2826
J. Phys. Chem. B 2009, 113, 2826–2831
Superoxide Protonation by Weak Acids in Imidazolium Based Ionic Liquids Alice Rene´,† Didier Hauchard,† Corinne Lagrost,‡ and Philippe Hapiot‡,* Sciences Chimiques de Rennes (Equipe MaCSE), UniVersite´ de Nationale Supe´rieure de Chimie de Rennes, CNRS, UMR N° 6226, Campus de Beaulieu, Bat 10C, 35042 Rennes Cedex, France ReceiVed: NoVember 21, 2008; ReVised Manuscript ReceiVed: December 26, 2008
The reactivity of the superoxide anion versus a series of substituted phenols was investigated in a common ionic liquid, 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([BMIm][TFSI]) and for comparison in dimethylformamide (+0.1 mol · L-1 of Bu4NPF6 as supporting electrolyte). On the whole, the mechanism of the reduction of O2 in the presence of the different phenols was found to be very similar in [BMIm][TFSI] and in DMF: A 2-electron mechanism involving a succession of electrochemical and protonation steps. These steps are accompanied by the production of the corresponding phenolate that was identified through its oxidation potential. The reactivities of the phenols were observed to slightly differ in the two media. A qualitative analysis of the voltammogram allows a classification of the reactivities of the superoxide as a function of the phenols. As previously found in organic solvents, the protonation of superoxide by phenol is an uphill reaction that is rendered possible thanks to a subsequent irreversible electron transfer. Its pKa is estimated to be around 4-5 units lower than that of unsubstituted phenol. Introduction Room-temperature ionic liquids (RTILs) have been proposed for numerous applications in the general field of chemistry.1,2 Owing to their special properties, including good intrinsic conductivity and wide electrochemical windows, electrochemical methods allow detailed investigations of their physical and chemical behaviors.3 Several experimental observations have first evidenced a strong stabilization of the electrogenerated species that have important consequences on the kinetics of homogeneous4 and heterogeneous electron transfer5 and on the associated chemical reactions.6 Another remarkable result concerns the mass transport of small charged molecules in RTILs. It was found that the diffusion coefficients of the oxidized and reduced forms of a redox couple could considerably differ in RTILs.3,7 The most spectacular dissimilarities were revealed by the one-electron reduction of dioxygen to superoxide ion in commonly used RTILs in which the diffusion coefficients of O2-• were reported to be 1/30 to 1/50 of the oxygen values.8,9 These observations suggest that the solvated O-• 2 diffuses in the RTIL with several associated cations, resulting in a much larger entity.3c In the framework of the Stokes-Einstein relationship, such unusual variations would correspond to a “solvated” O2-• 10000 times heavier than the neutral O2.9 In summary, all these observations show that O2-• in an ionic liquid is a very different species from O2-• in a conventional solvent.10 Because of the importance of the reactive oxygen species (ROS), the electrochemistry of O2 has been extensively investigated in conventional media for these last three decades (in water or in organic solvent)11 and more recently in roomtemperature ionic liquids.12-16 Indeed, superoxide ions could be produced in RTIL, and quasi-reversible cyclic voltammograms, recorded at relatively low scan rates, have been reported in several ionic liquids.12 The reactivity of superoxide toward the cations of RTILs was recently examined on a long time * To whom correspondence should be addressed. Email: philippe.hapiot@ univ-rennes1.fr. † Ecole Nationale Supe´rieure de Chimie de Rennes. ‡ Universite´ de Rennes 1.
SCHEME 1: [TFSI] Anion and [BMIm] Cation Used to Prepare the Ionic Liquid
scale and O2-• was found to slowly react with the cations of the RTIL.15 In the same work, the authors point out the role of “ion pairing” on the reactivity of ROS for explaining the unusual diffusion O2-• in RTILs.15 All these previous investigations open new questions about the possible modification of reactivity of this important intermediate with the media. An important issue concerns the nucleophilicity or the basicity of O2-•, since the protonation of this intermediate is a key step for many chemical and biological processes where ROS are engaged.11 In the present work, we have focused on the reactivity of O2-• toward different weak acids. As a homogeneous family of acid, we have chosen four different phenolic compounds (4-chlorophenol, phenol, p-cresol, 2,4-dimethylphenol) as proton donor. Phenols are interesting compounds because their acidity and redox properties could be easily tuned from their substitution while keeping the same acidic structure.17 Experiments in RTIL were performed in a common ionic liquid, 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([BMIm][TFSI]) (see Scheme 1), and compared with the data obtained in dimethylformamide (DMF + 0.1 mol · L-1 of NBu4PF6 used as supporting electrolyte) taken as an example of organic solvent. The reduction of O2 and the reactivity of O2-• have been extensively studied in DMF, and the reduction mechanism in the presence of weak acid is now clearly established (see the discussion below) allowing easy comparisons between both media.18 Especially, the rate constants for the protonation of O-• 2 by phenolic acid have been measured in DMF. Experimental Section Chemicals. N,N-dimethylformamide (DMF), extra dry quality ([H2O] e 0.005%, stored over molecular sieve 3Å) and the
10.1021/jp810249p CCC: $40.75 2009 American Chemical Society Published on Web 02/10/2009
Superoxide Protonation by Weak Acids tetrabutylammonium hexafluorophosphate (Bu4NPF6) of electrochemical grade were purchased from Fluka. The commercially available phenols were purchased with the highest purity grade (Acros Organics or Alfa Aesar) and used without further purifications. 1-Butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([BMIm][TFSI]) ionic liquid was prepared from LiN(CF3SO2)2 (Solvionic, France) and BMImCl aqueous solutions according to previously published procedures.19 The samples were purified by repeated washing with H2O, filtered over neutral alumina and silica. Prior to each experiment, vacuum pumping carefully dried RTILs overnight and the amount of residual water was measured with Karl Fischer titration (Karl Fischer 652 Metrohm). The amount of water measured in our samples ranged from 100 to 200 ppm. Before bubbling gas into the cell, argon, air, and oxygen were dried using a molecular sieve-drying column. Indeed, [BMIm][TFSI] is hygroscopic and the presence of even low amounts of water could have a dramatic effect on the electrochemical behavior of these media, including a shrinking of the potentials window, changes of the diffusion coefficients or a possible reaction with the electrogenerated superoxide, hence leading to the irreversibility of the O2/O-• 2 system and/or a modification of its reactivity versus other acids.20 Electrodes and Cyclic Voltammetry Measurements. In RTIL, experiments were carried out in 1 mL [BMIm][TFSI]. The ionic liquid was saturated by bubbling in the cell dry air or oxygen for 15 min prior to experiments. In the case of O2saturated solutions, the solubility of oxygen was estimated around 4.5 × 10-3 mol · L-1 at room temperature. This value was calculated from the peak current considering the values of the diffusion coefficients for O2 and O2-•, DO2 and DO-• , 2 previously measured in [BMIm][TFSI] ) 6.1 × 10-6 and 3.1 × 10-7 cm2 · s-1, respectively.9 A full simulation of the voltammogram was used to take into account the effect of a slow electron transfer. Voltammograms were recorded with a homemade potentiostat using a three-electrode setup equipped with a positive feedback circuit for the ohmic drop compensation.21 The resistance of RTILs required a careful treatment of the ohmic drop to avoid artifacts on the recorded signal.3c The working electrode was a 1-mm diameter disk gold electrode, the reference electrode was a polypyrrole quasi-reference prepared according previously published procedure,22 and a platinum wire served as the counter electrode. The Au disk electrode was polished using a 1-µm diamond paste, then washed with acetone and dried. For measurements involving the oxidation of phenolate, it was necessary to clean the electrode because of the possible reactivity of the phenoxyl radical with other radicals or the electrode surface. For cyclic voltammetry experiments performed in DMF (+ 0.1 mol · L-1 Bu4NPF6, used as supporting electrolyte), a similar three-electrode setup was used. The potentiostat was a PGSTAT302 (Autolab instrument, Eco Chemie B.V., Utrecht, The Netherlands) with a gold disk electrode (1 mm diameter) as the working electrode, an Ag/AgCl reference and a platinum wire as the counter electrode. The DMF solution was saturated by bubbling dry air or pure O2 for 10 min. Solubility of oxygen was taken as respectively [O2] ≈ 0.98 × 10-3 and [O2] ≈ 4.8 × 10-3 mol · L-1.23,24 All experiments were performed at room temperature (20 ( 2 °C) and potential values are given against the ferrocene/ferrocenium couple that was added at the end of the measurements.
J. Phys. Chem. B, Vol. 113, No. 9, 2009 2827
Figure 1. Cyclic voltammograms of dioxygen (A) in [BMIm][TFSI] (CO2 ≈ 10-3 mol · L-) and (B) in DMF + 0.1 mol · L-1 NBu4PF6 (CO2 ) 0.94 × 10-3 mol · L-1). Scan rate V ) 0.2 V · s-1. Conditions: Au electrode (d ) 1 mm); air-saturated solutions.
Simulations of voltammograms were performed using the “Digisim Simulation 3b” software from Bioanalytical Systems.25 Results and Discussion First and Second Reduction Processes of O2 in [BMIm][TFSI] without Added Acid. Cyclic voltammograms of the reduction of dioxygen recorded under the same experimental conditions in [BMIm][TFSI] and in DMF are shown in Figure 1. At low scan rates, the electrochemical reduction of O2 in DMF displays two electrochemical processes corresponding first 11 to the reversible formation of O-• 2 (E1/2 ) -1.37 V) and second to its further and irreversible reduction to O2(E ) -2.5/-2.7 2 p V) that often appears as an ill-defined signal.26 In [BMIm][TFSI], the same general patterns are observed and the two processes are visible at E1/2 ) -1.02 and Ep ) -1.75 V, respectively. The anodic peak corresponding to the reoxidation of O2-• diminishes when the final potential is set sufficiently negative for reaching the irreversible formation of O22- by contrast with the case where the scan is reversed just after the first reduction process of the dioxygen. In addition, a new anodic peak appears during this reverse scan at about -0.8 V. This behavior confirms the attribution of the first and second reduction processes and shows that the dianion reacts rapidly to form HO2- and OH-, certainly with proton traces, in agreement with previous report.13b However, if the same general patterns are observed in DMF and in [BMIm][TFSI], several differences are noticeable. A first observation concerns the potential difference between the first and second reductions. The value is much smaller in the ionic liquid (around 0.5 V) than in DMF (more than 1.0 V). It is also noticeable that the second process displays a sharper peak in the ionic liquid than in the
2828 J. Phys. Chem. B, Vol. 113, No. 9, 2009 DMF. Similar effects have been already reported before in a closely related situation in DMSO.15 When an imidazolium salt (1-ethyl-3-methylimidazolium tetrafluoroborate, [EMIm][BF4]) is added to the oxygen solution in DMSO, the position and the shape of the second reduction peak are similarly modified. The phenomenon was explained by the occurrence of ion pairings between the charged species of oxygen and the [EMIm] cation and their equilibrium association constants were determined.15 The second observation concerns the first reduction process. In O2-saturated [BMIm][TFSI] solution, the reduction potential is less negative than in DMF, which is also indicative of a stabilization of O-• 2 certainly through ion associations between the superoxide and [BMIm] cation. Additionally, the difference between the cathodic and anodic peak potentials is larger than that observed in the DMF solution (compare Figure 1, panels A and B) corresponding to a slower kinetics of the heterogeneous electron transfer in the ionic liquid than in DMF. A rapid estimation based on the peak-to-peak separation gives for the standard heterogeneous electron transfer rate constant for the O2/O2-• couple, ks, a value in the range of 1.5-2.0 × 10-4 cm · s-1. This value is around 10 times lower than the value of ks derived from Figure 1B (ks ≈ 2 × 10-3 cm · s-1 in DMF)27 and agrees with previous investigations of the O2/O2-• couple in 1-ethyl-3-methylimidazolium tetrafluoroborate, [EMIm][BF4]15 or 1-butyl-3-methylimidazolium hexafluorophosphate, [BMIm][PF6].12 The slowness of electron transfers in RTILs versus classical organic solvent is a well-known phenomenon, which was reported for many electrochemical systems.3,28 In the framework of the Marcus-Hush model, ∆G*o scales linearly the solvation energy.29 A decrease of ks could be interpreted by a considerable increase of the solvation that in the case of RTILs, corresponds to an increase of the interactions between the electrogenerated radical and the cations of the ionic liquid. Thus, all these observations fall in line with the occurrence of strong ionic interactions, between O2-• and the cation of the RTIL. Effect of Phenols Addition on the Cyclic Voltammetry. For the four different phenols (2,4-dimethylphenol, phenol, p-cresol, and 4-chlorophenol), cyclic voltammetries of the O2 reduction were recorded both in the ionic liquid and in the DMF under the same experimental conditions (same phenol concentrations, same scan rates, etc.) (Figure 2). In both media, the O2/O2-• system loses its reversibility in the presence of phenol and the forward peak currents increases upon phenol addition. For large concentrations of phenol, the O2 reduction becomes totally irreversible and the intensity of the cathodic peak becomes almost twice than that in the absence of phenol, indicating that the electron stoichiometry passes from one to two electrons per O2 molecule. Simultaneously, as soon as phenol is added in the solution, a new irreversible anodic process appears in the -0.3 to -0.2 V range for which the precise oxidation potential depends on the nature of the added phenol. By comparison with an authentic sample, this oxidation peak could be ascribed to the oxidation of the phenolate ion obtained after deprotonation of the added phenol. This result is similar to the one reported by Hardacre et al. in a comparable ionic liquid for the unsubstituted phenolate.13b To complete the literature data, we also checked with the other substituted phenols, that a small addition of base (Me4NOH) to the solution of the substituted phenol in [BMIm][TFSI] results in the appearance of very similar peaks (same potential, same current intensity, and similar shape). The variation of the phenolate oxidation potentials in DMF and [BMIm][TFSI] with the substituent were compared through the
Rene´ et al. plot of their peak potentials. Values were measured on the reverse scan of the cyclic voltammograms of the O2 reduction and for the lowest concentrations of added phenol. As seen in Figure 3, a linear variation is obtained (slope 0.93) confirming the nature of these oxidation processes and showing that the introduction of a substituent on the aromatic ring provokes similar effects in both media, slightly smaller in the ionic liquid. By analogy with the known behavior of O2 in organic solvents,18 all these results show that its electrochemical reduction corresponds to the global following reaction:
O2+2e- + 2PhOH f H2O2+2PhO-
(0)
Mechanism of O2 Reduction in [BMIm][TFSI] in the Presence of a Phenol. Estimation of the reactivity of an electrogenerated intermediate, like O-• 2 , could be achieved from the measurements of the reversibility of the electrochemical process (estimated from the ratio between the peak currents measured for the forward and reverse scan)30 for a given phenol concentration.31 Qualitatively, for comparable experimental conditions, a larger decrease of the reversibility upon addition of phenol corresponds to a higher reactivity of O2-• toward the considered phenol. By using this simple approach, the reactivities of the four phenols toward O2-• could be directly compared through the amplitude of the return peak for given concentration of phenol and scan rate. As expected, the derived reactivity of these phenols is a function of their substitution.17,18 Using the curves recorded for the highest concentrations of phenols, to take into account the large concentration of O2 in the RTIL, the reactivity was found to increase in [BMIm][TFSI] as 2,4dimethylphenol < 4-chlorophenol < p-cresol < phenol. Conversely, in DMF, the relative reactivities are different: 2,4dimethylphenol < p-cresol < phenol < 4-chlorophenol. A more quantitative analysis requires the extraction of the kinetics for the rate-determining step and, thus, additional knowledge about the nature of the mechanism. In DMF, the mechanism of O2 reduction in the presence of phenol has been investigated in detail.18 It corresponds to a succession of electrochemical and chemical steps (ECE-DISP mechanism): ks
O2 + e h O2-•
(1)
k2
O2-• + PhOH h HO2• + PhOk-2
k3
(2)
O2-• + HO2• y\z O2 + HO2-
(3)
HO2- + PhOH f H2O2+PhO-
(4)
After the formation of O2-•, the species is protonated by the phenol to produce the radical HO2• which is much easier to reduce than the starting superoxide.11,18 It results that HO2• is immediately reduced in solution in the vicinity of the electrode by another O2-• (reaction 3), leading to an overall 2 electronprocess as presented in eq 0. This accounts for an electron transfer disproportionation since O2-• and HO2• have formally the same oxidation state. The analogy between the voltammo-
Superoxide Protonation by Weak Acids
J. Phys. Chem. B, Vol. 113, No. 9, 2009 2829
Figure 2. Cyclic voltammograms of O2 reduction in [BMIm][TFSI] (A-H) and in DMF (+0.1 mol · L-1 NBu4PF6) in the presence of increasing concentrations of 4-chlorophenol (A, E), phenol (B, F), p-cresol (C, G), 2,4-dimethylphenol (D, H). Concentrations of phenols: black line ) 0 mol · L-1, pink ) 10-3 mol · L-1, cyan ) 2 × 10-3 mol · L-1, red ) 4 × 10-3 mol · L-1, green ) 8 × 10-3 mol · L-1, blue ) 16 × 10-3 mol · L-1. Scan rate V ) 0.2 V · s-1. Conditions: 1 mm diameter disk Au electrode; O2-saturated solutions.
grams recorded in DMF and in [BMIm][TFSI] strongly supports that the reduction of O2 to H2O2 follows the same DISP mechanism. In the context of the DISP mechanism, two limiting behaviors could take place depending on the nature of the rate-determining step. If the backward reaction 2 (proton exchange) is slower than reaction 3 (electron exchange), the forward reaction (2) is
the rate-determining step and its rate constant controls the reversibility (DISP1 case). The kinetic parameter is equal to: k2CPhOHRT/FV (where k2 is the forward rate constant of reaction 2, CPhOH is the initial concentration of phenol, and V is the scan rate), and thus a value of the protonation rate could be derived from the measurement of the reversibility. Conversely, if backward reaction 2 is faster than reaction 3, reaction 2 acts as
2830 J. Phys. Chem. B, Vol. 113, No. 9, 2009
Rene´ et al. TABLE 1: Kinetic Parameter for the Reduction of O2 in [BMIm][TFSI] and DMF in the Presence of Phenols log(k2k3/k-2) acid
in RTIL
in DMF
pKa in DMF
4-chlorophenol phenol p-cresol 2,4-dimethylphenol
2.97 3.49 3.0 2.57
3.2 3.7 (3.74)a 3.4 (3.48)a 3.2
19.0 19.8 20.1 20.4
a
Values in brackets are from ref 18.
Figure 3. Oxidation of the different phenolates. Peak potentials in [BMIm][TFSI] as a function of peak potentials in DMF (+0.1 mol · L-1 NBu4BF4 at a scan rate 0.2 V · s-1 on a gold electrode: 4-chlorophenolate (1), phenolate (2), p-cresolate (3), 2,4-dimethylphenolate (4).
a pre-equilibrium versus reaction 3 (DISP2 case). In such a case, the kinetic parameter is equal to k2k3/k-2CPhOHRT/FV (where k2 and k-2 are the forward and backward rate constants of reaction 2 and k3 the rate constant of the disproportionation step). The discrimination between the two DISP subcases is not an easy task since both limiting kinetics display the same reaction orders versus added acid or O2 concentrations in unbuffered media.32 Experimentally, they only differ from how the process reversibility varies with its own kinetics parameters.18,32 In organic solvents like DMF or DMSO, the DISP2 case was found to prevail for the reduction of O2 in the presence of weak acids as those considered in this work.18 It means that reaction 2 (protonation of O2-• by the phenol) is an energetically unfavorable step followed by a highly exergonic irreversible electron transfer, reaction 3, that displaces the protonation equilibrium. In our experimental situation, a clear-cut answer between the DISP1 and DISP2 cases is difficult to obtain because of different experimental difficulties. Indeed, the available potential window is limited on the negative scan (see Figure 1), because we must prevent the interference of the second reduction of O2-• on the reversibility of the first process. Conversely, because the kinetics of electron transfer is rather slow in [BMIm][TFSI], the first reduction displays considerable peak-to-peak potential differences and requires a large potential scan for obtaining an almost complete conversion of O2 to O2-•. Moreover, superoxide may also slowly react with impurities or the RTIL cation itself.15,16 Thus, to choose between the occurrence of a DISP1 in opposition to a DISP2 mechanism in [BMIm][TFSI], we should discuss the evolution of the competition between the backward reaction 2 and the irreversible electron transfer (3) when passing in the ionic liquid.33 Concerning the electron transfers, several reports have emphasized that these reactions are significantly slower in ionic liquids than in organic solvents because of ion interactions between the electrogenerated species and the cation of the ionic liquid.3c Concerning the proton exchange reactions involving phenolates, there is no available kinetics data about their evolution between an organic solvent and the ionic liquid. However, interactions between [BMIm] cations and phenolates might also exist. They would result in a better stabilization of the charged phenolate compound than the phenol, which is a neutral species, and thus, a lower acidity of the phenol/phenolate couple is expected. This stabilization is confirmed by the higher oxidation potentials observed in [BMIm][TFSI] than in DMF (see Figure 3). However by analogy with the behavior described for the electron transfer, we could propose that the effects of ion interactions will be larger for a small species like O2-• than
Figure 4. Variation of the kinetics parameter k2k3/k-2 of the reduction of O2 for the different phenolates in DMF (9) and in [BMIm][TFSI] (b): 4-chlorophenolate (1), phenolate (2), p-cresolate (3), 2,4-dimethylphenolate (4).
for a phenolate, in which the negative charge is delocalized over the aromatic ring. It is thus very likely that the same kinetics situation (DISP2) prevails in [BMIm][TFSI] as previously found in organic solvents like DMF or DMSO.18 Comparative Reactivities of the Phenolate for the Protonation of O2-•. Assuming the occurrence of a DISP2 mechanism, the comparison of the experimental data with simulated curves (using the Digisim simulation software)25 allows us to derive the kinetic parameters both in DMF and in [BMIm][TFSI] (see Table 1).34 The obtained values are in good agreement with previously published data.18 In this previous work, the measurements were performed for the two phenols in DMF by double potential step chronoamperometry, hence validating our procedure.18 Table 1 displays the variations of the kinetic parameter k2k3/ k-2 for the different phenols. In the framework of the DISP2 mechanism, the parameter log(k2k3/k-2) is directly related to the pKa difference of the superoxide and the phenol through the relation log(k2/k-2) ) pKa(HO2-•/O2-•) - pKa(PhOH/PhO-) taking into account that the kinetics for the electron exchange, k3, is expected to remain the same when changing the nature of the phenol. It is thus interesting to compare the values of log(k2k3/k-2) with the pKa of the phenol. Even if data are not available in [BMIm][TFSI], to illustrate the effects of increasing the phenol acidity on the protonation rate of O2-•, log(k2k3/k-2) was plotted versus the pKa of the phenols in DMF (see Figure 4). In DMF, a linear fit with a slope close to unity (slope ) 0.9 without the 4-Cl-phenol value) is obtained as expected for a DISP2 mechanism.18,32 In [BMIm][TFSI], the three less acidic phenols, log(k2k3/k-2) follows a similar linear variation. The log(k2k3/ k-2) values are just below the DMF ones, which could be ascribed to a decrease of the electron transfer kinetics constant, k3. This similarity supports the occurrence of the DISP2 mechanism in DMF both in [BMIm][TFSI].
Superoxide Protonation by Weak Acids Similarly to conventional organic solvents, O-• 2 is a weak base and its protonation by phenol is a strongly uphill reaction. Assuming that reaction 3 is close to the diffusion limit in the RTIL (around 107-108 mol · L-1 · s-1),3c,4 we could estimate that the acidity of the HO2•/O2-• couple is 4-5 units of pKa lower than the phenol/phenolate couple. The case of 4-chlorophenol is not clear. The kinetics is clearly below the expected value from the linear fit. Its reactivity versus O2-• is in the range of the p-cresol suggesting the existence of more complex interactions of the RTIL cation with this acid. Conclusion Based on its diffusion coefficients and recent electrochemical investigations, superoxide appears as strongly associated to the cations of the ionic liquid (ion pairing). However, results obtained in this work show that these important associations do not significantly modify the reactivity of the superoxide ion, at least, for what it concerns its redox reactivity and ability to abstract protons. O2-• behaves as in a conventional organic solvent containing a supporting electrolyte. As established before in DMF or DMSO, the superoxide ion appears as a weak base in [BMIm][TFSI]. However, it can be protonated by a weak acid like phenol thanks to the occurrence of a following electron transfer that displaces the unfavorable proton exchange equilibrium. Conclusions obtained here could certainly be extended to the reduction of oxygen in many common room temperature (aprotic) ionic liquids. Acknowledgment. Authors gratefully acknowledge support of this research by “Re´gion de Bretagne” through a Ph.D. grant (A.R.). References and Notes (1) (a) Welton, T. Chem. ReV. 1999, 99, 2071. (b) Ionic Liquids in Synthesis; Wasserscheid, P., Welton, T., Eds. Wiley-VCH: Weinheim, 2003. (c) Rogers, R., Seddon, K. R., Eds. Ionic Liquids: Industrial Applications for Green Chemistry. ACS Symp. Ser. 2002, 818. (2) (a) Ionic Liquids in Synthesis; Wasserscheid, P., Welton, T., Eds.; Wiley-VCH: Weinheim, Germany, 2003. (b) Green Industrial Applications of Ionic Liquids; Roger, R. D., Seddon, K. R., Volkov, S., Eds.; NATO Sciences Series; Kluwer: Dordrecht, The Netherlands, 2002; Vol 92. (c) Chiappe, C.; Pieracinni, D. J. Phys. Org. Chem. 2005, 18, 275. (3) (a) Buzzeo, M. C.; Evans, R. G.; Compton, R. G. Chem Phys Chem. 2004, 5, 1106. (b) Silvester, D. S.; Compton, R. G. Z. Phys. Chem. 2006, 220, 1247. (c) Hapiot, P.; Lagrost, C. Chem. ReV. 2008, 108, 2238. (4) (a) Grodkowski, J.; Neta, P. J. Phys. Chem. A 2002, 106, 11130. (b) Skrzypczak, A.; Neta, P. J. Phys. Chem. A 2003, 107, 7800. (c) Wishart, J. F.; Neta, P. J. Phys. Chem. B 2003, 107, 7261. (5) Brooks, C. A.; Doherty, A. P. J. Phys. Chem. B 2005, 109, 6276. (6) (a) Lagrost, C.; Gmouh, S.; Vaultier, M.; Hapiot, P. J. Phys. Chem A 2004, 108, 6175. (b) Lagrost, C.; Hapiot, P.; Vaultier, M. Green Chem. 2005, 7, 468. (7) In contrast, this phenomenon is generally negligible in organic electrolyte because of the weak variation of the diffusion coefficient with the mass of the molecule: the Stokes-Einstein law predicts that D varies as a function of the inverse of the radius of the equivalent sphere and thus as a function of the third root of molecular mass D ≈ M-1/3.3 (8) Buzzeo, M. C.; Klymenko, O. V.; Wadhawan, J. D.; Hardacre, C.; Seddon, K. R.; Compton, R. G. J. Phys. Chem. A 2003, 107, 8872. (9) Ghilane, J.; Lagrost, C.; Hapiot, P. Anal. Chem. 2007, 79, 7383. (10) (a) The phenomenon is not limited to the oxygen reduction, several redox couples display similar effects but to a lower extent. Recently, the variations of diffusion coefficients were investigated for the reduction of a series of nitro-derivative. It was found that the variation is directly related to the localization and accessibility of the charge on the radical ion.10b. (b) Zigah, D.; Ghilane, J.; Lagrost, C.; Hapiot, P. J. Phys. Chem. B, 2008, 112, 14952.
J. Phys. Chem. B, Vol. 113, No. 9, 2009 2831 (11) (a) Sawyer, D. T.; Valentine, J. S. Acc. Chem. Res. 1981, 14, 393. (b) Sawyer, D. T.; Roberts, J. L. J. Electroanal. Chem. 1966, 12, 90. (c) Sawyer, D. T.; Chiericato, G.; Angelis, C. T.; Nanni, E. J.; Tsuchiya, T. Anal. Chem. 1982, 54, 1720. (12) (a) AlNashef, I. M.; Leonard, M. L.; Kittle, M. C.; Matthews, M. A.; Weidner, J. W. Electrochem. Solid-State Lett. 2001, 4, D16. (b) AlNashef, I. M.; Leonard, M. L.; Matthews, M. A.; Weidner, J. W. Ind. Eng. Chem. Res. 2002, 41, 4475. (13) (a) Evans, R. G.; Klymenko, O. V.; Saddoughi, S. A.; Hardacre, C.; Compton, R. G. J. Phys. Chem. B 2004, 108, 7878. (b) Villagran, C.; Aldous, L.; Lagunas, M. C.; Compton, R. G.; Hardacre, C. J. Electroanal. Chem. 2006, 588, 27. (14) (a) Kumelan, J.; Kamps, A. P.-S.; Urukova, I.; Tuma, D.; Maurer, G. J. Chem. Thermodyn. 2005, 37, 595. (b) Jacquemin, J.; Costa Gomes, M. F.; Husson, P.; Majer, V. J. Chem. Thermodyn. 2006, 38, 490. (15) Islam, Md. M.; Ohsaka, T. J. Phys. Chem. C 2008, 112, 1269. (16) Barnes, A. S.; Rogers, E. I.; Streeter, I.; Aldous, L.; Hardacre, C.; Wildgoose, G. G.; Compton, R. G. J. Phys. Chem. C 2008, 112, 13709. (17) Hapiot, P.; Pinson, J.; Yousfi, N. New J. Chem. 1992, 16, 877. (18) Andrieux, C. P.; Hapiot, P.; Save´ant, J.-M. J. Am. Chem. Soc. 1987, 109, 3768. (19) (a) Sun, J.; Forsyth, M.; MacFarlane, D. R. J. Phys. Chem. B 1998, 102, 8858. (b) Bonhoˆte, P.; Dias, A. P.; Papageorgiou, N.; Kalyanasundaram, K.; Gra¨tzel, M. Inorg. Chem. 1996, 35, 1168. (20) (a) Singh, P. S.; Evans, D. H. J. Phys. Chem. B 2006, 110, 637. (b) Costentin, C.; Evans, D. H.; Robert, M.; Save´ant, J.-M.; Singh, P. S. J. Am. Chem. Soc. 2005, 127, 12490. (21) Garreau, D.; Save´ant, J.-M. J. Electroanal. Chem. 1972, 35, 309. (22) Ghilane, J.; Hapiot, P.; Bard, A. J. Anal. Chem. 2006, 78, 6868. (23) Dapremont-Avignon, C.; Calas, P.; Commeyras, A.; Amatore, C. J. Fluorine Chem. 1991, 51, 357. (24) Tsushima, M.; Tokuda, K.; Ohsaka, T. Anal. Chem. 1994, 66, 4551. (25) Digisim 3b. BASi DigiSim Simulation Software for Cyclic Voltammetry, Bioanalytical Systems. http://www.bioanalytical.com/products/ ec/digisim. (February, 2004). (26) Costentin, C.; Evans, D. H.; Robert, M.; Save´ant, J.-M.; Singh, P. R. J. Am. Chem. Soc. 2005, 127, 12490. (27) Taking for DO2 and DO2-• in DMF: 4.7 × 10-5 and 1.5 × 10-5 cm2 · s1, respectively, that were estimated from ref 22 and a ratio of 3 for DO2/DO2-• (from ref 11b) as found in DMSO. (28) (a) A decrease by a ratio of 40-50 is in the range of what was reported for the reduction of unprotected nitro-aromatic compounds when passing from an organic solvent to a RTIL.28b. (b) Lagrost, C.; Preda, L.; Volanschi, E.; Hapiot, P. J. Electroanal. Chem. 2005, 585, 1. (29) (a) The value of ks and the activation Gibbs energies are related * from the following expression ks ) A exp[-(∆G*o + ∆Gin )/(RT)] where A * is a pre-exponential factor and ∆G*oand ∆Gin are the outer and inner 29b. reorganization energies. (b) Marcus, R. A. J. Electroanal. Chem. 2000, 483, 2. (30) Save´ant J.-M. Elements of Molecular and Biomolecular Electrochemistry. An Electrochemical Approach to Electron Transfer Chemistry; Wiley-Intersciences: Hoboken, NJ, 2006. (31) Le Bourvellec, C.; Hauchard, D.; Darchen, A.; Burgot, J. L.; Abasq, M. L. Talanta 2008, 75, 1098. (32) Amatore, C.; Gareil, M.; Save´ant, J.-M. J. Electroanal. Chem. 1983, 147, 1. (33) In principle, precise comparisons of the experimental data with the relevant theoretical curves calculated for the DISP1 or DISP2 mechanisms (variation of the reversibility with the scan rate in cyclic voltammetry or the phenol concentration) may be used to discriminate between the two limiting subcases.32 This strategy was used with success for characterizing the reduction mechanism of O2 in organic media (DMF or DMSO).18 However, the procedure requires very precise data and negligible interferences of other decay processes of the electrogenerated superoxide for allowing the extraction of the experimental data on a large range of the kinetic parameter (2-3 orders of variations) which was not possible in the present situation. (34) The method consists in comparing the experimental reversibility estimated as the ratio between the anodic and cathodic peak currents with simulated curves. (For a more detailed presentation, see page 85 in ref 30). This approach is very similar to the DSP technique used in ref 18 but simulations of the voltammograms require the injection in the calculations of an estimation of the intrinsic electron transfer rate constant, ks, and of thediffusioncoefficientsDO2 andDO2-•previouslymeasuredin[BMIm][TFSI].9
JP810249P
