Material characterization correlations over several scales - E

[PerennialII]

I'm occationally asked about correlations between macroscopically and microscopically measured material properties, the typical case in question being the elastic modulus (or more like why don't they always match). Say you measure it using nanoindentation and in the other end you do something more traditional, like a tensile test. In many cases it's a necessity to go for nanoindentation (the case I'm blabbing about here involves typically coatings of nano & micrometer scale so there really ain't an abundance of methods available), and if you happen to be in luck, and have experimental data from several sources you can use to verify your measurements at least to some extent (i.e. there is sufficient comparability ... there are a bunch of things even material wise that make the comparison meaningless), in many cases you easily get differences of the order of tens of percents. I'm probably going to take an in-depth look at this 'cos it's starting to bug me quite a bit, but was wondering whether anyone has any information / experience with respect to the uniqueness of material properties when characterized this way ? The case property is elastic modulus, since it's something which in many cases is a problem overall no matter what method you're using, and it's also quite fundamental if doing e.g. detailed wear analyses or something like that.

 

[Astronuc]

This is an issue that my colleagues and I deal with intimately with respect to developing constitutive models for the purposes of 'predictive analysis' of the behavior of a structure and its material composition. Basically, one must consider the difference between macro-mechanics of materials and micro-mechanics (its similar to the difference between classical physics and quantum physics).

By macro-mechanics, I refer to the traditional material properties and methods for measureing those properties. With traditional methods, one is measuring 'bulk' properties, which applies in a 'global' sense. Look at the tensile (or compressive) test for determining E (Young's or elastic modulus). Look at size of specimen as compared to grain size and morphology.

At the micro-mechanics level, there is a population of grain size with some mean value, and additionally there is the subgrain lattice structure, grain orientation, local variations of composition, imperfections, all of which have some general distribution.

There is a relatively new (in the sense of the last couple of decades) field of micro-mechanics and materials testing. ASTM (American Society for Testing and Materials) and ISO (International Organization for Standardization) have a lot of standards that address materials, their properties and the testing methods to determine those properties. And there is ANSI / NIST and a host of similar national organizations.

I cannot overemphasize - How one makes a material is crucial to how it performs.

PerennialII - I recommend that you look into membership in ASTM (www.astm.org[/url]), ISO ([url]http://www.iso.org/iso/en/ISOOnline.frontpage[/URL]), and/or other comparable organizations.

 

[PerennialII]

This is an issue that my colleagues and I deal with intimately with respect to developing constitutive models for the purposes of 'predictive analysis' of the behavior of a structure and its material composition. Basically, one must consider the difference between macro-mechanics of materials and micro-mechanics (its similar to the difference between classical physics and quantum physics).

I'm starting to think our turfs are pretty close to each other ... I'm making my living pretty much from the same field, developing constitutive models (primarily damage mechanical related) and all sorts of affiliated stuff.

There is a relatively new (in the sense of the last couple of decades) field of micro-mechanics and materials testing. ASTM (American Society for Testing and Materials) and ISO (International Organization for Standardization) have a lot of standards that address materials, their properties and the testing methods to determine those properties. And there is ANSI / NIST and a host of similar national organizations.

I cannot overemphasize - How one makes a material is crucial to how it performs.

PerennialII - I recommend that you look into membership in ASTM (www.astm.org[/url]), ISO ([url]http://www.iso.org/iso/en/ISOOnline.frontpage[/URL]), and/or other comparable organizations.
[/QUOTE]

I hadn't considered looking at the very microlevel aspect, and the tougher part of this problem, the nanoindentation, via ASTM or ISO. Got to start browsing my E-8 handbook first thing (don't know whether they've already included something about it there, remains to be seen), going through recent ASTM STPs & publs might be a first priority as well. Got to do that first thing. Problems such as this one seem to be surfacing ever more, all sorts of indirect methods are being developed at a growing rate to characterize materials from small material samples, non-destructively etc ... and the fine material - specific details as you pointed out become ever more important to tackle.

 

[Astronuc]

I want to get Clausius2, Gokul43201 and others to jump in on this one - because this represents one of the main challenges in engineering - particular nuclear and aerospace systems.

We used to design systems with large 'safety factors' (or large margin to failure) because of uncertainties. However, such an approach can make large systems (e.g. bridges, dams, aircraft, spacecraft (e.g. Space Shuttle)) much less economical.

So, with improvements in technology (that is in design and analytical methods, manufacturing of materials, construction, etc), we have reduced the margins in order to realize greater economy.

In some cases we have pushed the limits - and sometimes that has resulted in catastrophic failure (where, as Gene Kranz succinctly put it, "Failure is not an option").

There is a strong need for improvements and extensions in the states-of-the-arts in materials science and engineering (in fact in all engineering disciplines), non-destructive testing (NDT), and predictive analysis.

We also need to answer - General Engineering/Material Properties, which is somewhat relevant to this thread.

 

[Gokul43201]

Yes, the characterization method can drastically influence the measured property. It is for this reason that some properties are specified in terms of the techniques used to measure it. Hardness testing comes immediately to mind. The different techniques in that case are really measuring different definitions of 'hardness' and hence comparison between results is meaningless.

This applies to other properties as well, and especially when using techniques like micro-indentation. Such techniques typically measure very surface dominated behaviors while a tensile test performed on a UTM measures mostly the bulk behavior. It would be unwise to, in general, compare the two results and expect a very high correlation.

Usually, the accepted method of testing (if more than one is possible) is the one (or all those) that most closely replicates the expected field conditions on the material. Very often, this luxury does not exist, and you have to go with what you can get. I believe that micro-indentation techniques work very well for determining properties (hardness, E, TS, etc.) of coatings.

Astronuc, as far as that other thread is concerned, that looks like homework to me. I don't think we should be doing someone else's homework, especially if it's something like a paper. I could have suggested references, for solid behavior, but I really didn't know how/why you use Hooke's law with liquids (most liquids behave incompressibly) and plasmas (while these are compressible, this is an electrostatics problem, and I would solve it as such). That part struck me as odd.

 

[PerennialII]

We used to design systems with large 'safety factors' (or large margin to failure) because of uncertainties. However, such an approach can make large systems (e.g. bridges, dams, aircraft, spacecraft (e.g. Space Shuttle)) much less economical.

So, with improvements in technology (that is in design and analytical methods, manufacturing of materials, construction, etc), we have reduced the margins in order to realize greater economy.

In some cases we have pushed the limits - and sometimes that has resulted in catastrophic failure (where, as Gene Kranz succinctly put it, "Failure is not an option").

There is a strong need for improvements and extensions in the states-of-the-arts in materials science and engineering (in fact in all engineering disciplines), non-destructive testing (NDT), and predictive analysis.

Yeah, the way to go and where everything is going is to increasingly rely and apply predictive analysis methods. The increased application has led to discovery of quite a number of trap holes, which haven't been apparent when e.g. simplistic constitutive models have been applied, and quite a bit of work is still required in order to get the means from qualitative predictibility to quantative scale, at current even in a single case that often requires quite a bit of effort. But the structural and material upturn you can get when able to ditch many of the current outdated methods ... .

Usually, the accepted method of testing (if more than one is possible) is the one (or all those) that most closely replicates the expected field conditions on the material. Very often, this luxury does not exist, and you have to go with what you can get. I believe that micro-indentation techniques work very well for determining properties (hardness, E, TS, etc.) of coatings.

Typically comparing results which have some "geometric similitude" and similar loading conditions, characterized for example by the membrane to bending ratio, yield pretty good results. Overall, results I've gotten using micro-indentation tests work pretty well, the problems arise when you got information from a multitude of sources and got to make a decision upon them (which comes about often when doing multiscale analyses and related stuff which is under scrutiny nowadays). The differences I've seen usually e.g. in E have been of / to the order of 25%.

With respect to this particular case I'm probably first going to do some more detailed browsing of recent journals, someone has in all likelihood done something related to this and at least I can get some more experimental material. I've some indentation tests coming up which I can use for more detailed analyses of the problem, and as a last thing I might resort to doing some contact analyses of the indentation event itself (introducing material properties as distributions if required, got to see how the size scales match up). I'm essentially going to be working with problems where nanonindentation is applied to derive material property gradients in thickness direction (for a coating case) and as such the problem gets a bit more complicated material property wise.