Aging Behavior and Precipitation Analysis of Cu-Ni-CoSi Alloy

Cu-Ni-Si alloy with a different Co content was prepared by inductive melting and hot rolling. The alloy was solution treated at 950 ◦C for 1.5 h and aged at 450 ◦C, 500 ◦C, and 550 ◦C for different times. The phase diagram calculation and transmission electron microscopy was used to investigate the effect of Co addition on the aging precipitation behavior of the Cu-Ni-Si alloy. The phase transformation kinetics equation was calculated as well. The results show that, with the increase of aging temperature, the two-phase region of Fcc + Ni2Si in the Cu-Ni-Si ternary diagram would get wider. Some NixSiy phases would also form in the Cu-rich isothermal section. The addition of Co would replace part of Ni to form the (Ni, Co)2Si phase, which inhibits the spinodal decomposition process of the Cu-Ni-Si alloy during the aging process. The precipitated phase of the Cu-Ni-Si alloy with a high content of the Co element is more likely to grow with the extension of aging time. The phase transformation kinetic equations of the Cu-Ni-Si alloy at 450 ◦C and 500 ◦C showed good agreement with the experimental results. Furthermore, it can be seen from the precipitation kinetic curve the addition of the Co element accelerates precipitation in the aging process.


Introduction
The Cu-Ni-Si alloy is a typical high-strength and age-hardening alloy that has been successfully applied to various industries. It has been widely used and developed since Dr. Corson had found that it has a precipitation strengthening effect at the beginning of the 20th century [1][2][3][4][5]. The development of alloys such as C64710, C64730, C70250, and C70350 has further promoted the development of Cu-Ni-Si alloy in recent years [6][7][8]. This series of alloys has attracted much attention due to its high strength, good electrical conductivity, and stress relaxation resistance. Some of these alloys have even been investigated to replace beryllium copper [9][10][11].
The strengthening mechanism of Cu-Ni-Si alloys includes solid solution strengthening, aging strengthening, and work hardening [12][13][14][15]. Since the strengthening power of this alloy is mainly due to the fine, dispersed Ni 2 Si that precipitated during the aging treatment, the strength and electrical conductivity are closely related to the distribution, number, and size of the precipitated phases. Deformation can further increase the strength of the alloy, but, at the same time, it can also slightly decrease electrical conductivity. So far, the research on Cu-Ni-Si alloys is limited to the ternary alloy system. In other words, there are few research studies focused on the influence of the fourth trace alloying element addition [16][17][18]. Lei Jingguo et al. [19][20][21] studied the effect of Cr addition on mechanical properties and electrical properties of Cu-Ni-Si alloy and showed that the addition of a trace amount of Cr can produce a dual alloy aging strengthening effect, which can increase the hardness and strength of the Cu-Ni-Si alloy while ensuring that the conductivity is not greatly affected. Wang Li et al. [22,23] added 0.03% P to the Cu-3Ni-0.6Si alloy for comparison. It was found that the strength property was improved while the conductivity was sacrificed significantly. In this paper, a small amount of Co is added to the Cu-Ni-Si alloy to study its influence on the early precipitation phase during the aging treatment. The objective is to distinguish the phase transformation in the Cu-Ni-Co-Si alloy during aging at a wide aging temperature range.

Experimental
High-purity electrolytic copper, nickel ingots, cobalt ingots, and silicon ingots were melted in a 50 kg medium frequency induction furnace and cast at 1100 • C to 1200 • C. The four-component alloys shown in Table 1 were obtained. The alloy ingot obtained after casting was hot-rolled in multiple passes into a plate with a thickness of 4 mm. Then, the SX2-18-16T box resistance furnace produced by Shanghai Shiyan Electric Furnace Co., Ltd. was used for solution and aging treatment. The solid solution was taken into the temperatures of 950 • C for 1.5 hours, which was followed by water quenching at room temperature. The aging temperatures were 450 • C, 500 • C, and 550 • C. The different isothermal cross-section phase diagrams have been optimized via the CALPHAD approach using Pandat software and the version of the software is CT-1067. Electrical conductivity was carried out on a 7501 eddy current conductivity meter. A transmission electron microscope (TEM) observation of the aged samples was carried out by a JEM-2000 transmission electron microscopy (200 kV) and selected area electron diffraction analysis was carried out to determine the type and structure of the precipitated phase and the orientation relationship between the precipitated phase and the matrix. The transmission sample was mechanically ground to 100 µm, then punched into a small piece of Φ3 mm, and then thinned to about 50 µm. Lastly, a thin zone was obtained by double-spray electrolytic thinning and the electrolyte ratio was 25% HNO 3 + 75% CH 3 OH. The electropolishing temperature was −30 • C and the voltage was 15~20 V.  Figure 1 shows the calculation results of different isothermal cross-section phase diagrams in the copper-rich region of the Cu-Ni-Si ternary alloy. With the increased temperatures, the Fcc + Ni 2 Si phase region gradually get wider. There are Fcc + NiSi, Fcc + Ni 3 Si 2 , Fcc + Ni 2 Si, Fcc + Ni 3 Si, and Fcc + Ni 5 Si 2 two-phase regions and three-phase regions such as Fcc + Ni 3 Si + NiSi, Fcc + Ni 2 Si + Ni 3 Si 2 , Fcc + Ni 2 Si + Ni 5 Si 2 , Fcc + Ni 3 Si 2 +NiSi in the copper-rich section of the Cu-Ni-Si ternary alloy aging at 450 • C, 500 • C, and 550 • C. Therefore, it can be inferred that the Cu-Ni-Si alloy with these components in these regions may have NiSi, Ni 2 Si, Ni 3 Si, Ni 3 Si 2 , and Ni 5 Si phases as the composition and aging temperature of the alloy change, which explains the reason for the formation of so many types of precipitation during the aging treatment of Cu-Ni-Si alloys.  Figure 2 shows the calculated Cu-Ni-Co-Si quaternion phase diagram with the variation of Ni and Co contents when the Si content is fixed at 0.6wt.%. It can be seen that, when the content of Ni and Co is less than 5 wt.%, the precipitated phase mainly includes CoSi, Co2Si, Ni2Si, and Ni5Si2.  Figure 3 shows the influence of the temperature on the mass fraction of the equilibrium phases. The isothermal sections of the alloy at 450 °C, 500 °C, and 550 °C are located between the two phase regions of Fcc + Ni2Si and Fcc + Co2Si. The theoretical calculations indicate that the Ni2Si and Co2Si are the main precipitates and, as the Co content increases, the proportion of the Co2Si phase also increases. The above suggests that the addition of Co would combine with some Si atoms to form the Co2Si phase.   Figure 2 shows the calculated Cu-Ni-Co-Si quaternion phase diagram with the variation of Ni and Co contents when the Si content is fixed at 0.6 wt.%. It can be seen that, when the content of Ni and Co is less than 5 wt.%, the precipitated phase mainly includes CoSi, Co 2 Si, Ni 2 Si, and Ni 5 Si 2 .  Figure 2 shows the calculated Cu-Ni-Co-Si quaternion phase diagram with the variation of Ni and Co contents when the Si content is fixed at 0.6wt.%. It can be seen that, when the content of Ni and Co is less than 5 wt.%, the precipitated phase mainly includes CoSi, Co2Si, Ni2Si, and Ni5Si2.  Figure 3 shows the influence of the temperature on the mass fraction of the equilibrium phases. The isothermal sections of the alloy at 450 °C, 500 °C, and 550 °C are located between the two phase regions of Fcc + Ni2Si and Fcc + Co2Si. The theoretical calculations indicate that the Ni2Si and Co2Si are the main precipitates and, as the Co content increases, the proportion of the Co2Si phase also increases. The above suggests that the addition of Co would combine with some Si atoms to form the Co2Si phase.   Figure 3 shows the influence of the temperature on the mass fraction of the equilibrium phases. The isothermal sections of the alloy at 450 • C, 500 • C, and 550 • C are located between the two phase regions of Fcc + Ni 2 Si and Fcc + Co 2 Si. The theoretical calculations indicate that the Ni 2 Si and Co 2 Si are the main precipitates and, as the Co content increases, the proportion of the Co 2 Si phase also increases. The above suggests that the addition of Co would combine with some Si atoms to form the Co 2 Si phase.  Figure 2 shows the calculated Cu-Ni-Co-Si quaternion phase diagram with the variation of Ni and Co contents when the Si content is fixed at 0.6wt.%. It can be seen that, when the content of Ni and Co is less than 5 wt.%, the precipitated phase mainly includes CoSi, Co2Si, Ni2Si, and Ni5Si2.  Figure 3 shows the influence of the temperature on the mass fraction of the equilibrium phases. The isothermal sections of the alloy at 450 °C, 500 °C, and 550 °C are located between the two phase regions of Fcc + Ni2Si and Fcc + Co2Si. The theoretical calculations indicate that the Ni2Si and Co2Si are the main precipitates and, as the Co content increases, the proportion of the Co2Si phase also increases. The above suggests that the addition of Co would combine with some Si atoms to form the Co2Si phase.   Figure 4 is the TEM bright field images and selected area electron diffraction pattern (SADP) of the S1 alloy after aging at 500 • C for 1 hour. The spots have a modulated structure, which produced satellite reflections. Aged for 1 hour, since spinodal decomposition happened in the supersaturated solid solution, in the following process of aging, the affinity of Ni to Si was greater. Therefore, the process of the ordering was stronger and the super-lattice diffraction peak appeared. The super-lattice diffraction spots are between (000) m and (220) m in Figure 4c (at the position of ( 1 2 10) m ), which means that the ordering occurred at the early stage of the aging process. From Figure 4c, it can be distinguished that the type of the ordering is DO 22 ordering because the typical ( 1 2 10) m is unique to DO 22 ordering [24]. A small amount of secondary precipitated particles with a size of approximately 5 nm can also be observed in the matrix. From the SADP in Figure 4b, the crystal structure of the precipitates was orthorhombic and consists of the δ-Ni 2 Si phase. From the analysis in Figure 4b,c, the crystal orientations between the copper based matrix and precipitates was determined as:  Figure 4 is the TEM bright field images and selected area electron diffraction pattern (SADP) of the S1 alloy after aging at 500 °C for 1 hour. The spots have a modulated structure, which produced satellite reflections. Aged for 1 hour, since spinodal decomposition happened in the supersaturated solid solution, in the following process of aging, the affinity of Ni to Si was greater. Therefore, the process of the ordering was stronger and the super-lattice diffraction peak appeared. The superlattice diffraction spots are between (000)m and (220)m in Figure 4c (at the position of ( 1 2 10)m), which means that the ordering occurred at the early stage of the aging process. From Figure 4c, it can be distinguished that the type of the ordering is DO22 ordering because the typical ( 1 2 10)m is unique to DO22 ordering [24]. A small amount of secondary precipitated particles with a size of approximately 5 nm can also be observed in the matrix. From the SADP in Figure 4b, the crystal structure of the precipitates was orthorhombic and consists of the δ-Ni2Si phase. From the analysis in Figure 4b and Figure   TEM images and the electron diffraction patterns of selected areas and the calibration results of the S2, S3, and S4 alloy aged at 500 °C for 1 hour are shown in Figure 5. We can see that, compared with the traditional ternary Cu-Ni-Si alloy, more precipitates are formed in the early stage of the aging process in the Cu-Ni-Co-Si alloy part of the Ni atoms of which are replaced by Co addition. There are many dispersed precipitates in the matrix that are about 10 nm in size and are slightly larger than the precipitates of the S1 alloy. These phases are completely coherent with the matrix, which indicates that the movement of dislocations would be hindered by this type of precipitate. However, no satellite spots or ordered spots were found in the electron diffraction spots in the selected area at this stage. Therefore, we can speculate that the addition of Co may delay the formation of spinodal decomposition while the Ni element slightly accelerates the spinodal decomposition. According to the literature, this is because Co has a small solid solubility in Cu and easily combines with vacancies, which inhibits the movement of vacancies required for the formation of spinodal decomposition. The binding energy of Co and Si to form the second phase is larger than the binding energy of Ni and Si. In the early stage of aging of the Cu-Ni-Co-Si alloy, no microstructure was generated due to spinodal decomposition. Its corresponding satellite spots were found. TEM images and the electron diffraction patterns of selected areas and the calibration results of the S2, S3, and S4 alloy aged at 500 • C for 1 hour are shown in Figure 5. We can see that, compared with the traditional ternary Cu-Ni-Si alloy, more precipitates are formed in the early stage of the aging process in the Cu-Ni-Co-Si alloy part of the Ni atoms of which are replaced by Co addition. There are many dispersed precipitates in the matrix that are about 10 nm in size and are slightly larger than the precipitates of the S1 alloy. These phases are completely coherent with the matrix, which indicates that the movement of dislocations would be hindered by this type of precipitate. However, no satellite spots or ordered spots were found in the electron diffraction spots in the selected area at this stage. Therefore, we can speculate that the addition of Co may delay the formation of spinodal decomposition while the Ni element slightly accelerates the spinodal decomposition. According to the literature, this is because Co has a small solid solubility in Cu and easily combines with vacancies, which inhibits the movement of vacancies required for the formation of spinodal decomposition. The binding energy of Co and Si to form the second phase is larger than the binding energy of Ni and Si. In the early stage of aging of the Cu-Ni-Co-Si alloy, no microstructure was generated due to spinodal decomposition. Its corresponding satellite spots were found.  Figure 6 shows the transmission electron micrographs and select-area diffraction patterns of the S1 alloy after aging at 500 °C for 4 hours. Many precipitates can be observed in Figure 6a. From the diffraction pattern in Figure 6b, the modulated structure, which caused satellite reflections with the beam direction paralleling to [001]Cu, can be seen. A computer simulation (Figure 6c Figure 6 shows the transmission electron micrographs and select-area diffraction patterns of the S1 alloy after aging at 500 • C for 4 hours. Many precipitates can be observed in Figure 6a. From the diffraction pattern in Figure 6b, the modulated structure, which caused satellite reflections with the beam direction paralleling to [001] Cu , can be seen. A computer simulation (Figure 6c Figure 6 shows the transmission electron micrographs and select-area diffraction patterns of the S1 alloy after aging at 500 °C for 4 hours. Many precipitates can be observed in Figure 6a. From the diffraction pattern in Figure 6b, the modulated structure, which caused satellite reflections with the beam direction paralleling to [001]Cu, can be seen. A computer simulation (Figure 6c

Phase Transformation Kinetics Equation
The electrical conductivity of the alloy in aging mainly depends on the purity of the copper matrix. Because of the sensitivity of electrical conductivity to the precipitation behavior, the kinetics of precipitated phase can be studied by measuring the change of the conductivity and the phase transformation kinetics can be calculated. During aging, the solute atoms precipitate from the solid solution to form the precipitates. Since the solute atoms such as Ni and Si cannot be completely extracted out of the copper matrix, the volume fraction of the precipitated can be expressed as [25]: Vf is the equilibrium volume of the new phase in a unit volume of the matrix when precipitation is over and V is the volume of the new phase formed in a unit volume of the matrix during aging for a certain time t. Clearly, before phase transformation, V = 0 and f = 0. The initial state of electrical conductivity is σ0. After a long aging time, the phase transformation process ended and the electrical conductivity hardly increases and reaches the maximum σmax. So V = Vf and f = 1. According to Matthissen's rule, the electrical conductivity and volume fraction of precipitates has a linear relationship, which can be expressed as Equation (2) below [26].
On the hypothesis of linearity, when transformation is finished, σ = σmax, A = σmax − σ0, and σ can be expressed by the equation below.
The kinetics Avrami equation of phase transformation can be expressed by Equation (4) [27].
In this scenario, f is the volume fraction of precipitates, t is the aging time, and b and n are constants. According to Equation (2) and Equation (4), the electrical conductivity equation can be expressed by the formula below.

Phase Transformation Kinetics Equation
The electrical conductivity of the alloy in aging mainly depends on the purity of the copper matrix. Because of the sensitivity of electrical conductivity to the precipitation behavior, the kinetics of precipitated phase can be studied by measuring the change of the conductivity and the phase transformation kinetics can be calculated. During aging, the solute atoms precipitate from the solid solution to form the precipitates. Since the solute atoms such as Ni and Si cannot be completely extracted out of the copper matrix, the volume fraction of the precipitated can be expressed as [25]: V f is the equilibrium volume of the new phase in a unit volume of the matrix when precipitation is over and V is the volume of the new phase formed in a unit volume of the matrix during aging for a certain time t. Clearly, before phase transformation, V = 0 and f = 0. The initial state of electrical conductivity is σ 0 . After a long aging time, the phase transformation process ended and the electrical conductivity hardly increases and reaches the maximum σ max . So V = V f and f = 1. According to Matthissen's rule, the electrical conductivity and volume fraction of precipitates has a linear relationship, which can be expressed as Equation (2) below [26].
On the hypothesis of linearity, when transformation is finished, σ = σ max , A = σ max − σ 0 , and σ can be expressed by the equation below.
The kinetics Avrami equation of phase transformation can be expressed by Equation (4) [27].
In this scenario, f is the volume fraction of precipitates, t is the aging time, and b and n are constants. According to Equations (2) and (4), the electrical conductivity equation can be expressed by the formula below.
Taking logarithms of both side of Equation (5), the following is calculated.
ln (ln (1/(1 − f))) = lnb + nlnt (6) With the date of Tables 2 and 3, the relationship of ln (ln (1/(1 − f))) and lnt is shown in Figure 8. We can observe that the curve fitting diagram of ln (ln (1/(1 − f))) − lnt shows an approximately straight line. The n is the slope coefficient of the line and the lnt is the intercept.  Taking logarithms of both side of Equation (5), the following is calculated.
With the date of Tables 2 and 3, the relationship of ln (ln (1/(1 − f))) and lnt is shown in Figure  8. We can observe that the curve fitting diagram of ln (ln (1/(1 − f))) − lnt shows an approximately straight line. The n is the slope coefficient of the line and the lnt is the intercept.    Therefore, the phase transformation kinetics equation of the Cu-Ni-Co-Si alloy solid solution aged at 450 • C and 500 • C can be expressed as seen in Table 4.
The phase transformation kinetics curves are shown in Figure 9 and are based on Table 4. It can be seen that the experimentally obtained values agreed with the theoretical analysis. Therefore, the phase transformation kinetics equation of the Cu-Ni-Co-Si alloy solid solution aged at 450 °C and 500 °C can be expressed as seen in Table 4.  The phase transformation kinetics curves are shown in Figure 9 and are based on Table 4. It can be seen that the experimentally obtained values agreed with the theoretical analysis.

Conclusions
(1) There are various combinations of NixSiy phases in the Cu-rich isothermal section of the Cu-Ni-Si ternary alloy. With the increase of temperature, the Ni2Si two-phase region is widened. After the addition of the Co element, a part of the Ni atom is replaced to form a Co2Si phase, which increases as the content of the Co element increases. (2) The addition of the Co element will delay the formation of spinodal decomposition during the aging of the Cu-Ni-Si alloy. After adding the Co element to the Cu-Ni-Si alloy to form the δ-

Conclusions
(1) There are various combinations of Ni x Si y phases in the Cu-rich isothermal section of the Cu-Ni-Si ternary alloy. With the increase of temperature, the Ni 2 Si two-phase region is widened. After the addition of the Co element, a part of the Ni atom is replaced to form a Co 2 Si phase, which increases as the content of the Co element increases. (2) The addition of the Co element will delay the formation of spinodal decomposition during the aging of the Cu-Ni-Si alloy. After adding the Co element to the Cu-Ni-Si alloy to form the δ-Ni 2 Si phase, which has the [001] m //[110] p , (010) m //(001) p orientation relationship with the matrix and will precipitate the (Ni,Co) 2 Si phase, the relationship between the matrix and (Ni,Co) 2 Si