Techniques for Operando Battery Analysis

Battery technology is ubiquitous in modern life. Everyday electrical goods from phones and laptops, to cars and power tools employ hundreds of different battery types, each with chemistries tailored to the specific demands of the device. Developers are constantly working on new designs to improve the efficiency, lifetime and performance of these batteries and over the last 20 years have made impressive progress in this field.¹

Understanding a system is the key to improving it and it is their access to powerful analytical tools that allows these researchers to carefully hone the properties of different electrochemical devices. But this is no simple matter. “Batteries typically operate in an extreme thermodynamic condition. They essentially always run outside of their stability window,” explained Prof. Dr. Philipp Adelhelm, a professor of physical chemistry at Humboldt-University Berlin. “This means their components must be carefully optimized and batteries have to be sealed to avoid any entry of impurities.”

Opening up the battery, known as post-mortem analysis, can completely change the internal chemistry, offering a limited understanding of the processes occurring when the battery is in operation.² In situ analysis is far less invasive, pausing the usual operation of the battery for a longer time to allow researchers to take a measurement before continuing use. But likewise, this snapshot of the battery does not truly reflect the chemistry at play when the system is running normally.² Ideally, researchers need to be able to analyze different parameters during typical use conditions to build a realistic picture. Such methods are known as operando techniques, however while these techniques are the most informative, they are also the most challenging to apply practically.

“For most techniques, you need to somehow get through that metal layer to see what's happening inside which is especially difficult for optical techniques,” said Jason Porter, a professor of mechanical engineering at Brigham Young University. “We're also working at very small length scales. You have layers between maybe 10 and 200 microns – less than the width of a human hair – so trying to image what's happening at that scale while simultaneously operating under as realistic conditions as possible is a real challenge.”

Despite these difficulties, there are already multiple complementary techniques available, each exploring different processes integral to the performance of a battery.³

Differential electrochemical mass spectrometry (DEMS)

“Why do batteries age? Why do they lose capacity? It's because we abuse them – fast charging, extreme temperatures, they all can induce side reactions which corrode the materials inside the battery,” explained Adelhelm. Identifying and understanding these unwanted processes is the focus of Adelhelm's research.

Commercial cells can reach values exceeding 99.9% efficiency.⁴ But even with laboratory research cells easily reaching 99%, developing techniques sensitive enough to probe the 1% wasted in side reactions is a complex task.⁵

Optimizing Battery Technology With Ball Mills

Explore the essential applications of laboratory ball mills in battery production.
View Whitepaper

Advertisement

Adelhelm’s group specializes in a technique called differential electrochemical mass spectrometry (DEMS) which involves connecting a mass spectrometer to an electrochemical cell in order to analyze the gaseous side products released during different operational conditions. “If you use a too-high charging voltage, you decompose, for example, your liquid electrolyte. A flowing carrier gas transports any gaseous decomposition products into the mass spectrometer where we then identify them by mass,” Adelhelm said.

This technique is a prime example of an operando method, with the group designing a specialist experimental setup to enable the trapping and transporting of gaseous products to the mass spectrometer while the battery continuously charges and discharges.⁶ The design is a careful balance between keeping the system as realistic as possible while also accommodating the necessary alterations to the cell which facilitate measurement.

“The carrier gas dries out the cell because the liquid electrolyte is an organic solvent, so we have to work with excess electrolyte,” said Adelhelm. “The other challenge is with the generation of gas bubbles which can get stuck in the very porous electrode, rather than going straight to the mass spectrometer. This can cause a random spike in your mass spec data which is then tough to analyze.”

The team reviews the overall mass spec data holistically, looking for characteristic combinations of side products which indicate certain degradation processes, rather than focusing on the individual compounds themselves. This wider approach enables them to cut through the noise and dramatically simplifies the analysis process.

While not a perfect replica of a commercial battery, Adelhelm’s method does give profound insight into the side processes occurring within a battery under standard operating conditions. Ultimately, the Adelhelm Group hope that understanding these unwanted reactions will enable researchers to design better, degradation-resistant materials, leading to longer battery lifetimes and improved overall performance.

Fourier transform infrared spectroscopy (FTIR)

As an optical technique – one that measures the interaction of chemical compounds with light from optical wavelengths – probing the internal workings of a battery with Fourier transform infrared spectroscopy (FTIR) requires careful engineering. “We actually build coin cells on top of a diamond window,” Porter said.⁷ “As we cycle the battery, we direct an infrared beam up into this diamond window where it will reflect and come back down. But at the interface between the window and whatever's touching it, there's an interaction. So we can see where the light absorbs in the material and from that we can get an absorption spectrum and correlate that to different bonds in the molecules of the electrolyte.”

Porter's work investigates ion transport within battery electrolytes, looking at factors influencing how the ions move between the electrodes and the conditions affecting the stability of the electrolyte itself.⁸ Both have implications for the rate of charging – the faster and more easily lithium ions can move, the quicker and more efficiently the battery will charge.

FTIR is typically used to probe the bonds within organic molecules; lithium ions themselves are not IR active. To overcome this limitation, Porter’s team has developed several tricks to measure the concentration of these crucial particles indirectly.⁸ “In the electrolyte there's always a cation and an anion, for example, a lithium cation and a PF₆ (hexafluorophosphate) anion,” Porter explained. “Because of charge neutrality, anions and cations are always going to be at the same concentration everywhere in the battery, so if we can measure the anion via those phosphorus-fluorine bonds, we can infer the concentration of the lithium cation.”

Likewise, the interaction of these lithium ions with the organic solvent also leaves a detectable trace. “The carbonate solvents arrange themselves around the lithium ion and this shifts the vibrational modes of the solvent molecules,” said Porter. “We've been able to calibrate these shifts to the concentration of lithium ions so we've got a couple of ways to measure these IR inactive ions.”

By probing their movement under different conditions, Porter can determine the particular factors that limit the diffusion of ions and therefore the rate of charging, insights that the team hope will inform the design of superior materials in the future.

Complementary methods to create complete data sets

The limitation of all these methods is that no single technique works on all types of battery, Porter adds. “They all have strengths and weaknesses but by combining different people's expertise across these techniques we can get a fuller picture and accelerate battery development,” he said.

X-ray absorption spectroscopy is a powerful example.⁹ Similarly to Porter’s optical methods, the technique measures the interaction of atoms with X-rays and provides detailed information about the local environment and electronic state of atoms. This radiation is sufficiently penetrating that these measurements can sometimes be taken without alterations to the battery. However, these experiments often require synchrotron time, which is difficult, expensive and often slow to access. “We really need more data and more reproducibility with synchrotron methods,” said Adelhelm. “With actively limited time, students understandably want to look at different samples to see trends between them, rather than measuring one sample 20 times.”

Sample reproducibility will also be key for developing effective AI tools to help streamline battery analysis studies further. “Some groups are building big datasets from voltage and current responses after cycling thousands and thousands of batteries under different conditions,” said Porter.^10,11 “Machine learning can correlate these responses to what the changes were in that battery and try to interpret why they perform better or worse.” Porter believes this could be the beginning of an important symbiosis, with advances in operando techniques feeding better data to AI models, which in turn can support the development of more refined batteries and techniques.

For the present, the challenge remains of making these existing methods applicable to commercial batteries. The demand for high-performance batteries shows no sign of slowing down and, as battery chemistry grows in complexity, so too will the complexity of the techniques used to analyze them. “It goes hand in hand,” said Adelhelm. “As batteries become better and more complex, the analytics will keep following behind them.”