Chromatography Forum

Here is a question left over from another thread-maybe someone can explain it.

Why are highly hydrated inorganic ions like fluoride less retained in anion exchange hplc (for instance using weak anion exchange chromatography)?

Victor,

Sorry, I haven't been ignoring you but I've been embroiled in a lot of administrative duties lately so I haven't been able to respond (I also dug up some interesting information on the protein precipitation part of your question which I will post in the other thread). I'll try to post my answer to your question in the next 24 hours.

The simple explanation is that a small and concentrated charge like fluoride ion is shielded by water molecules organised around the ion. That creates a "protective layer" that keep the charge away from the corresponding charge on the ion exchanger (also hydrated).

As opposite for a large ion like bromide that may be polarised and the water molecules layer does not shield the charge to that extent. These are sometimes called "soft ions" and are normally more strongly retained.

The cations generally have a different (mostly opposite) sequence in lyotropy (salting out) and chaotropy (salting in) lists in comparison to IX selectivities and energy of hydration series. The shielding by water (also mentioned by Chris?) doesn´t seem to explain that.

Victor,

The reason that highly hydrated ions are weakly retained in ion exchange is a direct consequence of the hydrated ionic radius. For monoatomic ions such as simple halogen anions, alkali metal cations or alkaline earth cations there is a simple relationship between position in the periodic table and hydrated ionic radius. Ionic radius increases as one proceeds to higher atomic weight ions in the same column. Hydrated ionic radius follows the inverse trend. Thus, in the case of halogen anions, fluoride has the smallest ionic radius and the largest hydrated ionic radius. Since, to a first approximation, the forces involved in ion exchange retention are directly related to the distance between a given ion exchange site and a given counterion (with the force being greatest when opposite charges are the smallest distance apart) one can predict elution order based on hydrated ionic radius. In the case of the simple halogen anions, the prediction is for fluoride, the anion with the largest hydrated ionic radius to elute first and astatide, the anion with the smallest hydrated ionic radius to elute last (not that I ever had access to this anion to determine its retention time). Likewise, in the case of alkali metal cations there is a similar relationship with the prediction being: lithium, the cation with the largest hydrated ionic radius eluting first and francium, the cation with the smallest hydrated ionic radius eluting last.

Extending this relationship beyond simple monoatomic anions and cations is more difficult, though. Unfortunately, there is no readily available resource with accurate tabulated data for the hydrated ionic radius of polyatomic ions. For this reason, hydrated ionic radius information cannot be used in general to predict selectivity. In addition, cross-linked ion exchange materials add a new dimension to the selectivity equation, further complicating matters. The selectivity of a given material can be varied significantly depending upon cross-link because cross-linked phases must expand to accommodate larger ions. The opposing forces of the stationary phase cross-link impair the ability of highly hydrated ions to enter the matrix. Hydrated ionic radius is more closely correlated to entropy of hydration whereas enthalpy of hydration is a better indicator of the ease with which in ion can shed its waters of hydration in order to enter a cross-linked matrix. By changing the cross-link one can shift the relative importance of entropy and enthalpy of hydration. For that reason, elution order can be varied by changing the cross-link of the stationary phase to favor specific ions. For example, it requires a relatively large amount of energy for fluoride with its extremely large hydrated radius to shed a significant fraction of its waters of hydration in order to enter a highly cross-linked matrix. For that reason, fluoride is nearly unretained in highly cross-linked materials but exhibits significantly more retention relative to chloride as the cross-link approaches zero.

Chris, this is an excellet explanation and thank you very much for it and the time and trouble you have taken. I was wondering about the relative influence of coulombic and size exclusion effects on the retention, but now you have explained it.

Just one final point, is it possible to put some general figures on the pore sizes of the "highly crosslinked" and "zero crosslinked" phases which you refer to? I have hardly worked with polymeric phases, although I guess these figures are less meaningful with such phases than silica based phases, because they shrink and swell dependent on conditions.

Victor,

When it comes to discussing pore size in cross-linked ion exchange gels, the problem is that such materials don't have a true pore in the sense one typically thinks of with, for example, a silica based chromatographic material. In the latter case, there are distinct pores is that can be measured and characterized by a variety of analytical techniques. On the other hand, characterization of the porosity of an ion exchange gel is a bit trying to characterize the pore size of Jell-O. An ion exchange gel is better thought of as a three-dimensional fishnet where the various elements of the network are randomly moving around relative to one another. The pores are neither fixed in location nor fixed in size and so they can only be estimated as an average pore size without a lot of information about the range of pore sizes that might exist in the material. Having said all that, you can still obtain size exclusion estimates in the published literature for bulk media based on the same chemical composition. As a rough rule of thumb, an 8% cross-link ion exchange gel typically has an exclusion limited in the range of around 1000 molecular weight. In ion chromatography, anyway, we rarely use such highly cross-linked media. At 2% cross-link, the exclusion limit approaches 3000 molecular weight. This is closer to the average for most cross-linked materials used in ion chromatography applications.

One additional point to remember regarding silica based ion exchange materials and cross-link: because the silica based material is rigid and does not swell in the mobile phase, such an stationary phase is effectively the same as a cross-linked gel in the sense that the ligands attached to the surface of the silica have a well defined average distance in much the same way that the ion exchange sites in a cross-linked three-dimensional gel do. In essence, such a material can be thought of as a two-dimensional cross-linked gel. Of course, the actual pore size of the base silica defines the exclusion limit but the selectivity of the phase for polyvalent species is defined by the bonding density on the surface in exactly the same way that polyvalent species are affected by the cross-link of a gel.

Chris, many thanks again for these clear and knowledgeable explanations.

Chris, your last sentence of your post, Aug 28, 10:25pm, on the retention reversal of Cl- and F- depending on cross link had me wondering whether a reversal takes place in a highly crosslinked material if the flow is slowed considerably. If F- is retained more strongly in a low crosslinked material it would mean that the relevant interaction of F- is thermodynamically stronger than that of Cl-. Since water of hydration exchanges very rapidly the thermodynamics should be determining the retention behavior in a very slow running highly crosslinked column as well?

Hans,

I'm afraid my wording in the earlier post wasn't clear. I didn't mean to imply that there is a fluoride-chloride elution order crossover when the cross-link approaches zero. I only meant that the relative k' becomes significantly more similar for the two analytes. With higher cross-link materials the k' for fluoride can easily fall to a value of 1/10 the that of the k' for chloride whereas with low cross-link materials, the k' for fluoride approaches one half that of the k' for chloride. In essence, elevated cross-link tends to "exclude" highly hydrated anions based on the thermodynamics of dehydration. As the cross-link approaches zero this thermodynamics penalty is virtually gone and in this case the determining factor in is the hydrated radius of the ion. Since fluoride is the most hydrated, it still elutes first regardless of the thermodynamics of dehydration.

Chris,
OK, if I understand this correctly you are saying that the energy of hydration + the energy of ammonium-halogen ion interaction is such that it favors F- to be in the mobile phase in relation to Cl-? If this is so there must be some kinetics involved as I can´t see pore size influencing this equilibrium. If this thinking is correct one should get results depending on flow rate for the highly cross linked species.
This is of interest to me as there are cases here where ion exchange in bulk (rather than in flow mode) is advantageous. The thought is that cross linking might be irrelevant for bulk separation of small ions?

Another way of looking at this issue is to examine the equilibria involved in the ion exchange reaction, which is

solution ion X(solvated) + site ionY(solvated) =

solution ion Y(solvated) + site ion X(solvated)

Looking at the reaction this way makes apparent the great importance of ion and site solvation.

During the exchange both the solvation of X and Y change, maybe substantially, and I suppose the site solvation could change as well. How much (delta H and delta S) depends on the size and charge of X and Y as well as the details of the ion exchange site. The details of the ion exchange site depend, among other things, on the free volume of the polymer(composition and crosslinking), as well as any other factors that influence the ability of the site to accomodate solvation of the counter ion.

Maybe Chris can shed further light on how column manfacturers diddle with the composition of the resin beyond crosslinking to affect site solvation.