Effects of rhodium dispersion on catalytic behavior of Rh/active-carbon catalysts for H/D exchange reaction between and CH4 and D2

Article abstract of DOI:10.1039/b100687h

The H/D exchange reaction between CH₄ and D₂ was carried out over Rh/active-carbon catalysts, which were prepared from RhCl₃ and Rh(NO₃)₃. In the case of the catalysts prepared from RhCl₃, Rh species were homogeneously dispersed on the support from external surface to the inside of pores. Metallic particles of Rh were found to be the predominant species on the catalysts prepared from Rh(NO₃)₃ in the low Rh-loading region of < 2 wt.%, whereas the number of highly dispersed Rh species remarkably increased with an increase in the amount of Rh-loading in the region of > 2 wt.%. The reaction rate per unit gram of catalyst and the product distribution in methane reflected well the Rh-dispersion on the catalysts. The catalysts which contained the highly dispersed Rh species as predominant species were found to be more active for the H/D exchange reaction than the catalysts with relatively large metal particles of Rh. On the former, the ratio of CH₃D/CD₄ was observed to be much higher than that on the latter.

Full text of DOI:10.1039/b100687h

View Article Online / Journal Homepage / Table of Contents for this issue
E†ects of rhodium dispersion on catalytic behavior of
Rh/active-carbon catalysts for H/D exchange reaction between
CH and D
4
2
Hirotoshi Sakagami, Shingo Ogata, Nobuo Takahashi* and Takeshi Matsuda
Department of Materials Science, Kitami Institute of T echnology, 165 Koencho, Kitami,
Hokkaido 090-8507, Japan
Received 18th January 2001, Accepted 8th March 2001
First published as an Advance Article on the web 5th April 2001
The H/D exchange reaction between CH and D was carried out over Rh/active-carbon catalysts, which were
4
2
prepared from RhCl and Rh(NO ) . In the case of the catalysts prepared from RhCl , Rh species were
3
3 3
3
homogeneously dispersed on the support from external surface to the inside of pores. Metallic particles of Rh
were found to be the predominant species on the catalysts prepared from Rh(NO ) in the low Rh-loading
3 3
region of \2 wt.%, whereas the number of highly dispersed Rh species remarkably increased with an increase
in the amount of Rh-loading in the region of [2 wt.%. The reaction rate per unit gram of catalyst and the
product distribution in methane reÑected well the Rh-dispersion on the catalysts. The catalysts which
contained the highly dispersed Rh species as predominant species were found to be more active for the H/D
exchange reaction than the catalysts with relatively large metal particles of Rh. On the former, the ratio of
CH D/CD was observed to be much higher than that on the latter.
3
4
X-ray photoelectron spectroscopy (XPS). In addition, the H/D
exchange reaction between CH and D was studied on the
two kinds of catalyst, Rh(C)/AC and Rh(N)/AC.
Introduction
4
2
For the evaluation of noble metal dispersion on the support,
the chemisorption of carbon monoxide (CO) has frequently
been used, where adsorption of one CO molecule on each
surface metal atom was assumed. Although such 1 : 1 stoichi-
ometry has been pointed out to be questionable in the case of
a highly dispersed phase of noble metal,1h3 the amount of CO
adsorbed would strongly reÑect the dispersion. We have
already reported that Rh supported on active-carbon pre-
pared from rhodium trichloride (denoted by Rh(C)/AC)
adsorbed a very large amount of CO compared with that from
rhodium trinitrate (denoted by Rh(N)/AC).4h6 The transmis-
sion electron microscope (TEM) measurements showed that
the dispersion of Rh species on Rh(C)/AC was very high com-
pared with Rh(N)/AC.7 On the other hand, metallic Rh
species were found to be the main species on Rh(N)/AC. Thus,
the Rh compound used in the preparation of the Rh/AC cata-
lyst strongly a†ected the dispersion of Rh. However, the e†ects
of the amount of Rh-loading on the dispersion with respect to
the two kinds of Rh compounds, RhCl and Rh(NO ) , have
Experimental
The active-carbon support (obtained from Wako Pure Chemi-
cal Ind., Ltd, speciÐc surface area \ 930 m2 g~1, average pore
diameter (d ) \ 4 nm) was immersed in 0.1 mol dm~3 of
av
hydrochloric acid solution and then thoroughly washed with
de-ionized water. Rh/AC catalysts with Rh loading from 1.0
wt.% (97 ] 10~6 mol g(cat)~1) to 4.0 wt.% (389 ] 10~6 mol
g(cat)~1) were prepared by a conventional impregnation
method, where rhodium trichloride trihydrate (RhCl É 3H O)
3
2
or rhodium trinitrate (Rh(NO ) ) was deposited on the
3 3
support from the appropriate aqueous solution. The solid was
dried at 383 K for 12 h in an oven.
The samples for the CO chemisorption measurements were
prepared as follows. The solid pretreated with H at 573 K for
2
3 h in a conventional Ñow system was transferred into a glass
3
3 3
not been clariÐed.
Kemball and co-workers have studied extensively the H/D
exchange reaction between methane (CH ) and D over
various kinds of single-component metal catalysts and
bimetallic alloy catalysts.8 In the literature, CD was found to
be the main product at the early stage of the reaction on Rh-
cell, and it was re-treated with H at 573 K for 1 h before the
2
chemisorption measurement. The amounts of CO adsorbed
on the catalysts were measured at 296 K by a conventional
static method using a glass-made apparatus connected to a
vacuum line. The Ðrst CO adsorption run was followed by the
second run after the evacuation of the sample for 10 min at
room temperature. The sum of physical and chemical adsorp-
tion was estimated by extrapolating the Ñat part of the Ðrst
adsorption isotherm to the zero of the equilibrium pressure
for adsorption. The amount of physically adsorbed CO was
estimated from the second isotherm in the same way for the
estimation of the total adsorption. The amount of chemically
4
2
4
foil. On the other hand, CH D was the main product on
3
Rh/SiO catalysts.9,10 In the case of highly dispersed sup-
2
ported noble metal catalysts, e.g., Pt/Al O , CH D was found
2
3
3
to be the predominant product at the early stage of the reac-
tion.3 Multiple and stepwise exchange processes were pro-
posed to be the reaction routes for CD and CH D formation,
respectively.8 Thus, the catalytic behavior in the exchange
reaction between CH and D may reÑect the di†erence in the
Rh dispersion on the Rh/AC catalysts.
In the present work we have studied the dispersion of Rh
on active-carbon support by means of CO chemisorption and
4
3
adsorbed CO (denoted by q ) was estimated by subtracting
CO
the physical adsorption from the total adsorption. The value
4
2
for q on the active carbon support itself (\1 ] 10~6 mol
CO
g(cat)~1) was found to be much less than those on the Rh/AC
catalysts.
1930
Phys. Chem. Chem. Phys., 2001 3, 1930È1934
This journal is ( The Owner Societies 2001
DOI: 10.1039/b100687h

View Article Online
The H/D exchange reaction between CH and D was
carried out using a glass-made apparatus with a re-circulation
(XRD) and TEM measurement.7 In the case of Rh(N)/AC, q
4
2
CO
on the samples with 1 and 2 wt.% Rh was found to be much
system (350 cm3) connected to a vacuum line. The catalyst
less than those on the corresponding samples of Rh(C)/AC.
On the Rh(N)/AC samples, the formation of metallic particles
(0.100 g, pretreated with H at 573 K for 3 h in the Ñow
2
system) placed in the reactor was treated with H at 573 K for
of Rh was observed. However, q increased signiÐcantly in a
2
CO
1 h, and cooled to the temperature at which the reaction was
region of high Rh-loading, as shown by open triangles in Fig.
performed. Introduction of 0.50 ] 10~3 mol each of CH , D
2, indicating that the amount of highly dispersed Rh species
increased in this region on Rh(N)/AC samples.
The e†ects of Rh-loading on the XPS intensity (peak area
for Rh) are shown in Fig. 3. Although all samples were
4
2
and Ar (as diluent) into the system was followed by the reac-
tion. A small portion of the reaction gas mixture was period-
ically taken out from the reaction system to analyze the
D-distribution in methane by mass spectrometry, using a
ULVAC-STANDOM-200R spectrometer.
pretreated with H , they were exposed to air during the trans-
2
fer of them from the sample tube to the sample holder. There-
XPS measurements were carried out using a Rigaku XPS-
7000 spectrometer employing Al Ka X-radiation; adhesive
tape on the sample holder was used to support a Rh/AC
sample in a Ðne powder form. Binding energy values are refer-
enced to the peak at 285.0 eV for C 1s. To evaluate the Rh
concentration by XPS, the peak intensities for C 1s were nor-
malized to be 100 000 counts per second (cps).
fore, we focussed solely on the concentration of Rh by XPS,
but not on the oxidation number of Rh on the samples. In the
case of Rh(C)/AC samples, the intensity was increased with an
increase in the amount of Rh-loaded. In the case of Rh(N)/AC,
however, the increase in the peak intensity was very small in
the region from 1 to 4 wt.% Rh. The concentration of Rh by
XPS on the Rh(C)/AC(1 wt.%) was observed to be much
lower than that on the Rh(N)/AC(1 wt.%), indicating that the
dispersion of Rh species on the former was very high through-
out the whole active-carbon support, i.e., from the external
surface to the inside of the pores. On the other hand, Rh
species on the Rh(N)/AC(1 wt.%) were mainly present at the
external surface and/or in the pores near the entrance at low
Rh contents. However, the amount of CO chemisorption
remarkably increased with an increase in Rh-loading as
shown in Fig. 2, suggesting an increase in the amount of a
highly dispersed phase of Rh species in the pores at high Rh-
loading on Rh(N)/AC.
Results and discussion
Results on the CO adsorption measurements are shown in
Figs. 1 and 2 for Rh(C)/AC and Rh(N)/AC, respectively. In the
case of Rh(C)/AC, the amount of CO chemisorbed per unit
gram of catalyst (q ) almost linearly increased with an
CO
increase in Rh-loading, as shown by open circles in Fig. 1. The
value for (q )/(Rh-loaded) obtained was around 1.3, indicat-
CO
ing that the Rh dispersion on Rh(C)/AC was very high. In fact,
no formation of metallic particles of Rh was observed on
The H/D exchange reaction between CH and D was per-
Rh(C)/AC even after the H pretreatment by X-ray di†raction
4
2
2
formed on the two kinds of Rh/AC catalyst. Typical time
courses of the exchange reaction over Rh(C)/AC(2 wt.%) and
Rh(N)/AC(2 wt.%) are shown in Figs. 4 and 5, respectively.
The main product at the early stage of the reaction was found
to be CH D, indicating that the stepwise exchange process
3
was predominant on these catalysts. On Rh(N)/AC, however,
the amount of CD was more than the amounts of CH D
4
2 2
and CHD , which is similar to the results on Rh/SiO report-
3
2
ed by Kemball and co-workers,9,10 indicating the presence of
the Rh sites for the multiple exchange process. It is note-
worthy that the formation of CD was much less than CH D
on Rh(C)/AC. Namely, the amount of relatively large metal
particles, which contribute to the multiple exchange process to
4
2 2
form CD , was extremely low on Rh(C)/AC. These results
4
correspond well with the results of CO chemisorption and
XPS. In the present work, the rates for the exchange reaction
between CH and D per unit gram of catalyst (denoted by r)
4
2
are estimated from the data for the products formed during
Fig. 1 E†ects of Rh-loading on the CO adsorption on Rh(C)/AC
the reaction time of 1 h using the equation &iCH D .
4~i
i
(L, reduced; …, treated with CO after reduction with H ).
2
Fig. 2 E†ects of Rh-loading on the CO adsorption on Rh(N)/AC
Fig. 3 E†ects of Rh-loading on the XPS intensity of Rh(C)/AC (L)
and Rh(N)/AC (|).
(|, reduced; >, treated with CO after reduction with H ).
2
Phys. Chem. Chem. Phys., 2001, 3, 1930È1934
1931

View Article Online
The e†ects of reaction temperature on r over the Rh/AC(2
wt.%) catalysts are shown in Fig. 6. The rate at 433 K on
Rh(C)/AC (r \ 9.8 ] 10~6 mol min~1 g(cat)~1) was observed
to be very high compared with that on Rh(N)/AC(2 wt.%)
(r \ 1.8 ] 10~6 mol min~1 g(cat~1)). The value of the appar-
ent activation energy (denoted by E) calculated on Rh(N)/AC
is 90 kJ mol~1. This value is close to that on Rh/SiO ,
2
already reported to be 103 kJ mol~1.10 On the other hand,
the value for E on Rh(C)/AC (51 kJ mol~1) was observed to
be much lower than that on Rh(N)/AC. Partial pressure
dependencies of r on CH and D were observed to be ca. 0.6
4
2
and [0.3, respectively, on Rh(C)/AC as shown in Table 1. On
Rh(N)/AC, the D partial pressure dependency at 453 K was
obtained as [0.4. This value is slightly negative compared
with that on Rh(C)/AC at 413 K, indicating the stronger
2
adsorption of D on Rh(N)/AC. These results indicate that the
2
nature of the active sites on Rh(C)/AC is di†erent from those
on Rh(N)/AC. Although we have not determined the number
of highly dispersed Rh species as well as the number of the
surface Rh atoms on the metal particles, turnover frequency
(TOF) values at 433 K may roughly be estimated as follows.
When the value for %-exposed of Rh was assumed to be
100%, the TOF was calculated to be 5.1 ] 10~2 min~1 for the
highly dispersed Rh species. In the case of Rh(N)/AC(2 wt.%),
the value for %-exposed of Rh was estimated to be 20% on
the assumption that each surface Rh atom could adsorb one
molecule of CO. By using the value for %-exposed, the TOF
at the Rh site on metallic particles was estimated to be
4.6 ] 10~2 min~1. Thus, the di†erence in TOF values
between the two types of active site is not large, indicating
that the di†erence in the activity (r) at the reaction tem-
peratures around 433 K mainly corresponded to the di†erence
in the number of active sites.
Fig. 4 Exchange reaction between CH and D over Rh(C)/AC(2.0
4
2
wt.%) at 413 K (L, CH ; …, CH D; >, CH D ; @, CHD ; =, CD ).
4
3
2
2
3
4
As previously described, the amount of CO adsorbed on
Rh(C)/AC was observed to be very large. However, the cata-
lyst once treated with CO at temperatures [433 K did not
show a high ability for the CO adsorption.5,7 On Rh/Al O
2
3
and Rh/SiO catalysts, agglomeration of dispersed Rh species
2
into metal particles took place when the sample was exposed
to CO.11h15 We also showed by means of TEM that the
transformation of dispersed Rh species into metallic particles
of Rh was induced by the treatment of the samples with CO at
473 K.7 As can be seen from Fig. 1, the amount of CO chemi-
sorbed on the Rh(C)/AC with the CO treatment was observed
to be much lower than that on the corresponding sample
without the CO treatment. The e†ects of the CO treatment on
r are shown in Fig. 7. It is noteworthy that the treatment of
Rh(C)/AC with CO at 473 K remarkably reduced r. In addi-
tion, the apparent activation energy on Rh(C)/AC with the
CO treatment (68 kJ mol~1) was higher than that on the cata-
lyst without the CO treatment (51 kJ mol~1). As shown in
Fig. 5 Exchange reaction between CH and D over Rh(N)/AC(2.0
4
2
wt.%) at 453 K (L, CH ; …, CH D; >, CH D ; @, CHD ; =, CD ).
4
3
2
2
3
4
Fig. 8, the amount of CD was observed to be more than the
4
amounts of CH D and CHD on the former catalyst. These
2
2
3
results clearly showed that some portions of the dispersed Rh
species were converted into the relatively large metallic par-
ticles of Rh through the CO treatment.
The e†ects of Rh loading on r are shown in Fig. 9. The rate
on Rh(C)/AC was almost proportional to the amount of Rh-
loading. The rate on Rh(N)/AC with 2.0 wt.% Rh was much
low compared with that on Rh(C)/AC with the same Rh-
loading. The rate on Rh(N)/AC, however, remarkably
Fig. 6 E†ects of reaction temperature on the rate of exchange reac-
tion between CH and D (L, Rh(C)/AC; |, Rh(N)/AC).
4
2
Table 1 Comparison of kinetic behavior of Rh/AC catalysts for the exchange reactions between CH and D
4
2
Partial pressure dependency
Catalyst
Reaction temp./K
CH
D
E/kJ mol~1
4
2
Rh(C)/AC(2.0 wt.%)
Rh(N)/AC(2.0 wt.%)
413
453
0.6
0.6
[0.3
[0.4
51
90
1932
Phys. Chem. Chem. Phys., 2001, 3, 1930È1934

View Article Online
Fig. 7 E†ects of the CO treatment of Rh(C)/AC on the rate for the
exchange reaction between CH and D (L, Rh(C)/AC reduced; …,
Fig. 10 Comparison of catalytic behavior of Rh(N)/AC(2.0 wt.%) (|)
with Rh(N)/AC(4.0 wt.%) (>) for the exchange reaction between CH
4
2
4
Rh(C)/AC treated with CO at 473 K).
and D .
2
increased with an increase in the amount of Rh-loading as
shown in Fig. 9. As shown in Fig. 10, the apparent activation
energy on Rh(N)/AC with 4.0 wt.% Rh was calculated to be
70 kJ mol~1. This value is much lower than the value of E on
Rh(N) with 2.0 wt.% Rh (90 kJ mol~1), suggesting the pres-
ence of highly dispersed Rh species. The CO chemisorption
and XPS results on Rh(N)/AC indicate the presence of highly
dispersed Rh species in the pore of active-carbon support at
the high Rh-loading (see Fig. 2). In addition, the decrease in
the amount of CO chemisorbed was induced by the treatment
of Rh(N)/AC(4 wt.%) with CO as shown in Fig. 2, which was
due to the conversion of the dispersed Rh species into metal
particles. The product distribution of methane on Rh(N)/AC(4
wt.%) is compared with that on Rh(N)/AC(2 wt.%) in Fig. 11.
The content of CD for the former was found to be less than
4
that for the latter, indicating the higher ratio of highly dis-
persed Rh species against relatively large particles of Rh.
Thus, the remarkable increase in the activity (r) for H/D
exchange reaction between CH and D on Rh(N)/AC corre-
4
2
sponded well to the formation of the highly dispersed Rh
species in the pores.
In conclusion, the dispersion proÐle of Rh species on Rh(C)/
AC with respect to the change in Rh-loading was found to be
very di†erent from that on Rh(N)/AC by means of CO
adsorption and XPS techniques. On the former, Rh species
are mainly present in a highly dispersed form throughout the
Fig. 8 Product distribution in methane formed during the exchange
reaction on Rh(C)/AC (reaction temperature \ 413 K, CH conver-
4
sion 10.1%) and on Rh(C)/AC treated with CO at 473 K (433 K,
7.1%).
Fig. 11 Product distribution in methane formed during the exchange
Fig. 9 E†ects of Rh-loading on the rate for exchange reaction
between CH and D at 433 K (L, on Rh(C)/AC; |, on Rh(N)/AC).
reaction on Rh(N)/AC(2.0 wt.%) (reaction temperature \ 443 K, CH
4
conversion \ 10.1%) and on Rh(N)/AC(4.0 wt.%) (443 K, 23.1%).
4
2
Phys. Chem. Chem. Phys., 2001, 3, 1930È1934
1933

View Article Online
4
5
6
7
N. Takahashi, T. Takeyama, T. Fujimoto, A. Fukuoka and M.
Ichikawa, Chem. L ett., 1992, 1441.
N. Takahashi, Y. Sato, T. Uchiumi and K. Ogawa, Bull. Chem.
Soc. Jpn., 1993, 66, 1273.
N. Takahashi, Y. Takabatake, H. Sakagami, Y. Imizu, N.
Okazaki and A. Tada, J. Catal., 1996, 159, 491.
H. Sakagami, N. Ohta, S. Endo, T. Harada, N. Takahashi and T.
Matsuda, J. Catal., 1997, 171, 449.
pores of active-carbon support. On the latter, however, the
predominant Rh species were found to be relatively large
metal particles of Rh in the region of low Rh-loading, whereas
the amount of Rh species in a highly dispersed form increased
with an increase in Rh-loading. The catalytic behavior in the
exchange reaction between CH and D reÑected well the dif-
4
2
ference in Rh-dispersion on the catalyst. The catalyst contain-
ing the higher concentration of the dispersed phase showed
the higher apparent activation energy and the higher value for
the ratio of CH D/CD formed.
8
9
C. Kemball, Catal. Rev., 1971, 5, 33.
T. C. Wong, L. F. Brown, G. L. Haller and C. Kemball, J. Chem.
Soc., Faraday T rans. 1, 1981, 77, 519.
10 A. C. Faro, Jr. and C. Kemball, J. Chem. Soc., Faraday T rans.,
1995, 91, 741.
3
4
11 F. Solymosi and M. Pasztor, J. Phys. Chem., 1985, 89, 4789.
12 M. I. Zaki, G. Kunzman, B. C. Gates and H. Knozinger, J. Phys.
Chem., 1987, 91, 1486.
References
1
2
3
Z. Zhang, H. Chen and W. M. H. Sachtler, J. Chem. Soc.,
Faraday T rans., 1991, 87, 1413.
13 F. Solymosi and H. Knozinger, J. Chem. Soc., Faraday T rans.,
1990, 86, 389.
Z. Zhang, H. Chen, L.-L. Sheu and W. M. H. Sachtler, J. Catal.,
1991, 127, 213.
A. Khodakov, N. Barbouth, Y. Berthier, J. Oudar and P. Schulz,
J. Chem. Soc., Faraday T rans., 1995, 91, 569.
14 D. A. Buchanan, M. E. Hernandez, F. Solymosi and J. M. White,
J. Catal., 1990, 125, 456.
15 F. Solymosi and T. Bansagi, J. Phys. Chem., 1993, 97, 10133.
1934
Phys. Chem. Chem. Phys., 2001, 3, 1930È1934