The Effect of Physical Structural Properties on Electrochemical Properties of Ruthenium Oxide for Neural Stimulating and Recording Electrodes

Abstract

Recently, there has been a growing interest in ruthenium oxide (RuO_x) as an alternative mixed-conductor oxide to SIROF as an electrode coating. RuO_x is recognized as a faradic charge-injection coating with high CSC_c, long-term pulsing stability, and low impedance. We examined how the structural properties of sputter-deposited RuOx influence its electrochemical performance as an electrode coating for neural stimulation and recording. Thin film RuO_x was deposited under various pressures: 5 mTorr, 15 mTorr, 30 mTorr, and 60 mTorr on wafer-based planar test structures. Electrochemical characterizations, including electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV), and voltage transient (VT), were employed. The structure of RuO_x films was characterized by scanning electron microscope (SEM). Our findings revealed that the sputtering pressure significantly influences the growth of the RuO_x film, subsequently affecting its electrochemical performance. The results indicate that the electrochemical performance of RuO_x can be optimized by adjusting the deposition conditions to achieve a favorable balance between electronic and ionic conductivity.

I. Introduction

The use of chronically implanted multi-electrode arrays and devices for electrical stimulation and recording of neural activity is increasing, with anticipated applications in sensory restoration, paralysis treatment, epilepsy, and depression, among others [1] – [6]. The invasive nature of these arrays induces a tissue foreign body response (FBR), which is implicated in the ongoing deterioration of electrode performance during chronic implantation [7] – [10]. The introduction of microfabrication techniques has led to the miniaturization of these implants, reducing surgical insertion trauma, and increasing the spatial selectivity of neural stimulation and recording [8], [9]. In addition, miniaturized devices with ultramicroelectrode dimensions have been associated with a reduced chronic FBR, thereby improving the quality of neural recording and stimulation, as well as the longevity of the electrical neural interface [8] – [11].

However, a major technical challenge for these microscale implants is the compromise between the electrochemical performance of the electrode coatings and the physical size of the electrodes for neural recording and stimulation. As the physical size of the implants gets smaller, the size of the electrodes will decrease, which inevitably results in higher impedance, and lower charge storage and injection capacity. Thus, it is necessary to optimize electrode coatings to achieve high charge injection capacity for neural stimulation and low impedance for neural recording while maintaining a relatively small size.

Recent studies have investigated the feasibility of ruthenium oxide (RuO_x) as an alternative electrode coating for neural recording and stimulation [12], [13]. In the present study, we further investigated the relationship between the physical structure of RuO_x electrodes, which is characterized by scanning electron microscopy (SEM), and their charge storage (CSC_c) and stimulation charge injection capacity (Q_inj) using electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV), and constant current pulsing. The objective is to establish structure-property relations that enable the selection and optimization of electrode coatings for electrical microstimulation.

II. Methods

A. Test Structure Fabrication

Wafer-based test structures were fabricated using microfabrication techniques in the cleanroom at the University of Texas at Dallas. A test-grade, single-side polished silicon wafer underwent thermal oxidization in a Tystar Diffusion/Oxidation Furnace to grow a 1 μm SiO₂ layer. Subsequently, a 2-μm layer of amorphous silicon carbide (a-SiC) was deposited on top of the SiO₂ via plasma-enhanced chemical vapor deposition (PECVD), serving as the bottom encapsulation layer. The a-SiC was deposited at a substrate temperature of 300 °C, 200 W RF power, and a chamber pressure of 800 mTorr, with 164 sccm Ar, 600 sccm 2% SiH₄ diluted in Ar, and 36 sccm CH₄ as the reactive gases [16]. A tri-layer Ti-Au-Ti metal layer, with thicknesses of 50-200-50 nm, respectively, was sputtered using a customized AJA 2200 magnetron sputtering system. Next, this tri-layer underwent patterning through photolithography processes and a chemical wet etch process to create the contacting pads and metal traces for the testing structure. A second 2-μm layer of a-SiC was deposited over the bottom a-SiC and metal traces to form the top encapsulation layer. The electrode sites and contacting pads were patterned by photolithography and etched by an SF₆ plasma based reactive ion etching to remove the top layer a-SiC. The geometric surface area (GSA) of the electrode sites ranged from 400 μm² to 4000 μm². RuO_x films with Ti adhesion layers were then sputtered using reactive DC magnetron sputtering from metal targets in an AJA 2200 magnetron sputtering system. The RuO_x films were sputtered with 20 sccm Ar, 7.5 sccm O₂, 22.5 sccm H₂O and 100 W DC power (power density 4.93 W/cm²) at 5 mTorr, 15 mTorr, 30 mTorr, and 60 mTorr. The deposition time varied for RuO_x sputtered at four different pressures. The RuO_x film was patterned using a bi-layer photoresist-based lift-off process. Film thickness was measured using scanning electron microscopy (SEM) of fractured cross-sections.

B. Electrochemical Characterization

Electrochemical characterization was conducted at 37 °C in Ar-sparged pH balanced (pH = 7.4) phosphate buffered saline (PBS). Electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV) were carried out using a Gamry Instruments 600+ potentiostat in a 3-electrode configuration with a Pt counter electrode and Ag|AgCl reference electrode. EIS measurements were taken between 0.05 to 10⁵ Hz, employing a 10 mV RMS sinusoidal voltage signal at 10 data points per decade at the open circuit potential (E_oc) of RuO_x. CVs were recorded over a potential range of +0.5 V and −0.5 V vs. Ag|AgCl at sweep rates of 50 mV/s and 50,000 mV/s. This potential range is a conservative approximation of the electrolysis limit of water for RuO_x at pH = 7.4. Cathodal charge storage capacity (CSC_c) was determined from the CV by integrating the cathodal current over time and normalizing it by the geometric surface area (GSA).

Voltage transients (VT) were performed using a PlexStim stimulator with symmetrical biphasic constant current pulsing at a 200 μs pulse width per phase at 50 Hz. The charge injection capacity (Q_inj) was calculated from the VT by integrating the cathodal current pulse and normalizing it by the GSA. The RuO_x films were polarized to the maximum cathodic potential −0.7 V vs. Pt (−0.5 V vs. Ag|AgCl) to avoid water electrolysis. This value is determined by 50 mV/s CV.

III. Results

Fig. 1 shows fracture cross-section SEM images of RuO_x films as a function of sputtering pressure. The columnar growth of the RuO_x, which approximates Zone 1 of structure zone models, changes as the sputtering pressure increases [17]. From the SEM images, we observed that RuO_x deposited at 5 mTorr (Fig. 1a) exhibits a more compact structure compared to films deposited at 15 mTorr (Fig. 1b), 30 mTorr (Fig. 1c), and 60 mTorr (Fig. 1d). The columnar structure of RuO_x is more pronounced as the deposition pressure increases, which increases the electrochemical surface area of electrolyte by promoting voids between the columns.

Figure 1.

The fractured cross-sectional SEM image of RuO_x film sputtered deposited on a-SiC at various sputtering pressures (a) 5 mTorr (b) 15 mTorr (c) 30 mTorr (d) 60 mTorr. White arrow shows the RuO_x layer. Yellow arrow shows the Ti adhesion layer. The scale bar is 100 nm.

Table I provides the deposition time, deposition rate, and film thickness for each deposition pressure. The deposition rate of RuO_x films deposited at different pressures was different: 8.4 nm/min for 5 mTorr films, 6.8 nm/min for 15 mTorr films, 3.5 nm/min for 30 mTorr films, and 0.56 nm/min at 60 mTorr.

A. Electrochemical Impedance Spectroscopy

Fig. 2 presents a representative EIS spectra of 2000 μm² RuO_x electrodes deposited at 5 mTorr, 15 mTorr, 30 mTorr, and 60 mTorr. Notably, the RuO_x deposited at 5 mTorr exhibits higher impedance across the entire frequency range compared to those deposited at higher pressures, despite having the highest thickness (380 nm) of the films tested. We attribute the high impedance to lower porosity in the low deposition pressure film, as suggested by the SEMs in Fig. 1a.

Figure 2.

Representative impedance spectra of RuO_x deposited at different pressure. Measurement taken at 10 mV RMS and open circuit potential.

The 15 mTorr RuO_x film has a similar impedance magnitude as the 30 mTorr RuO_x film. This observation suggests that increased porosity enhances the accessibility of RuO_x to the electrolyte, leading to a decrease in impedance despite an approximate difference of 50 nm in thickness. Interestingly, the 60 mTorr film exhibits higher impedance compared to the 15 mTorr (310 nm) and 30 mTorr (230 nm) films, likely due to the much lower thickness (153 nm) of the 60 mTorr film.

B. Cyclic Voltammetry

Representative CVs and CSC_c of RuO_x electrodes deposited at different pressures are shown in Fig. 3. At a sweep rate of 50 mV/s (Fig. 3a), the CVs for 30 mTorr and 60 mTorr RuO_x electrodes are similar and indicate normal accessibility of electrochemical surface area in the porous films. In contrast, the 5 mTorr RuO_x displays a narrow and tilted CV, likely due to the highly packed nature of the film that reduces the available electrochemical surface area, despite the film having the highest thickness. The tilt in the 5 mTorr CV likely results from the resistive nature of the film due to its high density.

Figure 3.

(a) Representative 50 mV/s sweep rate cyclic voltammograms of 2000 μm² RuO_x deposited at different pressures at, (b) 50,000 mV/s sweep rate. (c) CSC_c of RuO_x electrodes with different GSA. Electrolyte: Ar-sparged PBS at 37°C.

After examining the CVs at a sweep rate of 50,000 mV/s (Fig. 3b), the 5 mTorr film still displays a narrow CV. Notably, the 30 mTorr film exhibits a larger CV than the 60 mTorr film at 50,000 mV/s, suggesting the involvement of a time constant in the accessibility of the internal area within the film. The film deposited at 15 mTorr (310 nm) produces the widest CV at both sweep rates, indicating the highest CSC_c among all tested films. CVs on electrodes with various sizes yield similar results to the representative CVs. Therefore, when calculated from the CVs at a sweep rate of 50 mV/s (Fig. 3c), we observe similar CSC_c values for 30 mTorr (33.9 ± 2.0 mC/cm², n = 10) and 60 mTorr RuO_x (31.6 ± 2.8 mC/cm², n = 8) electrodes, while 15 mTorr RuOx electrodes exhibit the highest CSC_c (53.70 ± 2.3 mC/cm², n = 20). It is noteworthy that the CSC_c of RuO_x electrodes sputtered at the same pressure does not vary significantly with the electrode’s GSA.

C. Voltage Transient

Fig. 4 shows representative VTs and Q_inj of RuO_x electrodes deposited at different pressures. When polarized to the maximum cathodic potential to avoid water electrolysis (−0.7 V vs. Pt), indicated by the crosshair in Fig. 4b, the Q_inj values of 15 mTorr, 30 mTorr, and 60 mTorr RuO_x electrodes are 3.5 ± 0.2 mC/cm² (n = 20), 3.0 ± 0.1 mC/cm² (n = 10), and 1.6 ± 0.2 mC/cm² (n = 8), respectively. The Q_inj of 15 mTorr and 30 mTorr RuO_x is comparable with SIROF [18] and higher than TiN [19]. Despite 30 mTorr and 60 mTorr RuO_x films exhibiting very similar CSC_c, as shown in Fig. 3c, we observed that the 60 mTorr film has a significantly lower Q_inj than the 30 mTorr film, as demonstrated in Fig. 4b and 4c. Meanwhile, the 30 mTorr RuO_x film displays a Q_inj similar to that of 15 mTorr RuO_x. Similar to CSC_c value, the Q_inj of RuO_x electrodes sputtered at the same pressure does not vary significantly with the electrode GSA.

Figure 4.

(a) Representative current waveform for 400 μm² RuO_x voltage transient, (b) representative voltage response for 400 μm² RuO_x voltage transient, (c) Q_inj of RuO_x electrodes with different GSA. Electrolyte: Ar-sparged PBS at 37°C.

IV. Discussion and Conclusion

We demonstrated the ability to produce RuO_x with different structures by varying chamber pressures for each sputtering. At all sputtering pressures, RuO_x exhibits the porous columnar growth morphology described in Zone 1 of structure zone models [17]. With an increase in deposition pressure from 5 mTorr to 60 mTorr, the deposition rate decreases, with the development of a more pronounced columnar structure as pressure increases. This enhancement is observable in EIS, CV, and VT. The reported values of CSC_c and Q_inj are lower than previously reported data by Chakraborty et al. [12], [13], likely due to a more conservative scan range and a conservative maximum cathodic polarization in CV and VT measurements, respectively.

As the RuO_x films become more porous with increasing sputtering pressure, the accessibility of internal surface area should increase and the impedance of the RuO_x electrodes should decrease. However, the 60 mTorr RuO_x electrodes exhibit higher impedance than both the 15 mTorr and 30 mTorr RuO_x electrodes. We attribute this observation to the difference in film thickness, although multiple factors may contribute to the impedance of the RuO_x coatings.

As indicated by the 50 mV/s CVs, 30 mTorr and 60 mTorr RuO_x have very similar CSC_c (shown in Fig. 3c), even though 60 mTorr film is roughly 100 nm thinner than the 30 mTorr film. This suggests that higher porosity improves the accessibility of internal surfaces to the electrolyte, enhancing the CSC_c. Interestingly, the film deposited at 15 mTorr (310 nm) has the highest CSC_c at both sweep rates. It remains unclear whether the higher CSC_c of the 15 mTorr film is due to electronic and ionic kinetics within the film or the difference in film thickness.

Meanwhile, the deviation of 30 mTorr and 60 mTorr CVs at 50,000 mV/s, and voltage transient results suggest that electronic conductivity may be hindered as the film density decreases. It’s noteworthy that VTs and 50,000 mV/s CVs are conducted at a much higher current density than 50 mV/s CVs. At higher current density, the electrochemical and charge injection performance of RuO_x may be limited by electronic transport within the film. These results indicate that an optimized film should balance ionic and electronic conductivity to maximize CSC_c and Q_inj. We are currently investigating the effect of RuO_x thickness and sputter pressure to identify conditions that maximize Q_inj. We are also conducting long-term saline pulsing studies to assess the stability of RuO_x films exhibiting a high Q_inj.

TABLE I.

RuO_x Deposition Time, Rate, and Film Thickness

Pressure (mTorr)	Time (min)	Rate (nm/min)	Thickness (nm)
5 (Fig.1a)	45	8.4	380
15 (Fig.1b)	45	6.8	310
30 (Fig.1c)	66	3.5	230
60 (Fig.1d)	270	0.56	153

Clinical Relevance—

This research underscores the potential for optimizing the structural properties of RuO_x to enhance its electrochemical capabilities for neural stimulation and recording.