Concentration levels. Aromatic levels are generally low - in the ppm, ppb or ppt range. Therefore, it is not only necessary to isolate the components but to concentrate them by several orders of magnitude.

Matrix. The sample frequently contains nonvolatile components such as lipids, proteins and carbohydrates. These may create problems of foaming and emulsification during isolation procedures and will create artifacts if injected into a hot GC injection port.

Complexities of food aromas and flavors. The aromatic composition of foods is frequently complex. For example, coffee currently has almost 800 identified compounds. Complicating the picture is the fact that the flavor compounds present may cover a wide range of polarities, solubilities, and pH values.

Variation in volatility. The flavor-important chemicals may have a boiling point from well below room temperature, or they may be solid at fairly high temperatures (such as vanillin, with a melting point of 81°C.)

Instability. Many food aroma components are unstable and may be oxidized by air or degraded by heat or extremes of pH.

Shortcomings of traditional test methods

  Common sample preparation methods include solvent/solvent extraction, distillation, and simultaneous distillation and solvent extraction. More recently developed methods such as static headspace (SHS), purge-and-trap (referred to as P&T or dynamic headspace), and direct thermal desorption (usually used for solid samples) have gained widespread acceptance among today's flavor chemists.

  While the list of sample preparation methods is long, none of the methods is appropriate for every type of analyte or sample matrix. In fact, they all suffer from significant disadvantages. Methods that use solvent extraction and concentration are time consuming and expensive; not only are solvent reagents expensive to buy, but they are also expensive to dispose of once they are used. Distillation methods are time consuming, and heating can degrade analytes and other food chemicals, generating artifact peaks.
  Even the two most commonly used methods, SHS and P&T, aren't perfect. For example, static headspace does a poor job of detecting semi-volatile chemicals - some of which are critically important to food flavor. P&T is time consuming, difficult to automate, and does a poor job with some important highly volatile flavor chemicals.

Advantages of SPME

  A relatively new sample preparation technique, solid-phase microextraction (SPME), promises to offer numerous advantages over other sample preparation methods for food aroma analysis. SPME allows for the rapid, solventless extraction and pre-concentration of volatile and semi-volatile organic compounds. It uses the partitioning of organic components between a bulk aqueous or vapor phase and thin polymeric films coated onto fused silica fibers in the SPME apparatus. (See sidebar for more details on how the technique works.) The technique was first described by Berlardi and Pawliszyn for the analysis of environmental chemicals in water (Water Pollution Research Journal of Canada, 24:179, 1989).

SPME has two general operation modes:

sampling the analyte from the equilibrated headspace gas phase in a sealed vial

immersing the fiber to collect the analyte from the liquid portion of the sample directly.The major advantages of SPME include:

It can be easily automated for increased sample throughput.

It is fast and solvent free. (Equilibration times for sampling onto fiber range from 2 to 15 minutes.)

It can be used manually with any GC or GC/MS (gas chromatography/ "mass spec").

It is quantitative.

It is sensitive, with ppt detection limits for some compounds.

It is ideal for quick screening.

It is inexpensive.

Solving real problems with SPME

  Solid-phase microextraction has found many applications in the food industry. Here are some specific examples of problems that this technique has helped to clear up.

Potatoes contaminated with paint solvents. Ron Skinner, food chemistry laboratory manager of the Prince Edward Island (PEI) Food Technology Center, PEI, Canada, says that SPME is one of his favorite laboratory tools. He has used SPME to solve a variety of analytical problems, including the resolution of off-flavors in potato-based products.

  In one case, SPME analysis revealed that potatoes were tainted with paint fumes. According to Skinner, SPME testing was more sensitive than taste paneling and, based on the chromatographic profile of various paint solvent volatiles, he was even able to identify the specific type of paint that caused the problem. By pinpointing the source of the food taint quickly, further contamination problems and unnecessary expenses associated with tainted products were avoided.

  Skinner, who has been using SPME for the past two years, likes the technique because it's fast and doesn't alter the chemical analytes of interest like other extraction methods. He sees SPME as a general-purpose, problem-solving tool. He has used it for analyzing food products and soil contaminated with gasoline and other fuels, as well as for analyzing potatoes for pesticides. "For some pesticide testing, it works better than traditional techniques," says Skinner.

  One minor problem Skinner notes is the tendency of the large-bore SPME needle to break off small pieces of the GC injector septa into the GC injection port.

Wine taints from corks. Gordon Burns, president of ETS Labs (St. Helena, CA), uses SPME to monitor off-flavor taints from corks used for wine bottles. ETS Labs is an independent laboratory that specializes in testing for the wine industry along with testing of distilled spirits and nonalcoholic beverages.

  Burns uses SPME to monitor wine and cork samples for levels of 2,4,6-trichloroanisole, a bacterial and fungal metabolite that can contaminate the cork anywhere in its life cycle - from trees in Portugal to a processing stage in a California winery. The 2,4,6-trichloroanisole, which has an extremely low odor threshold-of-perception (approximately 5 to 25 ppt), can leach out from the cork, imparting a woody, dank, acrid aroma and flavor to the wine.

  Prior to the development of SPME, the analytical method for monitoring 2,4,6-trichloroanisole consisted of extracting a wine sample with a liter of methylene chloride, concentrating the extract down to about 200 µL, and injecting the concentrated extract into a GC. The test, which cost clients $250 per sample, required three days to complete.

  Cork venders and wine makers need to test more corks than the expensive and time-consuming solvent extraction method can handle. For example, a medium-size winery produces approximately 300,000 cases of wine annually. This would involve some 360 different production lots of corks. Since approximately five samples should be tested from each lot to get a statistically representative sampling, one medium-size winery would require about 1,800 tests per year in a QC program designed to prevent occurrence of the taint!

  Using a Varian automated SPME instrument, which has recently been modified by the supplier to incorporate sample agitation features, Burns can analyze a sample every 25 minutes. "For this application, no other analytical method can match SPME's sample throughput," he says.

  While the SPME fiber is absorbing volatiles from the wine headspace, the previous sample is chromatographing. Since the volatiles are absorbed in only a few minutes, the 25-minute GC run is the time-limiting step in the analysis. Burns uses the headspace approach instead of immersion because the lifetime of the SPME is longer. "Thousands of injections with the same polydimethylsiloxane fiber are possible," says Burns.

  ETS analyzes mostly wine samples, but it has analyzed cork samples, as well. The minimum detectable quantity of 2,4,6-trichlorophenol is 5 ppt and a typical standard deviation is less than 1 ppt - far better than the precision demonstrated by P&T analysis. For this application, SPME is not only significantly more affordable to perform, it is perhaps one of the only methods possible for such a large volume of testing.

  Burns is currently working on SPME methods to measure the concentration levels of important flavor chemicals that are imparted to wine by oak barrels during aging.

Analysis of flavored beers. Marc Constant and John Collier of Miller Brewing Co., Milwaukee, have used SPME for about three years to monitor competitors' samples and Miller's production samples of beer flavored with raspberries, apricots, cherries, cranberries, and other fruits and fruit combinations. Analyzing esters and other fruit flavor components is faster and more precise with SPME than with other techniques.

  SPME proved to be easier and faster than P&T. It is approximately 20 times more sensitive than SHS methods for the mid-boiling-point analytes, and it can quantitate high-boilers (up to 300°C), which don't even show up in chromatograms of samples analyzed by SHS.

  Extremely volatile components like acetaldehyde were not detected as well by SPME as by SHS. However, cryofocusing volatiles with liquid nitrogen prior to their release into the GC column greatly improved sensitivity and peak shapes for these compounds.

  So far, Miller hasn't applied SPME to routine QC testing applications. "We could because it's simple enough," says Constant, manager of Miller's Research QA Method Development Laboratory. One impediment to applying SPME as a QC tool, he says, is that "it provides almost too much information." There could easily be over 100 peaks in one GC chromatogram, and only a few of them might be important to monitor for flavor characteristics.

  Like Burns, Constant prefers to use SPME to sample volatiles in the headspace above the sample rather than immerse the SPME fiber in the liquid portion of the sample. Constant particularly appreciates that SPME injections are clean and don't foul the GC column or mass spectrometry detector.

  Based on his experience, Constant says the key to getting accurate and reproducible results with SPME is to be sure to perform sampling in exactly the same way each time. The actual position of the fiber in the headspace (e.g., 2 mm from the surface of the sample or nearer to the top of the vial) can affect results, so it's important to place the fiber in the same position each time when sampling. Reasonable temperature control during the extraction process, controlling extraction times, proper selection of the type of fiber to use, and using the method of additions (spiking) standard calibration technique with an internal standard are important steps to obtaining accurate and reproducible results.

  More effective control of the analysis can be assured using automated sampling. The software associated with the automated version of SPME by Varian allows precise control of the absorption and desorption times, a preabsorption delay, controlled sample agitation, and multiple sampling from each vial using either headspace or direct liquid sampling.

  "SPME fills the gap between SHS and P&T methods," says Constant. "It gives the same sensitivity as P&T, but provides better reproducibility and shorter analysis time. The advantage it has over SHS is that it's far more sensitive, especially for high boilers." Like Burns, Constant most commonly uses polydimethylsiloxane fibers and finds that they "last forever."

Additional application examples. Alan Harmon, a research chemist with McCormick & Co. Inc., Hunt Valley, MD, presents several excellent food flavor applications of SPME in the book "Techniques for the Analysis of Food Aroma." Examples include shelf-life determination of seafood cocktail sauces by monitoring concentration losses of allyl isothiocyanate and phenylethyl isothiocyanates, the primary volatiles in horseradish; isolation of aroma volatiles from cantaloupes, bananas, Bartlett pears, and other types of fresh fruits; analysis of volatiles from spices like curry powder and black peppercorn; and comparison of flavor volatiles in colas, root beer, and other types of soft drinks.

  Complete quantitative analysis of every volatile and semi-volatile flavor-contributing chemical in a food product may not be possible when using SPME as the only isolation/ sample preparation technique. Still, SPME appears to come closer to attaining this goal than any other single analytical sample preparation method. Furthermore, it is rapid, less expensive, and offers analytical precision as good as or better than that of most sample preparation techniques widely used today for flavor analysis.

  Even though SPME is in its infancy, it's already clear that the technique is likely to continue to grow in popularity and be a prominent and irreplaceable tool in the flavor chemist's toolbox.

How SPME Works

  The accompanying figure illustrates the extraction and desorption procedure for a manual SPME device introduced by Supelco, Bellefonte, PA. A similar device designed for automated techniques is available for use with Varian 8100 and 8200 CX series GC autosamplers (Varian Chromatography Systems, Walnut Creek, CA.)

  The manual device is essentially a modified syringe that has a spring-loaded plunger and a barrel with a detent to allow the plunger to be held in an extended position during the extraction phase and during the injection/desorption period. Also contained within the barrel is a modified 24-gauge stainless-steel needle which encloses another length of stainless steel tubing fitted tightly to a short piece of solid-core fused silica fiber. The bottom 1 cm of the fused silica fiber is coated with a relatively thin film of any of several stationary phases. This film serves as the organic "solvent" during the absorption of the volatile compounds from the matrix. The needle functions to puncture the septa that seals the sample container and the GC injection port, and to protect the fragile fused silica fiber during storage and use.

  Fibers coated with non-polar polydimethylsiloxane and the more polar polyacrylate are commercially available now. For most analyses (e.g., volatile flavor compounds), a fiber with a 100 micron coating of polydimethylsiloxane is preferred. For analysis of high-boiling-point components (e.g., polyaromatic hydrocarbons), fibers with a 7 micron thickness of polydimethylsiloxane usually work best.

  According to Robert Shirer, Ph.D., senior research chemist for Supelco, Inc., new fiber materials are being developed for SPME. For example, the company is currently working on a porous carbon SPME filter. This new fiber would improve sensitivity for low-molecular-weight alcohols and other highly volatile components (with carbon numbers from about C2 to C6) by 10 to 100 times over currently available fibers. Porous carbon filters are ideal for analyzing volatile organic sulfur compounds, which are difficult-to-analyze chemicals with extremely low flavor and aroma thresholds.

Back to top