Although the deadline for RoHS compliance for medical device manufacturers is nearly a year away (July 22, 2014), there is more than ten years of testing and in-service data that can be used to ensure a smooth transition to lead-free medical devices and equipment.
The July 2006 enactment of the European Union’s (EU) Restriction of Hazardous Substances (RoHS) Law corresponded closely with the explosion of personal electronic devices, such as the smartphone. These new devices were subject to a new failure mode: drop shock. In addition, the challenges of lead-free assembly, edicted by RoHS, and the much finer pitches demanded by these small, electronically dense devices were a trial for electronic assemblers.
By the late 2000s, most process engineers had settled on SAC105 (Sn98.5Ag1.0Cu0.5) as the alloy of choice for electronics that required drop shock performance. However, SAC105 was not robust in traditional thermal cycle testing, typically required for large electronics such as personal computers, televisions, etc. In these applications, SAC305 (Sn96.5Ag3.0Cu0.5) was used. SAC305 has admirable thermal cycle performance as compared to SAC105, but is inferior in drop shock performance.
Lead-free soldering has been widely adopted by the electronics industry, with a tin-silver-copper alloy, known as SnAgCu (SAC) and its high silver (Ag) content (approximately 3.0%) being the initial mainstream alloy. This selection was challenged later by the fragility of the solder joints during drop shock testing and the high cost of silver. Low Ag SAC was considered a solution for resolving both issues. However, this approach compromised temperature cycling performance, and therefore was not acceptable for high-end applications in which temperature cycling performance is critical.
To address this need for an alloy that would perform well in both drop shock and thermal cycle testing, a low Ag SAC alloy doped with manganese (Mn), called SACM, was evaluated against eutectic SnPb, SAC105, and SAC305 in the JEDEC drop, dynamic bending, -40/125°C temperature cycling, and 1Hz/2mm cyclic bending tests. SACM is a patent-pending alloy consisting of 0.5 to 1% Ag, 0.5 to 1% Cu, and <0.1% Mn. SACM has a composition similar to SAC105, but its small addition of manganese results in an alloy that provides both superior drop shock and thermal cycle performance. This article will provide an overview of this analysis.
Four solder sphere alloys were evaluated: the new SACM alloy and three controls: Sn63Pb37 (SnPb), Sn98.5Ag1.0Cu0.5 (SAC105), and Sn96.5Ag3.0Cu0.5 (SAC305).
For the JEDEC drop test, thermal cycling test, and cyclic bending test, two no-clean solder pastes were used: SnPb and SAC305. The former was used for SnPb thin-film ball grid array (TFBGA) assemblies, while the latter was for leadfree TFBGA assemblies. For the dynamic bending test, Sn95.5Ag3.8Cu0.7 (SAC387) no-clean solder paste was used. The details of the components, PWBs, and assembly parameters can be found in a paper, “Achieving High Reliability Low- Cost Lead-Free SAC Solder Joints via Mn or Ce Doping,” delivered at the Electronic Components and Technology Conference, in San Diego, CA, in May 2009.
The JEDEC Drop Test (JDT): As shown in Figure 1, the board-level test vehicle was affixed to the drop table at the four corners with the mounted packages facing downward, according to JESD22-B111. The drop table was then released and dropped freely at a certain height to impact on the strike surface repetitively, each time creating a halfsine wave impact acceleration pulse of a peak acceleration of G0 (1500Gs) and duration of τ (0.5 ms).
For TFBGA (NiAu finished solder balls) devices assembled on organic solderability preservative (OSP) pretreated PCBs with 250 cycles of TCT, the JDT performance of various sphere alloys was charted and the reliability can be ranked in the following order: SACM ≥ SnPb > SAC105 > SAC305.
Thermal Cycling Test (TCT): The samples were subjected to TCT (-40°C to 125°C, 42 min/cycle, ramp up/down: 11 min., dwell time 10 min.) with real time resistance monitoring. A failure was defined when a 20% resistance increase was recorded.
For devices with TFBGA (NiAu) assembled on OSP treated PCBs and aged at 150°C for 250 hrs., the TCT performance of various sphere alloys was recorded ranked in this order: SACM ≥ SAC305 > SAC105 > SnPb. Although SAC305 has a good C-Life, the poor showing in the first two failures might be troubling in high-reliability applications.
Cyclic Bending Test (CBT): Nine packages were mounted on a 132mm x 77mm x 1mm standard 8-layer PC board with layout regulated by JESD22B113. The test board and mounted packages were daisy chain designed so that the overall electrical resistance of the daisy chain solder joints could be individually measured in each package. Each cell was subjected to a cyclic bend test, as shown in Figure 2, at a 1 Hz/2mm testing condition until all components failed. The number of cycles were recorded for failed units when resistance exceeded 1,000 ohms.
For TFBGA devices with NiAu finished solder balls assembled on OSP treated PCB pads and aged at 150°C for 250 hrs., the CBT performance of various sphere alloys was charted and ranked in the following order: SAC305 > SACM, SAC105 > SnPb.
Dynamic Bending Test (DBT): A high strain rate drop test was adopted in this work to measure failure behavior of the second level package reliability for mobile applications. This test was found to duplicate the failure mode of solder joints in surface mount devices found in mobile phone drop shock tests.
The test apparatus was composed of the 4-point bending setup shown in Figure 3. A steel ball was dropped from various heights onto the top span fixture to induce various levels of strain in order to control stress levels at the solder joints. The strain gauge was mounted on the back of the PCB, as shown.
The board strain was increased incrementally, and each unit was impacted only once. After the dynamic bending test, joint failure was identified by a dye and pry process. The number of joint failures for each unit was collected for the given board strains, and the data set was fit to a Weibull curve to obtain the board strain level required to generate one solder joint failure.
For devices with TFBGA (NiAu) assembled on OSP treated PCBs, the reliability of the devices on the PCB after reflow can be ranked in the following order: SACM > SAC105 > SAC305. However, for thermally aged (150°C/ 250 hrs.) devices, the performance gap between SACM and SAC105 increases, and the ranking of alloys is altered to: SACM > SAC305 > SAC105.
The Mn-doped SAC alloy, SACM, achieved higher drop test and dynamic bending test reliability than SAC105 and SAC305, and exceeded results for SnPb for many test conditions.
More significantly, SACM matched SAC305 in thermal cycling performance. In other words, SACM achieved better drop test performance than the low Ag SAC alloys plus the desired thermal cycling reliability of high Ag SAC alloys. The mechanism for high drop performance and high thermal cycling reliability can be attributed to a stabilized microstructure with uniform distribution of fine intermetallic compound (IMC) particles, presumably through the inclusion of Mn in the IMC. The cyclic bending results showed SAC305 being the best, and all lead-free alloys are equal or superior to SnPb.
Considering the strong performance of SACM, it may be a good choice for medical electronics, as these critical electronic devices conform to RoHS in the near future.
As the medical electronics assembly industry moves to support a growing variety of hand-held, in-home devices, the strong performance of SACM can provide conformance to RoHS standards and reliability in these critical electronic devices.
This article was written by Professor Ronald C. Lasky, PhD, PE, Director, Cook Engineering Design Center, Thayer School of Engineering at Dartmouth College, Hanover, NH; and Carol Gowans, Incubator Market Manager, Indium Corporation, Clinton, NY. For more information, visit http://info.hotims.com/45607-168.