Riassunto analitico
The problem of qualify in a fast and reliable way the chemicals during the production process is increasingly pressing. This is true for all the industrial sectors but for pharmaceutical companies, because of the intended end use of their products, this issue is especially relevant. This thesis reports the use on the field of some (almost) new-generation tools under the framework of Process Analytical Technology (PAT). The experimental issues faced during the internship conduct with a Danish international pharmaceutical company are addressed by means of Near Infrared Spectroscopy (NIRS). This spectral technique owes its widespread diffusion to the multivariate methods collected under the name of chemometrics. In short, chemometrics is the science of making sense of data in chemistry. In consideration of the data explosion that followed the computerized laboratory automation11, chemometrics is an extremely precious resource, both for industries and universities, in the developing of Quality by Design (QbD) processes and experiments. The PAT theoretical tools used throughout this work make possible the development of simulation models whose mathematical relationships mimic physicochemical phenomena and, therefore, can be used as predictive tools. Prediction capability is probably the most noteworthy attainment of the PAT approach to industrial problems: models give us the opportunity to make crucial decisions for process quality and product conformance based on rigorous statistical principles.
|