1
Voltammetry of immobilised microdroplets containing p-chloranil on basal plane pyrolytic graphite electrodes

2
The electron transfer properties of p-chloranil (2,3,5,6-tetrachloro-1,4-benzoquinone, TCBQ) were investigated in both homogeneous and heterogeneous media.

3
For the homogeneous study, the electrochemical reduction of TCBQ was carried out in different aprotic solvents (namely: benzonitrile (BN), N,N,dimethylformamide (DMF), propylcyanide (PrCN) and dimethylsulfoxide (DMSO)) and revealed two successive one-electron reductions according to a quasi-reversible EE mechanism.

4
For the heterogeneous study, cyclic voltammetry with basal plane pyrolytic graphite electrodes modified with microdroplets of benzonitrile/TCBQ was employed.

5
The droplets were found to be randomly dispersed with a degree of overlapping and average diameters of 5 μm giving the microdroplets individual volumes of ca.

6
33 fL.

7
The redox processes within the electrically insulating microdroplets were shown to be very sensitive to the nature and concentration of ions in the surrounding aqueous phase as, in order to retain electroneutrality within the unsupported oil phase, electric field-induced migration of ions likely occurs during the Faradaic current flow.

8
Depending on the lipophilicity of the aqueous electrolyte cation uptake into or electrochemical generated anion expulsion from the organic phase containing the electroactive specie TCBQ was induced electrochemically.

9
Alkali metal cation uptake into the microdroplet environment was not observed.

10
However less hydrophilic tetraalkylammonium cations NR4+ (R+ = Bu and Pe) inserted.

11
Proton insertion into the oil phase was also shown to occur as the current|voltage shifted to more positive potentials, making the reductive process more facile, as the pH of the buffer solution was decreased.

12
The higher efficiency of proton insertion as compared with Group I cations insertion was explained in terms of the formation of strong O–H covalent bonds which outweighs the ion phase transfer thermodynamics.

13
Finally, the cross-phase electron transfer across the benzonitrile|water interface was examined when the TCBQ microdroplets were purposely made conductive by addition of a hydrophobic, nonpartitioning electrolyte in the oil phase.

14
Again, the resulted voltammetry was found to change depending on the identity and concentration of the salt dissolved in the surrounding aqueous environment.

Introduction

15
The investigation of electrochemically driven ion-transfer processes across liquid|liquid interfaces has been shown to be of fundamental interest1–21 as well as of interest for specific ion sensing in aqueous media22,23 and for the replication of redox chemistry occurring in both chemical (i.e., aqueous emulsion system, etc.) and natural systems.24–27

16
Several studies which deal with electrodes modified with microdroplets of water-immiscible compounds have been reported and we refer the reader to a review article recently published in this journal28 for a more thorough description of the techniques used and the results obtained by different authors.

17
Based on the seminal work of Marken et al2–4. our group has, in the past six years, investigated the voltammetric processes of random arrays of unsupported electroactive microdroplets which are deposited on the surface of a basal plane pyrolytic graphite electrode (bppg) via solvent evaporation of a solution of the electroactive oil.12–16,25,28

18
This approach has provided a simple methodology to interrogate the electrochemistry at the liquid|liquid interface of several compounds that cannot be studied homogeneously in aqueous media due to their low solubility.

19
Initial work was mainly focused on organic hydrophobic oils containing a phenylenediamine backbone,12–16,29 although recently the biphasic redox chemistry of n-butylferrocene,16 4-nitrophenylnonylether,25 vitamin K126 and vitamin E27 in the form of microdroplets has also been explored.

20
The voltammetry associated with microdroplets has been shown to be very sensitive to the nature and concentration of ions in the surrounding aqueous phase as, in order to preserve charge neutrality within the unsupported microdroplets oil environment, ion insertion into or expulsion from the oil phase must take place during the Faradaic current flow.

21
When the transfer of the aqueous-based counterion from aqueous solutions into the oil droplet is thermodynamically more favourable than the transfer of the electrogenerated ion into the aqueous phase, electrochemically induced ion insertion leading to the formation of a novel ionic liquid was shown to occur at the triple phase boundary of electrode|oil|aqueous electrolyte.3,4,30

22
However, when the transfer of the aqueous based counterion into the organic deposit is energetically unfavourable, expulsion of the electrogenerated charge species into the adjacent aqueous phase was observed.

23
Dissolution processes of this sort have shown to take place at the liquid|liquid interface.30

24
Finally, reactive chemistry between the electrogenerated and the inserted ions was also shown to occur on some occasions.30

25
The research presented in this paper falls into two section.

26
In the first part, the well-documented electrochemical reduction31–35 of a hydrophobic benzoquinone derivative, p-chloranil (2,3,5,6-tetrachloro-1,4-benzoquinone, TCBQ), will be examined in homogeneous solutions of various aprotic solvents (namely: benzonitrile (BN), N,N,dimethylformamide (DMF), propylcyanide (PrCN) and dimethylsulfoxide (DMSO)).

27
In the second part, the redox chemistry of TCBQ will be further explored in the form of TCBQ/BN microdroplets immobilized at a bppg electrode surface immersed in various electrolyte solutions.

28
The literature has many examples of anion insertion,2,12,14–16,22,23,28,30,36 however, much fewer examples of cation insertion have, so far, been reported.11,25

29
It will be shown that, depending on the lipophilicity of the cation present in the aqueous phase, cation uptake into, or anion expulsion from, the organic deposit can take place to restore the electroneutrality within the microdroplet environment.

30
Protonation across the liquid|liquid interface and the effect of the addition of a supporting electrolyte in the oil phase will also be examined.

Experimental

Chemical reagents

31
Aqueous electrolyte solutions have been prepared with KCl (Riedel–de Haen,), LiCl, RbCl, tetramethylammonium chloride, (TMeACl) tetrapropylammonium chloride, (TPrACl) tetrabutylammonium bromide (TBuABr), tetrabutylammonium nitrate (TBuANO3), (all Aldrich), and tetrabutylammonium chloride, (TBuACl) (Fluka).

32
All these reagents were of the highest grade available and used without further purification.

33
The solutions and subsequent dilutions were carried out using deionised water from an Elgastat (Elga, UK) UHQ grade water system with a resistivity of not less than 18 MΩ cm.

34
The electrochemical reduction of p-chloranil (2,3,5,6-tetrachloro-1,4-benzoquinone, TCBQ, Aldrich, 99%) was investigated in the aprotic solvents benzonitrile (BN), N,N,dimethylformamide (DMF), propylcyanide (PrCN) (all Aldrich) and dimethylsulfoxide (DMSO, BDH).

35
Tetrabutylammonium perchlorate (TBuAP, Fluka) was used as supporting electrolyte at 0.1 M concentration.

36
For the deposition procedure followed for the microdroplets study (see 2.3, Procedure), acetonitrile (MeCN) (Fisher Scientific, dried and distilled) was used without further purification.

37
Buffered solutions were prepared using 0.05 M Na2HPO4/NaH2PO4, 0.1M NaH2PO4 and 0.1M boric acid solutions, and adjusted to the required pH using NaOH (all Aldrich).

38
Potassium hexacyanoferrate(ii) trihydrate (99%) was purchased from Lancaster.

39
Solutions were deoxygenated by outgassing with oxygen-free nitrogen (BOC Gases, Guildford, Surrey, UK) for at least 15 min prior experimentation and a nitrogen atmosphere was maintained over the solutions during each experimental run.

40
All experiments were carried out at room temperature of 20 ± 2°C.

Electrochemical assembly

41
All electrochemical measurements were recorded using a PG STAT 20 computer controlled potentiostat (Eco-Chemie, Netherlands) with a standard three electrode configuration.

42
For the homogeneous electrochemistry section, a glassy carbon (GC, 3 mm diameter, BAS Technicol, UK) or a bare basal plane pyrolytic graphite (bppg, 5 mm diameter, Le Carbone Ltd., Sussex, UK) electrode acted as working electrode, a platinum wire wound into a spiral provided the counter electrode and a platinum pseudo-reference electrode completed the cell assembly.

43
The glassy carbon electrodes were mechanically polished between each set of experiments with diamond paste (Kemet, UK) of decreasing particle size (15 to 1 μm) and rinsed thoroughly before use.

44
For the microdroplets study, the bppg electrode (vide supra) was used as working electrode with a platinum wire wound into a spiral (counter electrode) and a saturated calomel reference electrode (SCE, Hg/Hg2Cl2, Radiometer, Copenhagen, Denmark).

45
All potentials are quoted with respect to the SCE, unless otherwise specified.

Procedures

46
Several methods in which the surface of a bppg electrode can be modified with oil droplets containing an electroactive compound have been reported.28

47
In the present study, we investigated droplets of the electroactive hydrophobic yellow material TCBQ.

48
An inert organic liquid was required to dissolve the solid TCBQ and benzonitrile (BN) was chosen as it was found to be electrochemically inert in the potential window of interest.

49
The bppg working electrode was modified with microdroplets of TCBQ dissolved in BN by solvent evaporation of a volatile solvent (MeCN) from an aliquot (typically 5 μL) of TCBQ/BN stock solution of accurately known concentration that had been pipetted onto the electrode surface prior the immersion in aqueous solution.

50
The concentration of TCBQ in the BN oil phase was ca. 35 mM (ca. 1 nmol of TCBQ in a typical volume of 2.7 × 10−8 L of BN).

51
This technique of depositing the organic phase containing the electroactive species in the form of microdroplets rather than a continuous film11 gives the opportunity to omit supporting electrolytes in the organic phase as, due to the presence of the contact lines, electrode|BN|water, charge compensation during electrochemical reduction or oxidation can always be provided at the three-phase boundary.12

52
However, a study of the behaviour of supported microdroplets was also conducted by preparing stock TCBQ/BN in MeCN solution containing the required amount of TBuAP so that the concentration of TBuAP in the microdroplets was either 0.1, 1 or 2 M (see Section 3.2.4).

53
In order to obtain reproducible results, careful bppg electrode preparation was found to be important.

54
The bppg electrode was cleaned prior to each scan by rinsing with MeCN and water, and the surface was renewed by polishing on carborundum paper (P1000 grade, Acton and Bormans, Stevenage, UK) prior to each experimental run.

55
All the experiments were undertaken as soon as possible after immersion of the modified bppg electrode into the aqueous phase.

Computation

56
The voltammetric response of microdroplet partially blocked electrodes was simulated using the same method described in reference.14

57
Simulations for naked electrodes were carried out with DigiSim 3.0©,37 a commercially available computer code, distributed by Bioanalytical systems (2701 Kent Ave., West Lafayette, IN 47906).

Results and discussion

Homogeneous voltammetry

Electrochemical reduction of TCBQ in benzonitrile (BN) at glassy carbon (GC) and basal plane pyrolytic graphite (bppg) electrodes

58
First, we investigated the voltammetric reduction of TCBQ in a homogeneous solution of benzonitrile (BN) containing 0.1 M TBuAP as supporting electrolyte at a GC (3 mm diameter) electrode.

59
The corresponding cyclic voltammetric responses obtained at increasing scan rates (from 0.1 to 1 V s−1) for the reduction of 1 mM TCBQ in BN are shown in Fig. 1a and the potential data are summarised in Table 1.

60
Two well-separated one-electron reduction processes were observed at −0.273 V (vs.

61
Pt) and −1.069 V (vs.

62
Pt), respectively.

63
These can be attributed to the formation of the radical anion, TCBQ˙, which can be further reduced in a second one-electron transfer to the dianion, TCBQ2−.

64
The latter is re-oxidizable in two successive one-electron transfers up to the quinone state.

65
Therefore, the reduction mechanism can be denoted as a quasi-reversible process as expressed by the reactions:The ratios of the forward and backward peak heights of both signals were ≈1 corroborating the formation of a stable radical anion capable of undergoing further reduction and then re-oxidation to the parent.

66
Both the reduction and re-oxidation peaks currents of both couples were found to increase with scan rate and a plot of the reduction current peaks (Ipr1and Ipr2) against the square root of scan rate,ν1/2, (insert plot Fig. 1a) was found to be linear suggesting diffusion-controlled processes.

67
The peak potentials were found to be slightly dependent on the scan rate.

68
With increasing scan rate, the first and the second reduction peaks (Epr1and Epr2) were shifted toward more negative potentials.

69
In contrast, the first and the second re-oxidation peaks (Epo1and Epo2) moved towards more positive potentials.

70
As a consequence, the potential separation ΔEp (=Ep°−Epr) increased with increasing scan rate (see Table 1), consistent with quasi-reversible behaviour.

71
Analysis of the peak currents using the Randles–Sevčik equation and the peak-to-peak separation data using Nicholson’s treatment38 yielded a value for the diffusion coefficient, D, of the radical anion TCBQ˙ in BN of D = 3.6 × 10−6 ± 0.2 cm2 s−1.

72
The reduction of TCBQ was also studied at a bppg electrode (5 mm diameter) and the corresponding voltammograms obtained at increasing scan rate (0.1, 0.2, 05, 0.8 and 1 V s−1) are shown in Fig. 1b and the potential data are summarised in Table 2.

73
Similar to that seen at the GC electrode, the electrochemical reduction of TCBQ revealed two successive one-electron transfer according to a quasi-reversible EE mechanism.

74
However, ΔEp at bppg electrode was larger than ΔEp obtained at the GC electrode (Tables 1 and 2) suggesting more sluggish electron transfer kinetics operate at a bppg electrode as compared to a GC electrode.

75
Furthermore, ΔEp2 was found to be slightly larger than ΔEp1 (Table 2) suggesting that the second electron transfer step is slower than the first one at a bppg electrode.

76
The electrochemical reduction of TCBQ was also carried out in other aprotic solvents, namely DMF, DMSO and PrCN in the presence of 0.1 M TBuAP as supporting electrolyte.

77
The corresponding voltammograms are depicted in Fig. S1 of the electronic supplementary information (ESI) and reveal a quasi-reversible EE mechanism similarly to that seen with BN as solvent.

78
Solvent effects are also discussed in the ESI.

Heterogeneous voltammetry

79
The study of the redox chemistry of TCBQ was extended and its electron transfer properties were investigated at the liquid|liquid interface via immobilisation of electrically insulating microdroplets of TCBQ at a bppg electrode surface as described in the experimental section.

Determination of the average size of microdroplets deposited on a bppg electrode

80
First, the average size of the microdroplets deposited on a bppg electrode surface was determined using a cyclic voltammetric method recently developed by our group.14

81
For a particular redox couple, by measuring the decrease in peak current when the electrode is modified with inert microdroplets compared to when the electrode is naked, we can obtain an estimate of the microdroplet size and total coverage of the electrode surface.14

82
Fig. 2 illustrates cyclic voltammograms recorded at 0.2 V s−1 with a 5.0 mm diameter bppg electrode in 1.5 mM ferrocyanide/0.1 M KCl solution where the electrode is (a) naked and (b) partially covered with 2.7 × 10−5 cm3 of benzonitrile, corresponding to 10 μL of 26 mM benzonitrile (BN) in acetonitrile (MeCN).

83
The voltammogram for the modified electrode shows a clear decrease in peak current, Ip, and increase in peak to peak separation, ΔEp, when compared to the response of the naked electrode.

84
Both observations are consistent with an inertly blocked surface indicating the microdroplets are inert in this particular medium and potential window.

85
Analysis of the naked electrode data (for a range of scan rates) with DigiSim 3.0© gave D = 6.5 × 10−6 cm2 s−1 and k0 = 0.01 cm s−1 for the Fe(CN)64−/Fe(CN)63− redox couple, which agrees well with that determined previously for carbon electrodes.39

86
We can model the microdroplet partially blocked electrode surface as a random array of overlapping hemispheres of the same radius.14

87
This is based on the experimental observation that at low surface coverage when microdroplets are well spaced, their distribution is approximately monodisperse, whereas at high surface coverage there is a larger range of droplet radii.14

88
Using the method described in ref. 14 we can simulate the dependence of peak current with droplet radii for a range of scan rates.

89
This is illustrated in Fig. 3 for the modified electrode discussed above at scan rates of 0.1, 0.2, 0.5, 1.0 and 2.0 V s−1.

90
Overlaid as dashed horizontal lines are the experimental peak currents recorded at each scan rate.

91
The points where the dashed and solid lines cross are highlighted in the figure and represent the best agreement between theory and experiment.

92
As observed, the results suggest an average droplet radius of 2.5 μm (experiments where the benzonitrile droplets contained 0.1 M TBuAP gave a similar result).

93
Fig. 4 illustrates a 100 × 100 μm section of the corresponding model electrode surface that best fits the experimental results, i.e. where the droplet radius is 2.5 μm.

94
It should be noted that Fig. 4 corresponds to how the model perceives the electrode surface.

95
In reality, overlapping droplets would coalesce, forming larger droplets of different radii.

96
Our approach to modelling random arrays of droplets ignores the resulting geometry from hemisphere/droplet overlap allowing us to treat the surface as a random array of overlapping discs, which is far easier to simulate the voltammetry of.14

97
Indeed, inclusion of droplet coalescence in the modelling process would result in a problem requiring far more computation time.

98
For this particular block volume there is a large amount of overlap resulting in a wide range of observable droplet radii.

99
We can calculate the fraction of the uncovered electrode area to be Θ = (uncovered electrode area)/(total electrode area) = 0.44.

100
Note that only 44% of the electrode is available for the electrochemical redox reaction yet the peak current of the modified electrode greatly exceeds 44% of that for the naked electrode.

101
Thus for redox reactions of this type, the uncovered surface area of the microdroplet modified electrode behaves more like an array of microelectrodes than a macro electrode of the same total area.

Voltammetry of unsupported microdroplets of TCBQ deposited on a bppg electrode immersed in aqueous 0.1 M Y+Cl (Y = Li+, K+, Rb+, TMeA+, TPrA+, TBuA+ or TPeA+)

102
In this section we discuss the voltammetric processes that occur at a bppg electrode (5 mm diameter) covered with a random arrays of purposely unsupported electroactive microdroplets of TCBQ dissolved in BN (formed as described in 2.3 Procedure) and immersed into an aqueous electrolyte where the droplet organic phase in insoluble and in which the SCE reference and the Pt counter electrodes are placed.

103
In these experiments TCBQ is reduced to the respective anion radical TCBQ˙.

104
Hence the electroneutrality of the oil environment can be retained by either the uptake of a cation from the aqueous phase or the expulsion of a newly generated anion into the aqueous phase.

105
It can be anticipated that as the aqueous cation hydrophilicity increases, it becomes increasingly energetically more difficult to transfer the cation into the oil phase, until the cation can no longer thermodynamically transfer and the organic phase anion is expelled into the aqueous phase.

106
We have examined aqueous chloride solutions containing the hydrophilic alkaline metal cations Li+, K+ and Rb+ and the less hydrophilic tetraalkylammonium cations NR4+ (R = Me, Pr, Bu and Pe).

107
We first consider the response obtained in 0.1 M A+Cl (A+ = K+, Li+, Rb+) aqueous solutions.

108
Fig. 5a shows 5 successive scans (0.1 V s−1) obtained at a bppg electrode modified with 1 nmol of TCBQ immersed in 0.1 M KCl aqueous solution.

109
The first cycle exhibits a single reduction peak at −0.052V (vs.

110
SCE) and a small back oxidative peak at +0.008V (vs.

111
SCE) (ratio of the anodic to cathodic charge Q°/Qr = 0.26) giving rise to a peak-to-peak-potential separation of 60 mV.

112
With increasing potential redox cycling on the same electrode, the prominent reductive peak magnitude decreases and almost no anodic response remains.

113
A slightly shift toward less negative reductive potential is observed on the second scan.

114
This might be due to an electrochemically induced partitioning of the electrolyte into the droplets causing the electrical resistance of the droplets to diminish which in turn makes the reduction process easier or, more likely, to the reduction of the smaller size droplets.

115
A scan rate dependence study was also carried out.

116
The corresponding initial (first cycle) voltammograms recorded at different scan rate (0.05, 0.1, 0.2, 0.5 and 1.0 V s−1) are shown in Fig. 5b and the voltammetric data summarised in Table 3.

117
A single reductive wave with a corresponding small oxidative wave is observed at all scan rates.

118
The ratio anodic charge to cathodic charge (see Table 3) increases slightly with increasing scan rate, however data obtained at scan rate larger that 0.2 V s−1 should be treated carefully since ohmic losses in the unsupported microdroplets are not negligible.

119
A plot of the reductive peak current versus scan rate (insert plot of Fig. 5b) is shown to be linear, consistent with an essentially complete electrolysis with a thin layer of deposited material.

120
Fig. 5c exhibits the voltammograms obtained when the potential was cycled twice at very slow scan rate (i.e, 10 mV s−1).

121
The first scan exhibits a reductive wave at −0.0011 V and no appreciable anodic counterpart.

122
The second scan shows little further than resistive behaviour suggesting that complete reduction of the whole droplet has occurred during the first cycle.

123
Analysis of the amount of charge passed (i.e, integrated area of the reductive peak) on the first reductive scan potential reveals the transfer of a single electron per TCBQ molecule (see Table 4).

124
A similar response was observed for other very hydrophilic metal cations such as Li and Rb and the corresponding voltammograms obtained at 0.01 V s−1 are overlaid for comparison in Fig. 6a.

125
A single reductive peak of similar magnitude is observed at ca. +0.0 V for all cations.

126
Furthermore, plots of the reductive peak current against scan rate (Fig. 6b) are linear for each different metal cation and the amount of charge passed suggested, in all cases, the discharge of one electron per molecule of TCBQ (Table 4).

127
This type of electrochemical behaviour is somehow not surprisingly considering the high transfer free energy possessed by the alkali metal40 which makes the insertion into the oil deposit a thermodynamically unfavourable process.

128
Hence the electroneutrality of the organic phase is maintained by expulsion of the electro-generated radical anion into the aqueous phase as expressed by:TCBQ(oil) + e → TCBQ˙(aq)As the radical anion diffuses out of the diffusion layer during reduction, little radical anion remains to be oxidised and therefore little oxidative back peak is observed.

129
Furthermore, this process causes depletion of TCBQ in the oil phase and hence, on successive scans, the peak height decreases as shown in Fig. 5a.

130
This has been described as the electrochemically induced anion dissolution mechanism.12

131
It has been shown that, in the case of anion dissolution, the electron transfer might be initiated at the three phase boundary, but the dissolution may occur from the liquid|liquid interface.

132
The large currents observed further support this interpretation.

133
We now turn to the voltammetric characteristics of the microdroplets in the presence of less hydrophilic cations, i.e the quaternary alkylammonium cations NR4+ (R = Me, Pr, Bu and Pe) in the aqueous phase.

134
Fig. 7a, 7b, 7c and 7d shows the cyclic voltammograms (0.1 V s−1) obtained for the reduction of TCBQ deposited onto a bppg electrode and immersed in aqueous 0.1 M TMeACl, TPrACl, TBuACl and TPeACl, respectively.

135
By increasing the length of the alkyl chain of the cation the hydrophobicity of the aqueous-based cation increases and this has profound effects on the electrochemical behaviour of the TCBQ microdroplets.

136
In 0.1M TMeACl solutions (Fig. 7a) the response is qualitatively similar to that observed in aqueous alkali metal chloride solutions, with a single reductive peak at −0.05 V and a very little back oxidative peak.

137
The signal was found to diminish on increasing redox cycling and charge integration at low scan rates (0.01 V s −1) reveals the transfer of one electron per molecule of TCBQ suggesting that the dissolution of the electrogenerated radical anion is still more favourable process than the insertion of TMeA+.

138
The dissolution rather than the cation insertion mechanism appears to be the pathway followed also in 0.1 M TPrACl aqueous solution.

139
However, in this case, the associated voltammetry (Fig. 7b) is more complex as it was found to be dependent on the scan rate.

140
At low scan rate (0.01 V s−1) a single reductive wave can be seen at −0.04 V, consistent with the dissolution of the radical anion.

141
However, based on the charge under the peak the number of electrons associated with this signal is only 0.5, suggesting that incomplete droplet conversion takes place under these conditions.

142
Furthermore, voltammograms recorded at scan rate faster than 0.050 V s−1 (also overlaid in Fig. 7b) exhibit two reductive peaks at +0.048V and at −0.1 V (vs.

143
SCE), respectively.

144
These could be tentatively attributed to the reduction of TCBQ to the radical anion and to the dianion, respectively, upon the uptake of TMeA+ from the solution to balance the charge for TCBQ being reduced.

145
However, the fact that the anodic to cathodic charge ratio is much less than one for both peaks, suggests that insertion pathway is only a minor contribution to the overall electrochemical process which is thought to proceed mainly via dissolution of the electrogenerated species.

146
Fig. 7c shows three successive redox cycles obtained in 0.1M TBuACl solutions at 0.1 V s−1.

147
The first scan exhibits a single peak at +0.219 V (vs.

148
SCE) with a corresponding oxidative peak at +0.289 V (vs.

149
SCE) giving rise a peak-to-peak separation of 70 mV.

150
The re-oxidative peak current is of the same magnitude of the current associated with the reductive process and the anodic charge passed on the return of the scan is equal to the cathodic charge passed in the preceding negative going scan.

151
Finally, analysis of the charge passed at 0.01 V s−1 suggests the transfer of 1 electron per molecule of TCBQ.

152
All these observations are consistent with TBuA+ insertion into the oil phase as expressed by:TCBQ(oil) + TBuA+ (aq) + e → [TCBQ˙TBuA+](oil)As TCBQ is reduced to the radical anions TCBQ˙, TBuA+ cations transfer across the liquid|liquid interface (specifically at the three phase boundary3,4,30) to preserve the electroneutrality of the electrically insulating oil microdroplets.

153
A comparison of the reduction peak potential observed in aqueous TBuACl and in alkali metal chloride solutions reveals a shift versus less negative potential of ca. 0.27 V which further supports the assignment of the insertion mechanism.

154
A slightly dissolution, probably via ion-pairs,15,16,33,41–44 is thought to be responsible for the diminished signal obtained upon continual redox cycling.

155
The possibility of cation adsorption followed by its slow transfer across the liquid|liquid interface cannot be ruled out.

156
We finally turn our attention to the current|voltage characteristics of TCBQ microdroplets in 0.1 M TPeACl.

157
The corresponding initial voltammogram recorded at 0.05 V s−1 (Fig. 7d) reveals the appearance of a sharp reductive peak at +0.3 V (vs.

158
SCE), a broad reductive peak at ca. −0.10 (vs.

159
SCE) and a small third reductive wave at −0.245 V (vs.

160
SCE).

161
Upon reversal of the scan three corresponding oxidative back peaks are observed at +0.319 V, −0.215 V and −0.34 V, respectively.

162
As shown in Fig. 8, a stable response of all three peaks upon continual redox cycling is obtained at faster scan rate (0.5 V s−1).

163
Therefore, we speculate the reversible TPeA+ insertion and the assignment of the two reductive peaks as expressed by:TCBQ(oil) + TPeA+ (aq) + e → [TCBQ˙TPeA+](oil)[TCBQ˙TPeA+](oil) + TPeA+(oil) + e → [TCBQ2−(TpeA+)2](oil)This response is analogous to that observed in homogeneous solution of BN (see section 3.1, Homogeneous voltammetry).

164
Voltammograms recorded when the droplets where purposely made conductive via the addition of TBuAP in the oil phase (see section 3.4) exhibited the same voltammetric characteristics of the unsupported droplets, which further suggests the insertion of TPrA+ as given in eqns. (5) and (6).

Effect of increasing the aqueous phase proton concentration on the voltammetry of TCBQ microdroplets

165
The pH dependence on the reduction of unsupported microdroplets of TCBQ was examined in buffer solutions with pH values of 4, 7 and 10 (see the Experimental section for a detailed specification of each buffer composition).

166
Fig. 9 exhibits the corresponding cyclic voltammograms (scan rate 0.01 V s−1) for the reduction of 1.1 nmoles of TCBQ immobilised on a bppg electrode immersed in each different buffer solution.

167
It can be seen that the response obtained at pH 10 is similar to that obtained in 0.1 M alkali metal chloride solutions, with a single reductive peak at −0.013 V and almost negligible anodic counterpart.

168
Analysis of the amount of charge passed at 0.01 V s−1 reveals the transfer of 1 electron per molecule of TCBQ.

169
This, together with a linear relationship of the magnitude of the reductive peak with scan rate, suggests that the radical anion TCBQ˙ dissolves out of the surface confined organic droplet phase.

170
However, as shown in Fig. 9, by increasing the concentration of protons in the aqueous phase, the reduction of TCBQ occurs at less negative potentials (the observed reductive peak potential were +0.042 V at pH 7 and +0.173 V at pH 4, respectively).

171
A closer examination of this figure also reveals that, in the presence of a high proton concentration, the voltammetry is qualitatively different form that observed in the absence of protons.

172
At lower pH values the signal shape is more symmetric and the re-oxidative peak current is of the same magnitude of the forward going reductive peak current.

173
This pH dependence can be rationalised considering that protons are positively-charged species that are able to move into the oil phase to neutralised any negative charge that develops during the reduction of TCBQ.

174
The liquid|liquid protonation process is well-documented in the literature11,25 and protons are perhaps the most common type of electrolyte ions partitioning into an organic phase from an aqueous phase.

175
The sensitivity of quinone species to protons is also well-know.45,46

176
Thus, we might write for the reductive peak observed:TCBQ(oil) + mH+(aq) + ne → [TCBQnHm+](oil)We can then apply the Nernst equation to the proton insertion process described above in eqn. (7):The identity of the [TCBQnHm+](oil) species could, in principle, be inferred by the analysis of the variation of the reduction peak potential Epr with proton concentration together with the quantification of the amount of charge passed during the reductive process which should yield n for the process given in eqn. (7).

177
In the range 4 ≤ pH ≤ 7, Epr for the single reductive peak observed varies as 51 mV per decadic change in proton concentration, indicative (within the experimental error resulting from the analysis of distorted voltammograms associated with the resistive nature of the droplets) of m = n in the process given in (7).

178
Monitoring of the charge passed at low scan rate (0.01 V s−1) yields values of n = 1.2 ± 0.2 at pH 4 and n = 1.15 ± 0.2 at pH 7.

179
However, this data should be treated with caution as second scans on the same electrode reveals that exhaustive electrolysis did not occur at either pH 7 or pH 4, under these conditions.

180
A quantitative analysis of this data is further complicated because of difference in the composition of the liquid|liquid interface determined by different identity and concentration of anions present in each different buffer solutions.

181
Moreover, reactions between the semiquinone and the dianion of various quinone derivatives are well known46 and thus, a cascade of follow-up chemistry via protonation can be anticipated.

182
Nevertheless, the present results are qualitatively not inconsistent with either the transfer of one proton to the oil phase to protonate the electrogenerated radical anion TCBQ˙ (and produce a neutral free radical anion TCBQH˙) or, more likely, the transfer of two protons and a two–electron reduction of TCBQ to the hydroquinone structure TCBQH2.

183
This field-induced proton migration into and across the oil phase provides the ionic current required for the electron transfer to occur thereby ensuring electroneutrality within the microdroplets environment is maintained.

184
The observed selectivity towards proton insertion as compared to alkali metal cation insertion discussed above is contrary to thermodynamic considerations for which the free energy of transfer across the BN|water interface is least favourable for protons.40

185
However, the higher efficiency of proton insertion, as compared with Group I cations insertion, reflects the basicity of the reduction products of TCBQ.

186
Thus the proton phase is induced by the formation of strong O–H bonds within the hydroquinone molecules.

Voltammetry of purposely supported microdroplets of TCBQ deposited on a bppg electrode immersed in aqueous 0.1 M Y+Cl (Y = Li+, TPeA+ or TBuA+)

187
Having assessed the properties of unsupported TCBQ microdroplets, we now discuss the behaviour of microdroplets made conductive by addition of a hydrophobic, nonpartitioning supporting electrolyte such as TBuAP dissolved in the oil phase.

188
As the microdroplets are now supported, electron transfer may occur within the microdroplets or at the liquid|liquid interface.

189
We have studied the voltammetric characteristics of supported (with TBuAP) droplets of TCBQ immobilised on the bppg electrode immersed in three different aqueous chloride solutions containing the non-inserting cation Li+ or the inserting TBuA+ and TPeA+.

190
We first examined the effect of the presence of TBuAP dissolved in the oil phase microdroplets on the reduction of 1 nmole TCBQ immobilised at a bppg electrode immersed in aqueous 0.1 M LiCl.

191
Shown in Fig. 10a is the effect of increasing the concentration of TBuAP in the oil phase (from 0 to 2 M) on the response of TCBQ microdroplets in 0.1 M LiCl solutions at 0.05 V s−1.

192
It can be seen that the presence of a supporting electrolyte within the oil deposit has a profound effect on the resulting voltammetry of TCBQ.

193
The current|voltage profiles are now characterised by two reductive peaks (as opposed to the single reductive peak observed in the case of unsupported droplets) and two small corresponding oxidative back peaks.

194
Furthermore, as the TBuAP concentration increases, the first reductive peak shifts versus less negative potential meanwhile the second reductive peak shifts versus more negative potential so that the potential separation between the two reductive peaks increases as the droplets are made more and more conductive.

195
Voltammograms were also recorded for the supported droplets (with 2 M TBuAP) when the potential was reversed after the first reductive process (Fig. 10b).

196
In this case, the ratio of anodic to cathodic peak current was found to be ≈1 and the signal was found to decrease only slightly upon continual redox cycling.

197
Furthermore, a scan rate dependence study reveals that the first cathodic peak scales linearly with the scan rate and the peak-to-peak separation relative to the first reductive wave increases with scan rate, from 20 mV at 0.05 V s−1 to 200 mV at 1 V s−1.

198
Hence, the two reductive peaks can be attributed to the reduction of TCBQ to the radical anion TCBQ˙ and to the dianion TCBQ2−, respectively, as observed in homogeneous solution of BN (vide supra).

199
Next we examine the electrochemical behaviour of supported TCBQ microdroplets immersed in an aqueous solution containing 0.1 M TPeACl.

200
Overlaid in Fig. 11 are the responses obtained in the absence and in the presence of 2 M TBuAP in the oil phase.

201
No substantial change in the peaks appearance or in the peak shape is observed.

202
Hence, addition of a nonpartitioning hydrophobic electrolyte such at TBuAP to the oil phase causes no change in the reduction of the immobilised microdroplets of TCBQ when 0.1M TPeACl is dissolved in the aqueous phase.

203
Thus, it can be inferred that the insertion of the TPeA+ cation is unaffected by the addition of an electrolyte in the oil phase.

204
However, when the supporting electrolyte dissolved in the aqueous phase still contains a more hydrophilic inserting cation than TPeA+ the addition of a hydrophobic salt (i.e., TBuAP) to the oil phase microdroplets results in an anomalous and more difficult to explain voltammetric behaviour.

205
In Fig. 12a, 12b and 12c are overlaid the first four current|voltage profiles obtained for the reduction of TCBQ microdroplets immersed in 0.1 M TBuACl in the absence and in the presence of 0.1, 1 and 2 M TBuAP, respectively, in the oil phase.

206
It can be seen that, as the TBuAP concentration in the oil phase increases, the reductive peak moves versus more negative potential and the anodic counterpart of the voltammograms is almost lost as the droplets are made more conductive.

207
The response from the TCBQ/TCBQ˙ couple is stable upon continual potential cycling.

208
The fact that it is more difficult to reduce TCBQ as the resistance within the droplets decreases, is, at first, surprising.

209
However, this might be explained considering that the reduction of TCBQ occurs upon TBuA+ insertion, the latter being more and more difficult process to occur as the TBuAP concentration in the oil phase is increased.

210
Alternatively, it might be suggested the electroneutrality within the oil deposit is maintained by expulsion of the perchlorate anion into the aqueous phase.

211
The fact that perchlorate anion is toward the hydrophobic end of the Hofmeister scale, might account for the observed stretched voltammograms (i.e., perchlorate anion must be “dragged” out of the oil phase into the aqueous phase).

Conclusions

212
The method of immobilising microdroplets of TCBQ on the surface of a bppg electrode and then investigating the electrochemical reduction when the electrode is placed in aqueous electrolyte solutions has given new insights into the redox chemistry of TCBQ because, in contrast to the case of its reduction in homogeneous media, the redox conversion of TCBQ in the form of microdroplets requires the uptake or expulsion of ions dictated by the requirements of electroneutrality.

213
Hence, this study has provided new data on the nature of the processes occurring at the liquid|liquid interface involved in the reduction of microdroplets containing dissolved TCBQ.

214
The variety of experimental data collected reveals that the response changes depending on the identity and concentration of the cation dissolved in the aqueous phase.

215
Upon reduction, the electroneutrality within the microdroplet environment, was shown to be preserved by the occurrence of two different type of charge neutralisation processes depending on the hydrophilicity of the cation in the aqueous phase: cation electroinsertion (for the large organic cations TBuA+ and TPeA+) or electrogenerated-anion expulsion (for the hydrophilic alkali metal cations).

216
Finally, the large shift versus less negative reductive potential observed in acidic media was explained in terms of the basicity of the reduction products of TCBQ and it showed that cross-phase proton transfer can control the potential where Faradaic processes take place.