Die DPG-Frühjahrstagung in Dresden musste abgesagt werden! Lesen Sie mehr ...

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

MM: Fachverband Metall- und Materialphysik

MM 30: Poster Session II

MM 30.40: Poster

Dienstag, 17. März 2020, 18:15–20:00, P4

X-ray reflectivity of thin films evaluated by neural networks — Thorben Finke and •Uwe Klemradt — II. Physik. Inst., RWTH Aachen University, Germany

X-ray reflectivity (XRR) is a widespread method for the structural analysis of thin films on the nanometer scale. We studied the application of a neural network based on Tensorflow for the evaluation of reflectivity curves from metal films of several 10 nm thickness on Si substrates. The focus was on the automatic fitting of layer thickness and surface / internal interface roughness. The network was trained using 900k simulated XRR curves and provided highly accurate results in subsecond computation time, resulting in thickness and roughness errors below 0.1 nm in 95% of the cases when applied to data not known to the network. The results will be discussed in the context of automatic fitting with minimum user interference and high-throughput experiments.