INFLUENCE OF THE FABRICATION PROCESS OF COPPER MATRIX COMPOSITES ON CAVITATION EROSION RESISTANCE

Copper matrix composites reinforced with ZrB2 particles were produced in two ways: by hot pressing (HP) and laser-sintering process. Powder mixture Cu-Zr-B was mechanically alloyed before densification processes. Variations in the microstructure of treated samples obtained during cavitation test were analyzed by scanning electron microscopy (SEM). Cavitation erosion resistance was investigated with the standard test method for cavitation erosion using vibratory apparatus. Changes in mechanical alloying duration show a strong influence on cavitation erosion resistance of Cu–ZrB2 composites regardless the number of reinforcements. Laser-sintered samples show better cavitation erosion resistance than hot-pressed samples.


Introduction
In past decades, considerable attention was paid to the production of composite materials with the best combination of mechanical and physical properties. The copperbased composites are being examined as potential candidate materials in aerospace, automotive, military and electrical industry [1][2][3]. Low content of alloying elements in the copper matrix enables the highest possible ratio of mechanical/physical properties sufficient to withstand demanding operating conditions even in nuclear technology and rocket industry [4,5]. Previous studies [6][7][8] showed that addition of ZrB2 as a hardening phase into the copper matrix significantly improved its mechanical properties, with maintaining high electrical and thermal conductivity. Combining mechanical alloying (MA) with hot pressing process enables proper distribution and in situ formation of ZrB2 particles in the copper matrix. This method shows excellent influence on mechanical and tribological properties since strong interfacial bonding between the coarse ZrB2 particles and Cu-matrix is avoided [3][4][5]. On the other hand, by using laser-sintering, high super saturation occurs providing better formation of finer ZrB2 reinforcing particles in the copper matrix. During MA particles of Zr and B are mechanically activated which enables in situ formation of ZrB2 in a copper matrix in the course of hot pressing process or lasersintering [6]. All these facts enable excellent mechanical properties of the Cu-ZrB2 composite. The occurrence of cavitation erosion is widespread in many engineering materials. This fact justifies an investigation of this kind of erosion of Cu-ZrB2 composite.
The phenomenon of cavitation involves the formation, growth, and collapse of bubbles in a liquid due to the local pressure changes. When the pressure in liquid drops below the vapor pressure at a given temperature, microscopic bubbles are formed [9]. These bubbles can collapse suddenly if the liquid is subsequently subjected to a higher hydrostatic pressure. Collapsing bubbles produce shock waves and microjets which can cause high temperatures (5000˚C) and pressures (103 MPa) on the surfaces of small dimensions (10 -10 m 2 ) in a short period (< 1 μs) [10,11]. If these collapses repeated close to the solid surface, it is subjected to fatigue which can cause cavitation erosion. Material response to cavitation erosion is usually described as mass loss versus exposure time, and ideally the plotted data form the S-shaped curve. Formation of S-curve consists of several stages: incubation period, accelerating rate period, constant rate period, followed by oscillating rate periods and, in some cases, a final lower constant rate period [12]. In practice, the shape of this curve depends on material properties and its response to cavitation attack, i.e., cavitation resistance.
Several research studies have been devoted to understanding the phenomenon of cavitation erosion of conventional materials, which are used in the production of hydraulic machinery and other mechanical parts. They proved that the cavitation erosion resistance depends on mechanical parameters (hardness, tensile strength, Young's modulus, fatigue strength), microstructure (grain size, some material defects, present phases), and also on surface roughness [13][14][15][16]. Hence, many studies investigated the cavitation erosion resistance of different materials such as gray cast irons, ductile irons, stainless steels, nonferrous metals and alloys such as aluminum and copper alloys [17][18][19][20][21].
The present work includes analysis of the influence of mechanical alloying duration on the cavitation erosion resistance of copper matrix composites as well as the influence of different fabrication processes. Cavitation test was performed on hot-pressed samples obtained from powders after 5 and 30 hours of MA, and on laser-sintered samples obtained from powders after 30 hours of MA. The best results of cavitation testing for both processes were compared. Microscopic investigations support discussion of the results.

Preparation of hot-pressed and laser-sintered copper matrix composites
Hot pressing of mechanically alloyed mixtures (Cu1.1Zr-0.3B (wt.%) and Cu-4.1Zr-1.1B (wt.%)) was carried out in Astro furnace with graphite mold (10 mm diameter). The heating rate was 15˚C/min, at a temperature of 950 °C and under a pressure of 35 MPa. Retention time was 2.5 hours. Obtained compacts were 5 mm in height. For laser-sintering method, mechanically alloyed powders (Cu-4.1Zr-1.1B (wt%), and after 30 hours of MA) were cold pressed by the applied pressure of 180 MPa, to produce green compacts. Laser-sintering of green compacts was carried out with Nd: YAG pulsed laser radiation in the nitrogen atmosphere. Samples were distinguished by a number of scans from 1 up to 4. Except for the number of scans, parameters of laser sintering for all samples were the same: frequency 3 Hz, pulse duration 10 ms (for the sample with 4 scans 8 ms), pulse energy 22 J (for the sample with 4 scans 18.5 ms). Obtained laser-sintered samples were 300 μm in height. Properties of obtained copper matrix composites are given in Table 1. As was reported in previous studies [6][7][8] in situ forming of ZrB2 particles was achieved during hot-pressing and laser-sintering process. The density of the samples was determined by Archimedes method, and the rule of mixture determined theoretical density.

Cavitation test
The ultrasonic vibratory cavitation test set up (stationary sample methodaccording to the ASTM G32 standard) was used for cavitation erosion analysis [22]. The frequency of vibration and peak-to-peak displacement amplitude of the horn were 20 kHz and 50 μm, respectively, with separation of 0.5 mm between the sample and the horn tip. The test liquid was water maintained at 25°C [22][23][24][25]. The mass loss measurements were performed after each exposure to the cavitation (every 10 min) for a test period of 120 min for HP compacts and laser-sintered samples every 10 min for a test period of 60 min. Differences in cavitation test duration are due to a thickness of the sample. Before and after each test interval the samples were cleaned and dried with hot air. Mass losses of the tested samples were measured using an analytical balance with an accuracy of ±0.1 mg. Cavitation erosion rate was calculated as a slope of the line obtained after a linear fit of mass loss vs. exposure time curve.

Microscopic examination
The microstructure of samples before and after cavitation test was investigated by JEOL-JSM 5800LV scanning electron microscope (SEM) at an accelerating voltage of 20 kV. Detailed analysis of microstructure before cavitation test was reported in our previous studies [6][7][8].

Results and discussion
Size and distribution of reinforcements in copper matrix depend on their percentage and mechanical alloying duration or number of scans in case of laser-sintering and show a strong influence on mechanical properties [8,26]. The highest densities and hardness have samples obtained by hot-pressing from powders mechanically alloyed for 30 h or by 4 scans during laser-sintering. Compared to pure copper sample, the full densification of the Cu-ZrB2 alloy was not achieved because of hardening effects due to the presence of ZrB2 particles and agglomerates with varying sizes [6]. Therefore, it was noticed that longer milling time provides more homogeneous distribution of reinforcing particles and better densification which have a substantial influence on the hardness of this composite. In previous studies [7,8] amount of present reinforcements in the copper matrix was determined as 1% ZrB2 for system Cu -1.1 wt.% Zr -0.3 wt.% B; 7% ZrB2 for system Cu -4.1 wt.% Zr -1.1 wt.% B after hot-pressing; around 3.5% for lasersintered samples of a green compact with Cu -4.1 wt.% Zr -1.1 wt.% B composition. As was reported [6,7], the presence of higher amount of submicron reinforcing particles in the copper matrix was observed after 30h of MA comparing to 5h, which leads to higher hardness. In Fig. 1. results of cavitation erosion testing of hot-pressed (Fig.1a) and lasersintered (Fig.1b) samples are shown. Hot-pressed samples with the same duration of mechanical alloying show similar cavitation resistance despite the various amount of present reinforcements. However, observing the overall behavior of hot-pressed samples, the influence of reinforcement amount on cavitation resistance can be noticed, especially after 30 min of cavitation erosion test. Sample 30hHP-H which shows the highest cavitation erosion resistance also shows the highest mass loss at the beginning of cavitation test, but after 30 min the mass loss rate decreases, i.e., steady-state rate occurs due to surface roughness. The behavior of laser-sintered samples during cavitation erosion test depends on the number of scans. According to the density and hardness values of the laser-sintered samples, it was expected that LS4x shows the highest cavitation erosion resistance and LS2x the lowest.

Fig. 1. Mass loss during cavitation erosion testing of a) hot-pressed and b) lasersintered samples.
All samples do not show any incubation period during cavitation erosion testing, probably due to the presence of agglomerates or coarse ZrB2 particles which provide increasing of some local peeling of the copper matrix. Also, ZrB2 particles can sustain higher loads compared to the copper matrix because of their higher hardness and consequently local strain of the copper matrix is higher. On the other hand, the presence of reinforcements in copper matrix increases the dislocation density and decreases the grain size since the reinforcing particles act as sites of nucleation during solidification. The grain size decreases due to the formation of subgrains with a high dislocation density. Also, it is supposed that in this composite material the subgrains are formed and surrounded by ZrB2 particles which can act as barriers to the dislocations movement. As can be noticed, samples 30hHP-H and with LS4x show the best results in performed cavitation erosion testing which confirms the correlation between cavitation resistance and mechanical and microstructural parameters. Therefore, comparing these two samples, it can be concluded that laser-sintering technique provides samples with slightly better cavitation erosion resistance than hot-pressed samples. This fact is supported by similar values of cavitation erosion rate after 60 min of exposure (0.32 mg/min for 30hHP-H and 0.41 mg/min for LS4x) despite the significant difference in the sample thickness (5 mm for 30hHP-H and 300 µm for LS4x) as well as the difference in amount of ZrB2 particles (7% in 30hHP-H and 3.5% in LS4x). Lower values of cavitation rates indicate higher cavitation erosion resistance (Fig. 2).

Fig. 2. Cavitation erosion rates of hot-pressed and laser-sintered samples.
The SEM investigation has shown evenly spread corrugated surface of the copper sample and randomly distributed pits in the corrugated surface of other samples. Since material response to cavitation erosion attack differs depending on the production technique, mechanisms of damage also vary. As was reported [21], the final surface damage caused by cavitation erosion is a consequence of two mechanisms: the individual collapse of bubbles and simultaneous collapses of the whole cloud of bubbles close to the solid surface. In ultrasonic vibratory cavitation erosion test set these mechanisms can occur individually or jointly. The progress of cavitation erosion on the surface of hotpressed samples observed by SEM is shown in Fig. 3. Damaged surfaces of CuHP, 5hHP-L, and 30hHP-H after 120 min of cavitation erosion test are shown in Fig. 3a-c, respectively. These SEM micrographs show mainly corrugated surface. Since copper is a ductile material with a low hardness, it shows a high material loss. It is assumed that the dominant mechanism is a simultaneous collapse of the whole cloud of bubbles. On the other hand, sample (30hHP-H) with the lowest material loss shows the lesser number of shallow pits probably due to the occurrence of mixed ductile/brittle-mode cracks caused by the presence of ZrB2 particles, which indicates that individual collapse of bubbles is the more dominant mechanism. Eroded surface of CuHP sample shows plastically deformed grains due to the presence of cleavage as dominant fracture mode, but the ductile fracture was also noted (Fig. 3a). Addition of ZrB2 particles induces mixed fracture (Fig. 3b,c), both ductile and brittle mode cracks occurred, and cavitation attack is primarily on grain boundaries (a few mentioned cracks occurred on the grain boundary were marked with arrows). The corrugated surface is observed in all samples, and distribution and depth of pits are in correlation with the distribution of reinforcing particles. Hot-pressed samples show grain boundary erosion and distribution of pits inside the grains depends on amount and distribution of ZrB2 particles as well as grain size. Observing the surface of 30hHP-H sample shown in Fig. 3c, it can be noticed the presence of non-damaged areas (marked with arrows) as a consequence of relatively high hardness due to the presence of submicron ZrB2 particles and probably small grains. Laser-sintered samples show similar behavior which is presented in Fig. 4. Sample LS2x (Fig. 4a) shows high material loss between scanning lines after 60 min of cavitation erosion testing due to the presence of pores or unmelted areas. Sample LS3x (not shown in this paper) exhibits the similar behavior as LS2x, i.e., erosion along areas between scanning lines as the dominant damaged surface but with the lower material loss. Observing surface of LS4x sample (Fig. 4b) after 60 min only a few pits were noticed surrounded mainly by non-damaged areas. With increasing the number of scans, material porosity decreases (as well as unmelted regions) and cavitation resistance increases. Also, the presence of ZrB2 increased the work-hardening rate and combined with all facts mentioned above enables the individual collapse of bubbles as a more dominant mechanism. It may be concluded that cavitation erosion resistance depends on mechanical alloying duration, as well as production process. Overall, as was expected, samples with higher hardness show better cavitation erosion resistance.

Conclusion
In this paper mass loss was used as a measure of cavitation erosion resistance of copper matrix composites. By adding the reinforcing ZrB2 particles in the copper matrix, cavitation erosion resistance of the matrix improves. In the first 30 min mass loss has almost identical trend for samples with the same amount of ZrB2 particles, and after 30 min greater influence on the trend of mass loss vs. time curve shows mechanical alloying duration. Also, laser-sintered samples show slightly better cavitation erosion resistance compared to hot-pressed samples (taking into account the significant difference in the sample thickness), and it can be concluded that production techniques influence cavitation erosion resistance. On the other hand, duration of mechanical alloying, distribution and size of ZrB2 particles in the copper matrix increase its ability to work-hardening and absorption of cavitation energy. The lowest cavitation erosion resistance of Cu-ZrB2 composites shows laser-sintered sample after 2 scans due to high porosity and unmelted regions. With increasing the number of scans, the cavitation resistance also increases.