Investigation of the microstructure and tribological properties of CNTs/Ni composites prepared by electrodeposition

To improve the dispersity of CNTs in electroplating liquid, the CNTs were acid treated. Oxidation forms a variety of functional groups such as carboxylic (–COOH), carbonyl (–C=O), and hydroxyl (–OH) groups on the surface of CNTs. In the plating bath, nickel ions were adsorbed on the surfaces of CNTs and surrounded the carboxylic groups on the surface of the CNTs to stabilize the system [27]. These functional groups promoted the chemical reactivity between CNTs and the matrix [28], which could enhance the dispersibility of CNTs in solution and allow them to be directly used for composite fabrication.

Figure 4(a) shows the FT-IR spectra of pristine carbon nanotubes (P-CNTs) and surface-treated carbon nanotubes (ST-CNTs). Different functional groups formed on the CNTs surface after functionalization. The peaks between 2800 cm⁻¹ and 3500 cm⁻¹ were characteristic of C–H and O–H bonds. The broad peak at approximately 3445 cm⁻¹ was associated with the stretching vibration of O–H bonds in the carboxyl groups. The peak at 2800 cm⁻¹ to 2900 cm⁻¹ was attributed to the C–H stretching vibration. By comparing the P-CNTs and ST-CNTs spectra, the absorption peak intensity of the latter was higher due to the oxidation of amorphous carbon in the carbon nanotubes by H₂SO₄. The peak at 1641 cm⁻¹ was attributed to carbonyl groups. The peak at 1394 cm⁻¹ was related to C–O bonds [29], and the peaks at 1028 cm⁻¹ to 1165 cm⁻¹ were related to C–O–C bonds. These peaks appeared due to the oxidation of the amorphous carbon in the CNTs by H₂SO₄, which changed the bonding pattern. Since the electric dipoles in C–C bonds were extremely weak, no peaks associated with these bonds were observed [30]. The FTIR results show that acid treatment did not drastically change the functional groups in the CNTs, but the amorphous carbon and carbon nanoparticles in the CNTs were oxidized. Acid treatment removed impurities and graft carboxyl, hydroxyl, and other groups onto the surfaces of the CNTs, which further improved their dispersion in solution [31].

Zoom In Zoom Out Reset image size

Because the Raman spectra of CNTs were highly similar to those of graphite [32], Raman spectroscopy was used to characterize CNTs. Two characteristic peaks at approximately 1335 cm⁻¹ and 1559 cm⁻¹ are called the D and G peaks, respectively. The D band is related to structural defects including vacancies, impurities, and disorder, while the G band corresponds to the in-plane stretching vibrations of sp²-hybridized carbon. The relative intensity of the D and G peaks (I_D/I_G) is related to the defect and disorder degree of CNTs. A higher value of I_D/I_G represents more structural defects in the carbon atomic crystal. A higher D peak intensity implies the presence of defects at the edges of carbon nanotubes, which results in decreased symmetry. As illustrated in figure 4(b), the D and G peaks of CNTs are around 1330 cm⁻¹ and 1575 cm⁻¹, respectively. By comparing P-CNTs with ST-CNTs, it can be seen that the I_D/I_G values after treatment were higher than before treatment. In conclusion, the oxidation of carbon nanotubes broke some of their bonds. These results also indicated the presence of defects inserted in the nanotube structure, such as functional groups.

XPS was used to determine the contents and chemical states of oxygen-containing functional groups. The XPS survey spectra of P-CNTs and ST-CNTs are shown in figures 5(a) and (b). As shown in figure 5, in terms of the contents of chemical bonds, the concentration ratio of O=C increased because nitrate ions (NO₃ ⁺) attacked and broke some of the C=C bonds. Functional groups such as hydroxyl, carboxylic, and carbonyl groups were formed on the tube walls of CNTs. This reaction produced defects on the surface of CNTs and even fractured long CNTs into shorter ones [33]. Collectively, the XPS results are consistent with the FT-IR and Raman spectroscopy results. It was concluded that the CNTs were purified by acid treatment, and the surfaces and edges of the CNTs were modified by a large number of oxygen-containing functional groups. This improved the chemical activity and dispersibility of CNTs in aqueous solutions.

Zoom In Zoom Out Reset image size

Figure 6 shows images of the microstructure of P-CNTs and ST-CNTs. As shown in figure 6(a), most of the CNTs were entangled with each other, which negatively impacted the dispersion of carbon nanotubes in the plating solution. By comparing figures 6(d) to (a), it can be seen that the ST-CNTs were more stretched than P-CNTs, and the agglomeration of CNTs was decreased. TEM was used to further study the effect of purification of the CNTs. Figures 6(b) and (c) and (e) and (f) show TEM images of P-CNTs and ST-CNTs, respectively. As shown in figures 6(b) and (c), the surfaces of 18-walled P-CNTs with a diameter of 20–40 nm were covered with large amounts of amorphous carbon, confirming that the sample was multi-walled carbon nanotubes. Nevertheless, the surfaces of the ST-CNTs were smooth in figures 6(e) and (f). Because most of the impurities were removed during the pretreatment process, there were no obvious amorphous carbon, carbon nanoparticles, catalysts, or other impurities on the CNTs surface. In addition, the ends of the CNTs were opened.

Zoom In Zoom Out Reset image size

The pretreatment process removed the impurities from the surface of CNTs, giving them a clean surface, shortening the carbon tube length, and introducing defects, such as vacancies. However, due to the large specific surface area of CNTs, it was also difficult to avoid mutual entanglement. Consequently, surfactant was added to the plating solution, and it was stirred.

The surface roughness for samples was further investigated with a three-dimensional AFM image, as shown in figures 7(a)–(e). Here, a specific spot (10 μm × 10 μm) was identified during the analysis to provide accurate and consistent analysis to study the surface roughness analysis. The figures revealed that the 0.3 g l⁻¹ CNTs/Ni composite coating has a more relatively smooth and compact surface morphology than another coating. A dimpled appearance was observed on the 2 g l⁻¹ CNTs/Ni surface, possibly due to the aggregation of the CNTs. Furthermore, the average roughness (Ra) of the (0.3 g l⁻¹–2 g l⁻¹) CNTs/Ni coatings were evaluated to be about 129 nm, 223 nm, 318 nm, 473 nm, and 703nm, respectively. It can be concluded that the CNTs in the Ni matrix increased the surface roughness, and the roughness of the composite surface increased with the increase of carbon tube content.

Zoom In Zoom Out Reset image size

Figure 8 shows the microhardness of the composite at different concentrations of CNTs in the plating solution. As shown in the figure, the hardness value of pure nickel material was 99.9 Hv, and the hardness of the plated layer increases with the increase of carbon tube concentration when the concentration of CNTs is relatively low. The hardness reached a maximum value of 260.21 Hv when carbon nanotubes were added at 1 g l⁻¹. Maintained adding CNTs, and the hardness started to drop. The decrease in hardness value when the carbon nanotube content stands 2 g/l was because too high CNTs caused entanglement, and agglomerates of carbon nanotubes lead to defects when deposited into the composite plating, thus reducing the hardness of the plated layer.

Zoom In Zoom Out Reset image size

In this experiment, to further improve the dispersion of CNTs via the cavitation effect of ultrasound, CNTs/Ni composites were prepared using ultrasonication and stirring. To investigate the effect of carbon nanotubes on the nucleation and growth of nickel, the CNTs/Ni composite surface was analyzed by FE-SEM, as shown in figures 9(a) and (b). The shape of the middle part of the sample revealed step-like growth, but, at the edges, nickel grew into a spherical shape. Many defects appeared on the exterior surface of CNTs such as vacancies [34]. Due to the tip effect of CNTs and the existence of defects on their surface, nickel ions were selectively adsorbed at the tips or defects. The resistance after adsorption was lower, so the nickel ions were more likely to accept electrons and be deposited. In other words, the tips and defects were higher in energy and in an active state than other sites, which was more conducive to the adsorption and reduction of nickel ions. This resulted in the spherical particles.

Zoom In Zoom Out Reset image size

To further explore the dispersion of CNTs in CNTs/Ni composites and the bonding of CNTs to the substrate interface, TEM was used to investigate the CNTs/Ni composites. As presented in figure 9(c), the CNTs were in the coatings and were uniformly dispersed due to ultrasonication and stirring. At the same time, small-scale nickel crystals were observed. Furthermore, the addition of CNTs refined the grain size. The cross-sectional HRTEM image of the CNTs is shown in figure 9(d). It is clear from the image that the CNTs were approximately 20–40 nm in diameter had a circular texture, which is typical of multi-walled carbon nanotubes. Due to the surface treatment of CNTs, the wettability and tight bonding at the interface between CNTs and Ni were substantially improved, which promoted the overall performance of the composites.

Figure 10 shows the XRD patterns of the CNTs/Ni composites and pure Ni prepared by electrodeposition. From the XRD patterns (figure 10(a)), the diffraction peaks of Ni (JPCDS card No. 04-0850), it can be seen that the CNTs changed the preferential growth orientation of nickel particles from (200) to (111). Furthermore, the peak intensity gradually decreased as the CNTs content increased. Diffraction peaks of CNTs were not detected due to their low loading. From the partially-enlarged details (figure 10(b)), the half-peak width of the CNTs/Ni composites increased, which was related to grain refinement and internal stress. In addition, the introduction of CNTs caused the splitting of the nickel diffraction peak (222) in contrast to pure nickel, which was related to the generation of twin crystals. According to Qu's work [35], suspended carbon nanoparticles adsorbed ions in the electrolyte, such as Ni²⁺. When carbon nanoparticles were wrapped into the nickel-based growth center, the adsorbed ions freed the growth center from the cation in the electrolyte and further prevented the growth of nickel particles [36]. In this case, the probability of nucleation on the (220), (311), and (222) planes increased so that their peak intensities were higher than those of pure Ni.

Zoom In Zoom Out Reset image size

Finally, relative texture coefficient (RTC) of each diffraction crystal plane (h k l) of the CNTs/Ni composites and pure nickel were calculated by equations (1) and (2).

$$\begin{eqnarray}&&RT{C}_{\left(hkl\right)}=\displaystyle \frac{{R}_{\left(hkl\right)}}{\displaystyle \sum {R}_{\left(hkl\right)}}\times 100 \% \end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&{R}_{\left(hkl\right)}=\displaystyle \frac{\left.{I}_{S\left(hkl\right)}\right)}{{I}_{P\left(hkl\right)}}\end{eqnarray} \tag{ 2 }$$

where I_s and I_p are the diffraction peak intensity of the CNTs/Ni composites and the non-deformed standard nickel powder, respectively. The correlated RTC_{(h k l)} values are listed in table 3. Furthermore, the pure nickel coating had a strong [200] texture, and RTC₍₂₀₀₎ was calculated to be about 74.5%. However, RTC₍₂₀₀₎ of Ni/CNTs composites were lower than those of pure Ni. Moreover, upon increasing the CNTs concentration, the RTC ₍₁₁₁₎ and RTC ₍₂₀₀₎ of Ni/CNTs composites gradually decreased. This indicated that the CNTs played an important role during the co-deposition process.

The scratch test method has been used for a long time and is now capable of making quantitative measurements. The scratch test is based on the observation of the layer failures induced by a loaded indenter that scratches the surface of a sample. To evaluate their frictional behaviors, changes in the coefficient of friction (COF) of 0 g l⁻¹ CNTs/Ni and 0.2 g/l CNTs/Ni are shown in figure 11(a). When a load was applied, the COF quickly reached a peak, and then the indenter began to slide from the maximum static friction. At this point, the COF represented kinetic friction, which gradually decreased to a minimum as the load was increased to 20 N. In the subsequent process, the COF gradually increased until it was essentially constant. Throughout the loading process, the curves of the CNTs/Ni composites remained below the pure nickel curve (figure 11(a)). At the end of loading, the COF were 0.5 and 0.45 for pure nickel and composite materials, respectively. Composite's COF was lowered by 10 percent compared to pure nickel, which revealed that the friction performance of the materials was improved.

Zoom In Zoom Out Reset image size

Figures 12 and 13 show that the actual length of the scratch was approximately 5 mm, and the depth of the scratch was shallow without cracks at the initial phase of the scratch. As the load F_n increased, a large compressive stress was created at the top of the head, and then the scratch depth and width increased as the load increased. As F_n increased, the deformation become prominent, resulting in a sharp increase in the local stress and energy [37]. When the stress concentration exceeded the critical stress value when the load reached 100 N, the internal phase interface or crystals released the internal stress, which produced a large radial crack, as shown in figures 12(d) and 13(d).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

For pure nickel, the partial magnification shows that the sample surface had a large number of micropores (figures 12(a)–(c)) as well as residual nickel metal debris with radial cracks at the maximum load (figure 12(d)). In contrast, after the introduction of CNTs, crumbs of metallic nickel accumulated at the edges of the scratch (figures 13(a)–(c)), and no microcracks appeared at the head of the scratch (figure 13(d)). The presence of a huge number of micropores may be one of the reasons for the reduced frictional properties of pure nickel.

Figure 11(b) shows the trend of COF with different CNT contents. 0.3 g l⁻¹ CNTs/Ni, 0.4 g l⁻¹ CNTs/Ni, and 0.5 g l⁻¹ CNTs/Ni were unstable and fluctuated greatly over time, which may have been caused by the hierarchical structure generated from the deposition conditions. Moreover, Mohamed [38] et al argue that it was observed that the COF oscillates significantly with time owing to the stick-slip phenomenon during the sliding motion. The COF-T curves of 1 g l⁻¹ CNTs/Ni and 2 g l⁻¹ CNTs/Ni samples were quite similar, with no large jump after reaching the maximum value and remaining relatively stable thereafter. Due to the higher CNTs enhancing the self-lubricating layer formation on the rubbing surfaces, the COF of 1 g l⁻¹ CNTs/Ni and 2 g l⁻¹ CNTs/Ni composite had a slight fluctuation of the friction behavior. Table 4 shows the average friction coefficient of the samples. It can be clearly observed (table 4) that there was a decrease in COF when carbon nanotube mass concentration was increased from 0.3 g l⁻¹–2 g l⁻¹, and the COF was only 0.343 when the carbon nanotube content was increased to 2 g l⁻¹. It can be seen that CNTs improved the friction wear resistance of the CNTs/Ni composites. When there were fewer CNTs in the Ni matrix, friction mainly occurred between the Ni matrix and friction pair, resulting in a higher COF. As the CNT content increased, the second phase of the surface increased, which reduced the coefficient of friction. The results showed that the wear resistance was closely related to the CNT content in the CNTs/Ni composites. Due to the formation of a carbon film on the wear surface, the self-lubricating effect of the carbon film decreased the friction coefficient of the composites, which is consistent with Choi's [39] work on Al-CNTs composites.

To further investigate the wear mechanism, SEM characterization was performed. Figures 14 and 15 show the surface after 30 min wear at a constant load of 30 N and at 10m min⁻¹ sliding speeds. Figure 14 shows FE-SEM images of the tribometer tracks of the 0.5 g l⁻¹ CNTs/Ni composites. As shown in figure 14(a), stress was concentrated on the composite surface, producing microcracks and plastic deformation. The specimen presented the coarse worn surface, and it can be observed that spalling pits and peeling are adhesive wear characteristics [40]. As the wear time increased, peeled particles (figures 14(b)–(d)) were crushed and appeared at the wear interface, giving rise to distinct furrows on the wear surface. However, most of the frictional wear area remained as smooth furrows (figure 14(a)), despite the presence of sheet debris. Since the content of CNTs was low, the surface peeling was more severe during the wear process, cracks were formed, and severe plastic deformation occurred on the grinding surface (figure 14(c)). The area of plastic deformation of the surface visible in the figure was mainly the shell layer formed by the tearing and flaking of the nickel matrix during the friction process. Meanwhile, it can be seen from figures 14(b) and (d) that the scratches were deeper. There were more abrasive chips, and no CNTs were observed to be distributed on the sample surface, so only a tiny amount of carbon-based lubricating film was formed during the friction process, and the specimen was worn drastically. The wear mechanism was mainly manifested as abrasive wear and delamination wear.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

As shown in figure 15, the corresponding FE-SEM image (figure 15(a)) showed a small-width track with many worn metallic particles, which was pushed down under shear and smeared onto the contact surface to form a thick and continuous carbon film that reduced the friction. Exposed CNTs are shown in figure 15(d). One edge of these CNTs was deeply embedded in the nickel-based metal. In contrast, the other ends were exposed outside the nickel-based metal with a maximum diameter significantly smaller than the middle part, which was related to the loading, which ground most carbon nanotube into pieces and even broke some of them. The FE-SEM images (figures 15(b) and (c)) show that both surfaces produce large parallel furrows, mainly because abrasive dust was embedded in the sample surface, suggesting the occurrence of abrasive wear in the composite. At a low CNTs co-deposition, the surface roughness caused micro-contact between the ball and substrate, which caused wear on the asperities. However, increasing the CNTs content in the deposited coating resulted in both grain refinement and improved friction, which decreased the real contact area.

Judging from the abrasion's appearance, both 0.5 g l⁻¹ CNTs/Ni and 1 g l⁻¹ CNTs/Ni mainly underwent stick wear and abrasive wear, but the details were slightly different. With the increase of CNTs content, compared with the 0.5 g l⁻¹ CNTs-Ni sample, the worn surface of the 1 g l⁻¹ CNTs-Ni sample became smooth and delicate, and the texture was directional. The spalling layer was not found, indicating that the content of CNTs was an essential factor affecting the lubrication performance of the composites. The carbon-based lubricating film formed during the friction process of CNTs can effectively inhibit the spalling of the sample during the friction process and reduce the plastic deformation and scratch of the worn surface. The wear mechanism was mainly changed from delamination wear to abrasive wear.

As can be easily seen in figures 14(b) and (d), the material's surface has deep and wide furrows and some lamellar abrasions on the surface, which were caused by repeated plastic deformation of the material as a result of strain hardening. Figures 15(b) and (c) show the SEM with the addition of 1 g l⁻¹ CNTs, where the plow grooves on the material's surface had become shallow. There was no noticeable presence of lamellar abrasive debris, indicating that the addition of CNTs had improved the plastic rheology resistance of the composite, which in turn had changed the wear resistance of the composite. This was due to the self-lubricating effect of CNTs. During the friction process of the composite material, CNTs will be extruded from the material to reduce the direct contact between the matrix and the friction pair. As the CNTs content increases, the extruded CNTs gradually increase, further reducing the direct contact between the matrix and the friction pair, thus gradually reducing the friction factor of the composite material. In addition, because the CNTs were tubular, there was rolling friction in the sliding friction of the material, thereby reducing the friction factor.

To further understand the wear mechanism of the CNTs/Ni composites, EDS analysis of the wear chips was carried out by FE-SEM (figures 14(c) and (d) and 15(c) and (d)). The main elemental composition of the scrap included Ni, C, O, and Si. The Si may have originated from the indenter or from the incomplete cleaning of the sample after grinding and polishing. The wear track showed a significant amount of carbon across the wear surface, which may be due to the formation of a carbon film on the sample during the wear test. The self-lubricating effect of this film reduced the coefficient of friction of the material.

Friction wear is usually a complex mechanism that involves several different mechanisms occurring simultaneously. When the composite surface first contacted the friction pair, adhesive wear occurred, and then the sample surface was peeled and broke under the positive pressure and friction resistance, thereby forming grinding debris. Furthermore, CNTs were pulled out under the tangential force and then produced free CNTs debris. During friction pair sliding, the free CNTs fragments were constantly enriched on the friction subsurface, avoiding the direct contact between the composites and the friction pair, and providing a good lubrication effect. Kim [41] also believed that the carbon film acted as a solid lubricant when it covered the sample surface, thus reducing the coefficient of friction. When the CNTs content was low (0–0.5 g l⁻¹), there were not enough damaged CNTs to form a carbon film on the friction subsurface. Meanwhile, a small amount of free CNTs fragments constantly scratched the sample surface and then aggravated damage to the sample surface, which increased the roughness of the sample surface, thus increasing the COF of the composites. Upon increasing the CNTs content, broken CNTs fragments were constantly enriched on the friction subsurface. When there were enough CNT fragments, the enriched CNTs fragments formed a continuous carbon film on the subsurface, which prevented the direct contact between the composites and the friction side, thus greatly reducing the COF. Furthermore, because some of the carbon nanotubes protrude from the surface, the direct contact between the metals was cut off, and the friction coefficient will be decreased, while the carbon nanotubes still adhere to the coating and the friction pair after breaking off, which can still play the role of lubrication. In addition, the CNTs in the coating were distributed in the form of network and winding, which increased the bonding force of the composite phase and made it less likely to be pulled out and fall off during wear, thus improving the wear resistance of the material.