Nondestructive metallographic examination of materials frequently must be performed on-site when the component in question cannot be moved or destructively examined. Often, it is imperative that specific microstructural information (i.e., material type, heat treatment condition, homogeneity, etc.) be obtained either before initial use of a component, or before the use of a component can be safely resumed. In this paper, the use of standard metallurgical laboratory equipment, and the procedures required to conduct nondestructive on-site metallographic analyses of engineering materials, is presented. As an example, the materials and metallographic techniques employed in an actual on-site investigation of a gas tungsten-arc weldment joining two large diameter Ti-6Al-4V alloy cylinders are discussed in depth to illustrate what can be accomplished.

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.

This article provides a discussion on the metallographic techniques used for failure analysis, and on fracture examination in materials, with illustrations. It discusses various metallographic specimen preparation techniques, namely, sectioning, mounting, grinding, polishing, and electrolytic polishing. The article also describes the microstructure examination of various materials, with emphasis on failure analysis, and concludes with information on the examination of replicas with light microscopy.

Metallographic examination is one of the most important procedures used by metallurgists in failure analysis. Typically, the light microscope (LM) is used to assess the nature of the material microstructure and its influence on the failure mechanism. Microstructural examination can be performed with the scanning electron microscope (SEM) over the same magnification range as the LM, but examination with the latter is more efficient. This article describes the major operations in the preparation of metallographic specimens, namely sectioning, mounting, grinding, polishing, and etching. The influence of microstructures on the failure of a material is discussed and examples of such work are given to illustrate the value of light microscopy. In addition, information on heat-treatment-related failures, fabrication-/machining-related failures, and service failures is provided, with examples created using light microscopy.

A 2–3 mm thick electroformed nickel mold showed early cracking under thermal load cycles. To determine the root cause, investigators obtained monotonic and cyclic properties of electroformed nickel at various temperatures and identified possible fatigue mechanisms. With the help of finite element modeling, they analyzed the material as well as the design and in-service application of the mold. They discovered that overconstraining the mold, while it was in service, caused excessive thermal stresses which accelerated crack initiation and propagation. Investigators also proposed remedies to prevent additional failures.

Several type 304 stainless steel fire truck water tanks developed through-wall leaks after being in service for approximately two years. One representative tank underwent a comprehensive laboratory analysis, which included metallographic examinations and chemical analyses. The examinations revealed a classic case of microbiologically influenced corrosion (MIC), which preferentially attacked the heat affected zones of the tank welds, resulting in the leaks.

The curved parts of exit pigtails made of wrought Incoloy 800H tubing used in steam reforming furnaces failed by performance after a period of service shorter than that predicted by the designers. Examination of a set of tubes consisting of both curved (perforated) and straight parts revealed that the cracks initiated at the outer surface by a combined mechanism of creep and intergranular embrittlement. A smaller grain size resulting from cold bending fabrication procedures for the curved parts was responsible for accelerating the embrittlement. It was recommended that hot bending be used for fabrication of the curved parts. A change of alloy to a low-alloy chromium-molybdenum allay to protect against heat was also suggested.

The fractured end of a piston rod of a hydraulic press failed in line with the leading face of the piston retaining nut. Although the nut apparently had been seated uniformly, the face was polished, indicating that relative movement between it and the piston had taken place. Failure resulted from the culmination of two principal fatigue cracks which developed on approximately parallel planes from the roots of adjacent threads. A longitudinal section through the screw thread on the piston rod showed it had been carburized but not hardened, and that subsequent surface de-carburization to a depth of approximately 0.001 in. had occurred. It was concluded that insufficient tightening, as evidenced by the polish markings, was the main reason for failure, the portion of the rod therefore being subjected to a greater variation of cyclic stress during operation. The presence of the de-carburized layer lowered its resistance to the initiation of a fatigue crack to that of iron, considerably less than the resistance of the mild steel from which the rod was made and well below that shown by the carburized layer.

Spalled fragments from the work rolls of a steel bar straightening machine were received for failure analysis. Visual inspection coupled with optical and scanning electron microscopy were used as the principal analytical techniques for the investigation. Fractographic analysis revealed the presence of a characteristic fatigue crack propagation pattern (beach marks) and radial chevron marks indicating the occurrence of final overload through a brittle intergranular fracture. The collected evidence suggests that surface-initiated cracks propagated by fatigue led to spalling, resulting in severe work roll damage as well as machine downtime and increased maintenance costs.

The handle from a consumer product exhibited an apparent surface defect. The handle had been injection molded from a medium viscosity grade ABS resin. The anomalous appearance was objectionable to the assembler of the final product and resulted in a production lot being placed on quality-control hold. Investigation included visual inspection, 24x micrographs, and FTIR in the reflectance mode. The spectrum obtained on the included material was characteristic of polybutadiene, the rubber-modifying agent present in ABS. This supported the conclusion that the inclusion's most likely source was an undispersed gel particle formed during the production of the molding resin.

Failure analysis was performed on threaded Ti-6Al-4V fasteners that had fractured in the threads during installation. Scanning electron microscopy (SEM) and optical metallography revealed that the fractures initiated in circumferential shear bands present at the thread roots. The fractures propagated by microvoid coalescence typical of that observed in notched tensile specimen fractures of the same material. For comparison, Ti-6Al-4V fasteners from various commercial sources were tested to failure in uniaxial tension and examined in the SEM. In all cases, the fracture appearances were similar to that exhibited by the fasteners that failed during installation. In addition, results of optical microscopy indicated that the geometry and extent of the shear bands appeared to depend on the fabrication process employed by the individual manufacturers. Causes of shear band formation are discussed along with potential methods to eliminate these microstructural in homogeneities.

A 60.3 mm (2.375 in.) diam drive shaft in the drive train of an overhead crane failed. The part submitted for examination was a principal drive shaft that fractured near a 90 deg fillet where the shaft had been machined down to 34.9 mm (1.375 in.) to serve as a wheel hub. A 9.5 mm (0.375 in.) wide x 3.2 mm (0.125 in.) deep keyway was machined into the entire length of the hub, ending approximately 1.6 mm (0.062 in.) away from the 90 deg fillet. A second shaft was also found to have cracked at a change in diameter, where it was machined down to serve as the motor drive hub. Investigation (visual inspection, inspection records review, optical and scanning electron microscopy, and fractography) supported the conclusion that the fracture mode for both shafts was low-cycle rotating-bending fatigue initiating and propagating by combined torsional and reverse bending stresses. Recommendations included replacing all drive shafts with new designs that eliminated the sharp 90 deg chamfers in favor of a more liberal chamfer, which would reduce the stress concentration in these areas.

Cracking occurred in type 304L stainless steel sheaths on nichrome wire heaters that had been in storage for about 5 years in a coastal atmosphere. The cracks were discovered when the heater coils were removed from storage in their original polyethylene packing materials and straightened for use. Fractography established that fracture occurred by stress-corrosion cracking. The cracks originated at rusted areas on the cladding that occurred under iron particles left on the surface during manufacture. High hardness values indicated that solution annealing following cold working had not been carried out as specified. It was recommended that the sheathing material be fully annealed and that the outer surface be pickled and passivated.

The fatigue failure of a wire rope used on a skip hoist in an underground mine has been studied as part of the ongoing research by the Bureau of Mines into haulage and materials handling hazards in mines. Macroscopic correlation of individual wire failures with wear patterns, fractography, and microhardness testing were used to gain an understanding of the failure mechanism. Wire failures occurred predominantly at characteristic wear sites between strands. These wear sites are identifiable by a large reduction in diameter; however, reduction in area was not responsible for the location of failure. Fractography revealed multiple crack initiation sites to be located at other less noticeable wear sites or opposite the characteristic wear site. Microhardness testing revealed hardening, and some softening, at wear sites.

Piping and structural components used in space launch facilities such as NASA's Kennedy Space Center and the Air Force's Cape Canaveral Air Station face extreme operating conditions. Launch effluent and residue from solid rocket boosters react with moisture to form hydrochloric acid that settles on exposed surfaces as they are being subjected to severe mechanical loads imparted during lift-off. Failure analyses were performed on 304 stainless steel tubing that ruptured under such conditions, while carrying various gases, including nitrogen, oxygen, and breathing air. Hydrostatic testing indicated a burst strength of 13,500 psi for the intact sections of tubing. Scanning electron microscopy and metallographic examination revealed that the tubing failed due to corrosion pitting exacerbated by stress-corrosion cracking (SCC). The pitting originated on the outer surface of the tube and ranged from superficial to severe, with some pits extending through 75% of the tube's wall thickness. The SCC emanated from the pits and further reduced the service strength of the component until it could no longer sustain the operating pressure and final catastrophic fracture occurred. Corrosion-resistant coatings added after the investigation have proven effective in preventing subsequent such failures.

A spiral bevel gear set in the differential housing of a large front-end loader moving coal in a storage area failed in service. The machine had operated approximately 1500 h. Although the failure involved only the pinion teeth, magnetic particle inspection was performed on each part. The 4817 NiMo alloy steel pinion showed no indication of additional cracking, nor did the 4820 NiMo alloy steel gear. The mode of failure was tooth bending fatigue with the origin at the designed position: root radius at midsection of tooth. The load was well centered, and progression occurred for a long period of time. The cause of failure was a suddenly applied peak overload, which initiated a crack at the root radius. Progression continued by relatively low overstress from the crack, which was now a stress-concentration point. This was a classic tooth bending fatigue failure.

Carbon steel axle forgings were rejected due to internal cracks observed during final machining. To determine the cause of the cracks, the preforms of the forging were analyzed in detail at each stage of the forging. The analysis revealed a large central burst in the intermediate stage of the forging preform, which subsequently increased in the final stage. A high upset strain during forging, especially in the final stage, accentuated the center burst by high lateral flow of the metal. It was concluded that the center burst of the axle forging resulted from a high concentration of nonmetallic inclusions in the central portion of the raw bar stock rather than the usual problem of improper forging temperature. Strict control over the inclusion content in the raw material by changing the vendor eliminated the problem.

A portion of two large spur tooth bull gears made from 4147H Cr-Mo alloy steel that had spalling teeth was submitted for evaluation. The gears were taken from a final drive wheel reduction unit of a very large open-pit mining truck. The parts had met the material and initial heat treat hardening specifications. The mode of failure was tooth profile spalling. By definition, spalling originates at a case/core interface or at the juncture of a hardened/nonhardened area. The cause of this failure was either insufficient or no induction-hardened case along the active profile. The cause was activated by a nonfunctioning induction hardening coil that did not or was not allowed to harden the midprofile of several teeth.

The shaft of an exciter that was used with a diesel-driven electric generator broke at a fillet after ten hours of service following resurfacing of the shaft by welding. The fracture surface contained a dull off-center region of final ductile fracture surrounded by regions of fatigue that had been subjected to appreciable rubbing. The fracture appeared to be typical of rotary bending fatigue under conditions of a low nominal stress with a severe stress concentration. It appeared that the fatigue cracks initiated in the surface-weld layer. The weld deposit in the original keyway displays a lack of fusion at the bottom corner. Fatigue fracture of the shaft resulted from stresses that were created by vibration acting on a crack or cracks formed in the weld deposit because of the lack of preheating and postheating. Rebuilding of exciter shafts should be discontinued, and the support plate of the exciter should be braced to reduce the amount of transmitted vibration. Also, the fillet in the exciter shaft should be carefully machined to provide an adequate radius.

Two outboard main-wheel halves (aluminum alloy 2014-T6 forged) from a commercial aircraft were removed from service because of failure. One wheel half was in service for 54 days and had made 130 landings (about 1046 roll km, or 650 roll mi) when crack indications were discovered during eddy-current testing. The flange on the second wheel half failed after only 31 landings, when about 46 cm (18 in.) of the flange broke off as the aircraft was taxiing. Stains on the fracture surfaces were used to determine when cracking was initiated. The analysis (visual inspection, liquid penetrant inspection, and micrographs with deep etching in aqueous 20% sodium hydroxide) supported the conclusion that failure on both wheel halves was by fatigue caused by a forging defect resulting from abnormal transverse grain flow. The crack in the first wheel half occurred during service, and the surfaces became oxidized. Because the fracture surface of the second wheel half had chromic acid stains, it was obvious that the forging defect was open to the surface during anodizing. No recommendations were made except to notify the manufacturer.