Plastics: yield strength higher than breaking strength? 7

[CB77]

I've noticed while reviewing many plastic datasheets, that they will often list tensile strength at yield values that are higher than tensile strength at break values. How does this make sense? Yield should always come before fracture, no? (Except maybe for really brittle things where they'll essentially be the same).

And how does one measure yield at these higher values anyway without the parts first breaking?

What am I missing here?

 

[EdStainless]

Once they start to yield the strength of some materials drops significantly, but you still have to stretch them a lot before they break.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[GregLocock]

Is it as simple as after the peak strength is reached at yield, the structure breaks down internally until at some greater strain it snaps?

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[3DDave]

Where it's useful to think of metals in terms of crystal slip behaviors, in plastics it is inter-molecular behavior that typifies the performance.

If one considers that the majority of the strain is from unaligned molecular chains becoming aligned with the direction of the applied load then it is possible that there are obstacles to initial yield if that yield requires inter-molecular forces to be overcome; but after that alignment takes place those disrupted molecules which served as anchors no longer contribute to resisting strain, leaving a smaller number of molecules to resist further strain.

Below yield those anchor molecules remain in place with mostly elastic strain (not entirely as thermal motion and weak links cause failures below yield - aka creep)

So, why doesn't it just immediately break at the yield? Because the machine that is applying the load is typically providing a constant rate of strain with force being a fallout.

 

[rothers]

Learn the difference between engineering and true stress and strain.

 

[HTURKAK]

The mentioned stress and strains are engineering or nominal stresses and strains..If you calculate the true stress ( with reduced section due to elongation) , true stress at tensile strength would be higher than yield value..
The following figure is useful to get the concept..

 

[Nereth1]

My experience as a mixed metals/plastics engineer, is that for tensile results, the plastics industry runs a constant displacement/second test (as does metals industry), and then critically, much of it defines yield as "the highest load it was able to react" (what we would call ultimate strength in metals) and strength at break as "what it was reacting just before it snapped". That ultimate strength number usually occurs after it has necked down to half, stretched out to 3x its original length, and has a fracture halfway through it. It's pretty far gone.

Please see below an exemplary image of this happening, where I got told on a datasheet a UTS of basically 0 MPa, because, as it turned out upon inspection of the test curves, 3 out of 4 the samples basically held on by a few strands of glass fibre until that fibre finally got pulled all the way free and the machine recorded a break.

This frustrates me to no end, to the extent that when I received that one, I wrote a big email to the polymer supplier about how I wanted them to stop sending me datasheets by that definition, because;

1) The plastic is usually well and truly cooked and very permanently deformed once it reaches "Yield" so it's not useful to figure out a repeatable strength
2) You can't use the polymer for any kind of static load between its yield and its ultimate strength - this is on the backside of its stress-strain curve and if you operate there, it will be necking down and failing momentarily.
3) There's ambiguity as to which definition is used on a datasheet, as some polymers that don't appreciably neck or weaken from ultimate strenght to final failure (for example some fibre filled stuff), so yield will record similar to UTS, which I will interpret as e.g. a ceramic not yielding at all and then shattering, but in reality there may have been significant yield earlier in the curve, not picked up because of this definition.

So in general, yield can be used as UTS, and ultimate strength can be used for nothing in my field at least, and if you don't know by which method your datasheet has been labelled, then nothing can be used for anything because everything is up in the air.

Good luck!

 

[LittleInch]

Nereth1,

Great post and explanation.

Those curves show exactly the point that "plastic" - that's a very vague term - exhibits different stress strain characteristics to metals and is often where the confusion lies when trying to compare the two. Once you put fibres in there it all changes again.

There is very little that could be described as an "elastic" zone there and at certain levels you can get sustained strain (creep) at a constant load or stress value.

So it's a world of its own and as noted you need to look very carefully at the data and how the testing was actually performed to get any real idea of what those numbers mean.



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[EdStainless]

With moth polymers and elastomers you have materials that by our traditional definition of Yield Strength (remove load and get full return) have a yield strength that is less than 10% of their UTS. But we use them at much higher loads than that. Because at those loads and the temperatures that we operate the creep rates are slow enough that they retain enough strength and rigidity to be useful.
But creep is a PIA to measure, and often the change in elongation after high temperature (environmental) exposure is critical.
The three things to really watch on testing polymers are elongation, strain rate, and temperature.
And you need to know the GTT also (glass transition temperature) since this defines an irreversible change in properties.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[Nereth1]

LittleInch - yes you're right, although 0.2% proof strain could be useful as a standard measure of "yield", and used in combination with UTS (defined as the highest point on the curve). That way if 0.2% proof is very low compared to UTS, at least you know that the thing is acting very non-plastic/irrecoverably right from the start. And you still have the UTS to use. If it's higher, then you know you might have some more useful elastic properties.


EdStainless - GTT (Tg in my parlance?) isn't irreversible - it's discontinuous but it's quite reversible when temperature returns?

Agree creep is a PITA to measure. Time temperature superposition in general makes polymers a whole freaking mess. As does the fact that mixing your own polymer compounds/composites is so common compared to doing the same with metals. The net result of all that is that there just doesn't seem to be "established data" for polymers to remotely the same extent as what's available for metals (especially ferrous metals), and whatever you do find isn't going to be at the right strain rate, the right number of cycles, right environmental conditions (temperature, UV... humidity for gods sake), and on and on. A few years ago my company (an OEM that makes money selling widgets that have a mix of polymers and metals - by no means a materials company or R&D outfit or consultancy) threw its hands in the air and now has our own materials test lab, because it was the most practical way to be able to see where the hell you are going/achieve continuous improvement when polymers are involved.