Go back to
Project
Sample case-study
Title
Resonant beam pressure sensor
Principal source
"Resonant beam pressure sensor fabricated with silicon fusion bonding,"
K. Petersen, et al., Proceedings of Transducers 91, San
FRancisco, CA, June 24-27, 1991, pp. 177-180.
Purpose of the device
It is a pressure sensor. Its specific purpose is to measure pressure with
an accuracy of 0.01% of absolute pressure. The applications are air
pressures in aerospace systems and in engine control systems.
Principle of operation
It consists of a resonant beam on top of which piezo resistors (red) are
formed. Underneath the beam, there is a diffused electrode (orange) that
actuates the beam in resonance. There is also a pit underneath the beam
as shown. Below the pit is a square disphragm. When the pressure is
applied from bottom throught the pressure access hole, the diaphragm and
the portion of the device above it experience change in tension. This
changes the resonant frequency of the beam. This is detected via the
piezo resistors. The change in the resonance frequency is correlated to
the pressure under the diaphragm.
The beam is kept in resonance by a simple analog maintainer circuit while
the change in resonance frequency is monitored by a microprocessor. These
are outside the packaged sensor.
Microfabrication
-
This process begins with an n-type silicon wafer.
-
A shallow pit is etched as shown. One can see this as a rectangle in the
top view.
-
p-type diffusion is performed to define electrodes in the shallow pit.
These electrodes actuate the resonant beam that will be formed later.
Diffused electrodes are connected to the top of the wafer as shown so
that electrical connection can be easily made to them.
-
A second wafer is bonded and then ground to 6 microns. This allows for
"released" mechanical structures in the portion of the top wafer above
the shallow pit in the first wafer. This is where a resonant beam will be
formed later.
-
A passivation layer (blue) is formed for electrical isolation. Piezo
resistors (red) are created by ion implantation.
-
Contact holes are etched through the passivation layer and top silicon
wafer to enable electrical contact with the diffused electrodes.
-
Bond pads and electrical interconnections are formed using metallization.
Not see in the figure are electrical connection made to the piezo
resistors. These will be seen another cross-section as they go
perpendicular to the plane of this image.
-
The bottom wafer is then etched back from below to leave a square
diaphragm. The square shape an can be seen if we view from top.
-
Two slots are etched in the top wafer to form the resonant beam. The
piezoresistors are on top of this including the metallic interconnects.
This is a top-view of the relevant portion of the wafer. The actuating
electrode underneath the beam can also be seen.
-
A glass wafer with a pressure access hole is bonded to the bottom.
Modeling challenges
This device involves five energy domains.
- Elasto-static and elasto-dynamic (diaphragm deflection and beam
vibration)
- Electrostatic (beam actuation)
- Electrical (power supply to actuating electrodes and resistance
measurement in piezoresistors
- Fluidic (squeezed film effects under the beam)
- Thermal (as a secondary effect in piezoresistors)
These energy domains are also coupled in many different ways.
Electrostatic force and deflection of the beam are coupled. The
mechanical strain in the beam is coupled to the resistance change in the
piezoresistors, which is the basis for sensing mechanism in this device.
The squeezed film effects influence the natural frequency of the beam.
The current passed through the piezoresistors may have a second order
effect on their resistance and thus interfering with the signal being
measured. Of course, the external pressure influences the tension in the
beam and its natural frequency, which is the basis for the device operation.
Modeling tasks
- Modeling the piezoresistor behavior as a lumped model
- Dynamic modeling of the beam to gets its requency response. This can
be done first as a lumped 1 dof model and later as a distributed beam
model. In both cases, squeezed film effects and external pressure effects
are to be included.
-
Modeling the diaphragm deformation to see how the pressure on it changes
the natural frequency of the beam. 3-D finite element analysis could be
used to do this.
- Secondary influence of temperature raise on the mechanical behavior
(due to thermal loads) and electrical resistance can be studied. Again
both as a lumped model as well as distributed model.
- System level model of the whole device so that the influence of the
ambient pressure and temperature can be studied.
Required background and governing equations
The following aspects need to be understood in order to model the device.
- Piezeoresistive effect
- 3-D elastic deformation
- Simulation of dynamic behavior of beams
- Electrostatic forces on elastic beams
- Squeezed film effects
- Change in electrical resistance due to temperature raise
Other references
- Kovacs, G.A., "Resonant pressure sensors," Micromachined
Transduers, WCB-McGraw-Hill, 1998, Boston, pp. 260-261.
Model analysis using the finite
element modeling
In order to compute the first natural frequency of the resonant
beam, a solid model of the device was built in I-DEAS and its
model analysis was performed. The solid model was constructed in
accordance with the dimensions given in the principal reference for this
case-study. The following material properties were assumed.
- Young's modulus = 160 GPA
- Poisson's ratio = 0.29
- Density = 2300 kg/m^3
Only a quarter model was used by taking advantage of symmetry.
The reuslts are as follows.
Quarter model of the device
Close-up of the quarter model of the device
Meshed model. Tetrahedron elements were
used. Fine mesh was used in the beam portion.
First mode . Frequency was 280 kHz
Second mode . Frequency was 1.25 MHz. It
can be seen that the diaphragm also participates in this mode.
Go back to
Project