Catalytic reduction of NO by CO over rhodium catalysts. 3. The role of surface isocyanate species

Article abstract of DOI:10.1006/jcat.2000.2892

Catalytic reduction of NO by CO or hydrocarbons has been the subject of intense study in recent years due to the importance of these reactions for automotive emission control. The formation, stability, and reactivity of surface isocyanate species with CO,

Full text of DOI:10.1006/jcat.2000.2892

Journal of Catalysis 193, 303–307 (2000)
doi:10.1006/jcat.2000.2892, available online at http://www.idealibrary.com on
Catalytic Reduction of NO by CO over Rhodium Catalysts
3. The Role of Surface Isocyanate Species
1
2
Dimitris I. Kondarides, Tarik Chafik, and Xenophon E. Verykios
Department of Chemical Engineering, University of Patras, GR-26500 Patras, Greece
Received January 28, 2000; revised April 11, 2000; accepted April 17, 2000
�
�
namely Rh–NO (high), Rh–NO (low), Rh(NO)2, and Rh–
+
The formation, stability and reactivity of surface isocyanate NO . It was proposed that production of N2 and N2O in
6
+
)
species with CO, NO, and O2 has been examined over Rh/TiO2(W
the gas phase involves the negatively charged and dini-
trosyl species, respectively, while the positively charged
species are inactive. The formation and reactivity of sur-
face species present under NO–CO reaction conditions as
well as the effect of oxygen on their nature and population
was also examined over Rh/TiO2 catalysts. It was found that
formation of isocyanate species is favored on the catalyst
surface under conditions where NO conversion to reduc-
tion products is observed. However, no straightforward ev-
idence on the exact role of isocyanates was obtained due
to their low thermal stability, which did not enable detailed
examination.
catalysts, under conditions of NO reduction by CO, using FTIR and
6
+
transient MS techniques. It has been found that W doping of TiO2
results in stabilization of the Rh–NCO species and in expansion of
the temperature window of N2O formation toward lower tempera-
tures by ca. 50 C. Surface isocyanates react with NO to yield gas-
phase N2O, thus providing an alternative route for the production
�
of nitrous oxide.
�c 2000 Academic Press
Key Words: reduction of NO; rhodium catalyst; surface iso-
cyantes; in situ FTIR; transient reactivity.
1
. INTRODUCTION
In the present study, the formation, stability, and re-
activity of isocyanates are examined over an Rh catalyst
Catalytic reduction of NO by CO or hydrocarbons has
been the subject of intense investigation in recent years
due to the importance of these reactions for automotive
emission control [(1) and reference therein]. The reaction
mechanism has been investigated in many laboratories,
over different catalysts. However, open questions still exist
regarding the nature and role of surface intermediates and,
especially, of adsorbed isocyanates. The latter species have
been observed in several FTIR studies dealing with reduc-
tion of NO by CO (2–7) and have been mostly regarded as
spectator species, not directly involved in the formation of
the reaction products. On the other hand, studies on the re-
duction of NOx by hydrocarbons and ethanol showed that
there is a correlation between the formation of NCO(a) and
catalytic activity (8–10).
dispersed on W6 -doped TiO2, employing transient MS and
FTIR techniques. The present catalyst seems to stabilize
surface isocyanates, thus enabling the examination and
elucidation of their role in the chemistry of NO reduction
by CO.
+
2
. EXPERIMENTAL
The W6 -doped TiO2 carrier, containing 0.45 at.% of the
+
dopant cation, was prepared with the solid-state diffusion
technique. The 0.5 wt% Rh/TiO2(W6 ) catalyst was pre-
pared by the incipient wetness impregnation method us-
ing RhCl3 � 3H2O (Alfa) as metal precursor. The catalytic
performance of the NO–CO reaction was examined in the
+
�
temperature range of 100 to 500 C. Experiments were per-
formed using 50 mg of catalyst and a feed composition of
.25% NO and 0.50% CO (balance He). The total feed flow
In our recent studies (11, 12), the adsorption and dis-
placement characteristics of NO and CO have been ex-
amined over Rh catalysts supported on undoped and
-doped TiO2. It has been found that four kinds of ni-
trogen oxide species may coexist in the adsorbed mode,
0
3
3
6+
was 150 cm /min (W/F = 0.02 g s/cm ).
W
FTIR spectra were obtained using a Nicolet 740 FTIR
spectrometer equipped with a TGS detector and a KBr
beam splitter. The catalyst, in finely powdered form, was
placed into a holder, which enables gas flow through the
sample, and its surface was carefully flattened to increase
the intensity of the reflected IR beam. The transient MS
1
Permanent address: Department of Chemistry, Faculty of Science and
Technology, University of Tangier, P.O. Box 416, Tangier, Morocco.
2
To whom correspondence should be addressed. Fax no: +30-61-991
5
27. E-mail: verykios@chemeng.upatras.gr.
303
0021-9517/00 $35.00
Copyright �c 2000 by Academic Press
All rights of reproduction in any form reserved.

3
04
KONDARIDES, CHAFIK, AND VERYKIOS
experiments were carried out in an independent reactor by
switching the feed composition to the reactor with the use
of chromatographic valves equipped with electronic actua-
tors. The concentrations of the gases used in the transient
MS experiments were the same as those used in the corre-
sponding FTIR experiments, but were diluted with He in-
stead of Ar. In all experiments presented here, an amount of
Ar was added to the feed. This was done because the Ar re-
sponse curve represents the flow pattern and the residence
time distribution of gaseous reactants flowing through the
reactor and the transportation lines, since Ar does not in-
teract with the catalyst. Since the principal peaks of CO
and N2 (m/z = 28) and CO2 and N2O (m/z = 44) cannot
be separated, isotopic 15NO was used in the transient MS
experiments reported here.
The flow rate in all experiments reported here was
5 ml/min (ambient). It should be mentioned that all ex-
perimentalconditions(gasconcentrationsflowrate, massof
catalyst, pretreatment conditions, etc.) were kept the same
as those in the corresponding FTIR experiments and, there-
3
FIG. 2. (A) FTIR spectra recorded following interaction of 0.25%
6
+
NO–0.5% CO (balance Ar) with reduced Rh/TiO2(W ) in the temper-
�
ature range of 25–550 C. (B) FTIR spectra obtained as a function of
time-on-stream following interaction of the reduced catalyst with 0.25%
�
NO–0.5% CO (in Ar) at 250 C (a–e) and subsequent isothermal treatment
fore, a direct comparison between the transient FTIR and with Ar flow (f–h).
MS results can be made. Further details on the procedures
and techniques employed are discussed elsewhere (11, 12).
abruptly increases with increasing reaction temperature
�
�
above 200 C and reaches 100% at 270–280 C. In the tem-
�
3
. RESULTS AND DISCUSSION
perature range of ca. 200–300 C the catalyst is highly selec-
tive toward N2O formation; the yield of N2O goes through a
The catalytic performance of 0.5% Rh/TiO2 (W6+), ex-
pressed in terms of steady-state conversion of NO to reduc-
tion products and in terms of yield of N2O, as functions of
the reaction temperature, is presented in Fig. 1. For com-
parison, the corresponding results obtained over the un-
doped catalyst [from Ref. (12)] are also shown. Regarding
the doped catalyst, it is observed that the conversion of NO
�
maximum in the temperature range of 250–270 C and drops
�
to zero at temperatures above 350 C. Comparison with the
results obtained from the undoped 0.5% Rh/TiO2 catalyst
shows that W6 doping of TiO2 results in lowering of the
temperature at which 100% conversion of NO is observed
+
�
by 40–50 C, mainly by widening the temperature window
for N2O formation toward lower temperatures.
FTIR spectra obtained following interaction of the cat-
alyst with the NO/CO mixture in the temperature range
�
of 25 to 550 C are presented in Fig. 2A. It should be
noted that prior to each experiment the catalyst was re-
�
duced in an H2 flow at 300 C. Spectra were obtained af-
ter steady state was reached. It is observed that interac-
tion of the reaction mixture with the catalyst leads to the
formation of several surface species, the nature and popu-
lation of which depend on the reaction temperature. The
assignment of the FTIR bands has been discussed in detail
elsewhere (11, 12). Briefly, these bands are attributed to iso-
�
1
� 1
cyanates (2210 cm ), carbonyls (2150–2020 cm ), and ni-
�
1
trosyls (1920–1660 cm ) bonded to rhodium, and to C- and
N-containing species associated with the support (bands
below 1660 cm ).
As observed in Fig. 2A, at 250 C, i.e., just above the
light-off temperature (compare with Fig. 1), the spectrum
is dominated by an intense band at 2210 cm , attributed
to Rh–NCO, and is accompanied by a weak band, due to
�
1
�
�
1
FIG. 1. Conversion (% ) of NO and yield (% ) of N2O obtained over
6
+
Rh catalystssupported on undoped and W -doped TiO2 asfunctionsofre-
action temperature. Feed composition: 0.25% NO, 0.5% CO (balance He),
0
� 1
Rh –CO, located at 2025 cm . Increasing the reaction tem-
perature to 350 C results in a decrease of the intensity of
3
�
W/F = 0.02 g s/cm .

CATALYTIC REDUCTION OF NO BY CO OVER Rh CATALYSTS
305
the Rh–NCO band and in an increase of the intensity of the gested that the surface coverages of these species are closely
0
Rh –CO band. Further increase of the reaction tempera- related.
�
ture to 450 and 550 C leads to the progressive disappear-
ance of all bands from the spectra (Fig. 2A).
The intensity of the Rh–NCO band continuously in-
creases with time and reaches a plateau after ca. 60 min
0
Comparison of the results presented in Figs. 1 and 2A on stream. During this period, the intensity of the Rh –CO
�
1
shows that there is a correlation between the population of band at 2025 cm does not significantly change. Isothermal
the surface isocyanate species and the yield of N2O in the switch of the flow from CO/NO to Ar (Fig. 2B, f–h) results in
gas phase. In particular, it is observed that both the yield of a slow decrease of the intensity of the Rh–NCO band, which
�
1
0
N2O and the intensity of the Rh–NCO band at 2210 cm
is initially accompanied by an increase of the Rh –CO band,
increase with increasing reaction temperature, go through a giving additional evidence that there is an interconversion
�
maximum above the light-off temperature (ca. 250 C), and between these species, as reported previously (12). The iso-
�
then diminish at temperatures above 350 C.
cyanate species are very stable and cannot be removed from
In order to elucidate the role of the isocyanate species, the catalyst surface even after 30 min on stream. This is also
their formation, stability, and reactivity have been further the case for the various species associated with the support.
�
examined at the reaction temperature of 250 C, at which Comparison with the corresponding experiments obtained
the population of this species is maximized. The IR spectra over the undoped catalyst (12), where isocyanates were re-
obtained as a function of time-on-stream following expo- moved from the surface after ca. 5min on stream, showsthat
�
6+
sure of the reduced catalyst to the NO/CO mixture at 250 C
W
doping of TiO2 stabilizes Rh–NCO species. It should
are shown in Fig. 2B. It is observed that immediately after be noted that the thermal stability of isocyanate species was
exposure to the reaction mixture, weak bands at 1660 (sh) also examined with the transient MS technique by isother-
�
1
and 2025 cm appear in the spectral region of interest, mal switching of the feed composition from CO/NO/He to
�
0
�
attributed to Rh–NO and Rh –CO species, respectively He at 250 C, under conditions identical to those reported in
11, 12). Increasing time of exposure results in the devel- Fig. 2A. However, under the experimental conditions em-
opment of additional bands in the frequency region below ployed, the possible formation of reduction products, origi-
650 cm , attributed to N- and C-containing species asso- nating from the slow decomposition of isocyanates, was not
ciated with the support. The weak band at 1660 cm dis- possible.
appears in less than 2 min on stream. It is only then that The interaction of the reduced catalyst with the NO/CO
the Rh–NCO band at 2210 cm starts to develop. This mixture, at 250 C, was also examined with the transient
behavior is in accordance with the results of Hecker and MS technique, in order to investigate the transient behav-
Bell (3) who found that the intensities of the bands for ior of reactants and products in the gas phase. As observed
(
�
1
1
�
1
�
1
�
�
Rh–NO and Rh–NCO varied inversely in response to in Fig. 3A, purging with He and switching to NO/CO re-
changes of the NO and CO partial pressures and sug- sults in the immediate appearance of CO2, CO, and N2 at
�
FIG. 3. Transient MS responses of reactants and products obtained at 250 C after the switch (A) He → 0.25% NO–0.5% CO (in He) over the
reduced catalyst and after the switches (B) He → 0.25% NO/He and (C) He → 1% O2/He over catalysts previously treated with 0.5% CO–0.25% NO
�
mixtures at 250 C.

3
06
KONDARIDES, CHAFIK, AND VERYKIOS
tion of NO appears at the exit of the reactor immediately
after the He → 0.25% NO/He switch, due to adsorption on
the catalyst surface and to formation of reduction products
in the gas phase, and it rapidly increases after ca. 2.0 min on
stream. The responses of CO2 and N2 follow the same trend
exhibiting two maxima, one immediately after the switch
and a second after ca. 2 min on stream. The responses of
both species then decrease and drop to zero when the con-
centration of NO in the gas phase reaches that of the
feed. N2O formation also takes place immediately after the
switch, reaches a plateau, and drops following the responses
of N2 and CO2. The amounts of N2, N2O, CO, and CO2
formed during the first 10 min on stream were 12.4, 12.9,
1
4.9, and 4.3 �mol/g-cat., respectively.
It is of interest to note that in the case of the undoped
�
�
1
Rh/TiO2 catalyst (12), where isocyanates are not stable and
are almost completely removed from the catalyst surface
FIG. 4. Variation of the normalized intensity of the 2210 cm band,
formed under 0.5% CO–0.25% NO reaction conditions at 250 C, as a
function of time on stream under isothermal flow of He and following upon purging with He, significantly different results were
interaction with 0.5% CO, 0.25% NO, and 1% O2.
observed in a similar experiment. Switching the flow to NO
initially resulted in the observation of only N2 and CO2 in
the gas phase. It was only when the response of N2 started to
decrease that N2O and NO were observed at the exit of the
reactor. In that case, formation of N2O was attributed solely
to interaction of surface dinitrosyl species with NO and ad-
jacent reduced Rh sites (12). In the present case, N2O, CO,
and NO appear at the exit of the reactor immediately upon
switching to NO (Fig. 3B). The appearance of NO indicates
the exit of the reactor. The response of N2 progressively
decreases with time of exposure and reaches a plateau
after ca. 2.0 min on stream. The response of N2O fol-
lows the opposite trend, i.e., it gradually increases with
time on stream and reaches a plateau, also after ca. 2.0
min. Comparison with the corresponding FTIR experi-
ments of Fig. 2B shows that the plateau is reached when
0
that there are fewer vacant Rh sites, compared with the un-
�
� 1
the Rh–NO band at 1660 cm disappears from the spec-
tra and the isocyanate species start to populate the catalyst
surface.
doped catalyst, for the dissociative adsorption of NO. This
is probably because these sites are in the present case occu-
pied by Rh–NCO species, which are relatively more stable
on the doped catalyst. It is then reasonable to suggest that
isocyanates are responsible for the production of N2O and
CO, thus providing an additional route for the production
of nitrous oxide:
The reactivity of the adsorbed species, formed under
�
CO–NO reaction conditions at 250 C, was investigated by
monitoring the change of the IR bands after purging with
Ar for 1 min and switching to 0.5% CO, or 0.25% NO, or
1
% O2 (in Ar) flow. Results are summarized in Fig. 4 where
�
1
the normalized intensity of the Rh–NCO band at 2210 cm
Rh–NCO + NO → Rh + N2O + CO.
[1]
(
I/I0) is plotted as a function of time-on-stream. For com-
parison, the corresponding curve obtained upon isothermal
This is further supported by the ratio of N2O and CO
switch to He (results obtained from Fig. 2B, f–h) is also plot- produced, which is close to unity (see above), as predicted
ted. It is observed that the isocyanates do not interact with by Eq. 1. This reaction could explain the expanded temper-
the incoming CO, as indicated by the rate of disappearance ature window for N2O formation observed under steady-
�
1
6+
of the 2210 cm band, which is similar with that observed state conditions over Rh/TiO2(W ) (Fig. 1) compared with
under He flow. However, in our previous study it was shown the undoped catalyst. This behavior may be related to the
that there is an interconversion between the Rh–NCO and higher ability of the present catalyst to stabilize isocyanates
0
Rh –CO species accompanied by the formation of nitride under given reaction conditions.
(
12).
Regarding the reactivity of isocyanates with oxygen,
On the other hand, interaction of isocyanates with O2 or the transient MS results obtained following the switch
�
1
NO results in faster disappearance of the 2210 cm band He → 1% O2 over a catalyst previously exposed to the
(
Fig. 4). In order to further examine this behavior, similar CO–NO mixture for 10 min are shown in Fig. 3C. Imme-
experiments have been conducted employing the transient diately after the exposure to oxygen, CO2 is produced ac-
MS technique. Results are presented in Figs. 3B and 3C, companied by evolution of N2 and small amounts of N2O.
where the transient responses of reactants and products are Formation of all products stops after less than ca. 0.3 min.
plotted as functions of time-on-stream under 0.25% NO (B) The amounts of N2, N2O, and CO2 formed were 6.9, 0.8,
and 1% O2 (C) flow. It is observed that a small concentra- and 5.2 �mol/g-cat., respectively. Comparison with the

CATALYTIC REDUCTION OF NO BY CO OVER Rh CATALYSTS
ACKNOWLEDGMENTS
307
corresponding FTIR results of Fig. 4 clearly shows that
isocyanates are not responsible for the formation of
these products since the intensity of the Rh–NCO band
varies much more slowly. Similar results were reported by
Hadjiivanov et al. (13) who studied by FTIR the forma-
tion and reactivity of isocyanate species on Fe–ZSM-5 and
found that they are inert toward oxygen. However, the au-
thors observed that isocyanates disappeared in an NO + O2
atmosphere (i.e., in the presence of NO2) even at ambient
temperature.
This work was partly funded by the Commission of the European Com-
munity, under contract BRPR-CT97-0460.
REFERENCES
1. Parvulescu, V. I., Grange, P., and Delmon, B., Catal. Today 46, 233
(
1998).
2. Hyde, E. A., Rudham, R., and Rochester, C. H., J. Chem. Soc.,
Faraday Trans. 80, 531 (1984).
3
4
5
. Hecker, W. C., and Bell, A. T., J. Catal. 85, 389 (1984).
. Dictor, R., J. Catal. 109, 89 (1988).
. Krishnamurthy, R., Chuang, S. S. C., and Balakos, M. W., J. Catal.
4
. CONCLUSIONS
1
57, 512 (1995).
(
1) Isocyanate species adsorbed on Rh are significantly
6
. Solymosi, F., and Bansagi, T., J. Catal. 156, 75 (1995).
7. Chuang, S. S. C., and Tan, C.-D., J. Catal. 173, 95 (1998).
6+
stabilized over the W -doped Rh/TiO2 catalyst and are
present on the catalyst surface under all experimental con-
ditions which lead to formation of N2O in the gas phase.
As a result, reduction of NO by CO over doped catalysts
occurs at lower temperatures (ca. 50 C) compared to the
undoped catalyst mainly due to expansion of the tempera-
ture window of N2O formation.
8
. Ukisu, Y., Sato, S., Muramatsu, G., and Yoshida, K., Catal. Lett. 11,
77 (1991).
. Ukisu, Y., Sato, S., Abe, A., and Yoshida, K., Appl. Catal. B 2, 147
1993).
1
9
(
�
1
1
1
1
0. Kameoka, S., Chafik, T., Ukisu, Y., and Miyadera, T., Catal. Lett. 55,
211 (1998).
1. Chafik, T., Kondarides, D. I., and Verykios, X. E., J. Catal. 190, 446
(
2000).
2. Kondarides, D. I., Chafik, T., and Verykios, X. E., J. Catal. 191, 147
2000).
(
2) Rh–NCO species do not interact directly with CO
and O2. In contrast, isocyanates react with NO to yield N2O,
thus providing an alternative route for the production of
nitrous oxide.
(
3. Hadjiivanov, K., Kn o¨ zinger, H., Tsyntsarski, B., and Dimitrov, L.,
Catal. Lett. 62, 35 (1999).