Nucleation & growth
Nuclei appear, grow, and merge while the current traces a hump.
Step the potential and watch the current: nuclei form, grow, and merge while the transient traces a hump. The shape of that hump, sharp and early versus broad and late, is the fingerprint of how the deposit is forming, before you ever image it.
What it measures
In a potentiostatic deposition experiment the current first rises, new nuclei appearing and each one growing, then peaks as the nuclei’s diffusion zones overlap, then decays toward planar-diffusion (Cottrell) behavior. The Scharifker-Hills analysis makes that shape quantitative by plotting the dimensionless transient (i/iₘ)² against t/tₘ and comparing it to two limiting models:
- Instantaneous nucleation: all nuclei form at once, then grow: (i/iₘ)² = 1.9542/(t/tₘ) · {1 − exp[−1.2564·(t/tₘ)]}². A sharper, earlier peak.
- Progressive nucleation: nuclei keep forming during growth: (i/iₘ)² = 1.2254/(t/tₘ) · {1 − exp[−2.3367·(t/tₘ)²]}². A broader, later peak.
The distinction matters because it predicts morphology: a fixed population growing uniformly versus a continuously seeded surface with a wide size distribution.
How to read the output
Normalize the measured transient by its maximum (iₘ, tₘ) and overlay the two theoretical curves. Data hugging the instantaneous curve means the nucleus population was set early, expect a more uniform deposit. Data tracking the progressive curve means nucleation continued throughout, expect a broader particle-size distribution and rougher growth. Transients that fall between the curves, or cross from one to the other, are telling you the conditions sit near a transition; small changes in overpotential or additive concentration will tip the mode. Check the tail: at long times the current should approach Cottrell decay, and a tail that refuses to fall usually means convection or a side reaction is contaminating the transient.
A real use case
An anode-free cell program needs lithium to plate densely and uniformly on bare copper, morphology is the whole game, and waiting for post-mortem SEM on every condition is slow. Plating transients run across a matrix of current collector treatments and electrolyte additives, pushed through the Scharifker-Hills analysis, sort the conditions in minutes: the baseline electrolyte tracks progressive nucleation (continuously seeded, mossy growth risk), while the fluorinated-additive cells shift cleanly to instantaneous behavior, a fixed nucleus population growing densely. SEM on the two extremes confirms the prediction, and the transient analysis becomes the screening proxy for the rest of the matrix.
Common mistakes
- Including the double-layer charging spike at t ≈ 0 in the analysis. The first instants of the transient are capacitive, not faradaic; the nucleation signal starts after it.
- Classifying from the raw transient instead of the dimensionless (i/iₘ)² versus t/tₘ plot, peak position alone, unnormalized, reflects concentration and overpotential as much as mechanism.
- Running with stirring or vibration present. Scharifker-Hills assumes diffusion-only transport; convection flattens the tail and breaks the fit.
- Forcing a verdict on a transient that sits between the limiting curves. “Mixed or transitional” is a real answer; the limiting models are bounds, not bins.
- Generalizing one overpotential’s result. Nucleation mode is potential-dependent; classify across the operating window before drawing conclusions.
Waveform recognition first, then the right analysis
Drop in a chronoamperometry file and Niobia identifies the waveform and classifies the experiment, recognizing a deposition transient as deposition, rather than forcing a Cottrell analysis onto it. For nucleation studies it locates the transient maximum, builds the dimensionless Scharifker-Hills plot, fits both limiting models, and reports the classification, instantaneous versus progressive, with a confidence assessment rather than a hard verdict on ambiguous data. Assumption violations it can detect from the data, like a tail inconsistent with diffusion-only transport, are flagged alongside the result, so a contaminated transient does not silently become a morphology claim.
Frequently asked
Why does nucleation mode matter for battery manufacturing?
Because it predicts deposit morphology. Instantaneous nucleation tends toward uniform, dense deposits; progressive nucleation toward broad size distributions and rougher growth, which, for lithium on copper, is the difference between dense plating and mossy, dendrite-prone deposits.
Can the analysis distinguish anything besides the two limiting modes?
The limiting curves are bounds. Real transients often fall between them or transition with potential; the defensible readout there is 'mixed/transitional,' supported by where the data sits relative to both curves.
What experimental conditions does Scharifker-Hills assume?
Potentiostatic control, diffusion-limited growth of hemispherical nuclei, and no convection. Violate those, stirring, gas evolution, large IR drop, and the dimensionless plot will misclassify; the fix is the experiment, not the fit.