A novel mechanism for poisoning of metal oxide SCR catalysts: Base-acid explanation correlated with redox properties

Article abstract of DOI:10.1039/c4cc02991g

A novel mechanism is proposed for the poisoning effect of acid gases and N₂O formation on SCR catalysts involving base-acid properties correlated with redox ability of M-O or M-OH (M = Ce or V) of metal oxides and the strength of their basicity responsible for resistance to HCl and SO ₂ at medium and low temperatures. This journal is the Partner Organisations 2014.

Full text of DOI:10.1039/c4cc02991g

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
A novel mechanism for poisoning of metal oxide
SCR catalysts: base–acid explanation correlated
with redox properties†
Cite this: Chem. Commun., 2014,
50, 10031
Received 22nd April 2014,
Accepted 9th June 2014
Huazhen Chang,a Junhua Li,*ab Wenkang Su,a Yuankai Shaoa and Jiming Haoab
DOI: 10.1039/c4cc02991g
www.rsc.org/chemcomm
A novel mechanism is proposed for the poisoning effect of acid extent as compared with the acidity. It was reported that Ti4+
gases and N2O formation on SCR catalysts involving base–acid ions exposed on (001) and (010) faces of TiO2 P25 particles are L
properties correlated with redox ability of M–O or M–OH (M = Ce acid centers and a fraction of O2À ions on the surface of TiO2 P25
or V) of metal oxides and the strength of their basicity responsible behave as basic centers, actually, the behavior of Ti4+–O2À acid–
for resistance to HCl and SO2 at medium and low temperatures.
base pairs depends on the surface structure.8 For low-
temperature SCR reaction, the active sites for NO adsorption
Important features in the development of SCR catalysts for and oxidation act indeed as basic sites. Except for non-oxide
stationary and automotive applications are the enhancement of catalysts, the basic sites are believed to be surface O atoms.9
resistance to toxicants and an increase of N2 selectivity at Further consideration of the phenomena of acid and base
medium and low temperatures. It is commonly recognized that catalysis has contributed to the understanding of the mecha-
acid and redox functions coexist and they are both important nism of chemical reactions in general. Another important reason
on SCR catalysts.1 The acidity of catalysts is enhanced after is that the basic sites are easily poisoned by acidic molecules
poisoning by SO2, but the activity decreased at low tempera- such as HCl and SO2; characterization and understanding of the
tures.2 The research on the correlation between basicity of basic sites would be helpful to improve the resistance to
catalysts and SCR performance is scarce, especially for deacti- acid gases.
vation by acid gases and N2O formation.
Carbon dioxide is believed to adsorb on metals and basic
The most widely used catalyst system for SCR of NOx is the oxides in many forms, such as the monodentate, bidentate, and
V-based catalyst supported on TiO2.1 It was suggested that the bridged forms.9 The strength and the amount of basic sites are
redox properties are the key factors for the high reactivity of the reflected in desorption temperature and the peak area, respec-
V-based catalyst at low temperatures.3 In recent decades, many tively, in a CO2–TPD plot. In this study, the basicity of CeMo/Ti
neoteric catalysts were developed in medium and low tempera- and VW/Ti catalysts was characterized by DRIFT spectra for CO2
ture ranges, such as FeTiOx, MnOx–CeO2, Ln–Na–Cu–O and Co/ adsorption and CO2–TPD. Based on surface basicity, the
MFI.4 In the development of novel SCR catalysts, Ce-based poisoning mechanism of two catalysts by HCl was revealed.
oxides with high oxygen storage capacity and excellent redox The correlation between base–acid and redox properties was
properties attracted much attention.5 The enhancement of proposed in which basic sites could convert to acid sites for
acidity upon Ce addition was recognized by previous study.6
SCR reaction and the redox properties were inhibited after
Acids and bases are related to each other. Acid and base active poisoning by acid gases. The latter contributes mainly to the
surface sites, in particular, can be relevant in determining the deactivation effect of acid gases due to the existence of strong
interaction mechanisms between the surface and the reagents, basic sites on the catalyst surface. In addition, the effect of
or between the surface and the reaction intermediates.7 basic sites on the formation of N2O was also proposed.
Whereas, the basicity of SCR catalysts was studied to a lesser
The catalysts were prepared by an impregnation method,
which is similar to that reported in the literature.5 The detailed
preparation procedures are described in ESI.† The CeO2–MoO3/
TiO2 (the molar ratio of Ce : Mo = 2 : 1, loading 10 wt% totally)
catalyst is denoted as CeMo/Ti. The V2O5–WO3/TiO2 catalyst
with 1 wt% V2O5 and 5 wt% WO3 is denoted as VW/Ti. The V
and W loading has been selected since it roughly corresponds
to that of commercial catalysts. The HCl poisoning experiment
a State Key Joint Laboratory of Environment Simulation and Pollution Control
(SKLESPC), School of Environment, Tsinghua University, Beijing, 100084, China.
E-mail: lijunhua@tsinghua.edu.cn
b State Environmental Protection Key Laboratory of Sources and Control of Air
Pollution Complex, Beijing 100084, China
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4cc02991g
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 10031--10034 | 10031

View Article Online
Communication
ChemComm
Fig. 2 (a) The effect of HCl on the NO reaction rate in the SCR reaction, and
(b) the effect of HCl and NH4Cl-poisoning on the NO reaction rate with GHSV =
70 000 hÀ1 (STP) over CeMo/Ti and VW/Ti catalysts. Reaction conditions:
T = 200 1C, 500 ppm NO, 500 ppm NH3, 5% O2, 9% H2O, 200 ppm HCl
(when used). (c) CO2–TPD curves of two catalysts after HCl poisoning.
Fig. 1 (a) DRIFT spectra for CO2 adsorption on CeMo/Ti and VW/Ti
catalysts and the number of –OH basic sites vs. temperature on two
catalysts (inset); (b) CO2–TPD curves of CeMo/Ti and VW/Ti catalysts.
was carried out by employing 250 ppm of HCl and 10% of H2O
in inlet gas to poison the catalysts on stream at 200 1C. The spectrum of CeMo/Ti consists of four peaks attributed to strong and
NH4Cl-poisoned catalysts (NH4Cl content of 4.0 wt%) were very strong basic sites. The calculated area of the latter peaks is much
prepared by impregnating samples in aqueous solution of bigger indicating that very strong basic sites are dominant on the
NH4Cl. The SCR activity test was estimated in a fixed-bed quartz CeMo/Ti catalyst. Meanwhile, much more strong basic sites exist on
reactor with the following reaction conditions: 500 ppm NO, the VW/Ti catalyst and the total number of basic sites (calculated
500 ppm NH3, 5% O2, 9% H2O (when used), 200 ppm HCl from CO2–TPD) is much higher than that on the CeMo/Ti catalyst.
(when used), and balance in N2. When investigating the effect Combining with the results obtained from the DRIFT study, we
of HCl addition on SCR activity, the sample was kept at a steady proposed that the basic O2À adjacent to metal ions, which showed
state for 12 h. The concentrations of NO, NO2 (NOx = NO + NO2) much strong interaction with CO2, can be attributed to very strong
and other components were measured using a FT-IR gas basic sites; meanwhile, the –OH basic sites, which are abundant on
analyzer (Gasmet Dx-4000).
The basic centers of CeMo/Ti and VW/Ti catalysts were char- CO2, can be considered as strong basic sites in CO2–TPD. The
acterized by DRIFT spectra for CO2 adsorption (see Fig. 1(a)). existence of basic sites is supposed to facilitate the adsorption of
The spectra of both catalysts show complex patterns mainly acid gases. The different features of basic sites on two catalysts ought
the VW/Ti catalyst and showed relatively weaker interaction with
due to monodentate (bands at 1580 and ca. 1360 cmÀ1) and to represent different behavior in acid gas poisoning. The basic O2À
,
bidentate (bands at 1670, 1537 and 1246 cmÀ1) carbonate i.e. very strong basic site, is more likely being attacked by acid gases
groups, produced by nucleophilic attack of basic O2À ions (i.e., HCl and SO2).
exposed to CO2 molecules. In addition, components at 1630,
Fig. 2(a) shows the effect of HCl on the NO reaction rate of
1434 and 1230 cmÀ1 are assigned to bicarbonate species, which CeMo/Ti and VW/Ti catalysts at 200 1C. The reaction rate over
should be produced by reaction of CO2 with some –OH basic CeMo/Ti was much higher than that on VW/Ti in the presence of
groups left on sites in the defect position.10
H2O. With the addition of HCl, the NO reaction rate decreased over
It is generally considered that the appearance of the character- both catalysts. Whereas, the NO reaction rate over CeMo/Ti was still
istic band near 1230 cmÀ1 [d(OH)] confirmed the formation of higher than that on VW/Ti. The CeMo/Ti and VW/Ti samples after
HO–CO2À (i.e. bicarbonate) species. The intensity of the 1230 cmÀ1 the SCR activity test (with H2O and HCl) were characterized by
band hereby has been taken as a measure of the number of –OH ICP-AES. The molar ratio of ClÀ/NH4+ was 1.5 and 0.056 for CeMo/
basic species.11 In this study, the number of –OH basic groups on Ti and VW/Ti, respectively. It indicated that more chloride formed
CeMo/Ti and VW/Ti were calculated from the peak areas of the on CeMo/Ti besides NH4Cl; whereas, formation of large amount of
characteristic band (shifted to 1240 cmÀ1 on VW/Ti) and they are NH4+ on VW/Ti demonstrated that more Brønsted (denotes as B)
shown in Fig. 1(a). It revealed that more –OH basic sites exist on the acid sites formed over VW/Ti catalysts in the presence of HCl. The
VW/Ti catalyst and they can still be detected at 400 1C. Meanwhile, literature focused on the poisoning effect of acid gases proposed
these groups decreased rapidly as the temperature increased and that formation of ammonium or metal salts (i.e., chloride or
disappeared at 300 1C on CeMo/Ti, indicating that the interaction sulfate) is the main reason for deactivation of SCR catalysts.13
between –OH basic sites and CO2 was unstable on this catalyst. But
HCl and NH4Cl poisoning was performed over two catalysts
the carbonate species were stable at high temperatures on two and the SCR performance was tested afterward (see Fig. 2(b)).
catalysts, which means that the interaction between basic O2À sites After HCl poisoning, the NO reaction rate decreased on the
and CO2 was much strong.
CeMo/Ti catalyst, indicating that HCl poisoning leads to the
Fig. 1(b) shows the CO2–TPD curves of CeMo/Ti and VW/Ti deactivation effect for this catalyst. It is noteworthy that the NO
catalysts. The weak, medium, strong and very strong basic sites on reaction rate increased to a certain extent on the VW/Ti catalyst,
catalysts are characterized from the peaks in the temperature which might be due to the formation of new acid sites on its
ranges of 100–250, 250–400, 400–650, and 4650 1C.12 The CO2–TPD surface. Poisoning by NH4Cl restrains the reaction rate severely
10032 | Chem. Commun., 2014, 50, 10031--10034
This journal is ©The Royal Society of Chemistry 2014

View Article Online
ChemComm
Communication
over both catalysts; whereas, the reaction rate was higher on the and the latter showed even higher SCR activity after poisoning
poisoned CeMo/Ti than the poisoned VW/Ti catalyst. The by HCl. Cl prefers to interact directly with surface oxygen
deposition of NH4Cl on the catalysts blocks the active sites atoms, instead of V atoms which could be available through
apparently at low temperatures.13 The different behavior in HCl the presence of oxygen vacancies, thus, much less chloride
poisoning can be contributed to the difference in the relative could be detected on the VW/Ti catalyst.16 This can also explain
ratio of very strong and strong basic sites on two catalysts. HCl why much more NH4+ is detected on the VW/Ti catalyst after the
prefers to react with very strong basic sites immediately, which activity test with HCl. On the CeMo/Ti catalyst, new acid–base
was the main reason for deactivation of the CeMo/Ti catalyst. sites (–OH) formed when HCl adsorbed on basic O2À. These
The interaction between HCl and the VW/Ti catalyst was weaker –OH sites facilitate the adsorption of NH3, which was evidenced
due to the weaker intensity of basic sites. As shown in Fig. 2(c), after by DRIFT study.
poisoning by HCl, the intensity of peaks of very strong basic sites
In past decades, great efforts were also made to reveal the
decreased on CeMo/Ti, meanwhile, the intensity of peaks of strong poisoning effect of SO2 on SCR catalysts.17 The effect of SO2 on
basic sites increased apparently. It indicated that the number of CeMo/Ti was also tested in this work and the SCR activity decreased
very strong basic sites (O2À) decreased and strong basic sites (–OH) after poisoning, especially at low temperatures (see Fig. S2, ESI†).
increased. It was possible that HCl adsorbs on basic O2À sites and The peak attributed to very strong basic sites shifted at lower
forms new –OH sites. Oxygen (maybe neighboring Ce) was so basic temperature in CO2–TPD (see Fig. S2, ESI†), indicating that the
that HCl prefers to interact with it, thus, much less NH4Cl could basicity decreased. The –OH basic sites increased apparently at the
form. The intermediate, Ce–(OH)2Cl2, even though could not be same time after SO2 poisoning, which will provide new B acid sites.
detected, was much likely to form. The increase of –OH basic sites The decrease in basicity and raise in the number of –OH sites by
after HCl poisoning could also be detected on the VW/Ti catalyst SO2 poisoning are similar to those by HCl poisoning. The decrease
(see Fig. 2(c)).
in SCR activity at low temperatures was mainly related to redox
On the other side, the –OH basic sites could also act as B properties.
acid sites for NH3 adsorption. In order to investigate the acidity
H2-TPR was applied to distinguish the redox properties of
and the effect of HCl on NH3 adsorption, DRIFT study of the the CeMo/Ti catalyst after HCl poisoning (see Fig. S3, ESI†). A
effect of HCl on NH3 adsorption was carried out on the CeMo/Ti peak at 547 1C disappeared for the poisoned sample, indicating
catalyst at 200 1C (see Fig. 3(a)). In the presence of HCl, the that the reducibility was inhibited after HCl poisoning. Fig. 3(c)
intensity of the bands assigned to B acid sites (at 1652, 1443, presents NO2 production in NO oxidation before and after HCl
and 1413 cmÀ1) and L acid sites (at 1600 and 1202 cmÀ1
)
poisoning on the CeMo/Ti catalyst. At low temperature, metal
increased noticeably compared with that in the absence of ions are considered as the active sites for NO oxidation, the
HCl.5,14 These results suggested that the addition of HCl oxidation of NO to NO2 is an important step in the SCR
obviously increases the number of both B and L acid sites on reaction.18 It showed that NO oxidation was suppressed
the CeMo/Ti catalyst. Meanwhile, the broad bands appearing severely over the CeMo/Ti catalyst after HCl poisoning. The
+
due to the stretching vibration of NH4 species are detected formation of cerium chlorides resulted in the disruption of the
over the VW/Ti catalyst,15 revealing that much more B acid sites redox cycle of active sites (Ce3+ 2 Ce4+ + e). Both the increase
appeared on this catalyst (see Fig. S1, ESI†). Contributions of B in acidity and the decrease in redox ability affect the SCR
acid sites on two catalysts at different temperatures were performance of CeMo/Ti catalysts. It is noteworthy that HCl
calculated by DRIFT spectra and the results are shown in poisoning would inhibit the over oxidation of NH3, the for-
Fig. 3(b). It showed that the number of B acid sites on the mation of N2O was suppressed and the N2 selectivity was
VW/Ti catalyst is larger than that on the CeMo/Ti catalyst. The improved subsequently (see Fig. 3(c) and Fig. S4, ESI†). The
existence of abundant B acid sites promoted the SCR reaction modification of basic sites and repression of redox properties
which benefit the N2 selectivity were also evidenced by the
MnOx–CeO2 catalyst (see Fig. S5, ESI†).
Based on the present work and previous study, the correlation of
base–acid and redox properties of M–O or M–OH (M = Ce or V) can
be incorporated into a cross and four quadrants for the SCR
reaction over metal oxide catalysts (Scheme 1). According to this
scheme, there are essentially two related functions on the catalyst,
that is, base–acid and redox functions. Acidity is indispensable for
NH3 adsorption in the SCR reaction. With moderate redox proper-
ties, the SCR reaction could take place which is initiated by NH3
adsorption on a B acid site at high temperatures (1st quadrant). As
the redox ability increases, NO could be oxidized to NO2 or nitrate
Fig. 3 (a) The DRIFT spectra for the effect of HCl on NH3 adsorption over
the CeMo/Ti catalyst at 200 1C; (b) the number of surface acid sites vs.
temperature on CeMo/Ti and VW/Ti catalysts; and (c) the production of
species, and then reacts with adsorbed NH3 or NH4+ at low
temperatures (2nd quadrant). For the latter, the reaction involving
the formation of ammonium nitrate as an intermediate has been
proposed as the Fast SCR reaction.18 With the acid sites, the basic
NO2 vs. temperature in NO oxidation and N2O formation vs. temperature
in the SCR reaction on CeMo/Ti catalysts.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 10031--10034 | 10033

View Article Online
Communication
ChemComm
This work was financially supported by the National Natural
Science Foundation of China (Grant No. 21307071), the National
Basic Research (973) Program of China (Grant No. 2013CB430005),
and the National High-Tech Research and Development (863)
Program of China (Grant No. 2013AA065401, 2012AA062506).
Notes and references
1 N. Y. Topsøe, Science, 1994, 265, 1217.
2 F. D. Liu, K. Asakura, H. He, W. P. Shan, X. Y. Shi and C. B. Zhang,
Appl. Catal., B, 2011, 103, 369.
3 L. J. Alemany, L. Lietti, N. Ferlazzo, P. Forzatti, G. Busca, E. Giamello
and F. Bregani, J. Catal., 1995, 155, 117.
4 (a) G. Qi and R. T. Yang, Chem. Commun., 2003, 848; (b) X. Y. Chen
and S. Kawi, Chem. Commun., 2001, 1354; (c) J. Liu, Z. Zhao, C. M. Xu
and A. J. Duan, Appl. Catal., B, 2008, 78, 61.
5 (a) F. Esch, S. Fabris, L. Zhou, T. Montini, C. Africh, P. Fornasiero,
G. Comelli and R. Rosei, Science, 2005, 309, 752; (b) W. Shan,
F. Liu, H. He, X. Shi and C. Zhang, Chem. Commun., 2011,
47, 8046; (c) Z. Liu, S. Zhang, J. Li and L. Ma, Appl. Catal., B, 2014,
144, 90.
6 Y. Peng, R. Y. Qu, X. Y. Zhang and J. H. Li, Chem. Commun., 2013, 49,
6215–6217.
7 (a) J. N. Brønsted, Chem. Rev., 1928, 5, 231; (b) A. Corma, Chem. Rev.,
1995, 95, 559.
Scheme 1 Proposed scheme for the correlation of base–acid and redox
properties of metal oxide catalysts in the SCR reaction (M = Ce or V).
sites (O2À or –OH) are co-existence on the catalyst surface. In the
3rd quadrant, the basic sites and strong redox properties would
enhance the over oxidation of NH3 and yield N2O, which will
decrease the N2 selectivity in the SCR reaction. However, the basic
sites are more likely being attacked by acid gases, i.e. HCl and SO2.
8 G. Martra, Appl. Catal., A, 2000, 200, 275.
9 H. Hattori, Chem. Rev., 1995, 95, 537.
To a certain degree, these sites could transform to acid sites for the 10 (a) F. Solymosi and H. Knozinger, J. Catal., 1990, 122, 166;
(b) L. Ferretto and A. Glisenti, Chem. Mater., 2003, 15, 1181.
11 J. C. Lavalley, Catal. Today, 1996, 27, 377.
12 Istadi and N. A. S. Amin, J. Mol. Catal. A: Chem., 2006, 259, 61.
SCR reaction after poisoning. But the redox cycle is always being
disturbed, which contributes to deactivation of SCR catalysts.
Moreover, in the presence of H2O and NH3, ammonium chloride 13 (a) J. Chen, M. Buzanowski, R. Yang and J. Cichanowicz, J. Air Waste
Manage. Assoc., 1990, 40, 1403; (b) L. Lisi, G. Lasorella, S. Malloggi
and G. Russo, Appl. Catal., B, 2004, 50, 251; (c) F. Y. Chang,
J. C. Chen and M. Y. Wey, Fuel, 2010, 89, 1919.
or sulfate forms and therefore blocks the active sites (4th quadrant).
The formation of these salts seems to be inevitable and only
regeneration is possible to eliminate the deposition.
14 H. Z. Chang, M. T. Jong, C. Z. Wang, R. Y. Qu, Y. Du, J. H. Li and
J. M. Hao, Environ. Sci. Technol., 2013, 47, 11692.
15 (a) N. Y. Topsøe, H. Topsøe and J. A. Dumesic, J. Catal., 1995,
151, 226; (b) G. Busca, L. Lietti, G. Ramis and F. Berti, Appl. Catal., B,
1998, 18, 1.
16 A. S. Negreira, S. Aboud and J. Wilcox, Phys. Rev. B, 2011, 83.
17 (a) V. I. Parvulescu, P. Grange and B. Delmon, Catal. Today, 1998,
46, 233; (b) G. G. Park, H. J. Chae, I. S. Nam, J. W. Choung and
K. H. Choi, Microporous Mesoporous Mater., 2001, 48, 337.
18 (a) G. S. Qi, R. T. Yang and R. Chang, Appl. Catal., B, 2004, 51, 93;
(b) C. Ciardelli, I. Nova, E. Tronconi, D. Chatterjee and B. Bandl-
Konrad, Chem. Commun., 2004, 2718; (c) J. Liu, Z. Zhao, C. M. Xu,
A. J. Duan and G. Y. Jiang, J. Phys. Chem. C, 2008, 112, 5930.
In conclusion, the base–acid properties of M–O or M–OH (M = Ce
or V) were correlated with the redox ability of metal oxide SCR
catalysts. The strength of their basicity is responsible for deactivation
behaviour in HCl and SO2 poisoning at medium and low tempera-
tures. HCl could react with the very basic sites (O2À) easily, resulting
in disturbance of the redox cycle on the CeMo/Ti catalyst and
formation of –OH base–acid sites. The former leads to a decrease
of SCR activity at low temperature. Meanwhile, the over oxidation of
NH3 and N2O formation are suppressed after poisoning.
10034 | Chem. Commun., 2014, 50, 10031--10034
This journal is ©The Royal Society of Chemistry 2014