An ESC power supply is the high-voltage DC source that drives an ESC. It delivers the stable clamp voltage, typically 600–2,000 V DC, that holds the wafer flat against the ESC during processing. It also limits the output current, monitors leakage, and manages the controlled dechuck sequence that releases the wafer at the end of a run. This article explains what the supply does and how monopolar and bipolar outputs differ. It also explains why leakage current matters for maintenance, and how the supply integrates with the tool.
For the full scope of ESC design, materials and applications, see our complete guide to ESCs.
What the ESC power supply does
The job of the power supply is to put a stable, well-regulated DC voltage on the ESC electrode. It must then hold that voltage for the whole process step. The induced clamping force scales with the square of that voltage. So any drift or ripple in the output directly changes the holding force. That in turn changes helium backside cooling and wafer temperature. A production supply regulates its output tightly across the working range. It holds the output steady against changing chamber conditions. It also feeds back the actual voltage and current it is delivering. The tool controller can then confirm the clamp is healthy before the recipe proceeds. The table below summarises the parameters a maintenance or process engineer works with.
| Parameter | Typical value / role |
|---|---|
| Output voltage | 600–2,000 V DC, regulated to the recipe setpoint |
| Output configuration | Single output (monopolar) or dual opposite-polarity outputs (bipolar) |
| Polarity | Selectable / reversible, for clamp and dechuck |
| Current limit | Sub-milliamp to low-milliamp ceiling, capped in hardware |
| Leakage current monitor | Reports steady-state current through the dielectric; a health signal |
| Voltage ramp rate | Controlled rise on clamp, controlled fall on dechuck |
| Dechuck function | Ramp-down plus timed reverse-polarity pulse |
| Interlocks | Door, vacuum, lift-pin and tool-state signals that inhibit HV |
Single output versus dual outputs: monopolar and bipolar
The number of outputs the supply provides follows directly from the electrode design of the ESC it drives. A monopolar ESC has one electrode and needs the surrounding plasma to complete the circuit. So it is driven by a single high-voltage output referenced to the tool. A bipolar ESC has two electrodes held at opposite polarity. It needs a supply with two outputs, one positive and one negative, of equal magnitude. The bipolar circuit closes through the wafer itself rather than through the plasma. So that ESC clamps with or without plasma present, and it keeps the wafer near ground potential. This is why most modern tools pair a bipolar ESC with a dual-output supply. The clamping mechanics behind both arrangements are covered in our guide to ESC clamping force.
Current limiting and why leakage current matters
Electrically, an ESC is a high-value capacitor. Once the wafer is clamped and the voltage has settled, almost no current should flow. The supply is current-limited in hardware to a low ceiling. This means an arc, a flashover or a wafer-handling fault cannot dump large energy into the ESC or the wafer. That same low ceiling is what makes leakage current such a useful diagnostic. In normal operation a small, steady leakage current flows through the dielectric. The magnitude depends on the regime. A slightly conductive Johnsen–Rahbek ceramic passes more current than a near-insulating Coulombic one.
For maintenance, the trend matters more than a single reading. A leakage current that rises over time at a fixed voltage and temperature signals that the dielectric is degrading. Typical causes are pinholes, micro-cracks, surface erosion under plasma, or contamination tracking across the surface. Left unchecked, rising leakage ends in a flashover or a hard short that takes the ESC out of service. Log the leakage current the supply reports, and watch for a climb from run to run. This gives an early warning that an ESC will soon need to come off the tool. Persistent high leakage is one of the clearest triggers for refurbishment. In refurbishment the dielectric surface is reconditioned or rebuilt, and the ESC is returned to as-new clamping and particle performance with a full functional test.
The controlled dechuck sequence
Releasing the wafer cleanly is as much the supply’s job as clamping it. Simply switching the high voltage off leaves charge trapped in the dielectric. The wafer then stays stuck, and it can crack or fly off when the lift pins push. To avoid this, the supply runs a controlled dechuck sequence. It first ramps the clamp voltage down rather than cutting it. It then applies a short reverse-polarity pulse sized to neutralise the residual charge. This is often combined with a plasma-assisted discharge that gives the trapped charge a path to ground. The aim is to bring the net charge on the dielectric to near zero before the pins move.
Johnsen–Rahbek ESCs need this sequence handled more carefully than Coulombic ones. They store charge in the ceramic surface by design, and they release more slowly. A well-tuned dechuck clears the wafer in under one to five seconds. Sticking that the standard sequence cannot clear points to an ageing dielectric, not a supply fault. The failure mode and its remedies are detailed in our guide to dechucking and residual clamping.
Monitoring, interlocks and high-voltage safety
The supply works at up to 3000 V DC, so it is built around layered protection. It continuously monitors its own output voltage and current, and it reports both to the tool controller. The controller compares them against window limits and faults the recipe if the clamp is out of spec. Interlocks tie the high voltage to the state of the tool. The chamber door, the vacuum level, the lift-pin position and the overall tool-ready signal all have to be correct before the output can energise. If any one of them drops out, the high voltage is inhibited or shut down. Three further protections complete the design: hardware current limiting, output bleed resistors that discharge stored energy when the supply is off, and clear lockout/tagout points for service. For a maintenance engineer the practical rule is simple. Treat the ESC and its cabling as live until the supply is off, interlocked and confirmed discharged.
Integration with the process tool
The power supply does not work alone. It is one node in the tool’s control architecture. The recipe sets the clamp voltage, ramp rates and dechuck timing. The supply executes them on command from the system controller, while streaming voltage, current and fault status back. Clamp and dechuck are sequenced against wafer handling, plasma state and helium delivery. So the wafer is only clamped when it should be, and only released once it is safe to lift. High-voltage feedthroughs and shielded cabling carry the output to the buried electrode. A marginal feedthrough shows up as erratic leakage current or unstable clamp force. When a supply, an ESC or the cabling between them is suspect, the first step is to isolate which of the three is at fault. A known-good replacement ESC makes that much quicker.
Maintenance, calibration and troubleshooting
Most preventive work on an ESC power supply comes back to one habit. Trend the numbers the supply already reports, rather than reading them once. Leakage current is the single most useful trend. A clamped ESC is electrically a capacitor. So at a fixed voltage and temperature, the steady-state leakage should sit at a near-constant value. When that value climbs from run to run, the dielectric is degrading. Typical causes are pinholes, micro-cracks, plasma erosion of the surface, or contamination tracking. The climb usually appears weeks before a hard failure. Log leakage against voltage and ESC temperature, and set an alarm band around the established baseline. This turns the supply into an early-warning instrument for the ESC, not just a clamp source.
High-voltage arcing is the failure the current limit exists to contain, and its causes are worth recognising. Arcing and flashover tend to come from a breached or eroded dielectric, or from contamination bridging the clamp surface or a feedthrough. Moisture, poor vacuum at the high-voltage path, or a cracked ceramic are other causes. Each arc event injects a fast current transient that the supply should clamp in hardware. Repeated arcs at a stable recipe point are a symptom, not a nuisance. They justify pulling the ESC for inspection before the dielectric fails completely.
Output voltage calibration keeps the rest of the system accurate. Clamping force scales with the square of the applied voltage. So a supply reading 2000 V DC while delivering 1900 V DC under-clamps by roughly ten percent in force. That can let the helium backside pressure lift the wafer. Calibrate the supply periodically against a traceable high-voltage divider. Check both the setpoint accuracy and the reported feedback. This confirms the supply is delivering what the recipe asks, and that its monitor channel can be trusted for diagnostics. The dechuck sequence deserves its own verification. Confirm that the ramp-down, the reverse-polarity pulse and any plasma-assisted discharge actually clear the wafer within the expected one-to-five-second window. Confirm also that lift-pin force stays low. A dechuck that has quietly drifted is a wafer-breakage risk that no single clamp reading exposes.
Interlock and safety checks complete the routine. The door, vacuum, lift-pin and tool-ready signals that inhibit the high voltage should be tested as part of preventive maintenance, not assumed. The output bleed path that discharges stored energy on power-down should be confirmed before anyone touches the ESC or cabling. When a fault does appear, the practical question is whether it points to the supply or to the ESC itself. Some symptoms follow the supply. These include the output failing to reach setpoint, and voltage ripple or instability seen on the monitor independent of the chamber. Faults that persist when a known-good ESC is substituted also point to the supply. Other symptoms point to the ESC or its cabling. These include leakage current that rises with ESC temperature or run count. They also include sticking that a healthy dechuck sequence cannot clear, and arcing localised to one clamp zone. Swapping in a known-good replacement ESC is the fastest way to separate the two. This is why keeping one on the shelf shortens every high-voltage troubleshooting session.
Matching a supply to the ESC and tool
A power supply and an ESC are a matched pair, and the match has to hold on several points at once. The first point is the clamping regime and voltage range. A Coulombic ESC is built around a near-insulating dielectric. It typically needs the upper part of the 600–2,000 V DC band to reach working clamp force. A Johnsen–Rahbek ESC reaches the same force at lower voltage, but it draws more leakage and releases more slowly. A supply has to cover the regime the ESC actually uses. It needs enough range, tight regulation, and a current limit set for the leakage that regime produces. The difference between the two regimes is explored in our guide to Coulombic versus Johnsen–Rahbek ESCs.
The second point is electrode design, and there is no flexibility here. A monopolar ESC has a single electrode and completes its circuit through the plasma. So it needs a single high-voltage output referenced to the tool. A bipolar ESC has two opposite-polarity electrodes. It needs a supply with two outputs of equal magnitude that close the circuit through the wafer. The two are not interchangeable. Driving a bipolar ESC from a single output will not clamp correctly, and it can stress the dielectric. The same applies to a monopolar ESC driven from a mis-wired dual-output supply. The trade-offs between the two electrode layouts are covered in our comparison of monopolar and bipolar ESCs.
Configuration is where mismatches usually appear. The highest-risk moment is when an ESC is replaced or refurbished. The recipe setpoint, ramp rates, current limit and dechuck timing all have to be re-confirmed against the new ESC’s regime and electrode count. They must not be carried over without checking, because a refurbished ESC can behave slightly differently from the unit it replaces. We supply OEM-equivalent replacement ESCs, and we refurbish worn ones back to as-new clamping and particle performance with a full functional test. That test data gives the engineer the voltage, leakage and dechuck behaviour needed to set the supply correctly on the first install. Confirm the regime, the output configuration and the dechuck parameters before the first production wafer. This is the cheapest insurance against a stuck or under-clamped wafer.
Frequently asked questions
What voltage does an ESC power supply output?
Typically 600–2,000 V DC, regulated to the recipe setpoint. Coulombic ESCs tend to run at the higher end, while Johnsen–Rahbek ESCs reach a given clamp force at lower voltage. The supply holds that voltage steady throughout the process step, because clamp force scales with the square of the applied voltage.
Why does an ESC power supply monitor leakage current?
A clamped ESC behaves like a capacitor, so only a small steady leakage current should flow through the dielectric. A leakage current that rises over time at the same voltage and temperature signals that the dielectric is degrading. That is an early warning of a coming short. Logging the trend from run to run flags an ESC for inspection before it fails.
Does a bipolar ESC need a different power supply from a monopolar one?
Yes. A monopolar ESC uses a single high-voltage output and relies on plasma to complete the circuit. A bipolar ESC needs two opposite-polarity outputs of equal magnitude, so the circuit closes through the wafer. This lets it clamp with or without plasma and stay near ground potential.