ANALYST, AUGUST 1986, VOL. 111 883 Simplified Technique for the Preparation of Glassy Carbon Electrodes Russell Moy* Union Carbide Corporation, Battery Products Division, 25225 Detroit Road, Westlake, OH 44 145, USA Glassy carbon electrodes, encapsulated i n a chemically inert Teflon FEP resin, have been prepared. The preparation of these electrodes is simple, fast and inexpensive. These electrodes are believed to be better suited than wax- or epoxy-impregnated electrodes when used in organic electrolytes or hostile environments. Electrodes prepared in this manner were found to be free from memory effects and leakage. Three identically prepared electrodes were found to behave similarly, facilitating the comparison of data obtained from different electrodes. These electrodes can be used in the evaluation of electrocatalysts, as sensors and in conventio na I vol tam metric analyses.Keywords: Glassy carbon electrodes; voltammetry Carbon electrodes are routinely employed in voltammetric and polarographic analyses as indicator electrodes. One advantage that carbon has over metals such as gold and platinum is its ability to resist surface oxide passivation.1 Metallic surface oxides can passivate the electrode surface and change its electrochemical properties. Different types of carbon have been used as indicator electrodes , with glassy carbon and pyrolytic graphite offering the most consistent behaviour, i.e., their electrochemical properties are not heavily dependent on their origin.2 The surface of the pyrolytic graphite was found to be susceptible to mechanical cleavage.Cleaved graphite can entrap materials, resulting in electrodes exhibiting memory effects. Glassy carbon has been used in the evaluation of electro- chemically active functional groups covalently bound to the electrode.3 Some covalently attached functional groups exhibit electrocatalytic activity,4 and the electrochemical evaluation of these functional groups is usually accomplished by standard voltammetric and polarographic techniques. The chemical modification of porous carbon substrates such as acetylene black and porous graphite can be accomplished, but the electrochemical evaluation of these materials is difficult. Voorhies and Davis5 and Elving and Smith6 reported that large background currents were obtained when these porous materials were used as working electrodes.Similar results were obtained in this laboratory when acetylene black and spectroscopic graphite electrodes were used in cyclic voltam- metry. These voltammograms are shown in Figs. 1 and 2. Wax impregnation is effective in reducing these currents.6 It should be noted that the wax may be soluble in, or reactive with, a number of non-aqueous electrolytes. Conventional glassy carbon electrodes available from PAR (Princeton Applied Research Corp. , Princeton, NJ, USA) or prepared by Panzer and Elving2 consist of amorphous carbon sealed in a glass tube with epoxy or melted polyethylene. These sealants may be attacked by aggressive or organic solvents. Other investigators have forced glassy carbon rods into tight-fitting Teflon sleeves'-9; however , electrolyte seep- age may occur when this procedure is employed. Electrolyte seepage may also occur when single-layer Teflon shrink tubing is used to seal the electrodes.Experimental The electrodes prepared for use in corrosive solvents consist of 3 mm diameter amorphous carbon rods (Grade V10, * Present address: Department of Chemical Engineering, University of Michigan, 2135 Dow Building, Ann Arbor, MI 48109, USA. Atomergic Chemetals, Plainview, NY , USA). The carbon is sealed to a two-layer Teflon shrink tube (Grade S-130, Norton/Chemplast, Wayne, NJ, USA). The outer layer is a heat-shrinkable Teflon TFE shell and the inner layer is a meltable Teflon FEP resin. On heating, the outer layer contracts and the inner layer forms a hermetic seal to the carbon. This procedure eliminates electrolyte seepage into the electrode.The carbon rods are cut into 1-cm lengths and dried for several hours at 150 "C. The carbon is inserted into 8-cm lengths of tubing. Electrical contact to the carbon is accom- plished with a nickel wire inserted between the carbon and tubing so that the end of the wire is at the midpoint of the carbon. In order to promote the adhesion of the FEP to the carbon, the electrodes are sealed under vacuum in a tube furnace. The carbon is polished with silicon carbide paper and diamond paste. Final polishing of the electrodes is carried out with a 0.25 pm diamond paste. The electrodes are then cleaned with methanol in an ultrasonic bath. Electrochemical pre-treatment of these electrodes was not used as this step has been reported to be deleterious to the reversibility of the hexacyanoferrate(I1) - hexacyano- ferrate(II1) couple on carbon electrodes.8,' Potassium hexa- cyanoferrate(I1) electrolytes were used for the evaluation of these electrodes. 3.0 2.0 1.0 0.0 -1.0 -2.0 PotentialiV vs.Ag - AgN03 Fig. 1. Voltammetry of acetylene black - Teflon electrode in a supporting electrolyte-. Supporting electrolyte: 0.1 M tetraethylammo- nium tetrafluoroborate in acetonitrile. Reference electrode: Ag - 0.1 M AgN03. Counter electrode: Pt gauze. Scan rate: 0.100 v sANALYST, AUGUST 1986, VOL. 111 884 t c 2 3 0 TI .- 5 0 Fig. 2. I I I , I I 2.0 1.0 0.0 -1.0 -2.0 -3.0 PotentiaVV vs. Ag - AgN03 Voltammetry of Teflon-sealed spectroscopic graphite in a supporting electrolyte.Supporting electrolyte: 0.1 M tetraethylammo- nium tetrafluoroborate in acetonitrile. Reference electrode: Ag - 0.1 M AgN03. Counter electrode: Pt gauze. Scan rate: 0.100 V s-l t c 2 3 V n .- 5 0 I I I 1 I I I I -0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0 PotentiaW vs. SCE Fig. 3. Voltammetry of freshly prepared glassy carbon electrodes in a supporting electrolyte. Supporting electrolyte: 0.1 M H2S04 in water (deaerated). Reference electrode: SCE. Counter electrode: Pt gauze. Scan rate: 0.050 V s-1 I I I I I I I I -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 PotentialN vs. SCE Fig. 4. Voltammetry of glassy carbon electrodes after K4Fe(CN 6 soaking. Supporting electrolyte: 0.1 M H2S04 in water (deaerated]. Reference electrode: SCE. Counter electrode: Pt gauze.Scan rate: 0.050 V s-1 Results and Discussion Voltammograms for five randomly selected electrodes are shown in Fig. 3. The supporting electrolyte was a deaerated 0.1 M solution of sulphuric acid. Two of the electrodes, designated A and E , were found to exhibit significant background currents. These currents could possibly be attri- buted to incomplete polishing of the electrode surface. Additional polishing of these electrodes should reduce the background current. Electrodes B, C and D produce feature- less voltammograms with low background currents. For comparison, the voltammogram for a commercially available mV s-’ m1 /loo c t 4- 2 3 0 V TI .- z 0.0 0.2 0.4 0.6 0.8 1.0 PotentialiV vs. SCE Fig. 5. Anodic oxidation of Fe(CN),4- on electrode B. Electrolyte: 1.0 mM K,Fe(CN)6 - 0.1 M H2S04 - HzO.Reference electrode: SCE. Counter electrode: Pt gauze 0 0.1581 0.3162 0.4743 Fig. 6. Anodic peak current versus (scan rate)$ for the oxidation of Fe(CN) 4- on glassy carbon electrodes. H, Electrode B; A, electrode C; and b, electrode D (Scan rateN s-1)i glassy carbon electrode (PAR) is also shown in Fig. 3. The larger background current obtained with the PAR electrode is expected as its surface area is 5.5 times greater than that of those electrodes prepared in-house. The electrodes were stored overnight in a saturated solution of potassium hexa- cyanoferrate(I1) in order to test for memory effects. The electrodes were rinsed with distilled water and additional cyclic voltammograms were obtained (Fig. 4). These voltam- mograms do not differ significantly from those shown in Fig.3, indicating that the electrodes prepared in the above manner can be considered to be free from memory effects. Linear sweep and cyclic voltammograms were obtained for electrodes B, C and D at different scan rates in an aqueous solution consisting of 1 .O mM potassium hexacyanoferrate( 11) and a 0.1 M sulphuric acid supporting electrolyte. A typical voltammogram is shown in Fig. 5. A graph of the anodic peak height as a function of the square root of the scan rate is linear. An estimate of the theoretical slope and intercept of this line can be obtained from the equation ip = (2.69 x lo5) n:A D& v4 C,, . . . . (1) where n is the number of equivalents per mole of reactant, A cm2 is the electrode area, Do cm2 s-1 is the diffusion coefficient of the reacting ion, v V s-1 is the scan rate and Co mol cm-3 is the bulk solution concentration of the reactant.lo The diffusion coefficient of the hexacyanoferrate(I1) ion has been reported to be 3.9 x 10-6 cm* s-1.11 The theoretical slope of the peak current as a function of the square root of theANALYST, AUGUST 1986, VOL. 111 885 Table 1. Statistical analysis of voltammetric data Anodic peak height/A X 106 Scan rate/V s- l (Scan rate); Electrode B Electrode C Electrode D 0.010 0.100 2.76 2.72 2.99 0.020 0.1414 3.62 3.48 4.06 0.050 0.2236 5.20 4.80 5.55 0,100 0.3162 6.93 6.57 7.09 0.200 0.4472 9.25 8.66 9.13 Slope/A(Vs-’)-* x 105 . . . . 1.86 1.72 1.74 Intercept/A x lo7 . . . . . . 9.68 10.2 14.9 R . . . . . .. . . . . . 0.9997 0.9996 0.9973 Student t-ratio (intercept) . . 13.62 12.53 7.34 Student t-ratio (slope) . . . . -70.28 -68.74 -27.33 Expected slope: 3.76 x Expected intercept: 0 A A (V s- I)-$ t3.0.05 = 2.353 H,: experimental slope (or intercept) does not significantly differ from theoretical H I : experimental slope (or intercept) is significantly different than theoretical slope Test: reject H, if It1 > t3, o,Os (at the 90% level of significance) slope (or intercept) (or intercept) scan rate is therefore calculated to be 3.76 x 10-5 A(V s-+)-i and the theoretical intercept is zero. Student t-tests at the 90% level of significance showed that the experimental slope and intercept differed from the predicted values. The statistical calculations are shown in Table 1.This deviation from the predicted line suggests that there is some electrochemical irreversibility in the system. Wightman and co-workers12J3 have reported on the irreversi- bility of hexacyanoferrate(II1) reduction on carbon electrodes and similar results have been reported for hexacyano- ferrate(I1) oxidation.* Different polishing procedures may improve the reversibility of this couple on these electrodes.9 Nevertheless, Fig. 6 indicates that the peak heights for the different electrodes are almost identical. Hence, a comparison of data obtained from different electrodes will be meaningful. The preparation of these electrodes is simple, fast and inexpensive. They may be useful in situations requiring expendable electrodes. These electrodes can also be used for the evaluation of electrocatalysts and as indicator electrodes in conventional voltammetry. Electrolyte seepage into the elec- trode is eliminated when two-layer Teflon shrink tubing is used.Encapsulated electrodes should, however, be checked for memory and leakage effects in the manner described above. The author thanks Andrew Webber for helpful comments used in the preparation of this manuscript and Paul Dunn for his assistance in the preparation and evaluation of these electrodes. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. References Gunasingham, H., and Fleet, B . , Analyst, 1982, 107, 896. Panzer, R. E., and Elving, P. J., J. Electrochem. SOC., 1972, 119, 864. Rocklin, R. D., and Murray, R. W., J . Electroanal. Chem., 1979, 100,271. Yamana, M., Darby, R., and White, R. E., Electrochim. Acta, 1984, 29, 329. Voorhies, J . D., and Davis, S. M., Anal. Chem., 1960, 32, 1855. Elving, P. J., and Smith, D. L., Anal. Chem., 1960, 32, 1849. Rusling, J. F., Anal. Chem., 1983, 55, 1719. Engstrom, R. C., and Strasser, V. A., Anal. Chem., 1984,56, 136. Kamau, G. N., Willis, W. S., and Rusling, J. F., Anal. Chem., 1985, 57, 545. Bard, A. J., and Faulkner, L. R., “Electrochemical Methods,” Wiley, New York, 1980, p. 218. Eisenberg, M., Tobias, C. W., and Wilke, C. R., 1. Electro- chem. SOC., 1954, 101,306. Wightman, R. M., Kovach, P. M., Kuhr, W. G., andstrutts, K. J . , in Sarangapani, S . , Akridge, J. R., and Schumm, B . , Editors, “The Electrochemistry of Carbon,” Proceedings of the Electrochemical Society, 84-5 Electrochemical Society, Pen- nington, NJ, 1984, p. 510. Wightman, R. M., Deakin, M. R., Kovach, P. M., Kuhr, W. G., and Strutts, K. J., J . Electrochem. SOC., 1984,131,1578. Paper A51445 Received December loth, 1985 Accepted February 17th, 1986