Stanford Researchers Discover How Exactly Li-Ion Batteries Degrade

May 11, 2020 by Gary Elinoff

By knowing more about lithium-ion battery degradation, engineers can define steps to improve lifetime and performance.

One of the great struggles of lithium-ion batteries, especially for EV developers, is lifespan. Now, new research out of the U.S. Department of Energy’s SLAC National Accelerator Laboratory at Stanford University reveals the actual electrochemical events in lithium-ion batteries that cause degradation.

While the researchers don't suggest a solution for stopping or slowing down degradation, we can break down the results of this new study by first assessing what we already understand about Li-ion battery (LIB) performance and lifespan.


What Do We Already Know About LIB Lifespan?

Battery University outlines the factors that affect a LIB's lifespan, including storage and operating temperature. Another factor is a battery's percentage of electrical capacity when it is charged and its charge percentage when discharged. The total number of charge/discharge cycles the battery has undergone also degrades a LIB over time.

The example below graphs the number of DST (dynamic stress tests), or charge-discharge cycles versus the percentage of total energy storage capacity the battery retains.


An LIB's loss in capacity as a function of charge/discharge bandwidth

A LIB's loss in capacity as a function of charge/discharge bandwidth. Note that DST stands for "dynamic stress tests" and SoC, in this case, stands for "state of charge." Image used courtesy of Battery University


The orange data points indicate that batteries charged to 75% capacity and discharged at 10% to 65% capacity last the longest. The black data points correspond to a deep 100% to 25% duty cycle typical of smartphone usage. Predictably, the battery loses power much more quickly and lasts through fewer cycles. 

Clearly, users pay a price for a more convenient deep charge/discharge cycle.


How LIB Cathodes Degrade

To learn more about LIB lifespan, scientists at the SLAC National Accelerator Laboratory delved into the gradual degradation of LiB cathodes made of nickel-manganese-cobalt, or NMC, using two assets: SLAC’s Stanford Synchrotron Radiation Lightsource (SSRL) and the European Synchrotron Radiation Facility (ESRF). According to SLAC, the investigators combined X-ray tomography data with machine learning to provide a detailed picture and analysis of the process. 

The results of the study were published in Nature. The researchers honed in on the possibility that the NMC particles, breaking away from the conductive carbon matrix that hold them together, might be a cause of the degradation.


New computer vision algorithm identifies NMC particles from an X-ray tomography image.

New computer vision algorithm identifies NMC particles from an X-ray tomography image. Image used courtesy of Yijin Liu/SLAC National Accelerator Laboratory

They then employed computer vision algorithms and X-ray tomography to identify the specific NMC particles breaking apart and away from the matrix. One problem the researchers had to overcome with machine learning was the tendency of the algorithm to zero in on the boundaries around particles, making it difficult to distinguish the difference between several small particles stuck together and one large particle with fractures in it. 

The researchers addressed this issue by using an algorithm designed for hierarchical objects—entities made of smaller parts. The researchers “taught” the algorithm to distinguish different kinds of particles. They were then able to develop a three-dimensional picture of how NMC particles break away from the cathode, regardless of whether they were fractured or whole, large or small.


The Results of the Study

The research produced two major results.

The researchers first found that when consumer electronics are under typical usage conditions, a LIB’s charge capacity is significantly affected by particle detachment from the carbon matrix.


The difference among the original NMC particle data

The difference among the original NMC particle data (top), traditional segmentation (middle), and machine learning segmentation (bottom). Image used courtesy of Zhisen et. al

The second result was that, yes, large NMC particles are more likely to break away than small ones. According to Yijin Liu, a staff scientist at SLAC and a senior author the Nature paper, smaller particles also break away. However, there is more variation in small particle behavior. It had been previously assumed that by devising smaller NMC particles, developers could yield longer-lasting batteries.


LIB Preservation Tips

So, how do we slow down LIB degradation in a practical sense? Researchers at the University of Michigan have compiled studies from academic and corporate literature searching for information about the use and maintenance of LIBs.

These sources included ten cellphone manufacturers, ten laptop manufacturers, and ten electric vehicle manufacturers, as well as four power tool manufacturers. Unlike the SLAC study, this report was not scientific research, but a compendium of observations.

Many of the suggestions from U of M researchers paralleled the findings from Battery University, published ten years ago: 

  • Avoid extreme temperatures and high-moisture environments
  • Avoid leaving the battery at either 100% charge or 0% charge
  • Avoid puncturing LIBs 
  • When possible, opt for standard charging over fast charging


University of Michigan tips for preserving EV batteries.

University of Michigan tips for preserving EV batteries. From your experience, what about these tips affect LIBs at the circuit level? Image (modified) used courtesy of the University of Michigan


Keep in mind that the inconsequential problems of LIBs in laptops are not so readily gainsaid in the hundreds of millions of EV LIBs, each weighing more than the average human. More research like that conducted by SLAC may help us to remedy the wear and tear on LIBs over time.


The Next Step: Research and Recycling? 

The primary question raised in the SLAC study—namely, How do we slow down the breakdown and breakaway of NMC particles from the matrix?—appears to be the next step for future LIB research. In the meantime, EV developers will need to find power-efficient solutions to meet the demand for LIBs.

Lithium is expensive. According to a Beroe report on the price trend and cost structure of LIBs, a LIB cell for, say, a smartphone ranges from $2 to $4. But in an EV, LIBs can cost anywhere between $7,000 and $20,000. In this instance, it's possible that recycling may be the solution. 



Lithium-ion batteries appear in a host of electrical applications—from laptops to EVs. What's your experience with them? Share your thoughts in the comments below.