Use of NIRS technology with a remote reflectance fibre-optic probe for predicting mineral composition (Ca, K, P, Fe, Mn, Na, Zn), protein and moisture in alfalfa

• Published in: Analytical and bioanalytical chemistry Vol. 387; no. 6; pp. 2199 - 2205
• Main Authors: González-Martín, I, Hernández-Hierro, J. M, González-Cabrera, J. M
• Format: Journal Article
• Language: English
• Published: Germany Berlin/Heidelberg : Springer-Verlag 01-03-2007
• Subjects: alfalfa; Calibration; Fibre-optic probe; Medicago sativa - chemistry; Mineral composition; Minerals - analysis; near-infrared spectroscopy; Plant Proteins - analysis; proteins; Spectroscopy, Near-Infrared - instrumentation; Spectroscopy, Near-Infrared - methods; Water - analysis; water content
• ISSN: 1618-2642; 1618-2650
• DOI: 10.1007/s00216-006-1039-4

In the present work we study the use of near-infrared spectroscopy (NIRS) technology together with a remote reflectance fibre-optic probe for the analysis of major (Ca, K, P) and minor (Fe, Mn, Na, Zn) elements, protein and moisture in alfalfa. The method allows immediate analysis of the alfalfa without prior sample treatment or destruction through direct application of the fibre-optic probe on ground samples in the case of the mineral composition and on-ground and compacted (baled) samples in the case of protein and humidity. The regression method employed was modified partial least-squares (MPLS). The calibration results obtained using samples of alfalfa allowed the determination of Ca, K, P, Fe, Mn, Na and Zn, with a standard error of prediction (SEP(C)) and a correlation coefficient (RSQ) expressed in mg/kg of alfalfa of 1.37 x 10³ and 0.878 for Ca, 1.10 x 10³ and 0.899 for K, 227 and 0.909 for P, 103 and 0.948 for Fe, 5.1 and 0.843 for Mn, 86.2 and 0.979 for Na, and of 1.9 and 0.853 for Zn, respectively. The SEP(C) and RSQ values (in %) for protein and moisture in ground samples were 0.548 and 0.871 and 0.150 and 0.981, respectively; while in the compacted samples they were 0.564 and 0.826 and 0.262 and 0.935, respectively. The prediction capacity of the model and the robustness of the method were checked in the external validation in alfalfa samples of unknown composition, and the results confirmed the suitability of the method.