Off-Line Coupling of Subcritical Water Extraction with Subcritical Water

Anal. Chem. 2003, 75, 2237-2242

Off-Line Coupling of Subcritical Water Extraction with Subcritical Water Chromatography via a Sorbent Trap and Thermal Desorption Lori J. Lamm and Yu Yang*

Department of Chemistry, East Carolina University, Greenville, North Carolina 27858

In this study, the off-line coupling of subcritical water extraction (SBWE) with subcritical water chromatography (SBWC) was achieved using a sorbent trap and thermal desorption. The sorbent trap was employed to collect the extracted analytes during subcritical water extraction. After the extraction, the trap was connected to the subcritical water chromatography system, and thermal desorption of the trapped analytes was performed before the SBWC run. The thermally desorbed analytes were then introduced into the subcritical water separation column and detected by a UV detector. Anilines and phenols were extracted from sand and analyzed using this off-line coupling technique. Subcritical water extraction of flavones from orange peel followed by subcritical water chromatographic separation was also investigated. The effects of water volume and extraction temperature on flavone recovery were determined. Because a sorbent trap was used to collect the extracted analytes, the sensitivity of this technique was greatly enhanced as compared to that of subcritical water extraction with solvent trapping. Since no organic solvent-water extractions were necessary prior to analysis, this technique eliminated any use of organic solvents in both extraction and chromatography processes. Water is environmentally benign, but it is too polar to serve as an extraction fluid or chromatographic eluent at ambient conditions. Fortunately, the polarity of water is dramatically decreased at elevated temperature and pressure,1 which makes it possible to apply subcritical water (sometimes also termed as pressurized hot water, superheated water, high-temperature water, or hot water) for both extraction1-22 and chromatography23-36 * Corresponding author. Tel.: 252-328-1647. Fax: 252-328-6210. E-mail: [email protected]. (1) Yang, Y.; Belghazi, M.; Lagadec, A.; Miller, D. J.; Hawthorne, S. B. J. Chromatogr., A 1998, 810, 149. (2) Hawthorne, S. B.; Yang, Y.; Miller, D. J. Anal. Chem. 1994, 66, 2912. (3) Yang, Y.; Bowadt, S.; Hawthorne, S. B.; Miller, D. J. Anal. Chem. 1995, 67, 4571. (4) Yang, Y.; Hawthorne, S. B.; Miller, D. J. Environ. Sci. Technol. 1997, 31, 430. (5) Hartonen, K.; Inkala, K.; Riekkola, M.-L. J. Chromatogr., A 1997, 785, 219. (6) Jimenz-Carmona, M. M.; de Castro, M. D. L. Anal. Chim. Acta 1997, 342, 215. (7) Field, J. A.; Reed, R. L. Environ. Sci. Technol. 1999, 33, 2782. (8) Fernandez-Perez, V.; Jimenez-Carmona, M. D.; de Castro, L. Analyst 2000, 125, 481. 10.1021/ac020724o CCC: $25.00 Published on Web 04/18/2003

© 2003 American Chemical Society

processes. Recent studies have shown that water at elevated temperature and pressure is an excellent extraction fluid for many classes of compounds from environmental samples, natural products, and other matrixes.1-22 However, organic solvents were used to collect the extracted analytes in most of these subcritical water extraction processes. In solvent trapping, the analytes are extracted by subcritical water and then collected in an organic solvent. Thus, organic solvent-water extractions have to be performed prior to GC analysis. Another drawback to solvent (9) Lagadec, A. J.; Miller, D. J.; Lilke, A. V.; Hawthorne, S. B. Environ. Sci. Technol. 2000, 34, 1542. (10) Curren, M. S. S.; King, J. W. J. Agric. Food Chem. 2001, 49, 2175. (11) Curren, M. S. S.; King, J. W. J. Chromatogr., A 2002, 954, 41. (12) Kuba´tova´, A.; Lagadec, A. J.; Hawthorne, S. B. Environ. Sci. Technol. 2002, 36, 1337. (13) Basile, A.; Jimenez-Carmona, M. M.; Clifford, A. A. J. Agricultural and Food Chemistry 1998, 46, 5205. (14) Pawlowski, T. M.; Poole, C. F. J. Agric. Food Chem. 1998, 46, 3124. (15) Kuba´tova´, A.; Miller, D. J.; Hawthorne, S. B. J. Chromatogr., A 2001, 923, 187. (16) Hageman, K.; Mazeas, L.; Grabanski, C. B.; Miller, D. J.; Hawthorne, S. B. Anal. Chem. 1996, 68, 3892. (17) Hawthorne, S. B.; Grabanski, C. B.; Miller, D. J. J. Chromatogr., A 1998, 814, 151. (18) Yang, Y.; Li, B. Anal. Chem. 1999, 71, 1491. (19) Li, B.; Yang, Y.; Gan, Y.; Eaton, C. D.; He, P.; Jones, A. D. J. Chromatogr., A 2000, 873, 175. (20) Young, T. E.; Ecker, S. T.; Synovec, R. E.; Lomber, J. P.; Wai, C. M. Talanta 1998, 45, 1189. (21) Kuosmanen, K.; Hyo ¨tyla¨inen, T.; Hartonen, K.; Riekkola, M.-L. J. Chromatogr., A 2002, 943,113. (22) Andersson, T.; Hartonen, K.; Hyo ¨tyla¨inen, T.; Riekkola, M.-L. Anal. Chim. Acta 2002, 466, 93. (23) Smith, R. M.; Burgess, R. J. Anal. Commun. 1996, 33, 327. (24) Yang, Y.; Jones, A. D.; Eaton, C. D. Anal. Chem. 1999, 71, 3808. (25) Pawlowski, T. M.; Poole, C. F. Anal. Commun. 1999, 36, 71. (26) Chienthavorn, O.; Smith, R. M. Chromatograhia 1999, 50, 485. (27) Wilson, I. D. Chromatographia 2000, 52, S-28. (28) Yang, Y.; Lamm, L. J.; He, P.; Kondo, T. J. Chromatogr. Sci. 2002, 40, 107. (29) Lamm, L. J.; Yang, Y. Subcritical water extraction coupled with subcritical water chromatography. Proceedings of The 10th International Symposium on Supercritical Fluid Chromatography/Extraction, Myrtle Beach, SC, August 2001. (30) He, P.; Yang, Y. J. Chromatogr., A 2003, 989, 55. (31) Miller, D. J.; Hawthorne, S. B. Anal. Chem. 1997, 69, 623. (32) Ingelse, B. A.; Janssen, H.-G.; Cramers, C. A. J. High-Resolut. Chromatogr. 1998, 21, 613. (33) Hooijschuur, E. W. J.; Kientz, C. E.; Brinkman, U. A. T. J. High-Resolut. Chromatogr. 2000, 23, 309. (34) Yang, Y.; Jones, A. D.; Mathis, J. A.; Francis, M. A. J. Chromatogr., A 2001, 942, 231. (35) Smith, R. M.; Chienthavorn, O.; Wilson, I. D.; Wright, B. Anal. Commun. 1998, 35, 261. (36) Smith, R. M.; Chienthavorn, O.; Wilson, I. D.; Wright, B.; Taylor, S. D. Anal. Chem. 1999, 71, 4493.

Analytical Chemistry, Vol. 75, No. 10, May 15, 2003 2237

trapping is its limited sensitivity, since only a tiny fraction of the sample is used for chromatographic analysis. To overcome these drawbacks, coupled techniques have been developed to avoid the use of organic solvent trapping. These include subcritical water extraction coupled with solid-phase microextraction (SPME)-GC,16,17 HPLC,18,19 LC using ambient water as the eluent,20 and LC/GC.21 Subcritical water extraction followed by SPME is instrumentally simple. A static extraction vessel is filled with solid samples and water and heated to a desired temperature. The SPME fiber coated with an organic layer is then submerged into the cooled water to extract the analytes. Once the aforementioned step is complete, the fiber is then injected into the GC for analysis.16,17 Disadvantages of this technique include nonexhaustive extraction and difficult calibration in terms of trying to simulate the sample matrix for standard solutions. Subcritical water extraction has been recently coupled to HPLC, in both off-line18 and on-line modes.19 In both approaches, sorbent trapping was used to collect the extracted organic compounds, followed by connection with a HPLC system for analysis. An obvious disadvantage of this method is the common use of organic solvents during chromatographic separation. It has been reported that subcritical water extraction can be coupled with liquid chromatography using ambient water as the eluent.20 This coupling was achieved by using an array of complex instrumentation: two pumps, several high-pressure valves, and a switching valve. Once the extraction was complete, the switching valve was used to change the flow of the water mobile phase, flushing the sampling loop and introducing the SBWE effluent into the separation column. Since heart cuts of the SBWE effluent stream were made every few minutes, only a fractional analysis was achieved. Pressurized hot water (subcritical water) extraction was recently coupled on-line with LC/GC for the analysis of brominated flame retardants in sediments.21 Solid-phase trapping was used during pressurized hot water extraction. The collected analytes were eluted to a LC column, where cleanup, concentration, and fractionation of the extract were performed. The LC fraction was then introduced into the GC system via an on-column interface. Subcritical water chromatography has also received attention.23-36 By adding an oven and a pressure restrictor, a traditional HPLC system can be converted into a subcritical water chromatography system. UV detection has been used mainly in subcritical water chromatography;23-30 however, other means of detection have also been explored. For example, subcritical water chromatography coupled with FID has been investigated,31-34 and it has also been reported that NMR was used to detect analytes following subcritical water separation.35,36 The use of superheated water for thermal desorption of the analytes after solid-phase extraction (SPE) has been investigated.37 The extracted analytes were eluted from the SPE cartridge using pure water at high temperatures, and the water eluent was analyzed for quantitation. Very recently, the thermal desorption approach was used to link superheated water extraction with superheated water chromatography.38 In this study, three ovens

were employed to provide heating for the extraction cell, the trapping column, and the analytical column. Unfortunately, no recovery data were reported in this paper. Our initial studies demonstrated that it was feasible to couple SBWE off-line with SBWC via a sorbent trap and thermal desorption using a simple arrangement of extraction and chromatography systems.29 Therefore, a homemade system was developed to perform both subcritical water extraction and subcritical water chromatography in this study. After hot water extraction was performed, the sorbent trap was connected to a hot water chromatography system using high-pressure valves. The analytes were thermally desorbed from the sorbent trap using a heating tape and then introduced into the separation column using hot water. Since both subcritical water extraction and subcritical water chromatographic separation of anilines and phenols were previously reported,19,24 they were used to test this new technique. The effects of extraction temperature and water volume on the recoveries of flavones39 were also investigated.

(37) Bone, J. R.; Smith, R. M. Anal. Commun. 1999, 36, 375. (38) Tajuddin, R.; Smith, R. M. Analyst 2002, 127, 883.

(39) Lazarides, L. The Nutritional Health Bible; Harper Collins Publishers: London, 1997.

2238

Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

EXPERIMENTAL SECTION Reagents. Aniline, 3-chloroaniline, 2,4-dichloroaniline, phenol, 2-chlorophenol, 2,3-dichlorophenol, flavone, and 7-hydroxyflavone were purchased from Aldrich (Milwaukee, WI). Deionized water (18 MΩ cm) was obtained from a NANOpure-A four-holder system (Sybron/Barnstead Co., Boston, MA). HPLC grade methanol, methylene chloride, and acetone were from Fisher Scientific (Fair Lawn, NJ). Analyte solutions were prepared in methanol at concentrations ranging from 100 to 1000 ppm. Sorbent Trap Packing. The sorbent trap used in the flavone experiments was packed in our laboratory. An empty analytical guard (trapping) column was purchased from Upchurch Scientific (20 cm × 2.0 mm i.d., Oak Harbor, WA). Stainless steel frits (0.5µm porosity; frit diameter, 0.25 in.; thickness, 0.062 in.) were obtained from Upchurch Scientific. The slurry was prepared by using ∼2 g of the stationary phase from a ZirChrom-PS column (polystyrene-coated zirconia, 3 µm, 300 Å, ZirChrom Separations, Inc., Anoka, MN) and mixing it with 40 mL of methanol. The slurry was sonicated for 30 min before filling the reservoir (1/4in. inlet and outlet, 40 mL; Alltech Associates, Inc., Deerfield, IL). The trapping column was then packed using a high-capacity slurry packer (95551U; Alltech Associates, Inc.) equipped with an air compressor (C350; Alltech Associates, Inc.). The packed trapping column was then rinsed with water before use. Sample Spiking. Ottawa sand (Fisher Scientific) was cleaned using methylene chloride followed by acetone and then dried in an oven at 100 °C for 1 h. The extraction vessel (50 × 4.6 mm i.d., Keystone Scientific, Bellefonte, PA) was filled with ∼1.2 g of clean sand in about 0.24 g increments. Between each layer of sand, 2 µL of analyte solution (anilines or phenols) was added to the extraction vessel until a total of 10 µL was achieved. The extraction vessel was then sealed with the appropriate frit and end cap. Orange peel was used as the spiking matrix for flavones. Dried orange peel was cut into long strips using a razor knife and then placed in the extraction vessel. The extraction vessel containing the sliced orange peel was heated at 200 °C for 20 min at a flow rate of 1.5 mL/min to extract the orange peel before being used

Figure 1. Subcritical water extraction system with sorbent trapping.

for spiking. A blank extraction was performed using the preextracted orange peel and showed no interfering peaks for flavones. Once the orange peel was preextracted, 10 µL of flavone solution was added on the top of the preextracted orange peel at the inlet of the extraction vessel. The extraction vessel was then sealed with an appropriate frit and end cap. Subcritical Water Extraction. The homemade system for performing hot water extraction with sorbent trapping is shown in Figure 1. The inlet of the loaded extraction vessel was connected to a preheating coil, which was attached to a pump (LDC Analytical, Riviera Beach, FL) using a three-way valve (V1/ V2). All valves used in this work were purchased from HighPressure Equipment Co. (Erie, PA). The extraction vessel and the preheating coil were then placed into a Fisher Isotemp oven. The outlet of the extraction vessel was connected to a sorbent trap packed with either ValuPak C18 (20 × 4 mm i.d., Keystone Scientific) or ZirChrom-PS (20 × 2 mm i.d.) materials. After checking for leaks, the oven was set to the desired temperature with V2 closed. After the desired extraction temperature was reached, V1 was closed and V2 and V5 were opened. A constant flow rate was then set for the pump, and the extraction was performed. The sorbent trap was placed in an ice water bath during the hot water extraction, and the analytes were collected in the sorbent trap. After the extraction was completed, the pump was stopped, and the sorbent trap was disconnected from the extraction system then connected to the hot water chromatography system. Anilines and phenols were extracted at 100 °C with 2 mL of water at 0.5 mL/min. The extractions of flavones were carried out both at 100 and 150 °C with three different volumes of water at 0.5 mL/min. Subcritical Water Chromatography. Hot water chromatographic separation was performed using the same instrument with minor adjustments as shown in Figure 2. The same LDC pump and Fisher Isotemp oven that were used for hot water extraction were utilized for hot water chromatography. The outlet of the pump was connected to a Valco injector fitted with a 2-µL sample loop (purchased from Keystone Scientific). A 50 × 4.6 mm i.d. ZirChrom-PS separation column (polystyrene-coated zirconia, 3 µm, 300 Å) was used for separating the anilines and phenols. A Discovery HS PEG separation column (silica-based poly(ethylene glycol), 150 × 4.6 mm i.d., 5 µm, 120 Å, Supelco, Bellefonte, PA) was used for separation of flavone and 7-hydroxyflavone. Detection of the eluted solutes was accomplished with a UV detector (Shimadzu, Japan) set at a wavelength of 254 nm. Approximately

Figure 2. Subcritical water chromatography system coupled with the sorbent trap.

20 min after the desired temperature was reached, the separation of analytes using pure water at elevated temperature was carried out by opening V1 and V3 (V2 and V4 being closed). The chromatographic separation temperatures were 80 °C for anilines/ phenols and 100 °C for flavones. The flow rates were 0.4 and 2 mL/min for anilines/phenols and flavones, respectively. Coupling of Subcritical Water Extraction Trap with Subcritical Water Chromatography. After a subcritical water extraction, the sorbent trap was connected with V2 and V4, as demonstrated in Figure 2. The trap was heated at 130 °C (for anilines and phenols) or 150 °C (for flavones) for 20 min using a heating tape to thermally desorb the collected analytes from the trap. Valves V2 and V4 were closed while thermal desorption was performed, V1 and V3 being opened. This allowed the equilibration of the SBWC system. After 20 min, V1 and V3 were closed, and V2 and V4 were opened. Deionized water was pumped through the sorbent trap, and the desorbed analytes were introduced into the column for separation and analysis. Sonication Extraction. After SBWE, the sand residue in the extraction vessel was quantitatively removed into a glass vial, and 1.5 mL of methylene chloride was added. The vials were sealed and wrapped with Parafilm and sonicated for 2 h. The vials were checked frequently to avoid evaporation of the solvent. Internal standard (p-xylene) was added to the methylene chloride extract after the sonication extraction but prior to the GC analysis. Gas Chromatographic Analysis. A Hewlett-Packard 6890 GCFID (Wilmington, DE) was used for analyzing the sonication extract of the sand residues. The sample was introduced into a HP capillary column (HP-35, 30 m × 0.25 mm, i.d., 0.25-µm film thickness) using the splitless mode. The initial oven temperature was 80 °C, and it was increased to a final temperature of 320 °C at 15 °C/min. RESULTS AND DISCUSSION Anilines. A blank extraction of sand was performed to investigate whether the heating of the packing material in the sorbent trap (ValuPak C18 as described in the Experimental Section) would cause any problems in the chromatograms. The chromatogram produced from a blank hot water extraction using 2 mL of 100 °C water followed by hot water chromatography at 80 °C showed that the baseline was stable, and no significant peaks were encountered, demonstrating that the thermal desorption did not cause significant degradation of the packings in the trap. The Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

2239

Table 1. Recoveries of Anilines and Phenols after the Extraction and Chromatography analyte

% recovery (% RSD, n ) 3)

aniline 3-chloroaniline 2,4-dichloroaniline phenol 2-chlorophenol 2,3-dichlorophenol

85 (3) 81 (1) 80 (1) 82 (2) 86 (5) 79 (3)

breakthrough volume of the ValuPak C18 for anilines and phenols is ∼3 mL of ambient water. As shown in Table 1, recoveries of 80% or higher were achieved for both aniline and chloroanilines from the sand matrix. The low % RSDs in Table 1 demonstrate that this off-line coupling technique was reproducible. The chromatograms of anilines obtained from SBWC and from the off-line coupling of extraction with chromatography are shown in Figure 3. One of the concerns in this study was whether the heating of the sorbent trap was an efficient way to thermally desorb the analytes collected in the trap prior to introducing them into the separation column. Therefore, immediately after a complete run of the coupling (SBWE-thermal desorption-SBWC), the sorbent trap was connected as a precolumn to a traditional HPLC system with a mixture of methanol and water as the mobile phase to determine whether there are analytes remaining in the trap after the thermal desorption. The resulting chromatogram showed no analyte peaks, demonstrating that the thermal desorption was efficient. GC analysis of the sand residue revealed that ∼5-10% of each analyte was found in the sand residue. On the basis of previous studies,18,19 the extraction efficiency can be improved by increasing either the temperature or the volume of water. Since the focus of this work was to investigate the feasibility of the offline coupling SBWE with SBWC, the extraction efficiency of anilines was not further optimized. The peak widths at half peak height were compared to verify whether peak broadening would occur after the off-line coupling. As shown in Figure 3, the peak widths of aniline and 3-chloroaniline after the coupling are almost one-half the size of the peak widths found in SBWC calibration. We believe this to be due to the difference in injection temperature between the two injection modes. In the off-line coupling mode, the sorbent trap was heated at 130 °C, which was significantly higher than the separation temperature of 80 °C. This high temperature should result in better mass transfer, thus producing narrower peaks after the coupling of the trap with SBWC. Since the trap was not used in the calibration mode of the SBWC and the separation column was exposed only at 80 °C during the course of separation, the mass transfer might be poorer than that in the off-line coupling mode. In addition, the hot injection at 130 °C onto a colder column (80 °C) in the off-line coupling mode could cause some focusing, whereas a cold solvent injection in the SBWC calibration mode may result in peak broadening. The injection temperature difference between the two modes was also responsible for the slightly shorter retention times in the off-line coupling mode. Phenols. Like the aniline mixture, the efficiency of the coupling technique was examined by comparing the chromato2240 Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

Figure 3. Chromatograms of anilines obtained from SBWC calibration (top) and after the off-line coupling of the extraction with chromatography (bottom). Peak identification: 1, aniline; 2, 3-chloroaniline; 3, 2,4-dichloroaniline.

grams of phenols obtained from a calibration run (Figure 4, top) and the coupled system (Figure 4, bottom). The phenols were extracted from sand at 100 °C using 2 mL of water. As shown in Table 1, phenol recoveries of 79% or higher have been achieved under the quoted experimental conditions. The peak widths of phenol after the coupling were generally narrower than that obtained using the SBWC calibration run, as shown in Figure 4. Flavones. Since optimization of subcritical water extraction of flavones has not been previously performed, an experiment was carried out to assess the breakthrough volumes of the flavones from the ZirChrom-PS trap during subcritical water extraction. The results showed that flavones started to elute from the trap after ∼17 mL of cold water running through it at a flow rate of 1

Table 2. Effect of Extraction Volume on Recoveries of Flavones at 150 °C water volume (mL) 10 15 20

% recovery (% RSD, n ) 3) flavone 7-hydroxyflavone 70 (10) 82 (10) 76 (13)

69 (11) 92 (13) 64 (17)

Figure 5. Chromatograms of flavones obtained from SBWC calibration (top) and after the off-line coupling of the extraction with chromatography (bottom). Peak identification: 1, flavone; 2, 7-hydroxyflavone. Figure 4. Chromatograms of phenols obtained from SBWC calibration (top) and after the off-line coupling of the extraction with chromatography (bottom). Peak identification: 1, phenol; 2, 2-chlorophenol; 3, 2,3-dichlorophenol.

mL/min. Therefore, the volume of 17 mL was considered as the breakthrough volume for the ZirChrom-PS trap. To evaluate whether the preextracted orange peel can be used as the spiking material for the extraction of flavones, a blank subcritical water extraction of the preextracted orange peel was performed at 150 °C. The resultant chromatogram displayed peaks at the beginning of the elution pattern, but as time progressed, the baseline became stable. Since the retention times of flavone and 7-hydroxyflavone were relatively long, the initial peaks did not interfere with the separation of flavones. Flavone and 7-hydroxyflavone were then extracted and analyzed by hot water extraction and chromatography. The hot water extraction of flavones was optimized in this study. Effect of Water Volume on Recoveries. The percent recoveries of flavone and 7-hydroxyflavone were determined using three

different volumes of water. The extraction temperature for these experiments was 150 °C. As shown in Table 2, the highest percent recoveries of flavone and 7-hydroxyflavone, 82 and 92%, respectively, were obtained by using an extraction volume of 15 mL. When the extraction volume was increased to 20 mL, the percent recovery actually decreased for both analytes. This decrease was expected, since the breakthrough volume was 17 mL for both analytes, as mentioned earlier. The chromatograms shown in Figure 5 represent the SBWC calibration and the coupling using 15 mL of water for SBWE. Effect of Extraction Temperature on Recoveries. The temperature effect on the recoveries of flavone and 7-hydroxyflavone was studied using an extraction volume of 15 mL. Two different temperatures, 100 and 150 °C, were used. As shown in Table 3, the percent recovery was increased by raising temperature. Since the solubility of organic compounds is enhanced with increasing water temperature, the improved recoveries were expected at higher temperatures. Due to the higher desorption temperature (150 °C) used for flavones, the peaks were narrower and the Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

2241

Table 3. Effect of Extraction Temperature on Recoveries of Flavones Using 15 mL of Water temperature (°C) 100 150

% recovery (% RSD, n ) 3) flavone 7-hydroxyflavone 22 (15) 82 (10)

16 (28) 92 (13)

retention times were shorter in the off-line coupling mode, as shown in Figure 5. Discussion on Trapping Material and Thermal Desorption Temperature. It should be noted that the key issue in this offline coupling technique is the packing material of the trap. There is a need to optimize analyte retention within the trap during the SBWE process, efficiency of thermal desorption of the analytes from the trap, and the thermal stability of the trapping material during the thermal desorption and chromatographic separation steps. With respect to trapping, a longer retention is preferred, since long retention times allow larger breakthrough volumes, which may improve subcritical water extraction efficiency. On the other hand, analytes with longer retention (larger breakthrough volumes) require higher temperature to thermally desorb them from the trap, introducing the possibility of damaging the trapping materials. On the basis of previous research,30 silica-based bonded phase could be stable at 100 °C but definitely show degradation at 150 °C. Therefore, 130 °C was chosen as the desorption temperature for the C18 trap. It has been reported that the ZirChrom columns could tolerate much higher temperatures;40 therefore, 150 °C was used as the desorption temperature for the ZirChrom-PS trap. CONCLUSIONS In this study, hot water was the only fluid used for both extraction and chromatography processes, that is, hot water extraction was coupled off-line to hot water chromatography via (40) Li, J.; Hu, Y.; Carr, P. W. Anal. Chem. 1997, 69, 3884.

2242 Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

a sorbent trap and several high-pressure valves. The extractions of anilines and phenols from sand yielded ∼80% recovery, demonstrating that the off-line coupling technique was efficient and successful. The two flavones were then extracted from orange peel with subcritical water at 100 and 150 °C. The optimal extraction volume for flavones was determined to be 15 mL. Since high temperatures enhanced organic solubility, the recoveries of flavones were improved at 150 °C. The peak widths obtained after coupling were mostly narrower than those obtained by SBWC calibration. This is an additional advantage of this green extraction and chromatography coupling technique. In previous studies, subcritical water extractions were performed using solvent trapping, which required organic solventwater extractions. Because only a very small fraction of extracts can be analyzed with either GC or HPLC, the sensitivity is poor for the solvent trapping approach. Coupling hot water extraction with hot water chromatography via a sorbent trap greatly enhances the sensitivity, since every extracted and collected analyte molecule can be detected using this technique. Although subcritical water extraction was coupled to HPLC in previous study, the use of organic solvents was still required in the chromatographic separation process. Since pure water was used as both the extraction and chromatographic fluid in this study, organic solvents were completely eliminated. The results gathered in this work have proven that the off-line coupling of subcritical water extraction with subcritical water chromatography via a sorbent trap and thermal desorption can be a useful and efficient green separation technique. ACKNOWLEDGMENT This research was supported by an award from Research Corporation. The authors also thank ZirChrom Separations Inc. for providing ZirChrom-PS columns. Received for review November 27, 2002. Accepted March 12, 2003. AC020724O