydride (deuteride)
IV. CONCLUSIONS
Our results indicate that paraffin isomerization on
transfer steps in Scheme 1, including that of the 3-hexyl WOx /ZrO2-based catalysts takes place mostly on mono-
cation (I) to regenerate deuterated n-hexane. However, for functional acid sites via intermolecular, hydride-transfer
this to explain the almost complete loss of C6H14 hexane reactions between neutral molecules and carbenium ions.
with the Pt/FeO y/WOx /ZrO2 catalyst (Fig. 3), it would re- The highly acidic character of this material is such that high
quire that all of the feed be converted to the 3-hexyl cation conversions are achieved without the need for a hydrogena-
(I), of which only 20% proceeds on to isomerize, while tion/dehydrogenation component.
the remaining 80% reverts via deuteride transfer to linear
The presence of Pt is found to specifically facilitate the
C6H14 x Dx . Thus the pool of spillover deuterium generated approach to equilibrium of the 2,2-dimethylbutane iso-
by Pt(20), would have to provide enough of a driving force mer. This isomer is shown to form from the most ster-
to shift the kinetics to greatly favor the deuteride transfer ically hindered carbenium ion intermediate. The results
reaction relative to the facile conversion of the highly reac- suggest that Pt-catalyzed hydrogen activation generates
tive sec-3-hexyl carbenium ion to the more stable tertiary substantial amounts of surface atomic hydrogen which
intermdiates (II & III), and on to the isomerized products. readily undergoes hydrogen transfer with this hindered
This is unlikely.
intermediate, which would otherwise undergo very slow
It has been reported that H–D exchange between n- hydrogen transfer with neutral molecules. Our results also
C7H16 and D2 occurs much more rapidly than isomeriza- suggest that Pt-catalyzed hydrogen activation may play a
tion on Pt/WOx /ZrO2 (17, 23). This requires an alternative more classical role in keeping the surface concentration of
exchange pathway which does not require a carbenium ion hexyl cations low, and consequently secondary reactions,
intermediate, such as the direct deuteration of the n-hexane oligomerization/cracking, that give rise to cracked products,
feed via C-H bond activation by Pt (24, 25). Deuterium ex- are inhibited.
change with alkane adsorbed on metal surfaces is known to
The results of deuterium exchange measurements in D2
be very easy (24). The rate of exchange between the pool at low conversion are best explained by invoking both
of H atoms bound on the surface of the Pt particles, and acid- and metal-facilitated deuteration pathways. The pres-
the H atoms of the chemisorbed hydrocarbon radical, is ence of Pt results in a moderately higher level of deuterium

THE ROLE OF Pt IN Pt/FeO y/WOx /ZrO2 CATALYSTS
33
exchange in the isohexane products, but a large increase 10. Iglesia, E., Soled, S. L., and Kramer, G. M., J. Catal. 144, 238–253
(1993).
in the deuteration of the normal hexane starting mate-
rial. At 20% conversion, nearly all of the unreacted nor-
mal hexane molecules contain at least one deuterium atom
over Pt/FeO y/WOx /ZrO2, while only 4–5% of the hexane
11. Larsen, G., Lotero, E., and Parra, R. D., in “Proceedings, 11th Interna-
tional Congress on Catalysis, Baltimore, 1996” (J. W. Hightower, W. N.
Delgass, E. Iglesia, and A. T. Bell, Eds.), p. 543. Elsevier, Amsterdam,
1996.
is deuterated over Pt-free FeO y/WOx /ZrO2. These results 12. Chu, H. Y., Rosynek, M. P., and Lunsford, J. H., J. Catal. 178, 352
(1998).
can be attributed to branched isomer exchange with deuter-
ated Brønsted sites, supplemented by non-isomer-selective
C–H bond activation by Pt.
13. Weitkamp, J., Jacobs, P. A., and Martens, J. A., Appl. Catal. 8, 123
(1983).
14. Sommer, J., Jost, R., and Hachoumy, M., Catal. Today 38, 309
(1997).
15. Buchanan, J. S., Santiesteban, J. G., and Haag, W. O., J. Catal. 158, 279
ACKNOWLEDGMENTS
(1996).
16. Eliel, E. L., in “Steric Effects in Organic Chemistry” (M. S. Newman,
Ed.), p. 76. Wiley, New York, 1956.
17. Iglesia, E., Barton, D. G., Soled, S. L., Miseo, S., Baumgartner,
The authors thank Mobil Technology Company for supporting this
work.
J. E., Gates, W. E., Fuentes, G. A., and Meitzner, G. D., in “Proceed-
REFERENCES
ings, 11th International Congress on Catalysis, Baltimore, 1996” (J. W.
Hightower, W. N. Delgass, E. Iglesia, and A. T. Bell, Eds.), p. 533.
Elsevier, Amsterdam, 1996.
1. Hsu, C.-Y., Heimbruch, C. R., Armes, C. T., and Gates, B. C., J. Chem.
Soc. Chem. Commun., 1645 (1992).
2. Arata, K., and Hino, M., in “Proceedings, 9th International Congress
on Catalysis, Calgary, 1988” (M. J. Phillips and M. Ternan, Eds.),
p. 1727. Chem. Institute of Canada, Ottawa, 1988.
18. Price, G. L., and Iglesia, E., Ind. Eng. Chem. Res. 28, 839 (1989).
19. Sommer, J., Habermacher, D., Hachoumy, R. J., and Reynaud, A.,
Appl. Catal. A 146, 193 (1996).
20. van Santen, R. A., and Kramer, G. J., Chem. Rev. 95, 637 (1995).
21. Sommer, J., Hachoumy, M., and Garin, F., J. Am. Chem. Soc. 116, 5491
(1994).
3. Hino, M. and Arata, K., J. Chem. Soc. Chem. Commun., 1259
(1988).
4. Santiesteban, J. G., Vartuli, J. C., Han, S., Bastian, R. D., and Chang,
C. D., J. Catal. 168, 431 (1997) .
22. Kubelkova, L., and Novakova, J., Collect. Czech. Chem. Commun.
41(2), 325 (1976).
5. Weisz, P. B., in “Advances in Catalysis” (D. D. Ely, H. Pines, and P. B.
Weisz, Eds.), Vol. 13, p. 137. Academic Press, New York, 1963.
6. Gault, F. G., in “Advances in Catalysis” (D. D. Ely, H. Pines, and P. B.
Weisz, Eds.), Vol. 30, p. 1. Academic Press, New York, 1981.
7. Pines, H., in “The Chemistry of Catalytic Hydrocarbon Conversions.”
p. 13. Academic Press, 1981.
23. Barton, D. G., Soled, S. L., Meitzner, G. D., Fuentes, G. A., and Iglesia,
E., J. Catal. 181, 57 (1999).
24. Kemball, C., in “Advances in Catalysis” (D. D. Ely, H. Pines, and P. B.
Weisz, Eds.), Vol. 11, p. 223. Academic Press, New York, 1959.
25. Paal, Z., and Menon, P. G., “Hydrogen Effects in Catalysis.” Chaps. 15
and 17. Dekker, New York, 1988.
8. Guisnet, M., Garcia, J. J., Chevalier, F., and Maurel, R., Bull. Soc.
Chim. Fr. 1657 (1976).
9. Ebitani, K., Konishi, J., and Hattori, H., J. Catal. 130, 257 (1991).
26. Santiesteban, J. G., Calabro, D. C., Borghard, W. S., Chang, C. D.,
Vartuli, J. C., Tsao, Y. P., Natal-Santiago, M. A., and Bastian, R. D.,
J. Catal. 183, 314 (1999).