Question 1: Tell us about yourself. What is your background,
and where were you educated?

I am one of the rare Silicon Valley natives. Second
generation computer engineer, learned to program when
I was 10. BS/Comp. Sci. from UCSB, MS/Comp. Sci. from
UCLA. But computers aren't the only thing I know.

My family has a tradition of researching anything of
concern or interest to us - in any field. For
example: my mother, to quell her fears of the unknown
as she grows old, is currently getting a Master's in
gerontology. Likewise, I studied nanotechnology and
quantum physics in my spare time, to see what
practical near-term applications could come out of
these fields that have been getting so much press in
recent years. I am careful not to style myself as an
expert just because I've studied a little bit; I
acknowledge there are things I do not know - so that I
can learn them so they do not remain unknown. For
example, the experiment this article is about is first
and foremost an experiment: its primary objective is
to further my own knowledge and maybe others'. It so
happens that we might get a new power source as well,
but that's not guaranteed.

Question 2: What is the casimir effect?

A result of the fact that virtual particles -
short-lived pairs of particles and antiparticles - are
constantly appearing and then annihilating each other.
In most cases this is of little consequence to our
macroscopic world. Only in extreme cases does it give
rise to measurable effects, like the slow evaporation
of black holes, or if you create an extremely confined
space.

Back in 1948, physicist Hendrik B. G. Casimir
predicted that if you place two electrically neutral,
perfectly parallel sheets of metal very close to each
other - less than a micron, preferably less than a
hundred nanometers - the longer wavelengths of virtual
particle pairs would not be able to appear between the
sheets. But they would continue to appear outside the
sheets, thus producing a pressure that would force the
sheets together. Technically, this could be said to
exist even if the sheets are more than a micron apart,
but the force only becomes significant with small
separations (thus more wavelengths excluded, thus
higher pressure).

It appears to have taken until the 1990s to actually
prove this theory, though, mainly for want of
instruments sensitive enough and machining capability
with the necessary tolerances. But this has now been
demonstrated by multiple experiments in different
labs, and appears to be one of the more generally
agreed-upon facets of quantum mechanics.

So, of course, many people have been trying to tap it
for energy - but in the classical formulation of
parallel metal plates, the Casimir force is
conservative: you have to put as much energy in to
separate the metal plates as you got out from them
coming together. There have been useful experiments
in using this to moderate some other energy source, or
in using this to store energy, but to my knowledge, no
one has been able to successfully demonstrate actual
energy production from the Casimir force alone. (Many
people claim to have done this, but every one I've
seen shows all the signs of fraud, especially never
actually managing to show the claimed success to
independent outside observers. They could be
deliberate frauds, or they could have deceived
themselves into believing, but the fact remains that
they don't apparently actually have what they say they
have. I might not have succeeded yet, but then I
don't yet claim that I have - or promise that I will
when I don't yet know for sure.)

Question 3: Describe your vision for exploiting the casimir
effect for energy production. How did you conceive of this concept?

I conceived of it by studying the failures, and why
they had failed. At first I was thinking about
parallel plates too, trying to come up with a way
where you wouldn't have to put energy in to get the
plates apart. One of the ideas I considered was
linking a series of motors like cylinders in an
internal combustion engine: one unit would be pulling
its plates together, which force would power the other
unit to pull the plates apart. But, of course, that
would not work either. I wondered if there was a way
to insert something to soak up the Casimir force while
the plates were being pulled apart.

That was when I stumbled across my current concept: a
metal ring, rotating around a metal core, with shields
of a different material to cancel or at least lessen
the Casimir force in one direction (clockwise or
counterclockwise). The result: a small, but nonzero,
torque that would cause the ring to spin. There are
a number of ways to tap this mechanical energy, if it
in fact would exist; the simplest and most efficient
seems to be to place it in a magnetic field and place
wires to draw off the resulting electric current.

I have attached a diagram, where the ring (the gray
circle on the outside) rotates clockwise. The Casimir
force at point B can be broken into two components, as
illustrated: most of it tries to pull the point
inwards (and is resisted by the structure of the ring,
and thus can be ignored), but a small portion pulls
the point clockwise. There is no counterclockwise
force to balance this: point A, and the part of the
circle immediately clockwise from it, is shielded by
the non-metal material (designated by black; I haven't
settled on what material to use here, but ideally it
should be some kind of blackbody). The shields are
fixed in place relative to the square metal core, but
again, the ring is free to rotate around the rest of
the device. (Leaving the ring free proved to be one
of the most difficult requirements when figuring out
how to manufacture this.)

Note especially that there is no external energy input
save for the Casimir force itself. This is
deliberate, so I can not delude myself into thinking
I'm getting more energy out than I put in when I'm
actually just getting a measurement error (which has
derailed quite a lot of proposed alternate energy
sources; an equivalent for reactions that do require
energy would be powering the reactions from their own
output). If I get anything out, it will be in excess
of zero.

While there is an equation stating the magnitude of
the Casimir force for parallel plates, similar
equations for other cases (like this) are hard to
find. There is even some speculation that the Casimir
force would actually be repulsive instead of
attractive in this situation, in which case the arrows
on the diagram would reverse. But all of the studies
I have seen thus far indicate that, whatever the
actual force is, it should at least be larger than
static friction in my current experimental setup, so I
should get some rotation.

For the record: I registered appropriate IP
protections for this long before this interview. Not
that I'm that afraid of IP theft - to successfully
implement this, there are a few quantum mechanics
details one would need to know that I haven't
mentioned here, and then of course you need enough
practical knowledge of nanotech to be able to actually
build the thing. For instance, the diagram here is
for an idealized version if you could manipulate
things down to the atom level; in practice, the
diagrams I'm actually feeding to the equipment I'm
working with are blocky approximations (which blocks
could affect the experiment).

Question 4: Approximately how efficiently would your casmir
machine operate? What proportion of the energy gained from
harnessing casimir torque could be harnessed for useable energy?

To be honest, I haven't looked very far into this.
I'm just seeing whether I can tap any energy there at
all. That said, if it does work, the best initial
gains will probably come more from increasing the raw
output than by increasing efficiency. The output
energy per unit volume goes up roughly with the fifth
power of the feature size - that is, if you could make
the rings half as big (in each of X and Y - changes to
Z largely cancel out in my current configuration), the
output energy per unit volume would go up by a factor
of 32. This is a combination of the Casimir effect's
equations and the fact that smaller feature size, and
thus smaller rings, means you can have more rings in
the same space.

That is all theoretical, and again, I have not
thoroughly studied that part, so there could be
factors I have not yet accounted for. I've been
focusing on whether it works at all: if it does not,
then efficiency and so forth becomes moot.

Question 5: Have you come across any technical arguments
which couldpotentially invalidate your casimir energy machine?
How many individuals have examined your proposals for technical accuracy?

I've lost count of the number of quantum mechanics
academics I've asked to review my idea. The response
from every one of them has been essentially the same:
while they do not see why this would not work, they do
not wish to endorse it since all other proposals
they've seen to tap the Casimir effect have run afoul
of its conservative nature. They acknowledge my
explanation of why this looks like it won't*, but
they'd still prefer a demonstration before endorsing
it. Which is fine by me: I'm not going to claim it
will work for sure either, unless and until I actually
get it to work.

* See question 3. More technically, the other
proposals ran afoul of the Second Law of
Thermodynamics, but this would be an open system with,
in theory, continual energy input from the virtual
particles. In practice, there has never been a
device that could affect the energy that powers the
Casimir force, so we don't know how said energy would
flow from one point to another. It is certainly
possible that the ramifications of that would reduce
this to a scientific curiosity (useless as a power
generator) even if it does work, though I would still
count that as a success.

In short, the expert opinion I've received is, "We
don't know if this will or will not work."

Question 6: Would a portable casimir machine be feasible,
or even possible?
How small could a casimir machine be made?
Could this technology ever be used to power portable electronics?

As a matter of fact, the experiments I'm doing are
for units only about a micron or so across. Billions
would fit on a standard 4 inch diameter silicon wafer.
So, technically, this would be very portable, and the
non-portable versions would just be very large arrays
of the portable versions.

The flip side of that is that each individual unit
produces an extremely tiny amount of power even under
the most optimistic theoretical calculations. I do
not yet know exactly how much that would be, and I
won't know until I find out the rotational speed et al
(on which there is theoretical disagreement, so this
is one of the things my experiments are intended to
find out - immediately after determining whether there
is any rotation in the first place). So while you
could certainly have a generator small enough to
carry, it is not yet known whether it would produce
enough power to be worth anything (again, assuming
this works at all). Even if everything goes very
well, larger arrays (like for power plants, or at
least home cogeneration facilities) would probably be
more feasible at first than portable ones, although
one application I would like to aim for is using this
in place of batteries for electric vehicles.

Question 7: It appears that virtually all of the expense resulting
from generating energy from the casimir effect would come from
materials and component fabrication costs. Have you made any
preliminary analysis regarding the price per kilowatt of electricity
derived from casimir torque?

No. Any analysis I tried to make at this stage would
be misleading at best, a result of both uncertainty in
how much power would be generated and in having not
yet honestly explored the mechanisms for mass
manufacture. At this level of uncertainty, I'd rather
not provide made up data for people to pin false hopes
on. The only thing for certain is that, if this does
work, mass production techniques would be the best way
to go for even the smallest craft shop production
facility. Creating a few units by hand to try various
configurations, I can do by hand for a reasonable
budget. Industrial fabrication of enough units to
produce significant power is another thing entirely.

Question 8: Your technique for component fabrication relies
extensively on sophisticated 3-dimensional lithographic etching.
But lithography equipment is extremely costly. Can your
energy-generating machines be made with less expensive methods?

Actually, I'm using 2-dimensional etching with
precisely controlled processing to create the
3-dimensional effect, mainly because I do not have
access to 3-dimensional direct fabrication methods
with the necessary resolution. I have thought of some
relatively cheap ways to manufacture them in quantity,
once I have identified which (if any) configuration
produces the most power, but these setups do not lend
themselves well to the kinds of experiments I need to
do to figure out which (if any) configuration works
best. Again, though, development of that is more
appropriate for a later stage, after I have determined
whether this works at all.


Question 9: Have you had difficulty getting funding for your
company, Winged Cat Solutions? Do you have any angel investors
or corporate financial support?


Winged Cat Solutions is funded out of my own pocket.
You'd be surprised how little funding you truly need
if you go to the right places. For example, the
Stanford Nanofabrication Facility has agreed to allow
me to use their equipment for my experiments, for at
most a few thousand dollars per month (the exact
amount depending on how much I actually do; I've had
months where I've spent less than a hundred dollars).
The same offer is open to anyone with similar
experiments to run; my operation is smaller than they're
used to dealing with, but it certainly fitswithin their scope.
(Of course, you have to know what
you're doing, but they offer - and require - safety
and equipment training for anyone who wishes to run
experiments using their facilities.) A number of
other facilities are coordinating with Stanford to
offer similar services elsewhere in the United States,
under the banner of the National Nanotechnology
Infrastructure Network.

Once my experiments are complete and I'm ready to go
into production, that will be a different story. SNF
and the NNIN are for research and development only,
emphasis on research. Then again, if and when I can
demonstrate that this works (necessary before I could
possibly be ready for production), I doubt I will have
much difficulty raising funding at that time.

SNF has asked that I add their formal acknowledgement
statement: "Work was performed in part at the Stanford
Nanofabrication Facility (a member of the National
Nanotechnology Infrastructure Network) which is
supported by the National Science Foundation under
Grant ECS-9731293, its lab members, and the industrial
members of the Stanford Center for Integrated
Systems."

Question 10: How great is the potential for generating energy
from thecasimir effect? Could machines based on casimir torque
obviate the need for fossil fuels or nuclear energy?

If it works, yes. The usage model would be much like
solar energy, except not affected by weather or
day/night or needing to be outdoors. (For the
record, most of my home's power needs are provided by
solar power - and that's only "most" because I have a
lot of computer servers, and thus use enough power to
qualify for industrial rates. Or rather, I did before
the panels were installed. Suffice it to say that
they've already paid for themselves, and it's been
less than 5 years since installation.)

Question 11: Your arguments for casimir energy generation
are stilltheoretical. Have you done any computer simulations
that would indicate that your machines should work?

Yes. But those are only as good as the data,
including assumptions, that one feeds into them. If
this doesn't work in practice, I'm pretty sure I know
which one of the assumptions I used will turn out to
have been wrong.

Question 12: What are your plans for the next decade?

Finish the experiment to see if this works. Then,
based on the results, make plans. ;)

Seriously, though, I don't tend to plan that far out.
My life tends to be too chaotic to reliably plan more
than several months in advance, on average. But it
shouldn't take a decade to see whether or not this
device works. If it weren't for Murphy and some
equipment failures, I'd already have the initial
experiments done by now. As it is, said experiments
are currently planned for sometime around October.
Even if it fails, I might write an article about that,
in case anything of what I've done is of use to
others.

 

Copyright © 2004 Sander Olson