Several mercury diffusion pump stages in the Tritium Purification process at the Savannah River Site (SRS) have been removed from service for scheduled preventive maintenance. These stages have been examined to determine if failure has occurred. Evidence of fatigue around the flange portion of the pump has been seen. In addition, erosion and cavitation inside the throat of the venturi tube and corrosion on the other surface of the venturi tube has been observed. Several measures are being examined in an attempt to improve the performance of these pumps. These measures, as well as the noted observations, are described. Six stages [two machined (MP) and four electron beam (EB) welded] from the mercury diffusion pumps operating in the Tritium Purification process at SRS have been analyzed to determine their condition after nine months of usage. Several cracks were found around the necked region of the two MP stages. The EB welded stages, however, seemed to perform better in service; only two of four stages showed cracking. The cracking is caused by fatigue that has been enhanced by high stresses and tritium in the flange area. The EB welded stage appears to be a step in the right direction. Since the EB weld is a shrink fit, the surface is in compression, thereby eliminating crack propagation. In addition, shot peening has been employed to produce a compressive material surface since fatigue usually originates at the surface. Pitting was observed down the throat of the venturi. This pitting was caused by cavitation and erosion along the length of the venturi tube. Corrosion and pitting was seen on the exterior walls of the diffuser tubes. Stress-corrosion cracks were observed emanating from these corrosion pits. The corrosion likely occurred from the chloride ions present in the process cooling water. Shot peening is now being used in an attempt to place the outside of the diffuser tube in compression to eliminate the stress-corrosion cracking.

The failure mode of through-wall cracking of a butt weld in a 5083-O aluminum alloy piping system in an ethylene plant was identified as mercury liquid metal embrittlement. As a result of this finding, 226 of the more than 400 butt welds in the system were ultrasonically inspected for cracking. One additional weld was found that had been degraded by mercury. A welding team experienced in repairing mercury contaminated piping was recruited to make the repairs. Corrective action included the installation of a sulfur-impregnated charcoal mercury-removal bed and replacement of the aluminum equipment that was in operation prior to the installation of the mercury-removal bed.

An interstage radiator gas coil began leaking after only 45 days of service. The original brass coil with several aluminum fins was replaced three times but each replacement lasted less than a day. After removing the fins, leaks were found at circumferential cracks. A section of a tube was removed and split, revealing a series of cracks, evenly spaced. Crack spacing coincided with fin spacing, indicating that stresses incurred during installation of the fins promoted failure. Metallographic examination showed intergranular, branched cracking, characteristic of stress corrosion failures, with the cracks starting on the inside surfaces of the tubes. There was no known corrosive agent in the system, and no other corrosion damage could be found. Qualitative tests and spectrographic analysis gave a positive indication for mercury. The spacing of the cracks, the branched intergranular cracking, the rapid failure, and presence of mercury led to the conclusion of stress-corrosion cracking. It was impossible to remove mercury from the system so carbon steel coils were substituted for the brass ones. The carbon steel coils gave failure-free service for over nine years.

Failure of three C22000 commercial bronze rupture discs was caused by mercury embrittlement. The discs were part of flammable gas cylinder safety devices designed to fail in a ductile mode when cylinders experience higher than design pressures. The subject discs failed prematurely below design pressure in a brittle manner. Fractographic examination using SEM indicated that failure occurred intergranularly from the cylinder side. EDS analysis indicated the presence of mercury on the fracture surface and mercury was also detected using scanning auger microprobe (SAM) analysis. The mercury was accidentally introduced into the cylinders during a gas-blending operation through a contaminated blending manifold. Replacement of the contaminated manifold was recommended along with discontinued use of mercury manometers, the original source of mercury contamination.

A 1984 Chevrolet Blazer was being pushed by three youths after it ran out of gas when it was hit from behind by a 1979 Mercury Cougar. One of the youths was crushed between the two vehicles and killed. Optical microscopy was used to examine the tungsten filaments from the headlamps of the Cougar and from the four signal/emergency/parking lights of the Blazer to determine whether the vehicles lights were in use at the time of the accident. Based on degree of distension and brittle fragmentation, it was determined that the low-beam headlights of the Cougar and the parking lights of the Blazer were on at the time of the collision.

Several cases of embrittlement failure are analyzed, including liquid-metal embrittlement (LME) of an aluminum alloy pipe in a natural gas plant, solid metal-induced embrittlement (SMIE) of a brass valve in an aircraft engine oil cooler, LME of a cadmium-plated steel screw from a crashed helicopter, and LME of a steel gear by a copper alloy from an overheated bearing. The case histories illustrate how LME and SMIE failures can be diagnosed and distinguished from other failure modes, and shed light on the underlying causes of failure and how they might be prevented. The application of LME as a failure analysis tool is also discussed.

Metal-induced embrittlement is a phenomenon in which the ductility or fracture stress of a solid metal is reduced by surface contact with another metal in either liquid or solid form. This article summarizes the characteristics of solid metal induced embrittlement (SMIE) and liquid metal induced embrittlement (LMIE). It describes the unique features that assist in arriving at a clear conclusion whether SMIE or LMIE is the most probable cause of the problem. The article briefly reviews some commercial alloy systems where LMIE or SMIE has been documented. It also provides some examples of cracking due to these phenomena, either in manufacturing or in service.

Metal-induced embrittlement is a phenomenon in which the ductility or the fracture stress of a solid metal is reduced by surface contact with another metal in either the liquid or solid form. This article summarizes some of the characteristics of liquid-metal- and solid-metal-induced embrittlement. This phenomenon shares many of these characteristics with other modes of environmentally induced cracking, such as hydrogen embrittlement and stress-corrosion cracking. The discussion covers the occurrence, failure analysis, and service failures of the embrittlement. The article also briefly reviews some commercial alloy systems in which liquid-metal-induced embrittlement or solid-metal-induced embrittlement has been documented and describes some examples of cracking due to these phenomena, either in manufacturing or in service.