Application Research of Near-infrared Spectroscopy (NIRS) in Online Monitoring of the Mixing Uniformity of Intermediate Granules

Abstract

This study focuses on the application of near-infrared spectroscopy (NIRS) in the online monitoring of the mixing uniformity of intermediate granules, taking compound cold tablets (referring to Yabang "Yabang Pharma") and Loratadine, Paracetamol, and Pseudoephedrine Sulfate Sustained-release Tablets ("Yabang Lite®") as the research references. By comparing NIRS with high-performance liquid chromatography (HPLC), using the coefficient of variation (CV) and standard deviation (SD) as evaluation indicators, the results show that the average CV of NIRS is 5.2%, and that of HPLC is 6.1%. The average SD of NIRS is 0.08, and that of HPLC is 0.11. After a paired t-test, P>0.05, indicating that the two methods are similar in the accuracy of evaluating the mixing uniformity. However, NIRS has the ability to perform real-time monitoring, while HPLC takes 30-60 minutes to complete one test. Taking the mixing process of the intermedi-ate granules of compound cold tablets as the research object, the changes in the intensity and shape of the absorption peaks of the NIRS spectrum at different mixing times are analyzed. It is found that the intensity of the absorption peaks of the spectrum fluctuates significantly within 0-10 minutes of mixing, gradually stabilizes, and the peak shape tends to be symmetrical within 10-30 minutes. The spectral characteristics are basically constant after more than 30 minutes. Based on this, the mixing process and the optimal mixing time can be accurately determined. Different chemometric methods are used to construct the NIRS model. Among them, the root mean square error of calibration (RMSEC) of the PCA-PLS model is 0.04, the root mean square error of prediction (RMSEP) is 0.06, and the coefficient of determination (R²) is 0.97, with the best prediction accuracy. The study also finds that the particle size distribution has a significant impact on the detection sensitivity of NIRS. When the particle size distribution is in the narrow range of 50-100μm, the CV of NIRS detection is 4.5% and the SD is 0.07. When the particle size distribution is in the wider range of 20-200μm, the CV rises to 7.2% and the SD is 0.10, indicating that a narrow particle size distribution is more conducive to improving the detection sensitivity of NIRS. This study provides a solid theoretical basis and practical guidance for the widespread application of NIRS in the online monitoring of the mixing uniformity of granule intermediates during the production process.