Electrochemical Properties of a Verdazyl Radical in Room

14 Room temperature ionic liquids (RTILs) have been widely investigated as alternative electrochemical solvents 15 for a range of dissolved species over the past two decades. However, the behaviour of neutral radicals dissolved 16 in RTILs is relatively unexplored. In this work, the electrochemistry of a stable verdazyl radical – 1,5-dimethyl17 3-phenyl-6-oxoverdazyl (MPV) – has been studied on a platinum thin-film electrode using cyclic voltammetry 18 and chronoamperometry in ten different RTILs. The organic solvent propylene carbonate is also employed as a 19 comparison. The nature of the solvent system was found to have a large effect on the electrochemical behavior, 20 particularly on the reduction reaction of the verdazyl radical. Chronoamperometry on a microdisk electrode was 21 used to calculate diffusion coefficients (D’s), and plots of D vs the inverse of viscosity were linear, suggesting 22 typical hydrodynamic diffusional characteristics of the radical, in line with the behaviour of dissolved neutral 23 and charged compounds (e.g. ferrocene and cobaltocenium) in RTILs. Overall, this study demonstrates that 24 different RTILs have a significant influence on the electrochemistry of MPV, and therefore careful selection of 25 the solvent system for electrochemical applications is advised. 26 Aust. J. Chem. xxx https://doi.org/xxx


Introduction
Room-temperature ionic liquids (RTILs) are a unique class of solvents that are made up entirely of cations 2 and anions. They are increasingly being recognised as ideal alternative electrolytes in a multitude of traditional 3 and emerging electrochemical applications, such as batteries, sensors, actuators, capacitors, fuel cells, and 4 photovoltaics. [1][2][3][4][5][6] They can effectively be described as an 'electrolyte solvent' as they serve as both the solvent 5 and electrolyte in electrochemical experiments. RTILs have many key advantages over conventional solvents, 6 including intrinsic ionic conductivity, high thermal and chemical stability, low volatility, good solvating 7 properties, and wide electrochemical windows. [7][8][9][10] To understand their viability for use in different 8 electrochemical applications, the fundamental electrochemical behaviour of dissolved species in RTILs must 9 first be understood. Many electrochemical studies have been performed in RTILs, and the voltammetry has been 10 compared and contrasted to conventional solvents in an extensive review paper. [11] 11 One type of dissolved species that has received less attention in RTILs are neutral radical compounds. Evans 12 et al. [12] studied the stable free radical 2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPO) by cyclic voltammetry in 13 five RTILs, observing differences in the voltammetry in two tetraalkylphosphonium ionic liquids. This was 14 attributed to the formation of a more ordered bilayer structure consisting of alternating ionic and lipophilic 15 regions. The mechanisms and reactions of some charged radical speciesincluding the N,N,N′,N′-tetramethyl- 16 para-phenylenediamine (TMPD) radical cation, [13] and the 1-bromo-4-nitrobenzene radical anion [14] have been 17 studied in RTILs. We note that many electrogenerated species in RTILs are radicals (e.g. superoxide radical 18 anion, [15] N,N-dimethyl-p-toluidine (DMT) radical cation [16] ) but their behaviour is often difficult to assess 19 because of their short-lived nature at the electrode. In this work, the electrochemistry of a stable neutral verdazyl 20 radical is studied in a range of RTILs to expand the knowledge of radical electrochemistry in these solvent 21 systems. 22 Neutral radicals are molecules that contain one (or more) unpaired electrons. These species do not obey 23 conventional bond valence theory and as a result have fewer bonds than predicted, leading to the molecules being 24 highly reactive and short lived. [17] However, there are a number of stable organic radicals that exist. [18] Stability 25 in these systems tends to arise from the resonance delocalisation of unpaired spin and in some cases the steric 26 protection of the high spin electron density. A number of applications are being developed based on stable 27 radicals due to their unique redox properties. [19] Examples include both the development of species for use as 28 single component conductors [20][21] and organic batteries. [22][23] It is the electron-transfer chemistry of these 29 molecules which determines the usefulness of radical species, such as redox potentials. [24][25] Electrochemistry is 30 an essential tool used to understand these processes. [26][27] 31 Verdazyls (Type I and II) are one such highly stable organic radical system (Fig. 1). [28][29][30] These heterocyclic 32 molecules have an unpaired spin density that is delocalised over the nitrogen atoms. These molecules are stable 33 in both air and water making them ideal candidates for understanding the electrochemical properties associated 34 with the unpaired electrons. Type II verdazyls have been shown to be difficult to oxidize due to the electron 35 withdrawing nature of the carbonyl functionality. [24,31] The twisting of the molecule around the N-substituents 36 can further add to this as it is associated with a decrease in the delocalisation of the unpaired electron. [32] Based 37 on our literature search, the electrochemistry of various verdazyl (Type I and II) molecules has been studied 38 3 previously in various conventional solvents and mixtures (e.g. THF, triethylamine, acetonitrile, etc.), [24-25] but 1 not in RTILs. In conventional solvents, the verdazyls were observed to undergo a one-electron oxidation from 2 the neutral 7π radical to the 6π cation, and a one-electron reduction to the 8π anion (Fig. 2). [24] The oxidation and 3 reduction potentials are significantly affected by the electron donating and withdrawing substituents present on 4 the radical. [24] The 1,5-dimethyl-3-phenyl-6-oxoverdazyl (MPV) molecule, with methyl as both the R1 and R2 5 substituents, will be the focus of this work (Fig. 3). This radical has specifically been explored as a mediator in 6 RAFT polymerization of styrene and n-butyl acrylate. [33][34] Fig. 3 Chemical structure of the Type II verdazyl, 1,5-dimethyl-3-phenyl-6-oxoverdazyl (MVP) used in this work. 13 Given the obvious advantages of RTILs as solvent systems, and the potential applications of the family of 14 verdazyl radicals, we present a first look into the electrochemical mechanism of MPV ( Fig. 3) in RTILs. We 15 discuss the relationship between diffusion coefficient and solvent viscosity in terms of hydrodynamic theory and 16 Stokes-Einstein behaviour. The findings in this study will be valuable in the understanding of the effect of 17 solvation on the electrochemical behavior of a neutral radical, and to guide the selection of solvents for different 18 electrochemical applications of these emerging and interesting materials. Additionally, understanding the 19 electrochemistry of verdazyl radicals in RTILs is of great significance in future applications where they may be 20 used concomitantly. An example is in batteries, where improved energy density can be achieved by increasing 21 the concentration of the redox species in the catholyte. The use of RTILs with a stable and soluble redox 22 compound is one solution for the development of new battery materials with higher energy storage 23 4 capabilities. [35] The stable verdazyl radical is a good candidate for such applications and warrants greater 1 investigation to understand the underlying electrochemical behaviour of this species in RTILs. All reagents for the synthesis of MPV (see Supporting Information) were purchased from Sigma-Aldrich or 6 Alfa Aesar and used without further purification, unless otherwise stated. between each CV scan to allow the system to fully recover. The electrochemical behaviour of MPV in ten different RTILs was first investigated using cyclic voltammetry 3 (CV) and compared with the behaviour in a conventional aprotic solvent, propylene carbonate (PC) with 0.1 M 4 TBAP (Fig. 4). PC was selected due to its very low volatility (boiling temperature = 240 °C) and its properties 5 as a 'versatile solvent for electrochemistry'. [41] For example, it has been used as a solvent for various applications 6 including ammonia gas sensing. [42] The voltammetry of this particular verdazyl radical (Fig. 3) has not yet been 7 reported in the literature, so its behaviour in a conventional solvent system such as PC is important to study. 8 The potentials are scanned positively over the oxidation peak, before reversing to scan over the reduction 9 peak. The CVs in Fig. 4 were taken at a relatively fast scan rate of 2000 mV s -1 , since the electrode was observed 10 to 'foul' during cathodic CV measurements, particularly at lower scan rates in some RTILs. The fouling is likely As can be seen in Fig. 4, the CVs typically show one oxidation peak and one reduction peak, consistent with 17 the general mechanism reported for verdazyl radicals in acetonitrile and dichloromethane (Fig. 2). The verdazyl radical reduction peak, however, exhibits vastly different voltammetry in the different solvents. 25 In the conventional solvent, PC, the reduction is chemically reversible (Fig. 4a), although the CV shape is much 26 broader compared to the oxidation peak, suggesting that the reduction step has more sluggish electrochemical 27 kinetics (see discussion below). In most of the RTILs, the oxidative back-peak following the reduction process 28 (oxidation of the 8 verdazyl anion back to the neutral radical) is absent, suggesting that the 8 verdazyl anion 29 in unstable and probably reacts with the RTIL cation. The imidazolium cation is known to undergo proton 30 abstraction from the C(2) position, in the presence of a strong base. [ (Fig. 4). 32 The reduction peak in some RTILs shows a split-wave feature, further hinting of a more complicated mechanism 33 at play. The only RTIL that displays a clear reverse peak is [C4mim][PF6], but the current for the reduction peak 34 is much larger than the oxidation peak, decreasing the likelihood that the reduction is a one-electron reversible 35 reduction as described in Fig. 2 times. [44][45][46] We note that water may also play a role in the mechanism, since water is known to be present at ppm 39 8 levels in RTILs even after rigorous drying procedures (e.g. under vacuum). [47] Our experiments were carried out 1 under a constant stream of dry nitrogen at a high flow rate of 1000 sccm to ensure that the water content was 2 kept as low as possible during the measurements, and that this effect is minimal.
3 Table 1 summarises the data extracted from the CVs in Fig. 4, including the peak potentials (Ep vs. Fc/Fc + ), 4 peak currents (Ip) and the peak-to-peak separations (∆Ep) of the oxidation and reduction peaks. For easy 5 visualization of the differences between the solvents, histogram plots of these values are also provided in the 6 supporting information (Fig. S2 and Fig. S3).   18 the CVs due to the absence of a clear peak. The 8π verdazyl anion therefore appears to be much less stable in 19 RTIL environments compared to the 6π verdazyl cation. 20 21 Table 1 Peak potentials, Ep, peak currents, Ip, and peak-to-peak separations, ∆Ep, for the CV of 10 mM MPV at 2000 mV s -1 1 . Data extracted from CVs in Fig. 4 and Fig. S1. Numbers in brackets are estimated because of the absence of a clear peak 2 The peak-to-peak separations give an indication of the kinetics of the electrochemical step. For the oxidation 4 reaction, a ∆Ep of 81 mV in PC implies a moderately fast electrode process, but not as fast as the ∆Ep of 59 mV 5 expected for an ideal one-electron process. [48] In RTILs, the peak-to-peak separations are similar but slightly 6 larger than PC, ranging from 83-110 mV, suggesting more moderate kinetics in some of the RTILs. The RTIL 7 [P14,6,6,6][NTf2] displayed a much larger ∆Ep of 248 mV for the oxidation peak, indicating the slowest kinetics 8 out of all the solvents studied. In contrast to the oxidation peaks, the reduction processes are mostly irreversible 9 such that a ∆Ep cannot be measured. Only three RTILs show a reverse peak, and these display much larger peak- To understand more about the diffusional behaviour of the verdazyl radical in the different solvents, CV was 15 carried out at a range of scan rates from 100 to 2000 mVs -1 (Fig. 5). The number of CV scans were minimised 16 due to obvious 'fouling' of the electrode that was observed, particularly at scan rates < 100 mVs -1 , likely because 17 of adsorption of electrogenerated products on the electrode. was left for ~15 mins between scans, presumably due to the slow diffusion of the electrogenerated products away 21 from the surface. Plots of peak current vs scan rate were linear for both oxidation and reduction peaks in all 22 RTILs (see Fig. S4  The peak-to-peak potentials (∆Ep) for the oxidation peak do not significantly change with scan rate (data not   26 shown), consistent with a relatively fast electrode process. However, in the solvents where reverse peaks are 27 observed, the redox couple for the reduction peak becomes more separated (i.e. increased ∆Ep) with increasing 28 scan rate, consistent with slower kinetics of the reduction process.  to the average potentials of ferrocene/ferrocenium (Fc/Fc + ) peaks. 6 Due to the good reversibility of the reduction peak in PC, further analysis was undertaken. Fig. 6 shows a plot 11 1 Fig. 6 Plots of the background corrected reverse divided by forward peaks currents of the first reduction peak of 10 mM 2 MPV at different scan rates (Fig. 5) on Pt-TFE in 0.1 M TBAP in PC.
3 Chronoamperometric analysis 4 Potential step chronoamperometry was carried out to determine the diffusion coefficient, D, of the verdazyl 5 radical in the different solvents. The oxidation peak was chosen for analysis due to the more ideal CV shape 6 compared to the reduction peak, and the oxidative chronoamperometric transient was iteratively fitted to the 7 Shoup and Szabo expression. [40]   where kB is the Boltzmann constant, T is temperature, α the hydrodynamic radius of the diffusing particle, j = 4 15 is for the 'slip' limit, and j = 6 is for the 'stick' limit modes of diffusion. In RTILs, a linear relationship between 16 D and the inverse of viscosity is generally observed for neutral compounds such as ferrocene [46] , 2,4,6-17 trinitrotoluene [45] , 2,4-dinitrotoluene [54] , a rhenium tetrazolato complex (fac-[Re(CO)3(phen)L]) [55] , the charged 18 cobaltocenium cation [46] and the stable free radical, 2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPO) [12] . In most 19 cases, the 'stick' mode of diffusion is followed in RTILs, [45][46]55] with some exceptions where j is closer to 4, 1 and this behaviour is widely considered to be a "topic of interest". [56][57][58] For sufficiently small species (e.g. 2 oxygen [44] , hydrogen [59] , and sulfur dioxide [60] ) deviations become more pronounced in RTILs as the molecule is 3 able to move through the small dynamic channels and pores, leading to D values that are only partially dependent 4 on the self-diffusion of the solvent itself. [61] 5 Fig. 7 shows a plot of D against the inverse of viscosity, ɳ -1 , in the ten RTILs. The diffusion coefficients were 6 1-2 order of magnitudes smaller in the RTILs than in PC, which is not unexpected, considering the much higher 7 viscosities of the RTILs (Table 2). A reasonably linear fit (R 2 = 0.907) was obtained, suggesting that the diffusion 8 of 1,5-dimethyl-3-phenyl-6-oxoverdazyl in RTILs agrees reasonably well with the Stokes-Einstein relationship. 9 From the slope of the fit, the hydrodynamic radius, α, was calculated to be 2.7 Å for the slip limit (j = 4), and

20
The electrochemical properties of a Type II verdazyl radical -1,5-dimethyl-3-phenyl-6-oxoverdazylhas 21 been studied in range of different RTILs and PC. It was found that the solvent environment significantly affected 22 the shapes of the reduction peak, andto a lesser extentthe oxidation peak. The oxidation peak was chemically 23 reversible in most solvents, but the reduction peak was generally chemically irreversible in the RTILs, also 24 displaying more sluggish kinetics of the electrochemical step. The oxidation peak was analysed by results show that different solvent environments significantly influence on the electrochemistry of 1,5-dimethyl-1 3-phenyl-6-oxoverdazyl, and therefore careful selection of the RTIL is recommended for electrochemical 2 applications using these verdazyl radicals.
3 Supporting Information 4 Chemical procedures for the synthesis of 1,5-dimethyl-3-phenyl-6-oxoverdazyl, along with full 5 characterisation of the intermediates. 1 H/ 13 C NMR and IR spectra of all the intermediates and final product. ESR 6 spectrum of the final radical. Chemical structures of RTIL cations and anions employed in this work. Photo of 7 the verdazyl radical synthesized powder and dissolved in two RTILs. Cyclic voltammetry of the verdazyl radical, 8 showing the wider scan range in PC and the 10 RTILs. Histograms of the peak current and peak potential in the 9 different solvents in order of increasing viscosity. Plots of peak current vs. scan rate in all solvents, plot of charge 10 ratios for the reduction peak as a function of scan rate in PC, and plot of ratio of oxidation to reduction peak 11 current as a function of scan rate in all solvents. 12

13
The authors declare no conflicts of interest.