UHPC Method
A high-precision measurement is only as good as its weakest control. The signal — a coulombic efficiency a few parts per million away from 1 — is small enough that temperature drift, an inconsistent cutoff, or a noisy channel can erase it.
Setup
- Reference and cell. For mechanism work, a lithium-metal reference holds a stable potential; keep the cell and its connections low-resistance so the sensed voltage tracks the electrode, not the leads.
- Temperature. Current sensing has a temperature coefficient, so hold the chamber steady — a fraction of a degree of drift can move the reading more than the effect you are chasing. Fix the setpoint (for example 40.0 ± 0.1 °C) and log it.
- Rate. Low rates (C/20–C/10) give near-equilibrium curves and cleaner dV/dQ; they also lengthen each cycle, which feeds directly into the CIE/h reading.
- Cutoff. Use a consistent voltage window and avoid ending a cycle on a flat voltage plateau, where a small shift moves the true endpoint and drags CE with it.
Running the measurement
- Record charge and discharge capacity every cycle at the instrument's full resolution; the analysis uses the sensed values, not the set current.
- Ten to fifteen cycles are usually enough to read a trend that ordinary equipment needs hundreds of cycles to show.
- Keep conditions identical across channels — slippage depends on channel-to-channel accuracy, not just single-channel precision.
Reading the data
- CE curve. How fast CE converges and how steady it stays. A cell whose CE climbs above 99.9% within a few cycles and holds is forming a stable interface; a slow or drifting CE flags continued lithium loss.
- dV/dQ (differential voltage). Peaks mark phase transitions: a peak losing area points to active-material loss; a peak shifting along the capacity axis points to lithium loss.
- EIS. Run in parallel to separate ohmic, interfacial (SEI + charge transfer), and diffusion resistances, and to watch each grow over cycling.
Pitfalls
- Over-attribution. CE, fade, and slippage say a class of mechanism is present, not which reaction. Naming a specific reaction needs a controlled experiment (for example removing the cathode, or a symmetric cell).
- The FSR-vs-reading trap. Current accuracy is quoted against full scale, not reading; dividing an absolute error by a small set current invents a false error rate. See Measurement & precision.
- Over-extrapolation. Early-cycle data predicts long-term life only within the linear region; extrapolating far past the measured range (roughly beyond 6× the tested cycles) grows the error fast.