Article pubs.acs.org/cm
Atmosphere Controlled Processing of Ga-Substituted Garnets for High Li-Ion Conductivity Ceramics Carlos Bernuy-Lopez,*,† William Manalastas, Jr.,† Juan Miguel Lopez del Amo,*,† Ainara Aguadero,*,†,‡ Frederic Aguesse,† and John A. Kilner†,‡ CIC Energigune, Parque Tecnológico de Á lava, 48, 01510 Miñano, Á lava, Spain Department of Materials, Imperial College, London SW7 2AZ, U.K.
† ‡
S Supporting Information *
ABSTRACT: Ga-substituted La3Zr2Li7O12 garnet is shown to be a promising Li-ion conducting electrolyte material. The strategy adopted in this study is the substitution of Li by Ga, thereby creating Li vacancies and enhancing the Li conductivity. Solid State Magic Angle Spinning Nuclear Magnetic Resonance (MAS NMR) measurements have been used to identify the location of the substituted Ga in the structure and its effect on the Li distribution and mobility. In addition MAS NMR was used to follow the effect of protonation due to atmospheric moisture on the sintering behavior of these materials. In particular, it is shown that the Ga atoms are located in tetrahedral positions promoting the random distribution of lithium over the available sites, hence promoting an increase in the conductivity. Control of the sintering conditions by using a dry O2 atmosphere leads to the formation of dense ceramic materials and avoids the degradation process due to the exchange of Li+ by H+ from atmospheric moisture. Electrochemical Impedance Spectroscopy data show total conductivities as high as 1.3 and 2.2 mS cm−1 at 24 and 42 °C, respectively, which are among the highest Li ion conductivities reported for garnet-structured materials to date.
■
INTRODUCTION The continued development of Li-ion batteries depends upon the discovery of new materials to address the problems found with current technologies, of which safety is an important concern. In current Li-ion battery technology the liquid organic electrolytes show poor compatibility with electrodes and pose a safety hazard due to their flammability.1,2 These issues can be solved by the use of a solid electrolyte instead of a liquid, with the added advantage of shock and vibration resistance. Solid electrolytes also have the added capability of being compatible with thin film techniques necessary for applications involving miniaturization. At the same time, these electrolytes must be produced from abundant, cheap, and environmentally friendly elements with low processing costs. Recently very high ionic conductivity has been reported for Li10GeP2S12 with a total Liion conductivity as high as 12 mS cm−1 at 27 °C,3 the highest conductivity reported to date. Unfortunately, this material has a high cost due to the high price of some of its elements. Replacing Ge by Sn has been shown to give a total Li-ion conductivity as high as 4 mS cm−1 at 27 °C.4 Unfortunately, these compounds rapidly decompose forming toxic H2S.5 A good alternative to these materials are the garnet-type oxides with general formula A3B2C3O12 where the A-, B-, and Csites have dodecahedral, octahedral (Oh), and tetrahedral (Td) coordinations, respectively. The stoichiometric garnet (A3B2Li3O12) has a full occupancy6,7 of the Td crystallographic positions and shows a very low Li-ion conductivity. Increasing the Li content (Li > 3) promotes a reorganization of the Li-ions within the available crystallographic and interstitial positions © XXXX American Chemical Society
with a resultant increase of the Li-ion conductivity. La3Zr2Li7O12 has been shown to be a very good Li-ion conductor making it a potential solid electrolyte for Li batteries. 8 However, this material only presents good conductivity values when a cubic structure is stabilized instead of the more thermodynamically stable tetragonal structure. An obvious way to attempt to stabilize this cubic phase and to hence obtain larger conductivity values is by means of substitution strategies. Most reports of Li conducting materials follow processing steps performed in air. Air processing is clearly desirable if the cost and ease of processing of these materials is to be maintained. We will comment later on the effect of air processing; however, suffice it so say at this point that, unless stated, the materials discussed have been processed in a laboratory air atmosphere with all that this entails. The most common element used to stabilize the cubic phase is Al.9−12 This Al substitution is either produced intentionally or unintentionally by the combined use of high sintering temperatures, long sintering times, and alumina crucibles. It has been shown that without any Al substitution the cubic phase cannot be stabilized. However, only very small concentrations are needed to stabilize this cubic phase. This stabilization seems to originate from increased Li disorder between the different crystallographic sites due to Al-incorporation in the tetrahedral Received: December 5, 2013 Revised: May 20, 2014
A
dx.doi.org/10.1021/cm5008069 | Chem. Mater. XXXX, XXX, XXX−XXX
Chemistry of Materials
Article
of vacancies in the C-site of the garnet structure. Figure 1a shows the crystal structure of the La3Zr2Li7−3xGax□2xO12
24d site in the garnet structure. It has also been shown that a significant increase of concentration above the level needed for phase stabilization leads to a substitution in the Zr octahedral sites which might lead to phase decomposition and, therefore, a decrease in Li-ion conductivity. Substitution on the C-site (Td coordination) with Al has led to values of total conductivity as high as 0.49 mS cm−1 at 25 °C. Ga has been also used as a substituting element on the C-site although total conductivities as low as 0.02 mS cm−1 have been reported, due to a large grain boundary resistance.13 Alternatively, substitution strategies in the B-site (Oh coordination) of La3Zr2Li7O12 with Ta14 and Nb15 have recently been shown to improve the total Li-ion conductivity to 1 mS cm−1 and 0.8 mS cm−1 at 25 °C, respectively, which are the best reported conductivities to date, to the best of our knowledge. Simultaneous substitutions have also been reported in the literature,16 although conductivities remain at the levels mentioned above. It has been shown that the improvement of the density of garnet materials leads to an improvement in the total Li-ion conductivity. It seems rather difficult to stabilize the cubic phase and to obtain high density materials without any undesired impurities or the presence of the undesirable tetragonal phase. Expensive methods such as hot-press or spark plasma sintering lead to stable cubic phases and values close to 90% theoretical density although they are still far from the 98−99% value desired for applications.17 However, it seems possible to obtain very high densities for the tetragonal La3Zr2Li7O12 phase by means of the use of a hot-press as it has recently been shown by Wolfenstine et al.18 In a recent study, El Shinawi19 et al. reported a slightly different strategy; using Ga as a sintering aid rather than a substituent, obtaining densities close to 92%. In this work, they did not find any clear evidence of the insertion of Ga in the garnet structure, but this strategy led them to stabilize the cubic phase and to obtain total Li-ion conductivities as high as 0.5 mS cm−1 at 20 °C. Similar effects in the increase of density have been shown by Li et al.20 when using lithium oxide as additive in a Ta-substituted garnet. In this case, an increase of the relative density from 91.5% to 97.3% was shown that led them to conductivities of 0.64 mS cm−1. Furthermore, all these problems (i.e., processing of high density samples, high grain boundary resistance, etc.) in La3Zr2Li7O12 seem to be related with its reactivity with atmospheric moisture. A recent study by Larraz et al.21 has shown the different structural changes that this material can undergo when it is in contact with different levels of atmospheric moisture. The authors of this report claim that these changes are probably due to the high reactivity of the Li+ with the H+ from the water or the introduction of water molecules into the structure, which would lead to a large Li-ion conductivity decrease. A very recent report by Li et al.22 shows the improvement in the total Li-ion conductivity of an unintentionally Al-substituted La3Zr2Li7O12 material when the sintering was done under pure O2 showing conductivities as high as 0.74 mS cm−1 at 25 °C, although no precautions were taken to avoid contact with moisture. Therefore, in this work we have taken into account all the issues related to the stabilization of the cubic phase with high Li content and to the control of densification and contact with moisture to produce high Li-ion conducting materials. First, we aim to substitute Li by small amounts of Ga in the Td positions of the La3Zr2Li7O12 material in order to promote the creation
Figure 1. (a) Garnet structure of La3Zr2Li7−3xGax□2xO12 where the blue circles are Zr in octahedral (Oh) coordination, the dark red circles are La in dodecahedral coordination, the gray circles are Li in Oh coordination, the turquoise blue and green circles are Li and Ga in tetrahedral (Td) coordination, respectively, and the arrows represent a vacancy (□). (b) Vacancy representation (□) of an enlarged area of (a) (dotted lines) showing the Li pathway in the garnet structure.
garnet (□ represents generated vacancies; 0 ≤ x ≤ 0.3) where the Li path created by extra vacancy generation is shown in Figure 1b. Second, we aim to optimize the sintering conditions with minimal moisture contact for our samples, in particular using dry O2 as sintering gas. Following this scheme we have synthesized a series of Ga-substituted La3Zr2Li7O12 materials with the general formula La3Zr2Li7−3xGax□2xO12 and studied in detail the effects of atmospheric moisture during processing on the Li-ion conductivity.
■
EXPERIMENTAL SECTION
Syntheses and Sintering of the Materials. La3Zr2Li7−3xGax□2xO12 (x = 0 (LZLO), x = 0.150 (LZLGO_150), 0.200 (LZLGO_200), and 0.300 (LZLGO_300)) were synthesized in 10 g batches, by a citric acid-nitrate route. The starting materials were Ga2O3 (≥99.99%, Sigma-Aldrich), La(NO3)3 (≥99.99%, SigmaAldrich), Zr(C5H7O2)4 (>98%, Alfa Aesar), and LiNO3 (>99.0%, Sigma-Aldrich)). These materials were mixed in stoichiometric quantities (+10% Li excess) and consequently dissolved in a citric acid solution with some drops of HNO3. The resulting gel was fired at 600 °C for 12 h in air to burn off the organic components. After this treatment, the powder was ground and reheated to 800 °C for 12 h in B
dx.doi.org/10.1021/cm5008069 | Chem. Mater. XXXX, XXX, XXX−XXX
Chemistry of Materials
Article
metal grade, Fischer Scientific) and subsequently diluted in deionized water (Millipore, registering 13.1 MΩcm at 25 °C).
dry O2 to obtain single phase powders. In order to measure the Li-ion conductivity of these materials, the synthesized powders were uniaxially pressed into 6 mm in diameter pellets and sintered in dry O2 at 1085 °C for 6 h. Reaction and sintering steps were carried out in alumina boats. The pellets were thoroughly buried in identical powder to mitigate losses of volatile components and accidental contamination. For comparison purposes, all the synthesized and sintered materials were also produced in ambient air. Relative densities of the sintered pellets were obtained geometrically after the final sintering step at 1085 °C in either dry O2 or air. The O2 sintered samples were quickly moved into a glovebox in order to avoid contact with atmospheric moisture. Values no higher than 70% were obtained for the samples sintered in air; however, values as high as 94% of theoretical density were obtained for the samples sintered in dry O2, close to those reported in the literature for the most conductive garnet materials.9,10 It is important to note at this point the need to minimize as much as possible the contact with atmospheric moisture of the samples sintered in dry O2. All the synthesized powders and pellets processed in dry O2 were kept in a glovebox under an Ar atmosphere with very low water and oxygen levels (
