Selective catalytic oxidation of ammonia to nitrogen over Fe2O3-TiO2 prepared with a sol-gel method

Article abstract of DOI:10.1006/jcat.2002.3545

Fe₂O₃-Al₂O₃, Fe₂O₃-TiO₂, Fe₂O₃-ZrO₂, and Fe₂O₃-SiO₂ were prepared with a sol-gel method and showed high activities for selective catalytic oxidation (SCO) of ammonia to nitrogen in the presence of excess oxygen. The catalysts prepared from iron sulfate were more selective for N₂ than those prepared from nitrate. The N₂ selectivity for the SCO reaction was in good agreement with their surface acidity and SCR activity. More than 92% of N₂ yields were obtained on the 10% Fe₂O₃-TiO₂ SO₄/^2- and 20% Fe₂O₃-TiO₂ (SO₄/^2-) at 400°-450°C under the condition of GHSV = 2 × 10⁵/hr. The gaseous, adsorbed, and lattice oxygen all participated in the SCO reaction.

Full text of DOI:10.1006/jcat.2002.3545

Journal of Catalysis 207, 158–165 (2002)
doi:10.1006/jcat.2002.3545, available online at http://www.idealibrary.com on
Selective Catalytic Oxidation of Ammonia to Nitrogen over Fe O –TiO
2
3
2
Prepared with a Sol–Gel Method
R. Q. Long and R. T. Yang1
Department of Chemical Engineering, University of Michigan, Ann Arbor, Michigan 48109-2136
Received May 17, 2001; revised January 30, 2002; accepted January 30, 2002
4
N2 + 6H2O). The commercial catalysts that are used to-
Fe2O3–Al2O3, Fe2O3–TiO2, Fe2O3–ZrO2, and Fe2O3–SiO2 were day are V2O5 + WO3 (or MoO3) supported on TiO2 (12).
prepared with a sol–gel method and they showed high activities In order to control ammonia slip, most processes are car-
for selective catalytic oxidation (SCO) of ammonia to nitrogen in
the presence of excess oxygen. The Fe2O3–TiO2 catalysts prepared
from iron sulfate yielded a higher selectivity for N2 than those pre-
ried out under the conditions such that NH3/NO < 1 (≈0.9),
which results in a decrease in NO reduction efficiency.
For improving the NO reduction efficiency, the use of a
pared from nitrate. More than 92% of N2 yields were obtained on the
stoichiometric or excess amount of ammonia is desirable.
The SCO of ammonia can be applied to the SCR of NO
in a secondary bed to oxidize the residual ammonia to
N2, without introducing other reactants into the gas mix-
ture. The SCO process can also be applied to the com-
bustion of biomass-derived gases for removing the NH3
2
−
2−
1
4
0wt%Fe2O3–TiO2 (SO4 )and20wt%Fe2O3–TiO2 (SO )at400–
4
◦
5
−1
50 C under the condition of GHSV = 2.0 × 10 h . Also, after
the Fe2O3–TiO2 prepared from nitrate was treated with SO2 + O2
◦
at 450 C, the N2 selectivity and yield were enhanced significantly
◦
at 450–500 C, suggesting a promoting role by SO2. But H2O de-
creased the N2 yield slightly. The N2 selectivity for the SCO reac-
tion is in good agreement with their surface acidity and the activity impurity (3, 4).
for selective catalytic reduction (SCR) of NO with ammonia. This
Several types of materials have been reported to be
further supports the two-step SCO mechanism in which NH3 is active for SCO of ammonia to N , such as Pt, Rh, and Pd
2
first oxidized to NO and then NO is reduced to N2 by unreacted
exchanged to ZSM-5 (2); Ni, Fe, and Mn oxides supported
on γ -Al2O3 (3, 4), V2O5/TiO2, CuO/TiO2, and Cu–ZSM-5
(5); CuO/Al O (6, 7); Cu–Mn/TiO (8); Fe-exchanged
2 3 2
TiO2-pilllared clay (9); and Fe-exchanged ZSM-5 and
other zeolites (10, 11). These catalysts exhibited activities
for N2 formation under various conditions. Amblard et al.
NH3 adsorbed species through a SCR reaction. The presence of sul-
2
4
−
fate species on Fe2O3–TiO2 (SO ) increased surface acidity and
thus improved SCO performance. Temperature-programmed des-
orptionandtemperature-programmedsurfacereactionofammonia
showed that gaseous, adsorbed, and lattice oxygen may participate
in the SCO reaction. ꢀc 2002 Elsevier Science (USA)
Key Words: selective catalytic oxidation (SCO) of NH3; selective
catalytic reduction (SCR) of NO; Fe2O3–TiO2; sol–gel.
(
3) reported that among transition metal oxides supported
on γ -Al2O3, Ni/Al2O3, Mn/Al2O3, and Fe/Al2O3 were the
most active and selective catalysts for the SCO reaction. In
our previous work (10), we studied the SCO reaction on a
series of transition-metal (Cr, Mn, Fe, Co, Ni, Cu, and Pd)
ion-exchanged zeolites. Results showed that the catalytic
performance (i.e., NH3 conversion and N2 selectivity)
increased in the trend of Co–ZSM-5 ≈ Ni–ZSM-5 <
Mn–ZSM-5 < H–ZSM-5 < Pd–ZSM-5 < Cr–ZSM-5 < Cu–
ZSM-5 < Fe–ZSM-5 at a high gas hourly space velocity
INTRODUCTION
The removal of ammonia from waste streams is becom-
ing an increasingly important problem. It is known that
many chemical processes use reactants containing ammo-
nia or produce ammonia as a by-product. They are all
plagued with ammonia slip problem. Selective catalytic oxi-
dation (SCO) of ammonia to nitrogen is potentially an ideal
technology for removing ammonia from oxygen-containing
waste gases and consequently it has become of increasing
interest in recent years (1–11). Moreover, ammonia is used
effectively in power plants for NOx (x = 1, 2) abatement by
selective catalytic reduction (SCR, 4NH3 + 4NO + O2 =
5
−1
(
GHSV = 2.3 × 10 h ). In particular, near 100% of NH3
◦
conversion to N2 was obtained at 450 C on the Fe–ZSM-5.
H2O and/or SO2 decreased the NH3 conversion only
slightly (10). Also, we found that there existed a good
correlation between the N2 selectivity for the SCO reaction
and the activity for the SCR of NO with ammonia for the
Fe-exchanged zeolites, i.e., the higher the SCR activity, the
higher the N2 selectivity in the SCO (11). This supported
the SCO mechanism involving NO as an intermediate for
1
To whom correspondence should be addressed. E-mail: yang@
umich.edu.
2
N formation.
158
0
ꢀ
021-9517/02 $35.00
c 2002 Elsevier Science (USA)
All rights reserved.

CATALYTIC OXIDATION OF AMMONIA TO NITROGEN OVER Fe2O3–TiO2
159
In this work, we studied the SCO performance over
Fe2O3–TiO2 mixed oxides. The catalysts were prepared
through a sol–gel route, using iron nitrate and iron sul-
fate as iron precursors. It is known that the sol–gel tech-
nique can result in a better iron dispersion than conven-
tional impregnation methods. Also, the effect of water
on the SCO performance and the correlation between
SCO performance (N2 selectivity) and SCR activity were
studied. Temperature-programmed desorption (TPD) and
temperature-programmed surface reaction (TPSR) were
used to investigate ammonia adsorption and the reaction
between ammonia adsorbed species and oxygen.
EXPERIMENTAL
Preparation of catalyst. The mixed oxides were pre-
pared by a one-step sol–gel technique. In each case, a cer-
tain amount of iron nitrate or sulfate was first dissolved
in 70 ml of methanol (1.7 mol) and then 0.05 mol of tita-
nium butoxide (or zirconium butoxide, aluminum butox-
ide, and silicon ethoxide) was added into the solution with
stirring. Subsequently, 3 ml of a 0.2 M HNO3 solution was
added dropwise into the mixture and stirred for hydroly-
sis and gelation. After the gelation was completed, the gel
was aged for 2 days at room temperature and then cal-
cined at 500 C for 6 h in flowing O2 (150 ml/min). Finally, (d) 1% Fe2O3–TiO2 (SO4 ) catalysts at room temperature.
the samples were crushed and sieved to 60–100 mesh. The
FIG. 1. FTIR spectra of sulfate species on (a) 20, (b) 10, (c) 5, and
◦
2−
Fe2O3 amounts in the mixed oxides were controlled at 1–
tral resolution of 4 cm−1, using the empty cell as back-
ground.
2
0% by weight. Since significant amounts of sulfate species
still existed on the catalysts that were prepared from iron
sulfate (see Fig. 1), the catalysts are expressed as Fe2O3–
SCO performance. The SCO activity measurement was
carried out in a fixed-bed quartz reactor. The reaction
temperature was controlled by an Omega (CN-2010) pro-
grammable temperature controller, and in this work 0.2 g
catalyst was used. The reactant gas was obtained by blend-
ing different gas flows. The typical reactant gas composition
2
4
−
TiO2 (SO ). Fe(NO3)3 · 9H2O (98%), Fe2(SO4)3 · 7H2O
(
98%), methanol (99.9%), and aluminum butoxide (97%)
were supplied from Aldrich. Titanium butoxide (98+%),
zirconium butoxide (78% in n-butanol), and silicon ethox-
ide (98%) were obtained from Strem Chemicals.
Characterization of catalyst. A Micromeritics ASAP was as follows: 1000 ppm NH , 2% O , 2.5% H O (when
3
2
2
2
010 micropore size analyzer was used to measure the N2 used), and balance He. The total flow rate was 500 ml/
adsorption isotherms of the samples at liquid N2 temper- min (ambient conditions) and thus a high gas hourly space
◦
5
−1
ature (−196 C). The specific surface area was determined velocity (GHSV = 2.0 × 10 h ) was obtained. Premixed
fromthelinearportionoftheBETplot(P/P0 = 0.05–0.20). 1.07% NH in He (with 0.05% H O as impurity) was
3
2
The pore size distribution was calculated from the des- supplied by Matheson. Water vapor was generated by
orption branch of the N2 adsorption isotherm using the passing He through a gas-wash bottle containing deion-
Barrett–Joyner–Halenda (BJH) formula. Prior to the sur- ized water. A magnetic-deflection-type mass spectrome-
face area and pore size distribution measurements, the sam- ter (AERO VAC , Vacuum Technology, Inc.) was used
ples were degassed in vacuum at 350 C for 6 h.
TM
◦
to monitor continuously the effluent gas from the reactor,
The mixed oxides were also investigated by FTIR spec- which contained NH (m/e = 17 minus the contribution
3
troscopy on a Nicolet Impact 400 FTIR spectrometer with of H O), H O (m/e = 18), N (m/e = 28), NO (m/e = 30),
2
2
2
a TGS detector. In each experiment, the catalyst was first O (m/e = 32), and N O (m/e = 44). NO (m/e = 46) was
2
2
2
mixed with KBr at a ratio of 1/9 (by weight). A self support- not detectable with this mass spectrometer. The concentra-
ing wafer with a 1.3-cm diameter was prepared by press- tions of NH and formed NO were also continually mon-
3
x
ing 30 mg of the mixture and was then loaded into an itored with a chemiluminescent NO/NO analyzer (Model
x
IR cell with BaF2 windows. The spectra were recorded at 42C, Thermo Environmental Instruments, Inc.), in which a
room temperature by accumulating 100 scans at a spec- high-temperature converter converted NH to NO by the
3
x

1
60
LONG AND YANG
reaction NH3 + O2 → NOx + H2O. The NH3 conversion
was calculated by ([NO] + 2[N2] + 2[N2O])/[NH3]0 ×
TABLE 1
Characterization of the Catalysts
1
00%, where [NH3]0 is the initial NH3 concentration. The
BET surface Pore volume Average pore
selectivity is defined as the percentage conversion of amm-
onia to N2, N2O, and NO. The data were collected when
the SCO reaction reached the steady state, typically after
2
3
Sample
area (m /g)
(cm /g)
diameter (nm)
5
% Fe2O3–Al2O3
245
104
106
29
97
98
1.10
0.14
0.14
0.04
0.18
0.14
0.19
0.14
11.9
3.7
3.8
6.0
4.6
3.9
3.6
3.9
2
0 min at each temperature.
5% Fe2O3–TiO2
% Fe2O3–ZrO2
5% Fe O –SiO
5
SCR performance. The SCR activity measurement was
2
3
2
2
−
carried out in the same fixed-bed quartz reactor. The reac-
tion conditions were as follows: 0.2 g of catalyst, 1000 ppm
NO, 1000 ppm NH3, 2% O2, and balance He. The total flow
rate was 500 ml/min (ambient conditions). Premixed 1.01%
NO/He was supplied by Matheson. The NO and NO2 con-
centrations were continually monitored by the chemilumi-
1% Fe2O3–TiO2 (SO4 )
% Fe2O3–TiO2 (SO2
−
−
−
5
)
)
)
4
2
1
2
0% Fe2O3–TiO2 (SO4
0% Fe2O3–TiO2 (SO2
106
92
4
nescent NO/NOx analyzer. To avoid errors caused by the nated to Fe and/or Ti sites (13). This indicates that signifi-
oxidation of ammonia in the converter of the NO/NOx ana- cant amounts of sulfate species still existed on the catalysts
◦
lyzer, an ammonia trap containing phosphoric acid solution after calcination at 500 C. Since sulfated TiO2 and sulfated
was installed before the sample inlet to the chemilumines- Fe2O3 show similar IR bands at 1210–1230, 1130–1150, and
−
1
cent analyzer. All the data were obtained after 20 min, when 1030–1060 cm , it is difficult to determine to which sites
the SCR reaction reached steady state.
the sulfate species were bonded by IR spectroscopy (13).
However, our previous work indicated that Fe2O3 was sul-
fated more easily than TiO2 (14). Also the sulfate amount
decreased with decreasing Fe2O3 content. Hence it is rea-
TPD and TPSR analyses. TPD and TPSR experiments
were carried out in a fixed-bed quartz reactor. Before the
experiment, 0.5 g of sample was pretreated in He at 500 C
◦
sonable to conclude that the sulfate species were bonded
for 30 min to remove adsorbed H2O and other gases. After
the sample was cooled to room temperature, the He flow
was switched to a flow of 1.07% NH3/He for 60 min (100 ml/
min). Mass spectra showed that the intensity of ammonia
had stabilized. The sample was then purged by He (200 ml/
min) for another 30 min. For the TPD experiment, He was
passed through the reactor and ammonia TPD was per-
formed by ramping the temperature at 10 C/min to 500 C.
For the TPSR experiments, subsequent to the ammonia
adsorption step, the He flow was switched to 5% O2/He
2−
mainly to iron sites of Fe2O3–TiO2 (SO ). By comparison,
4
−1
no IR bands at 1500–1300 cm due to nitrate species were
observed on the Fe2O3-containing mixed oxides that were
prepared with iron nitrate. Iron nitrate was decomposed to
iron oxide in the process of calcination.
SCO performance on Fe2 O3-containing catalysts. The
SCO performance of Fe2O3-containing catalysts prepared
with iron nitrate is summarized in Table 2. Under the condi-
◦
◦
5
−1
tions of 1000 ppm NH3, 2% O2, and GHSV = 2.0×10 h
,
the mixed oxides showed various catalytic performance at
different temperatures. NH3 conversion increased with re-
action temperature and all of the catalysts showed very high
NH3 conversions at high temperatures. In all cases, N2, NO,
and H2O were the products for ammonia oxidation. N2O
formation was not observed. At low temperatures, N2 was
the main product. With increasing temperature, N2 selec-
tivity decreased, whereas NO selectivity increased signi-
ficantly. The maximum N2 yield on the Fe2O3-containing
catalysts decreased in the order of Fe2O3–Al2O3 > Fe2O3–
TiO2 > Fe2O3–ZrO2 > Fe2O3–SiO2.
(
200 ml/min). At the same time, the reactor was heated lin-
◦
◦
early at 10 C/min to 500 C. The magnetic-deflection-type
mass spectrometer was used to monitor continuously the
effluent gas from the reactor, which contained NH3, H2O,
N2, NO, O2, and N2O.
RESULTS
Main characteristics of catalysts. The BET surface area,
pore volume, and pore size of Fe2O3-containing mixed ox-
ides are summarized in Table 1. The mixed oxides prepared
with the sol-gel method showed high surface areas, except
2
4
−
SCOperformanceon Fe2 O3–TiO2(SO )catalysts. The
Fe2O3–SiO2. The surface area and pore volume decreased SCO performance of 1–20% Fe2O3–TiO2 prepared from
in a sequence of Fe2O3–Al2O3 > Fe2O3–TiO2, Fe2O3– iron sulfate is summarized in Table 3. These catalysts also
ZrO2 > Fe2O3–SiO2. The average pore diameters in these showed near 100% NH3 conversions at high temperatures.
catalysts were 3.6–11.9 nm.
N2 and NO were the oxidation products. By comparison,
The IR spectra of the Fe2O3–TiO2 prepared with iron these catalysts showed much higher N2 selectivity than
sulfate are shown in Fig. 1. Four IR bands were observed those prepared from iron nitrate (Table 2). This was re-
−1
−1
at 1620, 1218, 1135, and 1039 cm . The 1620-cm band lated to the fact that sulfate species are present on the cata-
is assigned to adsorbed H2O on the surface, whereas the lyst, as shown by the above IR spectra (Fig. 1). This point
other bands are attributed to the bidentate sulfate coordi- has been verified by a separate experiment. When the 5%

CATALYTIC OXIDATION OF AMMONIA TO NITROGEN OVER Fe2O3–TiO2
161
TABLE 2
Fe O –TiO prepared from iron nitrate was treated with
2
3
2
◦
5
00 ppm SO2 + 10% O2 (250 ml/min) at 450 C for 1 h, N2
Catalytic Performance of 5% Fe2O3-Containing Catalysts for
◦
a
selectivity increased from 62–76% to 90–92% at 450–500 C
SCO of NH3
(
with a slight decrease in NH3 conversion from 97–99% to
87–98%). This results in an increase in N2 yield from 61–
4% to 80–89%, suggesting a promoting role by sulfation
NH3
conversion
(%)
N2
Selectivity (%)
Temperature
yield
NO (%)
7
◦
Catalyst
( C)
N2
2
4
−
for N2 formation. For Fe2O3–TiO2 (SO ), with increasing
Fe2O3–TiO2
300
39
65
89
97
99
91
78
74
76
62
9
22
26
24
38
35
51
66
74
61
Fe2O3 content (also sulfate content) from 1% to 10%, N2
selectivity and yield increased significantly (Table 3). Fur-
ther increase in iron and sulfur contents did not increase N2
selectivity and yield at high temperatures. More than 92%
3
4
4
5
50
00
50
00
2
4
−
of N2 yields were obtained on the 10% Fe2O3–TiO2 (SO
)
Fe2O3–ZrO2
Fe2O3–Al2O3
Fe2O3–SiO2
300
34
62
83
92
95
94
80
74
77
61
6
20
26
23
39
32
50
61
71
58
2
−
◦
and 20% Fe2O3–TiO2 (SO ) at 400–450 C.
3
4
4
5
50
00
50
00
4
Effect of H2O on SCO performance. It is known that the
waste streams usually contain water vapor. We further stud-
ied the effects of H2O on the SCO performance of Fe2O3–
300
38
68
85
96
97
93
80
78
81
71
7
20
22
19
29
35
54
66
78
69
2
4
−
TiO2 (SO ) catalysts (Table 4). When 2.5% H2O was
3
4
4
5
50
00
50
00
added to the reactants, it can be seen by comparing Tables 3
and 4 that N selectivities were almost unchanged at low
2
temperatures, but decreased slightly at high temperatures.
300
24
47
84
95
98
94
79
70
65
51
6
21
30
35
49
23
37
59
62
50
The NH3 conversion was increased by water vapor at low
temperatures. Overall, the N2 yield decreased slightly in the
presence of H2O. Also, in the presence of water vapor, the
maximum N2 yield increased with increasing iron and sulfur
3
4
4
5
50
00
50
00
2
4
−
contents over the 1–20% Fe2O3–TiO2 (SO ) catalysts.
a
Reaction conditions: 0.2 g of catalyst, [NH3] = 1000 ppm, [O2] = 2%,
5
−1
.
He = balance, total flowrate = 500 ml/min, and GHSV = 2.0 × 10 h
TABLE 4
TABLE 3
2−
4
Effect of H2O on Catalytic Performance of Fe2O3–TiO2(SO
)
a
2
−
Catalysts for SCO of NH3
Catalytic Performance of Fe2O3–TiO2(SO ) Catalysts
4
a
for SCO of NH3
Selectivity
NH3
N2
(
%)
Selectivity
NH3
Temperature conversion
yield
NO (%)
N2
yield
NO (%)
(
%)
◦
Catalyst
( C)
(%)
N2
Temperature conversion
◦
Catalyst
( C)
(%)
N2
1
% Fe2O3–TiO2 (SO2
−
−
−
)
)
)
)
300
350
44
87
97
99
99
91
75
74
73
51
9
25
26
27
49
40
65
72
72
50
4
% Fe2O3–TiO2 (SO2
−
1
)
)
300
37
83
98
99
99
91
78
80
75
52
9
22
20
25
48
34
65
78
74
51
4
4
4
5
00
50
00
3
4
4
5
50
00
50
00
2
4
5% Fe O –TiO (SO
2
3
2
300
32
78
93
96
99
94
83
85
77
59
6
17
15
23
41
30
65
79
74
58
−
3
4
4
5
50
00
50
00
5
% Fe2O3–TiO2 (SO2
300
19
62
98
99
99
95
86
91
77
59
5
14
9
23
41
18
53
89
76
58
4
3
4
4
5
50
00
50
00
2
4
10% Fe O –TiO (SO
2
3
2
300
350
400
450
500
36
51
97
99
99
95
92
86
71
65
5
8
14
29
35
34
47
83
70
64
−
1
2
0 % Fe2O3–TiO2 (SO2
)
300
17
51
97
99
99
94
92
96
95
72
6
8
4
5
28
16
47
93
94
71
4
3
4
4
5
50
00
50
00
2−
2
0% Fe2O3–TiO2 (SO4
300
350
400
450
500
21
52
94
99
99
97
89
87
87
71
3
11
13
13
29
20
46
82
86
70
0% Fe2O3–TiO2 (SO2
−
)
300
18
52
98
99
99
94
93
94
93
72
6
7
6
7
28
17
48
92
92
71
4
3
4
4
5
50
00
50
00
a
Reaction conditions: 0.2 g of catalyst, [NH3] = 1000 ppm, [O2] = 2%,
a
Reaction conditions: 0.2 g of catalyst, [NH3] = 1000 ppm, [O2] = 2%,
[H2O] = 2.5%, He = balance, total flow rate = 500 ml/min, and GHSV =
5
−1
5
−1
.
He = balance, total flow rate = 500 ml/min, and GHSV = 2.0 × 10 h
.
2.0 × 10 h

1
62
LONG AND YANG
of Fe2O3–Al2O3 (0.28 mmol/g) > Fe2O3–TiO2 (0.19 mmol/
g), Fe2O3–ZrO2 (0.18 mmol/g) > Fe2O3–SiO2 (0.05 mmol/
g). This is in line with their surface area (Table 1). During
the ammonia TPD experiments, only a trace amount of N2
◦
was produced above 350 C, but other oxidation products
(
such as N2O or NOx ) were not detected.
The ammonia TPD profiles on 1–20% Fe2O3–TiO2
2
4
−
(
SO ) are shown in Fig. 4. Similar to Fe2O3–TiO2 prepared
with iron nitrate, these samples also showed a wide tem-
perature range for ammonia desorption. However, a large
◦
amount of N2 formation was observed at around 430 C.
The amount of N2 formation increased with the amounts
of iron and sulfur (Fig. 5). But other oxidation products,
e.g., N2O or NOx , were not detected. This indicates that
lattice oxygen can oxidize ammonia adsorbed species to ni-
trogen at high temperatures. The amount of ammonia des-
orption (including oxidation product N2) increased from
0.20 to 0.38 mmol/g with increasing Fe2O3 content from 1 to
2
0% on the catalysts (Figs. 4 and 5). Also, 5% Fe2O3–TiO2
2
−
(
SO ) (0.23 mmol/g) showed more ammonia desorption
4
than 5% Fe2O3–TiO2 (0.19 mmol/g).
TPSR between O2 and ammonia adsorbed species.
Figure 6 shows the TPSR results between ammonia adsor-
2
−
FIG. 2. SCR performance on (a) 10% Fe2O3–TiO2 (SO ), (b) 5% bed species and O2 on 5% Fe2O3–TiO2. NH3 desorption
4
Fe2O3–TiO2 (SO2 ), and (c) 5% Fe2O3–TiO2 catalysts under the condi- was significant at low temperatures (i.e., below 300 C).
−
◦
4
5
−1
.
tions of 1000 ppm NO, 1000 ppm NH3, 2% O2, and GHSV = 2.0×10 h
At high temperatures, the reaction between O2 and NH3
adsorbed species occurred, producing N2 and H2O. The
SCR performance on Fe2 O3–Ti O2 catalysts. Since
Fe2O3–TiO2 catalysts are known to be active for the SCR
of NO with ammonia, we also investigated the SCR activ-
ities on the Fe2O3–TiO2 catalysts prepared with the sol-
2
4
−
gel method. Fe2O3–TiO2 (5%), 5% Fe2O3–TiO2 (SO ),
and 10% Fe2O3–TiO2 (SO2 ) were chosen for this study.
−
4
NO conversions on these catalysts are shown in Fig. 2. Un-
der the conditions of 1000 ppm NO, 1000 ppm NH3, 2%
5
−1
O2, and GHSV = 2.0 × 10 h , NO conversions increased
in the following order: 5% Fe2O3–TiO2 < 5% Fe2O3-TiO2
2
4
−
2−
(
SO ) < 10% Fe2O3–TiO2 (SO4 ). This is consistent with
N2 selectivity for the SCO reaction over these catalysts
Tables 2 and 3). The maximum SCR activity was obtained
(
◦
2−
at 350 C on 10% Fe2O3–TiO2 (SO4 ).
TPD of ammonia. The TPD profiles of ammonia on
Fe2O3-containing catalysts are shown in Fig. 3. On 5%
Fe2O3–SiO2, onlyasmallamountofammoniawasdesorbed
◦
at 115 C. This may be related to its low surface area. For
the other samples, ammonia desorption was observed over
◦
the wide temperature range of 50–450 C, indicating a broad
distribution of sites for ammonia adsorption. The NH3 des-
orbed at low temperatures may be assigned to physically
adsorbed NH3 and that desorbed at high temperature may
be related to the chemisorbed NH3. A quantitative analysis
using 1.07% NH3/He as standard gas indicated that the
FIG. 3. TPD profiles of ammonia from (a) 5% Fe2O3–Al2O3, (b) 5%
amount of NH3 desorption decreased in the relative order Fe2O3–TiO2, (c) 5% Fe2O3–ZrO2, and (d) 5% Fe2O3–SiO2.

CATALYTIC OXIDATION OF AMMONIA TO NITROGEN OVER Fe2O3–TiO2
163
FIG. 4. TPD profiles of ammonia from (a) 20, (b) 10, (c) 5, and (d) 1%
FIG. 6. TPSR profiles in 5% O2 following exposure of 5% Fe2O3–
TiO2 to 1.07% NH3/He for 60 min.
Fe2O3–TiO2 (SO2 ) catalysts.
−
4
◦
maximum N2 production appeared at 330 C. Also, a small
amount of NO was formed at high temperatures. N2O was
not detected in this experiment. The observation of H2O at
◦
below 200 C is due to adsorbed H2O (an impurity in 1.07%
NH3/He) because no other oxidation product of ammonia
(
e.g., N2 or NO) was observed.
Similar results were obtained on 10% Fe2O3–TiO2
2
−
(
SO ). NH3 desorption dominated at low temperatures.
4
◦
At above 300 C, O2 oxidized ammonia adsorbed species
to nitrogen and water. The oxidation peak temperature ap-
peared at 370 C. Almost no N2O or NO was observed in
◦
the entire process.
DISCUSSION
The above results showed that Fe2O3-containing mixed
oxides prepared from the sol-gel route were highly active
for the SCO of ammonia to nitrogen. They could convert
all of the ammonia to N2 and NO at high temperatures,
with N2 as the major product. The N2 selectivity and yield
were found to increase in a trend of 5% Fe2O3–SiO2 < 5%
Fe2O3–ZrO2 < 5% Fe2O3–TiO2 < 5% Fe2O3–Al2O3 <
2
−
2−
5
2
% Fe2O3–TiO2 (SO4 ) < 10% Fe2O3–TiO2 (SO4 ),
2
−
0% Fe2O3–TiO2 (SO4 ) (Tables 2 and 3). The catalysts
prepared with iron sulfate were more selective for N2
than those prepared with nitrate. More than 92% of N2
2
4
−
yields were obtained on the 10% Fe2O3–TiO2 (SO
)
FIG. 5. N2 formation during TPD of ammonia from (a) 20, (b) 10,
c) 5, and (d) 1% Fe2O3–TiO2 (SO4 ) catalysts.
2−
2−
◦
(
2 3 2
4
and 20% Fe O –TiO (SO ) at 400–450 C under the

1
64
LONG AND YANG
5
−1
condition of GHSV = 2.0 × 10 h . They were more active action, the NO generating from NH3 oxidation by O2 on the
than Ni/Al2O3 (3) but less active than Fe-exchanged latter would be lower than that on the former. The decrease
zeolites (10, 11) for the conversion of NH3 to N2. H2O only in NO formation on the sulfur-containing catalyst will
decreased N2 yield slightly due to a decrease in N2 sele- also “increase” the apparent NO conversion. In the SCO
ctivity (Table 4). The ammonia TPD profiles showed reaction, when NO (the main by-product) is generated
that NH3 could adsorb on a variety of acid sites of these during the SCO reaction, it can be further reduced to N2 by
catalysts and they desorbed over a wide temperature range the unreacted ammonia adsorbed species through the SCR
(
Figs. 3 and 4). The ammonia adsorbed species include reaction. This finding supports the two-step SCO mech-
physically adsorbed NH3 (which was desorbed at low anism in which NO is an intermediate for N2 formation
temperatures) and chemisorbed NH3. The coordinated (4, 11). NH3 was first oxidized to NO by O2. This reaction
NH3 is probably the main form of chemisorbed ammonia occurs either on the catalyst surface or in the gaseous phase,
because only Lewis acid sites were found on Fe2O3, TiO2, or both. Our empty-tube results showed that NH3 conver-
◦
and Fe2O3/TiO2 (15). However, when sulfated, strong sions were 23–55% at 350–450 C under the condition with
Brønsted acid sites were formed on TiO2 (16) and Fe2O3 500 ml/min of total flow rate, with NO as the predominant
+
(
14). Therefore, NH ions are probably present on the product (10). Gaseous NH3 mainly contributes to NO for-
4
2−
Fe2O3–TiO2 (SO ) catalysts. The foregoing IR spectra mation, which was proven on Fe-exchanged pillared clays
4
showed that sulfate species were present on the Fe2O3– in our previous work (9). Subsequently, the formed NO re-
2
4
−
TiO2 (SO ) catalysts after calcination (Fig. 1). In the case acts with NH3 adsorbed species to produce N2 through the
of sulfate ions, S == O has a covalent double bond and has a SCR reaction. Therefore, good SCR catalysts are expected
much stronger affinity to electrons compared with that of a to have high N2 selectivities for the SCO reaction. An
simple metal sulfate; hence, the Lewis acid strength of metal increase in surface acidity enhances ammonia adsorption
ion becomes substantially stronger by the inductive effect and thus decreases the concentration of gaseous NH3. Since
of S == O in the complex (14, 17). When a water molecule is NO is mainly generated from gaseous NH3 oxidation by O2
bonded to the Lewis acid site, the Lewis acid site becomes (9, 10), less gaseous NH3 will result in less NO formation.
a Brønsted acid site. Also, sulfate species provided new Simultaneously, more NH3 (i.e., NH3 adsorbed species)
+
sites for ammonia adsorption, forming NH4 ions. All of will be used to reduce NO. Of course this will improve
these will increase surface acidity and acid strength of the SCR activity and N2 selectivity for the SCO reaction.
catalysts containing Fe2O3 and TiO2 after sulfation. By However, water also adsorbs on the catalyst competitively
comparing surface acidity (via ammonia TPD, Figs. 3 and and enhances sulfur removal under the reaction conditions.
4
) and SCO performance (Tables 2 and 3), we can see that This will increase the concentration of gaseous NH3 and
the surface acidity has no bearing on the NH3 conversion, thus increase NH3 oxidation to NO and decrease N2
while it is consistent with the N2 selectivity and yield, i.e., selectivity, as shown in Table 4.
the higher the surface acidity, the higher the N2 yield.
In addition, the above TPD results also showed that lat-
It is significant that the Fe2O3-containing catalysts also tice oxygen could oxidize the ammonia adsorbed species to
2
4
−
showed good activities for SCR of NO with ammonia. N2 on the Fe2O3–TiO2 (SO ) catalysts (Fig. 5), indicating
The SCR activities increased with the sequence of 5% that lattice oxygen may be one of the active oxygen species
Fe2O3–TiO2 < 5% Fe2O3–TiO2 (SO2 ) < 10% Fe2O3– for the SCO reaction at high temperatures. Over the cata-
−
4
2
−
TiO2 (SO ) (Fig. 2), which is in good agreement with their lysts prepared with iron nitrate, only a trace amount of
4
surface acidity (Figs. 3 and 4) and N2 selectivity for the N2 was formed during TPD experiments. This is related to
SCO reaction (Tables 2 and 3). This observation suggests the fact that most of the ammonia desorbed from the sur-
a correlation among them. It is noted that a high Fe face at high temperatures, prior to oxidation, due to their
content on the 10% Fe2O3–TiO2 (SO2 ) also has beneficial weaker acidities. The reaction temperature (430 C) with
−
◦
4
effect on the SCR activity at low temperatures (14). The lattice oxygen during TPD experiments (Fig. 5) was higher
◦
SCR of NO with NH3 has been widely studied on many than those (330 and 370 C) in the presence of gaseous oxy-
materials, such as mixed oxides and molecular sieves. Most gen during TPSR experiments. This suggests that gaseous
researchers believe that strong surface acidity is beneficial and/or adsorbed oxygen species may also participate in the
to NH3 adsorption and SCR activity, although which SCO reaction at lower temperatures. The variable valence
+
ammonia adsorbed species (NH or coordinated NH3) is of iron cations on the Fe2O3-containing catalysts might be
4
involved in the reaction is still under debate (12, 15). The beneficial to oxygen adsorption and activation. The oxygen
−
δ−
2−
higher surface acidity results in a higher SCR activity on adsorbed species, e.g., O , O (1 < δ < 2), and O , were
2
2
2
2−
5
% Fe2O3–TiO2 (SO ) than on 5% Fe2O3–TiO2. Also, observed on O2-adsorbed Fe2O3 by IR spectroscopy (18).
4
the 5% Fe2O3–TiO2 showed a higher NO selectivity in the TheadsorptionandactivationofoxygenontheFe2O3–TiO2
2−
SCO reaction than did 5% Fe2O3–TiO2 (SO4 ), especially SCO catalysts need to be verified by further spectroscopic
◦
at 350 and 400 C (Tables 2 and 3). Hence in the SCR re- characterization.

CATALYTIC OXIDATION OF AMMONIA TO NITROGEN OVER Fe2O3–TiO2
165
CONCLUSIONS
3. Amblard, M., Burch, R., and Southward, B. W. L., Appl. Catal. B 22,
L59 (1999).
4
5
. Amblard, M., Burch, R., and Southward, B. W. L., Catal. Today 59,
65 (2000).
. Sazonova, N. N., Simakov, A. V., Nikoro, T. A., Barannik, G. B.,
Lyakhova, V. F., Zheivot, V. I., Ismagilov, Z. R., and Veringa, H., React.
Kinet. Catal. Lett. 57, 71 (1996).
The Fe2O3-containing mixed oxides prepared from a sol-
gel route were highly active for the SCO of ammonia to ni-
trogen. The catalysts prepared from iron sulfate were more
selective for N2 than those prepared from nitrate. The N2
selectivity for the SCO reaction was in good agreement with
their surface acidity as well as SCR activity. More than 92%
3
6. Gang, L., van Grondelle, J., Anderson, B. G., and van Santen, R. A.,
J. Catal. 186, 100 (1999).
7. Curtin, T., O’Regan, F., Deconinck, C., Knuttle, N., and Hodnett, B. K.,
Catal. Today 55, 189 (2000).
2
4
−
of N2 yields were obtained on the 10% Fe2O3–TiO2 (SO
)
and 20% Fe2O3–TiO2 (SO2 ) at 400–450 C under the con-
−
◦
4
8. Wollner, A., and Lange, F., Appl. Catal. A 94, 181 (1993).
5
−1
dition of GHSV = 2.0 × 10 h . The gaseous, adsorbed,
9. Long, R. Q., Chang, M. T., and Yang, R. T., Appl. Catal. B 33, 97 (2001).
and lattice oxygen all participated in the SCO reaction.
10. Long, R. Q., and Yang, R. T., Chem. Commun. 1651 (2000).
1
1
1
1. Long, R. Q., and Yang, R. T., J. Catal. 201, 145 (2001).
2. Bosch, H., and Janssen, F., Catal. Today 2, 369 (1988).
3. Arata, K., in “Advance in Catalysis,” Vol. 37, p. 165. Academic Press,
San Diego, 1990.
4. Long, R. Q., and Yang, R. T., J. Catal. 186, 254 (1999).
5. Busca, G., Lietti, L., Ramis, G., and Berti, F., Appl. Catal. B 18, 1
ACKNOWLEDGMENT
This work was supported by the Electric Power Research Institute and
NSF Grant CTS-0095909.
1
1
(
1998).
16. Chen, J. P., and Yang, R. T., J. Catal. 139, 277 (1993).
7. Yamaguchi, T., Appl. Catal. 61, 1 (1990).
REFERENCES
1
1
2
. II’chenko, N. I., Russian Chem. Rev. 45, 1119 (1976).
. Li, Y., and Armor, J. N., Appl. Catal. B 13, 131 (1997).
18. Al-Mashta, F., Sheppard, N., Lorenzelli, V., and Busca, G., J. Chem.
Soc., Faraday Trans. I 18, 979 (1982).