Failure Analysis Data

Types of Embrittlement in Steels

Some fractures arrive with no overload, no wear and a clean certificate, and the cause is a change in the steel itself that the process record explains. This page collects the embrittlement types the source lists for ferrous alloys, the temperature window or environment that triggers each one, and the way each is avoided.

The embrittlement types and what triggers each one The table below is the summary the source gives for ferrous alloys, with the susceptible steels, the trigger and the result. The trigger column is the part to read against a process record, because most of these conditions are visible in the heat treatment or service log rather than in the fracture.
Embrittlement typeSteels affectedWhat triggers itWhat it does
Strain-age embrittlementLow carbon steelPrecipitation after a deformation process, with aging at room or slightly elevated temperatureStrength rises, ductility falls. About 15 percent cold reduction gives the maximum effect, and it can take months at room temperature or minutes at 400 °F (200 °C).
Quench-age embrittlementLow carbon steelQuenching from below the lower critical temperature, followed by precipitation during agingStrength rises, ductility falls. Steels between 0.04 and 0.12 percent carbon are the most susceptible, and the quench has to start above about 1040 °F (560 °C).
Blue brittlenessCarbon and alloy steelsExposure between 450 and 700 °F (230 and 370 °C)Strength rises and both ductility and impact strength drop sharply. It is an accelerated form of strain-age embrittlement, and deformation inside the temperature band makes it worse.
Stress-relief embrittlement and reheat crackingAlloy and stainless steelsPostweld heat treatment of a precipitation hardening alloy systemToughness falls in the heat affected zone or the weld metal, and severe cases crack along grain boundaries. Vanadium, molybdenum and boron raise the risk.
Temper embrittlementCarbon and low alloy steelsExposure between 660 and 1060 °F (350 and 570 °C), or slow cooling through that band after high temperature temperingThe ductile to brittle transition temperature rises, by as much as 360 to 540 °F (200 to 300 °C) in reported cases, while strength and hardness stay where they were.
Tempered martensite embrittlementHeat treated alloy steelTempering between 400 and 700 °F (200 and 370 °C), with service at room temperatureImpact energy falls, worst near 600 °F (315 °C), while tensile strength, yield strength and elongation are unchanged. Also called one-step temper embrittlement.
400 to 500 °C embrittlementStainless steel above 15 percent chromiumLong exposure between 750 and 930 °F (400 and 500 °C), most severe near 885 °F (475 °C)Strength rises and ductility falls, and corrosion resistance drops beside the grain boundaries. The reaction reverses on heating above the band.
Sigma phase embrittlementFerritic and austenitic stainless steelLong exposure between 1050 and 1800 °F (565 and 980 °C)Toughness falls and notch sensitivity rises, and the loss of toughness is almost complete once the part cools below 500 °F (260 °C). Corrosion resistance and creep resistance fall with it.
GraphitizationCarbon and low alloy steelLong service above about 800 °F (425 °C), typically 850 to 1100 °F (450 to 600 °C)Toughness falls where graphite forms along a continuous path through a load carrying member. Aluminium deoxidized steels are the most exposed, and chromium above 0.7 percent resists it.
Intermetallic compound embrittlementAll steelElevated temperature contact with a metal that forms an intermetallic phase, such as zinc on galvanized steelA brittle intergranular phase forms at the grain boundaries. No melting is involved, and the damage is irreversible.
Neutron embrittlementAll steelNeutron irradiation in nuclear serviceThe ductile to brittle transition temperature rises, by as much as 360 °F (200 °C), and intergranular fracture becomes more likely. Nickel content and copper or phosphorus residuals raise the sensitivity.
Hydrogen embrittlementCold worked or heat treatment hardened steelHydrogen charging during steelmaking, pickling, electroplating, welding or corrosion in serviceDuctility falls and the part cracks in a delayed manner under static load, often below the yield strength and with no visible plastic deformation.
Liquid metal induced embrittlementAll steelContact with a molten low melting point metal, under tensile stress and wetting of the surfaceDuctility falls sharply and a crack runs along grain boundaries at speeds measured in inches per second. Small parts fail almost instantly once the stress is high enough.
Solid metal induced embrittlementAll steelContact with a near molten or solid low melting point metal, under tensile stressDuctility falls and the part fails in a delayed manner under static load, with time to failure rising as the stress falls. Crack growth is temperature dependent.

Source: ASM Handbook, Vol. 11, Failure Analysis and Prevention (ASM International), the summary of the types of embrittlement experienced by ferrous alloys in the article Overload Failures.

  • The list is a summary of ferrous alloys, and the source notes that nonferrous metals have their own embrittling phenomena. It also notes that several of these mechanisms are still being described in the literature, so the absence of a mechanism from a table is not evidence that it cannot happen.
  • A part that cracked with no overload and no visible deformation is the population these mechanisms come from. The reading of the fracture itself is covered on the damage mode identification chart.
The temperature windows in one view The same triggers read by temperature, which is how they are met in a furnace schedule or a service log. A tempering temperature or a working temperature that sits inside one of these bands deserves a second look at the time the part spends there.
WindowWhat happens in itWhere it is met
400 to 700 °F (200 to 370 °C)Tempered martensite embrittlement in heat treated alloy steel, and blue brittleness in carbon and alloy steels. Both work on the impact energy rather than on the tensile properties.Any part tempered inside this band, and any part that works at this temperature. A Charpy V-notch series against tempering temperature is the usual way to find it.
660 to 1060 °F (350 to 570 °C)Temper embrittlement, fastest around 850 to 900 °F (455 to 480 °C). It is driven by phosphorus, antimony, tin and arsenic segregating to the grain boundaries.Alloy steels that combine chromium with nickel or manganese, and thick sections that cool slowly through the band after a high temperature temper.
750 to 930 °F (400 to 500 °C)The high chromium stainless band, worst near 885 °F (475 °C). Chromium rich phases form and corrosion resistance drops beside the grain boundaries.Ferritic, martensitic and duplex stainless parts held in the band, including stress relief and tempering cycles.
800 to 1025 °F (425 to 550 °C)Graphitization in carbon and low alloy steel, and creep embrittlement in ferritic steels over 800 to 1100 °F (425 to 595 °C). Both depend on time as much as on temperature.Long service at moderate temperature, with the heat affected zones of welds as one of the likely locations.
1050 to 1800 °F (565 to 980 °C)Sigma phase in stainless steels, forming fastest around 1550 °F (845 °C). The phase is hard and brittle and it removes toughness once the part cools.Stainless steel that is solution annealed, stress relieved or welded in this band, and any high temperature service inside it.

Source: ASM Handbook, Vol. 11, the thermally induced and environmentally induced embrittlement sections of the article Overload Failures.

  • The window and the time matter together. Temper embrittlement is diffusion driven and depends on the impurity content and the time spent in the band as well as on the temperature, while blue brittleness can appear in minutes.
  • The metallurgical reactions that fire in each window during a heat treatment cycle, and the dimensional change each one carries, are tabulated separately on the size change in heat treatment page.
Temper embrittlement, the mechanism that reverses Temper embrittlement is the one mechanism on this page that can be removed from a finished part, and its chemistry is specific enough to read from a heat analysis.
ItemDetail
Elements responsiblePhosphorus, antimony, tin and arsenic segregate to the grain boundaries. Silicon, germanium, selenium, tellurium and bismuth are also identified as embrittling agents, and manganese both causes and facilitates it.
Steels most at riskNickel chromium steels are particularly prone. Plain carbon steels are not susceptible, and plain carbon steels below 0.3 percent manganese are not considered susceptible at all. Molybdenum additions reduce the susceptibility.
Where it is fastestAbout 850 to 900 °F (455 to 480 °C) inside the 660 to 1060 °F (350 to 570 °C) band. Heat treated martensitic and bainitic structures are more susceptible than ferritic and pearlitic ones.
How it is detectedA rise in the ductile to brittle transition temperature with strength and hardness unchanged, so toughness testing finds it and a tensile test does not. Electron fractography adds the area fraction of intergranular facets.
Susceptibility indicatorsWelding research produced two simple indicators. The X factor combines phosphorus, antimony, tin and arsenic with a weighting on phosphorus, and the Watanabe number J is (Si + Mn) x (P + Sn) x 10,000, with a value below 200 taken to mean the metal is not susceptible.
How it is removedRetemper above the critical band and cool rapidly. Thick sections that embrittle by slow cooling through the band can be quenched or cooled fast instead, or given an extra temper above the band followed by a rapid cool.

Source: ASM Handbook, Vol. 11, the temper embrittlement section of the article Overload Failures. The X factor and the Watanabe number come from the welding research cited there.

  • Tempered martensite embrittlement sits in a different window and behaves differently. It is produced by tempering inside 400 to 700 °F (200 to 370 °C), it is worst near 600 °F (315 °C), and it is the transition from epsilon carbide to cementite at the prior austenite grain boundaries that drives it. Tensile properties do not move, so a Charpy V-notch series against tempering temperature is the way to see it. The source calls this one-step embrittlement, against two-step for temper embrittlement.
  • A tool steel heat treatment schedule that lands inside these bands is common, because hardness and toughness trade against each other along the tempering curve. The mistakes that band produces on one grade are collected on the D2 heat treatment mistakes page, and the symptom led route through a furnace problem is on the heat treatment troubleshooting chart.
Hydrogen embrittlement Hydrogen is the mechanism that most often leaves no mark on the certificate, because a sound heat of the right grade can still carry enough hydrogen to crack after processing.
ItemDetail
The three requirementsTensile stress, whether residual or applied, a susceptible steel, and hydrogen above a threshold level. A few parts per million of embrittling hydrogen is enough.
Where the hydrogen comes fromSteelmaking, acidic cleaning and pickling, electroplating, welding, and service reactions from gas exposure, lubricant breakdown or corrosion. Cathodic poisons increase the quantity absorbed.
What fails and whenCracking or fracture in a delayed manner under static load, frequently below the yield strength with no plastic deformation. Cracks branch little and are often intergranular, though highly stressed parts can show transgranular fracture.
SusceptibilityRising hardness and rising cold work both increase susceptibility. Steels below 100 ksi (690 MPa) tensile strength are often considered resistant to cracking, and remain susceptible to blistering. Austenitic grades are far less susceptible than martensitic or cold worked ferritic grades.
What helpsControl the processes that generate hydrogen, bake after electroplating, and stress relieve before plating. Baking does not remove all of the hydrogen, and the source records a case where properly baked fasteners still cracked, with the failure at the head to shank fillet of a screw at 180 ksi (1,220 MPa) minimum tensile strength.
The steelmaking sideHydrogen flaking is a potential problem for carbon tool steels and for some medium carbon low alloy steels, and vacuum degassing is the usual way to bring the hydrogen down to a safe level. In one reported case an O1 die cracked after heat treatment with flakes confined to the center of the section, the shallow ones carrying temper color and the deeper ones not.

Source: ASM Handbook, Vol. 11, the hydrogen damage and embrittlement section of the article Overload Failures, with the flaking case from the steel quality section of Failures of Tools and Dies.

  • Vacuum degassing and clean steelmaking practice belong on the purchase order for grades where hydrogen matters, rather than in the tool room. The reading of a heat analysis and the checks that come with a delivered lot are covered further down this page.
Contact with a low melting point metal Liquid and solid metal induced embrittlement both need three things at once, a susceptible steel, wetting of the surface by the embrittling metal, and tensile stress. Commercial service rarely supplies a pure embrittler, and solder, tin and lead base bearing alloys and other alloys all embrittle.
Structural steelMetals shown to embrittle it
Steel, ferritic or martensiticMercury, gallium, indium, lithium, tin, cadmium, lead, zinc, tellurium, copper
Steel, austeniticMercury, lithium, zinc, copper

Source: ASM Handbook, Vol. 11, metals that have been shown to cause liquid metal induced embrittlement, solid metal induced embrittlement, or both, of commercial alloys. The table in the source also covers aluminium, copper, nickel and titanium base metals.

  • The failure is fast where the metal is molten, with crack growth plateau velocities up to about 0.1 m per second measured in fracture mechanics tests, and small components fail almost instantly once the stress is high enough. The path is usually intergranular, and in a liquid metal case the cracks and their branches are filled to the tip with resolidified metal, which is what energy dispersive spectroscopy is used to confirm.
  • The mechanism is a contact phenomenon rather than an alloy property, so the answer sits in the tooling arrangement. Where a die works against molten zinc or aluminium, or where a fixture, a bearing alloy or a coating brings one of these metals into contact under load, the pair belongs in the risk review.

Finding the mechanism before choosing a fix

An embrittled part can be sound by every routine check and still fail below its yield strength. The stage by stage causes, the examination sequence and the case evidence are collected on the tool and die failure analysis page.

A crack that opened with no load applied

Delayed cracking after quenching is a retained austenite and stress problem rather than an embrittlement problem, and it has its own signature on the crack wall. The mechanisms and the prevention steps are set out on the quench cracking page.

Residual elements and what to ask the mill

Temper embrittlement depends on phosphorus, antimony, tin and arsenic, and these are the residuals a heat analysis reports at the bottom of the range. Certificate checks and the test report are covered on the quality verification page.

Tempering ranges by grade

Every grade has a tempering range that lands the working hardness where the job needs it, and some of those ranges sit inside an embrittlement band. The ranges and the resulting hardness are tabulated on the tempering chart.

Confirm before quoting

Composition data is for general reference only. Actual values vary by standard, mill and heat number, and the susceptibility of a heat depends on residuals that are not part of a normal specification. Confirm against the material test certificate or contact Aobo Steel.

The windows and thresholds on this page describe general steel behaviour. A conclusion on a specific part rests on its heat treatment record, its service history and a metallurgical examination.

Source: ASM Handbook, Volume 11, Failure Analysis and Prevention (ASM International), the embrittlement sections of the article Overload Failures, the article Liquid Metal and Solid Metal Induced Embrittlement by William R. Warke, and the steel quality section of Failures of Tools and Dies. Reference data for comparison only. Confirm the mechanism and the process change with your own metallurgist before altering a specification. Aobo Steel supplies tool steel in the annealed condition.