Characterization of Nanocarbon Copper Composites Manufactured in Metallurgical Synthesis Process

The base material for the melting experiments was high-purity oxygen-free high conductivity (OFHC) copper in the form of 8-mm diameter wires obtained directly from UPCAST continuous casting line. The chemical composition of the base metal is shown in Table I. For metallurgical synthesis, two kinds of carbon were used: CWZ-14 activated carbon and multiwall carbon nanotubes (IGMWNT). The properties of both carbon forms are shown in Table II.

Tests of the metallurgical synthesis method required conditions of electrical current flow and stirring of the molten copper–carbon mixture. This novel method required the design and fabrication of a special device, shown in Figure 1. Figure 1 shows the laboratory stand, which consisted of an induction furnace (1), graphite crucible (not visible, inside induction furnace), current supply device for applied current to the electrode (2), power supply for induction furnace (not labeled), stirring device (3), and inert gas supply (4). Figure 2 shows a detailed schematic of the crucible, electrode for applied current, carbon powder feed tube, stirring impeller and thermocouple. All graphite equipments (crucible, crucible lid, electrode, rotor, blender type hollow shaft) were made from the purest, commercially available, R4550 graphite. Crucible with 110 mm inner diameter was electrically insulated from the sidewalls so that electrical pathway was directly from the upper electrode, positioned in metal melt just beneath the surface, through the melt and to the connection at the bottom of the crucible. In all tests, temperature was measured with the use of type S thermocouple as shown in the Figure 2.

Several trials were necessary before the equipment operated as desired. Each trial produced casts which were designated sequentially from Cast W1 to Cast W10. In this paper, results from cast W5, W9, and W10 are presented.

The procedure for synthesis is as follows:

Tests were carried out using the following instruments and methodology:

During synthesis of covetic, casts with cylindrical shape (diameter 110 mm) were obtained and shown in Figure 3. Cast W5 had surface porosity, while casts W9 and W10 had good (smooth) surface. Additional research results (not described in this paper) confirmed that surface quality depends on (a) the type of mixing device used for stirring and (b) the method of carbon addition. A mixing device that creates a vertical vortex increases the Cu–C mixture stirring intensity which allows for achievement of better surface quality of casts. The ingots were sectioned for testing.

Chemical analysis results are shown in Table IV. An increased content of Fe, Ni, and S in the casts came from known impurities in the carbon materials. Casts W9 and W10 had an increased content of Fe (5.7 ppm—W9, 10.2 ppm—W10). Cast W5 had an increased content of Ni (16.1 ppm). All experimental heats exhibited increased sulfur levels, from 1.2 ppm in the base metal up to 5.6 to 7.6 ppm. The 8 to 9 ppm levels of Ag were consistent with the base metal level of 9 ppm. The very low level of O₂-max 2 ppm is consistent with the Cu OFHC base material. Maximum carbon content value—C_max, attached in Table IV is consistent with the amount of carbon added to the molten metal during each trial.

Research results of density, Vickers hardness (HV5), and electrical conductivity are shown in Table V. Cast W5 had density 8.932 g/cm³ which is the density of measured Cu OFHC base material. Casts W9 and W10 had lower density than Cu OFHC base material (W9—8.906 g/cm³, W10—8.914 g/cm³). Hardness tests indicated that cast W5 had the lowest average hardness (HV5—42.8) and W10 the highest average hardness (HV5—58.1). Tests results of electrical conductivity of casts show that sample W5 had conductivity at the level of 59.4 MS/m (102.4 pct IACS), W9 58.84 MS/m (101.4 pct IACS) and W10—58.1 MS/m (100.2 pct IASC). However, it should be noted that both electrical conductivity and hardness tests results may be influenced by the micro-porosity of produced samples which occurred in casts volume. Nevertheless surprising is the fact that sample W5 has relatively high electrical conductivity despite of high level of contamination with Ni. Cast W9 and W10 contaminated with Fe and S had lower electrical conductivity than the base material used for synthesis. Nevertheless, it should be noted that each ppm of Fe, Ni, and S has very negative effect for electrical properties of copper and high electrical conductivity copper alloys. Therefore, it is very surprising that obtained materials had relatively high electrical conductivity which normally should be at the level of 57 MS/m or even lower.

Photos of cast structures are shown in Figure 4. Test were carried out on samples cut from obtained casts from the center of casting axis. Preparation of samples included macroetching with the use of nitric acid (diluted with water to a 50 to 50 pct ratio). Research results shown that all samples had dendrite structure which is typical for standard casts (low amount of relatively big grains). The pictures also indicate grain boundary areas and dendrite morphology of samples.

Analysis methods of carbon content in copper are very difficult and little known. Quantitative determination of the carbon content in copper using EDS (EDX) or XPS methods is very problematic for samples with low quantities of carbon. In those types of tests carbon will be always present in the microscope chamber and will give “false” carbon peaks on EDS (EDX) or XPS diagram. Also performed during the experiment, carbon measurements with the use of Leco method (ASTM E1019-08) did not give a good result (intense discrepancy of carbon content between the individual tests). SIMS method on the other hand gives a good result in qualitative determination of carbon presence in copper but cannot be used for quantitative determination. Using the SIMS method, we can compare carbon presence (ion yield) between different samples.

Research results of carbon presence and distribution in produced casts using SIMS method show that in sample cast W5 very small amounts of carbon were detected whereas in the samples from W9 and W10 there was much more carbon (in the cast W9 in the form of columns and in the cast W10 it is evenly distributed in the sample volume). Figure 5 shows the visualization of carbon content in W5, W9, W10 casts samples (C₂ presence is marked in white). Reference point is Cu OFHC base material where we do not observe any C₂ presence. Figure 6 shows the relative magnitudes of ion yields Cu, 65Cu, C, C₂, CuC vs depth of W9 and W10 samples.

Figure 7 presents photos of microstructure and diffraction of W9 and W10 samples. Research results of TEM microstructure show that in the structure of casts W9 and W10 there was a second phase. In the picture A and B, visible dark features possibly correspond to copper–carbon crystalline particles. In the picture B also dislocation structures were being observed. Visible in the picture B individual white particle is a void that comes from an inclusion that was dissolved during the sample preparation. Diffraction pictures confirm that close to the Cu pattern there is a second pattern which actually may come from carbon structure incorporated into Cu lattice (possibly a graphene-type carbon structure). Similar research results can be found in Reference 17. In addition, a strong stretch of diffraction spots along arcs indicates a large distortion of crystalline structure.