ESC failure modes are the distinct ways an ESC degrades in service. They include loss of clamping force, surface erosion, helium leak-up, residual clamping, electrical degradation, thermal non-uniformity and coating wear. Each one shows up as a recognisable symptom at the tool. Correct diagnosis is what separates a quick recondition from a scrapped ESC or a wrongly blamed process excursion. This article works through the main failure modes in turn. For each one, it gives the symptom an engineer actually observes, the likely physical cause, the diagnostic check to confirm it, and the typical remedy.
It is a supporting article in our ESC repair and refurbishment cluster. It focuses on diagnosis at the tool before an ESC is pulled for repair or replacement.
Why disciplined diagnosis matters
An ESC is a consumable. Most of the ways it wears out are gradual rather than sudden. An ESC rarely stops working overnight. Far more often, clamp force drifts down over months, particle counts rise slowly, or a single zone starts running warm. These symptoms overlap with chamber, gas-panel and power-supply faults. So on a busy fab floor, teams often swap the ESC based on suspicion alone. That wastes a serviceable ESC body and the tool time needed to fit a replacement. It also does not fix the problem if the real fault was upstream.
Good diagnosis answers three questions in order. Is the symptom real and repeatable? Is the ESC actually the cause? Can the damage be recovered by a recondition (L1 surface recondition, roughly 25% of new list price)? Or does it need a full L2 rebuild (roughly 70%)? An ESC can typically be refurbished more than three times across its life. Confirming the failure mode before acting protects the ESC body. It also protects the cost-of-ownership case for repairing rather than replacing it.
ESC failure modes: symptom, cause, check and remedy
The table below is the working reference. It lists the eight failure modes a maintenance engineer meets on legacy 150 mm and 200 mm tools. For each mode it shows what is seen at the tool, the underlying cause, the check that confirms it, and the usual remedy. Each mode is then explained in the sections that follow.
| Failure mode | Symptom at the tool | Likely cause | Diagnostic check | Typical remedy |
|---|---|---|---|---|
| Dielectric wear / flatness loss | Drifting clamp force, poor He sealing, edge cooling errors | Eroded or out-of-flat clamping surface | Flatness/profilometry vs single-digit µm spec | L1 re-lap and re-surface; L2 if dielectric thinned |
| Surface erosion & mesa wear | Rising backside particle counts, defectivity excursions | Worn or chipped mesas; plasma-eroded finish | Mesa-height inspection, particle/defect map | L1 re-surface and re-mesa, re-test particles |
| Clamping-force loss | Wafer slips or won’t hold against He | Surface wear, dielectric ageing, charge leakage | Clamp-force-vs-voltage curve; He leak-up test | L1 recondition; L2 if dielectric or electrode at fault |
| Helium leak-up | He flow high, backside pressure won’t hold, wafer runs hot | Lost seal at clamp surface, cracks, O-ring/seal wear | He leak-back test at set clamp voltage | L1 re-surface; investigate seals and body |
| Residual clamping / slow dechuck | Wafer sticks, slow release, lift-pin events, breakage | Trapped charge; ageing or contaminated dielectric | Dechuck-time measurement; surface inspection | Tune dechuck sequence; L1/L2 if dielectric aged |
| Electrical degradation | Rising leakage current, arcing, HV trips | Feed-through or electrode breakdown; cracks; contamination | Leakage-current and HV insulation test | L2 electrode/feed-through rework |
| Thermal non-uniformity | CD or film non-uniformity, hot/cold zones | Bond voids, blocked He channels, zone faults | Thermal-uniformity map; He distribution check | L2 strip and rebuild bond/channels |
| Coating / finish degradation | Discoloration, flaking, drifting performance | Plasma-eroded protective coating or finish | Visual and surface inspection | L1 strip and re-coat / re-finish |
Dielectric wear and loss of flatness
The clamping surface of an ESC is held flat to single-digit micrometres. This lets the wafer sit flat and keeps the helium gap uniform. Years of plasma exposure and thermal cycling slowly erode and distort that surface. The engineer sees three signs. Clamp force has drifted below spec. Helium no longer seals cleanly. Edge or zone cooling errors appear that were not there a year ago. The cause is loss of flatness and thinning of the dielectric where erosion is worst. Confirm it with a flatness measurement or surface profilometry against the original specification. An ESC that has lost its flatness is the most common reason for slowly degrading performance. The remedy is usually an L1 surface recondition. The clamping face is re-lapped and re-surfaced back to flat. This works as long as the dielectric still has enough thickness left. Sometimes erosion has thinned the dielectric so much that re-surfacing would compromise it. In that case the ESC moves to an L2 rebuild.
Surface erosion and mesa wear (rising particles)
The clamping surface carries mesas or a textured pattern. This pattern defines the real wafer contact area and the volume available for helium. As those mesas erode, chip or wear round, two things happen. The contact geometry changes, which shifts clamp force and helium behaviour. And the worn ceramic begins shedding material onto the wafer backside. The first symptom engineers notice is usually a rising backside particle count. It may also be a defectivity excursion that tracks back to specific chambers. Confirm it in two ways. Inspect mesa height and condition. Then map particles or defects to the ESC rather than the process. The remedy is an L1 re-surface that restores the mesa pattern and finish. A particle and defectivity re-test follows before the ESC returns to service. Catching mesa wear early is one of the practices covered in our guide to extending ESC lifetime. An ESC reconditioned on a planned schedule never reaches the stage where particle counts rise out of control.
Clamping-force loss
Loss of clamping force is the failure mode with the most operational consequences. Clamp force is what holds the wafer flat against the helium backside pressure pushing it up. When force drops, the wafer slips or fails to seal. In the worst case it lifts. The cause is usually some combination of surface wear, dielectric ageing and charge leakage through a degraded ceramic. The definitive check is a clamp-force-versus-voltage curve. Raise the helium backside pressure until the wafer begins to leak or lift. The pressure at that point is the usable clamp force at that voltage. Compare it with the original figure. An ESC that once held 20–80 Torr-equivalent at 600–2,000 V DC but now leaks well below its old figure has lost force measurably. Our note on ESC clamping force describes the measurement in full. The remedy depends on the root cause. Surface-driven loss recovers with an L1 recondition. Loss traced to the dielectric or electrode needs an L2 rebuild.
Helium leak-up and poor backside cooling
Helium is fed into the micro-gap behind the clamped wafer, typically at 5–40 Torr. There it conducts heat from the wafer to the cooled ESC body. If the ESC can no longer seal that gas, three things happen. Helium flow climbs to hold pressure. Backside pressure will not stabilise. The wafer runs hot, which shows up as drifting critical dimensions or film properties. The cause is a lost seal at the clamping surface, surface cracks, or worn perimeter seals and O-rings. Confirm it with a helium leak-back test performed at the normal clamp voltage. With the wafer clamped, isolate the helium supply. Then watch how fast backside pressure decays. The fix for a surface-seal problem is an L1 re-surface. Cracks or body and seal faults are investigated during strip-down. They may require an L2 rebuild. The cooling mechanism, and why force must exceed helium pressure, are covered in helium backside cooling.
Residual clamping and slow dechuck
Residual clamping is charge that stays trapped in the dielectric after the clamp voltage is switched off. The wafer then sticks instead of releasing cleanly. The engineer sees slow or failed dechuck and lift-pin events. In the worst cases there is wafer breakage on lift. A healthy ESC releases in under 1–5 seconds. An ESC whose dechuck time has grown, or that the standard sequence can no longer clear, is signalling a problem. Rule out sequence tuning first. A reverse-polarity dechuck pulse and plasma-assisted discharge should drain the stored charge before the lift pins move. Johnsen–Rahbek ESCs store charge in the ceramic surface by design. So they are more prone to this than Coulombic ESCs. Sometimes a correctly tuned sequence still leaves the wafer stuck. The cause is then usually an ageing or contaminated dielectric. The ESC needs an L1 recondition or, if the dielectric is genuinely aged, an L2 rebuild. The full failure mechanism is set out in dechucking and residual clamping.
Electrode and feed-through electrical degradation
Beneath the dielectric sit the buried electrode or electrodes. A high-voltage feed-through connects them to the supply. Over time, ceramic cracks, moisture ingress, contamination or feed-through breakdown raise the leakage current. They can also cause arcing. The symptoms are clear at the tool. There are rising leakage current readings, audible or visible arcing, and high-voltage supply trips that interrupt processing. Confirm electrical degradation with a leakage-current measurement and a high-voltage insulation test at the rated clamp voltage. An ESC drawing more leakage than its baseline has an insulation fault. So does one that flashes over below its rated voltage. This is not a surface problem, so it is not an L1 fix. It requires an L2 strip and rebuild to rework the electrode or feed-through and restore the insulation. Catch a rising-leakage trend before it becomes hard arcing. This avoids further damage to the dielectric and to the power supply.
Thermal non-uniformity
An ESC removes heat through a cooled base plate and a network of helium and coolant channels. These are often split into radial zones for tighter uniformity. The ESC stops cooling evenly when a bond layer develops voids, a helium channel partially blocks, or a zone control fails. The process result is critical-dimension or film non-uniformity that maps to a fixed pattern on the wafer. Examples are a warm centre, a cold edge, or a zone that no longer responds. Confirm it with a thermal-uniformity map and a check of helium distribution across the zones. This separates an ESC fault from a chamber or gas-panel fault. Thermal non-uniformity from bond voids or channel blockage is an internal problem. So it generally calls for an L2 strip and rebuild to remake the bond and clear the channels. The symptom looks like several chamber faults. So confirming the ESC is the cause before pulling it is especially important here.
Coating and finish degradation
Many ESCs carry a protective coating or engineered surface finish. This layer absorbs most of the plasma exposure. As that layer erodes, it discolours, flakes or wears through. ESC performance then drifts in ways that follow surface condition rather than any single electrical or thermal parameter. The check is largely visual and surface inspection. Discoloration, flaking and exposed substrate are clear signs the finish is worn out. The remedy is an L1 strip and re-coat or re-finish. This restores the protective surface before the underlying dielectric is attacked. Treated early, coating degradation is a routine recondition. Left too long, it lets plasma reach and erode the dielectric. That escalates an L1 job into an L2 one.
Turning a diagnosis into a repair decision
Once the failure mode is confirmed, the decision is straightforward. Surface-led failures are handled by an L1 recondition at roughly 25% of new list price. These include flatness loss, mesa wear, helium-seal loss at the surface, and coating degradation. Failures that reach the dielectric, electrode, feed-through or bond call for an L2 rebuild at roughly 70%. These include hard electrical degradation, deep dielectric ageing, and thermal non-uniformity from voids. Either way, a properly refurbished ESC is returned to as-new clamping and particle performance, and warranted. This follows a full functional test covering clamp force versus voltage, helium leak-back, particle and defectivity, and thermal uniformity. On legacy 150 mm and 200 mm tools, a new OEM ESC may carry a long lead time or a high price. It may also be discontinued outright. So confirming the failure mode and repairing the ESC you already own is usually the lowest-cost route back to spec.
Frequently asked questions
What are the most common ESC failure modes?
The most common are clamping-force loss, surface erosion and mesa wear that drive up backside particles, and helium leak-up with poor cooling. Others are residual clamping and slow dechuck, electrode or feed-through electrical degradation, thermal non-uniformity, and coating wear. Most are gradual. So they show up as drifting performance rather than a sudden stop.
How do I diagnose loss of clamping force at the tool?
Run a clamp-force-versus-voltage check. With the wafer clamped, raise the helium backside pressure until the wafer leaks or lifts. Read the usable clamp force at that point. An ESC that once held 20–80 Torr-equivalent at 600–2,000 V DC but now leaks well below its old figure has measurably lost force.
Does a confirmed failure mode mean L1 or L2 repair?
Surface-led failures are usually recovered by an L1 recondition at about 25% of new list price. These include flatness loss, mesa wear, surface helium-seal loss and coating degradation. Failures reaching the dielectric, electrode, feed-through or bond need an L2 rebuild at about 70%. These include electrical degradation and thermal non-uniformity from voids.