Effect of Normal Load on the Current-Carrying Friction Performance of Copper–10% Graphite Composites

Abstract

A copper-10 wt.% graphite composite was paired with QCr0.5 to investigate the effects of normal load on current-carrying friction behavior. Arc discharges were monitored using a high-speed camera and photodiodes. The results indicate that, under the given experimental conditions, normal load predominantly influences the tribological performance of the material. As the c normal load increases, the wear rate decreases rapidly at first and then increases gradually. The optimal normal load was found to be 70 N, at which the wear rate reached a minimum of 0.46 mg/m. Material degradation was found to consist of mechanical damage—mainly plowing and plastic deformation—as well as arc-induced erosion characterized by melting and spattering. With increasing normal load, arc erosion decreased progressively, and the overall damage was minimized at 70 N. Arc erosion contributed to surface wear non-uniformity. Moreover, particular attention should be paid to high-current, long-duration arcs, which can pose serious localized threats to material integrity.

1. Introduction

Normal load is a critical service condition influencing the current-carrying friction performance of copper–graphite composites [1,2,3]. At the microscale, normal load affects both the number of contact spots and the degree of stress concentration at these sites; it also governs adhesion and tearing processes, thereby influencing the initiation, evolution, and extinction of associated electrical arcs. At the macroscale, normal load determines the operational stability of the friction pair, subsequently affecting both mechanical wear and arc erosion during conductive sliding [4].

There has been considerable research on the impact of normal load on current-carrying friction. Research by Yasar et al. [5] showed that at normal loads below 0.03 MPa and above 0.12 MPa, severe material wear occurs, with arc erosion dominating below 0.03 MPa and mechanical wear predominating above 0.12 MPa. Xinlin Xie et al. [6] found that at low normal loads, instability occurs, leading to severe arc erosion, while at higher pressures, the actual contact area increases, reducing normal load drop and electrical noise. Zhenying Huang et al. [7] demonstrated that the stability of the friction process decreases with a reduction in normal load; below a critical value, wear increases, contact stability deteriorates, and arc ablation area expands. Liu Shuhua et al. [8] pointed out that the wear rate follows a “U-shaped” trend as normal load increases. Hui Yang et al. [9] showed that an increase in normal load improves the contact quality between friction pairs, while both arc ignition rate and arc energy decrease accordingly. However, the materials used in these studies were not optimized, and the investigation of the effects of associated electrical arcs and their damage primarily relied on wear surface observations.

This study aims to use optimized copper–10 wt.% graphite material [10] paired with QCr0.5, and employ high-speed cameras and photodiodes to observe arcs, in order to investigate the effect of normal load on the current-carrying friction process.

2. Experiment

The experiment used custom-made copper–graphite composite pins paired with commercially available QCr0.5 chromium bronze disks (with Cr content ranging from 0.4% to 1.1%, total impurities ≤ 0.5%, and the remainder being Cu). The copper–graphite composite was prepared using powder metallurgy, with electrolytic pure copper powder (purity greater than 99 wt.%, particle size of 200 mesh) and copper-coated graphite powder (graphite content ranging from 49.7 wt.% to 50.3 wt.% and the remainder being copper, particle size of 200 mesh).

The powder metallurgy process includes mixing, pre-pressing, sintering, and re-pressing. The mixing process involved adding equal mass copper balls and mixing for 18 h in a V-shaped mixer (Zhuangcheng Equipment Technology Co, Ltd., Wuxi, China). Pre-pressing was carried out using a custom-made mold on an MY-100 universal hydraulic press (Abrasives and Grinding Tools Research Institute Co, Ltd., Zhengzhou, China).under a load of 360–380 MPa, with a holding time of 3 min. The sintering process was conducted under hydrogen protection at 860 °C for 1 h, with a heating rate of 3 °C/min, followed by furnace cooling. The re-pressing process was performed at 320–360 MPa with a holding time of 3 min.

The current-carrying friction tests were conducted using a custom-made HST-100 high-speed current-carrying friction tester (China Academy of Machinery Wuhan Research Institute of Materials Protection Co, Ltd., Wuhan, China) (Figure 1). The tester uses a pin-and-disk configuration, with the disk sample driven by a motor via a belt, rotating and creating relative sliding with the pin sample. Current flows into one pin sample, passes through the disk sample, and exits through the other pin sample. A photodiode is used to measure the arc light intensity, and an NAC HX-5 high-speed camera (Nac Image Technology, Tokyo, Japan) is employed to capture the arc morphology.

In the experiment, the disk sample had a diameter of Φ180 mm, and the pin sample had a diameter of Φ10 mm and a length of 25–30 mm, with a pin-to-pin distance of 160 mm. The applied current in the experiment was 120 A, with a relative sliding speed of 30 m/s. Prior to the experiment, both the pin and disk specimens were pre-polished using 600-grit sandpaper for no less than 10 min. After the preparation, the surface roughness of the specimens was Ra 0.8 μm. Each experiment was repeated three times, and the average value of the three trials was used in the paper, except for some dynamic data.

The characterization parameters used in the experiment mainly include the coefficient of friction (the ratio of the real-time frictional force to the normal load), wear rate (the mass of material worn per unit sliding distance), current-carrying efficiency (the ratio of the measured current passing through the friction pair to the applied current), current-carrying stability (the ratio of the standard deviation of the dynamic current to its mean value), relative arc energy (the sum of real-time arc light intensity values), and arc ignition rate (the percentage of time the arc is present during the total test duration).

3. Result

3.1. Frictional Performance Under Different Normal Load Conditions with Current Flow

Figure 2 presents the dynamic curves of the friction coefficient under different normal load conditions. It is evident that the single friction coefficient curve initially exhibits significant fluctuations, and then stabilizes and fluctuates around a lower value. As the normal load increases, both the duration and range of the initial severe fluctuations decrease, and no significant fluctuations are observed at 90 N. It is also evident that as the normal load increases, both the average value and standard deviation of the friction coefficient show a downward trend, with the trend leveling off after 60 N.

Figure 3 presents the histogram of the wear rate. It can be observed that as the normal load increases, the wear rate first decreases, and then increases, with the decrease occurring at a significantly faster rate than the increase. The wear rate is 70 N at its minimum, approximately 0.46 mg/m.

Figure 4 presents the current-carrying efficiency and stability curves of the pairing under different normal load conditions. It can be observed that as the normal load increases, both the current-carrying efficiency and stability remain at a good level with minimal variation. The current-carrying efficiency values range from 0.819 to 0.830, and the stability values range from 0.818 to 0.831.

In summary, from the perspective of current-carrying frictional performance, the primary factor limiting the service life of the pairing is the wear performance.

3.2. Arc Under Different Normal Load Conditions

Figure 5 shows an image of the arc captured by a high-speed camera at 20,000 FPS (Frames Per Second) under a 70 N condition. It can be observed that the arc is generated between the friction surfaces and shows a downward motion, possibly pulling the friction surface. It is also evident that material is ejected from the friction surface during the current-carrying friction process.

Figure 6 presents the statistical curve of arcs with different burnout durations under a 70 N condition. It can be seen from the figure that a total of 3025 arcs occurred during the experiment, with an average burnout duration of 0.70 ms. The burnout durations of the arcs are primarily concentrated below 1.5 ms, with the longest lasting 18.3 ms. Overall, it can be observed that the number of arcs is inversely proportional to the burnout duration of each individual arc.

Figure 7 shows the statistical chart of the arc ignition points under a 70 N condition. The pin sample is divided into 10 equal segments along the friction direction, and the number of arcs ignited in each segment is counted. It can be seen from the figure that the arcs primarily originate from the latter half of the friction surface of the pin sample.

Figure 8 shows the curves of arc energy, arc ignition rate, and number of arcs under different normal load conditions. It can be observed that as the normal load increases, the arc energy, arc ignition rate, and number of arcs all decrease, with the rate of decrease initially fast and then slowing down.

3.3. Material Damage During the Current-Carrying Friction Process

Figure 9 shows the percentage curve of the area of the arc severely eroded zone relative to the total wear area under different normal load conditions. It can be observed that the worn surface can be divided into the arc’s severely eroded zone (the region surrounded by the dashed line in the figure) and the primarily mechanical wear region. Moreover, the arc erosion zone is generally located at the rear part of the wear surface. It can also be observed that as the normal load increases, the area of the arc’s severely eroded zone tends to decrease. At 50 N, the area of the arc’s severely eroded zone is the largest, accounting for 16.06% of the total area, and at 90 N, it is the smallest, accounting for 4.30% of the total area.

Table 1. Surface roughness of different regions on the wear surface under various normal load conditions (Sa, μm).

Regions	50 N	60 N	70 N	80 N	90 N
Front region	5.88	4.63	5.93	7.06	5.44
Rear region	40.18	26.10	35.34	29.78	15.05

shows the surface roughness of different regions of the worn surface under various normal load conditions. It can be seen that in the front region of the worn surface (dominated by mechanical wear), the surface roughness ranges from 4.63 to 7.06 μm; in the rear region (dominated by arc erosion), the surface roughness ranges from 15.05 to 40.18 μm. On the worn surface of the same specimen, the surface roughness in the rear region is approximately 2.8 to 7.2 times that in the front region. Moreover, with increasing normal load, the surface roughness in the rear region shows a decreasing trend.

Figure 10a–e show SEM images of the mechanically worn region near the front part of the worn surface under different normal load conditions. It can be observed that the main damage features on the worn surface are plowing grooves and plastic deformation, and the degree of plastic deformation increases with increasing normal load. Figure 10f shows a cross-sectional SEM image of the mechanically worn region of the worn surface under a normal load of 70 N. It can be seen that plastic deformation occurred in the surface layer of the worn area (with deformation twins observed in the copper grains) [11], with a thickness of approximately 40 μm.

Figure 11a presents an SEM image of molten spatter traces caused by arc erosion under a normal load of 50 N, and Figure 11b shows a magnified view of Figure 11a. It can be observed that multidirectional spatter traces are distributed within a localized area. The authors attribute this morphology to the intense heat released by the arc, which generates a molten pool at the arc root. Under the influence of the electromagnetic field, this molten material is ejected to form the observed spatter [12]. Temporally, the movement of the arc root causes displacement of the molten pool, resulting in the stacking of spatters. Figure 11c shows another type of arc erosion captured in an SEM image. As seen in the image, spherical and near-spherical particles are formed. Given that graphite in the copper–graphite composite does not melt, these particles are presumed to be composed of copper. The authors propose two possible mechanisms for the formation of spherical or near-spherical particles: (1) The arc releases a large amount of heat, locally melting the metal on the friction surface. Due to the poor wettability between copper and graphite, the molten metal contracts under surface tension, forming spherical or near-spherical shapes [13]. (2) The molten metal in the pool is ejected as spatter. If the spatter cools mid-air and falls back, it forms more regular spherical shapes; if it does not fully detach from the surface, it forms elongated or irregular spheres [14].

4. Discussions

Under high-speed current-carrying conditions, the Cu–10 wt.% graphite powder metallurgy composite exhibits favorable current-carrying tribological performance [15]. In this study, the current-carrying efficiency of the friction pair fluctuated between 0.819 and 0.830, and the current-carrying stability ranged from 0.818 to 0.831. The area of severe arc erosion on the worn surface accounted for 4.30% to 16.06% of the total wear area, further supporting this conclusion. Therefore, the primary factor limiting the service performance of the friction pair is its wear resistance.

Under high-speed current-carrying friction conditions, the damage mechanisms of Cu–10 wt.% graphite powder metallurgy composites are primarily mechanical—characterized by plowing and plastic deformation [16,17]—and arc erosion dominated by melting and spattering [18,19]. From the perspectives of the arc energy, arc burning rate, and total number of arcs, all three parameters exhibit a consistent trend with increasing normal load: a rapid decline before 70 N, followed by a more gradual decrease. Therefore, it can be inferred that the sharp deterioration in wear resistance with increasing load is mainly due to the suppression of arcing. After 70 N, however, the increase in wear rate is no longer associated with arcing but rather results from intensified mechanical wear. Consequently, 70 N can be considered the optimal load for achieving the lowest wear rate for this material system.

However, the arc energy, arc burning rate, and total arc count represent the cumulative effects of all arcs, without accounting for the differences between individual arc events. In a single test, high-speed camera footage revealed a large number of arc events. Given the absence of a clear boundary between spark discharge and arc discharge, all observed discharge phenomena in this study are collectively referred to as arc discharges. Under a normal load of 70 N, the average arc burning duration was 0.70 ms, with the longest lasting 18.3 ms. A total of 2054 arcs burned for less than 1 ms, 2588 for less than 2 ms, and 128 arcs exceeded 5 ms in duration, accounting for 4.2% of all arcs. These arcs varied in terms of initiation points, instantaneous intensity, and burning duration, resulting in varying levels of material damage. Thus, analyses based solely on aggregate quantities are inherently limited.

Arc discharges result in spatially and temporally non-uniform material damage [20]. In terms of area, regions dominated by mechanical wear occupy the majority of the worn surface, while arc erosion primarily occurs near the trailing edge of the wear track. Furthermore, most arc initiation points are located in the rear half of the contact surface (as shown in Figure 7). The smoothness of the friction pair’s operation influences arc initiation on the contact surface, and this smoothness is reflected in the fluctuation of the friction coefficient [21]. Except under a 90 N load, all other conditions exhibited significant initial fluctuations in the friction coefficient, followed by periodic oscillations. This suggests that the likelihood of arc initiation is also temporally non-uniform. Therefore, the material damage is considered to be both spatially and temporally heterogeneous [22]. Compared to the relatively uniform damage observed in typical dry sliding wear, this heterogeneity is likely attributable to arc erosion mechanisms involving localized melting and splattering.

A single arc can cause material damage in the form of melting, splattering, and the formation of a heat-affected zone [23]. These effects differ markedly depending on arc parameters. According to the arc energy expression (Equation (1)), individual arcs with high current and extended burning duration possess disproportionately high energy levels [24].

$$\mathrm{E}={\int }_{t_1}^{t_2}{I}^{2}Rdt$$

(1)

In the equation, t₁ and t₂ represent the start and end times of the arc, respectively (ms); I is the instantaneous current passing through the arc (A); and R denotes the instantaneous resistance of the arc (Ω).

In terms of temperature, the surface temperature of typical dry friction is several times lower than that of current-carrying friction. Arcs with high current and long burning duration produce arc root temperatures that differ even more drastically.

From an engineering perspective, numerous arcs occur in the sliding contact between high-speed rail contact wires and collector strips, and individual arcs can inflict severe damage, such as melting and severing the contact wire.

Therefore, in practical applications, it is essential not only to optimize the contact load but also to account for the potential hazards posed by high-energy arcs with large current and long burning duration.

5. Conclusions

• Under the test conditions, the wear rate initially decreased rapidly and then increased gradually with increasing normal load. The current-carrying efficiency and stability of the friction pair remained high, fluctuating slightly within a narrow range. A normal load of 70 N was identified as optimal, primarily based on the minimum wear rate observed (approximately 0.46 mg/m).
• Material damage consisted of mechanical wear, characterized by plowing and plastic deformation, and arc erosion, marked by localized melting and splattering. With increasing normal load, arc erosion gradually decreased while mechanical damage intensified. The overall damage was minimized at 70 N.
• Arc erosion led to pronounced spatial non-uniformity on the worn surface. Particular attention should be paid to high-energy arcs—those with high current and long burning duration—due to their potential to cause severe localized damage.