Anal. Chem. 2004, 76, 6156-6159
A Solid-Phase Microextraction Device for the Analysis of Electrochemical Reaction Products by Gas Chromatography/Mass Spectrometry Robert C. Sparrenberger, Catheryn K. Cross, and Eric D. Conte*
Department of Chemistry, Western Kentucky University, Bowling Green, Kentucky 42101
Presented is a solid-phase microextraction syringeelectrode assembly that may be used to identify electrode reaction products. After an electrochemical experiment, the electrode within this syringe-electrode assembly can be introduced into the injection port of a gas chromatograph. Electrochemical reaction products can be analyzed, provided they adhere to the electrode surface and are amenable to gas chromatographic/mass spectrometric analysis. We highlight the potential usefulness of this device using well-known electrochemical reaction of quinones. In cyclic voltammetric experiments, peaks appearing during positive and negative scans reveal oxidation or reductions occurring, respectively. However, identification of the resulting products requires additional instrumentation. Mass spectrometry has been proven to be a useful instrument for identifying electrode reaction products. Bruckenstein is considered the early pioneer in this area using mass spectrometry for identifying products of electrode reactions.1-2 Very volatile species from an electrode reaction could be detected as they passed through a porous electrode and Teflon frit and entered a mass spectrometer. Later designs by Heitbaum allowed for quicker analysis times because the electrode/frit assembly was located very close to the ion source.3-4 Houk et al.5 provides a more detailed review of the progress in identifying electrode products by mass spectrometry. Since this review, the groups of Brajter-Toth,6,7 Stassen,8 and Van Berkel9,10 have presented other on-line designs for the interfacing of electrospray mass spectrometry with electrochemical techniques. We present in this paper an off-line coupling of cyclic voltammetry and gas chromatography/mass spectrometry (GC/MS). Electrode reactions proceed and the resultant products, provided * To whom correspondence should be addressed. E-mail: eric.conte@ wku.edu. (1) Bruckenstein, S.; Gadde, R. R. J. Am. Chem. Soc. 1971, 93, 793-794. (2) Gadde, R. R.; Bruckenstein, S. J. Electroanal. Chem. 1974, 50, 163-174. (3) Willsau, J.; Wolter O.; Heitbaum J. J. Electroanal. Chem. 1985, 185, 163170. (4) Hambitzer, G.; Heitbaum, J. Anal. Chem. 1986, 58, 1067-1070. (5) Chang, H.; Johnson, D. C.; Houk, R. S. Trends Anal. Chem. 1989, 8, 328333. (6) Zhang, T.; Palii, S. P.; Eyler, J. R.; Brajter-Toth, A. Anal. Chem. 2002, 74, 1097-1103. (7) Regino, M. C. S.; Brajter-Toth, A. Anal. Chem. 1997, 69, 5067-5072. (8) Hambitzer, G.; Heitbaum, J.; Stassen, I. Anal. Chem. 1998, 70, 838-842. (9) Deng, H.; Van Berkel, G. J. Anal. Chem. 1999, 71, 4284-4293. (10) Zhou, F.; Van Berkel, G. J. Anal. Chem. 1995, 67, 3643-3649.
6156 Analytical Chemistry, Vol. 76, No. 20, October 15, 2004
that they adhere to the electrode, can be identified by GC/MS. The device, used to hold the electrode for introduction into a GC injection port, is a modification of the popularly used solid-phase microextraction (SPME) device introduced by Pawlisyn and coworker.11 A detailed review of SPME is avaliable.12 We present a cyclic voltammetric test procedure using quinones with our presented device. EXPERIMENTAL SECTION Hydroquinone (1,4-dihydroxybenzene), benzoquinone (1,4benzoquinone), gentisic acid (2,5-dihydroxybenzoic acid), and citric acid were all purchased from Aldrich (Milwaukee, WI). The syringe device, presented in Figure 1, is similar in design to one previously described;13 however, a modification was made to allow an easily connected conductive path to the electrode. A 10 cm3 syringe having a blunted 22 gauge needle (Becton Dickinson, Rutherford, NJ) housed the electrode, which was a 0.3 mm pencil lead B refill (Pentel, Japan). The pencil lead was conditioned at 250 °C in a muffle furnace for 4 h. Openings were drilled through the plunger of the syringe to accommodate a stainless steel tube having a 1/16 in. o.d. and a 0.02 in. i.d. The exposed stainless steel tube on the plunger tip was used to attach the 0.3 mm o.d. pencil lead electrode. Quick drying epoxy glue mixed with -140/+325 mesh 304L stainless steel powder (Carpenter Powders, Bridgeville, PA) was used. The pencil lead must have physical hand resistance when being placed into the stainless steel tube containing the adhesive mixture to ensure conductivity from the tube to the electrode. A septum was placed in the needle to prevent carrier gas from exiting the injection port during injections in the same manner as reported.13 A BAS (West Lafayette, IN) Epsilon electrochemical workstation was used for all electrochemical experiments. The threeelectrode setup consisted of the syringe-pencil lead working electrode, a Ag/AgCl reference electrode, and a Pt counter electrode. A conductive path to the pencil lead electrode was made at the outside connection of the stainless steel tube that extended outside of the plunger. Gas chromatographic/mass spectrometric analysis was conducted using a Varian Saturn 2000. A 30 m poly(dimethylsiloxane) phase was used with chromatographic-grade helium as the mobile phase. This column was temperature(11) Arthur, C. L.; Pawliszyn, J. Anal. Chem. 1990, 62, 2145. (12) Lord, H.; Pawliszyn, J. J. Chromatogr. A. 2000, 885, 153-193. (13) Conte, E. D.; Miller, D. W. J. High Resolut. Chromatogr. 1996, 19, 292297. 10.1021/ac0490090 CCC: $27.50
© 2004 American Chemical Society Published on Web 09/17/2004
at an electrode, the difference in oxidation and reduction peak potentials should equal 59.2 mV/n, where n is the number of electrons involved in the electrochemical reaction. Both species undergo a two-electron oxidation according to the reactions below.
Figure 1. Syringe device assembly. (a) Pencil lead electrode, (b) stainless steel tube (working electrode connection made on the portion of this tube extending out of the plunger), (c) septum, (d) needle, and (e) syringe plunger.
programmed from 40 to 300 °C at 10 °C/min. The injection port was held at 300 °C. To prolong the life of the injection port septum, before introducing the syringe device, we opened the septum nut a quarter turn and then hand-tightened it after syringe needle introduction. Teflon tape was placed around the septum nut to prevent burns. RESULTS AND DISCUSSION Hydroquinone (1,4-dihydroxybenzene) and gentisic acid (2,5dihydroxybenzoic acid) were used to test the usefulness of this device. In electrochemistry, quinones are well-studied because they are among the few organic molecules that have electrochemical reversibility. For electrochemical reversibility to occur
These molecules were chosen to test the concept of this device (i.e., being able to identify electrode reaction products adhering to the electrode). Both substances at 10 mM in a pH 3.5, 0.1 M citric acid buffer were examined by cyclic voltammetry. The scan range was from to -200 to 700 mV. Both molecules did not display electrochemical reversibility in this setup, according to both ∆Es being greater than 59.2 mV/n. The ∆Es were 317 and 273 mV for hydroquinone and gentisic acid, respectively. A pre-anodization procedure was attempted by applying to the working pencil lead electrode 1500 mV vs Ag/AgCl while in 0.1 M NaOH. Much improved electrochemical reversibility was achieved for both molecules after running cyclic voltammagrams in the same citric acid buffer solution as seen in Figures 2 and 3. Solid traces represent CVs taken after preanodization. A measured ∆E for gentisic acid of 43 mV was close to the theoretical reversible target of 30 mV. Although reversibility for hydroquinone improved after pre-anodization, having a ∆E of 181 mV versus 317 mV, it was not near the calculated theoretical value of 30 mV. The better electrochemical reversibility of gentisic acid compared with that of hydroquinone may be the result of a hydrogen-bonding cyclictype association with ortho-substituted groups at the electrode
Figure 2. Cyclic voltammagrams of 10 mM hydroquinone in pH 3.5, 0.1 M citric acid buffer. Solid trace is after pre-anodization.
Analytical Chemistry, Vol. 76, No. 20, October 15, 2004
6157
Figure 3. Cyclic voltammagrams of 10 mM gentisic acid in pH 3.5, 0.1 M citric acid buffer. Solid trace is after pre-anodization.
Figure 4. Chromatograms of (a) linear sweep (-200 to 700 mV) in blank citric acid buffer and (b) after electrolysis at 600 mV for 5 min.
surface as reported by Zen and co-workers.14,15 These experiments were repeated with 4H, 3H, H, HB type pencil leads; however, no improvement for hydroquinone electrochemical reversibility was observed (data not shown). (14) Zen, J.-M.; Chung, H.-H.; Yang, H.-H.; Chiu, M.-H.; Sue, J.-W. Anal. Chem. 2003, 75, 7020-7025. (15) Zen, J.-M.; Kumar, A. S. Anal. Chem. 2004, 76, 205A-211A.
6158
Analytical Chemistry, Vol. 76, No. 20, October 15, 2004
For both substances the electrode protruding from the syringe was allowed to contact a stirred solution of each substance individually in citric acid buffer. The electrode was withdrawn into the syringe and then introduced into the injection port of a gas chromatograph after piercing the injection port septum with the syringe needle. Resulting chromatograms for each subsequent injection clearly revealed peaks for the individual substances. The
obtained mass spectrum for each peak was compared to the reference available at the NIST Chemistry WebBook.16 A positive scan from -200 to 700 mV of these substances in buffer was run under the same solution conditions. For both substances, a clearly observable additional peak was not observed after this positive linear scan. However, by running a constant potential electrolysis experiment at 600 mV for 5 min, a pronounced new peak was observed from the hydroquinone experiment. This peak was confirmed to be benzoquinone according to the obtained mass spectrum and through the injection of a standard. Chromatograms of a blank experiment (0.1 M, pH 3.5 citric acid buffer) and an experiment with 10 mM hydroquinone in buffer after linear scans are depicted in Figure 4. Predominant peaks for the starting molecules were observed in the respective chromatograms before electrochemical oxidation (not shown). Additional peaks in the blank experiment are unknown oxidation products from the solution, electrode, or a combination of the two. The peak area of the starting substance, hydroquinone, was essentially unchanged before and after the linear scans. The oxidized product of gentisic acid could not be observed using GC/MS even after trying a variety of inlet column temperatures. The oxidized product is most likely too active to be identified by GC/MS. Since a commercially available standard of the oxidized form of gentisic acid is not available, we attempted to make our own standard by placing 30% H2O2 in a test tube containing 10 mM gentisic acid in deionized water. The solution turned yellow, strongly suggesting oxidation took place. This oxidized product was extracted with ethyl acetate, and then an aliquot was injected into the GC/MS. Again, a response for the oxidized product was not observed; however, the peak area of
gentisic acid compared with a control experiment greatly decreased.
(16) http://webbook.nist.gov/chemistry/.
AC0490090
CONCLUSION This presented device holds promise as a tool for the easy identification of electrode reaction products provided that the species formed are stable, adhere to the electrode, and are amenable to gas chromatographic analysis. For stable species that cannot be identified by GC/MS, such as the oxidized form of gentisic acid, the possibility exists for using LC/MS as an identification tool. A device that allows for the coupling of SPME devices to LC is commercially available (Supelco, Bellefont, PA). While this procedure will not be able to detect short-lived species as reported for on-line approaches, it does offer the possibility of providing useful information such as the product identification of environmental electrochemical remediation reactions and electrode fouling and adsorption studies. ACKNOWLEDGMENT This work was supported by the National Science Foundation (CHE 0132181). Further support was provided from Western Kentucky University’s Materials Characterization Center, which is part of the Ogden College Applied Research and Technology Program. We appreciate valuable discussions with professor J. M. Zen (NCHU, Taichung, Taiwan) and D. W. Miller (NCTR, Jefferson, AR). We would also like to thank W. Scott Brown, Jason Seay, and Tabitha Hocker for their work on this project. C.K.C., a visiting high school student from Jefferson City, MO, was supported through a WKU summer research scholarship. Received for review July 6, 2004. Accepted August 10, 2004.
Analytical Chemistry, Vol. 76, No. 20, October 15, 2004
6159
