Reading results
A capacitance–voltage sweep steps the applied bias and reports the small-signal capacitance at each point, then fits it to pull out a doping concentration and a built-in potential.
01Setting up a sweep
Tick CV calculation and press Run Simulation. A C-V Sweep Settings dialog opens before the sweep starts:
- V start (V), V stop (V), V step (V) — the applied terminal voltage is V = Vtop − Vrear. Enter a positive step magnitude; the dialog chooses the sign needed to move from start to stop.
- Contact type — Schottky–Ohmic, Ideal Ohmic–Ohmic or HF MOS / MIS. This chooses the model for the analysis.
- Schottky barrier ΦBn (eV), with Auto-set n-type flat-band ΦBn — the conduction-band electron barrier at the top interface, used for the Schottky case only. The auto option computes it as Ec − Ef in the neutral top layer; for a p-type Schottky, enter the electron barrier by hand.
- Stop after failed points and Max sweep time (s) — guards that end the sweep early if too many bias points fail to converge, or a time cap is reached. When either is activated, the Summary tab says so.
The sweep is a high-frequency C–V. In each doped region the minority carriers are frozen out, so no inversion layer forms: the band bending and depletion width at each voltage are set by the majority carriers and the ionised dopants alone, and the minority carriers add nothing to the capacitance. Keeping inversion out of the picture is what holds 1/C² linear for the doping extraction, and it is why the Summary reports Minority carriers: frozen for C-V.
What counts as “high frequency” — and why it is almost always met. The assumption holds whenever the minority carriers cannot be thermally generated fast enough to form an inversion layer or follow the test signal. In the wide-gap compounds this tool is built around (GaAs and above) that response is very slow — the small intrinsic carrier density makes generation take seconds or longer — so the threshold sits well below a hertz and every practical measurement is high-frequency.
The exception is the narrow-gap elemental semiconductors, where the larger intrinsic carrier density speeds the minority response and lifts the threshold: silicon into the hertz-to-kilohertz range, and germanium highest of all, up to a few kilohertz. Even that is low by the standards of modern electronics — C–V is normally measured from tens of kilohertz up to a megahertz — so the frozen-minority picture still applies. Only a deliberately low-frequency measurement on germanium or silicon, below that threshold, would show the minority carriers following the signal and the inversion capacitance rising back up, which this tool does not reproduce.
02The results window
The C–V results open in their own window with four tabs:
- Summary — point count and any early-stop status, the bias start / stop / step, the contact configuration, the permittivity used, and the capacitance (nF/cm²) and free charge density (C/cm²) at the two ends of the sweep.
- 1/C² Analysis — a straight-line fit to 1/C² against bias over the reverse-bias region. Reports the fit window, slope, intercept and R², then the Extracted doping concentration (from the slope) and the Built-in potential (from the intercept, plus kT/q per depleted side for the majority-carrier tail — one term for a one-sided or Schottky junction, two for a symmetric pn). This extraction assumes a bare depleted semiconductor surface, so it is not applied to MOS/MIS structures (there the doping needs the oxide capacitance de-embedded and the intercept is a flat-band, not a built-in, quantity); for those, read the oxide capacitance from accumulation and use the depletion width instead.
- Depletion / Field — the depletion width estimated from the capacitance (W ≈ εs/C) at each end of the sweep, and the range of swept terminal free charge. For a MOS/MIS structure the oxide capacitance is de-embedded first (W = εs/Cdep with 1/Cdep = 1/C − 1/Cox), so the width is the semiconductor’s; accumulation points show no depletion.
- Raw Data — the full sweep as tab-separated columns, V, capacitance, free charge and 1/C², ready to copy into a spreadsheet.
03What the extraction assumes
Every quantity here is per unit area — F/cm² and C/cm², not a total capacitance. The doping extraction assumes depletion happens in one uniformly doped region (a metal–semiconductor or strongly asymmetric junction); the ideal ohmic–ohmic case instead treats a symmetric pn junction where both sides deplete. Read the extracted doping and built-in potential as a first-order figure for that geometry.
The extraction is deliberately conservative about sign. It expects 1/C² to fall as the applied bias increases; if the fitted slope comes out positive — a wrong-polarity or reversed-orientation sweep — it reports the fit but declines to extract, rather than hand back a negative doping concentration.
A C–V run does not draw the band diagram. The plot would show only the first, biased sweep point, which is not the equilibrium picture. To see the bands, run once with CV calculation unticked first, then enable it for the sweep. For copying the results text and saving to a file, see Copying, Saving and Exporting.
