27.2\ Case Study: Aluminum Wire Failures in Residential Wiring

Background: In the early 1970s, aluminum wire was used extensively as a substitute for more expensive copper wire in residential and commercial wiring, specifically for 110 V electrical outlets that used steel screw compressive clamping of the wire against a brass or steel plate. The aluminum wire– steel screw junctions were observed to fail catastrophically through a process of overheating, leading in extreme cases to initiation of structural fires (Meese and Beausoliel 1977; Rabinow 1978). .  Figure 27.6 shows an example of the damage to the wire-screw junction and the surrounding plastic housing and wire insulation caused during an overheating event observed in a laboratory test. This failure was a puzzling occurrence, since aluminum is an excellent electrical conductor and was long used successfully in high voltage electrical transmission lines. Moreover, the vast majority of Al wire–screw connections provided proper service without overheating. However, those connections that did fail in service often produced such catastrophic effects that the critical evidence of the initiation of the failure was destroyed. Capturing an event like that shown in .  Fig. 27.6 required intensive laboratory studies in which thousands of junction boxes were tested and continuously monitored with thermal sensors until a failure initiated, which was then automatically interrupted to prevent complete destruction of the evidence.

This problem illustrates the “macro to micro” sampling problem. The failure mechanism was eventually discovered by SEM/EDS characterization to have a microscopic point of origin, but this microscopic failure origin with micrometer dimensions was hidden within a complex macroscopic structure with centimeter dimensions. Solving the problem required a careful sample preparation strategy to locate the unknown feature(s) of interest. The metallographer mounted the entire Al-wire/steel screw/brass plate assembly in epoxy, as shown in .  Fig. 27.7, and sequentially ground and polished­

Thermal damage to

plastic case

. Fig. 27.6  Residential electrical outlet wired with aluminum. The laboratory test was interrupted after the thermal event initiated and was automatically detected, but significant thermal damage to the plastic casing and wire insulation still occurred

. Fig. 27.7  Metallographic mount (2.5-cm diameter) showing the cross section of the steel screw, aluminum wire, and brass plate

27.2 · Case Study: Aluminum Wire Failures in Residential Wiring

Steel screw

Galvanized

coating

. Fig. 27.8  SEM-BSE image of an anomalous zone of contact between the Al wire and the Fe screw

through the structure until an anomalous region was revealed (.  Fig. 27.8). As shown with SEM/BSE imaging and elemental mapping in .  Fig. 27.9, in this anomalous region the aluminum and iron had reacted to form two distinct Al-Fe zones (Newbury and Greenwald 1980; Newbury 1982). Fixed

beam quantitative X-ray microanalysis with NIST DTSA II and pure element standards (analyses performed during a revisiting of the 1980 specimens) produced the results shown in .  Fig. 27.10, where zone 1 is found to correspond closely to the intermetallic compound FeAl3, while zone 2 corresponds to Fe2Al5. The presence of these intermetallic compounds is significant because of their resistivity. FeAl3 and Fe2Al5 have electrical resistivities of approximately 1 μΩ–m, similar to that of the alloy nichrome (1.1 μΩ–m), which is used for resistive heating elements and which is a factor of 38 higher than pure Al and 10.3 higher than pure Fe. The formation of these intermetallic compounds at the screw-wire contact was initiated when electrical arcing occurred because the connection became mechanically loose due to creep of the Al wire and the poor compliance (springiness) of the wire– screw clamp. Once the local formation of the intermetallic compounds had been initiated by arcing followed by local welding of the Al wire and the steel screw, the increased resistivity caused localized resistive heating that stimulated the interdiffusion of Al and Fe, leading to the further intermetallic compound growth in a runaway positive feedback. Eventually this intermetallic compound zone expanded to dimensions of several hundred micrometers, as seen in

.  Fig. 27.7, creating a resistive heating element that caused

\476 Chapter 27 · X-Ray Microanalysis Case Studies

Al 0.759 a/o

Fe 0.241 a/o

FeAl3

Mounting epoxy

Al 0.719 a/o Fe 0.281 a/o Fe2Al5

Steel screw

the damage seen in .  Fig. 27.6. (Note that the practical solution to this problem was to modify the wire connections to provide much greater springiness to eliminate the opening of gaps that allowed arcing to occur.)

27.3\ Case Study: Characterizing

the Microstructure of a Manganese

Nodule

“Manganese nodules” are rock concretions that form on the deep sea floor through the action of microorganisms that precipitate solid chemical forms from metals dissolved in the water, often in close association with hydrothermal vents.

The elemental composition of a polished cross section of a manganese nodule, shown in .  Fig. 27.11, was examined by SEM/BSE and by SEM/EDS X-ray spectrum imaging elemental mapping. The SEM/BSE image in .  Fig. 27.12 reveals a complex layered microstructure that suggests non-uniform deposition of the precipitated minerals over time. This non-­uniform deposition is confirmed by the elemental maps for O, Mn, and Ni and color overlay shown in .  Fig. 27.13 and for the Mn, Fe, and Ni maps shown in .  Fig. 27.14. Note the oxygen-­rich areas (green) in .  Fig. 27.13. These regions correspond to silica and aluminosilicate grains within the manganese nodule, as revealed in .  Fig. 27.15. The composition measured with a fixed beam placed at the center of the field of view is listed in .  Table 27.1, showing the high abundance of Mn as a major constituent and the presence of other transition elements (e.g., Fe, Ni, and Cu) as minor constituents. .  Figure 27.16 shows the results of quan-

. Fig. 27.11  Manganese nodule

titative processing of the XSI with the k-ratio/matrix correction protocol using DTSA-II. The resulting concentration maps have been encoded with the logarithmic threeband color scheme shown in .  Fig. 27.16, enabling quantitative comparison of the constituents, using NIST Lispix.

Note that some features in the elemental maps are a result of artifacts. Thus, the cracks noted in the SEM/BSE of

.  Fig. 27.11 are also seen in the O elemental map, but not in the Mn or Ni maps. The origin of this artifact is the difference in the photon energies of these elements. The O K-shell X-ray