On the effect of deactivation on the kinetics of CO oxidation by NO over Pt-Rh catalysts

Full text of DOI:10.1006/jcat.1998.2086

JOURNAL OF CATALYSIS 177, 147–151 (1998)
ARTICLE NO. CA982086
RESEARCH NOTE
On the Effect of Deactivation on the Kinetics of CO Oxidation
–
by NO over Pt Rh Catalysts
P. Granger,1 L. Delannoy, L. Leclercq, and G. Leclercq
Laboratoire de Catalyse He´te´roge`ne et Homoge`ne, URA CNRS 0402, Baˆt. C3, 59650 Villeneuve d’Ascq Ce´dex, France
Received January 23, 1998; revised March 10, 1998; accepted March 12, 1998
The CO + NO reaction has been extensively studied over
Previous investigations on the kinetics of the CO + NO
a wide variety of catalysts, in particular on noble metals reaction performed in our laboratory at 300ꢁC under reduc-
based catalysts (1, 2). The replacement of platinum and ing or moderate oxidizing conditions over Pt-only (14), Rh-
–
rhodium by less expensive materials exhibiting similar cata- only, and Pt Rh catalysts (15) agree with this mechanism.
lytic properties in three-way catalysts (TWC) is of greatest The reactant partial pressure dependencies of the experi-
importance from an economical viewpoint, consequently, mental reaction rate has been correctly modelled by Eq. [8]
many investigations focused on this practical interest have derived from this mechanism. Previous assumptions such
been reported in the literature (3, 4). However, both met- as: (i) competitive adsorptions of NO and CO on a single
als are still the most active components in TWCs. Further site and, (ii) NO dissociation (step [3]) as rate determining
developments of TWCs seem related both to a better un- have been considered to establish the rate expression
derstanding of elementary surface reactions on atomic scale
k3ꢀNOPNO
(1 + ꢀNOPNO + ꢀCOPCO
where ꢀCO and ꢀNO are respectively the adsorption equilib-
rium constants of CO and NO and k3, the NO dissociation
rate constant.
In the case of Pt Rh/Al2O3 a more relevant kinetic model
has been proposed (15), based on previous infrared obser-
vations dealing with the co-adsorption of CO and NO on
Pt Rh/SiO2. Van Slooten and Nieuwenhuys (16) have sug-
gested that NO is preferentially adsorbed on Rh and CO on
Pt. This alternative mechanism can fit more accurately our
and the existing relationship between surface processes and
r =
,
[8]
2
)
structure and composition of TWCs. In order to clarify
these two fundamental considerations many kinetic and
spectroscopic studies were devoted to the CO + NO reac-
tion. Several mechanisms have been proposed to describe
the CO + NO reaction in various operating conditions
(5–7) because TWCs never operate at steady state. How-
ever, there seems to be presently a general agreement on
the following mechanism to describe the activity and selec-
tivity of TWCs (8–13) in a wide range of operating condi-
tions. This mechanism includes the preliminary dissociation
of NO according to step [3].
–
–
–
kinetic data on Pt Rh/Al2O3, freshly prepared, according
to the rate equation [9]
NO + ꢀ ⇔ NOꢀ
[1]
[2]
[3]
[4]
[5]
[6]
[7]
k3ꢀNOPNO
CO + ꢀ ⇔ COꢀ
r =
.
[9]
(1 + ꢀNO
P
NO)(1 + ꢀCOPCO
)
NOꢀ + ꢀ → Nꢀ + Oꢀ
NOꢀ + Nꢀ → N2 + Oꢀ + ꢀ
NOꢀ + Nꢀ → N2O + 2ꢀ
Nꢀ + Nꢀ → N2 + 2ꢀ
COꢀ + Oꢀ → CO2 + 2ꢀ,
Furthermore the adsorption equilibrium constants of NO
and CO are respectively similar to those obtained on Rh-
only and Pt-only catalysts which is in excellent agreement
with the proposals of Van Slooten and Nieuwenhuys (16). It
was concluded that Pt and Rh preserve their individual ad-
sorption properties on a fresh catalyst. Hence, the average
–
activity of Pt Rh/Al2O3 has been ascribed to a dilution ef-
where ꢀ denotes a vacant adsorption site.
fect by Pt which considerably lowers the surface Rh activity.
There are presently evidences that the surface composi-
tion of bimetallic catalysts is very sensitive to deactivation
whichoccursonTWCsathightemperatureand/orwhilethe
1 To whom correspondence should be addressed. E-mail: catalyse@
univ-lille1.fr.
147
0021-9517/98 $25.00
c
Copyright ꢂ 1998 by Academic Press
All rights of reproduction in any form reserved.

148
GRANGER ET AL.
TABLE 1
–
Bulk and XPS Composition of Fresh and Sintered Bimetallic Pt Rh/Al2O3 Catalysts
XPS analysis
Bulk composition
Surface composition
B.E. (eV)
ꢀ
ꢁ
ꢀ
ꢁ
b
b
a
a
Rh
Rh+Pt
Rh
Rh+Pt
Pt
Rh
Catalysts
Pt 4d3/2
Rh 3d5/2
Al
Al
Fresh
Sintered
0.27
0.27
0.33
0.50
2.1 ꢃ 10ꢄ3
1.0 ꢃ 10ꢄ3
1.0 ꢃ 10ꢄ3
1.0 ꢃ 10ꢄ3
332.1
332.3
307.8
307.0
a Atomic ratios calculated from the intensities of the photopeaks Al 2p from Al2O3, Pt 4d3/2, and Rh 3d5/2
.
b Atomic Rh composition.
feedstream composition varies. Several studies mentioned tial fixed bed flow reactor at 300ꢁC (15) in order to avoid
that deactivation leads to both particle growth and surface external heat and mass transfer limitations. The space ve-
Rh enrichment (17–19).
locity was adjusted to 25000 hꢄ1 during the catalyst testing
–
A sample of a fresh Pt Rh/Al2O3 catalyst was prepared to obtain NO and CO conversions below 14%. It is notice-
according to the same procedure used by Kacimi and able that the thermal sintering does not significantly alter
–
Duprez (20) who found some indications for the formation the specific activity Pt Rh/Al2O3 (see Table 2) despite the
–
of bimetallic Pt Rh particles and then was deactivated by decrease of the metal dispersion by a factor ꢆ2.5. The ex-
submitting it to a thermal treatment at 800ꢁC in a mixture of perimental rates measured at various NO and CO partial
10 vol% H2O in N2 for 16 h. This thermal treatment induced pressures are listed in Table 3.
a slight decrease of the specific area from 95.5 to 71 m2 gꢄ1
.
The reaction orders obtained from these data by means
As seen in Table 1, XPS measurements performed on these of linear regression analysis are ꢄ0.33 and ꢄ0.12 respec-
two catalysts do not reveal any significant changes in the tively for CO and NO within the reactant partial pressure
binding energies of Pt 4d3/2 and Rh 3d5/2 levels. Their com- ranges used in this study (Table 2). Obviously sintering has
parison with the current values reported in the literature a major effect on the formal kinetics of CO oxidation by
(21) shows that Pt and Rh are mainly in their metallic state NO, mainly on the order in NO, but a limited effect on
–
on both catalysts. The results of the quantitative analysis by the global reaction rates (Table 2). Sintered Pt Rh behaves
–
XPS data of the two bimetallic Pt Rh catalysts using a pro- more like Rh/Al2O3 which has been shown to exhibit posi-
cedure already reported in Ref. (22) (Table 1) show clearly tive order in NO in the range of PNO between 0.5 ꢃ 10ꢄ3 and
a Rh enrichment in the layer analyzed by XPS for the sin- 2.0 ꢃ 10ꢄ3 atm, but negative orders for higher PNO (the max-
tered sample mainly due to a decrease in the Pt signal. As imum rate was obtained for PNO = 2.5 ꢃ 10ꢄ3 atm). Conse-
expected, this thermal treatment resulted in a decrease of quently in the PNO range studied here, the apparent order
–
the metal dispersion from 0.64 to 0.27, as determined by hy- in NO is negative for the sintered Pt Rh/Al2O3.
drogen chemisorption. The corresponding average bimetal- The negative order for CO is compatible with both
lic particle sizes are 1.4 and 3.4 nm, respectively, on the fresh Eqs. [8] and [9], but the negative NO order for the sintered
–
and sintered catalysts.
Pt Rh/Al2O3 is inconsistent with Eq. [9].
The kinetics of the CO + NO reaction has been investi-
We have attempted to calculate the rate constant of NO
gated on the sintered Pt Rh/Al2O3 catalyst in a differen- dissociation, k3, and the adsorption equilibrium constants
–
TABLE 2
Rates and Reaction Orders for the CO + NO Reaction at 300ꢁC
PNO
PCO
Specificc,e
activity
Catalyst
(10ꢄ3 atm)
(10ꢄ3 atm)
TOFd,e
1423
na
ma
2.6 ꢃ 10ꢄ2
0.80
<0
0.40
ꢄ0.12
ꢄ0.32
ꢄ0.32
ꢄ0.38
ꢄ0.33
b
0.2 wt% Rh/Al2O3
0.5–2.0
2.5–3.2
1.5–5.6
2.0–8.2
3.0–13.6
3.0–13.6
3.0–8.0
4.5–8.8
b
Fresh 1 wt% Pt-0.2 wt% Rh/Al2O3
Sintered 1 wt% Pt-0.2 wt% Rh/Al2O3
2.1 ꢃ 10ꢄ3
46
101
1.9 ꢃ 10ꢄ3
a Rate = k ꢃ PNn O ꢃ PCmO
.
b From Ref. (15).
c CO mol hꢄ1 gꢄ1 catalyst.
d CO molec ꢅ hꢄ1 (surface metal at)ꢄ1
.
e At T = 300ꢁC with initial PNO = PCO = 5 ꢃ 10ꢄ3 atm.

THE EFFECT OF DEACTIVATION
149
–
Pt Rh/Al2O3 sample, the value of ꢀCO is closer to that ob-
tained on Rh/Al2O3 than to that on the fresh one, which had
been shown to be approximately the same as that obtained
on Pt/Al2O3 (15).
TABLE 3
Influence of the Partial Pressures of NO and CO on the Rate
of CO Oxidation by NO on Sintered 1 wt% Pt-0.2 wt% Rh/Al2O3
(T = 300ꢁC, Space Velocity = 25000 hꢄ1
)
–
It may seem surprising that for the fresh Pt Rh/Al2O3
PNO, atm
PCO, atm
ra, mol gꢄ1 hꢄ1
the model with noncompetitive adsorptions is favored (see
Ref. (15) for arguments) while for the sintered sample only
competitive adsorptions are consistent with our experimen-
tal results. This apparent contradiction could be removed
if sintering strongly modifies the surface of the metal par-
ticles and transforms the metal surface where both Pt and
Rh coexist together in a metal surface mainly composed of
Rh atoms.
2.0 ꢃ 10ꢄ3
4.2 ꢃ 10ꢄ3
4.4 ꢃ 10ꢄ3
7.4 ꢃ 10ꢄ3
8.2 ꢃ 10ꢄ3
4.3 ꢃ 10ꢄ3
4.3 ꢃ 10ꢄ3
4.3 ꢃ 10ꢄ3
4.5 ꢃ 10ꢄ3
4.5 ꢃ 10ꢄ3
4.4 ꢃ 10ꢄ3
4.5 ꢃ 10ꢄ3
4.6 ꢃ 10ꢄ3
4.7 ꢃ 10ꢄ3
4.7 ꢃ 10ꢄ3
4.7 ꢃ 10ꢄ3
2.6 ꢃ 10ꢄ3
4.7 ꢃ 10ꢄ3
4.7 ꢃ 10ꢄ3
5.8 ꢃ 10ꢄ3
7.7 ꢃ 10ꢄ3
8.8 ꢃ 10ꢄ3
(2.4 ꢇ 0.35) ꢃ 10ꢄ3
(1.9 ꢇ 0.3) ꢃ 10ꢄ3
(2.0 ꢇ 0.3) ꢃ 10ꢄ3
(1.95 ꢇ 0.3) ꢃ 10ꢄ3
(2.0 ꢇ 0.35) ꢃ 10ꢄ3
(2.1 ꢇ 0.35) ꢃ 10ꢄ3
(1.9 ꢇ 0.3) ꢃ 10ꢄ3
(1.9 ꢇ 0.3) ꢃ 10ꢄ3
(1.9 ꢇ 0.3) ꢃ 10ꢄ3
(1.6 ꢇ 0.2) ꢃ 10ꢄ3
(1.6 ꢇ 0.2) ꢃ 10ꢄ3
A comparative infrared study of CO adsorbed on
–
Pt Rh/Al2O3 freshly prepared and after thermal sintering,
reveals modifications in the IR band shape of adsorbed CO
species. Figure 1 shows the IR spectra recorded on pre-
a Rate of CO oxidation by NO, expressed per gram of catalyst, calcu-
lated from the CO conversion, TCO according to the equation: TCO
TN +TN O/2, where TN and TN O are respectively the conversion of NO
=
–
reduced Pt Rh/Al2O3 samples obtained after contact with
CO at 25ꢁC for 1 h and outgassing at the same tempera-
ture. The assignments of IR bands have been carried out
from previous IR studies on Pt-only and Rh-only catalysts
(23). For Pt/Al2O3, a major band and a less intense one
were detected at 2070 and 1850 cmꢄ1 which characterised
respectively linear and bridged CO species on Pt metal.
These corresponding IR bands were located at 2060 and
1890 cmꢄ1 on Rh/Al2O3; in addition two bands appeared at
2030–2040 and 2100 cmꢄ1 due to rhodium gem-dicarbonyl
species, Rh+(CO)2. All these IR observations are in agree-
ment with IR studies already published in the literature
2
2
2
2
into N2 and N2O.
of NO and CO (ꢀCO and ꢀNO) by solving Eqs. [8] and [9], us-
ing the graphic and optimisation methods earlier described
in Ref. (14). The accuracy on the calculated parameters is
about ꢇ20%. Only Eq. [8] leads to positive values for k3,
ꢀNO, and ꢀCO (Table 4) from the graphic method and con-
verges towards a unique set of values using the optimisation
method. In addition, comparable results are obtained from
these two methods which allow us to conclude that only
competitive adsorptions of CO and NO enable us to model
the reactant partial pressure dependencies of the reaction
rate. Comparison of ꢀNO on fresh and sintered bimetallic
–
(24–27). The spectrum on the fresh Pt Rh/Al2O3 sample
(Fig. 1.A) is dominated by an IR band at 2070 cmꢄ1. A
broad band with a weak maximum at 1840 cmꢄ1 is also ob-
served which indicates the formation of respectively linear
and bridged CO species mainly on Pt metal. On the sin-
tered sample (Fig. 1.B), the IR band corresponding to the
–
Pt Rh/Al2O3 shows no significant change in ꢀNO after age-
ing, it remains close to that obtained on Rh/Al2O3, accord-
ing to the margin of error. On the contrary, on the sintered
TABLE 4
Kinetic and Thermodynamic Constants for the Oxidation of CO by NO on Rh and Pt Based Catalysts at 300ꢁC
Rate
equation
Slope
ꢁ
Intercept
ꢀNO
atmꢄ1
ꢀCO
atmꢄ1
c
b
Catalyst
ꢂ
k3
k30
225 ꢃ 10ꢄ3
4.74 ꢃ 10ꢄ3
13.2 ꢃ 10ꢄ3
11.9 ꢃ 10ꢄ3 h
12315
469
472
505
71
122
97
a
0.2 wt% Rh/Al2O3
[8]
[9]
[8]
a
Fresh 1 wt% Pt-0.2 wt% Rh/Al2O3
Sintered 1 wt% Pt-0.2 wt% Rh/Al2O3
42.2d/174.5e
144.7f/523.6g
1.14d/0.65e
691
402
623h
448h
85h
93
[9]
1.55f/ꢄ0.15g
ꢄ3491
a From Ref. (15).
b Specific rate constant for the dissociation of NO (mol hꢄ1 g1).
c Intrinsic rate constant for the dissociation of NO (molec ꢅ hꢄ1 metal atꢄ1).
d From the linear plot (PNO/r)0.5 vs PCO
.
e From the linear plot (PNO/r)0.5 vs PNO
f From the linear plot PNO/r vs PCO
g From the linear plot PNO/r vs PNO
h Calculated from the optimisation method.
.
.
.

150
GRANGER ET AL.
several explanations. First, a true rhodium surface segre-
gation could be considered if bimetallic particles exist at
–
the surface of Pt Rh/Al2O3. For the fresh catalyst, with a
metal dispersion of 0.64, the global composition (27.5 at%
Rh) clearly indicates that the metal surface must contain
a majority of Pt atoms (at least 58%), even if all the Rh
atoms were segregated at the surface. On the contrary, for
the sintered sample, its lower metal dispersion (0.27) would
allow a metal surface mainly composed of Rh atoms if these
atoms were segregated at the surface of metal particles. Of
course such a total segregation would be a borderline case
and, if such an explanation is correct, the segregation is
likely to be incomplete. Moreover some separated Pt and
Rh particles could be present. Alternatively, if, on the fresh
catalyst, bimetallic particles were not formed, different de-
activation processes for Pt and Rh could occur. For example
the metal Pt particles could sinter to large particles, whereas
the Rh particles could sinter less, leading to a global catalyst
exhibiting a Rh-like behaviour. Let us note that this latter
explanation could still hold even if bimetallic particles are
present in the fresh catalyst if sintering in the presence of
water leads to dealloying of Pt and Rh atoms.
Going further, a good comparison of the catalytic pro-
perties of these two catalysts can only be achieved from
the examination of the intrinsic NO dissociation rate con-
stant, k03. Let us notice the low value of k30 on the sintered
–
Pt Rh/Al2O3 which is unexpected since it should be com-
parable to that of Rh/Al2O3, whereas, in fact, k03 is approx-
FIG. 1. Infrared spectra of CO adsorbed on a fresh bimetallic
–
–
–
imately 20 times lower on sintered Pt Rh/Al2O3 than on
Pt Rh/Al2O3, A and on a sintered Pt Rh/Al2O3 catalyst, B. The samples
were reduced “in situ” in flowing hydrogen at 500ꢁC, then outgassed under
vacuum at 350ꢁC. At the end of this treatment the cell was cooled at room
temperature before introducing 1.6 ꢃ 10ꢄ2 atm CO. The catalysts were
maintained in these operating conditions for 1 h before CO evacuation.
Rh/Al2O3. Such a variation in k03 could arise from the struc-
ture sensitivity of the dissociation of adsorbed NO that was
already mentioned by Peden et al. on Rh (111) and (110)
single crystals (30).
To conclude, both the physicochemical characterization
and the kinetic study of the CO + NO reaction evidenced a
linear carbonyl species significantly decreases in intensity
and is shifted towards lower frequencies (2060 cmꢄ1) while
the CO-bridged band is shifted towards higher frequencies
(1890 cmꢄ1). Moreover, the IR bands at 2025 and 2095 cmꢄ1
corresponding to the gem-dicarbonyl species clearly ap-
pear, but they remain less intense than that corresponding
to the linear CO species. All these absorptions are charac-
teristic of CO adsorbed on Rh. On this metal, the relative
intensities of the gem-linear bands have been shown to be
a function of the metal dispersion (27, 28). The results re-
ported here suggest that Rh is not in very small particles in
the sintered sample. As a matter of fact, the spectrum ob-
tained resembles very closely that reported by Barbier et al.
(29) on Rh/Al2O3 with medium metal dispersion (0.48). Fi-
nally, all these changes in both location and intensity of IR
bandssuggestthatCOadsorbsmainlyonRhonthesintered
–
major surface composition modification of a Pt Rh/Al2O3
catalyst after thermal sintering in the presence of water.
Such a treatment seems to lead to a Rh enrichment of the
surface of the metal particles. This shows that kinetic stud-
ies can provide relevant information by deriving a rate ex-
pression and calculating parameters which can further be
used, not only in the field of kinetic modelling, but also as a
tool for the characterization of the surface composition of
bimetallic catalysts.
ACKNOWLEDGMENTS
We thank Drs. G. Mabilon and M. Prigent for their contribution to
this work and also Dr. H. Praliaud for the interpretation of the infrared
measurements.
–
Pt Rh/Al2O3 catalyst.
Consequently both kinetic and IR results seem to point
out a Rh-like behaviour of the sintered Pt Rh/Al2O3 cata-
REFERENCES
–
lyst, in contrast to the fresh sample at the surface of which
both Pt and Rh coexist. Such a phenomenon could have
1. Shelef, M., and Graham, G. W., Catal. Rev. Sci. Eng. 36, 433 (1994).
2. Taylor, K. C., Catal. Rev. Sci. Eng. 35, 144 (1993).

THE EFFECT OF DEACTIVATION
151
3. Adams, K. M., and Gandhi, H. S., Ind. Eng. Chem. Prod. Res. Dev. 17b. Kim, S., and D’Aniello, M. J., Jr., Appl. Catal. 56, 45 (1989).
22, 207 (1983).
18. Powell, B. R., and Chen, Y. L., Appl. Catal. 53, 233 (1989).
4. Gandhi, H. S., Yao, H. C., and Stepien, H. K., in “ACS Sympos. Ser.,”
Vol. 178 (A. T. Bell and L. Hegedus, Eds.), p. 143. Am. Chem. Soc.,
Washington, DC, 1982.
5. Klein, R. L., Shwartz, S. B., and Schmidt, L. D., J. Phys. Chem. 89,
4908 (1985).
6. Aldoch, W., and Linz, H. G., Surf. Sci. 78, 69 (1978).
7. Dubois, L. H., Hansma, P. K., and Somorjai, G. A., J. Catal. 65, 318
(1980).
8. Masel, R. I., Catal. Rev. Sci. Eng. 28, 335 (1986).
9. Hecker, W. C., and Bell, A. T., J. Catal. 85, 389 (1984).
10. Hecker, W. C., and Bell, A. T., J. Catal. 59, 223 (1979).
11. Cho, B. K., Shank, B. H., and Bailey, J. E., J. Catal. 115, 486 (1989).
12. Oh, S. E., Fisher, G. B., Carpenter, J. E., and Goodman, D. W.,
J. Catal. 100, 360 (1986).
19. Malm, J. O., and Bovin, J. O., Microanal. Microstruct. 1, 387 (1990).
20. Kacimi, S., and Duprez, D., in “Catalysis and Automotive Pollution
Control II” (A. Crucq, Ed.), Vol. 71, p. 581. Elsevier, Amsterdam,
1991.
21. Wagner, C. D., Riggs, W. M., Davis, L. E., and Moulder, J. F., “Hand
Book of X-Ray Photoelectron Spectroscopy” (G. E. Muilenberg,
Ed.), Perkin Elmer, Palo Alto, CA, 1979.
22. Grimblot, J., in “Analyse des surfaces des solides,” p. 125. Masson,
Paris, 1995.
23. Unpublished results obtained at the Institut de Recherche sur la
Catalyse.
24. Anderson, J. A., and Rochester, C. H., Catal. Today 10, 275 (1991).
25. Alikina, G. M., Davidov, A. A., Sazonova, I. S., and Popovskii, V. V.,
React. Kinet. Catal. Lett. 27, 279 (1985).
13. Ng, K. Y. S., Belton, D. N., Schmieg, S. J., and Fisher, G. B., J. Catal.
146, 394 (1994).
14. Granger, P., Dathy, C., Lecomte, J. J., Prigent, M., Mabilon, G.,
Leclercq, L., and Leclercq, G., J. Catal. 173, 304 (1998).
15. Granger, P., Dathy, C., Lecomte, J. J., Leclercq, L., and Leclercq, G.,
J. Catal., in press.
26. Yang, A. C., and Garland, C. W., J. Phys. Chem. 61, 1504 (1957).
27. Yates, D. J. C., Murel, L. L., and Prestridge, E. B., J. Catal. 84, 41
(1979).
28. Solymosi, P., J. Phys. Chem. 89, 4789 (1985).
29. Barbier, J., Dumas, J. M., Ge´ron, C., and Hadrane, H., Appl. Catal.
67, L1 (1990).
16. Van Slooten, R. F., and Nieuwenhuys, B. E., J. Catal. 122, 429 (1990).
17a. Kim, S., and D’Aniello, M. J., Jr., Appl. Catal. 56, 23 (1989).
30. Peden, C. H. F., Belton, D. N., and Schmieg, S. J., J. Catal. 155, 204
(1995).