GC-MS two peaks for pure compound

[egy]

I have a pure compound as determined by NMR. The compound gives two very close peaks by GC-MS with a peak-to-peak separation of 0.25 to 0.5 min, depending on the concentration. The MS suggests that the two peaks are identical. A comment for a previous post from Sept 22, 2008, "GC/MS identifies 2 peaks of the same compound ?" mentioned that depending on the way the sample of injected, there could be peak splitting in the GC. Can anyone expand on this comment? How would you overcome this type of peak splitting? And, would you integrate both peaks for purposes of making a standard curve for quantitation?
Thanks.

[reeftom12]

The most simple observation is the possibility that you have a stereoisomer that is not easily distinguished by H1 NMR. The GC was able to seperate the 2 isomers.

[egy]

The most simple observation is the possibility that you have a stereoisomer that is not easily distinguished by H1 NMR. The GC was able to seperate the 2 isomers.

Thank you for the reply. However, it is definitely not stereoisomers, since the molecule has no stereocenters or potential rotomers or cis/trans olefins. It is a very simple bis-aryl ether. However, I see two peaks by GC. Some concentrations worse than others. I used isopropanol as the solvent to do serial dilutions, which gave two peaks, and then I tried acetonitrile, which gave two peaks sometimes, but was much better. From my search of the forums, it seems someone else noted splitting of peaks (see original post for reference). I just don't know how to explain this.

[tom jupille]

It's been decades since I did any GC, but if you'll tolerate naive speculation:
- temporary formation of an adduct? a dimer maybe? in the injection port. Have you tried changing the injection port temperature to see if that has any effect?
- quenching of the injection port temperature from a too-large injection (IPA has a higher boiling point than ACN). Try changing the injection volume and see if that has any effect.

[Don_Hilton]

What GC conditions are you using? (Temperatures, column, liner, flow rate, mode of injection, etc.)

Methanol is famous for splitting up peaks when injected onto a non-polar column at too low a temperature - the explanation being that methanol does not wet the column and disperses into droplets along the column rather than forming a continuous band. Both isopropanol and acetonitrile have a good degree of polarity, particularly if wet. But, let's know what you are doing, otherwise some of us can wander off into numerous speculations...

And if you can tell what the compund is, someone familiar with the compund or related compunds may be able to give some advice based on experience.

[egy]

Thank you for the information. I was able to solve the problem simply by taking the compound up in ethanol and then diluting with water to give a mostly aqueous solution for injection. Sorry I cannot disclose the compound I am working with, but it is a bis-aryl ether, which is not very soluble in water, but seemed to give stable solutions at low concentrations. As a matter of reference, the conditions I used were 175 degrees to 210 degrees over a 6 min time interval, with a column flow rate of 1.03 mL/min, on a 30 m Restek RTX-XLB capillary column (a proprietary non-polar phase), with an autoinjector, with injection temperature set at 210 degrees C, 5 microL injection, and the compound came out at 4.5 min.
Thanks for your help.

[Don_Hilton]

In the event you need to explore the chromatography again: If you can avoid water in the GC inlet, do so. Less polar solovents give lower expansion volumes and typically have no problems in interacting with the GC column. And for a typical GC inlet system (such as split/splitless), a 5 microliter injection is way, way too much sample, particularly with water in the solvent mix. This size of injection tends to give expansion of the sample out of the inlet and into the gas supply tubing - which can result in contaminiation in subsequent injections. If you are usign a PTV inlet, you can use this size of sample.

[Peter Apps]

Thank you for the information. I was able to solve the problem simply by taking the compound up in ethanol and then diluting with water to give a mostly aqueous solution for injection. Sorry I cannot disclose the compound I am working with, but it is a bis-aryl ether, which is not very soluble in water, but seemed to give stable solutions at low concentrations. As a matter of reference, the conditions I used were 175 degrees to 210 degrees over a 6 min time interval, with a column flow rate of 1.03 mL/min, on a 30 m Restek RTX-XLB capillary column (a proprietary non-polar phase), with an autoinjector, with injection temperature set at 210 degrees C, 5 microL injection, and the compound came out at 4.5 min.
Thanks for your help.

You originally asked about integration and calibration, which suggests that you need quantitative results. The conditions that you give are not likely to yield good repeatability. In particular a 5 ul aqueous injection, even if split, is likely to cause all sorts of trouble. Am I right to suspect that separating your target compound from potential interferences is not a problem ? If so there are probably better ways of solving the peak splitting. What else did you try ?

Peter

[egy]

Yes, you are right, I want to quantify. My experiment is measuring the levels of a bis-aryl ether after exposure to liver microsomes. The test solution is mostly water, which I ran and was able to see a single peak (no peak splitting). Therefore, I made my standard curve using a mostly aqueous solvent. I originally tried to make my standard solutions using IPA or ACN, but saw considerable peak splitting. I do not have a problem with separating the compound from interferences. However, I am working at fairly low concentrations, so I used 5 microL injections, and stayed with 5 microL to be consistent throughout the standard curve. From your comment, I assume that my results and standard curve will not be very reproducible. As you can tell, I am very much an amateur. I appreciate your feedback.

[Peter Apps]

What concentration is "fairly low" ? What solvents is your ether soluble in ?, I assume anything non-polar. What clean-up treatments have your aqueous samples been through between microsomes and injection ?

Peter

[egy]

On the low end I am trying to measure approximately 100 nM solution. The ether is soluble in most any organic solvent such as DMSO, diethyl ether, chloroform, methylene chloride, methanol, acetonitrile, etc. The clean-up is fairly simple, an 100 microL aliquot is removed from the medium and added to 100 microL of ice cold acetonitrile. The solution is then centrifuged and the supernatant was used directly for GC-MS analysis. I said earlier it was mostly aqueous, but I guess it is actually 50% aqueous.

[Don_Hilton]

A couple of things to watch out for here:

If your compund is poorly soluable in water, it will tend to parition well into organic phases, cell debris, or onto surfaces - particulary plastics. The way to check this is to spike medium mix gently and immediately sample to see if you recover the right quantity of material. Do this at several concentrations, particularly at low levels, to see if you have significant losses close to the lower levels you will be measuring.

If your compund makes it into the medium, it may be better to shake the 100 uL medium with something like 100 uL DCM and inject from the DCM layer (bottom layer). Yo0u may want to do multiple extractions with DCM, in which case you will want to remove the DCM layer to a separate vial, reextract, and move the DCM layer, combining it with the first DCM removed from the sample. You can easily evaporate the DCM to obtain a more concentrated sample -- and you can keep the injection volume to 1 microliter.

[Peter Apps]

On the low end I am trying to measure approximately 100 nM solution. The ether is soluble in most any organic solvent such as DMSO, diethyl ether, chloroform, methylene chloride, methanol, acetonitrile, etc. The clean-up is fairly simple, an 100 microL aliquot is removed from the medium and added to 100 microL of ice cold acetonitrile. The solution is then centrifuged and the supernatant was used directly for GC-MS analysis. I said earlier it was mostly aqueous, but I guess it is actually 50% aqueous.

Don's advice is good. Another thing that you can try is to add salt to the AcN/water mix - the AcN will separate and although AcN is not commonly used as a solvent in GC it will be much less susceptible to strange inlet effects than a 1:1 mix with water.

A back of envelope calculation gives you about 15 ng per ul in what you inject to the GC - I'm guessing a MW of 300 for your analyte. If your MS is working right you could do a 1 ul injection with a 10:1 split and still have a very respectable signal:noise ratio at the lower end of your calibration.

Peter

[egy]

Thanks very much for the advice in such details. I will give these techniques a try. The education is much appreciated!

[Camisotro]

Don's advice is good. Another thing that you can try is to add salt to the AcN/water mix - the AcN will separate and although AcN is not commonly used as a solvent in GC it will be much less susceptible to strange inlet effects than a 1:1 mix with water.

A back of envelope calculation gives you about 15 ng per ul in what you inject to the GC - I'm guessing a MW of 300 for your analyte. If your MS is working right you could do a 1 ul injection with a 10:1 split and still have a very respectable signal:noise ratio at the lower end of your calibration.

Peter

Right, the salting out is the basis of the QuEChERS extraction method that I'll be working with in my lab for pesticides - MgSO4 and NaCl are shaken with the mix. Apparently salting out also works with isopropanol. The ACN layer after salting out and centrifuging is approximately 16% water, which can be further diminished by a second MgSO4 treatment if desired.