MOSFET LF Noise under large signal excitation: measurement, modelling and application
Promotion date: 15 April 2005
| To very briefly summarize: it is about noise in transistors. Roughly speaking there are two types of transistors: bipolar transistors and MOSFETs. The latter is most common these days, because the computer industry solely uses MOSFETs, (CMOS technology comprises of two types of MOSFET) and because they are so common, the rest of the industry has also started to use them. One of the problems of MOSFETs is their low frequency noise. The cause of this noise is not exactly known. We found that the noise decreases and there is a slight change in character if you switch the transistor off and on, an interesting phenomenon that has never properly been investigated. That is what I did. |
What was your thesis about?
To very briefly summarize: it is about noise in transistors. Roughly speaking there are two types of transistors: bipolar transistors and MOSFETs. The latter is most common these days, because the computer industry solely uses MOSFETs, (CMOS technology comprises of two types of MOSFET) and because they are so common, the rest of the industry has also started to use them.
One of the problems of MOSFETs is their low frequency noise. The cause of this noise is not exactly known. We found that the noise decreases and there is a slight change in character if you switch the transistor off and on, an interesting phenomenon that has never properly been investigated. That is what I did. Half a year after starting we discovered at a conference in Leuven that two researchers from Israel had done measurements on the same phenomenon a few years earlier, but had not followed it up. We met them at that conference and discussed our re-discovery of the effect.
Why is it important to reduce the noise in transistors if they are mainly used for computers?
For the computers themselves it is not very important to reduce the LF noise. But noise plays an important role in the peripherals of computers: such as monitors, Bluetooth and other wireless systems, beamers, microphones etc.
And you now know what causes the noise?
The current fluctuates because of irregularities in the transistor. The ‘jump’ of single electrons in and out of 'traps' prevent an even flow of current. This noise is sometimes called 'popcorn noise' because it sounds somewhat like the popping of popcorn. By switching the transistor off and on, the electrons that are stuck in traps are forcibly removed. That is why the noise decreases. By quickly switching on and off you can reduce the noise of the transistor.
So that is possible now?
No, not quite. On the one hand too much is still unknown for practical applications. We don’t know if every transistor is the same and what other factors play a part, and on the other hand there are less complicated technologies for noise reduction.
A large part of my thesis deals with the development of new, very delicate, measurement techniques and modelling of the findings. Now at least we understand more about the phenomenon itself, and that is why you do research in the first place. Negative answers are important too. Moreover, understanding these electron traps is of increasing importance, as new gate oxides will probably have lots of traps.
What did you enjoy most about your research?
The fundamental physics, being able to study the behaviour of single electrons. It is absolutely fascinating. And the measurements I did are completely new.
What didn’t you like?
I would have liked to have made the first transistor without noise whatsoever. Although I am sure I have contributed significantly with regard to insight and understanding, my research does not contribute much to the concrete development of new types of transistors.
What are you going to do next?
I’d like to try industry. I think the experience will contribute to my educational abilities. I would like to diversify my knowledge by specializing in a new field.