Triple-Barrel Ultramicroelectrodes for Multipurpose, Submilliliter Electroanalysis

Abstract

Here, we have developed and applied a triple-barrel microelectrode. This device incorporates a platinum disk working electrode, a platinum disk counter electrode, and a low-leakage Ag/AgCl reference electrode into a small probe. We demonstrate that the incorporated low-leakage reference electrode shows similar voltammetry, potentiometry, and drift when compared to a commercial reference electrode in bulk solution. We also demonstrate the versatility of such a small three-channel system via voltammetry in nanoliter droplets and through electroanalysis of captured aerosols. Finally, we demonstrate the probe’s potential utility in single-cell electroanalysis by making measurements within salmon eggs.

The electrode has a long history of evolution. One need only look at the first electrophore developed by Johan Wilcke (and significantly improved by Alessandro VoltayA-lessandro Volta¹) to study static electricity and compare it to modern nanometer-sized electrodes that can measure cellular activity² to see how far this technology has come in both its form and function. Not only have electrodes seen significant miniaturization, but they have also undergone a transition from single to multichannel probes. Examples of two or more probes can be readily found in the literature.^3–5 The applications of these devices range from neuroscience⁶ to scanning electrochemical microscopy³ to corrosion studies.⁴ While there have been many applications of multichannel electrodes demonstrated, the full range of their utility has not yet been elucidated. Thus, we introduce here a triple-barrel ultramicroelectrode (UME). We have termed this UME the Frankenelectrode, which takes inspiration from Frankenstein’s monster⁷ as an effort to push forward the scientific endeavor through a piece-wise compilation that is greater than the sum of its individual parts.

The triple-barrel UME (Figure 1) consists of three channels, wherein all channels contain a platinum wire heat sealed using a blowtorch to form three 25 μm platinum disks sealed within borosilicate glass. Electrical connection is made in two of these channels to provide a working and a counter electrode. In the third channel, a solution of KCl and a AgCl-coated silver wire are inserted to form a low-leakage reference electrode, as previously reported by Walker and Dick.⁸ This differentiates it from other multibarreled electrodes, which use either an external commercial reference electrode,³ a quasi-reference electrode,⁶ or an internal reference electrode that leaks KCl into the sample during the measurement.^4,5 All of these options are undesirable within small volumes.

Figure 1.

(a) Schematic and (b) photograph of the triple-barrel UME, consisting of a triple-barrel capillary tube with 25 μm platinum disks sealed into one end, forming the three electrode surfaces. One disk is the working electrode (green lead), one is the counter electrode (red lead), and the third is the tip of the bipolar Ag/AgCl reference electrode (white lead). (c) 5X magnified image of the tip of the triple-barrel UME, showing the three platinum disks sealed in a triangle.

Using this triple-barrel UME, we demonstrate three applications: measurements in nanoliter droplets, aerosol electroanalysis, and intracellular measurements. Measurements inside of small droplets allow for the study of enhanced reaction rates as previously seen in other confined systems.^9–13 Intracellular measurements taken with the reference electrode within the cell avoid confounding of the measurement from the membrane potential.¹⁴ The triple-barrel UME demonstrates a proof-of-concept for these kinds of measurements, but further investigations will consider the miniaturization of this device to probe even smaller droplets and biological systems. Lastly, aerosols contribute to many environmental processes¹⁵ and health concerns.¹⁶ Thus, their detection is of significant importance. Using a previously reported method,¹⁷ the triple-barrel UME can collect and analyze aerosols. Our applications of the triple-barrel UME not only show the value of this tool, but they also pave the way for scientists to use this tool in other creative ways.

EXPERIMENTAL SECTION

Chemicals and Reagents.

Potassium ferricyanide (99%), potassium ferrocyanide (99%), potassium chloride (certified ACS), and hydrochloric acid (ACS grade) were obtained from Thermo Fisher Scientific. Dulbecco’s phosphate buffered saline (PBS) (1X) with calcium and magnesium (pH 6.8) was purchased from Corning. Salmon eggs were bought from Marky’s Caviar Company. All solutions were prepared in ultrapure water or PBS and sonicated using an ultrasonic cleaner (VWR International, Radnor, PA).

Instrumentation.

All electrochemical experiments were performed on a CHI model 601D potentiostat (CH Instruments, Austin, TX, input impedance = 1 TΩ). A Ag/AgCl reference electrode (stored in 1 M KCl) or the triple-barrel UME’s inbuilt homemade Ag/AgCl reference electrode was used in every experiment. All electrodes were purchased from CHI (CH Instruments, Austin, TX).

Triple-Barrel UME Fabrication.

Triple-barrel borosilicate glass capillary tubes were used with outer diameter = 1.0 mm, inner diameter = 0.50 mm, length = 5 cm (Sutter Instruments, Novato, CA). One platinum wire (d = 25 μm, length) from Goodfellow Cambridge Ltd. (Huntington, UK) was carefully inserted into each barrel of the tube such that it was flush with the edge before it was sealed using a propane torch (BernzOmatic, Newark, NJ). Electrical connection was made to two of the sealed platinum disks using 22-gauge copper wire (Plusivo, Miami, FL) and gallium. The third became a reference electrode through a method previously described.⁸ Briefly, the capillary was filled with 1 M KCl and a Ag/AgCl wire made using silver wire (d = 0.25 mm) anodized in 1 M HCl. All three wires were hot glued at the top to keep them in place.

Bulk and Droplet Measurements.

Bulk and droplet measurements were made in solutions of 50 mM ferricyanide and 50 mM ferrocyanide in 1 M KCl. For bulk measurements, the triple-barrel UME was placed in a vial containing 10 mL of this solution before the electrical connections were made, and the measurement was taken. For the droplet measurements, a 750 nL droplet was pipetted onto the tip of the inverted triple-barrel UME before electrochemical connection was made and the measurement began.

Aerosol Measurements.

Adapted from a previous method,¹⁷ a 500 nL droplet of 1 M KCl was pipetted onto the tip of the inverted triple-barrel UME before electrical connection was made. Then, aerosols were generated by flowing compressed air through a Hudson RCI 1724 nebulizer filled with a solution of 300 mM ferricyanide and 300 mM ferrocyanide in ultrapure water at the droplet for 5 s. Then the measurement was taken.

Intracellular Measurements.

Fish eggs were incubated in a solution of 10 mM ferricyanide and 10 mM ferrocyanide in 1 M KCl for 3 h. Eggs were then gently dried using a paper towel and placed in a 3.5 mL cell culture dish (VWR) modified to hold the egg steady (Supporting Information, Figure S1). Then, two needles were used to make a hole in the top of the egg, and the triple-barrel UME was inserted into the hole. Then, the electrical connections were made (Supporting Information, Figure S2), and a measurement was taken. In a modification, the triple-barrel UME was used as the working and counter electrode, and a commercial Ag/AgCl reference electrode was placed in a PBS solution surrounding the egg (Supporting Information, Figure S3).

RESULTS AND DISCUSSION

The triple-barrel UME (Figure 1a, b) was fabricated from three fused borosilicate glass capillary tubes. Into one end of each tube, a platinum wire (d = 25 μm) was sealed to create three 25 μm platinum disks in a triangle (Figure 1c). Two were converted into a working electrode and a counter electrode by making electrical connection to the inner portion of the sealed platinum wire with a copper wire and gallium. The third was converted into a very low leakage Ag/AgCl reference electrode following a method we developed previously.⁸

After fabrication, the triple-barrel UME was tested by placement in a bulk solution of 50 mM ferricyanide and 50 mM ferrocyanide in 1 M KCl. Both the oxidized and reduced forms of the mediator were used, as the best performance was observed when both species were present in solution. The cyclic voltammogram showed the characteristic shape (Figure 2a, blue) of an ultramicroelectrode. When compared to the cyclic voltammogram taken using the triple-barrel UME as the working and counter electrode and a commercial Ag/AgCl reference electrode (Figure 2a, orange), the voltammograms were identical. This indicates that the reference electrode barrel in the triple-barrel UME, designed based on our BPRE (bipolar reference electrode) is measuring the same potential as a commercial Ag/AgCl reference electrode. The half-wave potential (E_1/2) of the ferri:ferrocyanide redox couple was determined to be 0.2484 ± 0.0002 V vs Ag/AgCl (n = 3), and the E_1/2 using a commercial Ag/AgCl reference electrode was determined to be 0.2544 ± 0.0002 V vs Ag/AgCl (n = 3). This 6 mV difference between the two can be attributed to slight differences in the AgCl coatings on the homemade vs the commercial Ag/AgCl wires. The commercial Ag/AgCl wire could have been made through a variety of methods, of which the thermoelectrolytic method is most common,¹⁸ and the in-house ones are made by an electrolytic method. The differences in fabrication method, temperature during AgCl formation, and speed of AgCl layer formation can lead to small variances in the AgCl coating, which, in turn, results in small differences in the measured potential.¹⁸

Figure 2.

Cyclic voltammograms taken in (a) a bulk solution of 50 mM ferricyanide and 50 mM ferrocyanide in 1 M KCl using the triple-barrel UME (blue) and the triple-barrel UME when the reference lead is attached to a commercial Ag/AgCl reference electrode (orange) at 50 mV/s, (b) a 750 nL droplet of 50 mM ferricyanide and 50 mM ferrocyanide in 1 M KCl pipetted onto the tip of the triple-barrel UME at 100 mV/s, (c) a 1 M KCl droplet pipetted onto the tip of the triple-barrel UME after 5 s of an aerosol of 300 mM ferricyanide and 300 mM ferrocyanide in ultrapure water was sprayed into the droplet at 100 mV/s, and (d) a single salmon egg after 3 h of incubation in a solution of 10 mM ferricyanide and 10 mM ferrocyanide in 1 M KCl at 100 mV/s.

Further, the limiting current of a cyclic voltammogram (i_lim) taken at an inlaid disk on an ultramicroelectrode can be related to the radius using the following equation:¹⁹

$$i_{\text{lim}}=4\mathit{\text{nFDC}}*r$$ (1)

where n is the number of electrons contributing to the reaction, F is the Faraday, D is the diffusion coefficient, C* is the bulk concentration of the redox species, and r is the radius of the electrode. Using the diffusion coefficient²⁰ for ferrocyanide, the limiting current measured gives a diameter of 25.8 μm. Magnified images of the surface of the electrode show that the working electrode disk is 26.7 μm in diameter (Supporting Information, Figure S4). The overall diameter of the probe, including the glass sheath, is approximately 1 mm. We have used the equation for an inlaid disk because the voltammetry and microscopy show no evidence of recession (i.e., no peaks observed in the voltammograms). Thus, we believe this equation adequately describes the behavior at the working electrode.

Next, open circuit potentiograms were taken using the triple-barrel UME in 50 mM ferricyanide and 50 mM ferrocyanide in 1 M KCl (Supporting Information, Figure S5). The average open circuit potential in this system was 0.2607 ± 0.0006 V vs Ag/AgCl (n = 3). On average, there was a drift of 1.9 ± 0.4 mV over 8 h (n = 3) (Supporting Information, Table S1). These values are comparable to those previously measured for drift in the homemade reference electrode.⁸ For comparison, a 10 min open circuit potentiogram was recorded in the same solution using the triple-barrel UME as the working and counter electrodes and a commercial Ag/AgCl reference electrode. This measured 0.2619 V vs Ag/AgCl, which is within measurement error of the potential measured using the triple-barrel UME’s reference electrode. This setup measured a drift of 0.4 mV over 10 min. These drifts, and those done previously,⁸ are comparable to the drift calculated here for the triple-barrel UME.

For comparison, the triple-barrel UME’s drift was also calculated in the same manner as previously described, except the reference lead was clipped to the counter electrode, to determine drift in systems similar to other multibarreled electrodes in which a quasi-reference counter electrode is used. In this way, the system took 4.8 h to equilibrate, after which the equilibration potential was rapidly reached. If the equilibration time is included, the average potential measured was −0.0638 ± 0.057 V vs Ag/AgCl over 8 h (n = 3), and the drift was 168 ± 291 mV over that time. If the equilibration period is ignored, the average potential measured was 0.0052 ± 0.00002 V vs Ag/AgCl over 8 h (n = 3), and the drift was 0.6 ± 0.9 mV over that time. After the long equilibration time, the Pt disk quasi-reference/counter electrode is comparably stable to the internal Ag/AgCl reference electrode. Thus, the leakless BPRE shows excellent stability compared to a simple Pt quasi-reference.

However, when both the internal Ag/AgCl reference electrode and the Pt disk quasi-reference/counter electrode are used to gather cyclic voltammograms on the triple-barrel UME, there is substantially more drift using the latter. The internal Ag/AgCl reference electrode exhibits 13.5 ± 7.3 mV drift over the 100 consecutive cyclic voltammograms in 50 mM ferricyanide and 50 mM ferrocyanide in 1 M KCl. The Pt disk quasi-reference/counter electrode measured 34.7 ± 33.3 mV drift over the 100 consecutive cyclic voltammograms (Table 1).

Table 1.

Measured Potential and Drift

	Measured OCP (V)	Drift (mV) during OCP	Drift (mV) during 100 CVs
Internal Ag/AgCl reference	0.2607 ± 0.0006	1.9 ± 0.4	13.5 ± 7.3
Pt disk quasi-reference/counter	−0.0638 ± 0.057	168.2 ± 291.2	34.7 ± 33.3
Pt disk quasi-reference/counter, ignoring equilibration	0.0052 ± 0.00002	0.6 ± 0.9	N/A

Of note, the Pt disk quasi-reference/counter system altered the shape and current mangitude of the cyclic voltammogram (see Supporting Information, Figure S6).

Next, experiments were done to demonstrate the utility of such a small three-electrode system under a variety of different conditions: droplet, aerosol, and intracellular.

A droplet of 750 nL was pipetted onto the tip of the triple-barrel UME while it was held inverted (tip facing up, leads facing down) to form a hemisphere of radius 710 μm. This droplet contained 50 mM ferricyanide and 50 mM ferrocyanide in 1 M KCl, and when a cyclic voltammogram was taken, it shows the characteristic shape and an E_1/2 of the ferri:ferrocyanide redox couple on an ultramicroelectrode (Figure 2b). Interestingly, the current increases over time as the droplet evaporates, concentrating the redox molecules. Using eq 1, the previously determined radius of 26.7 μm, the concentration increases an average of 1.18 ± 0.08 times over 100–160s, indicating an evaporation rate of 0.55 ± 0.17 nL/s (n = 8).

There is a growing understanding of the effects of confinement of chemical reactions.^9–13 Provided that at least one component of the reaction is electroactive, the triple-barrel UME is a powerful tool for directly measuring what is occurring in a droplet, and the smaller the tip of the triple-barrel UME that can be fabricated, the smaller the droplet that can be studied. Efforts are currently underway to employ laser pulling methods similar to those utilized to manufacture nanoelectrodes^21–23 to achieve smaller and sharper triple-barrel UMEs.

For aerosol detection, a 500 nL droplet of 1 M KCl was pipetted onto the tip of the triple-barrel UME to act as an aerosol collector volume, in a system similar to the aerosol detection method developed previously.¹⁷ Into this collector volume, the analyte, a solution of 300 mM ferricyanide and 300 mM ferrocyanide, was sprayed for 5 s using a nebulizer. The resultant cyclic voltammogram (Figure 2c) again has the characteristic sigmoidal shape and E_1/2 value of the ferri:ferrocyanide redox couple on an ultramicroelectrode. Again, as the collector volume evaporated, it concentrated the ferri- and ferrocyanide, causing successive scans of the cyclic voltammogram to increase in current. Using eq 1 and the previously determined radius of 26.7 μm, the concentration increases an average of 1.63 ± 0.03 times over 140 s, indicating an evaporation rate of 2.1 ± 0.07 nL/s (n = 6).

In this manner, the triple-barrel UME provides a highly portable method for making aerosol measurements in real time. This could be useful for detection of toxic electroactive species such as lead, by performing anodic stripping voltammetry on the triple-barrel UME,¹⁷ or even indirect detection of nonelectroactive environmental contaminants like perfluorooctanesulfonic acid by modification of the working electrode with a molecularly imprinted polymer.^17,24 More fundamentally, this miniaturized version of the previously designed aerosol detection system can be used to make real-time measurements of electrospray droplets to learn about their charged state without interrupting the mass spectrometric analysis.²⁵

Finally, experiments were done in which the triple-barrel UME was inserted into salmon eggs to demonstrate their usage for single cell intracellular studies. The eggs had been incubated in a solution of 10 mM ferricyanide and 10 mM ferrocyanide in PBS for 3 h prior to the measurement after which the eggs were dried off gently and the triple-barrel UME was inserted inside. The system used for these proof-of-concept studies was chosen because the probe is still quite large so a large cell was required. Also, salmon eggs are commercially available and can be used without culturing facilities, while still containing similar components as found in mammalian cell lines.²⁶ Ferri/ferrocyanide was used as a model mediator for consistency throughout the different experiment types and because it has been shown to have biological relevancy.^27–30

The resultant cyclic voltammogram (Figure 2d) again shows the characteristic shape and E_1/2 value of the redox couple on an ultramicroelectrode. The current magnitude is lower than would be expected of a solution of 10 mM ferricyanide and 10 mM ferrocyanide for two reasons. First, that after 3 h, not all of the ferri- and ferrocyanide has been passively taken up by the cell. Since the rate of passive uptake is dependent on the ability of the redox moelcules to pass through the egg’s membrane,³¹ the charge state of the redox species affects the efficacy of their passive uptake. Second, the viscosity inside of a cell is higher than that of 1 M KCl; thus, the diffusion coefficient of the redox pair is lower inside the cell.¹⁴ Other examples of cyclic voltammograms taken inside multiple salmon eggs are shown in Supporting Information, Figure S7.

There is a great deal of interest in making intracellular measurements of various species in single cells to further our understanding of cell-to-cell heterogeneity and how it affects various disease states.^32–34 While the majority of these studies are done using mass spectrometry methods,³³ there are a growing number of direct electrochemical methods being designed using modified nanoelectrodes.³⁵ However, these methods all use either quasi-reference electrodes or external reference electrodes and thus ignore the membrane potential of the cell, which can lead to ambiguity in measurements of potential. By directly placing our reference electrode inside of the cell when making the measurement, we are able to avoid this issue.

To obtain the membrane potential, cyclic voltammograms were taken in a modified manner by placing a small amount of PBS into the cell culture dish surrounding the salmon egg, in which a commercial Ag/AgCl reference electrode was placed. In this manner, cyclic voltammograms were first taken inside of the salmon egg, as described above (see Supporting Information, Figure S2). Then, the reference electrode lead was moved to the external commercial Ag/AgCl reference electrode (see Supporting Information, Figure S3), and another measurement was taken. The cyclic voltammograms in both were the characteristic sigmoidal shape, but those taken using the external commercial Ag/AgCl reference electrode displayed a shifted E_1/2 (Supporting Information, Figure S8) due to the membrane potential of the salmon egg.¹⁴ These measurements show that when the homemade Ag/AgCl reference electrode in the triple-barrel UME is used, the average E_1/2 measured is 0.2561 ± 0.009 V vs Ag/AgCl (n = 8), whereas it is measured to be 0.2108 ± 0.007 V vs Ag/AgCl (n = 8) on average when the external commercial Ag/AgCl reference is used. This is an average shift in the ferri/ferrocyanide redox couple’s E_1/2 value of 45.3 ± 13.0 mV more negative (n = 8) (Figure 3 normalized, Supporting Information, Figure S9 not normalized). These experiments demonstrate that the membrane potential of a cell can cause substantial shifting of the measurement and must be taken into account for accurate intracellular measurements. The triple-barrel UME thus avoids the membrane potential and the resistance to charge transfer across the membrane, which influence the mesurement response time.

Figure 3.

Cyclic voltammograms taken in a single salmon egg after 3 h of incubation in a solution of 10 mM ferricyanide and 10 mM ferrocyanide in 1 M KCl at 100 mV/s using the (blue) internal Ag/AgCl reference or (orange) external Ag/AgCl reference electrode to demonstrate the effects of the cellular membrane potential.

In another study, the membrane potentials of MCF-10A cells were estimated to be ΔE_1/2 = 46 ± 4 mV and 39 mV through measurements of ferrocenemethanol or Wurster’s reagent in solution surrounding the cells, as well as inside of the cells.¹⁴ In all cases, the reference electrode was placed in the external PBS solution. They argue that the sharpness of their nanoelectrode probe ensures no leakage occurs through the pierced section of the cell and demonstrate that the redox species is not moving through but do not account for other ion movements. Our method does not allow for other ions to pass through the pierced section of the cell, as the liquid level into which the external reference electrode is placed is far lower than where the hole is made. Thus, the use of the triple-barrel UME provides a less ambiguous method for quantification of the membrane potential of a cell and allows one to avoid complications of electronically communicating between a working electrode and a reference electrode through a membrane (e.g., ohmic drop and associated time constants).

We acknowledge that the system used in this proof-of concept study is quite large. The probe is 1 mm in diameter, and the cells are ~2 mm in diameter. However, laser pulling the tips to make them sharper and smaller, as discussed above, would allow for these measurements to be taken in even smaller cells and with less damage done to the cell upon insertion. Optimization to create smaller probe sizes will be the subject of future work. Additionally, functionalization of the working electrode surface will allow the user to selectively measure species of interest within cells, including those that are redox inactive. Sensitivity of these methods would depend on the particular method of functionalization used, but a wide variety of methods have already been designed for nano-electrochemical single cell measurements.³⁵ Our probe could be a platform that these sensors could be integrated onto for specific, sensitive measurements of biologically relevant species within single cells.

CONCLUSIONS

The triple-barrel UME is a unique, three-electrode system that allows for the insertion into small droplets and cells such that direct electrochemical measurements can be made in such confined volumes. The potentiometric measurements are comparable to commercial reference electrodes, indicating that the incorporation of the reference electrode into the third channel of the glass capillary does not negatively impact its measurement capabilities. The triple-barrel UME can also be used to entrap and detect aerosols of interest, as well as to directly quantify the membrane potential of a single cell. Using this system creatively will add insight into a wide set of interesting questions in small volumes.

ABBREVIATIONS

PBS

Dulbecco’s phosphate buffer saline (1X)

CV

cyclic voltammogram

OCP

open circuit potentiometry

Ag/AgCl

silver/silver chloride

E_1/2

half-wave potential

PFOS

perfluorooctanesulfonic acid

UME

ultramicroelectrode