Chromatography Forum

Hi

We have a method for determination of impurities using RP-HPLC, previously we used grace vydac C18 column,4.6 X 250mm ,300A and had 3 related, well resolved impurities

Recently we shifted it to a different brand of column with same column specification (in this case “ACE-C18 4.6 X 250mm ,300A) after this change we found that the number of impurities resolved were much higher than pervious column.i.e we have 7 related well resolved impurities in comparison with old column which had 3 impurities.

I would like to know what is the reason for such an observation..any one else have same experience??..one of my colleague told me that the purity of silica may be the reason behind such a change..is it true..? if so how do I investigate it..? (I would like to know do you look into specification such as purity of silica when choosing a column?if so how??)

Thank you

any one else have same experience

Everyone has had that experience.

The selectivity of a reversed-phase column is affected by (among other things):
- the purity of the underlying silica
- the "ligand" (e.g., C18, C8, cyano, etc.)
- the pore size of the packing (e.g., 300 Ã…)
- the pore size distribution
- the specific surface (m²/g) (this is a related to the pore size and pore size distribution)
- the surface coverage by the bonded phase (μmoles/m²)
- whether the reactive silane used for bonding was mono-, di-, or tri-functional
- what was the leaving group on the silane (e.g., Cl-, MeO, EtO, etc.)
- whether the bonded phase is a monomeric layer or polymeric (this is related to the type of silane used)
- the side groups on the silane (for mono- or difunctional silanes).

You have the "same specification" on two of those parameters.

I looked at the PQRI database of column selectivity similarities and differences on the USP web site:
http://www.usp.org/USPNF/columnsDB.html
The Ace 5 C18 300 does not show up in the 10 closest matches to any of the Vydac columns (by the way, you did not say which Vydac column you are using; I'll assume 201TP). I was involved in the original study and still have a copy of the actual database, so I went in and looked a bit deeper. The Ace 5 C18-500 ranks number 205 in similarity to the Vydac 201TP (out of 360 columns in the database).

Is the difference due to the purity of the silica? That is almost certainly a contributor; the Vydac 201TP has been around for decades while the Ace is based on much newer technology, so you would expect it to be a better column in general. That is probably not the only difference, however. Note that "better" is not always a good thing. There are many methods that only work correctly because of impurities in some of the older-technology (so-called "Type A") columns.

The lesson to be learned from all of this is that you cannot realistically expect to substitute one C18 column for another without seeing changes in the method. If you must do it, the similarity databases on the USP website can provide a more accurate choice of a "similar" column. Those databases use standard text mixes to actually measure the selectivity of the columns, rather than relying on trying to sort out interactions among the dozen or so parameters that control it.

When you start comparing Vydacs, you also have to consider particle shape in addition to the other properties listed. Vydacs of old frequently employed silica particles that were not at all close to the nice, uniform spheres we expect current columns to be packed with. I'm not sure if they use "odd" shaped particles in newer offerings or not.

Hi

Frist of thanks to DR and tom jupille reply….
Well both replies have put new dimensions…could you tell me how the following factors actually contribute in difference in impurity profile??

(a) shape deformation of silica (as suggested by DR)
(b) the pore size distribution
(c) what was the leaving group on the silane
(d) the purity of the underlying silica

Shape usually affects resolution and results in broader (irregular shape particles) or narrower (spherical, even particles) peaks, since is much easier to pack golf balls than Lego bricks!

Pore size distribution has the same rationale: even particles provides better and more reproducible results.

Purity of silica has to do with interactions of impurities with your molecules and the uneveness it might cause to the phase bonding.

And, finally, I didn't get the "leaving group" question!!!

A lot has to do with the chemistry of (and access to) the underlying silica surface. The following is a huge oversimplification:

Shape (and particle size) pe se, probably has little if any effect on selectivity, but does have a big effect on efficiency (plate count). Smaller particles = more plates = better ability to resolve compounds given the same selectivity.

Pore size affects:
- efficiency (especially for larger molecules)
- retention (larger pores = lower surface area = less retention)
- selectivity (an octadecyl group is about 30Ã… long -- assuming about 1.5Ã… for a C-C bond; for a small pore size, say 100Ã…, a substantial fraction of the pore is filled with stationary phase, which limits access to the underlying silanols)

The details of bonding chemistry (including whether the leaving group was chlorine, methoxy, or whatever) will affect the fraction of the silanol groups derivatized, as well as whether the stationary phase is a monolayer or not.

The purity of the underlying silica (especially with regard to residual heavy metal cations) will affect the pKa and pKa distribution of the silanols.

All of these factors have been the topic of papers, books, and PhD theses for the past three decades. They interact in ways that make it virtually impossible to use "seat of the pants" judgment in matching columns. That's why the databases I referred to earlier were developed.

Hi

First..thank you u all for the contribution

Well I just went back and browsed through the different column catalogue. And found that most of the catalogue doesn’t give information about purity of silica (most of them claim to use HIGH PURITY SILICA), what is the leaving group on silica (tom mention this on his answer), the specific surface (m2/g) and the surface coverage of bonded phase (µmoles/ m2)…

Is there a way to check all this for a particular brand of column??

…and what r the parameters all of you look at before deciding on a particular brand of column ( I am talking about situation when same type of column is available from different brands )

What would you do with that information if you had it?

If you want more detailed information, I'd suggest you get a copy of Uwe Neue's book on HPLC columns.

With regard to "what to look for", a reasonably new-technology ("type B") silica based C8 or C18 column is a good general-purpose starting point for operation at pH 2 - 8. For higher pH, start with one of the "hybrid silica" based columns. Column chemistry is one of about 6 parameters that will affect selectivity. If you can't get the selectivity you need, then switch to as different a column as you can find (you can use the databases on the USP web site for possibilities); generally, this will be an "embedded polar group" or a cyano column.

Once a method has been developed and validated on a particular column, it is specific to that column. The similarity / differences data bases will point you to the most nearly equivalent column, but there is no guarantee that you will get the same results without at least tweaking some of the other parameters. If you want to get into more of the details, the Snyder, Glajch, and Kirkland book Practical HPLC Method Development is probably the best reference.

I apologize if it sounds like I am doing nothing but giving book references; the questions you are asking are excellent ones, but they do not have simple answers.

In principle, there is no reason for hiding relevant information. I know that web sites tend to be less informative than catalogues, and I am also aware that the information in some catalogues is rather incomplete. Anyway, I can find all relevant information that I need in the Waters catalogue. There is another source of information that you can consider, and this is the certificate of analysis that comes with the columns. It tends to give you even more precise information than what you get in a catalogue.

With respect to your question on the purity of the silica: This was a big deal in the early to mid 1990s, when people discovered that bonded phases based on a high purity silica work much better than than those based on the older silicas. Within a few years, all silica manufacturers that wanted to stay in business had created a high-purity line.

The difference between a high-purity silica and one of the older silicas is really the raw material. High-purity silicas are made from an organic silane, and with proper precautions, the high purity of such a silane can be maintained in the process of making silica. I am sure that there can be differences in the level of "high purity", but there is a night-and-day difference between any high purity silica and a classical silica.