316 Analyst March 1983 Vol. 108 $$. 316-321 Effect of pH on the Response of Glassy Carbon Electrodes Hari Gunasingham" and Bernard Fleet? Imperial College of Science and Technology London SIV7 2A Z The use of glassy carbon as an electrode material engenders a number of practical problems owing to the presence of C-0 functionalities on its surface. One such problem is the susceptibility of the electrode t o pH changes. Highly surface-active glassy carbon electrodes having a high proportion of irreversible C-0 groups are particularly prone to variations in pH compared with ones having mainly quinoidal species This is reflected in the perform-ance of glassy carbon in the cyclic voltaninietry of hydroquinone. Keywords Glassy carbon electrode ; surface groups ; pH ; hydroquinone For routine analytical applications it is often desirable to have an electrode material that is relatively insensitive to the effects of pH.In practice however electrodes show a marked change in response. The pH response of electrodes can be traced to acid - base reactions taking place at the surface. For example the rest potentials of a number of metal and semiconductor electrodes have been reported to show a Nernstian response to pH with the rest potential changing by about 59 mV per pH Such behaviour has been attributed to electrochemically reversible redox reactions involving the surface oxides and H+(as.,. The effect of pH on carbon electrodes has been previously rep~rted.~-ll The effect has mainly been thought to be a consequence of surface C-0 functionalities in particular quinoidal groups formed on free valence carbon sites by the chemisorption of.oxygen. I t has been observed that the equilibrium potential of glassy carbon electrodes varied by about 59 m\.' per pH unit.1° The hypothesis put forward to explain this apparent Nernstian beliaviour is that the quinoidal redox couple is reversible with respect to reactions involving H+(aq.). An important consideration with respect to the effect of pH is its influence on electrode reactions. There is the direct effect on the species (undergoing electrolysis) itself; this has been the subject of considerable research. What is less clear is the indirect influence on elec-trode reactions as a result of changes in the electrode surface with pH. Change in pH could occur in the bulk solution or more subtly in the region in the immediate vicinity of the elec-trode where the actual electrode processes take place.In the latter instance depletion or enhancement of hydrogen ions (or hydroxyl species) could be the result of the electrode reaction itself. In a previous paper5 it was reported that glassy carbon showed varying surface character-istics depending on the degree of compactness of its bulk structure the more compact the structure the fewer the free valence carbon sites available on the surface and hence the fewer the C-0 functionalities formed. Here and in the previous work the glassy carbon used was of two types Tokai glassy carbon characterised by low surface activity (mainly quinoidal) ; and Plessy glassy carbon having a higher surface activity including irreversibly formed functionalities.This paper considers the effect of pH on the performance of glassy carbon electrodes in the pH range 0.3-8.4. At higher pH values a distinct difference in behaviour was observed. Differences in the electrochemical behaviour of glassy carbon at high and low pH have been noted by other workers.6-8 Our own findings in this respect will be the subject of a separate paper.B These reactions often involve surface oxides. Experimental Gold and platinum electrodes were similarly made using 3-mm discs encased in a Kel-I; (331 USA) body. The Plessy glassy carbons are classified as before namely GC1 and GC2. * Present address Department of Chemistry National University of Singapore Kent Ridge Singapore 051 1 .t Present address HSA Reactors Fesken Drive Rexdale Toronto Canada. The fabrication of glassy carbon electrodes has been described el~ewhere. GGNASINGHAM AKD FLEET 317 Glassy carbon electrodes were polished to a mirror finish with a 1 - p i diamond paste rinsed with ethanol and distilled water and then soaked overnight in distilled de-ionised water. After this treatment the clironopotentiometric and voltammetric experiments described below were performed with no further polishing. Electrodes were rinsed with distilled water prior to each analysis that required change of buffer solution. The major reason for not polishing the electrode between analyses was the likelihood of drastically changing the physi-cocheniical characteristics of the carbon surface which would have defeated the purpose of this study.Precautions were taken to keep potential limits at which electrodes were operated, to within the range -0.5 to + 1.5 V ~ ! E Y S U S S.C.E. which ensured that significant alteration of the carbon surface did not take place. Background cyclic voltammograms were routinely run in 0.5 ?tl sulphuric acid between experiments to check the state of the surface. From these voltammograms it appeared that both Plessy and Tokai glassy carbon surfaces remained reasonably constant throughout the entire course of the experiments. This conclusion was based on the background current as well as the background-peak potentials of the voltammo-grams. The former is indicative of surface area as well as the concentration of surface functionalities.I t should also be mentioned that no sign of adsorption was seen for the cyclic voltammetric studies of hydroquinone. Gold and platinum electrodes were cleaned with concentrated nitric acid polished with 1-pm diamond paste and then rinsed with ethanol and distilled de-ionised water. The electrodes were soaked overnight in de-ionised water. The following buffer solutions were used in this work sulphuric acid pH 0.3-0.8; citrate buffer (citric acid - sodium citrate) pH 2.2-5.5; and phosphate buffer (Na,HPO - NaH,PO,) pH 5 . 8 4 4 . Chrono-potentiometric studies were carried out with a Model PAR 173 potentiostat - galvanostat (Princeton Applied Research). Cyclic voltammograms were obtained with a Model PAR 174 polarograph. Purified nitrogen was used to de-aerate the solutions.Open-circuit potentials were measured with a Corning EEL Model 112 pH meter. Results were plotted on a Servoscribe plotter. Results and Discussion Open Circuit Potential uersus pH Typical cyclic voltammograms of the two carbons obtained for 0.5 M sulphuric acid given in Fig. 1 show the relative differences in surface activity. The cyclic voltammograms were obtained immediately prior to the open-As already mentioned GC1 is more compact than GC2. I 0.1 0.3 0.5 0.7 0.9 -0.1 +0.3 +0.7 +1.1 E N vs. S.C.E. Fig. 1. Background of (a) GC1 and (b) GC2 in 0.5 M sulphuric acid. Scan rate 20 mv s-1 318 GUNASINGHAM AND FLEET EFFECT OF pH Analyst "01. 108 circuit potential - pH measurements described below. The cyclic voltammogram of GC1 on the basis of the earlier reasoning shows the dominance of the quinoidal redox couple whereas GC2 shows evidence of irreversible functionalities.12 Meas-urements were made 2 min after immersion to ensure that the electrode had approached its equilibrium-potential value.As can be seen the graph for GC1 shows a near Nernstian be-haviour with a slope of about 60 mV per pH unit. This slope is consistent for a 2e/2H+ process which would be expected for the surface quinoidal redox couple; the slope for GC2 has a significantly lower value. The difference in the slopes could be explained on the basis that the acid- base redox reaction at the surface of GC2 is irreversible a consequence of the irreversible C-0 functionalities dominant on the surface of this carbon.Fig. 2 shows graphs of open-circuit potential versus pH obtained for GC1 and GC2. pH 8.4 pH 7.5 pH 6.7 pH 5.5 pH 4.1 pH 2.3 pH 0.8 ULL 600 71 0 2 4 6 8 PH Fig. 2. Open-circuit vmsus pH plots for (A) GC1 and (B) GC2. f cri E & +50 I 0 5 10 HCl/ml Fig. 3. Use of glassy carbon as an indicator electrode in acid - base titration: (a) GC1; and (b) GC2. Acid - Base Titration The use of carbon electrodes as indicators in acid - base titrations has been described and affords an interesting demonstration of the differing response of GC1 and GC2. Fig. 3 shows the titration of sodium hydroxide by hydrochloric acid as monitored by the two carbon elec-trodes. Each measurement was made 2 min after addition of the acid. As can be seen GC2 shows a poorer response at the end-point ; again this is consistent with the irreversible nature of the C-0 groups found on the surface of this carbon.Tokai glassy carbon previously oxidised at + 1.5 V showed a behaviour similar to GC1 with respect to the open-circuit potential veYsus pH graphs. This could be expected as the surface C-0 functionalities of this carbon are mainly of the quinoidal type. 1 .o 1 - I 2 min -Time -1.0 Fig 4. Effect of pH on cathodic charging curves of Plessy glassy carbon. Charging current = - 10 PA March 1983 ON THE RESPONSE OF GLASSY CARBON ELECTRODES 319 Chronopotentiometric Studies According to Vetter,13 the anodic and cathodic charging curves for platinum are the same regardless of pH and the only apparent change was a displacement of the charging curves by 59.2 mV per pH unit.This result is indicative of the reversible nature of the acid - base redox reactions involving adsorbed oxides on platinum surfaces. With glassy carbon the effect of pH on the charging curves is more complex. Fig. 4 shows the cathodic-charging curves ob-tained for GC2 between a pH of 0.8 and 8.4. It can be seen that as pH increases the charging curve becomes broader having less defined arrests. If corresponding points for each potential wwus time charging curve are plotted against pH interesting trends are found as shown in Fig. 5. Plot (A) representing corresponding points of the cathodic charging curves 5 min 0.6 , 4 0.4 uj 0.2 G o > > 0 2 4 6 8 PH Fig. 5. Corresponding potential uevsus pH plots (A) points taken 1 min after start of cathodic charging curve; and (B) points taken 5 min after start.after the start of each curve has a slope of about 60 mV per pH unit. This value reflects the response of the reversible quinoidal couple. A similar observation was reported by Evans and Kuwana,ll for oxygen plasma treated pyrolytic graphite though here the “surface quinone” potential was evaluated by cyclic voltammetry and differential pulse voltammetry. Plot (R), representing points 1 min after the start of the cathodic charging curves has a significantly lower slope that is close to 25 mV per pH unit. As the points are from the extreme negative region of the charging curves it is plausible to surmise that the small slope is the result of reactions involving irreversible C-0 functionalities and hydrogen adsorption.Hence, hydrogen adsorption appears to be an irreversible process on glassy carbon. Comparable trends were observed for anodic charging curves of GC2 as shown in Fig. 6. \Vith increase in pH arrests appeared to become broader and less well defined than with the cathodic charging curves. pH 4.1 I pH 6.7 pH 7.5 pHb.4 pH 2.3 pH 5.5 I‘ Time T;ig. 6. Effect of pH on anodic charging curve of Plessy glassy carbon. Charging current = - 10 u.1 320 GUNASINGHAM AND FLEET EFFECT OF pH Analyst Vol. 108 0.2 V vs. S.C.E. I Potential + Fig. 7. Cyclic voltammo-grams of hydroquinone for platinum electrode. pH 7.9 pH 6.8 - 0 . 4 2 pH 4.9 0.2 V vs. S.C.E. -0.2 Potential -b Fig. 8. Cyclic voltammograms of hydroquinone for gold electrode.Cyclic Voltammetry of Hydroquinone According to Adams,14 the cyclic voltammetry of hydroquinone on solid electrodes suggests a high degree of irreversibility. The over-all reaction involves a two-electron transfer and is highly sensitive to pH. Vetter15 showed that the reaction involved consecutive one-electron transfers. I t led to the conclusion that the reaction mechanism varies with pH; one at low pH and the other at high pH. Hydroquinone was chosen for our investigations on account of its well studied electrochemistry and because the response of the quinhydrone redox couple to pH should parallel that of surface quinoidal species on glassy carbon. Figs. 7-10 show cyclic voltammograms of 5 mM hydroquinone at different pH for platinuni, gold Tokai glassy carbon and GC2 electrodes.The scan rate for all voltammograms is 20 mV s-1. It can be seen that for the Tokai carbon platinum and gold electrodes as pH is increased the shape of the cyclic voltammogram does not change as significantly as it does with 0.2 V vs. S.C.E. -0.4 if-Potential __+ Fig. 9. Cyclic voltammograms hydroquinone for Tokai glassy carbon. of < ~ D H 2.2 I -0.3 / I \ I -0.i v Potential Fig. 10. Cyclic voltammograins of hydroquinonc for l'lcssy glassy carbon (GCB) March 1983 ON THE RESPONSE OF GLASSY CARBON ELECTRODES 32 1 GC2. Also the peak current decreases (as pH increases) most appreciably for GC2. Tokai glassy carbon in fact shows the least susceptibility to pH change in the way of peak shape as well as peak current.A plot of peak potential versus pH given in Fig. 11 shows that anodic and cathodic peak potentials become more negative with increase in pH. This has been described by Adams.14 The line drawn through the plotted points marks the average poten-tial versus pH change in the low pH range; the slope of this line for GC2 is significantly closer to the Nernstian value of 60 mV per unit than for the other electrodes. This is indicative of the greater reversibility of the hydroquinone electrode reaction in the GC2 electrode; a point further substantiated by the fact that the separation of the cathodic and anodic peak poten-tials as shown in the cyclic voltammograms of Fig. 10 are less for this carbon. The deviation at pH 6 for all the electrodes seen in Fig.11 could be ascribed to an anion effect caused as a result of changing from citrate to phosphate buffers. Another reason could be the difference in the quinhydrone redox process at high and low pH as mentioned before. 0.8 a d 2 0.4 s 0 0 4 8 0 4 8 PH Fig. 11. Variation of anodic (left) and cathodic (right) peak potentials with pH. Conclusion The results presented in this paper confirm the assertion previously made that the greater the surface activity of glassy carbon electrodes the greater the susceptibility to pH. Plessy glassy carbon of the GC2 type having a significantly greater proportion of irreversible C-0 functionalities at its surface appears to be more susceptible in this respect. It may therefore, be desirable from the analytical point of view to use a less surface active carbon such as the Tokai type in favour of a more active one such as GC2.The criterion would be electrocata-lytic performance (that is sensitivity) zlersuus reproducibility (in this example with change in pH). Moreover Tokai carbon affords some measure of predictability in regard to its response. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. References Trasatti S. Editor “Electrodes of Conductive Metallic Oxides Part A,” Elsevier Amsterdam 1981, Hoare J . P. Adv. Electrochem. Electrochem. Eng. 1967 6 202. Hoare J . P. J. Electrochem. Soc. 1962 109 858. Bockris J . 0. M. Conway B. E. and Yeager E. Editors “Comparative Treatise of Electrochemistry,” Gunasingham H. and Fleet B. Analyst 1982 107 896. Taylor R. J. and Humfray A. A. J. Electroanal. Chem. 1975 64 63. Taylor R. J. and Humfray A. A. J. Electroanal. Chem. 1975 64 85. Taylor R. J. and Humfray A. A. J. Electroanal. Chem. 1975 64 95. Gunasingham H. in preparation. Dodson A. and Jennings V. J. Anal. Chim. Ada 1974 72 205. Evans T. and Kuwana T. Anal. Chem. 1977 49 1632. Laser D. and Ariel M. J. Electroanal. Chem. 1974 52 291. Vetter K. J . “Electrochemical Kinetics,” Academic Press New York 1967. Adams K. N. “Electrochemistry at Solid Electrodes,” Marcel Dekker New York 1969. Vetter K. J. J. Electrochem. 1952 56 797. p. 333. Plenum Press New York 1980 p. 320. Received September 8th 1982 Accented Sehtenzher 27th. 198