Electrochemical Impedance Spectroscopy (EIS): Nyquist Plots and the Randles Circuit Explained

What EIS Is and Where It Applies

EIS probes an electrochemical interface by measuring its impedance — the AC analog of resistance — across a wide frequency range. Impedance Z(ω) generalizes Ohm's law to sinusoidal signals: for an applied potential E(t) = E₀·sin(ωt) and current response I(t) = I₀·sin(ωt + φ), the impedance is a complex number Z = |Z|·e^(jφ) = Z′ − jZ″, where Z′ is the resistive (real) part and Z″ the reactive (imaginary) part, and ω = 2πf is the angular frequency.

The power of EIS is that different physical processes dominate at different frequencies. At high frequency (kHz–MHz), fast capacitive charging of the double layer and the ohmic electrolyte resistance dominate. At low frequency (mHz–Hz), slow processes — charge-transfer kinetics and mass-transport (diffusion) — emerge. A single spectrum therefore deconvolutes several phenomena that a DC measurement blurs together.

• Batteries and supercapacitors — state-of-health, SEI resistance, ion diffusion
• Corrosion science — coating integrity, polarization resistance (Stern–Geary)
• Fuel cells — membrane and cathode losses
• Biosensors — label-free detection of binding events

The Mechanism: Deriving the Randles Circuit Response

The Randles circuit models the interface as: solution resistance Rs in series with a parallel combination of the double-layer capacitance Cdl and the Faradaic branch (charge-transfer resistance Rct in series with a Warburg element Zw).

Step 1 — the Faradaic branch. A redox reaction O + ne⁻ ⇌ R proceeds at rate governed by Butler–Volmer kinetics. Linearizing at small perturbation gives the charge-transfer resistance Rct = RT/(nFi₀), where i₀ is the exchange current density; a facile reaction (large i₀) means small Rct.

Step 2 — diffusion. If reactant supply is limited, a Warburg impedance Zw = σ·ω^(−1/2)·(1 − j) appears in series with Rct, where σ = RT/(n²F²A√2)·(1/(C*_O√D_O) + 1/(C*_R√D_R)).

Step 3 — the double layer. Cdl charges in parallel, shunting current at high frequency.

Step 4 — combine. The parallel Faradaic branch and Cdl, added to series Rs, give the total impedance. At high ω the capacitor shorts the branch (Z → Rs); at low ω the capacitor is open and Z → Rs + Rct + Warburg tail. The crossover produces the characteristic semicircle plus 45° Warburg line.

Key Quantities and a Worked Nyquist Example

On a Nyquist plot (−Z″ vs Z′), the Randles circuit produces a semicircle whose left intercept = Rs and right intercept = Rs + Rct. The semicircle diameter therefore reads Rct directly. Below it, a straight line at 45° (slope 1) is the Warburg diffusion tail.

The frequency at the top of the semicircle, ω_max = 2πf_max, satisfies ω_max = 1/(Rct·Cdl), so the double-layer capacitance is Cdl = 1/(ω_max·Rct).

Worked example. Suppose Rs = 20 Ω, the semicircle spans from 20 Ω to 220 Ω (so Rct = 200 Ω), and the apex occurs at f_max = 80 Hz. Then:

• ω_max = 2π·80 = 503 rad/s
• Cdl = 1/(503 × 200) = 9.9 × 10⁻⁶ F ≈ 10 µF
• If area = 1 cm², Cdl ≈ 10 µF/cm² — a plausible smooth-metal value
• From Rct = RT/(nFi₀), with n=1, T=298 K: i₀ = (8.314·298)/(96485·200) ≈ 1.3 × 10⁻⁴ A/cm²

Real surfaces flatten the semicircle; the capacitor is replaced by a constant phase element Z_CPE = 1/[Q(jω)ⁿ], with n = 1 for an ideal capacitor and n ≈ 0.85 for rough electrodes.

How EIS Is Measured in Practice

The instrument is a potentiostat coupled to a frequency response analyzer (FRA). The potentiostat holds the working electrode at a chosen DC bias (often the open-circuit potential, OCP) and superimposes a small AC perturbation. The FRA correlates the current response with the excitation at each frequency to extract magnitude and phase.

• Three-electrode cell — working, reference (e.g., Ag/AgCl, SCE), and counter electrode isolate the interface of interest.
• Small amplitude — typically 5–10 mV RMS to stay in the linear (pseudo-linear) regime, so Butler–Volmer can be linearized.
• Frequency sweep — logarithmically spaced, commonly 100 kHz down to 10 mHz; very low frequencies require long, stable measurements.

Data quality is checked with the Kramers–Kronig (KK) transform, which relates Z′ and Z″ for any linear, causal, stable, time-invariant system; large KK residuals flag drift or nonlinearity. The spectrum is then fit to an equivalent circuit by complex nonlinear least squares (CNLS), reporting each element with confidence intervals. Presentation is via the Nyquist plot and the two Bode plots (|Z| vs f, and phase φ vs f).

How EIS Compares to Related Electrochemical Methods

EIS is a small-signal, frequency-domain, steady-state method, which distinguishes it from its DC cousins:

• vs. Cyclic voltammetry (CV) — CV sweeps potential over a large range (hundreds of mV to volts) and is a large-signal, time-domain method that maps redox potentials and reaction reversibility. EIS instead perturbs by only millivolts and resolves resistances and capacitances quantitatively. CV tells you what reacts and roughly how fast; EIS tells you the interfacial resistances and capacitances.
• vs. Chronoamperometry / potential-step — a step in the time domain contains all frequencies at once; EIS is its frequency-resolved equivalent, trading speed for spectral separation of processes.
• vs. Linear polarization resistance (LPR) — LPR gives a single Rp near OCP; EIS separates Rs from Rct so the polarization resistance is not corrupted by solution ohmic drop.

The Randles circuit itself is one of many equivalent circuits. Coated metals use nested R–CPE ladders; porous electrodes use transmission-line models; finite diffusion layers replace the semi-infinite Warburg with a finite-length (bounded) Warburg element.

Limits, Pitfalls, and Significance

EIS is deceptively powerful but has real caveats. Its foundational assumptions are linearity, causality, stability, and time-invariance — a battery that drifts during a slow 10 mHz sweep violates stability, and the Kramers–Kronig test will show it.

• Equivalent-circuit ambiguity — different circuits can fit the same data equally well. A good fit does not prove a mechanism; the circuit must be justified physically, not just statistically.
• The CPE puzzle — the constant phase element fits non-ideal capacitance elegantly but its exponent n and admittance Q have debated physical origins (surface roughness, distributed time constants, electrode geometry).
• Warburg vs. finite diffusion — the ideal 45° line assumes semi-infinite diffusion; thin cells or convection bend it into a finite-Warburg arc.

Significance: EIS underpins modern battery diagnostics — the growth of the solid-electrolyte interphase (SEI) shows up as an added semicircle, and rising Rct signals aging. In corrosion, EIS non-destructively grades coating breakdown years before visible rust. Randles's 1947 insight, later formalized with the Warburg element (Emil Warburg, 1899) and the Butler–Volmer framework, remains the reference model taught in every electrochemistry course.

Randles circuit elements and their physical meaning (typical values for a metal electrode in aqueous electrolyte)

Element	Symbol	Impedance	Typical value	Physical meaning
Solution resistance	Rs (or RΩ)	Rs (real, frequency-independent)	1–50 Ω·cm²	Ohmic drop of electrolyte between reference and working electrode
Charge-transfer resistance	Rct	Rct (semicircle diameter)	10–10⁵ Ω·cm²	Kinetic barrier to the Faradaic redox reaction; inversely proportional to i₀
Double-layer capacitance	Cdl	1/(jωCdl)	10–100 µF/cm²	Charge stored in the electrical double layer at the interface
Warburg impedance	Zw	σ·ω^(−1/2)·(1 − j)	σ ≈ 10–100 Ω·s^(−1/2)	Semi-infinite linear diffusion of redox species to the electrode
Constant phase element	CPE (Q, n)	1/[Q·(jω)ⁿ]	n = 0.8–0.95	Non-ideal capacitor accounting for surface roughness/heterogeneity