Failure analysis is an investigative process in which the visual observations of features present on a failed component and the surrounding environment are essential in determining the root cause of a failure. This article reviews the basic photographic principles and techniques that are applied to failure analysis, both in the field and in the laboratory. It discusses the processes involved in visual examination, field photographic documentation, and laboratory photographic documentation of failed components. The article describes the operating principles of each part of a professional digital camera. It covers basic photographic principles and manipulation of settings that assist in producing high-quality images. The need for accurate photographic documentation in failure analysis is also presented.

The goal of using nondestructive evaluation (NDE) in conjunction with failure analysis is to obtain the most comprehensive set of data in order to characterize the details of the damage and determine the factors that allowed the damage to occur. The NDE results can be used to determine optimal areas upon which to focus for sectioning and metallography in order to further investigate the condition of the component. This article provides information on the inspection method available for failure analysis, including standard methods such as visual testing, penetrant testing, and magnetic particle testing. It covers the effects of various factors on the properties of the part that may impact failure analysis, describes the characterization of damage modes and crack sizes, and finally discusses the processes involved in application of NDE results to failure analysis.

Component slippage in the left-side final drive train of a tracked military vehicle was detected after the vehicle had been driven 13,700 km (8500 miles) in combined highway and rough-terrain service. The slipping was traced to the mating surfaces of the final drive gear and the adjacent splined coupling sleeve. Specifications included that the gear and coupling be made from 4140 steel bar oil quenched and tempered to a hardness of 265 to 290 HB (equivalent to 27 to 31 HRC) and that the finish-machined parts be single-stage gas nitrided to produce a total case depth of 0.5 mm (0.020 in.) and a minimum surface hardness equivalent to 58 HRC. Investigation (visual inspection, low-magnification images, 500X images of polished sections etched in 2% nital, spectrographic analysis, and hardness testing) supported the conclusion that the failure occurred by crushing, or cracking, of the case as a result of several factors. Recommendations included reducing the high local stresses at the pitch line to an acceptable level with a design modification. Also suggested was specification of a core hardness of 35 to 40 HRC to provide adequate support for the case and to permit attainment of the specified surface hardness of 58 HRC.

An irrigation pipe made of medium-density PE failed during service. This pipe was subjected to severe cyclic-bending strain of the order of 6% while under tensile stress of approximately 6.9 MPa (1000 psi) and a hoop stress of approximately 6.2 MPa (900 psi), far more stringent conditions than those encountered in most applications of PE pipes. Visual inspection and reflected-light optical micrographs were used to plot bandwidth as a function of crack length. The conclusion was that, contrary to the dominant belief that pipe failure initiates from surface defects, a critical size flaw within the pipe wall can also initiate failure as it did in this case. Recommendations included that similarity criteria should be established between the fracture behavior of a component in service and that observed in the laboratory.

Two broken ball-mill liners from a copper-mine ore operation were submitted for failure analysis. These liners failed prematurely, having reached less than 20% of their expected life. The chemical composition of the liners was within specifications for high-chromium white cast iron. The two broken liners were sand blasted for visual inspection and subsequent metallography and hardness testing. Many cracks were found externally and on the undersides. There were also signs of mechanical damage that occurred inside the mill before detection of the failures. The underside cracking is significant because the user advised that the liners were not backed in the installation. Cracking was present in the microstructures of both liners. These cracks tend to fracture the brittle carbide phase first; once nucleated, the sharp cracks can propagate and grow to critical dimensions, which eventually induces complete failure to the load-bearing section. The premature failure of these liners was caused by severe localized overstress conditions due to localized impact in service. Proper backing of shell liners should be ensured to reduce the effect of impact forces in the ball mill.

Both halves of a gray cast iron transmission housing from a 50-ton dump truck were found to contain numerous cracks. The housing material was possibly G3000 grade designation for automotive gray cast iron. No service duration or material specifications were provided. Investigation (visual inspection, tensile testing, 2% nital etched 59x cross sections, and metallographic analysis) supported the conclusion that failure was due to applied stresses sufficient to fracture the castings which exhibited brittle overload cracks at highly stressed locations. No recommendations were made.

The upper frame from a large cone crusher failed in severe service after an unspecified service duration. The ductile iron casting was identified as grade 80-55-06, signifying minimum properties of 552 MPa (80 ksi) tensile strength, 379 MPa (55 ksi) yield strength, and 6% elongation. Investigation (visual inspection, chemical analysis, unetched 30x images, and 2% nital etched 30x images) was difficult because the fracture surface of the frame section was obliterated by postfracture corrosion. Repeated attempts at cleaning using progressively stronger chemicals revealed that no telltale fracture morphology remained. However, the investigation supported the conclusion that the crusher frame failed via brittle overload fracture, likely due to excessive service stresses and substandard mechanical properties. Recommendations included additional quality-control measures to provide better spheroidal graphite morphology at the frame surface.

Visual examination, optical and scanning electron microscopy were used to determine the cause of failure in the connector groove of a marine riser coupling. The specified steel was AISI 4142 (0.40 to 0.45% C; 0.75 to 1.00% Mn; 0.20 to 0.35% Si; 0.80 to 1.10% Cr; 0.15 to 0.25% Mo) normalized from 9000C. Microscopic examination revealed the crack's initiation point and subsequent propagation. SEM examination of chemically stripped corrosion showed that corrosion fatigue and stress corrosion might have contributed to the initial slow crack growth. Impact tests revealed a fracture transition temperature in excess of 1000C. The sequence of events leading to failure was detailed. The main recommendation was to quench and temper existing couplings and to use a lower carbon quenched and tempered steel for new couplings.

Electrical contact-finger retainers blanked and formed from annealed copper alloy C65500 (high-silicon bronze A) failed prematurely by cracking while in service in switchgear aboard seagoing vessels. In this service they were sheltered from the weather but subject to indirect exposure to the sea air. About 50% of the contact-finger retainers failed after five to eight months of service aboard ship. Investigation (visual inspection, 250x images etched with equal parts NH4OH and H2O2, emission spectrographic analysis, and stereoscopic views) supported the conclusion that the cracking was produced by stress corrosion as the combined result of: residual forming and service stresses; the concentration of tensile stress at outer square corners of the pierced slots; and preferential corrosive attack along the grain boundaries as a result of high humidity and occasional condensation of moisture containing a fairly high concentration of chlorides (seawater typically contains about 19,000 ppm of dissolved chlorides) and traces of ammonia. Recommendations included redesign of the slots, shot-blasting the formed retainers, and changing the material to a different type of silicon bronze-copper alloy C64700.

A pushrod made by inertia welding two rough bored pieces of bar stock installed in a mud pump fractured after two weeks in service. The flange portion was made of 94B17 steel, and the shaft was made of 8620 steel. It was disclosed by visual examination that the fracture occurred in the shaft portion at the intersection of a 1.3 cm thick wall and a tapered surface at the bottom of the hole. The fatigue crack was influenced by one-way bending stresses initiated at the inner surface and progressed around the entire inner circumference. A heavily decarburized layer was detected on the inner surface of the flange portion and sharp corner was found at the intersection of the sidewall and bottom of the hole. It was concluded that the stress raiser due to the abrupt section change was accentuated by decarburized layer. As a corrective measure, the design of the pushrod was changed to a one-piece forging and circulation of atmosphere during heat treatment was permitted through a hole drilled in the flange end of the rod to avoid decarburization.

After about two years in service, a 303 stainless steel valve in contact with a carbonated soft drink in a vending machine occasionally dispensed a discolored drink with a sulfide odor. According to the laboratory at the bottling plant, the soft drink in question was strongly acidic, containing citric and phosphoric acids and having a pH of 2.4 to 2.5. Investigation (visual inspection, chemical analysis, immersion testing in the soft drink, and 100x unetched micrographs) supported the conclusion that the failure was caused by the size and distribution of sulfide stringers in the alloy used in the valve. Manganese sulfide stringers in the valve were exposed at end-grain surfaces in contact with the beverage. The stringers, which were anodic to the surrounding metal, were subject to corrosion, producing a hydrogen sulfide concentration in the immediately adjacent liquid. Recommendations included changing the valve material to type 304 stainless steel.

Socket head cap screws used in a naval application were failing in service due to delayed fracture. The standard ASTM A 574 screws were zinc plated and dichromate coated. Investigation (visual inspection, 1187 SEM images, chemical analysis, and tension testing) of both the failed screws and two unused, exemplar fasteners from the same lot supported the conclusion that the cap screws appear to have failed due to hydrogen embrittlement, as revealed by delayed cracking and intergranular fracture morphology. Static brittle overload fracture occurred due to the tension preload, and prior hydrogen charging that occurred during manufacturing. The probable source of charging was the electroplating, although postplating baking was reportedly performed as well. Recommendations included examining the manufacturing process in detail.

An austenitic stainless steel (type 316/316L stainless steel, schedule 40, 64 mm (2.5 in.) diam and larger) piping network used in the fire-sprinkler system in a large saltwater passenger and car ferry failed by rapid leaking. Operating conditions involved stagnant seawater at ambient temperatures. The pipe was in service for four weeks when three leaks appeared. Investigation (visual inspection and photographic images) supported the conclusion that the failure was caused by attack and corrosion damage of Cl ions in conditions that were ideal for three modes of highly accelerated pitting of austenitic stainless steel: the bottom surface, weld or HAZ pits, and crevices. Recommendations included proper material selection for piping, flanges, and weld rods with greater corrosion resistance. Proper filtering to prevent entrained abrasives and timely breakdown inspections were also advised.

A motorboat engine connecting rod forged from carbon steel fractured in two places and cracked at the small end during service. The analysis (visual inspection, 50x micrographs of sections etched with 2% nital, magnetic-particle inspection, and metallographic examination) supported the conclusion that the connecting rods were rendered susceptible to fatigue-crack initiation and propagation by the notch effect of coarse folds formed during the forging operation. One fracture was caused by fatigue resulting from operating stresses, and the other was a secondary tensile fracture. No recommendations were made.

Cracks occurred in a new ship hull after only three months in service. It was noted that the 5xxx series of aluminum alloys are often selected for weldability and are generally very resistant to corrosion. However, if the material has prolonged exposure at slightly elevated temperatures of 66 to 180 deg C (150 to 350 deg F), an alloy such as 5083 can become susceptible to intergranular corrosion. Investigation (visual inspection, corrosion testing, SEM images) supported the conclusion that the cracks occurred because during exposures to chloride solutions like seawater, galvanic couples formed between precipitates and the alloy matrix, leading to severe intergranular attack. No recommendations were made.

During a routine inspection, cracks were discovered in several aluminum alloy (similar to either 2014 or 2017) coupling nuts on the fuel lines of a missile. The fuel lines had been exposed to a marine atmosphere for six months while the missile stood on an outdoor test stand near the seacoast. A complete check was then made, both visually and with the aid of a low-power magnifying glass, of all coupling nuts of this type on the missile. Investigation (visual inspection, spectrographic and chemical analysis, and metallographic examination) supported the conclusion that the cracking of the aluminum alloy coupling nuts was caused by stress corrosion. Contributing factors included use of a material that is susceptible to this type of failure, sustained tensile stressing in the presence of a marine (chloride-bearing) atmosphere, and an elongated grain structure transverse to the direction of stress. The elongated grain structure transverse to the direction of stress was a consequence of following the generally used procedure of machining this type of nut from bar stock. Recommendations included changing the materials specification for new coupling nuts for this application to permit use of only aluminum alloys 6061-T6 and T651 and 2024-T6, T62, and T851.

This paper describes the metallurgical investigation of a broken spindle used to attach an antenna to the mast of a naval vessel. Visual inspections of both failed and intact fastener assemblies were carried out both on-board ship and in the laboratory followed by metallographic and fractographic examinations. Simulations were also performed on stressed material in a suitable environment to assess the relative importance of postulated failure mechanisms. Factors contributing to this failure including assembly procedures and applied preloads, service loading and environment, and material selection and specification. The discussion considers whether this failure was an isolated incident or is likely to be a fleet-wide problem, and suggests ways to prevent reoccurrence.

A brazing-furnace muffle 34 cm (13 in.) wide, 26 cm (10 in.) high, and 198 cm (78 in.) long, was fabricated from nickel-base high-temperature alloy sheet and installed in a gas-fired furnace used for copper brazing of various assemblies. The operating temperature of the muffle was reported to have been closely controlled at the normal temperature of 1175 deg C (2150 deg F); a hydrogen atmosphere was used during brazing. After about five months of continuous operation, four or five holes developed on the floor of the muffle, and the muffle was removed from service. Analysis (visual inspection, x-ray spectrometry, and metallographic examination) supported the conclusion that the muffle failed by localized overheating in some areas to temperatures exceeding 1260 deg C (2300 deg F). The copper found near the holes had dripped to the floor from assemblies during brazing. The copper diffused into the nickel-base alloy and formed a grain-boundary phase that was molten at the operating temperature. The presence of this phase caused localized liquefaction and weakened the alloy sufficiently to allow formation of the holes. No recommendations were made.

Three radiant tubes, made of three different high-temperature alloys, were removed from a carburizing furnace after approximately eight months of service when they showed evidence of failure by collapsing (telescoping) in a region 30 cm (12 in.) from the tube bottoms in the vicinity of the burners. The tubes had an original wall thickness of 3.0 mm (0.120 in.) and were made of three different alloys: the first was Hastelloy X; the second alloy was RA 333, a wrought nickel-base heat-resistant alloy; and the third was experimental alloy 634, which contained 72% Ni, 4% Cr, and 3.5% Si. The three radiant tubes had been operated at a temperature of about 1040 deg C (1900 deg F) to maintain furnace temperatures of 900 to 925 deg C (1650 to 1700 deg F). Analysis (visual inspection and micrographic examination) supported the conclusion that all three tubes failed by corrosion. Recommendations included replacing the material with an alloy, such as RA 333, with a higher chromium content and with an additional element, like silicon, resistant to carburization-oxidation.

Severely reduced wall thickness was encountered at the liquid line of a lead-bath pan that was used in a continuous strip or wire oil-tempering unit. Replacement of the pan was necessary after six months of service. The pan, 6.9 m (22.5 ft) long, 0.6 m (2 ft) wide, and 38 cm (15 in.) deep with a 2.5-cm (1-in.) wall thickness, was a type 309 stainless steel weldment. Operating temperatures of the lead bath in the pan ranged from 805 deg C (1480 deg F) at the entry end to 845 deg C (1550 deg F) at the exit end. Analysis (visual inspection. metallographic analysis, moisture testing, and etched micrographs using Murakami's reagent) supported the conclusions that thinning of the pan walls at the surface of the molten lead resulted from using coke of high moisture content and from the low fluctuating coke level. Recommendations included reducing the supply of oxygen attacking the grain boundaries and the hydrogen that readily promoted decarburization with the use of dry (2 to 3% moisture content) coke. Maintaining a thick layer of coke over the entire surface of molten lead in the pan would exclude atmospheric oxygen from the grain boundaries.