Enhancing Military Systems with Plastic Encapsulated Microcircuits

1. DMSMSKissimmee, Florida December 2nd – 5th, 2013 Joseph L. HoltVice President Business Development Integra Technologies LLC 3450 N Rock Road Building 100Wichita, Kansas 67226Ph 316-630-6808Email: Joe.Holt@Integra-tech.comWeb: www.integra-tech.com Use of Plastic Encapsulated Microcircuits (PEMs) in Military and Space Systems: A review of 15 years of actual PEM reliability data.

2. How Can PEMs Increase Material Availability? • How can use of Plastic Encapsulated Microcircuits (PEMs) reduce diminishing material availability issues and add flexibility to parts management challenges? • The short answer is that because inexpensive consumer electronics drive todays semiconductor manufacturers, inexpensive plastic parts are by far the dominant package type – which means suppliers are more likely to support these technologies for a longer period of time. • It does not mean they will support plastic devices as long as many Aerospace and Defense contractors need, but all other things being equal, plastic parts will likely be available longer than hermetic.

3. Other (Better) Reasons to use PEMs • Other reasons to use PEMs in addition to greater availability: • Generally lower cost materials • Greater investment in supplier product improvement programs • Greater product variety • Mechanically more rugged • Lighter weight • Available in smaller/thinner packages • More automated assembly methods • Higher volume/more cost effective production • Increasing reliability due to supplier competition • All of the above items will help reduce end system costs. And with the relentless pressure to reduce costs (sequestration, budget deficits, congressional gridlock, recessions in US and Europe, etc.) the ability to reduce costs are at least on par with availability as a reason to use PEMs and arguably even more important.

4. So why bother with a PEM Qual • If PEMs are so good, why bother qualifying them at all? • PEMs have predominately been designed for commercial and industrial applications – not Aerospace and Defense (A&D). • Higher military temperature ranges are not generally available. • Screening and qualification testing by the semiconductor suppliers are not always as stringent as A&D end applications require. • Reliability of plastic packages varies dramatically from supplier to supplier • Any of these reasons should be justification by themselves to perform PEMQuals, but the last reason makes it essential to intelligently evaluate the use of plastic parts. There is a substantial variation from supplier to supplier in PEMs reliability, as our data will show.

5. What is a PEM Qual • Example PEMQual flow (per NASA PEM-INST-001): • External Visual and Serialization: Qty=32 • Baseline CSAM: Qty: 32 • Pre-Conditioning: Qty=32 • Moisture Soak per appropriate MSL Level • Reflow at 235C • Electrical Testing at 25C, min and max operating temp; Qty=32 • Sub-Group 1 Qual: Qty=22 • 125C Life Test for 1000 hours • Electrical Testing at 25C, min and max operating temp • Temp Cycle; 200 cycles • Electrical Testing at 25C, min and max operating temp • CSAM; 22 units • DPA : 5 units (per sec.5.3/ Fig 4 of PEM-INST-001 spec) • Subgroup 2 Qual: Qty=10 • Unbiased HAST; 96 hours at 130C/85%RH • Electrical Testing at 25C, min and max operating temp

6. What is a PEM Qual • PEMQual flow notes: • Most customers customize PEMQual flows to be more stringent or less stringent depending upon their application needs. • This is important because over qualifying adds evaluation cost while eliminating devices that might otherwise be reliable enough and lower priced. • Even worse, under qualifying may lead to inadequate device reliability and system field failures. • Clearly understand how comprehensive the electrical test coverage is that your test supplier is providing. Insufficient electrical test coverage will allow failing devices to be counted as passing and could lead to poor system and field reliability. • Thorough and intelligent assessment of the intended application is essential for the successful use of PEM devices.

7. What are the Typical PEM Qual Failure Types • Typical PEMQual Failure Types: • Cracked Die • Wire breaks • Wire Lift • Delamination • Corrosion • Contamination • Moisture Ingress • Cracked Package • Cracked Passivation • Die Attach Voids • Most PEMQuals start by assessing the die level qualification data already available from the manufacturer. Once the die level reliability data is deemed acceptable, the PEMqual can evaluate the package versus the application needs.

8. Integra PEM Qual Data Overview • Notes: • No qualifications conducted by semiconductor manufacturers are included. • Plastic packaged semiconductor devices only - no passives. • Predominant test temperatures are -40, 25 85 and -55, 25 125. • Testing temperature order is usually room, cold, hot. • Once a qual fails it is usually stopped. • Failures are for electric test only (no mechanical failures). • Vast majority of testing performed to manufacturers datasheet limits. • Virtually all electrical test programs written by Integra Technologies.

9. Overall PEM Pass/Fail Rate This data suggests that PEMs are getting more reliable over time or that the customers are getting smarter about which suppliers to select for PEM quals. We actually believe it to be both.

10. Overall Customer PEM Qual Success Rate Customers who are more experienced with PEM Quals tend to have better success. These 4 customers averaged over 100 PEM Qual lots each over the 15 year analysis period. Average Pass Rate = 80% The average pass rate over the entire population is 63%. Customers who are less experienced with PEM Quals tend to have less success. These 4 customers averaged ~4 PEM Qual lots each over the 15 year analysis period. Average Pass Rate = 42%

11. Overall PEM Qual Success Rate by Package There are meaningful differences between package types in their ability to pass a PEM Qual.

12. Overall PEM Qual Success Rate by Package Pin Count Despite the sensitivity to package type shown on the previous page, there does not appear to be a meaningful sensitivity to package pin count.

13. Overall PEM Qual Success Rate by Device Technology There is not a great deal of sensitivity to technology, with the exception of memory. It should be noted that the memory devices we evaluated tended to more often come in packages that were previously shown to be less reliable.

14. Overall PEM Qual Success Rate by Device Technology by Supplier

15. Overall PEM Qual Success Rate by Device Technology by Supplier Many suppliers have good PEM Qual results….

16. Overall PEM Qual Success Rate by Device Technology by Supplier Many suppliers have good PEM Qual results…. …But many suppliers do not.

17. Overall PEM Qual Success Rate by Device Technology by Supplier Many suppliers have good PEM Qual results…. Acceptable suppliers can be found for most technologies. …But many suppliers do not.

18. Overall PEM Qual Success Rate by Supplier This is arguably the most important point of this data analysis – that PEM Qual success rates vary dramatically from supplier to supplier.

19. Overall PEM Qual Conclusions • There is a great variation in the success rate depending upon the supplier that is being evaluated, so it is prudent to evaluate multiple suppliers. • Success rate is also influenced by the package being evaluated, although not to as great an extent as by the supplier. • Acceptable PEMQual success rates can be obtained with most technologies and pin counts. • Understanding your application needs is essential for selecting the best PEMQual flow. By matching the PEMQual flow to the application needs, an accurate assessment can be made at the lowest evaluation cost.

20. Overall PEM Qual Conclusions (continued) • Clearly understand how comprehensive the electrical test coverage is that your test supplier is providing. Insufficient electrical test coverage will allow failing devices to be counted as passing and could lead to poor system and field reliability. • PEM Quals are very complex flows with hundreds of processing steps and thousands of data points. Assure that trained project management staff is present at your test supplier to manage the flow execution and assure data integrity. • Overall, PEM Quals can be used to effectively select devices for use in non-commercial environments. In this study, approximately two-thirds of industry devices will pass PEM Qualification. • These conclusions are consistent with the conclusions reached in 1998 when we last analyzed our database, although our data indicates that PEM Qual success rates are improving.

21. Upscreening – The Simplest Form of PEM Qual • Upscreening is where a device is not subjected to any environmental testing (Temp Cycle, HAST, etc.), only to electrical testing at temperatures beyond where the original manufacturer tested. • Upscreening is the simplest form of a PEM Qual even though it is not technically a “qualification.” • Many customers combine upscreening and PEM qualification when assessing the suitability of a device for given application. • Integra also has considerable upscreening data.

22. Integra Technologies Historical Upscreening Data Overall Summary of Upscreening Data Time period data taken 13 years (1999-2012) # of electrical test steps* 19,489 # of devices tested 11,678,154 # manufacturer part numbers 1,403 # manufacturers 14 # of customers 176 Overall failure rate 0.93% Minimum failure rate 0.00% Maximum failure rate 100% * A test step is group of customer devices tested at a given temperature. The same devices are typically tested again at a different temperature. These are separate “test steps”.

23. Upscreening Data by Supplier Supplier failure rates vary significantly from supplier to supplier.

24. Upscreening Data by Test Temp by Technology Some technologies show better performance than others at extended temperature ranges.

25. Upscreening Data by Pin Count Device failure rates tend to be higher for higher pin count devices.

26. Upscreening Data by Package Device failure rates vary meaningfully by package type.

27. Upscreening Data by Year An initial look at the data would seem to indicate that device failure rates are going up……

28. Upscreening Data – Lots Tested & # Zero Fail Lots The number of lots tested is going down…. As customers eliminate devices that do not fail.

29. Upscreening Data by # Failing Devices But the actual number of gross failing devices is actually increasing slightly. Conclusion – Customers are becoming more adept at upscreening only those devices that really need it, but many devices continue to need to be screened – the trick is to know which ones.

30. Upscreening Data Conclusions Upscreening Conclusions • Customers are reducing testing when data shows that no failures are being found. • While the number of lots being tested is going down, the number of gross failing devices is staying constant and even increasing. This suggests that certain devices definitely still need to be screened, but others may not require screening or can go to sample testing. • Failure rates still vary significantly by supplier and by technology. • Clearly understand how comprehensive the electrical test coverage is that your test supplier is providing. Insufficient electrical test coverage will allow failing devices to be counted as passing and could lead to poor system and field quality/reliability. • Upscreening can be successfully used to yield devices that can perform at extended temperature ranges.

31. Upscreening Data Notes Important Notes for Upscreening Data The data only contains information for devices tested for users of integrated circuits (IC’s). No test results are included for devices tested for an original semiconductor manufacturer. No test results are included for custom fabricated devices manufactured exclusively for a specific end user. The failure rates in the data tables should be considered minimum failure rates because if testing of a device reveals high fallout, it is usually designed out (or an alternate part is selected by the customer) therefore eliminating the highest failing devices from the tested population. In addition, many customers characterize smaller samples at the system level prior to upscreening so they can determine if the yields will be acceptable. The vast majority of the lots were tested at one of the following temperatures: +125C, +85oC, 25C, -40oC & -55oC. The vast majority of the tests performed were to the original manufacturers data sheet. The vast majority of the test software used was developed independently by Integra, without help from the original manufacturer.

32. Upscreening Data Notes Important Notes for Upscreening Data ~97% of the lots tested were for Military and Avionics customers. No lots designated as Counterfeit were included, as electrical testing is often very abbreviated. The failures noted are for electrical failures only. No mechanical failure information is included.

33. Integra Technologies Thanks You Thank you from the Employee Owners of Integra Technologies!! Integra Technologies LLC, along with the recently acquired Analytical Solutions, has been providing one of the broadest ranges of test and evaluation services in our industry for over 30 years. Our services include: Test Development PEM Qualifications Final Test Qualification Services (HTOL, HAST, Temp Cycle, etc.) Characterization Assembly/Repackaging (outsourced to qualified partners) Wafer Probe MIL-STD 883 and 750 Testing Upscreening Volume Production Test Failure Analysis Destructive Physical Analysis Counterfeit Detection Obsolescence Management We are approved by DLA for MIL-STD-883 & 750 processing, ITAR, ISO 9001, AS9100 and DMEA Category 1 “Trusted” www.Integra-Tech.com