In2O3/Al2O3 catalysts for NOx reduction in lean condition

Article abstract of DOI:10.1006/jcat.2002.3667

The lean NO_x performance and catalytic properties of In₂O₃/Al₂O₃ catalysts were investigated. High lean NO_x activity was observed when propene was used as a reductant in the presence of 9% O₂ and 7% H₂O at a space velocity of 30,000 h^-1. The optimum lean NO_x activity of In₂O₃/Al₂O₃ catalysts was observed at a loading of 2.5 wt% indium on γ-Al₂O₃ which was prepared by a sol-gel technique (230 m²/g). When propane was used as a reductant, the In₂O₃/Al₂O₃ catalyst did not promote NO_x reduction compared to the alumina substrate. X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), and temperature-programmed reduction (TPR) have been used to characterize a series of In₂O₃/Al₂O₃ catalysts to increase understanding of the surface structure of indium oxide species on the alumina support. The XRD data indicated that crystalline In₂O₃ was present at In₂O₃ loadings >5 wt% and the quantity of the crystalline phase increased as a function of indium loading. XPS results suggested that indium oxide existed as a well-dispersed phase up to 10 wt% indium. The well-dispersed or reducible indium oxide species below 400°C in TPR experiments were assigned as the sites which activate propene to oxygenated hydrocarbons such as acetaldehyde and acrolein. Alumina sites readily utilize the oxygenated hydrocarbons to reduce NO_x. A bifunction mechanism was proposed to explain high NO_x reduction over In₂O₃/Al₂O₃ catalysts.

Full text of DOI:10.1006/jcat.2002.3667

Journal of Catalysis 210, 97–105 (2002)
doi:10.1006/jcat.2002.3667
In2O3/Al2O3 Catalysts for NOx Reduction in Lean Condition
∗
∗
,2
Paul W. Park,∗,1 Christie S. Ragle, Carrie L. Boyer, M. Lou Balmer,†
Mark Engelhard,† and David McCready†
∗
Caterpillar Inc., Tech Center E/854, P.O. Box 1875, Peoria, Illinois 61656; and †Pacific Northwest National Laboratory,
P.O. Box 999, Richland, Washington 99352
Received December 18, 2001; revised April 30, 2002; accepted May 8, 2002
nia. In this respect, a lean NOx catalyst using onboard fuel
to supply hydrocarbon as a reductant has attracted much
attention as an alternative technique to conventional NH3–
SCR processes. Cu–ZSM-5 was discovered to be a catalytic
materialveryeffectiveinreducingNOtoN2 (1, 2);however,
it has been a great challenge to develop more-effective and
-durable catalysts to remove NOx from diesel or lean-burn
gasoline engines.
Zeolite materials show excellent NOx reduction effi-
ciency especially at high space velocity; however, cation
migration and hydrothermal stability limits the application
ofthematerialsinpracticalapplications. Noblemetal-based
catalysts show high NOx conversion at lower temperatures;
however, these catalysts inefficiently utilize reductants and
produce N2O, which is a stable pollutant and a strong green-
house gas. Base metal oxides supported on alumina cata-
lysts have drawn many researchers’ attention because these
catalytic materials are highly stable at elevated temperature
as well as very selective for NOx reduction while consuming
a minimal amount of reductant.
Indium oxide has shown promise for lean NOx cataly-
sis. In2O3 doped on alumina has been reported to show
high-activity NOx reduction as well as high resistance to
water and SO2 poisoning in lean conditions (3). Kikuchi
and co-workers (4–6) introduced indium into various zeo-
litematerialsandreportedthatIn/H–ZSM-5showedhigher
catalytic activity and selectivity for NOx reduction than
other In-exchanged zeolitic materials when CH4 was used
as a reductant. Hamada and co-workers (7, 8) studied
In2O3-doped sol–gel alumina catalysts and reported that
the catalyst showed better catalytic performance than In-
or Cu-exchanged ZSM-5 catalysts. In addition, the indium
catalysts showed higher activity for NO reduction over the
broad temperature window compared to various metal ox-
ides, such as Sn, Ag, Co, Cu, Fe, and Ni, supported on
30wt%Ga2O3/Al2O3 catalysts, especiallyinthepresenceof
H2O at low temperature (≤350◦C). The authors concluded
that the presence of H2O enhanced the catalytic perfor-
mance for NO reduction due to suppression of undesirable
propene oxidation and to removal of surface carbonaceous
materials.
The lean NOx performance and catalytic properties of
In2O3/Al2O3 catalysts were investigated. High lean NOx activity
was observed when propene was used as a reductant in the presence
of 9% O2 and 7% H2O at a space velocity of 30,000 h−1. The opti-
mum lean NOx activity of In2O3/Al2O3 catalysts was observed at a
loading of 2.5 wt% indium on γ -Al2O3 which was prepared by a sol–
gel technique (230 m2/g). When propane was used as a reductant,
the In2O3/Al2O3 catalyst did not promote NOx reduction compared
to the alumina substrate. X-ray photoelectron spectroscopy (XPS),
X-ray diffraction (XRD), and temperature-programmed reduction
(TPR) have been used to characterize a series of In2O3/Al2O3 cata-
lysts to increase understanding of the surface structure of indium
oxide species on the alumina support. The XRD data indicated
that crystalline In2O3 was present at In2O3 loadings >5 wt% and
the quantity of the crystalline phase increased as a function of in-
dium loading. XPS results suggested that indium oxide existed as
a well-dispersed phase up to 10 wt% indium. The well-dispersed or
reducible indium oxide species below 400◦C in TPR experiments
were assigned as the sites which activate propene to oxygenated hy-
drocarbons such as acetaldehyde and acrolein. Alumina sites read-
ily utilize the oxygenated hydrocarbons to reduce NOx. A bifunc-
tion mechanism was proposed to explain high NOx reduction over
c
In2O3/Al2O3 catalysts. ꢀ 2002 Elsevier Science (USA)
Key Words: NOx; reduction; lean; In2O3; Al2O3; XPS; XRD; TPR;
acrolein; acetaldehyde.
1. INTRODUCTION
Nitrogen oxides (NOx ) emitted from internal combus-
tion engines are a serious environmental concern. For sta-
tionary lean-burn engine applications, NOx has been ef-
fectively controlled by selective catalytic reduction (SCR)
with ammonia. However, ammonia SCR technology may
not be suitable for mobile applications because it requires
a nationwide infrastructure of ammonia or urea supply and
needs to address safety issues for handling toxic ammo-
1 To whom correspondence should be addressed. E-mail: park paul w@
cat.com.
2 Current address: Caterpillar Inc., Tech Center E/854, P.O. Box 1875,
Peoria, IL 61656.
97
0021-9517/02 $35.00
c
ꢀ 2002 Elsevier Science (USA)
All rights reserved.

98
PARK ET AL.
Although promising results of indium-based catalysts γ -Al2O3, two individual reactors containing each compo-
have been recently reported, the detailed surface struc- nent were installed to a bench system with three-way valves
ture of indium species on alumina and the role of in- to pass or bypass the reaction mixture so that the products
dium in NOx reduction have not been investigated. In produced over each catalyst bed individually or in series
this paper, we show that an In2O3/Al2O3 catalyst is one were analyzed.
of the most active catalysts for lean NOx catalysis using a
2.3. BET
propene reductant in simulated engine exhaust (9% O2, 7%
H2O, and 30,000 h−1 space velocity). XRD (X-ray diffrac-
tion), XPS (X-ray photoelectron spectroscopy), and TPR
(temperature-programmed reduction) were used to char-
acterize the properties of indium species on alumina. The
information derived from the characterization techniques
has been correlated with NOx reduction (NO and NO2) re-
sults using different hydrocarbons (C3H6 and C3H8). The
reaction mechanism and identity of the active sites on these
In2O3/Al2O3 catalysts are proposed.
Surface area, pore volume, and pore size distribu-
tion measurements were performed using a Micromeritics
ASAP 2010 system. Catalyst samples were outgassed under
vacuum at 450◦C overnight prior to taking adsorption mea-
surements. The surface area was determined by multipoint
measurements using several relative pressures of N2 in the
range 0.06–0.2 (N2 surface area, 0.162 nm2) at 77 K.
2.4. X-Ray Diffraction (XRD)
X-ray powder diffraction patterns were obtained with a
Philips PW3040/00 X’Pert MPD system which employed a
PW3050/10 vertical theta–theta goniometer and a PW3373
long fine-focus, ceramic X-ray tube with fixed Cu anode.
The X-ray source was operated at 2000 W (40 kV, 50 mA).
The optical train of the diffractometer consisted of pro-
grammable divergence, antiscatter, and receiving slits, a
curved graphite monochromator, and a Xe proportional
counter detector. The divergence and antiscatter slits were
operated in 2-theta compensating mode. Combined with a
fixed, 10-mm-wide brass beam mask, the study area on the
mounted specimens was limited to 10 mm × 10 mm. The
receiving slit was set at 0.2 mm height. The samples were
fine-grained, dry powders mounted in 12 mm diameter ×
0.5 mm deep cavities cut in off-axis single-crystal quartz
plates. The range of diffraction data collection was 10–80◦
(2-theta), which was scanned at a rate of 0.04◦/2 s.
2. EXPERIMENTAL
2.1. Catalyst Preparation
The alumina support was prepared by a sol–gel method
using alumina isopropoxide and 2-methyl-2,4-pentanediol
as a complexing agent. The procedure for the alumina pre-
paration was described previously (9). In2O3/Al2O3 cata-
lysts were prepared using the incipient wetness technique
with γ -alumina powder (surface area, 230 m2/g) and aque-
ous solutions of indium(III) nitrate pentahydrate (Aldrich,
99.99%). In (2.5 wt%) doped on SiO2 powder (Aerosil
Degussa, 200 m2/g) was also prepared by an incipient wet-
ness technique to study reaction mechanisms. The impreg-
nated samples were dried in air at 100◦C for 24 h and cal-
cined at temperatures up to 600◦C (ramp rate, 1.2◦C/min)
for 5 h under flowing air (5 L/min). The indium content of
the In2O3/Al2O3 series was varied from an In/Al atomic ra-
tio of 0 to 0.128 (0–20 wt% based on In metal). In2O3/Al2O3
catalysts will be designated “Iny,” where y is the In content
(weight percent).
2.5. X-Ray Photoelectron Spectroscopy (XPS)
XPS measurements were made on a Physical Electronics
Quantum 2000 Scanning ESCA Microprobe. This system
uses a focused monochromatic Al Kα X-ray (1486.7 eV)
source for excitation and a spherical section analyzer. The
X-ray beam used was a 100-W, 100-µm-diameter beam that
is rastered over a 1.5 mm by 0.2 mm rectangle on the sample.
The X-ray beam is incident normal to the sample and the X-
ray detector is at 45◦ away from normal. The collected data
were referenced to an energy scale with binding energies
for Cu 2p3/2 at 932.67 0.05 eV and for Au 4f at 84.0
0.05 eV. Low-energy electrons and argon ions were used
for specimen neutralization.
2.2. Catalytic Performance Test
The catalytic performance of samples was measured us-
ing a quartz microtube reactor system with a feed gas of
0.1% NOx , 0.1% propene, 9% O2, and 7% H2O balanced
with He. The total gas flow rate was 200 cc/min, and 0.2 g of
catalyst sample was used, unless otherwise noted. The re-
action products were measured and identified by GC, GC–
MS, or NOx chemiluminescence analyzer. NOx conversion
was calculated from the amount of N2 produced. Since no
by-product such as N2O was detected in any of the experi-
ments, the measured N2 yield corresponded directly to NOx
conversion. The catalyst samples were initially heated to
600◦C with 15% O2/He. When the temperature reached
2.6. Temperature-Programmed Reduction (TPR)
Temperature-programmed characterization was per-
600◦C, H2O, NOx , and propene were introduced into the formed on an Advanced Scientific Designs Inc. (ASDI)
reactionmixture. Activitydatawasrecordedfromhightem- catalyst test stand. Catalyst powder samples (0.1–0.2 g)
peraturetolowtemperature. TostudytherolesofIn2O3 and were placed on a glass frit in a flow-through tube that was

In2O3/Al2O3 CATALYSTS FOR NOx REDUCTION IN LEAN CONDITION
99
evacuated at 10−3 torr. The samples were pretreated to
a
100
80
60
40
20
0
105◦C at 10−3 torr for 1 h, then cooled to room temper-
ature under vacuum. For temperature-programmed reduc-
tion studies, the sample was exposed to a mixture of 4%
H2 in argon at a flow rate of 10 cc/min. Samples were heat
treated at a rate of 10◦C/min to 850◦C then held at 850◦C
for 1 h under the flowing H2/Ar mixture. Hydrogen uptake
during heat treatment was measured using a thermal con-
ductivity detector (TCD).
Al2O3
In1
In2.5
In5
In10
3. RESULTS
3.1. NOx Reduction over In2O3/Al2O3 Catalysts
350
400
450
500
550
600
650
3.1.1. The effect of indium loading on NOx reduction.
In order to determine the optimum loading of indium on
the sol–gel alumina support, NO reduction with C3H6 was
examined over a series of indium-doped alumina catalysts.
Figure 1a shows N2 yields and Fig. 1b integral N2 produc-
Temperature (oC)
5
4
3
2
1
0
b
Al2O3
In1
In2.5
In5
In10
80
a
Al2O3
In1
In2.5
In5
In10
60
40
20
0
350
400
450
500
550
600
650
Temperature (oC)
FIG. 2. The effect of indium loading on (a) the C3H6 conversion and
(b) the integral C3H6 consumption rate normalized by the catalyst surface
area for NO reduction over In2O3/Al2O3 catalysts (0.1% NO, 0.1% C3H6,
9% O2, 7% H2O, SV = 30,000 h−1).
350
400
450
500
550
600
650
Temperature (oC)
12
10
8
b
tion rates as a function of reaction temperature for NO
reduction over 1, 2.5, 5, and 10 wt% indium-doped Al2O3
catalysts. The NO reduction performance for undoped alu-
mina is also shown in Fig. 1 for comparison. Tmax (tempera-
ture at maximum NOx conversion) of the alumina for NO
reduction was 575◦C and the NO conversion at Tmax was
51%. All of the indium-doped alumina catalysts showed
lower Tmax and exhibited a broader temperature window
of operation than those of the undoped alumina. The In2.5
catalyst showed optimum NO reduction performance with
the highest N2 yield (60%). The performance of NO reduc-
tion over the In2O3/Al2O3 catalysts decreased with increas-
ing indium loading, from 2.5 to 10 wt%.
Al2O3
In1
In2.5
In5
In10
6
4
2
350
400
450
500
550
600
650
Temperature (oC)
Figure 2a illustrates propene conversions and Fig. 2b in-
tegral propene consumption rates as a function of reaction
temperature for NO reduction over the catalysts. Undoped
alumina, shown on the far right of the graph, converted
propene at the rate of 20–100% in the temperature range
FIG. 1. The effect of indium loading on (a) the N2 yield and (b) the
integral N2 formation rate normalized by the catalyst surface area for NO
reduction over In2O3/Al2O3 catalysts (0.1% NO, 0.1% C3H6, 9% O2, 7%
H2O, SV = 30,000 h−1).

100
PARK ET AL.
70
60
50
40
30
20
10
0
100
80
60
40
20
0
Equilibrium
In2.5
Blank reactor
NO reduction
C3H6 conversion for NO reduction
NO2 reduction
C3H6 converion for NO2 reduction
300
350
400
450
500
550
600
650
300
350
400
450
500
550
600
650
700
Temperature (oC)
Temperature (oC)
FIG. 3. The effect of NOx species on NOx reduction over In2.5 (0.1%
NOx , 0.1% C3H6, 9% O2, 7% H2O, SV = 30,000 h−1).
FIG. 4. NO oxidation to NO2 over In2.5 (0.1% NO, 10% O2, SV =
25,000 h−1).
525–625◦C, respectively. The addition of up to 2.5 wt% in-
dium to the alumina significantly lowered the activation
energy for propene conversion, as shown by lower reaction
temperatures. Increasing the indium beyond 2.5 to 5 and
10 wt% levels reversed this trend and increased the reac-
tion temperature, thereby increasing the activation energy.
However, notice that 100% conversion of propene still oc-
curs at the elevated temperature of 600◦C for all catalysts
tested.
tures by 150◦C. In addition, the maximum NO conversion
decreased from 60 to 42% when C3H6 was replaced with
C3H8.
Figure 6 illustrates N2 yield as a function of reaction tem-
perature for NOx reduction with propane reductant over
the indium-doped alumina catalysts. The NO reduction per-
formance for undoped alumina is also shown in Fig. 6 for
comparison. In1 and In2.5 catalysts showed NO reduction
similar to that of alumina. Increasing the indium to 10 wt%
levels decreased NOx reduction performance to levels even
lower than that of alumina. The results of NO2 reduction
with propane (Fig. 6b) over the series of indium-doped alu-
minas shows that the addition of indium oxide did not pro-
mote NOx reduction and in fact lowered the activity for
NO2 reduction.
3.1.2. The effect of NOx species. Figure 3 compares
NOx reductions and C3H6 conversions as a function of tem-
perature over the In2.5 catalyst for NO and NO2. While the
Tmax and peak shape was similar for NO and NO2 feeds, the
N2 yield from NO2 reduction was higher over the entire
temperature range. Alternatively, the propene conversion
profiles were the same with NO and NO2 feeds despite the
fact that NO2 is a stronger oxidation agent than NO.
The extent of NO oxidation reaction was measured over
the In2.5 catalyst with a feed of 0.1% NO, 10% O2, and an
SV of 25,000 h−1 (Fig. 4). The theoretical equilibrium value
of NO to NO2 conversion and the experimentally deter-
mined values in an empty reactor under the same reaction
conditions are also shown for comparison. The NO oxida-
tion activity over the In2.5 catalyst was negligible and was
similar to the blank reactor below 500◦C. The NO2 con-
centration remained below the equilibrium value even at
550◦C.
120
NO reduction with C3H6
C3H6 conversion
NO reduction with C3H8
100
C3H8 conversion
80
60
40
20
0
3.1.3. The effect of hydrocarbon species. Figure 5 com-
pares the NO reduction and hydrocarbon conversions as
a function of reaction temperature over the In2.5 catalyst
withtwohydrocarbonspecies:C3H6 andC3H8. BothNOre-
duction and hydrocarbon conversion were suppressed sig-
nificantly when C3H6 was replaced with C3H8. As shown
in Fig. 5, the NO reduction peak profile with propane as
the reductant was narrower and shifted to higher tempera-
350
400
450
500
550
600
650
700
Temperature (oC)
FIG. 5. The effect of hydrocarbon species on NO reduction over In2.5
(0.1% NO, 0.1% C3Hx , 9% O2, 7% H2O, SV = 30,000 h−1).

In2O3/Al2O3 CATALYSTS FOR NOx REDUCTION IN LEAN CONDITION
101
were analyzed. SiO2 was used to support the In2O3 because
itisexpectedtobeinertforhydrocarbonoxidationandNOx
reduction.
60
50
40
30
20
10
0
a
Al2O3
In1
In2.5
In10
Table1showsreactionproductsoverasingle-bedcatalyst
of 0.1 g In2.5/Al2O3, 0.3 g In2.5/SiO2, and 0.1 g Al2O3, as
well as over dual-bed catalysts, which the reaction mixture
passed through 0.3 g In2.5/SiO2 catalyst and then through
0.1 g Al2O3. The 0.1-g In2.5/Al2O3 catalyst demonstrated
excellentNOreduction(54%)andhydrocarbonconversion
(67%)inthepresence of9% O2 and7% H2Oata high space
velocity of 60,000 h−1
.
However, when the same loading of indium (2.5 wt%)
was introduced on SiO2 support, no N2 formation was de-
tected despite 23% propene conversion. Instead, 16 ppm
acrolein (C3H4O) and 22 ppm acetaldehyde (CH3CHO)
were observed over the In2.5/SiO2 catalyst.
525
550
575
600
625
650
Temperature (oC)
Alumina was tested at higher temperatures (500◦C) than
were In2.5/Al2O3 or In2.5/SiO2 catalysts; however, the alu-
mina converted only 4% NO to N2 and 3% propene to CO
and CO2. No acrolein or acetaldehyde was observed. This
indicates that alumina alone did not contribute significantly
to NO and propene conversions at the reaction tempera-
ture.
When the alumina was placed downstream of the
In2.5/SiO2 catalyst, NO reduction to N2 (59 ppm) increased
by a factor of 3 compared to that with alumina alone.
Acrolein and acetaldehyde produced over the In2.5/SiO2
catalyst was not observed after the alumina. The overall
propene conversion over a dual-bed catalyst was similar to
In2.5/SiO2; therefore, most of hydrocarbon was activated
over indium species.
Table 2 shows major reaction products over the
In2.5/SiO2 catalyst when each oxidant (O2, NO, or NO2)
was introduced individually in the gas stream in order to
identify the origin of oxygen in the produced acetaldehyde
and acrolein. Substantial propene conversions (>16%) and
production of acetaldehyde, acrolein, CO, and CO2 were
observed when O2 or NO2 was used as oxidation agents.
However, no oxygenated hydrocarbon was observed with
a minimum propene conversion when NO was used. The
results indicate that the oxygen atoms in the oxygenated
hydrocarbons are originally from O2.
80
60
40
20
0
b
Al2O3
In1
In2.5
In10
500
525
550
575
600
625
650
Temperature (oC)
FIG. 6. The effect of indium loading on N2 yield for (a) NO reduction
and (b) NO2 reduction with C3H8 over In2O3/Al2O3 catalysts (0.1% NO,
0.1% C3H8, 9% O2, 7% H2O, SV = 30,000 h−1).
3.1.4. The role of indium oxide. In order to elucidate
the individual roles of In2O3 and of Al2O3, an experiment
was performed using a dual-bed reactor system in which
the components were separately loaded in two catalyst beds
arranged in series and the products over each catalyst bed
TABLE 1
The Results of GC/MS Analysis for NO Reduction over In Catalysts and γ -Al2O3
Product concentration (ppm) Conversion (%)
Temperature (◦C)
==
Catalyst
N2
CO
CO2
C3H6
CH3CHO
CH2 CHCHO
C3H6
NO
In2.5/Al2O3
In2.5/SiO2
Al2O3
In2.5/SiO2 + Al2O3
272
0
21
59
469
96
75
1314
432
19
300
824
1040
811
0
22
0
0
16
0
67
23
3
54
0
4
475
475
500
475
500
332
444
0
0
24
12
Note. Reaction condition: 1000 ppm NO, 1060 ppm C3H6, 9% O2, 7% H2O, balanced with He; total flow, 200 cc/min.

102
PARK ET AL.
TABLE 2
The Results of GC/MS Analysis over an In2.5/SiO2 Catalyst
Product concentration (ppm)
==
Gas mixture
CO
CO2
C3H6
CH3CHO
CH2 CHCHO
Conversion (%)
O2 + C3H6 + He
NO + C3H6 + He
NO2 + C3H6 + He
246
0
115
286
6
265
854
1019
865
9.7
0
2.8
12.7
0
21.2
17.6
1.6
16.5
Note. Reaction condition: 940 ppm NO, 970 ppm NO2 or 9.3% O2 in the presence of 1040 ppm C3H6,
balanced with He; total flow, 200 cc/min; reaction temperature, 475◦C.
In2O3
3.2. Catalyst Characterization
Al2O3
Table 3 shows BET results of the In2O3/Al2O3 catalysts.
The surface area, pore volume, and pore diameter of the
catalystssignificantlydecreasedasafunctionofindiumcon-
tent. The decrease in surface area, pore volume, and pore
diameter was not proportional to the increase in indium
content on the alumina support. For example, for the In15
catalyst, thepercentagelossofsurfacearea(37%), porevol-
ume (69%), and pore diameter (52%) based on the alumina
was more than the percentage increase in indium loading
(15%).
In20
In10
In5
In2.5
Figure7showspowderXRDpatternsoftheIn2O3/Al2O3
catalysts. The diffraction patterns for samples containing
up to In2.5 are characteristic of the alumina support. How-
ever, XRD patterns obtained for samples with indium load-
ings higher than 5 wt% contained peaks characteristic of
In2O3 in addition to alumina peaks. The intensities of the
In2O3 XRD peaks increased with increasing indium con-
tent. XRD patterns that were measured after the same cata-
lystswereusedforleanNOx reactionsshowednodifference
between fresh and used catalysts (not shown). Therefore,
exposure to reaction conditions did not affect the In2O3
crystalline phase or the alumina support.
In1
Al2O3
0
20
40
60
80
100
Degrees (2θ)
FIG. 7. XRD patterns of In2O3/Al2O3 catalysts.
Figure 8 shows the variation of the XPS In3d5/2/Al2p
intensity ratio measured for the In2O3/Al2O3 catalysts as
a function of In/Al atomic ratio. The intensity ratio of
In3d5/2/Al2p calculated for theoretical atomic dispersion is
also shown for comparison. The theoretical value was cal-
4
theoretical monolayer line
3
2
1
0
TABLE 3
Surface Area, Pore Volume, and Pore Diameter of In-Doped
Alumina Catalysts
Surface area
(m2/g)
Pore volume
(cc/g)
Pore diameter
(nm)
Catalysts
Al2O3
In1
In2.5
In5
In10
In15
230
223
209
188
170
146
1.1
14
10
11.3
9.2
7.9
6.7
0.75
0.79
0.58
0.46
0.34
0.00
0.02
0.04
0.06
0.08
0.10
0.12
Atomic Ratio (In/Al)
FIG. 8. XPS In3d5/2/Al2p intensity ratios measured for In2O3/Al2O3
catalysts. The theoretical monolayer line is included for comparison.

In2O3/Al2O3 CATALYSTS FOR NOx REDUCTION IN LEAN CONDITION
103
sions of indium species were determined by comparing the
XPS intensity ratios obtained from the experiments
with the values calculated from the theoretical monolayer
line. The XPS results indicated that the dispersion of in-
dium species was 100% for In10. However, the dispersion
decreased at higher-indium-loading catalysts: 48 and 42%
dispersion for In15 and In20, respectively.
Figure 9 shows temperature-programmed reduction
spectra for the In2O3/Al2O3 catalysts. Detectable H2 up-
take was observed at 300 and 675–700◦C for In1 and In2.5
catalysts. For the higher-indium-loading catalysts (In5 and
In10), an additional H2 uptake peak was observed in the
temperature range of 550–600◦C. The H2 consumption in-
creased as a function of indium loading. From the compar-
ison with the TPR spectra of standard mixtures of In2O3
and alumina, the H2 uptake peaks around 550–600◦C can
be assigned to the reduction of the large-grained In2O3
crystalline phase. The H2 uptake peak at 300◦C was as-
signed to the reduction of the highly dispersed indium ox-
ide species. The calculation of the amount of H2 update
for each catalyst indicated that 100% of the indium oxide
species was reduced to indium metal up to In5. In10 showed
83% of indium oxide to be reduced to the metal during TPR
measurements.
Figure 10 shows SEM backscattered images of the
In2O3/Al2O3 catalysts. At indium loadings of less than
2.5 wt%, only the alumina was imaged with the SEM. How-
ever, EDS quantitative analysis verified that the concen-
tration of indium was consistent with as-prepared values.
At 5 wt% indium loading, well-dispersed particles of in-
dium smaller than 20 nm in diameter (lighter phase) were
FIG. 9. TPR spectra of In2O3/Al2O3 catalysts.
culated using the model proposed by Kerkhof and Moulijn
(10). The photoelectron cross sections and the mean escape
depth of the photoelectrons used in these calculations were
taken from Scofield (11) and Penn (12), respectively.
The In/Al intensity ratio increased linearly with increas-
ing In/Al atomic ratio up to 0.049 (In10), consistent with a
theoretical monolayer. For higher-indium-loading catalysts
(In15 and In20), the In/Al intensity ratios deviated from
the linear trend. The In/Al intensity ratios of the catalysts
which were tested for the lean NOx reaction showed sim-
ilar values and a similar trend within the range of error.
Therefore, exposure to reaction conditions did not affect
the indium dispersion over the alumina support. The disper-
FIG. 10. SEM backscattered images of In2O3/Al2O3 catalysts.

104
PARK ET AL.
imaged on the surface of the alumina. The size of the in- change in catalytic performance for different indium load-
dium species increased with increasing indium content in ings due to the reduction of alumina sites would be difficult
the sample. At 15 wt% indium loading, large (200 nm)- to observe.
facetedcubicindiumoxidecrystalliteswereobservedonthe
surface.
The formation of NO2 and its subsequent reaction with
hydrocarbons has been demonstrated to be a key reaction
mechanism in lean NOx catalysis (15–17). It is expected that
NO2 should enhance hydrocarbon conversion compared to
NO because NO2 is a stronger oxidizer. However, C3H6
conversion over In2.5 is unchanged when NO is replaced
with NO2 in the gas stream (Fig. 3). This observation sug-
gests that NO2 does not play a critical role in hydrocarbon
activation over this catalyst. Indeed, there is enhancement
of NOx reduction over the entire temperature range when
NO is replaced with NO2. However, this enhancement can
be attributed to activity on the Al2O3 sites because pure
Al2O3 is known to be active in promoting the reduction of
NO2 to N2 by C3H6 (9). Al2O3 sites should be readily avail-
able in the In2.5 catalyst due to the low metal loading. Poor
NO oxidation to NO2 over the In2.5 catalyst (Fig. 4) also
suggests that NO2 plays a minimum role in enhancing cata-
lytic NOx reduction over the In2.5 composition compared
to undoped alumina.
Indium-doped alumina catalysts were further examined
with propane in order to understand an essential step in
the lean NOx process (Figs. 5 and 6). The NOx reduction
and hydrocarbon oxidation activity of the In2.5 catalyst was
comparedusingpropaneandpropeneasreductants(Fig. 5).
ThelowerleanNOx activityforpropaneovertheIn2.5cata-
lyst indicates that the catalyst does not promote NO reduc-
tion with C3H8. With increasing indium loading, the NO re-
duction performance with C3H8 decreased to values even
lower than those observed for the alumina support (Fig. 6a).
This result indicates that the indium oxide does not utilize
propane for NOx reduction and that it only blocks active
sites on the alumina. When NO is replaced with NO2 with
a propane reductant, the catalytic performance did not im-
prove (Fig. 6b). It has been reported that the NOx reduction
over In/H-ZSM-5 catalyst with methane was affected by the
species of NOx and NO2 was involved in the important re-
action mechanism (5). The results obtained in this study
indicate that type of hydrocarbon reductant is more critical
than NOx species to achieve high catalytic performance for
NOx reduction over the In2O3/Al2O3 catalysts.
4. DISCUSSION
The NOx reduction activity of the most active compo-
sition identified in this study, In2.5/Al2O3, is among the
highest reported for the relatively aggressive reaction con-
ditions used in this study (7% H2O, 9% O2, and space ve-
locity of 30,000 h−1). The lean NOx performance of the
catalysts was strongly dependent on indium loading simi-
lar to other lean NOx catalysts reported in the literature
(Fig. 1). For example, NO conversions over CoOx /Al2O3
and AgO/Al2O3 catalysts decrease with increasing metal
loading(13, 14). Typically, highmetalormetaloxideloading
accelerates hydrocarbon oxidation and depletes the reduc-
tant, which in turn decreases the catalytic NOx reduction
performance. However, In2O3/Al2O3 catalysts examined in
this study show both a decrease in propene oxidation and a
NO reduction at higher indium loadings (Fig. 2). The cata-
lytic performance and characterization results suggest that
small, amorphous indium oxide clusters which are present
at lower loadings promote NOx reduction and hydrocarbon
oxidation, while larger, crystalline In2O3 observed above
loadings of 5 wt% (Fig. 7) are less effective for these reac-
tions. As shown in Table 3, alumina parameters such as sur-
facearea, porevolume, andaverageporediameterdecrease
significantly with increasing indium loading. While the sur-
face area of the metal oxide is expected to be very small
compared to that of the alumina substrate, the observed
decrease in the physical properties is larger than predicted
by a simple rule of mixtures. This indicates that much of
the previously accessible alumina surface area and porosity
is blocked by the indium oxide. Therefore, the decrease in
catalytic NOx reduction performance at high indium oxide
loadings can be attributed to the reduction in active alu-
mina sites rather than to the acceleration of hydrocarbon
oxidation.
In contrast to the findings in this study, Hamada and
co-workers found relatively little effect of indium load-
ing on NO reduction with C3H6 (7, 8). They observed no
significant change in the maximum NO conversion for a
range of indium loadings on alumina. However, as shown
in Fig. 1, in our study an optimum indium loading (2.5 wt%
indium on alumina) was clearly observed for both NO re-
duction and C3H6 conversion. The discrepancy between the
two studies can be explained by the difference in reaction
conditions. In Hamada’s experiments, the contact time was
0.18 g · s−1 · cm−3, which is three times longer than that used
in our work (0.06 g · s−1 · cm−3). Due to the relatively mild
reaction conditions used in Hamada’s study, a significant
The role of In2O3 in hydrocarbon activation was fur-
ther proved by quantitative analysis using GC mass spec-
troscopy over In2O3 on an inert substrate (SiO2). In these
experiments, acrolein and acetaldehyde were detected in
the exhaust stream after the In2.5/SiO2 catalyst. An alu-
minacatalystdownstreamthenconvertedtheseoxygenated
hydrocarbons in a reaction with NO to make N2 and CO
(Table 1). The oxygenated hydrocarbons have also been
reported as key intermediates in the selective reduction of
NOx over SnO2/Al2O3 catalysts (18). The authors proposed
a bifunction mechanism in which SnO2 converted propene

In2O3/Al2O3 CATALYSTS FOR NOx REDUCTION IN LEAN CONDITION
105
to the oxygenated intermediates and Al2O3 reduced NOx mum NOx reduction activity was achieved with 2.5 wt%
to N2 utilizing these intermediates as reductants. In this indium loadings on alumina (230 m2/g) prepared by a sol–
study, the same bifunction mechanism is operative over gelmethod. Thisstudyshowedthatitisnecessarytobalance
In2O3/Al2O3 catalysts. Observation of the formation of the well-dispersed indium species and alumina active sites
acrolein and acetaldehyde from hydrocarbon reductants in order to achieve the optimum catalytic performance. A
verifies that partial oxidation of the hydrocarbon is more bifunction mechanism where the indium oxide species par-
essential for the promotion of NOx reduction over indium- tially oxidizes propene to acrolein and acetaldehyde and
doped alumina catalysts than is the oxidation of NO to NO2. the alumina utilizes the oxygenated hydrocarbons to re-
XRD, XPS, and TPR results showed that the dispersion duce NOx to N2 was identified. Characterization results in-
as well as the phase of indium oxide changed as a function dicated that well-dispersed and readily reducible indium
of indium loading. The XPS peak intensity ratio of In3d5/2
/
oxide clusters are the active sites for converting propene
Al2p suggested the indium oxide species were atomically to the oxygenated hydrocarbons. However, the lean NOx
dispersed up to concentrations as high as 10 wt%. However, performance decreases at high indium loadings (>5 wt%)
crystalline In2O3 was observed in compositions containing where active alumina sites responsible for NOx reduction
more than 5 wt% indium, indicating that the dispersion was are blocked by well-dispersed indium oxide species.
compromised. This can be explained by considering the na-
ture of indium oxide species that can contribute to both
XPS and XRD intensities. SEM backscattered images of
ACKNOWLEDGMENTS
the catalysts (Fig. 10) showed that dispersed indium clusters
The research described in this paper was performed in part at the W. R.
and large crystalline indium species were present together
on the In5 and In10 catalysts. TPR results confirmed the
presence of two different indium oxide species on the alu-
mina support for compositions containing greater than 2.5
wt% indium. For the In5 and In10 catalysts, the presence
of the large high-temperature H2 uptake (550◦C) is consis-
tent with XRD data showing that large crystalline In2O3 is
present in these samples. The low-temperature H2 uptake
(300◦C) was assigned to the well-dispersed indium oxide
species. The well-dispersed phase increased with increas-
ing indium loading. This finding was consistent with the
XPS intensity ratio that increased as a function of indium
loading up to the In10 catalyst. XPS and TPR results con-
cluded that the atomically dispersed indium oxide species
increased with indium loading up to the In10 catalyst; how-
ever, crystallites of In2O3 are also present in the catalysts
containing higher than 5 wt% indium.
Wiley Environmental Molecular Sciences Laboratory, a national scien-
tific user facility sponsored by the U.S. Department of Energy’s Office of
Biological and Environmental Research, located at the Pacific Northwest
National Laboratory in Richland, WA. A part of the study was financially
supported by the Office of Heavy Vehicle Technologies in the U.S. De-
partment of Energy. The authors also acknowledge Prof. Harold Kung
and Prof. Mayfair Kung at Northwestern University for their comments
as well as technical support.
REFERENCES
1. Held, W., Konig, A., Richter, T., and Puppe, L., SAE Paper 900496
(1990).
2. Iwamoto, M., Yahiro, H., Yuu, Y., Shundo, S., and Mizuno, N.,
Shokubai (Catalyst) 32, 430 (1990).
3. Miyadera, T., and Yoshida, K., Chem. Lett. 1483 (1993).
4. Ogura, M., and Kikuchi, E., Chem. Lett. 1017 (1996).
5. Kikuchi, E., Ogura, M., Aratani, N., Sugiura, Y., Hiromoto, S., and
Yogo, K., Catal. Today 27, 35 (1996).
6. Ogura, M., Hayashi, M., and Kikuchi, E., Catal. Today 45, 139
(1998).
7. Haneda, M., Kintaichi, Y., and Hamada, H., Catal. Lett. 55, 47
(1998).
8. Maunula, T., Kintaichi, Y., Inaba, M., Haneda, M., Sato, K., and
Hamada, H., Appl. Catal. B 15, 291 (1998).
9. Park, P. W., Kung, H. H., Kim, D.-W., and Kung, M. C., J. Catal. 184,
440 (1999).
10. Kerkhof, F. P. J. M., and Moulijn, J. A., J. Phys. Chem. 83, 1612 (1979).
11. Scofield, J. H., J. Electron. Spectrosc. Relat. Phenom. 8, 129 (1976).
12. Penn, D. R., J. Electron Spectrosc. Relat. Phenom. 9, 29 (1976).
13. Yang, J., Kung, M. C., Sachtler, W. M. H., and Kung, H. H., J. Catal.
172, 178 (1997).
14. Bethke, K. A., and Kung, H. H., J. Catal. 172, 93 (1997).
15. Sasaki, M., Hamada, H., Kintaichi, Y., and Takehiko, I., Catal. Lett.
15, 297 (1992).
16. Bethke, K. A., Li, C., Kung, M. C., Yang, B., and Kung, H. H., Catal.
Lett. 31, 287 (1995).
17. Lukyanov, D. B., Sill, G., d’Itri, J. L., and Hall, W. K., J. Catal. 153, 265
(1995).
The correlating NOx performance results and character-
ization data indicate that the well-dispersed indium oxide
species promotes the lean NOx reaction via partial oxida-
tion of hydrocarbons. However, high concentrations of the
well-dispersed indium species on the surface decrease the
NOx reduction activity by blocking active alumina sites.
Therefore, it is important to optimize the indium and alu-
mina sites to achieve the best catalytic performance over
this class of catalysts.
5. CONCLUSION
Indium (1–10 wt% In) supported sol–gel alumina cata-
lysts converted as much as 60% NO and 70% NO2 to N2
in lean exhaust conditions (1000 ppm NOx , 9% O2, 7%
H2O, and 30,000 h−1 space velocity) with 1000 ppm propene
as a reductant. However, with propane as the reductant
the activity was diminished to 42% NO conversion. Opti-
18. Yezerets, A., Zheng, Y., Park, P. W., Kung, M. C., and Kung, H. H.,
Stud. Surf. Sci. Catal. 130, 629 (2000).