effect of internal diameter of column??

[mardexis]

Uwe,

I think you are referring to the heat generated by the large deltaP introducing radial temperature gradient on the column and thus an additional broadening effect. This is actually very similar to a multiplicity of path term except that there is a uniform distribution... faster at the center and slower at the wall. The opposite of wall effect.

I believe that the data from Jorgenson showed that compounds with very different capacity factors benefitted equally from decreasing diameter. If this were a thermal gradient effect then you would expect the effect to be proportional to capacity factor if the compounds have similar van't Hoff behavior. In fact indoing a calculation I would not expect to be any effect to RTG in this data. I would expect to see loss of theoretical performance by RTG on a 2.1 x 150 mm column packed with 1.7 um particles and run over 15KPSI.

Danko,

the book Uwe mentioned is excellent but a better source on this particular topic is "Dynamics of Chromatography" which is still available at Amazon for $68. Look into Giddings theory on coupling of A and B terms. I think that is what we are seeing in the Figure from Jorgenson's paper.

[danko]

Hi Uwe,

May I recommend to you to read a book on HPLC Columns? There is one available with this title, published by Wiley-VCH. It has a pretty good section on theory.

Thank you for the recommendation. I’ll look at it when I get the opportunity. Though, I wonder how you arrived to the conclusion that I needed to read this particular book in order for me to understand the A-term better.
But anyway, I gave you my version, or should I say the conventional version of the A-term interpretation.
And I have the feeling, you disagree with it. If so, would you mind telling me what’s your version – just what happens with the molecules ones they move into the column? Do they move straight through, in a lump, or do they diffuse in different directions, as well as taking the shortest path? If they diffuse, then how far from the center do they reach? As far as the column walls allows them? Needless to say I’m open for innovative ideas.

As I stated before, there is no reason to think about the detector.

[danko]

Sorry I submitted too early by mistake. I'll try again.

Hi Uwe,

May I recommend to you to read a book on HPLC Columns? There is one available with this title, published by Wiley-VCH. It has a pretty good section on theory.

Thank you for the recommendation. I’ll look at it when I get the opportunity. Though, I wonder how you arrived to the conclusion that I needed to read this particular book in order for me to understand the A-term better.
But anyway, I gave you my version, or should I say the conventional version of the A-term interpretation.
And I have the feeling, you disagree with it. If so, would you mind telling me what’s your version – just what happens with the molecules ones they move into the column? Do they move straight through, in a lump, or do they diffuse in different directions, as well as taking the shortest path? If they diffuse, then how far from the center do they reach? As far as the column walls allows them? Needless to say I’m open for innovative ideas.

As I stated before, there is no reason to think about the detector.

Let’s forget the detector for a while. I’ll save it for the punch line – many posts from now.

PS.: I read your post carefully.

Really? And you overlooked my questions? Since you did me a favor recommending me a book, I decided to return you the favor by recommending you to read my previous post even more carefully.
Because if you had read it very carefully, you’d had observed that my mental experiment corresponds directly to yours or vice versa (see the quote bellow).

If I inject 10 micrograms on the fatter column, and 5 micorgrams on the thinner column, I will get the exact same chromatogram, no larger peak, no smaller peak, just the same peak.

Isn’t it the same as: If you inject 5 μg on both columns you’ll get the half peak height on the fatter column and the same peak widths on both columns?
In my mind, these two mental experiments and the resultant outcomes correspond to the following equation: 2 + 2 = 4.
Please don’t tell me it’s not so, because I’ll have to sue my primary school for teaching me the wrong mathematics.

Anyway, you can choose to address either one of these two questions: Why the 10 and the 5 μg loads (your “showcaseâ€

[mardexis]

I'll take a crack at the injection volume vs mass part of the last post.

If, for some reason, you only have 5 ug of a compound and you can dissolve it in say 200 nL of solvent and then inject it by partial-fill injection onto a column then you will get twice the peak height for the thinner column.

If you take the same 5 ug and dissolve it in 1 mL of solvent then inject a volume scaled to the column, 2:1 in this case, then the peak heights will be roughly the same as will the plate counts.

There should always be dilution taking place on column or you are going to be doing horrible chromatography. If one is not worried about using up a few micrograms of sample it is almost always best to use larger columns and injection volume to scale. That said, I do believe that capillary columns offer some theoretical advantage in plate counts but good luck buying columns and equipment to prove this.

[Uwe Neue]

Dancho;

So we agree that one gets the same peak height on the 2x fatter column by injecting 2x the amount.

So where does that leave the eddy diffusion or the van Deemter A-term? We got the same retention time, the same peak width, thus the same plate count, the same theoretical plate height... Seems to me that this means that the eddy diffusion and the A term are the same also.

[Uwe Neue]

To address the question of the radial dispersion or the origin of the van Deemter A-term or eddy dispersion: it is nothing but the relaxation of flow rate differences in the radial direction of the column. You can imagine the column best as a type of Galton board, where there is radial dispersion because of random flow over the randomly arranged particles. This random radial dispersion relaxes the random local flow differences.

The reason why the A-term declines with very small column diameters, i.e. under about 30 particle diameters, is the fact that the radial relaxation is large. For standard column diameters, the relaxation reaches a large part of the column, but it does not span the entire column width. But in the standard diameter range, the relaxation is the same, independent on the column width. Therefore the A-term is the same, unless you do unusual things with the column.

[danko]

Hi Uwe,

So we agree that one gets the same peak height on the 2x fatter column by injecting 2x the amount.

Absolutely! And that’s what my “showcaseâ€

[HW Mueller]

Danko, here is the dicussion on turbulent flow:
http://www.sepsci.com/chromforum/viewto ... 4&start=15

It is quite evident what you are up to, at the end your statement will again be something like: "Just wanted to mix up the forum a bit".

The more I read your statements the more confused I get about what you are trying to say. Regarding flow:
If one changes the flow rate on the same system then you get a narrower peak, same peak hight, and lower area (for instance, 2x the flow rate will half the area, reduce the peak width and the rt). Now if you change the system (here different columns) this doesn´t hold. If you take the larger column (of the example here) and adjust linear velocity such that rt is the same, as well as peak width (as compared to the smaller diameter column), it doesn´t matter what the flowrate is through the cell. Even if you changed the cell so flow would be 4x it wouldn´t change peak width. Now if one could devise a hypothetical concentrator which does not increase post column volume one could concentrate the eluate of the thick column to half and get half the flow rate and double the peak hight. The peak width can not be effected.
It seems that you are arguing intensity when the rest is talking about separation efficiency and vice versa?

[danko]

Hi Hans,

I’ll address your comments later – after Uwe has answered my questions. I just don’t want this discussion to be “dilutedâ€

[Kostas Petritis]

Danko,

According the chromatographic theory, the concentration of a compound in the Apex of a peak is given by:

Cmax= (Q0/Vr) * (N/2pi)^1/2

where Q0 is the injected amount and Vr is the retention volume (Vr=tr * F where tr retention time and F is the flow rate). As a result for "concentration detectors" such as UV, fluorescence the concentration of a compound in the Apex of a peak will depend on the flow rate. So if you adjust everything (for different ID columns) except the injection amount you will get decreased intensities for larger ID columns. An example can be found in F.J. Yang in Journal of High Resolution Chromatography, 6, 348, 1983 where a mixture of aromatic hydrocarbones are injected in columns of ID 4.6, 2.1, and 1 mm. It is isocratic separation, flow rate at 1000, 200, 50 uL/min respectively, detection cell at 0.5 uL, column length at 15 cm and injection volume of 1 uL of the same injection amount every time.

I hope that this clarifies things...

Kostas

[mardexis]

Kostas,

This is a nice description but why would you carefully scale all parameters but forget to scale the sample volume?

Sorry to rant but I have a drawerful of $1000 2.1 mm columns from round-eyed RA's who saw Franks article or equivalent and were then disappointed that their optimized 4.6 mm assay was even worse.

I'm back with Uwe and Bryan... you will see little or no change in decreasing ID so stick with bigger columns. Again, except for the minor and esoteric effect in capillaries and unless you have only a miniscule amount of sample.

[Kostas Petritis]

Hi mardexis,

Why would you care for the injection loop volume? It was kept into a minimum so it is taken out of any consideration. Especially in the case of gradient chromatography, and assuming that you get good retention for all your compounds, you really do not care about your injection volume. In proteomic experiments we use 150 um Id columns with 10 uL injection loops without any side effects (after column loading the valve is switched to load before the gradient begins).

Capillary chromatography can have advantages when coupled with mass spectrometry and if splittless you can have significant savings in terms of solvent consumption and the associated wastes. Assuming that you have an optimized and robust HPLC system going to lower ID columns (within your confort level) is easy and provides some advantages... while maintaining the chromatographic resolution...

[danko]

Hi Kostas,

According the chromatographic theory, the concentration of a compound in the Apex of a peak is given by:

Cmax= (Q0/Vr) * (N/2pi)^1/2

But the question is: Why larger columns reduce the intensity (= concentration) if the peak width is constant???
So, take a look at the equation. We are talking about a single variable Vr (Q0 is constant, 2pi is constant and finally we designated N to be constant – or is it?).
Now, Vr=tr * F (where tr is retention time and F is the flow rate). We decided that the retention time should be constant; it means that the mobile phase velocity in the column should be constant, which in turn means that wee need to apply higher flow rate (volume per minute) when switching from a smaller to a larger diameter column in order to keep the retention time constant. So, the only thing changing is the volume. And amount per volume spells concentration in my world.
All this means larger column diameter = lower concentration = more dilution = less concentrated bands. But matter does not disappear in nature!!! Wouldn’t the resultant peaks become broader?
The rest will come later as mentioned in my previous post. I’ll await the answers to my questions I put to Uwe.

Best Regards

[Kostas Petritis]

Danko,

I think that several people have replied your question in different ways, you are just not listening...

If you inject 1 mg of one compound in two columns with different ID's, 1 mg will pass through the detector in both cases. The difference is that as these columns operate at different flow rates, and as the UV detector is the same in both cases and reports "concentration" at regular time intervals, it will report smaller values for the larger ID column than for the smaller ID column. The peak won't be any broader as the larger ID column operates at higher flow rates. So if the peak width in the base is 1 min it means that you will have 1 mg of your compound in 200 uL with a 2.1 mm column and 1 mg of your compound in 1000 uL with a 4.6 mm column. But as the UV detector cell has the same dimensions it will report smaller values for the 4.6 mm column than for the 2.1 mm column (because again it reports "concentration").

As you see, matter does not dissapear and everything works as it should... If you still do not understand what is going on, I can only assume that you are arguing for the sake of arguing...

[danko]

Hi Kostas,

The difference is that as these columns operate at different flow rates, and as the UV detector is the same in both cases…….

This is exactly, what I’ve been saying all the time. Please read my earlier posts. But people kept saying that flow rate and detector didn’t matter. That’s why I chose to save these two parameters for later use. So we are coming nearer the facts.

The peak won't be any broader as the larger ID column operates at higher flow rates.

This is also true, but as I pointed out earlier this is what the detector “seesâ€