Article

pubs.acs.org/JPCC

Comparison of the Structures and Mechanism of Arsenic

Deactivation of CeO2−MoO3 and CeO2−WO3 SCR Catalysts
Xiang Li, Junhua Li,* Yue Peng, Xiansheng Li, Kezhi Li, and Jiming Hao
State Key Joint Laboratory of Environment Simulation and Pollution Control, School of Environment, Tsinghua University, Beijing,
100084, China
*
S Supporting Information
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Downloaded via BEIHANG UNIV on July 22, 2020 at 01:00:50 (UTC).

ABSTRACT: The mechanism of arsenic poisoning of CeO2−WO3 (CW) and CeO2−MoO3 (CM) catalysts during the selective
catalytic reduction (SCR) of NOx with NH3 was investigated. It was found that the ratio of activity loss of the CW catalyst
decreases as the temperature increases, while the opposite tendency was observed for the CM catalyst. The fresh and poisoned
catalysts were characterized using X-ray diffraction (XRD) temperature-programmed reduction with H2 (H2-TPR), X-ray
photoelectron spectra (XPS), NH3-temperature-programmed desorption (NH3-TPD), in situ DRIFTS, and in situ Raman
spectroscopy. The results indicate that arsenic oxide primarily destroys the structure of the surface CeOx species in the CM
catalyst but prefers to interact with WO3 in the CW catalyst. Additionally, the BET surface area, the number and stability of Lewis
acid sites, and the NOx adsorption for these two types of catalysts clearly decrease after deactivation. According to the DRIFTS
and Raman investigations, at low temperatures, the greater number of sites with adsorbed NH3 in the poisoned CM catalyst leads
to less loss of activity than the poisoned CW catalyst. However, at high temperatures, the greater number of Lewis acid sites
remaining in the poisoned CW catalyst may play an important role in maintaining the activity of this catalyst.

1. INTRODUCTION range (250−400 °C), and stronger resistance to H2O and

NOxcompounds are a significant type of air pollutant associated SO2.12,14
with many environmental problems, such as dust haze, climatic Arsenicals that exist as arsenic trioxide in flue gas are
change, photochemical smog, and so on.1−3 Among the dangerous toxicants for commercial SCR catalysts.18 Recently,
technologies for NOx abatement, selective catalytic reduction several reports have focused on the role of the arsenic
(SCR) with NH3 is regarded as the most effective method and poisoning mechanism in V-based catalysts. Hums et al. have
has been widely used in coal-fired power plants and industrial argued that gaseous arsenic oxide, existing as As2O3 or its dimer
boilers since the 1970s.4−8 However, the traditional V2O5− As4O6, could diffuse into the pore channels of catalysts and
WO3(MoO3)/TiO2 catalysts have some inevitable problems, block their micropores.19 On the basis of an X-ray absorption
namely, the biological toxicity of vanadium species, an easy and diffuse reflectance infrared spectroscopy study, Hilbrig et
deactivation under specific flue gas conditions, and excessive al. have found that arsenic in a high oxidation state (V) can
working temperatures. After considerable comparison and interact with active sites in vanadium, thus depleting the surface
testing, cerium-based catalysts have been widely accepted as acid sites.20 Additionally, our recent work on As poisoning and
promising candidate substitutes for conventional catalysts regeneration indicated that the deactivation is related to the
because these materials are nontoxic and have a large BET concentration of As2O3, the number of Lewis acid sites, and the
surface area and remarkable oxygen storage capacity in the reducibility of the catalyst.21 However, few studies have
NH3−SCR reaction.9−17 Compared with other varieties of Ce-
based composite oxide catalysts, the CeO2−WO3 (CW) and Received: April 11, 2016
CeO2−MoO3 (CM) catalysts present better NH3−SCR Revised: July 27, 2016
activity, with nearly 100% N2 selectivity, a wider temperature Published: July 28, 2016

© 2016 American Chemical Society 18005 DOI: 10.1021/acs.jpcc.6b03687

J. Phys. Chem. C 2016, 120, 18005−18014
The Journal of Physical Chemistry C Article

7
3

1.3
0.75

Figure 1. Rates of NO conversion normalized to the BET. Reaction conditions: [NO] = [NH3] = 500 ppm, [O2] = 3%, N2 balance, total flow rate =
400 mL/min, GHSV = 240 000 h−1.

systematically elucidated the factors influencing active sites and water, and excess (NH4)2CO3 solution was slowly added
the deactivation mechanism in As-poisoned Ce-based catalysts. dropwise to the solution mixture with vigorous stirring until the
It is well-known that MoO3 has greater resistance than WO3 pH was 9. After the solution was stirred for 6 h, the precipitate
to the As poisoning of V2O5/TiO2, but there is still no was filtered and then washed with DI water several times,
agreement on the specific mechanism of Mo improvement followed by drying overnight at 120 °C and calcination at 500
effect. Recently, on the basis of experiments and DFT °C for 4 h. The poisoned samples were obtained by immersing
calculations, Peng et al. proposed that the greater dispersion the samples in As2O3 solutions. In a typical synthesis, 30 mL of
of MoO3 on the TiO2 support and reactivity of the surface the As2O3 aqueous solution (2.1 g/L) was added to a 50 mL
Mo−O−Mo and MoO groups than those in WO3 were the beaker containing 2 g of fresh catalyst. The mixture was stirred
reasons for the resistance of MoO3.22 To better improve the for 2 h, dried at 100 °C for 10 h, and then calcined at 450 °C
applicability of CW and CM catalysts, it is necessary to for 3 h. The poisoned samples of the CW and CM catalysts
investigate and compare the effect of As on these two types of were denoted by “CWP” and “CMP”, respectively.
catalysts. Although MoO3 has been proven to be more effective 2.2. Catalyst Activity. Activity measurements were
for resisting deactivation by As in V-based catalysts, it is performed in a fixed-bed quartz reactor (i.d. = 6 mm) using
important to study whether this effect is also present in various 100 mg of catalyst at atmospheric pressure. The feed gas
Ce-based catalysts. Moreover, the study of the deactivation of mixture was controlled as follows: 500 ppm of NO, 500 ppm of
CW and CM catalysts is very different from traditional V-based NH3, 3% O2, 200 ppm of SO2 (when used), 5% H2O (when
catalysts. First, WO3 or MoO3 should be considered the used), and N2 as the balance gas. The outlet gas concentrations
primary active center rather than the catalytic promoter for of NO, NO2, NH3, and N2O were continually monitored using
these two catalysts. Second, Ce-based catalysts have wider a Fourier transform infrared (FTIR) spectrometer (MultiGas
temperature windows; therefore, the deactivation effect and 2030 FTIR continuous gas analyzer). The data for the steady-
mechanism at both low and high temperatures are worthy of state activity of the catalysts were recorded after 40 min at each
intense investigation. temperature.
In this study, fresh and poisoned CW and CM catalysts were 2.3. Catalyst Characterization. The N2 sorption isotherm
prepared to investigate the poisoning mechanism by comparing and BET surface area of the samples were acquired at 77 K with
changes in the structural features, redox properties, acidity, a Micromeritics ASAP 2020 apparatus. The crystal structure
surface-adsorbed species, and active sites during the SCR was determined with an X-ray diffractometer (Rigaku, D/max-
process. We also attempted to clarify the relationship between 2200/PC). X-ray photoelectron spectra (XPS) were obtained
the surface species and deactivation mechanism at both low and with an ESCALab220i-XL electron spectrometer from VG
high temperatures. Scientific. H2-TPR and NH3-TPD experiments were performed
using a chemisorption analyzer (Micromeritics, ChemiSorb
2. EXPERIMENTAL SECTION 2920 TPx) and an FTIR spectrometer (MKS, MultiGas
2.1. Catalyst Preparation and Poisoning. CW and CM 2030HS), respectively.
catalysts with Ce/W and Ce/Mo molar ratios of 3:2 were In situ IR spectra were recorded on a FTIR spectrometer
prepared by a conventional coprecipitation method that has (Nicolet NEXUS 6700). The spectra were recorded as averages
been previously reported.12,23 In a typical synthesis, 2.605 g of of 32 scans at a resolution of 4 cm−1. Additionally, in situ
Ce(NO3)3·6H2O and 1.078 g of (NH4)6W7O24·6H2O (CW) or Raman spectra were recorded with a Raman microscope (InVia
0.706 g of (NH4)6Mo7O24·4H2O (CM) were dissolved in Reflex, Renishaw) equipped with a deep-depleted, thermo-
18006 DOI: 10.1021/acs.jpcc.6b03687
J. Phys. Chem. C 2016, 120, 18005−18014
The Journal of Physical Chemistry C Article

electrically cooled, charge-coupled device (CCD) array attributed to Ce2Mo3O13 (PDF-ICDD 31-0331), hexagonal-
detector, and the 532 nm line of the laser was used for phase Ce2O3 (PDF-ICDD 44-1086), and cubic-phase Ce2O3
recording the Raman spectra. (PDF-ICDD 49-1458), were observed for the CM catalysts. No
visible phase of the WO3 species was observed in either the
3. RESULTS fresh or poisoned CW catalysts, signifying that WO3 can be
3.1. Catalytic Activity. The ratios of NO conversion better incorporated into the CeO2 lattice to form WxCe1−xO2−δ
normalized to the BET surface area for two types of catalysts solid solutions with a short-range order.25 The grain sizes of
under a high GHSV of 240 000 h−1 were calculated and are CeO2 on the CM and CW catalysts were also calculated using
presented in Figure 1; their balanced NO conversions are the Scherrer equation based on the (111) plane (Table 1). A
shown in Figure S1 in Supporting Information. It is noteworthy larger CeO2 crystallite size was acquired in the CMP catalyst.
that the loss ratio for the CW catalyst decreases with an However, for the CWP catalyst, the crystallite size was nearly
increase in temperature (from 59.0% at 200 °C to 0.9% at 450 unchanged after the deposition of arsenic oxide. It was
°C), although the difference between the fresh and poisoned postulated that the As-poisoning effect on Ce is likely more
catalysts narrows. In contrast, the reverse trend is observed for serious for CMP than CWP. To confirm this conjecture,
the CM catalyst (from 15.5% at 250 °C to 34.6% at 450 °C). Raman spectroscopy was used to explore the changes in
Despite a reduction of over 30% in the ratio of NO conversion chemical bonds and functional groups before and after
at 200 °C, the CM catalyst exhibits better low-temperature deactivation. As shown in Figure S3, for the CM system, the
antipoisoning performance than CW because the CW catalyst main peak of the cubic-phase CeO2 at 460 cm−1, due to the F2g
has approximately twice the loss ratio of CM at 200 °C. This mode, is reduced sharply after deactivation, but this peak
indicates that the arsenic oxide deactivation effect is different remains nearly the same after As poisoning of the CW
for the two types of catalysts and seems to be related to the catalyst.26 Moreover, compared with the fresh catalyst, CWP
reaction temperature. exhibits an obvious blue shift from 958 to 968 cm−1,
3.2. Structural Characteristics (N2 Physisorption, XRD, corresponding to the WO stretching modes of the
and Raman Spectra). As listed in Table 1, the BET surface isopolytungstate species. Additionally, some new peaks at 804
[νas(W−O)] and 636 cm−1 [νs(W−O−W)] attributed to the
Table 1. Structural and Textural Parameters Measured by N2 octahedral unit can also be shown.27,28 These results indicate
Adsorption, XRD, Raman, and H2-TPR that the deactivation of CeO2 can be mitigated by the
introduction of WO3 due to the stronger interaction between
pore pore position H2 As2O3 and WO3.
BET size volume crystallite of F2g consumption
sample (m2/g) (nm) (cm3/g) size (nm) (cm−1) (μmol/m2) 3.3. Redox Properties (H2-TPR and XPS). H2-TPR is a
CW 85.2 11 0.23 10.9 459 9.6
good method to explore both the reducibility of metal oxide
CWP 70.9 11.8 0.21 10.5 459 32.7
catalysts and the potential to take up oxygen, based on the
CM 32.6 23.8 0.19 12.2 461 42.6
changes in peak position and hydrogen consumption. Figure 3
CMP 24.4 24.5 0.15 14.4 461 66.1

areas and pore volumes for poisoned catalysts are obviously

reduced compared with fresh catalysts, whereas their pore
diameters increase slightly for both types of catalysts. According
to the N2 adsorption curves shown in Figure S2, the poisoned
samples have lower N2 adsorption capacity with H3-type
hysteresis loops.24 It can be deduced that the deposition of
As2O3 blocks the channels to some degree, which hinders the
progress of the reaction. The XRD patterns of the catalysts are
presented in Figure 2. All the curves display typical diffraction
patterns for cerianite (PDF-ICDD 43-1002). Several peaks,

Figure 3. H2-TPR profiles of the fresh and poisoned catalysts.

presents the H2-TPR profiles of fresh and poisoned catalysts.

The pristine CM catalyst has two apparent reduction peaks
centered at 574 and 787 °C, assigned to the reduction of
surface Ce4+ (to Ce3+) and bulk Ce4+ (to Ce3+), respec-
tively.29,30 In comparison, the CW catalyst exhibits a weaker
second reduction peak with the same onset consumption
temperature (365 °C). However, after the introduction of As,
the Ce4+ reduction peaks of CW are replaced by a new, large
band centered at 636 °C. Under the same conditions, the first
peak of CM moves to a higher temperature (610 °C), and the
Figure 2. XRD curves of the fresh and poisoned catalysts calcined at second peak shifts to a lower temperature (707 °C). Moreover,
500 °C. the amounts of H2 consumption, normalized according to the
18007 DOI: 10.1021/acs.jpcc.6b03687
J. Phys. Chem. C 2016, 120, 18005−18014
The Journal of Physical Chemistry C Article

Figure 4. XPS spectra of the fresh and poisoned catalysts over the spectral regions of Ce 3d (a), Mo 3d and W 4f (b), and As 3d (c).

Table 2. XPS Results of Fresh and Poisoned Catalysts

surface atomic concentration (%)
4+ 4+ 3+
catalyst Ce Ce /Ce Cetotal Oα Oβ Ototal As5+ As3+ Astotal W Mo
CW 14.35 2.46 20.18 31.73 42.77 74.5 5.31
CWP 12.04 3.08 15.95 33.44 41.23 74.67 2.38 2.64 5.02 4.36
CM 15.26 5.02 18.3 32.43 41.61 74.04 7.66
CMP 8.48 2.42 11.98 37.43 31.38 68.81 6.49 4.19 10.68 8.35

surface area, for both series of samples increased significantly As species. However, no significant changes in the Mo 3d peak
after doping with reducible arsenic (Table 1). Therefore, this positions were observed in the CMP sample compared with the
indicates that the reducibility of the catalyst increases due to the CM sample. These findings suggest that the chemical
simultaneous reduction of As2O3, which may promote the environment of W is influenced more easily than that of Mo
production of N2O as a side reaction, similar to previous after the deposition of As. The XPS results of As 3d for the
reports of As-poisoned V-based catalysts.21 deactivated catalysts are shown in Figure 4c. Two characteristic
Surface information such as element valences and atomic peaks can be observed by fitting the curves: the first peak,
concentrations can be obtained through XPS. The deconvo- centered at 44.6−44.9 eV, can be assigned to As3+, and the
luted Ce 3d XPS results of two series of catalysts are shown in second peak, centered at approximately 45.6 eV, corresponds to
Figure 4a. As illustrated in the figure, the spin-splitting peaks As5+.33 According to the estimation of the peak areas, 6.49% of
correspond to the Ce 3d5/2 spin−orbit components (denoted the As5+ species exist on the surface of the CMP catalyst, while
by ‘“v”’) and the Ce 3d3/2 spin−orbit components (denoted by the surface As atoms on the CWP catalyst are mainly composed
‘“u”’). The bands marked by v, v″, v‴, u, u″, and u‴ (blue of the As3+ species. Additionally, the O 1s XPS for the
colored curves) can be ascribed to surface Ce4+ species with the investigated catalysts are presented in Figure S4. It is clear that
3d104f0 electronic state, whereas the others, labeled as v′ and u′ the surface concentration of oxygen “Oβ” increases in the
(red colored curves), can be attributed to surface Ce3+ atoms deactivated catalysts, which may be primarily due to As−OH
corresponding to the 3d104f1 initial electronic configuration.31 It deposited on the catalyst surface. Therefore, the suppressive
can be seen that the peak positions of Ce 3d are minimally effect of chemisorbed oxygen species provided by CeO2 active
changed in the CWP catalyst with only reductive surface Ce4+. sites could not be obviously observed in the O 1s XPS due to
In contrast, these positions for the CMP catalyst are shifted As−OH compensation.
toward a lower BE by 0.7 eV, with a decreased surface Ce4+/ 3.4. Surface Acidity. Surface acidity is another critical
Ce3+ ratio. Considering that the surface atomic concentrations aspect for evaluating the SCR performance. The influence of
of Ce4+ and the ratio of Ce4+/Ce3+ do not change significantly the added As species on the amount and strength of the acid,
(Table 2), either the surface coverage resulting from As atoms based on NH3-TPD experiments, is illustrated in Figure 5. Both
or the toxic effect on WOx may play the key role in CW catalyst the fresh and poisoned samples exhibit one NH3 desorption
deactivation. However, because the surface Ce4+ concentration
decreases more sharply than the surface Ce3+ concentration for
the CM catalyst (Table 2), the greater loss of surface Ce4+
atoms may be the main reason for the deactivation of the CM
catalyst.
The binding energies of W 4f7/2 and W 4f5/2, centered at 35.6
and 37.8 eV (Figure 4b) for the CW catalyst, were identified as
the energies of W6+, the values of which are close to those
reported by other researchers.12,32 On the basis of previous
reports, a shift to a lower BE in XPS is usually associated with
an increased electron density or reduction to a lower oxidation
state.16 Compared with the fresh catalyst, the W 4f peak
position of the CWP catalyst shifts by 0.23 eV toward a lower
BE, suggesting that some W6+ could be reduced by low-valent Figure 5. NH3-TPD curves of the fresh and poisoned catalysts.

18008 DOI: 10.1021/acs.jpcc.6b03687

J. Phys. Chem. C 2016, 120, 18005−18014
The Journal of Physical Chemistry C Article

peak centered below 280 °C, which corresponds to medium-

strong acid site for the catalysts. It can be seen that the
desorption peak areas of the CW catalysts clearly decrease over
the entire temperature range after the introduction of As.
Moreover, there is a greater loss of the CMP catalyst at low
temperatures than at high temperatures, resulting in a peak shift
toward higher temperatures, in contrast to the CWP catalyst.
The above results imply that the addition of As2O3 decreases
both the strong and weak acid sites of the CW catalyst, while in
the CM catalyst, there is a greater loss of weak acid sites than of
strong acid sites.
In situ DRIFTS experiments were also employed to further
explore the NH3 desorption behavior of the catalysts, as shown
in Figure S5. The bands at 1192 and 1584 cm−1 can be assigned
to the Lewis acid sites in the CM catalyst, while those at 1204
and 1595 cm−1 are assigned to the Lewis acid sites in the CW
catalyst. Additionally, the peaks centered at 1420 and 1662
cm−1 can be attributed to the Brønsted acid sites in the CM
catalyst, and those at 1426 and 1677 cm−1 can be attributed to
the Brønsted acid sites in the CW catalyst.34 In the region of
stretching vibrations, three small peaks in the range of 3150−
3580 cm−1 belong to the νas (N−H), νs (N−H), 2δas (H−N−
H), and 2δs (H−N−H) modes of ammonia adsorbed onto the
Lewis acid sites.35,36 Moreover, the band at 1540 cm−1 is
ascribed to the amide species, which is an important Figure 6. DRIFTS spectra of the NOx desorption of CW (a), CWP
intermediate species in the SCR process and can be observed (b), CM (c), and CMP (d) catalysts at 100−350 °C.
for both types of catalysts.37 Two negative bands are also found
at approximately 3620 and 3665 cm−1, and these may be nitrate exist. Furthermore, the bands at 1218, 1453, and 1597
ascribed to doubly coordinated and bridging O−H stretching cm−1 are assigned to chelating nitro compounds, monodentate
modes, respectively.38,39 On the one hand, after poisoning by nitrite, and bridging nitrate, respectively.11,44,45 After the
As2O3, the peaks corresponding to Lewis acid sites and amide introduction of As2O3, only the bridging nitrate species and
species diminished remarkably and almost vanished as the nitro compounds were observed below 300 °C; the adsorption
temperature continually increased. Considering the NH3-TPD of NOx species was thus obviously suppressed. Above 300 °C,
results, the loss of weakly adsorbed ammonia species (weak the band corresponding to bridging nitrates shifts to a higher
acid sites) may be related to the Lewis acid sites. Additionally, a wavenumber (1618 cm−1), with a weaker signal intensity.
new characteristic peak at 1534 cm−1 was observed above 250 Unlike the CW catalyst, the CM catalyst has a single peak at
°C, which may be associated with the oxidation intermediates 1564 cm−1 for bidentate nitrate and a new band at 1209 cm−1,
of NH3, originating from arsenic oxide species.40 These results which belongs to bidentate nitrite.34 In contrast, the inhibitory
indicate that poisoning by As2O3 destroys the number and effect of As oxide on NOx adsorption appears to be more
stability of Lewis acid sites as well as the reaction mechanism. evident in the CMP than in the CWP catalyst because all the
On the other hand, the bands of the poisoned catalysts ad-NOx nearly vanished above 200 °C. However, previous
attributed to Brønsted acid sites seemed to increase slightly at studies have reported that stably adsorbed nitrate species can
low temperatures (below 100 °C), although these diminished suppress the NH3 activation and weaken the low-temperature
more easily at elevated temperatures. At the same time, the SCR process with the CM catalyst. Therefore, this may be a
negative band, attributed to the bridging −OH stretching reasonable explanation for the higher low-temperature SCR
frequency, was preserved up to 250 °C for the CWP catalyst. In activity of the CMP catalyst.
contrast, the doubly coordinated −OH of the CMP catalyst was 3.6. Investigation of Reaction Mechanism and
susceptible to temperature. Consequently, newly formed As− Deactivation Details. Further studies on changes in the
OH sites from As oxide deposition can supplement a portion of reaction mechanism were carried out using in situ DRIFTS and
the Brønsted acid sites, but these As−OH sites break the in situ Raman spectroscopy. Figure 7 and Figure 8 display the
original surface −OH and Brønsted acid sites. Furthermore, the DRIFTS spectra of two series of catalysts under a reactive
stability of the Brønsted acid sites is decreased after atmosphere at 200 °C. As illustrated in Figure 7a, when the CW
deactivation. catalyst is treated with NH3, coordinated NH3 species (1172
3.5. NO + O2 Adsorption. Because adsorbed NOx can and 1593 cm−1) and NH4+ species (1455 and 1677 cm−1) form
participate in the low-temperature SCR process through the on the CW surface.38 After switching of the gas to NO + O2,
Langmuir−Hinshelwood mechanism, the NOx adsorption both Lewis and Brønsted acid sites rapidly vanish in 3 min and
behavior over the two types of catalysts was investigated are then replaced by new bands, attributed to NOx species
using in situ DRIFTS.41−43 As shown in Figure 6, after the (1242, 1488, 1577, and 1605 cm−1). When the reactants are
adsorption of NO + O2 onto CW, seven bands appeared in the introduced in a reversed order (Figure 7b), the chelating nitro
range of 1100−1700 cm−1, at 1597, 1579, 1546, 1527, 1453, compounds (1236 cm−1), monodentate nitrate (1523 cm−1),
1301, and 1218 cm−1. The bands centered at 1301 and 1527 and bidentate nitrate (1542 and 1579 cm−1) disappear within 1
cm−1 can be ascribed to monodentate nitrate, while those at min after the NH3 is purged. These results indicate that both
1546 and 1579 cm−1 may imply that two types of bidentate Lewis and Brønsted acid sites can participate in the SCR
18009 DOI: 10.1021/acs.jpcc.6b03687
J. Phys. Chem. C 2016, 120, 18005−18014
The Journal of Physical Chemistry C Article

NH3 appears at 1302 cm−1, which may be related to the arsenic

species. When the catalyst is purged with NO + O2, this band
remains almost constant as the Brønsted acid sites are slowly
consumed. Therefore, arsenic oxide not only suppresses the
adsorption of NH3 on both Lewis and Brønsted acid sites but
also provides nonreactive ad-NH3 sites. Furthermore, when the
gaseous order is reversed, only a few weak bands due to
monodentate (1296 cm−1), bidentate (1583 cm−1), and
bridging nitrate (1605 cm−1) can be observed after the
adsorption of NO + O2.34 These bands are soon overlapped
by the adsorbed NH3. It can thus be concluded that the
adsorption of nitrate species is also vastly inhibited due to the
As-poisoning effect, and these species thus have minimal
participation in the low-temperature SCR process.
The DRIFT spectra of the CM and CMP catalysts were also
recorded under the same reaction conditions (Figure 8). For
the fresh sample, NH3 adsorption onto Brønsted and Lewis
acid sites and a few chelating species of nitro compounds were
involved in the low-temperature SCR, which primarily
underwent the Eley−Rideal process between the adsorbed
NH3 and gaseous NOx. However, for the CMP catalyst, the
reaction rate of ad-NH3 on the Brønsted acid sites was slowed.
Additionally, Lewis acid sites and nitro compound species
disappeared, and only weak monodentate and bridging nitrates
Figure 7. Sequential DRIFTS spectra of CW (a, b) and CWP (c, d) could be seen. Considering that the major active sites, Brønsted
recorded under various atmospheres. The dehydrated catalyst is first acid sites, remain, the influence of As on low-temperature
treated by NH3, then NO + O2 is added and the reversed order at 200 activity is less serious than that in the CWP catalyst, which is in
°C. accordance with the activity data shown in Figure 1.
Figure S6 presents the DRIFTS spectra of two series of
catalysts in a flow of NO + O2 after the catalyst was pre-
exposed to a NH3 atmosphere for 30 min at 350 °C; the Eley−
Rideal process is widely regarded as the high-temperature SCR
mechanism for CW and CM catalysts.14,25 When exposed to
NH3, the CW, CWP, and CM catalysts exhibited both Brønsted
and Lewis acid sites, whereas CMP had only Brønsted acid
sites. However, when NO and O2 were introduced, chelating
nitro compounds, monodentate nitrate, bidentate nitrates, and
bridging nitrate (1236, 1481, 1544/1576, and 1598 cm−1)
gradually emerged. Additionally, NH4NO3 (1368 cm−1) and
H2O (1619 cm−1) also appeared as intermediates during the
SCR reaction, with adsorbed NH3 species decreasing on the
CW and CWP catalyst surfaces.41,44 It must be noted that
Lewis acid sites rarely participate in the reaction, and the
reaction rate on Brønsted acid sites slows down with the
emergence of the characteristic peak of NH3 oxidation (1534
cm−1) in the CWP catalyst. Despite the few remaining
Brønsted acid sites on the CMP catalyst, the acid sites are
greatly reduced because of the deposition of arsenic oxide.
Therefore, the smaller amount of acid may play the key role in
the greater loss of activity at high temperatures with the CMP
catalyst.
To further investigate the changes in active sites on the
surface during the reaction mechanism, in situ Raman spectra
Figure 8. Sequential DRIFTS spectra of CM (a, b) and CMP (c, d) were are also collected following the same procedure. For the
recorded under various atmospheres. The dehydrated catalyst is first
CW catalyst (Figure 9a), when the catalyst was purged with
treated by NH3, then NO + O2 is added and the reversed order at 200
°C. NO + O2 after pretreatment with NH3, the υs mode of the W
O vibrations due to WOx (998 cm−1) reemerged, while the
characteristic crystalline CeO2 vibrations (457 and 248 cm−1)
reaction with ad-NOx species through the L−H process at low were obviously enhanced, indicating that NH3 can be
temperatures. However, although Brønsted acid sites (1438 and chemisorbed onto the surface of both WOx and CeOx at low
1671 cm−1) still exist in poisoned catalysts (Figure 7c), Lewis temperatures. Moreover, when the gaseous order was reversed,
acid sites (1172 and 1593 cm−1) decrease significantly. After the band at 930 cm−1, attributed to the WO vibrations of
the catalyst is exposed to NH3, a small new band of adsorbed Ce2(WO4)3 compounds, and the band at 817 cm−1 due to the
18010 DOI: 10.1021/acs.jpcc.6b03687
J. Phys. Chem. C 2016, 120, 18005−18014
The Journal of Physical Chemistry C Article

Figure 9. Sequential Raman spectra of CW (a, b) and CWP (c, d) recorded under various atmospheres. The dehydrated catalyst is first treated by
NH3, then NO + O2 is added and the reversed order at 200 °C.

oxotungstate species were strengthened to some extent after

the NH 3 was introduced (Figure 9b). 25,26 Therefore,
Ce2(WO4)3 compounds and W−O bonds from [WO4] or
[WO6] units may also be active sites during the SCR reaction.
Nevertheless, the changes described above become incon-
spicuous after deactivation by As oxide because the peak for the
WO vibration of WOx weakens. Compared to the results of
the IR spectra, we tentatively assigned the Lewis acid sites to
CeO2 and WO from amorphous WOx. Correspondingly, we
assigned the Brønsted acid sites to the hydroxyl formed on the
W−O or WO modes of both WOx and Ce2(WO4)3. On the
basis of the acidic nature of tungsten along with the basic
nature of ceria, these materials have different adsorption
functions during the SCR reaction. Therefore, after the
introduction of basic As2O3, the acid−base balance of fresh
catalysts can be destroyed, thus affecting the NH3 adsorption
onto WO3.
The in situ sequential Raman spectra of the CM catalyst were
also acquired under the same conditions (see Figure S7), and Figure 10. Sequential Raman spectra of CM (a, b) and CMP (c, d)
similar results were observed, apart from the substitution of recorded under various atmospheres. The dehydrated catalyst is first
treated by NH3, then NO + O2 is added and the reversed order at 200
Ce2Mo3O13 (935 cm−1) for Ce2(WO4)3.46 Additionally, it °C.
should be noted that the band for the MoO stretching
vibration at 962 cm−1 decreases with As deactivation. This
decreasing trend is in accordance with the loss of Lewis acid reaction rate and peak intensity of the CMP catalyst weakened
sites shown before (Figure 10).47 Therefore, the Lewis acid despite a sustained reaction, suggesting a loss of Lewis acid sites
sites of the catalysts may be mainly provided by MoO bonds, originating from MoO that are essential for the high-
although these sites seem to be less important than Brønsted temperature activity of the CM catalyst. These results also
acid sites for the low-temperature SCR process. Figure S7 convincingly reflect the difference between the two catalysts in
shows the in situ sequential Raman spectra of the two types of the high-temperature activity by deactivation. In particular, the
catalysts in a flow of NO + O2 after the catalyst is pre-exposed deactivation effect of the CMP catalyst is more severe than that
to a NH3 atmosphere at 350 °C. As shown in the figure, all the of the CWP catalyst.
CeO2, WOx or MoOx and Ce2(WO4)3 compounds and
Ce2Mo3O13 still participate in the high-temperature SCR 4. DISCUSSION
reaction as NH3 adsorptive sites. After deactivation, the 4.1. Influence of As on Structure and Properties.
intensity of the Raman spectra of the CWP catalyst decreased Unlike our previous research on the As poisoning of the
under the same reaction atmosphere. Additionally, both the CeO2−WO3/TiO2 catalyst, no characteristic XRD peaks
18011 DOI: 10.1021/acs.jpcc.6b03687
J. Phys. Chem. C 2016, 120, 18005−18014
The Journal of Physical Chemistry C Article

belonging to the arsenic oxide species occurred with the CMP to its basic surface. Therefore, the lower NOx adsorption of
and CWP catalysts (Figure 2), suggesting that the arsenic oxide CMP than of CWP at low temperatures due to the structural
species can be highly dispersed on the surface of catalysts.48 change of CeOx mitigates the deactivation of CMP. When the
Considering that the radius of As3+ (0.58 Å) or As5+ (0.46 Å) is reaction temperature is above 300 °C, the NOx adsorptive sites
less than that of Ce4+ (0.87 Å) combined with W6+ (0.62 Å) or become unimportant, and the reaction between ad-NH3 and
Mo6+ (0.64 Å), the replacement of Ce4+ combined with W6+ or gaseous NO and O2 (the E−R process) is the crucial SCR
Mo6+ by As3+ or As5+ to form a solid solution is unexpected. mechanism for both the CW and CM catalysts (Figure S6 and
Moreover, on the basis of the Raman spectra (Figure S3), the Figure S7). Although some Brønsted acid sites provided by As
Ce−O bonds of the CM catalyst are remarkably influenced by species can be found, these sites are inactive during the SCR
As species; therefore, the deactivation effect on CeOx should be reaction. Therefore, the loss of acid sites may play a key role in
the major reason for the loss of activity. In contrast, for the the high-temperature deactivation. Furthermore, for the CMP
CWP catalyst, the WO bond is shortened, with the catalyst, the reduced quantity of acid and Lewis acid sites,
appearance of a more distorted octahedral geometry and originating from the damage of the surface Ce species and
augmented isopolytungstate groups, indicating that WO3 is MoO bonds, may play the major role in the greater loss of
preferentially affected by As2O3. In other words, WO3 can be activity of this catalyst than CWP.
sacrificed to partly prevent the deactivation of CeO2. Therefore,
a reasonable explanation for the structural changes is that a 5. CONCLUSIONS
portion of the arsenic oxide can be primarily bonded as The mechanism of poisoning of CM and CW catalysts by
amorphous structures to the surface CeO2 (CM) and WO3 arsenic oxide was investigated, compared, and analyzed in this
(CW) via Ce−O−As and W−O−As linkages, respectively. work. It was found that for the CMP catalysts, the degree of
Additionally, because the electronegativity of As (2.18) is activity loss increases with temperature, while the tendency is
stronger than that of Ce (1.12), a portion of the surface Ce4+ reversed for the CWP catalysts. By exploring the changes in the
species of CMP can be reduced to the Ce3+ species after the physical structure, redox properties, acidity, surface-adsorbed
introduction of As due to the electron-withdrawing nature of species and reaction mechanism before and after deactivation,
As. These results, combined with the XPS results (Figure 4 and we determined that the reasons for these results are as follows:
Table 2), which indicate the formation of less excess surface (1) As2O3 has a major impact on CeOx by oxidizing a
Ce3+ on CMP, suggest that a redox process between Ce4+ and portion of the Ce3+ into Ce4+ species for the CMP catalyst.
As3+ may occur and inhibit the self-redox process of the Ce However, As2O3 prefers to interact with WO3 through W−O−
species, hindering the SCR reaction. In contrast, W(IV) hinders As linkages in the CWP catalyst. (2) As2O3 decreases the
the reduction of Ce4+ by As3+, and the strong interaction strength, number, and stability of acid sites for the two types of
between W and As may be the dominant reason for the catalysts. Moreover, CMP exhibits a more serious loss of Lewis
deactivation of CWP. Previous research on CW and CM acid sites with an increase in temperature compared with CWP.
catalysts had indicated that surface Ce atoms and unsaturated (3) According to the results of in situ DRIFTS and Raman
Wn+ cations are the main contributors of Lewis acid sites, while spectra, more Brønsted acid sites and reduced NOx adsorption
WOx (with an appropriate doping amount) or polymeric MoOx mitigate the deactivation levels in CMP catalysts compared with
units promoted the generation of Brønsted acid sites.14,25 CWP at low temperatures. At high temperatures, the greater
Combined with the structure effects discussed above, the loss of quantity of remnant Lewis acid sites in CWP originates from
Lewis acid sites derived from a decreased number of surface Ce the protection of the surface Ce species by WO3 and may play
atoms and distorted WOx species may be the key reason for the the key role in the reduced loss of activity of this catalyst
deactivation of the CMP and CWP catalysts, respectively. Thus, compared with CMP.

■
it is concluded that the surface structures of CWP and CMP are
influenced by As species in two different ways. ASSOCIATED CONTENT
4.2. Influence of As on the Reaction Mechanism. For * Supporting Information
S
the CM catalyst, when the reaction temperature is below 250
The Supporting Information is available free of charge on the
°C, previous research and our DRIFTS results (Figure 8) reveal
ACS Publications website at DOI: 10.1021/acs.jpcc.6b03687.
that NH3 adsorption is vital at low temperatures (via the E−R
process).14 Therefore, excessive NOx adsorption has an adverse Figures of activities, N2 adsorption curves, Raman
effect on low-temperature activity because of competitive spectra, XPS spectra, and additional DRIFTS spectra
adsorption. However, when As is introduced to the CM (PDF)

■
catalyst, the NOx adsorption of the poisoned sample becomes
weaker (Figure 6). Some of the ad-NH3 may be released to AUTHOR INFORMATION
retard the loss of activity due to deactivation. In contrast, for Corresponding Author
the CW catalyst, NOx adsorption can also participate in the *E-mail: [email protected]. Phone: +86 10 62771093.
low-temperature SCR process (via the L−H process).49 The Fax: +86 10 62771093.
loss of ad-NOx sites has a great impact on the deactivation of
the CW catalyst. In addition, on the basis of the structural Notes
The authors declare no competing financial interest.

■
analysis of the CMP catalyst, the deposition of arsenic atoms
leads to the loss of surface Ce4+ and thus suppresses the
conversion cycle of Ce4+ to Ce3+. Nevertheless, the existing ACKNOWLEDGMENTS
W(IV) atoms on the CW surface hinder the reduction of Ce4+ This work was financially supported by National Natural
by As3+ because of the formation of stronger interactions Science Foundation of China (Grants 21325731, 21407088,
between W and As. However, the NOx adsorption sites for both and 51478241), National High-Tech Research and Develop-
types of catalysts are mainly provided by the CeO2 species due ment (863) Program of China (Grants 2013AA065401 and
18012 DOI: 10.1021/acs.jpcc.6b03687
J. Phys. Chem. C 2016, 120, 18005−18014
The Journal of Physical Chemistry C Article

2013AA065304), and China Postdoctoral Science Foundation (18) Senior, C. L.; Lignell, D. O.; Sarofim, A. F.; Mehta, A. Modeling
(Grant 2013M530643). Arsenic Partitioning in Coal-Fired Power Plants. Combust. Flame 2006,

■
147, 209−221.
(19) Hums, E. A Catalytically Highly-Active, Arsenic Oxide Resistant
REFERENCES V-Mo-O Phase  Results of Studying Intermediates of the
(1) Liu, J.; Li, X.; Zhao, Q.; Hao, C.; Wang, S.; Tadé, M. Combined Deactivation Process of V2O5-MoO3-TiO2 (Anatase) DeNOx Cata-
Spectroscopic and Theoretical Approach to Sulfur-Poisoning on Cu- lysts. Res. Chem. Intermed. 1993, 19, 419−441.
Supported Ti−Zr Mixed Oxide Catalyst in the Selective Catalytic (20) Hilbrig, F.; Göbel, H. E.; Knözinger, H.; Schmelz, H.; Lengeler,
Reduction of NOx. ACS Catal. 2014, 4, 2426−2436. B. Interaction of Arsenious Oxide with DeNOx-Catalysts: An X-Ray
(2) Meng, D.; Zhan, W.; Guo, Y.; Guo, Y.; Wang, L.; Lu, G. A Highly Absorption and Diffuse Reflectance Infrared Spectroscopy Study. J.
Effective Catalyst of Sm-MnOx for the NH3-SCR of NOx at Low Catal. 1991, 129, 168−176.
Temperature: Promotional Role of Sm and Its Catalytic Performance. (21) Li, X.; Li, J.; Peng, Y.; Si, W.; He, X.; Hao, J. Regeneration of
ACS Catal. 2015, 5, 5973−5983. Commercial SCR Catalysts: Probing the Existing Forms of Arsenic
(3) Habib, H. A.; Basner, R.; Brandenburg, R.; Armbruster, U.; Oxide. Environ. Sci. Technol. 2015, 49, 9971−9978.
Martin, A. Selective Catalytic Reduction of NOx of Ship Diesel Engine (22) Peng, Y.; Si, W.; Li, X.; Luo, J.; Li, J.; Crittenden, J.; Hao, J.
Exhaust Gas with C3H6 over Cu/Y Zeolite. ACS Catal. 2014, 4, 2479− Comparison of MoO3 and WO3 on Arsenic Poisoning V2O5/TiO2
2491. Catalyst: DRIFTS and DFT Study. Appl. Catal., B 2016, 181, 692−
(4) Burch, R.; Scire, S. Selective Catalytic Reduction of Nitric-Oxide 698.
with Ethane and Methane on Some Metal Exchanged ZSM-5 Zeolites. (23) Chang, H.; Jong, M. T.; Wang, C.; Qu, R.; Du, Y.; Li, J.; Hao, J.
Appl. Catal., B 1994, 3, 295−318. Design Strategies for P-Containing Fuels Adaptable CeO2-MoO3
(5) Busca, G.; Lietti, L.; Ramis, G.; Berti, F. Chemical and Catalysts for DeNOX: Significance of Phosphorus Resistance and
Mechanistic Aspects of the Selective Catalytic Reduction of NOx by N2- Selectivity. Environ. Sci. Technol. 2013, 47, 11692−11699.
Ammonia over Oxide Catalysts: A Review. Appl. Catal., B 1998, 18, 1− (24) Sing, K. S. W. Reporting Physisorption Data for Gas/Solid
36. Systems with Special Reference to the Determination of Surface Area
(6) Cristiani, C.; Bellotto, M.; Forzatti, P.; Bregani, F. On the and Porosity. Pure Appl. Chem. 1985, 57, 603−619.
Morphological Properties of Tungsta-Titania De-NOxing Catalysts. J. (25) Peng, Y.; Li, K.; Li, J. Identification of the Active Sites on
Mater. Res. 1993, 8, 2019−2025. CeO2−WO3 Catalysts for SCR of NOx with NH3: An in Situ Ir and
(7) Martens, J. A.; Cauvel, A.; Francis, A.; Hermans, C.; Jayat, F.; Raman Spectroscopy Study. Appl. Catal., B 2013, 140−141, 483−492.
(26) Mamede, A. S.; Payen, E.; Granger, P.; Florea, M.; Pârvulescu,
Remy, M.; Keung, M.; Lievens, J.; Jacobs, P. A. Nox Abatement in
V. I. WOx-CeO2 and WOx-Nb2O5 Catalysts Deactivation During
Exhaust from Lean-Burn Combustion Engines by Reduction of NO2
Hexane Isomerization. AIChE J. 2008, 54, 1303−1312.
over Silver-Containing Zeolite Catalysts. Angew. Chem., Int. Ed. 1998,
(27) Picquart, M.; Castro-Garcia, S.; Livage, J.; Julien, C.; Haro-
37, 1901−1903.
Poniatowski, E. Structural Studies During Gelation of WO 3
(8) Topsoe, N. Y.; Topsoe, H.; Dumesic, J. A. Vanadia-Titania
Investigated by in-Situ Raman Spectroscopy. J. Sol-Gel Sci. Technol.
Catalysts for Selective Catalytic Reduction (SCR) of Nitric-Oxide by
2000, 18, 199−206.
Ammonia. 1. Combined Temperature-Programmed in-Situ Ftir and (28) Ross-Medgaarden, E. I.; Wachs, I. E. Structural Determination
Online Mass-Spectroscopy Studies. J. Catal. 1995, 151, 226−240. of Bulk and Surface Tungsten Oxides with Uv-Vis Diffuse Reflectance
(9) Qi, G. S.; Yang, R. T.; Chang, R. MnOx-CeO2 Mixed Oxides Spectroscopy and Raman Spectroscopy. J. Phys. Chem. C 2007, 111,
Prepared by Co-Precipitation for Selective Catalytic Reduction of NO 15089−15099.
with NH3 at Low Temperatures. Appl. Catal., B 2004, 51, 93−106. (29) Boningari, T.; Ettireddy, P. R.; Somogyvari, A.; Liu, Y.;
(10) Li, F.; Zhang, Y. B.; Xiao, D. H.; Wang, D. Q.; Pan, X. Q.; Yang, Vorontsov, A.; McDonald, C. A.; Smirniotis, P. G. Influence of
X. G. Hydrothermal Method Prepared Ce-P-O Catalyst for the Elevated Surface Texture Hydrated Titania on Ce-Doped Mn/TiO2
Selective Catalytic Reduction of NO with NH3 in a Broad Catalysts for the Low-Temperature SCR of NOx under Oxygen-Rich
Temperature Range. ChemCatChem 2010, 2, 1416−1419. Conditions. J. Catal. 2015, 325, 145−155.
(11) Li, P.; Xin, Y.; Li, Q.; Wang, Z.; Zhang, Z.; Zheng, L. Ce−Ti (30) Reiche, M. A.; Maciejewski, M.; Baiker, A. Characterization by
Amorphous Oxides for Selective Catalytic Reduction of NO with NH3: Temperature Programmed Reduction. Catal. Today 2000, 56, 347−
Confirmation of Ce−O−Ti Active Sites. Environ. Sci. Technol. 2012, 355.
46, 9600−9605. (31) Zhang, S.; Zhong, Q.; Shen, Y.; Zhu, L.; Ding, J. New Insight
(12) Chen, L.; Li, J.; Ablikim, W.; Wang, J.; Chang, H.; Ma, L.; Xu, J.; into the Promoting Role of Process on the CeO2−WO3/TiO2 Catalyst
Ge, M.; Arandiyan, H. CeO2-WO3 Mixed Oxides for the Selective for NO Reduction with NH3 at Low-Temperature. J. Colloid Interface
Catalytic Reduction of NOx by NH3 over a Wide Temperature Range. Sci. 2015, 448, 417−426.
Catal. Lett. 2011, 141, 1859−1864. (32) Baek, Y.; Yong, K. Controlled Growth and Characterization of
(13) Shan, W.; Liu, F.; He, H.; Shi, X.; Zhang, C. Novel Cerium- Tungsten Oxide Nanowires Using Thermal Evaporation of WO3
Tungsten Mixed Oxide Catalyst for the Selective Catalytic Reduction Powder. J. Phys. Chem. C 2007, 111, 1213−1218.
of NOx with NH3. Chem. Commun. 2011, 47, 8046−8048. (33) Lange, F.; Schmelz, H.; Knözinger, H. An X-Ray Photoelectron
(14) Peng, Y.; Qu, R.; Zhang, X.; Li, J. The Relationship between Spectroscopy Study of Oxides of Arsenic Supported on TiO2. J.
Structure and Activity of MoO3−CeO2 Catalysts for NO Removal: Electron Spectrosc. Relat. Phenom. 1991, 57, 307−315.
Influences of Acidity and Reducibility. Chem. Commun. 2013, 49, (34) Hadjiivanov, K. I. Identification of Neutral and Charged NXOY
6215−6217. Surface Species by Ir Spectroscopy. Catal. Rev.: Sci. Eng. 2000, 42, 71−
(15) Qu, R.; Gao, X.; Cen, K.; Li, J. Relationship between Structure 144.
and Performance of a Novel Cerium-Niobium Binary Oxide Catalyst (35) Centeno, M. A.; Carrizosa, I.; Odriozola, J. A. NO−NH3
for Selective Catalytic Reduction of NO with NH3. Appl. Catal., B Coadsorption on Vanadia/Titania Catalysts: Determination of the
2013, 142, 290−297. Reduction Degree of Vanadium. Appl. Catal., B 2001, 29, 307−314.
(16) Liu, Z.; Yi, Y.; Li, J.; Woo, S. I.; Wang, B.; Cao, X.; Li, Z. A (36) Mamede, A. S.; Payen, E.; Grange, P.; Poncelet, G.; Ion, A.;
Superior Catalyst with Dual Redox Cycles for the Selective Reduction Alifanti, M.; Prvulescu, V. I. Characterization of WOx/CeO2 Catalysts
of NOx by Ammonia. Chem. Commun. 2013, 49, 7726−7728. and Their Reactivity in the Isomerization of Hexane. J. Catal. 2004,
(17) Liu, Z.; Zhang, S.; Li, J.; Ma, L. Promoting Effect of MoO3 on 223, 1−12.
the NOx Reduction by NH3 over CeO2/TiO2 Catalyst Studied with in (37) Wu, Z.; Jiang, B.; Liu, Y.; Wang, H.; Jin, R. Drift Study of
Situ Drifts. Appl. Catal., B 2014, 144, 90−95. Manganese/Titania-Based Catalysts for Low-Temperature Selective

18013 DOI: 10.1021/acs.jpcc.6b03687

J. Phys. Chem. C 2016, 120, 18005−18014
The Journal of Physical Chemistry C Article

Catalytic Reduction of NO with NH3. Environ. Sci. Technol. 2007, 41,

5812−5817.
(38) Sellick, D. R.; Aranda, A.; García, T.; López, J. M.; Solsona, B.;
Mastral, A. M.; Morgan, D. J.; Carley, A. F.; Taylor, S. H. Influence of
the Preparation Method on the Activity of Ceria Zirconia Mixed
Oxides for Naphthalene Total Oxidation. Appl. Catal., B 2013, 132−
133, 98−106.
(39) Ramis, G.; Busca, G.; Cristiani, C.; Lietti, L.; Forzatti, P.;
Bregani, F. Characterization of Tungsta-Titania Catalysts. Langmuir
1992, 8, 1744−1749.
(40) Peng, Y.; Li, J.; Si, W.; Luo, J.; Wang, Y.; Fu, J.; Li, X.;
Crittenden, J.; Hao, J. Deactivation and Regeneration of a Commercial
SCR Catalyst: Comparison with Alkali Metals and Arsenic. Appl.
Catal., B 2015, 168−169, 195−202.
(41) Yang, S.; Li, J.; Wang, C.; Chen, J.; Ma, L.; Chang, H.; Chen, L.;
Peng, Y.; Yan, N. Fe-Ti Spinel for the Selective Catalytic Reduction of
NO with NH3: Mechanism and Structure-Activity Relationship. Appl.
Catal., B 2012, 117, 73−80.
(42) Yang, S.; Xiong, S.; Liao, Y.; Xiao, X.; Qi, F.; Peng, Y.; Fu, Y.;
Shan, W.; Li, J. Mechanism of N2O Formation During the Low-
Temperature Selective Catalytic Reduction of NO with NH3 over Mn-
Fe Spinel. Environ. Sci. Technol. 2014, 48, 10354−10362.
(43) Li, X.; Li, J.; Peng, Y.; Zhang, T.; Liu, S.; Hao, J. Selective
Catalytic Reduction of NO with NH3 over Novel Iron-Tungsten
Mixed Oxide Catalyst in a Broad Temperature Range. Catal. Sci.
Technol. 2015, 5, 4556−4564.
(44) Chen, L.; Li, J.; Ge, M. Drift Study on Cerium−Tungsten/
Titiania Catalyst for Selective Catalytic Reduction of NOx with NH3.
Environ. Sci. Technol. 2010, 44, 9590−9596.
(45) Ma, Z.; Wu, X.; Si, Z.; Weng, D.; Ma, J.; Xu, T. Impacts of
Niobia Loading on Active Sites and Surface Acidity in NbOx/CeO2−
ZrO2 NH3−SCR Catalysts. Appl. Catal., B 2015, 179, 380−394.
(46) Wu, Y.; Hu, G.; Xie, Y.; Guo, M.; Luo, M. Solid State Reaction
of MoO3−CeO2 Complex Oxide Studied by Raman Spectroscopy.
Solid State Sci. 2011, 13, 2096−2099.
(47) Ohler, N.; Bell, A. T. A Study of the Redox Properties of
MoOX/SiO2. J. Phys. Chem. B 2005, 109, 23419−23429.
(48) Li, X.; Li, J.; Peng, Y.; Chang, H.; Zhang, T.; Zhao, S.; Si, W.;
Hao, J. Mechanism of Arsenic Poisoning on SCR Catalyst of CeW/Ti
and Its Novel Efficient Regeneration Method with Hydrogen. Appl.
Catal., B 2016, 184, 246−257.
(49) Chen, L.; Li, J.; Ge, M.; Ma, L.; Chang, H. Mechanism of
Selective Catalytic Reduction of NOx with NH3 over CeO2-WO3
Catalysts. Chin. J. Catal. 2011, 32, 836−841.

18014 DOI: 10.1021/acs.jpcc.6b03687

J. Phys. Chem. C 2016, 120, 18005−18014