Following a freight train derailment, part of a fractured side frame was retained for study because a portion of its fracture surface exhibited a rock candy appearance and black scale. It was suspected of having failed, thereby precipitating the derailment. Metallography, scanning electron microscopy, EDXA, and x-ray mapping were used to study the steel in the vicinity of this part of the fracture surface. It was found to be contaminated with copper. Debye-Scherrer x-ray diffraction patterns obtained from the scale showed that it consisted of magnetite and hematite. It was concluded that some copper was accidentally left in the mold when the casting was poured. Liquid copper, carrying with it oxygen in solution, penetrated the austenite grain boundaries as the steel cooled. The oxygen reacted with the steel producing a network of scale outlining the austenite grain structure. When the casting fractured as a result of the derailment, the fracture followed the scale in the contaminated region thus creating the “rock candy” fracture.

A forged 4130 steel cylindrical permanent mold, used for centrifugal casting of gray- and ductile-iron pipe, was examined after pulling of the pipe became increasingly difficult. In operation, the mold rotated at a predetermined speed in a centrifugal casting machine while the molten metal, flowing through a trough, was poured into the mold beginning at the bell end and ending with the spigot end being poured last. After the pipe had cooled, it was pulled out from the bell end of the mold, and the procedure was repeated. Investigation supported the conclusion that failure of the mold surface was the result of localized overheating caused by splashing of molten metal on the bore surface near the spigot end. In addition, the mold-wash compound (a bentonite mixture) near the spigot end was too thin to provide the proper degree of insulation and to prevent molten metal from sticking to the bore surface. Recommendations included reducing the pouring temperatures of the molten metal and spraying a thicker insulating coating onto the mold surface.

A fan support casting failed unexpectedly while running at 1800 rpm in pulp at 65 deg C (150 deg F). The leading edge of the blade exhibited deep spongy holes leading to reduced section and finally to fracture of the part when the remaining section size was insufficient to support the load. Analysis showed the support casting to be a standard 8620 type composition with a hardness of 311 HRB. The design of the casting was not streamlined. There were several square corners present where great pressure differences could be generated. This was a case of erosion-corrosion with the classic spongy appearance of cavitation. Two changes were proposed: streamlining the part to avoid abrupt changes in fluid flow; and a change in alloy to a more corrosion-resistant material (304 or preferably 316) to increase the tenacity of protective films.

In a steel foundry, tensile and bend specimens of castings made in a 2-ton basic arc furnace showed, at irregular intervals, regions with coarse-grained fractures where the specimens broke prematurely, so that the specified strength and toughness values could not be reached. Several cast tensile specimens and some forcibly-broken pieces of the flanges of armature yokes made of cast steel GS C 25 according to DIN 17 245 were investigated. Microscopic examination showed that the cause of damage was the superabundant use of aluminum as deoxidizer. According to recommendations, the aluminum addition was reduced by one-half. Since then, there have been no additional rejects due to insufficient tensile and bend values.

In the EMD-2 Joint Directed Attack Munition (JDAM), the A357 aluminum alloy housing had been redesigned and cast via permanent mold casting, but did not meet the design strength requirements of the previous design. Mechanical tests on thick and thin sections of the forward housing assembly revealed tensile properties well below the allowable design values. Radiology and CT evaluations revealed no casting defects. Optical microscopy revealed porosity uniformly distributed throughout the casting on the order of 0.1 mm pore diam. Scanning electron microscopy revealed elongated pores, which indicated turbulent filling of the mold. Spherical pores would have indicated the melt had been improperly degassed. Based on these findings, it was recommended that the manufacturer analyze and redesign the gating system to eliminate the turbulent flow problem during the permanent mold casting process.

This article focuses on the general root causes of failure attributed to the casting process, casting material, and design with examples. The casting processes discussed include gravity die casting, pressure die casting, semisolid casting, squeeze casting, and centrifugal casting. Cast iron, gray cast iron, malleable irons, ductile iron, low-alloy steel castings, austenitic steels, corrosion-resistant castings, and cast aluminum alloys are the materials discussed. The article describes the general types of discontinuities or imperfections for traditional casting with sand molds. It presents the international classification of common casting defects in a tabular form.

A cast housing, part of a multi-shaft yoking mechanism, failed during assembly and installation of the equipment in which it was to be used. The housing, or yoke body, was cast from AISI 420 grade ferritic stainless steel. Analysis revealed that the failure was caused by the presence of shrinkage cavities, which lowered the load-bearing capability. The failure occurred at the location where there was an abrupt change in the section thickness. A redesign to provide a smooth contour at the section junction was recommended along with optimization of casting parameters to avoid shrinkage cavities.

Threads of a bone screw (Co-Cr-Mo alloy, type ASTM F75) had broken off, and other threads had cracked. 15x sectioning showed porosity, and 155x magnification showed gas holes, segregation, and dissolved oxides. This supports the conclusion that manufacturing defects caused the failure.

A segment from a premium-quality H13 tool steel die for die casting of aluminum failed after only 700 shots. The segment was subjected to visual, macroscopic, hardness, and metallographic testing. The investigation revealed that failure occurred as a result of fatigue at an electrical-discharge-machined surface where the resulting rehardened layer had not been removed. This rehardened layer had cracked, providing a source for fatigue initiation.

A femoral knee implant was returned to the casting vendor for analysis after exhibiting poor bond strength between the cast substrate and a sintered porous coating. Both the coating and the substrate were manufactured from a cobalt-chromium-molybdenum alloy. Metallographic analysis indicated that a decarburized layer existed on all surfaces of the casting, which prevented bonding during the sintering thermal cycle. Bead-to-bead bonding within the coating appeared sufficient, and no decarburized layer was present on the bead surfaces. It was concluded that the decarburization did not occur during the sintering thermal cycle. It was recommended that the prosthetic manufacturer investigate atmosphere controls for all thermal cycles prior to coating.

A new supplier for aluminum die castings was being evaluated, and the castings failed to meet the durability test requirements. Specifically, the fatigue life of the castings was low. Initial inspection of the fatigue fracture surfaces revealed large-scale porosity visible to the naked eye. New castings with reduced porosity also failed the durability tests. The fatigue fracture surfaces of additional casting fragments were very rough and contained multiple ratchet marks along the inner fillet. These observations indicated the fatigue process was heavily influenced by the presence of surface imperfections. Improving the surface finish or choosing a stronger alloy, were more likely to improve part durability than reducing the porosity.

Two sand-cast low-alloy steel equalizer beams (ASTM A 148, grade 105-85) designed to distribute the load to the axles of a highway truck broke after an unreported length of service. Normal service life would have been about 805,000 km (500,000 mi) of truck operation. Investigation (visual inspection, chemical analysis, tensile testing, unetched 65x and 1% nital etched 65x magnification) supported the conclusions that the steel was too soft for the application – probably due to improper heat treatment. Fracture of the equalizer beams resulted from growth of mechanical cracks that were formed before the castings were heat treated. Recommendations included the following changes in processing: better gating and risering in the foundry to achieve sounder castings; better shakeout practice to avoid mechanical damage; better inspection to detect imperfections; and normalizing and tempering to achieve better mechanical properties.

Near-surface porosity in zinc die castings that were subsequently plated with copper, nickel, and bright chromium was causing blemishes in the plating. Identifying die casting turbulence and hot spots were keys to process modifications that subsequently allowed porosity to be greatly minimized.

A 10-cm (4-in.) chain link used in operating a large dragline bucket failed after several weeks in service. The link was made of cast low-alloy steel (similar to ASTM A487, class 10Q) that had been normalized, hardened, and tempered to give a yield strength of approximately 1034 MPa (150 ksi). A hydrogen flake approximately 5 cm (2 in.) in diam was observed at the center of the fracture surface. Beach marks indicative of fatigue encircled the hydrogen flake and covered nearly all of the remaining fracture surface. The failure of this linkways caused by an excessive hydrogen content. Two steps were taken to combat this type of failure. First, when service conditions did not require high hardness to combat wear, the links were produced of a steel having a yield strength of about 690 MPa (100 ksi) rather than 1034 M

A cast steel bracket manufactured in accordance with ASTM A 148 grade 135/125 steel failed in railroad maintenance service. Ancillary property requirements included a 285 to 331 HB hardness range and minimum impact energy of 27 J (20 ft·lbf) at -40 deg C (-40 deg F). The conditions at the time of failure were characterized as relatively cold. Investigation (visual inspection, chemical analysis, and unetched 119x and 2% nital etched 119x SEM images) supported the conclusion that the bracket failed through brittle overload fracture due to a number of synergistic factors. The quenched-and-tempered microstructure contained solidification shrinkage, inherently poor ductility, and type II Mn-S inclusions that are known to reduce ductility. The macro and microscale fracture features confirmed that the casting was likely in low-temperature service at the time of failure. The composition and mechanical properties of the casting did not satisfy the design requirements. Recommendations included exerting better composition control, primarily with regard to melting, deoxidation, and nitrogen control. Better deoxidation practice was recommended to generate the more desirable Mn-S inclusion morphology, and reevaluation of the casting design was suggested to minimize shrinkage.

The front wall of a cast iron crankcase cracked at the transition from the comparatively minor wall thickness to the thick bosses for the drilling of the bolt holes. Metallographic examination showed the case was aggravated by the fact that the casting had a ferritic basic structure and the graphite in part showed a granular formation, so that strength of the material was low. In a second crankcase with the same crack formation the structure in the thick-wailed part was better. But it also showed granular graphite in the ferritic matrix in the thin-walled part between the dendrites of the primary solid solution precipitated in the residual melt. A third crankcase had fractures in two places, first at the frontal end wall and second at the thinnest point between two bore holes. In all three cases casting stresses caused by unfavorable construction and rapid cooling were responsible for the crack formation. A fourth crankcase had cracked in the bore-hole of the frontal face. In this case the cause of the fracture was the low strength of a region that was caused by a bad microstructure further weakened by the bore hole.

A die-cast zinc adapter used in a snowthrower failed catastrophically in a brittle overload manner. The component had a chemical composition similar to standard zinc alloy ZA-27 (UNS Z35840), although the iron content was much higher and the copper slightly lower. The mechanical properties and alloy designation were not specified. Investigation (visual inspection, 187x SEM images, unetched 30x images, hardness testing, and chemical analysis) of both the failed adapter and an exemplar casting from known-good lot supported the conclusion that the casting failed as a result of brittle overload fracture due to excessive iron-zinc phase and gross porosity. These conditions acted synergistically to reduce the strength of the material. The composition was nonstandard, and the inherent brittleness suggested that it was unlikely that this material was an intentional proprietary alloy. No recommendations were made.

Cluster bomb tailcone assemblies each containing two aluminum die-cast components were rejected because of the poor surface condition of the die castings. Numerous heat checks were found on the surfaces of the tailcones and radiographic inspection revealed inclusions, gas holes, and shrinkage defects in the castings. Most of the components failed to meet required mechanical properties because of these casting defects.

Three sprinkler system dry pipe valve castings (class 30 gray iron), two that had failed in service and one that had been rejected during machining because of porosity, were submitted for examination. The two failures consisted of cracks in a seating face. All three were from the same heat. Visual examination showed that the casting had cracked through a thin area in the casting sidewall. Evidence of a sharply machined corner at the fracture site was also discovered. Tensile testing and metallographic analysis revealed no metallurgical cause for the failure. It was recommended that the manufacturer work with the foundry to evaluate the criticality of core placement and to eliminate the undesired thin section.

A failed crosshead of an industrial compressor was examined using optical and scanning electron microscope. The crosshead was an ASTM A148 grade 105-85 steel casting. On the basis of the observations reported and available background information, it was concluded that the failure began with the initiation of cracks at slag inclusions and sharp fillets in weld-repair areas in the casting. The weld-repair procedures were unsatisfactory. The cracks propagated in a fatigue mode. he casting quality was judged unacceptable because of the presence of excessive shrinkage porosity. It was recommended that crosshead castings be properly inspected before machining. Revision of foundry practice to reduce or eliminate porosity was also recommended.