Langmuir 2007, 23, 7847-7852
7847
Investigating the Thermodynamic Causes Behind the Anomalously Large Shifts in pKa Values of Benzoic Acid-Modified Graphite and Glassy Carbon Surfaces Poobalasingam Abiman,† Alison Crossley,‡ Gregory G. Wildgoose,† John H. Jones,§ and Richard, G. Compton*,† Physical and Theoretical Chemistry Laboratory, UniVersity of Oxford, South Parks Road, Oxford, OX1 3QZ, United Kingdom, Materials Department, UniVersity of Oxford, Parks Road, Oxford, OX1 3PH, United Kingdom, and Chemical Research Laboratory, UniVersity of Oxford, Mansfield Road, Oxford, OX1 3TA, United Kingdom ReceiVed February 22, 2007. In Final Form: April 30, 2007 The difference between the values of 4-carboxyphenyl groups, covalently attached to either graphite (BAcarbon) or glassy carbon (BA-GC) surfaces, and benzoic acid in solution is explored using potentiometric titration and cyclic voltammetry. In solution, benzoic acid has a pKa of 4.20 at 25 °C. However, the observed pKa value on the graphitic surfaces shows significant deviations, with BAcarbon exhibiting a large shift to higher pKa values (pKa ) 6.45) in contrast to BA-GC, which is shifted to lower pKa values (pKa ) 3.25). Potentiometric titrations at temperatures between 25 and 50 °C allowed us to determine the surface pKa of these materials at each temperature studied and hence to determine the enthalpy, entropy, and Gibbs’ energy changes associated with the ionization of the carboxylic acid groups. It was found that the enthalpic contribution is negligible and that the changes in surface pKa values are entropically controlled. This suggests that solvent ordering/disordering around the interface strongly influences the observed pKa value, which then reflects the relative hydrophobicity/hydrophilicity of the different graphitic surfaces.
1. Introduction For many years it has been well known that the pKa value of a chemical species at an interface, usually a solid-liquid interface, can differ markedly from the solution-phase value. This effect has been observed for a wide number of relevant systems (e.g., phenyl carboxylic acids on glassy carbon electrodes,1-3 azo dyes,4 and self-assembled monolayers of thiols5-8 and thioctic acids9 on gold electrode surfaces). Usually such shifts are relatively small, being on the order of less than 1 pKa unit, and are usually shifted to a higher pKa value on the surface than for the same species in bulk solution. The interesting exception is found in Saby et al., who report a shift in the pKa of benzoic acid from a value of 4.2 in solution to a value of around 2.8 when the benzoic acid was covalently attached to the surface of a glassy carbon electrode.2 The reasons for the shifts in pKa are not fully understood but have been attributed to a variety of factors. These include such considerations as whether the interface is charged, the nature of the surface such as its hydrophobicity, lateral interactions between adsorbed species (e.g., hydrogen bonding), electronic interactions between surface-bound species and the bulk solid electrode, and * Corresponding author. E-mail:
[email protected]. Tel: +44 (0)1865 275413. Fax: +44 (0)1865 275410. † Physical and Theoretical Chemistry Laboratory. ‡ Materials Department. § Chemical Research Laboratory. (1) Liu, J.; Cheng, L.; Liu, B.; Dong, S. Langmuir 2000, 16, 7471. (2) Saby, C.; Ortiz, B.; Champagne, G. Y.; Be´langer, D. Langmuir 1997, 13, 6805. (3) Yang, G.; Shen, Y.; Wang, M.; Chen, H.; Liu, B.; Dong, S. Talanta 2006, 68, 741. (4) Wang, H.; Callahan, P. M. J. Chromatogr., A 1998, 828, 121. (5) Cao, X.-W. J. Raman Spectrosc. 2005, 36, 250. (6) Smalley, J. F. Langmuir 2003, 19, 9284. (7) Smalley, J. F.; Chalfant, K.; Nahir, T. M.; Bowden, E. F. J. Phys. Chem. B 1999, 103, 1676. (8) Sugihara, K.; Shimazu, K.; Uosaki, K. Langmuir 2000, 16, 7101. (9) Rooth, M.; Shaw, A. M. Phys. Chem. Chem. Phys. 2006, 8, 4741.
differing molecular environments (such as the solvation shell, solvent accessibility, etc.) of the species on the surface. Recently we have observed that the pKa values of 1-anthraquinonyl groups on the surface of graphite powder and carbon nanotubes of various morphologies (“bamboo-like” and “hollow-tube” multiwalled carbon nanotubes and single-walled carbon nanotubes) exhibit anomalously large changes in their pKa values compared to those of free anthraquinone in solution.10,11 These shifts are typically greater than 3 pKa units, which, to the best of our knowledge, is the largest change reported in the literature for carbon surfaces. This work has prompted us to investigate changes in the surface pKa value of species on graphitic surfaces further. In this report, we covalently modify both graphite powder and glassy carbon powder with 4-carboxyphenyl groups via the chemically activated reduction of the corresponding diazonium salt of benzoic acid. Potentiometric titration over a range of temperatures, from 25 to 50 °C, has allowed us to determine the pKa value of the 4-carboxyphenyl groups on the two different graphitic surfaces at each temperature. This then allows us to determine the thermodynamic parameters associated with this process such as the enthalpy, entropy, and Gibbs’ free energy change upon ionization of the carboxyl groups. The observed pKa values of 4-carboxyphenyl-modified graphite are found to be shifted by more than 2 pKa units to higher values; in contrast, the pKa of the 4-carboxyphenyl groups on the glassy carbon surface are found to be shifted by just over 1 pKa unit to lower values than that of benzoic acid in solution, in agreement with the work of Saby et al.2 and Liu et al.1 These changes in pKa were also observed by using the cyclic voltammetry of potassium ferricyanide in solutions of different pH (1.0-10.0) at a 4-carboxyphenyl-modified glassy carbon (10) Heald, C. G. R.; Wildgoose, G. G.; Jiang, L.; Jones, T. G. J.; Compton, R. G. ChemPhysChem 2004, 5, 1794. (11) Masheter, A. T.; Abiman, P.; Wildgoose, G. G.; Wong, E.; Xiao, L.; Rees, N. V.; Taylor, R.; Attard, G. A.; Baron, R.; Crossley, A.; Jones, J. H.; Compton, R. G. J. Mater. Chem. 2007, [online early access] DOI: 10.1039/6702492d.
10.1021/la7005277 CCC: $37.00 © 2007 American Chemical Society Published on Web 06/09/2007
7848 Langmuir, Vol. 23, No. 14, 2007
electrode and a composite electrode consisting of graphite powder modified with 4-carboxyphenyl groups and epoxy resin in a 1:3 ratio. Above the pKa of the 4-carboxyphenyl groups on the electrode surface, the peak current of Fe(CN)63- rapidly decreased as a result of the electrostatic screening of the charged surface carboxylate groups. The enthalpic contribution in all cases is found to be negligibly small, indicating that lateral interactions and/or hydrogen bonding is not responsible for these changes. However, the change in the surface pKa value is found to be controlled by the entropy change upon ionization of the 4-carboxyphenyl groups. This is interpreted as being due to a change in solvent ordering/disordering at the interface upon ionization, and the different behavior on graphite and glassy carbon surfaces is attributed to the different relative hydrophobicity/hydrophilicity of the two surfaces.
Abiman et al. Scheme 1. Reaction Scheme for the Synthesis and Subsequent Derivatization of Carbon Powder with 4-Carboxyphenyl Groups
2. Experimental Section 2.1. Reagents and Equipment. All reagents were purchased from Aldrich (Gillingham, U.K.) with the exception of glassy carbon powder (10-20 micrometers, spherical particles, type 1, Alfa Aesar, Heysham, U.K.) and were of the highest grade available and used without further purification. All solutions were prepared using deionized water from an Millipore (Vivendi, U.K.) UHQ-grade water system with a resistivity of not less than 18.2 MΩ cm at 25 °C. Solutions of known pH were prepared as follows: pH 1.0 - 3.5, 0.1 M HCl; pH 4.0 - 5.5, 0.05 M CH3COONa; pH 6.0 - 8.0, 0.05 M Na2HPO4 + 0.05 M NaH2PO4; and pH 8.5 - 10.0, 0.05 M disodium tetraborate. In addition, each solution contained 0.1 M KCl as the supporting electrolyte and was adjusted to the appropriate pH using 0.1 M HCl or 0.1 M NaOH. The synthetic graphite powder (Aldrich) consisted of irregularly shaped particles that were 2-20 µm in diameter. pH measurements were performed using a Hannah pH 213 pH meter. To calibrate the pH meter at elevated temperatures, a two-point calibration was made prior to each experiment using IUPAC standard buffer solutions (0.5 M sodium acetate + 0.5 M acetic acid, pH 4.6 at 25 °C; 0.05 M disodium tetraborate, pH 9.2 at 25 °C), which have a known pH at each temperature studied.12 Electrochemical measurements were carried out using a computercontrolled potentiostat, µAutolab computer (Eco-Chemie, Utrecht, Netherlands) with a standard three-electrode configuration. Either a glassy carbon electrode (GC, 3 mm diameter, BAS Technicol, U.K.) or a graphite composite electrode was used as the working electrode. The graphite composite electrode consisted of benzoic acid-modified carbon (BAcarbon) mixed with epoxy resin in a 1:3 ratio packed into a plastic holder, 3 mm in diameter. Electrical connection was made using a wire inserted through the rear of the electrode assembly. A bright platinum coil (99.99% Goodfellow, U.K.) and a saturated calomel reference electrode (SCE, Radiometer, Copenhagen, Denmark) acted as the counter and reference electrodes, respectively. For nonaqueous electrochemical experiments, a silver wire pseudoreference electrode replaced the SCE. Room temperature and elevated temperature experiments were performed in a doublewalled glass cell, of 20 cm3 volume, which was connected to a heated water bath at the appropriate temperature. X-ray photoelectron spectroscopy (XPS) was performed on a VG Clam 4 MCD analyzer system at the OCMS Begbroke Science Park, University of Oxford, U.K. using X-ray radiation from the Al KR band (hν ) 1486.7 eV). All XPS experiments were recorded using an analyzer energy of 100 eV with a takeoff angle of 90°. The base pressure in the analysis chamber was maintained at not more than 2.0 × 10-9 mbar. Each derivatized carbon sample studied was mounted on a stub using double-sided adhesive tape and then placed in the ultrahigh vacuum analysis chamber of the spectrometer. The analysis of the resulting spectra was performed using Origin 6.0, and the assignment of spectral peaks was determined using the UKSAF13 and NIST14 databases. (12) Covington, A. K.; Bates, R. G.; Durst, R. A. Pure Appl. Chem. 1985, 57, 531. (13) http://www.uksaf.org/.
2.2. Synthesis of 4-Carboxybenzenediazonium Tetrafluoroborate. The synthesis of 4-carboxybenzenediazonium tetrafluoroborate was achieved as follows:2 4-aminobenzoic acid (0.01 mol) was dissolved in 18 cm3 of warm hydrochloric acid, after which the solution was cooled to 0 °C and a solution of sodium nitrite (0.011 mol in 4 cm3 of water) was added dropwise for 20 min with stirring. The solution was filtered, and 1.48 g (0.013 mol) of sodium tetrafluoroborate was added in several small portions to the filtrate, which was stirred for another 15 min. After this time, the solution was filtered by suction and washed with ice water and cold ether. The powder was dried and recrystallized with a mixture of acetonitrile and ether. The resulting 4-carboxybenzenediazonium salt was stored under vacuum in a desiccator over dry alumina. 2.3. Chemically Activated Modification of Graphite or Glassy Carbon Powder with Benzoic Acid Groups. The chemically activated derivatization of graphite powder or glassy carbon powder was achieved by the following method (Scheme 1):15-18 Graphite or glassy carbon powder (0.5 g) was stirred into 10 cm3 of a 0.1 M solution of 4-carboxybenzenediazonium tetrafluoroborate (synthesized as described above) to which 10 cm3 of hypophosphorous acid (H3PO2, 50% w/w in water) was added. The solution was then gently stirred for 30 min at room temperature. Next the solution was filtered by water suction and washed with deionized water to remove any excess acid and finally with acetonitrile to remove any unreacted diazonium salt. This procedure was carried out twice on each sample of carbon powder to ensure that maximum derivatization was achieved. The graphite or GC powder derivatized with benzoic acid groups (BAcarbon and BA-GC, respectively) was then air dried by placing it inside a fume hood for 12 h prior to use. 2.4. Electrochemical Modification of a Glassy Carbon Electrode with Benzoic Acid Groups. First the glassy carbon electrode was cleaned by successively polishing with diamond paste (3.0-0.1 µm, Buehler). After being polished, the electrode was washed with water and sonicated in acetonitrile for 10 min. Electrochemical modification was carried out in acetonitrile containing 5 mM 4-carboxybenzenediazonium tetrafluoroborate, with 0.1 M tetrabutylammonium perchlorate as the supporting electrolyte. Cyclic voltammetry was performed from 0.5 to -1.0 V versus Ag with a scan rate of 100 mV. Upon first scanning in a reductive fashion, an irreversible reduction wave was observed at -0.15 V versus Ag (Figure 1), which can be attributed to the formation of the 4-carboxyphenyl radical from the parent diazonium salt with subsequent covalent bond formation between the 4-carboxyphenyl (14) http://srdata.nist.gov/xps/. (15) Pandurangappa, M.; Lawrence, N. S.; Compton, R. G. Analyst 2002, 127, 1568. (16) Pandurangappa, M.; Lawrence, N. S.; Jiang, L.; Jones, T. G. J.; Compton, R. G. Analyst 2003, 128, 473. (17) Leventis, H. C.; Streeter, I.; Wildgoose, G. G.; Lawrence, N. S.; Jiang, L.; Jones, T. G. J.; Compton, R. G. Talanta 2004, 63, 1039. (18) Wildgoose, G. G.; Pandurangappa, M.; Lawrence, N. S.; Jiang, L.; Jones, T. G. J.; Compton, R. G. Talanta 2003, 60, 887.
Anomalously Large Shifts in pKa Values
Langmuir, Vol. 23, No. 14, 2007 7849
Figure 1. Overlaid cyclic voltammograms of a glassy carbon electrode in a 5 mM solution of 4-carboxybenzenediazonium tetraflouroborate in acetonitrile containing 0.1 M TBAP. The scan rate was 100 mV s-1. group and the glassy carbon electrode surface. The observed reduction wave in the first scan disappears on consequent scans, indicating that maximum surface coverage is achieved and further polymer formation does not occur. After the modification, the electrode was washed by rinsed and sonicating in acetonitrile to remove any unreacted or physisorbed species. 2.5. Potentiometric Titration of Benzoic Acid, BAcarbon, and BA-GC Powders. Potentiometric titrations at each temperature studied were carried out as follows: 10 cm3 of an aqueous 0.01 M solution of benzoic acid was pipetted into a 20 cm3 double-walled glass cell connected to a temperature controlled water bath and allowed to come to thermal equilibrium. After which the initial pH was recorded using a potentiometric glass pH meter. The solution was then titrated using aliquots of a known volume of 0.1 M NaOH. The pH of the solution was measured 30 seconds after each addition of NaOH. The potentiometric titration of either BAcarbon or BA-GC powders was performed using 0.5 g of the modified carbon powder stirred into 10.0 cm3 of pure water for 2 min prior to titration against NaOH. Again the suspension was stirred for 30 s after each addition of NaOH. The pH of the suspension was then recorded. Note that the pH meter was calibrated at each temperature before taking readings using two IUPAC standard buffers that have a known pH at each temperature studied.12
3. Thermodynamic Analysis Benzoic acid is a weak acid and reacts with NaOH according to eq 1. Ka
PhCOOH + H2O y\z PhCOO- + H3O+ PhCOOH + NaOH f PhCOO- + Na+ + H2O
(1)
The pH of the solution depends on both the equilibrium dissociation constant, Ka, and the concentration of the dissociated benzoate and undissociated benzoic acid remaining in solution, as described by eq 2.
(
pH ) pKa + log
[PhCOO-] [PhCOOH]
)
(2)
Note that that the autoionization of water has a negligibly small effect and is therefore ignored. If the concentrations of benzoate and benzoic acid are known (or can be calculated), then the pKa of the species under consideration can be determined from the intercept of the appropriate plot of pH versus log([PhCOO-]/ [PhCOOH]). Having obtained the ionization constant, Ka, one can then determine the values of thermodynamic parameters
Figure 2. XPS spectra (wide scans 0-1296 eV) of blank graphite and BAcarbon (offset for clarity) for comparison.
such as the changes in entropy, enthalpy and Gibbs’ energy for the dissociation of benzoic acid via the integral form of the van’t Hoff isochore (eq 3a) and standard thermodynamic relationships given in eqs 3b and 3c.
ln K ) -
+ const (∆H RT )
(3a)
∆G ) -RT ln Ka
(3b)
∆G ) ∆H - T∆S
(3c)
4. Results and Discussion 4.1. XPS Characterization of Benzoic Acid-Modified Carbon. A wide scan (0-1200 eV) of both unmodified and benzoic acid-modified graphite powders (Figure 2) and unmodified glassy carbon and BA-GC powders (Supporting Information, Figure 1) revealed two major peaks with binding energies of ca. 285.6 and 533.6 eV corresponding to emission from the C 1s and O 1s levels, respectively. Two additional spectral features are observed at ca. 745 and 997 eV corresponding to the OKLL and CKLL Auger emissions. Oxygen-containing surface functional groups such as carboxyl, quinonyl, phenol ether, and sigma carbonyl groups are known to decorate the surfaces of graphitic carbons, to which the O 1s signal of blank graphite or GC corresponds. The presence of titratable protons on such a group is addressed below. After modification, the atomic percentage of oxygen on the surface, as measured from the O 1s signal increased by 3.4% for both BA-modified graphite and glassy carbon powder. Thus the introduction of benzoic acid onto these carbon surfaces accounts for ca. 1.7% of the total oxygen-containing functionality on the surface. Furthermore, this implies that the benzoic acid contributes to ca. 10% of the C 1s emission from the graphitic surface. A crude calculation using the molecular dimensions of benzoic acid and the lattice parameters for a perfect sheet of graphite indicate that this is equivalent to an almost monolayer coverage of benzoic acid on the graphite and the glassy carbon surface (10-11 mol cm-2 for both). Detailed XPS spectra were also recorded over the C 1s and O 1s regions of blank graphite, BAcarbon, blank GC, and BAGC powders (10 cumulative scans over each region). Unfortunately, the spectral emissions from the different relevant chemical species on the surface, such as carboxyl, hydroxyl, or quinonyl groups, are separated by less than 2 eV. This prevents us from performing a quantitative deconvolution of the spectra into the relevant contributions from each group, and only the total C 1s or O 1s emissions can be measured.
7850 Langmuir, Vol. 23, No. 14, 2007
Abiman et al.
Figure 4. Plots of ln Ka versus 1/T for (a) benzoic acid in solution and (b) BAcarbon.
Figure 3. (a) Plot of pH vs volume of NaOH added at 25 °C and (b) the corresponding plot of pH vs log{[PhCOO-]/ [PhCOOH]}at 25 °C for BAcarbon.
4.2. Potentiometric Titration of Benzoic Acid in Solution and in BAcarbon and BA-GC Powders. The potentiometric titration of benzoic acid (described in section 2.5) produced a sigmoidal plot of pH against volume of NaOH (Supporting Information, Figure 2). The exact concentration of benzoic acid initially present in solution was calculated from the equivalence point volume of NaOH, where stoichiometric amounts of reactants have been mixed together. After each addition of NaOH, the concentration of benzoate ions and hence the concentration of benzoic acid remaining in solution was obtained from the concentration of hydroxonium ions, as determined by the solution pH and the amount of NaOH added. The pKa of benzoic acid was determined graphically using eq 2 and was found to be 4.18 at 25 °C (Supporting Information, Figure 2b), in agreement with the literature value of 4.20.19 Next a potentiometric titration of the BAcarbon and BA-GC powders was performed at 25 °C, which also produced sigmoidal titration curves, as shown in Figure 3a for BAcarbon and Supporting Information Figure 3a for BA-GC. In the case of BAcarbon, the surface pKa value was found to be 6.45 at 25 °C using eq 2 (Figure 3b). This value is significantly higher than that of benzoic acid in solution. In contrast, the surface pKa of BA-GC powder (Supporting Information, Figure 3b) was found to be 3.25 at 25 °C, which is lower than that of benzoic acid in solution. Note that titrations were also performed on both blank graphite (Supporting Information, Figure 4) and glassy carbon powders; however, the number of titratable protons was negligable (>1%, see below) compared to those for the BA-modified powders. Liu et al. reported that the pKa of aminobenzoic acid covalently grafted onto a GC electrode through the amine group is 3.0,1 and (19) Petrov, J. G.; Mobius, D. Langmuir 1996, 12, 3650.
Saby et al. reported a pKa value of 2.8 for benzoic acid covalently attached to a GC electrode surface in a similar manner to that used here;2 these are reassuringly similar values to those reported here. Both of these reported pKa values are lower than for the free species in solution. (pKa values for aminobenzoic acid and benzoic acid are 4.001 and 4.20,2,19 respectively.) The ionexchange capacity of BAcarbon and BA-GC was also estimated from the titration curves as 5 ( 1 × 10-5 mol g-1 for both materials, whereas for blank graphite and GC powder the ionexchange capacity was less than 1% of these values at ca. 2 × 10-7 mol g-1. Both of these ion-exchange capacities and the surface coverage of BA molecules on each carbon surface indicate that the BA groups are at less than a monolayer coverage, implying that lateral interactions between the groups are unlikely to be the cause of such large pKa changes. Clearly the substrate material that the benzoic acid is covalently attached to plays some role in changing the surface pKa value. To investigate this effect further, a series of potentiometric titrations were carried out on benzoic acid, BAcarbon, and BAGC powder at elevated temperatures (25-50 °C). The surface pKa values for each material were calculated at each temperature studied as described above and were used to calculate the enthalpy, entropy, and Gibbs’ energy of the ionization of benzoic acid using eqs 3a-3c. The enthalpy change, ∆Ha, for the ionization of benzoic acid, BAcarbon, and BA-GC was determined by plotting ln Ka versus 1/T (Figure 4). In all three cases, the plots were linear, with an R2 value of not less than 0.997. The enthalpy changes were found to be negligibly small at -0.02, -0.21, and 0.16 kJ mol-1, respectively. Table 1 details the Gibbs’ energy and entropy changes for the ionization of benzoic acid in solution, which are found to be reassuringly close to the literature values obtained using conductivity measurements.20 Tables 2and 3 list the entropy and Gibbs’ energy changes for BAcarbon and BAGC powder, respectively. In the case of BAcarbon, the entropy change upon ionization is significantly greater than that of benzoic acid in solution, whereas for BA-GC powder the entropy change is slightly lower. Therefore, we conclude that it is the entropic (20) Read, A. J. Sol. Chem. 1981, 10, 437.
Anomalously Large Shifts in pKa Values
Langmuir, Vol. 23, No. 14, 2007 7851
Table 1. Tabulated Values of the Change in Gibbs’ Free Energy and Entropy for the Ionization of Benzoic Acid in Solution from 25 to 45 °C T/°C
∆G° (kJ mol-1)
∆S° (J K-1 mol-1)
25 30 35 40 45 25 (ref 20)
23.8 24.2 24.7 25.1 25.5 24.0
-79.8 -79.9 -80.0 -80.1 -80.2 -81.0
Table 2. Tabulated Values of the Change in Gibbs’ Free Energy and Entropy for the Ionization of BAcarbon Powder from 25 to 45 °C T/°C
∆G° (kJ mol-1)
∆S° (J K-1 mol-1)
25 30 35 40 45 50
36.8 37.8 38.7 39.6 40.6 41.3
-124.0 -125.5 -126.6 -127.2 -128.2 -128.4
Table 3. Tabulated Values of the Change in Gibbs’ Free Energy and Entropy for the Ionization of BA-GC Powder from 25 to 45 °C T/°C
∆G° (kJ mol-1)
∆S° (J K-1 mol-1)
25 30 35 40 45 50
18.5 18.5 18.6 18.8 18.8 19.0
-61.0 -60.5 -59.9 -59.4 -58.6 -58.4
factor that is responsible for the large shifts in the surface pKa values of BAcarbon and BA-GC. The negligibly small enthalpic changes indicate that lateral interactions and/or hydrogen bonding do not play a significant role in altering the pKa value, in agreement with the surface coverage calculated from XPS and the ionexchange capacities of these materials. The entropy change is found to be the dominant factor affecting the observed pKa value. It is well known that graphite surfaces are relatively hydrophobic in nature, whereas GC surfaces are slightly more hydrophilic, probably because of its structure and the relatively larger number of polar/hydrogen-bonding oxygen-containing groups on its surface compared to those for graphite. Because the ionization process and subsequent charge generated on both surfaces are likely to be similar in both cases, we can speculatively say that the entropic factor most likely reflects a change in the solvent ordering around the interface, implying that it is the relative hydrophobicity/hydrophilicity of the graphite and GC surface that is the major factor influencing the changes in surface pKa values. Because we can measure only relative changes in entropy rather than absolute values, we are unable to state whether the solvent is more or less ordered at the GC surface than at the graphite surface and before and after ionization for each case, but the relative differences in hydrophobicity/hydrophilicity between graphite and GC can certainly be used to explain why the entropy changes have different values and cause the pKa to then shift in opposite directions compared to the value of benzoic acid in free solution. 4.3. Voltammetry of Fe(CN)63- at a Benzoic Acid-Modified Glassy Carbon Electrode. To explore the effect of the pKa of benzoic acid on a glassy carbon surface further, a GC electrode was modified with benzoic acid as described in section 2.4. Cyclic voltammetry was performed on the modified GC electrode in a 1 mM aqueous solution of potassium ferricyanide in buffer solutions of differing pH (pH 1.0-7.0, Figure 5a). From Figure 5a, it is apparent that as the pH of the solution increases, the peak
Figure 5. (a) Overlaid cyclic voltammograms of a benzoic acidmodified glassy carbon electrode in 1 mM Fe(CN)63- from pH 1 to 7. The scan rate was 10 mV s-1. (b) Corresponding plot of anodic peak current vs pH.
current decreases markedly at the modified electrode. Figure 5b details a plot of oxidative peak current versus pH for the ferricyanide couple. A large decrease in the observed peak current occurs when the solution pH reaches values above ca. pH 3. This observation agrees with the surface pKa value of 3.25 for BAGC powder determined using the potentiometric titration method above. This “blocking” behavior of the benzoic acid-modified GC electrode can be explained by considering electrostatic interactions between the modified surface and the electroactive probe and electrolyte.21-28 Above the pKa of the BA-GC surface, the 4-carboxyphenyl groups are deprotonated. These negatively charged surface groups repel the Fe(CN)63- anion from the surface, thus preventing any outer-sphere electron transfer from taking place. 4.4. Voltammetry of Fe(CN)63- at a BAcarbon Composite Electrode. To investigate whether a benzoic acid-modified graphite electrode exhibited similar behavior in the presence of the ferricyanide anion as the modified GC electrode, the BAcarbon powder was combined into a composite electrode by mixing the BAcarbon powder with epoxy resin in a 1:3 ratio. Again, cyclic voltammetry was performed using 1 mM Fe(CN)63- in buffer solutions of varying pH (1.0-10.0), and the voltammetric (21) Cheng, Q.; Brajter-Toth, A. Anal. Chem. 1995, 67, 2767. (22) Cheng, Q.; Brajter-Toth, A. Anal. Chem. 1996, 68, 4180. (23) Finklea, H. O.; Avery, S.; Lynch, M.; Furtsch, T. Langmuir 1987, 3, 409. (24) Godinez, L. A.; Castro, R.; Kaifer, A. E. Langmuir 1996, 12, 5087. (25) Madoz, J.; Kuznetzov, B. A.; Medrano, F. J.; Garcia, J. L.; Fernandez, V. M. J. Am. Chem. Soc. 1997, 119, 1043. (26) Miller, C.; Cuendet, P.; Gratzel, M. J. Phys. Chem. 1991, 95, 877. (27) Miller, C.; Cuendet, P.; Gratzel, M. J. Phys. Chem. 1991, 95, 5225. (28) Redepenning, J.; Tunison, H. M.; Finklea, H. O. Langmuir 1993, 9, 1404.
7852 Langmuir, Vol. 23, No. 14, 2007
Figure 6. (a) Overlaid cyclic voltammograms of a BAcarbon composite electrode in 1 mM Fe(CN)63- from pH 1 to 10. The scan rate was 10 mV s-1. (b) Corresponding plot of anodic peak current vs pH.
response toward Fe(CN)63- decreased with increasing pH as shown in Figure 6a. The plot of anodic peak current with pH (Figure 6b) revealed a dramatic decrease in the anodic peak current at pH values greater than ca. 6. This is again in agreement with the measured surface pKa value of 6.45 for BAcarbon.
5. Conclusions Graphite powder and glassy carbon powder covalently modified with 4-carboxyphenyl groups exhibit significantly large changes
Abiman et al.
in the measured pKa values compared to those for benzoic acid in solution (pKa ) 6.45, 3.25, and 4.18, respectively, at 25 °C). Potentiometric titration and voltammetric experiments, using ferricyanide as a redox probe, indicate that the pKa of the 4-carboxyphenyl groups on the surface of graphite powder is shifted by more than 2 pKa units higher than that of benzoic acid in solution. In contrast, when the 4-carboxyphenyl groups are covalently attached to a glassy carbon surface the pKa is shifted to lower values by more than 1 pKa unit. The underlying thermodynamic parameters such as the Gibbs’ energy, enthalpy, and entropy of the ionization of the 4-carboxylphenyl groups on the different carbon surfaces, as well as benzoic acid in solution, were determined. The enthalpic contribution to this process was negligibly small in all cases, indicating that any lateral interactions on the carbon surface and/or hydrogen bond formation were not responsible for the observed changes in pKa. Instead, the different surface pKa values were found to be controlled by the entropic term. This suggests that the ordering/disordering of solvent molecules at the interface between the carbon substrate and the solution is likely responsible for the variation in the observed pKa values, reflecting the different hydrophobicity/hydrophilicity of the graphite and glassy carbon surfaces. The significant changes in pKa values of substrates on carbon surfaces should be an important consideration when designing synthetic reactions to modify these materials further for a variety of applications including advanced material design. Acknowledgment. P.A. acknowledges support via a Dorothy Hodgkin Postgraduate Award (DHPA). G.G.W. thanks St John’s College, Oxford, for a Junior Research Fellowship. Supporting Information Available: Supplementary figures. This material is available free of charge via the Internet at http://pubs.acs.org. LA7005277
