Case-Hardened Alloy Steels in Planetary Final Drive Gearing

2. Why carburized low-alloy steel dominates

A loaded gear tooth is asked for two incompatible things. At the flank, it must resist Hertzian contact fatigue (pitting, micropitting and, deeper down, sub-case fatigue), which demands very high hardness and a compressive residual stress field. At the root fillet, and through the tooth as a whole, it must resist bending fatigue and, in a machine that shock-loads its travel drives every time a track climbs a curb, must absorb overload without brittle fracture. Hardness and toughness trade against each other in any single homogeneous microstructure.

Case hardening resolves the conflict by making the material a graded composite in a single continuous piece. Carbon is diffused into the austenitized surface to roughly 0.75–0.90 %, then the part is quenched. The high-carbon surface transforms to high-carbon martensite at 58–62 HRC, while the low-carbon core, typically 0.15–0.23 % C, transforms to a much softer low-carbon martensite or bainite that retains substantial ductility and impact toughness. There is no interface to debond; the property change is a carbon gradient.

A second, less obvious benefit is residual stress. Martensite formed from high-carbon austenite is less dense than the surrounding material and transforms later during the quench, because a higher carbon content depresses the martensite start temperature. The core therefore transforms and expands first, and the case transforms last against an already rigid core. The result is a self-generated compressive residual stress in the case, commonly reported in the range of a few hundred MPa and often quoted around 200–400 MPa before any mechanical treatment. Because fatigue cracks are driven by tensile stress, this offsets the applied bending stress at the root and raises the endurance limit. Nitriding produces hardness by a different route and yields a far thinner hardened layer, which is why it is common on lightly loaded or dimensionally critical gears but rare on heavily loaded planetary sets.

3. Grade selection: hardenability and core toughness

3.1 What the alloying elements do

Carburizing grades are low-carbon steels; their base carbon content sets the core strength, and the alloy content is there almost entirely to buy hardenability, that is, to suppress the diffusional transformation products (ferrite, pearlite, upper bainite) long enough for martensite to form at the required depth and section size during a practical quench. Chromium and molybdenum are potent hardenability agents and both resist tempering softening; molybdenum in particular suppresses temper embrittlement and improves toughness in heavier sections. Nickel contributes hardenability but its principal role is core toughness, lowering the ductile-to-brittle transition and improving impact resistance. Manganese is cheap hardenability. Boron, in the treated grades, gives a large hardenability increment at very low cost but is sensitive to nitrogen control and is uncommon in premium gearing.

Compositions for the standard SAE grades are given in SAE J404¹ and, on the bar procurement side, in ASTM A322¹⁰ with general requirements in ASTM A29/A29M⁹. The ISO equivalents for case-hardening steels are in ISO 683-3²⁰. Hardenability itself is measured by the Jominy end-quench test per ASTM A255¹¹, and the guaranteed-hardenability "H" grades are bounded by the bands in SAE J1268². For any section thicker than a thin-walled ring, the H-band is more useful information than the composition.

3.2 Common grades

Two points follow from the table. First, all carburizing grades reach essentially the same surface hardness, because surface hardness is set by case carbon content, not by alloy content; paying for 4320 or 9310 buys core properties and section capability, not a harder flank. Second, 4140 appears not as a gear material but as a shaft material: through-hardened and tempered to a uniform 28–34 HRC it offers good torsional fatigue strength and machinability, and it is a mistake to carburize a shaft simply because the gear next to it was carburized.

Section size governs. A grade whose hardenability is adequate for a 30 mm planet will produce a soft, partly bainitic core in a 120 mm ring gear rim, and the case depth achieved on the tooth flank of a large part will be accompanied by a core that never reached its specified hardness. This is the most common root cause of "the heat treat was to print but the part failed": the print specified surface hardness and case depth, which were met, and said nothing enforceable about core hardness at a defined location.

4. The carburizing cycle

4.1 Carbon potential, boost and diffuse

Carburizing is carried out in the austenite field, commonly 900–950 °C, because austenite dissolves carbon readily and diffusion is fast. In endothermic-atmosphere gas carburizing the furnace atmosphere is controlled to a carbon potential C_p, the carbon content that a plain steel surface would reach at equilibrium with that atmosphere, inferred from dew point, CO₂ infrared analysis or an oxygen probe. The dominant modern practice is a two-stage boost–diffuse cycle: a boost stage at elevated C_p, often 1.05–1.20 %, which drives carbon in quickly, followed by a diffuse stage at a reduced C_p of roughly 0.80–0.90 %, during which the excess surface carbon redistributes inward. Boost–diffuse reaches a given case depth faster than a single-stage cycle at the lower potential, and it avoids the two defects of an over-carburized surface: a continuous or networked carbide film at the grain boundaries, which is brittle, and an excessive retained austenite fraction.

Case depth grows approximately with the square root of time at temperature, as expected for a diffusion-controlled process, so doubling the case depth costs roughly four times the cycle. Raising the temperature accelerates diffusion strongly, but above roughly 950 °C austenite grain growth becomes a concern unless a fine-grain-treated (aluminum-killed) steel is used, and grain size should be verified per ASTM E112¹⁸ or ISO 643²².

4.2 Atmosphere versus low-pressure carburizing

Endothermic-atmosphere carburizing is oxygen-bearing, and the strong oxide formers in the steel, chiefly manganese, chromium and silicon, oxidize preferentially along prior austenite grain boundaries at the surface. This intergranular oxidation, typically a few micrometers to a few tens of micrometers deep, locally depletes the matrix of hardenability elements, producing a thin non-martensitic surface layer. Because bending fatigue cracks initiate at the root surface, this layer is a recognized debit on bending fatigue strength, and gear metallurgical specifications generally limit its depth and its proportion of the case. Low-pressure (vacuum) carburizing with acetylene or a similar hydrocarbon, usually paired with high-pressure gas quenching, eliminates the oxidizing species and therefore the intergranular oxidation, and gives better control of uniformity in blind holes and tight root fillets. It is more capital-intensive, and gas quenching is less severe than oil, so it interacts with grade selection: a gas-quenched part often needs a higher-hardenability steel than the same part quenched in oil.

4.3 Quench and temper

After the diffuse stage the part is usually cooled to a lower austenitizing temperature, around 820–850 °C, and equalized before quenching. This reduces thermal shock and distortion and refines the structure. Quenching may be direct into agitated oil, into a quench press or plug quench for ring gears and thin sections where roundness and flatness must be controlled mechanically during transformation, or into high-pressure gas. Quench severity, agitation and fixturing are as much a part of the specification as the steel grade, because distortion determines how much stock must be removed afterwards, and stock removal comes off the case.

Tempering follows, conventionally 150–200 °C for at least two hours. This relieves the most brittle condition of as-quenched martensite and stabilizes dimensions at a small cost in hardness. Tempering higher than roughly 200 °C is generally avoided on case-hardened gearing because the hardness loss is not recovered. Aerospace and premium practice for carburized parts is codified in AMS 2759/7.

5. Case depth: definition, measurement and sizing

5.1 Effective versus total case depth

Two definitions are in use and they are not interchangeable. Total case depth is the perpendicular distance from the surface to the point where the case is metallurgically indistinguishable from the core, which is a judgment call and poorly reproducible. Effective case depth, called case hardening depth (CHD) in ISO usage and Eht in German usage, is the distance from the surface to the point at which the hardness falls to a defined limit value. For carburized steel the conventional ISO limit is 550 HV. Long-standing North American practice commonly uses 50 HRC instead, which is a lower hardness rather than a Rockwell equivalent of 550 HV, so it reads a deeper case on the same part. It is the effective depth that belongs on a drawing.

The limit value is a convention, not a physical constant, and it should be stated explicitly. For a steel with a high core hardness, say above 450 HV in a heavy alloy grade, the 550 HV crossing moves deeper, toward the slowly falling tail of the traverse, where it depends increasingly on the core rather than the case and grows more sensitive to scatter in the readings; as the core approaches 550 HV the criterion can become meaningless. ISO 18203²³ addresses this by permitting an alternative, higher limit hardness, which must then be recorded with the result. A case-depth callout without its limit hardness and test force is ambiguous.

5.2 Measurement

The reference method is a microindentation hardness traverse on a metallographically prepared transverse section, normal to the surface, following ASTM E384¹⁴ (or ASTM E92¹³ for Vickers and Knoop generally), with the case-depth construction and reporting per ISO 18203²³ or SAE J423³. Practical requirements that are routinely violated: the section must be truly perpendicular to the surface, indent spacing must respect the minimum center-to-center distance so that adjacent work-hardened fields do not interact, and the measurement location must be specified. On a gear tooth, case depth at the pitch line, at the root fillet and at the tooth tip are different quantities, because carbon diffuses into a convex tip from two sides and into a concave fillet from one. A specification that does not name the location has not specified anything. Bulk surface hardness is measured separately by Rockwell per ASTM E18¹², on a surface prepared so that the indenter is not reading decarburized or ground-affected material.

5.3 Sizing the case: module and subsurface shear

Two independent load paths set the required case depth. The bending path scales with tooth size, and the widely used guideline expresses effective case depth as a fraction of normal module, commonly quoted in the range 0.15–0.25 m_n at the pitch line for general industrial gearing. This is a starting point, not a rating.

The contact path is governed by the Hertzian stress field. For line contact the maximum orthogonal shear stress lies at roughly 0.5 a below the surface and the maximum octahedral (von Mises) shear at roughly 0.7 a, where a is the contact half-width; for typical gear-tooth contacts this places the peak a few tenths of a millimeter down. If the case is too thin, the shear stress at depth exceeds the local shear fatigue strength of the transition zone or the core, and a crack initiates below the case and propagates parallel to the surface until a large plate of case spalls away. This failure mode is called case crushing or sub-case fatigue, and it is not prevented by making the surface harder; it is prevented by making the case deeper or the core stronger. Figure 3 shows the construction.

Deeper is not automatically better. An excessively deep case reduces the tough core cross-section, raises distortion and grinding cost, increases the risk of carbide networks and retained austenite, and in extreme cases embrittles the tooth so that a shock overload breaks it rather than yielding. The optimum is bounded on both sides. ISO 6336-5²⁶ and the AGMA material and metallurgical documents, notably AGMA 923-B05³⁰, are the proper sources for turning a rating into a metallurgical requirement.

6. Surface hardness, core hardness and retained austenite

Surface hardness for carburized gearing is commonly specified in the 58–62 HRC band. Below roughly 58 HRC the flank loses contact fatigue capacity disproportionately; above roughly 62 HRC the material is usually indicating either excess carbon with carbide networks or an unusually low retained austenite fraction, and grindability and impact resistance suffer. Core hardness is typically specified in the 30–45 HRC band, again as general published practice, and it should be called out at a defined location, usually the tooth centreline at mid-face or at the root circle, because that is where sub-case fatigue is decided.

Retained austenite is the untransformed austenite left after quenching, present because the high carbon content of the case depresses the martensite finish temperature below room temperature. As-quenched case fractions of 15–30 % by volume are commonly reported for conventional cycles, and specifications frequently cap it at a value in that range, tighter for higher metallurgical quality grades. It is measured by X-ray diffraction per ASTM E975¹⁷, with quantitative metallography as a cross-check.

This is a genuine point of engineering disagreement and should be presented as one. Retained austenite is soft, reduces measured surface hardness, is dimensionally unstable because it can transform in service under stress or over time, and reduces bending fatigue strength; those arguing for tight limits emphasize these effects. On the other hand a moderate retained austenite fraction can blunt and deflect microcracks, work-hardens and transforms under contact loading, accommodates debris indentation without cracking, and is associated in the published literature with improved rolling contact fatigue life under contaminated lubrication. The practical resolution used in industry is grade-dependent limits rather than a universal target, with sub-zero treatment (typically to around −80 °C, followed by tempering) available where dimensional stability dominates, as in precision aerospace gearing. There is no single correct number.

7. Steel cleanliness and non-metallic inclusions

In a properly heat-treated, correctly case-depth-sized gear operating in clean oil, the fatigue life is frequently set by the steel's non-metallic inclusion population. Hard, angular, non-deformable oxide inclusions, principally alumina and calcium aluminates from deoxidation practice, and complex oxides from refractory or slag entrainment, act as internal stress raisers. When such an inclusion lies within the highly stressed subsurface zone, it can initiate a fatigue crack below the surface, often surrounded by the characteristic white-etching "butterfly" structure. The failure then presents as a spall with no visible surface origin, which is why fractographic evidence matters when diagnosing a returned unit.

Sulphide inclusions behave differently. Manganese sulphides are plastic at rolling temperature, elongate into stringers, and are far less damaging than oxides in the transverse direction, though they do introduce anisotropy: transverse fatigue and impact properties in a heavily stringered bar are lower than longitudinal, so the relationship between grain flow and the loaded direction matters in forged gear blanks.

Cleanliness is graded by micrographic comparison against standard chart series. ASTM E45¹⁵ defines the classical A (sulphide), B (alumina), C (silicate) and D (globular oxide) inclusion types with thin and heavy series severity levels; ISO 4967²¹ provides the equivalent international method and EN 10247²⁸ the European standard-picture method. Chart methods characterize the typical population, but fatigue is governed by the largest inclusion in the stressed volume, not the average, so extreme value analysis per ASTM E2283¹⁶ is the more appropriate tool where the objective is to bound the worst-case defect. For the most demanding gearing, cleanliness is bought at the melt shop through vacuum degassing, vacuum arc remelting or electroslag remelting, and verified by procedures such as the magnetic-particle cleanliness practice of AMS 2300 for premium aircraft-quality steel. ISO 6336-5²⁶ formalizes the connection between all of this and rating: its material quality grades ML, MQ and ME embody progressively more demanding requirements on composition control, cleanliness, hardness, case depth, microstructure and inspection, and a higher permissible stress number may only be claimed if the corresponding evidence exists.

8. Post-hardening operations that consume the case

Carburized gears distort during quench and are normally finish-ground or hard-finished afterwards. Every micrometer of stock removed from the flank removes case, and removes it from the highest-hardness, highest-compressive-stress region. Grinding stock must therefore be part of the heat-treat plan, not an afterthought: the specified case depth must be the depth remaining after finishing, and distortion control during quench is what keeps the required stock allowance small.

Abusive grinding is a distinct hazard. Excessive wheel loading or inadequate coolant can re-austenitize a thin surface layer that then re-quenches as untempered martensite over an over-tempered layer, producing a soft band, tensile residual stress and, frequently, grinding cracks. Detection is by chemical surface temper etch inspection, covered by ANSI/AGMA 2007³¹ and ISO 14104²⁷; it is a cheap inspection relative to the cost of the failure it catches. Controlled shot peening, per the Almen-strip framework of SAE J442⁴ and J443, is often applied to root fillets to restore and increase compressive residual stress; published work generally reports meaningful bending fatigue improvements, with the magnitude depending strongly on intensity, coverage and the baseline surface condition, so a specific percentage should not be assumed without process-specific evidence.

9. What a defensible specification contains

Bringing the above together, a metallurgical callout for a carburized planetary gear member should state, at minimum: the steel grade and, for anything but a light section, its hardenability requirement (H grade or a Jominy limit); melt and cleanliness requirements with the rating method named; the required surface hardness range and the scale and method; the required core hardness range and the location at which it is verified; effective case depth after finishing, with the limit hardness, test force and measurement location stated; limits on retained austenite, surface carbide morphology and intergranular oxidation; grain size; and the required post-grind temper etch inspection. Each of these maps to a published test method, so each is auditable. A specification that lists only "carburize and harden, 58–62 HRC case" is not auditable and does not distinguish a good part from a bad one.

Finally, an honest statement of limits. The value ranges quoted throughout this paper are conventional published practice and vary between application sectors, between rating standards and between individual gear geometries. They are useful as sanity checks on a specification, and they are not a substitute for a rating calculation performed to ISO 6336²⁴ or the AGMA methods with material quality grade evidence to match.