Microstructural, thermal and mechanical properties of Co added Sn–0.7Cu lead-free solder alloy

This study investigated the effects of cobalt microalloying addition on the microstructural features, thermal characteristics and mechanical behavior of eutectic Sn–0.7wt%Cu lead-free solder alloys. The results show that minor cobalt addition of ~ 0.05 wt% causes significant grain refinement of β-Sn, facilitates the formation of fine fibers (Cu,Co)6Sn5 phases and preventing the formation of η′-Cu6Sn5 phases, whereas a large amount of Co (~ 0.5 wt%) additions accumulated in the (Cu,Co)6Sn5 IMCs and clearly changed into coarse fibers. The precipitation strengthening mechanisms of fine fibers (Cu,Co)6Sn5 in the β-Sn matrix increased the ultimate tensile strength (UTS) and Young’s modulus (Y) of the alloy from 30.5 MPa and 15 GPa to 44.6 MPa and 22.3 GPa, respectively, but the ductility decreased from 60 to 45.7%. The coarse fibers (Cu,Co)6Sn5 in eutectic alloys is of interest from not only increased UTS and Y to 38.7 MPa and 16.3 GPa but also maintaining the ductility at the same level, allowing for unique microstructure design. Furthermore, 0.05wt% of Co significantly reduce the onset, eutectic temperatures and undercooling, although pasty rang has been slightly raised, which may enhance the thermal characteristics. This presumably has important implications for the reliability of solders as well as their performance in electronic service.


Introduction
One of the most important challenges in the design and reliability evaluation of the soldered joints is the mechanical characteristics of the solder alloy. Therefore, solder selection is critical when considering the joint quality and reliability performance of the assembly. Over the past few decades, Pb-containing solders have been widely used in soldering electronic applications owing to their acceptable mechanical properties, which include excellent corrosion resistance, higher creep properties, low eutectic temperatures, better wettability and low cost [1][2][3]. However, because Pb is toxic and harmful to environment and human health, the electronic devices containing Pb have been banned by following the WEEE and RoHS directives. After that, the idea of being environmentally friendly and Pb-free spread all over the world. Consequently, the choice of leadfree solder has become an imperative topic of research in the electronics manufacturers to replace the traditional PbSn solder, prompting scholars about worldwide to discover and develop lead-free solder alloy. These efforts have resulted in the development of new generation of Pb-free alloys that have the potential to maintain or improve the reliability of the solders such as Sn-Zn, Sn-Sb, Sn-Ag, Sn-Bi, Sn-Ni-Mn, Sn-Zn-Bi, Sn-Ag-Cu and Sn-Zn-Ag for electronic applications [4][5][6][7][8][9][10][11][12]. Sn-Cu alloys have attracted attention when compared to conventional Sn-Pb solder because of its excellent solderability, low melting point, reasonable cost advantages, good fluidity, and narrow crystallization temperature range [13,14]. Consistent with the phase diagram of Sn-Cu ( Fig. 1 [15]), below 300°C, the Cu 3 Sn and Cu 6 Sn 5 phases are the stable intermetallic phases. However, at 227°C, microstructures of Sn-0.7 wt%Cu are distinguished by b-Sn surrounded by the b-Sn ? g-Cu 6 Sn 5 IMCs. When the melt is cooled further to 186°C, another major reaction occurs: the Cu 6 Sn 5 phase transition from g-Cu 6 Sn 5 [hexagonal (highertemperature)] to g 0 -Cu 6 Sn 5 [monoclinic (low-temperature)] [16][17][18]. One of the most important factors influencing solder joint reliability is the phase transition from g-Cu 6 Sn 5 to g 0 -Cu 6 Sn 5 . Microalloying significantly impacts phase transformation, phase formation, solder alloy microstructural stability, and solder joint reliability. In view of the results of adding alloying elements to the Sn-Cu system, additionally to being significantly soluble in Sn, alloying elements possess the capacity to create new intermetallic (IMCs) within Cu 6 Sn 5 or exist besides Cu 6 Sn 5 , substituting atoms of Cu or Sn [19][20][21][22]. In terms of the soluble nature of mico-alloying elements in Cu 6 Sn 5 IMCs, Laurila et al. [16] classified these elements into two groups: (i) materials that do not dissolve completely in the Cu 6 Sn 5 intermetallic (P, Al, Fe, Ag, Bi, Ti, S and rare-earth elements); and (ii) materials with a pronounced solubility in the Cu 6 Sn 5 intermetallic (Zn, Pd, Pt, Co, Ni, In, Au and Sb). Zeng et al. [23] indicated that Sn-0.7 wt% Cu microstructure was enhanced by addition Zn and Ni at the same time, and the growing of Cu 3 Sn and Cu 6 Sn 5 IMCs were simultaneously reduced. Also, at the solder/Cu boundary, Ni and Zn were uniformly dispersed in Cu 6 Sn 5 IMCs and inhibited the phase change of Cu 6 Sn 5 . Yang et al. [24] reported that by adding trace Al into Sn-0.7 wt% Cu alloy, the Cu 6 Sn 5 IMCs may be converted to Al 2 Cu, preventing dislocation slippage, and increasing tensile strength and creep resistance. Chen et al. [25] reported the effect of Sb additions on the Sn-0.7Cu interface. When the Sb concentration reached 0.5 wt%, the IMCs shape of the solder joints significantly altered and the scallop IMCs were refined. A small amount of Zn was added to enhance the reliability and mechanical properties of the Sn-0.7Cu solder [26]. Co is one of the most attractive alloying elements with several benefits such as reducing undercooling and enhancing the solder's mechanical properties [27]. Some researchers have indicated that Co is used as an additional elements in the adaptation of SnCu alloys, hence, the copper atoms can be replaced by Co atoms, and (Cu,Co) 6 Sn 5 phases can be formed in the primary phase, which can improve the primary phase while preventing the formation of g 0 -Cu 6 Sn 5 phases, that may enhance mechanical reliability of the solder alloy [28][29][30].  Therefore, this investigation aims to shed more light on the effect of minor/trace addition of Co microalloying on the microstructure, thermal properties, and mechanical performance of Sn-0.7Cu solder alloy.
2 Experimental details

Sample preparation
High purity metals of Cu, Co, and Sn (* 99.986%) were used to prepared three different lead-free solder alloys, namely Sn-0.7 wt%Cu (SnCu), Sn-0.7 wt%Cu-0.05 wt%Co (SnCu0.05Co), and Sn-0.7 wt%Cu-0.5 wt%Co (SnCu0.5Co). Chemical composition of the examined solders is tabulated in Table 1. The initial metals were put into a graphite crucible and the melting process was performed in a vacuum arc melting furnace at 700°C, where the liquid alloys were held for 60 min. The molten ingots were stirred with a graphite rod during the melting process to ensure that the metals were fully mixed. After  ensuring a homogeneous mixture, the molten ingots poured into a stainless-steel mold to form a 10-mmdiameter cylindrical rod and then left to cool slowly to room temperature. The cylindrical rods were then mechanically pulled into specimens of 4 cm gauge length and 0.8 mm diameter for mechanical and thermal investigation. Finally, all specimens were heat-treated at 120°C for 50 min to diminish internal stresses [31].

Analysis details
The as-solidified solder alloys were cross-sectioned and polished before being inspected, and then examined by field emission scanning electron microscopy (Quanta 250 FEG high resolution) and optical microscopy (OM, Carl-Zeiss-Jena-JENAPOL). An X-ray diffractometer (XRD, Model Phillips X'pert) operating at 40 kV was used to identify the phase of the various solders. Differential scanning calorimetry (DSC-(STA PT 1000)) was used to determine thermal characteristics of studied alloys. Finally, tensile measurements were carried out by a tensile testing machine at constant temperature 25, 70 and 110°C under a strain rate ranging from 10 -5 to 10 -3 s -1 for all solder compositions to study the effects of Co additions on the mechanical parameters (YS, UTS and EL.) of investigated alloys.

Results and discussion
3.1 Identification of the IMCs XRD analysis was achieved to evaluate and identify the existence of the phase's structures of IMCs in the solder alloys. As shown in Fig. 2a, the principal peaks are indexed by b-Sn, g 0 -Cu 6 Sn 5 , and no additional phases were established in SnCu solder alloy. Meanwhile, new main peak of (Cu, Co) 6 Sn 5 IMCs was detected, as well as b-Sn phases in SnCu0.05Co alloy and SnCu0.5Co alloy (Fig. 2b, c). It was found that adding Co to SnCu solder promotes the formation of (Cu,Co) 6 Sn 5 IMCs, where more Co atoms replace Cu atoms in the lattice and all g 0 -Cu 6 Sn 5 IMCs are transformed into (Cu,Co) 6 Sn 5 IMCs. This result is consistent with other studies in the literature, whereas the Co solubility is very high in the Cu 6 Sn 5 phase but insoluble in b-Sn phase [27]. The appearance of thermodynamic stable IMCs such as (Cu, Co) 6 Sn 5 particles inside the melt is supposed to be the foremost prerequisite for heterogeneous nucleation, which is known to strengthen the solder matrix, increased reliability and improve its mechanical characteristics.

Microstructural characterization
The microstructure evolution of the examined alloys using OM is shown in Fig. 3a, b. The addition of Comicroalloying to SnCu solder alloy resulted in a significant alteration in the morphology of the solder alloy. Firstly, minor Co additions (0.05 wt%) can introduce a fine fibers of (Cu, Co) 6 Sn 5 IMCs formation within the alloy matrix, and also the grain size of b-Sn were refined obviously with random regular texture as shown in Fig. 3a. On the other hand, a small amount of Co (0.5 wt%) addition is capable of completely changing the mode of solidification for SnCu solder alloy, i.e. coarsening of fiber-shaped (Cu, Co) 6 Sn 5 phase and more refine of b-Sn grains as shown in Fig. 3b. This has also been reported by Kim et al., which revealed that the phases of (Cu,Co) 6 Sn 5 is the only stable intermetallic compounds that present in Sn-Cu-Co solder alloy [28]. To further confirm the formation of (Cu,Co) 6 Sn 5 phases, SEM-EDX analysis were performed on SnCu0.05Co and SnCu0.5Co alloys as revealed in Figs. 4, 5 and 6. The results show that the phases have compositions close to (Cu,Co) 6 Sn 5 . However, the phases contain Co, Cu and Sn. In light of the EDS analysis, high volume fraction of Co is present within the (Cu,Co) 6 Sn 5 IMC, which preventing the formation of g 0 -Cu 6 Sn 5 solidphase and thereby preventing mechanical degradation and crack initiation. Elemental mapping characterizations showed that the structures of fine fibershaped in Fig. 5 are (Cu,Co) 6 Sn 5 IMCs, which is much different from the coarse fiber (Cu,Co) 6 Sn 5 IMCs particles revealed in Fig. 6. This new IMC precipitates could act as nucleation sites for b-Sn grains, leading to significant grain refinement as reported by Lauro et al. [32]. Notably, the small grains of b-Sn are beneficial for ductility of SnCu0.5Co alloy. The refinement of b-Sn grains besides the coarsening of (Cu, Co) 6 Sn 5 IMCs are expected to resist dislocation movement in SnCu solder, this called dislocation hardening mechanism. According to theory of heterogeneous nucleation, presence of (Cu,Co) 6 Sn 5 IMCs, prevents the nucleation of g 0 -Cu 6 Sn 5 , increasing nucleation rate and preventing the IMC phase from ripening sequentially. Understanding the microstructural changes can help to explain differences in the thermal characteristic and mechanical properties of solder systems.

Thermal properties of Co-containing Sn-0.7Cu solders
The melting temperature, pasty range, and degree of undercooling are known to be important thermal parameters for the development of new solder alloys and for improving their reliability. Figure 7 shows a thermal profile obtained during the heating and cooling of SnCu, SnCu0.05Co and SnCu0.5Co solder alloys, respectively. The peak temperature, starting/ end points of each reaction, pasty range and the amount of undercooling were determined and summarized in Tables 2 and 3. It is noticeable that only a single endothermic peak (T m ) was detected at 229.2°C for Co-free Sn-0.7Cu solder, while the solidus temperature (T onset ) was reduced to 227.2°C during the heating process. Notably, the T onset and T m have positive relationship with Co content as revealed in Table 2. The addition of 0.05 wt% Co reduces T onset and T m to 222.6°C and 226.3°C respectively. Owing to the reaction between Sn, Cu and Co, the fine fibershaped of (Cu, Co) 6 Sn 5 IMCs has been formed. The solidification of Cu 6 Sn 5 phase due to Co addition can be written as: Co ? Cu 6 Sn 5 ? (Cu,Co) 6 Sn 5 . However, adding 0.5 wt% Co could markedly increase both T onset and T m of SC solder to 228°C and 231°C. The change in T onset and T m of SnCu0.5Co solder can be attributed to the effect of (Cu, Co) 6 Sn 5 IMCs on the rate of solidification, which could modify the surface instability and physical properties of grain boundary/interfacial characteristics. The pasty range of solders obtained from the DSC curve calculated as DT = T L -T S , where T L and Ts are the liquidus and solidus temperature during the heating process. The pasty range of SnCu, SnCu0.05Co and SnCu0.5Co, respectively, is 7.3, 9.2 and 6.6°C, which were smaller than the pasty range of Sn-37Pb solder alloy (11.5°C) [33]. This narrow pasty range facilitates the rapid soldering and reduces the possibility of fillet lifting and porosity. The degree of undercooling is critical in solidification process of solder alloys. The primary goal of undercooling is to control the degree of growth of IMCs and other alloy phases. The undercooling in DSC is calculated as the difference between the onsets temperatures in the heating and cooling curve (DT = T onset , heating -T onset , cooling ) and is related to the difficulty of nucleating the solid phase in a liquid state [29]. As shown in Fig. 7 and Table 3, undercooling of Sn-0.7Cu alloy is noticeably reduced from 8.6 to 6.8°C but increased to 12.2°C with 0.05Co and 0.5Co additions, respectively. This means that the solidification processes of Co-containing Sn-0.7Cu alloys were initiated with the nucleation of new (Cu,Co) 6 Sn 5 IMC particles. These particles may be the primary phases that form upon cooling, and could act as an additional nucleation sites, enhancing the rate of nucleation, leading to decreased undercooling effect. Hence, the 0.05% Co addition has been assumed to suppress the undercooling by inserting potent heterogeneous nucleant particles of fine fiber of (Cu,Co) 6 Sn 5 . Since the undercooling is associated with the complication of nucleating and growing solid phases within liquid phase. The lower degree of undercooling at 6.8°C for Sn-0.7Cu-0.05%Co alloy indicates rapid solidification and faster nucleation rate and this will further diminish the time allowed for g 9 -Cu 6 Sn 5 IMCs formation to grow. With increasing Co content, the degree of undercooling is slightly increased from 8.6 to 12.2°C. That is to say, the primary coarse fiber (Cu,Co) 6 Sn 5 IMCs had a sufficient solidification time to nucleate and grow, that eventually appear as enlargements thanks to the modification in the liquid density levels with 0.5% Comicroalloying. In other words, if higher nucleation is induced, the rate of solidification of b-Sn will be faster and this will further reduce the time allowed for the g 9 -Cu 6 Sn 5 IMCs formation to grow in the Sn-0.7Cu alloys. Instead, the (Cu,Co) 6 Sn 5 phase has been formed.

Mechanical properties of Co-containing
Sn-0.7Cu solders

General tensile tests
Tensile stress-strain testing was performed at constant strain rate of 3.1 9 10 -3 S -1 as shown in Fig. 8. The average UTSs of the investigated samples were 30.5, 44.6, and 38.7 MPa for SnCu, SnCu0.05Co, and SnCu0.5Co, respectively, while the corresponding elongation ranges were 60, 45.7, and 53%, respectively. It was found that as more Co element is added, the elongation was decreased (see Fig. 8a), whereas the UTS increased. Generally, microstructural changes and formation new IMCs inside the matrix alloy could increase the strength and decrease the ductility [34]. Microstructure observations from (SEM) and helpful of EDX and XRD showed refinement of size and distribution of b-Sn grains beside the large amount of homogenous fine fibers (Cu,Co) 6 Sn 5 . These fine fibers IMCs have a strengthening mechanism and may behave as pinning locations, inhibiting the mobility of dislocations that concentrate near grain boundaries, resulting in higher tensile strength which results in increased reliability of solder alloy.

Effect of strain rate on tensile parameters
In tensile tests, the mechanical properties of solder alloy show a considerable strain rate sensitivity and strong temperature dependence [35]. Figure 9 depicts the stress-strain curves of the investigated alloys at   dependent. The UTS of Sn-0.7Cu solder alloy clearly increased with rising strain rate, from 15.8 to 30.6 MPa, from 24.7 to 44.5 MPa for SnCu0.05Co, and from 22 to 38.6 MPa for Sn-0.7Cu-0. 5Co. Dislocation reproduction and dislocation-interaction [30] can shed light on this behavior; hence strain rate is directly related to the dislocation density [36]. As the strain rate continues to increase, the alloy solder deforms, the density of its dislocations grows, and its movement of dislocations is impeded dramatically. As a result, dislocation intersections arise, which may increase the mechanical properties of the solder [37]. Generally, the correlation between UTS and strain rate is described in the following [38,39]: where r denotes the tensile strength, A is a constant that indicates the stress of materials, e is the strain rate and m is sensitivity of strain rate. Figure 10 illustrates the mechanical parameters (UTS, YS and El.) dependence of strain rate for the alloys under investigated. As demonstrated, 0.05 wt% Co microalloying and strain rate improve UTS and YS, but 0.5 wt% Co improves ductility when compared to SnCu0.05Co. The microstructural changes seen in Fig. 4 reflect such changes in ductility and UTS. The SnCu0.05Co alloy showed uniform microstructure with fine fibers (Cu,Co) 6 Sn 5 phases distributed in the b-Sn matrix due to 0.05wt%Co addition. The fine fibers shaped (Cu, Co) 6 Sn 5 IMCs may promote tensile strength. However, the addition of 0.05wt%Co enhanced the ductility of Sn-0.7Cu solder alloy. The coarse fiber-shaped (Cu, Co) 6 Sn 5 IMCs and the presence of b-Sn phases is believed to contribute to the ductility of the solder.

Constitutive equations
The mechanism of tensile deformation investigated with several parameters, such as activation energy (Q) and stress exponent (n), which is related by the Garofalo law [34,37] as: where A is a material parameter, a is the stress level parameter, T is the temperature in kelvin, and R is the b Fig. 10 Effect of strain rate on a yield stress (YS), b ultimate tensile strength (UTS), and c elongation (El. %) at T = 25°C for Sn-0.7Cu, Sn-0.7Cu-0.05Co and Sn-0.7Cu-0.5Co alloys at 25°C and different strain rates The value of a can obtained from the relation a = b/n 1 , where b and n 1 are the average slopes of ln (e)-r and ln (e)-ln(r) respectively, at constant temperature, as revealed in Figs. 11 and 12. These parameters are listed in Table 4. The stress exponent values 'n' of the entire solders are determined from the slope of the ln e against ln[sinh(ar)] in range of 25°C -110°C (Fig. 13), however the Q values can be approximately calculated as the slope of ln[sinh(ar)] against 1/T (Fig. 14). It can be seen that the values of n decreases from 7.6 to 5.2, 7.9 to 5.9 and 7.8 to 5.7 for SnCu, SnCu0.05Co and SnCu0.5Co solder alloys, respectively, as the temperature rises from 25 up to 110°C. The decreases in n value with increasing temperature indicate that the microstructure is unstable at high deformation temperatures. The experimental values of n have frequently been used to determine the mechanism governing the deformation process. In general, dislocation mechanism happens as a consequence of pipe diffusion (n values around 1), grain boundary sliding (n close to 2), viscous glide (n = 3), dislocation climb (n C 4) and strengthening mechanism (n = 7) [38]. In the current study, the fine microstructure and distributed IMCs in low Co-content solder generated a stress exponent of about 7 due to the dispersion strengthening mechanism. These IMCs particles could act as obstacles which restrict the motion of dislocation, as is common in precipitation-strengthened materials. Additionally, the considerable increase UTS and the material parameters value appears to be caused by further homogenous distributions of the eutectic mixture Table 5 shows that the Q values were found to be 47.8, 75.6 and 66.8 kJ/mol for SnCu, SnCu0.05Co and SnCu0.5Co solder alloys, respectively (Table 6). According to these values, it is suggested that the deformation mechanism is dislocation climb controlled by short-circuit pipe diffusion mechanism [40].

Conclusions
The effects of microalloying Co addition on the microstructure formation, thermal properties, and mechanical behavior of Sn-0.7wt% Cu solder alloy were examined in this study. The findings are described below: (1) By microalloying with 0.05% Co to Sn-0.7wt% Cu results in the synthesis of new fine fiberslike morphology (Cu,Co) 6 Sn 5 IMCs and a reduction in the amount of b-Sn phase, whereas 0.5% Co results in the formation of coarse fibers-like morphology (Cu,Co) 6 Sn 5 phases. (2) It was discovered that the percentage of Co added to Sn-0.7wt% Cu controls the melt crystallization of solder alloy. A small amount of Co (0.05wt%) has a significant influence on the pasty range, eutectic temperature, and undercooling of Sn-0.7Cu solder alloy, while a large amount of Co (0.5wt%) has unfavorable effects, resulting in manufacturing issues for saving energy.   mol -1 with Co additions, which propose that the dominant deformation mechanism is dislocation climb governed by short-circuit pipe diffusion mechanism in both solders.

Data availability
Raw data were generated at Physics department-Faculty of Science-Zagazig University. Derived data supporting the findings of this study are available from the corresponding author upon request.

Declarations
Conflict of interest They have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Research involving human and animal rights This article does not contain any studies involving human and animals participants performed by any of the authors.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licen ses/by/4.0/.