Effect of addition of non-functionalized graphene oxide in a commercial epoxy resin used as coating

Samples were prepared by curing commercial epoxy resin DER 331 (epoxide equivalent weight of 182–192 Da), with a commercial epoxy hardener 2767, both received from the Mexican Polymer Corporation. GO was synthesized by a modified Hummers’ method [27] using graphite from Sigma-Aldrich. As substrates, cold-rolled carbon steel coupons were used; they were washed with detergent and rinsed with deionized water and acetone, and dried at 60°C for 15 min.

GO was dispersed in DER 331 at concentrations of 0.1 wt.%, 0.5 wt.%, 1 wt.%, and 3 wt.%, and the samples were labeled as EG0, EG01, EG05, EG1, and EG3, respectively; EG0 was treated as the reference sample. The procedure of the GO-epoxy dispersion was as follows: GO powders were dispersed in acetone by ultrasonic treatment for 1 h at room temperature. Then, GO suspension was mixed with the epoxy resin followed by sonication for 30 min. Next, acetone was removed via evaporation, and finally the stoichiometric amount of hardener 2767 was added to the mixture (the weight ratio of resin to hardener was 1:1) with mild stirring. The obtained solution was coated onto cold-rolled carbon steel panels using a wire-wound applicator. Then, the panels were stored at room temperature for 24 h and cured at 120°C for 2 h.

Scanning electron microscopy (SEM) images of the nanocomposite coatings were obtained by a JEOL JSM-6610LV instrument equipped with an Oxford Instruments energy dispersive X-ray spectroscopy device. The samples were placed on conductive aluminum tape and coated with gold in a sputtering unit (Denton Vacum, DESK v SCD-030) before their analysis. The analyses were carried out using a secondary electron detector working in high vacuum mode and at 20 kV of accelerating voltage. X-ray diffraction (XRD) data were collected on a Bruker model diffractometer, D8 Discover, in a Bragg–Brentano geometry using CuKα radiation (λ = 0.15418 nm) and step-scan mode (range: 5–80° of 2θ, step-time: 0.60 s, step-size: 0.04°). Fourier transform infrared spectroscopy (FTIR) was carried out on a Bruker spectrometer model, Vertex 70, in the interval of 500–4,000 cm⁻¹.

The anticorrosive properties were determined by electrochemical impedance spectroscopy (EIS). Coated coupons were immersed into a 5 wt.% NaCl solution for 500 h. A three-electrode cell arrangement, with the coated carbon steel substrates as a working electrode, a graphite bar as a counter electrode, and a saturated calomel electrode as the reference electrode, was used. AC impedance measurements were conducted on open circuit potential using a Gamry, Reference 600 potentiostat/galvanostat/ZRA over a frequency range of 100 kHz–10 mHz, and the amplitude of sinusoidal voltage was 10 mV. Interpretation of the impedance data was performed using the Gamry Instruments Echem Analyst fitting program, version 6.25.

GO dispersion in DER 331 matrix was first investigated by SEM (Figure 1). The micrographs revealed that the EG0, EG01, and EG05 epoxy coatings exhibited smooth and homogeneous surfaces without cracks or discontinuities. Nevertheless, the surface became rougher when the GO increased, as can be seen in EG1 and EG3 (Figure 1). Moreover, GO clusters were clearly identified in EG3 indicating that an incorporation and dispersion limit is achieved, then when the content of GO increases, good dispersion and exfoliation of GO is not obtained in the samples.

Fig. 1

Figure 2 shows the coatings’ cross-section, where an adhesion in the coating–metal interface can be observed for all samples, reaching the best results each with the EG0, EG01, EG05, and EG1 samples. The average of the measured coatings’ thicknesses was found to be in the 26–35 μm range. It would be relevant to mention in this context that a plausible explanation for the poor coating–metal adhesion disclosed by EG3 is the GO cluster formation (Figure 1E).

Fig. 2

The synthesized samples were analyzed by FTIR spectroscopy (Figure 3) to monitor the curing reaction of epoxy resin by considering the absorbance decrease of the oxirane group band located in the 970–914 cm⁻¹ region. After curing, this band disappears in the coatings’ samples, indicating a good curing reaction as reported by Grundmeier et al. [28] and Zheng et al. [29].

Fig. 3

GO dispersion in epoxy resin was also studied by the XRD technique to determine the structural transformation in GO, if any, after its incorporation in the resin matrix (Figure 4). Pristine graphite and GO were also included for comparison. The characteristic reflections of pristine graphite were identified at 26.5, 44.5, and 54.5 of 2θ degrees (PDF # 41-1487). The interlayer spacing was calculated from the (002) plane corresponding value of 0.33 nm. On the other hand, the XRD pattern of GO showed a characteristic reflection at 2θ = 10.7°, which corresponded to an interlayer spacing of about 0.83 nm. The interlayer spacing observed in the case of GO is larger than that of graphite (0.33 nm), and the larger interlayer spacing of GO is attributable to the “decoration” given to it by oxygen-containing groups, which expand the interlayer distance, as indicated in the researches of Liao et al. [30] and Pei and Cheng [31].

All samples exhibited a broad signal centered at 2θ = 18.40, which is related to the amorphous structure of cured epoxy resin. It is important to note that GO characteristic reflections were not detected, which is indicative of its good dispersion in the resin matrix.

Fig. 4

However, the XRD spectra of samples EG05, EG1, and EG3 also showed small reflections centered at 25.07°, 25.39°, and 25.67°, respectively, which were attributed to multilayer graphene, which is formed during the thermal curing process [32]. These results showed that after the curing treatment, part of the GO structure transformed into multilayer graphene, achieving interlayer spacing values of 0.355 nm, 0.351 nm, and 0.347 nm (corresponding to EG05, EG1, and EG3, respectively). Besides, a broadening effect was also noticed with GO increase (which is more evident in EG1), indicating that graphene stacking is not well ordered [33]. By the Scherrer equation – L₍₂₀₀₎ = kλ/β cos θ, where k is a dimensionless shape factor with a value of 0.9, λ is the X-ray wavelength, β is the line broadening determined by the full width at half maximum (FWHM), and θ is Bragg's angle – GO thickness was determined, obtaining a value of 7.1 nm. In this sense, if d₍₀₀₂₎ is considered as an individual GO layer, a multilayered graphene consisting of ∼ 9 (8.54) graphene layers can be assumed.

Figure 5 shows the electrical equivalent circuit that models the electrochemical behavior of the coated steel panels exposed to 5 wt.% NaCl solution. The equivalent circle consists of R_E, constant phase element (CPE)_c, R_C, CPE_dl, and R_CT, which are solution resistance, coating CPE, coating pore resistance, double layer constant phase, and charge transfer resistance, respectively. The presented model employs CPEs instead of capacitors; CPE represents the deviation from ideal capacitive behavior and is used in the equivalent circuit to minimize the systematical error and obtain more precious fitting results. The CPE is associated with an exponent (0 ≤ N ≤ 1) which indicates the deviation from an ideal dielectric behavior; if N = 1, it represents an ideal capacitor behavior and if N = 0, it behaves as an ideal resistor, as reported by Mondal et al. [34].

Fig. 5

EIS measurement is used to evaluate the corrosion protection performance of nanocomposite coatings with different GO wt.%. The Bode and Nyquist plots of samples after immersion in 5 wt.% NaCl solution for 0 h, 150 h, 300 h, and 500 h are presented in Figures 6 and 7.

Figure 6 shows that the impedance value of nanocomposite coatings significantly depends on GO wt.%. The impedance value of nanocomposite coatings decreases in the order of 0.5 wt.% GO > 0.1 wt.% GO > 0.0 wt.% GO > 1 wt.% GO > 3 wt.% GO during immersion time. Furthermore, the impedance values of coatings decrease by increasing immersion time.

Fig. 6

Fig. 7

The Bode plot for EG0, EG01, and EG05 is a straight line with a negative slope (−1) in the first times, revealing that the coating behaved as a perfect capacitor (Figure 6A), showing extremely high resistance of the pore (R_p) for nanocomposite coatings. The impedance for EG1 and EG3 were lower and the little plateau observed in the Bode plot confirmed that these coatings had begun water absorption.

After 150 h of immersion, it was observed that the impedance value was lower than its early state. For instance, the EG05 impedance value decreased to 7.1 GΩ·cm⁻² but it maintained a linear relationship with frequency and the slope did not change, suggesting water absorption. In the same way, the impedance of neat epoxy (EG0), EG01, and EG1 coated samples also decreased to 28.2 MΩ·cm⁻², 27.5 MΩ·cm⁻², and 26.6 MΩ·cm⁻², respectively. As revealed by the horizontal intersections of the graph with the impedance modulus axis in the Bode plot (Figure 6B), this behavior is indicative of the electrolyte's penetration of the coatings. It is worth mentioning that sample EG3 registered an impedance value of 6.9 MΩ·cm⁻², which was the lowest among the samples; this observation is in good agreement with the XRD and SEM analysis, which revealed cluster formation, structure distortion, and lower substrate-coating adhesion.

After 300 h of immersion, sample EG05 provided a better barrier protection (4.95 GΩ·cm⁻²) than the others. The impedance value of sample EG3 was the lowest; this was signaled by a more complex Bode plot, which exhibited two time constants, indicating the initiation of metallic substrate corrosion. After 500 h of immersion, the Bode plot for sample EG05 showed the highest impedance (3.71 GΩ·cm⁻²), while there was a decrease in the protective properties of the EG01, EG0, EG1, and EG3 samples, meaning that water had diffused into the coating.

3.2.1

Variations of corrosion electrochemical parameters

Coating resistance (R_p) is related to the barrier property of coating, and it was thus used to evaluate the protective performance of the coatings. The coating capacitance (C_c) is closely related to the diffusion behavior of the electrolyte in the coating, which reflects the barrier properties of the coating [35]. When the electrolyte penetrates the coating, its dielectric constant changes, resulting in a C_c variation.

As can be seen from Table 1, the R_p and R_CT of all coatings decreased gradually until the end of the test. During the exposure time, the R_p and R_CT values of sample EG05 decreased slowly; also, these values were significantly higher than those of the other samples, including the reference sample EG0, throughout the entire investigated time. This behavior can be explained in terms of the micropores’ filling by GO, leading to significant enhancement of the barrier properties. When the corrosive electrolyte reaches the coating-metal interface, electro-chemical reactions, including oxidation of active sites on the metal surface and reduction of solution species (O₂ and H⁺ ions), take place, leading to the formation and dissolution of corrosion products affecting the coating's electrochemical parameters.

Table 1

The electrochemical parameters extracted from EIS data for nanocomposite epoxy coatings containing 0 wt.%, 0.1 wt.%, 0.5 wt.%, 1 wt.%, and 3 wt.% GO after immersion in 5 wt.% NaCl solution for different durations

Sample	time (hours)	CPEC		RP (Ω·cm2)	CPEdl		RCT (Ω·cm2)	x2
		Y0 (Ω−1·cm−2·sn)	n		Y0 (Ω−1·cm−2·sn)	n
EG0	0	8.70E−10	9.77E−01	4.89E+10	7.11E−10	9.52E−01	1.45E+10	4.64E−04
	150	2.32E−09	9.56E−01	9.80E+06	1.37E−07	4.00E−01	6.07E+08	2.42E−04
	300	1.96E−09	9.15E−01	9.50E+06	7.44E−07	2.89E−01	2.36E+07	3.96E−04
	500	5.79E−09	8.71E−01	321000	4.67E−06	5.61E−01	8.71E+05	2.59E−03
EG01	0	6.96E−10	9.66E−01	5.45E+10	2.92E−10	6.00E−01	1.49E+12	2.39E−03
	150	1.80E−09	9.34E−01	8.24E+07	5.53E−09	8.40E−01	4.64E+08	4.40E−03
	300	2.38E−09	9.51E−01	3.90E+07	1.77E−07	3.97E−01	1.12E+07	2.82E−04
	500	4.46E−09	9.38E−01	431000	1.82E−06	2.63E−01	5.58E+08	9.06E−04
EG05	0	7.97E−10	9.82E−01	9.60E+10	1.72E−09	8.23E−01	5.30E+09	6.68E−04
	150	9.28E−10	9.77E−01	8.73E+09	3.53E−10	4.35E−01	1.38E+12	3.56E−04
	300	7.13E−10	9.69E−01	7.26E+09	8.56E−10	6.19E−01	1.27E+10	4.45E−04
	500	9.13E−10	9.59E−01	4.70E+09	2.66E−11	1.80E−01	6.21E+07	1.08E−02
EG1	0	1.06E−09	9.70E−01	5.29E+09	1.34E−13	9.59E−01	6.32E+09	5.01E−03
	150	2.97E−09	8.99E−01	3.94E+06	1.62E−07	6.20E−01	2.96E+10	8.81E−02
	300	3.14E−09	8.96E−01	1.54E+07	2.36E−06	4.46E−01	5.84E+06	7.93E−04
	500	7.05E−09	9.57E−01	138000	3.62E−06	4.12E−01	4.11E+06	3.24E−04
EG3	0	2.41E−09	8.13E−01	3.16E+09	2.42E−09	9.23E−01	1.30E+09	2.31E−02
	150	4.18E−09	9.19E−01	1.32E+06	3.35E−07	3.00E−01	3.38E+07	4.36E−04
	300	1.50E−08	8.80E−01	22330	5.37E−06	7.99E−01	1.89E+05	1.56E−03
	500	2.48E−08	8.54E−01	2270	2.81E−05	7.46E−01	2.22E+04	1.91E−03

CPE, Constant phase element; EIS, Electrochemical impedance spectroscopy; GO, Graphene oxide

The overall loss of coating resistance for sample EG3 is much sharper (6 and 5 orders of magnitude) than the other samples. These findings indicate that delamination occurred in sample EG3, which is consistent with the visual examination of the exposed panels (Figure 8).

Fig. 8

Table 1 shows that the EG0, EG01, and EG1 samples’ CPE_C ascended slowly in the initial stage of exposure and then declined to a stable amount, indicating that the homogeneity of the coatings decreased along with the water uptake process, which continued until the coatings were saturated with water. Overall, the highest coating capacitance values were obtained for sample EG3, where the amount of GO was 3 wt.%, suggesting a higher water uptake; this is due to functional groups in GO making it hydrophilic; and as a result of increasing the GO, more electrolytes permeate into the coating [36]. Moreover, GO aggregation occurs when the amount of GO in nanocomposite coatings (ascertained based on SEM observations) is increased, leading to decrease of GO aspect ratio and therefore acceleration of the corrosion phenomenon [37].

This is consistent with the results from visual examination of the exposed panels (Figure 8), where sample EG03 showed evident corrosion spots and delamination. In turn, the lowest coating capacitance values were obtained for sample EG05, indicating a lower water absorption than in the case of epoxy coating.

All of these observations demonstrate that GO wt.%, and therefore GO distribution in the polymer matrix, have an important influence on coating protective performance. According to SEM results, GO nanosheets tend to agglomerate in the polymer matrix by increasing GO wt.% to >0.5 wt.%, and this agglomeration is attributed to their high specific surface area [38]. Also, nanocomposites containing 0.5 wt.% GO have the best GO distribution in the polymer matrix.