Tensile and bending test of carbon/epoxy and carbon/geopolymer composites after temperature conditioning

Composites with epoxy matrix cannot be used in high temperature, while geopolymer matrix excel in high temperature resistance. First, prismatic specimens were subjected to conditioning temperature. Second, the tensile and bending test were performed at room temperature. This paper present comparison of mechanical properties of carbon/epoxy and carbon/geopolymer composites. Numerical simulations of tensile and bendidng tests were performed in finite element system Abaqus.


Introduction
Currently, composite materials are applied in many industrial areas. Usually, composites are made with carbon or glass fibers and epoxy matrix. These composites have very good mechanical properties, for example low weight, high strength and high stiffness, but they cannot be used in high temperature environments. Mechanical properties of these composites significantly degrade with increasing temperature. This shortcoming can be removed using a geopolymer matrix. Geopolymer matrix is an inorganic polymer material. Preparation of this material is based on aluminosilicate alkali activation. Geopolymers excel in many properties. Generally, the papers present high temperature resistance [1 -5] or resistance against acids and organic solvent agents. Any paper, which present specific mechanical properties of carbon/geopolymer composites, was not found.
This paper presents analysis of mechanical properties of carbon/epoxy and carbon/geopolymer composites. Composites with geopolymer matrix were subjected to high temperature. Two types of geopolymer matrix were used. Tensile and bending test were performed for identification of mechanical properties. Numerical simulations of these tests were performed in finite element system Abaqus.

Materials and specimens
The composites plates were made from 10 layers of plain weave carbon fabric ( Table 1). The prismatic specimens were cut using diamond blade. Epoxy matrix L285 and hardener 285 MGS was used for specimens with label CE1. Two types of geopolymer matrix were used in this work. Geopolymer matrix FC4 consists of potassium water glass, potassium hydroxide (KOH), silica fume, constituent with high content of metakaolinite, and boric acid. Molar ratios of the components are presented in Table 2. Geopolymer matrix B3P1 consists of potassium water glass, constituent with high content of metakaolinite, and ingredients with calcium. Constituent with metakaolonite Mefisto L 05 (produced by České lupkové zavody, a.s.) was used for both types of geopolymer matrix.  The specimens with geopolymer matrix were subjected to conditioning temperature of 23 °C, 200 °C, 400 °C or 600 °C, specimens with epoxy matrix only of 23 °C. Temperature resistance of the used epoxy matrix is lower than 200 °C. Afterwards, the tensile or bending test were performed using universal testing machine Zwick/Roell Z050 at room temperature. Designation of specimens is presented in Fig. 1.

Fig. 1. Designation of specimens
The geometric properties of all specimens with epoxy matrix were: width W e = 25 mm, thickness H e = 2.4 mm, total length L e = 180 mm. The geometric properties of specimens with geopolymer matrix were: width W g = 25÷26 mm, thickness H g = 2.9÷3.6 mm, total length L g = 150 mm. The exact dimensions for each specimen are given in [5].

Tensile test
The force-displacement (F-Δl) dependencies were obtained from tensile test complying with ASTM D 3039. The specimen size was modified according to possibilities resulting from the plate size. An initial grip distance was l j = 100 mm. An extensometer was used for measuring the elongation. Gage length was l e = 60 mm. The load velocity (crosshead displacement) was v T = 2 mm/min.
The stress-strain dependencies were calculated using (1) where F is loading force, W is width of specimen, H is thickness of specimen, Δl is elongation and l e is initial gage length. The effective modulus was identified according to the standard ASTM D 3039 on interval of strain ⟨0.1 ,0.3 ⟩ The stress-strain dependencies of CB4 and CF1 specimens for all temperature conditioning are shown in Fig. 2 and Fig. 3, for CE1 specimens at 23 °C in Fig. 4. The values of maximum stress and effective modulus are presented in Table 3 and Table 4.    The stress-strain dependencies are linear for specimens with epoxy matrix and they are nonlinear for specimens with geopolymer matrix. The effective modulus in tension for CB4 specimens with B3P1 geopolymer matrix is approximately half that of the CF1 specimen with FC4 geopolymer matrix. The effective modulus in tension of CE1 specimen with epoxy matrix is larger by 60% that of CF1 specimen for 23 °C temperature conditioning.

Numerical simulation of the tensile test
The finite element system Abaqus was used for the numerical simulation of the tensile test. Quadratic hexahedral elements with 20 nodes were used in a parametrically created model (Fig. 5). The loading was controlled by the displacement of the upper crosshead. One numerical model was created for each group of specimens (material of matrix + temperature conditioning). This model had averaged geometric parameters. Isotropic material model with average effective modulus ( Table 2) was used in numerical analysis. The stress-strain dependencies obtained using finite element analysis (FEA) are presented in Fig. 2 -Fig. 4.

Bending test
The force-displacement (F-Δl) dependencies were obtained from 3-point bending test complying with ASTM D 7264. The support span was l s = 80 mm. An extensometer was used for measuring the displacement of loading nose (deflection of specimen). The load velocity (crosshead displacement) was v B = 2 mm/min.

Numerical simulation of the bending test
The finite element system Abaqus was used for the numerical simulation of the bending test. Quadratic hexahedral elements with 20 nodes were used in a parametrically created model One numerical model was created for each group of specimens (material of matrix + temperature conditioning). This model had averaged geometric parameters. Isotropic material model with average effective modulus ( Table 2) was used in numerical analysis. The force-displacement dependencies obtained using finite element model are presented in Fig. 6 -Fig. 8.   Fig. 6. Force-displacement dependencies for CB4 specimens

Conclusion
The composite specimens were composed of carbon fibers and geopolymer or epoxy matrix. Two types of geopolymer matrix were used. The specimens were subjected to conditioning temperature at 23 °C, 200 °C, 400 °C or 600 °C. Afterwards the tensile or bending tests were performed at room temperature. The numerical simulation of tensile and bending tests were performed in finite element system Abaqus.
The stress-strain dependencies were calculated for tensile test. The stress-strain depedencies are linear for specimens with epoxy matrix and they are nonlinear for specimens with geopolymer matrix. The effective modulus in tension for CB4 specimens with B3P1 geopolymer matrix is approximately half that of the CF1 specimen with FC4 geopolymer matrix. The effective modulus in tension of CE1 specimen with epoxy matrix is larger by 60% that of CF1 specimen for 23 °C temperature conditioning.
In case of bending test, the force-displacement dependencies were obtained from 3point bending test. The effective modulus in tension was used in numerical simulation of bending tests. These numerical models showed higher stiffness than the experiment.
The multilayer composites show different value of the effective elasticity modulus for tension and bending. Therefore, in the 3D solid finite element model, the effective elasticity for bending has to be used in the 3 point bending model. This publication was supported by the project LO1506 of the Czech Ministry of Education, Youth and Sports. The publication is a result of the project Development of the UniCRE Centre (project code LO1606) which was financially supported by the Ministry of Education, Youth and Sports of the Czech Republic under the National Programme for Sustainability I. The authors would like to thank Robert Zemčík for helpful comments and proofreading of the manuscript.