How surface finish affects titanium fatigue strength

A titanium component can meet tensile specifications yet fail prematurely under cyclic loading if its surface condition is poorly controlled. Roughness, machining damage, and treatment-induced residual stresses all influence crack initiation, making surface finish a critical variable in titanium fatigue strength. For technical evaluators, the challenge is distinguishing a visually smooth surface from one that delivers reliable fatigue performance.

That distinction matters when qualifying rotating equipment, robotic joints, lightweight machinery, or safety-critical energy hardware. A polished specimen and a production component may share the same alloy designation but behave differently under repeated loading. Meaningful evaluation therefore starts with how the surface was produced, what lies beneath it, and whether fatigue testing represents the intended service conditions.

Why fatigue often starts at the surface

Under cyclic loading, small regions of a component experience repeated local plastic deformation even when the nominal stress remains below the material’s yield strength. Surface valleys, scratches, pits, and machining defects concentrate stress, creating favorable sites for crack initiation. Once a crack forms, its growth depends on loading, microstructure, environment, and component geometry—not surface finish alone.

Titanium alloys can be sensitive to these surface discontinuities, particularly in high-cycle fatigue, where initiation may occupy a substantial fraction of total life. Removing a sharp defect can therefore improve performance without changing the alloy’s tensile strength. Conversely, a finishing operation that introduces microcracks or tensile residual stress can undermine an otherwise sound component.

The relationship is not universal. Internal defects may control failure in some materials and loading regimes, especially in additively manufactured titanium. Low-cycle fatigue, dwell-sensitive loading, fretting, and elevated-temperature service also introduce mechanisms that cannot be reduced to surface roughness.

A roughness number is not a fatigue qualification

Average roughness, usually reported as Ra, is useful for manufacturing control. It is not a complete description of a fatigue-critical surface. Two surfaces with the same Ra can have very different valley depths, notch radii, texture directions, and isolated defects.

Consider a uniformly textured surface and a polished surface containing one deep scratch. Their average roughness values might be similar, yet the scratch may dominate crack initiation. The relevant question is not simply “How smooth is it?” but “What is the most damaging feature in the highly stressed region?”

Grooves running transverse to the principal tensile stress are often more damaging than grooves aligned with it, although multiaxial loading complicates that comparison. Measurement direction matters too: a stylus trace taken along machining lay may miss the profile severity revealed by a transverse trace.

A defensible surface requirement combines texture limits with defect controls and process evidence. For complex surfaces, areal measurements can reveal features that a few profile traces miss. Neither method eliminates the need to inspect critical edges, fillets, holes, and tool-entry or tool-exit regions.

Machining and polishing: the finish beneath the appearance

Machining can leave more than tool marks

Titanium’s low thermal conductivity and tendency to interact with cutting tools make machining conditions important. Tool wear, cutting speed, feed, coolant delivery, and fixturing can affect surface integrity. An operation may leave tearing, smeared material, burrs, or unfavorable residual stresses even when dimensional tolerances are satisfied.

For fatigue-critical parts, the finishing pass deserves explicit qualification. A supplier’s successful roughing procedure does not establish that its final tool path is suitable for a loaded fillet. Tool-life limits also matter: results obtained with a fresh insert may not represent surfaces produced near the end of an approved tool-use interval.

Grinding requires thermal control

Grinding can achieve tight geometry and low roughness, but poorly controlled heat input may introduce tensile residual stresses, microstructural alteration, or cracking. A fine-looking ground surface is therefore not automatically superior to a well-controlled machined surface. Abrasive condition, dressing practice, coolant delivery, and removal rate should form part of the qualified process.

Polishing helps only if it removes the relevant damage

Mechanical polishing can reduce notch severity and remove shallow machining damage. Its fatigue benefit depends on removal depth, consistency, and access to critical geometry. Smoothing the peaks while leaving a deep valley does not resolve the original risk.

Electropolishing or chemical finishing can remove material without conventional cutting forces, but neither is automatically fatigue-neutral. Process control must address uneven removal, localized attack, dimensional change, and any risk of hydrogen uptake associated with the particular chemistry and operating conditions. Thin sections and small radii require special attention.

Why a rougher treated surface can perform better

The apparent contradiction becomes clearer when residual stress is considered. Compressive residual stress near the surface can reduce the effective tensile loading experienced by small cracks and delay their initiation or early growth. A treatment that produces beneficial compression may improve fatigue performance even if it increases measured roughness.

Shot peening is a common example. Controlled impacts introduce near-surface plastic deformation and compressive residual stress. However, excessive processing can create folds, sharp impact features, or damage that offsets the benefit. Peening intensity, coverage, media condition, and treatment sequence all require control; a certificate stating only “shot peened” is insufficient.

Laser shock peening and deep rolling can also create beneficial compressive stress fields. Their effectiveness depends on geometry, material condition, process parameters, and the depth of the resulting stress profile relative to likely crack origins. No treatment should be ranked solely by its name.

Residual stress is not necessarily permanent. Temperature exposure, cyclic plasticity, and subsequent manufacturing steps may relax or redistribute it. Likewise, polishing after peening can remove part of the beneficial layer. Evaluation must address the final delivered condition, including cleaning, coating, thermal exposure, and material removal.

Hardness and wear resistance are different acceptance criteria

Nitriding, anodizing, coatings, and other surface modifications may improve wear, corrosion behavior, or contact performance without improving plain fatigue strength. Brittle layers, coating cracks, process-induced defects, and changes in residual stress can introduce competing effects.

In titanium joints exposed to fretting, reducing contact damage may be valuable even if the treatment offers little benefit in a smooth-specimen fatigue test. The test method must match the failure mechanism. For hydrogen-related service, surface-finish qualification must also remain separate from environmental compatibility assessment.

Reading titanium fatigue strength data without overclaiming

A statement such as “polishing improves fatigue strength” is incomplete unless it identifies the comparison. Improvement relative to a rough as-built surface is not equivalent to improvement over a carefully machined baseline.

Request the following information before using supplier data in a technical decision:

• Material identity: alloy, product form, heat treatment, microstructure, and sampling orientation.
• Manufacturing route: wrought, cast, or additively manufactured, including relevant thermal and consolidation processes.
• Surface condition: measured texture, defect acceptance criteria, finishing sequence, and treatment parameters.
• Loading: axial, bending, torsional, or multiaxial; stress ratio; waveform; frequency; and any dwell periods.
• Test conditions: temperature, environment, specimen geometry, and stress calculation method.
• Statistical basis: specimen count, scatter, runouts, and confidence or tolerance bounds.

Stress ratio, R, is the minimum stress divided by the maximum stress in a cycle. Results from fully reversed loading should not be substituted directly for results under a tensile mean stress. Likewise, rotating-bending data cannot automatically establish the axial fatigue capability of a component with a different stressed volume and stress gradient.

A reported fatigue strength should identify the associated cycle count and test conditions. A runout means the specimen survived the specified test duration; it does not establish unlimited life. For titanium, assuming a universal endurance limit is not a defensible shortcut.

Which standards support the assessment?

Standards provide repeatable methods, but they do not supply a universal surface-finish correction factor for titanium. Evaluators should confirm the applicable edition, scope, and any sector-specific requirements.

• ASTM E466 addresses force-controlled, constant-amplitude axial fatigue testing of metallic materials. It is relevant to many comparisons between surface conditions under predominantly elastic loading.
• ASTM E606/E606M addresses strain-controlled fatigue testing, useful when cyclic plastic strain and low-cycle fatigue are important.
• ASTM E739 provides guidance for statistical analysis of linear or linearized stress-life and strain-life fatigue data, subject to its assumptions and limitations.
• ISO 21920 addresses profile surface texture, while ISO 25178 addresses areal surface texture. Specifications should identify the required parameters and measurement conditions.
• SAE AMS2430 may be relevant where a procurement specification requires controlled shot peening. Compliance with a peening process specification does not, by itself, establish component fatigue life.

Material specifications, such as ASTM B348 for titanium bars and billets, serve a different purpose. Conforming feedstock is essential, but its certification does not qualify the fatigue behavior of the finished surface. Material acceptance, process acceptance, and fatigue substantiation should remain separate, traceable elements of the evidence package.

Turning surface-finish requirements into a procurement decision

The most useful specification begins with the component’s critical locations rather than a single roughness limit applied everywhere. Identify regions of high cyclic tensile stress, geometric concentration, contact motion, and environmental exposure. Then assign surface requirements and inspection methods proportionate to those risks.

For a loaded fillet, this may mean controlling tool-mark direction, excluding sharp scratches, verifying burr removal, and qualifying the final machining sequence. For a peened region, it may require documented coverage and intensity controls, supported by fatigue evidence and residual-stress characterization where warranted.

When comparing suppliers, prioritize tests that hold material, geometry, and loading constant while changing the surface process. Representative coupons are useful, but transfer to a production component needs justification. Coupons can miss restricted tool access, thin-wall distortion, local treatment variation, or damage introduced during assembly.

A practical acceptance package links four records: the drawing’s surface requirements, the qualified manufacturing route, inspection results from the delivered condition, and fatigue evidence relevant to service. Process changes—new tooling, altered polishing depth, different peening media, or a revised coating cycle—should trigger a documented review of that linkage.

The decisive question is not whether titanium looks smooth, or whether one treatment usually improves fatigue. It is whether the final surface has controlled defects, an understood stress state, and test evidence appropriate to the actual component. That is the basis for a credible titanium fatigue strength assessment—and for avoiding a polished part with an unrecognized fatigue risk.