Figure "Diffusion processes" (Ref. 5.4.4-3): Hydrogen that has been absorbed by metals can considerably increase the growth rate of cracks (e.g. also corrosion cracks, see Fig. "Role of hydrogen in corrosion processes"). Experts worldwide do not agree in their estimations of the role of hydrogen on the embrittlement process. To the author, one, albeit old, process seems to describe the damage process in steels so plausibly, that it will be explained in the following. It must not be denied that at least the explanation of the pore development phase is not considered to be proven by much technical literature (Ref. .5.4.4-4).
At the crack tip, a typically shaped plastic deformation zone forms (top left diagram), in which a lattice expansion occurs under the influence of the 3-axis tensile stress. Atomic hydrogen diffuses into this area. The hydrogen diffusion occurs (top right diagram, Refs. 5.4.4-3 and 5.4.4-6) in the direction of the greatest tensile stress conditions (largest lattice expansion) present in a “germinal point”. These are flaws in the transition of the plastic zone to the elastic zone. The hydrogen collects at the relatively small “germination point”. In less hard steels with hardnesses below about 30 HRC, the hydrogen can recombine to molecules that can no longer diffuse into the metal. In this way, hydrogen pressure p_n (10⁴ - 10⁸ bar) rapidly builds up. When the gas pressure in the expanded (toward the inner crack) pores around the “germination point” in front of the crack has risen sufficiently, the “explosive effect” causes it to combine with the crack tip and increase crack growth.

However, this damage model is unsatisfactory for explaining the behavior of high-strength and case-hardened steels. For example, in these materials, damage has been observed to decrease again after longer times of conditioning (Ref. 5.4.4-14). The hydrogen is trapped at lattice deformations in reversible or irreversible “Traps”. This model also explains why long times of conditioning diminish the embrittlement effect, since the hydrogen is able to distribute better into the lattice from the reversible traps. However, the reversible traps can also act as hydrogen sources and provide hydrogen to zones with lattice expansions (e.g. in front of a crack tip). This explains the grain boundary cracks typical for hydrogen embrittlement.
The tensile stresses which cause hydrogen embrittlement must not necessarily be caused by external loads. Internal stresses from heat treatments, welds, coatings, and machining processes are frequently sufficient to cause hydrogen to collect in resting parts, such as during storage. Once hydrogen has collected at the “germination points” and created micro-cracks/pores, the damage will be irreversible.
As long as the hydrogen in reversible traps is atomically dissolved in the lattice, the damage is reversible (bottom diagram); suitable heat treatment (expulsion of hydrogen - degasification) can prevent the embrittlement process (Fig. "Disembrittlement").

Figure "Effects of hydrogen on the structure": The crack-promoting effect of hydrogen on metals can be caused by different effects (top diagram) that can occur individually or in combination with one another:

Deformation of the metal lattice: This embrittlement mechanism is dominant in high-strength and case-hardened steels (Ref. 5.4.4-14). The metal lattice is deformed by included hydrogen atoms. This reduces the mobility of dislocations, resulting in a loss of ductility. The hydrogen is caught in reversible or irreversible traps, depending on whether or not diffusible oxygen can be released. These traps are caused by lattice inhomogeneities and are located primarily in the grain boundary zone. This explains the intercrystalline crack development typical for hydrogen embrittlement. The hydrogen diffuses primarily in already-deformed atomic lattices. These include, for example, areas that have been cold-deformed under mechanical tensile loading, as is the case at crack tips (Fig. "Diffusion processes"). But even hardened martensite steels have suitable deformed lattices. These influences are especially effective in high-strength steels.

Recombination and gas pressure buildup: The hydrogen can diffuse out of the lattice into flawed zones (e.g. cast porosity, micro-cracks, deformation pores, inclusions, Ref. 5.4.4-6), around which the lattice is expanded due to the stress concentration, e.g. under internal stresses. In less high-strength steels (with hardnesses below about 30 HRC), pore formation is explained by a recombination to molecular hydrogen. This hydrogen gas can no longer escape through diffusion due to its relatively large molecular diameter. In the present or forming hollow spaces, extremely high pressures (top right diagram, Ref. 5.4.4-4) build up with an “explosive effect” which can cause brittle fractures without any major external loads.
A typical phenomenon is the microscopically small pores on the grain boundaries (with no sign of ductility of the crack honeycombs) of the fractured surface (bottom left diagram), which is REM-detectable evidence that hydrogen influenced the damage mechanism (Ref. 5.4.4-4). Hydrogen embrittlement is usually intercrystalline (bottom right diagram, Ref. 5.4.4-7). However, there are also cases of transcrystalline fractures. Hairlines (so-called crow`s feet) are flat crack honeycombs due to the remaining ductility of the material.

Hydride formation: Hydrogen can be chemically bonded by metals as a brittle hydride. Hydrides in the form of surface coatings or platelets in the structure promote brittle fractures. This embrittlement form is especially well known in Ti alloys (see).

Figure "Absorption scenarios": Atomic hydrogen can develop on metal surfaces and be absorbed by diffusion under various different scenarios:

Metal melts (“1”): At the high temperatures, the steam is split and atomic hydrogen develops at the surface of the melt. Especially large amounts of hydrogen can dissolve in melts, which leads to pronounced crack development in the frozen material.

Welding (“1a”): Contact between water and the melt is caused by

• damp electrodes/electrode coverings resulting from improper storage and failure to bake out a sufficiently short time before welding.
• damp protective gas (e.g. argon). Hydrogen absorption resulting from unsuitable permeable steam feed hoses has been observed.
• part surfaces dampened by a film of condensation water (see Fig. "Weld seams and turbine housings").

Casting (“1b”): In this case, also, hydrogen absorption occurs due to water from a damp environment. The causes are similar to those in welding. For example, as with damp electrode coverings, the use of damp shell moulds can provide a hydrogen supply.

Processing baths (“2”): In galvanic baths (e.g. coating baths for silver, chrome, nickel, cadmium, or zinc) or chemical baths (e.g. etching baths, electroless coating baths for nickel, etc.) with cathodic hydrogen creation at the part surface (cathodic reduction, H₂ → 2 H⁺ + 2 e). These reactions also occur during certain corrosion processes (Fig. "Role of hydrogen in corrosion processes"). These include corrosive processes, especially at crack tips, where there is increased possibility for diffusion of the atomic hydrogen due to lattice expansion (Fig. "Diffusion processes") (elastic deformations, ductile cold strain).

“Contact diffusion” (“3”): Hydrogen diffusion can occur in case of intense metallic contact with a hydrogen-charged surface, such as screwed or shrink-fit part surfaces.

Thermal decomposition of hydrogen compounds (“4”): If a compound containing hydrogen, such as water, is heated to high temperatures, it may thermally decompose and result in the creation of atomic hydrogen. This hydrogen can diffuse into the metal walls of, for example, pipes or boilers.
During case-hardening through carbo-nitriding (Fig. "Gear failure"), the gas decomposes, creating atomic hydrogen (e.g. 3 Fe + CH₄ → Fe₃C + 4 H) that diffuses into the part being carburized.

Diffusion of catalytically dissociated hydrogen (“5”): From the molecular state (e.g. natural gas or hydrogen as engine fuel), dissociation due to a catalytic effect can occur at the surface of a part (e.g. a pipe). The dissociation can be explained by plastic deformations in the micro-zone, in which active fresh metal surfaces are created. These plastic deformations can be expected, for example, under LCF loads.

Figure "Role of hydrogen in corrosion processes" (Ref. 5.4.4-8): Hydrogen embrittlement plays a deciding role in the damage mechanism SCC (Chapter 5.4.2.1). During the anodic partial reaction, the actual erosive/dissolving corrosion process of the metallic material is accompanied by reaction coating development (Fig. "Coating damage I"). This creates positively charged metal ions when electrons separate (oxidation process).
In the case of anodic SCC the material is in a passive state, in which dissolving takes place only at the crack tip. At the surface of the part and crack, the negative charge of the material leads to an anodic partial reaction and atomic hydrogen creation due to the charge equilibrium with the positively charged hydrogen ions of the surrounding electrolyte (reduction process). This hydrogen can diffuse into the part and accelerate crack growth at the crack tip (Fig. "Diffusion processes"). The anodic and cathodic partial reactions area always coupled. In corrosion, the cathodic partial reaction usually determines the speed.
In cathodic SCC the material dissolution at the part surface occurs in zones with active corrosion behavior (anodic zones). The dissolving of metal and creation of positively charged metal ions frees electrons, which move to cathodic (passive re. corrosion) surface zones where charge balancing with the hydrogen ions of the electrolyte creates atomic hydrogen, which diffuses into the metal lattice and promotes SCC.
The creation of passive reaction coatings during the corrosion process can result in the development of protective conversion coatings (passive coatings). This allows the same metal to occur in both the passive and active state, depending on the corrosion conditions.
The influence of hydrogen embrittlement components in SCC can be seen on the fracture surfaces through characteristics such as micro-pore formation, quasi crack surfaces, and “rooster feet” (Fig. "Effects of hydrogen on the structure", Ref. 5.4.4-4). The passive state retards corrosion and is therefore usually desirable, but it is also more prone to deep localized corrosive attack, the dreaded pitting corrosion (Fig. "Tensile stress").
In Ti alloys, etching and signs of hydrogen embrittlement were discovered in connection with the influence of typical aviation-use hydraulic oil above 130°C (Ref. 5.4.4-13). At these temperatures, the hydraulic fluid evidently forms an acid that corrodes titanium and causes hydrogen embrittlement. According to the literature, only a certain beta-titanium alloy is immune to this damage. This alloy is used on the tail cone and thrust jet of the newest cargo aircraft.

Figure "Sensitivity to hydrogen embrittlement": Generally, the greater the strength of a material, the greater its sensitivity to hydrogen embrittlement (top diagram, Ref. 5.4.4-9). With increasing hardness, only a fraction of the amount of hydrogen that is diffused into the material is necessary to have a pronounced effect in reducing the contraction at fracture.
The bottom diagram shows the decrease in strength and life span in notched samples of a high-tempered steel under static loads. The samples were charged with hydrogen and subjected to varying degasification times at a heating temperature of 150°C. The influence of the heating times on the usable strength can be clearly seen.

Figure "Verification of hydrogen embrittlement": The effect of hydrogen on crack development in titanium alloys depends on the creation of brittle hydrides (see). If relatively slow diffusion processes occur in metal (e.g. steels), embrittlement (expansion of the lattice) is delayed (Fig. "Diffusion processes"). Therefore, hydrogen embrittlement is not usually (except for embrittlement of a-titanium through hydride formation) verifiable through notch impact tests or tensile tests (usually at high speeds). The high loading speed does not allow sufficient time for embrittling diffusion processes.
A decrease in crack strain and contraction at fracture can be expected from strain-controlled tensile tests with <0.05 mm/mm.min.
This damage in a creep test can be recognized by the decrease in strength and operating life.
The tensile notch test does not give any clear results.
On the other hand, influence on the crack development (critical loads for crack initiation and crack growth rates) can be expected from fracture mechanical CT samples with typical high stress concentrations (Fig. "Tensile stress").

Figure "Disembrittlement" (Ref. 5.4.4-5): As long as no recombination of atomic hydrogen to molecules has occurred (Fig. "Diffusion processes"), several hours of heating in air at a few hundred °C can expel the hydrogen or distribute it in the lattice in such a way that it sufficiently prevents the crack development effect. It is also possible to “disembrittle” Ti alloys that have become embrittled through hydride formation (see) by pyrolizing the hydrides at high temperatures (about 200° - 300°C). Because of the danger of recombination and/or bonding in irreversible “traps”, disembrittlement (see Fig. "Effects of hydrogen on the structure") must be done immediately (within a few hours) after hydrogen absorption in order to be successful.

Experience has shown that this is where the most mistakes are made, when the parts are allowed to sit for too long before disembrittlement or degasification is forgotten entirely.

The diagrams show that with increasing degasification temperatures and longer hold times, considerable reduction of the original hydrogen content occurs. The remaining amount of dissolved hydrogen depends largely on the amount of dissolved hydrogen before the disembrittlement process. There is no great difference between the degasification behaviors of chromed and cadmium-plated parts. However, it is obvious that cadmium-plated steels are less able to expel hydrogen than are hard-chromed parts.The difference may lie in the gas permeability of the coatings with a different sealing effect.

Figure "Crack growth in strong materials": Typical parts in which hydrogen embrittlement damage is frequently reported are hardened or high-tempered parts made from low- and high-alloy steels with galvanic coatings, especially cadmium coatings (Ref. 5.4.4-15). The supply of atomic hydrogen seems to be especially intense during this process. Typical parts include disk springs, thread inserts, spring washers, plain washers, and threaded rods (top diagrams).
The lower diagram shows a hydraulic piston from a thrust jet adjuster made from a high-strength heat-treated steel. After a restorative chrome-plating procedure, certain areas (arrows) that were not chrome-plated developed cracks with typical characteristics of hydrogen embrittlement (Fig. "Effects of hydrogen on the structure").

Figure "Fracture of compressor disk": This integrally cast compressor disk from a helicopter turbine made precipitation-hardened high-strength steel (type 17 4 PH) was in an unfavorable heat-treatment state with too great a hardness. Hydrogen absorption probably occurred in an etching bath during new part manufacture. The disk failure occurred spontaneously during operation, and the disk broke into many pieces (compare with Fig. "Material behavior"). The crack development was transcrystalline with typical signs of hydrogen embrittlement such as micro-pores on material inhomogeneities.

Figure "Gear failure": Teeth fractures on case-hardened gears due to inner crack development following hydrogen embrittlement. The gears evidently absorb hydrogen during case-hardening (see Fig. "Absorption scenarios"). The damage then seems to occur as follows (Ref. 5.4.4-11): the high internal stresses in the case-hardened layer at the tooth surface induce high tension residual stresses in the tooth core (top left diagram). The smaller and more filigreed the tooth, the greater the danger that a circular hydrogen embrittlement crack will develop at internal notches within several hours. An intercrystalline hydrogen-induced brittle fracture then originates in this crack, and can later lead to a dynamic fatigue fracture under the operating loads, destroying the gearing.

Figure "Weld seams and turbine housings": Older engine types often have compressor housings that are welded constructions made from low-alloy heat-treated steels. Similar housings are found in modern engines near the outlet of the low-pressure turbine. The complex shape of these housings causes them to develop high tension residual stresses that are induced by welding during new part manufacture or repair procedures. If moisture (e.g. in the shielding gas or as condensation water) is present during the welding process (see Fig. "Absorption scenarios"), then hydrogen from the moisture can be absorbed by the melt (Ref. 5.4.4-12). This can lead to delayed crack initiation in the weld seam area during storage, manufacturing processes, or operation (arrows show typical crack-prone part zones). For this reason, if this type of crack is suspected, REM inspections of the fracture surface should be conducted first of all, in order to confirm characteristic signs of hydrogen embrittlement (see Fig. "Effects of hydrogen on the structure").

Figure "Surface contaminants": It has been discovered that, under certain circumstances, hydride formation can cause massive hydrogen embrittlement in high-strength titanium alloys even in oxidizing etching baths (such as HNO₃). Important influences on this include:

• Deposits of iron oxide. Ni-deposits may also play a part.

• Structural changes due to overheating, such as can occur at the edges of compressor rotor blades during a serious stall (heating-up due to agitation losses).

• Roughening of the surface due to erosion.

In light of this, there is a question as to dangerous hydrogen embrittlement on titanium part surfaces in cleaning baths. Steel shot peening or machining with steel tools (e.g. polishing brush) can leave iron wear products/iron oxide on the part surface.
However, no such cases have been reported yet.