Researchers at the University of Michigan-Ann Arbor have combined recent developments in reducing interface resistance and optimizing microstructure to develop a solid electrolyte that demonstrates Li cycling at current densities competitive with Li-ion.
Next generation electric vehicles and consumer electronics demand volumetric energy densities beyond that capable of state-of-the-art Li-ion technology. All-solid-state-batteries (ASSBs) offer the potential for meeting these demands with energy densities > 1000 W h L−1 while improving safety. These high energy densities are enabled through the use of a metallic Li anode, with a volumetric capacity of 2 A h cm−3. Metallic Li has long been explored for use as a negative electrode in secondary batteries but the well-known challenge of stabilizing the Li-electrolyte interface has impeded widespread implementation. Solid electrolytes offer a prospective solution by physically stabilizing the Li-electrolyte interface.
The Monroe and Newman model predicts resistance to Li dendrite initiation if the shear modulus G of a solid electrolyte is greater than two times that of metallic Li (∼4.5 GPa). This criterion has been extended to and is easily met by most solid electrolytes, provided they are homogeneous ionic conductors. Yet, Li metal propagation occurs across nearly all known solid electrolytes in spite of their high shear moduli.
… While the mechanism that enables Li metal propagation in solid electrolytes is not well understood, understanding interface transport phenomena could guide approaches to increase the maximum tolerable current density. … The objective of this work was to integrate recent progress in understanding solid-state electrolyte interface resistance and micro-structure control into one culminating effort to demonstrate up to now unprecedented cycling stability of metallic Li.—Taylor et al.
The team selected the garnet, LLZO (Li7La3Zr2O12) as the model solid electrolyte due to its stability against metallic Li and high ionic conductivity.
The chemical and mechanical treatments of the electrolyte material provide a pristine surface for lithium to plate evenly, effectively suppressing the formation of dendrites or filaments. Not only does this improve safety, it enables a dramatic improvement in charging rates, said Jeff Sakamoto, a U-M associate professor of mechanical engineering who led the work.
Up until now, the rates at which you could plate lithium would mean you'd have to charge a lithium metal car battery over 20 to 50 hours (for full power). With this breakthrough, we demonstrated we can charge the battery in 3 hours or less.
We’re talking a factor of 10 increase in charging speed compared to previous reports for solid state lithium metal batteries. We’re now on par with lithium ion cells in terms of charging rates, but with additional benefits.—Jeff Sakamoto
The study found that Li|LLZO|Li cells are capable of cycling at up to 0.9 ± 0.7 mA cm−2, 3.8 ± 0.9 mA cm−2, and 6.0 ± 0.7 mA cm-2 at room temperature, 40 and 60 °C, respectively. Extended stability was shown in Li plating/stripping tests that passed 3 mAh cm−2 charge per cycle for a cumulative capacity of 702 mAh cm−2 using a 1 mA cm−2 current density.
We did the same test for 22 days,” he said. “The battery was just the same at the start as it was at the end. We didn’t see any degradation. We aren’t aware of any other bulk solid state electrolyte performing this well for this long.—Nathan Taylor, lead author
Bulk solid state electrolytes enable cells that are a drop-in replacement for current lithium ion batteries and could leverage existing battery manufacturing technology. With the material performance verified, the research group has begun producing thin solid electrolyte layers required to meet solid state capacity targets.
Nathan J. Taylor, Sandra Stangeland-Molo, Catherine G. Haslam, Asma Sharafi, Travis Thompson, Michael Wang, Regina Garcia-Mendez, Jeff Sakamoto (2018) “Demonstration of high current densities and extended cycling in the garnet Li7La3Zr2O12 solid electrolyte,” Journal of Power Sources, Volume 396, Pages 314-318 doi: 10.1016/j.jpowsour.2018.06.055