Chromatography Forum

Does 3.5 micron prep column has the same column efficiency as analytical one if everything keeps constant?

In principle, yes. In practice, it is more difficult to pack small particles in wide columns.

Thanks Tom. so it means that even if we pack prep column with small particles, then it might comprise column efficiency.

In principle, yes. In practice, it is more difficult to pack small particles in wide columns.

That depends on whether you are interested in efficiency for "analytical-scale" (non-overload) or "prep-scale" (overloaded) separations.

For analytical-scale separations, packing your own 3.5-micron material in *any* size column is challenging. It generally takes a lot of experimentation to optimize the combination of solvent, concentration of packing, temperature, and time/flow/pressure profile. That's the reason that very few people pack their own columns; a manufacturer can amortize that effort over hundreds of columns. I suspect the problems are more severe for wider columns because it's harder to maintain a consistent flow profile with the larger cross-section.

For prep-scale separations, the peak width depends much less on particle size than it does on the degree of overload. What that means in practice is that a heavily-overloaded separation will look about the same whether it's run on 3 micron or 10 micron particles. Which brings up the practical question: why go to the bother of packing a prep column with 3 micron particles when you can do almost as well with 10 micron particles which are easier to pack and run at a lower back pressure?

If I can get up on my soapbox and generalize a bit: equations are a useful tool to help us understand what's going on, but they are merely simplified approximations ("The map is not the territory"*). If you want an analogy, think about the Henderson-Haselbalch equation that we use to predict pH. As with most thermodynamic equations, it assumes that activity = concentration, which is strictly true only at infinite dilution. It's close enough to true to be useful at "reasonable" concentrations, but it can be wildly incorrect at high concentrations. The van Deemter and Knox equations relating peak width the particle size, flow rate, and diffusivity likewise assume linear distribution isotherms, which is strictly true only at infinite dilution. They are close enough at "analytical" levels to be useful, but become incorrect at high sample load.

*http://en.wikipedia.org/wiki/Map%E2%80%93territory_relation

Really thanks for your explanation! although I used around 21mm diameter column, I always run it under linear isothermal condition or slightly overload (smaller than 20% of peak width owing to mass overload).
I guess my problem will only be solved when I actually get the real column and run at the same condition for 4.6mm and 21mm ID column.

That depends on whether you are interested in efficiency for "analytical-scale" (non-overload) or "prep-scale" (overloaded) separations.

For analytical-scale separations, packing your own 3.5-micron material in *any* size column is challenging. It generally takes a lot of experimentation to optimize the combination of solvent, concentration of packing, temperature, and time/flow/pressure profile. That's the reason that very few people pack their own columns; a manufacturer can amortize that effort over hundreds of columns. I suspect the problems are more severe for wider columns because it's harder to maintain a consistent flow profile with the larger cross-section.

For prep-scale separations, the peak width depends much less on particle size than it does on the degree of overload. What that means in practice is that a heavily-overloaded separation will look about the same whether it's run on 3 micron or 10 micron particles. Which brings up the practical question: why go to the bother of packing a prep column with 3 micron particles when you can do almost as well with 10 micron particles which are easier to pack and run at a lower back pressure?

If I can get up on my soapbox and generalize a bit: equations are a useful tool to help us understand what's going on, but they are merely simplified approximations ("The map is not the territory"*). If you want an analogy, think about the Henderson-Haselbalch equation that we use to predict pH. As with most thermodynamic equations, it assumes that activity = concentration, which is strictly true only at infinite dilution. It's close enough to true to be useful at "reasonable" concentrations, but it can be wildly incorrect at high concentrations. The van Deemter and Knox equations relating peak width the particle size, flow rate, and diffusivity likewise assume linear distribution isotherms, which is strictly true only at infinite dilution. They are close enough at "analytical" levels to be useful, but become incorrect at high sample load.

*http://en.wikipedia.org/wiki/Map%E2%80%93territory_relation

Wow! Packing 3.5-micron material in a 21mm column is waaay outside the scale of anything I've ever tried. It's is going to be a huge challenge to get an even distribution of the bed with that cross-section.

for 5 micron, they did a good job. I do not see it is a big jump from 5micron to 3.5 micron. Might be I am totally wrong.

Wow! Packing 3.5-micron material in a 21mm column is waaay outside the scale of anything I've ever tried. It's is going to be a huge challenge to get an even distribution of the bed with that cross-section.

i wonder how heat from the friction with the phase will affect the chromatography in a 21 ID column.
and why not go for a 5u prep column and simply change the loading. it is also cheaper for sure.
i wonder also if the prep system will be capable of handing the pressure or will you go for less flow rate to compensate?

the reason why I want choose 3u column is to able to increase peak capacity production rate. As you know, in 2dlc, second dimensional separation should be completed in a relative short time period.
For pressure compensation, I will choose short column(50mm) so pressure will not be so high at its optimum flow rate.

i wonder how heat from the friction with the phase will affect the chromatography in a 21 ID column.
and why not go for a 5u prep column and simply change the loading. it is also cheaper for sure.
i wonder also if the prep system will be capable of handing the pressure or will you go for less flow rate to compensate?