Deuterium in the Gate Dielectric of CMOS Devices
[Financed by Philips Semi conductors]
Promotion date: 29 October 2004
| Electronic switching on chips these days primarily comprises complementary MOS*-switches called CMOS (*Metal Oxide Silicon, a method to make transistors). The gate dielectric is an insulating layer of two nanometres, which is about ten atomic layers, so in fact hardly anything at all. So when exposed to an electric current of a few Volts the insulating capacity is questionable. When the device is used as a flash memory, it should not be allowed to ‘leak’ under normal conditions, but it does. The defects caused by repeated writing and erasing actions are believed to be created by the hydrogen atoms. |
What was your thesis about?
Electronic switching on chips these days primarily comprises complementary MOS*-switches called CMOS (*Metal Oxide Silicon, a method to make transistors).
The gate dielectric is an insulating layer of two nanometres, which is about ten atomic layers, so in fact hardly anything at all. So when exposed to an electric current of a few Volts the insulating capacity is questionable. When the device is used as a flash memory, it should not be allowed to ‘leak’ under normal conditions, but it does. The defects caused by repeated writing and erasing actions are believed to be created by the hydrogen atoms.
Of the about 108 chemical elements Hydrogen is the lightest, comprised of one electron and a proton. Of almost all elements have a heavier version: the heavier version of hydrogen is Deuterium comprised of a proton, an electron and a neutron. The neutron makes Deuterium two times as heavy than hydrogen. Tritium is three times as heavy, but radioactive. The idea was to use Deuterium instead of hydrogen to counteract the occurrence of defects.
Is Deuterium an element that exists in nature?
Yes, in seawater for instance. But you can fabricate it. It is of course more expensive than hydrogen, but if the economic demand increases, it gets cheaper.
We used heavy water as a base to get our Deuterium.
Did you succeed?
Yes and no. Deuterium gives a noticeable better result on the interface, but that is all.
And of course I had hoped for more. But it is a conclusion, albeit a negative one.
Deuterium is evidently not the ultimate broad solution, it may work for a certain type of flash memory.
What did you enjoy most about your research?
I started it because there were so many sides to it. It involved work in the clean room making your own devices apart from the electrical characterization. So modelling also came into it and made it a bit of everything.
What didn’t you like?
We had hoped for a better outcome, that’s for sure. And it creating an optimum environment for our tests took longer than I expected.
Have you worked with other groups?
Well, because of the disappointing results I did not go to conferences much. I have found that negative results are very hard to publish! But I have been to Ireland and next week I will go to the US. So people are interested after all.
What are your plans for the future?
I would like to continue in research, in a different field perhaps. I have the practical problem that my wife has a good job in Groningen, and in that area there are not many research jobs for technical research. If I cannot find anything there we’ll have to find two jobs I another part of the country or abroad.
You were also a member of the Faculty Board?
Yes. But I find it is increasingly hard to find people for it. Starting scientists are very busy with their own research. But it is important, and it does not look bad on your CV.