EP0237154A1

EP0237154A1 - Insulated aluminium wire

Info

Publication number: EP0237154A1
Application number: EP87300595A
Authority: EP
Inventors: Alexander Joseph Otto; Harry Sang
Original assignee: Alcan International Ltd Canada
Current assignee: Rio Tinto Alcan International Ltd
Priority date: 1986-02-06
Filing date: 1987-01-23
Publication date: 1987-09-16
Also published as: JPS62217507A; GB8602927D0; CN87100618A

Abstract

Insulated aluminium wire has many applications in the microelectronics field. The invention provides insulated aluminium wire in which the insulation comprises either a single layer of barrier-layer (non-porous) anodic aluminium oxide, or two layers, an inner barrier-layer anodic oxide film and an outer layer of hydrated alumina. The insulated wire may be made by continuously anodizing wire of Al or an Al/Mg alloy in a barrier-layer electrolyte such as tartaric acid, and then exposing the anodized wire to hot water to partly hydrate the anodic oxide film.

Description

This invention concerns aluminium wire, electrically insulated by means of an anodic oxide film. Such wire has many different uses in the microelectronics field. For example, very fine aluminium and gold based wires are commonly used to connect integrated circuit terminals to lead frames. Connection is achieved by either wedge-wedge bonding or ball-wedge bonding. Neither the gold nor the aluminium wires currently used in bonding are insulated. However, there is an expressed need for insulated bonding wire, especially with the large number of chip to lead frame connections required for VLSI circuits.
This need exists because bare bonding wires must be connected from chip to lead frame in a way that prevents wire to wire contact and wire to circuit contact. This requirement severely restricts the numbers of leads possible from a chip, because the wires must by spaced far apart and carefully looped to isolate them from their environment. A suitably insulated bonding wire would effectively relax the strict geometric requirements of looping and spacing, because wire to wire and wire to circuit contact would not lead to circuit malfunction.
Insulated wire is not now used in bonding, for two reasons. First, insulating materials on wire surfaces are known to severely compromise wire bondability in the same way that dirt on the wire surfaces compromises bondability. Second, insulating very fine wires acceptably with conventional wire insulating material such as plastics has proven to be very difficult.
DE 3335848 A describes a method of connecting an insulated aluminium wire to a contact of an electronic circuit, the method involving the use of ultrasonic vibrations. The insulated aluminium wire is briefly described as being an anodized aluminium wire, comprising an aluminium core with a diameter of 30 to 60 microns and an insulating alumite layer having a thickness of 0.1 to 1 microns. The layer of "alumite" is understood to refer to a sealed porous anodic oxide film. In our hands, the insulation provided by such a layer has generally proved inadequate for microelectronics applications.
In order to explain the development which constitutes this invention, it is necessary first to discuss briefly the nature of the anodizing process. When an aluminium substrate is placed in an electrolyte and made anodic, an oxide film is built up on the surface, and increases in thickness by conversion of metal to oxide at the metal/oxide interface. Two different kinds of anodic oxide film may be formed depending on the circumstances:-

i) When the electrolyte is one which has a significant solvent effect on alumina, a porous anodic oxide film is formed. This consists of a thin barrier layer adjacent the metal/oxide interface, and a much thicker porous layer in which pores extend from the oxide/electrolyte interface to the barrier layer. Continued oxide film growth involves transport of species up and down the pores, and final film thickness is primarily dependent on anodizing time.
ii) When an electrolyte is used which has no significant dissolving effect on alumina, a non-porous or barrier-layer anodic oxide film is formed. The thickness of this film is primarily dependent upon anodizing voltage, and is typically from 1.0 to 1.4nm per volt.

Nearly all commercial anodizing is carried out using chromic acid, sulphuric acid or oxalic acid under conditions to give rise to porous anodic oxide films. For most purposes, the pores in these films need to be sealed, and this is commonly done by exposing the films to boiling water which hydrates the surface alumina and causes it to swell so as to block the pores at their outer ends. Alumite is a term of art applied to such sealed porous anodic oxide films.
Barrier layer films are widely used as insulators in electrolytic capacitors. This invention is based on the discovery that non-porous or barrier-layer anodic oxide films can be formed on aluminium wires to provide not only excellent electrical insulation, but also other advantageous properties. In one aspect, the invention provides insulated aluminium wire wherein the insulation comprises a barrier-layer anodic oxide film. Under favorable circumstances, such insulated aluminium wire can satisfy the needs of the microelectronics industry. The insulating performance can, however, be further improved by converting part of the barrier layer to hydrated alumina. Thus in another aspect, the invention provides insulated aluminium wire wherein the insulation comprises two layers, an inner barrier-layer anodic oxide film and an outer layer of a hydrated alumina. This product shows improved insulating performance in a dry atmosphere, and performance in a humid atmosphere is less adversely affected.
The aluminium wire generally has a diameter in the range from 10 to 1000 microns. Wires currently used in the microelectronics industry typically have diameters of 25 to 30 or 100 to 400 microns. Wires below 10 microns diameter are not easy to make. Above 1000 microns diameter, other methods of insulation may prove more attractive.
The wire may be of pure aluminium or of an aluminium alloy. Pure aluminium may be used where the strength of the wire is not of paramount importance, and has the advantage that an even unblemished anodic oxide film can be formed thereon. Al-1%-Si wire is commonly used in the microelectronics industry, and can be used to form the insulated wire of this invention. However, it contains silcon phase particles, which may form defects on the wire surface during anodizing. In many cases, such defects will not matter. When they do, it may be preferable to use a single-phase aluminium alloy, such as for example Al-1%-Mg, which is stronger than pure aluminium and is the preferred alloy for use in this invention. An added advantage of the Al-Mg alloys is that they form anodic oxide films which are rather brittle, which may assist bonding as described below.
To achieve significant electrical insulation, the barrier-layer anodic oxide film should be at least 0.01 microns thick. It is difficult to make barrier-layer films more than about 0.3 microns thick, because of dielectric breakdown at the high anodizing voltages required. Preferred barrier layer film thicknesses are generally in the range 0.1 to 0.25 microns.
As noted above, when a barrier-layer anodic oxide film is exposed to hot or boiling water, the outer surface becomes hydrated and undergoes considerable swelling. The chemical changes are complex, but may be approximately summarised as:-
Gamma-Al₂-O₃ + H₂O → Al₂O₃. H₂O.
The resulting film has a two-layer structure, comprising an inner barrier-layer anodic oxide film, which should be preferably from 0.01 to 0.15 microns thick although there is no critical lower limit on thickness; and an outer layer of a hydrated alumina which will generally be from 0.01 to 0.8 microns, more usually from 0.1 to 0.5 microns, thick. Very thin hydrated alumina layers do not significantly improve the insulating properties of the barrier-layer film. It is difficult to form a hydrated alumina layer more than 0.8 microns thick without completely hydrating the inner barrier layer with the associated danger of spalling.
The insulated wire may be formed by anodizing aluminium wire under suitable conditions. The electrolyte is one which, under the anodizing conditions chosen does not have a significant dissolving effect on alumina. We have found it convenient to use to a weak (up to 5% by weight) tartaric acid solution buffered with ammonium hydroxide to a pH of from 5 to 7 but other acids, such as oxalic acid, citric acid and boric acid, may be used as is well known, and for these other buffers, other concentrations and other pH ranges may be appropriate. The technique of barrier-layer anodizing is described, for example, in S.Wernick and R.Pinner, "Surface Treatment of Aluminium".
The electrolyte is conveniently left at ambient temperature. The applied voltage should be high enough to ensure rapid film growth, without being so high as to cause dielectric breakdown of the film. The electrolysis time should be sufficient to enable the anodic film thickness to approach the theoretical maximum, and may typically be in the range of 15 to 60 seconds. Anodizing under these conditions is readily performed on a continuous basis, with the wire being passed over guides, one of which also acts as a current carrier, through a bath of the electrolyte.
After anodizing, the wire is washed and may then be subjected to hot water in order to hydrate the outer surface of the alumina barrier layer. The water temperature is generally at least 80°C, but the water preferably not boiling in order to avoid possible damage by bubbles to the anodic oxide layer. As is well known in the art of sealing porous anodic oxide films, additives may be present in the water to hasten hydration. The extent of hydration of the alumina depends among other things on the duration of its exposure to hot water. Exposure times of from 5 or preferably 10 to 120 seconds are likely to give satisfactory layers of hydrated alumina, but without completely destroying the inner barrier-layer anodic oxide film.
For example, wire having a primary oxide film thickness of 0.17 microns was immersed in boiling water for about 40 s to form a hydrated second oxide layer about 0.5 microns thick on top of the remaining 0.08 microns of primary oxide.
The invention is illustrated by the accompanying drawings, in which:-

Figure 1 is an idealized section, on an enlarged scale, through part of an insulated aluminium wire according to the invention; and
Figure 2 is a top view of a microcircuit whose terminals are joined by means of fine wires to lead frames.

Referring to Figure 1, a fine metal wire 10 is made of an Al-1%-Mg alloy, and carries insulation on its surface comprising two layers. An inner barrier-layer anodic oxide film 12 is covered by an outer layer 14 of a hydrated alumina, approximately four times as thick as the inner layer. During anodizing, conversion of metal to oxide took place at the metal/oxide interface 16. Hydration of the alumina took place starting from the oxide/air or oxide/liquid interface 18.
Figure 2 shows a chip 20 containing a microcircuit 22 surrounded by a series of small contacts 24, each of which is connected by means of a fine aluminium wire 26 to a bonding pad 28 on a lead frame 30. The wires are mostly parallel to one another, but two pairs are shown crossed at 32, and in contact with one another, for which purpose they need to be insulated to avoid malfunction.
Bonding of the wires 26 to their respective contacts 24, 28 are routinely performed using a wedge shaped head, through which the wire is threaded, which is pressed down on the contact and then vibrated to spread the metal of the wire and cause it to adhere to the contact. Adhesion cannot be reliably effected in the presence of insulation on the wire. There is therefore required a form of insulation which is sufficiently brittle to break off when required, so as to leave a bare metal end which can be spread by the wedge bonder and caused to adhere to the contact. The insulating films described herein are found to be sufficiently brittle for this purpose. Indeed, anodic oxide films formed on aluminium-magnesium alloys are known to be relatively brittle.
The electrical resistance requirements of the microelectronics industry can be described in relation to wires which overlap in an X pattern as shown at 32 in Figure 2, such that the top wire is in contact with, and forces down upon, the bottom wire with enough force to bow the bottom wire. The two wires which make contact in this way should not conduct more than 10⁻¹²A at 20V, and the insulation breakdown voltage should be above a minimum figure such as 20 V but which depends upon the intended application.
The following example illustrates the invention.
As-drawn 35 micron diameter Al-1%-Mg wire was electrochemically cleaned and thinned (electropolished) in a perchloric solution until it was 30 microns diameter. This wire was anodized at 105V in 3% by weight tartaric acid/ammonium hydroxide solution with a pH of 6 at ambient temperature to form a barrier-layer anodic oxide film about 0.13 microns thick. During this continuous anodizing procedure the wire was charged by pulling it over a charged pulley prior to entry into the solution. The wire was pulled through the solution with a 40 second residence time.
Upon leaving the solution, the wire was rinsed in water. Then, still in series, the wire was pulled through water at 98°C with a residence time of 35 seconds. This washed and hydrated the barrier-layer anodic oxide film. One set of samples was not hydrated.
Samples of these wires were wedge-bonded onto thick film gold with several bonding machines. The wedge-wedge bond strengths were about 15g - well above the 12g lower limit required by the microelectronics industry for high quality devices. (Depending on application, lower wedge bond strength down to 8g or even to 2.5g may be acceptable. Leakage currents between contacting wires were tested at 40 volts for both hydrated and unhydrated barrier-layer anodic oxide films.
The wire to wire leakage current after 60 seconds was 7 x 10⁻¹³A in dry nitrogen for the hydrated sample. The wire to wire leakage current after 60 seconds was 2 x 10⁻¹²A in dry nitrogen for the unhydrated sample.

Claims

1. Insulated aluminium wire wherein the insulation comprises a barrier-layer anodic oxide film.

2. Insulated aluminium wire as claimed in claim 1, wherein the average thickness of the barrier-layer film is from 01 to 0.25 microns.

3. Insulated aluminium wire wherein the insulation comprises two layers, an inner barrier-layer anodic oxide film and an outer layer of a hydrated alumina.

4. Insulated aluminium wire as claimed in claim 3, wherein the inner barrier-layer is at least 0.01 microns thick and the outer layer of hydrate alumina is from 0.001 to 0.8 microns thick.

5. Insulated aluminium wire as claimed in any one of claims 1 to 4, wherein the wire is of pure aluminium or of a single-phase aluminium-rich alloy.

6. Insulated aluminium wire as claimed in claim 5, wherein the wire is of a single-phase Al/Mg alloy.

7. A microelectronic device comprising at least one integrated circuit and at least one lead frame, characterized in that these are connected to one another by means of insulated aluminium wire as claimed in any one of claims 1 to 6.

8. A method of making an insulated aluminium wire, which method comprises subjecting an aluminium wire to anodizing in an electrolyte which has no significant dissolving power for aluminium oxide so as to form a barrier-layer anodic oxide film thereon, and then exposing the anodized wire to hot water to partly hydrate the film and form an outer layer of a hydrated alumina thereon.

9. A method as claimed in claim 8, wherein the wire is continuously passed successively through a bath containing the anodizing electrolyte and then through a hot water bath.