Catalytic studies on ceria lanthana solid solutions I. Oxidation of methane

Article abstract of DOI:10.1016/S0021-9517(03)00044-7

The catalytic combustion of methane over ceria, lanthana, and the entire range of mixed oxides has been studied under reaction conditions free from limitations set by the transfer of heat and mass. Bulk structures of the catalysts were determined by X-ray diffraction and their surface compositions were monitored by XPS. Their surface areas were measured by BET methodology. Surface basicity was measured by temperature-programmed desorption of carbon dioxide. The combustion rate was close to first order on methane partial pressure, close to zeroeth order on oxygen partial pressure, and strongly retarded by carbon dioxide. A particular study was made of the solid solutions of lanthana in ceria where, operated at temperatures just above that of threshold activity, the two metal oxides act synergistically in the catalytic combustion of methane. It is proposed that cerium and lanthanum interact to form the catalytic active centre providing both redox and base functions.

Full text of DOI:10.1016/S0021-9517(03)00044-7

Journal of Catalysis 219 (2003) 286–294
www.elsevier.com/locate/jcat
Catalytic studies on ceria lanthana solid solutions
I. Oxidation of methane
M.F. Wilkes, P. Hayden, and A.K. Bhattacharya ∗
Warwick Process Technology Group, Department of Engineering, University of Warwick, Coventry CV4 7AL, UK
Received 9 January 2003; accepted 27 January 2003
Abstract
The catalytic combustion of methane over ceria, lanthana, and the entire range of mixed oxides has been studied under reaction conditions
free from limitations set by the transfer of heat and mass. Bulk structures of the catalysts were determined by X-ray diffraction and their
surface compositions were monitored by XPS. Their surface areas were measured by BET methodology. Surface basicity was measured by
temperature-programmed desorption of carbon dioxide. The combustion rate was close to first order on methane partial pressure, close to
zeroeth order on oxygen partial pressure, and strongly retarded by carbon dioxide. A particular study was made of the solid solutions of
lanthana in ceria where, operated at temperatures just above that of threshold activity, the two metal oxides act synergistically in the catalytic
combustion of methane. It is proposed that cerium and lanthanum interact to form the catalytic active centre providing both redox and base
functions.
2003 Published by Elsevier Inc.
Keywords: Methane; Combustion; Carbon dioxide; Ceria; Lanthana; Mixed oxide; Synergism
1. Introduction
and Flytzani-Stephanopoulos [7] where the activity of ce-
ria was increased slightly by doping with lanthana in atomic
fractions of 1 and 4.5 cation %. This study was extended
by Kundakovic and Flytzani-Stephanopoulos [8] who con-
cluded that the sintering characteristics of ceria were mod-
ified by doping, especially with up to 10% of lanthanum.
The higher activity achieved with lanthana as dopant was
attributed to various factors, including smaller ceria crystal
size, increased reducibility, and the introduction of extrinsic
oxygen vacancies. The higher activity achieved with zirco-
nia as dopant, which creates no extrinsic oxygen vacancies,
was attributed by others to increased oxygen mobility in the
defective fluorite structured mixed oxides [1].
This plethora of mechanistic proposals may be due not
only to the wide range of oxidation conditions practised but
also to the wide extent of catalyst preparative procedures
adopted. Most studies were confined to specific reaction con-
ditions or to narrow ranges of catalyst composition. Despite
the recognition of catalysis being a surface phenomenon,
all of these studies presented their findings in terms of the
composition of the bulk oxides. The present paper extends
the study of the lanthana–ceria catalysts to the complete
range of compositions, in terms of both bulk and surface
compositions, from pure ceria to pure lanthana. It identi-
The catalytic activity of pure and doped lanthanides has
been variously attributed to the presence of structural de-
fects [1], the participation of lattice oxide [2], the basicity
of the surface [3], a redox cycle [4], and surface oxygen
species [5]. Different though these classifications may be,
they are not completely unrelated.
Hattori et al. [4] correlated the catalytic activity of pure
lanthanide oxides in butane oxidation with conductivity. Op-
erating with an oxygen deficit relative to the stoichiometry
required for combustion, activity was limited by the ox-
idation of cations in their tertiary valency and hence re-
lated to the fourth ionisation potential. Ceria, praseodymia,
and terbia were the most active catalysts. Of lanthanide ox-
ide catalysts for the combustion of methane, Mackrodt et
al. [6] classified ceria as the most active with lanthana and
praseodymia its most active promoters: used together, these
two promoters acted synergistically. The optimum dopant
atomic fraction in the bulk oxide was between 12.5 and 25%.
The promoter action was confirmed in an initial study by Liu
*
Corresponding author.
E-mail address: akb@warwick.ac.uk (A.K. Bhattacharya).
0021-9517/$ – see front matter
2003 Published by Elsevier Inc.
doi:10.1016/S0021-9517(03)00044-7

M.F. Wilkes et al. / Journal of Catalysis 219 (2003) 286–294
287
fies those compositions forming solid solutions based on the
face-centred cubic (fcc) and hexagonal close-packed (hcp)
structures of ceria and lanthana, respectively. Surface com-
position and surface basicity are measured over the largest
range of mixed oxide compositions forming a single phase,
i.e., the fluorite structure of ceria as solvent. For these com-
positions, the roles of cerium and lanthanum in the catalytic
action are explored.
The conversion of methane and the formation of carbon
dioxide were measured: the mass balance was 100 3%
with 95% confidence limits. At 560 ◦C, tests showed that
the oxidation rate increased with linear gas velocity up to
about 12 cm/s and thereafter remained invariant with gas
rate up, at least, to 20 cm/s. Additionally, halving the cata-
lyst grain size, sieved from the same population of catalyst
grains, showed no significant change in conversion.
2.4. Temperature-programmed desorption of carbon
dioxide
2. Experimental
2.1. Catalyst preparation and characterisation
With the aim of characterising the adsorption/desorption
characteristics of the catalyst materials under conditions ap-
proximating to those of the catalytic reaction, temperature-
programmed desorptions were made under 20% oxygen in
helium. Preliminary tests showed that, for an adsorption pe-
riod of an hour, the resultant desorption, between 27 and
727 ◦C was essentially independent of adsorption tempera-
ture between room temperature and 200 ◦C.
These are presented in full elsewhere [9]. Mixed ox-
ide catalysts of the general formula Ce1−xLaxO2−x/2 were
prepared by co-precipitation from nitrates (Ce(NO3)3·6H2O
(99.99%, Alfa); La(NO3)3·6H2O (99.999%, Aldrich)) with
ammonium bicarbonate, washed free of alkali, dried in air
at 110 ◦C, crushed, decomposed in air at 450 ◦C, calcined at
800 ◦C for 8 h, and sized to 200–400 µm. The calcination
regime was selected as a compromise between vigorous cal-
cinations, which achieve slow sintering during the catalyst
tests, and mild calcinations, which form catalysts with ac-
curately measurable surface areas. Before catalytic reaction
testing, heating overnight to 600 ◦C in air cleaned the stored
samples. The bulk structures and surface compositions were
determined by XRD and ESCA, respectively. Catalyst sur-
face areas were measured by the single-point technique us-
ing the Brunauer–Emmett–Teller isotherm method both be-
fore and after the catalyst tests.
2.5. Diffuse reflectance infrared FT spectroscopy (DRIFTS)
In situ DRIFTS measurements were made by flowing gas
from the microreactor to a temperature-controlled environ-
mental cell (Graseby Specac) monitored with a spectrometer
(Mattson Galaxy 7020) operated in transmittance mode. Fol-
lowing the recording of background scans over KBr, catalyst
samples were introduced to the cell as finely ground powders
and exposed to the standard gas mixture at 500 ◦C for 1 h be-
fore recording the spectrum.
2.2. Reactor
3. Results
Gases of known composition are introduced from cylin-
ders via Brooks 5850 mass flow controllers. The microreac-
3.1. Oxidation rate as a function of oxide composition:
quantitative study of the effect of x in Ce1−xLaxO2−x/2
1
tor consists of 14 -, 81 -, and 16 -in. stainless steel tubes con-
nected using Swageloc fittings. The system is operated at
atmospheric pressure, with negligible pressure drop across
the catalyst bed. The reactant gases were mixed in the mani-
fold section that included a 7-µm filter, pre-heated to 120 ◦C,
and delivered to a heated quartz reaction tube. The fixed cat-
alyst bed consisted of a cylindrical plug of 200- to 400-µm
catalyst particles, 10-mm length and 4-mm diameter. Prod-
uct gases were analysed by gas chromatography (Hewlett-
Packard Model 5890) using a Carboxen 1000 column and
mass spectroscopy (Fisons Instruments).
Details of the physical characterisations of these catalysts
have been presented elsewhere, where it is concluded that
the pure oxides of this study are only partially soluble in
each other [9]. The fcc fluorite structure of ceria is retained
up to an ionic fraction of about 0.6 in lanthanum whilst
the hcp structure of lanthana is formed above an ionic frac-
tion of about 0.9. The intermediate compositions form mixed
phases. Fig. 1 and Table 1 present the XRD data. Whereas
we shall, in the first instance, present results from the entire
range of compositions, we have focused on the phase formed
over the largest range of compositions, i.e., the fluorite struc-
ture where ceria is the solvent.
2.3. Catalyst testing
The catalyst bed comprised a weighed amount (about
0.2 g) of particles (200–400 µm) located within a quartz
tubular reactor (4-mm i.d.) packed, with tapping, into a
length of at least 20 mm. The reacting gases were delivered
as 10% CH4 in helium (25 ml/min) and 21% O2 in helium
(118 ml/min), generating a linear gas velocity of 19 cm/s.
Specific activity is shown in Figs. 2a and 2b as functions
of bulk and surface compositions, respectively. The detailed
data on surface composition are presented and treated quan-
titatively elsewhere [9]. Doping either ceria or lanthana with
low levels of the other oxide promotes catalyst activity. Up
to the limit of the lanthana-in-ceria solid solutions at about

288
M.F. Wilkes et al. / Journal of Catalysis 219 (2003) 286–294
Fig. 2. Specific activity as a function of the composition of (a) the catalyst
bulk and (b) the catalyst surface.
Fig. 1. XRD patterns for Ce
La O
x
where 0 < x < 1.
1−x
2−x/2
bulk x = 0.6 [9], catalyst activity responds to progressive
additions of lanthana in the manner of a volcano plot. In the
lanthana-rich regions above the solution limit in ceria, activ-
ity exhibits a second, poorly defined, volcano plot response
as the bulk composition approaches that of pure lanthana.
However, in these regions the character of the mixed oxide
is extremely sensitive to its composition; e.g., solubility lim-
its are exceeded and the ceria-in-lanthana solution is formed
and moreover the data on surface composition, catalytic ac-
tivity, and adsorption of carbon dioxide are inadequate for
treating in the proposed semiquantitative treatment adopted
here. Accordingly, beyond some description of the pure lan-
thana case, we shall concern ourselves with the behaviour of
the system only up to the limit of the bulk lanthana-in-ceria
solid solution. Following our intention to relate specific ac-
tivity to the character of the surface phase, we shall present
the results in terms of the surface fraction of lanthanum
ions, xs.
Fig. 3. (a) The apparent energy of activation and (b) apparent
pre-exponential factor as functions of the surface ionic fraction of lan-
thanum.
3.2. Oxidation rate as a function of oxide composition:
qualitative study
Mackrodt et al. [6] claimed synergistic promotion of ce-
ria catalysts by, for example, praseodymia and lanthana with
optimal total promoter atomic fractions being about 20%
by metal atom. Since this is in the mid-range of composi-
tions exhibiting promotion in Fig. 2a, it forms a convenient
dopant level at which to search for promotion and synergism
amongst some mixed oxides in the present study. The results
are presented in Table 2.
The response of the reaction rate to temperature may be
expressed as an apparent energy of activation, which fur-
ther allows evaluation of the pre-exponential factor. These
are presented as a function of xs in Figs. 3a and 3b, respec-
tively.
Lanthana, preaseodymia, and gadolinia all promoted the
catalytic activity of ceria, both in terms of activity per unit
weight and activity per unit surface area. Whereas the in-
clusion praseodymia alone or in combination with lanthana
increased the activity per unit weight, the index adopted by
Table 1
Lattice parameters of Ce
La O
x
for compositions 0 < x < 1
1−x
2−x/2
x
d
111
d
200
d
220
d
311
a
111
(Å)
a
200
(Å)
a
220
(Å)
a (Å)
311
Average LP (Å)
0
3.1200
3.1295
3.1581
3.1642
3.1625
3.1894
3.2108
3.2809
3.3370
3.3389
2.7046
2.7062
2.7305
7.7366
2.7480
2.7637
2.7863
2.8330
2.8899
2.8911
1.9122
1.9189
1.9313
1.9389
1.9432
1.9610
1.9763
2.0095
2.0516
2.0505
1.6311
1.6357
1.6492
1.6546
1.6641
1.6733
1.6889
1.7191
1.7487
1.7507
5.404
5.420
5.470
5.481
5.478
5.524
5.561
5.683
5.780
5.783
5.409
5.412
5.461
5.473
5.496
5.527
5.573
5.666
5.780
5.782
5.409
5.427
5.463
5.484
5.496
5.547
5.590
5.684
5.803
5.800
5.410
5.425
5.470
5.488
5.519
5.550
5.601
5.702
5.800
5.806
5.408
5.421
5.466
5.481
5.497
5.537
5.581
5.684
5.791
5.793
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9

M.F. Wilkes et al. / Journal of Catalysis 219 (2003) 286–294
289
Table 2
Catalytic combustion activity as a function of promoter
Ce
La
Pr
Gd
Nb
Surface area
Rate
Specific rate
2
−1
−1 −1
◦
−1 −3
3
◦
(1 − x)
1
0.8
0.8
0.8
0.8
0.98
(m g
)
(mol of CO h kg 560 C)
2
(mol of CO h
m
(×10 ) 560 C)
2
0
0.2
0
0.1
0
0
0
0
0.2
0.1
0
0
0
0
0
0.2
0
0
0
0
0
0
0.02
15.2
1.1
2.8
3.9
3.9
4.0
0.21
0.07
0.20
0.22
0.20
0.22
0.01
13.9
17.9
19.0
18.6
15.0
0
Mackrodt et al., relative to the promoting effect of lanthana
alone, when expressed in terms of the rate per unit area,
no synergism was detected. Adopting specific activity as
the preferred measure of activity, it is concluded that there
is equal promoting effect among lanthana, preaseodymia,
and gadolinia. Access to good quality salts of lanthanum
favoured the study of promotion amongst its mixed oxides
with ceria.
the retarded kinetics from such a narrow study but it is
clear that the rate comprises at least two components: one
that is strongly retarded by carbon dioxide whilst the other
is unaffected or, at most, weakly retarded. The model dis-
played is
Rate = 1.25 + 1.3/(1 + 0.9[CO2])2.
3.3.2. Over lanthana
Within the range of compositions in Table 2, Fig. 2a
shows that lanthana invariably acts as a promoter. In marked
contrast to the other dopants in Table 2, niobia, in a dilute
solution, acts as a strong retarder.
Fig. 5a shows the catalytic combustion rate exhibits
lower than first-order dependence on the partial pressure of
methane at constant oxygen partial pressure. The solid line
is drawn according to a Langmuir-style expression:
3.3. Kinetics of catalytic oxidation
Rate = 4.22 × [CH4%]/(1 + 0.58[CH4%]).
3.3.1. Over ceria
The response of combustion rate to changing the partial pres-
sure of oxygen at constant methane partial pressure is pre-
sented in Fig. 5b, where the results are modelled according
to a Langmuir-style expression:
Fig. 4a shows the dependence of the combustion rate on
the partial pressure of methane at constant oxygen partial
pressure: the order of the response is one. The response of
combustion rate to changing the partial pressure of oxygen
at constant methane partial pressure is presented in Fig. 4b,
where the results are modelled according to a Langmuir-
style expression:
Rate = 1.4 × [O2%]0.5/(1 + 0.1 × [O2%]0.5).
As with the ceria case, carbon dioxide strongly retards the
catalytic combustion over lanthana. The experimental data
is presented in Fig. 5c where it is compared with the model:
Rate = 1.44 × [O2%]/(1 + 0.4 × [O2%]).
A strong retarding effect of added carbon dioxide is illus-
trated in Fig. 4c where the combustion rate at constant partial
pressures of methane (1.5%) and oxygen (15%) is presented
as a function of the partial pressure of added carbon diox-
ide. It is not possible to derive a rigorous expression for
Rate = 0.97 + 2.8/(1 + 0.92 × [CO2%]).
Again, the combustion rate may be divided into two parts:
one in which the retardation by carbon dioxide is strong and
one in which it is weak or nonexistent.
Fig. 4. Kinetics over the ceria catalyst: the effect of the partial pressure of (a) methane, (b) oxygen, and (c) carbon dioxide.

290
M.F. Wilkes et al. / Journal of Catalysis 219 (2003) 286–294
Fig. 5. Kinetics over the lanthana catalyst: the effect of the partial pressure of (a) methane, (b) oxygen, and (c) carbon dioxide.
Fig. 6. Kinetics over the Ce
La
O
catalyst: the effect of the partial pressure of (a) methane, (b) oxygen, and (c) carbon dioxide.
0.95 0.05 1.925
3.3.3. Over Ce1−xLaxO2−x/2
Table 3
The apparent responses to changes in the partial pressure of methane, oxy-
gen, and carbon dioxide in the catalytic combustion of methane over the
pure and mixed oxides Ce La O
Responses similar to those above to changes in the partial
pressure of these three gases were measured over a series
of mixed oxide catalysts. Those for the case x = 0.05 are
shown in Figs. 6a–c for methane, oxygen, and carbon diox-
ide as the independent parameters, respectively.
x
1−x
2−x/2
Bulk Surface Mean order Mean order % retardation on addition
x
x
on CH
on O
of 15% CO
2
4
2
0
0
1.0
0.8
0.8
0.8
0.8
0.7
0.2
0.2
0.3
0.3
0.3
0.4
42
55
57
60
63
70
In light of the complicated response to lanthanum dopant
above bulk ion fractions of 0.6, the limit of solubility of lan-
thana in ceria, the detailed study of the mixed oxide catalysts
has been restricted to below this point. It is of relevance to
define the responses to the partial pressures of methane, oxy-
gen, and carbon dioxide over the range of solid solutions in
which lanthana acts as a promoter. These are summarised in
Table 3 where the atomic fraction of lanthanum in the bulk
composition (x) was varied through 0 to 1.0. With increas-
ing levels of lanthana in the catalyst, it shows a decreasing
order on methane, an increasing order on oxygen, and an en-
hanced retardation by carbon dioxide.
0.001
0.05
0.3
0.4
1
0.044
0.176
0.504
0.609
1
solid solutions: the intermediate compositions exhibit inter-
mediary profiles. Table 4 shows the quantitative response of
the adsorption of carbon dioxide to the atomic fraction of
lanthanum in the bulk mixed oxide and the effect of dif-
ferent lanthanides. The volume of carbon dioxide adsorbed
increases as the concentration of lanthana is raised and that
the adsorption by the mixed oxides of ceria with lanthana,
preaseodymia, and gadolinia are broadly similar, although in
detail the lanthana mixed oxide absorbs more carbon dioxide
than those based on preaseodymia and gadolinia.
3.4. Temperature-programmed desorption of carbon
dioxide from ceria–lanthana mixed oxides,
Ce1−xLaxO2−x/2
Fig. 7 illustrates the TPD profiles of carbon dioxide des-
orbed from the two end members of the lanthana-in-ceria
On the basis that the adsorption is a surface phenomenon,
Fig. 8 shows it as a function of the atomic fraction of lan-

M.F. Wilkes et al. / Journal of Catalysis 219 (2003) 286–294
291
of carbon dioxide increases smoothly with the atomic frac-
tion of lanthanum in the surface.
3.5. Crystal size of Ce1−xLaxO2−x/2 catalysts
To test the conclusion drawn by Kundakovicand Flytzani-
Stephanopoulos [8] that crystallite size is a factor in the
determination of this catalytic activity, data on the crystal
size of this range of pure and mixed oxide catalysts, re-
ported elsewhere [9], are displayed in Table 5. It shows that
increasing the lanthanum atomic fraction results in progres-
sively smaller crystal sizes until a minimum is reached in the
range x = 0.6–0.7 after which additional lanthanum results
in enlarged crystallites.
Any correlation between catalytic activity and crystal-
lite size will be sought by comparing Fig. 2a and Table 5.
However, we shall seek to model catalytic specific activ-
ity in terms of the composition of the surface of the mixed
oxides and to include any response to crystallite size. Com-
bining data relating bulk and surface compositions of these
catalysts [9] with that in Table 5 enables the correlation of
crystallite size, D, with the surface cationic fraction of lan-
thanum, [xs], in the empirical equation
Fig. 7. TPD of carbon dioxide from Ce
Table 4
The adsorption of carbon dioxide on Ce
0.6, 0.8, and 1)
La O
for x = 0 and 0.6.
(x = 0, 0.2, 0.4,
2−x/2
x
1−x
2−x/2
La O
x
1−x
2
Bulk
Atomic fraction
No moles adsorbed per m
(mol of CO m (×10 )
−2
6
composition
of surface lanthanum
2
CeO
0
1.1
1.8
3.4
2
Ce La
O
O
O
O
0.41
0.61
0.79
0.88
1.0
0.8 0.2 1.9
Ce La
0.6 0.4 1.8
0.2
D = 478.2 − 341.7 × [xs]
.
(a)
Ce La
5.1
0.4 0.6 1.7
Ce La
10.7
13.6
1.5
0.2 0.8 1.6
La O
2
3
Ce Pr
O
0.8 0.2 1.9
4. Discussion
Ce Gd
O
1.6
0.8
0.2 1.9
4.1. On the selection of the promoter
On the question of synergistic promotion by praseodymia
and lanthana, it is concluded that although catalysts compris-
ing praseodymia are more active in weight terms, the units
adopted by Mackrodt et al. [6], this does not extend to spe-
cific activities expressed relative to surface areas. The higher
activity claimed by Mackrodt et al. is wholly attributable to
the increase in catalyst surface area due, in turn, to retarding
effect of the dopants on the sintering of the mixed oxide dur-
ing its calcination. Accordingly, the present study focused
on the promoting effect of a single dopant, lanthana, cho-
sen because of the detailed published data on the electrical
conductivity of ceria–lanthana mixed oxides.
4.2. Oxidation over pure ceria catalysts
Fig. 8. Amount of carbon dioxide desorbed as a function of atomic fraction
of surface lanthanum.
Although the main aim of this study was to characterise
the promotion of the catalytic combustion of methane over
lanthana-doped ceria, the results may be put into some per-
spective following an outline of the kinetics over the two
pure oxides.
Over ceria, under the conditions of the standard test, the
reaction rate is first order on methane, is zero order on oxy-
gen, and is retarded by carbon dioxide.
thanum in the surface. The shape suggests that the amount
of adsorbed carbon dioxide undergoes a step change when xs
exceeds 0.6, i.e., beyond the limiting solubility of lanthana
in ceria [9] where the bulk and, probably, the surface contain
segregated lanthana. It is a region richer in lanthanum than
is the subject of the present report. Of the mixed oxides with
compositions forming solid solutions in ceria, the adsorption
The form of the parametric responses may be accommo-
dated in the following reaction scheme because of Li and

292
M.F. Wilkes et al. / Journal of Catalysis 219 (2003) 286–294
Table 5
Crystallite size as a function of the composition of Ce
La O
x
catalysts
1−x
2−x/2
Bulk atomic fraction of lanthanum, x
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
1
Crystallite size (nm)
47.7
20.9
19.0
19.3
17.0
18.0
12.8
12.3
18.7
32.0
Xin [10]:
but carbon dioxide adsorption on the unsaturated sites only
then
Ce(IV)–O–Ce(IV) + CH4
k × PCH
4
↔ Ce(IV)–O(···CH4)–Ce(IV),
(1)
(2)
(3)
(4)
(5)
(6)
(7)
rate =
2
(1 + KCH × PCH + KCO × PCO
)
2
4
4
2
Ce(IV)–O− + Ce(III) + CH4
kꢀ × PCH
4
↔ Ce(IV)–O−(···CH4) + Ce(III),
+
,
(1 + KCꢀ H × PCH
)
2
4
4
Ce(IV)–O(···CH4)–Ce(IV)
where the first term refers to the cus and the second to
the ccs.
→ Ce(III)–OH− + Ce(III)–CH3
Ce(IV)–O−(···CH4) + Ce(III)
→ Ce(III)–OH− + Ce(III)–CH3
Ce(IV)–O− + Ce(III) + CO2
,
+
At constant partial pressures of two of the three gases in-
volved, this relationship reduces to a reasonable description
of each of the three experimental responses of the oxidation
rate over ceria to variations in the partial pressure of the third
gas.
+
,
↔ Ce(IV)–O−(···CO2) + Ce(III),
Ce(III)–OH− + Ce(III)–CH3+ + nO2
→ 2Ce(III) + CO2 + H2O,
2Ce(III) + mO2
4.3. Oxidation of pure lanthana catalysts
Over lanthana, under the conditions of the standard test,
the reaction rate is almost first order on methane, is about
half-order on oxygen, and is strongly retarded by carbon
dioxide. Reactive oxygen species on lanthana have been var-
iously attributed to associated and dissociated forms [11].
The current measurements marginally favour concepts based
on atomic oxygen,
→ Ce(IV)–O–Ce(IV) + Ce(IV)–O− + Ce(III).
Li and Xin identified two separate sites on ceria for the ad-
sorption of methane: one coordinatively saturated and the
other unsaturated. These are involved in reactions (1) and
(2), respectively: the second site is alternatively described as
surface lattice oxygen. A competitive adsorption of methane
and carbon dioxide on the (more basic) coordinatively un-
saturated sites is proposed in reactions (2) and (5). It is
proposed that reactions (3) and (4) are slow, whilst those in-
volving molecular oxygen, reactions (6) and (7), are fast.
In the present study only the characteristic IR bands at-
tributable to methane adsorbed on coordinatively saturated
sites (ccs) were observed under methane oxidation reaction
conditions in the in situ DRIFTS study (Fig. 9). No IR bands
attributable to methane adsorption on coordinatively unsat-
urated sites (cus) were detected. Operating at very low tem-
peratures, Li and Xin [10] had detected methane adsorbed
on both ccs and cus. It is relevant to this discussion that
Li and Xin [10] reported competition between adsorption
of methane and carbon dioxide on the cus but not on the
ccs. Of the two sites cus is assumed to be the most ba-
sic, facilitating heterolytic scission of a C–H bond, thereby
activating it to oxidation processes. It would seem that, in
contrast to the very low temperature experiments by Li and
Xin [10], the lifetime of any methane adsorbed on cus un-
der the high-temperature oxidation conditions used here is
too short compared with the time resolution of the DRIFTS
study.
SCH + CH4 ↔ SCH –CH4,
(8)
(9)
4
4
2SO + O2 ↔ SO–O,
SCH + CO2 ↔ SCH –CO2,
(10)
(11)
(12)
4
4
SO + CO2 ↔ SO–CO2,
SCH − CH4 + SO–O → products,
4
where SCH is a site adsorbing methane and carbon dioxide
4
competitively and SO is a site adsorbing oxygen. Were the
rate of reaction limited by an interaction between methane
and oxygen adsorbed on randomly distributed sites, then
1/2
2
rate = kKSCH PCH KS PO
4
O
O
4
CH
4
ꢀꢁꢁ
ꢂ
1 + KCH PCH + KS PCO
SCH
CH
4
4
4
2ꢂꢂ
CO2
4
ꢁ
1/2
SO
× 1 + KOS PO + KCO PCO
O
2
2
2
where KCSCHH4 , KOSO , KCSCOH4 , and KCSOO are the equilibrium
4
2
2
constants for reactions (8)–(12), respectively. At constant
partial pressures of two of the three gases involved, this re-
lationship reduces to a reasonable description of each of the
three experimental responses of the oxidation rate over lan-
thana to variations in the partial pressure of the third gas.
Assuming adsorption according to Langmuir isotherms,
the requirement of an empty site adjacent to that activating
methane, activity from both saturated and unsaturated sites

M.F. Wilkes et al. / Journal of Catalysis 219 (2003) 286–294
293
◦
under methane–oxygen gas at 500 C.
Fig. 9. DRIFTS of Ce
La O
x
1−x
2−x/2
4.4. Oxidation of mixed oxides of ceria and lanthana
between them:
β
δ
φ
ϕ
specific activity = α[Ce] + χ[La] + ε[Ce] [La] . (13)
Over the mixed oxides, the responses of the combustion
rate to changes in the partial pressures of the three gases
were broadly similar to those over the pure oxides but, as
shown in Table 3, there are distinct trends in the differences.
Restricting the discussion to the lanthana-in-ceria solid so-
lutions, introduction of lanthana induces a fall in the order
on methane, a rise in the order on oxygen, and an increase in
the retardation coefficient of carbon dioxide.
In summary, the dependence of reaction rate on gas com-
position suggests that catalytic activity is controlled mainly
by the activation of methane competing against its retarda-
tion by carbon dioxide. Increasing the concentration of lan-
thana in the mixed oxides progressively increases the activity
of adsorbed methane relative to that of adsorbed oxygen.
This increment in methane activity is reflected in a corre-
sponding rise in the retardation effect of carbon dioxide.
The TPD results parallel these trends in that, on the mixed
oxides with compositions forming solid solutions, the ad-
sorption of carbon dioxide increases in approximate propor-
tion to the atomic fraction of lanthanum in the surface. With
this background, we now discuss the effect of composition
of the solid solution on the catalytic specific activity.
Up to the limit of the formation of solid solutions of lan-
thana in ceria, the volcano form of the specific rate as a
function of atomic fraction of lanthanum in the surface of
these solutions presented in Fig. 1b suggests simultaneous
activity contributions from both cerium and lanthanum. In a
nonrigorous formulation of a parametric response model for
rate and surface composition over the range of the lanthana-
in-ceria solid solution, we may begin with the assumption
of contributions from cerium, lanthanum, and an interaction
Computer curve fitting Eq. (13), where the atom fractions
refer to surface values, to the experimental data simulates
Fig. 1a up to a bulk composition of x = 0.6 generating a
modelled peak maximum activity xs = 0.46. Omission of the
cerium term worsened the fit whereas omission of the lan-
thanum term had little effect. Accordingly, the cerium term
is retained whilst that for lanthanum is omitted. This is con-
sistent with the structure of the solid solutions as deduced
from the XRD patterns that comprise lines attributable to
ceria but none to lanthana. The simplified model is
0.9
0.8
0.8
specific activity = 0.07[Ce] + 0.5[Ce] [La]
.
(14)
Eq. (14) and general chemistry invites further simplification
of the model by rounding up the exponents to 1, as in
specific activity = 0.1[Ce] + 0.7[Ce][La].
(15)
Specific rate data evaluated using this model is compared
with the experimental data in Fig. 10. Over the range of com-
position forming lanthana-in-ceria solid solutions, the exper-
imental specific activities exhibit a weakly defined symmet-
rical curve peaking at xs in the range 0.4–0.6. The modelled
function maximises at xs = 0.44.
Such analysis of the compositional dependency of the
specific rate has tacitly assumed an unchanging regime of
kinetics. Table 3 shows that the orders on methane, oxygen,
and carbon dioxide do, in practice, change slightly over the
range of the lanthana-in-ceria solid solution. Inspection of
the apparent fundamental kinetic parameters, the energy of
activation and the pre-exponential factor, adds a further di-
mension to the analysis. Fig. 3 shows that these parameters

294
M.F. Wilkes et al. / Journal of Catalysis 219 (2003) 286–294
methane is adsorbed in a molecular state that is activated by
polarisation of its C–H bond which, given adequate strength
of the acid–base pair, leads to heterolytic scission with for-
mation of a carbanion. In the present case of lanthana-in-
ceria solid solutions, lanthanum provides the population and
strength of the acid–base pair to form carbanions that are
subsequently oxidised by cerium. This proposal is consistent
with the contribution from the cerium–lanthanuminteraction
term maximising at equi-atomic fractions.
At maximum catalytic activity in the present study, the
concentrations of both bulk and surface lanthanum are con-
siderably in excess of that required for maximum anionic
conductivity [14], casting doubt on a requirement of either
elevated oxygen ion mobility or enhanced reduction of the
bulk oxide [8], although these may well be factors deter-
mining the promotion of ceria by oxides less basic than
lanthana [1].
Fig. 10. Experimental and modelled specific rates as a function of atomic
fraction of lanthanum in the surface phase.
exhibit compositional dependencies approximately similar
to that of the specific reaction rate, all three exhibiting max-
ima in the range xs = 0.4–0.6.
5. Conclusion
From this broad investigation of the system it is con-
cluded that in lanthana-in-ceria solid solutions the two
metals synergistically catalyse the combustion of methane.
Cerium contributes a redox capability to the reaction, whilst
the basicity of lanthanum activates adsorbed methane The
fundamentals of the reaction would be better understood by
further adsorption studies, particularly of the co-adsorption
of methane and carbon dioxide. The range of catalysts might
be suitably extended to stronger bases than lanthana and the
scope of the reaction to the oxidation and co-oxidation of
ammonia.
In their study of methane combustion over doped transi-
tion metal–ceria catalysts, Liu and Flytzani-Stephanopoulos
[7] concluded that the activity enhancement induced by lan-
thanide dopants could not be explained on the basis of the
association energies of the dopant ion-oxygen vacancy pair.
Kundakovic and Flytzani-Stephanopoulos [8], whilst recog-
nising they are not independent parameters, list crystal size,
defect formation, and reducibility of surface oxygen species
as factors important for the catalytic activity of doped ce-
ria. Kundakovic and Flytzani-Stephanopoulos’s conclusion
that activity exhibits a reciprocal relationship with crystal
size was based on mixed oxides in which the atomic frac-
tion of lanthanum was less than 0.1. The data presented here
in Fig. 2a and Table 5 show no simple correlation between
catalytic specific activity and crystallite size holds over the
range of solubility of lanthana in ceria. Furthermore, inclu-
sion into Eqs. (13)–(15) of a term comprising crystallite size,
D, as a function of xs (see Eq. (a)), appreciably lowered their
correlation coefficients. Consequently, we reject significant
contributions from crystallite size and explore the scope for
an alternative determinant of specific activity.
References
[1] F. Zamar, A. Trovarelli, C. de Leitenburg, G. Dolcetti, J. Chem. Soc.
Chem. Commun. (1995) 965.
[2] A. Trovarelli, Catal. Rev. 38 (4) (1996) 439.
[3] V.R. Choudhary, V.H. Rane, J. Catal. 130 (1991) 411.
[4] T. Hattori, J.-I. Inoko, Y. Murukami, J. Catal. 42 (1976) 60.
[5] K.D. Campbell, M. Zhang, J.H. Lunsford, J. Phys. Chem. 92 (1988)
750.
[6] W.C. Mackrodt, M. Fowles, M.A. Morris, US Patent 4,940,685.
[7] W. Liu, M. Flytzani-Stephanopoulos, J. Catal. 153 (1995) 304.
[8] Lj. Kundakovic, M. Flytzani-Stephanopoulos, J. Catal. 179 (1998)
203.
[9] M.F. Wilkes, P. Hayden, A.K. Bhattacharya, Appl. Surf. Sci., in press.
[10] C. Li, Q. Xin, J. Phys. Chem. 96 (1992) 7714.
[11] A.M. Maitra, Appl. Catal. A 104 (1993) 11;
D.J. Ilett, M.S. Islam, J. Chem. Soc. Faraday Trans. 89 (1993) 3833.
[12] A.M. Maitra, J. Therm. Anal. 36 (1990) 657.
It is proposed that the catalytic activity originates at
cerium sites but mainly from a synergistic interaction be-
tween cerium sites and lanthanum sites. It is further pro-
posed that Eq. (15) provides a qualitative expression of this
activity. With increasing lanthana content, the contribution
from the pure cerium term decreases, whilst that from the
interaction term maximises at equi-atomic fractions. Studies
of the distribution of acid and base strengths on ceria and
lanthana have concluded that lanthana comprised the most
and strongest basic site [12,13]. As Choudary and Rane have
said of the catalytic action of ceria [13], it is proposed that
[13] V.R. Choudhary, V.H. Rane, J. Catal. 130 (1991) 411.
[14] T. Takahashi, in: J. Hladik (Ed.), Physics of Electrolytes, Volume 2—
Thermodynamics and Electrode Processes in Solid State Electrolytes,
Academic Press, London, 1972, ISBN 0-12-349802-3.