Voltage Control of Capillary Electrophoresis High-Voltage Power Supplies in Drug Screening and Sequencing Support

Capillary electrophoresis separates analytes by the ratio of charge to hydrodynamic radius, and the migration time is proportional to the applied voltage divided by the capillary length. A thirty-kilovolt supply driving a fifty-centimetre capillary at fifty microlitres of total volume produces migration times of a few minutes, and a voltage error of one percent shifts every peak by the same percentage. In a screening assay with fifteen peaks, that shift is enough to move a peak outside its identification window.

Current is tiny and the load is capacitive. A capillary filled with buffer is a resistor of a few microsiemens in series with a capacitance of tens of picofarads, and the supply sees a load that changes as the buffer heats. Joule heating is the dominant thermal effect: at thirty kilovolts and twenty microamps, the capillary dissipates six hundred milliwatts over a volume of a few microlitres, and the temperature rise changes the viscosity and the mobility by several percent.
Voltage stability is therefore the primary requirement, and it must be stability at the capillary, not at the supply terminal. Resistive drop along the electrode and the capillary connection appears as a voltage error that grows with current, and current grows with temperature. A four-terminal connection, with the sense pair brought to the capillary ends, removes most of the error and makes the reading meaningful.
Ramp control matters more than steady-state accuracy for separation quality. A fast ramp at the start of a run produces a pressure wave in the capillary that disturbs the sample plug; a slow ramp takes minutes and reduces throughput. The usual solution is a controlled ramp of a few seconds, with the shape defined by the supply, and a delay before injection to let the plug settle.
Polarity reversal is a routine operation in methods that use both electroosmotic and electrophoretic transport. The reversal must happen with the output at zero, and the zero must be verified rather than assumed. A reversal performed with residual voltage produces a current surge that damages the capillary inlet and creates bubbles that terminate the run.
Bubble formation is the failure mode that links the supply to the fluidics. A bubble at the detection window produces a step change in current and a discontinuity in the trace, and the cause is usually a local hot spot from a voltage transient. Supplies with current limiting that acts slowly allow the transient to persist; supplies with fast limiting prevent it. The current signal itself is the best diagnostic, since a bubble appears as a characteristic current drop.
Multi-channel instruments put several capillaries under one supply, and the channels interact. A single supply feeding eight capillaries through individual current limiters produces cross-talk when one channel arcs, because the common output droops and the other channels see a voltage dip. Independent supplies per channel, or a supply with per-channel regulation, avoids the coupling at a cost.
Safety design is dominated by the interface between high voltage and liquid. The buffer reservoir is at high potential, and any leak path through the fluid to the operator is a hazard. The supply must be interlocked with the sample compartment, the enclosure must be grounded, and the high-voltage electrode must be isolated from the operator by a physical barrier that cannot be opened with the output live.
Current measurement resolution determines how well the separation can be monitored. At twenty microamps full scale, a resolution of twenty nanoamps allows the current trace to be used for quality assessment; a resolution of two microamps only shows gross changes. The sense resistor and the amplifier's noise floor set this, and both are cheap to improve at the design stage and expensive afterwards.
Temperature control interacts with the electrical loop. A capillary held at twenty-five degrees has a stable mobility, and the supply's power dissipation becomes the only heat input. A capillary without active cooling sees the supply's output power heat the buffer, reduce the resistance, increase the current, and increase the heating further. The loop is unstable in a way the supply cannot correct, and the cure is thermal rather than electrical.
Sequencing support applications use higher voltages and longer capillaries, which shifts the requirements toward stability over hours rather than minutes. A run of several hours at thirty kilovolts with a drift of a tenth of a percent is acceptable, but a drift of one percent moves the peaks far enough to corrupt base calling. Drift over hours is dominated by the divider and the reference, and by thermal transients in the supply itself.
Electrode behaviour is a slow variable. The anode in contact with buffer evolves gas and corrodes, and the contact resistance changes over a run; the supply's voltage regulation absorbs the change and the current tells the story. Replacing electrodes on a schedule derived from current drift, rather than from run failures, keeps the separation consistent.
Diagnostics should distinguish a supply fault from a capillary fault. A supply in voltage mode with current falling to near zero indicates a broken capillary or a bubble; a supply whose voltage falls while current is constant indicates a supply fault. Recording both quantities continuously makes this distinction available without touching the instrument.
Calibration of the voltage reading is usually neglected. The supply's internal divider is calibrated at manufacture, and the capillary connection resistance is not. Periodic calibration with an external divider at the capillary end, at the operating voltage and current, closes the gap between displayed and delivered voltage and removes a systematic error that affects every method.
Electromagnetic compatibility is modest by high-voltage standards but still relevant. The supply sits next to a detector with a photomultiplier or a fluorescence channel, and switching noise coupled into the detector appears as baseline noise. Shielding the supply, using a linear post-regulator for the low-noise channel, and keeping the detector's signal leads away from the high-voltage cable are the practical measures.
Cost trade-offs centre on the regulation and measurement chain. A supply with a precision divider, a four-terminal connection and a high-resolution current sense costs more than a general-purpose unit, and the difference is recovered by fewer failed runs and less method revalidation. In a screening laboratory, a single failed plate costs more than the difference.
Limits should be stated. At thirty kilovolts the capillary's own Joule heating sets the practical current limit, and no supply can deliver more current without changing the separation. Similarly, a capillary with a partially blocked inlet will show low current and unstable migration regardless of the supply's quality. The supply provides a stable field; the separation depends on the chemistry and the capillary.
An acceptance test can be built around a standard test mixture. Run a mixture of known analytes at three voltages, record migration times and current, and compare the ratios; then run the same mixture six times in succession at one voltage and compare the reproducibility. The first test checks the voltage accuracy, the second checks the stability, and both are more informative than a bench measurement of ripple.
Sample injection is where the supply and the fluidics meet. Electrokinetic injection uses the same voltage that drives the separation, and a supply that cannot hold a precise low voltage for a few seconds delivers an injection plug whose size varies. Pressure injection avoids the dependency but adds a pump, and the choice between the two determines how much of the separation repeatability rests on the supply.
Buffer composition changes the load in ways that are predictable but rarely characterised. Ionic strength sets the resistance, and resistance sets the current at a given voltage; a method transferred from one buffer to another without rechecking the current will run at a different Joule heating level and a different mobility. The supply's current display is the fastest way to detect the change.
Detection coupling is a mechanical as well as an electrical concern. The detection window sits a few centimetres from the high-voltage electrode, and the electric field at that point can couple into the detector's front end. A grounded guard between the capillary and the detector, and a supply with a well-controlled output, together keep the baseline quiet enough for trace analysis.
Method transfer between instruments fails most often on voltage calibration. Two instruments whose displayed voltages differ by half a percent produce migration times that differ by half a percent, and a method validated on one will fail its acceptance criteria on the other. Calibrating both against an external divider, and recording the correction in the method, makes transfer predictable.
Throughput pressure encourages higher voltages and shorter capillaries, both of which push the Joule heating limit. The supply's role is to hold the voltage while the capillary heats, and a supply with a slow current limit will allow the buffer to boil before it reacts. Fast current limiting, set below the boiling threshold for the buffer, protects the run and the capillary.
Records of current and voltage per run support both troubleshooting and method validation. A run whose current trace shows a step indicates a bubble or a blockage; a run whose voltage trace sags indicates a supply or connection problem; a run whose current rises steadily indicates heating. Reading the two traces together answers most questions before any hardware is touched.