PMT Voltage Optimization: Walking Through Voltration and the Peak-2 Method

A dim activation marker that resolved cleanly on the core facility’s Fortessa came up as an unseparated smear on the second instrument — same panel, same antibody, voltages copied over from the first machine. Copied voltages do not transfer, because every detector has its own noise floor and its own response curve. PMT voltage optimization fixes the detector gain at the point where dim populations separate from electronic noise without wasting dynamic range, and the standard way to find that point is voltration using the peak-2 method. This walks the procedure through with real numbers.

What PMT voltage actually sets

The photomultiplier tube voltage is the gain applied to the photons hitting a detector. Too low, and dim signal falls into the electronic noise of the instrument and cannot be separated from a true negative. Too high, and you gain nothing — signal and noise amplify together — while bright populations climb toward the top of scale and risk going off-scale. The optimum is the lowest voltage at which electronic noise no longer limits resolution. Below it you lose dim populations; above it you only spend dynamic range.

The peak-2 voltration method

The approach published by Maecker and Trotter uses multi-peak rainbow calibration beads — the same 8-peak beads used for daily QC with calibration beads. You run a dim reference peak (conventionally the second-from-bottom peak, “peak 2”) across a range of voltages for one detector and record the robust CV of that peak at each step. Peak 2 is chosen because it is dim enough to sit near the noise floor, so it is sensitive to exactly the region you are trying to optimize.

Plot rCV against voltage and you get a curve that drops steeply, then flattens. The flat region is where electronic noise has stopped dominating and added voltage no longer tightens the peak. The optimum sits just past the knee of that curve.

Confirming the choice with stain index

The voltration curve finds where noise stops limiting resolution; stain index confirms that the chosen voltage actually separates a real stained population from its negative. Stain index normalizes the gap between positive and negative by the spread of the negative:

where the denominator uses the robust standard deviation of the negative population. Take a single-stained control at the candidate 500 V: a positive median of 8200, a negative median of 230, and a negative rSD of 92.

Now push the same control to 600 V. The positive median rises to 11000 and the negative to 360, but the negative rSD widens to 140:

Higher voltage did not improve separation — the stain index actually fell from 43 to 38, because signal and noise scaled together past the optimum. That is the same conclusion the voltration plateau gave, confirmed on a real stained population. Stain index is the same metric used to choose antibody concentration by titration; here it is read against voltage instead of dilution.

The lower bound: the rSD method

The peak-2 curve finds the knee, but you can also set the floor directly. The rSD method puts the minimum useful voltage where the robust SD of unstained cells reaches roughly 2.5× the robust SD of the electronic noise — written as \(2.5 \times rSD_{en}\). Below that, electronic noise dominates the negative and dim events are unrecoverable. The two methods agree in practice: the rSD floor and the voltration knee land close together, and the optimum is the higher of the two.

Where this breaks

Voltration assumes a tunable PMT with a noise floor you can climb off of. Several common instruments do not fit that model. The Beckman CytoFLEX uses avalanche photodiodes with fixed, factory-set gains, so you adjust gain rather than voltrate a PMT, and its noise characteristics differ. Spectral instruments set a single detector configuration across the full array, optimized at the system level rather than per fluorochrome.

And on any instrument, peak-2 voltration optimizes resolution at the dim end — it does not protect bright populations from going off-scale, which is a separate dynamic-range check you make on the stained sample. Treat the procedure as the right tool for conventional PMT-based cytometers and verify the assumptions before applying it elsewhere.

The full procedure — rainbow-bead voltration, the rSD floor, and stain-index confirmation — is laid out in the advanced-cytometry methods literature; the practical advanced flow cytometry guide and the original methods in Cytometry Part A are the primary references. Optimize each detector once at panel setup, record the curve, and from then on let daily QC beads flag drift rather than re-voltrating by hand — the curve is a setup artifact, not a daily one.