Two different ways materials fail over time — and neither one is a simple overload.
A part can fail years into service without the load on it ever changing, and without the load ever cycling. Those are two entirely different failure stories, and engineers routinely mix them up because both involve "damage that builds up slowly." Creep is slow, permanent stretching under a constant stress, driven by temperature-activated diffusion inside the material — it only matters when things are hot. Fatigue is crack growth driven by a repeated, cycling stress — it happens at any temperature, including room temperature, and has nothing to do with how long a steady load has been sitting there. Picking the wrong one when diagnosing a failure means picking the wrong fix.
Creep needs two things to be present at once: a sustained, essentially constant stress, and a temperature that is high relative to the material's absolute melting point — typically above roughly 0.3–0.4 Tm. Above that threshold, atoms and dislocations inside the crystal lattice become mobile enough to rearrange under load, and the material continuously, permanently stretches — even at a stress well below its yield strength — purely because it is hot and the load has been there a long time. A gas-turbine blade running for years at a bright red-orange metal temperature under centrifugal load is the textbook case. Fatigue needs a completely different ingredient: a stress that goes up and down, or reverses, over and over. It has no minimum temperature requirement — a steel bridge expansion joint fatigues from millions of ordinary traffic-load cycles at outdoor ambient temperature. Fatigue is about a microscopic crack initiating and slowly advancing a little further with every cycle; creep has no cracks at all until its final tertiary stage — it is a continuous, diffusion-driven stretching of the whole material.
The x-axis is the tell. A creep curve plots strain against time, under one unchanging stress, and it only becomes a real design concern once the material is hot enough — roughly above 0.3–0.4 Tm on the absolute (Kelvin) scale — for diffusion and dislocation climb to let the lattice rearrange under sustained load. A fatigue curve plots stress against cycles, and the temperature is almost beside the point; what matters is that the stress goes up and down. The secondary, steady-state stage of a creep curve is the one engineers design against, because it sets the long-term minimum creep rate; for fatigue, the corresponding design number is where a crack initiates and how many cycles it takes to grow to failure. Some components — a turbine blade is the classic one — see both a sustained high-temperature centrifugal stress and thermal/mechanical cycling every startup and shutdown, and fail from creep-fatigue interaction, a combined mechanism that has to be evaluated on its own terms rather than as "creep or fatigue."
False — and mixing the two up leads directly to the wrong fix. Slow, continuous deformation under a steady, non-cyclic load at elevated temperature is creep: a diffusion- and dislocation-climb-driven process with no cracking until its final tertiary stage. Fatigue is an entirely different mechanism — cyclic-stress-driven crack initiation and growth — and it can happen just as easily at room temperature as it can at high temperature; time alone is not the ingredient, repetition is. Adding vibration damping or reducing stress concentrations will do essentially nothing for a pure creep problem, because there's no cyclic stress or crack to address. Likewise, a high-temperature creep-resistant coating or a coarser, creep-resistant grain structure won't fix a cyclic-fatigue crack that has nothing to do with sustained load or temperature. Correctly identifying which mechanism produced the failure is the whole ballgame— get it wrong and the "fix" addresses a mechanism that was never actually happening.
Explains why creep and fatigue are two entirely separate time-dependent failure mechanisms — creep being slow, permanent deformation under a constant stress at elevated temperature (above roughly 0.3-0.4 of the material's absolute melting point), and fatigue being crack initiation and growth driven by repeated, cyclic stress at any temperature — and why confusing the two leads to picking the wrong mitigation.
Both creep and fatigue are described as failures that happen gradually, over a long service life, rather than from a single overload event, which makes it tempting to lump them together as "slow failure." But the loading condition that drives each one is fundamentally different. Creep requires a sustained, essentially constant stress and a temperature high enough (typically above ~0.3-0.4 Tm, the homologous temperature) for atomic diffusion and dislocation climb to let the material continuously deform. Fatigue requires a repeated or reversing stress and has no minimum temperature — it is driven by crack initiation and growth at a stress concentration, one small increment per cycle.
A creep curve (strain vs. time at constant stress and temperature) begins with an instantaneous elastic strain, then passes through three stages: primary creep, where the creep rate decreases as the material work-hardens; secondary (steady-state) creep, where the creep rate is roughly constant and at its minimum — this is the stage engineers design against, since it governs long-term dimensional stability; and tertiary creep, where the rate accelerates (often due to internal void formation or necking) rapidly to rupture. Creep only becomes a practical design concern once the operating temperature exceeds roughly 0.3-0.4 of the material's absolute (Kelvin) melting temperature — well below that threshold, the same sustained stress produces negligible time-dependent strain.
A gas turbine blade sees a sustained centrifugal stress at a very high metal temperature for its entire operating life — the textbook creep scenario. But it also sees thermal and mechanical cycling every time the engine starts up and shuts down, which is a fatigue loading. Components like this can fail from creep-fatigue interaction, where the two damage mechanisms combine and accelerate each other, and that interaction has to be assessed as its own design case rather than as "creep OR fatigue."
For most engineering metals, no — not in a practically significant way. Creep becomes significant above roughly 0.3-0.4 of the material's absolute melting temperature, so a steel component at room temperature (a small fraction of its melting point in Kelvin) shows negligible creep, while the same steel at a few hundred degrees Celsius can creep measurably. Some very low-melting-point materials, like solder alloys or lead, can show meaningful creep even near room temperature because room temperature is already a large fraction of their melting point.
Yes — fatigue itself has no temperature requirement and occurs at any temperature where a component sees cyclic loading. At high temperature, however, fatigue can combine with creep (creep-fatigue interaction), producing faster damage accumulation and shorter life than either mechanism would predict on its own.
Primarily the secondary, steady-state stage, since it represents the constant minimum creep rate the material will sustain over most of its service life. Design life calculations typically extrapolate the steady-state creep rate to estimate how much total deformation will accumulate, or use it to set a stress/temperature limit that keeps the material from ever reaching the accelerating tertiary stage within the intended service life.
Creep rupture is the tertiary-stage failure of a creeping component: after the accelerating creep rate produces enough internal damage (often void formation and coalescence, or localized necking), the remaining load-bearing cross-section can no longer support the constant applied load and the part fractures. It is analogous to how fatigue's final fast fracture ends a slowly-growing fatigue crack, but the mechanism getting there — diffusion-driven deformation versus cyclic crack growth — is completely different.
Fractography is the usual tool. A fatigue fracture surface typically shows beach marks or striations recording each cycle of crack growth, plus a distinct final fast-fracture zone. A creep fracture surface more often shows intergranular cracking, cavitation (voids) at grain boundaries, and evidence of general, distributed deformation rather than a single dominant crack front — reflecting that creep damage accumulates throughout the material's grain structure rather than at one localized crack tip.
Try our Materials Science & Metallurgy Studio
More calculators, simulators, and guides for this discipline.