N2O Removal by Catalytic Decomposition and Reduction with CH4 over Fe/Al2O3

Article abstract of DOI:10.1246/bcsj.76.2329

N2O removal by catalytic decomposition and reduction with CH4 over Fe/Al2O3 was investigated. The catalytic conversion of N2O to N2 over Fe/Al2O3 was effectively promoted by the addition of CH4. The activity of Fe/Al2O3 normalized on a TOF basis for N2O decomposition increased linearly with the Fe loading, whereas that for N2O reduction increased with an increase in Fe loading up to 2 wt percent, then subsequently decreased. Structural characterization by UV-vis and TPR revealed the presence of two types of Fe species, Fe2O3 and FeAlO3, in Fe/Al2O3. It was proposed that the catalytically active species for N2O decomposition and N2O reduction are Fe2O3 and FeAlO3, respectively.

Full text of DOI:10.1246/bcsj.76.2329

Ó 2003 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 76, 2329–2333 (2003) 2329
N O Removal by Catalytic Decomposition and Reduction
2
with CH over Fe/Al O
3
4
2
ꢀ
y
y
Masaaki Haneda, Manabu Shinriki, Yukinori Nagao, Yoshiaki Kintaichi, and Hideaki Hamada
Research Institute for Green Technology, National Institute of Advanced Industrial Science and Technology (AIST),
AIST Tsukuba Central 5, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8565
yFaculty of Science and Technology, Tokyo University of Science,
2641 Yamazaki, Noda, Chiba 278-8510
Received July 11, 2003; E-mail: m.haneda@aist.go.jp
N2O removal by catalytic decomposition and reduction with CH4 over Fe/Al2O3 was investigated. The catalytic
conversion of N2O to N2 over Fe/Al2O3 was effectively promoted by the addition of CH4. The activity of Fe/Al2O3
normalized on a TOF basis for N2O decomposition increased linearly with the Fe loading, whereas that for N2O reduction
increased with an increase in Fe loading up to 2 wt %, then subsequently decreased. Structural characterization by UV–vis
and TPR revealed the presence of two types of Fe species, Fe2O3 and FeAlO3, in Fe/Al2O3. It was proposed that the
catalytically active species for N2O decomposition and N2O reduction are Fe2O3 and FeAlO3, respectively.
Nitrous oxide (N2O) has a strong greenhouse effect, with a
global-warming potential per molecule of about 310-times that
of CO2, and contributes to catalytic depletion of the strato-
N2O by decomposition and reduction with CH4 over Fe/Al2O3
in which the formation of a highly dispersed Fe species can be
expected. The objective of using CH4 as a reductant was to
achieve a simultaneous abatement of greenhouse gas, due to
the fact that its global-warming potential per molecule is about
21-times that of CO2. To specify the active species on Fe/
Al2O3 for N2O decomposition and N2O reduction with CH4,
the structural properties of supported Fe species were charac-
terized by means of UV–vis and temperature-programmed
reduction (TPR).
1
spheric ozone layer. N2O originates from both natural sources
and human contributions, such as the production of adipic acid
2
and nitric acid. According to a recent study, the contribution
from vehicles (especially gasoline-fueled) has increased by a
factor of 3 over the past 10 years. In this regard, catalytic re-
moval of N2O has attracted much attention recently. The sim-
plest way to remove N2O is decomposition to N2 and O2, for
which various types of catalysts, such as ion-exchanged
3
–6
7–9
9–11
Experimental
zeolites, metal oxides, and noble metals,
have been
3
reported. For example, Li and Armor examined various ion-
exchanged zeolites, and found that Cu- and Co-exchanged
ZSM-5 are active in N2O decomposition. A high catalytic per-
The alumina used in this study was supplied by Mizusawa
Chemical (GB-45). Fe/Al2O3 catalysts were prepared by impreg-
nating Al2O3 with an aqueous solution of Fe(NO3)2 9H2O, fol-
ꢁ
4
lowed by drying at 383 K and calcination at 873 K for 5 h in air.
The loading of Fe was changed from 1 to 10 wt %. Unsupported
Fe2O3 was prepared by mixing an aqueous solution of
formance of Co–ZSM-5 was also reported by Tabata et al. and
5
11
Kapteijn et al. Kapteijn et al. also reviewed the decomposi-
tion of N2O over solid catalysts, especially unsupported and
supported metal oxides. They showed that some Co-containing
mixed oxides, such as Co–La and Co–Mg binary oxides, are
catalytically more active than Co–ZSM-5.
Fe(NO3)3 9H2O with an aqueous solution of Na2CO3, followed
ꢁ
by washing the resulting precipitation with distilled water, drying
at 383 K, and calcination at 873 K for 5 h in air.
The catalytic activity was measured with a fixed-bed flow
reactor. The reaction gas containing N2O (1000 ppm), CH4 (0
or 1000 ppm), and helium as a balance gas was fed to 0.05 g of cat-
As another catalytic N2O removal method, the reduction of
N2O with hydrocarbons is attracting great interest in both
applied and fundamental research. Since the first reports of
3
ꢂ1
alyst at a rate of 60 cm min . Before the activity measurement,
the catalysts were treated in situ at 873 K for 1.5 h in flowing
helium. The reaction temperature was first decreased from 873
to 673 K in steps of 50 K, and then increased from 873 to 973
K. The effluent gas was analyzed by two on-line gas chromato-
graphs equipped with a Molecular Sieve 5A column (for the anal-
ysis of O2, N2, CH4, and CO) and a Porapak Q column (for that of
N2O and CO2). The reaction rates of N2 formation were also mea-
sured under nearly differential reaction conditions, giving N2O
conversion of less than 30% by controlling the catalyst weight.
The amount of chemisorbed NO was measured by a pulse
method. A sample (0.05 g) was first treated at 873 K in flowing
12–15
Segawa and coworkers
on the high catalytic performance
of Fe–ZSM-5 for N2O reduction with C3H6 in the presence
of H2O and O2, numerous investigations have focused on Fe
16
ion-exchanged zeolites. In contrast to N2O decomposition,
however, few investigations of N2O reduction on metal oxide
catalysts have been performed so far. In view of a recent report
1
7
3þ
on Fe/ZSM-5 catalysts, suggesting that isolated Fe ions are
responsible for the catalytic activity in N2O reduction with
C3H6, supported catalysts, including highly dispersed Fe spe-
cies, would also show good performance for N2O reduction.
In the present study, we investigated the catalytic removal of
Published on the web December 15, 2003; DOI 10.1246/bcsj.76.2329

2
330 Bull. Chem. Soc. Jpn., 76, No. 12 (2003)
Catalytic Removal of N2O over Fe/Al2O3
1
00
100
80
60
40
20
0
(
A)
(B)
80
60
40
20
0
5
00 600 700 800 900 1000
500 600 700 800 900 1000
Temperature / K
Temperature / K
Fig. 1. Temperature dependence of N2O conversion to N2 over Fe/Al2O3 in (A) decomposition and (B) reduction with CH4 in the
absence of O2. Reaction conditions: N2O = 1000 ppm, CH4 = 0 or 1000 ppm, catalyst weight = 0.05 g, gas flow rate = 60
3
ꢂ1
cm min . ( ) Al2O3, ( ) 1% Fe/Al2O3, ( ) 2% Fe/Al2O3, ( ) 5% Fe/Al2O3, and ( ) 10% Fe/Al2O3.
helium for 0.5 h, followed by cooling to room temperature. Sev-
eral pulses of NO were introduced onto the sample until no more
adsorption was observed. The crystal structure was identified
by measurements using an X-ray diffraction (XRD) system
80
(
Shimadzu, XD-D1) with Cu Kꢀ radiation at 40 kV and 40 mA.
UV–vis diffuse reflectance spectra of the catalysts were recorded
in the range of 200–800 nm with a resolution of 2 nm using a
UV–vis spectrometer (Shimadzu, UV-2400PC) equipped with an
integrating sphere attachment. The reduction behavior of support-
ed-Fe species was examined by TPR with H2. Before each TPR
measurement, a sample (0.05 g) was heated in flowing air at 873
K for 1 h, and then cooled to room temperature in the same
60
40
atmosphere. The sample was then heated to 1173 K at a rate of
ꢂ1
20
0
1
of 30 cm min . Reduction of CuO to metallic copper was mea-
0 K min in a flow of 10 vol % H2–Ar gas mixture at a flow rate
3
ꢂ1
sured to calibrate the amount of H2 consumed.
0
2
4
6
8
10
Results and Discussion
Fe loading / wt%
N2O Decomposition and N2O Reduction with CH4. N2O
decomposition and N2O reduction with CH4 were carried out
over Fe/Al2O3 with different Fe loadings. Figures 1(A) and
Fig. 2. Changes in the activity of Fe/Al2O3 for N2O decom-
position ( ) and N2O reduction with CH4 ( ) in the ab-
sence of O at 723 K as a function of Fe loading. Reaction
1
(B) show the temperature dependence of N2O conversion
2
for the decomposition and reduction, respectively. N2, CO2,
and traces of CO were detected as the products in N2O reduc-
tion with CH4. Although Al2O3, itself, catalyzed both reactions
at high temperatures, the addition of Fe caused an enhancement
of N2O conversion at low temperatures. Figure 2 shows the Fe
loading dependence of the reaction rates of N2 formation for
N2O decomposition and N2O reduction at 723 K. The reaction
rate for N2O reduction is clearly much faster than that for N2O
decomposition, indicating that the addition of CH4 is effective
in promoting the catalytic conversion of N2O. It was also found
that the N2 formation rates for both reactions increased with in-
creasing Fe loading. This suggests that the supported Fe acts as
catalytically active sites for both reactions. However, the effect
of Fe loading on the reaction rate was quite different in the two
reactions. An almost linear correlation was observed up to 10
wt % for N2O decomposition. For N2O reduction, on the other
hand, the N2 formation rate increased with Fe loading up to 5
conditions: N O = 1000 ppm, CH = 0 or 1000 ppm, cat-
2
4
3
ꢂ1
alyst weight = 0.015–0.2 g, gas flow rate = 60 cm min
.
wt %, while a prominent increase was not observed above 5
wt % Fe loading. This suggests that the number of active sites
for N2O reduction does not increase linearly with Fe loading.
Therefore, the active sites for N2O reduction should be differ-
ent from that for N2O decomposition.
In order to obtain information about the active sites for N2O
decomposition and N2O reduction, the turnover frequency
(TOF) for N2 formation was calculated based on the number
of surface Fe atoms. Here, the number of surface Fe atoms
was estimated by NO chemisorption, because NO is known
to be adsorbed on Fe (b ) and Fe(c) in Fe/Al2O3 as mononitrosyl
18,19
species.
Since a small amount of NO was adsorbed on
Al2O3, the number of exposed Fe atoms was calculated by sub-
tracting the amount of chemisorbed NO on Al2O3 from that on

M. Haneda et al.
Table 1. Physical Properties of Fe/Al2O3
Catalyst BET surface area
Bull. Chem. Soc. Jpn., 76, No. 12 (2003) 2331
Amount of NO
chemisorption/mol g
Number of exposed Fe
ꢂ1
2
ꢂ1
ꢂ1
/m g
atoms/Fe-atoms g
ꢂ5
Al2O3
176
169
165
163
156
1:4 ꢃ 10
—
1:2 ꢃ 10
ꢂ5
19
1% Fe/Al2O3
2% Fe/Al2O3
5% Fe/Al2O3
10% Fe/Al2O3
3:4 ꢃ 10
ꢂ5
19
3:8 ꢃ 10
1:4 ꢃ 10
ꢂ5
19
6:5 ꢃ 10
3:1 ꢃ 10
ꢂ5
19
9:6 ꢃ 10
4:9 ꢃ 10
1
1
5
0
5
0
540
(
e)
300
(d)
(
(
c)
b)
(a)
0
2
4
6
8
10
Fe loading / wt%
2
00 300 400 500 600 700 800
Wavelength / nm
Fig. 3. Changes in the TOF for N2O decomposition ( ) and
N2O reduction by CH4 ( ) over Fe/Al2O3 in the absence
of O2 at 723 K as a function of Fe loading. Reaction con-
ditions: N2O = 1000 ppm, CH4 = 0 or 1000 ppm, catalyst
Fig. 4. Diffuse reflectance UV–vis spectra of (a) 1% Fe/
Al2O3, (b) 2% Fe/Al2O3, (c) 5% Fe/Al2O3, (d) 10% Fe/
Al2O3, and (e) Fe2O3.
3
ꢂ1
weight = 0.015–0.2 g, gas flow rate = 60 cm min
.
Fe/Al2O3. Table 1 summarizes the amounts of chemisorbed
NO on Fe/Al2O3 with different Fe loadings as well as their sur-
face areas. The BET surface area of Fe/Al2O3 was slightly de-
creased as Fe loading was increased, suggesting a blockage of
the pores of Al2O3 by supported Fe. Based on the number of
surface Fe atoms, the TOFs for N2 formation in N2O decompo-
sition and N2O reduction were calculated. As shown in Fig. 3,
the dependency of the TOF on Fe loading was quite different in
the two reactions. The TOF for N2O decomposition linearly in-
creased with Fe loading, while that for N2O reduction increased
with an increase in Fe loading up to 2 wt %, then subsequently
decreased. This suggests that the active site for N2O reduction
is different from that for N2O decomposition.
Structural Characterization of Fe/Al2O3. The XRD pat-
terns of the Fe/Al2O3 samples were measured to identify their
crystal structures. Only the diffraction peaks assignable to ꢁ-
Al2O3 were observed for all of the Fe/Al2O3 samples with
up to 10 wt % Fe loading (results not shown). This suggests
that the Fe phases are present in a highly dispersed state or in
a solid solution with the support.
the spectrum of ꢀ-Fe2O3 (Fig. 4(e)), the absorption band at
540 nm would be assigned to Fe2O3 particles. Centi and
17
Vazzana observed absorption bands at 270 and 330 nm in
the UV–vis spectra of Fe/ZSM-5 samples prepared by CVD,
3
þ
and assigned these bands to isolated Fe species interacting
strongly with the support. Accordingly, we can consider the
3þ
band at 300 nm to be due to Fe species interacting strongly
þ
3
with Al2O3. A possible candidate for the Fe is Fe–Al com-
20
posite oxide. Hoffmann et al. recognized the presence of
FeAlO3 in Fe/Al2O3 using XRD, EXAFS, M o¨ ssbauer, and
ESCA techniques. They also observed an increase in the
amount of FeAlO3 as the Fe loading increased. In accordance
with their report, the band intensity at 300 nm increased with an
increase in Fe loading. Consequently, we can conclude that the
band at 300 nm is assignable to FeAlO3. It is noteworthy that
the absorption band at 300 nm slightly shifted to a higher wave-
length as the Fe loading increased to 10 wt %. This is probably
due to overlapping of the absorption bands for Fe2O3 particles,
suggesting that 10 wt % Fe/Al2O3 includes a relatively large
amount of Fe2O3 particles compared with the other Fe/Al2O3
samples.
UV–vis diffuse reflectance spectra were taken to learn about
the details of the state of supported Fe. Figure 4 shows the
spectra of Fe/Al2O3 with different Fe loadings after subtracting
the Al2O3 spectrum, along with that of ꢀ-Fe2O3. Fe/Al2O3
showed a distinct absorption band at 300 nm as well as a
shoulder band at around 540 nm. From a comparison with
It is known that the reduction behavior of a supported Fe spe-
cies changes depending on its dispersion and metal-support
21–23
interaction.
TPR measurements were made to obtain infor-
mation on the dispersion state of the Fe species in the Fe/Al2O3
catalysts. Figure 5 shows the TPR profiles of the Fe/Al2O3

2
332 Bull. Chem. Soc. Jpn., 76, No. 12 (2003)
Catalytic Removal of N2O over Fe/Al2O3
1
00
8
6
4
2
0
0
0
0
0
(
d)
(
(
(
c)
b)
a)
0
2
4
6
8
10
200
400
600
800
1000 1200
Fe loading / wt%
Temperature / K
Fig. 6. Percentage composition of Fe O ( ) and FeAlO3
2
3
(
) present in Fe/Al2O3 catalysts.
Fig. 5. TPR profiles of (a) 1% Fe/Al2O3, (b) 2% Fe/Al2O3,
c) 5% Fe/Al2O3, and (d) 10% Fe/Al2O3.
(
ferent structures, Fe2O3 and FeAlO3, in the catalysts. The same
20
conclusion was also reported by Hoffmann et al. If Fe2O3 and
FeAlO3 have different catalytic functions, good correlation be-
tween the TOF and the percentage composition of Fe2O3 and
FeAlO3 in Fe/Al2O3 catalysts should be obtained. Figure 6
shows the change in the percentage composition of Fe2O3
and FeAlO3 as a function of Fe loading. Here, the percentage
composition was calculated by dividing the amounts of Fe2O3
and FeAlO3, which were estimated from the H2 uptake in TPR,
by the total Fe content in the catalyst. A comparison of Fig. 3
with Fig. 6 shows a good correlation between the TOF for N2O
reduction and the percentage composition of FeAlO3. It should
be noted that the former reflects the surface profile and the latter
the subsurface, from the feature of impregnation, whereby Fe is
deposited on the surface of Al2O3. If Fe2O3 and FeAlO3 phases
are present separately on the Al2O3 support, the surface and
samples with different Fe loadings. The Fe/Al2O3 samples
gave very complicated TPR profiles consisting of two or three
reduction peaks, with the peak temperatures differing depend-
21–23
ing on the Fe loading. According to previous reports,
the
peaks located below 773 K can be ascribed to the reduction
of Fe2O3. The reduction of supported oxide species has been
reported to consecutively proceed from the surface to the
bulk. For example, CeO2/Al2O3 shows two or three reduction
peaks at around 873 and 1073 K, assignable to the reactions of
oxygen atoms located at the surface of and inside the CeO2 par-
24
ticles, respectively. A similar phenomenon would take place
for Fe2O3 supported on Al2O3. Accordingly, the low-temper-
ature peak at around 573 K is due to the surface reduction of
Fe2O3 particles, and the high-temperature peak at around 673
K is due to bulk reduction. The fact that 1 wt % Fe/Al2O3 gave
one reduction peak below 773 K suggests that this sample con-
tained only small Fe2O3 particles. This is in good agreement
with the results of the UV–vis measurements (Fig. 4(a)).
On the other hand, the reduction peaks located above 773 K
3þ
subsurface compositions of the two Fe species should be
20
the same. Hoffmann et al. prepared Fe/Al2O3 samples using
the same impregnation method as we employed in the present
study, and observed the presence of both Fe2O3 and FeAlO3
phases on the surface of Al2O3 by means of ESCA. This indi-
cates that Fe2O3 and FeAlO3 phases are present separately.
Taking into account their report, the composition of Fe2O3
and FeAlO3 on the surface of Fe/Al2O3 prepared in the present
study can be expected to be the same as that in the subsurface.
The fact that a good correlation between the TOF and the per-
(Fig. 5(a)–(d)) would be due to the reduction of iron–aluminum
oxide species.
21,23
Taking into account the results of UV–vis
measurements mentioned above, the peaks above 773 K can
be ascribed to the reduction of FeAlO3. It should be noted that
the apparent peak area increased with an increase in Fe loading,
suggesting an increase in the amount of FeAlO3. This is in
good agreement with the results of the UV–vis measurements.
Accordingly, we can conclude that there are two types of Fe
species, namely Fe2O3 and FeAlO3, in Fe/Al2O3 samples.
Specification of Active Sites. From the relationship be-
tween the TOF for the conversion of N2O and Fe loading shown
in Fig. 3, the active site for N2O reduction was presumed to be
different from that for N2O decomposition. Since Al2O3, itself,
was almost inactive for the conversion of N2O, supported Fe
should be the active sites. Therefore, there seem to be at least
two types of Fe species in the Fe/Al2O3 catalysts. In accord-
ance with this idea, UV–vis (Fig. 4) and TPR (Fig. 5) measure-
ments revealed the presence of two types of Fe species with dif-
3
þ
centage composition of Fe species was observed thus leads
us to conclude that the catalytically active species for N2O re-
duction is FeAlO3, whereas Fe2O3 behaves as a catalytically
active species for N2O decomposition.
Conclusions
The present study has demonstrated that the catalytic remov-
al of N2O over Fe/Al2O3 can be effectively promoted by the
addition of CH4. It was revealed by UV–vis and TPR measure-
ments that Fe2O3 and FeAlO3 phases are present in Fe/Al2O3
catalysts. The former acts as a catalytically active species for
N2O decomposition, while the latter acts as that for N2O reduc-
tion.

M. Haneda et al.
Bull. Chem. Soc. Jpn., 76, No. 12 (2003) 2333
oporous Mater., 21, 549 (1998).
C. Pophal, T. Yogo, K. Tanabe, and K. Segawa, Appl. Cat-
al., B, 16, 177 (1998).
15 K. Yamada, S. Kondo, and K. Segawa, Microporous Meso-
porous Mater., 35/36, 227 (2000).
References
1
4
1
2
H. Rodhe, Science, 248, 1217 (1990).
T. Salvesen, S. Roesch, P. A. Sermon, and P. Kaur, Top.
Catal., 16/17, 381 (2001).
3
4
Y. Li and J. N. Armor, Appl. Catal., B, 1, L21 (1992).
M. Tabata, H. Hamada, Y. Kintaichi, M. Sasaki, and T. Ito,
16 S. Kameoka, T. Suzuki, K. Yuzaki, T. Takeda, S. Tanaka,
S. Ito, T. Miyadera, and K. Kunimori, Chem. Commun., 2000,
745, and references therein.
17 G. Centi and F. Vazzana, Catal. Today, 53, 683 (1999).
18 D. G. Retwisch and J. A. Dumesic, J. Phys. Chem., 90, 1625
(1986).
19 M. Jiang, N. Koizumi, and M. Yamada, J. Phys. Chem. B,
104, 7636 (2000).
20 P. Hoffmann, M. Houalla, A. Proctor, M. J. Fay, and D. M.
Hercules, Appl. Spectrosc., 46, 208 (1992).
21 X. Gao, J. Shen, Y. Hsia, and Y. Chen, J. Chem. Soc., Fara-
day Trans., 89, 1079 (1993).
22 A. J. H. M. Kock, H. M. Fortuin, and J. W. Geus, J. Catal.,
96, 261 (1985).
23 R.-Y. Tang, S. Zhang, C. Wang, D. Liang, and L. Lin, J.
Catal., 106, 440 (1987).
24 H. C. Yao and Y. F. Yu-Yao, J. Catal., 86, 254 (1984), and
references therein.
Chem. Express, 7, 77 (1992).
F. Kapteijn, G. Marb a´ n, J. Rodriguez-Mirasol, and J. A.
Moulijn, J. Catal., 167, 256 (1997).
R. S. da Cruz, A. J. S. Mascarenhas, and H. M. C. Andrade,
Appl. Catal., B, 18, 223 (1998).
5
6
7
8
E. R. S. Winter, J. Catal., 19, 32 (1970).
A. Satsuma, H. Maeshima, K. Watanabe, and T. Hattori,
Energy Convers. Manage., 42, 1997 (2001).
J. Oi, A. Obuchi, G. R. Bamwenda, A. Ogata, H. Yagita, S.
Kushiyama, and K. Mizuno, Appl. Catal., B, 12, 277 (1997).
K. Yuzaki, T. Yarimizu, K. Aoyagi, S. Ito, and K.
Kunimori, Catal. Today, 45, 129 (1998).
F. Kapteijn, J. Rodriguez-Mirasol, and J. A. Moulijn, Appl.
Catal., B, 9, 25 (1996).
C. Pophal, T. Yogo, K. Tanabe, and K. Segawa, Catal. Lett.,
4, 271 (1997).
K. Yamada, C. Pophal, and K. Segawa, Microporous Mes-
9
1
0
1
1
1
2
4
1
3