SENSORS
This invention relates to sensors. In particular to gas sensors and to pH sensors. The invention also relates to methods of making sensors and electrical/electronic apparatus.
The invention is particularly, though not exclusively applicable to the measurement of metabolic gases (e.g.02 and C02) in blood. Monitoring metabolic gases in blood is particularly important in pre-term infants where inadequately developed lungs or breathing control mechanisms may lead to inadequate take up of oxygen. If the arterial partial pressure of oxygen (pa02) falls too low for even a short while brain damage or death may result: if however pa02 goes too high as a result of administered oxygen, damage to the eyes can occur.
The measurement of the arterial partial pressure of carbon dioxide (paC02) is also important as there is evidence that rupture of blood vessels in the brain may result as a consequence of fluctuations in cerebral blood flow caused partly by changes in paC02.
One technique that has been used in measurement of both pa02 and paC02 is known as the transcutaneous technique. In this method a sensor applied to the skin measures oxygen partial pressure of the skin (ps02) and/or carbon dioxide partial pressure at the skin (psC02). At normal temperatures ps02 is near zero. However if the skin temperature is raised ps02 can approach Pa02; and by choice of sensor and skin temperature (due to a fortuitous cancellation of errors) the measured ps02 can match pa02• In new born infants a sensor temperature of 43.5 ± 0.5°C has been found optimal for this purpose. co2 can be measured in a similar way, but in this case psC02 does not match paC02 and it has been found that if measured at 43.5 ± 0.5°C (the optimum temperature for Pa02 measurement) , psC02 = m.paC02 + c where m and c are constants that may have to be determined for a given sensor geometry.
The sensors currently used for Ps02 and PsC02 differ in mechanism but can be included in one apparatus. Such
apparatus has been reviewed by one of the inventors in Journal of Physics E: Scientific Instruments 2_0 (1987) pages 1103- 1112. A known combined 02/C02 sensor described in this review comprises:- a polarographic 02 sensor having a platinum wire cathode and Ag/AgCl anode, a glass pH electrode, an electrolyte, and a membrane.
All these electrodes (platinum wire cathode, Ag/AgCl anode, and pH electrode) contact the electrolyte which is retained by the membrane. The membrane may be placed and sealed against the skin so that 02 and C02 can diffuse through the membrane into the electrolyte. 02 is consumed by the polarographic 02 sensor which gives a current proportional to dissolved 02 in the electrolyte. C02 reacts with the electrolyte to produce carbonic acid which is measured using the pH electrode with the Ag/AgCl anode used as a reference electrode.
Such sensors have been extensively used but they suffer several disadvantages:- i) the glass pH electrode is sensitive to shock and easily damaged ii) the electrolyte has to be periodically replaced since the aqueous electrolytes used have a short shelf life due to loss of water iii) replacement of the electrolyte is a skilled job since it must be free of air bubbles to give consistent and accurate results. An 02 sensor is described in U.S.Patent Specification 4534356 and comprises a pair of electrodes deposited on a ceramic substrate, an insulating layer overlaying the electrodes, part of the electrodes remaining exposed and an oxygen permeable polymer electrolyte. This electrode has several important drawbacks:- i) Comparatively large cathode areas are exposed (90mm2, Example 1; 24.2mm2, Examples 2 and 3). The problem with using such large electrodes is that
they consume a great deal of oxygen and so it cannot be assumed that the bulk oxygen concentration in the electrolyte is identical with that overlaying the electrode, ii) For reproducability the overlaying insulating layer must maintain its integrity so that the electrode area remains constant. This is difficult to ensure using the geometry of U.S. Patent Specification 4534356 in particular since the leads to the electrodes are mounted on the same face of the substrate as the electrodes themselves, iii) A long stabilisation time (about hour) is required because of the large diffusion zone around the cathode, iv) Such an electrode cannot be used in a combined 02/C02 probe since the high current consequent on such large electrode areas will produce hydroxyl ions that will alter the pH of the electroyte. Also known as oxygen sensors are devices such as that described in GB 2055476-A. This sensor has a re-usable base assembly comprising a heater and means for supplying electrical current to a disposable assembly; the disposable assembly includes a gas permeable membrane and electrolyte is placed between this membrane and the body of the disposable assembly either by permeation of electrolyte through the membrane; or by permeation of water through the membrane forming an electrolyte with a coating of electrolyte salt disposed between the membrane and the body of the disposable part.
A further known sensor is described in GB 2073891-B. This comprises in one embodiment an alumina body having electrodes, heaters and contacts formed thereon, the body being secured in an epoxy mount. A membrane is securable over the electrodes (to trap electrolyte between the membrane and body) by an O-ring.
Both of the above sensors have the great disadvantage that they require the user to apply the electrolyte solution.
In the case of the sensor shown in GB 2073891B this is by applying electrolyte to the sensor and fixing the membrane thereover with the O-ring. (This has the disadvantages mentioned above) . In the case of the sensor shown in GB 2055476 either the electrolyte has to penetrate the membrane (in which case the membrane is also permeable to sweat leading to uncertainties in electrolyte composition) ; or water has to penetrate the membrane, giving a long preparation time (20 - 50 minutes) and uncertainty as to electrolyte compositions.
The inventors have found a way of overcoming some or all of these difficulties. Accordingly the present invention provides apparatus for measuring the concentration of gases comprising:- i) a disposable sensor assembly comprising; a plurality of electrodes for one or more polarographic or electrochemical cells, the electrodes being disposed on one face of a ceramic substrate; leads passing through the substrate to electrical contacts disposed on another face of the substrate; a membrane, permeable to the chemical species or derivatives thereof, secured to overly the electrodes on said one face of the substrate; and an electrolyte medium provided between the membrane and said one face and ii) a reusable base part comprising complementary electrical contacts for engagement with the electrical contacts on the disposable sensor assembly. Further features of the invention are set out in the claims and in the following illustrative description of two forms, of a combined 02/co2 sensor and methods of making such sensors. The construction techniques and design features of each sensor described may be used interchangeably on both alternative substrates described.
Reference is made in the following to the drawings in which:-
Fig. 1 shows in section (la) and plan (lb) a ceramic substrate for use in the present invention,
Fig. 2 shows in section (2a) and plan (2b) a mount for the substrate of Fig.l,
Fig. 3 shows in bottom plan the substrate-of Fig. 1 with applied electrical contacts,
Fig. 4 shows in top plan the substrate of Fig.3 with applied anode,
Fig. 5 shows in section the substrate of Fig.4 showing a lead from anode to contact,
Fig. 6 shows in plan the inverted substrate of Fig.5 mounted in a jig for insertion of cathode leads,
Fig. 7 is a sectional view of line A-A of Fig. 6.
Figs 8 & 9 show the substrate of Fig. 7 and sealing of the bores with a glass paste.
Fig. 10 shows a substrate printed for soldering and for eventual mounting of a copper ring,
Figs. 11 & 12 show deposition of glass materials on the ring and substrate respectively prior to their joining,
Fig. 13 shows a jig and assembly of the substrate and ring,
Fig. 14 shows the assembled substrate and ring in a later stage of processing,
Fig. 15 shows the finished assembly of substrate, ring, electrolyte and membrane.
Fig. 16 shows an alternative unitary construction for the substrate and mount.
A sensor, for blood 02 and C02/ and in accordance with this invention may consist of two parts:
A) A non-disposable head consisting of a heating element and thermistor temperature sensors mounted in a thermal load which extends to the front face of the head, together with electrical contacts for a disposable component, and cable and plug. This is not shown in the drawings..
B) A disposable component, which consists of 02 and C0 sensing elements, electrolyte and membrane together with a housing by which the disposable element can be attached to and used with the non-disposable head.
The construction of part A above is by means of conventional materials and techniques, and will not be
enlarged upon here.
The disposable element uses ceramic materials, glass and precious metals in its construction, with hybrid techniques which are at present used in the electronics industry. The inertness of the materials greatly improves electrode stability over known electrodes by minimising side effects which together with the more uniform production method gives more reproducable performance from one electrode to another than is the case with presently known sensors.
The 'gas sensitive' electrodes consist of a polarographic oxygen electrode capable of detecting 02 quantitatively; and a pH sensitive electrode which, in conjunction with a bicarbonate containing electrolyte, is capable of detecting C02 quantitatively.
Both these electrodes are mounted on an alumina substrate in the form of a disc 1, which also carries a layer of silver 2 which acts as a reference electrode for the pH electrode and as an anode for the oxygen cathode. The alumina substrate may for example comprise HILOX 961 (Trade Mark) , a dense sintered 96% alumina supplied by Morgan Matroc Limited, of Stourport-on-Severn, England. A suitable size of disc is of diameter 8 mm, thickness 1.0 mm, ± 0.01 mm.
The alumina substrate is mounted in a ring 3, which forms a mounting for the disc, and also holds a membrane mounting ring 4. The ring 3 may be of the same material as disc 1.
A major part of the disposable sensor production is concerned with the processing of the alumina substrate l to produce the three electrodes. The sequence of operations on the ceramic materials is largely determined by the processing temperatures involved at each stage.
The first step is to print electrical contacts 5,6,7, shown in Figure 3, on the rear of the alumina disc. This is done ing ESL9635B (Trade Mark) , a silver palladium ink supplied by Electro-Science Laboratories, of New Jersey, U.S.A. and applied using a screen printer. This may be hand loaded, or automated with machine feed of components as is conventionally known. The components are conventionally dried and then fired at 850°C to 900°C in a belt furnace.
This again can be a one-off or continuous process.
The anode electrode 2 is then formed in an e.g. 3mm diameter 0.5mm depth circular depression 8 in the front of the disc. An ESL 9601 (Trade Mark) pure silver ink supplied by Electrode Science Laboratories, is used in a conventional dispensing machine such as is made by Page Precisma Industries Inc., and using an e.g. 1mm nozzle. Using a single nozzle in a one-off manual process takes about five seconds per disc, but with a multinozzle dispensing head this process would be slightly faster and could be automated. This process puts a layer of wet silver/glass paste in the recess around the central hole 9 (which is tapered, having a diameter of e.g. 3mm, in the front face and 0.6mm on the back face) shown in Figure 4. Before this paste is fired the discs are vacuumed to draw the wet paste down through the central hole '9 and over the periphery of contact 5 as shown in Figure 5. This process also takes a few seconds in handling and process time. Having vacuumed the paste through to make connection, the discs are dried and fired at 850°C, producing the front anode and annular feed through 10 depicted in Figure 5. Of a typical batch of 25 discs processed manually using this process all 25 anodes made connection through the disc first time. The contact 5 to anode 2 resistance is of the order of 0.1 - 0.2 ohms.
The next step is to put leads in the form of gold wires 11, 12 in the two outer holes 13, 14 of the alumina disc. This is done with a wire bonding machine having automatic feed, using e.g. 0.025mm gold wire. A suitable machine is a TSB 460 (Trade Mark) supplied by Hughes Aircraft of Carlsbad, California. The machine control sequence is modified to bypass the first bond and feed through before the second bond. A disc is placed face down on the machine in a circumferential jig 5 to give clearance at the front of the disc (see Fig. 6). For each hole 13, 14 the first operation is to feed about 2mm of wire (rather than make a first bond as such bonding machines usually do), and this wire ii,- 12 is then placed through the appropriate hole 13, 14 (having diameter of e.g. 0.6mm) in the alumina discs and the wire 11,
12 is then bonded to the electrode 6, 7 adjacent to the appropriate hole. The wire 11, 12 is then conventionally cut off. Figure 7 shows the gold wire feedthroughs in place.
The disc remains in the jig 15 which gives front clearance until the next stage of sealing the holes has been completed. All three holes in the disc are filled and sealed with R250 (Trade Mark) glass from EMCA of Sawston, Cambridge, United Kingdom. This is a vitreous glass compatible with alumina. The glass at present is obtained as a dry frit and is normally used as a dielectric coating ink. The amount of binding agent is too high for the present application however and for these purposes is modified to reduce the level of binding agent (and the likelihood of bubbling during firing of the glass) by adding a modified organic solution to the dry glass powder. This solution is solution of propylene carbonate containing 1-2% collodion (which is a solution of nitro-cellulose and pyroxlin in ether and alcohol) to obtain a paste of suitable consistency. The glass is injected in a controlled quantity using the same machine as used for the depositing of the anode, with e.g. a 1mm or 0.8mm nozzle, from the back of the disc. This is the same side as the previous operation, the discs do not have to be turned over.
The seals then have to be dried and fired, but because of the thickness of glass the times required are extended. The drying time is overnight at 40°C, but this could probably be reduced. The discs, when dry, are burnt out at 260°c to 280°C for an hour and fired at 750°C for 15 minutes. These discs should remain face down throughout these processes to ensure that the glass sets near the front of the disc and that any bubbling of the glass occurs at the back of the disc where it is not so important. The above operations produce an alumina disc with gold wire feed-throughs in two holes and a central silver anode in the recess. This is shown in cross section in Figure 9.
As an alternative to using a suction deposited thick film, the electrical connection to the anode from the rear contact pad may be made by gold wire bonding. Wire bonding is used to produce the cathode and C02 sensing connections, and a
gold wire for the anode is welded on during the same stage. This will also use the modified bonding cycle, which omits the first bond, produces a wire feed and then does a second bond and pull off. The connection for the anode may be made anywhere on the annular contact ring around the central hole. As with the 02 and C02 connections, this will leave a length of wire protruding through the hole, and any jigs used to hold the alumina disc for wire bonding should leave space for these wires.
The disc is then turned over, i.e. face up so that the ends of the gold wires are protruding through the three holes. The anode will have already been deposited/printed and fired on at this stage, and the gold wire protruding through the anode hole must be bonded to the silver anode deposit. This will require the facilities and mechanisms of the wire bonder, but does not require extra gold wire. The substrate is heated and the gold wire pressed against the surface and ultrasonic vibrations applied for a few seconds. Under the correct conditions which are only determinable by experiment this will bond the gold wire to the silver anode. This can be done using the gold wire bonder with the gold wire removed from the nozzle. This, however, requires exact placement of the nozzle on the protruding gold wire, after it has been bent over to lay over the silver surrounding the hole. For large scale production a much larger head (2mm diameter) would be more practical.
When the gold wires are in place the three holes in the disc are sealed, by filling with a glass frit, drying and firing. With this method of anode connection the gold wire must also be sealed off from contact with the electrolyte. This is most easily done using the same glass frit to cover the welded end of the gold wire. This is assisted by having a shallow circular depression (1.5 - 2mm diameter) around the anode hole, so that the second weld takes place in the depression and the well and underlying hole are filled with glass frit. Despite shrinkage this will still seal off the gold wire.
Before assembly of the disc 1 and ring 3 takes place the
ring 3 undergoes deposition of substrate for a solder ring, the purpose of which is detailed later in this description. The ring 3 is printed with ESL 9635B on the rear surface flange, shown in Figure 10, and dried, burnt out and fired in the same way as the ESL 9635B printing on the back of the alumina disc. When this has been done the recess in the ring and the back of the alumina disc can be prepared for welding.
The disc 1 remains face down in the jig 15 from the previous hole sealing process. A ring of glass paste ESL 4032 (Trade Mark) supplied by Electrode Science Laboratories is printed on the back of the disc around the outer edge. The annular ring is for example 0.6mm wide. The printed disc is then dried, at 250°C for 15 minutes and fired at 450°C for 10 minutes, to give a ring of bubble free glass on the back of the disc. A deposit of ESL 4032 is also put down on the inside front lip of ring 3 shown in Figure 11, and fired similarly. In this case the glass must be placed by a dispensing system, for example as described above for the anode, since normal screen printing cannot be accomplished in the recess. For production purposes an annular dispensing head of the correct diameter with an opening of 0.4mm would be used. The disc 1 and ring 3 are then assembled and held together in a stainless steel spring loaded jig (Fig. 13) which will press the two parts together as the glass softens. The assembly is refired at 450°C giving adequate time for the assembly and jig to reach temperature (e.g. 1 hour) and for the glass to flow, giving a single ceramic unit. Drying and burn out are not required in this refire operation.
With the ceramic ring and disc assembled the next stage is to grind off the front of the assembly to expose the gold wire feedthroughs. The depth of grinding is sufficient to cut a flat surface on the glass in the two outer holes in the alumina disc. With the glass described above and the firing times and temperatures given this will mean removing about one quarter of the thickness of the disc. The grade of wheel required is a diamond grinding wheel of from 280 to 400 grit. Because of the depth of cut a multipass operation is required, but the ceramic assemblies would be done in batches. The
assemblies are fixed down to a flat sheet by suitable means such as with wax or double sided adhesive tape, so that there is no movement during the grinding process. Upon- completion all wax must be thoroughly cleaned from the assemblies. It should also be noted that after the grinding process the front of the alumina disc is flat except for the anode recess 8 which should be below the surface still. Also, the front of the disc should be protruding above the front of the surrounding ceramic ring. The front face does not require any further shaping or polishing operation beyond this diamond grinding. An assembly processed to this stage is shown in figure 14.
The next step is to sputter a 2mm diameter disc of iridium and iridium oxide over one of the outer feedthroughs 11, 12 so that the electrode comprised by the end of the wire becomes pH sensitive. The choice of which wire to sensitise is arbitrary, but must obviously correspond to the non- disposable head contacts, and must be consistent. The front face of the disc may be masked with adhesive tape, but permanent metal masks have been used with success. The ceramic assemblies must be thoroughly cleaned for sputtering, and the use of organic solvents for final washes should be avoided. The disc is placed in a chamber and pumped down to 10~5 Torr to remove harmful trace material. Pressure is then brought up to 0.015 Torr under argon and the masked disc sputtered with an iridium oxide/iridium electrode.
Iridium electrodes are used at a spacing of 25mm and sputtered at 1500V dc at 10mA per 10cm diameter target or at 1000V ac at 50 - 100W per 10cm diameter target for 2 hours. The atmosphere is then changed to 50% 02, 50% Argon gas at same pressures as above. Sputtering continues for a further 2 hours at 1500V dc at 20mA per 10cm diameter target or 1000V ac at 50-100W per 10cm diameter target.
After sputtering, the components may be aged by immersing in 0.1% sodium sulphite solution for 8 hours.
When sputtering and ageing are complete the ceramic assembly can be soldered to a copper washer. This washer serves as mount for the disposable part of the overall sensor
and as a good thermal contact to the heater held in the non- disposable part of the sensor.
The palladium silver metallizing on the rear of the ring 3 is tinned with a 2% Ag lead tin solder such as Frys 270 (Trade Mark) obtainable from Fry Metals of London, and which may be obtained in fluxed wire form. For automatic production the solder may be screen printed on and treated in the normal way. The copper washers are similarly treated and the parts assembled by reflow soldering in suitable jigs.
The next step is to convert the silver central anode 2 into a silver/silver halide anode. Until now the sensor electrolyte has been used for this, but a simpler bromide only electrolyte would do. Electrical connection may be made to the anode via the rear connection pad, and a current of l/2mA at 400 mV for 5 minutes will do. The sensor face should be washed off afterwards to make sure that the wetness does not carry copper ions over into the working electrolyte. Sufficient of this electrolyte is then added and a p.t.f.e. 25μm membrane applied. Note that the clean alumina wets very easily with electrolyte, so air bubbles are not a problem. The electrolyte used may be 0.03M KBr, 0.1M NaHC03, 0.1M NaN03 in 5% H20/95% propylene glycol solution. The membrane is applied over the whole of the ceramic front face of the sensor and held under tension and in place with a mounting ring around the tapered seating on the outer ceramic ring. An assembly to this stage is shown in Figure 15. Excess electrolyte should be dried off. On prototypes a nylon mounting ring has been used, but for moulded production ceramic rings the dimensional tolerances will not be as tight as machined prototypes. To accommodate this a more flexible material such as a grade of ABS or high density polyethylene could be used.
The membrane assembly can now be inserted into the plastic housing. It is held by means of circumferential clamping of the outside of the copper washer, which allows the central components to move slightly. Thus when the disposable sensor is loaded onto the non-disposable head the spacing ring brings the ceramic parts parallel to and
correctly spaced from the thermal block. The outer plastic housing of the disposable sensor is assembled from an inner and outer ring which are clamped up and glued together. Alternatively a latching fit may be used. The mounted sensor is shown in Figure 15.
The ceramic comp'onents must be correctly aligned, on an angular basis, with the connecting pins in the non-disposable head. These pins are lined up with the thread on the non- disposable head, so that the ceramic components must be aligned to the thread in the inner plastic ring of the disposable sensor. For large scale production this may not be the best system, and an alignment notch can be moulded directly into the ceramic components (see below). For a push on rather than a screw on disposable such a notch would align directly with a pin in the non-disposable head.
Fig. 16 shows in plan, and in two sections along lines A-A and B-B, a unitary substrate and mount formed by moulding alumina powder and binder in a mould, removing the green compact from the mould, and firing. This substrate includes moulded-in through holes 30,31,32 to accomodate lead throughs for the oxygen sensor, the anode, and the carbon dioxide sensor respectively of the electrochemical cells of the two gas sensors.
The front of the ceramic disc incorporates an annular recess 33 which serves as an electrolyte reservoir in the final product. The front face of the ceramic disc also has a central raised portion 34 with a shallow radial slot 35 which has a dual function. Firstly, the radial slot 35 enables the formation of the anode and the connection to it to be formed during the high temperature processing of the ceramic disc and secondly it serves as an orientation reference during such processing. The rear face of the ceramic disc has a flat annular ring 36 around its periphery which is used as a mounting surface. The central area 37 of the rear of the disc is thereby recessed in order to minimise the thickness of material between the heated block in the re-usable base part and the skin. At the outer edge of the central recess is a small cylindrical recess 38 which serves for angular
orientation during processing of the front face of the sensor. The outer edge of the ceramic disc is formed with a slight taper, being narrower at the rear than at the front. The taper is sufficient to lock a flexible membrane ring to trap a membrane against the front face of the sensor. The front outer edge of the ceramic disc is slightly rounded to enable push-fitting of the membrane ring and should be smooth to reduce the risk of membrane damage.
The essential elements of the oxygen and carbon dioxide gas sensors are formed by processing the ceramic disc as described below.
The first step is to print the electrical contacts for the anode, cathode and carbon dioxide sensor in the recess 37 on the rear of the disc. These electrical contacts are segmental in form in order to give an angular tolerance in mounting the sensor on the re-usable base part and to facilitate wiping contact during that operation. The profile and location of the contacts is such as to place part of their areas in close proximity to the feedthrough holes 30,31,32 in the disc. The contacts are formed by off-set printing from a cliche using a conformable silicone rubber pad. The ink used is a particulate silver palladium composition with a conventional glass frit bonding medium, such as ESL 9635B, which requires a small degree of dilution (approximately 10% of type 400 thinners as supplied by ESL) to bring it to a suitable consistency for this printing process. The printing takes place with the ceramic disc heated to promote the deposition of a sufficient quantity of ink and a dwell period of a few seconds may also be incorporated into the deposition cycle for this same purpose. After printing the ink is dried by one of several conventional techniques, but not yet burnt out or fired. The printing process is such that several discs may be printed in one operation, with a multiple cliche and holding jig. The holding jig should incorporate a locating pin for each disc to mate with the radial anode recess 35.
After drying the contacts, the next step is to produce the anode on the front face. The discs are held face up in a
jig either singularly or in multiples, with the angular location determined by a mating pin in the cylindrical recess 38 in the rear face of the ceramic disc. The anodes may be produced by e.g. pressurized ink deposition, or be printed in the same way as the rear face electrical contacts. The ink used for the anode must be pure silver and should have thermal processing requirements close to those of the electrical contact ink. A suitable material is ESL9601, which again may require some dilution to obtain a viscosity suitable for cliche off-set printing. The anode is consumed during the operation of the oxygen sensor, as well as during the formation of the reference electrode during sensor manufacture. However, because the unit is a disposable item, the amount of silver required is finite and for continuous operation for one month at a cathode current of 5 nA approximately 10 microgrammes of silver would be consumed. This is very much less than would be deposited over the available anode surface in a single printing operation.
After printing, the anodes on the discs are again dried. The discs are then burnt out and fired in a belt furnace with a maximum temperature of 850°C for 10 minutes, which makes both the contacts on the rear and the anode on the front of the disc permanently adherent to the disc.
The next stage is to produce the lead throughs in the three holes 30,31,32 in the ceramic disc. This is done using a gold wire bonding machine with a modified bond cycle and 0.001" gold wire. The disc is placed face down in a holding jig which allows a clearance of 2-3 mm over the raised central portion 34 of the front face. The bonding operation takes place with the discs at 150°C so time must be allowed for heating which takes perhaps 30 seconds. The other parameters of the process, i.e. pressure, bond time, height and power can be preset on the bonding machine. The wire bonding cycle is modified to give a short feed of approximately 2 mm followed by a conventional second bond and pull-off, followed by the electronic flame-off cycle to cut the wire to a predetermined length. The operation for the installation of a single lead through wire then consists of placing a pre-fed length through
the appropriate feedthrough hole, tracking to the contact close to the hole, wire bonding and separating. This is done for each of the cathode, anode and carbon dioxide feedthroughs 30,31,32, with two wires placed in each of the latter two holes.
The next operation is to turn the disc over and bond the anode lead throughs to the deposited anode. This is again a heated ultrasonic operation using the gold wire bonder but with no feed wire. The two anode lead through wires from the previous operation are welded to the anode and assuming that they are of the correct length, the weld will occur close to the hole with no removal of excess wire being required. The gold wires for the oxygen and carbon dioxide sensors will also be protruding through their respective holes and before the next stage these wires may be pulled straight and laid flat, by lightly brushing the front face of the disc.
The three feedthrough holes 30,31,32 are then sealed using a specially prepared glass frit. This is EMCA R250 which is supplied as a dry powder. It is mixed with propylene carbonate containing 2% colloidon as a binding agent, to a suitable consistency. A measured quantity of the resulting paste is injected into each hole from the front surface, taking care not to allow any excess to contaminate the rear contact pads. During this operation the lead through-welds on the anode are also coated. The seals are then dried and fired under vacuum. The complete cycle is to pump down to 0.1 mm Hg whilst heating to 375°C for hour which will dry out the glass frit. The discs are then heated to 650°C still under vacuum to burn off the binding agent and melt the glass. This condition is held for 15 minutes and the pressure is then slowly returned to ambient whilst heating to 700°C for a further 15 minutes to flow the glass and collapse any remaining voids. The process is carried out with the discs face down and should leave a solid void-free glass seal in each hole close to the front surface and covering the weld to the anode.
The central portion of the disc is then ground back by 0.2 mm with a 300 grit diamond wheel. This should leave the
oxygen and carbon dioxide lead throughs flat and level with the surface. Small voids in the lead through seals may be tolerated provided they are not close to the gold wires.
The discs must then be thoroughly cleaned prior to sputtering. This is done ultrasonically in a series of washes:- dilute detergent followed by rinsing and several passes in distilled water. During this and subsequent cleaning and processing stages the discs should not be handled with ungloved hands.
The discs are mounted face up in a suitable jig for the sputtering of the pH sensitive surface over the carbon dioxide lead through. This requires that the remainder of the disc be masked off, which may be done by several means. Adhesive tape or a close-fitting metal or plastic mask are suitable methods. If a plastic mask is used it facilitates subsequent recovery of iridium.
If a separate re-usable mask is used the disc orientation relative to the sputtering jig needs to be controlled, which may be done by means of the recess 38 in the rear of the ceramic disc. The sputtering process generates heat in the substrate and is important that the temperature is limited by sufficient thermal conductivity of the substrate holder and discs. The pH sensitive coating may be deposited by d.c. magnetron sputtering in a 0.1 Tesla field, considerably reducing the time required over simpler d.c. sputtering. The pressure should be in the range 0.2 to 0.3 mm Hg and sputtering is done for half an hour under pure argon followed by half an hour under 50% to 70% oxygen, balance argon. The current should be approximately 20 A for a 4 inch target. It is probable that the current could be increased several fold over this figure provided the heat so generated can be dissipated.
After sputtering the discs are read for mounting, membraning and use.
The discs are placed face down in a suitable jig and rigid pvc washers are fitted to the mounting surface 36 with adhesive. The adhesive may be a two part methylmethacrylate adhesive, the monomer being placed on one part and the
hardening agent on the other. The discs and washers are brought together and some pressure applied for about 15 seconds, after which they may be removed from the jig. Full hardening may well take longer than this and can be ensured by baking the assembly at 60°c for 2 hours.
The discs are then ready for membraning. The membrane may comprise a 12 micron, cellulose membrane (such as Cupraphan [trade mark of .Medicell International Limited of London, U.K.]) as a spacing membrane, 8 mm in diameter to cover the central portion of the disc; followed by a 12 micron p.t.f.e. membrane covering the whole of the disc face. To apply the membrane to the disposable sensor sufficient electrolyte is deposited on the front face of the disc which is then covered by the twin membrane layer, cellulose first. A slight pressure is applied to the face of the disc to exclude excess electrolyte and a membrane ring is then pressed over the membrane and disc to- lock on to the peripheral tapered edge as mentioned above. The excess p.t.f.e. is then cut off from around the membrane ring and the plastic washer wiped dry. The electrolyte consists of 10% water, 90% glycerol with 0.03 molar sodium bicarbonate and potassium bromide and 0.1 molar potassium nitrate.
The membrane ring may comprise any flexible material having a suitable resilience to lock onto the ceramic substrate.
The disc is then mounted in a two-part mounting ring by means of the plastic washer which extends beyond the periphery of the ceramic substrate and membrane ring combination. The mounting device may comprise an outer holder and an internal threaded ring. The two halves of the mounting may be held together with cyanoacrylate adhesive which also fixes the plastic ring, and hence the disc, in place. The angular orientation of the disc with respect to the threaded ring is critical since it determines the position of the contact pins on the re-usable base part relative to the contacts on the disc. The orientation of the disc after membraning may be determined mechanically by means of the cylindrical recess 38 in the rear of the disc and a similar reference point may be
built into the threaded ring.
After wetting, the iridium oxide coating on the disc will begin an aging process which produces an exponentially reducing rate of drift in C02 signal for a. constant C02 concentration. It 'has been found that this process can be speeded up by immersion in 0.1. sodium sulphite which will reduce the drift rate to an acceptable value in 24 hours. However the same process will occur in approximately one month of wet storage, i.e. membraned, if the artificial aging is not done.
All of the above description has related to use of the present invention as applied to transcutaneous sensors. However the invention is not so limited and it is clear that sensors of the type described may be used in a wide variety of other applications. Further, the invention is not restricted to the particular form of sensors shown and other gases or chemical species may be measured using the technology described.