Hydrogen-induced cracking doesn't announce itself during welding. The joint looks sound when the welder walks away; the crack initiates in the HAZ during cooling and may not be detectable until hours or days later. By then the structure is often partially assembled, making repair expensive and disruptive.
The good news is that hydrogen-induced cracking is the most predictable and preventable serious weld failure mode in structural and pressure vessel fabrication. Every contributing factor is visible in advance, and every contributing factor is controllable through WPS design and shop practice. The shops that have eliminated HIC from their nonconformance records didn't do it by accident — they baked the prevention into the procedure.
The three-factor model
HIC requires all three of the following conditions simultaneously:
- Diffusible hydrogen — hydrogen atoms that have entered the weld or HAZ metal and can move through the lattice under stress
- Susceptible microstructure — a hard, high-carbon-equivalent martensitic or bainitic HAZ that provides the crack initiation sites
- Residual stress — tensile stress from the welding thermal cycle, joint restraint, or both
Eliminate any one factor and cracking cannot occur. In practice, WPS controls target all three, because no single measure is 100% reliable on its own. Controlling hydrogen source, preheat, and heat input simultaneously produces defense-in-depth against cracking.
Where hydrogen comes from
The primary sources of diffusible hydrogen in a weld are:
Moisture in consumables. Covered electrode coatings, flux (SAW), and cored wire flux fill absorb atmospheric moisture. Low-hydrogen electrodes like E7018 are manufactured to low moisture content but must be stored in dry conditions — electrode ovens at 250–300°F (120–150°C) for opened cans. Electrodes left in open air even for a few hours can absorb enough moisture to elevate diffusible hydrogen well above the levels measured at the factory. Rebaking (typically 700–800°F/370–425°C for one to two hours per AWS A5.1 requirements) restores moisture content after inadvertent exposure.
Contamination on the base metal or joint. Mill scale, rust, paint, cutting oil, moisture from rain or condensation, and hydrogen-bearing coatings all contribute to diffusible hydrogen in the weld deposit. Joint cleanliness immediately before welding is a process variable, not just a housekeeping note. A WPS for high-strength steel should explicitly require surface preparation and dryness of the joint.
Shielding gas quality. For GMAW and FCAW-G, shielding gas moisture (dew point) can contribute if cylinder storage or delivery system conditions allow condensation. This is less common but relevant on outdoor field operations in humid conditions.
Low-hydrogen consumable selection
The most direct hydrogen control is the filler metal selection on the WPS. AWS A5.1 (covered electrodes for SMAW) designates low-hydrogen electrodes with an "8" in the usability position of the classification: E6018, E7018, E8018, E9018. The "18" specifically indicates a low-hydrogen iron-powder coating.
Cellulosic electrodes (E6010, E6011) and rutile coatings (E6013) are not low-hydrogen. They are appropriate for specific applications — E6010 on pipeline root passes, for example — but should not appear in a WPS designed to prevent HIC on restrained joints or high-strength material.
For FCAW, the low-hydrogen advantage is built into metal-cored and basic-flux cored wires but not into all flux compositions. The WPS should specify the AWS A5.20 or A5.36 electrode classification, not just a trade name, so the low-hydrogen characteristic is document-traceable.
For GMAW with solid wire, moisture in the shielding gas is the primary hydrogen vector — the wire itself contributes minimal hydrogen when clean and dry. Solid wire GMAW on properly cleaned joints with dry shielding gas produces very low diffusible hydrogen levels, which is one reason GMAW is preferred on high-strength applications where weld metal hydrogen is a concern.
For a detailed look at SMAW low-hydrogen electrode WPS documentation, see SMAW WPS with E7018 low-hydrogen electrode.
Preheat: what it does and what it doesn't
Preheat raises the base metal temperature before welding begins. The mechanism is cooling rate control: a hot base metal cools more slowly after the arc moves on, which does two things:
- Reduces the tendency to form hard martensite and bainite in the HAZ (because slower cooling allows more favorable transformation products to form)
- Extends the time for hydrogen to diffuse out of the joint before the temperature drops below the level where diffusion becomes negligible (roughly 200°F / 95°C)
AWS D1.1:2025 specifies minimum preheat temperatures based on the combination of process (low-hydrogen vs. non-low-hydrogen), base metal type, and plate thickness. These minimums are the floor, not the target — high-restraint joints, thick sections, and high-carbon-equivalent materials benefit from preheat above the minimum.
What preheat does not do: it does not reduce the amount of hydrogen entering the weld in the first place. A joint welded at proper preheat with a moisture-saturated electrode still introduces elevated hydrogen — the preheat just gives it more time to leave. This is why preheat and low-hydrogen filler are both required on high-risk joints.
For preheat documentation on the WPS form, see preheat and interpass temperature on a WPS.
Rule library based on AWS D1.1:2025; verify against your governing edition.
Carbon equivalent and susceptibility
Carbon equivalent (CE) is a single-number measure of a steel's hardenability — how likely it is to form a crack-susceptible martensitic HAZ. The IIW formula, widely used for carbon and low-alloy steels, is:
CE = C + Mn/6 + (Cr + Mo + V)/5 + (Ni + Cu)/15
(All values in weight percent from the mill certificate.)
As a rough guide:
- CE ≤ 0.40 — generally weldable without preheat on thin to moderate sections under low restraint; low HIC risk with low-hydrogen filler
- CE 0.40–0.60 — preheat required; the required temperature increases with thickness and restraint
- CE > 0.60 — high preheat (often 300°F / 150°C and above); possibly post-weld hydrogen release bake; may require controlled heat input range to limit HAZ width
Common structural steels have CE values in the 0.35–0.55 range. High-strength low-alloy steels used in heavy structural applications, pressure vessels, and crane structures can reach 0.55–0.65 CE. The mill certificate for each heat should be reviewed before finalizing preheat requirements on the WPS.
A WPS that specifies preheat for a generic base metal group without accounting for CE variation within that group is leaving risk on the table. When your material procurement sources vary, build preheat requirements to the highest CE material likely to be used — or require CE verification before reducing preheat below the conservative value.
Interpass temperature and heat input
Interpass temperature — the minimum and maximum temperature of the weld immediately before depositing the next pass — is the companion to preheat for multi-pass welds.
The minimum interpass temperature is typically the same as the preheat requirement: don't let the joint cool below preheat between passes. Letting the joint cool and then reheating introduces additional thermal cycling that can promote cracking and cause unnecessary hardening.
The maximum interpass temperature matters differently: excessive interpass temperature on high-strength steels can reduce yield strength and toughness in the HAZ. On steels with yield strength above 70 ksi (485 MPa) or applications with CVN toughness requirements, maximum interpass limits are an essential variable and must be documented and enforced.
See CVN impact testing and Table 6.8 supplementary essentials for when CVN toughness requirements make interpass maximum an essential variable trigger.
Post-weld hydrogen release bake
On thick sections, highly restrained joints, or base metal with CE above 0.50, a post-weld hydrogen release bake adds a final layer of protection. The mechanism: immediately after welding (before the joint cools below preheat), the joint is maintained at 400–450°F (205–230°C) for a period that allows remaining diffusible hydrogen to continue diffusing out before residual stress locks it in place.
The bake is typically specified as a minimum temperature and minimum hold time — for example, 400°F minimum for one hour per inch of weld throat. The joint must not be allowed to drop below preheat between completion of welding and the start of the bake.
If the post-weld bake is required on a given joint, it belongs on the WPS, not in a separate memo. An NDE requirement that defers inspection until the bake is complete should also be documented, so inspection does not occur on a joint that still contains elevated hydrogen.
Documenting hydrogen controls on the WPS form
Every hydrogen control measure should appear on the WPS document itself — not only in a shop procedure, SOP, or verbal instruction. An inspector reviewing the WPS should be able to read off:
- Filler metal classification (including low-hydrogen designation)
- Electrode storage and conditioning requirements (if not covered by a referenced shop procedure)
- Minimum preheat, minimum interpass, maximum interpass temperatures
- Post-weld hydrogen release bake parameters, if required
Controls that live only in separate documents or in supervisors' heads don't survive turnover, subcontracting, or audits. The WPS is the production instruction — if hydrogen prevention matters for a given joint, it belongs on that form.
For the complete picture of what a CWI reviews when signing a WPS, see CWI WPS review: a step-by-step checklist. For how these controls connect to the broader audit documentation picture, see common WPS deficiencies found in third-party audits.