J. CHEM. SOC. FARADAY TRANS., 1994, 90(11), 1533-1535 Quartz Crystal Microbalance Study of the Adsorption of Ions onto Gold from Non-aqueous Solvents Andrew P. Abbott," David C. Loveday and A. Robert Hillman Chemistry Department, University of Leicester, Leicester, UK LE1 7RH This work investigates the adsorption of ions onto a gold electrode surface using an electrochemical quartz crystal microbalance. It was found that in anisole solutions adsorption of a monolayer of tetrafluoroborate anions or tetrabutylammonium cations occurred at different potentials, whereas in ethanol only tetra-butylammonium cations were adsorbed. The implications that this has upon electron-transfer processes are discussed. The use of non-polar solvents has become widespread in elec- trochemistry because of their wide potential windows, chemi- cal inertness and high solubility for organic electroactive species.' High conductivity has been achieved in such media using quaternary ammonium salts as electrolytes.2 This has led to their use in studying solvent effects on electron-transfer kinetics by electrochemical method^.^,^ Recent work by Fawcett and Fedurco' has shown that the ionic radius of quaternary ammonium ions used as an electrolyte has a marked effect upon the kinetics of electron transfer in a range of non-aqueous solvents.It is suggested that this is caused by the adsorption of the ions at the electrode surface which would present a barrier to electron transfer. The measure- ment of ion adsorption is difficult, particularly by the usual method of measuring electrode capacitances. The aim of the present work is to investigate ion adsorp- tion using an electrochemical quartz crystal microbalance (EQCM).This technique measures the resonant frequency variation (An of a quartz crystal oscillator from its base value cfo)that accompanies a change in mass (AM) attached to the crystal. When the additional mass is small and rigidly coupled,6 (AflHz) = -(2/Pv)f;(AM/g cm -2, (1) where p is the density of the quartz and v is the wave velocity in the quartz. The EQCM is an extremely sensitive device for measuring in situ processes at the electrode/solution interface with submonolayer resolution and has found a variety of electrochemical application^.'.^ Although this technique has been extensively applied to the study of polymer films, very little work has been carried out on the adsorption of ions onto electrodes.One example of this, however, was work by Deakin et aL9 which studied the adsorption of bromide and iodide onto gold and found that the weight gain corres-ponded well to a complete close-packed monolayer of bromide. We are interested in studying the specific adsorption of ions onto gold from electrolytes with different characteristics, as determined by solvent relative permittivity, acidity and electron-donating ability. In this work, we describe the adsorption of tetrabutylammonium tetrafluoroborate (TBABF,) from anisole (E = 4.3), ethanol (E = 24.3) and meth- anol (E = 32.6). Experimental Anisole (Aldrich, 99%) was distilled at reduced pressure under a nitrogen atmosphere.Ethanol and methanol (BDH, AnalaR) were used as received. Tetrabutylammonium tetra- fluoroborate (TBABF,) (Fluka, puriss, electrochemical grade) was used as received. The EQCM instrumentation and cell configuration have been described elsewhere. lo All of the experiments described were performed in potentiostatic mode controlled by an Oxford electrodes modular potentiostat. Data were acquired and stored by an IBM-ATX computer using a Keithley series 570 data acquisition workstation and a Hewlett-Packard 5334B frequency counter. Gold-coated 10 MHz AT-cut quartz crystals were supplied by International Crystal Manufacturing Company (Oklahoma City, OK, USA).The crystals had a 900 A thick gold layer deposited on the surface in a keyhole shape with a central disc (area, 0.211 cm') sensitive to mass changes. All potentials given are with respect to the Ag I AgBF, reference electrode" which was in contact with 0.1 mol dm-3 TBABF, ,separated from the bulk solution by a glass frit. All solutions were bubbled with argon for at least 10 min before use. Results and Discussion Fig. l(a) shows a plot of weight gain against potential for a gold electrode in a 0.1 mol dm-3 TBABF, in anisole solution at 23°C at a sweep rate of 1 mV s-'. Note that the results presented here are from a single scan, unlike those previously presented by Deakin et al.' which were an average of several scans.When the potential was scanned from -0.2 to +0.7V there was an increase in mass which can be attributed to the adsorption of a negatively charged species onto the gold elec- trode. This could either be BF4- ions or the triple ion [TBA(BF,),] -. While the concentration of free BF,- ions is small (about mol dm-3), the repulsive force between the positively charged electrode and the TBA ion would prob- + ably cause dissociation of the triple ion somewhere close to the electrode surface. When the potential was scanned in the negative direction the adsorbate was desorbed to a potential of cu. -0.3 V, whereupon the electrode again started to increase in mass. This was ascribed to the adsorption of TBA' ions {or, less likely, the triple ion [(TBA),BF,]+). When the potential was swept in the positive direction from -1.2 V the adsorbate was again desorbed from the electrode surface.The adsorp- tion processes are seen to reach a maximum on both the positive and negative potential scans. Before proceeding to interpret the data, we make some comments on possible experimental artefacts. First, the observed frequency (mass) changes are not a consequence of impurity adsorption. Experiments employing less pure electrolytes resulted in monotonic mass increases that were (a)independent of poten- tial and potential scan direction and (b) quantitatively differ- ent (cu. lo00 ng cm-2 on the timescale of the experiment in Fig. 1) to the mass changes for 'clean' electrolytes.Secondly, the discrepancy in the weight change at the start and finish of I I I I I I -1200 -800 -400 0 400 800 E/V vs. Ag 1 AgBF, 5-a 40 -1200 -800 -400 0 400 800 E/V vs. Ag IAgBF, Fig. 1 Mass change us. potential for a gold electrode in a 0.1 mol dm-’ solution of TBABF, in anisole at 23 “C at a sweep rate of (a) 1 and (b)10 mV sP1 the experiment is small and probably due to temperature changes during the long timescale of each run (>1 h). The observation that the drift is so small is proof of the stability of the system. Knowing the ionic radius of the two adsorbed ions (BF,-= 2.02 A and TBA+ = 4.13 A ’), and assuming each ion occupies a box corresponding to its diameter, d, the mass of an adsorbed monolayer, M, ,can be calculated : M, = (d-2M,)/N, (2) where M, is the molar mass and N, is Avogadro’s number.The corresponding masses of an adsorbed monolayer of TBA’ and BF,- are 59 and 89 ng cm-’, respectively. These agree well with the values of ca. 50 and 90 ng cm-’ shown in Fig. l(a).This means that in media of such low relative per- mittivity, at potentials positive and negative of the potential of zero charge (pzc), the electrode is covered with a close- packed layer of specifically adsorbed electrolyte ions. Fig. l(b)shows a repeat of the above experiment at a faster scan rate (10 mV s-I). Pronounced hysteresis is observed between the forward and reverse scans. The adsorption and desorption of the charged species from the electrode surface reflect an equilibrium between the electrostatic and disper- sion interactions of the ion and the electrode and the solva- tion interactions of the ion and solvent molecules.The adsorption and desorption processes are further complicated by the ion-ion interactions in media of low permittivity. Close to the electrode surface, ion aggregates will dissociate due to the electric field. The ions with a charge opposite to J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 that of the electrode will experience a large electrostatic inter- action and hence be strongly attracted to the electrode. The desorption of the ion is difficult since the solvation of a single ion is thermodynamically unfavourable owing to the low sol- vation energy.Thus the ion will remain adsorbed until the concentration of counter-ions in the double layer at the electrode/solution interface becomes significant, when it can form an ion pair and be solvated, i.e. when the potential is closer to the pzc. Two interesting points arise from the above results. First, the minimum in the mass-potential plot at both scan rates occurs at ca. -0.3 V. This minimum in the mass-potential plot occurs in this region, irrespective of the potential from which the scan starts. It can be inferred that this must be the pzc of the electrode. Secondly, the above result has signifi- cant consequences for electrochemical experiments carried out in such media. For redox processes where El/’ < -0.8 V the electrode will be covered with a close-packed monolayer of TBA+ ions.This will block the electrode and any electron transfer will have to occur over a distance of cu. 8.3 A. This will lead to a decrease in apparent rate constants for pro- cesses studied in non-polar media, as observed by Fawcett and Fedurco.’ It would also lead to a marked electrolyte effect on the apparent rate constant of electron-transfer pro- cesses. The same would also be true, although to a lesser extent, of redox processes carried out at El/*> 0.5 V, where BF,-blocks the surface. Ion adsorption could also present problems for the study of phase transformation reactions such as metal deposition. Recently, a large amount of work has been carried out on the deposition of metals from aromatic and polyaromatic sol- vents (E < 5).’,’’ Metal ions with a relatively positive reduction potential, such as copper and zinc, could be depos- ited easily.Metal ions such as titanium and tungsten with deposition potentials < -1 V appeared to be reduced,” but no macroscopic deposits could be obtained. The above obser- vations would account for this anomaly because, although the metal ions could be reduced to the corresponding zero valence state, they could not be incorporated into the metal lattice because the electrode would be covered with a layer of quaternary ammonium ions. If the above ideas are correct then similar anomalies will also be observed for other com- monly used electrochemical solvents such as tetrahydrofuran, dichloroethane and dichloromethane.Fig. 2 shows a plot of mass gain of a gold electrode against potential in a 0.1 mol dm-3 TBABF, in ethanol solution. In this case, no mass change was observed on a potential sweep -20 1 -1000 -800 -600 -400 -200 0 200 E/V vs. Ag IAgBF, Fig. 2 Mass change us. potential for a gold electrode in a 0.1 mol dm-’ solution of TBABF, in ethanol at 23°C at a sweep rate of 5 mV s-l J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 12 -7 8-5 UJ 5.4-f a 0--1200 -1000 -800 -600 -400 -200 0 200 E/V vs. Ag 1 AgBF, Fig. 3 Mass change us. potential for a gold electrode in a 0.1 mol dm-3 solution of TBABF, in methanol at 23 "C at a sweep rate of 5 mV s-' from 0 up to +0.7 V. On sweeping the potential in a negative direction from 0 V, however, an increase in mass of the elec- trode was again observed.These observations can be inter- preted as showing no adsorption of BF,- over the range of positive potentials scanned, which probably results from a stronger solvation of the anion by the solvent molecules, making the desolvation/ion adsorption process more difficult. At potentials negative of -0.6 V the adsorption of TBA' is again observed with a coverage of ca. one monolayer. This shows that even in higher relative permittivity solvents, such as ethanol (E = 24.3), the specific adsorption of quaternary ammonium ions could still produce experimental artefacts. Fig. 3 shows a plot of mass gain of a gold electrode against potential in a 0.1 mol dm-3 TBABF, in methanol solution.Results similar to those observed in ethanol were obtained. Note, however, that the mass increase corresponds to only half of one monolayer. This suggests that the higher the polarity, the greater the solvation of the electrolyte ions and the lower the extent of adsorption on the electrode surface. The results of this work support the ideas of Fawcett and Fedurco that ions are adsorbed onto electrode surfaces from low relative permittivity solvents containing electrolytes. This has large implications for the study of electron-transfer kinetics from such solvents, as the apparent rate constants will be affected not only by the solvent, but also by the separation of the electroactive species from the electrode owing to the presence of adsorbed electrolyte ions.Conclusion In solvents of low relative permittivity, quaternary ammon- ium electrolytes are found to be adsorbed on a gold electrode surface forming a close-packed monolayer. We have no evi- dence for the adsorption of triple ions of either charge type. This adsorption process appears to be only partially reversible and this may result in experimental artefacts when measuring electron-transfer kinetics. It was found that in ethanol and methanol anion adsorption did not occur, but cation adsorption occurred in both solvents, although to a lesser extent in methanol. This work was funded by a grant from the Leicester/ Loughborough Research Fund. The authors would like to thank Mrs. G. Lonergan for her help with the experiments. References 1 A. P. Abbott, Chem. SOC.Rev., 1993,22,435. 2 A. P. Abbott and D. J. Schiffrin, J. Chem. Soc., Faraday Trans., 1990,86,1453. M. J. Weaver, Chem. Rev., 1992,92,463, and references therein. W. R. Fawcett, Lmgmuir, 1989,5,661. W. R. Fawcett and M. Fedurco, J. Phys. Chem., 1993,97,7075. G.Z. Sauerbrey, 2. Phys., 1959,155,206. D. A. Buttry, in Electroanalytical Chemistry, ed. A. J. Bard, Marcel Dekker, New York 1991, vol. 17, p. 1. 8 M. R. Deakin and D. A. Buttry, Anal. Chem., 1989,61, 1147A. 9 M. R. Deakin, T. T. Li and 0.R. Melroy, J. Electroanal. Chem., 1988,243, 343. 10 S. Bruckenstein and M. Shay, Electrochim. Acta, 1985,30, 1295. 11 A. P. Abbott, E. E. Long, A. Bettley and D. J. Schiffrin, J. Elec-troanal. Chem., 1989,261,449. 12 A. P. Abbott, A. Bettley and D. J. Schiffrin, J. Electroanal. Chem., 1993,347,153. Paper 3/07103K; Received 1st December, 1993