1 molecule that appears in two really distinct fractions !

[clairette31]

Hello everybody !!

I am doing a purification of an heterocyclic amine at semi-preparative scale. I find my molecule (1) in two different fractions that have very distinct retention times and my molecule is always with an other molecule (2) with a similar UV spectra.
Is it possible that one and two be the acid and basic forms of the same molecule ? Can these two forms have very different behavior with hplc ??

Thanks for your help,
Anne-Claire

[Bruce Hamilton]

Can you please give us more details, including sample solvent and loading, column type, mobile phase composition, etc.?

My guesses would be sample solvent isn't the same as the mobile phase, and/or both sample solvent and mobile phase aren't adequately buffered to a suitable pH.

Another possibly may be the product is precipitating on column due to low solubility in mobile phase at column temperature..

Please keep having fun,

Bruce Hamilton

[clairette31]

Hi Bruce,

Thanks for your interest in my problems !

My sample is in methanol and is acid since it is the result of an acid hydrolysis, dried and then solubilised in methanol.
I am doing my separation with a mobile phase water/acetonitrile on a column uptisphere ODB 5 (interchim) 250 * 7.8 mm.

I can not acidify the mobile phase with any anhydric acid because with water my molecule could react to form an amide link !

I try to adapt the following analytical method that I apply on the column Prontosil C18 SH 150 * 4.6 mm and that permits to separate my molecule.
t = 0 , 20 % acetonitrile 80 % water
t= 30 min, 50 % acetonitrile 50 % water

My product has a retention time of 8 min and some co-products go out at a retention time upper 20 min (23,25,26 and 27 min).

At a semi-preparative scale I inject 250 µL of a sample of approximatively 1.3 g/L.

On the preparative column, my molecule go out earlier (too early perhaps) at around 21 % of acetonitrile.
And the other co-product around from 33 to 37 % of acetonitrile.
I find in the two fraction 21 and 35 % my molecule and only one other that has a retention time of 25 min at the analytical scale.

I already have done preparative thin layer chromatography. I have the same problems. I recover bands from my TLC and I solubilise them in methanol, then I filter silica and I analyse bands with my analytical method. My bands are well separated but I get the molecule (Tr=25 min) in a big proportion with my product (Tr=8 min).

I hope, you will be able to help me !
Anne-Claire

[Mark Tracy]

If your substance does form amide linkages as easily as you say, then it probably is doing exactly that and your sample has both forms in it. This may be an artifact of the way you dry the hydrolysate. Amide bonds are usually slow to form and slow to hydrolyze but in your case the rates may be inconveniently high. I suspect you are separating the two forms, then in the workup of the fractions, the amide degrades back to the parent compound.

Another hypothesis is that your compound has some slow isomerization.

[Bruce Hamilton]

I suspect that Mark might have identified the problem.

I'd also try dilution of a small sample with your initial mobile phase at a suitable concentration, and analyse that to ensure that you are not seeing any solubility issues, which should be apparent by a change in the peak ratios.

The same issue may apply to the the preparative system. You may also have to investigate using a weak buffer to dissolve your samples, if the problem is not caused by the sample preparation.

Please keep having fun,

Bruce Hamilton

[clairette31]

Thanks for your answers !

I do not think that it is a problem of amide formation because I well know the amide compound (Tr, UV spectra) and I do not form the corresponding peak.

But the idea of isomerization is good, I really have the impression that the two peaks, that I do not manage to separate, form an equilibria (that’s why I thought to basic and acid forms). Moreover I really have problems to find in fractions the globality of the molecule injected even when I analyse all the fractions and the result of column wash !!!

But it can also be a problem of solubility because my compound is not soluble in water !!!
I will try to solubilise it in water/acetonitrile (80/20) but the problem is that my total sample volume will probably increase (I already have done little investigations on this side but solubility seem poor and it is always a lost of product to manipulate my sample).

If I dilute my sample in my initial mobile phase what kind of ratio should I use to simulate an injection of 250 µl at a flow rate of 1.5 ml/min ?? Could 1/5 be a good beginning ?

Can I buffer my sample preparation without using a totaly aqueous buffer (Since my sample is not soluble in pure water). Perhaps I will say a big mistake, but, is it possible to prepare a phosphate buffer (for exemple) in water/acetonitrile (90/10) ???

Thanks to you I have a lot of idea now ! But the way to cover is still long !!!

Anne-Claire

[Mark Tracy]

Isomerization is a possibility, but not simple acid-base chemistry because proton transfer reactions are nanosecond reactions.

There is no reason not to dilute a phosphate buffer with a little acetonitrile. Do check the solubility of the mixture before putting it in the HPLC.

[Uwe Neue]

There is another possibility to consider. I have seen it more often with DMSO injections, but it is also possible with other organic solvents. The thought is that you inject too large an amount of analyte dissolved in organic. Only a fraction of the analyte gets diluted out of the organic plug and is retained on the column, the rest gets through the column as an unretained peak. This seems to agree with the observation that it elutes in the starting composition of your mobile phase.

The best solution to this problem is at-column dilution. You need another pump for doing this.

[JA]

There's quite a few things which spring to mind, but as it's so very late my reponse will probably lack continuity...

As I'm not familiar with the term anhydric acid, how does an amine form an amide in it's presence? I can't say I've ever observed the formation of an amide in RP HPLC during the analysis of an amine in a mobile phase modified with a carboxylic acid (Formic, TFA). Esters yes, amides no.

I can't remember where or who I first got this from, but in my mind the requirement for a buffered mobile phase is a given when your analyte is ionisable.

It's true that there can be a large change in retention between an ionised and unionised form; perhaps a factor of 10-20 under isocratic conditions. I admit I don't know how this is reflected under gradient conditions. Is a gradient required... not unless you are trying to maximise peak capacity throughout the 30 minute run. Are you trying to collect all the co-products separately aswell? Isocratic or a step-up after the product has come out could lead towards an increase in throughput.

There's a fair bit in the literature to show that both peak shape (and thus the general chromatography under many circumstances) and preparative loadability for an analyte is improved in the unionised form. Ideally therefore, your work would take place in alkaline conditions - but will the Uptisphere ODB column allow it...

I'm wondering about the TLC work and if it offers any clues. Is it performed under reverse phase conditions - how is it run? You say you've analysed the two collected bands by HPLC but have you developed them by TLC aswell - do both prepTLC bands give two spots by TLC?

It's an interesting thread though. Good luck.

edit: I realised I took so long to write that I missed Uwe's post in the mean time. I also seem to be going on about changing everything you're doing rather than troubleshooting the problem like the other contributors. Expanding on the suggestions in Bruce's second post will allow you to check for the strong solvent effect. It would be real unfortunate (read: unusual) if your stuff only goes in methanol.

[clairette31]

Hi everybody !

I really think now that I have a problem of instability of my molecule (perhaps an imine formation because my molecule have both reactive amine and carbonyl groups). I Would like to try to buffer mobile phase at pH8 (imine formation takes place around pH5).

I am going to post a new topic about precautions to take, working with buffers ... because I heard that it can damage columns if correct washes are not done !!!

"See you in this new topic"

Anne-Claire