Lead-Free Brings Out Problem of Brittle Fracture

Though it has served the industry well, traditional solder joint integrity tests, such as solder ball shear and pull testing, do a poor job of consistently detecting microstructural weaknesses at the pad/solder interface, particularly those associated with a failure mode called brittle fracture failures. With the introduction of lead-free solders, brittle-fracture failures became more pronounced because of the composition of the solder and surface finishes. As a result, new equipment and test methodologies were needed.

In most tests, the solder ball itself is either sheared off or pulled to destruction (Fig. 1 ). This shows that the bond is at least as strong as the applied test force, but the test does not provide the actual ball-to-pad bond strength. Conventional shear and pull testing are performed in the range of 100-600 µm/sec and 500-1000 µm/sec, respectively. With current tests, brittle-fracture failures rarely occur. Though brittle-fracture failures do happen with SnPb solder, the ductility of SnPb solders means that these failures rarely occur. Later in the process, drop tests can be performed on the completed electronic assembly. However, such testing requires setup and fixturing, and occurs late in the manufacturing process.

Dage Precision Industries (Fremont, Calif.) and Sun Microsystems (Santa Clara, Calif.) worked together with a consortium of many of the major ball grid array (BGA) device manufacturers to develop a new bond test method that would qualify joint integrity for brittle fracture early in the manufacturing process. They decided to explore high-speed testing because the strength of the joint of both leaded and lead-free solder increases with increasing strain: At high test speeds, a higher force is transferred from the solder ball to the bond between the ball and pad interconnect, resulting in a test that examines bond reliability. In an 18-month project, Dage developed new bond test equipment and evaluated a variety of test and setup requirements, including shear testing and pull testing, test speed, peak force, impact energy absorbed at the bond, etc. This flexibility allowed the user to test under different conditions and determine the setup best suited to the application. Such testing allows a comparison of the reliability of different pad finish and solder alloy combinations.

1. Standard shear testing often results in shearing of the solder ball itself (left). A successful test for brittle fracture separates the ball from the pad (right) and can be used to improve product reliability.

Shear testing at high speeds requires an acceleration distance before contacting the ball. Samples were prepared by clearing all but the two outside perpendicular rows on the substrate. Shear testing involves alignment of the tool to ball, completing a normal land, completing a step back to the programmed shear height; moving the sample from tool to acceleration distance, accelerating the sample to programmed test speed; and contacting the ball to tool. Speed is held constant during impact.

Pull testing uses a similar alignment process. In this five-step process, first the jaws of the tester lower onto and grip the ball. They continue to descend, pushing the sample down against a spring, creating the acceleration distance. The jaws then accelerate upward to the test velocity, which it reaches at the starting height of the sample. Sample travel is restricted by a rigid, abrupt stop, and the jaws continue at test speed. Finally, the ball is pulled from the sample at the test speed (Fig. 2).

2. In high-speed pull testing, the jaw lowers and grips the ball, the sample is pushed down against a spring to create the acceleration distance, and then it is pulled at the test velocity.

The tests evaluated ball shear characteristics over the range of 100 µm/sec to 4.0 m/sec and pull characteristics from 500 µm/sec to 1.3 m/sec. Pull testing was conducted using a cold bump pull process. The tests determined:

• No significant difference in shear force as a function of tool shape or tool offset.
• Shear and pull strength increased with test speed.
• Interfacial solder joint fracture rate increased with test speed for shear and cold bump pull tests.
• Shorter time between reflow and test resulted in higher frequency of interfacial failure for SnPb solder on ENIG-plated substrates.
• Interfacial failure occurred more often for SnPb solder on ENIG-plated package substrates than on NiAu or bare copper-plated substrates.

Dage developed these tests for its recently introduced 4000HS high-speed bond tester. The industry is working toward developing standard tests for high-speed shear and pull.