Separations Anxieties

By Paul Thomas, Managing Editor

Chromatography is a continuously evolving science, and not a month goes by that some column or media manufacturer doesn’t tout the “next big thing.” In his 30 years in the business, Tom Jupille, founding consultant of Separation Science Associates, has seen enough to distinguish between the wheat and chaff. Jupille, who also presides over the eminently popular Chromatography Forum (www.chromforum.com), agreed to answer some of the pressing chromatography questions of the day.

P.M.:What have been the true innovations in HPLC equipment over the past year or so?T.J.: On the analytical side, the continuing trend is “faster and smaller.” The efficiency of HPLC (measured by the infamous “theoretical plates”) has been roughly constant for a generation. The difference is that we are now generating 10,000 plates in 3-minute separations, compared to 30 minutes a couple of decades ago. This trend is continuing with the jump in operating pressure capability from about 5,000 psi to 10-15,000 psi over the past couple of years.P.M.:Are you seeing more interest in chromatography for small molecules? What technologies (simulated moving bed, etc.) are driving this interest? What is limiting it?T.J.: Process-scale chromatography of small molecules has always been a last resort, in large part because of the costs of the solvents required (chromatography will typically consume 5 to 10 times more solvent than something like a recrystalization or an extraction). There is some evidence that process-scale supercritical fluid chromatography (SFC) is changing the rules of the game by allowing the use of cheap, “green” mobile phases based on CO2. SFC is proving itself today in chiral purification; it remains to be seen whether increasing familiarization with the technology will extend that to other applications.P.M.:Are new silicas significantly better than old ones? Which stand up best?T.J.: Yes. The big (conceptual) improvement occurred a decade and a half ago with the realization that the purity of the underlying silica (especially with regard to residual heavy metal ions) had a tremendous influence on the chromatographic properties. The resulting “type B” (basic) silicas have now become the standard. While the materials continue to get better, the improvements are evolutionary rather than revolutionary.

The exception to that last statement is the development of “hybrid” silicas which incorporate a certain amount of alkyl functionality in the silica backbone. That provides a material with higher stability at high pH and overcomes what has been a major limitation of silica-based column packings.

P.M.:Disposables are catching on, but what’s their future?T.J.: If I could accurately forecast the future, I’d be rich! That said, analytical HPLC columns, at $300 or so each, are actually “disposable” in the sense that it is usually more cost-effective to replace them rather than repack them when they die. If you assume a typical column lifetime of 1,000 injections, that works out to about 30 cents per sample in column costs.

If, as I suspect, you mean “one-shot” disposable: unless someone comes up with a 30¢ column, I think that their use will be confined to the “upstream” (sample prep) side of the analytical process: vials, solid-phase-extraction (SPE) cartridges, membrane filters, etc.

P.M.:Are membrane adsorbers up to par with traditional columns? What are their limitations?T.J.: They’re different animals. One way of looking at HPLC is as a technology that can reliably deliver separation efficiencies of 10,000 plates. Without bogging down in the math, this means that an HPLC column can completely separate two compounds whose chemical interactions (with the chromatography system) differ by about 10%. That’s one extreme. At the other extreme, if you have large differences in chemistry, you can do the separation with one “plate” using standard techniques like liquid-liquid extraction. Membranes occupy a middle ground of sorts, offering a few tens of plates. P.M.:What’s the latest you’re hearing regarding the enhancement and monitoring of column performance and retention? Is there a fountain of youth?T.J.: Columns are cheap; time is valuable. In my experience, “consumables” (solvents, columns, vial, filters, etc.), in total, amount to less than 25% of the cost of “people.” There are all kinds of “fountain of youth” column revitalization tricks, but most of them take more time than they are worth.

Monitoring is a different story. Any validated pharmaceutical method has “system suitability” specifications which must be met before results from the method can be accepted. Control charting of system suitability results is common sense for any method which is run regularly.

P.M.:Why is it that manufacturers of silica-based columns do not recommend using low-pH mobile phases, such as those with TFA (trifluoroacetic acid)?T.J.: The majority of reversed-phase HPLC columns use a “silyl ether” (Si-O-Si-) linkage to attach the bonded phase to the underlying silica. That Si-O-Si bond is subject to acid-catalyzed hydrolysis (the stoichiometry even works out: Si-O-Si + H2O → SiOH + SiOH). The hydrophobic bonded phase shields the bond to some extent, but the lifetime of most columns will drop as the pH gets much below 3. TFA is a strong enough acid that it drops the pH down to the low 2s. There are ways of shielding that silyl ether (for example, using hydrophobic side chains on the bonded phase), so some columns are reasonably stable down below pH 2, but the consequence is often decreased stability at high pH.P.M.:Manufacturers often recommend column storage in pure acetonitrile (ACN). Why? How about pure methanol or methanol-water as storage solvents?T.J.: The argument is that ACN is aprotic and therefore potentially less corrosive. I have personally not found that to make any difference (but then, I haven’t exhaustively studied the problem). P.M.:What’s the danger in flushing with ACN if buffers have not been thoroughly washed out of the column? What could happen?T.J.: Many of the buffers used in HPLC (potassium phosphate/phosphoric acid, for example) are only sparingly soluble in ACN. If you’ve ever had buffer precipitate in your system, you know that the results aren’t pretty (ask me how I learned that!), so it’s better to be safe. That said, in most cases, the buffers are sufficiently dilute (10 – 50 mM or so), that a quick flush with 10 column volumes of solvent (that would be about 15 mL for a standard analytical column) should suffice.P.M.:What sort of mobile phase do you recommend to combat microbial and bacterial growth during periods of storage?T.J.: That’s the topic of a recent discussion on Chromatography Forum. My experience is that anything over about 5% organic solvent (either MeOH or ACN) is pretty much bacteriostatic. Every now and then I hear anecdotes about bugs adapting to that or actually growing inside HPLC pumps, but those have always happened to “an acquaintance of a friend,” and I’ve never been able to confirm from the original source.

As for columns, the safest procedure is to find out what the column was filled with when the manufacturer shipped it to you, and use that. This is not necessarily the most cost-effective, however, especially if the column required extensive re-equilibration before it could be used. “Time is valuable; columns are cheap,” and it can make sense to sacrifice column lifetime on the altar of quick usability.

About Tom Jupille