Reactions of n-hexane on Pt catalysts: Reaction mechanism as revealed by hydrogen pressure and compensation effect

Article abstract of DOI:10.1006/jcat.2001.3438

The reactions of n-hexane + hydrogen was studied using two Pt black catalysts prepared from H₂PtCl₆ using two methods, i.e., reduction with formaldehyde and with hydrazine. The activities and selectivities were compared with those of EUROPT-1 (6.3% Pt/SiO₂). EUROPT-1 showed higher selectivities for isomers and methylcyclopentane and lower values for hydrogenolysis and benzene. Dehydrogenation to hexenes was negligible at low temperatures and became significant at low hydrogen pressure and high pressure. Different compensation lines and isokinetic parameters were obtained in the range of positive and negative hydrogen order. The isokinetic parameters showed that the rate-determining step in the positive hydrogen order might have been a Pt-C bond breaking while a C-C bond rupture could be the rate-limiting step in excess hydrogen. Aside from gas-phase products residual carbonaceous deposits were formed during hexane transformation. C₆-saturated product formation suffered mostly from carbon accumulation, and accordingly, their formation reduced significantly on carbonized catalysts.

Full text of DOI:10.1006/jcat.2001.3438

Journal of Catalysis 205, 86–96 (2002)
doi:10.1006/jcat.2001.3438, available online at http://www.idealibrary.com on
Reactions of n-Hexane on Pt Catalysts: Reaction Mechanism as
Revealed by Hydrogen Pressure and Compensation Effect
Attila Wootsch1 and Zolta´n Paa´l
Institute of Isotope and Surface Chemistry, Chemical Research Center of the Hungarian Academy of Sciences, H-1525 Budapest, P.O. Box 77, Hungary
Received May 16, 2001; revised September 17, 2001; accepted October 4, 2001
multiatomic active ensemble. This hypothesis resulted in
maximum rates as a function of hydrogen pressure (9).
The reactions of n-hexane + hydrogen have been studied us-
ing two Pt black catalysts of different preparation (Pt–N, reduced
by hydrazine; Pt–HCHO, reduced by formaldehyde). The activities
and selectivities were compared with those of EUROPT-1 (6.3%
Pt/SiO2). Maximum rates were observed as a function of hydrogen
pressure for the overall reaction as well as for hydrogenolysis, iso-
merization, C5-cyclization, and aromatization. Dehydrogenation to
hexenes decreased with increasing H2 pressure. Both Pt black cata-
lysts showed lower turnover frequencies than did EUROPT-1 and
the maximum rates appeared at higher hydrogen pressure. The val-
ues of Arrhenius parameters calculated in the negative and in the
positive hydrogen order range were different for the overall reaction
as well as for the individual processes. The apparent activation en-
ergies showed compensation effects. Different compensation lines
and isokinetic parameters were obtained in the range of positive and
negative hydrogen order. The isokinetic parameters indicated that
the rate-determining step in the positive hydrogen order may have
Bond and Slaa (10–12) used the Langmuir–Hinshelwood
bimolecular kinetic theory as an explanation of maximum
rates of n-butane hydrogenolysis over a Ru/Al2O3 cata-
lyst. Accordingly, at lower H2 pressures the—dissociatively
chemisorbed—reactive intermediates compete for hydro-
gen and the reaction exhibits a positive hydrogen order.
Having reached a certain H2 pressure, the hydrogen cov-
erage reaches an optimum value for the reaction, corre-
sponding to zero hydrogen order. On a further increase of
p(H2), the reacting molecules have to compete with hydro-
gen for the empty surface sites. Hence, the reaction shows
a negative hydrogen order (10).
Skeletal reactions of n-hexane over Pt catalysts include
four product classes: hydrogenolysis, isomerization, C5-
cyclization, and aromatization. These reactions are accom-
panied by dehydrogenation to hexenes. The reaction sys-
tem (1, 5) is more complicated than the hydrogenolysis
processes of smaller hydrocarbons (10); hence no single
model could be proposed to describe n-hexane reactions
over EUROPT-1 (6.3% Pt/SiO2) (13). Turnover frequen-
cies (TOF) and Arrhenius parameters could be calculated
for the overall reaction as well as for most of the indi-
vidual processes (i.e., hydrogenolysis, isomerization, C5-
cyclization, and aromatization). The Arrhenius plots cal-
culated for the overall rate at constant hydrogen pressures
“bent down” at higher temperatures (13, 14), while those
calculated from the maximum rates (corresponding to dif-
ferent hydrogen pressures) gave straight Arrhenius lines
(6, 13). We suggested that the latter parameters might ap-
proximate the “true” values, although their “apparent”
character could not be disproved (13). Large differences
were seen between the Arrhenius parameters in the range
of positive and negative hydrogen order (13).
–
been a Pt C bond breaking (“hydrogenative product desorption”),
–
while a C C bond rupture could be the rate-limiting step in excess
hydrogen. The compensation effects of different reactions under
different conditions contributes to our knowledge of the reaction
c
mechanism.
2002 Elsevier Science
Key Words: n-hexane reaction; turnover frequency; hydrogen
pressure effect; compensation effect; virtual isokinetic parameters;
Pt black; EUROPT-1.
INTRODUCTION
Hydrogen Effect on the Arrhenius Parameters of n-Hexane
Reaction over Pt Catalysts
We reported on the reactions of n-hexane over differ-
ently prepared, modified, and partly deactivated Pt black
catalysts (1–4). The conversion over various Pt catalysts
showed maxima as a function of hydrogen pressure in a
wide temperature range (1, 5, 6) (i.e., the hydrogen or-
der of the reaction changed from positive to negative with
increasing hydrogen pressure). Several explanations have
been proposed for this phenomenon. Frennet and cowork-
ers (7–9) postulated the “landing-site model,” assuming a
The Temkin equations (15) proposed that the apparent
activation energy is the sum of the real activation energy
and the adsorption enthalpy of the compound multiplied
by the reaction order:
1 To whom correspondence should be addressed. E-mail: wootsch@
alpha0.iki.kfki.hu.
Eapp = Et + 6ni1Hi.
[1]
86
0021-9517/02 $35.00
c
2002 Elsevier Science
All rights reserved.

KINETICS OF n-HEXANE REACTIONS ON Pt
87
Equation [1] shows that it is difficult to estimate true ac- where is the vibration mode of the reactant molecule (ex-
tivation energies from the measured values. We suggested pressed in cm 1) that distorts most strongly the structure
(13) using the maximum points (ni = 0), while Larsson (16) of the molecule toward the structure at the activated state
used zero limiting pressure or concentration (1Hi = 0) for (27–29).
calculating kinetic parameters closer to true ones.
The Arrhenius data for the overall consumption of n-
hexane over EUROPT-1 in the range of negative hydrogen
order showed a compensation effect (13) but the point of
intersection was out of the range of the real experimental
data. We called it a virtual isokinetic point (13).
Compensation Effect
The linear correlation between activation energy and the
preexponential factor in heterogeneous catalysis is called
the compensation effect. Compensation phenomena can be
observed either when the same reaction is studied on a se-
ries of different catalysts or when the same catalyst is used
for different reactions (17). Though compensation effect
is a widely applied and accepted phenomenon, there is no
general interpretation of this behavior (18). Galwey (19)
suggested that the main reason may be the energetically
heterogeneous surface of the catalyst. There is an expo-
nential dependence between the partial constants and the
activation energy on heterogeneous surfaces. The fraction
of the corresponding active site depends on the activation
energy and gives rise to a compensation effect (19). An-
other kind of explanation is that the heat and entropy of
adsorption are often related to each other in a linear man-
ner (19), because a greater binding energy of the molecule
to the surface considerably restricts its vibrational and ro-
tational freedom, as derived mathematically by Balandin
(20). Corma et al. (21) pointed out that the assumption of
the transition state model of the compensation effect means
a linear relationship between enthalpy and entropy of the
transition state.
The use of Langmuir–Hinshelwood kinetics can also lead
to a compensation effect (17, 19). Since the surface cover-
age of the catalyst by hydrogen and/or reactive hydrocar-
bon intermediates changes significantly as the temperature
increases, the activation energy measured using constant
hydrogen pressures can be regarded as “apparent” (Eapp),
while those based on constant surface coverage are labeled
“true” (Et) (17, 22, 23, 24). Compensation effects may oc-
cur because of the use of apparent rather than true kinetic
parameters (13, 17, 22–24).
The compensation effect can be used for characteriz-
ing heterogeneous catalyst surfaces (22) and can serve as
a basis for suggesting a reaction mechanism (19, 22), in
spite of the fact that its exact explanation is not proven
and that the phenomenon corresponds to apparent rather
than to true kinetic parameters (13, 22–24). The reactions
that define the same compensation line share common fea-
tures of reaction mechanisms, for example the same rate-
determining step, similar structure of surface intermediates,
or analogous types of active centers (19, 22). We reported
(13) the effect of hydrogen pressure on kinetic parame-
ters and their compensation phenomenon for EUROPT-
1, without drawing any conclusion about the reaction
mechanism.
The aim of the present paper is to obtain more infor-
mation about the effect of hydrogen pressure on kinetic
parameters of different Pt catalysts. We want to show
that the apparent kinetic parameters as well as their com-
pensation effects can be used for characterizing different
catalysts (22). We expect to collect new information on
the reaction mechanism of n-hexane transformation on Pt
catalysts.
METHODS
Catalyst Preparation
Pt black samples were prepared from H2PtCl6 using two
“classical” methods: reduction with formaldehyde (30) and
with hydrazine (31). We adopt an earlier notation (32) and
call them Pt–N and Pt–HCHO, respectively.
For preparation of Pt–N, 150 ml of 10% aqueous hy-
drazine hydrate was added dropwise to a solution of 10 g
of H2PtCl6 in 8 ml of water over 25 min. A dark-gray pre-
cipitate was formed. The liquid was thereafter boiled for
24 h to remove residual ammonia; the Pt was washed with
bidistilled water for several days (2).
For preparation of Pt–HCHO, 5 g of H2PtCl6 was dis-
solved in 4 ml of water and mixed with 10 ml of formalde-
hyde. Then, 16 ml of 50% aqueous KOH was added drop-
wise, with cooling to 290 K. The precipitated Pt was
washed with HNO3 to remove residual K and with bidis-
tilled water (3).
Arrhenius plots that obey compensation effects must in-
tersect each other at exactly the same point, defining an
“isokinetic temperature” (Tiso) and “isokinetic rate” (kiso
)
(19, 25). Rooney derived the isokinetic parameter from
the expanded Eyring equation (26). In his interpretation
(26), Tiso is correlated with kexp, a ratelike characteristic
frequency of the Arrhenius lines resulting in a compensa-
tion effect, since, under standard conditions, kexp = kTiso/h
(where k is Boltzmann’s constant and h is Planck’s con-
stant). Larsson (27, 28) and Karpinski and Larsson (29), in
turn, correlated Tiso with the vibration frequency between
the reactant molecule and the catalyst system in complete
resonance,
Both samples were pretreated in a H2 flow at 473 K,
resulting in a sintered but stable sample (33). The specific
[2] surface(BET, N2)aswellasthedispersion(H2–O2 titration,
Tiso 0.715 ,

´
WOOTSCH AND PAAL
88
TABLE 1
400 K) was measured after the hydrogen treatment. For
Pt–N, BET (N2) = 2.64 m2/g, and the dispersion was 0.89%
(2.73 1016 Pt atom on the surface of 1 mg of catalyst).
For Pt–HCHO, BET (N2) = 3.80 m2/g, and the dispersion
was 1.36% (4.20 1016 Pt atom on the surface of 1 mg of
catalyst).
XPS Surface Analysis of Pt Black Catalysts (atom %)
Catalyst
Pt–N
Pt
O
C
As received
84
78
66
5
2
1
11
20
33
Regenerateda
After n-hexane exposureb
Pt–HCHOc
As received
75
76
68
11
6
4
15
18
28
Catalytic Tests
Regenerateda
After n-hexane exposureb
Catalytic runs were carried out in a closed-loop reactor
(volume 70 ml) described earlier (34). It was filled with a
mixture of n-hexane (10 Torr) and hydrogen (60–480 Torr).
The temperature ranged between 513 and 633 K. The same
charge of catalyst was used (16.5 mg of Pt black reduced
by hydrazine—with 4.51 1017 surface Pt atoms—and
11.2 mg of Pt black reduced by HCHO—4.69 1017
surface Pt atoms). Sampling took place after a 5-min
reaction time in all cases. Product analysis was performed
gas chromatographically using a 50-m CP-Sil glass capillary
column (34). The TOFs (35) were calculated as the number
of hexane molecules reacted per one surface Pt atom, using
the length of the run as the “contact time” (13). Regen-
eration between the runs was carried out at the reaction
temperature with 30 Torr air for 2 min followed by evacu-
ation and a 3-min hydrogen treatment at 100 Torr. When
the reaction temperature was changed, two regenerations
took place, both at the previous and at the subsequent
reaction temperatures. Several catalytic runs were carried
out with n-hexane + H2 mixtures, and the catalyst was also
regenerated in between. The catalytic properties were well
reproducible (typically within 5%) after several regenera-
tions (36a).
a Regeneration was carried out, after evacuating the reaction mixture,
using 30 Torr of air for 2 min, then evacuating and subsequently using
100 Torr of H2 for 3 min.
b After exposure to p(nH) : p(H2) = 10 : 60 Torr at T = 603 K for t =
20 min.
c Ref. 36b.
RESULTS
Figure 1 shows the overall consumption of n-hexane on
Pt black catalysts as a function of hydrogen pressure. The
shape of these curves resembles earlier analogous results
on Pt blacks (2, 3) and on EUROPT-1 (13). The maxima
were shifted toward higher hydrogen pressures as the
temperature increased. The spline-type curves presented
in Fig. 1 are similar to those constructed by a more correct
computer simulation for the simpler reaction: butane
hydrogenolysis (17).
TOF values were calculated as the number of hexane
molecules transformed over one surface Pt (determined by
H2–O2 titration) per unit time (35). Due to the well-known
ensemble effect (37), however, the number of actual active
sites is very likely not equal to the number of surface Pt
atoms. Figure 2 summarizes the hydrogen pressure depen-
dence of the TOFs on both Pt blacks and EUROPT-1 (13)
Surface Analysis
The atomic surface composition of the Pt black catalysts at three selected temperatures (533, 573, and 603 K). The
(2, 3, 32, 36) was determined by X-ray photoelectron spec- reaction rate was highest on EUROPT-1 and the lowest on
troscopy in three states: “as received,” after exposure to the Pt–HCHO. The values of maximum rates represent “the
n-hexane + H2 mixture (20 min), and after regeneration catalytic activity,” while the hydrogen pressure of maxima
with 20 Torr of O2 (5 min) and 200 Torr of H2 (20 min). shows the optimum reaction conditions for each tempera-
These treatments were carried out in the preparation ture (Table 2). They were at somewhat higher values in the
chamber of the photoelectron spectrometer (36) and were case of Pt blacks than of Pt/SiO2 (EUROPT-1) (Table 2).
analogous to the reaction of the n-hexane + hydrogen
Thehydrogendependenceoftheyieldofindividualprod-
mixture and subsequent regeneration in a glass reactor. uct classes was also examined. The curves of benzene,
The results of surface analysis are presented in Table 1. methylcyclopentane (MCP), and skeletal isomers had max-
Pt–HCHO contained more carbon and oxygen than did ima over hydrogen pressure. The last two are closely related
Pt–N. The amount of carbon atoms increased on the surface reactions (38, 39); therefore, as an example, we show the hy-
after exposure to n-hexane + H2. Regeneration removed drogen pressure dependence of their sum as “C6 saturated
a fraction of the surface carbon. The surface composition, products” in Fig. 3 for Pt–N at all the examined temper-
such as catalytic properties, was well reproducible after atures. Analogous, but lower, values were seen with Pt–
several regenerations. The residual surface oxygen after re- HCHO. The optimum conditions were different for each
generation (on both Pt–HCHO and Pt–N) belonged mainly reaction (1) (Table 3). Skeletal isomerization needed the
to adsorbed water and chemisorbed Pt–OH ensembles, highest optimum hydrogen pressure, while the lowest value
with hardly any “Pt–O” and very little oxidized C entities appeared for aromatization (Table 3). Hydrogenolysis
(2, 3, 32).
showed positive and dehydrogenation showed negative

KINETICS OF n-HEXANE REACTIONS ON Pt
89
a
12
b
20
15
10
5
10
8
6
4
2
0
0
0
100
200
300
400
500
0
100
200
300
400
500
p(H2) (Torr)
p(H2) (Torr)
FIG. 1. Conversion (%) of n-hexane transformation over Pt black catalysts at different temperatures. (a) Pt–N, (b) Pt–HCHO. The ordinate scale of
the two graphs are different. The curves represent the closest fit (spline) curve. Encircled points are the measured ones that are nearest to maxima. The
dottedlineisasecond-orderpolynomforthemaximaoftheindividualcurves. ᭜, 513K;᭿, 533K;1, 543K; , 558K; , 573K;᭹, 583K;᭛, 603K;᭡, 663K.
hydrogen order under the experimental conditions; thus enes was negligible at low temperatures; it became im-
their optimum hydrogen pressure was not reached.
portant at low hydrogen pressure and high temperature.
The selectivities of individual product classes are shown Even the highest value (EUROPT-1 at 603 K, p(H2) =
in Fig. 4 at three temperatures (533, 573, and 603 K). 60 Torr) did not exceed 25%. Pt–N resulted in more hex-
EUROPT-1 exhibited higher selectivities for isomers and enes, somewhat more fragments of benzene, and less iso-
methylcyclopentane (MCP) and lower values for hydro- mers and MCP than did Pt–HCHO.
genolysis and benzene (34, 40). Dehydrogenation to hex-
Arrhenius plots are usually calculated at constant hy-
drogen pressures. Another possibility is to use maximum
rates representing zero hydrogen order. Figure 5 illustrates
these results, similar to those reported using EUROPT-
1 (13). Different Arrhenius lines could be constructed in
the range of negative and positive hydrogen order, re-
spectively. The points taken from the maxima defined an
enveloping curve (thick lines in Fig. 5). The calculated
160
Pt-N
Pt-HCHO
140
EUROPT-1
120
100
80
60
40
20
0
TABLE 2
Comparing the Maximum Rates and Their Hydrogen Pressures
of Pt Black Catalyst with Standard EUROPT-1
Pt–N
Pt–HCHO
EUROPT-1a
p(H2)
(Torr)
TOF
p(H2)
(Torr)
TOF
p(H2)
(Torr)
TOF
(h )
1
1
1
T (K)
(h
)
(h
)
533
543
558
573
583
603
118
148
192
315
397
11.6
115
147
201
320
375
6.4
11.5
21.3
23.7
41.5
—
97
133
214
257
294
432
22.5
30.1
45.5
62.5
74.7
135.0
17.1
22.9
39.1
60.0
—
0
100
200
300
400
500
p(H2) (Torr)
FIG. 2. Comparison of the hydrogen dependence of the overall rate
of n-hexane transformation over Pt blacks and EUROPT-1 (13) at 533
(᭹), 573 ( ), and 603 K (᭜).
>480
>480
a Ref. 13.

´
WOOTSCH AND PAAL
90
12
10
8
with the three constants (k1, KA, KB) having their own
temperature dependence. The Eapp calculated may elimi-
nate some but not all distorting factors and thus may be
closer to the true value.
We attempted to calculate Arrhenius plots for the in-
dividual processes, too. The Arrhenius plots of skeletal
isomers, MCP, and6C6-saturatedproductswere“inverted”
in the range of positive hydrogen order (Fig. 6) (14). Thus
the rate of their formation formally decreased with the
temperature so that even apparent Arrhenius parameters
could not be calculated. This behavior also appears in
Fig. 3, showing at low hydrogen pressure lower yields of
C6-saturated products when the temperature increased. In-
verse Arrhenius plots also appeared separately for skele-
tal isomerization and C5-cyclization. This phenomenon can
be explained by the “hydrogen effect” (1, 22, 42, 43) and
carbonaceous deposit formation (44): the hydrogen surface
coverage depends on both temperature and hydrogen pres-
sure. The yield and rate of formation of reaction interme-
diates is a function of the available hydrogen on the sur-
face. Its amount is proportional to the hydrogen pressure
at the same time; the “effective surface hydrogen pressure”
may be higher than the gas-phase value (45). Besides the
gas-phase products residual carbonaceous deposits were
also formed during hexane transformation (46). They are
the main cause of catalyst deactivation (47). Their appear-
ance is controlled by the available surface hydrogen (44)
(i.e., high temperature and low hydrogen pressure, range of
positive hydrogen order, favored their formation). Under
these severe conditions, high T and low p(H2), the initial
catalytic activity is directed for forming surface carbona-
ceous deposits rather than products (46). As a consequence
(Fig. 5), the Arrhenius curves for the overall reaction bend
down (42). C6-saturated product formation suffers mostly
from carbon accumulation, and accordingly their formation
reduced significantly on carbonized catalysts (40, 46). Thus
inversion of Arrhenius curves appears here (Fig. 6).
6
4
2
0
0
100
200
300
400
500
p(H2) (Torr)
FIG. 3. Yield of saturated C6 products over hydrogen pressure at var-
ious temperatures on Pt–N. ᭜, 513 K; ᭿, 533 K; ᭝, 543 K; , 558 K;
573 K; ᭹, 583 K; ᭛, 603 K; ᭡, 663 K.
,
Arrhenius parameters are presented in Table 4 and com-
pared with the results of EUROPT-1. These parameters are
apparent—they are mathematically calculated constants
of the Arrhenius lines—and do not have necessarily any
physicochemical meaning (13).
It is more difficult to calculate the true activation en-
ergies. As predicted from Eq. [1], the apparent energy is
independent of 1HH when the hydrogen order is zero,
but still depends on 1HnH. The value of the hexane order
(nnH) changed from zero to one, depending on tempera-
ture and hydrogen pressure (41). From the results reported
earlier, nnH 1 was between 543 and 633 K for EUROPT-1
at p(H2) = 120 Torr (41). Thus Emaximum Et + 1HnH. The
values can be regarded as close to true values if we apply
one single Temkin equation [Eq. 1] for the overall reac-
tion. However, this is a rather rough approximation. Differ-
ent reactions leading to each product class have their own
elementary steps and surface intermediates. Their formal
description involves individual Langmuir-type equations,
DISCUSSION
Compensation Behavior and Isokinetic Effect
TABLE 3
Figure 7 compares the results of Pt black catalysts with
the compensation lines reported for EUROPT-1 (13, 48).
The points of the overall reaction calculated in the nega-
tive hydrogen order range for both Pt black catalysts de-
termine a straight line parallel to those taken from the lit-
erature. The differences between the compensation lines
can be caused by various things. One of them might be the
different dispersion value. The dispersion was measured by
H2–O2 titration.
Arrhenius parameters calculated in the region of neg-
ative hydrogen order for MCP, for skeletal isomers, for
benzene, and for 6C6-saturated products were compared
with earlier results (13, 48). The same was done in the
H2 Pressures of the Maximum Rates for Individual Processes over
Pt–N (in Torr)
T (K)
Isomer
MCP
6C6 sat.a
Benzene
513
533
543
558
573
583
603
663
<60
100
142
150
342
396
>480
>480
<60
<60
120
132
285
363
>480
>480
<60
90
135
143
332
381
>480
>480
<60
78
117
130
197
262
362
>480
a 6C6 sat. = skeletal isomers + MCP.

KINETICS OF n-HEXANE REACTIONS ON Pt
91
70
60
50
40
30
20
10
0
70
60
50
40
30
20
10
0
70
60
50
40
30
20
10
0
a
0
200
400
0
0
0
200
400
400
400
0
0
0
200
400
400
400
p(H2) (Torr)
p(H2) (Torr)
p(H2) (Torr)
70
60
50
40
30
20
10
0
70
60
50
40
30
20
10
0
70
60
50
40
30
20
10
0
b
0
200
400
200
200
p(H2) (Torr)
p(H2) (Torr)
p(H2) (Torr)
70
60
50
40
30
20
10
0
70
60
50
40
30
20
10
0
c
70
60
50
40
30
20
10
0
200
200
0
200
400
p(H2)
p(H2)
p(H2)
FIG. 4. Selectivity of the individual product classes as a function of the hydrogen pressure (in Torr) over Pt–N (left), Pt–HCHO (middle), and
EUROPT-1 (right) at (a) 533, (b) 573, and (c) 603 K. +, Skeletal isomers; ᭛, MCP; ᭡, benzene; ᭹, fragments (<C6); ᮀ, hexenes.
region of positive hydrogen order for the overall process, products is on the same compensation line as the overall
hydrogenolysis, and benzene (Fig. 6). The slopes of the com- conversion in the negative hydrogen order for all three cat-
pensationlinesforthenegativeandpositivehydrogenorder alysts (Fig. 7). Thus, they are closely related reactions (34,
ranges were different for all examined catalysts. Further, 38, 39), and their formation (exhibiting highest selectivities)
some points for different reactions lie on the same com- maydeterminetheoverallrateofn-hexanetransformations
pensation line, while others do not. In particular, (a) skele- in the range of the negative hydrogen order (23).
tal isomers and 6C6-saturated products agree with those of
Arrhenius plots that obey the compensation effect ought
the n-hexane overall consumption in the negative hydrogen to intersect each other at exactly the same point, defining
order range, and (b) the data points for hydrogenolysis lie an isokinetic temperature and isokinetic rate (19, 25).
on a different curve, coinciding with that of aromatization The Arrhenius lines for EUROPT-1 intersect each other
in the positive hydrogen order range.
outside the physicochemical range of the actual range
Reactions that obey the compensation effect can be of measurements, resulting in so-called virtual isokinetic
called related reactions, since they have the same kind of a parameters (13). The same phenomenon appears in the case
surface intermediate and/or the same rate-determining step of Pt blacks. The calculated parameters are presented in
(19, 22). The formation of skeletal isomers and C6-saturated Table 5.

´
WOOTSCH AND PAAL
92
1000
100
10
a
1000
100
10
b
1
1
1.5
1.6
1.7
1.8
1.9
2
1.5
1.6
1.7
1.8
1.9
2
1000/T(K)
1000/T(K)
FIG. 5. Arrhenius plots calculated at constant hydrogen pressures over Pt black catalysts: (a) Pt–N, (b) Pt–HCHO. Encircled points correspond
to maximum of conversion in Fig. 1. 1, 480 Torr; +, 360 Torr; ᭜, 240 Torr; , 120 Torr; ᭿, 60 Torr.
–
The isokinetic parameters calculated from the intersec- stretch on Ir, Pt, and Ni and to that of a M C bond stretch
tion of compensation lines can also give some information on Co and Fe (49, 50). Thus a different reaction mecha-
–
about the likely reactions of adsorbed surface intermedi- nism was suggested for the different metals (50): C C bond
ates. Larsson found that the isokinetic temperatures are breaking was claimed to be rate determining on Pt, Ir, and
–
different over different metals in ethane hydrogenolysis re- Ni, and M C bond breaking was claimed on Fe and Co. This
actions (49, 50), being around 725 K for Ir, Pt, and Ni, and
330–320 K for Co and Fe. The vibration frequencies calcu-
lated using Eq. [2] were close to the frequency of a C C
100
–
TABLE 4
Comparison of Overall Kinetic Parametersa
10
Pt–N
Pt–HCHO
EUROPT-1b
Catalyst:
H2 order
p(H2)
Eapp ln(A) Eapp ln(A) Eapp ln(A)
Positive
60
120
240
360
11
11
21
30
4.8
5.5
8.1
6
15
20
33
3.2
5.5
7.3
8
13
17
—
4.9
6.4
7.8
—
1
10.4
10.1
Zero
Maximum
81
21
71
18
66
18
Negative
60
120
240
480
730
—
—
107
120
—
—
—
26.2
28.3
—
—
—
94
107
—
—
—
22.7
25.1
—
91
102
115
117
117
23.7
26.0
28.1
28.3
26.8
0.1
1.5
1.6
1.7
1.8
1.9
2
1000/T(K)
Literaturec
FIG. 6. Arrhenius plots calculated for saturated C6 at constant
hydrogen pressures over Pt–N. Encircled points correspond to maximum
of yields in Fig. 5. 1, 480 Torr; +, 360 Torr; ᭜, 240 Torr; , 120 Torr; ᭿,
60 Torr.
a
Eapp in kJ mol 1; A, in h 1; p(H2), in Torr.
b Ref. 13.
c Ref. 48.

KINETICS OF n-HEXANE REACTIONS ON Pt
93
50
45
40
35
30
25
20
15
10
5
Range of negative
hydrogen order
32
30
28
26
24
22
20
80
90
100
110
120
130
140
150
Range of positive
hydrogen order
0
0
50
100
150
200
250
Ea (kJ/mol)
FIG. 7. The compensation effect for n-hexane conversion as well as for individual reactions. Earlier results of several reactions (48)—lower dotted
line—and our earlier reaction (13)—upper dotted line—on EUROPT-1 are compared with those of the present work. Pt–N: ᭿, overall reaction;
,
6C6-saturated products; ᭹, skeletal isomers; ᭡, MCP; ᭜, benzene (hydrogenolysis). Pt–HCHO: ᮀ, overall reaction; , 6C6-saturated products; ᭺,
skeletal isomers; ᭝, MCP; ᭛, benzene (hydrogenolysis).
latter process represents the decomposition of rather stable hydrogen order, when hydrogen excess was present on the
–
surface species.
surface, a C C bond formation or cleavage appeared to be
The present system is more complicated than the ethane the rate-limiting step.
hydrogenolysis procedure but the similar isokinetic tem-
Hydrogen Effect and the Suggested Reaction Mechanism
peratures may indicate similar frequencies in the transition
states. According to Larsson’s suggestion, the frequencies
calculated from Tiso belong to the rate-limiting transforma-
tion of the adsorbed surface intermediates (49, 50). The
isokinetic parameters presented in Table 5 estimated a fre-
The suggested reaction mechanism is presented in
Scheme 1. The first step can be a—probably rate determin-
–
ing (52)—dissociation of the first C H bond of the saturated
reactant (3, 38, 39, 45). Alternatively “reactive adsorption,”
–
quency close to the value of a C C bond stretch in the nega-
–
tive hydrogen order and to that of a C Pt bond stretch in the
C6H14 + PtHPt → C6H13Pt + 2PtH,
[3]
positive hydrogen order, analogously to Larsson’s values on
different metals (49, 50). Thus when the hydrogen cover-
age was low (i.e., in the range of positive hydrogen order),
the overall rate-determining step might be the formation
can also be assumed (7). This can be the surface
–
or rupture of a C Pt bond (51). In the range of negative
TABLE 5
Isokinetic Parameters Calculated from the Compensation
Parameters in the Negative and Positive Hydrogen Order Range
C H
13
6
Negative hydrogen order
Positive hydrogen order
1
1
Catalyst
Pt–N
TOFiso (h
)
Tiso (K)
TOFiso (h
)
Tiso (K)
1760
365
850
690
680
650
8
6
5
440
465
320
Pt–HCHO
EUROPT-1a
SCHEME 1. The suggested reaction mechanism for the reaction of n-
hexaneoverPtcatalysts. Theboldlinessymbolizetheeffectofdeactivation
by carbonaceous deposits (3, 36, 40).
aRef. 13.

´
WOOTSCH AND PAAL
94
intermediate of the adsorbed C5-cyclic species, which can results calculated from the isokinetic parameters. On the
react to give MCP or skeletal isomers (Scheme 1) (3, 38, 39). other hand when much hydrogen is present (negative hy-
Their ratio is governed by the available surface hydrogen drogen order), dehydrogenation to unsaturated surface in-
(3, 45, 53, 54). Formation of isomers from the common in- termediates of aromatization can be hindered (1, 45, 59).
termediate needs the uptake of two hydrogen atoms; hence
The hydrogen pressure at maximum rates defines the
it should be more favored by ample hydrogen availability. optimum hydrogen coverage. This is higher over Pt blacks
Indeed, the isomer/MCP ratio was higher when the back- than over Pt/SiO2 (Fig. 2). Hydrogen spillover facilitates
spillover from the support could increase the amount of hy- hydrogen “traffic” to and from the silica support and this
drogen available on the metal (e.g., on graphite nanofiber- offers one possible explanation (63). It may not be as
supported Pt catalysts) (55).
pronounced as over alumina but is certainly impossible on
Arrheniusparametersofhydrogenolysislieonadifferent unsupported blacks.
compensation line. Thus there must be a separate reaction
route from the key intermediate (Scheme 1). It was sug-
gested (56) that a “one-point adsorption” of the alkane and
its subsequent attack by a neighboring hydrogen can be the
route of ethane hydrogenolysis over Pt catalysts. Larsson
(50) based his calculations on this assumption. This route
is analogous to the formation of the C6H13Pt key interme-
diate (Scheme 1 and Eq. [3]) first and its fragmentation
without further dehydrogenation. Internal rupture was re-
garded as closely related to isomerization and may involve
a not deeply dehydrogenated intermediate (37b, 57). How-
ever, several authors attribute a deeply dissociated (e.g.,
The optimum condition for the overall reaction is a
“weighted average” of the individual product classes. Their
selectivity values are controlled by both the hydrogen pres-
sureandthecatalyststructure. Electronmicroscopyshowed
different morphology for the two Pt black catalysts: Pt–N
containing “polyhedral” crystals, with more flat planes (2),
while Pt–HCHO, sintered at 473 K in H2 and appearing as
aggregates of rounded crystals (64), contained much high-
index “miniplanes” (65). The structure of EUROPT-1 was
approximated with cubo-octahedra containing 55 Pt atoms
(66); both this and the raftlike structures deduced from
EXAFS results (67) contain much smaller contiguous Pt
surfaces. These different catalyst structures favor different
reactions. Hydrogenolysis increased gradually with “step
and kink density” over single crystals (68) and hence was
attributed to crystal edges. The importance of the so-called
B5 sites (containing five atoms along a step: (110) or (311)
“ledge structures”) was pointed out for hydrogenolysis
on both single crystals and 10% Pt/Al2O3 catalysts (69).
This reaction was favored in the sequence Pt–N
Pt–HCHO EUROPT-1 (Fig. 4). The rather ordered crys-
tallites of EUROPT-1 contained fewer “ledge structures”
(present as imperfections) than did the big polyhedral parti-
cles of Pt–N. This, in turn, must have exposed more “ledges”
thandidtheroundedparticlesofPt–HCHO. Aromatization
,
, ) intermediate(s) to hydrogenolysis (58), the forma-
tion and decomposition of which must involve several ele-
mentary steps, more difficult to interpret.
Dehydrogenation to hexenes becomes pronounced (up
to 25%, Fig. 4) when a small amount of hydrogen is present
on the surface (3) and the number of contiguous metal sites
drops (37b). These unsaturated species can desorb as hex-
enes or form more unsaturated adsorbed molecules like
hexadienes or trienes (51, 59). Their cis trans isomeriza-
tion requires also hydrogen but can still take place with less
H(ads) present. Benzene forms via the cyclization of the cis
trienes (59, 60), while the trans isomers can polymerize to
carbonaceous deposits (1, 59).
Longer polymeric deposits—precursors of graphitic or was attributed to structures of threefold symmetry (37a)
Cx Hy entities (32, 36)—are formed more easily on the corresponding to the (111) planes of EUROPT-1 (41), as
larger contiguous planes of Pt blacks (D 1%) than on the confirmed by studies on single crystals of similar symmetry
well-dispersed EUROPT-1 (D = 60%). The different hy- (68). Benzene selectivity (Fig. 4) was highest on Pt–N (its
drogencontentofcarbonaceousdepositscouldalsobemea- larger “flat” planes being likely structurally favorable) and
sured by their deuterative removal (61): the carbonaceous lowest on disperse EUROPT-1 (40, 41). Isomerization and
deposits on EUROPT-1 contain more hydrogen than do MCPformationwasfoundtoberatherstructure-insensitive
those onPtblack. Thisresult agreeswellwiththe presentse- on single crystals (68). Their proposed flat-lying intermedi-
lectivity data (Fig. 4): more saturated C6 products—favored ate (1, 70) would require two adjacent metal atoms and
by a “Pt–C–H” state (62)—were formed over EUROPT-1. could equally take place on the (100) and (111) structures
More benzene and lower catalytic activity were, in turn, (41). Indeed, the selectivity values for the C6-saturated
found on Pt blacks, in agreement with its easier transfor- products (Fig. 4) were highest on EUROPT-1. Dehydro-
mation to a deactivated “Pt–C” state (62).
In the range of positive hydrogen order, benzene forma-
tion should need surface hydrogen. The rate-limiting step
genation to alkenes is a structure-insensitive reaction (34,
37b, 68), requiring single atom sites. Its selectivity depends
on the possibility of other reactions and plays an important
–
role on partly deactivated catalysts with many Pt C en-
may be “hydrogenative desorption” of C6H6 from C6H6
x
sembles (high temperature, low hydrogen pressure), where
other reactions are hindered (37b, 40, 41).
surface entities (51, 60) (i.e., the hydrogenolytic rupture
of the surface Pt–C bond). This is in agreement with the

KINETICS OF n-HEXANE REACTIONS ON Pt
95
We found only small differences in the optimum hydro- 17. Bond, G. C., Catal. Today 49, 41 (1999).
18. Bond, G. C., Catal. Rev. 42, 323 (2000).
19. Galwey, A. K., Adv. Catal. 16, 247 (1977).
20. Balandin, A. A., Dokl. Akad. Nauk. SSSR 93, 55 (1953).
gen pressure of individual product classes on the three
different catalysts (i.e., the maximum p(H2) of C6-
saturated product formation was a slightly lower over
EUROPT-1 than over blacks, while the optimum pressure
of benzene showed the opposite behavior). Assuming an
adsorbed C6H13Pt primary intermediate (Scheme 1) and
supposing a quasiequilibrium between a less dehydro-
genated intermediate of C5-cyclic reactions and the more
dehydrogenated intermediates producing benzene and
carbonaceous deposits, it follows that the abundance of
these surface entities is governed by the “effective surface
hydrogen pressure” (45). This might be about 10 times
higher than in the gas phase (45). The actual values of the
corresponding adsorption constants must be different on
the supported Pt than on Pt blacks, with the same p(H2)
resulting in a higher effective surface hydrogen pressure on
EUROPT-1.
21. Corma, A., Lopis, F., Monton, J. B., and Weller, S., J. Catal. 142, 97
(1993).
22. Bond, G. C., Appl. Catal. 191, 23 (2000).
23. Bond, G. C., Hooper, A. D., Slaa, J. C., and Taylor, A. O., J. Catal. 163,
319 (1996).
24. Bond, G. C., and Cunningham, R. H., J. Catal. 166, 172 (1997).
25. Bond, G. C., “Catalysis by Metals,” p. 139. Academic Press, London/
New York, 1962.
26. Rooney, J. J., J. Mol. Catal. 129, 131 (1998).
27. Larsson, R., J. Mol. Catal. 55, 70 (1989).
28. Larsson, R., React. Kinet. Catal. Lett. 68, 115 (1999).
29. Karpinski, Z., and Larsson, R., J. Catal. 168, 532 (1997).
30. Loew, O., Berichte 23, 289 (1890).
31. Paal, C., Berichte 49, 548 (1916).
32. Paa´l, Z., Schlo¨gl, R., and Ertl, G., J. Chem. Soc., Faraday Trans. 88,
1179 (1992).
33. Baird, T., Paa´l, Z., and Thomson, S. J., J. Chem. Soc., Faraday Trans.
69, 311 (1973).
The hydrogen pressure dependence of n-hexane trans-
formation shows the same picture but with different ap-
parent parameters on the catalysts studied. The conversion
and reaction rate values were lower on Pt–HCHO than on
Pt–N. Both the presence and position of surface impurities
(Table 1) and structural differences (2, 3, 64) can explain
the differences.
34. Paa´l, Z., Groeneweg, H., and Zimmer, H., Catal. Today 5, 199
(1989).
35. Boudart, M., in “Handbook of Heterogeneous Catalysis” (G. Ertl,
H. Kno¨zinger, and J. Weitkamp, Eds.), Vol. 3, p. 958. VCH, Weinheim,
1997.
36. (a) Paa´l, Z., Wild, U., Wootsch, A., Find, J., and Schlo¨gl, R., Phys.
Chem. Chem. Phys. 3, 2148 (2001); (b) Find, J., Paa´l, Z., Sauer, H.,
Schlo¨gl, R., Wild, U., and Wootsch, A., Stud. Surf. Sci. Catal. 130, 2291
(2000).
37. (a) Biloen, P., Helle, J. N., Verbeek, H., Dautzenberg, F. M., and
Sachtler, W. M. H., J. Catal. 63, 112 (1980); (b) Ponec, V., Adv. Catal.
32, 149 (1983).
ACKNOWLEDGMENTS
38. Maire, G., Plouidy, G., Prudhomme, J. C., and Gault, F. G., J. Catal. 4,
556 (1965).
39. Barron, Y., Maire, G., Muller, M., and Gault, F. G., J. Catal. 5, 428
(1966).
40. Paa´l, Z., Groeneweg, H., and Paa´l–Luka´cs, J., J. Chem. Soc., Faraday
Trans. 86, 3159 (1990).
The authors are grateful to Professor R. Larsson for valuable discus-
sion and advice. Inspiration given by Professor G. C. Bond is thankfully
acknowledged. The help of Professor R. Schlo¨gl in making the XPS mea-
surements possible and Ms. Ute Wild for her contribution to their real-
ization is acknowledged. The research received financial support from the
Hungarian Scientific Research Fund (Grant OTKA T 025599).
41. Bond, G. C., and Paa´l, Z., Appl. Catal. 86, 1 (1992).
42. Paa´l, Z., J. Catal. 91, 181 (1985).
REFERENCES
43. Paa´l, Z., and Somorjai, G., in “Handbook of Heterogeneous Catalysis”
(G. Ertl, H. Kno¨zinger, and J. Weitkamp, Eds.), Vol. 3, p. 1084. VCH,
Weinheim, 1997.
1. Paa´l, Z., Adv. Catal. 29, 273 (1980).
2. Paa´l, Z., Zhan, Z., Fu¨lo¨p, E., and Tesche, B., J. Catal. 156, 19 (1995).
3. Paa´l, Z., Xu, X. L., Paa´l-Luka´cs, J., Vogel, W., Muhler, M., and
Schlo¨gl, R., J. Catal. 152, 252 (1995).
4. Paa´l, Z., Matusek, K., and Muhler, M., Appl. Catal. 149, 113 (1995).
5. Paa´l, Z., Catal. Today 12, 297 (1992).
44. Sa´rka´ny, A., J. Chem. Soc., Faraday Trans. 85, 1523 (1989).
45. Paa´l, Z., Sze´kely, G., and Te´te´nyi, P., J. Catal. 58, 108 (1979).
46. Rodriguez, N. M., Anderson, P. E., Wootsch, A., Wild, U., Schlo¨gl, R.,
and Paa´l, Z., J. Catal. 197, 365 (2001).
47. Bond, G. C., Appl. Catal. 149, 3 (1997).
6. Paa´l, Z., Matusek, K., and Te´te´nyi, P., Acta Chim. Acad. Sci. Hung.
94, 119 (1977).
48. Bond, G. C., Maire, G., and Garin, F., Appl. Catal. 41, 313 (1988).
49. Larsson, R., Catal. Lett. 11, 137 (1991).
7. Parayre, P., Amir-Ebrahimi, V., Gault, F. G., and Frennet, A., J. Chem.
Soc., Faraday Trans. 76, 1704 (1980).
8. Frennet, A., Lienard, G., Crucq, A., and Degols, L., J. Catal. 53, 150
(1978).
9. Frennet, A., in “Hydrogen Effects in Catalysis” (Z. Paa´l and P. G.
Menon, Eds.), p. 399. Dekker, New York, 1988.
50. Larsson, R., Catal. Lett. 13, 71 (1992).
51. Zimmer, H., Rozanov, V. V., Sklyarov, A. V., and Paa´l, Z., Appl. Catal.
2, 51 (1982).
52. Te´te´nyi, P., Acta. Chim. Acad. Sci. Hung. 40, 157 (1964).
53. Falco, M. G., Canavese, S. A., Comelli, R. A., and F´ıgoli, N. S., Stud.
Surf. Sci. Catal. 130, 2393 (2000).
10. Bond, G. C., and Slaa, J., J. Mol. Catal. 98, 81 (1995).
11. Bond, G. C., and Slaa, J. C., J. Mol. Catal. 89, 221 (1994).
12. Bond, G. C., and Slaa, J. C., Catal. Lett. 23, 293 (1994).
13. Wootsch, A., and Paa´l, Z., J. Catal. 185, 192 (1999).
14. Davis, S. M., Zaera, M., and Somorjai, G., J. Catal. 85, 206 (1984).
15. Temkin, M., Acta Physicochim. URSS 3, 312 (1935).
16. Larsson, R., Catal. Lett. 36, 171 (1996).
54. Paa´l, Z., J. Catal. 156, 301 (1995).
55. Baker, R. T. K., Laubernds, K., Wootsch, A., and Paa´l, Z., J. Catal.
193, 165 (2000).
56. Gudkov, B. S., Guczi, L., and Te´te´nyi, P., J. Catal. 74, 207 (1982).
57. Gault, F. G., Adv. Catal. 30, 1 (1981).
58. (a) Anderson, J. R., Adv. Catal. 23, 1 (1973); (b) Sinfelt, J. H., Adv.
Catal. 23, 91 (1973).

´
96
WOOTSCH AND PAAL
59. Paa´l, Z., and Te´te´nyi, P., J. Catal. 30, 350 (1973).
66. Gnutzmann, V., and Vogel, W., J. Phys. Chem. 94, 4991 (1990).
60. Teplyakov, A. V., and Bent, B. E., J. Phys. Chem. B 101, 9052 67. Jackson, S. D., Keegan, M. B. T., McLellan, G. D., Meheux, P. A.,
(1997).
Moyes, R. B., Webb, G., Wells, P. B., Whyman, R., and Willis, J., Stud.
Surf. Sci. Catal. 63, 135 (1991).
68. (a) Somorjai, G. A., “Chemistry in Two Dimensions: Surfaces,”
p. 479. Cornell Univ. Press, London, 1981; (b) Blakely, D. W., and
Somorjai, G. A., J. Catal. 42, 181 (1976).
61. Wootsch, A., Descorme, C., Paa´l, Z., and Duprez, D., in “ Proceed-
ings, 12th International Congress on Catalysis, Granada, 2000” (A.
Corma, F. V. Mello, S. Mendioroz, and J. L. G. Fierro, Eds.). Elsevier,
Amsterdam, 2000.
62. Sa´rka´ny, A., Catal. Today 5, 173 (1989).
63. Duprez, D., Stud. Surf. Sci. Catal. 112, 13 (1997).
64. Paa´l, Z., Zimmer, H., Gu¨nter, R., Schlo¨gl, R., and Muhler, M., J. Catal.
119, 146 (1989).
69. (a) Garin, F., Aeiyach, S., Le´gare´, P., and Maire, G., J. Catal. 77, 323
(1982); (b) Dauscher, A., Garin, F., and Maire, G., J. Catal. 105, 233
(1987).
70. Zimmer, H., Dobrovolszky, M., Te´te´nyi, P., and Paa´l, Z., J. Phys. Chem.
90, 4758 (1986).
65. Ehrlich, G., Adv. Catal. 14, 364 (1963).