Dimensional Ranges

• 1 m < L < 300 m lateral dimensions

Surface micromachined structures … “classic MEMS”

• 300 m < L < 3 mm

Bulk silicon/wafer bonded structures … still call them MEMS
and cover them in this course

• 10 nm < L < 1 m
Nano electromechanical systems … NEMS
(overlap with MEMS … some coverage in this course)

EE C245 – ME C218 Fall 2003 Lecture 1

9
 What aren’t MEMS

It runs!

Cost?
• The Denso micro-car: circa 1991
http://www.globaldenso.com/ABOUT/history/ep_91.html
• Fabrication process: micro electro-discharge machining

EE C245 – ME C218 Fall 2003 Lecture 1

10
 Experimental Catheter-type Micromachine for
Repair in Narrow Complex Areas
Repairing Vision device
manipulator
Multi freedom bending tube

Welding device Monitoring device Posture Detecting Device

Japanese Micromachine Project 1991-2000

EE C245 – ME C218 Fall 2003 Lecture 1
11
 Batch Fabrication Technology
• Planar integrated circuit technology 1958 -
1. Thin-film deposition and etching
2. Modification of the top few m of the substrate
3. Lateral dimensions defined by photolithography, a process derived from offset printing

• Result: CMOS integrated circuits became the ultimate “enabling technology” by circa 1980

• Moore’s Law
Density (and performance, broadly defined) of digital integrated circuits increases by a factor of two every
year.

EE C245 – ME C218 Fall 2003 Lecture 1

12
 A Microfabricated Inertial Sensor
MEMSIC
(Andover, Mass.)

Two-axis thermal-bubble
accelerometer

Technology: standard
CMOS electronics with
post processing to form
thermally isolated sensor
structures

Note: I’m a technical advisor to MEMSIC

a spinoff from Analog Devices.
EE C245 – ME C218 Fall 2003 Lecture 1
14
 Other Batch Fabrication Processes
• Historically, there aren’t that many examples outside
of chemical processes

• However, that’s changing:

Soft (rubber-stamp) lithography

Parallel assembly processes 
enable low-cost fabrication of MEMS from micro/
nano components made using other batch
processes … “heterogeneous integration”

EE C245 – ME C218 Fall 2003 Lecture 1

15
 Microassembly Processes
Parallel Pick-and-Place
Parallel assembly processes promise
inexpensive, high-volume hetero-
geneous integration of MEMS, CMOS,
and photonics

www.memspi.com, Chris Keller, Ph.D. MSE 1998

Fluidic Self-assembly
Wafer-Level
Batch Many challenges:
Assembly > interconnect
> glue

www.microassembly.com
Michael Cohn, Ph.D. EECS, 1997

EE C245 – ME C218 Fall 2003 Lecture 1 Uthara Srinivasan, Ph.D., Chem.Eng. 2001
16
 A Brief History of MEMS:
1. Feynmann’s Vision
• Richard Feynmann, Caltech (Nobel Prize, Physics, 1965)
American Physical Society Meeting, December 29, 1959:
“What I want to talk about is the problem of manipulating and
controlling things on a small scale. …. In the year 2000, when they
look back at this age, they will wonder why it was not until the year
1960 that anybody began seriously to move in this direction.”

“… And I want to offer another prize -- … $1,000 to the first guy

who makes an operating electric motor---a rotating electric motor
which can be controlled from the outside and, not counting the lead-
in wires, is only 1/64 inch cube.”

… he had to pay the electric motor prize only a year later

• http://www.zyvex.com/nanotech/feynman.html

EE C245 – ME C218 Fall 2003 Lecture 1

17
 2. Planar IC Technology
• 1958 Robert Noyce – Fairchild and Jack Kilby (Nobel Prize, Physics,
2000) -Texas Instruments invent the integrated circuit
• By the early 1960s, it was generally recognized that this was the way
to make electronics small … and cheaper
Harvey Nathanson
and William Newell,
surface-micromachined
resonant gate transistor,
Westinghouse, 1965

Did Harvey hear about

Richard Feynman’s talk in
1959? I don’t think so …

EE C245 – ME C218 Fall 2003 Lecture 1

18
 Why Didn’t MEMS Take Off in 1965?
• Resonant gate transistor was a poor on-chip
frequency reference  metals have a high
temperature sensitivity and don’t have a sharp
resonance (low-Q) … specific application didn’t “fly”
• In 1968, Robert Newcomb (Stanford, now Maryland)
proposed and attempted to fabricate a surface
micromachined electromagnetic motor after seeing
the Westinghouse work
Energy density scaling for this type of motor indicated
performance degradation as dimensions were reduced …
Materials incompatibility with Stanford’s Microelectronics Lab
research focus on electronic devices became a major issue

EE C245 – ME C218 Fall 2003 Lecture 1

19
 Another Historical Current:
Silicon Substrate (Bulk) Micromachining
• 1950s: silicon anisotropic etchants (e.g., KOH)
discovered at Bell Labs
• Late 1960s: Honeywell and Philips commercialize
piezoresistive pressure sensor utilizing a silicon
membrane formed by anisotropic etching
• 1960s-70s: research at Stanford on implanted silicon
pressure sensors (Jim Meindl), neural probes, and a
wafer-scale gas chromatograph (both Jim Angell)
• 1980s: Kurt Petersen of IBM and ex-Stanford students
Henry Allen, Jim Knutti, Steve Terry help initiate Silicon
Valley “silicon microsensor and microstructures”
industry
• 1990s: silicon ink-jet print heads become a commodity
EE C245 – ME C218 Fall 2003 Lecture 1
20
 When the Time is Right …
• Early 1980s: Berkeley and Wisconsin demonstrate
polysilicon structural layers and oxide sacrificial
layers … rebirth of surface micromachining
• 1984: integration of polysilicon microstructures with
NMOS electronics
• 1987: Berkeley and Bell Labs demonstrate
polysilicon surface micromechanisms; MEMS
becomes the name in U.S.; Analog Devices begins
accelerometer project
• 1988: Berkeley demonstrates electrostatic
micromotor, stimulating major interest in Europe,
Japan, and U.S.; Berkeley demonstrates the
electrostatic comb drive

EE C245 – ME C218 Fall 2003 Lecture 1

21
 Polysilicon Microstructures
• UC Berkeley 1981-82

R. T. Howe and
R. S. Muller,
ECS Spring Mtg.,
May 1982

EE C245 – ME C218 Fall 2003 Lecture 1

22
 Polysilicon MEMS + NMOS Integration
• UC Berkeley 1983-1984
Transresistance
amplifier

Capacitively driven
and sensed 150 m-long
polysilicon microbridge

R. T. Howe and
R. S. Muller,
IEEE IEDM,
San Francisco,
December 1984

EE C245 – ME C218 Fall 2003 Lecture 1

23
 Polysilicon Electrostatic Micromotor

Self-aligned pin-joint, made

possible by conformal deposition
of structural and sacrificial layers

Prof. Mehran Mehregany,

Case Western Reserve Univ.

EE C245 – ME C218 Fall 2003 Lecture 1

24
 Electrostatic Comb-Drive Resonators
• W. C. Tang and R. T. Howe, BSAC 1987-1988
New idea: structures move laterally to surface

C. Nguyen and
R. T. Howe,
IEEE IEDM,
Washington, D.C.,
December 1993

EE C245 – ME C218 Fall 2003 Lecture 1

25
 Analog Devices Accelerometers
• Integration with BiMOS linear technology
• Lateral structures with interdigitated parallel-plate
sense/feedback capacitors

ADXL-05 (1995)

Courtesy of Kevin Chau,

Micromachined Products
Division, Cambridge

EE C245 – ME C218 Fall 2003 Lecture 1

26
 Surface Micromachining Foundries
1. MCNC MUMPS technology (imported from Berkeley) 1992-
2. Sandia SUMMiT-IV and -V technologies: 1998 –
4 and 5 poly-Si level processes

result: more universities, companies do MEMS

M. S. Rodgers
and J. Sniegowski,
Transducers 99

(Sandia Natl. Labs)

EE C245 – ME C218 Fall 2003 Lecture 1

27
 Self-Assembly Processes
Alien Technologies, Gilroy, Calif.
chemically micromachined
“nanoblock” silicon CMOS
chiplets fall into minimum energy
sites on substrate

nanoblocks being fluidically

self-assembed into embossed
micro-pockets in plastic antenna
substrate

Prof. J. Stephen Smith, UC Berkeley EECS Dept.

EE C245 – ME C218 Fall 2003 Lecture 1
28
 More Recent History
• Mechanical engineers move into MEMS, starting with
Al Pisano in 1987 … expand applications and
technology beyond EE’s chip-centric view
• DARPA supports large projects at many US
universities and labs (1994 – 200?) with a series of
outstanding program managers (K. Gabriel, A. P.
Pisano, W. C. Tang, C. T.-C. Nguyen, J. Evans)
• Commercialization of inertial sensors (Analog
Devices and Motorola polysilicon accelerometers
1991 → ) ≈ 108 by each company by 2002
• Microfluidics starts with capillary electrophoresis circa
1990; micro-total analysis system (-TAS) vision for
diagnosis, sensing, and synthesis
• Optical MEMS boom and bust: 1998 – 2002.

EE C245 – ME C218 Fall 2003 Lecture 1

29
 MEMS and Nanotechnology I
• Richard Feynmann’s 1959 talk:
“But it is interesting that it would be, in principle, possible
(I think) for a physicist to synthesize any chemical substance
that the chemist writes down. Give the orders and the
physicist synthesizes it. How? Put the atoms down where
the chemist says, and so you make the substance.”
• Eric Drexler, 1980s: visionary promoting a molecular
engineering technology based on “assemblers” …
had paper at first MEMS workshop in 1987
• Early 1990s: U.S. MEMS community concerned that
“far-out” nanotech would be confused with our field,
undermining credibility with industry and government

EE C245 – ME C218 Fall 2003 Lecture 1

30
 MEMS and Nanotechnology II
• Buckyballs, carbon nanotubes, nanowires, quantum
dots, molecular motors, … together with the atomic-
force microscope (AFM) as an experimental tool 
 Synthetic and “top-down” nanotechnology earns respect of
MEMS community
• Why is nanotechnology interesting?
 Molecular control of sensing interface (chemical detection)
 Synthetic processes promise to create new batch-fabrication
technologies
 Planar lithography is reaching into the nano regime (state-of-
the are is 50 nm line/space; spacer lithography has reached
7 nm)
 New computational devices: neural, quantum computing

EE C245 – ME C218 Fall 2003 Lecture 1

31
 1 GHz NEMS Resonator
Si double-ended tuning fork
tine width = 35nm
SOI length = 500 nm
Drive thickness = 50 nm
electrode
resonator
Interconnect parasitic
SOI
elements are critical 
need nearby electronics
Sense
electrode
Uses vertical channel
FINFET
process on SOI substrate
L. Chang, S. Bhave, T.-J. King, and R. T. Howe
UC Berkeley (unpublished)

EE C245 – ME C218 Fall 2003 Lecture 1

32
 MEMS (NEMS?) Memory: IBM’s Millipede
Array of AFM tips write and read bits:
potential for low and adaptive power

EE C245 – ME C218 Fall 2003 Lecture 1

33
 Electrostatic NEMS Motor
Alex Zettl, UC Berkeley, Physics Dept., July 2003

multi-walled carbon nanotube

rotary sleeve bearing

500 nm

EE C245 – ME C218 Fall 2003 Lecture 1

34
 Nanogap DNA Junctions
• Development of ultrafast and ultrasensitive dielectric DNA detection
• Applications to functional genomics or proteomics chips, as well as an
exploration of nanogap DNA junction-based information storage and
retrieval devices

Nanogap Inlet Electrodes Nanogap Electrodes

Junction arrays
Nanogap
Junction
Poly Si (I)

Poly Si (II)

Nanofluidic
Network for Insulator
Nanogap
DNA Trapping Outlet DNA Si3N4
(5 to 50 nm)

Luke P. Lee and Dorian Liepmann, BioEng.

Jeff Bokor, EECS
EE C245 – ME C218 Fall 2003 Lecture 1
36
 Opportunities in Blurring
the MEMS/NEMS Boundary
• Aggressive exploitation of extensions of “top-down”
planar lithographic processes
• Synthetic techniques create new materials and
structures (nanowires, CNT bearings)
• Self-assembly concepts will play a large role in
combining the top-down and bottom-up technologies
• Application: mainstream information technology with
power consumption being the driver
“Beyond CMOS” … really, extensions to CMOS > 2015
Non-volatile memories
Communications

EE C245 – ME C218 Fall 2003 Lecture 1

38