Particle size optimisation in development of near infrared microscopy methodology to build spectral libraries of animal feeds.

Abstract

Near infrared spectroscopy coupled to a microscope (NIRM) is a technique where the infrared beam is focused on each particle (< 50 μm) of a sample to collect NIR spectra. In this work, the experimental conditions (sample pre-treatment and instrumental) have been evaluated and compared for collecting spectral information in order to develop a strategy able to build large and robust spectral libraries to be transferred between NIRM instruments. To attempt this objective, we selected the most representative ingredients included in animal feeds: straw as low density forage, soybean as raw material and total mixed ration (TMR). Then, different mills, sieves and the number of spectra to be averaged were evaluated, selecting those that minimise noise and scanning time. The best results were obtained grinding through 1 mm sieve ingredients and TMR and collecting each particle spectra by scanning 70 individual spectra averaged as pre-treatment and instrumental conditions to obtain the most relevant information about each ingredient