Variable corrosion behavior of a thick amorphous carbon coating in NaCl solution

Abstract

The present work investigates a thick amorphous multi-layer carbon coating fabricated by a plane hollow cathode plasma-enhanced chemical vapor deposition technique. The thick amorphous multi-layer carbon coating included a F–Si doped multi-layer structure, and a silicon interlayer which was able to reduce internal stress, improve the bonding and adhesion, and bridge the film and substrate. This work mainly discusses the cause of the variable corrosion behavior of the coating. Corrosion resistance was assessed by potentiodynamic polarization tests and electrochemical impedance spectroscopy (EIS) in 3.5 wt% NaCl solution. The results showed that the mild electrochemical reactions and corrosion product film had a significant effect on determining the variable corrosion behavior of the coating.

---

1. Introduction

In the last few years, it has been shown that although TiN, CrN and polymers can improve corrosion resistance of a substrate,^1–3 ultimately, they were prone to failure in corrosive environments. However, diamond-like carbon (DLC) coatings have been extensively applied in electronic devices and vehicles due to their excellent corrosion resistance in corrosive environments.^4–12 It is well known that the corrosion resistance of DLC coatings is mainly determined by their chemical and physical properties.^13,14 Significant methods have been developed to improve the corrosion resistance of DLC coatings over the past years. For instance, incorporation of foreign atoms was found to be an effective way to increase the intrinsic properties of host materials. Thus, incorporation of foreign atoms into DLC coatings has been shown to have the potential to improve the corrosion resistance of these coatings.^15,16 However, there were many aspects restricting the improvement of corrosion resistance of DLC coatings in corrosive media. Defects such as nanopores are direct paths allowing the penetration of water, environmental oxygen and ions, and corrosive media, which would lead to the electrochemical dissolution of the substrate.^4,17

In addition, the corrosion resistance of DLC coatings could be well improved by increasing the thickness of the coatings.^10,18,19 Nevertheless there exists an inevitable inadequacy which has to be taken into account: internal stress would increase in in situ deposition processes with increasing thickness of the coatings, which limited the coating thickness to a range between 1 and 3 μm,²⁰ as a coating thickness more than 3 μm would result in cracking of the coating from the substrate. Therefore, reducing the nanopores and increasing the coating thickness were two keys to improving the corrosion resistance of the DLC coatings.

A possible method to solve these problems well is through fabricating a multi-layer coating (MLCC),^21,22 since they could not only increase the thickness of the coatings, but also reduce the intrinsic internal stress and the possibility of through-coating defects.^20,23 Moreover, the MLCC containing alternating interlayers and few nanopores not only exhibited good corrosion resistance in corrosive solution due to its better ability of corrosion prevention,^24–26 but also low internal stress and high adhesion as reported in our previous studies.^27,28 As we know, F or Si doped amorphous carbon coatings display superior corrosion resistance in corrosive environments as reported by most previous studies.^29,30 Arguably, F and Si codoped amorphous carbon coatings should exhibit good corrosion resistance in corrosive environments. Therefore, to design a simple and durable MLCC would be desirable for improving the corrosion resistance of the amorphous carbon coating.

Herein, the objective of the present work was to fabricate a novel MLCC using a plane hollow cathode plasma-enhanced chemical vapor deposition technique, which was aimed at improving the corrosion resistance of the thick MLCC. Interestingly, the results of electrochemical tests showed that the coating exhibited variable corrosion behavior in NaCl solution. The detailed discussion reveals that the variable corrosion behavior was attributed to the mild electrochemical reactions and corrosion product film.

2. Experimental procedure

Fig. 1a presents a schematic of the plane hollow cathode plasma-enhanced chemical vapor deposition system. Two parallel plates, 7.5 cm apart and made of stainless steel, served as the electrode bodies, and the lower one also served as a substrate holder. The substrates used for coating deposition were AISI 304 stainless steel plates (30 mm × 30 mm × 1 mm) that were polished to a mirror finish. The stainless steel substrates were ultrasonically cleaned in acetone and alcohol for 30 min and dried by N₂ gas blowing. Then they were placed in a vacuum chamber and the chamber was pumped down to a pressure of 1.0 × 10⁻³ Pa using a composite molecular pump. The deposition process for the MLCC covered with a DLC top-layer is shown in Fig. 1b. To grow these coatings, substrates were cleaned at a pressure of 1.5 Pa for 20 min with a constant flow of argon gas in order to remove the oxides. Then, a Si interlayer of about 0.2 ± 0.03 μm was deposited with SiH₄ gas of 50 sccm and Ar gas of 100 sccm (−15.0 kV bias voltage obtained from a high voltage power (voltage range: 0 to 20 kV), 15 Pa and 30% duty ratio) in order to improve the adhesion of the final coatings to the substrate. A multi-layer coating was deposited in a SiH₄, CF₄, C₂H₂ and Ar environment. The F_x1–Si_y1-DLC layers were deposited in SiH₄ (25 sccm), CF₄ (25 sccm), C₂H₂ (150 sccm) and Ar (100 sccm) at 4.0 Pa. The F_x2–Si_y2-DLC layers were deposited in SiH₄ (25 sccm), CF₄ (25 sccm), C₂H₂ (100 sccm) and Ar (100 sccm) at 2.8 Pa. Finally, a pure DLC layer was deposited on the F–Si-DLC layer surface. The pure DLC layer was deposited in C₂H₂ (150 sccm) and Ar (100 sccm) gases by the same deposition system. The substrate bias voltage was maintained at −800 V derived from low voltage power (voltage range: 0 to 1.5 kV), a duty cycle of 30%, and a repetition frequency of 1.5 kHz. No external heating of the substrate was employed, and the maximum temperature during deposition was about 180 °C.

Fig. 1 Schematic of (a) the PECVD deposition system and (b) the deposition process using the hollow cathode effect.

The surface and fracture cross-sectional microstructure of the coating was obtained using a thermal field electron emission scanning electron microscope (JSM 6701F, FEI Quanta FEG 250). TEM images were examined in detail by high-resolution transmission electron microscopy (HRTEM; TF20). The chemical composition of the coating was determined using time-of-flight elastic recoil detection analysis (TOF-ERDA). The chemical compositions of the original coating and after 5 min of OCP testing were examined using X-ray photoelectron spectroscopy (XPS: AXIS ULTRA DLD). The residual stress was measured by stress-induced bending on an interferometer surface profiler. The curvature radii of the substrate before and after coating deposition were measured by the observation of Newton’s rings using an optical interferometer system, and then the residual stress was calculated by the Stoney equation. The adhesion of the sample was tested by a scratch tester (CSEM Revetest) equipped with a diamond tip of radius 200 μm. The normal load was increased from 0 to 50 N at a loading rate of 50 N min⁻¹ and a scratching speed of 10 mm min⁻¹. Electrochemical tests were carried out using a computer controlled potentiostat/frequency response analyzer (Autolab PGSTAT302N) to evaluate the corrosion behavior of bare, and carbon coated, AISI 304 stainless steel. A typical three electrode cell, consisting of the working electrode (0.5 cm² exposed area), a saturated Ag/AgCl electrode (saturated with KCl) as a reference electrode, and platinum as the counter electrode, was used in the corrosion tests. The corrosive medium was a 3.5 wt% NaCl solution. All the solutions were prepared from deionized water with a pH of around 6.8 ± 0.2. Potentiodynamic polarization tests were carried out at scan rate of 0.5 mV s⁻¹ from −160 mV with reference to the open circuit potential (OCP) to a final anodic current density of 0.1 mA cm⁻² after an initial 30 min exposure to the test electrolyte for achieving a stabilized OCP.^31–33 Measurements of electrochemical impedance (EIS) were conducted at the open circuit potential with an AC amplitude of 10 mV after immersion of a sample into solution for 30 min. The frequency ranged from 0.01 Hz to 10⁵ Hz.^34,35 The material used for this study was commercial AISI 304 stainless steel, the chemical composition of which is listed in Table 1. The data was reported by Dagbert.³⁶ The polarization resistance (R_p) values were calculated using the following formula:³⁷

where R_p is in kΩ cm²; β_a and β_c are in terms of mV per dec; and i_corr is in mA cm⁻².

Table 1 The chemical composition (% weight) of AISI 304 stainless steel

Composition	C	Cr	Ni	Mn	N	S	Fe
Wt%	0.047	18.27	8.66	1.19	0.078	0.0007	Balance

---

Ahn reported that the porosity of protective coatings was an important factor for effective corrosion protection.³⁸ The porosity of the protective coatings can be calculated according to eqn (2).³⁹

A quantitative measurement of the effect of defects (pores or pinholes) on coating density is the packing factor (P). According to ref. 37, the sum of the porosity and the packing factor is 1.

3. Results

3.1. Microstructure and composition

The cross-sectional and surface morphologies of the MLCC are illustrated in Fig. 2a and b. The typical cross-sectional morphology (Fig. 2a) reveals a coating with a total thickness of about 8 ± 0.2 μm, including a 0.2 ± 0.03 μm silicon interlayer, 0.2 ± 0.02 μm F_x1–Si_y1-DLC layers, 0.5 ± 0.04 μm F_x2–Si_y2-DLC layers and a 0.5 ± 0.03 μm DLC layer. The surface morphology presented in Fig. 2b exhibits no micropores and cracks in the surface. In order to further investigate the microstructure of the coating, TEM was utilized. Fig. 3a shows the TEM cross-sectional interface image of the MLCC. It can be obviously observed that a compact multi-layer structure is found in the coating. Fig. 3b shows an HRTEM image of the surface coating, which exhibits an amorphous carbon network structure according to the SAED patterns.

Fig. 2 (a) SEM cross-sectional image of the thick multi-layer carbon coating. (b) Surface morphology of the thick multi-layer carbon coating.

Fig. 3 (a) The TEM image of the interface morphology of the thick multi-layer carbon coating. (b) High resolution TEM (HRTEM) image of the surface morphology of the thick multi-layer carbon coating (inset is the SAED pattern).

The elemental depth profiling of the coating examined by TOF-ERDA is shown in Fig. 4. Only the six top layers (total thickness of 2 μm) of the cyclical coating are detected due to the TOF-ERDA only getting a signal from the first 1 to 2 μm beneath the surface. The MLCC with a top DLC layer consists of 7.95 at% H, 0.11 at% O, 1.46 at% F, 0.78 at% Si and the remainder C. C is the major component, indicating that it is a typical carbon coating. In the scratch method, as shown in the typical scratch curve, the peeling-off value merely meant that the coating detached from the substrate, which was obviously the adhesion failure mode.⁴⁰ Fig. 5 shows the typical scratch curve of the MLCC, the critical adhesion load is more than 20 N, indicating a good adhesion between the coating and substrate, which is attributed to the low internal stress of the coating (about −0.5 GPa). The optical image (inset in Fig. 5) of the scratch trace for the MLCC shows that no chipping is observed at the border of inside the scratch of the MLCC until achieving the maximum load. The phenomenon suggests exceptional adhesion between the coating and substrate.

Fig. 4 Elemental depth profiling of the thick multi-layer carbon coating examined by TOF-ERDA.

Fig. 5 The scratch curve of the thick multi-layer carbon coating.

3.2. Corrosion behavior

Fig. 6 shows the potentiodynamic polarization curves of AISI 304 stainless steel and the simple DLC coating after 30 min exposure in 3.5 wt% NaCl. Although the i_corr of the simple DLC coating is larger than that of the bare 304 stainless steel, the value of E_corr of the simple DLC coating is more negative compared to the bare 304 stainless steel. Fig. 7 shows the SEM image of the simple DLC coating after the potentiodynamic polarization test. The coating peels off the substrate after the potentiodynamic polarization test, which indicates the poor corrosion resistance of the simple DLC coating. Thus, in this work, we mainly discuss the corrosive behavior of the thick multi-layer carbon coating. The open circuit potentials (E_OC) measured as a function of immersion time for the bare steel and thick multi-layer carbon coating are presented in Fig. 8. It is worth noting that the E_OC values for the thick multi-layer coating are lower than those for the bare steel after 5 min of immersion in the 3.5 wt% NaCl solution, then rapidly increase to become more positive than those of the bare steel. After 30 min of immersion, the E_OC values for the thick multi-layer coating are more positive than those for the bare steel. Potentiodynamic polarization curves of AISI 304 stainless steel and the MLCC after 30 min exposure in 3.5 wt% NaCl are presented in Fig. 9. The corrosion potential (E_corr) and the corrosion current density (i_corr) of the specimens derived from polarization curves are listed in Table 2. It can be seen that the E_corr and i_corr of 304 stainless steel is −0.149 V and 7.5 × 10⁻⁶ A cm⁻², respectively. For the MLCC, the E_corr (−0.133 V) shifts to the positive direction by about 0.016 V and i_corr (5.1 × 10⁻⁹ A cm⁻²) decreases by more than three orders of magnitude compared to the AISI 304 stainless steel. The anodic (β_a) and cathodic (β_c) Tafel slopes are determined at the same time. Calculated polarization resistance (R_p) values are listed in Table 2. It is noted that the polarization resistance of the MLCC increases by more than three orders of magnitude compared to the 304 stainless steel. The large value of R_p corresponded to excellent corrosion resistance in the potentiodynamic polarization test.⁴ The potentiodynamic polarization test result is a clear indication that the MLCC with a top DLC layer displays superior corrosion resistance in 3.5 wt% NaCl. In addition, the porosity and packing factor of the coating is calculated using eqn (2) and the values are given in Table 2. The beneficial effect of a greater packing factor was seen by the inhibition of the passage of the corroding solution to the substrate and the reduction of localized corrosion kinetics as demonstrated by a previous study.³⁷

Fig. 6 Open circuit potential vs. immersion time in 3.5 wt% NaCl solution for bare steel and the thick multi-layer carbon coating.

Fig. 7 Potentiodynamic polarization curves for 304 stainless steel and the simple DLC coating in 3.5 wt% NaCl.

Fig. 8 The SEM image of the simple DLC coating after the potentiodynamic polarization test in 3.5 wt% NaCl.

Fig. 9 Potentiodynamic polarization curves for 304 stainless steel and the thick multi-layer carbon coating in 3.5 wt% NaCl.

Table 2 Results of porosities and packing factors obtained from electrochemical experiments

Specimen	Ecorr (V)	icorr (A cm^−2)	βa (V per dec)	βc (V per dec)	Rp (Ω cm^2)	Porosity (α)	Packing factor (P)
Steel	−0.149	7.5 × 10^−6	0.418	0.354	1.1 × 10^4	—	—
Thick multiple-layer carbon coating	−0.133	5.1 × 10^−9	0.128	3.6	1.1 × 10^7	0.00078	0.99922

---

3.3. Electrochemical impedance spectroscopy

Electrochemical impedance spectroscopy (EIS) was employed to investigate the corrosion characteristics of AISI 304 stainless steel and the MLCC in 3.5 wt% NaCl solution. The resulting EIS plots of bare steel and the MLCC are shown in Fig. 10. Based on the EIS plots, appropriate equivalent circuits were proposed as shown in Fig. 11. R_s accounts for the solution resistance, R_i and R_ct can be assigned respectively to the pore resistance of the coating and the charge transfer resistance. CPE₁ and CPE₂ are the constant elements of the coating and electrical double layer (EDL), respectively. Constant phase elements are utilized here instead of pure capacitances because of the deviations from an ideal capacitive behavior. The CPE impedance may be calculated using:

Z_CPE = A(jω)⁻ⁿ (3)

where A is a constant and n is defined as the formula:⁴¹

n = 1 − 2χ/180 (4)

where χ is the depression angle (in degrees) that evaluated the semicircle deformation. In many cases, this impedance element was introduced formally only for fitting impedance data. But the CPE behavior has sometimes been attributed to the fractal nature of the electrode interface.⁴² The factor n, defined as a CPE power, is an adjustable parameter that always lies between 0.5 and 1. When n = 1, the CPE describes an ideal capacitor. For 0.5 < n < 1, the CPE describes a distribution of dielectric relaxation times in frequency space, and when n = 0.5 the CPE represents a Warburg impedance with diffusion character.⁴³ Fig. 10a and b show the Nyquist diagrams for the MLCC and bare steel. It shows that the impedance values of the multi-layer carbon coating are much larger than those for the bare steel. For the multi-layer carbon coating, a high impedance modulus (10⁷ Ω cm²) and two time constants are observed compared to the bare steel. Zhao et al. reported that the semicircle at the high frequencies indicated that the coating could be treated as a barrier to corrosive media.⁴⁴ Thus, the semicircle (Fig. 10a) at high frequencies revealed that the MLCC could function as a barrier that made interfacial charge transfer difficult; Montemor et al.⁴⁵ declared that the semicircle at high frequencies was characterized by a capacitive response. The corresponding Bode diagrams for the MLCC are shown in Fig. 10c and d. In Fig. 10c, in the high frequency regions (10³ to 10⁵ Hz), the relationship between impedance and frequency is almost linear with a slope close to −1. This capacitive behavior is related to the EDL at the corrosive solution/coating interface. The phase angle Bode plots in Fig. 10d show a capacitive response in the high frequency domain, which could be linked to the barrier properties of the multi-layer carbon coating.⁴⁶ There is a resistance response in the low frequency range, which is a consequence of formation of conductive pathways through the coating. Table 3 shows the impedance parameters of the corresponding equivalent circuits to fit the impedance data of 304 stainless steel and the MLCC after being exposed to 3.5 wt% NaCl. The larger values of R_ct and R_i had been conducive to improving the resistance to corrosion and decreasing the corrosion rate.^5,47 The results of EIS tests are consistent with the obtained results from potentiodynamic polarization tests.

Fig. 10 Experimental (a) and (b) Nyquist and (c) and (d) Bode plots of 304 stainless steel and the thick multi-layer carbon coating in 3.5 wt% NaCl solution.

Fig. 11 The equivalent circuits used to fit the impedance data of (a) the bare steel and (b) the thick multi-layer carbon coating in 3.5 wt% NaCl solution.

Table 3 Equivalent circuit data of the thick multi-layer carbon coating and 304 stainless steel in 3.5 wt% NaCl solution, and the respective fitting parameters obtained using ZView2

	CPE1-P or n1	CPE1-T (F cm^−2 s^(n−1))	Ri (Ω cm^2)	CPE2-P or n2	CPE2-T (F cm^−2 s^(n−1))	Rct (Ω cm^2)
304 stainless steel	—	—	—	0.84	2.4 × 10^−5	4.8 × 10^3
Thick multiple-layer carbon coating	0.68	7.1 × 10^−8	3.4 × 10^6	0.98	4.0 × 10^−7	3.1 × 10^8

---

The corroded surface (0.5 cm² area) of the MLCC after EIS tests in 3.5 wt% NaCl is shown in Fig. 12. The high magnification SEM image depicted in Fig. 12a shows that no evident corrosion damage is found on the MLCC surface, and the coating has not been decomposed, attributed to the strong adhesion between the coating and substrate. The EDS analysis was carried out at the region shown in Fig. 12b and the results (Fig. 12c–g) show that the materials on the coating surface mainly consist of C, Na and Cl, indicating that the coating provides superior protection to the steel substrate in the short term.

Fig. 12 Surface appearances of the thick multi-layer carbon coating in 3.5 wt% NaCl solution: (a) high magnification SEM morphology of the thick multi-layer carbon coating; (b) high magnification SEM image showing the localized corrosion of the thick multi-layer carbon coating. The elemental EDS maps taken from the whole areas shown in (b) are shown for (c) C K, (d) Si K, (e) F K, (f) Cl K and (g) Na K.

3.4. Long-term electrochemical impedance spectroscopy

Generally, the corrosion behavior of the coating obtained from short-term EIS tests does not well illustrate that the coating has really good or bad corrosion resistance in corrosive environments. Accordingly, long-term EIS tests are supposed to further understand the corrosion behavior of the MLCC. EIS diagrams with bare steel and the multi-layer carbon coating obtained with different times of immersion in 3.5 wt% NaCl are presented in Fig. 13 and 14. For the bare steel (Fig. 13), a low impedance modulus (10³ kΩ cm²) and only one time constant is observed throughout the plot, which is attributed to the formation of a passivating film. The increment in the corrosion resistance of the bare steel was ascribed to the passivating film.⁴⁸ For the multi-layer carbon coating (Fig. 14), the total impedance of the system is above 10⁴ kΩ cm², being independent of the immersion time. In the low frequency regions (Fig. 14a and b), the linear part, due to control of the diffusion process, has a particularity for the multi-layer carbon coated work electrode. At the initial 2.5 h of immersion, the plot reveals a capacitive response in the high frequency domain, which could be related to the barrier properties of the multi-layer carbon coating.⁴⁹ In the low frequency regions, there is a resistance response, which is a consequence of formation of conductive pathways through the coating. Fe is detected on the coating surface, indicating the occurrence of mild corrosion reactions between the corrosive solution and substrate, as shown in Fig. 15. It suggests that the corrosion processes occur in the amorphous carbon coated steel substrate. As the immersion time elapsed, after 32.5 h of immersion, all the spectra present similar behavior, showing that the impedance values are almost constant.

Fig. 13 EIS plots (experimental) and the fitting curves (solid lines) for bare steel: (a) Nyquist plots, and (b) and (c) Bode plots at different times of immersion in 3.5 wt% NaCl solution.

Fig. 14 EIS plots (experimental) and the fitting curves (solid lines) for thick multi-layer carbon coating: (a) and (b) Nyquist plots and (c) and (d) Bode plots at different times of immersion in 3.5 wt% NaCl solution.

Fig. 15 (a) Secondary electron imaging (SEI) of the surface of the thick multi-layer carbon coating immersed for 2.5 h in 3.5 wt% NaCl. (b) High magnification SEM image showing the localized corrosion of the thick multi-layer carbon coating. The elemental EDS maps taken from the whole areas shown in (b) are shown for (c) C K, (d) Si K, (e) F K, (f) Cl K, (g) Na K and (h) Fe K.

For the multi-layer carbon coating, although the overall process is expected to be dynamic, a single circuit can be used as the EIS response depends upon the duration of immersion time. For the equivalent circuits depicted in Fig. 16, R_s is the solution resistance, R_f represents the passivating film resistance, R_i is suggested to represent the pore resistance, and R_pf is the corrosion products film resistance. CPE₀, CPE₁ and CPE₂ are the constant elements of the passivating film, multi-layer carbon coating and corrosion products film, respectively. W_o is a Warburg element, producing a Warburg impedance, R_w. The Warburg element W_o represents the linear diffusion to the reduced electrode surface.^49,50 The circuit elements calculated from the fitting results of the bare steel and the MLCC are summarized in Table 4 and 5, respectively. Interestingly, from Table 5, in the range from 2.5 to 12.5 h of immersion, the decrease of the R_i of the multi-layer carbon coating resulted from the increase of immersion time due to the formation and growth of new pores.⁵¹ While beyond 12.5 h of immersion, the R_i increases to a maximum value of 49.7 kΩ cm² and the R_pf still remains a high value in Table 5, which was attributed to the compact structure allowing the corrosion products to plug the micro-corrosion holes more efficiently,⁵² as can be seen in Fig. 15. Importantly, the coating is closely adhesive to the steel substrate, as shown in Fig. 17.

Fig. 16 The equivalent electrical circuits for the impedance plots fitting of (a) the bare steel and (b) the thick multiple-layer carbon coating immersion in 3.5 wt% NaCl.

Table 4 EIS parameters corresponding to the equivalent circuit for the bare steel in 3.5 wt% NaCl after different times of immersion at the open circuit potential. The respective fitting parameters were obtained using ZView2

Time of immersion, hours	CPE0-P or n0	CPE0-T (F cm^−2 s^(n−1))	Rf (Ω cm^2)	CPE2-P or n2	CPE2-T (F cm^−2 s^(n−1))	Rct (Ω cm^2)
2.5	0.85	5.9 × 10^−6	1.4 × 10^4	0.54	6.1 × 10^−6	3.7 × 10^6
7.5	0.82	7.0 × 10^−6	2.22 × 10^5	0.71	7.7 × 10^−6	1.4 × 10^6
12.5	0.83	5.0 × 10^−6	3.83 × 10^5	0.63	5.3 × 10^−6	5.2 × 10^7
17.5	0.83	5.3 × 10^−6	2.5 × 10^5	0.73	5.5 × 10^−6	2.4 × 10^6
22.5	0.84	5.1 × 10^−6	2.25 × 10^5	0.74	5.2 × 10^−6	2.3 × 10^6
27.5	0.88	2.9 × 10^−6	7.45 × 10^4	0.54	3.0 × 10^−6	5.2 × 10^7
32.5	0.89	2.5 × 10^−6	5.0 × 10^3	0.55	2.7 × 10^−6	2.4 × 10^7

---

Table 5 EIS parameters corresponding to the equivalent circuit for the thick multiple-layer carbon coating in 3.5 wt% NaCl after different times of immersion at the open circuit potential. The respective fitting parameters were obtained using ZView2

Time of immersion, hours	CPE1-P or n1	CPE1-T (F cm^−2 s^(n−1))	Ri (Ω cm^2)	CPE3-P or n3	CPE3-T (F cm^−2 s^(n−1))	Rpf (Ω cm^2)	RW (Ω cm^2)
2.5	0.94	2.4 × 10^−9	2.02 × 10^4	0.64	2.8 × 10^−8	2.9 × 10^6	6.6 × 10^5
7.5	0.95	2.2 × 10^−9	5.5 × 10^3	0.63	9.8 × 10^−8	3.8 × 10^6	5.5 × 10^5
12.5	0.98	9.4 × 10^−11	1.4 × 10^3	0.66	9.5 × 10^−8	2.9 × 10^6	1.9 × 10^6
17.5	0.76	2.7 × 10^−8	4.97 × 10^4	0.63	9.7 × 10^−8	3.1 × 10^6	1.8 × 10^6
22.5	0.83	1.2 × 10^−8	1.55 × 10^4	0.64	1.5 × 10^−7	2.4 × 10^6	7.0 × 10^6
27.5	0.82	1.2 × 10^−8	1.27 × 10^4	0.64	1.7 × 10^−7	3.2 × 10^3	3.9 × 10^6
32.5	0.89	4.8 × 10^−9	3.0 × 10^3	0.66	1.8 × 10^−7	3.5 × 10^6	7.4 × 10^3

---

Fig. 17 The SEM image of between the coating and steel substrate after immersion for 32.5 h.

4. Discussion

A multi-layer amorphous carbon coating was successfully fabricated using a simple deposition technique. Generally, a multi-layer carbon coating could be expected to improve the corrosion resistance due to: (a) increasing coating thickness, which statistically reduces the possibility of through-coating defects such as pores; (b) alternating interlayers leading to different electrical behavior, which could redirect the current flow between coating and substrate, as reported by most previous investigations.^23,24 In this paper, the thick multi-layer carbon coating exhibits variable corrosion behavior. The interesting corrosion behavior is well discussed as follows.

The OCP of the thick multi-layer carbon coating exhibited variable behavior. We analyzed the chemical composition of the coating after 5 min of OCP testing. Fig. 18 shows the (a) XPS C 1s and (b) Si 2p spectra of the original coating and after the 5 min OCP test. Fig. 18a shows that the intensity of the C1s spectrum decreases after the 5 min OCP test, but the composition did not change. Fig. 18b exhibits that the intensity of the Si 2p spectrum changes from a low value to a large one after the 5 min OCP test. Moreover, the Si–O and Si–C bonds group sharply increases after the 5 min OCP test. Bunker et al.⁵³ reported that the relative reactivities of strained and unstrained Si–O bonds show that bond strain promoted bond rupture reactions that led to stress corrosion cracking. Thus, the potential drop after ∼5 min of OCP testing for the thick multi-layer carbon coating should be attributed to this reason. In addition, Maguire⁵⁴ reported that an amorphous carbon film containing Si–C bonds could exhibit good corrosion resistance. Thus, the increase in potential after the 5 min OCP test could be attributed to the effect of Si–C bonds.

Fig. 18 The XPS spectra, (a) C 1s and (b) Si 2p, for the original coating and after 5 min of OCP testing.

Alternated interlayer structures are found in the multi-layer amorphous carbon coating as shown in Fig. 2a and 3a. As reported in our previous investigations,^27,28 these structures could improve adhesion load between coating and substrate and reduce internal stress of the coating. Reducing internal stress and increasing adhesion force was an effective approach to improve the resistance to corrosion.^55–59 Based on a survey of this literature, in this investigation, an amorphous carbon coating containing a F–Si-doped multi-layer structure and possessing low internal stress (−0.5 GPa) and high critical load (>20 N) should exhibit good corrosion resistance in 3.5 wt% NaCl solution. However, Fig. 14 shows that after 2.5 h of immersion, the corrosion resistance of the thick multi-layer carbon coating tends to gradually decrease. As described in Fig. 15, Fe and small corrosion products are detected on the coating surface, indicating that mild corrosion reactions are observed for the amorphous carbon coated steel with an immersion time of 2.5 h in 3.5 wt% NaCl solution. We can therefore conclude that the decrease of corrosion resistance could be attributed to the mild corrosion reactions. It can also be seen from Fig. 17 that the coating does not peel off from the substrate, which strongly indicates adhesion between the coating and substrate. Thus, the mild corrosion reactions would be one of the reasons that lead to the variable corrosion behavior.

In addition, corroding solution penetration generally took place through real microscopic pores and virtual pores, which were regions of low cross-linking, and therefore high transport. A highly dense cross-linked structure made the multi-layer carbon coating less permeable to the corrosive medium leading to less delamination along the coating/substrate interface, since the coating was less permeable, as reported by Liu.²³ The multi-layer carbon coating possesses a high packing factor value (Table 2), thus, in the absence of macroscopic pores or other easily accessible defects for the corroding solution, the pore resistance R_i originates from the actual ionic conductivity of the multi-layer carbon coating. Papakonstantinou reported that the high value of pore resistance R_i of a DLC film reflected the large degree to which the film formed a barrier that hinders electrolytic conduction.⁵ In the range from 2.5 to 12.5 h, the decrease of R_i indicated the formation and growth of new pores, as reported by Zheludkevich.⁵¹ However, in the range from 17.5 to 32.5 h, the decrease of R_i indicates that the ability to inhibit corrosion by the solution becomes weak. Nevertheless, the coating still displayed good corrosion resistance after an immersion time beyond 12.5 h, which was attributed to the compact structure of the coating allowing the corrosion rust to plug the micro-corrosion holes more efficiently,⁵² leading to high R_pf values (Table 5) that could still improve the corrosion resistance of the coating. Thus, the above observation and discussion brings out clearly the fact that the corrosion product film has a significant effect on the corrosion resistance of the coating in 3.5 wt% NaCl solution, which could be another reason for the variable corrosion behavior.

5. Conclusions

A simple DLC coating and a thick multi-layer carbon coating are successfully fabricated by a plane hollow cathode plasma-enhanced chemical vapor deposition method. Compared to the simple DLC coating, the thick multi-layer carbon coating exhibits good corrosion resistance. Interestingly, the EIS results of the thick multi-layer carbon coating demonstrate that the thick multi-layer carbon coating exhibits variable corrosion behavior in 3.5 wt% NaCl solution. The variable corrosion behavior is discussed in detail, which shows that the variable corrosion behavior is attributed to the effects of the mild electrochemical reactions and the corrosion product film.

References

1. R. Bayón, A. Igartua, X. Fernández, R. Martínez, R. J. Rodríguez, J. A. García, A. De Frutos, M. A. Arenas and J. de Damborenea, Corrosion-wear behaviour of PVD Cr/CrN multilayer coatings for gear applications, Tribol. Int., 2009, 42, 591–599.
2. L. Major, J. M. Lackner and B. Major, Bio-tribological TiN/Ti/aC: H M\multilayer coatings development with built-in mechanism of controlled wear, RSC Adv., 2014, 4, 21108–21114.
3. J. Zhu, S. Wei, I. Y. Lee, S. Park, J. Wills, N. Haldolaarachchiqe, D. P. Young, Z. P. Luo and Z. H. Guo, Silica stabilized iron particles toward anti-corrosion magnetic polyurethane nanocomposites, RSC Adv., 2012, 2, 1136–1143.
4. N. W. Khun, E. Liu and X. T. Zeng, Corrosion behavior of nitrogen doped diamond-like carbon thin films in NaCl solutions, Corros. Sci., 2009, 51, 2158.
5. P. Papakonstantinou, J. F. Zhao, P. Lemonine, E. T. McAdams and J. A. McLaughlin, The effects of Si incorporation on the electrochemical and nanomechanical properties of DLC thin films, Diamond Relat. Mater., 2002, 11, 1074.
6. H. G. Kim, S. H. Ahn, J. G. Kim, S. J. Park and K. R. Lee, Corrosion performance of diamond-like carbon (DLC)-coated Ti alloy in the simulated body fluid environment, Diamond Relat. Mater., 2005, 14, 53.
7. A. Zeng, E. Liu, S. N. Tan, S. Zhang and J. Gao, Cyclic voltammetry studies of sputtered nitrogen doped diamond-like carbon film electrodes, Electroanalysis, 2002, 14, 1110.
8. W. Cai and J. H. Sui, Effect of working pressure on the structure and the electrochemical corrosion behavior of diamond-like carbon (DLC) coatings on the NiTi alloys, Surf. Coat. Technol., 2007, 201, 5194.
9. C. L. Liu, D. P. Hu, J. Xu, D. Z. Yang and M. Qi, In vitro electrochemical corrosion behavior of functionally graded diamond-like carbon coatings on biomedical nitinol alloy, Thin Solid Films, 2006, 496, 457.
10. R. Sharma, P. K. Barhai and N. Kumari, Corrosion resistant behaviour of DLC films, Thin Solid Films, 2008, 516, 5397.
11. H. G. Kim, S. H. Ahn, J. G. Kim, S. J. Park and K. R. Lee, Electrochemical behavior of diamond-like carbon films for biomedical applications, Thin Solid Films, 2005, 475, 291.
12. E. Liu and H. W. Kwek, Electrochemical performance of diamond-like carbon thin films, Thin Solid Films, 2008, 516, 5201.
13. L. X. Liu and E. Liu, Nitrogenated diamond-like carbon films for metal tracing, Surf. Coat. Technol., 2005, 198, 189.
14. A. Zeng, E. Liu, I. F. Annergren, S. N. Tan, S. Zhang, P. Hing and J. Gao, EIS capacitance diagnosis of nanoporosity effect on the corrosion protection of DLC films, Diamond Relat. Mater., 2002, 11, 160.
15. S. C. Trippe and L. Pereira, Electrical and photocurrent analysis of Si–fluorine doped DLC films heterojunctions, Diamond Relat. Mater., 2007, 16, 1349.
16. W. Dai, G. S. Wu and A. Y. Wang, Preparation Preparation, characterization and properties of Cr-incorporated DLC films on magnesium alloy, Diamond Relat. Mater., 2010, 19, 1307.
17. G. H. Song, X. P. Yang, G. L. Xiong, Z. Lou and L. J. Chen, The corrsive behavior of Cr/CrN multilayer coatings with different modulation periods, Vacuum, 2013, 89, 136–141.
18. A. Dorner-Reisel, C. Schürer, G. Irmer and E. Müller, Electrochemical corrosion behaviour of uncoated and DLC coated medical grade Co₂₈Cr₆Mo, Surf. Coat. Technol., 2004, 177, 830–873.
19. S.S. Hadinata, M. T. Lee, S. J. Pan, W. T. Tsai, C. Y. Tai and C. F. Shih, Electrochemical performances of diamond-like carbon coatings on carbon steel, stainless steel and brass, Thin Solid Films, 2013, 529, 412–416.
20. Z. Xu, Y. J. Zheng, F. Liang, Y. X. Leng, H. Sun and N. Huang, The microstructure and mechanical properties of multilayer diamond-like carbon films with different modulation ratios, Appl. Surf. Sci., 2013, 264, 207–212.
21. Y. Uematsu, T. Kakiuchi, T. Teratani, Y. Harada and K. Tokaji, Improvement of corrosion fatigue strength of magnesium alloy by multilayer diamond-like carbon coatings, Surf. Coat. Technol., 2011, 205, 2778–2784.
22. K. Feng, Z. Li, F. Lu, J. Huang, X. Cai and Y. Wu, Corrosion resistance and electrical properties of carbon/chromium–titanium–nitride multilayer coatings on stainless steel, J. Power Sources, 2014, 249, 299–305.
23. C. Liu, A. Leyland, Q. Bi and A. Matthews, Corrosion resistance of multi-layered plasma-assisted physical vapour deposition TiN and CrN coatings, Surf. Coat. Technol., 2001, 141, 164–173.
24. L. A. Dobrzański, K. Lukaszkowicz, D. Pakula and J. Mikula, Corrosion resistance of multilayer and gradient coatings deposited by PVD and CVD techniques, Arch. Mater. Sci. Eng., 2007, 28, 12.
25. E. Härkönen, I. Kolev, B. Díaz, J. Światowska, V. Maurice, A. Seyeux, P. Marcus, M. Fenker, L. Toth, G. Radnoczi, M. Vehkamäki and M. Ritala, Sealing of Hard CrN and DLC Coatings with Atomic Layer Deposition, ACS Appl. Mater. Interfaces, 2014, 6, 1893.
26. C. L. Liu, D. Z. Yang, G. Q. Lin and M. Qi, Corrosion resistance and hemocompatibility of multilayered Ti/TiN-coated surgical AISI 316L stainless steel, Mater. Lett., 2005, 59, 3813.
27. J. J. Wang, J. B. Pu, G. A. Zhang and L. P. Wang, Interface architecture for superthick carbon-based films toward low internal stress and ultrahigh load-bearing capacity, ACS Appl. Mater. Interfaces, 2013, 5, 5015.
28. J. J. Wang, J. B. Pu, G. A. Zhang and L. P. Wang, Tailoring the structure and property of silicon-doped diamond-like carbon films by controlling the silicon content, Surf. Coat. Technol., 2013, 59, 312.
29. F. R. Marciano, E. C. Almeida, D. A. Lima-Oliveira, E. J. Corat and V. J. Trava-Airoldi, Improvement of DLC electrochemical corrosion resistance by addiction of fluorine, Diamond Relat. Mater., 2010, 19, 537.
30. J. H. Sui, Z. G. Zhang and W. Cai, Surface characteristics and electrochemical corrosion behavior of fluorinated diamond-like carbon (F-DLC) films on the NiTi alloys, Nucl. Instrum. Methods Phys. Res., Sect. B, 2009, 267, 2475.
31. J. Liang, P. B. Srinivasan, C. Blawert and W. Dietzel, Influence of chloride ion concentration on the electrochemical corrosion behaviour of plasma electrolytic oxidation coated AM50 magnesium alloy, Electrochim. Acta, 2010, 55, 6802.
32. R. H. Zhang, J. Liang and Q. Wang, Preparation and characterization of graphite-dispersed styrene–acrylic emulsion composite coating on magnesium alloy, Appl. Surf. Sci., 2012, 258, 4360.
33. L. P. Wang, Y. L. Lin, Z. X. Zeng, W. M. Liu, Q. J. Xue, L. T. Hu and J. Y. Zhang, Electrochemical corrosion behavior of nanocrystalline Co coatings explained by higher grain boundary density, Electrochim. Acta, 2007, 52, 4342.
34. J. Liang, P. B. Srinivasan, C. Blawert and W. Dietzel, Comparison of electrochemical corrosion behaviour of MgO and ZrO₂ coatings on AM50 magnesium alloy formed by plasma electrolytic oxidation, Corros. Sci., 2009, 51, 2483.
35. L. P. Wang, J. Y. Zhang, Y. Gao, Q. J. Xue, L. T. Hu and T. Xu, Grain size effect in corrosion behavior of electrodeposited nanocrystalline Ni coatings in alkaline solution, Scr. Mater., 2006, 55, 657.
36. C. Dagbert, T. Meylheuc and M. N. Bellon-Fontaine, Corrosion behaviour of AISI 304 stainless steel in presence of a biosurfactant produced by Pseudomonas fluorescens, Electrochim. Acta, 2006, 51(24), 5221.
37. Z. T. Chang, B. Cherry and M. Marosszeky, Polarisation behaviour of steel bar samples in concrete in seawater. Part 2: a polarisation model for corrosion evaluation of steel in concrete, Corros. Sci., 2008, 50, 3078.
38. S. H. Ahn, J. H. Lee, J. G. Kim and J. G. Han, Localized corrosion mechanisms of the multilayered coatings related to growth defects, Surf. Coat. Technol., 2004, 177–178, 638.
39. B. Matthes, E. Broszeit, J. Aromaa, H. Ronkainen, S. P. Hannula, A. Leyland and A. Matthews, Corrosion performance of some titanium-based hard coatings, Surf. Coat. Technol., 1991, 49, 489.
40. K. Cheng, C. B. Ren, W. J. Weng, P. Y. Du, G. Shen, G. R. Han and S. Zhang, Bonding strength of fluoridated hydroxyapatite coatings: a comparative study on pull-out and scratch analysis, Thin Solid Films, 2009, 517, 5361.
41. J. Flis, Y. Tobiyama, K. Mochizuki and C. Chiga, Characterisation of phosphate coatings on zinc, zinc–nickel and mild steel by impedance measurements in dilute sodium phosphate solutions, Corros. Sci., 1997, 39, 1757.
42. A. Jimenez-Morales, J. C. Galvan, R. Rodriguez and J. J. de, Damborena Electrochemical study of the corrosion behaviour of copper surfaces modified by nitrogen ion implantation, J. Appl. Electrochem., 1997, 27, 550.
43. E. M. A. Martini and I. L. Muller, Characterization of the film formed on iron in borate solution by electrochemical impedance spectroscopy, Corros. Sci., 2000, 42, 443.
44. G. Zhao, J. J. Feng, J. J. Xu and H. Y. Chen, Direct electrochemistry and electrocatalysis of heme proteins immobilized on self-assembled ZrO₂ film, Electrochem. Commun., 2005, 7, 724.
45. M. F. Montemor, W. Trabelsi, M. Zheludevich and M. G. S. Ferreira, Modification of bis-silane solutions with rare-earth cations for improved corrosion protection of galvanized steel substrates, Prog. Org. Coat., 2006, 57, 67.
46. M. Taryba, S. V. Lamaka, D. Snihirova, M. G. S. Ferreira, M. F. Montemor, W. K. Wijting, S. Toews and G. Grundmeier, The combined use of scanning vibrating electrode technique and micro-potentiometry to assess the self-repair processes in defects on “smart” coatings applied to galvanized steel, Electrochim. Acta, 2011, 56, 4475.
47. Z. Xiao, Z. Li, A. Y. Zhu, Y. Y. Zhao, J. L. Chen and Y. T. Zhu, Surface characterization and corrosion behavior of a novel gold-imitation copper alloy with high tarnish resistance in salt spray environment, Corros. Sci., 2013, 76, 42.
48. G. Okamoto, Passive film of 18–8 stainless steel structure and its function, Corros. Sci., 1973, 13, 471.
49. O. Y. Posudievsky, O. A. Kozarenko, V. S. Dyadyun, S. W. Jorgensen, J. A. Spearot, V. G. Koshechko and V. D. Pokhodenko, Effect of host–guest versus core–shell structure on electrochemical characteristics of vanadium oxide/polypyrrole nanocomposites, Electrochim. Acta, 2011, 58, 442.
50. E. Casero, A. M. Parra-Alfambra, M. D. Petit-Domínguez, F. Pariente, E. Lorenzo and C. Alonso, Differentiation between graphene oxide and reduced graphene by electrochemical impedance spectroscopy (EIS), Electrochem. Commun., 2012, 20, 63.
51. M. L. Zheludkevich, R. Serra, M. F. Montemor, K. A. Yasakau, I. M. M. Salvado and M. G. S. Ferreira, Nanostructured sol–gel coatings doped with cerium nitrate aspre-treatments for AA2024-T3: corrosion protection performance, Electrochim. Acta, 2005, 51, 208.
52. C. L. Liu, P. K. Chu, G. Q. Lin and D. Z. Yang, Effects of Ti/TiN multilayer on corrosion resistance of nickel–titanium orthodontic brackets in artificial saliva, Corros. Sci., 2007, 49, 3783.
53. B. C. Bunker, D. M. Haaland, T. A. Michalske and W. L. Smith, Kinetics of dissociative chemisorption on strained edge-shared surface defects on dehydroxylated silica, Surf. Sci., 1989, 222, 95.
54. P. D. Maguire, J. A. McLaughlin, T. I. T. Okpalugo, P. Lemoine, P. Papakonstantinou, E. T. McAdams, M. Needham, A. A. Ogwu, M. Ball and G. A. Abbas, Mechanical stability, corrosion performance and bioresponse of amorphous diamond-like carbon for medical stents and guidewires, Diamond Relat. Mater., 2005, 14, 1277.
55. J. Choi, S. Nakao, J. Kim, M. Ikeyama and T. Kato, Corrosion protection of DLC coatings on magnesium alloy, Diamond Relat. Mater., 2007, 16, 1361.
56. H. J. Yasuda, Plasma polymerization, 1st edn, Academic Press, New York, 1985.
57. R. C. C. Rangel, M. E. P. Souza, W. H. Schreiner, C. M. A. Freire, E. C. Rangel and N. C. Cruz, Effect of the fluorination of DLC film on the corrosion protection of aluminum alloy (AA 5052), Surf. Coat. Technol., 2010, 204, 3022.
58. G. S. Wu, L. L. Sun, W. Dai, L. X. Song and A. Y. Wang, Influence of interlayers on corrosion resistance of diamond-like carbon coating on magnesium alloy, Surf. Coat. Technol., 2010, 204, 2193.
59. M. Fenker, M. Balzer, H. A. Jehn, H. Kappl, J. J. Lee, K. H. Lee and H. S. Park, Improvement of the corrosion resistance of hard wear resistant coatings by intermediate plasma etching or multilayer structure, Surf. Coat. Technol., 2002, 150, 101.

---