Bioelectronic control of chloride ions and concentration with Ag/AgCl contacts

Ranging from intercellular communication to organ function, ionic species play an important role in natural systems.¹ A majority of physiological processes, such as muscle contraction, neuronal signaling, and metabolism, involve the exchange of ions between cells.² On the other hand, electrons and holes dominate communication between electronic devices. Bioelectronics bridges natural systems and electronic devices by translating ionic signals in the body into electronic signals for sensing and control.^3–7 Iontronic devices, for example, address individual ions and small molecules to control specific physiological pathways.^4,8,9 These include the precise delivery of Ca²⁺, K⁺, and gamma-aminobutyric acid (GABA) to control epileptiform activities and treat neuropathic pain with low dosage and less side effect.^10,11 To this end, our group has demonstrated the control of H⁺ in field-effect transistors¹² and memories,¹³ enzymatic logic gates and bioluminescence,¹⁴ ion channel devices,^15–18 H⁺ modulators for delivery of cargo^19,20 and glucose sensing.²⁰ These devices are enabled by using Pd/PdH_x contacts, also known in the reversible H reference electrode, to translate electronic signals into H⁺ signals. This translation occurs according to the potential dependent reversible reduction, H⁺ + e⁻ ↔ H, at the Pd/PdH_x–solution interface and the subsequent physisorption of H onto Pd to form PdH_x.²¹ An electrical potential applied on the Pd/PdH_x contact shifts the reaction equilibrium and induces the transfer of H⁺ to and from the solution, which effectively controls [H⁺] in the solution. Cl⁻ is a major anion in the physiological environment and functions importantly in many fundamental biological processes, including the regulation of intracellular pH gradients, maintenance of intracellular volume, and resting membrane potential.^22–25 Controlling [Cl⁻] in solution has great significance in regulating extra- and intra-cellular distribution of chloride, which influences cell growth and differentiation, metastatic conversion, patterning of innervation, and neuronal excitability in the central nervous system.^25–28 

Here, we control [Cl⁻] in solution by exploiting perhaps the best-known reference electrode, Ag/AgCl.^23,29–32 We do so by driving the reaction

out of equilibrium condition with a potential (V_Cl−) on the Ag/AgCl contact to effectively transfer Cl⁻ across the contact/solution interface (Fig. 1). While Ag/AgCl is a well-known system, the use of Ag/AgCl to control [Cl⁻] as an e⁻ to Cl⁻ transducer has yet to be reported to the best of our knowledge. Furthermore, we adapt Ag/AgCl contact to a bioelectronic device, a chloride modulator with a chloride reservoir, and a target chamber that is able to host cells. We use this chloride modulator to demonstrate the proof-of-concept actuation of membrane voltage (V_mem) on human pluripotent stem cells (hiPSCs).

First, we use an Ag/AgCl wire as a chloride transducer to show the ability of Ag/AgCl to control [Cl⁻] in solution. We do so in a standard three-electrode configuration with the Ag/AgCl wire as the working electrode (W), glass Ag/AgCl as the reference electrode (R), and platinum (Pt) wire as the counter electrode (C). Figure 1(a) describes how an anodic voltage, $VCl−$⁠, transfers Cl⁻ from solution to the Ag/AgCl wire according to Ag + Cl⁻ → AgCl + e⁻ and, thus, reduces [Cl⁻]. The excess e⁻ flows out of the Ag/AgCl contact into the lead. We use (N-(ethoxycarbonylmethyl)-6-methoxyquinolinium bromide) (MQAE) to monitor [Cl⁻]. MQAE is a diffusion-limited collisional Cl⁻-quenched fluorescent dye whose intensity decreases when [Cl⁻] increases. We calibrate MQAE using the Stern–Volmer equation (Fig. S1) and quantify [Cl⁻].³³ Figure 1(b) shows [Cl⁻] decreases from 50 mM to 32 mM when $VCl−$ = 0.4 V for 100 s. Figure 1(c) describes how a cathodic voltage, $VCl−$⁠, transfers Cl⁻ from the Ag/AgCl wire to the solution according to AgCl + e⁻ → Ag + Cl⁻ and, thus, increases [Cl⁻]. Figure 1(d) shows that [Cl⁻] increases from 0 mM to 48 mM when $VCl−$ = −0.4 V for 100 s. These ranges and changes for [Cl⁻] are comparable to changes in [Cl⁻] plasma, interstitial fluids, and intracellular fluids,²² and they are, thus, relevant for biological applications.

One of the advantages of using AgCl to transfer Cl⁻ to and from solution is the fact that AgCl is selective to Cl⁻ and is able to affect [Cl⁻] under physiological conditions with little to no interference from other ions. To this end, we demonstrate that the Cl⁻ transducer works in a complex solution such as stem cell culture media (Neurobasal^TM-A Medium, ThermoFisher) that contains 79.6 mM NaCl, 26 mM NaHCO₃, and 0.9 mM KH₂PO₄. With $VCl−$ = 0.6 V for 100 s, we are still able to remove Cl⁻ from solution and reduce [Cl⁻] from 79.6 mM to 47 ± 2 mM. Repeating the experiment without HCO₃⁻ and H₂PO₄⁻ ions leads to the same result with the final [Cl⁻] = 46 ± 5 mM (Fig. S2). While we are not able to monitor the change in concentration of the other anions, these results indicate that the presence of other anions does not affect the Cl⁻ transducer indicating that AgCl is indeed specific to Cl⁻. From our work with the AgCl wire, we found that transferring Cl⁻ from the Ag/AgCl contact into the solution is more efficient than transferring Cl⁻ from the solution to the contact. One of the reasons can be that when applying an anodic voltage, Cl⁻ ions are absorbed into Ag/AgCl electrodes immediately, and a depletion area forms around the electrodes.³⁴ Then, the reaction is limited by Cl⁻ migration to the contact/solution interface. We further investigate the electrode kinetics of Ag/AgCl conversion in Sec. III to verify the hypothesis.

To better understand and optimize the Cl⁻ transducer, we investigate the kinetics of Cl⁻ transfer at different potentials. Surprisingly, not many investigations on the kinetics of the conversion between Ag/AgCl under dynamic polarization or under other non-equilibrium exist.³⁵ For a Ag/AgCl electrode in solution containing Cl⁻, the Cl⁻ in the AgCl are in equilibrium with the [Cl⁻] in solution. As such, a Ag/AgCl contact shows a lower electrode potential in a solution with higher [Cl⁻] and vice versa.²³ The well-known Ag/AgCl reference electrode takes advantage of this relationship by reading a constant potential when encapsulated in an internal electrolyte with a constant [Cl⁻].³⁶ AgCl has been used as the gate electrode in organic electrochemical transistors (OECTs), which utilized its faradaic reactions in halide electrolytes to achieve higher current modulation compared to Pt electrodes.³⁷ However, there are few studies exploring the dynamic change in the Ag/AgCl system. Little is known about the kinetics of the conversion between Ag and AgCl.^35,38 Although people have reported the open circuit potential (OCP) and cyclic voltammetry (CV) of Ag/AgCl, the analysis of the conversion between e⁻ and Cl⁻ is absent. Here, we explore the kinetics of Ag/AgCl conversion of miniaturized on-chip Ag/AgCl nanoparticle (NP) electrodes. We use Ag/AgCl NPs to create an interface with a higher surface area and capacitance for better conversion efficiency.³⁹ The conversion between Ag/AgCl under non-equilibrium conditions is consistent with Cl⁻/e⁻ transformation, which is critical for controlling [Cl⁻] by electronic means.

We electroplate Ag NPs on thin-film Au contacts and chlorinate the Ag NPs into Ag/AgCl NPs in a 50 mM KCl solution (Fig. S3). The bare Ag/AgCl NP electrode we make here is a “quasi-reference electrode,” whose electrical potential depends on [Cl⁻] in solution according to the thermodynamically based Nernst relationship,

where $VCl−0$ is the standard potential of Ag/AgCl.

Ag/AgCl conversion depends on [Cl⁻] in solution and the applied potential (⁠$VCl−$⁠). Here, we record the open circuit potential (V_OCP) of Ag/AgCl contact in various [Cl⁻] solutions under equilibrium conditions with a three-electrode configuration at room temperature. Figure 2(a) shows the V_OCP of Ag/AgCl in 1 mM, 10 mM, 100 mM, and 1M KCl, respectively. V_OCP and logarithmic scale of [Cl⁻] shows a linear relationship with a slope of −55.35 mV and R² = 0.999, as expected from the Nernst equation. We use the values of V_OCP to define the process map for the Cl⁻ transducer as a function of [Cl⁻] and $VCl−$ [Fig. 2(a)]. The straight line in the graphs joins the V_OCPs at different [Cl⁻] ions and represents equilibrium (as many Cl⁻ transfer from the solution into the electrode as they transfer from the electrode into the solution). This line also marks the threshold of the Ag/AgCl conversion. For a combination of $VCl−$ and [Cl⁻] in the gray area below the line, Cl⁻ transfer into the solution increasing [Cl⁻] and for a combination of $VCl−$ and [Cl⁻] above the line in the white area Cl⁻ transfer from the solution decreasing [Cl⁻].

To further study the electrode kinetics at the interface of the Ag/AgCl contact and solution, we investigate the limiting mechanism of this electrochemical reaction [Fig. 2(b)]. In detail, we cycle the voltage of Ag/AgCl NP electrodes between −0.8 V and 0.8 V vs a glass Ag/AgCl electrode. Pt wire is the counter electrode as usual. We use a series of scan rates, 20 mV/s, 40 mV/s, 60 mV/s, 80 mV/s, and 100 mV/s, respectively. Taking 40 mV/s as an example, the interfacial reaction can be explained as below. During the scan of $VCl−$ from 0 V to 0.8 V, Cl⁻ is absorbed onto Ag/AgCl and oxidizes Ag to AgCl. $ICl−$ increases to a peak with higher $VCl−$ and then becomes smaller by the delivery of Cl⁻ from the bulk solution. Thus, the mass transport of Cl⁻ becomes the limitation of the electrochemical reaction, and the current decays gradually.⁴⁰ After the scan direction is switched to negative, $VCl−$ is still sufficiently positive to oxidize Ag, so $ICl−$ continues until $VCl−$ becomes strong enough to reduce AgCl. Then, Cl⁻ starts to be released from the electrode, reaches a cathodic peak and decays as AgCl is consumed. With the increase in the scan rate, the diffusion layer becomes smaller, and it takes less time to record one cycle, inducing $ICl−$ increase.^41,42 The measured current in CV with a redox couple is mainly faradaic current, which is from the charge transfer at the interface of the electrode contact/solution and depends on the kinetics of charge transfer at the surface and the redox species diffusion to the surface.⁴¹ We plot the peak currents (⁠$ICl−,Peak$⁠) of CV vs the square root of scan rates of three electrodes in Fig. S4, which all show an extraordinary proportional relationship, and the standard error is shown in Fig. 2(c). According to the Randles–Sevcik equation [Eq. (2)], the analytes, Cl⁻, are not adsorbed on the electrode but freely diffused in the solution. Thus, we conclude that Ag/AgCl conversion is a Cl⁻-diffusion controlled reversible electrochemical reaction,⁴⁰ 

where $ICl−,Peak$⁠, R, F, and T are peak current of cathodic or anodic reaction, gas constant, Faraday constant, and absolute temperature, respectively. A, C, D, and v are electrode area, concentration of the analyte, diffusion coefficient, and scan rate of CV.

As the translation between Cl⁻ and e⁻ is in parallel with the conversion between Ag/AgCl, which depends on $VCl−$ and [Cl⁻] in solution, we can feasibly adapt the Ag/AgCl electrode to various bioelectronic devices and utilize it to selectively control [Cl⁻] in solution with $VCl−$⁠. Here, we design a chloride modulator, which transfers Cl⁻ from a reservoir electrolyte with a rich Cl⁻ source to a target electrolyte under control. The Ag/AgCl NPs act as the working electrode. The selectivity of AgCl to Cl⁻ ensures it a storage media specifically to Cl⁻ and allows Cl⁻ to move between reservoir and target depending on $VCl−$⁠. Thus, the chloride modulator can manipulate [Cl⁻] in a much more extensive range than a single Ag/AgCl contact. As proof-of-concept, we culture cells in the target chamber and study the effect of [Cl⁻] modulation on their function.

Figure 3(a) shows the schematic of the chloride modulator (side view), and the design is similar to previously reported proton modulator with Pd/PdH_x electrodes.¹⁹ The chloride modulator is composed of two independent chambers, referred to as reservoir (left) and target (right), respectively, connected by an anion exchange membrane (AEM). Figure 3(b) shows the optical image of the device. The reservoir has 0.1M KCl to provide a rich Cl⁻ source, and we change [Cl⁻] of the target solution, which can contain 100 μM MQAE in de-ionized (DI) water [Fig. 3(c)], or cells [Figs. 3(d) and 3(e)]. We use an Ag/AgCl NPs microelectrode as the working electrodes, providing multiple independent openings for Cl⁻ to enter/leave the target solution. Here, we define the potential difference between the Ag/AgCl and Pt electrode in the reservoir as $VCl−,R$⁠, and the potential difference between the Ag/AgCl and Pt electrode in the target as $VCl−$⁠. AEM is a polymer with high conductivity for anions due to their large amount of fixed cationic groups that support the storage and transport of mobile anions through the membrane, and its exclusion to cations ensures that the majority of measured currents are from the moving of anions.⁴³ Here, AEM is a mixture of polyvinyl alcohol (PVA) and poly(diallyldimethylammonium chloride) (PDDA), which work as the hydrophilic matrix and polycation, respectively. To avoid the undesirable passive diffusion current caused by the [Cl⁻] gradient between the reservoir and target, we pattern a thin cation exchange membrane (CEM), PVA and polystyrene sulfate acid (PSS) mixture, at the opening of the MEAs to form a Donnan exclusion barrier. The AEM and CEM function together to provide a bridge that allows for an electrophoretic Cl⁻ flow during stimulation with a voltage across the bridge and minimization of passive diffusion current. More details about device fabrication are shown in Fig. S5.

We design a 4-step process in Table I to control [Cl⁻] in target solution periodically by driving Cl⁻ from and back to the reservoir, which results in the [Cl⁻] changes shown in Fig. 3(c). The Cl⁻ modulator also achieves diffusion limited spatial resolution with larger [Cl⁻] changes occurring close to the contacts that are active, with little or no [Cl⁻] changes observed in the non-active region (Fig. S6).

As a proof-of-concept, we used the chloride modulator to study the effects of extracellular [Cl⁻] on human pluripotent mammalian stem cells (hiPSCs) and their V_mem [Figs. 3(d) and 3(e)]. A cell’s resting potential, V_mem, is an electrical control signal that occurs between the inside of the cell and the extracellular environment.⁴⁴ V_mem affects cell physiology and functions such as proliferation, differentiation, migration, and apoptosis, as well as cell–cell communication and large-scale morphogenesis.⁴⁵ 

Chloride is the major anion in extracellular environments. In hiPSCs, an increase in extracellular [Cl⁻] results in cell hyperpolarization (higher V_mem), and a decrease in extracellular [Cl⁻] results in cell depolarization (lower V_mem).⁴⁶ We measure V_mem of the hiPSCs using ArcLight, a fluorescent reporter that we expressed on the cell membrane, whose fluorescence intensity increases when V_mem increases.⁴⁷ By switching the environment between high [Cl⁻] and low [Cl⁻] conditions using the same protocol in Table I, we observe that V_mem of hiPSCs shifts with the expected pattern as measured by the ArcLight fluorescence intensity [Fig. 3(d)]. Figure 3(e) shows the fluorescence images of the cells at t = 75 s and t = 119 s, corresponding to high (hyperpolarized) and low V_mem(depolarized), respectively. This change in fluorescence occurs close to the activated area and not far away from the active electrodes, indicating that the induced change in [Cl⁻] is responsible for the measured V_mem change. While this is a proof-of-concept, it shows how the Cl⁻ modulator is able to affect the cell function.

Ag/AgCl is a well-known reference electrode that has been used in electrochemistry for decades. Here, we first demonstrate Ag/AgCl as a Cl⁻ transducer that controls [Cl⁻] by transferring Cl⁻ from and to solution controlled by an applied voltage, $VCl−$⁠. Furthermore, we investigate kinetics of the conversion between Ag/AgCl by using electrochemical techniques, which are consistent with Cl⁻/e⁻ transformation and critical for controlling [Cl⁻] by electronic means. This control paves the way to the integration of Ag/AgCl into complex bioelectronic devices for dynamic control of bioelectric signaling in vivo and in vitro models of physiology and morphogenesis. As a proof of concept, we adapt Ag/AgCl into a chloride modulator that precisely regulates [Cl⁻] in solution and affects the membrane voltage of pluripotent stem cells. Since Cl⁻ is a critical anion in bioelectricity, the accurate regulation of [Cl⁻] is currently in high demand and has potential in bioelectronic therapies.

See the supplementary material for the fluorescence dye calibration, Ag/AgCl NP electroplating, and fabrication of the chloride modulator.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

This research was sponsored by the Defense Advanced Research Projects Agency (DARPA), Army Research Office, under Cooperative Agreement No. W911NF-18-2-0104, and the Department of Interior, Award No. D20AC00003. The content of the information does not necessarily reflect the position or the policy of the Government, and no official endorsement should be inferred. Microfabrication was performed using equipment sponsored by the W. M. Keck Center for Nanoscale Optofluidics, the California Institute for Quantitative Biosciences (QB3), and the Army Research Office, Award No. W911NF-17-1-0460. We also acknowledge Dr. Tom Yuzvinsky for the help in performing electron microscopy and the W. M. Keck Center for Nanoscale Optofluidics for the use of the FEI Quanta 3D Dualbeam microscope.