Denoising strategies for three-dimensional visualization of multilayer thin-film interfaces via angle-resolved hard x-ray photoelectron spectroscopy
Source abstract
Abstract
Nondestructive three-dimensional visualization of buried interfaces in multilayer thin films is central to semiconductor device characterization. Angle-resolved hard x-ray photoelectron spectroscopy (AR-HAXPES) enables such depth-resolved analysis, but short-exposure measurements suffer from photon shot noise that degrades the reconstructed depth profiles—a problem that becomes acute as the technique is pushed toward spatiotemporal “4D-XPS” (spatially and temporally resolved XPS) at focused-beam instruments. Here, we benchmark two denoising strategies—statistical bin-pool (BP) resampling and self-supervised deep neural network (DNN) denoising—applied to AR-HAXPES data from a C/Al2O3/TiO2/Si multilayer film acquired with a laboratory Ga Kα source (9.25 keV) under two per-frame noise levels (24 s/frame × 2500 frames; 120 s/frame × 400 frames). Depth profiles are reconstructed per pixel via L1-regularized inversion and scored against a pseudoground truth using the depth-profile root mean square error (Depth RMSE), film-thickness fidelity Δthk, and the first-order Wasserstein distance W1. The two strategies turn out to be complementary rather than competing: under high-noise conditions, the self-supervised DNN dominates with up to a 27-fold equivalent-exposure gain, whereas under low-noise conditions, BP matches or slightly surpasses the DNN at roughly a six- to eightfold gain and retains a clear advantage in reconstructed layer-thickness fidelity. BP—training-free and computationally trivial—is, therefore, the preferred choice whenever the per-frame statistics are sufficient for the inversion to resolve interfaces, while the self-supervised DNN remains justified specifically when photon starvation is severe.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Electron and X-Ray Spectroscopy Techniques · Advanced Electron Microscopy Techniques and Applications · Machine Learning in Materials Science
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Advanced Equipment Division, Vacuum Products Corporation 1 , 16-2 Miyukicho, Kodaira-shi, Tokyo 187-0012,
School of Science and Technology, Meiji University 2 , 1-1-1 Higashimita, Tama-ku, Kawasaki, Kanagawa 214-8571,
Meiji Renewable Energy Laboratory, Meiji University 3 , 1-1-1 Higashimita, Tama-ku, Kawasaki, Kanagawa 214-8571,
References (27)
Show source-provided references
1. Sci. Rep. · 2024 · 14 · 1849710.1038/s41598-024-69495-0 ↗
2. Appl. Surf. Sci. · 2019 · 483 · 14010.1016/j.apsusc.2019.03.211 ↗
3. Surf. Interface Anal. · 1989 · 14 · 44310.1002/sia.740140804 ↗
4. Surf. Interface Anal. · 2006 · 38 · 59010.1002/sia.2216 ↗
5. J. Surf. Anal. · 2009 · 15 · 29110.1384/jsa.15.291 ↗
6. Surf. Interface Anal. · 2011 · 43 · 158110.1002/sia.3753 ↗
7. Surf. Interface Anal. · 2023 · 55 · 65810.1002/sia.7240 ↗
8. Vac. Surf. Sci. · 2021 · 64 · 8610.1380/vss.64.86 ↗
9. Faraday Discuss. · 2022 · 236 · 31110.1039/D2FD00021K ↗
10. Scienta Omicron
11. Vac. Surf. Sci. · 2024 · 67 · 11210.1380/vss.67.112 ↗
12. Surf. Interface Anal. · 1997 · 25 · 37410.1002/(SICI)1096-9918(199705)25:5<374::AID-SIA249>3.0.CO;2-L ↗
13. Surf. Rev. Lett. · 2002 · 09 · 21310.1142/S0218625X0200180X ↗
14. Rev. Sci. Instrum. · 2011 · 82 · 11370110.1063/1.3657156 ↗
15. J. Synchrotron Radiat. · 2023 · 30 · 101310.1107/S160057752300629X ↗
16. Proc. Mach. Learn. Res. · 2018 · 80 · 2965
17. 2019 · 2129
18. Proc. Mach. Learn. Res. · 2019 · 97 · 524
19. IEEE Trans. Comput. Imaging · 2022 · 8 · 58510.1109/TCI.2022.3176536 ↗
20. Sci. Adv. · 2025 · 11 · eads555210.1126/sciadv.ads5552 ↗
21. Vac. Surf. Sci. · 2025 · 68 · 6610.1380/vss.68.66 ↗
22. Cross-exposure transferability and failure boundaries of self-supervised DNN denoising for AR-HAXPES depth profiling10.26434/chemrxiv.15005615/v2 ↗
23. Appl. Phys. Lett. · 2025 · 126 · 07210310.1063/5.0245254 ↗
24. J. Electron Spectrosc. Relat. Phenom. · 1995 · 73 · 2510.1016/0368-2048(94)02270-4 ↗
25. At. Data Nucl. Data Tables · 2018 · 119 · 9910.1016/j.adt.2017.04.003 ↗
26. Surf. Interface Anal. · 1991 · 17 · 91110.1002/sia.740171304 ↗
27. At. Data Nucl. Data Tables · 1985 · 32 · 110.1016/0092-640X(85)90016-6 ↗