Help - unseparable compounds

[Fee]

Hi!

I am actually working on the isolation of 2 natural product of about 650 of molecular weight. I made an alcoholic extract that I then fractionated by liquid-liquid extraction with 1) hexane, 2) Ethyl acetate 3) CH2Cl2 4) BuOH.

I passed the CH2Cl2 extract on a silica gel column (with increasing polarity mobile phase) and analyzed every fraction by GC-MS. I found 2 major compounds in my fraction that I choose to isolate. The problem is that every compounds appeared in every single fraction (Fraction number 1 to 46!!!) isolated from the silica gel column.

Then I tried several time to isolate on of these two compound from each other using Sephadex LH-20 and even semi-prep HPLC. But the probleme was the same. When I was analyzing my fraction in GC, both compounds were present in every single isolated fractions!

Using HPLC did not solved the problem : while I was seeing separated peaks on the screen, every fraction took from the HPLC was containing both compound.

Did you ever see this problem ? I have no more idea of what I can do to achieve this work.

Thank you very much,

Fee (still Ph.D. student!)

[Uwe Neue]

If I understand you correctly, you see two separated compounds on the display screen showing you the chromatogram as it is seen by the detector. You collect the fractions, but when you analyze the fractions, most contain both compounds.

I suspect that the problem is the widening of the peaks in the tubing that connects your UV detector to the fraction collector. Since you are collecting a large number of fractions, it appears that your fraction collector is designed to work with HPLC columns. If this is the case, then the tubing connecting the detector to the fraction collector is the problem.

[Fee]

Thank you for the tip, but I actually am the fraction collector (I am not using an automatic fraction collector).

More clearly, the situation is that in RP-HPLC with UV detector (MeOH and water as mobile phases) I see a lots of peaks on the computer screen. I collected a lots of them (about 10 fractions equal to 10 different peaks). But when I am analyzing thoses fractions (that eluted with different retention times!!) with GC-MS, ALL fractions contained the two compounds I am trying to isolate.



It is actually the same situation that wath happened while I was using sephadex LH-20 or silica gel : all others compounds are separing, all but the 2 I am looking for!! I have already looked for contaminations in my solvants but that was not the solution.

Hope you will have some tips for me

Thank you,

Fee

[Uwe Neue]

Here is something to think about: the outlet tubing of your detector has some volume. In many detectors, the outlet tubing has a larger diameter then the inlet tubing. So if you collect peaks, you want to keep the volume of the tubing as small as possible, and you want to know the time difference between the time that the peak starts and ends on your display and the time that this event would occur for the vial in your hand. You can calculate / estimate this delay, or you can do an experiment (without the column) by injecting a colored band that you can see.

Note that all of this means that the bandspreading in the post-column tubing is not enormous!

Oh, here is another thought: Are you by any chance using some length of teflon or other plastic tubing for the fraction collection?

[Alex Buske]

Hi,

First jsut a commentz to your separation process: you extract with ethanol (?) and then you do Liquid/liquid partition Water / hexane, water/ethyl acetate, water/ethyl acetate, water/CH2Cl2 and water/butanol right? I dont know where this protocol comes from, i have used a similar protocol (without CH2Cl2) and i have suspected a big german pharma giant be responsible for that.
The problem is : if you leave a medium polar substance the chopice of going into water or hexane it will be found in both phases, concentrations depend on how many repetitions you do and the volumes used. its the same with the other steps.
Finally I have used 90% aqueous methanol for extraction, reduced it to 1/3 (then ca. 70% MeOH) and did a partition ageinst hexan. The methanol is still unpolar enough to hold the the medium polar compounds, while fats, chlorophyll, sterols and all that unpolar stuff went into the hexane.
The methanol is then further reduced to more or less pure water and partitioned against chloroform. but that was just due to my selectivities, subsequent ethyl acetate, methylen chloride and Butanol would work also.
I got the impression that it is faster do one of the medium polar extrats and do the gradient NP chromatography (or large scale SPE) and then NP chromatography with a different solvent. (snyder: Classification of solvent properties of common liquids)

Did you make sure that your peaks are No plastic additives. I have Phtalates in nearly every fraction. There was even a book about this plastik additives and their MS spectra.
Another possibility woulld be that compouds ionise easily are not readily retained in your unbuffered RP-system.

Alex

[HW Mueller]

You see this stuff in 46 consecutive fractions, spanning many minutes of a run?
If you are sure that it´s not from the solvents (how are you sure??) than it´s from the column (bleed or crud).

[Fee]

Ube neue : I am not sure that the tubing of the HPLC could have something to do with the phenomenon I have observed since I had the same probleme while I was doing manual column chromatography on silica gel or sephadex. However, I am using glass bottles with teflon cap to stock my samples. You think that this could be the cause of the probleme ?

Alex : I did a 80% MeOH 1L extract that I concentrated under vacuum to about 200 mL. Then I partitionned with hexane, EtOAc, CH2Cl2 and BuOH. The compounds I am looking for are not phtalates. I can easily identify phtalates with GC-MS. They are present each time something has been kept in either a plastic bag or plastic box.

HW : I have already checked for contamination in the CH2Cl2. Actually, all the answers I received here are about to convaince me to repeat the protocole from the beginning and check what's happening...

Thank you again for your answers,

Fee

[Uwe Neue]

On an HPLC, you have narrow peaks, and any teflon tubing will just widen the peaks until they are unresolved. This will happen even if there is adsorption on the tubing.

Collect the fractions directly at the detector outlet, as close as you can get to the detector cell, without teflon tubing.

[HW Mueller]

What about the column?

[Fee]

HW : I actually tried a lot of column (Silica gel, Sephadex LH-20, semi-prep HPLC) and also TLC separation, all of them being tried with several mobile phases, and I had the same problem : both compounds I am looking for appeared in almost all fractions.

I just don't know what to thing abou that.

I runned a GC-MS sequence with several old samples I have worked on before, from the liquid-liquid phases extractions to the fractions collected from LC columns. I will check the results this morning and I will come back with this.

Thank you again for your help.

Fee

[HW Mueller]

You have not characterized the two compounds? The MS doesn´t ring a bell? What makes you sure that you are not chasing "dirt"? (didn´t someone else mention something like that, above?)
Or are you vastly overloading everything?
In either case, the GC is one vast mess? (Still running after injecting this stuff?)

[Fee]

Hi Mueller,

I don't think that I am searching dirt. However, I looked for contamination in all my solvent, by concentrate them from 200ml to 2 ml and nothing appeared in the GC. Those two compounds does not come from the column. I tried to derivatize them with BSTFA but they remained the same (so there is probably no hydroxy function on them). I saponified them, I extracted the polar phase with hexane. Then I reacified the polar phase with HCl and I extracted it with 1) CH2Cl2 and 2) Ethyl acetate. None of those fraction contained my compounds. So I think they are saponifiables.

That all the new info I have. Any more idea ?

Thanks a lot,

Fee

[Tony_Montanari]

Until you can get these compounds to run as round spots on TLC, you are wasting your time. TLC is a quick screening method for using column and HPLC.

Start with RP TLC and try running it with acid, base and neutral pHs. Then try silica with hex / isopropanol mixtures. Sometimes Ammonium hydroxide in the silica mobile phase can be helpful.

But get round spots first on tlc before proceeding.

[Fee]

Hi Tony,

I did TLC several times with different mobile phases but I got the same probleme : each time I was analyzing (on GC-MS) each band obtained from TLC, both compounds were present in all bands.

I usually used neutral or acidic (acidified with CH3COOH) mobile phases. What can I use to have a basic non polar mobile phase ?

Thank you,

Fee

[Tony_Montanari]

For basic ion suppression you can try any number of amine and ammonium compounds. Try ammonium hydroxide first. Diethylamine and triethylamine are also possible. Try about 1 mL per L of mobile phase to start.

It is important to get round spots from TLC to start. If you streak with TLC then the same will happen when you get to HPLC or column chromatography.

Have you had any success with TLC not streaking? you mentioned that you are using acetate as a modifier. Does this give you uniform spots?

You may also want to try using ion pair reagents. Things like hexanesulfonic acid or tetrabutyl ammonium hydroxide or tert-butyl ammonium hydroxide. If you have an ionic compound, then there is a chance that ion suppression may not always work. It could be that adding a PIC reagent will give you tight bands and you can collect the compound. Then the PIC reagent can be removed by ion chromatography.