Changes in Admittance and Internal Structure of Coating Films by Environmental Testing

Abstract

The impedance of the paint film varied depending on the environmental test. It is considered that the impedance spectrum of the coating films changed by the effects of water content in the paint film, polymer structure, and elements that penetrated the paint film. However, no clear relationship could be found between the impedance spectrum of the coating film and its internal structure. To understand the change in the impedance spectrum, we performed D-SIMS and AFM-IR analyses on the paint films after the environmental test. The results suggest that the admittance spectra may reflect the state of infiltrated moisture and elements into the coating films.

 

This Paper was Originally Published in Japanese in Zairyo-to-Kankyo 72 (2023) 243–248. “the state of infracted moisture” in the abstract has been corrected as “the state of infiltrated moisture”.

Cross-sectional AFM topography and chemical mappings of coating films after environmental tests.

1. Introduction

Resin paints are widely used to protect automotive steel panels from corrosion. Conventionally, environmental tests (e.g., cycle corrosion test, salt spray test) that simulate corrosive environments have been used to evaluate the corrosion resistance of coated steel panels, as done for steel panels. In these environmental tests, it takes a long time for the coating film to deteriorate or age until rust can be detected. The evaluation criteria for deterioration and aging are based on the measurement of the rusting area and incidence of blisters (determined from observation) [1]. Exposure to water, heat, and ultraviolet rays deteriorates resins, changing their mechanical and electrical properties [2, 3]. The electrical properties, which are directly related to polarization resistance [4], dominate the corrosion rate and thus their changes are considered to be correlated with the corrosion protection performance degradation of coating films. We previously reported that the AC electrical properties of resins, including coating films, measured by electrochemical impedance spectroscopy (EIS) change after relatively short-time exposure tests [5, 6]. However, the relationship between the changes in the EIS spectra and the resin internal state was not clarified. An understanding of this relationship would allow the nondestructive measurement of the coating film internal state.

Fourier transform infrared spectroscopy (FT-IR) and electron probe micro analyzer (EPMA) are used to identify the structural changes of resins [7]. However, FT-IR has a low spatial resolution of only a few micrometers [7], making it difficult to detect initial changes in a narrow analysis area, such as the cross-section of coating films. In our preliminary study, EPMA did not detect minimal changes in the amounts of elements in the coating film in the initial stage of a corrosion test. Therefore, in this study, we use atomic force microscope infrared spectroscopy (AFM-IR), which has a high spatial resolution, and dynamic secondary ion mass spectroscopy (D-SIMS), which is expected to have a concentration resolution or sensitivity on the order of parts per billion to clarify the states in coating films that showed changes in EIS spectra [8].

2. Experimental Procedure

2.1 Environmental tests

Environmental tests were conducted on coated steel panels to change the internal state of the coating film. The coated steel panels were the same as those used in automobile manufacturing. The specimen panels were cold-rolled steel sheets coated with epoxy-based cationic electrodeposition paint film (thickness: 15 µm) after phosphate conversion treatment. Before the tests, the thickness of the coating film was confirmed to be 15–18 µm using an electromagnetic paint thickness gauge. The dimensions of the specimens were approximately 20 mm × 70 mm × 1 mm (thickness). Two types of test with different environmental conditions were conducted. In Test 1, which used the same conditions as those used in a previous study [5], a coated steel panel was coated (area: 40 mg·cm⁻²) with a simulated snow-melting salt (98 wt%NaCl-1 wt%CaCl₂-1 wt%MgCl₂), placed in a tray, and maintained at 80°C and 95% relative humidity (RH) for 10 days. In this test, the simulated snow-melting salt was deliquesced by the moisture in the environment and a water film formed on the coating film. Test 1 was thus an immersion test in a solution with a high salt concentration. In Test 2, after the specimen panels were covered with the simulated snow-melting salt using the same method as that used for Test 1, the environmental cycling test shown in Fig. 1 was conducted. The environment (80°C, 95% RH) was the same as that in Test 1. The water on the coated steel sheet evaporated during the transition to −10°C. During the holding period at −10°C, the water film on the surface became sorbet-like. The above cycle was repeated 10 times.

Fig. 1

Temperature and humidity control pattern of corrosion test.

2.2 EIS measurement

The EIS spectra of the coated steel plates (as the working electrode) before and after the exposure tests were measured using a previously reported method [5]. Figure 2 shows a schematic diagram of the measurement. The EIS measurements were performed with Modulab^® (Solartron Analytical).

Fig. 2

Schematic diagram of impedance measurement system.

A measurement cell with a circular hole was pressed against the coated steel panel to bring the coating film surface in contact with the 1 M NaCl solution in the cell. The measurement was performed at room temperature after it was confirmed that the change in the spontaneous potential had become less than 10 mV/min (i.e., the coating film surface had approached equilibrium with the solution). The required contact time prior to the measurement was usually about 50 minutes or longer depending on the condition of the coating film. The area of the circular hole was approximately 1 cm². The measurement was performed using the three-electrode method with a platinum wire counter electrode and a saturated Ag/AgCl reference electrode. A sinusoidal AC signal with a frequency range of 10⁻² to 10⁶ Hz and a potential amplitude of 100 mV around the open-circuit potential was applied to the measurement. Ten points per decade from high to low frequencies were measured.

The equivalent circuit of an ideal coating film without defects in the EIS measurement is shown in Fig. 3 [9]. The same electrochemical cell and electrolyte (1 M NaCl solution) were used throughout this study and thus the solution resistance R_s was assumed to be the same for all measurements. In order to measure R_s, a platinum plate instead of the coated steel panel (working electrode) in Fig. 2 was measured. When the impedance of the measurement system is small, the spectrum has to be analyzed assuming an inductive equivalent circuit for the effects of wiring and equipment [10]. In this case, the solution resistance can be considered to be approximated by a value close to the real impedance axis on the high-frequency side since the impedance caused by the inductance is small in this case. The Nyquist diagram in the vicinity of the origin when Pt was used as the working electrode is shown in Fig. 4. From the results, R_s is estimated to be around 0.3 to 0.5 Ω. These values can be used as the impedance of coating films since the values of coating films were 10³ to 10⁴ Ω even at 10⁶ Hz. The coating film in Fig. 3, if considered as a single electrical element, has impedance Z_c. An equivalent circuit that includes coating defects has been proposed for the corrosion under the coating film [11]. However, in the present study, it was difficult to fit the impedance spectrum with any equivalent circuit that includes defects, and no obvious corrosion was observed under the coating film. Accordingly, we considered the coating film to be a single electrical element with impedance Z_c. In this way, the measured impedance spectrum of the coating film can be considered to reflect the internal structure of the film through the corresponding frequency dependence of impedance Z_c. Therefore, we converted the measured impedance into admittance, which made it easy to consider the mobility of electric charges caused by moisture and ions in the coating film.

Fig. 3

Simple equivalent circuit assumed for coating film. R_s: resistance of solution, R_c: resistance of coating, C_c: capacitance of coating.

Fig. 4

Nyquist plot of Pt near origin.

Given an impedance Z_c for the coating film, the admittance Y_c is 1/Z_c. Assuming the equivalent circuit for the coating film in Fig. 3, the following equation can be obtained.

  

\begin{equation} Y_{c} = Y_{c}{}' + jY_{c}{}'' = 1/R_{c} + j\omega C_{c} = G_{a} + j\omega C_{a} \end{equation} (1)

Where R_c and C_c are the electrical resistance and capacitance for the coating film, respectively. Setting G_a = Y_c′ and C_a = Y_c′′/ω, G_a and C_a are the electrical conductivity and capacitance, respectively. They can be related to the state of electrical charges in the coating film and/or at the film/metal interface. It should be noted that in this calculation, G_a and C_a are the apparent values obtained for the whole coating film since the equivalent circuit in Fig. 3 is assumed, even though the components of Z_c are expected vary.

2.3 Dynamic secondary ion mass spectroscopy

An elemental analysis was performed using D-SIMS in the depth direction from the surface of the coated steel panel to the steel plate before and after the environmental tests. A quadrupole secondary ion mass spectrometer (PHI ADEPT1010, ULVAC Phi Inc.) was used for the analysis. Table 1 shows the elements analyzed and analysis conditions.

Table 1 Typical conditions of D-SIMS measurement.

2.4 Atomic force microscope infrared spectroscopy

The cross-sections of the coating films before and after the environmental tests were analyzed using AFM-IR. The assay samples were prepared as follows. Scratches were made with a cutter on the coated steel panel before and after the environmental tests. The scratches were widened by bending the steel panel and the coating film was mechanically peeled off from the scratches to obtain a lamina that was then embedded in resin to produce a cross-section of the coating film. AFM-IR analysis was performed using a nanoscale infrared spectrometer (Anasys nanoIR2, Bruker). The spatial resolution of this analysis (50–100 nm) is not affected by the wavelength of the infrared rays [12, 13]. In addition, since the spectra are equivalent to FT-IR spectra, the knowledge obtained from conventional IR analyses can be applied [12, 13].

3. Results and Discussions

3.1 Admittance analyses of coating films before and after environmental tests

Figures 5(a) and 5(b) show the frequency dependence of G_a and C_a obtained from the admittance analyses of the coating film before and after the environmental tests. G_a and C_a are considered to correspond to the mobile (conductive) and immobile (dielectric) charges, respectively, during a cycle of the potential amplitude [14]. For example, the charges (e.g., polar molecules) that move fast and travel short distances contribute to G_a at high frequencies but contribute to C_a at low frequencies. On the other hand, substances such as salt ions do not contribute to G_a and C_a at high frequencies. That is to say, these charges may contribute to G_a and C_a at each frequency depending on the ease and range of their movement. Regardless of the environmental test conditions, both G_a and C_a were larger than those for the pre-test coating film at all frequencies, indicating an increase in electric charges in the coating film during the tests. Based on the environmental test conditions, these charges are thought to be ions derived from the simulated snow-melting salt. In addition, penetration of water into the coating film is necessary for the ions to move.

Fig. 5

Frequency dependence of paint steel with (a) G_a and (b) C_a.

In view of the above, the differences between the spectra of G_a and C_a after Test 1 and Test 2 reflect the states of the electric charges that had penetrated into the film. Both G_a and C_a were larger in Test 2 in the range of 10⁻¹ to 10³ Hz, indicating that a large number of charges related to conductivity and dielectricity remained in the coating film in this frequency range. Moreover, the conductivity (G_a) at 10⁻² Hz changed from 10⁻¹¹ S to about 10⁻⁷ S after both Test 1 and Test 2. As the DC electrical resistance is an indicator of the corrosion resistance of coating films, and G_a at low frequencies (e.g., 10⁻² Hz) is close to the reciprocal of DC resistance, the numbers of charges that contributed to conductivity were same for the two tests. Therefore, the films showed the same degradation of corrosion resistance after the two environmental tests, but in the film after Test 2, many charges remained (i.e., charges that could not continue to move at low frequencies such as 10⁻² Hz), showing that the corrosion resistance of the film still did not deteriorate. Even though the coating film after Test 2 had a shorter exposure time to high temperature and high humidity than that after Test 1, it had a similar G_a at 10⁻² Hz as that after Test 1 and higher G_a and C_a at higher frequencies (above 100 Hz) than those after Test 1. These results suggest that there were more charges in the coating film after Test 2 and that the degradation mechanism in Test 2 was different from that in Test 1.

3.2 Dynamic secondary ion mass spectroscopy analyses

Figure 6 shows the results of the elemental analysis from the surface of the coated steel panel to the film/steel interface before and after the environmental tests. The concentrations of all elements except Mg increased in the coating films after the tests.

Fig. 6

Elements distributions in resin films.

Elements derived from cations of sodium, calcium and magnesium segregated at the film/steel interface in the initial state of the coated panel. These elements are considered to be the dry contaminants left by the dewatering/drying after the steel panel was rinsed before the electrodeposition of the epoxy resin film [15].

The secondary ion intensities of Na and Ca were about 10 times higher than those before the environmental tests, indicating that these salts had penetrated into the coating film during the environmental tests. After Test 1, which was conducted at high temperature and high humidity, Na and Ca were widely distributed in the coating film. Na clearly increased in the coating film even after Test 2. Thus, the elements of the salts penetrated into the coating films during the environment tests. The distribution of the elements that initially segregated at the resin/steel interface became smoother after Test 1 but remained almost unchanged after Test 2. In Test 1, there was no exposure to −10°C and thus the exposure to high temperature (80°C) was longer. The elements that initially segregated at the interface were thus homogenized by diffusion.

In both the tested films, the intrusion of Cl (corresponding to anions) was small compared to that of the elements corresponding to cations, such as Na and Ca. This indicates that the cations in the salt water reacted with the coating film through ion-exchange reactions. It is thought that some terminal groups of the resin are ion-exchangeable and that the ions in the aqueous solution replace the polymer’s hydrogen ions [16]. Therefore, not all the ions that penetrated into the film acted as ions responding to the potential amplitude. The reason that Ca in the coating film increased only in Test 1 may be that divalent metal ions are adsorbed more onto the resin as temperature rises when ion exchange occurs with the resin, as shown by Sasaki et al. [17].

The increase in only cations constituting the salt in the coating film after Test 1 suggests that most of the cations that penetrated into the coating film were probably bound to the resin by ion exchange. On the other hand, Cl also increased after Test 2, indicating that counter-cations likely existed without adsorption in the coating film. It is expected that more cations unbound to the resin. This does not contradict the fact that the coating film after Test 2 had larger G_a and C_a in the range of 10⁻¹ to 10³ Hz than that after Test 1.

Although it is unclear why Na and Cl were also observed in the steel region in the specimen after Test 1, it may be for the following reasons. When the coating film was peeled off from the steel panel to prepare samples for AFM-IR analysis (described in the next section), the film after Test 1 peeled off significantly when the steel plate was bent and was easier to peel than new films or that after Test 2. This indicates the possibility of the formation of a salt-water film near the interface or the alteration of the surface of the steel in Test 1. Hence, the analysis of the film/steel interface after Test 1 may not be accurate due to the influences of i) a water film between the film and the steel, and ii) alteration of the steel or the chemical conversion film by the salt water. Therefore, the equivalent circuit in Fig. 3 may have changed and the measured admittance (G_a, C_a) was not only the admittance of the film, but also that of the measurement system, containing information about the film/steel interface.

3.3 Atomic force microscope infrared spectroscopy analyses

AFM-IR analyses were performed on the cross-sections of the coating films before and after the environmental tests. Figure 7 shows an example of a spectrum obtained at each analysis point. The spectra were measured for a cross-section with a distance of approximately 1 µm from the surface of the coating film before any environmental test. These spectra were measured over the whole cross-section. In this study, 3400 cm⁻¹ was selected as the representative position of the broad peak between 3000 and 3600 cm⁻¹ (corresponding to moisture). The peak at 1724 cm⁻¹ (corresponding to esters) was chosen as the resin component. Although the fingerprint region could not be linked to a specific bond, 1118 cm⁻¹, which showed changes in the environmental tests, was also selected together. The absorbance at 1118 and 1724 cm⁻¹ on the low-wavenumber side was normalized by that at 1508 cm⁻¹ (corresponding to the aromatic ring that constitutes the main chain of the epoxy resin), and the absorbance at 3400 cm⁻¹ on the high-wavenumber side was normalized by that at 2956 cm⁻¹ (corresponding to -CH).

Fig. 7

AFM-IR spectrum initial resin film at 1 µm from surface of paint steel.

AFM-IR chemical mapping images of the film cross-section are shown in Fig. 8. In the figure, the region outside of the surface and the interface is the embedding resin used to make the thin section. The thickness of the film measured at the time of painting was 15–18 µm, but the AFM topography results show a thickness of 18–20 µm. It is thought that this difference is due to stretching during the mechanical removal of the coating film from the steel plate after the conversion treatment and the inclination of the analysis sample during the thin section preparation.

Fig. 8

AFM topography and chemical mappings.

The absorbance that originated from resin bonding at 1118 and 1724 cm⁻¹ was compared before and after the environmental tests. After Test 2, areas with a lower absorbance at 1724 cm⁻¹ (shown in blue in Fig. 8), which was attributed to esters, were scattered throughout the film cross-section and the absorbance at 1118 cm⁻¹ in the coating film was lower than that of the other films. This result suggests that the chemical bonds of the resin were altered by Test 2. The absorption at 3400 cm⁻¹ (corresponding to water) increases in the following order: untested < Test 2 < Test 1. The water contained in the films thus increased in this order. A comparison of the water absorption (Fig. 8) with the measured G_a and C_a (Fig. 5) indicates that the absorbed water in the films contributed to the increase in G_a and C_a at frequencies lower than 10³–10⁴ Hz. However, although the water absorption amount in Test 2 was lower than that in Test 1, G_a and C_a after Test 2 were higher than those after Test 1. Therefore, the spectral changes do not show a monotonic dependence on the water absorption amount.

Summarizing the above results, more water penetrated into the film in Test 1 and the changes in the chemical bonds of the film resin were greater in Test 2. The D-SIMS and EIS results indicate that there were more mobile charges in the coating film after Test 2. Nevertheless, the film after Test 2 showed a similar G_a value at 10⁻² Hz to that observed after Test 1 (i.e., the degradation of corrosion protection was similar). This is likely to be attributed to the higher water absorption in Test 1. In other words, the anti-corrosion performance of a coating film is governed by the extent of charge mobility, which is determined by the number of charges in the coating film, the moisture content, and the chemical bonding state of the resin. In an environment with constant high temperature and high humidity, as in Test 1, water and ions diffuse and penetrate into the coating film with increasing exposure time and the amount of water and the number of ions that penetrate dominate the degradation of the corrosion resistance. In contrast, as in Test 2, where stresses are expected to occur in the film or at the film/steel interface due to the thermal cycle and water freezing, the effect of volume expansion due to freezing of the moisture on the bonding state of the resin also affects the penetration of water and ions into the film, accelerating the degradation of the resin. In addition, the D-SIMS results suggest that the penetrating ions are not involved in ion exchange in Test 2.

From the above results, the frequency dependence of G_a and C_a is clearly affected by the water content in the coating film and the chemical bonding state of the resin in the film.

4. Conclusions

Coated steel panels were exposed to environments with simulated snow-melting salt. One test used high temperature and high humidity (80°C, 95%RH), and another test also included a freezing process (80°C, 95%RH ↔ −10°C). The changes in the coating films after the exposure tests were investigated by analyzing the frequency dependence of the admittance, the elements, and the infrared spectra. The following conclusions were obtained.

1. (1)    The spectra showing the frequency dependence of G_a and C_a obtained by analyzing the admittance before and after the environmental tests showed that the charges that contributed to conductivity and capacitance increased in both the environmental tests. The shapes of the spectra obtained after each environmental test differed.
2. (2)    The D-SIMS results showed that salt penetrated into the coating films during the environmental tests. The elements were distributed more uniformly after a longer exposure to the high-temperature and high-humidity environment. In addition, cation exchange seemed to have occurred in the coating film.
3. (3)    The G_a and C_a values for Test 2, where Cl penetrated into the film, were large from 10⁻¹ to 10³ Hz.
4. (4)    AFM-IR analyses showed that the increase in G_a and C_a cannot be explained only by the increase in moisture content, although a large amount of moisture penetrated into the coating film in Test 1. The results obtained after Test 2 suggest that the frequency dependence of G_a and C_a may reflect the amount of water in the coating film and the state of the elements that penetrate into the coating film.

REFERENCES

• 1)   T.  Asada,  T.  Shigenaga and  Y.  Isomoto: Evaluation of Corrosion Resistance for Electrodeposition Coating Steel with a High Voltage Supplied Method, Zairyo-to-Kankyo 68 (2019) 92–97. doi:10.3323/jcorr.68.92
• 2)   T.  Yagasaki and  Y.  Kimura: Effects of Degradation Factors on the Fatigue Characteristics of Polyetherimide Resin, J. Soc. Mat. Sci. Japan 45 (1996) 328–333. doi:10.2472/jsms.45.328
• 3)   S.  Kumagai,  X.  Wang and  N.  Yoshimura: Thermal Ageing, Water Absorption and Their Multiple Effects on Tracking Resistance of Epoxy for Outdoor Use, T. IEE Japan 118 (1998) 1255–1263. doi:10.1541/ieejfms1990.118.11_1255
• 4)   T.  Tsuru: Measurement Technology on Aquaous Corrosion, Materia Japan 50 (2011) 283–286. doi:10.2320/materia.50.283
• 5)   T.  Horie and  G.  Kitahara: Change in Electrochemical Impedance Spectrum of Coated Steel Plate by High Temperature·High Humidity-Cold Cycle test, Zairyo-to-Kankyo 68 (2019) 182–186. doi:10.3323/jcorr.68.182
• 6)   T.  Horie and  G.  Kitahara: Change in Electrochemical Impedance Spectrum of Al/CFRP Adhesive Lap Joint by High Temperature, High Humidity-Cold Cycle test, Zairyo-to-Kankyo 70 (2021) 365–369. doi:10.3323/jcorr.70.365
• 7)   H.  Sugitani and  Y.  Mukoyama: Analyses of Polymers Used for Electronic Industry, Jpn. Thermosetting Plastic Industry Association 10 (1989) 184–198. doi:10.11364/networkpolymer1980.10.184
• 8)  https://www.kobelcokaken.co.jp/contract/physics_analysis/pdf/pa-2.pdf [Accessed May 30, 2023].
• 9)  ISO 16773-4, Electrochemical impedance spectroscopy (EIS) on coated and uncoated metallic specimens Part4: Examples of spectra of polymer-coated and uncoated specimens, (2017).
• 10)  M. Itagaki: Denkikagaku inpidansu hou, 2nd ed., (Maruzen Shuppan, Tokyo, 2011).
• 11)   T.  Tsuru,  M.  Asari and  S.  Haruyama: Impedance characteristics and mass transport of underfilm corrosion, J. Metal Finishing Soc. Jpn. 39 (1988) 2–9. doi:10.4139/sfj1950.39.2
• 12)   A.  Dazzi,  C.B.  Prater,  Q.  Hu,  D.B.  Chase,  J.F.  Rabolt and  C.  Marcott: AFM–IR: Combining Atomic Force Microscopy and Infrared Spectroscopy for Nanoscale Chemical Characterization, Appl. Spectrosc. 66 (2012) 1365–1384. doi:10.1366/12-06804
• 13)   N.  Baden: Nanoscale Infrared Spectroscopy, J. Jpn. Soc. Tribologists 64 (2019) 661–666. doi:10.18914/tribologist.64.11_661
• 14)  T. Hanai: Fukinshitsukouzou-to-yudennritsu, (Yoshiokashoten, 2000).
• 15)   Y.  Masubuchi: Practice and Problems in the Applications of Cathodic Electropainting to Automotive Industry, Jitsumu Hyomen Gijutsu 26 (1979) 610–618. doi:10.4139/sfj1970.26.610
• 16)   Y.  Kanzaki and  N.  Suzuki: On the Selectivity of Lon Exchange Reaction. The Hydration in Solid Phase Determines the Selectivity, J. Ione Exchange 12 (2001) 57–66. doi:10.5182/jaie.12.57
• 17)   Y.  Sasaki,  S.  Tagashira,  Y.  Murakami,  I.  Fujiwara and  K.  Hayashi: Adsorption isotherms for bivalent cations on ion exchange resins, Bunseki Kagaku 43 (1994) 111–116. doi:10.2116/bunsekikagaku.43.111