[permanganate]

So i'm looking into deformulations of paints and i've noticed that GPC might be a good way to start, considering that i wouldn't have to separate each component of the formulation in order to do quantitative analysis (i would use FT IR if necessary for compound identification then proceed with the tests according to the compound).

Considering that paints formulations are more or less the same (for example nitrocellulose is a very common compound) i was thinking on making a calibration curve for this compound in specific, plotting Peak Area x Nitrocellulose concentration. However, on my research i did not find if the molecular weight affects this peak area or is it only dependent of the concentration of the sample (I'm dont have much knowledge about GPC), because there are NCs of varied molecular weight. I've found a paper but it didn't mention if the molecular weight varies to the detector response or if is it only the sample concentration.

So, is this quantitative method viable? Separating each compound individually and proceeding to techniques like FT IR/GC/HPLC would be better? Also, im looking for the maximum precision in the quantitative analysis.

[Borek]

GC works only for things volatile enough, I doubt it will work with nitrocellulose.

[wildfyr]

Borek, hes talking about gel permeation chromatography, also called size exclusion chromatography. It depends on the mean free path length of molecules in solution through a porous column as a function of hydrodynamic volume. The larger the molecule the shorter its path because it can't pass through some fraction of the pores which are under a certain size. High MW polymers elute before low MW polymers and small molecules come off last. Its how polymer molecular weight is usually determined.

permanganate, I've never heard of GPC being used for determining concentration. For one thing, GPCs work a few different ways. They can be UV detectors, light scattering detectors, or viscometers. UV detectors have a pretty linear response to concentration, but light scattering and viscocity are probably different ball games. These detectors are also very very sensitive, and I bet your injection volume will vary a little, so you would need a standard of some kind in there that is very separated from the other peaks. In all detector cases more polymer=more signal but deconvoluting it is what I would consider "very very tricky at best."

[Borek]

My bad, brain fart.

[marquis]

With polymer analysis, we started GPC with the columns and an RI detector.  They (RI detectors) are relatively insensitive, but also relatively rugged. Also, I would recommend a guard column on the GPC columns.  After you get the method down, you can remove it. Until then...

We eventually got away with straight rf  factors for each peak and having the system calculate the percentages of each ingredient. This was for PVC and the various plasticizers in each lot.   You may have to change your system, depending on your results. 

Don't know if you've decided on solvents yet..

BTW, Waters has a good tech department for more info.

Good luck