S. Haneklaus et al., APPLICATION OF LOW-RESOLUTION ENERGY-DISPERSIVE X-RAY-FLUORESCENCE SPECTROSCOPY TO TOTAL GLUCOSINOLATE DETERMINATION IN RAPESEED, Fett, 96(6), 1994, pp. 204-206
Due to the high standard of reproducibility and repeatability which is
already provided by wavelength X-RF spectrometers it is difficult to
see low resolution energy dispersive X-RF instruments entering the fie
ld of analyses for official certifications or research purposes. This
type of X-RF machines, however, seems to be suitable for the separatio
n of yield batches according to their glucosinolate content. The stand
ard error of estimation for that application is calculated to 3.5 mumo
l/g which is two times higher than those obtained from wavelength disp
ersive instruments (1.3 mumol/g) but, however, still much lower than t
hose derived from routine NIR instruments with fixed filters (8.9 - 18
.9 mumol/g). Thus this instrument might be an interesting alternative
to NIR machines, which are in the same cost range. Although the statis
tical values of the results obtained from energy dispersive X-RF are m
uch better than those derived from NIR it should be kept in mind, that
NIR provides more analytical possibilities for an agrochemical labora
tory (e.g. protein and oil determination). For the separation of seed
batches it is not necessary to cover the range of high glucosinolate c
oncentrations so that the settings of the instrument can be optimized
to the lower range of concentrations. This, however, requires that the
user need to be supplied with an adapted set of rape-seed reference m
aterials for the calibration in a certain range.