Induction Hardened, Carburised or Nitrided? Choosing Heat Treatment for Gear Racks

Heat treatment and precision grinding in production

Target keyword: gear rack heat treatment

Secondary: induction hardened gear rack, case hardening depth gears, rack hardness HRC

Meta description: Real hardness figures and case depths for induction hardened, carburised and nitrided racks - plus why hardening ruins accuracy and what the research says about optimum case depth.

There is a hardness number on your rack drawing. Where did it come from?

In our experience, about half the time it was copied from a previous drawing, and nobody remembers why. The other half it was picked because higher sounded safer. Neither is a good reason, and both can cost you either money or a failed rack.

Heat treatment on a gear rack is a set of trade-offs between surface hardness, core toughness, case depth, distortion and price. Get the balance wrong and you either wear out the flanks or crack the teeth off. Here is what the published data actually supports.

Induction hardening - the workhorse

Induction hardening runs a high-frequency current through a coil around the tooth area, heating only the surface layer, then quenches it. The core never reaches austenitising temperature, so it stays tough.

Figure 1. Surface hardness ranges and case depths by treatment method.

Achievable hardness, from published sources:

Two hard constraints govern whether a material can be induction hardened at all. The process requires steel with carbon content of 0.30% or more - below that there simply is not enough carbon to form martensite. At the top end, carbon should stay below about 0.55%, with low sulphur and phosphorus, to reduce the chance of quench cracking.

That window, 0.30% to 0.55% carbon, is precisely why S45C and 42CrMo4 (SCM440) dominate rack production. It is not tradition; it is metallurgy.

One limitation matters more than people expect: even when the tooth flank and tip harden properly, the tooth root often does not.

The tooth root is where bending fatigue cracks start. If your failure mode is root fatigue rather than flank wear, induction hardening may not be solving the problem you have. The FZG study confirms the geometry of this directly - for all induction-hardened gears tested, "the hardening contour is thickest at the tooth flanks and lowest in the area of the 30° tangent to the tooth root fillet."

The accuracy penalty nobody mentions in the quote

Hardening distorts the part. On a long slender rack, that distortion goes straight into pitch error.

When induction hardening is used, gear accuracy is reduced by strain after heating. Grinding the teeth afterwards is required if high accuracy is wanted.

The size of the effect is visible in how racks are graded by process:

The same rack that measures DIN 10 (< 0.200 mm/m) straight out of induction hardening reaches DIN 5–7 (< 0.026 to 0.052 mm/m) once the teeth are ground.

That is a factor of four to eight in accuracy, recovered by a finishing operation. The way it is done: "Extra material is left during the milling of the teeth so they can be ground after hardening."

So the real specification is not "induction hardened." It is "induction hardened and ground after hardening." If your drawing says the former and your application needs the latter, you will be disappointed by a rack that is technically compliant.

One practical consequence worth knowing before you plan machining: secondary operations are normally possible on racks in any condition except induction-hardened teeth. Drill your mounting holes before hardening, not after.

Carburising - when you need a hard case and a tough core

Carburising diffuses carbon into the surface of a low-carbon steel, then quenches. You end up with a high-carbon hard case over a low-carbon tough core.

Standard carburising practice for gear and rack production.

For 16MnCr5 (EN 10084, 1.7131), one of the standard European case-hardening grades, published figures after proper carburising and quenching (Otai Steel):

Surface hardness: 58–62 HRC

Effective case depth: 0.6–1.2 mm typical

Core hardness: 28–35 HRC

Core tensile strength (case hardened condition): 850–1000 MPa, yield ~550 MPa

The core hardness figure is the point of the whole exercise. At 28–35 HRC the core still absorbs shock and resists crack propagation, while the 58–62 HRC case resists pitting and wear. A through-hardened rack at 55 HRC has no such reserve - it is hard all the way through and correspondingly brittle.

Carburising costs more, takes longer, and distorts more than induction hardening. Specify it when contact stress is genuinely high and shock loading is present.

How deep should the case be? The research has a number

This is where most specifications are guesses, so it is worth quoting actual research.

Figure 2. Root bending strength falls when case depth exceeds 0.25 x module (FZG / TU Munchen research data).

A study by FZG at TU München (FVA 660 II, published via AGMA 22FTM14) tested 42CrMo4 gears at normal modules of 14 mm and 20 mm with gap-by-gap induction hardening, measuring tooth root bending strength against surface hardening depth (SHD400, the depth to the 400 HV1 limit).

Findings, quoted directly:

"there seems to be an optimal SHD400 in the range of 0.15 to 0.25 × m_n."

"The tooth root bending strength of the variant with an SHD of 0.35 × m_n reaches only 80 percent of the tooth root bending strength of the variants with lower SHD."

Deeper is worse. A case at 0.35 × module loses a fifth of its root bending strength compared with 0.15–0.25 × module.

The mechanism is residual stress. A shallow hardened case leaves the surface in compression, which suppresses crack initiation. Push the case too deep and the compressive layer weakens and the transition zone moves into a region of high bending stress.

Practically, for a module 4 rack, that range is 0.6 to 1.0 mm. For module 8, 1.2 to 2.0 mm. The study's recommended conditions, for gear sizes m_n = 14 to 20 mm: SHD400 at 0.15–0.25 × m_n, roughly equal on left and right flanks, surface hardness 650–750 HV, no hardening cracks.

A related figure from the same study is worth knowing if you have a fatigue-critical application: the shot-blasted variant reached a nominal bending stress number of nearly 600 N/mm², about 30 percent higher than the unblasted reference, and higher than the case-hardened comparison. Shot peening induces compressive residual stress and it works.

Set against that, ISO 6336-5 rates surface-hardened gears about 20 percent below case-hardened gears for maximum allowable stress. Carburising still wins on raw capability - induction hardening plus shot peening just closes the gap at lower cost.

Nitriding - hard, thin, and barely distorted

Nitriding diffuses nitrogen into the surface at low temperature, typically 520 °C (Bodycote). No quench, no phase transformation, minimal distortion.

The layer has two zones: a compound "white" layer giving "high resistance to wear, scuffing, galling and seizure," and a diffusion layer beneath providing "improved fatigue strength" and supporting the compound layer.

Two constraints decide whether nitriding is available to you:

Material. It is "most effective when applied to the range of steels containing nitride-forming elements such as chromium, molybdenum, vanadium and aluminium." All ferrous materials can be gas nitrided up to 5% chromium; above that, plasma nitriding. Critically, for plain carbon steel: "Due to its lack of alloying elements, AISI 1045 steel does not respond to the nitriding process" (AZoM). If your rack is C45, nitriding is not an option.

Case depth. Nitrided cases are thin. For 38CrMoAl, a nitriding-specific grade, published figures are 900–1100 HV (~65–70 HRC) with a case of only 0.3–0.6 mm.

That thin case is the limitation. Under heavy Hertzian contact stress, a thin hard layer over a softer substrate can suffer case crushing - the layer collapses into the core beneath it. Nitriding suits racks where wear and galling dominate and contact stress is moderate, and where distortion must be minimal because the rack is already finish-ground.

The material must be hardened and tempered before nitriding for optimum results.

Choosing

FAQ

Is 60 HRC better than 50 HRC on a rack?

Not automatically. Higher hardness means better wear resistance and worse toughness. A 60 HRC rack in a shock-loaded application will chip teeth that a 50 HRC rack would survive. Match hardness to the failure mode you are designing against.

Can I induction harden a C45 rack after machining the mounting holes?

You can, but distortion affects hole position and flatness as well as tooth pitch. Standard practice is to avoid secondary operations on induction-hardened racks. Machine first, harden second.

Why does my hardened rack have worse backlash consistency than the soft one it replaced?

Almost certainly hardening distortion, if the teeth were not ground afterwards. The published pitch error for induction-hardened non-ground racks is around ±0.200 mm/m against ±0.150 mm/m for a soft rack - the hardened rack is genuinely less accurate.

What hardness should the pinion be relative to the rack?

Equal or harder. Because the pinion sees the most cycles, its hardness and quality class should always equal or exceed that of the rack. The pinion makes many revolutions per metre of rack travel, so it accumulates contact cycles far faster.

Does deeper case depth always mean longer life?

No, and the FZG data is explicit: at 0.35 × module the root bending strength drops to 80% of the value at 0.15–0.25 × module. Over-specifying case depth costs money and reduces strength.

Tianjin OuNaiDa Transmissions Technology operates tooth hardening and grinding equipment in-house and supplies racks in S45C, 42CrMo4 and case-hardening grades. Tell us the load, the shock characteristics and the accuracy target, and we will recommend the treatment rather than just quoting the hardness you asked for.

Induction hardened racks, modules M5 to M8