Enriching solid-state batteries
Researchers at MIT have come up with a new pulsed laser deposition technique to make thinner lithium electrolytes using less heat, promising faster charging and potentially higher-voltage solid-state lithium ion batteries.
Key to the new technique for processing the solid-state battery electrolyte is alternating layers of the active electrolyte lithium garnet component (chemical formula, Li6.25Al0.25La3Zr2O12, or LLZO) with layers of lithium nitride (chemical formula Li3N). First, these layers are built up like a wafer cookie using a pulsed laser deposition process at about 300 degrees Celsius (572 degrees Fahrenheit). Then they are heated to 660 C and slowly cooled, a process known as annealing.
During the annealing process, nearly all of the nitrogen atoms burn off into the atmosphere and the lithium atoms from the original nitride layers fuse into the lithium garnet, forming a single lithium-rich, ceramic thin film. The extra lithium content in the garnet film allows the material to retain the cubic structure needed for positively charged lithium ions (cations) to move quickly through the electrolyte. The findings were reported in a Nature Energy paper published online recently by MIT Associate Professor Jennifer L. M. Rupp and her students Reto Pfenninger, Michal M. Struzik, Inigo Garbayo, and collaborator Evelyn Stilp.
“The really cool new thing is that we found a way to bring the lithium into the film at deposition by using lithium nitride as an internal lithiation source,” Rupp, the work's senior author, says. Rupp holds joint MIT appointments in the departments of Materials Science and Engineering and Electrical Engineering and Computer Science.
“The second trick to the story is that we use lithium nitride, which is close in bandgap to the laser that we use in the deposition, whereby we have a very fast transfer of the material, which is another key factor to not lose lithium to evaporation during a pulsed laser deposition,” Rupp explains.
Lithium batteries with commonly used electrolytes made by combining a liquid and a polymer can pose a fire risk when the liquid is exposed to air. Solid-state batteries are desirable because they replace the commonly used liquid polymer electrolytes in consumer lithium batteries with a solid material that is safer. “So we can kick that out, bring something safer in the battery, and decrease the electrolyte component in size by a factor of 100 by going from the polymer to the ceramic system,” Rupp explains.
Although other methods to produce lithium-rich ceramic materials on larger pellets or tapes, heated using a process called sintering, can yield a dense microstructure that retains a high lithium concentration, they require higher heat and result in bulkier material. The new technique pioneered by Rupp and her students produces a thin film that is about 330 nanometers thick (less than 1.5 hundred-thousandths of an inch). “Having a thin film structure instead of a thick ceramic is attractive for battery electrolyte in general because it allows you to have more volume in the electrodes, where you want to have the active storage capacity. So the holy grail is be thin and be fast,” she says.
Compared to the classic ceramic coffee mug, which under high magnification shows metal oxide particles with a grain size of tens to hundreds of microns, the lithium (garnet) oxide thin films processed using Rupp’s methods show nanometer scale grain structures that are one-thousandth to one-ten-thousandth the size. That means Rupp can engineer thinner electrolytes for batteries. “There is no need in a solid-state battery to have a large electrolyte,” she says.
Faster ionic conduction
Instead, what is needed is an electrolyte with faster conductivity. The unit of measurement for lithium ion conductivity is expressed in Siemens. The new multilayer deposition technique produces a lithium garnet (LLZO) material that shows the fastest ionic conductivity yet for a lithium-based electrolyte compound, about 2.9 x 10-5 Siemens (0.000029 Siemens) per centimeter. This ionic conductivity is competitive with solid-state lithium battery thin film electrolytes based on LIPON (lithium phosphorus oxynitride electrolytes) and adds a new film electrolyte material to the landscape.
“Having the lithium electrolyte as a solid-state very fast conductor allows you to dream out loud of anything else you can do with fast lithium motion,” Rupp says.
A battery’s negatively charged electrode stores power. The work points the way toward higher-voltage batteries based on lithium garnet electrolytes, both because its lower processing temperature opens the door to using materials for higher voltage cathodes that would be unstable at higher processing temperatures, and its smaller electrolyte size allows physically larger cathode volume in the same battery size.
Co-authors Michal Struzik and Reto Pfenninger carried out processing and Raman spectroscopy measurements on the lithium garnet material. These measurements were key to showing the material’s fast conduction at room temperature, as well as understanding the evolution of its different structural phases.
“One of the main challenges in understanding the development of the crystal structure in LLZO was to develop appropriate methodology. We have proposed a series of experiments to observe development of the crystal structure in the [LLZO] thin film from disordered or 'amorphous' phase to fully crystalline, highly conductive phase utilizing Raman spectroscopy upon thermal annealing under controlled atmospheric conditions,” says co-author Struzik, who was a postdoc working at ETH Zurich and MIT with Rupp’s group, and is now a professor at Warsaw University of Technology in Poland. “That allowed us to observe and understand how the crystal phases are developed and, as a consequence, the ionic conductivity improved,” he explains.
Their work shows that during the annealing process, lithium garnet evolves from the amorphous phase in the initial multilayer processed at 300 C through progressively higher temperatures to a low conducting tetragonal phase in a temperature range from about 585 C to 630 C, and to the desired highly conducting cubic phase after annealing at 660 C. Notably, this temperature of 660 C to achieve the highly conducting phase in the multilayer approach is nearly 400 C lower than the 1,050 C needed to achieve it with prior sintering methods using pellets or tapes.
“One of the greatest challenges facing the realization of solid-state batteries lies in the ability to fabricate such devices. It is tough to bring the manufacturing costs down to meet commercial targets that are competitive with today's liquid-electrolyte-based lithium-ion batteries, and one of the main reasons is the need to use high temperatures to process the ceramic solid electrolytes,” says Professor Peter Bruce, the Wolfson Chair of the Department of Materials at Oxford University, who was not involved in this research.
“This important paper reports a novel and imaginative approach to addressing this problem by reducing the processing temperature of garnet-based solid-state batteries by more than half — that is, by hundreds of degrees,” Bruce adds. “Normally, high temperatures are required to achieve sufficient solid-state diffusion to intermix the constituent atoms of ceramic electrolyte. By interleaving lithium layers in an elegant nanostructure the authors have overcome this barrier.”
After demonstrating the novel processing and high conductivity of the lithium garnet electrode, the next step will be to test the material in an actual battery to explore how the material reacts with a battery cathode and how stable it is. “There is still a lot to come,” Rupp predicts.
Understanding aluminum dopant sites
A small fraction of aluminum is added to the lithium garnet formulation because aluminum is known to stabilize the highly conductive cubic phase in this high-temperature ceramic. The researchers complemented their Raman spectroscopy analysis with another technique, known as negative-ion time-of-flight secondary ion mass spectrometry (TOF-SIMS), which shows that the aluminum retains its position at what were originally the interfaces between the lithium nitride and lithium garnet layers before the heating step expelled the nitrogen and fused the material.
“When you look at large-scale processing of pellets by sintering, then everywhere where you have a grain boundary, you will find close to it a higher concentration of aluminum. So we see a replica of that in our new processing, but on a smaller scale at the original interfaces,” Rupp says. “These little things are what adds up, also, not only to my excitement in engineering but my excitement as a scientist to understand phase formations, where that goes and what that does,” Rupp says.
“Negative TOF-SIMS was indeed challenging to measure since it is more common in the field to perform this experiment with focus on positively charged ions,” explains Pfenninger, who worked at ETH Zurich and MIT with Rupp’s group. “However, for the case of the negatively charged nitrogen atoms we could only track it in this peculiar setup. The phase transformations in thin films of LLZO have so far not been investigated in temperature-dependent Raman spectroscopy — another insight towards the understanding thereof.”
The paper’s other authors are Inigo Garbayo, who is now at CIC EnergiGUNE in Minano, Spain, and Evelyn Stilp, who was then with Empa, Swiss Federal Laboratories for Materials Science and Technology, in Dubendorf, Switzerland.
Rupp began this research while serving as a professor of electrochemical materials at ETH Zurich (the Swiss Federal Institute of Technology) before she joined the MIT faculty in February 2017. MIT and ETH have jointly filed for two patents on the multi-layer lithium garnet/lithium nitride processing. This new processing method, which allows precise control of lithium concentration in the material, can also be applied to other lithium oxide films such as lithium titanate and lithium cobaltate that are used in battery electrodes. “That is something we invented. That’s new in ceramic processing,” Rupp says.
“It is a smart idea to use Li3N as a lithium source during preparation of the garnet layers, as lithium loss is a critical issue during thin film preparation otherwise,” comments University Professor Jürgen Janek at Justus Liebig University Giessen in Germany. Janek, who was not involved in this research, adds that “the quality of the data and the analysis is convincing.”
“This work is an exciting first step in preparing one of the best oxide-based solid electrolytes in an intermediate temperature range,” Janek says. “It will be interesting to see whether the intermediate temperature of about 600 degrees C is sufficient to avoid side reactions with the electrode materials.”
Oxford Professor Bruce notes the novelty of the approach, adding “I'm not aware of similar nanostructured approaches to reduce diffusion lengths in solid-state synthesis.”
“Although the paper describes specific application of the approach to the formation of lithium-rich and therefore highly conducting garnet solid electrolytes, the methodology has more general applicability, and therefore significant potential beyond the specific examples provided in the paper,” Bruce says. Commercialization may be needed to be demonstrate this approach at larger scale, he suggests.
While the immediate impact of this work is likely to be on batteries, Rupp predicts another decade of exciting advances based on applications of her processing techniques to devices for neuromorphic computing, artificial intelligence, and fast gas sensors. “The moment the lithium is in a small solid-state film, you can use the fast motion to trigger other electrochemistry,” she says.
Several companies have already expressed interest in using the new electrolyte approach. “It’s good for me to work with strong players in the field so they can push out the technology faster than anything I can do,” Rupp says.
A low ride on processing temperature for fast lithium conduction in garnet solid-state battery films
Reto Pfenninger, Michal Struzik, Iñigo Garbayo, Evelyn Stilp & Jennifer L. M. Rupp
Nature Energy, volume 4, pages 475–483 (2019)
Assoc. Prof. in Materials Science and Engineering & Assoc. Prof. in the Department of Electrical Engineering and Computer Science at MIT
Massachusetts Institute of Technology (MIT)