Titania-doped alumina is an ESC dielectric. A controlled fraction of titania (TiO₂) is added to the alumina (Al₂O₃) matrix. The addition lowers the volume resistivity from the Coulombic band into the Johnsen–Rahbek band. That single change raises clamping force per volt. This is why doped alumina exists as a distinct ESC material rather than a curiosity. This article explains how titania doping shifts resistivity, and the temperature trade-off that comes with it. It explains how the doping level tunes the target band. It also explains why a fab would specify doped alumina over plain alumina or aluminium nitride.
Doped alumina is part of the wider dielectric-selection question covered in our pillar on ESC dielectric materials, alumina vs aluminium nitride.
Why resistivity sets the clamping regime
The volume resistivity of the ceramic between the buried electrode and the wafer decides the clamping mechanism. A near-perfect insulator above 10¹⁵ Ω·cm keeps all charge on the buried electrode. The wafer then clamps in the Coulombic regime across the full dielectric thickness. A slightly conductive ceramic in the band of roughly 10¹⁰–10¹³ Ω·cm lets charge migrate to the ESC surface. The effective gap collapses to the microscopic contact points, and the wafer clamps in the Johnsen–Rahbek regime. Electrostatic force depends on the inverse square of that gap. So the same applied voltage produces far higher force in the JR regime than in the Coulombic one.
Plain ESC-grade alumina is naturally a Coulombic material. At a typical purity of 99.5–99.9%, its bulk resistivity sits well above 10¹⁵ Ω·cm. So an undoped alumina ESC clamps in the Coulombic regime and needs higher voltage to reach a given force. Doping is the deliberate, repeatable way to move that resistivity down into the JR window. It does not change the base ceramic system, the bond stack or the ESC geometry. The link between resistivity band and clamping mechanism is set out in our companion piece on ESC volume resistivity and the clamping regime.
How titania lowers alumina’s resistivity
Titania is a semiconducting oxide. A measured amount is sintered into the alumina matrix. It adds conduction paths and donor-type behaviour to what is otherwise a highly insulating ceramic. The net effect is to lower the bulk resistivity of the composite by several orders of magnitude. The resistivity moves from the Coulombic band above 10¹⁵ Ω·cm down into the JR band of roughly 10¹⁰–10¹³ Ω·cm. Across the full family of ESC dielectrics, bulk resistivity spans roughly 10⁹–10¹⁶ Ω·cm. Titania doping is one of the practical ways to place a given ESC deliberately within that span.
The point of moving into the JR band is force per volt. A JR ESC clamps the wafer at its microscopic contact points rather than across the full dielectric thickness. So it reaches the production clamp-force target, typically 20–80 Torr-equivalent, at a lower chuck voltage than a Coulombic ESC of the same geometry. That matters because clamping force must always exceed the 5–40 Torr helium backside pressure used for cooling. If it does not, the gas lifts the wafer and temperature control collapses. Reaching the required margin at lower voltage eases the demand on the power supply. It also eases the dielectric breakdown budget, since the ceramic still has to withstand its rated dielectric strength of 10–20 kV/mm.
Before and after doping: the resistivity shift
The table below compares undoped ESC-grade alumina with titania-doped alumina engineered for the JR window. The base ceramic properties are essentially unchanged: thermal conductivity, CTE and dielectric strength stay the same. Doping moves resistivity, and with it the clamping regime and its consequences.
| Property | Plain alumina (undoped) | Titania-doped alumina |
|---|---|---|
| Volume resistivity | > 10¹⁵ Ω·cm | ≈ 10¹⁰–10¹³ Ω·cm |
| Clamping regime | Coulombic | Johnsen–Rahbek |
| Clamp force per volt | Lower | Higher |
| Voltage to reach force target | Higher | Lower |
| Release / dechuck speed | Fast | Slower (residual charge) |
| Temperature sensitivity | Low | Higher (resistivity falls as temperature rises) |
| Thermal conductivity | ≈ 30 W/m·K | ≈ 30 W/m·K (unchanged) |
| CTE | ≈ 7–8 ppm/K | ≈ 7–8 ppm/K (unchanged) |
| Dielectric strength | 10–20 kV/mm | 10–20 kV/mm |
The single intended change is resistivity. Everything in the right-hand column follows from it. The higher force per volt is the benefit. The slower release and greater temperature sensitivity are the costs of operating in the JR regime. For the underlying physics of the two mechanisms, see our guide to Coulombic vs Johnsen–Rahbek ESCs.
The temperature trade-off
The cost of moving into the JR band is that resistivity becomes temperature-dependent. The conduction that doping introduces is thermally activated. So the bulk resistivity of titania-doped alumina falls as the ceramic heats. Within a plasma process the ESC does not sit at a single uniform temperature. It warms during a run, varies across its radius, and changes between recipes. As resistivity drops with rising temperature, the ESC moves further into the JR band. Charge migrates more readily, clamp force rises and release slows. A Coulombic alumina ESC is comparatively stable in this respect. At above 10¹⁵ Ω·cm there is almost no conduction to begin with, so small temperature changes barely move the force.
This has a practical meaning for a process engineer. A doped-alumina ESC must be characterised across its real operating temperature window, not at a single set point. The clamp-force margin over helium backside pressure has to hold at the coldest point of the cycle. That is where resistivity is highest and force is lowest. The dechuck sequence has to clear residual charge reliably at the hottest point. That is where force is highest and release slowest. Multi-zone ESCs add a further complication. Radial temperature gradients translate directly into radial resistivity and clamp-force gradients across a single doped dielectric.
Tuning the doping level to the target band
The doping level is the design choice that places the resistivity where the process needs it. A light titania addition lowers resistivity only modestly. The ceramic then sits near the top of the JR band, at around 10¹³ Ω·cm. That is closer to Coulombic behaviour, with a more moderate force gain and a smaller temperature coefficient. A heavier addition pushes resistivity toward the bottom of the band, near 10¹⁰ Ω·cm. That gives the highest force per volt, but the strongest temperature dependence and the slowest release. The target is a balance. It weighs the force the process needs against the temperature stability and dechuck behaviour the process can tolerate.
Hitting that target repeatably is a manufacturing discipline, not a single recipe. The final bulk resistivity of a doped alumina depends on the dopant concentration. It also depends on how evenly the dopant is distributed through the matrix, and on the sintering conditions. All of these have to be controlled to keep every ESC inside the intended window. A dielectric that drifts out of band defeats the purpose. Too insulating, and it loses the force advantage that justified doping. Too conductive, and release and temperature stability suffer. This sensitivity is one reason doped-alumina ESCs are specified against a resistivity range rather than a single nominal figure. It is also why consistent material processing matters across a fleet of tools that must behave identically.
Doped alumina vs plain alumina vs AlN
Doped alumina occupies a specific position between the two mainstream ESC ceramics. Plain alumina is the cost-effective Coulombic standard. It is robust and well understood, with a thermal conductivity of around 30 W/m·K. It is the right choice where Coulombic behaviour suits the process: fast release and low temperature sensitivity. The required force must also be reachable at the available voltage. Sometimes the process needs more clamp force than plain alumina delivers at that voltage. Doping the same base ceramic with titania then moves it into the JR regime and raises force per volt. The ESC keeps alumina’s familiar thermal and mechanical properties, and its lower cost relative to AlN.
Aluminium nitride is the alternative route to JR operation, but it solves a different problem. Its thermal conductivity of roughly 150–180 W/m·K is around five to six times that of alumina. Its CTE of about 4.5 ppm/K is a closer match to silicon’s ~2.6–3.3 ppm/K. So it gives a flatter wafer temperature and less differential stress in thermally demanding processes. That thermal performance carries a higher material and fabrication cost. The practical distinction is this. A fab chooses titania-doped alumina when it wants the higher force per volt of the JR regime. In that case the thermal demands do not justify AlN, and the fab wants to stay on the cheaper, well-characterised alumina platform. Where wafer temperature uniformity is the binding constraint, aluminium nitride is the better-matched choice despite the cost.
The summary below frames the decision for an engineer specifying an ESC for a given process.
| Material | Regime | Thermal conductivity | Choose when |
|---|---|---|---|
| Plain alumina | Coulombic | ≈ 30 W/m·K | Force reachable at available voltage; fast release and low temperature sensitivity wanted; cost-driven |
| Titania-doped alumina | Johnsen–Rahbek | ≈ 30 W/m·K | Higher force per volt needed; thermal demands modest; staying on the alumina platform |
| Aluminium nitride | Coulombic or JR | ≈ 150–180 W/m·K | Wafer temperature uniformity and CTE match are the binding constraints; cost secondary |
Practical notes for legacy and 300 mm tools
Doped-alumina ESCs appear on both legacy 150/200 mm tools and 300 mm tools. They are used wherever a process needs JR-level force on an alumina platform. Older tools serve power devices, analogue, MEMS and compound semiconductors. The original equipment makers no longer prioritise much of that installed base. On those tools, the original dielectric specification may be a doped alumina that is no longer documented or readily available new. Reproducing that ESC means reproducing its resistivity band, not just its dimensions. A replacement that clamps in the wrong regime will not behave like the original on the tool.
This is where second-source and refurbishment capability matters. A worn doped-alumina ESC whose body is sound can often be reconditioned. Reconditioning restores its clamping and particle performance. An OEM-equivalent replacement can reproduce the original resistivity target rather than substituting a different dielectric. Where the ESC body is beyond recondition, the replacement should match the doped-alumina specification. That keeps the tool behaving as the process was qualified to expect. The reconditioning and replacement routes are covered in our overview of ESC repair and refurbishment.
Frequently asked questions
What is titania-doped alumina used for?
It is used as an ESC dielectric where a process needs higher clamping force per volt than plain alumina delivers. Adding titania lowers alumina’s volume resistivity from the Coulombic band into the Johnsen–Rahbek band of roughly 10¹⁰–10¹³ Ω·cm. This raises force per volt while keeping the ESC on the lower-cost alumina platform.
What volume resistivity is needed for a JR ESC?
A Johnsen–Rahbek ESC needs a slightly conductive dielectric in the band of roughly 10¹⁰–10¹³ Ω·cm. A Coulombic insulator is above 10¹⁵ Ω·cm. Titania doping is one way to place alumina deliberately within that JR window. The doping level sets where in the band the resistivity lands.
Why choose doped alumina over AlN?
Both can give Johnsen–Rahbek operation, but they solve different problems. Titania-doped alumina raises force per volt on the cheaper, well-characterised alumina platform when thermal demands are modest. AlN has about five to six times the thermal conductivity and a closer CTE match to silicon. It is chosen when wafer temperature uniformity is the binding constraint and justifies its higher cost.