Millions of electric vehicle (EV) batteries are on the road, in testing labs, or already retired, and researchers are finding new ways to study, extend, and repurpose them. Their work spans advanced imaging, material redesign, and real-world field testing. What’s emerging is a deeper understanding of battery materials and architectures that could redefine how energy is stored, charged, reused, and managed.
The following research projects explore how scientists and engineers are working hard to advance EV battery lifespans through real-time diagnostics, novel materials, and new approaches to reuse and reliability.
Making movies to diagnose battery failures
Cornell chemists, led by Yao Yang and Erik Thiede, have developed a new method to observe lithium-ion battery materials in action during charging and discharging, particularly under extreme temperatures. This is done using a specialized tool known as operando electrochemical transmission electron microscopy (EC-TEM).
Unlike traditional methods that analyze battery materials after they’ve failed or been removed from the system, operando EC-TEM enables researchers to make “movies” of the materials as they change in real time at the nanoscale during actual battery operation.
View a movie-like animation of operando electrochemical 4D-STEM of Cu dendrites here, which are structures that can cause short circuits in EV batteries. The animation captures an electron beam scanning around a 2D STEM image, with each pixel containing a 2D electron diffraction pattern. (Image: Journal of the American Chemical Society).
Using this electron microscope setup, the team created 4D movies showing the growth of copper dendrites, tiny metal structures that can form and lead to short circuits or battery degradation. These contour maps, similar to heat maps, revealed how uneven and fast the growth can be at different temperatures.
The project also used AI algorithms developed by Thiede’s lab to analyze the massive datasets from these movies, which contain millions of data points from nanometer-scale reactions.
The method enables material-level diagnostics from minus 50° to 300°C, offering insight into how batteries behave in real-world conditions and supporting the development of safer, more resilient EV battery systems. It offers EV engineers and battery scientists a real-time diagnostic tool to understand failure modes, improve thermal design, and validate new battery materials under realistic conditions.
Identifying thermal runaway risk
A new method for diagnosing EV battery faults was recently studied using vehicle usage history rather than real-time sensor data, marking a first in the field. Instead of monitoring live voltage, temperature, or current readings, the approach identifies risks accumulated from harmful charging behaviors and extreme environmental conditions over time.
By analyzing patterns such as high-rate discharging and fast charging at low temperatures, the system applies logistic regression to calculate risk scores. Tested on two EV models, it achieved diagnostic accuracies of 87.5 and 84.8%, highlighting its potential for fault prediction and design improvement.
Overview of the proposed method for identifying high-risk EVs using historical usage data. The four-step process includes data processing, risk feature extraction, logistic regression modeling, and accuracy evaluation without relying on real-time sensors. (Image: Heliyon)
Published in Heliyon, the work could transform how faults are detected in large EV fleets or vehicles lacking high-frequency sensing or real-time monitoring. Because it doesn’t rely on high-frequency measurements, it’s ideal for offline diagnostics, sensitivity analysis, and comparative design improvements.
The researchers, which are from the Beijing Institute of Technology and State Grid Tianjin Electric Power Company, are plan to refine the model by integrating more risk factors and applying advanced machine learning techniques to broaden its applicability and predictive power across diverse EV platforms.
Predicting battery degradation 1,000x faster
Researchers at the National Renewable Energy Laboratory (NREL) have developed a physics-informed neural network (PINN) model that predicts lithium-ion battery health nearly 1,000 times faster than traditional diagnostic models.
Unlike standard approaches, which are computationally intensive and often impractical for real-time diagnostics, the PINN model blends physical battery degradation principles with deep learning to simulate internal changes without requiring destructive analysis.
NREL’s battery researchers are turning to advanced AI models to optimize battery performance. (Image: Werner Slocum, NREL)
By embedding physical laws directly into the neural network’s training process, the model can infer key degradation metrics(such as electrode capacity and Li-ion transport) based solely on voltage behavior. This surrogate modeling approach was validated against widely used models like the Single-Particle Model and Pseudo-2D Model and demonstrated accuracy with drastically reduced computational cost.
Currently, in the validation stage using lab-cycled batteries, the PINN surrogate offers a path to onboard diagnostics for EV batteries, enabling systems to adapt charging behavior based on real-time degradation signals. Researchers aim to scale this method to future battery designs, advancing predictive maintenance, system optimization, and battery lifespan.
Ten-minute EV charging at subzero temps
It may be summer in North America, but engineers at the University of Michigan are already preparing for the cold. They’ve developed a modified manufacturing process for lithium-ion EV batteries that enables ten-minute fast charging at subfreezing temperatures without sacrificing energy density or battery lifespan.
This system tackles a key barrier to EV adoption: poor cold-weather performance. It addresses lithium plating and power loss that typically occur during winter fast charging.
A University of Michigan engineering student plugs in an EV for charging in snowy weather on the school’s North Campus. EV charging becomes less efficient in colder weather, but a new strategy for manufacturing battery electrodes could enable charging in ten minutes in temperatures, despite the cold. (Image: Marcin Szczepanski, Michigan Engineering)
The team combined micro-structural optimization with surface chemistry engineering to overcome diffusion and interface challenges at low temperatures. The anode is laser-patterned with micron-scale vertical channels (~40 µm) that improve ionic access deep within thick graphite electrodes. These 3D pathways reduce tortuosity and allow lithium ions to move uniformly across the electrode during charging.
Despite these improvements, cold conditions still triggered the formation of a resistive solid electrolyte interphase (SEI) layer, causing lithium plating and reduced capacity. To counter this, the team applied an ultrathin (~20 nm) amorphous lithium borate-carbonate coating to the electrode. This artificial interphase stabilizes electrode-electrolyte interactions and suppresses lithium nucleation, enabling uniform lithium intercalation even at –10° C.
Together, the 3D patterning and artificial SEI coating produced a 500% improvement in charge rate under cold conditions, with 97% capacity retention over 100 extreme fast-charging cycles.
Published in Joule, the work was led by Dasgupta with co-authors Tae Cho and Manoj Jangid, supported by the Michigan Translational Research and Commercialization (MTRAC) program. The team notes the method is compatible with current battery production lines, offering a viable path for commercial adoption. Prototypes were built at the U-M Battery Lab and analyzed using advanced microscopy at the Michigan Center for Materials Characterization.
Next, the researchers aim to optimize the method for large-scale manufacturing and explore how electrode architecture and interfacial chemistry can be tuned for next-generation EV batteries with better cold-weather resilience and energy throughput.
Restoring old EV batteries to perform like new
Researchers at the University of Chicago, in collaboration with UC San Diego, have discovered a new material with thermodynamic properties unlike anything previously observed. The material expands under pressure, contracts when heated, and exhibits inverse electrochemical behavior, defying conventional physics.
As part of a long-term collaboration, researchers in Prof. Y. Shirley Meng’s lab at the UChicago Pritzker School of Molecular Engineering and visiting scholars from UC San Diego found negative-thermal expansion in metastable oxygen-redox active materials, seemingly violating the laws of thermodynamics (Image: Jason Smith)
One of the most promising applications is in restoring aging EV batteries. The material shows an opposite voltage reaction during cycling, which the researchers believe could be used to reset a degraded battery’s performance. With a targeted voltage activation, an older EV battery could regain near-original range without replacement or recycling.
Published in Nature, the study suggests these materials could enable new battery designs that are more durable, heat- and pressure-resistant, and multifunctional. The researchers envision future energy systems where structural components also serve as energy storage, opening the door to lighter, more efficient EVs and electrified aircraft.
The next phase of the project involves studying how electron transfer processes affect the material’s behavior over time. By applying redox chemistry, the team hopes to fine-tune performance for different environments, battery chemistries, and structural applications. The goal is to define the material’s limits while unlocking its full potential across EVs and aerospace.
Giving EV batteries a second life on the grid
Millions of retired EV batteries are headed for scrapyards. However, a new project from the University of Michigan–Dearborn is testing whether they might instead become the backbone of tomorrow’s clean energy grid. Backed by a $1.48 million grant from the Michigan Department of Environment, Great Lakes, and Energy, the research team is preparing one of the country’s most ambitious real-world tests of second-life EV batteries.
Professors Xuan Zhou, Mengqi Wang, and Wencong Su are building a 500-kilowatt grid-connected storage system powered entirely by used EV battery packs. Unlike lab simulations that degrade new batteries under controlled conditions, this system will use real GM battery packs that have endured the stress of everyday driving, including thousands of charge cycles, vibrations, temperature shifts, and heat from nearby vehicle systems.
Associate Professor Xuan Zhou (right), Associate Professor Mengqi Wang (left), and Professor Wencong Su are heading up a new project that will tie used EV batteries to the electric grid. (Image: University of Michigan-Dearborn)
The team’s goal is to understand whether these aged batteries can reliably support key grid functions like peak shaving (discharging power during high-demand periods to offset fossil fuel use) and load shifting, where energy is stored during off-peak hours and dispatched when electricity is more costly.
To do so, they’ll need to develop custom converters, control algorithms, and a power management system capable of connecting multiple used packs safely to solar power and the grid without introducing harmonics or damaging infrastructure. Many of these components will need to be built from scratch, as no off-the-shelf solution exists.
Initial work includes detailed chemical and physical analysis of each battery, followed by small-scale proof-of-concept testing. Because the full system is too large for any current university lab, the researchers are also exploring partnerships to support high-power validation. Once testing concludes, their partner, ReCharge ReCycling, will help dismantle the batteries for material recovery, closing the loop by extracting valuable metals for future reuse.
The project offers what could be the most practical dataset to date on the safety, longevity, and economic feasibility of second-life batteries in grid storage. It may help solve two challenges at once by reducing battery waste and making renewable energy more affordable and resilient.
Filed Under: Batteries, FAQs