Decomposition and Oxidation of Dimethyl Ether on Rh Catalysts

Article abstract of DOI:10.1006/jcat.1997.1827

The adsorption, decomposition, and oxidation of dimethyl ether have been investigated on Rh catalyst using alumina, magnesia, titania, and silica supports. Transmission infrared spectroscopic measurements have revealed that adsorbed ether molecules are transformed into methoxy species at 200-300 K. This process proceeds mainly on the supports. The effect of Rh is exhibited at higher temperature, above 473 K, when it catalyses the decomposition of dimethyl ether to give CO, H₂, and a small amount of CH₄. The formation of strongly adsorbed species, such as CO, CH_x, and C, leads to the early deactivation of the catalysts. The complete oxidation of dimethyl ether to CO₂ and H₂O was studied in a flow system. The reaction started above 450 K and attained almost 100% conversion at 600-630 K for all supported Rh, except Rh/MgO. There are no significant differences in the activities of other Rh samples. Taking into account the dispersities of Rh, the most effective catalyst is the alumina-, followed by titania-, silica-, and magnesia-supported Rh. Whereas at lower temperatures different hydrocarbons in minor amounts were detected in the products, above 600 K only 0.1-0.2% methane was identified. Based on the results obtained on Rh(111) under UHV conditions, it is concluded that dimethyl ether interacts with adsorbed O atoms bonded to the Rh, and the oxidation proceeds via the transient formation of methoxy species on the Rh.

Full text of DOI:10.1006/jcat.1997.1827

JOURNAL OF CATALYSIS 171, 476–484 (1997)
ARTICLE NO. CA971827
Decomposition and Oxidation of Dimethyl Ether on Rh Catalysts
F. Solymosi, J. Csere´nyi, and L. Ova´ri
Institute of Solid State and Radiochemistry, A. Jo´zsef University and Reaction Kinetics Research Group of the Hungarian Academy of Sciences,1
P.O. Box 168, H-6701, Szeged, Hungary
Received March 17, 1997; revised July 1, 1997; accepted July 11, 1997
in use even after the application of DME as a fuel. In a
recent comparative work we found that alumina-supported
platinum metals catalyse the total oxidation of DME above
473 K, and 100% conversion can be attained at 673–723 K
(5). As regards the decomposition of DME, Rh/Al2O3 was
one of the most active catalysts, it deactivated, however,
very easily. The interaction of DME with clean and
K-promoted Rh(111) has been also examined by high res-
olution electron energy loss spectroscopy (HREELS) and
the primary dissociation products have been established
(6, 7).
The adsorption, decomposition, and oxidation of dimethyl ether
have been investigated on Rh catalyst using alumina, magnesia, ti-
tania, and silica supports. Transmission infrared spectroscopicmea-
surements have revealed that adsorbed ether molecules are trans-
formed into methoxy species at 200–300 K. This process proceeds
mainly on the supports. The effect of Rh is exhibited at higher
temperature, above 473 K, when it catalyses the decomposition of
dimethyl ether to give CO, H2, and a small amount of CH4. The for-
mation of strongly adsorbed species, such as CO, CHx, and C, leads
to the early deactivation of the catalysts. The complete oxidation of
dimethyl ether to CO2 and H2O was studied in a flow system. The
reaction started above 450 K and attained almost 100% conversion
at 600–630 K for all supported Rh, except Rh/MgO. There are no
significant differences in the activities of other Rh samples. Taking
into account the dispersities of Rh, the most effective catalyst is the
alumina-, followed by titania-, silica-, and magnesia-supported Rh.
Whereas at lower temperatures different hydrocarbons in minor
amounts were detected in the products, above 600 K only 0.1–0.2%
methane was identified. Based on the results obtained on Rh(111)
under UHV conditions, it is concluded that dimethyl ether inter-
In the present paper we performed detailed studies on
supported Rh catalysts. Attention is paid to the adsorption,
decomposition, and oxidation of DME. The effects of sup-
port materials are also explored for all these processes.
2. EXPERIMENTAL
The oxide supports were impregnated at room temper-
acts with adsorbed O atoms bonded to the Rh, and the oxidation ature in an aqueous solution of RhCl3 ꢁ 3H2O (Johnson–
proceeds via the transient formation of methoxy species on the Rh.
Matthey). The amount of RhCl3 ꢁ 3H2O used was calculated
c
ꢀ 1997 Academic Press
so, that if it will be transformed into metallic Rh (following
oxidation and reduction), it should correspond to 1 wt% of
the catalyst. The impregnated material was dried at 330 K
and stored over silica gel. The following supports were used:
Al2O3 (Degussa P 110 C1), TiO2 (Degussa P 25), SiO2 (Cab-
O-Sil) and MgO (DAB 6). The fragments of catalyst pellets
were oxidized at 673 K for 30 min and reduced at 773 K
in the catalytic reactor for 1 h. In the case of Rh/TiO2 the
effects of low and high temperature reduction were also ex-
amined. When the effects of oxidized Rh/Al2O2 was tested
the reduction step was omitted. XPS analysis showed that
in the oxidized sample, ꢂ70% of the Rh was in the state
of Rh3+. DME was the product of Aldrich. The gases used
were initially of commercial purity. The Ar (99.995) was
deoxygenated with an oxytrap. The other impurities were
adsorbed by a 5A molecular sieve at the temperature of
liquid nitrogen.
1. INTRODUCTION
The study of the adsorption and catalytic reaction of
dimethyl ether (DME) is motivated by several reasons.
DME is a key compound in the conversion of methanol
to gasoline (1, 2), which is one of the most successful in-
dustrial catalytic processes. Recently, it has been proposed
to use DME as an alternative fuel for diesel engines (3, 4).
It is claimed that the burning of DME produces much less
pollutants compared to diesel oil and satisfies worldwide
emission standards (1–3) .
In our laboratory we are mainly concerned with the inter-
action of DME with catalyst surface and with its complete
oxidation, assuming that exhaust catalysts may remain
Catalytic measurements were carried out in a fixed bed
continuous-flow reactor made of quartz (17 mm ID). The
amount of catalysts used was 0.3 g. Analyses of the exit
1 This laboratory is a part of the Center for Catalysis, Surface and Ma-
terial Science at the University of Szeged.
476
0021-9517/97 $25.00
c
Copyright ꢀ 1997 by Academic Press
All rights of reproduction in any form reserved.

DIMETHYL ETHER ON Rh CATALYSTS
477
During annealing the adsorbed layer over Rh/Al2O3, the
first spectral changes, indicating the formation of new sur-
face species, occurred at 200 K (Fig. 1A). The positions of
absorption bands were 2960, 2879, 2841, 1472, 1461, 1259,
1157, and 1061 cmꢃ1. Further annealing the sample up to
300 K, a slight attenuation of these bands was observed.
These bands differ from those due to adsorbed DME and
agree well with the absorption bands attributed to the vi-
bration of methoxy species (Table 2).
Similar features were experienced for Rh/TiO2 and
Rh/MgO with the difference that the new absorption bands
produced were weaker and appeared at somewhat lower
frequencies compared to Rh/Al2O3. This was more ex-
pressed for Rh/MgO (Figs. 1B and C). The absorption bands
of DME were also identified for Rh/SiO2 at 130–150 K,
which attenuated and finally disappeared after gradually
heating the adsorbed layer to 300 K without any change in
their positions.
TABLE 1
Some Characteristic Data for Supported Rh
Surface area Dispersion Mean particle
of the support
(m2/g)
of Rh
(% )
diameter
TOFb
(sꢃ1
˚
Supports
of Rh (A)
)
Al2O3
(Degussa P 110 C1)
TiO2
(Degussa P 25)
SiO2
(Cab-O-Sil)
MgO
DAB C
100
150
200
170
38.3
22.6
0.022
4.4
30.4a
45.5
—
28.4
19.0
—
0.021
0.019
4.8
—
(0.003)
a In this case the reduction temperature was 673 K.
b Turnover frequency for the oxidation of DME on supported Rh cata-
lysts under isotherm conditions at 523 K. Rates of CO2 formation were
related to the number of surface Rh atoms.
When the adsorption of DME was performed at 300 K,
we detected the same absorption bands as were observed
previously after heating the adsorbed layer from 150 to
250–300 K (Fig. 1). In addition, we found a weak band in the
region of CO stretching frequency at 2020–2042 cmꢃ1. Fol-
lowing heating the adsorbed layer under continuous evac-
uation led to a decay in the intensities of all the bands with
the exception of the CO band. The latter grew and attained
the highest intensity at 450–500 K, where all the other bands
disappeared or became extremely weak (Fig. 2). These fea-
tures very likely correspond to the complete decomposi-
tion of methoxy species located on supports. Adsorption
of DME on Rh/SiO2 at 300 K produced weak bands cor-
responding to adsorbed DME. After the adsorbed layer
is heated to 300–400 K, two very weak spectral features
remained at 2859 and 2947 cmꢃ1, and a CO band also de-
veloped (Fig. 3).
gases were performed with a Hewlett–Packard 5890 gas
chromatograph using Porapak QS column. The absence of
˝
diffusion limitation was confirmed by the method of Koro¨s
and Novak (8). The system was operated at a total pres-
sure of 1 atm. The carrier gas was argon. In the study of the
decomposition of DME it contained 10.0 vol% DME. In
the oxidation of DME, the carrier gas contained 1 vol%
DME and 5 vol% O2. In both cases the flow rate was
20 ml/min. Note that the experimental conditions were out-
side of both the lower (2 vol% ) and the higher (27 vol% )
explosion limit of DME in air.
Infrared spectroscopic studies were made in a vacuum
IR cell using self-supporting wafers of catalyst powders
(30 ꢄ 30 mm, ꢅ20 mg/cm2) which underwent the same pre-
treatments as before catalytic measurements. The IR cell
used made it possible to register the spectra between
100–300 K. Spectra were recorded with a Biorad (Digilab.
Div.) Fourier transform IR spectrometer (FTS 155).
The dispersity of the supported metals was determined
by hydrogen adsorption at 300 K. Data are listed in Table 1.
TABLE 2
Vibrational Frequencies (cmꢃ1) of Gaseous and Adsorbed DME
on Rh-free Supports
MgOc
2938
a
b
c
c
Mode
(CH3)2O(g)
(CH3)2O/Al2O3
Al2O3
2962
TiO2
3. RESULTS
ꢀas(CH3)
2996ꢆ
2925
2887
2984
2922
2890
2872
2821
1477
1459
(1252)
1166
1092
2952
—
3.1. Interaction of DME with Rh Catalysts
2ꢁ(CH3)
—
—
The interaction of DME with Rh catalysts was first inves-
tigated by means of infrared (IR) spectroscopy. In order to
minimize the thermal dissociation, DME was adsorbed on
the catalyst samples at 110–130 K. Afterwards the IR cell
was evacuated and the catalysts were warmed under con-
tinuous pumping conditions to higher temperatures. In the
presence of DME vapor we identified the characteristic ab-
sorption bands of gaseous DME (Table 2). These bands
were detected at somewhat different positions for all sam-
ples following the evacuation at 120–140 K.
ꢀs(CH3)
ꢁ(CH3)
2817
1470
1456
1244
1179
1102
2845
1473
1458
2839
1470
2821
1465
ꢂ (CH3)
ꢀas(CO)
(1261)
1157
(1250)
1156
1068
1155
1058
1060
a Gaseous DME (Ref. (10)).
b Adsorbed DME on Al2O3 at 150 K (Ref. (10)).
c After heating the adsorbed DME on Rh-free supports to 290 K
(Ref. (16)).

´ ´
SOLYMOSI, CSERENYI, AND OVARI
478
FIG. 1. Infrared spectra of DME adsorbed on supported Rh catalysts recorded as a function of temperature during continuous evacuation. DME
was adsorbed at 130 K: (A) Rh/Al2O3; (B) Rh/TiO2; (C) Rh/MgO. (a) 200 K, (b) 220 K, (c) 250 K, and (d) 300 K.
3.2. Thermal Decomposition of DME
the conversion of DME was also observed. A very low ac-
tivity was found for Rh/MgO. Note that a decrease in the
Measurable decomposition of DME was observed at reduction temperature of Rh/TiO2 from 773 to 673 K en-
523 K. Results for the four catalyst samples are presented in hanced its catalytic activity by a factor of 6.
Fig. 4A. The highest initial conversion (25% ) was exhibited
The catalytic influence of the support materials has been
by Rh/Al2O3, which decayed abruptly to 5% , and gradually also tested under the same conditions. The highest initial
to about 0.5% . The main products of the decomposition activity (0.18% DME conversion) was measured for alu-
were H2, CO, and CH4. The selectivity to CO varied be- mina. The other supports exhibited even less activity if
tween 34–55% and that to CH4 between 15–51% . In very at all.
small amounts, methanol, ethane, and propane were also
In order to establish the possible reasons of the deacti-
detected (Fig. 4B). The high initial activity can be restored vation of the Rh/Al2O3 several measurements have been
by the oxidation and reduction of the catalyst. Although performed. First we determined the uptake of CO on fresh
Rh/TiO2 and Rh/SiO2 were less active catalysts, a decay in and used catalysts. We found that after DME treatment for
FIG. 2. Infrared spectra of DME adsorbed on supported Rh catalysts recorded as a function of temperature during continuous evacuation. DME
was adsorbed at 300 K.

DIMETHYL ETHER ON Rh CATALYSTS
479
FIG. 3. Formation of the CO bands following the annealing of adsorbed DME on Rh catalysts: (A) Rh/Al2O3; (B) Rh/TiO2; (C) Rh/SiO2.
90 min the uptake of CO at 300 K decreased by 60–70% . (Tp = 605 K). The desorption of H2O was also observed
In harmony with this, while the adsorption of CO on an un- during the heating up of the sample, but no peak could be
treated catalyst gave intense absorption bands at 2107 and resolved. Some TPD curves are presented in Fig. 6A. After
2040 cmꢃ1 and a less intense one at 2078 cmꢃ1, after DME desorbing these products, TPR measurements were carried
treatment at 523 K for 90 min, the CO bands shifted to lower out in the presence of H2 flow. The evolution of methane
frequencies and were of much less intensities. It is impor- started at 390 K and did not cease up to 900 K. Two sharper
tant to note that there were no significant spectral changes peaks appeared at 445 and 780 K. The amount of C + CHx
in the presence of CO even after extended adsorption time was also determined by temperature-programmed oxida-
(Fig. 5).
tion (TPO). Independently of the nature of the support,
We also performed TPD and TPR measurements after the oxidation started at lower temperatures, above 330 K,
2 h of reaction on Rh/Al2O3. In the TPD we found the des- reached the highest rate at 450 K and yielded only CO2
orption of CH4 (Tp = 637 K), CO (Tp = 605 K), and H2 and H2O.
FIG. 4. (A) Thermal decomposition of DME on supported Rh catalysts at 523 K in a flow system. (B) The products of the decomposition of DME
on Rh/Al2O3 at 523 K.

´ ´
SOLYMOSI, CSERENYI, AND OVARI
480
achieved at the lowest temperature for Rh/SiO2 followed
by Rh/TiO2 and Rh/Al2O3. The catalytic performance of
Rh/TiO2 was somewhat higher when it was reduced at
673 K. Less active catalyst was again the Rh/MgO, where
the conversion of DME was only 36–38% at 673 K and
100% of conversion was attained only at 873 K.
According to the analysis, the total oxidation of DME
can be described by the equation
– –
CH3 O CH3 + 3O2 = 2CO2 + 3H2O.
[1]
At lower temperatures, below 573–600 K, small amounts
of CH4, C2H6, C3H8, and CH3OH were also produced.
Above ꢂ600 K, however, only CH4 was detected in traces.
At the conversion of 90–95% , the selectivity to CO2 forma-
tion (based on C-containing gaseous compounds) is around
96–98% on all catalyst samples. Note that the atomic bal-
ances on C and H2 correspond well to the calculated values
(95 and 97% ).
The oxidation of DME also occurred on pure supports
but at significantly higher temperatures (Fig. 7B). The most
active support was titania followed by alumina, silica, and
magnesia. In the empty reactor the reaction began only
above 873 K.
FIG. 5. FTIR spectra following CO adsorption at 300 K on unused
Rh/Al2O3 and with DME-treated Rh/Al2O3 samples at 523 K.
The oxidation of DME was also studied under isotherm
conditions in time on stream at 523 K. Data are presented in
Fig. 8A. Almost 65–70% conversion of DME was measured
for Rh/Al2O3 and Rh/SiO2. No decay in the performance of
the catalyst occurred in 4 h. High rate of oxidation was mea-
sured for Rh/TiO2 samples. Only a very slight oxidation of
DME was observed at this temperature on Rh/MgO. Apart
3.3. Oxidation of DME
In the first series of measurements the oxidation of DME
wasstudied bygraduallyheatingthe catalyst bed in the pres-
ence of reacting gas mixture from 300 K. The conversion-
temperature curves are shown in Fig. 7A. The reaction
began slightly above 450 K. A complete oxidation was
FIG. 6. (A) Temperature programmed desorption (TPD) following the decomposition of DME on Rh/Al2O3 at 523 K for 2 h. (B) After TPD, the
sample was cooled in He to room temperature, heated up in the H2 flow (TPR experiment), and the amount of methane produced was determined.

DIMETHYL ETHER ON Rh CATALYSTS
481
FIG. 7. (A) Oxidation of DME on supported Rh catalysts and (B) on Rh-free supports at different temperatures in a flow system as a function of
temperature.
from the very low activity of Rh/MgO, no support effect of reactants on the rate of oxidation of DME. We found
was observed: the turnover frequencies were practically the that an increase in the concentration of both compounds
same for the three active catalysts (Table 1).
increased the rate of oxidation. The kinetic order with re-
From the variation of the space velocity for Rh/Al2O3 we spect to O2 was found to be 0.4. The order with respect to
obtained that the rate of the oxidation of DME increased DME is about 0.8.
linearly with the increase of the space velocity (Fig. 8B). On
In order to knowmore about the interaction ofDME with
this catalyst we determined the effect of the partial pressure the catalyst, the reaction ofDME with reduced and oxidized
FIG. 8. (A) Oxidation of DME on supported Rh catalysts under isotherm conditions at 523 K. (B) Dependence of the rate of CO2 formation on
the space velocity in the oxidation of DME on Rh/Al2O3 at 523 K.

´ ´
SOLYMOSI, CSERENYI, AND OVARI
482
Rh/Al2O3 was also followed by pulse technique at 523 K. adsorbed on oxidic supports under the same conditions are
One pulse contained 1.56 ꢃmol of DME. Assuming that the of great use (16). The important absorption bands and their
oxidized Rh is in the form of Rh2O3, the complete oxidation assignements are listed in Table 2. The main feature of the
of one pulse of DME requires 24% of the oxygen of Rh2O3. adsorbed DME on oxides is its partial transformation into
In the case of reduced sample CO, H2, and CH4 were the methoxy species around 200 K. This process was first estab-
main products with minor amounts of ethane, methanol, lished on alumina surface byYatesetal. (10). The disappear-
and propane. The rates of their formation were practically ance of absorption bands at 2922, 2821, and 1092 cmꢃ1 and
constant during the use of 40 DME pulses. On the oxidized the appearance of bands at 2962, 2848, and 1059 cmꢃ1 were
sample the primary products were CO2 and H2O. With the considered as evidence for the transformation of adsorbed
progress of the reduction of rhodium oxide, the amount of DME into methoxy species (10). Following the adsorption
these compounds gradually decreased and the formation of DME on oxides at and above 300 K, this transformation
of CO, CH4, C2H6 and somewhat later methanol came into occurred instantaneously. Silica represents a special case as
prominence. The formation of the latter compound showed it exhibits a rather low tendency toward methoxy forma-
a maximum when the production of CO2 and H2O ceased. tion. The weak absorption bands at 2859 and 2947 cmꢃ1
observed after annealing the adsorbed DME to 350–400 K
agree well with the bands at 2860 and 2956 cmꢃ1 obtained
4. DISCUSSION
following the esterification of SiOH with CH3OH (17, 18).
4.1. Interaction of DME with Solid Surfaces
– –
These bands were attributed to the formation of Si O CH3
– –
surface complex (17). Note that the formation of Si O CH3
There is very little information concerning the interac-
tion of DME with oxides. The adsorption of DME on alu-
mina has been studied by Yakerson et al. (9) and Yates et al.
(10) by means of IR spectroscopy. It was found that DME
molecularly adsorbs on alumina, via hydrogen bonding to
the surface hydroxyl groups, at 150 K. Heating the adsorbed
layer to >250 K caused the desorption of the ether and the
formation of surface methoxy species (10). Recently, the
adsorption of DME on the protonated ZSM-5 zeolite was
investigated by FTIR spectroscopy (11). Results indicated
that DME molecules adsorbed on the OH groups by hydro-
gen bonding irrespective of the acidity of the OH groups
and that the oxonium ions of DME were not produced on
the studied surface.
DME is rather unreactive on metal surfaces. At low
temperature, at 100 K, it adsorbs molecularly on Pt(111),
Cu(100), and Al(111) surfaces and desorbs without under-
going dissociation (13, 14). On polycrystalline Pd, DME
underwent a thermally activated decomposition at 300 K
yielding CO and H (15). Recently the interaction of DME
with Rh(111) was studied in our laboratory by HREELS
(7). Following the low temperature adsorption, only molec-
ular adsorption and desorption were experienced. The sit-
uation was different when the adsorption temperature was
raised to 250–300 K. In this case CHO and/or CHxOCHx
were identified, which decomposed to CO and H2 at higher
temperatures. Vibration losses of adsorbed DME, CH3, and
CH3O were not seen in the HREEL spectra. On O-dosed
Rh(111), however, methoxy species was identified. The
presence of K adatoms promoted the dissociation of DME
and stabilized the primary products of its dissociation (8).
was recently observed in the interaction of CH3OH and
CH3Cl with silica and with silica supported metals (19–21).
In the presence of 1% Rh we found the same spectral
features as determined for Rh-free supports at 110–290 K
(Fig. 1). Accordingly, the adsorption and the transforma-
tion of DME into methoxy is practically unaltered by Rh
in this temperature range. This is not surprising taking into
account the high surface area of the supports. Performing
the adsorption on supported Rh at 300 K, and annealing
the adsorbed layer under continuous evacuation, we ob-
served that, in addition to the absorption bands of methoxy,
a CO band developed at 2020–2041 cmꢃ1 for all Rh sam-
ples (Fig. 3). Its intensity increased up to 450–500 K with
the concomitant attenuation of the methoxy bands (Fig. 2).
These features suggest that the methoxy species, located on
the oxides, migrated to the Rhx, where it underwent decom-
position
CH3O(a) = CO(a) + 3H(a)
.
[1]
We cannot exclude that a small fraction of methoxy species
may exist on Rhx itself. Following the adsorption of CH3OH
on Rh(111) surface, it was found that the stability region of
methoxy on Rh(111) is very narrow; it decomposes readily
below 300 K (22, 23).
In many respects similar processes proceeded during the
high temperature decomposition of DME under dynamic
conditions. The decomposition products identified by gas
chromatograph are H2, CO, and CH4 with minor amountsof
methanol, ethane and propane. The decomposition follows
first-order kinetics. In the decomposition we maycount with
the following steps
4.2. Adsorption and Decomposition of DME
on Supported Rh
CH3OCH3(g) = CH3OCH3(a)
CH3OCH3(a) = CH3(a) + CH3O(a)
[2]
[3]
For the interpretation of the IR spectroscopic results ob-
tained on supported Rh, the IR characteristics of DME

DIMETHYL ETHER ON Rh CATALYSTS
483
The evolution of a large amount of hydrogen deserves spe-
cial attention. Its peak temperature is 590–600 K, which is
much higher than that measured following the adsorption
of hydrogen on the same surface. This suggests that the hy-
drogen evolution from the used catalyst may stem from the
decomposition of strongly adsorbed hydrocarbon species,
which is also responsible for the release of methane at this
temperature.
CH3O(a) = CO + 1.5 H2
2CH3(a) = C2H6(g)
CH3(a) = C + 1.5 H2(g)
[4]
[5]
[6]
[7]
CH3(a) + H(a) = CH4(g)
.
The coupling of CH3 to C2H6 is very limited as only a
very small amount of ethane was identified in the prod-
ucts. It is very likely that the decomposition of CH3 to car-
bon proceeds through the transient formation of CH2 and
CH species. The formation of a small amount of propane is
probably the result of the combination of C2H6 with CH2.
An alternative route of the decomposition of DME is
its partial dehydrogenation via the transient formation of
4.4. Oxidation of DME
An interesting feature of the complete oxidation of DME
over supported Rh is the lack of the deactivation of the cata-
lysts which is in contrast with the catalytic decomposition
of DME. This suggests that the strongly adsorbed species
formed in the decomposition on Rh are not produced in
the oxidation reaction nor are they also oxidized. Another
behavior worth mentioning is that the specific activity of
Rh is practically the same for all samples, except that for
the Rh/MgO (Table 1).
In the oxidation of DME we may consider two major
routes. (i) The activation of DME, which consists of the
following steps: formation of methoxy species on the sup-
ports, migration of this species to the Rh, decomposition of
CH3O at the metal/support interface, and the oxidation of
its decomposition products. (ii) Another possibility is the
activation of DME by the adsorbed oxygen atoms bonded
to the Rh to give a transient surface compound
– –
[CHx O CHx] adsorbed complex which would decompose
through CH and CHO species to H, C, and CO. In the
study of the interaction of DME with a clean Rh(111) at
and above 250 K, where DME is not stable on the surface,
we detected CO and H2 without any CH4 (7). By means
of HREELS we identified a vibration loss at 1235 cmꢃ1
–
which we attributed to the C O stretching mode [ꢀ(CO)]
of ꢄ2-HCO. On the basis of these findings it was concluded
that the major pathway of the decomposition of DME on
Rh(111) is its partial dehydrogenation. The primary dissoci-
ation to adsorbed CH3 and CH3O represents only a minor,
undetectable step.
In the present case, however, working at higher pressure
and using oxidic supports the reaction (3) is the primary
process as the adsorption of DME on the supports gives
methoxy complex and we measured significant amount of
methane among the products.
The interesting feature of the decomposition of DME on
Rh/Al2O3 is the dramatic decay in the catalytic effect at the
very beginning of the reaction. The possible reasons are
(i) blocking the active area of Rh by strongly adsorbed the decomposition of which results in the formation of ad-
species, (ii) a decrease in the dispersion of Rh due to ag- sorbed CH3O on the Rh itself. This reaction channel is sup-
glomerization of Rh, (iii) deactivation of the metal/support ported by the results obtained on Rh(111) under UHV con-
interface particularly effective for the activation of DME. ditions (7). Following the adsorption of DME on O-covered
Let us examine these possibilities separately. From the Rh(111), we observed the characteristicvibrationsofCH3O
study of the uptake of CO we measured by about 60–70% complex. This surface compound may react with other ad-
less CO adsorption on used catalyst compared to the un- sorbed O atoms to give CO2 and H2O, or it decomposes to
used sample. This is consistent with the IR spectrum of adsorbed CO and H, which are subsequently oxidized. The
used sample which showed that a significant amount of CO fact that in situ FTIR spectra always showed the presence of
remained on the surface after DME decomposition. In ad- adsorbed CO even during the high temperature oxidation
dition, whereas the adsorption of CO produces an intense suggests the importance of the second possibility.
pair of the dicarbonyl (Rh1(CO)2) species on a fresh cat-
As regards the importance of the two main routes it
alyst, on used catalyst the development of the dicarbonyl should be kept in mind that the adsorption and dissocia-
species is very limited. This is a strong indication of the tion of DME on silica is very limited compared to other
agglomeration and deactivation of Rh particles (24, 25). It supports (16). At the same time, the catalytic efficiency of
has been established before that the formation of Rh1(CO)2 Rh/SiO2 is almost the same as that of Rh/Al2O3 (Figs. 7, 8
species from Rhx cluster is the result of the CO-induced dis- and Table 1). This feature suggests that the reactions occur-
ruption of Rhx crystallites (24–29). The combined TPD and ring on the supports (formation and migration of methoxy)
TPR measurements revealed that in addition to adsorbed do not play an important role in the complete oxidation of
CO other species are also present on the catalyst (Fig. 6). DME on supported Rh catalysts.

´
´
484
SOLYMOSI, CSERENYI, AND OVARI
CONCLUSIONS
7. Klive´nyi, G., and Solymosi, F., to be published.
8. Koro¨s, R. M., and Nova´k, J. M., J. Chem. Eng. Sci. 22, 470 (1967).
˝
9. Yakerson, V. I., Lafer, L. N., Danyushevski, V. Ya., and Rubinshtein,
A. M., Akad Nauk SSSR Ser. Khim. 11, 2246 (1967).
10. Chen, J. G., Basu, P., Ballinger, T. H., and Yates, J. T., Langmuir 5, 352
(1989).
11. Fujino, T., Kashitani, M., Kondo, J. N., Domen, K., Hirose, C., Ishida,
M., Goto, F., and Wakabayashi, F., J. Phys. Chem. 100, 11649 (1996).
12. Kanazawa, Y., and Nukada, K., Bull. Chem. Soc. Jpn. 35, 612 (1962).
[Kiselev, A. V., and Lygin, V. I., “Infrared Spectra of Surface Com-
pounds,” Wiley, New York, 1975]
13. Sexton, B. A., and Hughes, A. E., Surf. Sci. 140, 227 (1984).
14. Ng, L., Chen, J. G., Basu, P., and Yates, J. T., Jr., Langmuir 3, 1161
(1987).
15. Lu¨th, H., Rubloff, G. W., and Grobman, W. B., Surf. Sci. 63, 325 (1977).
16. Ova´ri, L., and Solymosi, F., to be published.
17. Borello, E., Zecchina, A., and Morterra, C., J. Phys. Chem. 71, 2938,
2945 (1967).
18. Kitahara, S., Bull. Chem. Soc. Jpn. 49, 3389 (1976). [Takezawa, N., and
Kobayashi, H., J. Catal. 25, 179 (1972)]
19. Rasko´, J., Bontovics, J., and Solymosi, F., J. Catal. 143, 138 (1993).
20. Rasko´, J., Bontovics, J., and Solymosi, F., J. Catal. 146, 22 (1994).
21. Driessen, M. D., and Grassian, V. H., J. Catal. 161, 810 (1996).
22. Solymosi, F., Berko´, A., and Tarno´czi, T. I., Surf. Sci. 141, 533 (1984).
[Solymosi, F., Berko´, A., and Tarno´czi, T. I., J. Phys. Chem. 88, 6170
(1984)]
(i) Dimethyl ether adsorbed on supported Rh is trans-
formed into methoxy species at 200–300 K. The reaction
occurs on the supports and is influenced only slightly with
the presence of Rh.
(ii) The decomposition of ether molecules is catalysed
by Rh above 473 K to give CO, H2, and small amounts of
hydrocarbon products. Rhodium loses its catalytic activity
due to the strongly adsorbed species formed in the decom-
position.
(iii) In contrast, supported Rh samplesexhibited high and
constant activity in the total oxidation of dimethyl ether;
100% conversion was reached at 600–650 K.
(iv) It is assumed that the oxidation reaction proceeds
through the activation of dimethyl ether by adsorbed O
atoms and the transient formation of methoxy species
on Rh.
ACKNOWLEDGMENTS
Financial support of this work by OTKA (contract No. T 022869) and
loan of rhodium chloride from Johnson–Matthey are greatfully acknowl-
edged.
23. Houtman, C., and Barteau, M. A., Surf. Sci. 248, 57 (1991). [Houtman,
C., and Barteau, M. A., Langmuir 6, 1558 (1990)]
24. Solymosi, F., and Pa´sztor, M., J. Phys. Chem. 89, 4789 (1985).
[Solymosi, F., and Pa´sztor, M., J. Phys. Chem. 90, 5312 (1986)]
25. Solymosi, F., and Ba´nsa´gi, T., J. Phys. Chem. 97, 10133 (1993).
REFERENCES
1. Meisel, S. L., McCullogh, J. P., Lechthaler, C. H., and Weisz, P. B., 26. Van’t Bilk, H. F. J., Van Zon, J. B. A. D., Huizinga, T., Vis, J. C.,
CHEMTECH 6, 86 (1976). [Chang, C. D., “Hydrocarbons from
Methanol,” Dekker, New York, 1983]
Koningsberger, D. C., and Prins, R., J. Amer. Chem. Soc. 107, 3139
(1985).
2. Spivey, J. J., Chem. Eng. Commun. 110, 123 (1991).
3. Rouhi, A. M., C&EN May 29, 37 (1995).
27. Zaki, M. I., Kunzmann, G., Gates, B. C., and Kno¨zinger, H., J. Phys.
Chem. 91, 1486 (1987).
4. Fleish, T., Basu, A., Gradassi, M. J., and Masin, J. G., Stud. Surf. Sci.
Catal. 107, 117 (1997).
28. Basu, P., Panayotov, D., and Yates, J. T., Jr., J. Amer. Chem. Soc. 110,
2074 (1988).
5. Solymosi, F., Csere´nyi, J., and Ova´ri, L., Catal. Lett. 44, 89 (1997).
6. Bugyi, L., and Solymosi, F., Surf. Sci., in press.
29. Solymosi, F., and Kno¨zinger, H., J. Chem. Soc. Faraday Trans. 86, 389
(1990).