Multivariate Analysis of Formalin Using UV-Vis Spectroscopy

Abstract

An investigation on formaldehyde detection using UV-Vis spectrophotometer has been explored in this work. A 37 % aqueous solution of formaldehyde known as formalin is categorized as potential carcinogen and a threat to human health. Formalin is a widely used food adulterant and excessive consumptions may adversely affect lifespan. In this study different concentrations of formalin solutions were prepared and examined using a UV-Vis spectrophotometer. During the experiment the absorbance spectra were recorded by UV Probe user interface within the wavelength range of 200400 nm. The absorbance peak obtained at a wavelength of 207 nm depicted a linear relation with concentration variation from $500-1000 \mu \mathrm{M}$. The qualitative data analysis tool, principal component analysis (PCA) was implemented and a successful data clustering was observed between four different concentrations $1000 \mu \mathrm{M}, 400 \mu \mathrm{M}, 200 \mu \mathrm{M}$ and $20 \mu \mathrm{M}$. A high measure of class separability index (SI) 87.7 was obtained from the PCA analysis. The quantitative estimation of formalin concentrations $1000 \mu \mathrm{M}, 400 \mu \mathrm{M}, 200 \mu \mathrm{M}$ and $20 \mu \mathrm{M}$ was carried out using partial least square regression (PLSR) analysis and principal component regression (PCR). In PLSR, the prediction analysis provided a high correlation factor 0.98 and a high average prediction accuracy of 84% while in case of PCR the correlation factor was 0.97 and average prediction accuracy was obtained as 83.4%. The experimental observations suggest that the UV-Vis technique may be applicable for qualitative analysis of formalin in food products.