Effects of the Composition and Morphology of Carbon Nanomaterial Additives on the Anticorrosive Properties of Polyvinyl Chloride-Based Paint Coatings

Abstract

The article investigates the role of carbon nanomaterials (CNMs), surface-oxidized carbon nanotubes (CNTs) and few-layer graphene fragments (FGFs), as well as FGFs hetero-doped with N and P atoms, as anticorrosive additives in industrial paints based on polyvinyl chloride. All CNMs were characterized by thermogravimetry, transmission electron microscopy, low-temperature nitrogen adsorption, and X-ray photoelectron spectroscopy. Corrosion resistance was determined using electrochemical tests and impedance spectroscopy. The surface and internal 3D structure of steel and coated steel were visualized using laser confocal microscopy and computed tomography. Coatings containing polyvinyl chloride with 0.05 wt% oxidized CNTs or FGFs show the highest electrochemical resistance and the best anticorrosive properties. The corrosion rate for coatings containing CNMs decreases by an average of 5–7 times compared to uncoated steel. It is shown that the improvement in anticorrosive characteristics is determined by the texture parameters and the composition of CNMs. The pores in CNMs act as a reservoir for the electrolyte and increase the corrosion rate. Oxygen-containing surface groups prevent corrosion by increasing the resistance of the materials.

1. Introduction

The corrosion and fouling of ship bottoms by living organisms are global problems that make their operation and maintenance expensive and environmentally unsafe [1]. To increase service life, reduce maintenance costs, and increase sailing speed, hull surfaces are coated with various coatings that reduce friction with water and prevent corrosion and fouling [2]. Metals, metal oxides, ceramic compositions, and organic compounds are used as coating materials [3]. Organic coatings are usually based on polymers such as polypyrrole [4], polyurethane [4,5,6], epoxy resins [7], polyaniline [8], acrylic polymers [9,10], polyvinyl chloride and its copolymers [11,12], and other compounds [13]. Cu(I), Ag, Au, Zn, Sn, Pb, Ce, Ti, As, and Cr compounds are also introduced into the polymer matrix, which increases the service life of the polymer coating due to their biocidal action and, as a result, reduces biocidal corrosion. However, these compounds are environmentally unsafe and limit the possible effects of surface self-restoration [14,15,16,17,18].

Recently, polymer coatings to which carbon nanomaterials (CNMs) are added, combining protection against corrosion, fouling, and rotting, have become of particular importance [18,19,20,21,22,23]. CNM–polymer composites form an effective insulating barrier due to their hydrophobicity and reduced diffusion rate of salt water to the metal [24,25,26,27]. CNMs adsorb oxygen molecules and radicals, OH⁻, Cl⁻, and SO₄²⁻ ions on their surface, which leads to their decreased concentration in protective coatings and helps mitigate corrosion [28,29]. Theoretically, it has been shown that, in the most general case, the effectiveness of CNMs as an additive is determined (in order of decreasing effect) by their hydrophilicity, particle geometry (morphology), and active surface area [30].

Graphene stabilized with polybutylaniline in an epoxy resin [26], as well as graphene oxide modified with polyaniline [31] or polyvinylpyrrolidone [32], have been described as CNMs. CNTs are also described as fillers for polymer matrices made of polyurethane [33,34,35] or polyvinyl chloride [36,37]. Carbon quantum dots have been proposed for use in polyvinyl chloride compositions [38]. In all cases, such additives provided a significantly reduced (by 5–7 times) corrosion rate compared to coatings without CNMs in their compositions. Among heterosubstituted CNMs, N-doped graphene oxide and N-CNTs have been described as additives in coatings and enamels that reduce the corrosion rate by seven times [39].

CNMs as additives to polymer matrices also contribute to improving the tribological characteristics of coatings [40]. Nanoparticles, especially ones with a 0D structure, such as nanodiamonds, fullerenes, and carbon dots, fill the pores in the polymer structure, making the coatings monolithic [25]. In the case of 1D structures, such as CNTs, the polymers are reinforced, and the coating gains additional mechanical strength [40,41]. Graphene sheets have the ability to form a cross-linked graphene conductive network with bulk conductivity resulting in the displacing of cathodic oxygen reduction sites to graphene from the metal body [25,42]. CNM hydrophilicity is a crucial characteristic of carbon additives resulting in uniform distribution in the polymer. The hydrophilicity is achieved through oxidation or surface modification [43,44]. For example, An and their co-authors oxidized and then fluorinated graphene oxide surface to achieve the balance between anticorrosion and hydrophilic characteristics [45].

The purpose of the current research is to study the effect of the composition and morphology of carbon nanomaterials (surface-oxidized carbon nanotubes, multilayer graphene fragments, and heteroatomic N- and P-atom-doped FGFs) on the anticorrosive properties of composite coatings based on polyvinyl chloride paints.

2. Materials and Methods

2.1. Synthesis of Carbon Nanomaterials

Few-layer graphene fragments (FGFs), nitrogen- and phosphorus-doped FGFs (N-FGFs and P-FGFs, respectively), were obtained by hexane CVD (chemically pure, JSC Reahim, Moscow, Russia), acetonitrile (99.85%, LLC Component Reaktiv, Moscow, Russia), or PPh₃ (98%, Sisco Res. Lab., Mumbai, India) in a toluene solution (99.5%, LLC Component Reaktiv, Moscow, Russia) in the presence of an MgO template at 900 °C, according to [46,47,48]. CNTs were synthesized by hexane CVD in the presence of a Co,Mo/MgO catalyst at 750 °C [49].

Quartz boats with an MgO template or a Co,Mo/MgO catalyst were placed in a tubular quartz reactor (55 mm diameter). N₂ (99.999%, LLC Logika, Moscow, Russia) was purged, after which hexane, acetonitrile, or a PPh₃ solution in toluene was passed through at the previously specified operating temperatures for 30 min. The reactor was then cooled in an N₂ current to room temperature, and the resulting products were boiled with HCl (chemically pure, JSC Reakhim, Moscow, Russia) to remove MgO and dried at 110 °C for 10 h.

The surfaces of FGFs and CNTs were oxidized by boiling them under reflux in a solution of 69% HNO₃ (extra pure 18-4, LLC Component-Reaktiv, Moscow, Russia) for 1.5 h for FGFs and 3 h for CNTs, after which they were filtered, rinsed with water to a neutral pH of the rinsing water, and dried at 110 °C for at least 10 h.

2.2. Physico-Chemical Analysis

The thermal stability of CNTs, the number of carbon phases, and the content of combustible impurities were evaluated using thermogravimetric (TG), differential thermal analysis (DTA), and differential scanning calorimetry (DSC) with a Netzsch STA 449 PC LUXX instrument (NETZSCH-Gerätebau GmbH, Selb, Germany). The samples were heated in atmospheric air from room temperature to 1000 °C at a rate of 10 °C·min⁻¹.

Surface morphology and particle size were studied by transmission electron microscopy (TEM) using a JEOL 2100 F/Cs device (JEOL, Tokyo, Japan) at an operating voltage of 200 kV.

The composition of the CNM surface was determined by X-ray photoelectron spectroscopy (XPS) on an Axis Ultra DLD instrument (Kratos Analytical, Great Britain, Manchester, United Kingdom) using monochromatic AlKα radiation (1486.7 eV). The survey XPS spectra were obtained with an analyzer transmission energy of 160 eV and a step of 1 eV, and the high-resolution spectra were obtained with a transmission energy of 40 eV and a step of 0.1 eV.

Low-temperature adsorption–desorption isotherms were obtained using an Autosorb-1C/QMS device (Quantachrome Inc., Boynton Beach, USA). The BJH model [50] was used to calculate the specific surface area S_BET and the pore size distribution. The samples were evacuated at 300 °C for 3 h before the measurements, after which the measurements were carried out.

2.3. Preparation and Coating of Steel Samples

CNM samples (0.05 wt%) were weighed and mixed with a commercially available vinyl chloride copolymer enamel of the HS–5226 brand (LLC NTC Spetsmal, St. Petersburg, Russia). The mixtures were placed in a glass reactor equipped with an external cooling jacket and treated with ultrasound (IL10-0.63 ultrasound generator, Inlab Ltd., Russia) for 3.5 h. The enamels modified in this way were applied using brushes, followed by leveling the layer with a crossbar to a specified thickness of 100 µm on PCA steel plates (PJSC Novolipetsk Metallurgical Combine, Lipetsk, Russia) measuring 100 × 135 × 4 mm in size. For comparison, HS–5226 enamel without CNMs was applied, forming coatings 50 and 100 µm thick. Before coating, the surface of the plates was sanded with sandpaper and wiped with alcohol. The applied enamel was dried at room temperature for 24 h. The thickness of the coating layer was additionally monitored using an RGK MC-25 micrometer (RGK JSC, Moscow, Russia).

2.4. Electrochemical Measurements

Electrochemical tests were performed in 50 mL three-electrode glass electrochemical cells. A 3.5 wt% NaCl solution in distilled water, simulating seawater, was used as the electrolyte. The working electrode consisted of coated steel plates, the counter electrode was Pt, and the reference electrode was a silver–silver chloride electrode. The area of the working and Pt electrodes was 1 cm². All tests were performed at room temperature.

A Biologic VSP potentiostat/galvanostat (BioLogic Science Instruments, Seyssinet-Pariset, France) was used to measure the voltage characteristics of materials, electrode polarization, and impedance spectroscopy. For corrosion monitoring, the linear polarization resistance method was used, specifically designed to determine the polarization resistance of a material and the corrosion current when passing through potential steps around the corrosion potential. During this procedure, the open-circuit potential (OC) E_oc of the sample was measured for 5–55 min to ensure system stability, determined by an E_oc drift of no more than 2 mV h⁻¹. The sample was then polarized to −30 mV relative to E_oc for 1 min. After that, the potential was scanned from −30 mV to 30 mV relative to E_oc at a rate of 0.1 mV s⁻¹. The values of R_p, E_corr, and i_corr were determined using EC-Lab 11.61 software (BioLogic Science Instruments, Seyssinet-Pariset, France).

Before recording the Tafelian curves, the E_oc value was continuously measured until its drift did not exceed 2 mV·s⁻¹. Next, the curve was recorded in the linear potential scanning mode in the range from −250 mV to 250 mV relative to E_oc at a scanning rate of 0.1 mV s⁻¹. The values of E_corr, i_corr, coefficients b_a, b_c, and corrosion rate were determined using EC-Lab 11.61 software.

Impedance measurements were carried out with a potential amplitude of ±10 mV, 1–10 weeks after the coating was impregnated with 3.5 wt% NaCl, in the frequency range from 50 × 10⁻³ to 10⁶ Hz. The data obtained were processed using OriginPro 2021 software (OriginLab Corporation, Northampton, USA).

2.5. Visualization of Coating Corrosion by Confocal Microscopy and Computed Tomography

Surface visualization before and after electrochemical corrosion etching was performed using an OLYMPUS LEXT OLS-3000 laser confocal microscope (Ryokosha, Ltd., Tokyo, Japan).

Additionally, steel plate samples 2 mm thick were prepared for tomography. The coatings were applied similarly to the method described above. Corrosion was induced by holding these plates for 4 weeks in a 3.5 wt% NaCl solution at room temperature.

Three-dimensional visualization of corrosion of steel plates was performed by computed tomography using an energy-sensitive MARS Bioimaging CT scanner (MARS Bioimaging Ltd., Christchurch, New Zealand) with semiconductor hybrid pixel detectors Medipix3RX (Medipix3 Collaboration, Geneva, Switzerland) with CdZnTe sensors 1 mm thick. Scanning was performed with a tube voltage of 120 kV, a current of 40 µA, an exposure time of 100 ms, an energy threshold of 30 keV, and a projection step of 0.5° (720 projections per revolution). During scanning, constant temperature conditions were maintained, with the detector temperature not exceeding 16 °C. Reconstruction was carried out for an energy window of 30–120 keV with a voxel size of 0.06 mm. The reconstructions were visualized using MARS Vision 2.0 software (MARS Bioimaging Ltd., Christchurch, New Zealand). At least four plates of each type were examined to obtain statistically acceptable results.

3. Results

3.1. Physical and Chemical Characteristics of Carbon Nanomaterials

The carbon phase content in the synthesized CNMs was confirmed by the TG method. In oxidized CNTs, oxidized FGFs, N-FGFs, and P-FGFs, the content of combustible impurities was 2.9, 3.7, 11.5, and 2.8 wt%, respectively (Figure 1).

According to the TEM data, particles of oxidized FGFs, N-FGFs, and P-FGFs are two-dimensional polyhedra consisting of several stacked graphene layers (Figure 2a–f). Oxidized CNMs are multi-walled tubes with an outer diameter of 15 nm and 12–15 carbon layers (Figure 2g,h).

The surface composition of the obtained CNMs was determined by the XPS (

Table 1. Chemical composition, according to the XPS data, specific surface area S_BET and textural characteristics are the total pore volume V and average pore radius R of adsorbents, according to the BJH model.

CNMs	Composition, at. % (XPS)			Textural Characteristics
	C	O	Heteroatom (N or P)	SBET, m2·g−1	V, cm3·g−1	Rpore, nm
Oxidized CNTs	94.0	5.8	0.2 a	181	0.929	24.02
Oxidized FGFs	87.5	11.9	0.6	355	0.856	12.65
N-FGFs b	90.2	1.8	7.8	639	3.543	31.00
P-FGFs	97.3	2.7	<0.1	1207	1.996	5.03

^a Nitrogen is present in the form of surface NO₂⁻, NO₃⁻, and NR₄⁺ species. ^b N-FGFs also contain 0.1 at. % Cl and 0.1 at. % Mg.

). In the high-resolution C1s XPS spectra, maxima corresponding to the C–C bond energies in the sp ² (284.9 eV) and sp ³ (285.2 eV) hybridized states, C–O (286.2 eV), C=O (287.4), and O=C–O (288.8) can be distinguished (Figure 3a). High-resolution O1s XPS spectra contain maxima corresponding to oxygen in the O⁻ (530.2 eV) state, as part of O=C (531.6 eV), O–C, and O–N (532.4 eV) bonds, O=C–O as part of carboxyl groups (533.5 eV), and non-conductive impurities (534.5) (Figure 3b). The high-resolution N1s XPS spectrum of N-FGFs contains maxima corresponding to nitrogen atoms in the pyridine-like state C=N–C (398.2 eV), in amino groups NC₃ (399.5 eV), the substitution state (401.1 eV), NC₄⁺ (402.2 eV), as well as nitrogen bound to oxygen atoms N–O (403.4 eV) and NO₂ (405.0) (Figure 4).

The low-temperature adsorption–desorption isotherms of all CNMs belong to type II (Figure 5), and the materials contain mesopores, according to the IUPAC classification [50]. The pore parameters and specific surface area values are shown in Table 1.

3.2. Anticorrosive Resistance of Coatings

Tafel curves showing the correlation between the overvoltage at the cathode and anode and lg(i) are given in Figure 6.

The corrosion current density (

Table 2. Values of E_corr., i_corr., R obtained by the Tafel method and the stationary potential method, anode and cathode coefficients of Tafel b_a, b_c, calculated corrosion rate values and coating effectiveness P.

Tested Coverage	Tafel Extrapolation Method				Stationary Potential (Calculated Using the Stern–Geary Equation)			Corrosion Rate, v.10−3, mm·year−1	P, %
	Ecorr., mV	icorr., µA	ba, mV	bc, mV	Ecorr., mV	icorr., µA	Rp, Ohms
Uncoated steel	−699.8	0.528	20.1	80.2	-	-	-	6.202	-
Enamel, layer thickness 50 µm	−632.4	0.218	356.9	208.7	−576.8	0.329	173.698	2.560	58.71
Enamel, 100 µm	−388.5	0.099	144.8	646.3	-	-	-	1.162	81.25
Oxidized CNTs	−379.1	0.085	275.5	275.8	−410.9	0.131	455.836	0.998	83.90
Oxidized FGFs	−431.5	0.075	300.6	287.5	−446.7	0.156	407.122	0.881	85.80
N-FGFs	−581.9	0.220	230	507	−592.2	0.721	95.282	2.584	58.33
P-FGFs	−624.6	0.089	251.3	126.2	−543.9	0.256	142.586	1.045	83.14

) was determined graphically by the intersection of the lines obtained by extrapolating the linear regions of the cathode and anode dependencies n = f(i) with the corrosion potential, the point where the overvoltage is zero (Figure S1, Supplementary Materials). Taking into account the well-known Tafel equation:

η_a(c) = a + b_a(c) lg(i_a(c)),

(1)

where η_a(c) is the overvoltage at the anode or cathode, expressed in volts, i_a(c) is the current at the electrode, expressed in amperes, a is the Tafel constant; b_a(c) is the anode or cathode Tafel coefficient [51], the values of the coefficients b_a(c) were determined (Table 2).

The polarization resistance was calculated using the equation (Table 2):

$${\mathrm{R}}_{\mathrm{p}}={\displaystyle \frac{1}{\frac{\mathrm{d}i}{\mathrm{d}\mathrm{E}}}},$$

(2)

where

$$i=i_{corr.}({\mathrm{e}\mathrm{x}\mathrm{p}}^{(\mathrm{l}\mathrm{n}10{\displaystyle \frac{(\mathrm{E}-{\mathrm{E}}_{\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{r}.})}{{\mathrm{b}}_{\mathsf{\alpha }}}}}-{\mathrm{e}\mathrm{x}\mathrm{p}}^{(\mathrm{l}\mathrm{n}10{\displaystyle \frac{(\mathrm{E}-{\mathrm{E}}_{\mathrm{c}\mathrm{o}\mathrm{r}\mathrm{r}.})}{{\mathrm{b}}_{\mathsf{\beta }}}}})$$

(3)

If E = E_corr., then Expression (3) is reduced to the well-known Stern–Geary equation:

$$i_{corr.}={\displaystyle \frac{{\mathrm{b}}_{\mathrm{a}}{\mathrm{b}}_{\mathrm{c}}}{2.303\left({\mathrm{b}}_{\mathrm{a}}{+\mathrm{b}}_{\mathrm{c}}\right){\mathrm{R}}_{\mathrm{p}}}},$$

(4)

where b_a(c) is the anode or cathode Tafel coefficient, determined by Equation (1), and R_p is the polarization resistance. The polarization resistance R_p was calculated as the coefficient of proportionality between current i and voltage E for small values of E (Table 2).

The values of E_corr. and i_corr. obtained from Equations (1) and (4) are different (Table 2), which is consistent with the data [37].

The corrosion rate was calculated using the equation:

$$\mathrm{C}\mathrm{o}\mathrm{r}\mathrm{r}\mathrm{o}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{R}\mathrm{a}\mathrm{t}\mathrm{e}={\displaystyle \frac{0.13i_{corr.}(\mathrm{E}.\mathrm{W}.)}{\mathrm{A}\mathrm{d}}},$$

(5)

where i_corr. is the corrosion current expressed in amperes, E.W. is the equivalent weight equal to 28 g∙equation⁻¹, A is the area equal to 1 cm², and d is the density of steel equal to 7.800 g∙cm⁻³.

The effectiveness of coatings (Table 2) was estimated using the equation:

$$\mathrm{P} = (1– {\displaystyle \frac{i_{corr.}}{i_{corr.}^{\prime }}}) 100,$$

(6)

where i_corr. is the corrosion current of steel, i_’corr. is the corrosion current of steel coated with a film of enamel grade HS–5226 or enamel grade HS–5226 with 0.05 wt% CNMs.

3.3. Impedance Spectroscopy

Impedance spectroscopy is a universal method for examining the metal–coating, metal–electrolyte, and similar interfaces. Experimental impedance hodographs of the obtained coatings are shown in Figure 7.

In the most general case, to study the corrosion of steel plates coated with paints and enamels, calculations were carried out using an equivalent circuit (Figure 8) [52,53,54], taking into account the resistive (R₁) and capacitive (Q₁, constant phase element, CPE) layer parameters. Resistance R₂ refers to pore resistance, while resistance R₃ and capacitance Q₃ (Warburg element) characterize the kinetics of the process associated with the diffusion of the electrolyte to the steel surface. The hodograph can be described by the equation:

$$Z={\displaystyle \frac{1}{\mathrm{A}{(j\mathsf{\omega })}^{\alpha }}},$$

(7)

where A is a preexponential multiplier, which is a frequency–independent parameter, α is an exponent that determines the nature of the frequency dependence (−1 ≤ α ≤ 1), and j = −1 is an imaginary unit.

When calculating the corresponding equivalent scheme for each coating, it was found that it satisfactorily describes the hodographs of all coatings, except for the 50 µm coating sample without CNM additives (Figure S2, Supplementary Materials). Based on these models, the parameters R₁, R₂, R₃, Q₁, Q₃, Q₄ and α₁, α₃, α₄ for each coating were calculated (

Table 3. Values of resistances R, capacitances Q, and exponent α obtained by impedance spectroscopy using the equivalent circuit shown in Figure 8.

Coating	R1, Ohm	Q1·10−6 F·s−α	α1	R2, Ohm	R3, Ohm	Q3·10−6 F·s−α	α3	Q4·10−6 F·s−α	α4
Enamel, 50 µm	0	0.140	0.66	1747	0	65	0.85	81	0.39
Enamel, 100 µm	0	0.137	0.61	7039	0.38	0	0.24	21	0.60
Oxidized CNTs	0	0.523	0.56	3213	4369	72	1	176	0.37
Oxidized FGFs	91	0	0.49	26965	0	0.027	0.71	27	0.77
N-FGFs	0	0.039	0.71	7568	0	13	0.14	530	0.48
P-FGFs	129	0.029	0.71	18357	560	47	0.25	32	0.83

).

4. Discussion

Seawater is an electrolyte due to the salts and oxygen dissolved in it (~8 mg/L), which contribute to the oxidation of steel surfaces, resulting in the formation of rust, oxides, hydroxides, and oxyhydroxide derivatives of iron of variable composition, FeO·Fe₂O₃·nH₂O. Chloride ions destroy and prevent the formation of passivating films of FeO·Fe₂O₃·nH₂O on the metal surface. Applied polymer coatings protect the steel surface from direct interaction with oxygen, water, and salt ions dissolved in water. Uneven or imperfect coatings, as well as possible chipping and damage due to temperature fluctuations, lead to the development of corrosion. Another possible cause of corrosion under a continuous coating layer is the ability of water, oxygen molecules, and electrolyte ions to diffuse through the coating and eventually reach the steel surface [55].

In the most general case, the dissolution of Fe occurs, according to the equation:

Fe⁰ – 2e⁻ → Fe²⁺

(8)

with simultaneous reduction of oxygen atoms dissolved in water:

O₂ + 2H₂O + 4e⁻ → 4OH⁻

(9)

At high external potentials, hydrogen gas is released under electrochemical testing conditions:

2H⁺ + 2e⁻ → H₂↑

(10)

The reaction products of Fe²⁺ and OH⁻, according to Equations (8) and (9), then diffuse back through the coating layer (Figure 9).

For an anticorrosive coating, good adhesion to the steel surface is crucial, since otherwise water and reaction products (7)–(9) can accumulate under the coating, violating the integrity of the interfacial boundary. In extreme cases, bubbles form, leading to peeling and cracking in the coating.

The nature of damage and corrosion in this study was investigated using optical and tomographic methods. Optical and 3D confocal images of the surface of coating samples after electrochemical corrosion tests are shown in Figure 10. Traces of surface grinding are clearly visible on the steel surface (Figure 10a), which do not practically change the height of the steel plate (Figure 10b). After corrosion tests, crater-like defects appeared on the surface—pitting corrosion (Figure 10c,d), associated with the dissolution of Fe, as well as the formation of rust. When steel is coated with films of polyvinyl chloride and/or polyvinyl chloride with CNMs, surface degradation and rust spots occur to a much lesser extent (Figure 10e–p). The morphology and nature of the damage differ from those of unprotected steel. It is likely that electrochemical oxidation products accumulate under the films according to Equations (7)–(9), resulting in swelling and rupture accompanied by cracking of the coating (Figure 10g,k,o). As a result, the electrolyte solution becomes freely accessible to the steel surface, and corrosion develops over the entire damaged area of the coating. It should be noted that the formation of crater-like defects in coatings was previously observed for a large number of polymers and polymer composites with CNMs, as confirmed by scanning electron microscopy [39], confocal microscopy [56], etc.

Similar results were obtained when visualizing corrosion using computed tomography (Figure 11). There were no distinct signs of corrosion on the tomograms of uncoated steel, presumably due to the superficial nature of the process, which did not lead to the formation of localized defects detectable by tomographic methods (Figure 11a,b). In corroded steel samples coated with 0.05 wt% oxidized CNTs, the formation of deep crater-like defects up to 0.5–1 mm in depth was observed, corresponding to a pitting corrosion process (Figure 11c,d). For enameled steel with the addition of 0.05 wt% N-FGFs, the formation of crater-like defects was also observed, but with smaller dimensions and depth (Figure 11e,f).

When water diffuses into the coating, its conductivity changes [36]. It has been experimentally found that the better the coating conducts electric current, the worse its anticorrosive properties [57].

In this paper, the conductivity of coatings was investigated by impedance spectroscopy. The samples were kept in a 3.5 wt% NaCl solution, and the films were completely impregnated for several days up to ~10 weeks, after which the conductivity was measured. Uncoated steel is a conductive material: there is only one semicircle on the impedance hodograph (Figure 7), and the corrosion rate is maximal at 6.202 mm·year⁻¹ (Table 2).

For polymer samples with CNMs, two insufficiently pronounced semicircles with frequency-independent parameters can be distinguished on the impedance hodographs (Figure 7 and Figure S2). This pattern is typical for layered structures, in which two or more interphase boundaries can be distinguished. In this case, these are the electrolyte coating (resistance Q₁ in Figure 8) and steel coating (resistance Q₃). The load Q₄, whose contribution is small, can be attributed to possible inhomogeneities in the film structure associated with the presence of CNM particles in the polymer film and the appearance of resistance at the polymer/CNM interface. The physical meaning of the observed hodographs lies in the fact that at low frequencies, OH⁻ ions diffuse to the steel electrode through the film and then deposit on it with charge transfer to an external circuit [55].

To evaluate the effectiveness of the coating in terms of forming a double electric layer, it is necessary to compare the values of the exponent α, which is the argument of the function Z in Equation (5). In this paper, the exponent of α₁, corresponding to the electrolyte/coating interface, varies from 0.49 to 0.71 (Table 3), which indicates the Warburg diffusion resistance of the element Q₁. When considering the coating/steel interface, only coatings with 0.05 wt% oxidized CNTs showed an α₃ value of 1, which corresponds to an «ideal supercapacitor» [58]. It is for this coating that one of the lowest corrosion rates was found, equal to 0.998 × 10⁻³ mm·year⁻¹ (Table 2).

The Tafel curves for all examined materials are shown in Figure 8. The appearance of all the obtained curves is typical for electrochemical processes during metal corrosion in marine conditions. The more positive the open-circuit potential, the better the corrosion process is inhibited, as previously shown for the addition of silanes to epoxy resins [59]. The highest value of E_corr. = −388.5 mV is typical for a coating without the addition of CNMs (Figure 8, Table 2). Depending on the additive to the polymer coating, the potential E_corr increases in the following range: coating without additives ~ N-FGFs > P-FGFs ~ oxidized CNTs > oxidized FGFs. Accordingly, the current and the corrosion rate decrease in the same order.

To substantiate the effect of CNMs on the anticorrosive properties of coatings, it is necessary to take into account their composition and morphology. Oxidation of the surface of CNMs, as a rule, leads to better dispersibility in polymer compositions [60], although, in general, dispersion of CNM particles in a PVC matrix under ultrasound is a commonly accepted technique leading to positive results [61].

It is known that polyvinyl chloride has a highly organized structure; the introduction of CNMs into its composition promotes reorganization such that the CNMs are not randomly distributed in the polymer [62,63]. The improvement of material properties is achieved due to the strong influence of the developed CNM surface on the ordering of elements in systems with a randomly formed structure. This occurs due to the formation of covalent bonds between CNMs and the polymer [22,39] or due to interactions via van der Waals forces [64].

In this paper, the introduction of surface-oxidized FGFs led to the most significant improvement in anticorrosive properties, which is probably determined by a combination of the textural characteristics of FGFs and the presence of oxygen-containing surface groups (Figure 12). It is well-known that graphene particles occupy pores in the polymer, which typically reduces the water diffusion into the composite volume [25,65]. As the pore volume V_pore in the CNM structure increases, the corrosion rate also increases (Figure 12a). The pores probably act as a reservoir for the electrolyte solution and contribute to its diffusion to the steel surface. At the same time, the corrosion rate decreases with increasing content of surface groups (Figure 12b). Oxygen-containing surface groups improve the adhesion of polymer-CNM composites to steel [66] and reduce the conductivity of both CNMs [67] and the composites containing them [63,68].

5. Conclusions

The development of anticorrosion additives has gone beyond the laboratory and become a routine practice. Currently, majority of paint and enamel manufacturers use CNMs as additives [69,70]. The anticorrosive properties of coatings based on commercially available polyvinyl chloride paints can be significantly improved by adding CNMs. In particular, surface-oxidized CNTs and FGFs, as well as FGFs hetero-doped with N and P atoms, introduced in an amount of 0.05 wt%, reduced the corrosion rate of steel by 5–7 times compared to the uncoated surface.

Impedance spectroscopy showed that, in the most general case, the coating contributes to the formation of two interfaces: electrolyte/coating and coating/steel. The highest resistance and, as a result, the lowest corrosion rate were observed for the coating with 0.05 wt% surface-oxidized CNTs. Oxygen-containing surface groups contribute to a decrease in the conductivity of CNMs and polyvinyl chloride/CNM composites in general. Heteroatoms in FGFs lead to the formation of different numbers of surface groups; therefore, samples with a higher content of surface groups are the most effective anticorrosive additives. In addition to surface modification, the porosity of the CNMs itself is critically important. These pores act as a reservoir for water and may contribute to corrosion. Therefore, we recommend choosing CNMs with micropores.