Stuart R. Gallant, MD, PhD
This post is the first in an occasional PharmaTopoTM series called “Technology Focus.” Today, we look at supercritical fluid chromatography (SFC). The most important point in this post is that if you have a background in liquid chromatography, most of the ideas in this post should be fairly intuitive. In other words, it’s not a huge leap to begin working in supercritical fluid chromatography.
SFC Fundamentals
SFC is a gradient chromatography technique in which supercritical carbon dioxide forms the low strength eluent (in gradient lingo “Solution A”). The carbon dioxide is pressurized to a condition near its critical point. In this condition, carbon dioxide starts to behave more like a liquid than as a gas:
The high strength eluate (in gradient lingo “Solution B”) is known as the “modifier” and the modifier may be supplemented with an “additive.” First, the modifier… Lesellier and West have a nice figure to demonstrate the effect of a modifier [1]:
Some observations regarding this figure are:
- Most of the solvents on the right side aren’t very green. Using Prat’s CHEM21 selection guide [2], some of the solvents (diethyl ether and chloroform) qualify as “highly hazardous” or (dichloromethane, hexane, and 1,4 dioxane) as “hazardous.” The remainder are “problematic,” except the alcohols. Methanol, ethanol, and isopropanol qualify as “recommended,” and of course water. Fortunately, the alcohols are quite useful.
- Focusing on the green methanol curve which plots Nile Red solvation strength versus percent methanol, we see that a small percent of methanol has a big effect. 10% methanol reduces ENR from 61.5 to 55.5—essentially ½ of the total methanol effect occurs with the first 10% of methanol added.
For these reasons (greenness and strong effect on retention), methanol is widely used as a modifier in SFC.
So, now that we have discussed the modifier, what about the “additive?” Additives, have secondary effects—for example improving tailing by acting as modulators of secondary interactions with the stationary phase. For example, the elution peak of an acid sample may be improved with a small amount of trifluoroacetic acid.
Columns
A good way to organize the possible column choices is based on polarity. A range of octanol-water partition coefficients are given as an example:
| log KOW | |
| Hexane | 3.13 |
| Dibutyl Ether | 2.77 |
| n-Butanol | 0.81 |
| Acetonitrile | -0.17 |
| Acetic Acid | -0.22 |
| Dimethylformamide | -0.63 |
| Water | -0.65 |
| Formamide | -1.08 |
After extensive testing, Hirose and developed the following guidance regarding column chemistry for SFC, based on log P (i.e., log KOW) [3]:
The essential point is that more nonpolar station phases are used with more nonpolar compounds, and more polar stationary phases are used with more polar compounds. This point is discussed further below in the method development section.
Because the viscosity of supercritical CO2 is low, expect that the pressure drop across the column will be lower at the same flow rate than in liquid chromatography. As a result, the column can be run faster. At the same time, the peaks will be more efficient (narrower). This can be seen in the following van Deemter plot typical of HPLC and SFC performance.
As can be seen in this plot which shows HETP data for 10 micron chromatographic media, the most efficient HPLC peaks are obtained in the plot at linear velocity of 0.2 cm/s. SFC can obtain the same peak efficiency at a linear velocity of 0.8 cm/s. So, the SFC system can be run 4 times as fast and deliver the same separation.
Equipment
The equipment required for SFC is relatively simple. The flow diagram is quite similar to that of a gradient liquid chromatography system:
Carbon dioxide is supplied either in pressurize cylinders or in dewars. Compressed gases are frequently used in liquid chromatography. The major difference in this application is that the volumes required are higher, and particularly for preparative SFC, dewars can have substantial economic advantages.
The supercritical or near critical carbon dioxide is supplied to Pump A in a two-pump system with the modifier and any additive supplied to Pump B. The sample may be injected using a traditional loop injector (as depicted in the illustration) or a dedicated pump may be employed for larger volume loads The column is held in an oven to prevent convective cooling of the column surface (and loss of column efficiency due to band broadening caused by radial temperature gradients). Many different types of detectors are possible (for example, UV if the molecules to be separated have chromophores or ELSD if not). The fraction collector may be a grid type or valve type collector, but pressure must be maintained on the fractions until they come to atmospheric pressure. Waste carbon dioxide can be discharged to a fume hood. Because large quantities of carbon dioxide are involved, a carbon dioxide detector should be placed close to the instrument for safety purposes in case a leak occurs.
Method Development
Putting it all together, here’s a good routine for method development:
- For non-polar compounds, consider C18 and CHO columns.
- For polar compounds, consider 2-ethylpyridine, cyano, diol, and bare silica columns.
- A methanol gradient from about 4% to 40% will allow the column to be evaluated for possible retention windows. Better to always keep a small amount of modifier in the column to improve run to run reproducibility.
- 1/10 to 1/2 percent additive (TFA, isopropylamine…) in the modifier may help with tailing
Once you have a reasonable starting point, further optimization can focus on smaller changes in column properties, temperature, and other refinements to improve selectivity. All of this can be evaluated relatively quickly because each run will be about 4x as fast as a similar HPLC run. The biggest challenge can be keeping up with the large amount of data that can be generated in a short period of time.
Advantages and Outlook
Having taken a quick stroll through SFC, let’s consider the advantages, compared to HPLC:
- SFC is greener. The quantities of solvents used are smaller than in HPLC.
- CO2 itself is green. It’s a byproduct gas of many industrial processes. It’s extraordinarily cheap compared to HPLC solvents—particularly when it is purchased in dewars.
- SFC is cheaper. Carbon dioxide is much cheaper as a solvent, compared with traditional HPLC solvents.
- Column life is substantially longer in SFC than in HPLC. HPLC columns eventually degrade and must be discarded. The same column used in SFC may not ever need to be discarded because contact with supercritical CO2 actually reconditions the column.
- SFC is about 4x faster than the equivalent HPLC separation.
SFC is still evolving:
- Industrial scale SFC is growing due to the favorable economics and greener nature of the technology. Major investments are being made to support pharmaceutical and high-value reagent production. Demand for equipment to support industrial processes will increase the user base in preparative and analytical SFC in the coming decades.
- In the last 2 decades, the quality of SFC equipment has been slower to improve than HPLC because of the smaller user base, compared to HPLC. I know of labs stockpiling Berger Instrument Multigram parts to keep these machines from the 1990s running because Multigrams are as good as any SFC machines available today. Expect this to change as more users move over to SFC for performance, economic, and environmental reasons. The larger user base will drive innovation in instrument design.
- More advanced SFC systems will reduce the impact of installation on the laboratory. Currently, high-pressure stainless-steel tubing is used to distribute CO2 in the laboratory. Next generation systems will used medium pressure tubing which is substantially easier and quicker to install.
All together, the future of SFC is bright.
[1] Lesellier, E. and West, C. “The many faces of packed column supercritical fluid chromatography – A critical review,” Journal of Chromatography A, 1382 (2015) 2–46.
[2] Prat, D., et al. “CHEM21 selection guide of classical- and less classical-solvents,” Green Chem., 2016, 18, 288.
[3] Hirose, T., et al. “Comparison of Retention Behavior between Supercritical Fluid Chromatography and Normal-Phase High-Performance Liquid Chromatography with Various Stationary Phases,” Molecules 2019, 24, 2425; doi:10.3390/molecules24132425.
Disclaimer: PharmaTopoTM provides commentary on topics related to drugs. The content on this website does not constitute technical, medical, legal, or financial advice. Consult an appropriately skilled professional, such as an engineer, doctor, lawyer, or investment counselor, prior to undertaking any action related to the topics discussed on PharmaTopo.com.