Rapid quantification of NaDCC for water purification tablets in commercial production using ATR-FTIR spectroscopy based on machine learning techniques

Loading...
Thumbnail Image
Files
978-3-031-26438-2_9.pdf(1.27 MB)
Published version
Date
2023-02-23
Authors
Asadi, Hamzeh
O’Mahony, Tom
Lambert, Julie
Brown, Kenneth N.
Journal Title
Journal ISSN
Volume Title
Publisher
Springer
Research Projects
Organizational Units
Journal Issue
Abstract
Accurate, fast and simple quantitative analysis of solid dosage forms is required for efficient pharmaceutical manufacturing. A spectroscopic analysis in ATR-FTIR (Attenuated Total Reflection-Fourier Transform Infrared) mode was developed for NaDCC (Sodium dichloroisocyanurate) quantification. This fast and low-cost method can be used to quantify NaDCC solid dosage forms using ATR-FTIR in absorbance mode in conjunction with partial least squares. A simple sampling procedure is included in the proposed experiment by just dissolving the samples in deionized water. An algorithm pipeline is also included for data cleaning, such as outlier removal, scatter correction, scaling, and mapping of the sample’s spectrum to a NaDCC concentration. In addition, a simple model based on Beer’s law was evaluated on a sub-range of 1220−1830cm−1. Furthermore, a variable selection algorithm shows minimum excipient interference from the sample matrix in addition to visual analysis. A statistical analysis of the proposed method shows that it demonstrates a promising result with a regression coefficient of 0.996 (R2=0.996) and recovery range of 95.5%–107%. As a result of the positive correlation of ATR-FTIR with NaDCC concentration, and in conjunction with the proposed method, this can serve as a clean, fast, affordable and eco-friendly method for pharmaceutical analysis.
Description
Keywords
Machine learning , ATR-FTIR , Chemometric
Citation
Asadi, H., O’Mahony, T., Lambert, J. and Brown, K.N. (2023) ‘Rapid quantification of nadcc for water purification tablets in commercial production using atr-ftir spectroscopy based on machine learning techniques’, AICS 2022, in L. Longo and R. O’Reilly (eds) Artificial Intelligence and Cognitive Science, Communications in Computer and Information Science, vol 1662, Cham: Springer Nature Switzerland, pp. 106–120. https://doi.org/10.1007/978-3-031-26438-2_9
Link to publisher’s version