Stimulation and Recording Electrodes: General Concepts

Abstract

In prosthetic devices, electrodes are the interfaces between the implant system and the body. Electrodes may be used for neural stimulation or neural signal recording according to the application. The signals to be recorded are typically small, i.e. some tens of microvolts for single pulse activity to a maximum amplitude of around 80 mV for intracellular potentials measured for example by fine-tipped electrolyte-filled glass micropipettes in cognitive studies in brain machine interface. On the opposite, neural stimulation sometimes needs relatively high electrode voltages and current densities, sometimes as high as several volts, so it may lead to a high enough energy transfer triggering chemical reactions that involve corrosion and changes in electrode properties.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

In prosthetic devices, electrodes are the interfaces between the implant system and the body. Electrodes may be used for neural stimulation or neural signal recording according to the application. The signals to be recorded are typically small, i.e. some tens of microvolts for single pulse activity to a maximum amplitude of around 80 mV for intracellular potentials measured for example by fine-tipped electrolyte-filled glass micropipettes in cognitive studies in brain machine interface. On the opposite, neural stimulation sometimes needs relatively high electrode voltages and current densities, sometimes as high as several volts, so it may lead to a high enough energy transfer triggering chemical reactions that involve corrosion and changes in electrode properties.

The electrodes used in neuroprosthetics appear in different sizes and areas. Regarding area, one could categorize electrodes into microelectrodes and macroelectrodes. Although there is not a clear boundary, an area lower than 10, 000 μm² corresponds to a microelectrode and above 100, 000 μm² (0. 001 cm²) to a macroelectrode [1]. Macroelectrodes are sometimes much larger, such as deep brain stimulation electrodes with an area of 0. 06 cm². Electrodes come in different shapes. Some electrodes used are flat and disk or square shaped, like the ones used in visual prostheses, and some like the ones used in cognitive control of prosthetic limbs for paralyzed patients have the shape of a shaft. The electrode of cardiac pacemaker is helical (Fig. 1.1).

Fig. 1.1

The electrode of a cardiac pacemaker

Except for the case of the micropipette electrode for which the conducting electrode material is an electrolyte, the electrodes are built from metals or conducting ceramics like titanium nitride (TiN). Therefore, the boundary between the electrode and the electrolyte forms a phase boundary. For a well conducting electrolyte, this phase boundary can be modeled by a capacitance whose dielectric consists of two consecutive water molecule layers. The conducting plates of this capacitor are the electrode and the electrolyte. The two layers are the water dipoles adsorbed on the electrode surface and the hydration envelope of the ions in the vicinity of the electrode-electrolyte phase boundary. This capacitor is called Helmholtz double layer capacitor. In practice, especially when the electrolyte has a low concentration and therefore a high resistance, another capacitance in series with the Helmholtz double layer is considered, the so called Gouy-Chapman capacitance. It corresponds to a diffuse area of space charge which neutralizes the immobilized charge directly on the electrode surface. This capacitor is neglected in the following.

Charge injection into the electrode may be through charging and discharging the Helmholtz capacitor. When this capacitor is charged, positive and negative charges (in solid electrode, electrons and ions; in solution, only ions) gather on its two plates, namely the electrode and the electrolyte. The electrode is said to be polarized. Metal and many ceramic electrodes are said to be polarizable, because the interface between their surface and the electrolyte can be regarded as a chargeable capacitor. The micropipette electrode mentioned above is non-polarizable. There is no phase boundary here and thus no interface capacitor can exist. The charge transfer is done by ions flowing into or out of the solution from the electrode body.

Depending on the material and especially at higher charge magnitudes, charge transfer can occur by electron transfer across the surface. This happens through redox reactions at the electrode surface. For example, the anions are oxidized and transformed into gas at anode (e.g. Cl⁻ into Cl₂) and the cations are reduced and precipitate on cathode (e.g. Cu²⁺ into Cu). If a considerable amount of electron transfer (ideally unlimited) occurs at already low electrode-electrolyte potential differences, the charges can not be separated across the phase boundary and no capacitor can exist. The material is again non-polarizable, even if there is a phase boundary. One example is the silver-silver chloride electrode which is used as a reference electrode and can be considered as ideally non-polarizable for low current magnitudes (this electrode will be explained below). In practice the electrodes lie somewhere between being perfectly polarizable or perfectly non-polarizable.

If the material is polarizable, like metals or TiN, for lower charge injection densities, capacitive currents flow into the tissue or the solution representing the tissue. Here only the electrode Helmholtz capacitance is charged to make current injection possible. Higher charge injections require a charge transfer across the electrode surface through the redox reactions.

When producing electrodes for neural stimulation within the body, choosing the best electrode material for the application is important. Several criteria must be considered when choosing these materials. Materials that consume very little power while still operating efficiently and safely are ideal. It is also important for the electrodes to be as small as possible so the implanted chip can maximize performance while minimizing intrusion to the patient [6].

Tow types of area are considered for the electrodes: real surface area and geometric surface area. Geometric surface area is the traditional idea of surface area (length × width for a rectangular electrode, π × (radius)² for a circular electrode, etc.). In practice, if the electrode surface is rough or porous, the area at which the electrode and electrolyte touch is much larger than the geometric surface area. This interface area is called real surface area. Real surface area (or electrochemical surface area) factors in the surface roughness of the electrode being used and can be orders of magnitude larger than the geometric surface area. A rough electrode has a higher Helmholtz capacitance compared to a smooth surface. In the case of faradaic current flow, more surface chemical reactions can occur on a rough electrode than a smooth one with the same geometric surface area. The real surface area is usually described with a surface roughness factor (the ratio of real surface area to geometric surface area) which can sometimes be higher than 2,000 [5]. Figure 1.2 helps to show the difference between the real surface area and the geometric surface area. Table 1.1 lists typical roughness factor ranges for common types of stimulating electrodes used in cardiac pacemakers.

Fig. 1.2

The porous (fractal) surface of TiN electrode enlarges the surface area; the fractal structure is visible at the bottom. The picture is from Multi Channel Systems GmbH, with permission

Table 1.1 Typical roughness factor ranges for common cardiac stimulation electrodes [5]

Fractal TiN, with a high electrochemical surface area, is applied extensively as coating for cardiac pacing electrodes, mainly because the electrode polarization during a pacing pulse is minimal [1]. However, under the high rate, high current density conditions of a neural stimulation pulse, access to all the available charge is limited by pore resistance [1]. The amount of the charge injected by a material before unwanted redox reactions occur is called charge injection capacity. A higher real surface area increases the double layer capacitance of the electrode. Therefore, the electrode-electrolyte capacitance voltage is smaller for the same amount of injected charge. This avoids occurrence of unwanted redox reactions, as they preferably occur at higher electrode-electrolyte interface voltage drops. For all porous electrodes, pore resistance exerts a geometric limitation on the increase in charge injection capacity that can be attained by increasing the “electrochemical surface area (ESA)”/“geometric surface area (GSA)” ratio. A schematic view of a pore and electrolyte in a porous electrode coating is illustrated in Fig. 1.3. The solution resistance and capacitance on the inner surface of the pore form a delay-line with a time-constant dependent on the pore geometry, electrolyte resistivity, and the interface double-layer capacitance. The result is that the total ESA of the electrode is not accessed at the current densities encountered during current and voltage edges. Narrower and deeper pores result in higher time-constants, and their charge-injection capacity is more difficult to access than that of electrodes with shallow pores. Both TiN and iridium oxide have higher measurable electrode-electrolyte capacitance per unit area with 0.5 ms pulses compared to 0.2 ms [6], because more real electrode area is accessible in lower frequencies and the diffusion rates of the charge carriers in the solution are limited.

Fig. 1.3

View of a pore profile illustrating the pore resistance (R1…R3) and double-layer capacitance (C1…C3) elements that make up a delay-line and time-constant for accessing all the ESA and its associated double-layer capacity, picture adapted from [1]

The surface roughness factor is experimentally determined by measuring the amount of chemicals that can be adsorbed onto the surface of the electrode [3]. As the porosity of an electrode increases, the real surface area increases and the surface impedance decreases. The real surface area of a porous electrode also increases when the thickness of the porous metal increases.

Another method to estimate the real surface area in materials like TiN which inject charge capacitively is cyclic voltammetry. Due to the intrinsic delays present in the topography of electrode surface structures (pores RC delays) explained above, the large interface capacitance of TiN measured with slow cyclic voltammetry is not available at fast rates of current injection [4].

Lower electrode impedances are favorable for both stimulating and recording electrodes. A certain amount of current density is required to initiate stimulation. A high electrode impedance will therefore mean high electrode voltages for a given current density which result in electrochemical reactions which may damage both the electrode and the tissue. In recording electrodes, the signals to be recorded are very small, in the order of millivolts to microvolts. The signal may be lost in the noisy ionic environment if the electrode impedance is not low enough [2]. Larger electrodes are preferred as the total resistance of the solution in the vicinity of the electrode there is smaller and therefore the thermal noise due to the resistance in series with the electrode double layer capacitance is decreased.

For extracellular signal recording, as the signal itself has almost no DC and contains frequencies between hundreds and thousands of Hertz, the electrode must feature a large interface capacitance, otherwise the electrode impedance is too high for this low frequency range. More porous electrodes are favorable.

Recording electrodes should be used in an input voltage amplitude range where they are linear, otherwise the nonlinearity would deform the shape of the recorded signal. In Chap. 5 it will be explained how to determine the linearity range of an electrode using a method called impedance spectroscopy.

1.1 Cyclic Voltammetry and Impedance Spectroscopy

Cyclic voltammetry (CV) uncovers a lot about the nature of charge transfer at the electrode surface. The standard measurement setup has three electrodes. The voltage of the working electrode (the electrode under test, here the stimulation or the recording electrode) is varied against a reference electrode. The reference electrode is used as a potential reference and sets the solution potential to a definite value. The result is a ramp voltage on the working electrode. The third electrode, the counter electrode, is used to close the circuit, i.e. ideally no current flows into the reference electrode. The current flowing from the working into the counter electrodes is measured. After a certain voltage limit is reached, the voltage is decreased till another lower limit is reached. This process may be repeated (i.e. the potential is cyclically varied). A cyclic voltammogram is the resulting current versus voltage difference measurement trace. The peaks in the graph correspond to different redox reactions. Also in case of purely capacitive charge injection, the amount of double layer capacitance can be assessed from the cyclic voltammogram. An example for cyclic voltammetry is shown for iridium oxide and TiN in Fig. 6.1.

The reference electrode is a very important component in CV. When a polarizable material like a metal is immersed in a solution, an electrode potential builds up at the electrode-electrolyte phase boundary at the thermodynamic equilibrium. This potential is due to the oxidation of the metal atoms as they turn into positive ions and enter the solution and the reduction of the ions as they are adsorbed on the metal surface. For example if a Zn electrode is immersed into ZnSO₄ solution, Zn atoms begin to solve in very small magnitudes into the solution and form Zn²⁺ ions:

$$\displaystyle{ Zn(s) \leftrightarrow Zn^{2+}(aq) + 2e^{-} }$$

(1.1)

The resulting electrons remain on the electrode. So a potential difference develops at the interface which is approximately 1 V.

Unfortunately it is not possible to measure this electrode potential in practice; therefore, the so called reference electrodes are used which feature a stable electrode potential. The reference electrode must be kept thermodynamically in equilibrium, i.e. very little or ideally no current must flow into it, otherwise the electrode potential deviates from what is expected from the equilibrium state. There are different kinds of reference electrodes like the standard hydrogen electrode and the silver-silver chloride (Ag | AgCl) electrode. Silver-silver chloride electrode is the most prevalent in the physiological studies and is shown in Fig. 1.4.

Fig. 1.4

Saturated Ag-AgCl reference electrode

In order to achieve a defined electrode-electrolyte potential difference (electrode potential), a solution with a defined concentration is contained inside a plastic or glass tube. The solution is usually chosen as a saturated chloride solution, e.g. KCl. In order to avoid the reference electrode from changing the main electrolyte environment, a limited electrolyte connection is available between the KCl solution and the main electrolyte through a porous plug or diaphragm at the electrode tip which is permeable for chloride ions. The corresponding reduction reaction is:

$$\displaystyle{AgCl(s) + e^{-}\leftrightarrow Ag(s) + Cl^{-}E^{0} = +0.23\,\mathrm{V}}$$

$E^{0} = +0.23\,\mathrm{V}$ is called the standard reduction potential which is measured against another reference electrode, the standard hydrogen electrode (SHE) which has by definition an electrode potential of zero. This definition is necessary as the electrode potential cannot be experimentally measured in practice. As the chloride solution is saturated (above 3.5 mol/kg concentration) and thus the concentration is not unity as in the standard condition, the saturated silver-silver chloride electrode has a potential of + 0. 197 V against SHE.

In some literature, instead of using a saturated chloride solution, another definite concentration is considered. With a 3 and 1 molal (mol/kg) KCL solution, the electrode potential is + 0. 21 V and + 0. 235 V versus SHE, respectively.

Due to the glass tube, the silver-silver chloride electrode can be put in the same solution vessel containing also the working and counter electrodes as shown in Fig. 1.5. If a reference electrode does not have this containing tube, it must be dipped in a separate solution vessel. This vessel is then electrically connected to the vessel containing the two other electrodes through a salt bridge.

The application of silver-silver chloride electrode is limited to outside the body, e.g. skin electrodes for ECG or EEG recording. In this case they are made wet by a chloride ion containing fluid or paste.

In order to perform CV a potentiostat is used. The basic three electrode potentiostat setup is shown in Fig. 1.5. V _IN is the input voltage from the potentiostat. As the reference electrode has a definite electrode potential, it is used to set the potential of the electrolyte. The counter electrode does not affect the solution potential and only carries current through redox reactions which are actually not a point of interest in CV. So carrying current and setting the solution potential are done by different electrodes.

Fig. 1.5

Three-electrode electrochemical cell

The three electrode system eliminates the dependence of the measured current on the counter electrode. If the amplifier has a very high amplification and zero input bias current, it can be shown that in the equivalent electrical circuit of the three cell setup shown in Fig. 1.6, $V _{IN} = V _{W} + V _{Ref}$. The current flowing into the working electrode is measured through the measurement resistor. The effect of the impedance of the counter electrode on the measured current is eliminated.

Fig. 1.6

The equivalent circuit of the three-electrode electrochemical cell

The same setup is used in impedance spectroscopy. Here, electrode impedance composed of amplitude and phase is measured by putting a sine voltage with varying frequency (V _IN in Fig. 1.6) on the electrode (a DC offset of V _Ref will exist at V _W ) and measuring the resulting current amplitude and phase shift over the frequency. The impedance spectrum of the electrode under test, i.e. Z _Work versus frequency of V _IN , is determined. An electrode model can be extracted from the resulting impedance data. It is usually composed of the prevalent linear components in electrical engineering like resistor and capacitor and components which are defined for the especial use in electrochemistry. One of these elements is introduced in Chap. 8, namely the constant phase element.

For monopolar stimulation arrangement with small microelectrode and a large distant counter electrode a two electrode system can be used for impedance spectroscopy. This will be investigated in Chap. 5.