Methods of in-situ monitoring of suspension polymerisation for process understanding and optimisation

  • Barry Thomas Robertson

Student thesis: Doctoral Thesis


Increasing Environmental and Health and Safety regulations mean that the demands for cost-effective, accurate and information-rich methods of process analysis are increasing exponentially. This project aimed to evaluate the use of technologies – Raman and Acoustic Emission Spectroscopies – coupled with a variety of multivariate analysis tools to provide a framework for the optimisation of polystyrene suspension polymerisation, allowing industrial partners to monitor the reaction progression, determine particle size information and unreacted monomer concentration. Spectral data collected during a series of lab-scale polymerisation reactions and basic model mixtures was used to determine the effectiveness of each method – including the use of a variety of probe configurations for Raman analysis. The data was treated with established pre-treatment methods (Savitzky-Golay filtering, SNV transformation and EMSC) and a novel method (OPLECm) to enhance the performance of mathematical models and investigate the effectiveness of the methods for this application. The results indicate that Raman and Acoustic Emission spectroscopies can provide monomer concentration and particle size information for this reaction, respectively. Offline Raman data is shown to be approximately 33% less variable than current offline HPLC methods, and in-situ analysis showing qualitatively similar results to offline gravimetric determination of residual monomer. Furthermore, the potential benefit of increasing laser diameter is shown. The pre-treatment of this data prior to modelling shows Savitzky-Golay derivatisation to provide the least improvement (11.4% RMSEp); with SNV, EMSC and OPLECm performing similarly (4.1, 3.8 and 4.1% respectively). Models built with EMSC and OPLECm pre-treatment provide best results overall, with just 4 and 3 latent variables, respectively. Finally, Acoustic Emission spectroscopy provided data which showed good correlation to offline sieving analysis, indicating a strong potential for its use in PSD determination during this reaction.
Date of Award4 Oct 2019
Original languageEnglish
Awarding Institution
  • University Of Strathclyde
SponsorsEU & University of Strathclyde
SupervisorAlison Nordon (Supervisor) & David Littlejohn (Supervisor)

Cite this