In the middle 1940s, when I developed the nickel-based brazing filler metals for the aircraft jet engine high-temperature applications, we ran many physical property tests. Among these was the test for the stress rupture strength of both butt-brazed test specimens and single-lap-shear specimens. These data were published in the August 1952 Welding Journal, along with the summary of the three years of testing, to qualify the nickel-based brazing filler metals for jet-engine applications.

At that time, the only high-temperature brazing filler metal was the 85Ag-15Mn (now BAg-23) silver brazing filler metal that had been used in the German jet engines during WW II to make hollow, high-temperature turbine blades. After WW II, the filler metal was introduced to the United States. In the middle 1940s, jet engines were running hotter, and if an engine overheated, the 85Ag-15Mn would melt and be blown out of the joint and jet engine. Therefore, need arose for a brazing filler metal that would braze at normal brazing temperatures that would have a high-remelt temperature. The nickel-based filler metal that was developed, when fully diffusion brazed, had a remelt temperature in excess of 2500°F (1371°C).

The paper that was published in the August 1952 Welding Journal is titled "Design Properties of Brazed Joints for High-Temperature Applications," by Robert L. Peaslee and Willard M. Boam.

The referenced paper included stress rupture data of AMS 5770 base metal tested at 1500°F (816°C). These data were for butt brazed test specimens that were fully diffusion brazed. The graph shows the comparison between the unprocessed base metal from the manufacturer's data, and the butt brazed specimens diffusion brazed with Nicrobraz 125 (AWS BNi-1, AMS 4775) at 2150°F (1121°C). The diffusion brazed and aged stress-rupture data followed the lower range of the base metal stress-rupture data from the manufacturer. This was to be expected, as the brazing temperature was higher than the recommended solution temperature for that base metal. The brazed specimens failed at around 220 hours when loaded to approximately 12,000 lb/in.2, at 1500°F (816°C). Thus the diffusion brazed test data were comparable to the base metal data.

A later stress-rupture test was run on a diffusion brazed V-type joint of A286, UNS S66286 iron base metal, brazed with BNi-1 at 2150°F (1177°C). A stress- rupture test was run at 1350°F (732°C) and the failure occurred at the same number of hours that the manufacturer's data predicted.

I have two papers in my file containing stress rupture data:

1. "Diffusion Brazed Inconel 713C," by A. Sokamoto and M. Ohsumi of Nagoya Aircraft Works, Mitsubishi Heavy Industry, Ltd. This is a 29-page manuscript presented at the Japanese Welding Society Meeting held April 8 and 9, 1980. The paper discusses the stress-rupture data, and contains butt-brazed-joint data, micrographs of joints, and chemical element traces across the joints. The paper is very detailed but, unfortunately, no comments on the data are included.

2. "High-Temp Stress-Rupture Properties of Brazed Joints in Heat-Resistant Alloys," Research Report A1250, September 1959, by A. Cibula, published by The British Non-Ferrous Metal Research Association. Stress rupture data were obtained using a plug-in socket-type brazed joint. Austenitic steels were tested at 500°C (932°F). Nimonic 90 tested at 800°C (1472°F) and 900°C (1652°F). Later tests were run on Nimonic 90 at 950°C (1742°F) and Nimonic100 at 800°C (1472°F) and 950°C (1742°F).

It is my belief, from the work that has been done, that when full diffusion brazing takes place, the interdiffusion between base metal and brazing filler metal make the joint area respond the same as the original base metal; thus the stress- rupture properties of the joint will equal those of the base metal.

When BNi-1, 1a, 2, 3, or 4 are used in a joint that is fully diffusion brazed, the filler metal will be completely dissolved, and only nickel-based metal grains will be apparent in the joint area.

In reference to nickel brazing filler metals containing silicon or phosphorus as the melting point depressant, I do not have any data. These brazing filler metals can also be diffusion brazed, but require special consideration. As an example, BNi-7 can be diffusion brazed with 304L, if the joint is a light-press fit and held at the brazing temperature of 1950°F or higher for 60 min. The resulting joint will exhibit 50 to 90% base metal. With a higher temperature, and/or longer diffusion time, the brazed joint area would be 100% base metal in appearance on a micro specimen when viewed at 100 or 500 power magnification.

While I do not have all the data that would be desirable, what I have seen indicate that a fully diffusion brazed joint with a boron-containing brazing filler metal will give the desired long-term stress-rupture properties.