Search

Neste trabalho 20 amostras comerciais sob a forma de comprimidos e 15sintéticas de cloridrato de metformina foram utilizadas e seus espectros adquiridos na faixa de 4000–650 cm-1 por espectroscopia difusa no infravermelho com transformada de Fourier (DRIFTS). Foram construídos e comparados modelos de calibração multivariada utilizando toda a região espectral analisada (modelo global PLS) e modelos otimizados empregando sub-regiões dos espectros FT-IR (modelos iPLS e siPLS). Os melhores resultados foram aqueles obtidos utilizando o algoritmo siPLS, dividindo o espectro em 40 intervalos e combinando 3 intervalos (37, 39 e 40), 4 variáveis latentes, e apresentou um valor de RMSEP inferior ao obtido utilizando o modelo global e o algoritmo iPLS. Assim, a aplicação das técnicas de seleção das regiões dos espectros de infravermelho (iPLS e siPLS) mostraram-se eficientes no desenvolvimento de metodologias mais simples, rápidas e não destrutivas para a análise das formulações farmacêuticas contendo metformida na forma de comprimidos, destacando-se a potencialidade destas técnicas no controle e fiscalização de medicamentos industrializados.In this work 20 metformin hydrochloride commercial pill samples and 14 synthetic samples were used and their spectra acquired in the 4000 – 650 cm-1 range by Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS). Multivariate calibration models were built and compared using all spectral region (global model PLS) and optimized models applying FTIR spectral sub-regions (iPLS and siPLS models). The best results were the ones obtained by using the siPLS algorithm, dividing the spectrum into 40 intervals and combining 3 intervals (37, 39 and 40) with 4 latent variables, which also presented an RMSEP value lower than the one obtained by using the global model and the iPLS algorithm. Finally, the application of selection techniques for infrared spectral regions (iPLS and siPLS) has been efficient to develop simpler, faster and non-destructive methodologies to analyze pharmaceutical pill formulations containing metformin, highlighting the potentialities of these techniques to control and inspect industrialized medicines.

In this work 20 metformin hydrochloride commercial pill samples and 14 synthetic samples were used and their spectra acquired in the 4000 – 650 cm-1 range by Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS). Multivariate calibration models were built and compared using all spectral region (global model PLS) and optimized models applying FTIR spectral sub-regions (iPLS and siPLS models). The best results were the ones obtained by using the siPLS algorithm, dividing the spectrum into 40 intervals and combining 3 intervals (37, 39 and 40) with 4 latent variables, which also presented an RMSEP value lower than the one obtained by using the global model and the iPLS algorithm. Finally, the application of selection techniques for infrared spectral regions (iPLS and siPLS) has been efficient to develop simpler, faster and non-destructive methodologies to analyze pharmaceutical pill formulations containing metformin, highlighting the potentialities of these techniques to control and inspect industrialized medicines.

In this work 20 metformin hydrochloride commercial pill samples and 14 synthetic samples were used and their spectra acquired in the 4000 – 650 cm-1 range by Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS). Multivariate calibration models were built and compared using all spectral region (global model PLS) and optimized models applying FTIR spectral sub-regions (iPLS and siPLS models). The best results were the ones obtained by using the siPLS algorithm, dividing the spectrum into 40 intervals and combining 3 intervals (37, 39 and 40) with 4 latent variables, which also presented an RMSEP value lower than the one obtained by using the global model and the iPLS algorithm. Finally, the application of selection techniques for infrared spectral regions (iPLS and siPLS) has been efficient to develop simpler, faster and non-destructive methodologies to analyze pharmaceutical pill formulations containing metformin, highlighting the potentialities of these techniques to control and inspect industrialized medicines.