Testing for BGA Resistance to Brittle Fracture
INDIVIDUAL BONDS CAN BE TESTED IMMEDIATELY AFTER REFLOW
BY BOB SYKES
Solder ball joint integrity has always been an issue, but never more so than with the introduction of lead-free solder. Research has shown that lead-free joints can be particularly susceptible to brittle fracture at the ball-to-pad interface. These failures can occur over the full life cycle of a solder joint, from manufacture to test and handling, and through its life as an end product.
With the introduction of lead-free solder alloys, brittle fracture failures have increased. Many electronics industry assemblers and OEMs agree that urgent attention is required to address the issue. Though the change to lead-free solders has brought more attention to the brittle fracture failure mode, it is a demonstrated reliability concern with many other solder alloys and pad surface finishes.
The problem with current shear and pull bond testing methodology is that the brittle fracture failure mode is a rare occurrence. It’s not because brittle fractures don’t occur in PbSn solders, it is because current test regimes are not capable of consistently applying force in a manner to demonstrate the failure mode.
Figure 1. Typical solder shear.
In most tests, the solder ball itself is either sheared off or pulled to destruction (Figure 1). This proves that the bond is at least as strong as the applied test force, but does not provide the actual ball-to-pad bond strength. Though conventional tests still have their place in checking for many manufacturing defects, they are not well-suited to testing for the brittle fracture failure mode (Figure 2).
Figure 2. Typical brittle fracture.
It is known that the strength of solder increases with strain rate. The same is true of a completed product subjected to a rapid load change, such as when a product is dropped.
Conventional bond testing is limited to low speeds, typically a few mm/sec. To test for resistance to brittle fracture, the test speed must be much higher, in the order of several m/sec. High test speeds are often referred to as impact tests, where the load on the bond rises in fractions of a millisecond. In an impact test, the solder ball is stronger and more of the load is transmitted to the bond. It stands to reason that problems occurring under impact or high strain rates should be tested by similar means, and it has been clearly shown that testing at high speed can produce many more brittle fractures and provide data that can be used to improve product reliability.
Testing at high strain rates is not a new test methodology. Drop testing, for example, has been used for many years. Although the drop test will remain a useful and definitive test, it can only be accomplished with a completed assembly and requires a great deal of set-up and fixturing. It was recognized that a test similar to conventional ball shear and pull that could qualify joint integrity to brittle fracture early in the manufacturing process was required. To this end, a research project* was expanded to include a consortium of many of the major manufacturers of BGA devices. The project lasted more than 18 months and resulted in a new bond test regime and approach to bond test equipment.
Brittle Fracture Test Equipment
The new test regime requires a new approach to bond testing and a new bond test machine capable of testing the BGA solder joint resistance to the brittle fracture failure mode.
The science of high-speed bond testing is new, and the test regime necessitates a wide range of test and setup requirements: shear testing and pull testing, test speed, peak force, impact energy absorbed at the bond, etc. This enables the user to test under different test conditions and find the setup best suited to the application.
High-speed Test Regime
Shear testing at speed requires an area in which the tool can accelerate (acceleration distance) before contacting the ball. To make this possible, the sample under test must be prepared for the test by clearing all but two rows of solder balls as shown in Figure 3.
Figure 3. High-speed test regime.
Shear Test Procedure:
- Align the tool to the ball in the conventional manner (Figure 4);
- Initiate the test
- complete a normal land
- complete a step-back to the programmed shear height;
- The sample must then move away from the tool to create the acceleration distance;
- From this position, the sample is accelerated to the programmed test speed and the ball contacts the tool. Having a region of constant velocity (acceleration distance), the speed is held prior to and during impact with the ball;
- Decelerate the sample to complete the test.
Figure 4. Shear test procedure.
Pull testing at high speed requires a different setup than conventional pull testing. Because the clamping jaws must have contact with the ball, an acceleration distance between the jaws and the sample is not possible.
Pull Test Procedure:
- In this test, the alignment is again identical to conventional low-speed testing (Figure 5);
- When initiating the test, the jaws descend and grip the ball;
- They then continue to descend by pushing the sample down against a spring, creating the acceleration distance;
- The jaws then accelerate upward to the test velocity that is held constant through the starting height of the sample;
- At the starting height, the sample travel is restricted by a rigid abrupt stop and the jaws continue at the test speed;
- The ball is then pulled from the sample at the test speed established during setup.
Figure 5. Pull test procedure.
It is very likely that we have much to learn and that this new technology will continue to evolve, leading the way to better understanding and improvements in both the testing regime and the equipment. However, initial results are encouraging and data collected to date shows:
- Many more brittle fractures occur at high test speed than at conventional speeds;
- Lead-free solder exhibits more brittle fracture failures when tested at high speed than traditional lead solder;
- SnPb solder on ENIG plating exhibits more brittle fracture failures than on NiAu or bare Cu pads;
- SnPb and SnPbAg solders perform similarly in both high-speed shear and pull.
Brittle fracture is a problem affecting many manufacturers. Until now, qualifying solder joint integrity has been difficult, unreliable, and time-consuming. Using the new test regime, individual bonds can be tested immediately after reflow under a range of consistent speeds and loading conditions - providing accurate performance data on solder joint resistance to brittle fracture.
* Founded by Sun Microsystems and Dage Precision Industries.
BOB SYKES, design manager, Dage Precision Industries, Rabans Lane, Aylesbury, Buckinghamshire HP 19 3 RG, England; 44 0 1296 317800.