Characterization of thermomechanical damage on tungsten surfaces during long-duration plasma transients

Abstract

A new experimental facility constructed at UCLA for the simulation of high heat flux effects on plasma-facing materials is described. The High Energy Flux Test Facility (HEFTY) is equipped with a Praxair model SG-100 plasma gun, which is nominally rated at 80kW of continuous operation, of which approximately 30kW reaches the target due to thermal losses. The gun is used to impart high intermittent heat flux to metal samples mounted within a cylindrical chamber. The system is capable of delivering an instantaneous heat flux in the range of 30–300MW/m2, depending on sample proximity to the gun. The duration of the plasma heat flux is in the range of 1–1000s, making it ideal for studies of mild plasma transients of relatively long duration. Tungsten and tungsten-copper alloy metal samples are tested in these transient heat flux conditions, and the surface is characterized for damage evaluation using optical, SEM, XRD, and micro-fabrication techniques. Results from a Finite Element (FE) thermo-elastoplasticity model indicate that during the heat-up phase of a plasma transient pulse, the majority of the sample surface is under compressive stresses leading to plastic deformation of the surface. Upon sample cooling, the recovered elastic strain of cooler parts of the sample exceeds that from parts that deformed plastically, resulting in a tensile surface self-stress (residual surface stress). The intensity of the residual tensile surface stress is experimentally correlated with the onset of complex surface fracture morphology on the tungsten surface, and extending below the surface region. Micro-compression mechanical tests of W micro-pillars show that the material has significant plasticity, failing by a “barreling” mode before plasma exposure, and by normal dislocation slip and localized shear after plasma exposure. Ongoing modeling of the complex thermo-fracture process, coupled with elasto-plasticity is based on a phase field approach for distributed fracture, and a discrete cracking approach, with cracks represented by Volterra dislocations.