I am a registered patent agent and I have a pending patent application based upon an experiment that I carried out over 25 years ago when I was on the technical staff at Watkins-Johnson Company. I was working on circuit assembly techniques for missile programs such as AMRAAM and HARM, and the eutectic solder attach of GaAsFETs to circuit substrates and carriers was one of the issues that I dealt with. At the time, the standard approach was to place a gold/tin solder preform on the bonding site, heat the circuit until the solder melted, and gently scrub the die into the molten solder. This operation was typically carried out under a blanket of a neutral or slightly reducing gas. This approach did not remove oxides that were present, but largely displaced them by the scrubbing operation. The integrity of the die attach process could be evaluated by applying a lateral force to the die. When tested to failure, the die typically sheared away from the circuit in one piece, or left a small amount of GaAs behind.
As a materials scientist, I was interested in finding a method that could remove the native oxides on the solder preform and enhance the solder bonding of the GaAsFET. I dissolved ammonium chloride in methanol and applied a drop of the solution to a GaAsFET die positioned on a gold/tin preform placed on a dummy circuit. The methanol was then evaporated, leaving a light "frost" of ammonium chloride. I selected ammonium chloride since it decomposes into ammonia and hydrogen chloride gases at a temperature below the melting point (283°C) of the eutectic gold/tin preform. Upon placing the assembly on a hot plate, the ammonium chloride decomposed and the gold/tin preform melted and completely wetted the bottom of the GaAsFET die. Inspection after cooling revealed a continuous fillet around the base of the die and a smooth shiny surface that was atypical of the conventional die attach method. During shear testing of the device, the die invariably cleaved, leaving 100% coverage of the die attach surface.
Watkins-Johnson wasn't interested in further investigation or development of the process I had demonstrated for two main reasons. First, the devices they were using had aluminum gate metallization and they were sure that the ammonium choride would attack the gate. Second, they were working with low power, low noise devices for which heat transfer was not a critical concern.
I subsequently worked at Hughes Aircraft's Microwave Products Division where they were manufacturing power GaAsFETs with gold metallization and back-etched vias. Since heat transfer was an important concern, I repeated my earlier experiment by attaching one of their GaAsFETs to a standard gold-plated copper rib. The die attach result was similar to that I had obtained with the Watkins-Johnson GaAsFET, with excellent flow and a shiny oxide-free surface. However, upon cooling I heard a slight "tink" sound. The Hughes GaAsFET had sheared in half due to the compressive forces induced by the copper rib. The die attach process had produced a much more robust bond that resulted in much greater mechanical coupling than was typical. The die had fractured under the compressive stress in part because of its size and in part because of the stress concentrations introduced by the back-etched vias. In spite of the excellent coupling, Hughes wanted no part of an assembly process that could destroy a die.
Obviously, since the basic process had been demonstrated at Watkins-Johnson, I had no intellectual property rights and aside from repeating my initial experiment with a Hughes GaAsFET, I did no further work on the process. In 2006 I was doing some circuit design for a garage research project and realized that the process might be useful. I contacted my supervisor from Watkins-Johnson and he told me that nothing further was ever done with the ammonium chloride die attach technique.
At this point I filed a patent application (11/625,345) which I eventually abandoned after restriction. The pending application (12/725,356) is a divisional of the abandoned application. I have avoided claiming the process as demonstrated over 25 years ago (i.e., in open air), but I have been ambivalent about prosecuting the patent application due to concerns about patentability related to diligence and the amount of time that has elapsed. The experiment was never recorded in writing by me, and to my knowledge never published by any of the handful of people that were aware of it.
I would very much like to hear the opinions of patent attorneys on this forum with regard to the patentability of above described process carried out in an enclosed chamber. Since I never performed the process in a chamber or discussed performing it in a chamber, I am assuming that my past activities would not be a bar. There are a number of other claim limitations that were developed when I recently decided to file. My main concern is that of diligence, but since my employer did not want me to pursue the process development, is there a lack of diligence on my part if I fail to develop something on my own when it would be the property of Watkins-Johnson? Also, would Watkins-Johnson have rights in an invention based upon a process that they chose to ignore for 25 years? I believe that the die attach process has value, but on the other hand, I do not want to obtain a patent that I would not be entitled to. Thanks to those who have taken the time to read this.
Author
Topic: 25 year old invention ignored by employer (Read 3133 times)
