Model‐Based Recognition of Mid‐Infrared Sensor Fouling in Paracetamol Crystallization

A procedure for the recognition of undesired solid deposition on the optical element of an attenuated total reflectance probe for online mid‐infrared spectroscopy during crystallization processes is presented. Unlike other approaches, the method exclusively uses intelligent spectral analysis for solid recognition and does not require additional devices. Indirect hard modeling enables the separation of the solid material's contribution to the mixture spectra from those contributions of the dissolved material and the solvents, expressed through the predicted component weights. The detection of the solid component weight is promising for implementation into process monitoring or control.