Graphene Oxide / Polyaniline Nanocomposites Used in Anticorrosive Coatings for Environmental Protection

: In recent years, metal corrosion causes serious threats to the economy of the world and the living environment. Hence, it is very important to seek non-toxic and environmentally friendly materials with metal anti-corrosion properties for the sustainable development of society. The barrier properties of graphene oxide (GO) and the special electrochemical property of polyaniline (PANI) can signiﬁcantly improve the corrosion resistance of metals. Herein, we developed an in-situ polymerization method to prepare graphene oxide / polyaniline (GO / PANI) nanocomposites with unique anti-corrosion properties. The obtained GO / PANI nanocomposites were characterized by Fourier transform infrared spectroscopy, X-ray di ﬀ raction, thermalgravimetric analysis, UV–vis spectroscopy and scanning electron microscopy. The as-prepared composite materials were uniformly dispersed in epoxy resin to prepare anticorrosive coatings and coated on the surface of steel. The anti-corrosion performance of the coatings was measured by electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization technique. The EIS results showed that the total impedance of epoxy / GO / PANI coatings is greater than epoxy / PANII coatings, and the impedance module value can reach 8.67 × 10 8 Ω · cm 2 . In general, it is concluded that the anti-corrosion performance of GO / PANI coating is signiﬁcantly higher than PANI coating and pure GO coating. both the surface of GO and the intercalation with the agglomeration phenomenon is signiﬁcantly reduced and the morphology of PANI has changed. It shows that the GO / PANI composite was successfully synthesized, and the dispersion of PANI was enhanced.


Introduction
Metal corrosion is a phenomenon that metallic material is destroyed by the surrounding medium. Nowadays, metal corrosion has been a common problem faced by all countries in the world, due to that it significantly reduces the mechanical properties of metal materials and the service life of metal equipment. Therefore, metal corrosion must be prevented and controlled. Among the numerous anti-corrosion methods, anticorrosive coatings are a kind of simplest, most extensive, effective and economical method [1,2]. Corrosion inhibitors containing heavy metals such as Pb, Cr, Zn, etc. are often added to traditional anticorrosive coatings. The residue of these heavy metals will cause great harm to human health and the ecological environment. Therefore, it is of great significance for the sustainable development of society to seek non-toxic and bio-environmentally friendly coatings with metal anti-corrosion properties.
In recent years, the addition of conductive polymers to organic coatings to inhibit metal corrosion has become a research hotspot. Among the conducting polymers, polyaniline (PANI) is most promising due to its ease of synthesis, non-toxic, low cost, excellent chemical and environmental stability, as well as reversible redox properties [3][4][5]. PANI has been widely used in the fields of batteries, catalysts, sensors, light-emitting diodes, electrochromic and anti-corrosion coatings [6][7][8], etc. Since DeBerry [9] discovered by compounding GO and PANI is significantly improved. The non-toxic and environmentally friendly GO/PANI coatings are multifunctional coatings for anti-corrosion.
In the present investigation, the GO/PANI nanocomposites were prepared by in-situ polymerization. By controlling the content of GO, different GO/PANI composites were prepared and dispersed uniformly in epoxy resin to prepare anticorrosive coatings. The resulting samples were characterized by Fourier transform infrared (FT-IR), UV-vis spectroscopy (UV-vis), and X-ray diffraction (XRD). The thermal properties of the materials were studied by thermogravimetric analysis (TGA) analysis, the anticorrosive performance of epoxy coatings was investigated using electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization. In this study, the PANI-GO (4 wt.%) coating shows the best anti-corrosion performance, which is much better than the PANI and pure epoxy coating.

Preparation of GO
We used a modified hummers method to prepare GO. Firstly, mixed 5 g of graphite powder and 2.5 g of NaNO 3 in a 1000 mL round bottom flask with 115 mL of concentrated H 2 SO 4 in an ice bath for about 15 min. Then, we added 15 g of KMnO 4 in batches within 1 h, and continuously stirred for 2 h. In the next step, raised the temperature to 35~40 • C, and continuously stirred for 4 h. After that, we added 150 mL of distilled water to a flask and set the temperature at 98 • C with continuous stirring for 30 min. We terminated the reaction after adding a large number of distilled water. Finally, we slowly added H 2 O 2 solution until the reaction no longer generated bubbles. The reactant was centrifuged and washed with dilute hydrochloric acid until the sulfate ions were completely removed, and then washed with distilled water to neutrality to obtain 0.02 wt.% GO suspension. The specific preparation process is shown in Figure 1.

Preparation of Graphene Oxide/Polyaniline (GO/PANI) Nanocomposites
The reaction was carried out at −5~0 °C under continuous stirring. Different content of GO (0, 1, 2, 3, 4, and 5 wt.%) was added into 25 mL deionized water and ultrasonicated for 30 min to get a uniform solution. Then, aniline (2.33 g) was added to the GO suspension. The APS (2.00 g) was

Preparation of Graphene Oxide/Polyaniline (GO/PANI) Nanocomposites
The reaction was carried out at −5~0 • C under continuous stirring. Different content of GO (0, 1, 2, 3, 4, and 5 wt.%) was added into 25 mL deionized water and ultrasonicated for 30 min to get a uniform solution. Then, aniline (2.33 g) was added to the GO suspension. The APS (2.00 g) was dissolved in 40 mL of 1 M HCl and stirred up for 10 min. Finally, the APS solution was slowly added to the flask dropwise. Under the ice bath and magnetic agitation condition, the reaction lasted for 6 h. The GO/PANI was repeatedly washed with distilled water and ethanol, and then dried at 60 • C in a vacuum for 24 h.

Preparation of Coatings
GO, PANI and GO/PANI nanocomposites were taken into 2 mL ethanol respectively, and ultrasonically dispersed for 20 min. Afterwards, 3 g of epoxy resin (E-44) was added to the above solution and dispersed at high speed for 30 min. We added 2 g of harder (polyamide 650) to this mixture, and then put the sample in a centrifuge tube, through which the bubbles of the coating are removed. Prior to coating, the steel plate (10 × 10 × 2 mm 3 ) was polished with emery papers (grade 500, 1000, and 2000), degreased with ethanol, and dried. The above mixture was uniformly coated on the surface of the steel plate with a wire rod coater to control the coating thickness to 60 µm. Finally, coatings were cured at room temperature for 24 h, then put in an oven at 60 • C to dry for 24 h. Different coatings were prepared by the same method, which are denoted as pure epoxy, GO, PANI, PANI-GO(1 wt.%)(PG1), PANI-GO(2 wt.%)(PG2), PANI-GO(3 wt.%)(PG3), PANI-GO(4 wt.%)(PG4) and PANI-GO(5 wt.%)(PG5), respectively.

Characterization
The surface morphology of the materials was characterized by a scanning electron microscope (TecnaiG220, FEI, Hillsboro, OR, USA). Fourier Transform Infrared spectroscopy (Avatar 360 Nicolet, Thermo Fisher Scientific, Waltham, MA, USA) was employed to investigate the surface functional groups of materials. X-ray powder diffraction tests were performed on Cu-Kα radiation (DY5261/Xpert3, PANalytical B.V., Malvern Panalytical, Malvern, UK), the scan range is from 10 • to 90 • and the scanning speed is 0.2 • /s. UV-vis spectroscopy was performed by UV-vis spectrophotometer (TU-1900, Beijing Puxi General Instrument Co., Ltd., Beijing, China) in the wavelength region from 200 to 800 nm. Thermogravimetric analysis was employed with an automatic sampling synchronous thermal analyzer instrument (STA449F5, Selb, Germany), with heating from room temperature to 900 • C at a rate of 10 • C/min under an N 2 atmosphere. For the corrosion studies, a conventional three-electrode cell system was assembled by the coated steel panel as working electrode, platinum as a counter electrode and a saturated calomel electrode (SCE) as a reference in an electrolyte of 3.5 wt.% NaCl solution. Electrochemical impedance spectroscopy (EIS) measurements and potentiodynamic polarization curves were done by the CHI-650E electrochemical workstation (Shanghai CH Instrument Company, Shanghai, China). Figure 2 shows the scanning electron microscopy (SEM) images of GO, PANI and GO/PANI composites. As shown in Figure 2a, the pure GO has a flaky structure with the surface is smooth and pleated. The SEM image of the pure PANI presents short-rod morphology with partially agglomerated in some sections. It can be seen from Figure 2c that the flaky structure of GO is still retained, while PANI accumulates on both the surface of GO and the intercalation with the agglomeration phenomenon is significantly reduced and the morphology of PANI has changed. It shows that the GO/PANI composite was successfully synthesized, and the dispersion of PANI was enhanced. pleated. The SEM image of the pure PANI presents short-rod morphology with partially agglomerated in some sections. It can be seen from Figure 2c that the flaky structure of GO is still retained, while PANI accumulates on both the surface of GO and the intercalation with the agglomeration phenomenon is significantly reduced and the morphology of PANI has changed. It shows that the GO/PANI composite was successfully synthesized, and the dispersion of PANI was enhanced.

XRD Analysis
The crystal structures of the different samples were studied by powder X-ray diffraction (XRD) measurements. Figure 3 shows the XRD patterns of GO, PANI, and GO/PANI composites. The XRD pattern of GO revealed an intense and sharp reflection peak at 2θ = 8.8° (001), corresponding to d spacing of 1.0 nm. This interlaminar distance is due to the presence of oxygen-containing groups (such as carboxyl, epoxy, hydroxyl) and water absorption in the GO structure [34]. In the case of PANI, the characteristic peaks at 8.4°, 15.1°, 20.1°, 25.3°, 26.8° and 29.2°, correspond to (001), (011), (020), (200), (121), and (022) diffraction planes of emeraldine salt form PANI, respectively [35,36]. The XRD patterns of synthesized GO/PANI nanocomposites are the same as that of PANI, which indicates that the molecular chain structure of PANI was not destroyed. No obvious peak of GO was observed in the XRD pattern of GO/PANI, indicating that the GO has almost no agglomeration, but fully reacted with the polymerization reaction substrate to synthesize the GO/PANI nanocomposite [37].

XRD Analysis
The crystal structures of the different samples were studied by powder X-ray diffraction (XRD) measurements. Figure 3 shows the XRD patterns of GO, PANI, and GO/PANI composites. The XRD pattern of GO revealed an intense and sharp reflection peak at 2θ = 8.8 • (001), corresponding to d spacing of 1.0 nm. This interlaminar distance is due to the presence of oxygen-containing groups (such as carboxyl, epoxy, hydroxyl) and water absorption in the GO structure [34]. In the case of PANI, the characteristic peaks at 8.  [35,36]. The XRD patterns of synthesized GO/PANI nanocomposites are the same as that of PANI, which indicates that the molecular chain structure of PANI was not destroyed. No obvious peak of GO was observed in the XRD pattern of GO/PANI, indicating that the GO has almost no agglomeration, but fully reacted with the polymerization reaction substrate to synthesize the GO/PANI nanocomposite [37]. agglomerated in some sections. It can be seen from Figure 2c that the flaky structure of GO is still retained, while PANI accumulates on both the surface of GO and the intercalation with the agglomeration phenomenon is significantly reduced and the morphology of PANI has changed. It shows that the GO/PANI composite was successfully synthesized, and the dispersion of PANI was enhanced.

XRD Analysis
The crystal structures of the different samples were studied by powder X-ray diffraction (XRD) measurements. Figure 3 shows the XRD patterns of GO, PANI, and GO/PANI composites. The XRD pattern of GO revealed an intense and sharp reflection peak at 2θ = 8.8° (001), corresponding to d spacing of 1.0 nm. This interlaminar distance is due to the presence of oxygen-containing groups (such as carboxyl, epoxy, hydroxyl) and water absorption in the GO structure [34]. In the case of PANI, the characteristic peaks at 8.4°, 15.1°, 20.1°, 25.3°, 26.8° and 29.2°, correspond to (001), (011), (020), (200), (121), and (022) diffraction planes of emeraldine salt form PANI, respectively [35,36]. The XRD patterns of synthesized GO/PANI nanocomposites are the same as that of PANI, which indicates that the molecular chain structure of PANI was not destroyed. No obvious peak of GO was observed in the XRD pattern of GO/PANI, indicating that the GO has almost no agglomeration, but fully reacted with the polymerization reaction substrate to synthesize the GO/PANI nanocomposite [37].   Figure 4 shows the FT-IR spectra of the GO, PANI and PANI/GO nanocomposites in the region of 3900-500 cm −1 , respectively. In the FT-IR spectra of the GO, there is a wide peak near at 3420 cm −1 , which is caused by the tensile vibration of the O-H bond group in GO [15]. The peaks at 1740 and 1620 cm −1 are attributed to the C-O, C=O in COOH, and peaks at 1210 and 1050 cm −1 are caused by the vibration of -OH, and C-O-C, respectively [38,39]. For the PANI, the peaks at 1580 and 1480 cm −1 correspond to the characteristic C=C stretching mode of the quinoid and benzenoid rings, respectively [40]. Peaks at 1130 and 804 cm −1 are assigned to in-plane and out-of-plane bending of C-H [41]. These absorption bands are significantly decreased in the FTIR spectra of the GO/PANI, indicating that most of the functional groups had been substantially removed by chemical reduction. The mainly characteristic peaks of the GO/PANI composite shows are similar to the peaks of PANI, and it shows the characteristic peak of O-H of GO at 3430 cm −1 . From the spectrum of the GO/PANI nanocomposite, all the characteristic peaks of PANI and GO can be found, indicating the nanocomposite successfully prepared.

FT-IR Analysis
of C-H [41]. These absorption bands are significantly decreased in the FTIR spectra of the GO/PANI, indicating that most of the functional groups had been substantially removed by chemical reduction. The mainly characteristic peaks of the GO/PANI composite shows are similar to the peaks of PANI, and it shows the characteristic peak of O-H of GO at 3430 cm −1 . From the spectrum of the GO/PANI nanocomposite, all the characteristic peaks of PANI and GO can be found, indicating the nanocomposite successfully prepared.  Figure 5 illustrates the UV-vis absorption spectra of GO, PANI and GO/PANI nanocomposites dispersion in ethanol solution. The UV-vis spectrum of neat GO includes a sharp peak at 227 nm is attributed to π-π* transition, and there is a shoulder peak at 295 nm is related to the n-π* transition, showing that there are some oxygen-containing functional groups (hydroxyl, carboxylic acid and epoxide) in GO, and these functional groups are highly consistent with the absorption peaks of GO in previous reports [31]. The two characteristic peaks of PANI appeared at 359 and 691 nm, which was attributed to the π-π* transition of an aromatic C−C bond and a quinonoid C−C bond, respectively [24]. The spectrum for GO/PANI composites is similar to that for the PANI sample, except a new absorption band at 235 nm, which corresponds to the presence of GO. This result clearly shows the successful in-situ compounding of GO in the composites.   Figure 5 illustrates the UV-vis absorption spectra of GO, PANI and GO/PANI nanocomposites dispersion in ethanol solution. The UV-vis spectrum of neat GO includes a sharp peak at 227 nm is attributed to π-π* transition, and there is a shoulder peak at 295 nm is related to the n-π* transition, showing that there are some oxygen-containing functional groups (hydroxyl, carboxylic acid and epoxide) in GO, and these functional groups are highly consistent with the absorption peaks of GO in previous reports [31]. The two characteristic peaks of PANI appeared at 359 and 691 nm, which was attributed to the π-π* transition of an aromatic C−C bond and a quinonoid C−C bond, respectively [24]. The spectrum for GO/PANI composites is similar to that for the PANI sample, except a new absorption band at 235 nm, which corresponds to the presence of GO. This result clearly shows the successful in-situ compounding of GO in the composites.

UV-Vis Analysis
indicating that most of the functional groups had been substantially removed by chemical reduction. The mainly characteristic peaks of the GO/PANI composite shows are similar to the peaks of PANI, and it shows the characteristic peak of O-H of GO at 3430 cm −1 . From the spectrum of the GO/PANI nanocomposite, all the characteristic peaks of PANI and GO can be found, indicating the nanocomposite successfully prepared.  Figure 5 illustrates the UV-vis absorption spectra of GO, PANI and GO/PANI nanocomposites dispersion in ethanol solution. The UV-vis spectrum of neat GO includes a sharp peak at 227 nm is attributed to π-π* transition, and there is a shoulder peak at 295 nm is related to the n-π* transition, showing that there are some oxygen-containing functional groups (hydroxyl, carboxylic acid and epoxide) in GO, and these functional groups are highly consistent with the absorption peaks of GO in previous reports [31]. The two characteristic peaks of PANI appeared at 359 and 691 nm, which was attributed to the π-π* transition of an aromatic C−C bond and a quinonoid C−C bond, respectively [24]. The spectrum for GO/PANI composites is similar to that for the PANI sample, except a new absorption band at 235 nm, which corresponds to the presence of GO. This result clearly shows the successful in-situ compounding of GO in the composites.

TGA Analysis
As shown in Figure 6, the thermal behavior of GO, PANI and GO/PANI nanocomposites samples was studied by TGA analysis. The initial mass loss of all materials approximately occurred at 120 • C due to the evaporation of moisture from the samples. For the GO, the first weight loss of~15% at 120 • C is due to the loss of the evaporation of physically adsorbed moisture in the samples. The weight loss in the next stage between 120 and 300 • C (about 45%) is due to the decomposition of carbon oxidation and oxygen-containing functional groups [42,43]. A lighter weight loss of 14% from 300 to 900 • C can be attributed to the removal of more stable oxygen-containing functional groups and the decomposition of GO ring chains. For the PANI, the main weight loss was divided into three steps. The first step of weight loss is observed with 12% weight loss at~120 • C, due to the vaporization of moisture and liquid elimination of unreacted monomer; the next one (120~300 • C) may be caused by the decomposition of doped acid and aniline oligomer produced in the reaction; and the third one (300~900 • C) is attributed to the complete decomposition of polymer back-bone. For the GO/PANI nanocomposites, exhibit similar decomposition patterns compared with PANI. The total weight loss of composites is 55%, the PANI is 63%. In general, the thermal stability of composites is found to be slightly higher as compared to PANI. 300 to 900°C can be attributed to the removal of more stable oxygen-containing functional groups and the decomposition of GO ring chains. For the PANI, the main weight loss was divided into three steps. The first step of weight loss is observed with 12% weight loss at ~120 °C, due to the vaporization of moisture and liquid elimination of unreacted monomer; the next one (120~300 °C) may be caused by the decomposition of doped acid and aniline oligomer produced in the reaction; and the third one (300~900 °C) is attributed to the complete decomposition of polymer back-bone. For the GO/PANI nanocomposites, exhibit similar decomposition patterns compared with PANI. The total weight loss of composites is 55%, the PANI is 63%. In general, the thermal stability of composites is found to be slightly higher as compared to PANI.

Electrochemical Impedance Spectroscopy
Electrochemical Impedance Spectroscopy (EIS) was used to study the corrosion behavior of the GO/PANI nanocomposite-based coatings. In the Bode plots, the high-frequency region responds to the protection of the coatings, while the middle and low-frequency region responds to the corrosion reaction between corrosive materials and metal. The Bode plots for the different coatings are shown in Figure 7a. As we know, the |Z| modulus displayed by the Bode diagram at low frequency (F = 0.01Hz) is determined to define the corrosion resistance of the coating. Correspondingly, the value of |Z|0.01Hz for different samples coated steel in 3.5 wt.% NaCl solution was found that increasing in the sequence epoxy < PANI < GO < PG1 < PG2 < PG5 < PG3 < PG4. It can be found that the corrosion resistance with GO, PANI and GO/PANI composites coatings is obviously better than the pure epoxy coatings. For the composite samples, withstanding corrosive competence of the coatings was increased first and then decreased as the GO loadings increased. Compared with all other samples, PG4 displayed the best anti-corrosion performance in 3.5 wt.% NaCl solution.

Electrochemical Impedance Spectroscopy
Electrochemical Impedance Spectroscopy (EIS) was used to study the corrosion behavior of the GO/PANI nanocomposite-based coatings. In the Bode plots, the high-frequency region responds to the protection of the coatings, while the middle and low-frequency region responds to the corrosion reaction between corrosive materials and metal. The Bode plots for the different coatings are shown in Figure 7a. As we know, the |Z| modulus displayed by the Bode diagram at low frequency (F = 0.01 Hz) is determined to define the corrosion resistance of the coating. Correspondingly, the value of |Z| 0.01Hz for different samples coated steel in 3.5 wt.% NaCl solution was found that increasing in the sequence epoxy < PANI < GO < PG1 < PG2 < PG5 < PG3 < PG4. It can be found that the corrosion resistance with GO, PANI and GO/PANI composites coatings is obviously better than the pure epoxy coatings. For the composite samples, withstanding corrosive competence of the coatings was increased first and then decreased as the GO loadings increased. Compared with all other samples, PG4 displayed the best anti-corrosion performance in 3.5 wt.% NaCl solution.   Figure 7b shows the Nyquist plots of the samples. Generally, the larger the diameter of the Nyquist curve is, the better the corrosion resistance of the sample will be. It shows that the diameters of the Nyquist curve has the same order as that of Bode plots, confirming the consistency of results. The impedance of the coatings is four to five orders of magnitude higher than that of the bare steel. With the increase of GO content in the composites, the coating demonstrates better anti-corrosion performance. However, the performance of the epoxy/PG5 coating decreased slightly, due to the agglomeration of PANI increased on the surface at higher GO content. In summary, the performance of GO/PANI composites is greater than PANI, GO, epoxy and uncoated coatings.   Figure 7b shows the Nyquist plots of the samples. Generally, the larger the diameter of the Nyquist curve is, the better the corrosion resistance of the sample will be. It shows that the diameters of the Nyquist curve has the same order as that of Bode plots, confirming the consistency of results. The impedance of the coatings is four to five orders of magnitude higher than that of the bare steel. With the increase of GO content in the composites, the coating demonstrates better anti-corrosion performance. However, the performance of the epoxy/PG5 coating decreased slightly, due to the Coatings 2020, 10, 1215 8 of 11 agglomeration of PANI increased on the surface at higher GO content. In summary, the performance of GO/PANI composites is greater than PANI, GO, epoxy and uncoated coatings. Figure 8 shows the potentiodynamic polarization curves of different coatings in 3.5 wt.% NaCl solution, and the values of corrosion potential (E corr ), corrosion current (I corr ) and corrosion rate (CR) of electrochemical measurements are summarized and shown in Table 1. The value of I corr and E corr can be determined via the intersection of the tangent of the anodic and cathodic polarization curves. The CR of the specimens was calculated using the following formula [44]:

Potentiodynamic Polarization Studies
where K is 3268.5 (mol/A), M is the molecular weight of iron 56 (g/mol), D is the density of iron (7.85 g/cm 3 ) and V is the valence of iron.
Coatings 2020, 10, x FOR PEER REVIEW 9 of 12 where K is 3268.5 (mol/A), M is the molecular weight of iron 56 (g/mol), D is the density of iron (7.85 g/cm 3 ) and V is the valence of iron.
As shown in Table 1, the Ecorr value of bare steel is seen to increase from −0.798 V for bare steel to more Ecorr as −0.704, −0.548, −0.569, −0.426, −0.419, −0.375, −0.403, and −0.378 V in epoxy-coated, epoxy/PANI, epoxy/GO, epoxy/PG1, epoxy/PG2, epoxy/PG3, epoxy/PG4 and epoxy/PG5 coated samples, respectively. It is obvious that the Ecorr of all the coating samples had a positive shift compared with bare steel. Ecorr mainly represents the tendency of corrosion reaction, where a sample with the higher Ecorr has a better corrosion resistance [45]. Therefore, it can be concluded that the addition of composite material significantly improves its anti-corrosion performance. According to formula (1), the Icorr and corrosion rate are in positive proportion to the relationship. Therefore, the sample with a lower Icorr value will have the lower corrosion rate. Table 1 indicates that the Icorr ranked by numerical value from the largest to smallest: uncoated steel, coatings in epoxy, epoxy/PANI, epoxy/GO, epoxy/PG1, epoxy/PG2, epoxy/PG5, epoxy/PG3 and epoxy/PG4. The corrosion current on the coated surfaces is apparently much lower (approx. 2-5 orders of magnitude) than the bare steel. It is obvious that coatings with epoxy/PG have better corrosion resistance than epoxy/PANI and epoxy/GO, demonstrating an increase in corrosion resistance of the GO/PANI nanocomposites coated substrates. Meanwhile, the decreased amplitude of corrosion current is positively correlated with the increase of GO content, indicating that GO plays a leading role in nanocomposites. It is easy to determine that the coating with PG5 composites have lower anti-corrosion property than the PG4, it may be that the excessive amount of GO leads to the deterioration of polymer dispersion performance, so that the barrier performance deteriorates. It can be seen that the results of electrochemical impedance spectroscopy are in good agreement with the potentiodynamic polarization. In particular, among all of the coatings, the epoxy/PG4 showed the best anti-corrosion performance and the lowest retained CR (5.46 × 10 −5 mm/year).  As shown in Table 1, the E corr value of bare steel is seen to increase from −0.798 V for bare steel to more E corr as −0.704, −0.548, −0.569, −0.426, −0.419, −0.375, −0.403, and −0.378 V in epoxy-coated, epoxy/PANI, epoxy/GO, epoxy/PG1, epoxy/PG2, epoxy/PG3, epoxy/PG4 and epoxy/PG5 coated samples, respectively. It is obvious that the E corr of all the coating samples had a positive shift compared with bare steel. E corr mainly represents the tendency of corrosion reaction, where a sample with the higher E corr has a better corrosion resistance [45]. Therefore, it can be concluded that the addition of composite material significantly improves its anti-corrosion performance. According to formula (1), the I corr and corrosion rate are in positive proportion to the relationship. Therefore, the sample with a lower I corr value will have the lower corrosion rate. Table 1 indicates that the I corr ranked by numerical value from the largest to smallest: uncoated steel, coatings in epoxy, epoxy/PANI, epoxy/GO, epoxy/PG1, epoxy/PG2, epoxy/PG5, epoxy/PG3 and epoxy/PG4. The corrosion current on the coated surfaces is apparently much lower (approx. 2-5 orders of magnitude) than the bare steel. It is obvious that coatings with epoxy/PG have better corrosion resistance than epoxy/PANI and epoxy/GO, demonstrating an increase in corrosion resistance of the GO/PANI nanocomposites coated substrates. Meanwhile, the decreased amplitude of corrosion current is positively correlated with the increase of GO content, indicating that GO plays a leading role in nanocomposites. It is easy to determine that the coating with PG5 composites have lower anti-corrosion property than the PG4, it may be that the excessive amount of GO leads to the deterioration of polymer dispersion performance, so that the barrier performance deteriorates. It can be seen that the results of electrochemical impedance spectroscopy are in good agreement with the potentiodynamic polarization. In particular, among all of the coatings, the epoxy/PG4 showed the best anti-corrosion performance and the lowest retained CR (5.46 × 10 −5 mm/year).

Conclusions
GO/PANI nanocomposites with unique anti-corrosion properties were synthesized by in-situ polymerization. Structural analyses were conducted through FTIR, UV-vis, XRD, and TGA to confirm that the PANI successfully synthesis on the surface or interlayers of GO. The epoxy/GO/PANI coatings were prepared to improve the corrosion resistance. The corrosion protection properties of the composite coatings in 3.5 wt.% NaCl solution was tested by potentiodynamic polarization and EIS measurements. Among all the coatings with different samples, the epoxy/PG4 coating shows the best anti-corrosion performance, the |Z| 0.01Hz value of which is 12.4 times higher than epoxy/PANI and 55.22 times higher than pure epoxy coating. The composite coating has excellent corrosion resistance because it achieves higher corrosion inhibition efficiency and protection efficiency compared with PANI and epoxy coatings. It is concluded that GO/PANI composite materials can expand the application in the field of anti-corrosion due to their simple preparation, stable performance, and good anti-corrosion effect, etc.