The pharmaceutical drying process is critical to the quality of the active
pharmaceutical ingredient or the drug product as it is often the final unit
operation. The non-uniformity of drying is particularly challenging as it renders
inaccurate the determined solvent content and drying endpoint, which is a
critical quality attribute. As process analytical technology further develops,
novel techniques are implemented to address process and product challenges.
Spatially resolved spectroscopy is a technique in which spectra are collected
from spatially resolved distances from the incident light. This technique is used
for the characterisation of non-uniform media. In this thesis, the use of spatially
resolved spectroscopy for the monitoring of the drying of an active
pharmaceutical ingredient is reported.
Pharmaceutical drying of model systems was monitored using spatially
resolved spectroscopy. In this work, three bespoke probes were used for the
collection of spatially resolved spectra. One probe allows the collection of
spatially and angularly resolved diffuse reflectance near-infrared
measurements (SAR-DRM), while the other was developed for the collection
of spatially offset Raman spectroscopy (SORS) measurements. The third
probe combines both techniques and was developed for the collection of both
spatially resolved near-infrared spectra and spatially offset Raman spectra and
is termed the combined probe. This thesis details the in-line and at-line
application for industrial process monitoring using these techniques, which to
our knowledge were not applied in this setting.
The drying of two grades of paracetamol, granular and powder, in the solvents
n-heptane and methyl tert-butyl ether was monitored using SAR-DRM. The
drying of granular and powder paracetamol in the solvents anisole and methyl
tert-butyl ether was monitored using SORS. Partial least squares regression
(PLSR) analysis was used for the estimation of the solvent content using spectra from the individual signal collection configurations, in addition to a
combination of the configurations. Results from both techniques suggest that
PLSR models of spectra collected from larger distances lead to more accurate
vii
estimations of the solvent content. This was attributed to the larger volume of
the drying powder cake probed by those techniques. Since the drying of
paracetamol in methyl tert-butyl ether could be monitored using both
techniques, the combined probe was used for monitoring this system, and
multi-block PLSR analysis was conducted using both near-infrared and Raman
spectra of combinations of the configurations. The multi-block PLSR model
performance was similar to that of the individual SORS spectra, which was
attributed to the stronger signals and spectral features of the Raman signal
compared to the near-infrared measurements. Since the application of SORS
for the monitoring of pharmaceutical drying was demonstrated and showed
improvement in PLSR model performance and solvent content estimation,
SORS was further used for the monitoring of the washing with methyl tert-butyl
ether a paracetamol filter cake wet with anisole. The results similarly showed
improved estimations of the content of both solvents in the filter cake from
spectra from larger offset distances. The outcomes of the studies in this thesis
demonstrate the advantage of the application of spatially resolved
spectroscopy for monitoring the solvent content in pharmaceutical drying. The
use of such novel process analytical technology offers potential for improved
process monitoring and accurate prediction of the process end point.
| Date of Award | 10 Sept 2024 |
|---|
| Original language | English |
|---|
| Awarding Institution | - University Of Strathclyde
|
|---|
| Sponsors | University of Strathclyde |
|---|
| Supervisor | Yi-Chieh Chen (Supervisor) & Chris John Price (Supervisor) |
|---|