Will your BGA fall off ?
07 February 2006
BIf a board is dropped, is the BGA likely to fall off? This extract from a paper by Keith Newman looks at research into 'BGA Brittle Fracture' and methods to test for it.
The many package configurations varied in substrate plating type (exposed Cu, electrolytic NiAu and electroless-Ni/immersion- Au), solder composition (SnPb, SnPbAg and SnAgCu), packaging construction and materials, solder and package geometries, package assembly location, and time between reflow and test. First, some background: Unlike fatigue failures, BGA brittle solder joint fractures typically occur during a monotonic stress event such as drop or high strain rate flexure. These stress events can occur during board assembly/test, shipment/handling operations, or actual end-use.
Recent investigations have documented voiding at the pad/solder interface for various plating/solder material combinations. The interfacial voids appear to relate to observed brittle solder joint fractures. The primary industry method to assess solder joint integrity, solder ball shear testing, does not consistently detect microstructural weaknesses at pad /solder interfaces that are associated with brittle failures.
Solder ball shear/pull strength (determined using conventional test equipment and procedures) appears to relate to solder cell (grain) coarsening but not to brittle fracture resistance. The solder ball shear/pull failure mode may relate to brittle fracture resistance but appears to be dependent on strain rate and time-after-reflow (time between solder ball reflow and shear/pull testing).
The strength characteristics of SnPb and Pb-free solder alloys are highly strain-rate dependent. Although the ultimate strength of solder typically improves with increased strain-rate, the higher plastic modulus at elevated strain levels can result in increased interfacial strain at the pad/solder region under flexural or impact loading conditions.
Alternative tests
Alternative solder joint integrity test methods, including drop, shock and high-speed monotonic bend, approximate actual conditions associated with BGA solder joint brittle fracture, and can yield more accurate assessments of brittle fracture resistance. These alternative characterisation methods provide useful tools for component manufacturers to optimise the brittle fracture resistance of their product, but they are an impractical replacement for in-line solder joint integrity monitoring. Thus, alternative assessment methods, high speed solder ball shear and pull, were investigated.
With respect to the test apparatus in general, several early prototype testing machines were developed by equipment manufacturer, Dage Precision, to provide the desired range of solder ball shear and pull test speeds.
The conclusions of the testing were revealing. Solder ball shear characteristics were evaluated over shear rates ranging from 0.0001 m/s (100 æm/s) to 4.0 m/s. Also, solder ball pull characteristics were assessed over pull speeds ranging from 0.0005 m/s (500 æm/s) to 1.3 m/s. Pull testing was conducted using a ''cold'' process and a ''hot'' process.
Although a less common solder joint integrity method than solder ball shear, pull testing using a tweezer-like clamp that grips an individual solder ball at room or ''cold'' temperature has been available for years. A high-speed cold pull tester was developed to achieve a maximum test speed of 1.3 m/s, several orders of magnitude higher than a typical pull speed of 0.00004-0.005 m/s (40-5000 æm/s) using conventional equipment. The package is positioned against the gripping head, and an individual solder ball is clamped. For high-speed pull testing, the package, gripping head and support platform are then lowered in the vertical direction to provide the necessary acceleration travel distance to achieve the desired pull speed at impact.
An alternative method for performing solder ball pull testing employs a heated copper pull wire or electrode, that is solder-attached to the solder ball. The electrode is initially heated well above the solder liquidus temperature, then attached to the solder ball by penetrating the copper tip into the solder ball and then cooling the electrode. This ''hot'' attachment technique has some advantages over the ''cold'' process: it avoids the need of customised clamping grips for each solder ball diameter, and eliminates the deformation of the solder ball owing to mechanical clamping. Disadvantages include potential alteration of solder morphology and intermetallic structure at the solder/package interface, and increased test time/cost.
Of particular interest in this study was whether interfacial fracture between the solder and pad would occur more frequently with increased shear/pull test speeds that are perhaps more commensurate with actual drop, impact or bend-loading conditions. Given the thousands of assessed failures, only simplified failure mode classifications were used.
Selected conclusions from the study include the fact that no significant differences in solder ball shear force as a function of tool shape (standard/cavity) or tool offset (5-50 æm) was detected among the many tested package configurations. A close examination of the source data indicates the slight differences in failure mode between standard and cavity tool shapes at elevated shear speeds appear dependent on test variables other than the tool type.
Another conclusion was shear-and-pull strength typically increased with increased test speed over the useful force transducer range. Since related studies on tensile and compressive loading of solder alloys also indicate a general increase in ultimate strength with increased strain rate, the observed increase in shear strength with increased shear speed is presumed to reflect the dynamic material properties of solder rather than solely an artifact of the solder ball geometry and shearing process itself.
Interfacial solder joint fracture rate generally increased with increased test speed for shear and cold pull test methods, while shortened time between reflow and test typically resulted in a higher frequency of interfacial failure for SnPb solder on ENIG plated substrates.
Other conclusions were that high interfacial fracture rates at lower test speeds were observed for hot pull testing (SnPb and SnAgCu solder), and for cold pull testing of short time-after-reflow samples (SnPb and SnPbAg solder over ENIG). SnAgCu solder balls demonstrated a higher interfacial fracture rate at high-test speeds than SnPb/SnPbAg solder balls.
Then there was the fact interfacial failure occurred more frequently for SnPb solder on ENIG plated package substrates than on either electrolytic NiAu or bare Cu plated substrates.
Finally, SnPb and SnPbAg solder joints performed nearly identical in shear/pull testing. The SnPb/ENIG hot pull samples displayed a high interfacial failure rate across all tested pull speeds. Whether the high interfacial failure mode incidence correlates to brittle fracture resistance has not been confirmed, but the use of hot-pull testing using conventional test equipment potentially appears to be an improved solder joint integrity test method.
Although correlation between interfacial failure rate and brittle fracture resistance has not been confirmed, higher shear/pull test speeds and shorter time between reflow and test may yield improved brittle fracture solder joint integrity test methods. Cold pull (with short time-after-reflow) and hot pull using conventional test equipment may offer improved methods for assessment of brittle fracture resistance.
Contact Details and Archive...
Related Articles...
Most Viewed Articles...