Comparaison des méthodes de détermination de la matière grasse dans les tourteaux d'oléagineux-Étude collaborative inter-laboratoires

Abstract

The data collected in France through a long-term survey on the quality of oilseed meals indicated that the use of unsuitable analytical methods may underestimate the residual oil content of rapeseed and sunflower meal. This misevaluation could make more difficult the optimizations of the oil extraction process and the formulation of feedstuffs. In order to evaluate the performance of current methods, a study was conducted in collaboration with most of the laboratories involved in the analysis of oilseed meal and feed within AFNOR V18A, the French committee for the standardization for feedstuffs. Eight samples were analysed. Five had a low oil content: rapeseed meals (3) from different oil mills, sunflower meal (1), soybean meal (1). Three had a high oil content: extruded full-fat soybeans (1), rapeseed cake (cold-pressed) (1) and linseed cake (cold-pressed) (1). The methods tested were based on three principles: (A) solvent extraction without hydrolysis (B) hydrolysis and solvent extraction (C) without hydrolysis and solvent extraction with a ball mill. The results showed that repeatability and reproducibility data were in agreement with those determined in previous ring-tests and mentioned in standards: NF V18-117 Part 1 for method A, NF V18-117 Part 2 for method B and ISO 734-2 method for C. Some differences for the determination of the oil content were observed between the three methods, particularly for samples that were heated during their production process. Methods A, C and B in that order gave increasing contents of residual oil in deoiled oilseed meals. The difference between results from methods A and B corresponding to the bound fat ranged from 0.7 (soybean meal) to 1.4 g/100 g (rapeseed meal) and represented around one half of the fat content determined by the method A. Recovery of the non linked residual oil was higher with method C than with method A by around 0.6 g/100 g in rapeseed meal, showing the ability of method C to extract the entrapped oil. This study showed that the three methods gave, as expected, different results depending on the principle of method and the process used for the meal or cake production. Finally, the choice of the most relevant method for fat determination should depend on how the results are used, for feed formulation or for the optimization of the extraction process. © 2013 A. Quinsac et al., published by EDP Sciences.