Hydrogen embrittlement is the failure mode that catches the strongest carbon steels, not the weakest: a bolt that passed every tensile test can crack in the warehouse 48 hours after pickling, a galvanized spring can snap during installation with no overload, and a welded pressure part can fracture days after hydrotest. The mechanism is invisible, the delay between cause and failure makes it look like a manufacturing defect rather than a process defect, and the cost lands on whoever cannot prove where the hydrogen came from. For buyers of medium and high strength carbon steel — fasteners above 1000 MPa, springs, cold drawn bar, Q&T parts above about 32 HRC — understanding hydrogen embrittlement carbon steel failure is not academic: it decides whether a plating order, a pickling line or a welding procedure is safe to approve. This guide explains where the hydrogen comes from, why high-hardness steel is the victim, how delayed cracking actually happens, which tests prove susceptibility, and — the part that matters commercially — the baking and process controls that prevent it, with the specific standards to cite on your purchase orders.
What Hydrogen Embrittlement Actually Is
Hydrogen embrittlement (HE) is the loss of ductility and fracture resistance that occurs when atomic hydrogen enters the steel lattice and accumulates at high-stress locations — notches, inclusions, grain boundaries, and above all the plastic zone at the tip of a crack or at a stress concentration under sustained load. Atomic hydrogen is small enough to diffuse through the iron lattice at room temperature, which is what makes the problem unique: unlike most embrittlement mechanisms, HE can develop after the part has left the factory, while the steel sits in service at ambient temperature. The hydrogen does not create a crack by itself; it assists crack growth in steel that is already under tensile stress, allowing a subcritical crack to propagate slowly until the remaining section can no longer carry the load — at which point the part fails suddenly, usually with a flat, bright, “brittle-looking” fracture face that looks out of place on a ductile steel.
The dependency on strength level is the single most important fact in the whole subject. Low hardness ferritic-pearlitic carbon steels — your Q235, A36, hot rolled 1018 — are essentially immune at normal service conditions, because atomic hydrogen is trapped harmlessly in microstructural sinks and recombines to molecular hydrogen without assisting crack growth. As hardness and tensile strength rise, susceptibility climbs steeply: general industry guidance treats steels below about 22 HRC (roughly 800 MPa tensile) as having negligible risk under most conditions, and treats everything above about 32–36 HRC (1100–1200 MPa) as requiring explicit hydrogen management. That is why the fastener and spring industries — whose high-strength parts pass through pickling and electroplating — are where HE claims concentrate.
| Hardness / strength band | Typical products | HE susceptibility | Required controls |
|---|---|---|---|
| ≤ 22 HRC (< ~800 MPa) | Mild steel bolts Gr 4.6–5.8, hot rolled bar, structural steel | Negligible in normal service | None specific to HE |
| 24–32 HRC (800–1100 MPa) | Gr 8.8 fasteners, Q&T 1045/C45, Stud bolts B7 | Low to moderate; risk with heavy plating and sustained load | Bake after plating per ISO 4042 / ASTM B850 |
| 32–40 HRC (1100–1400 MPa) | Gr 10.9 fasteners, springs, high strength studs | Significant; delayed cracking documented | Strict bake, process qualification, sometimes mechanical plating instead |
| ≥ 40 HRC (> 1400 MPa) | Gr 12.9 fasteners, valve springs, drill rod tools | High; many specs prohibit electroplating entirely | Mechanical galvanizing, hydrogen-free processes, or material substitution |
Where the Hydrogen Comes From: Six Industrial Sources
Hydrogen enters steel whenever the metal surface contacts a process that generates atomic hydrogen at the surface. The common industrial sources, in rough order of how often they cause real claims:
- Acid pickling. Hydrochloric or sulphuric acid dissolves scale and simultaneously generates atomic hydrogen at the steel surface. Inhibitors reduce but never eliminate the entry rate. Pickled wire rod and bar destined for cold drawing or plating is a classic carrier, and the risk scales with pickling time — heavily scaled or over-pickled material absorbs the most.
- Electroplating. Zinc, cadmium and nickel electroplating all produce hydrogen at the cathode — which is the part itself. The higher the plating current efficiency problems, the more hydrogen enters. This is why plated fastener specs (ISO 4042, ASTM F1941) carry mandatory bake requirements.
- Cathodic protection and stray current. Buried pipelines and offshore structures protected by impressed current or sacrificial anodes generate atomic hydrogen at the steel surface in service. Combined with sour environments this drives the wet H2S cracking family below.
- Wet H2S (sour service). H2S corrosion generates hydrogen and simultaneously poisons the recombination reaction, maximizing hydrogen entry. Carbon steel in sour service suffers the specific cracking types — hydrogen induced cracking (HIC), stress oriented hydrogen induced cracking (SOHIC) and sulfide stress cracking (SSC) — controlled by NACE MR0175 / ISO 15156 material selection.
- Welding. Moisture in coatings, damp flux and humidity on the joint introduce hydrogen during welding, causing hydrogen assisted cracking in the HAZ. This is the same mechanism with a different entry route; low hydrogen electrodes (E7018 stored in heated ovens), preheat and interpass control are the countermeasures.
- Corrosion in service. Even ordinary rusting releases some hydrogen at cathodic sites on high strength steel. Galvanic couples and crevice corrosion on high strength parts generate enough hydrogen to propagate cracks under sustained load — the reason high strength bolted joints in corrosive service require careful hardness limits.
The Delayed Cracking Mechanism, Explained Simply
The reason HE failures are delayed is diffusion. Atomic hydrogen absorbed during pickling or plating is initially distributed near the surface. Under a tensile stress — bolt preload, residual stress from cold drawing, welding residual stress — the hydrogen migrates toward the highest triaxial stress regions, which are notches, thread roots and crack tips. As hydrogen concentrates at a stress raiser, the local atomic bonding weakens, and at some threshold combination of hydrogen concentration and stress intensity a microcrack initiates or extends a short distance. The crack extension consumes the local hydrogen, so the process pauses while more hydrogen diffuses in from the surrounding lattice, then extends again. The result is crack growth in discrete steps over hours, days or weeks — the classic “delayed failure” — until the section ruptures. The fracture surface shows the diagnostic intergranular or quasi-cleavage facets near the origin, transitioning to normal ductile dimples where fast final overload occurred.
Three conditions must be present simultaneously, and removing any one of them removes the risk: a susceptible microstructure (high hardness, high tensile strength), atomic hydrogen in the steel, and sustained tensile stress (applied or residual). This is why a hard bolt in a drawer fails while an equally hard bolt under a properly controlled preload may not, and why a plating shop can bake the same batch twice with different outcomes — the process variable is the hydrogen inventory, not the steel.
Testing: How Susceptibility and Safety Are Proven
Because HE cannot be seen, the industry relies on standardized tests that either charge hydrogen deliberately or qualify the post-plating bake. The tests below are the ones that appear on purchase orders for high strength carbon steel parts:
| Test / standard | What it does | Where it applies |
|---|---|---|
| ASTM F519 (sustained load) | Notched AISI 4340 specimen under 75–80 % notch strength, 200 h; hydrogen absorbed from the environment must not cause failure | Plating shop process qualification; aerospace plating lines |
| ASTM F1624 (step loading) | Accelerated step-load version giving quantitative threshold for HE | Fastener and spring development, batch screening |
| Slow strain rate test (ASTM G129) | Tensile test at very low strain rate in hydrogen-charging environment; ductility loss vs air test quantifies embrittlement | Material qualification, sour service research |
| ASTM B850 bake effectiveness | Defines post-plate bake schedules by strength level (e.g. 190–220 °C, 4–24 h) | Electroplated parts order requirement |
| ISO 4042 Annex A | Bake schedules for metric fasteners plus pre-treatment rules; includes “parts over 1000 MPa” requirements | Export fastener orders quoting ISO |
| NACE MR0175 / ISO 15156 | Material limits (hardness, chemistry, HIC/SSC test per NACE TM0284 / TM0177) for sour service | Pipes, plates and fittings for oil and gas |
| HIC test, NACE TM0284 | Immersion in sour solution 96 h, sectioning, crack length ratios CSR/CLR/CTR measured metallographically | Line pipe, vessel plate in sour service |
One practical note about receiving inspection: HE failure cannot be reliably diagnosed by hardness or visual check alone. If a failure occurs, the fracture surface is the evidence — keep both mating halves clean, dry and un-damaged, photograph the origin region, and have a metallographic lab examine the initiation zone. The difference between an HE claim and a manufacturing defect claim is usually decided by that fracture face, and parts sent back loose in a box destroy the evidence.
Prevention: The Process Controls That Actually Work
Hydrogen embrittlement is fully preventable with boring, auditable process discipline. The controls that matter, in the order they occur in production:
- Minimize hydrogen entry. Use inhibited pickling acid at correct concentration and temperature; shorten pickling time to the minimum that removes scale; avoid over-pickling. Where possible, replace acid pickling with mechanical descaling (shot blasting) for wire and bar destined for high strength parts.
- Choose the plating process deliberately. For parts above 1200 MPa, specify mechanical galvanizing or zinc flake coating (non-electrolytic) instead of electroplating. Many fastener specifications prohibit electroplating above specific strength levels for exactly this reason; ISO 4042 marks such cases in its applicability tables.
- Bake immediately after plating. The standard schedules: 190–220 °C for 4 h (parts to 1000 MPa), 8 h (1000–1200 MPa), 12–24 h (above 1200 MPa or on suspect parts), starting within 4 hours of plating and always before any chromate conversion damage. Baking lets atomic hydrogen diffuse out and recombine; delay lets it migrate into stress raisers where removal becomes far slower.
- Control preload and geometry. Thread roots, sharp notches and heavy undercut raise local stress; the rolling of threads (versus cutting) improves resistance measurably by putting the root in compression. Proper torque control avoids gross over-preload that pushes a plated bolt past its threshold.
- Manage residual stress. Stress relief after cold drawing or heavy machining (typically 550 °C for 2 h on medium carbon steel) removes the driving force for delayed cracking in storage.
- In welding, use low hydrogen practice. Electrodes baked and stored in portable ovens, preheat per thickness, interpass temperature control, and post-weld dehydrogenation treatment on thick or high constraint joints.
- For sour service, buy to the standard, not to hardness. Specify NACE MR0175 compliance with HIC/SSC test certificates; hardness caps (22 HRC for most carbon steel weldments and components) are part of the standard’s compliance, not an optional extra.
The audit point that separates good suppliers from average ones is traceability of the bake: a plating line that logs oven temperature curves, load times and the interval between plating and baking can prove the process; a line that “bakes when the oven is full” cannot. When we source high strength fasteners or plated parts for export, the bake log is one of the first documents we request, and it is worth more than any certificate statement about “hydrogen embrittlement avoided.”
Standards to Cite on the Purchase Order
| Situation | Standard to reference | Key requirement |
|---|---|---|
| Electroplated metric fasteners | ISO 4042 | Bake schedules by property class; pre-treatment; hydrogen embrittlement test method |
| Electroplated inch fasteners | ASTM F1941 / B850 | Coating spec + bake by tensile strength band |
| Plating process qualification | ASTM F519 | Sustained load 200 h on notched specimens from the production line |
| Sour service components | NACE MR0175 / ISO 15156 | Hardness caps, chemistry limits, HIC/SSC testing per TM0284/TM0177 |
| Bolting for pipelines and flanges | ASTM A320 / A193 with notes | Grade-specific hardness and test requirements |
| Welding of carbon steel | AWS D1.1 / applicable code | Low hydrogen consumables, preheat tables, electrode storage |
Real Failure Patterns We See in Export Orders
Three recurring scenarios are worth knowing, because each has a different commercial lesson. First, the plated 10.9 bolt that cracks in the customer’s warehouse two days after arrival: the root cause is almost always a skipped or shortened bake at an intermediate plating house, and the defense is the bake log plus an F519 qualification on file. Second, the cold drawn 1045 shaft with a hydrogen crack at a keyway after pickling and storage — the lesson is that medium carbon steel in the 28–34 HRC band sits exactly at the susceptible boundary, and parts in that band need the same bake discipline as fasteners if they pass through pickling or plating. Third, the welded thick-wall pipe that cracked at the HAZ during hydrotest 72 hours after welding — this is hydrogen assisted welding cracking, prevented by electrode ovens and preheat, and it fails in the shop rather than in service, which at least keeps the claim in-house. In every case the supplier who can produce process records resolves the claim quickly; the supplier who cannot pays for the failure regardless of where the fault actually lay.
Telling Hydrogen Embrittlement Apart From Look-Alike Failures
When a high strength part breaks, three mechanisms get blamed interchangeably, and the differences decide who pays. True hydrogen embrittlement produces a flat fracture with intergranular or quasi-cleavage facets concentrated at the origin, initiates at a stress raiser such as a thread root or keyway, and shows no corrosion product inside the crack — because the crack grew while the part was dry and loaded. Stress corrosion cracking (SCC) looks similar metallographically but requires a specific corrosive medium (chlorides for austenitic stainless, nitrate or caustic for carbon steel) and usually shows branching cracks and corrosion products along the crack path. Fatigue starts at a visible origin point and grows in beach marks over thousands or millions of load cycles — completely different surface appearance, and rarely confused once the fracture face is examined under a magnifier. The practical sorting rule used by failure labs: origin location plus fracture morphology plus service environment history, in that order. For the buyer, one step matters most — preserve the fracture faces untouched and send both halves to the lab together, because the origin region is where the entire diagnosis lives, and it is the first thing destroyed by rust, handling or a well-meaning attempt to “clean up” the broken parts before photographing them.
FAQ
Does galvanizing cause hydrogen embrittlement? Hot dip galvanizing involves acid pickling before immersion, so the pickling step can introduce hydrogen, and high strength parts (above roughly 1000 MPa) are at risk. Electroplated zinc carries a higher risk than hot dip because the part is the cathode during plating. Both require baking for high strength steel; mechanical galvanizing avoids the problem entirely.
How long after plating can embrittlement failure occur? Most delayed failures happen within 24 hours to a few weeks of the hydrogen entry event, once the part is under sustained load. Failures months later usually indicate continued hydrogen generation from corrosion in service rather than the original plating.
Can baking remove hydrogen completely? It reduces the diffusible hydrogen inventory to a safe level for the strength class; it cannot extract hydrogen that has already concentrated at traps. That is why bake timing — starting within a few hours of plating — matters as much as bake duration.
Is low carbon steel ever affected? Below about 22 HRC, general engineering carbon steel is not considered susceptible under normal conditions. The exceptions are sour service (where HIC affects even low strength steel as internal blister cracking) and cathodically protected structures, which is why pipeline steels have their own test regime.
What hardness is safe for plated bolts? Industry practice treats parts up to 1000 MPa (about 320 HV / 34 HRC) as manageable with standard bakes, and increasingly prohibits electrolytic plating above 1200–1400 MPa. For anything in the upper band, mechanical plating or zinc flake is the safe order.
Buy High Strength Carbon Steel With Hydrogen Risk Controlled
Huaxia-Steel supplies Q&T bar, high strength stud stock, fastener steel and sour-service grades with the process controls that keep hydrogen risk out of your claims file: inhibited pickling or shot-blast descaling on bar stock, bake schedules per ISO 4042 and ASTM B850 coordinated with our plating partners, F519 qualification support for aerospace-grade lines, and NACE MR0175 compliant material with HIC/SSC certificates for oil and gas. Tell us the strength level, the surface treatment and the service environment, and our export team will propose the process route and the certificates that protect you — quotation within one working day.
