Chromatographic Peak Shape Changes?

[dkreller]

Hi,

I observed some odd changes in my chromatographic peaks during a recent experiment, and this has me scratching my head. In this type of experiment I am not doing separations chromatography. I am actually injecting a complex mixture (naturally occuring dissolved humic substances - complex weak organic acids with a range of structures and molecular weights) onto a sorbent (crystalline Al oxide - corundum), in a manually packed column (30x4.6mm). I know very well there is a range of interactions between the sample and the sorbent - including some strong interactions that lead to adsorption. Adsorption consumes reactive surface sites and leads to an irreversibly bound organic coating on the column packing solids. Yes, the surfaces seriously change over the course of my experiment, that is what I am looking at!

Here is a link to an image that I've created to feature these changes: http://chemistry1.che.georgiasouthern.e ... hanges.jpg

In each injection, molecules of sample injected do one of two things. They either adsorb/stick to the sorbent, or elute and are detected. However with each injection in which adsorption occurs some of the surface sites get consumed and with each successive injection, since there are fewer surface sites available, there is lower adsorption, and greater elution. The chromatograms increase in peak height/area and are a broad unresolved peak since the sample contains such a complex mixture.

Since the chromatograms really range in peak height/area I normalized each of the 15 chromatograms to their maximum peak height. This allowed me to focus on the changes in the peak shapes.

Notice how in injections 1-5 the peaks systematically broadened. I hypothesize that this occurred because the humic substance molecules are relatively hydrophilic, compared to the surface. As the organic surface coverage increased, the surface becames more hydrophilic. Then the humic molecules in solution had weak (H bonding, electrostatic?) interactions with the organic-modified surface that they didn't really have with the uncoated surface. I have other evidence for this that makes me feel somewhat confident in this hypothesis.

However, I am shaking my head trying to understand why, over injections 5-15, the peaks systematically seemed to become a little less broad. Part of me wants to explain this, but I learned (with the help of dearly missed U. Neue) not to get too carried away thinking about the surface chemistry if there is a physical/chromatographic/column hydrodynamic explanation for the phenomenon.

Any thoughts? Thanks for your consideration!

Dave

[tom jupille]

I'm not in Uwe's league (very few people are ), but I can take a "hand-waving" crack at this one.

If you're comfortable with distribution isotherms, visualize a situation with two types of active sites:
- strong active sites present at some concentration
- weaker active sites present at a much higher concentration.
(this is *gross* oversimplification!)



As long as your sample concentration stays in that steep isotherm region, you get a narrow peak with high retention.
If the sample concentration is such that a significant amount is on each isotherm, then you will get broad peaks, with retention decreasing at higher concentration.
Once the sample concentration gets so high that most of it is in the shallow isotherm, the peaks will get narrower again, but with less retention.

If you want to play with the idea, I've uploaded a crude model in a spreadsheet:
http://www.lcresources.com/sandbox/crai ... sition.xls (it's large and SLOOOWWW)
Change the "lf" value from 10 down to 0.0001 in factors of 10 (10, 1, 0.1, . . . ) and watch what happens to peak shape.

In your case, I suspect that the analyte concentration is staying the same, but that you are shifting the isotherms to the left with successive injections.

[dkreller]

Tom,

Thanks for your reply! I will take a close look and download and interact with the spreadsheet, and get back to you with my results and probably further questions.

Dave

[tom jupille]

Don't read more into that spreadsheet than is there. I generated it for my own use to get a feel for Tailing Factor / Asymmetry Factor comparisons, so the isotherms are arbitrary (someone who knows a lot more about adsorption phenomena would be able to use more "realistic functions!). It does illustrate the type of narrow - broad - narrow phenomenon you saw.