SEOUL, September 07 (AJP) - Some of the signals that battery researchers read as ions moving inside a material are produced by nothing more than an uneven surface, a team at the Korea Advanced Institute of Science and Technology found, and the correction is as simple as polishing the sample flat.
The team tested the effect on single-crystal silicon, a material in which no ions travel, after machining fine trenches into its surface, and a signal resembling ion movement appeared anyway. The same behavior showed up in a graphite anode of the kind used in lithium-ion batteries and in a sodium solid electrolyte. The Korea Advanced Institute of Science and Technology (KAIST) announced the findings Monday.
Batteries work by shuttling lithium or sodium ions back and forth inside the cell as they charge and discharge. How fast and how freely those ions travel determines how quickly a battery can be refilled and how long it survives, so laboratories building the next generation of cells spend much of their effort mapping where inside a material ions flow easily and where they stall.
One of the main tools for that mapping is electrochemical strain microscopy. The instrument drags an extremely fine probe across the surface of a battery material and applies a small alternating voltage through the tip. The voltage nudges the ions nearby, the material swells or contracts by an amount far too small to see as they crowd or thin out, and the probe registers the flexing. Ion traffic is never observed directly, only inferred from it.
That is what leaves the measurement exposed. A change in the height of the surface alters how firmly the tip presses against the sample, and that shift alone moves the reading. The instrument's control system, which adjusts continuously to keep the probe tracking the surface, lags a fraction behind a sudden step or dip, and during the lag the contact stiffens or softens. The result looks much like ions on the move.
To separate the two, the researchers needed a sample incapable of producing a real signal. Silicon in single-crystal form conducts no ions, so anything the instrument picked up had to come from the shape of the surface. Cutting trenches into it reproduced the grooves and steps found on a sliced battery electrode, and the height changes on their own generated a response.
The mechanism resembles what happens to a car on a rutted road. The body rises and falls with the surface and the load on the tires changes with it. A probe crossing a ridge on an electrode behaves the same way, and the varying pressure feeds straight into the measurement.
Real materials behaved no differently. In a graphite anode and in a sodium solid electrolyte known as Na2Zn2TeO6, the reading tracked the shape of the surface, which pointed to a problem that follows the instrument rather than any one class of material.
The sharpest case appeared at grain boundaries, the seams where the small crystals making up a battery material meet, similar to the lines between laid tiles. Ions are widely thought to move quickly along those seams, and before the samples were smoothed the signal there was strong. After smoothing, the extra signal was gone.
Smoothing was done with a cooling cross-section polisher, a machine that trims the cut face of a sample with a beam of argon ions. Argon reacts with almost nothing, which lets the beam shape a surface without changing the chemistry underneath. Roughness fell sharply, and the false readings fell with it.
The work stops short of showing that published measurements of fast ion movement along grain boundaries were wrong. It establishes that surface shape by itself can generate a comparable signal, and that in these samples the extra signal at the seams vanished once the face was flattened. How much earlier work carries the same error has not been tested, and the polishing method was demonstrated on a small set of materials.
Accurate readings at this scale also feed a second line of research, in which artificial intelligence and machine learning systems are trained on nanoscale data to predict how a new battery material will perform before anyone builds it. Those predictions are only as reliable as the measurements behind them.
Three groups at KAIST ran the study together. Hong Seung-bum and Yuk Jong-min, both in the Department of Materials Science and Engineering, worked with Choi Nam-soon in the Department of Chemical and Biomolecular Engineering.
Hong said he expects the result to help researchers "understand and design the operating principles of next-generation battery materials," including solid-state cells that replace liquid electrolyte with a solid and sodium-ion cells that swap lithium for a cheaper and far more plentiful metal.
The paper appeared online June 11 in the journal Small Methods. Dongyan Chen, a doctoral researcher in the Department of Materials Science and Engineering, is its first author.
[Reference Information]
Journal/Source: Small Methods
Title: Quantitative Analysis of Topographic Crosstalk in DART-ESM Arising from Feedback-Loop-Delay-Induced Contact Stiffness Variations in Battery Materials
Link/DOI: 10.1002/smtd.70763
AJP Takeaways
- KAIST said a widely used nanoscale imaging technique can produce signals resembling ion movement when the only thing changing is the height of the sample surface.
- KAIST researchers reproduced the effect on single-crystal silicon, which conducts no ions, and again in a graphite anode and the sodium solid electrolyte Na2Zn2TeO6.
- KAIST reported that the strong signal seen at grain boundaries disappeared once the sample face was flattened with an argon ion beam, and the team did not test how much earlier work carries the same error.
- Small Methods published the paper online June 11, three months before the announcement, tracing the false readings to a lag in the instrument's tracking system that changes how firmly the probe presses on the surface.
Copyright ⓒ Aju Press All rights reserved.