WELCOME TO THE COURSE

ON

MICRO MACHINING PROCESSES

BY

Dr. V.K.JAIN
MECHANICAL ENGINEERING DEPARTMENT
I.I.T KANPUR- 208016
e-mail : [email protected]
Prof. V.K.Jain, Mechanical Engineering
1
Department, I.I.T. Kanpur
ORGANISATION

•TEXT BOOKS

•CHAPTERS TO BE COVERED IN THIS COURSE ON “MMPs”

• WHY DO YOU NEED MMPs

•CLASSIFICATION OF MMPs

•EVOLUTION AND WORKING PRINCIPLES OF SOME AMPs

•APPLICATIONS OF MMPs

•CONCLUSIONS

Prof. V.K.Jain, Mechanical Engineering

2
Department, I.I.T. Kanpur
ADVANCED MACHINING PROCESSES

Prof. V.K.Jain, Mechanical Engineering

3
Department, I.I.T. Kanpur
 Grading Policy

Mid Sem. Exam--> 30 %

End Sem. Exam 40 %
Term Paper  15%
Lab. Visits  05%
Attendance  05%
Quizzes (unannounced)-- 05%

For getting “D” or better grade, one should have minimum 30 % marks
in both theory exams (Mid Sem. + End Sem.) and 35 % in total. For getting „E‟ ,
One should have 25 % in theory and 30 % in total.

Prof. V.K.Jain, Mechanical Engineering

4
Department, I.I.T. Kanpur
 TEXT BOOKS

Prof. V.K.Jain, Mechanical Engineering

5
Department, I.I.T. Kanpur
 BOOKS

 INTRODUCTION TO MICROMACHINING , V.K.JAIN (EDITOR) PUBLISHED BY

NAROSA PUBLISHERS, N EW DELHI (2009). (second Edition)

 MICROMANUFACTURING PROCESSES BY V. K. JAIN (Editor), CRC PRESS.

 ADVANCED MACHINING PROCESSES BY V.K JAIN , ALLIED PUBLISHERS,

NEW DELHI.

 NON- CONVENTIONAL MATERIAL REMOVAL PROCESSES BY V.K.JAIN,

BLOCK-4, INDIRA GANDHI NATIONAL OPEN UNIVERSITY (IGNOU), NEW
DELHI

Reference Books

 MICROMACHINING METHODS BY J.A. Mc GEOUGH, CHAMPAN AND HALL, LONODON

Prof. V.K.Jain, Mechanical Engineering

6
Department, I.I.T. Kanpur
TOPICSTO BE COVERED

Prof. V.K.Jain, Mechanical Engineering

7
Department, I.I.T. Kanpur
 SUBJECT AREAS

S.N TOPIC

1 INTRODUCTION TO THE COURSE & CLASSIFICATION OF MMPs

PART – 1 : MECHANICAL TYPE ADVANCED MICRO MACHINING
PROCESSES
2 ABRASIVE JET MICRO MACHINING (AJMM)

3 ULTRASONIC MICRO MACHINING (USMM)

4 ABRASIVE WATER JET MICRO MACHINING (AWJMM)

PART – 2 : ABRASIVE BASED NANO FINISHING PROCESSES

5 ABRASIVE FLOW FINISHING (AFF)

6 CHEMOMECHANICAL POLISHING (CMP)

Prof. V.K.Jain, Mechanical Engineering

8
Department, I.I.T. Kanpur
 SUBJECT AREAS

S.N TOPIC

7 MAGNETIC ABRASIVE FINISHING (MAF)

8 MAGNETORHEOLOGICAL FINISHING (MRF)

9 MAGNETORHEOLOGICAL ABRASIVE FLOW FINISHING (MRAFF)

10 MAGNETIC FLOAT POLISHING (MFP)

PART – 3 : THERMOELECTRIC TYPE MICRO MACHINING PROCESSES

11 ELECTRIC DISCHARGE MICROMACHINING (EDMM)

12 WIRE EDM , EDDG, ELID

13 LASER BEAM MICROMACHINING (LBMM)

14 ELECTRON BEAM MICROMACHINING (EBMM)

Prof. V.K.Jain, Mechanical Engineering

9
Department, I.I.T. Kanpur
 SUBJECT AREAS

S.N TOPIC

PART – 4 : CHEMICAL AND ELECTROCHEMICHAL TYPE ADVANCED

MACHINING PROCESSES

15 ELECTROCHEMICAL MICROMACHININIG (ECMM)

16 ELECTROCHEMICHAL MICRO DEBURRING

17 CHEMICHAL AND PHOTOCHEMICAL MICROMACHINING

PART-5 : TRADITIONAL MECHANICAL MICROMACHINING

PROCESSS
18 MICRO TURNING

19 MICRO MILLING

Prof. V.K.Jain, Mechanical Engineering

10
Department, I.I.T. Kanpur
 SUBJECT AREAS

S.N TOPIC

20 MICRO DRILLING

PART- 6 : MISCELLANEOUS TOPICS

21 FOCUSSED ION BEAM (FIB) MACHINING

22 SELECTION OF MICRO MACHINING PROCESSES

23 CONCLUDING REMARKS

Prof. V.K.Jain, Mechanical Engineering

11
Department, I.I.T. Kanpur
 INTRODUCTION
• In today’s high tech engineering industries, the designer’s requirement for the
component are stringent, such as:

 Extraordinary properties of materials (say, high Strength, high heat

resistant, high hardness, corrosion resistant etc.).

 Complex 3D component (say, turbine blade).

 Miniature features ( filters for food processing and textile industries having a
few tens of micrometer as hole diameter and thousands in numbers).

 Nano level surface finish on complex geometries which are impossible to

achieve by any traditional methods (say , thousands of turbulated cooling holes
in a turbine blade, making & finishing of microfluidic channels in the electrically
conducting and non-conducting materials (say, glass, quartz, ceramics)).

• Such features on a component can be achieved only through the advanced

manufacturing processes in general and advanced machining processes in
particular.

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 12

 PRESENT DAY DEMAND TRENDS IN INDUSTRIES
(AEROSPACE , MISSILES , AUTOMOBILES, NUCLEAR REACTORS, ETC.)

ENGG. MATERIALS ULTRAHIGH STRENGTH , HARDNESS

HAVING MUCH SUPERIOR VERY HIGH TEMPTURE RESISTANCE
PROPERTIES DIFFICULT TO MACHINE BY
CONVENVENTIONAL MACHINING
METHODS

?
WORK PIECE MATERIAL HARDNESS >> TOOL MATERIAL HARDNESS

HOW TO SOLVE THE PROBLEM

ADVANCED MICRO MACHINING PROCESSES

WORKPIECE HARDNESS DOES NOT MATTER IN AMMPs
Prof. V.K.Jain, Mechanical Engineering
13
Department, I.I.T. Kanpur
 METHODS OF MICRO FABRICATION

Material Metal forming

Material removal
deposition (etc)
Electro discharge Deposition Traditional material removal
(EDD) processes

Chemical vapor deposition Advanced material removal

(CVD) processes

Physical vapor deposition

(PVD)
Micro joining
Rapid prototyping/rapid Micro casting
tooling (RP/RT)
Etc.

LIGA

Electro chemical spark

deposition (ECSD)

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 14

WHY DO YOU NEED OF MMPs?

Prof. V.K.Jain, Mechanical Engineering

15
Department, I.I.T. Kanpur
 Why miniaturization?

Minimizing energy and materials use in manufacturing

Reduction of power budget

Faster devices
Increased selectivity and sensitivity
Improved accuracy and reliability
Cost/ performance advantages
Integration with electronics, simplifying systems

16
17
 MICRO-PRODUCTS

NOWADAYS , FOCUS IS ON MINIATURIZATION THROUGH DEVELOPMENT OF

NOVEL PRODUCTION CONCEPTS (SPECIALLY MICRO & NANO) FOR THE
PROCESSING OF NON-CERAMIC MATERIALS.

MICROFABRICATION DEALS WITH ALL KIND OF MANUFACTURING

PROCESSES BUT AT MICRO & NANO LEVEL.

THE REPLICATION OF MICROPARTS THROUGH MOLDING IS ONE OF THE

PREFERRED ROUTES FOR MICROMANUFACTURE BECAUSE OF ITS
MASS-PRODUCTION CAPABILITY AND RELATIVELY LOW COST.

HOWEVER IN THIS COURSE I WILL MAINLY

CONCENTRATE ON MICRO ATTRITIOUS PROCESSES :
“MICROMACHINING PROCESSES”

18
18
 Material Removal Processes

Micro/nano machining Micro/nano finishing

Traditional machining processes Traditional finishing processes

Advanced machining processes Advanced finishing processes

19
 Micro / nano finishing

traditional Advanced
Grinding AFM

Lapping MAF

Honing MRF

CMP

MRAFF

ELID

MFP

20
 CLASSIFICATION Micromachining Methods BASED ON THE KIND OF ENERGY USED : MACHANICAL,
THERMOELECTRIC, ELECTROCHEMICAL & CHEMICAL, BIOCHEMICAL

Traditional Advanced

µ-turning Mechanical Thermal Chemical

µ-milling µ-AJM µ-EDM ECMM

µ-drilling µ-USM µ-EBM PCMM

µ-grinding µ-AWJM µ-LBM

AFM, MAF,
MRAFF, MFP
µ-PAM

HYBRID

µ-ECSM µ-ECG µ-EDG ELID, etc 21

BASED ON THE PROPERTIES OF WORK MATERIAL TO BE
MACHINED

 APPLICABLE ONLY FOR ELECTRICALLY CONDUCTING

MATERIALS : ECM, EDM, EBM.

 APPLICABLE FOR BOTH ELECTRICALLY CONDUCTING & NON -

CONDECTING MATERIALS: USM , AJM, LBM, ETC.

 APPLICABLE FOR NON – MAGNETIC MATERIALS : MAF, MRF, ETC.

 THERMAL CONDUCTIVITY, REFLECTIVITY, ETC. ALSO PLAY AN

IMPORTANT ROLE IN SOME CASES: LBM

Prof. V.K.Jain, Mechanical Engineering

22
Department, I.I.T. Kanpur
EXAMPLES: WHY WE NEED MICROMACHINING / MICROMANUFACTURING?

Prof. V.K.Jain, Mechanical Engineering

23
Department, I.I.T. Kanpur
 MICROMACHINING
MICRO MACHINING

Removal of material at micron level

• Macro components but material removal is at micro / nano level

• Micro / nano components and material removal is at micro / nano level (Ex.
MEMS, NEMS)

Unfortunately , the present day notion is

Machining of highly miniature components

with miniature features. Literally it is NOT
correct

MORE CORRECT DEFINITION IS material removal is

micro/nano level
WITH NO CONSTRAINT ON THE SIZE OF THE COMPONENT
Size: 2mm×2mm
Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 24
SOME MICROMACHINED PARTS

LASER-CUT STENTLIKE PATTERN

IN MINIATURE STAINLESS- STEEL
TUBE WITH 1.25-mm OD.

LASER-DRILLED HOLE(DIA.
15 µm) PATTERN IN
POLYURETHANE TUBE

25
25
PROBLEMS IN MICROMACHINING
• MECHANICAL DEFORMATION
• THERMAL DEFORMATION
• SURFACE INTEGRITY
• GAP BETWEEN TOOL AND WORK PIECE
• COORDINATE SHIFT IN TOOL HANDLING

26
26
 ENGINEERING MATERIALS

METALS AND ALLOYS PLASTICS AND COMPOSITES CERAMICS

• GETTING MORE POPULARITY

• DEFINITE ADVANTAGES OVER
OTHERS

HOW TO
MICROMACHINE
THEM

?
SOLUTION

ADVANCED MACHINNG PROCESSES

Prof. V.K.Jain, Mechanical Engineering

27
Department, I.I.T. Kanpur
 WHY DO YOU
NEED ADVANCED
MICRO
MACHINING
PROCESSES ?

some examples

Prof. V.K.Jain, Mechanical Engineering

28
Department, I.I.T. Kanpur
 MICRO MACHINING OF COMPLEX SHAPED WORKPIECES?

ELECTROCHEMICAL MACHINING

HOLE NORMAL TO THE WALL

Turbine Blade with cooling Holes

Prof. V.K.Jain, Mechanical Engineering Department, 29

I.I.T. Kanpur
Contoured Hole Drilled In Inconel Using ECM
 PATTERN OF HOLES DRILLED BY EBM
HOLE  =0.006 mm (6 m);
PART OF A HELICOPTER HOLE=0.09 mm HOLES DENSITY = 4000/cm2 HOLES DENSITY = 200,000 / cm2 ;
TURBINE “ HOLES DRILLED WORKPIECE- S.S.;
BY EBM” THICKNESS = 0.12 mm; TIME= 2s / HOLE
THICK = 0.2 MM; TIME = 10 S/HOLE

AFF MEDIA ACTS AS A SELF-DEFORMABLE STONE

Prof. V.K.Jain, Mechanical Engineering

30
Department, I.I.T. Kanpur
WORKING PRINCIPLE OF SOME MMPs

Prof. V.K.Jain, Mechanical Engineering

31
Department, I.I.T. Kanpur
MECHANICAL ADVANCED MICROMACHINING PROCESSES

• Abrasive jet micromachining (AJMM)

• Abrasive water jet micromachining (AWJMM)

• Water jet micromachining (WJMM)

• Ultrasonic micromachining (USMM)

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 32

 How Abrasive JET Machining (AJM) Works?

F1
D1

Soft wall

F2
D2 F2>F1
D2>D1
Soft wall

• An young boy hits a ball twice on the wall with F1 & F2.

• The ball makes a crater of size D1 & D2 such that D2>D1 when F2>F1.

• D2 & D1 size = ø (Kinetic energy of the ball when hitting the wall).
= Force (or velocity of the ball with which it hits the wall,
and mass of the ball)
•Abrasive Jet Machining (AJM) works on the same principle. 33
WATER JET MACHINING (WJM), ABRASIVE WATER JET
MACHINING (AWJM) , AIR JET MACHINING (AJM)

34
APPLICATIONS OF ABRASIVE WATER JET CUTTING
(AWJC)

Granite cutting

35
COURTESY : IITM CHENNAI
PRINCIPLE OF MECHANICAL ADVANCED MICROMACHINING

 Fine abrasive particles with high kinetic energy (KE) hit the

workpiece at an angle and remove the material in the form of

micro/nano-chips.

 If the KE of the abrasive particle is high enough, then it will remove

the material by shear deformation in case of ductile workpiece

material and by brittle fracture if work piece material is brittle.

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 36

 APPLICATIONS

• Holes up to 66 μm deep can be drilled without employing special

techniques.

• Micro burrs are clearly visible.

• This process is also useful for producing micro cavity.

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 37

LASER BEAM MACHINING (LBM)

Prof. V.K.Jain, Mechanical Engineering

38
Department, I.I.T. Kanpur
ULTRASONIC MACHINING

Prof. V.K.Jain, Mechanical Engineering

39
Department, I.I.T. Kanpur
LASER BEAM MACHINING (LBM)

Prof. V.K.Jain, Mechanical Engineering

40
Department, I.I.T. Kanpur
ELECTRON BEAM MICROMACHINING

Prof. V.K.Jain, Mechanical Engineering

41
Department, I.I.T. Kanpur
 FOCUSSED ION BEAM MACHNING

 SPUTTERING OFF: KNOCKING OUT ATOMS FROM THE WORK-PIECE SURFACE

BY THE KINETIC MOMENTUM TRANSFER FROM INCIDENT ION TO THE TARGET
ATOMS
 REMOVAL OF ATOMS WILL OCCUR WHEN THE ACTUAL ENERGY TRANSFERRED
EXCEEDS THE USUAL BINDING ENERGY.

FIG. SCHEMATIC ILLUSTRATION

 AT SUFFICIENTLY HIGH ENERGY, THE CASCADING EVENTS WILL PENETRATE MORE DEEPLY INTO THE
SOLID, SEVERAL ATOMS OR MOLECULES WILL BE EJECTED OUT AND THE BOMBARDING ION WILL
BECOME IMPLANTED DEEP WITHIN THE MATERIAL.

Prof. V.K.Jain, Mechanical Engineering

42
Department, I.I.T. Kanpur
ELECTROCHEMICAL & CHEMICAL MICRO MACHINING PROCESSES

ELECTROCHEMICAL MACHINING

Prof. V.K.Jain, Mechanical Engineering

43
Department, I.I.T. Kanpur
WORKING PRINCIPLE OF NANO FINISHING
TECHNIQUES

Prof. V.K.Jain, Mechanical Engineering

44
Department, I.I.T. Kanpur
 Fig: Magnetic Abrasive finishing (MAF) AFF
Ref: MAF (Kremen,1994) MRF

R-MRAFF PROCESS

MRF AFM ROTATION

Fig: Chemo mechanical Polishing (CMP) MRAFF

Ref: T.H. Tsai et al. (2003) Ref: MRAFF (Jha and Jain, 2004)
Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 45
 MAF AFM MRAFF
FLEXIBLE MAGNETIC Ft
Magnetic Electromagnet pole
ABRASIVE BRUSH F
lines of
force n

Workpiece N

Fmn
S
N S
Abrasive Media
Chip Abrasiv
Fc Workpiece
e
Workpiece Workpiece
fixture
Fn Workpiece Motion
Ft surface CIP Chain

MRF CMP MFP

Spindle
Drive
Shaft

Magnetic
Fluid &
Abrasives

Ceramic balls

Float Al Base
NS N S N S
Steel yoke
S NS NS N
46
Magnets
 THEMAL ADVANCED MICROMACHINING PROCESSES

• Electro discharge micromachining (EDMM)

• Electro beam micromachining (EBMM)

• Laser beam micromachining (LBMM)

• Plasma arc micromachining (PAMM)

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 47

PRINCIPLE OF THERMAL ADVANCED MICROMACHINING
PROCESSES

 Localized intense heat is produced which increases

temperature of the workpiece in a narrow zone ( i.e. Beam
diameter), equal to its melting or vaporization temperature

 Which leads to removal of material at micro/ nano level in the

form of debris (irregular shaped particles or spherical globules).

*Note: In case of IBM. An ion hits an atom at the top surface of the
workpiece, and removes the material atom by atom or in the groups of
atoms. There is no thermal damage to the workpiece in IBM

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 48

 APPLICATION

 Some of the products include devices such as computer hard disc drive

heads, inkjet printer heads, sensors, infrared images.

 Micro electro-mechanical systems required fabrication with μm/ nm

tolerance which is possible with these techniques.

 Reduced hole diameter, lower hole pitch and longer head can be

manufactured by thermal micromachining processes.

 Marking and engraving.

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 49

 ELECTROCHEMICAL & CHEMICAL ADVANCED
MICROMACHINING PROCESSES

• Electrochemical micromachining (ECMM)

• Chemical micromachining (ChMM)

• Photo Chemical micromachining (PCMM)

• Electrochemical spark machining (ECSM)

• Electrochemical microdeburring (ECMDe)

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 50

 PRINCIPLE OF ELECTROCHEMICAL & CHEMICAL MICRO
MACHINING PROCESSES
Electrochemical micromachining:
Electrochemical metal removal is an alternate wet etching process
 The workpiece made an anode and the tool as cathode in an electrolyte cell in
which a nontoxic salt solution is used as an electrolyte
 Controlled metal removal takes place when the current (smooth D.C. Or Pulse
D.C.) flow through the electrolytic cell
Chemical micromachining:
 It is an ancient process being used for engraving the metal for making
ornaments and other products.
 It removes material in a controlled manner by the application of maskant and
etchant.
 Maskant does not allow etchant to reach & react with work piece to dissolve it.
 Etchant dissolves workpiece material by chemical action.
Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 51
 ADVANCED NANO FINISHING PROCESSES

• To finish surfaces to nano level, it is required to remove material

in the form of atoms or molecules individually or in groups

• Some processes such as Elastic Emission Machining (EEM) and

Ion Beam Machining (IBM) work directly by removing atoms and
molecules from the workpiece surfaces

• While other processes based on finishing by abrasives, remove

them (atoms and molecules) In clusters

• Most of the nano finishing processes are using abrasive particles

either suspended in liquid or held by the viscoelastic material,
carbonyl iron particles, or by magnetorheological fluid As a carrier

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 52

 Micro/Nano Finishing

Traditional Advanced /Non-traditional

AFM
Grinding
MRF

Lapping MAF

Honing MRAFF

CMP

MFP

ELID

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 53

CLASSIFICATION OF ADVANCED NANO-FINISHING PROCESSES

• Advanced abrasive finishing processes

 Abrasive Flow Machining (AFM)

• Magnetic field assisted advanced finishing processes (AFPs)

 Magnetic Abrasive Finishing (MAF)

 Magnetic Float Polishing (MFP)
 Magnetorheological Finishing (MRF)
 Magnetorheological Abrasive Flow Finishing (MRAFF)

• Magnetorheological fluids based processes

 Rotational Magnetorheological Abrasive Flow Finishing (R-MRAFF)

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 54

 Fig: Abrasive Flow maching (AFM), Fig: Magnetic Flot Polishing (MFP)
Ref: AFM (Jain and Jain, 1998) Ref: MFP (Komanduri et al., 1997)

Magnetorheological finishing (MRF), Magnetorheological Abrasive flow finishing (MRAFF)

Ref: MRF Kordonski and Jacobs1 et al., 1996 (QED technologies) Ref: MRAFF (Jha and Jain, 2004)

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 55

 MAGNETO RHEOLOGICAL ABRASIVE FINISHING (MRF)
• MRF was invented at the Luikov Institute of Heat & Mass
Transfer in Minsk, Belarus in the late 1960s by a team led
by William Kordonski and after some time they form a
company known as QED technology.

• MRF is a deterministic and magnetic field assisted

precision finishing Process.

• MRF uses MRP fluid which is invented by Rabinow in

late 1940s consist of
 CIP (Magnetic),
 Abrasive Particle (Non-magnetic)
 carrier liquid (Oil or water)
 additives (glycerol, grease ) Fig : (a) Vertical wheel MRF machine
Ref: Kordonski and Jacobs1 et al., 1996 (QED technologies)

• MRP fluid works as polishing tool.

Application:
•MRF has been used for finishing a large variety of brittle
material ranging from optical glasses to hard crystals.
Limitations:
 Internal and specially complex surfaces can’t be
finished.

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 56

 ABRASIVE FLOW MACHINING (AFM)
• AFM was developed by Extude Home
Corporation USA in1960 as a method to deburr
,polish and radius difficult to reach surface like
intricate geometries and edges by flowing a
abrasive laden viscoelastic medium over them

Key Component

Abrasive
Machine Tooling
Medium
Process Input Parameters
Abraisive Flow maching (AFM),
Ref: AFM (Jain and Jain, 1998)

Grit Composition Tolling &

Extrusion Pressure No. Of Cycle
& Type Fixture

Application:
 In industries such as Aerospace, medical , Limitations:
electronics, Automotive, Precision dies and  AFF is not deterministic process.
mould as part Of manufacturing activities.
Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 57
 CHEMO MECHANICAL POLISHING (CMP)

Chemical Mechanical Chemo

mechanical

• Basically CMP is used to polish the silicon wafer.

• Wafer-Pad-Slurry interactions
Why CMP:
 Local planarization
 Global planarization
Principle:
 CMP uses both chemical and mechanical type or
material removal mechanism.
 Chemical reaction to soften material and than
mechanically polish off this layer.
 Mechanical removal takes place due to abrading.
Limitations:
• This process is used for only flat surfaces.
• CMP is not deterministic in nature. Fig: Chemo mechanical Polishing (CMP)
Ref: T.H. Tsai et al. (2003)
• Endpoint of CMP is a difficult to control for a desired
thickness.

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 58

 MAGNETIC ABRASIVE FINISHING (MAF)
• In MAF finishing is performed by the application of
magnetic field across the gap b/w the workpiece surface
& the rotating electromagnet pole.

• The magnetic abrasive particles are attracted with

each other magnetically b/w magnetic poles along the
lines of magnetic force forming a flexible magnetic
abrasive brush.

• MAF was developed to produce efficiently and

economically good quality finish on the internal and
external surface of tubes as well as flat surface made
of magnetic or non-magnetic material.

 8-9 nm surface roughness value (High surface finish)

Magnetic Abrasive finishing (MAF)
Limitations: Ref: MAF (Kremen,1994)
 MAF can not be used for complex geometries.

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 59

 MAGNETIC FLOAT POLISHING (MFP)
• Magnetic Float Polishing is a technique based on
the Magnetohydrodynmic behaviour of the magnetic
fluid which in the presence of magnetic field can
levitate a non-magnetic float and abrasive particles
suspended in it.

• The forces applied by abrasives are extremely

small and controllable.

• When the magnetic field is applied the

ferromagnetic particle in the ferrofluid are attracted
downward to the area of higher magnetic field and Fig: Magnetic Float Polishing
(MFP)
upward buoyant force is exerted on all non-magnetic Ref: MFP (Komanduri et al.,
materials to push them to the area of lower magnetic 1997)
field.

• The balls are polished by the abrasive particles

mainly due to the action of the magnetic buoyancy
force when the spindle rotates

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 60

MAGNETORHEOLOGICAL ABRASIVE FLOW FINISHING (MRAFF)
• MRAFF is the hybrid finishing
process to take the advantage of
both the finishing process (MRF &
AFF)

• It is deterministic process.

• Any complex geometries can be

finished by this process.
MRF AFF
Limitations:
 Non-uniform surface Why finish in
low finishing
case of freeform surfaces. rate…..?

• Interaction
Low finishing rate. the NMAPs and w/p surface
between
is in straight line sweeping a small area.

• Only a small percentage of NMAPs which are on

the periphery of the slug actually participate in the
finishing operations. MRAFF
Ref: MRAFF (Jha and Jain, 2004)
Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 61
 ROTATIONAL - MAGNETORHEOLOGICAL ABRASIVE FLOW
FINISHING (R-MRAFF)
How to R-MRAFF PROCESS
increase the Solution
finishing rate
in MRAFF
MRF AFM ROTATION

• One way to rotate the medium itself and

other is to rotate the workpiece.

• Rotation of w/p is difficult in the existing

MRAFF setup with the possible chance of
leakage of medium.

• Therefore , rotation of the medium has

been adopted by rotating the magnetic field
and the process is named as (R-MRAFF).
Fig: (a) Schematic of Fe particles chains structure and
(b) force components acting on NMAP in R-MRAFF process
Ref: V K Jain, et al. (2012)

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 62

• The magnetic iron particles chains structure and force components acting on
each NMAP During R-MRAFF process.
• The magnitude of magnetic force (Fm) acting on the individual NMAP is given as
follows.
Where
Fm = mµo χm H∆H m = Mass of a magnetic (Fe) particle
χm = mass susceptibility of the magnetic particle
H = magnetic field intensity
∆H = Gradient of the magnetic field intensity
µo = Permeability in free space

• Axial force (Fa) and radial force (Fr) act on the NMAP due To the reciprocation of
the medium by hydraulic unit.

• Axial force (Fa) is responsible for shearing.

• Radial force (Fr) helps the NMAP in indenting the w/p surface.

• Axial force (Fa) and Radial force (Fr) are proportional to the hydraulic extrusion
pressure (P) and medium viscosity (Fr=kP) where k is constant.

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 63

 • The total normal indentation force is the sum of these three forces

Findentation = Fm + Fcen+ Fr

Where
Fm = Magnetic force
Fcen = Centrifugal force
Fr = Radial force
Ft = Tangential cutting
force

Fcen = mrῳ2
Fc = Fa+ Ft
Ft= 2m(ῳ×v)
• The resultant force Fc makes the active NMAP to follow a helical
path. It produces cross-hatch pattern on the finished Surface

• Helix angle and length of the helical path traversed by an active

NMAP around the w/p are the function of rotational speed and axial
velocity of the medium.

Higher finishing rate comparison to MRAFF.

Ref: V K Jain et al. (2012)

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 64

Thank You

65
 Methods of Micro fabrication

Material deposition Material removal

Traditional material
Electro chemical spark deposition
removal processes
Advanced material
Electro discharge deposition
removal processes

Chemical vapour deposition

Physical vapour deposition

Rapid prototyping / rapid tooling

LIGA

66
 Micro machining

Traditional Advanced

µ-turning µ-EDM

µ-milling µ-ECM

µ-drilling µ-AJM

LBM

EBM

µ-USM

67
CLASSIFICATION OF MICROMACHINING
µ-MACHINING
METHODS

TRADITIONAL ADVANCED

µ-MILLING MECHANICAL THERMAL CHEMICAL &

ELECTROCHEMICAL

µ-TURNING µ-USM µ-LBM ECMM

µ-DRILLING µ-AWJC µ-EBM PCMM

µ-GRINDING, etc µ-AFM, µ-MAF, µ- µ-PAM

MRAFF, µ-MPF

HYBRID

µ-ECSM µ-ECG µ-EDG ELID, etc. 68

CLASSIFICATION OF MICROMACHINING AND NANOFINING
PROCESSES
Advanced Micromachining processes

μ-machining μ-/nano finishing

AFM
Mechanical Beam energy based Chemical & electro-
μ-machining MAF
μ-machining chemical μ-machining
MRAFF
EBM PCMM
USM MFP
LBM ECMM EEM
AJM
EDM MRF
AWJM
IBM

WJM PBM

Hybrid μ-/nano machining

ECSM EDG ELID CMP Etc.

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 69

PRINCIPLE OF MECHANICAL ADVANCED MICROMACHINING

Fig: Abrasive water jet micromachining Fig: Water jet micromachining

Fig: Abrasive jet micromachining Fig: Ultrasonic micromachining

Prof. V.K. Jain Dept. Of Mechanical Engineering IIT Kanpur 70

 MEMS (Micro-electromechanical Systems)

MEMS has been identified as one of the most promising technologies for the 21 st Century and has
the potential to revolutionize both industrial and consumer products by combining silicon-based
microelectronics with micromachining technology. Its techniques and microsystem-based devices
have the potential to dramatically affect of all of our lives and the way we live.

This report presents a general introduction to the field of MEMS, with emphasis on its commercial
applications and device fabrication methods. It also describes the range of MEMS sensors and
actuators, the phenomena that can be sensed or acted upon with MEMS devices, and outlines the
major challenges facing the industry.

1. Introduction

This report deals with the emerging field of micro-electromechanical systems, or MEMS.
MEMS is a process technology used to create tiny integrated devices or systems that combine
mechanical and electrical components. They are fabricated using integrated circuit (IC) batch
processing techniques and can range in size from a few micrometers to millimetres. These
devices (or systems) have the ability to sense, control and actuate on the micro scale, and
generate effects on the macro scale.

The interdisciplinary nature of MEMS utilizes design, engineering and manufacturing

expertise from a wide and diverse range of technical areas including integrated circuit
fabrication technology, mechanical engineering, materials science, electrical engineering,
chemistry and chemical engineering, as well as fluid engineering, optics, instrumentation and
packaging. The complexity of MEMS is also shown in the extensive range of markets and
applications that incorporate MEMS devices. MEMS can be found in systems ranging across
automotive, medical, electronic, communication and defence applications. Current MEMS
devices include accelerometers for airbag sensors, inkjet printer heads, computer disk drive
read/write heads, projection display chips, blood pressure sensors, optical switches, micro
valves, biosensors and many other products that are all manufactured and shipped in high
commercial volumes.

MEMS has been identified as one of the most promising technologies for the 21 st Century and
has the potential to revolutionize both industrial and consumer products by combining silicon-
based microelectronics with micromachining technology. Its techniques and micro system-
based devices have the potential to dramatically affect of all of our lives and the way we live.
If semiconductor micro fabrication was seen to be the first micro manufacturing revolution,
MEMS is the second revolution.

This report introduces the field of MEMS and is divided into four main sections. In the first
section, the reader is introduced to MEMS, its definitions, history, current and potential
applications, as well as the state of the MEMS market and issues concerning miniaturization.
The second section deals with the fundamental fabrication methods of MEMS including
photolithography, bulk micromachining, surface micromachining and high-aspect-ratio
 micromachining; assembly, system integration and packaging of MEMS devices is also
described here. The third section reviews the range of MEMS sensors and actuators, the
phenomena that can be sensed or acted upon with MEMS devices, and a brief description of
the basic sensing and actuation mechanisms. The final section illustrates the challenges
facing the MEMS industry for the commercialization and success of MEMS.

2. Micro-electromechanical Systems (MEMS)

What is MEMS?

Micro-electromechanical systems (MEMS) is a process technology used to create tiny

integrated devices or systems that combine mechanical and electrical components. They are
fabricated using integrated circuit (IC) batch processing techniques and can range in size from
a few micrometers to millimeters. These devices (or systems) have the ability to sense,
control and actuate on the micro scale, and generate effects on the macro scale.

MEMS, an acronym that originated in the United States, is also referred to as Microsystems
Technology (MST) in Europe and Micro machines in Japan. Regardless of terminology, the
uniting factor of a MEMS device is in the way it is made. While the device electronics are
fabricated using ‘computer chip’ IC technology, the micromechanical components are
fabricated by sophisticated manipulations of silicon and other substrates using
micromachining processes. Processes such as bulk and surface micromachining, as well as
high-aspect-ratio micromachining (HARM) selectively remove parts of the silicon or add
additional structural layers to form the mechanical and electromechanical components. While
integrated circuits are designed to exploit the electrical properties of silicon, MEMS takes
advantage of either silicon’s mechanical properties or both its electrical and mechanical
properties.
In the most general form, MEMS consist of
mechanical microstructures, microsensors,
microactuators and microelectronics, all integrated
onto the same silicon chip. This is shown
schematically in Figure 1.
Microsensors detect changes in the system’s
environment by measuring mechanical, thermal,
magnetic, chemical or electromagnetic information
or phenomena. Microelectronics process this
information and signal the microactuators to react
Figure 1. Schematic illustration of
and create some form of changes to the environment.
MEMS components.
MEMS devices are very small; their components are usually microscopic. Levers, gears,
pistons, as well as motors and even steam engines have all been fabricated by MEMS (Figure
2). However, MEMS is not just about the miniaturization of mechanical components or
making things out of silicon (in fact, the term MEMS is actually misleading as many
micromachined devices are not mechanical in any sense). MEMS is a manufacturing
technology; a paradigm for designing and creating complex mechanical devices and systems
as well as their integrated electronics using batch fabrication techniques.

Figure 2. (a) A MEMS silicon motor together with a strand of human hair [1], and (b)
the legs of a spider mite standing on gears from a micro-engine [2 - Sandia National
Labs, SUMMiT *Technology, http://mems.sandia.gov].

From a very early vision in the early 1950’s, MEMS has gradually made its way out of
research laboratories and into everyday products. In the mid-1990’s, MEMS components
began appearing in numerous commercial products and applications including accelerometers
used to control airbag deployment in vehicles, pressure sensors for medical applications, and
inkjet printer heads. Today, MEMS devices are also found in projection displays and for
micropositioners in data storage systems. However, the greatest potential for MEMS devices
lies in new applications within telecommunications (optical and wireless), biomedical and
process control areas.
MEMS has several distinct advantages as a manufacturing technology. In the first place, the
interdisciplinary nature of MEMS technology and its micromachining techniques, as well as
its diversity of applications has resulted in an unprecedented range of devices and synergies
across previously unrelated fields (for example biology and microelectronics). Secondly,
MEMS with its batch fabrication techniques enables components and devices to be
manufactured with increased performance and reliability, combined with the obvious
advantages of reduced physical size, volume, weight and cost. Thirdly, MEMS provides the
basis for the manufacture of products that cannot be made by other methods. These factors
make MEMS potentially a far more pervasive technology than integrated circuit microchips.
However, there are many challenges and technological obstacles associated with
miniaturization that need to be addressed and overcome before MEMS can realize its
overwhelming potential.
Definitions and Classifications

This section defines some of the key terminology and classifications associated with MEMS.
It is intended to help the reader and newcomers to the field of micromachining become
familiar with some of the more common terms. A more detailed glossary of terms has been
included in Appendix A.

Figure 3 illustrates the classifications of microsystems technology (MST). Although MEMS

is also referred to as MST, strictly speaking, MEMS is a process technology used to create
these tiny mechanical devices or systems, and as a result, it is a subset of MST.

Figure 3. Classifications of microsystems technology

 Micro-optoelectro mechanical systems (MOEMS) is also a subset of MST and together with
MEMS forms the specialized technology fields using miniaturized combinations of optics,
electronics and mechanics. Both their Microsystems incorporate the use of microelectronics
batch processing techniques for their design and fabrication. There are considerable overlaps
between fields in terms of their integrating technology and their applications and hence it is
extremely difficult to categories MEMS devices in terms of sensing domain and/or their
subset of MST. The real difference between MEMS and MST is that MEMS tends to use
semiconductor processes to create a mechanical part. In contrast, the deposition of a material
on silicon for example, does not constitute MEMS but is an application of MST.

Transducer
A transducer is a device that transforms one form of signal or energy into another form. The
term transducer can therefore be used to include both sensors and actuators and is the most
generic and widely used term in MEMS.

Sensor
A sensor is a device that measures information from a surrounding environment and provides
an electrical output signal in response to the parameter it measured. Over the years, this
information (or phenomenon) has been categorized in terms of the type of energy domains but
MEMS devices generally overlap several domains or do not even belong in any one category.
These energy domains include:
 Mechanical - force, pressure, velocity, acceleration, position
 Thermal - temperature, entropy, heat, heat flow
 Chemical - concentration, composition, reaction rate
 Radiant - electromagnetic wave intensity, phase, wavelength, polarization
reflectance, refractive index, transmittance
 Magnetic - field intensity, flux density, magnetic moment, permeability
 Electrical - voltage, current, charge, resistance, capacitance, polarization [4,5,6,7]

Actuator
An actuator is a device that converts an electrical signal into an action. It can create a force to
manipulate itself, other mechanical devices, or the surrounding environment to perform some
useful function.

Applications

Today, high volume MEMS can be found in a diversity of applications across multiple
markets (Table 1).

Table 1. Applications of MEMS [10].

Automotive Electronics Medical Communications Defence

Internal Fibre-optic
Blood pressure Munitions
navigation Disk drive heads network
sensor guidance
sensors components
Air conditioning Muscle RF Relays,
Inkjet printer
compressor stimulators & drug switches and Surveillance
heads
sensor delivery systems filters
 Projection
Brake force
displays in
sensors & Projection
Implanted portable
suspension screen Arming systems
pressure sensors communications
control televisions
devices and
accelerometers
instrumentation
Fuel level and
Earthquake Voltage controlled Embedded
vapour pressure Prosthetics
sensors sensors oscillators (VCOs) sensors
Avionics Miniature
Splitters and
Airbag sensors pressure analytical Data storage
couplers
sensors instruments
Mass data
"Intelligent" tyres storage systems Pacemakers Tuneable lasers Aircraft control

As an emerging technology MEMS products are centred around technology-product

paradigms rather than product-market paradigms. Consequently, a MEMS device may find
numerous applications across a diversity of industries. For example, the MEMS inkjet printer
head nozzle in widespread use today has developed from a nozzle originally used in nuclear
separation. The commercialization of selected MEMS devices is illustrated in Table 2.

Table 2. Commercialization of selected MEMS devices [11].

Cost Reduction/
Discovery Full
Product Evolution Application
Commercialisation
Expansion
Pressure sensors 1954-1960 1960-1975 1975-1990 1990-present
Accelerometers 1974-1985 1985-1990 1990-1998 1998
Gas sensors 1986-1994 1994-1998 1998-2005 2005
Valves 1980-1988 1988-1996 1996-2002 2002
Nozzles 1972-1984 1984-1990 1990-1998 1998
Photonics/displays 1980-1986 1986-1998 1998-2004 2004
Bio/Chemical sensors 1980-1994 1994-1999 1999-2004 2004
RF switches 1994-1998 1998-2001 2001-2005 2005
Rate (rotation) sensors 1982-1990 1990-1996 1996-2002 2002
Micro relays 1977-1982 1993-1998 1998-2006 2006

It is not within the scope of this report to detail all the current and potential applications
within each market segment. Instead, a selection of the most established MEMS devices is
detailed along with the most potentially significant future applications.
Established MEMS Applications

i) Automotive airbag sensor

Automotive airbag sensors were one of the first commercial devices using MEMS. They are
in widespread use today in the form of a single chip containing a smart sensor, or
accelerometer, which measures the rapid deceleration of a vehicle on hitting an object. The
deceleration is sensed by a change in voltage. An electronic control unit subsequently sends a
signal to trigger and explosively fill the airbag.
Initial air bag technology used conventional mechanical ‘ball and tube’ type devices which
were relatively complex, weighed several pounds and cost several hundred dollars. They
were usually mounted in the front of the vehicle with separate electronics near the airbag.
MEMS has enabled the same function to be accomplished by integrating an accelerometer and
the electronics into a single silicon chip, resulting in a tiny device that can be housed within
the steering wheel column and costs only a few dollars (Figures 4 and 5).
The accelerometer is essentially a capacitive or piezoresistive device consisting of a
suspended pendulum proof mass/plate assembly. As acceleration acts on the proof mass,
micromachined capacitive or piezoresistive plates sense a change in acceleration from
deflection of the plates. The sense plates can be seen in Figure 4.

Figure 4. (a) The first commercial accelerometer from Analog

Devices (1990); its size is less than 1 cm2 (left) [12], and (b)
capacitive sense plates, 60 microns deep (right) [13].

Figure 5. Modern day MEMS accelerometer (left), and

the fully packaged device (right) [12].
The airbag sensor is fundamental to the success of MEMS and micromachining technology.
With over 60 million devices sold and in operation over the last 10 years and operating in
such a challenging environment as that found within a vehicle, the reliability of the
technology has been proven. An example of this success is today’s vehicles – the BMW 740i
has over 70 MEMS devices including anti-lock braking systems, active suspension, appliance
and navigation control systems, vibration monitoring, fuel sensors, noise reduction, rollover
detection, seatbelt restraint and tensioning etc. As a result, the automotive industry has
become one of the main drivers for the development of MEMS for other equally demanding
environments. Some of the leading airbag accelerometer manufacturers include Analog
Devices, Motorola, SensorNor and Nippondenso.
Accelerometers are not just limited to automotive applications. Earthquake detection, virtual
reality video games and joysticks, pacemakers, high performance disk drives and weapon
systems arming are some of the many potential uses for accelerometers.

ii) Medical pressure sensor

Another example of an extremely successful MEMS application is the miniature disposable
pressure sensor used to monitor blood pressure in hospitals. These sensors connect to a
patients intravenous (IV) line and monitor the blood pressure through the IV solution. For a
fraction of their cost ($10), they replace the early external blood pressure sensors that cost
over $600 and had to be sterilized and recalibrated for reuse. These expensive devices
measure blood pressure with a saline-filled tube and diaphragm arrangement that has to be
connected to an artery with a needle.

Figure 6. Schematic illustration of a piezoresistive pressure sensor.

The disposable sensor consists of a silicon substrate which is etched to produce a membrane
and is bonded to a substrate (Figure 6). A piezoresistive layer is applied on the membrane
surface near the edges to convert the mechanical stress into an electrical voltage. Pressure
corresponds to deflection of the membrane. The sensing element is mounted on a plastic or
ceramic base with a plastic cap over it, designed to fit into a manufacturer’s housing (Figure
7). A gel is used to separate the saline solution from the sensing element.

As in the case of the MEMS airbag sensor, the disposable blood pressure sensor has been one
of the strongest MEMS success stories to date. The principal manufacturers being Lucas
Novasensor, EG & G IC Sensors and Motorola with over 17 millions units per year. More
recently, the technology from the blood pressure sensor has been taken a step further in the
development of the catheter-tip pressure sensor. This considerably smaller MEMS device is
designed to fit on the tip of a catheter and measure intravascular pressure (its size being only
0.15 mm x 0.40 mm x 0.90 mm).

Pressure sensors are the biggest medical MEMS application to date with the accelerometer
MEMS a distant second. Although the majority of these accelerometer applications remain
under development, advanced pacemaker designs include a MEMS accelerometer device that
measures the patient’s activity. The technology, similar to that found in the airbag sensor,
enables the patient’s motion and activity to be monitored and signals the pacemaker to adjust
its rate accordingly.

Figure 7. (a) Disposable blood pressure sensor connected to an IV line [14],

(b) disposable blood pressure sensors (as shipped) [15], and (c) intracardial
catheter-tip sensors for monitoring blood pressure during cardiac
catheterisation, shown on the head of a pin [13].

iii) Inkjet printer head

One of the most successful MEMS applications is the inkjet printer head, superseding even
automotive and medical pressure sensors. Inkjet printers use a series of nozzles to spray
drops of ink directly on to a printing medium. Depending on the type of inkjet printer the
droplets of ink are formed in different ways; thermally or piezoelectrically.
Invented in 1979 by Hewlett-Packard, MEMS thermal inkjet printer head technology uses
thermal expansion of ink vapour. Within the printer head there is an array of tiny resistors
known as heaters. These resistors can be fired under microprocessor control with electronic
pulses of a few milliseconds (usually less than 3 microseconds). Ink flows over each resistor,
which when fired, heat up at 100 million ºC per second, vaporizing the ink to form a bubble.
As the bubble expands, some of the ink is pushed out of a nozzle within a nozzle plate,
landing on the paper and solidifying almost instantaneously. When the bubble collapses, a
vacuum is created which pulls more ink into the print head from the reservoir in the cartridge
(Figure 8). It is worth noting there are no moving parts in this system (apart from the ink
itself) illustrating that not all MEMS devices are mechanical.

Figure 8. Thermal inkjet print technology [16].

A piezoelectric element can also be used to force the ink through the nozzles (Figure 9). In
this case, a piezoelectric crystal is located at the back of the ink reservoir of each nozzle. The
piezoelectric crystal element receives a very small electric charge causing it to vibrate. When
it vibrates inwards it forces a tiny amount of ink out of the nozzle. As the element vibrates
back out, it pulls some more ink into the reservoir to replace the ink that was sprayed out.
Epson patented this technology but it is also used by the majority of the other leading printer
companies.

MEMS has enabled more and more heating elements and piezoelectric crystals to be
incorporated into a printer head. Early printers had 12 nozzles with resolutions of up to 92
dpi possible. Today, modern inkjet printers have up to 600 nozzles which can all fire a
droplet simultaneously enabling 1200 dpi. Epson, Lexmark, Hewlett-Packard, Olivetti, Xerox
and Canon all use a form of these MEMS in their inkjet printers. Over 350 million units were
sold in 2000.
Figure 9. Thermal inkjet print technology [16].
iv) Overhead projection display
One of the early MEMS devices used for a variety of display applications is the Digital Micro
mirror Device (DMD) from Texas Instruments. The device contains over a million tiny pixel-
mirrors each measuring 16 µm by 16 µm and capable of rotating by ±10º, over 1000 times a
second (Figure 10). Light from a projection source impinges on the pupil of the lens (or
mirror) and is reflected brightly onto a projection screen. DMD’s are used for displays for PC
projectors, high definition televisions (HDTV’s) and for large venues such as digital cinemas
where traditional liquid crystal technology cannot compete. MEMS has enabled the micro
mirrors to be only 1 µm apart, resulting in an image taking up a larger percentage (89 percent)
of space on the DMD chip's reflective surface, as compared to a typical LCD (12 to 50
percent). This reduces the pixilation and produces an overall sharper and brighter image.
Today over 30 manufacturers use the DMD (Kodak being the largest) and over 500,000
systems have been shipped.

Figure 10. The MEMS Digital Micromirror Device

(DMD) [17].
New MEMS Applications

The experience gained from these early MEMS applications has made it an enabling
technology for new biomedical applications (often referred to as bioMEMS) and wireless
communications comprised of both optical, also referred to as micro-optoelectromechanical
systems (MOEMS), and radio frequency (RF) MEMS.

i) BioMEMS
Over the past few years some highly innovative products have emerged from bioMEMS
companies for revolutionary applications that support major societal issues including DNA
sequencing, drug discovery, and water and environmental monitoring. The technology
focuses on microfluidic systems as well as chemical testing and processing and has enabled
devices and applications such as ‘lab-on-a-chip’, chemical sensors, flow controllers,
micronozzles and microvalves to be produced. Although many devices are still under
development, microfluidic systems typically contain silicon micromachined pumps, flow
sensors and chemical sensors. They enable fast and relatively convenient manipulation and
analysis of small volumes of liquids, an area of particular interest in home-based medical
applications where patients can use devices to monitor their own conditions, such as blood
and urine analysis.

One example of a new bioMEMS device is the microtitreplate on which a number of cavities
can be simultaneously filled accurately and repeatably by capillary force (Figure 11a). This is
a relatively simple MEMS product in the form of a piece of plastic with high-aspect-ratio
micromachined microchannels and is classified as a ‘lab-on-a-chip’ product. Its dimensions
are only 20 mm x 37 mm x 3 mm and enables automatic filling of 96 microwells by the use
of capillary action.

Figure 11. (a) Micromachined microtitreplate with 96 cavities filled by capillary

force [18,19], and (b) a bioMEMS device actuated with ‘microteeth’ to trap,
hold and release single red blood cells (unharmed). The little balls in the
channels are red blood cells [2].

Future lab-on-a-chip technology may include implantable ‘pharmacy-on-a-chip’ devices to

carefully release drugs into the body from tiny chambers embedded in a MEMS device,
eliminating the need for needles or injections. The delivery of insulin is one such application,
as is the delivery of hormones, chemotherapy drugs and painkillers. First generation devices
are being developed which release their medication upon signals from an outside source,
wired through the skin. Proposed second generation devices may be wireless and third
generation MEMS chips could interact with MEMS sensors embedded in the body to respond
to the body’s own internal signals.
One of the most recent MEMS microfluidic devices to emerge from development laboratories
incorporates a ‘Pac-Man’-like microstructure that interacts with red blood cells (Figure 11b).
The device from Sandia National Laboratories, U.S.A, contains silicon microteeth that open
and close like jaws trapping and releasing a single red blood cell unharmed as it is pumped
through a 20 µm channel. The ultimate goal of this device is to puncture cells and inject them
with DNA, proteins, or pharmaceuticals to counter biological or chemical attacks, gene
imbalances and natural bacterial or viral infections.

ii) MOEMS
Optical communications has emerged as the only practical means to address the network
scaling issues created by the tremendous growth in data traffic caused by the rapid rise of the
Internet. Current routing technology slows the information (or bit) flow by transforming
optical signals into electronic information and then back into light before redirecting it. All
optical networks offer far superior throughput capabilities and performance over traditional
electronic systems.

The most significant MOEMS device products include waveguides, optical switches, cross
connects, multiplexers, filters, modulators, detectors, attenuators and equalizers. Their small
size, low cost, low power consumption, mechanical durability, high accuracy, high switching
density and low cost batch processing of these MEMS-based devices make them a perfect
solution to the problems of the control and switching of optical signals in telephone networks.
An example of a MEMS optical connect is shown in Figure 12. Here a network of 256
MEMS micromirrors route information in the form of photons (the elementary particle that
corresponds to an electromagnetic wave) to and from any of 256 input/output optical fibres.

Figure 12. A MEMS optical cross connect consisting of an array of microscopic mirrors, each the size
of a pin head and able to tilt in various directions to steer light [20,21].

MEMS fabrication processes have reached the stage where mass manufacture of such devices
is now practical. A typical optical switch can cost over $1000, but using MEMS, the same
level of functionality can be achieved for less than a dollar. Agere Systems (previously
known as the microelectronics division of Lucent Technologies), Corning, JDS Uniphase and
Sycamore Networks are some of the leading companies in this field.

iii) RF MEMS
RF MEMS is one of the fastest growing areas in commercial MEMS technology. RF MEMS
are designed specifically for electronics in mobile phones and other wireless communication
applications such as radar, global positioning satellite systems (GPS) and steerable antennae.
MEMS has enabled the performance, reliability and function of these devices to be increased
while driving down their size and cost at the same time (Figure 13).
 Figure 13. (a) A miniature acoustic resonator, shown in the
foreground, is one-fifth the size of a traditional component used in
mobile phones and other wireless communications devices [22],
and (b) on-chip micro-microphones may make it possible to build
radios on a chip [20,21].

The technology includes circuit tuning elements (capacitors/inductors, resonators, filters,

microphones and switches). These low-loss ultra-miniature and highly integrative RF
functions can and will eventually replace classical RF elements and enable a new generation
of RF devices. As it can be seen today, if RF MEMS components continue to replace
traditional components in today’s mobile phones, then phones could become extremely small
(the size of wristwatch is not too far away), require little battery power and may even be
cheaper.

MEMS Market

The three most well known market studies are the Network of Excellence in Multifunctional
Microsystems (NEXUS) study (1998), the System Planning Corporation (SPC) study (1999)
and the Battelle study (1990) and there is discrepancy between each study [23, 24, 25
respectively]. The size of the MEMS market (M3) is contingent on how MEMS is defined
(M3 is shorthand for MEMS, Microsystems and Micromachining and although it is not yet
common, it is used as a reference for the entire MEMS market. Smaller M3 figures are
obtained if MEMS is considered as just micromachining, which is more elemental and at the
device level. Alternatively, much larger M3 figures arise if MEMS is examined at the system
or subsystem level (as in the case of NEXUS). Depending on the study under review, the M 3
market today ranges from $4.2 billion to $14.2 billion. Much of the current market centres on
read/write heads for computer disk drives, pressure sensors, inkjet printer heads and
accelerometers. Table 3 provides the NEXUS worldwide M3 market size in 1996 and
forecasts for 2002 for existing MEMS product types.

Table 3. Worldwide M 3 market size in 1996 and 2002 for existing MEMS product types
in $US millions [23].

1996
$ 2002 Units $
Product Types Units
(millions) (millions) (millions)
(millions)

HDD heads 530 4500 1500 12000

Inkjet print heads 100 4400 500 10000

 Heart pacemakers 0.5 1000 0.8 3700

In vitro diagnostics 700 450 4000 2800

Hearing aids 4 1150 7 2000

Pressure sensors 115 600 309 1300

Chemical sensors 100 300 400 800

Infrared imagers 0.01 220 0.4 800

Accelerometers 24 240 90 430

Gyroscopes 6 150 30 360

Magnetoresistive sensors 15 20 60 60

Microspectrometers 0.006 3 0.15 40

TOTAL 1595 $13,033 6807 $34,290

In the area of emerging MEMS products, Table 4 provides the NEXUS worldwide M 3 market
size in 1996 and forecasts for 2002. Drug delivery systems (microfluidic microdosing
systems), lab-on-a-chip devices and MEMS-based optical switches are predicted to reach
billion dollar market segments by 2002.

Table 4. Worldwide M 3 market size in 1996 and 2002 for emerging MEMS product types
in $US millions [23].

1996
$ 2002 Units
Product Types Units $ (millions)
(millions) (millions)
(millions)
Drug delivery systems 1 10 100 1000
Optical switches 1 50 40 1000
Lab on ship 0 0 100 1000
Magneto optical heads 0.01 1 100 500
Projection valves 0.1 10 1 300
Coil on chip 20 10 600 100
Micro relays 0.1 50 100
Micromotors 0.1 5 2 80
Inclinometers 1 10 20 70
Injection nozzles 10 10 30 30
Anti-collision sensors 0.01 0.5 2 20
Electronic noses 0.001 0.1 0.05 5
TOTAL 33 $107 1045 $4,205

A more recent market study by NEXUS/Roger Grace Associates, shown in Table 5, estimated
the M3 market to be $14.2 billion in 2000, increasing to $30.4 billion by 2004. This
corresponds to a compounded annual growth rate (CAGR) of 21%. Telecommunications is
forecast to be the major growth area, comprised of both optical MEMS and RF MEMS-based
devices.
 Table 5. Worldwide shipment of M3 products by application sector for
2000-2004 in $US millions [23,26].

Application Sector 2000 2004 CAGR(%)

IT/Peripheral $ 8,700 $13,400 11.5
Medical/Biochemical 2,400 7,400 32.5
Industrial/Automation 1,190 1,850 11.6
Telecommunications 130 3,650 128.1
Automotive 1,260 2,350 16.9
Environmental Monitoring 520 1,750 35.4
TOTAL $14,200 $30,400 21.0%

Miniaturization Issues

As previously stated, MEMS is not about miniaturization; it is a manufacturing technology

used to create tiny integrated microdevices and systems using IC batch fabrication techniques.
Similarly, miniaturization is not just about shrinking down existing devices (although there
have been some classic examples, namely the DENSO Micro-Car as shown in Figure 14); it’s
about completely rethinking the structure of a microsystem.

Figure 14. The DENSO Micro-Car is a miniature version of Toyota’s first passenger car. Fabricated
using MEMS, at 1/1000th the size of the original, it consists of a 0.67 mm magnetic-type working
motor and when supplied with 3 V 20 mA of alternating current through a 18 µm copper wire, the
engine runs at 600 rpm equivalent to 5-6 mm/s [27].

In order to manufacture a successful MEMS device basic physics and operating principles
including scaling laws need to be fully understood and appreciated at both a macro and
microlevel. Sometimes no advantages in terms of performance, size/weight, reliability and
cost can be gained with a MEMS device. Increased surface area (S) to volume (V) ratios at
microscales have both considerable advantages and disadvantages (Figure 15).

Figure 15. Effect of miniaturization on surface area and volume.

Some of these micro level issues include:

 Friction is greater than inertia. Capillary, electrostatic and atomic forces as well as
stiction at a micro-level can be significant.
 Heat dissipation is greater than heat storage and consequently thermal transport
properties could be a problem or, conversely, a great benefit.
 Fluidic or mass transport properties are extremely important. Tiny flow spaces are
prone to blockages but can conversely regulate fluid movement.
 Material properties (Young’s modulus, Poisson’s ratio, grain structure) and
mechanical theory (residual stress, wear and fatigue etc.) may be size dependent.
 Integration with on-chip circuitry is complex and device/domain specific. Lab-on-a-
chip systems components may not scale down comparably.
 Miniature device packaging and testing is not straightforward. Certain MEMS sensors
require environmental access as well as protection from other external influences.
Testing is not rapid and is expensive in comparison with conventional IC devices.
 Cost – for the success of a MEMS device, it needs to leverage its IC batch fabrication
resources and be mass-produced. Hence mass-market drivers must be found to
generate the high volume production.

3. MEMS Fabrication Methods

MEMS fall into three general classifications; bulk micromachining, surface micromachining
and high-aspect-ratio micromachining (HARM), which includes technology such as LIGA (a
German acronym from Lithography, Galvanoformung, Abformung translated as lithography,
electroforming and moulding).

Conventional macro scale manufacturing techniques e.g. injection molding, turning, drilling
etc, are good for producing three dimensional (3D) shapes and objects, but can be limited in
terms of low complexity for small size applications. MEMS fabrication, by comparison, uses
high volume IC style batch processing that involves the addition or subtraction of two
dimensional layers on a substrate (usually silicon) based on photolithography and chemical
etching. As a result, the 3D aspect of MEMS devices is due to patterning and interaction of
the 2D layers. Additional layers can be added using a variety of thin-film and bonding
techniques as well as by etching through sacrificial ‘spacer layers’. Figure 16 shows the
potential complexity of a MEMS system by the addition of independent structural layers.

Figure 16. MEMS device complexity by structural layers [2].

Photolithography

Photolithography is the photographic technique to transfer copies of a master pattern, usually

a circuit layout in IC applications, onto the surface of a substrate of some material (usually a
silicon wafer).

The substrate is covered with a thin film of some material, usually silicon dioxide (SiO 2), in
the case of silicon wafers, on which a pattern of holes will be formed (Figure 17). A thin
layer of an organic polymer, which is sensitive to ultraviolet radiation, is then deposited on
the oxide layer; this is called a photo resist. A photo mask, consisting of a glass plate
(transparent) coated with a chromium pattern (opaque), is then placed in contact with the
photo resist coated surface. The wafer is exposed to the ultraviolet radiation transferring the
pattern on the mask to the photo resist which is then developed in a way very similar to the
process used for developing photographic films. The radiation causes a chemical reaction in
the exposed areas of the photo resist of which there are two types; positive and negative.
Positive photo resist is strengthened by UV radiation whereas negative photo resists
areweakened. On developing, the rinsing solution removes either the exposed areas or the
unexposed areas of photoresist leaving a pattern of bare and photoresist-coated oxides on the
wafer surface. The resulting photoresist pattern is either the positive or negative image of the
original pattern of the photomask.

Figure 17. Photo resist and silicon dioxide patterns following photolithography [28].

A chemical (usually hydrochloric acid) is used to attack and remove the uncovered oxide
from the exposed areas of the photo resist. The remaining photo resist is subsequently
removed, usually with hot sulphuric acid which attacks the photo resist but not the oxide layer
on the silicon, leaving a pattern of oxide on the silicon surface. The final oxide pattern is
either a positive or negative copy of the photo mask pattern and serves as a mask in
subsequent processing steps.

At this point MEMS diverges from traditional IC fabrication. In processing IC devices, the
oxide pattern serves as a mask during the ‘doping’ of the wafer with impurities (such as boron
or phosphorous) that alter the local conduction necessary for microelectronic devices. In
MEMS, the oxide serves as a subsequent mask for either further additional chemical etching
 creating deeper 3D pits or new layers on which to build further layers, resulting in an overall
3D structure or device.

Materials for Micromachining

Substrates
The most common substrate material for micromachining is silicon. It has been successful in
the microelectronics industry and will continue to be in areas of miniaturization for several
reasons:

i) silicon is abundant, inexpensive, and can be processed to unparalleled purity

ii) silicon’s ability to be deposited in thin films is very amenable to MEMS

iii) high definition and reproduction of silicon device shapes using photolithography are
perfect for high levels of MEMS precision
iv) silicon microelectronics circuits are batch fabricated (a silicon wafer contains
hundreds of identical chips not just one)

Other crystalline semiconductors including germanium (Ge) and gallium arsenide (GaAs) are
used as substrate materials due to similar inherent features, but silicon is distinguished from
other semiconductors in that it can be readily oxidized to form a chemically inert and
electrically insulating surface layer of SiO2 on exposure to steam.

The homogeneous crystal structure of silicon gives it the electrical properties needed in
microelectronic circuits, but in this form silicon also has desirable mechanical properties.
Silicon forms the same type of crystal structure as diamond, and although the interatomic
bonds are much weaker, it is harder than most metals. In addition, it is surprisingly resistant
to mechanical stress, having a higher elastic limit than steel in both tension and compression.
Single crystal silicon also remains strong under repeated cycles of tension and compression.

The crystalline orientation of silicon is important in the fabrication of MEMS devices because
some of the etchants used attack the crystal at different rates in different directions (Figure
18).

Figure 18. Low crystallographic index planes of silicon [29].

Silicon is dominant as a substrate for MEMS but research and development is ongoing with
other non-semiconductor substrate materials including metals, glasses, quartz, crystalline
insulators, ceramics and polymers. The ability to integrate circuitry directly onto the substrate
is currently the underlying issue with today’s MEMS substrate materials; hence the success of
silicon.
Additive Films and Materials
The range of additive films and materials for MEMS devices is much larger than the types of
possible substrates and includes conductors, semiconductors and insulators such as:

 silicon - single crystal, polycrystalline and amorphous

 silicon compounds (SixNy, SiO2, SiC etc.)
 metals and metallic compounds (Au, Cu, Al, ZnO, GaAs, IrO x, CdS)
 ceramics (Al203 and more complex ceramic compounds)
 organics (diamond, polymers, enzymes, antibodies, DNA etc.)
Bulk Micromachining

Bulk micromachining involves the removal of part of the bulk substrate. It is a subtractive
process that uses wet anisotropic etching or a dry etching method such as reactive ion etching
(RIE), to create large pits, grooves and channels. Materials typically used for wet etching
include silicon and quartz, while dry etching is typically used with silicon, metals, plastics
and ceramics.

Wet Etching
Wet etching describes the removal of material through the immersion of a material (typically
a silicon wafer) in a liquid bath of a chemical etchant. These etchants can be isotropic or
anisotropic.

Isotropic etchants etch the material at the same rate in all directions, and consequently remove
material under the etch masks at the same rate as they etch through the material; this is known
as undercutting (Figure 19 a and b). The most common form of isotropic silicon etch is HNA,
which comprises a mixture of hydrofluoric acid (HF), nitric acid (HNO 3) and acetic acid
(CH3COOH). Isotropic etchants are limited by the geometry of the structure to be etched.
Etch rates can slow down and in some cases (for example, in deep and narrow channels) they
can stop due to diffusion limiting factors. However, this effect can be minimized by agitation
of the etchant, resulting in structures with near perfect and rounded surfaces (Figure 19a) [4].

Figure 19. Isotropic etching with (a) and without (b)

agitation, and anisotropic wet etching of (100) and (110)
silicon (c and d respectively) [9].

Anisotropic etchants etch faster in a preferred direction. Potassium hydroxide (KOH) is the
most common anisotropic etchant as it is relatively safe to use. Structures formed in the
substrate are dependent on the crystal orientation of the substrate or wafer. Most such
anisotropic etchants progress rapidly in the crystal direction perpendicular to the (110) plane
and less rapidly in the direction perpendicular to the (100) plane. The direction perpendicular
to the (111) plane etches very slowly if at all. Figures 19c and 19d shows examples of
anisotropic etching in (100) and (110) silicon. Silicon wafers, originally cut from a large
 ingot of silicon grown from single seed silicon, are cut according to the crystallographic
plane. They can be supplied in terms of the orientation of the surface plane.

Dopant levels within the substrate can affect the etch rate by KOH, and if levels are high
enough, can effectively stop it. Boron is one such dopant and is implanted into the silicon by
a diffusion process. This can be used to selectively etch regions in the silicon leaving doped
areas unaffected.

Dry Etching
Dry etching relies on vapour phase or plasma-based methods of etching using suitably
reactive gases or vapours usually at high temperatures. The most common form for MEMS is
reactive ion etching (RIE) which utilizes additional energy in the form of radio frequency
(RF) power to drive the chemical reaction. Energetic ions are accelerated towards the
material to be etched within a plasma phase supplying the additional energy needed for the
reaction; as a result the etching can occur at much lower temperatures (typically 150º - 250ºC,
sometimes room temperature) than those usually needed (above 1000ºC). RIE is not limited
by the crystal planes in the silicon, and as a result, deep trenches and pits, or arbitrary shapes
with vertical walls can be etched [4].

Deep Reactive Ion Etching (DRIE) is a much higher-aspect-ratio etching method that involves
an alternating process of high-density plasma etching (as in RIE) and protective polymer
deposition to achieve greater aspect ratios (Figure 20).

Figure 20. Deep Reactive Ion Etching (DRIE) [2,3].

Etch rates depend on time, concentration, temperature and material to be etched. To date
there are no universally accepted master equations to predict etch performance and behavior.

3.4 Surface Micromachining

Surface micromachining involves processing above the substrate, mainly using it as a

foundation layer on which to build. It was initiated in the 1980’s and is the newest MEMS
production technology. Material is added to the substrate in the form of layers of thin films
on the surface of the substrate (typically a silicon wafer). These layers can either by structural
layers or act as spacers, later to be removed, when they are known as sacrificial layers. Hence
the process usually involves films of two different materials: a structural material out of
which the free standing structure is made (generally polycrystalline silicon or polysilicon,
silicon nitride and aluminium) and a sacrificial material, deposited wherever either an open
area or a free standing mechanical structure is required (usually an oxide).
 Figure 21. Surface micromachining of a cantilever beam using a sacrificial layer [2,3].

These layers (or thin films) are deposited and subsequently dry etched in sequence, with the
sacrificial material being finally wet etched away to release the final structure. Each
additional layer is accompanied by an increasing level of complexity and a resulting difficulty
in fabrication. A typical surface micro machined cantilever beam is shown in Figure 21.
Here, a sacrificial layer of oxide is deposited on the silicon substrate surface using a pattern
and photolithography. A polysilicon layer is then deposited and patterned using RIE
processes to form a cantilever beam with an anchor pad. The wafer is then wet etched to
remove the oxide (sacrificial) layer releasing and leaving the beam on the substrate. More
complex MEMS structures can be achieved using structural polysilicon and sacrificial silicon
dioxide, including sliding structures, actuators and free moving mechanical gears. Figures 22
shows the process flow for the fabrication of a micromotor by the commercially available
Multi-User MEMS Process (MUMPS).

Figure 22. Surface micromachining of a MEMS micromotor using the Multi-

User MEMS Process (MUMPS) [30].

The levels of complexity achievable with MEMS has already been shown in Figure 16. In
this case, five mechanical levels of micromachined polysilicon can be achieved using Sandia
Ultra-Planar Multi-Level Technology (SUMMiT).

The success of the surface micromachining process depends on the ability to successfully
remove all of the sacrificial layers to free the structural elements so that they can be actuated.
This step is responsible for curtailing the yield (percentage of the devices on a wafer that
function properly) and reliability of fabricated MEMS due to the phenomenon known as
 stiction. Stiction refers to the sticking of structural elements either to the substrate or the
adjacent elements. Capillary forces from rinsing liquids, as well as electrostatic and van der
Waals forces can also produce permanent adhesion after the system has dried.

3.4.1 Fusion Bonding

In order to form more complex and larger MEMS structures, micromachined silicon wafers
can be bonded to other materials in a process known as fusion bonding. It is a technique that
enables virtually seamless integration of multiple layers and relies on the creation of atomic
bonds between each layer either directly (with heating and pressure in the case of glass to
wafer bonding), or through a thin film of silicon dioxide (Figure 23). The resulting composite
has very low residual stress due to matching coefficients of thermal expansion from each
layer. In addition, the mechanical strength of the bond is comparable to that of the adjoining
layers resulting in a very strong composite fabrication technique for enclosed cavities and
channels. High-aspect ratio structural layers can also be bonded to silicon substrates in a
similar manner. Photoresist and polymethylmethacrylate (PMMA) are used as MEMS fusion
bonding media and have proved very successful for the bonding of polyimide [4].

Figure 23. Formation of sealed cavity using fusion bonding

High-Aspect-Ratio Micromachining

High-aspect-ratio micromachining (HARM) is a process that involves micromachining as a

tooling step followed by injection molding or embossing and, if required, by electroforming to
replicate microstructures in metal from molded parts. It is one of the most attractive
technologies for replicating microstructures at a high performance-to-cost ratio and includes
techniques known as LIGA. Products micro machined with this technique include high-
aspect-ratio fluidic structures such as molded nozzle plates for inkjet printing and micro
channel plates for disposable microtitreplates in medical diagnostic applications. The
materials that can be used are electroform able metals and plastics, including acrylate,
polycarbonate, polyimide and styrene.

LIGA
LIGA is an important tooling and replication method for high-aspect-ratio microstructures.
The technique employs X-ray synchrotron radiation to expose thick acrylic resist of PMMA
under a lithographic mask (see Figure 24 below). The exposed areas are chemically dissolved
and, in areas where the material is removed, metal is electroformed, thereby defining the tool
insert for the succeeding moulding step. LIGA is capable of creating very finely defined
microstructures up to 1000 µm high.

LIGA is limited by the need to have access to an X-ray synchrotron facility. A compromise
which combines some features of LIGA with surface micromachining eliminating the need
 for exposure to X-rays has been developed and is known as SLIGA (Sacrificial LIGA) [31].
It replaces the thick PMMA photo resist with polyimide as the electroplating mould, thus
enabling compatible conventional IC batch processing. HARM production methods have
provided radically new ways to produce micro machined parts for MEMS devices at relatively
low cost. In particular, techniques such as SLIGA enable the production of MEMS
components with much lower manufacturing infrastructures in terms of investment, facilities
and access to advanced materials and technology.

Figure 24. The LIGA process [3].

Other microreplication techniques can be combined to generate a preform for the tool insert.
These include laser ablation, ultra-violet (UV) lithography and mechanical micromachining,
which includes electric discharge machining (EDM) and diamond milling. EDM is a
relatively new approach that uses machine shop production techniques and offers the
capability to make parts out of most conductive materials. Unfortunately, as a spark erosion
technique, it is slow and not ideal for batch processing but has found many applications for
MEMS prototype production. For example, EDM was used for the micromachining of the
DENSO Micro-car previously shown in Figure 14.

Laser Micromachining
Most laser micromachining processes are not parallel and hence not fast enough for effective
MEMS fabrication. Nonetheless, they have utility in specialty micromachining or making
moulds. Excimer laser micromachining is used particularly for the micromachining of
organic materials (plastics, polymers etc.) as material is not removed by burning or
vaporization. Hence, material adjacent to the machined area is not melted or distorted by
heating effects. Lasers have found other applications in MEMS but only in a limited capacity;
laser drilling, laser annealing and etching are the most common forms.

Computer Aided Design

Computer Aided Design (CAD) is generally used in MEMS for the design of
photolithographic masks. This is a straightforward process as MEMS structures are relatively
large in comparison to the sub-micron structures usually associated with silicon chip
components. As well as using CAD for mask design, CAD and finite element analysis (FEA)
are important simulation tools for the design of MEMS applications. Unfortunately, to date
 there is a lack of adequate advanced software based design tools to fully model, analyse and
simulate MEMS microstructures as well as integrated MEMS/IC devices. This has acted as a
barrier to the development of MEMS devices and systems.

One of the most successful and commercially available software design tools today is
MEMCAD, a package from Microcosm Technologies in North Carolina, USA. The
MEMCAD system defines device layout and process, constructs the three dimensional
geometry of the device, assembles a detailed 3D model and analyses device performance as
well as device sensitivity to manufacturing and design variations. MEMS Pro, a package
from Tanner Research in California, enables designers of MEMS to simulate the
growth/deposition, implantation/diffusion and etch steps in a MEMS fabrication process.

Assembly and System Integration

The MEMS fabrication process essentially uses the same process as the microelectronics
industry as shown in Figure 25.

Figure 25. Similarities between IC and MEMS microfabrication processes [3].

Despite the fact that MEMS uses some of the same tools as those used with ICs, the greatest
challenge facing the MEMS industry is system integration between the miniature mechanical
systems and the electronic interface. For the cost-effective production of MEMS devices it is
necessary to combine complex mechanical structures together with microelectronics to form
integrated mechanical and electrical systems on a single chip that can be batch fabricated with
high yield and no additional or subsequent assembly (Figure 26).

Figure 26. Integration of mechanical structures and microelectronics [2].

Despite certain successful high volume applications such as the airbag accelerometer and the
disposable blood pressure sensor, high yields are difficult with MEMS devices due to their
mechanical complexity and their integration with the necessary microelectronics. Assembling
and packaging complex microscopic parts is also extremely difficult. As conventional
automated assembly and packaging is not suited to such a microlevel, to date, many MEMS
devices require individual handling. As a result, the final cost of a device may be up to 100
times the cost of the actual component. For their successful commercial production these
challenges have to be overcome.

Over the years different approaches have been developed for the integration of the electronic
interface. These include hybrid integration using conventional wire bonding and flip-chips
(described later in Section 3.8) and monolithic integration. Monolithic integration offers
superior system integration performance to hybrid systems but at an overall higher price in
terms of involved technology and processing. Monolithic integration can be carried out in
three ways:

i) IC before MEMS
Monolithic integration by IC first has proved to be successful and relatively cheap; an
example is the technology in Texas Instrument’s DMD (Figure 10). The process relies
heavily on bulk micromachining and the addition of new layers through electroplating. It is a
relatively simple integrated system but suffers from residual stresses within the device
materials. To date, refractory metals need to be used within the IC components in order to
withstand the high temperature annealing cycles required to relieve the stress in structural
polysilicon.

ii) Mixed MEMS-IC fabrication

A typical example of MEMS and microelectronics being fabricated side by side is the airbag
accelerometer (Figures 4 and 5). Monolithic processing of this device as well as the reduced
number of parts enable a very compact device with high reliability at a very low cost. The
trade-off lies within its complexity as this process leads to a very rigid and constrained
process flow which is expensive, thus requiring very high volumes.

iii) MEMS fabricated prior to IC

The most promising monolithic integration technique includes fabricating the MEMS device
prior to the microelectronics. Using technology known as iMEMS (Integrated Micro-
electromechanical Systems) patented by Sandia National Laboratories, USA, MEMS
components are fabricated in trenches on a silicon substrate and then the standard electronics
are processed onto the same substrate as shown in Figure 27 [2].
 Figure 27. Integrated MEMS/IC (iMEMS) technology developed at
Sandia National Laboratories for the monolithic integration of IC control
circuits and MEMS sensors and actuators on the same silicon chip. To
handle the differences in surface topography, the MEMS components are
fabricated in a shallow trench below the wafer surface prior to IC
processing [2].

A significant improvement in device functionality and a high level of integration can be

obtained by integrating the control circuits and MEMS devices on the same silicon chip.
Benefits of trench integration include smaller, faster, less costly, lower power and higher
sensitivity integrated systems. There are no real trade-offs with the technology except the
potential for more parasitic noise from interconnects on the chip.

Packaging

The proper operation of MEMS devices depends critically upon the ‘clean’ environment
provided by the package and is considered an enabler for the commercialisation of MEMS.
Packaging of microsensors presents special problems as part of the sensor requires
environmental access while the rest may require protection from environmental conditions
and handling (Figure 28).

Figure 28. Schematic illustration of the packaging role of a MEMS microsensor [32].

Although there is no generic package for a MEMS device, the package should:
 provide protection and be robust enough to withstand its operating environment
 allow for environmental access and connections to physical domain (optical fibres,
fluid feed lines etc.)
 minimize electrical interference effects from inside and outside the device
 dissipate generated heat and withstand high operating temperatures (where necessary)
 minimize stress from external loading
 handle power from electrical connection leads without signal disruption
 The most commonly used packages for sensors are usually based on derivatives of
conventional semiconductor packages including plastic, ceramic and metal can packages
(Figure 29).

Figure 30. Types of traditional IC packaging [32].

Each of these packages has been adapted in one form or another for the packaging of silicon
sensors. Since the substrate on many integrated circuits requires an electrical connection to
bias it, sensor dies are usually mounted to a die attach pad in the package using a conductive
bond. The die attach pad is typically joined to a metal lead frame with wire bonds providing
the electrical connections to the lead frame fingers. Various bonding media include AuSi
eutectic bonding, epoxy bonding (conductive or insulating depending on filler material) and
glass usually loaded with silver. The package is subsequently formed by plastic moulding (as
in the case of moulded plastic packages), sealed ceramic or metal caps (ceramic packages), or
with a brazed metal cap to the base of a metal package.

Wire bonding is still the most common technique for electrically connecting the die and
consists of two main technologies; ultrasonic and thermosonic, their difference being in the
mechanical stress applied to the die, the minimum spacing, the used wire alloys and the
possibility of wire adjustment.

In flip-chip (FC) technology the chips are bonded face down to a substrate via bumps;
materials include solder, gold, copper and nickel. On heating, the bump material melts and
simultaneously forms all the electrical and mechanical connections between the chip and the
substrate.

Multi-chip Modules
Multi-chip modules (MCMs) enable the integration and packaging of MEMS devices on a
single substrate using traditional thick-film technology. Using ceramics, silicon and printed
circuit board laminates as substrate materials, a variety of die types can be attached to, or
embedded within, the substrate surface. The dies can be interconnected by wire bonds, flip
chips or direct metallisation. The close proximity of each die allows for improved system
performance by providing low-noise wiring and in some cases eliminating unnecessary
interconnections. Three-dimensional variations of this technology are emerging in which dies
and their holding substrates are stacked up on top of each other.
 Passivation and Encapsulation
In order to protect MEMS devices from external contamination as well as enable them to
dissipate generated heat, thin-film coatings can be deposited on the components. In a process
called passivation using plasma enhanced chemical vapour deposition (PECVD), thin-film
coatings of usually silicon dioxide or silicon nitride increase wear resistance and electrical
insulation.

Encapsulation is used to protect the sensor die against adverse influences from the
environment like contaminants, mechanical vibration and shock. Common encapsulants are
epoxies, silicones and polyurethanes. These materials need to adhere well to the substrate, be
crack free and minimize induced mechanical stress as well as stresses due to mismatching of
thermal expansion coefficients.

3.9 Foundry Services

Despite the many similarities between IC and MEMS fabrication, MEMS makers, or
foundries, are still in their adolescence. The widening variety and increasing complexity of
MEMS products make the MEMS foundry business extremely problematic. Although the
fabrication technology is similar, the technology is on a different scale. MEMS are 3D
products in comparison to the 2D level of IC’s. Furthermore, unlike a standard IC foundry,
which performs one or two standard processes, a MEMS foundry performs a wide variety of
processes.

From an economic standpoint MEMS foundries share a common characteristic with

semiconductor foundries in that they are often more cost-effective than internal
manufacturing. This is because there is an enormous economy of scale as MEMS foundries
can leverage the cumulative volume from multiple products from multiple companies to
achieve high equipment utilization rates. The most important difference though is that
MEMS produces a mechanical structure that moves and that is significantly more complex
and sensitive than traditional ‘stationery’ IC structures.

Expansion of the MEMS market has been restricted by the need for specialized MEMS
engineering knowledge. Until recently, the majority of global research and development
investment has been limited to only a few MEMS foundries. A couple of years ago Cronos
Integrated Microsystems of North Carolina was seen as the leading MEMS foundry in the
field. A spin-off from MCNC and sponsored heavily by DARPA, it developed MUMPS,
offering low-cost manufacturing solutions for prototype and small-series MEMS applications.
Other pioneering foundries that were able to leverage the large scale investment already made
in silicon semiconductor fabrication include the Metal Oxide Implementation Service
(MOSIS), operated by the Information Sciences Institute at University of Southern California;
and Sandia National Laboratories, USA, with their SUMMiT technology. Smaller original
foundries still exist but in technology specific areas: MEMSCAP (France) – IC based;
SensorNor (Norway) – bulk micromachining and fusion bonding; GEMAC (Germany) – bulk
micromachining; and Bosch (Germany) – DRIE process. But of these labs, none are able to
handle mass production, certainly on the scale usually associated with the IC chip industry.
The lack of market demand for MEMS devices is one of the factors holding the technology
back; the devices can not be manufactured in numbers high enough to bring the price of
MEMS chips down to the point where they make economic sense.
In 1999, the MEMS industry witnessed unprecedented growth and wealth creation as major
venture capital and corporate funding took place. Today, it is no longer necessary to build a
separate fabrication facility to create a MEMS product or invest hundred of millions of dollars
to become a MEMS player. A newly created infrastructure of smaller foundries is now
available to support both emerging and proven MEMS-based applications. Although the
biggest of these foundries are the risk-taking producers of specialized and niche chips such as
Motorola, Sony, Analog Devices and Texas Instruments, there are also many smaller start-up
MEMS foundries which offer more specialized services such as packaging, testing, reliability
analysis etc. Not only do these foundries offer the obvious manufacturing technology and
services to outside customers but they offer the concept of ‘shared learning’ in that lessons
learned when the foundry puts into volume production one product can be applied to the
subsequent volume production of another product. This can occur even when the products are
completely different because there will still be sharing of specific process modules. Today,
there are more than 40-50 MEMS makers or foundries worldwide (though none has broken
away from the pack as in the case, for example, of Intel with computer chips).

4. MEMS Transducers

Microsensors and microactuators are at the very core of a MEMS device or system. A
microsensor detects changes in the system’s environment; an ‘intelligent’ part processes the
information detected by the sensor and makes a decision in the form of a signal; and a
microactuator acts on this signal to create some form of changes in the environment.
Microelectronic components make up most of the intelligent part of the device and, as an
established technology, will not be discussed here.

Sensors and actuators are broadly termed transducers and are essentially devices that convert
one form of energy into another. Many of the MEMS sensors and actuators described in this
section have been developed within the microelectronics industry and do not all involve any
special micromachining techniques; they are based on conventional integrated circuits that,
through inherent mechanisms, sense light, temperature etc. However, many of these can be
enhanced by the use of MEMS.

Basic MEMS mechanisms and structures consist of both in-plane and out-of-plane
mechanisms as well as structural members to couple energy between the actuator and sensors
as well as with the physical interface of a mechanical system. Mechanisms such as joints,
linkages, gears and hinges are very typical.

This section concentrates on the phenomena that can be sensed or acted upon with MEMS
devices with a brief description of the basic sensing and actuation mechanisms. It is
important to note that although these devices are mechanical and have been categorized in
terms of their sensing domain (e.g. thermal, chemical, radiation), there are many overlaps, and
forms of mechanical transducer can be commonly found as intermediate mechanisms in other
devices.
 Mechanical Transducers
Mechanical Sensors

There is a tremendous variety of direct mechanical sensors that have been or could be
micromachined depending on their sensing mechanism (usually piezoresistive, piezoelectric
or capacitive) and the parameters sensed (typically strain, force and displacement).

i) Piezoresistive sensors
As a result of the piezoresistive effect (defined as the change in resistivity of the material with
applied strain), changes in gauge dimension result in proportional changes in resistance in the
sensor. The piezoresistive effect in semiconductors is considerably higher than in traditional
metals, making silicon an excellent strain sensor. MEMS piezoresistors are readily
manufactured using bulk silicon doped with p-type or n-type impurities.

ii) Piezoelectric sensors

Piezoelectric sensors utilize the piezoelectric effect in which an applied strain (or force) on a
piezoelectric crystal results in a potential difference across the crystal. Similarly, if the crystal
is subjected to a potential difference, a displacement, or strain, is produced. The effect can be
used to sense mechanical stress (i.e. displacement) and as an actuation mechanism, although
displacements are small even for large voltages. Common piezoelectric materials used for
MEMS applications include quartz, lead zirconate titanate (PZT), polyvinylidene fluoride
(PVDF) and ZnO, PVDF and ZnO being the most common. Silicon is not piezoelectric;
hence a thin film of a suitable material must be deposited on the devices.

iii) Capacitive sensors

Capacitive (or electrostatic) sensing is one of the most important (and widely used) precision
sensing mechanisms and includes one or more fixed conducting plates with one or more
moving conducting plates. Capacitive sensing relies on the basic parallel-plate capacitor
equation shown below. As capacitance is inversely proportional to the distance between the
plates, sensing of very small displacements is extremely accurate.

 0r A = permittivity of free space = 8.854x10 -12 Fm-1

C where: 0
d
r = relative permittivity of material between the plates

A = overlapping plate area (m)

d = plate separation (m)

iv) Resonant sensors

MEMS resonant sensors consist of micromachined beams or bridges which are driven to
vibrate at their resonant frequency. They can be attached to membranes or designed to adhere
to a particular substance (as in the case of a biosensor). Movement of the membrane or
increased build-up of the binding substance will affect the resonant frequency and can be
monitored using implanted piezoresistors.

Types of mechanical sensor include:

a) Strain gauge - a strain gauge is a conductor or semiconductor that is fabricated on or
bonded directly to the surface to be measured. An example of a polysilicon strain sensor
unable to be fabricated by any other method than MEMS is an implantable piezoresistive
strain gauge to measure forces in heart and brain tissue.

b) Accelerometer - accelerometers sense acceleration by using a suspended proof mass on

which external acceleration can act (Figure 30). Upon acceleration (or deceleration), a force
(F=ma) is generated on the proof mass resulting in displacement. The force or displacement
is usually measured by piezoresistive and capacitive methods.

Figure 30. Suspended proof mass in a piezoresistive accelerometer (not to scale).

c) Gyroscope – a gyroscope is a device that measures the rotation rate and detects inertial
angular motion. As a result it can be found, for example, in transportation, navigation and
missile guidance applications. It relies on measuring the influence of the Coriolis force on a
body in a rotating frame. MEMS gyroscopes typically use vibrating structures because of the
difficulty of micromachining rotating parts with sufficient useful mass.

d) Pressure sensor - MEMS pressure sensors are usually based around thin membranes with
sealed gas or vacuum-filled cavities on one side of the membrane and the pressure to be
measured on the other side. Piezoresistive and capacitive membrane deflection measurement
techniques are most commonly used in commercial pressure sensors.

Mechanical Actuators

i) Electrostatic actuation
The fundamental actuation principle behind electrostatic actuators is the attraction of two
oppositely charged plates. Their use is extensive in MEMS devices, since it is relatively
simple to fabricate closely spaced gaps with conductive plates on opposite sides. For a
parallel plate capacitor, the energy (W) stored at a given voltage (V) is equal to:
1
W  CV 2 where: C=capacitance between the plates
2

W 1 C 2
And the force between the plates is: F   V
x 2 x

Comb-drive-type actuators make use of a large number of fine interdigitated fingers that are
actuated by applying a voltage between them (Figure 31). As the capacitance is related to
area, the greater the number of fingers, the larger the force that can be generated by the
actuator.
 Figure 31. Comb-drive electrostatic actuator concept [2,33].

Electrostatic rotary motors are another good example of the success of MEMS sacrificial
oxide/polysilicon techniques. They rely on a central freely-moving rotor with surrounding
capacitive plates that can be driven in correct phase to cause the rotor to turn. Harmonic or
‘wobble’ motors rely on the principle of a rotor turning in a slightly larger stator ring, such
that it ‘wobbles’ around the central axis as it rotates (Figure 32). Reduction of sliding friction
and increased electrostatic forces can be achieved with these motors.

Figure 32. A MEMS electrostatic ‘wobble’ motor [30].

ii) Piezoelectric actuation

As previously described, the piezoelectric effect can be used in both sensors and actuators. In
piezoelectric actuation, the electrically induced displacement (or strain) is proportional to the
applied potential difference. Despite small displacements, relatively high forces (in the region
of tens of MPa) can be achieved using lower voltages than those required for comparable
electrostatic actuation. It should be noted, however, that it is dependent on the geometry of
the device components. The main disadvantages of piezoelectric actuation include high
complexity of fabrication, as well as small actuation displacements. Larger displacements can
be achieved using multiple piezoelectric layers known as piezoelectric bimorphs. Most
MEMS piezoelectric actuation is used where small strains are required (for example, the tip of
a scanning tunnelling microscope) [4].
Radiation Transducers
Radiation Sensors

Radiation sensors cover ionising radiation as well as visible light, infra-red (IR) and ultra-
violet (UV) radiation. Current ionising radiation sensors for high-energy particles and X-rays
include Geiger-Müller (GM) tubes and scintillators and although they have not been realized
using MEMS, their miniaturization is potentially feasible. Sensors for visible, IR and UV
radiation are generally categorized as either direct or indirect. Direct optical sensors detect
photons and result in an electronic signal. Indirect sensors convert the optical signals into an
intermediate energy form (e.g. thermal or chemical), which is then measured electrically.
There are a wide variety of both direct and indirect sensors and only the most common will be
described here.

i) Photodiodes
A photodiode is a semiconductor device for measuring light intensity based on the
photoconductive effect (increase in conductivity of a semiconductor on exposure to light).
Photodiodes are junction-based photoelectrodes which have a p-n junction. When visible or
near infra-red light falls on the device, additional charge carriers are generated resulting in
increased current flow.

ii) Charge-coupled devices

Charge-coupled devices (CCDs) are one of the most common photodetectors used in
handheld video recorders and many other consumer applications. They consist of a metal gate
(electrode) above a dielectric and a semiconductor substrate. This forms a metal oxide
semiconductor (MOS) capacitor, the charge on which arises from photogenerated carriers.
CCDs can be linear or made up of arrays of metal-insulator-semiconductor sensors arranged
so that photo-generated charge can be stored and transferred between elements by an
appropriate variation of control voltages applied to surface electrodes (memory/signal
processing approach).

iii) Pyroelectric sensors

Pyroelectric detectors are an example of indirect optical sensors and are essentially capacitors
whose charge can be altered by illumination or temperature changes. By converting incident
light into heat, which is then measured, pyroelectric sensors have a wide range of applications
in surveillance, military, security consumer markets etc. e.g. human motion detectors.
Pyroelectric sensors use piezoelectric and ferroelectric materials (varying dielectric constant
with applied voltage). ZnO is the most common in MEMS devices.

Radiation (Optical) Actuators

The two most common forms of optical actuation include light emitting diodes and light
modulators such as liquid crystal displays and reflective micromechanical light modulators
(technology used in Texas Instrument’s DMD projection system).

Optical devices can either be active or passive; active devices include laser emitting diodes,
photodiodes and optical switches. Passive devices include couplers, mirrors, wavelength
division multiplexers, polarisers etc. Their application is important in both optical MEMS
 devices and integrated optics to enable the control and analysis of optical acquired data. A
brief selection of optical MEMS components includes:

 optical waveguides to route optical energy from one region to another. MEMS
techniques are commonly employed to enable waveguide-to-photodiode coupling.
 fibre-optic couplers incorporate micromachined grooves and channels as mechanical
couplings for fibre-optic components. Anisotropic wet etching, DRIE and HARM
methods are typically used.
 micromirrors with flawless surfaces can be achieved using anisotropic etching to yield
perfect crystal plane surfaces.

Thermal Transducers
Thermal Sensors

i) Thermo-mechanical sensors
Thermo-mechanical sensing (and actuation) utilize the fact that all materials have a
coefficient of thermal expansion. Consequently, if two different materials are sandwiched
together and undergo a temperature change, movement in the sandwich assembly would
occur. This is the basis for the common bimetallic (or thermal bimorph) sensing and
actuation.

ii) Thermoresistive sensors

Thermoresistive sensors rely on the fact that the resistivity, , of most materials changes with
temperature and is shown by the equation   R(1  aT  bT 2 ) where R is the resistivity of the
material at a reference temperature, T (ºC), and a and b are constants specific to the material
being used. The rate of change of resistance with temperature is known as the temperature
coefficient of resistance. Most common materials exhibit an increase in resistance with
temperature (e.g. platinum is particularly linear with temperature). Certain materials – for
example, carbon, some ceramics and most semiconductors used in thermoresistors or
‘thermistors’ – exhibit a decrease in resistance with increasing temperature. They are not as
linear as platinum, but often cheaper to fabricate and easier to integrate with circuitry in
MEMS devices.

iii) Thermocouples
The thermocouple is probably the most common temperature transducer. It consists of a
junction between two different materials and measures the temperature-dependent voltage that
arises across the junction. Semiconductor materials often exhibit a better thermoelectric
effect than metals. Thermocouples have been used in a wide variety of MEMS sensors in an
array arrangement referred to as a thermopile.

Thermal Actuators
Thermal actuation in MEMS is usually as a direct result of incorporating tiny heaters, or
resistors. These resistors can be controlled to locally heat specific areas or layers as in the
case of a bilayer actuator. As already detailed, basic thermal actuation utilizes the difference
 in thermal coefficients for expansion of two bonded materials and is referred to as thermal
bimorph actuation. A typical thermal bimorph actuator is shown in Figure 33.

Figure 33. Example of a MEMS thermal bimorph actuator [34].

This can be applied to a volume of fluid (liquid or gases) in sealed cavities with a thin
membrane as a wall. By incorporating a heater the liquid can be heated causing it to expand
and deform the membrane outwards. Large forces can be achieved using thermally actuated
devices but power consumption can be high and it can take time for material to cool to its
original activation state.

i) Shape memory alloy actuation

Shape memory alloys (SMAs) exhibit considerable changes in their length (contraction) when
heated. These include titanium/nickel alloys, of which some, once mechanically deformed,
would return to their original undeformed state when heated. Being conductive they can be
heated simply by passing a current through them.

Magnetic Transducers
Magnetic Sensors

Most MEMS magnetic sensors are based on the Hall effect. They rely on the production of an
electric field across a material through which an electric current is flowing and where a
magnetic field is acting. The force applied to the charge carriers by the electric field exactly
balances a force from the magnetic field called the Lorentz force. Other ways to sense
magnetic fields use optical sensors that rely on the magneto-optical effect or materials that
exhibit magnetic anisotropy (shape, stress and crystalline anisotropy are particularly common
in MEMS magnetic applications). Magnetic coils are not common in MEMS as they tend to
be 2D, which is not useful for many applications.

The most sensitive magnetic sensors include superconducting quantum interference devices
(SQUIDS). The devices can detect diminutive magnetic fluxes produced by the electric
currents in heart and brain tissue. Advances in superconducting thin films as well as MEMS
techniques have accelerated growth with these devices.

Most of today’s magnetic sensors are silicon based, not only because of the ease of fabrication
and their ability to be readily integrated with circuitry, but also because their high volume
demand necessitates lower cost.
Magnetic Actuators

Magnetic actuation is based on the fact that a current-carrying conductor generates a magnetic
field. If this conductor is a wire (or coil) and interacts with another external magnetic field
(e.g. from a similar conductor or coil) a mechanical force is produced.

Despite the success of magnetic actuation on a macroscale, such as motors or solenoids,

MEMS magnetic devices are still relatively unestablished. This is due to the fact that 3D
coils are very difficult to fabricate by MEMS. A notable exception is the magnetic read/write
heads for computer disk drives. This device can be classified as both a sensor and actuator in
that it reads from and writes to magnetic media.

Choice of magnetic material is limited to those that can be easily micromachined and do not
suffer from high power consumption and heat dissipation. Thin film permanent magnets can
be fabricated using MEMS; LIGA is particularly common for polymers such as polyimide
which can be loaded with a magnetic powder and electroplated. Although attempts have been
made to fabricate MEMS magnetic actuators with wire-bonded coils, often in MEMS devices,
magnetic actuators compete with electrostatic devices which are stronger for the same
volume.

i) Magnetostrictive Actuators
These rely on the magnetostrictive effect, which is the change of shape or size of a
ferromagnetic material induced by a magnetic field, for example, the contraction of a nickel
rod under a longitudinal magnetic field.

Chemical and Biological Transducers

Chemical and Biological Sensors

Chemical and biological sensors encompass a large and wide variety of devices that interact
with solids, gases and liquids of all types and are therefore extremely diverse and
interdisciplinary. They are different from previously described sensors in that they must
directly interact with a chemical medium to connect the chemical and electrical domains.
Hence they require ‘openings’ within their packaging to enable this interaction (like pressure
sensors).

Most chemical and biological sensors do not require extremely sophisticated micromachining
but can require considerable interdisciplinary knowledge and sophistication for their actual
use. Chemical sensors must be very highly selective in order to make such identifications
among compounds without falsely responding to potential interfering species.

Chemical sensors can be categorized in many ways including passive chemical sensors, work
function based systems and electrochemical transducers. Passive chemical sensors include:

 Chemiresistors measure the resistance of a chemically sensitive layer between two

electrical contacts. Sensitivity can be increased using more micromachined
electrodes.
  Chemicapacitors are similar to chemiresistors but the capacitance of the sensitive layer
is measured. Chemicapacitors have found application in sensing humidity.
 Chemomechanical sensors rely on direct chemical-to-mechanical transduction (for
example, the expansion of a thin-film polymer in the presence of a substance being
analysed).
 Calorimetric sensors measure heat generated by chemical reactions.

i) Work function based sensors

This class of sensors, including the ion sensitive field effect transistor (ISFET) and the metal
oxide semiconductor field effect transistor (MOSFET), utilizes metal-insulator-semiconductor
junctions and the fact that the work function of the material at the interfaces can be
chemically modulated. The ISFET was developed as a direct result of the fact that metal-
oxide-semiconductor transistors were so sensitive to surface contaminants during their
manufacture. As a large proportion of chemical sensors are based on the ISFET, a more
detailed description of their operation is outlined.

ISFETs sense the concentration or activity level of a particular ion in solution. The ISFET is
a derivative of a common electronic component called a MOSFET. This consists of a silicon
semiconductor substrate (doped with impurities to make it p-type) and two electrical contacts
(source and drain) doped with impurities so that negatively charged electrons are the main
carriers in these small n-type silicon regions. A small distance separates source from drain
(Figure 34). Overlaying the substrate between the source and drain is a silicon dioxide
insulator which itself is overlaid with a metal electrode called a gate. When a potential is
applied to the gate of the MOSFET, the induced electrical field changes the freedom with
which the current flows between the source and the drain. In the case of an ISFET however,
there is no gate electrode and the insulator is in direct contact with an electrolyte solution to
be measured. With the ISFET, electric current flows from the source to the drain via a
channel. As in the MOSFET the channel resistance depends on the electric field
perpendicular to the direction of the current. Also it depends on the potential difference over
the gate oxide. Therefore the source-drain current is influenced by the interface potential at
the oxide/aqueous junction. When SiO2 is used as the insulator, the chemical nature of the
interface oxide is reflected in the measured source-drain current. With the selection of other
appropriate insulator material, such as silicon nitride or aluminium oxide, hydrogen ions will
reside at the surface of the insulator in proportion to the pH. Their positive charge produces
an electric field that modulates the current between the source and drain. In order to quantify
this effect, the control voltage is measured that must be applied (via a reference electrode) to
maintain the drain-source current at a constant value.

The chemical sensitivity of the ISFET is completely controlled by the properties of the
electrolyte/insulator interface. One significant problem in the design and fabrication of
ISFETs is ensuring that the selective membrane adheres to the device. If the integrity of the
membrane is compromised, then the device is useless.
 Figure 34. Schematic diagram of an MOSFET and ISFET [35].

ii) Biosensors
Biosensors is the term used for a whole class of sensors that utilize a biochemical reaction to
determine a specific compound. A biosensor is generally a transducer combined with an
immobilized enzyme or cell to monitor a specific change in the microenvironment. Typically,
an enzyme, antibody, polysaccharide or nucleic acid is used to interact with the substance to
be measured. The sensor itself can vary but a form of ISFET is typically used. The probe tip
is immersed in a liquid phase and is in contact with the process either directly or through a
membrane. They have not seen widespread use because as a class they exhibit many
disadvantages including oxidization reactions with product, extreme sensitivity and the need
for regular recalibration (not ideal for implantable devices).

One of the most promising applications of an amperometric enzyme-based biosensor is the

glucose oxidase based sensor for monitoring glucose levels in the blood. This is particularly
important for diabetes and also in the fermentation process.

Chemical Actuators

There is certainly potential to fabricate chemical actuators using MEMS, but to date, there has
been little work in this area with the exception of a few devices using microelectrodes. These
electrochemical transducers are based on the simple electrochemical electrode concept in
which current is transducer from the circuit domain into the chemical domain through
oxidation or reduction of chemical species at the electrode surface. These structures are
amongst the simplest (they can be as simple as a region of bare metal in solution) and play a
major role in biological interfacing (e.g. neurophysiologic probes).

4.6 Micro fluidic Devices

MEMS has many applications in micro fluidics with many of the key building blocks such as
flow channels, pumps and valves fabricated using mature micromachining techniques.
Chemical analysis, drug delivery, biological sensing, environmental monitoring and many
other applications typically incorporate MEMS micro fluidic devices. It should be noted that
in MEMS fluidic devices the type of flow (laminar or turbulent), effect of bubbles, capillary
forces, fluidic resistance and capacitance all have an effect on their final design.
i) Flow channels
A wide variety of micro fluidic channels have been fabricated using bulk micromachining
(wet and dry etching), surface micromachining and moulding techniques (Figure 35).

Figure 35. Selection of MEMS channels for microfluidic applications [36,37,38].

ii) Flow sensors

MEMS flow sensors can be fluid-dependent flow or fluid-independent. In a very basic form,
fluid-dependent flow sensors measure the flow rate by heating a fluid ‘upstream’ and then
recording its temperature ‘downstream’. The flow rate is proportional to the temperature
difference and transit time of the two actions. Fluid-independent flow sensors measure
pressure or force exerted on an object by the fluid. Figure 36 shows an example of a MEMS
bulk drag-force flow sensor. Using a piezoresistive sensing mechanism, flow measurements
are fairly linear; direction and magnitude can be sensed by this method. In addition, fluid-
independent flow sensors do not involve any form of heating and hence are more suited to
biological fluid applications.

Figure 36. Micromachined mechanical (drag-force) flow sensor [39].

iii) Valves
Valves are generally classified as either active or passive depending on whether or not they
have an external power or control source. One of the simplest MEMS valves is the passive
check valve shown in Figure 37.

Figure 37. Basic concept of passive silicon check valve [40].

Active MEMS valves can be actuated by many methods including thermal, piezoelectric,
electrostatic and shape memory alloy means. Thermal actuation is the most common.
 iv) Pumps
Pumps are generally an important part of micro fluidic devices. However, MEMS pumps are
very sensitive to fine particles which often cause contamination and leakage of the device.
Membrane, rotary and ultrasonic pumps are the most common types of MEMS pumps.

v) Rotary pumps
Figure 38 shows the basic concept of a magnetic rotary micro pump. LIGA is commonly used
as a fabrication technique for micro machined PMMA gears in MEMS micro fluidic systems.
These gears can be driven for example using electroplated NiFe bars mounted on one or both of
the gears. Fluid is pumped by the action of the turning gears.

Figure 38. Basic concept of magnetic rotary micropump [41].

vi) Droplet generators

The majority of droplet generators in commercial MEMS microfluidic devices are inkjet printer
heads as previously described in Section 2.4.1. These rely on either thermal or piezoelectric
actuation to eject ink droplets, thermal actuation being the most common. Piezoelectric ally
actuated valves offer the advantage of very high forces, but very small movement for even very
large voltages. They also find use in automotive fuel-injection valves and spray nozzles.

FABRICATION OF OUT-OF-PLANE POLYMER MICROLENSES

UV lithography method to directly fabricate out-of-plane microlens with negative tone photoresist SU-8 .
An out-of-plane microlens array with nearly-spherical micrlens pixels was successfully fabricated with this
method. However, there are two disadvantages with the previous technology. First, the use of an ultra- thick
(>500 m) SU-8 photoresist layer and multiple tilted UV exposures made this process a time consuming
one. It took more than two days to finish the whole process and expensive photo-lithography process
equipments was necessary.
Secondly, microlenses can only be made of SU-8 polymer which is essentially a photo sensitive material and
has limited optical transparency. It has been observed that aged SU-8 polymer turns to reddish color. A fast
and low cost out-of-plane microlens fabrication technology which can be used to make microlenses with a
broad selection of materials is therefore highly desirable.
PDMS is a silicone based elastomer. It was introduced into micro fabrication in late 90s [20]. Because it is
in a liquid state before curing and has a very low surface energy (essential ford molding process), PDMS has
quickly become arguably the most popular material in micro and nano-molding processes. The PDMS
molding process has also been adopted into micro lens fabrication [60]. However, the reported micro lens
fabrication processes using PDMS are all for in-plane microlens fabrication. Unlike in-plane micro lenses
whose relief structures are on the substrate plane, out-of-plane micro lens has relief structure on the sidewall.
This makes the molding of out-of-plane micro lens very challenging since the lens structures may easily be
damaged during the demolding process. PDMS, because of its unique properties, is a great candidate for out-
of-plane molding and replication process. Before curing, PDMS pre-polymer is in relatively low-viscosity
liquid state, so it can conformally cover the three-dimensional microstructures. After curing, the
polymerized PDMS has good elasticity, which makes it possible to deform during the demolding process
and come back to the original shape after demolding. So the negative mold can be released while keeping
the SU-8 structure on the master intact.

Fabrication of SU-8 Master Mold

The fabrication of an out-of-plane micro lens master mold was done using a tilted UV lithography process
with the widely used negative tone photo resist SU-8. Details of the fabrication process has been reported by
Yang and Wang .
UV lithography in water medium

nglass=1.52
nwater=1.33 nSU-8=1.65

UV light

Mask

SU-8

Substrate

DI water
To perform the tilted UV lithography process, we chosed to use a water immersion lithography method
similar to the one reported by Sato et al [62]. Comparing to prism assisted tilted UV lithography [6] [59],
the water immersion approach is relatively easy to set up, and the maximum inclined angle can reach 56.2o
which is sufficient for out-of-plane microlens fabrication because only +/- 45o inclined angles are required.
Before exposure, silicon substrate spin-coated with SU-8 was fixed together with a chromium mask using a
homemade clutch. As shown schematically in Figure 2-1, the clutch set was immersed in deionized (DI)
water in a tilted position. The relationship between the resulting structure inclined angle and the tilted angle
of the clutch set can be calculated using Snell’s law. The following equation can be derived:

Where , nSU =1.65 , nglass =1.52, nwater =1.33, are the refractive indices of uncured SU-8, glass (soda
lime), and water respectively. In our fabrication process, the final structures should have
-/+ 45o inclined angles with respect to the substrate.

Therefore, the clutch should be held at 61.26o angle with respect to the horizontal plane during exposure.
Since SU-8 has strong absorption to short wavelength UV light, short wavelength UV light may be
absorbed only in the top layer and cause exposure non- uniformity throughout the whole photoresist
thickness. For better exposure results, a piece of 1cm thick PMMA sheet (Plexiglas Grade G, Altuglas
International, Philadelphia, PA) was used as an optical fileter to eliminate short wavelength spectrum of the
UV light source. This PMMA sheet can block most of the UV light with wavelength shorter than 400nm.
With the filter, SU-8 was only exposed by the g-line (436nm) and the h-line (405nm). Absorption of the g-
line is much lower than that of the h-line, the exposure is dominated by the h-line. All exposure dosage
mentioned in this chapter is measured at the h-line, unless otherwise stated.

The fabrication process for micro lenses made of cured SU-8 is graphically illustrated in Figure 2-2.

(1) Substrate preparation and spin-coating: A 4 inch silicon wafer with one side polished was cleaned
in acetone, isopropyl alcohol (IPA) and DI water successively. After dehydration at 180 oC for 30
minutes, about 16g of SU-8 100 was dispensed onto the wafer and spincoated at 460 rpm for 30
seconds to obtain an 1100µm thick SU-8 layer. Since the layer was relatively thick, no edge beads
were observed after letting the sample set at room temperature for 20min. To reduce the bubble
generation during the spin-coating process, SU-8 was poured onto the wafer directly from the bottle
with care, and formed bubbles were pricked by a sharp needle.
 Unexposed SU-8

Exposed SU-8

Substrate

(b)
(a)

Double exposed
region
(c) (d)

Figure 2-2 SU-8 out-of-plane microlens fabrication process

(2) Soft bake: The spin-coated SU-8 was soft baked on a well-leveled hot plate to eliminate the solvent.
The hot plate has a heating part underneath the wafer which helps drive the solvent from the bottom to
the top and, thus, keep the top layer soft. This is important for thick SU-8 layer soft bake since a dried
top layer can prevent remaining solvent from being evaporated. To reduce the stress in the thick SU-8
layer and achieve better lithography results, multi-step ramping and stepping soft bake process is used.
Figure 2-3 shows the multi-step baking process used in this
fabrication. The SU-8 layer was slowly ramped up to 110oC in multi steps and kept at 110oC for 10 hrs (for
a thick SU-8 layer, the rule of thumb for soft bake time is about 1hr per100µm). The
Dwell for 10hrs
Ramp up to 110oC in 30m Ramp to 65oC in 30m Dwell for 15m
Dwell for 15m Ramp up to 75oC in 30m Ramp to 25oC in 3 hrs Relax at 25oC until exposure
Relax at 25oC for 1-2 hrs
Figure 2-3 Multi step soft baking process

SU-8 layer was then slowly cooled down to 650C. Since the glass transition temperature of uncured SU-
8 is about 60oC, the cooling at around and below glass transition temperature should be done in a very
slow rate to greatly reduce the internal stress in the thick SU-8 layer. 20oC per hour was used in this
process. After the soft bake, surface inspection was performed. It was found that the SU-8 surface had
relatively smooth surface, height deviation across the whole 4- inch wafer was less than 30µm for an
1100µm layer.

(3) Tilted UV exposure: The wafer a chromium mask were fixed together in a rotatable home- made
clutch (a drop of glycerol was applied between the mask and the wafer to serve as index matching
fluid) and immersed in a DI water container for the first 45 o tilted UV exposure as described in the
previous section (Figure 2-2 (a) shows the result after first exposure). After about 2 minutes
relaxing, the whole set was put in an opposite direction for the -45o exposure (Figure 2-2 (a) shows
the result after second exposure). At each intersection area of two exposure beams, a four faced
double exposed region was obtained as shown in Figure 2-2 (c). 4.5 J/cm2 exposure dosage at
405nm was used for each exposure. The dosage was experimentally optimized so that even the
double exposed regions were still slightly under exposed, which was important for resulting of
microlens with smooth surface. All UV exposures were performed using an Oriel UV exposure
station (Newport Stratford, Inc. Straford, CT).

Figure 2-4 Multi step post bake proces

Dwell for 30m

o
Ramp up to 95 C in 20m Ramp to 65oC in 30m Dwell for 15m
Dwell for 15m Ramp up to 75oC in 20m Ramp to 25oC in 3 hrs Relax at 25oC
Relax at 25oC for 30m
 (4) Post bake: After exposure, the SU-8 layer was post-baked to accelerate the polymerization process.
Similar to the soft baking process, rapid heating and cooling should also be avoided during post
baking process. As shown in Figure 2-4, the sample was slowly heated up to 96oC and kept there for
30m and then slowly cooled down to room temperature. To reduce the residual stress even more, a
post bake at a lower temperature for a much longer time could also be employed.
(5) Development: After post bake, the whole sample was developed in SU-8 developer (PGMEA,
propylene glycol methyl ether acetate). Because of the thickness of our SU-8 layer, a two-step
development process was performed. As the first step, the sample was held facing- down in the
developer for about 1 hr to solve most of the unexposed SU-8. Then, the sample was held face-up in a
fresh developer solution for about 40 minutes with very mild agitation. Since the double exposed
region was slight under exposed, the sharp edges of the four-faced region were smoothed out during
the development process and eventually spherical microlens pixels with smoothed surfaces were
formed as shown in Figure 2-2 (d). The developed sample was then rinsed in fresh SU-8 developer
and dried in the hood under mild airflow. For microlens fabrication, IPA rinsing step which is widely
used in SU-8 development process should be avoided since milky white spots formed on the partially
polymerized SU-8 microlens surface can
cause great deterioration of its optical performance.

(6) Flood exposure and hard bake: After being dried, the sample with SU-8 out-of-plane microlens
arrays is then flood exposed under the UV exposure station and baked at 95 oC for 1 hr. This treatment
gave the final structures better mechanical stress. However, it was found during experiment that fully
cross-linked SU-8 polymer turned to reddish color, which may negatively affect the optical
performed of the microlens if SU-8 microlenses were directly used. For molding and replication
process, the effect of this color change is negligible. SU-8 master mold fabrication was finished after
hard bake process.

Out-of-plane Microlens Replication

PDMS molding process was used to make the negative mold from the SU-8 microlens array. Other curable
polymers could then be used to cast out-of-plane microlens replica from the PDMS mold. In this work, a
solvent-free UV curable polymer (NOA 73, Norland Products Inc., Cranbury, NJ) was used as an example
curable polymer for demonstration purpose.

The microlens replication process is schematically shown in Figure 2-5. SU-8 out-of-plane microlens array
was ready for PDMS molding after hard bake (Figure 2-5. (a)). PDMS pre- polymer mixture (Sylgard 184
Dow Corning) was prepared by mixing the base with curing agent in a 10:1 ratio. After being thoroughly
stirred to ensure complete mixing, the PDMS pre-polymer mixture was placed in a vacuum chamber for 20
minutes to eliminate air bubbles. The PDMS pre-polymer mixture was then poured on the SU-8 master
mold to form a 3mm thick PDMS layer. The PDMS covered wafer was then placed in a vacuum chamber
for 20 minutes again to evaporate air bubbles and ensure that PDMS conformally covered the
microstructures on the master mold. PDMS was then cured at 85 oC for 3 hours (Figure 2-5 (b)). Since SU-
8 mold master with out-of-plane microlens array had micro relief structure on the sidewall, one would

Figure 2-5 Out-of-plane microlens replication process: (a)-(c) PDMS molding, (d)-(f) curable polymer
casting

Curable Polymer
SU-8 polymer
Substrate

PDMS Glassslide
(a) (d)

(b)
(e)

(c) (f)
have expected that demolding process might damage the master mold. However, the excellent flexibility of
PDMS facilitated the demolding and made this whole process possible. Before being peeled off from SU-8
master, the PDMS mold was pulled sideways to be separated from the SU-8 microlens surface.

The whole PDMS negative mold could then be peeled off from the master with ease (Figure 2-5 (c)). A
calculated amount of UV curable resin (NOA 73) was then poured onto the PDMS negative mold. After
vacuum treatment, excessive resin was scraped off using a clean razor blade (Figure 2-5 (d)). Comparing to
other NOA series product, NOA 73 resin had low viscosity (130cps, from NOA 73 datasheet), which was
very important for the resin to flow into the three-dimensional micro cavities in the PDMS mold.

A microscope slide was then covered on the PDMS mold filled with UV curable resin (Figure 2-5 (e)) and
exposed under a mercury lamp (Newport Cooperation, 9mw at 365nm) for 300s. After peeling off PDMS
negative mold, out-of-plane microlens array replicas made of NOA 73 UV curable polymer was
fabricated on glass slide (Figure 2-5 (f)). The whole replication process took less than 5 hours (not
including the fabrication of SU-8 master mold). Because the PDMS mold is reusable, multiple replications
can be done use the same negative mold. In the following operations, each replication process can be
completed in less than 30 minutes.
Optical Analysis

Optical analysis is important to study the performance of the optical detection system and optimize
itsdesign. Optical simulation was done using Zemax EE (ZEMAX Development Corp., Tucson, AZ). In our
system, the output multimode fiber core diameter was 200 m, which is more than 200 times greater than
typical fluorescent light wavelength. Therefore enough transverse modes can be supported in the output
fiber and geometrical ray tracing method can be used to evaluate the light coupling efficiency .

An optical system as shown in Figure 3-2 was built in Zemax. A sample cell/bead emitting fluorescent
light was modeled as a point light source [73, 74]; an imaginary aperture stop was placed in front of the
microlens to define the optical rays which were traced during the simulation; microlens was modeled as a
spherical lens with diameter of 450 m and radius of curvatures of 380 m for both surfaces; a circular
image surface with numerical aperture (N.A.) of 0.2 was defined in Zemax to model the optical fiber with
same N.A..

During the simulation, Zemax randomly generated a group of rays from the point light source to fill the
aperture stop; those rays were then bent by the microlens and reached the plane where the output fiber
end face was; part of the rays missed the defined output fiber end face and was obviously not supported
by the output fiber; among those rays that did hit the acceptance end of the optical fiber, some came in at
an angle larger than the optical fiber’s numerical aperture and were also be dropped; only those rays that
hit the optical fiber acceptance end at an angle within its numerical aperture were supported by the
output fiber and contributed to the final detection signal. By comparing the number of rays that were
supported by the output fiber to the original rays filling the aperture stop, the light coupling efficiency
was calculated. The optimization function built in Zemax automatically compared coupling efficiency
for different microlens focal lengths, light source to lens distances, lens to optical fiber distances, etc and
decided the best value for those parameters.

For comparison purposes, the light source to output fiber direct light coupling efficiency was also
studied. The direct coupling system was similar to the microlens aided system shown in Figure 3-2,
however, there was no focusing microlens and the optical fiber was moved to the microlens position.
The aperture stop size remained the same to ensure that the same total amount of rays was used in the
coupling efficiency calculation for both systems. Under the same

simulation conditions, a snap shot on the output fiber end face was taken in both systems (Figure 3-3).
Each small dot in the center areas of both pictures represents a ray which was supported by the output
fiber and contributed to the final optical signal. It is obvious to see that microlens aided system has
much higher coupled ray density (Figure 3-3 (a)) than the direct coupling system (Figure 3-3 (b)), thus
a higher coupling efficiency was achievable.

With a comprehensive consideration of fluid dynamics (personal communication with Dr. Udoetok),
optical simulation results and fabrication feasibility, the critical feature dimensions were decided as
following: all three inlets were 1000 m in both height and width; the injection nozzle was 100m by
10m; three slopes were inclined 30o with respect to the substrate plane; the outlet channel was 500m by
500m; the microlens array was placed about 450 m away from the outlet channel center; microlens
pixel diameters ranging from 250 m to 450 m were tried; the distance from the output fiber to the
microlens back surface was estimated at about mm for the highest coupling efficiency.

Although the distance between the microlens and the sample cell light source was optimized, practically,
it was very difficult to maintain the sample cells right at the center of the outlet channel. It is thus very
important to study how the sample cells being off from the outlet center can affect the coupling
efficiency.Two types of offset were studied. The first one was the longitudinal offset caused by the
slight asymmetry of two sheath flows from right and left sides. When the sheath flow from one side is
stronger, sample cells, instead of being focused at the center, are pushed to the opposite side.

system and the microlens aided system. For the direct coupling system, the coupling efficiency stays
low but slight increases as the light source moves toward the output fiber. In contrast, with the
microlens as the focusing element, the coupling efficiency is much higher. Coupling efficiency drops
from about 60% to about 25% when point light source moves 90 m closer to the microlens and drops
to about 40% when the point light source moves 90m away.

The other common offset was radial offset related to the hydro-focusing in vertical direction. Non-
sufficient vertical hydrodynamic force may cause sample cells being focused in a position lower than the
center of the flow. Detection system with microlens is very sensitive to radial offset, coupling efficiency
drops to nearly 0 for only 60m offset from the center. In the meantime, direct coupling system is not so
sensitive to radial offset, but the overall coupling efficiency is low.
It should be mentioned that only fluorescent signal was considered in this optical analysis. The
intensity distribution function for the scattering light may be greatly different from the light intensity
distribution from a fluorescent bead. More sophisticated light source model should be used to analyze
the coupling efficiency for scattering light. Also, since fluorescent bead is modeled as a point source,
this analysis works better for fluorescent beads with diameter smaller than the diffraction limit of the
optical detection system (about 1.5 m). For larger beads, the point source assumption may cause
error to the coupling efficiency, however, the efficiency trend remains the same and the result can be
used as a reference for the system design