-
- admin
- Site Admin
-
- Posts: 647
- Joined: Wed Aug 11, 2004 12:19 am
By Mark on Tuesday, August 24, 2004 - 05:20 pm:
I would like to pose the following question: What is the effect of flow rate on a chromatographic separation. First let me elaborate (the question is not as dumb as it may seem).
We all know that efficiency (i.e. peak width) is a function of flow rate, as described by the van Deemter equation. This is the kinetic part of the situation. But what about the effect of flow rate on the separation of two peaks: or what the text books define as alpha (the ratio of retention factors). This being the thermodynamic part of the issue.
It just seems, from a common sense perspective, that flow rate has to have some effect on the separation between peaks. Simply because it governs the amount of time that the analytes spend in the column. More time in the column - more opportunity for separation. At least one would think. But I can't find where this is covered in any theoretical discussions.
Any insight is appreciated.
-------------------------------------------------------------------------------------------------------
By A.Mouse on Tuesday, August 24, 2004 - 08:32 pm:
Unless strange things happen, the relative retention or alpha between two peaks does not change. Look at it in the following way: the thermodynamics do not change with time, i.e an equilibrium does not change with time. If the peaks spend more or less time in the column, their partitioning equilibrium between mobile phase and stationary phase does not change.
I have not seen a discussion on that either. I assume that the gurus think that this is "self-evident".
-------------------------------------------------------------------------------------------------------
By Andreas Neumaier on Wednesday, August 25, 2004 - 03:26 am:
Hi Mark,
let me add (or start with?) some thoughts to your theoretical discussion:
With increasing flow rate pressure is increasing.
Increasing pressure means increasing friction, which leads to increased delta temperature at the inner beginning of the column.
Higher temperature means less viscosity which leads to increases peak tailing.
But the increased peak tailing caused through higher pressure can be ignored compared with other factors causing peak tailing.
-------------------------------------------------------------------------------------------------------
By tom jupille on Wednesday, August 25, 2004 - 09:58 am:
A.Mouse basically nailed it: to a very good first approximation, selectivity is thermodynamics, peak broadening is kinetics.
As Andreas pointed out, you *can* see secondary effects because of temperature changes (equilibria can also shift slightly as a function of pressure), but in the vast majority of cases, these are negligible.
The kinetics/thermodynamics observation brings up another point: if you have two distinct species in equilibrium with a slow interconversion rate, you can see some "interesting" effects, ranging from peak broadening, through "chair-shaped" peaks to actual peak splitting. This is most often seen with partially denatured proteins (multiple conformers), although I've also seen it with glucose (the alpha- and beta- anomers). These situations, however, are very much the exception rather than the rule.
-------------------------------------------------------------------------------------------------------
By Mark on Wednesday, August 25, 2004 - 04:00 pm:
I don't know guys. I'm not feeling convinced. I agree with the statement that thermodynamic factors are generally not time-dependent. I also agree that alpha won't change with time (becuase of how it's defined).
But as a separation proceeds the spacing between two peaks defenitely increases (whether we want to think of it in terms of retention time or the actual distance between the peaks). So it seems to me that we are missing something.
Oh wait. I think I just figured it out. N = L/H. So N captures both the kinetic issues (H) as well as the effect of time-on-column: by virtue of column length (L). Which I don't know what to call but a "thermodynamics over time" effect. And since N is contained within the fundamental resolution equation, these effects are all captured by R.
I guess I should not have tried to force a connection between flow rate and separation - in my original question.
But there is an effect of *time-on-column* on separation. Aparently it is properly thought of in terms of column length, not flow rate.
Anyone disagree?
-------------------------------------------------------------------------------------------------------
By RKK on Thursday, August 26, 2004 - 07:32 pm:
Thanks for pointing this out. I never realized that N carried information about more than kinetics!
-------------------------------------------------------------------------------------------------------
By Uwe Neue on Thursday, August 26, 2004 - 09:13 pm:
Under normal circumstances (among other things - isocratic chromatography), the relative retention between peaks, i.e. alpha, does not change with the flow rate. The peak width changes - not only in time units, but also in units relative to the retention time. With other words, the plate count changes. The plate count is low at very high flow, goes through a maximum at some low flow rate, but then becomes lower again at very low flow rates. The underlying effect is covered by the van-Deemter equation or the Knox equation, i.e. hydrodynamics.
-------------------------------------------------------------------------------------------------------
By A.Mouse on Friday, August 27, 2004 - 05:56 pm:
The position of the tip of the peaks relative to each other does not change with a change in flow rate (unless you get temperature effects, as said before).
The width of the peak relative to the retention time changes with flow rate. This is not a function of column length on its own. It depends on the time that the peak spends in the column (=diffusion), the time that it gets mixed with the surrounding mobile phase (=eddy dispersion) and the time it has to get in and out of the particles (=mass transfer). Bottom line: sometimes a shorter column has more plates than a long column.
Maybe that is too much philosophy...
-------------------------------------------------------------------------------------------------------
By Mark on Saturday, August 28, 2004 - 06:35 pm:
Uwe and A. Mouse
To be honest, sometimes it seems you guys assume a post a wrong - and go about correcting it - when you haven't really read it carefully. If you look at my previous posts, I have already agreed that alpha is not a function of flow rate. And I have already spoken of the relationship between flow rate and peak shape (the kinetic effect - van Deemter effect - whatever you want to call it). But I am making the additional statement that plate count also carries information about how separation improves with column length simply because the separation has more time to develop.
I agree with A. Mouse that a shorter column could have a greater N than a longer column, but this would be because it had (for example) smaller particles, and this advantege outweighed the length advantage of the longer column.
Bottom line: N prop. L (and it has nothing! to do with kinetics).
I would like to pose the following question: What is the effect of flow rate on a chromatographic separation. First let me elaborate (the question is not as dumb as it may seem).
We all know that efficiency (i.e. peak width) is a function of flow rate, as described by the van Deemter equation. This is the kinetic part of the situation. But what about the effect of flow rate on the separation of two peaks: or what the text books define as alpha (the ratio of retention factors). This being the thermodynamic part of the issue.
It just seems, from a common sense perspective, that flow rate has to have some effect on the separation between peaks. Simply because it governs the amount of time that the analytes spend in the column. More time in the column - more opportunity for separation. At least one would think. But I can't find where this is covered in any theoretical discussions.
Any insight is appreciated.
-------------------------------------------------------------------------------------------------------
By A.Mouse on Tuesday, August 24, 2004 - 08:32 pm:
Unless strange things happen, the relative retention or alpha between two peaks does not change. Look at it in the following way: the thermodynamics do not change with time, i.e an equilibrium does not change with time. If the peaks spend more or less time in the column, their partitioning equilibrium between mobile phase and stationary phase does not change.
I have not seen a discussion on that either. I assume that the gurus think that this is "self-evident".
-------------------------------------------------------------------------------------------------------
By Andreas Neumaier on Wednesday, August 25, 2004 - 03:26 am:
Hi Mark,
let me add (or start with?) some thoughts to your theoretical discussion:
With increasing flow rate pressure is increasing.
Increasing pressure means increasing friction, which leads to increased delta temperature at the inner beginning of the column.
Higher temperature means less viscosity which leads to increases peak tailing.
But the increased peak tailing caused through higher pressure can be ignored compared with other factors causing peak tailing.
-------------------------------------------------------------------------------------------------------
By tom jupille on Wednesday, August 25, 2004 - 09:58 am:
A.Mouse basically nailed it: to a very good first approximation, selectivity is thermodynamics, peak broadening is kinetics.
As Andreas pointed out, you *can* see secondary effects because of temperature changes (equilibria can also shift slightly as a function of pressure), but in the vast majority of cases, these are negligible.
The kinetics/thermodynamics observation brings up another point: if you have two distinct species in equilibrium with a slow interconversion rate, you can see some "interesting" effects, ranging from peak broadening, through "chair-shaped" peaks to actual peak splitting. This is most often seen with partially denatured proteins (multiple conformers), although I've also seen it with glucose (the alpha- and beta- anomers). These situations, however, are very much the exception rather than the rule.
-------------------------------------------------------------------------------------------------------
By Mark on Wednesday, August 25, 2004 - 04:00 pm:
I don't know guys. I'm not feeling convinced. I agree with the statement that thermodynamic factors are generally not time-dependent. I also agree that alpha won't change with time (becuase of how it's defined).
But as a separation proceeds the spacing between two peaks defenitely increases (whether we want to think of it in terms of retention time or the actual distance between the peaks). So it seems to me that we are missing something.
Oh wait. I think I just figured it out. N = L/H. So N captures both the kinetic issues (H) as well as the effect of time-on-column: by virtue of column length (L). Which I don't know what to call but a "thermodynamics over time" effect. And since N is contained within the fundamental resolution equation, these effects are all captured by R.
I guess I should not have tried to force a connection between flow rate and separation - in my original question.
But there is an effect of *time-on-column* on separation. Aparently it is properly thought of in terms of column length, not flow rate.
Anyone disagree?
-------------------------------------------------------------------------------------------------------
By RKK on Thursday, August 26, 2004 - 07:32 pm:
Thanks for pointing this out. I never realized that N carried information about more than kinetics!
-------------------------------------------------------------------------------------------------------
By Uwe Neue on Thursday, August 26, 2004 - 09:13 pm:
Under normal circumstances (among other things - isocratic chromatography), the relative retention between peaks, i.e. alpha, does not change with the flow rate. The peak width changes - not only in time units, but also in units relative to the retention time. With other words, the plate count changes. The plate count is low at very high flow, goes through a maximum at some low flow rate, but then becomes lower again at very low flow rates. The underlying effect is covered by the van-Deemter equation or the Knox equation, i.e. hydrodynamics.
-------------------------------------------------------------------------------------------------------
By A.Mouse on Friday, August 27, 2004 - 05:56 pm:
The position of the tip of the peaks relative to each other does not change with a change in flow rate (unless you get temperature effects, as said before).
The width of the peak relative to the retention time changes with flow rate. This is not a function of column length on its own. It depends on the time that the peak spends in the column (=diffusion), the time that it gets mixed with the surrounding mobile phase (=eddy dispersion) and the time it has to get in and out of the particles (=mass transfer). Bottom line: sometimes a shorter column has more plates than a long column.
Maybe that is too much philosophy...
-------------------------------------------------------------------------------------------------------
By Mark on Saturday, August 28, 2004 - 06:35 pm:
Uwe and A. Mouse
To be honest, sometimes it seems you guys assume a post a wrong - and go about correcting it - when you haven't really read it carefully. If you look at my previous posts, I have already agreed that alpha is not a function of flow rate. And I have already spoken of the relationship between flow rate and peak shape (the kinetic effect - van Deemter effect - whatever you want to call it). But I am making the additional statement that plate count also carries information about how separation improves with column length simply because the separation has more time to develop.
I agree with A. Mouse that a shorter column could have a greater N than a longer column, but this would be because it had (for example) smaller particles, and this advantege outweighed the length advantage of the longer column.
Bottom line: N prop. L (and it has nothing! to do with kinetics).
