Robustness of the near infrared spectroscopy method determined using univariate and multivariate approach

摘要

The robustness assessment is a part of a method validation protocol, during which several characteristics of an analytical method are also evaluated (e.g. accuracy, repeatability, reproducibility, linearity, interme-diate precision, measurement uncertainty) in order to assess its fitness for purpose. The purpose of robustness assessment of the near infrared spectroscopy method (NIRS) is to indicate which factor signif-icantly influence the obtained results, as well as to point to the potential problems that might occur in the routine application of the method. The assessment of robustness of the NIRS method included variation of certain operational and environmental factors at three level (-1,0,1) by applying univariate (one-variable-at-a-time, OVAT) and multivariate (multivariate-at-a-time, MVAT) approach to the experimental design. Operational and environmental factors that were varied included the number of subsamples to be measured in the NIRS measurement (1), environmental temperature (2), sample temperature (3), environmental air humidity (4), instrument voltage (5) and lamp aging (6). Regardless the applied exper-imental design, external factors with significant influence on obtained NIRS results were indicated, as well as pointed the potential problems that might occur in the routine application of the method. In order to avoid them, every effort should be made to stabilize instrument and sample temperature and to standardize the homogeneity and number of subsamples to be measured in NIRS measurement. Moreover, the obtained results highlighted the necessity that the NIRS instruments should work through a voltage regulator.