A team from the Johns Hopkins University Applied Physics Laboratory (APL) and Stanford University took an important step toward safer and faster charging of lithium-ion batteries by advancing the capability for dynamic, noninvasive internal temperature measurement.
Knowing a battery's internal temperature is critical; temperatures warmer than 80°C (176°F) can initiate cascading exothermic reactions that lead to venting of toxic and flammable gases, combustion and explosion. But estimating internal temperature has been difficult for conventional surface-temperature sensors because of thermal inertia and noise.
Now, new research is prompting researchers to rethink battery charging, now with stronger guarantees on cell safety and faster charging times. A study published by the APL-Stanford team in the Journal of Power Sources shows that dynamic internal temperature measurements actually peak when a charging battery reaches around 61% of its full capacity, a safety consideration that is entirely obscured at the cell surface.
This work builds on the noninvasive battery internal temperature sensor (NIBITS) recently developed by APL's Rengaswamy Srinivasan, the lead author of "Graphitic carbon anode temperature excursions reflect crystallographic phase transitions in lithium-ion cells," along with colleagues at APL.
The technique works by first calibrating the cell's internal resistance (more precisely impedance) measurements across the cell terminals against premeasured internal temperatures obtained under static conditions with a thermal chamber. Subsequently, internal temperature measurements are dynamically generated during charging by combining impedance measurements with the calibration data. Experiments detailed in the paper suggest these internal temperatures are less noisy and more responsive than temperatures measured at the cell surface.
While the NIBITS technique obtains calibration data under static conditions, there are no gold-standard comparisons to directly validate the resulting dynamic temperature measurements. To better establish the validity of the technique under dynamic conditions, the team showed that NIBITS measurements related to known phase transitions in the lithium-intercalated graphite anode of the battery.
Employing unsupervised machine learning, Dr. Lakshminarayan Srinivasan, MD, Ph.D., principal investigator of Stanford's Neural Signal Processing Laboratory -- as well as coauthor and son of APL's Srinivasan -- was able to identify these physical state transitions from the internal temperature measurements alone. "These transitions were entirely obscured in cell-surface temperature measurements during fast charging," he notes, adding that the technique is still presently limited to measurements below 80°C (176°F).
Before the technology can reach consumers, the researchers say, challenges remain in scaling the technique to multi-cell batteries and designing the manufacturing process to integrate NIBITS into the battery form factor. The APL team has already begun exploring these scaling challenges.
"Extending these solutions to larger batteries would open the possibility for sensor-enabled closed-loop charging strategies that could simultaneously increase safety and drop charging times for electric cars, consumer electronics, and many other applications in today's lithium-ion battery-powered world," says Rengaswamy Srinivasan. "Meeting scaling challenges to unlock the potential of NIBITS…that's part of the fun."
Materials provided by Johns Hopkins University Applied Physics Laboratory. Note: Content may be edited for style and length.
Cite This Page: