Although innovative design practices in the past decade have brought higher levels of reliability in electronic equipment, failures still are a challenge for the system designer. Many techniques for improving the reliability of electronic equipment in harsh environments have evolved over the past two or three decades, based largely on experience gained during the development of military and space systems. Some of this knowledge has helped improve the reliability of industrial electronic devices. This is a reexamination of one of these techniques—environmental stress screening (ESS).
We can consider ESS for a system in two parts. First, component reliability should be evaluated by accelerated life testing to expose latent defects. Second, the entire system must be exposed to and successfully withstand stresses similar to those expected in the field.
Accelerated Life Testing
At the component level, accelerated life testing obtains reliability information in a compressed time frame to help predict failure modes. The life of the part is accelerated by applying high levels of stimuli and by simulating a normal lifetime in just a few days.
High levels of stimuli include high-temperature burn-in, temperature cycling, thermal shock, and vibration. Accelerated life testing cycles the component thousands of times in a short period of time. Such testing can expose latent defects and help predict the reliability of components in normal operation.
Accelerated tests use temperatures in the range of 75°C to 225°C and relative humidity (RH) in the range of 50% to 90%. The standard combination is 85°C and 85% RH. Failure mechanisms such as corrosion and metallic growth caused by migration of ions can be stimulated by these tests.
Where environmental stresses are applied in conjunction with high temperature, a number of procedures are used to model the results. They include the Arrhenius, Eyring, Reich-Hakim, Peck, and Lawson models.
ESS
In ESS, environmental stresses within the design capability of the system are accelerated to bring about failures that otherwise would likely occur in actual service. Using physics-of-failure-analysis methods, we can explain electronic component failures in terms of certain physical and chemical processes. Some failures are temperature-dependent; others are accelerated by electrical stresses such as the application of high voltage or high current.
By applying the appropriate excitations for a specific number of times and dwelling on certain points, we can generate a good generic ESS sequence. If the expected operating environment is atypical, a custom screen can be developed instead.
There are several benefits of an ESS program. These include the reduction in the number of equipment failures after products are delivered, a resulting decrease in warranty/repair costs, an identification of latent design deficiencies, and a better relationship between manufacturers and users.
Among the disadvantages of ESS is the cost of environmental chambers, and the labor to perform the tests. There also is the continuing expense of cables that wear out and fixtures that are unique for each item. Also, certain self-healing faults may not show up while performing these procedures.
Screening Methods
Several methods for electronic screening are summarized in Table 1. Procedures are based on MIL-STD-202F, Test Methods for Electronic and Electrical Parts. Burn-in standards are based largely on MIL-STD-883, Test Method Standard for Electronics.
High-Temperature Burn-In
In high-temperature burn-in, the UUT is subjected to a temperature stress that accelerates the aging process: 70°C for a commercial device or 125°C for a military device, for 24 hours to 168 hours. During burn-in, the equipment is powered with electrical inputs to simulate normal operation. After burn-in, a complete functional test is run in an ambient environment.
This sequence will eliminate errors caused by manufacturing process faults or incorrect handling during assembly. These may include wire-bond defects, oxide-layer faults, or imperfect metallization.
Temperature Cycling
Alternating cycles of hot and cold temperature, typically -40°C to +125°C for industrial equipment or -65°C to +150°C for military equipment, will uncover temperature-dependent failure mechanisms. These may include wire-bond defects, die-substrate attachment problems, cracks in the die, a mismatch of thermal coefficients of expansion, imperfect seals, or packaging problems.
The sequence should be 10 to 20 cycles, dwelling at each extreme for 30 minutes. The rate of change should be 5°C to 10°C in the -10°C to 70°C range.
After all cycling, the equipment should be tested at ambient temperature to verify that it still meets specifications.
Storage at High Temperature
Power-off exposure to a higher temperature than the burn-in sequence will uncover problems such as moisture entrapment, imperfect metallization, silicon defects, oxidation, or contact imperfections. Plastic-encapsulated components should be heated to 150°C and hermetically sealed devices to 250°C, each for 24 hours.
Life
Life testing helps to evaluate mechanical and electrical properties of the operating UUT at an elevated temperature for a fixed length of time. Capacitance, dielectric strength, insulation resistance, and surge current may be affected by this process.
Thermal Shock
The resistance of a component to temperature extremes and cyclic exposure is determined in thermal shock. Cracks in a surface, delamination, ruptured seals, changed electrical characteristics, or leakage of an electrolyte or other filler material are detected.
Salt Spray
The UUT is submitted to a water mist of 5% salt solution to determine susceptibility of metallic parts and their coatings to corrosion.
Humidity
In a steady-state humidity test, the UUT is subjected to 90% to 95% RH at 40°C for 96 hours to evaluate the absorption of moisture by components.
Moisture Resistance
Moisture resistance is an expansion of the steady-state humidity test. In this test, the parts are subjected to alternate cycles of dampness and drying while the temperature is cycled. This accelerates corrosion in susceptible components such as wires and contacts.
Immersion
To judge the effectiveness of component seals when subjected to hot and cold liquids, the equipment is immersed in a pure water or a salt-water bath at several temperatures between 0°C and 35°C.
Vibration
The UUT is vibrated from 20 Hz to 2,000 Hz or at the frequencies of expected operation. This very important procedure may loosen parts or cause motion between parts or noise. It can expose mechanical fatigue or failure, loose solder joints, improper bonding, or chip-level mechanical flaws.
Random Drop
Normal shipping and handling impacts are simulated by drop testing. This detects weaknesses that might not show up in shock and vibration.
Solderability
This screen determines the solderability of a termination. The level of wetting of the termination and the formation of a solder fillet after application of solder are evaluated. Accelerated aging also is included in this procedure.
Resistance to Soldering Heat
In resistance to soldering heat screening, the through-hole-mounting component is immersed in a 260°C solder bath for 10 s to verify that it can withstand the heat of soldering and cleaning processes. Possible ill effects include a change in the electrical characteristics of the component, loosening of mechanical terminations, softening of insulation, open solder seals, or weakening of mechanical joints.
Electrical Overstress
Electrical overstress is the application of a voltage that is increased in steps from normal to 25% or 50% above normal while the temperature is held above ambient. The sequence also is known as step-stress testing.
Conclusion
Failures in electronic equipment can be attributed to infant mortality, electrical or thermal overstress, or fatigue. Infant mortality problems may be caused by manufacturing defects, faulty handling, defective component parts, or mistakes in design. Failures during normal operation can be caused by sudden application of high stress, aging of components, chemical contaminants, moisture, or electromigration of metal ions. Each of these types of failures should be detected by ESS.
Defects uncovered by screening are corrected by redesign, improvement of manufacturing processes, or corrective action by component vendors. With the continuing development of smaller, higher-performance products that may operate in higher stress environments, the search for higher reliability will be an ever-present challenge, and screening techniques must be applied to meet the goal of high operating reliability in the coming years.
References
1. Jensen, F., Electronic Component Reliability, John Wiley and Sons, 1995.
2. Amerasekera, E.A., et al, Failure Mechanisms in Semiconductor Devices, John Wiley and Sons, 1987.
3. Doyle, E.A., How Parts Fail, IEEE Spectrum, October 1981.
4. Kales, P., Reliability Technology for Engineering and Management, Prentice-Hall, 1998.
5. Ramakumar, R., Engineering Reliability Fundamentals and Applications, Prentice-Hall, 1993.
6. Pollino, E., Microelectronic Reliability, Volume II, Artech House, 1989.
7. MIL-HDBK-202F, Test Methods for Electronic and Electrical Component Parts.
8. MIL-HDBK-217F, Reliability Prediction for Electronic Equipment.
9. MIL-STD-883E, Test Method Standard for Microcircuits.
About the Author
V. Lakshminarayanan is failure analysis and reliability coordinator at the Centre for Development of Telematics. He holds an M.E. degree in electrical communications engineering from the Indian Institute of Science. Centre for the Development of Telematics, 71/1 Miller Rd., Bangalore 560 052, India, [email protected].
Copyright 1999 Nelson Publishing Inc.
October 1999