Modeling of the Tensile Strength of Immiscible Binary Polymer Blends Considering the Effects of Polymer/Polymer Interface and Morphological Variation

Esmail Sharifzadeh

doi:10.1007/s10118-019-2274-4

Abstract

In this work, a unique model is proposed for predicting the tensile strength of binary polymer blends considering the effects of polymer/polymer interface and the morphological variation of the system. The modeling was performed based on the combination of analytical and artificial neural network (ANN) modeling methods. For the analytical part, Kolarik’s model was developed in accordance with the system requirements and ANN was simultaneously involved in order to interpret some effective model parameters using the tensile test result of an actual sample (e.g. the yield strength and thickness of the interface, etc.). Furthermore, the model accuracy was evaluated by comparing the tensile test results of differently prepared iPP/PA and PS/PMMA blend samples and also some other data from literature with the model predictions. It was revealed that the designed ANN perfectly elevates the capability of the analytical section in order to predict the tensile strength of binary polymer blends with different compositions (prediction error < 10%).

Highlights

Improving the mechanical and physical properties of an individual polymer phase via blending processes and adding nanoparticles is quite well-known.[1−10] the presence of many involving parameters in such processes makes it quite challenging to control the properties of the final products.[11]
It was revealed that the designed artificial neural network (ANN) perfectly elevates the capability of the analytical section in order to predict the tensile strength of binary polymer blends with different compositions
Mathematical modeling has always been considered as an efficient method for reducing unnecessary costs.[12−16] Focusing on the immiscible binary polymer blends, there are many practical models which are capable of providing reliable predictions for the mechanical properties.[16−19] The improvement of modeling procedures during the last decades has been increasingly elevating the prediction accuracy and reliability due to the consideration of more detailed parameters in the models.[16−19] for polymer blends, the morphological variation and the formation of polymer/polymer interface can be considered as the most effective parameters in determination of the mechanical/physical properties.[20,21]

Summary

INTRODUCTION

Improving the mechanical and physical properties of an individual polymer phase via blending processes and adding nanoparticles is quite well-known.[1−10] the presence of many involving parameters in such processes (e.g. random orientation of nanoparticles, properties of the polymer/nanoparticle interphase, properties of the polymer/polymer interface, blend morphology, etc.) makes it quite challenging to control the properties of the final products.[11]. Sci. 2019, 37, 1176–1182 taneously considered.[5,16,19] the designed ANN was appended to the procedure in order to interpret some effective model parameters using the tensile test result of an actual sample. Considering a polymer/polymer interface region, Kolarik’s COS model can be reformed as a new geometrical structure in which the morphological variation is involved (Fig. 1).[16,19] The presence of both phases in the interface region makes it crucial to consider its role in determining the final tensile strength of the system.[31] Eq (1) can be reformed based on the geometry of the model structure at different intervals and the phase inversion point. Where, φ2 denotes the actual (experimental) volume fractions of Phases (II), In the step, the designed ANN can be applied in order to interpret parameters “A”, “Sui”, and “yp” using an actual blend samples in the 1st (or 4th) interval.

80 Model predictions

RESULTS AND DISCUSSION

CONCLUSIONS