Keitaro Yamashita
General Manager
MEMS Department
Semiconductor Technology
Development Group
Semiconductor Business Unit
Sony Corporation |
 |
Dr. Koichi Ikeda
Senior Manager
MEMS Department
Section 2
Semiconductor Technology Development Group
Semiconductor Business Unit
Sony Corporation |
 |
Shun Mitarai
MEMS Department
Section 2
Semiconductor Technology Development Group
Semiconductor Business Unit
Sony Corporation |
 |
|
| . |
. |
Making the Motivation Clear
What has prevented the realization of practical MEMS filters so far?
Yamashita: Until now, research on MEMS
filters was mostly carried out at US universities. The earliest research was concerned
with explicating the principles of operation,
and either extremely high voltages were required, or the researchers were satisfied as
long as the phenomenon was verified, no
matter how small the transmitted signal.
Those works were not applied directly for
consumer electronics products. In contrast,
we had the explicit motivation of developing a device with practical characteristics,
and the filter characteristics that we were targeting were comparatively clear. Thus in our
work it was never a matter of saying “OK, I
understand” and that being the end of it when
we understood some phenomenon, but rather
always a matter of, “OK, what can I do with this now?”, that is, we always had a firm grasp
of the direction we wanted to go, namely of
always explicitly moving towards improvements. The idea of moving from a two-port
to a three-port structure was an example of
this.
Ikeda: Normally, we would have pursued
research and development limited to the two-port structure. It is easily expected that the
shorter mechanical beam shows the smaller
signal attenuation. However, we took the exactly opposite approach.
Mitarai: As for high-frequency region, it
was hard to design high signal-to-noise ratio
devices using two-port resonator.
Yamashita: The beam, which is a mechanical device, operates as a capacitor that
isn’t even needed and transmits signals as
leakage. If we could isolate the input and
output signal lines, the leakage signal could
be reduced relative to the vibrational component. We were not convinced of this before we started, rather, it was something we
found by trial and error.
Ikeda: At a group session, someone drew
the vibrating element on the whiteboard in
the form of a wave. When you draw a picture
of a waveform, you don’t draw a single wave,
but two or three or more. These correspond
to the second order and higher order oscillation modes. The key point here was that
instead of erasing them immediately, we
decided to prototype and measure them.
Yamashita: This process is the fascination of MEMS research and developments.
Even though the devices are on the micro
scale, in principle, they vibrate in ways that
are no different from a guitar string or a plank
that you can see with your eyes. We were able
to come up with new ideas by returning to
the basic concept that more than a semiconductor development issue, this was a vibrating board in the form of a beam clamped at
both ends.
The Significance of Imaging the Motion
The integration of mechanical motion and
electronic circuits also has stimulated a synergy between engineers from different specialties.
Yamashita: Although I was originally an
end-product person, in the MEMS department, we built a team that consisted of engineers from different specialties, including
process engineers, device designers, and
simulation and measurement specialists.
When someone would say something,
everyone else would think “Oh, so that’s how
it is”, and when someone else would interpret something from another viewpoint
propose new idea, everyone else would think
“Oh, right. Of course”. This is actually closer
to the way things happen in end-product design than in semiconductor groups.
Ikeda: Every day we’d seen interactions
such as the following. A device designer
would hover over Mitarai, commenting on his
work. “Hey, if you could hold the resonator
gap height to levels such as this, we’d get
even better characteristics. ”But Mitarai
would respond “That’s unreasonable. But I’ll
try to hold it to this level.”
Mitarai: The simulation software shows us
a result of vibration analysis with an animation (see the top figure at the right), and we’d
watch it and discuss, for example, what was
the most efficient position for the lower electrode would be.
Yamashita: It is extremely important to be
able to visually image the actual operation
of the device. Depending on their main field,
there are people who would search for the
cause of the electrical signal leakage, and
there are people who would view the system
as a pure mechanical system. Even if what
people would see in the system differed with
the person, as long as we had a common image, then we could discuss the system in terms
of that image. This can be said to be the
unique feature of MEMS. Inversely, even for
two systems that are both MEMS, for example, the GLV optical diffraction grating
and this bandpass filter, it goes without saying that the images we need to work with will
be different. For people who take pleasure in
this sort of thing, these are research projects
in which there is value in participating, even
if one’s area of specialization differs.
Ikeda: When I was involved with the GLV
project, I would often keep the image of the
product in my mind while walking around
TV showrooms. Since this filter development
project started, I’ve found home construction
sites and the structure of Tokyo’s Rainbow
Bridge more interesting. Because houses and
bridges and resonators all consist of pillars
and beams. When looking at these projects,
I’d be tempted to say something presumptuous, for example, “Hey, if you changed the
design like this, it would be able to withstand
vibrations better.” With MEMS devices, it’s
easy to see how their function directly links
to the final product.
Asking Once Again What
Semiconductors are Capable of
How will MEMS technology change our concepts of semiconductors, end products, and
creative production (“monozukuri”)?
Mitarai: Well, the first thing I’d like to do
soon for our MEMS devices is to be commercialized.
Ikeda: MEMS (micromachine) technology
already has a long history in consumer applications, for example it is widely used in printer
head nozzles and pressure sensors. Semiconductor technology has contributed some of the
manufacturing techniques used. This technology, that is, the manufacturing and material
technologies used, is now mature, and tools
unique to MEMS, such as sacrificial layer etching, are commercially available. In my opinion, “motion” will be introduced into system
LSIs as a newly emerging function.
Yamashita: I think that micromachines
have finally developed into practical components. MEMS devices have been applied as
discrete components in products such as the
acceleration sensors used in automotive air
bag systems. If the echnology progresses just
one step further, these components that were
previously external parts will come to be included on the LSI itself. Except for the interface with external systems, parts that operate
in places that can’t be seen will rapidly be
integrated onto single chip solutions. This will
increase the flexibility we have in end product design and allow us to provide end users
with products that are even easier to use.
I think that it is a natural progression for semiconductor engineers to be aiming for that.
Mitarai: If I were to say what has changed
after joining MEMS department, one thing
is that I grew to read not only the semiconductor magazines, but the electronics magazines as well. While it is also the case that I
am being asked to do things that I’d like to
do anyway, nowadays I look at a lot of things
and wonder how they could be improved by
using MEMS technology. I am sometimes
surprised at the degree to which I’ve been
caught up in this.
Yamashita: I think that it is important for
semiconductor engineers to imagine how
components that one has been in charge of
are taken advantage of in end products and
in what way they are provided to the end user.
Only being able to see the area that one is in
charge of is inimical to the spirit of creative
production (“monozukuri”), and is not a state
of mind from which revolutionary ideas will
be born. |
|
Sony exhibited the GxL imaging technology
that uses these concepts and technologies
at EXPO 2005, Aichi, Japan. (See CX-NEWS,
Vol. 40.) |
 |
 Embedded MEMS Filter Chip and its Fabrication for VHF Applications
Bandpass Filter Developed Using MEMS Technology |
|
Vol.43
