Tailoring Morphology and Size of Microstructure and Tensile Properties of Sn-5.5 wt.%Sb-1 wt.%(Cu,Ag) Solder Alloys

Transient directional solidification experiments, and further optical and scanning electron microscopy analyses and tensile tests, allowed the dependence of tensile properties on the micromorphology and length scale of the dendritic/cellular matrix of ternary Sn-5.5Sb-1Ag and Sn-5.5Sb-1Cu alloys to be determined. Extensive ranges of cooling rates were obtained, which permitted specific values of cooling rate for each sample examined along the length of the casting to be attributed. Very broad microstructural length scales were revealed as well as the presence of either cells or dendrites for the Ag-containing alloy. Hereafter, microstructural spacing values such as the cellular spacing, λc, and the primary dendritic spacing, λ1, may be correlated with thermal solidification parameters, that is, the cooling rate and the growth rate. While, for the Cu-containing Sn-Sb alloy, the β-Sn matrix is characterized only by the presence of dendritic arrangements, the Ag-containing Sn-Sb alloy is shown to have high-velocity β-Sn cells associated with high cooling rate regions, i.e., positions closer to the bottom of the alloy casting, with the remaining positions being characterized by a complex growth of β-Sn dendrites. Minor additions of Cu and Ag increase both the yield and ultimate tensile strengths when compared with the corresponding values of the binary Sn-5.5Sb alloy, with a small reduction in ductility. This has been attributed to the homogeneous distribution of the Ag3Sn and Cu6Sn5 intermetallic particles related to smaller λ1 characterizing the dendritic zones of the ternary Sn-Sb-(Cu,Ag) alloys. In addition, the Ag-modified Sn-Sb alloy exhibited an initial wetting angle consistent with that characterizing the binary Sn-5.5Sb alloy.


INTRODUCTION
2][3][4][5] The majority of research on this topic has been devoted to the examination of alloys of well-established binary systems such as Sn-Pb, 6 Sn-Ag, 7 Sn-Ni, 8 Sn-Cu 9 and Sn-Zn. 10During soldering processes, it is presumed that the experienced solidification conditions should affect not only the morphology but also the microstructure length scale of the b-Sn phase as well as of other microstructural phases.][13] For example, in a recent research work on the Sn-5.5 wt.%Sb solder alloy, 14 it was shown that the morphology of the b-Sn phase is strongly dependent on the cooling rate.Relatively high cooling rates were shown to be enough to guarantee the growth of high-cooling rate cells, which occurred for _ T > 1.2 K/s.Finer b-Sn cells have been associated with higher tensile strength, which has been accredited to the presence of very homogeneously distributed SbSn intermetallic particles (IMC) throughout the intercellular regions.
][17] The largest part of the studies refer to the Sn-5 wt.%Sb alloy as having good mechanical properties and a solder/substrate contact angle of about 43°.Such characteristics allow the consideration of such an alloy as a potential alternative to replace lead-containing solders.Indeed, several applications of Sn-5Sb alloy have been reported: the creation of hermetic seals in multichip modules, bonding a semiconductor device onto a substrate, 18 and attachment of input/output (I/O) pins to multilayer ceramic substrates. 19The replacement of other alloys with the Sn-5Sb alloy in the last application may provide the advantages of either lowering the joining temperature by about 50°C or reducing the stress induced by the joining process.In another application, the Sn-5Sb alloy was used as a corrosion protection coating on a steel plate, as well as an electrically conductive area that could be used for subsequent grounding. 20dditional improvements on the microstructure features and mechanical properties of the binary Sn-Sb alloys must require the addition of third alloying elements, such as In, 21 Cu, 22 Zn, 23 and Bi, [15][16][17] which are those most frequently employed.
According to Esfandyarpour and Mahmudi, 15 the addition of 1.5% Bi, and 1.5% Cu to the Sn-5Sb solder is able to improve both strength and ductility of the base alloy.In this research, these amounts have been considered as non-deleterious to the wettability and other soldering characteristics.These authors state that the more pronounced addition effect was found for the Bi-containing alloy.This gain in the properties was due to the solid-solution hardening effect provoked by bismuth within the b-Sn matrix.The addition of third elements is stated as the reason why structural refinement is obtained, which is mainly emphasized in the case of the Cu-containing alloys referring to the presence of Cu 6 Sn 5 IMCs.The samples used in the referred investigation were produced through casting into 120 mm 9 30 mm 9 13 mm slabs.There are no considerations regarding the effects generated on the microstructure by possible variations in cooling rate along the slabs.
In order to understand how the microstructure evolves during the formation of a solder joint, it is considered fundamental, firstly, to focus the analysis on an appropriate way of interpreting the original microstructures generated during the solidification step.Esfandyarpour and Mahmudi 15 affirmed correspondingly that this is an important issue, which affects the reliability of the manufactured solder joints. 15eranmayeh and co-authors 17 stated that the average dendritic spacing associated with the binary Sn-5Sb alloy ($ 35 lm) was higher than that determined for the other Sn-5 wt.%Sb-X alloys investigated, i.e., the Sn-5 wt.%Sb-1.5 wt.%Bi, and Sn-5 wt.%Sb-1.5 wt.%Ag alloys.The microstructure of the ternary Ag-containing alloy consisted of second phases almost randomly distributed throughout the Sn-rich matrix.These phases were identified as being the Ag 3 Sn and the Sb-Sn IMCs.Such results were obtained by energy-dispersive x-ray (EDX) and were shown to be consistent with additional findings by x-ray diffraction (XRD) analysis.
El-Daly et al. 24 examined the influence of the addition of silver (3.5Ag) and gold (1.5Au) on the physical properties of the Sn-5Sb-based lead-free solder alloy.In both cases, the addition allows significant changes on the latent heat of fusion (DH).The Sn-5Sb-1.5Aualloy, for instance, exhibited the lowest DH.This indicates that this alloy may be the most effective in terms of energy saving when compared with the binary Sn-5Sb and with the ternary Sn-5Sb-3.5Agalloys.Furthermore, the Sn-5Sb-1.5Auleadfree solder was observed to give a good combination of higher creep resistance and fracture time.
El-Daly et al. 25 examined the microstructure characteristics and their influences on the tensile properties of the Sn-5 wt.%Sb, Sn-5 wt.%Sb-0.7 wt.%Cu, and Sn-5 wt.%Sb-0.7 wt.%Ag alloys.The separated addition of a third element (Cu or Ag) to the Sn-Sb alloy provoked refinement of microstructure when compared with that observed for the non-modified alloy.Two types of intermetallic particles were distinguished in the case of the Cu-modified alloy, i.e., Cu 6 Sn 5 and Cu 3 Sn.However, only the Ag 3 Sn phase was observed in the microstructure of the Ag-containing alloy.The SbSn IMC was found to be precipitated within the b-Sn matrix.The higher level of refinement of the Cucontaining alloy microstructure was explained due to the presence of the two aforementioned IMCs, which are distinguishable in nature.The tensile strength was higher for the Sn-5 wt.%Sb-0.7 wt.%Cu alloy compared with that determined for the Sn-5 wt.%Sb, and Sn-5 wt.%Sb-0.7 wt.%Ag alloys.According to these authors, the growth and distribution of fine needle-like Cu-Sn IMCs were responsible for the strengthening effect.
In the present research work, two ternary Sn-5.5 wt.%Sb-1 wt.%Ag, and Sn-5.5 wt.%Sb-1 wt.%Cu alloys are directionally solidified (DS) under transient heat flow conditions.The study is focused on evaluating the resulting microstructural features, i.e., morphology, size and distribution of the formed phases along the length of the DS castings.Experimental growth laws relating cellular spacing and primary dendritic spacing with both cooling rate and growth rate are proposed.The effects promoted by minor additions of 1 wt.% of copper (Cu) and 1 wt.% of silver (Ag) in the binary Sn-5.5 wt.%Sb alloy are examined, emphasizing either the changes in morphology and size of the phases forming the microstructure or the impacts on the corresponding tensile mechanical properties, as well as the effects on wettability.

EXPERIMENTAL
Directionally solidified (DS) specimens of the Sn-5.5 wt.%Sb-1 wt.%Ag, and Sn-5.5 wt.%Sb-1 wt.%Cu alloys were obtained by using a watercooled solidification setup, which allows keeping a transient heat flow regime during the solidification processing. 26The Cu-containing and Ag-containing Sn-Sb base alloys were melted in situ by radial electrical wiring heating a cylindrical stainless steel split mold.When melt temperatures of around 10% above the liquidus temperatures (T p ) were achieved, the furnace coilings were disconnected and at the same time the external water flow at the bottom of the container was initiated through a water supply tube, initiating the cooling procedure.A number of J-type thermocouples were laterally placed along the length of the castings, where their tips were placed along the centerline of the container so that the thermal profiles could be acquired.The alloys of interest in the present investigation were directionally solidified against a low-C (SAE 1020) steelbottom mold.The inner surface of the bottom-part molds had been finished with a 1200-grit SiC abrasive paper.
The temperature-time records were the basis to experimentally determine the values of growth rate (V L ) and cooling rate ( _ T) along the length of the DS castings according to the positons in which the thermocouples have been connected.Experimental plots of position (P), from the water-cooled surface of the casting, and the corresponding time (t L ) of the liquidus isotherm (T L ) passing by each thermocouple, i.e., P = f(t L ), gave rise to experimental fitting functions for each alloy examined.A time-derivative of these fitting functions was carried out so that the growth rate (V L ) could be calculated in the form of V L = f(t L ).By replacing t L = f(P) with t L inside the equation for V L , resultant equations of the form V L = f(P) have been obtained.The thermal data recorded immediately after the passage of the liquidus isotherm by each thermocouple was considered to compute the cooling rates ( _ T L ) along the length of the castings.The following positions of thermocouples from the metal/mold interface were adopted in the experiments: 2 mm, 7 mm, 11 mm, 16 mm, 21 mm, 38 mm, 51 mm, and 68 mm for the Sn-5.5 wt.%Sb-1 wt.%Cu alloy, and 2 mm, 6 mm, 11 mm, 16 mm, 21 mm, 45 mm, 57 mm, and 69 mm for the Sn-5.5 wt.%Sb-1 wt.%Ag alloy.
Longitudinal and transverse samples were extracted for each DS Sn-5.5Sb-X alloy casting from at least nine positions along the length of the DS castings and were prepared using conventional metallography techniques.A light etching procedure with the solution of 92% (vol.)CH 3 OH, 5% (vol.)HNO 3 and 3% (vol.)HCl was applied during 2-5 s, so that microstructural characteristics regarding the morphology and size of the b-Sn matrix could be revealed.Also, such a procedure may provide enough contrast between Sn-rich and formed intermetallic compounds so that the observation of these phases could be attained by both scanning electron microscopy (SEM) and light microscopy (LM).
Therefore, the microstructures of the two Sn-Sb base alloys were investigated using not only an Olympus 41GX light microscope with a coupled optical image processing system (Olympus, Japan) but also using a field emission gun-scanning electron microscope (SEM) and a FEI/SEM-EDS (FEI Inspect F50).The triangle method 27,28 was applied to each of the prepared samples so that the primary dendritic arm spacing (k 1 ) and the cellular spacing (k c ) of the DS castings could be determined.At least 40 measurements were performed for each selected transverse position along the length of the DS castings.
The transverse specimens were prepared according to specifications of the ASTM Standard E 8 M/04 and tested in a MTS 810 testing machine at a strain rate of about 3 9 10 À3 s À1 at room temperature.In order to ensure reproducibility of the tensile results, three specimens were tested for each selected position from where the specimens were extracted from the DS castings, with a view to determining the ultimate tensile strength, the yield tensile strength and the elongation-to-fracture as a function of the position in the casting.The tensile properties of the four DS alloys castings were then examined and suitably correlated with the corresponding cellular spacing, k c , and the primary dendritic spacing, k 1 , which are considered representative microstructural aspect because the length scale is directly related to the distribution of the reinforcing IMCs within the microstructure.Some samples were also investigated by a fluorescence spectrometer (Rigaku model RIX-3100) to estimate local average Sb, Cu and Ag concentrations.The x-ray diffraction (XRD) patterns for phases formed in the ternary Sn-Sb-Cu and Sn-Sb-Ag alloys examined have been acquired by a PANanalytical X' Pert pro MRD XL diffractometer with a 2h range from 20°to 90°, CuKa radiation and a wavelength, k, of 0.15406 nm.
A Goniometer Kru ¨ss DSHAT HTM Reetz was employed to measure the contact angles (h) of the Sn-5.5 wt.%Sb, Sn-5.5 wt.%Sb-1 wt.%Ag ,and Sn-5.5 wt.%Sb-1 wt.%Cu alloys. 29The computational method ''tangent-2'' allowed both average values of h R and h L (R-right and L-left) to be determined.For each couple, solder alloy/carbon steel, triplicate wetting tests were applied.This means that each sample was heated, molten and examined individually.Each wetting run allows the shape of the molten droplet to be tracked and recorded during the tests.The determination of the contact angles was carried out continuously according to the variations observed in the form of the droplet.Melt superheats of 20% above the liquidus temperatures were adopted, followed by a natural cooling procedure inside the furnace.The initial times of the wetting tests (first 15 s) were considered to determine the initial contact angles (h i ), whereas the end part of the experiment defined the equilibrium contact angles (h e ).

RESULTS AND DISCUSSION
Figure 1 shows pseudo-binary diagrams of Sn-5 wt.%Sb-X alloys: (a) X = wt.%Cu,and (b) X = wt.%Ag,calculated by the computational thermodynamics software Thermo-Calc, where the vertical dashed lines indicate the two alloys examined in the present study.The information on the equilibrium phases of such diagrams evolving along solidification will permit a further comparison with those resulting of the non-equilibrium transient directional solidification of the alloys examined.
Figure 2 shows the cooling curves related to different positions from the water-cooled bottom surface of each examined Sn-Sb base alloy casting.Such cooling curves have been examined on the basis of the passage of the liquidus temperature by each thermocouple inserted.After that, experimental plots of cooling rate and growth rate of the Sn-Sb-Cu and Sn-Sb-Ag alloys varying with position in the casting were conceived.
It is worth noting in Fig. 3 that cooling rates and growth rates associated with the Cu-containing alloy are slightly higher than those found for the other alloy.A strong variation of both solidification thermal parameters is noted for the alloys examined, as shown in Fig. 3.
Figure 4 depicts the distribution of the solute elements considering a number of positions along the length of the castings.Almost constant chemistries of each component have been observed for the Cu-modified and the Ag-modified alloys.This means that the chemistry of all samples of a certain alloy examined for microstructure and tensile properties can be considered the same, avoiding further effects related to macrosegregation.Horizontal lines roughly represent the experimental evolutions of Sb, Cu and Ag.It seems that neither shrinkagedriven 30 nor solutal convection effects 31 were enough to provoke alteration on the resultant solute distribution along the length of the DS Sn-Sb based alloys castings.
Some characteristic microstructures of each alloy related to distinct levels of cooling rate can be seen in Figs. 5 and 6.These light microstructures prove the prevalence of the dendritic morphology for the Sn-Sb-Cu alloy with larger microstructural spacing associated with lower cooling rates.For the Sn-Sb-Ag alloy, the growth of cells occurred for samples related to higher cooling rates, and dendrites started to appear as the dominant morphology with the decrease in cooling rate.A similar occurrence of Dias, Costa, Soares, Silva, Cheung, Spinelli, and Garcia such morphological transition was also reported in a previous study focused on the base Sn-5 wt.%Sb solder alloy. 14The addition of Ag was not able to avoid the development of high-cooling rate cells, which have prevailed for cooling rates higher than 1.3°C/s, in quite good agreement with the critical value found for the Sn-5.5 wt.%Sb alloy, i.e., 1.2°C/ s.This type of morphology can be considered atypical and very rare in metallic alloys, [32][33][34] since additional increase in V L or in _ T L , instead of inducing finer dendritic trunks to grow, may change the solidification front back to cellular, characterizing a reverse morphological transition.
High-magnification light microstructures are inserted in Figs. 5 and 6 so that the formation of the intermetallic particles can be seen.These particles were found to be located within the interdendritic and intercellular zones of the formed structures.The Ag 3 Sn particles embedded into the interstitial regions of b-Sn cells and b-Sn dendrites have their shape varying from globular-like to fibrous-like with the decrease in cooling rate.
In the present investigation, experimental tendencies regarding the variations of k 1 and k c with cooling rate found in a previous research with the Sn-5.5 wt.%Sb alloy 14 were used for comparison purposes, since the experimental conditions are quite similar to those of the ternary Sn-5.5 wt.%Sb-1 wt.%Cu and Sn-5.5 wt.%Sb-1 wt.%Ag alloys.As can be seen in Fig. 7, the additions of either Cu or Ag have affected the microstructural spacing, when compared with those representing the binary Sn-5.5 wt.%Sb alloy.Even so, a À 0.55 exponent power law is able to properly represent the experimental points measured for all alloys examined.
The line representing the evolution of primary dendritic arms, observed for the Cu-containing alloy, remained between the differences in trend of the growth of aligned cells and of primary dendrites   trunks for the base alloy.For _ T L < 0.9°C/s one can affirm that the Cu-containing alloy exhibited lower k 1 values when compared with those of the other non-modified Sn-Sb alloy.The CuSn IMC formed in the dendritic microstructure of the Sn-Sb-Cu alloy is recognized to act as a nucleation site for both SbSn and b-Sn. 15As a consequence, the dendritic growth can be restricted due to increase in the nucleation rate of the b-Sn phase.However, the scattered points become higher for _ T > 1.3°C/s when compared with the inserted tendency related to the growth of b-Sn cells for the Sn-5.5 wt.%Sb alloy.
A certain range of experimental cooling rates is related to b-Sn cells whereas another range is related to dendrites for the ternary Sn-Sb-Ag alloy in Fig. 7b.Plots of cell spacing and primary dendritic spacing highlight both regions, establishing a morphological transition.Although two distinct morphologies have been identified for the Sn-rich phase of the Sn-Sb-Ag alloy, as can be seen in Fig. 7b, if one could consider only a single set of scattered points for this alloy (solid and open stars together), the experimental tendencies stay roughly in an intermediary position when compared with the other inserted tendencies of the Sn-5.5 wt.%Sb alloy.For cooling rates higher than 1.3°C/s, only cells prevailed in the microstructure of the Agcontaining alloy, which were characterized by higher k c values than those found in the binary Sn-Sb base alloy.In contrast, the k 1 values referring to the dendritic zone of the Sn-5.5 wt.%Sb-1 wt.%AgThe presence of fibers of Ag 3 Sn seems to be conducive for such a refinement on k 1 , 15,17 since such IMCs located on the b-Sn dendrite interstices (see Fig. 6) seem to effectively block further growth of primary dendrite stems.These results are in agreement with the reports by Geranmayeh and coauthors 17 who stated that the average dendritic spacing associated with the binary Sn-5 wt.%Sb was higher than that found for the Sn-5 wt.%Sb-1.5 wt.%Ag alloy.
Figure 8 shows the average k 1 and k c values versus growth rate, V L , together with the maximum and minimum spacing deviations for each plotted point of the ternary Sn-5.5 wt.%Sb-1 wt.%Cu, and Sn-5.5 wt.%Sb-1 wt.%Ag alloys.A single experimental power function trend was found to encompass all the experimental scatters.Even though the increase in V L may result in the decrease of either cellular or dendritic spacings, the microstructural spacing can be considered essentially the same when a single growth rate is considered, regardless of either the type of morphology of the b-Sn matrix or the added third element.However, it is worth noting that the soldering process occurs in transient heat flow conditions, where the microstructure is controlled by the cooling rate, _ T, which is given by the product of the growth rate and thermal gradient (G L ), that is _ T = G L V L .Since G L and V L vary differently in time, these thermal parameters cannot be dissociated when analyzing the microstructural length scale of solders, but rather they must be synthesized by the cooling rate.
The x-ray diffractograms for the Sn-Sb-Cu and Sn-Sb-Ag solder alloys examined are shown in Fig. 9. Five different positions were examined along the length of each alloy casting covering a representative range of cooling rates.The characteristic x-ray peaks have been identified as being associated with b-Sn, Sb-Sn, Cu 6 Sn 5 , and Ag 3 Sn phases.The Cu-rich and Ag-rich IMCs are related to the Sn-5.5 wt.%Sb-1 wt.%Cu alloy, and Sn-5.5 wt.%Sb-1 wt.%Ag alloys, respectively.
Although there are no clear tendencies regarding the intensities of peaks for the different cooling rates related to the Cu-modified alloy samples, the intensities of peaks of Ag 3 Sn and Sb-Sn IMCs increased for low cooling rate regions in the case of the Ag-containing alloy.This is based on the intensities of peaks associated with the following 2h angles as found for the SbSn phase: 29.1°, 41.7°, and Growth rate, V L (mm/s) Sn-5.5wt.%Sb-1wt.%Cu (λ 1 ) Sn-5.5wt.%Sb-1wt.%Ag (λ 1 ) Sn-5.5wt.%Sb-1wt.%Ag (λ C ) Fig. 8. Cellular and primary dendritic spacing as a function of the growth rate (V L ) for the directionally solidified Sn-5.5 wt.%Sb-1 wt.%Cu, and Sn-5.5 wt.%Sb-1 wt.%Ag alloys.
60.3°, and for the Ag 3 Sn: 37.6°, 60.3°, and 89.3°(see Fig. 9b).This result is in agreement with Fig. 6, which shows that thicker areas are occupied by the mentioned IMCs within the interdendritic zones of the microstructures related to slower cooling conditions (0.06°C/s) during directional solidification.Figures 10 and 11 show the elemental SEM-EDS mapping for the examined Sn-Sb base alloys.The Cu and Ag content (in blue in Figs. 10 and 11, respectively) show higher intensities within the interdendritic regions, while Sb (in green) is mainly almost randomly distributed within the Sn-rich matrix.The presence of the Cu 6 Sn 5 and Ag 3 Sn IMCs within the interdendritic zones is therefore verified through these findings.
Figures 12 and 13 show the results of tensile tests against the microstructural spacings, k 1 and k c , separated by three main parameters, i.e., ultimate tensile strength (r u ; Fig. 12), yield tensile strength (r y ; Fig. 12) and elongation to fracture (d; Fig. 13).Considering the ternary Sn-5.5 wt.%Sb-1 wt.%Cu alloy with dendritic micromorphology observed in all the samples, it can be seen in Fig. 12a that a decrease in the primary dendrite arm spacing is associated with increasing r u and r y .Lower k 1 triggers the Cu 6 Sn 5 fine particles to be better distributed throughout the microstructure.These particles are reinforcing ones, which result in increasing alloy strength.However, it seems that there is a limit for this reinforcing effect, which is associated with 1/k 1 1/2 < 0.10.Once this limit is achieved, the primary spacing becomes so large that the coarser IMCs may no longer be homogeneously distributed.The values of r u and r y remain   Dias, Costa, Soares, Silva, Cheung, Spinelli, and Garcia practically unaltered for 1/k 1 1/2 < 0.10.For the scattered points in which representative variation in strength is observed, Hall-Petch-type correlations are proposed.
As can be seen in Fig. 12b 1 , two collections of strength data versus k 1,c can be observed for the Ag-containing alloy.Firstly, roughly unaltered r u and r y values were observed for 1/k 1 1/2 < 0.13, representing the results related to the samples having dendritic morphology.Secondly, for the cellular tested samples, a decrease in the cellular spacing can be associated with increasing r u and r y .The growth of cells may be associated with both rapid cooling conditions and Ag 3 Sn globules within the intercellular regions, as can be seen in Fig. 6.Those characteristics in combination with very fine cells induce higher strengths. 35Hall-Petch-type correlations are proposed to represent r u and r y as a function of the cellular length scale.Similarly, finer cellular and cellular-dendritic microstructures observed for a Sn-10.2Sb 36alloy resulted in increased hardness and compressive strength.
So, the change in morphology and in the representative microstructural length scale plays an important role on the resultant properties in Sn-Sb base alloys.Parallel graphs (gray lines) can be observed in Fig. 12b 2 for the binary Sn-5.5 wt.%Sb alloy, which shows stabilized tensile strengths for the dendritic region and increasing values for lower k c , despite lower strengths in comparison with the Agcontaining alloy.Chemical properties, such as the electrochemical corrosion resistance, have also been reported to be influenced by the morphology and microstructural length scale of a Sn-based solder. 37s can be seen in Fig. 13, the combined results of the two ternary alloys and their respective morphologies show that the elongation-to-fracture increases slightly with the decrease in microstructural spacing of the DS samples (increase in k 1 ).The experimental tendency indicates higher ductility related to samples extracted from near the water-cooled bottom surface of the castings (finest microstructures).Further tendencies included for the binary Sn-5.5 wt.%Sb alloy in Fig. 13 (gray lines) allow the statement that the ductility of this alloy is superior to that of the ternary alloys, with an exception made for the results considering 1/k c 1/ 2 > 0.3, which correspond to the growth of very fine b-Sn cells.
It is certainly important to evaluate the changes in strength and ductility by adding third elements to the Sn-Sb base alloy, but wettability changes must correspondingly be checked even for minute additions of elements.Figure 14 shows the experimental evolutions of the contact angles for Sn-5.5 wt.%Sb-1 wt.%Cu and Sn-5.5 wt.%Sb-1 wt.%Ag solder alloys against steel substrates.Same fluctuations in the scattered points can be observed which have already been explained in the literature. 38his seems to be mostly related to the viscous flow of the melt during the wetting process.
The Sn-5.5 wt.%Sb-1 wt.%Ag solder alloy provided the best level of initial wetting, i.e., h i = 42.1 o , which can be considered a similar situation concerning soldering conditions, characterized by rapid cooling conditions during early stages of solidification of the alloy imposed by contact with the cold substrate.After that, the experimental contact angles tend to decrease, producing equilibrium steady-state conditions once 400 s of wetting is achieved (see Fig. 14).The equilibrium contact angles (h e ) are quite close when the values for both Cu-and Ag-modified alloys are compared.The lowest initial wettability level was observed in the profiles recorded for the Sn-5.5 wt.%Sb-1 wt.%Cu alloy.
According to Table I, the mean equilibrium initial contact angles experimentally determined for the Sn-5.5 wt.%Sb, Sn-5.5 wt.%Sb-1 wt.%Ag, and Sn-5.5 wt.%Sb-1 wt.%Cu alloys are 45.1°, 42.1°, and 55.0°, respectively.The level of wetting of the Sn-5.5 wt.%Sb alloy can be considered better than or comparable with findings reported in the literature for the same composition. 15able I synthetizes the main properties determined in the present investigation considering a range of cooling rates typically found in solder joints for microelectronic packages.Overall, the best balance of tensile strength, ductility and wettability among the alloys examined is that obtained for the ternary Sn-5.5 wt.%Sb-1 wt.%Ag alloy casting, especially for samples extracted from positions that are closest to the cooled bottom of the DS casting, i.e. associated with higher cooling rates.

Fig. 4 .
Fig. 4. Experimental profiles of the solute elements along the length of the ternary DS Sn-Sb-Ag and Sn-Sb-Cu alloys castings.

Fig. 5 .
Fig. 5. Tridimensional optical views of fully dendritic patterns observed in the ternary Sn-5.5 wt.%Sb-1 wt.%Cu alloy casting emphasizing longitudinal section features for different levels of cooling rate, that is 11.7°C/s, and 0.17°C/s.

Fig. 6 .
Fig. 6.Typical solidification microstructures of transverse and longitudinal sections of both dendritic and cellular patterns observed along the length of the Sn-5.5 wt.%Sb-1 wt.%Ag alloy casting.
alloy casting were observed to be around 2 times lower.

Fig. 7 .
Fig. 7. Variation in the range of the cellular spacing and of the primary dendrite arm spacing as a function of cooling rate for: (a) Sn-5.5 wt.%Sb-1 wt.%Cu alloy and (b) Sn-5.5 wt.%Sb-1 wt.%Ag alloys.R 2 is the coefficient of determination of the fitted curves.