Stress corrosion cracking: which materials, and how to avoid it

What is stress corrosion cracking and what three things cause it?

Stress corrosion cracking, or SCC, is a crack that grows in a metal when a tensile stress, a susceptible material and a specific environment are all present. Remove any one and the crack does not start. That definition is general engineering knowledge, not in our data. The Hendrix Group text in Free-cutting brass C36000 says brass needs a sufficiently high tensile stress together with a specific environment, usually ammonia or amines.

The stress does not have to come from the load. Cold working leaves internal stress in brass, and a stress-relief heat treatment removes it. Aluminium bronze C95400 lists stress from working, welding or assembly. Lean duplex stainless steel 2304 says moderate cold bending raises the SCC risk in neutral oxygen-bearing chloride and recommends heat treatment after bending, and Duplex stainless steel 2205 says heavily cold-bent sections should be fully annealed when SCC is possible.

Material and environment are paired, not general. Test stress varies too: NASA loaded 2024 up to 100 percent of yield, and the 2507 threshold is about 50 percent of proof strength. The chart gives yield strengths for the profiled alloys; NASA's actual 2024-T351 plate yield is not in our data.

Which stainless steels crack in chloride, and at what temperature?

The austenitic grades carry the chloride warning. The Stainless steel 304L profile quotes Wikipedia for 304, which it says has the same corrosion resistance as 304L: SCC above about 60 °C (140 °F). That is a 304 comparison, not a measured 304L value. The Alleima data in Duplex stainless steel 2205 extends the 60 °C figure to TP304L and TP316L, and Lean duplex stainless steel 2304 gives the test: 304/304L and 316/316L should not be used above 60 °C in dilute oxygen-bearing chloride at about 8 ppm oxygen.

The temperature is not a hard line. It comes from one maker's test conditions. SSINA text in Stainless steel 444 says that for stainless steels in general, evaporation on hot surfaces can cause cracking well below the immersion thresholds. The chart shows why service ratings do not help: the code limit for 304L sits far above the 60 °C chloride figure.

17-4 PH stainless steel is a martensitic precipitation-hardening grade, and its SCC risk depends on the heat treatment. Jacquet says it is more susceptible at peak strength and should be aged to the lowest hardness compatible with the use where chloride SCC is possible. Carpenter says good resistance comes from hardening at 1025 °F (552 °C) and higher.

The strength cost is in the profile. ASTM A564 bar minimum yield strengths are 1170 MPa for H900, 1000 MPa for H1025 and 725 MPa for H1150.

Do duplex and ferritic stainless steels resist chloride cracking better?

Yes, in the tests our profiles record, though the tests differ. ATI says the ferrite phase in Duplex stainless steel 2205 gives resistance substantially better than the 300-series, and the profile prints no temperature limit. In 40 percent CaCl2 at pH 6.5 and 100 °C, aerated, tests run to 500 hours at most, Alleima reports 2205 resisting much better than TP304L and TP316L.

Lean duplex stainless steel 2304 has the clearest stated limit. Alleima says it can be used in dilute oxygen-bearing chloride up to about 125 °C without risk of SCC, in a test at about 8 ppm oxygen for 1000 hours under a load above yield strength. Above 125 °C with more than about 10 ppm chloride in an oxygen-bearing solution, it says not to use 2304. At low oxygen, higher values are tolerated, but no numbers are printed.

Super duplex stainless steel 2507 data comes from seamless tube and pipe. Alleima reports no SCC up to 1000 ppm chloride at 300 °C and up to 10000 ppm at 250 °C. Those are test conditions, not service ratings: the same profile limits prolonged exposure above 250 °C because of impact strength, and limits sour service to 232 °C. In boiling 45 percent MgCl2 at 155 °C (ASTM G36), parent metal and welds have a threshold stress of about 50 percent of proof strength, and U-bends in hot 25 percent NaCl brine at 108 °C for 1000 hours did not crack.

Stainless steel 444 is the ferritic case. ATI tested U-bends to ASTM G30 in boiling 42 percent MgCl2, 33 percent LiCl and 26 percent NaCl for 1000 hours. Alloy 444 did not crack, while ATI 304 and ATI 316 cracked in every solution, and ATI calls 444 practically immune. SSINA adds that no stainless steel grade is totally immune. Caustic is the exception: in boiling 50 percent sodium hydroxide, ATI's table shows 444 corroding faster than 304 and 316 (293 against 183 and 123 mils per year).

Which aluminium alloys and tempers are prone to stress corrosion cracking?

Two high-strength alloys are flagged, and temper and grain direction decide the outcome. For Aluminium alloy 7075-T6, the NASA handbook says T6 has high stress-corrosion resistance in the longitudinal and long-transverse directions but is susceptible in the short-transverse direction, and that 7075 is attacked by industrial and marine environments. The Kaiser datasheet rates T6 with a C and T73 with a B, and NASA reports better resistance in T73. The profile stores one typical T6 yield, 503 MPa for tube and pipe, and no T73 yield, so the strength cost of the better temper is not in our data.

Aluminium alloy 2024 rests on a 1981 NASA seacoast programme (TM-82393). Short-transverse plate specimens, stressed up to 100 percent of yield, were exposed at the Kennedy Space Center seacoast and in salt water and synthetic seawater. The 2024-T351 temper failed after very short exposure in all three media, and 7075-T651 was also highly susceptible. 2024-T851 showed moderate resistance, with failures at 75 percent of yield. NASA cautions that the results rest on limited lots of plate and may not represent other lots or forms.

Which copper and nickel alloys crack in ammonia or caustic?

Free-cutting brass C36000 cracks in ammonia. The Hendrix Group and Wikipedia text in the profile say brass is susceptible, especially from ammonia or amines,. The crack pattern described is for simple alpha brass, not for C36000, and the profile holds no threshold stress or time to crack.

Aluminium bronze C95400 is the aluminium-bronze case. Hendrix rates aluminium bronzes much superior to brasses in ammonia, though not as good as copper-nickel. The profile grades the severity low, with tensile stress plus an ammonia or ammonia-like atmosphere as the trigger.

Monel 400 splits by environment. Special Metals says it resists SCC and pitting in most fresh and industrial waters, and ATI says it is exceptionally resistant to chloride SCC; neither prints a test stress, medium or time. ATI also says high temperature, high stress and high caustic concentration have produced caustic SCC, with no values printed. Special Metals recommends stress relief as a precaution.

Inconel 600 is more explicit. Its bulletin says the chloride cracking tendency falls as nickel rises. With at least 72 percent nickel, it is virtually immune to chloride-ion SCC. The profile also repeats the bulletin's claim of excellent resistance in high-purity reactor water with no chloride cracking detected. Read that as a chloride statement only: it does not rate primary-water SCC, a known problem for Alloy 600 that is not in our data (general engineering knowledge). The same bulletin says it is subject to SCC in high-temperature, high-strength caustic alkalies, and with mercury at elevated temperature. Material for caustic service should be fully stress-relieved at 1650 °F for 1 hr or 1450 °F for 4 hr, with operating stresses kept to a minimum.

How do you reduce the risk of stress corrosion cracking?

Remove one of the three ingredients: change the environment or material, lower the stress by stress relief, anneal or post-bend heat treatment, or change the temper (7075-T73 over T6, and 17-4 PH aged to the lowest hardness the job allows).

Some cracking that looks like SCC is a related, hydrogen-linked mode. Duplex stainless steel 2205 records sulphide stress cracking limits for sour service, and Super duplex stainless steel 2507 records hydrogen-induced stress cracking on cathodically protected subsea steel. This page does not cover either.

Last, test the real condition. The profiles lean on ASTM G30 U-bends, ASTM G36 boiling MgCl2 and constant-load tests, each narrower than a service environment.

Terms used

Stress corrosion cracking (SCC)

a crack that grows when a metal under tensile stress sits in a specific corrosive environment.

Austenitic stainless steel

the 300-series family, such as 304 and 316.

Duplex stainless steel

a stainless steel that mixes austenite and ferrite. Our profiles credit the ferrite for much of the chloride SCC resistance.

Ferritic stainless steel

a low-nickel stainless steel with a ferrite structure, such as 444.

Temper

the heat treatment or work-hardening state of an alloy, written as a letter and number such as T6 or T73. The same alloy can differ sharply in SCC resistance between tempers.

Short-transverse direction

the through-thickness direction of rolled plate or forged product, at right angles to the rolling and the width (general engineering knowledge, not in our data).

Peak-aged

a precipitation-hardened alloy aged to its highest strength. Overaging goes past that peak and, according to our 17-4 PH profile, improves SCC resistance.

Season cracking

the older name for SCC of brass in ammonia-bearing air.

What to check before you order

1. Write down the environment (chloride level, ammonia, caustic strength), the temperature range, and any wet and dry cycling.
2. List every source of tensile stress, including residual stress from forming and welding, and ask whether stress relief or a full anneal is planned.
3. For aluminium plate and forgings, ask for the temper (T6, T73, T351, T851), the direction the stress acts in, and a lot-specific short-transverse SCC test result. ASTM G47 is the usual method (general engineering knowledge, not in our data).
4. For 17-4 PH, ask for the aging condition (H900, H1025, H1150) and the SCC data behind it at your chloride level and temperature.
5. For stainless steel in chloride, ask which test sits behind any quoted limit (constant load, ASTM G36 boiling MgCl2 or ASTM G30 U-bends) and compare its chloride level, oxygen and temperature with yours.

What do we not know yet?

• Six candidate profiles (stainless 316 and 301, a 5xxx marine aluminium, Ti-6Al-4V, magnesium AZ31B and a quenched-and-tempered alloy steel) hold no SCC statement of their own; 316 appears only as a comparison inside other profiles. A few other profiles outside this article mention SCC and were not read for it.
• Most entries are qualitative. Apart from the stainless steel tests, the profiles hold no threshold stress, time to failure, crack-growth rate or stress intensity limit for any alloy here, so no figure a buyer can put on a purchase order (maximum chloride, temperature or stress ratio) comes from this page.
• No chloride limit is stored for 304L itself, 2205 or 444, and none for 17-4 PH, whose profile gives the aging rule but no temperature or chloride level.
• 7075 holds only a typical T6 tube yield, with no T73 strength. 2024 holds no T851 yield, and its SCC evidence is one limited NASA programme.
• Brass C360 and bronze C954 have no ammonia concentration, stress or time numbers for the alloy itself, and the Monel 400 caustic entry gives no temperature or concentration.
• Not covered by the profiles (general engineering knowledge only): sensitised or welded 304 intergranular and polythionic SCC, the small gain of 316 over 304 in chloride, primary-water cracking of Inconel 600, hydrofluoric-acid cracking of Monel 400, and stainless steel in hot caustic.

Questions people ask