Case Hardening

Why one hardness is never enough

A gear tooth has to do two contradictory jobs at the same time. Its flank is rubbed and pounded by the mating tooth hundreds of times a second, so the surface must be glass-hard to resist wear, scuffing, and the rolling-contact pitting that eats hardened steel away one microscopic flake at a time. But the same tooth is also a tiny cantilever beam that bends every time it takes up load, and if the whole tooth were as hard as its flank needs to be, it would be brittle — a single shock load would snap it clean off at the root.

Hardness in steel comes almost entirely from carbon. Quench a high-carbon steel and the carbon is trapped in a distorted body-centred lattice called martensite, which is extremely hard and extremely brittle. Quench a low-carbon steel and there is not enough carbon to lock the lattice, so it stays soft and tough. You cannot have both hardness and toughness from one uniform composition; the two properties pull in opposite directions along the carbon axis.

Case hardening sidesteps the contradiction by making the surface and the core out of different steel, in the same part. Start with a cheap low-carbon steel that is tough but soft everywhere. Then diffuse carbon into the outer skin until the surface is, locally, a high-carbon steel. Quench the whole part: the carbon-rich skin hardens to martensite while the low-carbon core, having no carbon to lock, stays soft. One part, two metallurgies — a hard case for wear, a tough core for shock. That is the entire idea, and it is why almost every gear, camshaft, bearing race, gudgeon pin, and crankshaft journal you will ever touch has been case hardened.

The mechanism — diffusion, then transformation

Case hardening is two physical events stacked in sequence. First a diffusion step at high temperature loads the surface with carbon (or nitrogen). Then a transformation step — usually a quench — converts only the enriched region to a hard phase.

During carburizing the part sits at 900–950 °C, hot enough that the iron has rearranged into the face-centred-cubic phase called austenite, which can dissolve up to about 2 % carbon in solid solution. The furnace atmosphere is rich in carbon, so carbon atoms adsorb onto the surface and then march inward, hopping from one interstitial site to the next. The deeper they have to go, the longer it takes — and that journey is governed by diffusion physics, not by any mechanical action.

The governing equation — Fick's second law and the √t rule

Carbon concentration as a function of depth and time obeys Fick's second law of diffusion. For a flat surface held at constant surface carbon concentration C_s, in a steel starting at uniform concentration C_0, the solution is the complementary error function:

C(x, t) = C_s − (C_s − C_0) · erf( x / (2·√(D·t)) )

  x  = depth below surface (m)
  t  = time at temperature (s)
  D  = carbon diffusivity in austenite (m²/s)
  C_s = surface carbon (set by atmosphere carbon potential)
  C_0 = base carbon of the steel

The diffusivity D itself rises exponentially with temperature in the Arrhenius form D = D_0·exp(−Q/RT), with an activation energy Q of about 148 kJ/mol for carbon in austenite. That is why temperature is the single most powerful lever: raising the furnace from 900 °C to 950 °C roughly doubles D, and therefore halves the time to reach a given depth.

The structure of the error-function solution gives the famous engineering shortcut. Because x appears divided by √(D·t), the depth at which the carbon reaches any chosen value grows in proportion to the square root of time:

case depth  ≈  K · √t

  At 925 °C, K ≈ 0.50 mm / √hour  (typical gas carburizing)

  Worked example — target 1.0 mm effective case:
    t = (depth / K)²  =  (1.0 / 0.50)²  =  4 hours

  Double it to a 2.0 mm case:
    t = (2.0 / 0.50)²  =  16 hours   (4× the time, not 2×)

That square-root penalty is the central economic fact of carburizing. A shallow case is cheap; a deep case is brutally slow because every doubling of depth quadruples the furnace hours. Heavily loaded wind-turbine and marine gear teeth needing a 4–6 mm case can sit in the furnace for 30 to 60 hours, which is why such cases are designed to be exactly as deep as the contact stress requires and no deeper.

The second concentration-setting lever is carbon potential, the effective carbon activity of the atmosphere, which fixes C_s. Modern sealed-quench furnaces hold carbon potential to ±0.05 % using an oxygen probe and a programmed gas mix, often running a high-potential "boost" stage to load carbon in fast, followed by a lower-potential "diffuse" stage that levels the gradient so the very surface does not become so carbon-rich that brittle carbides or retained austenite form.

The family of processes

"Case hardening" is an umbrella over several distinct routes that differ in the diffusing species, the temperature, and whether a phase transformation is involved.

• Pack carburizing. The original method: pack the part in a box of charcoal and an energizer (barium carbonate) and bake at 900 °C for hours. Carbon monoxide liberated in the box carburizes the surface. Crude, slow, hard to control surface carbon, but still used for one-off large parts. This is the process behind antique "case-hardened" colours on firearms.
• Gas carburizing. The industrial workhorse. The part sits in an endothermic gas atmosphere (CO, H₂, N₂) enriched with a hydrocarbon such as methane or propane at 900–950 °C, then is direct-quenched in oil. Carbon potential is closed-loop controlled. Case depth 0.5–2 mm, surface 60–64 HRC.
• Vacuum (low-pressure) carburizing. Carburizing under partial pressure of acetylene followed by high-pressure gas quench. No surface oxidation, cleaner parts, faster cycles, and gentler distortion than oil quench. The premium route for aerospace gears.
• Nitriding. Diffuses nitrogen at 500–530 °C, below the austenite transition, so there is no quench and no phase change — the part comes out at its finished size with almost no distortion. Nitrogen forms hard iron and alloy nitrides directly. Requires nitriding-grade alloy steels (Nitralloy 135, 31CrMoV9) containing aluminium, chromium, or molybdenum. Surface up to 1100–1200 HV, case a thin 0.1–0.6 mm.
• Carbonitriding. A hybrid at 800–870 °C adding both carbon and nitrogen, then quenched. Nitrogen improves hardenability so the case hardens even in plain steels and tolerates milder quenches; common on mass-produced low-cost parts (fasteners, small gears).
• Induction and flame hardening. Not diffusion at all — no carbon is added. A medium-carbon steel (1045, 4140) that already has enough carbon is heated locally and fast by an induction coil or oxy-fuel flame, then quenched. Only the heated skin austenitizes and hardens. Used for crankshaft journals, axle shafts, and gear teeth too large to put in a carburizing furnace.

The hidden bonus — residual compressive stress

The most valuable thing about case hardening is something the casual observer never sees: the case ends up squeezed in compression, typically 300–600 MPa, and that compression is worth more than the hardness itself for fatigue life.

The mechanism is a timing trick during the quench. Martensite is less dense than the austenite it forms from, so when steel transforms to martensite it expands by roughly 1 %. During the quench the cool, carbon-rich case reaches its martensite-start temperature and begins transforming and expanding while the core is still hot austenite. Later the core cools, tries to contract and transform, but is now restrained by the already-hardened skin. The net result locks the surface into compression and balances it with mild tension deep in the core.

Because fatigue cracks grow only under tension, a surface pre-loaded in compression resists crack initiation: an applied bending stress must first cancel the residual compression before it can open a surface crack at all. This is why a carburized-and-ground gear tooth can carry 1.5 to 2 times the bending-fatigue load of the same tooth merely through-hardened to identical surface hardness, and why shot peening (which adds compressive stress mechanically) is frequently stacked on top of carburizing for the most demanding aerospace and motorsport gears.

Case hardening vs through hardening

The clearest way to see what case hardening buys you is to compare it head-to-head with through hardening, where the entire cross-section is brought to the same high hardness.

Property	Case hardening (carburized)	Through hardening	Nitriding
Base steel	Low-carbon (0.1–0.25 % C)	Medium/high-carbon (0.4–1.0 % C)	Special alloy (Al/Cr/Mo)
Surface hardness	58–64 HRC	50–62 HRC	1000–1200 HV (~70 HRC eq.)
Core hardness	20–40 HRC (soft, tough)	Same as surface (brittle)	30–45 HRC (pre-hardened)
Toughness / shock	High — core absorbs impact	Low — brittle throughout	Good (tough core preserved)
Residual surface stress	−300 to −600 MPa (compressive)	Often slightly tensile	−400 to −800 MPa (compressive)
Process temperature	900–950 °C + quench	800–900 °C + quench	500–530 °C, no quench
Distortion	Moderate (quench) — needs grinding	Moderate (quench)	Very low — near net shape
Case depth control	Yes — 0.1 to 2+ mm	No — hardened throughout	Yes — thin, 0.1–0.6 mm
Typical parts	Gears, camshafts, bearing races	Springs, blades, hand tools	Crankshafts, dies, valve stems

Through hardening is simpler and cheaper, and it wins where the whole part genuinely needs to be hard and impact is not a concern — a leaf spring, a knife blade, a coil spring. But the moment a part has to be hard and survive shock, the uniform option fails: it is either too brittle in the core or too soft on the surface. Case hardening is the only way to put both properties in one piece of steel.

Where case hardening shows up

• Gears. The flagship application. Carburized-and-ground gears are standard in automotive transmissions, helicopter mains, wind-turbine gearboxes, and industrial reducers. The case resists tooth-flank pitting; the residual compression carries the root bending fatigue; the soft core survives shock loading and overload.
• Camshafts and crankshafts. Cam lobes and crank journals are usually induction hardened — the part is too big for a carburizing furnace and only the rubbing surfaces need hardness. Selective hardening leaves the webs and counterweights tough.
• Rolling-element bearings. Bearing races and balls see enormous Hertzian contact stress. Through-hardened bearing steel (52100) is common, but large bearing rings and wind-turbine bearings are case carburized so the soft core resists the brittle ring fracture that a fully hardened ring is prone to.
• Gudgeon (wrist) pins and gears in engines. Small low-carbon parts case hardened to take the slap of combustion loading while staying unbreakable.
• Firearms. Bolt faces, breech faces, and historically receivers were pack carburized. The mottled blue-and-straw "case colours" on fine shotguns are a cosmetic byproduct of the bone-charcoal pack process.
• Cutting and forming tools. Nitrided dies, taps, and extrusion screws gain a hard low-friction skin without the distortion a quench would cause, preserving precision-ground geometry.
• Stainless components. Low-temperature carbon "S-phase" treatments (Kolsterising) harden austenitic stainless surfaces — surgical instruments, valves, marine fittings — without forming chromium carbides that would destroy corrosion resistance.

Failure modes — where case-hardened parts actually break

• Case crushing / spalling. If the core is too soft or the case too shallow for the contact stress, the hard case is unsupported and flakes off in chunks. The fix is matching effective case depth to the depth of maximum shear stress under Hertzian contact — usually placing the case-core boundary below where the subsurface shear peaks.
• Grinding burn. Over-aggressive finish grinding re-heats the surface, re-tempering (softening) it or re-hardening it with a brittle untempered layer, and crucially flipping the beneficial compressive residual stress into damaging tension. Nital etch inspection and gentle grinding with adequate coolant are the standard guards; this is a common warranty failure on production gears.
• Intergranular oxidation (IGO). Oxygen in a gas-carburizing atmosphere oxidizes alloying elements at grain boundaries in the first 10–25 µm, depleting them and lowering local hardenability so a soft, weak skin sits on top of the hard case. Vacuum carburizing eliminates it; grinding it off afterward is the conventional remedy.
• Retained austenite. Too much surface carbon depresses the martensite-finish temperature below room temperature, so part of the case stays as soft austenite. It lowers hardness and can transform later in service, changing dimensions. Controlled by lower carbon potential, sub-zero (cryogenic) treatment, or tempering.
• Distortion. The quench warps the part; on precision gears the entire economic case for grinding after hardening exists because the part moves during heat treatment. Press quenching and vacuum gas quenching reduce it.
• Hydrogen embrittlement. Atmospheres rich in hydrogen, or downstream plating, can leave diffusible hydrogen in the hard case that causes delayed cracking under sustained load. High-strength cases are baked after processing to drive hydrogen out.

Common pitfalls when specifying case hardening

• Quoting hardness without depth. "60 HRC surface" is meaningless for a loaded part without an effective case depth (commonly defined as the depth to 50 HRC or 550 HV). Always specify both, plus the test method.
• Forgetting to protect soft zones. Threads, bores, and surfaces machined after hardening must be masked — copper electroplating or anti-carburizing paint (stop-off) — or they will carburize and become unmachinable.
• Choosing the wrong steel. Carburizing needs a low-carbon case-hardening grade (8620, 16MnCr5, 20MnCr5); nitriding needs a nitriding alloy; induction hardening needs a medium-carbon grade. Asking a process to harden a steel it cannot is the most common rookie error.
• Over-deep cases. Specifying more case than the contact stress requires quadruples furnace time for every doubling and can make the case brittle and prone to spalling. Depth should be designed to the stress field, not maximised.
• Ignoring distortion in the drawing. If post-hardening grinding is not planned and toleranced, the part will be out of spec after heat treatment. Leave grind stock and dimension the part as the ground, not the as-machined, geometry.