Electrochemical Impedance Spectroscopy (EIS): A Practical Guide to Measurement and Interpretation
Electrochemical impedance spectroscopy (EIS) is widely used to investigate charge transport and interfacial processes in electrochemical cells, solid electrolytes, functional oxides, batteries, fuel cells, and sensors. Its strength is frequency resolution: different processes can respond on different characteristic time scales.
EIS is also easy to overinterpret. A semicircle is not automatically a grain-boundary response, and a visually convincing equivalent-circuit fit does not prove that the selected circuit is physically unique. Reliable interpretation requires controlled measurements, a justified model, and consistency with the specimen geometry and material chemistry.
1. What does an impedance measurement describe?
For a sinusoidal perturbation, impedance is the ratio of the voltage response to the current response in the frequency domain:
\[ Z(\omega)=\frac{\widetilde{V}(\omega)}{\widetilde{I}(\omega)} =Z'(\omega)+jZ''(\omega) \]
where:
- \(Z'\) is the real component of impedance, in ohms (Ω).
- \(Z''\) is the imaginary component, in ohms.
- \(j=\sqrt{-1}\) is the imaginary unit.
- \(\omega=2\pi f\) is angular frequency, in rad s\(^{-1}\).
- \(f\) is frequency, in hertz.
The phase angle describes the phase difference between voltage and current. A purely resistive element has zero phase angle; capacitive and inductive responses introduce frequency-dependent phase shifts.
For a small-amplitude perturbation around a stable operating point, EIS is commonly treated as a linear-response measurement. If the amplitude is too large, the system may become nonlinear and the measured spectrum may no longer represent a small-signal impedance.
2. Nyquist and Bode representations
Nyquist plot
A Nyquist plot usually displays \(Z'\) on the horizontal axis and \(-Z''\) on the vertical axis. Resistive contributions appear along the real axis, while capacitive processes commonly produce arcs above the axis under this plotting convention.
The high-frequency intercept of an ideal, resolved semicircle can indicate a series or bulk resistance, while the low-frequency intercept can include additional resistive contributions. Real materials often show depressed arcs, overlapping processes, electrode polarization, or incomplete arcs within the measured frequency window.
Important: the number of visible arcs is not necessarily the number of physically distinct processes. Two processes can overlap, one process may lie outside the measured frequency range, and distributed relaxation times can produce non-ideal shapes.
Bode plots
Bode plots show impedance magnitude and phase as functions of frequency:
\[ |Z|=\sqrt{(Z')^2+(Z'')^2} \]
\[ \phi=\tan^{-1}\left(\frac{Z''}{Z'}\right) \]
Use the correct quadrant when calculating phase from measured real and imaginary components. Bode magnitude and phase plots are useful for identifying characteristic frequency regions that may not be obvious in a Nyquist representation alone.
3. Common equivalent-circuit elements
An equivalent circuit is a mathematical representation of the measured response. It should be selected because its elements have a defensible physical interpretation, not simply because it produces a low residual.
| Element | Impedance | Typical interpretation |
|---|---|---|
| Resistor, \(R\) | \(Z_R=R\) | Ohmic or transport resistance |
| Capacitor, \(C\) | \(Z_C=1/(j\omega C)\) | Ideal capacitive response |
| Constant-phase element, CPE | \(Z_{CPE}=1/[Q(j\omega)^n]\) | Distributed or non-ideal capacitive response |
| Inductor, \(L\) | \(Z_L=j\omega L\) | Inductive contribution, wiring or physical processes |
For a CPE, \(Q\) is its magnitude parameter and \(n\) is the exponent. When \(n=1\), the CPE behaves as an ideal capacitor; when \(n\) differs from one, the response is non-ideal. The units of \(Q\) depend on \(n\), so it should not automatically be reported as a capacitance.
A simple series resistance and parallel \(R-C\) element can be written as:
\[ Z=R_s+\left(\frac{1}{R^{-1}+j\omega C}\right) \]
This is a useful teaching model, not a universal circuit for ceramic electrolytes. Grain, grain-boundary, electrode, contact, and diffusion contributions may require different representations depending on the material and measurement configuration.
4. From resistance to ionic conductivity
For a specimen with parallel, planar electrodes, the conductivity associated with a fitted resistance is:
\[ \sigma=\frac{L}{RA} \]
where \(L\) is the distance between electrodes or specimen thickness, \(A\) is the electrode area, and \(R\) is the resistance assigned to the relevant process.
If a cell constant \(K=L/A\) is used:
\[ \sigma=\frac{K}{R} \]
For \(L\) in centimetres, \(A\) in cm² and \(R\) in ohms, conductivity is obtained in S cm\(^{-1}\). For SI dimensions, use metres and obtain S m\(^{-1}\).
Do not use the total low-frequency resistance automatically when the objective is bulk conductivity. First identify which fitted resistance corresponds to the intended transport contribution. In a polycrystalline ceramic, bulk and grain-boundary contributions may be separated only if the frequency range, data quality, and model support that separation.
Illustrative calculation
Suppose a pellet has thickness \(L=0.10\) cm and electrode area \(A=0.50\) cm². If a justified fit assigns a resistance of \(R=2000\) Ω to the process of interest:
\[ K=\frac{0.10}{0.50}=0.20\;\mathrm{cm^{-1}} \]
\[ \sigma=\frac{0.20}{2000} =1.0\times10^{-4}\;\mathrm{S\,cm^{-1}} \]
This example demonstrates the geometry conversion only. It does not identify the resistance as bulk, grain-boundary, or total conductivity; that assignment must come from the spectrum and physical model.
5. A reproducible EIS workflow
Before collecting data
- Verify electrical contacts, lead configuration, shielding and instrument calibration.
- Choose a perturbation amplitude small enough to remain in the linear regime; verify amplitude independence when necessary.
- Allow the specimen to equilibrate at each temperature and atmosphere.
- Record the frequency range, points per decade, integration settings and measurement amplitude.
- For temperature-dependent studies, document heating/cooling history and whether measurements were repeated.
During analysis
- Inspect raw real and imaginary data before fitting.
- Check for inductive artefacts at high frequency and electrode polarization at low frequency.
- Compare candidate circuits using residual structure, parameter uncertainty and physical plausibility.
- Avoid adding circuit elements solely to improve a fit statistic.
- Where appropriate, test consistency with Kramers–Kronig relations or other linearity, causality and stability checks.
6. Why ceramic impedance spectra can be difficult
In oxide ceramics, a measured response may contain contributions from the bulk grains, grain boundaries, electrodes, interfaces, and gas/solid reactions. Their characteristic frequencies depend on capacitance, resistance, temperature, microstructure, and electrode chemistry.
A high-frequency response is often associated with a smaller capacitance than a low-frequency electrode response, but frequency alone is not sufficient for assignment. Use capacitance estimates, temperature trends, specimen thickness, electrode changes, and independent measurements to test the interpretation.
For proton-conducting oxides, water-vapour pressure and gas composition matter. For mixed ionic-electronic conductors, the measured response may not be represented by a single ionic resistance. A blocking-electrode experiment, controlled atmosphere, or complementary transport measurement may be needed to distinguish contributions.
7. Common interpretation errors
- Calling every semicircle a grain boundary. Confirm the assignment using capacitance, temperature dependence and independent evidence.
- Reporting a fitted resistance without geometry. Conductivity requires a defined cell constant or specimen dimensions.
- Treating a CPE parameter as a capacitance. Convert or interpret it using the selected model and exponent.
- Ignoring electrode and lead artefacts. These can dominate parts of the spectrum.
- Overfitting. A more complicated circuit is not automatically a better physical explanation.
- Combining measurements from different conditions. Temperature, atmosphere, humidity and electrode history must be comparable.
8. What should be reported?
A useful EIS methods section includes:
- Specimen composition, preparation, dimensions, density and electrode material/area.
- Instrument and measurement configuration, including wiring and shielding where relevant.
- Frequency range, perturbation amplitude, points per decade and stabilization procedure.
- Temperature, atmosphere, gas flow, humidity and pressure where applicable.
- Equivalent-circuit diagram, fitting procedure, parameter uncertainty and residual assessment.
- The resistance used for conductivity, the cell constant, units and uncertainty propagation.
- Reproducibility across specimens or repeated measurements.
Further reading
- E. Barsoukov and J. R. Macdonald (eds.), Impedance Spectroscopy: Theory, Experiment, and Applications, 2nd ed., Wiley (2005).
- M. E. Orazem and B. Tribollet, Electrochemical Impedance Spectroscopy, 2nd ed., Wiley (2017). https://doi.org/10.1002/9781119363682.
- J. R. Macdonald, “Impedance spectroscopy,” Annals of Biomedical Engineering 20, 289–305 (1992).