Tuning of Atomic Layer Deposition Pulse Time through Physics-Informed Bayesian Active Learning
Pouyan Navabi, Christos G. Takoudis · 2026
Read Tuning of Atomic Layer Deposition Pulse Time through Physics-Informed Bayesian Active Learning by Pouyan Navabi, Christos G. Takoudis (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-18565.txt.
Description
Pouyan Navabi and one co-author tune atomic layer deposition faster. Their 2026 paper notes ALD development is hindered by slow, precursor-heavy tuning cycles to find saturation points. The team introduces a physics-informed Bayesian framework that autonomously sets pulse times by integrating a Langmuir model into the Gaussian Process kernel.
A two-stage strategy splits noise filtering from physical fitting: the Gaussian Process first smooths noisy data, then Langmuir parameters fit smoothed values, separating signal from noise. Against a standard Gaussian Process across four simulated cases, the framework converges in five rounds, improves accuracy fourfold, and cuts precursor usage two to fourfold. Lab tests with titanium dioxide coating confirms saturation times for high targets, with deviations at lower saturation showing desorption.
Catalog record
| CREATOR | Pouyan Navabi, Christos G. Takoudis |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-18565