Chapter 1: Nanoelectrochemistry at the liquid/liquid interfaces

Last 15 years have witnessed the successful transition of electrochemistry at the liquid/liquid interface to the “nano era”.^1–3  This development has built upon and also contributed to the electrochemical understanding of charge transfer (CT) reactions at the interface between two immiscible electrolyte solutions (ITIES). In contrast to solid/liquid electrochemistry, CT reactions at the ITIES include both electron transfer (ET) and ion transfer (IT) processes. After the discovery of the polarizable ITIES, rapid progress in liquid/liquid electrochemistry was achieved by adopting methodologies and concepts from other branches of electrochemistry, including microelectrode techniques.^4,5  More recently, the tools and approaches of liquid/liquid electrochemistry contributed to the development of the broader field of nanoelectrochemistry,⁶  including electrocatalysis, electrochemical imaging, and electroanalysis. These applications have been enabled by remarkable progress toward the miniaturization of liquid/liquid interfaces to the nanometre scale.

Nanoscale ITIES and their arrays can be formed by using nanopipets, nanopores, and porous membranes, some of which are created using modern nanofabrication techniques. Both nanoscopic and macroscopic ITIES can serve as a platform for studying the electrochemical behaviours of a variety of nanoscale entities, e.g., nanoparticles and biological macromolecules employed in electrocatalysis and electrochemical sensing. In this chapter, we survey recent progress in electrochemistry at the nanoscale liquid/liquid interfaces in the general context of nanoelectrochemistry.

All CT processes occurring at macroscopic ITIES can also be observed at a nanopipet-supported ITIES, including simple IT,^7–12  facilitated IT,^7,9,13,14  and ET reactions.¹⁵  A simple IT process is a one-step reaction in which an ion Iⁿ⁺ is transferred directly from one phase (e.g., water) to the second phase (e.g., organic):

This process can also involve ion pairs¹⁶  or ion clusters.¹⁷ 

Facilitated IT reactions require a ligand (L^m−) in the second phase (e.g., 1,2-dichloroethane, DCE), which can react with Iⁿ⁺ to form a complex, resulting in the transfer of Iⁿ⁺:

The ET reaction between redox molecules confined to two immiscible liquid phases can be described as:

Nanopipets can be fabricated by pulling borosilicate or quartz capillaries with a laser pipet puller (e.g., P-2000, Sutter Instrument Co.). When choosing the proper capillaries for different experiments, one needs to consider several factors, including the material (quartz or borosilicate) and properties of a specific capillary (thickness of the wall, with or without a filament, single or double barrel).¹⁸  Borosilicate glass has a low melting point and requires HEAT (one of the P-2000 parameters) between 300 to 400, while quartz requires HEAT between 550 and 900. Borosilicate glass is easier to work with because its properties change gradually with temperature, but it is difficult to use for producing ultra-small nanopipets with relatively short taper (which is essential for attaining a relatively small resistance). Quartz is preferred in most cases because it allows one to make very small and not exceedingly long pipets. The shortcoming of quartz is that it is very sensitive to uneven heating, which might result in asymmetrical pipets. In this case, using quartz capillaries with a thicker wall (≥0.5mm) can help.¹⁹ 

To support an ITIES, a nanopipet has to be filled with solution. Capillaries with filaments are preferred in most cases since they help to bring aqueous solution to the end of the nanopipet tip; otherwise it can be very difficult to remove the air and to fill the nanopipet completely. Capillaries without filaments were used to fill pipets with organic solution, which is relatively easy to inject in a glass or quartz pipet, and at the same time the solvent evaporation is slower in the absence of a filament.

The pulling process is controlled by adjusting five pulling parameters in the program, which are HEAT, FILAMENT, VELOCITY, DELAY and PULL. Generally speaking, to obtain smaller tips, one can increase the value of HEAT, VELOCITY or PULL, or decrease the value of FILAMENT or DELAY. To control the length of the taper while maintaining the nanometer-scale size, one can limit the value of VELOCITY and increase PULL at the same time.

The glass roughness after pulling might be an issue in some cases. It has been shown that the roughness of the pipet tip can be reduced by polishing²⁰  or by focused ion beam (FIB) milling.²¹  A potential problem is that the pipet orifice can be contaminated by polishing agent.

When a water-filled pipet is immersed in an organic solution a thin aqueous film forms on its hydrophilic outer wall, making the true area of the liquid/liquid interface much larger than the geometrical area of the pipet orifice.²²  The film formation can be avoided by silanizing the outer pipet wall to render it hydrophobic while keeping the interior wall non-silanized. In most previous publications this was done by dipping the pipet tip into a silanizing agent (chlorotrimethylsilane) while passing a flow of argon through the pipet, which is straightforward for micrometer-sized pipets, but not easy for nanopipets.^7,14  Silanization of smaller pipets must be done cautiously to avoid the formation of a film on the inner wall, which can partially block the pipet orifice and induce solvent penetration into its narrow shaft. A recently developed protocol for silanizing pipets in the vapour phase allows one to avoid oversilanization of relatively small (e.g., ∼10nm radius) pipets.¹¹  However, the possibility of silanizing even smaller (e.g., 1–5nm⁹) pipets is uncertain.

When the pipet is filled with organic solution and immersed in aqueous solution, the inner wall of the pipet needs to be silanized to avoid water getting drawn into the pipet. This can be done by dipping the pipet tip into chlorotrimethylsilane for 5−7 s.^20,23  In this case, both the outer and inner wall of the pipet get silanized, but unlike water, organic solution is not likely to form a layer on the outer wall even though it becomes hydrophobic.²²  A more controlled method for vapour silanization was reported recently.^24,25  The pipets were fixed in a mini-vacuum desiccator, which was first evacuated by the pump, and then the vapour of highly pure N-dimethyltrimethyl silylamine was delivered from the flask to the desiccator, where the pipets were exposed to it for about 15 minutes.

The radius of a nanopipet is too small to be measured by optical microscopy. The most commonly used methods for evaluating the radius are scanning electron microscopy (SEM) and electrochemical techniques–cyclic voltammetry and scanning electrochemical microscopy (SECM). SEM is a direct way of visualizing the geometry of a nanopipet, but it is limited by the resolution of the instrument. For pipets with diameters smaller than ∼50nm, it is difficult to see the orifice clearly since the glass wall is not conductive.

Steady-state voltammetry of IT from the external liquid phase to the filling solution can be used to evaluate the radius of a non-silanized nanopipet from Eq. 1.4a proposed by Beattie et al.²⁶  for the diffusion limiting current

where z, D₂ and c₂ are the charge of the transferred ion, its diffusion coefficient and bulk concentration in the external solution (phase 2). Sometimes the background subtraction is necessary to obtain accurate results.⁹ 

Silanized pipets give more reliable information about the interface since there is no leakage of aqueous solution from the tip, and the equation for the limiting current is:

where x is a function of the dimensionless parameter, RG=r_g/a (r_g is the pipet wall thickness at the tip). x was tabulated²⁷  and expressed by an analytical approximation for disk-shaped interfaces.²⁸ 

If a transferable ion is initially present in the filling solution, the IT current is determined by diffusion inside the pipet and should depend on its geometry, i.e., the orifice radius and the pipet angle at the tip (θ_p; Fig. 1). This dependence can be used to evaluate θ_p from the limiting current of the ion egress based on the following equation^10,11 :

where D₁ and c₁ are the diffusion coefficient and bulk concentration of the transferred ion in the filling solution (phase 1), and f(θ_p) is a function of the tip inner angle, θ_p, as given by^11,29 

SECM has also been used to evaluate the shape of a nanopipet tip (see Section 1.4).^7,15,20  Good fits of SECM negative feedback approach curves between the experimental and theoretical curves can confirm that the ITIES is essentially flat and not recessed, and also provide information about the RG value of the nanopipet. Pipets as small as ∼8nm radius with RG=1.6 have been characterized in this way.²⁰ 

Unlike macroscopic ITIES, in nanopipet voltammetry the interfacial CT current is very small (pA-range). Therefore, potentiostatic experiments at nano-ITIES are performed by applying voltage between two reference electrodes, and a four-electrode potentiostat is not required. Typically, the potential gradient and the ohmic potential drop inside a pipet are too small for significant electromigration or electroosmotic flow along its charged inner wall.^10,11,13  The electrostatic and double layer effects can be more significant for smaller nanopipets, e.g., a≤5nm.⁹ 

Choosing a proper scanning rate in voltammetry is essential for attaining a steady-state and sufficiently low charging current. Typically, the time required to attain a steady-state is determined by the mass transfer rate inside the pipet. If θ_p is not very small (e.g., ≥5°), sigmoidal forward and reverse waves that completely retrace each other can be obtained at moderate scan rates (e.g., v≤1V/s), thereby confirming that IT on either side of the nanopipet tip reaches a steady state. However, very slower diffusivities (e.g., in ionic liquid (IL)¹² ) can result in a significantly longer time required for the IT to reach a steady-state in the external solution. Fig. 2A shows two CVs obtained at a 500nm pipet At ν=1 mV/s, both egress and ingress currents is curve 1 attain a steady-state; however, at ν=1 V/s (curve 2), the egress wave remains essentially sigmoidal, while the ingress wave is peak-shaped. At higher ν, the charging current becomes significant (Fig. 2B). SECM can also be employed to probe CT processes at ITIES as well as for topographic²³  and electrochemical²⁵  imaging (Section 1.4).

Dual pipets (or double-barrel pipets or θ-pipets) based electrochemical generation/collection (G/C) technique was developed as a new tool for studying heterogeneous IT reactions and homogeneous chemical reactions of ionic species in solution.^30,31  This technique allows quantitative separation of different CT processes simultaneously occurring at the liquid/liquid interface (e.g., simple transfer of potassium, facilitated transfer of the same ion with a crown ether, and IT of supporting electrolyte). Another advantage of this technique is the possibility to overcome potential window limitations and study numerous important reactions occurring at high positive or negative potentials (e.g., transfers of alkali metals from water to organic media).

Figure 3 shows the schematic of a dual pipet with a θ–shaped tip. Such a pipet can be made from borosilicate θ-tubing (OD=1.5mm, Sutter Instrument Co.) using a laser puller.³⁰  The pulling procedure is similar to what has been described for conventional pipets. Typically, two barrels of the dual pipet are filled with water.^30,31  If one of the barrels (“generator”) contains a cation, it can be transferred to the outer organic solvent by biasing this pipet at a sufficiently positive potential (E_g). A significant fraction of ejected cations reaches the negatively biased second pipet (“collector”) and gets transferred back into the aqueous phase (Fig. 3B). The collection efficiency, η=i_c/i_g (i_c is the collector current and i_g is the generator current) can be used to investigate CT and chemical reactions occurring in the space between two channels.^30–32  The η value strongly depends on the collection potential and the geometry of the θ-pipet. In the absence of chemical reactions in solution, its maximum value, η_max, is obtained when all ions reaching the opening of the collector pipet are transferred into it. The η_max value depends only on the normalized distance between the centers of two barrels. Simulating mass transfer in such a system is challenging, and no exact theory is currently available for dual pipet–supported ITIES. Three-dimensional finite element simulation was recently reported for the dual carbon electrode based on a θ-nanopipet.³³ 

Shao et al. have demonstrated the possibility of fabricating submicrometer- and nanometer-sized dual pipets, and forming two independent ITIES at the orifice of these devices.³⁴  Figures 3C and 3D show the SEM images of two dual nanopipets. The K⁺ transfer at the W/DCE interface facilitated by DB18C6 was used as the model system to study the effects of geometric parameters of the pipet on collection efficiency. The larger, submicrometer-sized pipets showed higher collection efficiency, while nanometer-sized pipets produced better results for a system without supporting electrolyte.³⁴ 

Laforge et al. reported in 2007 that a nanopipet can also be used as an “electrochemical attosyringe” for controlled fluid delivery.³⁵  The prepared nanopipet was filled with an organic solvent and immersed in an aqueous solution. The ITIES at the pipet orifice was shown to move in response to variations in applied voltage. Water entered the pipets when the potential of the inner organic solution was made negative and was expelled at positive potentials. This phenomenon was used to sample and deliver attoliter-to-picoliter volumes of fluorescent dyes into human breast cells in culture. The injection volumes could be monitored and evaluated by measuring the pipet resistance and/or current vs. potential curves. Compared to other existing microinjectors, this device is inexpensive, easy to fabricate and use; it can be made very small and used repeatedly. Potential applications are in cell biology, nanolithography and microfluidics.

The kinetic parameters of a heterogeneous CT reaction can be determined electrochemically only if the CT rate is lower than or comparable to that of mass transfer.^36,37  Therefore, to investigate fast heterogeneous reactions under steady-state conditions one has to increase the mass-transfer rate by fabricating submicrometer-sized devices, e.g., metal nanoelectrodes. In this way, the effects of the resistive potential drop and double layer charging current can also be diminished. In liquid/liquid electrochemistry, similar advantages can be attained by using a nanopipet-supported ITIES. In earlier experiments,^7,13,14  the rates of simple and facilitated IT were determined from steady-state voltammograms obtained using nanopipets filled with an aqueous solution (the resistance of an organic-filled pipet is usually too high to attain the ohmic potential drop of <1 mV required for reliable kinetic measurements¹³ ). The ion of interest was initially present only in one phase (ether aqueous or organic) and its transfer across the ITIES produced sigmoidal voltammograms, which were used to extract kinetic parameters. In the case of a facilitated IT (Eq. 1.2), an excess amount of the transferred ion is added to the pipet to deplete a ligand in the external solution. The essentially spherical diffusion of a ligand species to the pipet orifice resulted in the true steady-state voltammogram that was analyzed using the equations developed earlier for quasi-reversible steady-state voltammograms at uniformly accessible solid microelectrodes^36–39 

where , , λ=k⁰/m, k⁰ and α are the standard rate constant and the transfer coefficient, respectively, and m is the effective mass-transfer coefficient. The kinetic parameters were extracted either by fitting the entire voltammogram to Eq. 1.7 or by using the three-point method based on the determination of the half-wave potential, E_1/2, and two quartile potentials, E_1/4 and E_3/4.⁴⁰  The same approaches could be used for analysis of steady-state voltammograms of ET obtained at the nano-ITIES.¹⁵ 

In the case of simple IT (Eq. 1.1), the asymmetry of the diffusion field at a pipet-based ITIES, where the diffusion inside a narrow shaft is essentially linear in contrast to the spherical diffusion of ions to the pipet orifice in the external solution, makes the mass transport more complicated.^5,7,29  A cyclic voltammogram of simple IT at a micropipet consists of an apparently steady-state, sigmoidal wave that corresponds to ingress of an ion into the pipet and a time-dependent, peak-shaped wave produced by egress of the same ion to the external solution. Depending on experimental conditions, simple IT at a nanopipet may produce either an asymmetrical transient (i.e., time-dependent) voltammogram⁷  or a sigmoidal and retraceable steady-state curve.^7–9  In earlier studies, such voltammograms were treated using simple steady-state theory (e.g., Eq. 1.7), assuming that their shape is independent of geometry of the pipet inside. More recent simulations^29,41,42  and experiments²⁹  showed that this simplification is not realistic, and the reversible half-wave potential of simple IT from the external solution to the pipet under steady state depends on pipet angle, θ_p. It was suggested that kinetic and thermodynamic parameters of simple ITs determined without taking into account effects of ion diffusion in the inner space of a nanopipet may not be accurate. A more realistic model developed in ref. 29 was used to extract kinetic parameters of a slow IT reaction as well as the formal potential and diffusion coefficients in both liquid phases from a single transient voltammogram obtained at a micropipet-supported ITIES.

Another issue complicating kinetic analysis of rapid CT reactions is a weak dependence of the shape of an almost reversible steady-state voltammogram on kinetic parameters and, consequently, the lack of the unique fit between the theoretical and experimental curves. The possibility to fit the same experimental curve using different combinations of k⁰ and α leads to significant uncertainties in extracted parameter values.¹⁰  This problem was addressed by using common ion voltammetry^10,11  (see Section 1.3.4).

To our knowledge, ET at nanometer-sized ITIES was subject of a single study.¹⁵  Such experiments are challenging because of interfering ion transfer reactions and/or interfacial precipitation. By exploring a number of combinations of aqueous and organic redox couples and different supporting electrolytes, two experimental systems—the reduction of TCNQ by Ru(NH₃)₆²⁺ and the ET between Fe(EDTA)²⁻ and TCNQ—were shown to be suitable for such studies, while all other systems failed to yield high quality voltammograms. Among the latter group was the reduction of TCNQ by Fe(CN)₆⁴⁻ at the water/DCE nano-interface, which has previously been employed in ET studies at macroscopic polarizable ITIES.⁴³  It was found that Fe(CN)₆³⁻ was transferred from water to DCE, and the similar half-wave potentials of ET and IT precluded quantitative studies of ET kinetics in this system.

Steady-state voltammetry was used to investigate ET reactions at the polarizable ITIES formed at the tip of 50-nm to 400-nm radius pipets.¹⁵  The pipet was filled with aqueous solution containing a mixture of two forms of redox species (O₁ and R₁) and immersed in organic solution containing water-insoluble redox species (O₂). The application of a sufficiently negative potential to the internal reference electrode with respect to the external (organic) reference resulted in the electric current across the nano-ITIES due to the interfacial ET between R₁ and O₂ species. The condition c_R1≫c_O2 was maintained in all experiments, so that the diffusion of R₁ species inside the pipet was negligible and did not control the overall current, and the aqueous phase showed a metal-like behaviour.⁴⁴  The kinetic parameters obtained in this way for the reaction

(=2.75 M⁻¹ cm s⁻¹ and α=0.53) were thought to be less reliable because of significant sensitivity of Ru(NH₃)₆²⁺ species to oxygen. The complete removal of oxygen during pipet filling and voltammetric experiments was not feasible, and, thus, it was difficult to ensure that the concentrations of Ru(NH₃)₆²⁺ and Ru(NH₃)₆³⁺ remained unchanged during the entire experiment.

An extensive set of data was obtained for the TCNQ reduction by Fe(EDTA)²⁻ in the following cell:

(TBA⁺ is tetrabutylammonium; TPBCl⁻ is tetrakis[4-chlorophenyl]borate).

High quality steady-state voltammograms obtained for this ET reaction (Fig. 4A, curve 1) were further improved by background subtraction (Fig. 4B) and fitted to theoretical curves calculated either for a microdisk geometry (Fig. 4C) or for a uniformly accessible ITIES (Eq. 1.7). Kinetic parameters were obtained for different concentrations of organic and aqueous redox species and for a wide range of radii (∼50nm to ∼350nm). While the determined α values were close to 0.5 and essentially independent of a and concentrations of redox species, the values were much larger than the rate constants previously measured for any ET at macroscopic polarized interfaces and at micrometer-sized non-polarized ITIES.¹⁵  More surprisingly, the apparent standard rate constant increased markedly with decreasing pipet radius (i.e., from ∼0.4cm/s at a=300nm to ∼1.8cm/s at a=50nm). The authors noted that this behavior is at variance with existing ET theories. They eliminated the possibilities of the recessed interface, incorrectly determined pipet radius, and other artifacts by thoroughly characterizing nanopipets (including SECM experiments with conductive and insulating substrates; see below). However, other factors, including the lack of the unique fit between the theoretical and experimental steady-state voltammograms (see above) and possible double-layer effects, may have affected the determined kinetic parameters. The experimental difficulties mentioned above and the scarcity of the available literature data precluded the comparison of the measured ET rates to those determined for the same ET reactions either at micrometer-sized polarizable or non-polarizable ITIES.

The rates of most IT processes are too fast to be accurately measured at either macroscopic or micrometer-sized interfaces. The first IT kinetics studied at the nanopipet-supported ITIES was that of potassium transfer from the aqueous filling solution to DCE facilitated by dibenzo-18-crown-6 (DB18C6)¹³ 

The mass-transfer rate was sufficiently high to measure the rate constant of potassium transfer under steady-state conditions using pipets with a≤250nm. Assuming uniform accessibility of the ITIES, k⁰ and α values were found by fitting the experimental data to Eq. 1.7. Additionally, the kinetic parameters were evaluated by the three-point method.⁴⁰  A number of voltammograms obtained at 5- to 250-nm pipets yielded k⁰=1.3±0.6cm/s, and α=0.4±0.1, and no apparent correlation was found between the measured rate constant and the pipet size. This k⁰ was significantly higher than rate constant values measured for this reaction at larger interfaces, thus providing the first evidence that the IT rates may be faster than it appeared from earlier experiments. The mass-transfer coefficient for a 10-nm-radius pipet was estimated to be≥10cm/s (assuming D=10⁻⁵ cm²/s) and the corresponding upper limit for the determinable heterogeneous rate constant was given as≥50cm/s.

Yuan and Shao investigated the kinetics of several alkali metal ion transfers facilitated by DB18C6 at the water/DCE nano-interfaces.¹⁴  Their measurements yielded the rate constant for potassium transfer similar to that reported in ref. 13. Well shaped steady-state voltammograms were also obtained for other alkali metal cations, but the kinetic parameters determined for Li⁺, Rb⁺ and Cs⁺ showed significant correlation with the pipet radius. A similar approach was used by the same group to measure the kinetics of alkali metal transfers across the water/DCE interface facilitated with N-(2-tosylamino)-isopentyl-monoaza-15-crown-5.⁴⁵  The association constants were measured for alkali metal complexes in DCE, and the selectivity of this ionophore was shown to follow the sequence Na⁺>Li⁺>K⁺>Rb⁺>Cs⁺. The standard rate constants determined from steady-state voltammograms were similar for all studied cations (∼0.5cm/s) and somewhat lower than those measured with DB18C6 for K⁺ and Na⁺.^13,14 

Two sources of error, which may have affected the accuracy of the results reported in refs. 13–15, were identified in later studies. One of them is the lack of silanization of the outer pipet wall. The formation of a thin aqueous film on the hydrophilic glass surface may have resulted in the true ITIES area significantly larger than that evaluated from the diffusion limiting current (see Section 1.2.2.2). This should result in overestimated values of the mass-transfer coefficient and standard rate constants calculated from the dimensionless parameter λ=k⁰/m. Another source of error—the uncertainty in fitting experimental IT voltammograms to the theory—is discussed below.

Some simple IT reactions are even faster than facilitated processes. The first attempt to measure kinetics of two rapid simple ITs at the water/DCE interface formed at the tip of a nanopipet was reported by Cai et al.⁷  Employing 10nm–300nm-radius pipets, k⁰=2.3cm/s was found from quasi-steady-state voltammograms of the TEA⁺ transfer from DCE to the aqueous filling solution, and a similar value (k⁰=2.1cm/s) was obtained by steady-state voltammetry for the reverse reaction. However, the corresponding transfer coefficients, α=0.70 and β=0.60, were larger than 0.5, and their sum was larger than the theoretically expected value of 1.0. Additionally, a noticeable inverse correlation between the k⁰ and a suggests that the data is not completely reliable. A slightly lower rate constant (k⁰=1.5±0.3cm/s) and α=0.60±0.04 were obtained for the tetramethylammonium transfer.

Jing et al. studied IT kinetics at the nanoscopic water/n-octanol (OC) interface, which is often used as a model system to mimic CT processes through biomembranes.⁸  Although the potential window (∼400 mV) was narrower than that observed with the same supporting electrolytes at the water/DCE interface, it was possible to obtain sufficiently well-defined steady-state voltammograms to determine partition coefficients and standard potentials for the transfers of tetraphenylarsonium, TBA⁺, and laurate from OC to water. This suggests the possibility of a more straightforward approach to investigating the transfers of ionizable drugs through cell membranes. Kinetic parameters were determined for laurate transfer at the water/OC and water/DCE nano-interfaces, and the rate constant measured at the former was about six times lower.

More recently,⁹  very large rate constants were measured for simple transfers of TEA⁺ (110±23cm/s) and ClO₄⁻ (35±8cm/s) and facilitated transfer of K⁺ with DB18C6 (95±31cm/s) from extremely small (1nm≤a≤5nm) water-filled pipets to DCE. However, it was noticed that the reported k⁰ values may have been significantly overestimated because of problems with the data analysis and lack of pipet silanization.¹¹  Additional factors that could have increased the apparent IT rate constant are double layer effects and possible deviations from the conventional theory at ultra-small pipets.

One reason for the inconsistency of kinetic data obtained from steady-state IT and ET voltammograms is that the shape of such a curve depends weakly on kinetic parameters. The possibility to fit the same experimental steady-state voltammogram using different combinations of k⁰ and α leads to significant uncertainties in extracted parameter values.¹⁰  In Fig. 5, very similar steady-state voltammograms were simulated by using different sets of α and λ from wide ranges of 0.20–0.80 and 0.15–7.0, respectively, and slightly adjusting the formal potential, . This result indicates that an experimental voltammogram with a “conventional” α value of 0.5 and a relatively large λ value of 1.0 (red line) can be fitted to the theory using an anomalously large (or small) α value of 0.8 (or 0.2) coupled with an underestimated (or overestimated) λ value of 0.15 (or 7.0). This problem can be addressed by having a transferrable ion present simultaneously in both liquid phases.¹⁰ 

Unlike conventional voltammetric setup, common ion voltammetry requires the initial addition of a transferable ion to both liquid phases, i.e., to the filling solution inside a nanopipet and the external solution. The resulting steady-state IT voltammogram comprises two waves corresponding to the ingress of the common ion into the pipet and its egress into the external solution (Fig. 6). When both processes are at steady state, the corresponding limiting currents can be expressed by Eqs. 1.10 and 1.5

Eq. 1.10 is identical to Eq. 1.4b for the diffusion to the disk-shaped electrode (or ITIES) in the external solution.

An important characteristic point in a common-ion voltammogram is a zero current point, at which the potential (equilibrium potential, Δϕ_eq) is given by the Nernst equation

The formal potential of the common ion transfer () can be determined directly from the Δϕ_eq value instead of being found from the fit of a conventional IT voltammogram to the theory. This is one of the reasons why kinetic measurements by common-ion voltammetry are more reliable.

If the D₁ and D₂ values are known, geometric parameters can be directly obtained from the two limiting currents in the same nanopipet voltammogram produced by ingress and egress transfers of the common ion, using Eqs. 1.5 and 1.10. Then, the kinetic parameters (k⁰ and α) for IT can be determined by fitting the entire common-ion voltammogram to the theory (Eq. 1.12)¹⁰ 

where and are the mass-transfer coefficients of ions in the inner and outer solutions, respectively; and the heterogeneous rate constants, k_f and k_b, are given by the Butler-Volmer-type model

The unique fit of the experimental steady-state voltammogram to the theory can be obtained when both ingress and egress IT waves are quasi-reversible.^10,11  The asymmetry of the diffusion field results in different extents of reversibility (i.e., kinetic vs. diffusion control) of the ion ingress and egress processes, which can be assessed using two dimensionless parameters, λ_ing=k⁰/m_ing and λ_eg=k⁰/m_eg. If the ratio of diffusion coefficients, D₂/D₁ is not very far from the unity, reliable kinetic parameters can be extracted from a common ion voltammogram if both λ_ing and λ_eg are smaller than 10. In this way, the unique combination of the kinetic parameters, α=0.50 and k⁰=6.5cm/s was obtain for the TEA⁺ transfer across the water/DCE interface from the best fit shown in Fig. 6.¹¹  Similar values (k⁰=6.1±0.9cm/s and α=0.49±0.09) were determined with various pipets (9.7nm≤a≤33nm) at different TEA⁺ concentrations. The α value very close to 0.5 (expected for a simple one-step IT process⁴⁶ ) and the independence of kinetic parameters from a were taken as an indication that IT was not complicated by double-layer effects produced either by the ITIES or by the negatively charged wall of a quartz pipet.

The values k⁰ reported in ref. 11 are much higher than those determined previously from conventional nanopipet voltammograms⁷  (k⁰ ∼ 2cm/s). In the latter case, the analysis of a nearly reversible voltammogram with λ_ing (or λ_eg)>1 did not give a unique combination of kinetic and thermodynamic parameters for rapid IT. An additional source of error was the neglected effect of ion diffusion in the internal solution.

The advantages of common ion voltammetry enabled the study of IT reactions at the water/ionic liquid (IL) interface.¹²  Kinetic measurements at such interfaces are challenging because of slow mass-transfer rates in IL. For example, the IL employed in ref. 12, [THTDP⁺][C₄C₄N⁻], is ∼700 times more viscous than water. Slow mass transfer in the IL phase results in a low diffusion current and necessitates the use of small nanopipets and very low potential sweep rates to attain a steady state. Kinetic parameters of the TBA⁺ transfer (k⁰=0.12±0.02cm/s and α=0.50±0.06) were extracted by fitting common ion voltammograms to the theory (Eq. 1.12). The pipet radii were between 21nm and 140nm. While larger pipets could not be used because of mass-transfer limitations, the ingress current measured with smaller pipets was too low to obtain high-quality voltammograms even with a very high concentration of TBA⁺ in IL (up to 200 mM). Because of the large ratio of diffusion coefficients (D₁/D₂=275), the λ_ing values were much larger than the corresponding λ_eg values; and almost all λ_ing values were ≥10. However, unlike water/organic interface, where D₁/D₂ ≈ 1 and λ_ing ≥10 corresponds to an essentially Nernstian IT, the ingress waves at the water/IL interface were quasi-reversible as long as λ_ing≤50.

Several factors that could affect the results of kinetic experiments at the water/IL nanointerface were investigated. Very similar IT rate constants were determined for TBA⁺ and similarly sized but asymmetric C₈mim⁺ ion. This result was taken as an evidence that ionic adsorption is not a major rate-determining factor in the studied system. The comparison of the diffusion currents produced by the egress of cations and anions from the water-filled nanopipets (a ≥ 11nm) to IL showed that the mass transfer inside the pipet shaft is not significantly affected by migration and other electrostatic effects. No correlation was found between the interfacial size and IT kinetics, which would be indicative of double layer effects.

The measured rate constants were more than an order of magnitude lower than those obtained previously for tetraalkylammonium transfers at the DCE/water interface. This difference was attributed to higher viscosity of IL as compared that of DCE. Possible origins of the viscosity effect on k⁰ considered in ref. 17 are lower diffusivities in the interfacial mixed solvent layer, slower formation of the interfacial protrusions and different ion solvation energies in IL.

The advantages of nanopipets—small interfacial area, fast mass-transfer rate and small resistive potential drop—were essential for the studies of IT between aqueous solutions and neat organic solvents. It was found that most metal cations and some strongly hydrated anions (F⁻ and OH⁻) cannot be transferred to purified organic solvents containing essentially no electrolyte and very little water.⁴⁷  In Fig. 7A, neither H⁺ nor Ca²⁺, Mg²⁺ and Fe³⁺ could be transferred to neat DCE at any applied voltage (up to+3 V; the absence of IT at even higher voltages, up to 9 V, was also reported⁴⁷ ). Conspicuously, Cl⁻ (and other moderately hydrophilic anions and cations) could easily be transferred under the same experimental conditions.^47,48  The addition of very low concentration of hydrophobic supporting electrolyte to DCE (e.g., <1 nM) greatly facilitated transfers of hydrophilic ions (Fig. 7B). A somewhat similar effect of water content in organic phase on IT was reported: no transfer of Li⁺ and other strongly hydrophilic ions to triply distilled DCE was observed, but adding small amounts of water (e.g., ∼100 µM) made such ITs possible (Fig. 7C).⁴⁹ 

The role played by an organic counter-ion in transfers of hydrophilic ions contradicts the generally accepted notion of those IT reactions as unassisted, one-step processes. It was suggested that unlike simple transfers of relatively hydrophobic ions, IT of strongly hydrophilic species has to be facilitated. The shuttling mechanism (Fig. 8) was proposed to describe such reactions.⁴⁸  Based on experimental results and molecular dynamics simulations,^50,51  aqueous and organic phases in Fig. 8 are separated by ∼1nm-thick mixed-solvent layer. An ion pair (“–+”) formed by a hydrophilic cation (“+”) and a hydrophobic organic anion (“–”) at the outer boundary of the aqueous phase diffuses across the mixed-solvent layer toward the organic phase and dissociates. (The process in Fig. 8 is a cation transfer, but the extension of this model to anion transfer is straightforward). The released cation is driven into the bulk of the organic phase by the electric field, while the anion travels (via diffusion/migration) across the mixed solvent layer and assists the transfer of the next cation. Besides the interfacial voltage, which carries the cation and the anion in opposite directions, the shuttling process is driven by the gradient of the cation concentration across the mixed layer, which is high (e.g., ∼0.1 M) on the aqueous side of the interface and low on the organic side. The shuttling mechanism explains how a miniscule amount of a hydrophobic counterion can produce measurable IT current.

The increase in IT current with the addition of organic electrolyte was shown to be unrelated to conductivity changes. The amount of hydrophobic salt added to DCE was much smaller than the effective concentration of ionic impurities initially present in the purified solvent. Moreover, the addition of a much larger amount of less hydrophobic counterion to distilled DCE (e.g., ClO₄⁻) did not result in any facilitation of cation transfer.

Similarly, the effect of adding water to distilled DCE on IT was initially thought to be related to organic phase conductivity. However, it was found that the DCE conductance decreases with increasing concentration of water in it.⁴⁹  The conductivity of water-saturated DCE was five times lower than that of essentially dry DCE. The diminished effective concentration of charges was attributed to the formation of water clusters into which the ionic species are extracted. The existence of such clusters in organic solvents and their role in solvation of ions were revealed by NMR studies⁵²  and molecular dynamics simulations.^16,17  It was concluded that strongly hydrophilic ions are transferred to water clusters dispersed in DCE rather than to the bulk organic solvent. This model explains why the transfers of hydrophilic metal ions and protons to water-saturated DCE occur at modest interfacial voltages. Hydrophobic ions undergo simple transfers from water to the organic solvents. Such reactions do not require the presence of organic electrolyte, and they are essentially unaffected by concentration of water in organic solvent.

It is interesting to notice that the studies of IT to neat DCE could not be carried out at a macroscopic liquid/liquid interface. One problem is a high ohmic potential drop in the neat organic solvent. The partitioning of water molecules to organic phase would also impair the study of IT to neat DCE. Although at the nano-ITIES water molecules also egressed from the pipet to DCE, they diffused rapidly from the interface into bulk solvent and therefore could not induce IT processes.

Synergy between SECM and electrochemistry at the ITIES has been explored for a couple of decades.^53,54  Originally, SECM was introduced as a powerful electrochemical tool to investigate CT dynamics at the ITIES using a metal ultramicroelectrode as an SECM tip. More recently, nanopipet-supported ITIES have found applications as SECM nanotips to enable SECM measurements that are impossible or have not been achieved using a metal tip. The knowledge and experience thus obtained through the applications of nanopipet-supported ITIES tips is vital also for the development of the SECM approach based on metal nanotips. Moreover, various nanosystems have been characterized by SECM at the micrometer scale using ITIES as a tip or a substrate.

An ITIES tip can probe both ET and IT reactions to offer more operation modes of SECM in comparison to a redox-sensitive metal tip. The conventional feedback mode of SECM utilizes an ET reaction at the tip and can be realized also using the pipet filled with the solution of a redox-active species (Fig. 9a).^15,55  In the ET mode, the ITIES supported by a pipet tip is externally biased to drive the ET reaction between the original inner species, O₁, and a species in the external solution, R₂, to generate species R₁ and O₂ in the respective phases. The tip-generated external species, O₂, is effectively reduced at a conductive substrate when the tip is positioned within a few tip radii from the substrate, i.e., feedback distance. Subsequently, the substrate-generated species, R₂, is oxidized again by the original species O₁ at the ITIES tip. When species O₁ is in excess with respect to R₂, the tip current depends on the diffusion of O₂ and R₂ across the tip–substrate gap and is enhanced with a narrower gap to yield a positive feedback response. On the other hand, the species R₂ is not regenerated at an insulating substrate and is originally present in the external solution. Thus, the tip current based on the oxidation of the external species R₂ decreases toward zero as the tip approaches the inert substrate. This negative feedback response is obtained because the substrate hinders the diffusional access of species R₂ to the tip.

The operation mode of SECM was also developed for an ingress IT reaction at the ITIES tip.⁵⁶  In the ingress IT mode, the tip reaction can be simple IT (Fig. 9b) or facilitated IT, where the excess amount of an ionophore must be present in the inner solution to deplete the target ion near the tip. This operation mode is simple and requires only a small amount of ionophores, which are usually costly or must be synthesized. The tip current can be enhanced when another ITIES is positioned under the tip to serve as the substrate that provides a target ion from the bottom solution. This operation mode is equivalent to the SECM-induced transfer mode at the ITIES⁵⁷  when the transfer of the target ion from the bottom phase is passively induced by the depletion of the target ion at the tip.⁵⁶  Alternatively, the target ion can be supplied from the bottom phase by externally controlling the phase boundary potential of the adjacent ITIES (see below). In contrast, a negative feedback response is obtained with an inert substrate and has been used for the topography imaging of the substrate.

Another IT mode is based on egress IT transfer, where a target ion is present in the inner solution and is transferred into the external solution with (or without) an ionophore (Fig. 9c).^7,58–60  In the egress IT mode, the ion–ionophore complex is formed at the tip and then dissociates at the substrate. A free ionophore is regenerated at the substrate to participate in the tip reaction again when the inner target ion is present in excess. In this case, the diffusion of the free ionophore and ion–ionophore complex between the tip and the substrate controls the tip current, thereby yielding a positive feedback response. The dissociation of the ion–ionophore complex may be driven by the transfer of the target ion across the ITIES as a substrate or by the reduction of a target heavy metal ion at a conductive solid substrate for metal deposition. An inert substrate does not regenerate a free ionophore, which results in the negative feedback response controlled by the diffusion of the free ionophore from the external bulk solution to the tip.

Recently, another operation mode was developed by coupling ET at the substrate with IT at the tip (Fig. 9d).²⁵  In the ET–IT mode, the ITIES tip is filled with the solution of a redox-active neutral species, R, e.g., ferrocenedimethanol, which partitions into the adjacent external solution to initiate an ET reaction at the conductive substrate. Then, the substrate-generated ion, O⁺, is amperometrically detected at the tip. Interestingly, the dependence of the tip current on the tip–substrate distance in this operation mode is very different from that in the other operation modes. The ET–IT mode gives no tip current response in the bulk external solution, which does not originally contain the target ion, O⁺. As the tip approaches a conductive substrate, the tip current increases from zero toward the maximum current that is limited by the steady-state diffusion of the “tip-generated” neutral species, R, through the tapered region of a nanopipet. Moreover, zero tip current is expected at any distance from an insulating substrate. These unique tip current responses were quantitatively studied using the finite element method and confirmed experimentally.

The nanopipet- and micropipet-supported ITIES have been successfully used as SECM tips to demonstrate several advantages in comparison to metal tips. An ITIES tip can probe both ET and IT reactions while a metallic tip is sensitive only to the former. The unique ion sensitivity of an ITIES tip has found a broad range of applications for SECM. Moreover, pipet-supported ITIES tips are readily fabricated and also miniaturized to a radius, a, of down to <10nm. Such a nanotip dramatically improves the spatial resolution of SECM, which depends on the tip size and the tip–substrate distance, d. The sharp ITIES tip surrounded by the thin wall of a glass pipet possesses a small outer diameter, r_g, of ∼1.5a and can approach very close to a substrate to further improve the spatial resolution. Additionally, mass transfer across a narrower tip–substrate gap is enhanced to enable the kinetic study of a faster reaction. In contrast to the success of sharp nanopipet tips, the multiple-step fabrication of a sharp metal tip has been successful only down to submicrometer size.⁶¹  In addition, a metal nanoelectrode is readily damaged electrostatically or electrochemically without appropriate protections to be recessed and, subsequently, to give a low current response.⁶²  A nanopipet-supported ITIES is somehow more robust or less amenable to such damage probably because ITIES is soft and renewable from bulk solutions.

The relatively rough tip of a heat-pulled glass pipet can be smoothened by mechanical polishing to achieve a shorter separation between the tip and the substrate,²⁰  which is pivotal for improving the spatial and time resolutions of SECM. With appropriate care, the tip of a borosilicate nanopipet can be polished using a micropipet beveller without breaking the tip or plugging the orifice with polishing material. SEM images clearly showed that a polished nanotip is smoother than an unpolished nanotip (Fig. 10a and b, respectively). A polished nanopipet tip with the radius of 8.1nm was positioned at 0.8nm from a glass slide as determined from a plot of tip current versus tip–substrate separation, i.e., an approach curve (Fig. 10c). The short tip–substrate distance also confirms that the nanoscale ITIES is flat and is flush with the surrounding nanopipet orifice. In fact, the feedback mode of SECM is essential for the in-situ characterization of the geometry and size of a nanopipet-supported ITIES tip, which is not compatible with electron microscopy in vacuum. For this purpose, the feedback effect on sharp SECM tips with small r_g/a values has been described quantitatively by empirical equations for a few⁵⁶  or any^63,64  values of r_g/a.

Alternatively, the tip of a heat-pulled pipet can be smoothened by focused ion beam (FIB) milling.⁶⁵  An FIB milled pipet tip is advantageous because the smoothened tip can approach closer to a substrate^65,66  and also gives more reproducible voltammograms.⁶⁷  Modern FIB instruments are equipped with SEM to allow for the dual-beam imaging of the milled tip from different angles. The size of each milled tip can be determined from FIB and SEM images to check the tip size determined electrochemically, most reliably, by SECM. ITIES tips based on FIB-milled pipets with 1–5 µm diameters were employed to precisely determine the high ion permeability of artificial⁶⁵  and biological⁶⁶  nanopore membranes. The FIB milling of a smaller tip is possible²¹  and will be ultimately limited by a resolution of ∼10nm or by the charging of the insulating glass surface without a conductive coating.

Nanopipet-supported ITIES tips were employed to probe the kinetics of fast CT reactions at the ITIES. For instance, fast ET reactions at the ITIES tip was studied under the positive feedback condition of the ET mode (Fig. 9a).¹⁵  Specifically, a 206nm-radius pipet was filled with the aqueous solution of Fe(EDTA)^2– and Fe(EDTA)⁻ and was immersed in the DCE solution of 7,7,8,8-tetracyanoquinodimethane (TCNQ). The tip reaction is given by

With an excess amount of Fe(EDTA)^2–, the tip current was controlled by the diffusion of TCNQ. Therefore, a positive feedback effect was observed as the tip approached close to a Au substrate (Fig. 11a), where TCNQ^•– was oxidized at the diffusion-limited rate to regenerate TCNQ. The high feedback current at short distances confirms the flat geometry of the nanopipet-supported ITIES tip. Tip voltammograms were obtained in the bulk solution and also at short tip–substrate distances of 103 and 41.2nm to enhance the mass transport of the TCNQ/TCNQ^•– couple across the nanogap (Fig. 11b). The quasi-reversible voltammograms were analyzed to yield nearly distance-independent kinetic parameters for this fast ET reaction with high k⁰ values of 0.86–1.26cm/s and normal transfer coefficients, α, of 0.42–0.56. These k⁰ values thus obtained at the nano-ITIES are free from a mass-transfer limit and are much higher than the values previously measured at macroscopic and micrometer-sized ITIES under lower mass transport conditions.

The fast kinetics of a facilitated IT reaction at the macroscopic DCE/water interface was studied using nanopipet-supported ITIES tips in the egress IT mode (Fig. 9c).^59,60  Elegantly, this challenging measurement was enabled by supporting the macroscopic ITIES on a Ag/AgCl electrode for external bias and mechanical stability.⁵⁹  Both tip and substrate reactions are given by Eq. 1.9. The tip reaction was amperometrically driven to the diffusion limit of the forward reaction. In contrast, the potential of the macroscopic ITIES, E_S, was externally changed around a formal potential, , for each approach curve (Fig. 12) to accelerate either forward or reverse reaction. The most positive approach curves at –0.15 V<E_S – <–0.1 V were controlled by the diffusion-limited dissociation of the ion–ionophore complex at the macroscopic ITIES. As E_S was changed toward , slower dissociation resulted in a less positive approach curve. At E_S > , the formation of K⁺–DB18C6 complexes at the macroscopic ITIES became faster than their dissociation to further decrease the positive feedback effect. Eventually, negative approach curves were obtained at E_S≫ owing to the shielding⁶⁸  of the tip from free ionophores, which were consumed as K⁺ complexes near the whole macroscopic ITIES. This time-dependent condition, however, gave apparently steady-state approach curves because the concentration of the free ionophore was nearly uniform within the short feedback distances of <500nm from the macroscopic ITIES.⁶⁹  In fact, the apparently quasi-steady-state approach curves fitted well with theoretical steady-state curves to yield very high CT rate constants, k_f, of 0.3–1.9cm for the forward reaction. These rate constants are not affected by high mass transfer across the nanometer tip–substrate gap. Moreover, a plot of rate constant versus interfacial potential of the macroscopic ITIES is linear around the formal potential, thereby yielding k⁰ values of 0.7±0.3cm/s and α values of 0.56±0.08. These kinetic parameters agree with those determined by nanopipet voltammetry.¹⁴  Nevertheless, some errors may be involved in these parameters because the invalid assumption of an irreversible IT reaction around was made for the analysis of the approach curves and the k_f versus E_s plot.

The first application of a nanopipet-supported ITIES tip for high-resolution SECM imaging was recently reported.²³  In this work, substrate topography was imaged by using the negative feedback effect of the ingress IT mode (Fig. 9b). Specifically, the 103nm-radius pipet filled with a DCE solution was scanned at a constant height over the IBM wafer built with the 90nm process technology. During the tip scan, a tip current response to TEA⁺ was affected by the distance-dependent negative feedback effect from the inert substrate. Approximately 10nm-height topographic features of the substrate were clearly resolved in the resultant image based on the plot of tip current versus tip position. Importantly, the sharp nanopipet tip was scanned over the non-flat surface features without crashing the tip, which is highly challenging in SECM imaging with a nanotip. High-resolution imaging with a nanotip in the feedback-mode requires tip–substrate distances within the nanoscale tip diameter while the contact of the tip with the substrate must be avoided to protect the fragile nanotip and the substrate from damage.

More recently, a nanopipet-supported ITIES tip was employed for simultaneous topography and reactivity imaging.²⁵  To enable this dual-mode imaging, a 270nm-radius pipet was filled with the DCE solution of ferrocenedimethanol as a redox-active probe for reactivity imaging and was immersed in the aqueous solution of PF₆⁻ as a redox-inert probe ion for topography imaging. While the 25 µm-diameter Pt disk substrate embedded in glass sheath was imaged, the perpendicular tip position was dynamically adjusted so that the negative feedback tip current based on the transfer of PF₆⁻ was maintained constant. Subsequently, the tip–substrate distance was also maintained constant, thereby yielding the topographic image of the substrate as the plot of z coordinate of the tip versus lateral tip position (Fig. 13a). The well-polished substrate appears flat at the nanoscale and does not show any topographic features at the Pt/glass boundary. An ∼100nm variation in the height between the lower left and upper right corners of the image (∼20 µm distance) represents the minor tilt of the substrate due to the imperfect tip/substrate alignment. During the topography imaging, the current at the substrate was also monitored to obtain a reactivity image (Fig. 13b). A higher substrate current results from the oxidation of ferrocenedimethanol released from the pipet tip when the tip was positioned over the Pt disk. Since the tip–substrate distance was maintained constant, the reactivity image is not affected by the topography or tilt of the substrate surface. In fact, the reactivity image details the geometry of the Pt/glass boundary including small protrusions of Pt into glass as also found in the optical micrograph (the inset of Fig. 13b).

In comparison to the aforementioned examples, an order of magnitude smaller ITIES tip was used to quantitatively image the single nanopores of a porous nanocrytstaline silicon (pnc-Si) membrane.⁷⁰  This nanoporous membrane (Fig. 14a) was imaged in the constant height mode by employing a 17nm-radius nanopipet tip (Fig. 14b). The nanopipet tip was scanned at a distance of 1.3nm from the impermeable region of the membrane, which resulted in a negative feedback response to TBA⁺ in the ingress IT mode. In comparison, a higher tip current was obtained when a tip was positioned over a nanopore as a source of TBA⁺. Overall, 13 pores were successfully resolved in the 280nm×500nm image to yield a high density of 93 nanopores/μm². The resolution of this nanopipet approach was limited by the tip size, which is comparable to the pore size. The effect of tip size on the SECM image was corrected by employing the finite element simulation to determine the real dimensions of the apparently elliptical nanopore, 7, i.e., major and minor axes of 53 and 41nm and a depth of 30nm. The density and dimensions of the nanopore thus determined from the SECM image (Fig. 14b) are consistent with the respective average values determined by TEM of the pnc-Si membrane in vacuum (Fig. 14a). Significantly, this good agreement confirms that nearly all pores were filled with water to mediate ion transport. This was achieved by wetting the nanopore membrane with isopropanol before its immersion into the aqueous electrolyte solution. Without the isopropanol treatment, nanopores were filled with air bubbles.⁷¹  The ability of SECM to assess single pore permeability in liquid is important because pnc-Si membranes found applications in liquid environments for cell culture, filtration, separation, etc.⁷²  The further development of high-resolution SECM imaging will enable imaging the intrinsic structural and transport properties of single biological nanopores in water without chemical fixation or physical contact.

The unprecedentedly high spatial resolution of SECM in the constant-height imaging of single silicon nanopores was achievable because the tip position was extremely stabilized by eliminating its thermal drift using a newly developed isothermal chamber.²⁴  Air temperature in the chamber changes only at ∼0.2mK/min to remarkably and reproducibly slow down the drift of tip–substrate distance to ∼0.4nm/min, which is measurable using a nanopipet-supported ITIES tip. Without the chamber, a thermal drift of 5–150nm/min under laboratory ambient conditions significantly affect the feedback current of an SECM tip with a diameter of <∼1 µm. The thermal drift is likely due to the expansion or contraction of the SECM stage upon slight temperature change and is rarely noticeable using micrometer-sized tips. It is well recognized for other types of high-resolution scanning probe techniques that the nanoscale thermal drift of the perpendicular and lateral position of a probe causes vertical and lateral image distortions even when the probe–substrate distance is feedback-controlled. The thermal chamber can solve these problems by reducing thermal drift, which was unavoidable and only correctable as practiced for atomic force microscopy unless a cryostat or fast scanning was employed.

The ion-selective permeability of a pnc-Si membrane through multiple nanopores was also determined by employing micropipet-supported ITIES tips.⁶⁵  The pipet tips were smoothened by FIB milling to obtain high-quality SECM approach curves for 6 small ions (quaternary ammoniums and ClO₄⁻) at the pnc-Si membrane in the ingress IT mode (Fig. 9b). The approach curves were analysed by the finite element method to determine membrane permeability, k, which was proportional to the aqueous diffusion coefficient of these ions, D_w. This linear relationship quantitatively agrees with the permeability of a nanoporous membrane as expected from effective medium theories when the ions freely diffuse through water-filled nanopores, i.e.,

with

where a pore length, l, of 16nm, an average pore radius, r, of 5.6nm, a pore density, N, of 67 pores/µm², a porosity, σ (=πNr²), of 0.0079, and f(σ)=1.01 were determined from the TEM image of the nanoporous membrane (note that the nanopores of this pnc-Si membrane are shorter, smaller, and denser than those shown in Fig. 14a). In contrast, membrane permeablity was lower than expected from Eq. 1.15 when it was measured for polyions, i.e., a synthetic polyanionic pentasaccharide, Arixtra (1.5 kDa and charges of –10), and a polycationic protein, protamines (4.1 kDa and charges of +20 or +21). In fact, the permeability to Arixtra further decreased at a lower ionic strength, where a thicker electrical diffuse double layer was developed at the negatively charged wall of the SiO₂-covered nanopores to exert stronger electrostatic repulsion on Arixtra. The lower permeability to protamines independent of ionic strength is due to their large Stokes radius of 2.0nm. Noticeably, this SECM method was applied also to measure the high ion permeability of the nuclear pore complexes at the nucleus of the Xenopus laevis oocyte using pipet-supported ITIES tips⁶⁶  as well as Pt tips.^73,74 

The redox activity of the monolayer-protected Au nanoclusters at the ITIES was studied by SECM using either organic nanoclusters⁷⁵  or the aqueous redox species⁷⁶  (Fig. 15a) as a mediator. In the latter study, the quantitative approach curves based on the feedback effect were obtained for the oxidation of Au₃₈(PhC₂S)₂₄⁺ by IrCl₆²⁻ as generated at the tip (Fig. 15b). The finite element analysis of the approach curves gave an extremely high rate constant of 76 M⁻¹ cm s⁻¹ for this irreversible bimolecular ET reaction with a high driving force of 0.315 V. Interestingly, this rate constant is as high as a half of the ET rate constant between decamethylferrocene and IrCl₆²⁻ with an even higher driving force of 0.68 V. The high redox activity of the Au nanocluster at the ITIES was ascribed to its large size, as predicted by Marcus theory.

The powerful combination of SECM with ITIES electrochemistry has enabled the spatially controlled deposition of metal particles, which is potentially extendable to nanoparticle (NP) deposition by using a nanotip. For instance, Ag particles were locally electrodeposited on a conductive substrate by employing a micropipet-supported ITIES tip in the egress IT mode (Fig. 9c).⁷⁷  The spatial resolution of the tip-induced electrodeposition is controlled by the tip size and the tip–substrate distance. A shorter distance can be maintained by monitoring a shear force between the micropipet tip and the substrate to improve the spatial resolution.⁷⁸  An even higher spatial resolution can be achievable by the shear-force-based control of a submicrometer-sized ITIES tip.⁷⁹ 

SECM was employed not only to induce the deposition of Ag particles at the ITIES but also to monitor their nucleation and growth dynamics at the nanoscale.⁸⁰  In this approach, a 25 µm-diameter Ag tip was oxidized to generate Ag⁺, which was reduced by decamethylferrocene at the DCE/water interface (Fig. 16a). The tip current based on Ag oxidation depends on the rate of Ag⁺ reduction at the ITIES, thereby enabling the kinetic study of Ag deposition. The phase boundary potential of the macroscopic ITIES was controlled by changing the aqueous and organic concentrations of a common ion, ClO₄⁻, to enable the modulation of the driving force. The numerical simulation of the chronoamperometirc tip current revealed that the high rate of Ag⁺ reduction (5–10cm/s) at the ITIES weakly depends on the Galvani potential difference across the ITIES. Moreover, the local deposition of Ag particles at the interface under the SECM tip was visualized by confocal microscopy (Fig. 16b). The deposition of much smaller Ag particles will be possible using a newly developed Ag nanoelectrode as an SECM tip.⁸¹ 

Electrochemistry at nano-ITIES arrays (or ensembles depending on their periodicity) was pioneered by Dryfe et al.^82,83  as summarized in an excellent review article.⁸⁴  Commercially available γ-alumina and track-etched polymer membranes were employed as nanoporous templates to support a nanometer-sized ITIES at each nanopore. Advantageously, these nanomaterials possess pore radii as low as 10nm and have been well characterized for various technological and nanoscience applications. IT reactions at these nano-ITIES arrays were studied by voltammetry under static and hydrodynamic conditions. Interestingly, these arrays can also serve as templates for the deposition of metal NPs. More recently, solid-state nanopore membranes were designed and nanofabricated to support nano-ITIES arrays for electrochemical sensing.⁸⁵ 

The transfer of a probe ion across nano-ITIES arrays has been characterized voltammetrically. For instance, static IT voltammetry at the nano-ITIES array templated by a γ-alumina membrane was used to determine membrane porosity.⁸⁶  In this work, a nanoscale interface was formed at the orifice of each nanopore filled with the aqueous solution of a probe ion (tetraethylammonium, TEA⁺) in contact with the external organic solution (Fig. 17a). Potential sweep rates were chosen such that the mass transport of the probe ion during the forward potential sweep was controlled by the linear diffusion confined within nanopores (Fig. 17b). The resultant peak current based on TEA⁺ transfer depends on the total area of the nanoscale interfaces, thereby yielding membrane porosity from the total area of the membrane exposed to the external solution. At slower scan rates, the voltammetric peak on the forward scan becomes less prominent and eventually plateau. This apparently steady-state behaviour is due to the radial diffusion of the probe ion at the mouth of each nanopore (Fig. 17c). Furthermore, the overlapping of the radial diffusion fields at even slower scan rates will develop a macroscopic linear diffusion field into the bulk inner solution (Fig. 17d). The peak-shaped response due to this linear diffusion process, however, was not observed experimentally even at the slowest scan rate employed. On the other hand, the voltammetric response on the reverse scan was peak-shaped independent of the scan rate because diffusion fields at the external-solution side of the nanoscale interfaces quickly overlap with each other without hindrance from the wall of a pore.

Hydrodynamic voltammetry was enabled using the nano-ITIES array stabilized by either γ-alumina or track-etched polymer membrane. A convective flow was induced by stirring the external solution using a rotating baffle (rotating baffle cell),^87,88  rotating the nanoporous membrane attached to the end of a glass tube (rotating disk cell),⁸⁷  or employing a channel flow cell.⁸⁹  The convection effect was seen as a sigmoidal steady-state voltammogram and was exerted on ion transport at either one side or both sides of the interface depending on the cell type (Fig. 18).⁸⁷  Hydrodynamic voltammetry at nano-ITIES arrays was applied for the kinetic study of fast IT reactions to take advantage of enhanced mass transfer conditions. The rate of mass transfer to the whole nano-ITIES array is enhanced several times when the fastest rotation speed of ∼300 rad/s (∼3000 rpm) is employed in the rotating baffle cell.⁸⁸  The mass-transfer rate to each nano-ITIES is even higher although it is inversely proportional only to membrane porosity (∼0.20) rather than to the size of each interface in contrast to a single nano-ITIES system. The enhanced mass transfer resulted in nearly Nernstian steady-state voltammograms for the fast simple transfer of TEA⁺ at the arrays of DCE/water nanointerfaces to yield standard CT rate constants, k⁰, of 0.82–3.92cm/s.⁸⁷  This hydrodynamic approach was also employed to determine k⁰ values of 0.3±0.2cm/s for the fast Na⁺ transfer facilitated by dibenzo-18-crown-6.⁸⁸  These high rate constants, however, are still lower than those determined by nanopipet voltammetry under even higher mass transport conditions^11,14  and are likely limited by mass transport.

IT voltammetry was also employed to characterize nano-ITIES arrays for sensing applications.^90–92  To support the arrays, solid-state nanopore membranes were designed and nanofabricated. Specifically, a 100nm-thick Si₃N₄ membrane was deposited on the <100> silicon substrate by low-pressure chemical vapour deposition, patterned with a nanopore array by electron-beam lithography, and perforated by magnetic zero-resonant incubation.⁹⁰  Fig. 19 shows the SEM images of 5 µm×5 µm arrays of 1, 95, or 390 nanopores with a uniform radius, r_a, of 25, 115, 125, or 225nm and a wide pore–pore separation, r_c, of 5 or 20r_a. The nano-ITIES arrays supported by the Si₃N₄ nanopore membranes give nearly steady-state IT voltammograms. Steady-state mass transport is achieved between the nanopore-supported ITIES and the bulk solution not only because radial diffusion fields at the widely separated nanopores barely overlap with each other to be confined over a small 5 µm×5 µm array. The contribution of each diffusion mode to the steady-state behaviour will be assessable by spatially and quantitatively resolving a diffusion field at each nanopore, which may be possible by high-resolution SECM imaging (see above). Noticeably, the linear diffusion of a probe ion through 100nm-long nanopores is efficient enough to reach a steady state in contrast to the micrometer-long nanopores in γ-alumina or track-etched polymer membranes.

An interesting application of nano-ITIES arrays is the templated deposition of metal NPs. A heterogeneous ET reaction between an aqueous metal ion and an organic reducing agent (Aⁿ⁺ and D, respectively, in Fig. 20) across the ITIES results in the deposition of metal particles at the interface.⁹³  This process can be carried out using a macroscopic ITIES, a nanoscopic ITIES, and its array. In this application, a nanopore membrane confines the growth of metal particles within a nanopore, thereby yielding NPs. By this way, Pd NPs were deposited using butylferrocene (BuFc) in the DCE phase as a reducing agent⁹⁴ 

This reaction is thermodynamically spontaneous because the standard potential of the Pd couple is more positive than that of the ferrocene couple. The equilibrium position and rate of the heterogeneous ET reaction at the ITIES, however, depends on the phase boundary potential, which can be controlled externally for electrodeposition^94,95  or by a common ion in the aqueous and organic phases for electroless deposition.⁹⁶  Similarly, Pt NPs were deposited at nano-ITIES arrays.^95,97  A more detailed discussion of particle deposition at the ITIES is given in the next section.

The formation of NPs has been intensely studied because of their broad applications in many areas such as catalysis and electronics. ITIES is potentially an ideal medium for depositing NPs because it can provide a clean environment without defects. The most common way of forming NPs at the ITIES is via the reduction of metal ion in one phase with a reducing agent present in the second phase. In 1996, Cheng and Schiffrin reported the electrochemical deposition of gold particles at water/DCE interface from tetraoctylammonium tetrachloroaurate by using potassium hexacyanoferrate(II) as the reducing agent.⁹⁸  Later, the Schiffrin group has investigated the nucleation at the liquid/liquid interface by cyclic voltammetry and galvanostatic experiments.⁹⁹  Selvakannan et al. reported a one-step synthesis of gold nanoparticles by vigorous stirring of a mixture of chloroform containing hexadecylaniline and aqueous chloroauric acid.¹⁰⁰  The size of gold nanoparticles was controlled by varying the molar ratio of the reducing agent and metal ion.

Metal NPs combined with conducting polymers were also produced. Johans et al. synthesized polyphenylpyrrole coated silver particles based on the E_rC_i mechanism, in which the reversible facilitated IT of Ag⁺ to the organic phase was followed by its slow irreversible reduction by phenylpyrrole.¹⁰¹  The process was studied by cyclic voltammetry and UV–VIS.

To better control the nucleation process, Unwin and co-workers deposited Ag particles on the micropipet- and nanopipet-supported DCE/water interfaces. The number of nuclei was evaluated by analyzing current–time transients, which were shown to be significantly affected by the size of the pipet. Single particles were generated with 0.5 µm-radius or smaller pipets while multi-particle nucleation was observed with larger pipets.¹⁰² 

The kinetics of NP deposition at the ITIES has been studied by the Samec group.¹⁰³  They investigated the reproducibility of the potential-step current transients measured under the same experimental conditions for the deposition of the Pt particles at the ITIES and found that the initial rate of the Pt deposition can vary within a broad (over two orders of magnitude) range of values and even approach zero. These findings reflected the random rate of the formation of nuclei with a critical size that is required for a stable growth to occur. Later on, Dryfe et al. noticed that the random nature of the deposition process may be related to the presence of contaminants and found that no solid phase formation could be observed in a perfectly clean environment without defects.¹⁰⁴  In these experiments, the deposition only occurred after adding artificial nucleation sites.

Most recently, a new analytical technique—spatial scanning spectroelectrochemistry—has been used to study the electrodeposition of Pd nanoparticles at the water/DCE interface.¹⁰⁵  The movable slit for the light beam enabled sampling at well-defined positions on both sides of the interface. It was observed that nanoparticles are not only deposited on the interface, but also diffuse into the bulk aqueous solution.

Liquid/liquid interface has been explored as an environment for self-assembly of NPs, with potential chemical and biological applications. It was pointed out that the assembly of spherical particles at the oil/water interface was driven by the decrease in the total free energy,¹⁰⁶  according to Eq. 1.18¹⁰⁷ :

where, ΔE is the energy change after placing a single particle with an effective radius r at the O/W interface. γO/W,γP/W and γP/O represent the interfacial energy associated with the oil/water interface, particle/water interface, and particle/oil interface, respectively. Three factors can be used to control the assembly process: the effective radius of the NP (larger particles adsorb more strongly than smaller particles); surface modification of the NPs through ligands to change γP/W and γP/O; and the nature of the O/W interface.¹⁰⁸ 

The effect of the particle size was confirmed by Lin et al.¹⁰⁹  They reported the direct assembly of tri-n-octylphosphine oxide covered CdSe nanoparticles at the interface between toluene and water droplets, and the photoinduced NP transport across the toluene/water interface. In this work, 4.6-nm particles were introduced in a dispersion containing water droplets in toluene that had been stabilized with 2.8-nm particles. The differences in adsorption were detected from fluorescence (525nm for the 2.8-nm particles, 610nm for the 4.6-nm particles). It was shown that the 4.6-nm particles displaced the 2.8-nm particles, as expected from the theory.

The second approach—surface modification—was demonstrated by modifying the NPs with ligands containing 2-bromopropionate group, which renders the contact angles of the NPs at the O/W interface close to 90°. The self-assembly of NPs (Au, Ag, γ-Fe₂O₃) at the interface resulted in the formation of closely packed arrays. Furthermore, mixtures of Ag and Au NPs, “nanoalloys”, could also be formed at the interface.¹¹⁰ 

The Vanmaekelbergh group reported spontaneous formation of gold nanocrystal monolayers at the water/oil interface by gradually reducing the surface charge of the nanocrystals. The remarkably robust monolayers could be easily transferred to substrates, suggesting the possibility of technological applications.¹¹¹  The assembly of Au NPs could also be induced by voltage changes, so that the number of particles at the interface was effectively controlled by tuning the Galvani potential difference.¹¹² 

Dai et al. investigated the self-assembled structure of Ag NPs at a trichloroethylene/water interface and reported the first direct observation of NPs in a liquid medium by the environmental transmission electron microscope, as shown in Fig. 21.¹¹³  The spontaneous assembly of Ag NPs into a multilayered, blue opalescent film at the O/W interface was also reported and confirmed by TEM and UV-vis.¹¹⁴ 

Metal nanoparticle-carbon nanotube composite materials were assembled at the diethyl ether/water interface.¹¹⁵  It was shown that carbon nanotubes can mediate the transfer of Ag and Au NPs from water to organic phase and enable the formation of novel nanocomposite films with the NPs bound to nanotubes.

Electrodeposition of NPs at the ITIES opens up a new way of preparing catalysts since NPs with a large surface area often exhibit catalytic properties. So far, only a few examples of using NPs as electrochemical catalysts at the ITIES have been reported. Girault and co-workers showed that electrochemically generated Pd NPs can be used as an ET mediator for photoinduced reduction of TCNQ in the organic phase by the photoactive water-soluble zinc porphyrin at a polarised ITIES.¹¹⁶  Catalytic effects of NPs on more complicated electrochemical reactions have also been shown. For instance, aqueous metal colloid (Au or Pd) prepared by citrate reduction acted as the electron transfer catalyst for the dehalogenation of 2-bromoacetophenone to acetophenone by decamethylferrocene at a DCE/water interface, and the possibility of catalyst recycling was demonstrated.¹¹⁷  More recently, Samec and co-workers presented the first example of using in situ deposited Pt NPs as the electrocatalyst for the oxygen reduction by decamethylferrocene at the polarised water/DCE.¹¹⁸  The convolution analysis of voltammograms showed that the catalytic oxygen reduction proceeds as a four-electron transfer reaction and the reduction rate increased by more than one order of magnitude in the presence of the interfacial Pt NPs.

Here we focus on direct voltammetry of biological macromolecules at the ITIES including proteins, oligosaccharides, and nucleic acids as important nanoparticles and nanoscale building blocks. This new application of the powerful voltammetric approach dramatically widened the range of the targets during the last decade.^85,119,120  Beforehand, the electrochemical behaviour of various biological macromolecules at the ITIES was investigated using impedance spectroscopy and potentiometry. For instance, Vanysek et al. demonstrated the potential-dependent adsorption of bovine serum albumin at the nitrobenzene/water interface, which produced changes in the interfacial capacitance.^121,122  Kakiuchi et al. found that phospholipases maintain their enzymatic activity at the nitrobenzene/water interface modified with phosphatidilecholine monolayers.¹²³  The potential-dependent hydrolysis of the monolayers changed the interfacial capacitance^124,125  and was also imaged by fluorescence microscopy.¹²⁶  Meyerhoff et al.^127–129  pioneered the development of the potentiometric polyion-sensitive electrodes based on the ion-exchange extraction of biological polyions such as protamines and heparins at the interface between water and a polymer membrane.

The direct voltammetric responses based on the extraction of biological macromolecules across the ITIES was demonstrated for the first time using protamines at the micropipet-supported ITIES.¹³⁰  Later, the extraction of the multiply charged protamines was confirmed by micropipet chronoamperometry.¹³¹  Protamines were extracted into a relatively polar nitrobenzene phase containing tetrakis(4-chlorophenyl)borate as an organic counterion. Manning–Oosawa counterion condensation predicts that the multiple charges of a protamine molecule are dense enough to be significantly screened by multiple ions of the tetraphenylborate.¹³²  The extraction of protamines into the less polar DCE phase required a charged ionophore, dinonylnaphthalenesulfonate (DNNS),¹³³  which possesses a sulfonate group to form two charged hydrogen bonds with the guanidinium group of protamine. The formation of a neutral and hydrophobic complex of a protamine molecule with a stoichiometric amount of DNNS molecules was confirmed by using chronoamperometry and cyclic voltammetry at the micropipet-supported ITIES. The quantitative analysis of the voltammograms indicated that the facilitated transfer of protamine by DNNS involves multiple steps, i.e., DNNS adsorption at the interface, interfacial complexation between protamine and DNNS, and complex desorption from the interface. The voltammetric protamine-selective micropipet electrode was also used as an SECM tip to investigate protamine transport across a pnc-Si membrane⁶⁵  (see above).

Arrigan et al.^85,119  and others extended the direct voltammetric approach for the label-free detection of various proteins at the ITIES. Examples are insulin,¹³⁴  haemoglobin,^135–139  α-chymotrypsin,^140–142  α-lactalbumin,¹⁴²  lysozyme,^{137,140,143–145}  myoglobin,^146,147  poly-L-lysine,¹⁴⁷  cytochrome c,^{141,142,147,148}  ribonuclease A,^147–149  albumin,^147,150  trypsins,¹⁴²  trypsin inhibitor,¹⁴⁰  oligopeptides,¹⁵¹  proteinase inhibitor,¹⁴¹  horseradish peroxidases,¹⁴¹  and protein digests.¹⁵²  In contrast to protamines, most of these larger proteins are adsorbed at the ITIES rather than extracted into the organic phase as confirmed by voltammetry and also by impedance spectroscopy.^153,154  Exceptions are α-chymotrypsin¹⁴²  (24 kDa) and cytochrome c^147,148  (12 kDa), which were voltammetrically extracted by bis(2-ethylhexyl)sulfosuccinate into n-octane and DCE/isooctane phases, respectively. The co-extraction of water molecules with cytochrome c was ascribed to the formation of a “reverse micelle” incorporating protein.¹⁴⁹  The protein extraction mediated by reverse micelle has been hypothesized for the separation and purification of proteins, enzymatic reactions in organic solvents, the refolding of denatured proteins, etc.¹⁵⁵ 

Significant progress in development of the electrochemical heparin sensors based on the ITIES has been made using both voltammetric and potentiometric approaches.¹²⁰  Direct voltammetry of heparin at the ITIES was first reported by Samec et al.¹⁵⁶  and quickly followed by the voltammetric detection of this widely used anticoagulant/antithrombotics in diluted¹⁵⁷  and undiluted¹⁵⁸  blood samples. The initial success of voltammetric heparin detection at the ITIES lead to the development of a sensor platform by coating a solid electrode with a plasticized poly(vinyl chloride) membrane as a robust organic phase.^159,160  The solid-supported polymer membrane can be rotated to enhance mass transport and, subsequently, lower a limit of detection. In addition, the current response to heparin is highly sensitive owing to the multiple charges carried by each heparin molecule (–75 as an average for unfractionated heparins) and can be further amplified in stripping voltammetry based on the adsorption and desorption of heparins at the ITIES.^158,159  The reported detection limit of 0.005 unit/mL (equivalent to 2.4 nM with the average molecular weight of 12 kDa for unfractionated heparin)¹⁵⁹  is much lower than therapeutic heparin concentrations (>3.6 unit/mL for cardiac surgery and 0.2–0.7 unit/mL for the treatment of thrombosis and embolisis). Moreover, the voltammetric response to heparins is selective and reversible owing to the external control of the phase boundary potential. These features render the voltammetric approach highly suitable for quantitative characterization of new heparin ionophores,^161,162  which lead to the discovery of the very strong hydrogen-bonding ionophore that can completely extract a heparin molecule (at least up to 20 kDa) into the non-polar DCE phase.¹⁶²  In contrast, the traditional potentiometric counterpart employs the non-polarizable ITIES to yield the mixed-potential response that is non-Nernstian and sensitive to both heparin and its aqueous co-ion.¹⁶³  Pulse chronopotentiometry at the polarizable ITIES gives a reversible non-Nernstian response to highly charged heparins, which is still cross-sensitive to co-ions.¹⁶⁴  Nevertheless, the potentiometric approaches are still useful and were quickly adopted for the detection of oversulfated chondroitin sulfate in biomedical grade heparin preparations^164–166  that has caused severe and even fatal inflammatory responses in patients.^167,168 

In a few reports, the interactions of DNA molecules with small organic molecules at the ITIES were investigated by voltammetry to observe the adsorption and desorption of their complexes. Specifically, the intercalations of DNA with N-methylphenantroline¹⁶⁹  and acridine-calix[4]arene¹⁷⁰  were probed voltammetrically by monitoring the transfer of the cationic intercalators. Electrostatic interactions between DNA and dimethyldioctadecylammonium at the DCE/water interface gave absorption waves although DNA extraction by the reverse micelles of this surfactant had been reported.¹⁷¹ 

Direct voltammetry and SECM were employed to investigate the catalytic activity of the redox enzymes added to the aqueous solution of ITIES systems. Williams et al.¹⁷²  carried out the SECM–ITIES study of glucose oxidase (GOx) to propose two reaction mechanisms. In one mechanism, GOx adsorbs at the ITIES and couples glucose oxidation with the reduction of dimethylferrocenium (DMFc⁺) generated from dimethylferrocene (DMFc) at the Pt tip in the DCE phase, i.e.,

Subsequently, DMFc is regenerated at the ITIES producing a feedback current response at the tip. A similar mechanism was also suggested for the oxidation of DMFc by cytochrome c at the ITIES.¹⁷³  Alternatively, DMFc⁺ may transfer to the aqueous side of the interface to participate in the enzymatic reaction homogeneously and catalytically. This mechanism was supported by Osakai and co-workers, who carried out quantitative voltammetric studies to propose a reaction-layer mechanism for glucose oxidase,¹⁷⁴  D-fructose dehydrogenase,¹⁷⁵  and cytochrome c (Fig. 22).¹⁷⁶ 

This work was supported by the National Institutes of Health (GM073439 for S.A.) and the National Science Foundation (CHE-1213452 for S.A.; and CHE-0957313 and CBET-1251232 for M.M.).