An investigation of promoter effects in the reduction of NO by H2 under lean-burn conditions

Article abstract of DOI:10.1006/jcat.2002.3596

The reduction of NO by H₂ has been investigated under lean conditions at temperatures representative of automotive cold-start conditions (<200°C) using MoO₃- and Na₂O-modified Pt/Al₂O₃ and Pt/SiO₂ catalysts. It has been found that small additions of sodium significantly increase the NO conversion while larger loadings of sodium have a severe poisoning effect. However, in the presence of excess O₂ no enhancement in nitrogen selectivity at low temperatures has been observed for all loadings of Na. Indeed, an adverse effect has been found at higher temperatures. Addition of molybdenum as a promoter results in increases in NO conversion and nitrogen selectivity for all loadings tested. The optimal formulation was determined to be 1% Pt/10% MoO₃/0.27% Na₂O/Al₂O₃. Steady-state isotopic-transient kinetic (SSITK) experiments were performed on this and a model Pt/MoO₃/Na₂O/SiO₂ catalyst using labelled nitric oxide in order to estimate the surface concentrations of species leading to N₂, N₂O, and retained NO. The data reveal significantly greater surface concentrations of N₂ precursors over the modified catalysts for both the Pt/Al₂O₃ and Pt/SiO₂ systems and this has been used to rationalise the increased selectivity to N₂. An additional significant effect of the molybdenum promoter has been proposed because when the concentrations of N₂ intermediates and the amount of available platinum in the modified catalysts are calculated, the storage of N₂ precursors on the MoO₃ seems to occur. This effect has been further explored using the modified SiO₂ catalyst in non-steady-state transient experiments where the reductant supply (i.e., the H₂) is cut off. It has been found that the decay in the production of N₂ is very significantly delayed in comparison with the unmodified catalysts, and it is proposed that this is consistent with the trapping on the Mo of a reduced intermediate that can form N₂ even in the absence of the normal H₂ reductant. The mechanistic consequences of these novel results are discussed.

Full text of DOI:10.1006/jcat.2002.3596

Journal of Catalysis 208, 435–447 (2002)
doi:10.1006/jcat.2002.3596
An Investigation of Promoter Effects in the Reduction of NO
by H under Lean-Burn Conditions
2
1
R. Burch and M. D. Coleman
School of Chemistry, Queen’s University Belfast, David Keir Building, Stranmillis Road, Belfast, BT9 5AG, N. Ireland
Received December 12, 2001; revised February 15, 2002; accepted February 15, 2002
resulted in increasingly stringent legislation (4). Whilst
The reduction of NO by H has been investigated under lean con- three-way catalysts (TWCs) have been successfully used
2
ditions at temperatures representative of automotive “cold-start” for several decades with engines that operate close to sto-
◦
conditions (<200 C) using MoO - and Na O-modified Pt/Al O
3
2
2
3
ichiometric the current technology cannot remove NO at
low temperatures or under lean conditions (5–7). As a re-
x
and Pt/SiO catalysts. It has been found that small additions of
sodium significantly increase the NO conversion while larger load-
ings of sodium have a severe poisoning effect. However, in the pres-
2
sult, there is a need to find catalytic systems able to operate
under these conditions (8). To this end there has been much
interest in using the hydrocarbons present either in the on-
board fuel or in the exhaust itself to carry out the reduction.
A variety of hydrocarbons, ranging from methane (9–
ence of excess O no enhancement in nitrogen selectivity at low
2
temperatures has been observed for all loadings of Na. Indeed, an
adverse effect has been found at higher temperatures. Addition of
molybdenum as a promoter results in increases in NO conversion
and nitrogen selectivity for all loadings tested. The optimal formu-
13), ethene (14–16), and propane/-ene (17–22) to larger
lation was determined to be 1% Pt/10% MoO /0.27% Na O/Al O . molecules (23–27), have been shown to selectively react
3
2
2 3
Steady-state isotopic-transient kinetic (SSITK) experiments were with NOx in the presence of excess oxygen but the lack
◦
performed on this and a “model” Pt/MoO /Na O/SiO catalyst us- of any significant activity below ∼200 C remains a seri-
3
2
2
ing labelled nitric oxide in order to estimate the surface concen-
trations of species leading to N , N O, and retained NO. The data
ous problem. Ammonia is a possible alternative and is of-
ten used as the reductant in stationary power sources (28).
However, due to ammonia slip and the lack of a distribu-
tion network it is still desirable to find a way to remove NOx
using the normal on-board fuel.
Hydrogen may offer a solution to these problems since
it can be generated on-board from the fuel and is active at
low temperatures. The NO/H2 reaction has been well in-
2
2
reveal significantly greater surface concentrations of N precursors
2
over the modified catalysts for both the Pt/Al O and Pt/SiO sys-
2
3
2
tems and this has been used to rationalise the increased selectivity
to N . An additional significant effect of the molybdenum promoter
2
has been proposed because when the concentrations of N interme-
2
diates and the amount of available platinum in the modified cata-
lysts are calculated, the “storage” of N precursors on the MoO
2
3
seems to occur. This effect has been further explored using the mod- vestigated over the past few decades (29–34). In the 1970s
ified SiO catalyst in non-steady-state transient experiments where Kobylinski and Taylor screened a number of supported
2
the reductant supply (i.e., the H ) is cut off. It has been found that noble metals for activity and found both palladium and
2
the decay in the production of N is very significantly delayed in
comparison with the unmodified catalysts, and it is proposed that
this is consistent with the trapping on the Mo of a reduced inter-
2
platinum to be particularly active (35). Furthermore, they
also reported that N2 and NH3 were the main reduction
products. More recently, both Harkness and Lambert, and
Marina et al. have reported that sodium can be used to
increase NO conversion and N2 selectivity over platinum-
based catalysts (36, 37). These authors proposed that these
observations could be explained by considering that since
sodium is electropositive the chemisorption strength of an
electron acceptor adsorbate (i.e., NO) would be increased.
It was further surmised that the increased strength of the
Pt–NO interaction caused a weakening of the N–O bond,
leading to increased dissociation. Sodium addition has also
been found to promote reaction characteristics over cata-
lysts based on platinum, palladium, and rhodium using both
CO and hydrocarbon reductants (38–47). However, when
using methane as a reductant sodium addition has been
mediate that can form N even in the absence of the normal H
2
2
reductant. The mechanistic consequences of these novel results are
discussed. ꢀc 2002 Elsevier Science (USA)
Key Words: SSITKA; NO/H ; lean burn; platinum; molybdenum
2
promoter.
INTRODUCTION
The adverse health and environmental effects caused
through NOx (defined as NO+NO2) release from both mo-
bile and stationary sources are well-known (1–3) and has
1
To whom correspondence should be addressed.
435
0
021-9517/02 $35.00
c 2002 Elsevier Science (USA)
All rights reserved.
ꢀ

4
36
BURCH AND COLEMAN
found to poison the activity of a palladium catalyst (48). In e.g., 1% Pt/10% MoO3/0.27% Na2O/Al2O3 is written as
this case it was proposed that since methane is a weaker ad- Pt/10Mo/0.27Na.
sorbate than NO (in contrast to the NO/propene system),
promoting the adsorption of NO by using sodium is unde- Characterisation
sirable since it results in oxygen poisoning of the catalyst.
BET surface areas were measured using a Micromerit-
Halasz et al. have reported the promotion of PdO/Al2O3
ics ASAP 2010 system. The SiO2 had a surface area of
withMoO3 (49, 50). However, whiletheactivitytowardsthe
2
−1
2
−1
2
83 m g and the Al2O3 had a surface area of 180 m g .
NO/H2 reaction was found to increase as a consequence of
this modification, no effect was observed under lean-burn
conditions (51).
With regard to investigations of the NO/H2 reaction in
an excess of oxygen there is a relative dearth of litera-
ture (52–55). In a previous paper we demonstrated that
hydrogen could be used to selectively reduce NO under
Pt surface areas were determined using pulses of H2 in a
Micromeritics AutoChem 2910 system—details are given
under the Results section.
Kinetic Measurements
The catalysts (typically 100 mg) were loaded into a quartz
lean-burn conditions over platinum-based catalysts (56). reactor and held in place with quartz wool plugs. A thermo-
However, in common with many such systems using a couple was located in the centre of the catalyst bed and the
hydrocarbon reductant (57) it was found that significant reactor was connected to a stainless steel gas handling sys-
amounts of undesirable N2O were produced in addition to tem fitted with Area mass flow controllers for the produc-
tion of the various gaseous reaction mixtures. Using He as
N2 (no NH3 was seen). Tanaka et al. have modified Pt/SiO2
with molybdenum and sodium oxide promoters, and for the
reduction of NO with C3H6, H2, and CO they have found in-
creasesinthenitrogenselectivity(58–60). Theyrationalised
the results on the basis that the enhanced selectivity was due
to the suppression of the oxidation of the platinum under
lean conditions.
In this paper we present the results of a kinetic investi-
gation into the activity and nitrogen selectivity of modified
Pt/Al2O3 catalysts as a function of MoO3 and Na2O load-
ings using H2 as the reductant under lean-burn conditions
in relation to cold-start or low-temperature operation. On
the basis of steady-state isotopic-transient kinetic analy-
sis (SSITKA) (61) for an optimised Pt/MoO3/Na2O/Al2O3
and a “model” Pt/MoO3/Na2O/SiO2 catalyst, we provide
a rationalisation for the mode of operation of these oxide
promoters.
3
−1
the makeup gas, the total flow rate was set at 200 cm min
,
3
−1 −1
corresponding to 120,000 cm g
the reactor was analysed by an online gas chromatograph
Perkin–Elmer Autosystem XL, with Nelson software) fit-
h . The effluent from
(
ted with Heysep-n and 13× molecular sieve columns, and
by a Signal chemiluminscence NOx analyser. The large ex-
cess of O2 made separate analysis of N2 problematical so,
in general, this was determined by difference since mass
spectrometric analysis had shown that NH3 was not formed
under lean conditions:
[
NOx in feed] − ([NOx in effluent] + 2[N2O in effluent])
[N2] =
.
2
The percentage NO conversion was calculated from
[
NO in feed] − [NO in effluent]
NOconv =
× 100.
[
NO in feed]
EXPERIMENTAL
The percentage N2 selectivity was calculated from
Catalyst Preparation
Supported Pt catalysts (nominal composition, 1 wt%)
were prepared by incipient wetness using aqueous so-
lutions of dinitrodiamine platinum(II) with acid-washed
silica (Grace 432) or alumina (Akzo CK300, supplied by
[
N2 in effluent]
SN =
× 100.
2
[
N2 in effluent] + [N2O in effluent]
Criterion) that had been precalcined overnight at 500 and
Transient kinetic studies, both using simple gas-switching
◦
7
00 C, respectively. The prepared catalysts were dried at techniques and steady-state isotopic-transient kinetic
◦
◦
1
20 C overnight and then calcined at 500 C for 2 h. In addi- (SSITK) methods, were undertaken using a specially de-
tion, the oxide supports were modified by addition of MoO3 signed, very low dead-volume apparatus with accurate con-
using ammonium heptamolybdate) and/or Na2O (using trolofpressures(monitoredbyhigh-precisionWIKAtrans-
(
sodium acetate) promoters, and also by incipient wetness, ducers) on either side of a fast-switching valve (Valco).
except in the case of very high MoO3 loadings where the Prior to a switch, whether using isotopically labelled gases
solubility limit of the ammonium heptamolybdate required or not, the flows and pressures were carefully balanced so
◦
excess solution to be used. After drying overnight at 120 C that when the switch was made there would be no upset to
◦
and calcinations for 2 h at 500 C, the Pt was added as be- the gas flow that could lead to erroneous readings on the
fore. For convenience, the catalysts are labelled as follows: mass spectrometer (VG Gaslab 300 Mass Spectrometer).

H2 NO REDUCTION UNDER LEAN-BURN CONDITIONS
437
RESULTS
Steady-State Experiments
1
00
9
8
0
0
ThesurfaceareasofthePtinthevariouscatalysts, asmea-
sured using the H2 pulse method, are given in Table 1. For
the alumina-supported catalysts, addition of Na appears ini-
tially to increase the exposed Pt surface area, but at higher
loading the uptake of H2 is significantly suppressed. Ad-
dition of Mo oxide has a clear effect and severely reduces
the amount of H2 chemisorbed. Surprisingly, with the silica-
supported catalyst even the addition of a very large amount
of Mo oxide has no significant effect on the H2 chemisorp-
tion. The calculated Pt dispersion ranges from 54 to 2.2%,
although it must be understood that partial coverage of the
PtparticlesbyNaand/orMooxidecouldreducetheamount
of H2 chemisorbed and cause the Pt dispersion to appear
to decrease.
70
6
5
4
3
0
0
0
0
The NO conversion as a function of temperature for the
Pt/Al2O3 and Na-modified catalysts is shown in Fig. 1. The
base catalyst has a good activity for the reduction of NO
under these lean-burn conditions with a conversion peak
20
◦
of 75% at 125 C. Addition of the smallest amount of Na
1
0
0
(
0.27%) results in a significant increase in the level of NO
reduction with a peak now at almost 90% conversion and
a much improved performance on the higher temperature
side of the maximum. However, further increases in the Na
loading first result in a loss of activity at the lower temper-
atures and then in a dramatic loss of activity at all tempera-
tures when the Na content is increased to 5 and then 10%.
4
0
90
140
190
240
Catalyst temperature / °C
FIG. 1. NO conversion as a function of temperature for various
Even allowing for the loss of Pt area, as measured by H2 sodium-modified catalysts. ꢀ, Pt/Al2O3; ꢁ, Pt/0.27Na/Al2O3; ꢂ, Pt/1Na/
Al2O3; ꢃ, Pt/5Na/Al2O3; ×, Pt/10Na/Al2O3. Reaction conditions:
chemisorption (see Table 1), there is a clear loss of activity
when high loadings of Na are used.
3
−1
1
000 ppm NO, 4000 ppm H2, 6% O2, 200 cm min total flow, 100 mg
of catalyst.
The H2 conversion (results not shown) corresponding
to the NO reduction profiles shown in Fig. 1 reflects the
activity of the catalysts. Thus, as the Na loading is in-
creased, the temperature for 20% conversion of H2 varies
with Fig. 1 shows is close to the temperature at which the
NO reduction lights off.
◦
◦
◦
as follows: Pt (100 C), Pt/0.27Na (100 C), Pt/1Na (140 C),
In the reduction of NO, the selectivity to N2 rather than
N2O is important since N2O is itself a pollutant. It has been
shown that under stoichiometric (or rich) conditions the
addition of sodium can markedly increase the selectivity to
N2 rather than to N2O (36–47). However, Fig. 2 shows that
under lean-burn conditions there is no change in the selec-
◦
◦
Pt/5Na (150 C), and Pt/10Na (>190 C), which comparison
TABLE 1
Pt Surface Areas as Determined Using the H Pulse Method
2
H2 uptake
Pt surface area Pt dispersion
◦
tivity in the temperature range from 90 to 140 C. Indeed,
−
1
2
−1
Catalyst
Pt/Al2O3
Pt/0.27Na/Al2O3
Pt/1Na/Al2O3
Pt/5Na/Al2O3
Pt/10Na/Al2O3
Pt/2.5Mo/0.27Na/Al2O3
Pt/10Mo/0.27Na/Al2O3
Pt/23Mo/0.27Na/Al2O3
Pt/46Mo/0.27Na/Al2O3
Pt/SiO2
(µmol g
)
(m g )
(%)
at higher temperatures, the selectivity to N2 is actually de-
creased somewhat when sodium is added. Clearly, there are
differences in the performance of our catalysts under lean-
burn conditions and those reported in the literature tested
under stoichiometric conditions.
Figure 1 has shown that the optimum effect of sodium on
activity under our lean-burn conditions is found with the ad-
dition of only 0.27% Na. Therefore, this was selected as the
reference material in order to examine the effect of adding
molybdenum. Figure 3 shows that addition of molybdenum
has a significant effect on the activity for NO reduction.
10.8
13.9
13.5
1.04
1.34
1.30
0.93
0.46
0.75
0.32
0.46
0.05
0.51
0.50
42.0
54.2
52.8
37.5
18.4
30.4
12.7
18.7
2.2
9.60
4.73
7.81
3.26
4.78
0.54
5.31
5.13
20.7
20.1
Pt/46Mo/0.27Na/SiO2

4
38
BURCH AND COLEMAN
1
00
molybdenum to the alumina-supported Pt catalyst, we first
investigated the NO/H2 reaction in the absence of oxygen.
One reason is that under oxidising conditions the alumina
support has a high affinity for adsorbing NO2 as a surface ni-
trateandthiscompletelymasksanytransientisotopeeffects
that might be occurring on the metal. Therefore, we tried to
circumvent this problem (see later) by first confirming that
the promoted alumina-supported catalysts do indeed show
differences even under oxygen-free conditions, and then by
investigating a model silica-supported catalyst that does not
have this NO2 adsorption problem.
9
8
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
0
Figure 5 shows the transient curves for the isotopic
1
4
switching experiment, in which NO is replaced at steady
1
5
◦
state with NO, using the Pt/Al2O3 catalyst at 70 C. The
1
4
15
N NO profile shows a small maximum (masked by other
profiles on figure) a very short time after the switch and then
declines rapidly to zero, indicating that the formation of this
mixedisotopeproductrequiresthepresenceofgaseousNO.
1
5
Similarly, the N2O profile rises rapidly to a constant value,
indicating that N2O precursors have a short lifetime on the
catalyst surface prior to desorption. On the other hand, the
100
4
0
90
140
190
240
9
8
7
6
0
0
0
0
Catalyst temperature / °C
FIG. 2. N2 selectivity as a function of temperature for various sodium-
modified catalysts. ꢀ, Pt/Al2O3; ꢁ, Pt/0.27Na/Al2O3; ꢂ, Pt/1Na/Al2O3; ꢃ,
Pt/5Na/Al2O3. Reaction conditions: 1000 ppm NO, 4000 ppm H2, 6% O2,
00 cm min total flow, 100 mg of catalyst.
3
−1
2
In all cases, the performance at lower temperatures is en-
hanced, and in the specific cases where 2.5 or 10% MoO3
has been added, the conversion of NO exhibits a maximum
that exceeds 90%. The 10% MoO3 catalyst also shows im-
proved performance at the higher temperatures where the
50
◦
conversion stays above 70%, up to at least 180 C. The con-
version of H2 (results not shown) parallels the NO reduc-
4
0
tion profiles, with the more active catalysts lighting off some
◦
30
20
2
0–30 C lower than for the unpromoted Pt catalyst.
TheproductselectivitywiththeseMo-promotedcatalysts
is shown in Fig. 4. In this case, there is a significant improve-
ment in the selectivity to N2 rather than to N2O. For the
most active catalysts, the selectivity to N2 has increased by
around 30% in the middle range of temperatures. In order
to try to understand the reason for the improved selectiv-
ity of the molybdenum-promoted catalysts under lean-burn
conditions, we have performed a comprehensive set of tran-
sient kinetic experiments that are now described.
1
0
0
50
100
150
200
Catalyst temperature / °C
Transient Kinetic Experiments on Alumina-Supported
Catalysts under Rich Conditions
FIG. 3. NO conversion as a function of temperature for various
sodium- and molybdenum-modified catalysts. ꢀ, Pt/Al2O3; ꢁ, Pt/2.5Mo/
0
.27Na/Al2O3; ꢃ, Pt/10Mo/0.27Na/Al2O3; ꢂ, Pt/23Mo/0.27Na/Al2O3; ×,
In order to further probe the nature of the changes in-
duced in the Pt catalysts by the addition of sodium and/or 6% O2, 200 cm min total flow, 100 mg of catalyst.
Pt/46Mo/0.27Na/Al2O3. Reaction conditions: 1000 ppm NO, 4000 ppm H2,
3
−1

H2 NO REDUCTION UNDER LEAN-BURN CONDITIONS
439
1
00
1.2
1
5
90
80
70
60
50
40
30
20
10
0
NO
1
0
0
0
0
0
.0
.8
.6
.4
.2
.0
1
5
N O
2
1
5
N
2
1
4
N
2
1
5
14
N N
1
4
15
N NO
1
4
NO
14
2
N O
-
500
0
500
Time / s
1000
1500
5
0
100
150
200
Catalyst temperature / °C
FIG. 5. Concentration profiles as a function of time after a SSITK
1
5
14
switch of NO for NO for a Pt/Al2O3 catalyst. Switch from (0.76%
FIG. 4. N2 selectivity as a function of temperature for various
sodium- and molybdenum-modified catalysts. ꢀ, Pt/Al2O3; ꢁ, Pt/2.5Mo/
14
15
◦
NO/Ar/1.08% H2) to (0.76% NO/He/1.08% H2) at 70 C for 100 mg of
3
−1
.
catalyst at a total gas flow rate of 100 cm min
0
.27Na/Al2O3; ꢃ, Pt/10Mo/0.27Na/Al2O3; ꢂ, Pt/23Mo/0.27Na/Al2O3; ×,
Pt/46Mo/0.27Na/Al2O3. Reaction conditions: 1000 ppm NO, 4000 ppm H2,
3
−1
6
% O2, 200 cm min total flow, 100 mg of catalyst.
further enhanced, with the NO retention being longer and
1
4
15
the N N decay being slower again.
Using standard SSITK curve analysis procedures (61) we
estimated the concentration of intermediates on the sur-
face of each catalyst that leads to the different nitrogen-
containing products. These are summarised in Table 2. It
1
4
N2 profile decays slowly to zero over a period of about
1
4
15
1
500 s, as also does the N N profile, indicating that these
14
products are formed from N that is retained on the cata-
lyst surface for a long period of time.
Figure 6 shows the corresponding results for the sodium-
modified catalyst (Pt/0.27Na). The results mainly differ in
TABLE 2
14
15
−1
the nitrous oxide profiles. In this case, both the N NO
Concentrations of Adsorbed Species (µmol g ) Leading to N2
1
4
and the N2O curves take longer to decay, and this seems and N2O Products and to NO Desorption as Estimated by Analysis
14
to parallel the slower displacement of NO from the system of the SSITK Profiles Shown in Figs. 5–12
(
compare Figs. 5 and 6, and see also the quantitative data
Catalyst
C(N2)
C(N2O) C(NO)
ꢀC
14
later in Table 2). On the other hand, the profiles for N2 and
1
4
15
N N are rather similar for both the Pt and the Pt/0.27Na Pt/Al2O3
15.0 ± 0.2
8.3 ± 0.5 27.4 ± 4.1 50.7 ± 4.8
Pt/0.27Na/Al2O3
Pt/10Mo/Al2O3
Pt/10Mo/0.27Na/Al2O3
Pt/SiO2
17.4 ± 0.3 17.2 ± 0.8 75.1 ± 3.6 109.7 ± 4.7
catalysts.
Much more significant differences in the profiles are seen
in Fig. 7 for the Pt/10Mo catalyst. In this case the N NO
and the N2O curves decay significantly faster than the NO Pt/46Mo/0.27Na/SiO2
curve whereas the N N curve shows a peak maximum Pt/SiO2
at a later time than before and, moreover, does not even
decay to zero even after 1500 s. Figure 8 shows that for the
doubly promoted Pt/10Mo/0.27Na catalyst these effects are
34.4 ± 6.9
56.0 ± 9.3
3.6 ± 0.2
3.4 ± 0.6 63.8 ± 3.8 101.6 ± 11.3
3.2 ± 0.6 95.7 ± 3.8 154.9 ± 13.7
3.5 ± 0.5 4.9 ± 4.2 12.0 ± 4.9
1.6 ± 0.6 19.9 ± 3.8 147.8 ± 5.9
1.5 ± 1.1 18.1 ± 2.3 30.5 ± 5.6
14
15
1
4
126.3 ± 1.5
10.9 ± 2.2
1
4
15
(
lean conditions)
Pt/46Mo/0.27Na/SiO2
57.1 ± 10.0 1.8 ± 1.4 14.2 ± 1.6 73.1 ± 13.0
(
lean conditions)

4
40
BURCH AND COLEMAN
1
1
0
0
0
0
0
.2
.0
.8
.6
.4
.2
.0
of the promoting effect of sodium and/or molybdenum we
have prepared a “model” silica-supported catalyst using a
high loading of molybdenum so as to amplify any effects
that might be occurring.
1
5
2
N O
Figure 9 shows first the effect of switching 15NO for 14NO
with the Pt/SiO2 catalyst in the absence of oxygen. Compar-
ison with Fig. 5 illustrates the complication of using alumina
as a support. Thus, with silica, which does not adsorb NO,
1
5
NO
14
Fig. 9 shows that the N2 profile (see magnified portion of
1
5
N
2
Fig. 9) decays to zero very rapidly after the switch is made.
1
4
15
Similarly, the N N mixed isotope decays to zero much
more rapidly with the silica-supported catalyst.
Repeating this switching experiment in the absence of
oxygen with the Pt/46Mo/0.27Na/SiO2 catalyst gives the re-
sults shown in Fig. 10. The most important point to note is
1
4
15
the very slow decay in the N N profile. Even after 600 s
this still gives a signal corresponding to about 50% of the
initial value. The corresponding estimated values for the
concentration of N2 and N2O precursors on the surface, and
the amount of retained NO, for these two silica-supported
catalysts are given in Table 2. These results show that
1
4
N
2
1
4
4
15
N
NO
1
5
14
N
N
1
N O
2
1
4
NO
1
1
0
0
0
0
0
.2
.0
.8
.6
.4
.2
.0
-
500
0
500
Time / s
1000
1500
1
5
2
N O
FIG. 6. Concentration profiles as a function of time after a SSITK
switch of NO for NO for a Pt/0.27Na/Al2O3 catalyst. Switch from
0.76% NO/Ar/1.08% H2) to (0.76% NO/He/1.08% H2) at 70 C for
00 mg of catalyst at a total gas flow rate of 100 cm min
1
5
14
1
4
15
◦
(
1
15
NO
3
−1
.
is clear that addition of sodium has a small positive effect
on the concentration of N2 precursors and a large positive
effect both on the concentration of N2O precursors and
on the retention of NO. Addition of molybdenum, on the
otherhand, increasesmarkedlytheconcentrationofN2 pre-
cursors and the retention of NO but decreases sharply the
concentration of N2O precursors. When both molybdenum
and sodium are added, there is a further sharp rise in the
concentration of N2 precursors and in the NO retention but
hardly any further change in the already small value for the
concentration of N2O precursors.
1
5
N
2
14
N
2
15 14
N N
1
4
15
N NO
Transient Kinetic Experiments on Silica-Supported
Catalysts under Rich Conditions
1
4
14
N
2
O
NO
As indicated earlier, all our attempts to perform transient
kinetic experiments using 15NO with alumina-supported
catalysts under lean-burn conditions have been unsuc-
cessful because the exchange of 15N with NOx species
adsorbedonthealuminasupportcompletelymaskanytran-
sient changes due to the Pt at the temperatures of our ex-
-500
0
500
Time / s
1000
1500
14
FIG. 7. Concentration profiles as a function of time after a SSITK
15
14
switch of NO for NO for a Pt/10Mo/Al2O3 catalyst. Switch from (0.76%
14
15
◦
NO/Ar/1.08% H2) to (0.76% NO/He/1.08% H2) at 70 C for 100 mg of
periments. Therefore, in order to further probe the details catalyst at a total gas flow rate of 100 cm min
3
−1
.

H2 NO REDUCTION UNDER LEAN-BURN CONDITIONS
441
1
1
0
0
0
0
0
.2
.0
.8
.6
.4
.2
.0
the switch). Figure 12 shows the corresponding results for
the promoted silica-supported catalyst. In this case it is no-
1
4
15
tablethat the N NOprofile decays very fast. On the other
1
4
15
1
5
hand, the N N curve decays much more slowly in the case
of the promoted catalyst. This trend parallels that shown
earlier for the switching experiments on the same cata-
lysts under rich conditions. Consequently, it seems reason-
able to assume that similar mechanisms are involved in both
cases.
N
2
O
1
5
NO
Using the standard SSITK peak profile analysis we arrive
at the concentrations of N2 and N2O precursors and the
amount of retained NO shown in Table 2. Addition of the
promoters results in a large increase in the concentration
of N2 precursors but, within experimental error, essentially
no change in the concentration of N2O precursors or the
amount of retained NO.
1
5
N
2
Non-Steady-State Transient Experiments on
Silica-Supported Catalysts under Lean Conditions
1
4
N
2
1
5
14
N N
The SSITK experiments have shown that promotion
1
4
15
of the Pt/SiO catalyst results in significant differences in
N NO
2
1
4
NO
1
4
2
N O
1
1
0
.2
.0
.8
-
500
0
500
Time / s
1000
1500
15
NO
FIG. 8. Concentration profiles as a function of time after a SSITK
1
5
14
15
switch of NO for NO for a Pt/10Mo/0.27Na/Al2O3 catalyst. Switch
2
N O
1
4
15
◦
from (0.76% NO/Ar/1.08% H2) to (0.76% NO/He/1.08% H2) at 70 C
for 100 mg of catalyst at a total gas flow rate of 100 cm min
3
−1
.
1
5
N
2
addition of the promoters has a major effect in increas-
ing the concentration of N2 precursors, a significant effect
on increasing the amount of NO retained, and a very small
negative effect on the concentration of N2O precursors.
An important distinction between the alumina- and the
0.2
1
5
NO
0.6
0.4
1
4
15
N
NO
1
4
2
N O
0
0
.1
.0
14
N
2
silica-supported catalysts is that in the case of the alumina-
14
supported catalysts the 14N2 peak decays quite slowly
NO
whereas in the silica-supported case the decay is very fast.
This seems to suggest, from the silica-supported catalyst re-
sults, that the formation of N2 requires gaseous NO, in line
with conclusions drawn from previous related work from
our laboratory (33, 34, 55).
0
10 20 30
1
5
14
0
0
.2
.0
N N
Transient Kinetic Experiments on Silica-Supported
Catalysts under Lean Conditions
-100
0
100 200 300 400 500 600
Figure 11 shows the results of switching 15NO for 14NO
under lean conditions (6% O2). The 14N15NO decays more
Time / s
rapidly than the NO, which, comparison with Fig. 9 shows, is
FIG. 9. Concentration profiles as a function of time after a SSITK
significantly delayed under lean conditions. Moreover, the
15
14
switch of NO for NO for a Pt/SiO2 catalyst. Switch from (0.76%
1
4
15
decay of the N N profile is much faster in the presence
of excess oxygen (ca. 200 as compared with >600 s after catalyst at a total gas flow rate of 100 cm min
14
15
◦
NO/Ar/1.08% H2) to (0.76% NO/He/1.08% H2) at 55 C for 100 mg of
3
−1
.

4
42
BURCH AND COLEMAN
to zero after about 150 s. Moreover, the increase in the con-
1
1
0
0
0
0
0
.2
.0
.8
.6
.4
.2
.0
centration of NO2 now also takes a similar length of time,
again suggesting that the Pt surface changes from a reduced
to an oxidised state after the hydrogen is removed. How-
ever, in this case, both the very large amount of N2 formed
and the long time before the Pt becomes oxidised suggest
that in the promoted catalyst there is a source of a reduc-
tant that can continue to convert NO to N2 in the absence
of H . Comparison between the unpromoted Pt/SiO cata-
1
5
2
N O
1
5
NO
2
2
lyst and the promoted catalyst suggests that this additional
reductant may be associated with the promoters. Since the
main promoter is molybdenum, it further seems reasonable
to associate the additional reductant with the molybdenum
oxide.
1
5
14
N
N
1
5
N
2
Finally, Figs. 15 and 16 show the reverse experiments,
in which the H is added to the NO/O mixture. It is im-
0
.3
.2
.1
.0
1
5
2
2
NO
portant to note that the rise in the concentration of H2O
and the fall in the concentration of NO2 are very similar
in both cases, suggesting that the reduction of the catalysts
occurs on the same time scale in both cases. However, very
marked differences are observed between the two catalysts.
14
NO
0
0
0
14
N
2
14
2
N O
14 15
N
NO
0
10 20 30
1
1
0
0
0
0
0
.2
.0
.8
.6
.4
.2
.0
1
5
NO
-
100
0
100 200 300 400 500 600
1
5
2
N O
Time / s
FIG. 10. Concentration profiles as a function of time after a SSITK
1
5
14
switch of NO for NO for a Pt/46Mo/0.27Na/SiO2 catalyst. Switch from
1
5
1
4
15
◦
(
1
0.76% NO/Ar/1.08% H2) to (0.76% NO/He/1.08% H2) at 80 C for
N
2
3
−1
.
00 mg of catalyst at a total gas flow rate of 100 cm min
1
5
14
N N
the concentrations of surface intermediates, leading to the
various reduced products. As a further probe of these mate-
rials, we performed switching experiments in which the H2
was cut off completely. To further increase the sensitivity
of the switching experiments, we used high concentrations
of both O2 and H2 so that the conditions were still lean.
Figure 13 shows the results with the Pt/SiO2 catalyst. As
soon as the H2 was switched off the concentration of N2O
immediately dropped to zero. Conversely, the N2 profile
showed an initial sharp, but short-lived, increase in con-
centration before decaying to zero after ca. 75 s. This much
greater formation of N2 compared to N2O is consistent with
the SSITKA data summarised for this catalyst in Table 2.
In parallel with the decreasing concentration of N2 is an in-
crease in the concentration of NO2, indicating, as expected,
that the Pt surface changes from a reduced to an oxidised
state when the H2 is removed.
0.4
0.3
1
5
NO
1
4
NO
14
15
N
NO
0
0
.2
.1
14
N O
2
14
N
2
0.0
0
10 20 30
-100
0
100
200
300
400
Time / s
Figure 14 shows the remarkable difference observed for
the promoted Pt catalysts when the hydrogen supply is cut
off. In this case, while the N2O still falls rapidly to zero, the
FIG. 11. Concentration profiles as a function of time after a SSITK
15
14
switch of NO for NO for a Pt/SiO2 catalyst. Switch from (0.76%
14
15
NO/Ar/1.08% H2/6% O2) to (0.76% NO/He/1.08% H2/6% O2) at
◦
3
−1
.
N2 concentration rises to a very high value and only decays 105 C for 100 mg of catalyst at a total gas flow rate of 100 cm min

H2 NO REDUCTION UNDER LEAN-BURN CONDITIONS
443
1
1
0
0
0
0
0
.2
.0
.8
.6
.4
.2
.0
• Addition of Mo oxide as a second promoter signifi-
cantly increases the NO conversion with peak values now
in excess of 90%.
1
5
NO
1
5
2
N O
•
Mo oxide promotes the selectivity to N2 relative to
N2O across the whole temperature range of the experi-
ments.
1
5
•
For the unpromoted Pt catalyst, SSITK experiments
N
2
1
4
14 15
show that N2O and N NO are only formed when
1
4
14 15
14
gaseous NO is present whereas N N and N2 continue
to be formed for at least 1500 s after the supply of NO is
1
5
14
N N
14
cut off.
•
Addition of Na delays the desorption of 14NO and
0
0
0
0
0
.4
.3
.2
.1
.0
15
causes a parallel increase in the decay time for the labelled
NO
1
4
N O products.
NO
2
14
• Addition of Mo oxide causes the labelled N O profiles
N
2
O
2
1
4
14 15
to decay faster than the NO profile and also significantly
N
NO
1
4
slows down the labelled N profiles so that the maximum
N2
2
14 15
in the N N curve occurs later than before and does not
decay to zero even after 1500 s.
0
10 20 30
• With the silica-supported catalyst under rich condi-
1
4
tions, the N2 profile decays very rapidly to zero after the
1
4
14 15
supply of NO is cut off; the N N mixed isotopic product
1
1
.6
.4
-
100
0
100
200
300
400
N
2
Time / s
FIG. 12. Concentration profiles as a function of time after a SSITK
1
5
14
switch of NO for NO for a Pt/46Mo/0.27Na/SiO2 catalyst. Switch from
1
4
15
1.2
(
0.76% NO/Ar/1.08%H2/6%O2)to(0.76% NO/He/1.08%H2/6%O2)
at 105 C for 100 mg of catalyst at a total gas flow rate of 100 cm min
◦
3
−1
NO
2
.
1
.0
In particular, the rise in the concentration of N2O and N2 is
much slower in the case of the promoted catalyst, indicating
that the species responsible for the enhanced production of
N2 when the H2 supply was cut off may also be formed
slowly when the H2 supply is restored. Thus, it seems that
while H2 is consumed to make H2O at similar rates in both
catalysts, some of the H2 on the Pt surface is not immedi-
ately available to produce N2. We discuss possible reasons
for this behaviour later.
0.8
0.6
0.4
H
2
2
N O
H O
2
DISCUSSION
0
0
.2
.0
To facilitate the discussion of the results we first sum-
marise the main observations.
•
Addition of small amounts of Na to a Pt catalyst en-
-25
0
25 50 75 100 125 150 175 200
hances significantly the activity in the NO/H2 reaction un-
der lean-burn conditions, but higher amounts of Na inhibit
the reaction.
Time / s
FIG. 13. Concentration profiles as a function of time after switching
off the H2 supply for a Pt/SiO2 catalyst. Switch from (0.5% NO/12.9%
H2/13.3% O2) to (0.5% NO/Ar/13.3% O2) at 110 C for 100 mg of catalyst
at a total flow of 180 cm min
•
Addition of Na at any level has no promoting effect
on the N2/N2O selectivity at lower temperatures but lowers
the selectivity to N2 at higher temperatures.
◦
3
−1
.

4
44
BURCH AND COLEMAN
1
1
1
1
0
0
0
0
0
.6
.4
.2
.0
.8
.6
.4
.2
.0
duction of NO and the selectivity to N2 rather than to N2O
for a Pt film deposited on Al2O3. The optimum coverage
of the Pt surface with sodium was 0.06. In our experiments
under lean conditions we find that while small amounts of
sodium enhance the activity, larger loadings lead to poison-
ing of the Pt catalyst. Moreover, sodium has no promoting
N
2
effect on N formation at low temperatures and actually re-
2
duces the selectivity at higher temperatures. Clearly under
NO
2
oxidising conditions the dissociation of NO is not enhanced
to the same extent as was found by Marina et al. under sto-
ichiometric conditions. This is consistent with the fact that
a partially oxidised Pt surface will contain fewer suitable
sites (for example, two adjacent free Pt atoms) where dis-
sociative adsorption of NO can occur.
The effect of sodium on a reduced Pt surface is thought
to be to enhance the adsorption and dissociation of NO so
that the formation of N2O via a N(ads)+NO(ads) reaction
is inhibited (37–44). However, while the role of sodium is
quite possibly to facilitate NO dissociation, this particular
mechanistic step is not consistent with our earlier work on
the NO/H2 reaction (33, 34, 55) nor with many of the new
H
2
O
H
2
2
N O
1
1
0
.2
.0
.8
H
2
-
25
0
25 50 75 100 125 150 175 200
Time / s
FIG. 14. Concentration profiles as a function of time after switching
off the H2 supply for a Pt/46Mo/0.27Na/SiO2 catalyst. Switch from (0.5%
NO/12.9% H2/13.3% O2) to (0.5% NO/Ar/13.3% O2) at 110 C for 100 mg
of catalyst at a total flow of 180 cm min
◦
3
−1
.
2
N O
also decays more rapidly as compared with the alumina-
supported catalyst.
•
Addition of Mo oxide produces no significant change
0.6
0.4
N
2
in the N2 profile but the 14N N profile now decays very
1
4
15
slowly indeed.
•
For the silica-supported catalyst under lean conditions,
1
4
15
2
H O
thedecayofthe N Nprofileismuchfasterinthepresence
of oxygen, but for the Mo-promoted catalyst the 14N N
decays more slowly. Note also that for the Mo-promoted
15
1
4
15
catalyst the N NO decays very rapidly.
In non-steady-state experiments the formation of N2O
0
0
.2
.0
•
immediately stops when the H2 supply is cut off whereas
there is a transient increase in N2 formation. For the Mo-
promoted catalyst, the N2O behaves similarly but there is
now a very large production of N2 even though there is no
gas-phase reductant present to react with the NO.
NO
2
-25
0
25 50 75 100 125 150 175 200
•
When the H2 supply is reconnected, the concentra-
tions of N2O and N2 recover much more slowly on the Mo-
promoted catalyst.
Time / s
FIG. 15. Concentration profiles as a function of time after switching
on the H2 supply for a Pt/SiO2 catalyst. Switch from (0.5% NO/Ar/13.3%
O2) to (0.5% NO/12.9% H2/13.3% O2) at 110 C for 100 mg of catalyst at
Under stoichiometric conditions, Marina et al. (37) have
reported that the addition of sodium promotes both the re- a total flow of 180 cm min
◦
3
−1
.

H2 NO REDUCTION UNDER LEAN-BURN CONDITIONS
445
1
1
0
0
0
0
0
.2
.0
.8
.6
.4
.2
.0
rapidly. We attribute this to the fact that alumina can store
H
2
NOx , possibly as a nitrite or nitrate, but this does not oc-
1
4
cur with silica. Therefore, the slow decay in the N2 profile
in the case of alumina could indicate that NOx species ad-
sorbed on the alumina can reverse spillover back to the Pt to
1
5
combine with N(ads) atoms, or to dissociate and combine
2
H O
1
5
with NO(ads). We propose this N(ads) + NO(ads) route
to N2 based on our earlier work (33, 34) and also on the fact
that with the silica-supported catalyst there is no formation
1
4
14
of N2 as soon as the NO is switched off even though
1
4
there is a large reservoir of N that can subsequently form
1
4
15
N N.
N
2
O
Figure 11 also confirms for the silica-supported catalyst
1
4
that under lean conditions the formation of N2 does not
1
4
occur in the absence of gaseous NO. Note also (see Fig. 13)
that in non-steady-state experiments N2O formation ceases
immediately when the reductant supply (hydrogen) is cut
off. ThissuggestseitherthatN2Oformationisblockedwhen
the surface becomes over oxidised even though N2 forma-
tion can still proceed, or that NO(ads) molecules simply
scavenge N(ads) atoms faster than they can couple with
other NO(ads) molecules to produce N2O. The fact that
NO2 is observed soon after the H2 is cut off is consistent
with a competition for NO, either by O(ads) to give NO2,
or by N(ads) to give N2, so that the NO(ads) +NO(ads) →
N2O + O(ads) reaction is inhibited, as observed in Fig. 13.
We now consider the role of Mo oxide as a promoter
of the NO reduction reaction. Mo oxide has the following
effects: enhances N2 selectivity; accelerates the decay of the
N
2
NO
2
-25
0
25 50 75 100 125 150 175 200 225
Time / s
FIG. 16. Concentration profiles as a function of time after switching
on the H2 supply for a Pt/46Mo/0.27Na/SiO2 catalyst. Switch from (0.5%
NO/Ar/13.3% O2) to (0.5% NO/12.9% H2/13.3% O2) at 110 C for 100 mg
of catalyst at a total flow of 180 cm min
◦
N O profile relative to that of NO; slows down the decay
2
3
−1
.
of the N2 profiles for both alumina- and silica-supported
Pt; under lean conditions greatly reduces the rate of decay
1
4
15
of the N N profile as compared to the unpromoted cata-
lyst; significantly increases the concentration of adsorbed
species leading to N2 but has no effect on the concentration
of adsorbed species leading to N2O (see Table 2); in non-
steady-state experiments allows the generation of a very
large N2 peak when the H2 is cut off and the recovery of
the N2 (and N2O) peaks when the H2 is restored is very slow.
The only related work in the literature is that of Tanaka
et al. (58), where for the reduction of NO by C3H6 they
report that NO is selectively reduced in the presence of O2
and attribute this to a lower affinity of Pt for oxygen when
the Pt is modified by Mo. In our work it seems that the
Mo oxide has a much more significant role and, in addition
to any small effect on the electronic properties of Pt, may
also be directly involved in the conversion of NO to N2.
Thus, it seems unlikely that electronic modifications of the
results presented in the present work. For example, we see
1
4
15
15
in Fig. 5 that the N NO profile decays and the N2O rises
at essentially the same rate as the 14NO and NO profiles
fall and rise, respectively. This shows that the formation
of N2O requires the presence of gas-phase NO, which is
not too surprising. However, the fact that the N2O profiles
change so rapidly also indicates that this product cannot be
formed from N(ads) and NO(ads) in significant amounts
since there is a large reservoir of N(ads) on the surface, as
demonstrated by the long delay in the various isotopically
labelled N2 profiles.
15
Previously we proposed (33, 34) that N2O is formed at
these lower temperatures through the following reaction:
NO(ads) + NO(ads) = N2O + O(ads).
The present SSITK experiments are consistent with this Pt could account for the extremely large changes in the
model. quantity of adsorbed species leading to the formation of
The N2 profiles for the alumina- and silica-supported Pt N2, for example, as shown in Table 2. We suggest that the
show significant differences. With the alumina-supported Mo oxide also plays a direct role. Analysis of the results
1
4
catalyst the N2 profile takes over 1000 s to decay whereas of the switching experiments (see, for example, Figs. 13 and
14
with the silica-supported catalyst the N2 decays extremely 14) shows that in the absence of H2 there is a temporary, but

4
46
BURCH AND COLEMAN
very large, production of N2 from NO in a Mo-containing
catalyst. Clearly, this catalyst has stored on its surface a
substantial amount of a species that can reduce NO. In our
experiments this has to be a hydrogen-containing species.
H(ads) itself is ruled out because even if this was present
under oxidising conditions, it would only be adsorbed on
the Pt and there is no reason why the amount of H(ads)
should be affected by the addition of Mo oxide (see the
relevant chemisorption results in Table 1).
A more probable explanation is that the Mo oxide cap-
tures a reduced form of nitrogen that may be either formed
on the Mo oxide or spillover from the Pt to adjacent Mo
oxide sites. NH3, or at least an NHx -type species, seems
the most likely candidate. Indeed ammonium molybdate
is a well-known compound and so adsorption of NH3 on
the surface of Mo oxide, leading even to some bulk forma-
tion if the kinetics are satisfactory, seems entirely plausible.
Preliminary FTIR experiments have failed to confirm the
presence of NHx -type species. However, in a separate study
REFERENCES
1
2
3
4
. Chiron, M., Stud. Surf. Sci. Catal. 30, 1 (1987).
. Armor, J. N., Appl. Catal. B 1, 221 (1992).
. De Kermikri, I., Recherche 279, 884 (1995).
. Adams, K. M., Cavataio, J. V., and Hammerle, R. H., Appl. Catal. B
10, 157 (1996).
5
. Bond, G. C., “Heterogeneous Catalysis,” 2nd ed. Oxford University
Press, Oxford, UK, 1987.
6
7
. Fritz, A., and Pitchon, V., Appl. Catal. B 13, 1 (1997).
. Hu, Z., Allen, F. M., Wan, C. Z., Heck, R. M., Steger, J. J., Lakis, R. E.,
and Lyman, C. E., J. Catal. 174, 13 (1998).
8. Harrison, B., Searles, R. A., and Wilkins, A. J. J., in “Car Exhaust
PollutionControl:Lean-BurnEngines&theContinuingRequirement
for Platinum-Containing Autocatalysts” (J. W. S. Longhurst, Ed.),
p. 327. British Library, Technical Communications, 1989.
9
. Gutierrez, L., Boix, A., and Petunchi, J. O., J. Catal. 179, 179 (1998).
1
1
1
1
0. Ribotta, A., Lezcano, M., Kurgansky, M., Miro, E., Lombardo, E.,
Petunchi, J., Moreaux, C., and Dereppe, J. M., Catal. Lett. 49, 77 (1997).
1. Cowan, A. D., Cant, N. W., Haynes, B. S., and Nelson, P. F., J. Catal.
76, 329 (1998).
2. Lombardo, E. A., Sill, G. A., d’Itri, J. L., and Hall, W. K., J. Catal. 173,
40 (1998).
3. Ohtsuka, H., and Tabata, T., Appl. Catal. B 21, 133 (1999).
1
4
(
62) the formation and capture of NH3 has been observed
with ceria–zirconia-supported Pt catalysts in the NO/H2 re- 14. Sullivan, J. A., and Cunningham, J., Appl. Catal. B 15, 275 (1998).
1
1
5. Long, R., and Yang, R. T., Catal. Lett. 52, 91 (1998).
6. Okazaki, N., Shiina, Y., Itoh, H., Tada, A., and Iwamoto, M., Catal.
Lett. 49, 169 (1997).
action and a similar process may operate with the Mo oxide-
promoted catalyst.
1
1
7. Halasz, I., and Brenner, A., Catal. Lett. 51, 195 (1998).
8. Burch, R., and Watling, T. C., J. Catal. 169, 45 (1997).
CONCLUSIONS
19. Burch, R., Sullivan, J. A., and Watling, T. C., Catal. Today 42, 13 (1998).
2
2
0. Burch, R., and Watling, T. C., Catal. Lett. 37, 51 (1996).
1. Hayes, N. W., Joyner, R. W., and Shpiro, E. S., Appl. Catal. B 8, 343
(
It has been shown that for the reduction of NO by H2
under lean conditions at temperatures representative of au-
tomotive “cold-start” conditions the activity and nitrogen
selectivity of platinum-based catalysts can be significantly
increased by addition of MoO3 and Na2O. Addition of
the latter was revealed to increase NO conversion activ-
ity at low loading; however, the catalyst was poisoned at
increased loadings. Moreover, the nitrogen selectivity re-
mained unaffected at low temperatures for all loadings
studied, with adverse effects becoming evident as the tem-
perature was raised. The MoO3 promoter significantly in-
creased both the catalyst activity and selectivity at all load-
ings. Isotopic switching experiments and subsequent SSITK
analysis showed that modification with MoO3 induced sub-
1996).
2. Kameoka, S., Chafik, T., Ukisu, Y., and Miyadera, T., Catal. Lett. 55,
11 (1998).
2
2
2
2
2
2
2
2
2
3. Delahay, G., Ensuque, E., Coq, B., and Figueras, F., J. Catal. 175, 7
(1998).
4. Shimizu, K.-I., Satsuma, A., and Hattori, T., Appl. Catal. B 25, 239
(
2000).
5. Burch, R., Fornasiero, P., and Southward, B. W. L., J. Catal. 182, 234
1999).
(
6. Burch, R., Fornasiero, P., and Southward, B. W. L., Chem. Commun.
739 (1998).
7. Burch, R., Fornasiero, P., and Southward, B. W. L., Chem. Commun.
625 (1998).
8. Ertl, G., Knozinger, H., and Weitkamp, J., Ed., “Environmental Cata-
lysis.” Wiley-VCH, New York, 1999.
9. Taylor, K. C., and Klimisch, R. L., J. Catal. 30, 478 (1973).
stantial increases in the concentration of surface species 30. Yao, H. C., Yu Yao, Y.-F., and Otto, K., J. Catal. 56, 21 (1979).
3
3
3
3
3
1. Otto, K., and Yao, H. C., J. Catal. 66, 229 (1980).
leading to N2 product. Moreover, in non-steady-state tran-
sient experiments when H2 was removed from the feed it
was shown that N2 could be produced for a significantly
longer time period over the modified catalyst in the ab-
2. Sharpe, R. G., and Bowker, M., Surf. Sci. 360, 21 (1996).
3. Burch, R., Shestov, A. A., and Sullivan, J. A., J. Catal. 186, 353 (1999).
4. Shestov, A. A., Burch, R., and Sullivan, J. A., J. Catal. 186, 362 (1999).
5. Kobylinski, T. P., and Taylor, B. W., J. Catal. 33, 376 (1974).
sence of any gas-phase reductant. These observations were 36. Harkness, I. R., and Lambert, R. M., J. Chem. Soc. Faraday Trans. 93
(
7), 1425 (1997).
consistent with the “storage” of a reduced nitrogen species
on the MoO3.
3
3
7. Marina, O. A., Yentekakis, I. V., Vayenas, C. G., Palermo, A., and
Lambert, R. M., J. Catal. 166, 218 (1997).
8. Palermo, A., Lambert, R. M., Harkness, I. R., Yentekakis, I. V.,
Mar’ina, O., and Vayenas, C. G., J. Catal. 161, 471 (1996).
ACKNOWLEDGMENTS
39. Yentekakis, I. V., Konsolakis, M., Lambert, R. M., Macleod, N., and
Nalbantian, L., Appl. Catal. B 22, 123 (1999).
We are grateful to the EPSRC for a postgraduate studentship for MDC.
Also, we are thankful for the assistance of S. T. Daniells.
40. Yentekakis, I. V., Palermo, A., Filkin, N. C., Tikhov, M. S., andLambert,
R. M., J. Phys. Chem. B 101, 3759 (1997).

H2 NO REDUCTION UNDER LEAN-BURN CONDITIONS
447
4
4
1. Macleod, N., Isaac, J., and Lambert, R. M., J. Catal. 198, 128 (2001).
2. Macleod, N., Isaac, J., and Lambert, R. M., Appl. Catal. B 33, 335
2001).
3. Williams, F. J., Tikhov, M. S., Palermo, A., Macleod, N., and Lambert,
R. M., J. Phys. Chem. B 105, 2800 (2001).
4. Williams, F. J., Palermo, A., Tikhov, M. S., and Lambert, R. M., Surf.
Sci. 482–485, 177 (2001).
5. Yentekakis, I. V., Konsolakis, M., Lambert, R. M., Palermo, A., and
Tikhov, M., Solid State Ionics 136, 783 (2000).
6. Lambert, R. M., Palermo, A., Williams, F. J., and Tikhov, M. S., Solid 58. Tanaka, T., Yokota, K., Isomura, N., Doi, H., and Sugiura, M., Appl.
State Ionics 136, 677 (2000). Catal. B 16, 199 (1998).
7. Williams, F. J., Aldao, C. M., Palermo, A., and Lambert, R. M., Surf. 59. Yokota, K., Fukui, M., and Tanaka, T., Appl. Surf. Sci. 121/122, 273
51. Halasz, I., Brenner, A., and Shelef, M., Catal. Lett. 18, 289 (1993).
52. Frank, B., Emig, G., and Renken, A., Appl. Catal. B 19, 45 (1998).
53. Costa, C. N., Stathopoulos, V. N., Belessi, V. C., and Efstathiou, A. M.,
J. Catal. 197, 350 (2001).
54. Ueda, A., Nakao, T., Azuma, M., and Kobayashi, T., Catal. Today 45,
135 (1998).
55. Burch, R., Shestov, A. A., and Sullivan, J. A., J. Catal. 188, 69 (1999).
56. Burch, R., and Coleman, M. D., Appl. Catal. B 23, 115 (1999).
57. Burch, R., and Millington, P. J., Catal. Today 26, 185 (1995).
(
4
4
4
4
4
4
Sci. 412/413, 174 (1998).
8. Yentekakis, I. V., Lambert, R. M., Konsolakis, M., and Kiousis, V.,
Appl. Catal. B 18, 293 (1998).
(1997).
60. Tanaka, T., Yokota, K., Doi, H., and Sugiura, M., Chem. Lett. 409
(1997).
4
5
9. Halasz, I., Brenner, A., and Shelef, M., Catal. Lett. 16, 311 (1992).
0. Halasz, I., Brenner, A., and Shelef, M., Appl. Catal. B 2, 131 (1993).
61. Shannon, S. L., and Goodwin, J. G., Jr., Chem. Rev. 95, 677 (1995).
62. Breen, J. P., Burch, R., and Lingaiah, N., Catal. Lett., in press.