From spectroscopic data variability to optimal preprocessing: leveraging multivariate error in almond powder adulteration of different grain size.

Abstract

Analysing samples in their original form is increasingly crucial in analytical chemistry due to the need for efficient and sustainable practices. Analytical chemists face the dual challenge of achieving accuracy while detecting minute analyte quantities in complex matrices, often requiring sample pretreatment. This necessitates the use of advanced techniques with low detection limits, but the emphasis on sensitivity can conflict with efforts to simplify procedures and reduce solvent use. This article discusses the shift towards green analytical methods, focusing on portable spectroscopic techniques in the near-infrared (NIR) region. A case study involving the prediction of adulteration in almond flour with bitter almond flour illustrates the importance of particle size and the integration between the sample and the instrument. The study emphasizes the necessity of investigating the multivariate error associated with raw data to enhance data preprocessing strategies. This research provides valuable insights for professionals in the field, presenting a methodology applicable to a broad range of analytical applications while underscoring the critical role of raw data analysis in achieving accurate and reliable results.