Palmer, Kelly Emma; (1995) Transferability of near-infrared spectral libraries for the identification of pharmaceutical actives and excipients. Doctoral thesis (Ph.D.), University College London (United Kingdom).

Text Transferability_of_near-infrar.pdf Download (24MB)

Abstract

Chemometrics has been applied to near-infrared (NIR) spectroscopic data to facilitate the rapid identification and qualification of pharmaceutical materials. The procedure has clear advantages over traditional chemical methods: speed, non-destructive nature, and ease of sample preparation to mention a few. However, variations between instruments introduce small differences in spectra making it difficult to transfer spectral libraries between different instruments. For the technique to remain advantageous in the pharmaceutical industry this problem needs to be overcome. The aim of this research was to determine the best method for the construction and transfer of NIR spectral libraries between different instrumental setups. Direct transfer i.e. using a spectral library constructed on one instrument and used to identify a substance from its spectrum measured on a different instrument, was found possible between optically similar instruments for a limited selection of compounds. For larger libraries popular in the industry, identification could only be accomplished by applying a cascading procedure of a number of identification algorithms which were used depending on the nature of the similarities between compounds. When transferring spectral libraries between optically similar instruments of different manufacturers (Fourier transform technologies) it was found that the spectral resolution of each instrument and sample presentation was paramount to the success of transfer. A set of commercially available standards was utilised to correct for shifts in the wavelength and differences in photometric scale. However, when samples and transfer standards differed in their specular reflection component, transfer was less successful. Using a set of transfer standards matched to that of the samples allowed better transfer between optically similar dispersive instruments. The knowledge obtained through robust library construction, correction procedures and a thorough understanding of spectral differences demonstrated that library transfer can be achieved. However, limitations have been identified and a comprehensive list of these has been produced.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as XMLCSVJSON

Downloads by country - last 12 months

Export as XMLJSONCSV

Archive Staff Only

View Item