*3.3. Scratch Tests*

Adhesion is a must-attend feature of protective coatings. In spite of its relevance, adhesion properties of Ni–P/MWCNT composite coatings are not commonly reported in the literature. The results obtained by scratch tests of the Ni–P and Ni–P/MWCNT coatings are shown in Figure 4. The CLSM 3D views of the scratched regions are shown as well as the transverse profiles along the lines marked in the 3D micrographs. The width (average of ten measurements over the scratch length) and maximum depth of the scratched regions were determined from these lines. The results are shown in Table 2. By evaluating the transverse profiles, it is evident that MWCNT loading greatly affected the scratched region. Both the width and the depth were reduced as the MWCNT loading in the plating bath increased. The shallowest scratch was measured for the CNT-1.0 sample which was also the narrowest one. The conventional Ni–P film, in turn, presented the deepest and widest scratch. This result points to the strong hardening effect of MWCNT addition into the Ni–P matrix, confirming the results obtained by other authors [43].

**Figure 4.** 3D views of the scratched region and the corresponding transverse profile along the lines marked in the micrographs: (**a**) Ni–P; (**b**) CNT-0.25; (**c**) CNT-0.5; (**d**) CNT-1.0.


**Table 2.** Scratch dimensions of the Ni–P and Ni–P/MWCNT coatings.

The critical role played by MWCNTs in the adhesion strength of the Ni–P layer to the pipeline steel substrate is also perceived from the 3D views presented in Figure 4. The conventional binary Ni–P coating was delaminated over the scratch length, as can be seen from the relatively wide dark region in the vicinity of the scratched area in Figure 4a. In spite of the reduction in both the scratch width and depth with respect to the Ni–P film, the CNT-0.25 film also presented a relatively wide dark region along the scratch, indicating that coating delamination (Figure 4b) had occurred in a similar way of that observed for the conventional unfilled Ni–P matrix. As the MWCNT loading increased, though, different features could be perceived along the scratch line. As seen in Figure 4c, the delaminated area was greatly reduced for the CNT-0.50 sample in comparison with the CNT-0.25 and Ni–P. Such a trend was also observed for the CNT-1.0 sample (Figure 4d). For this condition, the spalling area in the surroundings of the scratch scar is confined within a narrower region. Spalling avoidance is associated with a better mechanical load accommodation during scratch tests of nickel-based coatings [44]. In this respect, our results point that the CNT particles had a beneficial effect on the adhesion properties of the Ni–P/MWCNT composite coatings.

#### *3.4. Global Electrochemical Tests*

EIS results are represented as Nyquist plots, as shown in Figure 5. All samples are characterized by capacitive loops in the medium to low frequencies, whose diameter depends on the MWCNT loading in the coating. As pointed out in the literature, the corrosion resistance is associated with the diameter of the Nyquist plots, since it is associated with the polarization resistance of the electrode [45,46]. The uncoated substrate presented very low impedance values when compared to the coated samples. As a consequence, its Nyquist plot is only seen when the impedance scales are expanded, as shown in the inset of Figure 5. The impedance values were greatly enhanced for the coated samples, indicating its beneficial effect on the corrosion resistance of the steel substrate.

The Nyquist plot of the Ni–P exhibits a bigger diameter when compared to the uncoated sample, revealing the increased corrosion resistance imparted by the electrolessly deposited film. A progressive increase in the diameter of the Nyquist plot is observed by incorporating MWCTs into the Ni–P. The CNT-1.0 is the most corrosion-resistant, as suggested by its large capacitive loop.

The corrosion resistance of the Ni–P and Ni–P/MWCNT coated samples was further evaluated by potentiodynamic polarization tests after 1 h of immersion in 3.5 wt.% NaCl solution at room temperature. The results are shown in Figure 6. The uncoated API 5L X80 steel substrate was also tested for comparison purposes. The corrosion potential (Ecorr) and corrosion current densities (icorr) were determined from these curves by means of the Tafel extrapolation method. The results are displayed in Table 3, along with the protection efficiency (P%) of the different coatings, as calculated from Equation (2).

$$P\% = \left(1 - \frac{i\_{corr}^{\*}}{i\_{corr}^{0}}\right) \times 100\tag{2}$$

where *i* ∗ *corr* and *i* 0 *corr* are the corrosion current densities of the coated and uncoated substrate, respectively.

**Figure 5.** Nyquist plots of the Ni–P and Ni–P/MWCNT coatings after 1 h of immersion in 3.5 wt.% NaCl solution at room temperature.

**Figure 6.** Potentiodynamic polarization curves of the Ni–P and Ni–P/MWCNT coatings after 1 h of immersion in 3.5 wt.% NaCl solution at room temperature.

**Table 3.** Corrosion parameters of the uncoated substrate, Ni–P and Ni–P/MWCNT coatings.


The corrosion potential (Ecorr) of the conventional Ni–P coating was shifted to the nobler direction with respect to the uncoated substrate. This trend was also observed for the composite Ni–P/MWCNT films. Notwithstanding, there is no clear tendency of increasing Ecorr with the MWCNT loading. The electrochemical stability of the electrode surface is associated with Ecorr, which increases as this parameter is shifted towards more anodic values [47]. The intrinsic low reactivity of carbon nanotubes [48] is likely to be responsible for the nobler Ecorr values, compared with the uncoated substrate and the conventional Ni–P film.

All coatings markedly decreased the corrosion current density of the API 5L X80 steel substrate. The reduced icorr scaled with the carbon nanotube was loaded in the plating bath. The protection efficiency of each coating is also presented in Table 3, showing the higher protectivity of the CNT-1.0 condition. Moreover, the anodic currents are lower for the composite Ni–P/MWCNT films when compared to the conventional Ni–P layer. It is also noteworthy that the polarization curves of the composite films presented a well-defined passive region that is not seen for the uncoated substrate or the Ni–P film. Both the intrinsic chemical inertia of carbon nanotubes and the possibility of blocking small pores in the Ni–P matrix were reported as the main causes of the increased corrosion resistance of electroless Ni–P–CNT composite coatings [49].

#### *3.5. Scanning Electrochemical Microscopy (SECM)*

SECM 2D maps of the uncoated API5LX80 steel, Ni–P, and composite Ni–P/MWCNT coatings are shown in Figure 7. The maps were recorded in 0.1 M NaCl at room temperature. The tip was biased at +600 mVAg/AgCl to sense the formation of Fe2+ ions generated at the sample surface. The sample was unbiased.

During the corrosion process of this type of material, the oxidation of Fe2+ to Fe3+ occurs in the anodic sites, in which the Fe2+ ions are originated from the material dissolution. Thus, Fe2+ ions produced from the corrosion process are available on the surface. Thereby, when the Pt tip (polarized at +600 mVAg/AgCl) passes over the anodic sites, Fe2+ ions are oxidized to Fe3+, according to Equation (1). Hence, since Fe2+ production is a primary feature of the anodic regions on the surface, the SECM maps shown in Figure 6 display the electrochemical activity of the studied surfaces related to possible active domains of Fe2+ ions. Higher oxidation current values indicate greater electrochemical activity of the surface.

The SECM maps show current spikes where the electrochemical activity is higher at the probed surface. The uncoated substrate presented the highest currents over the whole area, indicating that the electrochemical activity is more intense when compared to the coated material. There is no preferential site for current spikes, suggesting that corrosion does not occur by a localized attack, which is in agreemen<sup>t</sup> with the potentiodynamic polarization curve shown in Figure 6 for the bare substrate.

The currents were significantly reduced for the Ni–P film, indicating the protective character of the electrolessly deposited layer, leading to a decrease in the electrochemical activity probed by the tip near the sample surface. The cathodic values of the measured currents confirm the low activity for Fe2+ oxidation.

Izquierdo et al. [50] reported that the cathodic currents can be due to the fact that once the concentration of Fe2+ ions is low at the metallic surface, these species are likely to be easily oxidized to Fe3+ before diffusing to the bulk electrolyte. The cathodic current would then be probed at regions where Fe3+ ions are formed according to reaction (1). The presence of such regions would indicate that the film, although presenting low electrochemical activity, is prone to corrosion at its defective sites. It is possible to see in the SECM map shown in Figure 7B that the current values vary throughout the surface, suggesting that it is not homogeneous with respect to the sites where Fe2+ oxidation occurs.

**Figure 7.** SECM 2D maps obtained with the tip biased at +600 mVAg/AgCl and the sample at the open circuit potential: (**A**) uncoated substrate; (**B**) Ni–P layer; (**C**) CNT-0.25; (**D**) CNT-0.50; (**E**) CNT-1.0. Electrolyte: 0.1 M NaCl solution at room temperature.

The probed anodic currents are low on the surface of the CNT-0.25 sample (Figure 7C) when compared to the uncoated substrate, revealing the beneficial effect of carbon nanotubes in reducing the electrochemical activity of the metallic substrate. The current values are homogeneous over the probed area. The CNT-0.50 film, in turn, presented several current fluctuations on the SECM map and values that are slightly cathodic (Figure 7D). These low cathodic currents are likely due to the absence of the electroactive probed species (Fe2+ ions), indicating the lower activity of the CNT-0.50 film when compared with the CNT-0.25 film. This result agrees well with the global electrochemical behavior described in the previous section.

The currents are predominantly anodic over the surface of the CNT-1.0 sample, as shown in Figure 7E. The maximum currents are low over most part of the probed area, as indicated by the blue color scale at the right part of the SECM map. Hence, the increment of CNT loading in the coating layer has led to a decrease in the electrochemical activity. This is unequivocally perceived by comparing CNT-1.0 and CNT-0.25 samples. Notwithstanding, the difference is not so obvious when CNT-1.0 and CNT-0.50 samples are compared, since their SECM maps indicate low electrochemical activity in both cases. This result is also in agreemen<sup>t</sup> with the evaluation of the global electrochemical behavior the by potentiodynamic polarization curves (Figure 6 and Table 3). SECM proved to be sensitive to the electrochemical activity of the composite Ni–P–CNT films. Our results sugges<sup>t</sup> that CNT-0.50 would give a suitable performance with respect to the adhesion and corrosion properties of the composite films. CNT-1.0 gave the best overall performance.
