Cyclic thermoresistivity of resin-based carbon fiber bars modified with mixed carbon powder and nano-silica

The addition of conductive fillers to the carbon fiber bar reduces its resistivity and improves electrical stability. Unstable conductivity and susceptibility to ambient temperature change impede the application of this technique in engineering, unless such influence can be eliminated by technical means or precisely predict. In this paper, modified epoxy Resin based carbon fiber bars with 4 sets of different mixed fillers have been designed to evaluate the temperature-resistance effect under different temperature cycles. Results show that the initial volume resistivity reduces due to the incorporation of carbon powder(CP) and nano-silica(NS), and meets the lowest when the CP and NS mass ratio is 1:0.6. The volume resistivity increases linearly with the temperature rising, and reaches the maximum temperature sensitivity coefficient of 78.8%. During the temperature cycle process, the volume resistivity of all specimens first decreases and then increases with the increasing temperature uniformly. The three groups (CP/NS ratio 1:0.2, 1:0.6, and 1:0) share the same PTC effect transition temperature range, from 30 to 60 °C. And for the 1:1 group, the transition temperature is about 0 °C, which is the lowest. Altogether, these enhancements provide avenues for future self-sensing carbon fiber composites in engineering structures.


Introduction
Carbon fiber reinforced polymer (CFRP) has been widely used in the field of advanced engineering technology as a functional material of high performance [1,2]. It combines excellent properties of lightweight, high-strength, good conductivity, and processibility [3][4][5].
Recently, significant attention has been paid to the conductivity of this composite [6]. The self-sensing function to physical parameters such as strain and temperature can be realized by establishing the corresponding relationship between environmental change factors (such as force field, strain field, temperature field, etc) and CFRP dynamic electrical response [7]. Since the piezoresistive effect of CFRP was first reported by Schulte in 1989 [8], the piezoresistive behavior and electric conduction mechanism of CFRP materials with different matrix material, layout, preparation methods, conductive fillers modification have been explored [9][10][11].
In exploring the piezoresistive effect of CFRP, the researchers found that the conductivity of CFRP was also significantly affected by ambient temperature [12][13][14]. The ambient temperature of some practical engineering structures, such as trains and airplanes, will fluctuate greatly in their service process. For satellites, space shuttles, rockets, etc serving in outer space, the temperature range may even exceed 100°C within one hour. CFRP composites will not be used as the structural health monitoring sensors of such structures unless such influence can be eliminated by technical means or precisely predict. It is found that the temperature-resistance effect is presented in two forms: positive temperature coefficient (PTC), that is, resistance rises as temperature rises, and Negative temperature coefficient (NTC), resistance decreases with temperature rises [15][16][17]. The main reason for this phenomenon is that the CFRP's thermal expansion coefficient changes in the process of temperature rising, which leads to internal thermal stress deformation and resistance change [18] anisotropy of the composite material, the results of this temperature-resistance effect are very discrete [19], and it is necessary to improve the discrete characteristics of the measured values by adding a certain percentage of conductive fillers to the resin matrix to improve the sensitivity to temperature changes [20].
The most extensive studies on this subject are mostly concerned with determining ways to increase the temperature sensitivity of composite materials by including a certain quantity of conductive particles or short conductive fibers. Increasing the thermo-resistivity of continuous conductive carbon fiber composites by improving the conductivity of the matrix has not been extensively studied. Therefore, it is of interest to explore the influence of temperature change on continuous carbon fiber composites and how to stabilize them by matrix modification. Our previous research results have established the relationship between the stress deformation and the resistance change rate of continuous carbon fiber composites, and confirmed that the modified CFRP material with a 6% content of carbon powder can improve the piezoresistive sensitivity [21]. Moreover, the existing research shows that the piezoresistive stability of composite modified by 3% nano-silica (in weight) achieves optimal [22].
The objective of this paper is to further investigate the effect of temperature on the resistivity of these CFRP bars. For this purpose, a systematic experimental investigation on the resistance response of CFRP bars with continuous carbon fiber and epoxy resin was prepared by adding different proportions of carbon powder and nano-silica under fluctuating temperature conditions.

Materials
The PAN-based carbon fibers (HF10-3K) used as the reinforcement were obtained from Hengshen company in Zhenjiang China. Their tensile strength is 3562 MPa, the carbon content is 95.6%, the elongation percentage is 1.5%, and the volume resistivity is 1.4953 × 10 −3 Ω·cm −1 . The epoxy resin component A with the density of 1.1 Figure 1. The temperature circulating process: (a) Preheating test, (b) Temperature circulating test from 30°C to 100°C, (c) Temperature circulating test from 20°C to −40°C ∼ 1.2 g cm −3 and Brookfield viscosity of 1000 ∼ 1300 cPs and components B with the density of 0.9 ∼ 1.0 g cm −3 and Brookfield viscosity of 10 ∼ 50 cPs under the temperature of 25°C were procured from Wuhao Co Ltd in Nanjing China. And the ratio of component A to component B by weight is 100:30. The nano-silica and carbon powder with the grain size of 2500 nm and 3.8 × 10 4 nm separately was purchased from Shenzhen Lanboshi Technology Co Ltd and Yancheng Xiangsheng Carbon Fiber Products Co Ltd respectively. The conductive silver paste (Ag/Ni-based Silicone adhesive) used for bonding electrodes with CFRP bar is produced by Shenzhen Xinwei Electronic Materials Co., Ltd.

Preparation of CFRP bars
The optimal sequence of modified resin matrix preparation was as follows: mixed the component B and the conductive fillers, stirred under the rotational speed range of 500-720 rpm for 15 min, then added component A, stirred under the speed of 500-720 rpm for 5 min. The total conductive filers mass fraction in the epoxy was 4%. And the ratio by weight of total nano-silica and carbon powder was set at 1:0, 1:0.2, 1:0.6, and 1:1 separately.
Considering the modified CFRP bars should meet both the mechanical and electrical properties, the preparation process was optimized several times according to the mechanical and electrical test results using ASTM D3039 and D4496-13. The detail of the CFRP bar preparation and curing process have been published in previous studies [23]. The mass proportion of reinforcement carbon fibers and total modified epoxy resin matrix was 65:35.

Process of temperature circulating test
As illustrated in figure 1(a), a preheating test is performed on all specimens using an oven with a ramp rate of 10°C /10 min from 30°C to 80°C. For two hours, the temperature is held at 80°C before naturally dropping to 30°C . Then, the specimens are subjected to three cycles of heating and cooling tests in the temperature range of 30°C to 100°C, and the heating rate is 10°C/10 min, the cooling rate is 5°C/10 min, as shown in figure 1(b). Finally, the specimens are placed in a freezing machine until the temperature falls from 30°C to 20°C.
Following the steps in figure 1(c), three more cycles of heating and cooling tests are conducted in the range of 20°C to −40°C following the procedure shown in figure 1(c). Electric resistance was also recorded at various temperatures during the lower temperature cycle.

Characterization techniques
The impact of high and low temperatures on the conductivity of electrode materials must be taken into account when choosing them due to the broad range of test temperatures (−40°C to 100°C). In this test, a metal electrode made of purple copper foil is employed, and conductive silver paste is utilized to bind the electrode to the test subject. Four equispaced copper electrodes are attached to the surface of the CFRP bars for resistance measurement. (as shown in figure 2). The test specimens' surface is uniformly covered with an epoxy resin coating of protection after the conductive silver paste has dried and hardened. The four-electrode approach, depicted in the figure 2(d), is used to measure the test in order to lessen the impact of the electrodes' contact resistance. Constant current is input through the outer two electrodes, and the output real-time voltage is continuously measured through the inner two electrodes using the self-made CFRP data collecting system during the test. The electrical resistance of the specimens can be calculated by dividing the voltage signals by the current [24]. The morphology of the modified carbon fiber bars was imaged by HITACHI SU8000 field emission scanning electron microscope (FE-SEM) operating at 10 kV.

Initial volume resistivity of CFRP bars
To explore the effect of mixed filler with the different mass ratios on the initial volume resistivity of CFRP bars at 20°C and the discrete characteristics of each set of data, the resistivity of each specimen was measured by the  four-electrode method. The four-probe method can eliminate the contact resistance between electrodes and the CFRP bars, which can boost the efficiency and reliability compared to the two-probe method [25]. The initial volume resistivity of CFRP bars was evaluated and is shown in figure 3. The group of CFRP-4 wt% CP: NS = 1:1 and CFRP-4 wt% CP: NS = 1:0.6 have the highest and lowest average initial volume resistivity r 0 measuring 0.105Ω·cm and 0.087Ω·cm, respectively. And the mean value of r 0 for group CFRP-4 wt% CP: NS = 1:0 is 0.091Ω·cm, and for CFRP-4 wt% CP: NS = 1:0.2 is 0.098Ω·cm.
Generally, adding insulating particles such as nano-silica to conductive materials will increase the average initial volume resistivity. But the test results show that r 0 of CFRP-4 wt% CP: NS = 1:0.6 decreases by 4.54%, 11.66%, and 17.68% respectively in comparison to CFRP-4 wt% CP: NS = 1:0, CFRP-4 wt% CP: NS = 1:0.2 and CFRP-4 wt% CP: NS = 1:1. It can be seen that adding a specific proportion of nano-silica can effectively reduce the initial volume resistivity of the CFRP bars. The standard deviations for the four sets are 0.016, 0.028, 0.008, and 0.007 respectively. A similar finding has also been observed by Wei et al [20].
Due to the large specific surface area of modified particles, strong van der Waals force among those particles makes it difficult to disperse uniformly in resin matrix, as shown in figure 4(a). The high viscosity of the resin matrix makes it challenging to scatter the particles, which is another major factor contributing to the agglomeration effect. It should be noted that the SEM observation is focused on the longitudinal middle layer of the carbon fiber. As can be seen in figure 4(b), in the CFRP-4 wt% CP: NS = 1:1 group, the distribution of modified particles adhered between CFRP fibers was relatively uniform. Experimental results have shown that the addition of nano-silica can reduce the agglomeration of particles in the matrix to a certain extent.
The better the dispersion of conductive fillers, the more conductive pathways between the fibers inside the CFRP bars, the lower the average resistivity, and the smaller the standard deviation. This is the reasonable explanation for the minimum average resistivity and standard deviation of CFRP-4 wt% CP: NS = 1:0.6 and CFRP-4 wt% CP: NS = 1:1 samples.

Volume resistivity variation in the preheating process
To verify the relationship between the volume resistivity of CFRP bars and temperature change during the monotonic heating process, a preheating test was carried out [26]. Another objective of this test was to eliminate the interference caused by uneven heating and residual temperature stress during the initial heating process [27]. Figure 5 Displays the variation of volume resistivity with temperature from 30°C to 80°C in four sets of CFRP bars. An identical trend can be observed for all specimens that volume resistivity increases linearly with raising temperature. This also follows the PTC effect reported in the literature [28]. Linear regression analysis between ρ and t shows that the correlation coefficients of the fitted curve are more than 0.86. This phenomenon can be explained that the difference in thermal expansion coefficient between the resin matrix and CFRP fiber results in the uncoordinated deformation of the two fibers, which blocks the conductive pathway between fibers and reduces the number of effective contact points [26]. Figure 5(e) is the average volume resistivity change value fitting curve of different proportion specimens during the preheating process. The resistivity change of the specimens, although the same in linearly increasing trend, vary with different growth ranges. And these differences are quantified by introducing a temperature sensitivity coefficient(s T ), as shown in equation (1).
Where the r , T r 0 are the volume resistivity at a given temperature and the initial volume resistivity respectively. The Maximum resistivity occurs at 80°C. Hence the average temperature sensitivity can be calculated for the three sets of specimens.
Based on figure 5(e), the average s T for the four sets are 42.5%, 20.2%, 78.8%, and 51.6%. The conclusion can be drawn that adding carbon powder or nano-silica into epoxy resin can increase the average temperature sensitivity. This is caused by the thermal expansion effect which can decrease the conductive pathways. The CP: NS = 1:0.2 and CP: NS = 1:0.6 have the lowest and highest temperature sensitivity, decreased by 52.5% and increased by 85.4% compared with CP: NS = 1:0, respectively. Compared with the temperature sensitivity of the group CP: NS = 1:1 and CP: NS = 1:0, the difference between the two groups is smaller than that of the other groups. This is such that adding nano-silica will result in a greater reduction in the number of conducting routes owing to thermal expansion than adding carbon powder will. The test results show that the reasonable mass ratio can effectively improve the temperature sensitivity of CFRP bars. When the mass ratio of carbon powder to nano-silica is 1:0.6, the CFRP bars are most sensitive to temperature. Figure 6 shows the comparison of the initial volume resistivity at 30°C and the volume resistivity at the end of the preheating cycle at 30°C. It can be seen from figure 6 that the volume resistivity of all samples increased after the preheating test. Besides, the volume resistivity after preheating and cooling is further increased than that at 80°C in figure 3, which is consistent with the phenomenon described in the literature [29]. The standard deviations of CFRP samples with the mass ratio of 1: 0, 1: 0.2, 1: 0.6, and 1: 1 are 0.016, 0.028, 0.008, and 0.007 The fractional change in volume resistivity before and after testing of 4 sets of the specimen are 24.8%, 20.7%, 91.0% and 27.4%, corresponding to CP: NS = 1: 0, 1: 0.2, 1: 0.6, and 1: 1, respectively. From the above results, compared with the other three groups, the change in volume resistivity of CP: NS = 1: 0.6 is particularly significant before and after the preheating cycle test.
The mechanism behind this could be polymer chain oxidation, residual stress from thermal expansion, and conductive path rearrangement [11]. Another possible reason is that a specific amount of nano-Silicon is more likely to cause small spatial changes among the carbon fibers and conductive particles within the CFRP tendons, leading to the rearrangement of the conductive network within the CFRP bars and significant changes in the volume resistivity.

Volume resistivity variation in the temperature circulating process
The temperature circulating test was carried out after the preheating test, and it consists of two independent stages. Three heating and cooling cycle tests were conducted within the temperature range of 30°C to 100°C first, followed by a three-cycle trial from 20°C to −40°C. For regularity, the heating and cooling data in figures 7(a) to (d) were represented by the average values of three heating processes or cooling processes, respectively. As shown in figures 7(b) to (d), the volume resistivity of each specimen are highly consistent with temperature, that is, first decreases and then increases with the increase of temperature. The volume resistivity The regression analysis of the data in figure 7 shows that the correlation coefficients of the specimens are between 0.83 and 0.95, which indicates the fitting results meet well with the experimental data. In addition, figure 7(e) shows the variation of the average volume resistivity of the four sets of samples with temperature under three temperature circulating. The average volume resistivity for a straightforward CFRP specimen is not sensitive to temperature changes. Contrarily, CFRP specimens with mixed fillers added are more sensitive to temperature changes. Based on formula (1) and the PTC sections of figure 7(e), the average temperature sensitivity for CP: NS = 1:0 is 7.5%, for CP: NS = 1:0.2 is 23.30%, for CP: NS = 1:0.6 is 3.87%, and for CP: NS = 1:1 is 30.29%.
Combined with the temperature sensitivity coefficient and the temperature range of the PTC effect, it can be seen that the temperature corresponding to the PTC effect of CP: NS = 1:0, CP: NS = 1:0.2, and CP: NS = 1:0.6 is basically in accordance: the transition temperature ranges from 30 to 60°C. In the above three groups, the temperature sensitivity coefficient of CP: NS = 1:0.2 is the largest, which is 23.30%. But for CP: NS = 1:1 group, the transition temperature of the PTC effect is about 0°C, the lowest in all four test groups. Therefore, the PTC temperature range of CP: NS = 1:1 is the widest, and the temperature sensitivity coefficient is 30.29%. Considering practical engineering applications, when the working temperature range of CFRP specimen is 0°C ∼ 80°C, the optimum mass ratio of carbon powder and nano-silica is 1:1; and the working temperature range is 40 to 80°C, the optimum mass ratio is 1:0.2.
The possible reasons for the increase of volume resistivity of modified CFRP bars are: (1) the decrease of effective conductive contact point and pathway caused by the thermal expansion effect; (2) the tensile strain of some fibers within the CFRP bars caused by residual stress of thermal expansion; (3) the oxidation of polymer chain with the formation of groups like -OH, -CHO, -COOH caused by the long period of cooling and heating cycling process [30].

Analysis of test results and thermal-electrical model
The above results show that the CFRP materials exhibit both NTC and PTC effects in the temperature range of −40°C to 80°C when carbon powder and nano-silica are mixed into the matrix. For CP: NS = 1: 0 and 1: 0.2 specimens, the PTC effect temperature ranges from 40°C to 80°C, for 1: 0.6 specimens from 45°C to 80°C, and for 1: 1 specimens from 0°C to 80°C. We previously presented a coupled thermal-electrical model to characterize the thermo-resistivity effect of CFRP materials as below [28].