Definition of Computer Aided Design

CAD Application to MEMS in Microsystems Technology

Technology In MEMS technology, CAD is defined as a
tightly organized set of cooperating computer
Bruce K. Gale, University of Utah
programs that enable the simulation of
manufacturing processes, device operation
Acknowledgment:
and packaged Microsystems behavior in a
Nora Finch, Intellisense
continuous sequence, by a Microsystems
M. Mehregany, Case Western Reserve engineer.
John R. Gilbert, Microcosm Technologies
Abdelkader Tayebi, Louisiana Tech

Commercially Available Software MEMS CAD Motivation

• Coventorware from Coventor

128 Transmitters
– http://www.memcad.com • Match system specifications : MEMS micro-mirrors
:
• IntelliSuite from Intellisense Inc. (Corning) – Optimize device performance

128 Receivers
Collimator Lens Arrays :
– Design package
– http://www.intellisense.com :
– Validate fabrication process
• MEMS ProCAETool from Tanner Inc. • Shorten development cycle
– http://www.tanner.com • Reduce development cost
• MEMScap from MEMScap Inc.
– http://www.memscap.com
• SOLIDIS from ISE Inc.
– http://www.ise.com
 Example: IntelliSuite System Example: IntelliSuite Advantages
• Design for manufacturability
– Fabrication database
Thermo-electro- – Thin-film materials engineering
Materia mechanical
ls databa
se – Virtual prototyping

Electromagnetic

Piezoelectric
Fabrication database
Solid Modeling
• Ease of use
simulator
& Meshing – Consistent user interface
Anisotropic etch
– Communication with existing tools
d itor
Mask e
Thermal-fluid- • Accuracy
structure
– MEMS-specific meshing and analysis engines
Fabrication Simulation Performance Analysis – In-house code development
– Validated by in-house MEMS designers
– ISO certified for quality
Model courtesy of Auburn University

The Design Process MEMS CAD System Flow

• All systems have some common threads to

their design Layout
Layout

– Device design System

SystemModeling
Modeling
Process
Process
Design
Design
• Design a manufacturable component Manufacturing
Manufacturing
Data
Dataand
andQA
– Package design QA
Structures
Structures
Device
DeviceModeling
Modeling
• Design a practical package
– System design
• Design the system into which the device fits.
Package
PackageDesign
Design
• Goal: concurrent design at these levels
 Types of MEMS Design Who Designs?
• Custom Level
• Design New MEMS in New Process Digital
Analog
• Goal: A New MEMS component System Design

• Semi-Custom
• Design Existing MEMS in New Process
MEMS
• Goal: A Better MEMS component
• Standard/IP System
Architect
• Re-Use Existing MEMS and MEMS Process
• Making Existing MEMS Available to IC level Packaging
Designers to Build new systems

What is Top Down Design Implementing Top Down Design

• System Architect • Iterative design in each subsystem Implementing the
– Designs and Simulates Mixed Technology System at a Architect to Designer Loop
high level – Behavioral Model to Layout (Design)
• Subsystem Designers – Layout to Behavioral Model (Verify)

– Receive subsystem target specs in Hardware Description • Enable Communication in the Design Team
Language (HDL) form from SA • Interoperability (Composite CAD VHDL-AMS working
– Design and pass back HDL model of realizable subsystem group)
HDL
– Iterate with SA until realizable design is acceptable Specification
• Top Down Design MEMS
– Enables SA to make tradeoffs among subsystem design System Subsystem Analog
Digital
teams Design Design
– Enables Design teams and SA to quantitatively
communicate their goals and constraints Verification
against HDL
 MEMS IC Design Flow Cornering the Design Space
System
Requirements

Stimulus Response
Bottom-up 0.004 Central Location

Change in relative Capacitance

Model Verification Bottom Location
Sub-System
0.002
Behavior Requirements

Stimulus Response 0

Model Model -0.002

Behavior Extraction
-0.004

Simulated von Mises stress

Top-down Layout 3D Physics in Analog Devices ADXL76 -0.006
Design Model 200 250 300 350 400
Temperature [K]

Outline of the Task Sequence Layout and Process Resources

Accomplished by a CAD Tool
• First Resource: The Process Description of the
interface and the driving circuitry:
• Layout and process
– Can be acompished using a layout file editor (eg.
• Topography simulation CADENCE, http://www.cadence.comor L-Edit,
http://www.tanner.com)
• Boundaries, IC process results and Material
• Second Resource: The Process flow description
properties
file:
• Mesh generation – Relates a processing step to each lithography mask in
• Device simulation the layout file
– Can be optimized by using the MISTIC software from
• System-Level Simulation the University of Michigan
• MEMS Control CAD (http://www.eecs.umich.edu/mistic/)
 Layout Editor Topography Simulation
• Layout process • Goal: Obtain a realistic topography of the
– Multi-layer mask sets
– Cell hierarchy considered device by:
– Boolean operations – Realistically representing complex 2D and 3D
– Curved shapes
structures to simulate the manufacturing
• MEMS-specific features process
– Any-angle feature creation
– Multi-copy by translation
or rotation
• Links directly to process
simulation and mesh
generation
• Compatible with GDSII &
DXF

Anisotropic Etch Simulation

Process Simulation
(AnisE®)
• Document & validate • Etch rate databases
process steps or process flows • Single & double sided
• Model creation directly etching
from fabrication process • Multiple etch stops
• Link process & design • Real time etch visualization
to reduce prototype runs • 3D geometry visualization
• Process database • Direct measurements of
– MEMS process steps etch depths and feature
– Standard foundry templates sizes
– Expandable for custom • Study process deviations
steps or templates
Above: Examples of corner compensation
Below: Rounded edge after 1 hour (left) and 5 hours (right)

Model courtesy of the University of Windsor

SEM courtesy of IME Singapore Lower models courtesy of OptIC
 Boundaries, IC Process Results
Virtual Prototyping
and Material Properties
• Description of the material interface
boundary
• Dopant Distribution within each layer of the
device
• Distribution of residual stresses
Surface micromachining simulation
• Optimization of the Material Properties (eg.
MEMCAD from Microcosm Inc.)
• Validate process
• Verify mask set Anisotropic etch simulation

• View 3D geometry after each process step

Model (left) courtesy of Tennessee Technological University

Thin-film Material Expertise Mesh Operations

• Accurate material property
estimation for device • Generate a computational mesh for device
analysis simulation by either using boundary
• Provide insight into element methods or finite element methods
material behavior
or coupling of both
• Expandable for custom
materials or processes
• Reduce number of
materials characterization
fabrication runs
• Increase device
performance
Young’s Modulus variation in deposited layer due to process
• Improve yields temperature and film thickness
 Automatic Mesh Generation Interactive Mesh Refinement
• From fabrication simulation • Mesh optimization provides faster simulation times
– 3D model based on mask set and process sequence – 100% Automated or 100% user-driven
– Material properties transferred to analysis
– Local or global
• Import or export ANSYS, ABAQUS, PATRAN models
• Mesh optimization results in greater accuracy
– Independent refinement of electrostatic & mechanical
meshes

Models courtesy of the University of Windsor (left), Raytheon (center),

and Tennessee Technological University(right) Model (right) courtesy of DSI, Singapore

Device Simulation Modeling of All Contributing Factors

• Process induced effects
• Compute the coupled response of a MEMS – Deformation
– Stiffening
device using numerical methods • Micro-assembly &
post-contact behavior
• Also provide many coupling effect that • Coupled dynamic analysis
MEMS rely on (eg. electromechanical, – Frequency vs. voltage bias
– RF switching time
thermomechanical, optoelectrical, and • Macro-model extraction
optomechanical coupling behaviors) • Electrostatic force vs.
Displacement characterization
• Extract behavioral models for system-level • Coupled boundary element &
finite element analysis
simulation. • Large & small displacement theory
• 3D static & dynamic analysis

Model courtesy of Auburn University

 3D Device Modeling Coupling Effects
• Structural Mechanics (including contact)
• Electrostatics & Capacitance Extraction
• Thermo-mechanics
• Coupled Electro-Thermo-Mechanics (including contact)
• Thermal Flow Analysis
• Piezoresistive Devices
• Electro-Thermal Devices
• CFD for Compressible and Incompressible Flow
• Electrokinetics and Chemical Transport in Liquids
• Inductance (RL) and RL-Thermo-Mechanics
• Damping of complex structures Electrokinetic Switching A. K. Noor and S. L Venneri, bulletin for the international association for computational
for Chemical Transport o
mechanics, n 6, summer 1998

System-Level Simulation Device to System

• Conversion of a numerical matrix to an XL76 Layout Process
Flow
equivalent subcircuit
• Translate specific changes in device
configuration, dimensions, and material MemBuilder
3-D
properties into the circuit-equivalent Solid
Model
behavioral model System
Performance

Sensor Finger Capacitance (fF)

95.00

94.50

94.00

93.50
Cross-
93.00
axis
92.50
sensitivity
92.00
0 0.5 1 1.5 2 2.5 3 3.5
System Model Time (ms)
 HDL (Macromodel) Generation
System Modeling
from Device Modeling
60
Input Acceleration • Extract from 3D model:
40
Auto Fit of Behavior Curves
Acceleration (g's)

20

-20

-40

-60
– Mechanical Spring
– Electrostatic Forces
0 0.5 1 1.5 2 2.5 3 3.5
Time (ms)

Sensor HDL Model z

– Mass
pos pos1
zposi t i on_m ass

pos1 k : 0. 65

ADXL76 finger-cell 3D model

m ass

m: 2e-10

pos2 d: . 5 7e-6

– Damping Coefficients
pos2 y

zpo s2 zpos2

zpos i t i on_ mass ypo si t i on _out

zpos1 zpo s1

ypos 1 ypo s2 ypos2 ypos1

pos1
posi t i on
vi n ypo si t i on _mass vout

e1 e2 e2 e1 12 0u

pos2
100M eg

v initial:0
pulse:30
CLK
period:50m
tr:.1u vbi as

• Auto generation of 6-DOF

tf:.1u di el e ct ri c2 d_mi c roc
pos2 d i el ect ri c2d_ mi cro c

del t a0: 60u pos1 d el t a0: 60u

v
50

d: . 23u 50

del t a0: 30 u

pos2 pos2 del t a0: 30 u

pos1 pos1
pos1 pos1
posi t i on pos i t i on

30 u 90u

k : 0. 325 k: 0. 3 25

Demodulator Output
pos2 pos2

pos
mass

3.00

HDL Model
m : 2e-10

De
mo
dul2.50
Sensor Finger Capacitance (fF)

95.00 ato
r
94.50 Ou 2.00
tpu

• Industry standard
t
94.00 (V) 1.50

93.50 Cross Axis

93.00
1.00
Capacitance
Sensitivity system/circuit modeling characterization
0.50
92.50
0.00
92.00
0 0.5 1 1.5 2 2.5
0 0.5 1 1.5 2 2.5 3 3.5

tools: SABER, SPICE,

Time (ms)
60
Time (ms)

Matlab, etc.

Effect Of 10% Tether Misalignment Effect Of A +/- 10% Variation Of

On Response Tether Spring Constants
 Packaging Simulation Package to Device
• Automated package-device interaction
simulation by: Device

– Separating FEA of both the package and the

device Coupled Package
and Device Effects
– Coupling the results through parametric Compact
behavioral package models (MEMCAD from Package
Microcosm Inc. Model

Package

Package Model Calibration Packaging Sensitivity Analysis

Package Mechanical
Analysis
Silicon die,
wirebonds,
and leadframe
of plastic
package
µm Temperature BC MEMCAD

Displacement along the sensitive 0 .0 0 4 C e n t r a l L o c a t io n

Change in relative Capacitance

B o t t o m L o c a t io n
axis of the resistor element 0 .0 0 2
Displacement
0

- 0 .0 0 2 Extraction
- 0 .0 0 4

- 0 .0 0 6
200 250 300 350 400
T e m p e ra tu r e [ K ]

Device Mechanical /
Potential distribution in the metal foil Electrical Analysis

Metal foil strain gauges

 Summary
• MEMS/MST tools exist today.
• The Tools can support the design of RF devices
and systems.
IntelliSuite Application
• The Design Process needs to support the
integration of MEMS and ASIC subsystems. Examples
• ALL players in the design process (Architect,
Analog, Digital, MEMS, Package) must
communicate.
• Communications are enabled by specific layers in
the design tool set which allow models from one
subsystem to influence the others.

Raytheon Systems NASA

• RF switch • Radiation detectors
– Corrugated geometry contact analysis
– Electrostatically actuated

Von Mises Stress

CAD

Device model Fabricated device SEM

CAD – stress results Fabricated array
 Ford Microelectronics, Inc. Gyro / Accelerometer
• Capacitive pressure • Natural frequency shift
sensor 4.4
Ford Experimental Data
– Electrostatic or thermal frequency tuning
– Capacitance as a function 4.2 IntelliSuite Simulation Data
– Only 3D simulation available

Device Capacitance (pF)

of applied pressure 4

– Accounts for levitation & other 2nd order

3.8

3.6
effects Natural Frequency vs. Voltage

Norm. Frequency (48854.3 Hz)

3.4

3.2

3
0 50 100 150

Differential Pressure (PSI)

Comparison between IntelliSuite simulation and
Ford’s experimental results
Voltage (V)

*Ford Microelectronics, Inc. Colorado Springs, CO, JMEMS, June’96, p 98

Corning IntelliSense Integrate With System-level Design

• Mirror array packaging analysis • Electro-mechanical output
as input to optical model
– 3D mirror profile
– Maximum mirror angle
– Jitter angle associated with
mirror stability
Micro-mirror radius vs.fraction of geometric encircled

– Surface material
energy in an 8 micron diameter

0.8

0.7

encircled energy in an 8 micron diameter

0.6

Fraction of geometric
0.5

0.4

0.3
Range of silicon wafer Range of silicon wafer
natural radius natural radius
0.2

0.1

0
-6000 -4000 -2000 0 2000 4000 6000

Radius of micro-mirror (mm)

IntelliSense Packaging Group

Model courtesy of NASA, Goddard Space Flight Center
 Fluidic analysis overview Electrophoresis channels
• 3D Navier-Stokes solution
• Incompressible, laminar,
single-phase flow
• Heat transfer
• Steady-state and transient
• Squeeze-film damping
• Electro-kinetic phenomena
– Electro-osmosis
Clockwise from top: AnisE simulation of channels,
– Electro-phoresis velocity vectors, flow profile

• Finite element & finite

volume solvers

Electro-osmosis
• Cross-channel fluid flow

Ambient Pressure at all Ports

5V

Injection Port 0V
10 V

5V

Pressure Plot Vector Plot

Ref.: Patankar and Hu; Analytical Chemistry, Vol. 70, No. 9, May 1, 1998