The mechanism of methane reforming with carbon dioxide: Comparison of supported Pt and Ni (Co) catalysts

Article abstract of DOI:10.1023/A:1024494918755

CH₄ reforming with CO₂ is one of the promising processes for natural gas conversion. Since the chemical properties of Pt radically differs from those of Ni/Co, the interaction of the catalyst 5.16 wt % Pt/α-Al₂O₃ with CH₄, O₂, CO₂, and CH₄ + CO₂ pulses was investigated. CH₄ activation occurred via a common pathway via dissociative chemisorption on the metal surface with the formation of H₂ and carbon on all the catalysts. CO₂ activation on Pt/Al₂O₃ differed from its activation on Ni(Co)/Al₂O₃. Pt/Al₂O₃ was graphite-like in contrast to carbide carbon on Ni(Co)/Al₂O₃. This graphite carbon was more stable and less reactive. This prevented it from being an active intermediate of CO₂ reforming of CH₄.

Full text of DOI:10.1023/A:1024494918755

Kinetics and Catalysis, Vol. 44, No. 3, 2003, pp. 353–359. Translated from Kinetika i Kataliz, Vol. 44, No. 3, 2003, pp. 384–390.
Original Russian Text Copyright © 2003 by Bychkov, Tyulenin, Korchak.
VI RUSSIAN CONFERENCE
ON CATALYTIC REACTION MECHANISMS
(Moscow, October 1–5, 2002)
The Mechanism of Methane Reforming with Carbon Dioxide:
Comparison of Supported Pt and Ni (Co) Catalysts
V. Yu. Bychkov, Yu. P. Tyulenin, and V. N. Korchak
Semenov Institute of Chemical Physics, Russian Academy of Sciences, Moscow, 117977 Russia
Received October 9, 2002
Abstract—The interaction of the catalyst 5.16 wt % Pt/α-Al2O3 with ëç4, ëé2, é2, and ëç4 + ëé2 pulses is
studied using a setup involving the differential scanning calorimeter DSC–111 and a system for chromato-
graphic analysis. Comparison of the results obtained with analogous data on Ni/Al2O3 and Co/Al2O3 suggests
that methane activation occurs via a common pathway via dissociative chemisorption on the metal surface with
the formation of ç2 and carbon on all the catalysts studied. Carbon dioxide activation on Pt/Al2O3 differs from
its activation on Ni(ëÓ)/Al2O3. It follows from the enthalpy of formation that carbon on Pt/Al2O3 is graphite-
like in contrast to carbide carbon on Ni(Co)/Al2O3. This graphite carbon is more stable and less reactive.
enthalpies of formation ∆HCf of these forms of carbon.
INTRODUCTION
Methane reforming with carbon dioxide
It was found that these values are noticeably higher
than the values of ∆HCf for graphite and are close to
ëç4 + ëé2 = 2ëé + 2ç2
(I)
that for carbide carbon. Since the chemical properties
of platinum radically differs from the properties of
nickel/cobalt, it was interesting to carry out an analo-
gous study of the mechanism of CO2 reforming of
methane on Pt/Al2O3 and compare results with data for
Ni (Co) catalysts.
is one of the promising processes for natural gas conver-
sion. The mechanism of this reaction was studied by
many authors (see, for instance, the review papers [1, 2]),
but some important details of the process remain debat-
able. Among the catalysts for reforming methane, sup-
ported platinum-group metals or nickel/cobalt are the
most active. Earlier we studied the mechanism of CO2
reforming of methane on Ni/Al2O3 [3] and Co/Al2O3
catalysts [4], using a method combining the measure-
ment of sample reactivity toward the pulses of separate
reactants (ëç4, ëé2, ç2, and é2) and their mixtures
and continuous measurements of the calorimetric effect
of these reactions. The results obtained made it possible
to compare the rates of various routes of methane and
ëé2 activation discussed in the literature [1, 2]. It was
shown that, in contrast to the widespread opinion in the
literature, ëé2 activation in the course of methane
reforming does not occur via the oxidation of metallic
nickel/cobalt and further interaction of surface oxygen
with surface carbon,
EXPERIMENTAL
A sample 5.16 wt % Pt/α-Al2O3 was prepared by the
impregnation of alumina with a solution of H2PtCl6.
Initial γ-Al2O3 (Ssp = 220 m2/g) was calcined at 1150°ë
for 20 h. According to XRD data, the resulting product
contained only α-Al2O3 without admixtures γ-Al2O3
and θ-Al2O3 and had a specific surface area of 5 m2/g.
The necessary amount of H2PtCl6 was dissolved in
water, and the resulting solution was used to impreg-
nate Al2O3. Then, the sample was dried at 120°C for 1 h
and calcined in air for 1 h at 350°C.
The interaction of the sample with gaseous reactants
was studied using a setup involving a differential scan-
ning calorimeter DSC-111 and two gas chromato-
graphs. The solid sample (100–150 mg) was loaded
into a flow-type quartz cell, which was placed in a mea-
suring channel of the calorimetric unit of DSC-111.
Preliminary treatment of the samples included heating
in a flow of air at 700°C for 60 min. To study the
å + ëé2
å–é + ëé,
2å + ëé,
(II)
å–é + å–ë
(III)
but rather occurs via the direct interaction of ëé2 with
carbon:
ëé2 + å–ë
2CO + å.
(IV)
Moreover, it was shown that the state of carbon formed reduced state of a catalyst, the sample was treated in a
in methane decomposition on nickel/cobalt changes flow of hydrogen at 700°C for 30 min. The interaction
with time from a monolayer coverage to clusters and of the samples with reactants was carried out in flow
greater particles. Using calorimetric measurements of and pulse regimes. In the case of the pulse regime, the
the heat of reaction (IV), we managed to determine the sample was purged with specially purified helium pass-
0023-1584/03/4403-0353$25.00 © 2003 MAIK “Nauka /Interperiodica”

354
BYCHKOV et al.
Table 1. Data on oxygen consumption in alternating reduction (by an H2 flow) and oxidation (air pulses) of the fresh sample
Pt/Al2O3
Temperature, °C
Q , kJ/mol O2
Run
O2 amount, µmol
O*/PtO, mol %
O2
reduction
oxidation
1
2
3
4
5
6
500
500
500
500
700
500
500
500
100
500
700
500
22.6
19.0
12.0
14.8
5.0
–
17.0
14.4
9.0
291
264
293
377
395
11.2
3.8
5.0
3.8
* The ratio of oxygen g-atoms consumed to the amount of PtO formed in the complete oxidation of supported platinum (264.5 µmol/g), %.
ing at a rate of 30 ml/min. The pulses of reaction gases decrease in the amount of oxygen consumed (run 5). A
(0.6 ml) were supplied into a flow of helium using a six- decrease in the temperature to 500°C (run 6) did not
way valve. The flow that passed through a sample was restore the previous value. At the same time, the spe-
analyzed using two chromatographs equipped with col- cific heat of oxidation drastically increased. However,
umns packed with Porapak N and zeolite 5 Å and ther- taking into account a small amount of absorbed oxy-
mal conductivity detectors. The system of analysis gen, the unusually high (~380 kJ/mol é2) specific heat
allowed us to determine the concentrations of CO2, H2, may be a result of measurement error.
O2, N2, CH4, CO, C2H6, and C2H4.
To characterize the concentration of gaseous reac-
Interaction of Pt/Al2O3 with Methane
tants and reaction products, we used the ratio of each of
In the interaction of Pt/Al2O3 with methane pulses at
700°C, the formation of hydrogen via the following
reaction is observed:
the components to the initial pulse volume (as a per-
centage). This value was convenient for further calcula-
tions, but the sum of all the gaseous components after
the reaction was not equal to 100%.
ëç4 + å
ë–å + 2ç2.
(V)
After the reaction with methane, Pt/Al2O3 contains
coke, which is rapidly removed in the form of ëé2
when oxygen pulses are admitted. Platinum surface
blocking with coke leads to a rapid decrease in the
amount of hydrogen evolved with the number of pulses
(Fig. 1). After the first methane pulse, CO is also
formed via the reaction of surface coke with the resid-
ual oxygen of the catalyst:
RESULTS
Reduction of Pt/Al2O3 by Hydrogen
and Reoxidation by Oxygen
It was shown earlier [3, 4] that a comparison of
redox properties of a catalyst at different temperatures
makes it possible to characterize the state (the oxidation
state) and the dispersity of the supported metallic com-
ponent of the catalyst (the approach used by us is
described in the subsequent discussion of the results
below). Reduction was carried out in a flow of ç2 at
500 and 700°ë. Then, the reduced sample was oxidized
by a series of pulses of air, and oxygen consumption
and the reaction heat were measured. The results are
shown in Table 1.
ë–å + /é/
å + ëé.
(VI)
Preliminary treatment of the sample affects the for-
mation of ç2 and CO. Figure 1 shows data on the for-
mation of ç2 and CO in the reaction of Pt/Al2O3 with a
series of methane pulses after preliminary reduction of
the sample by hydrogen (1), reduction and reaction
with seven pulses of ëé2 (2), reduction and reaction
It follows from this table that the amount of oxygen with one air pulse (3). It is seen that the results of runs 1
consumed was maximal (22.6 µmol é2/g) in the course and 2 are almost identical: the yield of hydrogen drasti-
of the first reduction–oxidation run at 500°C. In the cally decreases from the first methane pulse to the sec-
course of further runs at 500°C (Table 1, runs 2, 4), this ond. Despite long treatment in a flow of ç2, the sample
value continuously decreased to 14.8 µmol é2/g. A contained a certain amount of bound oxygen, which
decrease in the oxidation temperature (run 3) also led to reacts with methane to form CO. Additional treatment
some decrease in oxygen consumption. The heat of oxi- with ëé2 pulses leads to practically no increase in the
dation practically remained constant (runs 2, 4). An concentration of oxygen in the sample. Conversely,
increase in the temperature to 700°C led to a drastic after the reaction with an air pulse, the amount of bound
KINETICS AND CATALYSIS Vol. 44 No. 3 2003

THE MECHANISM OF METHANE REFORMING WITH CARBON DIOXIDE
355
Product concentration, vol %
oxygen in the sample increases, which leads to an
increase in the amount of CO (Fig. 1, run 6). Simulta-
neously, the yield of ç2 in the first ëç4 pulse decreases,
but it increases (compared to runs 1 and 2) when the
second methane pulse is supplied (curve 3).
25
1
20
15
Figure 2 shows the results of the study of the effect
of the time interval between ëç4 pulses on the amount
of H2 formed (we took a fresh portion of the catalyst for
this run; therefore, the dependence of H2 yield on the
pulse number differs from the results presented in
Fig. 1). It is seen that the dependences of H2 on the
pulse number when the time intervals between CH4
pulses are 20 s and 10 min are very close.
3
10
6
5
4
2
5
0
1
2
3
Pulse number
Interaction with CO2
Fig. 1. Formation of (1–3) H and (4–6) CO in the interac-
2
tion of Pt/Al O with ëç pulses at 700°C: (1, 4) sample
2
3
4
When ëé2 pulses are supplied to the Pt/Al2O3 sample
treated in a standard manner at 700°C, ëé2 consumption
or CO formation was not seen (Fig. 3, pulses 1, 2). Pre-
liminary treatment with methane leads to a drastic
increase in ëé2 consumption and CO formation. Fig-
ure 3 shows CO formation in a series of ëé2 pulses
(Fig. 3, pulses 6–8, 12–14) after the preliminary reac-
tion of Pt/Al2O3 with three ëç4 pulses at 700°C (Fig. 3,
pulses 3–5, 9–11). Carbon monoxide is formed in an
amount that is twice as large amount as the ëé2 con-
sumed, in agreement with the reaction stoichiometry
preliminarily treated with ç ; (2, 5) reduced sample reacted
2
with seven ëé pulses; (3, 6) reduced sample reacted with
2
an air pulse.
Amount of H2, vol %
25
20
15
10
ë + ëé2
2ëé.
(VII)
At a higher concentration of carbon, the concentra-
tion of CO increases (Fig. 4) but remains lower than
4.5–5.0%. When CO evolution is completed (Fig. 4,
curve 1), further air pulses do not lead to ëé2 forma-
tion, which indicates that Pt/Al2O3 is cleaned of surface
carbon by a series of ëé2 pulses.
5
2
1
0
1
2
3
4
5
6
7
8
Pulse number
Fig. 2. Formation of ç in the reaction of Pt/Al O with the
2
2 3
Interaction with the ëç4 +ëé2 Mixture
series of ëç with intervals between pulses of (1) 20 s and
4
(2) 10 min.
Figure 5 shows the results of interaction of Pt/Al2O3
with pulses of the mixture 49%ëç4 + 49%ëé2 + 2%N2
at 700°C. The maximal ç2 and CO formation is
observed in the first pulse and then noticeably
decreases. The admission of two air pulses (Fig. 5,
pulse 11) restores the initial catalyst activity, whereas a
pulse of 100% ëç4 (pulse 14) leads to a drastic
decrease in the amount of ç2 and CO (pulse 15). In fur-
ther pulses, the catalyst activity increases but does not
reach the initial value.
Product concentration, vol %
8
7
6
2
2
5
4
3
2
1
Determination of the Enthalpy of Formation
of Surface Carbon
In the preceding paper [4], to determine the enthalpy
of surface carbon formation ∆HCf on Co/Al2O3, we
1
1
1
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14
measured the heat of reaction (VII). In this work, we
made similar attempts, but they were unsuccessful
because of the low intensity of the thermal signal in
these runs (this is due to the lower reaction rate on
Pulse number
Fig. 3. Formation of (1) CO and (2) ç in the interaction of
2
Pt/Al O with the pulses of (1) ëé and (2) ëç at 700°C.
2
3
2
4
KINETICS AND CATALYSIS Vol. 44 No. 3 2003

356
BYCHKOV et al.
Product concentration, vol %
CO concentration, vol %
5
80
70
60
50
40
30
20
10
2
1
2
4
3
2
1
3
4
2
1
O2
O2
CH4
1
2
3
4
5
6
7
0
0
4
6
8
10 12 14 16 18 20 22 24
Pulse number
Pulse number
Fig. 4. Formation of CO in the interaction of ëé pulses
2
Fig. 5. Amounts of (1) CO, (2) ç , (3) ëç , and (4) ëé in
with Pt/Al O at 700°ë after catalyst treatment by (1) two
2
4
2
2
3
the products of reaction of Pt/Al O with pulses of the mix-
methane pulses at 700°ë and (2) a methane flow at 700°ë
2
3
ture 49%ëç + 49%ëé + 2%N at 700°C.
for 10 min.
4
2
2
DISCUSSION
Pt/Al2O3 at a relatively low specific heat of the reac-
tion). Therefore, to determine ∆HCf , we measured the
The State of Supported Platinum
heat of carbon oxidation QÓx:
In previous papers [3, 4], it was shown that, at tem-
peratures of CO2 reforming of methane (~700°C), sup-
ported nickel/cobalt can transform into oxides (NiO
and CoO) or reduces to metallic particles by treatment
with air or hydrogen, respectively. In an inert atmo-
sphere, the resulting surface compounds are stable and
allow reliable measurements of the amount of oxygen
consumed and the oxidation heat to be made. The ratio
of the amount of é2 consumed by oxidation of nickel
and cobalt catalysts (preliminarily reduced by hydro-
gen) at 100°C to the amount of é2 consumed at 700°C
characterizes the fraction of surface metal atoms in sup-
ported particles. Specifically, for the Co/Al2O3 sample,
whose properties are compared in this work with the
properties of Pt/Al2O3, the fraction of surface cobalt
atoms measured in this way was 25%. Analogous
experiments carried out using platinum catalysts show
ë + é2
ëé2 + QÓx.
(VIII)
Because for the Pt/Al2O3 system the intensity (and, cor-
respondingly, the heat) of é2 interaction with supported
platinum at 700°C is rather low, it can be neglected and
the thermal effect can be assigned to reaction (VIII) of
coked Pt/Al2O3. The value of ∆HCf was calculated by
the formula ∆HCf = QÓx + ∆HCf O , where ∆HCf O is the
2
2
enthalpy of ëé2 formation at 700°C (taken from hand-
book [5]). Table 2 shows the experimental and calcu-
lated data for ∆HCf . The average value over the two
measurements is –4 kJ/mol and close to zero taking into
account the scatter of values. This is typical of graphite another pattern. Even the heating of the initial oxidized
sample in a flow of helium to 500–700°C leads to the
reduction of platinum particles. Correspondingly, the
amount of oxygen consumed in the interaction of the
reduced Pt/Al2O3 sample with air pulses at 700°C (see
Table 1) is significantly smaller than the amount
required for the complete oxidation of supported plati-
num into PtO. Oxidation at a low temperature (100°C)
gives almost the same amount of consumed oxygen
(Table 1). According to the conventional methods for
measuring the surface area of supported platinum [6],
the amount of chemisorbed ç2, é2, or CO at 20–100°C
corresponds to the number of surface metal atoms. It is
likely that in our experiments on Pt/Al2O3 oxidation at
both 100 and 700°C, é2 chemisorbs without platinum
oxide formation. Because we failed to oxidize all of the
supported platinum in the experiment, in order to esti-
mate the fraction of the surface atoms, the amount of
carbon.
Table 2. O2 consumption and heat evolution Q measured in the
oxidation of Pt/Al2O3 with surface carbon at 700°C and the cal-
culated values of the specific heat of oxidation Qox and ∆Hf
C
∆HfC , kJ/mol
O2, µmol
Q, mJ
Qox, kJ/mol O2
1.67
1.29
671.5
487.6
403
379
8
–16
Average
–4
KINETICS AND CATALYSIS Vol. 44 No. 3 2003

THE MECHANISM OF METHANE REFORMING WITH CARBON DIOXIDE
357
adsorbed oxygen was divided by the amount of oxygen
in PtO calculated from the amount of platinum sup-
ported in the course of preparation. The results pre-
sented in Table 1 show that on the alumina used in these
experiments the dispersity of supported platinum is
only 17% after heating the fresh catalyst at 500°ë. It
continuously decreases in the course of oxidation–
reduction treatments at 500°ë and drastically decreases
to ~4% after heating the sample to 700°ë. Such an
effect of a decrease in the dispersity of supported plati-
num in the course of sample heating above 700°ë is
known from the literature [7, 8].
Earlier we showed [3, 4] that in the process of ëé2
reforming of methane on Ni/Al2O3 and Co/Al2O3, reac-
tion (V) is the main route of methane activation, while
reaction (XIV) is the secondary route. Carbon dioxide
activation on these catalysts occurs via the second route
(IV). Results obtained in this work make it possible to
draw the same conclusion for Pt/Al2O3 catalysts as
well.
The reactivities of Pt/Al2O3 and Ni(Co) catalysts
toward methane are different [3, 4]. The latter are char-
acterized by a rather low reactivity (complete oxidation
of methane) in the oxidized state and high reactivity
(dissociative chemisorption of methane) in the reduced
state. Pt/Al2O3 actively reacts with methane after both
reductive and oxidative treatments. The amount of ç2
formed in the first pulse of methane ëç4 is comparable
with the amount of ç2 in the reaction of Pt/Al2O3 with
the pulses of the ëç4 + ëé2 mixture. That is, as in the
case of Ni(Co) catalysts, reaction (V) is the main route
of methane activation. Treatment with air at 700°ë
(even for a long time) does not lead to a decrease in the
conversion of ëç4, and a small amount of adsorbed
oxygen held by Pt/Al2O3 at 700°ë is removed after the
first pulse of ëç4 in the form of CO and ç2é
(a decrease in the amount of ç2) (Fig. 1, pulse 1). In
this case, the residence time of CO in chromatography
increases by ~1 min compared to the standard time
required for CO analysis under the given conditions.
This delay means that the CO evolved noticeably later
than the time of contact of the ëç4 pulse with the sam-
ple (~1 s). It is likely that the relatively low rate of CO
formation is due to diffusional limitations of the reac-
tion of surface carbon and catalyst oxygen, as in the
case of Ni/Al2O3 [3]. This reaction retards platinum
surface coking, and this is clear from the higher activity
of Pt/Al2O3 in the second pulse of ëç4.
Data on the heat of oxidation (Table 1) characterize
the energetics of this surface oxygen on Pt/Al2O3.
According to the data on the enthalpies of platinum
oxide formation [9], the heats of their formation in plat-
inum oxidation should be
Pt + O2
2Pt + O2
PtO2, 134 kJ/mol é2;
2PtO, 142.4 kJ/mol é2;
Pt3O4, 134 kJ/mol é2;
(IX)
(X)
3Pt + 2O2
(XI)
Pt + 0.675O2
3Pt + 2.085O2
PtO1.35
,
84.4 kJ/mol é2; (XII)
81.9 kJ/mol é2. (XIII)
Pt3O4.17
,
The values 260–290 kJ/mol é2 measured by us are
at least two times higher than those expected for possi-
ble platinum oxides. This fact suggests that the state of
chemisorbed oxygen in the oxidized state of supported
platinum differs from its state in platinum oxides. Sen
and Vannice [10] compiled a set of data on the experi-
mental heats of oxygen adsorption on Pt/SiO2, which
belong to the interval 210–290 kJ/mol é2. Results
obtained in this work agree well with these data.
The Mechanism of CO2 Reforming
In the reactions in the pulses of the ëç4 + ëé2 mix-
ture, the catalytic activity of Pt/Al2O3 and
Ni(Co)/Al2O3 also has different responses to the addi-
of Methane on Pt/Al2O3
The mechanisms of ëé2 reforming of methane dis-
cussed in the literature differ in the degree of detail pro- tion of oxygen. Air pulses increase the catalytic activity
vided and in the choice of routes for methane and ëé2
activation. Usually, methane activation is considered as dis-
sociative adsorption of methane on metallic particles (V) or
reaction of ëç4 oxygen in various states on the catalyst
surface:
of Pt/Al2O3 toward ëé2 reforming of methane (Fig. 5),
but they lead to a decrease in the conversion of ëé2 on
Ni/Al2O3 [3] in an analogous experiment. This differ-
ence is explained by the fact that in the series of succes-
sive pulses of the ëç4 + ëé2 mixture, the catalytic
activity of Pt/Al2O3 continuously decreased (Fig. 5,
pulses 1–10) because of catalyst poisoning by coke.
The introduction of oxygen leads to the fast oxidation
of coke and the restoration of the initial catalytic activ-
ity. In the case of the Ni/Al2O3 system [3], ëç4 + ëé2
pulses do not lead to catalyst coking and a decrease in
its activity under analogous conditions. In contrast, a
pulse of air leads to the considerable oxidation of the
nickel surface. This bound oxygen efficiently oxidizes
surface carbon in the subsequent pulses of the
ëç4 + ëé2 mixture and thus retards the activation of
ëé2 via reaction (IV).
ëç4 + /é/
ëé + 2ç2.
(XIV)
Three variants are possible for ëé2 activation:
(1) Two-step oxidation of metal and the further reac-
tion of bound oxygen with surface carbon (reactions (II)
and (III));
(2) Immediate interaction of ëé2 with surface car-
bon (IV); and
(3) The reaction of ëé2 with adsorbed hydrogen:
ëé2 + 2/ç/
ëé + ç2é.
(XV)
KINETICS AND CATALYSIS Vol. 44 No. 3 2003

358
BYCHKOV et al.
At 700°C the reactivity of Pt/Al2O3 toward ëé2 is nickel according to which the process occurs via the
intermediate formation of nickel carbide. In our exper-
iments with a small amount of coke, there are only pre-
cursors of these bulk carbon particles, but the results
agree well with the published data.
Different reactivities of carbon on Pt/Al2O3 and
Ni(Co)/Al2O3 reveal themselves in different catalytic
stabilities of these samples. In the pulse regime at
700°ë, the catalytic activity of Ni(Co)/Al2O3 in CO2
reforming of methane was stable and high [3, 4]. The
activity of Pt/Al2O3 noticeably decreases with the num-
much lower than the reactivity toward reduced Ni(Co)
catalysts for which a 2–3% conversion of ëé2 into CO
was observed [3, 4]. In the case of Pt/Al2O3, no CO for-
mation was detected. This is probably due to the insuf-
ficiently high value of the binding energy of oxygen
with the platinum surface, which can be judged from
the value of the heat of oxidation, kJ/mol O2: Qox =
260–290 (Pt/Al2O3), 470–490 (Ni/Al2O3) [3], and 450–
470 (Co/Al2O3) [4]. This means that, as in the case of
the Ni(Co) system, the activation of ëé2 via reactions (II)
and (III) does not play an important role in ëé2 reform- ber of pulse (Fig. 5). Additional coking of the Pt/Al2O3
catalyst due to a 100% ëç4 pulse decreases the yield of
CO and ç2 (Fig. 5, pulse 15). In further pulses of the
ëç4 + ëé2 mixture, the catalytic activity increases but
does not restore the initial value. In the case of
ëÓ/Al2O3 in an analogous experiment [4], additional
catalyst coking only lead to a slight reversible decrease
ing of methane.
The rate of ëé2 reaction with surface carbon (IV)
on Pt/Al2O3 is much higher than the rate of reaction (II).
However, it is an order of magnitude lower than the rate
of the analogous reaction on the Ni(Co) samples, and
the maximal conversion of ëé2 (~2.5%) is much lower in the catalytic activity.
than the ëé2 conversion under conditions of catalysis
at the same temperature (70–80%). It is likely that this
is due to different properties of the surface carbon on
Earlier we concluded that the interaction of ëé2
with the carbide carbon (IV) is the main route of ëé2
activation on Ni(Co) catalysts [3, 4]. In the case of
Pt/Al2O3. Because a decrease in the rate of methane Pt/Al2O3, the rate of reaction between ëé2 and graphite
carbon cannot be responsible for the apparent conver-
sion of ëé2 under catalytic conditions. The carbon spe-
cies that has a shorter lifetime (less than a second) on
Pt/Al2O3 is probably more reactive toward ëé2 than
graphite carbon, and such short-lived carbon is the
main intermediate in ëé2 reforming of methane.
decomposition with an increase in the number of ëç4
pulses is due to site blocking with coke on the surface
of metallic particles, the effect of the time interval
between ëç4 methane pulses on the dependence of the
ç2 yield on the number of pulses would point to the
motion of surface carbon for the time between pulses.
It was shown for ëÓ/Al2O3 [4] that, in the interval 10–
20 min after the formation of surface carbon in the
decomposition of methane, carbon diffusion with the
formation of its more stable form and a cleaning of the
cobalt surface are observed. In this case the value of the
enthalpy of formation (30–60 kJ/mol [4]) for this car-
bon is close to carbon in cobalt carbide. Conversely, in
the case of Pt/Al2O3, the formation of ç2 in the series
of ëç4 pulses with intervals between pulses 20 s and
10 min is practically identical (Fig. 2). This means that
the form of carbon formed during first seconds after
methane decomposition is rather stable under the reac-
tion conditions. Calorimetric measurements suggest
Thus, in the course of ëé2 reforming of methane, car-
bon formed by methane decomposition (reaction (V)) par-
ticipates in two competing processes: (1) a reaction
with ëé2 (IV) to form CO and (2) an agglomeration
with the formation of stable carbon. In the case of
Ni(Co)/Al2O3 catalysts, this stable form of carbon is
carbide carbon. It is highly reactive toward ëé2 and is
an active intermediate of ëé2 reforming of methane.
Moreover, carbide carbon does not block the surface of
metallic particles. In the case of Pt/Al2O3, carbon is sta-
bilized right on the platinum surface in a graphite-like
state and blocks the catalytic sites. The reactivity of
graphite carbon toward ëé2 is an order of magnitude
lower than the reactivity of carbide carbon. This pre-
vents it from being an active intermediate of CO2
reforming of methane.
that the value of ∆HCf for this carbon is close to zero,
and this value characterizes it as graphite carbon, in
contrast to carbide carbon on ëÓ/Al2O3.
ACKNOWLEDGMENTS
Different states of carbon on platinum and nickel
catalysts have been observed many times using micros-
copy, although in all cases the sample contained a large
amount of carbon after long treatment under reaction
conditions. These results suggest that, when the amount
of coke in nickel and cobalt systems is large, the forma-
tion of carbon filaments and nanotubes is observed [11,
12]. On platinum, large agglomerates with an undefined
indeterminate shape and containing graphite-like impu-
rities are formed [13, 14]. Zaikovskii et al. [11] pro-
posed a mechanism for carbon nanotube formation on
This work was supported by the Russian Foundation
for Basic Research (grant no. 01-03-32554).
REFERENCES
1. Bradford, M.C.J. and Vannice, M.A., Catal. Rev.–Sci.
Eng., 1999, vol. 41, no. 1, p. 1.
2. Krylov, O.V., Ross. Khim. Zh., 2000, vol. 44, no. 1, p. 19.
3. Bychkov, V.Yu., Krylov, O.V., and Korchak, V.N., Kinet.
Katal., 2002, vol. 43, no. 1, p. 94.
KINETICS AND CATALYSIS Vol. 44 No. 3 2003

THE MECHANISM OF METHANE REFORMING WITH CARBON DIOXIDE
359
4. Bychkov, V.Yu., Tyulenin, Yu.P., Krylov, O.V., and Kor-
9. Fiziko-khimicheskie svoistva okislov. Spravochnik
(Physicochemical Properties of Oxides: Handbook),
Samsonov, G.V., Ed., Moscow: Metallurgiya, 1978.
chak, V.N., Kinet. Katal., 2002, vol. 43, no. 5, p. 775.
5. Zhorov, Yu.M., Termodinamika khimicheskikh pro-
tsessov (Thermodynamics of Chemical Processes), Mos-
cow: Khimiya, 1985.
10. Sen, B. and Vannice, M.A., J. Catal., 1991, vol. 129,
no. 1, p. 31.
11. Zaikovskii, V.I., Chesnokov, V.V., and Buyanov, R.A.,
Kinet. Katal., 2001, vol. 42, no. 6, p. 890.
6. Benson, J.E. and Boudart, M., J. Catal., 1965, vol. 4,
no. 6, p. 704.
12. Yang, R.T. and Chen, J.P., J. Catal., 1989, vol. 115,
no. 1, p. 52.
7. Sushumna, I. and Ruckenstein, E., J. Catal., 1988,
13. Espinat, D., Freund, E., Dexpert, H., and Martino, G.,
vol. 109, no. 2, p. 433.
J. Catal., 1990, vol. 126, no. 2, p. 496.
8. Beck, D.D. and Carr, C.J., J. Catal., 1988, vol. 110, 14. Chang, T.S., Rodriguez, N.M., and Baker, R.T.K.,
no. 2, p. 289.
J. Catal., 1990, vol. 123, no. 2, p. 486.
KINETICS AND CATALYSIS Vol. 44 No. 3 2003