Sie sind auf Seite 1von 7

View Article Online / Journal Homepage / Table of Contents for this issue

Journal of Dynamic Article Links < C


Materials Chemistry
Cite this: J. Mater. Chem., 2012, 22, 25388
www.rsc.org/materials PAPER
Structural and defect properties of the LaPO4 and LaP5O14-based proton
conductors
Satyajit Phadke,*a Juan C. Ninob and M. Saiful Islam*c
Published on 29 October 2012. Downloaded by CONRiCYT on 7/19/2018 6:14:19 AM.

Received 9th May 2012, Accepted 13th September 2012


DOI: 10.1039/c2jm32940a

Atomistic simulation techniques are used to perform a comparative study of intrinsic defects, dopant
incorporation and protonic groups in two lanthanum phosphate compounds, namely, the
orthophosphate (LaPO4) and the ultraphosphate (LaP5O14). The suitability of dopant incorporation
predicted from the dopant solution energies (with Ca and Sr the most favorable) is in excellent
agreement with trends in ionic conductivity from recent experimental investigations. The defect
chemistry of the phosphates related to protonic defects and oxygen vacancies created from extrinsic
doping is investigated. The results indicate favorable orientations for the protonic defect within the
structures. The binding energies for proton–dopant interactions indicate that defect association may
occur. In LaPO4 it was observed that the relaxed local atomic structure around an oxygen vacancy is
analogous to the formation of a P2O7 pyrophosphate anion.

1. Introduction Proton incorporation into these materials is accomplished by


acceptor doping on the cation site which leads to the formation
Proton-conducting oxides have attracted considerable attention of charge compensating oxygen ion vacancies. The vacancies
owing to the potential electrochemical applications in fuel cells, can be readily filled to form mobile protonic defects in the
gas sensors, and ceramic membranes. In particular, proton- presence of ambient water vapor. Previous investigations on the
conducting solid oxide fuel cells (SOFCs) or proton ceramic fuel defect chemistry of proton conducting phosphate compounds
cells (PCFCs) can operate within an intermediate temperature have been limited to orthophosphates (such as LaPO4) and
range (400–800  C).1–3 there is no reported study which discusses the defect chemistry
Several investigations (in the last decade or so) have focused of more complex phosphate compounds at the microscopic
on ceramic based materials such as perovskite oxides (AMO3) level.
and phosphates (APO4).3–18 Within phosphates, the experi- Using contemporary computational techniques a compara-
mental investigations have mainly focused on orthophosphates tive study of orthophosphate LaPO4 and ultraphosphate
such as LaPO4, Ba3Ce(PO4)3, CePO4, CsH2PO4, RbH2PO4, LaP5O14 has been performed in order to understand the defect
CsH5(PO4)2, and some pyrophosphate materials.4,19–26 In chemistry in complex phosphate compounds. These methods
general, phosphates tend to form compounds of higher are well suited to investigating the local defect structure in
complexity at higher phosphorus contents by corner sharing an materials at an atomic level which is crucial to develop a deeper
oxygen ion between adjacent PO4 tetrahedra. For example, understanding of proton incorporation and proton dynamics
within the La2O3–P2O5 binary phase system, three different within the material. In particular, intrinsic defect formation
phosphate compounds exist namely La2P4O13, La(PO3)3, and (such as Schottky and Frenkel), dopant incorporation, and
LaP5O14 (in increasing order of structural complexity) in addi- proton–dopant interactions in both LaPO4 and LaP5O14 have
tion to LaPO4 which is a well investigated proton conducting been investigated.
material.9,10,27–30 The conductivity of 1 mol% Sr doped LaPO4 is
about 106 S cm1 at 550  C while that of LaP5O14 is about
2  106 S cm1.31,32 2. Simulation methods
Comprehensive reviews of the atomistic simulation techniques
a
Department of Materials Science and Engineering, Massachusetts
are available elsewhere.33–35 In general, the simulation involves
Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA calculation of the potential energy of the system in terms of
02137, USA. E-mail: Phadke@mit.edu; Tel: +1 617 253 8468 atomic coordinates, with ion–ion interactions described as ionic
b
Materials Science and Engineering Department, University of Florida, 172 pairwise potentials. In accordance with the Born model, the
Rhines Hall, Gainesville, FL 32611, USA. E-mail: jnino@mse.ufl.edu; Fax:
+1 352 846 3355; Tel: +1 352 846 3787 interatomic forces are separated into the long-range Coulombic
c
Department of Chemistry, University of Bath, Bath BA2 7AY, UK. terms and short-range terms that correspond to electron
E-mail: m.s.islam@bath.ac.uk cloud overlap (Pauli repulsion) and attractive van der Waals

25388 | J. Mater. Chem., 2012, 22, 25388–25394 This journal is ª The Royal Society of Chemistry 2012
View Article Online

interactions. The short range interactions are described by the Table 2 Parameters for the O–H interaction
following Buckingham interatomic potential:36
Morse potential Dc (eV) 1)
a (A 
ro (A)
  Cij
Vij ðrÞ ¼ A exp  r=rij  6 (1) O/H 7.0525 2.1986 0.9485
r
where Aij, rij and Cij are potential parameters for each set of ion– Buckingham potential A (eV) 
r (A)  6)
C (eV A
ion interactions in the material. In addition to the above short
O/H 311.97 0.25 0
range interactions, three-body interaction terms were also used
to account for the angle dependent nature of the O–P–O and P–
O–P bonds in the system. The P–O–P interaction term is required
only for LaP5O14 to maintain the angle between the corner
where D, b, and ro were obtained from ab initio quantum
sharing PO4 tetrahedra in the structure. In LaPO4, the tetrahedra
Published on 29 October 2012. Downloaded by CONRiCYT on 7/19/2018 6:14:19 AM.

mechanical cluster calculations (developed by Saul et al.44), with


are isolated from each other and hence there are no direct P–O–P
a point charge representation of the surrounding lattice. These
interactions between adjacent tetrahedra. Such three body
values are summarized in Table 2. The dipole of the hydroxyl
interaction terms have been widely employed previously in the
species was distributed across both ions with O assigned 1.4263
successful modeling of phosphates, apatite-materials, silicates
and H assigned +0.4263 to give an overall charge of 1 for the
and aluminosilicates.37,38,40,41 The three body term takes the
protonic defect. This charge distribution reproduces the correct
form of a harmonic angle-bending potential about the central ion
dipole moment of the O–H group. To describe the interaction
(P or O):
between the hydroxyl group and the lattice ions, an additional
1 Buckingham potential was included which was taken from
V3-body ¼ kðq  qo Þ2 (2)
2 previous studies of water incorporation into silicates.45 We note
where k is the force constant and qo is the equilibrium bond that such simulation techniques have been applied successfully to
angle. Ionic polarizability is accounted for through the shell the atomic-scale study of a wide variety of proton-conducting
model developed by Dick and Overhauser.42 The interatomic oxides for SOFCs including BaMO3 perovskites.46–49 As indi-
potential and shell model parameters used in this study of LaPO4 cated in the Introduction, the present work has focused on
and LaP5O14 are listed in Table 1. The parameters for La–O and structural, defect and dopant properties of these La-phosphate
O–O interactions have been transferred from previously pub- materials, which are not fully characterized. Nevertheless, this
lished work; those for P–O were derived for this study employing study is a valuable precursor to future work on the pathway of
empirical fitting procedures using observed structural data.36 proton migration which will require the use of ab initio simula-
The existence of a charged defect in a lattice can cause tion methods.
significant displacement or relaxation of the positions of the
surrounding lattice ions. Modeling of such effects is carried out 3. Results and discussion
using the two region Mott–Littleton approach, which partitions
the crystal into two spherical concentric regions.43 The inner 3.1 Structural modeling
region (consisting of >500 atoms) surrounding the defect is
The basic difference between the ultraphosphate (Fig. 1) and the
relaxed explicitly. However, the ions in the outer region are
orthophosphate (Fig. 2) lies in the condensed versus isolated
treated by more approximate quasi-continuum methods since the
nature of the phosphate anion in these materials. The structure of
defect forces are much weaker in this region.
LaPO4 consists of isolated PO4 tetrahedra that are corner sharing
For the protonic defect, the O–H interaction was modeled
only with LaO9 polyhedra as shown in Fig. 2. However, as we
using the following Morse potential:
move towards the phosphate rich compounds within the La2O3–
P2O5 phase diagram the materials become structurally more
V(r) ¼ D[1  exp(b(r  ro))]2 (3)
complex due to the condensation of the phosphate anion. In
LaP5O14 (which is structurally the most complex among the
Table 1 Interatomic potentials and shell model parameters for LaPO4 lanthanum phosphates), the anion consists of a three dimen-
and LaP5O14 sional network of corner sharing PO4 tetrahedra as shown in
Short-range Fig. 1(a). Although the basic building blocks comprising the
phosphate anion are still PO4 tetrahedra, they are arranged
Interaction A (eV) 
r (A)  6)
C (eV A Ref. forming a regular ribbon-like morphology. A small segment of
the anion is shown in Fig. 1(a) and shows the manner in which
La /O
3+ 2
4579.23 0.3044 0.000 37 and 38
P5+/O2 1250.00 0.3240 0.000 — the tetrahedra are oriented.
O2/O2 22 764.30 0.1490 27.879 37 and 38 In LaP5O14 the P–O bonds can be classified as either bridging
(P–Ob) or terminal (P–Ot) oxygen bonds. If the oxygen ion is
Shell model
located at the vertex of two corner sharing PO4 tetrahedra then
Species Y (e) 2)
K (eV A Ref. the P–O bond is termed as a bridging oxygen bond. However, if
the oxygen ion is located at the vertex of a PO4 tetrahedron
La3+
0.25 145.0 37 and 38 which is corner sharing that particular oxygen ion with an LaO8
P5+ 5.00 99 999 39
O2 2.96 27.879 39 polyhedron, then the P–O bond is termed as a terminal oxygen
bond. In all the five different tetrahedra in LaP5O14, the

This journal is ª The Royal Society of Chemistry 2012 J. Mater. Chem., 2012, 22, 25388–25394 | 25389
View Article Online

Table 3 Experimental (Exp.) and calculated (Calc.) lattice parameters



for LaPO4 and LaP5O14. All values are in A

LaPO4 LaP5O14

Exp. Calc. D Exp. Calc. D

a 6.8313 6.8302 0.02 8.8206 8.8809 0.06


b 7.0705 6.9922 0.08 9.1196 9.0823 0.03
c 6.5034 6.6258 0.12 13.1714 12.9774 0.19
b 103.27 104.42 1.15 90.66 90.20 0.46
Published on 29 October 2012. Downloaded by CONRiCYT on 7/19/2018 6:14:19 AM.

Table 4 Experimental (Exp.) and calculated (Calc.) average bond


lengths for LaPO4 and LaP5O14


Bond length (A)
Difference
System Bond type Exp. Calc. 
(A)

LaPO4 P–O 1.539 1.564 0.03


La–O 2.579 2.621 0.04
LaP5O14 P–Ot 1.474 1.519 0.02
P–Ob 1.593 1.615 0.02
La–O 2.423 2.503 0.08

In LaPO4 all the four P–O bonds in the PO4 tetrahedron are
equivalent and are of the terminal type due to the isolated
arrangement of the tetrahedra. From the experimental values
Fig. 1 (a) View of LaP5O14 along the [001] direction. La3+ ions are
shown in orange (large spheres) and PO4 tetrahedra are shown in
shown in Table 4 it can be seen that the average P–O bond length
green. (b) Segment of ribbon composed of PO4 tetrahedra as viewed in LaPO4 has an intermediate value between the terminal and
along the [010] direction. (O is the origin, A is the x-axis and B is the bridging P–O bond lengths in LaP5O14. The calculated structure
y-axis). of LaPO4 is in excellent agreement with respect to the bond
lengths and the difference between the calculated and experi-
mental values is small (0.03 A). The other difference in the
structures of LaPO4 and LaP5O14 is that in the former the
structure is composed of LaO9 polyhedra while the latter has
LaO8 polyhedra. The average experimental bond length for
La–O is longer for LaPO4 and matches well with the calculated
values (Table 4).

3.2 Intrinsic defects


The energies of isolated point defects such as vacancies and
interstitials were calculated for all the anions and cations in the
LaPO4 and LaP5O14 structures. The calculated energies of the
isolated defects were then combined according to the defect
Fig. 2 LaPO4 structure viewed along the c-axis. La3+ ions are shown in
equations to evaluate the energy of formation of Frenkel and
orange (large spheres) and PO4 tetrahedra shown in green. (O is the
Schottky defects that are summarized in Table 5. To evaluate the
origin, A is the x-axis and B is the y-axis.)
partial Schottky defect energies (La-Schottky type and P-
Schottky type) the lattice energy for La2O3 and P2O5 was
terminal P–O bonds are considerably shorter than the bridging obtained by modeling their respective crystal structure using
P–O bonds. the same interatomic potentials that was employed to model the
The experimentally observed crystal structures of LaPO4 and lanthanum phosphates. The results in Table 5 indicate that the
LaP5O14 were reproduced prior to carrying out the defect energies of formation of Frenkel and Schottky defects are very
calculations and the calculated lattice parameters are compared high suggesting that formation of such defects is highly unlikely
with the experimental values in Table 3. The differences in the in the temperature range of stability of the materials.
observed and calculated lattice parameters are small indicating The high defect formation energies can also be taken as an
that the potential model accurately represents these complex indication of the structural integrity of the PO4 tetrahedra in the
structures, thus supporting the validity of the potentials used for phosphate materials. For example, the concentration of intrinsic
the subsequent defect calculations. Also, the differences in the oxygen ion vacancies can be expected to be almost negligible.
P–O terminal and bridging bond lengths in LaP5O14 are correctly This is consistent with experimental defect models in which there
reproduced (listed in Table 4). is negligible ionic conductivity in the undoped materials. Thus

25390 | J. Mater. Chem., 2012, 22, 25388–25394 This journal is ª The Royal Society of Chemistry 2012
View Article Online

Table 5 Energies of intrinsic defect formation in LaPO4 and LaP5O14. All values are shown in eV (S – Schottky and F – Frenkel)

Type Defect equilibrium LaPO4 LaP5O14


__
LaXLa + PP + 4OO / VLa + VP + 4VO + LaPO4
X X /// /////
(S) 46.09 114.46
__
LaXLa + 5PP + 14OO / VLa + 5VP + 14VO + LaP5O14
X X /// /////
_____
P (F) PX
P / V /////
P + P i 41.33 19.76
___
LaXLa / Lai + VLa
///
La (F) 17.06 14.69
__
O (F) OO / VO + Oi
X //
12.74 7.31
__
2LaX La + 3OO / 2VLa + 3VO + La2O3
X ///
La (S) 29.94 40.95
__
P (S) 2PXP + 5O X
O / 2V /////
P + 5V O + P2O5 74.51 34.23
Published on 29 October 2012. Downloaded by CONRiCYT on 7/19/2018 6:14:19 AM.

Table 6 Inter-ionic distances in the [P2O7]4 group compared with inter- According to Amezawa et al. 1 mol% alkali earth metal doped
ionic distances in [PO4]3 LaPO4 shows the following trend in decreasing order of
[P2O7]4 [PO4]3 [HPO4]2
conductivity Sr > Ca > Ba > Mg which is in agreement with the
dopant solution energy obtained for the various dopants in this
Distance Distance Distance study.31 We recognise, however, that the correlation between
Ion–ion 
(A) Ion–ion 
(A) Ion–ion 
(A) dopant solution energetics and the proton conductivity may not
P1–P2 3.175 P–Padj 4.187 P–Padj 4.179 be straightforward since the low solubility or small amount of the
P1–Ob 1.719 P–O1 1.574 P–Oproton 1.882 dopants in the structure can simply relate to the low proton
P2–Ob 1.719 P–O2 1.576 P–O2 1.558 uptake. Another key factor is possible dopant trapping effects
P1–O2 1.550 P–O3 1.585 P–O3 1.542 which are addressed below.
P1–O3 1.564 P–O4 1.605 P–O4 1.541
P1–O4 1.552 Acceptor doping on the cation site produces oxygen vacancies,
P2–O2 1.547 which can then lead to the formation of protonic defects by the
P2–O3 1.553 dissociative absorption of ambient water as represented by the
P2–O4 1.572
following equation:
__ _
H2O + VO O / 2OHO
+ OX (5)
the extrinsic defects formed upon acceptor doping are likely to
dominate the defect chemistry of these phosphate materials, and Examination of the local relaxed structure around the oxygen
are considered in the next section. vacancy in LaPO4 revealed that the formation of an oxygen
vacancy is analogous to the generation of a [P2O7] pyrophos-
phate group in the structure. This is accomplished by corner
3.3 Dopant incorporation and vacancy defects sharing an oxygen ion between two neighboring PO4 tetrahedra,
and leads to a significant distortion of the tetrahedra involved
It is known that acceptor dopants (Ca and Sr) are used to arising from the stretching of the bridging P–Ob bonds
promote the observed proton incorporation and conduction in  as illustrated in Fig. 4(b). For comparison, the length
(1.718 A)
these phosphate materials. Although the prediction of the solu- of the remaining terminal P–Ot bonds of the pyrophosphate
bility limit of a particular dopant is a much more demanding group is listed in Table 6 where it can be seen that these bonds are
task, the relative energies of dopant incorporation provide useful much shorter (1.55 A).  In undoped LaPO4 the P–Ot bond
information about trends in dopant solubility between different 
lengths (1.574–1.605 A) are also quite similar to the P–Ot bond
species. Here, the energetics of divalent dopant (D ¼ Mg2+, Ca2+, lengths in the pyrophosphate as shown in Fig. 4(a).
Sr2+, Ba2+) substitution on the La3+ site were investigated. The
defect equation for dopant incorporation for both lanthanum
phosphate materials can be denoted in the Kr€ oger–Vink notation
as shown below:

3OXO þ 2LaXLa ! 2D=La þ 2OXO þ V$$


2DO
O þ La2 O3 (4)

By combining the appropriate lattice energy and defect energy


terms the dopant solution reactions were then evaluated. In eqn
(4), ‘DO’ is the divalent dopant oxide substituting into the lattice
structure. The interatomic potentials for the dopant cations are
obtained from the respective metal oxides.50
The resulting dopant incorporation (or solution) energies for
LaPO4 and LaP5O14 are shown in Fig. 3 plotted as a function of
the dopant ionic radius. Of these four dopant ions, Ca2+ and Sr2+
are the most favorable while Ba2+ and Mg2+ seem to be unfa-
vorable. This result correlates well with experimental studies
reported by Norby and Christiansen who found proton
conductivity in 5 mol% Ca and 5 mol% Sr doped LaPO4.9 Fig. 3 Calculated solution energies as a function of dopant ionic radius.

This journal is ª The Royal Society of Chemistry 2012 J. Mater. Chem., 2012, 22, 25388–25394 | 25391
Published on 29 October 2012. Downloaded by CONRiCYT on 7/19/2018 6:14:19 AM. View Article Online

Fig. 4 (a) Orthophosphate group [PO4]3 and (b) pyrophosphate group [P2O7]4 in LaPO4 showing bridging oxygen ion (Ob) and P–Ob bond distances.
(c) Protonated orthophosphate group [HPO4]3 in LaPO4 also showing the P–O bond lengths.

The formation of a pyrophosphate group in LaPO4 upon the protonated orthophosphate group indicates that there is
introduction of an oxygen vacancy is in good agreement with the significant distortion of the tetrahedron due to the presence of an
experimental observation by Amezawa et al. using MAS-NMR additional proton. The presence of the proton on the oxygen ion
studies.10 Given the similarity of the double-tetrahedral units we leads to stretching of the P–O bond (1.882 A)  as shown in
find here to those which form in the La1xBa1+xGaO4x/2 ionic Fig. 4(c), whereas the other P–O bonds are much shorter as
conductor we conjecture that oxide-ion transport in LaPO4 may summarized in Table 6.
also involve the breaking and reforming of the [P2O7] units in a Such O–H calculations were performed for all the 14 crystal-
cooperative ‘‘cogwheel-type’’ process, as in the gallate.37 Further lographically different oxygen ions in the structure of LaP5O14.
examination of the conduction process, for example, by molec- In the relaxed structure an equilibrium O–H distance of 0.98 A 
ular dynamics methods, while beyond the scope of the present was obtained for all the calculations. However, it was found that
study, needs to be undertaken. in the relaxed structure the O–H bond was directed in such a way
that the equilibrium position of the proton was located in either
of the two hollow channels located either along the a-axis or the
3.4 O–H configuration and proton–dopant interactions c-axis. These hollow channels exist in the ultraphosphate due to
Atomistic simulations were used to probe the local environment the open framework structure of the material. The equilibrium
around a proton to determine the most favorable configuration positions for the protons bonded to oxygen ions O1 and O14 are
within the structure. For the doped LaPO4 system, upon addition shown in Fig. 5(a) and (b) respectively. Along both crystallo-
of a water molecule to the pyrophosphate group the oxygen graphic directions it can be seen that the proton is located within
vacancy is filled by a hydroxyl ion and it leads to the formation of the hollow channels (shown as dotted region in Fig. 5). The
two protonated orthophosphate groups as indicated by the existence of these channels may provide a low energy pathway
following equation proposed by Amezawa et al.10 for proton conduction resulting in enhanced conductivity in the
ultraphosphate compared to the orthophosphate as experimen-
[P2O7]4 + H2O / 2[HPO4]2 (6) tally observed previously.41
If there is dopant–OH association between the two oppositely
The above equation is analogous to eqn (5) which is relevant to charged defects it can adversely affect the proton conductivity in
doped oxide materials. Analysis of the relaxed structure of a the material and hence it is of interest to calculate the energetics of

Fig. 5 View of LaP5O14 along (a) c-axis and (b) a-axis showing the location of hollow channels (dotted) in the structure. In the relaxed structure the
O–H bond is directed towards these channels and the proton preferentially tends to lie in these channels.

25392 | J. Mater. Chem., 2012, 22, 25388–25394 This journal is ª The Royal Society of Chemistry 2012
View Article Online

Table 7 Proton–dopant interactions in LaPO4 and LaP5O14 1. High formation energies were found for all the types of
intrinsic Frenkel and Schottky defects suggesting that the degree
System D 2+
Ebinding (eV) 
D–H (A)
of intrinsic disorder in both LaPO4 and LaP5O14 is negligible,
LaPO4 Mg 0.75 2.94 and confirms the structural integrity of the PO4 tetrahedra in the
Ca 0.59 3.04 undoped structures. Hence, extrinsic defects formed upon
Sr 0.57 3.19 acceptor doping are likely to dominate the defect chemistry in
Ba 0.56 3.37
LaP5O14 Mg 0.44 2.50 these phosphate materials, as observed. The trend for dopant
Ca 0.50 2.69 incorporation energetics (with oxygen vacancy compensation)
Sr 0.39 2.82 was as follows (Sr > Ca > Ba > Mg), which is in agreement with
Ba 0.45 3.30 experimental conductivity results for doped LaPO4 where Ca and
Sr are the most commonly used dopants.
Published on 29 October 2012. Downloaded by CONRiCYT on 7/19/2018 6:14:19 AM.

such an interaction. The cluster binding energy for such an 2. Examination of the local relaxed structure around the oxygen
interaction can be calculated by considering the energy of the vacancy in LaPO4 revealed the creation of a P2O7 pyrophosphate
cluster (Ecluster) and comparing it with the sum of the energies of group. This is accomplished by a significant relaxation of two
isolated individual defects as shown in the general equation below: neighboring PO4 tetrahedra so as to share an oxygen ion, and is in
X  excellent agreement with previous experimental NMR observa-
Ebind ¼ Ecluster  Eisolated-defect (7) tion of the pyrophosphate anion in doped LaPO4.
3. The local environment around protonic defects in LaP5O14
A negative Ebind in such a case would indicate that dopant–OH was examined to predict the energetically favorable O–H
association is energetically favorable. In the case of the configuration. The calculations suggest that irrespective of the 14
lanthanum phosphate compounds the relevant equation would crystallographically different oxygen sites, the O–H bond was
be as follows: directed within either of the two hollow channels along the a- or
c-axes. For all the dopants examined, dopant–OH association
Ebind ¼ E(OHO_ /D/La)  [E(OHO_ ) + E(D/La)] (8) was found to be favorable in both materials. The weakest asso-
ciation is found for Sr in LaP5O14 and the strongest association
The calculated binding energies for the divalent dopant cations
for Mg in LaPO4 indicating that local proton trapping effects
for both phosphates are listed in Table 7. It is found that for all
and proton mobility are sensitive to the type of acceptor dopant.
the dopants the values are negative which implies that such defect
association is favorable. In LaPO4, the highest (strongest)
binding energy calculated was for the Mg dopant (0.75 eV) Notes and references
while for all the other dopants the values were in the range of 1 S. M. Haile, D. A. Boysen, C. R. I. Chisholm and R. B. Merle, Nature,
about 0.55 eV. The lowest (weakest) energy calculated in 2001, 410, 910–913.
LaP5O14 was for the Sr dopant (0.39 eV) while Ca dopant 2 S. Ishimaru, M. Togawa, R. Ikeda, T. Shimizu, E. Shinohara and
Y. Umemura, Bull. Chem. Soc. Jpn., 2006, 79, 656–659.
exhibited the highest binding energy (0.50 eV). The calculated
3 L. Malavasi, C. A. J. Fisher and M. S. Islam, Chem. Soc. Rev., 2010,
dopant ion–proton distances (D–H) are also listed in Table 7. It 39, 4370–4387.
can be seen that the D–H distance increases with increase of the 4 G. J. Zhang, E. D. Wachsman, J. C. Nino, S. J. Song and
ionic radius of the dopant cation (RBa > RSr > RCa > RMg). J. M. Rhodes, J. Electrochem. Soc., 2007, 154, H566–H571.
5 S. J. Song, E. D. Wachsman, J. Rhodes, S. E. Dorris and
The high binding energy for the Ca dopant (in LaP5O14) and U. Balachandran, Solid State Ionics, 2004, 167, 99–105.
Mg dopant (in LaPO4) suggests a greater change in the 6 E. Fabbri, T. K. Oh, S. Licoccia, E. Traversa and E. D. Wachsman,
conduction activation energy with an increase of dopant J. Electrochem. Soc., 2009, 156, B38–B45.
concentration in these materials. Indeed, high binding energies 7 K. D. Kreuer, Solid State Ionics, 1999, 125, 285–302.
8 K. D. Kreuer, W. Munch, M. Ise, T. He, A. Fuchs, U. Traub and
indicate strong proton trapping effects, which would be detri- J. Maier, Ber. Bunsen-Ges., 1997, 101, 1344–1350.
mental to proton conductivity. As already indicated, another 9 T. Norby and N. Christiansen, Solid State Ionics, 1995, 77, 240–243.
factor to be considered is the low Mg dopant incorporation into 10 K. Amezawa, H. Maekawa, Y. Tomii and N. Yamamoto, Solid State
Ionics, 2001, 145, 233–240.
both structures, which would affect the degree of proton uptake. 11 H. Iwahara, T. Esaka, H. Uchida and N. Maeda, Solid State Ionics,
We note that a recent DFT study of Ba-doped LaPO4 has found 1981, 3–4, 359–363.
that Ba-doping stabilizes a proton at a distance of 2.7 angstrom 12 H. Uchida, H. Yoshikawa, T. Esaka, S. Ohtsu and H. Iwahara, Solid
from the dopant by up to 0.2 eV relative to positions far from the State Ionics, 1989, 36, 89–95.
13 R. C. T. Slade and N. Singh, Solid State Ionics, 1993, 61, 111–114.
dopant.29,30 14 A. S. Nowick, S. Q. Fu, W. K. Lee and L. A. Boatner, Solid State
Ionics, 1988, 26, 168.
15 A. S. Nowick and A. V. Vaysleyb, Solid State Ionics, 1997, 97, 17–26.
4. Conclusions 16 T. Scherban, W. K. Lee and A. S. Nowick, Solid State Ionics, 1988,
28, 585–588.
A comparative study of two lanthanum phosphate proton 17 T. Scherban and A. S. Nowick, Solid State Ionics, 1989, 35, 189–194.
18 T. Scherban and A. S. Nowick, Solid State Ionics, 1992, 53–56, 1004–
conductors, namely the orthophosphate (LaPO4) and the ultra- 1008.
phosphate (LaP5O14), was performed using atomistic simulation 19 E. G. del Moral, D. P. Fagg, E. Chinarro, J. C. C. Abrantes,
techniques. The defect chemistry of the phosphates was investi- J. R. Jurado and G. C. Mather, Ceram. Int., 2009, 35, 1481–1486.
20 R. H. Chen, C. C. Yen, C. S. Shern and T. Fukami, Solid State Ionics,
gated along with the examination of the local relaxed structures
2006, 177, 2857–2864.
around protonic defects and oxygen vacancies. The main 21 H. Muroyama, T. Matsui, R. Kikuchi and K. Eguchi, J. Electrochem.
conclusions are as follows. Soc., 2008, 155, B958–B962.

This journal is ª The Royal Society of Chemistry 2012 J. Mater. Chem., 2012, 22, 25388–25394 | 25393
View Article Online

22 T. Matsui, T. Kukino, R. Kikuchi and K. Eguchi, Electrochem. Solid- 36 A. K. Cheetham and P. Day, Solid State Chemistry: Techniques,
State Lett., 2005, 8, A256–A258. Clarendon Press, Oxford, 1987.
23 T. Kukino, R. Kikuchi, T. Takeguchi, T. Matsui and K. Eguchi, Solid 37 P. M. Panchmatia, A. Orera, E. Kendrick, J. V. Hanna, M. E. Smith,
State Ionics, 2005, 176, 1845–1848. P. R. Slater and M. S. Islam, J. Mater. Chem., 2010, 20, 2766–2772.
24 O. Paschos, J. Kunze, U. Stimming and F. Maglia, J. Phys.: Condens. 38 P. M. Panchmatia, A. Orera, G. J. Rees, M. E. Smith, J. V. Hanna,
Matter, 2011, 23. P. R. Slater and M. S. Islam, Angew. Chem., Int. Ed., 2011, 50,
25 V. Nalini, R. Haugsrud and T. Norby, Solid State Ionics, 2010, 181, 9328–9333.
1264–1269. 39 C. A. J. Fisher, V. M. H. Prieto and M. S. Islam, Chem. Mater., 2008,
26 I. Hammas, K. Horchani-Naifer and M. Ferid, J. Rare Earths, 2010, 20, 5907–5915.
28, 321–328. 40 D. Bougeard and K. S. Smirnov, Phys. Chem. Chem. Phys., 2007, 9,
27 N. Hatada, Y. Nose, A. Kuramitsu and T. Uda, J. Mater. Chem., 226–245.
2011, 21, 8781–8786. 41 G. R. Gardiner and M. S. Islam, Chem. Mater., 2009, 22, 1242–1248.
28 M. T. Colomer, J. A. Diaz-Guillen and A. F. Fuentes, J. Am. Ceram. 42 B. G. Dick and A. W. Overhauser, Phys. Rev., 1958, 112, 90–103.
Soc., 2010, 93, 393–398. 43 M. F. Mott and M. J. Littleton, Trans. Faraday Soc., 1938, 34, 0485–0499.
Published on 29 October 2012. Downloaded by CONRiCYT on 7/19/2018 6:14:19 AM.

29 R. Yu and L. C. De Jonghe, J. Phys. Chem. C, 2007, 111, 11003– 44 P. Saul, C. R. A. Catlow and J. Kendrick, Philos. Mag. B, 1985, 51,
11007. 107–117.
30 N. Adelstein, J. B. Neaton, M. Asta and L. C. De Jonghe, J. Mater. 45 K. P. Schroder, J. Sauer, M. Leslie, C. R. A. Catlow and
Chem., 2012, 22, 3758–3763. J. M. Thomas, Chem. Phys. Lett., 1992, 188, 320–325.
31 K. Amezawa, Y. Tomii and N. Yamamoto, Solid State Ionics, 2005, 46 S. J. Stokes and M. S. Islam, J. Mater. Chem., 2010, 20, 6258–6264.
176, 135–141. 47 M. Cherry, M. S. Islam, J. D. Gale and C. R. A. Catlow, Solid State
32 S. R. Phadke and J. C. Nino, J. Am. Ceram. Soc., 2011, 94, 1817– Ionics, 1995, 77, 207–209.
1823. 48 G. C. Mather, C. A. J. Fisher and M. S. Islam, Chem. Mater., 2010,
33 Computer Modeling in Inorganic Crystallography, ed. C. R. A. Catlow, 22, 5912–5917.
Academic Press, San Diego, CA, 1997. 49 E. Kendrick, J. Kendrick, K. S. Knight, M. S. Islam and P. R. Slater,
34 J. D. Gale and A. L. Rohl, Mol. Simul., 2003, 29, 291–341. Nat. Mater., 2007, 6, 871–875.
35 J. P. Perdew, K. Burke and Y. Wang, Phys. Rev. B: Condens. Matter, 50 G. V. Lewis and C. R. A. Catlow, J. Phys. C: Solid State Phys., 1985,
1996, 54, 16533. 18, 1149–1161.

25394 | J. Mater. Chem., 2012, 22, 25388–25394 This journal is ª The Royal Society of Chemistry 2012

Das könnte Ihnen auch gefallen