Kinetic role of Cu content in reaction process, behavior and their relationship among Cu-Zr-C system

1. Introduction

Resistance spot welding has been widely used in the manufacturing industry. However, the repeated heating and pressing operations during the welding process would contribute to the deformation and softening of the Cu alloy electrodes [1]. Therefore, the production efficiency and welding quality decreased. One of the solutions is to develop particle-reinforced Cu matrix composites. ZrC possesses high hardness, a high melting point, excellent thermal stability, and superior electrical conductivity [[2], [3], [4]]. Furthermore, ZrC can increase the recrystallization temperature and high-temperature strength of composites [5]. These properties make ZrC a good candidate for acting as the reinforced particles in a Cu matrix for spot welding electrode materials.

As a combination of the self-propagating high-temperature synthesis (SHS) and traditional casting route, a novel SHS-casting technique has recently been developed to prepare composites. In this process, an in situ master alloy (e.g., TiC-Al) is initially synthesized by SHS reaction from a predetermined composition (e.g., Ti-C-Al) and is then added into a metal liquid through stirring casting to fabricate composites [6]. This technique combines the ease and high reaction rate of SHS technology, as well as the high efficiency of the casting process [7,8]. Furthermore, composites have many advantages, including a reduced degradation of reinforcements at elevated temperature, a relatively uniform distribution of particles, and a clean reinforcement/matrix interface [9]. To date, the technique has been widely used to produce Al and Mg matrix composites [6,9,10], but there is a lack of knowledge on Cu matrix composites.

Generally, the strength and ductility of composites increases with a decrease in the size of the reinforcing particles [[11], [12], [13], [14]]. Since the reinforcement originates from the SHS reaction, knowledge about reaction kinetic characteristics is required to control the size of the reinforcements. In recent years, the effect of the kinetic parameters (e.g., the composition of the reactants) on the reaction behavior and the obtained products has been extensively investigated [[15], [16], [17]]. It was found that the adjustment of the kinetic parameters could effectively optimize SHS products. As is well-known, SHS technology is a self-sustaining chemical reaction of powders, which only requires the external ignition energy at the initial stage. In terms of performance, it is the difference in the reaction behavior, which is largely due to the influence of the kinetic parameters on the reaction process, which causes the differences in the final products. Therefore, it is necessary to reveal the relations between the reaction process, reaction behavior and resultant products through the investigation of the kinetic parameters. Unfortunately, detailed studies regarding the above aspects remain rather limited due to the rapid reaction of SHS and the complex reaction process.

In this study, we explore the feasibility of preparing ZrC/Cu composites through the SHS-casting route. The main purpose is to investigate the influence of Cu content on the reaction process, behavior and obtained microstructure in the Cu-Zr-C system to elucidate their relationships. These results could provide insights into the SHS reaction and would be beneficial for improving the production practice of ZrC/Cu composites.

2. Experimental

Commercial Cu (99.9% purity, $\widetilde{0}$.5 μm), Zr (98% purity, <48 μm), and graphite (99% purity, <100 nm) powders were used to synthesize the ZrC-Cu master alloy. It should be noted that graphite generally displays lamellar morphology. Therefore, the particle size of graphite refers to its thickness. The reactant powders were made from Zr and C at the ratio of the stoichiometry of ZrC with the addition of 10-50 wt% Cu. The blended powders were ball-milled in a stainless-steel container by using ZrO₂ balls at a low speed ($\widetilde{5}$0 rpm) for 10 h. Subsequently, the mixtures were uniaxially pressed to form compacts with an approximate 65% theoretical density in a cylindrical die (20 mm in diameter). SHS experiments were performed in a stainless-steel glove box under a 99.999% Ar atmosphere, as described in a previous work [16]. The compact was then placed on a thin graphite plate and ignited from the bottom by an arc heating source with a strong current of 60 A. The temperature in the center of the compacts was measured by a W-5%Re/W-26%Re thermocouple, and the signals were recorded and processed by a data acquisition system. The combustion wave velocity was approximately estimated by recording the entire combustion event. The obtained SHS products were crushed into powders and then analyzed by X-ray diffraction (XRD) (Model D8 Advance, Bruker, Germany). Fractography of the SHS samples were observed under a field emission scanning electron microscope (SEM, Model S-4800, Hitachi, Japan) equipped with an energy dispersive X-ray spectrometer (EDS, Model Link-ISIS, England).

The reaction process of the Cu-Zr-C system was studied by a differential scanning calorimeter (DSC, Model 200 F3 Maia, Germany). A small amount of loose packed powder (∼40 mg) was held in an alumina crucible and heated to a defined temperature. The heating was conducted in an argon flow gas (flow rate: 50 ml/min) at a heating rate of 30 °C/min. The DSC experiment was carried out twice for each sample condition, and the obtained DSC results were consistent and repeatable. Phase constituents and microstructures of the DSC products were examined by XRD and SEM, respectively.

In a vacuum arc-melting furnace equipped with a magnetic stirring device (Model DHL-1250, China), ZrC/Cu composites were fabricated through dispersing a ZrC-Cu master alloy into a molten oxygen-free Cu liquid. To accelerate the dispersion of the master alloy, the SHS products were broken into debris by a hammer. The melting current and stirring electric current were approximately 200 A and 10 A, respectively. After being melted six times, the button-like ingot with a diameter at approximately 20 mm was obtained. These button-like samples were cross-sectioned, ground, polished and etched by a nitric acid solution (HNO₃ 20 ml + deionized water 60 ml) for 20 min. Then, the microstructures were examined by SEM.

3. Results and discussion

3.1. Effect of Cu content on the reaction process

3.1.1. Effect of Cu content on the reaction process in Cu-Zr systems

Numerous studies showed that the reaction mechanism of a given system (e.g., Cu-Ti-C) in DSC and SHS seemed to be similar [[18], [19], [20], [21]]. To unveil the influence of Cu content on the reaction sequence, DSC was applied to investigate the heating process of a Cu-Zr-C system with 20 wt%, 30 wt% and 50 wt% Cu contents, in which the corresponding molar ratios of Cu:Zr are 1:2.5, 1:1.5 and 1:0.69, respectively. According to the Cu-Zr binary phase diagram, there are many copper zirconium compounds [22]. Considering the complex reactions in the Cu-Zr-C system, DSC experiments were first performed on Cu-2.5Zr, Cu-1.5Zr and Cu-0.69Zr systems.

Fig. 1, Fig. 2 illustrate the DSC curves and XRD patterns, respectively, of the DSC products in Cu-0.69Zr, Cu-1.5Zr and Cu-2.5Zr systems. As shown in Fig. 1, an exothermic peak appeared near 650 °C in Cu-0.69Zr. In addition to Cu and α-Zr, CuZr₂, Cu₁₀Zr₇ and Cu₅₁Zr₁₄ were detected in the sample quenched from 650 °C (Fig. 2(a)). Moreover, the diffusion bonding between spherical Cu particles and irregular Zr particles was found in the corresponding microstructure (Fig. 3(a)). The above results imply that the exothermic peak at 650 °C should be the Cu-Zr solid-state reaction and the transition layer formation of Cu_xZr_y compounds at the Cu/Zr interface, i.e., Cu-Cu₅₁Zr₁₄-Cu₁₀Zr₇-CuZr₂-Zr. In contrast, the formation temperature of these Cu_xZr_y compounds in Cu-1.5Zr and Cu-2.5Zr systems reduced to 616 °C and 599 °C as shown in the DSC curves (Fig. 1) and XRD patterns (Fig. 2(b) and (c)). Since the Cu particle (0.5 μm) is notably smaller than the Zr particle (45 μm), the increasing Zr content enhances the number of Zr particles per unit volume, and thus increases the contact area between Cu and Zr. Therefore, Cu_xZr_y compounds were synthesized at relatively low temperatures. However, low temperature reduces the diffusion rate. As a result, the resultant Cu_xZr_y compounds in the Cu-0.69Zr system at 650 °C are the most, while those in Cu-2.5Zr system at 599 °C are the least.

View original graphic| Download| PPT slide

Fig. 1.   DSC curves of (a) Cu-0.69Zr system, (b) Cu-1.5Zr system and (c) Cu-2.5Zr system.

View original graphic| Download| PPT slide

Fig. 2.   XRD patterns for DSC products of (a) Cu-0.69Zr system, (b) Cu-1.5Zr system and (c) Cu-2.5Zr system quenched at different temperatures.

View original graphic| Download| PPT slide

Fig. 3.   Microstructures for DSC products of Cu-0.69Zr system quenched at different temperatures (a) 650 °C, (b) 910 °C, (c) 983 °C and (d) 1200 °C, respectively.

Further increasing the temperature in the DSC experiment revealed an endothermic peak at approximately 900 °C in the Cu-0.69Zr and Cu-1.5Zr systems, corresponding to the melting of Cu₁₀Zr₇ (melting point: 895 °C). By contrast, there is no obvious endothermic phenomenon in the Cu-2.5Zr system, possibly due to the small amount of Cu₁₀Zr₇ synthesized by the prior Cu-Zr reaction.

The presence of a Cu-Zr liquid could enhance the packing density of the powders and the contact area between the particles (Fig. 3(b)). This effect is beneficial for the further synthesis of Cu₅₁Zr₁₄ and Cu₈Zr₃ during the solid-state Cu-Zr reaction. Furthermore, with the wetting and spreading of the Cu-Zr liquid, sub-μm Cu particles can quickly dissolve into the Cu-Zr liquid compared with Zr particles, which increases the Cu concentration in the Cu-Zr liquid. Hence, in the DSC spectrum of Cu-0.69Zr quenched from 910 °C, a new Cu₈Zr₃ phase was detected, and the peak intensities of Cu₅₁Zr₁₄ and Cu₁₀Zr₇ increased while those of Cu and α-Zr decreased (Fig. 2(a)). As the content of Zr increases in the Cu-Zr system, the Cu-Zr solid-solid reaction tends to form Cu_xZr_y compounds with high Zr content and the Zr-content in the liquid increases. As a result, in the DSC spectra of Cu-1.5Zr quenched from 909 °C and Cu-2.5Zr quenched from 931 °C, the diffraction peak intensity for CuZr₂ increased significantly, while those of Cu₈Zr₃ and Cu₁₀Zr₇ decreased (Fig. 2(b) and (c)).

With a further increase in temperature, the effect of Cu content on the reaction process is more distinct. As shown in Fig. 1, there is an exothermic peak at 983 °C and two endothermic ones at 957 °C and 1021 °C in a Cu-0.69Zr system. In contrast, only endothermic peaks near 949 °C and 1024 °C emerge in Cu-1.5Zr and Cu-2.5Zr systems, respectively.

For Cu-0.69Zr, the continuous heating leads to the dissolution of CuZr₂ and Cu₈Zr₃ into Cu-Zr liquid, corresponding to the endothermic peak at 957 °C in Fig. 1. Therefore, an increased amount of the liquid phase was obtained, which further enhanced the packing density of powder mixtures. Subsequently, the residual Cu reacted with Zr to form Cu₅₁Zr₁₄, which corresponded to the exothermic peak near 983 °C (Fig. 1). As a result, Cu and CuZr₂ disappeared in the sample quenched from 983 °C. In addition, the peak intensities of Zr and Cu₈Zr₃ decreased, while that of Cu₅₁Zr₁₄ increased, as shown in Fig. 2(a). Another endothermic peak emerged at 1021 °C, which came from the dissolution of Cu₅₁Zr₁₄ and residual Zr into the Cu-Zr liquid. The dissolution of Zr may dilute the Cu concentration in the liquid and lead to the crystallization of Cu₈Zr₃ and Cu₁₀Zr₇ with less Cu atoms during cooling process. As a consequence, in Cu-0.69Zr quenched from 1200 °C, Zr was not observed by XRD, and the peak intensity of the Cu-rich phase Cu₅₁Zr₁₄ decreased while those of Cu₈Zr₃ and Cu₁₀Zr₇ increased (Fig. 2(a)). Fig. 3(c)-(f) shows the microstructures and corresponding EDS results for Cu-0.69Zr systems quenched from 983 °C and 1200 °C. With an increased degree of dissolution of the residual solid phase (e.g., Zr) into the liquid, the porosity level in the DSC sample and the Cu ratio in the solidification structures decreased, which agreed with the DSC and XRD results.

For Cu-1.5Zr, an endothermic peak at 949 °C corresponded to the dissolution of CuZr₂ and Cu₈Zr₃. However, the Cu-1.5Zr system did not display the exothermic event at 983 °C and endothermic phenomenon at 1021 °C seen in the Cu-0.69Zr system (Fig. 1). The disappearance of the exothermic peak (983 °C) may be observed because the free Cu has been exhausted by the Cu-Zr reaction at 909 °C (Fig. 2(b)). The disappearance of the endothermic peak implied that the further dissolution of the Zr-rich phases (e.g., Zr) into the Cu-Zr liquid was restricted due to the high Zr concentration in the liquid. Fig. 2(b) provides direct evidence that Zr existed in Cu-1.5Zr at the high temperature of 1200 °C.

For Cu-2.5Zr, the Zr-content is the highest in this work. At 949 °C, CuZr₂ did not dissolve into the Cu-Zr liquid. When the temperature was increased to 1000 °C, CuZr₂ began to melt, which contributed to an endothermic peak at 1024 °C. Moreover, the dissolution of Zr into the Zr-rich liquid was limited, which is similar to what was seen with Cu-1.5Zr. Hence, in Cu-2.5Zr quenched from 1200 °C, Zr was detected according to Fig. 2(c).

Based on the results described above, decreasing the Cu content in the Cu-Zr system reduced the amounts of Cu₅₁Zr₁₄, Cu₁₀Zr₇ and CuZr₂ synthesized in the early solid diffusion stage. During the heating process, the amount of Cu-Zr liquid originating from the melting of Cu₁₀Zr₇ reduced, and the residual Cu reacted with Zr to fabricate Cu_xZr_y compounds with high Zr-content. These phenomena restricted the following diffusion of Zr-rich phases into the Cu-Zr liquid and decreased the amount of Cu-Zr liquid at elevated temperatures.

3.1.2. Effect of Cu content on the reaction process in Cu-Zr-C systems

Fig. 4, Fig. 5 present the DSC curves and XRD patterns in 20 wt% Cu-Zr-C (20 wt% Cu content), 30 wt% Cu-Zr-C (30 wt% Cu content) and 50 wt% Cu-Zr-C (50 wt% Cu content) systems, respectively. When a 50 wt% Cu-Zr-C system was heated to 661 °C, a weak exothermic peak appeared. In addition to Cu, Zr and C, new CuZr₂, Cu₁₀Zr₇ and Cu₅₁Zr₁₄ phases were detected in the cooling sample (Fig. 5(a)). Hence, the exothermic peak at 661 °C is the solid diffusion reaction between Cu and α-Zr. With decreasing Cu content, the peak temperature dropped down to 645 °C in the 30 wt% Cu-Zr-C system and 627 °C in the 20 wt% Cu-Zr-C system. In addition, the peak intensities of the Cu_xZr_y compounds were gradually weakened (Fig. 5(b) and (c)), which were consistent with the variation trends of the Cu-Zr system. It should be noted that the addition of C into Cu-Zr mixed powders would decrease the contact area between Cu and Zr and restrain the Cu-Zr reaction. Although Cu_xZr_y compounds were formed at elevated temperature in Cu-Zr-C systems compared with the Cu-Zr systems, the peak intensities of the Cu_xZr_y compounds decreased. This finding is supported by the disappearance of Cu₅₁Zr₁₄ in a 20 wt% Cu-Zr-C system quenched from 627 °C (Fig. 5(c)).

View original graphic| Download| PPT slide

Fig. 4.   DSC curves of (a) 50 wt% Cu-Zr-C system, (b) 30 wt% Cu-Zr-C system and (c) 20 wt% Cu-Zr-C system.

View original graphic| Download| PPT slide

Fig. 5.   XRD patterns for DSC products of (a) 50 wt% Cu-Zr-C system, (b) 30 wt% Cu-Zr-C system and (c) 20 wt% Cu-Zr-C system quenched at different temperatures.

The DSC analysis of the Cu-0.69Zr system revealed that Cu₁₀Zr₇ melted into Cu-Zr liquid at approximate 900 °C according to Fig. 1, which enhanced the contact area between particles and promoted the synthesis of Cu₈Zr₃. On the other hand, it provided the desired condition for the synthesis of ZrC. With the diffusion and dissolution of C into the Cu-Zr melt, a Cu-Zr-C ternary liquid was obtained. ZrC precipitates were formed once the Cu-Zr-C liquid reached saturation according to the microstructure measurements (Fig. 6(a)) and EDS results (Fig. 6(d) and (e)). These two processes contributed to an exothermic peak at 907 °C in Fig. 4. Therefore, in 50 wt% Cu-Zr-C quenched from 907 °C, new ZrC and Cu₈Zr₃ phases were observed (Fig. 5(a)). With decreasing Cu content in the Cu-Zr-C system, the amount of Cu₁₀Zr₇ formed in the prior stage was reduced. This weakens the endothermic effect coming from the melting of Cu₁₀Zr₇, and leads to the disappearance of the endothermic peak near 900 °C in the 30 wt% and 20 wt% Cu-Zr-C systems. Less liquid decreased the dissolution rate of C into Cu-Zr liquid. As a result, the ZrC-formation temperatures were increased to 937 °C in 30 wt% Cu-Zr-C system (Fig. 5(b)) and 990 °C in 20 wt% Cu-Zr-C system (Fig. 5(c)). On the other hand, when Cu content was decreased in Cu-Zr-C, Cu reacted with Zr and formed Zr-rich compounds. Therefore, the amount of Cu₈Zr₃ decreased while the CuZr₂ content increased, which is similar to the Cu-Zr system discussed in Section 3.1.1.

View original graphic| Download| PPT slide

Fig. 6.   Microstructures for DSC products of 50 wt% Cu-Zr-C system quenched at different temperatures (a) 907 °C, (b) and (c) 1250 °C.

With a further increase in the temperature, the difference between the DSC curves of different samples was more distinct. An endothermic peak at 989 °C in 50 wt% Cu-Zr-C system was the dissolution of Cu₈Zr₃. This enhanced the density of the powder mixtures and triggered the synthesis of Cu₅₁Zr₁₄ by a Cu-Zr solid-state reaction in turn. Meanwhile, the increased amount of liquid promoted the dissolution of C into Cu-Zr melt and the formation of ZrC. As a result, a large amount of Cu₅₁Zr₁₄ and ZrC were formed at 1067 °C (Fig. 5(a)). The formation of Cu₅₁Zr₁₄ and ZrC would release heat and increase the temperature surrounding, which improved the flow and solubility of the liquid. Subsequently, the residual Zr and Cu₅₁Zr₁₄ dissolved into the liquid, resulting in an exothermic peak at 1076 °C (Fig. 4). By this means, the amount of liquid increased once more, which further promoted the dissolution of residual C into the liquid. At 1096 °C, the ZrC-forming reaction was triggered once again. The synthesis of ZrC completely consumed the Zr and C atoms in the liquid and thus the free Cu was released. When the 50 wt% Cu-Zr-C sample was heated to 1096 °C, the full conversion of Zr and C to ZrC was achieved, and only ZrC and Cu were detected by XRD (Fig. 5(a)). Fig. 6(b) and (c) show the microstructures of 50 wt% Cu-Zr-C quenched from 1250 °C. Fig. 6(b) provides direct evidence of an increase in the amount of liquid, and Fig. 6(c) demonstrates that some amount of ZrC was formed and precipitated from the liquid.

For 30 wt% Cu-Zr-C, a strong exothermic peak emerged at 1009 °C. In the cooling sample, the peak intensities of Cu₁₀Zr₇, Cu₈Zr₃ and ZrC were enhanced, while that of CuZr₂ was weakened (Fig. 5(b)). The Zr concentration in the melt of the 30 wt% Cu-Zr-C system is higher than those of the 50 wt% Cu-Zr-C system. This property may prevent the rapid dissolution of Cu_xZr_y compounds into the Cu-Zr liquid. As a consequence, CuZr₂ was transformed into Cu-Zr liquid until the temperature was raised to its melting point (1000 °C). Then, C dissolved into the increasing Cu-Zr liquid and reacted with Zr to form ZrC, which contributed to an exothermic peak at 1009 °C. On the other hand, the subsequent precipitation of ZrC improved the Cu concentration of the melt. During the cooling stage, Cu-rich phases Cu₁₀Zr₇ and Cu₈Zr₃ were crystallized out of the liquid. Therefore, in 30 wt% Cu-Zr-C quenched from 1009 °C, the peak intensities of Cu₁₀Zr₇, Cu₈Zr₃ and ZrC increased while that of CuZr₂ decreased. With increasing temperature, the residual Zr continuously dissolved into the Cu-Zr melt, which resulted in an endothermic peak at 1035 °C. In this way, more liquid was obtained, which further promoted the following diffusion of C into liquid. At 1077 °C, an exothermic peak originating from a ZrC-forming reaction appeared (Fig. 4). This finding is supported by the disappearance of Zr and the increase of the ZrC peak intensity of 30 wt% Cu-Zr-C quenched from 1077 °C (Fig. 5(b)). With the dissolution of Cu₅₁Zr₁₄ into the liquid, an endothermic peak emerged at approximately 1088 °C. Subsequently, the residual C dissolved into the melt and reacted with Zr to form ZrC. The complete formation of ZrC contributed to the exothermic event at 1094 °C and released Cu atoms from liquid. Thus, only ZrC and Cu were detected in the sample quenched from 1094 °C (Fig. 5(b)).

For 20 wt% Cu-Zr-C, less Cu-Zr liquid and more CuZr₂ were formed in the prior stage. CuZr₂ was melted into a Cu-Zr liquid at approximately 1000 °C. Subsequently, an exothermic peak emerged at 1002 °C, which resulted from a ZrC-forming reaction. Hence, in 20 wt% Cu-Zr-C quenched from 1002 °C, the peak intensity of ZrC significantly increased while that of CuZr₂ was reduced (Fig. 5(c)). Similar phenomena occurred in the 30 wt% Cu-Zr-C system as well. With increasing temperature, the residual Zr and C continuously dissolved into liquid to synthesize ZrC and led to the exothermic event at 1106 °C. As a result, there were only ZrC and Cu in the sample quenched from 1106 °C (Fig. 5(c)).

In the light of the above results, it is believed that ZrC was formed through the diffusion and dissolution of C into the Cu-Zr liquid. With decreasing Cu content in the Cu-Zr-C system, the amounts of Cu₅₁Zr₁₄, Cu₁₀Zr₇ and CuZr₂ synthesized in the solid diffusion reaction stage were reduced. During the heating process, the amount of Cu-Zr liquid arising from the melting of Cu₁₀Zr₇ decreased, which increased the initial formation temperature of ZrC (904-990 °C). Meanwhile, the residual Cu reacted with Zr to form Cu_xZr_y compounds with high Zr-content. This decreased the subsequent dissolution rate of the Zr-rich phase into the Cu-Zr liquid, and increased the final formation temperature of ZrC (1094-1106 °C). However, decreasing the Cu content accordingly increased the C-content in Cu-Zr-C system. In the intermediate reaction stage, the amount of C dissolving into the Cu-Zr liquid at unit time increased. As a result, the explosive formation temperature of ZrC (1002-1067 °C) decreased.

3.2. Internal relation among reaction process, reaction behavior and reaction products

Fig. 7 displays the variation of the combustion temperature, ignition time and wave velocity in Cu-Zr-C systems. The combustion temperature gradually descends with increasing Cu addition. Referring to the thermodynamic data [23], the change of Gibbs free energy of Zr-C reaction is more negative than those of Cu-Zr reactions. Therefore, the contents of Zr and C in the Cu-Zr-C system decrease, and the heat liberated from the ZrC-forming reaction decreases, with increasing Cu content. This effect leads to a decrease in combustion temperature in turn. In addition, exothermic reactions mainly originate from the synthesis of Cu_xZr_y compounds and ZrC, according to the DSC analysis of Cu-Zr-C samples. In 20 wt%, 30 wt% and 50 wt% Cu-Zr-C systems, the exothermic peak areas are determined at 13,468 J/g, 9838 J/g and 8979 J/g, respectively, by the Netzsch proteus thermal analysis software, which offers circumstantial evidence that the increased Cu content would decrease the heat generated by the SHS reaction.

View original graphic| Download| PPT slide

Fig. 7.   Variations of the combustion temperature, wave velocity and ignition time with Cu content.

The time from the start of heating to the occurrence of the SHS reaction is called the ignition time. It can be used to quantify the difficulty for the reaction to occur. As indicated in Fig. 7, the ignition time first decreases and then increases with an increased Cu content. Similar behaviors existed in the SHS reaction of Co-Zr-B₄C, Ni-Ti-B₄C and Al-Ti-B₄C systems as well [16,24,25]. Referring to references [16,19], the incorporation of metallic powder to the reactants can accelerate the SHS reaction by the prior formation of liquid at low temperatures. DSC results of the Cu-Zr-C samples revealed that with increasing Cu content, the amount of Cu₁₀Zr₇, CuZr₂ and Cu₅₁Zr₁₄ compounds formed by the Cu-Zr solid-state reaction increased. The amount of Cu-Zr liquid formed by the melting of Cu₁₀Zr₇ also increased, which facilitates the dissolution rate of C into the Cu-Zr liquid and thus leads to a fast ignition of the SHS reaction. Therefore, the ignition time in the 20 wt% Cu-Zr-C sample is shorter than that in the 10 wt% Cu-Zr-C sample. However, the ignition time gradually increases as the Cu addition further increases. The change of Gibbs free energy of Zr-C reaction is more negative than those of Cu-Zr reactions. This implies that heat released by the ZrC-forming reaction should be responsible for the ignition of the Cu-Zr-C powder compacts. An increased Cu content dilutes the relative contents of Zr and C in the Cu-Zr-C system. As a result, the amount of C dissolving into the liquid at unit time reduces, and the ignition time gradually increases. The DSC results may provide a direct proof of the ignition time. The initial formation temperatures of some amount of ZrC are 1002 °C, 1009 °C and 1067 °C in the 20 wt%, 30 wt% and 50 wt% Cu-Zr-C mixtures, respectively. One could expect that the heating time required for igniting the Cu-Zr-C powder compacts increases with an increased Cu addition.

The combustion wave velocity refers to the migration rate of the combustion front. The expression of the wave propagation velocity (V) is derived as follows [26]:

v²=f(n) $\frac{kC_{p}}{\rho Q} \frac{RT_{c}^{2}}{E} k_{0} exp\frac{E}{RT}$ (1)

where f(n) is a kinetic function of the order of the reaction. k, ρ, C_p, Q, R, T_c, T, E, and k₀ are thermal conductivity, density, heat capacity, reaction heat, gas constant, combustion temperature, reaction activation energy, the temperature and temperature constant, respectively. According to Formula (1), the combustion wave velocity largely depends on the combustion heat and thermal conductivity of the compact [20,21]. As shown in Fig. 7, the combustion wave velocity gradually decreases with increasing Cu content, displaying a similar tendency with combustion temperatures. This suggests that the combustion temperature plays a more important role in affecting the wave velocity. Based on the DSC results of the Cu-Zr-C system, the variation trend of the combustion wave velocity may be more reasonable. As is well-known, a chemical exothermic reaction sustains the spontaneous propagation of a combustion wave. The ignition for the unreacted region of the compact can be triggered by the heat generated in the adjacent reacting region. The initial formation temperatures of a considerable amount of ZrC are 1002 °C, 1009 °C and 1067 °C in the 20 wt%, 30 wt% and 50 wt% Cu-Zr-C mixtures, respectively. From another perspective, one could expect that the energy required for igniting the neighboring unreacted regions of the compact increases with an increased Cu content. However, the heat generated by the SHS reaction of reacting region decreased with increasing Cu additions, as previously mentioned when discussing combustion temperature. Apparently, these phenomena are not consistent with a propagating of combustion wave.

Fig. 8 presents the XRD spectra of the SHS samples. For Cu content below 40 wt%, the final products were only composed of Cu and ZrC. In the 50 wt% Cu addition sample, Cu_xZr_y and C were detected in addition to Cu and ZrC. Munir shows that the formation of metastable phases is based on the rapid cooling rate and reactions process [27]. According to the analysis of the reaction process and reaction behavior, the synthesis of ZrC is controlled by the diffusion of C into the Cu-Zr liquid. An increased Cu content decreases the combustion temperature. As a result, the diffusion rate of C atoms and the solubility of C in the melt decrease. This may prevent the full conversion of C and Cu_xZr_y into ZrC. On the other hand, the ZrC-forming reaction is a multistage process. Once ignited, the heat released from the reacting region would transfer to the adjacent unreacted region. With increasing Cu contents, the dwell time at the elevated temperature of the reacting region would be shortened due to the good thermal conductivity of Cu and Cu_xZr_y compounds [23]. A short holding time at high temperatures restricts the subsequent diffusion and dissolution of C into the liquid.

View original graphic| Download| PPT slide

Fig. 8.   XRD patterns of SHS products with different Cu content (a) 10 wt%, (b) 20 wt%, (c) 30 wt%, (d) 40 wt% and (e) 50 wt%, respectively.

Fig. 9 shows the typical fractured microstructure of the SHS samples. It shows that Cu content plays an important role in the particle size of ZrC. When the Cu content is between 10 wt% and 30 wt%, the obtained ceramic particles are at a micrometer scale. Upon increasing the Cu content to 40 wt%, the particle size of ZrC is reduced to a submicrometer scale. With further increasing Cu content, residual C can be observed in the 50 wt% Cu-Zr-C sample, which indicates an incomplete SHS reaction. Numerous studies report that the reduction of particle size is mainly derived from a decrease in the combustion temperature and an increase in the cooling rate [2,16,20]. This agrees with the results in the current work. It should be noted that ZrC particles display inhomogeneity in size, which is mainly caused by the multistage formation process. For the synthesis of ZrC particles in the 30 wt% Cu-Zr-C system, there are four formation temperatures including 937 °C, 1009 °C, 1077 and 1094 °C. A small amount of ZrC synthesized at 937 °C tends to grow during the following reaction process. By contrast, there is inadequate time and a lack of Zr and C atoms for the growth of ZrC particles fabricated at 1094 °C. This may be the reason for the inhomogeneity in particle size.

View original graphic| Download| PPT slide

Fig. 9.   Microstructures of SHS products with different Cu content (a) 10 wt%, (b) 20 wt%, (c) 30 wt%, (d) 40 wt% and (e) 50 wt%, respectively.

3.3. Microstructure of ZrC/Cu composites

The SHS sample of the 40 wt% Cu-Zr-C system were used as a master alloy for the fine sized ZrC particles. Fig. 10 illustrates the microstructure of the ZrC/Cu composites. As the ZrC content in the ZrC/Cu composites ranges from 0.1 wt% to 0.3 wt%, the ZrC particles disperse in the Cu matrix. The relatively uniform distribution of ZrC implies that the addition of ZrC-Cu master alloy into the Cu liquid is an effective way to prepare ZrC/Cu composites.

View original graphic| Download| PPT slide

Fig. 10.   Microstructures of ZrC/Cu composites with different ZrC content (a) 0.1 wt%, (b) 0.2 wt% and (c) 0.3 wt%, respectively.

4. Conclusions

In this study, the relationship among reaction process, reaction behavior and resultant products in a Cu-Zr-C system with different Cu contents was investigated. In addition, we attempted to produce ZrC/Cu composites by a combined SHS-casting method. The following conclusions were drawn from this work.

(1) With an increased Cu addition, the heat generated is reduced in the Cu-Zr-C samples, which results in a decrease in the combustion temperature. Meanwhile, the energy required for igniting the neighboring unreacted region increases. These lead to a decrease in the combustion wave velocity. As the formation of ZrC is mainly controlled by the diffusion of C into Cu-Zr liquid, additional Cu content enhances the volume fraction of the Cu-Zr liquid formed at the early stage but decreases the amount of C atoms diffusing and dissolving into the liquid at unit time. Therefore, the ignition time first decreases and later increases with Cu concentrations.

(2) The reaction for the ZrC synthesis is a multistage process, which results in a varied ZrC particle size. In addition, an increased Cu content decreases the combustion temperature and enhances the thermal conductivity of the compact. As a result, the diffusion rate of the C atoms and the solubility of C in the melt decrease, as well as the dwell time of the reacting region at elevated temperature. These contribute to a decline in the ZrC particle size and the incomplete synthesis of ZrC in 50 wt% Cu-Zr-C system.

(3) Cu matrix composites reinforced with submicrometer sized ZrC particles can be prepared through adding an optimized ZrC-Cu master alloy into a Cu liquid.

Acknowledgements

This work is supported by the National Key Research and Development Program (No. 2017YFB0305300), the National Natural Science Foundation of China (Nos. 51404157, 51374144), Public Welfare Projects of Science and Technology Department of Zhejiang Province (Grant No. 2017C31118), the Natural Science Foundation of Zhejiang Province (Grant No. LY17E050003).