Supercritical Fluid Chromatography Seeks New Users

Supercritical fluid chromatography, or SFC, has been around for decades. During that time, the separation technique has fallen in and out of and back in favor. But it has consistently been relegated to niche status. Could that finally be changing? Instrument manufacturers certainly hope so.

In the past several years, instrument companies Agilent Technologies and Waters Corp. have both brought out new analytical SFC instruments. The companies are optimistic that these instruments, which are more robust and more sensitive than previous generations of SFC models, will attract new users. They see SFC’s potential going far beyond the large, preparative-scale purification of chiral drug compounds, where the method long ago proved its value.

Indeed, the method is gaining ground in other types of separations, including small, analytical-scale ones of achiral compounds. But other experts question whether the improvements will be enough to give SFC staying power this time.

In SFC, carbon dioxide dominates the mobile phase that carries a mixture’s components through the column where they are separated. The CO₂ is usually combined with smaller concentrations of a modifier, which helps target analytes dissolve in the mobile phase. Modifiers are typically organic solvents; the most common one is methanol.

In true SFC, as the name suggests, that CO₂ is supercritical, which means that it’s at temperatures and pressures at which there are no phase boundaries between the liquid and gas phases. But most of the time SFC is a misnomer.

“Technically, we rarely do supercritical fluid chromatography,” says SFC expert Larry T. Taylor, an emeritus chemistry professor at Virginia Tech. “Just about every application I’m aware of uses modified carbon dioxide. The temperature is usually near ambient. Under those conditions, the mobile phase is not supercritical.”

SFC expert Pat Sandra, president of the Research Institute for Chromatography, located in Kortrijk, Belgium, and Lille, France, agrees. “We are exploiting the unique properties of carbon dioxide but not the unique properties of supercritical fluids,” he says. “It’s not SFC, but it’s working.”

“The terminology has probably held back the acceptance of the technology,” Taylor says. “In the 1980s, people thought the word supercritical implied a dangerous situation in the laboratory. Maybe that terminology wasn’t very useful. Those who work in the area just call it SFC for lack of a better term.”

Nowadays, CO₂ should be considered just another mobile phase for liquid chromatographic separations, adds Martin Vollmer, Agilent’s marketing director for analytical liquid chromatography and preparative LC, in Waldbronn, Germany.

The majority of LC separations come in two flavors: normal and reversed phase. In both versions, a pump is used to force a mobile phase through a column, so the methods are known as high-performance LC, or HPLC. Normal phase is the original incarnation of LC, in which compounds are separated using silica stationary phases and nonpolar solvents such as hexane or heptane. Normal phase typically separates compounds on the basis of their polar functionality. It is usually performed with an unchanging mobile-phase composition.

In contrast, reversed-phase chromatography uses water mixed with a polar solvent and separates compounds on the basis of their nonpolar functionality. The ratio of the water and polar solvent changes with time, resulting in a solvent gradient. Such gradients speed up overall separation time.

SFC brings the power of gradient separations to normal-phase separations. “In the same way that water is a universal solvent that makes gradient separations work in reversed-phase HPLC, CO₂ is a universal solvent that makes gradient separations work in normal phase,” says Mark Baynham, program director of UPC² separations technology at Waters.

“Normal phase is much more powerful than reversed phase,” Baynham says. “For structurally similar compounds, it’s better than reversed phase. For a wider range of compounds, it’s better than reversed phase.”

And SFC is “normal phase on steroids,” says SFC guru Terry Berger of SFC Solutions. SFC separations can be much faster than reversed-phase HPLC. They also require lower pressures.

In ultra-high-performance LC, stationary phases with particle diameters less than 2 μm require high pressure to force the mobile phase through the column. By contrast, in SFC with similar particle sizes, pressure higher than 400 bar—a pressure associated with conventional HPLC rather than UHPLC—is rarely needed.

“With just a little bit of training, anybody who does LC nowadays will also be able to successfully do SFC,” Agilent’s Vollmer says.

But those same folks might find SFC requires a little more method development than the reversed-phase LC they’re used to. Most reversed-phase separations are done on variations of the octadecylsiloxane-bonded—called C18—column with a simple gradient of water and a polar organic solvent such as acetonitrile, methanol, or tetra­hydrofuran. Although this combination doesn’t work for everything, it works well enough for most compounds that it’s the default method.

There is no such one-size-fits-most solution in SFC. “With SFC there are more options” to test, says Geoffrey Cox, president of PIC Solution, an SFC company based in Media, Pa., and Avignon, France. “You need to run a sample through a screen of several columns and several solvents in order to be comfortable that you’ll actually have a separation you can use.”

“We need the universal stationary phase for SFC,” says Christopher J. Welch, senior principal scientist in process and analytical chemistry at Merck & Co., Rahway, N.J. “The stationary phase that rivals C18 is not there yet.” Researchers at Merck are collaborating with Myung Ho Hyun of Pusan National University in South Korea to develop SFC stationary phases.

One application for which the power of SFC has long been recognized is the separation of racemic mixtures into their enantiomers. When Berger started his own SFC company in the mid-1990s, the eponymous Berger Instruments (which was later bought by Mettler-Toledo, Thar Instruments, and ultimately Waters) focused its efforts “on the one niche where it was absolutely obvious that SFC just killed LC—chiral separations,” Berger says.

The company quickly realized that SFC-based chiral separations could be scaled up, and SFC has since come to dominate such preparative-scale separations and purifications in the pharmaceutical industry.

“Chiral separations were always normal phase, so it was easy to switch to SFC,” says Holger Gumm, managing director of Sepiatec, a separations company in Berlin. “If you work in chiral separations, you must have SFC,” says Sandra of the Research Institute for Chromatography.

“I think normal-phase LC has had its time,” Baynham says. “To continue using it for chiral separations when the alternative is so compelling is ludicrous.”

But even in chiral separations, normal-phase LC has been surprisingly tenacious. “We amazingly find customers still doing normal-phase LC for chiral work,” says D. J. Tognarelli, a chromatography product specialist at Jasco, another SFC manufacturer. But, he adds, those users of normal-phase LC are becoming increasingly rare.

Researchers at pharma companies and instrument manufacturers are examining whether SFC can make similar inroads with achiral separations and purifications. Such separations are typically done with reversed-phase chromatography, Welch says. “To get your sample back, you have to dry it down, usually overnight. There’s half a day for dry-down. In SFC with just CO₂ and methanol, you can dry down in 30 minutes or less.

“All other things being equal, if SFC could do those jobs, the game would go to SFC,” Welch continues. But the game has not yet been won. Even those who already use SFC for achiral separations continue to use conventional reversed-phase HPLC as well because SFC separations don’t work for everything, Welch says.

SFC’s inability to separate all compounds equally is not the only problem that has hindered its wider adoption. A major barrier to acceptance of analytical SFC in the past was poor instrument sensitivity. That problem could be attributed to the compressibility of CO₂ and to the pumps, Berger says.

In previous generations of instruments, the pumps caused noise in the chromatogram baseline that Berger describes as awful. “It was all due to the compressibility of the fluid. The fluid was being pumped accurately, but you were getting pressure and flow spikes on every stroke,” he explains. Those spikes caused refractive index changes that wreaked havoc on the ultraviolet detector and translated into noisy chromatograms.

This noise has been greatly reduced in the newest analytical SFC instruments. Though they vary in the details, Agilent and Waters both use multiple pumps to separate the compression of the CO₂ from the mixing with the modifier solvent.

Agilent’s instruments—the 1260 Infinity Analytical SFC system and the 1260 Infinity Hybrid SFC/UHPLC system—use an add-on component originally developed by Aurora SFC, a company started by Berger and subsequently purchased by Agilent. “Our box has its own pump that precompresses the fluid to just below the pressure of the metering pump,” Berger says. “We’ve separated metering from compression. That eliminates all that noise.”

In the Waters instrument, the Acquity UltraPerformance Convergence Chromatography (or UPC²) system, the compressible CO₂ and the noncompressible modifier each has its own pump. The two are mixed only after the CO₂ has been compressed and both have been individually metered. The strategy allows the instrument to achieve precise gradients, Baynham says.

But it remains to be seen whether these improvements in sensitivity and precision are going to translate into new quantitative applications. For example, an area in which SFC has never been used is quality control. QC is an area where U.S. Pharmacopeia methods dominate.

Many USP methods are based on normal-phase LC, which would seem to make them ripe for conversion to SFC methods. The LC methods can take as long as two hours, depending on the compounds and the conditions. With SFC, “every single one is usually done in two minutes with a generic gradient,” Baynham says. “The fact that you can take QC from 90 minutes to two minutes means that you can do more batches.”

Berger is skeptical, however, that SFC will find its way into QC quickly. Because the existing methods work, many organizations may not find it worth the effort and expense to validate new methods. “If they’ve got an awful method that’s really tough by LC and it works well by SFC, they might switch over, but I don’t see any other compelling reason.”

Some people hope that SFC’s reputation for being environmentally friendly will give it the boost it needs to garner more use in QC and other quantitative applications. “When you look at it as a normal-phase replacement, the green bit really stands out,” Baynham says. Because it obviates the need for solvents such as heptane or hexane, SFC results in a significantly smaller carbon footprint. Plus, it eliminates other toxic solvents such as dichloromethane and tetrahydrofuran, which are used in reversed-phase HPLC.

“SFC took off in the pharmaceutical development labs,” Cox says. “People were able to do their purifications more quickly than with HPLC and using much less solvent. In the early days, that meant that you had less solvent to get rid of when you evaporated it. These days, with the emphasis on green chemistry, you get lots of ‘greenie points’ if you’re using CO₂ rather than hexane as your solvent.”

Proponents of SFC now see application areas outside the pharmaceutical industry. For example, Taylor has collaborated with scientists at Waters to use SFC to separate components of biodiesel, particularly fatty acid alkyl esters and impurities such as tri-, di-, and monoacyl­glycerols, free fatty acids, and glycerol itself. Other applications are in food and vitamin analysis.

But some applications are not practical for SFC. Sandra notes that most biomolecules, which are often polar or ionic in nature, aren’t easily solubilized in SFC’s CO₂-based mobile phases. Peptide separations with SFC have been reported in the literature, but such separations are limited to mixtures containing just a handful of different peptides.

For example, Tognarelli separated a five-peptide mixture for a customer who was already using SFC for other applications. “It worked pretty well and didn’t take a lot of effort, but the SFC didn’t really offer any more speed than reversed-phase HPLC,” he says. “I told the customer: ‘You’re going to want to stick with HPLC for peptides.’ ”

That conclusion doesn’t surprise Sandra. “You will never convince people in life sciences, in proteomics, in genomics, not even in metabolomics, to prefer SFC over other techniques,” he says.

Moving forward, the biggest hurdle to SFC adoption is education, Berger says. “Who teaches it?” he asks. “Nobody. In the U.S., I doubt there are more than two dozen—and I bet it’s less than one dozen—SFCs operating in academic laboratories.”

Taylor also worries about academia ignoring SFC. “Academic labs have not embraced SFC for the last 20 years,” he says. “It’s disturbing to me that nobody is being trained at the undergraduate and graduate level in this type of technology. In order for it to really take off, we’re going to have to have considerable training so that people will have the same comfort with SFC as exists with LC. It’s depressing that there are companies who want to hire in this area, but people aren’t available.”

Even if the new generation of SFC instruments is adopted, it’s possible that SFC will remain a niche technique. “Eighty percent of the chromatography market is reversed phase,” Sandra says. Of the other 20%, “normal phase can only be about 5%, and maybe half of that will be SFC. The market will be small. I’m not pessimistic. I’m just realistic.”

Berger, however, thinks SFC has the potential for a much bigger market share. The improved sensitivity of UV detection with SFC will have a significant impact on the broader QC market over time, he says.

But he still worries that this wave of the technique could recede like the others before it. The large companies have shown their willingness to jump in and out of the field. “I am hopeful but have been disappointed in the past,” he says.

Still, company reps remain optimistic. Waters is already seeing major accounts at global pharmaceutical companies purchase additional instruments. “If it’s good technology, they’re the ones who should be buying multiple units the fastest,” Baynham says. Quite a few companies have bought in the double digits. “In two years, that’s not bad,” he adds.

“I’m convinced that it will stay because we really see uptake and interest,” Agilent’s Vollmer says. “It’s perhaps a little slower than anybody expected, but we see a lot of interest even from fields or markets where we didn’t expect it.”