Cambridge University Engineering Department
Scientific Imaging Group
In the years 1972-1987 this lab was developing techniques to dynamically image semiconductor devices. This work was pioneered by WC Nixon and Graham Plows .
The technique uses a principle referred to as Stroboscopy. The electron beam is blanked in synch with the device clock rate. A Scanning Electron Microscope (SEM) is used to image a semiconductor device ( An optical micrograph shown here with the lid taken off). The device is powered up as it would be normally and clocked at its design frequency. The SEM is operated in normal secondary mode but with an reduced accelerating voltage of only about 1-2kV.
In the years 1982-86 John Thong ( now at National University of Singapore) and I were involved in a research project to find the limitations of Stroboscopy in an SEM. We used a Cambridge S200, interfaced with a Lintech. This was a great combination of instruments and produced visually stunning results that never ceased to amaze visitors to our lab. We were also doing some collaboration with Texas Instruments and worked alongside design engineers in troubleshooting prototype devices that did not achieve specs. This gave us considerable experience in developing techniques to optimise the hardware and software. The final result was that we achieved a rate of 64 Gigahertz. ( Ref: Ultra Highspeed electron beam testing system. Microelectronic Engineering 1987; 6 : pp 683-688). We didn't realise until some weeks later that we might just have been able to achieve 100 Gigahertz, however it was much too late as some of the specialised equipment had been very kindly loaned to us and had now been returned. What was even more interesting about this feat, is that there were no devices that we could get that would operate at that speed. So what does John Thong do about this....he makes one !!!! We had that at time an Electron Beam Lithography machine, courtesy of Texas Instruments. John and I were primarily responsible for the operation and maintenance of this machine and so it was a relatively simple task for John to design and fabricated a device. Well...everything is relative.
The next section contains a few AVI files (these are extracts from a much longer video sequence but they would take ages to download, so instead I will extract bits from time to time and place them here). which are all about 10 seconds length and about 230Kbytes in size.
The clip is best viewed with Firefox and the WindowsMedia Plugin, there is also audio.
The image of the device is displayed in a unique manner, as the video clips show. Observe (AVI video clip of Microchip in dynamic Voltage contrast mode) in "real" time while in the SEM.
This AVI video clip (low mag image) is a device operating at approx. 50Mhz and with the stroboscopic mode of operation we can in effect slow the chip down (higher mag image) such that we can determine whether the device is operating at its design specifications.
All images and video clips on this page are Copyright © and may not be used without permission from the author.