Reproducibility Problems with Menthol

[tcjones10]

I am running menthol in MeOH at about 3.5mg/ml on a TR-Wax 0.32mm x 30m x 0.5um column on a 6890. I am using a 5:1 split at a column flow of 10ml/min. I am using a restek cyclosplit liner at 300 degrees. I am getting RSD's of nearly 10% in replicate std injections. I am stuck with the Wax column, because it is the only column found that can seperate some other interferences. Any suggestions to improve my RSD?

[chromatographer1]

How much are you injecting?

With only a 5:1 split you must be looking for very small amounts of impurities.

I would suggest that you use a direct injection if you want tight RSD values.

The most likely cause of the problem you are experiencing right now is from the installation of the column tip in the injector. Even a small displacement from optimum can cause the high RSD values.

There are kits to change a split injector to a on-column or a direct injection system. If you are looking for trace amounts a large sample injection system might be an addition you might find useful.

You are right, 10% is too large, 2% would be acceptable.

best wishes,

Rod

[tcjones10]

I am injecting 1 uL.

The split was a typo should have read 15:1.

I have tried lowering the split and I have tried in splitless, this only results in fronting in the menthol peak, I can remedy that by increasing flow, but I have two other analytes that will co-elute with the increase. This is an assay for menthol and two preservatives in a matrix that has tons of extractables. I will check the column depth in the injector, and tighten the furrels (high injector temp my have them leaking) But other than that I am at wits end.

[chromatographer1]

Methanol has almost 4 times the expansion volume as does hexane for example. Perhaps you are exceeding the volume capacity of your injection liner. This will give you large RSDs.

I would first check this by manually injecting a large volume of methanol, enough to flood your inlet (5µL) and wash out any backflashed menthol from your injector. Watch and see if any menthol elutes after the methanol baseline recedes. Of course, make sure your syringe is clean before you do this.

The next thing I would try would be to reduce my injection volume to 0.2µL and see if the RSDs improve. Also if you are performing manual injections I would require a Chaney adapter. This is another indication that you may have been injecting too large a sample.

Check that you are getting an exact amount in your needle for each injection, ie that your syringe is actually taking from your sample vial the same amount each time. Worn needle plungers can be a problem.

If reducing the size of your sample injection did not improve your RSD values, and your syringe is picking up a reproducible amount of sampke, you might reduce the rate of your injection if you are using an autosampler and see if this improves the performance.

Let us know how things work out.

best wishes,

Rod

Oh yeah, don't forget the obvious. Install a new septa in your injection port, just in case it may have been leaking.

[AICMM]

Can I suggest one thing that would be an easy starting point to look at your problem. Instead of 1:15 split, dilute your sample 1:5 or 1:10 in hexane and try it again splitless. I think your problem is flash (especially with cycloliner.) According to the math you should end up about the same sensitivity but I think you will have much better reproducability.

I have a customer who does menthol all the time but the extraction solvent is hexane rather than methanol. Method is pretty rugged.

Best regards.

[Peter Apps]

Presuming that you have an Agilent autoinjector on the 6890 you need to change the injector parameters - the default fast cold needle settings are no good for solvents like methanol that have a high boiling point for their molecular weight and generate a lot of vapour. Try a pre-injection dwell.

A small plug of deactivated glass wool at the injection point might give more repeatable solvent vaporization.

Also, I am puzzled by the 10 ml/min column flow, which is WAY too high, or is this another typo ?

Peter

[Tomasz]

I suppose you can not dilute because you have preservatives at low level.
Moreover, methanol is not soluble with hexanel in 1:1o (or 1:5) ratio as AICMM suggested.
Adding internal standard should help a little.

[AICMM]

Tomasz, TCJones10,

I honestly did not even think of the methanol not dissolving in the hexane, sorry. I was going to propose using that to your advantage and extracting 1:1 hexane and methanol but menthol is noted to be very soluble in alcohols so that probably would not work. (The method I am aware of extracts menthol from water with hexane but no methanol is present...)

Thus, I am forced to suggest diluting in a solvent that methanol is soluble in and which has a much better flash volume like MTBE or ethyl acetate. The idea here is to determine if it is the flash volume is the cause of the poor reproducability (which I think it is...)

You might also go away from the cyclosplitter and use a more conventional liner with more volume.

Best regards.