1. Mechanical characterization of lead- free solder joints J. Cugnoni*, A. Mellal*, Th. Rütti @, J. Janczak @, Pr. J. Botsis* * LMAF / EPFL ; @ EMPA Switzerland Project funded by OFES (CH) Cost 531 WG 5 & 6 Meeting, Vienna 17.01.05

2. LMAF / EPFL Objectives and tasks Objectives: identify the nature of irreversible deformation and damage; identify the nature of irreversible deformation and damage; correlate the role of micro structure on the deformation and damage mechanisms correlate the role of micro structure on the deformation and damage mechanisms examine the role of interface on deformation and damage of a joint; examine the role of interface on deformation and damage of a joint; identify appropriate constitutive equations; identify appropriate constitutive equations; characterise the role of the thermo-mechanical loading histories on the constitutive behaviour of the material and durability of various joints; characterise the role of the thermo-mechanical loading histories on the constitutive behaviour of the material and durability of various joints; compare the results with those of the standard alloy (Sn63Pb37). compare the results with those of the standard alloy (Sn63Pb37).Tasks design of experiments design of experiments optical strain field measurement optical strain field measurement observation of microstructural effects observation of microstructural effects identify constitutive laws for the lead-free alloy identify constitutive laws for the lead-free alloy construct numerical models construct numerical models comparison and validation comparison and validation

3. LMAF / EPFL Mechanical characterization The elasto-plastic constitutive law may depend on: strain rate and temperature strain rate and temperature microstructure and thermal history (processing / ageing) microstructure and thermal history (processing / ageing) geometrical / mechanical constraints geometrical / mechanical constraints characteristic size and scale effects characteristic size and scale effectsCharacterization: should be carried out on real solder joints should be carried out on real solder joints temperature, strain rate and joint thickness are independent parameters and must be changed temperature, strain rate and joint thickness are independent parameters and must be changed a correlation between thermal history, microstructure and constitutive behaviour must be found a correlation between thermal history, microstructure and constitutive behaviour must be found

4. LMAF / EPFL Lead-free solder joints specimens Specimen specifications Dimension: 120 x 20 x 1 mm, joint thickness from 0.1 to 1 mm Dimension: 120 x 20 x 1 mm, joint thickness from 0.1 to 1 mm Solder: ECOREL Sn-4.0Ag-0.5Cu Solder: ECOREL Sn-4.0Ag-0.5CuProduction: joint cast in a special jig joint cast in a special jig temperature cycle: heated at 40 K/min up to melting point, held 60s in liquid phase, and then rapid cooling of the jig (water). temperature cycle: heated at 40 K/min up to melting point, held 60s in liquid phase, and then rapid cooling of the jig (water).

5. LMAF / EPFL Mechanical testing Mechanical testing: displacement control, 1  m/s up to rupture displacement control, 1  m/s up to rupture 50 mm extensometer => average strain in the specimen 50 mm extensometer => average strain in the specimen Effects of the joint thickness on mechanical properties decreased solder gap width increases yield and tensile strengths and decreases strain (ductility) decreased solder gap width increases yield and tensile strengths and decreases strain (ductility) large scatter probably mostly due to gas porosity and the averaging effect of the strain measurements large scatter probably mostly due to gas porosity and the averaging effect of the strain measurements Solder gap width Yield strength Tensile strength Young’s modulus Strain at fracture [µm][MPa] [GPa][%] 181 41.8 ± 0.1 42.9 ± 3.3114.4 ± 13.10.042% ± 0.007% 204 40.9 ± 3.0 44.5 ± 2.7100.6 ± 4.50.050% ± 0.001% 395 36.3 ± 5.6 42.4 ± 6.1107.6 ± 9.10.048% ± 0.004% 526 34.7 ± 4.2 41.0 ± 3.7101.8 ± 6.30.056% ± 0.004% 61129.4 ± 1.242.5 ± 0.595.3 ± 7.60.078% ± 0.008% 79539.6 ± 5.646.9 ± 5.492.7 ± 4.70.072% ± 0.015% 93539.3 ± 1.149.5 ± 1.5105.7 ± 9.00.077% ± 0.018% 110735.8 ± 1.346.4 ± 1.390.8 ± 19.30.076% ± 0.005%

6. LMAF / EPFL Ageing Test matrix effect of solder gap width: 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.8 and 1.0mm effect of solder gap width: 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.8 and 1.0mm effect of room temperature ageing (T H = 0.6): 1day, 2 days, 1 week, 2 weeks, 1 month, 2 months effect of room temperature ageing (T H = 0.6): 1day, 2 days, 1 week, 2 weeks, 1 month, 2 months effect of ageing at elevated temperatures: 1 week and 2 weeks at T H =0.75 and T H = 0.9 effect of ageing at elevated temperatures: 1 week and 2 weeks at T H =0.75 and T H = 0.9 Effects of ageing no visible influence of ageing at room temperature no visible influence of ageing at room temperature ageing at high temperatures reduces yield and tensile strengths and increases strain (ductility) ageing at high temperatures reduces yield and tensile strengths and increases strain (ductility)

7. LMAF / EPFL A first modelling approach The elasto-visco-plastic model (Garofalo) of classical lead solders (Shi et al., 1999 ) has been adapted to lead-free solders: yield stress and Young's modulus adjusted for lead-free solders yield stress and Young's modulus adjusted for lead-free solders hardening parameters from the classical lead solders hardening parameters from the classical lead solders Young modulus (GPa) Poisson’s ratio Elastic behavior 560.35 Plasticity Yield stress = 32.5 (MPa) Linear hardening up to rupture: Ultimate stress = 33 (MPa) Ultimate strain = 0.02 (-) Creep behavior A = 96200 (sec -1 ) B = 0.087 (MPa -1 ) n = 3.3 Q = 67437 (J mol -1 ) R=8.314 (J mol -1 K -1 )

8. LMAF / EPFL A first modelling approach Finite element simulation of real experiments to test the "adjusted" constitutive law: modelling of both copper and solder joint modelling of both copper and solder joint real recorded (extensometer) displacements are applied to the FEM => simulated loads real recorded (extensometer) displacements are applied to the FEM => simulated loads Constitutive law shows a good agreement with experiments for thick joints (1mm) but must be improved for thin joints (0.15 mm) Constitutive law shows a good agreement with experiments for thick joints (1mm) but must be improved for thin joints (0.15 mm)

9. LMAF / EPFL Bulk solder properties Preliminary results: specimens of pure solder produced in several ways specimens of pure solder produced in several ways important effects of thermal history and processing important effects of thermal history and processing properties must be characterized "in-situ" properties must be characterized "in-situ"

10. LMAF / EPFL Mechanical characterization of constrained joints Objectives characterize the stress - strain law of lead-free solders in a real joint (constrained) characterize the stress - strain law of lead-free solders in a real joint (constrained) optical strain measurement technique to measure the real strains of the solder only (not the average strains of the joint) optical strain measurement technique to measure the real strains of the solder only (not the average strains of the joint) Optical measurement technique a grid of fine dots (pitch = 0.2 mm) is glued on the surface of the specimen a grid of fine dots (pitch = 0.2 mm) is glued on the surface of the specimen the deformation of the grid is observed with a microscope (24x) and recorded through a high resolution video camera (1.3 MPixels) at 1 fps the deformation of the grid is observed with a microscope (24x) and recorded through a high resolution video camera (1.3 MPixels) at 1 fps video extensometry by motion tracking based on a Normalized Cross Correlation algorithm (NCC) video extensometry by motion tracking based on a Normalized Cross Correlation algorithm (NCC) Resolution: displacement 0.2  m, strain 0.01% Resolution: displacement 0.2  m, strain 0.01%

11. LMAF / EPFL Mechanical characterization of constrained joints Preliminary results: Solder joint properties showing the constraining effects: Solder joint properties showing the constraining effects: Yield stress, ultimate stress and ultimate strain are modified by the constraints Properties must be determined in the most realistic conditions

12. LMAF / EPFL Future work Characterization of the solder Compare the experimental stress-strain curve with the predictions of a FEM based on the bulk solder properties to evaluate the possibility to use directly the bulk solder stress-strain curve in real applications Compare the experimental stress-strain curve with the predictions of a FEM based on the bulk solder properties to evaluate the possibility to use directly the bulk solder stress-strain curve in real applications Identify the elasto-visco-plastic constitutive parameters by a mixed numerical-experimental identification procedure Identify the elasto-visco-plastic constitutive parameters by a mixed numerical-experimental identification procedure at a given strain rate and room temperature, with variable joint thickness (size / constraining effects) at different strain rates and temperatures Microstructure evolution (in collaboration with EMPA, Switzerland) Correlate the mechanical properties with the microstructure of the solder Correlate the mechanical properties with the microstructure of the solder Evaluate the evolution of micro structure and mechanical properties in function of the thermal history Evaluate the evolution of micro structure and mechanical properties in function of the thermal history Improve the mechanical properties by inclusion of strengthening particles Improve the mechanical properties by inclusion of strengthening particles