Battery Testing and Characterisation

Cell, energy and power

A battery cell consists of two electrodes (positive/cathode, negative/anode), a separator and an electrolyte. The cell voltage is the potential difference between the two electrodes; 3–4.2 V for Li-ion. Capacity (Ah) is the total charge the cell can deliver, energy (Wh) the product of voltage and capacity, power (W) the instantaneous voltage×current. Values normalised to mass and volume (Wh/kg, Wh/L, W/kg) are the common language for comparing chemistries; the Ragone plot shows the energy–power trade-off.

C-rate, SOC, DOD and efficiency

The C-rate expresses the current relative to the nominal capacity: 1C is full discharge in one hour, C/10 in ten hours. State of charge (SOC) is the percentage of remaining charge, depth of discharge (DOD) its complement. Coulombic efficiency (CE) is the ratio of discharge charge to charge charge in one cycle; energy efficiency is lower because it also includes the voltage hysteresis. In commercial Li-ion cells CE is above 99.9 %; CE differences therefore cannot be measured meaningfully without 1 ppm-level current accuracy — this is the rationale of 'high-precision coulometry'.

Cell formats and test cells

In research, coin cells, three-electrode test cells (with reference electrode), Swagelok-type and small pouch cells are used; in industry cylindrical (18650, 21700, 4680), prismatic and pouch formats. A half cell (Li metal counter) gives the properties of a single electrode, a full cell the realistic balance. Electrode loading (mg/cm²), N/P ratio and electrolyte amount are parameters that determine the results and must be reported.

Electrode chemistry and intercalation

Li-ion electrodes are intercalation compounds in which Li⁺ enters and leaves a host structure (graphite, LCO, NMC, LFP). The voltage–composition curve reflects the phase diagram: single-phase regions give slopes, two-phase regions flat plateaus. At the negative electrode the electrolyte is reduced and forms the solid electrolyte interphase (SEI); formation of the SEI in the first cycle is the main cause of irreversible capacity loss, and its long-term growth the principal mechanism of impedance rise and capacity fade.

Galvanostatic cycling (GCD / GCPL)

Charge and discharge at constant current are defined by voltage limits (e.g. 3.0–4.2 V). In each cycle capacity, CE, average voltage and energy are calculated; the capacity–cycle curve is the indicator of life, the voltage profile that of the chemistry. Rest steps (OCV relaxation) are used to separate hysteresis and polarisation.

CC-CV charging

After the upper voltage limit is reached at constant current, the voltage is held until the current falls to a cut-off value (e.g. C/20). The CV step completes the capacity left incomplete at high C-rate because of kinetic and ohmic limitations; the charge passed in the CV step is an indirect measure of the cell's internal resistance and diffusion limitation. It is the standard of commercial chargers.

Rate capability, HPPC and duty cycles

The rate test measures how capacity drops with C-rate (C/10, C/5, 1C, 2C, 5C…). HPPC (Hybrid Pulse Power Characterization) determines internal resistance and pulse power with short discharge/charge pulses at different SOCs; the results feed the equivalent-circuit models of electric-vehicle battery management systems (BMS). Duty-cycle (drive-cycle) tests apply real usage profiles (e.g. vehicle driving cycles) as a current/power sequence; constant-power (CPW) discharge produces the Ragone plot.

Life testing and acceleration

Cycle life and calendar life (storage at constant SOC and temperature) trigger separate mechanisms. High temperature, high upper voltage and high C-rate accelerate ageing; for extrapolation of results to real conditions, periodic reference performance tests (RPT: low-C-rate capacity, EIS, HPPC) are added to the protocol. Consistent current accuracy and time base across channels is the prerequisite of comparability in multi-cell life tests.

Separating thermodynamics and kinetics

The voltage measured on an intercalation electrode while current flows adds the ohmic, charge-transfer and diffusion overpotentials on top of the equilibrium (thermodynamic) voltage. GITT and PITT separate these contributions with short excitations and long rests: the voltage at the end of the rest gives the equilibrium OCV (against composition), the transient response during the excitation the diffusion kinetics.

GITT (galvanostatic intermittent titration)

A small current pulse (e.g. C/20, 10–30 min) is followed by a rest (hours) until the OCV stabilises; at each step the composition advances by a small amount. According to the Weppner–Huggins analysis, in the short-pulse limit (t ≪ L²/D) the voltage varies linearly with √t and the chemical diffusion coefficient is calculated as D = (4/π)·(I·V_m/(z·F·A))²·[(dE/dδ)/(dE/d√t)]². The result depends strongly on correct knowledge of the geometric area and particle size; this is the main reason why literature values are scattered over orders of magnitude.

PITT (potentiostatic intermittent titration)

The potential is changed in small steps (5–10 mV) and the decay of the current is recorded. In the long-time limit the current decreases exponentially: the slope of ln(I) vs t gives −π²D/(4L²); in the short-time limit Cottrell behaviour is seen. The total charge passed at each step gives the differential capacity (dQ/dE) directly; phase transformations appear as peaks in dQ/dE and minima in D.

Comparison with EIS

The D calculated from the low-frequency Warburg region rests on the same Fick model as GITT and PITT; consistency of the three methods tests the validity of the model. In two-phase regions (flat plateau) the concept of 'diffusion coefficient' is debatable; here the movement of the phase boundary determines the rate and results should be reported as apparent values.

Anatomy of a battery spectrum

In a typical Li-ion Nyquist plot one sees at high frequency an inductive tail from the cables, at the real-axis intercept the ohmic resistance (electrolyte, current collector, contact), at mid-high frequency the SEI/contact-resistance semicircle, at mid-low frequency the charge-transfer semicircle (strongly temperature-dependent, Arrhenius) and at low frequency solid-state diffusion (Warburg, followed by a capacitive rise). The semicircles usually overlap; DRT or a temperature/SOC sweep helps to separate them.

Measurement practice

The impedance of batteries is of the order of mΩ; four-terminal connection, short twisted cables and galvanostatic mode (to avoid OCP drift) are required. The amplitude is chosen in current such that the cell's voltage response does not exceed 5–10 mV. Since the spectrum depends on SOC and temperature, the measurement conditions must be fixed and stated; rest before the measurement lowers the NSD indicator.

dQ/dV and dV/dQ analysis

Differential capacity (dQ/dV vs V) turns the plateaus of the charge/discharge curve into peaks; each peak corresponds to a phase transformation or a stage in the electrode pair. Shift, weakening or disappearance of peaks with ageing points to the mechanism: loss of Li inventory (LLI), loss of active material (LAM) and loss of conductivity leave different signatures. For data recorded at constant time intervals, differential coulometry (DCS) gives the same information in a more noise-resistant form.

Ageing mechanisms

SEI growth (Li consumption, impedance rise; progresses as t^(1/2)), Li metal plating (at low temperature/high C-rate; safety risk), cathode structural degradation and transition-metal dissolution, loss of binder/conductive network, electrolyte depletion and gas formation. Growth of the SEI semicircle and rise of the ohmic resistance in EIS, peak shifts in dQ/dV and the 'knee' in the capacity–cycle curve are the observable consequences of these mechanisms.

Solid electrolytes and interfaces

Solid-state batteries use a ceramic (garnet LLZO, NASICON, sulfides) or polymer electrolyte instead of a liquid one. The total resistance is the sum of bulk and grain-boundary conductivity and the electrode/electrolyte interface resistance; EIS separates these components by frequency (bulk MHz, grain boundary kHz, interface of the order of Hz). Ionic conductivity shows Arrhenius behaviour; a temperature sweep gives the activation energy. Dendrite formation, contact loss and chemical instability are the main problems; measurements with blocking (Au/Pt) and reversible (Li) electrodes separate the ionic and electronic contributions.

Module and pack measurement

In a module of series-connected cells the total voltage reaches tens to hundreds of volts; the voltage of every cell must be monitored separately because the weakest cell limits the module. Multichannel simultaneous EIS measures the impedance of each cell separately while the current passes through the whole stack and identifies the faulty/unbalanced cell directly. At high voltage, isolation, floating measurement and safety (contactor, fuse, temperature cut-off) are part of the experiment design.

Flow batteries

In redox flow batteries (e.g. vanadium, VRFB) energy is stored in the electrolyte tanks and power in the cell stack; the two scale independently. Characterisation is done by constant-current charge/discharge (CE, VE, EE efficiencies), polarisation curve (area-specific resistance, ASR) and EIS (membrane, electrode kinetics, mass transport); electrolyte flow rate and SOC must be stated for every measurement.