Can you change flow rate during a run?

[DoryFish]

This is quite a strange question, so I'll ask it first then explain why:

Can I change the flow rate during an analysis to speed up the elution of slower peaks? Pumps nowdays are capable of doing this, column packing more robust and "re-equilibration" time would be so much faster than with gradient elution. However I've googled and can't find examples of it being done. Am I missing something?

My question is part curiosity but also I've been struggling to develop a method to analyse a sample with two components of very different hydrophobicity. On a short C8 column, one component elutes close to the solvent front and the second at around 10 mins. That's with 95% ACN so it doesn't give me much to play with in terms of increaseing organic with a gradient. pKa of the second compound is very high, so no luck there.

I've tried different columns and am happy to keep working, but always on my mind is the thought of ramping up the flow rate after the first peak has eluted. I'm in a research lab so not too worried about loss of column life (within reason of course). Wondered what people thought of the idea?

Thanks

[Rndirk]

I think you need to worry about 2 things here:

- Can your system/column handle the pressure bump? Does your software allow you to set a flow ramping rate?
- Can your detection handle it? For example, if you perform MS with ESI, you impact the stability of the spray with increasing the flow during the run.

I would say try it out. An alternative would be increasing the column oven temperature. For more detailed suggestions we need some details: analytes, matrix, detection,..

[tkubowicz]

Hello

My question is part curiosity but also I've been struggling to develop a method to analyse a sample with two components of very different hydrophobicity. On a short C8 column, one component elutes close to the solvent front and the second at around 10 mins. That's with 95% ACN so it doesn't give me much to play with in terms of increaseing organic with a gradient. pKa of the second compound is very high, so no luck there.

I've tried different columns and am happy to keep working, but always on my mind is the thought of ramping up the flow rate after the first peak has eluted. I'm in a research lab so not too worried about loss of column life (within reason of course). Wondered what people thought of the idea?

Thanks

You can change flowrate during run if your LC can handle it. I'd consider 3 things:

- Is it worth to change flow if equlibration time will be long
- Isn't better to increase organic solvent content to flush everything out from column
- different column need more time to equlibrate

In my opinion if you run with 95% ACN method is poorly optimized. I'd play with different buffers and check "k"vs. "pH"

Regards

Tomasz Kubowicz

[lmh]

Yes, but if you can run the flow faster to get the second peak off, why would you run it slower for the first peak? Yes, running faster throughout the whole run means that your first peak comes off embarrassingly early, but if it's already nearly in the injection peak anyway, you may find that everything just gets narrower, with no loss of resolution. But maybe your detector can't go fast enough to cope with the narrowness.
Personally I don't much like barely-retained peaks, but it's up to you whether you feel you've got enough retention for the job.

[KM-USA]

I have increased the flow during a run, isocratic with RI detector where I had low pressure and after my peak of interest eluted, so I could get unwanted stuff off the column sooner for next injection.

[DoryFish]

Thanks for your replies. In this instance, I can easily increase flow rate without any back pressure problems, also the time window between the two peaks is sufficient for the baseline to settle, re-equilibration is very quick.

I agree my separation isn't yet optimized. At this stage, my question is more about the concept of changing the flow rate during a run. Why is changing the mobile phase viewed as a useful tool for achieving separation but manipulating flow rate seen as a distateful option? With modern instrumentation (and columns) we can technically do both. In fact I have done just this and my two peaks are well resolved but obviously this is a chromatographic no-no. I'm curious why this is so?

Is it simply a case of practice being out of sync with what modern technology is capable of, or is there a fundamental reason why it's to be avoided?

[Gerhard Kratz]

For sure you can run a flow gradient, or temperature gradient. But you need to monitor it carefully and implement it in your validation. In the 80s were some publications about that.

[Rndirk]

Thanks for your replies. In this instance, I can easily increase flow rate without any back pressure problems, also the time window between the two peaks is sufficient for the baseline to settle, re-equilibration is very quick.

I agree my separation isn't yet optimized. At this stage, my question is more about the concept of changing the flow rate during a run. Why is changing the mobile phase viewed as a useful tool for achieving separation but manipulating flow rate seen as a distateful option? With modern instrumentation (and columns) we can technically do both. In fact I have done just this and my two peaks are well resolved but obviously this is a chromatographic no-no. I'm curious why this is so?

Is it simply a case of practice being out of sync with what modern technology is capable of, or is there a fundamental reason why it's to be avoided?

I wouldn't say it's a no-no but between all the things you can play with in LC, it's more of a last resort; depending on the stability your detector needs. Running an RP method with 95% organic to start with and 2 peaks that are separated that much is not a common situation, like you said it's poorly optimized.

[HPLCaddict]

Why is changing the mobile phase viewed as a useful tool for achieving separation but manipulating flow rate seen as a distateful option? With modern instrumentation (and columns) we can technically do both. In fact I have done just this and my two peaks are well resolved but obviously this is a chromatographic no-no. I'm curious why this is so?

Most HPLC applications today are still run using UV detectors, which are susceptible to flow-rate changes. Meaning peak areas will change if you change flow-rate! Double the flow-rate and the peaks will be half as big. This can have serious implications on quantitation, especially if you use a flow-rate gradient with peaks eluting during that gradient - peak areas will depend upon how accurately and reproducible the pump can generate that flow-rate gradient.
In your case, with two widely separated peaks, this is less of a concern - you just may ramp up the flow between the two peaks, so you have a well defined flow-rate for both peaks. But still, quantitation will be different for both.

I myself would rather go for a classic gradient separation instead of flow-rate change. Call me bourgeois...

[danko]

@ DoryFish,

If you elute with 95% ACN it is obvious to me that you have chosen a wrong stationary phase (column).

Flow rate gradients are not practical and are typically avoided because change of flow rate changes the separation/resolution.
Besides, methods like that are nort quite robust in terms of reproducability.

What is the K value for the first eluting peak BTW?

Best Regards

[lmh]

I'd reiterate, the main reason for not doing it is merely that it's generally pointless.

There is an optimum flow-rate for a column (van Deemter plots and all that). Ideally, for best resolution, you would run your column at the minimum of the van Deemter plot. Often this isn't achievable because the minimum happens at too high a back-pressure, or (opposite situation) you want a shorter method and you're prepared to sacrifice some resolution by running faster.

You can't improve separation by changing the flow-rate (unless you're working at a really silly flow-rate, in which case why are you working at that flow rate at all?). Flow-rate isn't the same as concentration. You change the concentration of the solvent in order to change the degree to which the analyte is retained, and you need to do this if you're looking at two wildly-different analytes (as you are! One is barely retained, so you might not want to use 95% ACN at that stage). Changing the flow rate has two effects: it stretches/compresses the time-axis, and it gives more/less time for diffusion effects in the column, which means it influences peak-width. That's all it does.

If you are running at less than the optimal flow-rate, then speeding up is a good idea if you can. If you are running at the optimal flow-rate, then slowing down is always a bad idea (unless you are obliged to for non-chromatographic reasons, like a very slow detector). If you have enormous separation between peaks, then slowing down is pointless. If you're worried about the separation between your injection peak and your first analyte peak, then changing solvent composition will give you better retention, while, if you're working at a relatively flat bit of the van Deemter curve, or you're already at the optimum, slowing down will either not affect, or will harm, resolution (even if it increases the distance between the peaks).

There are useful applications for increasing flow, for example speeding up washes and re-equilibrations.

The issue of it affecting peak area is only relevant if you're trying to compare peaks at one retention time with peaks at another. If you're calibrating each peak against a standard, or you're comparing the same peak between multiple samples, it will have eluted under the same flow conditions.

You wrote about changing flow as a chromatographic no-no. There are a handful of situations where it really is a no-no. For example, to run a method fast until peak 1 has eluted, slow the flow, and then speed it up again just before peak 2 elutes, is definitely bad practice, to the point of fraud. It makes it look as though the selectivity of the method is much higher than it really is.

[DoryFish]

Thanks everyone for your comments and time taken to reply. I've very much enjoyed reading them all. For some time I've been curious about the concept and when google didn't come up with any hits then I wondered about the "why not". Thanks again

[tjbr]

From a validation standpoint keeping the flow rate constant is preferable, modifying it mid injection will complicate an already complicated DOE for robustness, a multi step gradient would be preferable. Failing that using a more powerful shorter column or UPLC would achieve faster run times.

[JI2002]

Change of flow rate can potentially cause selectivity change in gradient RPLC, but it does not apply here, so this can be used to shorten the run time. However, is it possible to do a step gradient to 100% ACN after the 1st peak elutes? If the second peak still shows undesirable long retention time, is it possible to try a stronger organic solvent, like THF or IPA?