Quench Cracking Heat Treatment Distortion Case Study

Quenching is often specified as if the only question were whether the part becomes hard enough. In real production, the quench must also control temperature gradients, phase transformation strain, residual stress, distortion, surface flaws, and inspection evidence. A quench that meets hardness can still fail the part.

This case study follows an alloy-steel drive sleeve that develops surface-breaking cracks after heat treatment. The cracks appear near a shoulder and keyway relief after an oil quench. The first reaction is to accept the process because the hardness target is met. The engineering review shows that the quench is too severe for the geometry and surface flaw population.

The purpose is to connect quench severity, thermal stress, fracture toughness, hardness, distortion, non-destructive testing, and release validation into one manufacturing decision.

Case Context

The component is a machined alloy-steel drive sleeve used in a rotating power-transmission assembly. It is a medium-section part with a shoulder, bore, and keyway relief. The part is austenitized, quenched, and tempered before final grinding.

Item	Value
Batch size	200\ \text{parts}
Required final surface hardness	50 to 54\ \text{HRC}
Original quench medium	agitated oil at 40^\circ\text{C}
Original transfer time to quench	18\ \text{s}
Original temper	180^\circ\text{C}
Cracked parts found by magnetic particle inspection	36 parts
Reject fraction	18\%
Bore ovality limit after heat treatment	40\ \mu\text{m}
Measured bore ovality, original process	92\ \mu\text{m}
Suspect surface flaw depth at shoulder	1.0\ \text{mm}
Estimated as-quenched effective fracture toughness	24\ \text{MPa}\sqrt{\text{m}}

The cracks are short, surface-breaking indications concentrated near the shoulder and keyway relief. Metallography shows a hard martensitic surface layer and no welding, plating, or service damage. The evidence points to heat treatment and geometry, not field overload.

Why Hardness Alone Misled the Team

The original batch passes the hardness target:

Location	Measured hardness
Surface near shoulder	54 to 55\ \text{HRC}
Bore surface	52 to 54\ \text{HRC}
Mid-wall section	50 to 52\ \text{HRC}

The local surface hardness is slightly high, but the main problem is not simply hardness. The part cracked because high surface cooling rate, geometric restraint, and pre-existing surface flaws combined to create a fracture condition during or shortly after quenching.

A heat-treatment release therefore cannot rely on one scalar hardness number. It must check microstructure, distortion, residual stress, flaw detection, and whether the geometry can tolerate the chosen quench severity.

Thermal Strain Screen

A simplified thermal-strain screen starts with:

where \alpha is the coefficient of thermal expansion and \Delta T is the temperature difference driving local restraint.

The process review estimates that the surface and core differ by about:

near the critical cooling interval at the shoulder. Use:

The thermal strain scale is:

or:

This is large enough that geometry, phase transformation, and local restraint cannot be ignored.

Thermal Stress Screen

A first-pass restrained thermal stress can be screened as:

where E is elastic modulus and \nu is Poisson’s ratio. This is not a full heat-treatment simulation. It is a conservative screening model for whether the process is plausibly crack-driving.

Use:

Then:

The number should not be treated as the exact residual stress. Transformation plasticity, local yielding, temperature-dependent properties, and stress redistribution matter. The useful engineering point is that the stress scale is hundreds of megapascals at a surface location that already contains a flaw-like discontinuity.

Fracture Screen at the Shoulder

For a surface flaw, a simple fracture screen is:

where Y is a geometry factor, \sigma is tensile stress, and a is flaw depth.

Use:

Then:

Compare with the estimated as-quenched effective toughness:

The utilization is:

The screen predicts that the original quench can drive an existing surface flaw unstable. That agrees with the observed surface-breaking indications.

Corrected Quench Severity

The corrective heat-treatment package changes the cooling severity and fixture condition rather than accepting cracks as random defects.

The revised process is:

1. reduce transfer time variation by staging baskets closer to the quench station;
2. replace high-agitation oil with a polymer quench selected from coupon trials;
3. raise quench bath control and agitation discipline to reduce local hot spots and cold jets;
4. add shoulder-radius polish and burr removal before heat treatment;
5. temper at a higher qualified temperature to reduce brittle as-quenched response while keeping hardness within specification;
6. inspect the first three validation batches with magnetic particle inspection and hardness traverse before release.

The revised cooling curve reduces the estimated critical temperature difference to:

The corrected thermal stress screen is:

The corrected fracture screen becomes:

The corrected utilization is:

The corrected process has fracture margin on the same simplified flaw basis:

or about 32\% before considering improved surface preparation.

Distortion and Hardness Check

The correction must not solve cracking by making the part too soft or too distorted. The validation batch therefore checks both hardness and bore ovality.

Metric	Original process	Corrected process	Requirement
Surface hardness	54 to 55\ \text{HRC}	51 to 53\ \text{HRC}	50 to 54\ \text{HRC}
Mid-wall hardness	50 to 52\ \text{HRC}	50 to 52\ \text{HRC}	engineering drawing range
Bore ovality	92\ \mu\text{m}	24\ \mu\text{m} mean	\leq40\ \mu\text{m}
Magnetic particle indications	36/200 parts	0/150 parts	no rejectable cracks

For the corrected ovality data, assume the validation sample has:

The one-sided capability index is:

This supports release for distortion, provided the measurement system and sample conditions are representative.

Crack Escape Screen

The corrected validation run inspected:

and found:

With zero observed failures, a rough 95% upper bound on the defect fraction is the rule of three:

So the validation does not prove the defect probability is zero. It supports that the crack fraction is likely below about 2\% under the tested conditions. If the application is safety-critical, the release plan should require more production evidence, stronger NDT coverage, or a tighter process capability target.

Release and Validation

The corrected heat-treatment route should be released only if process evidence, part evidence, and inspection evidence agree.

Acceptance evidence should include:

1. controlled austenitizing temperature, soak time, furnace uniformity, and load pattern;
2. measured transfer time distribution to the quench medium;
3. quench bath temperature, concentration, agitation, contamination, and maintenance records;
4. hardness traverse from surface to core for representative section thicknesses;
5. metallography confirming acceptable microstructure and absence of abusive overheating;
6. bore ovality and shoulder runout before and after heat treatment;
7. magnetic particle inspection for all validation parts at the crack-prone shoulder;
8. X-ray diffraction or other residual-stress evidence when the component risk justifies it;
9. confirmation that machining marks, burrs, and shoulder radius meet the pre-heat-treatment condition;
10. documented reaction plan if hardness, ovality, quench bath condition, or crack indications drift.

Release criteria should include:

with:

and:

for the accepted flaw-screen basis, with no rejectable magnetic particle indications in the validation lots.

Engineering Lessons

The first lesson is that hardness is necessary but not sufficient. Heat treatment also creates residual stress, distortion, microstructure gradients, and crack risk.

The second lesson is that quench severity must match geometry and flaw population. A fast quench that is acceptable for a simple coupon may be too severe for a shoulder, bore, keyway, thread, or thin wall.

The third lesson is that inspection must close the loop. Magnetic particle inspection, hardness traverse, dimensional checks, metallography, and process records must tell the same story before a cracked batch is called corrected.

Good materials engineering therefore treats heat treatment as a controlled manufacturing route, not a black-box hardness operation. The quench medium, part geometry, surface condition, stress screen, toughness basis, distortion data, and validation evidence must all support the release decision.