With supply chains originating anywhere in the world today, tracking down the root causes of problems can be tricky.
Achieving profitability in electronics manufacturing rests heavily on avoiding defects in parts and materials or, when a defect does occur, finding the cause and taking corrective action as quickly and inexpensively as possible. But a manufacturer really depends on his suppliers to keep defects at a minimum. In today�s globally sourced manufacturing environment, eradicating a defect may not be as simple as it once was.
In this case history, a board assembler had been using the same plastic-encapsulated microcircuit (PEM) on one of its boards for some years without incident. The PEM came from a supplier of long standing, and there had been no significant problems involving the device.
This uneventful pattern continued until the day when functional test suddenly revealed an unacceptable number of board failures. Internal failure analysis quickly discovered that the failure was occurring in the previously trouble-free PEM.
How widespread was the problem? Quality-control personnel established that the failures were isolated to a small number of production runs, all of which used a recently received lot of the PEMs in question. Boards using the PEMs from previous lots were unaffected. But in the meantime, production of the board was suspended.
The PEM had long been classified as a moisture sensitivity level 1 (MSL-1) part, meaning that the time during which the PEM could be exposed to ambient humidity on the assembly floor was unlimited (Table 1). That is, the part was very unlikely to uptake enough moisture to cause damage during the reflow process. A review of the handling procedures during production and the reflow profiles showed no anomalies. The smoking gun had not yet been found.
| Moisture Sensitivity Level |
Floor Life |
At These |
| 1 | Unlimited | <30�C @ 85% RH |
| 2 | 1 Year | <30�C @ 60% RH |
| 2a | 4 Weeks | <30�C @ 60% RH |
| 3 | 168 Hours | <30�C @ 60% RH |
| 4 | 72 Hours | <30�C @ 60% RH |
| 5 | 48 Hours | <30�C @ 60% RH |
| 5a | 24 Hours | <30�C @ 60% RH |
| 6 | Time on Label | <30�C @ 60% RH |
Table 1. MSL Class
Because this internal review found no cause for the failures, the assembler contacted the supplier of the PEM. The supplier reported back that the parts had passed the supplier�s internal quality-control evaluations.
The supplier also asked for some of the PEMs to be returned for their own analysis. The assembler sent some of the parts to the supplier and, at the same time, some samples of the failed boards along with a small number of unmounted PEMs to Sonoscan�s laboratory for evaluation by acoustic micro-imaging (AMI).
The first break came from the nondestructive AMI, which showed internal delaminations at the die face and the lead frame�but only in the mounted PEMs that had gone through the reflow process (Figure 1). The few unmounted PEMs had no delaminations, but the number of unmounted PEMs imaged was fairly small.
The red areas are the very high acoustic echo signals sent back by delaminations.
Tentative conclusion: the internal delaminations were related to the electrical failures and probably caused by heightened sensitivity to moisture/reflow conditions, something that should not happen if the PEMs were living up to the MSL-1 classification.
At this point, it appeared that moisture had collected in the part and had been flashing into steam during the heat of reflow. The volumetric expansion of the moisture as it transforms into steam can cause a delamination.
The ways in which internal delaminations can interact with a PEM to cause electrical failure are complex, but there are at least two well-known mechanisms:
- A delamination can break the wire bond when the delamination first occurs during reflow or later when the delamination expands as a result of thermal cycling during normal use. In this case, the electrical failures probably happened as soon as the delaminations occurred during reflow.
- A delamination is a natural collecting point for moisture and contaminants during the service life of the PEM. A layer of moisture as thin as 10 molecules can form an electrolytic cell that initiates the internal corrosion that can break a wire.
AMI is good at detecting and imaging internal delaminations because the method is very sensitive to internal gaps in material. At the same time, AMI is relatively insensitive to the thickness of a gap or delamination. Recent research by Sonoscan showed that delaminations revealed by AMI had thicknesses of only 100 to 1,000 �.
Since the data suggested that the PEMs might not be living up to the MSL-1 classification, a group of the unmounted PEMs was tested for moisture/reflow sensitivity per J-STD-020, which is based on the thickness and volume of the PEM and the anticipated factory environment. In this case, the test parameters for MSL-1 called for a 168-h soak time at 85�C/85% relative humidity. The unmounted PEMs failed this test (Figure 2).
The red areas are delaminations.
The assembler reported the results of AMI and the MSL testing to the supplier. The supplier agreed that internal delaminations were present but expressed disbelief that the delaminations could be causing the electrical failures.
This point of view is not without some basis in fact. An internal delamination that does not initially break bond wires may never cause a later electrical failure through corrosion or expansion. In this case, however, there were too many electrical failures, and they appeared to be tied directly to the occurrence of the delaminations.
At this point, the conversation turned to the production processes involved in manufacturing the PEMs. Had the supplier made any recent changes in the design of the PEM or in the materials used? After a bit of discussion, the supplier revealed that it no longer manufactured this PEM. Instead, manufacture had been subcontracted to an outside source.
The failed lot of PEMs was the first received by the assembler from the outside source. It seemed clear that there had been some modification in the PEM that had degraded its MSL.
What could the assembler do at this point to salvage something from this train wreck? Fortunately, the same AMI that revealed the delaminations in the first place could be used to evaluate the large number of remaining unmounted PEMs. They were imaged to determine the presence, size, and location of internal delaminations and, based on the results of acoustic imaging, sorted into three categories: accept, reject, and marginal. Some of the PEMs were usable with modified handling procedures, permitting the assembler to resume production.
Conclusion
The investigation carried out in this case did not reveal the precise cause of the delaminations. It seems probable that the new maker of the PEM did something different, such as used a different molding compound or modified the manufacturing process. Whatever the change, it was substantial and should have been accompanied by retesting to determine whether the MSL had changed. In this case, MSL retesting would have avoided long downtime for the assembler.
But the changes carried out by the subcontractor may, in fact, have been minimal. For example, minor surface contamination of the lead frame could cause delaminations. Since harmful internal defects such as delaminations sometimes are caused by subtle process changes, it makes sense to perform routine acoustic screening of incoming parts. The savings in dollars and time by using screening to prevent a production shutdown can be substantial.
About the Author
Tom Adams is a freelance writer and photographer who has written more than 500 articles for semiconductor and microelectronics trade magazines.
Sonoscan, 2149 E. Pratt Blvd., Elk Grove Village, IL 60007, 847-437-6400
FOR MORE INFORMATION
on AMI
www.rsleads.com/404ee-176
Return to EE Home Page
Published by EE-Evaluation Engineering
All contents � 2004 Nelson Publishing Inc.
No reprint, distribution, or reuse in any medium is permitted
without the express written consent of the publisher.
April 2004