Tensile creep and microstructural characterization of bulk Sn3.9Ag0.6Cu lead-free solder

The microstructural and creep behavior of bulk 63SnPb37 and the Pb-free solder alloy Sn3.9Ag0.6Cu are reported and compared. The Sn3.9Ag0.6Cu alloy showed much lower absolute creep rates than 63SnPb37. The size and distribution of the intermetallic compound (IMC) coarsened with increasing creep temperature. A number of coarsened precipitates of Cu₆Sn₅ segregate around β-Sn grain boundaries. After creep at 80°C and 115°C. the β-Sn particles in the Sn3.9Ag0.6Cu alloy are strongly aligned at approximately 45° to the uniaxial tension, parallel to the maximum shear-stress planes. The powerlaw-defined stress exponent significantly increases with increasing stress in both the 63Sn37Pb and Sn3.9Ag0.6Cu alloys; therefore, the Dorn model is unsuitable for these materials over large stress and temperature ranges. Both sets of experimental data were successfully fit with the present power-law stress-dependent energy-barrier model and the Garofalo model. However, the application of the present power-law stress-dependent energy model resulted in a significantly lower estimated variance as compared to the Garofalo model.     

Anomalously high tensile creep rates from thin cast Sn3.9Ag0.6Cu lead-free solder

The present paper compares the creep and microstructural changes during creep behavior of bulk and thin cast forms of Sn3.9Ag0.6Cu. The processing parameters of the thin cast material were selected to result in a very fine microstructure analogous to what occurs in very small size solder electronic interconnections. We found that the thin cast material is less creep resistant than the bulk material. A comparison of Ag element maps between as-crept bulk and thin cast material shows that the relevant climb process occurs in a very different environment in the bulk material as compared to the thin cast material. In the bulk material, the relevant climb process occurs within a finely dispersed intermetallic compound (IMC) eutectic, which covers broad areas within the material. In the thin cast material, the relevant climb process occurs primarily in the beta-Sn grains that continuously surround isolated, coarse IMC particles. This resulted in the activation energy of the bulk material being larger than that for the thin cast material. Finally, it is important to note that the strength deficiency of the thin cast material is persistent—once the material is cast in thin cast form, it will remain weak in comparison to the bulk material. Therefore, using data obtained from bulk material samples for the construction of thermomechanical models of very small scale solder interconnections is likely to result in significant, intrinsic errors.     

Creep behavior of the ternary 95.5Sn-3.9Ag-0.6Cu solder: Part II—Aged condition

Compression creep tests were performed on the 95.5Sn-3.9Ag-0.6Cu (wt.%) solder. The specimens were aged prior to testing at 125°C, 24 h or 150°C, 24 h. Applied stresses were 2–40 MPa. Test temperatures were −25°C to 160°C. The 125°C, 24-h aging treatment caused the formation of coarsened Ag₃Sn particle boundaries within the larger ternary-eutectic regions. The 150°C, 24-h aging treatment resulted in contiguous Ag₃Sn boundaries in the ternary-eutectic regions as well as a general coarsening of Ag₃Sn particles. The 125°C, 24-h aging treatment had only a small effect on the strain-time curves vis-à-vis the as-cast condition. Negative creep was observed at 75°C for time periods >10⁵ sec and stresses of 3–10 MPa. The creep kinetics exhibited a sinh term (stress) exponent, p, of 5.3±0.6 and an apparent-activation energy, ΔH, of 49±5 kJ/mol when data from all test temperatures were included. A good data correlation was observed over the [−25–125°C] temperature regime. Steady-state creep kinetics exhibited a greater variability in the [125–160°C] regime because of the simultaneous coarsening of Ag₃Sn particles. The aging treatment of 150°C for 24 h resulted in a more consistent stress dependence and better reproducibility of the creep curves. Negative creep was observed in samples aged at 150°C for 24 h when tested at −25°C, 25°C, and 75°C. The values of p and ΔH were 4.9±0.3 kJ/mol and 6±5 kJ/mol, respectively. Only a slight improvement in the data correlation was observed when the analysis examined separated [−25−75°C] and [75–166°C] temperature regimes. Creep testing did not cause observable deformation in any of the sample microstructures.     

Creep behavior of the ternary 95.5Sn-3.9Ag-0.6Cu solder—Part I: As-cast condition

The compression creep behavior was studied for the ternary solder alloy 95.5Sn-3.9Ag-0.6Cu in the as-cast condition. Samples were tested under stresses of 2–45 MPa and temperatures of −25–160°C. There was a significant variability in the creep curve shape, strain magnitude, and steady-state strain-rate properties. A multivariable linear-regression analysis of the steady-state strain-rate data, using the sinh-law stress representation, indicated two mechanisms distinguished by low- and high-temperature regimes of −25–75°C and 75–160°C, respectively. The sinh-law stress exponent (n) and apparent-activation energy (ΔH) in the −25–75°C regime were 4.4 ± 0.7 kJ/mol and 25 ± 7 kJ/mol (63% confidence intervals), respectively. Those same parameters in the 75–160°C regime were 5.2±0.8 kJ/mol and 95±14 kJ/mol, respectively, for the high-temperature regime. The values of ΔH suggested a short-circuit diffusion mechanism at low temperatures and a lattice or bulk-diffusion mechanism at high temperatures. The stress dependency of the steady-state strain rate did not indicate a strong power-law breakdown behavior or a threshold stress phenomenon. Cracks and grain-boundary sliding were not observed in any of the samples. As the creep temperature increased, a coarsened particle boundary and particle depletion zone formed in the region of fine Ag₃Sn particles that existed between the Sn-rich phase areas. The coarsened particle boundary, as well as accelerated coarsening of Ag₃Sn particles, were direct consequences of the creep deformation process.     

Solid-state intermetallic compound layer growth between copper and 95.5Sn-3.9Ag-0.6Cu solder

Long-term, solid-state intermetallic compound (IMC) layer growth was examined in 95.5Sn-3.9Ag-0.6Cu (wt.%)/copper (Cu) couples. Aging temperatures and times ranged from 70°C to 205°C and from 1 day to 400 days, respectively. The IMC layer thicknesses and compositions were compared to those investigated in 96.5Sn-3.5Ag/Cu, 95.5Sn-0.5Ag-4.0Cu/Cu, and 100Sn/Cu couples. The nominal Cu₃Sn and Cu₆Sn₅ stoichiometries were observed. The Cu₃Sn layer accounted for 0.4–0.6 of the total IMC layer thickness. The 95.5Sn-3.9Ag-0.6Cu/Cu couples exhibited porosity development at the Cu₃Sn/Cu interface and in the Cu₃Sn layer as well as localized “plumes” of accelerated Cu₃Sn growth into the Cu substrate when aged at 205°C and t>150 days. An excess of 3–5at.%Cu in the near-interface solder field likely contributed to IMC layer growth. The growth kinetics of the IMC layer in 95.5Sn-3.9Ag-0.6Cu/Cu couples were described by the equation x=x_o+Atⁿexp [−ΔH/RT]. The time exponents, n, were 0.56±0.06, 0.54±0.07, and 0.58±0.07 for the Cu₃Sn layer, the Cu₆Sn₅, and the total layer, respectively, indicating a diffusion-based mechanism. The apparent-activation energies (ΔH) were Cu₃Sn layer: 50±6 kJ/mol; Cu₆Sn₅ layer: 44±4 kJ/mol; and total layer: 50±4 kJ/mol, which suggested a fast-diffusion path along grain boundaries. The kinetics of Cu₃Sn growth were sensitive to the Pb-free solder composition while those of Cu₆Sn₅ layer growth were not so.     

The solid solubility of Ag and Cu in the Sn phase of eutectic and near-eutectic Sn-Ag-Cu solder alloys

Early studies of Ag-Sn and Cu-Sn binary alloys showed very low values, 0.04 wt.% for Ag and 0.0063 wt.% for Cu, for the solid solubility of these elements in Sn at the eutectic temperature. In recent work on “as-cast” Sn-Ag-Cu solder alloys, much higher values have been reported for the Ag and Cu content of the Sn phase. In the present study, wavelength dispersive x-ray microprobe measurements made on a near-equilibrium sample confirmed the earlier solubility values. It was concluded that higher values, some of which are reported in the current paper, represent nonequilibrium, supersaturated solid solutions.     

The anomalous microstructural,tensile, and aging response of thin-cast Sn3.9Ag0.6Cu lead-free solder

In this study, bulk and thin-cast samples were produced with an identical Sn3.9Ag0.6Cu composition. The thin-cast material exhibited a much finer as-quenched microstructure than the bulk material with the intermetallic compound (IMC) phase restricted to a thin network. Both the bulk and thincast materials continually softened during room-temperature aging, while both materials initially softened and then subsequently hardened when aged at 120°C and 180°C. The thin-cast material was in all cases significantly softer than the bulk material, and responded to aging as if it were bulk material aged at a higher temperature. These results have significant implications for the elevated temperature application of Sn3.9Ag0.6Cu.     

Time-independent mechanical and physical properties of the ternary 95.5Sn-3.9Ag-0.6Cu solder

The development of a constitutive model for predicting the thermal-mechanical fatigue (TMF) of 95.5Sn-3.9Ag-0.6Cu (wt.%) Pb-free solder interconnects requires the measurement of time-independent mechanical and physical properties. Yield stress was measured over the temperature range of −25–160°C using strain rates of 4.2 × 10⁻⁵ s⁻¹ and 8.3 × 10⁻⁴ s⁻¹. The yield-stress values ranged from approximately 40 MPa at −25°C to 10 MPa at 160°C for tests performed at 4.2 × 10⁻⁵ s⁻¹. The faster strain rate and specimen aging had a limited impact on the yield stress. The true stress/true strain curves indicated that dynamic-recovery and dynamic-recrystallization processes took place in as-cast samples exposed to temperatures of 125°C and 160°C, respectively, while tested at a strain rate of 4.2 × 10⁻⁵ s⁻¹. Aging the sample prior to testing, as well as a faster strain rate, mitigated both phenomena. Dynamic Young’s modulus values ranged from 55 GPa at −50°C to 35 GPa at 200°C, while the coefficient of thermal expansion (CTE) increased from approximately 12 × 10⁻⁶°C⁻¹ to 24 × 10⁻⁶°C⁻¹ for the same temperature range. The aging treatment had little effect on either Young’s modulus or the CTE.     

低银系无铅焊料

概述了低银系无铅焊料的开发和特性。     

Zn-Al-Mg-Ga alloys as Pb-free solder for die-attaching use

Zn-based alloys have been investigated to replace Pb-5%Sn solder for die-attaching use. We have found that a Zn-4%Al-3%Mg-3%Ga alloy has a 309°C solidus and a 347°C liquidus. A die-attaching test was done with preforms of this alloy, Ag-plated lead-frames, and Au-plated dummy dies. Good die-attaching with a small amount of voids can be achieved at 320°C or higher. In subsequent reliability tests, no failure was observed until 1000 cycles between −65°C and 150°C or until 1000 h at 85°C and 85% humidity. Although the poor workability and poor ability of stress relaxation at room temperature of this alloy may somewhat limit its application areas, this solder is the first Pb-free solder for die-attaching use to our knowledge.     

Creep-constitutive behavior of Sn-3.8Ag-0.7Cu solder using an internal stress approach

The experimental tensile creep deformation of bulk Sn-3.8Ag-0.7Cu solder at temperatures between 263 K and 398 K, covering lifetimes up to 3,500 h, has been rationalized using constitutive equations that incorporate structure-related internal state variables. Primary creep is accounted for using an evolving internal back stress, due to the interaction between the soft matrix phase and a more creep-resistant particle phase. Steady-state creep is incorporated using a conventional power law, modified to include the steady-state value of internal stress. It is demonstrated that the observed behavior is well-fitted using creep constants for pure tin in the modified creep power law. A preliminary analysis of damage-induced tertiary creep is also presented.     

Effects of Cu contents in Pb-free solder alloys on interfacial reactions and bump reliability of Pb-free solder bumps on electroless Ni-P under-bump metallurgy

Using the screen-printed solder-bumping technique on the electroless plated Ni-P under-bump metallurgy (UBM) is potentially a good method because of cost effectiveness. As SnAgCu Pb-free solders become popular, demands for understanding of interfacial reactions between electroless Ni-P UBMs and Cu-containing Pb-free solder bumps are increasing. It was found that typical Ni-Sn reactions between the electroless Ni-P UBM and Sn-based solders were substantially changed by adding small amounts of Cu in Sn-based Pb-free solder alloys. In Cu-containing solder bumps, the (Cu,Ni)₆Sn₅ phase formed during initial reflow, followed by (Ni,Cu)₃Sn₄ phase formation during further reflow and aging. The Sn3.5Ag solder bumps showed a much faster electroless Ni-P UBM consumption rate than Cu-containing solder bumps: Sn4.0Ag0.5Cu and Sn0.7Cu. The initial formation of the (Cu,Ni)₆Sn₅ phase in SnAgCu and SnCu solders significantly reduced the consumption of the Ni-P UBM. The more Cu-containing solder showed slower consumption rate of the Ni-P UBM than the less Cu-containing solder below 300°C heat treatments. The growth rate of the (Cu,Ni)₆Sn₅ intermetallic compound (IMC) should be determined by substitution of Ni atoms into the Cu sublattice in the solid (Cu,Ni)₆Sn₅ IMC. The Cu contents in solder alloys only affected the total amount of the (Cu,Ni)₆Sn₅ IMC. More Cu-containing solders were recommended to reduce consumption of the Ni-based UBM. In addition, bump shear strength and failure analysis were performed using bump shear test.     

Electromigration studies on Sn(Cu) alloy lines

The Cu alloying effect in the Sn(Cu) solder line has been studied. The Sn0.7Cu solder line has the most serious electromigration (EM) damage compared to pure Sn and Sn3.0Cu solder lines. The dominant factor for the fast EM rate in Sn0.7Cu could be attributed to the relatively small grain size and the low critical stress, i.e., the yielding stress of the Sn0.7Cu solder line. Also, we found that the shortest Sn0.7Cu solder line, 250 μm, has the most serious EM damage among three solder lines of different lengths. The back stress induced by EM might not play a significant role on the EM test of long solder lines. A new failure mode of EM test was observed; EM under an external tensile stress. The external stress is superimposed on the stress profile induced by EM. As a result, the hillock formation was retarded at the anode side, and void formation was enhanced at the cathode.     

Effect of Cu additives on Sn whisker formation of Sn(Cu) finishes

Sn whisker formation on Sn(Cu) finishes has been studied. (1) With respect to the thickness effect, we found that Sn whisker density for pure Sn and Sn0.7Cu finishes has a linear relationship with the finish thickness. The safety thickness for Sn and Sn0.7Cu finishes is about 10 μm and 20 μm, respectively. (2) With respect to the alloying effect, we found that Sn whisker formation could be retarded by increasing Cu content in the Sn(Cu) finishes. We conclude that the Cu additives could reduce the two major driving forces of the Sn whisker formation, i.e., metal underlayer dissolution and thermal stress. The Cu additives self-formed a Cu-Sn compound barrier layer, which effectively prevents the reaction and dissolution with the metal underlayer. On the other hand, the Cu additives precipitated out as Cu-Sn compound in the Sn(Cu) finish layer, which is believed to be the reason for smaller values of the coefficient of thermal expansion (CTE) for Sn(Cu) alloys. The smaller CTE values results in a lower thermal stress level in the Sn(Cu) finishes.     

Influence of Au addition on the phase equilibria of near-eutectic Sn-3.8Ag-0.7Cu Pb-free solder alloy

This paper illustrates the influence of Au addition on the phase equilibria of Sn-Ag-Cu (SAC) near-eutectic alloys and on the interface reaction with the Cu substrate. From the thermal and microstructural characterization of Sn-3.8Ag-0.7Cu alloys containing various amounts of Au, it is found that the Au promotes the formation of a quaternary-eutectic reaction at 204.5°C ± 0.3°C. The equilibrium phases in the quaternary-eutectic microstructure are found to be AuSn₄, Ag₃Sn, βSn, and Cu₆Sn₅. While the addition of Au to Sn-3.8Ag-0.7Cu alloys is also found to increase liquidus temperature and the temperature ranges of the phase equilibria field for primary phases, such influences from Au are found to be less pronounced when the alloys were reacted with the Cu substrate. Because of the formation of the Au-Cu-Sn-ternary interface intermetallic, it is found that a majority of Au added to the solder is drained from the melt. The drainage of Au reduces the impact of Au on the phase equilibria of the solder alloys in the joint. It is further found that the involvement of Au in the interface reaction results in a change of the interface phase morphology from the conventional scallop structure to a compositelike structure consisting of (AuCu)₆Sn₅ grains and finely dispersed, βSn islands.     

镧对Sn3.5Ag0.5Cu钎料组织性能影响

运用莱卡显微镜、扫描电镜和能谱分析等手段,研究了稀土元素La的添加量对Sn3.5Ag0.5Cu钎料及其与Cu基体焊接后微观组织及性能的影响。结果表明:添加不同含量的稀土La均能使钎料及其与Cu基体焊接后组织与性能得到改善,其中以w(La)达到0.05%时为最优,显微硬度及剪切强度分别提高14%和10.7%。键参数函数计算结果表明La具有"亲Sn"倾向,可细化钎料组织,降低IMC(界面金属间化合物)的长大驱动力。     

Correlation between interfacial reactions and mechanical strengths of Sn(Cu)/Ni(P) solder bumps

The correlation between interfacial reactions and mechanical strengths of Sn(Cu)/Ni(P) solder bumps has been studied. Upon solid-state aging, a diffusion-controlled process was observed for the interfacial Ni-Sn compound formation of the Sn/Ni(P) reaction couple and the activation energy is calculated to be 42 KJ/mol. For the Sn0.7Cu/Ni(P), in the initial aging, a needle-shaped Ni-Sn compound layer formed on Ni(P). Then, it was gradually covered by a layer of the Cu-Sn compound in the later aging process. Hence, a mixture layer of Ni-Sn and Cu-Sn compounds formed at the interface. For the Sn3.0Cu/Ni(P), a thick Cu-Sn compound layer quickly formed on Ni(P), which retarded the Ni-Sn compound formation and resulted in a distinct Cu-Sn compound/Ni(P) interface. The shear test results show that the mixture interface of Sn0.7Cu bumps have fair shear strengths against the aging process. In contrast, the distinct Cu-Sn/Ni(P) interface of Sn3.0Cu solder bumps is relatively weak and exhibits poor resistance against the aging process. Upon the reflowing process, the gap formation at the Ni(P)/Cu interface caused a fast degradation in the interfacial strength for Sn solder bumps. For Sn0.7Cu and Sn3.0Cu solder bumps, Ni₃P formation was greatly retarded by the self-formed Cu-Sn compound layer. Therefore, Sn(Cu) solder bumps show better shear strengths over the Sn solder bump.     

Microstructures in solder joints between Sn95.5Ag4Cu0.5 solder and Ag/Pd thick film

A joint between Sn95.5Ag4Cu0.5 (mass%) solder and an Ag/Pd thick film was soldered by dipping at 260°C for 3–30 sec. Shrinkage voids and Sn grain growth were characterized as well as their transformation kinetics. Void shrinkage occurred in the zone near the top surface of the joint. Shrinkage was always accompanied by colonies of ternary/quaternary meta-eutectic that were the regions solidified last in the joint. The Sn grains accumulated into two bands across the joint during solidification: one was transverse through the thickness and the other was parallel to the solder pad.     

Solderability testing of 95.5Sn-3.9Ag-0.6Cu solder on oxygen-free high-conductivity copper and Au-Ni-Plated kovar

The solderability of 95.5Sn-3.9Ag-0.6Cu solder on oxygen-free high-conductivity (OFHC) copper and Au-Ni-plated Kovar was examined as a function of flux and process temperature. The three solder fluxes included a rosin-based (R) material, a rosin mildly activated (RMA) flux, and a low-solids (LS) flux. The solderability metric was the contact angle, θ_C, measured by a meniscometer and wetting-balance techniques. The wetting rate and time to maximum force parameters were also documented. In most cases, the contact angles for the 95.5Sn-3.9Ag-0.6Cu solder alloy, regardless of the type of flux or temperature, were higher then those for the 63Sn-37Pb eutectic-solder alloy, indicating a less “solderable” surface.     

Evaluation of the mechanical properties of a ternary Sn-20In-2.8Ag solder

The Sn-20In-2.8Ag solder alloy is a potential lead-free solder for replacing the traditional Sn-Pb solders. In this study, the mechanical properties of the bulk material are reported by tensile test at various strain rates and temperatures. The Sn-20In-2.8Ag solder possessed a solidus and liquidus between 170.8°C and 195.5°C. The ultimate tensile strength (UTS) and elongation were 59.3 MPa and 50.2% at a strain rate of 10⁻³ s⁻¹ at room temperature. Moreover, the UTS of this alloy decreased, but its elongation increased, with increasing testing temperature. Stress exponents of Sn-20In-2.8Ag alloy varied from 6.5 at room temperature to 4 at 100°C, and the activation energy for creep was 51.0 kJ/mol at the higher temperature range from 50°C to 100°C. The typical intergranular creep fracture mode was observed in Sn-20In-2.8Ag solder during tensile creep deformation.