Non Volatile Technologies Pty Ltd

SYDNEY  AUSTRALIA   2227    (ABN 96 078 508 264)    Ph:+61 2 9016 4695 

"Helping others to have a future assures our own."

Plasma Reactor for Depositing Thin Polymer Films to Make Capacitors

Introduction


From 1993 to 1997 I was involved in a project, the aim of which was to develop RFID technology suitable for use in airline baggage labels.  It was, by any measure, an ambitious project.  The technology was based on an RFID transponder that utilised a new type of Ferro-Electric Non Volatile Memory (called FRAM™) and was powered at 132kHz.  The transponder broadcast its signal on 32 different frequencies randomly from around 1MHz to 30MHz with a data transfer rate of around 100kBits/Second.  To do this required two antennae; one for powering and one for transmission.  The silicon chip (called an Application Specific Integrated Circuit - ASIC), comprising the transponder, required there be two capacitors; one 5.6nF and one 330nF. (See Figure 1)  The smallest temperature stable ceramic chip capacitors I could find were from a company called Novacap founded by Dr Andre Galliath (a truly exceptional person, on many levels, I'm pleased to call a friend).  As thin as Novacap's capacitors were, they still made incorporation of the RFID device into labels and credit cards a challenge.  Being ceramic, they were also susceptible to damage. 

In 1994 I learnt about work the CSIRO had done laying down ultra-thin, polymeric films in a plasma for the purposes of placing a coating on long-wearing contact lens.  Before approaching the CSIRO, I filed for a provisional patent describing how it was possible to manufacture ultra-thin (and hence flexible), high value, temperature stable capacitors to a precise value.  The CSIRO already had some experience in laying down polymers in a plasma using hexa-methyl-di-siloxane as the monomer to coat the surface of contact lenses so as to make them suitable for extended wear by users.  My idea was to sputter silver, copper, aluminium, platinum or gold down between a mask.  After laying down the metal, a broader mask would be employed to lay down a thin layer of HMDSO.One of the materials they used was hexa-methy-di-siloxane (HMDS0).  HMDSO was significant because it contains silicon and therefore had a higher dielectric coefficient than compounds that were purely organic.  (My eventual aim was to find a monomer that incorporated a titanate, such as Tetrabutyl Titanate, and therefore would yield a higher dielectric constant.)  The concept of laying down ulta-thin films attracted my attention for the following reasons:
  • The capacitance resulting from having a dielectric between two metallic plates is dependent on:
    • the area of the plates facing each other,
    • the electrical properties of the dielectric, and
    • the separation distance of the plates or, put another way, the thickness of the dielectric.
  • HMDSO would have a higher dielectric than a simple plastic film; and
  • The ability to lay it down thinly would be useful in achieving the highest possible capacitance for any given area of plate provided the film could stand the voltage.
  • Because the Bagtag operated at very low voltages, ie, no more than 7 Volts, it was possible I could create thin films that could stand this voltage.
  • A capacitor using a semi-plastic dielectric would be more flexible and therefore less likely to suffer damage in a package such as a credit card or a label than a ceramic chip capacitor.

Personalities at the CSIRO

When making contact with the CSIRO, I met with Dr Albert Mau, the person heading that Division.  He introduced me to Dr Hans Griesser, the person who had been active in the field of plasma polymerisation for some time.  When approval for the project was given by the Board of Intag, I was introduced to Dr Robert Fleming, a physicist who, like myself, had an interest in plasma polymerisation.  The lab technician who built the reactor was Zoran Vasic.  Zoran was particularly useful in that he required very little direction and exercised his initiative on many occasions.  I really only had to sketch out my concepts and Zoran turned them into a functioning reality. 

A Snapshot of what was achieved

Figure 2 shows a micro-photograph of capacitors made, using the process I had conceived, at the Ian Wark Laboratories, Clayton, Victoria with the help of the CSIRO's Division of Chemicals and Polymers. 
Figure 3 is a picture of an actual thin film capacitor that was made using the processes we developed and which worked as a tuning capacitor on a BagTag ASIC. 
Figure 4 is a picture of the Plasma Reactor that was built to my design by Zoran Vasic.


A Primer on Polymerisation in Plasma

A Plasma is created by having a highly evacuated space into which a small portion of a gas such as Argon is inserted and a high voltage is put across two parallel plates separated by some distance.  The effect of doing this is that electrons flow from the negatively charged plate to the positively charged plate using the gas molecules as their "stepping stones".  When this happens the space will glow with light of a frequency related to the gas that is providing the stepping stones for the electrons. 

If you then inject into this highly evacuated chamber a monomer, the electrons travelling from one plate to the other collide with the molecules of monomer, knocking electrons off them and causing the creation of points on the molecule that want to bond with other similarly affected molecules.  (These areas are called "active polar groups".)  When these monomer molecules combine with other activated monomer molecules, the form a polymer (many monomer molecules now chained together).  As the polymeric molecules form, they precipitate out the gas and fall towards earth, landing on whatever surface is below.  The layers formed are very thin (as in nanometres) but because the molecules are still very active, they continue to chain together as they lay on the surface; joining up any holes that may be present.  This results in a hyper-crosslinked plastic membrane that is free from pin-holes.  In the case of HMDSO, if a high voltage was used to cause the plasma, the material that landed on the surface contained a higher percentage of silicon.  This meant it had superior dielectric properties.  Next the hyper-crosslinked nature of substances meant they would have mechanical properties far superior to what was possible in a normal chemical reaction.  This had the promise of making plastic substances that could withstand high temperatures and would have significantly superior tensile strength.  Highly cross-linked molecules were also known to exhibit an very good memory and as a consequence high resistance to permanent deformation.

Concept for Fabricating Capacitors

Process used to fabricate a capacitor

Fig 5. Process used to fabricate a capacitor.

The concept for making the capacitors is shown in Figure 5.  In the case of our experiments, the capacitor was built on a very thin layer of a durable and dimensionally stable polyimide tape; polyimide being sold under the trademark of Kapton™.  In production, it is possible that this Kapton tape would be coated with a couple of nanometres of film that would easily separate from the Kapton when the fabrication of the capacitor was fully complete.  In this way, the Kapton tape could be reused and the thickness of the capacitor would not be added to by the presence of the Kapton film. 
Illustration of a Cross-section of a Thin Film Capacitor

Fig 6. Cross-section of a Thin Film Capacitor
created per Figure 5.

The first (foundation) layer of the capacitor can be any sort or metal or even an inert oxide.  The capacitor's metal electrodes are created by sputtering down a conductive metal through a mask.  Because the electrodes are so thin in cross section, they will offer a significant resistance if the current flow is substantial.  For that reason, highly conductive metals should be considered; metals such as Silver, Copper and Aluminium for example.  The layers of dielectric are deposited through a mask using the process of plasma polymerisation in such a manner that one side of the electrode is exposed and the other side is entirely covered by dielectric as shown.  After that, a layer of metal is sputtered down through a mask to form another electrode on the opposite side to the first.  This process is repeated as many times as is desired; creating alternating electrodes that are joined together.  Figure 6 illustrates the cross-section that develops when this procedure is repeated a number of times.
Process to fabricate capacitors on a continuous band

Fig 7. Process to fabricate capacitors on a continuous band

Figure 7 illustrates how this process would work depositing many layers of electrode and dielectric on a band passing through three stations.  The first station sputters down a metal electrode, for example, the left one.  The next station deposits the dielectric using plasma polymerisation.  The next station sputters down another electrode, in this example, the right electrode.
Creating multiple strips of capacitors on the one band of tape

Fig 8. Creating multiple strips of capacitors
on the one band of tape

Figure 8 illustrates how it would be possible with this process to have numerous rows of capacitors being fabricated at the same time on the one strip of Kapton.
Volume production of capacitors using a reel to reel process

Fig 9. Volume production of capacitors using a reel to reel process

Figure 9 illustrates a further development of this idea where the tape travels from reel to reel.  In this manner, it would be possible to achieve a method of low-cost, volume production of capacitor tape.

Cutting the Capacitor Tape to Create Individual Capacitors


Cutting off capacitors to achieve a set value with precision and little wastage

Fig 10. Cutting off capacitors to achieve a set value
with precision and little wastage


Cutting the Tape

In order to create individual capacitors it is necessary to cut the tape and to seal off the edges of the cut.  When the tape is cut into segments, each segment consists of a left and right electrode and two cut edges which have most probably been traumatised by the cutting action.  From the outset, it was appreciated that, with cutting, there was the possibility of smearing of metal from one electrode to the next.  This would result in short-circuiting the capacitor.  One simple way of possibly solving this would be to apply a high current to the circuit in an oxygen-rich atmosphere and simply burn the metal, turning it into a non-conducting oxide.  Cutting the capacitor and the treatment of the cut edges became one of the more challenging aspects of this project.  It was never properly resolved due to the project being terminated because of a lack of funds and the unwillingness of the Intag Board to proceed with further R&D.;

Achieving Precise Values

Producing the capacitors in a tape then enables the creation of capacitors to a precise value with minimal wastage.  Because the tape has a reasonably consistent value of capacitance per unit of length, it is possible to cut the tape to a particular length, measure the resulting capacitance, and then make an adjustment to the length of the next cut to ensure that the value derived is absolutely precise.  In this manner, it is believed it would be possible to produce capacitors to +0.1% accuracy (See Figure 10.)

Minimising Wastage

A by-product of achieving high levels of precision in the manufacture of this type of capacitor is that one also has little wastage.  The accuracy achieved by this method of production is far superior to that which is possible with ceramic chip capacitors.  For example, it is normal when making ceramic chip capacitors for up to 30% of the capacitors not to fit within the desired tolerance.  The tighter the tolerance, the greater the wastage.  When that happens, those capacitors that fail badly are likely to fit into the tolerance band of another value of capacitor.  Where that is not the case, the capacitors are retained in the store, in the hope of satisfying a special order where the customer does not want a standard value. 

Cutting Methods

The method used to cut the tape is not only fundamental to the successful operation of the capacitor but also will determine, to some extent, the wastage.  Shearing the tape using, for example, a scisors or cutting action involves virtually no wastage at all but unfortunately, experiments indicated that, even with a diamond knife(as can be seen in Figures 11 and 12), the results of this type of cut were not good. 
Having successfully made a prototype capacitor on a continuous tape, the real challenge was therefore to work out a means for cutting off set lengths of tape to form a discrete capacitor without damaging the striated layers and, even if the layers were not damaged, how to prevent metal smearing and therefore short circuiting of the capacitor.  Having found that the sharpest scissors and a diamond knife did not yield the desired results a water jet was tried and though better than a diamond knife, (See Figure 13) it appeared as though the layers were separated.  Using a laser appeared to be the most promising (See Figure 14 and 15)  This, too, had its problems in that there was some contamination by disassociated metal that could potentially short the plates. 
It is likely the best approach would be to cut the capacitors, using very fine abrasive saw suspended in an inert cleaning fluid, with the capacitor squeezed between two platens to prevent edge separation.(Figure 16&17)  A water jet may also yield acceptable or better results.  Because the capacitor is firmly clamped between the platens, separation of the layers cannot occur.  The water jet may eliminate the problem of smearing of metal that could occur if an abrasive cutting saw is used.  If an abrasive saw is used then, after the cutting is completed, nitric acid could be introduced to eat away metal smearing such that the electrodes became slightly recessed between the dielectric layers.(Figure 18)  At that point, the edge could be washed and dried before being coated with an epoxy or acrylic compound to seal it.(Figure 19)  It might also be possible to use a heat-shrink plastic sleeve into which the cut capacitor is placed.  This could be done quite cheaply in a volume production process.  In summary, the matter of cutting the capacitor tape and then sealing its edges is considered to be solvable and, indeed, there could be a number of ways that would prove to be feasible in achieving this.

Experimentation

Experimentation took the form of firstly, manually creating single capacitors through a mask.  The second step was to have a travelling tape and lay down alternate layers of polymer and metal using plasma polymerisation and sputtering.  Of interest, it was possible with some adjustments to conduct both operations in the same chamber at the same time.  It was also found that, as one increased the energy of the plasma, the HMDSO's composition became more heavily biased towards silicon.  The other interesting phenomena observed was that the plastic film had mechanical properties far superior to those possible with a normal chemical reaction.  For example, the tensile strength of the finished product was around 7 times that of "normal" silastic.  The maximum temperature the substance could withstand was also considerably greater, being around 240C as opposed to 110->125C. 

Interesting Possiblities for Additional Research

This work led onto the means of creating armour plate from micro-layers of polymer and amorphous diamond. Because the resultant polymers are hyper-cross-linked, the process also has potential for the creation of super micro-fibres for use in highly specialised textiles.

Copyright © NVTech 2005-2011