A Nyquist plot is a graph that maps how a system responds to different frequencies, displayed as a curve on a two-dimensional plane. The horizontal axis represents the real part of the response, and the vertical axis represents the imaginary part. Each point on the curve corresponds to a specific frequency, but frequency itself isn’t shown on either axis. It’s an implicit variable, meaning you trace along the curve as frequency increases. Nyquist plots are used primarily in two fields: control systems engineering (to determine whether a feedback system is stable) and electrochemistry (to characterize batteries, sensors, and corrosion).
How the Plot Works
To understand a Nyquist plot, it helps to think about what “frequency response” means. When you feed a signal into a system at a particular frequency, the output comes back with a certain strength (magnitude) and a certain delay (phase shift). Together, these define a complex number with a real and an imaginary component. Plot that complex number as a single dot. Then repeat the process across a range of frequencies, and you get a trail of dots that form a curve.
The result is a snapshot of the system’s entire frequency behavior compressed into one shape. Low frequencies typically appear on one end of the curve and high frequencies on the other, but you can’t read the frequency directly off the axes. In practice, measured data covers only a finite frequency range with limited resolution, so you always see a portion of the complete theoretical curve rather than a perfectly closed shape.
Stability Analysis in Control Systems
The most well-known use of a Nyquist plot is determining whether a closed-loop control system will be stable, using only the open-loop frequency response. The key idea is simple in principle: you look at how the curve behaves relative to a single critical point on the plot, located at -1 on the real axis (formally written as -1+j0).
The Nyquist stability criterion says: count how many times the curve wraps around that -1 point in the clockwise direction. Call that number N. Then count how many unstable poles the open-loop system already has, and call that P. The number of unstable poles in the closed-loop system is Z = N + P. If Z equals zero, the closed-loop system is stable. If Z is greater than zero, it’s unstable.
This is where the Nyquist plot has a clear advantage over its main alternative, the Bode plot. A Bode plot splits the same frequency response information into two separate graphs (magnitude and phase versus frequency), which works well for straightforward systems. But when the open-loop system itself is unstable, or when stability is ambiguous from the Bode diagram, the Nyquist plot gives a definitive answer. As MIT course materials on feedback control put it, if stability is unclear from the Bode diagram, always revert to the Nyquist plot. The encirclement count captures subtleties that are hard to read from magnitude and phase curves alone.
Gain Margin and Phase Margin
Beyond a simple stable-or-not verdict, the Nyquist plot also reveals how much room a system has before it becomes unstable. These safety buffers are called gain margin and phase margin.
Phase margin is found where the curve crosses the unit circle (the set of points at distance 1 from the origin). The angle between that crossing point and the negative real axis tells you how many additional degrees of phase lag the system can tolerate before the curve starts encircling -1. A phase margin of 0° means the system is right on the edge of instability. Higher values mean more breathing room.
Gain margin works similarly but focuses on where the curve crosses the real axis. It tells you how much the system’s gain could increase before the curve reaches -1. A gain margin of zero means the system is marginally stable. Engineers typically want both margins to be comfortably positive, and reading them off a Nyquist plot is a visual, intuitive process.
Nyquist Plots in Electrochemistry
Outside of control theory, Nyquist plots are a standard tool in electrochemical impedance spectroscopy (EIS). Here, instead of analyzing a feedback controller, you’re probing a physical system like a battery cell, a fuel cell, or a coated metal surface. A small alternating voltage is applied across a range of frequencies, and the impedance (resistance to alternating current) is measured at each one. The real and imaginary parts of that impedance form the Nyquist plot.
The shapes that appear on these plots have direct physical meaning. A semicircle in the mid-frequency range typically represents charge transfer resistance, which is the opposition to electrons moving between an electrode and the surrounding solution. The diameter of that semicircle corresponds to the size of that resistance. Where the curve first touches the real axis at high frequency, that intercept gives you the ohmic resistance of the solution and electrode material combined.
A straight line angled at roughly 45 degrees, usually appearing at lower frequencies, indicates diffusion-limited behavior. This is called a Warburg impedance and reflects ions slowly migrating through the electrolyte or electrode material. If the diffusion path has a defined length (as in a thin film), you may see a second semicircle instead of a straight line. The angle of the line can also shift: wider or shallower pores in an electrode produce a steeper angle, while branching or widening pores flatten it.
Equivalent Circuit Modeling
Electrochemists interpret Nyquist plots by fitting the data to an equivalent circuit, a simple arrangement of resistors, capacitors, and other elements that mimics the system’s electrical behavior. The most common starting point is the Randles circuit, which combines four components: the ohmic resistance of the solution, a capacitor representing the electrical double layer that forms at the electrode surface, a resistor representing charge transfer, and a Warburg element for diffusion.
In practice, real electrodes rarely behave like perfect capacitors. The double-layer capacitance is often replaced with a constant phase element (CPE), which uses an exponent between 0 and 1 to capture the imperfect, “leaky” capacitor behavior seen in rough or porous surfaces. An exponent of 1 means a perfect capacitor; values closer to 0.5 or below suggest something more complex is happening and should be interpreted carefully.
For battery diagnostics, this approach is especially useful. The Nyquist plot of a lithium-ion battery electrode typically shows features for charge transfer resistance, diffusion impedance, and capacitive storage. Twin semicircles at intermediate frequencies can appear when multiple processes overlap, such as ionic resistance in the electrolyte and electron transfer at the electrode surface. Tracking how these features change over charge cycles gives engineers a non-destructive way to monitor battery health and aging.
Nyquist Plots vs. Bode Plots
Both Nyquist and Bode plots display the same underlying frequency response data, just organized differently. A Bode plot uses two separate charts with frequency on the horizontal axis: one for magnitude and one for phase. This makes it easy to read off specific values at specific frequencies and is generally more intuitive for designing controllers.
The Nyquist plot sacrifices that frequency readability for a compact, single-curve view that excels at stability analysis. Its main strengths are the encirclement criterion for determining closed-loop stability and its role as the foundation for robustness analysis in control theory. Its main weakness is that it can be harder to visualize, especially for complex systems where the curve loops and crosses itself multiple times. The general rule in engineering practice: use Bode plots for design and synthesis, and switch to the Nyquist plot when you need a definitive stability verdict.