ion pair for atypical samples

[david s]

hi
I have developed an ion pair method (diethylammine phosphate) for the seperation of corticosteroids. In particular it was developed to exploit a critical pair that has very similar physical properties i.e solubility (logp octanol:water is the same value) but different chemical arrangment. Ring A for compound A has 1 more double bond that compound B. My theory went a little like this :- compound A has a larger electron density around ring A than does compound B, and hence a bigger dipole within the molecule with that in mind, using the above ion pair, compound A should retain on the column ( luna C18) for longer and separate the critical pair. Now here is the issue i have seen good separation of compound A and compound B but compound A comes off first, not second as predicted in my theory. does anyone have any ideas what is going on? the other consituents of the mobile phase is THF:Water:meoH in the quantity 20:75:5. although again in theory as the solubility logp octanol:water are the same i doubt that these contribute to the critical pair separation and infact when you remove the ion pair, the separation is no where near as good as when the ion is there so the ion pair is doing something, but i am not sure what

[HW Mueller]

One can not comment on the effect of the ion pairing agent, with this info., but if you think that A is more polar than B than one would normally suspect A to have the lower r.t.

[david s]

hi,
the two molecules are hydrocortisone derivatives, they are both missing the seconday keto group of ring A, but compound A has an additional double bond on the ring A ( conjugated to the typical double bond pi electrons). everything else is the same as hydrocortisone.

[JA]

I would add a couple of thoughts:
- What are the RTs of compounds A & B with and without the diethylamine phosphate?
- Can DEA phosphate act as an IP reagent in the stated conditions?
- Do true IP reagents influence the retention of dipolar compounds in the same way as formally charged species? i.e. IP+ δ-(compound)
- Could it just be down to solubility differences in the MP? (not your logp values)

[david s]

hi to clarify with some data

for no ion pair, compound A elutes at 8.1 minutes. compound B elutes at 8.8 minutes resolution is 2.6

for ion pair, the compounds elute at the same times, resolution is 3.2 usp tailing is observed to be the same at 1.1 in all cases peaks do increase ( sharpen in height as ion pair conc increases

the methanol was added to affect some other related compounds that appear later in the chromatography (solubility wise) but removal of the methanol does not unduly affect retention of compounds A and B. although it was noted that without the methanol, resolution was further increased to about 3.6 ( although this would give rise to a different critical pair). i do not have any valuble solubility data for the compounds other than the octanol : water system and although i do not have octanol in the mobile phase. the fact that they have the same value of 1.6 suggests (to me) that their solubility in most solvents would be the same.

[Bryan Evans]

For neutral compounds, you may find it helpful to use an ODS phase
that is not fully endcapped (as opposed to a fully endcapped phase).

In the chromatogram below (adreno-corticosteroids), Cadenza CL-C18 (partially endcapped) provides better
separation for the critical pair than the fully endcapped ODS column (Cadenza CD-C18):

http://www.imtakt.com/TecInfo/TI183E.pdf

[HW Mueller]

Cortisol is not ionized under such HPLC conditions, if this holds true for your derivatives I don´t see how you can call this ion pair chromatography. My guess would be that you just changed the surface properties of your column with the ion pair agent. Strange effect, though? Don´t really see how it is possible. I suspect that you are not telling us everything of relevance, like amounts injected., integration settings.....

[david s]

HW Mueller

I agree entirley with you, although an ion pair reagent is used, it is not ion pair chromatography- the approach was to exploit the additional double bond and hence the percieved dipole within the molecule. the initial theory being that the more ionic character (induced dipole) percieved in molecule A should give rise to a larger retention and come off after molecule B. The reverse is true and so i am not 100% sure what is going on (on column). FYI the injection is 20µl and consists of 6µg of each substance ( other related substances are also included including cortisone, 4 androsten etc). the resolution is such that intergration is at baseline. injection solvent is MP without the ion pair reagent. cannot think of anything else to add. you may be right in that this is just a surface phenomenom or it could be a solubility thing bearing in mind that it does not follow that because the octanol : water values are the same that the same solubility would exist in all sovents this is an assumption made on my part. but i am ever interested in trying to understand what is going on. I really do not like accepting something because it works and not knowing why it works

[JA]

david s,
You have put a lot of faith in the logP values of compounds A & B. May I ask which technique you used to determine them and suggest that perhaps they are not exactly the same? Your empirical evidence of nonequivalent retention on a fully endcapped C18 phase tends to suggest, as I personally expect, that they differ in hydrophobicity.

Is the understanding of your observations important to your work or research or are you simply looking to improve upon the separation? If the latter, it probably opens this thread up to a lot more input.

[Bryan Evans]

What is your value of k'? If you have plenty of retention - then I think
you can do without the IP reagent. Diethylamine is classicly used
for anionic compounds.

For RP, you can try playing with a fully endcapped C18, a partially endcapped C18
(such as Cadenza CL-C18), or a Phenyl phase.

A: H2O (or neutral buffer)
B: ACN, MeOH, or THF (or combination of the 3)

Provided you have enough retention - I would think that would be a better choice.

[david s]

thank you for all your replies. i believe the co-efficient was determined by the flask method using UV as the measure.

Value of K is approximatley 8 so i can play around with it. the reason for the method development "enhancement" was that an exisiting method using 20% THF in water yielded unressolved peaks.i did play around with THF,MEOH and ACN and could not improve greatly on the peaks ( not to give the ressolution i am now seeing) this was an adaption of the EP method for hydrocortisone. now as the unressolved peaks were unsucessful i set about my theory looking at the molecularity. now as it is improved ( and reproducible across columns of different batches of silica) i am pleased but as the result does not fit my theory i would like to know what is going. i dont mind opening the string up for more input and all suggestions are welcome. Based upon the Hydrophobicity idea JA, as you run from no ion pair to approx 27 mM improved ressolution is seen so it must be doing something you would guess although the retention time does not change peak height is seen to increase and from an overlay of chromatography the peak width at half height does decrease

[HW Mueller]

Where does your theory come from that a more polar compound will be retained more strongly on a C-18 phase??
I suspect some change, which you are not mentioning, is responsible for your peak shape differences.

[david s]

my theory was not so much that it is retained more strongly on a C18 column, but i was "expecting" a polar molecule to be retained for longer on a C18 that has diethylammine across ajoining the C18 adducts. than on a C18 alone
The thought process was that the positive amine group from the ion pair, that is sitting across the C18 would have an affinity for molecule A 's extra electron density around ring A- in turn the equilibrium between molecule A retention on the stationery phase against mobile phase would be greater than molecule B resulting in a longer retention time than molecule B.

Please note i am not looking to improve the chromatography as that has been shown to be fine in terms of repeatability, and the ressolution has improved over what was the original method i inherited, but
I do not understand what is going on and it frustrates me

[Alex Buske]

hi David,

Having played with corticosteroids some time ago I do not understand your ion pairing approach. They are neutral.
Chances for separation come up with pentafluorophenyl columns or with C18 at low temperatures (ACN) or with C18 with THF as far as i know.
Differnces in logP are rather small and as your solvent isnt pure water extra effects come in (assuming that the ODS plays the octanol part).

Alex

[david s]

hi Alex

The approach went like this , an existing method utilising THF and Water did not separate to baseline two synthetic analogues of cortisone. A quick bit of analysis using THF,Acetonitrile and methanol at various concentrations did not brink about any real meaningful improvment to the critical pair without seriously affecting or generating any other peaks in the chromatography. examination of the molecularity shows the only difference in the structure ( as confirmed by NMR and Mass sec) of the two molecules is the appearance of an additional double bond in ring A - This lead me to think that my best option is to try and exploit this difference in structure. It follows that 2 double bonds on ring A in a conjugate system must have a higher electron density at ring A and therefore experimenting with ion pair may bring about the separation, not through a truly ion pair chromatography but applying the principles of change affiliation the theory also being that molecule A which has the extra double bond ( and hence the stronger affilitiation due to localised electron density) should retain on the column for longer and come off molecule B. In practice a good separation is observed however molecule A does come off first. This is the bit that is confusing.

Now when you run on a new column, without the ion pair, the separation is not so good- introduction of the ion pair on the same column does improve the chromatography. and so the system is shown to be reliable and reproducible but i do not understand ( and hence my problem) the mechanics as to what is going on

hopefully my logic makes sense