Interpreting Dynamic Interfacial Changes at Carbon Fiber Microelectrodes Using Electrochemical Impedance Spectroscopy

Abstract

Carbon-fiber microelectrodes are instrumental tools in neuroscience used for the electroanalysis of neurochemical dynamics and recordings of neural activity. However, performance is variable and dependent on fabrication strategies, the biological response to implantation, and the physical and chemical composition of the recording environment. This presents an analytical challenge, as electrode performance is difficult to quantitatively assess in situ, especially when electrodes are permanently implanted or cemented in place. We previously reported that electrode impedance directly impacts electrochemical performance for molecular sensing. In this work, we investigate the impacts of individual components of the electrochemical system on impedance. Equivalent circuit models for glass- and silica-insulated carbon-fiber microelectrodes were determined using electrochemical impedance spectroscopy (EIS). The models were validated based on the ability to assign individual circuit elements to physical properties of the electrochemical system. Investigations were performed to evaluate the utility of the models in providing feedback on how changes in ionic strength and carbon fiber material alter impedance properties. Finally, EIS measurements were used to investigate the electrode/solution interface prior to, during, and following implantation in live brain tissue. A significant increase in impedance and decrease in capacitance occur during tissue exposure and persist following implantation. Electrochemical conditioning, which occurs continually during fast-scan cyclic voltammetry recordings, etches and renews the carbon surface, mitigating these effects. Overall, the results establish EIS as a powerful method for characterization of carbon-fiber microelectrodes, providing unprecedented insight into how real-world factors affect the electrode/solution interface.

Figure 1.

Introduction to EIS and FSCV. A) Typical FSCV waveform (left). Background CVs (middle) are shown for a potential-naïve and an electrochemically conditioned carbon-fiber microelectrode in the absence and presence of dopamine. A background-subtracted CV for dopamine is also shown (right). B) 10, 100, and 1000 Hz EIS potential waveforms (left) and current responses (right). An overlay of a 100 Hz potential waveform (left axis) and the generated current (right axis) are shown in the middle panel, displaying the phase shift. C) Impedance (left) and phase angle (right) Bode plots, as well as a Nyquist plot (bottom) comparing EIS measurements made using a three-electrode (black) versus a two-electrode (red) system (n=6 electrodes).

Figure 2.

EIS performed from 1 Hz to 1 MHz at carbon-fiber microelectrodes insulated in glass (red; n=16) and silica (black; n=11). A) Graphic depicting physical differences between these electrode subtypes. B) Average impedance (left) and phase (right) Bode plots. C) Average Nyquist plot. Inset is an expanded view from 100 Hz to 1 MHz (dashed box). D) Plot of capacitive reactance versus frequency, as well as relevant equations for determining capacitance.

Figure 3.

Equivalent model circuits used to fit EIS data. A) Model circuits and a description of the physical significance of each resistance (R) and constant phase (Q) element. B) Average impedance Bode (left), phase Bode (middle), and Nyquist (right) plots for the glass-insulated (top; n=16) and silica-insulated (bottom; n=11) microelectrodes, overlaid with fits performed using both equivalent circuits. C) Quantitative comparison of equivalent circuit values generated in fitting impedance data. D) Specific capacitance values (left) and RC time constants (right) calculated using R1 and Q1 fit values.

Figure 4.

The equivalent circuit model can be used to accurately report on controlled manipulations to the electrochemical system. A) Constant phase element and resistance values for glass-insulated microelectrodes as a function of ionic strength (left and right), and the approximated Debye length (middle); all x-axes plotted on a log10 scale. B) Constant phase element value as a function of geometric surface area. Black and red data points were generated using glass-insulated cylindrical T-650 and disc P-55 microelectrodes, respectively. C) Fit values for pristine (black) and electrochemically conditioned (gray) T-650 microelectrodes, and for conditioned P-55 microelectrodes (white) of equal length.

Figure 5.

Tissue exposure impacts impedance properties. A) Impedance (top-left) and phase (top-right) Bode plots, as well as Nyquist (bottom-left) and capacitive reactance versus frequency (bottom-right) plots for silica-insulated microelectrodes (n=5). Data were collected in buffer prior to implantation in the striatum, while implanted in tissue, in buffer following removal from tissue, as well as following electrochemical conditioning to clean the electrode surface. B) 100 Hz impedance (top) and capacitive reactance (bottom) values extracted from the data displayed in A). C) Equivalent circuit fit values to EIS data.