Improved N2 selectivity for platinum on alumina prepared by sol-gel technique in the reduction of NOx by propene

Article abstract of DOI:10.1006/jcat.1998.2246

We present results which demonstrate that a 1% Pt/alumina catalyst prepared by a sol-gel technique is superior in its N₂ selectivity in the catalytic reduction of NO_x by propene to similarly loaded Pt/alumina catalysts prepared by the more commonly used impregnation technique. The selectivity of the sol-gel catalyst to N₂ at maximum conversion is 83% as compared to 43% and 60% selectivity observed with Pt catalysts prepared using the impregnation technique. The sol-gel made catalyst does not change its activity and selectivity even in the presence of 2.5% water.

Full text of DOI:10.1006/jcat.1998.2246

JOURNAL OF CATALYSIS 179, 339–342 (1998)
ARTICLE NO. CA982246
PRIORITY COMMUNICATION
Improved N2 Selectivity for Platinum on Alumina Prepared by Sol-Gel
Technique in the Reduction of NOx by Propene
Erol Seker and Erdogan Gulari1
Chemical Engineering Department, The University of Michigan, Ann Arbor, Michigan 48109-2136
Received June 29, 1998; revised July 21, 1998; accepted August 8, 1998
EXPERIMENTAL
We present results which demonstrate that a 1% Pt/alumina cata-
Catalyst Preparation and Activity Tests
lyst prepared by a sol-gel technique is superior in its N2 selectivity
in the catalytic reduction of NOx by propene to similarly loaded
Pt/alumina catalysts prepared by the more commonly used impreg-
nation technique. The selectivity of the sol-gel catalyst to N2 at
maximum conversion is 83% as compared to 43% and 60% selec-
tivity observed with Pt catalysts prepared using the impregnation
technique. The sol-gel made catalyst does not change its activity and
c
In our sol-gel process (10), aluminum tri-sec butoxide,
ethanol, excess amount of water, and the necessary amount
of hexachloroplatinic acid (as a precursor for Pt loading, to
get 1 wt% Pt in the final catalyst) were used to make a gel.
The gel was dried at 100ꢁC for 12 h, followed by calcina-
tion at 600ꢁC for 24 h. This calcined catalyst was tested and
characterized without further pretreatment. In addition to
this sol-gel made platinum catalyst (SG-Pt), an impregna-
tion catalyst with 1 wt% Pt loading (IM-Pt) was prepared
using Degussa alumina (100 m2/g.) and hexachloroplatinic
acid. A second impregnation catalyst (Pt-IMP-Al2O3(SG))
was also prepared with alumina produced by the same sol-
gel synthesis technique used in our single step synthesis.
Both of the impregnation catalysts were dried at 100ꢁC to
remove water and calcined at 500ꢁC for 16 h. These cal-
cined catalysts were reduced with hydrogen prior to activity
studies.
The reactant gas mixture was blended by using four inde-
pendent mass flow controllers to get 750 ppm NO, 750 ppm
propene, 6.5% oxygen, 2.5% water (when used) and he-
lium as balance. In addition, all catalysts were ground and
sieved to 120–180 mesh size prior to catalytic studies. The
total flow rate of reactants and the amount of the catalyst
used were 72 ml/min and 0.1 g, respectively.
selectivity even in the presence of 2.5% water.
ꢀ 1998 Academic Press
Key Words: N2 selectivity; lean burn engines; nitric oxide reduc-
tion; platinum/alumina; propene.
INTRODUCTION
Of all the catalysts tested so far alumina-supported Pt
has shown the highest activity at the lowest temperatures
for selective NOx reduction from exhaust gas streams (1–7).
High activity at low temperatures makes Pt/alumina a de-
sirable catalyst. However, Pt/alumina and Pd/alumina cata-
lysts have very high selectivities for unregulated but highly
undesirable nitrous oxide. Based on literature (1,8), Pt on
alumina catalysts typically exhibit at least 50% selectivity to
N2O, the remainder being N2. Thus, it would be very desir-
able to develop a catalyst with the low temperature activity
of Pt but without its high selectivity to N2O for use in diesel
engine vehicles, which have cooler exhaust than gasoline
engines.
Recently we have prepared (9) high activity and N2 se-
lective silver/alumina catalysts for selective NOx reduction,
using a single step sol-gel synthesis. In order to see if we can
achieve the same gain that we have seen with silver cata-
lysts with platinum-based catalysts we also prepared single
step sol-gel platinum/alumina catalysts. In this study, we re-
port the preliminary results of experiments carried out with
some of the 1% Pt/alumina catalysts.
Catalyst Characterization and Product Analysis
All catalysts were analyzed by using a Rigaku XRD sys-
tem for the determination of bulk properties and Micro-
meritics ASAP 2010 for the BET surface area measure-
ment. The effluent gas mixture from the reactor was
analyzed by a Thermo Environmental 42CHL NOx chemi-
luminescence analyzer and by FTIR with a 10-cm path
length gas cell for quantitative determination of CO2, CO,
1 To whom proofs should be mailed. E-mail: gulari@engin.umich.edu. and N2O.
339
0021-9517/98 $25.00
c
Copyright ꢀ 1998 by Academic Press
All rights of reproduction in any form reserved.

340
SEKER AND GULARI
RESULTS AND DISCUSSION
Figure 1 shows the NOx reduction activities of the
three catalysts as a function of temperature. When the
combustion of propene is around 95% , as seen in Fig. 2,
the maximum in NOx conversion is observed for all three
catalysts. This is a general phenomenon within lean-NOx
catalysis. The temperature at thispoint isnot the same for all
the catalysts. For the catalyst made with the impregnation
technique using Degussa alumina (IM-Pt), the maximum in
NOx conversion (37% ) occurred at 254ꢁC. The impregna-
tion catalyst made with our sol-gel alumina reached a peak
conversion of 35% at 255ꢁC and for the single step sol-
gel made catalyst (SG-Pt), a maximum NOx conversion of
34% was observed at 269ꢁC. Interestingly there is a signif-
icant difference in the activity of the three catalysts below
their peak conversion temperatures. At 150ꢁC, all have low
conversions but at 200ꢁC the impregnation catalysts have
conversions of ꢂ30% and ꢂ25% versus ꢂ2% for the sin-
gle step sol-gel catalyst. Above the peak conversion tem-
perature the shape of the conversion versus temperature
is similar for all three. Above 350ꢁC, the single step sol-
gel catalyst has higher conversions. Also, below the peak
conversion temperature, NO2 formation was not observed
for all three catalysts. However, above the peak conver-
sion temperature up to 23% (at 350ꢁC) of NO was con-
verted to NO2 for all the catalysts. This can be explained
by a lack of sufficient reductant and platinum catalysts
being good oxidizing catalysts under oxygen-rich reaction
conditions.
FIG. 2. Propene combustion for IM-Pt catalyst and SG-Pt catalysts
as measured by CO2 concentration in the reactor effluent.
a broad conversion curve showing detectable activity from
150ꢁC to 350ꢁC with a full width at half peak of almost 150ꢁ.
The Pt-IMP-Al2O3(SG) catalyst conversion curve has a full
width at half maximum of almost 130ꢁC. These widths can
be compared to a maximum width of 100ꢁC observed by
Burch and co-workers (2) with their Pt catalysts. In com-
parison the shape of the SG-Pt conversion curve is simi-
lar to those observed by Burch and co-workers (2). It is
sharply peaked and has a full width at half peak conversion
of ꢂ75ꢁC. We speculate that the unusual widths of the ac-
tivity curves of the impregnation catalysts are most likely
due to broader particle size distribution of these catalysts
compared to the single step sol-gel catalyst, but as of now
we do not have any conclusive evidence. Figure 2 shows
that the combustion of propene is slower on the SG-Pt
catalyst and there is a difference of ꢂ50ꢁC between the two
catalysts. The combustion rate increasesalmost linearlybet-
ween 150ꢁC and 240ꢁC for the two impregnation catalysts,
whereas the rate increase seen with the SG-Pt catalyst is
more exponential in nature, indicative of a higher activation
energy combustion. The difference in combustion catalysis
is obviously the reason for the broader activity peaks of the
impregnation catalysts.
The widths of the NOx conversion curves for the cata-
lysts are also significantly different. The IM-Pt catalyst has
The selectivities of the three catalysts are shown in Fig. 3.
With platinum catalysts using hydrocarbons as reductants
there are two main products of NO reduction: N2O and N2.
N2 is the desired product as N2O will eventually decom-
pose to produce NO in the atmosphere. Based on litera-
ture reports (1,2), at least 50% of NO is reduced to nitrous
oxide by alumina-supported platinum catalysts. This was
also seen with our IM-Pt catalyst as shown in the Fig. 3.
FIG. 1. Steady state catalytic activity of the three catalysts.

IMPROVED N2 SELECTIVITY FOR Pt/Al2O3
341
and/or covered by a thin layer of alumina perhaps simi-
lar to what is observed with titania-supported platinum af-
ter high temperature reduction. These strong interactions
with alumina may change the oxidation state of some of
the Pt atoms on the surface and lead to the formation of
more N2, either by only dissociatively adsorbing NO, or
by changing the surface recombination rates between N
and NO molecules. We also carried out some preliminary
FTIR investigation of surface species seen on impregna-
tion and sol-gel single step catalysts. FTIR in situ reaction
studies show that organic nitro and organic nitrite species
appear to be present under reaction conditions on the
SG-Pt catalyst surface. In comparison, on impregnation
made Pt catalysts, isocyanate species were also present
(11,12). So far the results obtained with all the catalysts
appear to be very reproducible within experimental error
(ꢄ2% ) over a period of several months. The reproducibility
error in selectivity at low conversions (<10% ) is higher be-
cause of the very low concentrations of N2O measured. We
have also successfully synthesized several different batches
of the SG-Pt catalyst which show the same enhanced se-
lectivity with respect to N2 in NOx reduction by hydrocar-
bons.
FIG. 3. Comparison of conversions of the three catalysts to N2 and
N2O. Note that conversion here is defined as the percentage of the original
NOx reduced to N2 or N2O. The total conversion is the sum of the two
conversions.
CONCLUSIONS
We have shown that it is possible to make a highly se-
lective alumina-supported Pt catalyst for the reduction of
NOx by propene under oxidizing conditions by changing
the preparation procedure. This new procedure narrows
the temperature window on the low temperature side while
increasing the selectivity dramatically. Preliminary studies
showthat addition ofwater widensthe temperature window
on the high temperature side without negatively affecting
the conversion or selectivity. We are continuing our stud-
ies of N2 selectivity enhancement for Pt catalysts prepared
by the sol-gel method. We are also in the process of using
other surface science techniques such as XPS and electron
microscopy to further understand the differences between
the catalysts and will report on our other findings in the
near future.
The N2 selectivity of the Pt-IMP-Al2O3(SG) is better than
about 60% at peak conversion. However, our SG-Pt cata-
lyst with the same metal loading is better than either of
the impregnation catalysts and shows ꢂ83% N2 selectivity
at peak conversion. Note that the percentage of selectiv-
ity, in the reduction of NOx (NO + ꢂ3% NO2 present in
the NO as an impurity), is defined as 100 ꢃ the amount of
N2/(amount of N2 + amount of N2O). We also tested the
effect of water on total conversion and selectivity of our
sol-gel-made catalyst and did not observe any changes in
the total conversion or the selectivity to nitrogen.
In order to explain the difference in the selectivites and
activities observed between the two catalysts we measured
the particle sizes of the IM-Pt and SG-Pt catalysts by XRD.
XRD results indicated that we had metallic platinum in
both catalysts and that the average crystallite size was
ꢂ11 nm for the impregnation catalyst and 17 nm for the
sol-gel catalyst. BET measurements revealed that the De-
gussa alumina we used had a surface area of 100 m2/g and
the sol gel catalysts both had surface areas of 250 m2/g
with a narrow pore size distribution sharply peaked at
ꢂ2.7 nm. Thus, based on the limited information provided
by XRD and BET, it is not possible to explain the reasons
for the large difference in selectivity. One possible specu-
lative explanation is that since the SG-Pt catalyst had Pt
incorporated into the alumina gel before the gel was dried
and calcined, some of the Pt may be atomically dispersed
ACKNOWLEDGMENTS
We thank YOK of the Republic of Turkey for financial support of Erol
Seker by a Ph.D. fellowship.
REFERENCES
1. Obuchi, A., Ohi, A., Nakamura, M., Ogata, A., Mizuno, K., and
Ohuchi, H., Appl. Catal. B: Environmental 2, 71 (1993).
2. Burch, R., Millington, P. J., and Walker, A. P., Appl. Catal. B: Envi-
ronmental 4, 65 (1994).
3. Hamada, H., Kintaichi, Y., Sasaki, M., Ho, T., and Tabata, M., Appl.
Catal. 75, L1 (1991).

342
SEKER AND GULARI
4. Inaba, M., Kintaichi, Y., and Hamada, H., Catal. Lett. 36, 223
(1996).
8. Burch, R., and Watling, T. C., Appl. Catal. B: Environmental 11, 207
(1997).
5. Miyadera, T., Appl. Catal. B: Environmental 13, 157 (1997).
6. Ogasawara, K., Muramatsu, G., and Miakino, S., JSAE Rev. 16, 61
(1995).
9. Lorpongpaiboon, P., Osuwan, S., Seker, E., Cavatio, J., and Gulari, E.,
J. Appl. Catal. A: General, submitted.
10. Seker, E., and Gulari, E., in preparation.
7. Demicheli, M. C., Hoang, L. C., Menezo, J. C., and Barbier, J., Appl. 11. Okuhara, T., Hasada, Y., and Misono, M., Catal. Today 35, 83 (1997).
Catal. A: General 97, L11 (1993). 12. Rasko, J., and Solymosi, F., J. Catal. 71, 219 (1981).