The influence of chromium substitution on an iron-titanium catalyst used in the selective catalytic reduction of NO

Article abstract of DOI:10.1016/j.jcat.2012.04.007

A series of Cr-substituted Fe-Ti compounds were prepared by the co-precipitation method and were systematically investigated as catalysts for the selective catalytic reduction (SCR) of NO by NH₃. A variety of analytical techniques revealed that the Cr substitution amount affects the N₂ selectivity, SCR activity, redox behavior of NH ₃/NO_x, adsorption ability, and structure of catalysts in terms of surface properties, porosity, mobility of lattice oxygen, oxidative ability of Cr species, ratio of Bronsted acid sites and Lewis acid sites, NO_x adsorption capacity, and structural disorder and distortion. In a series of Fe_aCr_1-aTiO_x (a = 1, 0.75, 0.5, 0.2, 0) catalysts, Fe_0.5Cr_0.5TiO_x showed the highest activity because of the optimized interactions of Fe, Cr, and Ti species in this catalyst.

Full text of DOI:10.1016/j.jcat.2012.04.007

Journal of Catalysis 292 (2012) 32–43
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
The influence of chromium substitution on an iron–titanium catalyst
used in the selective catalytic reduction of NO
Asghar Karami , Vajiheh Salehi b
a,
⇑
a
Department of Chemical Engineering, Amirkabir University of Technology (Tehran Polytechnic), 424 Hafez Ave., Tehran, Iran
Department of Chemistry, Sharif University of Technology, Azadi Ave., Tehran, Iran
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 17 October 2011
Revised 7 April 2012
Accepted 11 April 2012
Available online 23 June 2012
A series of Cr-substituted Fe–Ti compounds were prepared by the co-precipitation method and were sys-
tematically investigated as catalysts for the selective catalytic reduction (SCR) of NO by NH
analytical techniques revealed that the Cr substitution amount affects the N selectivity, SCR activity,
redox behavior of NH /NO , adsorption ability, and structure of catalysts in terms of surface properties,
porosity, mobility of lattice oxygen, oxidative ability of Cr species, ratio of Brønsted acid sites and Lewis
acid sites, NO adsorption capacity, and structural disorder and distortion. In a series of Fe Cr1ꢀaTiO
a = 1, 0.75, 0.5, 0.2, 0) catalysts, Fe0.5Cr0.5TiO showed the highest activity because of the optimized inter-
actions of Fe, Cr, and Ti species in this catalyst.
3
. A variety of
2
3
x
x
a
x
Keywords:
Selective catalytic reduction
Redox behavior
(
x
Ó 2012 Elsevier Inc. All rights reserved.
Cr substitution
x
NO adsorption
1
. Introduction
enhances the low-temperature activity, probably as a result of syn-
ergistic effects between Fe and Mn species. It has been reported
that the introduction of lanthanide elements (such as La, Ce, and
Pr) and the main-group element In can also improve the activity,
Selective catalytic reduction (SCR) of NO with NH
3
is an efficient
from the exhaust
and economical technique for the removal of NO
x
gases of stationary and mobile sources, such as coal-fired power
plants and lean-burn diesel engines. Nowadays, the most widely
2 3
stability, and SO durability of SCR catalysts, when NH or hydro-
carbons are used as reducing agents. Based on our Fe titanate
catalyst, we partially replaced Fe with other elements to adjust
its physicochemical properties, expecting to enhance its low-tem-
perature SCR activity [10,11].
In this work, the influence of Cr substitution amounts on the
catalyst structure and catalytic activity was investigated, using
various characterization methods. First, X-ray photoelectron spec-
troscopy (XPS) and H temperature-programmed reduction (H
2 2
TPR) methods were used to evaluate variations in redox properties
during the substitution process. Then the structural properties
used catalyst system is V
has some problems, such as low N
tures as a result of formation of N
SO [1], the toxicity of V to the environment, and the adverse
O
2 5
–WO
3
(MoO
selectivity at high tempera-
O, high conversion of SO to
3 2
)/TiO , but this system
2
2
2
3
2
O
5
effects of V on human health (such as causing irritation to the skin,
eyes, and respiratory tract). Many studies have therefore been
performed to find new SCR catalyst systems that can be used as
substitutes for conventional V-based catalysts [2–4]. Among these,
Fe-based catalysts usually show good SCR activity and N
2
selectiv-
ity at relatively high temperatures.
2
were characterized using N physisorption, powder X-ray diffrac-
tion (XRD), UV–vis diffuse-reflectance spectroscopy (UV–vis DRS),
and X-ray absorption fine-structure (XAFS) methods. Finally, tem-
In general, most Fe-based catalysts are Fe3+-exchanged zeolites
or Fe
HBEA [6], Fe–TiO
8], usually show good SCR activity and N
ally used as a support for SCR catalysts because it has excellent SO
durability because it does not react with either SO or SO above
00 °C [9].
Previous studies have shown that in Fe-containing SCR cata-
lysts, the introduction of elements such as Mn significantly
2
O
3
-loaded types. These catalysts, such as Fe/ZSM-5 [5], Fe/
–PILC [7], Fe /WO /ZrO [4], and Fe–Mn/USY
selectivity; TiO is usu-
2
O
2 3
3
2
perature-programmed desorption of NH
NO TPD) and in situ diffuse-reflectance infrared Fourier-transform
spectroscopy (in situ DRIFTS) of NH and NO adsorption were car-
3 x 3
and NO (NH TPD and
[
2
2
x
2
3
x
3
2
ried out to reveal the evolution of the adsorption abilities of the
reactants, which is important for SCR reactions. A mechanism for
Cr promotion of the low-temperature SCR activity of Fe titanate
catalysts is proposed. Finally, a proposed overall mechanism of
2
a x
SCR reactions catalyzed by Fe Cr1ꢀaTiO is presented.
⇑
Corresponding author. Fax: +98 2155345806.
E-mail address: asghar54@jdsharif.ac.ir (A. Karami).
0
021-9517/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2012.04.007

A. Karami, V. Salehi / Journal of Catalysis 292 (2012) 32–43
33
2
. Experimental
collected XAFS data were calibrated using standard Fe
CrO
2
O
3
and
2
samples and then analyzed using the Viper software package
2
.1. Catalyst synthesis and activity tests
according to standard procedures [12]. During the EXAFS data pro-
cessing, the back-subtracted EXAFS function was first converted
into a k space and weighted by k to compensate for the diminishing
3
Fe
a = 1, 0.75, 0.5, 0.25, 0) were prepared by the co-precipitation
method using Fe(NO O, Ti(SO , and Cr(NO )3ꢁ9H O as pre-
ꢁ9H
cursors, and NH O (25 wt.%) as the precipitator. The precipitate
a
Cr1ꢀaTiO
x
catalysts with different Cr substitution amounts
(
amplitude caused by decay of the photoelectron wave. Fourier
3
3
)
3
2
4
)
2
3
2
transformation of the k -weighted EXAFS data was performed in
ꢀ1
ꢁH
3 2
the range k = 2–11.01 Å for both Fe K-edge and Cr K-edge with a
Hanning function window. In the least-squares fitting procedure,
the possible scattering paths were calculated using the IFEFFIT soft-
ware package [13].
cake was filtered and washed using distilled water, followed by
desiccation at 100 °C for 12 h and calcination at 400 °C for 6 h in
an air-conditioned environment. Pure oxides, namely Fe
and TiO , were prepared from Fe(NO O, Cr(NO ꢁ9H
ꢁ9H
Ti(SO using the same precipitation method for comparison of
SCR activity. state-of-the-art SCR catalyst, 4.5 wt.%
0 wt.% WO /TiO , was also prepared, using the conventional wet
impregnation method, as a reference in the SCR activity tests.
The NH SCR, NO oxidation, and NH oxidation tests were car-
2
O
3
, CrO
O, and
x
,
2
3
)
3
2
3
)
3
2
UV–vis DRS was performed using a V-550 UV–vis spectropho-
tometer (JASCO, Easton, MD, USA) equipped with an integrating
sphere attachment (ISV4469) under ambient conditions. The spec-
tra were collected for the wavelength range 200–900 nm, using
4 2
)
A
2 5
V O –
1
3
2
BaSO
XPS were recorded using a scanning X-ray microprobe (PHI
Quantera, ULVAC-PHI, Inc.), using Al K radiation (1486.7 eV).
4
as a reference material.
3
3
ried out over 0.6 mL of catalyst (ca. 200–350 mg, depending on
the catalyst density) in a fixed-bed quartz-tube reactor under the
following reaction conditions: 500 ppm NO and (or) 500 ppm
a
The binding energies of Fe2p, Cr2p, Ti2p, and O1s were calibrated
using the C1s peak (binding energy = 284.8 eV) as a standard. The
powdered samples were mounted with copper double-sided adhe-
sive tape on Al foil, and then the foil was fixed to a sample holder.
The sample was degassed overnight at room temperature at a pres-
NH
3
, 5 vol.% O
2
, 1000 ppm CO (when used), 5 vol.% CO
2
(when
used), 5 vol.% H
2
O (when used), N balance, and gas hourly space
2
ꢀ1
velocity (GHSV) 50,000 h . The effluent gas was analyzed using
a Fourier-transform infrared (FTIR) spectrometer (Nexus 670, Nico-
let, Madison, WI, USA) equipped with a heated, low-volume, multi-
ꢀ7
sure on the order of 10 Torr.
Prior to the H TPR experiments, the samples (100 mg) were
pretreated at 300 °C in a flow of 20 vol.% O /He (30 mL/min) for
2
ple-path gas cell (2 m). NO
x
conversion (XNox ) and N
2
selectivity
2
(
S
N2 ) were calculated as follows:
0.5 h and cooled to room temperature (30 °C). The temperature
was then raised linearly to 900 °C at a rate of 10 °C/min in a flow
of 5 vol.% H /Ar (30 mL/min). The H O, O , and H signals (m/
z = 2) were monitored online using a quadrupole mass spectrome-
ter (HPR20, Hiden Analytical Ltd.).
ꢀ
ꢁ
½
NO
x
ꢂ
out
ꢂin
2
2
2
2
X
NOx 1 ꢀ
ꢃ 100 % with ½NO
x
ꢂ ¼ ½NOꢂ þ ½NO
2
ꢂ
ð1Þ
½
NO
x
½
NOꢂin þ ½NH
3
ꢂ
ꢀ ½NO
2
ꢂ
ꢀ 2½N
2
Oꢂout
NH
pole mass spectrometer to record the signals of NH
NH and m/z = 15 for NH) and NO (m/z = 30 for NO and m/z = 46
for NO ). Prior to the TPD experiments, the samples (100 mg) were
pretreated at 300 °C in a flow of 20 vol.% O /He (30 mL/min) for
.5 h and cooled to room temperature (30 °C). Then the samples
3
TPD and NO
x
TPD were performed using the same quadru-
in
out
S
N2
ꢃ 100 %
ð2Þ
½
NOꢂ þ ½NH
3
ꢂin
3
(m/z = 16 for
in
2
x
2
2
.2. Characterizations
2
0
N
2
adsorption–desorption isotherms were obtained at 77 K
3
were exposed to a flow of 2500 ppm NH /Ar or 2500 ppm NO + 10 -
using a QuantachromeAutosorb-1C instrument. Prior to N adsorp-
2
2
vol.% O /Ar (30 mL/min) at 30 °C for 1 h, followed by Ar purging for
tion, the samples were degassed at 300 °C for 4 h. The surface areas
were determined using the Brunauer–Emmett–Teller (BET) equa-
tion in the relative pressure range 0.05–0.35. The pore volumes,
average pore diameters, and pore size distributions were determined
by the Barrett–Joyner–Halenda (BJH) method from the desorption
branches of the isotherms. The reproducibility of the BET data
1 h. Finally, the temperature was raised to 500 °C in an Ar flow at a
rate of 10 °C/min.
The in situ DRIFTS experiments on NH
Cr1ꢀaTiO catalysts were performed using an FTIR spectrometer
(Nicolet Nexus 670) equipped with an MCT/A detector cooled with
liquid N . An in situ DRIFTS reactor cell with a ZnSe window (Nexus
Smart Collector) connected to a purging/adsorption gas control sys-
tem was used for the NH /NO in situ adsorption experiments. The
temperature of the reactor cell was controlled precisely using an
Omega programmable temperature controller. Prior to NH /NO
adsorption, the samples were pretreated at 400 °C in a flow of
20 vol.% O /N for 0.5 h and cooled to 30 °C. The spectra of different
catalysts at 30 °C were collected in flowing N and set as back-
grounds, which were automatically subtracted from the final spectra
after NH /NO adsorption. The samples were then exposed to a flow
of 500 ppm NH /N or 500 ppm NO + 5 vol.% O /N (300 mL/min) at
30 °C for 1 h, followed by N purging for 0.5 h. All spectra were re-
3 x
/NO adsorption over
Fe
a
x
2
ꢀ2
presented is 15 m /g. The error limit of the BJH method is ±2%.
Powder XRD measurements were carried out using a computer-
3
x
ized D/max-RB diffractometer (Rigaku, Tokyo, Japan) with a Cu K
a
radiation resource. The 2h data from 10° to 90° were collected at a
rate of 4°/min with a step size of 0.02°.
3
x
XAFS experiments were performed using the U7C beamline at
the Synchrotron Radiation Laboratory; the storage ring was oper-
ated at 0.8 GeV with a maximum current of 300 mA. The hard X-
ray beam was from a three-pole superconducting Wiggler with a
magnetic field intensity of 6 T. A fixed-exit Si(111) double-crystal
monochromator was used to reduce the harmonic content of the
monochromatic beam. The incident and output beam intensities
were monitored and recorded using ionization chambers filled with
2
2
2
3
x
3
2
2
2
2
ꢀ1
corded by accumulating 100 scans with a resolution of 4 cm
.
2
Ar/N . A Model 6517 electrometer (Keithley, Cleveland, OH, USA)
was used to collect the electronic charge directly. Before the XAFS
measurements, the catalyst samples were crushed into fine pow-
ders above 200 mesh and applied to transparent adhesive tape as
a coating. The XAFS, X-ray absorption near-edge structure (XANES),
and extended X-ray absorption fine-structure (EXAFS) spectra of
the Fe K-edge and Cr K-edge were recorded in transmission mode
at room temperature in an air-conditioned environment. The
3. Results and discussion
3.1. Catalytic performance
3.1.1. SCR activity of Fe
Fig. 1A shows the results of NO
over Fe Cr1ꢀaTiO (a = 1, 0.75, 0.5, 0.25, 0) catalysts and pure
a
Cr1ꢀaTiO
x
catalysts
x
conversion in the SCR reaction
a
x

3
4
A. Karami, V. Salehi / Journal of Catalysis 292 (2012) 32–43
oxides, namely Fe
from 75 to 400 °C, using the following reaction conditions:
NO] = [NH ] = 500 ppm, [O ] = 5 vol.%, balance, and
GHSV = 50,000 h . As we can see, pure TiO showed no SCR activ-
ity below 350 °C, and only 30% NO conversion was obtained at
00 °C. Pure Fe and CrO showed very narrow operating-tem-
perature windows in the high- and low-temperature ranges,
respectively, and neither of the maximum NO conversions
reached 100%. Furthermore, the N selectivities over these two
samples were rather low, as shown in Fig. 2. The coexistence of
Fe and Ti in FeTiO greatly enlarged the operating-temperature
window, and NO was completely reduced in the temperature
range 225–350 °C, with high N selectivity. When Fe was partially
replaced with Cr, the NO conversion at relatively low tempera-
tures increased significantly. Fe0.5Cr0.5TiO with a molar ratio of
Fe/Cr = 1:1 showed the best activity, and NO was completely re-
2 3 x 2
O , CrO , and TiO , as a function of temperature,
[
3
2
N
2
ꢀ
1
2
x
4
2
O
3
x
x
2
x
x
2
x
x
x
2 3 a x
Fig. 2. N selectivity in the NH SCR reaction over Fe Cr1ꢀaTiO catalysts.
duced at about 175 °C. However, for greater replacement of Fe,
the conversions at 250 °C were lower than that over Fe0.75Cr0.25-
TiO
in the following sequence: FeTiO
< Fe0.5Cr0.5TiO . This sequence is a result of the
x
. The apparent SCR activities at low temperatures increased
tures below 300 °C, for comparison with state-of-the-art SCR cata-
lysts, including the traditional V –WO /TiO catalyst, and Fe-
and Cu-exchanged zeolites. As shown in Fig. 1B, Fe/ZSM-5 [14]
and Fe/HBEA [15] catalysts showed good SCR activity at relatively
ꢄ CrTiO
x
< Fe0.25Cr0.75TiO
x
<
2
O
5
3
2
x
Fe0.75Cr0.25TiO
x
x
strong interactions between Fe and Cr, which led to high dispersion
of the active phases, and therefore high SCR activity.
high temperatures, and maximum NO/NO
tained above 350 and 250 °C, respectively. In the case of our
catalyst, the low-temperature SCR activity was
x
conversions were ob-
Although the replacement of Fe with Cr in an Fe titanate cata-
lyst could enhance the SCR activity, the N
significantly as a result of production of N O, especially at temper-
atures above 200 °C. The N selectivity at high temperatures de-
creased in the following sequence: FeTiO
Cr0.5TiO . For NO
ꢅ CrTiO
selectivity, we chose Fe0.75Cr0.25TiO as a model catalyst, over
which N selectivities of above 90% could be obtained at tempera-
2
selectivity decreased
Fe0.75Cr0.25TiO
x
much higher than that of Fe-exchanged zeolites, although the
SCR activity above 300 °C showed a sharp decrease, resulting in a
narrow operating-temperature window. Remarkably, the tempera-
2
2
x
x
> Fe0.75Cr0.25TiO > Fe0.5
ꢅ Fe0.25Cr0.75TiO
x
x
x
conversion and N
2
ture for 50% NO
er than that over the traditional V
would be possible to use it for the removal of NO
gases with low exhaust temperatures. To fully compare the SCR
activity of Fe0.75Cr0.25TiO with that of Cu/ZSM-5 reported by Park
x
conversion over Fe0.75Cr0.25TiO
x
was ca. 50 °C low-
catalyst, so it
from actual flue
x
2
O
5
–WO /TiO
3
2
x
2
x
x
et al. [16], we also performed another activity test under reaction
conditions identical to those reported in the literature, that is,
(
(
A)
B)
ꢀ1
GHSV 100,000 h and 10 vol.% H
concentration, the operating-temperature window of
shifted greatly toward high temperature, resulting
2
O. Under this high GHSV and
2
H O
Fe0.75Cr0.25TiO
x
in SCR activity below 300 °C lower than that of Cu/ZSM-5. The
GHSV is therefore an important factor in catalyst design, and the
H
2
O durability of Fe0.75Cr0.25TiO
ture work. The influences of GHSV and O
conversions over Fe0.75Cr0.25TiO were also investigated (Figs. S1
and S2 in the Supporting information). NO conversions of 100%
could still be obtained over this catalyst at temperatures above
x
still needs to be improved in fu-
2
concentration on NO
x
x
x
ꢀ1
2
50 °C, even at a high GHSV of 100,000 h , which is beneficial in
actual industrial applications. The influence of O on NO conver-
2
x
sion was more obvious at low temperatures than at high tempera-
tures, implying that different SCR reaction mechanisms might
operate in different temperature ranges, as in the case of an unsub-
stituted Fe titanate catalyst [17].
3
.1.2. Influence of CO, CO
activity
For practical purposes, the Fe0.75Cr0.25TiO
tested in the presence of CO, CO , and H O. The reaction conditions
for these tests were [NO] = [NH ] = 500 ppm, [O ] = 5 vol.%,
CO] = 1000 ppm (when used), [CO ] = 5 vol.% (when used),
O] = 5 vol.% (when used), N balance, total flow rate 500 mL/
2 2
, and H O on selective catalytic reduction
x
catalyst was also
2
2
3
2
[
[
2
H
2
2
ꢀ1
min, and GHSV = 50,000 h ; the results are shown in Fig. 3. As
we can see, the existence of CO and CO in the feed gas did not
obviously influence the SCR activity below 200 °C. However, the
NO conversion above 300 °C showed a slight increase, probably
as a result of reduction of NO by CO when unselective oxidation
of NH occurred extensively at high temperatures: 2CO + 2N-
O ? N + 2CO To confirm this speculation, an additional
2
Fig. 1. (A) NO
x 2 3
conversion and N selectivity in the NH SCR reaction over
x
Fe Cr1ꢀaTiO catalysts and pure oxides including Fe
a
x
2
O
3
,
CrO
x
,
and TiO
2
.
(B)
1
ꢀ1
x
Comparison of apparent SCR activity over Fe0:75Cr0:25TiOx (GHSV 50,000 h without
2
ꢀ1
H
2
O), Fe0:75Cr0:25TiOx (GHSV 100,000 h with 10 vol.% H
2 2 5
O), 4.5 wt.% V O –10 wt.%
3
WO /TiO , e/ZSM-5, Fe/HBEA, and Cu/ZSM-5.
3
2
2
2
.

A. Karami, V. Salehi / Journal of Catalysis 292 (2012) 32–43
35
O
2
+ 2NH
2
? 2N
2
+ 3H
2
O [10,18]. The effects of NO
2
and the ‘‘fast
SCR’’ reaction mechanism over conventional V
2 5 3 2
O –WO /TiO and
Fe-zeolite catalysts (such as Fe/HBEA and Fe/ZSM-5) have been
studied extensively by many researchers [15,19,20]. In this study,
the effects of Cr substitution amounts on the NO oxidation activi-
ties of Fe titanate catalysts were also investigated with the same
reaction conditions as those in Fig. 1A, along with a total flow rate
of 500 mL/min; the results are shown in Fig. 4A. With increasing Cr
substitution amounts, the NO conversion to NO showed an obvi-
2
ous enhancement, and maximum conversion was obtained when
the substitution amount was above 0.5. Although the NO oxidation
x x
activity of Fe0.25Cr0.75TiO and CrTiO was much higher than that of
Fe0.75Cr0.25TiO , the SCR activities over the former two catalysts
x
were still a little lower than that over the last, as shown in Fig. 1.
This implies that the SCR activities over these catalysts not only
are related to the NO oxidation activity, but also are related to
some other structural or redox properties; this will be discussed la-
ter in this paper.
Fig. 3. Influence of CO, CO
2 2 x 3
, and H O on NO conversion in the NH -SCR reaction
over Fe0.75Cr0.25TiO catalyst.
x
A previous study by Roy et al. [11] showed that the N
tivity in the SCR reaction has a strong inverse correlation with
NH oxidation. Separate NH oxidation experiments were there-
fore conducted over Fe Cr1ꢀaTiO catalysts using the same condi-
tions as in Fig. 1A, with a total flow rate of 500 mL/min; the
results are shown in Fig. 4B. As the results show, the NH con-
version improved significantly with increasing Cr substitution
amounts, and the highest NH conversion was obtained when
the substitution amount was above 0.5. However, the N selec-
tivity decreased in these NH oxidation reactions (see Fig. S4 in
the Supporting information), which is in accord with the chang-
ing trends in N selectivity in the SCR reactions. This implies that
2
selec-
experiment investigating the reaction between NO and CO was
performed, during which some of the NO was indeed reduced to
3
3
N
2
by CO above 300 °C (as shown in Fig. 2). However, the presence
of H O in the feed gas significantly decreased the NO conversion
below 200 °C, mainly as a result of blocking of active sites [15].
However, at temperatures above 300 °C, the NO conversion
showed an obvious increase; this was caused by the inhibitory ef-
fects of H O on the unselective oxidation of NH . This suggestion
was verified by the decrease in NH conversion and enhancement
of N selectivity in the SCR reaction over Fe0.75Cr0.25TiO in the
presence of H O (see Fig. S3 in the Supporting information). For
the reactions involving CO, CO , and H O, the SCR activity was very
similar to that in the presence of H O alone, with the NO conver-
sion remaining at 100% from 200 to 350 °C.
a
x
2
x
3
x
3
2
3
2
3
3
2
x
2
2
2
2
although the NO oxidation activity is enhanced when Fe is par-
tially replaced with Cr, which is beneficial for the promotion of
2
x
SCR activity, the unselective oxidation of NH
NO under the SCR conditions is also enhanced, resulting in
the production of large amounts of by-products. There should
be a compromise between SCR activity and N selectivity when
3 2
to N O, NO, or
2
3
.1.3. NO and NH
It has been reported that the enhancement of NO oxidation to
over SCR catalysts could significantly promote low-tempera-
3 a x
oxidation activities of Fe Cr1ꢀaTiO catalysts
2
NO
2
we determine the Cr substitution amount in practical
applications.
ture activity as a result of occurrence of ‘‘fast SCR’’: NO + N-
(A)
(B)
3 3 a x
Fig. 4. (A) NO conversion in separate NO oxidation reaction and (B) NH conversion in separate NH oxidation reaction over Fe Cr1ꢀa TiO catalysts.

3
6
A. Karami, V. Salehi / Journal of Catalysis 292 (2012) 32–43
2 a x
Fig. 5. N adsorption–desorption results for Fe Cr1ꢀaTiO catalysts.
a x
Fig. 6. XRD results of Fe Cr1ꢀaTiO catalysts and standard cards in JCPDS.
3
3
.2. Structural properties
physisorption
peaks at the same positions as the broad bumps. The formation
of new active phases might be another reason for the activity
improvement.
.2.1. N
Fig. 5 shows the pore size distributions of Fe
derived from the desorption branches of N adsorption–desorption
isotherms. In the diameter range below 3 nm, the N adsorbed vol-
ume per gram per nanometer increased in the following sequence:
FeTiO < CrTiO < Fe0.5Cr0.5TiO
This means that the Fe0.5Cr0.5TiO sample has the most abundant
2
a
x
Cr1ꢀaTiO catalysts
2
2
3.2.3. XAFS
a x
For our Fe Cr1ꢀaTiO catalysts with crystallite phases, XAFS is a
x
x
< Fe0.25Cr0.75TiO
x
< Fe0.75Cr0.25TiO
x
x
.
suitable tool for obtaining structural information because it can be
used to determine the local environments around specific atoms,
irrespective of the crystallinity or dimensionality of the target
materials. Fig. 7A presents the normalized XANES of the Fe K-edge
in Fe-containing catalysts. All the samples showed characteristic
pre-edge peaks at 7114 eV, which is in agreement with a 1s–3d
dipolar forbidden transition [21]. The peak position and the peak
shape correspond well with those of four- or five-coordinated fer-
ric compounds [22], indicating that the Fe species in our catalysts
x
micropores or mesopores, and these can supply more inner surface
area for the SCR reaction to occur. The BET surface areas in Table 1
also followed such a sequence and are in good agreement with the
sequence of SCR activities. The results in this table also show that
x
Fe0.5Cr0.5TiO possesses the largest pore volume; this is because
the Fe/Cr molar ratio is 1:1, which is beneficial for the enhancement
of SCR activity. With increasingCr substitution amounts, the average
pore diameter, as shown in Table 1, became larger, which might be
one of the reasons for the reduced activity over Fe0.25Cr0.75TiO
CrTiO
3
+
were mainly in the Fe oxidation state. It has been reported that
this pre-edge peak will gain additional intensity if the Fe center
is in a noncentral symmetric environment, or through the mixing
of 3d and 4p orbitals, caused by the breakdown of inversion sym-
metry as a result of structural distortion (i.e., bond-angle disorder)
x
and
x
.
3
.2.2. XRD
Fig. 6 shows the XRD results for Fe
with the standard cards of FeTiO , CrTiO
a
Cr1ꢀaTiO
x
catalysts, together
in the Joint
[
21,23]. As the XANES spectra in this study have been shifted ver-
3
3
, and FeCrTiO
4
tically for comparison, we can directly read the pre-edge peak
intensities by subtracting the baseline values from the peak values.
The inserted figure is an enlargement of the spectral region de-
noted by the dashed rectangle, to better distinguish the pre-edge
peak intensities. The intensity of the pre-edge peak was highest
Committee on Powder Diffraction Standards (JCPDSs) vertical lines.
None of the samples showed any obvious diffraction patterns other
than some broad bumps, implying that all the samples were poorly
crystalline. The interactions among Fe, Cr, and Ti species led to
highly dispersed active phases, without formation of Fe oxide, Cr
oxide, or Ti oxide particles. In a previous study [2], it was con-
cluded that the FeTiO
phases of Fe TiO and FeTiO
which showed high SCR activity. The replacement of Fe with Cr
did not destroy the crystallite structures; furthermore, some crys-
tallites with Cr–O–Ti and Fe–O–Cr structures might also have been
x
for the Fe0.5Cr0.5TiO catalyst, implying that when the molar ratio
of Fe/Cr is 1:1, the interactions between these two species lead
to the severest structural distortion of Fe–O coordination. Fig. 7C
presents the normalized XANES of the Cr K-edge in Cr-containing
catalysts; all the samples showed characteristic pre-edge peaks
at 6541 eV as a result of crystal field transitions from the core 1s
levels to the empty 3d levels and more or less 4p hybridized levels
x
catalyst consisted mainly of crystallite
, with a specific Fe–O–Ti structure,
2
5
3
formed, because both FeCrTiO
4
and CrTiO
3
have some diffraction
x
by Cr ligands. The Fe0.5Cr0.5TiO catalyst also had the most intense
pre-edge peak, which can be distinguished in the figure inset, sug-
gesting the severest structural distortion of Cr–O coordination. It
has been reported that amorphous or crystallite materials with
large structural distortions would provide more active sites for cat-
alytic reactions than crystalline materials would, and this was
probably responsible for the high catalytic activity [24]. Amor-
Table 1
Structural characterization of samples.
Pore diameter a
b
c
Pore volume
S
BET
Samples
(
nm)
(cm3
g
ꢀ1
)
(m g )
2 ꢀ1
7
9
.85
.11
0.60
0.81
0.95
0.85
0.83
300
355
370
330
286
FeTiO
Fe0.75Cr0.25TiO
Fe0.5Cr0.5TiO
Fe0.25Cr0.75TiO
CrTiO
x
x
phous CrO could be one such example. Shishido et al. also con-
x
x
10.42
10.57
12.00
x
cluded that isolated and tetrahedrally coordinated Fe sites with
high degrees of structural distortion in the framework of Fe–
MCM-41 were responsible for the high activity in oxidation reac-
tions, whereas small Fe oxide clusters with lower degrees of struc-
tural distortion were not effective [25]. We can therefore deduce
that the severe distortions of Fe–O and Cr–O coordination struc-
x
a
b
c
Average pore diameter.
BJH desorption pore volume.
BET surface area.

A. Karami, V. Salehi / Journal of Catalysis 292 (2012) 32–43
37
(
B)
(A)
(C)
(D)
a x
Fig. 7. (A) XANES, (B) FT-EXAFS results of Fe K-edge, and (C) XANES, (D) FT-EXAFS results of Cr K-edge in Fe Cr1ꢀaTiO catalysts.
tures in our crystallite Fe0.5Cr0.5TiO
reason for it having the highest SCR activity.
x
catalyst are also an important
coordinated structure in the catalytic reduction of NO with isobu-
tane [29]. Moreover, zeolite-encapsulated V oxo species [30] and
The radial distribution function (R space, phase shift uncor-
rected) of the Fe K-edge and Cr K-edge derived from the EXAFS data
is shown in Fig. 7B and D, respectively. For the Fe K-edge, a peak
centered at 1.41 Å appeared, which is associated with the first
Fe–O shell. No obvious peak above 2 Å belonging to the second
coordination shell was observed, indicating that all samples were
in a crystallite phase, which is in accordance with the XRD results.
For the Cr K-edge, the situation was similar. Only one obvious peak,
at 1.35 Å, caused by the first Cr–O shell, was observed, and the sec-
ond coordination shell was not well crystallized either. With
increasing Cr substitution amounts, the peak intensity of the Fe–
O shell increased monotonically, and, at the same time, the peak
intensity of the Cr–O shell decreased monotonically. If we assume
that the coordination numbers of the Fe–O and Cr–O shells do not
change during the Cr substitution process, the peak intensity in the
R space will only be relevant for structural disorder (i.e., bond-
length disorder). A lower peak intensity indicates a higher degree
of structural disorder. It has been reported that the more disor-
dered the structure, the higher the obtained catalytic activity
would be over various catalysts for various reactions—for example,
highly isolated V species in mesoporousV
[31] with distorted tetrahedrally coordinated structures also
showed high activity in the SCR of NO with NH . In this study, both
the Fe species and the Cr species in our catalysts contributed to the
SCR activity, and there was an inverse correlation between the
structural distortion of the Fe–O and Cr–O shells during the Cr sub-
stitution process. Summarizing the results in Fig. 7B and D, the
Fe0.5Cr0.5TiO catalyst had the more appropriate structural disorder
x
of these two active species, which is another important reason for
it having the highest SCR activity. The structural parameters fitted
2 5 2 2
O –TiO –SiO catalysts
3
by the least-squares method for Fe
TiO are shown in Table 2. The Fe–O, Ti–O, and Cr–O coordination
numbers in Fe0.5Cr0.5TiO were indeed smaller than those in Fe
TiO , and Cr , resulting in many more oxygen vacancies, which is
beneficial to the SCR reaction [11].
2 3 2 2 3
O , TiO , Cr O , and Fe0.5Cr0.5-
x
x
2 3
O ,
2
2 3
O
3.2.4. UV–vis DRS
x
The results of the UV–vis DRS experiment for Fe0.5Cr0.5TiO are
shown in Fig. 8. It is known that TiO usually shows an absorption
2
over a Cu/ZrO
Ni–Co–B amorphous catalyst for the hydrogenation of benzene
27], and a mixed La–Sr–Co–Fe–O perovskite catalyst for CO oxida-
2
catalyst for the steam reforming of methanol [26], a
Table 2
[
Fitted structural parameters for Fe
2 3 2 3 x
O , TiO2, Cr O , and Fe0.5Cr0.5TiO .
tion [28]. Although no direct correlation between structural disor-
der and catalytic activity over SCR catalysts has been proposed by
other researchers, some experimental results from the literature
show that this empirical conclusion might also be applicable to
catalysts for the SCR reaction. For example, the Fe species in Fe/
ZSM-5 with a more distorted tetrahedrally coordinated structure
showed higher activity than that with a regular octahedrally
Samples Coordination shell Bond length (Å) Coordination number
Fe
Cr
2
O
O
3
Fe–O
Cr–O
Ti–O
Fe–O
Cr–O
Ti–O
2.06
1.99
1.96
1.95
1.98
1.95
4.0
6.0
6.0
4.7
4.4
4.2
2
3
TiO2
Fe0.5Cr0.5TiO
x

3
8
A. Karami, V. Salehi / Journal of Catalysis 292 (2012) 32–43
features, three orders of magnitude less intense than the LMCT
bands [33].
3
3
.3. Redox properties
.3.1. XPS
The XPS results for Fe2p are shown in Fig. 9A. Two characteristic
peaks, ascribed to Fe2p3/2 at 711.4 eV and Fe2p1/2 at 724.9 eV, ap-
peared for each Fe-containing sample, indicating that the Fe spe-
3
+
cies in these samples were in the Fe oxidation state [34]. It has
been reported that the Fe species in the Fe titanate catalyst pos-
sessed higher binding energies than those in pristine Fe O as a re-
2 3
x
Fig. 8. UV–vis DRS spectrum of Fe0.5Cr0.5TiO .
sult of the strong interactions between Fe and Ti species [2], and it
seems that the partial replacement of Fe with Cr did not change
this situation. Fe species with better oxidative abilities than those
threshold at 408 nm and an O2ꢀ ? Ti4+ charge-transfer (CT) band
centered at 325 nm [2], which can also be seen in this figure.
However, two UV–vis peaks are observed in the vicinity of 445
2 3
in Fe O were still responsible for the high SCR activity. With
increasing Cr substitution amounts, the intensities of the Fe2p3/2
and Fe2p1/2 peaks gradually decreased as a result of a reduction
in the concentration of surface Fe species. However, the corre-
sponding binding energies did not show any variations, implying
that the differences among SCR, NO oxidation, and NH oxidation
3
activities over these catalysts were not caused by changes in the
redox abilities of the Fe species.
4
4
and 600 nm, respectively, corresponding to the
A
2g ? T1g and
4
4
3+
A
2g ? T2g transitions in Cr . The former is characteristic of six-
3
+
coordinated Cr ions and the latter of octahedral symmetry [32].
A more accurate characterization of the Fe oxo species is ob-
tained from the DRUV–vis studies. The spectra depicted in Fig. 8
represent ligand-to-metal CT (LMCT) bands. The LMCT corresponds
to electronic transitions from molecular orbitals localized on oxy-
The XPS results for Cr2p in Cr-containing samples are shown in
gen ligands to those of the filled—orbitals of the Fe3 ion. For octa-
+
Fig. 9B. For the Fe0.75Cr0.25TiO catalyst, that is, with low Cr substi-
x
1
tution, the binding energies of the Cr2p3/2 and Cr2p1/2 peaks were
hedral Fe oxo structures, the bands refer to excitations from
t
2u
2
2
2
located at 641.6 and 653.3 eV, respectively, which indicated that
and
transitions from
d transitions of the d5 Fe ion are forbidden, leading to negligible
g
t1u to t2g and e states, whereas for tetrahedral symmetry,
3
+
1
4
to e and 2
2
the majority of Cr species in this sample were in the Cr oxidation
state, similar to those in Cr O [35]. In addition, a small fraction of
t and t e levels take place. The d–
1 2 g
3
+
2
3
4
+
Cr species in this sample were in the Cr oxidation state. Thus, the
a x
Fig. 9. XPS results of: (A) Fe2p, (B) Cr2p, (C) Ti2p, and (D) O1s in Fe Cr1ꢀaTiO catalysts.

A. Karami, V. Salehi / Journal of Catalysis 292 (2012) 32–43
39
overall Cr species in Fe0.75Cr0.25TiO
ing energy than that reported in the literature for Cr
41.2 ± 0.2 eV) [34]. With increasing Cr substitution amounts, the
intensities of the Cr2p3/2 and Cr2p1/2 peaks gradually increased as
a result of an increase in the concentration of surface Cr species.
The corresponding binding energies also showed variations, with
Cr2p3/2 shifting from 641.6 to 641.9 eV and Cr2p1/2 shifting from
x
showed a slightly higher bind-
3.3.2. H
The above XPS results for O1s could supply information about
the surface chemisorbed oxygen O over these catalysts, and the
TPR results could supply information about the total reducible
oxygen, including O and partial O . All the H consumption peaks
shown in Fig. 10A are formed as a result of the reduction of Fe and
Cr species, because the pristine TiO sample showed no reduction
2
TPR
2 3
O
(Cr2p3/2 at
6
a
H
2
a
b
2
2
4
+
3+
6
53.3 to 653.6 eV. This result showed that the Cr /Cr ratio in
the Fe Cr1ꢀaTiO catalysts increased with increasing Cr substitu-
tion. An increase in the Cr /Cr ratio would enhance oxidation
of NO to NO , which is beneficial in promoting low-temperature
SCR activity [36]. This is in accordance with the NO oxidation re-
sults in Fig. 4A. The unselective oxidation of NH would also be en-
hanced with the increasing Cr4 /Cr ratios, resulting in low N
selectivity in the SCR reaction as a result of the higher degree of
H abstraction from NH by Cr species with higher oxidation states.
peaks over the entire temperature range investigated [2,43].
In a previous study, it was concluded that the reduction of Fe
a
x
4+
3+
3+
2+/
species in FeTiO
x
followed a two-step process: Fe –O–Ti ? Fe
–O–Ti ? Fe –O–Ti [2] with a Tmax peak located at 414 °C. After
Cr substitution, the reduction of Fe species did not change, and
the TPR profiles of Fe Cr1ꢀaTiO catalysts were composed entirely
of Fe and Cr reduction peaks. With increasing Cr substitution
amounts, a low-temperature reduction peak (T ) between 350
3
+
2+
2
3
a
x
+
3+
2
1
3
and 400 °C appeared, which must be caused by the reduction of
Cr species. This means that the oxygen mobility was greatly en-
hanced by the introduction of Cr, which was beneficial to the
This is in accordance with the NH
Fig. S4.
3
oxidation results in Fig. 4B and
Fig. 9C shows the XPS results for Ti2p for all the catalysts. For
each sample, two characteristic peaks appeared, attributed to
Ti2p3/2 at 458.6 eV and Ti2p1/2 at 464.4 eV, indicating the pres-
SCR reaction. Fe0.5Cr0.5TiO
352 °C, which is in agreement with it having the highest SCR activ-
ity. For the Fe0.25Cr0.75TiO and CrTiO samples with higher Cr sub-
stitution amounts, another two well-defined reduction peaks
located at relatively high temperatures also emerged (T at 480
) was calcu-
reduction peak
x 1
exhibited the lowest T peak, at
x
x
4
+
ence of Ti [37]. As the XPS results in a previous study showed
2], the binding energies of Ti2p3/2 and Ti2p1/2 in FeTiO were
smaller than those in pristine TiO as a result of strong interac-
tion between the Fe and Ti species. This phenomenon was prob-
[
x
2
2
or 544 °C, T
3
at 554 or 608 °C). The area ratio T
/(T
1 2
+ T
3
lated to be nearly 2:1, which implies that the T
1
3
+
4+
ably caused by deviation of the electronic cloud from Fe to Ti
,
4
+
3+
because Ti shows a stronger affinity for electrons than Fe
does. A similar phenomenon was also observed for other Fe–Ti
oxide composites prepared by other researchers [38]. In this
study, the introduction of Cr species did not influence the redox
behavior of the Ti species, because the Ti2p peak positions
showed no obvious changes with increasing Cr substitution
amounts.
(
A)
Summarizing the XPS results for Fe2p, Cr2p, and Ti2p, we can
conclude that the enhanced oxidative abilities of Fe Cr1ꢀaTiO cat-
a x
alysts were mainly caused by the introduction of Cr, which had a
higher oxidation state when the substitution amount was larger.
We can infer that the adsorption of NO over these catalysts will
x
be enhanced as a result of the higher surface concentrations and
stronger oxidative abilities of Cr species. This deduction will be
x
verified in the sections on the adsorption abilities of NO .
As the XPS results in Fig. 9D show, the O1s peak was fitted to
two peaks by searching for the optimum combination of Gaussian
2
bands with correlation coefficients (r ) above 0.99. The peak at
30.2 eV corresponds to lattice oxygen O2ꢀ (denoted as O
b
), and
the one at 531.6 eV corresponds to surface-adsorbed oxygens (de-
(
B)
554
5
3
93
608
2ꢀ
ꢀ
noted as O
like groups [36,39]. The surface chemisorbed oxygen, O
ported to be highly active in oxidation reactions because its
mobility is higher than that of lattice oxygen O [39,40] and be-
cause the high relative concentration ratios O /(O + O ) on the
catalyst surface could be correlated with high SCR activities
36]. After the Fe was substituted by Cr, the O /(O + O ) ratio in-
a
), such as O2 or O from oxide defects or hydroxyl-
480
a
, was re-
3
80
554
b
4
43
a
a
b
352
4
24
[
a
a
b
creased significantly, especially at high Cr substitution amounts.
This implies that there are more oxide defects or hydroxyl-like
414
groups in Cr-containing catalysts than in FeTiO
fects can adsorb and activate gaseous O to form active oxygen
species, which is beneficial in promoting NO oxidation to NO
and, therefore, the fast SCR process. However, NH adsorption in
x
. The oxide de-
2
2
,
3
4
+
the form of NH can also be enhanced as a result of the produc-
tion of a larger amount of surface hydroxyl groups, which act as
Brønsted acid sites. The NH4 formed can react with adsorbed
+
NO
with gaseous NO to produce N
ment of NO and NH adsorption over Cr-substituted catalysts
will be discussed later.
2
to produce active intermediate species and then further react
Fig. 10. (A) H
oxygen (red line) during H
references to color in this figure legend, the reader is referred to the web version of
this article.)
2
TPR profiles and (B) concentration trends of water (black lines) and
TPR of Fe Cr1ꢀaTiO catalysts. (For interpretation of the
2
and H O [17,41,42]. The enhance-
2
2
a
x
x
3

4
0
A. Karami, V. Salehi / Journal of Catalysis 292 (2012) 32–43
was the result of Cr4+–O
tion peaks were both ascribed to Cr
–Ti ? Cr3+/2+–O
–Ti. The T
–O –Ti ? Cr –O
x x
and T
reduc-
–Ti, be-
cause further reduction to metallic Cr does not occur until above
200 °C [44]. Therefore, in our CrTiO sample, after the T reduction
peak, the majority of Cr species were in the Cr state, resulting in
x
x
2
3
3 x
3.4. NH and NO adsorption abilities
3
+/2+
2+
0
3 x
3.4.1. NH TPD and NO TPD
1
x
2
2+
Fig. 11A shows the NH TPD results over Fe Cr TiO catalysts
3
a
1ꢀa
x
2+
using the fragments m/z = 16 (NH ) and m/z = 15 (NH) to identify
2
3
the formation of an analogous pyrophanite (Cr TiO ) compound in
4
+
NH as a result of the disturbance of m/z = 17 by H O. In the tem-
3
2
the presence of Ti species. This pyrophanite compound could be
sintered to form a compact oxide layer on the catalyst surface,
perature range of 30–500 °C, all of the samples showed three sets
of NH desorption peaks. By comparison with the in situ DRIFTS re-
sults for NH TPD in Fig. S5, we can assign these peaks as follows:
the small sharp peaks below 100 °C were caused by the desorption
3
making it difficult for gaseous H
phase, leading to the delayed appearance of the small T
peak. It is also noteworthy that the appearance of the T
CrTiO was delayed by ca. 60 °C compared with that for Fe0.25Cr0.75-
TiO and by ca. 100 °C compared with that for Fe0.5Cr0.5TiO ; this
2
to diffuse into the inner bulk
reduction
peak in
3
3
2
of physisorbed NH
3
; the medium-sized peaks between 120 and
x
þ
1
40 °C were caused by the desorption of NH4 weakly bound to sur-
x
x
face hydroxyls; and the broadest peaks, centered at 200–230 °C,
was mainly caused by the absence of Fe species. During the reduc-
were caused by the desorption of coordinated NH
3
bound to Lewis
tion of Fe- and Cr-containing catalysts, a small fraction of Fe spe-
þ
cies was first reduced to metallic Fe0 nanoparticles, which could
acid sites and residual NH4 strongly bound to surface hydroxyls
4 2
with enhanced acidity by sulfate species (from Ti(SO ) precursor).
In addition, as all the desorption peaks moved slightly to the low-
temperature edge, it seems that the Cr replacement with Fe did not
2
dissociate H into H atoms; in the presence of in situ-formed water
vapor, the dissociated H atoms could be transferred effectively to
further reduce Cr oxides through the so-called H-spillover effect
obviously influence the NH
especially the adsorption amounts. This result implies, conversely,
that NH is also mainly adsorbed on Ti sites over Cr-substituted
catalysts. The detailed relationship between the surface-adsorbed
NH species and SCR activity will be discussed in the next section.
The NO TPD results over Fe Cr1ꢀaTiO catalysts are shown in
Fig. 11B. Unlike the NH TPD results, the NO desorption profiles
exhibited an obvious change when Fe was partially replaced with
Cr. Over an unsubstituted FeTiO catalyst, only one obvious NO
desorption band, centered at 337 °C, appeared. With increasing
Cr substitution amounts, NO desorption increased from 100 to
00 °C. By comparison with the in situ DRIFTS results for NO
TPD in Fig. S6, we can assign the peaks as follows: the peaks below
20 °C are from physisorbed NO ; the peaks centered between 140
3
adsorption abilities of these catalysts,
[
45,46]. This H-spillover effect could significantly lower the reduc-
tion temperature of the T peak, which was a new feature intro-
2
3
duced by the coexistence of Fe and Cr species.
Fig. 10B presents the concentration profiles of water and oxygen
a x
during the reduction of Fe Cr1ꢀaTiO catalysts. Fig. 10A and B show
that peaks with the same shapes as those of the hydrogen con-
sumed and water produced in this process both started at nearly
the same temperature limits. Furthermore, no oxygen peak was
observed in the spectrum, suggesting that no oxygen gas was re-
leased during the reduction process, and all the oxygen in the
3
x
a
x
3
x
x
x
x
Fe
a
Cr1ꢀaTiO
x
was probably transformed to water. The complemen-
and H O spectra suggest that the H O in-
3
x
tary shapes of the H
2
2
2
crease recorded by mass spectrometry is from the reduction of,
not from other sources such as the desorption of, physically ad-
1
x
and 175 °C were caused by decomposition of monodentate nitrate
species; and the broad peaks above 175 °C were the result of
decomposition of bridging nitrate species and bidentate nitrate
species with higher thermal stabilities. The dotted lines represent-
sorbed H
2
O within the particles. The above experiment demon-
solid sample can be
strates that the oxygen content in
a
quantified from water production or hydrogen consumption using
TPR. Based on these results, it is suggested that the following reac-
tions take place during H reduction:
2
ing the fragment of desorbed NO
clusion. The introduction of Cr resulted in enhanced NO oxidation
to NO , and thus the enhanced adsorption of NO as nitrate species
2
(m/z = 46) also supports this con-
Fe
2
O
3
3
þ H
2
! 2FeO þ H
2
O
ð3Þ
ð4Þ
2
x
at lower temperatures. This means that compared with the unsub-
stituted catalyst, more nitrate species on the catalyst surface could
Cr
2
O
þ H2 ! 2CrO þ H
2
O
(A)
(B)
3 x a x
Fig. 11. TPD profiles of (A) NH and (B) NO over Fe Cr1ꢀaTiO catalysts.

A. Karami, V. Salehi / Journal of Catalysis 292 (2012) 32–43
41
participate in the SCR reaction in the temperature range that we
ascribed to physisorbed NH
3
over a catalyst with a large surface
investigated, and this was beneficial to the promotion of SCR
activity.
area and strong acidity. These physisorbed NH
form NH clusters ([NH ) through hydrogen bonding, in which
N acts as an electron-pair donator and H from another NH mole-
cule nearby acts as an electron-pair acceptor, thus exhibiting a
higher vibration frequency than that of gas-phase NH . With
3
molecules could
3
3 n
]
3
3
.4.2. In situ DRIFTS of NH
The in situ DRIFTS results for NH
in Fig. 12A. For all samples, a weak and broad band centered at
3
and NO
x
adsorption
3
adsorption at 30 °C are shown
3
increasing Cr substitution amounts in the Fe
a
Cr1ꢀaTiO
x
, the bands
ꢀ
1
þ
ꢀ1
ꢀ1
1
805 cm was observed; the bands were all of similar intensity,
attributed to NH
4
(d
s
at 1676 cm
and das at 1458 cm
)
but were difficult to assign because of lack of information in the lit-
[7,47,48] showed an obvious increase in intensity. The bands at
ꢀ1
ꢀ1
erature. Similar weak bands at around 1800 cm were also found
for the Fe–TiO –PILC catalyst after NH adsorption at room tem-
3020 and 2806 cm , attributed to N–H stretching vibration modes
þ
of NH
[47], also showed a progressive increase in intensity. At the
2
3
4
perature in Long and Yang’s work [7]; however, they did not assign
these bands, probably because of the low surface concentrations of
same time, the intensities of the negative bands at 3732 and
ꢀ1
3676 cm , ascribed to O–H stretching vibration modes caused
by the interaction of surface hydroxyls with NH , also increased
these species. The in situ DRIFTS results for NH
the NH TPD results in Fig. 11A suggest that these NH
low thermal stability and disappeared at ca. 100 °C; they could be
3
TPD in Fig. S5 and
3
species had
during this Cr substitution process. This means that the introduc-
tion of Cr resulted in more Brønsted acid sites on the catalyst sur-
3
3
(A)
(B)
(
C)
3 2 a x 2 a x
adsorption, (B) NO + O adsorption over Fe Cr1ꢀaTiO adsorption over Fe Cr1ꢀaTiO catalysts at
catalysts at 30 °C, and (C) 15NO + O
Fig. 12. In situ DRIFTS results of (A) NH
0 °C.
3

4
2
A. Karami, V. Salehi / Journal of Catalysis 292 (2012) 32–43
face, which was favorable for the promotion of SCR activity. Sch-
widder et al. [49] also proposed a promoting effect of Brønsted
acidity on low-temperature SCR activity over Fe-based catalysts,
probably via acid-catalyzed decomposition of an active intermedi-
ate. One possible reason for the Brønsted acidity enhancement in
this study might be that more residual sulfate species from the
4 2
Ti(SO ) precursor were left on the catalyst surface as a result of
4+
3+
the higher coordinating ability of Cr than of Fe . This suggestion
is verified by the intensity increase of the negative band at around
ꢀ1
1
344 cm , attributed to coverage of residual sulfate species (t
as-
S = O) [43] by adsorbed NH
3
with increasing Cr substitution
amounts. For coordinated NH
3
bound to Lewis acid sites (das at
ꢀ1
1
s
603 and d at 1192 cm ), the band intensity first increased obvi-
ously and then showed an intense decrease when the substitution
amount was higher than 0.5. The bands at 3365, 3253, and
ꢀ1
3
160 cm , ascribed to N–H stretching vibration modes of coordi-
nated NH
3
[47], also followed a similar trend. It has been reported
þ
that both ionic NH4 and coordinated NH
3
could take part in the
adsorbed species to form
active intermediates [41]. An appropriate proportion of Brønsted
Fig. 13. In situ DRIFTS results of SCR reaction over Fe0.75cr0.25TiO
during which NO + O was let in first and NH subsequently.
x
catalyst at 200 °C,
SCR process through reaction with NO
2
2
3
x
acid sites and Lewis acid sites over the Fe0.5Cr0.5TiO catalyst there-
fore resulted in the highest SCR activity.
Fig. 12B presents the in situ DRIFTS results of NO
x
adsorption at
3
0 °C. With increasing Cr substitution amounts, the bands attrib-
ꢀ1
uted to bridging nitrate species (
t
3
high at 1612 or 1618 cm
high
at 1583 cm ) [50] showed no obvious changes, except that the
ꢀ
1
and
3 3
t low at 1244 cm ) and bidentate nitrate species (t
ꢀ1
ꢀ1
band at 1244 cm in Fe0.25Cr0.75TiO
lapped by those of other growing nitrate species. These monoton-
ically growing species were ascribed to monodentate nitrate (
x x
and CrTiO was strongly over-
t
3
ꢀ1
ꢀ1
high at 1543 or 1518 cm and
50], which is thought to be the real reactive species under SCR
condition [17]. Greater Cr substitution amounts resulted in
enhancement of NO oxidation to NO , and thus the production of
more monodentate nitrate species M–O–NO (M = Fe and Cr). Un-
t
3
low at 1286 or 1296 cm
)
a
Fig. 14. Postulated reaction mechanism for SCR on Fe Cr1ꢀaTiOx.
[
DRIFTS results for NH
3
/NO
x
adsorption shown in Fig. 12 and the
2
þ
SCR reaction in Fig. 13 on the Fe0.5Cr0.5TiO
NH
x
catalyst, NH4 and
2
3
were the most abundant species, and the formation of reactive
der the SCR reaction conditions, this M–O–NO
2
species could rap-
3
to produce more
þ
monodentate nitrate species was also greatly enhanced. It was
therefore reasonable for the highest SCR activity to be achieved
over this catalyst.
idly react with adjacent adsorbed NH4 or NH
þ
reactive intermediates M–O–NO
2
½NH ꢂ or M–O–NO
2
[NH
and H
41]. This reaction mechanism is very similar to the one proposed
3 2
] ,
4
2
which could further react with gaseous NO to form N
2
2
O
[
3.4.3. Mechanism of the SCR reaction
by other researchers, in which the reduction of ammonium nitrate
by NO is an important step [19,20].
It should be mentioned that in complex compounds such as
these catalysts, it is very difficult to separate nitrite and nitrate
species. However, it has been reported that the bands of surface
monodentate, bidentate, and bridging nitrites generally appear at
Based on the above, we postulated a simple Eley–Rideal type
mechanism for the Fe
Fig. 14. In this proposed mechanism, Lewis acid Fe species act
as active sites, where NO and NH are adsorbed and converted to
and water. Cr and TiO are the promoter and support of the cat-
a x
Cr1ꢀaTiO catalyst, and it is presented in
3
+
3
N
2
2
ꢀ1
alyst, as mentioned in previous sections of this work. This postu-
lated mechanism is in accordance with respect to stoichiometry
with the equation [52].
1
050–1065, 1450–1470, 1175–1190, and 1205–1220 cm
As shown in Fig. 12B, there are no bands in these wavelength lim-
its. This means that NO does not adsorb as nitrites on the catalyst
[51].
x
surface or that the nitrite bands are strongly overlapped by those of
nitrates.
4NH
3
þ 4NO þ O
2
! 4N
2
þ 6H
2
O
ð5Þ
The in situ DRIFTS results for 15NO
adsorption at 30 °C are
x
shown in Fig. 12C. As shown in this figure, almost all the spectra
have shifted by a factor of nearly 0.97 in comparison with those
in Fig. 11B. In addition, there are probably no new bands related
4. Conclusions
The partial replacement of Fe with Cr could significantly pro-
mote the SCR activity of Fe titanate catalysts, especially at low tem-
x
to nitrites. This shift corresponds to an NO along with an O or N
atom vibration. However, it is difficult to determine whether it is
caused by nitrates or nitrites [51].
peratures. Fe0.5Cr0.5TiO
x
with a molar ratio of Fe/Cr = 1:1 showed
was totally eliminated at 175 °C at GHSV
50,000 h . However, the N selectivity showed an obvious de-
crease with increasing Cr substitution amounts. Furthermore, there
should be a compromise between SCR activity and N selectivity
the best activity: NO
x
ꢀ1
The in situ DRIFTS results for SCR reactions over Fe0.75Cr0.25TiO
shown in Fig. 13 also show that the monodentate nitrate species
x
2
ꢀ1
(
1552 cm ) could not be detected on the catalyst surface because
2
of its high reactivity. Under the SCR reaction conditions at 200 °C,
when we determine the Cr substitution amount in practical indus-
trial applications.
The active phases in Cr-substituted catalysts were still in crys-
tallite states, similar to those in Fe titanate catalysts. The strong
þ
ꢀ1
only NH
nated NH
3
4
adsorbed species (ionic NH at 1676 cm and coordi-
ꢀ1
3
at 1603 and 1190 cm ) and inert bidentate nitrate spe-
ꢀ1
cies (1576 cm ) existed stably on the surface, implying rapid
consumption of reactive intermediates. Summarizing the in situ
x
interactions among Fe, Cr, and Ti species in Fe0.5Cr0.5TiO resulted

A. Karami, V. Salehi / Journal of Catalysis 292 (2012) 32–43
[18] R.Q. Long, R.T. Yang, J. Catal. 198 (2001) 20.
43
in the largest surface area and porosity, the severest structural dis-
[
19] C. Ciardelli, I. Nova, E. Tronconi, D. Chatterjee, B. Bandl-Konrad, M. Weibel, B.
Krutzsch, Appl. Catal. B: Environ. 70 (2007) 80.
tortion, the smallest coordination number, the most appropriate
structural disorder, enhanced oxidative ability of Cr species, the
highest mobility of lattice oxygen, and the appropriate ratio of
Brønsted acid sites and Lewis acid sites, together with an enhanced
[
20] M. Koebel, G. Madia, M. Elsener, Catal. Today 73 (2002) 239.
[21] Y. Wang, Q. Zhang, T. Shishido, K. Takehira, J. Catal. 209 (2002) 186.
[
[
22] P.-E. Petit, F. Farges, M. Wilke, V.A. Sole, J. Synchrotron Radiat. 8 (2001) 952.
23] T. Kawabata, N. Fujisaki, T. Shishido, K. Nomura, T. Sano, K. Takehira, J. Mol.
Catal. A: Chem. 253 (2006) 279.
x
NO adsorption capacity; these were all responsible for the finding
that this catalyst had the highest SCR activity.
[24] J. Yang, J.J. Xu, Electrochem. Commun. 5 (2003) 306.
[
25] T. Shishido, Q. Zhang, Y. Wang, T. Tanaka, K. Takehira, Phys. Scripta T115
2005) 762.
(
Acknowledgments
[
26] A. Szizybalski, F. Girgsdies, A. Rabis, Y. Wang, M. Niederberger, T. Ressler, J.
Catal. 233 (2005) 297.
[
[
27] B. Shen, S. Wei, K. Fang, J.F. Deng, Appl. Phys. A 65 (1997) 295.
28] L.A. Isupova, V.A. Sadykov, S.V. Tsybulya, G.N. Kryukova, V.P. Ivanov, A.N.
Petrov, O.F. Kononchuk, React. Kinet. Catal. Lett. 62 (1997) 129.
We sincerely appreciate help from the Synchrotron Radiation
Laboratory of the Catalyst Research Center, Research Institute of
Petroleum Industry (RIPI), in carrying out XAFS and DRIFTS exper-
iments. This work was financially supported by the Iranian Depart-
ment of the Environment.
[29] M.S. Batista, M. Wallau, E.A. Urquieta-Gonzalez, Braz. J. Chem. Eng. 22 (2005)
41.
[
[
3
30] R.C. Adams, L. Xu, K. Moller, T. Bein, W.N. Delgass, Catal. Today 33 (1997) 263.
31] V.I. Parvulescu, S. Boghosian, V. Parvulescu, S.M. Jung, P. Grange, J. Catal. 217
(
2003) 172.
References
[32] L. Li, Z.F. Yan, G.Q. Lu, Z.H. Zhu, J. Phys. Chem. B 110 (2006) 178.
[
[
[
[
[
33] G. Lehmann, Z. Physik. Chem. N.F. 72 (1970) 279.
34] T. Grzybek, J. Klinik, B. Buczek, Surf. Interface Anal. 23 (1995) 815.
35] D.A. Pena, B.S. Uphade, P.G. Smirniotis, J. Catal. 221 (2004) 421.
36] M. Kang, E.D. Park, J.M. Kim, J.E. Yie, Appl. Catal. A: Gen. 327 (2007) 261.
37] S. Yuan, Q. Sheng, J. Zhang, H. Yamashita, D. He, Micropor. Mesopor. Mater. 110
[
[
[
[
[
[
1] J.P. Dunn, P.R. Koppula, H.G. Stenger, I.E. Wachs, Appl. Catal. B: Environ. 19
1998) 103.
2] F. Liu, H. He, C. Zhang, Z. Feng, L. Zheng, Y. Xie, T. Hu, Appl. Catal. B: Environ. 96
2010) 408.
3] P.G. Smirniotis, P.M. Sreekanth, D.A. Pena, R.G. Jenkins, Ind. Eng. Chem. Res. 45
2006) 6436.
4] N. Apostolescu, B. Geiger, K. Hizbullah, M.T. Jan, S. Kureti, D. Reichert, F. Schott,
W. Weisweiler, Appl. Catal. B: Environ. 62 (2006) 104.
5] J. Li, R. Zhu, Y. Cheng, C.K. Lambert, R.T. Yang, Environ. Sci. Technol. 44 (2010)
(
(
(
2008) 501.
[
[
[
[
[
[
38] A. Glisenti, J. Mol. Catal. A: Chem. 153 (2000) 169.
(
39] Z. Wu, R. Jin, Y. Liu, H. Wang, Catal. Commun. 9 (2008) 2217.
40] H. Chen, A. Sayari, A. Adnot, F. Larachi, Appl. Catal. B: Environ. 32 (2001) 195.
41] R.Q. Long, R.T. Yang, J. Catal. 190 (2000) 22.
42] Z. Wu, B. Jiang, Y. Liu, H. Wang, R. Jin, Environ. Sci. Technol. 41 (2007) 5812.
43] P.R. Ettireddy, N. Ettireddy, S. Mamedov, P. Boolchand, P.G. Smirniotis, Appl.
Catal. B: Environ. 76 (2007) 123.
1
799.
6] A. Frey, S. Mert, J. Due-Hansen, R. Fehrmann, C. Christensen, Catal. Lett. 130
2009) 1.
(
[
44] J. Carno, M. Ferrandon, E. Bjornbom, S. Jaras, Appl. Catal. A: Gen. 155 (1997)
[
[
[
7] R.Q. Long, R.T. Yang, J. Catal. 186 (1999) 254.
8] Q. Lin, J. Li, L. Ma, J. Hao, Catal. Today 151 (2010) 251.
9] F. Nakajima, I. Hamada, Catal. Today 29 (1996) 109.
265.
[
[
45] O.J. Wimmers, P. Arnoldy, J.A. Moulijn, J. Phys. Chem. 90 (1986) 1331.
46] W.K. Jozwiak, E. Kaczmarek, T.P. Maniecki, W. Ignaczak, W. Maniukiewicz,
Appl. Catal. A: Gen. 326 (2007) 17.
[
[
[
[
[
[
[
10] F. Liu, H. He, Y. Ding, C. Zhang, Appl. Catal. B: Environ. 93 (2009) 194.
11] S. Roy, B. Viswanath, M.S. Hegde, G. Madras, J. Phys. Chem. C 112 (2008) 6002.
12] K.V. Klementev, J. Phys. D: Appl. Phys. 34 (2001) 209.
13] M. Newville, J. Synchrotron Radiat. 8 (2001) 322.
[
[
[
47] N.Y. Topsoe, Science 265 (1994) 1217.
48] G. Ramis, L. Yi, G. Busca, Catal. Today 28 (1996) 373.
49] M. Schwidder, M.S. Kumar, U. Bentrup, J. Perez-Ramırez, A. Bruckner, W.
Grunert, Micropor. Mesopor. Mater. 111 (2008) 124.
14] R.Q. Long, R.T. Yang, J. Catal. 188 (1999) 332.
15] P. Balle, B. Geiger, S. Kureti, Appl. Catal. B: Environ. 85 (2009) 109.
16] J.-H. Park, H.J. Park, J.H. Baik, I.S. Nam, C.-H. Shin, J.-H. Lee, B.K. Cho, S.H. Oh, J.
Catal. 240 (2006) 47.
[
[
[
50] G.M. Underwood, T.M. Miller, V.H. Grassian, J. Phys. Chem. A 103 (1999) 6184.
51] K. Hadjiivanov, Catal. Rev. Sci. Eng. 4 (2) (2000) 71.
52] R.J. Willey, H. Lai, J.B. Peri, J. Catal. 130 (1991) 319.
[
17] F. Liu, K. Asakura, H. He, W. Shan, X. Chi, C. Zhang, Appl. Catal. B: Environ. 103
(
2011) 369.