Stress corrosion cracking (SCC)

Stress corrosion cracking or "season cracking" occurs only in the simultaneous presence of a sufficiently high tensile stress and a specific corrosive environment. For brasses the environment involved is usually one containing ammonia or closely related substances such as amines, but atmospheres containing between 0.05% and 0.5% of sulphur dioxide or nitrites by volume can also cause stress corrosion cracking. Cracking caused by liquid mercury is the basis of the mercurous nitrate test for excessive internal stress in tubes and components of brass and other copper alloys specified, for example, in BS2871: part 3, BS EN ISO 196 and ASTM B 154. Mercury stress corrosion cracking of brass components can also occur in service due to contamination from broken thermometers.

Recognition

Stress corrosion cracking in brass is usually localised and, if ammonia has been involved, may be accompanied by black staining of the surrounding surface. The fracture surface of the crack may be stained or bright, according to whether the crack propagated slowly or rapidly. The cracks run roughly perpendicular to the direction of the tensile stress involved. For example, drawn brass tube that has not been stress relief annealed has a built-in circumferential hoop-stress; consequently exposure to an ammoniacal environment is liable to cause longitudinal cracking. Stress corrosion cracking in pipes that have been cold bent without a subsequent stress relief anneal occurs typically along the neutral axis of the bend. Stress corrosion cracking due to operating stresses is transverse to the axis of the applied stress.

Examination of metallographic sections through cracked areas will usually show a markedly intergranular crack pattern in simple alpha brasses. In aluminium brass the cracking is transgranular and much branched and in admiralty brass either or both forms of cracking may be observed. Stress corrosion cracks in alpha-beta brasses run transgranularly through the beta phase or, occasionally, along the alpha-beta interface. The cracks look discontinuous in metallographic sections, as they divert above or below the plane of the section to pass round the alpha phase.

Influence of zinc content and stress level

D H Thompson and A W Tracey made a detailed study of the effect of stress level and zinc content on the time for failure by stress corrosion cracking to take place in axially loaded specimens exposed to air containing 10% ammonia and 3.7% water vapour at 35oC. This is an accelerated test giving failures in much shorter times than would be experienced under most service conditions; the results, presented in figure 8, are therefore to be taken as indicative of trends but should not be used to predict service life. It does show that the higher the copper content, the better the resistance to stress corrosion cracking.

Figure 14 Effect of zinc content on stress corrosion susceptibility of brass

Accelerated tests in an ammoniacal atmosphere at three different stresses

In another series of experiments, D H Thompson () used loop specimens to study the effect of adding a third element on the stress corrosion behaviour of various brasses in a moist ammoniacal atmosphere. The results showed marked beneficial effects of nickel - the 10% nickel, 25% zinc, nickel silver tested being superior to 15% zinc brass without additions. Addition of silicon to a 17% zinc brass was also beneficial. Similar results to these have been found by other researchers and are supported by practical experience.

A further point of interest arising from Thompson’s and Tracy’s loop tests is that aluminium brass was shown to have better stress corrosion resistance than admiralty brass. This was confirmed in atmospheric stress corrosion tests of various copper alloys carried out by J M Popplewell and T C Gearing () . U-bend specimens of aluminium brass exposed to industrial atmospheres at Newhaven and Brooklyn failed in times ranging from 221 to 495 days while Admiralty brass specimens failed between 41 and 95 days. Both materials were in the 40% cold rolled condition.

It has occasionally been suggested that arsenic levels near the 0.06% maximum permitted by ISO and most national standards may increase the susceptibility of aluminium brass to stress corrosion, but a survey of relevant publications by H S Campbell (), concluded that reducing the maximum arsenic content from 0.06 to 0.03% would have only a marginal effect on stress corrosion susceptibility and would reduce the reliability of the arsenic addition as an inhibitor of dezincification. Consequently, no change in the standards was considered desirable.

The test results and practical experience outlined above refer to alpha or alpha-beta brasses and principally to ammoniacal environments, though sulphur dioxide may have been the more important corrosive factor in the industrial atmospheric exposure tests. All-beta brass (the only important commercial example of which is the cast high tensile brass HTB3) is susceptible to stress corrosion cracking also in environments containing chlorides and is therefore much more restricted in use.

Avoidance

Provided that service and manufacturing process requirements permit, improved resistance to stress corrosion cracking can be achieved by selecting the less susceptible brasses - low zinc rather than high zinc alloys; nickel silver rather than simple brass; aluminium brass rather than admiralty; HTB1 rather than HTB3 for example. Since, however, all brasses are susceptible to stress corrosion cracking to a greater or less extent it is more important to control manufacturing, assembly and operating conditions to avoid the combination of high stress and unfavourable environment that may cause stress corrosion.

Cold working operations such as pressing, spinning, drawing and bending leave internal stresses which, unless removed or substantially reduced by stress relief heat treatment, can lead to stress corrosion cracking. The optimum time and temperature for stress relief depends upon the alloy but will lie within the range ½ to 1 hour at 250-300oC. A second, avoidable source of dangerously high stress levels that can induce stress corrosion cracking is careless fitting in assembly and installation. Poor alignment, gaps at joints and overtightening of bolts are obvious examples of bad practice in this respect. One that is not so often recognised is the practice of screwing taper-threaded connectors into parallel-threaded brass valves. Especially when ptfe tape is used to seal the thread, it is all too easy to overtighten such joints to a point where a very high circumferential "hoop" stress is generated in the female member. There have been many examples of subsequent longitudinal stress corrosion cracking of the valve ends as a result of contact with quite low concentrations of ammonia in service.

The control of the environment in which brass is used may seem an impractical way of ensuring freedom from stress corrosion cracking in service, in view of the wide range of service conditions under which brass articles and components are in daily use, but it is possible to avoid unnecessary exposure to ammoniacal contamination. One source of such contamination that has caused brass fittings, overstressed in assembly, to crack in service is some varieties of foamed plastic insulating material in which amines or other ammonia-related chemicals are used as foaming or curing agents. Chilled water valves in air conditioning units are most likely to be affected since these are subjected to condensed moisture as well as the ammoniacal chemicals. More common, but usually less harmful, sources of ammonia are latex cements used to fix wall and floor tiles and certain household cleaners (which usually advertise their ammonia content as one of their great advantages). The best advice regarding these possible sources of trouble is to provide good ventilation after using latex cement, so that any stressed brass articles in the room have only a short period of exposure to ammonia, and to wash away ammoniacal household cleaner residues after use.

< < < click here to go back to the previous page  ::