Information Technology Reference
In-Depth Information
Lent:
I think if either we had Bob's chain switching, or molecules switching each
other, that would be a big step.
Another big step, next to that, is integrated detection. It's really very for-
tunate that there are scanning probes, or it would be hard to see how anything
would be possible at this stage. Scanning probes are a big help; they are a sort of
scannable electrometer. That's not exactly right, but they let one see very small
structures and see changes in electronics. Chemists learn an empirical base of
knowledge that allows them to synthesize structures, and they have a battery
of tools to try to figure out what it is they made, spectroscopically, in solution.
And the gold standard, without which we would have made hardly any progress,
is X-ray crystallography, which allows you can see something of what you made
in a crystalline structure. But it's much more dicult to tell the
charge state
of
what you've made, or to detect something within the molecule that has changed.
X-ray crystallography relies upon really indirect kinds of arguments, so scanning
probes have been enormously helpful as a way to see. Without that, it would be
very hard. We'd like that integratable, and we'd like to be able to see more.
But the short answer is probably device-device interactions.
Wo lkow:
So we have a whole list of things. One of the challenges - but advan-
tages - of trying to start a company is that you must be organized and you must
have a list of marching orders and things that you would deliver against money
received. We're looking at the I/O problem. The “I” and the “O” are quite sep-
arate, the atom fabrication has to be better and better. We also have to address
really dull-sounding things like laser cleaning processes to make more perfect
surfaces because of defects, which we have for two reasons: (1) because we make
things incorrectly, and (2) because there was a preexisting defect that prevents
us from ever making things correctly. We need new algorithms. If, for example,
this room was my substrate that I was going to build on, where the table tops
are ideal regions and the regions in between were not ideal, I would have an
algorithm that preselects the ideal regions and then knits those together. We are
working on all of these things in parallel. So I wouldn't pick one technology, I
would say there's a whole bunch.
But, echoing the earlier conversation, we have to be, hopefully, clever enough
to deliver something impressive sooner than later. We have to stop the tendency
to be researchers and wild explorers only, and really constrain ourselves to solve
some of these kinds of dull problems so that we make tangible advances soon.
Anderson:
Wolfgang, do you have a favorite problem that you'd love to solve
tomorrow?
Porod:
There are two sets of issues, in different directions. One issue is exactly
along the lines that Bob mentioned: at some point, you have to demonstrate
something. So I think it would be great if we were able to demonstrate an
integrated system with electronic input and output and magnetic “something”
inside. As far as I can see, there are no scientific barriers to doing that. It's just
a hard thing to do. Industry creates large teams to work on things like that, and
they make it happen. Like MRAM, for example. Something like that is very hard
Search WWH ::
Custom Search