May 1, 1996

18 Min Read
SFE for Food Analysis:  Extracting the Hype from Reality

SFE for Food Analysis:
Extracting the Hype from Reality
May 1996 -- QA/ QC

By: Ray Marsili
Contributing Editor

  When the first commercial supercritical fluid extraction (SFE) instruments appeared in 1988, vendors promised the world. They touted the instruments as the ultimate in extraction technology, capable of extracting nearly any type of analyte from any type of matrix in just a few minutes without the use of expensive, toxic solvents.

  Unfortunately, in many cases, the instruments didn't deliver on the promises. Perhaps the most frustrating aspect was the tendency of the capillary restrictors to clog, resulting in substantial downtime.

  Although many of the problems with early instruments have been worked out and more published SFE applications for food testing are appearing in scientific journals, today vendors themselves are sending potential users mixed messages about the technology.

  Consider Hewlett-Packard. In the late 1980s, the company was years ahead of competitors when it developed the first automatic variable restrictor (which the company refers to as a "nozzle"), and it was the first to offer analyte trapping on solid-phase extraction media and other types of solid support materials. But the company has done little since then to further its SFE technology. Furthermore, the company has sold off its supercritical fluid chromatography (SFC) technology and is no longer making SFC instruments.

  Although Dionex still sells its SFE instrument, it no longer markets it aggressively. The company has developed an alternate instrument, the ASK(tm) 200 Accelerated Solvent Extractor, which it says is superior to SFE for most applications.

  "The ASK 200 was developed as a result of our experience with SFE," says Dale Felix, manager of Dionex's Salt Lake City Technical Center. "We soon realized co-solvents were more important than the supercritical carbon dioxide in affecting extraction."

  The ASK 200 uses traditional solvents, which it heats and pressurizes. Pressures are low (700 psi to 3,000 psi) compared with pressures used in SFE; this serves to prevent heated solvents from vaporizing and to push the hot liquid solvents through the extraction cells. Since no supercritical fluids are used, the instrument design is simpler than SFE instruments. For example, no restrictors are needed because no phase change from supercritical carbon dioxide fluid to carbon dioxide gas is involved in the extraction. Rapid and highly efficient analyte extraction using only a few milliliters of traditional solvents is accomplished because the solvents are heated. Elevated temperatures accelerate the extraction kinetics.

  So far, the ASK 200 has been used primarily for environmental analysis, but more and more food companies are showing interest. According to Felix, Nabisco, Nestle, and Procter & Gamble have ordered instruments.

  "We sold more ASK 200s in the first six months than we did SFEs in the first year they were offered," says Felix. The instrument costs approximately $44,000 and comes with 12 extraction cells. Twelve more cells can be added.

  SFE makers and some users say that although the technique used in the ASK 200 and similar instruments has merit with some environmental testing applications, it suffers from several significant disadvantages for food analysis. Extracts with the ASK 200 tend to be "dirtier" than SFE extracts. As a result, subjecting extracts to off-line analysis (e.g., GC or HPLC analysis) may not work without additional sample cleanup steps because of interferences.

  Furthermore, the technique is not as selective as SFE. For example, some SFE users claim they can fractionate cholesterol from triglycerides in food samples with SFE. The ASK 200 is more of a "brute force" technique that extracts cholesterol, triglycerides and other lipid material at the same time. In addition, accelerated solvent extraction techniques using hot solvents tend to decompose thermally labile analyses like fat-soluble vitamins.

Routine analysis, or research?

  Larry Taylor, Ph.D., professor of chemistry at Virginia Polytechnic Institute and State University, has focused his research efforts on SFE/ SFC applications for the past several years, and he is one of the most respected researchers in the field. Taylor, who uses instruments from all major SFE vendors, has published over 80 papers on the subject of SFE and SFC.

  Can SFE be used as a routine analyte extraction technique by food testing labs? According to Taylor, SFE method development can be time-consuming and complicated since there are numerous experimental variables to optimize, including flow rates, temperatures (of the extraction chamber, restrictor, etc.), pressure, type of modifier, type of analyte trapping technique, and so on. However, once an appropriate method has been developed, SFE analysis is a routine matter with today's robust instruments, he says.

  The problem usually isn't extracting the analyte from the matrix, according to Taylor. By selecting appropriate modifiers, chemists can extract most types of chemical analytes from food samples. Taylor has even been successful extracting highly polar compounds from samples. For example, ionic substances like drugs can be extracted from different types of samples by using ion pairing agents as modifiers. Taylor also can extract metals by using the appropriate chelating-agent modifiers. He says that in most cases the real culprit responsible for low extraction recoveries is inefficient trapping of the analyte once it's extracted.

  One problem is that modifiers can form aerosols and sweep analyte molecules through the system without the molecules being deposited or collected in traps. Low recoveries result. In some cases, tandem trapping techniques using one trap with a solid support and a second trap with a liquid phase are required for efficient recoveries.

  Taylor sees tremendous potential for fat analysis as well as many other types of food analysis by SFE. Choosing the proper system is important for good results. For fat analysis, Taylor says an automatic variable restrictor and modifier pump are crucial. For doing trace analyses like pesticides or fat-soluble vitamins, large sample extraction cells and automation features are important. (Automation is important because it allows users to more easily investigate optimum parameters for temperature, pressure, flow rates, etc.)

  Taylor is bullish on SFE. He predicts more hyphenated SFE techniques will be developed: SFE-GC, SFE-MS, SFE-FTIR, and so on. He also sees application potential for supercritical fluids other than carbon dioxide. For example, he has had good results selectively removing pesticides from meat samples while leaving behind fat. To accomplish this, he has used supercritical fluid fluoroform -- a recommendation first proposed by Jerry King, Ph.D., another highly respected SFE researcher at the Food Quality & Safety Research Unit, National Center for Agricultural Utilization Research, Agriculture Research Service, USDA, Peoria, IL.

  King and his group have recently developed an elegant procedure for analyzing fats in meats by SFE. King's technique goes beyond the normal gravimetric test in that it satisfies the new NLEA definition for fat -- i.e., it uses SFE coupled with GC to measure the individual fatty acids from mono-, di-, and triglycerides; free fatty acids; phospholipic fatty acids; and sterol fatty acids. Total fat is then calculated as the sum of all fatty acids, expressed as triglycerides, with the corresponding saturated and cis-monounsaturated fat content computed similarly from the sums of the appropriate fatty acid methyl esters associated with these types of fat.

  The new test is being evaluated in round-robin testing for approval as an official test method. Calculations of total fat based on fatty acid content usually result in lower total fat values compared to solvent extraction-based gravimetric test methods. And, of course, this makes the nutrition labels on food products appear more appealing to consumers.

  King says multi-residue pesticide analysis, fatty acid testing and total fat determination are now being routinely performed by SFE for most types of food products. Areas that are not so routine and require more research include testing for vitamins, cholesterol and flavor chemicals.

Well suited to pesticide testing

  Bob Epstein, Ph.D., deputy director at the Science and Technology Division of the USDA in Washington, DC, supports SFE for pesticide testing. His group is conducting an extensive study comparing SFE and conventional extraction for the analysis of pesticides in fruits and vegetables. Four state labs are involved in the project. Each lab will analyze different commodities: sweet potatoes (Michigan), apples (North Carolina), green beans (Texas), and oranges (California). According to Epstein, preliminary testing shows SFE provides results equivalent to conventional solvent extraction, but with far less operator labor.

  Two different SFE procedures must be used on each sample in order to analyze all the pesticides of interest, according to Epstein. One procedure works for non-polar pesticides, while the second technique must be used for polar pesticides, which are less soluble in supercritical carbon dioxide. One aspect that impresses Epstein is that the SFE extracts are so clean that no additional sample cleanup steps are required before injection into the gas chromatograph (GC). Off-line GC analysis is most often performed with ion-trap mass spectroscopy detection (electron impact mode). However, flame photometric detectors and other halogen-specific detectors also are used. Based on preliminary results, Epstein is confident that SFE will work. He believes that SFE eventually will replace solvent extraction techniques for most types of chemical analysis because the technique uses only minimal amounts of hazardous solvents, saves analyst labor, and costs less per analysis. He predicts traditional solvents like methylene chloride, chloroform, etc., will become more expensive because they are so detrimental to the environment.

  Steve Lehotay, Ph.D., a research chemist at the USDA in Beltsville, MD, is developing the SFE techniques for Epstein's study. Although methods have been developed, some "tweaking" is necessary for some types of samples. Sweet potatoes, for example, are a difficult matrix.

  Lehotay has successfully extracted 75 different pesticides from 12 different types of fruit and vegetable products using two different SFE techniques. He says SFE provides clean, selective extraction of pesticides without extracting other interfering chemicals from the food matrix. According to Lehotay, fruit and vegetable extracts obtained with the ASK 200 were less desirable, requiring more sample cleanup before they could be analyzed.

  Since many of Lehotay's samples contain 90% water, he binds up the free water in his samples with Hydromatrix prior to placement in the extraction cells. Details of his methods describing SFE of polar and non-polar pesticides in fruits and vegetables will be published in the near future.

  SFE is not a perfect technique, however. Lehotay is critical of the costs associated with the high purity carbon dioxide that must be used. He also would like to see more reliable parallel extraction systems developed instead of the sequential (one-at-a-time) instruments that are more commonly sold. The ability to analyze several samples at once in the parallel mode would save time.

SFE is Matrix-Dependent

  Tim Norden, a chemist at the Federal Grain Inspection Packers and Stockyards Administration, Kansas City, MO, has been using SFE for four years. He uses the technique to extract some 42 different pesticides from wheat, corn, barley, soybeans and other grains prior to GC. He has reduced analytical time considerably by rigging up Isco Inc.'s SFE equipment to run six sample extractions in parallel.

  Unlike Lehotay, Norden says his SFE pesticide extracts require further cleanup before they can be analyzed by GC or HPLC. Also while Lehotay has had success with polar pesticides, Norden has had only limited success -- even with modifiers. Norden prefers to use accelerated solvent extraction for polar pesticide testing.

  Why the differences between Lehotay's and Norden's SFE methods for pesticide analysis? One major reason probably is that the two chemists are analyzing significantly different types of foods. Norden's samples are grains which contain considerably more oil than Lehotay's fruits and vegetables. The results of the two chemists illustrate an important fact: Success with SFE not only depends on the types of analyses but also on the sample matrix involved. So, try before you buy an instrument.

  Today's SFE instruments are reliable, rugged, user-friendly, and capable of delivering on many of the promises that were once made. Companies like Suprex Corp., Isco Inc., and Applied Separations have remained committed to the technology and have made remarkable upgrades and improvements in their instruments in the past two years. For example, Suprex, the only company whose sole business is supercritical fluid extraction/chromatography, sells seven different SFE models, all optimized for specific types of applications. The company has targeted the food industry as the primary market for its instruments.

  Most users say the nightmares associated with chronic clogging of restrictors are a thing of the past. Today, food chemists can select from a wide range of dependable SFE instruments, ranging in price from under $16,000 to over $80,000, depending on the degree of automation desired, trapping techniques required, etc. While the instruments are not suited to every type of analyte or sample matrix, food chemists have successfully used the instruments to replace laborious and dangerous solvent extraction methods for a variety of food testing applications, including total lipids in snack foods and meat products; fat soluble vitamins (including beta-carotene) in fruits and vegetables; pesticides in fruits, vegetables, grains and meat; aflatoxins in corn and grains; and flavor compounds in a variety of food materials.

  If you've been waiting for SFE to become a more reliable, user friendly technique before purchasing an instrument, the time is right to start evaluating potential candidates.

Examples of Published Articles on Using SFE for Food Analysis

  • "Supercritical Fluid-Based Cleanup Technique for the Separation of Organochlorine Pesticides from Fat," John France, Jerry King and Janet Snyder, J. Agric. Food Chem., 1991, 39, 1871.

  • "Extraction of Fat from Ground Beef for Nutrient Analysis Using Analytical SFE," Jerry King, et al., J. Agric. Food Chem. (in print).

  • "Enhanced Supercritical Carbon Dioxide Extraction of Pesticides from Foods Using Pelletized Diatomaceous Earth," J. Assoc. Off. Anal. Chem., 1991, 74, 661.

  • "Development of Analytical SFE of a Polar Drug From an Animal Food Matrix," D.C. Messer and L.T. Taylor, J. High Resol. Chromatogr., 1992, 15, 238.

  • "Extraction and Detection of Sulfamethazine in Spray-Dried Milk," S. Malik, S. Duncan, J.R. Bishop, and L.T. Taylor, J. Dairy Sci., 1994, 77, 418.

  • "Comparison of a Liquid Solvent Extraction Technique and SFE for the Determination of Alpha- and Beta-Carotene in Vegetables," Ray Marsili and Dan Callahan, J. Chromatographic Sci., 1993, 31, 422.

  • "Determination of PCBs in Fish Tissue Using SFE," Robert Hale and Michael Gaylor, Environmental Science & Technology (in print).

  • "Analysis of Wine Aroma by Off-Line and On-Line SFE-GC," Gracia Blanch, Guillermo Reglero and Marta Herraiz, J. Agric. Food Chem., 1995,43, 1251.

    SFE: How it Works

      SFE instruments extract analyses from foods and other matrices by using supercritical fluids, usually carbon dioxide, as the extracting solvent. Carbon dioxide is an ideal candidate to use as a supercritical fluid because its supercritical temperature is near ambient (31.3°C) and its critical pressure is moderate (72.9 atm or 1,070 psi). It is nontoxic and inexpensive. Pressurization of carbon dioxide is accomplished with a pump (syringe or reciprocating dual piston) similar to an HPLC pump.

  • Modifiers and modifier pump. Supercritical carbon dioxide has solubility/dissolving characteristics similar to liquid hexane but with the penetrating properties of a gas. To enhance extraction of more polar components, small amounts of modifiers (e.g., ethanol, methanol, ethyl acetate, etc.) are usually required in addition to the supercritical carbon dioxide.

      Modifier addition can be accomplished in three ways: ] ) by adding it (usually 500 microliters or less) directly to the sample in the extraction cell or by mixing it with the sample prior to loading in the extraction cell; 2) by using carbon dioxide gas cylinders containing premixed amounts of added modifier; or 3) by adding it in a continuous fashion with an external modifier pump. In general, the modifier pump approach is the most desirable and convenient way.

      In many cases, recovery of analyses from food samples is significantly improved with modifier addition.

  • Extraction vessels. Solid or semi-solid samples are placed in the extraction vessel. While liquid samples are problematic, some researchers have been able to extract analyses from liquid samples by first mixing them with Hydromatrix, sodium sulfate or some other solid support prior to loading into the extraction cell. Large extraction vessels (i.e., 20 milliliters and larger) are advantageous because they accommodate the larger sample sizes which are needed for trace analyte analysis. AISO, in the case of liquid samples which are diluted with substantial amounts of solid support to absorb free moisture, larger extraction cells are a necessity.

      In recent years, instrument vendors have developed innovative cells that don't require wrenches to assemble and disassemble, are chemically inert and can withstand high pressures.

      Two extraction modes may be employed: dynamic extraction (in which fresh supercritical fluid is passed over the sample on a continuous basis) and static extraction (in which the extraction cell is filled with sample, modifier and carbon dioxide at a desired density and temperature and allowed to equilibrate before it is released to the collection vessel). In some cases, combining an initial static extraction followed by a dynamic extraction works more effectively

  • Restrictors. Once the supercritical carbon dioxide and analyte leave the extraction vessel, the supercritical carbon dioxide must be depressurized in order to change from a supercritical fluid state to a gas state. Pressure regulation is accomplished with a restrictor, which is by no means a trivial part of the instrument. Early instruments used fixed restrictors -- essentially fused silica capillary tubes. Fixed restrictors tend to plug and have limited lifetimes. In addition, flow rate variations between restrictors can cause irreproducible results.

      Now all major vendors have automatic variable restrictors which can open and close to maintain flow rates at desired levels. The decompression control offered by automatic variable restrictors has been a major factor contributing to the reduction of clogging. This is particularly important in the analysis of high-fat foods.

  • Collection vials and solid-phase traps. After passing through the restrictor, the supercritical carbon dioxide prior to loading into the extraction cell. Large extraction vessels (i.e., 20 milliliters and larger) are advantageous because they accommodate the larger sample sizes which are needed for trace analyte analysis. AISO, in the case of liquid samples which are diluted with substantial amounts of solid support to absorb free moisture, larger extraction cells are a necessity.

      In recent years, instrument vendors have developed innovative cells that don't require wrenches to assemble and disassemble, are chemically inert and can withstand high pressures.

      Two extraction modes may be employed: dynamic extraction (in which fresh supercritical fluid is passed over the sample on a continuous basis) and static extraction (in which the extraction cell is filled with sample, modifier and carbon dioxide at a desired density and temperature and allowed to equilibrate before it is released to the collection vessel). In some cases, combining an initial static extraction followed by a dynamic extraction works more effectively

  • Restrictors. Once the supercritical carbon dioxide and analyte leave the extraction vessel, the supercritical carbon dioxide must be depressurized in order to change from a supercritical fluid state to a gas state. Pressure regulation is accomplished with a restrictor, which is by no means a trivial part of the instrument. Early instruments used fixed restrictors -- essentially fused silica capillary tubes. Fixed restrictors tend to plug and have limited lifetimes. In addition, flow rate variations between restrictors can cause irreproducible results.

      Now all major vendors have automatic variable restrictors which can open and close to maintain flow rates at desired levels. The decompression control offered by automatic variable restrictors has been a major factor contributing to the reduction of clogging. This is particularly important in the analysis of high-fat foods.

  • Collection vials and solid-phase traps. After passing through the restrictor, the supercritical carbon dioxide reverts to carbon dioxide gas and the analyte is deposited in a collection vial, which may be filled with glass wool or an organic trapping solvent. Some vendors' instruments have the capability to cool the collection vial, which for some analytes has been shown to improve recoveries.

      Some instruments allow for alternative analyte trapping techniques. For example, cryogenically controlled solid-phase extraction (SPE) traps are useful for some applications. When SPE traps are used, the analyte is deposited in the trap and the trap is then rinsed with a small volume (usually 1 to 2 milliliters) of hexane, chloroform, iso-octane, ether or some other solvent. Alternatively, the solid-phase trap can be filled with glass beads, alumina or other types of solid supports, depending on the specific application.

      Once the analyte is recovered, its level in the food product can be determined by simple gravimetric analysis (e.g., weighing the amount of fat extracted from a known quantity of snack food), or a portion of the extract can be removed with a syringe and analyzed by gas chromatography (GC) or high-performance liquid chromatography (HPLC). This type of analysis is referred to as off-line analysis, or off-line collection.

  • On-line analysis. On-line analysis is a more advanced technique that involves coupling the GC or HPLC to the restrictor of the SFE so the analyte can be transferred directly into the analytical instrument. It is advantageous for trace analysis. Since all the analyte is transferred to the analytical instrument, on-line SFE can be 1,000 times more sensitive than off-line SFE. Off-line analysis, however, is far simpler to perform.  On-line analysis will undoubtedly be the next major SFE area that will grow in importance as vendors create instrument accessories to simplify the cumbersome complexities of the technique. Furthermore, the development of more reliable, non-plugging SFE instruments increases the likelihood of coupling SFEs to other instruments.

    Back to top


Subscribe and receive the latest insights on the health and nutrition industry.
Join 37,000+ members. Yes, it's completely free.

You May Also Like