2019 Volume 60 Issue 1 Pages 2-9
The effect of solidification conditions on the tensile deformation behavior of pure copper castings for electrical parts was investigated. Two main types of tensile deformation properties were distinguished on the basis of the difference in uniform elongation. For the castings fabricated under a superheat of 100°C or 150°C, larger and smaller uniform elongation types corresponded to the absence and presence, respectively, of the Cu–Cu2O eutectic phase in the microstructure. Meanwhile, for the castings fabricated under a superheat of 50°C, greater uniform elongation was sometimes obtained when the eutectic phase was present. In addition, irrespective of the presence or absence of the eutectic phase, greater uniform elongation was always obtained when chills were used. Cross-sectional observations showed the existence of considerable nonspherical porosity when the eutectic phase was present; the porosity was reduced when the pouring was conducted under the superheat of 50°C and when the chills were used because of lower hydrogen content in the melt and supersaturation of the hydrogen by rapid cooling, respectively. These results suggest that not only the presence of the eutectic phase but also the inferior casting soundness due to the existence of the porosity is a dominant factor responsible for the decrease in the uniform elongation. The findings presented here indicate that a decrease in the hydrogen content in the melt and/or the rapid cooling during solidification are effective measures to stably achieve practically sufficient deformation properties along with superior casting soundness.
Crimp terminals and branched sleeves are used for electrical wire connections in applications such as electrical transmission and substation equipment and electric railways. Because these products require high electrical conductivity, they are made from pure copper and are manufactured by sand-mold casting to achieve near-net-shape and high-mix low-volume production. However, pure copper exhibits substantial thermal and solidification shrinkage and absorbs large amounts of hydrogen and oxygen during the casting process. These effects lead to casting defects such as shrinkage cavities, hot tears, blow holes and porosity.1–6) Thus, both degas (de-hydrogen) and deoxidization treatments and/or the optimization of the casting design are required in the manufacturing processes for pure copper castings.1–4)
The crimp terminals and branched sleeves are connected to wires by crimping them using a specialized hydraulic tool. To avoid cracking during the crimping process, the terminals and the sleeves should be ductile. Therefore, the casting process for these products should also employ a method to impart the products with adequate ductility. In previous work, we investigated the relationship between the solidification structure of pure copper castings and their oxygen content.5) The results showed that trace oxygen at concentrations of 0.01–0.02 mass% led to the formation of a Cu–Cu2O eutectic phase in the microstructure of the pure copper casting.5) In addition, the castings with the Cu–Cu2O eutectic phase exhibited higher electrical conductivity than those without the phase.5,6) However, the presence of the eutectic phase likely makes pure copper castings brittle.5) To ensure both high electrical conductivity and ductility of pure copper castings, the new casting process must be improved to make to impart pure copper containing the eutectic phase with sufficient ductility.
Through the aforementioned investigation, we also found that the differences in crystal grain morphology and/or hydrogen behavior due to differences in the cooling conditions during the solidification process affect the ductility of the pure copper castings.5,6) In the present study, to clarify these effects, we prepared pure copper castings under several solidification conditions, where various degrees of superheat were employed and some conditions included chills, whereas others did not, and subsequently conducted tensile tests using the castings. On the basis of the results, we propose a casting method that can stably provide ductile pure copper castings for electrical parts.
No. 1 copper wire (JIS H 2109, high purity scrap, ≥99.9 mass%) and electrolytic copper with a total weight of approximately 1.1 kg were melted under atmospheric conditions using a high-frequency induction furnace and a graphite crucible. The mixing ratio of the electrolytic copper was approximately 10–20 mass%. Some charcoal pieces were charged into the crucible during melting to prevent oxygen absorption. The melt was sequentially treated with a (C2F4)n-based commercial degasifying agent and phosphor copper deoxidizer (Cu-15 mass%P) in the temperature range from 1200 to 1250°C. The amount of added phosphor copper was approximately 4.4 g, which was controlled to 0.4 mass% of the raw materials. After the temperature was maintained at 1150–1250°C for approximately 2 min to stabilize the reaction,4) the deoxidized melt was poured into a sand mold (room temperature) under a specific superheat condition (50°C, 100°C, or 150°C). Figure 1 shows the casting design, where a sand mold with two-piece cavities was employed.5,6) The sand mold was made from silica sand and water glass by the CO2 process. Plates of spheroidal graphite cast iron (room temperature) were used as chills. The molds and the plates were coated with graphite powder spray. After the pouring, the melt surface at the sprue was covered with charcoal powder to prevent oxygen absorption. The four sets of solidification conditions employed in this study are shown in Table 1.
The tensile properties were evaluated using tensile tests at room temperature (25°C). JIS Z 2241 No. 8B specimens were used in the tests; these specimens were prepared from the positions of the castings specified in Fig. 1. The lengths of reduced and gauge sections of the specimen were 14.6 mm and 10 mm, respectively. The outside of the gauge section was painted black for clear delineation of the edges, and the elongation of the specimen (i.e. the width change between the edges of the blacked sections) was then measured in real time using a vision sensor (a real-time image-processing technique using a camera). The crosshead speed was set to 0.1 mm/s. The corresponding actual strain rates (at the gauge section) ranged from 0.15 to 0.25% s−1. The strain at the onset of the test force change of less than −5000 N/s was recorded as the rupture elongation. In the case of not only fracture at the outside of the gauge section but also approximately no necking, the rupture elongation was obtained in the same manner as the fracture at the inside of the section.
After each test, the fracture surface of the specimens and the cross-section in the vicinity of the fracture surface were observed by optical microscopy and scanning electron microscopy (SEM). The samples of the cross-section in the vicinity of the fracture surface were electropolished; these were used to compare the solidification structures of samples cast under different conditions. The electropolishing was conducted using an electrolytic aqueous solution of phosphoric acid (53 mass%) under an applied voltage of 2 V.7) The electrical conductivity of the castings was measured by an eddy-current method, through which the converted value at 20°C was obtained. The chemical compositions of the castings were also analyzed by optical emission spectroscopy (Shimadzu Corporation, PDA-7000). The oxygen content was separately analyzed by an inert-gas fusion infrared absorption method (JIS H 1067). The portions of the castings in the vicinity of the specimens were used for these measurements and analyses.
Figure 2 shows typical stress–strain curves of samples prepared under the various solidification conditions. As shown in Fig. 2, curves showing different stress levels and rupture elongations were obtained, with no clear correlation with the solidification conditions. In particular, the rupture elongation values varied more than those in a previous study,5) which were difficult to classify according to the magnitude of the elongation.
Typical stress–strain curves (smoothed).
Figures 3(a) and 3(b) show the relationships between tensile strength and uniform elongation and between necking and uniform elongation, respectively. Here the uniform elongation refers to elongation at the maximum stress, which is a practically important tensile property because some cracks may occur after the maximum stress.5,6) The solid line in Fig. 3(a) indicates the minimum tensile strength (155 MPa) of an industrial standard for electrical pure copper castings, JIS H 5120 CAC102. The dashed lines in Figs. 3(a) and 3(b) also indicate the requisite uniform elongation (26%) to satisfy the minimum rupture elongation of CAC102 (35%), which was evaluated using JIS Z 2241 No. 10 specimens.6) Both the rupture elongation measurement according to JIS and a direct comparison with the minimum of CAC102 are difficult because of the short reduced section of the No. 8B specimens. Therefore, the requisite uniform elongation was used as an alternative standard in this study.
Relationship between uniform elongation and (a) tensile strength and (b) necking elongation.
As shown in Fig. 3(a), two main relationships between tensile strength and uniform elongation were distinguished for specimens on the basis of the difference in the uniform elongation, where the threshold between types was approximately 26%. The two relationships are referred to here as the smaller elongation type and the larger elongation type. For the larger uniform elongation type, practically sufficient tensile strengths and uniform elongation were obtained, satisfying the standards and the requirements specified in CAC102, respectively. Both uniform elongation types were observed among the specimens fabricated under superheats of 50°C, 100°C and 150°C (without chills). By contrast, the specimens fabricated using chills were consistently categorized as the larger uniform elongation type. The results of the two castings (four specimens) fabricated using a one-side chill are also included in Fig. 3. Although the no-chill sides of the specimens belonged to the smaller uniform elongation type, the chill sides of the specimens belonged to the larger uniform elongation type. These results suggest that the use of chills is effective for stably ensuring practically sufficient tensile properties.
Similar to Fig. 3(a), Fig. 3(b) shows two main types of relationships between necking and uniform elongation, where the two types are differentiated on the basis of the difference in the necking elongation. Samples that exhibited larger uniform elongation belonged to the both larger and smaller necking elongation types, where the specimens fabricated under a superheat of 150°C consistently belonged to the smaller necking elongation type. Although Fig. 3(a) and 3(b) also show that tensile strengths greater than 200 MPa were sometimes obtained, these specimens were always categorized as both larger uniform elongation and smaller necking elongation types. Meanwhile, as shown in Fig. 3(b), the samples categorized as exhibiting smaller uniform elongation always belonged to the smaller necking elongation type irrespective of the solidification conditions; no sample exhibited a combination of smaller uniform elongation and larger necking elongation. Thus, three types of tensile properties were comprehensively classified according to the combination of the uniform and necking elongations, which did not depend directly on the solidification conditions.
3.2 Electrical conductivity and chemical compositionsTable 2 shows the typical electrical conductivity and chemical compositions of the castings prepared under various solidification conditions. The presence or absence of the Cu–Cu2O eutectic phase as a result of variations in the oxygen content, as determined on the basis of the cross-sectional microstructure observations, is indicated in Table 2.5) The results in Table 2 indicate that the castings exhibited higher electrical conductivity and greater purity than the practical standards of CAC102. Castings with lower phosphorus contents tend to contain more oxygen, which leads to formation of the Cu–Cu2O eutectic phase,5,8) and to exhibit slightly higher electrical conductivity. These tendencies were confirmed in previous studies,4–6) which suggests that the presence of the eutectic phase promotes stably higher electrical conductivity.
Figures 4(a) and 4(b) show the relationships between uniform elongation and oxygen content and between necking elongation and oxygen content, respectively. As shown in Fig. 4(a) as well as Table 2, there was an overall tendency that lower oxygen contents of less than 0.017 mass% always lead to the both absence of the eutectic phase and larger uniform elongation. For the castings fabricated under a superheat of 100°C or 150°C, the larger and smaller uniform elongation types correspond to the absence and presence of the eutectic phase, respectively. By contrast, for the castings fabricated under a superheat of 50°C, greater uniform elongation was sometimes obtained when the eutectic phase was present. In addition, irrespective of the oxygen content and the presence or absence of the eutectic phase, greater uniform elongation was always obtained when the chills were used.
Relationship between oxygen content and (a) uniform and (b) necking elongations.
As shown in Fig. 4(b) as well as Table 2, there was also an overall tendency that higher oxygen contents of more than 0.024 mass% always lead to the both presence of the eutectic phase and smaller necking elongation. The larger and smaller necking elongation types approximately correspond to the absence or presence of the eutectic phase. For the castings fabricated under a superheat of 50°C or 100°C, smaller necking elongation was infrequently obtained when the eutectic phase was absent. Meanwhile, for the castings fabricated under a superheat of 150°C, smaller necking elongation was always obtained despite the oxygen content and the presence or absence of the eutectic phase.
3.3 Fracture surface of specimensFigure 5 shows the appearances of the fracture surface of the specimens. As shown in Figs. 5(a), 5(b), 5(e) and 5(f), cavities were sometimes observed on the fracture surface. Several cavities have elongated shapes, as shown in Figs. 5(a), 5(b) and 5(e), which were approximately along the conceivable outside-in solidification direction. This observation suggests that the cavities were blow holes, as observed in lotus-type porous copper.9,10) For the specimens lacking both the eutectic phase and cavities, substantial area reduction along with larger necking elongation was observed (Figs. 5(c) and 5(g)). In particular, for the specimens fabricated using the chills, the fracture surface was shaped like an ellipse (Fig. 5(g)), and no cavities were observed. Meanwhile, as shown in Figs. 5(b) and 5(e), the specimens without the eutectic phase and with cavities were classified as smaller necking elongation type. These tendencies are similar to those observed for the specimens prepared under a superheat of 50°C. In addition, for the specimen with constantly smaller necking elongation fabricated under a superheat of 150°C, cavities were always observed (Fig. 5(e)). For the specimens with larger necking elongation and in which the eutectic phase was present, smaller area reduction was observed despite the difference in the superheats (Figs. 5(a) and 5(d)). By contrast, for the specimens fabricated using the chills (Fig. 5(f)), the area reduction was larger than in the specimens fabricated without using the chills (Figs. 5(a) and 5(d)), which only exhibited smaller cavities.
Appearances of the fracture surface of specimens.
Figure 6 shows SEM micrographs of the fracture surface of the specimens. For the specimens without the eutectic phase, dimpled ductile fracture surfaces were observed, as shown in Figs. 6(c), 6(e), 6(f) and 6(h), irrespective of whether the necking elongation was larger or smaller. In particular, for the specimens fabricated using the chills, fine streaky dimples were observed (Fig. 6(h)). However, for the specimens with the eutectic phase and fabricated under a superheat of 100°C, brittle grain-boundary fracture was observed (Fig. 6(d)). This tendency was similar to that observed for the specimens fabricated under a superheat of 150°C. For the specimens with larger uniform elongation fabricated under a superheat of 50°C or using the chills, dendritic phase-boundary fractures between the Cu primary phase (solid solution) and the eutectic phase were also observed (Figs. 6(b) and 6(g), respectively). In particular, for the specimens fabricated using the chills, a directional fine structure of columnar dendrite was observed (Fig. 6(g)). Meanwhile, for the specimens with smaller uniform elongation fabricated under a superheat of 50°C, intermediate behavior between the grain and phase-boundary fractures was observed (Fig. 6(a)), irrespective of finer dendritic structure than that shown in Fig. 6(b).
SEM micrographs of the fracture surface of specimens.
Figure 7 shows the macrostructure of the electropolished cross-section in the vicinity of the fracture surface of the specimens. For the specimens without the eutectic phase, a radial structure of columnar crystal grain along the solidification directions was observed (Figs. 7(c) and 7(e)). For the specimens fabricated using the chills, the unidirectional structure was also observed (Fig. 7(g)). These results indicate that the cooling effect of the chills was sufficient, suggesting that the crystal grain structure has less effect on the uniform and necking elongation types. As shown in Figs. 7(c) and 7(g), microsegregation was sometimes observed inside the crystal grains.6) However, for the specimens with the eutectic phase, a nondirectional structure of equiaxed crystal grains was observed (Figs. 7(a), 7(b) and 7(d)). For the specimens fabricated using the chills, the directional fine structure was also partially observed (Fig. 7(f)). In addition, for the specimens with constantly smaller uniform elongation fabricated under a superheat of 100°C, considerable nonspherical porosity was observed (Fig. 7(d)). Figures 7(a) and 7(b) show that the porosity was reduced in samples fabricated under a superheat of 50°C. The porosity was also stably reduced in samples fabricated using the chills (Fig. 7(f)). These results imply that the porosity corresponds to the magnitude of the uniform elongation.
Macrostructure of the electropolished cross-section in the vicinity of the fracture surface of specimens.
Figure 8 shows SEM micrographs of the electropolished cross-section in the vicinity of the fracture surface of the specimens without the eutectic phase. As shown in Fig. 8(a), spherical microporosity was observed between the columnar dendrites. By contrast, for the specimens fabricated using the chills, fine streaky microporosity was observed (Fig. 8(b)). The difference in these morphologies approximately corresponds to the difference in the dimple morphologies in the fracture surfaces shown in Figs. 6(f) and 6(h).11)
SEM micrographs of the electropolished cross-section in the vicinity of the fracture surface of specimens.
The aforementioned results suggest that not only the presence of the Cu–Cu2O eutectic phase but also the inferior casting soundness is a dominant factor affecting both uniform and necking elongations. Figure 9 shows a schematic of the plausible fracture mechanisms under the three classifications of tensile properties discussed in Section 3.1. As shown in Fig. 9(a), brittle grain-boundary fracture occurred in the specimens with both the eutectic phase and nonspherical porosity. This behavior is attributed to cracking of the eutectic phase in the vicinity of the grain-boundary being promoted by stress concentration at the pores. As a result, smaller uniform and necking elongations were obtained. However, in the specimens with the eutectic phase and with less porosity, dendritic phase-boundary fracture occurred (Fig. 9(b)), leading to both larger uniform elongation and larger area reduction. These results indicate that the presence of the eutectic phase does not adversely affect the ductility. In this case, tensile strengths greater than 200 MPa were sometimes obtained, which suggests that the dispersion strengthening was caused by the presence of the Cu2O12) and/or that crack propagation was often deflected by both the substantially lower porosity and the dendritic phase boundary.13) In addition, as shown in Fig. 9(c), ductile fracture associated with the formation of the dimples occurred for the specimens without the eutectic phase but with rounded defects such as the cavities and microporosity. The rounded defects have less stress concentration despite their size, which leads to greater uniform elongation. Meanwhile, large defects such as the cavities reduce the cross-sectional area, possibly leading to slightly smaller tensile strength and uniform elongation as well as smaller necking elongation.5,6,9,14) For the specimens fabricated using the chills, the ellipse-shaped fracture surface also suggests that the streaky microporosity caused an easy deformation path along the array.
Schematic of the fracture mechanisms (a) when both the Cu–Cu2O eutectic phase and pores are present and when the eutectic phase is (b) present and (c) absent.
The aforementioned casting defects such as cavities and pores are generally caused not only by higher hydrogen and oxygen contents in the melt but also by a slower cooling rate and/or a lower temperature gradient during the solidification process.6) The hydrogen and oxygen in the melt form water vapor during the solidification process.1,2,4,10) Lower superheats lead to a lower hydrogen content of the melt, which can in turn lead to fewer defects by reducing the amount of water vapor generated.1,2,4,15) Meanwhile, the difference in fracture morphologies observed in Figs. 9(a), 9(b) and 9(c) obviously depends on the oxygen content, which is responsible for the presence or absence of the eutectic phase. Because the hydrogen and oxygen contents in the melt show an inverse relationship,1,2,4) the hydrogen content can be increased in the case of lower oxygen content (Fig. 9(c)), particularly under higher superheats. These observations correspond to the frequent cavity generation observed in samples fabricated under a superheat of 150°C.9,10) However, both the porosity and the amount of eutectic phase suggest that the oxygen content in the specimen represented in Fig. 9(a) is lower than that in the specimen in Fig. 9(b) and that the hydrogen content in the specimen represented in Fig. 9(a) is higher than that of the specimen in Fig. 9(b). In the present case of lower oxygen content and higher hydrogen content, a greater amount of water vapor is generated, which leads to the generation of dispersed porosity between the equiaxial primary phases.16) Vapor nucleation during the solidification process can also be promoted by an increase in the liquid concentration and/or the existence of Cu2O.6,17) As a result, the inferior casting soundness results in the fracture morphology observed in Fig. 9(a). By contrast, in the case of higher oxygen content and lower hydrogen content, a greater amount of the eutectic phase with lower porosity tends to be generated,16) particularly under lower superheats. This effect causes the fracture morphology shown in Fig. 9(b). In addition, the use of the chills causes rapid cooling, leading to the generation of both less water vapor and fewer casting defects, along with supersaturation of the hydrogen, irrespective of the presence or absence of the eutectic phase.6,8,10,18) Although the temperature gradient during the solidification process is lower because of the higher thermal conductivity of pure copper, the use of chills also induces a greater temperature gradient as well as solidification directionality. These effects lead to both a narrower liquid–solid mushy zone and the entrapment of less water vapor between the primary solid phases during the solidification process. Thus, the fracture morphology shown in Fig. 9(a) can be prevented by the use of chills, irrespective of the hydrogen and oxygen contents.
The aforementioned findings indicate that a reduction of the hydrogen content in the melt and/or rapid cooling during the solidification process are effective measures to stably achieve practically sufficient deformation properties along with superior casting soundness irrespective of the presence or absence of the eutectic phase. In particular, the rapid cooling using chills is easy to implement and is less likely to result in undesirable effects caused by differences in the superheats than processes where chills are not used. On the basis of this study, both sufficient deformation properties and greater electrical conductivity of approximately 100% IACS can be achieved when the eutectic phase is present. Such conditions are likely to be stably ensured by not only the use of chills but also by melting under an oxidative atmosphere and/or no deoxidization treatment.2,5,8)
In this study, the effect of solidification conditions on the tensile deformation behavior of pure copper castings for electrical parts was investigated. The following conclusions were drawn from the results of this study:
This work was supported in part by the Yamaguchi Educational and Scholarship Foundation.
