Doping and defect association in AZrO
3
(A Ca, Ba) and
LaMO
3
(M Sc, Ga) perovskite-type ionic conductors †
M. Saiful Islam,* Peter R. Slater, Julian R. Tolchard and Tim Dinges ‡
Materials Chemistry Laboratory, Chemistry Division, University of Surrey, Guildford,
UK GU2 7XH. E-mail: m.islam@surrey.ac.uk
Received 20th February 2004, Accepted 5th April 2004
First published as an Advance Article on the web 23rd July 2004
Computer simulation techniques have been used to investigate the defect chemistry of perovskite-structured
ionic conductors based upon AZrO
3
(A = Ca, Ba) and LaMO
3
(M = Sc, Ga). Our studies have examined dopant
site-selectivity, oxide ion migration and dopant–defect association at the atomic level. The energetics of dopant
incorporation in AZrO
3
show strong correlation with ion size. We predict Y
3+
to be one of the most favourable
dopants for BaZrO
3
on energetic grounds, which accords with experimental work where this cation is the commonly
used acceptor dopant for effective proton conduction. Binding energies for hydroxy–dopant pairs in BaZrO
3
are
predicted to be favourable with the magnitude of the association increasing along the series Y < Yb < In < Sc.
This suggests that proton mobility would be very sensitive to the type of acceptor dopant ion particularly at higher
dopant levels. Oxygen vacancy migration in LaScO
3
is via a curved pathway around the edge of the ScO
6
octahedron.
Dopant–vacancy clusters comprised of divalent dopants (Sr, Ca) at the La site have significant binding energies in
LaScO
3
, but very low energies in LaGaO
3
. This points to greater trapping of the oxygen vacancies in doped LaScO
3
,
perhaps leading to higher activation energies at increasing dopant levels in accord with the available conductivity
data.
1 Introduction
Oxygen ion and proton conductivity in perovskite-structured
oxides have attracted considerable attention owing to the range
of electrochemical applications (e.g. fuel cells, gas sensors, cer-
amic membranes) and the fundamental fascination with trans-
port phenomena in solid state materials. A range of perovskite
ceramics, particularly cerates (ACeO
3
) and zirconates (AZrO
3
),
exhibit significant proton conductivity.
1
An important example
is the development of a potentiometric gas sensor for hydrogen
in molten metal based upon doped CaZrO
3
as the proton-con-
ducting electrolyte.
2
In terms of defect chemistry, the oxide is
typically doped by a trivalent cation at the Zr site resulting in
the formation of charge-compensating oxygen vacancies, which
are readily filled by hydroxy ions in the presence of water
vapour.
More recently, there has been renewed interest in acceptor-
doped BaZrO
3
whose combination of high proton conductivity
coupled with good chemical stability,
3–7
makes this material a
promising candidate for solid oxide fuel cell (SOFC) appli-
cations. Bohn and Schober
5
find that the proton mobility in
Y-doped BaZrO
3
is among the highest ever reported for a per-
ovskite-type proton conductor. The Y-doped BaZrO
3
system
has the potential to operate at lower temperatures than the con-
ventional SOFC electrolyte, and hence recent research has
involved attempts to optimise the materials’ properties. An
electrochemical reactor with a ceramic proton-conducting
membrane based on doped BaZrO
3
has also been used to study
the electrochemical promotion of catalysis.
6
Also in this field of conducting solids, the oxygen transport
properties of the LaGaO
3
-based perovskite have been widely
investigated,
8–18
owing to the higher oxygen ion conductivity
than the conventional Y/ZrO
2
electrolyte at lower temperatures.
The incorporation of cation dopants to form the system La
1-x
-
Sr
x
Ga
1-y
Mg
y
O
3-δ
(often termed LSGM) gives rise to the highly
mobile oxygen vacancies that are responsible for the observed
† Based on the presentation given at Dalton Discussion No. 7, 5–7th
July 2004, University of St Andrews, UK.
‡ Present address: Fachhochschule Gelsenkirchen, D45665 Reckling-
hausen, Germany.
ionic conductivity. More recently, the oxygen ion and proton
conduction properties of the related LaScO
3
material have been
investigated, particularly systems doped with alkaline-earth
dopants at either La or Sc sites.
19–21
For example, Lybye
and Bonanos
19
have investigated the La
0.9
Sr
0.1
Sc
0.9
Mg
0.1
O
3-δ
material and showed mixed conductivity at low oxygen partial
pressure; proton conduction was dominant at temperatures
below 700 °C while above 800 °C oxygen ion conduction
became increasingly dominant with temperature. More
recently, Kato et al.
21
have investigated the electrical conductiv-
ity of Al-doped La
1-x
Sr
x
ScO
3
as a potential SOFC material.
It has become increasingly clear that the investigation of
defect phenomena and atomistic diffusion mechanisms under-
pins the fundamental understanding of macroscopic behaviour.
However, there is often limited atomic-scale information on
complex ceramic oxides, such as lattice defects, dopant-site
selectivity and the extent of defect-dopant clustering. There is
also debate as to whether there is any significant interaction
between the dopant ion and the protonic defect leading to
possible proton “trapping”.
This study attempts to provide further insight into these
problems by using computer simulation techniques, which are
now well established tools in solid state chemistry. The reliabil-
ity of such an approach has been demonstrated by our simu-
lation studies of defects, ion transport and surface structures of
other perovskite oxides (e.g. LaMnO
3
, LaCoO
3
).
22–25
This paper
presents recent computational studies of topical oxygen ion
and proton-conducting perovskites based upon LaScO
3
and
BaZrO
3
, with direct comparison with related work on LaGaO
3
(ref. 23) and CaZrO
3
(ref. 22), respectively. Emphasis here is
placed on probing dopant site-selectivity, defect association
and oxygen ion migration, which have assisted in the further
understanding of these complex oxides on the atomic-scale.
2 Computational methods
Our description of the computational techniques will be brief
since comprehensive reviews are given elsewhere.
26–29
In this
paper, two main classes of technique have been employed in the
study of the perovskite materials: atomistic (static lattice) and
quantum mechanical (ab initio) methods.
DOI: 10.1039/ b402669C
This journal is © The Royal Society of Chemistry 2004 Dalton Trans. , 2004, 3061–3066 3061
Downloaded by University of Bath on 12 November 2012
Published on 23 July 2004 on http://pubs.rsc.org | doi:10.1039/B402669C
View Online / Journal Homepage / Table of Contents for this issue