Numerical analyses and loss tangent measurements for the W7-X ECRH gyrotron diamond output windows
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Repository KITopen (Karlsruhe Institute of Technology) |
| Published | 2026-07-31 |
| Authors | G. Aiello, A. Meier, H. P. Laqua, T. Scherer, S. Schreck |
| DOI | 10.5445/ir/1000195809 |
Water cooling was the preferred option for the W7-X ECRH gyrotron diamond output window at 1.5 MW beam power. The numerical analyses predicted lower temperatures and stresses than DC 200(R) silicon oil. The reported thermal case evaluated 1.5 kW absorbed power against an assumed 250°C limit for diamond. The structural case used 1.2 kW absorbed power and an assumed 150 MPa limit. The authors recommend temperature-dependent material properties and plan further characterization for 2 MW beam transmission.
Loss tangent measurements set the absorption input
Section titled “Loss tangent measurements set the absorption input”The authors measured the loss tangent of diamond disks using Fabry-Perot resonators in spherical and hemispherical setups. The measured loss tangent was an input to numerical analyses of power absorption in the disks. The measurement work served the absorption model and the degradation assessment. It supplied the dielectric-loss input for calculating absorbed power. It also provided a basis for checking thermal reserve against possible disk-quality degradation. The stated degradation concern is a loss of microwave transmission capability.
The reported material results identify the setups but give no numerical loss tangent values, specimen count, or measurement uncertainty. They report no separate experiment that causes or tracks degradation. The reported work therefore provides the measurement route and its use in the model, but not a numerical loss-tangent benchmark. This missing data prevents a reader from linking a specific loss tangent to an absorbed-power case or temperature margin.
Operating point and CFD model
Section titled “Operating point and CFD model”The modeled operating point combined a 1.5 MW beam with a frequency of 140 GHz. The diamond disk input set used a relative permittivity of 5.67 and a thickness of 1.8 mm. The millimeter-wave beam was Gaussian with a 20 mm radius. The CFD model used symmetry and temperature-dependent material properties.
| Parameter | Reported value | Applied condition |
|---|---|---|
| Beam power | 1.5 MW | W7-X operating point |
| Frequency | 140 GHz | ECRH beam |
| Diamond thickness | 1.8 mm | Disk model |
| Relative permittivity | 5.67 | Diamond input |
| Beam radius | 20 mm | Gaussian beam |
| Coolant mass flow | 0.2 kg/s | Inlet |
| Coolant inlet temperature | 25°C | Inlet |
| Outlet pressure | 0 Pa | Cooling outlet |
| Silicon oil viscosity | 5 cSt | DC 200(R) |
| Material | CVD diamond | Disk |
The cooling setup supplied 0.2 kg/s at a 25°C inlet and 0 Pa at the outlet. The liquid alternatives were water and DC 200(R) silicon oil with a viscosity of 5 cSt. The model varied absorbed power while keeping the cooling boundary conditions fixed. This separation matters because beam power and power absorbed in the dielectric are different quantities. The 1.5 MW value is the beam power, while later sections use absorbed power in kilowatts. Temperature-dependent properties were applied to the diamond and the thermal model.
Thermal comparison at 1.5 kW absorption
Section titled “Thermal comparison at 1.5 kW absorption”The CFD comparison at 1.5 kW absorbed power separated the water and oil cooling cases. The authors assumed 250°C as the diamond temperature limit. That value is a design assumption, not a measured failure threshold. They selected water cooling for high-power operation and stated that oil cooling was not an option at this absorbed power. The result supports a clear coolant choice, although the reported results do not list either peak temperature.
The water case was presented as providing substantially more temperature reserve. The authors also require temperature-dependent properties, particularly for diamond, to show window performance at these power levels. The reported comparison does not give peak temperatures, coolant outlet temperatures, or numerical reserve margins. It therefore identifies the preferred coolant but does not quantify how much thermal margin remains at the design point. A fixed-property calculation would not meet the authors’ stated modeling requirement.
Loss tangent sensitivity and thermal reserve
Section titled “Loss tangent sensitivity and thermal reserve”The parametric study examined loss tangent through cases with different absorbed-power values. Its purpose was to test whether water cooling retains thermal reserve as dielectric loss and absorbed power increase. Water cooling provided significant temperature reserve margins even at higher absorbed power. The authors connected those margins to potential disk-quality factors that could reduce microwave transmission capability.
The reserve was intended to absorb uncertainty in dielectric loss and window quality. The presentation gives neither the loss tangent assigned to each case nor the numerical temperature margin. It also does not identify a loss-tangent threshold at which cooling becomes inadequate. The sensitivity result therefore remains qualitative. Within the reported comparison, water cooling still provided the stronger thermal margin. The design implication is to retain water cooling when evaluating uncertain dielectric loss, rather than treating the nominal loss tangent as fixed.
Structural response at the design absorption
Section titled “Structural response at the design absorption”The structural analysis used a plastic steady-state model. Its main water cooling case used 1.2 kW absorbed power, with 150 MPa assumed as the diamond stress limit. The water option produced lower stresses than the oil option. The authors report no significant changes when absorbed power increased to 1.5 kW.
The 150 MPa value should not be read as a measured failure stress because the source labels it an assumed limit. The report gives no peak stress, stress components, deflection, safety factor, or contact pressure. It also does not quantify what “no significant changes” means in stress or deformation. These omissions limit independent reproduction of the structural result.
The thermal and structural comparisons also require careful reading. The main thermal case used 1.5 kW, while the structural baseline used 1.2 kW. The 1.5 kW structural update is qualitative rather than a reported stress value. The supported comparison is therefore lower stress with water, checked against an assumed 150 MPa limit.
Power context and engineering decisions
Section titled “Power context and engineering decisions”The power context places W7-X between the 1 MW ITER beam and the 2 MW EU DEMO beam. The study therefore addresses an intermediate high-power case within the EC H&CD examples. The authors plan later characterization of the window for 2 MW beam transmission. No 2 MW performance result appears in the reported work.
The loss tangent measurements matter because absorbed power can rise as dielectric loss changes under the same beam power. The reported margin argument is intended to protect microwave transmission against potential window-quality degradation. It is not a demonstrated service-life result. For device engineering, the supported choice is water cooling at the reported high-power condition. The analysis does not support the oil option for the 1.5 kW absorption case.
A final device design still needs numerical loss tangent data, peak temperatures, peak stresses, and uncertainty information. The reported results do not provide enough numerical output to reproduce the exact reserve margins. Using temperature-dependent properties remains a stated requirement for the final assessment. The planned 2 MW characterization should also precede any performance claim at that higher beam power.
Materials and support
Section titled “Materials and support”The W7-X window calls for a CVD diamond disk 1.8 mm thick, a 140 GHz, 1.5 MW beam, and water cooling at 0.2 kg/s with a 25°C inlet. The 6CCVD diamond materials catalog lists single-crystal and polycrystalline diamond grown by MPCVD, with plates and wafers up to 125 mm. The disk thickness and cooling interface should be reviewed through the engineering support service for custom fabrication.
View Original Abstract
The 140 GHz 1 MW gyrotron for the electron cyclotron resonance heating (ECRH) system at the stellarator Wendelstein 7-X (W7-X) is being upgraded to 1.5 MW continuous wave operation to increase the total heating power for achieving operating regimes with high plasma beta and low collisionality. The gyrotron features a chemical vapor deposition (CVD) polycrystalline diamond window with an aperture of 88 mm and a disk of 1.8 mm thickness and 106 mm diameter. In this work, numerical analyses of the window are shown with loss tangent values provided by experimental measurements on 25 bare diamond disks. Computational fluid dynamics (CFD) conjugated heat transfer and structural analyses were carried out to check the window performance at 1.5 MW operation, when cooled by water and silicon oil Dow Corning 200(R) 5cSt.
Single Crystal Cvd Diamond Plate
View Specs →