WO2013014206A1 - Method and device for stimulation/recording of cells - Google Patents

Abstract

In order to reliably integrate vertically-aligned carbon nanotubes (CNTs) as an electrode material in future high-resolution implantable neural probes, proper embedding of the CNT electrodes is essential to maintain their mechanical integrity during implantation. A new approach is described to fabricate functional CNT microelectrodes encased in a protective coating stack comprising Si0₂ and Parylene C with lithographically defined electrode openings. The vertically-aligned CNTs were grown on passive arrays of individually-addressable microelectrodes using low-temperature (425°C) plasma enhanced chemical vapor deposition compatible with wafer-scale Cu-CMOS processing.

Description

Method and Device for Stimulation/Recording of Cells

This present invention is situated in the field of implantable devices for the stimulation and recording of cells. More in particular it is situated in the field of implantable neural microsystems.

Background of the invention

Electrical recording and stimulation of the central nervous system is used in basic neuroscience to study brain functions and in clinical practice to treat conditions such as epilepsy, Parkinson's disease, and chronic pain. Device engineers are increasingly confronted with demands from neuroscientists and clinicians alike to develop implantable microsystems that may provide both stimulation and recording capabilities on one platform. Resolutions down to cellular (10-50 um) and even subcellular dimensions (<10 μιτι) should allow bidirectional interaction with the brain at different levels of neural organization. Ultimately, dense arrays with μιτι-sized electrodes for parallel and high-fidelity recording and selective stimulation of single neurons or neuronal populations are envisaged. Eventually, such strict device specifications can only be met by using advanced Cu-CMOS technology capable of addressing large electrode arrays in a reliable way.

Commonly used thin-film materials, including Pt, Ir, and TiN, face serious limitations in terms of their potential for further electrode miniaturization. The electrode impedance, which scales inversely with the electrochemical interface capacitance, may be too high - on the order of several MOhms - for small electrodes based on thin-film materials to offer sufficient recording sensitivity. In addition, the high driving voltages required to supply sufficient stimulation charge with small electrodes may damage the electrode and surrounding tissue. Solutions to mitigate these limitations aim at increasing the effective surface area (and hence capacitance) of the electrodes by coating them with rough or porous materials such as Pt black, iridium oxide (IrO,), conducting polymers, and recently also carbon nanotubes (CNTs). Pt black suffers from poor mechanical stability, while IrO, and conducting polymers, such as poly(3,4-ethylenedioxythiophene) and polypyrrole, can degrade under electrical stimulation leading to impedance fluctuations and loss of charge injection capacity. In contrast, the long-term electrode viability is expected to improve significantly by using CNTs, which are chemically inert and stable against degradation under prolonged potential cycling. They furthermore exhibit excellent electrical conductivity and, most importantly, biocompatibility towards neurons.

Recent studies have demonstrated that CNT-coated microelectrodes exhibit superior performance compared to uncoated ones. These reports describe various methods to deposit the CNTs: chemical vapor deposition (CVD), drop-coating, pyrolysis of acetylene, and microcontact printing. Among these, only CVD can readily and reliably be applied at wafer-scale and provide subcellular electrode resolution. However, conventional thermal CVD of CNTs requires temperatures far in excess (»450°C) of Cu-CMOS-compatible processing. Driven by the need to replace standard Cu-based interconnect technology, wafer-scale CVD- synthesis of CNTs has been optimized to satisfy the temperature constraints of Cu-CMOS processing. The resulting CNTs exhibit excellent electrical and structural properties and hold much promise regarding their potential applications in various fields. While low-temperature CVD at 400°C has already been employed to grow CNTs on microelectrode arrays, a viable process integration scheme for CNTs on implantable neural probes is still lacking.

The major challenge in this regard relates to the mechanical stability of the CNTs during implantation in tissue such as e.g. brain tissue. Due to their vertical structure, CVD-grown CNTs may be susceptible to lateral abrasive forces during probe implantation, leading to damaged or collapsed CNTs. The latter is problematic for dense electrode arrays where collapsed CNTs may form electric shorts across neighboring electrodes. In the usual top-down approach described for microelectrode arrays, CNTs are selectively grown in electrode openings defined in the dielectric that covers the interconnects. Here, the CNT growth can be the last fabrication step, which is advantageous for preserving the as-grown CNT properties. The major drawback of this approach, however, is that the catalytic metal film required for the CNT growth cannot reliably and selectively be deposited within electrode openings of subcellular dimensions. Consequently, the wafer surface will in part or wholly be coated with the potentially toxic metal catalyst and eventually a thin sheet of amorphous carbon. While in vitro neuronal survival of a few days has been shown on such structures, it is unclear whether catalyst traces will not leach into the physiological medium or tissue over prolonged immersion times.

Summary of the invention

In a first aspect of the present invention, an appropriate embedding and lateral confinement of the CNTs helps to overcome the problem of lateral abrasive forces during probe implantation which could damage the CNTs.

An electrode for stimulation and recording of tissue, e.g. brain tissue, is presented.The electrode comprises:

a substrate;

an insulating layer, e.g. an oxide layer such as a Si0₂ layer, atop the subtrate;

- an electrode layer atop the insulating layer, the electrode layer forming one or more electrodes;

a first layer located atop the electrode layer;

a plurality of carbon nanotubes atop the first layer;

a second layer atop the insulating layer and atop the plurality of carbon nanotubes (5); and an opening (8) in the second layer for providing access to the plurality of carbon nanotubes (4).

In an embodiment of the first aspect of the present invention, a hermetic electrode for stimulation and recording of tissue is presented. The electrode comprises:

a substrate;

- an insulating layer, e.g. an oxide layer such as a Si0₂ layer, atop the subtrate.;

an electrode layer atop the insulating layer, the electrode layer forming one or more electrodes;

a first layer atop the electrode layer;

a plurality of carbon nanotubes atop the first layer;

a second layer which is located atop the insulating layer and atop the plurality of carbon nanotubes; - atop the second layer, a third layer, e.g. a biocompatible layer such as a layer comprising or consisting of Parylene C or SiN or any other suitable material that is resistang to sodium ions and moisture,; and an opening in the second layer and the third layer for providing access to the plurality of carbon nanotubes.

It is an advantage of the third layer that it protects the second layer from disintegration, at least when the electrode is implanted.

In a second aspect of the present invention, a implantable device is presented. The device comprises:

a shaft;

- a tip;

- a plurality of carbon nano tube electrodes for cell stimulation and recording in accordance with the first aspect of the present invention being located on the shaft; and

a plurality of bondpads connected to the plurality of cell stimulation electrodes via a plurality of interconnections.

In an embodiment of the second aspect, the implantable device may comprise additional (packaged) circuitry to drive the bondpads.

The implantable device may be a neural probe for electrical stimulation and recording of brain tissue.

In a third aspect of the present invention, the fabrication of individually addressable passive microelectrodes of cellular and subcellular dimensions coated with vertically-aligned CNTs is presented.

A method is presented to manufacture an electrode for stimulation and recording of tissue as described in the first aspect of the present invention. The method comprises:

providing a substrate;

providing, e.g. creating, an insulating layer, e.g. an oxide layer such as a Si0₂ layer, atop the subtrate; providing, e.g. creating, an electrode layer atop the insulating layer, the electrode layer forming one or more electrodes;

providing, e.g. creating, a first layer atop the electrode layer;

growing atop the first layer, a plurality of carbon nanotubes;

- providing, e.g. creating, atop the insulating layer and atop the plurality of carbon nanotubes a second layer; and

providing, e.g. creating, an opening in the second layer for providing access to the plurality of carbon nanotubes.

In an embodiment of the third aspect of the present invention, a method is provided to manufacture a hermetic electrode for tissue stimulation and recording as described in the first aspect of the present invention. The method comprises:

providing a substrate;

providing, e.g. creating, an insulating layer atop the subtrate;

providing, e.g. creating, an electrode layer atop the insulating layer, the electrode layer forming one or more electrodes;

providing, e.g. creating, a first layer atop the electrode;

growing atop the first layer, a plurality of carbon nanotubes;

providing, e.g. creating, atop the insulating layer and atop the plurality of carbon nanotubes a second layer;

- providing, e.g. creating, atop the second layer a third layer;

creating an opening in the third layer at the location of the carbon nanotubes; and

creating an opening in the second layer for providing access to the plurality of carbon nanotubes.

The third layer provides a hermetic layer to the electrode device. The third layer is preferably made of a biocompatible material such as SiN, Parylene C or any other suitable material that is resistant to to sodium ions and moisture. The third layer protects the second layer from disintegration.

The second layer serves also as a protection layer for the CNTs during the plasma etch of the third (hermetic) layer for creating the opening. The second layer may be a Si0₂ layer or another suitable material.

Brief description of the drawings

The present invention will now be described further, by way of example, with reference to the accompanying drawings, in which:

FIG. 1 is a schematic of a process flow for the CNT (g), TiN (g⁷) and Pt (g") electrode arrays;

FIG. 2 (a) is an optical micrograph of an electrode array after CNT growth on top of the electrodes; FIG. 2(b) is a cross-sectional scanning electron micrograph (SEM) image of the as-grown CNTs; FIG. 2(c) is a SEM image of an electrode and FIG.2(d) is a close-up of the same after conformal CVD Si0₂ coating;

FIG. 2(e) is an overview SEM image o an electrode after etching of the Parylene C and removal of the CVD Si0₂ in buffered HF;

FIG. 2(f) is an overview SEM image of a typical electrode edge partially encased in Parylene C;

FIG. 2(g) is a combined FIB/SEM image showing details of the final CNT-Parylene C interface at an electrode opening;

FIG. 2(h) is a combined FIB/SEM image showing details of the final CNT-Parylene C interface at the outermost electrode edge;

FIG.3 illustrates cyclic voltammograms of Ft, TiN and CNT electrodes of 5 mm diameter;

FIG.4 illustrates average Bode plots with impedance magnitude, | Z | , and phase angle, q, for the Pt, TiN and CNT electrodes of 25, 10 and 5 mm diameter measured between 1-105 Hz in phosphate buffered saline;

FIG.5(a) illustrates impedance magnitude (with standard deviation) at 1 kHz a a function of the electrode diameter for the Pt, TiN and CNT electrodes;

FIG.5(b) illustrates an equivalent circuit to fit the impedance spectra;

FIG.6 is a graph illustrating the baseline-corrected transmittance infrared spectra of (a) as-grown CNTs, (b) of Si0₂-coated CNTs after a BHF treatment, and (c) of CNTs after BHF treatment. The corresponding contact angles for the three substrates are shown on the righthand side;

FIG.7 illustrates SEM images of primary rat hippocampal neurons 5 days in vitro on fabricated CNT substrates;

FIG.8 schematically illustrates a structural CNT electrode according to embodiments of the present invention;

FIG.9 schematically illustrates a structural and hermetic CNT electrode according to embodiments of the present invention;

FIG. 10 illustrates an implantable probe for stimulation/recording of cells according to embodiments of the present invention; and

FIG. 11 illustrates implementations of probes.

The drawings are only schematic and are non-limiting. In the drawings, the size of some of the elements may be exaggerated and not drawn on scale for illustrative purposes. The dimensions and the relative dimensions do not necessarily correspond to actual reductions to practice of the invention.

Any reference signs in the claims shall not be construed as limiting the scope. In the different drawings, the same reference signs refer to the same or analogous elements.

Detailed Description of illustrative embodiments

The present invention will be described with respect to particular embodiments and with reference to certain drawings but the invention is not limited thereto but only by the claims.

The terms first, second and the like in the description and in the claims, are used for distinguishing between similar elements and not necessarily for describing a sequence, either temporally, spatially, in ranking or in any other manner. It is to be understood that the terms so used are interchangeable under appropriate circumstances and that the embodiments of the invention described herein are capable of operation in other sequences than described or illustrated herein.

Moreover, the terms top, under and the like in the description and the claims are used for descriptive purposes and not necessarily for describing relative positions. It is to be understood that the terms so used are interchangeable under appropriate circumstances and that the embodiments of the invention described herein are capable of operation in other orientations than described or illustrated herein.

It is to be noticed that the term "comprising", used in the claims, should not be interpreted as being restricted to the means listed thereafter; it does not exclude other elements or steps. It is thus to be interpreted as specifying the presence of the stated features, integers, steps or components as referred to, but does not preclude the presence or addition of one or more other features, integers, steps or components, or groups thereof. Thus, the scope of the expression "a device comprising means A and B" should not be limited to devices consisting only of components A and B. It means that with respect to the present invention, the only relevant components of the device are A and B.

Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Thus, appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment, but may. Furthermore, the particular features, structures or characteristics may be combined in any suitable manner, as would be apparent to one of ordinary skill in the art from this disclosure, in one or more embodiments.

Similarly it should be appreciated that in the description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. This method of disclosure, however, is not to be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.

Furthermore, while some embodiments described herein include some but not other features included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the invention, and form different embodiments, as would be understood by those in the art. For example, in the following claims, any of the claimed embodiments can be used in any combination.

It should be noted that the use of particular terminology when describing certain features or aspects of the invention should not be taken to imply that the terminology is being re-defined herein to be restricted to include any specific characteristics of the features or aspects of the invention with which that terminology is associated.

In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the invention may be practiced without these specific details. In other instances, well- known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.

Methodology

For bottom-up embedding according to embodiments of the present invention, which has not yet been demonstrated for CNT microelectrodes, an insulation layer is deposited after the CNT growth, and electrode openings are defined lithographically or possibly also by chemo-mechanical polishing.

It is an advantage of embodiments of the present invention that by depositing an insulation layer afther the CNT growth, the metal catalyst on the wafer surface is covered, thereby creating a biocompatible surface and thereby prolonging the viability of cells. The major challenges in this approach are avoiding any damage and/or chemical modifications to the CNTs during the electrode opening and maintaining the biocompatibility of the CNTs throughout the fabrication process. For the latter, exposure of the CNTs to potentially toxic organic compounds, including organic substances such as photoresists and solvents, and metal contaminants must be prevented as they may become entrapped within the CNT matrix. As an additional advantage of embodiments of the present invention, the insulation material serves a protective purpose whereby contamination of theCNTswith toxic materials during fabrication is avoided.The insulation prevents the contamination of the CNTs with toxic materials such as photoresist or metal traces during fabrication. In an aspect of the present invention, the fabrication of individually addressable passive microelectrodes of cellular and subcellular dimensions coated with vertically-aligned CNTs is demonstrated. In view of the long-term objective to incorporate CNTs in neural implants with high-density electrode arrays, preserving the mechanical integrity of the CNTs during prospective implantation becomes a key objective. Therefore, a bottom-up approach is presented to encase the vertical CNT microelectrode for improved mechanical stability.

A final device as obtained in accordance with embodiments of the present invention is schematically illustrated in FIG.8 and FIG.9. Both embodiments illustrate a CNT electrode in accordance with embodiments of the present invention, comprising a substrate 1, e.g. a Si substrate, an insulating layer, e.g an oxide such as S1O2, an electrode layer 3 which may be suitably patterned so as to comprise one or more electrodes, a seed layer 4 to grow CNTs, e.g. a TiN or TiN/Ni layer, and a plurality of CNTs 5 grown on the seed layer 4. A dielectric layer 6, such as an oxide layer, e.g. a Si0₂ layer, is provided aside and optionally partially atop the CNTs 5. Such dielectric layer 6 protects the CNTs 5 against mechanical impacts when introducing the electrodes into tissue for stimulation and recording. The embodiment illustrated in FIG.9 furthermore shows, on top of the dielectric layer 6, a protective layer 7. This protective layer 7 preferably is made from biocompatible material, and protects the dielectric layer 6 from disintegration when implanted and in contact with sodium ions and moisture. In both embodiments, an opening 8 is provided, in the layer 6 and optionally, when present, in the layer 7, for providing access to the CNTs.

It is an advantage of embodiments of the present invention that the CNTs are protected by the encasing formed by the dielectric layer 6 and optionally the protective layer 7. The encasing provides protection for the CNTs. As an additional advantage, the encasing provides a solution for inserting the CNTs in tissue. In accordance with embodiments of the present invention, this was achieved using a coating stack comprising a dielectric layer such as e.g. S1O2, and a protective layer such as Parylene C, with lithographically defined electrode openings. The CNTs were grown using low-temperature (425 C) plasma-enhanced CVD (PECVD) optimized for back-end-of-line (BEOL) Cu-CMOS processing.

The CNT electrodes were characterized by cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) and benchmarked against co-fabricated Pt and TiN electrodes. The impact of processing on the CNT functional chemistry and wettability was analyzed by Fourier-transform infrared spectroscopy (FTI ) and contact angle measurements, respectively. Finally, the biocompatibility of the fabricated CNT substrates was demonstrated using primary rat hippocampal neurons and scanning electron microscopy (SEM).

FIG. 10 illustrates an implantable probe for stimulation and recording of cells according to embodiments of the present invention. Such implantable probe comprises a tip 13 for introducting the probe into tissue of which cells are to be stimulated and of which cell activity is to be recorded. This tip 13 is provided at an extremity of a shaft 10 comprising at least one, preferably a plurality of, CNT cell stimulation and recording electrodes 12 in accordance with embodiments of the present invention. The shaft 10 may be connected to a wider portion of the probe, which is not to be implanted, and which is provided with bondpads 14 for making electrical connections to other pieces of circuitry, either integrated on the probe or external thereto. Electrical interconnections 11 are provided on the probe for electrically interconnecting the CNT electrodes 12 with the bondpads 14.

FIG. 11 illustrates implementations of such probes. FIG. 11 (c) illustrates a probe comprise additional packaged circuitry 15 for driving the bondpads 14.

Experimental Results

In an embodiment of the present invention, microelectrode arrays were fabricated on semiconductor wafers 1, e.g. Si wafers, such as 200 mm Si, wafers, according to the scheme depicted in FIG. 1. Briefly, for all investigated electrode materials, interconnects 3 were defined by lift-off of sputter- deposited Pt (FIG. la). Electrode arrays based on TiN 4 and CNTs 5 had an additional TiN layer on top of the electrode areas of the interconnects 3, the additional TiN layer being patterned by means of a second lift-off step (FIG. lb). Following reported procedures, vertically-aligned CNTs 5 were selectively grown on the electrodes by a suitable low-temperature process (such as below 600°C, e.g. 425 C), for example a PECVD process using a 2 nm (nominal) thick Ni layer 4 as catalyst (FIG. lc and FIG. 2a). The resulting multiwalled CNTs had a homogeneous height of ~ 2 μιτι, an average diameter of 34 nm (determined by SEM), and a density of approximately 2-10¹¹cm^"2 (FIG. 2b). This CNT height was opted for in order to avoid too high topographies on the wafer surface and hence to ensure a reliable lithography. The CNTs 5 were embedded in a dielectric layer 6, e.g. a si0₂ layer, such as a 300 nm Si0₂ layer, to serve as a protection and etch stop layer during subsequent fabrication steps. In the experiments, the Si0₂ was deposited by CVD at 150 C (FIG. Id) and formed a conformal coating on top of the CNTs as seen from the SEM images in FIG.2c and FIG. 2d.

In an embodiment of the present invention, an additional biocompatible layer 7 was provided, e.g. evaporated on top of the dielectric layer 6. As an example biocompatible and hermetic insulation material, 1 um of Parylene C was evaporated (FIG. le). The bond pads 9 were opened by suitable methods, e.g. reactive ion etching (RIE) of the biocompatible layer 7, e.g. Parylene C, and the dielectric layer 6, e.g. Si0₂ (FIG. If). In a second RIE step (FIG. lg), the biocompatible layer 7, e.g. Parylene C, on the electrode areas was opened. The etch time was adjusted to stop on top of the dielectric layer 6, e.g. oxide layer, and thus prevent damage to the embedded CNTs. The corresponding stack profiles for the TiN and Pt electrodes are shown in FIG. lg") and FIG. lg"), respectively. For the fabrication of the Ft electrodes, steps b), c) and d) of FIG. 1 were omitted. For the fabrication of theTiN electrodes, steps c) and d) were omitted.

After providing access to the electrodes, the wafers were diced and the chips were wirebonded onto custom printed circuit boards. Prior to experimentation, the Si0₂ layer on top of the CNT electrodes was removed by dipping the packaged chips in buffered hydrofluoric acid (BHF). The SEM image in FIG. 2e displays an electrode after RIE of Parylene C and removal of the Si0₂ in BHF. Due to this wet treatment and subsequent drying of the chips, capillary action caused a clustering of the CNTs, resulting in the formation of dense microbundles. A typical electrode edge partly encased in Parylene C is shown in the overview SEM image in FIG. 2f. As an advantage of embodiments of the present invention, it can clearly be seen that the Parylene C encasement is higher than the CNTs and should hence provide sufficient protection of the CNT electrodes from abrasive forces that may arise during implantation. Focussed ion beam (FIB) combined with SEM was employed to analyze the details of the CNT-Parylene C interface at an electrode edge. A cross-sectional FIB/SEM image of the final electrode edge shows the CNT-Parylene C interface at the electrode opening (FIG. 2g). Dotted lines indicate the different layers, which are (from top to bottom): Parylene C, CNTs, TiN, Ft, and thermal Si0₂. Also clearly visible are larger voids underneath the Parylene C stemming from the removal of the CVD Si0₂ by BHF. These voids are also discernible in the cross-sectional FIB/SEM image in FIG. 2h, which displays the CNT-Parylene C interface at the outermost electrode edge. Here, the indicated layers are (from top to bottom): CVD Si0₂ and thermal Si0₂. Clearly, the BHF etch time was sufficient to remove the CVD Si0₂ on top of the CNTs and preserve the Si0₂ stack outside the perimeter of the electrode. This aspect is essential for maintaining an optimal insulation of the metal interconnects. From the FIB/SEM analysis it also becomes clear that the embedding and subsequent electrode opening did not lead to evident morphological changes of the CNT forest.

All electrode materials were characterized electrochemically by cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) in phosphate buffered saline (PBS). Cyclic voltammetry is generally used to identify the nature and extent of faradaic and non-faradaic electrode processes and to determine the material-specific voltage limits for hydrogen and oxygen evolution, also termed the water window. During neural stimulation it is advised to avoid the occurrence of faradaic reactions in general and gas evolution in particular because they may damage the electrode and/or release chemical species which may alter the tissue beyond physiologically tolerable levels. The charge density that an electrode can supply within the water window is the charge-storage capacity (CSC). It corresponds to the area under the CV curve divided by the voltage sweep rate. Mostly, only the cathodic branch of the CV curve is taken into consideration to account for the fact that cathodic neural stimulation is preferred over anodic stimulation because cathodic activation thresholds are lower. This cathodic CSC (CSC,) is frequently employed to estimate the ability of an electrode material to supply sufficient charge for neural excitation. It has to be noted, however, that voltage limits determined by CV may differ significantly from the electrode polarization that is tolerable under stimulation. The reason is that the transient response of electrodes under stimulation is characterized by a highly nonlinear voltage-time behavior with a differential "sweep rate" of more than 1 kV/s. It is well-known that the (over)voltages determined by CV depend on the sweep rate and hence the kinetic and diffusional dynamics of the underlying electrode processes. Translating electrode (over)voltages obtained under CV to the case of stimulation is therefore strictly speaking not correct. Nonetheless, the CSC, is a useful figure of merit to compare the charge-storage capability of various materials under well-defined conditions.

FIG. 3 shows typical voltammograms obtained with Pt, TiN, and CNT electrodes of 5 μιτι diameter. Scans were performed at 0.1 V/s in phosphate buffered saline. For better clarity. For better clarity, a vertical offset is introduced for the Ft curve (right y-axis), and the values of the TiN curve (left y-axis) have been multiplied by 50. The Pt voltammogram exhibits the typical features of H adsorption and desorption between - 0.6 V and -0.3 V and a broad Pt oxidation band between 0.25-0.9 V. The large current step in the cathodic sweep between 0-0.2 V is due to the reduction of dissolved 0₂. Electrodes based on TiN show a strong blocking behavior in the potential region -0.5-1.1 V where current flow is mostly capacitive in nature. In the anodic scan, two large bands centered around -0.85 V and -0.4 V can be observed and are attributed to the oxidation of H₂ gas and chemisorbed H, respectively. Due to the lack of distinct H adsorption features in the cathodic scan, the onset potential for H₂ evolution cannot be identified. Thus, this is set at -0.75 V for the TiN electrodes. Further, in the anodic scan, an oxidative wave near 1.2 V was attributed to the formation of a surface oxynitride and/or oxide phase. The reductive shoulders visible at approximately 0 V and -0.25 V are considered to correspond to the reduction of this oxynitride and/or oxide layer. The signal at 1.1 V in the cathodic sweep may be related to the reduction of 0₂ which evolved in the preceding anodic scan. Electrodes coated with CNTs show a strong capacitive response as evidenced by the large area under the CV curve and the lack of large faradaic features. Besides a well-defined reductive peak at -0.25 V and a weaker shoulder at -0.6 V, only ill-defined broad oxidation bands can be discerned. The overall redox activity of the CNT electrodes including the reductive peak at -0.25 V, bears a strong resemblance to the voltammetric response of the bare TiN electrodes. This may indicate that the CNTs do not form a fully closed layer above the TiN substrate and that the electrolyte is able to enter the CNT matrix and spread over the TiN. This result is of great importance since it implies a good wettability of the CNT matrix. On the other hand, similar voltammetric features have also been reported for singlewalled CNT sheets and were attributed to oxygen-containing functional groups present in the CNT sidewalls. Thus, it remains unclear whether the observed redox activity is solely due to the TiN underlayer or is also caused by the CNTs. Only a small fraction of the overall charge available with CNTs is faradaic in nature, while most of the charge arises from the charging and discharging of the electrochemical double-layer. Given this rather inert nature of the CNTs, fewer chemical modifications and hence degradation of the electrodes are expected during prolonged stimulation.

Visual comparison of the CV curves reveals that the CSCc of the CNT electrodes is much larger than that of the Pt and TiN electrodes. Table 1 summarizes the CSCc values (mean ± standard deviation) for all investigated electrode sizes and materials obtained from ne different electrodes. Also provided are the CSCc values normalized by the respective value of the 25 μιτι-diameter electrode, CSC25c, and the voltage ranges of the water window. The CSCc obtained with the CNT electrodes is more than 2 orders of magnitude larger than that of the Pt and TiN electrodes of the same diameter. It also appears that more charge per unit area is available with decreasing electrode diameter as indicated by the CSC²^.

This behavior can be explained by the potentiodynamic current response of ultramicroelectrodes.

Here, the steady-state diffusion-limited current density, i|, scales inversely with the electrode radius:

. AnDC^k

l[ =

7 Γ,.

^' (i)

where n is the number of electrons, D is the diffusion coefficient, C* is the bulk analyte concentration, and re is the electrode radius. Hence, for electrodes with 5 and 10 um diameter, the current density is expected to increase by a factor of 5 and 2.5, respectively, compared to a 25 um-diameter electrode. This is in good agreement with the range of CSC^c values in Table 1 except for the 5 μιτι diameter CNT electrodes, which accommodate a larger charge density than expected. This is most likely caused by slight deviations of the actual geometric electrode area from the nominal one. The increase in CSC_C with decreasing electrode area is an interesting aspect for neural stimulation at subcellular scales. The water window for Pt was considerably smaller than for the other materials, which can be attributed to the electrocatalytic nature of Pt towards hydrogen and oxygen evolution. For efficient stimulation, it is preferred to use an electrode material with a wide water window and a high CSC_C. The CNTs provide the best combination of both properties.

Table 1: Values for the CSC_C, CSC^c, and water window of Pt, TiN and CNT electrodes.

Material 0 csc_c esc-¹⁵ n_e Water window

[μηι] [mC cm^~2] [a.u.] vs. Ag|AgCl [V]

25 1.6 ± 0.2 1.0 ± 0.2 6

Pt 10 3.5 ± 1.6 2,2 ± 1 ,3 6 -0.60 - 1 .00

5 13.5 ± 3.0 8.6 ± 2.9 8

25 3.0 ± 1.3 1.0 ± 0.7 8

TiN 10 5.4 ± 1.1 1.8 ± 1.4 5 -0.75 - 1.10

5 19.8 ± 1.7 6.7 ± 2.8 6

25 513.9 ± 61.6 1.0 ± 0.2 12

CNTs 10 2012.0 ± 208.2 3.9 ± 0.9 13 -0.75 - 1.10

5 6612.1 ± 394.5 12.9 ± 2.3 12 Electrochemical impedance spectra were recorded for all electrode sizes and materials at their respective equilibrium potentials. The EIS data was analyzed by fitting of physically meaningful equivalent circuit models and by parameter extraction. The linear frequency response of an electrode-electrolyte interface can be modeled in terms of an equivalent circuit where the individual circuit elements describe the various relaxation phenomena occurring at the interface. In a first approach, the interface behaves like a parallel C circuit, where the resistance represents the faradaic response of the system, termed the charge- transfer resistance, and the capacitance provides information on the interfacial charge distribution known as the double layer capacitance, C_d. The latter is largely responsible for the neural signal transduction across the electrode-tissue interface and should be designed to be as large as possible in order to detect the minute neuronal signals. In general, more complex circuit models should be employed to account for diffusion phenomena, chemical modifications at the electrode surface, and distributed relaxation phenomena arising with electrode materials that exhibit a certain porosity or surface roughness. In this last case, the capacitive nature of the interface is represented by a constant-phase element (CPE), which is characterized by a frequency-independent phase angle. Its impedance, ¾_¾ shows anempirically determined power law frequency dependence:

1

ZCPE— -. .— 7

A_CPE {iCO)^a (2)

where ω is the angular frequency, A_CPE is a measure of the magnitude of Z_CPE, and 0 <a< 1 is a measure for the systems deviation from an ideal capacitive response for which a = 1. In general, a is considered to depend on the roughness of an electrode material and decreases with increasing roughness.

FIG. 4 shows the Bode plots of all measured electrode sizes and materials averaged over n_e measurements from n_e different electrodes (see Table 2 for the values of n_e). The amplitude of recorded neural signals is adversely affected by a large electrode-electrolyte interface impedance. A qualitative indicator for the efficiency of a recording electrode is its impedance at 1 kHz, which is the characteristic frequency in the power spectrum of a neural action potential with a duration of 1 ms. As shown in FIG. 5a, for all electrode sizes, the CNT electrode impedance at 1 kHz is 3 orders of magnitude lower compared to the Pt and TiN electrodes. In contrast, impedances of previously reported CNT electrodes of subcellular dimensions were 10 times higher (60-100 kOhm) and were achieved at a higher growth temperature of 600 C. Hence, despite the low CNT growth temperature employed in the process according to embodiments of the present invention, electrodes with an improved performance were achieved. Reasons for this improvement are manyfold and can be related to the CNT density and structure, and the chemical properties of the CNTs, and the substrate, among others. Furthermore, both the impedance and CSQ of the CNT electrodes remained unchanged after long-term storage in a nitrogen dry box for more than 3 months post-processing.

For all Pt electrode sizes in FIG. 4, the phase angle, Θ, remains close to the value of an ideal capacitor for which Θ = -90 . In contrast, the TiN electrodes exhibit a clear minimum of | θ | between 10²-10³ Hz, which implies a slightly stronger faradaic contribution than with Pt. Electrodes coated with CNTs behave capacitively at low frequencies and resistively at higher frequencies similar to an electronic high-pass filter. Least-square fitting of the spectra reveals that this effect originates from various relaxation processes involving chemical modifications at the CNT surface and a low R_a combined with the occurrence of diffusion. The equivalent circuit is provided in FIG. 5b. The series resistance, R_s, accounts for the ohmic behavior of the solution and metal wires. For the TiN and CNT electrodes, an additional capacitance, Co is included in series with R_a, In the case of CNTs, it has been recognized that a small fraction of the interfacial capacitance is faradaic rather than electrostatic in nature and is termed the pseudocapacitance. It is considered to originate from functional groups (defects) present at the tip and sidewalls of the CNTs which are redox active. In general, pseudocapacitances can arise due to chemisorption or redox reactions occurring at electrode surfaces. The charge that is transferred in these electrode processes is some function of the electrode potential, and as such resembles the behavior of a capacitance. Importantly, this kind of capacitance is faradaic in nature, rather than being associated with a potential-dependent distribution of electrostatic charge as for the C_dL The pseudocapacitance is represented as a capacitance, C_¾ in series with R_a, In the case of TiN, C_x may be related to the presence of an insulating oxide and/or oxynitride layer on top of the electrodes introduced during the electrode opening by RIE. In particular for the CNT electrodes, an additional distributed element, the Warburg impedance, Z_w, was required to account for diffusion effects in the porous matrix. It is connected in series with R_a and C_x. In general, diffusion occurs as a consequence of charge-transfer reactions taking place at an electrode and depleting the adjacent electrolyte from reactants. It is therefore directly related to the magnitude of ¾ which is a measure of the kinetic feasibility of a faradaic reaction. If charge- transfer is hindered due to sluggish reaction kinetics and thus a large _d diffusion will not become a limiting factor. Its contribution to the observed impedance will hence be negligible. On the other hand, facile reaction kinetics, or equivalently a small _d is accompanied by a fast depletion of reactants at the interface. The measured reaction rate becomes diffusion-limited and will be registered as an additional impedance. The Z_w is described by the same equation as Ζ_σΕ (Eq. (2)) but with a constant power factor, a = 0.5. Parasitic capacitances due to the Parylene C insulation are described by the lumped capacitance, QTable 2 provides the fit parameters for all electrode sizes and materials. Table 2: Average fitting parameters (with standard deviations) for the Pt, TiN, and CNT electrodes of 25, 10, and 5 mm diameter.

Material 0 a Aw c_x c₍- n_e

[Mm] [kOlini] [kOhm] IpFJ [pF]

25 20.38 ±2.65 (6.23 ±0.31) E6 0.32 ±0.05 0.92 ±0.00 n/a n/a 56.46 ±4.33 7

Pt 10 80.81 ±17.97 ( 1.36 ± 1.09) E6 0.13 ±0.05 0.89 ±0.03 n/a n/a 32.21 ±2.82 6

5 (4.69 ±2.70) E2 (2.12 ±1.58) E5 0.47 ±0.16 0.89 ±0.04 (5.70 ±2.49) E-4 n/a 23. 3 ±1.43 7

25 {0.83 ±0.10) E3 (0.53 ±0.15) E4 0.86 ±0.18 0.84 ±0.02 n/a (2.49 ±0.74) E2 49.38±2.51 7

TiN 10 (2.56 ±1.51) E3 (1.78 ±0.82) E4 0.39 ±0.14 0.78 ±0.06 n/a 75.83 ±25.38 38.79 ±4.52 5

5 {7.02 ±0.97) E3 (2.31 ±0.79) E4 0.30 ±0.08 0.80 ±0.06 n/a 67.74±8.53 43.19 ±2.78 7

25 0.63 ±0.05 1.23 ±0.14 (1.39 ±0.32) E2 0.83 ±0.02 11.73 ±1.49 (1.92±0.35)E5 (6.72 ±0.53) E2 16

CNTs 10 1.15 ±0.33 1.47 ±0.13 (0.81 ±0.24) E2 0.85 ±0.03 11.43 ±2, 25 (1.52 ±0.20) E5 (4.24 ± 2.33) E2 15

5 2.37 ±0.33 1.51 ±0.21 (0,65 ±0.12) E2 0.86 ±0.02 10.58 ±2.47 (0.87 ±0.13) E5 (2.24 ±0.31)E2 14

The R_s shows a strong dependence on the electrode material and is up to 3 orders of magnitude higher for the Pt and TiN electrodes as compared to the CNT electrodes. The series resistance R_s of the CNT electrodes is largely determined by the low solution resistance (the metal resistance can be neglected), which is a direct consequence of the large distributed surface area of the CNTs, and can be described in terms of a porous electrode:

where p is the solution resistivity, I is the pore length, r_p is the pore radius, and n_p is the number of pores. For comparison, the solution resistance in the case of a recessed disc electrode with radius r_e and a recess height h is:

R - + P^

4^r*^> (4)

where h = 1 μιτι corresponds to the Parylene C thickness. The average diameter of a nanotube in this context is 34 nm. Assuming that the CNTs are arranged in a dense square lattice, the pores corresponding to the void space between the nanotubes can be approximated by cylinders with a radius r_p = 7 nm. Further, the total number of pores equals the number of nanotubes, i.e. n_p = 9.8 · 10^s for the 25 μιτι-diameter electrode, and the pore length corresponds to the height of the CNT film, I ~ 2 μιτι. Using these values and a PBS resistivity of p = 70 Ohm cm, one obtains for a 25 μιτι-diameter electrode _S;P =9.3 kOhm and R_¾e =15.4 kOhm. The latter is in good agreement with the value for a Pt electrode in Table 2. In contrast, the actual series resistance R_s for the CNT electrodes is more than an order of magnitude lower than the theoretical series resistance R_S;P. Considering that the CNT length and density are reliably controlled during the growth process, the most uncertain factor in Eq. (3) is the pore radius r_p. It can strongly be affected by the fabrication process and most likely covers a broad range of values. This also becomes clear from the SEM images in FIG. 2b and f. The actual pore radius r_p for a 25 μιτι-diameter electrode is therefore about 4 times larger than estimated. For the 10 and 5 μιτι-diameter CNT electrodes one obtains 7 and 10 times larger pore radius r_p values, respectively, than estimated. The bare TiN electrodes show an unexpected high series resistance Rs. This may be due to a higher intrinsic resistivity of the TiN layer possibly caused by oxidation during RIE.

The charge-transfer resistance R_a for the Pt and TiN electrodes is 4-6 orders of magnitude higher than for the CNT electrodes. Thus, charge-transfer occurs much easier at the CNTs than at the Pt or TiN electrodes, which also explains why diffusional involvement in the form of the Warburg impedance Z_w was mostly observed for the CNT electrodes. The extent to which this enhanced reactivity of the CNTs may or may not be beneficial for neural recording and stimulation remains to be determined.

The low impedance of the CNT electrodes is a direct consequence of their large interfacial capacitance. This is reflected by A_CPE values in Table 2 which are 2-3 orders of magnitude higher than for the Pt and TiN electrodes. The constant-phase element C_PE power factor, a, of the Pt electrodes is close to unity and thus resembles the behavior of an ideal capacitor. The TiN and CNT electrodes have a smaller power factor a, which indicates a higher roughness/porosity. The roughness of TiN is caused by its microcolumnar morphology.

The CNT pseudocapacitance, C_X, is of the same order of magnitude as AQ>E and therefore directly related to the large specific surface area of the CNTs. Although faradaic and not electrostatic in nature like the constant-phase element C_PE, a large pseudocapacitance C_X may similarly improve the quality of the signal transduction across the electrode-tissue interface and enhance charge injection during stimulation. On the other hand, the influence of the Warburg impedance, Z_W, on the performance of neural electrodes is not evident. Finally, the C, is one order of magnitude higher for the CNT electrodes. Here, additional parasitic coupling compared to the planar Pt and TiN electrodes may occur through the Parylene C encasement of the vertical electrodes (see FIG. lg and FIG. 2f ).

It is known that chemical modifications by acid treatment and/or exposure to oxygen plasma can profoundly change the properties of CNTs, including hydrophobicity and biocompatibility. Avoiding chemical changes of the CNTs during device fabrication was an important aspect of process reliability in this work. Such a restriction ensures that post-processing treatments can be tailored to specifically modify the CNTs. In order to investigate the impact of the different processing steps on the chemical properties of the CNTs, an FTI analysis and contact angle measurements have been performed. FIG. 6 shows the baseline-corrected FTIR spectra and contact angles determined from as- grown CNTs, BHF-treated CNTs, and BHF-treated Si0₂-coated CNTs similar to the fabrication sequence of the real devices. The exposure to BHF was always 1.5 min. All spectra contain a broad band with a maximum near 3258 cm-¹ originating from stretching modes related to various O-H containing surface groups. A weak band centered around 2700 cm-¹ may be caused by O-H and C-H stretching modes. The signals at 1813 cm-¹, 1720 cm-¹, and 1585 cm-¹ are assigned to C=0 stretching modes. In addition, for the BHF-treated sample, a strong peak appeared at 1447 cm-¹, which also describes C=0 stretching vibrations. Signals in the region 1400-1000 cm-¹ are assigned to O-H bending and/or C-0 stretching modes. A strong peak in this region at 1097 cm-¹ was observed for the BHF-treated sample. The band at approximately 850 cm-¹ is attributed to C-H bending modes Clearly, the BHF treatment introduces additional O-containing functional groups. In contrast, the spectrum of the BHF-treated Si02- coated CNTs does not reveal any significant changes to the as-grown CNTs, which indicates that the BHF exposure time is sufficient to remove the oxide and leave the CNTs unaltered. The results also indicate that functional groups are to some extent present in the as-grown CNTs. The contact angle is lowest for the BHF-treated Si0₂-coated CNTs. The BHF treatment of the as-grown CNTs only marginally improved their wettability, which is unexpected considering the largest I signal increase due to hydrophilic functional groups. Therefore, the improved wettability of the BHF-treated Si0₂-coated CNTs is rather physical than chemical in nature. It was indeed shown that a better hydrophilicity can be achieved with CNT films that have an open microtexture as opposed to closed unmodified CNT films. The appearance of microbundles on the BHF-treated Si0₂-coated CNTs may hence account for the reduced contact angle. On the other hand, no microbundles were observed after BHF treatment of the as-grown CNTs. This demonstrates that the presence of the Si0₂ is advantageous for mediating the BHF into the CNT matrix and eventually forming the microbundles which improve the wettability.

Cell culture experiments were performed to assess the biocompatibility of the low- temperature fabrication process and thus demonstrate that the CNTs were neither affected by toxic chemical modifications nor comprised any toxic substances. Primary rat hippocampal neurons have been used due to their high sensitivity towards toxic compounds. FIG. 7 shows SEM images of neurons grown on fabricated CNT-based substrates after 5 days in vitro. The neurons readily formed equally branched neural networks on top of the Parylene C and on the exposed CNT areas as shown in FIG. 7a and FIG. 7b. Visible neurite ruptures are an artifact caused by sample preparation. The neuronal bodies did not make conformal contact with the underlying CNT microbundles, but rather seemed suspended on the surface (FIG. 7b and FIG. 7c). This behavior is only of secondary importance for in vivo applications. Here, neurons are not expected to attach to the electrode due to the presence of a fibrous scar tissue layer between the implant and healthy tissue. The biocompatibility of implanted microelectrodes on the other hand is of utmost importance to prevent excessive inflammatory responses and thus ensure a long-term tissue integrity and device reliability. These initial experiments demonstrate that the proposed low-temperature fabrication process is biocompatible.

Conclusions

Low-temperature, e.g. below 600°C, (such as 425°C) PECVD was employed to grow vertically- aligned multiwalled CNTs on passive arrays comprising microelectrodes with cellular and subcellular dimensions of 25, 10, and 5 μιτι diameter. In view of the long-term objective to incorporate CNTs in implantable neural microsystems, the demand for high-density and high-resolution electrode arrays can only be met by using advanced Cu-CMOS technology. Therefore, all process temperatures employed in experiments relating to embodiments of the present invention were compatible with BEOL Cu-CMOS processing (<450°C). Using vertically-aligned CNTs for implantable devices has raised concerns related to their mechanical stability during the insertion procedure where lateral abrasive forces may damage the CNTs. In order to address these concerns, as an advantage of embodiments of the present invention, a novel approach to encase and hence confine the vertical CNT electrodes has been demonstrated. A coating stack comprising a dielectric, e.g. Si0₂, and a biocompatible material, e.g. Parylene C, was employed with lithographically defined electrode openings where the biocompatible material, e.g. Parylene C, partly overlapped the CNT film. The coating stack also provided an effective coverage of the potentially toxic catalytic film, e.g. Ni film, present on the wafer surface. As an additional advantage of embodiments of the present invention, this coating stack provides also a biocompatible packaging, therefore making it suitable to be implanted. The functionality and performance of the fabricated and packaged devices was evaluated by CV and EIS and compared against co-fabricated Pt and TiN microelectrodes. For all electrode diameters, the CNT electrode impedance at 1 kHz was reduced by 3 orders of magnitude compared to the Pt and TiN electrodes. In addition, an improvement by a factor of 10 could be achieved for the 5 μιτι-diameter electrodes compared to reported CNT electrodes of similar size which were realized at a higher CVD growth temperature than in the process according to embodiments of the present invention. This allows the conclusion that high-quality CNT electrode can be realized at low processing temperatures. Furthermore, the CNT electrodes outperformed the Pt and TiN electrodes in terms of the CSCc by 2 orders of magnitude. Considering the overall electrode performance, CNT electrodes even with sub-micron dimensions are a realistic scenario for future neural implants. Furthermore, FTI revealed that the different fabrication steps did not alter the chemical fingerprint of the CNTs. This will provide more flexibility with post-processing treatments for specific modifications of the CNTs. Finally, cell viability on fabricated CNT substrates was successfully demonstrated with primary rat hippocampal neurons indicating that CNTs were not afflicted by toxic compounds or modifications introduced during processing. The proposed fabrication process can therefore be employed to realize future biocompatible neural implants.

Materials and Methods Device Fabrication

For the experiments, as illustrated in FIG. 1, processing was carried out on 200 mm Si wafers 1 with 300 nm thermal oxide 2 on both the front and back sides. Metal interconnects 3, bond pads 9, and electrode areas were defined by lift-off of sputterdeposited Ti/Pt/Ti (20 nm/200 nm/20 nm) (Nimbus XP, Nexx Systems, USA) using LOR10A (Microchem, USA) as the under-layer and 1X845 (JSR-Micro, USA) as the photoresist. In a second lift-off step, TiN (100 nm; Nimbus XP, Nexx Systems, USA) was selectively patterned on top of the electrode areas followed by F sputtering of a thin, 2 nm thick (nominal) Ni layer (A610 sputter system, Alcatel, USA), thus forming a TiN/Ni stack 4. Multiwalled CNTs 5 were grown in a 200 mm microwave (2.45 GHz) PECVD chamber (TEL, Japan). The microwave plasma source was located remotely from the wafer surface to avoid excessive ion-bombardment favoring CNT formation over fiber-like structures. The CNT growth temperature was fixed at 425°C to be compatible with Cu- BEOL CMOS processing. Furthermore, at this temperature, Ni is an effective catalyst for CNT growth because it is reduced to its catalytically active metallic state. In a typical CNT growth experiment, this Ni film was transformed into active metal nanoparticles in a NH₃ plasma for 5 min. The CNTs 5 were then grown in a gas atmosphere of C2H4/H2 at 3 Torr for 30 min resulting in a CNT density of 2 · 10¹¹ cm^"2, a height of 2 μιτι, and an average nanotube diameter of 34 nm. A 300 nm thick Si0₂ layer 6 deposited by CVD at 150°C (Plasmalab 80, Oxford Instruments, UK) served as the protection of the CNTs 5. As an effective and biocompatible insulation material, 1 μιτι of Parylene C 7 (PDS 2010 Labcoater 2, SCS, USA) has been evaporated. The bond pads 9 were opened by RIE (SPTS, UK) of the Parylene C 7 (130 s; 30 seem SF₆, 200 seem 0₂, 100 torr, 250 W) and Si0₂ 6 (90 s; 10 seem 0₂, 100 seem SF₆, 100 torr, 200 W) using a resist etch mask (1X845, JSR-Micro, USA). In a second RIE step, the Parylene C 7 on top of the contact areas was etched, thus providing openings 8. An etch time of 130 s was sufficient to remove the Parylene C and leave the Si0₂ coating intact. After wafer dicing, the singulated chips were wire-bonded onto custom PCBs and sealed with epoxy (353ND-T, Epotek, USA). Prior to electrochemical experimentation, the Si0₂ layer on top of the CNT electrodes was removed by dipping the packaged chips in BHF for 1.5 min. Inspection was performed by SEM (SU8000, Hitachi, Japan) and combined FIB/SEM analysis (Nova 600 Dual-Beam, FEI, USA).

Electrochemical Characterization

Electrochemical measurements were performed in PBS (0.150 M NaCI, 0.016 M Na₂HP0₄, 0.004 M KH₂P0₄, pH 7.4). All chemicals were analytical grade and used as delivered (Sigma-Aldrich, USA). Experiments were conducted in a glass beaker using a three-electrode configuration placed inside a Faraday cage. A commercial double-junction Ag|AgCI (3 M KCI) reference electrode (Radiometer Analytical, France) was used together with a large-area Pt counter electrode. Measurements were made using an Autolab PSTAT302N potentiostat with integrated frequency response analyzer controlled by the NOVA software (Ecochemie, Netherlands). An additional low-current amplifier module (ECD, Ecochemie, Netherlands) was used for low-current CV measurements with the 5 μιτι-diameter electrodes. OriginPro 8.5 (Originlab, USA) and ZView (Scribner, USA) have been used for the data analysis and the nonlinear least-square fitting of the EIS spectra, respectively. The experimental sequence for all electrodes consisted of an initial electrochemical cleaning step where the electrode potential was cycled at 2 V/s between the respective limits of gas evolution until a stable and reproducible response was observed, followed by 10 slow-sweep CV cycles at 0.1 V/s and the EIS with a 10 mV (rms) AC signal applied between 1 -10^s Hz. The CSCc was obtained from the 10th slow-sweep CV cycle.

Analysis by FTIR

FTIR analysis has been performed on as-grown CNTs synthesized as described above on wafers uniformly coated with TiN/Ni (100 nm/2 nm), on CNTs immersed in BHF for 1.5 min, and on CNTs coated with CVD Si0₂ (300 nm) and immersed in BHF for 1.5 min. All spectra were collected in transmittance (%) mode (IFS 66 v/S, Bruker Optics, Germany) over the wavenumber range of 500 - 4000cm^"1. Recorded spectra were corrected for baseline and analyzed using OriginPro 8.5 (Originlab, USA).

Cell Culture and SEM Analysis

For biocompatibility testing, chips have been prepared according to the process sequence displayed in FIG. 1; however, the two lift-off steps ((a) and (b)) have been omitted and an unpatterned metal stack of Ti/Pt/Ti/TiN (20 nm/200 nm/20 nm/100 nm) has been used instead. The CNTs were grown according to the procedure described above. Cell viability was tested with primary hippocampal neurons isolated from embryonic rats. Unpackaged chips were sterilized in 70% ethanol and incubated with poly L-lysine prior to cell plating. After 5 days in vitro, samples were removed from the incubator, fixated with formaldehyde, and treated with a 2% Os0₄ solution (Sigma-Aldrich, USA) for increased contrast during SEM analysis (SU8000, Hitachi, Japan). Finally, the samples were rinsed with PBS, dehydrated in solutions of increasing ethanol concentration, and subjected to critical point drying (Tousimis, USA).

Claims

Claims

1 - An electrode for stimulation and recording of tissue, the electrode comprising:

a substrate (1);

an insulating layer (2) atop the substrate (1);

an electrode layer (3) atop the insulating layer (2), the electrode layer (3) forming one or more electrodes;

a first layer (4) located atop the electrode layer (3);

a plurality of carbon nanotubes (5) atop the first layer (4);

a second layer (6) atop the insulating layer (2) and atop the plurality of carbon nanotubes (5); and an opening (8) in the second layer for providing access to the plurality of carbon nanotubes (5).

2 - An electrode according to claim 1, furthermore comprising a third layer (7) atop the second layer (6), and an opening in the third layer for providing an opening at the location of the plurality of carbon nanotubes (5).

3. - An electrode according to claim 2, wherein the third layer (7) is a biocompatible layer.

4. - An electrode according to claim 3, wherein the third layer (7) comprises Parylene C or SiN.

5. - An electrode according to any of the previous claims, wherein the insulating layer (2) is an oxide layer.

6. - An electrode according to any of the previous claims, wherein the second layer (6) is an oxide layer.

7. - An electrode according to claim 6, wherein the second layer (6) comprises Si0₂.

8. - An implantable device comprising

- a shaft (10);

- a tip (13);

at least one carbon nano tube cell stimulation and recording electrode (12) according to any of the previous claims located on the shaft (10); and

at least one bondpad (14) connected to the at least one cell stimulation electrode via at least one electrical interconnection (11).

9. - An implantable device according to claim 8, furthermore comprising additional circuitry (15) for driving the bondpads (14).

10. - An implantable device according to any of claims 8 or 9, wherein the implantable device is a neural probe for electrical stimulation and recording of brain tissue.

11 - A method for manufacturing an electrode for stimulation and recording of tissue, the method comprising

providing a substrate (1);

providing an insulating layer (2) atop the substrate (1); providing an electrode layer (3) atop the insulating layer (2), the electrode layer (3) forming one or more electrodes;

providing a first layer (4) located atop the electrode layer (3);

growing a plurality of carbon nanotubes (5) atop the first layer (4);

- providing a second layer (6) atop the insulating layer (2) and atop the plurality of carbon nanotubes (5); and

providing an opening (8) in the second layer for providing access to the plurality of carbon nanotubes (5).

12. - A method according to claim 11, furthermore comprising providing a third layer (7) atop the second layer (6), and providing an opening in the third layer layer at the location of the plurality of carbon nanotubes (5).

13. - A method according to any of claims 11 or 12, wherein providing an insulating layer comprises providing an oxide layer.

14. - A method according to any of claims 11 to 13, wherein providing the second layer comprises providing an oxide layer.

15. - A method according to any of claims 11 to 14, wherein providing the third layer comprises providing a biocompatible layer.