Retention Time and Concentration

[Foobs]

I am trying to identify an unknown peak (acetaldehyde, I think) on a chromatogram (TCD, packed column). To that end, I injected a saturated vapor of the component that I believe the peak to be. However, the saturated vapor elutes a minute or so (65 minutes for the more concentrated "standard" verse 66 minutes in the real sample. Would an increase of the concentration by an order of magnitude decrease the residence time?

Thank You

[Don_Hilton]

Typically a peak apex is retarded if the peak is overloaded on the GC column. On thing to do is to inject a sample from the authentic and the unknown, mixed in the same injection - if thre are two peaks or the peak is flatter than either peak individually, you have two copunds with similar retention times.

Or try to confirm the separation with a different stationary phase.

[Foobs]

The problem is that the unknown sample was a gas that I was unable to save. I did something like what you suggested a few months ago with a liquid sample, but it isn't possible here. Water in the unknown gas sample could be an issue, though the peaks of current interest are all later than the water peak.

I think I'm going to try to make my "standard" more dilute (I bubble a gas through my liquid "standard" and inject the vapor through a 6-way valve).

[larkl]

It's pretty typical for the peak to come out earlier at higher concentrations. If you look, the end of the peak is usually right at the same place. Can't explain it, but I've seen it many times.

Do you have a gas bulb, say 500 or 1000ml? You could inject some of the acetaldehyde in this and it would dilute then in the air already present. I've done this successfully before using the saturated vapor from the acetaldehyde bottle. It's tough stuff t deal with given the extreme volatility.

[HW Mueller]

larkl, typical? I would ask myself whether I made some sort of mistake.

[Suresh Seethapathy]

HW Mueller,

I have observed minor retention time changes as larkl indicates, but hardly has any correlation with concentration. However these changes were mostly for peaks eluting at around 30 or more minutes, meaning mostly semivolatile compounds. I have wondered if small changes in the position of the glass wool in the liner may be the cause.

Suresh.

[larkl]

I'm probably wasting my breath, but I've seen it over and over again. With pneumatic instruments and EPC instruments. Mostly with PLOT columns, maybe that 's part of the phenomena?

[Ron]

Could you please list your conditions? I am a little surprised that the retention time for the single analyte of interest is 65 or 66 minutes. That is a long time for the analysis of a single component.

If there is a high concentration of water in your ample that may have an effect on retention times. Water is tightly bound to some column packings, and if there is enough water in the column this can shift the retention time to an earlier value.

[HW Mueller]

If a wrong stat. phase chemistry is not the culprit, I can think of one other reason for a correlation between a rise in analyte amt/concentration and a decrease in rt: Strong tailing.

[BB65]

Correct, tailing will result in an smaller RT for a given analyte. On PLOT columns, overloading a peak will result in tailing, thus shorter retention times.

The tailing occurs because of the mechanism of separation (adsorption vs diffusion on std. columns).

Bill

[HW Mueller]

BB65, tailing will delay the peak. The more you inject (to a certain point) the less will the tailing influence the peak apex, which will thus come earlier at higher injections. See what happens when you overload in Don_Hilton`s contribution.

[BB65]

HW Mueller,

I think we are talking about two different scenarios. On a PLOT column, overloaded peaks will tail. Therefore, if more sample is introduced onto the column (lower split flow or larger injection volume) the result will be even greater tailing, and lower reported retention times.

This is opposite what you would normally expect for standard bonded phase columns due to the way the separation is achieved (adsorption instead of diffusion).

Foobs, the original poster was using a packed column, but the type of packing was not noted. I can't say for sure if the adsorption separation mechanism is why Foobs saw the earlier retention time (packed GC is not an area that I have a lot of expertise in).

Bill

[HW Mueller]

Sorry, can´t follow here.

[BB65]

This may be getting a little off topic, so please indulge me.

When a PLOT column is overloaded for a particular analyte, all of the adsorption sites are tied up. The excess analyte has nothing to slow it's progress through the column so the amount in excess elutes first, followed by a decreasing signal as the adsorbed analyte gradually desorbs from the stationary phase. The result is a peak maximum followed by a tail. When even more sample is introduced into the column, the adsorption sites become tied up faster and the peak maximum occurs even earlier, followed by a longer tail.

It initially seems counter-intuitive, but it is a phenomenon that can be easily demonstrated (in fact, in the past I had to do it to see it myself).

Also, the degree by which the retention time "advances" isn't all that large (k changed from 5.02 to 4.94 in the overloaded condition for the analyte of interest).

Bill

[JI2002]

For adsorption chromatography, distribution coefficient K usually decreases with increase in sample concentration, in this case, shorter retention time and tailing are observed.

For partition chromatography, partition coefficient K mostly increases with larger sample concentration, so longer retention time and fronting are observed.

If K doesn't change with sample concentration, in theory a perfect symmetrical peak should be formed.