*Article* **An Investigation for Minimizing the Wear Loss of Microwave-Assisted Synthesized g-C3N4/MoS2 Nanocomposite Coated Substrate**

**Mukul Saxena 1, Anuj Kumar Sharma 1,\*, Ashish Kumar Srivastava 2,\*, Narendra Singh <sup>3</sup> and Amit Rai Dixit <sup>4</sup>**


**Abstract:** Mechanical components frequently come into contact against one another causing friction that produces heat at the contact area and wear of the components that shortens part life and increases energy consumption. In the current study, an attempt was made to optimize the parameters for the pin-on-disc wear tester. The experiments were carried out in ambient thermal conditions with varying sliding speeds (0.5 m/s, 0.75 m/s, and 1.0 m/s) and applied loads (5 N, 10 N, and 15 N) for pure molybdenum disulfide with 9% and 20% weight percentage of graphitic carbon nitride (g-C3N4) in molybdenum-disulfide (MoS2)-nanocomposite-coated steel substrate. Analysis of variance (ANOVA) was used to determine the outcome of interaction between various constraints. To identify the minimum wearing conditions, the objective was defined as the criterion 'smaller is better'. The maximum impact of the applied load on the coefficient of friction and wear depth was estimated to be 59.6% and 41.4%, respectively, followed by sliding speed. The optimal condition for the minimum coefficient of friction and wear was determined to be 15 N for applied load, 0.75 m/s for sliding speed, and weight percentage of 9 for g-C3N4 in MoS2 nanocomposite. At the 95% confidence level, applied load was assessed to have the most significant effect on the coefficient of friction, followed by sliding speed and material composition, whereas material composition considerably impacts wear, followed by loading and sliding speed. These parameters show the effect of mutual interactions. Results from the Taguchi method and response surface methodology are in good agreement with the experimental results.

**Keywords:** optimization; tribology; ANOVA; transition metal dichalcogenide; graphitic carbon nitride; nanocomposites

#### **1. Introduction**

The fundamental reasons for breakdown in many moving parts employed in mechanical work are friction and wear. A significant amount of energy is expended to eliminate frictional resistance. A result of friction is wear and heat, which can lead to noise emissions, material fatigue, mechanical losses, surface deterioration, and shorter component service life [1]. The cost of maintenance, machinery, and fitting due to wear and tear with frictional problems has an impact on a company's economy [2]. Enhancing the tribological characteristics of the contacted surfaces is a primary technique to reduce energy usage. Enhancing the tribological characteristics of the mating surfaces is a primary technique to reduce energy usage. The need for improved lubricants is increasing due to its properties including the capacity for use throughout a wider temperature range, higher loads and speeds, and durability with operational life [3]. For fulfilling the demands of machinery in

**Citation:** Saxena, M.; Sharma, A.K.; Srivastava, A.K.; Singh, N.; Dixit, A.R. An Investigation for Minimizing the Wear Loss of Microwave-Assisted Synthesized g-C3N4/MoS2 Nanocomposite Coated Substrate. *Coatings* **2023**, *13*, 118. https:// doi.org/10.3390/coatings13010118

Academic Editor: Manuel António Peralta Evaristo

Received: 14 December 2022 Revised: 3 January 2023 Accepted: 5 January 2023 Published: 8 January 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

harsh situations, unique and efficient friction-resistant lubricants with high-load-bearing additives must be developed [4].

Two-dimensional nanoparticles have a larger specific surface area than other nanomaterials for absorbing onto a substrate's surface, which reduces or eliminates friction between the contacted surfaces [5]. Molybdenum disulfide (MoS2) was discovered as a potential 2D transition-metal chalcogenide (TMD) due to its chemical stability and can be used as a coating material. Lower fiction is caused by a weakened Van der Waals attraction connecting the sulfur lamella [6]. The lubrication effect is produced by the molybdenum disulfide's layers, which effectively slide across each other and align parallel to the relative movement while sliding. The lamella, on the other hand, is particularly resistant to asperity penetration due to the strong ionic interaction between molybdenum and sulfur atoms [7]. Molybdenum disulfide in pure form, on the other hand, easily absorbs moisture and could be oxidized in any environment containing either molecular or atomic oxygen [8]. This produces a rapid growth in the frictional coefficient and deterioration in the durability of the frictional surface [9,10]. As a result of these limitations, its practicality is constrained and limited. Improving the coefficient of friction and wear life of molybdenum sulfide used as a lubricant (solid) in various domains of areas is becoming a serious problem. Increasing the wear behavior of MoS2 for practical usage as a lubricant in various industries while retaining a low frictional coefficient is currently recognized as a significant concern [11].

Using a four-ball tribometer, the tribological characteristics of the nanoparticle mixture consisting of lubricating oil and MoS2 under different friction conditions were examined and modeled. The MoS2 particles' influence on the coefficient of friction and severe load of the lubricating mixture was due to the easier adsorption of MoS2 particles on the sliding surfaces of the ball and the creation of MoO3, which was a protective and lubricating coating. This was made feasible by the relatively simple oxidation of MoS2 during the sliding process, which occurred after its release and passage through the valley onto the contact metal surfaces and their separation at the interface [12]. Nanoparticles developed on the highly-oriented pyrolytic (HOP) MoS2 and graphite were utilized for a model to examine the dependency of frictional forces on the contact area. The power dissipation required for nanoparticle translation was linearly dependent on the contacting area in between the substrate and MoS2 nanoparticles [13].

Graphitic carbon nitride (g-C3N4), which has weakened forces (Van Der Waals) in between the layer and tri-s-triazine unit, is now used in a range of disciplines [14,15]. For improving frictional performance, graphitic carbon nitride is often added to the lubricating medium. The bonding of graphitic carbon nitride with the octadecylamine, for example, led into the creation of boundary layers on surfaces, which improved the resistance to wear of material [16]. Duan et al. [17] selected g-C3N4 as a basic oil additive since it significantly increases thermoset polyimide wear resistance. Zhu et al. [18] developed g-C3N4/PVDF composites and observed that the g-C3N4 filler improved composite wearing.

Austenitic steel, which has a strong corrosion resistance, is the most commonly utilized substrate material. However, when analyzed in a tribological investigation, performance suffers due to various types of wear. Surface modification is probably the most often utilized strategy for mitigating these issues. Carbonitriding, nitriding, carburizing, coating or cladding, and other methods are commonly employed to improve the surface qualities of steel substrate [19]. Microwave radiation heats the substance at the molecular scale, resulting in uniform volume heating. Because the heating begins at the molecule across the bulk, the process is far faster than conventional methods of heating in which heating of the material is dependent on the traditional modes of transmission of heat [20]. However, microwave energy's adaptability in processing metal components is difficult due to the coefficient of absorption for metallic substance radiation at 2.45 Ghz being substantially lower at ambient temperature [21]. Cladding is often defined by partial substrate dilution and the formation of metallurgical adhesion between the deposit and substrate. Gupta et al. [22] developed a novel technique in surface improvement techniques via microwave cladding. Microwave hybrid heating (MHH) was used to create clads using tungsten-carbide-based

powder (WC10Co2Ni) over steel substrate [22]. The homogeneous distribution of the clad element across the substrate eliminates interfacial flaws and fractures. This contributes to improving the material's microstructure [23].

To analyze the tribological characteristics of friction set, a scientific method is required due to the intricacy of the wearing process. Design of the experiment (DOE) is among the most essential statistical analyses for investigating various process parameters by minimizing the number of multiple trials. The Taguchi method provides the most costeffective technique for the design of experiments since it is possible to construct a strong design by using expensive parts with high quality materials or even adjusting process factors, but these solutions are rarely effective [24]. Dr. Genichi Taguchi created a range of unique statistical tools, concepts, and methodologies for increasing product quality, most of which are subjected to the statistical theory of DOE [25]. Taguchi describes how he developed his methodology by using a design of experiment to develop a system that can endure a wide range of conditions and fluctuations as well as lowering the target value to limit fluctuation [26]. The Taguchi method was chosen because it is a cost-effective method [27]. Optimization is the process of determining the effective or most efficient use of any condition or resource [28]. The Taguchi design method removes superfluous experimentation from the process. ANOVA analysis is then performed. As a result, the critical characteristics that determine the wear rate are identified. ANOVA is also required to identify how much of any wear process parameter contributes to material wear loss [29,30]. As a consequence, when combined with analysis of variance (ANOVA), Taguchi's testing method provides a robust instrument for analyzing the effect of various process parameters [31–34].

The aim of this study is to achieve the optimal results in terms of establishing a minimum coefficient of friction and wear depth by analyzing the impact of various process parameters, such as applied load and sliding speed, on different coating material compositions under dry sliding conditions for g-C3N4/MoS2-nanocomposite-coated substrate with different weight percentages of g-C3N4 in the nanocomposite, which has never been performed before. The findings of this study give insight on the selection of a combination of parameters to achieve the minimum wear and the coefficient of friction.

#### **2. Materials and Methods**

For the tribological investigations on coating, a pin-on-disc POD testing apparatus (TR20LEPHM400, DUCOM Instruments, Bangalore, India) was used. Figure 1 depicts the experimental setup for the testing. The counterpart pin is made of AISI304 grade stainless steel and listed in Table 1 together with the substrate's composition, which was determined by energy dispersive X-ray analysis.

**Figure 1.** The experimental setup for tribological investigation.


**Table 1.** Composition of disc and pin material [35].

The ambient temperature was nearly 24 degrees Celsius, and the relative humidity (RH) was around 30%. This environment was used for all of the studies. Microwave-assisted synthesized pure molybdenum disulfide (MoS2) with 9% and 20% weight percentage of graphitic carbon nitride (g-C3N4) in MoS2 nanocomposite was taken as coating material for the analysis. Separately, synthesized MoS2 powder and g-C3N4/MoS2 nanocomposite were dissolved and dispersed in absolute ethanol (99.9%). After 30 min of ultrasonication, the solution was then deposited on the substrate with a spin coater (Holmarc, CAS, Lucknow, India).

At low rotating speed, the coating solution spreads across the substrates, and at high rotating speed, coated films were formed. First, the substrate was placed on the spin chuck. Turning on the vacuum liner kept the substrate in place. A preset volume of coating material was dispensed onto the substrate disc with a disposable pipette to coat it. After applying the coating solution, the substrate was spun, and the lid was fixed on the spin-coating. The MoS2 and g-C3N4/MoS2 nanomaterial suspensions were coated on the substrate once the solution was completely suspended. The thickness of the coating on substrate was then calculated from SEM analysis of the cross section of the coated substrate and found to be approximately 4.5 μm as shown in Figure 2.

**Figure 2.** Coating thickness measurement from scanning electron microscopy.

Tribological experiments were carried out on POD tester with varying sliding speeds (0.5 m/s, 0.75 m/s, and 1.0 m/s) and applied loads (5 N, 10 N, and 15 N). The radius of the disc and pin utilized in the test were 31.75 mm and 4 mm, respectively.

After studying past research on the effect of input parameters on output parameters and the prediction of responses, the Taguchi method and response surface method was chosen for the prediction of response parameters. The objective of this analysis was to determine the correlation between the operating and the response parameters of the nanocomposite-coated disc and counterpart pin. As shown in Table 2, the signal-to-noise

(SN) ratio was selected as a performance metric. This ratio evaluates the output's convergence to such an objective under various noise situations. The following was the formula:

$$SN = -10\log\frac{1}{n}\left(\sum y\_i^2\right) \tag{1}$$

where '*n*' represents the observation number and '*yi*' represents the data. The noise is denoted by the variable '*N*.' The letter '*S*', on the other hand, represents the signal. SN ratios were determined with the experiment purpose of 'smaller is better' in order to reduce frictional coefficient and depth of wear between the surfaces. The response was calculated using the results of adjusting a process parameter. The optimal wear conditions of a microwave-aided synthesized nanocomposite-coated steel disc were identified. The noise was decreased by altering the dependent variables. It was challenging to adjust the external variables to alter noise. These consist of applied load, sliding speed, and material composition. The procedure determines the optimal process variables for decreasing the wear depth and coefficient of friction. The interactions between the factors were also taken into account.


**Table 2.** Signal-to-noise (SN) ratios with the objectives and their meanings [36].

The orthogonal arrays were utilized to build the experiment design and engineering optimization. The detected input factors and related levels are listed in Table 3. The levels were chosen as per the specification of the instrument available at the center.

**Table 3.** Control factors with respective levels for analyzing wear behavior of the coating material.


#### **3. Results and Discussion**

#### *3.1. Design of Experiment (DOE)*

The most fundamental strategy for simultaneous examination of the numerous components affecting the process is the design of experiment. It not only limits the number of trials that must be conducted, but it also specifies the research projects required to reach the objective. Effective factor identification was required to evaluate the methodological approach [32–34]. Employing a Taguchi L-9 array, the combination of parameters used in the process for each experiment in the present investigation was obtained. Wear depth was estimated using the pin-on-disc tribometer, which was equipped with a linear-variable differential transformer (LVDT) sensor with a range of up to 2000 μm with least count of 0.1 micron and accuracy of 1 ± 1% of calculated wear. Table 4 contains the parameters of all 27 trials as well as the experimental results for coefficient of friction and wear in a previous investigation [35].


**Table 4.** Design of experiment (DOE) with experimental results.

#### *3.2. Signal-to-Noise (SN) Analysis of Wear Depth and Coefficient of Friction*

The purpose of the study was to determine the critical parameters impacting the wear mechanism as well as the related conditions for minimum wear depth and coefficient of friction. Figure 3 depicts the normal probability plot of the experiment. Because there was no indication of skew in the probability line, the line indicates the normality of the data distribution. Aside from the factors of interest, the graph shows no slope. This means that there was no influence of any unknown variables or other substantial variables influencing the response.

**Figure 3.** Normal probability plot of residuals when response is SN ratio (**a**) COF and (**b**) wear depth.

Figure 4 shows the main effects plot for the experiment. The slope of the plots showed that the applied load and weight percentage of g-C3N4 in the composite have the significant impact on the COF and wear, respectively, for the current experimentation state. As a result, any small variation in these parameters results in a significant difference in the COF and wear of the substrate.Therefore, substantial changes to this parameter are restricted. It can be seen from Figure 4 that the sliding speed was the least impacting parameter. With the associated optimum environment and wear mechanism, the effect of wear on control limits was investigated. Based on the SN ratios and data means, as reported in Table 5, the impact of input conditions was examined. The levels for every factor were established with the concept that they would indicate the range needed for the analysis, from low loading conditions to high. The 'data means' were the factor means for every factor and level combination, as presented in Table 5. The quantity of effects or the delta was calculated by subtracting the highest to lowest averages for any factor. When examining rank in the response table, it is simpler to understand which factor has the greatest impact.

**Figure 4.** The main effect plots for SN ratios (**a**) COF and (**b**) wear depth at different parameters, such as applied load, sliding speed, and g-C3N4 weight percentage in nanocomposite.

**Table 5.** Response table for SN ratio for COF, wear depth, and the corresponding rank of process parameters.


The factor with the largest delta value has been assigned rank 1, followed by the factor with the second largest delta and so on. The SN ratio for input parameter behavior was statistically significant. In Table 5, each delta, rank, and factor level's signal-to-noise ratio is given in each row. The table has a column for each factor. The same method as before was used to calculate delta and rank. In the experiment, the applied load had the greatest influence on the coefficient of friction, followed by sliding speed and g-C3N4 weight percentage. Figure 4a,b and Figure 5 depict the principal design of control for the SN ratio, data means, and the interaction effects with controlling inputs. The main impression curves' bends indicate how each parameter has an impact. The most important factor was the one with the highest elevation of the line. The applied load for COF and g-C3N4 weight percentage for wear depth were shown to be the most important factors in Figure 4a,b.

**Figure 5.** Interaction effect of load, sliding distance, and sliding speed on the wear rate of friction material at different operating conditions (**a**) COF and (**b**) wear depth.

The interaction plots of the present investigation are shown in Figure 5. All of the input variables have an interaction effect on the output variable. Because of this, the variations in the results generated is not due to one of the input parameters alone but rather to the combined effects of all input parameters that are taken into account. Any interaction plot can be used to determine the existence of effects from any non-parallel factor. A non-parallel interaction indicates weak interaction; complimentary interaction indicates a strong relationship. According to the findings of this investigation, the applied load for

coefficient of friction and wt.% of g-C3N4 in nanocomposite had the most significant effects on response. Figure 5 and the contour plots in Figure 6a–c show how the various inputs interact and have an impact on wear depth and coefficient of friction, respectively. Figure 5 illustrates how the applied load surpassed the variance in other inputs. Examining how each factor's deviation affects the COF and wear depth is simple and intuitive when using the contour map.

**Figure 6.** Contour plot for wear depth and coefficient of friction for (**a,d**) applied load vs. sliding speed; (**b,e**) applied load vs. weight percent of g-C3N4 in composite; (**c,f**) applied load vs. weight percent of g-C3N4 in composite.

#### *3.3. Analysis of Variance (ANOVA) for Wear Depth and Coefficient of Friction*

ANOVA is a statistical method used to estimate processes and examine mean differences. It is used to determine the test's statistical significance. It is used to investigate the effect of applied load, sliding velocity, and graphitic carbon nitride weight percentage in

nanocomposite on output COF and wear depth in a controlled manner. The difference between averages divided by the difference across yields was used to calculate the F-value (factor value). A significant difference among specimen average with the variation of the parameters confirms the higher F-value. The related probability value was lower if the factor value was higher. The *p*-value reflects the probability of any error. Table 6 shows the correlations between the data after they were corrected at three levels. Sources from Table 6 were utilized to figure out which parameter regulates the other parameter and how significantly each single factor contributed. A confidence level of 95% was employed in this study. The source that contributed to this performance metric was statistically significant, with *p*-values < 0.05. The following equations were utilized in the analysis of variance [36–38]:

$$PQ\_T = PS\_L + PS\_{SS} + PS\_{wt} \tag{2}$$

$$PQ\_T = \sum\_{i} {n \choose n} {d\_i^2} - \frac{d^2}{n} \tag{3}$$

$$PS\_v = \sum\_{k=1}^{t} \left(\frac{Sd\_i^2}{t}\right) - \frac{d^2}{n} \tag{4}$$

where *n* is the number of repititions, *Sdi <sup>2</sup>* is the addition of the experimental trials involving constants *v* at a level *k, d* is the resultant data for all the test trails, *PQT* is the total addition of squares, *PSL* is the applied load addition on squares, *PSSS* is the sliding speed addition on squares, and *PSwt* is the weight percentage of g-C3N4-addition of squares.

**Table 6.** Analysis of variance (ANOVA) for signal-to-noise (SN) ratios.


The analysis of variance (ANOVA) table can be seen in Table 6, and it is evident from table that each of the parameters taken into consideration has a significant effect on wear behavior. The much more significant parameters for COF and wear depth were applied load and graphitic carbon nitride in nanocomposite, respectively. This could be attributed to the reason that as the applied load rises, the pressure at the contact between the pin and disc rises, resulting in a lubricating characteristic of g-C3N4/MoS2. Additionally, the wt.% of g-C3N4 enhances to the wear preservation of the coating as the MoS2 could be quickly oxidized due to theheat generated between the surfaces. Because the material of the coating (i.e., nanocomposite) was removed in powder form, the eliminated material adhered to the disc surface, reducing the raw surface contact and reducing wear. Finally, due to the support of molybdenum disulfide and its nanocomposite with graphitic carbon nitride in the lubricating mechanism, the sliding speed has the lowest impact on wear.The wear of the coating material was most significantly impacted by the applied load. As a consequence, during the wear, the applied load, followed by the other parameters, were critical control components to consider. The COF and wear depth were only slightly impacted by the interaction between the different inputs, i.e., applied load, sliding speed, and nanocomposite composition.

#### *3.4. Modeling through Response Surface Methodology (RSM)*

RSM is a multipurpose technique that can be used to construct mathematical models to predict responses, analyze surface responses using response surface curves to help explain how an input parameter affects a response parameter, analyze variance in process parameter values, and determine the optimal parameter. In order to analyze the data, determine the significance of parameters for the model, calculate the mean response, and find the optimal operating condition for the control variables which assist in achieving a minimum or maximum response over a particular interested region, a linear and seconddegree model was used in this paper.

As a result, the model was created for analyzing variance to assess the significance and stability of response as well as the process parameters. This was performed after obtaining the response parameters (Table 4).

The equations below address the mathematical model which was developed by using the MINITAB-19 to analyze the response parameters:

$$\begin{array}{l} \text{COF} = 462-\text{(7.10 } \times \text{Load\\_Applied, (N))}-\text{(0.142 } \times \text{Słiding\\_Speed (m/s))}\\ \text{ -- (1.85 } \times \text{g-C}\_3\text{N}\_4 \text{ wt.\% (\%))} \end{array} \tag{5}$$

Wear = 1069 − 48.3 Load\_Applied, (N) − 1.25 Sliding\_Speed (m/s) − 29.6 gC3N4 wt.% (%) + 2.06 Load\_Applied, (N) × Load\_Applied, (N) + 0.000740 Sliding\_Speed, (m/s) × Sliding\_Speed, (m/s) + 1.363 g-C3N4 wt.% (%) × g-C3N4 wt.% (%) (6)

> The aforementioned regression model aids in the prediction of the response parameters, i.e., wear depth and coefficient of friction. The influence and significance of the parameters and related factors on the parameters for response must now be examined using the variance analysis (ANOVA) for response surface methodology. The probability value (*p*-value) for the factors must be below 0.05 to fulfill the criterion of a factor that was significant criteria as the ANOVA was performed at 95% confidence level. Table 7 provides the ANOVA analysis for response surface methodology, which summarizes the degree of freedom, sum of squares, *p*-value, and F-value of response parameters in MoS2 and g-C3N4/MoS2 nanocomposite.



The ANOVA analysis for the coefficient of friction and wear depth is shown in Table 7 to examine the significance of the process factors and their influence on response parameters, namely frictional coefficient and wear depth. Load applied on coated substrate has

a significant effect on COF, roughly 59.6%, which was the greatest between all parameters with their factors. Additionally, in Table 7, the ANOVA analysis for wear depth is shown for analyzing the significance of process parameters with the impact on the response parameter, i.e., wear depth. It can be seen from Table 7 that the applied load and sliding speed makes a significant impact on wear depth, approximately 41.4 % and 41.2%, which was the greatest among the other process parameters with their factors.

As analysis of variance studied the effect of parameters used for process on the parameters for response, similarly, variation in the parameters for response by changing inputs can be studied by response surface plots. Figure 7 depicts the response surface curve at applied load, sliding speed, and g-C3N4 weight percentage in nanocomposite. Figure 7a–c shows variation in wear depth with applied load, sliding speed, and weight percentage of g-C3N4 in g-C3N4/MoS2 nanocomposite, and it can be analyzed that at applied load 15N, sliding speed 0.75 m/s, and 9 wt.% of g-C3N4, minimum wear depth was found. Figure 7d–f show variation in coefficient of friction with the factors and as the same combination for that of wear depth, and it was found minimized at 15 N applied load, 0.75 m/s sliding speed, and 9% of weight percentage in g-C3N4/MoS2 nanocomposite due to the combined effects of response parameters and their interactions as the study was to optimize the wear depth with COF generated in the wear process.

**Figure 7.** Response surface for wear depth and coefficient of friction for (**a,d**) applied load vs. sliding speed; (**b,e**) applied load vs. weight percent of g-C3N4 in composite; (**c,f**) applied load vs. weight percent of g-C3N4 in composite.

Attributed to the reason that molybdenum disulfide can quickly oxidize at high temperatures between the pin and disc surfaces while used in pure form, and MoS2 predominates

when the wt.% of g-C3N4 in the composite increases, the reduction in wear depth and coefficient of friction were higher in the case of g-C3N4 at a wt.% of 9 in nanocomposite.

A response optimizer for experimentation was developed by response surface method and is depicted in Figure 8. The optimization of the process parameters to achieve the lower COF and wear depth is illustrated by the red-colored line, which was approximately same as obtained by the Taguchi method.

**Figure 8.** Plot for optimization for wear COF and wear.

The optimum results for response factors (i.e., coefficient of friction and wear depth) obtained from the Taguchi method and response surface methodology (RSM) with experimental results are summarized in Table 8. In all methods, the results are very comparable with each other. A minor difference was found while predicting the responses through Taguchi and RSM. The coefficient of friction varies by approx. 13% and 15% and wear depth by approx. 5% and 12% from the experimental findings while predicting the COF and wear depth through Taguchi and response surface methods, respectively.

**Table 8.** Calculated coefficient of friction and wear depth from different methods.


#### *3.5. Wear Mechanism*

Figure 9 shows FESEM images of the worn surfaces of an uncoated steel substrate disc and a g-C3N4/MoS2 (9 wt.%)-nanocomposite-coated steel substrate disc, after a wear test at 0.75 m/s sliding speed and 15 N applied stress. Figure 9 depicts the amount of wear loss caused by plastic deformation and ploughing. The wearing was heavily influenced by abrasive wear. Several deep scratches and craters, as well as small micro cracks with sheared off asperities of pin material, were detected while inspecting the worn surface. Figure 9a also shows some sheets that indicate the creation of an oxide layer during wear.

**Figure 9.** FESEM images of worn disc surface for (**a**,**b**) pure MoS2 disc and (**c**,**d**) g-C3N4/MoS2 (9 wt.%) nanocomposite coated disc.

Figures 10 and 11 show the EDX elemental analysis results of a worn disc coated with pure MoS2 and g-C3N4/MoS2 (9 wt.%) nanocomposite, respectively. The results of EDX in Figure 11 corroborate the conclusion that the groove included g-C3N4/MoS2 and that the elements were carbon, nitrogen, molybdenum, and sulfur.

Some oxygen elements were also detected in the EDX mapping in Figure 10 as a result of the oxidation of molybdenum disulfide into molybdenum oxide owing to excessive heating while wearing, however it can be seen from Figure 11 that oxidation was decreased to some extent when g-C3N4 was added in the nanocomposite.

Figure 12 shows the FESEM image of worn out of corresponding pin. Worn surfaces of the counterpart pin used for the tribo test against the g-C3N4/MoS2 (9 wt.%) nanocomposite-coated disc (Figure 12a) and FESEM images of the counterpart pin used for the pure MoS2-coated disc (Figure 12b) are depicted. The images indicate that rubbing a pin against a pure MoS2-coated disc causes it to wear off unevenly and roughly. This is because the MoS2 was highly oxidized by the heat created by the rubbing of the pin against the disc and the creation of its oxide. The depiction also illustrates ploughing and abrasive wear grooves on the pin. However, when g-C3N4 was included in the nanocomposite, the wear of the pin was found to be less than when the pin was rubbed against the disc coated with pure MoS2, and a comparably smaller quantity of coating was transferred from disc to pin, ensuring adequate adhesion. The worn area on the pin used for the wear test against

the g-C3N4/MoS2 (9 wt.%)-nanocomposite-coated disc was about 7.671 mm2, which was 23% less than the worn area on the pin used for the pure MoS2-coated substrate. This demonstrated that the inclusion of g-C3N4 has a considerable impact on wear reduction.

**Figure 10.** EDX elemental mapping of MoS2-coated disc after wear test at 15 N applied load and 0.75 m/s sliding speed.

**Figure 11.** EDX elemental mapping of g-C3N4/MoS2 (9 wt.%) nanocomposite coated disc after wear test at 15 N applied load and 0.75 m/s sliding speed.

The results of an energy-dispersive X-ray spectroscopy examination of the pin surface that was rubbed against the g-C3N4/MoS2 (9 wt.%)-nanocomposite-coated disc are depicted in Figure 13, confirming that the material transferred from the disc to the pin surface after the wear test was the g-C3N4/MoS2 nanocomposite.

**Figure 12.** FESEM images of worn pin surface (**a**) of counterpart pin for g-C3N4/MoS2 (9 wt.%) nanocomposite-coated disc and (**b**) of counterpart pin for MoS2 coated disc.

#### **4. Conclusions**

The current effort was aimed at the determination of appropriate process parameters for a pin-on-disc (POD) wear tester to analyze the coating material that may result in a low frictional coefficient between the mating surfaces of pin and disc and minimum wear of the substrate disc. For this, discs were coated at different weight percentages of g-C3N4 in molybdenum disulphide (MoS2), and coated discs were tested using a pin-on-disc machine wear testing machine under different operating parameters, such as applied load ranges from 5 to 15 N and sliding speed from 0.5 to 1.0 m/s. The Taguchi method was used to develop a design of experiment and the ANOVA method was employed to find the significance of the process parameters. The results obtained from the RSM method were used to build the mathematical model. After a prolonged wear test, worn surfaces showed evidence of abrasive wear with ploughing, and there was a transfer of coating from the substrate disc to the counterpart pin. The following are the major findings of this study:


**Author Contributions:** M.S.—Conceptualization, Writing; A.K.S. (Anuj Kumar Sharma)—Review and editing; A.K.S. (Ashish Kumar Srivastava)—methodology, validation, N.S.—Review and editing; A.R.D.—supervison; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

**Han Wu 1,2, Ying Jiang 1, Wenjing Hu 1,\*, Sijing Feng 3,\* and Jiusheng Li 1,\***


**Abstract:** To develop a high-performance additive that can meet different operating conditions, three liquid crystals (LCs) were developed as additives for a base oil. The structures and thermal stabilities of the obtained LCs were characterized by nuclear magnetic resonance (NMR), Fourier transform infrared (FT-IR) spectroscopy, mass spectroscopy (MS), and thermogravimetric analysis (TGA). The effects of mesogenic-phase temperature ranges on tribological properties were analyzed using differential scanning calorimetry (DSC) and polarized optical microscopy (POM). UMT-TriboLab friction and wear tester was used to study the friction-reducing properties of LCs. The width of wear marks was observed by a Contour GT-K 3D profiler to illustrate the anti-wear performance of LCs. The friction surface was characterized by scanning electron microscopy (SEM) and Raman spectroscopy. It was demonstrated that, in comparison with the base oil, the addition of LCs caused a remarkable reduction in the coefficient of friction (21.57%) and wear width (31.82%). In addition, LCs show better tribological abilities in the mesogenic-phase temperature ranges. According to the results, we demonstrated that LCs can be used as lubricant additives, especially for several operating conditions under specific temperatures.

**Keywords:** liquid crystals; mesogenic-phase temperature ranges; lubricant additives; tribological property

#### **1. Introduction**

With the increased mechanization of modern society, energy loss caused by friction and wear has been constantly increased [1–3]. Friction can be greatly reduced by using lubricants, thus lowering the energy waste [4–6]. Lubricants can form a stabilized film of lubrication on the surface of the mechanical part and prevent direct contact with the microconvexity of the metal surface [7]. Additives, which can tune the tribological behavior of the oil-based lubricant, are a type of important materials that help develop different lubricants for different working scenarios [8,9]. Therefore, there is a need to explore high-performance lubricant additives for meeting the growing lubrication requirements resulting from the development of mechanical society.

Liquid crystals (LCs) are special materials with rheological properties, having the fluidity of a liquid-crystal (LC) and the anisotropy of a single crystal, which can be a long-range ordered arrangement at the molecular level [10–12]. Moreover, due to the low viscosities of LCs, they can effectively fill the contact surface during the sliding process, resulting in reduced friction [13]. Because of these properties, LCs are considered new tribological materials with applications in the field of lubrication. LCs are characterized by excellent heat resistance and chemical stability, as well as an orderly transformation of molecular orientation and arrangement order when stimulated by an external electric or magnetic field or temperature [14]. Due to LCs' controllable nature, they can serve as

**Citation:** Wu, H.; Jiang, Y.; Hu, W.; Feng, S.; Li, J. Effect of Mesogenic Phase and Structure of Liquid Crystals on Tribological Properties as Lubricant Additives. *Coatings* **2023**, *13*, 168. https://doi.org/10.3390/ coatings13010168

Academic Editor: Ashish Kumar Srivastava

Received: 9 December 2022 Revised: 9 January 2023 Accepted: 10 January 2023 Published: 12 January 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

potential materials for tribological properties in a specific temperature range. Existing research reported that the use of LCs can significantly improve the tribological properties of a base oil when used as lubricant additives. Yang's team [15] reported that 1,3-diketone EPND (1-(4-ethyl phenyl) nonane-1,3-dione), a nematic LC lubricant additive with excellent tribological properties, is particularly suitable for the severe operating conditions of low load and friction. Additionally, Mokshin [16] reported that cholesteryl stearate and fatty acid cholesterol esters LCs were used as lubricant additives. The tribological properties of LCs were evaluated, and the addition of LCs was found to significantly reduce the coefficient of friction (COF). Ghosh and colleagues [17] blended three different series of LCs with polydecyl acrylate as lubricant additives. The study indicated that the addition of very small amounts of LCs could significantly improve the additive properties of polydecyl acrylate. Among the various types of LCs, cyanobiphenyl-type LCs were focused on due to their ability to form ordered molecular films on the friction surface [18].

The rigid structure in the molecule of a cyanobiphenyl-type LC provides orderliness, and the alkyl chains in the molecule offer the mobility of mesocrystals [19]. Related experiments have demonstrated that cyanobiphenyl-type LCs subjected to an applied electric field can increase their viscosity in the direction perpendicular to the wear surface, which facilitates the formation of a boundary lubrication state [18]. In addition to the electric field, the tribological properties of LCs have been affected by their structure. Polar molecules such as benzene rings, ester groups with longer alkyl chains, or ether groups in the structure usually exhibit better tribological properties, and the addition of flexible groups also improves the solubility of the compounds in base oil [20,21]. Additionally, the fluorinated substituent can be placed in a terminal position and within a terminal chain in the LC structure, and the addition of a fluorinated substituent can cause a remarkably steric effect by improving the melting point and mesogenic-phase temperature ranges [22]. On this basis, we designed and synthesized cyanobiphenyl-type LCs to evaluate their effects on tribological properties.

In this research, the objective was to investigate the effect of the structure of cyanobiphenyltype LC molecules and mesogenic-phase temperature ranges on their tribological properties. Three cyanobiphenyl-type LCs were designed and synthesized. The structures of the LCs were verified with Fourier transform infrared (FT-IR) spectroscopy, nuclear magnetic resonance (NMR) spectra, and mass spectroscopy (MS). The thermal stability was analyzed using a thermogravimetric analyzer (TGA). The characteristics of LCs were obtained using differential scanning calorimetry (DSC) and polarized optical microscopy (POM) to investigate the influence of the mesogenic-phase temperature range on the tribological properties. The influence of the structures and mesogenic-phase temperature ranges of LCs were evaluated on their tribological properties and the related lubrication mechanism was analyzed. This work provides a novel idea for developing new LC friction materials.

#### **2. Experimental**

#### *2.1. Chemicals*

Ethyl 6-bromohexanoate, 4 -hydroxy-4-biphenylcarbonitrile, 4 -hydroxy-4-biphenylcarboxylic acid, ethanol, dichloromethane (DCM), N, N-dimethylformamide (DMF), 4- Dimethylaminopyridine (DMAP), (S)-2-Octanol, hydrogen peroxide, potassium sulfate, magnesium Sulfate, ethyl acetate (EA), petroleum ether (PE) triethylamine (TEA), tetrahydrofuran (THF), and 1-(3-Dimethylaminopropyl)-3-ethylcarbodiimide Hydrochloride (EDCI) were obtained from Titan Scientific Co., Ltd., Shanghai, China. The 4-(Trifluoromethyl) phenol, methyl chloroformate, triphenyl phosphite, sodium hydroxide, potassium carbonate, ammonia water, and potassium iodide (KI), were purchased from Sinopharm Chemical Reagent Co., Ltd., Shanghai, China. The 4-(2-Methoxyethyl) phenol was purchased from Shanghai Aladdin Bio-Chem Technology Co., Ltd., Shanghai, China. Zinc dialkyl dithiophosphates (ZDDP) were obtained from Huihua Chemical Co., Ltd., Dongguan, China. The base oil used in the experiment was palmester 3970 (TTO) from Taiko Palm-Oleo Co. Ltd., Shanghai, China. All the above reagents can be used directly without further purification treatment.

#### *2.2. Measurements*

FT-IR was achieved using a Paragon 1000 (PerkinElmer, Waltham, MA, USA) with a measurement range from 4000 to 1000 cm<sup>−</sup>1. Bruker Avance III HD (400 MHz) spectrometer (Bruker, Billerica, MA, USA) for the analysis of 1H and 13C NMR spectrometer. Mass spectroscopy analysis was characterized using a MALDI TOF 7090 mass spectrometer (Bruker, Billerica, MA, USA). The SDT-Q600 Simultaneous TGA/DSC (TA instrument, New Castle, DE, USA) was used to characterize the thermal properties in a nitrogen atmosphere. The temperature was increased from 30 ◦C to 800 ◦C at a speed of 10 ◦C/min. The mesogenic-phase temperature range of the LCs was achieved with a TA DSC 25 (TA instrument, New Castle, DE, USA), and the textures were obtained with a XP-330C POM (Shanghai Caikon, Shanghai, China). The width of the wear marks on the steel plates surface after the test were measured with a Contour GT-K 3D profiler (Bruker, Billerica, MA, USA). The analysis of deposits on the wear surface following experimentation was performed using a LabRAM HR800 Raman spectrometer (Horiba Jobin Yvon, Pairs, France). The scanning electron microscope (SEM) was obtained from SU8010 (Hitachi, Tokyo, Japan).

#### *2.3. Synthesis of ethyl 6-(4-cyanobiphenyl-4-yloxy) Hexanoate (Intermediate 1)*

Anhydrous potassium carbonate (15 g, 108.5 mmol) was powdered and dissolved in 50 mL DMF. It was then poured into glassware and stirred under nitrogen protection to 50 ◦C. KI was added and stirred at 50 ◦C for 48 h. Next, 4-hydroxy-4-biphenylcarbonitrile (20 g, 102.5 mmol), ethyl 6-bromohexanoate (27.3 g, 123 mmol), and KI (0.5 g, 3 mmol) were added and stirred at 50 ◦C with 48 h. At the end of the reaction, the DMF was removed by spinning. The resultant residue was washed three times with water and finally recrystallized using 100 mL of ethanol to obtain a white powder.

#### *2.4. Synthesis of 6-(4-cyanobiphenyl-4-yloxy) Hexanoate Acid (Intermediate 2)*

Intermediate 1 (31.9 g, 94.54 mmol) was dissolved in ethanol in glassware and stirred at room temperature. Sodium hydroxide (6 g, 150 mmol) was then added, stirred for 10 min, and raised to reflux. This reaction lasted for 6 h. After finishing, the pH was adjusted to 2–4, and a large number of solids were precipitated. After extraction, the result was washed with 500 mL of water and extracted again to achieve the crude product. A white powder was obtained by recrystallizing it with 200 mL of ethanol.

#### *2.5. Synthesis of (R)-octan-2-yl 4'-((methoxycarbonyl) oxy-[1,1-biphenyl]-4-carboxylate (Intermediate 3)*

4 -hydroxy-4-biphenylcarboxylic acid (30.4 g, 141 mmol) was dissolved in 250 mL of water in glassware and stirred under an ice bath. The aqueous sodium hydroxide solution was poured into a dropping funnel and slowly dripped into the glassware. Subsequently, 30 mL of methyl chloroformate was added on a dropwise basis over 30 min and stirred overnight at room temperature. This reaction was followed by filtration to obtain the residue, which was recrystallized with 500 mL of EA to give a white powder (4 -(methoxycarbonyl) biphenyl-4-carboxylic acid).

4 -(methoxycarbonyl) biphenyl-4-carboxylic acid (16.32 g, 63.75 mmol), (s)-2-octanol (7.1 g, 54.52 mmol), and triphenyl phosphite (21 g, 67.68 mmol) were added to 120 mL of THF and stirred at −10 ◦C under nitrogen. DIAD was dissolved in 60 mL of THF and slowly dropped into glassware. After 5 h, the reaction gradually rose to room temperature and lasted for 36 h. The product was dissolved with 250 mL of DCM, extracted using 200 mL of 15% hydrogen peroxide, 200 mL of potassium sulfate solution, and 200 mL of water, in turn. After drying with anhydrous magnesium sulfate and removing the solvent, the purified product ((R)-octyl-2-yl 4 -((methoxycarbonyl) oxy)-[1,1 -biphenyl]-4-carboxylic acid ester) was obtained using column chromatography.

(R)-octyl-2-yl 4 -((methoxycarbonyl) oxy)-[1,1 -biphenyl]-4-carboxylic acid ester was dissolved in ethanol in glassware. The reaction was monitored by thin-layer chromatography with the addition of 100 mL of ammonia water. After the reaction, the solvent was

removed, and 200 mL water was added for extraction. The residue was then recrystallized using EA and PE. It was then placed in a refrigerator and filtered to obtain a white powder ((R)-octyl-2-yl 4 -((methoxycarbonyl)oxy)-[1,1 -biphenyl]-4-carboxylate, intermediate 3).

#### *2.6. Synthesis of LCs*

#### 2.6.1. Synthesis of LC-A

Intermediate 2 (2.27 g, 7.35 mmol) and EDCI (2.38 g, 12.26 mmol) were added to 100 mL of DCM with stirring. The reaction solution was white and turbid. TEA was added dropwise and warmed up to 30 ◦C to improve solubility; it stopped heating to room temperature after dissolution. Intermediate 3 (2 g, 6.13 mmol) and DMAP (0.22 g, 1.84 mmol) were dissolved in 50 mL of DCM, heated, stirred until completely dissolved, and poured into a dropping funnel. The above solution was slowly dropped into glassware to maintain the reaction for 24 h. Next, 5 mL of water was added and stirred for 30 min, and 300 mL of water was added to adjust the pH to 2 after removing the solvent. Finally, the LC-A (3.28 g, 72.36%) was obtained by column chromatography. 1H NMR and 13C NMR spectra are shown in Figure 1a,b and FT-IR is shown in Figure 2. 1H NMR (CDCl3, 500 MHz) δ 8.13–8.08 (m, 2H, 30, 32), 7.68 (d, J = 8.4 Hz, 2H, 4, 6), 7.66–7.59 (m, 6H, 1, 3, 8, 12, 29, 33), 7.56–7.51 (m, 2H, 24, 26), 7.21–7.16 (m, 2H, 9, 11), 7.03–6.97 (m, 2H, 23, 27), 5.17 (h, J = 7.0, 6.3 Hz, 1H, 36), 4.05 (t, J = 6.3 Hz, 2H, 14), 2.65 (t, J = 7.4 Hz, 2H, 18), 1.95–1.83 (m, 4H, 15, 17), 1.81–1.71 (m, 1H, 46'), 1.70–1.58 (m, 4H, 16, 45), 1.35 (d, J = 6.3 Hz, 3H, 37), 1.50–1.20 (m, 7H, 46", 44, 43, 42). 13C NMR (CDCl3,125 MHz) δ 172.04 (19), 166.03 (34), 159.67 (10), 150.72 (22), 145.23 (5), 144.49 (28), 137.82 (25), 132.58 (1, 3), 131.42 (7), 130.08 (29, 33), 129.82 (31), 128.37 (4, 6), 128.34 (30, 32), 127.09 (8, 12), 126.93 (24, 26), 122.06 (23, 27), 119.11 (2), 115.09 (9, 11), 110.09 (40), 71.85 (36), 67.76 (14), 36.11 (43), 34.29 (18), 31.77 (46), 29.19 (15), 28.92 (44), 25.67 (17), 25.45 (45), 24.66 (16), 22.61 (42), 20.13 (37), 14.09 (41). MS(ESI): calcd C40H43NO5 [M + H]+ 618.31, found 618.32. FT-IR (ATR): 2928, 2866, 2226, 1754, 1718, 1604, 1496, 1380, 1282, 1181, and 1116 cm<sup>−</sup>1.

#### 2.6.2. Synthesis of LC-D

LC-D was obtained in a method analogous to LC-A using intermediate 2 (5.5 g, 17.78 mmol) and 4-(Trifluoromethyl) phenol (3.46 g, 21.33 mmol) as reactants with a white powder product (8.06 g, 73.28%). 1H NMR and 13C NMR spectra are shown in Figure 1c,d and FT-IR is shown in Figure 2. 1H NMR (CDCl3, 500 MHz) δ 7.69 (d, J = 8.4 Hz, 2H, 4, 6), 7.67–7.61 (m, 4H, 1, 3, 8, 12), 7.56–7.50 (m, 2H, 25, 27), 7.21 (d, J = 8.4 Hz, 2H, 9, 11), 7.03–6.97 (m, 2H, 24, 28), 4.05 (t, J = 6.3 Hz, 2H, 14), 2.64 (t, J = 7.4 Hz, 2H, 19), 1.94–1.80 (m, 4H, 15, 18), 1.70–1.57 (m, 2H, 16). 13C NMR (CDCl3, 125 MHz) δ 171.51 (20), 159.63 (10), 153.17 (23), 145.22 (5), 132.58 (1, 3), 131.46 (7), 128.37 (4, 6), 128.07 (q, J = 32.8 Hz, 26), 127.09 (8, 12), 126.79 (q, J = 3.7 Hz, 25, 27), 123.85 (q, J = 273.1 Hz, 31), 122.06 (24, 28), 119.11 (2), 115.07 (9, 11), 110.11 (30), 67.70 (14), 34.21 (19), 28.90 (15), 25.64 (18), 24.55 (16). MS(ESI): calculated C26H22F3NO3 [M + Na]+ 476.16, found 476.14. FT-IR (ATR): 2948, 2866, 2226, 1752, 1604, 1494, 1328, 1174, 1126, and 1062 cm<sup>−</sup>1.

#### 2.6.3. Synthesis of LC-Z

LC-Z was obtained in a method analogous to LC-A using intermediate 2 (2.03 g, 6.56 mmol) and 4-(2-Methoxyethyl) phenol (1.2 g, 7.87 mmol) as reactants with a white powder product (2.08 g, 71.5%). 1H NMR and 13C NMR spectra are shown in Figure 1e,f and FT-IR is shown in Figure 2. 1H NMR (CDCl3, 500 MHz) δ 7.70–7.67 (m, 2H, 4, 6), 7.65–7.61 (m, 2H, 1, 3), 7.55–7.50 (m, 2H, 8, 12), 7.24–7.20 (m, 2H, 24, 26), 7.04–6.97 (m, 4H, 9, 11, 23, 27), 4.04 (t, J = 6.3 Hz, 2H, 14), 3.58 (t, J = 7.0 Hz, 2H, 29), 3.35 (s, 3H, 31), 2.87 (t, J = 7.0 Hz, 2H, 28), 2.60 (t, J = 7.4 Hz, 2H, 18), 1.93–1.78 (m, 5H, 15, 17), 1.69–1.56 (m, 3H, 16). 13C NMR (CDCl3, 125 MHz) δ 172.17 (19), 159.68 (10), 149.09 (22), 145.25 (5), 136.58 (25), 132.57 (1, 3), 131.39 (7), 129.79 (24, 26), 128.36 (4, 6), 127.09 (8, 12), 121.37 (23, 27), 119.12 (2), 115.09 (9, 11), 110.07 (33), 73.47 (29), 67.77 (14), 58.69 (28), 35.60 (31), 34.27 (18), 28.91 (15),

25.65 (17), 24.68 (16). MS(ESI): calculated C28H29NO4 [M + NH4] <sup>+</sup> 461.21, found 461.24. FT-IR (ATR): 2932, 2864, 2224, 1754, 1604, 1494, 1386, 1294, 1250, 1204, 1144, and 1108 cm<sup>−</sup>1.

**Figure 1.** (**a**) 1H NMR of LC-A; (**b**) 13C NMR of LC-A; (**c**) 1H NMR of LC-D; (**d**) 13C NMR of LC-D; (**e**) 1H NMR of LC-Z; and (**f**) 13C NMR of LC-Z.

#### *2.7. Tribological Test and Characterization of Worn Surface*

To verify the effect of the mesogenic-phase temperature ranges on the tribological properties of LCs, LCs were evaluated in a point-to-point contact mode with a UMT-TriboLab friction and wear tester (Bruker, Billerica, MA, USA). The steel ball and steel plate used in the experiment were made of 304 stainless steel, and the diameter of the steel ball was 8 mm. The experimental parameters are shown in Table 1 and the model schematic is shown in Figure 3. Tests were conducted at 25 ◦C, 50 ◦C, 75 ◦C, 100 ◦C, 125 ◦C, 150 ◦C, 175 ◦C, and 200 ◦C. Before each test, the steel ball and steel plate were sonicated using a mixture of ethanol and petroleum ether; the procedure lasted for 30 min. A computer was used to record the curve of the coefficient of friction and the sliding time. During the test, the friction-pair was completely immersed in the base oil with or without additives. Each experiment was repeated at least twice to make the experimental results accurate.

≡

≡

≡

**Figure 2.** Fourier transform infrared of LC-A, LC-D, and LC-Z.

**Table 1.** Experimental parameters of the UMT-TriboLab friction and wear tester.


**Figure 3.** Model schematic of UMT-TriboLab friction and wear tester.

After the test, the steel plates were cleaned using petroleum ether. The width of the wear surface was characterized by a Contour GT-K 3D profiler. Finally, the wear surfaces were characterized using Raman and SEM spectrometers to illustrate the mechanism of action of LCs as lubricant additives.

#### **3. Results and Discussion**

*3.1. Mesogenic-Phase Behaviors of the Synthesized LCs*

DSC and POM were used to characterize the LCs properties. DSC thermograms of the three LCs, from which the three LCs exhibit characteristics, are shown in Figure 4. LC-D exhibits the corresponding LC properties only in the heating cycle, and the mesogenic-phase temperature range is narrow. LC-A and LC-Z display mesophases as well as wide mesogenic-phase temperature ranges in both heating and cooling cycles. Therefore, compared to LC-D, LC-A and LC-Z have almost symmetrical mesophases and wider mesogenic-phase temperature ranges. From Figure 4a, LC-A shows three heat absorption peaks during the heating process. The first peak corresponds to the melting temperature (Tm), which refers to the temperature at which LC-A was melted by heat into the

LC phase; the second peak indicates that the phase transition occurred in the smectic C phase of LC-A; and the third peak shows the disappearance of the LC phase while entering the isotropic liquid phase, which is indicated using the clearing point (Td). LC-D and LC-Z have two heat absorption peaks, corresponding to Tm and Td, respectively. The phase-transition temperature and the mesogenic-phase temperature range of the three LCs during the heating process are shown in Table 2. The Tm of LC-A, LC-D, and LC-Z are 33.7 ◦C, 100.5 ◦C, and 37.8 ◦C, respectively. Depending on the mesophase temperature range, the order of the LCs is LC-A > LC-Z > LC-D, which is closely related to the molecular structure of the LCs. When compared with LC-D, LC-A and LC-Z show superior LC properties due to their longer chain lengths. Moreover, the two ester groups in the LC-A structure make it easier to form a smectic phase with a wider mesogenic-phase temperature range. Figure 5 shows the optical textures of the LCs, which were observed using POM. From Figure 5a, LC-A shows a schlieren texture at 65.9 ◦C. From Figure 5b, LC-D exhibits a fan-shaped focal conic texture at 100.6 ◦C, while Figure 5c shows the focal conic texture of LC-Z at 56.5 ◦C. The three synthetic LCs exhibit a colorful LC texture.

**Figure 4.** DSC thermograms of LCs: (**a**) LC-A; (**b**) LC-D; and (**c**) LC-Z.



**Figure 5.** (**a**) Schlieren texture of LC-A at 65.9 ◦C; (**b**) fan-shaped focal conic texture of LC-D at 100.6 ◦C; and (**c**) focal conic texture of LC-Z at 56.5 ◦C.

#### *3.2. Analysis of Thermal Stability*

Thermal stability is an index used to evaluate the anti-decomposition and anti-aging qualities of a lubricant, and good thermal stability is necessary for lubricant additives. The thermogravimetric analysis curve of TGA is shown in Figure 6. The initial decomposition temperatures of LC-A, LC-D, and LC-Z were 309.3 ◦C, 271.9 ◦C, and 340.7 ◦C, respectively, mainly due to the decomposition of ether bonds and ester groups in the structure of LCs. When compared with LC-D and LC-Z, LC-A contains two ester groups in its structure, which may be the reason for its second decomposition starting at 400 ◦C. LC-D starts its second decomposition around 340 ◦C, probably due to the presence of trifluoromethyl in its molecular structure. In addition, the final decomposition temperatures of the three LCs were 431.5 ◦C, 354.1 ◦C, and 409.9 ◦C, respectively. The analytical results show that LCs have excellent thermal stability due to their molecular structure, having the potential to be used as lubricant additives.

**Figure 6.** Thermogravimetric analysis curve of LC-A, LC-D, and LC-Z.

#### *3.3. Anti-friction Analysis*

The feasibility of LCs as lubricant additives was investigated using a UMT friction and wear tester. TTO was selected as the base oil, and 2 wt% of LCs were added as an additive. A comparison was made with the commercial lubricant additive ZDDP, thus illustrating the anti-friction effect. Figures 7 and 8 display the friction coefficient curves and average friction coefficient over a temperature range for pure TTO and TTO containing 2 wt% ZDDP, LC-A, LC-D, and LC-Z, respectively. The tested temperature points were 25 ◦C, 50 ◦C, 75 ◦C, 100 ◦C, 125 ◦C, 150 ◦C, 175 ◦C, and 200 ◦C. From Figures 7 and 8, pure TTO exhibited bad friction-reduction performance in the tested temperature range and its performance was less affected by an increasing temperature. When compared to the samples of TTO with LCs added, the friction coefficient curves show an increasing trend in general with large fluctuations. In contrast, the oil containing LCs had a lower average friction coefficient in a specific temperature range, and the friction coefficient curves showed less fluctuation and an overall decreasing trend, especially in the range of 75–150 ◦C. It is noteworthy that there are differences in the mesogenic-phase temperature ranges in which the three LCs have friction-reducing effects. As can be seen from Figure 8, LC-A can reduce the average friction coefficient of TTO at 25–175 ◦C; LC-D has the ability to reduce friction from 75–200 ◦C, and LC-Z can achieve a certain friction reduction effect across the whole temperature range of the test. Additionally, LC-A and LC-D have the best friction-reduction effects at 100 ◦C, with reductions of 19.25% and 21.57% when compared to the base oil, respectively. LC-Z showed the best performance in anti-friction at 75 ◦C, with a reduction of 19.59%. When combined with the DSC, the anti-friction properties of the LCs were very closely connected with the mesogenic-phase temperature ranges, and the optimum temperatures for their tribological properties were close to or within this range. Among them, LC-D has a mesophase temperature range of 100.5 ◦C to 100.9 ◦C. The narrower intermediate-phase temperature range may limit the range in which it can exert its frictionreducing effect. Compared with LC-D, LC-A and LC-Z have a wider intermediate-phase temperature range and can be used as additives to reduce the average friction coefficient over a wider temperature range. According to the analysis, it can be speculated that the

phase-change temperature range of LCs is an influential factor on their friction-reduction performance, and the temperature range which exerts the friction-reduction effect will be wider than the phase-change temperature range in general. In addition, ZDDP is the most widely used and effective lubricant additive. As can be seen from Figures 7 and 8, the trend of tribological performance with temperature remains insignificant when ZDDP is added to the base oil. The LC additives displayed superior friction-reduction performances compared to ZDDP, and the effect of friction-reduction was influenced by temperature. Even at 175 ◦C, the LCs' average friction coefficients remained superior when compared with ZDDP, which demonstrates that LCs can be used in lubricant applications. It is worth noting that each LC has its corresponding optimal temperature interval, within which the LC additives can improve the friction-reduction performance of the oil and yet can be applied to several operating conditions under specific temperatures, thus enabling the controllability of the oil temperature.

#### *3.4. Anti-Wear Analysis*

The width of the wear scar was characterized by a Contour GT-K 3D profiler to illustrate the anti-wear performance of LCs as lubricant additives. The average of the three measurements was calculated to illustrate the width of the wear scars. The results are shown in Figure 9. All LC additives exhibited superior anti-wear performance at different temperatures when compared with TTO. Even at 50–175 ◦C, the anti-wear performance of the three LC additives was significantly better than that of the commercially available additive ZDDP. It is noteworthy that, compared with LC-D, the addition of LC-Z resulted in a smaller width of wear marks on the steel plate over a wide temperature range of 25 to 75 ◦C. Among them, the width of wear marks at 75 ◦C was reduced by 31.82%. The addition of LC-A can also significantly improved the anti-wear performance of the oil, with a 25.78% reduction in the width of wear marks at 75 ◦C, which may be caused by the longer chain length of LC-A and LC-Z, resulting in wide-mesophase temperature ranges. It can be seen from Figure 9 that LC-D has good anti-friction properties throughout the temperature range, especially at 125 ◦C, with a 30.07% reduction in width of the wear scars, which is probably due to the presence of trifluoromethyl in its structure. In addition, the three LC additives can dramatically enhance the anti-wear performance of TTO in a wide temperature range of 25 to 175 ◦C, which is consistent with the following conclusion: the temperature interval in which the LC additives have tribological performance effects is wider than their mesophase temperature ranges. Therefore, the chemical structure and mesophase temperature ranges of LC additives are critical factors affecting their anti-wear performance.

#### *3.5. Surface Analysis*

#### 3.5.1. SEM Analysis

The morphology of the wear surface was further observed with SEM. Figure 10 shows the SEM images and the width of wear scars at 100 ◦C. As can be seen from Figure 10, the wear scars were deeper with pure TTO lubrication and an obvious furrow appeared, indicating that no stable lubrication film was formed on the wear surface. The addition of LC additives can reduce the degree of wear on the wear surface and form a stable lubricating film, so as to improve the anti-wear properties of the base oil. The width of the wear marks of pure TTO is 381.2 μm, and the width of the wear marks of the three LC additives are 305.4 μm, 297.7 μm, and 308.3 μm, respectively. Compared to pure TTO, the anti-wear effects of LCs are 19.89%, 21.88%, and 19.11%, respectively. The anti-wear effect of LC-A is better than LC-Z, which was in line with the mesophase temperature ranges. The anti-wear effect of LC-D was significantly better than the other two LC additives, which was probably due to the presence of trifluoromethyl in its structure, which forms a stable lubricating film and leads to the improvement of the anti-wear properties of the oil. It can be seen from Figure 10c,c', that there was a slight corrosive wear on the wear surface of LC-D, which may be due to the chemical reaction between the free fluorine atoms and iron elements on the wear surface.

**Figure 7.** The friction coefficients of TTO and TTO containing 2 wt% ZDDP, LC-A, LC-D, and LC-Z at different temperature points: (**a**) at 25 ◦C; (**b**) at 50 ◦C; (**c**) at 75 ◦C; (**d**) at 100 ◦C; (**e**) at 125 ◦C; (**f**) at 150 ◦C; (**g**) at 175 ◦C; and (**h**) at 200 ◦C.

**Figure 8.** Average friction coefficient of pure TTO and TTO with 2 wt% ZDDP, LC-A, LC-D, and LC-Z under different temperature points.

**Figure 9.** Width of wear scars of pure TTO and TTO with 2 wt% ZDDP, LC-A, LC-D, and LC-Z under different temperature points.

#### 3.5.2. Raman Spectroscopy Analysis

The lubrication mechanism of the worn surface was analyzed using Raman spectroscopy. After the test, we cleaned the steel plates with petroleum ether. The wear marks on the surfaces of the steel plates were characterized with Raman spectroscopy to illustrate the mechanism of action of LCs as lubricant additives. Figure 11 shows the Raman spectra after the UMT-TriboLab friction and wear tester, excited by the 532 nm laser. We can see that the strong peak at 663 cm−<sup>1</sup> is Fe3O4, the peaks at 1357 cm−<sup>1</sup> and 1573 cm−<sup>1</sup> belong to the D and G peaks, respectively, and the strong peak at 2890 cm−<sup>1</sup> is associated is with the 2D peak. From Figure 11a, the surface lubricated by base oil mainly contains Fe3O4, and the signals of the D peak and G peak are weak, which indicates that the carbon content in the base oil is relatively low. The surface of the friction marks after adding LCs is dominated

by Fe3O4, D peak and G peak, and has an obvious 2D peak. It can be confirmed that a carbon film and iron oxide film were formed on the friction surface, which played a role in reducing friction and increasing anti-wear properties. Significantly, the characteristic group with LC additives can be observed on the wear surface lubricated with 2 wt% LC-A, which indicates that the LCs has a strong adsorption ability on the metal surface and can generate a lubricant film with the chemical reaction on it. Figure 12 shows the schematic representation of the lubrication mechanism. It can be hypothesized that LC additives can form a lubricant film on the surface, preventing the direct contact of the rough surfaces and thus achieving the effect of friction-reduction and anti-wear.

**Figure 10.** SEM images of the wear surface after UMT tests at 100 ◦C: (**a**,**a'**) are TTO; (**b**,**b'**) are 2 wt% LC-A added to TTO; (**c**,**c'**) are 2 wt% LC-D added to TTO; and (**d**,**d'**) are 2 wt% LC-Z added to TTO.

**Figure 11.** Raman spectra at wear marks after UMT: (**a**) TTO; (**b**)TTO + 2 wt% LC-A; (**c**) TTO + 2 wt% LC-D; and (**d**)TTO + 2 wt% LC-Z.

**Figure 12.** The schematic representation of the lubrication mechanism.

#### **4. Conclusions**


3. To analyze the lubrication mechanism of LCs, it is found that the addition of LCs additive can reduce the width of wear marks and the surface furrow, which can form a lubricant film on the wear surface and prevent the direct contact of frictional pairs on the sliding surface, leading to the improvement of tribological properties of the base oil.

**Author Contributions:** Conceptualization, W.H. and J.L.; methodology, W.H. and S.F.; validation, H.W.; formal analysis, H.W. and Y.J.; investigation, H.W.; data curation, H.W.; writing—original draft preparation, H.W. and W.H.; writing—review and editing, W.H. and J.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Youth Innovation Promotion Association (2019288), the Shanghai Pudong New Area Science and Technology Development Fund (PKJ2019-C01), and the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA21021202).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data is contained within the article.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

**Chao Zhang 1,2, Mingfang Wu 1, Juan Pu 1,\*, Jiawei Rao 1, Weimin Long <sup>3</sup> and Yuanxun Shen <sup>3</sup>**


**Abstract:** The cold metal transfer (CMT) welding-brazing process was chosen to join Al alloy and Nicoated steel using AlSi12 as the filler wire. The macrostructure and microstructure of the joints were tested by using an optical microscope (OM), scanning electron microscope (SEM), energy dispersive spectrometry (EDS), and X-ray diffraction (XRD). The tensile properties and corrosion properties of the joints were also tested. The results showed that Ni coating could improve the wettability and spreadability of molten AlSi12 filler metal on the steel surface, resulting in a good appearance for the Al alloy/steel joint. Ni coating could hinder the chemical metallurgical reaction between Al atom and Fe atoms to inhibit the formation of brittle Fe-Al intermetallic compounds (IMCs) and reduce the thickness of the IMCs layer. Meanwhile, the Ni atom reacted with the Fe and Al atoms to form Al3Ni2, (Fe, Ni) Al3 and (Fe, Ni)2Al3, which improved the tensile strength of the joints. All joints with Ni coating cracked near the Al alloy. When the Ni-coating thickness was 5 μm, the tensile strength of the joint reached a maximum of 202.5 MPa. The addition of Ni could also improve the corrosion resistance of the joints. Significantly, when the Ni-coating thickness was 10 μm, most of the Ni coating was still solid, and the interface reaction layer was mainly composed of α-Ni solid solution and some (Fe, Ni)2Al3.

**Keywords:** interface reaction mechanism; Ni coating; microstructure; property; CMT

#### **1. Introduction**

The "Made in China 2025" development strategy proposes that the manufacturing industry changes from a traditional to an intelligent and green industry. Realizing weight reduction is considered a vital way to achieve this change [1,2]. Al alloys, which have the advantages of light weight, high specific strength, corrosion resistance, and good comprehensive performance, are widely used in aerospace, railway transportation, automobiles, and refrigeration [3–5]. They partially replace steel to form Al alloy/steel structures for achieving lightweight structures. Al alloy/steel parts have the characteristics of the lightweight of Al alloy together with the advantages of high strength and low cost of steel. They are widely used in the covering parts and chassis parts of automobiles. Undoubtedly, the Al alloy/steel hybrid body has become the trend of future development in the industry. Therefore, the connection of Al alloy/steel parts has high requirements.

For the joining of Al alloy and steel, some of the difficulties are as followings [6]: Firstly, the thermal expansion coefficients are different, which results in the severe deformation of the joints and thus produce considerable residual stress in the welding process. Secondly, due to their poor metallurgical compatibility, a series of brittle IMCs (i.e., FeAl3, Fe2Al5, FeAl2, FeAl) form, which deteriorates the mechanical properties of the joints. Thirdly, the potential of Al alloy is different from that of steel, and the localized electrochemical

**Citation:** Zhang, C.; Wu, M.; Pu, J.; Rao, J.; Long, W.; Shen, Y. Effect of Ni Coating on Microstructure and Property of Al Alloy/Steel CMT Welding-Brazing Joints. *Coatings* **2023**, *13*, 418. https://doi.org/ 10.3390/coatings13020418

Academic Editors: Ashish Kumar Srivastava and Amit Rai Dixit

Received: 6 January 2023 Revised: 2 February 2023 Accepted: 6 February 2023 Published: 12 February 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

corrosion near the steel occurred in a humid atmosphere. At the same time, porosity and defects in the joints would accelerate the process of electrochemical corrosion [7].

In order to prepare composite materials with excellent performance, many studies have been carried out. Farhad Ostovan et al. [8–10] believed that Gas Tungsten Arc Welding (GTAW) had the unique advantages due to its easy control of heat input. Qin et al. [11] compared and analyzed the advantages of several welding methods to join Al alloy and steel, such as mechanical connection, solid phase connection, liquid-solid diffusion, brazing, GTAW brazing and so on. They found that laser welding brazing, electron beam welding brazing and arc+laser hybrid welding could achieve high-quality Al alloy/steel joints by controlling the welding heat input. Maryam Roudbari et al. [12] reinforced the Al plates with steel wires by using the method of explosive welding. They determined the optimal parameters through the numerical simulation. The simulation results agreed very well with the experimental data. Then they studied the mechanical properties of aluminum base composite materials [13]. The results showed that the tensile strength of the reinforced composite was 8% higher than that of the unreinforced material. However, the explosive welding was limited by the weather and gave rise to the air pollution.

Our research team studied the joining of Al alloy and steel by using welding-brazing technology. Shao et al. [14] used different Al-Si filler metals to join Al alloy and steel by CMT welding brazing, respectively. They reported that Si element could inhibit the formation of brittle IMCs at the interface reaction layer. Shi et al. [15] showed that TIG welding brazing could control welding heat input precisely. The addition of Zn element could improve the wettability of AlSi5 filler metal and also enhance the corrosion resistance of the joint. Chao et al. [16] studied the effect of Cu coating on the microstructure and properties of the joints by CMT welding brazing. They found that with the increase in Cu-coating thickness, the thickness of the IMCs layer decreased and the corrosion resistance of the Al alloy/steel joint increased. When the Cu-coating thickness was 10 μm, the tensile strength and corrosion resistance of the joints were up to the optimum. Other researchers [17] have reported that Ni coating on the steel surface could change the IMCs composition of joints by laser welding brazing. However, they did not discuss the formation mechanism of IMCs compounds in detail, nor did they study the corrosion resistance of the Al/steel joints.

Therefore, in this paper, different thicknesses of Ni layer were designed to be coated on steel surfaces. The CMT welding-brazing process of Al alloy/steel was conducted using AlSi12 as the filler wire. The effects of different Ni-coating on the macrostructure, microstructure, tensile strength, and corrosion resistance of the joints were studied. We focused on exploring the action mechanism of Ni coating on the interface reaction layer of joints.

#### **2. Experimental Materials and Methods**

#### *2.1. Experimental Materials*

The purchased 6082 Al alloy and Q235B steel with dimensions of 150 mm × 100 mm × 2 mm were used as the base metals. AlSi12 flux-cored wire with a diameter of 1.6 mm was used as the filler material, and it was composed of AlSi12 wire and NOCLOCK flux. The chemical composition and tensile properties of the above materials are listed in Table 1.

**Table 1.** Chemical compositions and tensile properties of base metal and flux-cored wire.



**Table 1.** *Cont.*

#### *2.2. Coating Process of Ni on the Steel Surface*

Ni layers that ranged from 0 μm and 10 μm were coated on the surface of Q235B steel by the electroplating process. At least five samples were prepared and tested for each Ni-coating thickness. Before electroplating, all steel plates were polished with sandpaper to remove burrs. Subsequently, the steel plates were immersed into an alkaline solution at 80 ◦C for 20 min to remove oil stains. The composition of the alkaline solution was 30 g/L NaOH, 35 g/L NaCO3, 20 g/L NaPO4, and 7.5 g/L OP-10. Then, they were immersed in the acid solution for 2 minutes, which was composed of 5% H2SO4, 15% HCl, and 80% distilled water. After chemical cleaning, the steel plates were washed in the acetone solution with an ultrasonic cleaner to be electroplated. The electroplating solution containing 200 g/L NiSO4, 30 g/L NiCl2, 30 g/L H3BO3, 0.1 g/L C12H25NaSO3 and 0.5 g/L saccharin was prepared and stirred for 30 min. Finally, the treated steel plates were placed into the electroplating solution. During the electroplating process, the nickel plate was used as the anode, while the steel plate was used as the cathode. The process parameters of electroplating were a pH of 4, a temperature of 50 ◦C, and a current density of 1.5 A/min. The schematic diagram of the electroplating process is shown in Figure 1.

**Figure 1.** Schematic diagram of the electroplating process for Ni coating on the steel surface.

#### *2.3. CMT Welding-Brazing Process to Join Al Alloy/Steel*

A CMT5000i welding machine (Fronius company) was used to join Al alloy and Nicoated steel with the AlSi12 filler wire. Before CMT welding brazing, we cleaned the base metals with the acetone. The Al alloy plates had no groove while the steel plates were equipped with a single V-shaped groove of 30◦. The root gap was 0.1 mm for assembly, and we used the copper backing to solidify the welding joint. The inclination angle of the torch was set to 70◦ and the distance from the tip of the welding wire to the workpiece was 2 mm. The welding process was protected by Argon gas with 99.99% purity. The schematic diagram of the CMT welding-brazing process is illustrated in Figure 2. Its detailed parameters are listed in Table 2.

**Figure 2.** Schematic diagram of CMT welding-brazing process.



#### *2.4. Analysis of Microstructure and Tensile Properties of Al Alloy/Steel Joints*

After CMT welding brazing, samples for the investigation of microstructure and tensile properties were prepared. All the microstructure analysis samples were mechanically ground with sandpaper and polished with diamond suspension, and then etched by Keller's reagent for 10–15 s. The macrostructure was observed by a ZEISS optical microscope (OM, Oberkochen, Germany). The microstructure was analyzed by a JSM-6480 scanning electron microscope (SEM, JEOL, Tokyo, Japan), and the element distribution was identified by energy dispersive spectroscopy (EDS). We stripped the Al alloy/steel joint along the side of the steel, and the phase composition of the interface reaction layer and fusion zone were confirmed by an XRD-6000 X-ray diffractometer instrument (XRD, Shimadzu, Kyoto, Japan) with a scanning angle (2θ) ranging from 10◦ to 90◦ and a scanning speed of 3 ◦/min.

According to GB/T 2651-2008 standard, the tensile properties of the joints were tested on an ETM605D mechanical tester at a constant rate of 1 mm/min. The fracture surfaces were observed by SEM.

#### *2.5. Electrochemical Corrosion Measurements of Al Alloy/Steel Joints*

An Electrochemical Workstation (No. EGM283) was used to carry out the electrochemical experiments in a 3.5 wt. % NaCl solution. We used a saturated calomel electrode (SCE) as a reference electrode, a platinum plate as an auxiliary electrode, and the Al alloy/steel joint as the working electrode. The specimens including Al alloy, steel, and the Al alloy/steel joint with dimensions of 10 mm × 10 mm × 2 mm were prepared. The potentiodynamic polarization curves were recorded with a scanning rate of 2 mV/s and a scanning range starting at −1 V up to 1.5 V.

#### **3. Results and Discussions**

#### *3.1. Effect of Ni Coating on Macrostructure of Al Alloy/Steel Joints*

Figure 3 shows the weld seam appearance and cross-section appearance of Al alloy/steel joints with different thicknesses of Ni coating. For Al alloy/steel joints without a Ni coating, the top weld seam shape (see Figure 3a) and the bottom weld seam shape (see Figure 3b) were not continuous and not smooth. When the Ni-coating thickness was 5 μm and 10 μm, the shape of the top weld seam (see Figure 3d,g, respectively) and the bottom weld seam (see Figure 3e,h, respectively) appeared better and smoother.

**Figure 3.** Macrostructure of joints under different thicknesses of Ni coatings: (**a**–**c**) without Ni coating; (**d**–**f**) with 5 μm Ni coating; (**g**–**i**) with 10 μm Ni coating.

Seen from the cross-section appearance of the joints, Ni coating has a significant influence on the wettability and spreading of AlSi12 filler metal on the steel surface. For Figure 3c,f,i, the wetting angle of molten AlSi12 filler metal on the steel surface decreased from 37.3◦ to 28.3◦ and then 21.9◦. Generally, the smaller the wetting angle, the better the wettability. Thus, the results implied that the addition of Ni could improve the wettability and spreadability of molten Al-Si metal on the steel surface. Yang et al. [18] had reported that Ni could reduce the melting point of Ag-based filler metal and improve its wettability. Sun et al. [19] found Ni coating on the surface of Al2O3 particles could improve the wettability of Al2O3/Al-10Si composites. These results are consistent with our research results.

#### *3.2. Influence of Ni Coating on Microstructure of Al Alloy/Steel Joints*

In the process of CMT welding brazing, part of the Al alloy base metal together with AlSi12 filler metal preferred to melt under the action of arc heat due to their low melting point (below 660°C), and the fusion zone formed at the side of Al alloy base metal. Then, the molten Al alloy metal wet and spread on the steel surface, which was still solid because of its high melting point (above 1500°C), so the brazing interface reaction zone formed at the side of the steel base metal [20,21]. The fusion zone and brazing interface reaction zone were analyzed by SEM with EDS point and line scanning. The results are shown in Figure 4 and Table 3.

**Figure 4.** The microstructure and corresponding EDS line scanning results of joints: (**a**–**c**) without Ni coating; (**d**–**f**) with 5 μm Ni coating; (**g**–**i**) with 10 μm Ni coating.


**Table 3.** The EDS points scanning results and possible phase of corresponding characteristic points in Figure 4.

Figure 4a exhibits SEM images of the Al alloy/steel joint without Ni coating. The fusion zone near the Al alloy is composed of needle-like compounds and granular-shaped compounds. Figure 4b shows the enlarged view of zone B in Figure 4a. The brazing interface reaction layer at the steel side is composed of dark gray compounds. Points 1-4 were analyzed by EDS point scanning and the results were listed in Table 2. The interface reaction layer was analyzed by EDS line scanning as shown in Figure 4c. From Figure 4c, the thickness of the IMCs layer was about 3.83 μm. The content of Al element decreased, whereas the content of Fe element increased along the path from the fusion zone to the brazing interface reaction zone. According to the EDS points analysis results and Al-Fe-Si ternary alloy phase diagram, in the fusion zone, the thin needle-like phase (point 1) was a τ5-Al7.2Fe1.8Si and the dark gray phase (point 2) was an α-Al solid solution. Correspondingly, the interface reaction layer consisted of FeAl3 (point 3) formed near the fusion zone and θ-Fe (Si, Al)3 (point 4) formed at the steel side.

Figure 4d shows the microstructure of the joint with 5 μm Ni coating. Figure 4e shows the enlarged view of zone E in Figure 4d. As can be observed, a large number of thin needle-like phases are distributed uniformly in the fusion zone, and the thickness of the IMCs layer decreased to 3.52 μm. Figure 4f shows that the Ni content increases at the interface reaction layer, and Ni element diffuses into the fusion zone. Therefore, in the fusion zone, Al atoms react with the Ni atoms to form Al3Ni2 (point 5) and the thin needle-like phase changed from τ5-Al7.2Fe1.8Si into τ5-Al7.2(Fe, Ni)1.8Si (point 6). The α-Al solid solution (point 7) also contained Ni element. Meanwhile, Ni atoms replace some Fe atoms and then react with Al atoms to form (Fe, Ni) Al3 (point 8) at the interface reaction layer. The gray compound (point 9) was still θ-Fe (Si, Al)3.

When the Ni-coating thickness was 10 μm, the fusion zone contained some small blockshaped tissue, some white dendritical tissue, and a dark-grey matrix. Significantly, the thickness of the interface reaction layer was 12.12 μm, comsisting of about 10 μm Ni coating and 2.12 μm IMCs layer, as shown in Figure 4g,h. Based on the results listed in Table 2, the small block-shaped tissue (point 10) was determined to be Al3Ni2, the white dendritical tissue (point 11) was τ5-Al7.2(Fe, Ni)1.8Si, and the dark-grey matrix (point 12) was α-Al solid solution. The interface reaction layer was composed of (Fe, Ni)2Al3 (point 13) near the fusion zone and α-Ni solid solution (point 14). Figure 4i exhibits the elements distribution of A, Si, Fe, and Ni between the Al alloy and steel. Ni element was mainly concentrated in the interface layer. Meanwhile, a small amount of Al and Si elements in the fusion zone passed through the Ni coating to the steel side, and a few Fe elements arrived at the side of the fusion zone. Most of the Ni coating was still solid, so it hindered the diffusion between

Al element and Fe element. Hence, the interface reaction layer was mainly composed of α-Ni solid solution, and some (Fe, Ni)2Al3 compound formed.

To confirm the phase composition, the Al alloy/steel joint was stripped along the side of the steel, and then the XRD analysis of the fusion zone and the interface reaction zone were performed. The results are shown in Figure 5. For the joint without Ni coating, compounds of α-Al solid solution and τ5-Al7.2Fe1.8Si appeared. When the thickness of the Ni coating was 10 μm, the Al3Ni2 compound and the α-Ni solid solution formed in the joints. Some other phases were hard to be determined.

**Figure 5.** XRD results of Al alloy/steel joints.

Based on the above investigations, using AlSi12 filler metal to join Al alloy and steel by CMT technology, the joint was mainly composed of α-Al solid solution and τ5-Al7.2Fe1.8Si in the fusion zone, and FeAl3 together with θ-Fe (Si, Al)3 in the interface reaction zone. After coating 5 μm Ni on the steel surface, Ni participated in the metallurgical reaction, and reacted with Al, Fe, and Si elements to form some new phases of τ5-Al7.2(Fe, Ni)1.8Si and Al3Ni2 in the fusion zone. Meanwhile, (Fe, Ni) Al3 replaced FeAl3 in the interface reaction zone. It can be predicted that the change in microstructure composition will improve the mechanical properties of the joints. When the Ni-coating thickness increased to 10 μm, most of the Ni coating did not melt, so the diffusion between Al element and Fe element was blocked. Interestingly, the FeAl3 phase disappeared, and a new phase of (Fe, Ni)2Al3 appeared in the IMCs layer near the Al alloy. The interface reaction zone was mainly composed of α-Ni solid solution. Hence, the mechanical properties of the joint will not be poor.

#### *3.3. Action Mechanism of Ni Coating in Interface Reaction Layer*

Generally, ΔG<sup>0</sup> is a criterion to predict the spontaneity of a chemical reaction. When the value of ΔG0 is negative, the chemical reaction occurs easily. Moreover, compounds with the lowest ΔG0 are most likely to be formed.

Based on thermodynamic data of Al-Fe compounds [22,23], the functions of several compounds are calculated as followings:

$$
\Delta \mathbf{G}\_{\text{FeAl}^3}^0 = -142770 + 50.58 \mathbf{T} \tag{1}
$$

$$
\Delta \mathbf{G}^{0}\_{\text{Al7.2}^{\text{Fe}\_{1.8}\text{Si}}} = -295355 + 94.59 \text{T} \tag{2}
$$

$$
\Delta \mathbf{G}\_{\mathrm{Fe(Al/Si)}}^{0} = -142770.0 + 50.8 \mathbf{T} \tag{3}
$$

where T represents temperature. During CMT welding brazing, the temperature ranges from 900 K to 1300 K, and the result is ΔG0 Al7.2Fe1.8Si <sup>&</sup>lt; <sup>Δ</sup>G0 FeAl3 , which indicates that <sup>τ</sup>5- Al7.2Fe1.8Si preferentially forms when Al, Fe and Si atoms interact at the interface of the Al alloy and steel.

Figure 6 shows the schematic diagram of the interface reaction layer growth mechanism of an Al alloy/Steel joint with Ni coating by CMT technology. At the initial stage, as shown in Figure 6a, under the action of a pulsed arc, AlSi12 filler metal and a small amount of Al alloy base metal start to melt, while Ni coating remains semi-solid. At this time, Fe and Al atoms cannot form compounds because the Ni coating acts as a barrier layer. However, with the melting of the Ni coating under the action of arc heat, Fe atoms in the steel passed through Ni coating to react with Al and Si atoms, which come from the AlSi12 filler metal and Al alloy base metal, and then τ5-Al7.2Fe1.8Si formed preferentially. Subsequently, FeAl3 also formed as shown in Figure 6b.

**Figure 6.** Interface reaction layer growth mechanism of Al alloy/steel joint with Ni coating: (**a**) stage I; (**b**) stage II; (**c**) stage III; (**d**) stage IV.

Generally, the possible compounds between Al and Ni can be calculated by the following formulas:

$$
\Delta \mathbf{G}^{0}\_{\text{Al}^{3\text{Ni}\_2}} = -71545.2 + 13.7 \text{T} \tag{4}
$$

$$
\Delta \mathbf{G}^{0}\_{\text{Al}^{3\text{Ni}}} = -48483.8 + 12.7 \text{T} \tag{5}
$$

Based on formulas (4) and (5), Al3Ni2 is easier to generate. When the Ni coating starts to melt, Al3Ni2 had already appeared in the weld seam zone according to the above microstructure analysis. Fe element is a relative of Ni, so Fe atoms can occupy the positions of Ni atoms to form (Fe, Ni)2Al3, (Fe, Ni) Al3 and τ5-Al7.2(Fe, Ni)1.8Si as a result of their similar electronic structures [24]. Therefore, (Fe, Ni)2Al3, (Fe, Ni) Al3, τ5-Al7.2Fe1.8Si, and τ5-Al7.2(Fe, Ni)1.8Si formed in the interface reaction layer, as shown in Figure 6c.

With the wetting and spreading of molten liquid filler metal on the steel surface, Fe atoms reacts with τ5-Al7.2Fe1.8Si as follows:

$$\text{Fe} + \pi\_5\text{-Al}\_{7.2}\text{Fe}\_{1.8}\text{Si} = \text{\(\text{\(Al,Si\)}\)}\text{}\tag{6}$$

Finally, the interface reaction layer consists of τ5-Al7.2Fe1.8Si and τ5-Al7.2(Fe, Ni)1.8Si in the weld seam zone together with θ-Fe (Al, Si)3 and (Fe, Ni)2Al3 at the steel side, as shown in Figure 6d.

#### *3.4. Effect of Ni Coating on Tensile Properties of Al Alloy/Steel Joints*

Figure 7 shows the fracture location and tensile strength of joints. For the joint without Ni, it cracked at the brazing interface reaction layer, as shown in Figure 7a. When the Ni-coating thickness was 5 μm and 10 μm, the fracture mode was located in the fusion zone at the Al alloy side, as shown in Figure 7b,c, respectively. With the increase in Ni-coating thickness from 0 μm to 10 μm, the tensile strength first increased and then decreased, and the tensile fracture displacement gradually increased. When the Ni-coating thickness was 5 μm, the tensile strength value reached the maximum of 202.5 Mpa, as shown in Figure 7d. Figure 7e indicates that the addition of Ni coating improves the plastic toughness of the joints.

**Figure 7.** Tensile properties and fracture section morphology of joints: (**a**) fracture location without Ni coating; (**b**) fracture location with 5 μm Ni coating; (**c**) fracture location with 10 μm Ni coating; (**d**) tensile strength and fracture displacement; (**e**) stress-strain curves.

For the joint without Ni coating, FeAl3 formed in the interface reaction layer. The brittle FeAl3 tended to become the source of crack propagation, which resulted in the weakness of the joint during the tensile process. Therefore, the joint cracked at the interface reaction layer. For the joint with Ni coating, the Ni coating melted to occupy the position of the Fe atoms and then generated (Fe, Ni) Al3, (Fe, Ni)2Al3 and τ5-Al7.2(Fe, Ni)1.8Si under the heat action of the arc. The tensile strength of the interface reaction layer was improved as a result of adding Ni into Fe-Al-Si compounds [25,26]. Moreover, when the thickness of the Ni coating was 10 μm, it was too heavy to melt and hindered the diffusion of Fe atoms and Al atoms to avoid the generation of the brittle Fe-Al IMCs and prompt the formation of Ni solid solution. Thus, the tensile strength of the interface reaction layer was high, and the joint cracked at the Al alloy side.

The fracture morphologies of the joints were observed by SEM and are shown in Figure 8. In Figure 8a, tearing ridges and cleavage steps were observed in the fracture surface, which exhibited the brittle fracture mode. Based on the EDS result for point 15, the brittle IMC in the fracture surface was FeAl3, which easily became a source of crack propagation and promoted the crack growth, so the corresponding tensile strength of the joint was the lowest. In Figure 8b,c, a large number of dimples were found in the fracture surface, and they showed a typical ductile fracture mode. In addition, the dimples in the fracture with 5 μm Ni coating were more uniform and finer than that with 10 μm Ni coating. The tensile strength of the former joint reached up to the maximum.

**Figure 8.** Tensile fracture morphology of Al alloy/steel joints under different thicknesses of Ni coating: (**a**) without Ni coating; (**b**) with 5 μm Ni coating; (**c**) with 10 μm Ni coating.

#### *3.5. Effect of Ni Coating on Corrosion Resistance of Al Alloy/Steel Joints*

The potentiodynamic polarization curves of joints were measured to investigate the corrosion resistance and the results are shown in Figure 9. Figure 9a shows the variation curve of open-circuit potential with time. When the open circuit potential stabilized, the polarization curves were recorded. As can be seen in Figure 9b, all the curves of joints showed an obvious passivation area, and this indicated the passive film on the surface of joints formed spontaneously to delay their corrosion, but their potential range of the passivation area was different. In order to obtain the related data about the corrosion potential (Ecorr) and the self-corrosion current density (Icorr), CView software was used to fit the curves, and the fitting results are listed in Table 4.

**Figure 9.** Polarization curves of Al alloy/steel joints with different thicknesses of Ni coating: (**a**) open circuit potential–time curves; (**b**) polarization curves.


**Table 4.** Ecorr and Icorr fitted for the curves in Figure 9.

As seen from Figure 9 and Table 4, the Ecorr of the Al alloy plate was the highest, that of the steel plate was the lowest, and that of the joint was between them. Nevertheless, their corrosion current was in reverse order. Meanwhile, compared with the joints without Ni coating and with Ni coating, the Ecorr of the joint with Ni coating was higher while the Icorr of the joint with Ni coating was smaller. Some studies [27] have shown that Ecorr reflects the corrosion tendency of materials, and Icorr reflects the corrosion rate of materials. The greater Ecorr and the smaller Icorr, the better the corrosion resistance of materials.

For the Al alloy/steel joints, galvanic corrosion occurred between the weld seam zone and the steel. The steel with a low corrosion potential can protect the weld seam from corrosion. It was also found that the IMCs layer can accelerate the corrosion of the joint. After adding Ni into the joint, the corrosion resistance of the joint was improved. One reason was that Ni coating hindered the diffusion reaction between Al and Fe, which resulted in a reduction in IMCs thickness. The other reason was that Ni atoms reacted with Al to form Ni2Al3 and replaced Fe atoms to form (Fe, Ni)2Al3, and these compounds improve the corrosion resistance of the joints [28–30].

#### **4. Conclusions**

In this paper, a CMT welding-brazing process was conducted to join Al alloy and the steel under different Ni coatings by using AlSi12 as a filler wire. The effects of Ni coating on macrostructure, microstructure, tensile properties, and corrosion resistance of the joints were investigated. The main conclusions can be summarized as follows:


10 μm thicknesses, the weld seam zone of joints consisted of α-Al solid solution, τ5-Al7.2(Fe,Ni)1.8Si, and newly formed Al3Ni2.


**Author Contributions:** Conceptualization, C.Z., J.P. and M.W.; methodology, C.Z., Y.S. and J.R.; validation, C.Z. and M.W.; writing—original draft preparation, C.Z. and J.P.; writing—review and editing, W.L. and Y.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work is sponsored by the National Natural Science Foundation of China (No. 51675249), Natural Science Research General Project of Jiangsu Province (No. 19KJB46001), the Excellent Scientific and Technological Innovation Team of Universities in Jiangsu Province and Engineering Research Center Program of Development & Reform Commission of Jiangsu Province (Grant No. [2021] 1368) and Changzhou Science and Technology Plan Project (No. CJ20190028).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data used to support the findings of this study are available from the corresponding authors upon request.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
