Calendar Aging of Next Generation Batteries: A Trivial Issue or the Final Frontier?
NREL researcher, Ankit Verma, takes us through the complications of a battery’s shelf life, and how research is aiming to solve the problem of next generation calendar aging.
Introduction to calendar aging
When we think about battery metrics, the most reported values are capacity (Ah), energy density (Wh/kg), power density (W/kg) and cycle life (number of cycles before capacity fades to 80% of pristine capacity). In the context of electric vehicles (EVs), cycle life translates to mileage and having a decent cycle life >1000 cycles with a 300-mile range per charge at beginning of life translates to >250,000 miles before end of life which is at par with internal combustion engine cars.
An often overlooked but equally important aspect when talking about batteries is calendar life which, as the name suggests, is “how long will a battery take to reach 80% of its beginning-of-life (BOL) nominal capacity when not in use?”. This is a mode of degradation that is time dependent instead of the energy throughput dependent cycle life. In the context of current lithium-based batteries that power our EVs, thermodynamic non-equilibrium drives parasitic (electro)chemical reactions that lead to continued loss of active lithium to the solid electrolyte interphase (SEI) even when the battery is at rest.
Most batteries sit at rest for the majority of their lives. Take EVs for example; the average household vehicle is driven for ~60 minutes per day and charged overnight, so EV batteries are undergoing calendar aging for nearly two-thirds of the day. We want our batteries to last 10 years in conditions that can range from sunny summers to freezing winters; this will directly translate to EV life of 10 years before the battery needs to be replaced. Battery packs cost upwards of $10k in EVs; it’s too pricey and resource intensive to replace it every few years.
Calendar aging in current graphite-based lithium-ion batteries is a solved problem
Why don’t we hear much about calendar life when it comes to current graphite-based lithium-ion batteries (LIBs)? Because it’s a well understood problem with clear solutions. Calendar life is a function of the ambient temperature and battery state of charge (e.g. fully charged (100% SOC), fully discharged (0% SOC), 50% SOC etc.). An equivalence can be made with the shelf life of food products; while cooked meat might spoil in a couple of hours outside the fridge, inside it can last for days, even weeks inside the freezer.
Similarly, if we keep LIBs near low SOCs in cold temperatures, the driving force for the parasitic lithium ion-electrolyte solvent decomposition reactions happening at the graphite anode are minimal. These parasitic reactions are what lead to continuous cyclable lithium inventory consumption from graphite active material into the solid electrolyte interphase (SEI) and contribute to capacity fade.
LFP and NMC batteries paired with graphite anodes show good calendar life even at medium SOCs if the temperature is appropriately maintained. The general rule of thumb is to store at <50% SOC and keep the batteries below 20°C to achieve the optimum 10-year calendar life metric.
Why is calendar aging trickier for next generation batteries?
When we move towards the next generation LIBs with varying anodes like silicon and lithium as we try to push the energy density metrics higher towards 350-500 Wh/kg, things get more interesting. Initially, researchers were primarily focused on solving the cycle life challenges associated with the immense volume expansion of these anodes during cycling which can fracture both the active material and the SEI and lead to accelerated cycle fade. These efforts have borne fruit; strategies like nano-sizing, coatings etc. have enabled >1000 cycles life at C/3 for Si based anodes (see 1Nate Neale, BAT 498 The Silicon Consortium Project: Next Gen Materials for Silicon Anodes), while lithium metal batteries have reached 600 cycles with the use of novel electrolytes, electrode architectures (see 2Jun Liu, BAT317 Progress and Status of Battery500 Consortium Phase II). This progress was recently presented in the Vehicle Technologies Office Annual Merit Review in June 2024. However, calendar lives are yet to breach the 2-year mark for Si-based LIBs; for lithium metal cells a detailed report of calendar life testing is yet to be seen.
Why is time dependent degradation not showing up in the cycle life performance of silicon anodes? To run 1000 cycles at C/3, the total testing duration is only around 250 days (~8 months) which obfuscates the time-based losses. Furthermore, during cycling the battery is swinging through all SOCs instead of a nearly fixed SOC in calendar aging. The time spent at bad SOCs is minimal in cycle aging which was hiding this issue.
Calendar life is distinctly bad for silicon anodes as compared to graphite because of a multitude of reasons: silicon surface is inherently more reactive than graphite and can react with the binder, hydrolytic cycles continuously form HF that can etch the silicon and the SEI etc. These issues also result in fascinating observations like SEI “breathing” which is thickening/thinning of the SEI during Si lithiation/delithiation respectively due to a complex mechanics/dissolution interplay.
Lithium metal anodes also have lower reducing potential than graphite and suffer severely from calendar fade. With liquid electrolytes, they can lose 2-3% of capacity after only 24 hours of aging. Furthermore, while the general trend of increased calendar aging with temperature remains for next generation cells, the SOC dependence is not straightforward; calendar aging can be worse in the discharged state.
How do we measure calendar aging?
The typical way to measure calendar aging is using the United States Advanced Battery Consortium Open Circuit Voltage – Reference Performance Test (OCV-RPT) protocol. As the name suggests, it involves keeping the battery at rest for month-long intervals at a desired SOC and then doing capacity checks through full cycles between the working voltage window. The process is repeated till the capacity drops to 80%. This is a slow but quantitative approach towards estimating calendar life. A calendar aging test for a good performing cell can ideally last for multiple years; generally, we can get an idea of the fade trend and fit a regression line with a couple of years of data that can be extrapolated with good confidence for calendar life.
New techniques for accelerated calendar aging testing
The time consuming and resource intensive nature of the traditional OCV-RPT tests has spurred the need for accelerated tests for rapid calendar life screening. Checking if a new material or cell design is going to improve calendar life for rapid iteration towards the optimal configuration cannot be sustained by OCV-RPT tests. To this end, the Silicon Consortium Project has devised the potentiostatic hold (voltage hold) protocol. This incorporates holding the Si containing cell at a constant potential for upwards of a week instead of letting it rest for a month like in the OCV-RPT protocol. This protocol’s advantage lies in giving time resolved current and capacity data during the hold which can subsequently be deconvoluted to obtain the parasitic losses. The accelerated degradation observed under the hold conditions allows us to compare calendar life between two cells in a faster fashion allowing for rapid material/design iteration (for more on this read this recent paper). Another way to accelerate the calendar fade is to run the test at higher temperatures (e.g. 45-60oC), however care must be taken to ensure that we are not adding a new high temperature degradation mechanism that is insignificant in normal calendar aging.
In summary, calendar aging performance of next generation battery systems is equally important to other battery metrics, if not more. There is no point having an excellent cycling battery that loses all its capacity when sitting still for a few days or months.
Dedicated research and development of accelerated calendar aging protocols for next generation battery systems like lithium metal batteries with liquid/solid electrolytes, lithium-sulfur batteries, sodium-ion batteries etc. is needed for rapid calendar life assessment. And the next time we see one of the innumerable battery startups hitting battery energy/power density, cycle life metrics; a question should pop-up in our minds “what’s the calendar life of these cells?”
🌞 Thanks for reading!
📧 For tips, feedback, or inquiries - reach out
📣 For newsletter sponsorships - click here
🌐 Follow us on Twitter, LinkedIn, and our website
Drop us an email on issy@intercalation.co for the PDFs.
Really nice comprehensive and informative article.
Super good article explaining the difficulty of calendar aging test. Now it would be great if companies like Amorous would publish numbers so we can see if there are progress in the area. It also explain why silicon cell still are not mainstream.