Hydrogen Embrittlement High-Strength Bolt Failure Case Study

This case study examines delayed fracture of high-strength plated bolts after installation. The bolts passed dimensional inspection, hardness checks, and initial torque acceptance. Several failed hours later while the joint was under sustained preload. The fracture surfaces and process records pointed to hydrogen embrittlement introduced during surface preparation and plating.

The case is useful because a fastener can pass a static strength check and still be unsafe when high strength, tensile stress, a sharp thread root, surface processing, hydrogen uptake, and delayed cracking act together. The engineering decision is not only whether the broken bolts are strong enough. It is whether the entire lot can be trusted after the process history is understood.

Case Summary

Item	Engineering relevance
Component	M12 high-strength steel bolts used in a clamped structural bracket.
Process	Acid cleaning followed by electroplated corrosion-protection coating.
Symptom	delayed fracture at the first engaged thread after installation.
Static result	preload stress below proof strength, so simple static design appeared acceptable.
Root cause	hydrogen-assisted cracking from plating process exposure, delayed baking, high hardness, and sustained tensile stress.
Corrective action	quarantine the lot, change the coating route or bake control, validate by sustained-load testing, and update purchasing and release requirements.

The numbers below are simplified but realistic. They are used to show how an engineer connects fastener stress, crack-tip severity, process evidence, and lot-release decisions.

Field Data

Quantity	Value
nominal bolt size	M12
tensile stress area	A_t=84.3\ \text{mm}^2
installation preload	F_p=55\ \text{kN}
specified proof stress	\sigma_p=970\ \text{MPa}
specified ultimate tensile strength	\sigma_{UTS}=1220\ \text{MPa}
measured hardness range	43 to 45\ \text{HRC}
fracture location	first engaged thread
surface flaw depth from sectioning	a=0.20\ \text{mm}
geometry factor for screening	Y=1.10
hydrogen-assisted threshold screen	K_{TH,H}=15\ \text{MPa}\sqrt{\text{m}}
inert fracture toughness screen	K_{IC}=65\ \text{MPa}\sqrt{\text{m}}
required post-plating bake start	within 1\ \text{h} by internal specification
actual bake start for failed lot	10 to 14\ \text{h} after plating

The internal specification used here is a project control requirement. Actual fastener acceptance must follow the applicable product standard, coating specification, design authority, and safety classification.

Failure Mechanism

Hydrogen embrittlement in high-strength steel fasteners usually requires three conditions:

1. hydrogen enters the steel during manufacturing, cleaning, corrosion, cathodic protection, pickling, plating, or service exposure;
2. the material is susceptible, often because strength and hardness are high;
3. tensile stress is present, commonly from preload, residual stress, or service load.

The fracture can be delayed because hydrogen diffusion, trapping, crack initiation, and crack growth continue after installation. A bolt may tighten normally and fail later without any overload event.

Step 1: Check Nominal Preload Stress

Nominal tensile stress from preload is:

Use:

Because:

the nominal preload stress is:

Compare with proof stress:

The preload is about:

of proof stress.

Engineering Comment

The static screen passes. The bolt was not simply overloaded in nominal tension. That is exactly why hydrogen embrittlement is dangerous: the local crack-tip condition can be critical even when the gross-section stress looks acceptable.

Step 2: Screen Thread-Root Crack Severity

Use a simple mode-I stress-intensity screen:

where:

• Y is a geometry factor;
• \sigma is nominal tensile stress;
• a is flaw depth in meters.

Use:

Then:

Compare with the hydrogen-assisted threshold screen:

Since:

the thread-root flaw is credible for delayed hydrogen-assisted crack growth under sustained preload.

Engineering Comment

The calculation is a screening model, not a final fracture-mechanics assessment. Thread geometry, residual stress, plating defects, real crack shape, preload scatter, material heat treatment, and hydrogen distribution all matter. The screen is still valuable because it explains why a small thread-root flaw can become dangerous in a susceptible high-strength bolt.

Step 3: Check Why Immediate Brittle Fracture Did Not Occur

Use the inert fracture toughness screen to estimate a critical crack size without hydrogen-assisted degradation:

Solve for critical crack depth:

Use:

Then:

The observed flaw was:

which is much smaller than:

Engineering Comment

This explains the delayed nature of the failure. In a benign environment, the small flaw would not necessarily cause immediate unstable fracture at the installed preload. In the presence of hydrogen, the effective cracking threshold can be much lower, allowing crack growth until final fracture occurs.

Step 4: Review Process Evidence

The investigation compared the failed lot with an earlier accepted lot.

Evidence	Failed lot	Earlier accepted lot	Interpretation
coating route	acid clean and electroplate	mechanical coating route	failed lot had higher hydrogen uptake risk
bake start time	10 to 14\ \text{h} after plating	within internal limit	delayed bake reduced hydrogen-removal confidence
fracture location	first engaged thread	no failures	highest tensile and notch severity location
fracture appearance	intergranular and brittle regions	not applicable	consistent with hydrogen-assisted cracking
installation preload	within specification	within specification	preload alone did not explain difference
hardness	high-strength range	similar nominal strength	susceptibility controlled by process plus stress

Engineering Comment

The most persuasive evidence is not one observation. It is the convergence of mechanism, location, process route, delayed bake, sustained preload, and fracture morphology. A single broken bolt could be mishandling. A pattern tied to one process lot is a release-control problem.

Step 5: Decide Lot Disposition

The failed lot cannot be released by replacing only the broken bolts. The same process history applies to unbroken bolts from that lot.

The engineering disposition was:

1. quarantine all bolts from the affected plating batch;
2. remove installed bolts from critical joints;
3. preserve failed bolts for fracture analysis;
4. review plating, cleaning, bake timing, hardness and material certificates;
5. check whether the design truly requires the selected strength class;
6. select a lower-hydrogen-risk coating route or enforce immediate bake control;
7. validate replacement bolts with sustained-load testing and lot traceability.

Engineering Comment

Hydrogen embrittlement is a lot and process problem, not only a part problem. If unbroken bolts have the same material, coating, hydrogen exposure, and preload, they cannot be assumed safe because they have not failed yet.

Step 6: Define Replacement Controls

The replacement strategy used a controlled fastener specification:

Control	Engineering purpose
coating route review	avoid unnecessary hydrogen-generating process steps
bake start and duration record	prove hydrogen relief followed the specified window
hardness limit or strength-class review	reduce material susceptibility when design allows
thread-root quality inspection	reduce flaw severity at the highest-stress location
sustained-load lot test	expose delayed cracking before release
torque-preload method review	reduce preload scatter and over-tightening
lot traceability	prevent mixed fasteners from hiding the affected process route
fracture-analysis trigger	preserve evidence after any future delayed failure

Engineering Comment

The strongest correction may be changing the coating route or strength class, not only adding a bake line to a weak process. If corrosion protection, strength, and embrittlement risk conflict, the fastener specification must be treated as a design decision.

Step 7: Sustained-Load Validation

A validation screen was defined for replacement lots.

Sample bolts were loaded to:

First calculate proof load:

Then:

The sustained-load test therefore holds sample bolts above the installation preload:

Acceptance required no fracture, no visible cracking, no abnormal preload loss, and no process-record deviation during the hold period.

Engineering Comment

A sustained-load test does not prove every future environment or service condition. It is a release screen that targets the observed delayed failure mode. It must be combined with process control, traceability, and design review.

Corrective Actions

The accepted corrective actions were:

1. quarantine and replace the affected plated fastener lot;
2. require coating-process approval for high-strength fasteners;
3. define maximum time from plating to hydrogen-relief bake;
4. record actual bake temperature, duration, load time, and part count;
5. review whether a lower strength class or different coating can meet design requirements;
6. add sustained-load lot testing for susceptible fasteners;
7. require fracture analysis after any delayed fastener breakage;
8. update purchasing documents so coating substitutions cannot occur without engineering approval.

Final Decision

The defensible engineering decision was:

Do not release the affected high-strength plated bolt lot. Replace installed critical fasteners and accept future lots only when coating route, bake timing, hardness, sustained-load validation, and traceability support the hydrogen-embrittlement risk assessment.

The main lesson is that high-strength fasteners are material-process systems. Strength, coating, thread geometry, preload, hydrogen exposure, inspection timing, and lot control must be engineered together. A static stress calculation is necessary, but it is not enough to release a hydrogen-susceptible fastener.