University of Helsinki researchers have found through exceptionally large computer simulations that tungsten, a leading material for the walls facing plasma in future fusion reactors, may sustain more primary radiation damage at reactor-relevant energies than established models predict. The finding, reported by Phys.org on August 15 and accepted by Physical Review Letters, could affect how engineers estimate the durability of components exposed to fusion neutrons.
Fusion reactions release high-energy neutrons that can strike a tungsten atom and knock it from its lattice position. That recoiling atom can hit others, producing a collision cascade and leaving defects in the metal. Quantifying these first moments of damage is important because tungsten must withstand the extreme environment around the superheated, electrically charged plasma while retaining its structural performance.
The team led by first author Jesper Byggmästar used machine-learning-driven molecular dynamics adapted to run efficiently on graphics processors. Its calculations followed recoil energies as high as 2 megaelectronvolts and systems containing as many as one billion atoms. The paper says earlier full atomistic studies generally stopped at a few hundred kiloelectronvolts, leaving a gap between simulations and the much higher energies routinely reached in irradiation experiments.
Across that expanded range, the researchers identified four regimes in the production of primary damage. Existing descriptions generally expect defect numbers to rise sublinearly with recoil energy before becoming linear. The new calculations instead reveal an additional high-energy behavior that departs from previous models, with the transition beginning near 300 kiloelectronvolts, the maximum recoil energy that fusion-emitted neutrons can transfer to tungsten atoms.
The simulations therefore indicate that tungsten components approaching fusion-relevant recoil energies may deteriorate more than earlier projections assumed. The result does not mean tungsten is unsuitable, nor does it measure the lifetime of an operating reactor. It supplies a revised model for the initial defects created by one atomic recoil, which researchers can feed into longer-scale calculations used to forecast accumulated damage and component service life.
The Helsinki group plans to extend the approach to higher radiation doses, greater length scales and more accurate simulations. It also intends to examine how grain size, grain boundaries and alloying elements change the outcome. Experimental comparison and multiscale work will be needed before engineers can translate the result into specific reactor designs, but the billion-atom calculation gives fusion-material researchers a new basis for testing more durable walls. Those next steps will show how closely the simulated first-impact damage corresponds to the behavior of engineered tungsten over sustained exposure.
Comments