MEMS

1

Today’s Topics

 MEMS Design Strategy

 System approaches

 Design specifications • Typical specifications for sensors

• Specifications for actuators

• Examples of some sensor parameters

7/3/2019ECE 5134 Dr. H. Qu

MEMS Design Strategy

 Design of MEMS as a System  Sensor design  Actuator design  Interface design

 Packaging design

• Integration approaches and levels

• Packaging design

Physical parameters

User Power Supply and Management

I/O Channel and Protocol

Processing and Control Circuitry

Sensors and (Actuators)

Interface and Conditioning

7/3/2019ECE 5134 Dr. H. Qu

1

2

2

MEMS Design Strategy

 Design of MEMS as a System  Individual blocks

• Parasitics  Signal attenuation

• Noise  Signal-to-Noise ratio

• Packaging approach and device footprint

Physical parameters

UserPower Supply and Management

I/O Channel and Protocol

Processing and Control Circuitry

Sensors and Actuators

Interface and Conditioning

7/3/2019ECE 5134 Dr. H. Qu

MEMS Design Strategy

 Design of MEMS as a System  Integrated Devices (blocks can be integrated)

Outside World

UserPower Supply and Management

I/O Channel and Protocol

Signal Processing and Control Circuitry

Sensors and Actuators

Interface and Conditioning

• Somewhat improved performance

• Cost reduction

7/3/2019ECE 5134 Dr. H. Qu

3

4

3

MEMS Design Strategy

 Design of MEMS as a System  Integrated System

Outside World

UserPower Supply and Management

I/O Channel and Protocol

Processing and Control Circuitry

Sensors and Actuators

Interface and Conditioning

• Highly integrated

• Low cost

• Increased Reliability

7/3/2019ECE 5134 Dr. H. Qu

MEMS Design Strategy

 Design of MEMS as a System  System Integration Levels

Physical World

Processing and Control Circuitry

Sensors and Actuators

Interface and Conditioning

• Assembling

• Hybrid packaging

• Monolithic integration

Assembling

Sensor chip ASIC chip

Hybrid packaging

Monolithic Integration

Example: 3-axis accelerometers

5

6

4

MEMS Design Strategy

 Technology-Driven or Market-Driven?  Technology demonstration (concept and idea verification)

 Research tools

 Commercial products

 High-level Design Issues  Market

 Impact

 Competition

 Technology

 Manufacturing Category Market Impact Competition Technology Manufacturing

Technology Demonstration

** ***

Research Tools ** ** * *** **

Commercial Product *** *** *** *** ***

7/3/2019ECE 5134 Dr. H. Qu

MEMS Design Strategy

 MEMS Design Procedure

 Valid to all electronic devices/systems

Innovative ideas are of the most important!

Source: Stephen Senturia, Microsystem Design 7/3/2019ECE 5134 Dr. H. Qu

7

8

5

MEMS Design Strategy

 MEMS Design Approaches

System

Device

Physical

Process

S im

u la

tio n

V e

ri fi c a

tio n

 Top-down Design • System level modeling

• Device: Macro models

• Physical: numerical modeling, FEM

• Process: Technology CAD, CAM

 Bottom-up • Reverse procedure for design verification

 Different Modeling Levels • System level

» State variables

» State equations

» ODEs

• Physical level » PDE’s for FEM

7/3/2019ECE 5134 Dr. H. Qu

MEMS Design Strategy

 MEMS Design Approaches  Analytical or Numerical?

Source: Stephen Senturia, Microsystem Design

Design tools

7/3/2019ECE 5134 Dr. H. Qu

9

10

6

MEMS Design Specifications

 MEMS Specifications  Requirements

• Linear response is mostly desired.

Physical Parameters

Power Supply and Management

I/O Channel and Protocol

Processing and Control Circuitry

Sensors Interface and ConditioningP(t)

Pmea.(t)

• P(t): Physical variable (input)

• X(t): Sensor excitation

• Y(t): sensor response (output)

Y(t)

X(t)

Pmea. (t) = P(t) ?

Calibration: To determine the function that relates Y(t) to known P(t). Higher grade references are needed.

7/3/2019ECE 5134 Dr. H. Qu

Span

Range

 Full Scale Output (FSO, Span) • Ymax -Ymin

 Range: The input signal that causes the output to reach sensor span (maximum input).

 Linearity • Closeness of calibration curve to a specified straight line (maximum

deviation of calibration point from a regressing straight line as percentage of FSO).

 Offset • Y(t) under normal excitation and zero applied input (P(t)=0).

MEMS Specifications

 MEMS Specs (Sensors)

max l l

Y

FSO  

 

max lY 

Simple Linearization: least square regression

7/3/2019ECE 5134 Dr. H. Qu

11

12

7

MEMS Specifications

 MEMS Specs (Sensors)  Hysteresis

• Maximum difference in Y(t) when the value is approached first with increasing input and second with decreasing input, expressed in percentage of FSO.

 Error • Difference between measured Pmea.(t) and true value of P(t) obtained

by calibrated devices (normally measured as percentage of FSO.) » Mean square root of linearity, hysteresis, etc.

2 2 l h   

ΔYmax max h

Y

FSO 

 

7/3/2019ECE 5134 Dr. H. Qu

MEMS Specifications

 MEMS Specs (Sensors)  Sensitivity

• Magnitude of change of Y(t) with respect to change in P(t).

• S=Y(t)/P(t)

 Accuracy • Another expression of Error in percentage

 Repeatability • Agreement between independent measurements made under the

identical conditions (maximum difference in output readings given as percentage of FSO)

 Resolution • Smallest change in P(t) that results in a detectable change in Y(t) (called

“Threshold” if increment is from zero).

 Frequency response (Bandwidth) • Change of output/input magnitude ratio (sensitivity) with frequency

(=2f) and phase difference for sinusoidal varying input

7/3/2019ECE 5134 Dr. H. Qu

13

14

8

MEMS Specifications  MEMS Specs (Sensors)

 Cross talk, interference (Cross-axis sensitivity for accelerometers) • Interference from other physical variables • Sensitivity of sensor to transverse inputs (also known as Transverse

Sensitivity)

 Noise (a dedicated topic) • Energy (power) spectrum (density) of noise signal (over frequency) • Related to resolution (over certain interested bandwidth)

, _ ,6.6

17.82 V/

n p p n rmsv v

Hz

 



111.284

20 ,

6

10

2.7 10 /

n rmsv

V Hz







 

What is read from a spectrum analyzer

Why?  Normal distribution of noise

around its mean value.  RMS value is actually .  3 theory

2.7 V/Hz

7/3/2019ECE 5134 Dr. H. Qu

MEMS Specifications  MEMS Specs (Sensors)

 Noise (a dedicated topic) • Energy spectrum (density) of noise signal (over frequency) • Related to resolution (over certain interested bandwidth)

2

2

Probability density function

1 1 ( | , ) exp[ ( ) ]

22 : mean value of the expected x;

: standard deviation of x (variance: ).

x f x

  

  

 

  

Normal distribution:

7/3/2019ECE 5134 Dr. H. Qu

2 2

1

2

1 ( ) standard deviation of x. (variance: ).

for individual values x;

= ( ) ( ) , with = ( )

for continous variable x.

N

i i

X X

x N

x p x dx x p x dx

  

  



 

 



 

f( x/

µ ,δ

)

15

16

9

MEMS Specifications  MEMS Specs (Sensors)

 Noise (a dedicated topic) • Energy spectrum (density) of noise signal (over frequency) • Related to resolution (over certain interested bandwidth)

Cumulative distribution function of normal distribution:

7/3/2019ECE 5134 Dr. H. Qu

2

2

2

2

0

1 ( )

2

2 ( )

1 ( ) ( ) [1 ( )]

2 2 ( ) ( ) ( )

x t

x t

x e dt

erf x e dt

x x F x erf

P X x x



  

 







 



     

    





For µ-3δ  x  µ+3δ

( ) ( ) ( ) 2 ( ) 1

3 2 ( ) 1

3 1 ( ) 1 (2.12) 0.99728 99.728%

2

P X F x F x F x

erf erf

   

       

  

     

( ) 1 ( )x x    

MEMS Specifications  MEMS Specs (Sensors)

 Noise (a dedicated topic) • Energy spectrum (density) of noise signal (over frequency) • Related to resolution (over certain interested bandwidth)

3, 6 control

Cumulative distribution function of normal distribution:

Range Probability x falls in the range Expected frequency outside range

μ ± 1σ 0.682689492137086 1 in 3

μ ± 1.5σ 0.866385597462284 1 in 7

μ ± 2σ 0.954499736103642 1 in 22

μ ± 2.5σ 0.987580669348448 1 in 81

μ ± 3σ 0.997300203936740 1 in 370

μ ± 3.5σ 0.999534741841929 1 in 2149

μ ± 4σ 0.999936657516334 1 in 15787

μ ± 4.5σ 0.999993204653751 1 in 147160

μ ± 5σ 0.999999426696856 1 in 1744278

μ ± 5.5σ 0.999999962020875 1 in 26330254

μ ± 6σ 0.999999998026825 1 in 506797346

μ ± 6.5σ 0.999999999919680 1 in 12450197393

μ ± 7σ 0.999999999997440 1 in 390682215445

μ ± mσ erf(m/√2) 1 in 1/(1-erf(m/√2))

6-Sigma strict quality management for many companies

3.3σ  0.99906

7/3/2019ECE 5134 Dr. H. Qu

17

18

10

MEMS Specifications  MEMS Specs (Sensors)

 Signal-to-Noise • S/N = (Y/nrms )

2 where Y is the output magnitude (rms) and nrms is the root-mean-square noise (normally in dB, dB=10·log10(S/N_d).)

• Determined by operational condition • Shannon-Hartley Theorem (C = B·log2(1+S/N))

 Resolution • Smallest detectable signal in certain bandwidth

 Dynamic range • Ratio between the highest and lowest detectable signal magnitude

(normally in dB) • In some applications it’s interchangeable with S/N.

 Selectivity • Ability to measure one input (measurand) in the presence of other inputs

 Overload characteristics • Maximum magnitude of input that can be applied to the sensor without

changing the sensor response

 Stability • Ability of sensor to reproduce output for identical input and conditions

over time (percentage of FSO)

C: channel capacity B: Bandwidth

7/3/2019ECE 5134 Dr. H. Qu

MEMS Specifications

 MEMS Specs Example (Actuators - Switches)  Motion range

 Force and torque output capacity

 Dynamic response speed

 Bandwidth

 Power consumption

 Linearity of response (displacement, force..)

 Cross-sensitivity

 Stability

 Footprint

An electrothermal actuator

A bidirectional scanning electrostatic micromirror

7/3/2019ECE 5134 Dr. H. Qu

19

20

11

MEMS Specifications

 Exemplified Inertial Sensor Test Setup with Reference Devices

Rotary table

Hand-held shaker

Spectrum analyzer

shaker

Reference accelerometer

Kistler 8638B5

DUT

7/3/2019ECE 5134 Dr. H. Qu

MEMS Specifications

 Exemplified Inertial Sensor Test Results

 Derivation of some parameters

-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 1 -0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

Acceleration in y-axis (g)

S e n

s o

r o

u tp

u t

(V )

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0

50

100

150

200

250

300

350

400

input acceleration (g)

o u

tp u

t (V

)

Noise spectrumWaveforms 7/3/2019ECE 5134 Dr. H. Qu

21

22

12

MEMS Specifications

 Exemplified Inertial Sensor Test Results

 Derivation of some parameters (for this particular device)

103.9 62010 6.38 10 (V/ Hz )

  

• Maximum detectable input acceleration (determined by the linear range):

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0

50

100

150

200

250

300

350

400

input acceleration (g)

o u

tp u

t (V

)

offset

• Sensitivity:

2.2 g

• Measured output noise floor:

(320-10)/2.2 = 140 mV/g = 0.14 V/g

-103.9 dB Vrms/Hz.

This is converted to

7/3/2019ECE 5134 Dr. H. Qu

MEMS Specifications

 Exemplified Inertial Sensor Test Results

 Derivation of some parameters (for this particular device)

-6 -5Measured noise 6.38 10 V/ Hz= = 4.56 10 g/ Hz

Sensitivity 0.14 V/g

 

• Input-referred noise floor:

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0

50

100

150

200

250

300

350

400

input acceleration (g)

o u

tp u

t (V

)

• Resolution: (Smallest detectable input) This is from the total noise power to be excluded.

5 4

Input_referred noise floor Bandwidth

4.56 10 g/ Hz 400 9.1 10 gHz  

    

2 min

Total noise power = Noise power density Bandwidth

( ) H

L

f

n n n H L n

f

v P v df v f f v BW



     

7/3/2019ECE 5134 Dr. H. Qu

23

24

13

MEMS Specifications

 Exemplified Inertial Sensor Test Results

 Dynamic range calculation

10 10 4

Dynamic range:

input 2.2 DN=20 log ( ) 20 log ( )

detectable input 9.1 10

67.6 (dB)

Smallest detectable input = resolution

Maximum

smallest    

 

7/3/2019ECE 5134 Dr. H. Qu

Assignments

 MEMS device specifications

 Find and understand specifications of Analog Devices’ ADXL330 3-axis accelerometer.

 Watch the video clip on Semiconductor Fabrication Processes (from Global Foundry) post on MOODLE.

 Find and watch other similar video clips (such as those on memscentral and even Youtube).

7/3/2019ECE 5134 Dr. H. Qu

25

26