Thursday, September 29, 2011

Presenting Rheology Data

I posted a new page today that provides a very brief overview of dynamic mechanical analysis and time-temperature superposition. You can read it here if you are unfamiliar with rheology testing or want a review.

One of the most difficult aspects of rheology is that the data, particularly that generated by dynamic mechanical analysis, can be plotted in so manner different, but more or less equivalent manners.

As you saw from Monday's post, I prefer working the storage and loss moduli (G' and G" respectively) vs. frequency, although I most often supply clients with viscosity vs. frequency since that is a simpler concept to explain. Other rheologists have a built-in bias towards working with compliances, J' and J", while others are hung up on tan δ. The strangest set of plots I ever saw was from a guy who always plotted G* vs. tan delta. The information is always the same, it's just that sometimes you have to work a little bit more to figure it out.

Part of the reason I like G' and G" is that the generic shape of these curves is well established. If something is varying from it, I can immediately pick up on it to probe for further details. If you are working with tan δ, you only have the ratio of G"/G', and the only universally significant measure then is whether it is greater than, less than or equal to 1. You have a similar issue with G*, where it can hide many details that are otherwise noticeable.

In some ways, all these different plots are fun - you need to be on your toes in order to run the mental manipulations in your head to translate the data back to your preferred reference frame. I imagine that it really would be best to work within the other frameworks much like it is best to be thinking in a foreign language that you are speaking, rather than translating it all into your native tongue, preparing your response and then translating it back into the foreign language. I'm just not sure that this would happen anywhere in the rheology world as a given lab will have a given preferred format. It's only when you start working with others labs or researchers that you might start running into this issue, but those conversations don't happen all that often - you're much more likely to talk with fellow labmates than with people in other labs.


Andrew Sun said...

You mentioned two different issues in your post.

First you actually emphasized the fact that two parameters are needed to describe under linear viscoelastic condition the stress information from a sinusoidal shear strain. The stress amplitude and its phase lagging. For the former one may use the norm of complex modulus |G*| and the latter the phase angle delta. But other combinations are of course mathematically numerous. Using G' and G'' also suffices because you can get |G*| and delta from them. The same holds for J' and J'', but not, for example, for a plot of |eta*| ~ freq, which does not give the information of phase angle.

The second issue you actually mentioned in you post is whether to plot the 2 needed parameters independently against frequency, or mutually each other. |G*| and delta are two sufficient parameters. But plotting one against another sometimes show "hidden" information. The same case exist for G' and G'', which when plotted mutually is called the Cole-Cole plot.

I follow your blog for quite a long time so I am sure you must know the above things very well. I am not very confident about them, instead. That's why I commented here.

John said...


Thanks for your comments. It's proof of further ways to plot the data.