Micromachined electrode array

[0001] This is a continuation-in-part of application Ser. No. 10/426,152 filed Apr. 28, 2003, now U.S. Pat. No. 7,127,301.

[0002] This invention was made with Government support under Contract No. DE-AC04-94AL85000 awarded by the U.S. Department of Energy. The Government has certain rights in the invention.

[0003] The present invention relates in general to electrode arrays, and in particular to a micromachined electrode array that comprises a plurality of electrodes that can be addressed independently or in sets. The micromachined electrode array has applications for use in prosthetic devices including an implantable retinal prosthesis.

[0004] Efforts are currently underway in a number of groups worldwide to develop an implantable retinal prosthesis to restore at least partial sight to persons suffering from certain forms of blindness due to end-stage photoreceptor degenerative diseases such as age-related macular degeneration and hereditary retinitis pigmentosa. Such retinal prostheses, which are currently being developed, are based on the electronic transfer of visual information from a camera located in front of the eye to receiver circuitry which will be implanted within the eye to feed the visual information in pixelized form an electrode array that will be placed either in front of the retina (epiretinal) or beneath the retinal tissue (sub-retinal). Electrical currents from the electrode array will then provide artificial stimulation of neural tissue (e.g. ganglion cells) in the retina to generate the visual perception of dots of light corresponding to the pixelized visual image, with the goal of being able to restore a measure of sight to a blind person. Further details of implantable retinal prostheses to which the electrode array of the present invention can be applied can be found, for example, in U.S. Pat. Nos. 5,476,494; 5,836,996; 5,935,155; and 6,393,327 which are all incorporated herein by reference; and in an article by John Wyatt et al. entitled “Occular Implants for the Blind” published in IEEE Spectrum, pp. 47-53, May 1996; in another article by Mark S. Humayun et al. entitled “Pattern Electrical Stimulation of the Human Retina” published in Vision Research, vol. 39, pp. 2569-2576, 1999; and in yet another article by M. Schwarz et al. entitled “Single Chip CMOS Imagers and Flexible Microelectronic Stimulators for a Retinal Implant System” published in Sensors and Actuators, vol. 83, pp. 40-46, 2000.

[0005] Although progress has been made in the development of retinal prostheses, there still remains a need for an improved electrode array which can provide from hundreds to tens of thousands or more individual electrodes, each of which can conform to the curvature of the retina while providing a gentle, uniform contact pressure on the retina to prevent damage to the underlying neural cells.

[0006] The present invention represents an advance in the art by providing a micromachined electrode array which comprises, in certain embodiments, a plurality of spaced-apart electrodes, each of which is flexibly attached to a supporting substrate by a plurality of springs to allow independent movement of the individual electrodes, and to allow a spring constant of each spring to be tailored during design so that each electrode will provide substantially the same low contact force when urged into contact with the curved surface of the retina.

[0007] In other embodiments, the micromachined electrode array of the present invention can be formed with a plurality of spaced-apart electrodes which are attached to a substrate in a fixed position, and which can be independently electrically contacted, or contacted in sets.

[0008] The micromachined electrode array of the present invention can also be adapted for neural stimulation or sensing applications for many different types of neural tissue including neural tissue associated with visual, auditory and sensory systems, and for neural tissue associated with the control of muscles (e.g. for bladder function or the activation of paretic limbs).

[0009] In certain embodiments of the present invention, the electrode array can be used to sense the contact force of one or more electrodes in contact with a neural surface to ensure that the contact force does not exceed a predetermined limit, or to ensure that the contact force provided by the electrodes is substantially the same.

[0010] These and other advantages of the present invention will become evident to those skilled in the art.

[0011] The present invention relates to a micromachined electrode array which in certain embodiments comprises a silicon-on-insulator substrate that further comprises a monocrystalline silicon body, and a monocrystalline silicon layer supported above the monocrystalline silicon body and separated therefrom by an intervening electrically-insulating layer of silicon dioxide. This electrode array also comprises a plurality of electrodes spaced apart from each other and formed, at least in part, from the silicon-on-insulator substrate, with the plurality of electrodes protruding outward substantially normal to a major surface of the silicon-on-insulator substrate, and with each electrode being electrically isolated from adjacent electrodes while being mechanically and electrically connected to the silicon-on-insulator substrate. Each electrode also includes an electrically-conducting via extending through the electrically-insulating layer of silicon dioxide to connect a monocrystalline silicon layer portion of that electrode to a monocrystalline silicon body portion thereof.

[0012] Some embodiments of the micromachined electrode array can further comprise a plurality of springs to mechanically and electrically connect each electrode to the silicon-on-insulator substrate. In these embodiments, each spring can be formed, at least in part, from the monocrystalline silicon layer. Alternately, each spring can comprise polycrystalline silicon, a metal or a combination thereof. Each spring can also be a folded spring.

[0013] The electrically-conducting via, which extends through the electrically-insulating layer of silicon dioxide to connect the monocrystalline silicon layer portion of each electrode to the monocrystalline silicon body portion thereof, can comprise a material such as polycrystalline silicon, a metal (e.g. tungsten) or a combination thereof.

[0014] Each electrode in the micromachined electrode array can be optionally overcoated with a metal or an electrically-conducting metal oxide. A majority of each electrode can also be optionally overcoated with an electrically-insulating biocompatible material (e.g. parylene, silicon nitride, or silicon dioxide).

[0015] The micromachined electrode array can also comprise a flexible frame to hold the silicon-on-insulator substrate. The flexible frame can be used to position the electrodes in contact with a surface (e.g. an epiretinal surface), and can be used to attach one or more electrode arrays to that surface. One or more of the electrodes can optionally include means for sensing a contact force when the micromachined electrode array is urged into contact with the surface.

[0016] Addressing circuitry can be provided on the silicon-on-insulator substrate for electrically addressing each electrode. Additionally, a plurality of transistors can be formed on the silicon-on-insulator substrate (e.g. for addressing the electrodes individually or in sets, or for amplifying a signal sensed by one or more of the electrodes).

[0017] The present invention further relates to an electrode array for neural stimulation which comprises a silicon-on-insulator substrate; and a plurality of electrodes arranged in an array and protruding outward substantially normal to a major surface of the silicon-on-insulator substrate and connected thereto. The electrodes are formed, at least in part, from the silicon-on-insulator substrate, with each electrode comprising a monocrystalline silicon layer portion, a monocrystalline silicon body portion, and an intervening electrically-insulating layer of silicon dioxide. Each electrode further comprises an electrically-conducting via formed through the intervening electrically-insulating layer of silicon dioxide to electrically connect the monocrystalline silicon layer portion to the monocrystalline silicon body portion.

[0018] A plurality of springs (e.g. folded springs) can be optionally used to connect each electrode to the silicon-on-insulator substrate. When springs are used, each spring can comprise a material such as monocrystalline silicon, or alternately polycrystalline silicon, a metal or a combination thereof.

[0019] Each electrode can be overcoated with a metal or an electrically-conducting metal oxide. If needed, a majority of each electrode can also be overcoated with an electrically-insulating biocompatible material.

[0020] Addressing circuitry can be provided on the silicon-on-insulator substrate for electrically addressing each electrode. A plurality of transistors can also formed on the silicon-on-insulator substrate.

[0021] A flexible frame can be provided to hold the silicon-on-insulator substrate in contact with a neural surface. Means for sensing a contact force for one or more of the electrodes when the electrode array is urged into contact with the neural surface can also be provided.

[0022] The present invention also relates to a micromachined electrode array which comprises a monocrystalline silicon body; a plurality of electrodes formed, at least in part, from the monocrystalline silicon body, with each electrode extending outward substantially perpendicular to a plane of the monocrystalline silicon body; an electrically-insulating layer disposed over the monocrystalline silicon body to attach each electrode thereto; and addressing circuitry located above the electrically-insulating layer and connected to the plurality of electrodes through an electrically-conducting via formed through the electrically-insulating layer at the location of each electrode. A flexible frame can be provided to hold the substrate of one or more electrode arrays to permit contact of the electrodes with a surface such as an epiretinal surface.

[0023] The micromachined electrode array can further comprise a monocrystalline silicon layer disposed above the electrically-insulating layer, with a portion of the monocrystalline silicon layer being electrically connected to each electrode through the electrically-conducting via. A plurality of transistors can also formed in the monocrystalline silicon layer and operatively connected to the plurality of electrodes through the electrically-conducting via connected to each electrode. These transistors can be used to electrically address each electrode, or to amplify a signal sensed by each electrode, or both. In other embodiments of the present invention, a plurality of transistors can be formed in the monocrystalline silicon body and operatively connected to the plurality of electrodes through the electrically-conducting via connected to each electrode. These transistors can function as described above to electrically address each electrode, or to amplify a signal sensed by each electrode, or both.

[0024] The electrically-conducting via can comprise a material selected from the group consisting of polycrystalline silicon, metals, and combinations thereof. The electrically-insulating layer can comprise silicon dioxide or silicon nitride. Each electrode can be optionally overcoated with a metal or an electrically-conducting metal oxide. A majority of each electrode can also be overcoated with an electrically-insulating biocompatible material.

[0025] Additional advantages and novel features of the invention will become apparent to those skilled in the art upon examination of the following detailed description thereof when considered in conjunction with the accompanying drawings. The advantages of the invention can be realized and attained by means of the instrumentalities and combinations particularly pointed out in the appended claims.

[0049] Referring to FIG. 1, there is shown a schematic plan view of an example of the electrode array 10 of the present invention. Although FIG. 1 illustrates a limited number of electrodes 12 in the array 10, those skilled in the art will understand that the electrode array 10 of the present invention can include up to hundreds, thousands or even millions of individual electrodes 12 depending upon the size and spacing of the individual electrodes 12 and the particular application for the device 10.

[0050] In FIG. 1, the electrode array 10 comprises a plurality of the electrodes 12 spaced apart from each other by a distance of, for example, 100-250 μm, with each electrode 12 being flexibly attached to a supporting substrate 14 by a plurality of springs 16. The springs 16 allow each electrode 12 to move independently in a direction which is substantially normal to a surface of the substrate 14 (i.e. perpendicular to the plane of FIG. 1). This is illustrated in FIGS. 2A and 2B which show schematic cross-section views along the section line 1-1 in FIG. 1.

[0051] The substrate 14 can comprise a semiconductor such as silicon which can be doped for electrical conductivity (e.g. n-type doped to about 10¹⁹ cm⁻³ or more to provide a resistivity of, for example, 0.001-0.004 Ω-cm). The substrate 14 can initially be, for example, about 500 μm thick prior to fabrication of the electrodes 12 and can be thinned down to a final thickness of, for example, 100-300 μm, with the exact final thickness of the substrate 14 depending upon how the electrodes 12 are formed.

[0052] FIG. 2A shows the electrode array 10 in an as-fabricated position with each electrode 12 comprising a bulk-micromachined electrode seat 18 formed from the substrate 14 and an elongate electrode tip 20 which is attached to the electrode seat 18. The electrode tips 20 can be either pressed into the electrode seats 18 or attached with an electrically-conductive adhesive 22 (e.g. an electrically-conductive epoxy). In other embodiments of the present invention as will be described in detail hereinafter, the electrode tips 20 can be formed integrally with the electrode seats 18 by electroplating, or by etching the electrode seats 18 and the tips 20 out of the semiconductor substrate 14, or a combination thereof.

[0053] In FIG. 2A, each electrode 12 is flexibly attached to the surrounding substrate 14 by a plurality of springs 16 which can be formed from polysilicon (i.e. polycrystalline silicon), or a metal, or a combination of both. In other embodiments of the present invention, when the substrate 14 comprises a silicon-on-insulator substrate, the springs 16 can be formed from monocrystalline silicon, with or without an overlayer of a metal. The electrodes 12 can be electrically insulated from the substrate 14 by an electrically-insulating layer 24 which generally comprises silicon nitride, or alternately silicon dioxide when a silicon-on-insulator substrate 14 is used. When the springs 16 comprise polysilicon or polycrystalline silicon, the springs 16 can be doped for electrical conductivity (e.g. n-type doped with phosphorous up to about 10¹⁹ cm⁻³ or more). The springs 16 can also be used to electrically connect each electrode 12 to addressing circuitry 26 (i.e. wiring) located on the substrate 14.

[0054] The addressing circuitry 26 can include a plurality of switching transistors 28 formed in the silicon substrate 14 or in a monocrystalline silicon layer provided on a silicon-on-insulator substrate 14 to allow row and column addressing of the individual electrodes 12 in the array 10. A plurality of bond pads (32) can be provided around the periphery of the substrate 14 to allow the addressing circuitry 26 to be connected to other electronic circuitry (e.g. for providing addressing information to the switching transistors 28, for providing electrical signals to the electrodes 12 for neural stimulation, for sensing electrical signals from neural tissue using the electrodes 12, etc.). Alternately, the electronic circuitry can be located on a separate substrate which is attached to the electrode array 10 (see FIG. 7) with the electrical connections to the addressing circuitry 26 being made through a plurality of electrical interconnections between the substrates.

[0055] Each electrode 12 can be flexibly attached to the substrate 14 by a plurality of springs 16 to allow an extended range of displacement of the electrode 12 of up to about 500 μm depending upon a particular application for the electrode array 10. This extended range of movement is made possible by the use of springs 16 which have an overall length that can be in the range of 0.2-1 millimeter and lateral dimensions in the range of 1-5 μm. To save space, the springs 16 can be curved around the electrodes 12 (see FIGS. 10C and 10D) or else folded. The term “folded” as used herein refers to a spring 16 that has a shape which folds back on itself one or more times (e.g. a serpentine shape as shown in FIGS. 1 and 10B). The term “folded” as used herein also refers to a spring 16 that includes one or more 90° bends (see FIG. 10A).

[0056] In FIG. 2B, operation of the electrode array 10 is schematically illustrated for contacting a curved surface 100. Here, the springs 16 allow the electrodes 12 to move independently and thereby conform to the surface 100 as the electrode array 10 is urged into contact therewith. An advantage of this example of the electrode array 10 of the present invention is that each electrode 12 is moveable independent of the other electrodes 12 so that the electrode array 10 can be used to make contact with surfaces of arbitrary shape. In particular, the electrode array 10 can be used to contact a neural surface such as the retina inside an eye without a need to bend the substrate 14 or to shape the electrodes 12 to fit a curved surface presented by the inside of the retina (although shaping of the electrodes 12 can be performed for certain embodiments of the electrode array 10 as will be described hereinafter).

[0057] Another advantage of the electrode array 10 of the present invention is that a spring constant, k, for each spring 16 in the electrode array 10 can be tailored to provide a substantially equal contact force, F, when the electrode array 10 is urged into contact with a surface of a known shape (e.g. a curved surface such as an epiretinal surface). This can be done, for example, by designing a photomask pattern that will be used to form the springs 16 by surface micromachining to provide a predetermined width or length for each spring 16 connected to a particular electrode 12 so that the spring constant, k, for each spring 16 will be inversely related to an expected displacement, x, for that electrode 12. Then the force, F=−nkx, provided by each electrode 12 due to a stretching of the number, n, of the springs 16 attached to that electrode 12 will be substantially the same. The electrodes 12 can each provide a contact force that is on the order of a few microNewtons up to tens of microNewtons depending upon a particular application of the device 10.

[0058] Control over the contact force of each electrode 12 can be important when the electrode array 10 is to be used as part of a retinal prosthesis since nerve cells in the retina can be damaged by excessive pressure. The retina's sensitivity to pressure can be inferred from the onset of glaucoma in human eyes which can occur for chronic excess fluid pressures of about 10 mm of mercury. Thus, when used as part of a retinal implant for the artificial stimulation of neural cells (e.g. ganglion cells), the electrode array 10 should preferably provide a controllable and well-defined contact pressure that is substantially the same for all electrodes 12 when the electrode array 10 is urged into contact with the curved epiretinal surface. This can be done, for example, by decreasing the spring constant, k, with distance radially outward from the center of the electrode array 10, with the spring constant, k, for each spring 16 attached to an electrode 12 located at a given radius being inversely proportional to the expected displacement, x, of that electrode 12.

[0059] Yet another advantage of the electrode array 10 of the present invention is that the contact force for one or more of the electrodes 12 can be sensed while the electrode array is in contact with a surface of arbitrary shape. This can be done by utilizing the piezoresistivity of the polysilicon or monocrystalline silicon used to form the springs 16 which provides a change in resistance for each spring 16 as a function of a deflection of that spring 16. FIGS. 3A and 3B illustrate this method for providing an in situ or in vivo sensing of the contact force of one or more electrodes 12 of the array 10.

[0060] In FIG. 3A, a pair of the springs 16′ are electrically insulated from the electrode 12 by an electrically-insulating layer 24 disposed between the springs 16′ and the electrode 12; and these springs 16′ are used to form variable resistors R₁ and R₂, with the resistance of each resistor R₁ and R₂ varying in proportion to the displacement of the electrode 12 and springs 16′ according to the piezoelectric effect within one or more layers of polysilicon or monocrystalline silicon used to form the springs 16′, and an initial resistance of springs' without any displacement thereof. One or more of the remaining springs 16 in FIG. 3 can be electrically connected to the electrode 12 through openings formed in the electrically-insulating layer 24 thereby forming a current path between the electrode 12 and the addressing circuitry 26 (see FIG. 1).

[0061] In FIG. 3A, a pair of fixed resistors 30 denoted as R₃ and R₄ can be formed over the electrically-insulating layer 24 on the substrate 14 and connected in series with the variable resistors R₁ and R₂. The resistors R₁ through R₄ can be further connected to bond pads 32 as shown in FIG. 3A or alternately to electrical wiring formed on the substrate 14, with the resistors R₁ through R₄ preferably being arranged to form a Half-Wheatstone-Bridge circuit as shown in FIG. 3B. The Half-Wheatstone-Bridge circuit, when connected to a voltage source, V_S, provides an output voltage, V_OUT, that is responsive to changes in the displacement of the electrode 12 and springs 16′ thereby allowing the displacement and contact force to be determined from the output voltage, V_OUT. The resistors R₁ through R₄ can all have substantially the same initial resistance when the springs 16′ are at rest in an undeflected position. A calibration of the Half-Wheatstone-Bridge circuit can be performed, for example, by measuring the output voltage, V_OUT, with the springs 16′ in the undeflected position and then providing a known displacement or contact force on the electrode 12 and measuring a change in the output voltage, V_OUT, from that in the undeflected position.

[0062] The ability provided by the electrode array 10 in certain embodiments of the present invention to sense the contact force for one or more electrodes 12 is particularly useful for implanted neural prosthesis since this can provide a way of monitoring the pressure exerted by the individual electrodes 12 on neural tissue which has not heretofore been possible. Such monitoring can be performed, for example, during implantation of a retinal prosthesis to allow forces exerted on the retina to be sensed during implantation and to be monitored periodically or continuously throughout the use of the retinal prosthesis.

[0063] FIGS. 4A-4R show schematic cross-section views along the section line 1-1 in FIG. 1 to illustrate a process for fabricating a first embodiment of the electrode array 10 of the present invention using surface and bulk micromachining as known to the art. Surface and bulk micromachining processes are based on conventional IC processing steps, including repeated steps for material deposition, photolithography, masking, etching, mask stripping, and cleaning. Only the essential steps for fabricating the electrode array 10 of the present invention will be described in detail herein.

[0064] In FIG. 4A, a heavily-doped (≧10¹⁹ cm⁻³) n-type silicon substrate 14 is initially prepared for fabrication of the electrode array 10 by blanketing the substrate 14 with a layer of a thermal oxide (not shown) which can be 630 nanometers thick and formed by a conventional wet oxidation process at an elevated temperature (e.g 1050° C. for about 1.5 hours). A low-stress electrically-insulating layer 24 comprising silicon nitride (e.g. 800 nanometers thick) can then be blanket deposited over the substrate 14 using low-pressure chemical vapor deposition (LPCVD) at about 850° C. The thermal oxide layer and the silicon nitride layer 24 provide electrical isolation from the substrate 14 for certain elements of the electrode array 10 (e.g. the springs 16, the addressing circuitry 26 and the fixed resistors 30). A plurality of openings 34 can be formed through the thermal oxide layer and the silicon nitride layer 24 in preparation for electrically connecting the springs 16 to the electrode seats 18.

[0065] In FIG. 4B, a first polysilicon layer 36 (denoted Poly-0) can be blanket deposited over the substrate 14 by LPCVD at a temperature of about 580° C. and with a layer thickness of, for example, 300 nanometers. The Poly-0 layer 36 can be used to form part of the addressing circuitry 26 together with a subsequently-deposited second polysilicon layer 40 (denoted Poly-1) described hereinafter with reference to FIGS. 4E and 4F. Additionally, the Poly-0 layer 36 is used to connect the springs 16 to the electrode seats 18 which will be formed from the substrate 14 in later process steps. Phosphorous doping can be used to make the Poly-0 and Poly-1 layers 36 and 40 electrically conductive.

[0066] In FIG. 4C, the Poly-0 layer 36 is patterned to form a first layer of the addressing circuitry 26 (see FIG. 1) on the electrically-insulating layer 24, and to leave portions of the Poly-0 layer 36 on the substrate 14 at the locations of the openings 34 through the electrically-insulating layer 24, thereby forming an electrically-conducting via through the electrically-insulating layer 24. The term “patterning” as used herein refers to a sequence of well-known semiconductor integrated circuit processing steps including applying a photoresist to the substrate 14, prebaking the photoresist, aligning the substrate 14 with a photomask, exposing the photoresist through the photomask, developing the photoresist, baking the photoresist, etching away the surfaces not protected by the photoresist, and stripping the protected areas of the photoresist so that further processing can take place. The term “patterning” can further include the formation of a hard mask (e.g. comprising about 500 nanometers of a silicate glass deposited from the decomposition of tetraethylortho silicate, also termed TEOS, by low-pressure chemical vapor deposition at about 750° C. and densified by a high temperature processing) overlying a polysilicon, metal or sacrificial material layer in preparation for defining features into the layer by etching.

[0067] In FIG. 4D, a first layer of a sacrificial material 38 can be blanket deposited over the substrate 14 by LPCVD and patterned to form openings 34 at the locations of each end of the springs 16. The sacrificial material 38, which can be about 2 μm thick, can comprise silicon dioxide (SiO₂) or a silicate glass (e.g. TEOS which is deposited from the decomposition of tetraethylortho silicate by low-pressure chemical vapor deposition at about 750° C. followed by densification at a higher temperature).

[0068] In FIG. 4E, the second polysilicon layer 40 (Poly-1) is blanket deposited over the substrate 14 by LPCVD, filling in the openings 34 for use in building up the springs 16 and for providing a second layer of the addressing circuitry 26. In FIG. 4F, the Poly-1 layer is patterned using reactive ion etching to form the springs 16 and the second layer of the addressing circuitry 26. The Poly-1 layer 40 can be 1-2 μm thick. If needed, additional layers of polysilicon (not shown) can be deposited over the Poly-1 layer 40 and laminated thereto to build up the springs 16 to a greater thickness.

[0069] In other embodiments of the present invention, a metal (e.g. aluminum, tungsten or an alloy thereof) can be substituted for or deposited over the polysilicon in the layers 36 and 40 to form the addressing circuitry 26 and the springs 16.

[0070] In FIG. 4G, once the springs 16 have been formed, they are encapsulated in another layer of the sacrificial material 38. A thermal annealing step can then be provided to anneal out any residual stress in the polysilicon layers 24 and 40. This annealing step can be performed at an elevated temperature of about 1100° C. for several hours (e.g. 3 hours).

[0071] In FIG. 4H, a patterned etch mask 42 can be formed on a backside of the substrate 14 in preparation for etching through the substrate 14 to form the electrode seats 18. The etch mask 42 includes a plurality of shaped openings 34 at locations where the substrate 14 will be etched away. The etch mask 42 can comprise a hard etch mask as described previously and can be formed in part from the layers of silicon nitride, polysilicon and the sacrificial material 38 which are also generally deposited on the backside of the substrate 14 when LPCVD is used, although these layers have been omitted from FIGS. 4A-4G for clarity.

[0072] In FIG. 4I, the substrate 14 can be etched from the backside to remove portions of the substrate 14 that are exposed by the openings 34 through the etch mask 42 and thereby begin to form the electrode seats 18. The electrode seats 18 can have lateral dimensions of, for example, 100 μm. The etching process used to remove material from the substrate 14 can comprise a deep anisotropic plasma etching process which has been termed a Bosch etch process. The Bosch etch process is disclosed in U.S. Pat. No. 5,501,893 to Laermer, which is incorporated herein by reference. This Bosch etch process combines multiple anisotropic etching steps with steps for simultaneously depositing an isotropic polymer/inhibitor to minimize lateral etching thereby allowing openings to be etched through the substrate 14 to a depth of up to the thickness of the substrate 14 while retaining substantially uniform lateral dimensions (i.e. straight sidewalls) for the etched openings.

[0073] In FIG. 4J, the etch mask 42 can be re-patterned or replaced by another etch mask 44 which is used for a second Bosch etch step which can be terminated upon reaching the electrically-insulating layer 24, or which can be used to etch through both the substrate 14 and the electrically-insulating layer 24 as shown in FIG. 4K. The second Bosch etch step forms the electrode seats 18 and separates the electrode seats 18 from the remaining substrate 14. Additionally, the second Bosch etch step can be used to form a receptacle 46 in each electrode seat 18 wherein an electrode tip 20 can later be inserted and attached. Each receptacle 46 can be, for example, 10-100 μm in diameter depending upon the size of the electrode tips 20 to be used in the array 10. Although the receptacles 46 are shown in FIG. 4K as having straight sidewalls 46, the receptacles 46 can be optionally be formed with sloping sidewalls 46 (i.e. tapered inward with increasing depth of etching). Tapering of the receptacles 46 can be performed, for example, using a conventional reactive ion etching step.

[0074] The electrode tips 20 can, in turn, be sized to fit a particular type of neuron which is to be stimulated by or sensed with the electrode array 10. As an example, ganglion cells in the retina are typically about 10-20 μm in diameter, so that each electrode tip 20 can be about this size if single ganglion cells are to be stimulated with the electrode array 10 in a retinal prosthesis, and the diameter of the electrode tips 20 can be scaled upward from this size when multiple ganglion cells are to be stimulated by each electrode tip 20.

[0075] In FIG. 4K, after the second Bosch etch step is completed, the etch mask 44 can be removed. As shown in FIG. 4K, the remaining substrate 14, which is not covered by the etch mask 44, can be thinned (e.g. down to 100-300 μm thickness) by the second Bosch etch step. This allows the electrode seats 18 to protrude outward from the backside of the substrate 14. Although the electrically-insulating layer 24 is shown removed underneath the springs 16 in FIG. 4K after the second Bosch etch step, the electrically-insulating layer 24 can also be left in place after the second Bosch etch step is completed and removed in a later step (e.g. by etching from a topside of the substrate 14). Additionally, the electrically-insulating layer 24 can be optionally patterned in later steps to form a stop (not shown) underneath each spring 16 to limit a downward movement of the spring 16 (e.g. when the spring 16 is loaded by the added weight of the electrode tips 20).

[0076] In FIG. 4L, formation of the electrode tips 20 is commenced in another series of processing steps by providing a sacrificial substrate 48 whereon the electrode tips 20 can be formed by electroplating. A release layer 50 can be formed over a surface of the sacrificial substrate 48, with the release layer 50 being used to later separate the electrode tips 20 which will be formed by electroplating from the sacrificial substrate 48. The sacrificial substrate 48 can comprise, for example, silicon, glass, fused silica, alumina, sapphire or metal. The material used to form the sacrificial substrate 48 is generally not critical since it serves only as a support whereon the electrode tips 20 can be electroplated.

[0077] The release layer 50 can comprise, for example, copper or polymethyl methacrylate (PMMA). An electroplating seed layer (not shown) can be provided over the release layer for use in electroplating a metal such as platinum, iridium, titanium or gold or a combination thereof for forming the electrode tips 20. The seed layer can also be used to make an electrical connection for electroplating, especially when an electrically-insulating release layer such as PMMA is used. Those skilled in the art will understand that, while the above-cited metals are preferred for reasons of biocompatibility when the electrode array 10 is to be used for neural stimulation, other metals and electrically-conductive metal oxides can be used to form the electrode tips 20 for applications of the electrode array 10 other than neural sensing.

[0078] The process steps described herein for forming the electrode tips 20 in FIGS. 4L-4N are based on LIGA (an acronym based on the first letters for the German words for lithography and electroplating) which is well-known in the art. LIGA utilizes deep x-ray lithography (e.g. using synchrotron radiation) to form an electroplating mask 52 from a sheet resist material (e.g. PMMA) which can be patterned with micron-sized features and which can have a large aspect ratio as shown in FIG. 4M. The electroplating mask 52, which includes a plurality of shaped openings 54 at the locations wherein the electrode tips 20 are to be formed, can then be used to electroplate the electrode tips 20 on the release layer 50 as shown in FIG. 4N. The electrode tips 20 are formed as an array with a spacing corresponding to that of the receptacles 46 formed in the electrode seats 18 in FIG. 4K and with a length, for example, of up to 500 μm so that an overall length of the electrodes 12 is generally about one millimeter or less. The electrode tips 20 are generally sized to be slightly smaller in width than the receptacles 46 so that the electrode tips 20 can later be inserted therein all at once and attached to the electrode seats 18 by an adhesive or a friction fit.

[0079] In FIG. 4O, the electroplating mask 52 can be removed in part (e.g. by solvent dissolution or plasma etching) to expose the ends of the electrode tips 20 in preparation for inserting the electrode tips 20 into the electrode seats 18.

[0080] In FIG. 4P, an electrically-conductive adhesive 22 (e.g. an electrically-conductive epoxy) can be applied to the exposed ends of the electrode tips 20 immediately prior to mating the electrode tips 20 with the electrode seats 18 as shown in FIG. 4Q. In other embodiments of the present invention, the electrode tips 20 can be inserted into the electrode seats 18 and held in place by a friction fit. This can be done, for example, by heating the electrode seats 18 and cooling the electrode tips 20 and then mating the two together. Alternately, the electrode tips 20 can be urged into the electrode seats 18 with a backing (e.g. PMMA or a silicone polymer such as poly-dimethylsiloxane, also termed PDMS) being provided on a topside of the substrate 14 for protection and added stiffness. The electrode tips 20 can be rounded or pointed (e.g. by electropolishing or etching), if needed, to facilitate mating of the electrode tips 20 to the electrode seats 18.

[0081] In FIG. 4R, once the electrode tips 20 have been inserted into the electrode seats 18 and the adhesive 22 has cured, the electrode tips 20 can be separated from the sacrificial substrate 48 by selectively etching away the release layer 50 (e.g. using an acid for a copper release layer 50) or by dissolving the release layer 50 in a solvent such as acetone (e.g for a PMMA release layer 50). During this step, the electroplating seed layer can also be removed from the electrode tips 20.

[0082] If needed, an additional layer of metal or a metal oxide (e.g. iridium oxide) can be deposited over the ends of the electrode tips 20 for improved electrical conductivity or for compatibility when interfacing with a particular type of neural tissue or other surface with which the electrode array 10 is to be used. Such an additional layer of metal or metal oxide (not shown in FIG. 4R) can be deposited over the electrode tips 20 by a conventional deposition process (e.g. sputtering or evaporation), or by electroplating. An iridium oxide surface can be formed on electrode tips 20 comprising iridium by a thermal oxidation step whereby the iridium on exposed surfaces of the electrode tips 20 is converted to iridium oxide at an elevated temperature.

[0083] The layers of the sacrificial material 38 in FIG. 4R can then be removed with a selective etchant comprising hydrofluoric acid (HF) to release the individual electrodes 12 for movement and to complete the fabrication of the electrode array 10. This can be done by immersing the electrode array 10 in an HF solution for a time period ranging from several minutes up to several hours.

[0084] The formation of a plurality of switching transistors 28 on the substrate 14 for addressing the individual electrodes 12 in the array 10 can be performed during the fabrication of the other elements of the electrode array 10 described with reference to FIGS. 4A-4C. The transistors 28 can be formed in the substrate 14 by a series of standard integrated circuit (IC) processing steps including ion implantation or dopant diffusion to form drain and source regions for the transistors 28 which are preferably field-effect transistors (FETs). The Poly-0 layer 36 can be used to form a gate between the source and drain of each transistor, with the thermal oxide being used as a gate oxide to separate the gate from a channel region formed in the substrate 14 between the drain and source regions. The Poly-0 layer 36 can also be used for electrical interconnections to the gate and the drain and source regions. Electrical isolation of each transistor 28 from the substrate 14 can be provided by forming the drain, source and channel regions with a dopant type which is opposite the doping type used for the substrate 14 (i.e. by forming reverse-biased diode junctions between the drain, source and channel regions and the substrate 14).

[0085] Although the materials (e.g. silicon, silicon dioxide, silicon nitride, silicate glass, platinum, iridium, gold, titanium and/or iridium oxide) used to fabricate the electrode array 10 are generally biocompatible, in certain embodiments of the present invention used for neural stimulation (e.g. for use in a retinal prosthesis), a thin layer (e.g. up to about 0.1 μm) of an electrically-insulating biocompatible material such as parylene (i.e. a para-xylene polymer), silicon dioxide or silicon nitride can be conformally deposited over all exposed surfaces of the electrode array 10 and then selectively removed at the ends of electrode tips 20 (e.g. from the end and backwards for a distance of 1-10 μm). This conformal deposition can be performed by LPCVD in the case of silicon dioxide or silicon nitride. The selective removal of the electrically-insulating biocompatible material at the ends of the electrode tips 20 can be performed using laser ablation, solvent dissolution, or etching depending on the particular biocompatible material used.

[0086] Parylene, which is produced by the condensation and polymerization of a gaseous monomer, para-xylylene, can be conformally deposited at room temperature using a vapor deposition polymerization (VDP) process in commercial VDP apparatus. The VDP process is disclosed by Gorham in U.S. Pat. No. 3,342,754, which is incorporated herein by reference. In the VDP process, a parylene dimer (e.g. di-para-xylylene) is heated in the VDP apparatus to about 150° C. resulting in its conversion to a gaseous dimer. This causes the gas pressure in a vaporization zone wherein the parylene dimer is heated to rise, forcing the dimeric gas downstream into a pyrolysis zone where it is then heated further to about 650° C., splitting the dimer molecules into highly reactive monomer molecules (e.g. para-xylylene). The monomer molecules continue to respond to pressure, flowing into a room-temperature deposition chamber of the VDP apparatus where the monomer molecules disperse and grow as a clear linear-polymer film on all surfaces to which the monomer molecules are exposed. The thickness of the resultant parylene coating can be controlled by the volume of the parylene dimer that is vaporized and by the dwell time in the deposition chamber. Since the parylene deposition process is gaseous, the coating thickness is uniform and conformal, covering all exposed surfaces of the electrode array 10 and any other elements (e.g. electronic circuitry, wiring, etc.) attached thereto, without any associated cure stress.

[0087] In other embodiments of the present invention, the electrode tips 20 can be formed directly on the electrode seats 18 as described hereinafter with reference to FIGS. 5A-5C and FIGS. 6A-6F.

[0088] In FIGS. 5A-5C, the electrode tips 20 can be electroplated directly onto the electrode seats 18. This can be done, for example, by providing a thick patterned mask 56 over the backside of the substrate 14 as shown in FIG. 5A after the process step described with reference to FIG. 4K, with openings 58 defined photolithographically or by deep x-ray lithography at the locations of each electrode seat 18 wherein the electrode tips 20 are to be formed. The mask 56, which can have a thickness that is substantially equal to the length of the electrode tips 20 to be formed, can comprise one or more layers of a photoresist, or alternately a polymer material such as PMMA or PDMS. In some embodiments of the present invention, the mask 56 can be shaped (e.g. by reflowing the photoresist, by molding the PMMA or PDMS, etc.) to form the electrode tips 20 with different lengths across the electrode array 10 (e.g. to form the electrode tips 20 with a convex shape). This can be advantageous for equalizing or limiting the displacement of the electrode tips 20 when the electrode array 10 is to be urged into contact with a curved surface, and thereby equalizing a contact force of the electrodes 12 when the springs 16 all have substantially the same spring constant.

[0089] Although a plurality of receptacles 46 are shown in FIG. 4K, the formation of receptacles 46 is optional when the electrode tips 20 are to be electroplated directly onto the electrode seats 18. Omitting formation of the receptacles 46 and etching the electrode seats 18 to the same thickness as the remainder of the substrate 14 can facilitate the formation of the mask 56 especially when a sheet resist material such as PMMA is used to allow patterning of the mask 56 by deep x-ray lithography, with the sheet PMMA material being attached to the substrate 14 with an adhesive.

[0090] Once the patterned mask 56 has been provided over the backside of the substrate 14 as shown in FIG. 5A and the openings 58 formed at the locations wherein the electrode tips 20 are to be formed, the electrode tips 20 can be electroplated to a predetermined length as previously described with reference to FIG. 4N. This can involve the deposition of an electroplating seed layer on a surface of each electrode seat 18 to initiate (i.e. seed) the electroplating, or for use in improving the adhesion of the electroplated electrode tips 20 to the electrode seats 18. Once the electroplating of the electrode tips 20 has been completed as shown in FIG. 5B, the patterned mask 56 can be removed (e.g. with a solvent such as acetone or a commercial photoresist remover) to leave the electrode tips 20 permanently attached to the electrode seats 18 as shown in FIG. 5C. Removing the layers of the sacrificial material 38 by selective etching then releases the electrodes 12 for movement and completes the formation of the electrode array 10.

[0091] Although the electrode tips 20 are shown being formed on an underside of the electrode seats 18 in FIGS. 5A-5C, those skilled in the art will understand that the electrode tips 20 can also be formed on a topside of the electrode seats 18. This can be done, for example, by etching an opening down through the layers of the sacrificial material 38 at a center of each electrode seat 18 to expose the electrically-insulating layer 24 covering the electrode seat 18, or by etching the opening down to the surface of each electrode seats 18. Then, an electroplating seed layer can be deposited in each opening on the electrically-insulating layer 24 or on the exposed surface of each electrode seat 18. A patterned mask 56 can then be provided over the sacrificial material 38, with a plurality of openings 58 therein at the locations where each electrode tip 20 is to be formed. The electrode tips 20 can then be electroplated to fill the openings 58; and the patterned mask 56 and the layers of the sacrificial material 38 can be removed to complete the electrode array 10.

[0092] In yet other embodiments of the present invention, the electrode array 10 can be formed with the electrode tips 20 comprising the semiconductor substrate material (e.g. silicon). This can be done by utilizing the process steps described previously with reference to FIGS. 4A-4I and then utilizing a series of substitute process steps described hereinafter with reference to FIGS. 6A-6F.

[0093] In FIG. 6A, a patterned etch mask 60 can be provided on the backside of the substrate 14 at the locations where each electrode tip 20 is to be formed. The etch mask 60 can be formed by re-patterning the etch mask 42 in FIG. 4I, or alternately by removing the etch mask 42 in FIG. 4I and providing a new patterned etch mask.

[0094] In FIG. 6B, a the substrate 14 can be etched completely through to separate the electrode seats 18 from the remainder of the substrate 14, and to define the shape of the electrode tips 20. This etching step can utilize the Bosch process as described previously.

[0095] In FIG. 6C, the etch mask 60 can be removed (i.e. stripped). At this point, the layers of the sacrificial material 38 can be etched away as described previously to release the individual electrodes 12 for movement, thereby completing an electrode array 10 in which the electrode seats 18 and electrode tips 20 are formed integrally from the semiconductor substrate material. Since the semiconductor substrate material is electrically conductive, the electrodes 12 in FIG. 6C are fully functional and suitable for certain applications.

[0096] In other embodiments of the present invention, a coating of a metal or metal oxide can be provided over the electrodes 12 in FIG. 6C for improved electrical conductivity. This can be done in different ways as described hereinafter.

[0097] FIG. 6D shows a partial coating 62 of a metal (e.g. platinum, iridium, titanium, gold, or a combination thereof) or a metal oxide (e.g. iridium oxide) which can be provided over the exposed ends of the electrode tips 20, and which can also extend upward from the ends of the electrode tips 20 for a distance of, for example, 1-10 μm. Such a partial coating 62 can be formed, for example, by evaporation or sputtering, or by electroplating (e.g. by immersing the electrode tips 20 partway into a plating bath). An optional thermal oxidation step can be used, if needed, to convert an iridium metal coating 62 into an iridium oxide metal coating 62.

[0098] Once the partial coating 62 has been formed over the electrode tips 20, the electrically-insulating layer 24 can be etched away underneath the springs, and the layers of the sacrificial material 38 can be removed as described previously. This releases the electrodes 12 for movement. The electrically-insulating layer 24 can be left intact for other embodiments of the present invention wherein the electrodes 12 are to be stationary as described hereinafter with reference to FIGS. 13-16.

[0099] An optional electrically-insulating biocompatible coating 64 (e.g. parylene, silicon dioxide or silicon nitride) can be provided over the entire electrode array 10 using a conformal deposition process (e.g. LPCVD or VDP) as described previously. The coating 64 can then be removed from over the partial coating 62 as shown in FIG. 6E (e.g. by laser ablation, selective etching or solvent dissolution).

[0100] FIG. 6F shows a full-surface coating 66 of a metal or metal oxide formed over the exposed portions of the electrode seats 18 and electrode tips 20. The full-surface coating 66, which can be deposited by evaporation or sputtering, or alternately electroplated, can comprise a metal such as platinum, iridium, titanium or gold, or a combination thereof. An iridium oxide full-surface coating 66 can be formed, for example, by first depositing or electroplating an iridium coating 66 and then thermally oxidizing at least a portion of the iridium to form iridium oxide.

[0101] Once the full-surface coating 66 has been formed and the layers of the sacrificial material 38 have been removed, an optional electrically-insulating biocompatible coating 64 can be provided over the entire electrode array 10 and then removed from the ends of the electrode tips 20 as previously described with reference to FIG. 6E.

[0102] The electrode array 10 of the present invention can be used in conjunction with electronic circuitry 110 (e.g. neural stimulation circuitry which forms a part of a neural prosthesis such as an implantable retinal prosthesis) which can be formed on a separate substrate. The electronic circuitry 110, which can include a plurality of interconnected complementary metal-oxide-semiconductor (CMOS) transistors and one or more photodetectors, can be electrically connected to the electrode array 10 through a plurality of wires, or be directly attached to the electrode array 10 as shown in FIG. 7.

[0103] FIG. 7 schematically illustrates in cross-section view the attachment of electronic circuitry 110 to the electrode array 10. This can be done using an intermediary substrate 112 which includes a plurality of electrically-conductive posts 114 that can be aligned to features (e.g. the addressing circuitry 26 or bond pads 32) on the electrode array 10 and attached thereto using a plurality of solder bump bonds 116, or an electrically-conductive adhesive.

[0104] The intermediary substrate 112 can comprise, for example, a semiconductor substrate 112 that has been bulk micromachined to shape, or a substrate 112 whereon the electrically-conducting posts 114 are formed by LIGA to protrude outward from the substrate 112 or to extend therethrough, or a ceramic substrate 112, etc. The essential characteristic of the intermediary substrate 112 is that it provides a plurality of electrical connections between the electronic circuitry 110 (e.g. an integrated circuit chip) and the electrode array 10 while holding these two parts together. Electrically-conducting vias 118 can be provided through the intermediary substrate 112, as needed, to provide electrical conduction paths from a top surface of the substrate 112 to an underside thereof where the electrically-conducting posts 114 are located.

[0105] Once the electronic circuitry 110 has been attached to the electrode array 10 to form the apparatus shown in FIG. 7, the apparatus can be optionally coated for biocompatibility when this is needed (e.g. for forming a neural prosthesis). This biocompatibility can be provided, for example, by a conformal coating of parylene using VDP as described previously. Since the parylene can be deposited from a vapor at room temperature, all surfaces of the electronic circuitry 110, intermediary substrate 112, electrode array 10, solder bump bonds 116, external wiring (not shown), etc., can be provided with a biocompatible coating of parylene that can be, for example, up to 0.1 μm thick or more.

[0106] The electrode array 10 of the present invention can further include a flexible frame 70 attached to the substrate 14 for use in handling the electrode array 10, or for attaching the electrode array 10 onto a surface (e.g. a neural surface such as the epiretinal surface) to be electrically contacted by the electrode array 10. One example of such a flexible frame 70 for use with the electrode array 10 for use in an implantable retinal prosthesis 200 is schematically illustrated in FIGS. 8A and 8B. Although not shown in FIGS. 8A and 8B, the flexible frame 70 can optionally include electrical wiring (e.g. for forming a plurality of electrical connections to additional electronic circuitry not directly attached to the electrode array 10).

[0107] In the schematic plan view of FIG. 8A and the cross-section view of FIG. 8B, the flexible frame 70 can be attached to the electrode array 10 and to any electronic circuitry 110 that is directly attached to the electrode array 10 to form the implantable retinal prosthesis 200. The flexible frame 70 can include one or more wings 72 which extend outward from the electrode array 10, with each wing 72 including one or more through-holes 74 for use in attaching the flexible frame 70 and electrode array 10 to an epiretinal surface 130 shown in FIG. 9B.

[0108] Further details of the flexible frame 70 can be seen in FIG. 8B which shows a schematic cross-section view along the section line 2-2 in FIG. 8A. In FIG. 8B, the flexible frame 70 can be attached to the electrode array 10 at an outer edge thereof, and can further be attached to the outer edge of any electronic circuitry 110 included with the electrode array 10. The flexible frame 70 can comprise a biocompatible polymer such as PDMS that can be molded into shape and cured (e.g. at an elevated temperature of about 80° C. for one hour in the case of PDMS). The flexible frame 70 can include one or more recesses 76 adapted to receive and secure the electrode array 10 and any electronic circuitry 110 at the outer edges thereof. Alternately, the flexible frame 70 can be molded around the outer edges electrode array 10 and any electronic circuitry 110 and can optionally cover a topside of the electronic circuitry 110.

[0109] The flexible frame 70 provides a convenient way of handling the electrode array 10 and any electronic circuitry 110 during implant surgery when these elements form a part of an implantable retinal prosthesis 200. For implantation of the retinal prosthesis 200, the wings 72 can be folded as shown in FIG. 9A to allow manipulation of the assembled electrode array 10 and electronic circuitry 110 by a surgeon using a an insertion tool 120 (e.g. a pair of forceps, or a specially-designed tool) which grasps and holds the ends of each wing 72. This allows the retinal prosthesis 200 to be inserted through a surgical incision in a patient's eye with minimal concern about damaging the electrode array 10 or the patient's eye.

[0110] Once the electrode array 10 and other intraocular portions of the retinal prosthesis 200 have been implanted into a patient's eye using conventional ophthalmic surgical techniques and positioned near the epiretinal surface 130, the wings 72 which have been secured by the insertion tool 120 can be released and allowed to unfold thereby aiding in placement and of the retinal prosthesis 200. This is schematically illustrated in FIG. 9B. The resiliency of the wings 72 can also help to urge the electrode array 10 into contact with the epiretinal surface 130.

[0111] Once the retinal prosthesis is in place as shown in FIG. 9B, each wing 72 can be secured to the epiretinal surface 130 using one or more sutures, or with a titanium tack 140 inserted in each through-hole 74. Each titanium tack 140 can pierce the retina, choroid and sclera to hold the electrode array 10 in place against the epiretinal surface 130. The ability of each electrode 12 in the array 10 to move independently upon a plurality of springs 16 allows the electrode array 10 to gently conform to the curvature of the epiretinal surface 130 while at the same time allowing each individual electrode 12 to maintain substantially the same low contact force on the epiretinal surface 130.

[0112] Operation of the retinal prosthesis 200 can utilize a radio-frequency (rf) antenna or alternately a photodetector located within the eye and connected to the electronic circuitry 110 for powering the electronic circuitry 110 and for providing visual information to the electronic circuitry 110 and the electrode array 10. The rf antenna and receiving electronic circuitry, which are not shown in FIGS. 9A and 9B, can be located remotely from the electronic circuitry 110 and connected thereto using flexible wiring. The rf antenna, receiving electronic circuitry and flexible wiring can all be folded up within the folded wings 72 in FIG. 9A to facilitate implanting of the retinal prosthesis 200 in a patient's eye. When the wings 72 are released as shown in FIG. 9B, the rf antenna, receiving electronic circuitry and flexible wiring cab unfold and assume a position in the patient's eye away from the electrode array 10 to help alleviate power dissipation and tissue heating concerns which might otherwise occur if the antenna and receiving circuitry were directly located on the electrode array 10.

[0113] The retinal prosthesis 200 generally provides a balanced biphasic stimulation (i.e. a positive current pulse paired with a slightly-delayed negative current pulse of equal and opposite charge) through the electrode array 10 to the underlying ganglion cells in the retina since this provides a needed electrical stimulation required for visual perception while eliminating any net charge accumulation on the electrodes 12 that might otherwise lead to electrochemical reactions and possible dissolution of the electrodes 12 in the array 10. The electrical stimulation provided through the electrode 12 to the underlying ganglion cells in the retina can be on the order of up to a few microamperes current with a pulse width of, for example, 0.1-2 milliseconds and a pulse repetition rate of up to several tens of Hertz. The electrical stimulation requirements for other types of neural tissue can be learned from practice of the present invention.

[0114] In other embodiments of the present invention, the springs 16 in the electrode array 10 can be formed from other electrically-conductive materials (e.g. metals), and can have other shapes as shown in the examples of FIGS. 10A-10D. In FIG. 10A, the springs 16 comprise folded springs 16, but with a single 90° fold for each spring 16. In FIG. 10B, the springs 16 can be folded about a circular electrode seat with each folded spring 16 comprising a plurality of arcuate sections connected together by folds at the ends thereof. In FIG. 10C, each spring 16 can be curved with optional 90° folds where the spring 16 is attached to the substrate 14 and to the electrode 12. In FIG. 10D, each spring 16 can be formed as a curved spring without any folds if the electrode 12 and substrate 14 are shaped as shown. FIGS. 10A-10D are provided as examples and are not intended to limit the present invention in any way. Those skilled in the art will understand that there are many other ways of forming the plurality of springs 16; and that the electrodes can have an arbitrary shape (e.g. circular, square, polygonal, etc.).

[0115] FIG. 11 shows a schematic plan view of another example of the electrode array 10 of the present invention formed using a silicon-on-insulator substrate 14. The silicon-on-insulator substrate 14 comprises a monocrystalline silicon layer 80 disposed on a monocrystalline silicon body 82 and separated therefrom by an intervening electrically-insulating layer 24 of silicon dioxide. Silicon-on-insulator substrates 14 are available commercially with the monocrystalline silicon layer 80 and the electrically-insulating layer 24 formed integrally with or bonded to the monocrystalline silicon body 82. The monocrystalline silicon layer 80 and the electrically-insulating layer 24 can each have a layer thickness of up to a few microns (e.g. 0.2-2 μm); and the monocrystalline silicon body 82 can have an initial thickness of, for example, 0.5 millimeters (mm) which can be thinned down to 0.2-0.3 mm with the electrodes 12 in the completed device 10 being about 0.5 mm long. In this example of the electrode array 10, the springs 16 can be formed by patterning the monocrystalline silicon layer 80 using reactive ion etching. This patterning step can also form a monocrystalline silicon layer portion 80′ which overlies each electrode 12 and is attached to the remainder of the monocrystalline silicon layer 80 via the springs 16. Additionally, a plurality of transistors 28 can be formed in the monocrystalline silicon layer 80, or the layer portion 80′ or both for use in addressing the individual electrodes 12 in the array 10, to form an amplifier for each electrode 12, or to form electronic circuitry 110 for operating the electrode array 10 (e.g. for providing biphasic stimulation signals to individual electrodes 12 in the array 10 in an implantable retinal prosthesis incorporating one or more electrode arrays 10).

[0116] One or more electrically-conducting vias 84 can be formed through the electrically-insulating layer 24 to electrically connect the monocrystalline silicon layer portion 80′ to a monocrystalline silicon body portion 82′ which provides the seat 18 and tip 20 for each electrode 12. The electrically-conducting vias 84 can be formed by depositing a layer of polysilicon, or a metal (e.g. tungsten, aluminum, titanium, gold, platinum, iridium or alloys thereof), or both into one or more openings etched down through the layers 80 and 24 to contact the monocrystalline silicon body portion 82′. In this way, the monocrystalline silicon body portion 82′ can be electrically connected through the via 84 to the monocrystalline silicon layer portion 80′ and therefrom through one or more of the springs 16 to the switching transistors 28 and addressing circuitry 26.

[0117] In FIG. 12A, the monocrystalline silicon body 82 can be patterned to form the body portion 82′ comprising an electrode seat 18 and tip 20. This can be done by etching the monocrystalline silicon body 82 from the backside of the substrate 14 as previously described with reference to FIGS. 6A-6C. The electrically-insulating material 24 can be selectively etched away from below the springs 16 and between the monocyrstalline silicon body 82 and body portion 82′ using the Bosch etch process as described previously with reference to FIGS. 4I-4K and 6A-6C, or alternately using a selective etchant comprising HF. If the electrically-insulating material 24 is removed using the selective etchant comprising HF, the etching step can be timed to remove the material 24 between the portions 80′ and 82′ while not substantially removing the material 24 which is directly above the portions 80′ and 82′. In some embodiments of the present invention, the vial 84 can be annular in shape to form an etch stop to limit etching of the electrically-insulating material 24 directly above the monocrystalline silicon body portion 82′ since the electrically-conducting material (e.g. polysilicon and/or certain metals such as aluminum, tungsten, platinum, gold, etc.) used for the via 84 is substantially resistant to etching by HF. Such an annular via 84 can be made relatively narrow (e.g. 1-2 μm wide) so that the annular via 84 can be completely filled with the electrically-conducting material. Any of the electrically-conducting material extending above the monocrystalline silicon layer portion 80′ can be optionally removed (e.g. using a chemical-mechanical polishing step).

[0118] Although not shown in FIG. 12A, a coating of a metal or metal oxide can optionally be provided over the electrodes 12 for improved electrical conductivity. A partial coating 62 over the tip 20 of each electrode 12 can be formed as previously described with reference to FIG. 6D; and a full-surface coating 66 can be formed as described with reference to FIG. 6F. A electrically-insulating biocompatible coating 64 can also be provided over the entire electrode array 10 of FIGS. 11 and 12A and then removed from the ends of the electrode tips 20 as previously described with reference to FIG. 6E.

[0119] Operation of this example of the electrode array 10 is schematically illustrated in FIG. 12B which shows a schematic cross-section view of a portion of the device 10 urged into contact with a curved surface 100 (e.g. an epiretinal surface). An independent movement of each electrode 12 in the array 10 is made possible by the springs 16. This allows the electrode array 10 to conform to the curved surface 100, or to other surfaces of arbitrary shape. The springs 16 can be used to sense a contact force for one or more of the electrodes 12 when the electrode array 10 is urged into contact with the curved surface 100. This can be done, for example, using piezoresistive sensing as described previously with reference to FIGS. 3A and 3B.

[0120] Another example of an electrode array 10 formed from a silicon-on-insulator substrate 14 is shown schematically in plan view in FIG. 13. In this example of the electrode array 10 of the present invention, each electrode 12 is fixed in place on the silicon-on-insulator substrate 14 without any springs 16. Each electrode 12 in this example of the present invention can be made independently addressable, or addressable as part of a set of electrodes 12.

[0121] A substantially rigid (i.e. immovable) attachment of each electrode 12 to the silicon-on-insulator substrate 14 can be provided by leaving the electrically-insulating layer 24 intact between the monocrystalline silicon body portion 82′ of each electrode 12 and the remainder of the monocrystalline silicon body 82 after forming the base 18 and tip 20 of each electrode 12 as previously described with reference to FIGS. 6A-6C. This can be done, for example, by omitting a selective etching step with HF, or alternately by limiting the depth of the Bosch etching so that the electrically-insulating layer 24 is not substantially removed. The substantially rigid attachment of each electrode 12 to the silicon-on-insulator substrate 14 can be seen in FIG. 14 which is a schematic cross-section view taken along the section line 4-4 in FIG. 13. The exact rigidity of each electrode 12 in the example of FIGS. 13 and 14 will depend upon the thickness of the electrically-insulating layer 24 and the monocrystalline silicon layer 80. Each of these layers 24 and 80 will generally have a thickness in the range of 0.2-2 μm, but can be made thicker, as needed, for particular applications. Additionally, the rigidity to the electrodes 12 in this example can be increased by depositing an electrically-insulating material (e.g. parylene, silicon nitride, silicon dioxide or a silicate glass) within an annular gap 86 formed about each electrode 12. A layer of the electrically-insulating material up to a few microns thick can also be deposited over exposed surfaces of the electrode array 10 to provide added rigidity, electrical insulation and biocompatibility.

[0122] One or more transistors 28 can be formed in the monocrystalline silicon layer portion 80′ of each electrode 12 in the array 10 and electrically connected to the body portion 82′ for that electrode 12 by an electrically-conducting via 84 formed down through the electrically-insulating layer 24. The provision of one or more transistors 28 on the monocrystalline silicon layer portion 80′ of each electrode 12 can save space and allow the electrodes 12 in the array 10 to be spaced close together. The transistors 28 can be switching transistors for addressing each electrode 12 in the array as schematically illustrated in FIG. 13. In other embodiments of the present invention, a plurality of transistors 28 can be arranged to form an amplifier circuit which can be located on each electrode 12 to amplify an electrical signal being sensed with the electrode array 10 (e.g. a neural signal), or being supplied thereto.

[0123] The formation of transistors 28 on a silicon-on-insulator substrate 14 is well-known in the art and need not be repeated here in great detail. Each transistor 28 is electrically isolated from the substrate body portion 82′ by the electrically-insulating layer 24 comprising silicon dioxide, and can be further isolated from the remainder of the monocrystalline silicon layer 80 and layer portion 80′ by an isolation trench 88 formed thereabout through the monocrystalline silicon layer portion 80′ as shown in FIG. 14, with the isolation trench 88 being, for example, 1-2 μm wide. The isolation trench 88 can be filled with silicon dioxide, silicon nitride or a silicate glass (e.g. TEOS). An isolation trench 88 can also be used to electrically isolate the via 84 from the remainder of the monocrystalline silicon layer portion 80′. Ion implantation or dopant diffusion can be used to form drain and source regions for the transistors (e.g. FETs) in the monocrystalline silicon layer portion 80′.

[0124] A gate oxide (e.g. a thermal oxide) can be formed from the monocrystalline silicon layer portion 80′ using a conventional wet oxidation process at an elevated temperature. Alternately, the gate oxide can comprise a deposited layer of silicon dioxide or silicon nitride. A gate 90 for each transistor 28 can be formed from a deposited polysilicon layer (e.g. Poly-0). Additional layers of polysilicon, or a metal or both can be used to form the addressing circuitry 26 which can be electrically isolated from the monocrystalline silicon layer 80 by another electrically-insulating layer 92 which can comprise the gate oxide with a thickness of, for example, 0.6 μm and a layer of silicon nitride which can be, for example, 0.8 μm thick.

[0125] Multiple levels of the addressing circuitry 26 can be separated and encapsulated by one or more additional electrically-insulating layers (not shown) which can comprise, for example, a silicate glass (e.g. TEOS). This can also protect the transistors 28 and addressing circuitry 26 during patterning of the backside of the substrate 14 to form the bases 18 and tips 20 for the electrodes 12.

[0126] Although not shown in FIG. 14, a coating of a metal or metal oxide can optionally be provided over the electrodes 12 for improved electrical conductivity. A partial coating 62 can be formed over the tip 20 of each electrode 12 as previously described with reference to FIG. 6D. Alternately, a full-surface coating 66 can be formed over each electrode 12 as previously described with reference to FIG. 6F. An electrically-insulating biocompatible coating 64 can also be provided over the entire electrode array 10 of FIGS. 13 and 14 and then removed from the ends of the electrode tips 20 as previously described with reference to FIG. 6E. The electrically-insulating biocompatible coating 64 can also cover external electrical wiring (not shown) which can be connected to the electrode array 10 through a plurality of bond pads 32.

[0127] The electrode array 10 of FIGS. 13 and 14 can be formed with up to hundreds or thousands of individual electrodes 12 in an overall size of, for example, two millimeters across. Furthermore, a plurality of electrode arrays 10 of arbitrary shape (e.g. polygonal, circular or elliptical) can be arranged side-by-side. This can be done, for example to form a retinal prosthesis 200 as shown in FIGS. 15A and 15B which includes a plurality of electrode arrays 10 which are arranged as tiles in a flexible frame 70. The flexible frame 70 can allow the tiles to move to conform to a curved surface such as the epiretinal surface 130 shown in the schematic cross-section view of FIG. 15B. By using a plurality of electrode arrays 10 with fixed electrodes as small-size tiles in a flexible frame 70 as shown in FIGS. 15A and 15B, the amount of displacement and pressure on the epiretinal surface 130 can be reduced compared to a single larger-size electrode array 10 having fixed electrodes. This can also provide an alternative to the use of an electrode array 10 having spring-loaded electrodes 12 as shown previously in FIGS. 9B and 12B. In the case of an epiretinal surface 130 having a radius of curvature of 11.5 mm, the electrodes 12 nearest the edge of an electrode array 10 of 2-mm size will extend about 29 μm further into the epiretinal surface 130 than the electrodes 12 at the center of the array 10. Conversely, if the edge electrodes 12 are just touching the epiretinal surface 130, the center electrodes will be this same distance (29 μm) away from the retinal tissue. In certain embodiments of the present invention, the electrodes 12 can be pointed so that they can pierce the retinal tissue or other types of neural tissue.

[0128] FIG. 15A shows the retinal prosthesis 200 with a plurality of hexagonal electrode arrays 10, with each electrode array 10 having a structure as shown, for example, in FIGS. 13 and 14. The electrode arrays 10 can be assembled into a flexible frame 70 which can comprise a biocompatible polymer such as PDMS. The frame 70 can include one or more wings 72 that extend outward to allow the frame 70 and electrode arrays 10 to be attached to an epiretinal surface 130. This can be done with titanium tacks 140 as shown in FIG. 15B, or with sutures. Although not shown in FIGS. 15A and 15B, electrical wiring can be provided in the retinal prosthesis 200 to electrically connect each electrode array 10 to additional electronic circuitry (e.g. to address the various electrodes 12 in each array 10 and to provide biphasic electrical stimulation signals to each electrode 12 as needed).

[0129] An electrode array 10 similar to that shown in FIG. 13 can also be formed using a silicon substrate 14 without a monocrystalline silicon layer 80. A schematic cross-section view of one electrode 12 of such an array 10 is shown in FIG. 16. In this example of the present invention, the electrically-insulating layer 24 can comprise a layer of a thermal oxide, or a layer of silicon nitride or both. The thermal oxide layer, which can be formed over the entire top surface of the substrate 14 using a conventional wet oxidation process at an elevated temperature (e.g 1050° C. for about 1.5 hours), can be about 0.6 μm thick. The silicon nitride layer can be deposited by LPCVD at about 850° C. with a layer thickness of, for example, 0.8 μm. Openings can be formed down through the electrically-insulating layer 24 at locations wherein the transistors 28 are to be formed and also at the locations wherein the electrically-conducting vias 84 will be formed. One or more transistors 28 can be formed in a substrate portion 14′ which is patterned to define the base 18 and tip 20 for each electrode 12, or alternately on the remainder of the substrate 14. The transistors 28 can comprise switching transistors for addressing each electrode 12, or can comprise amplifying transistors for amplifying a signal to be received by each electrode 12.

[0130] After forming the transistors 28 in the silicon substrate 14, the addressing circuitry 26 and a plurality of bond pads 32 can be formed above the electrically-insulating layer 24. The addressing circuitry 26 and the bond pads 32, together with a gate 90 for each transistor 28 can be formed from a deposited polysilicon layer (e.g. Poly-0), or a layer of a metal (e.g. aluminum or tungsten), or a combination thereof. One or more additional electrically-insulating layers 24′ can then be blanket deposited over the substrate 14.

[0131] The backside of the substrate 14 can then be patterned as described previously with reference to FIGS. 6A-6C to form the electrode base 18 and tip 20 while retaining the electrically-insulating layer 24 substantially intact. This electrically isolates the substrate portion 14′ from the remainder of the substrate 14 and makes each electrode 12 substantially rigid. As previously discussed, with reference to FIG. 14 an electrically-insulating material (e.g. parylene, silicon nitride, silicon dioxide or a silicate glass) can be deposited within the annular gap 86 formed about each electrode 12 to provide added stiffness for the electrodes.

[0132] The electrode tip 20 can also be optionally metallized to form a partial coating 62 as previously described with reference to FIG. 6D, or a full-surface coating 66 as previously described with reference to FIG. 6F. The partial coating 62, which is shown in FIG. 16, can extend partially over the sides of the electrode tip 20. An electrically-insulating biocompatible coating 64 as previously described with reference to FIG. 6E can also be provided over the entire electrode array 10 of FIG. 16 and then removed from the ends of the electrode tips 20.

[0133] In other embodiments of the present invention, an electrode array 10 can be formed as described with reference to FIG. 16 except with the transistors 28 omitted. In this case, each electrode 12 or set of electrodes 12 will generally be directly connected to a separate bond pad 32.

[0134] Although the electrode array 10 of the present invention has been described in relation to use in an implantable retinal prosthesis 200, those skilled in the art will understand that the electrode array 10 described herein has applications for use in stimulating or sensing many different types of neural tissue including neural tissue associated with visual, auditory and sensory systems and neural tissue associated with the control of particular muscles (e.g. for bladder function or the activation of paretic limbs) or organs. Other applications and variations of the present invention will become evident to those skilled in the art. For example, other embodiments of the electrode array 10 of the present invention can be formed with the individual electrodes 12 being pointed for piercing neural tissue. Such pointed electrodes 12 can be formed by etching or electroforming. The matter set forth in the foregoing description and accompanying drawings is offered by way of illustration only and not as a limitation. The actual scope of the invention is intended to be defined in the following claims when viewed in their proper perspective based on the prior art.