Two numbers on the same datasheet, one of which design actually uses — and it isn't the bigger one.
Every material datasheet lists two headline stress values, and it's tempting to treat the larger one as "how strong the part really is." That larger number is the ultimate tensile strength (UTS) — the single highest stress the material reaches before it starts necking down toward fracture. But structural and mechanical design doesn't design to that number. It designs to the yield strength — a lower value that marks the end of purely elastic, fully-reversible deformation and the start of permanent, plastic deformation. The reason isn't caution for its own sake: a part stressed anywhere near its UTS has already yielded, already deformed permanently, and for essentially every engineering purpose is already considered failed— even though it hasn't snapped in two.
Below the yield strength, a loaded part is deforming elastically— stress and strain are proportional (Hooke's law), and removing the load returns the part to its exact original shape. Nothing has permanently changed; the atomic bonds simply stretched and relaxed back. The yield strength is the specific stress value at which that stops being true — the point where the stress-strain curve departs from its initial straight line, and dislocations begin gliding through the crystal lattice, permanently rearranging it. Cross that line, even slightly, and some fraction of the deformation will remain after the load is removed. That's why yield strength — not UTS — is the number that shows up in allowable-stress tables and factor-of-safety calculations: it's the actual boundary between "still the original part" and "permanently changed part."
Keep loading a ductile material past yield, and it doesn't immediately break — it keeps carrying more stress, because plastic deformation itself work-hardens the material and temporarily increases its resistance to further deformation. That resistance peaks at the ultimate tensile strength, the single highest stress value the material reaches in the entire test, well beyond yield. Immediately past that peak, necking begins — deformation localizes into one narrowing region — and the specimen heads toward fracture. UTS is a real, meaningful number: it's the material's absolute peak load-carrying capacity, useful for things like predicting ductile rupture or sizing a fastener's ultimate pull-out load. What it is not is a safe design limit, because by the time stress reaches anywhere near it, the material has already been permanently, often severely, deformed for a long time.
A factor of safety is a margin between the maximum stress a part will actually see in service and the stress at which the part stops behaving the way it's supposed to. For essentially all static structural and mechanical design, that reference point is yield strength — allowable stress = σy / FS— because the part's job is to hold its shape and geometry, not merely to avoid snapping in half. A bracket, shaft, or pressure vessel that has yielded has permanently changed dimensions, however slightly, and a structure full of parts that have quietly, permanently deformed is a structure that no longer fits together, no longer holds its tolerances, and in many cases has already lost the function it was designed for. Designing to UTS instead would mean deliberately operating in a region where the part is guaranteed to already be permanently bent, stretched, or distorted — an outcome that is only ever acceptable in very specific contexts (some designed-to-yield crush structures, sacrificial elements, and post-yield reserve-capacity checks in seismic design) that are explicitly engineered around it, not the general case.
Fracture is a dramatic, final failure mode — but it is not the only one, and for the vast majority of designed parts, it isn't even the relevantone. A part is generally considered to have failed the moment it can no longer reliably do its job, and permanent deformation alone is usually enough to disqualify it: a shaft that has yielded no longer holds its bearings in alignment, a bolt that has yielded no longer maintains its clamp load, a pressure vessel that has yielded no longer holds its designed dimensions. All of that damage is already locked in at any stress above yield strength — including everywhere between yield and UTS. UTS tells you where the material eventually runs out of load-carrying capacity altogether; yield strength tells you where the part stops being the part it was designed to be. Design margins protect against the failure mode that actually matters first, which is why they're built around yield strength, not the higher and later UTS number.
UTS is real and it is genuinely higher, but it isn't a hidden reserve of usable strength that caution is leaving on the table. UTS is the peak stress the material reaches afterit has already yielded and permanently deformed throughout the entire span between yield and that peak — by definition, every stress value on the curve at or above UTS comes with substantial, already-locked-in plastic strain. Treating UTS as the "real" strength and yield as an overly conservative buffer gets the relationship backwards: yield strength is the last point at which the part is still, dimensionally, the part that was designed. Everything past it — UTS included — describes how much load a now-permanently-altered, already-functionally-failed piece of material can still carry on its way to eventually fracturing. That can be a genuinely useful number for predicting rupture loads or crash behavior, but it was never a safe operating point, and using it as one means accepting permanent deformation as the price of admission before the part has even been put into service.
Explains the difference between yield strength (the stress at which a material transitions from fully-reversible elastic deformation to permanent plastic deformation — the value engineering design actually applies a factor of safety to) and ultimate tensile strength (the single highest stress the material reaches before necking begins, well beyond yield for ductile materials) — and why treating UTS as a safe design limit ignores that substantial, permanent deformation has already occurred by the time stress reaches anywhere near it.
Datasheets list ultimate tensile strength as a bigger, more impressive-looking number than yield strength, and it's tempting to treat it as the material's "true" strength — with yield strength read as an overly conservative buffer underneath it. That gets the relationship backwards. Every stress value at or above the yield strength already corresponds to permanent, irreversible deformation, and UTS sits well past that boundary, at the very peak of the plastic region. A part loaded anywhere near UTS has already been permanently, often substantially, deformed throughout the entire climb from yield up to that peak — it has not been quietly holding strength in reserve.
Below the yield strength, stress and strain are linearly related (Hooke's law, σ = Eε): atomic bonds stretch elastically and fully recover when the load is removed. At the yield strength, dislocations begin gliding through the crystal lattice, permanently rearranging atomic planes relative to one another — deformation that does not reverse when the load is removed. Beyond yield, ductile metals typically keep gaining apparent strength for a while through work hardening, as the increasing dislocation density itself impedes further dislocation motion, until that hardening can no longer outpace the resulting reduction in cross-sectional area. That point is the ultimate tensile strength — the highest stress the material reaches — immediately after which deformation localizes into a neck and the material heads toward fracture.
Static structural and mechanical design applies its factor of safety to yield strength (allowable stress = σy / FS), because a part's job is almost always to retain its designed shape and dimensions, not merely to avoid physically breaking. A shaft, bolt, bracket, or pressure vessel that has yielded has already permanently changed geometry — misaligned bearings, lost clamp load, distorted tolerances — well before it approaches UTS. UTS remains useful for specific purposes such as predicting ultimate rupture loads, crashworthiness, and deliberately-designed sacrificial or crush structures, but using it as a general-purpose safe operating limit means accepting permanent deformation as a built-in feature of the design rather than an unintended failure.
Because yield strength marks the boundary beyond which deformation stops being reversible. A part is generally considered to have failed once it permanently changes shape, even if it hasn't fractured — misaligned bearings, lost bolt preload, and out-of-tolerance geometry are all consequences of exceeding yield. Ultimate tensile strength is reached well after that boundary has already been crossed, so designing to it means deliberately accepting permanent deformation before the part enters service.
No — it's genuinely useful, just not as a safe operating limit. UTS is the right reference for predicting a material's absolute peak load-carrying capacity, useful in contexts like ultimate rupture prediction, crashworthiness analysis, and deliberately sacrificial or crush structures that are explicitly engineered to yield or deform in a controlled way. What it isn't suited for is representing a stress level a part can be routinely loaded to without consequence.
Yes, in specific, deliberately engineered contexts — crumple zones and crush structures designed to absorb energy through controlled plastic deformation, certain seismic-design reserve-capacity checks, and some one-time-use or sacrificial components. These are exceptions built around the fact that permanent deformation (or even fracture) is the intended behavior, not evidence that UTS is generally usable as a routine design limit.
Not always unsafe in the sense of imminent fracture, but it does mean the part has permanently changed shape, which for most engineering purposes counts as a functional failure even if the part continues to carry load. Whether that's acceptable depends entirely on the application — a part explicitly designed to yield in a controlled way (like a crush structure) is behaving as intended, while a structural member that was assumed to stay elastic is not.
Brittle materials typically show little or no distinct yield point and correspondingly little plastic deformation before fracture, so their yield strength and ultimate tensile strength (or fracture strength) sit much closer together than in ductile materials. The large gap between yield and UTS — and the extended "already functionally failed but not yet fractured" zone it creates — is characteristic of ductile materials specifically, which is also where the misconception about UTS being a safe design number is most consequential.
Try our Materials Science & Metallurgy Studio
More calculators, simulators, and guides for this discipline.