The dielectric material is the ceramic layer between the buried electrode and the wafer. It is the single component that decides how an ESC clamps, cools and survives in a plasma chamber. Its volume resistivity sets the clamping regime. Its thermal conductivity controls wafer temperature uniformity. Its expansion behaviour and plasma resistance decide how long the ESC lasts. This article compares the two dominant ESC dielectrics, alumina (Al₂O₃) and aluminium nitride (AlN), across the properties that matter. It explains where doped and composite ceramics fit. It also links to the detailed material guides in this cluster.
For the wider context of how the ESC itself works and where the dielectric sits in the assembly, see our complete guide to ESCs.
Why the dielectric material is the defining choice
An ESC is a layered assembly of ceramic and metal. It has a dielectric clamping surface over one or more buried electrodes, a bond layer, and a temperature-controlled metal base plate. The base plate carries helium and coolant channels. Almost every functional specification a process engineer cares about traces back to the dielectric ceramic. This includes clamp force per volt, release behaviour, temperature flatness, particle performance and service life. Choosing the dielectric is therefore not a minor decision. It fixes the operating envelope of the whole tool. Five material properties drive that envelope: volume resistivity, thermal conductivity, coefficient of thermal expansion (CTE), dielectric strength and plasma resistance.
The five properties that matter
Volume resistivity — sets the clamping regime
Volume resistivity decides whether an ESC clamps in the Coulombic or the Johnsen–Rahbek (JR) regime. A Coulombic dielectric is a near-perfect insulator above 10¹⁵ Ω·cm. The wafer and electrode behave as the plates of a capacitor. Force scales with the inverse of dielectric thickness. A JR dielectric is deliberately slightly conductive, in the band of roughly 10¹⁰–10¹³ Ω·cm. Charge migrates to the surface, and the effective gap shrinks to the microscopic contact points. This gives far higher force per volt. Across the range of ESC dielectrics in service, bulk resistivity spans roughly 10⁹–10¹⁶ Ω·cm. The relationship between resistivity, regime and temperature is set out in our guide to volume resistivity and clamping regime.
Thermal conductivity — controls temperature uniformity
Plasma deposits significant heat into the wafer. That heat is removed by conduction through the dielectric into the cooled base plate. Helium in the backside gap assists this. The higher the dielectric’s thermal conductivity, the flatter the wafer temperature across the clamp. Alumina sits at roughly 30 W/m·K. Aluminium nitride reaches 150–180 W/m·K, which is around five to six times higher. That difference is why AlN is the default where temperature uniformity is critical. Examples are high-power etch and demanding deposition. It works alongside helium backside cooling, not instead of it.
Coefficient of thermal expansion — controls differential stress
The wafer and the dielectric heat and cool together through every process cycle. If they expand at different rates, the wafer slips or builds stress at the interface. A silicon wafer has a CTE of about 2.6–3.3 ppm/K. Alumina sits at roughly 7–8 ppm/K. AlN is closer at about 4.5 ppm/K. So AlN is the better thermal match. It produces less differential stress and slip across temperature. CTE matching matters most on processes that cycle over a wide temperature range. The mechanism is covered in CTE and wafer matching.
Dielectric strength — sets the safe voltage
Dielectric strength is the breakdown field the ceramic can withstand before it conducts and fails. It is typically 10–20 kV/mm for ESC-grade ceramics. It limits how thin the dielectric layer can be made. It also limits how much chuck voltage the design can safely apply. Production ESCs run at 600–2,000 V DC. So the layer must hold that field with margin across its full service life. It must allow for surface erosion and ageing. A high, consistent dielectric strength is one thing that separates a high-purity ESC-grade ceramic from a general-purpose technical ceramic. ESC-grade ceramic is typically 99.5–99.9% pure.
Plasma resistance — sets service life and particles
The clamping surface is exposed to the same plasma that processes the wafer. So it erodes over time. That erosion drives particle generation and clamp-force drift. Both alumina and AlN resist plasma well by nature. This is why ceramics dominate ESC construction. Aggressive fluorine or chlorine etch chemistries need more protection. A plasma-resistant coating or finish on the exposed surfaces further reduces erosion and particle counts. Yttria (Y₂O₃) is the common choice. Surface finish and coatings are detailed in our guide to plasma-resistant coatings.
Alumina vs aluminium nitride: the property comparison
The two ceramics sit at different points on the same trade-off. Alumina is the robust, cost-effective standard choice. AlN provides better thermal performance and a closer CTE match at higher cost. The table below summarises the properties that separate them.
| Property | Alumina (Al₂O₃) | Aluminium nitride (AlN) |
|---|---|---|
| Volume resistivity | Usually > 10¹⁵ Ω·cm (Coulombic); doping lowers it into the JR band | Naturally in/near the JR band (≈ 10¹⁰–10¹³ Ω·cm) |
| Thermal conductivity | ≈ 30 W/m·K | ≈ 150–180 W/m·K |
| CTE | ≈ 7–8 ppm/K | ≈ 4.5 ppm/K (closer to silicon) |
| Dielectric strength | 10–20 kV/mm | 10–20 kV/mm |
| Typical clamping regime | Coulombic (JR when doped) | Johnsen–Rahbek |
| Relative cost | Lower (standard choice) | Higher |
| Best fit | General etch/PVD/CVD, cost-sensitive and legacy tools | High-power, thermally demanding processes needing tight uniformity |
Read either side in depth in the dedicated guides to alumina (Al₂O₃) ESCs and aluminium nitride (AlN) ESCs.
When alumina is the right choice
Alumina is the default for good reason. It is well characterised, mechanically robust, widely available and cost-effective. As a near-perfect insulator above 10¹⁵ Ω·cm, it operates cleanly in the Coulombic regime. That brings fast, predictable release and low temperature sensitivity. It suits general plasma etch, PVD and CVD where heat loads are moderate and uniformity targets are not extreme. In those cases alumina meets the requirement at the lowest cost. It is also the practical choice for much of the installed 150 mm and 200 mm capacity. This covers power devices, analogue, MEMS and compound semiconductors. There the process is mature, and the priority is reliable, economical clamping rather than leading-edge thermal performance.
When aluminium nitride is worth its higher cost
Aluminium nitride is selected when thermal performance, not cost, sets the requirement. Its thermal conductivity of 150–180 W/m·K removes heat from the wafer far faster than alumina. It holds a flatter temperature profile under high plasma power. This directly protects critical dimensions and film uniformity. Its CTE of about 4.5 ppm/K is closer to silicon. So it generates less differential stress across wide temperature swings. AlN also sits naturally in or near the JR resistivity band. This gives higher clamp force per volt, which suits high-force, thermally demanding etch and PVD. The cost is higher than alumina. So AlN is specified where the process genuinely needs the thermal and force advantages. It is not a default upgrade for every process.
Where doped and composite ceramics fit
Between plain alumina and AlN there is a tuned middle option. Alumina in its pure form is too resistive for JR operation. Adding a dopant such as titania (TiO₂) lowers its volume resistivity out of the Coulombic band. It moves into the JR window of roughly 10¹⁰–10¹³ Ω·cm. That raises clamp force per volt. An ESC can then run at lower voltage for the same hold. The cost is greater temperature sensitivity, because JR resistivity falls as the ceramic heats. Doped and composite ceramics therefore let designers place the resistivity exactly where a process needs it. They combine alumina’s robustness and cost with JR-level force. The detail is in our guide to titania-doped alumina.
Polymer and elastomer dielectrics exist as well. They are used in some low-temperature, large-area or non-semiconductor clamps. They tolerate far less plasma and heat than ceramics. So they are generally unsuited to aggressive wafer-fab plasma. For production etch and deposition, the choice stays within the ceramic family. The ceramic-versus-polymer trade-off is covered in ceramic vs polymer ESC dielectrics.
How resistivity maps to clamping regime
The dielectric choice and the clamping regime are part of the same decision. So it helps to see how the resistivity bands line up. Force per volt, release speed and temperature behaviour all follow from where the ceramic sits on this scale.
| Volume resistivity | Regime | Typical dielectric | Behaviour |
|---|---|---|---|
| > 10¹⁵ Ω·cm | Coulombic | Undoped alumina | Lower force/volt, fast release, low temperature sensitivity |
| ≈ 10¹⁰–10¹³ Ω·cm | Johnsen–Rahbek | AlN, titania-doped alumina | Higher force/volt, slower release, more temperature-sensitive |
The regime decision drives clamp force, dechuck time and temperature stability together. That is why it is treated as the central design choice. The full physics is set out in our comparison of Coulombic vs Johnsen–Rahbek ESCs. The relationship to hold-down is covered in ESC clamping force.
Matching the dielectric to the process
In practice the dielectric is chosen by working back from the process. Consider a thermally demanding, high-power etch that must hold tight critical dimensions. That points toward AlN for its conductivity and CTE match. A yttria finish is often added where the chemistry is strongly halogen-based. Now consider a mature, moderate-power process on a legacy 200 mm tool, where cost and availability dominate. That points toward alumina. It can be doped into the JR band if higher force per volt is needed. These trade-offs are most visible in etch. The full reasoning for that case is worked through in our guide to the best ESC dielectric for plasma etch. Whichever ceramic an ESC uses, the dielectric is a wear surface. Once it erodes or its clamping degrades, the ESC can often be returned to as-new performance through ESC repair and refurbishment rather than replaced outright.
Frequently asked questions
Alumina or aluminium nitride for an ESC?
Alumina is the robust, cost-effective standard choice and usually runs Coulombic. It suits general etch, PVD and CVD and most legacy 150/200 mm tools. Aluminium nitride conducts heat five to six times better (≈ 150–180 vs ≈ 30 W/m·K) and has a CTE closer to silicon. So it is chosen for high-power, thermally demanding processes that need tight temperature uniformity, at a higher cost.
What volume resistivity is needed for a JR ESC?
A Johnsen–Rahbek dielectric sits at roughly 10¹⁰–10¹³ Ω·cm. It is slightly conductive, so charge reaches the surface and clamp force per volt rises. Above about 10¹⁵ Ω·cm the ceramic behaves as a Coulombic insulator instead. Across ESC dielectrics, bulk resistivity spans roughly 10⁹–10¹⁶ Ω·cm.
What is titania-doped alumina used for?
Adding titania (TiO₂) to alumina lowers its resistivity from the Coulombic band into the JR window of about 10¹⁰–10¹³ Ω·cm. This raises clamp force per volt, which is useful where higher hold-down is needed at lower voltage. The cost is greater temperature sensitivity, since JR resistivity falls as the ceramic heats.