Anal. Chem. 1996, 68, 4114-4118
Solid Phase Microextraction for Quantitative Headspace Sampling of Apple Volatiles Adam J. Matich,*,†,‡ Daryl D. Rowan,† and Nigel H. Banks‡
The Horticulture and Food Research Institute of New Zealand Ltd., Private Bag 11030, Palmerston North, New Zealand, and Centre for Postharvest and Refrigeration Research, Department of Plant Science, Massey University, Private Bag 11222, Palmerston North, New Zealand
Solid phase microextraction (SPME) was evaluated for use in the quantification of aroma volatile production by Granny Smith apples during cool storage. Particular attention was paid to quantifying r-farnesene (3,7,11trimethyldodeca-1,3(E),6(E),10-tetraene) due to its involvement in superficial scald, a disorder of cool stored apples. Comparison between SPME and solid phase extraction (SPE) showed that the SPME fiber had greater adsorption of high molecular weight (MW) volatiles such as r-farnesene. When sampling by SPME, these higher MW volatiles did not equilibrate between apples, headspace, and fiber within sampling times as long as 90 min, while lower MW volatiles equilibrated within 5 min. This behavior was also shown by a simple model system consisting of five selected volatiles dissolved in an involatile, lipophilic liquid (squalane). The less volatile high MW aroma compounds evaporated slowly from the surface of the apples and were depleted from the headspace because of very rapid adsorption by the SPME fiber. The amount of r-farnesene adsorbed by the fiber increased with air movement through the system. In a static headspace system, the amount of r-farnesene adsorbed by the fiber decreased nonlinearly with increasing distance from the apples, due to adsorption onto the glass walls. While SPME is ideal for rapid, qualitative determination of apple headspace volatiles, the slower equilibration of higher MW volatiles limits its use for quantification in more complex systems. Solid phase microextraction (SPME) is a recently developed technique1 of headspace sampling which concentrates analytes by adsorption onto a polymer-coated silica fiber prior to their thermal desorption in the injection port of a gas chromatograph. SPME was developed for sampling organic contaminants (chlorinated hydrocarbons) in water by direct immersion of the fiber into the sample2 but has more recently been applied to headspace sampling above solid and liquid samples.3 This article describes experiments undertaken to determine the suitability of the SPME technique for the direct, noninvasive sampling of aroma volatiles in the headspace of stored apples. Particular attention was paid to quantification of the polyunsaturated sesquiterpene, R-farnesene (3,7,11-trimethyldodeca-1,3(E),6(E),10-tetraene), which is impli†
The Horticulture and Food Research Institute. Massey University. (1) Belardi, R. P.; Pawliszyn, J. B. Water Pollut. Res. J. Can. 1989, 24, 179191. (2) Arthur, C. L.; Pawliszyn, J. Anal. Chem. 1990, 62, 2145-2148. (3) Page, B. D.; Lacroix, G. J. Chromatogr. 1993, 648, 199-211. ‡
4114 Analytical Chemistry, Vol. 68, No. 23, December 1, 1996
cated as the causal agent of superficial scald, a storage disorder of apples and pears. Quantification of R-farnesene by SPME and SPE has highlighted the limitations of current techniques for quantifying higher molecular weight volatiles in complex systems. EXPERIMENTAL SECTION Solid Phase Extraction. The solid phase microextraction (SPME) device (Supelco Co., Bellefonte, PA) consisted of a 10 mm long, 100 µm diameter silica fiber, coated with a 100 µm thickness of poly(dimethylsiloxane). The fiber was sheathed in a syringe needle, from which it was extended for sampling and thermal desorption. Before sampling, the fiber was conditioned for 30 min in the gas chromatograph injection port at 200 °C. Conditioned fibers were used immediately or protected from contamination by inserting the end of the syringe needle sheath into a GC injection port septum before use. Granny Smith apples (early season harvest, air-stored for 8 months at 1 °C) were used for all experiments involving apples. Volatiles were sampled by inserting the sheathed fiber through a Teflon-coated silicone septum into a glass headspace jar and then extending the fiber from its sheath. Typically, headspace sampling was for 5 min, and desorption in the GC injection port was for 2 min. Sampling times of 30 min were needed for GC/MS analysis and 90 min for comparison between SPE and SPME sampling. Tenax traps (0.3 g of 60/80 mesh Tenax-TA (Alltech Associates, Deerfield, IL), packed into a 4 mm i.d. glass tube between plugs of silanized glass wool) for SPE were conditioned before use by flushing with 30 mL min-1 of nitrogen for 4 h at 250 °C. Dry air (100 mL min-1) was passed through a headspace jar for 36 h at 20 °C, and apple aroma volatiles were collected in a Tenax trap attached to the jar’s effluent port. Each trap was eluted with 2 mL of redistilled diethyl ether, and the washings were evaporated by a gentle stream of nitrogen to ∼0.3 mL before a 3 µL aliquot was analyzed by GC/MS. This concentration step results in minimal losses of volatiles (unpublished results from this laboratory). Analysis of Volatiles. Quantification of volatiles was by flame ionization detection on a Hewlett Packard 5840A GC using a 30 m, 0.25 mm i.d., 0.25 µm film thickness, SE-30 Alltech Econocap capillary column with a N2 head pressure of 72.4 kPa (0.66 mL min-1). Oven temperature was 50 °C for 2 min, increasing by 10 °C min-1 to 220 °C, and held for 2 min. The injection port and detector temperatures were 200 and 300 °C, respectively. A modified splitless injection port was used so that both the septum and inlet purges were interrupted during SPME injections. Volatiles were identified by GC/MS analysis using a HP 5890 Series II GC coupled directly to a VG70-250S mass spectrometer (VG Instruments, Manchester, UK) and by mass spectral comS0003-2700(96)00454-4 CCC: $12.00
© 1996 American Chemical Society
parison with library spectra.4,5 The He head pressure was 14.5 kPa, and the oven temperature program was 40 °C for 5 min, increased to 200 °C at 5 °C min-1, and held for 5 min. Comparison between SPE and SPME. This was performed using a HP 5890A GC with the same conditions as for GC/MS, except that the N2 head pressure was 72.4 kPa. To assist comparison between the two methods, the SPME sampling time was increased to 90 min so that the area of the R-farnesene peak was nearly equal (3.5% larger) to that resulting from injection of 3 µL of ether washings from the Tenax trap. Dependence of Analyte Uptake on Sampling Time. Two apples were equilibrated in a 1.5 L glass jar (10 cm i.d.) overnight at 20 °C. SPME sampling times of 5, 10, 20, 40, and 90 min were used. This experiment was repeated using a system developed to model the wax surface layer on the apples. In this system, 0.5 mL of a solution of apple volatiles was placed in the bottom of a 1.5 L jar and equilibrated overnight. The solution comprised 7.2 mg of propyl butanoate, 6.5 mg of hexyl butanoate, 5.4 mg of hexyl 2-methylbutanoate, 5.3 mg of ethyl butanoate, and 5.7 mg of R-farnesene6 dissolved in 1 mL of squalane (2,6,10,15,19,23hexamethyltetracosane). Sampling of headspace concentrations of R-farnesene, hexyl butanoate, and hexyl 2-methylbutanoate was repeated using polar (85 µm polyacrylate coating) and nonpolar (7 µm coating) SPME fibers. Depletion of Volatiles by Sampling. Peak areas obtained using two headspace sampling regimes above the squalane model solution were compared. In the first, the headspace was sampled for 5 min with the sampling fiber, and the volatile levels were determined by GC analysis. In the second, a SPME fiber was exposed to the headspace for 15 min, followed immediately by a 5 min sampling with a second fiber. This process was carried out three times in random order. Dependence of Analyte Uptake on Air Movement. Fifteen apples were equilibrated overnight in a dark, 5 L jar with a 100 mL min-1 flow of water-saturated air. Headspace samples were then collected for 5 min by SPME at the effluent port at flow rates of 0, 50, and 100 mL min-1. Dependence of Analyte Uptake on the Sampling Distance from the Apples. Two apples were equilibrated overnight in each of two separate 1.5 L jars. A 0.8 m length of glass tubing, 8.5 mm i.d., was attached to each jar, one length open to the air while the other was sealed. Sections of tubing were joined with Quickfit “T” connectors at eight different distances from the fruit in each jar. The atmosphere inside the tubing was sampled through Teflon-coated septa mounted in each “T” connector. Sampling distances were between 5 mm and 1 m from the upper apple in each jar, and three sets of random samplings were obtained. Adsorption of r-Farnesene onto Glass. Two apples were equilibrated in a 1.5 L glass jar overnight. A 250 µL portion of pentane (HPLC grade) was drawn into a 5 mL gas-tight glass syringe (pentane washed, oven-dried), followed by 4.75 mL of apple headspace bubbled through the pentane, which was then shaken in the syringe barrel. The pentane washings were injected onto a plug of silanized glass wool in a 4 mm i.d., 0.6 mL doublegooseneck GC injection port insert. The pentane was evaporated (4) NIST/EPA/NIH Mass Spectral Library, Standard Reference Data; National Institute of Standards and Technology: Gaithersburg, MD, 1995. (5) Registry of Mass Spectral Data with Structures, 5th ed.; Wiley: New York, 1989. (6) Fielder, S.; Rowan, D. D. J. Labelled Compd. Radiopharm. 1994, 34 (11), 1075-1085.
a
b
Figure 1. (a) GC trace of apple headspace volatiles sampled with a Tenax trap (300 mg) through which the headspace was drawn at a rate of 100 mL min-1 for 36 h. (b) GC trace of a repeat sampling of the apple headspace volatiles, but using the SPME fiber exposed to the headspace for 90 min. Peak identities were determined by GC/ MS.
off with a gentle flow (5 mL min-1) of N2, and the insert was placed in the GC injection port for analysis. Alternately, 4.75 mL of headspace was sampled with the glass syringe, the sample was ejected, and the walls of the syringe were rinsed with 0.25 mL of pentane. The pentane washings were then injected into the insert for GC analysis as above. RESULTS AND DISCUSSION Comparison between SPE and SPME. The R-farnesene peak areas obtained by SPE and SPME trapping were adjusted to be within 3.5% of each other, but peak areas of the other apple volatiles differed markedly between the two chromatograms. The relative peak areas (Figure 1 and Table 1) show that SPME sampling favored adsorption of higher MW volatiles. To more clearly show the relatively smaller peaks obtained by SPME for the low molecular weight volatiles, that chromatogram (Figure 1b) was plotted with a smaller y-axis range (4000-20 000) than that used for the plot of the SPE chromatogram (Figure 1a). For peaks numbers 1 (MW ) 60) to 23 (MW ) 172), the relative peak areas obtained by SPME sampling were between 5 and 20% of those obtained by SPE, but as the MW increased above 180, the disparity gradually decreased until, for R-farnesene, the levels were the same. The fiber’s bias toward adsorption of high MW volatiles was further indicated by the R-farnesene peak area being 82% of the total integrated peak areas for the sample, whereas for SPE it constituted only 36%. This bias is underestimated, as Analytical Chemistry, Vol. 68, No. 23, December 1, 1996
4115
Table 1. Relative Peak Areas of the Apple Headspace Compounds Sampled by SPME and SPE Expressed as a Proportion of the Area Obtained for (E,E)-r-Farnesene by Each Methoda relative peak area peak no.
compound
SPE
SPME
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
acetic acid butanol ethyl propanoate 2-methylbutanol ethyl 2-methylpropanoate methyl 2-methylbutanoate ethyl butanoate ethyl 2-methylbutanoate hexanol propyl butanoate ethyl pentanoate methyl hexanoate 2-methylpropyl butanoate methylpropyl butanoate ethyl hexanoate hexyl acetate butyl 2-methylbutanoate propyl hexanoate hexyl propanoate 2-methylpropyl hexanoate butyl hexanoate hexyl butanoate ethyl octanoate hexyl 2-methylbutanoate 2-methylbutyl hexanoate hexyl pentanoate hexyl hexanoate (E,Z)-R-farnesene (E,E)-R-farnesene farnesol
nd 0.44 1.7 0.8 0.34 0.33 24 19 16 1.3 1.1 2.4 2.7 0.4 55 1.8 0.22 0.65 1.2 0.21 0.23 7.5 1.9 3.5 0.17 0.4 5.3 0.75 100 0.094
0.049 0.037 0.14 0.1 0.027 0.04 1.5 1.2 0.5 0.034 0.042 0.072 0.047 0.03 2 0.067 0.021 0.061 0.11 0.022 0.023 1.3 0.29 0.98 0.05 0.16 3.2 0.66 100 0.14
Figure 2. Effect of sampling time on the uptake of different molecular weight apple volatiles by the SPME fiber. The headspace was sampled above two Granny Smith apples in a 1.5 L jar at 20 °C. Each point is the mean of three replicates. Error bars represent standard errors of the means.
(Figure 2), while the higher MW volatile compounds (hexyl butanoate, hexyl 2-methylbutanoate, and R-farnesene) had not equilibrated after 90 min. An exponential model of the type used to describe adsorption of gases onto solid surfaces8 was fitted to the data:
A ) B(1 - exp(-Ct))
(1)
aThe
peak area for (E,E)-R-farnesene obtained by SPME sampling was 3.5% larger than that obtained by SPE. nd, ) not detected.
results presented in the next section of this paper suggest that the high MW volatiles would not yet have all equilibrated between the skin of the apple, the headspace, and the fiber. In addition, in their study of the trapping performance of Tenax, Brown and Purnell7 found that, for low MW organic compounds such acetic acid (Figure 1, peak 1), the sample breakthrough volume on ∼0.1 g of Tenax was less than 1 L. In the present work, ∼360 L of headspace was sampled, and so breakthrough of the lower molecular weight volatiles would be expected. Thus, SPME is even more biased toward high molecular weight volatiles than the data in Table 1 suggest. Increased adsorption with increasing molecular weight has been ascribed3 to a concurrent decrease in water solubility (polarity) of the analyte, which increases the partitioning of higher molecular weight analytes into the headspace above the liquid, where they can be sampled by the fiber. In the present case, apples have a natural wax coating in which low-polarity analytes such as R-farnesene would be expected to be highly soluble. Regardless of relative analyte solubilities in the apple coating, the information presented in Table 1 indicates that partitioning of analytes between the headspace and the fiber favors the uptake of lower volatility analytes. Dependence of Analyte Uptake on Sampling Time. Levels of the two lowest MW apple volatiles (ethyl butanoate and propyl butanoate) achieved equilibrium on the fiber within 5-10 min (7) Brown, R. H.; Purnell, C. J. J. Chromatogr. 1979, 178, 79-90.
4116
Analytical Chemistry, Vol. 68, No. 23, December 1, 1996
where A is the amount of volatile (as measured by GC peak area) adsorbed on the fiber at time t, B is a constant which represents the equilibrium coverage of the adsorbate on the fiber, and C is a combination of the sticking coefficient and the flux of the adsorbate. A good fit to the data was obtained only when the sum of two such exponential equations was used, indicating that the rate of adsorption was dependent on a more complex process than just equilibration of the fiber with a fixed concentration of each volatile. One assumption made in eq 1 is that the fiber is subject to a constant flux of adsorbate. However, the levels of aroma volatiles are depleted from the air during sampling (see the next section), and so the adsorption profile would differ from that described by eq 1. A model system consisting of four butanoate esters and R-farnesene dissolved in squalane showed behavior analogous to that shown by the apples. The two lowest MW volatiles equilibrated between the squalane solution, the atmosphere, and the fiber in 5-10 min, whereas the other three volatiles had not equilibrated after sampling times of up to 90 min. Delayed equilibration was not peculiar to the 100 µm fiber. The adsorption/time profile for a polar phase fiber (85 µm thick polyacrylate coating) was similar to that of a nonpolar fiber, but the polar fiber absorbed only 35-60% of the amount of each volatile adsorbed by the nonpolar fiber. Equilibration with a 7 µm coating nonpolar phase fiber was reached within 5-10 min for all volatiles except R-farnesene, which did not equilibrate even after 90 min. These results confirm the importance of sampling duration for quantita(8) Somorjai, G. A. Introduction to Surface Chemistry and Catalysis; Wiley: New York, 1994.
Table 2. Effect of Presampling on the Subsequent Uptake of Headspace Volatiles by the SPME Fibera
apple volatile (MW) ethyl butanoate (116) propyl butanoate (130) hexyl butanoate (172) hexyl 2-methylbutanoate (186) R-farnesene (204) a
GC peak area ( SEM normal sampling after 15 min presample 44 600 ( 1720 146 100 ( 2670 31 700 ( 1330 19 700 ( 340 760 ( 14
44 800 ( 1600 140 200 ( 4220 19 400 ( 750 11 880 ( 460 450 ( 13
differences between the means (%) (level of significance (%)) 0.3 (ns) -4.0 (ns) -38.8 (0.02) -39.7 (0.001) -41.1 (0.0002)
Five minute sampling time at 20 °C. Each value is the mean of four measurements.
tive estimation of volatiles. For example, in an apple headspace experiment with a sampling time of 5 min (100 µm fiber), R-farnesene constituted ∼65% of the total integrated peak areas, whereas for a sampling time of 90 min, its proportion rose to 82%. Equilibration between SPME fibers and solutions containing simple aromatic hydrocarbons (benzene to xylenes) is complete within 15 min.9 Compounds with MWs similar to or higher than that of R-farnesene, such as tri- to hexachlorobenzenes (MW ) 180-285), equilibrated within 90 min.3 For headspace sampling, equilibration above a model solution of these polychlorinated benzenes was achieved within 30 min. While there was a dependence of equilibration time on MW, equilibration was achieved more rapidly than in the present system. Slow equilibration is not due to diffusion-limited transport of gaseous analytes to the fiber or to slow adsorption by the fiber itself. Zhang et al.10 noted that equilibration with the fiber requires less than 1 min due to the large diffusion coefficients of gases. Zhang and Pawliszyn11 pointed out that, while sensitivity of the fiber to the less volatile compounds is high, low partition coefficients between the sample and the headspace would result in long equilibration times. In the present case, slow equilibration between the apple surface, or a model solution of apple volatiles, and the headspace appears to be the rate-determining step for sampling the higher MW apple volatiles. Field et al.12 reported long equilibration times for SPME sampling of two sesquiterpenes in the headspace above ground-dried plant material. This suggests that slow equilibrium of ∼200 MW volatiles may be a general phenomenon. Page and Lacroix3 found that SPME uptake of halogenated volatiles from an aqueous solution decreased as increasing volumes of vegetable oil were added to the solution and that the decrease was largest for the less polar, less volatile analytes. Thus, the ratio of the fiber and oil volumes determined the amount of analytes adsorbed by the fiber. Similarly, if the apple wax is the oil phase, then the amount of apple volatiles adsorbed by the fiber would depend on the volume of the fiber. The differences between the polar (85 µm coating) and nonpolar (7 and 100 µm coating) fibers relate to both the volume and the chemical composition of the coatings on the different fibers.13 The more rapid equilibration of the 7 µm coating fiber may have two explanations. First, diffusion of volatiles through the coating on the fiber is slow and determines the rate at which equilibrium is reached. Second, slow evaporation of volatiles from the squalane solution (or apple wax) (9) Potter, D. W.; Pawliszyn, J. J. Chromatogr. 1992, 625, 247-255. (10) Zhang, Z.; Yang, M. J.; Pawliszyn, J. Anal. Chem. 1994, 66, 844A-853A. (11) Zhang, Z.; Pawliszyn, J. Anal. Chem. 1993, 65, 1843-1852. (12) Field, J. A.; Nickerson, G.; James, D. D.; Heider, C. J. Agric. Food Chem. 1996, 44, 1768-1772. (13) Yang, X.; Peppard, T. LC-GC 1995, 13, 882-886.
is rate determining. Because the 7 µm thick coating adsorbs 2-6% of the volatiles (and has only 4% of the volume) adsorbed by the 100 µm coating, mass transport equilibrium is reached sooner. Given that equilibration between the atmosphere and fiber occurs within 1 min,10 evaporation of volatiles from the squalane (fruit) appears limiting, as discussed below. Depletion of Apple Volatiles by Sampling. If the rate of desorption from the liquid/solid phase is slow compared with the rate of adsorption of a volatile by the fiber, then the fiber may materially deplete the level of that component in the headspace. In this case, presampling of the atmosphere by a separate fiber may result in a discernible reduction of volatile levels. The data presented in Table 2 show that, for the two lowest molecular weight volatiles, presampling with a separate fiber for 15 min did not affect the amount of volatile adsorbed by the second fiber during a subsequent 5 min sampling. However, as the molecular weight increased, the amount of each volatile adsorbed by the second fiber significantly decreased. This behavior is consistent with the hypothesis that the rate of evaporation of the volatiles from the squalane solution (or fruit) decreased with increasing molecular weight to the extent that, during sampling, the saturated vapor pressure of higher MW volatiles was not maintained. Dependence of Analyte Uptake on Air Movement. If diffusion from the fruit surface is the rate-limiting step in equilibration of higher MW fruit volatiles with the fiber, then increased air movement (increased evaporation from the surface of fruit) should increase the amount of volatile adsorbed by the fiber. This experimental situation mirrors that used for forced cooling of fruit in cold stores. The results from such an experiment are shown in Figure 3. After equilibration in an air flow rate of 100 mL min-1, a constant amount of R-farnesene was adsorbed onto the fiber in a fixed sampling time (Figure 3). However, immediately after halving the flow rate, the amount of R-farnesene adsorbed onto the fiber was 24% lower for the same sampling time. By commencing sampling immediately after turning the air flow (100 mL min-1) off, the amount adsorbed by the fiber decreased by 33% and continued to decrease to a basal level which was 57% below the amount adsorbed at the 100 mL min-1 flow rate. This strong dependency of volatile uptake on the rate of air movement through the system supports the hypothesis that the rate-limiting step for volatile uptake is evaporation from the fruit surface. The rate of evaporation of R-farnesene would increase with the rate of air flow through the system; therefore, the amount of R-farnesene available per unit time for adsorption by the fiber would also increase. The low and reducing uptake of R-farnesene in the system with no air flow may have two causes. First, due to the slow rate of evaporation from the fruit, there was depletion Analytical Chemistry, Vol. 68, No. 23, December 1, 1996
4117
Figure 3. Dependence of R-farnesene uptake by the SPME fiber on the rate of air flow through the headspace jar. Sampling time was 5 min.
of R-farnesene levels in the air by the fiber. Second, diffusion of R-farnesene produced a concentration gradient so that R-farnesene levels in the air were lower at the effluent port, where the volatiles were sampled, than they were immediately adjacent to the apples in the jar. Dependence of Analyte Uptake on Distance from the Apples. According to Fick’s law of diffusion,14 the concentration of R-farnesene in an open tube should be linearly dependent on the distance from the source (apples). However, for both open and closed tubes (Figure 4), a curvilinear dependence of R-farnesene levels on the distance from the apples was obtained. This behavior is consistent with a system in which the diffusate was removed along the length of the diffusion path by adsorption onto the walls of the vessels and is qualitatively similar to a model which describes the removal of oxygen as it diffuses along the length of plant roots.14 In separate experiments, ∼45% of the R-farnesene in a 4.75 mL headspace sample was found to be adsorbed onto the walls of a glass syringe. Significant adsorption onto glass of both hexyl 2-methylbutanoate and ethyl butanoate was also observed. CONCLUSIONS Solid phase microextraction has found widespread use as a simple, solventless method of headspace analysis useful for the rapid, multiple determination of airborne compounds.1,3,10,11,13 (14) Armstrong, W. Adv. Bot. Res. 1979, 7, 225-332.
4118 Analytical Chemistry, Vol. 68, No. 23, December 1, 1996
Figure 4. Amount of R-farnesene adsorbed onto the SPME fiber at different sampling distances from two Granny Smith apples in a 1.5 L glass jar at 20 °C. Each point is the mean of three replicates. Error bars represent standard errors of the means.
However, quantification of higher MW volatiles by SPME was hindered by the slow transport of analytes into the gaseous phase, which results in long equilibration times and headspace depletion of analyte during sampling, and by adsorptive losses onto walls of containers. These difficulties are not exhibited by all volatiles, and, for lower MW compounds which equilibrate rapidly between fruit and fiber (see Dependence of Analyte Uptake on Sampling Time and Depletion of Apple Volatiles by Sampling in the Results and Discussion), SPME offers possibilities for rapid, nonintrusive quantitative sampling of apple volatiles. For the higher MW volatiles, one strategy to overcome headspace depletion of volatiles during sampling is to allow sufficient time (>90 min) for the system to reach equilibrium (Figure 2). Such sampling times remove the advantage of “rapid sampling” by SPME. Isolating apples from the headspace immediately prior to sampling would circumvent the problem of slow equilibration between the apple and the fiber and would also enable quantification of headspace volatiles against standard atmospheres. The adsorption of high MW volatiles onto glass (see Dependence of Analyte on Distance from the Apples in the Results and Discussion) might be circumvented by modifying the surface of the glass (e.g. polyethylene glycol to increase the hydrophilicity) or by using headspace containers of other materials. Received for review May 8, 1996. Accepted September 14, 1996.X AC9604548 X
Abstract published in Advance ACS Abstracts, October 15, 1996.
