Mechanics of Materials
Change in Electrochemical Impedance Spectrum of Al/CFRP Adhesive Lap Joint by High Temperature, High Humidity-Cold Cycle Test
2022 年 63 巻 8 号 p. 1127-1132
詳細
2022 年 63 巻 8 号 p. 1127-1132
Weight reduction in automobiles is an effective way to reduce carbon dioxide emissions. To reduce the weight, adhesive joints made from light materials are expected to be widely used. However, the deterioration evaluation of adhesion in joints exposed to corrosive environments, such as the salt-damaged area, is performed using destructive evaluation methods such as tensile tests.
In this study, to develop a non-destructive method for detecting the deterioration of aluminum/carbon fiber-reinforced plastic lap joints by adhesives, we investigated the change in the electric properties of the joints before and after corrosion tests through electrochemical impedance spectroscopy. It was observed that the impedance and frequency dependence of the dielectric loss tangent of the joints were changed by the corrosion test.
This Paper was Originally Published in Japanese in Zairyo-to-Kankyo 70 (2021) 365–369.
The realization of a low-carbon society is required to prevent global warming, protect the environment, and save resources. Weight reduction of automobiles can contribute to a low-carbon society by suppressing carbon dioxide emissions during driving. Utilization of multi-material parts composed of various lightweight materials is an effective way to reduce the weight of automobiles. Especially, metal/resin adhesive bonding is necessary in the manufacture of multi-material parts and has a profound effect on weight reduction. From this point of view, metal/resin structural materials such as carbon fiber-reinforced plastic (CFRP) and aluminum are expected to be widely used.1) To ensure the long-term reliability of metal/resin adhesive bonding under the environmental conditions that automobiles are exposed, the degradation of a wide variety of new material combinations in a corrosive environment must be quantitatively evaluated.
Resins generally age owing to water absorption from the environment2) that results in deterioration of the mechanical properties3) and a decrease in impedance4,5) (deterioration of anticorrosion effect on metals). In addition, the impedance of epoxy resins immersed in saltwater changes with temperature.6) The infiltration of water and ions from saltwater affects the impedance and dielectric properties of the resins, and the changes in the impedance and dielectric properties are related to the deterioration of strength and anticorrosion performance. In contrast, automobiles are used all over the world and are exposed to various environmental conditions such as temperature, humidity, and salt content; thus, the properties of adhesives, which are resin materials, and the adhesive joints are expected to change.
Electrochemical impedance spectroscopy (EIS) and the current interrupter method are widely used to evaluate the corrosion resistance of coated steels.7,8) Meanwhile, dielectric analyses by applying an AC voltage on substances,9) similar to the EIS method, are also performed as a non-destructive method for investigating the internal structures of solids or liquids, as well as the interfaces between them. Therefore, the EIS data reflect the states of the coating film and interface between the coating film and substrate, and the frequency dependence of the impedance and dielectric loss tangent (tan δ) obtained through EIS lead to an understanding of the inhomogeneous structure in materials (mixed phases of water, ions, and resins).9) The Nyquist diagrams obtained from EIS can show the electrical time constants, chemical reactions, and adsorption phenomena associated with the substances and concerned interfaces.10)
Although EIS is considered useful for assessing the property changes of materials, there are few studies that use EIS to evaluate the changes in the impedance and dielectric properties of adhesive joints using various adhesives exposed to an environment that can deteriorate the physical properties. Examining the degradation of adhesive joints in severe corrosive environments such as cold regions is particularly useful, as the corrosion of coated steel plates was proved to be promoted in such an environment in a previous study.11) Therefore, we investigated the changes in the impedance and dielectric loss tangent of CFRP and metal (aluminum) joints by various adhesives, which are indispensable for the multi-materialization of automobiles, in corrosive environments such as cold regions through EIS.
The joint shown in Fig. 1 was used in the corrosion tests. CFRP plates were joined to A6061 aluminum plates using an adhesive. The CFRP plates were made by impregnating a 3 K plain weave of continuous carbon fiber with a bisphenol A-type epoxy resin. The diameter and the volume fraction of the carbon fiber were approximately 7 µm and 52–53%, respectively. The lengths, widths, and thicknesses of both the aluminum alloy and CFRP plates were 50, 20, and 2 mm, respectively. Epoxy, urethane, and silicone-based adhesives were used for the joints. An electrical insulation spacer with length, width, and thickness of 20, 15, and 1 mm, respectively, was inserted between the aluminum alloy and CFRP plates, and one of the adhesives was applied within a range of 10 mm from the end of the aluminum alloy plate. Therefore, the thickness of the adhesive layer was 1 mm. After the two plates were overlapped, the adhesive protruding from the end of the aluminum alloy plate was removed and the joint was stored at room temperature for 1 d or more such that the adhesive was sufficiently cured.
Structure of lap joint.
The temperature and humidity control patterns of the corrosion test are shown in Fig. 2. The corrosion test was repeated between 80°C, and −10°C for 10 cycles (10 Cy) daily. The humidity was set at 95% at 80°C and was not controlled at −10°C. The tests were performed in constant low temperature/humidity chamber (THE051FA) made by Toyo Engineering Works, Ltd. Similar to a previous report,11) a mixture of 98 mass% NaCl–1 mass% CaCl2–1 mass% MgCl2 (40 mg·cm−2) that imitated the snow-melting salt used in cold regions was applied to the joint.
Temperature and humidity control pattern of corrosion test.
EIS measurements were performed as soon as possible after the corrosion test (within 6 h). Figure 3 shows a schematic of the EIS measurements. In an actual environment, the adhesive is sandwiched between objects (e.g., CFRP and aluminum alloy plates) at the joint. Therefore, EIS measurements were also performed using the joined objects. A ModuLab® (Solartron analytical) was used to perform the measurements. The measurement region was exposed to 1 M NaCl for approximately 10 to 30 min before the measurement. The measurement region was approximately 1 cm2 on the CFRP side of the joint. The reference, counter, and working electrodes were a silver/silver chloride electrode (in a saturated KCl solution), platinum wire, and a joint, respectively. An iron needle was inserted into the aluminum alloy plate of the joint as a lead wire to reduce the contact resistance to an insignificant value and connected to the ModuLab®. The measurement frequency range was 10−2–105 Hz (10 points/decade), and an AC voltage of 100 mV amplitude was applied at open circuit potential.
EIS measurement.
EIS measurements were performed on CFRP for comparison with the joints. The iron needle shown in the previous section was used to connect the CFRP plates. In the measurement of the joints, the input voltage is exerted through an aluminum alloy plate that has a low electrical resistance to reduce the voltage drop in the measurement. However, in this experiment, the input voltage was exerted on the measurement section through the CFRP plate with a high electrical resistance.
Figure 4 shows the frequency dependence of the AC electric properties after the corrosion test with CFRP. The complex plane of the impedance Z (Nyquist diagram), frequency dependence of |Z| (the absolute value of the impedance), and the dielectric loss tangent tan δ are shown in Fig. 4(a), (b), and (c), respectively.
Electrochemical spectra of CFRP (a) Nyquist plot, (b) and (c) Frequency dependence of |Z| and tan δ.
The Nyquist diagram shown in Fig. 4(a) is straight without an arc. In other words, the dielectricity is dominant; hence, the time constant is considered so large that no arc can be observed. Figure 4(b) shows that at a low frequency, |Z| slightly decreased after the corrosion test. In Fig. 4(c), a tan δ peak was observed in the range of 100–101 Hz. This indicates that permittivity and conductivity changed suddenly at this frequency range, and these changes are suggested to be caused by the adsorption of water by the carbon fibers in the epoxy resin.9,12) Therefore, it is suggested that water penetrated the CFRP in the corrosion test, and impedance decreased. However, the impedance of the CFRP did not change as significantly as that of the joint, which will be described later. Hence, the changes in the impedance and tan δ (above 102 Hz) are associated with the changes in the interface between the adhesive and joint materials, or the deterioration of the adhesive.
Hereafter, the results of the adhesive joints are described. Figure 5 shows the Nyquist diagram of the Al/CFRP joint with the epoxy adhesive. The diagram in the vicinity of the origin in Fig. 5(a) is enlarged and shown in Fig. 5(b) and (c). Figure 5(a) and (b) show that the impedance was significantly decreased by the corrosion test. The Nyquist diagram changed to an elliptical arc shape from the straight line, which overlapped the imaginary axis. This indicates that the equivalent circuit has changed from a simple capacitor to a parallel circuit of a constant phase element (CPE) and a resistor. However, the time constant of the parallel circuit could not be determined within the measurement range. Therefore, the diagram shown in Fig. 5(b) might illustrate the result of a phenomenon with substance diffusion or a small time constant. From the further enlarged diagram shown in Fig. 5(c), an arc was observed after the corrosion test. The frequency corresponding to the vertex of the arc was 31623 Hz and the time constant was determined to be approximately 5 µs. According to Sato et al.,13) the time constants of the corroded metal and resin surfaces were approximately 10 s and 1 ms, respectively. Thus, the diagrams in Fig. 5 show a phenomenon different from those reported because of the different time constants of 5 µs.
Nyquist diagrams of joints with epoxy adhesive (a) 0–30 GΩ, (b) 0–0.2 GΩ, and (c) 0–0.4 MΩ.
Figure 6 shows the frequency dependences of |Z| and tan δ of the joint with the epoxy adhesive. The negative sign of tan δ indicates the direction of the phase of the output signal, that is, a phase lag with respect to the input signal. Hereinafter, the absolute value of tan δ is expressed as tan δ. According to Fig. 6(a), |Z| in the frequency range below 103 Hz decreased by approximately three orders in the corrosion test, indicating that the properties of the joint changed significantly. A peak of tan δ was observed at approximately 104 Hz after the corrosion test, as shown in Fig. 6(b). It is known that dielectric relaxation occurs in the range of 103–105 Hz in a thin resin film in water or an emulsion consisting of water and organic matters.9) When the above dielectric relaxation occurs, tan δ changes significantly, it is suggested that water infiltrated into and was absorbed on the adhesive side of the interface between the adhesive and metal. In other words, the frequency of the peak and the magnitude of tan δ in the high-frequency range indicate the absorption of water at the interface between the adhesive and metal.
Frequency dependence of joints with epoxy adhesive (a) |Z| and (b) tan δ.
Figure 7 shows the Nyquist diagram of the joint with the urethane adhesive. The impedance of the joint with the urethane adhesive before the corrosion test was less than that of the joint with epoxy adhesives after the corrosion test. The spectrum parallel to the imaginary axis, such as an ideal insulator (dielectric), could not be obtained, and the plot points in the low-frequency range approached the real axis. It seems that hard-to-react charges at high frequencies (ions and polar groups in the resin) moved in the initial stage of the corrosion test. The Nyquist diagram after 10 cycles is shown in Fig. 7(b). The diagram shows an elliptical arc that can be represented by a parallel circuit of resistance and CPE, and the frequency at the vertex of the elliptical arc was approximately 40 Hz. The corresponding time constant was approximately 4 ms. This is approximately equal to the time constant of 1 ms for the acrylic coating film reported by Sato et al.;13) thus, the time constant can be considered to have been obtained from urethane-based adhesives, which are made up of similar organic matter as acrylic.
Nyquist diagrams of joints with urethane adhesive (a) 0–10 GΩ and (b) 0–0.5 MΩ.
The frequency dependencies of |Z| and tan δ of the joint with the urethane adhesive are shown in Fig. 8. The value of |Z| at 10−1 Hz was as large as 1010 Ω before the corrosion test, but it decreased to about 105 Ω after 10 cycles, as shown in Fig. 8(a). The reason why |Z| after 10 days was smaller than |Z| shown in Fig. 4 is due to the anisotropy of electrical resistance of CFRP in the thickness direction and longitudinal direction (due to the fiber orientation), and other reason is considered the difference in the length of CFRP to the electrical connection point (when measuring EIS of CFRP, the distance from the measurement point to the electrical connection point is about 30 mm, on the other hand it is about 2 mm when the specimen containing the adhesive is measured) so that the impedance of CFRP in Fig. 4 is expected to be about 1/15 times larger in the measurement of the bonded range, because the bonded range is measured through CFRP.
Frequency dependence of joints with urethane adhesive (a) |Z|, (b) tan δ (10−2–105 Hz), and (c) tan δ (101–105 Hz).
The value of tan δ also increased, particularly below 100 Hz where it increased rapidly, as shown in Fig. 8(b), indicating that the electricity had become easier to flow. This change was not observed in the case of the epoxy and silicone adhesives described later; hence, it is considered as a phenomenon peculiar to urethane adhesives or significantly deteriorated materials. Further studies are required to clarify this. Figure 8(c) shows the enlargement of Fig. 8(b) in the range of 101–105 Hz. A tan δ peak was observed at approximately 104 Hz after 10 cycles. The frequency corresponding to the peak is approximately equal to that in the case of the epoxy adhesive after the corrosion test; thus, it probably resulted from the infiltration of water into the adhesive interface.
Figure 9 shows the Nyquist diagram of the joint with the silicone adhesive. The region near the origin of Fig. 9(a) is enlarged and shown in Fig. 9(b). Before the corrosion test, the trajectory was a curve comprised of a semicircle connected with a 45° tilted straight line at 0.79 Hz or less. In general, such a trajectory is expressed by an equivalent circuit in which a series circuit of a Warburg impedance representing the diffusion of chemical species and a resistance is parallel to the CPE.10) After the corrosion test, the trajectory of the impedance became closer to the origin of the Nyquist diagram, and the inclination angle was approximately 45°. According to Maruyama,14) silicone rubber is more permeable to gases than other types of rubbers such as natural rubber, and silicone adhesives are thought to have similar properties as silicone rubber, for example, both silicone rubber and silicone adhesive remain elastic even after curing. Therefore, Warburg impedance in the Nyquist diagram is presumed to correspond to mass transfer phenomena such as the dissolution of oxygen in the adhesive arising from the permeability of silicon adhesive.
Nyquist diagrams of joints with silicone adhesive (a) 0–20 and (b) 0–0.15 GΩ.
The frequency dependences of |Z| and tan δ of the joint with silicone-based adhesives are shown in Fig. 10. The value of |Z| decreased after the corrosion test for all frequencies, as observed in Fig. 10(a), indicating that the electrical properties of the joint were changed significantly by the corrosion test. In addition, the peak of tan δ appeared at approximately 105 Hz after the corrosion test, and the frequency corresponding to the peak was approximately equal to those in Figs. 6(b) and 8(c). Therefore, the appearance of the peak is probably a common phenomenon of adhesive structures by resins, for instance, the infiltration of water into the adhesive or the adhesive interface during the corrosion test.
Frequency dependence of joints with silicone adhesive (a) |Z| and (b) tan δ.
From the described results, |Z| decreased and a tan δ peak was observed in the range of 104–105 Hz after the corrosion test for all the adhesive joints. These phenomena are considered to be related to the infiltration of water into the adhesive and adhesive interface during the corrosion test. Therefore, the deterioration of the adhesive joints can be evaluated by monitoring the frequency dependence of AC electrical properties.
Environmental cycle corrosion tests on aluminum/adhesive/CFRP joints were performed under repeated high temperature, high humidity, and freezing conditions in the presence of salts, and the changes in the frequency dependence of the AC electrical properties were investigated. The following results were obtained:
