Solid Phase Microextraction Associated with Microwave Assisted Extraction of Food Products Yuwen Wang, Marisa Bonilla, and Harold M. McNair* Department of Chemistry, Virginia Tech, Blacksburg, Virginia 2406 1, USA

Maha Khaled Technical Service Division, Cultor Food Science, Groton, CT 06340, USA

Key Words:

GCMS Solid phase microextraction Microwave assisted extraction VeltoP VeItol-PIus'

Summary

Veltol@ @-methyl 3-hydroxy 4-pyrone) and Veltol-Plus@ (2-ethyl 3-hydroxy 4-pyrone) are patented flavor ingredients in food prod- ucts. Only Veltol@ can occur naturally, but both Veltol@ and Vel- tol-Plus@ are often added to food products. In order to monitor the use of these compounds in food products, lower detection limits were needed. The Solid Phase Microextraction (SPME) technique for beverages and SPME coupled with Microwave Assisted Extrac- tion (MAE) for solid food samples were studied. The influence of the pH, salt content, SPME adsorption time, GC inlet conditions, and the conditioning of the SPME fiber were investigated. Different food products were tested including coffee, beverages, chewing gums and potato chips. The coupled MAE and SPME shows good results for solid food samples. The reproducibility of the technique was less than 13%RSD and the detection limit was 10 ppb for Veltol@ and 2 ppb for Veltol-Plus@ using the SIM mode in GCNS. The technique also shows good selectivity for the target compounds investigated in different food samples.

1 Introduction

Sample preparation techniques based on adsorption have been widely used to pre-concentrate the analytes from different ma- trices for trace analysis. Solid Phase Microextraction (SPME) developed by Pawliszyn and co-workers [ 13 is a new variation of these adsorption techniques which has been used mainly for the analysis of pollutants in environmental water samples [2,3]. The key component of SPME is a fused silica fiber (-1 cm in length) coated with an adsorbent such as polydimethylsiloxane. When the SPME fiber is immersed in an aqueous solution, a partitioning of organic compounds between the aqueous phase and the hydrophobic fiber surface takes place. By removing the fiber and inserting it into a heated GC inlet, the adsorbed target compounds can then be thermally desorbed directly into a GUMS.

The hydrophobic polydimethyl siloxane fiber coating, being the most popular, has limited SPME, to date, to an aqueous extraction media. Organic solvents if used would compete with the target compounds for the sorption sites on the fiber and, probably, saturate the fiber. This would dramatically decreases the extrac- tion recovery of the target compounds. On the other hand, Mi- crowave Assisted Extraction (MAE) is an extraction technique which uses polar solvents, such as water, to extract target com- pounds primarily from solid matrices. Water absorbs the micro- wave energy, and when temperature and pressure are increased, the target compounds are more rapidly desorbed from the matrix.

The basic mechanism of this extraction is discussed in reference [4]. SPME can then be used to concentrate these compounds and introduce them into a GCMS. The MAE followed by SPME is a useful combination which combines extraction speed with concentration. The combined use of the two techniques has not yet been described.

Figure 1. Structures of Veltol@ and Veltol-Plus'.

Veltol@ (3-hydroxy-2-methyl4-pyrone) and Veltol-Plus@ (3-hy- droxy-2-eth&4-pyrone) are flavor ingredients atented by Cultor Ltd. Veltol is found naturally and both Veltol and Veltol-Plus@ are added to food products. In order to quantitate these com- pounds in food, a solvent extraction, packed column GC method was developed by Gonner et al. [5]. The tedious solvent extrac- tion, low sensitivity, and the need for derivatization make this method impractical. A GUMS detection method was developed in our lab [6] recently. It can determine Veltol-Plus@ down to 200 ppb and Veltol@ down to 400 ppb. These detection limits however are not adequate to effectively monitor the use of these compounds in some food products. For this reason, SPME for beverages and SPME in combination with MAE for solid food samples were developed.

8

2 Experimental

2.1 SPME /MAE extractions

The SPME device was purchased from Supelco Co. (Bellefonte, PA). The fibers used were 100 pm polydimethyl siloxane coat- ings. For liquid sampling, the SPME fiber was inserted directly into a 10 mL vial containing 4 mL of sample. The fiber remained inside the liquid for 10 min with magnetic stirring of the solution. For solid samples, Microwave Assisted Extractions were per- formed on a MES Model 1000 system (CEM, Matthews, NC) before SPME extraction. The extraction temperature was set to 100 "C and the pressure was 100 psi. After 10 minutes, the vessel was cooled to room temperature and pressure. The extract was

.I. High Resol. Chromatogr VOL. 20, MARCH 1997 213

SPME Associated with Microwave Assisted Extraction of Food Products

T filtered under low vacuum through a GFJA binder-free glass fiber filter (Fisher Scientific, Pittsburgh, PA).

2.2 Analysis of Food Products

2.2. I Beverages

In a 10 mL vial, 1.5 g of Na2S04 was dissolved in 4 mL of beverage. The pH was adjusted to 2 by adding 0.5 mL of 0.1 M HCl. SPME was performed at ambient temperature for 10 min. The SPME fiber was then introduced into GUMS inlet in the splitless mode.

2.2.2 Solid Food Products

Solid food products such as potato chips were ground before weighing. One gram of sample and 10 mL of water (HPLC grade, Mallinckrodt Chemical, Inc., Paris, KY) were placed in a MAE vessel. A 4 mL aliquot from the filtered MAE extract was placed into a 10 mL SPME vial for SPME concentration.

2.3 GC/MS Analysis

An HP model 597 1 GC/MS system (Hewlett Packard, Palo Alto, CA) was used. The inlet temperature was set to 260 "C. A 30 m, 0.25 mm i.d., 0.5 pm film thickness Rtx-20 (80% methyl and 20% phenyl polysiloxane) (Restek, Bellefonte, PA) capillary column was used. The column temperature was held initially 70 "C for 1 min, increased to 240 "C at 15 "/min and held for lmin. A 30 m, 0.25 mm i. d., 0.25 pm film Stabilwax (100% bonded polyethylene glycol) (Restek, Bellefonte, PA) column was also used for sample analysis. For this column, the oven temperature was programmed as follows: 40 "C for 1 min, in- creased to 230 "C at 15 "/min, then 30 "/min to 250 "C. For thermal desorption, the SPME fiber was left in the injector for 1 min. For direct injection, 1 pL of sample solution was injected by using an HP auto sampler. A splitless mode was used for both the SPME and direct injection with the purge valve closed for 1 min.

3 Results and Discussion

3. I Optimization of SPME adsorption conditions

The initial step was to optimize the SPME extraction conditions. The parameters investigated were pH, ionic strength, adsorption time, GC inlet conditions and fiber conditioning.

3.1.1 pH

Veltol@ and Veltol-Plus@ are both weak acids. At pH 7, afraction of the molecules will be ionized. This does not favor hydrophobic adsorption by the fiber. A series of pH solutions (7, 6,4, and 2) were prepared and 1 ml added to 4 ml of a 1.5 ppm Veltol@ and Veltol-Plus@ solution. SPME extractions were made from each solution. Figure 2 shows the results. The peak areas increase with decreasing pH. At pH 2 the adsorption efficiency is 4 times that as at pH 7. Veltol-Plus has a higher adsorption efficiency due to the increased hydrophobicity of the ethyl group relative to the methyl group in Veltol@.

d

mmVelt \

0 4 I 1 2 3 4 5 6 I

PH Figure 2. Influence ofpH on adsorption efficiencies of Veltol@ and Veltol-Plus@'.

3.1.2 Ionic Strength

The effect of decreasing solubility of organic compounds with the addition of salt is known as a "salting out" effect [7]. The presence of sodium sulfate decreases the solubility of both Vel- tol@ and Veltol-Plus@ in water and enhances their adsorption onto the SPME fiber. As can be seen in Table 1, the detection

Table 1. Effect of injector temperature on peak area.

Direct inj. Direct inj. SPME SPME Temperature VeltoP Veltol-Plus@ VeltoP Veltoi-Plus@ ("C) ( ~ 1 0 ~ ) ( ~ 1 0 ~ ) ( ~ 1 0 ~ ) (X105)

250 84 78 0.53 1.7

300 64 59 0.55 1.8

limit for Veltol at pH 2 without Na2S04 was 50 ppb while with Na2SO4 a limit of detection of 10 ppb was obtained. A 5 fold increase in the sensitivit is obtained for Veltol@ by adding Na2S04' For Veltol-Plus , limits of detection of 30 and 2 ppb are obtained without and with Na2S04. In this case, a 15 fold increase in sensitivity is obtained for the "salting out" effect. Fast stirring must accompany the addition of this salt to prevent recrystallization. Because the solvation of sodium sulfate is en- dothermic, the temperature of the sample solution decreased. This also may enhance the adsorption efficiency of the fiber. Sodium chloride was also evaluated, but the extraction efficiency was less than with Na2S04.

a .

3.1.3 Adsorption Time

SPME extraction is a dynamic partitioning process of the target compounds between the SPME fiber and the sample solution. With stirring, diffusion of the target compound to the fiber is increased, and 10 minutes was found necessary to reach equilib- rium (see Figure 3).

3.1.4 Fiber Conditioning

Replicate adsorption trials were carried out by successively ad- sorbing six fresh samples using the same fiber. The results show an increase in peak area with each consecutive injection then

214 VOL. 20, MARCH 1997 J. High Resol. Chromatogr.

SPME Associated with Microwave Assisted Extraction of Food Products

\r Veltol Plus@ 2000

1800

1600

1400

1200

2 1000

3 800 II

600

Veltol@ :/

200 400 I 0 4 I

0 5 10 1 5 20 2 5 30

time (min)

Figure 3. Influence of adsorption time on adsorption efficiencies of Veltol@ and Veltol-Plus@.

equilibrationwas reached after 4th or 5th adsorption. This means the fiber needs to be conditioned in the sample matrix before use. The conditioning can also be achieved by rinsing the fiber in the sample for 20 to 30 minutes and then thermally desorbing the fiber in the heated injection port.

3.1.5 GC Inlet

A major benefit of SPME for GCMS is the absence of solvent. This advantage eliminates: 1) the peak area variations with sol- vent type or lnlet temperature; and 2) the solvent delay time necessary for MS. SPME thus allows detection of early eluting volatile compounds not possible with direct injection. When splitless mode with 1 min purge time off is used, peak areas were found to decrease dramatically with an increase in the inlet temperature when conventional syringe injection was employed. This is probably due to sample loss through the septum purge with a fast solvent evaporation. The higher the inlet temperature, the greater the volume of the vaporized analytes, the greater the sample lose through the septum purge. Using SPME no signifi- cant changes in peak areas were observed (see Table 1).

3.2 Comparison of SPME with Conventional Syringe Injection

Evaluation of this SPME technique was conducted by comparing the sensitivity and reproducibility with the conventional syringe injection.

3.2. I Sensitivity

The sensitivity was represented by detection limits and was calculated as that concentration corresponding to 3 times the noise level. Duplicate injections were made for each trial. For direct injection, I pL Veltol@ and Veltol-Plus@ standard solutions were directly injected into the GC/MS. SPME of the original aqueous standard solution is represented by SPME (1) (see Table 2). SPME (2) is the original solution at pH 2. SPME (3) represents the original solution at pH 2 and saturated with sodium sulfate. The combination of salt and pH increased the sensitivity of SPME by a factor of 10 for Veltol@ and 25 for Veltol-Plus@. This method can be used to determine trace amounts of Veltol@ and Veltol- Plus@ in food products such as coffee where trace amounts of Veltol@ can be extracted.

Table 2. Comparison of the detection limit (ppb)

Compounds Direct SPME (1) SPME (2) SPME (3) injection. standard pH 2 pH2 and

NazS04

Veltol@ (ppb) 200 100 50 10 Veltol-Plus@ (ppb) 400 50 30 2

3.2.2 Precision

The precision of the method was determined by conducting 6 consecutive injections for both the SPME and direct injection techniques. The relative standard deviations (%RSDs) from both GC area counts and retention times were determined for each analyte in the mixture. For conventional syringe in'ections, con- centrations between 6 ppm and 120 ppm for Veltol and 5.1 ppm to 100 ppm for Veltol-Plus@ were investigated. For SPME, 1.5 ppm of Veltol@ and Veltol-Plus@ were tested. The %RSD of both the peak area and retention times are listed in Table 3. Direct syringe injection has better reproducibility for both peak area and retention time. Poor precision for SPME peak areas resulted from the variation of the adsorption and the desorption of the target compounds. It should be kept in mind that the precision of these results represents the entire method including the sampling stage, not only the chromatographic separation itself. Peak area preci- sion by SPME is not as good as direct injection, but is reasonable for ppb level analyses. The variation of the retention time make the identification of the target compounds difficult based solely on retention time. Use of a mass spectrometer is recommended for confirmation.

0

Table 3. Comparison of precision.

Compounds Direct inj. Direct inj. SPME SPME Area tR Area LR (RSD%) (RSD%) (RSD%) (RSD%)

Veltol@ 2.2 0.005 13 0.10 VeItol-Plus@ 2.0 0.005 11 0.05

3.3 Application of SPME in Food Products

These SPME and MAE-SPME techniques were applied to sev- eral kinds of food products including coffee, Cola, potato chips, canned food, and chewing gums.

3.3. I Beverages

Most of the beverages are aqueous solutions which make direct SPME extraction possible. Figure 4 shows a comparison of gas chromatograms obtained from Gevalla coffee by conventional direct injection and SPME. As shown, the SPME method shows higher sensitivity. Direct injection shows no peak for Veltol@. These results were confirmed by mass spectroscopy. These re- sults also demonstrate the selectivity of this SPME method. The unidentified peak at retention time 7.85 min has a peak height of 1000 for direct injection and only 150 for SPME. SPME mini- mized the interference of this peak with Veltol@ which has a retention time at 7.80 min. An additional advantage of SPME sampling is that it prevents water from entering the GC, avoiding possible damage to the GC column.

8

J. High Resol. Chromatogr. VOL. 20, MARCH 1997 215

SPME Associated with Microwave Assisted Extraction of Food Products

Column: Ra-20,30 m x0.25 MS mode: SIM 126, and 140 amu

Figure 4. Chromatograms of coffee by both SPME and direct injection.

3.3.2 Chips and Corn

Chips and corns are solid samples. Instead of performing head- space SPME [8], a MAE technique was examined. Water is the best solvent for MAE as it possess a high dielectric constant, hence it is characterized by a high ability to absorb microwave energy. MAE has the potential to extract compounds from diffi- cult matrices. It can be operated at high temperature (even higher than the boiling point of the extraction solvent) and high pressure. Therefore, MAE is a complimentary method to combine with SPME for solid samples. In support of the MAE extraction effi- ciency, MAE was compared with conventional solvent extraction (see Table 4).

Table 4. Comparison of extraction efficiency of solvent with MAE (integrator area counts).

Compound Solvent MAE MAE extraction (1 00 "C) (150 "C)

10 min 10 min

Figure 5 shows the chromatograms of potato chips with MAE- direct injection and MAE-SPME injection. Obviously MAE- SPME exhibits higher sensitivity for Veltol@. The concentration of Veltol@ in this sample is 0.3 ppm and Veltol-Plus@ was not detected. The significant peak height of Veltol@ in MAE-SPME compared with other unidentified components having the same ion in MAE-direct injection shows the selectivity of this SPME technique. The drawback of the MAE-SPME technique for solid food is that the particulate matter in the sample solution could coat the extraction fiber and interfere with the extraction. Further filtration prior extraction may help.

MAE-direct injection of potato chips

400

3 0 0

6 8 10 12 14 16 18

6 8 10 12 14 16 18

Column: Stabilwax

Figure 5. Chromatograms of potato chips by direct MAE-injection and MAE- SPME. Column: Stabilwax.

4 Conclusion

SPME is a useful tool for trace analysis of Veltol@ and Veltol- Plus@ in food products and can be considered superior to com- monly used methods such as solvent extraction. SPME is very simple to perform; it takes only a few minutes to complete and uses no solvent. It can be applied successfully to the determina- tion of trace amount of Veltol@ and Veltol-Plus@ in a variety of food products. It is selective for certain compounds. The combi- nation of MAE and SPME make solid sample SPME possible. It compensates for the disadvantage of Headspace SPME which can only detect volatile compounds [9].

Acknowledgment The authors acknowledge Cultor Ltd., Technical Service Division, Food Science Group, Groton CT for providing the samples. One author, M.B., acknowledges the financial support of R.J. Reynolds in the form of a post-doctoral fellowship.

References C.L. Aithur and J. Pawliszyn, Anal. Chem. 62 (1990) 2145.

C. L. Arthur, L. M. Killam, S. Motlagh, M. Lim, D. W. Potter, and J. Pawliszyn, Environ. Sci. Technol. 26 (1992) 979.

C. L. Arthur, D. W. Potter, K. D. Buchholz, S. Motlagh, and J. Pawliszyn,

J.R. Jocelyn Par6 and Jacqueline M.R. BClanger, Trends Anal. Chem. 13(4) (1994) 176.

S.W. Gunner, B. Haxd and M. Sahasrabudhe, J. Assoc. Off. Anal. Chem. 51(5) (1968) 959.

Y. Wang, M. Khaled, and H. McNair, Eastern Analytical Symposium & Exposition, poster #464, 1995, Somerset, NJ, Nov. 1995.

R. Fessenden, and J. Fessenden (eds), "Organic Laboratory Techniques", BrookdCole Publishing, Monterey, CA, (1984).

Z. Zhang and J. Pawliszyn, Anal. Chem. 65 (1993) 1843.

F. Pelusio, T. Nilsson, L. Montanarella, R. Tilio, B. Larsen, S. Facchetti, and J. Madsen, J. Agric. Food Chem. 43 (1995) 2138.

Ms received: March 25, 1996; accepted: November 18, 1996

LC-GC 10 (1992) 656.

216 VOL. 20, MARCH 1997 J. High Resol. Chromatogr.