GITT and PITT — getting a diffusion coefficient out of a battery
Diffusion of ions inside the electrode often sets the rate limit, and GITT is how you measure it. The trick is counterintuitive: the useful information is in the relaxation after the pulse, not the pulse itself.
GITT (galvanostatic intermittent titration technique) applies a series of current pulses, each followed by a long rest, marching across the state of charge as a staircase of voltage. During each pulse the voltage jumps (an instant IR drop plus a slope); during the relaxation it drifts toward equilibrium. The relaxation transient — voltage versus √time — yields the diffusion coefficient D at that state of charge. PITT is the potentiostatic counterpart using voltage steps and current decay. D typically varies strongly with state of charge, dipping near phase transitions.
Why the relaxation, not the pulse
When you apply the pulse, the voltage change has several overlapping contributions — ohmic IR drop, charge transfer, and diffusion — tangled together. When you stop and let the cell rest, the fast contributions vanish quickly and what remains is the slow equilibration of the concentration gradient inside the particle. That relaxation, analyzed as voltage versus √(time), isolates solid-state diffusion, which is why the technique brackets each measurement with a long rest.
How to read it well
1 · The staircase maps D across state of charge
Each step gives one D value at one composition. Plotting D against state of charge reveals where transport is slow — usually dips at phase transitions, where the diffusion mechanism changes.
2 · The √t window matters
The simple analysis assumes a linear voltage-vs-√t region during the pulse, valid only for short times before the gradient reaches the particle center. Fitting outside that window biases D.
3 · Rest long enough
If the relaxation hasn't reached quasi-equilibrium before the next pulse, the baseline is wrong and so is D. The rest must be long relative to the diffusion time — which is exactly why GITT is slow.
4 · Cross-check the number
D from GITT should be in the same ballpark as D from EIS (Warburg) and CV (Randles–Sevcik). Large disagreement flags an assumption breaking somewhere.
Common pitfalls
- Rests too short. The single most common error — an incomplete relaxation corrupts the equilibrium baseline and the extracted D.
- Wrong particle geometry / area. D depends on the assumed diffusion length; a wrong particle radius scales the answer.
- Fitting the whole transient. Only the early linear √t region is valid for the simple formula.
- Ignoring the SOC dependence. Quoting a single D for a material hides the dips that actually limit rate.
Where this gets slow by hand
GITT is slow to run and slow to analyze: dozens of pulse-relaxation steps per scan, each needing the equilibrium voltage, the IR drop, and a √t fit over the correct window to extract one D — repeated across state of charge and across cells. Doing it manually is hours per dataset, and the SOC-dependent D curve that matters only emerges once every step is processed consistently.
D as a function of state of charge, automatically
Niobia segments the GITT staircase into its pulse and relaxation phases, identifies the valid linear √t window for each step, and extracts the diffusion coefficient across the full state-of-charge range — producing the D-vs-SOC curve and flagging steps where the rest was too short to trust. It reconciles the result against D from EIS and CV into one defensible number with its uncertainty. Hours of consistent fitting become a single curve and a confidence band.
Frequently asked
What is GITT used to measure?
GITT (galvanostatic intermittent titration technique) is used to measure the solid-state diffusion coefficient of ions in an electrode as a function of state of charge, along with the equilibrium (open-circuit) voltage. It applies current pulses separated by long rests and analyzes the voltage relaxation.
Why does GITT use the relaxation period?
During a current pulse, the voltage response mixes ohmic, charge-transfer, and diffusion effects. During the rest, the fast effects decay and the slow equilibration of the internal concentration gradient dominates, so analyzing the relaxation (voltage versus the square root of time) isolates the diffusion coefficient.
What is the difference between GITT and PITT?
GITT applies constant-current pulses and watches the voltage relax; PITT (potentiostatic intermittent titration technique) applies voltage steps and watches the current decay. Both extract the diffusion coefficient as a function of state of charge, using current or voltage control respectively.