separation of two isomers

[mueller]

Can anyone of you guys help me out?
I have the ester of a carbonic acid and I'm not able to separate it from another isomer (with LC/MS)
I tested a lot of more or less "normal" silica C18 columns. I use MeOH and Water because they are the best.
I need a straight tip for a good column, because I guess that is the key for my problem - any suggestions?
Thanks

[Uwe Neue]

Please describe you LC method first, and why you prefer MeOH over MeCN. It could for example be that your gradient is too steep...

[mueller]

the problem is, that I couldn´t generate a good ion (fragment) with Acetonitrile/ water - with MEOH/ water the intensity is much (!) greater. I also tried it with MeCn, intensity was poor and no separation could become obvious (same result with MeOH)
Method: Doesn´t matter if gradient or isocratic flow - the eluation time of the two isomers is more or less always the same.

[tom jupille]

The word "isomers" encompasses a lot of things. I've argued (while being too lazy to dig up proof ) that bonded-phase columns inherently give up selectivity for isomers compared to adsorbents because the bonded phase is "squishy" (how's that for a technical term?) and can to a certain extent deform to accommodate different analyte geometries, whereas an adsorbent surface is "rigid".

That argument suggests something like porous graphite: you can use it with MeOH water, it provides reversed-phase retention behavior, and as a rigid surface it has been shown to provide unique isomer selectivity (most pronounced for analytes that can interact with pi-electrons). Another possibility is to move to a shorter-chain bonded phase column (less squishy, proportionally more silanol accessibility).

[Uwe Neue]

In the absense of good information...

Try a phenyl column or a column with an embedded polar group (XBridge Phenyl, XBridge RP18).

Also, significant (!!!) differences in ligand density can influence the selectivity of closely related analytes. Try Atlantis T3 with less than half the ligand density of a standard C18.

[bhuvfe]

Maybe normal phase using silica type-C? see http://www.microsolvtech.com/hplc/silica-c.asp

[Alex Buske]

What kind of isomers? If the carbonic acids have different chain lengths, try a C8.

Alex

[mueller]

Thanks for all your advices:

Well, I have two constitutional isomers - the only difference is -CH2-CH2- and -CHCH3- .
Waters xBridge didn´t help. C8 looks better, but a real separation couldn´t be observed.

Has anybody of you experience with THF as MP or will cyclodextrine help???

Other advises are welcome too...

[Uwe Neue]

I had suggested XBridge RP18, which is the packing with the embedded polar group, not just an XBridge packing.

To go from one random C18 to another one, or from a C18 to a C8 gives the LOWEST amount of selectivity change, no matter what the problem is. The only exception specifically for your problem at hand is to change the ligand density drastically.

I can send you a few publications on selectivity of different packings, if you want.

[Peter Apps]

Who wants to take a bet that the mobile phase is so strong that everything is coming out at the dead time It wouldn't be the first time.

Peter

[HW Mueller]

Branched chain not separable from straight chain? Is this part of a huge substituent or did Peter hit on the answer?

[Bryan Evans]

What is the value of k' (or rough estimate)?
Does the molecule have any ionizable functional groups?
Large aromatic substituents? Is peak shape ok?