Peak splitting on Agilent but not Waters HPLC

[fkm]

Does anyone know why peaks would split on Agilent (1200 and 1100) but not Waters 2685. This is an isocratic (0.3 ml/min) chiral separation using 0.02M ammonium acetate:ACN (50:50), column is chiralpak AD-RH 150x4.6mmx5um, 20 ul injection volume, sample diluent is ethanol and UV detection at 220 nm. So far I have tried to increase the length and diameter of precolumn and post column tubing with no success. Reduced injection volume (10 or 5 ul) eliminates the splitting suggesting column overload but still doesn't explain why splitting is present on Agilent system with 20 ul injection but now Waters.

[tkubowicz]

Hello

Waters LC and Agilent 1100/1200 samplers have different design so the injection technique is different.
In Waters you have syringe and in Agilent you have sample loop with metering device.

Are you using any vial with washing solvents for Agilent LC? Or you're running method with standard injection (no washing vial).

Also your sample solvent is stronger than mobile phase which can lead to peak distortion.

Also have you got the same UV parameters? - data rate/slit etc?

Regards

Tomasz Kubowicz

[Andy Alpert]

Are all of the components of your mobile phase in a single bottle or are you generating it online by blending an aqueous salt solution with 100% acetonitrile?

[Consumer Products Guy]

Different designs, and you know what the issue really is: mobile phase 50% aqueous, and sample solvent is ethanol.

And you know what the solution is: inject smaller amount.

At our company, decades ago - before internet - we figured this out on our own, that the injected solvent was functioning as minute amount of mobile phase.

The slow flow rate likely contributes a little as well, typically 1ml/min ballpark flow rate for 4.6mm id columns.

[fkm]

Are all of the components of your mobile phase in a single bottle or are you generating it online by blending an aqueous salt solution with 100% acetonitrile?

Thanks a lot. They are pre-mixed and adjusted to PH7.7.

[Andy Alpert]

It still isn't clear why there's a disparity between the two HPLC systems (hence my question, which involved online mixing, if any), but it isn't surprising that you're getting peak splitting. Consumer Products Guy is dead on target here. Dilute your sample 1:1 with the aqueous component of your mobile phase (making it necessary to inject 2x the volume in order to get the same amount of component in the analysis) and see if that reduces or eliminates the peak splitting. You might find it necessary to dilute 1 vol. of the ethanolic solution with 3 vols. mobile phase; experiment a little in this regard and see which approach works best.

This is a mixing problem in which the disparity in composition between sample solvent and mobile phase results in a difference in viscosity. Result: Streamers from the edges of the sample plug move downstream faster than does the bulk of the sample plug, hence the extra peak. If you reduce the problem but don't eliminate it, then instead of an extra peak you will get a badly skewed peak shape.

[fkm]

Hello

Waters LC and Agilent 1100/1200 samplers have different design so the injection technique is different.
In Waters you have syringe and in Agilent you have sample loop with metering device.

Are you using any vial with washing solvents for Agilent LC? Or you're running method with standard injection (no washing vial).

Also your sample solvent is stronger than mobile phase which can lead to peak distortion.

Also have you got the same UV parameters? - data rate/slit etc?

Regards

Tomasz Kubowicz

Thanks Tomasz, Injection is done with ethanol rinse on Agilent. Yes, ethanol is definitely is an issue. This is an old method with a lot going and needs improvement.
For UV settings, I cannot tell the slit settings on Waters but humming filter is used, sampling rate is 1. Slit width is 2 nm and sampling rate is 2.

[fkm]

Different designs, and you know what the issue really is: mobile phase 50% aqueous, and sample solvent is ethanol.

And you know what the solution is: inject smaller amount.

At our company, decades ago - before internet - we figured this out on our own, that the injected solvent was functioning as minute amount of mobile phase.

The slow flow rate likely contributes a little as well, typically 1ml/min ballpark flow rate for 4.6mm id columns.

Thanks for looking into this. Too much injection volume in a strong solvent, coupled with slow flow rate are the main issues. My quest is to understand why the poor method conditions are still giving almost good data on Waters but not Agilent. May be there is a way Agilent can be configured to give same chromatography as Waters, eliminating the need for method re-validation at least short term.

[Peter Apps]

On the Agilents replace the tubing between the injector and column with a longer length of wider diameter - the extra mixing volume will dilute the ethanol plug slightly before it hits the column and distorts the peaks.

Peter

[Consumer Products Guy]

Maybe there is a way Agilent can be configured to give same chromatography as Waters, eliminating the need for method re-validation at least short term.

USP <627> (I believe that's the number, Chromatography, I'm retired now) lists parameters which can be altered for validated assays for USP products and for which revalidation IS NOT REQUIRED (injection volume, column length, temperature, etc.). Changing to a smaller injection volume is permitted, and listed.

And if this is for a finished product for which your company has in-house validated the assay, FDA ORA 4.5 lists the same parameters and states that revalidation is not required.

It's there, in black and white; I don't know if European Pharmacopeia (etc.) has the same wording, wouldn't surprise me. I do know that my supervisor showed reluctance to accept those, even though it was there in writing, but he was ultra-ultra conservative and timid.

[fkm]

On the Agilents replace the tubing between the injector and column with a longer length of wider diameter - the extra mixing volume will dilute the ethanol plug slightly before it hits the column and distorts the peaks.

Peter

Thanks Peter, I am working on this. Initial attempts did not provide promising results. I was expecting changing from a 105 mm to a 400 mm to provide an improvement but that was not the case. I will be trying the maximum diameter and longest pre-column tubing available.

[James_Ball]

Another thing to be careful of on the Agilent autosampler is the small piece of teflon waste tubing that comes from the switching valve and slides into a slot just above the overflow port where any leaks drain out of the front of the instrument. This tubing allows a few microliters of sample to be pushed to waste which usually just evaporates or if a large amount will drip into the leak discharge tubing. If this tubing does not terminate above the level of the autosampler vial, you can have it form a gravity flow which will suck in a small plug of air at the needle when it is moving the vial out of the way before injection.

I found this out the hard way once when I tried to extend the line to the waste container on the floor, I had horrible chromatograms because it was injecting mostly air.

[DR]

One last, as yet unmentioned possibility.
Did someone swage a SS ferrule on the column connector of the Agilent? If so, I'd bet that they put it on too close to the tip of the tubing, leaving a gap at the column inlet. Bad connections can certainly lead to peak splitting, or at least some broadening. Compare the lengths of tubing after the end of the ferrules on the 2 systems...

[tkubowicz]

Hello

One last, as yet unmentioned possibility.
Did someone swage a SS ferrule on the column connector of the Agilent? If so, I'd bet that they put it on too close to the tip of the tubing, leaving a gap at the column inlet. Bad connections can certainly lead to peak splitting, or at least some broadening. Compare the lengths of tubing after the end of the ferrules on the 2 systems...

You can use non-swaged capillary and use PEEK fingertight fitting. you don't have to worry about a gap because you will adjust the distance every time.

Regards

Tomasz Kubowicz

[James_Ball]

Hello

One last, as yet unmentioned possibility.
Did someone swage a SS ferrule on the column connector of the Agilent? If so, I'd bet that they put it on too close to the tip of the tubing, leaving a gap at the column inlet. Bad connections can certainly lead to peak splitting, or at least some broadening. Compare the lengths of tubing after the end of the ferrules on the 2 systems...

You can use non-swaged capillary and use PEEK fingertight fitting. you don't have to worry about a gap because you will adjust the distance every time.

Regards

Tomasz Kubowicz

For this I prefer the Lite Touch fittings from Upchruch, they work great and can be used with either their PEEK nuts or the standard SS Agilent nuts.