Doubts about Ion pair chromatography

[Tequila]

Dear members of this forum,

I have read a lot of discussions about ion pair chromatography and I like a lot this subject and I would like to understand and clarify more about certain points.
The triethylamine works avoiding the tailing of the peaks becuase it join to the free silanols but not interact with the analytes through hidrophobic unions in other words it does not work as ion pair agent, I think.
The tetrabutylamine works as joining to the stationary phase and works as ion exchange chromatography because has quaternary nitrogen free to interact with anions (for example, anions of organic acids).
The sodium salts of sulphonic acid works in the same manner as tetrabutilamine but it has free anion to interact with cations (for example, cations of organic bases or protonated amines).
In other articles I noted that perchlorates salts and TFA work as ion pair agent too but I do not understand how works because I do not see an important hidrophobic part to interact with the stationary phase.
As I look the interactions between ion pair agent and the stationary phases is strong and I would like to know if it is easy to clean the columns with strong solvents as acetonitrile or methyl alcohol after working with it and start with other ion pair agent in the next analysis. It could be better works with columns exclusive to each ion pair agent. I noted people have columns only for ion pair chromatography but they can work with diferent agents.
I am writting this because I would like to understand some important features about ion pair chromatography and I can see in other interesting discussions, a lot of members can help me in order to understand more about that and excuse me for a long text.
Thanks a lot in advance,

Diego

[Chris Pohl]

Tequila,

There are a lot of different topics in your post and I'm not sure if it's reasonable to try to address all of them in such a format since some are rather complex but let me take a stab at a couple of them anyway.

First off, as Bidlingmeyer showed (J. Chromatogr. 186, (1979) 419), there are actually no ion pairs present in solution in "ion pair" chromatography. For that reason he proposed changing the name to ion-interaction chromatography but that nomenclature has not been widely accepted since the ion pair chromatography name was already in wide use when he published that paper.

Second, the retention process can differ significantly depending upon operating parameters and the nature of the analyte. For example, with tetrabutylammonium ion as the ion pair reagent and chloride as the analyte, the retention process is nearly exclusively based on an electrostatic interaction (it might be fair to describe this process as dynamic ion exchange). However with tetrabutylammonium ion as the ion pair reagent but with SDS as the analyte the retention process more nearly approximates the ion pair model in that retention is induced more by lowering the solubility of the analyte in the mobile phase than by increasing the electrostatic attraction of the analyte to the stationary phase. Furthermore, in the latter case chromatographic conditions are generally such that there is significantly more solvent in the mobile phase resulting in much more rapid transport of ion pair reagent between the stationary phase and the mobile phase. In essence, one can consider both the analyte and the ion pair reagent to be exhibiting similar mobility in the second mode (even if they aren't true ion pairs in this system, the system behaves as if ion pairs are moving freely from mobile phase to stationary phase and back again) whereas in the first mode the mobility of the ion pair reagent is much lower than the mobility of the analyte ion. One can consider a static coating of ion pair reagent on a reverse phase column as the extreme case of this mode of operation where the ion pair reagent is only on the stationary phase and only the analyte is moving between the two phases.

Ion pair reagents such as TFA or perchlorate are largely incapable of operating in the first mode because they exhibit very little "native" retention on normal reversed phase media (especially in the case of perchlorate) but they can still be used effectively in the second mode.

Finally, regarding the desirability of transitioning back and forth between reversed phase and ion pair separations on a given column this depends upon the ion pair reagent and the nature of the stationary phase. In the case of silica based reversed phase columns, rapid and quantitative removal of cationic ion pair reagents is rather difficult and so in this case it's generally not recommended to switch between ion pair and reversed phase. But in the case of anionic ion pair reagents similar problems are not observed and in this case switching back and forth shouldn't be an issue except in the case of columns having residual anion exchange capacity as a byproduct of their synthesis. Likewise in the case of polymeric reversed phase media switching back and forth usually isn't an issue because there aren't any residual silanols or other minor ionic functionalities to bind ion pair reagents electrostatically.

[HW Mueller]

Couldn´t find a copy of a Bidlingmeyer.... article in which, if I recall correctly, he (they) disprove(s) the presence of ion pairs via conductimetry. Now I wonder, if he misses 1%, or 0.1, or even only 0.01%...... ion pairs with that method how such remaining concentrations could effect retention.
I have a formular k = (K*n)/V, where k is the retention factor, n is the total number of adsorption sites, K is the phase distribution constant (an association constant), and V is the mobile phase volume in the column. Any math wizzes out there which can give an idea on how few ion pairs are required to effect reasonable retention?
I am, of course, shooting at the possibility that ion pairs are underestimated.
Wasn´t there a similar discussion before?

[Uwe Neue]

Hans,

while Bidlingmeyer disproved the existence of ion-pairs in the mobile phase, he demonstrated that there is a co-migration of the oppositely charged ions as well as a gap or dip in the concentration of the ion-pair reagent at the retention of the reagent in proportion to the concentration of the analyte. He interpreted this as stemming from the need to maintain charge balance as the analyte adsorbs on or into the stationary phase.

The experiments were rather logical and conclusive. Have to say though that there were some simplifications compared to the everyday practice, if I remember correctly.

[Bill Tindall]

it was established long ago that ion pairing begins in methanol at about 60% methanol. I don't have access to the B. paper at this moment, but if he concluded that ion pairing in methanol never occurs I'll wager his data is flawed. I haven't made the measurments but I'll put my money with the classical solution folks that studied this stuff 2 generations ago.

[HW Mueller]

Well, maybe most of the stuff existed in absorbed and/or in "free" ionized form, but 100%??? I am affraid that Bill is right.

[Uwe Neue]

Bidlingmeyer indeed claims the opposite, says that he has measured it up to 100% methanol, but only shows the data for 35% methanol.

"Additional conductance titrations were carried out looking for evidence of ion-pair formation in methanol water solvents ranging from 0 to 100% methanol. ... No evidence of ion-pair formation was found in any of these conductance studies."

[HW Mueller]

That´s what I remembered, that he hinged the stuff on conductimetry, so I still wonder: How many % ion pair can you miss that way?

[Chris Pohl]

Hans,

Since I am the guilt party in bringing up this ugly subject once again, I should give a bit more background on what I know about this issue surrounding Bidlingmerer's paper. If you dig up the original references that are the basis for the claim that no ion pairs are found in solution under standard HPLC conditions, these references date back to the 30s (J. L. Whitman, et al., J. Amer. Chem. Soc. 52 (1930) 4762 C. A. Kraus et al. J. Amer. Chem. Soc. 55 (1933) 93 and C. A. Kraus et al. J. Amer. Chem. Soc. 55 (1933) 1019). They show the effect of dielectric constant on dissociation constant for ion pairs in solution using tetraisoamylammonium nitrate. Their data show that there can be no ion pair formation with the above electrolyte when the dielectric constant is greater than 43.6 (based on extrapolation from measured dissociation constants at different dielectric constant values). The latter technique allowed the researchers to avoid the problem of accurately making measurements under conditions where the magnitude of ion pair formation is minimal. A dielectric constant of 43.6 corresponds to 24% water-76% methanol or 14% water-86% acetonitrile. For the electrolyte mentioned above further increases in the water content produced conditions where no ion pairs can be detected.

Now, to complicate matters further the above analysis only relates to the specific electrolyte mentioned above. The tendency of specific electrolytes to form ion pairs is dependant not only upon the dielectric constant but also upon the size of the solvated ion. Other ions may well form ion pairs at even higher water contents. To support this, an interesting paper which was written around the time of the Bidlingmerer paper was written by none other than Csaba Horvath (Hovath et al., J. Chromatogr., 201 (1980) 201). In this paper he disputes the conclusions of Bidlingmerer arguing that the conductivity method that he used is incapable of detecting ion pairs with a formation constant of less than 100. Instead, Horvath advocates the use of Job's plots which are considerably more sensitive for measurement of small formation constants. He used this approach to measure a formation constants for catecholamine ion pairs with alkylsulfates and measured formation constants that ranged from 8 to 20 using 100% aqueous media. His data suggest the existence of weak ion pairs even in aqueous media but by his own admission the magnitude of these formation constants is too small to solely account for the observed retention. To quote Horvath: "Nevertheless, abundant experimental evidence suggests that neither the simple ion pairing nor the dynamic ion-exchange model describes adequately the mechanism of the chromatographic process over the full range of practical operating conditions."

So where does this leave us? I guess it would be fairer to say that ion pairing may play a significant role under certain conditions but in general it is not the dominant factor in "ion pair" chromatography.

[tom jupille]

Chris, I'll stick my neck out here: my impression is that, in general, both the "ion pair" model and the "dynamic ion exchange" model result in pretty much the same predictions about retention behavior and selectivity (that's why the discussion about the mechanism persists after a quarter century).

[HW Mueller]

Maybe the Horvath paper has some math relating the K(phase distribution) to the K(ion/ion pair equilibrium), just don´t have the time to "power" into this.

The next ugly phenomenon: I have seen some exceedingly long "memory" effects of pH on silanol behavior, too early to mention any details, though. Might be related to this (controversial?) ion pair (detergent) interaction with stationary phases.

[Chris Pohl]

Tom,

Well, although it's generally true that the predictions for the are similar, there are some areas of significant difference between the two models. Most notable is the predicted effect of the counterion. In the ion pair model, the retention process is pretty much dominated by the relative "hydrophobicity" of the ion pair. In this model, one wouldn't expect the counterion to play a major role in retention and in fact one would expect added electrolyte to lower the solubility of the ion pair in the mobile phase and thus enhance retention. But in fact, addition of electrolyte to the mobile results in reduced retention and the relative effect of the counter ion on analyte retention is exactly as one would expect based on ion exchange selectivity order (i.e. with equimolar solutions of two different ion pair salts, say for example tetrabutylammonium chloride versus tetrabutylammonium nitrate, an analyte will experience the longest retention time with the eluent containing the most weakly retained an ion, chloride in the example provided).

So, I really think the dynamic ion exchange model is a better prediction of behavior but I believe that many prefer the ion pair model because it's easier to conceptualize more than because it's more accurate.

[HW Mueller]

Chris, your last example shows how complicated the matter is, all I am saying is that there does not seem to be unequivocal evidence either way. For instance, the salting out effect to which you appear to allude: At low concentrations even the most lyotropic salts are chaotropic (salting in). Also on Cl- vs. NO3-, could it be that Cl- displaces the analyte ion to a lower extent from the ion pair than does NO3-?.
To be honest, though, the older I get the less sure I become. In a related matter, the nucleophilicity of, as an example the halide anions, is reversed in going from water to a much lower polarity, etc., solvent (a shock to me when I recently saw this, having grown up with all studies done in highly polar solvents, with iodide being THE electrophile). So if one does not know about the thermodynamics (especially the entropy part is often counterintuitive) of everything invoved it is difficult.... and with the knowledge one might have a whopper of a math problem at hand.

[Victor]

There is a series of papers in the Journal of Chromatography by Guiochon which suggests that ion pairing of solutes such as protonated bases occurs with chloride ion using C18 columns. If this is true, I would find it hard to imagine that dynamic ion exchange could take place with the chloride ion providing the dynamic ion exchange site. I don't think chloride ions are retained on a RP surface. I could only imagine this happening somehow by the formation of ion pairs. On the other hand, I would have no problem with the idea that a TBA cation sticks on a RP surface and acting as a dynamic ion exchange site. It is possible that the mechanism of "ion pair" chromatography changes dependent on the nature of the ion pair reagent.

[Uwe Neue]

I don't know... My first reaction would be to say that you can't call everything ion-pairing. On the other hand, due to the requirement of electroneutrality, no ions are ever on their own, and they are always "paired". Maybe this points to the thought that the idea of an ion-pair is really a misnomer, at least in an aqueous system at the typical concentrations that are used in "ion-pair" chromatography.