INTERNATIONAL JOURNAL OF ELECTRIC AND HYBRID VEHICULES
I.J.E.H.V.
contact : b.maisseu@ehcar.net
I.J.E.H.V.
Conferences
Congress - Meeting
Library
News
News Records

2017-05-08

A Rechargeable Li-Air Fuel Cell Battery Based on Garnet Solid Electrolytes


Figure 1: Schematic image and first cycle performance.
(a) Schematic image and (b) cross-section image of a typical SSLAB. (c,d) Typical first discharge and charge cycles of SSLABs consisting of PI:LiTFSI and PPC:LiTFSI at 200?C and 80?C, respectively.

Non-aqueous Li-air batteries have been intensively studied in the past few years for their theoretically super-high energy density. However, they cannot operate properly in real air because they contain highly unstable and volatile electrolytes. Here, we report the fabrication of solid-state Li-air batteries using garnet (i.e., Li6.4La3Zr1.4Ta0.6O12, LLZTO) ceramic disks with high density and ionic conductivity as the electrolytes and composite cathodes consisting of garnet powder, Li salts (LiTFSI) and active carbon. These batteries run in real air based on the formation and decomposition at least partially of Li2CO3. Batteries with LiTFSI mixed with polyimide (PI:LiTFSI) as a binder show rechargeability at 200?C with a specific capacity of 2184?mAh g-1carbon at 20?A cm-2. Replacement of PI:LiTFSI with LiTFSI dissolved in polypropylene carbonate (PPC:LiTFSI) reduces interfacial resistance, and the resulting batteries show a greatly increased discharge capacity of approximately 20300?mAh g-1carbon and cycle 50 times while maintaining a cutoff capacity of 1000?mAh g-1carbon at 20?A cm-2 and 80?C. These results demonstrate that the use of LLZTO ceramic electrolytes enables operation of the Li-air battery in real air at medium temperatures, leading to a novel type of Li-air fuel cell battery for energy storage.

Introduction

Rechargeable Li-air (Li-O2) batteries have attracted intensive research interest in the past few years since they can achieve a much greater energy density than other electrochemical storage devices, which is now in high demand with the rapid development of extended-range electric vehicles and energy storage applications. Li-O2 batteries based on non-aqueous electrolytes have been widely investigated because of their rechargeability and potential cycle performance. However, these batteries face critical challenges for operation in air due to the influences of moisture and carbon dioxide in air, exhausting of the liquid electrolyte in open cells, attacks of peroxide or superoxide on the electrolyte and carbon, sluggish kinetics for oxygen reduction reactions (ORR) and oxygen evolution reactions (OER) at room temperature, as well as dendrite formation of lithium on the anode side. One potential solution is to replace the liquid electrolyte with a solid-state electrolyte that may protect the Li anode and suppress formation of the lithium dendrite, allowing operation of the Li-air battery in real air without destruction of the electrolyte. However, studies on solid-state Li-air batteries (SSLAB) are still in their infancy, mainly due to the lack of available solid electrolytes. So far, several solid-state electrolytes have been used to fabricate SSLAB, such as the NASICON-type lithium-ion conducting ceramics Li-Al-Ge-PO4 (LAGP)22 and Li-Al-Ti-PO4 (LATP)23,24 and polymer electrolytes such as polyethylene oxide (PEO)25. However, inorganic LAGP and LATP ceramic electrolytes become unstable when they come into contact with the lithium metal anode26,27. As a result, a protective layer must be added to prevent this reaction, which increases the internal resistance of the SSLAB. Polymer electrolytes such as polyethylene oxides (PEO) mixed with LiX (X, anion) are normally unstable at high potentials and are not resistant to the attack of O2- species.
A new class of solid-state electrolytes, garnet-type Li7La3Zr2O12 (LLZO) ceramics, were first reported by Murugan et al. in 2007. LLZO ceramics have several promising properties, including chemical stability against Li metal, large electrochemical windows (above 5?V), and high Li+ conductivity (above 10-4?S cm-1 at room temperature). One problem associated with LLZO is the difficulty in sintering high-density ceramics due to the volatility of Li. Recently, through hot-press sintering and optimized processes, we successfully obtained lamellar Li6.4La3Zr1.4Ta0.6O12 (LLZTO) electrolytes with densities as high as 99.6% and improved the ionic conductivity to 1.6??10-3?S cm-1 at room temperature (as depicted in Figure S1). Based on these LLZTO disk electrolytes (~0.1?cm in thickness), we designed a new type of SSLAB architecture. An air cathode mainly composed of Ketjen black (KB), LLZTO particles and LiTFSI in polyimide (PI:LiTFSI) or polypropylene carbonate (PPC:LiTFSI) was coated on one side of the LLZTO disk. The metallic Li anode was pressed on the other side of the LLZTO disk, which was subsequently sealed airtight. The operating temperature was elevated to improve the interface contact between the Li and LLZTO as well as between the cathode and LLZTO. These two-part SSLABs can be operated in real air with good rechargeability. By replacing PI:LiTFSI with PPC:LiTFSI as the conductive binder, the batteries can be operated at reduced temperatures while displaying improved performance.

Results

Figure shows the schematic configuration of a typical SSLAB. A 0.1-cm-thick lamellar LLZTO disk is used as the electrolyte. On the anode side, a Li foil is attached on the LLZTO, which is sealed in a stainless steel container with high-temperature resistive sealant. This part can withstand up to 250?C as indicated by the thermogravimetric-differential scanning calorimetry (TG-DSC) analysis. On the cathode side, 0.2-m LLZTO particles are used as the ion conducting framework. As demonstrated, the plate consisting of LLZTO particles and the PI:LiTFSI composite shows an ionic conductivity of 1.9??10-5 S cm-1 at 25?C and 3.5??10-5 at 80?C, while the PPC:LiTFSI counterpart shows 9.3??10-5?S cm-1 and 1.6??10-4?S cm-1, respectively. The KB used for the cathode is responsible for constructing the electronic network and the occurrence of both ORR and OER. Figure shows a cross-section image of a typical battery. Composition analysis of the air cathode by energy dispersive X-ray analysis (EDX) reveals that La, Zr and C are homogeneously distributed in the whole region. Figure shows the discharge and charge behaviours of the batteries with PI:LiTFSI as the binder. These batteries show rechargeability only at 200?C. Nevertheless, it is worth noting that the coulombic efficiency can be as high as 97.1%, which is likely due to the influence of elevated temperature on the Li anode and the interfacial resistance. To lower the operation temperature of the batteries, PI:LiTFSI is replaced with PPC:LiTFSI, which can reduce the interfacial resistance as discussed above. The resulting batteries can be cycled at 80?C with a current density of 20?A cm-2, as shown in Figure. After a large discharge plateau at approximately 2.78?V, the curve drops suddenly at a specific potential. The specific discharge capacity reaches approximately 20300?mAh g-1carbon, which is approximately 6430?mAh g-1cathode when being calculated according to the overall cathode mass. Upon charging, a plateau at 3.01?V with capacity of approximately 650?mAh g-1carbon is initially observed. Then, a large plateau at 3.87?V with a capacity of approximately 15000?mAh g-1carbon appears. To clarify the origin of the charge reaction at 3.87?V, the composites with PPC:LiTFSI and KB as the cathodes were charged in air at 80?C (curve displayed in figure). These composites display decomposition only when the charge potential is greater than 4.45?V. This indicates that the capacity during charging at 3.87?V is solely related to decomposition of the discharge products rather than decomposition of the composite cathode.
The morphology and composition of the reaction products were characterized by both ex-situ and in-situ techniques. Figure shows SEM images of changes in the air cathode during cycling. Compared to the pristine cathode, the average size of granular particles increases at the half discharge state. After full discharge, rod-like particles can be observed, indicating that the discharge products grow upward and merge with each other to form agglomerated particles. Upon charging, the large particles identified in Figure clearly shrink, as shown in Figure. After the fifth discharge, rod-like particles similar to those formed at the initial discharge and during the shrinkage can also be observed, indicating repetitive formation and decomposition of the discharge product.


Morphology of the air cathode(a) Pristine air cathode, (b) first discharge to a capacity of ~10000?mAh g-1carbon, (c) first discharge to a capacity of ~20000?mAh g-1carbon (~2.0?V), (d) first charge to a capacity of 20000?mAh g-1 (~4.5?V), (e) fifth discharge to 2.0?V, and (f) fifth charge to 4.5?V at 20?A cm-2. The white scale bar represents 500?nm for all images.

Fourier transform infrared spectroscopy (FTIR) was also performed on the air cathodes after discharge and charge processes at 80?C. Considering that PPC has complex FTIR peaks between 2000?cm-1 and 1250?cm-1?33, where the characteristic peaks of Li2O2 and Li2CO3 lie, we used the batteries with PI:LiTFSI in the FTIR measurement. As shown in figure, the peak intensity corresponding to the Li2CO3 reference greatly increases after discharge, which indicates that the identified discharge product is Li2CO3. The same phenomenon was observed by transmission electron microscopy (TEM) of the air cathodes, in which carbon nanotubes (CNT) were used as the active material instead of KB to clearly characterize the product composition upon cell operation. As shown in figure, the Li2CO3 grows around the CNT during discharging and decomposes upon charging.


FTIR measurementFTIR spectra of air cathodes after the 1 st and 5th discharge (black lines) and 1st and 5th charge (red lines) at 80?C. Regions are highlighted by the dotted lines corresponding to the peaks of Li2CO3. The spectra of pristine air cathodes, pure Li2CO3 (olive line) and Li2O2 (dark cyan line) powders are provided for comparison.

To obtain more information on the change in chemical composition on the cathode surfaces, in-situ X-ray photoelectron spectroscopy (XPS) of the discharge and charge process was performed under ambient pressure of 0.1 mbar. The cathodes used in the in-situ XPS measurement were carbon materials that were deposited on the LLZTO disk by sputtering (details can be found in Methods section). As shown in figure, Li2O2 peak intensity first increases and then decreases upon discharge. Meanwhile, the Li2CO3 and Li-C-O peak intensities steadily increase with increasing depth of discharge. These results indicate that the Li2O2 is formed at the initial stage of discharge, and reacts with trace H2O and CO2 in ambient air to form a large quantity of Li2CO3 on the cathode surfaces. The appearance of the Li-O-C peak can be attributed to the decomposition of the carbon cathode (i.e., KB) when the cell is charged. Note that in-situ measurement allows the observation of transient products (i.e., Li2O2). The ex-situ measurement can only provide information on the stabilized products in air, such as Li2CO3. While charging, the peak intensity of Li2CO3 greatly decreases, indicating that the Li2CO3 is most probably decomposed upon charging, at least partially. Rechargeable Li-CO2 batteries based on the formation and decomposition of Li2CO3 have already been demonstrated in previous reports, which also indicated that the decomposition of Li2CO3 upon charge was dependent on the type of carbon materials in the cathodes. The KB carbons used here are most likely not the optimum components, causing partial decomposition of Li2CO3 upon charge and thus accumulation of passivated Li2CO3 products after multiple cycles.


In-situ XPS analysis.
In-situ XPS of O 1s collected under air atmosphere of 10-4?atm for the carbon cathodes during the first discharge/charge cycle of the SSLAB cell.

Based on the existing data and previous reports, we propose the following reaction route in the present Li-air battery. During discharging, O2 is electrochemically reduced to form Li2O2 (equation 1) in the cathode initially. Li2O2 then reacts with CO2 in air to form Li2CO3 via a chemical reaction (equation 2). During charging, Li2CO3 is electrochemically decomposed at least partially, according to the release of CO2 during charge.
In fact, preliminary DEMS (differential electrochemical mass spectrometry) results indicate that the reaction mechanism upon charging is very complicated. Further work with isotope labelling is necessary and ongoing in our laboratory.
According to the above analysis, the discharge products grow on the cathode surfaces during discharge and partially decompose during charging. This situation is similar to Li-O2 batteries based on Li2O2 formation and decomposition. The insulating deposits have a passivating influence on active sites for ORR. In the case of full discharge, a number of thick discharge products form and coat the carbon surfaces, which allows only a few cycles, as shown in figure To increase the cycle number, limiting the depth of discharge is necessary. As shown in figure, the batteries can be run for more than 50 cycles at 20?A cm-2 when the discharge capacity is cut off at 1000?mAh g-1carbon, which is approximately 316?mAh g-1cathode if calculated according to the overall cathode mass.


Cycle performance of SSLAB.
(a) The PPC:LiTFSI cell operated at 80?C in real air with a cutoff discharge capacity of 1000?mAh g-1carbon at 20?A cm-2 and (b) the specific capacity as a function of cycle number.

Discussion

Here we demonstrate the fabrication of rechargeable SSLABs that can be operated in real air at elevated temperatures. The newly synthesized LLZTO ceramic electrolytes play essential roles in these batteries. They are dense (99.6% in relative density), highly conducting (1.6??10-3?S?cm-1 at room temperature)42, and are able to prevent gas species from diffusing through and reacting with the Li anodes. LLZTO ceramic electrolytes are also chemically and electrochemically stable against Li, which allows direct contact with Li without introducing an additional resistive intermediate layer. Previous reports found that LLZTO might react with H2O in air, causing degradation of ionic conductivity, but the moisture content in air can be greatly reduced when the surrounding temperature is elevated to above 80?C. This means that the problem of LLZTO degradation can be greatly relieved at elevated temperatures.
Other important issues include construction of the air cathode and interfacial concerns. Air cathodes cannot be wetted in the solid-state electrolytes as they do in the liquid electrolytes. As a result, it is necessary to artificially build the electronic and ionic network. Because the discharge products are solids deposited on the cathode surfaces, sufficient space to accommodate these solids is also required for large capacities, which is why porous composite cathodes are needed. The LLZTO particles provide a framework that is ionically conducting and able to create space for air diffusion, while PPC:LiTFSI enhances ion conduction at the LLZTO interfaces.
The use of elevated temperatures is especially advantageous in terms of improving ion conduction at the Li/LLZTO and electrode/LLZTO interfaces. However, too high operating temperature may cause the decomposition of carbon-based cathodes. As a result, selecting suitable carbon materials, and optimizing cathode configuration to reduce the interfacial resistance must be considered to improve battery performance. Moreover, searching for suitable materials with effective catalysts that can promote the decomposition of Li2CO3 upon recharge is also crucial for the development of SSLABs.

Conclusions

A novel type of solid-state Li-air battery based on LLZTO ceramic electrolytes is demonstrated here. After optimizing the interfacial resistance, the batteries show a specific capacity of approximately 20300?mAh g-1carbon at 20?A cm-2 (or 6430?mAh g-1cathode if referring to the mass of cathode) in real air at 80?C. With a cutoff discharge capacity of 1000?mAh g-1carbon at 20?A cm-2, the batteries can be operated for 50 cycles while maintaining their initial capacity. Analysis of the cell chemistry reveals that the batteries operate through formation and decomposition of Li2CO3 at least partially. Cathode design related to the interfacial issue and materials research to promote decomposition of the discharge product are two key problems worth investigating in the near future. The performance achieved in the SSLABs presented here demonstrates the possibility of operating Li-air batteries in real air, which opens new avenues for the use of Li-air fuel cell battery for energy storage.