Failure analysis and prevention are important functions to all of the engineering disciplines. The materials engineer often plays a lead role in the analysis of failures, whether a component or product fails in service or if failure occurs in manufacturing or during production processing. In any case, one must determine the cause of failure to prevent future occurrence, and/or to improve the performance of the device, component or structure.
One of the most visible examples of the application of failure analysis involves the aerospace industry. Aircraft accidents are remembered by the public because of the unusually high loss of life and broad extent of damage that is typical of this type of accident. On December 19, 2005, just off Miami Beach, Florida, a Grumman G73T Turbo Mallard seaplane crashed as it was taking off. The aircraft was carrying 15 passengers, 3 infants and a crew of two. There was an explosion and fire and the right wing separated prior to the plane impacting the water. Examination of the wreckage revealed fatigue cracks in the right wing. The cause of the accident is still under investigation. However, structural failure, initiated by fatigue is suspect. To see how this accident investigation unfolds, go to this Aircraft Maintenance Technology (AMT) website.
Another aircraft accident recently in the news was the TWA Flight 800 disaster that occurred when the Boeing 747 exploded off Long Island on July 17, 1996. There has been an immense effort to find the cause of this accident. To this date, the specific cause of the disaster has not been determined, but the most probable cause involved frayed sensor wires in the fuel tank of this airliner.
Most people have seen photos of the inside of the aircraft hanger where the Boeing 747 was re-constructed. This re-construction is a major step in aircraft accident investigation. A classic aircraft accident investigation involves the De Havilland Comet accidents of the early 1950's. The Comet was the world's first commercial jet airliner. A design defect resulted in the structural failure of many Comets. The loss of Comet I G-ALYP lead to the finding of 'design defect' as the cause. Metallurgical failure analysis plays a critical role in determination of sequence of failure, and ultimately, in identification of the cause of aircraft accidents. The interested reader can review a chronology of the Comet Mystery at the following web page . A simulation of one of the Comet aircraft crashes, off the coast of Italy, is available at this Youtube address. The full story of the de Havilland Comet is presented on the following Wikipedia web pages.
More recent aircraft accidents, which have (or appear to have) involved structural issues, include the tragedies of Alaska Airlines Flight 261 , a Boeing/McDonnell Douglas MD-83 which crashed into the Pacific Ocean off California on January 31 2000, and American Airlines Flight 587, an Airbus A-300, which crashed nose-first into a southern Queens neighborhood of New York city Monday, November 12, 2001. At the time of the AA Flight 261 accident, I wrote a brief article (unpublished) about aircraft accident investigation. If you have further interest in this subject, please link here.
Another recent example of an accident investigation in the transportation sector is the delrailment of the German high-speed passenger train that occurred on June 3, 1998. The train was travelling from Munich to Hamburg when it derailed just before 11 a.m. in Eschede, 35 miles north of Hanover. Post-crash reports focused on rattling that alarmed some passengers in the minutes before the crash of InterCity Express 884. The reason for the rattling now seems clear: investigators say that a broken wheel was the most likely cause of the disaster. Metal fatigue is suspect.
Failure analysis is not limited to train and aircraft accident investigations. On these pages I hope to provide one a sense of what failure analysis (sometimes called failure prevention and referred to in the electronics industry as reliability physics) is all about. Since materials engineers play a major role in the failure analysis process, one may consider Materials Science and Engineering education as a door to the world of the Failure Analyst
Venezuelan Natural Gas Pipeline Rupture
A natural gas pipeline in Venezuela ruptured next to a major highway in September, 1993. The subsequent gas jet ignition resulted in an inferno that killed at least 50 people. Within hours of the initial contact, Failure Analysis Association (a commercial firm) engineers with expertise in materials, combustion, and pipeline failure mechanisms arrived in Venezuela to start investigating. Such rapid response is essential for examining conditions as close as possible to the time of the incident. The reference source URL is a commercial failure analysis firm, Failure Analysis Associates, now known as Exponent
NEW (2/27/11): A SPECIAL OVERVIEW OF MECHANICAL CONCEPTS PERTINENT TO THE SAN BRUNO NATURAL GAS TRANSMISSION PIPELINE FAILURE OF SEPTEMBER 9, 2010, by p. pizzo. Link here .
Shown below is the Heverill Fire Department aerial ladder failure. Structural failure of a ladder is not at all an uncommon event. Failure can result, for example, from poor design, use of inferior material or fabrication methods, or from a phenomenon called fatigue. Fatigue is a failure mode which occurs in structural materials and is driven by repeat loading, which when you think about it is a necessary requirement of a ladder, a drive shaft on an automobile, and the wing on an airplane. Further into this web page, you will see how one can sort a fatigue failure from various other modes of failure. The reference source URL is the System Engineering Group at the NASA-Ames Research Center in Mountain View, CA.
In the analysis of structural failure, mechanical testing is often a requirement of the process. For example, let us say that one suspects fatigue to be involved in the failure of a garage door spring. To predict life of such a spring, one would need to know what load would be applied to the spring in its service to a garage door; and how many times this load might be cycled in a typical year. There are analytical methods and computer modeling methods to address these concerns and predict the time and/or cycles to failure. These models are based on, and confirmed by, empirical data from fatigue tests conducted both on uniaxial tensile specimens and on actual springs. A typical system for testing structural materials is shown below. A portable Instron Model 8511 which has been dedicated for fatigue and designed for both low force, cyclic fatigue applications and for tensile/compression testing is shown. The system was reportedly used by the Orthopedic Bio-Mechanics Laboratory at the Harvard Medical School for testing things like artificial hips, and stainless steel rods used to correct scoliosis. This site is no longer available; but a similar Bio-Mechanics Laboratory exists at the University of California at Davis, and you are encouraged to visit this site.
The possibility of failure of a structural ladder by fatigue was mentioned above. How can one tell if failure is by the fatigue mode rather then, for example, by some other metallic-embrittlement mode? The answer, in part, can be found by performing fractography. Fractography is simply microscopy of the fracture surface. But until the advent of the scanning (SEM) and transmission (TEM) electron microscopes, fractography was rather difficult to perform. Shown below-left is a detailed inspection [at approximately 5000X] of a fracture surface using SEM. The presence of a series of marks approximately parallel to the crack front are revealed. The marks are called fatigue striations and are characteristic of the growth of a fatigue crack in a ductile material. This confirms crack growth by the fatigue process. Visit the web pages of the The Materials Society (TMS), the award-winning paper An Introduction to Failure Analysis for Metallurgical Engineers by Thomas Davidson, to get a better idea of how features on the surface of a crack can reveal the underlying failure mode .
Now let's step back from the striations on a typical fatigue fracture surface to have a look-see at the "macro" features of the fracture plane of a typical engineering component. The subject is the fatigue failure of a splinded solid-shaft under torsional loading. Link here and you will see the shaft, with the fracture plane in profile. Note the interesting multi-step characteristic of the fracture profile. This characteristic is due to the presence of many crack origins along the splines. Multi-crack origins are a common feature of a fatigue failure. The refernece source URL for this failure is the Engineering Materials Group at Southampton University
Now recall the crash of the Grumman G73T Turbo Mallard mentioned above? This is what investigators found on a rear spar of the wing structure; the wing that separated in flight. Notice the tale-tale fatigue characteristics? Follow-on SEM fractography is needed to confirm the fatigue crack-growth mode. Go to this National Aerospace Laboratory (Amsterdam) report to see how fractography is integrated into aircraft accident investigation.
A good starting point to learn about failure analysis from a metallurgical viewpoint is the ASM-Handbook on Failure Analysis and Prevention which is available from ASM-International. If you are interested in failure modes which may occur in a variety of engineering structures, please visit these Southwest Research Institute (SwRI) pages. Also, try the Component Failure Museum web pages of the UK's Open University.
Inspection of Non-Metallic Components
Let us now consider a non-metallic component that the materials engineer must be prepared to analyze to optimize performance. By far the most common type of dynamic seals in use today is the oil seal (or rotary shaft seal). While its initial cost is minimal, its impact on maintenance time and labor can be significant. An early seal failure will throw even the best program off schedule. On the web pages (or screens) of Chicago-Rawhide were examples of the most common seal failures found when investigating field problems. However, Chicago-Rawhide was recently bought-out by SKF of Stockholm, Sweden, and the excellent failure analysis pages have been dropped. However, another failure guide for polymer materials can be found at MERL web pages (Great Britain).
Here are two images from the former Chicago-Rawhide web pages.
The first image conveys an important message of any failure analysis. Examine carefully, by eye and with low power lenses (5X to 10X), any failure or fracture to begin the failure analysis procedure.
The second image illustrates one particular failure mechanism. When operating speeds increase, seal lip temperatures may soar. One indication of high heat is a dry, brittle lip. Flexing the lip may reveal fine axial cracks around the entire circumference. Another indicator is a thin band of carbonized oil along the seal lip that results when heat causes the lubricant to break-down. These are clues to look for in examining failed seals. Remember too to look at other seals of similar life in similar situations to gain more knowledge about a particular failure mode; and to learn something about the extent of the problem!
Failure Analysis of Devices [also termed 'Reliability Physics']
Now, let us switch gears to solid state device failure analysis methods. Electronic, magnetic and optic devices too, can fail. The cause of failure must be ascertained to improve reliability and to correct errant process steps. Here are examples which may give you the idea of reliability physics or device failure analysis.
The application of Scanning-Probe Microscopy (SPM) in the failure analysis of a finished and packaged integrated circuit is discussed on these University of Rhode Island web pages. The image below was generated on a NanoScope AFM (courtesy of Digital Instruments, Europe) and shows current leakage points on a failing IC. The figure shows two points of emission which can be investigated in detail by Atomic Force Microscopy (AFM) and Electric Force Microscopy (EFM). These two techniques allow us to look for obvious mis-processing which might show up in the topography when measured by AFM or for electrical problems which may be shown by tracing applied potentials with EFM. To see the AFM and EFM images, refer to these ParkSystems webpages.
In the failure analysis of devices, it is often necessary to remove over-layers, such as the passivation (glass-like) layer that is used to protect the device from moisture andbad actor, mobile-ion species. One must do this in such a way as to maintain the integrity of the under-lying defect or contaminant information. One device used to remove passive-films is the plasma etcher. The source URL for the comparison image set (before plasma etch vs. after plasma etch) is Structure Probe, Inc.
The two SEM micrographs below are copyright photos of SPI SUPPLIES of West Chester, PA. Left: Before etching, original device with glass passivation layer. Right: After etching (anisotropic) with SPI Plasma Prep X etcher for 90 minutes.
Micro-electronics packages deliver integrated circuit technology to printed circuit boards. They are the means by which internal electrical information from the IC can be delivered to the outside world. One can imagine the various failure possibilities and consequences here. Various analytical, non-destructive inspection techniques are available for failure analysis of micro-electronic packages and they include CSAM, which is C-mode Scanning Acoustic Microscopy (C-SAM). Non-destructive failure analysis of IC packages using C-SAM can identify critical defects in three dimensions within the package. Scanning Infrared Microscopy (SIR) can perform non-contact surface temperature measurements of powered ICs and thermal impedance measurements of packages and is another powerful, non-destructive tool to the failure analyst. Below is a C-SAM image showing delamination (red regions) in a micro-electronics package. The reference URL is University of Maryland - College Park Campus.
I hope this has given you some insight into the world of the Failure Analyst. If you want to learn more, surf the internet using some of the terms and concepts introduced on this page. Good Luck!
Metals and Alloys
Concept of Structure
Return to the Materials Engineering Home Page at San Jose State University