WO2020035657A1

WO2020035657A1 - Forming electrical connection between wire electrode and metallic contact surface

Info

Publication number: WO2020035657A1
Application number: PCT/GB2019/052141
Authority: WO
Inventors: Ciprian BULZ; Ravi Desai; Andreas Schaefer; Romeo-Robert Racz
Original assignee: The Francis Crick Institute Limited
Priority date: 2018-08-14
Filing date: 2019-07-31
Publication date: 2020-02-20
Also published as: GB201813235D0; GB2576498A

Abstract

An electrical connection is formed between a wire electrode (30) comprising a metallic core (32) surrounded by a sheath (34) of insulating material and a metallic contact surface (20) of an integrated circuit. Mechanical force is applied to break the sheath (34) of insulating material and expose part of the metallic core (32) of the wire electrode. A bonding wire (8) made of the same metallic material as the metallic core (32) is provided in a capillary (6), and at least one of electrical charge, heat and ultrasonic vibration is applied to a tip of the bonding wire (8) and the capillary (6) is pressed towards the wire electrode, to melt the tip of the bonding wire to form a bond of metallic material which connects the exposed core of the wire electrode to the metallic contact surface (20) of the integrated circuit. This helps to reduce bond resistance.

Description

FORMING ELECTRICAL CONNECTION BETWEEN WIRE ELECTRODE AND METALLIC

CONTACT SURFACE

The present technique relates to the field of wire bonding. More particularly, it relates to forming an electrical connection between a wire electrode and a metallic contact surface of an integrated circuit.

Recently it has been proposed that wire electrodes comprising a metallic core surrounded by a sheath of insulating material can be used as an electrochemical probe. For example, such probes have been proposed for use in the brain for making neural measurements or performing neural stimulation for therapeutic purposes. These are just some examples of potentially use cases for such wire electrodes and further examples are discussed below.

Such probes may be bonded to an integrated circuit which may provide electronics for reading out signals measured using the probes or for generating stimulation signals to be applied using the probes. The resistance at the bond between the wire electrode and a metallic contact surface of the integrated circuit may be a significant factor in the overall performance of the probe. Existing bonding techniques may result in relatively high resistance at the interface between the wire electrode and the metallic contact surface of the integrated circuit.

At least some examples provide a method of forming an electrical connection between a wire electrode and a metallic contact surface of an integrated circuit, the wire electrode comprising a metallic core surrounded by a sheath of insulating material; the method comprising: positioning the wire electrode over the metallic contact surface; applying mechanical force to the wire electrode to break the sheath of insulating material and expose part of the metallic core of the wire electrode; providing, within a capillary, a bonding wire made of the same metallic material as the metallic core of the wire electrode; andapplying at least one of electrical charge, heat and ultrasonic vibration to a tip of the bonding wire and pressing the capillary carrying the bonding wire towards the wire electrode, to melt the tip of the bonding wire to form a ball bond of metallic material which connects the exposed core of the wire electrode to the metallic contact surface of the integrated circuit.

At least some examples provide a bonding machine comprising: a wire-carrying capillary for carrying a bonding wire; at least one bonding control element to provide at least one of electrical charge, heat and ultrasonic vibration to a tip of the bonding wire carried within the capillary; in which: the capillary has a tip with a sharpened edge.

Further aspects, features and advantages of the present technique will be apparent from the following description of examples, which is to be read in conjunction with the accompanying drawings, in which:

Figure 1 schematically illustrates an example of a bonding machine; Figure 2 illustrates a wire electrode having a metallic core and an insulating sheath;

Figure 3 shows example images of the tip of the wire electrode, with gold nano structures and an iridium oxide functionalisation layer deposited on the tip of the electrode;

Figure 4 is a graph showing how impedance at the front end of the wire electrode is reduced by including the gold nano-structures;

Figures 5A to 5F illustrate a process of bonding the wire electrode to a metallic contact surface of an integrated circuit using a capillary of the bonding machine;

Figure 6 is a flow diagram illustrating a method of bonding the wire electrode to the metallic contact surface;

Figures 7A to 7C show images of the capillary used for different steps of the bonding method;

Figures 8 and 9 are scanning electron microscope (SEM) images showing the bond between the wire electrode and the integrated circuit;

Figure 10 is a graph showing various measurements of the resistance at the bond for different example wires of different core metallic materials and diameters; and

Figures 1 1 to 18 show a number of examples of modifications of the capillary of the bonding machine.

It has been found that when bonding a wire electrode comprising a metallic core and a sheath of insulating material to a metallic contact surface of an integrated circuit, the presence of the insulating material may make it difficult to form a useful connection to the integrated circuit with sufficiently low resistance. In the method discussed below, the wire electrode is positioned over the metallic contact surface, and a mechanical force is applied to the wire electrode to break the sheath of insulated material and expose part of the metallic core of the wire electrode. A bonding wire made of the same metallic material as the metallic core of the wire electrode is provided within a capillary. At least one of electrical charge, heat and ultrasonic vibration is applied to a tip of the bonding wire carried by the capillary, and the capillary carrying the bonding wire is pressed towards the wire electrode. The charge, heat and/or ultrasonic vibration melts the tip of the bonding wire, and the pressure from the capillary causes the molten material to form a bond of metallic material which connects the exposed core of the wire electrode (which was exposed by breaking the sheath of insulating material) to the metallic contact surface of the integrated circuit. Flence, by breaking part of the sheath of insulating material this allows more of the metallic core to be exposed so that there is greater surface area for forming the bond between the wire electrode and the metallic contact surface. By using a bonding wire made of the same metallic material as the core of the wire electrode, a bond with lower resistance can be provided as there is no amalgam of different metallic materials at the bond location. In some examples, the capillary may be a capillary of a ball bonder machine. A ball bonder machine may have a capillary (a hollow tube with an internal bore passing along the length of the tube) through which a wire can be passed, a position control unit to control the position of the capillary relative to the integrated circuit (both laterally in the plane of the integrated circuit and vertically towards and away from the integrated circuit), and a heating element for heating a tip of the wire carried in the capillary to melt it (e.g. the heating element can supply an electrical arc to heat the wire tip). In the field of wire bonding, a ball bonder machine would normally be used for bonding a metallic wire to an integrated circuit, where the wire carried by the capillary would be the actual wire which is intended to form the electrical connection in the integrated circuit. However, as described below, the inventors recognised that the ball bonder machine can be repurposed so that instead of carrying the actual wire intended to form the electrical connection, the capillary carries a bonding wire which is molten to form the metallic material which is deposited in the bond between the metallic contact surface of the integrated circuit and the wire electrode, so that a bond with lower resistance can be formed for wire electrodes having an insulating sheath. An advantage of using a ball bonder machine to perform the bonding is that the ball bonder machine may already have control circuitry for precisely controlling the position of the capillary relative to the integrated circuit, and so can provide fine control over the location at which the bonding is performed. Hence, by reusing the ball bonder machine already used for integrated circuit bonding of entirely metallic wires, this simplifies the control of the bonding operation. However, by repurposing the ball bonder machine so that instead of the capillary carrying the electrical wire itself, the capillary carries a bonding wire which is molten to form a bond with the actual wire electrode to carry the electrical signals, this enables improved connections between wire electrodes having a sheath of insulating material and the integrated circuit contact surface.

The mechanical force applied to the wire electrode to break the sheath of insulating material could be applied by a separate element from the capillary of the ball bonder machine.

However, in some implementations the mechanical force may be applied to the wire electrode using the capillary itself. This may simplify the design as it reduces the number of mechanical elements required. Also, this enables the position control used to control the location of the capillary to also be used for controlling the location at which the sheath of insulating material is broken. This use of the capillary to provide a mechanical breaking force is completely different to the normal purpose of the capillary to carry the wire into position.

The capillary could be used to break the sheath of the insulated material, before then providing the bonding wire inside the capillary and performing the bonding. However, in some examples the bonding wire may already be provided inside the capillary when the mechanical force is applied to the wire electrode. This avoids the need to move the capillary into position twice, once for applying the mechanical force and another time for moving the bonding wire into position. Instead, the bonding wire can be provided in said capillary once and then the capillary can be moved down to break the sheath and then subsequently to perform the bonding. This can reduce the overall processing time for making the electrical connection between the wire electrode and the integrated circuit.

In some examples the capillary may have a tip with a sharpened edge. For example the sharpened edge may extend around all or part of a circumference of the tip of the capillary. While the capillary of a ball bonder machine would not normally be expected to be used for mechanical impact, by providing a modification where the tip of the capillary is sharpened, this makes the capillary better at breaking the sheath of insulating material for a given amount of force applied, and so can enable better exposure of the metallic core.

The method may include a step of removing loose pieces of broken insulating material, prior to forming the bond. The breaking of the sheath of insulating material may result in loose shards of broken insulating material in the vicinity of the exposed core, which if remaining when the ball bond is formed may increase the resistance of the bond. Hence, by including a step of removing such loose pieces of insulator before melting the tip of the wire electrode using a charge, heat, and/or ultrasonic vibration, this can reduce resistance and hence improve performance of the device comprising the electrode and integrated circuit.

The step of removing the loose insulating material can be performed in different ways. In one example the capillary of the ball bonder machine may be used to mechanically remove the loose pieces of broken insulating material. For example the capillary could be displaced from side to side or up and down so as to move the loose pieces of broken insulating material away from the region to be bonded.

In another example, the step of removing may comprise supplying at least one removal agent for removing the loose pieces of broken insulating material. For example the removal agent could comprise compressed air, so that a jet of compressed air is directed into the region to be bonded so as to blow away loose pieces of broken insulating material. In another example the at least one removal agent could comprise a solvent (e.g. a chemical in the liquid phase) for dissolving the insulated material. For example, the solvent can be any chemical capable of dissolving an oxide, for example hydrofluoric acid or sodium hydroxide. Also, organic solvents could be used as he chemical removal agent. In some examples a combination of different types of removal agents could be used, for example a jet of compressed air in combination with a supply of a chemical solvent.

In some implementations the removal agent could be supplied through a separate structure which is not attached to or formed integrally with the capillary. However, in other implementations the at least one removal agent may be supplied through a removal agent supply capillary which is either attached to, or is integrally formed with, the capillary used to carry the bonding wire. For example a second channel may be attached to the side of the capillary used to carry the bonding wire, or multiple channels could be supplied, one for compressed air and another for the solvent for example. By fixing at least one removal agent supply capillary to the wire-carrying capillary, the same position control mechanism can control the position of both capillaries so that they move together, which reduces the hardware scale of the bonding machine.

The metallic core of the wire electrode and the metallic material used in the bonding wire may be a metal or alloy. For example, the metallic core and bonding wire may be made of copper, silver, gold, iron, platinum, lead or other metals, or crystalline or amorphous alloy compositions such brass, bronze, platinum-iridium, lead-silver or magnetic alloys such as FeSiB.

Similarly, the metallic contact surface of the integrated circuit may be a metal or alloy. The metallic contact surface of the integrated circuit may be of the same metallic material as the metallic core of the wire electrode, or a different material. The metallic contact surface could be the substrate of the integrated circuit, or could be a contact pad or other conductive element disposed on top of the integrated circuit substrate.

The insulating material of the wire electrode may comprise an oxide or other non- metallic material. In one particular example the insulating material comprises glass. Other examples could use plastics as the insulating material.

The wire electrode may have a diameter less than or equal to 25 microns. Hence the wire electrode may comprise an ultramicroelectrode (UME) which is extremely narrow in comparison to probes normally used for neural measurements of stimulation or other electrophysiological applications. For example, it has recently been found that glass- ensheathed UMEs offer a promising platform for a range of electrophysiological applications. For example the wire electrodes may be Taylor-Ulitovsky wires made using the Taylor- Ulitovsky method. Such wire electrodes can be extremely useful for integration into tissue because of their form factors, smooth outer shank and (in some implementations) tapered tips. However, they can be challenging to interface with the integrated circuit at the back end of the wire electrode, as the exposed surface area of the metallic core at the back end of the wire is very small in comparison to the insulating material surrounding the core. By using the method discussed above to break part of the insulating sheath and use the capillary carrying the bonding wire to form the bond with the integrated circuit, this can provide much lower resistance at the back end of the wire than alternative techniques. This can be particularly useful for many medical or electrophysiological devices, because if the resistance at the interface between the wire electrode and the integrated circuit can be reduced, then this may allow signals of lower amplitudes to be detected when the probe is used for making recordings or measurements, enabling greater sensitivity. Alternatively, if the wire electrode is being used for delivering stimulation currents or voltages to a sample, then the reduced resistance at the interface between the wire electrode and the integrated circuit means that lower currents or voltages is needed to provide a certain amount of signal at the biological sample, and so lower power electronics can be used which can reduce power consumption and heat generated by the electronics, which can be particularly important for embedded implants for example.

In another example, a bonding machine may be provided which comprises a wire carrying capillary for carrying a bonding wire and at least one bonding control element to provide at least one of electrical charge, heat and ultrasonic vibration to a tip of the bonding wire carried within the capillary. The capillary may have a tip with a sharpened edge. As discussed above this makes the capillary better at breaking the insulating sheath of the wire electrode when the bonding machine is used to bond the wire electrode to a metallic contact surface of an integrated circuit. For example, the bonding machine may be a ball bonder machine for ball bonding of integrated circuit connections.

In some examples the bonding machine may also comprise at least one removal agent supply capillary attached to or integrally formed with the wire-carry capillary. This enables supply of a removal agent such as compressed air or a solvent for the purpose of removing broken insulating material when the ball bonder machine is used to bond the wire electrode to the integrated circuit. The attachment or integration of the removal agent supply capillary with the wire-carrying capillary means that a single position control mechanism may regulate the position of both capillaries relative to the integrated circuit.

In another example, a bonding machine may comprise a wire-carrying capillary for carrying a bonding wire; at least one bonding control element to provide at least one of electrical charge, heat and ultrasonic vibration to a tip of the bonding wire carried within the capillary; in which at least one removal agent supply capillary is attached to, or integrally formed with, the wire-carrying capillary.

Figure 1 schematically illustrates an example of a ball bonder machine 2 for performing bonding of wired connections to an integrated circuit 4 which may supported on a platform (not shown in Figure 1 ). The ball bonder machine 2 has a capillary 6 (a narrow tube with an internal bore for carrying a bonding wire 8 inside the capillary). It will be appreciated that the integrated circuit 4 and bonding wire 8 are provided when the machine 2 is in use but are not part of the machine itself. The capillary 6 is attached to a support 10 whose position relative to the integrated circuit 4 can be controlled by position control circuitry 12. For example the position control circuitry may control the lateral movement of the capillary 6 across the integrated circuit 4, and the vertical movement of the capillary towards and away from the integrated circuit 4. The support 10 may include a piezoelectric disk 14 and a horn 16 for transmitting ultrasonic vibrations generated by the piezoelectric disk 14 to the capillary 6, to provide vibration of the capillary 6 relative to the integrated circuit 4. The machine may also include an electronic flame off (EFO) wand 18 for providing a high voltage electric charge to the tip of the wire 8 carried inside the capillary 6. The combination of the electrical charge applied by the EFO wand 18 and the ultrasonic vibration provided by the horn 16 heats the wire tip and melts the tip to melt to form a ball, and the pressure of the molten wire tip against the metallic contact 20 of the integrated circuit (combined with ultrasonic vibrations to perform welding) bonds the wire to the metallic contact 20. For normal use of such a ball bonder machine 2, the wire 8 inside the capillary would be the actual electrical wire which is to be bonded with the integrated circuit 4 and which, in the manufactured circuit, is intended to carry electrical charge across the integrated circuit 4. Flowever, in the technique discussed below, the capillary 6 is instead used to carry a separate bonding wire and the actual electrode is a separate wire electrode not carried by the capillary 6.

Figure 2 shows an example of a wire electrode 30 to be bonded with the metallic contact 20 of the integrated circuit 4. The electrode 30 has a metallic core 32 which is surrounded by a sheath 34 of insulating material. While Figure 2 shows an example with a single metallic core 32 inside the sheath 34, other examples could provide multiple core wires inside the same sheath 34. In this example the metal core 32 is made of gold, but other examples of conducting materials which could be used include copper, silver, gold, iron, platinum, lead or other metals, as well as crystalline or amorphous alloy compositions such as brass, bronze, platinum-iridium, lead-silver and magnetic alloys such as FeSiB. In this example, the insulating material 34 is glass, but other examples could use plastics or other insulators. The electrode 30 is an ultramicroelectrode (UME) having a diameter less than or equal to 25 mhi. In other examples, the wire electrodes may be even narrower, for example with a diameter less than or equal to 20 mhi, less than or equal to 15 mhi, less than or equal to 10 mhi or less than or equal to 5 mhi.

The wire electrode can be used as an electrochemical probe. An electrochemical probe may be a probe for current and/or voltage measurements or injection in biological samples and a range of electrophysiological applications, or a probe for determination of the presence and/or quantity of one or more biologically and/or toxicologically significant substances in biological and/or liquid samples. A front end of the wire electrode is used as the end of the probe for interfacing with the sensing/stimulation target, and a back end of the wire electrode is bonded to the integrated circuit 4 for transmitting signals measured from the sensing target to read out electronics or data processing equipment, or for receiving stimulation signals from the integrated circuit 4 for applying to the target at the front end. As shown in Figures 2 and 3, at the front end, the wire electrode 30 has an impedance reducing layer of gold nano-structures 35 deposited on the tip of the wire electrode. Also, a functionalisation layer 37 can be deposited on top of the gold nano structures 35, of a material chosen depending on the particular purpose for which the probe is intended to be used. For example, the functionalisation layer could be an iridium oxide (IrOx) functionalisation layer comprising a layer of iridium oxide nano-structures deposited on top of the gold nano-structures, or could be other metal oxides such as titanium dioxide or manganese oxides, or carbon nanotubes, graphene, ATP, DNA or proteins, etc. An advantage of the layer of gold nano-structures 35 is that this helps to reduce the impedance between the wire electrode and the tissue at the front end of the wire, and also that the gold nano-structures 35 provide a good platform for a range of different functionalisation layers. Figure 3 shows images illustrating the various layers deposited at the tip of the wire electrode at the front end. The left hand image of Figure 3 shows the bare polished metal core surrounded by insulator, prior to depositing either of the layers onto the tip of the wire. The middle image shows the electrode after depositing the gold nano-structure layer. The layer of gold nanostructures has a flaky consistency, providing a large surface area for charge transfer which helps to reduce the impedance at the tips of the wire. The right hand image of Figure 3 shows an image of the wire electrode after depositing the iridium oxide functionalisation layer on top of the gold nano-structure layer. The iridium oxide layer has a spongey consistency and provides a surface modification suitable for a range of biosensing or electrochemical applications. For example, the IrOx layer facilitates pH sensing. Also, the chemical properties of iridium oxide provide increased charge storage capacity which enables current injection and amperometric analyte detection (detection of ions in a solution based on electrical current), e.g. for detecting dopamine. This can be useful for neuronal sensing and stimulation for example (stimulation refers to the stimulation of specific parts of the brain with electrical impulses delivered by a neural probe, which can be useful for treatment of neurological diseases for example). Other examples may not have a functionalisation layer at all.

Also, in some examples the front end of the wire electrode 30 may not have any nano-structures 35, and instead the front end of the wire electrode 30 could simply comprise the end of the metallic core 32 and insulating sheath 34, either bare cut or polished. For example, in applications where a low ohmic connection between the wire electrode 30 and the tissue sample is not needed (e.g. if there is no electric charge transfer dependent function), then the nano-structures 35 could be omitted.

The nanostructures in the respective layers at the front end of the probe may have a unit width less than or equal to 500 nm. More particularly, the unit width may be less than or equal to 400 nm, less than or equal to 300 nm, less than or equal to 200 nm, less than or equal to 100 nm or less than or equal to 50 nm. The term“unit width” refers to the width across the longest dimension of an individual nano-structure (e.g. an individual flake, grain or nanoparticle), not the width of the mass of nano-structures as a whole. In some cases the nano-structures at the front end may have a unit width which is less than or equal to 20% of the wire diameter of the electrode, less than or equal to 15% percent of the wire diameter, less than or equal to 10% of the wire diameter, or less than or equal to 5% of the wire diameter. Note that the different nano-structures within the layer will in practice have different unit widths to each other, but all the nano-structures may have a unit width defined within the thresholds described above. Similarly, the functionalization layer may also comprise a layer of nano-structures (e.g. of Iridium oxide, or another material), which may have unit widths as defined within the thresholds described above. The nano-structures in the functionalization layer may be of a different size to the nano-structures in the impedance reducing layer or connection layer.

For the impedance reducing layer 35 of nano-structures, the nano-structures can be deposited on the tip of the wire electrode by electrodeposition. In one particular example, gold micro-hemispheres were deposited from a two-part aqueous cyanide bath containing 50 gL ¹ potassium dicyanoaureate(l) (K₂[Au(CN)₂]) and 500 gl_ ¹ KH₂P0₄ dissolved sequentially in deionized water (18 MOhm) (Tech, UK) at 60 °C. All reagents were supplied by Sigma- Aldrich, UK, and were used without further purification. Prior to electrodeposition the polished substrate was washed with ethanol (90%), rinsed with deionized water, wiped with a lint-free cloth (Kimwipes, Kimtech, UK) and dried at 50 °C for 1 hour in an autoclave. The electrodeposition protocol was carried out with a VSP 300 potentiostat-galvanostat (Bio- Logic, France) controlled with EC-Lab (Bio-Logic, France) in a three-electrode cell setup composed of a gold UME bundle as working electrode (W_E), a coiled platinum wire (99.99%, GoodFellow, US) as counter electrode (C_E) and a Ag/AgCI|KCI/_3.5M reference electrode (REF) supplied by BASi, USA (E vs. NHE = 0.205V). The REF was kept separated from the bath by a glass tube containing the support electrolyte and a porous Vycor glass separator. During gold deposition the W_E potential was kept at E_red =-1 .1 V vs. REF for a time determined according to the desired size of the gold hemisphere to be formed. During electrodeposition the bath was thermostated at 60 °C under vigorous (500 rpm) stirring. This technique has been successful for many different types of metal conductor material, including gold, platinum, tin, copper, brass, bronze, silver and lead. It will be appreciated that this is just one example of a possible electrodeposition protocol for making the wire electrode.

At the front end, although not shown in Figure 2, the tip of the wire electrode may be sharpened, or have a tapered end which meets at a point, to facilitate insertion into the sample. The length of the wire electrodes may be greater than or equal to 1 cm. More particularly, the length of the wire electrodes may be greater than or equal to 1 .5 cm; or greater than or equal to 2 cm; or greater than or equal to 2.5 cm; or greater than or equal to 3 cm; or greater than or equal to 3.5 cm; or greater than or equal to 4 cm; or greater than or equal to 4.5 cm; or greater than or equal to 5 cm. A probe with electrodes of length 3-5 cm can be particularly useful. In the examples above, the nano-structures are made of gold, but other examples could provide nano-structures of a different metal or metal oxide at the front end of the electrodes. The provision of a layer of metal or metal oxide nano-structures on the tips of the electrodes enables a reduction in impedance at the contact with the sample so that signal-to-noise ratio can be increased and hence it is practical to provide longer probes, including probes longer than 5 cm, to enable sensing or stimulation at deeper locations within the sample.

For glass ensheathed ultramicroelectrodes (UMEs) to be used in any electrophysiological application which involves reception and transmission of electrical signal through any length, the following characteristics are advantageous:

• a controllable frequency response input representing one side usually the one in contact with the biological/and or liquid sample,

• a well-insulated and electrical conductive length/body and a low-ohmic connection on the other side, usually the connection side or back-end.

• UME connection to any microscopic and/or macroscopic conductor by mechanical means either reversibly or non-reversibly preferably features a repeatedly deformable positively protruding mass from surface.

The brain produces in 30 seconds as much electrical data as the Hubble telescope will produce in its lifetime, with the vast amount of data resulting from chemical, biochemical and electrochemical events at cellular, tissue and system levels. Understanding how the sum of these interactions result in behaviour is a major topic of interest, but current technical limitations regarding probe size, geometry, recording capability, channel number and versatility towards other types of information keep the advancement of our understanding of how the brain works at a slow pace. Glass ensheathed UMEs represent an ideal platform for brain activity mapping from both extracellular and intracellular space because of their small size, massive scalability and ability to be interfaced with emerging high-channel count read-out technologies all considered solutions to current tech bottlenecks in experimental neurosciences. UMEs feature small stray capacitances (e.g. less than 0.5 pFcm ²) given the high insulator-conductor ratio, mechanical workability, broad material choice and commercial availability. UMEs usually have one dimension in the micrometre or nanometre domain and at least one the millimetre or centimetre region, thus the properties of the electrified interfaces are to be carefully considered when high frequency electrical signal need to be passed by micron-sized or nano-sized interfaces.

Current challenges in neural sensor miniaturisation are signal coupling and transport from the electrogenic cells trough the interfaces, on different conductor lengths towards the resistive junction to finally be delivered and processed by the read out circuitry. The smaller the sensors are, the higher impedances (Z) in aqueous electrolytes become, resulting in significantly weakened signals and high noise levels. The interface’s electrical coupling properties consequently bring limitations in the design of the read-out systems. Firstly, more amplifier stages and higher amplifier gains are required to condition recorded signals. Secondly, pre-filter and impedance matching circuitry are included to reduce ambient noise and pick up small-signals. Thirdly, power consumption of these additional amplifier stages could easily be a critical issue when limited power sources are available i.e. for battery powered chronic neural implants - thus improving signal strength while keeping electrode dimensions in the micron and sub-micron domain is of paramount importance.

These issues can be addressed using the wire electrode 30 discussed above as an electrochemical probe. By performing a surface modification of the tips of the UMEs at the front end, to include a highly fractalized flake-like gold nano-structure layer, and optionally a second layer of highly porous metal oxide (e.g. iridium oxide), the impedance at the front end of the electrodes can be greatly reduced. This is shown in the graph of Figure 4 which compares the impedance across different frequencies for three probes:

• “polished Au” - a probe made of bare gold metal wires without surface modification of the tips at the front end

• “IrOx modified” - a probe where the tips of the wires at the front end have the IrOx layer but not the intervening gold nano-structure layer

• “jULIE” - a probe as in the example of Figures 2 and 3 with both the gold nano structure layer and the IrOx layer at the front end tips of the wires.

As shown in Figure 4, the impedance at the front end of the jULIE wires is an order of magnitude lower than the other types of wires. To test jULIEs we performed recordings in the olfactory bulb (OB) of anaesthetized mice (4-6 weeks old, Ketamine/Xylazine anaesthesia) using a Tucker Davis RZ2 amplifier with a PZ2 pre-amplifier and RA16AC-Z headstage. Extracellular spikes were reliably recorded with amplitudes of up to 1.6 mV. Consistent with this, when jULIEs were lowered several mm into the brain and returned to a superficial recording position, extracellular units were reliably recorded throughout the olfactory bulb. Due to the small size of the recording site and minimal damage to the tissue, jULIEs were found to be exceptionally suited for recording large amplitude (500-1500 mn), well isolated signals from the close vicinity of neurons (20-30 mhi). It was found that the amplitudes recorded using the jULIE probe are much larger than the amplitudes obtained by a commercial probe. Hence, signal to noise ratio can be improved and there is less need for additional amplification, helping to reduce power in battery-powered implants for example.

Other advantages of the probes include:

(i) the dimensions of the penetrating wires are 2x to 5x times smaller and recording sites can be up to 50 times smaller (e.g. 1 pm) than in conventional probes. Also, the use of the Taylor-Ulitovsky method as discussed below results in wires with smoother sides than in conventional probes. This results in reduced tissue displacement and damage as well as in highly localized recordings with better unit separation (better identification of signals from individual neurons).

(ii) the nanostructured interface represents an excellent platform for further improved electrical coupling characteristics with the extracellular media, for example the nanosized gold/lrOx interface allows for substantially higher signal-to-noise with amplitudes of up to 1 .5 mV compared to typically 200-500 pV with conventional electrodes.

(iii) the material choice enables semi-automatic preparation for recording sites pre arrangement to fit anatomical structures; and needle-like sharpening for seamless penetration of the neural tissue.

(iv) there are also substantially improved charge transfer capabilities i.e. enabling current injection for stimulation purposes and neurotransmitters or other analyte monitoring (e.g. alcohol, paracetamol), in a highly localized manner.

The wire electrode 30 can be manufactured using the Taylor-Ulitovsky method. The Taylor-Ulitovsky method is a technique for forming glass-sheathed wire electrodes with a very fine diameter, e.g. as small as a few microns. In the Taylor-Ulitovsky process, the metal or alloy conducting material is placed inside a glass tube which is closed at one end and the other end of the tube is heated to soften the glass to a temperature at which the conductor melts. The glass can then be drawn down to produce a fine glass capillary with the metal core inside the glass. Hence, metal cores of diameters in the range 1 to 120 microns can be coated with a glass sheath a few microns thick with this method. In particular, wires with a core in the range 1 -10 pm surrounded in 10-40 pm of glass can be useful for electrical and electrochemical sensing. The metal used can include copper, silver, gold, iron, platinum, lead or other metals as well as crystalline or amorphous alloy compositions such as brass, bronze, platinum-iridium, or magnetic alloys. The wire electrode 30 may be flexible, so that it can bend, which can help with insertion of the electrodes into biological tissue.

Hence, a wire electrode 30 of the type described above can be very useful for a range of applications, for example sensing, recording or mapping neural activity (e.g. as a neural probe), brain mapping, tumour scanning, cardiovascular scanning, monitoring spinal cord lesions, scanning electrochemical microscopy, and/or sensing, recording or mapping capacitance or resistance measuring (e.g. in label-free affinity impedimetric biosensing). However, current bonding techniques for bonding the electrode 30 to an integrated circuit 4 providing the read out / stimulation electronics tend to provide a high resistance at the join between the electrode 30 and the integrated circuit 4.

Figures 5A to 5F show an example using the capillary 6 of the wire bonding machine 2 to perform bonding of the wire electrode 30 of the type shown in Figure 2 to a metallic contact surface 20 of an integrated circuit 4. This method helps to reduce the bond resistance between the wire electrode 30 and the integrated circuit 4, which enables the probes using the electrode to provide greater sensitivity and/or use smaller stimulation voltage/currents to save power and reduce heat dissipation requirements.

Although in the examples shown in Figures 5A to 5F, the metallic contact surface 20 is a pad provided on the substrate 22 of the integrated circuit 4, in other examples, the wire electrode could instead be bonded to the substrate 22 itself or to other metallic regions of the integrated circuit. While Figures 5A to 5F show the wire electrode being bonded at one location on the integrated circuit 4, in some cases multiple bonds could be formed on the same electrode 30. However, in some examples it can be useful to bond the wire electrode to the integrated circuit near the back end of the electrode, and leave the front end of the wire electrode loose for insertion into a sample.

As shown in Figure 5A, initially the wire electrode 30 is positioned over the metallic contact surface 20 of the integrated circuit 4. The upper part of Figure 5A shows a view from above and the lower part of Figure 5B shows a view in cross-section. The wire electrode may, in the region corresponding to the metallic contact surface 20, run substantially parallel to the plane of the substrate 22, so that the insulating sheath 34 touches the metallic contact surface 20 on one side of the electrode 30. Although in the vicinity of the metallic contact surface, the electrode 30 may run parallel to the substrate 22, in other parts of the wire electrode this is not essential. For example in some examples the wire electrode may be flexible and so the part of the electrode further from the metallic contact surface 20 may not be parallel to the substrate. In some examples the wire electrode 30 may simply be laid in position, or may be held in place using an external implement. However, in the example shown in Figure 5A a bond 36 is formed over the wire electrode 30 to hold the electrode 30 in position. The bond may be at a position separate from the location at which the wire electrode is actually bonded to the metallic contact surface 20. For example the bond 36 may be ball-wedge bond which passes over the electrode 30 to hold it against the integrated circuit 4. The ball wedge bond could be formed by any conventional wire bonding technique. For example, the ball bonder machine 2 itself could be used to supply the bonding wire used to form the ball wedge bond 36 and bond each end of the bonding wire to the integrated circuit either side of the wire electrode 30. As shown in Figure 5B, the capillary 6 of the ball bonding machine 2 is brought down and used to apply mechanical force to the electrode 30, to break the insulating sheath 34 of the wire electrode 30 in the region corresponding to the metallic contact surface 20. For example the capillary 6 may be lowered in increments of a certain distance, for example in 10 mhi increments, until the wire is broken. As shown in Figure 5B, the bonding wire 8, which is made of the same metallic material as the core 32 of the electrode 30, may be carried within the capillary 6 at the time when the capillary 6 is forced against the wire to break the sheath 34. The bonding wire 8 may be retracted inside the capillary 6 when the capillary is used for mechanical impact.

The impact of the capillary 6 against the sheath 34 of the electrode 30 causes breaking of the sheath on both the upper and lower sides of the metallic core 32. This results in loose pieces of broken insulator 38 detaching from the electrode. If these loose pieces of insulator are allowed to remain then they could compromise the integrity of the bond as the remnants of the insulator may increase resistance.

Flence, as shown in Figure 5C a step of removing the loose pieces of broken insulator 38 may be performed. This may include mechanically displacing the capillary 8, either in a side to side direction in the plane of the substrate or in the up and down direction towards and away from the substrate, to mechanically move pieces of insulator away from the bonding region corresponding to the location of the metallic contact surface 20 on the integrated circuit. For example by imaging the bonding region using a scanning electrode microscope the locations of the pieces 38 of insulator can be identified and the corresponding displacements required to remove the insulator material can be selected. Alternatively, or in addition to the mechanical displacement of the capillary 6, loose insulating material 38 can be removed by supplying at least one removal agent into the bonding region for removal of the loose pieces of insulator 38. For example a jet of compressed air can be directed into the bonding region to blow away pieces of loose insulator. Also, it is possible to use a solvent in the liquid phase which dissolves the insulating material to remove some of the loose pieces of insulator. It will be appreciated that any combination of these techniques (mechanical displacement, compressed air or solvent based methods) can be used. In cases where a solvent is used, the particular solvent chosen may depend on the insulator material used for the insulating sheath 34 of the wire electrode 30. For example, for silica glasses, the solvent could be one or more of: buffered oxide etch (BOE); HF+HCI; FIF+FINO₃; KOFI; NaOFI; NFI₄OFI (other strong alkaline); FI₃PO₄; or K₂FIPO₄. For fluoride glasses, the solvent could be one of HF and BOE. For phosphate glasses, the solvent could be one of HF; BOE; HCI. For fluorberyllate, tellurite, or germanate glasses, the solvent could be dilute HF. As shown in Figure 5D, the bonding wire 8 is fed through the capillary 6 to expose the tip of the bonding wire 8 outside the capillary, and the EFO wand 18 is then used to supply an electric change to the exposed tip of the bonding wire 8 carried by the capillary 6, which causes the tip to melt. The molten tip forms a ball 40 of molten metallic material at the end of the bonding wire 8, due to the surface tension in the molten material.

As shown in Figure 5E the capillary 6 is then lowered and some bonding pressure applied to the electrode 30 (e.g. bond forces of around 25 to 50 cN (centinewton) may be applied. Also ultrasonic vibration may be applied using the piezo electric disk 14 and horn 16, to provide a welding of the ball bond in the bonding region 20. As shown in Figure 5E the molten material 40 spreads out into the gap formed by the breaking of the insulating sheath 34 and contacts the exposed metallic core 32 on both the upper and lower sides of the metallic core 32. As shown in Figure 5F the resulting ball bond 42 extends outside the diameter of the wire electrode 30 so that there is a bond of metallic material which contacts both the metallic contact surface 20 of the integrated circuit 4 and the exposed metallic core 32 of the electrode 30. As the bonding wire 8 carried by the capillary 6 is of the same metallic material as the core 32 of the electrode 30, there is no amalgam of different metals and so a clean bond with low resistance can be formed with this technique, with reduced insulating material in the bonding region.

While the Figures discussed above show an example of using a capillary 6 of a ball bonder machine 2 to perform the bonding, it would also be possible to perform a similar method using a capillary 6 in a different type of machine which is not intended for ball bonding. Flowever, it can be useful to reuse the ball bonding equipment 2 provided for ball bonding of wired connections in integrated circuits for the purpose of bonding wire electrodes of the type discussed above. This can reduce the cost of implementing the bonding method,

Figure 6 is a flow diagram illustrating a method of bonding a wire electrode 30 having a metallic core and insulating sheath to a metallic contact surface 20 of an integrated circuit 4. At step 50 the wire electrode is positioned over the metallic contact surface. Optionally the wire electrode may be bonded to the integrated circuit to fix the relative position of the wire electrode relative to the metallic contact surface 20 (this bond may not provide any electrical connection between the wire electrode 30 and the integrated circuit 4, but merely holds the electrode 30 in position).

At step 42 the bonding wire 8 is provided inside the capillary 6, for example a capillary of a ball bonding machine 2. At step 54 mechanical force is applied to the insulating sheath 34 of the electrode 30 to break the sheath 34 and expose the metallic core 32 of the wire electrode 30. Although step 54 could be performed separately from the bonding equipment 2 (e.g. by reordering step 54 to be performed before step 50 by a separate device to break through the insulating sheath 34), it can be more convenient to apply the mechanical force once the wire electrode is already in position as this may enable better control of the location at which the insulating sheath is broken, and avoids any need to subsequently align the part of the wire electrode 30 at which the sheath was broken with the metallic contact 20. A separate mechanical element could be used at step 54 to apply force to the wire to break the sheath, separate from the capillary 6 used to carry the bonding wire. However, if at step 54 the capillary 6 itself is used to apply the force, this can reduce the number of hardware elements required, hence reducing the cost of the apparatus, and also has the benefit that existing position control mechanisms 12 provided to control the location of the capillary 6 relative to the circuit 4 can be reused for also controlling the location at which the insulating sheath is broken.

The breaking of the insulating sheath may result in shards of glass or other insulator being left in the bonding region and so at step 56 a step of removing the loose pieces of broken insulating material is performed. This may involve mechanically displacing loose pieces of broken insulator 38 using the capillary 6 or another mechanical element, and/or supply of a removal agent, such as compressed air and/or a solvent provided as a chemical in the liquid phase, such as hydrofluoric acid or sodium hydroxide. The particular solvent to use may depend on which material is used for the insulating sheath 34. In general any solvent which may dissolve the insulator but not the metal could be used.

At step 58 at least one of electrical charge, heat and ultrasonic vibration (that is, either electrical charge alone, heat alone or ultrasonic vibration alone, or any two or more of these) is applied to the tip of the bonding wire 8 which is carried in the capillary 6. The capillary 6 is pressed towards the wire electrode 30. The electrical charge, heat and/or vibration causes the tip of the bonding wire 8 to melt, to form a ball of molten metallic material which is pressed towards the wire so as to form a ball bond 42 connecting the exposed core of the wire electrode to the metallic contact surface 20.

In summary, a classic ball-bonding machine can be reused for bonding glass insulated Taylor-Ulitovsky microwires to metallic substrates/pads. With classic bonding methods, the glass ensheathing the microwire prevents the formation of useful connection. Hence, the bonding-tip of the capillary of the bonding machine can be used in a controlled fashion to crush the glass. We have found that by adding a puff of compressed air locally and/or chemical solvent this improves the separation/cleanliness of the bonding regions thus increasing success rates. This can be particularly useful for with wires using core metals which are significantly smaller, where there is significantly more insulation to eliminate in order to expose the core.

Figures 7A to 7C show images taken during the bonding method in one example. In this example, glass stripped microwires with 9pm diameter gold core were bonded to a gold pad using 25-50 cN bond forces, ultrasonic power at 45 kHz with a duration of 25 ms. However, these parameters will vary with substrate and wire type. Figure 7A shows the microwire secured to the pad using a ball-wedge bond over the fibre (similar to Figure 5A). Figure 7B shows how after breaking the microwire by mechanical impact with the capillary 6, displacement of the tip of the capillary 6 can be used to remove the broken insulation. Note that Figure 7B was taken during a different bonding operation to the operation shown in Figure 7A and 7C as the loose pieces of insulating are more visible in Figure 7B. Figure 7C shows the subsequent bonding step of heating the tip of the bonding wire 8 using the arc heating and pressing the ball of molten material into the microwire 30 where the previous removal of glass insulating sheath means that there is a larger surface area of exposed metallic core to bond to.

Figure 8 shows an SEM (scanning electron microscopy) micrograph of a bonded gold Taylor-Ulitovsky (TU) wire 30 to a gold substrate 20 using the technique discussed above. Region A shows the gold bump that connects the exposed TU core 32 to the substrate 20. Region B shows the TU wire 30 with glass insulation intact in the region further from the bond. Region C shows the ball-wedge bond formed at step 50 to loop over the TU wire 30 to hold the wire in place during the formation of the bond in region A.

Figure 9 shows a magnification of region A shown in Figure 8, which shows the connection bump in more detail. Before the bump is applied, the TU wire insulation is broken down mechanically using the capillary 6 at which later the bump is formed. Region 1 of Figure 9 shows the gold bump pressing the exposed TU wire core against the substrate. Region 2 shows the exposed TU core 32 continuing underneath the gold bump. Region 3 shows a portion of the TU wire with the insulating sheath still intact. Region 4 shows loose pieces of broken glass insulation left around the bond. The insulator removal steps help to reduce the amount of glass insulation which is left inside the bond itself.

Figure 10 is a graph showing the resistance measured for bonds formed using the method discussed above, for a number of different bonding attempts on wires of different metallic core materials and different diameters. Although there is some variation in resistance, depending on the extent to which the insulating material is removed from the bonding region, in all cases the resistance of the bond between the wire electrode 30 and integrated circuit 4 is in the order of 1 s to 10s to 100s of Ohms. In contrast, using conventional bonding techniques to solder the end of the UME electrodes to the integrated circuit, e.g. using a solder material such as an alloy of gallium, indium and tin, the resistance at the junction would be of the order of kiloOhms. Hence, this technique can provide a significant improvement in the resistance at the bond.

In some examples the ball bonder machine 2 shown in Figure 1 which is used to perform the method may be an existing ball bonding machine provided for a conventional ball bonding, with no further modification performed. Hence the method can be performed by using an existing machine in a different way, but without any hardware modifications.

Alternatively, a bespoke designed machine may be provided, with some hardware modifications aimed at improving the bonding of the UMEs discussed above. Figures 1 1 to 18 show some potential hardware modifications of the capillary 6 which can help with the removal of the insulating sheath 34 and the supply of removal agent.

As shown in Figure 1 1 , the tip of the capillary 6 may have a sharpened edge 70. The sharpened edge 70 may extend along only part of the circumference of the tip of the capillary 6 as shown in Figure 1 1 , or could extend along the entire circumference as shown in Figure 12. As shown in Figure 1 1 , the tip of the capillary has an outer diameter 72 and an inner diameter 74, where the inner diameter 74 corresponds to the width of the internal bore for carrying the bonding wire 8 and the outer diameter 72 corresponds to the overall width of the tip of the capillary 6 including the surrounding material outside the bore. In the examples of Figures 1 1 and 12, the sharpened edge 70 extends around the outer diameter of the capillary 6. However, in other examples the sharpened edge could be at the inner diameter 74.

As shown in Figure 13, another modification of the capillary design may be to attach to the main wire-carrying capillary 6 a second capillary 80 which acts as a removal agent supply capillary for supplying the removal agent for removing loose insulator material following its breakage. For example, the removal supply capillary 80 may be used to supply a jet of compressed air or a chemical solvent. The removal supply capillary 80 could be integrally formed with the main capillary 6, as a single piece of integrated material, or could be formed separately and then attached to the capillary 6 by an adhesive or mechanical fixing. In the example of Figure 13 the removal agent supply capillary 80 is provided but there is no sharpened edge 70 as in Figures 1 1 and 12. However as shown in Figures 14 and 15 it is also possible to provide the removal supply capillary 80 in embodiments which include a sharpened edge 70 at the tip of the wire bonding capillary 6, either around part of the circumference of the wire bonding capillary 6 as shown in Figure 14 or around the full circumference as shown in Figure 15.

Also, as shown in Figures 16 to 18, it is possible to provide the wire-carrying capillary 6 with two separate removal agent supply channels 80, 82, one for compressed air and the other for a chemical solvent, so that both types of removal agent may be supplied. Again, three versions are shown in Figures 16 to 18, with Figure 16 showing the case where the wire carrying capillary 6 has a flat tip and Figures 17 and 18 showing cases with a partial or fully sharpened edge 70 similar to the examples of Figures 1 1 and 12 respectively.

Hence, with the examples of Figures 13 to 18 the channels for carrying the removal agent are attached to or integrally formed with the wire carrying capillary 6 so that any position adjustment to the wire carrying capillary also moves into position the corresponding channels for supplying any removal agent, which can reduce the hardware scale of the device.

Although illustrative embodiments of the invention have been described in detail herein with reference to the accompanying drawings, it is to be understood that the invention is not limited to those precise embodiments, and that various changes and modifications can be effected therein by one skilled in the art without departing from the scope and spirit of the invention as defined by the appended claims.

Claims

1. A method of forming an electrical connection between a wire electrode and a metallic contact surface of an integrated circuit, the wire electrode comprising a metallic core surrounded by a sheath of insulating material; the method comprising:

positioning the wire electrode over the metallic contact surface;

applying mechanical force to the wire electrode to break the sheath of insulating material and expose part of the metallic core of the wire electrode;

providing, within a capillary, a bonding wire made of the same metallic material as the metallic core of the wire electrode; and

applying at least one of electrical charge, heat and ultrasonic vibration to a tip of the bonding wire and pressing the capillary carrying the bonding wire towards the wire electrode, to melt the tip of the bonding wire to form a bond of metallic material which connects the exposed core of the wire electrode to the metallic contact surface of the integrated circuit.

2. The method of claim 1 , in which the capillary comprises a capillary of a ball bonder machine.

3. The method of any of claims 1 and 2, in which the mechanical force is applied to the wire electrode using the capillary.

4. The method of claim 3, in which the bonding wire is provided inside the capillary when the mechanical force is applied to the wire electrode.

5. The method of any of claims 3 and 4, in which the capillary has a tip with a sharpened edge.

6. The method of claim 5, in which the sharpened edge extends around all or part of a circumference of the tip of the capillary.

7. The method of any preceding claim, comprising a step of removing loose pieces of broken insulating material prior to forming the bond.

8. The method of claim 7, in which the step of removing comprises mechanically moving the loose pieces of broken insulating material using the capillary.

9. The method of any of claims 7 and 8, in which the step of removing comprises supplying at least one removal agent for removing the loose pieces of broken insulating material.

10. The method of claim 9, in which the at least one removal agent comprises compressed air.

1 1 . The method of any of claims 9 and 10, in which the at least one removal agent comprises a solvent for dissolving the insulating material.

12. The method of any of claims 9 to 1 1 , in which the at least one removal agent is supplied through a removal agent supply capillary attached to, or integrally formed with, the capillary used to carry the bonding wire.

13. The method of any preceding claim, in which the insulating material comprises glass.

14. The method of any preceding claim, in which the wire electrode has a diameter less than or equal to 25 mhi.

15. A bonding machine comprising:

a wire-carrying capillary for carrying a bonding wire;

at least one bonding control element to provide at least one of electrical charge, heat and ultrasonic vibration to a tip of the bonding wire carried within the capillary; in which: the capillary has a tip with a sharpened edge.

16. The bonding machine of claim 15, in which at least one removal agent supply capillary is attached to, or integrally formed with, the wire-carrying capillary.