EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Deep Learning Approach for Thickness Determination of AZO Thin Films

Víctor Manuel Guerrero-Sosa, Angélica Carrillo-Verduzco iD, Juan Carlos Alonso-Huitrón iD, Alan Gabriel Amaro-Colin, Miguel Esquivel-García, Xim Bokhimi iD

ACS Omega · 10.11.2026 · 11 (44) · 69373-69380

DOI: 10.1021/acsomega.6c10558 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ACS Omega, online erschienen am 27.10.2026. Verfasst von Víctor Manuel Guerrero-Sosa und 5 weiteren Autor:innen. Beteiligte Einrichtung: Instituto de Investigaciones en Materiales, Universidad Nacional Autónoma de México, Ciudad de México C.P. 04510, Mexico. Die Arbeit nennt 60 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Universidad Nacional Aut?noma de M?xico, Laboratorio de Refinamiento de Estructuras Cristalinas, Instituto de F?sica, Universidad Nacional Aut?noma de M?xico.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We developed a database and a machine learning model to predict the thickness of aluminum-doped zinc oxide (AZO) thin films from their optical transmittance spectra. We measured the spectra of 145 thin films with thicknesses between 100 and 1500 nm and fitted each with a 12-parameter function, one parameter being the film thickness. We divided the thickness range into five regions and modeled the distribution of each fitting parameter in each region as a sum of Gaussian functions. Random sampling from these distributions produced parameter sets used to construct synthetic transmittance spectra (X), each paired with its film thickness (Y) to form one database sample (X, Y). We generated 240,000 samples per region, yielding a database of 1.2 million samples for supervised training with thickness as the target. The initial machine learning model, a fully connected neural network, showed poor predictive accuracy. We therefore improved the architecture by incorporating an autoencoder with skip connections to extract the most informative features from the transmittance spectra. This change substantially improved performance, yielding mean absolute errors of 15.8(3) nm on the training set and 15.7(3) nm on the validation set. When applied to the transmittance spectra of the experimental AZO films, the trained model predicted thicknesses with a mean absolute percentage error of 6.91%. Because the model was trained using data derived from AZO measurements, it applies only to AZO thin films. The methodology, however, is general and can be applied to thin films of any composition.

Autor:innen und Institutionen

  • Instituto de Investigaciones en Materiales, Universidad Nacional Autónoma de México, Ciudad de México C.P. 04510, Mexico.
  • Instituto de Física, Universidad Nacional Autónoma de México, Ciudad de México C.P. 04510, Mexico.

ORCID-Profile: Angélica Carrillo-Verduzco, Juan Carlos Alonso-Huitrón, Xim Bokhimi

Förderung

  • Universidad Nacional Aut?noma de M?xico
  • Laboratorio de Refinamiento de Estructuras Cristalinas, Instituto de F?sica, Universidad Nacional Aut?noma de M?xico

Zitieren

APA

Guerrero-Sosa, V. M., Carrillo-Verduzco, A., Alonso-Huitrón, J. C., Amaro-Colin, A. G., Esquivel-García, M., & Bokhimi, X. (2026). A Deep Learning Approach for Thickness Determination of AZO Thin Films. ACS Omega, 11(44), 69373-69380. https://doi.org/10.1021/acsomega.6c10558

Vancouver

Guerrero-Sosa VM, Carrillo-Verduzco A, Alonso-Huitrón JC, Amaro-Colin AG, Esquivel-García M, Bokhimi X. A Deep Learning Approach for Thickness Determination of AZO Thin Films. ACS Omega. 2026;11(44):69373-69380. doi:10.1021/acsomega.6c10558

BibTeX
@article{GuerreroSosa2026,
  author = {Guerrero-Sosa, Víctor Manuel and Carrillo-Verduzco, Angélica and Alonso-Huitrón, Juan Carlos and Amaro-Colin, Alan Gabriel and Esquivel-García, Miguel and Bokhimi, Xim},
  title = {A Deep Learning Approach for Thickness Determination of AZO Thin Films},
  journal = {ACS Omega},
  year = {2026},
  volume = {11},
  number = {44},
  pages = {69373--69380},
  doi = {10.1021/acsomega.6c10558}
}
Kostenlos starten →