Improvement of SOx-resistance of silver lean-DeNOx catalysts by supporting on CeO2-containing zirconia

Article abstract of DOI:10.1006/jcat.2002.3614

Efficient NO_x abatement from oxygen-rich exhausts, such as those originating from lean-fuelled engines, has become one of the major challenges for environmental catalysis. A comparison of the lean-deNO_x activity of Ag/Al₂O₃, Ag/Ce_0.16Zr_0.84O₂, and Ag/ZrO₂ showed that by supporting Ag on ZrO₂-containing supports, low-temperature deNO_x activity is greatly enhanced compared to Ag/Al₂O₃. The support was at the origin of this shift of activity to low temperature rather than in the presence of large Ag particles. The use of Ce_0.16Zr_0.84O₂ and ZrO₂ as supports conferred enhanced sulfur resistance compared to Al₂O₃-based materials in that regeneration of the SO₂-poisoned catalyst might be realized under moderate conditions.

Full text of DOI:10.1006/jcat.2002.3614

Journal of Catalysis 209, 271–274 (2002)
doi:10.1006/jcat.2002.3614
RESEARCH NOTE
Improvement of SOx-Resistance of Silver Lean-DeNOx Catalysts
by Supporting on CeO2-Containing Zirconia
∗
∗
∗
N. Hickey, P. Fornasiero, J. Kasˇpar,∗,1 M. Graziani, G. Martra,†
S. Coluccia,† S. Biella,‡ L. Prati,‡ and M. Rossi‡
∗
Dipartimento di Scienze Chimiche, Universita` di Trieste, Via Giorgieri 1, I-34127 Trieste, Italy; †Dipartimento di Chimica IFM,
Universita` di Torino, Via P. Giuria 7, I-10125 Torino, Italy; and ‡Dipartimento di Chimica Inorganica, Metallorganica e
Analitica—Centro CNR, Universita` di Milano, Via G.Venezian 21, I-20133 Milan, Italy
Received February 2, 2002; revised March 12, 2002; accepted March 12, 2002
ance of the system are improved compared to that of
Ag/Al2O3.
Thelean-deNOx activityofAg/Al2O3, Ag/Ce0.16Zr0.84O2, andAg/
ZrO2 was compared. The activity data showed that by supporting
Ag on ZrO2-containing supports, low-temperature deNOx activity
is greatly improved compared to Ag/Al2O3. TEM investigation sug-
gested that the support is at the origin of this shift of activity to low
temperature rather than in the presence of large Ag particles. More
important, use of Ce0.16Zr0.84O2 and ZrO2 as supports confers im-
proved sulphur resistance compared to Al2O3-based materials in
EXPERIMENTS AND METHODS
Ag was supported either by incipient wetness (IW) (us-
ing AgNO3) or by depositing a preformed Ag sol (8) (SOL)
ontothesupport (Agloading, 2wt%;Ce0.16Zr0.84O2, 100m2
g
−1; ZrO2, 40 m2 g−1; and γ -Al2O3, 100 or 180 m2 g−1).
that regeneration of the SO2-poisoned catalyst may be achieved
Samples were calcined at 650◦C before reaction. Activity
tests were performed using a reaction mixture of 1000 ppm
NO, 1000 ppm C3H6, and 5% O2 (balance He) and, when
present, 50 ppm SO2 and 10% H2O; the weight/flow rate
(W/F) was 0.10–0.025 g s ml−1. Activity profiles were ob-
tained either by increasing the temperature in steps of 25◦C
and monitoring activity for at least 40 min at each tempera-
ture, or by run-up experiments (heating rate, 10◦C min−1).
Analysis was performed by gas chromatography (C3H6,
CO2, CO, O2, N2O, N2) and a chemiluminescence detec-
tor (NO, NO2). Transmission electron micrographs (TEM)
were obtained with a Jeol 2000EX microscope. The samples
were ultrasonically dispersed in 2-propanol and deposited
on a copper grid covered with a lacey carbon film.
c
ꢀ 2002 Elsevier Science (USA)
under moderate conditions.
Key Words: deNOx catalysts; silver catalysts; sulphur poisoning;
ceria–zirconia catalysts.
INTRODUCTION
Efficient NOx abatement from oxygen-rich exhausts,
such as those originating from lean-fuelled engines, has be-
come one of the major challenges for environmental cata-
lysis. Literally hundreds of catalysts have been tested (1),
with transition-metal-containing zeolites and Pt/Al2O3 be-
ing among the most effective systems. They, however, re-
spectively suffer from poor hydrothermal stability and a
narrow range of activity (200–270◦C). Ag-containing cata-
lysts have attracted attention as inexpensive—compared
to Pt catalysts—and highly active systems (2–7). How-
ever, even these catalysts suffer from some major draw-
backs, showing activity only at relatively high tempera-
tures (around 500◦C) and high sensitivity to sulphur
poisoning, which makes them unsuitable for typical lean-
deNOx applications such as in diesel engines. Here, we
show the key role of ZrO2-based supports in modify-
ing the catalytic properties of supported Ag. Both low-
temperature lean-deNOx activity and the sulphur toler-
RESULTS AND DISCUSSION
Figure 1 compares NOx and C3H6 conversion curves for
Ag/Al2O3 (IW, 100; IW, 180) Ag/ZrO2 (IW, 40), and Ag/
Ce0.16Zr0.84O2 (SOL, 100) as a function of temperature.
There is a remarkable difference between Ag/Al2O3 and
the zirconia-based systems, as the range of temperatures
of NOx conversion is shifted down by about 150◦C for
the latter catalysts. Similar shifts of activity to low tem-
peratures have previously been reported for an Ag/ZrO2
catalyst compared to Ag/Al2O3 (4) and also for highly
loaded Ag/Al2O3 (>6 wt%) (2). Such shifts of deNOx ac-
tivity were generally attributed to the presence of large Ag
1 To whom correspondence should be addressed. Fax: +39-040-5583960.
E-mail: kaspar@units.it.
271
0021-9517/02 $35.00
c
ꢀ 2002 Elsevier Science (USA)
All rights reserved.

272
HICKEY ET AL.
species are active at high temperatures and form almost
60
50
40
30
20
10
0
exclusively N2 (2). The product distributions obtained over
the present Ag/Al2O3 samples are comparable to those
reported in the literature; for example, less then 1% of
N2O is formed at a NO conversion of 51% over Ag/Al2O3
(IW, 180). At Tmax (Table 1) 9 and 12% N2O was observed
over, respectively, Ag/Ce0.16Zr0.84O2 and Ag/ZrO2. With
regard to the conversion levels achieved, direct compari-
son of data obtained in different laboratories is difficult,
with space velocity, Ag loading, concentration of reduc-
tant, concentration of O2, and preparation method all re-
ported to be significant factors in the conversions achieved.
Rates at 450◦C of 1.8 × 10−7 mol of NOx converted g−1 s−1
and 3.3 × 10−7 mol of NOx converted g−1 s−1 were mea-
sured for Ag/Al2O3 (IW, 100) and Ag/Al2O3 (IW, 180),
respectively. These compare well with the values of 3.1 ×
10−7 mol of NOx converted g−1 s−1 at 440◦C and 2.5 ×
10−7 mol of NOx converted g−1 s−1 at 450◦C for the same re-
action under similar reaction conditions, calculated from
the data reported in Refs. (4, 2), respectively. The presence
of a “plateau” in the C3H6 conversion curve for Ag/Al2O3
(IW, 100) is an unexpected feature. We believe that this
may be attributed to the inhomogeneous Ag particle distri-
bution of this sample (see Table 1). As has been suggested
in previous reports (2, 4), the presence of large Ag particles
should promote unselective combustion at lower temper-
(a)
(b)
100
80
60
40
20
0
100
300
500
700
Temperature (°C)
FIG. 1. (a) NOx and (b) C3H6 conversions measured as a function
of temperature over (ꢀ) Ag/Ce0.16Zr0.84O2 (SOL, 100), (ꢁ) Ag/Al2O3
(IW, 180), (ꢂ) Ag/ZrO2 (IW, 40), and (ꢃ) Ag/Al2O3 (IW, 100) catalysts.
W/F = 0.05 g s ml−1
.
metallic particles under reaction conditions, which are se- atures. As such unselective combustion occurs at the ex-
lective towards N2O formation, e.g., 27% of the NO fed pense of the deNOx activity, this would also contribute to
was reduced to N2O at a total NO conversion of 43% the difference in deNOx activities of the two alumina-based
over 6 wt% Ag/Al2O3 (2). More dispersed oxidised Ag materials.
TABLE 1
DeNOx Activity and Effects of SO2 Addition to the Feed/H2 Regeneration on Steady-State DeNOx Activity at 450◦Ca
DeNOx activity
Maximum DeNOx
activity
DeNOx activity
at 450◦C (%)
BET surface
area
Ag particle
size
Activity
range
(◦C)
Preparation
method
Sample
(m2 g−1
)
(nm)
Tmax (◦C)
Conv. (%)
SO2-free
+SO2
After H2
Ag/Ce0.16Zr0.84O2
IW
SOL
100
100
See noteb
200–600
200–600
375
350
13
39
—
26
—
6
—
23
28c
22
9
25
5
28c
17
15c
0
Ag/ZrO2
Ag/Al2O3
IW
IW
IW
SOL
40
100
180
100
4
8
2.5
250–600
325–600
325–600
400–600
375
510
500
540
36
35
51
18
4.0d
21
6
11 7.5e
13 8g
50 f
20
25 f
6
31 f
14
a Reaction conditions as reported in the text. W/F = 0.05 g s ml−1. H2 regeneration at 450◦C in flow of H2 (5% in Ar) for 10 min. DeNOx
activity: conversion of NOx (NO + NO2). Three other Ag/Ce0.16Zr0.84O2 and Ag/ZrO2 catalysts were tested, giving analogous activity results (data not
reported).
b Due to poor contrast with the support no Ag particle could be detected at the various magnifications employed. The presence of Ag was analytically
detected by the TEM/EDS technique, which suggests that Ag should be highly dispersed with particles <1 nm, even though the presence of some very
large particles undetectable by TEM cannot be discounted.
c 10% of water was added to the feed.
d Inhomogeneous particle distribution was observed with a few large particles, at 24–25, 29–31, and 35 nm, accounting for about 60 mass%.
e Bimodal particle distribution was observed with about 8 mass% of particles both at 21 and at 27 nm.
f
Reaction temperature, 500◦C, the temperature of maximum deNOx activity.
g Bimodal particle distribution was observed with about 50 mass% of particles at 32 3 nm.

SOx RESISTANCE OF Ag/CeO2–ZrO2 DeNOx CATALYSTS
273
In an attempt to clarify the origin of this high activity at
30
20
10
(1)
(2)
(3)
low temperatures of the zirconia-based catalysts, a number
of different catalysts were tested (Table 1). The differences
in preparation were intended to create variations in the Ag
particle size. It can be seen that, irrespective of the synthesis
method or surface area of the support, the temperatures
where the deNOx activity is observed over the zirconia-
containing materials is always lower by ca. 150◦C compared
to Ag/Al2O3. It should also be noted that under ramped
reaction conditions (10◦C min−1), no differences in the ac-
tivity ranges of Ag/Al2O3 (IW, 100) and Ag/Ce0.16Zr0.84O2
(SOL, 100) were observed upon varying W/F by a factor
of four. Attempts to detect the Ag particle size by XRD
failed; likewise, the reducible nature of the Ce0.16Zr0.84O2
precluded the use of O2 or N2O chemisorption. Therefore,
TEM data for selected samples were obtained (Table 1).
The particle size distributions measured suggests that the
presence of large Ag particles is not the major factor re-
sponsible for the observed behaviour of the zirconia-based
materials. Rather, an interaction between the support and
the Ag particles appears at the origin of the high activity at
low temperatures.
Knowledge of the ability to withstand the poisoning ef-
fects of sulphur is an important factor in the assessment of
any material with potential deNOx application. Here, the
effects of addition of SO2 to the feed on the deNOx activity
were compared at 450◦C, where all the systems show appre-
ciable activity (Fig. 2). All catalysts suffer from a significant
deactivation as 50 ppm SO2 is added to the feed. However,
the response of the catalysts to SO2 poisoning/catalysts re-
generation is fundamentally different. A remarkably good
recovery of the deNOx activity could be achieved in the
case of zirconia-based materials by exposure to 5% H2 at
450◦C for 10 min at a flow rate of 60 ml min−1 (Fig. 2).
Effective catalyst regeneration could be achieved for sev-
eral deactivation/regeneration cycles. It is very important
that even milder regeneration procedures were quite ef-
fective: 50–80% of the initial activity of Ag/Ce0.16Zr0.84O2
could be recovered, for example, by pulsing 1 ml of H2
over the catalyst, or even by simply removing the SO2 from
the feed at 450◦C. In contrast, only limited regeneration of
the Ag/Al2O3 catalysts could be achieved under the same
gas-phase composition and time of reduction for tempera-
tures as high as 600◦C. This is also true for the most active
Ag/Al2O3 (IW, 180), where the effect of SO2 addition was
tested at 500◦C, i.e., at the temperature of maximum deNOx
activity. For comparison, a significant degree of regenera-
tion of SOx -poisoned Ag/Al2O3 could be achieved only by
employingrathersevereconditions(9). Thenoticeablyslow
deactivation of Ag/Ce0.16Zr0.84O2 (SOL, 100)—the activity
halves after 140 min—upon SO2 addition deserves some
comment:immediately afterthe additionofSO2 tothefeed,
very slow deactivation of the deNOx activity is observed as
long as NO2 is produced over the catalyst. Once production
of NO2 is totally suppressed, a constant conversion of 22%
NOx
(a)
NO2
0
0
0
0
100
200
300
400
30
20
10
0
NOx
NO2
(b)
100
200
300
400
30
20
10
0
NOx
(c)
200
NO2
150
50
100
Time (min)
FIG. 2. DeNOx activity at 450◦C: (1) activity of the fresh catalysts,
(2) effect of SO2 addition (50 ppm) to the feed on the activity, and (3) ac-
tivity in the absence of SO2 after regeneration in H2 (5% in He for 10 min)
over (a) Ag/Al2O3 (IW, 100) (b) Ag/Ce0.16Zr0.84O2 (SOL, 100), and (c) Ag/
ZrO2 (IW, 40). NOx , conversion of NO and NO2 (formed in the reaction)
to N2 and N2O; NO2, conversion of NO to NO2. Less then 4% of NO fed
is converted to N2O on each catalyst. Total NO conversion (not reported)
is sum of NOx and NO2 conversions. W/F = 0.05 g s ml−1
.
is observed for about 20 min; the activity then dropped to
6%. In contrast, over both Ag/ZrO2 and Ag/Al2O3 deac-
tivation due to SOx poisoning started almost immediately
after the SO2 addition to the feed, with the activity halving,
respectively, after 25 and 45 min. Note that differences in
BET area of the support cannot account for the different
activity decay. Upon regeneration of the poisoned catalysts
with H2 at 450◦C, NO2 formation is irreversibly suppressed
in the case of Ag/Ce0.16Zr0.84O2, whereas it is reversible
in the case of Ag/ZrO2. No significant amount of NO2 is
observed over Ag/Al2O3, due to the fact that the reaction
temperature of 450◦C is below the region where NO2 is
observed in the gaseous phase over this catalyst.
Experiments carried out also in the presence of water
(10%)confirmedthispromotingeffectoftheCe0.16Zr0.84O2

274
HICKEY ET AL.
support on the Ag catalyst and the easy regeneration from
SOx poisoning (Table 1). The addition of water to the feed
did not affect the catalyst activity or regeneration character-
istics over the time-on-stream investigated (12 h). Finally,
preliminary temperature-programmed reduction experi-
ments carried out over the SO2-poisoned catalysts showed
that significantly higher amounts of sulphur-containing
species are stored over Ag/Ce0.16Zr0.84O2 compared to
Ag/Al2O3. This suggests that the Ce0.16Zr0.84O2 support
may act as a sink for these species, accounting for the im-
proved sulphur resistance.
In summary, it is shown that use of Ce0.16Zr0.84O2 as sup-
port strongly modifies the deNOx activity of the Ag cata-
lysts, the activity range being decreased to lower tempera-
ture. A change of Ag particle size does not appear to be a
major factor responsible for this observation. However, of
more importance is the observation that the sulphur toler-
anceisimprovedfromthepointofviewthatfacileregenera-
tion from SO2 poisoning can be achieved. While the overall
activity of the catalyst investigated is not sufficient to pro-
vide a practical solution to the problem, these results may
provide new directions for research into lean-deNOx cata-
lysis, as, for example, it has been demonstrated for Ag/
Al2O3 that appropriate synthesis may result in improved
activity (10).
ACKNOWLEDGMENTS
University of Trieste, Fondo Trieste—1999, and MURST (Roma) “Pro-
getti di Ricerca di Rilevante Interesse Nazionale—2000,” Consortium
“INCA—Progetto Atmosfera Urbana,” are gratefully acknowledged for
financial support.
REFERENCES
1. Shelef, M., Chem. Rev. 95, 209 (1995).
2. Bethke, K. A., and Kung, H. H., J. Catal. 172, 93 (1997).
3. Kameoka, S., Ukisu, Y., and Miyadera, T., Phys. Chem. Chem. Phys.
2, 367 (2000).
4. Meunier, F. C., Ukropec, R., Stapleton, C., and Ross, J. R. H., Appl.
Catal. B 30, 163 (2001).
5. Meunier, F. C., Breen, J. P., Zuzaniuk, V., Olsson M., and Ross, J. R. H.,
J. Catal. 187, 493 (1999).
6. Martens, J. A., Cauvel, A., Francis, A., Hermans, C., Jayat, F., Remy,
M., Keung, M., Lievens, J., and Jacobs, P. A., Angew. Chem. Int. Ed.
37, 1901 (1998).
7. Masuda, K., Tsujimura, K., Shinoda, K., and Kato, T., Appl. Catal.
B 8, 33 (1996).
8. Porta, F., Prati, L., Rossi, M., Coluccia, S., and Martra, G., Catal. Today
61, 165 (2000).
9. Meunier, F. C., and Ross, J. R. H., Appl. Catal. B 24, 23 (2000).
10. Seker, E., Catvataio, J., Gulari, E., Lorpongpaiboon, P., and Osuwan,
S., Appl. Catal. A 183, 121 (1999).