shapes such as irregular, hexagonal and round as well as ‘small’
were also observed.
particles on the support than by sintering because no sign of
particle agglomeration was observed. The initially ‘clear
square’ Pt nanoparticles were completely converted to round
( ≈ 25.5%), irregular ( ≈ 47.3%) and hexagonal ( ≈ 17.2%)
shaped particles (see Table 1). A small amount ( ≈ 6.4%) of
‘unclear square’ Pt particles could be still identified after the
restructuring of the cubic Pt particles. In other words the low
index facets were gradually shifted in high temperature reaction
conditions to higher index planes.
The supporting procedure of the colloidal Pt nanoparticles on
alumina has already been described.4 TEM and XRD observa-
tions confirmed that the morphology of the Pt nanoparticles did
not change after supporting them on alumina. The size of the
supported Pt crystallites was estimated from the broadening of
the Pt(111) XRD peak at 2q ≈ 40° by using the Debye–Scherer
equation. As can be seen in Table 1, the dTEM of 13.3 nm is in
fairly good agreement with dXRD of 11.7 nm. The determination
,5
From the clean surface chemistry, performed in high vacuum
conditions, it is well known that the low index Pt planes, i.e.
2
of the metal surface area by CO or by H chemisorption is not
7
relevant because the exposed surface of the large Pt particles is
very small. Theoretical calculations, by considering an average
particle size of 13 nm, gives a platinum dispersion of around
Pt(100), are less active for NO dissociation as compared with
8
the higher index planes, i.e. Pt(410). The surface defect (steps,
9
kinks) sites are also very efficient for NO decomposition. On
6
%, that is too small to be accurately measured.
The catalytic tests for the NO/CH reaction were performed
in the temperature range 350–600 °C by using a quartz reactor
loaded with 0.05 g of catalyst (1% Pt/Al ). The gas hourly
space velocity (GHSV) of the reactant mixture (0.41% CH , 1%
the low index Pt nanoparticles, having low concentrations of
surface defects, only a fraction of the adsorbed NO dissociates
4
2
in Nads and Oads. The yield of N O on the surfaces with low
2
O
3
activity for NO decomposition is high because the reaction
between NOads and Nads is favored.10 The high activity of the
4
2
1
NO and balance Ar) was 60000 h (the total flow rate of the
high index planes and of the surface defects for NO decomposi-
3
21
8,9
reactant mixture was 50 cm min STP).
tion explains well the decrease in the selectivity to N
2
O, in
The NO was practically completely converted to reaction
products over the fresh catalyst starting from 350 °C. After the
first test, the catalyst was subjected to an accelerated thermal
parallel with the increase in N
? N (Fig. 2).
The small Pt particles strongly interact with the oxygen in
2
yield via the reaction Nads + Nads
2
aging at 950 °C for 4 h in the NO/CH
4
reaction mixture. This
x
reaction conditions forming PtO species (bulk oxide) with low
allowed us to observe the morphological changes taking place
with the well-structured (manly cubic) Pt nanocrystals during
aging as well as the effect of this evolution on the catalytic
performances. The catalytic behaviour of the aged catalyst for
oxidation activity.11 On the other hand, the surface of the Pt
nanoparticles will be covered only with a layer of reactive
oxygen12 because the properties of the large metal particles (d >
13
5 nm) resembles those of the bulk metal. The CH
formed by the dissociative adsorption of CH will be further
oxidized to CO by oxygen resulting from the dissociation of
x
species
the NO/CH
4
reaction was then checked again in the low
4
temperature region (350–600 °C). The observed modifications
can be summarized as follows: (i) the catalytic activity for NO
conversion decreased slightly (the temperature for the total
conversion of NO was shifted from 350 to 400 °C), (ii) the
x
NO. The low activity of well-structured (cubic) Pt nanocrystals
for NO decomposition reduces the supply of active oxygen for
CH oxidation and therefore the catalyst will exhibit a higher
x
yield to CO as compared with the restructured Pt nanoparticles
(Fig. 2). The high activity of the restructured Pt nanoparticles
for NO decomposition increases the supply of active oxygen
which rapidly removes the surface carbonaceous species as
CO . The above-proposed mechanism is sustained also by the
2
fact that the reaction selectivity to hydrogen (not presented
here) decreased for the restructured Pt particles.
production of N
2
O and CO decreased significantly and (iii) the
was prevented.
formation of NH
3
Fig. 2 illustrates the aging effect on the catalytic behaviour of
alumina supported Pt nanoparticles. For simplicity only the
yields to the harmful products (N
compared (the selectivity to N , CO and H
depressed the formation of N O in the whole temperature
2
O, CO, and NH
3
) are
2
2
2
is omitted). Aging
2
domain investigated. Another positive benefit of the morpho-
logical evolution of the Pt nanoparticles is the complete
It is accepted that NH
3
is formed in the reactions between
N
ads (resulted from NO decomposition) and a hydrogen source
suppression of NH
3
formation and the significant decrease in
x 4
which can be either CH or Hads (both resulting from CH
CO production. The yield to CO becomes negligible below 550
splitting).14 The larger amount of oxygen available on the
surface of the restructured Pt favours the removal, as oxidation
°
C whereas for higher temperatures it decreases sharply
compared to the fresh catalyst.
products (CO
ads) responsible for ammonia formation. Therefore the
reaction between Nads and Nads becomes a favoured one in
comparison with the reaction of ammonia formation (i.e. Nads
).
2 2 x
and H O), of the hydrogen sources (CH and
The data presented in Table 1 shows that: (i) the size of Pt
nanocrystals seems to be not significantly altered by aging and
H
(
ii) the Pt nanocrystals are undergoing a significant facet
restructuring during the high temperature treatment. The larger
TEM (16.3 nm) as compared with dXRD (10.6 nm) for the aged
catalysts can be explained rather by the flattening of the Pt
+
[H]source ? NH
3
d
A research fund from the Japanese Society for the Promotion
of Science (No. P00136) is greatly appreciated.
Notes and references
1
2
B. C. Gates, Chem. Rev., 1995, 95, 511.
J. M. Rickard, L. Genovese. A. Moata and S. Nitsche, J. Catal., 1990,
1
21, 141.
3
4
5
6
7
A. Miyazaki, I. Balint and K-i. Aika, J. Catal., 2001, 204, 364.
I. Balint, A. Miyazaki and K-i. Aika, Chem. Lett., 2001, 1024.
I. Balint, A. Miyazaki and K-i. Aika, Appl. Catal. B, 2002, 37, 217.
A. Miyazaki and Y. Nakano, Langmuir, 2000, 16, 7109.
C. M. Comrie, W. H. Weinberg and R. M. Lambert, Surf. Sci., 1976, 57,
6
19.
8
9
W. F. Banholzer and R. I Masel, J. Catal., 1984, 85, 127.
T. Zambelli, J. Wintterlin, J. Trost and G. Ertl, Science, 1996, 273,
1
688.
1
1
0 T. W. Root and L. D. Schmidt, Surf. Sci., 1983, 134, 30.
1 R. F. Hicks, H. Qi, M. L. Young and R. G. Lee, J. Catal., 1990, 122,
Fig. 2 Comparison between the NO/CH
2, 5) and NH (., /) for the fresh (open symbols) and aged (closed
symbols) alumina-supported Pt nanoparticles in the 350–600 °C tem-
perature range (reactant mixture: 0.41% CH , 1% NO and balance Ar).
4
reaction yield to N
2
O (Ω, :), CO
280.
(
3
12 E. S. Putna, J. M. Vohs and R. J. Gorte, Surf. Sci., 1997, 391, L1178.
13 M. Boudart, J. Mol. Catal., 1985, 30, 27.
14 R. Burch and A. Ramli, Appl. Catal. B, 1998, 15, 63.
4
CHEM. COMMUN., 2002, 1044–1045
1045