| Created: 8/1/95 | Updated: 11/13/96 |
So, how does one develop products in a more timely manner? The CALCE EPRC is conducting various research efforts to identify the bottlenecks and explore alternative methods and techniques to remove the time-delay barriers. As part of these efforts, the CALCE EPRC has been investigating product flows, with particular emphasis on the flow associated with product qualification and quality assurance. Our view is that most electronics companies do neither of these tasks particularly cost effectively or efficiently. For example, many companies still conduct Mil-Std-883 qualification tests and quality assurance screens which are impractical or need modifications for new technologies, and add little or no value for mature technologies.
The CALCE EPRC has determined various methods to reduce qualification and quality assurance time and ensure greater reliability. The foundation of these methods for microelectronics packages can be found in the book titled Quality Conformance and Qualification of Microelectronic Packages and Interconnections, John Wiley, 1994, edited by Michael Pecht and Abhijit Dasgupta from the CALCE EPRC and John and Jillian Evans from NASA, with contributions from some twenty industry, government and university experts.
Emphasis at the CALCE EPRC is now being placed on examining timely product assurance for specific technologies such as plastic encapsulated microcircuits, ball grid arrays (PBGA)s, multichip modules and printed wiring board assemblies.
Health management is an affordable, efficient, and reliable on-line, non-destructive technique that involves development of health monitors incorporating sensor technologies, heuristic, mechanistic, and stochastic on-line diagnosis algorithms, and physics-of-failure reliability assessment for in-situ defect detection and reliability assessment. The technique provides information about the actual stresses responsible for failure and ensures compatibility of the test conditions with application environment. Health monitoring also enables early detection and warning of an imminent potential failure, critical in an electrical circuit, as it is the key to the safety of personnel and equipment. Finally, health monitoring enables higher operating efficiency to be achieved, because it provides information regarding the need for maintenance (repair or replacement).
Successful health management requires multiple integrated sensors for monitoring the physical condition of the electronic system. Accelerometers, strain sensors, stress sensors, resistive sensors, capacitive sensors, thermistors, all can provide needed sensory input. The data then must be analyzed, integrated and classified into a number of inference classes, and then coupled with computational algorithms capable of locating failure sites and predicting the extent of degradation.
Physics-of-failure approach is integrated with health management to estimate the remaining life of the system. The approach uses mathematical models based on the fundamental physical and chemical processes by which failure occurs. With these models, it is possible to estimate the time-remaining-to-failure using the information about stresses acting on the system, sensory inputs, relevant material properties, and original design. This methodology is enhanced by the use of health monitoring fuses. This involves incorporating on-line health monitors having design geometries that accelerate failure by a particular mechanism. The acceleration factor for a test structure is determined by the physics-of-failure models. This acceleration factor, together with the monitored time-to-failure for the test structure is used to estimate the time-remaining-to-failure of the actual part. For further information, please contact Dr. Michael Pecht at (301) 405-5323.
Through a cooperative research and development agreement (CRADA) with the Army Research Laboratory, CALCE EPRC is pursuing efforts to assess the reliability of PEMs on military Sonobuoys and unassembled PEMs stored for 8-12 years. A three phase approach is being used for this assessment. In the first phase, CADMP-II simulations are being conducted on the parts to determine the dominant failure mechanisms, modes, and sites, to guide the degradation analysis. This phase also includes the calculation of physics-of-failure based acceleration transforms for these dominant failure mechanisms. In the second phase, materials analysis techniques are being used to identify and characterize the degradation. This includes (1) using visual and optical microscopic examination to observe external cracks, corroded leads, or other signs of damage; (2) electrical testing to identify parametric shifts; (3) scanning acoustic microscopy to observe package delamination and cracking, and die attach voiding; (4) decapsulation and E-SEM/EDS to see evidence of corrosion, passivation cracking, electromigration, stress driven diffusive voiding, wire fatigue and fracture, and die cracking; and (5) cross-sectioning and E-SEM/EDS to characterize intermetallic growth. The degradation in stored PEMs is being compared to the degradation in ceramic parts stored for an equivalent period of time. These analyses are providing the first conclusive information on the location, mechanisms, and extent of degradation in PEMs caused by long term storage.
In the third phase, CALCE EPRC is addressing the critically important issue of estimating the "remaining life" of PEMs. This is being accomplished using a combination of the physics-of-failure models for the dominant failure mechanisms and accelerated testing on both the old and new components. The testing includes temperature-humidity stress tests, highly accelerated stress tests (HAST), and temperature cycling tests. The time-to-failure determined by these tests will be used together with physics-of-failure acceleration transforms calculated in phase I to assess "remaining life" for PEMs in various storage environments.
If you are interested in contributing to this effort and in obtaining the results of these investigations, please contact Dr. Patrick McCluskey at (301) 405-0047.
Background
Epoxy molding compounds are hygroscopic; the ingressed moisture accumulates at the interfaces inside the PEM. During reflow soldering, the entire PEM is heated to temperatures as high as 230øC, well above the glass transition temperature of most molding compounds. Stresses are developed in the PEM as it is heated to reflow temperatures. The factors contributing to the stresses include the rapid vaporization of absorbed moisture resulting in the build-up of vapor pressure, thermal expansion mismatches at the interfaces of the PEM, and a decrease in the adhesive strength and fracture toughness of the molding compound. When the hygrothermal stresses and vapor pressure exceed the adhesion strength and fracture toughness of the molding compound, they become the driving forces behind the growth of the delamination and the formation and propagation of cracks.
Popcorning may not result in the immediate failure of the package. However, the long-term reliability of the package could be jeopardized if surface-breaking delaminations and cracks serve as a path for the entry of ionic contaminants into the package. Delaminations and cracks that do not appear externally can also lead to reliability problems. During temperature excursions, the delaminated plastic may be free to move relative to the die surface, possibly stressing and damaging passivation and metallization. Cracking and delamination between the molding compound and the die surface may also lead to sheared or cratered wires and ball bonds, causing immediate failure or intermittent electrical failures during temperature excursions. Information on PEM composition, reliability and qualification can be found in the book titled Plastic Encapsulated Microelectronics by Pecht, Nguyen and Hakim, John Wiley, 1995.
A test methodology to effectively assess the susceptibility of PEMs to popcorning
The test system designed, built, and in-production use at the CALCE EPRC, consists of an X-ray radiography unit and probe tester embedded in an infrared heating system. The PEM can be precisely heated through any desired infrared reflow temperature profile, while simultaneously monitoring its X-ray image and out-of-plane deformation, in real time, for popcorning. The video image of the X-ray is processed and displayed on a monitor and is recorded on video cassette or photographic film. The probe is set up against the face of the PEM, and is used to quantify the doming of the encapsulant prior to package cracking.
The equipment can be used to identify the specific stage of the reflow profile at which doming of the encapsulant and package cracking occurs. It is an effective tool to study the effect of variations in parameters such as reflow temperature ramp rates and moisture content of the device on popcorning in PEMs. The build-up of vapor pressure is responsible for the formation of the characteristic dome-shaped bulge in the encapsulant. Popcorning is characterized by the growth of this bulge and eventual cracking of the device. The probe is used to physically measure the extent of doming that occurs in PEMs, as they are heated through the reflow cycle. The device thickness is continuously monitored, and the total deformation of the device just prior to cracking is detected with an estimate of the critical level of device doming, beyond which cracking occurs.
An acoustic microscope is then used to quantify delamination and popcorn cracks in PEMs. At the CALCE EPRC, we use the Sonix reflection-type microscope that generates images by mechanically scanning a transducer in a raster pattern over the sample. The transducer alternatively acts as sender and receiver. A very short acoustic pulse enters the sample, and return echoes are produced at the sample surface and at specific interfaces within the part. An oscilloscope display of the echo pattern clearly shows these levels and their time-distance relationships from the sample surface. Software converts this information into a pseudo color map of the sample interface. The contrast of the image is due to the acoustic impedance mismatch of the different materials at the interface. The image is used to quantify the extent of delamination, in terms of percentage area.
In some cases, such as with PBGAs, selected devices are cross-sectioned and examined using an environmental scanning electron microscope (E-SEM), in order to confirm the results obtained. The E-SEM can be used to study materials under controlled environmental conditions. The advantage of using an E-SEM is that uncoated specimens, free of surface charge and high vacuum damage can be examined. For information on service costs, please contact the CALCE Center at (301) 405-5323.
With support from the Deputy Under Secretary of the Army (Operations Research) and the Office of the Assistant Secretary of the Army (Research, Development and Acquisition), the Headquarters of the Army Material Command authorized the Electronic Equipment Physics-of-Failure (EEPOF) project. Involvement is in the defense conversion and dual-use arenas, where physics-of-failure concepts are being used to revamp the reliability technologies associated with design, engineering, development, production, and life-cycle support.
Some key completed efforts include:
Certain key parameters have been identified which influence overall heat transfer most significantly for such problems. Numerical calculations and experiments to study the effects of these parameters and their interactions are in progress. General correlations for the maximum temperature within the components, as a function of the system physical parameters are also being explored. For further information, please contact the CALCE Center at (301) 405-5323.