Doubts about Ion pair chromatography

[Chris Pohl]

Sorry, but because I've been doing research in this area for more than 25 years-I feel obligated to confuse matters further by adding an additional fact that flies in the face of the dynamic ion exchange hypothesis. While I previously mentioned the fact that the counterion effect is not predicted by the ion pair model, there is also a major experimental observation not readily explained by the "dynamic ion exchange" model: the amount of the ion pair reagent absorbed on the surface is counterion dependent, even in a case of counterions which are unlikely to be able to form ion pairs because of extreme hydrated radius (e.g. fluoride and chloride counterions of tetrabutylammonium ion). In a paper by Frederic Cantwell (Cantwell et al., Anal. Chem. 51 623 (1979)) from about the same time as the paper from Bidlingmeyer and the paper from Horvath, the presence of the electrical double layer at the surface of the stationary phase is proposed as an explanation for the observed behavior with "ion pair reagents". This model explains both the counterion effect and the variable surface concentration of the ion pair reagent without requiring a postulated "ion pair". The counterions are present in the diffuse region of the electrical double layer and allow for the effect of these counterions on the retention process. In essence, an ion which allows for a more condensed electrical double layer will allow a correspondingly larger concentration of the ion pair reagent to be present on the surface of the stationary phase. The "dynamic ion exchange" model is insufficient to explain this effect (it predicts that retention would be counterion dependent but does not predict that the ion pair reagent surface concentration would be variable).

To cloud matters further, here is a "bone" for those of you who prefer to think of this retention mode in terms of ion pair formation. The physical chemistry at the surface of a hydrophobic material leaves open the possibility that while ion pairs may generally be unlikely to form in the mobile phase, they may well be present in the stationary phase. For example, in a previous post I mentioned a paper which showed that a dielectric constant of less than 43.6 was required in order to have any measurable ion pair formation in the case of tetraisoamylammonium nitrate. In that specific case, more then 76% methanol or more than 86% acetonitrile are required in order to achieve that dielectric constant. While such conditions are uncommon in ion pair chromatography, the dielectric constant of the stationary phase under conditions commonly employed in ion pair chromatography might easily be below that critical limit. For example, the dielectric constant of the stationary phase in the case of C18 bonded phase silica is a function of the bonded phase, the solvent content and the water content. Although I'm not aware of any specific stationary phase dielectric constant data measured under reversed phase conditions, it seems fairly likely that the dielectric constant would be low enough to support ion pair formation since the stationary phase should be fairly rich in organic solvent relative to the mobile phase, poorer in water content relative to the mobile phase with the added dielectric constant reducing effect of the bonded phase. If this is so, then perhaps ion pairs actually form as they approach the stationary phase even if they can't form in the mobile phase.

[Uwe Neue]

A bonded phase is a monomolecular layer. Being a C18, it is a bit thicker than a standard probe molecule, and a bit thicker than a stretched surface active molecule such as SDS, but this does not make it a bulk hydrocarbon liquid. I think there is no "bone" ...

[HW Mueller]

Well, if one can invoke a double (or if convenient, higher) layer of ions to explain things then why not a double or more layer of organic modifyer on the C-18, with bulk solvent properties? Laminar flow can be invoked to indicate that.
Can something be learned here from interface or lipid bylayer studies?

Anyway, I am beginning to loose oversight so it seems to me that Tom "hit the nail on the head", as we would say here, when he pointed out that the sundry explanations seem to leave possibilities open, maybe wide open?

[tom jupille]

I've found this thread very instructive; more understanding is always better than less. As a matter of practice, however, simplified explanations such as "dynamic ion-exchange" or "ion-pair" are usually sufficient even if they range from incomplete to incorrect. The trick is knowing how far to push the model.

As a trivial example, "sunrise" is a useful concept for day-to-day living, even if we all (well, almost all) recognize that it's based on an incorrect model and that the earth, sun, and planets actually rotate around a common center of gravity. If I want to send a probe to Mars, on the other hand . . .

[Bill Tindall]

I have been following this discussion with interest, especially Chris's learned discussion.

The dielectric constant of a solvent has a huge effect on the pKa of acids, especially if there is a double charge involved (eg phosphate). If the dielectric constant of what I will call the retention volume (place where retention occurs) differs from that of the bulk solution then there should be great differences in retention from that predicted from bulk solution pH and analyte pKa( in the elution solvent). No doubt someone has done this work. In fact I think Roses, etc. were able to model retention and the models fit well (from memory, I may be inspired to go look the paper up to see if it is relevant to the discussion).

In my graduate research (Kolthoff) group it was accepted as fact that ion pairing was important above a methanol concentration of 60%. It was so accepted that I don't recall a reference ever being mentioned, or the "fact" questioned. I will see if I can find the source of this "fact". I do have in hand a reference(Grunwald) to a paper reporting ion pair formation at a dielectric constant larger than 43.5.