Identifying functional groups in IR spectra using an artificial neural network

Fessenden, Ralph J.; Györgyi, László

doi:10.1039/p29910001755

Cited by 37 publications

(30 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The information contained in a spectrum is most often presented to chemist as a 2D image, therefore it is desirable to develop models that learn via similar spectral visualization. 22 Previous implementations of FTIR ML for functional group identification have limited, 23 averaged, 24 and segmented 23,25 spectral data to reduce information used during training. The computational resources available today make this an unnecessary and limiting feature.…”

Section: Introductionmentioning

confidence: 99%

Functional Group Identification for FTIR Spectra Using Image-Based Machine Learning Models

et al. 2021

View full text Add to dashboard Cite

Fourier Transform Infrared Spectroscopy (FTIR) is a ubiquitous spectroscopic technique. Spectral interpretation is a time-consuming process, but it yields important information about functional groups present in compounds and in complex substances. We develop a generalizable model via a machine learning (ML) algorithm using Convolutional Neural Networks (CNNs) to identify the presence of functional groups in gas phase FTIR spectra. The ML models will reduce the amount of time required to analyze functional groups and facilitate interpretation of FTIR spectra. Through web scraping, we acquire intensity-frequency data from 8728 gas phase organic molecules within the NIST spectral database and transform the data into images. We successfully train models for 15 of the most common organic functional groups, which we then determine via identification from previously untrained spectra. These models serve to expand the application of FTIR measurements for facile analysis of organic samples. Our approach was done such that we have broad functional group models that inference in tandem to provide full interpretation of a spectrum. We present the first implementation of ML using image-based CNNs for predicting functional groups from a spectroscopic method. File list (2) download file view on ChemRxiv ML functional group ID in FTIR spectra.docx (545.84 KiB) download file view on ChemRxiv ML functional group ID in FTIR spectra.pdf (645.08 KiB) Functional group identification for FTIR spectra using image-based machine learning models

show abstract

Section: Introductionmentioning

confidence: 99%

Functional Group Identification for FTIR Spectra Using Image-Based Machine Learning Models

et al. 2021

View full text Add to dashboard Cite

show abstract

“…42 Here, we also used an encoder to create a corresponding latent space based on spectra to predict functional groups which may also be useful to design molecules for specic spectral properties. A few ML techniques to analyse spectra have been used previously [48][49][50][51][52] but such attempts for functional group prediction used only one type of spectral data; the training data were specic to the application and classied groups separately as a multiple binary classication problem. 51,52 Binary classiers are not optimal for a large number of classes and are sensitive to class imbalances during training resulting in problems in identifying all functional groups in a molecule or mixtures.…”

Section: Introductionmentioning

confidence: 99%

Spectral deep learning for prediction and prospective validation of functional groups

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Two additional methods for dimension reduction of FTIR spectra, published by Fessenden [21] and by Van Est [22], were also implemented in this study. Fessenden performed data reduction by reducing the spectrum regarding equidistant point selection, while the approach of Van Est calculated the mean of several data points.…”

Section: Discussionmentioning

confidence: 99%

Prediction of Thermal Exposure and Mechanical Behavior of Epoxy Resin Using Artificial Neural Networks and Fourier Transform Infrared Spectroscopy

2019

View full text Add to dashboard Cite

Thermal degradation detection of cured epoxy resins and composites is currently limited to severe thermal damage in practice. Evaluating the change in mechanical properties after a short-time thermal exposure, as well as estimating the history of thermally degraded polymers, has remained a challenge until now. An approach to accurately predict the mechanical properties, as well as the thermal exposure time and temperature of epoxy resin, using Fourier-transform infrared spectroscopy (FTIR)-spectroscopy, data processing, and artificial neural networks, is presented here. Therefore, an epoxy resin has been fully cured and exposed to elevated temperatures for different time periods. A FTIR-spectrometer was used to measure molecular changes, using mid-IR (MIR)-FTIR for film samples and near-IR (NIR)-FTIR for bulk samples. A quantitative analysis of the thermally degraded film samples shows oxidation, chain-scission, and dehydration in the FTIR spectra in the MIR-range. Using NIR spectroscopy for the bulk samples, only minor changes in the FTIR spectra could be detected. However, using data processing, molecular information was extracted from the NIR range and a degradation model, using an artificial neural network, has been trained. Even though the changes due to thermal exposure were small, the presented model is capable of accurately predicting the time, temperature, and residual strength of the polymer.

show abstract

Identifying functional groups in IR spectra using an artificial neural network

Cited by 37 publications

References 20 publications

Functional Group Identification for FTIR Spectra Using Image-Based Machine Learning Models

Functional Group Identification for FTIR Spectra Using Image-Based Machine Learning Models

Spectral deep learning for prediction and prospective validation of functional groups

Prediction of Thermal Exposure and Mechanical Behavior of Epoxy Resin Using Artificial Neural Networks and Fourier Transform Infrared Spectroscopy

Contact Info

Product

Resources

About