Nitrided Steel Is Not a Salt-Spray Shield: 38CrMoAlA’s 70% Fatigue Drop

Failure Analysis · Field Note from FUSHUN METAL

Nitrided Steel Is Not a Salt-Spray Shield: What Five Wet-Heat / Salt-Fog Cycles Did to 38CrMoAlA

A nitrided surface is hard, wear-resistant, and dimensionally stable. It is not, by itself, a marine service rating. Recent published data on 38CrMoAlA — the workhorse nitrided steel for aero-engine blower drive shafts — shows a median fatigue limit of 752 MPa falling to 218 MPa after five multi-cycle wet-heat / salt-fog cycles. That is a 70 % drop, and it happened in a test protocol that mimics a real marine atmosphere. Specifiers who read only the hardness number will not see it coming.

1. Grade cross-reference: 38CrMoAlA and its international equivalents

The article under review uses the Chinese GB/T 3077 designation 38CrMoAlA. The same alloy chemistry is sold under different names in other systems, and the compositional window is tight — the aluminium content in particular defines the grade as a nitriding steel, not a generic Cr-Mo alloy. The table below is the minimum an international buyer or specifier needs to recognise before quoting a substitute.

Aluminium is the grade-defining element. Below 0.70 % the nitrided compound layer thins and surface hardness falls; above 1.30 % coarse AlN precipitates embrittle the case. The international equivalents above all sit in the same window — they are interchangeable for the purposes of the test discussed here.

2. The intuition: a hard surface should be a durable one

38CrMoAlA is selected for one reason above all: it takes a deep, hard, dimensionally stable nitrided case at low temperature (500–540 °C) without the distortion of through-hardening. The published mechanical floor is ≥ 980 MPa tensile, ≥ 835 MPa yield, surface hardness 950–1100 HV at the standard 0.3–0.5 mm case depth. For a blower drive shaft, a high-precision spindle, a hydraulic piston, or an extrusion screw, those numbers are why the grade is on the drawing in the first place.

The intuition — sometimes spoken, sometimes just assumed — is that a surface that hard and that wear-resistant is also a surface that survives aggressive environments. The intuition is wrong for marine service, and the recent published data make the size of the error impossible to ignore.

3. The test: marine service conditions, simulated

The data come from a multi-cycle accelerated corrosion study published in Chinese Journal of Engineering (Engineering Science, 2024) by Yang, Li, Huang, Liu, Zhu and Zhan at AECC Beijing Institute of Aeronautical Materials and the University of Science and Technology Beijing. The protocol was built to mimic the in-service environment of an aero-engine blower drive shaft exposed to a tropical marine atmosphere:

• Wet-heat phase — 7 days. Per GJB 150.9A-2009, 43 ± 2 °C, 95 ± 5 % RH.
• Neutral salt-fog phase — 4 days. Per GJB 150.11A-2009, 35 ± 2 °C, 5 ± 0.1 % NaCl, pH 6.5–7.5.
• Acidic salt-fog phase — 3 days. Same temperature and NaCl, pH adjusted to 3.5–4.5 with dilute H₂SO₄.
• One cycle = 14 days of combined exposure. Specimens were pulled at 1, 3, and 5 cycles for tensile testing (per GB/T 228.1-2021) and rotating-beam fatigue testing (per GB/T 26077-2021, R = −1, 10⁷ cycles).

The acidic fog phase is the part most lab protocols leave out. It is also the part that makes the test relevant: industrial coastal air carries SO₂ from shipping and refining, and the pH 3.5–4.5 window approximates what an acidified salt film does to a real shaft.

4. The numbers: a 70 % fatigue drop in five cycles

The headline is the fatigue number. Median fatigue limit was measured by the staircase method at 10⁷ cycles, R = −1:

The tensile side tells the same story from a different angle. The as-received condition is 980 MPa tensile, 842 MPa yield, 18 % elongation, 63 % reduction of area. After 5 cycles the numbers are 870, 307, ~6 %, and ~33 % respectively. Yield strength collapses by 64 %; ductility by two thirds. Reduction of area is cut in half.

Notice that the biggest single drop is between cycle 0 and cycle 1: fatigue − 47 % in the first 14 days, yield − 30 % in the first 14 days. Most of the damage is done early. Cycles 3 and 5 add relatively little, because the corrosion-product layer thickens into a partial diffusion barrier. If a part is going to fail, it is going to fail in the first exposure window — the rest of the test is just confirmation.

5. The mechanism: why pits kill fatigue

The corrosion side of the story is comparatively mild. Average corrosion rate climbs in the first cycle, then flattens at roughly 1.47 mm/y as a dense Fe₃O₄ / Fe₂O₃ layer forms. Localised pitting, on the other hand, keeps going. Laser-confocal measurements of the cleaned surface give:

A 0.5 mm pit is, on its own, a fatigue crack starter. The fracture-surface SEM in the published study shows the classic signature: fatigue cracks nucleate at the pit root, run as intergranular cracks around the pit, and accelerate once they re-enter the ductile matrix. The corrosion pit does three things at once — it reduces the effective load-bearing cross-section, it acts as a stress concentrator with a notch radius of order tens of microns, and it furnishes a pre-existing crack that needs no incubation period under cyclic loading.

The chloride chemistry underneath is straightforward. Cl⁻ penetrates the porous Fe-oxide film, locally breaks it down, and lets Fe³⁺ enter solution as FeCl₃. The bare metal beneath becomes the anode of a self-sustaining micro-cell. The acidified salt-fog phase is what keeps the pH low enough for the attack to persist past the first few days.

6. The ASTM G217 paradox

A neat detail in the published data is that the average corrosion rate stabilises while the mechanical properties keep falling. The authors apply ASTM G217’s ratio a = (local corrosion rate) / (average corrosion rate), and report a < 2 from cycle 3 onwards — which, by the standard’s own criterion, suggests the material is still in acceptable shape for service. The fatigue data say the opposite: a 70 % drop in the endurance limit is not a material in acceptable shape, by any criterion that matters to a drive shaft.

The lesson is that a single average-corrosion-rate number can pass a screening test while the part underneath is being eaten by a few deep pits. For a fatigue-loaded component, the depth and distribution of the deepest pits — not the average mass loss — are the variable that matters. The ASTM G217-style ratio is a useful screen for uniform-corrosion-resistant alloys; it is the wrong screen for a nitrided, fatigue-loaded shaft.

7. What this means for buyers and specifiers

If you are buying or specifying 38CrMoAlA (or its SAE 135 Mod / EN 41CrAlMo7-10 / JIS SACM645 equivalents) for a marine-exposed application, the published data argue for three changes to the usual conversation:

1. Stop asking only for hardness. A 950–1100 HV nitrided case is not a corrosion-rating certificate. Ask for the test protocol the supplier actually used, and ask for data at the end-of-life condition the part will see, not at receipt.
2. Treat fatigue limit, not tensile strength, as the design variable. A 11 % drop in ultimate tensile strength is easy to absorb with a safety factor. A 70 % drop in the median fatigue limit is not. For shafting, the staircase fatigue number at 10⁷ cycles, R = −1, in the relevant environment, is the number that should drive the design margin.
3. Specify the protection you actually need. A marine-exposed 38CrMoAlA shaft should ship with a documented barrier: a hard chrome or electroless nickel underlayer, a topcoat appropriate to the chloride / SO₂ environment, and a re-inspection interval tied to the pit-depth growth rate rather than to a calendar. Where the design cannot tolerate a 70 % fatigue reduction over a 70-day exposure, the right answer is sometimes a different alloy — a 316-class austenitic, a 2205 duplex, or a nickel-base alloy such as Alloy 625 — not a thicker coat of nitrided case on the same steel.

8. The point of view

Nitriding is a surface-engineering answer to a surface problem — wear, contact fatigue, dimensional drift. It is not a corrosion answer. A shaft that lives in a wet-heat, salt-fog, acidified atmosphere will lose roughly 70 % of its fatigue limit in the first 70 days, regardless of how hard the case is, and the loss is driven by a handful of deep pits rather than by the average mass loss on the certificate. Specifiers who treat nitriding as a marine rating are buying a 752 MPa shaft and designing for a 218 MPa one. The data are now public; the spec just has to catch up.

Source: Yang L, Li L, Huang K, Liu F, Zhu J, Zhan Z. Corrosion damage behavior of 38CrMoAlA steel under alternating hot-humid and salt-fog conditions and its impact on mechanical properties for aircraft engine applications. Chinese Journal of Engineering, 2024 (accepted manuscript).