A 4140 shaft missing core hardness: hardenability, quench severity or tempering?
A Ø60 mm × 600 mm shaft in AISI 4140 (0.40 C, 0.90 Mn, 0.95 Cr, 0.20 Mo) is austenitised at 850 °C, quenched in agitated oil of Grossmann severity about 0.4, and tempered at 450 °C for 2 h. The drawing specifies 45 HRC minimum at the quarter point and 48 HRC minimum at the surface. A sectioned shaft shows 51 HRC at the surface, 44 HRC at 7.5 mm depth and 38 HRC at the centre, with roughly 20% bainite at the centre. A Jominy test on the same heat gives 50 HRC at 13 mm. Three consecutive heats behave the same. A water-quenched trial shaft met the hardness but cracked at the keyway, and the part is dynamically loaded through the keyway.
1. Decide whether the limiting factor is hardenability, section size, quench severity or the tempering schedule. Support the answer with the Jominy and section data.
Read the worked answer
The tempering schedule is not the cause of the hardness gradient. Tempering is uniform across a 60 mm section at 2 h and 450 °C; it lowers the whole curve, it does not create a 13 HRC drop from surface to centre. The surface already reads 51 HRC, so the temper is not over-softening the part.
The gradient is a cooling-rate problem. The key comparison is the cooling rate at the centre of a Ø60 mm oil-quenched bar against the Jominy curve. For H ≈ 0.4, the centre of a 60 mm bar cools at a rate roughly equivalent to a Jominy distance of 15–18 mm. This heat reaches 50 HRC only at 13 mm, so the centre never crosses the critical cooling rate - hence the 20% bainite and 38 HRC. Even the best point of this heat's Jominy curve is short of the centre's equivalent position.
So the primary limit is hardenability of this heat, worsened by low quench severity. The heat is lean within the 4140 band. The surface is fine because surface cooling is fast; the deficit grows with depth. A faster quench would shift the effective Jominy distance at the centre down to about 10–12 mm, which this heat would just satisfy - which is exactly what the cracked water-quenched trial demonstrated.
Second point worth raising: 450 °C is an unfavourable tempering temperature for 4140 because it sits in the tempered-martensite embrittlement range. Do not try to recover hardness by tempering lower; fix the through-hardening instead.
2. Evaluate the options: polymer quenching, upgrading to 4340, reducing the section, or accepting 38 HRC at the centre. Address distortion, cracking, fatigue and cost.
Read the worked answer
Polymer quench (H ≈ 0.7–0.9 with good agitation): the cheapest change. It converts the centre to near-full martensite, giving roughly 46–48 HRC at 7.5 mm and the mid-40s at the centre. Take-up is the keyway - the water-quenched trial cracked there because of the section change, so fixture, plug or re-profile the keyway, and quench vertically with controlled polymer concentration and temperature. Cracking risk is real but manageable; distortion will increase and may require a grinding allowance.
Upgrade to 4340 (0.40 C, 1.8 Ni, 0.8 Cr, 0.25 Mo): its Jominy curve holds 50 HRC to roughly 25–30 mm, so a 60 mm oil-quenched section through-hardens comfortably. It also gives better toughness at the same strength, which is what a dynamically loaded keyed shaft needs, and allows a 550 °C temper for the same strength with far lower embrittlement risk. Cost: higher material price, a new heat-treatment and machining qualification, longer lead time.
Reduce the section (bore out or step down to about Ø45 mm): metallurgically the most reliable fix, since the centre cooling rate rises sharply, and it also reduces rotating mass. But it is a design change requiring a stress re-check and customer approval, and a bored shaft may not fit the coupling.
Accept 38 HRC at the centre: reject. The problem is not the number alone, it is 20% bainite in a fatigue-loaded shaft - mixed microstructure gives non-uniform residual stress, a soft core that cannot support the case in bending, and unpredictable fatigue initiation. It will pass a hardness stamp at the surface and fail in service.
3. What verification test would you specify before releasing the fix, and what acceptance criteria would you set?
Read the worked answer
Verification should be destructive, on a sacrificial shaft from the first production lot, not on a test coupon. A coupon does not reproduce the section size or the keyway.
Tests and criteria:
- Hardness traverse at 5 mm intervals along a radial line, plus a longitudinal traverse. Criterion: 45 HRC minimum at the quarter point and 48 HRC at the surface, on the section the drawing actually calls out, and no point below 44 HRC anywhere in the load-bearing section.
- Metallography at the centre and quarter point. Criterion: at least 90% martensite, no continuous bainite network, and prior-austenite grain size finer than ASTM 5 (a coarse grain raises hardenability but hurts toughness).
- Magnetic particle inspection at the keyway and section changes after quenching and again after tempering. Criterion: no linear indications; any indication is a reject, not a blend-out, because the trial bar already showed the keyway is the cracking site.
- Dimensional and runout check after quench, against the grinding allowance, to prove the polymer quench has not pushed distortion beyond salvage.
- Incoming Jominy on every heat if the fix is a tighter 4140 specification: require 50 HRC at 15 mm minimum so that the heat is inside the top half of the band. This is the cheapest permanent control, because three heats failing identically says the purchase specification, not the heat treater, is the root cause.