www.mrs.org/publications/bulletin
bond pads, and formation of MEMS-to-
electronics interconnects.
Sandia National Laboratories has demon-
strated a modular poly-Si integrated MEMS
process that solves the topography prob-
lem by first etching a trench in the Si wafer
and then fabricating the MEMS device in
I
ntegrated MEMS
1
this trench. The trench is then filled with
Technologies
additional sacrificial-oxide depositions,
planarized by chemical–mechanical pol-
ishing (CMP), and sealed with a nitride
deposition. The electronics are then fabri-
cated on the pre-processed wafer using
a standard CMOS process. Contacts are
opened to poly-Si studs at the perimeter of
the trench prior to aluminum deposition
to form the MEMS–electronics interconnec-
tions. The electronics are then passivated
by an HF-resistant material for protection
during the sacrificial-oxide etching step
that releases the microstructures.
Andrea E. Franke, Tsu-Jae King,
and Roger T. Howe
Introduction
While microelectromechanical systems
MEMS) technology has made a substan-
Several issues must be considered in de-
veloping an integrated MEMS process. The
first is the process-temperature constraint
imposed by Al- or Cu-based integrated-
circuit (IC) metallization. Depending on
the particular interconnect metallurgy
used, the maximum post-metallization
temperature is in the range of 400–550ꢀC.
Asecond issue is the relatively thick MEMS
structural and sacrificial layers (2–10 ꢁm
thick), which create significant height
variations on the wafer surface. High-
resolution lithography, which is needed
for submicron CMOS (complementary
metal oxide semiconductor) transistors is
not possible on such rough surfaces. Elec-
trical interconnects between the MEMS
and electronics are also an important con-
sideration, since they should be of a
minimum length and made in a highly
conductive layer to minimize parasitic re-
sistance and capacitance. Finally, the pro-
tection of on-chip electronics from the
MEMS release etchant may be necessary.
For example, hydrofluoric acid (HF) is
typically used to remove sacrificial oxide
layers to release poly-Si microstructures.
Since the oxide layers and interconnect
metals employed in the electronic devices
are also etched in HF, they must be pro-
tected during the release etch.
(
tial impact over the past decade at the
device or component level, it has yet to
realize the “S” in its acronym, as complex
microsystems consisting of sensors and ac-
tuators integrated with sense, control, and
signal-processing electronics are still be-
yond the current state of the art. There are
several incentives to co-fabricate MEMS
devices and electronics on a single silicon
chip, which apply to applications such as
inertial sensors. The parasitic resistance
and capacitance associated with the inter-
connects between the MEMS devices and
electronics degrade electrical signal quality
and hence degrade system performance.
This problem is critical for thin-film, surface-
micromachined devices in which changes
in position are sensed as a small, capaci-
tively induced current. If the MEMS de-
vices can be co-fabricated with electronics
on a single chip, the parasitic resistance
and capacitance of these components can
be greatly reduced. The elimination of chip-
to-chip interconnections can also lead to
more reliable packages. This is especially
important for large arrays of MEMS de-
vices that require independent addressing
of each element, such as the micromirrors
in an optical cross-connect switch, for which
the monolithic integration of switching
electronics can reduce the number of off-
chip connections by orders of magnitude.
Depending on the integration strategy
that is adopted, the co-fabrication of mi-
crostructures and electronics on the same
substrate may provide significant overall
cost savings.
Another MEMS-first integration strategy
involves a different approach to planarize
the wafer after the deposition and pattern-
ing of the poly-Si MEMS layers. Rather
than etching a recessed well, the MEMS
film stack is formed on the wafer surface,
and then a layer is selectively grown by
epitaxy on the surrounding area to recover
a planar surface for electronics fabrica-
2
tion. Figure 1 is a scanning electron mi-
crograph of a 6-ꢁm-thick poly-Si lateral
accelerometer fabricated using this ap-
proach. The surface interconnects in Fig-
ure 1 are aluminum lines, which contact
poly-Si vertical feedthroughs at the pe-
rimeter of the MEMS area.
Interleaved MEMS/Electronics
Processes
By interleaving the MEMS and electronics
processing steps, the process designer has
maximum flexibility. With the exception
of the metallization, electronic device struc-
tures can tolerate temperatures as high as
900ꢀC without significant impact on their
characteristics. An attractive strategy for
poly-Si MEMS, therefore, is to postpone
the metallization steps until after the depo-
sition and annealing of the MEMS layers.
Interconnections between the electronics
and MEMS are relatively straightforward
in an interleaved process. However, the
process must allow for the topography of
the MEMS layers, which may require pla-
narization or relaxed layout design rules
for the metallization layers. In addition,
the electronics layers must be protected
during the release etch.
“MEMS-First”Modular Processes
By depositing, annealing, and pattern-
ing the MEMS layers before fabricating
the electronics, the thermal-budget con-
straint imposed by metallization is elimi-
nated for the MEMS process. This approach
is attractive for a MEMS technology that
requires high-temperature anneals for
residual-stress annealing or dopant acti-
vation, such as surface-micromachined
poly-Si technology. Integration issues
that must still be addressed are the topog-
raphy of the MEMS structures, passi-
vation of the electronics during the
release etch while allowing access to
This article surveys three approaches
to integrating MEMS with electronics by
means of co-fabrication in a single process
sequence: MEMS process steps completed
before the electronics process steps, MEMS
process steps interleaved with the electron-
ics process steps, and MEMS process steps
done after the electronics process steps.
In several cases, the initial demonstration
of an integrated MEMS technology has
used the flexible, interleaved approach.
Metal microstructures were fabricated as
gates of field-effect transistors in pioneer-
ing work at Westinghouse in the 1960s.3
MRS BULLETIN/APRIL 2001
291