alloyed

The mechanical reliability of the future miniaturized interconnects is mainly governed by the intermetallic compounds such as Cu 6 Sn 5 . Alloyed Cu 6 Sn 5 with various elements, including Co and In, have been introduced and attracted attention for different reasons, such as the enhancing mechanical reliability and lowering the bonding temperature. Hence, this work aimed to evaluate the microstructural and mechanical properties of Cu 6 Sn 5 -, Cu 6 (Sn,In) 5 -, (Cu,Co) 6 Sn 5 -, and (Cu,Co) 6 (Sn,In) 5 -interconnects. The grain size, grain orientation, and crystal structure of the pure and alloyed Cu 6 Sn 5 phases were analyzed using electron backscatter diffraction. The results revealed that all the joints contained monoclinic and hexagonal crystal structures arbitrarily formed across the bond-line. Furthermore, the Cu 6 Sn 5 grains exhibited random grain orientation and there was no discernible difference between the pure and alloyed Cu 6 Sn 5 interconnects other than Cu 6 (Sn,In) 5 grains elon-gated along the perpendicular direction to the bonding interface. However, it was found that alloying elements altered the grain sizes. In alloying refined and elongated the Cu 6 Sn 5 grains while the Co alloying enlarged the Cu 6 Sn 5 grains. The mechanical properties of the interconnects were examined using nanoindentation test. The results indicated that the hardness (H) and Young ’ s modulus (E i ) values of Cu 6 Sn 5 is increased with the alloying elements. (Cu,Co) 6 (Sn,In) 5 showed the highest E i /H value which indicates its highest plasticity.


Introduction
Recently, three-dimensional (3D) integration, wafer stacking or chips with vertical interconnections, has been developed to overcome the electronic components scaling limits [1][2][3].To meet the demand for higher integration, miniaturized fine pitch interconnections are necessary.Smaller interconnects would, however, result in a higher volume proportion of intrinsically brittle intermetallic compounds (IMCs) in 3D micro-joints, such as those produced by the Solid Liquid Interdiffusion (SLID) bonding method [1,4,5].If the bumps size is less than 10 μm, the entire micro-joint might only consist of a few IMCs grains [4,5].Therefore, micro-joint reliability in future electronic packages are greatly influenced by the characteristics of the IMCs formed in the bond-line [4,6].Understanding the microstructural and mechanical properties of these IMCs is of utmost importance, as they directly influence the reliability, durability, and performance of micro-joints.By delving into the microstructural characteristics, crystal structure, grain orientation, and mechanical behavior of these IMCs, we can gain valuable insights into their response to external stresses and uncover strategies to enhance the mechanical reliability of future micro-joints.
Cu 6 Sn 5 and Cu 3 Sn are the main IMCs formed in the Sn-based solders/ Cu interconnects [7].Cu 6 Sn 5 based joints are potentially thermomechanically more superior compared to Cu 3 Sn based joints due to reliability challenging Kirkendall void formation and growth associated with Cu 3 Sn formation.Therefore, intensive research [8][9][10][11] has been conducted to stabilize Cu 6 Sn 5 and inhibit the formation of Cu 3 Sn.For example, the addition of Co to the Cu-Sn system stabilizes the Cu 6 Sn 5 phase while preventing the formation of Cu 3 Sn [12][13][14].Nonetheless, Cu-Sn IMC joints are often unsuitable for temperature sensitive components due to their high bonding temperature (T > 250 • C) [15].Cu-Sn-In, for instance, has been demonstrated as an alternative to the Cu-Sn system by lowering the bonding temperature, stabilizing the Cu 6 Sn 5 phase, and preventing the Cu 3 Sn phase formation [15,16].To effectively use Co and In as alloying elements in Cu-Sn systems, a thorough knowledge of the mechanical characteristics of pure Cu 6 Sn 5 interconnects as well as the effects of the alloying elements is required.
Intensive research has been conducted to investigate the mechanical properties (Young's modulus (Ei) and hardness (H)) of the Cu 6 Sn 5 IMC [4,[17][18][19][20][21][22].The Young's modulus and hardness values for Cu 6 Sn 5 varies across sources, the results, however, can be impacted by several variables, including sample preparation, measurement methods, and measurement conditions.Moreover, the Cu 6 Sn 5 has a variety of crystal structures that could each have a unique mechanical characteristic.It has been shown that Cu 6 Sn 5 has at least two crystal structures in the solid state, including a hexagonal structure with space group of P6 3 /mmc and a monoclinic structure with space group of C2/c [1,7].However, Cu 6 Sn 5 -P6 3 /mmc is thermodynamically expected to transform to the Cu 6 Sn 5 -C2/c at the temperature range below 189 • C.This transformation is partially hampered because of insufficient time during the cooling down process [1].Cu 6 Sn 5 is thus observed at room temperature in both hexagonal and monoclinic crystal structures.The mechanical properties of the Cu 6 Sn 5 IMC in relation to its crystal structure has been previously studied, and has been found that the hexagonal Cu 6 Sn 5 has a comparable hardness to monoclinic Cu 6 Sn 5 , but with a higher Young's modulus [23].
According to the literature, hexagonal-Cu 6 Sn 5 shows intrinsic anisotropic mechanical properties [19,22,24].As a result, the orientation of the Cu 6 Sn 5 grain can impact the observed Young's modulus and hardness.According to V. Kumar et al., Cu 6 Sn 5 formed at Sn-based solder/polycrystalline Cu pads has no preferred orientation [25].While K. H. Prakash reported that Cu 6 Sn 5 films exhibits a strong grain orientation in the directions <102> and <101>.According to Y. Tian, both scallop shape and roof shape Cu 6 Sn 5 has a preferred orientation of {001} plane parallel to polycrystalline Cu surface [26].Cu 6 Sn 5 IMC can have a variety of grain orientations.Therefore, it is crucial to have a comprehensive understanding of how the mechanical properties of Cu 6 Sn 5 relate to grain orientation to assess the mechanical reliability of the joint.For instance, (001) plane of hexagonal Cu 6 Sn 5 has the highest elastic modulus and hardness than the other crystal planes [22,27].The presence of even a single Cu 6 Sn 5 grain with weak mechanical properties may deteriorate the reliability of future micro-joints containing a few Cu 6 Sn 5 grains.Hence, a consistent mechanical response of Cu 6 Sn 5 grains to the external stresses is desired [21].However, achieving a micro-joint with constant mechanical response of Cu 6 Sn 5 grains demands defined crystal structure and orientation, which complicates the fabrication process.
It is possible to assess the mechanical behavior of Cu 6 Sn 5 joints in part by measuring the E i /H value.It has been proposed that the plasticity of the intrinsic brittle materials such as Cu 6 Sn 5 intermetallic compounds can be evaluated by the modulus/hardness ratio (E i /H) [28].According to J. Song, the critical value for E i /H is 17.3-17.5,below which the indent morphology of Cu 6 Sn 5 turns brittle [23].Given that the brittleness of the IMCs is the main factor in IMCs micro-joints failure, Cu 6 Sn 5 -based joints with higher E i /H values are preferred.However, the minimum required E i /H value depends on the specific application.The E i /H ratio for Cu 6 Sn 5 is calculated to be 17 and 19.5, respectively, according to F.X. Che et al. [29] and V. Marques [30].Nevertheless, when third elements are added to Cu-Sn systems, they can take part in the Cu 6 Sn 5 formation and alter its mechanical properties.As a result, when examining Cu 6 Sn 5 's characteristics, the additive elements must be considered.
The effects of third elements such as Ni and Zn on the microstructural and mechanical characteristics of Cu 6 Sn 5 have been investigated [31][32][33][34][35][36][37][38].However, there is insufficient data on how cobalt and indium affect the mechanical and microstructural characteristics of Cu 6 Sn 5 .Grain refinement has been observed by dissolution of Co in the Cu 6 Sn 5 IMC, while Co does not alter the Cu 6 Sn 5 grain morphology [39,40].Furthermore, it has been found that Co and In additive stabilizes the hexagonal-Cu 6 Sn 5 down to room temperature [15,41].A. Yang et al. [11] reported that indium alloying element change the microstructure of the Cu 6 Sn 5 IMC as follows: a) refining the grains and b) increasing the crystal cell volume.The Young's modulus of Cu 6 Sn 5 with a small amount of indium (Cu 54.8 Sn 38.5 In 6.7 ) was measured to be 102.48GPa [42].According to W. Huang [43], the Young's modulus of Cu 6 Sn 5 increase by In alloying while A. Luktuke et al. [44] reported to the contrary (a decrease in Cu 6 Sn 5 Young's modulus by In alloying).In our previous work, we have briefly studied the mechanical and microstructural properties of Cu 6 Sn 5 formed in the Cu-Sn, Cu-Sn-Co, Cu-Sn-In, and Cu-Sn-In-Co SLID systems [13,14,16,45].However, there is no comprehensive study on both microstructural and mechanical properties of the Cu 6 Sn 5 IMC alloyed by Co and In.Hence, in this study we aimed to investigate the impact of third element, Co, In, and Co/In, elements on the microstructural and mechanical properties of the low-temperature Cu-Sn-based SLID joints primarily composed of Cu 6 Sn 5 phase.The crystal structure, grain size and grain orientation of the Cu 6 Sn 5 , Cu 6 (Sn,In) 5 , (Cu,Co) 6 Sn 5 , and (Cu,Co) 6 (Sn,In) 5 joints were characterized using EBSD technique.Furthermore, the mechanical properties (Ei and H) of the abovementioned joints were measured using nanoindentation.

Specimen preparation
All samples were prepared on thermally oxidized (300 nm SiO2) 100 mm Si〈100〉wafers.The device and cap chips for various systems (Cu-Sn, Cu-Sn-In-Cu, Cu-Sn-Co, and Cu-Sn-In-Co) were prepared as

Table 1
The fabrication process for device and cap chips in different SLID systems.

System
Fabrication process Bonding condition

Cu-Sn-Cu
Cap and Device chips: A 60 nm thick TiW adhesion layer was sputtered on the Si wafer, and it was followed by a 100 nm thick copper seed layer sputtering.4 μm of copper was electroplated utilizing NB Semi plate Cu 100 bath, followed by 2 μm of the electroplated tin using NB Semi plate Sn 100 solution from NB technologies.Then the wafer was cut into 1*1 cm pieces.

Cu-Sn-Co
Cap chip: A 60 nm thick TiW adhesion layer was sputtered on the Si wafer, and it was followed by a 100 nm thick copper seed layer sputtering.4 μm of copper was electroplated utilizing NB Semi plate Cu 100 bath, followed by 2 μm of the electroplated tin using NB Semi plate Sn 100 solution from NB technologies.Then the wafer was cut into 1*1 cm pieces.

• C, 1 h
Device chip: A 60 nm thick Ti adhesion layer was sputtered on the Si wafer, and it was followed by a 200 nm thick Mo barrier layer and 80 nm Co sputtering.The wafer was diced into 1*1 cm pieces.

Cu-Sn-In-Cu
Cap and Device chips: A 60 nm thick TiW adhesion layer was sputtered on the Si wafer, and it was followed by a 100 nm thick copper seed layer sputtering.5 μm of copper was electroplated utilizing NB Semi plate Cu 100 bath, followed by 1.7 μm of the electroplated tin using NB Semi plate Sn 100 solution from NB technologies and 1.7 μm of the electroplated indium using indium sulfamate plating bath.Finally, the wafer was cut into 1*1 cm pieces.

Cu-Sn-In-Co
Cap chip: A 60 nm thick TiW adhesion layer was sputtered on the Si wafer, and it was followed by a 100 nm thick copper seed layer sputtering.  1 shows a schematic illustration of the process flow for Cu-Sn sample (as an example).

EBSD measurements
EBSD maps were collected using a scanning electron microscope equipped with electron backscatter diffraction (EBSD) (JIB-4700F).For cross-section analysis using EBSD, samples were prepared in three steps: 1) using standard metallographic methods, 2) chemical etching in a solution with hydrochloric acid and nitric acid diluted in distilled water, and 3) ion beam polishing using JEOL Ion Beam Cryo Cross Section Polisher with ion beam accelerating voltage of 4 kV for 5 min.The EBSD analysis was carried out in the bond-line and focused on the Cu 6 Sn 5 IMC using SEM/EDX analysis prior to the EBSD measurements.The following EBSD detector settings were used during the EBSD data acquisition: 70 • sample tilt; 20 kV accelerating voltage; 14 nA probe current; 10 mm working distance; and 20 nm EBSD step size.The lattice parameters of the phases used for the EBSD data collection are summarized in Table 2.The band contrast map, phase color map, inverse pole figure Y color map, and (001) pole figures of both monoclinic and hexagonal Cu 6 Sn 5 were collected using AZtecCrystal EBSD Processing Software.In this paper, the cross-section surface perpendicular was the normal direction (ND).The transvers direction (TD) and rolling direction (RD) were along the joint cross-section surface and TD was parallel to the bond-line.

Mechanical properties characterization
Nano-indentation test was performed within the IMC layers at the joint area of Cu-Sn, Cu-Sn-In, Cu-Sn-Co, and Cu-Sn-In/Co stacks using a CSM Instruments Nanoindentation tester.A Berkovich diamond indenter was used with following load function parameters: 5 mN maximum load, 10 mN/min loading and unloading rate, and 15 s hold time.The samples were measured using the continuous data of the applied load (P) and indenter displacement (h) during the test.The phase composition of the indentation marks was analyzed using a JSM-6330F field emission scanning electron microscope (SEM; JEOL Ltd.) and INCA X-sight energy-dispersive X-ray spectroscopy (EDX) system (Oxford Instruments).The mechanical characteristics analysis was done using the data for the indents within the Cu 6 Sn 5 phase.For example, Fig. 2 shows the BSE-EDX micrograph of the indents on the Cu-Sn-Co Fig. 1.A schematic illustration of fabrication process of the Cu-Sn SLID bonded sample as an example.joint bonded at 250 • C for 1 h.The continuous data of the indents 1 to 6 located over the (Cu,Co) 6 Sn 5 phase were used to determine the mechanical properties of (Cu,Co) 6 Sn 5 .The Nano-Indentation hardness (H), reduced elastic modulus (Er), and Young's modulus were calculated using equations ( 1)-(3), respectively.
Where P, A, and dP/dh signify the maximum indentation load, the projected area, and the onset slope of the unloading curve, respectively.
And ν i and E i are the Poisson coefficient and the elastic properties of the diamond indenter, respectively, and ν s is the Poisson coefficient of the measured phase.For a diamond tip, E i = 1140 GPa and v i = 0.07.And v s was assumed to be for Cu 6 Sn 5 phase with and without alloying elements (Co, In, and Co/In).[13,14,45].As a result, the Co/In co-alloying does not exhibit consistent effect on the size of Cu 6 Sn 5 grains and is drastically changed by Co content, lower Co content results in smaller grains and higher Co content results in bigger grains.(Cu, Co) 6 (Sn,In) 5 with higher Co content displays roughly the same grain shape and size as (Cu,Co) 6 Sn 5 .However, with less Co, (Cu,Co) 6 (Sn,In) 5 exhibits completely different grain morphology and size, as they are considerably smaller and more rounded.

EBSD measurements and analysis
Fig. 3i-l show the phase color maps for the four different joints.The Cu, Cu 3 Sn and hexagonal and monoclinic crystal structures of Cu 5 Sn 5 are shown in red, blue, green, and yellow, respectively.The hexagonaland monoclinic-Cu 6 Sn 5 were randomly distributed across the bond-line.A small portion of the region that was identified by SEM/EDX as Cu 6 Sn 5 was incorrectly recognized by EBSD analysis as Cu and Cu 3 Sn.It is plausible that the discrepancy is attributed to poor data and diffraction patterns obtained from those specific regions of the samples.This issue is probably caused by the challenging sample preparation process for such samples due to several reasons.These include the narrow bond-line (under 10 μm) with small grains, and complex crystal structures of IMCs which requires better surface quality.
Additionally, there was no clear difference between the preferred crystal structure of Cu 6 Sn 5 formed in the pure Cu-Sn SLID joint and the alloyed Cu-Sn SLID systems.The proportion of hexagonal and monoclinic Cu 6 Sn 5 determined for each joint are listed in Table 3. Cu 6 Sn 5 , Cu 6 (Sn,In) 5 , (Cu,Co) 6 (Sn,In) 5 , and (Cu,Co) 6 (Sn,In) 5 interconnects all has hexagonal to monoclinic ratio of approximately 0.5-0.6,while (Cu, Co) 6 Sn 5 interconnects has a ratio of about 1.However, it was expected to have only hexagonal-Cu 6 Sn 5 by Co and In alloying.According to the literature [48], the Cu 6 Sn 5 formed in Cu-Sn system comprises of both monoclinic and hexagonal crystal structures due to the inadequate transferring time for hexagonal to monoclinic, however, the monoclinic Cu 6 Sn 5 is the thermodynamic stable phase at room temperature.Nevertheless, upon long-term annealing at 150 • C, the Cu 6 Sn 5 hexagonal will entirely transform to the monoclinic crystal structure.According to the findings of the current work, the same phenomenon exists for Cu 6 Sn 5 with either Co, In, or Co/In alloying elements, displaying both hexagonal and monoclinic crystal forms at ambient temperature.However, further investigation is needed to determine how the aging affect the crystal structure transformation for the Cu 6 Sn 5 phase alloyed with Co, In, or Co/In.
The inverse pole figure (IPF) maps are presented in Fig. 3 i-l.The IPF maps reveal that Cu 6 Sn 5 grains have varying orientation with no preferential orientation.To confirm the Cu 6 Sn 5 orientation distribution observed in IPF maps, hexagonal and monoclinic (001) pole figures of the Cu 6 Sn 5 , Cu 6 (Sn,In) 5 , (Cu,Co) 6 Sn 5 , and (Cu,Co) 6 (Sn,In) 5 IMCs were investigated (Fig. 5).According to the pole figures, the Cu 6 Sn 5 grains formed in the Cu-Sn, Cu-Sn-Co, and Cu-Sn-In-Co SLID systems exhibited a random orientation distribution.However, the poles of Cu 6 (Sn,In) 5 formed in Cu-Sn-In system was mostly aligned with the RD and the C-direction of the hexagonal and monoclinic Cu 6 Sn 5 was parallel to the RD.This is also clearly visible on the correlated IPF map (Fig. 3-j), the grains are mostly columnar and elongate in RD direction.
In general, the grain orientation diversification and fine grains strengthen the mechanical properties and improve the mechanical reliability of the joints due to hindering a direct stress propagation to adjacent crystal planes and grain boundary strengthening mechanism.The diversification of the grain orientation, however, could result in the failure of some bumps and possibly the entire joint in the future microbumps, where each interconnection would contain only a few Cu 6 Sn 5 grains.Since a bump only consists of a small number of randomly oriented Cu 6 Sn 5 grains, if one of the grains is oriented so that its mechanical characteristics are poor, it can significantly impact the mechanical properties of the entire joint.Therefore, engineering the grain orientation for extremely small bumps or to refining the grain size such that there is enough variety in the grain orientation might be beneficial.

Mechanical characterization
Representative nanoindentation load-displacement curves during the indentation on Cu 6 Sn 5 , Cu 6 (Sn,In) 5 , (Cu,Co) 6 Sn 5 , and (Cu,Co) 6 (Sn, In) 5 interconnects are shown in Fig. 6.The corresponding measured hardness (H), elastic modulus (Er), Young's modulus (Ei), and Ei/H are listed in Table 4.For the Cu 6 Sn 5 interconnects the Young's modulus and hardness are measured to be 113.6 ± 1.1 and 6.7 ± 0.5 GPa.The average values of Cu 6 Sn 5 are in good agreement with previously reported values [1,6,[49][50][51][52].Both the elastic modulus and hardness of (Cu,Co) 6 Sn 5 and Cu 6 (Sn,In) 5 interconnects are greater than their corresponding values for Cu 6 Sn 5 interconnects.However, the co-alloying of Co and In into Cu 6 Sn 5 exhibited a different impact on the mechanical properties of the Cu 6 Sn 5 interconnects compared to the single Co or In

Table 3
The percentage of Hexagonal and Monoclinic Cu6Sn5 in the studied joints.4 as well as the scatter of Young's modulus and hardness values may be caused by several factors.These factors include changes in the crystal structure of the Cu 6 Sn 5 from one sample to another or from one analyzed point to another, different growth texture for each joint and anisotropy in the mechanical properties, and the chemical composition and the grain size change by alloying elements (Co, In, and Co/In) [19,53].For instance, it has been shown that the hexagonal-Cu 6 Sn 5 has higher Ei and H values than monoclinic-Cu 6 Sn 5 [54].Additionally, it has been previously claimed that the mechanical properties of the Cu 6 Sn 5 is strongly anisotropic and related to the crystal orientation [22] and alloying elements.For instance, according to S. Mu et al., the average values and standard deviations in both elastic modulus and hardness for (Cu,Ni) 6 Sn 5 are higher than Cu 6 Sn 5 due to the growth of the preferred (101) texture and the anisotropy in mechanical properties of (Cu, Ni) 6 Sn 5 [19].In addition, the IMCs' mechanical characteristics can be impacted by grain size, Young's modulus increases with increasing grain size.However, the grain size has a considerable impact on the mechanical properties of materials when the grains size is in the nanoscale [55][56][57].
It was determined from the EBSD analyses of the various samples that all of the interconnects contained hexagonal and monoclinic Cu 6 Sn 5 , randomly distributed across the bond-line.Furthermore, it was observed that the Cu 6 Sn 5 in four studied joints exhibited randomly orientated polycrystalline grains (however, the Cu 6 (Sn,In) 5 shows grains with c-axis elongated in the RD direction).As a result, it is difficult to correlate the crystal structure and grain orientation of various joints, which could account for their varying mechanical properties.However, the varying mechanical properties for different joints and analyzed sites in each joint may be explained by the chemical composition change from one joint to another or across the bond-line.The chemical composition change by the alloying elements can lead to the solid solution strengthening.For instance, it has been claimed that Ni substitution with Cu in Cu 6 Sn 5 lattice reduces the Cu 6 Sn 5 unit cell volumes, generates inter-atomic stresses around the Ni atoms, and subsequently increases Young's modulus and hardness [22,58].Therefore, the same

Conclusion
The microstructural and mechanical properties of pure and Co, In, and Co-In alloyed Cu 6 Sn 5 -interconnects were investigated using EBSD and nanoindentation.The following conclusions can be drawn from this study: 1.The morphology of Cu 6 Sn 5 interconnects can be modified by Co and In addition.Among the interconnects, Co containing Cu 6 Sn 5 and In containing Cu 6 Sn 5 have the largest and smallest average grain sizes, respectively.Notably, the grain size and morphology of (Cu,Co) 6 (Sn, In) 5 interconnect vary across the bond-line due to fluctuations in Co content.Higher Co content corresponds to larger grains.2. Cu 6 Sn 5 , Cu 6 (Sn,In) 5 , (Cu,Co) 6 Sn 5 , and (Cu,Co) 6 (Sn,In) 5 all exhibit both hexagonal (H) and monoclinic (M) crystal structures.Cu 6 Sn 5 grains formed in all interconnects showed a random orientation distribution.However, Cu 6 (Sn,In) 5 grains in the Cu-Sn-In system were mostly aligned with C-direction of the hexagonal and monoclinic Cu 6 Sn 5 grains.Moreover, the (Cu,Co) 6 (Sn,In) 5 grains morphology can be changed by changing the Co content.Therefore, it is possible to design the layout of the interconnects so that either small or large grains are present to take use of the mechanism for reinforcing grain boundaries or the consistent mechanical response of the joints, respectively.In addition, indium can lower the bonding temperature when compared to pure Sn, and Co can be employed as a contact metallization for Cu-Sn-In systems.As a result, Cu-Sn-In-Co interconnects have a lot to offer electronics integrations.

Fig. 2 .
Fig. 2. BSE-SEM image of the indents on Cu-Sn-Co joint bonded at 250 C for 1h

Fig. 3 Fig. 3 .
Fig. 3 gives the BC (Band Contrast), phase color, and IPF-Y (Inverse Pole Figure Y) maps measured from the cross-sectional surfaces of the Cu-Sn, Cu-Sn-In, Cu-Sn-Co, and Cu-Sn-In-Co joints.All the phases across the bond-line were identified using SEM-EDX analyzing, and the Cu 6 Sn 5 -containing region is indicated by the dashed-lines in Fig. 3.The results of the Band Contrast maps (Fig. 3 a-d) reveal that the different alloying elements (Co and In) have a distinct impact on the Cu 6 Sn 5 morphology.By adding In into Cu 6 Sn 5 interconnects, the grains are refined and elongated in RD.While the addition of Co as an alloying element increased the size of Cu 6 Sn 5 grains.The average grain size (both width and length of the grains) of Cu 6 Sn 5 , Cu 6 (Sn,In) 5 , (Cu,Co) 6 Sn 5 , and (Cu,Co) 6 (Sn,In) 5 interconnects calculated using BC maps is shown in Fig. 4. The (Cu,Co) 6 Sn 5 showed the largest average grain size.Cu 6 (Sn, In) 5 on the other hand, had grains with smallest size.In general, the alloying elements involved in phase (such as Cu 6 Sn 5 ) formation lower the critical nucleus energy, which enhances the nucleus and may subsequently refine the grain sizes [46].However, according to Y. Wang et al.Co (which occupy the Cu site [47]) promotes the Cu 6 Sn 5 IMC

Fig. 4 .
Fig. 4. The average grain size (width and length of the grains) of Cu 6 Sn 5 phase formed in Cu-Sn, Cu-Sn-In, Cu-Sn-Co, and Cu-Sn-In-Co joints.

F.
Emadi et al.   alloying elements, the hardness decreased, whereas the Young's modulus increased slightly.Compared to Cu 6 Sn 5 interconnect, the alloyed Cu 6 Sn 5 interconnects had larger scatter in hardness and Young's modulus values and (Cu,Co) 6 (Sn,In) 5 interconnect showed the highest scatter in the Young's modulus values measured from various locations along the bond line.The different mechanical properties of Cu 6 Sn 5 interconnects listed in Table

Table 1 .
They were then soldered in an air muffle furnace.Because Sn-In has a lower melting point than Sn, samples with Sn and In were bonded at a lower bonding temperature (200 • C) than those with just Sn (T b = 250 • C).Fig.
5 μm of copper was electroplated utilizing NB Semi plate Cu 100 bath, followed by 1.7 μm of the electroplated tin using NB Semi plate Sn 100 solution from NB technologies and 1.7 μm of the electroplated indium using indium sulfamate plating bath.The wafer was cut into 1*1 cm pieces.Device chip: Co foil (purity: 99.99%, Goodfellow Ltd.) 1 mm in thickness was cut into pieces 1*1 cm in size.The pieces were mechanically ground to 2400 papers, cleaned with acetone, and airdried before bonding.200 • C, 1 h F. Emadi et al. described in

Table 2
Lattice parameters of the phases used for EBSD data collection.

Table 4
The mechanical properties of the IMCs.mightoccur when Cu 6 Sn 5 is alloyed with Co and In.As mentioned, Cu and Sn atoms in Cu 6 Sn 5 would be replaced by Co and In, respectively.Co, Cu, In, and Sn all have distinct atomic radii, hence any substitutions would result in substitutional defects.Additionally, Co-Sn and Cu-In interact more strongly than Cu-Sn based on their binary phase diagram.Therefore, by adding Co or In alloying elements into Cu 6 Sn 5 , the Young's modulus and hardness values may increase because of these effects (substitutional defects and greater atoms interaction).It is crucial to note that, in the current study, approximately 50% of the Sn atoms were substituted by In in Cu-Sn-In, while in Cu-Sn-Co a small number of Cu atoms were substituted by Co atoms.Therefore, the substitutional defects impact in the studied Cu-Sn-In could be significantly higher than in the Cu-Sn-Co.Hence, it could be the reason why indium alloying showed greater impacts on the increasing of the hardness than Co alloying.However, Cu 6 Sn 5 with co-alloying cobalt and indium showed completely different results, its hardness was less than while its Young's modulus was comparable to Cu 6 Sn 5 .To find the impact of the co-alloying Co and In on the Cu 6 Sn 5 lattice, first-principles calculations of structural and mechanical properties of Cu 6 Sn 5 with Co and In is required as we have various bonds: Cu-Sn, Co-Sn, Cu-In, and Co-In.As listed in Table 4, the Ei/H value for Cu 6 Sn 5 , Cu 6 (Sn,In) 5 , (Cu, Co) 6 Sn 5 , and (Cu,Co) 6 (Sn,In) 5 interconnects were measured to be 17.0 ± 0.8, 16.4 ± 2.3, 19.7 ± 1.4, and 20.7 ± 0.9.The (Cu,Co) 6 (Sn,In) 5 interconnect has the greatest Ei/H value, and accordingly highest plasticity which is preferable for interconnects. phenomenon 3. The addition of Co and In into Cu 6 Sn 5 interconnects alters their mechanical properties.Incorporating Co or In reduces the Young's modulus (E i ) and hardness (H).However, co-alloying Co and In in Cu 6 Sn 5 leads to decreased hardness but a slight increase in Young's modulus.The Ei/H value for Cu 6 Sn 5 , Cu 6 (Sn,In) 5 , (Cu,Co) 6 Sn 5 , and (Cu,Co) 6 (Sn,In) 5 interconnects are measured to be 17.0 ± 0.8, 16.4 ± 2.3, 19.7 ± 1.4, and 20.7 ± 0.9. 4. The (Cu,Co) 6 (Sn,In) 5 interconnect has the greatest Ei/H value and accordingly highest plasticity, which is preferable for interconnects.