Parenteral pharmaceutical products are delivered into the human body by injection and infusion. Due to the different properties of the active pharmaceutical ingredients, parenteral products exist in different forms: either as a powder or a solution. All pharmaceutical products must meet certain quality standards prescribed by healthcare organizations. Parenteral products must also meet the requirement for sterility, as they are by definition intended for direct application into the human body past the digestive system. In addition to the requirement for sterility, parenteral products must meet the pharmacopoeial requirements for the absence of visible particles. Quality control of parenteral products is rigorous: each container is 100% visually inspected before packaging. Containers with particles are rejected during 100% control. In order to track the incidence of particles, visible particles found in the rejected containers need to be identified. These are mostly particles that are in direct contact with the product during the production process: e.g. primary packaging, stoppers, filling needles, etc. Identification is performed on several levels, and the final stage is identification based on various spectroscopic techniques, the most commonly used of which are Raman spectroscopy, FTIR spectroscopy and LIBS spectroscopy. Identification is usually performed by comparing the spectra of particles of unknown materials with those of the reference materials whose source is known. A successful spectral library is crucial for successful identification of particles based on spectra comparisons with reference spectra.
The master's thesis covers the construction of a spectral library for the purpose of particle identification by IR spectroscopy in the software environment Spectrum 10 (Perkin Elmer). Each library requires an appropriate framework within a specific software package, which I have adapted to the needs of the Laboratory for Particle Analysis in the Lek company. Reference materials were obtained from the production unit and I recorded their infrared spectra using various techniques. Most often, I used the ATR mode with diamond crystal. The materials that were thinner and more transparent were also recorded in transmission mode. In this way, I compared the methods with each other, especially the differences in the recorded spectra and the appropriateness of the method to a particular material. I found that the transmission mode fails for thicker specimens, so I mostly focused on the ATR method. I added the spectra to the library and tested its suitability. The library can be upgraded with new entries in the future and represents an important contribution to the identification of particles in the Laboratory for Particle Analysis in Lek company.