Fast fitting of reflectivity data of growing thin films using neural networks

Greco, Alessandro; Starostin, Vladimir; Karapanagiotis, Christos; Hinderhofer, Alexander; Gerlach, Alexander; Pithan, Linus; Liehr, Sascha; Schreiber, Frank; Kowarik, Stefan

doi:10.1107/s1600576719013311

Cited by 38 publications

(44 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Advanced modeling and computer simulations are indispensable for deriving quantitative structural and dynamical information from scattering experiments. In this respect, the machine learning techniques are likely to be more exploited in the analysis of scattering data in the future [243,244]. The ultimate goal is to derive the real space images of functional systems with nanometer scale resolution from the scattering data by exploiting the coherence and inherent features within the scattering patterns.…”

Section: Discussionmentioning

confidence: 99%

Synchrotron Scattering Methods for Nanomaterials and Soft Matter Research

Narayanan

Konovalov

2020

Materials

View full text Add to dashboard Cite

This article aims to provide an overview of broad range of applications of synchrotron scattering methods in the investigation of nanoscale materials. These scattering techniques allow the elucidation of the structure and dynamics of nanomaterials from sub-nm to micron size scales and down to sub-millisecond time ranges both in bulk and at interfaces. A major advantage of scattering methods is that they provide the ensemble averaged information under in situ and operando conditions. As a result, they are complementary to various imaging techniques which reveal more local information. Scattering methods are particularly suitable for probing buried structures that are difficult to image. Although, many qualitative features can be directly extracted from scattering data, derivation of detailed structural and dynamical information requires quantitative modeling. The fourth-generation synchrotron sources open new possibilities for investigating these complex systems by exploiting the enhanced brightness and coherence properties of X-rays.

show abstract

Section: Discussionmentioning

confidence: 99%

Synchrotron Scattering Methods for Nanomaterials and Soft Matter Research

Narayanan

Konovalov

2020

Materials

View full text Add to dashboard Cite

show abstract

“…2 a, the conventional method of training for an ANN algorithm is shown, matching those in previous works 15 – 19 . A multilayer perceptron (MLP) type ANN algorithm was constructed and trained using Python, similarly to the previous work 15 , 17 , 19 . In the wavelength range of the spectrometer to be used for the CRM measurement, a wavelength range in which the intensity of the measured light is sufficiently greater than noise was selected, and the number of samples for that range was established as the number of input nodes.…”

Section: Methodsmentioning

confidence: 99%

“…Because the proposed method can basically offer strict validation of the ANN algorithms using certified values of the CRMs, it is not necessary to use validation data separately from ideally created reflectance spectra used for training datasets. The design parameters of the ANN algorithm were number of hidden layers (L = 1, 2, 3) and number of nodes in each layer (N = 50, 100, 150, 200), which were selected as simple cases based on the previous works 15 , 17 , 19 . Therefore, with the combination of these two parameters, 12 ANN algorithms were developed and then trained.…”

Section: Methodsmentioning

confidence: 99%

A novel method to design and evaluate artificial neural network for thin film thickness measurement traceable to the length standard

Lee

Jin

2022

Sci Rep

View full text Add to dashboard Cite

The artificial neural networks (ANNs) have been often used for thin-film thickness measurement, whose performance evaluations were only conducted at the level of simple comparisons with the existing analysis methods. However, it is not an easy and simple way to verify the reliability of an ANN based on international length standards. In this article, we propose for the first time a method by which to design and evaluate an ANN for determining the thickness of the thin film with international standards. The original achievements of this work are to choose parameters of the ANN reasonably and to evaluate the training instead of a simple comparison with conventional methods. To do this, ANNs were built in 12 different cases, and then trained using theoretical spectra. The experimental spectra of the certified reference materials (CRMs) used here served as the validation data of each trained ANN, with the output then compared with a certified value. When both values agree with each other within an expanded uncertainty of the CRMs, the ANN is considered to be reliable. We expect that the proposed method can be useful for evaluating the reliability of ANN in the future.

show abstract

“…Previous studies showed that deep learning methods with convolutional neural network (CNN) models are highly effective for decreasing the noise from digital signals and images 18 – 23 ; therefore, a similar approach would also work for the NR experiments. A few applications of the machine learning in NR have been intended to directly determine the structure of the surface and interface of a specimen 24 – 26 . The current study introduces a deep learning approach that uses neural networks to restore the reflection profile hidden in the large statistical noise to overcome the limitation in the conventional NR methods.…”

Section: Introductionmentioning

confidence: 99%

Deep learning approach for an interface structure analysis with a large statistical noise in neutron reflectometry

Aoki

Liu

Yamashita

2021

Sci Rep

View full text Add to dashboard Cite

Neutron reflectometry (NR) allows us to probe into the structure of the surfaces and interfaces of various materials such as soft matters and magnetic thin films with a contrast mechanism dependent on isotopic and magnetic states. The neutron beam flux is relatively low compared to that of other sources such as synchrotron radiation; therefore, there has been a strong limitation in the time-resolved measurement and further advanced experiments such as surface imaging. This study aims at the development of a methodology to enable the structural analysis by the NR data with a large statistical error acquired in a short measurement time. The neural network-based method predicts the true NR profile from the data with a 20-fold lower signal compared to that obtained under the conventional measurement condition. This indicates that the acquisition time in the NR measurement can be reduced by more than one order of magnitude. The current method will help achieve remarkable improvement in temporally and spatially resolved NR methods to gain further insight into the surface and interfaces of materials.

show abstract

Fast fitting of reflectivity data of growing thin films using neural networks

Abstract: Artificial neural networks trained with simulated data are shown to correctly and quickly determine film parameters from experimental X-ray reflectivity curves.

Cited by 38 publications

References 28 publications

Synchrotron Scattering Methods for Nanomaterials and Soft Matter Research

Synchrotron Scattering Methods for Nanomaterials and Soft Matter Research

A novel method to design and evaluate artificial neural network for thin film thickness measurement traceable to the length standard

Deep learning approach for an interface structure analysis with a large statistical noise in neutron reflectometry

Contact Info

Product

Resources

About