In 1981 my PhD project entered the last phase : design of a CCD line array, fabricate the device with the newly developed ITO gates and the newly developed polyimide isolation layer. In the 3 years before, I researched the implementation of ITO in an existing CCD process, and the fabrication of a new CCD would be the proof of the pudding.
Being a PhD student, we had to do everything ourselves. Starting with the lay-out of the CCD. This was almost a daily fight to get access to the lay-out system (Applicon), because many people had to make use of the same system and only one machine was available at our university. But finally, when the lay-out was ready, the layers of the chip had (literally) to be cut into the top part of a two-film foil (we called it “stabylene” but I do not know whether this is an official name) on a scale of 200:1. Next the structures were manually pealed off with a tweezer. This was a very time-consuming task and needed nerves of steel. Because every single failure made, resulted in a new start from scratch for that particular mask lay-out.
When the masks came back from the mask shop (thanks Piet) the processing of the devices could start. At that time we were working with silicon wafers of 2 inches. Thanks to the crew of the clean room (Viviane, Eddy, Rita), the base processing of the CCD flow chart took only a couple of weeks. Then I could start with the deposition, anneal, etching, contacting of the ITO layers. Everything had to be “first time right”, time because money was very limited in a PhD project. Believe it or not, but a full lot contained exactly 4 2”-wafers !
And then the testing (thanks Tony), also here everything was done manually. Soldering, de-soldering, re-soldering till finally the big moment : a functional device with the new material incorporated ! A 256 pixel CCD line array with ITO gates ! What a great moment. The device realized a much larger QE than the older technology with poly-Si gates. If I remember well the QE was increased by a factor of 2 overall, and even a factor of 4 increase in the blue part of the spectrum. Unfortunately the dark current of the new device was also much larger than the devices fabricated in old technology. We were not able in reducing the radiation damage introduced during the deposition of the ITO film. Because the temperature of any annealing step after ITO deposition needed to be kept low, the interface damage could not be repaired. A CVD deposition technique for the ITO layer could be a better alternative because CVD does not introduce interface radiation damage. But a CVD machine to deposit ITO was not available at those times.
All together : as far as the optical characteristics of the devices were concerned the project objectives were met, but a hard price has to be paid in the form of increased leakage current and consequently a larger noise floor. Several years later, the ITO technology was successfully and industrially incorporated by Kodak in their CCD process. It is not known whether Kodak could benefit from the ITO work I did, but it would surprise me if it was not, because my supervisor (Gilbert) was also working as a consultant for Kodak. And actually I hope that Kodak indeed did learn something from my PhD project, then my work has contributed to the development of the CCD technology in general.
Albert, 08-08-2021.