INDUSTRY KNOWLEDGE SERIES
1. Every Joint Has an Order of Failure
Engineers spend a great deal of time deciding how strong a bolt should be, and very little deciding which part is allowed to break. Yet every bolted joint has an order of failure. Something gives first: the bolt stretches and necks down, the nut threads shear off, the flange face deforms, or the gasket leaks before any of that. The order is not an accident of dimensions. It is the result of how the fasteners were specified, and the nut is the part that decides it.
This article is about a counter-intuitive requirement: in a correctly designed joint, the nut is deliberately made slightly softer than the bolt. Not weaker in load capacity, but lower in hardness, and often lower in tensile class than the shop floor would guess. That difference is what makes the joint fail predictably, and it is the first thing lost when a customer buys fasteners on price alone.
2. The Specification Decides Who Breaks First
Nut property classes exist to be matched to bolt property classes rather than to exceed them. In ISO 898-2, a class 8 nut is rated at a proof load stress of about 800 MPa, class 10 at about 1040 MPa and class 12 at about 1150 MPa for coarse threads in the common diameter range. A class 8.8 bolt, in ISO 898-1, reaches a tensile strength of 800 MPa but is proof tested at only about 580 MPa.
Read that again, because it is the whole design intent in two numbers: the nut is proof tested at a stress equal to the bolt's ultimate strength, while the bolt is proof tested at a stress well below its own breaking point. The nut is given a margin the bolt is not. When the joint is overloaded, the bolt reaches its limit while the nut still has reserve, so the bolt breaks and the nut survives.
That is why a nut which is merely strong enough is correct, and a nut which is much stronger than necessary is not an improvement. If the nut cannot fail, the failure moves into the bolt thread, into the flange face, or into a mode nobody inspected for.
3. Why the Bolt Should Be the One That Breaks
Both failure modes end the joint, but they do not end it equally.
- A breaking bolt is visible in advance. It necks down before it parts, it shows up as a torque that will not hold, as an elongation measurement out of range, or as a leaking flange before the break. There are inspection methods for it: torque checks, ultrasonic bolt meters, and simple visual inspection of thread run-out.
- A stripping nut is silent. Thread shear begins on the first engaged threads, hidden inside the nut, and progresses turn by turn under load. Nothing shows on the outside. By the time the joint moves, the nut thread and the bolt thread are both destroyed and the joint cannot be re-tightened.
There is also a repair argument. A broken bolt can be extracted and replaced in the field, with the flange intact. A stripped nut usually takes the bolt thread with it, and in a heavily loaded or corroded assembly that can mean cutting the fastener out.
So the design puts the visible, measurable, repairable failure in the bolt, and gives the nut enough margin that it never becomes the failure location.

Figure 1. On the inspection bench: a bolt that necked down and broke, next to a nut whose internal thread has been sheared flat. Both ended a joint; only the bolt gave any warning.
4. Why the Nut Is Softer, in Three Mechanisms
4.1 It protects the bolt thread during tightening
The nut is the part that travels along the bolt thread, under load, every time the joint is assembled. If the nut is harder than the bolt, its thread flanks scrape and cold-work the bolt thread instead of bedding onto it. Repeated assembly then removes material from the bolt - the part you cannot easily replace - and each removal changes the friction coefficient the torque was calculated for.
A slightly softer nut deforms locally and beds in. The thread surface of the bolt stays intact, and the torque-tension relationship stays repeatable.
4.2 It spreads the load over more threads
Thread load is not distributed evenly. Classical analyses put roughly one third of the total load on the first engaged thread of a conventional nut, with the rest falling away quickly towards the unloaded end. A nut wall that is slightly compliant deforms enough to let the later threads share the work; a very hard, very stiff nut concentrates the load on the first two or three threads and overloads them locally.
This is also why standards put an upper limit, not just a lower limit, on nut hardness: above that limit the nut stops bedding and starts concentrating.

Figure 2. Tightening a heavy hex nut on a pipe flange. The nut travels along the bolt thread under load every time the joint is assembled, which is why its hardness relative to the bolt decides whether that thread survives.
4.3 It buys toughness where cracks start
Hardness and toughness trade off. The places where a nut cracks are the sharp transitions: the chamfer at the bearing face, the corner of the hex, the root of the internal thread. A nut pushed to the top of the hardness range has less capacity to absorb a burr, a misalignment or an impact load at those points. A slightly softer nut bends where a hard one would crack.
5. The Costly Illusion That Harder Is Better
The most common purchasing mistake in this category is upgrading the nut. A buyer who wants a stronger joint specifies a class 12 nut onto a class 8.8 bolt, expecting the assembly to become stronger. It does not. The strength of the joint is set by the bolt, because the bolt is the tension member; the nut only has to survive the bolt.
What the upgrade actually buys is cost, a harder and more brittle part, more difficulty in tightening to a repeatable torque, and in some processes a higher risk of cracking after heat treatment. What it does not buy is a single extra newton of clamp load.

Figure 3. Nuts from one production run. What a purchase order has to ask for is the class, the hardness range and the proof load certificate; none of that can be seen from the outside.
The reverse mistake is just as common and more dangerous: a class 8 nut on a 10.9 bolt, or a thin nut in a low class on a high-strength bolt, chosen because it is cheaper. Now the nut is the weak link, the failure mode becomes thread stripping, and the joint loses its early-warning system.
6. Height and Engagement Are Part of the Same Decision
Hardness alone does not fix a short nut. Thread shear capacity depends on engaged thread area, which depends on nut height. Standards recognise this by allowing thinner nuts only in higher property classes: the shorter the nut, the higher the class needed to keep the proof load match.
Two shop-floor habits work against this. Over-tapping a nut to make assembly easier reduces the engaged area on one flank and shifts load to the other. Chamfering a nut heavily to remove a burr removes thread at the very point where the first thread carries its third of the load. Both are invisible after assembly, and both convert a correct joint into a stripping joint.
7. How to Check It Before It Becomes a Claim
The requirement is easy to state and easy to verify, provided the purchase order asks for the right evidence.
- Proof load test. The nut must pass the proof load of its matched bolt class without stripping or permanent distortion beyond the limit. This is the test that proves the failure order.
- Hardness, checked at three places: thread, mid-wall and hex corner. A large drop between surface and core means the nut is not through-hardened, and its effective strength is lower than its marking.
- Thread gauges. GO and NO-GO on the internal thread, after coating. A nut that passes before plating and fails after plating will strip earlier than calculated.
- Assembly data. Torque versus turn, or torque versus elongation, on a sample joint. If the curve flattens before the target preload, the nut is deforming, not the bolt stretching - the failure order is already wrong.
8. A Purchasing Rule That Fits on One Line
Match the nut to the bolt, do not exceed it. Specify the nut class together with the bolt class, ask for the hardness range and the tempering condition, and require the proof load certificate. If a supplier offers to upgrade the nut for the same price, ask what hardness it will actually be, because a very hard nut is a liability, not a bonus.
9. Conclusion
A bolted joint does not need the strongest possible nut. It needs a nut that will not fail, and a bolt that will fail visibly if the joint is ever overloaded. The slightly softer nut, proof tested above the bolt's breaking stress, is how that order is built in.
The difference is a few points of hardness and one property class, invisible on a certificate if nobody asks. It is also the difference between a joint that warns you and a joint that lets go.
Summary: The nut is specified with a margin the bolt does not have, so that an overload breaks the bolt - visible, measurablerepairable - instead of stripping the nut silently.




