Reaction mechanisms of Li0.30La0.57TiO3 powder with ambient air: H+/Li+ exchange with water and Li2CO3 formation

Article abstract of DOI:10.1039/b924684c

The proton/lithium exchange property of the lithium lanthanum titanate Li_0.30La_0.57TiO₃ (named LLTO) is shown to occur at room temperature under ambient air. The ¹H and ⁷Li MAS NMR, TGA analysis and IR spectroscopy techniques are used to probe reaction mechanisms. XRPD analysis gives evidence of the topotactic character of this exchange reaction. As for exchange in aqueous solution, it is shown that Li _0.30La_0.57TiO₃ is able to dissociate water on the grain surface and then to exchange H⁺ for Li⁺ into the perovskite structure. Lithium hydroxide is then formed on the grain surface and afterwards reacts with CO₂ contained in air to form Li ₂CO₃. It is shown that this mechanism is reversible. When the aged sample (aging in air for 5 months at room temperature) is annealed at 400°C for two hours, the initial LLTO sample is totally recovered, a mass loss is observed and the carbonate signal in IR spectra disappears, demonstrating the reversibility of the carbonation reaction process.

Full text of DOI:10.1039/b924684c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
+
+
Reaction mechanisms of Li La TiO powder with ambient air: H /Li
0
.30
0.57
3
exchange with water and Li CO formation
2
3
a
a
a
a
a
Anthony Boulant, Jean Francois Bardeau, Alain Jouanneaux, Jo e¨ l Emery, Jean-Yves Buzare and
Odile Bohnke*
b
Received 24th November 2009, Accepted 15th February 2010
First published as an Advance Article on the web 16th March 2010
DOI: 10.1039/b924684c
The proton/lithium exchange property of the lithium lanthanum titanate Li0.30La0.57TiO (named
3
1
7
LLTO) is shown to occur at room temperature under ambient air. The H and Li MAS NMR, TGA
analysis and IR spectroscopy techniques are used to probe reaction mechanisms. XRPD analysis gives
evidence of the topotactic character of this exchange reaction. As for exchange in aqueous solution, it is
+
shown that Li0.30La0.57TiO
3
is able to dissociate water on the grain surface and then to exchange H for
+
Li into the perovskite structure. Lithium hydroxide is then formed on the grain surface and afterwards
reacts with CO
2
contained in air to form Li
2
CO
3
. It is shown that this mechanism is reversible. When
◦
the aged sample (aging in air for 5 months at room temperature) is annealed at 400 C for two hours,
the initial LLTO sample is totally recovered, a mass loss is observed and the carbonate signal in IR
spectra disappears, demonstrating the reversibility of the carbonation reaction process.
Introduction
show CO
2
sorption. All these ceramics react with CO
2
at high
◦
◦
temperature, i.e. above 450 C for zirconates, above 650 C for
silicates and above 850 C for titanates.
The effect of CO on cathode materials used for secondary
lithium batteries has been also extensively investigated. Indeed, a
chemical instability of the cathodes has been shown during storage
in air. In particular, the reaction of CO
generally leads to a reduction of the batteries performances. In
the case of Li1+zNi1-x-yCo
The titanium-based perovskite phases (Li3xLa2/3-x
hereafter named LLTO, have attracted much attention since the
papers of Belous et al. in 1987 and Inaguma et al. in 1993
who reported a bulk lithium conductivity as high as 10 S cm
ꢀ
1/3-2x)TiO
3
,
◦
2
1
2
-
3
-1
at room temperature, for x = 0.10. This is one of the highest
2
with cathode materials
ionic conductivity in crystalline lithium-ion conductors reported
3
at the time. Ion migration occurs through the conduction paths
x
M
y
O
2
(M = Al, Mn), it has been shown
+
formed by Li ions and vacancies ꢀ, present in the A-sites of the
that Li
2
O is present on the sample surface and is able to react
ABO
3
perovskite structure. Structural data have been published by
17
with CO
2
in air to produce Li
2
CO
3
.
But carbon dioxide can also
4
5
Fourquet et al. and Robertson et al. These authors agree to say
that a pure solid solution exists in the composition range 0.04 <
produce lithium deintercalation and lithium carbonate formation,
as in LiNi0.81Co0.16Al0.03
temperature with a conversion of 8% after 500 h in air at 55%
18,19
O
2
.
This reaction can occur at room
x < 0.14. These phases (Li3xLa2/3-x
ꢀ
1/3-2x)TiO
3
can be exchanged
6
7
in nitric acid, but also in water with the same efficiency. The
RH. This conversion can reach up to 70% under atmospheric CO
2
ionic exchange reaction in water performed on Li0.3La0.57TiO
leads to the formation of the protonated phase H0.25Li0.05La0.57TiO
demonstrating the affinity of this material for water.
3
◦
at 675 C. M e´ n e´ trier et al. have suggested that the deintercalation
3
,
of lithium ions is accompanied with an intercalation of protons,
proving that an ionic exchange reaction occurs upon hydration of
Li(Ni1-y-z)Co
It is known from the literature that lithium containing materials
can be used for CO retention. In this context, it has been shown
that lithium hydroxide, which is a hygroscopic compound, can
be used in its monohydrated form for retention of CO with
can also occur
on lithium oxide, efficiency as high as 86 mol% conversion of
20
y
Al
z
2
O sample.
2
In a recent paper, Simon et al. showed that lithium titanate,
21
Li
4
Ti
5
O12, can be exchanged in acid solution. Electrochemical
2
+
tests indicate that Li ions are intercalated then deintercalated
from the material and a sample exposed to air more than 6
months also exhibits potential curves which characterize a Li /H
exchange. After annealing at 900 C, the received commercial
sample feature was recovered. A suspected behavior with air was
suggested and care in storage was therefore advised.
8
,9
high efficiency. The chemical sorption of CO
2
+
+
Li
2
O to Li
2
CO
3
is reported by Mosqueda et al. after heating
◦
◦
10
at 600 C for 2 h. Some lithium ceramics, such as lithium
zirconates,
11-14
15
16
lithium silicates and lithium titanates can also
In this study, it will be shown that an ionic exchange occurs
a
Laboratoire de Physique de l’Etat Condens e´ (UMR 6087 CNRS), Institut
de Recherche en Ing e´ nierie Mol e´ culaire et Mat e´ riaux Fonctionnels (FR 2575
CNRS), Universit e´ du Maine, Avenue O. Messiaen, 72085, Le Mans Cedex
in ambient air at room temperature in Li0.3La0.57TiO sample. A
3
mechanism is proposed to explain the aging reaction process. An
intermediate protonated phase reacts with carbon dioxide from air
leading to the formation of lithium carbonate Li CO . To highlight
9
, France
b
Laboratoire des Oxydes et Fluorures (UMR 6010 CNRS), Institut de
Recherche en Ing e´ nierie Mol e´ culaire et Mat e´ riaux Fonctionnels (FR 2575
CNRS), Universit e´ du Maine, Avenue O. Messiaen, 72085, Le Mans Cedex
2
3
the aging reaction process, X-Ray Powder Diffraction (XRPD)
analysis was used to show that a topotactic exchange reaction
occurs. Differential thermal analysis (DTA) was performed and
9
, France. E-mail: odile.bohnke@univ-lemans.fr; Fax: 33-2-43 83 35 06;
Tel: 33-2-43 83 33 54
3
968 | Dalton Trans., 2010, 39, 3968–3975
This journal is © The Royal Society of Chemistry 2010

View Article Online
1
Thermogravimetric Analysis (TGA) measurements enabled the
mass loss to be quantified. Infra Red (IR) spectroscopy revealed
the presence of carbonate whereas the properties of the hydrogen
and lithium nucleus are illustrated using H and Li Magic Angle
Spinning Nuclear Magnetic Resonance (MAS NMR).
and only H spectra recorded in synchronized mode are shown.
In the non synchronized mode, the spectral width is 100 kHz
and the spinning frequency is 10 kHz. In this case, the spectrum
contains both isotropic lines and their spinning side bands. In the
synchronized mode, the spectral width equals the rotor spinning
frequency (10 kHz) and the spectrum exhibits the isotropic lines
1
7
7
Experimental
only. Li spectrum of commercial Li CO (Aldrich, 99.997%) was
also recorded. The DMFIT software was used to reconstruct the
spectra and obtain the line widths, the peak positions (in Hz or
2
3
Li0.3La0.57TiO samples were analyzed after synthesis (the corre-
3
sponding sample is named “AS” in the following), and after an
29
◦
ppm) and the percentage of each contribution.
aging process in air at room temperature (~ 20 C) for 5 months
(
4
named “AGED”). The aged sample was further annealed at
00 C in air for two hours to form the sample named “OA”.
◦
Results
Preparation of LLTO powder
The result of the Rietveld refinement for the AS sample is
presented in Fig. 1a. All the peaks in the diffraction pattern
can be unambiguously indexed in a tetragonal unit cell (space
The lithium lanthanum titanate, Li0.30La0.57TiO , was prepared by
3
22
the Pechini method. The synthesis procedure was described in
ref. 23. The titanate powder was obtained after a thermal treatment
4
group P4/mmm). As suggested by Fourquet et al., the reflections
◦
were divided into two sets of lines and the “two phases” option
of FULLPROF software was used in the refinement. One set is
formed of (hkl) reflections with l = 2n while superstructure peaks
with l = 2n+1 are included in the second set. The same structural
model is applied to both “phases”. The lanthanum content was
refined between the two possible crystallographic sites (1a) (0,0,0)
and (1b) (0,0,1/2) but constrained to give a total number of
lanthanum atoms consistent with the x = 0.10 value.
of the precursor powder at 900 C in air. A pellet of this powder
◦
was sintered at 1000 C in air for two hours. After sintering, the
pellet was ground in an agate mortar to obtain the powder studied
in this work. All samples (AS, AGED, OA) came from the same
synthesis. The aging of the samples has been performed in air at
room temperature under a relative humidity close to 50%.
Characterization of the powder
XRPD, TGA, IR spectroscopy and MAS NMR techniques were
used to characterize the samples. XRPD patterns were collected
at room temperature with a Philips X’Pert PRO diffractometer
◦
(
Cu-Ka radiation) in the 2q range from 8 to 134.2 with an
◦
increment step of 0.0167 and a data acquisition of 7 h 33 min.
2
4
XRPD patterns were analyzed by the Rietveld method using the
25
FULLPROF software.
IR spectra were obtained with a FTIR spectrometer (Nicolet
AVATAR 370 DTGS) in the spectral range from 400 to 4000 cm
with a resolution of 2 cm . The sample consisted of pellets
prepared by pressing a mechanically homogenized mixture of
LLTO powder with dehydrated KBr (1 wt%).
-
1
-
1
TGA of LLTO powder was carried out using a TA Instruments
◦
-1
SDT 2960 system, with a heating rate of 5 C min in the
◦
temperature range from 25 to 1000 C under a flowing air
atmosphere. A Pt crucible was used.
NMR spectra were recorded with a Bruker Avance 300 spec-
7
trometer working at a frequency n
0
= 116.64 MHz for Li and
1
n
0
= 300.13 MHz for H. The experiments were performed on
powdered samples spinning at the magic angle with a standard
7
4
mm MAS probe. Li spectra were referenced from LiCl
saturated solution, and the proton ones were referenced from
TMS (tetramethylsilane). The p/2 pulse duration was 4 ms for
both nuclei. This pulse duration was sufficient to obtain a non
7
selective spectrum in the case of quadrupolar Li nucleus. For
1
the H nucleus, the recycle time was 200 s and 32 scans were
7
accumulated. For Li nucleus, the recycle time was 2000 s and
8
scans were accumulated. Magic Angle Spinning (MAS) one
7
dimension (1D) pulse experiments were performed on Li and
1
26-28
H nuclei with frequency rotation n
R
= 10 kHz. A Hahn echo,
Fig. 1 Observed, calculated and difference profiles of the final Rietveld
refinement for AS sample. Inset is a zoom of the 2q range from 75 to
1
with echo delay equal to 10 ms, was used for H nucleus. Prior
each sequence, a saturation comb is applied. In this paper Li
7
◦
134.2 . (b) Comparison of XRPD patterns for AS and AGED samples in
◦
spectra recorded in non synchronized and synchronized modes
the 2q range from 18 to 34 . The arrows indicate the filter apparatus effect.
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 3968–3975 | 3969

View Article Online
Table 1 Unit cell parameters and reliability factors obtained after Rietveld refinement of XRPD patterns of the AS and AGED samples
a = b/A˚
c/A˚
R
(l = 2n)
R
(l = 2n + 1)
R
R
R
I
I
p
wp
exp
AS
AGED
3.87335(3)
3.87335(3)
7.76258(6)
7.76604(8)
11.2
7.5
4.1
4.9
12.6
12
12.3
11.3
3.2
3.6
The final refinement involved 14 parameters for 194 reflections
(
1
98 and 86 in phase 1 (l = 2n) and 2 (l = 2n+1) respectively) up to
◦
34.2 . The R-factor values are gathered in Table 1. Results are in
good agreement with those obtained in ref. 4.
Fig. 1b shows the XRPD patterns of both AS and AGED
◦
samples in the 2q range from 18 to 34 . The vertical bars “|” on
the bottom of the figure indicate the position of the LLTO Bragg
peaks. The peak marked with the symbol “*” corresponds to the
most intense diffraction peak (102) of LLTO associated with the
˚
K
b1 radiation (1.39222 A) of the copper anticathode. These spectra
◦
◦
are very similar except in the 21.5 and 31 2q regions, shown in
the dashed frames. The diffraction lines that appear clearly in the
aged sample do not belong to the LLTO structure. However, the
lattice constants and atomic coordinates are similar in both AS
and AGED samples. As shown in Table 1, all samples have cell
parameters with a = b = a
P
and c ª 2a
parameter of the ideal perovskite structure.
P
, where a
p
is the cell
◦
TG curves measured in the temperature range from 25 C to
◦
1
000 C on AS and AGED samples of LLTO powders are shown
Fig. 3 IR spectra of the AS and AGED LLTO samples recorded in the
-
1
in Fig. 2. After synthesis, it is shown that the powder is thermally
stable in the temperature range investigated since a small weight
loss (about 0.4%) is observed between 150–300 C. On the other
wavenumber range from 400 to 4000 cm .
◦
-1
The Dn
of carbonate.
bending (n , p(CO
totally symmetric n
The bands observed at 1640 and 3460 cm are respectively
attributed to the bending mode d(H O) and the stretching mode
n(H O). They may be attributed to adsorbed water. However
3
= 54 cm splitting characterizes a unidentate structure
30,31
-1
hand, the TG curve obtained on aged powder reveals a larger
The band at 871 cm is related to an out of plane
)) and the band at 1088 cm-1 is assigned to
(CO).
◦
◦
weight loss between 150 C and 500 C, which accounts for 3.55%,
accompanied by an endothermic peak observed on the DTA curve.
The amount of adsorbed water on the surface of the powder is
small, of the order of 0.02 wt% (not visible in the figure).
2
3
S
-
1
2
2
these lines are also observed in the IR spectrum of the LLTO
7
sample exchanged in water, where protons are inserted in the
perovskite structure. In this case, protons are near oxygen atoms,
32
and form Ti(OH
modes observed are similar to those of adsorbed water. The
same observation is done in the Li Ti 12 sample, exchanged in
acid solution. In this case, the insertion of protons results in a
2
) groups as in titanyl phosphate. Then, vibration
4
5
O
21
-
1
+
vibration at 938 cm , due to a lattice coupling of the H -form
spinel.
21,33,34
+
In the case of H insertion in the perovskite structure,
this lattice coupling mode of the proton could explain the vibration
-
1
at 922 cm , observed in AGED LLTO and samples exchanged in
water (Li0.30-y
H
y
) La0.57TiO .
3
7
Fig. 4 shows the superposition of Li MAS NMR spectra for AS
and AGED LLTO samples and Li CO sample. The AS sample
◦
◦
2
3
Fig. 2 TG curves in the temperature range from 25 C to 1000 C of
spectrum evidences very small spinning side bands whereas the
AGED sample spectrum shows spinning side bands which are
the AS and AGED LLTO powder samples. The rise of temperature was
◦
-1
5
C min . The blue dashed line is for the TD curve for the AGED sample.
strictly identical to those observed for the Li
inset displays the superposition of isotropic lines of the AGED
LLTO sample and Li CO , obtained after normalization so as the
amplitudes of the spinning side bands of both samples (AGED
LLTO and Li CO ) are strictly identical. Then, this inset is a direct
view of the relative percentage of Li CO in the AGED LLTO
sample. The Li MAS NMR spectrum of lithium carbonate is
2
CO sample. The
3
Fig. 3 displays IR spectra for AS and AGED LLTO samples
-
1
recorded in the range from 400 to 4000 cm . These spectra
2
3
-
1
are both composed of bands centered at 593 cm (related to
(Ti–O)). For the AGED sample, 4 bands (1491, 1437, 1088
n
S
2
3
-
1
and 871 cm ) are straightforwardly attributed to the presence of
2
3
-
1
7
carbonates. Bands at 1491 and 1437 cm are associated with the
2
0
degeneracy of asymmetric stretching vibration mode (nAS(CO)).
similar to the one obtained by M e´ n e´ trier et al. After aging in air,
3
970 | Dalton Trans., 2010, 39, 3968–3975
This journal is © The Royal Society of Chemistry 2010

View Article Online
7
Fig. 4 Superposition of Li MAS NMR (10 kHz) spectra (spectral
window of 100 kHz) of the AS and AGED LLTO powder samples and of
Li CO . The inset is the superposition of the isotropic line of Li CO and
2 3 2 3
the AGED sample spectra, normalized so that the spinning side bands of
both samples are perfectly superposed.
the anisotropy is more significant in the AGED sample than in
the AS LLTO one, proving that the lithium environment changes
during aging. The fast Li motion in the AS sample averages
35-36
7
quadrupolar interaction,
whereas in the AGED sample, the
Fig. 5 (a) Li MAS NMR 10 kHz spectra (spectral window of 10 kHz)
of AS and AGED LLTO powder samples. Spectra have been normalized
to the powder mass. The inset shows the reconstruction of the AS sample
+
averaging process does not act, suggesting that the Li mobility
has decreased.
7
7
spectrum. (b) Reconstruction of the AGED sample Li NMR spectrum
with three lines (grey area is for Li CO ).
2 3
The Li synchronized MAS NMR spectra are presented in Fig. 5
and their characteristics are given in Table 2. Fig. 5a displays
7
the superposition of Li MAS NMR synchronized spectra of AS
and AGED LLTO samples (these spectra have been normalized
according to the mass of the powder sample). The spectra of AS
and AGED samples are different, the AGED spectrum shows an
asymmetric form. The intensity of the line at 1.51 ppm decreases
with aging, whereas a new line appears between 0 and -10 ppm.
The inset displays the reconstruction of the AS LLTO sample
spectrum. The narrow line called Li1 has an isotropic chemical
shift of 1.51 ppm and accounts for 51% of the signal intensity. The
broad line, at 1.41 ppm, accounts for 49%. These percentages are
in good agreement with those previously determined for samples
35-36
synthesized by the solid state reaction.
1
The H synchronized MAS NMR spectra are shown in Fig. 6.
The characteristics of the different line contributions are reported
1
in Table 3. Fig. 6a shows H MAS NMR spectra of AS and
AGED samples. The intensity of the AGED sample spectrum
is about 20 times higher than the AS sample one. The AGED
sample spectrum exhibits three lines, (H1, H2 and H3) as shown
in the left inset. The right inset shows the reconstruction of the
7
Table 2 Isotropic chemical shifts d (ppm), linewidths D (ppm) and line intensities (%) of the Li NMR spectra for AS, OA, AGED and Li
2
CO
3
samples
%
AS
OA
AGED
Li CO
2 3
Line
d
D
%
d
D
%
d
D
%
d
D
Li1
Li2
Li3
1.51
1.41
2.83
17.7
51
49
1.49
1.40
2.73
15.8
49
51
1.6
1.61
-0.03
1.94
18.2
9.79
18
16
66
-0.03
9.79
100
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 3968–3975 | 3971

View Article Online
1
Table 3 Isotropic chemical shifts d (ppm), linewidths D (ppm) and line intensities (%) of the H NMR lines in AS, AGED and exchanged samples
AS
AGED
Exchange in water [7]
Line
d
D
%
d
D
%
d
D
%
H1
H2
H3
H4
9.22
5.70
3.92
4.99
1.96
1.62
1.2
1.6
4.4
15.9
76.3
8.89
6.21
3.94
2.35
2.04
1.37
8.1
60.5
31.4
8.83
6.35
3.98
2.61
2.08
1.83
9.4
54.3
36.3
3.2
sample synthesized by solid state reaction after exchange (85%) in
7
ultra pure water (black dashed line).
The comparison of IR spectra of AS, AGED and OA samples
is illustrated in Fig. 7. After annealing, the lines attributed to
carbonate and hydroxyl groups inside the structure or close to the
surface disappear and IR spectra of OA and AS samples are then
◦
very similar (an annealing at higher temperature than 400 C will
lead to spectra without the Li
2
CO lines).
3
Fig. 7 IR spectra of the AS, AGED and OA LLTO samples in the
-1
wavenumber range from 400 to 4000 cm .
1
Fig. 6 (a) H MAS NMR (10 kHz) synchronized Hahn Echo spectra
recorded on AS (dotted line), and AGED (full line) LLTO powder samples.
Left inset shows reconstruction lines (H1, H2 and H3) of AGED curve
and right inset is the reconstruction of AS sample. (b) Superposition of
1
H MAS NMR (10 kHz) synchronized Hahn Echo spectra recorded on
AGED (full line) LLTO and LLTO synthesized by solid state reaction and
exchanged in water (85%) [7].
1
Fig. 8 H MAS NMR (10 kHz) synchronized Hahn Echo spectra
AS sample spectrum which can be done with four lines. The three
lines, H1, H2 and H3 previously observed in the AGED sample, are
observed but another predominant line appears, called H4. This
line H4 in the AS sample can be attributed to physically adsorbed
recorded on LLTO powder AGED, OA. Inset shows the superposition
of H MAS NMR (10 kHz) Hahn Echo spectra of AS and OA samples.
1
1
Fig. 8 shows the superposition of H MAS NMR spectra of
37-40
1
water.
As shown in Fig. 6b, the H MAS NMR spectrum of
AGED and OA samples. The intensity of spectra after annealing
1
the AGED LLTO is the same as the one obtained for the LLTO
strongly decreases. As shown in the inset, the weak intensity of H
3
972 | Dalton Trans., 2010, 39, 3968–3975
This journal is © The Royal Society of Chemistry 2010

View Article Online
NMR signal after annealing of the AGED sample is the same as
for the sample after synthesis.
Fig. 9a shows the superposition of Li MAS NMR spectra of
Discussion
1
7
H MAS NMR of the AGED sample shows clearly the presence
of proton sites already observed in the LLTO sample exchanged
AS, AGED and OA LLTO samples. After the thermal treatment,
the spinning side bands intensity strongly decreases, proving
that the particular lithium environment found in AGED sample
7
in water (Fig. 6b). In a previous work, it has been shown that
LLTO can undergo an exchange reaction with water in solution,
as follows:
7
disappears. This fact is confirmed by the superposition of Li MAS
+
NMR (10 kHz) mass normalized and synchronized spectra of AS,
AGED and OA LLTO samples (Fig. 9b). The spectrum of AS and
OA samples are identical and both can be reconstructed with two
Lorentzian lines, the characteristics of which are given in Table 2.
However some residual spinning side bands slightly more intensive
than the ones of AS sample are observed. A similar result can also
be observed in IR spectra (Fig. 7). All these experiments shown
in Fig. 7, 8 and 9 on AS, AGED and OA samples have been
performed several times and on several samples. They showed the
same characteristics confirming the reproducibility of the observed
Li0.30La 0.57TiO + y H O → (Li
y
H ) La TiO + y Li
3
2
(l)
0.30-y
-
0.57
3
+ y OH
(1)
This reaction can be either total when y = 0.3 or partial when
y < 0.3. In the exchanged compound three types of protons were
clearly evidenced in the substituted phase. These three protons
1
7
gave three lines in the H MAS NMR spectrum.
In a first approximation, as proton environments in the AGED
sample are identical to the ones exchanged in water, eqn (1) can
be rewritten as eqn (2) for reaction with water from air.
41
behavior.
Li0.30La0.57TiO + y H O → (Li
0.30-y
H ) La TiO + y
3
2
(g)
y
0.57
3
LiOH(s)
(2)
This equation shows the presence of four protons, three protons
from the substituted phase and one proton from the lithium
hydroxide present on the surface of the titanate after aging. The
three former protons are clearly observed in the spectrum of the
aged sample but the presence of a fourth proton, which should
be in a large amount, is not observed as clearly shown in the left
inset of Fig. 6a. This suggests that lithium hydroxide is not the end
product of the aging reaction.
The IR analysis performed on the AGED sample gives evidence
of the presence of carbonates (Fig. 3). Therefore, when LLTO is
in contact with humid air, the compound reacts with water. The
exchanged phase is formed, but the LiOH formed on the sample
surface reacts with carbon dioxide present in air to form Li
as described in eqn (3).
2
CO ,
3
8,9
2
LiOH(s)+ CO2(g) → Li
2
CO3(s) + H
2
O
(g)
(3)
Finally, the reaction of the LLTO sample with air can be written
as in eqn (4).
Li0.30La0.57TiO
3
+ y/2 H
La0.57TiO
2
O
(g) + y/2 CO2(g) → (Li0.30-y
y
H )
3
+ y/2 Li
2
CO3(s)
(4)
This equation explains why only proton sites of the exchanged
1
phase (Li0.30-y
H
y
) La0.57TiO are present in the H MAS NMR
3
spectrum of the AGED sample. It is worth noting that such a
behavior is linked to the high ionic lithium mobility in LLTO.
Therefore our experimental data can be unambiguously ex-
plained by the presence of Li
2
CO
3
in the AGED sample. The
infrared-active vibration modes observed in the IR spectrum at
-
1
1
491, 1437, 1088 and 871 cm can be assigned to the Li
2
CO
3
42-44
species.
Further, the diffraction peaks observed in the XRD
pattern at 21.3 and 30.6 (Fig. 1b) which appeared only in the
AGED sample can be attributed to the Li CO intense (1 1 0)
◦
◦
2
3
and (-2 0 2) diffraction peaks respectively (ICSD data 66941).
7
Finally, the new line that appears in the Li MAS NMR spectrum
between 0 and -10 ppm suggests a new Li environment. By keeping
constant the position and the width of the line found in the
7
Fig. 9 (a) Li MAS NMR (10 kHz) spectra (spectral window of 100 kHz)
7
Li CO sample (grey area), the AGED sample spectrum can be
of AS, AGED and OA LLTO powder samples. (b) Superposition of Li
2
3
MAS NMR synchronized (10 kHz) spectra of AS, AGED and OA LLTO
powder.
reconstructed with three lines, as shown in Fig. 5b. The two dashed
lines are attributed to the two kinds of lithium present in the LLTO
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 3968–3975 | 3973

View Article Online
structure. The relative percentage of the Li
to be 66%.
2
CO line is then found
3
Furthermore, the presence of lithium carbonate on the powder
of LLTO may explain the difference of the spin–lattice relaxation
45
time T
increase the T
1
found in the literature. This shell of carbonate may
value. This result points out the importance of
1
storage of the LLTO powder after synthesis. Further studies are
on the way to clarify this point.
It is well known from the literature that carbonation of lithium
hydroxide proceeds in a two-step reaction with the formation of the
intermediate hydrate (LiOH·H
2
O) which is a necessary precursor
of the CO reaction. A relative humidity between 20 and 70% is
8,9
2
required for the reaction to proceed. This is the storage condition
of LLTO. After synthesis, adsorbed water is present on the oxide
1
surface and can be detected in the weak intensity H MAS NMR
spectrum as the predominant compound (Fig. 6a right inset).
38-41
Line H4 can then be attributed to this adsorbed water.
As
carbonation proceeds, water is consumed, the lithium carbonate
and the exchanged phase are formed leading to the disappearance
of the H4 line and to the increase of the three H1, H2 and H3 lines
2 2
Fig. 10 Scheme of the mechanism proposed for CO and H O adsorption
on LLTO.
1
of (Li0.30-y
H
y
) La0.57TiO
3
phase in the H MAS NMR spectrum
(
Fig. 6a left inset).
It is then shown that LLTO reacts at room temperature with
+
+
shown in Fig. 9b where the OA sample displays a spectrum without
water and CO contained in air through a H /Li ionic exchange.
2
any line between 0 and -10 ppm characteristic of the presence
In a previous work, we already showed that LLTO was unstable
+
+
of Li CO . The thermal decomposition of (Li
H )La0.57TiO
in aqueous solutions and that exchange H /Li occurs in water
2
3
0.30-y
7
y
3
6
◦
◦
observed in sample exchanged in acid or water at 350 C and
with an efficiency of 85% after 24 h at 70 C. The present study
◦
the melting point of Li
2
CO
3
(710 C) are not observed, proving
shows that this ionic exchange also occurs in humid air at room
temperature. The kinetics are slower than the one in aqueous
solution since an efficiency of 66% is found after 5 months of
aging. Therefore, the carbonation of LLTO proceeds in a two-step
◦
◦
that both species react together from 150 C to 500 C to form
Li0.30La0.57TiO .
3
These results demonstrate the reversibility of the carbonation
◦
+
+
reaction. At 250 C, the mass loss can be written as in eqn (5).
reaction: first the H /Li ionic exchange with water and later the
reaction with CO from air.
2
(
Li0.30-y
H
y
) La0.57TiO
3
+ y/2 Li
2
CO3(s) → Li0.30La0.57TiO
3
In the literature, other lithium containing ceramics also react
with carbon dioxide from the air. But, for lithium zirconate,
+
y/2 H
2
O(g) + y/2 CO2(g)
(5)
1
1-16
silicates or titanates
the efficiency maximum of the reaction
◦
is reached at very high temperature (up to 600 C), the kinetic is
considerably slower at room temperature. The process is described
in two steps: First, a fast reaction forming a shell of lithium
carbonate on the surface which prevents the lithium reaction with
For a total efficiency (y = 0.3) of this reaction, the theoretical
mass loss would be equal to 5%. By comparison with the
experimental mass loss observed for the AGED sample (3.55%), it
can be concluded that the efficiency of the aging process is about
CO
occurs and depends strongly on the temperature. For LLTO, we
can propose a mechanism for CO and H O adsorption (Fig. 10).
It illustrates the fact that a shell of lithium carbonate is formed
on the surface of the grain with H0.30La0.57TiO .The interior of the
2
from air. Secondly, a slow process of diffusion through Li
2
CO
3
68%. This value can be compared to the value of 66% of Li
previously determined from Fig. 5b. Both similar values show that
eqn (4) seems to be very realistic for the reaction of LLTO with
2
CO
3
2
2
CO under humid air at room temperature.
2
1
7
3
It is the first time, to our knowledge, that H and Li NMR are
associated with IR spectroscopy to reveal such a two step behavior
in the exchangeable phase LLTO.
grain would be made of non exchanged LLTO. The lines Li1 and
7
Li2 in the Li MAS NMR spectrum of the AGED LLTO sample
2
1,46
are attributed to these non exchanged sites.
Previous work on Li
exchange process may occur under humid air, but no reaction
mechanism was clearly proposed. The Li CO formation was
explained by the conversion of surface Li O with CO
study, the efficiency of the process (66%) and H MAS NMR
4
Ti
5
O
12
had already suggested that the
1
It is clearly shown in the IR spectra of Fig. 7 and in the H
7
+
+
and Li MAS NMR of Fig. 8 and 9 that the H /Li exchange
2
3
10
process and the formation of Li
2
CO
3
are reversible. After a thermal
2
2
. In our
◦
1
treatment at 400 C for two hours, the LLTO sample is recovered.
This is confirmed by the mass loss observed during heating of the
7
6
spectra (which is identical to LLTO exchange in water or acid )
is not compatible with exchange only from surface lithium. This
is also confirmed by the small value of the specific surface (0.48
◦
aged sample above 100 C (Fig. 2). After annealing, IR spectra
1
do not show evidence of any carbonate signal. H MAS NMR
2
-1
spectrum of sample OA is the same as the one for LLTO after
synthesis (inset Fig. 8). The signature of the inserted proton in
m g ). Lithium ions which react with carbon dioxide also emerged
from the structure because of the high ionic lithium mobility and
instability of the lithium containing phase (the protonated phase
being thermodynamically more stable).
7
the perovskite structure disappears. Moreover Li MAS NMR
spectra of the OA sample are also similar to the AS sample one, as
3
974 | Dalton Trans., 2010, 39, 3968–3975
This journal is © The Royal Society of Chemistry 2010

View Article Online
Conclusions
12 K. Essaki, K. Nakagawa and M. Kato, J. Ceram. Soc. Japan, 2001, 109,
8
29.
1
3 E. Ochoa-Fernandez, M. Ronning, T. Grande and D. Chen, Chem.
This work clearly demonstrates that the LLTO powder sample
+
+
Mater., 2006, 18, 1383.
reacts with a humid atmosphere leading to H /Li ionic exchange.
1
4 H. Pfeiffer, C. Vasquez, V. H. Lara and P. Bosch, Chem. Mater., 2007,
+
+
It proves that this ionic exchange H /Li is the first step of
the mechanism and occurs with water vapor from air at room
temperature. Then, lithium ions emerge from the structure and
1
9, 922.
15 C. Gauer and W. Heschel, J. Mater. Sci., 2006, 41, 2405.
1
1
1
1
6 N. Togashi, T. Okumura and K. Oh-ishi, J. Ceram. Soc. Jpn., 2007, 115,
24.
3
react with CO
and a stable partial protonated phase (Li0.30-y
2
from ambient air, to form lithium carbonate
with
7 K. Shizuka, C. Kiyohara, K. Shima and Y. Takeda, J. Power Sources,
H
y
) La0.57TiO
3
2007, 166, 233.
8 K. Matsumoto, R. Kuzuo, K. Takeya and A. Yamanaka, J. Power
Sources, 1999, 81–82, 558.
9 G. V. Zhuang, G. Chen, J. Shim, X. Song, P. N. Ross and J. T.
Richardson, J. Power Sources, 2004, 134, 293.
the same perovskite structure as sample LLTO. An exchange
efficiency of 66% is found after 5 months of aging at room
temperature.
The second relevant result is the reversibility of the carbonation
reaction. An annealing at 400 C allows LLTO recovering. This is
20 M. M e´ n e´ trier, C. Vaysse, L. Croguennec, C. Delmas, C. Jordy, F.
Bonhomme and P. Biensan, Electrochem. Solid-State Lett., 2004, 7,
A140–A143.
◦
a very significant result that opens a large avenue of research for
2
1 D. R. Simon, E. M. Kelder, M. Wagemaker, F. M. Mulder and J.
Schoonman, Solid State Ionics, 2006, 177, 2759.
+
+
all exchangeable phases Li /H , to which LLTO belongs.
This study provides an explanation for understanding exchange-
22 U. S. Pat. 3,330,397, 1967.
2
3 M. Vijayakumar, Y. Inaguma, W. Mashiko, M. P. Crosnier-Lopez and
C. Bohnk e´ , Chem. Mater., 2004, 16, 2719.
4 H. M. Rietveld, J. Appl. Crystallogr., 1969, 2, 65.
able lithium phases. Such samples could be used to retain CO
air to limit the greenhouse effect.
2
from
2
2
5 J. Rodriguez-Carvajal, FULLPROF program: Rietveld Pattern Match-
ing Analysis of Powder Patterns, ILL, Grenoble, 1990.
6 E. L. Hahn, Phys. Rev., 1950, 80, 580.
2
2
2
2
Acknowledgements
7 J. Lowe, Bull. Am. Phys. Soc., 1957, 2, 344.
8 S. R. Hartmann and E. L. Hahn, Phys. Rev., 1962, 128, 2042.
9 D. Massiot, F. Fayon, M. Capron, I. King, S. Le Calv e´ , B. Alonso,
J. O. Durand, B. Bujoli, Z. Gan and G. Hoaston, Magn. Reson. Chem.,
2002, 40, 70.
The authors would like to thank the laboratory “Unit e´ de Chimie
Organique Mol e´ culaire et Macromol e´ culaire” (UCO2M, UMR
CNRS 6011) for access to the IR spectrometer. They would also
like to thank Dominique Massiot, Christian Bonhomme and
Michel M e´ n e´ trier for useful discussions.
3
3
3
0 G. Busca and V. Lorenzelli, Mater. Chem., 1982, 7, 89.
1 J. C. Lavalley, Catal. Today, 1996, 27, 377.
2 S. Benmokhtar, A. El jazouli, J. P. Chaminade, P. Gravereau, M.
M e´ n e´ trier and F. Bour e´ e, J. Solid State Chem., 2007, 180, 2713.
3 P. Lavela, L. Sanchez, J. L. Tirado, S. Bach and J. P. Pereira-Ramos,
J. Solid State Chem., 2000, 150, 196.
3
3
References
4 P. Aitchison, B. Ammundsen, T. Bell, D. Jones, J. Rozi e` re, G. Burns, H.
Berg, R. Tellgren and J. Thomas, J., Phys. B, 2000, 276–278, 847.
1
2
3
4
5
6
7
8
9
A. G. Belous, G. N. Novitskaya, S. V. Polyanetskaya and Y. I. Gornikov,
Zh. Neorg. Khim., 1987, 32, 283.
35 J. Emery, O. Bohnk e´ , J. L. Fourquet, J. Y. Buzar e´ , P. Florian and D.
Massiot, C.R. Chimie, 2001, 4, 845.
36 J. Emery, O. Bohnk e´ , J. L. Fourquet, J. Y. Buzar e´ , P. Florian and D.
Massiot, J. Phys.: Condens. Matter, 2002, 14, 523.
37 M. Jalbring, D. E. Sandstr o¨ m, O. N. Antzutkin and W. Forsling,
Langmuir, 2006, 22, 11060.
38 Y. A. Nosaka, T. Fujiwara, H. Yagi, H. Akutsu and Y. Nosaka,
Langmuir, 2003, 19, 1935.
39 G. Scholz, I. D o¨ rfel, D. Heidemann, M. Feist and R. St o¨ sser, J. Solid
State Chem., 2006, 179, 1119.
40 J. Soria, J. Sanz, I. Sobrados, J. M. Coronado, F. Fresno and M. D.
Hernandez-Alonso, Catal. Today, 2007, 129, 240.
41 A. Boulant, PhD Dissertation, Universite du Maine, Le Mans (2009).
42 P. Pasierb, S. Komornicki, M. Rokita and M. Rekas, J. Mol. Struct.,
2001, 596, 151.
43 K. Buijs and C. J. H. Schutte, Spectrochim. Acta, 1961, 17, 927.
44 M. H. Brooker and J. B. Bates, J. Chem. Phys., 1971, 54, 4788.
45 Y. Hirakoso, Y. Harada, J. Kuwano, Y. Saito, Y. Ishikawa and T. Eguchi,
Key Eng. Mater., 1999, 169–170, 209.
Y. Inaguma, L. Chen, M. Itoh, T. Nakamura, T. Uchida, H. Ikuta and
M. Wakihara, Solid State Commun., 1993, 86, 689.
S. Stramare, V. Thangadurai and W. Weppner, Chem. Mater., 2003,
1
5(21), 3974.
J. L. Fourquet, H. Duroy and M. P. Crosnier-Lopez, J. Solid State
Chem., 1996, 127, 283.
A. D. Roberston, A. R. West and A. G. Ritchie, Solid State Ionics,
1
997, 104, 1.
N. S. P. Bhuvanesh, O. Bohnk e´ , H. Duroy, M. P. Crosnier-Lopez, J.
Emery and J. L. Fourquet, Mater. Res. Bull., 1998, 33, 1681.
A. Boulant, P. Maury, J. Emery, J. Y. Buzar e´ and O. Bohnk e´ , Chem.
Mater., 2009, 21, 2209.
D. D. Williams and R. R. Miller, Ind. Eng. Chem. Fundam., 1970, 9,
4
54.
D. A. Boryta and A. J. Mass, Ind. Eng. Chem. Proc. Des. Dev., 1971,
0, 489.
0 H. A. Mosqueda, C. Vasquez, P. Bosch and H. Pfeiffer, Chem. Mater.,
006, 18, 2307.
1
1
1
2
1 A. Iwan, H. Stephenson, W. C. Ketchie and A. A. Lapkin, Chem.
Eng. J., 2009, 146, 249.
46 M. Q. Snyder, W. J. DeSisto and C. P. Tripp, Appl. Surf. Sci., 2007, 253,
9336.
This journal is © The Royal Society of Chemistry 2010
Dalton Trans., 2010, 39, 3968–3975 | 3975