Formulation optimization for thermoplastic sizing polyetherimide dispersion by quantitative structure–property relationship: experiments and artificial neural networks

• Published in: Journal of materials science Vol. 50; no. 1; pp. 420 - 426
• Main Authors: Malho Rodrigues, A., Franceschi, S., Perez, E., Garrigues, J.-C.
• Format: Journal Article
• Language: English
• Published: Boston Springer US 01.01.2015; Springer; Springer Nature B.V; Springer Verlag
• Subjects: Accuracy; Artificial neural networks; Carbon fiber reinforced plastics; Characterization and Evaluation of Materials; Chemical Sciences; Chemistry and Materials Science; Classical Mechanics; Coatings; Control stability; Crystallography and Scattering Methods; Dispersions; Emulsions; Formulations; Materials Science; Mechanical properties; Neural networks; Optimization; Original Paper; Parameter identification; Particle size; Particle size distribution; Polyetherimide; Polyetherimides; Polymer Sciences; Polymers; Sizing; Solid Mechanics; Stability; Surface active agents; Thermoplastic resins; Photon Correlation Spectroscopy; Artificial Neural Network; Carbon Fibre; Aqueous Dispersion; Short Carbon Fibre
• ISSN: 0022-2461; 1573-4803
• DOI: 10.1007/s10853-014-8601-9

The main function of a sizing in a composite is to fill interface between fibre and polymer matrix. This coating process contributes to increase adhesion between fibre and matrix, and therefore improve mechanical properties of the composites. The aim of this study was to optimize sizing formulations by identifying the experimental parameters influencing the particle size, distribution size and stability of various aqueous emulsions, used in the coating process of carbon fibres with polyetherimide as thermoplastic sizing polymer. A quantitative structure–property relationship (QSPR) method with artificial neural networks was used to determine the main parameters involved in the different formulation steps. The results indicated three recurrent parameters: stirring speed, surfactant concentration and type of reactor which control the particle size, stability and distribution size of the dispersions. With a reduced dataset constituted of 36 entries, this QSPR method was able to predict the stability of the aqueous dispersion of polymer and the particle size with an accuracy of 200 nm for an average diameter ranging from 330 to 2700 nm. The distribution size could be predicted with an accuracy of 0.047 for an experimental size distribution of 0.2.