Investigation on near-infrared and Raman spectral variations of solid samples depending on their particle size and subsequent influence on quantitative analysis

Abstract

This thesis aims to systematically investigate the particle size-dependent spectral variation in Raman and NIR measurements. For this purpose, both pure lactose powders and compressed pellets with five different particle sizes were prepared. Further, two types of binary composition samples, powders of ambroxol/lactose and pellets of lactose/polyethylene (PE), were prepared. In the ambroxol/lactose powders, only the particle sizes of lactose varied with maintaining the particle size of ambroxol unchanged. Only the particle size of lactose was varied in the lactose/PE pellets. In chapter I, the results observed in Raman spectra of the designed samples were discussed. The intensity of lactose peaks elevated with the increase of particle size in both cases when the diameter of detecting window was 25. 4 mm. The broadly generated Raman photons in the samples with the larger lactose particles were still reachable to the detector, since the detecting window was large. In contrast, the trend was reversed in the measurement of lactose powders when the detecting window decreased to 10.0 mm. Since the photon distribution became larger than the size of detecting window with the increase of particle size, the loss of photons in detection occurred and the decrease of intensity was resultant. In the case of binary composition samples, the particle-size dependent intensity variations were dissimilar when the size of detecting window varied. To discuss the particle-size dependent intensity variation, it was critically necessary to consider distribution of Raman photons in a sample which was related with many factors including particle size and specify the size as well as location of a detector. Depends on these factors, the trends of intensity variations according to particle size could be different and difficult to generalize. In chapter II, NIR absorbances of lactose powder and pellet elevated with the increase of particle size since the number of photons reaching to a detector (equivalent to attenuation of intensity) decreased due to the longer travel distance inside the samples to the detector. In a meanwhile, in the case of lactose/PE pellets, the absorbance of lactose diminished with the increase of lactose particle size. In these pellets composed of larger lactose particles, the particles were much more remotely located with each other. Under this situation, the photons might not interact with the big lactose particles on a certain occasion or partially interact with them, when the volume of photon propagation was small or limited due to the smaller particle size of surrounding PE. Therefore, the decrease of lactose absorbance with the increase of lactose particle size was resultant. In both Raman and NIR measurements, the accuracies for the determination of component concentrations degraded as the particle size became larger. Since the larger particle made the photons diffused more broadly, the uncertainty in photon propagation subsequently elevated and degradation of reproducibility in spectral acquisition was thereby consequential. The potential degradation of accuracy should be cautiously considered for reliable quantitative analysis based on Raman as well as NIR spectroscopy. In final, Monte Carlo was effective to probe photon propagation inside the samples and helpful to support the experimental observations.