Ask what a cell's internal resistance is and the correct answer begins with a question: over what timescale, at what state of charge, at what temperature, and at what age. A cell does not have one resistance. It has a response that unfolds over time, and every measurement method takes a slice of it.
What contributes
Three families of process oppose current flow, and they act on different timescales.
Ohmic resistance comes from the current collectors, tabs, welds, terminals and the ionic resistance of the electrolyte in the separator and pores. It responds effectively instantly, within microseconds.
Charge transfer resistance arises at the electrode and electrolyte interfaces, where lithium ions cross between the electrolyte and the solid. It responds over milliseconds to seconds, follows Arrhenius temperature dependence, and depends on state of charge.
Diffusion limitation covers lithium transport through the electrolyte and through the solid particles. It responds over seconds to minutes and is what causes voltage to keep drifting under a long constant load.
Apply a current step and you see all three arrive in sequence: an immediate ohmic jump, then a faster-then-slower curve as interfacial and diffusion effects build. That sequence is why the pulse duration defines the number.
DCIR, the workshop method
Direct current internal resistance is the voltage change divided by the current step, over a stated duration.
Worked example. A cell rests at 3.700 V. A 20 A discharge pulse is applied. After 1 second the terminal voltage reads 3.640 V; after 10 seconds it reads 3.610 V.
The 1 second DCIR is 0.060 V divided by 20 A, which is 3.0 milliohms. The 10 second DCIR is 0.090 V divided by 20 A, which is 4.5 milliohms.
Both are correct. They describe different things: the shorter pulse is weighted towards ohmic and fast interfacial behaviour, the longer one includes more diffusion. A supplier quoting 3.0 milliohms and a test house measuring 4.5 milliohms have not disagreed; they have used different definitions. Standards such as IEC 62660-1 define pulse conditions precisely for exactly this reason.
Datasheets often quote AC internal resistance (ACIR) instead, measured with a small 1 kHz alternating current as specified in IEC 61960. At that frequency the measurement captures mostly ohmic resistance, so ACIR reads lower than any DC pulse value from the same cell, and the two are not interchangeable. For orientation at room temperature and mid state of charge: large prismatic and pouch EV cells commonly sit below 1 milliohm, 21700 cylindrical cells around 10 to 25 milliohms, and 18650 cells around 20 to 40 milliohms.
DCIR is what most pack engineering runs on, because it is measured with the hardware already present and it maps directly to the voltage drop the system will see under load.
EIS, the diagnostic method
Electrochemical impedance spectroscopy applies a small sinusoidal current or voltage across a frequency sweep, typically from tens of kilohertz down to millihertz, and records the magnitude and phase of the response. Because each physical process has its own characteristic timescale, the frequency axis separates them.
Plotted on a Nyquist diagram, with negative imaginary impedance against real impedance, a typical lithium-ion cell shows:
| Frequency region | Feature | Physical process |
|---|---|---|
| Above roughly 1 kHz | Below the axis, inductive | Cabling, current collectors, cell geometry |
| Around 1 kHz, axis crossing | Intercept with the real axis | Ohmic resistance |
| Hundreds to tens of hertz | First depressed semicircle | Ion transport through surface films, including the SEI |
| Tens of hertz to around 1 Hz | Second semicircle | Charge transfer with double layer capacitance |
| Below roughly 0.1 Hz | Sloped tail, near 45 degrees | Diffusion, the Warburg element |
The two semicircles frequently overlap and need model fitting to separate. Fitting uses an equivalent circuit, most often a resistor in series with one or two parallel resistor-capacitor pairs and a diffusion element, with constant phase elements standing in for ideal capacitors to account for surface heterogeneity.
The diagnostic value is that ageing moves different features. A growing first semicircle points at surface film growth. A growing second semicircle points at charge transfer degradation, often from cathode surface reconstruction. A shifting ohmic intercept points at electrolyte depletion, connection degradation or drying out. A single DCIR number moving upward tells you none of this.
Temperature and state of charge dependence
Charge transfer resistance can rise by an order of magnitude between 25 degrees Celsius and minus 20. Electrolyte conductivity falls at the same time. Together they explain the cold weather experience: a pack that still holds its energy but cannot deliver or accept power, recovering as it warms.
State of charge dependence is smaller across the middle of the range and rises at both ends, particularly at low state of charge where the cathode becomes less conductive in some chemistries. Any resistance measurement intended for comparison must be taken at a fixed state of charge, or the comparison is noise.
Age raises all three contributions, but not equally, and the pattern is chemistry and duty dependent.
Why it matters commercially
Power capability. Available power is bounded by how much current the cell can pass before the terminal voltage hits a limit. That is a resistance question, not a capacity question, and it is why a pack can pass a capacity test and fail a power test.
Heat. Ohmic heat generation is current squared multiplied by resistance. A pack whose resistance has risen 50 per cent generates 50 per cent more heat at the same current, which loads the thermal system, which raises temperature, which accelerates the ageing that raised the resistance. Designers should check the thermal case at end of life resistance, not beginning of life.
Round-trip efficiency. For a grid storage asset, resistive loss is a direct revenue item. A percentage point of round-trip efficiency across a twenty-year asset is a large number, and it degrades over life.
State of health. Resistance-based state of health is the appropriate metric for power-limited applications, and it diverges from capacity-based state of health in the same cell. Both should be tracked.
Measurement pitfalls
Use a four-wire connection. Two-wire measurement includes contact and lead resistance, which on a large low-resistance cell can exceed the quantity being measured.
Let the cell stabilise thermally and electrically. A cell measured shortly after cycling reads differently from a rested one, and the difference can be larger than the ageing signal you are chasing.
Fix and record the state of charge, the temperature, the pulse duration and the current amplitude with every result. A resistance value without those four is not comparable to anything.
Keep the excitation small for EIS. The method assumes linear response, and too large an amplitude breaks that assumption and distorts the spectrum.
Watch for cable and fixture artefacts in the high frequency region. The inductive tail often says more about your test setup than about the cell.
Informational and educational content only. Not professional, financial, legal, or engineering advice.