| Abstract: | The development of a simple and cost-effective method for the determination of the release of coated urea has significant implications. Fourier-transform infrared spectroscopy with attenuated total reflectance was employed to determine the release of urea through univariate and multivariate calibration. The results indicated that univariate calibration did not accurately predict the release of urea, whereas partial least squares based on multivariate calibration performed significantly better. Partial least squares had the highest accuracy when the band located at 1420–1520 per centimeter was employed as the input. Moreover, the accuracy was further improved when segmented partial least squares models were developed at low and high urea concentrations. Unsegmented and segmented partial least squares models were employed, and release values were comparable to those measured by colorimetry. This work demonstrated the use of infrared spectroscopy and partial least squares to characterize the release of coated urea. |
