Observations on determination of dysprosium in labeled digesta samples by near infrared spectroscopy

The objective of this work was to determine whether the Dy content of ruminal digesta samples labeled by pulse-dosing intra-ruminally with Dy-labeled forages could be determined by near infrared reflectance spectroscopy. Ruminal grab samples were obtained from four steers. Equal numbers of alfalfa and orchardgrass samples were collected at 17 times (0 to 165 h) after dosing and analyzed (total of 272 samples). Dysprosium was determined by direct current plasma emission spectrometry. Spectra were obtained by scanning ground samples on a scanning monochromator. Calibrations for Dy (0 to 349.5 ppm; mean = 44.5 ppm) were carried out using both stepwise and partial least square regressions. Wavelength selections using stepwise regression included characteristic Dy wavelengths and resulted in R2 and standard error of analysis of approximately .9 and > or = 20 ppm, respectively. Partial least squares regressions, although more robust, were not significantly better. Sample ashing to increase relative Dy concentrations did not improve the results. Further investigations indicated that interference by Co marker, also present in the samples, was a problem. Although near infrared reflectance spectroscopy is capable of determining Dy in labeled forages, further work is needed to increase its accuracy.