Catalytic Activity of Alkali Metal Cations for the Chemical Oxygen Reduction Reaction in a Biphasic Liquid System Probed by Scanning Electrochemical Microscopy

Abstract Chemical reduction of dioxygen in organic solvents for the production of reactive oxygen species or the concomitant oxidation of organic substrates can be enhanced by the separation of products and educts in biphasic liquid systems. Here, the coupled electron and ion transfer processes is studied as well as reagent fluxes across the liquid|liquid interface for the chemical reduction of dioxygen by decamethylferrocene (DMFc) in a dichloroethane‐based organic electrolyte forming an interface with an aqueous electrolyte containing alkali metal ions. This interface is stabilized at the orifice of a pipette, across which a Galvani potential difference is externally applied and precisely adjusted to enforce the transfer of different alkali metal ions from the aqueous to the organic electrolyte. The oxygen reduction is followed by H2O2 detection in the aqueous phase close to the interface by a microelectrode of a scanning electrochemical microscope (SECM). The results prove a strong catalytic effect of hydrated alkali metal ions on the formation rate of H2O2, which varies systematically with the acidity of the transferred alkali metal ions in the organic phase.


Introduction
Liquid j liquid interfaces formedb etween two immiscible electrolyte solutions represent ab iomimetic reaction system for advanced oxidation of organic and metallorganic substrates, the formationo fr eactive oxygen species, and integrated extraction of reactionproducts. [1] Although the conduction of the reactionm ay be as simple as stirring ab iphasic liquid mixture and separating the two phases, such systems have much more to offer if they are combined with acontrol of the transfer processes of ions, electrons, and neutrals pecies across the liquid j liquid interface.
In this realm, ion-coupled electron transfer reactions such as dioxygen (O 2 )reduction to hydrogen peroxide (H 2 O 2 )and (photo)generation of H 2 were studied with model compounds as molecular electron donors, including decamethylferrocene (DMFc), [1c, 2] ferrocene( Fc), [1c, 2c] 1,2-diferrocenylethane, [3] decamethylosmocene, [4] andd ecamethylruthenocene, [5] dissolved in the organic phase. In the case of osmocene and ruthenocene derivatives, H 2 was evolved under light exposure from H + supplied in the aqueous phase (aq.). In all studied cases, the reaction proceeds by pumpingt he H + into the organic phase (o). The ion transfer (IT) could be controlled precisely either by using ap otentiostat to supply the free energy of ion transfer from the aqueous to the organic phase or by addition of a phase transfer catalyst, for example, lithium tetrakis(pentafluorophenyl)b orate (LiTB), to the acidic aqueous phase. Subsequently,t he reduction reactiono ccurred in the presenceo fa n appropriate electron donor species.
Although the transfer of ah ydrophilic ion from the aqueous to the organic phase is ak ey step in H 2 evolution in liquid j liquid systems, the oxygen reduction reaction( ORR) by DMFc can occur even in the absence of acidity in the aqueous phase. [6] It is suggested that instead of ah ydrated proton, a solvated alkali metal cation M + (aq.) is transferred, which is then able to provideaprotonf or the reaction according to the following scheme [ This sequence has been demonstrated for Li + ,i nw hich the hydrophilic cation polarizest he water molecules of its hydration shell, making thema cidic in an aprotic polar solvent like dichloroethane (DCE). The slightly acidic water of the hydration shell can then donate protonsf or both the oxygen reduction and hydrogen evolution reactions with DMFc as electron donor. [6] In this context, as ystematic understanding of the catalytic action of other metal cationsa sw ell as aq uantitative characterization of their reactivity effects on ion-coupled electron transfer reactions such as biphasic ORR is of fundamental importance for the evaluation of interfaces in fluidic systems for novel controlled reagentd elivery systems, energy-related systems, or advanced oxidation of organic substrate. [1a, b, d, f, g, 7] This may also concern reactions at the liquid j liquid interface either to control polymer microstructures, [8] to study reaction kinetics, [9] the effects of counter ions and doping, [10] or to prepare smart nanocarrierss uch as synthetic polymer shells with an aqueousc ore. [11] This also includes the use of phase-transfer catalysts to conduct the reaction selectively in one phase of choice, for example, to protect the product from hydrolysis in the aqueous phase, [12] or to design molecular click reactions at al iquid j liquid interface. [13] Interfacial reaction and mass transport processes have been disentangled by placing an microdisk electrode as as ensor close to any liquid j liquid interfaces and using the instrumentation for scanning electrochemical microscopy (SECM). [14] Reaction products that were detected by SECM include H 2 O 2 , [15] O 2 , [16] H 2 [15f, 17] in the substrate-generation/tip collection (SG/TG) mode. Recently,t he feedback mode of SECM was used to study chemically polarized liquid j liquid interfaces, that is, systems in which aG alvanip otentiald ifference D w o f between the two immiscible electrolyte solutions was formed by partitioning of ac ommoni on. Such systemsa lso facilitate the spontaneous assembly of charged photoactive nanoparticles, for example,B iVO 4 .S ECM was used to study the reaction of photogenerated holes and conduction band electrons at nanoparticle-decorated liquid j liquid interfaces. [16,18] The surface interrogation mode (SI-SECM), which can be considered as at ransient feedback experiment, was used to assess the amounta nd the decay kinetics of photogenerated surface-bound intermediates of the water oxidation reaction at BiVO 4 -decorated liquid j liquid interfaces. [19] Here, we use am odified setup for SECM for operando studies of the catalytic behavior of alkali metal ions during oxygen reduction ata ne xternally biased liquid j liquid interface. For that purpose, the liquid j liquid interface under study is mechanically stabilized at the orifice of am icropipette( MP, Figure 1). This setup also enablest he application of aw ell-defined potential differencea cross the liquid j liquid interface by ap otentiostat in at wo-electrode arrangementb etween working electrode 1(WE1) inside the pipette and acombined auxiliary and reference electrode (Aux1-RE1) in the aqueous solution. Thisa djustable potential is used to drive the transfer of alkali metal ions across the interface. Here, ion-transfer cyclic voltammetry (ITCV) is recorded for different ions at the MP with the liquid j liquid interface. The colinearly positioned Pt microelectrode (ME, WE2) is biased by as econd potentiostat in at hree-electrodec ell and is used for detection of ORR products (e.g.,H 2 O 2 )i nt he SG/TG mode.

Ion-transfer cyclic voltammetry
The compositiono ft he cell comprising an aqueous reference solution,a na queous phase, and an organic phase is outlined in Figure 2. The interface under study is formed between phases II and III. Phase II contains5m m DMFc as reductant and 5mm of very hydrophobic electrolyte bis(triphenylphosphoranylidene) ammonium tetrakis(pentafluorophenyl)borate (BATB). The aqueous phaseIII contains 100 mm of either hydrochlorica cid (HCl), lithium chloride (LiCl), sodium chloride (NaCl), or potassium chloride (KCl). Figure 3shows the ITCV obtained at the water j DCE interface at the orifice of the MP when using the two-electrode electrochemicalc ell outlined in Figure 2. The electrochemical response in Figure 3c urve 1e xhibitsap otential window of about 0.8 Vf or the background electrolytes, that is, Li + and Cl À ions in the aqueous phase and BA + and TB À in DCE. There is an asymmetric diffusionl ayer for an ion-transfer on both sides of the liquid j liquid interface. For ar eversible IT from  inside the pipette to the outside (egress transfer), linear diffusion inside the elongated taper of the MP controls the mass transport, whereas hemispherical diffusion is the limiting process for the ion-transfer from outside the pipette to the inside (ingress transfer). [14e] Thet wo diffusion regimes are associated with qualitatively different shapes of the resulting voltammetric curves and allow an assignment of the signals to specific transfer processes, which is usually difficult formacroscopic liquid j liquid interfaces. [20] Thei ngress transfer leads to as teadystate currenta nd the egress transfer resultsi napeak-shaped wave. At negative potentials at WE1 inside the MP versust he externalA ux1-RE1, the potential window is limited either by the ingress transfer of Li + to the MP,o rb yt he egresst ransfer of TB À from the MP.T here is ap eak at the reverse scan indicating that the current is due to an IT of from inside to the outside of the pipette. Consequently,t his side of the potential window is limited by the transfer of Li + ,a st he return peak must correspond to the transfer of an aqueous cation back from the MP after having transferred during the forward scan. The positive end of the potential window is determined by the transfer of either BA + or Cl À .Ap eak current is observed when the potential scan is reversed. This is caused by al inear diffusion processo fa ni on inside the MP,w hich has entered the pipette in the forwards can. Thus, this side of the potential window is limited by the transfer of Cl À .
As shown in Figure 4a,t he current wave at the negative potential limit shifts to different values for Li + ,N a + ,a nd K + and H + for identical initial concentrationso f[ MCl] = [HCl] = 0.1 m for all of the cations in the aqueous phase. Consequently,d ifferent potentialw indows are availablei nt he corresponding electrolytes, which is in good agreement with the tabulated transfer potentials (D w o f)o ft hose cations that are in the se-quenceN a + > Li + > H + > K + , [21] unless, the experimental conditions, for example, the concentration of electrolytes, size of liq-uid j liquid interface, reverset he sequence between Na + and Li + . [22] Upon addition of DMFc as electron donort ot he DCE phase, the negative current wave at D w o f = À0.65 Vi ncreases for the Li + transfer from the aqueous to the organic phase inside the MP ( Figure 3, curve 2). Additionally,t he onset of the current wave shifts by approximately 0.2 Vt owards positive potentials. The peak-shapedr esponse in curve 1d isappears from the reverse half scan of curve 2. The peak in curve 1i sc aused by the back transfer of Li + from the inside of the pipette to the outside. Thei rreversible transfer of aL i + from the aqueous to the organic phase in the presence of DMFc demonstrates that the solvated Li + ions enter into the ORR in this biphasic system, which may be associatedw ith the production of H 2 O 2 ,i n agreement with previousr eports about the catalytic role of Li + in ORR. [6] The Li + ion facilitates the transfer of H + needed for ORR. Therefore, its role has been described as that of ap hasetransfer catalyst. Thep ositive currentw ave for transfer of Cl À  from the aqueous to organic phase also increases in the presence of DMFc without change in the peak potential. Furthermore, the peak-shaped response for the backt ransfer of Cl À to the aqueous phase does not disappear and does not diminish. This is in strong contrast to the observation for the cation transfer.T herefore, we conclude that Cl À ions do not have any catalytic effect in the presence of DMFc. This is also not expected because Cl À acts as aL ewisb ase and cannot transfer a H + from aw ater molecule of the hydrations hell to DMFc despite the fact that Cl À can transfer into the organic phase along with its hydration shell (similar but not equal to Li + ). [23] The currenti ncrease in the both half scans may have the following reasons:i on-pairing between Cl À and Li + as ar esult of higher ion-transfer currents in the presence of DMFc for both half cycles. Alternatively,a ni ncreasedc apacitance of the liquid j liquid interface owing to the larger flux of Li + in presence of DMFc could also explain the increased currents in both half cycles.
As imilar catalytic behavior was also observed by us in ITCV of HCl and other aqueous alkali chloride solutions such as NaCl and KCl in the presence of DMFc (Figure 4b). The results demonstrate the catalytic role of alkali cations and H + in ORR in this biphasic system. However,t he comparison of the catalytic behavior is not informative or reliable when only using iontransfer CV.A ccordingly,t he SECM setup is developed for the miniaturized liquid j liquid interface as explained below.T he biphasic ORR with DMFc as reductant could be demonstrated for the first time with aqueous Li + and in the absence of an excess H + in the aqueous phase by using the capillarym ethodology and as imple two-electrode setup. This extends the scope of previous reports in whicht his reactionw as conducted with transfer of H + from the aqueous to the organicp hase in am ore conventionalf our-electrode liquid j liquid electrochemicalcell or dropletelectrode configuration. [2b, 15a] SECM measurements:ins itu electrochemicald etection of hydrogenp eroxide To corroborate the ITCV results, we aimed for the directd etection of the ORR reaction product. In ap revious report, H 2 O 2 was detected as ap roduct of O 2 reduction by DMFc at the liquid j liquid interface with an acidic aqueous solution by using a three-electrode droplet configurationi naSECM setup. [15a] The detector for H 2 O 2 was the positionable ME of the SECM instrument with the specific advantage that H 2 O 2 is collectedb efore it can be diluted in the aqueous phase. This greatlye nhances the sensitivity of the detection method. In our setup, the H 2 O 2 oxidation current is recorded at the ME, whichi sc olinearly alignedw ith the MP (Supporting Information, sectionS3) and moves from the bulk of the aqueous solution towards the liquid j liquid interface while recording the current. The potential appliedt ot he MP drives solvated cations over the interface. Surprisingly,H 2 O 2 is also detected, and hence ORR proceeds in the absence of as urplus of H + in the aqueous phase (Figure 5a,c urve 2). The ORR can still occur provided that alkali metal cations are transferred from water to the DCE phase, a processt hat can be externally controlled by applying as uit-able negative potential of D w o f = À0.65 Va tt he liquid j liquid interface. The precise value is obtained for diffusion-controlled transfer of alkali metal cations by ITCV such as in Figure 4a.I n contrast, the ME current changes only negligibly during the approacht ot he liquid j liquid interface if no potentiald rop is appliedacross the liquid j liquid interface under otherwise identical conditions (Figure 5a,c urve 1). In this situation, no Li + can transfer from the aqueous to the organic phase and hence ORR and H 2 O 2 production cannot happen.
On approachingt he interface under the conditions of ion transfer in Figure 5a,c urve 2, the H 2 O 2 oxidation current increasesb ecause of the higher H 2 O 2 concentration close to the interface. This marks the liquid j liquid interfacea st he local source of H 2 O 2 .A dditional approach curves in Figure S4 (in the Supporting Information) show negligible oxidation current upon approach of the MP to the ME for liquid j liquid interfaces biased at different potentials within the available potential window from ITCV in Figure 3, curve 1. This confirms the necessity of Li + transfer from the aqueous phase to the organic phase to facilitate the ORR by DMFc in the organic phase. Figure 5a provest he dependence of ORR on the presence of Li + transferred from the aqueous to the organic phase. Recently,G irault and co-workers [6] proposed am echanism for this reaction[ Eqs. (3)-(5)] Briefly,t he mechanism considers the hydrophilic alkali ion (e.g.,L i + (aq.) )a sL ewis acidic towardsw ater molecules of their solvation shell. When transferred to the organic phase, Li + (aq.) can transfer as lightly acidic protont o DMFc, forming [DMFc-H] + .T his speciesc an enter into the ORR similarly to solvated protons. [2a, b] ½LiðH 2 OÞ n þ þ DMFc !½LiOHðH 2 OÞ nÀ1 þ½DMFc-H þ ð3Þ All the above reactions take place in the DCE phase. However,o wing to the presence of the liquid j liquid interface, LiOH and H 2 O 2 will transfer into the aqueous phase. Under these conditions, the calculated total Gibbs free energies for the reaction [Eqs. (3)-(5)] are À111, À108, À96 kJ mol À1 for Li + ,N a + , and K + ,r espectively,w hich are clearly thermodynamically favorable (Supporting Information, sectionS5). However,t he Gibbs free energy in the absence of M + is + 57.3. [6a] In this comparison, the term D w for the back transfer of the alkali metal cation (as an ion-pair with OH À )m akes the energetic differencet hat favors the chemical ORR after the enforced transfer of alkali metal ions. The differentiation between the cations is in line with the solvation of those cations according to the non-Bornian solvation model taking into account the charge, hydration radius,a nd hydration number of each cation. [23a-c] As imilar sequence is reached by lookinga tt he hydration enthalpies ( Figure 6) and the hydration entropy, which make only as mall modification to the trend from the enthalpies (DS Li þ = À142 Jmol À1 K À1 , DS Na þ = À103 Jmol À1 K À1 , DS K þ = À88 Jmol À1 K À1 ). [24] These calculations confirm that the presence of M + is essential for such reactions to proceed.
The procedure was also applied to other alkali metal cations to compare their catalytic activity for biphasic ORR. Figure 5b shows ac omparison of the SG/TC approach curves based on H 2 O 2 oxidationa tt he ME for different aqueous alkali metal chlorides olutions. The oxidation currents increase withaclear trend:K Cl < NaCl < LiCl. The different approachc urves must originate from differentH 2 O 2 generation rates caused by the presenceo ft he different solvated alkali metal cations transferred to the organic phase. This sequence of K + < Na + < Li + measured for the different electrolyte solutions at ap articular distance is in excellent agreement with the sequence of their acidity and standard hydration enthalpy D hyd H8. [25] Figure 6 shows quantitativelyt he relationship between the H 2 O 2 oxidation current at the ME at ad istance d ME-MP = 2 mmb etween the ME and the MP (from Figure 5b)a nd D hyd H8 from the literature. [25] The position at which d ME-MP = 0i se vident when ME and MP touch each other and the ME current changes abruptly.T he H 2 O 2 oxidation current is an indicator of the ORR rate in the organic phase. Consequently,t he observationc onfirms the role of the hydration shell of alkali metal ions and their surprising acidity within the mechanism for catalytic reduction reaction of O 2 by DMFc in DCE [Eq. (3)].T he observed linear correlation for valueso fL iCl, NaCl, and KCl allows ap rediction for the rate of ORR when Rb + or Cs + ions are to be transferred. A linear extrapolationy ields an expectation of 28 pA for Rb + and 22 pA for Cs + .
Depending on the hydration shell surrounding the cations, [23] different ORR catalytic modes of action could be expectedf or cations in the organic phase. Cations of high charge density interactm ore strongly with the negative charge centers of water (oxygen atoms) in their hydration shells.T his results in higher acidity and hence stronger facilitation of DMFc-H + formation as the first step in the ORR. However,h ydrophobic and semihydrophobic cations strip the hydration shell upon transfer and are solvated predominately by DCE molecules in the organic phase. [23] Hence, their ORR activity is expectedt obeb elow those of the alkali metal ions.
Interestingly,t he current observed in the presence of HCl in the aqueous phase (Figure 5b,c urve 1) is much smaller than those of the tested alkali metal chloride solutions. Several effects may contribute to this phenomenon. The equilibrium concentration is 0.1 m H 2 Od issolvedi nD CE in aH 2 O j DCE biphasic system. [26] In the case of the HCl aqueous electrolyte, the transferred protons are consumed, leaving behind an unbalanced Cl À excess in the aqueousp hase. In case of MCl, the transfer of M + and subsequent reaction leads to an alkalization of the organic phase. M + can be transferred back to the aqueous phase, restoring the initially stable electrolyte and counteracting the Cl À excess. In such as ituation also [DMFc + ][OH À ] could transfert ot he aqueous phase in the presence of the produced OH À in the organic phase (D w DCE G o;w!DCE DMFc þ = 24.1 kJ mol À1 [6a] ). This may facilitate an ongoingr eaction. However,o wing to the multitudeofpossible transfer processes and associated free energy contributions, it is very difficult with the availabletechniques to completely disentangle the different effects.

Chronoamperometrymeasurements:quantitative detection of hydrogen peroxide
During the recording of an approach curve,t he diffusion layer above the MP may not attain ac omplete steady state but expand slightly duringt he experiment. Am ore defined situation is obtained by chronoamperometry,i nw hicht he ME and MP are kept at af ixed distance d ME-MP and the potential of the MP is changed at the start time t 0 = 0f rom av alue at which no ion transfer occurs,t ot he potential D w o f = À0.65 V, which causes iont ransfer from the aqueous to the organic phase (Figure 4a). At the same time, the ME potential E T is changed from 0Vto + 0.8 V( vs. Ag j AgCl j Cl À )t od etect H 2 O 2 by oxidation over several hundred seconds. Afterwards, the potentials are switched back to stop the reactions. After aw aiting period of 70 st oa llow the diffusion layers to relax, d ME-MP is incremented and the procedure is repeated. The potential jump at the MP defines an exact onset for the ion transfer across the liquid j liquid interface relative to the currentm easurement at the ME. Figure 7a shows the chronoamperometric transientsa t the ME in the four electrolytes and at differentd istances d ME-MP between 2 mmt o1 00 mm. The steady-state currents of each transienta re plotted together in Figure 7b for different d ME-MP . The oxidation current for each alkali chloridee lectrolyte solution increases as d ME-MP decreases. This behavior is in accordance with the approach curvesi nFigure 5b,w hich capture a condition very close to the steady-state current. It also reproduces the sequence of catalytic activity of the three tested alkali metal ions in Figure 5b.I tt hus corroborates the proposed mechanism forc atalytic ORR in the presence of those hydratedc ations and DMFc in the organic DCE phase [Eqs. (3)- Interestingly,t here are qualitative differences in the current transients both at different distances and between different aqueous electrolytes. For LiCl and the smallest distance,t he transients tarts with the highestc urrents and decayst ot he steady-state value.I nt his case, the collection efficiency is very high and the transition time between the onset of the reaction and product detection at the ME is not resolved. The current decays owing to the fast depletion of oxygen in the reaction zone in the DCE phase. The steady-state current is then controlled by the mass transport of O 2 into the reaction zone. At larger distances, the collection efficiency of the ME gradually decreases, thusd ecreasing the steady-state currents. Also, the transients rise more steadily because there is ac onsiderable time delay between the onset of the reaction and the detection at the microelectrode. For NaCl, the same trend is observed with as light modification.F or very short distances, the initially recordedc urrent is lower than forL iCl electrolyte but the transientd ecays slower to the steady-state values. Because Na + (aq.) is less activet han Li + (aq.) ,t he initial reactionr ate is slower and it takes al onger time until the mass transport of O 2 becomes as ignificant limitation. For KCl and HCl, the current transients are always decaying. The ORR reaction rate is slower so that other limitingf actorsare not important.
Please note that the order of reactivity H + < K + < Na + < Li + does not follow the sequence of the onset of IT in Figure 4a. The highestv oltage is required for transfer of Na + followed by Li + ,H + ,and K + .Ifi tw ere just the magnitude of cation transfer to the DCE phase caused by the applied potentialo fD w o f = À0.65 Vd uring the experiments, the sequence of H 2 O 2 production rates should follow the order NaCl < LiCl < HCl < KCl, which is qualitativelyd ifferent from our experimentalo bservation. This adds stronge vidence that the acidity of the hydrated cations in the organic phase is indeed the decisive factor for the observedb ehavior.

Conclusion
The oxygen reduction reaction using DMFc as the electron donor at aw ater j DCE interface represents an example of ioncoupled electron transfer reactions at al iquid j liquid interface. The reactionc an proceed by catalysis of hydrated alkali metal ions transferred from the aqueous to the organicD CE phase. Ion-transfer cyclic voltammetry demonstrates the transfer of H + ,L i + ,N a + ,a nd K + across the water j DCE interfacei ft he externallya pplied Galvani potential difference exceeds the ionspecific value of the transferp otential D w o f.T oc ompare the catalytic activity of different alkali metal ions, the coupling of the micropipette and an amperometricm icroelectrode is demonstrated. The MP mechanically stabilizes the liquid j liquid interface, whereas the ME is used for product detection in the substrate-generation/tip collection mode of SECM. With this combination,t he catalytic action of the transferred hydrated alkali metal ions was compared. The activity increases in the sequence K + < Na + < Li + in strong correlation with literature values for the standard hydration enthalpies of those ions. It demonstrates the catalytic biphasic ORR for H 2 O 2 formation in the presence of differenta lkali metal cations in aqueous electrolyte solutions. Accordingly,t he cationsw ith higher hydration enthalpy can induce ah igher driving force for ORR and, therefore, higher rate for H 2 O 2 generation.I nterestingly,t ransferred hydrated protons do not acceleratet he ORR to the same extent as alkali ions, probably as ar esult of the higher energy of the resulting water cluster in DCE.
Ta ken together,t he demonstrated methodology provides very useful combined informationf or ion and electron transfer reactions, whichi saprerequisite for optimizingb iphasic fluidic systemsw ith respect to the compositiono ft he aqueous and organic phases.T his also will allow us to generalize the chosen example of H 2 O 2 generation by utilization of aw ider range of electron donors and oxidants species with the ultimate goal of carrying out vital ion-coupled electron transferr eactions in energy-relatedo rs ynthetic fluidics ystems in which ah eterogeneousl iquid system facilitates the work-up and regeneration of the solutions. This also refers to organic synthetic protocols "on water", [27] in which the amount of organic solvents can be reduced. This is interesting for the reduction of organic waste but also for the synthesis of water-soluble materials,f or which optimized biphasic systems can be advantageous in terms of improved kinetics, higherselectivity,and higher yields.

Experimental Section
Chemicals NaCl (> 99.99 %, Carl Roth), LiCl (> 9999 %, VWN chemicals), KCl (> 99.99 %, Carl Roth), HCl (35 %, VWN chemicals), DMFc (Sigma-Aldrich), LiTB (Sigma-Aldrich), bis(triphenylphosphoranylidene) ammonium chloride (BACl, Sigma-Aldrich), and DCE (Sigma-Aldrich) were used as received. BATB was prepared by metathesis of BACl and LiTB in am olar ratio of 1:1a nd recrystallized in am ixture of 2:1m ethanol/water before use as the supporting electrolyte in the DCE phase of the biphasic liquid system. [28] Preparation of microelectrode and micropipette The Pt ME was fabricated and polished as described elsewhere. [29] In short, aP tw ire (25 mmd iameter,G oodfellow,C ambridge, UK) was sealed in ab orosilicate capillary (outer diameter (O.D.)/inner diameter (I.D.) = 1.0 mm/0.5 mm, 100 mm length) under vacuum. After connection to aC uw ire by using silver-epoxy glue (EPO-TEK, John P. Kummer GmbH, Augsburg, Germany) and heat treatment inside an oven at 60 8Cf or 10 h, the ME was polished and shaped into ac one by aw heel with 180-grid Carbimet paper disks, then polished sequentially with 0.3 mma nd 0.05 mma lumina powder on am icropolishing cloth (Buehler,L ake Bluff, IL, USA) for 5min, interrupted by rinsing with water after each polishing step. MPs with r MP 50 mmw ere fabricated from quartz capillaries (O.D./I.D, 1.0/ 0.7 mm, Sutter Instrument, Navato, CA, USA) by using al aser puller (P2000, Sutter Instrument) and by adjusting the pulling parameters (heat, filament, velocity,d elay,p ull) to obtain MPs with short tapers. These MPs yield undistorted voltammograms owing to the small iR drop. The inner wall of the pipette was silanized by inserting as mall syringe from the back of the pipette and placing a small drop of trimethylchlorosilane close to the end of the pipette. The solution was removed from the pipette after 30 min by as yringe, and the silanized pipette was allowed to dry in the air over 8h.S ilanizing allows the filling of the pipette with the organic electrolyte solution and the formation of the liquid j liquid interface with the outside aqueous electrolyte at the orifice of the pipette.

Apparatus and procedure
The ITCV and SECM approach curves were obtained by utilizing a home-built instrument [30] operated under SECMx. [31] The SECM setup designed for the study of the liquid j liquid interface at the MP is shown in Figure S1 a( in the Supporting Information). The ME was mounted by means of ac hromatography fitting to the bottom of the SECM cell body made from polytetrafluoroethylene. The ME (bottom, facing up) and MP (top, facing down) were aligned to each other as shown in Figure S1 (in the Supporting Information), by moving the MP in the xÀyÀz directions with ap iezoelectric motor (Scientific Precision Instruments, Oppenheim, Germany,d etailed in section S3 in the Supporting Information). To avoid crashing, the process was monitored with two cameras whose optical axes were at ar ight angle to each and parallel to the surface of the ME. The body of the electrochemical cell had walls made from microscopic slides to allow undisturbed optical observation of the relative position of the MP and ME inside the electrochemical cell in x and y directions. After the initial positioning, the relative x and y positions between the MP and the ME were fixed during the SECM experiments.
The composition of the aqueous reference solution, the aqueous and the organic phases is outlined in Figure 2. The silanized MP was partially filled with 5mm DMFc as an electron donor agent and 5mm BATB as as upporting electrolyte in DCE as solvent. The remaining part of the MP was back-filled by the aqueous 1mm BACl and 10 mm LiCl reference solution. The solution on each side of the liquid junction inside the MP (phases Ia nd II) shared BA + as common ion. The aqueous reference solution contained LiCl for establishing the potential at the Ag j AgCl j Cl À electrode used to polarize the liquid j liquid interface (II/III) under study (WE1). The filled MP was immersed in the aqueous phase containing 100 mm HCl or MCl (M = Li, Na, K). Another Ag j AgCl j Cl À electrode immersed in the outer aqueous solution served as Aux1 j RE1. The reaction under study (ORR) occurs at or close to the liquid j liquid interface (II/III) formed at the opening of the pipette, which is externally polarized by at wo-electrode setup ( Figure 1) and is used for record-ing ITCV.T he ionic current resulting from the transfer of cations from the aqueous to DCE phase (at negative potential) is defined as anegative current.
Simultaneously,athree-electrode configuration was integrated into the setup described above for recording cyclic voltammograms, chronoamperograms, and SECM approach curves for the detection of H 2 O 2 as ap roduct of ORR. AP tM E, aA g j AgCl j Cl À ,a nd aP t wire in the aqueous phase served as WE2, RE2, and Aux2, respectively.T he Pt ME (r T = 12.5 mm) was biased at E T =+0.8 Vv s. Ag j AgCl j Cl À for diffusion-controlled oxidation of H 2 O 2 ( Figure S2 in the Supporting Information). The SECM approach curves were obtained by moving the biased MP (supporting the liquid j liquid interface, D w o f = À0.65 Vf or transfer of alkali metal ions) toward the biased ME and recording simultaneously the H 2 O 2 oxidation current at the ME as af unction of d ME-MP .T he approach curves were repeated at different D w o f applied at the liquid j liquid interface with the ME at E T =+0.8 Vv s. Ag j AgCl j Cl À .F or recording chronoamperograms, the biased Pt ME was kept at fixed distances from the MP (d ME-MP = 2-100 mm) for detection of H 2 O 2 ,w hereas the liquid j liquid interface was biased at D w o f = À0.65 V.