Effect of surface composition of yttrium-stabilized zirconia on partial oxidation of methane to synthesis gas

Article abstract of DOI:10.1016/j.jcat.2004.09.025

Catalytic partial oxidation of methane to synthesis gas (CPOM) over yttrium-stabilized zirconia (YSZ) was studied within a wide temperature window (500-1100°C). The catalysts were characterized by X-ray fluorescence (XRF) and low-energy ion scattering (LEIS). The influence of calcination temperature, Y₂O₃ content, and especially impurities such as CaO, TiO₂, and Na₂O on catalytic performance were investigated. Creation of active sites by doping with Y₂O₃ improves the catalytic performance of ZrO₂ significantly. The surface composition rather than the bulk composition determines the catalytic performance of the catalysts in CPOM. As long as YSZ catalyst is not contaminated, the composition of the outermost surface of calcined YSZ is independent of both the concentration of Y₂O₃ in the bulk and calcination temperature; the surface always contains 12 ± 2 mol% Y₂O ₃ due to segregation of Y₂O₃. Calcination at higher temperatures creates more active sites per square meter, while the catalyst loses surface area via sintering. The same sintering treatment causes the activity of YSZ containing traces of (earth) alkali oxides to collapse. The effect is probably due to segregation of the impurities to the surface, which either blocks the active surface of YSZ catalyst or forms new phases with different catalytic properties. However, it cannot be ruled out that enhanced segregation of Y₂O₃ contributes to this effect as well. Heterogeneous reactions occur concurrently with homogeneous reactions at temperatures above 950°C during CPOM over YSZ. At such high temperatures, CPOM, steam- and CO₂ reforming, and reverse water-gas shift occur in competition during CPOM. These reforming reactions of methane result in a significant increase in synthesis gas selectivity, although the catalyst activity is still too low to reach thermodynamic equilibrium.

Full text of DOI:10.1016/j.jcat.2004.09.025

Journal of Catalysis 230 (2005) 291–300
www.elsevier.com/locate/jcat
Effect of surface composition of yttrium-stabilized zirconia on partial
oxidation of methane to synthesis gas
Jianjun Zhu a, Jan G. van Ommen a, Arie Knoester b, Leon Lefferts a,∗
a
Catalytic Processes and Materials, Faculty of Science and Technology, Institute of Mechanics, Processes and Control Twente (IMPACT),
University of Twente, PO Box 217, 7500 AE Enschede, The Netherlands
b
Calipso B.V., PO Box 513, 5600 MB Eindhoven, The Netherlands
Available online 13 January 2005
Abstract
Catalytic partial oxidation of methane to synthesis gas (CPOM) over yttrium-stabilized zirconia (YSZ) was studied within a wide tem-
◦
perature window (500–1100 C). The catalysts were characterized by X-ray fluorescence (XRF) and low-energy ion scattering (LEIS). The
influence of calcination temperature, Y O content, and especially impurities such as CaO, TiO , and Na O on catalytic performance were
2
3
2
2
investigated. Creation of active sites by doping with Y O improves the catalytic performance of ZrO significantly. The surface compo-
2
3
2
sition rather than the bulk composition determines the catalytic performance of the catalysts in CPOM. As long as YSZ catalyst is not
contaminated, the composition of the outermost surface of calcined YSZ is independent of both the concentration of Y O in the bulk and
2
3
calcination temperature; the surface always contains 12 2 mol% Y O due to segregation of Y O . Calcination at higher temperatures
2
3
2 3
creates more active sites per square meter, while the catalyst loses surface area via sintering. The same sintering treatment causes the activity
of YSZ containing traces of (earth) alkali oxides to collapse. The effect is probably due to segregation of the impurities to the surface, which
either blocks the active surface of YSZ catalyst or forms new phases with different catalytic properties. However, it cannot be ruled out that
enhanced segregation of Y O contributes to this effect as well. Heterogeneous reactions occur concurrently with homogeneous reactions at
2
3
◦
temperatures above 950 C during CPOM over YSZ. At such high temperatures, CPOM, steam- and CO reforming, and reverse water–gas
2
shift occur in competition during CPOM. These reforming reactions of methane result in a significant increase in synthesis gas selectivity,
although the catalyst activity is still too low to reach thermodynamic equilibrium.
2004 Elsevier Inc. All rights reserved.
Keywords: Partial oxidation of methane; Synthesis gas; Yttrium-stabilized zirconia; Contamination; LEIS
1. Introduction
The catalytic partial oxidation of methane (CPOM) to
synthesis gas has been studied intensively for about two
decades [1–5]. The first row of transition metals (Ni, Co) and
noble metals (Ru, Rh, Pd, Pt, and Ir) have been reported as
active catalysts for CPOM. However, these metal catalysts
are also good catalysts for CH4 decomposition to carbon
and H2. The problem of carbon deposition on these catalysts
remains to be solved. Recent studies have focused on devel-
oping a highly active and stable catalyst for CPOM. Mixed
metal oxides, NiO–MgO solid solution [6], Ni–BaTiO3 [7],
Ni–Mg–Cr–La–O [8], and Ca0.8Sr0.2Ti0.2Ni0.2 [9], were re-
ported to be highly active and selective catalysts at high
space velocities (102–103 m3/(kg h)) and high temperatures
(> 700◦C) with improved carbon resistance. A significant
There has been a growing interest in the direct conver-
sion of natural gas to CO and H2 as a potential alternative
to steam reforming of methane. The H2/CO ratio of synthe-
1
2
sis gas obtained by direct oxidation, CH4 + O2 → CO +
2H2 (ꢀH = −22.2 kJ/mol, 1000 K), is more favorable for
methanol synthesis and the Fischer–Tropsch process. More-
over, direct partial oxidation of methane to synthesis gas is
mildly exothermic, whereas the steam reforming reaction is
extremely endothermic.
*
Corresponding author. Fax: +31 53 4894683.
E-mail address: l.lefferts@utwente.nl (L. Lefferts).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.09.025
2004 Elsevier Inc. All rights reserved.

292
J. Zhu et al. / Journal of Catalysis 230 (2005) 291–300
temperature gradient is observed in the catalyst bed that is
due to highly exothermic combustion of methane at the en-
trance, followed by reforming reactions in the rest of the
catalyst bed (indirect mechanism). On the other hand, direct
partial oxidation of methane to synthesis gas was claimed
over a very short contact time in the order of a millisec-
ond at extremely high temperatures (> 1000◦C) by Schmidt
and co-workers [1–3]. Evaporation of metal in the form of
volatile metal oxides formed at high temperatures, especially
in presence of oxygen, causes deactivation of the metal cat-
alysts. Metal loss in the form of volatile oxide, which results
in deactivation of the catalyst, is also a serious problem in
the ammonia oxidation process operated under almost the
same conditions [10,11].
Selective partial oxidation of methane to synthesis gas
over oxide catalysts has often been mentioned as a major
side reaction in both oxidative coupling of methane over rare
oxides [12] and partial oxidation of methane to formalde-
hyde on titania [13]. In our laboratory, CPOM over some
irreducible metal oxides such as TiO2, La2O3, and ZrO2-
based mixed oxides was studied by Steghuis et al. [14,15].
Yttrium-stabilized zirconia (YSZ) showed the best catalytic
performance among these oxide catalysts. Compared with
metal catalysts, activity and selectivity are lower for YSZ.
However, catalyst stability is superior and the lower selec-
tivity can be dealt with by the introduction of a second
metal-based reforming catalyst that can be kept stable be-
cause contact with oxygen at high temperatures is avoided,
as described in our previous work [16].
respectively. Table 1 gives the details for all catalysts used in
the present work.
2.2. Catalyst characterization
BET surface area was determined by nitrogen adsorption
at 77 K with a Micromeritics ASAP 2000 instrument. Prior
to each measurement the sample was degassed at 300 ◦C for
2 h under 0.1 Pa.
Chemical composition was measured by X-ray fluores-
cence (XRF) with a PW1480 (Philips) apparatus.
Surface composition of fresh samples was measured by
low-energy ion scattering (LEIS) with the Calipso setup [17],
which has a double toroidal electrostatic energy analyzer.
This gives a sensitivity that is a factor of 1000 higher than
that of the cylindrical mirror analyzer used in the past [18].
Another advantage of this LEIS is that the signals are vir-
tually independent of the specific surface area of the cata-
lyst [19]. As described in Ref. [20], before surface analysis,
contaminations such as water and hydrocarbons were re-
moved from the sample surface by annealing to 300 ◦C and
oxidation with atomic oxygen (Oxford Applied Research
Atom/Radical Beam source MPD21). This treatment results
in a clean YSZ surface without a change in its inherent com-
position and structure [20]. The LEIS measurements were
4
made at room temperature with 3 keV He+ ions to deter-
mine concentrations of the impurities on the surface. Since Y
and Zr are neighbors in the periodic system, it is impossible
to separate them with He ion scattering. The Y2O3 content
is determined, therefore, with 5-keV 20Ne+ ions. Both pure
ZrO2 (Alfa, 99.9%) and Y2O3 (Alfa, 99.999%) were used
as references. The Y2O3 concentration on the surface was
estimated by fitting LEIS spectra of YSZ.
The objective of the present work is to determine whether
the catalytic performance of YSZ can be further improved in
terms of selectivity and activity by an increase in the opera-
tion temperature or by a change in the catalyst surface area.
The roles of consecutive steam and CO2 reforming reactions
are investigated as well. The influence of the catalyst surface
area and the surface composition is studied in detail.
2.3. Catalytic measurement
The catalytic reactions were carried out in a fixed-bed
reactor operated at atmospheric pressure. The catalyst (ca.
0.3 g) diluted with the same amount of α-Al2O3 was held by
quartz wool. It was pretreated in an alumina reactor (inner
diameter 4 mm) with a gas flow containing He (50 ml/min)
and O2 (30 ml/min) at 800 ◦C for 1 h and subsequently
cooled down to reaction temperature in the flow of helium.
To minimize the contribution of gas-phase reactions, two
thin alumina sleeves were put in front of and behind the cat-
alyst bed; these were used as thermocouple wells as well.
The blank experiments were performed either with 0.6 g
α-Al2O3 or in the empty reactor. The typical reaction con-
2. Experimental
2.1. Catalysts
All catalysts were made of ZrO2-based powder stabi-
lized with different amounts of Y2O3, which were provided
by TOSOH (Japan) and GIMEX (the Netherlands). Except
for pure ZrO2, which was calcined at 900 ◦C, YSZ sam-
ples were calcined at different temperatures (900, 1000, or
1100 ◦C) in air for 15 h. After calcination, the powder was
pressed, crushed, and sieved to 0.3–0.6 mm particles. Ac-
cording to the Y2O3 content as weight percentage and calci-
nation temperature, the catalysts are referred to as YSZ5A,
YSZ9A, YSZ12A, YSZ12B, YSZ12C, YSZ14A, YSZ14B,
and YSZ14C. In the sample code, YSZ means yttrium-
stabilized zirconia, the number refers to the weight percent-
age of Y2O3 in the sample, and A, B, and C represent the
different calcination temperatures, 900, 1000, and 1100 ◦C,
ditions were as follows, unless specified otherwise: PCH
=
4
0.118 bar, PO = 0.059 bar, balance with helium; the total
2
flow rate was kept at 170 ml/min (STP). The temperature
gradient of the catalyst bed caused by the reaction is about
2 ◦C at 800 ◦C when diluted CH4 and O2 are converted for
40 and 100%, respectively. For the purpose of adding steam
to the reaction, the reactant mixture can be passed through
a steam saturator with the switch of a four-port valve. The

J. Zhu et al. / Journal of Catalysis 230 (2005) 291–300
293
Table 1
Compositions, surface areas and suppliers of the samples used in the experiments
a
Catalyst
Calcination
temperature ( C)
S
Composition (mol%)
Surface Y O
Supplier
g
2
3
b
◦
2
(m /g)
concentration (mol%)
ZrO
Y O
TiO
HfO
2
2
2
3
2
ZrO
2
900
900
900
900
1000
1100
15.4
14.9
13.1
13.1
10.2
7.0
98.76
95.63
93.74
90.46
0.023
3.19
5.13
8.46
0.048
0.005
0.005
0.005
1.16
1.17
1.13
1.08
–
14
–
11
14
12
Tosoh, Japan
YSZ5A
YSZ9A
YSZ14A
YSZ14B
YSZ14C
YSZ12A
YSZ12B
YSZ12C
900
1000
1100
22.0
13.0
4.5
90.98
7.82
0.279
0.92
19
–
29
Gimex,
The Netherlands
a
Composition of the sample, measured with XRF.
Surface composition, measured with LEIS.
b
partial pressure of steam was determined by temperature of
the saturator. The reaction temperature was varied, and the
system was kept at steady-state for 2 h for analysis. On-
line gas chromatography with Carboxan 1000 and Haysep
N columns was used to analyze the effluent gases from the
reactor. N2 was used as an internal standard. Conversions
(X) and yields (Y) were calculated according to:
CHi4n − CHo4ut
Oi2n − Oo2ut
XCH
=
;
XO
=
;
4
2
Oi2n
CHi4n
COo2ut − COi2n
COout
YCO
=
;
YCO
=
=
;
2
CHi4n + COi2n
CHi4n
Ho2ut
H2Oout − H2Oin
YH
=
;
YH O
.
2
2
2CHi4n + H2Oin
2CHi4n
4
+
4
+
Fig. 1. LEIS spectra (3 keV, He ) of fresh YSZ12C with different He
ion doses.
3. Results
3.1. Surface composition
(0.1 × 1015 ions/cm2). Contaminating elements, such as
F, Na, Al/Si, K/Ca, and Cr/Mn, are detected on the outer-
most surface of YSZ12C. After extended sputtering (21.4 ×
1015 ions/cm2), intensities of the impurities decrease signif-
icantly, though they are still detectable, and at the same time,
the intensity of the Y–Zr LEIS signal increases by a factor of
2.4. So, it is roughly estimated that more than half of the out-
ermost surface of YSZ12C is covered with impurities in the
form of oxides. In contrast to YSZ12C (calcined at 1100 ◦C),
the spectrum of YSZ12A (calcined at 900 ◦C), as shown in
Fig. 2, shows less impurity on its outermost surface, and the
shielding of Y and Zr is much weaker.
LEIS spectra of YSZ14A (calcined at 900 ◦C) and
YSZ14C (calcined at 1100 ◦C) are shown in Fig. 3. Unlike
on the surfaces of YSZ12A and YSZ12C, hardly any impuri-
ties were detected on the surfaces of YSZ14A and YSZ14C.
The high purity of YSZ14 samples was confirmed by the
fact that removal of impurities via sputtering caused only a
modest increase in the Y–Zr signal, as shown in Fig. 4. This
effect was much smaller than that observed in Figs. 1 and 2.
Comparable results were obtained for YSZ5A as well.
All samples studied in the present work are listed in Ta-
ble 1. The major contaminations detected by XRF were
TiO2 and HfO2, which were found in all samples. However,
compared with the samples provided by TOSOH (Japan),
more TiO2 was found in YSZ12A, YSZ12B, and YSZ12C,
which were provided by GIMEX (the Netherlands). Surface
composition was measured with LEIS. All YSZ catalysts
from TOSOH showed identical surface Y2O3 concentrations
within experimental error, independent of the Y2O3 content
in bulk of the catalyst and of the calcination temperature. In
contrast, high Y2O3 concentrations were detected in the out-
ermost layer of YSZ samples provided by GIMEX. More-
over, it was also observed in these samples that more Y2O3
segregates to the outermost surface when the sample is cal-
cined at higher temperatures.
Fig. 1 shows LEIS spectra of YSZ12C sputtered with dif-
4
ferent He+ ion intensities. A weak signal due to Y and/or
Zr, which cannot be separated, is observed in the spec-
trum of YSZ12C after sputtering with a low 4He+ ion dose

294
J. Zhu et al. / Journal of Catalysis 230 (2005) 291–300
4
+
4
+
Fig. 4. LEIS spectra (3 keV, He ) of fresh YSZ14C with different He
ion doses.
4
+
Fig. 2. LEIS spectra (3 keV, He ) of fresh YSZ12A and YSZ12C.
4
+
Fig. 3. LEIS spectra (3 keV, He ) of fresh YSZ14A and YSZ14C.
Fig. 5. Conversion/yield as a function of reaction temperature for CPOM
on YSZ14A. Catalyst: 0.3 g YSZ14A diluted with 0.3 g α-Al O ;
2
3
CH :O :He = 2:1:14, F
= 170 ml/min.
4
2
total
The major bulk impurities TiO2 and HfO2, detected by
XRF, were not observed on the outermost surface of any of
the samples with LEIS.
YSZ9A. In contrast, pure ZrO2 and YSZ12B showed quite
different catalytic activities and selectivities in the whole
temperature window (500–900 ◦C). For comparison, activ-
ities and selectivities of five ZrO2-based catalysts at 600 ◦C
are presented in Fig. 6. Except for YSZ12B, YSZ catalysts
are much more active than pure ZrO2. The methane and
oxygen conversions over YSZ5A, YSZ9A, and YSZ14A are
twice as high as they are over ZrO2 at 600 ◦C. However, YSZ
catalysts show higher selectivities for CO and H2 than does
ZrO2. Similar catalytic performance is observed for YSZ5A,
YSZ9A, and YSZ14A at 600 ◦C.
Compared with other YSZ catalysts, YSZ12B shows, in
Fig. 6, much lower activity at 600 ◦C, which is even lower
than that of ZrO2, but higher selectivity for CO and H2.
However, at 800 ◦C, when oxygen is completely consumed,
similar methane conversions are observed for all YSZ cata-
lysts, including YSZ12B. However, the product distribution
over YSZ12B is quite different. In a comparison of YSZ12B
3.2. CPOM over YSZ
3.2.1. ZrO2 and YSZ
Catalytic tests were carried out over pure ZrO2 and
yttrium-stabilized ZrO2 (YSZ) catalysts with comparable
surface areas. The catalytic performance of YSZ14A is
shown in Fig. 5 for a temperature window from 500 to
900 ◦C. Methane conversion reaches 31.5% at 650 ◦C, when
oxygen is consumed completely. CO, CO2, H2, and H2O are
major products of CPOM over YSZ14A, in addition to a
small amount of hydrocarbons (C2H4 and C2H6) formed via
oxidative coupling in the high-temperature region. Selectivi-
ties to the major products vary with reaction temperature. In
contrast to temperatures above 600 ◦C, selectivity to CO is
higher than that for CO2 at temperatures below 600 ◦C. Sim-
ilar catalytic performances were observed for YSZ5A and

J. Zhu et al. / Journal of Catalysis 230 (2005) 291–300
295
◦
Fig. 8. CH conversion rate at 600 C as a function of calcination temper-
4
ature. Catalyst: 0.3 g, diluted with 0.3 g α-Al O ; CH :O :He = 2:1:14,
2
3
4
2
◦
F
= 170 ml/min.
total
Fig. 6. Catalytic performances of ZrO and YSZ catalysts at 600 C.
2
Catalyst: 0.3 g, diluted with 0.3 g α-Al O ; CH :O :He = 2:1:14,
2
3
4
2
F
= 170 ml/min.
total
than the initiation temperature of CPOM over YSZ12A.
Compared with the YSZ catalyst, minor activity of α-Al2O3
is observed at temperatures below 850 ◦C. As shown in
Fig. 7, the initiation temperature and T90, the temperature
at which 90% oxygen is converted, decrease with increas-
ing catalyst surface area. Complete oxygen conversion is
reached over YSZ12A with a surface area of 22 m2/g at
700 ◦C; in contrast, for example, on YSZ12C with a surface
area of 4.5 m2/g, the oxygen conversion is well below 10%
at the same temperature. CH4 conversion increases dramat-
ically over all catalysts with reaction temperature as long as
oxygen is not exhausted. Only a slight increase in CH4 con-
version is observed after oxygen is consumed completely.
3.2.3. Methane conversion rate
Experimental evaluation, for example, via the activation
energy, is not feasible because the heat of reaction influences
the reaction temperature and because catalysts are too ac-
tive to allow differential experiments at the space velocities
that can be achieved. Therefore, mass transfer limitation was
evaluated by calculation in this work. According to the usual
criteria (the Carberry number, and the Wheeler–Weisz mod-
ulus [21]), it can be concluded that mass transfer limitation
can be neglected under the reaction conditions described in
Fig. 6.
Fig. 7. CH and O conversions as a function of reaction tempera-
4
2
ture and surface area. Catalyst: 0.3 g, diluted with 0.3 g α-Al O ;
2
3
2
CH :O :He=2:1:14, F
YSZ12B 13 m /g; YSZ12C 4.5 m /g.
= 170 ml/min. Surface area: YSZ12A 22 m /g;
4
2
total
2
2
with other YSZ catalysts, significantly higher selectivities to
CO and H2O are observed, and selectivities for CO2 and H2
are lower. Therefore, the effect of the secondary reactions on
the product distribution was studied for YSZ12 catalysts in
more detail at temperatures above 900 ◦C.
Fig. 8 shows CH4 conversion rates in terms of molCH
4
m
−2 s−1 at 600 ◦C over YSZ12 and YSZ14 catalysts with
varying surface area via sintering at different temperatures.
The methane conversion rates over YSZ14 catalysts are sig-
nificantly higher than those over YSZ12 catalysts. The cal-
cination temperature influences the CH4 conversion rate sig-
nificantly. However, the trends are quite different for these
two series of catalysts. For YSZ12 catalysts, the CH4 con-
version rate (in molCH4 m−2 s−1) decreases by a factor of
5 when the calcination temperature is increased from 900
to 1100 ◦C. In contrast, the CH4 conversion rate per square
3.2.2. Influence of surface area
Fig. 7 shows the variation of CH4 and O2 conversions
with the catalyst surface area and reaction temperature.
A significant influence of the surface area on conversions
of methane and oxygen is observed over YSZ12 catalysts.
Homogeneous reactions in the gas phase started at around
900 ◦C (empty reactor), which is more than 400 ◦C higher

296
J. Zhu et al. / Journal of Catalysis 230 (2005) 291–300
Fig. 10. Conversion/yield as a function of reaction temperature for oxidation
in the empty reactor. CH :O :He = 2:1:14, F = 170 ml/min.
Fig. 9. Conversion/yield as a function of reaction temperature for CPOM
on YSZ12C. Catalyst: 0.3 g YSZ12C diluted with 0.3 g α-Al O ;
4
2
total
2
3
CH :O :He = 2:1:14, F
= 170 ml/min.
4
2
total
meter over YSZ14 catalysts increases by a factor of 1.5 after
calcination at 1100 ◦C.
3.3. CPOM over YSZ at temperatures above 900 ◦C
CPOM was carried out over YSZ12 series of catalysts at
temperatures above 900 ◦C to investigate the effect of the
secondary reactions, such as CO2 and steam reforming of
methane and the water–gas shift reaction, on product distri-
bution. Fig. 9 shows conversions and yields as a function
of reaction temperature for CPOM over YSZ12C. Because
of the small surface area and high impurity coverage al-
ready mentioned, a low activity (XCH < 5%) was observed
4
at temperatures below 800 ◦C, whereas conversions of CH4
and O2 increased significantly above 800 ◦C. The highest
yields of water and CO2 were obtained at 950 ◦C; oxygen
was exhausted, and about 45% of the methane was con-
verted. Yields of CO and H2 increased dramatically at the
expense of CO2, H2O, and methane when the temperature
was increased further.
Fig. 11. Influence of CO or H O addition to the feed on conversion and
2
2
◦
yields in partial oxidation of methane on YSZ12C at 950 C. Catalyst: 0.2 g
YSZ12C diluted with 0.2 g α-Al O ; F
= 170 ml/min. Normal feed:
2
3
total
CH :O :He = 2:1:14; H O addition: CH :O :H O:He = 2:1:0.6:13.4;
4
2
2
4
2
2
CO addition: CH :O :CO :He = 2:1:0.4:14.6.
2
4
2
2
Naturally, homogeneous reaction in the gas phase con-
tributes more when the temperature is increased. As shown
in Fig. 10, homogeneous reactions can be neglected at tem-
peratures below 950 ◦C. However, they will certainly con-
tribute at temperatures above 950 ◦C.
conversion and yields of CO and H2O increase significantly,
whereas the H2 yield remains unchanged.
3.5. Steam reforming of methane
The steam reforming reaction was carried out over
YSZ12C under the same gas hourly space velocity (GHSV)
of methane as used in the normal CPOM. As shown in
Fig. 12, the initial reaction temperature for steam reform-
ing of methane is about 850 ◦C. Methane conversion reaches
18% at 1100 ◦C. In addition to the major reforming products
(CO and H2), CO2 is also detected in the product mixture at
temperatures above 1000 ◦C.
3.4. Coupling CPOM with steam or CO2 reforming of CH4
Effects of the addition of H2O and CO2 to the feed
of CPOM over YSZ12C at 950 ◦C are shown in Fig. 11.
The product distribution obviously changed when H2O was
added to the feed, whereas the methane conversion increased
slightly. Selectivities for CO and H2 increased, and CO2 se-
lectivity decreased.
Steam reforming of methane was also carried out in an
empty alumina reactor under identical reaction conditions.
Methane conversion is only 3.8% at 1100 ◦C, which is much
Compared with the addition of H2O, adding CO2 to the
feed significantly influences not only the product distribution
but also methane conversion. Fig. 11 shows that methane

J. Zhu et al. / Journal of Catalysis 230 (2005) 291–300
297
4. Discussion
4.1. Active surface of YSZ
Comparing ZrO2 with three YSZ catalysts, YSZ5A,
YSZ9A, and YSZ14A, Fig. 6 clearly shows that yttrium-
stabilized ZrO2 is more active and selective than ZrO2.
These four catalysts were provided by TOSOH and have
comparable surface areas and impurity contents. Before we
discuss this improved catalytic performance in more detail,
the possible nature of the active sites will be discussed first.
Improvement of catalytic performance has been related to
the presence of vacancies in many oxidation processes over
mixed oxides [22], for example, the oxidation of propane
over ZrO2 and YSZ [23] and oxidative coupling of methane
over rare earth oxides doped with strontium fluoride [24],
Y2O3-CaO [25], and SrO- or ZnO-doped La2O3 [26]. The
positive effect of doping is attributed to an increase in con-
centration of oxygen vacancies, which favors the adsorption
and activation of oxygen. The material nature of YSZ has
been studied extensively by both experimental and theoreti-
cal methods [27,28], although much less is known about its
surface structure. It is well known that a lot of oxygen vacan-
cies can be created in ZrO2 with the addition of ions with a
low valence, such as Y3+, even though pure ZrO2 contains
oxygen vacancies in significant concentration [29,30]. Al-
though the catalytic performance of a ZrO2-based catalyst is
strongly influenced by the concentration of oxygen vacan-
cies, the precise role of the vacancies in the activation of
methane and/or oxygen is still not clear. This is therefore the
subject of ongoing work in our laboratory.
Fig. 12. Steam reforming reaction of methane on YSZ12C at different
temperatures. Catalyst: 0.2 g YSZ12C diluted with 0.2 g α-Al O ; feed:
2
3
CH :H O:He = 20:6:174 (ml/min).
4
2
Obviously, the concentration of oxygen vacancies in YSZ
increases with the Y2O3 concentration, which would in-
crease the activity of YSZ. However, in a comparison of
YSZ5A with YSZ9A and YSZ14A, it is obvious that these
three catalysts have similar activities and selectivities at
600 ◦C (Fig. 6). The results of LEIS measurements listed
in Table 1 show enrichment of Y2O3 in the outermost sur-
face of these three catalysts. Almost identical composi-
tions of the outermost surface are observed for YSZ5A and
YSZ14A (12 2 mol% Y2O3), despite the difference in
their bulk compositions. Comparable segregation phenom-
ena were also reported by Theunissen et al. [31], who stud-
ied the composition of grain boundaries and the surfaces of
dense (ZrO2)100 − x(Y2O3)x (x = 2–16) by AES and XPS.
de Ridder et al. [32] arrived at the same conclusion with
LEIS. Therefore, the similarity in the catalytic performance
of YSZ5A, YSZ9A, and YSZ14A can be attributed to the
similar surface composition (Y/Zr ratio), which appears to
be independent of the bulk Y2O3 concentration. In the range
of Y2O3 concentrations used in this work, the activity and
selectivity of YSZ catalysts apparently are not influenced by
the bulk composition.
Fig. 13. CO reforming reaction of methane on YSZ12C at different tem-
2
peratures. Catalyst: 0.2 g YSZ12C diluted with 0.2 g α-Al O ; feed:
2
3
CH :CO :He = 20:7.73:174 (ml/min).
4
2
smaller than that measured over YSZ12C. Except for small
amounts of CO and H2, no CO2 was detected as a product in
the effluent of the reactor.
3.6. CO2 reforming of methane
Fig. 13 shows methane conversion and product distribu-
tion as functions of reaction temperature for CO2 reform-
ing of methane over YSZ12C. Under the same GHSV of
methane as used in normal CPOM, CO2 reforming reac-
tion of methane starts at 800 ◦C and occurs significantly at
temperatures above 1000 ◦C. H2O is also observed in the
product at temperatures above 900 ◦C, though CO and H2
are the major products. Methane conversion over YSZ12C
reaches 20% at 1100 ◦C; in contrast, only 4% of methane is
converted in the empty alumina reactor.

298
J. Zhu et al. / Journal of Catalysis 230 (2005) 291–300
4.2. Effects of contamination
al. [37], who found stronger segregation in YSZ for cations
with low valence, such as Na+, Ca2+, and Mg2+, compared
with the high-valence impurities, like Ti4+ and Ce4+ [37].
The catalytic performance of YSZ12B is quite different
from that of YSZ14A, as shown in Fig. 6, though both sur-
face area and the Y2O3 concentration in the bulk are almost
identical. As discussed above, the catalytic performance of
the ZrO2-based catalyst is strongly influenced by the sur-
face composition. The major differences between YSZ12B
and YSZ14A are the surface concentration of contaminants,
which will be discussed first, and the Y2O3 concentration in
the surface layer, which will be discussed second.
4.3. Effects of calcination temperature
As discussed above, calcination at high temperature re-
sults in high coverage of impurities on the surface. The
conversion rate of methane per surface area increases with
the calcination temperature for the YSZ14 catalysts (Fig. 8),
on which only minor surface contamination is detected with
LEIS (Figs. 3 and 4). Moreover, these three catalysts have
almost identical surface Y2O3 concentrations (Table 1). Ap-
parently, high-temperature calcination increases the number
of active sites via modification of the surface structure. Work
is ongoing to elucidate this effect. The same phenomenon
was also observed in the oxidation of CO over La–Ca–Mn–
O perovskites by Isupova et al. [35]. Characterization of the
concentration of the active sites on the surface of YSZ is un-
der way in our laboratory.
The surface of the YSZ12 catalysts is significantly con-
taminated with oxides, mainly CaO, as detected by LEIS
(Figs. 1 and 2). The surface of the YSZ14 catalysts is much
less contaminated (Figs. 3 and 4). Based on LEIS measure-
ments, it is estimated that about half of the outermost surface
of YSZ12B is covered with these contaminants, mainly CaO.
Significant segregation of traces of contaminants is consis-
tent with the work of Brongersma and co-workers [33,34].
These authors reported that the surface of YSZ was com-
pletely covered by contaminants after calcination at 1000 ◦C,
although the coverage with impurities was limited to 16%
after thermal treatment at 700 ◦C for 1 h. Unlike the dense
YSZ samples used by Brongersma et al., the surface area
of the YSZ samples used in this work is up to 22 m2/g. The
surface-to-bulk ratio of the samples in this work is thus much
higher, resulting in less surface contamination. On the other
hand, the diffusion distance that must be traversed for con-
taminants to reach the surface is much shorter, which would
lead to more surface contamination. Moreover, calcination
at high temperature not only results in the segregation of the
impurities, but also causes the sample to lose surface area via
sintering (Table 1), thus decreasing the surface-to-bulk ratio.
These effects together lead to higher impurity coverage on
the surface of YSZ12C compared with YSZ12A, as shown
in Fig. 2. The segregation of CaO to the surface either simply
blocks the active surface of YSZ or forms new phases, such
as calcium-stabilized zirconia (CSZ), which should have cat-
alytic properties different from those of YSZ. This explains
the significant difference in catalytic performance between
YSZ12B and pure YSZ catalysts (Fig. 6) and the declining
methane conversion rate (Fig. 8) with increasing calcination
temperature for YSZ12 catalysts. Blocking of the active sur-
face by CaO was also suggested by Isupova et al. [35] to
cause the decline in activity of La–Ca–Mn–O perovskites
for CO oxidation.
4.4. CPOM at high temperatures
The extremely high CO selectivity of YSZ12B at 800 ◦C
invoked more detailed investigation of the catalytic perfor-
mance of the YSZ12 series of catalysts at higher tempera-
tures. As discussed above, because of the small surface area
and high impurity coverage, YSZ12C shows a relatively low
activity (XCH < 5%) at temperatures below 800 ◦C (Fig. 9).
4
After all oxygen is exhausted, the product distribution varies
with further increasing temperature, though only a slight in-
crease in the CH4 conversion is observed. This indicates that
reactions between CH4 and products (CO2, H2O), such as
steam reforming and CO2 reforming of methane, or among
products, that is, the reverse water–gas shift reaction, occur
at higher temperatures. Oxidation of methane in the empty
reactor, as shown in Fig. 10, occurs significantly at tem-
peratures above 950 ◦C. However, in a comparison of the
product distribution over YSZ12C (Fig. 9) with the product
distribution in the empty reactor (Fig. 10), significant differ-
ences are obvious, especially in the yields of H2 and CO2.
Moreover, a significant difference in the level of conversions
was observed at temperatures up to 1100 ◦C with and with-
out catalyst. These observations indicate that the catalyst
still plays an important role, even at very high temperatures.
The simultaneous occurrence of homogeneous and heteroge-
neous reactions was also reported, for example, by Leveles
et al. [38] for oxidative dehydrogenation of propane over
Li/MgO and by many other authors for oxidative coupling of
methane [39,40]. In addition to homogeneous and heteroge-
neous reactions taking place in parallel, the actual situation
is much more complicated, because the catalyst may both
generate and scavenge radical intermediates involved in the
homogeneous reaction. This results in a complex reaction
scheme, which is outside the scope of this work.
However, at this stage it cannot be ruled out that high
Y2O3 concentration in the surface would result in a decrease
in activity, for example, via clustering of the resulting oxy-
gen vacancies. Clustering of oxygen vacancies is known to
decrease the mobility of vacancies in the bulk of YSZ at con-
centrations above 8–12 mol% [36].
Two major impurities, TiO2 and HfO2, detected in the
bulk of YSZ12B by XRF (shown in Table 1) have not been
detected on the surface of YSZ12B by LEIS (see Figs. 1
and 2). This is in agreement with the results of Ross et

J. Zhu et al. / Journal of Catalysis 230 (2005) 291–300
299
The increase in CO and H2 yields after the addition of
H2O to the feed (Fig. 11) indicates that steam reforming of
methane (CH4 +H2O → CO+3H2) occurs during CPOM at
high temperatures, which is further confirmed by the direct
observation of steam reforming of methane over the catalyst
(Fig. 13). The conversion of methane over YSZ12C (about
18% at 1100 ◦C in Fig. 13) is much higher than in the empty
reactor, which gives only 3.8% methane conversion.
The activity of YSZ12C for CO2 reforming of methane is
obvious from the significant methane conversion observed
during CO2 reforming of methane over YSZ12C (Fig. 13),
which is 5 times higher than in the empty reactor. Water
formation during CO2 reforming of methane over YSZ12C
clearly indicates that the reverse water–gas shift reaction
(CO2 + H2 → CO + H2O) occurs as well. The change in
the product distribution that occurs when CO2 is added to
the feed of CPOM (Fig. 11) confirms that both CO2 reform-
ing of methane (CH4 + CO2 → 2CO + 2H2) and the reverse
water–gas shift reaction take place during CPOM.
namic equilibrium is not even approached, even at very high
temperatures, because of insufficient reforming activity.
Acknowledgments
This work was performed under the auspices of NIOK,
the Netherlands Institute of Catalysis Research. Stichting
Technische Wetenschappen (STW, Dutch Foundation of Ap-
plied Sciences) is gratefully acknowledged for financial sup-
port under project number UCP-5037. The authors thank
Prof. H.H. Brongersma, and Dr. M. de Ridder (Calipso LEIS
Expertise Center) for LEIS measurements and fruitful dis-
cussions. Thanks to Ing. V. Skolnik for the measurement of
the BET surface area and to Ing. J.A.M. Vrielink for XRF
analysis.
References
However, the reforming activity of YSZ is not sufficient
to reach equilibrium, even at 1100 ◦C. Many homogeneous
and heterogeneous reactions are involved in CPOM over
YSZ at temperatures above 900 ◦C, as discussed before. The
activity of the empty reactor for reforming reactions, al-
though limited, suggests that these reforming reactions pro-
ceed via an interweaved homogeneous–heterogeneous reac-
tion network, similar to CPOM as discussed above.
[1] D.A. Hickman, L.D. Schmidt, Science 259 (1993) 343.
[2] D.A. Hickman, E.A. Haupfear, L.D. Schmidt, Catal. Lett. 17 (1993)
223.
[3] D.A. Hickman, L.D. Schmidt, AIChE J. 39 (1993) 1164.
[4] V.R. Choudhary, V.H. Rane, A.M. Rajput, Appl. Catal. A 162 (1997)
235.
[5] A.T. Ashcroft, P.D.F. Vernon, M.L.H. Green, Nature 344 (1990) 319.
[6] V.R. Choudhary, B.S. Uphade, A.S. Mamman, J. Catal. 172 (1997)
281.
[7] R. Shiozaki, A.G. Andersen, T. Hayakawa, S. Hamakawa, K. Suzuki,
M. Shumizu, K. Takehira, J. Chem. Soc., Faraday Trans. 93 (1997)
3235.
5. Conclusions
[8] P. Chen, H.B. Zhang, G.D. Lin, K.R. Tsai, Appl. Catal. A 166 (1998)
343.
[9] T. Hayakawa, H. Harihara, A.G. Andersen, K. Suzuki, H. Yasuda, T.
Tsunoda, S. Hamakawa, A.P.E. York, Y.S. Yoon, M. Shimizu, K. Take-
hira, Appl. Catal. A 149 (1997) 391.
[10] M.M. Karavayev, A.P. Zasorin, N.F. Kleshchev, Catalytic Oxidation
of Ammonia, Khimia, Moscow, 1983.
[11] T.H. Chilton, Chem. Eng. Progr. Monogr. Ser. 3 (1960) 56.
[12] K.D. Campbell, H. Zhang, J.H. Lunsford, J. Phys. Chem. 92 (1988)
750;
Catalytic partial oxidation of methane to synthesis gas
was studied over ZrO2, pure YSZ, and YSZ containing ox-
ide contaminations in a wide temperature window. Doping
ZrO2 with Y2O3 generates active sites, resulting in improved
catalytic performance. The composition of the outermost
surface rather than the bulk of YSZ clearly determines its
catalytic performance.
The composition of the outermost surface of calcined
YSZ is independent of both the concentration of Y2O3 in
the bulk and the calcination temperature as long as the YSZ
catalyst is not contaminated; the surface always contains
12 2 mol% Y2O3 due to segregation of Y2O3. More ac-
tive sites are created per square meter after calcination at
higher temperatures.
S. Wada, H. Imai, Catal. Lett. 8 (1991) 131.
[13] A. Parmaliana, F. Arena, J. Catal. 167 (1997) 57.
[14] A.G. Steghuis, PhD thesis, University of Twente, The Netherlands
(1998).
[15] A.G. Steghuis, J.G. van Ommen, J.A. Lercher, Catal. Today 46 (1998)
91.
[16] J. Zhu, M.S.M. Mujeebur Rahuman, J.G. van Ommen, L. Lefferts,
Appl. Catal. A 259 (2004) 95.
[17] H.H. Brongersma, A. Gildenpfennig, A.W. Denier van der Gon, R.D.
van de Grampel, W.P.A. Jansen, A. Knoester, J. Laven, M.M. Viitanen,
Nucl. Instrum. Meth. B 190 (2002) 11.
Sintering causes the activity of YSZ containing traces of
(earth) alkali oxides to collapse. The effect is probably due
to segregation of the impurities to the surface. However, it
cannot be ruled out that enhanced segregation of Y2O3 con-
tributes to this effect.
[18] H.H. Brongersma, N. Hazewindus, J.M. van Nieuwland, A.M.M. Ot-
ten, A.J. Smets, Rev. Sci. Instr. 49 (1978) 707.
[19] W.P.A. Jansen, A. Knoester, A.J.H. Maas, P. Schmit, A. Kytökivi, A.W.
Denier van der Gon, H.H. Brongersma, Surf. Interface Anal. (2004),
in press.
[20] M. de Ridder, R.G. van Welzenis, H.H. Brongersma, Surf. Interface.
Anal. 33 (2002) 309.
Heterogeneous reactions occur concurrently with ho-
mogeneous reactions at temperatures above 950 ◦C during
CPOM over YSZ. The catalytic performance of YSZ can
be improved by a rise in the reaction temperature to 950 ◦C
and above, due to dry reforming, steam reforming, and the
reverse water–gas shift reactions. Nevertheless, thermody-
[21] R.A. van Santen, P.W.N.M. van Leeuwen, J.A. Moulijn, B.A. Averill,
Catalysis: An Integrated Approach, Elsevier, Amsterdam, 1999.

300
J. Zhu et al. / Journal of Catalysis 230 (2005) 291–300
[22] E.N. Voskresenskaya, V.G. Roguleva, A.G. Anshits, Catal. Rev.-Sci.
Eng. 37 (1) (1995) 101.
[23] M. Labaki, S. Siffert, J.F. Lamonier, E.A. Zhilinskaya, A. Aboukais,
Appl. Catal. B 43 (2003) 261.
[24] R. Long, Y. Huang, W. Weng, H. Wan, K. Tsai, Catal. Today 30 (1996)
59.
[25] Y. Osada, S. Koike, T. Fukushima, S. Ogasawara, T. Shikada, T.
Ikariya, Appl. Catal. 59 (1990) 59.
[26] H. Borchert, M. Baerns, J. Catal. 168 (1997) 315.
[27] J.P. Goff, W. Hayes, S. Hull, M.T. Hutchings, K.N. Clausen, Phys Rev.
B 59 (1999) 14202.
[28] S. Fabris, A.T. Paxton, M.W. Finnis, Acta Materialia 50 (2002) 5171.
[29] N. Mommer, T. Lee, J.A. Gardner, W.E. Evenson, Phys. Rev. B 61
(2000) 162.
[30] E. Karaoetrova, R. Platzer, J.A. Gardner, E. Schutfort, J.A. Sommers,
W.E. Evenson, J. Am. Ceram. Soc. 84 (2001) 65.
[31] G.S.A.M. Theunissen, A.J.A. Winnubst, A.J. Burggraaf, J. Mater. Sci.
Lett. 8 (1989) 55.
[32] M. de Ridder, A.G.J. Vervoort, R.G. van Welzenis, H.H. Brongersma,
S. Wulff, W.F. Chu, W. Weppner, Nucl. Instrum. Methods Phys. Res.
B 190 (2002) 732.
[33] M. de Ridder, A.G.J. Vervoort, R.G. van Welzenis, H.H. Brongersma,
Solid State Ionics 156 (2003) 255.
[34] M. de Ridder, R.G. van Welzenis, H.H. Brongersma, U. Kreissig, Solid
State Ionics 158 (2003) 67.
[35] L.A. Isupova, S.V. Tsybulya, G.N. Kryukova, G.M. Alokina, N.N.
Boldyreva, I.S. Yakovleva, V.P. Ivanov, V.A. Sadykov, Solid State Ion-
ics 141–142 (2001) 417.
[36] A. Bogicevic, C. Wolverton, Phys. Rev. B 67 (2003) 024106.
[37] I.M. Ross, W.M. Rainforth, D.W. McComb, A.J. Scott, R. Brydson,
Scripta Materialia 45 (2001) 653.
[38] L. Leveles, K. Seshan, J.A. Lercher, L. Lefferts, J. Catal. 218 (2003)
296.
[39] K.M. Dooley, S.Y. Chen, J.R.H. Ross, J. Catal. 145 (1994) 402.
[40] G.C. Hoogendam, K. Seshan, J.G. van Ommen, J.R.H. Ross, Catal.
Today 21 (1994) 333.