Benefits and Challenges of In-Field EIS

Even as batteries permeate almost every aspect of our modern lives, we still don't know that much about them : how they age, their real limits, how to most effectively use them. At ReJoule, we're driven to de-mystify the battery so we can use them more efficiently and effectively. We are doing that by making battery diagnostics lower cost and more accessible for battery packs, from electric vehicles to stationary storage.

Electrochemical impedance spectroscopy (EIS) is a powerful tool used by electrochemists around the world to evaluate battery characteristics. It is fast (can be less than 1 minute), nondestructive (doesn't require cutting open a battery), and can give rich electrochemistry information about the battery such as charge transfer, internal resistance, and diffusion characteristics.

EIS measurements can be visualized in a Nyquist plot, then fit to basic circuit elements. An example of the application of the battery management system (BMS) could be on the battery pack of an electric vehicle. By analyzing the behavior of these elements over time, we can paint a very clear picture of things like battery state-of-health (SOH). Below, I highlight a few key challenges to implementing and interpreting EIS data for a field application such as a BMS.

Key Challenges:

Sensitive voltage measurements

To perform EIS on a battery, an ac perturbation current is injected, and a resulting ac voltage response (or vice versa) signal is measured. The impedance of an EV lithium-ion cell is typically in the range 0.1 to 1mΩ. This means that a 5A sine current imposed on a battery cell with 1mΩ impedance will yield a voltage response a 5mV, often the accuracy of an existing BMS voltage measurement!

Wide frequency range

Another major challenge is that a complete EIS spectra spans 3 decades of frequencies, from 10mHz to 10kHz! Most electronics are designed to operate in a relatively narrow bandwidth. Many world class EIS analyzers use multiple parallel filter circuits to maintain accuracy across that frequency range. This adds complexity and cost to any electronic system.

Generating high current sine waves

If you think about it, to apply a sine wave current to a cell is to essentially rapidly charge and discharge a battery in a repeatable amplitude and frequency. An EIS analyzer must supply both positive and negative current, switch back and forth between the two rapidly, and also do so with high precision! Most power supplies are unidirectional, and are designed to supply a constant voltage.

Interpreting EIS spectra

A battery’s voltage and EIS spectra changes as a function of load, state-of-charge (SOC), cabling inductance, and temperature. This makes EIS a challenge, particularly for field battery applications where loading profiles are often unpredictable, and environmental conditions are not controllable.

Modeling EIS spectra

After all of the above, we finally get to the good stuff: linking battery impedance spectra to SOH, and remaining useful life (RUL). While many academic studies have suggested a strong correlation between certain elements of impedance spectra and capacity, most are limited studies. To truly build a strong aging model, much more data is needed for different battery chemistries. The reality is battery aging takes time, resources, and money.

At ReJoule, we're developing the technology and taking the steps to overcome these very real technical challenges. Please follow us on LinkedIn to get notifications as we update our blog!