Effect of carbon nanotube content and annealing temperature on corrosion performance of carbon nanotube/Ni composite layer

Adding carbon nanotubes (CNTs) to metal composites changes their corrosion resistance, which is significantly affected by the distribution of CNTs. In this study, the effect of the content and distribution of CNTs on the corrosion resistance of composites was investigated by changing the electrodeposition process. The results indicated that could inhibit grain growth and act as an elemental channel for passivation film formation, which positively enhanced the corrosion resistance of the material. However, the annealing used to improve the bonding strength of CNTs to the matrix increased the grain size of the material, which had a weakening effect on the corrosion resistance. Using ultrasonic in electrodeposition had an obvious promoting effect on the uniform distribution of CNTs. The composites with 0.1 g/l CNT showed the best corrosion resistance after annealing for 30 min at 600 °C.

CNTs can be used in different ways to enhance the corrosion resistance of materials.CNTs can change the crystal microstructure of the deposited layer to increase the fraction of low-angle grain boundaries (LAGBs) and low-Σ coincidence site lattice (CSL) (Σ 29) to 40% and 30%, respectively [17].Additionally, the CNT addition during the electrodeposition process can improve the hydrophobic angle of the film layer and reduce the area in contact with the corrosive medium, making the contact angle increase from 88°to 144°, thus effectively enhancing the corrosion resistance [18,25].Furthermore, due to the polarization phenomenon of CNTs, the appropriate CNT content can effectively enhance the corrosion resistance of the material [26][27][28][29].In addition, CNTs can also be used with other elements to strengthen the properties through synergistic interactions or by refining the grain [30][31][32][33][34]. CNTs can also strengthen the thermal stability of materials by filling the pores [35].Homogeneous distribution and well-adhesion of CNTs to the metal matrix are the main factors in enhancing the material properties.If CNTs are distributed uniformly, homogeneous corrosion can occur, and localized corrosion is prevented.In contrast, the agglomeration of CNTs can deteriorate the properties.This is the reason why the high CNT content leads to degradation of properties [36][37][38][39][40]. Annealing can change the defects and the distribution uniformity of CNTs, improve the mechanical properties of the composites effectively [41][42][43][44].However, there are few studies on the effect of corrosion properties.Previous studies on the effect of CNT content on their properties had generally shown that the best performance was achieved using ∼0.2 wt% CNTs [27], but there was no further research on more detailed concentration changes.Moreover, performance characteristics of CNTs/Ni composites have not yet been explored systematically even though there have been a few reports on their corrosion to wear properties.This requires further research to explore the interconnections.Therefore, it is crucial to investigate the corrosion resistance of the composites with different CNT contents to promote a wider application range for CNTs/Ni composites.
In this study, the uniform distribution of CNTs was achieved using ultrasound and stirring during the electrodeposition process.In addition, the electrodeposition process was changed cyclically to form multilayered composites, and the effect and mechanism of different CNT contents on the corrosion resistance were explored.Thus, multilayered composites with better corrosion resistance could be prepared.

Process
In this study, the nickel sulfamate solution system was employed to prepare CNTs/Ni composites by electrodeposition.Sodium dodecyl sulfate (SDS) was used as a surface leveling agent, hydroxypropyl cellulose (HPC) was used as a dispersant, and boric acid was used to maintain the pH of the plating solution between 4 and 5. Different CNT contents were added to study the corrosion resistance of composites.The specific parameters are shown in table 1, where T on is the pulse-on time during the experiment and T off is the pulse-off time.All chemical reagents were analytical reagents, and the solutions were prepared with distilled water.All the CNTs in the experiments were multi-walled CNTs (MW-CNTs) that were 10-30 μm in length.They were purchased from Chengdu Organic Research Institute, China, without any other treatment.The copper plate was used as a substrate for depositing CNTs/Ni composites, and the pure nickel plate was used as the anode.In this experiment, pulse current (PC) power and direct current (DC) power were used for electrodeposition.The specific experimental parameters are shown in table 2. When the experimental condition used stirring with a PC power supply, the electrodeposition was carried out using a constant temperature heating magnetic stirrer at 50 °C with a magnetic stirrer speed of 1000 r min −1 .When the experimental condition was ultrasonic with a PC power supply, the electrodeposition was carried out in an ultrasonic bath and sonicated at 40 Hz.Electrodeposition under a DC power supply was carried out in a thermostatic water bath.Before formal deposition, trial plating for 5 h was carried out to remove impurities from the electrodeposition solution.Electrodeposition was carried out in 24 h cycles.The first 12 h cycle was carried out under stirring conditions with a PC power supply.The last 12 h cycle was alternated by DC power electrodeposition for 30 min and PC power electrodeposition with ultrasonic for 30 min, totaling 12 times.The electrodeposition was carried out for a total of 3.5 cycles, lasting 84 h.After electrodeposition, the samples were annealed at 200, 400, and 600 °C for 30 min to investigate the effect of temperature on the corrosion properties of the CNTs/Ni composites.

Sample testing
The x-ray diffraction (XRD) analysis was employed to test the crystal orientation and microstrain of the samples.
In the XRD test, Cu-Kα target (r = 1.5418Å) was used with an operating voltage of 40 kV, scanning speed of 5°/ min, step width of 0.01°, and scanning angle of 10°to 90°.The CS2350H model electrochemical workstation was employed to test the electrochemical properties.The sample was cut into 10 mm × 10 mm square pieces by a wire-cutting machine and connected with a wire by welding.A three-electrode system was formed with a platinum electrode, saturated calomel electrode (SCE), and the sample.The solution of 3.5 wt% NaCl was employed as a corrosion medium.The sample was placed in an open circuit for 30 min to stabilize the corrosion potential before electrochemical testing.The electrochemical impedance spectroscopy (EIS) test was performed in a frequency range of 10 5 -10 −2 Hz, and the results obtained (repeated thrice) were analyzed using ZsimpWin 3.5 software.The potentiodynamic scan range was −1.0 to 1.0 V, with a scan rate of 1 mV s −1 .The Tafel extrapolation method was used to process the data.Scanning electron microscopy (SEM) was employed to observe the morphology of the sample after corrosion and scratching experiments.The SEM model was VEGA3 TESCAN and Nova NanoSEM 450.Transmission electron microscopy (TEM), the Tecnai 12 model, was employed to analyze the structure of composites.

Surface and microhardness properties
The surface state of materials greatly influences the properties of materials, and the corresponding microscopic characteristics can be preliminarily evaluated from the surface state of materials.The surface micrographs of samples with different CNT contents are shown in figures 1(a)-(d).It can be seen from figure 1(a) that the surface of the sample without CNTs was smooth and flat.However, the CNT addition made the sample surface rough and uneven with nodules, as shown in figures 1(b)-(d).It can be seen that the step-like structure gradually flattened, the edges and corners gradually rounded, and the number of step-like growths gradually decreased.This can be attributed to the stepped pyramidal nickel growth in electrodeposition [45] and the CNT addition which allowed new phases to continue to grow on the originally formed steps, leading to rough surface formation with nodules.Table 3 shows the densification of samples with different CNT contents.It can be seen from table 3 that the densification of the samples decreased slightly with the CNT addition.This can be attributed to the fact that the shape of CNTs was tubular and hollow and that CNTs were segregated on the surface.
It is necessary to observe the metallographs of the cross-section of the material to analyze the electrodeposition microstructure of the material ulteriorly.The samples were annealed to reduce the microstrain during electrodeposition and control of grain size.As shown in figure 1(e), the samples had a layered structure which was caused by the change in electroplating conditions.It can be seen from figures 1(e)-(h) that in the pure nickel sample with the straight shape of the alternating layer, the CNT addition made the alternating layer of the sample wavy, and some black spots appeared.This was caused by the stepped pyramid Ni growth, corresponding to the nodules on the surface micrographs.As can be seen from figures 1(e)-(p), with the increase in the CNT content, a large number of black clusters appeared on the cross-section and proceeded toward clustering together.It can be deduced that the black clusters in the metallographs were CNTs, and this may be due to the non-uniform distribution of CNTs during electrodeposition, which was not relieved by annealing.There were significantly more black clusters in the non-alternating layer than in the non-alternating layer.It can be determined that the ultrasound effect was significantly better than that of stirring, which resulted in the uniform distribution of CNTs.With the increase in the annealing temperature, the black clusters in the crosssection metallographs became dispersed, and at 600 °C the clusters were hardly visible in the low content of CNTs.It can be inferred that annealing samples at higher temperatures could effectively alleviate the inhomogeneous distribution of CNTs.
Figure 2 shows the microhardness maps of CNTs/Ni composites with different CNT contents, which were annealed at 200, 400, and 600 °C for 30 min.Because of the layered structure of the sample, the hardness test of the alternating layer (Ultra) and the non-alternating layer of the sample was carried out separately to analyze the influence of different structures on the microhardness.Based on figure 2(a), it is evident that the hardness of the alternating layer could slightly higher than the non-alternating layer when it was not annealed.Furthermore, with the increase in the CNT content, the hardness value of the alternating layer also increased, while the hardness value of the non-alternating layer was the highest when the CNT content was 0.1 g l −1 , as shown in figure 2(a).The change in hardness proved that CNTs had an enhancing effect, but the hardness of the non-alternating layer was highest at 0.1 g l −1 that perhaps due to the non-uniform distribution of the higher content of CNTs.After annealing at 200 °C for 30 min, as shown in figure 2(b), the hardness value of the alternating layers under different CNT contents increased.This may be due to the dispersion of the CNTs clusters during the annealing process.Figures 2(c)-(d) shows the hardness value of CNTs/Ni composites after annealing at 400 and 600 °C for 30 min, respectively.With the further increase in the annealing temperature, the hardness value of the CNTs/Ni composites was reduced, and the phenomenon that the hardness of the alternating layer was smaller than the non-alternating layer appeared.It can be inferred that the grain size and the distribution of CNTs changed during the annealing process, which synergistically caused the hardness to show a changing trend.From figures 2(a)-(d), it can be seen that the overall hardness of the composites decreased with the increase in the annealing temperature, but the hardness showed different changes with the addition of CNTs.This is because the addition of CNTs made the dislocation density and internal stress in the composites much higher than that of pure metal.The grains grew during heat treatment, but the presence of CNTs restricted the movement of grain boundaries and hindered grain growth.The dislocations moved during the annealing process and subgrain boundaries appeared inside the grains, which enhanced the hardness of the material.The CNTs in the alternating layers were more uniformly distributed and the limiting effect on grain size was more pronounced, resulting in higher hardness.The interaction between the decrease in hardness due to grain size growth and the inhibitory effect of CNTs on grain size made the hardness of the material fluctuate up and down.

XRD results
The physical phases searched by Jade 6.As can be seen from the XRD pattern of the pure nickel sample, the deposited nickel had the strongest diffraction peak for the (2 0 0) crystal plane and the weakest for the (3 1 1) crystal plane with only a weakly convex diffraction peak showing a strong (2 2 0) texture.The preferred orientation of nickel changed from the (2 0 0) crystal plane to the (1 1 1) crystal plane due to CNT addition, as shown in figure 3(a).Similar results were reported by Ramazan and coworkers [46], where the addition of CNTs caused a change in the preferred orientation from (2 2 0) to (1 1 1).addition of CNTs, which changed the preferred orientation of the crystals from (220) to (111).In contrast, the texture coefficient was slightly affected due to the change in the content of CNTs.This is consistent with previous findings that adding CNTs could change the preferred crystal orientation [46].As shown in figure 3(d), the samples with a CNT content of 0.1 g/l showed a weak change in the texture coefficient during the annealing process and kept the grains in a preferred orientation of (111).Figure 3(e) demonstrates the grain size variation of different CNTs, and figure 3(f) demonstrates the grain size variation of samples with a CNT content of 0.1 g/l for different annealing temperatures.The addition of CNTs resulted in smaller grain sizes as shown in figure 3(e).This can be attributed to the addition of CNTs, which promoted the crystal nucleation during electrodeposition [45].The grain growth by the increase in the annealing temperature is shown in figure 3(f), which corresponds to the microhardness in figure 2. Thus, it can be concluded that grain size was the reason for the decrease in microhardness of the samples during the annealing process.Figures 3(g)-(h) shows the effect on microstrains by the addition of CNTs and the effect of annealing on microstrains, respectively.As can be seen in figure 3(g), the addition of CNTs reduced the microstrain.This may be because CNTs were deformed by internal stresses, thus reducing internal stresses.From figure 3(h), it can be seen that the microstrain gradually decreased with the increase in the annealing temperature, due to the weakening of internal stresses during the annealing process.

Microstructural characteristics
The as-deposited composites were analyzed by TEM. Figure 4

Potentiodynamic polarization curve
Figure 5 shows the polarization curves of the composites with different CNT contents in a 3.5 wt% NaCl solution.As shown in figure 5(a), it can be seen that both pure nickel and composites exhibited typical activated corrosion behaviors.Figure 5(b) reveals the local enlargement of the Tafel region of the polarization curve.In the strongly polarized region, self-corrosion potential and current could be obtained by Tafel extrapolation to calculate the corrosion rate of the material and characterize the corrosion properties.Figure 5(c) shows the local enlargement of the anode area.It can be seen that the passivation zone width was about 0.7 V (as shown in section AB in figure 5(c)).In the passivation zone, the corrosion current density was basically unaffected by the potential.In this process, a passivation film was formed, which resulted in uniform corrosion of the materials.The potential was further increased to 0.2 V, arriving at point B on the curve, and the current density sharply increased.The breakdown potential was at 0.2 V.At this point, the passivation film began to dissolve in the over- passivation zone, and the dissolution rate was higher than the forming rate.This resulted in the rupture of the passivation film, which exhibited the pitting characteristics [50].From point B, the curve entered the oxygen release zone.Based on figure 5(c), it can be seen that the passivation zone of the composites had a greater range compared to pure nickel.The formation of the passivation film requires the diffusion of elements and the nucleation growth of oxides.CNTs are hollow tubes, that can provide an elemental diffusion channel and nucleation points for the passivation film formation, thus effectively promoting the formation of the passivation film and increasing the corrosion resistance of the material [51].
Figures 6(a)-(c) shows the polarization curves for as-deposited composites with different CNT contents, which were annealed at 200, 400, and 600 °C for 30 min, respectively.It can be seen from figures 6(a)-(c) that the composites had different passivation zones generated in the anodic region during the corrosion process in the 3.5 wt% NaCl solution.There was barely a change in the polarization curves of the samples after annealing.However, the passivation platform of the samples appeared to be significantly shortened after annealing at 600 °C for 30 min.As can be seen in figure 3(f), the grain growth was due to annealing.The grain growth decreased the elemental diffusion channels and the location of passivation film nucleation; thus, the passivation platform was shortened.Figure 6(d) shows the schematic diagram of the data processing when the CNT content was 0.1 g l −1 .The value corresponding to the horizontal coordinate at the intersection of the two green lines in figure 6(d) indicates the self-corrosion potential.Moreover, the value of the vertical coordinate indicates the selfcorrosion current.Since the differences in the polarization curves were insignificant, the electrochemical parameters were obtained by fitting the polarization curves to analyze the differences in the corrosion resistance in table 4. The fitting results include polarization resistance (R p ), corrosion potential (E corr ), and corrosion current (I corr ).A positive E corr value indicates a decreasing corrosion ability and a small I corr value implies a low corrosion rate.The corrosion rate (CR) and the corrosion inhibition efficiency (IE, %) were calculated using equations (4) and (5) [52-54], respectively.where I corr is the self-corrosion current density, the constant K = 3270 mol A −1 , M is the molar mass (Ni = 58.69g mol −1 ), n is the number of charges involved in the transfer during the electrochemical reaction, and ρ m is the density derived from table 3.
As shown in table 4, for as-deposited composites, with the increase in the CNT content, the CR values of composites decreased and then increased.However, the overall corrosion rate was lower than that of pure nickel, and the IE values were all positive, except for the samples annealed at 400 °C for 30 min.The CR values were minimized, and the corrosion resistance was best when the CNT content was 0.1 g l −1 .It shows that the CNT addition improved the corrosion performance of the composites to different degrees.It may be because the addition of CNTs refined the grains of the composites.During electrodeposition, when CNTs were wrapped around the nickel matrix, the adsorbed ions made the growth center unaffected by the cations in the electrolyte and further prevented the growth of nickel particles, thereby refining the grain size [55].The electronic activity of fine grain boundary is strong, and a stable passivation film is easily formed on the surface of the active metal [56].The poor corrosion rate with 0.2 g l −1 CNTs was attributed to the partial clustering of CNTs, which coincided with figure 1.The corrosion resistance of the samples was improved by annealing at 200 °C for 30 min.This can be attributed to the uniform distribution of CNTs.This is also consistent with the hardness increase in figure 2. For the composites annealed at 400 °C for 30 min, the grains grew.The increase in the grain size reduced the nucleation sites and diffusion of elements, making the formation of the passivation film difficult; thus, the corrosion resistance was reduced.Counterintuitively, the corrosion resistance of the samples with CNTs was enhanced when the annealing temperature was increased to 600 °C for 30 min, while the R p value of pure nickel decreased.At higher annealing temperatures, the grain grew, and CNTs were distributed uniformly, as shown in figure 1, which was the reason for the enhanced corrosion resistance.The decrease in corrosion performance of pure nickel samples without the addition of CNTs also indicated that CNTs could effectively enhance the corrosion resistance of the material.As can be seen from figure 1, samples with 0.05 and 0.1 g/l CNTs had a more uniform distribution of CNTs than samples with 0.2 g l −1 CNTs; thus, the corrosion resistance was higher.The results indicated that in this experiment, the corrosion resistance of the samples was better when the sample with 0.1 g l −1 CNTs was annealed at 600 °C for 30 min.Excess amounts of CNTs made it difficult to distribute them uniformly in the material, while too few CNTs had no significant enhancement effect on the corrosion properties.

Impedance analysis
AC impedance tests were performed on the samples to test the corrosion resistance under different process conditions.The studied samples included 0.1 g/l CNTannealed at 200 and 600 °C for 30 min and 0.05 g l −1 CNT annealed at 600 °C for 30 min.The results are shown in figure 7.As shown in figure 7(a), the impedance spectrum exhibited an unfinished half-arc feature.This implies that the sample was passivated and the exfoliation of the corroded surface did not occur during the corrosion process [57].All low-and high-frequency regions of the impedance spectrum were covered by capacitive semicircular arcs, meaning there was no change in the EIS characteristics by adding CNTs.In this experiment, the impedance spectrum was fitted using ZsimpWin, and the fitting results are shown in table 5.As shown in table 5, R s is the solution resistance, CPE is the passivation film capacitance, and R t is the charge transfer resistor between the passivation film and the substrate.The charge transfer resistance corresponding to the semicircle contained the charge transfer resistance of the Ni layer and the passivation film.The charge transfer resistance to form a passivation film arose from the electrochemical reaction between the oxide and the electrolyte, indicating the resistance to the generation of oxides or hydroxides.It can be concluded in table 5 that when the CNT content was 0.1 g l −1 with annealing at 600 °C for 30 min, the sample had the maximum charge transfer resistance value of 6.098 × 10 5 Ω•cm −2 and the lowest charge transfer rate.This indicates excellent passivation film stability and better corrosion resistance of the sample.However, the corrosion performance significantly decreased when the CNT content was 0.05 g l −1 with annealing at 600 °C for 30 min, and the charge transfer resistance value of the sample was 3.870 × 10 5 Ω•cm −2 .This indicates that the CNT content was the most influential factor in the corrosion resistance in this experiment.The grain refinement and the promotion of passivation film formation due to changes in the CNT content effectively enhanced the corrosion resistance.
Figure 7(b) shows the Bode plot, representing the relationship between the impedance mode, frequency, and phase angle.As shown in figure 7(b), maximum phase angle values were less than 90°, indicating the non-ideal capacitance of the passivated film.As can be seen from the Log F-Phase diagram, the curve gradually rose from the high-frequency zone to the mid-frequency zone.However, the maximum phase angle separation was not obvious enough, so it could not accurately determine the peak situation.In this study, the time constant under this system could not be clearly distinguished according to the curve.However, combined with the Nyquist plot, which shows that there was only one tolerance arc semicircle, it can be deduced that there was one time constant under this system [58,59].The Log F-Log|Z| diagram is usually used to evaluate the stability of the passivation film.It can be seen in figure 7(b) that the curve started from the high-frequency zone; the region was flat at the beginning, and then it increased and reached the maximum value gradually.The flat phase corresponds to the solution resistance (R s ) of the system, the rising phase of the CPE, and the maximum value corresponds to the charge transfer resistance (R t ) of the system.It can be judged from figure 7(b) that when the CNT content was 0.1 g l −1 and annealing was at 600 °C for 30 min, the passivation film formed on the sample was the most stable.After the kinetic potential polarisation curve and AC impedance analysis, it can be concluded that the sample with 0.1 g l −1 CNT annealed at 600 °C for 30 min had the best corrosion resistance.

Corrosion morphology analysis
The surface morphology of the samples was observed by SEM to analyze their corrosion mechanism.The electrochemical corrosion morphologies of the composites with different CNT contents in the deposited state are shown in figure 8.As is shown in figure 8, both pure Ni in figure 8(a) and composites in figures 8(b)-(d) had pitting characteristics with typical local corrosion characteristics.It is a common type of localized corrosion of passivated metals in which corrosion damage is highly concentrated.The main factor of pitting corrosion was Cl -, which significantly affected the pitting breakdown potential and pitting rate [60].The pitting generation due to the existence of Cl -resulted from competitive adsorption of oxygen and Cl -and localized rupture of the passivated film [61].The Cl -could cause the local breakdown of the passivation film and cause pitting or repassivation due to easy adsorption [62,63].Interfacial corrosion occurred at the agglomeration zone of CNTs due to the existence of composite interfaces and high-density grain boundaries, as shown in figures 8(b)-(d).The irregular shape of corrosion boundaries was due to corrosion product shedding.As shown in figure 8, the corrosion pit of the sample changed from a regular round shape (figure 8(a1)) to an irregular shape (figures 8(b1)-(d1)) as the CNT content increased, and multiple corrosion pits were connected.This is because the CNT addition produced gap corrosion, and multiple corrosion pits appeared at the CNT segregation zone.After the corrosion products fell off, the irregular shape of the corrosion pit edge was formed and joined together, as shown in figures 8(b1)-(d1).
Figure 9 shows the corrosion morphology of the samples annealed at 200 °C for 30 min.It can be observed that local pitting still occurred on the samples.The corrosion craters of pure nickel were relatively uniformly distributed, as presented in figures 9(a) and (a1).As shown in figures 9(b) and (b1), the corrosion pit edges of the composites with a CNT content of 0.05 g l −1 were stepped.It should be noted that when the CNT content was 0.1 and 0.2 g l −1 , the corrosion holes were more numerous and enriched, as seen in figures 9(c)-(d).The size of the corrosion craters was relatively small in figures 9(c1) and (d1) compared to that of as-deposited composites in figures 8(c1) and (d1).In general, incorporating CNTs led to the formation of many micropores in the composite layer.At the micropores or damages of the coating, the exposed portion of the metal surface rapidly dissolved and developed into corrosion holes or pits.However, the better corrosion resistance of the annealed samples was mainly attributed to the grain refinement, which competed with the uniformity of CNT distribution.Furthermore, when pitting was coupled with stress loading, the stress changed the corrosion potential and the corrosion current density of the metal.While most of the stress is eliminated in the annealing process, reducing the chance of stress corrosion, the higher passivation film content is also important for good corrosion resistance.
The surface morphology of the composite layer annealed at 400 °C for 30 min is shown in figure 10.Figures 10(a) and (a1) shows that the corrosion pit edges of the pure nickel were relatively smooth.At the same time, the laminar corrosion pit corroded out is also observed in figures 10(a1) and (d1).The laminated samples prepared by the alternating process were used in this experiment.After the corrosion test, it was found that the corrosion resistance of samples varied from layer to layer, and the anisotropy of the corrosion resistance of each layer could be seen at the corrosion pit edges.Morphological feathers highlight corrosion morphology in three  dimensions.Moreover, the increase in the CNT content led to further aggravation of the agglomeration phenomenon and microporosity, thus the corrosion pit edges were extremely irregular in figure 10(d1).
The annealing temperature of the samples was further increased to study the corrosion resistance of the material prepared at higher temperatures.Figure 11 shows the morphology of the samples annealed at 600 °C for 30 min.It can be seen that both pure nickel in figure 11(a) and composite materials in figures 11(b)-(d) showed localized pitting corrosion with small corrosion pits and deeper development toward its interior.In general, the corrosion current strength is related to the number of particles, surface area, and element type in the microstructure of the alloy involved in pitting [17].As the corrosion process proceeded, the pitting current density varied with time, and the three-dimensional(3D) morphological characteristics of the pits changed.As observed in figures 11(a) and (a1), the corrosion pits of pure nickel material remained with rounded edges.As shown in figures 11(b) and (b1), the number of corrosion pits of the composites was less, which were uniformly distributed as shown in the low-magnification image in figure 11(b).So, the material had better corrosion resistance, and the stepped corrosion pit edges were observed at high-magnification images in figure 11(b1).From figures 11(c) and (c1), it can be seen that multiple neighboring pitting pits expanded and fused into one pitting pit.From figures 11(d) and (d1), it can be seen that after annealing at a higher temperature, a large number of micropores still existed on the surface of the samples that had a CNT content of 0.2 g l −1 , which was detrimental to the corrosion resistance of the material.

Passivation film analysis by XPS
The polarization curves in figure 5 show that the sample formed a semiconductor-type passivation film on the surface during the corrosion process.Moreover, the composition of the film was generally aqueous or nonaqueous metal oxides or salts of metals.XPS analysis was performed on the samples to analyze the composition of the passivation film on the surface of the samples.Pure nickel and samples with 0.1 g l −1 CNTs were selected for XPS testing to analyze the effect of CNTs on the composition of the surface passivation film.
As shown in figure 12, for both samples, the Ni 2 P 3/2 peaks were mainly deconvoluted into five peaks: NiO, Ni(OH) 2 , Ni, Ni°, and Ni°s at .For the O 1s peak with the split-peak fitting, the peaks were mainly deconvoluted into three peaks: NiO, Ni(OH) 2 , and other valence states of hydroxides.It can be observed that the Ni°peak of the sample with 0.1 g l −1 CNTs was significantly decreased in the Ni 2 P 3/2 plot compared to pure nickel.This can be attributed to the fact that the addition of CNTs could promote oxygen diffusion and NiO generation.The intensity of the different fitted peaks in the XPS data was directly related to the content of the corresponding ions.Therefore, it is possible to quantify the percentage of ions in different valence states for each element and to calculate and analyze the fitted peak areas [64].
Figure 13 shows a statistical plot of the content ratio in each chemical valence state after peak fitting for both samples.As shown in figure 13(a), the NiO and Ni(OH) 2 contents of the sample with 0.1 g l −1 CNTs were at 43.29% and 9.62%, respectively.The NiO and Ni(OH) 2 contents for the pure nickel were decreased compared to the sample with 0.1 g l −1 CNTs.This can be attributed to the effect of CNTs that promoted the generation of passivation film and effectively improved the corrosion resistance of the material.As shown in figure 13(b), compared with pure nickel, the percentage content ratio of the passivation film on the surface of samples with 0.1 g l −1 CNTs increased by 14.4%.This indicates that the thickness of the passivation film increased after the CNT addition, and the passivation film was more stable, resulting in better corrosion resistance of the material.This is because the CNT incorporation caused the diffusion capacity of the sample to rise.
XPS tests were performed on samples with 0.1 g l −1 CNT that were annealed at different temperatures to investigate the effect of annealing temperature on the composition and thickness of passivation films.Based on figure 14, it can be seen that the intensity of the peaks slightly changed with temperature.This was caused by grain growth, uniform distribution of CNTs and other defects during the annealing process.The contents of each component were counted to scrutinize the changes.The results are shown in figure 15.As can be seen from figure 15(a), the sample after annealing at 200 °C for 30 min had the highest NiO content ratio of 44.56%, followed by samples annealed at 400 °C for 30 min with a content of 38.81%.This indicates that the samples annealed at 200 °C for 30 min had the strongest passivation film protection, consistent with the electrochemical  test results.Annealing enhanced the distribution uniformity of CNTs in the material, which was the reason for the better protection of the passivation film of the samples when annealed at 200 °C.In contrast, the performance decrease after annealing at 600 °C was attributed to the larger grain size.As seen in figure 15(b), after annealing at 400 °C for 30 min, the sample had the highest NiO and Ni(OH) 2 and the thickest passivation film, but its corrosion performance was not the best.From figure 1, it can be seen that CNT clusters were widely distributed on the cross-sectional metallographs of the samples annealed at 400 °C, and the inhomogeneous distribution was the reason for the poor corrosion resistance.The results after XPS testing showed the excellent corrosion resistance of the samples (0.1 g l −1 CNTs/Ni composite) when annealed for 30 min at 200 °C.From the results of this experiment, it is clear that the best corrosion resistance of the composites was achieved at an annealing temperature of 600 °C when the CNT content was 0.1 g l −1 .

Conclusion
(1) With the CNT addition, CNTs combined closely with the matrix nickel and refined the grains, which enhanced the hardness.Electrodeposition under an ultrasonic environment could enhance the CNT distribution uniformity, thus effectively improving the corrosion resistance of the material.In addition, annealing could lead to a more uniform distribution of CNTs in the material, but it would result in grain growth.
(2) The corrosion resistances of the samples with the CNT content of 0.1 g l −1 , which were annealed at 600 °C for 30 min, were better.Moreover, the corrosion type of the samples was dominated by localized pitting, with the corrosion pits extending to the interior in a stepped shape.The corrosion resistance of the material depended on the competition between the uniformity of CNT distribution and grain size.

5 software are shown in figure 3 .
As can be seen in figure3, there were five distinct diffraction peaks of nickel, corresponding to (1 1 1), (2 0 0), (2 2 0), (3 1 1), and (2 2 2) crystal planes of the face-centered cubic (FCC) structure.No corresponding diffraction peak of CNTs was found due to the low content of CNTs.In addition, the precision of the device made it difficult to detect the diffraction peak of CNTs.Figures3(a)-(b) shows the XRD patterns of the as-deposited samples.

Figure 3 (
b) shows the XRD patterns of the CNTs/Ni composites with 0.1 g l −1 CNTs, which were annealed at different temperatures.Based on figure 3(b), it can be seen that annealing had no effect on the preferential orientation of the composites, and the effect of different annealing temperatures on the diffraction peaks was difficult to observe.The texture coefficients, average particle size, and microstrain were calculated by equations (1)-(3)[47][48][49] to further analyze the effect of CNT addition and annealing temperature on the composites.

3 S
where I (hkl) and I 0(hkl) represent the diffraction intensities of Ni deposits and standard Ni powder respectively, D is the particle size, k (k = 0.9) is the shape factor, λ (λ = 0.154060 nm) is the radiation wavelength, β D is the full

Figure 3 .
Figure 3. XRD patterns.(a) The CNTs/Ni composites with different CNT contents; (b) The CNTs/Ni composites with 0.1 g l −1 CNTs at different annealing temperatures; (c) Texture coefficients of as-deposited sample; (d) Texture coefficients of the annealed sample; (e) The crystallite size of the as-deposited sample; (f) The crystallite size of the annealed sample; (g) Microstrain of the asdeposited sample; (h) Microstrain of the annealed sample.
shows the TEM image of 0.1 g l −1 CNTs/Ni composites annealed at 200 °C for 30 min.As shown in figure 4, to observe the distribution of CNTs in the composites and the bonding between CNTs and Ni at the interface.As shown in figure 4(a), the part of the figure surrounded by the red line was CNTs.The CNTs were distributed in random directions, and a more uniform appearance of CNT clusters was not found in the figure.

Figure 4 (
b) shows the structure of the grown CNTs with welled-fined multi-walled carbon layers.The demarcation line between the CNTs and the substrate was not clear, which indicating the close bonding between the CNTs and the substrate.

Figure 4 .
Figure 4. (a) The microstructural image of the composites; (b) The high-resolution image of the composite interface.

Figure 5 .Figure 6 .
Figure 5. (a) Scanning curves of the dynamic potential of the CNTs/Ni composites; (b) and (c) the magnified images of the corresponding boxes in (a).

Table 1 .
Main chemical reagents and ratios.

Table 2 .
An overview of the process parameters for electrodeposition.

Table 3 .
Densification of materials with different CNT contents.

Table 5 .
Equivalent circuit element parameters for electrochemical AC impedance spectroscopy.