Estimation of the Number of Scans Required per Hard-to-Clean Location and Establishing the Limit of Quantification of a Partial Least Squares Calibration Model When the FTIR Is Used for Pharmaceutical Cleaning Verification.

Abstract

This study aims to identify two critical components required for pharmaceutical cleaning verification when an FTIR is used: (a) the number of scans required per hard-to-clean location, and (b) the limit of quantification (LOQ) of the FTIR instrument when measuring the surface contamination. The current practice in pharmaceutical manufacturing does not require multiple samples as it is standard practice to collect a single swab sample from a 25 × 25 cm area from a difficult-to-reach area of the manufacturing equipment. However, since the FTIR will only scan a tiny portion of the surface compared to the swab, a sufficient number of samples (data points) are required to provide enough confidence to ensure that the measurement results are close to the true value with a maximum degree of certainty. Similarly, calculating the LOQ for a linear regression could be straightforward. However, complexity arises when the experimental data are complex; in this case, the complexity arises due to the nature of the measurement and the lack of the defined peak in the pre-processed spectra. Therefore, this study uses the practical approach of calculating the sample size and the LOQ.