2. Synthesis, Structure and Morphology
Historically, La
1−xSr
xGa
1−yMg
yO
3−δ (LSGM) oxides were the first well-studied doped materials in the LaGaO
3 system. In 1998, Huang, Tichy and Goodenough determined the existence of single-phase La
1−xSr
xGa
1−yMg
yO
3−0.5(x+y) perovskites while studying a LaO
1.5-SrO-GaO
1.5-MgO quasi-quaternary diagram
[54]. This was possible due to variations in both
x and
y contents in a composition range of 0.05–0.30 with a step of 0.05. Sr- and Mg- co-doped LaGaO
3 samples were prepared from La
2O
3, SrCO
3, Ga
2O
3, and MgO using solid-state reaction technology. The obtained powders were pressed into pellets and calcined at 1250 °C for 12 h. After remilling and repressing, the final pellets were finally sintered in air at 1470 °C for 24 h and quenched in a furnace at 500 °C.
Similar conventional techniques for synthesizing La
1−xSr
xGa
1−yMg
yO
3−δ were used in other studies
[55][56]. La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ samples were obtained from La
2O
3, SrCO
3, Ga
2O
3 and MgO sources, which were mixed and sintered in a platinum crucible at 1350 °C for 12 h
[55]. The annealed powder was milled with zirconia balls and dried. Then, the powder was pressed into disks and sintered at 1350 °C in air, nitrogen or oxygen atmospheres for various times ranging from 20 min to 5 h. Moure et al.
[56] obtained La
0.8Sr
0.2Ga
0.85Mg
0.15O
3−δ and La
0.8Sr
0.15Ga
0.85Mg
0.2O
3−δ samples from La
2O
3, SrCO
3, Ga
2O
3 and MgO, which were mechanochemically activated in a Pulverizette 6 Fritsch planetary mill with stainless steel balls. The mixtures were synthesized at 1300 °C for 16 h; then after milling for 2 h and sieving with a 100-μm sieve, the powders were pressed into pellets and finally sintered at 1550 °C to form the desired ceramic samples.
For the synthesis of La
0.9Sr
0.1Ga
1−xNi
xO
3−δ, Colomer and Kilner
[57] grinded a mixture of La
2O
3, SrCO
3, Ga
2O
3 and NiO in an agate mortar with acetone medium and then calcined them at 1000 °C for 6 h. After sieving with a 65-μm sieve, milling for 1 h, drying and secondary sieving to 65 μm, the finishing powders were pressed into disks and sintered at 1450–1500 °C for 48 h in air. The researchers chose nickel as element for gallium substitution in La
0.9Sr
0.1GaO
3−δ owing to the proposal about achieving a hopping conductivity among the Ni-sites.
Al-substituted La
0.95Sr
0.05Ga
0.9Mg
0.1O
3−δ and La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ derivatives were prepared using La
2O
3, Ga
2O
3, SrO, MgO and Al
2O
3 [58]. Mechanosynthesis was employed in a planetary mill (Retsch PM100, PM200) with tetragonal zirconia balls. The powders were pressed into disks that were sintered at 1300–1450 °C for 2–24 h.
As can be seen, the aforementioned methods (solid-state reaction synthesis and the mechanochemical route) that were conventionally used for the preparation of La
1−xSr
xGa
1−yMg
yO
3−δ and its derivatives have two considerable disadvantages. First, high sintering temperatures (above 1450–1500 °C) are required for full densification of the pressed pellets
[51]. This can influence the production cost of the final electrolyte materials. Second, the appearance of Sr
3La
4O
9, SrLaGa
3O
7 and/or SrLaGaO
4 impurity phases in La
1−xSr
xGa
1−yMg
yO
3−δ samples was frequently observed. This was due to gallium evaporation
[59], which resulted in the deterioration of the gallate material’s ionic conductivity
[51]. To solve the problems that arise during La
1−xSr
xGa
1−yMg
yO
3−δ preparation, techniques based on co-precipitation
[60][61], organic-nitrate precursors combustion
[55][62][63][64][65][66][67][68], self-propagating, high-temperature synthesis
[69][70] and spray-pyrolysis
[71] were developed.
For example, La
0.8Sr
0.2Ga
0.8Mg
0.2O
3−δ samples were prepared with carbonate co-precipitation from La(NO
3)
3·6H
2O, Sr(NO
3)
2, Ga(NO
3)
3·xH
2O and Mg(NO
3)
2·6H
2O starting reagents
[60]. The resulting aqueous solution containing La
3+, Sr
2+, Ga
3+ and Mg
2+ cations was gradually dropped into an aqueous (NH
4)
2CO
3 solution with heating at 70 °C. After 2 h of homogenization with continuous stirring, the formed sediments were washed, dried at 25 °C for 24 h in a N
2 atmosphere, and finally calcined in air at 900–1300 °C for 12 h.
Huang and Goodenough
[63] have reported the use of wet synthesis techniques (the sol-gel technique and the Pechini method) for forming single-phase La
0.8Sr
0.2Ga
0.83Mg
0.17O
3−δ materials. Solutions of La(CH
3COO)
3, Sr(CH
3COO)
2 and Mg(CH
3COO)
2 acetates and La(NO
3)
3, Sr(NO
3)
2, Ga(NO
3)
3 and Mg(NO
3)
2 nitrates were used in these preparation methods. During synthesis with sol-gel technology, the required amounts of metal acetates and gallium nitrate solutions were mixed by stirring. An ammonia solution was then added, forming a white gel. This was aged at 25 °C for 72 h and heated at 150 °C for 8 h upon full water evaporation. The resulting product was fired at 300, 500 and 700 °C at varying times. Using the Pechini method, La
0.8Sr
0.2Ga
0.83Mg
0.17O
3−δ samples were prepared from a mixture of the necessary amounts of metal nitrate solutions at 25 °C: citric acid was then added. The citric acid was used to fulfil a mole ratio of citric acid/total cations around 1.5/1. After stirring the precursor solution, ethylene glycol was added in an equal amount to the citric acid. The obtained solution was heated at 150 °C for 12 h and resulted in a polymer-like solid material. This resin was slowly heated to 300 °C and, after several sintering stages, it was finally calcined at 1400 °C for 4 h
[63]. The pressed La
0.85Sr
0.15Ga
0.8Mg
0.2O
3−δ samples were found to be single-phase after they were obtained via the Pechini method and annealed at 1400 °C for 6 h
[64].
A La
0.8Sr
0.2Ga
0.85Mg
0.15O
3−δ sample was also obtained via the glycine-nitrate combustion method
[65]. Ga, La
2O
3, MgO and SrCO
3 powders were dissolved in strong HNO
3 and mixed with water. Glycine was then added with a molar ratio of glycine/nitrate ions equal to 1:1. The glass beaker with the precursor glycine–nitrate solution was heated on a hot plate with spontaneous burning, which resulted in a white powder. Dense samples were formed at a temperature range of 1400–1550 °C for 6 h at each stage
[65]. A similar method was used in
[66] for the synthesis of La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ. The experimental procedure included the heating of the precursor glycine–nitrate solution at 550 °C upon combustion, initial calcination of voluminous oxide powders at 800 °C for 3 h, annealing the powders at 1000 °C and final annealing at 1300 °C for 2 h. It should be noted that the researchers of
[66] could not achieve single-phase sample. Huang and Goodenough also concluded that a La
0.8Sr
0.2Ga
0.83Mg
0.17O
3−δ single-phase material cannot be formed via hydrothermal treatment synthesis
[63].
In
[69], Ishikawa et al., prepared La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ and La
0.9Sr
0.1Ga
0.7Mg
0.3O
3−δ samples via self-propagating high-temperature synthesis from La
2O
3, SrCO
3, Ga
2O
3, Mg and NaClO
4. An initial powder mixture was supplied to a self-propagating synthesis reactor: it was then ignited with a disposable carbon foil in contact with the sample. The obtained powders were washed with water to remove NaCl. The samples were pressed into disks in vacuum and then sintered at a temperature range of 1000–1500 °C for 6 h in air. An alternative process for La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ synthesis based on a preliminarily mechanically activated powder mixture was proposed by Ishikawa et al.
[70]. The initial mixture was grinded in a planetary mill with stainless steel balls. The powder sample was pressed into a disk, which was placed in a self-propagating synthesis reactor: the aforementioned algorithm
[69] was then used.
It is worth noting that the crystal structure of the obtained LSGM samples depends on the strontium and manganese dopant contents. Basic LaGaO
3 at room temperature has an orthorhombic structure
[72] but varying the doping contents can change the crystal structure symmetry
[63][73]. Generally, the substitution of La
3+-ions with Sr
2+-ions increases the tolerance factor
t , while Ga-with-Mg substitution decreases it. Therefore, the
t factor for La
1−xSr
xGa
1−yMg
yO
3−δ is nearly equal to that calculated for undoped LaGaO
3.
The
t factor is equal to 1 for La
0.8Sr
0.2Ga
0.8Mg
0.2O
3−δ, which exhibits an ideal
Pm-3m cubic structure with a unit cell parameter of
a = 3.9146(1) Å
[73]. According to
[73], the crystal structure of La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ and La
0.9Sr
0.1Ga
0.9Mg
0.1O
3−δ samples was refined in a
I2/a monoclinic space group.
The crystal structure of LaGaO
3 and La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ samples was investigated via powder neutron diffraction at 25, 800 and 1000 °C in
[74]. According to the Rietveld refinement analysis of the diffraction data collected at 25 °C, an orthorhombic structure was observed for both samples: fitting was provided in the
Pnma space group for LaGaO
3 (unit cell parameters were equal to
a = 5.4908(1),
b = 7.7925(1) and
c = 5.5227(1) Å) and in the
Imma space group for La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ (unit cell parameters were equal to
a = 5.5179(1),
b = 7.8200(1) and
c = 5.5394(1) Å). The high temperature measurements
[74] show that the LaGaO
3 sample possessed a rhombohedral structure in the
R-3c space group (unit cell parameters were equal to
a = 5.5899(1) Å and
a = 5.5987(1) Å at 800 and 1000 °C, correspondingly), whereas La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ exhibits a cubic structure in the
Pm3m space group (unit cell parameters were equal to
a = 3.9760(1) Å and
a = 3.9866(1) Å at 800 and 1000 °C, correspondingly). Similar data at 25 °C (the
Imma space group
, a = 5.5056(9),
b = 7.8241(7),
c = 5.5387(5) Å) for a La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ sample obtained via solid-state route and sintered at 1350 °C for 2 h was reported in
[75].
Comparative analysis of the microstructural parameters for La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ disks sintered at 1400 °C for 6 h obtained via the self-propagating high-temperature and solid-reaction synthesis techniques showed that the first sample was denser
[69]. The relative densities of the samples were 98 and 92%, respectively, despite the fact that the sintering temperature for the first disk was 100 °C lower than that for the second one. These SEM images testify that mechanically activated self-propagating synthesis provided the high-grade powders with nano-size particles. The specific surface areas of the samples were 3.36 and 2.06 m
2 g
−1, respectively. Based on both studies, Ishikawa et al.
[69][70] concluded that this proved the advantages of using self-propagating high-temperature synthesis (especially with mechanical activation of the starting mixture) in comparison with the solid-reaction method.
The evolution of a La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ sample’s density against temperature was provided in by Batista et al.
[75]. Based on dilatometry experimental results, the researchers separated the process into three steps: an insignificant increase of relative density at 25–1000 °C; gradual densification at 1000–1300 °C; and, finally, a fast densification above 1300 °C. According to
[76], a relative density of over 99% was achieved after calcination at 1450 °C for 6 h.
To sum up, which was devoted to the synthesis methods of Sr, Mg-doped LaGaO3 oxides as electrolyte materials, the self-propagating high-temperature synthesis with mechanical activation of the starting mixtures can be identified as one of the most optimal techniques. The above-mentioned method can obtain the single-phase La0.9Sr0.1Ga0.8Mg0.2O3−δ powders with high specific surface areas, a narrow distribution of nano-size particles, and high relative densities for the sintered ceramic samples.
3. Functional Properties
In 1994, Ishihara et al.
[51] were the first to show that the La-substitution of LaGaO
3 with strontium and gallium with magnesium increased the electrical conductivity of doped materials owing to the formation of oxygen vacancies in La
1−xSr
xGa
1−yMg
yO
3−δ [77].
The measurements of Ishihara
[51], Stevenson
[78] and Goodenough
[54] demonstrate that the La
1−xSr
xGa
1−yMg
yO
3−δ samples possess maximal electrical conductivity values at
x = 0.15/0.2 and
y = 0.2, as can be seen in
Table 1. It should be also noted that conductivity of nominally similar materials can be varied over a wide range. This confirms that the microstructural parameters of ceramics, as well as the presence of insulating impurity phases, considerably affect the transport properties of gallates, encouraging the continuous search for their new synthesis and fabricating techniques.
Table 1. Total conductivities of LaGaO3-based materials depending on their compositions, preparation methods and temperatures.
|
|
|
|
|
LaGaO3 |
Solid-state route; 1500 |
950 |
0.02 |
[51] |
La0.9Sr0.1Ga0.9Mg0.1O3−δ |
Solid-state route; 1500 |
950 |
0.20 |
[51] |
La0.9Sr0.1Ga0.85Mg0.15O3−δ |
Solid-state route; 1500 |
950 |
0.27 |
[51] |
La0.9Sr0.1Ga0.8Mg0.2O3−δ |
Solid-state route; 1500 |
950 |
0.29 |
[51] |
La0.9Sr0.1Ga0.7Mg0.3O3−δ |
Solid-state route; 1500 |
950 |
0.28 |
[51] |
La0.9Sr0.1Ga0.6Mg0.4O3−δ |
Solid-state route; 1500 |
950 |
0.10 |
[51] |
La0.9Sr0.1Ga0.8Mg0.2O3−δ |
Glycine-combustion method; 1400 |
1000 |
0.26 |
[51] |
La0.85Sr0.15Ga0.8Mg0.2O3−δ |
Glycine-combustion method; 1400 |
1000 |
0.36 |
[51] |
La0.8Sr0.2Ga0.85Mg0.15O3−δ |
Glycine-combustion method; 1400 |
1000 |
0.31 |
[51] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Glycine-combustion method; 1400 |
1000 |
0.40 |
[51] |
La0.9Sr0.1Ga0.9Mg0.1O3−δ |
Solid-state route; 1470 |
800 |
0.116 |
[54] |
La0.9Sr0.1Ga0.85Mg0.15O3−δ |
Solid-state route; 1470 |
800 |
0.127 |
[54] |
La0.9Sr0.1Ga0.8Mg0.2O3−δ |
Solid-state route; 1470 |
800 |
0.132 |
[54] |
La0.9Sr0.1Ga0.7Mg0.3O3−δ |
Solid-state route; 1470 |
800 |
0.096 |
[54] |
La0.85Sr0.15Ga0.8Mg0.2O3−δ |
Solid-state route; 1470 |
800 |
0.150 |
[54] |
La0.8Sr0.2Ga0.85Mg0.15O3−δ |
Solid-state route; 1470 |
800 |
0.149 |
[54] |
La0.8Sr0.2Ga0.83Mg0.17O3−δ |
Solid-state route; 1470 |
800 |
0.17 |
[54] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Solid-state route; 1470 |
800 |
0.14 |
[54] |
La0.7Sr0.3Ga0.8Mg0.2O3−δ |
Solid-state route; 1470 |
800 |
0.109 |
[54] |
La0.9Sr0.1Ga0.8Mg0.2O3−δ |
Self-propagating high-temperature synthesis; 1500 |
800 |
0.11 |
[69] |
La0.9Sr0.1Ga0.8Mg0.2O3−δ |
Carbonate co-precipitation; 1400 |
800 |
0.045 |
[61] |
La0.9Sr0.1Ga0.9Mg0.1O3−δ |
Solid-state route; 1450 |
800 |
0.071 |
[73] |
La0.9Sr0.1Ga0.8Mg0.2O3−δ |
Solid-state route; 1450 |
800 |
0.1095 |
[73] |
La0.9Sr0.1Ga0.8Mg0.2O3−δ |
Glycine-combustion method; 1500 |
800 |
0.092 |
[79] |
La0.9Sr0.1Ga0.8Mg0.2O3−δ |
Glycine-combustion method; 1400 |
800 |
0.0395 |
[80] |
La0.85Sr0.15Ga0.85Mg0.15O3−δ |
Acrylamide polymerization technique; 1432 |
800 |
0.093 |
[81] |
La0.85Sr0.15Ga0.8Mg0.2O3−δ |
Mechanochemical route; 1380 |
600 |
0.016 |
[56] |
La0.85Sr0.15Ga0.8Mg0.2O3−δ |
Glycine-combustion method; 1300 |
800 |
0.053 |
[82] |
La0.85Sr0.15Ga0.8Mg0.2O3−δ |
EDTA-combustion method; 1300 |
800 |
0.06 |
[82] |
La0.85Sr0.15Ga0.8Mg0.2O3−δ |
Glycine-combustion method; 1400 |
800 |
0.096 |
[64] |
La0.85Sr0.15Ga0.8Mg0.2O3−δ |
Pechini method; 1400 |
800 |
0.135 |
[83] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Carbonate co-precipitation; 1300 |
600 |
0.014 |
[60] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Glycine-combustion method; 1300 |
700 |
0.022 |
[68] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Glycine-combustion method; 1400 |
700 |
0.085 |
[68] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Spray pyrolysis; 1400 |
500 |
0.0029 |
[71] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Solid-state route; 1450 |
800 |
0.126 |
[84] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Solid-state route; 1400 |
800 |
0.035 |
[84] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Hydrothermal urea hydrolysis precipitation; 1400 |
800 |
0.056 |
[84] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Carbonate co-precipitation; 1400 |
800 |
0.137 |
[85] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Solid-state route; 1250 |
727 |
0.019 |
[86] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Sol-gel technique; 1300 |
450 |
2.9 × 10−4 |
[87] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Solid-state route; 1400 |
800 |
0.132 |
[88] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Thin film deposited by vacuum cold spray; 200 |
750 |
0.043 |
[89] |
La0.8Sr0.2Ga0.8Mg0.2O3−δ |
Step-wise current-limiting flash sintering process; 690 |
850 |
0.072 |
[90] |
Hayashi et al.
[91] concluded that the electrical conductivity of La
1−xSr
xGa
1−yMg
yO
3−δ becomes greater when approaching the tolerance factor of the doped sample to
t for LaGaO
3 and decreases when the tolerance factor for the doped samples differed from
t for LaGaO
3. It was established that increasing the Sr, Mg-doping levels led to the association of oxygen vacancies
[51][78][91]; for this reason, further electrical investigations of the doped-LaGaO
3 oxides were performed on La
1−xSr
xGa
1−yMg
yO
3−δ samples with a fixed content of Sr and Mg dopants (nearly 20 mol.%, i.e.,
x =
y = 0.2). The literature on the transport properties of La
1−xSr
xGa
1−yMg
yO
3−δ ceramic samples is summarised in
Table 1.
It was shown in
[78] that the ion-transfer numbers were nearly equal to 1. For La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ and La
0.8Sr
0.2Ga
0.8Mg
0.2O
3−δ ceramic samples, the oxygen-ion transference numbers were found to be equal 1 at 700–1000 °C
[66], confirming the presence of electrolyte-type behaviour. Savioli and Watson
[92] studied the defect structure of LaGaO
3 upon the use of various doping strategies using DFT calculations. They confirmed that Sr-, Ba-, and Mg-doping should result in the greatest improvements to the ionic conductivity of the LaGaO
3 parent phase, while the Ni
2+-, Co
2+-, Fe
2+-, and Zn
2+-doping is responsible for the generation of a mixed ionic-electronic conducting behaviour. Sr- and Mg- co-doped LaGaO
3 complex oxides are predominantly oxygen-ionic conductors, for which the electronic conductivity levels are 3–4 magnitudes lower compared to the oxygen-ionic conductivity levels
[93].
According to
[82], the dependence ln(
σT) vs. 1/
T had a break at 700 °C for La
0.85Sr
0.15Ga
0.8Mg
0.2O
3−δ, which indicates that the activation energy value of oxygen-ion conductivity at a low-temperature range was higher than that at a high-temperature range.
A linear correlation between hardness and total ionic conductivity was revealed in
[83] for La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ and La
0.85Sr
0.15Ga
0.8Mg
0.2O
3−δ samples. It was shown that the electrical and mechanical properties of La
1−xSr
xGa
1−yMg
yO
3−δ are strongly defined by microstructural peculiarities and the presence of low-conductive LaSrGaO
4 and LaSrGa
3O
7 impurity phases
[80]. The LaSrGaO
4 phase exhibits a tetragonal structure K
2NiF
4-type and crystalizes in the
I4/mmm space group; its conductivity is found to be around 2·10
−7 S cm
−1 at 900 °C
[94]. The LaSrGa
3O
7 phase belongs to a melilitestructure described in the
P421m space group; its ionic conductivity level is around 2·10
−6 S cm
−1 at 800 °C
[95]. The maximum values of ionic conductivity and hardness were achieved for single-phase La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ (LSGM1020) and La
0.85Sr
0.15Ga
0.8Mg
0.2O
3−δ (LSGM1520) samples with a high relative density. With a significant amount of impurity phases at the grain boundaries, the samples exhibited a gradual decrease in hardness and the grain boundary conductivity, which resulted in a decreasing total conductivity.
The thermal expansion of La
1−xSr
xGa
1−yMg
yO
3−δ was studied by Baskaran et al.
[96]. The TEC values measured for the La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ sample were equal to 10 × 10
−6 K
−1 over a low-temperature range and 13.5–14.0 × 10
−6 K
−1 above 600 °C. Lee et al.
[62] reported about an average TEC of 12.1 × 10
−6 K
−1 for La
0.8Sr
0.2Ga
0.8Mg
0.2O
3−δ at a temperature range of 25–1000 °C, which is close to 12.3 × 10
−6 K
−1 for a La
0.65Sr
0.3MnO
3−δ electrode at the same temperatures
[97].
The expansion behaviour for La
1−xSr
xGa
1−yMg
yO
3−δ is correlated with its crystal structure in the observed temperature range. Therefore, the presence of a phase transition from an orthorhombic phase to a cubic one for La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ [74] and the existence of an ideal perovskite cubic structure for La
0.8Sr
0.2Ga
0.8Mg
0.2O
3−δ [73] are responsible for the aforementioned variations in their thermal expansion behaviour.
Datta et al.
[98] observed that the temperature of phase transition from an orthorhombic to a rhombohedral structure for La
1−xSr
xGa
1−yMg
yO
3−δ increased as Mg content increased at a fixed Sr content and decreased with increasing Sr content at a fixed Mg content. The effect of Sr and Mg co-doping on TEC values was explained for La
1−xSr
xGa
1−yMg
yO
3−δ in terms of the amount of generated oxygen vacancies. It was concluded that TEC values increased as oxygen vacancies increase, regardless of the dopant type. This was the result of the binding energy weakening as a result of oxygen vacancy formation.
Shkerin et al.
[99] analysed the structure and phase transitions of La
0.88Sr
0.12Ga
0.82Mg
0.18O
3−δ using dilatometry, XRD and Raman spectroscopy. According to the obtained data, La
0.88Sr
0.12Ga
0.82Mg
0.18O
3−δ exhibited two phase transitions of the second order at 502 and 607 °C. The first transition was attributed to a phase transition from an orthorhombic phase to a cubic one, while the second phase transition was attributed to the ordering of the oxygen vacancies.
Wu et al.
[100] studied transport properties of La
0.85Sr
0.15Ga
0.8Mg
0.2O
3−δ upon the partial or full Sr-substitution with calcium or barium. Their analyses have shown that both types of substitution result in a decrease in ionic conductivity by 20–30%. However, at the same time, the Ca-substituted ceramic materials showed higher conductivities compared to the Ba-substituted analogues. This confirms that strontium is an ideal dopant (from the steric and energetic viewpoints) to be introduced into the La-sublattice of LaGaO
3-based phases.
Chemical interactions between a La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ electrolyte and cathode materials such as La
0.65Sr
0.3MnO
3−δ, La
0.7Sr
0.3CoO
3−δ, La
0.65Sr
0.3FeO
3−δ, La
0.65Sr
0.3NiO
3−δ and La
0.6Sr
0.4Co
0.2Fe
0.8O
3−δ are demonstrated in
[101]. The LSGM/cathode powders were mixed at a weight ratio of 1:1, pressed into disks and annealed at 1300 °C for 3 h in air. The XRD data revealed that impurity phases were not formed in the LSGM mixed with La
0.65Sr
0.3MnO
3−δ, La
0.7Sr
0.3CoO
3−δ, and La
0.65Sr
0.3FeO
3−δ, but appear in the calcined mixtures with La
0.65Sr
0.3NiO
3−δ and La
0.6Sr
0.4Co
0.2Fe
0.8O
3−δ. The absence of reactivity between La
0.8Sr
0.2Ga
0.8Mg
0.2O
3−δ and La
0.8Sr
0.2MnO
3−δ was also confirmed during calcination at 800 °C
[102].
Sydyknazar et al.
[103] showed that La
0.83Sr
0.17Ga
0.8Mg
0.2O
3−δ exhibited good chemical compatibility with a novel cathode material, Sr
0.9Ba
0.1Co
0.95Ru
0.05O
3−δ, after joint calcination at 1100 °C for 12 h. According to the literature, La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ does not react with the following cathodes: La
0.4Sr
0.6Co
0.9Sb
0.1O
3−δ after heat treatment at 1150 °C for 6 h
[104], SrCo
0.8Fe
0.1Nb
0.1O
3−δ at 950 °C for 10 h
[105], BaCo
0.7Fe
0.2Ta
0.1O
3−δ at 950 °C for 10 h
[106] and Sr
2Ti
0.8Co
0.2FeO
6−δ after at 950 °C for 10 h
[107]. According to Tarancón et al.
[108], La
0.8Sr
0.2Ga
0.8Mg
0.2O
3−δ interacted with a GdBaCo
2O
5+δ cathode at temperatures above 900 °C, forming BaLaGa
3O
4 and BaLaGa
3O
7 secondary phases.
An analysis of works devoted to Ruddlesden–Popper phases demonstrates that La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ and Pr
2−xLa
xNi
0.85Cu
0.1Al
0.05O
4+δ (
x = 0, 0.2, 0.5, 1.0) have no interactions at 1000 °C for 5 h
[109], but La
0.95Sr
0.05Ga
0.9Mg
0.1O
3−δ reacted with Nd
2NiO
4+δ after annealing at 1000 °C for 5 h
[110]. Equally, La
0.85Sr
0.15Ga
0.85Mg
0.15O
3−δ reacted with Pr
2−xCa
xNiO
4+δ after annealing at 900 °C for 10 h (
x = 0, 0.5)
[111] and at 1200 °C for 1 h (
x = 0, 0.3)
[112].
Zhang et al.
[114] showed that a La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ electrolyte reacted with the nickel component in a Ni-SDC anode. The chemical interaction between LSGM and the composite was due to the interface diffusion of nickel from the anode to the LSGM electrolyte; this led to the formation of La-based poor-conductive secondary phases, which block oxygen-ion transport. The unit cell design with a buffer layer of SDC was suggested as an effective way of avoiding the problem of interface diffusion
[115]. However, chemical reactivity was observed between La
1−xSr
xGa
1−yMg
yO
3−δ and buffer layers of Gd
0.1Ce
0.9O
1.95, scandia-doped zirconia
[116] and Gd
0.8Ce
0.2O
1.9 [117].
An alternate solution to the problem of nickel interface diffusion from a Ni-based anode is to find novel anode materials. A study of the chemical compatibility between La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ and Fe
2O
3, Co
2O
3, NiO as anode materials is provided in
[118]. Powder mixtures of LSGM with metal oxides at a weight ratio of 1:1 were mixed in ethanol, pressed into pellets and annealed at 1150, 1250 and 1350 °C for 2 h. The obtained XRD data showed that the LSGM reacted with NiO and Co
2O
3 at 1150 °C, while a detectable reaction with Fe
2O
3 occurred only after calcination at 1350 °C.
Du and Sammes
[119] reported good chemical compatibility between La
0.8Sr
0.2Ga
0.8Mg
0.2O
3−δ and an alternative La
0.75Sr
0.25Cr
0.5Mn
0.5O
3 anode at a temperature range of 1100–1500 °C. However, the researchers note that a low-conductivity phase formed if the annealing time was more than 6 h or the annealing temperature was greater than 1500 °C.
Good chemical compatibility between LSGM and anodes with a double perovskite structure was shown for: La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ and Sr
2TiMoO
6−δ after calcining the samples at 1000 °C for 10 h in an atmosphere of 5% H
2/Ar
[120], La
0.8Sr
0.2Ga
0.8Mg
0.2O
3−δ and Sr
2Fe
1.5Mo
0.5O
6−δ after heat treatment at 1200 °C for 24 h in air
[121], La
0.88Sr
0.12Ga
0.82Mg
0.18O
3−δ with Sr
2NiMoO
6−δ at 1000 °C for 20 h
[122][123] and Sr
2Ni
0.75Mg
0.25MoO
6−δ at 1100 °C for 20 h
[124] and at 1250 °C for 2 h
[123]. The formation of secondary phases between LSGM and double perovskite anodes was observed for La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ and Sr
2MgMoO
6−δ after calcining at 1100 °C
[125], for La
0.88Sr
0.12Ga
0.82Mg
0.18O
3−δ and Sr
2ZnMoO
6 at 1000 °C for 20 h
[126] and for La
0.8Sr
0.2Ga
0.8Mg
0.2O
3−δ at 1300 °C for 10 h with Sr
2Ni
0.7Mg
0.3MoO
6−δ [127] and, after heat treatment at 1200 °C for 24 h, with Sr
2CoMoO
6−δ [121], Sr
2NiMoO
6−δ [121] and Sr
2MgMoO
6−δ [128].
According to Takano et al.
[125], La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ did not react with Ce
0.8La
0.2O
1.8 after annealing at 1300 °C for 1 h; therefore, it was concluded that La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ and Ce
0.8La
0.2O
2−δ might be recommended as SOFC electrolyte and buffer materials, respectively, with Sr
2MgMoO
6−δ used as the anode material. However, a comprehensive investigation of the chemical compatibility between various compositions of La
1−xSr
xGa
1−yMg
yO
3−δ and lanthanum-doped CeO
2, provided in
[129], showed that only a La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ/Ce
0.6La
0.4O
2−δ mixture did not result in additional phases after being annealed twice at 1350 °C for 2 h at each stage.
4. Applications in SOFCs
The problem of reactivity between the LSGM and SOFC electrode materials during sintering can be solved by reducing sintering temperatures or/and using the SDC buffer layer as a barrier, eliminating lanthanum- and nickel-cation diffusion. Several unit cell designs have been proposed in the literature.
Table 2 presents a summary of electrochemical performances for different types of hydrogen-fuelled SOFCs with LSGM-based electrolytes. These data testify that enhanced power densities were achieved for electrolyte-supported SOFCs when the LSGM electrolyte thickness was in a range of 100–300 μm. Buffer layers of doped ceria were used between the electrolyte and anode: Ce
0.8Sm
0.2O
2−δ [104][105][109][115][120][127], Ce
0.8Gd
0.2O
2−δ [130] and Ce
0.6La
0.4O
2−δ [131][132].
One can see that the SOFCs’ power density tends to increase with a decrease in the electrolyte’s thickness (due to a corresponding decline in the ohmic resistance) despite the existence/absence of CeO2-based buffer layers. Nevertheless, the performance of the compared SOFCs varies greatly, even for close electrolyte thicknesses, indicating that other functional components (cermets, oxygen electrodes) have a significant effect on the achievable output characteristics.
In
[131], it was shown that the OCV values were equal to 1.07 and 1.15 V at 800 °C and 700 °C, respectively, and there was no significant difference in the thickness of the Ce
0.6La
0.4O
1.8 interlayer. This LSGM-supported cell yielded up to 2200 and 1350 mW cm
−2 at 850 and 800 °C, respectively. The typical
I–
V curve and power densities at different temperatures for the LSGM-supported cell are shown, which is based on the Ni-Ce
0.8Gd
0.2O
2−δ/Ce
0.8Gd
0.2O
2−δ/(La
0.9Sr
0.1)
0.97Ga
0.9Mg
0.1O
3−δ/La
0.6Sr
0.4Fe
0.8Co
0.2O
3−δ cell tested in
[130]. The maximum power density of the aforementioned cell reached 540 mW cm
−2 at 800 °C, while the maximum power density of a cell containing a La
0.9Sr
0.1Ga
0.9Mg
0.1O
2.9 electrolyte reached 450 mW cm
−2 at 800 °C. The electrode polarization resistance values of the La
0.9Sr
0.1Ga
0.9Mg
0.1O
3−δ and (La
0.9Sr
0.1)
0.97Ga
0.9Mg
0.1O
3−δ based cells were equal to 0.34 and 0.30 Ω cm
2 at 800 °C, respectively.
Table 2 shows that, for electrode-supported SOFCs with thin-film LSGM electrolytes, a barrier layer between the electrolyte and the electrodes is not necessary
[134][135][136][144][145]. An anode-supported cell containing a La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ film deposited on an anode supported substrate using radio-frequency magnetron sputtering was fabricated in
[134]. The anode substrate was composed of a Ni-Sm
0.2Ce
0.8O
2−δ functional layer and a Ni collector layer; an LSGM-La
0.6Sr
0.4Co
0.2Fe
0.8O
3−δ composite layer was used as a cathode. The obtained SOFC revealed no cracking, delamination or discontinuity. The polarization resistance of an anode-supported cell containing a La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ film decreased from 0.41 to 0.05 Ω cm
2 as the temperature increased from 600 to 800 °C. The OCV and
Pmax values were in the range of 0.85–0.95 V and 650-1420 mW cm
−2, respectively, at a temperature range of 600–750 °C.
Combining the two approaches for SOFC design can be found in
[138][139][140][141]. Bi et al. deposited a Ce
0.6La
0.4O
2−δ(LDC)/LSGM bi-layer film on a Ni-Ce
0.9Gd
0.1O
2−δ anode. Therefore, the cell design allowed for high OCVs (1.02 and 1.043 V at 800 °C) and high power density values (1100 and 1565 mW cm
−2 at 800 °C) to be achieved at a LDC/LSGM bi-layer thickness of 100 and 65 μm, respectively
[138][139]. Ju et al.
[141] reached a paramount performance of 1790 mW cm
−2 at 700 °C for a SOFC based on an LSGM film with a thickness of 6 μm: this used an SDC buffer layer with a thickness of 500 nm, which was deposited on a Ni–Fe porous anode support. After a thermal cycle going from 700 to 25 °C, the fabricated cell showed an OCV of 1.1 V and
Pmax of 1620 mW cm
−2, which was almost the same as the first cycles.
According to a number of investigations
[139][142][143][158], the most effective design for SOFCs composed of barrier layers is the LDC/LSGM/LDC tri-layered electrolyte. Bi et al. reported
[139] that an anode-supported SOFC with an LDC/LSGM/LDC tri-layered electrolyte film significantly increased when using a cell with an LDC/LSGM bi-layered electrolyte film with the same thickness
[138]. Guo et al.
[143], depositing an LDC/LSGM/LDC tri-layer with thickness of 30 μm on a Ni-Ce
0.8Sm
0.2O
2−δ anode, fabricated a cell with a 75 mL min
−1 H
2 flow rate that generated 1230 W cm
−2 at 800 °C. The specific ohmic resistance across the LDC/LSGM/LDC tri-layer electrolyte film was measured to be equal to 0.086 Ω cm
2 at 800 °C. The obtained data showed that the polarization resistance was higher than the ohmic resistance at temperatures below 700 °C. A long-term stability experiment was performed on the aforementioned cell with a current density of 1000 mA cm
−2 and a 30 mL min
−1 H
2 flow rate at 800 °C. The results of 95 h-test demonstrated that the maximum power density values decreased from 1.08 to 0.81 W cm
−2. Researchers of
[143] suggest that there was little diffusion of the transition metal from the electrodes to the electrolyte during the test.
Serious efforts have been made to replace traditional cermet anodes with single-phase oxide materials: this is in an attempt to avoid chemical interactions. Complex oxides with double perovskite (Sr
2MMoO
6−δ (M = Mg, Ti, Ni, Fe)
[120][122][127][128][146][147][151][156]), layered
[150][153] and perovskite
[132][149] structures were successfully tested as alternative anode materials for SOFCs with LSGM electrolytes. A buffer layer of doped ceria was used to avoid chemical interactions between an LSGM electrolyte and double perovskites
[120][127][128], as well as between an LSGM electrolyte and an oxide cathode
[123][147][159]. The composite electrodes Sr
2Fe
1.5Mo
0.5O
6−δ-La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ [151], Sr
2CoMoO
6−δ-La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ and Sr
2Co
0.9Mn
0.1NbO
6−δ-La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ [160] have been proposed to solve the thermo-mechanical incompatibility between an electrolyte and an electrode due to a mismatch in the materials’ thermal expansion
[134][142][143][144][151][160][161][162][163].
An analysis of recent studies illustrates that LSGM can be used as a base matrix for the formation of both composite electrodes and new composite electrolytes
[160][164][165][166][167][168][169][170]. Xu et al.
[160] fabricated a cell based on a La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ-Ce
0.8Gd
0.2O
1.9 electrolyte, with Sr
2CoMoO
6−δ-La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ as the anode and Sr
2Co
0.9Mn
0.1NbO
6−δ-La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ as the cathode. For this cell, obtained with a 95 wt.% La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ-5 wt.% Ce
0.8Gd
0.2O
2−δ electrolyte, the OCV,
Pmax and current density values at 800 °C were equal to 1.08 V, 192 mW cm
−2, and 720 mA cm
−2, respectively
[160].
The electrochemical investigations in
[171][172][173][174][175] for LSGM-based SOFCs confirm that these cells can operate in both fuel cell and electrolysis cell modes. Reversible cells were fabricated in
[175] with NiO–YSZ-substrate as an anode, La
0.9Sr
0.1Ga
0.8Mg
0.2O
3−δ film as an electrolyte and Sm
0.5Sr
0.5CoO
3−δ as an air electrode. It was established that the infiltration of cerium nitrate into the substrate was an effective means of increasing cell performance. The maximum power density of this cell at 3 M Ce nitrate infiltration achieved 950 mW cm
−2 at 600 °C.
5. Conclusions
Complex oxides based on LaGaO3 offer a convenient basis for designing oxygen-conducting electrolytes that can be employed in intermediate-temperature solid oxide fuel cells (SOFCs). A rational combination of appropriate dopants incorporated at various sublattices of LaGaO3 allows for co-doped derivatives (La1–xSrxGa1–yMgyO3−δ, LSGM) with superior transport properties. LSGM materials are considered one of the most conductive oxygen-ionic electrolytes, enabling a decrease in SOFC operation temperatures by 100–300 °C compared to YSZ-based SOFCs. As a result, very high SOFC performances (from 0.5 to 1.5 W cm–2 at 700 °C) were reported for lab-type electrochemical cells. However, to efficiently place laboratory studies on a manufacturing scale, several issues remain, including the development of simple and low-cost technologies for electrolyte preparation (including thin-film forms), searching for strategies to improve the chemical stability of LSGM with other SOFC components (especially, with nickel) and the design of new electrochemically active electrodes. In this regard, the studies serves as the starting point for further research in fields like solid state chemistry, physical chemistry, electrochemistry and the technology of LaGaO3-based materials and electrochemical cells.