Dept.

of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

UNIT 1
INTRODUCTION

Intrinsic Characteristics of MEMS – Energy Domains and Transducers- Sensors and

Actuators Introduction to Micro fabrication - Silicon based MEMS processes – New
Materials – Review of Electrical and Mechanical concepts in MEMS – Semiconductor
devices – Stress and strain analysis Flexural beam bending- Torsional deflection.

THE INTRINSIC CHARACTERISTICS OF MEMS

There is no doubt that MEMS will continue to find major new applications in the future. The
reason for technology development and commercialization may vary by case. Nevertheless,
there are three generic and distinct merits for MEMS devices and micro fabrication
technologies: Miniaturization, Microelectronics Integration, and Parallel fabrication with
high precision. MEMS products will compete in the market place on the grounds of
functional richness, small sizes, unique performance characteristics (e.g., fast speed), and/or
low cost. For advanced students of MEMS, it is important to realize that the three merits will
not automatically lead to product and market advantages. One must understand the complex
interplay between these elements to fully unleash the power of MEMS technology.

Miniaturization

The length scale of typical MEMS devices generally ranges from 1 mm to 1 cm. (However, a
large array of MEMS devices or an entire system may occupy much bigger footprint or
volume.) Small dimensions give rise to many operational advantages, such as soft springs,
high resonance frequency, greater sensitivity, and low thermal mass. For example, the heat
transfer to and from a micro machined device is generally fast. A case in mind is the ink jet
printer nozzle, with the time constant of droplet ejection being on the order of 20 ms. Small
size allows MEMS devices to be less intrusive in biomedical applications (e.g., neuron
probes). Being small also means that MEMS devices can be integrated nonintrusively in
crucial systems, such as portable electronics, medical instruments, and implants (e.g., capsule
endoscopes). From a practical point of view, smaller device footprint leads to more devices
per wafer and greater economy of scale. Hence the cost of MEMS devices generally scales
favorably with miniaturization. However, all things do not work better when miniaturized.
Some physical phenomena do not scale favorably when the dimensions are reduced, while
certain physical phenomena that work poorly at the macroscale suddenly becomes very
practical and attractive at the microscale. Scaling laws are observation about how physics
work at different sizes. A well-known example is that fleas can jump dozens of times its own
height whereas elephants cannot jump at all, even though an elephant has far more muscle
mass than a flea. A rigorous scaling-law analysis starts with the identification of a
characteristics length scale (denoted L) for a device of interest. For example, the length of a
cantilever or the diameter of a circular membrane may be denoted as L of the respective
element. The remaining pertinent physical dimensions are assumed to scale linearly with the
characteristics length scale, with locked ratios. A performance merit of interest (e.g., stiffness
of a cantilever or resonant frequency of a membrane) is expressed as a function of L, with
dimension terms other than the characteristic length scale term expressed as a fraction or
multiples of L. The expression is then simplified to extract the overall effect of L.

Microelectronics Integration
Circuits are used to process sensor signals, provide power and control, improve the signal
qualities, or interface with control/computer electronics. MEMS products today are

1
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

increasingly being embedded with computing, networking, and decision-making capabilities.

By integrating micromechanical devices with electronics circuitry and offering the combined
system as a product, significant advantages can be produced in a competitive market place.
The ability to seamlessly integrate mechanical sensors and actuators with electronics
processors and controllers at the single wafer level is one of the most unique characteristics
of MEMS. This process paradigm is referred to as monolithic integration—fabrication of
various components on a single substrate in an unbroken, wafer-level process flow. (The
word “monolithic” means “one stone”. Hence “monolithic fabrication” means fabrication on
one piece of wafer.) Though not all MEMS devices should adhere or have adhered to the
monolithic integration format, it is observed that silicon circuits that are monolithically
integrated with mechanical elements have been involved in several successful commercial
MEMS applications, such as Analog Devices accelerometers, digital light processors, and ink
jet printer heads. Monolithic processes do not involve hybrid assembly methods such as
robotics pick-and-place or any manual attachment of individual parts. Dimensions and
precision of placement are guaranteed by lithography. Monolithic integration improves the
quality of signals by reducing the length of signal paths and noise. Monolithic integration
with circuits is arguably the only way by which a large and dense array of sensors or
actuators can be addressed. In the case of DLP, for example, each mirror is controlled by a
CMOS logic circuit that is buried directly underneath.Without the integration of circuits, it is
impossible to address individual mirrors in such a large and dense array.

Parallel Fabrication with Precision

MEMS technology can realize two- or three-dimensional features with small dimensions and
precision that cannot be reproducibly, efficiently, or profitably made with traditional
machining tools. Combined with photolithography, MEMS technology can be used to realize
unique threedimensional features such as inverted pyramid cavities, high aspect ratio
trenches, throughwafer holes, cantilevers, and membranes.To make these features using
traditional machining or manufacturing methods is prohibitively difficult and inefficient.
MEMS and Microelectronics are also different from traditional machining, in that multiple
copies of identical dies are made on a same wafer. This practice can contribute to lowering
the cost of individual units. Modern lithography systems and techniques provide not only
finely defined features, but also uniformity across wafers and batches.

Energy Domains and Transducers

MEMS technology enables revolutionary sensors and actuators. In general terms, sensors are
devices that detect and monitor physical or chemical phenomenon, whereas actuators are
ones that produce mechanical motion, force, or torque. Sensing can be broadly defined as
energy transduction processes that result in perception, whereas actuation is energy
transduction processes that produce actions. Sensors and actuators are collectively referred to
as transducers, which serve the function of transforming signals or power from one energy
domain to another. There are six major energy domains of interests: (1) electrical domain
(denoted E); (2) mechanical domain (Mec); (3) chemical domain (C); (4) radiative domain
(R); (5) magnetic domain (Mag); and (6) thermal domain (T). These energy domains and
commonly encountered parameters within them are summarized in Figure.The total energy
within a system can coexist in several domains and can shift among various domains under
right circumstances.

2
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Sensors generally transform stimulus signals in various energy domains to one that is
detectable by humans or into the electrical domain for interfacing with electronics controllers,
recorders, or computers. For example, a thermal-couple temperature sensor transforms a
thermal signal, temperature, into an electrical signal (e.g., voltage) that can be read
electronically. Often, more than one sensing principles can be used for a transduction
task.Temperature variation can be perceived via such phenomenon as resistance changes,
volume expansion of fluids, increased radiation power of an object, color change of
engineered dyes, shifted resonance frequency of resonant beams, or greater chemical
reactivity. Energy transduction pathways for particular sensor and actuation tasks do not have
to involve only two domains. Rather, the transduction process may incorporate multiple
domains. Direct transduction pathways that involve the minimal number of domains do not
necessarily translate into simpler device, lower cost, or better performances.
Energy and signal transduction is a vast space of research and development and a continuing
source of innovation. The desire to discover and implement efficient, sensitive, and lowcost
sensing principles transcends the boundary of scientific and technological disciplines.
Because many sensing tasks can be achieved in more than one ways, either directly (from one
energy domain to another) or indirectly (hopping through intermediate energy domains),
there is essentially unlimited number of transduction pathways for achieving one sensor or
actuator need. Each transduction pathway entails different sensing material, fabrication
method, design, sensitivity, responsivity, temperature stability, cross-sensitivity, and cost,
among others.A tradeoff study must be conducted, taking account of= performance, cost,
manufacturing ease, robustness, and, increasingly more important these days, intellectual
property rights. The development of sensors and actuators is a rich and rewarding research
experience.To invent a new sensor principle for a particular application involves selecting or
inventing the energy transduction paths, device designs, and fabrication methods that yield
simple transduction materials, high performance, and low-cost fabrication. I will discuss a

3
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

few specific examples of sensors to illustrate the richness of this field and to exemplify the
excitement involved with research and development activities. In many cases, new sensing
methods resulted in new device capabilities and industrialization opportunities.
Acceleration sensing (Mec--E transduction). Acceleration can be sensed in many different
ways. A micromachined proof mass suspended by cantilevers will experience an inertial
force under an applied acceleration. The force will cause movement of the suspended proof
mass.The movement can be picked up using piezoresistors, resistor elements whose
resistance change under applied stress (Mec--E). The displacement can also be sensed with a
capacitor (Mec--E). This is the principle of Analog Device accelerometers. These two
methods involve moving mechanical mass. Can one build accelerometers without moving
parts? The answer is yes. I will illustrate one example in the following. Inertial force can also
move a heated mass, whose ensuing displacement can be picked up by temperature sensors
(Mec--T--E).Thermal sensing does not provide as good a performance as capacitive sensing
of moving air mass, but the fabrication is readily compatible with integrated circuits. This is
the principle of a lowcost acceleration sensor (manufactured by MEMSIC Corporation)
designed for low-sensitivity applications. No moving mass is required, eliminating concerns
of mechanical reliability. Since no moving mass is needed, the device is compatible with
mass batch microelectronics foundries, reducing the time to market significantly.
Olfactory sensing (C--E transduction). Information about the presence and concentration of
certain molecules responsible for smell or pertaining to environmental monitoring can be
obtained using a number of strategies.A carbon-based material can be designed to specifically
absorb certain molecules and alter the electrical resistivity (C→E direct transduction).The
absorbance of certain molecules in the path of surface acoustic wave devices can alter
mechanical properties such as frequency of surface acoustic wave transmission (C→M→E).
These methods generally involve sophisticated electronics or algorithms. Can one build
olfactory sensors that are simpler and more intuitive? I will illustrate one example below.The
binding of chemical molecules can also alter the color of a specially designed chemical
compound, which can be detected using low-cost optoelectronics diodes (C→R→E
transduction) or directly by human beings without electronics (C→R). Sensors based on this
strategy are being made by ChemSensing Corporation. DNA sequence identification (C→E
transduction). DNA molecules consist of a chain of base pairs, each with four possible
varieties—A,C,G, or T.The sequence of base pairs in a DNA chain determines the code of
synthesizing proteins. The ability to decipher base pair sequences of DNA molecules rapidly,
accurately, and inexpensively is of critical importance for pharmaceutical and medical
applications.There are a wide variety of innovative methods for detection of DNA sequence
through their telltale binding (hybridization) events. Certain DNA molecules may be
chemically modified to incorporate (tagged) fluorescence reporters that lights or dims upon
binding with another DNA strand. In the most widely practiced case today, chemical binding
events are turned into optical signals first before transduced to the electrical domain
(C→R→E).The fluorescent image is captured using high power fluorescent microscopes.
However, fluorescent imaging requires sophisticated microscope and is not suitable for
portable, field applications.DNA molecules attached to gold nanoparticles can report the
event of hybridization through aggregation of gold particles, which can result in changes of
optical reflectance (C→R→E) [83] or electrical resistivity (C→E) [84]. The detection
method with gold nanoparticles provides better sensitivity and selectivity compared with
fluorescence methods while eliminating the need of cumbersome fluorescent imaging
instruments. It is, therefore, amendable for miniaturization and remote deployment.This
principle is the technological basis of Nanosphere Corporation.

Sensors Considerations

4
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Sensors fall into two categories, physical sensors and chemical/biological sensors. Physical
sensors are used to measure physical variables such as force, acceleration, pressure,
temperature, flow rate, acoustic vibration, and magnetic field strength. Chemical sensors are
used to detect chemical and biological variables including concentrations of chemicals, pH,
binding strength of biological molecules, protein–protein interactions, and so forth.
Many sensing principles might be available for a given application. Sensor developers
generally must evaluate a number of transducing pathways and designs according to many
sensor performance metrics.The most important sensor characteristics of concern are
summarized in the following:
1. Sensitivity. The sensitivity is defined as the ratio between the magnitude of output signal
and that of the input stimulus. Note that the sensitivity values may be a function of the input
amplitude and frequency, temperature, biasing level, and other variables. When electronics
signal amplification is used, it is meaningful to distinguish values of sensitivity before and
after amplification.
2. Linearity. If the output signal changes proportionally with respect to the input signal, the
response is said to be linear. Linear response of a sensor alleviates the complexity of signal
processing.
3. Accuracy. The ability of a sensor to provide results close to the true value.
4. Precision. The ability of a sensor to give the same reading when repeatedly measuring the
same quantity under the same conditions. Repeatability is the precision of a device over a
short term, whereas reproducibility is the precision of a device over a long term.
5. Responsivity, or resolution. It is also known as the detection limit or minimal detectable
signal (MDS). This term signifies the smallest signal a sensor can detect with confidence. It is
generally limited by noise associated with the transduction elements and circuits.
6. Noise. Noise can be applied to anything that obscures a desired signal. Noise can itself be
another signal (“interference”); most often, however, we use the term to describe “random”
noise of a physical (often thermal) origin.While interference noise can be corrected or
eliminated, such as by careful electrical shielding, random noises are ubiquitous and have
much more fundamental origins. Noise can also arrive from the circuits. Amplifiers, resistors,
parasitics capacitors and inductors in the circuitry can all generate noise with their unique
signatures.
7. Dynamic range. The dynamic range is the ratio between the highest and the lowest
detectable signal levels. In many applications, a wide dynamic range is desired.
8. Bandwidth. The bandwidth characterizes sensor ability to measure fast-changing signals.
Sensors behave differently to constant or time-varying signals. Oftentimes, sensors may cease
to respond to signals of extremely high frequencies.The effective frequency range is called
the bandwidth.
9. Drift. Drift may occur because electrical and mechanical properties of materials vary over
time. Sensors with large drift cannot be used successfully to detect slow changing signals,
such as monitoring stress building up in a civil structure over time.

Actuators Considerations
Actuators generally transform energy in non-mechanical energy domains into the mechanical
domain. For a particular actuation task, there could be several energy transduction
mechanisms. For example, one can generate a mechanical movement by using electrostatic
forces, magnetic forces, piezoelectricity, or thermal expansion. There are many other
actuation methods, including pneumatics shape memory alloys, thermal expansion, phase
change, electrochemical reaction and energetic combustion, and friction drag by moving
fluids. Microstructure can also be coupled to mesoscopic drivers (e.g., with interlocking
mechanisms).

5
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The following are general criteria when considering actuators designs and selections:
1. Torque and force output capacity. The actuator must provide sufficient force or torque for
the task at hand. For example, micro optical mirrors are used to deflect photons. Because
photons are lightweight, low levels of force provided by the mirror actuator is sufficient. In
some cases, micro actuators are used for interacting with a fluid (air or water) to actively
control the fluid. Such actuators must provide greater force and power to produce appreciable
effects.
2. Range of motion. The amount of translation or angular movement that the actuator can
produce under reasonable conditions and power consumption is an important concern. For
example, the DLP micromirrors are required to move within a 15-degree range. For optical
switches used for dynamic network routing, larger angles of displacement (30–45degrees) are
needed.
3. Dynamic response speed and bandwidth. The actuator must be able to produce sufficiently
fast response. From the point of view of actuator control, the intrinsic resonant frequency of
an actuator device should be greater than the maximum oscillation frequency.
4. Ease of fabrication and availability of materials. To reduce the potential costs of MEMS
actuators, there are two important strategies. One is to reduce the costs of materials and
processing time. Another is to increase the process yield for a given process in order to
produce more functional units in each batch.
5. Power consumption and energy efficiency. Many microactuators are envisioned for use in
small and mobile systems platforms. The total available power for such systems is generally
limited. In this and many other MEMS applications, low-power actuators are preferred to
increase the duration of operation.
6. Linearity of displacement as a function of driving bias. If the displacement varies with
input power or voltage in a linear fashion, the control strategy would be simplified.
7. Cross-sensitivity and environmental stability. The actuator must be stable over the long
term, against temperature variation, humidity absorption, and mechanical creep. Longterm
stability of such actuators is extremely important for ensuring commercial competitiveness
and success. A mechanical element may produce displacement, force, or torque in a
nonintended axis.
8. Footprint. The footprint of an actuator is the total chip area it occupies. In cases of dense
actuator arrays, the footprint of each actuator becomes a primary point of consideration.

OVERVIEW OF MICROFABRICATION
MEMS and IC devices are generally made on single crystal silicon wafers. Figure diagrams
the overall process from the production of such wafers to packaging of individual device
chips. Bulk silicon with crystalline consistency does not exist in nature, and must be prepared
through laborious industrial processes. To make bulk crystal silicon, one starts with a perfect
single crystal silicon seed. It is dipped into a molten silicon pool and slowly drawn out of the
liquid. Silicon crystallizes when drawn into the atmosphere and establishes crystallinity
consistent with that of the initial seed. Rods of single crystal with various diameters and
longitudinal crystal orientation can be formed this way.The rods are sawed into thin, circular
slices and polished to form wafers.
A wafer goes through a multi-step fabrication process in a clean room, where dust, particles,
and even ions in water are tightly controlled. The cleanliness of air in a clean room is
classified according to the concentration of air-borne particles (with sizes larger than 0.5
mm).According to a standard method for characterizing the cleanliness level of a cleanroom,
a class 1 cleanroom has fewer than 1 particle and a class 100 cleanroom has fewer than 100
particles per ft3 of air sampled. As a reference, average outdoor air contains more than
400,000 particles per ft3. Generally speaking, a class 1000, 100, 10, 1, and 0.1 cleanroom can

6
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

support production down to approximately 4 mm, 1.25 mm, 0.7 mm, 0.3 mm, and <0.1 mm,
respectively. State of the art integrated circuits uses linewidth smaller than 0.18 mm since
1999.

Water (for rinsing) and chemical solutions must go through stringent, costly manufacturing
and conditioning processes. Ions in water (e.g., sodium ions), even in trace amount, will
migrate into silicon and thin film materials upon direct contact. These ions may become
trapped charges in dielectrics, and hurt device performance. De-ionized water (more broadly
speaking, ultrapure water) used in semiconductor manufacturing has resistivity in excess of
18 Mohm, compared to a resistivity of less than 50 kohm for tap waters. Precision patterns
are made using photolithographic patterning method. Collimated light passes through a mask
and an image-reduction lens before hitting the wafer, in a process akin to taking a photograph
of an object through a telephoto length and recording the image on a photosensitive film. In
this case, the object being photographed is the mask and the film is the wafer, coated with a
photosensitive film. Various wavelength of light can be used. Light with
higher energy (and hence smaller wavelength) is capable of producing smaller linewidths.
The ultimate resolution is dictated by the diffraction limit. Using a machine-automated
photolithography process called step-and-repeat, many identical units can be made on a same
wafer with high linewidth resolution (0.1 mm or smaller in commercial processes).The
machine used for performing the step-and-repeat process is called a stepper.The way a
stepper works is described as follows. After a reduced image of the mask is printed to an area
on a wafer, the wafer is translated by a precise distance, and another exposure is made. Many
identical devices, called dies, are made on one given wafer in a single pass. This
differentiates MEMS process from conventional manufacturing technology, which generally
deals with one part at a time.

The larger the wafer, the more dies are made in a given batch. According to Figure , if the die
size is 1x1cm2, there could be roughly 21 dies on a 2 wafer, 82 on a 4 wafer, 178 on a 6
wafer, 314 on an 8 wafer, and 770 on a 10 one. The economy of scale when moving to a
larger wafer is compelling. However, readers should note that there is cost associated with
upgrading the wafer sizes.A piece of machine designed to handle a 4 dia wafer will not be
able to handle a 6 dia wafer, and so on.When the decision to upgrade the wafer size is made,
almost every piece of equipment in the entire fab line needs to be purchased, easily
amounting to investments of tens of millions of US dollars for a modest manufacturing
operation. Further, it is noteworthy that only large volume products would warrants the full
benefits of larger wafers. For small volume products, the benefit of parallel fabrication is
easily offset by the fixed cost of setup process and making high-resolution masks.

7
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

These dies have spacings between them so that they can be mechanically cut and separated.
Each cut die, called a chip, can then be electrically connected and encapsulated for
commercial resale. The process of incorporating a loose die to a housing and a system is
called packaging.
In a given process, a percentage of dies with satisfactory performance is called the process
yield.The yield deeply affects the economy of fabrication and the cost of final devices. It is
determined
by the choices of design, materials, and processes in complex ways. For industry, the yield
figure is a guarded secret. Should you approach a person from industry and ask him/her about
the yield of their process, the best answer you can get is a polite “no”. In a number of
important ways, the MEMS process and conventional macroscale machining differs
drastically:
1. Silicon, a principal substrate material for MEMS and integrated circuit, is mechanically
brittle and cannot be shaped by machine cutting tools;
2. MEMS and integrated circuits are made on planar crystalline wafers.The planarity is not
just a matter of convenience for automated wafer processing.When lithography patterning is
conducted on a planar substrate, a consistent focus distance is ensured, leading to uniformity
and resolution.The planarity of a wafer is also important to ensure the entire wafer surface
has identical crystal orientation.
3. Loose dies are generally too small and numerous to be handled by human labor. They must
be compatible with automated sorting and pick-and-place machines. A chip is placed into a
package, which is then mounted on a circuit board (Figure).

Take the example of a smart phone (Figure), a consumer electronics system with many
components, including display, battery, computer chip, camera, and sensors (e.g., inclination

8
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

sensors, motion sensors, microphone, touch screen, etc). A microphone chip is highlighted in
the diagram and shown in zoomed view in the lower half of the figure.The chip consists of a
package casing within which the diced silicon die is mechanically mounted and electrically
connected by bonding wires (Figure). There are many possible variants for integrating
mechanical elements with circuitry.

Photolithography
The goal of photolithography is to produce fine features on wafer surfaces.A most common
lithography process involves depositing photo-sensitive chemicals (called photoresists, or
simply resists) on a silicon wafer, exposing it with light through a mask, and removing
(develop) photoresist material that has been modified by light.The starting point of a
lithography process is to coat a wafer with photoresist through spin coating (Figure ). A wafer
is held on a rotating stage. Photoresist is applied to the center of the wafer at rest position.The
wafer is then spun at high speed, causing the photoresist to move towards the edge of the
wafer under centrifugal forces. After the wafer spinning is stopped, a uniform thin layer of
photoresist is coated on the front surface of a wafer. Process variables include the wafer
spinning speed, the viscosity of the resist, and the types of resists (e.g., target wavelength,
sensitivity).Typical thickness of photoresist is generally 1–10 mm. A lithography patterning
procedure involves multiple steps (Figure ). A wafer is first covered with a uniform thin layer
of resist (step a).A mask, consisting of a transparent substrate (e.g., glass or quartz) with
opaque features, are brought close to the resist-coated wafer (step b). High energy, collimated
light rays strikes the mask-wafer assembly. Resist regions that are not covered by opaque
features are exposed, changing the chemical composition of the resist. For positive resist, the
exposure by light causes the resist to be more soluble in a wet chemical developer (step c).
This allows the opaque features on the mask to be faithfully transferred to the wafer (step d).
A pattern in photoresist can be further transferred to an underlying layer, using the
photoresist as a mask layer.The process is shown in Figure . It starts with a wafer with a thin

9
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

film coating (step a).The wafer is covered with a spin-coated layer of photoresist (step b),
which is photolithographically patterned similar to the method discussed in Figure .The wafer
is then immersed in a chemical solution that preferably etch the thin film but not the
photoresist (step c). (Alternatively, the thin film may be etched with dry etch methods.) With
proper timing control, the thin film covered by the photoresist would stay intact, whereas the
thin film in areas not covered by the photoresist would be removed (step d). The photoresist
is then preferentially removed, leaving the thin film behind as patterned (step e).

Thin Film Deposition

Functional materials, conductors and insulators can be incorporated on a wafer through
additive deposition process. One such deposition process is direct transfer of the material
from a source to the wafer surface in an atom-by-atom, layer-by-layer fashion (Figure ).
Examples include metal evaporation and metal sputtering.The process is generally conducted
in a lowpressure environment so that atoms may travel from the source to the wafer surface
without interruptions caused by air molecules. One such system, an evaporator, is
diagrammed in . A wafer and a metal source are both placed inside a vacuum jar.The metal
can be transferred either by heating it (evaporation) or by bombarding it with high-energy
ions (sputtering). The achieved thickness is proportional to the power and time. In practice,
the routine thickness of metal thin films ranges from 1 nm to 2 mm. A second method for
placing thin film materials on a wafer surface is chemical vapor deposition. Two or more
active species arrive at the vicinity of the wafer surface (step a). They react under favorable
conditions (with energy provided by heating or plasma). The reaction of these species
produces a solid phase, which is absorbed onto the nearby wafer surface (step b). The
byproducts of the reaction (if any) may be removed by the surrounding media. Continuous
reaction causes a layer of material to be built on the wafer surface (step c). Typically the
average thickness of thin film deposited by CVD, evaporation or sputtering is below 1
mm.To deposit films of greater thickness is typically too time consuming or impractical.

10
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

SILICON-BASED MEMS PROCESSES

MEMS devices were first developed on silicon wafers because of the easy availability of
mature processing technologies that had been developed within the microelectronics industry,
and the availability of expertise in process management and quality control. Silicon actually
comes in three general forms: single crystal silicon, polycrystalline silicon, and amorphous
silicon. In a single crystal silicon (SCS) material, the crystal lattice is regularly organized
throughout the entire bulk. Single crystal silicon is often encountered in three cases: (1)
single crystal silicon wafer grown from a high-temperature melt/recrystallization process ; (2)
epitaxially grown silicon thin films; (3) single crystal silicon obtained from recrystalizing
polycrystalline or amorphous silicon by global or local heat treatment. A polycrystalline
silicon material (so called polysilicon, polySi, or poly) is made of multiple crystalline
domains.Within each individual domain, the crystal lattice is regularly aligned. However,
crystal orientations are different in neighboring domains. Domain walls, also referred to as
grain boundaries, play important roles in determining electrical conductivity, mechanical
stiffness, and chemical etching characteristics. The polysilicon material can be grown by low
pressure chemical vapor deposition (LPCVD), or by recrystalizing amorphous silicon through
global or local heat treatment. Amorphous silicon, on the other hand, exhibits no crystalline
regularity.Amorphous silicon films can be deposited by chemical vapor deposition methods
(CVD), at a lower temperature than that required to deposit polysilicon. Due to low
temperature, atoms do not have enough vibrational energy to align themselves after they are
incorporated into the solid. Amorphous silicon can be grown using the LPCVD method. (In a

11
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

typical, horizontal, low-pressure reactor, the transition temperature above which

polycrystalline structure forms during deposition is580°C. Amorphous silicon can be formed
by plasma enhanced chemical vapor deposition (PECVD) method as well.The two most
fundamental classes of fabrication technologies are bulk micromachining and surface
micromachining. Bulk micromachining processes involve selectivelyremoving the bulk
(silicon substrate) material in order to form certain three-dimensionalfeatures or mechanical
elements, such as beams and membranes. Bulk micromachining maybe combined with wafer
bonding to create even more complex three-dimensional structures.We use the example of a
micromachined pressure sensor to illustrate a representativeMEMS process.The process
involves two wafers—a bottom wafer is etched toform a cavity whereas a top wafer is used to
make the membrane.

A description for each step inthe diagram follows.Although the process diagramed in may be
quite lengthy for beginners, this is still a compressed process description because many
detailed steps are skipped, especially the routine photolithography sequence, which typically
consist of several steps themselves: photoresist coating, curing, exposure, development,
subsequent etching of underlying layers using the photoresist as a mask, and removal of the
photoresist.
Step a. The process starts with a bare silicon wafer. To create desired cavity shapes, the
wafer must be of a certain crystallographic orientation. The wafer is cleaned thoroughly to
remove any large particles, dirt particles, and invisible organic residues.A combined
mechanical wash and oxidizing acid bath may be used, followed by a rinse by ultrapure
water.
Step b. The cleaned wafer is placed inside a high-temperature furnace filled with running
oxygen gas or water vapor. Oxygen atoms present in the air or dissociated from the water
molecule will react with silicon to form a protective silicon dioxide thin film. Note the oxide
is grown on both sides of the wafer as well as on the edges. (For the sake of completeness, all
layers on both side of the wafer are shown in .)
Step c. The wafer is removed from the furnace and cooled to room temperature. It will be
very clean, because any organic molecules would have been decomposed in the
hightemperature oxidation step. A layer of thin film photoresist is deposited on the front
surface of the wafer. (A chemical called hexamethyldisilazane, or HDMS, is sometimes spin-
or vapor-coated to help increase the adhesion between the photoresist and an oxide surface.)
The photoresist is typically spin coated. Alternatively, photoresist thin film can be deposited
by vapour coating, mist coating, or electroplating. The wafer is baked in a convection oven to
remove some portion of the solvent from the photoresist (PR) layer to establish firmness.
This step is generally called “soft bake”. Alternatively, the moisture can be driven off with
infrared lamp, or vacuum.

12
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Step d. The photoresist is exposed through a mask with a high-energy radiation (such as
ultraviolet ray, electron beam, or X-ray). The entire wafer is then placed inside a developing
solution (often called developer) that removes loosely bound photosensitive polymer. In the
case of positive photoresist, regions hit by light will be dissolved. In the case of negative
photoresist, regions hit by light will stay.The soft bake process in step c ensures that the
photoresist will not be indiscriminately
stripped by the developer.
Step e. The photoresist needs to be baked again, this time at a higher temperature and often
for a longer duration than the soft bake. This second baking step, called “hard bake”, removes
remaining solvents and makes photoresist that remains on the wafer stick to the wafer even
stronger. The extent of the hard bake will depend on the nature of the subsequent step.
The photoresist mask is here used to selectively mask the underlying layer, the silicon oxide,
against a hydrofluoric acid etchant bath. A HF etchant attacks oxide within the exposed
window, but has negligible etch rate on the underlying silicon and the photoresist mask.
Step f. The photoresist is removed using an organic solvent etchant such as acetone (at room
temperature or elevated temperatures). The hard-baked photoresist is chemically resistant to
the HF etchant but not to acetone.The organic solvent does not etch the oxide and the silicon.
Step g. The silicon wafer is immersed in a wet silicon etchant, which does not attack the
silicon oxide. Only the silicon in the open oxide window is etched, resulting in a cavity with
sidewalls defined by crystallographic planes. The cavity may reach the other side of the wafer
if the open window is large enough for the given wafer thickness.
The wet silicon etching involves an elevated temperature (70–90°C). The etchant would
attack hard-baked photoresist, hence it is impossible to use photoresist directly as mask
in this step. The oxide is removed using HF etchant again.
Step h. The wafer at the end of stage (g) is tilted to provide a clear view of the throughwafer
cavity.
Step i. A second silicon wafer is firmly bond to the frontside of the bottom wafer processed
through step (g). It is important that the environment in which the wafers are processed be
very clean, because tiny particles adhering to the bonding surfaces of either wafer will
prevent good bond strength from being reached.
Step j. The bonded top wafer is thinned by using mechanical polishing or chemical etching.
The remaining thickness of the top wafer determines the thickness of the membrane. Thin
membranes are desired to have high sensitivity.
Step k. Strain sensors are then made on the prepared membrane.A thin film layer (e.g., oxide)
is deposited and patterned. It serves as a barrier layer to ion implantation. Areas on the silicon
wafer hit directly by energetic dopant ions will become doped and form a piezoresistor,
which changes its resistance upon applied stress due to membrane bending under pressure
difference. The wafer is tilted to present another perspective view of the through-wafer cavity
(m). To keep the description succinct, a few detailed steps are skipped between steps j and k.

13
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

PACKAGING AND INTEGRATION

The two terms integration and packaging are closely related but they are in fact different.
Integration refers to the act of combining mechanical and electrical functionalities. Packaging
refers to the act of placing loose (diced) chips into human- or machine handlable modules
that can be directly assembled on circuit boards and into systems. Packaging refers to dicing,
die assembly, encapsulation, and testing. Packaging is a crucial technological issues highly
relevant for MEMS device design. Although it is not discussed in detail in this book, it is
important to draw attention of beginning students to this very exciting, important, and
dynamically changing area. “Packaging is part of the process. Packaging is not what you do
after the process.” should be the motto learned by all students of MEMS. The cost of
packaging often accounts for 30–90% of the total cost of the finished system, and it deeply
affect the performance, cost, reliability, and ultimately, the competitiveness of a MEMS
device in the market place.

Integration Options
In a purest definition of MEMS, the circuits and the mechanical elements are co-fabricated
monolithically on the same silicon die. In practice, there are actually multiple options for
integrating the circuits with the mechanical components. These are diagrammed in . Circuits
can be integrated with MEMS devices at the wafer level, at the package level, or at the board
level.

Wafer-level integration Micromechanical components and integrated circuits are either

placed on the same wafer (monolithic) or integrated through die attachment (e.g., wafer
bonding). For monolithic integration, there are a number of variations:
(1) microelectronics and micromechanical elements are placed side by side (e.g., Analog
Devices accelerometers);

14
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

(2) microelectronics and micromechanical elements are placed top-bottm (e.g., digital micro
mirror array).
The monolithic integration of mechanical and circuit element, if necessary, may follow one
of the following major approaches:
1. Post-processing approach. Micro electromechanical elements are fabricated on the top
surface of a semiconductor wafer with preexisting circuits.
2. Pre-processing approach. The micro electromechanical elements are fabricated on a
wafer first, followed by IC fabrication.
3. Side-by-side processing approach. Micro electromechanical elements and integrated
electronics are created simultaneously. One major advantage for wafer-level integration lies
in the fact that the circuits and mechanical components are close, hence there is little chance
for electromagnetic noise and other noises to develop. The potential for creating small
dimensions of package is the highest for wafer-level integration.
Disadvantages for monolithic integration include:
1. chip foot print mismatch. The micromechanical components could be much larger than the
circuit elements, thus increasing the overall footprint of the die, and reducing the number of
dies produced on a wafer.
2. material complexity and possible reduction of yield. Since the circuit and micromechanical
elements are fabricated on the same wafer, their materials and processing are interwoven.
Often this introduces complexity for fabrication and manufacturing, and may increase costs.
Chip-level integration For this packaging scheme, micromechanical and circuit dies are
made separately.The dies are inserted into the same package and connected electrically with
bonding wires. Hence the materials and processing issues are not coupled. This multiple-
chip-module (MCM) approach involves longer distance between circuitry to
micromechanical elements. However, it avoids issues that often plagues the wafer-level
packaging scheme.
Board-level integration This package scheme involves connecting micromechanical
packages and circuit packages on the circuit board level. It involves the longest distance of
wiring and largest electronics noise, and offers the least potential for miniaturization.

Encapsulation
Many MEMS devices cann’t work in atmospherical environment.They must be placed
hermatically into a stable and safe environment (e.g., vacuum, or controlled
pressure/humidity). The act of sealing the MEMS devices into a controlled environment is
referred to as encapsulation (or sealing, vacuum packaging). Commercially successful
MEMS products, such as DLP micro mirrors, accelerometers, gyros, and resonators, are all
encapsulated in controlled environment or low pressure environment. The quality and cost of
the packaging plays a decisive role about the cost, reliability, and competitiveness of the
products. Encapsulation can be performed at the wafer level or the package level.Wafer-level
encapsulation allows the greatest potential for miniaturization and automation. However, the
technology for wafer-level encapsulation is challenging.
When designing a MEMS product, the decision on integration and packaging should be
made prior to the design and process flow is drawn.

NEW MATERIALS AND FABRICATION PROCESSES

In recent years, silicon micromachining techniques are being rapidly augmented with new
materials and processes. Silicon, a semiconductor material, is mechanically brittle. It is also
expensive or unnecessary for certain applications. New materials such as polymer and
compound semiconductors can fill the gap of performance. Polymer materials are being
incorporated into MEMS because of their unique materials properties (e.g., biocompatibility,

15
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

optical transparency), processing techniques, and low costs compared to silicon. Polymer
materials that have been explored in recent years include silicone elastomers, Parylene, and
polyimide, among others.
Many sensors and actuators are needed to operate in harsh conditions, such as direct exposure
to environmental elements, high temperature, wide temperature swing, or high shock.
Delicate microstructures made of silicon or inorganic thin film materials are not suited for
such applications. Several inorganic materials are being introduced for MEMS applications in
harsh environments. Silicon carbide, in both bulk and thin film forms, are explored for
applications, including high-temperature solid-state electronics and transducers. Diamond
thin films provide the advantage of high electrical conductivity and high wear resistance for
potential applications including pressure sensors and scanning electron microscopy probes.
Other compound semiconductor materials including GaAs are also being investigated. Other
metal elements have also been involved in MEMS devices, including nickel, titanium, etc.
New material processing techniques are being developed for fabrication both on silicon
substrates and other materials. New processes for MEMS include laser-assisted etching for
material removal and deposition, stereo lithography for rapid prototyping, local
electrochemical deposition, photo-electroforming, high aspect ratio deep reactive ion etching,
micro milling, focused ion beam etching, X-ray etching, micro electro discharg, ink jet
printing (e.g., of metal colloids), micro contact= printing, in-situ plasma , molding (including
injection molding,), embossing, screen printing, electrochemical welding, chemical
mechanical polishing, micro glass blowing, and guided and self-directed self-assembly in two
or three dimensions. Microfabrication processes have also been expanded to reach nanoscale
resolution to realize nanoelectromechanical systems (NEMS). Reliable and economical
fabrication of electromechanical elements with nanoscopic feature sizes or spacing represents
new challenges and methodologies.Traditional lithography does not offer sub-100-nm
resolution readily, at low cost, and with parallelism. A variety of nanostructure patterning
techniques, often drastically different from the photolithographic approach, are developed in
the physics and chemistry communities for producing nano-meter scale patterns. Readers
who are interested in exploring this class of techniques may start by reading literatures on
nanoimprint lithography, nano whittling, and nanosphere lithography.

Semiconductor Materials
The unique electrical properties of semiconductors stem from their atomic structures. In this
section, let us look at how conductivity of semiconductor silicon is originated. Silicon is a
Group IV element in the periodic table. Each silicon atom has four electrons in its outermost
orbit. As a consequence, each silicon atom in the crystal lattice shares four covalent bonds
with four neighboring atoms. Silicon atoms reside in a crystal lattice.The inter-atomic spacing
between atoms is determined by the balance of atomic attraction and repulsive forces.The
density of silicon atoms in a solid is 5x1022 atoms/cm3at 300 K. Electrons that covalently
bond to the orbits of silicon atoms cannot conduct current and do not contribute to bulk
conductivity. The conductivity of a semiconductor material is only related to the
concentration of electrons that can freely move in the bulk. An electron bond to a silicon atom
must be excited with enough energy to escape the outermost orbit of the atom for it to
participate in bulk current conduction.The statistically minimal energy needed to excite a
covalently bonded electron to become a free charge carrier is called the bandgap of the
semiconductor material. It corresponds to the energy necessary to break a single covalent
bond between two atoms.The bandgap of silicon at room temperature is approximately 1.11
eV.Silicon is not the only semiconductor material used for MEMS. Other semiconductor
materials, such as germanium, polycrystalline germanium, silicon germanium, gallium
arsenide (GaAs), gallium nitride (GaN), and silicon carbide (SiC), have also been used.

16
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Calculation of Charge Carrier Concentration

The electrical conductivity of a semiconductor material is determined by the number of free

charge particles in a bulk and their agility under the influence of electric field. In this section,
we will discuss basic formulation for determining the number (or volumetric concentration)
of free charge carriers. There are two types of free charge carriers—electrons and holes.
Electrons that are freed by breaking covalent bonds are not the only participants in bulk
current conduction. The bond vacancy left by an escaped electron, called a hole, can facilitate
bulk current movement as a site for successive electron hopping.A hole carries a positive
charge. In a given material, the abilities of the electrons and holes to conduct electricity are
different, but they are often in the same order of magnitude.
The electron concentration in a semiconductor is denoted as n, while the hole concentration is
denoted as p. The concentrations of electrons and holes under steady state, thermal
equilibrium conditions (i.e., no external current and no ambient light) are commonly referred
as no and po respectively. The subscript 0 indicates values under thermal equilibrium.
A perfect semiconductor crystal has no impurities and lattice defects in its crystal structure. It
is called an intrinsic semiconductor material. In this type of material, electrons and holes
are created through thermal or optical excitation.When a valence electron receives enough
energy, either thermally or optically, it is freed from the silicon atom and leaves a hole
behind. This event is called electron–hole pair generation.
The number and concentrations of electrons and holes, however, do not increase without limit
over time. Free electrons and holes can recombine and give up energy along the way.This
process, called recombination, competes with the generation process. Under steady state
conditions, the generation and recombination rates are identical. Since the electrons and holes
are created in pairs, the electron concentration (n) is equal to the concentration of holes (p).
For intrinsic materials, their common value is further denoted as ni with the subscript i
standing for the word intrinsic.The relation is summarized as n=p=ni.

Conductivity and Resistivity

The conductivity of a bulk semiconductor is a measure of its ability to conduct electric
current. In this section, I will discuss formula for calculating electrical conductivity when
carrier concentrations of both types (electrons and holes) are known. The overall conductivity
of a semiconductor is the sum of conductivities contributed by these two types individually.
Free charge carriers move under the influence of an electric field. This mode of carrier
transport is called drift. The agility of charge carriers drifting under the influence of a given
field affects the conductivity of the bulk. How fast can a free charge carrier move then? A
free charge carrier in a crystal lattice does not reach arbitrarily high speed over time when it
is placed in an uniform and constant electric field. Rather, it is frequently slowed or halted
when it collides with lattice atoms and other free charge carriers.A charge carrier reaches a
statistical average velocity ( ) between collision events.The magnitude of the velocity is the
mathematical product of the magnitude of the local electric field (E) and the agility of the
carrier, usually represented by a term called the carrier mobility µ, The mobility is defined as

The values of the mobility are influenced by the doping concentration, temperature, and
crystal-orientation, in a complex manner. Certain materials (such as GaAs) offer higher
electron and hole mobilities than silicon and are used in high-speed electronic circuits. The
statistical average distance a charge carrier travels between two successive collision events is
called its mean free path, .The average time between two successive collision events is the

17
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

mean free time, .The average velocity, mean free path and mean free time are linked by the
relation

Armed with knowledge about carrier concentration and their speed, we are set to derive the
expression for the conductivity of a bulk semiconductor. We start from the familiar Ohm’s
law, which states that the bulk resistivity associated with a material is the proportionality
constant between an applied electric field and the resultant current density J,

The current density equals current divided by the cross-sectional area.The conductivity is the
reciprocal of the resistivity The relationship between the current density and the applied
electric field can be rewritten as

The conductivity ơ is explicitly defined as

CRYSTAL PLANES AND ORIENTATIONS

Silicon atoms in a crystal lattice are regularly arranged in a lattice structure. Materials
properties (such as the Young’s modulus of elasticity, mobility, and piezoresistivity) and
chemical etch rates of silicon bulk often exhibit orientation dependency.The cross-sectional
views of the silicon crystal lattice from several distinct orientations are shown in .3.
Obviously, the atom packing density is different according to different planes, giving rise to
crystal anisotropy of electrical and mechanical properties and etching characteristics.
A set of common notations, called Miller Indexes, has been developed for identifying and
visualizing planes and directions in a crystal lattice. Let us first discuss the procedure for
naming planes.A crystal plane may be defined by considering how the plane intersects the
main crystallographic axes of the solid. Silicon lattice belongs to the cubic lattice family. In a
cubic lattice, materials properties exhibit rotational symmetry. Hence (010) and (001) planes
in the lattice are equivalent to (100) plane in terms of material properties. To represent a
family of such equivalent planes, a set of integers are enclosed in braces { } instead of
parentheses ( ). For example, crystal planes (100), (010), and (001) are said to belong to the
same {100} family (.6).
The Miller Index is also used to denote directions in a crystal lattice.The Miller Index of the
direction of a vector consists of a set of three integers as well, determined by following a two-
step process outlined below:
Step 1. In many cases a vector of interest does not intercept with the origin of the coordinate
system. If this is the case, find a parallel vector that start at the origin. The Miller Indexes of
parallel vectors are identical.
Step 2. The three coordinate components of the vector intercepting with the origin are
reduced to the smallest set of integers while retaining the relationship among them. For
example, the body diagonal of the cube in .4 extending from point (0,0,0) to point (1a, 1a, 1a)
consists of three X, Y, and Z components, all being 1a. Therefore, the Miller Index of the
body diagonal consists of three integers, 1, 1, and 1, enclosed in a bracket [111]. In a cubic
lattice a vector with Miller Index [hkl] is always perpendicular to a plane (hkl). Vectors that
are rotationally symmetric belong to a family of vectors. The notation of a family of vectors
consists of three integers enclosed in <>.

18
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Definitions of Stress and Strain

Mechanical stresses fall into two categories—normal stress and shear stress.For the simplest
case of normal stress analysis, let us consider a rod with uniform crosssectional area
subjected to axial loading. If we pull on the rod in its longitudinal direction, it will experience
tension and the length of the rod will increase. The internal stress in the rod is exposed if we
make an imaginary cut through the rod at a section. At any chosen cross section, a
continuously distributed force is found acting over the entire area of the section. The intensity
of this force is called the stress. If the stress acts in a direction perpendicular to the cross
section, it is called a normal stress. The normal stress, commonly denoted as ơ is defined as
the force F applied on a given area (A),

A normal stress can be tensile (as in the case of pulling along the rod) or compressive (as in
the case of pushing along the rod). The polarity of normal stress can also be determined by
isolating an infinitesimally small volume inside the bar. If the volume is pulled in one
particular axis, the stress is tensile; if the volume is pushed, the stress is compressive. The
unit elongation of the rod represents the strain. In this case, it is called normal strain since
the direction of the strain is perpendicular to the cross section of the beam. Suppose the steel
bar has an original length denoted Lo.Under a given normal stress the rod is extended to a
length of L.The resultant strain in the bar is defined as

The strain is commonly denoted as in mechanics. However, it could easily be mistaken with
the notation reserved for electrical permitivity (dielectric constant) of a material. In most
cases, this is very clear based on the context of the discussion. In others, such as the
discussion of piezoelectricity where the constitutive equations involve both strain and
permitivity, the strain and permitivity terms must be assigned different notations to avoid
confusion. In reality, the applied longitudinal stress along the x-axis not only produces a
longitudinal elongation in the direction of the stress, but a reduction of cross-sectional area as
well. This can be explained by the argument that the material must strive to maintain constant
atomic spacing and bulk volume.The relative dimensional change in the y and z directions
can be expressed as ᶓx and ᶓy. This general material characteristic is captured by a term called
the Poisson’s ratio, which is defined as the ratio between transverse and longitudinal
elongations:

19
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Stress and strain are closely related. Under small deformation, the stress and the strain
terms are proportional to each other according to the Hooke’s law, i.e.,

The proportion constant, E, is called the modulus of elasticity. The general relation between
stress and strain over a wider range of deformation, however, is much more complicated. The
modulus of elasticity, often called the Young’s modulus, is an intrinsic property of a
material. It is a constant for a given material, irrespective of the shape and dimensions of the
mechanical element. Atoms are held together with atomic forces. If one imagines inter-
atomic force acting as springs to provide restoring force when atoms are pulled apart or
pushed together, the modulus of elasticity is the measure of the stiffness of the inter-atomic
spring near the equilibrium point.

General Scalar Relation Between Tensile Stress and Strain

The relationship between tensile stress and strain has been studied both theoretically and
experimentally for many materials, especially metals. Relationship over a wider range of
deformation is discussed in this section. In order to determine the stress–strain relation, a
tensile test is commonly used. A rod with precise dimensions, calibrated crystalline
orientation and smooth surface finish is subjected to a tension force applied in the
longitudinal direction.The amount of relative displacement and the applied stress are plotted
on a stress-strain curve until the beam breaks. A generic stress-strain curve is illustrated in
figure below. At low levels of applied stress and strain, the stress value increases
proportionally with respect to the developed strain, with the proportional constant being the
Young’s modulus.This segment of the stress-strain curve is called the elastic deformation
regime. If the stress is removed, the material will return to its original shape.This force
loading can be repeated for many times. As the stress exceeds a certain level, the material
enters the plastic deformation regime.

In this regime, the amount of stress and strain does not follow a linear relationship
anymore. Furthermore, deformation cannot be fully recovered after the external loading is
removed. Bend a metal paper clip wire slightly, it will always return to its original shape. If
the wire is bent beyond a certain angle, the clip will never return to original shapes again.
Plastic deformation is said to have occurred. Stress-strain curves for materials in compression
differ from those in tension. The stress-strain curve has two noticeable points—yield point
and fracture point. Before the yield point is reached, the material remains elastic. Between the
yield point and the fracture point, the specimen undergoes plastic deformation. At the fracture
point, the specimen suffers from irreversible failure.The y-coordinate of the yield point is the

20
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

yield strength of the material. The y-coordinate of the fracture point is designated the
ultimate strength (or the fracture strength) of the material.

For many metals the generic relationship depicted in figure is true. However, not all
materials exhibit this generic stress–strain relationship. Some representative curves for
different classes of materials are shown in figure below, including brittle materials (such as
silicon) and soft rubber, both are used extensively in MEMS.

There are several common qualitative phrases used to describe materials—strong, ductile,
resilient, and tough. These terms can be explained much more clearly by relating to the
stressstrain curves.
A material is strong if it has high yield strength or ultimate strength. By this account, silicon
is even stronger than stainless steel. Ductility is an important mechanical property. It is a
measure of the degree of plastic deformation that has been sustained at the point of fracture.A
material that experiences very little or no plastic deformation upon fracture is termed brittle.
Silicon is a brittle material, which fails in tension with only little elongation after the
proportional limit is exceeded. Ductility may be expressed quantitatively as either percent
elongation or percent reduction in area. Toughness is a mechanical measure of the ability of a

21
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

material to absorb energy up to fracture. For a static situation, toughness may be ascertained
from the result of the tensile stress-strain test. It is the area under the stress-strain curve up to
the point of fracture. For a material to be tough, it must display both strength and ductility.
Resilience is the capacity of a material to absorb energy when it is deformed elastically and
then, upon unloading, to have this energy recovered.

General Stress—Strain Relations

The formula given earlier treats stress and strain as scalars. In reality, stress and strain are
tensors. Their relationship can be conveniently expressed in matrix form in which stress and
strain are expressed as vectors. The general vector representation of stresses and strains will
be discussed in this section. To visualize vector components of stress and strain, let us isolate
a unit cube from inside a material and consider stress components acting on it.The cube is
placed in a rectangular coordinate system with axes marked as x, y, and z. For reasons of
convenience that will become apparent later, the axes x, y and z are also labeled as axes 1, 2,
and 3, respectively. A cube has six faces. Consequently, there are 12 possible shear force
components—two for each face.These are not all independent. For example, each pair of
shear stress components acting on parallel faces but along the same axis have equal
magnitude and opposite directions for force balance (Newton’s first law). This reduces the
number of independent shear stress components to six.These six shear stress components Ƭ
are graphically illustrated in .13. Each component is identified by two subscript letters.The
first letter in the subscript indicates the normal direction of the facet on which the stress is
applied to, while the second letter indicates the direction of the stress component. Based on
torque balance, two shear stress components acting on two facets but pointing towards a
common edge have the same magnitude. Specifically, and In other words, equal shear
stresses always exist on mutually perpendicular planes. The independent number of shear
stress components is reduced to three. There are six possible normal stress components—one
for each face of a cube. Under equilibrium conditions, the normal stress components acting
on opposite facets must have

the same magnitude and point to opposite directions.Therefore, there are three independent
normal stress components. Normal stress components are labelled ơ with two subscript
letters.Overall, in a rectangular coordinate system under motion equilibrium, there are three
independent normal stresses and three shear ones.

FLEXURAL BEAM BENDING ANALYSIS UNDER SIMPLE LOADING

CONDITIONS
Flexural beams are commonly encountered in MEMS as spring support elements. Essential
skills and common practices for a MEMS researcher include calculating the bending of a
22
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

beam under simple loading conditions, analyzing induced internal stress, and determining the
resonant frequency associated with the element.
Types of Beams
A beam is a structure member subjected to lateral loads, that is, forces or moments having
their vectors perpendicular to the longitudinal axis. In this textbook, we focus on planar
structures— beams that lie in single planes. In addition, all loads act in that same plane and
all deflections occur in that plane. Beams are usually described by the manner in which they
are supported. Boundary conditions pertain to the deflections and slopes at the supports of a
beam. Consider a two-dimensional beam with movement confined in one plane. Each point
along the length of the beam can have a maximum of two linear degrees of freedom (DOF)
and one rotational degree of freedom.Three possible boundary conditions are summarized
below according to their restrictions on DOFs:
1. The fixed boundary condition restricts both linear DOFs and the rotational DOF. No
movement is allowed at the support. At the fixed support, a beam can neither translate nor
rotate. Representative examples include the anchored end of a diving board or the ground end
of a flagpole.
2. The guided boundary conditions allow two linear DOFs but restrict the rotational DOF.
3. The free boundary conditions provide for both linear DOFS and rotation. At a free end, a
point on beam may translate and rotate.A representative example is the free end of a diving
board.
Flexural beams can be classified according to the combination of two mechanical boundary
conditions associated with it. For example, a beam fixed at one end and free at another is
conveniently referred to as a fixed-free beam, commonly called a cantilever. In MEMS
research, the following types of beams are the most frequently encountered: fixed-free
(cantilevers), fixed-fixed (bridges), and fixed-guided. It is important to correctly identify the
boundary conditions associated with a beam.The best way to learn and recognize boundary
conditions is through examples.

The boundary conditions for these beams are summarized below.

(a) A fixed-free cantilever parallel to the plane of the substrate, with the free tip capable of
moving in a direction normal to the substrate. Lateral, in-plane movement of the free end
would encounter much significant resistance.
(b) A fixed-fixed beam (bridge) parallel to the substrate plane.

23
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

(c) There are actually two ways to classify this beam. It can be considered a fixed-fixed
cantilever parallel to the substrate plane with a thick and stiff part in the middle.Alternatively,
this beam can also be considered as two fixed-guided beams jointed in parallel to support a
rigid part.
(d) A fixed-free cantilever with the free end capable of movement perpendicular to the
substrate. Its behavior is similar to that of (a).
(e) A fixed-free cantilever with the free end capable of movement within the substrate
plane.With its thickness greater than its width, the movement of the free end in a direction
perpendicular to the substrate plane would encounter much greater resistance.
(f) A fixed-fixed beam (bridge).
(g) A fixed-free cantilever carrying a stiff object at the end.The stiff object does not undergo
flexural bending due to increased thickness.
(h) This beam is very similar to case (c) except for the fabrication method.
(i) A fixed-free cantilever with folded length. It consists of several fixed-free beam segments
connected in series.The free end of the folded cantilever is capable of movement in a
direction parallel to the substrate surface. Movement of the free end perpendicular to the
substrate would encounter much greater resistance.
(j) Two fixed-free cantilever connected in parallel.The combined spring is stiffer than any
single arm.
(k) Four fixed-guided beam connect to a rigid shuttle, which is allowed to move in the
substrate plane but with restricted out-of-plane translational movement.
TORSIONAL DEFLECTIONS
Whereas beams, especially cantilevers, are frequently used in MEMS for producing linear
displacements and small angular rotations, torsional beams are often used to create large
angular displacement.A case in mind is the digital micro mirror device.The mirror plate is
supported by torsional bars to facilitate rotation. We begin our discussion of torsion by
considering a prismatic bar of circular cross section twisted by torques T acting at the ends.
Since every cross section of the bar is identical, and since every cross section is subjected to
the same internal torque T, the bar is said to be in pure torsion. It may be proved that cross
sections of the bar do not change shape as they rotate about the longitudinal axis. In other
words, all cross sections remain plane and circular and all radii remain straight. Furthermore,
if the angle of rotation between one end of the bar and the other is small, neither the length of
the bar nor its radius will change. To aid in visualizing the deformation of the bar, imagine
that the left-hand end of the bar is fixed in position.Then, under the action of a torque T, the
right-hand end will rotate through a small angle Ǿ known as the angle of twist.
Because of this rotation, a straight longitudinal line on the surface of the bar will become a
helical curve. The angle of twist changes along the axis of the bar, and at intermediate cross
sections it will have a value Ǿ(x) that is between zero at the left-hand end Ǿ and at the
righthand end.The angle Ǿ(x) will vary linearly between the ends. Point a on the right-hand
side will move by a distance d to a new location a’. The torsion induces shear stress
throughout the bar. Stress distribution shows radial symmetry. The magnitude of the shear
stress is zero in the center of the cross section and reaches maximum at the outer surface of
the bar. The maximum stress in the bar is denoted The expression of the maximum shear
strain is

Further, the magnitude of the stress is linearly proportional to the radial distance to the
center.The distribution of shear stress along the radial direction is superimposed on the
crosssectional view of the bar in .22.

24
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The relationship between a torque and the maximum shear stress is found by torque balance
at any given section.

UNIT II
SENSORS AND ACTUATORS-I

Electrostatic sensors – Parallel plate capacitors – Applications – Interdigitated Finger

capacitor –Comb drive devices – Micro Grippers – Micro Motors - Thermal Sensing and
Actuation – Thermal expansion – Thermal couples – Thermal resistors – Thermal Bimorph -
Applications – Magnetic Actuators – Micromagnetic components – Case studies of MEMS in
magnetic actuators- Actuation using Shape Memory Alloys.

INTRODUCTION TO ELECTROSTATIC SENSORS AND ACTUATORS

A capacitor is broadly defined as two conductors that can hold opposite charges. It can be
used as either a sensor or an actuator. If the distance and relative position between two
conductors change as a result of applied stimulus, the capacitance value would be
changed.This forms the basis of capacitive (or electrostatic) sensing. On the other hand, if a
voltage (or electric field) were applied across two conductors, an electrostatic force would
develop between these two objects.This is defined as electrostatic actuation.
Electrostatic forces are not often used for driving macroscopic machinery. However,
micro devices have large surface-area-to-volume ratios and their masses are generally very
small, thus making electrostatic force, which is a surface force, an attractive candidate for
micro actuation.

25
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The electrostatically driven micro motor was one of the earliest MEMS actuators . The motor,
schematically diagrammed in Figure , consists of a rotor that is attached to the substrate with
a hub, and a set of fixed electrodes on the periphery, called stators.The stators are grouped
together such that each group of four electrodes is electrically biased simultaneously.Three
such groups are present for the motor shown in Figure and are identified by different fill
patterns. To explain the operation principle of the electrostatic motor, let’s begin with the
rotor at an arbitrary angular rest position (Figure a). One group of stator electrode is first
biased. (The electrodes that are biased at a particular stage are identified by arrows placed
next to them.) An in-plane electric field develops between any given stator electrode in this
group and the closest rotor tooth next to it. This generates an electrostatic attractive force that
aligns the said tooth with the said stator electrode.The torque values are on the order of pico
Nm for voltages of the order of 100 V, large enough to overcome static friction. A small
angular movement of the rotor is made to reach a new configuration diagramed in Figure
b.The electric bias is shifted to the next group of stator electrodes (Figure c), resulting in
another small angular movement in the same direction. Continuous motion of the rotor can be
achieved by activating the stator electrodes by groups in succession. Later works in the area
of electrostatic actuators have provided new designs for greater power and torque output. In
this chapter, I will discuss the principles and governing equations of capacitive sensors and
actuators. Major advantages of electrostatic sensing and actuation are summarized in the
following.
1.Simplicity. The sensing and actuation principles are relatively easy to implement, requiring
only two conducting surfaces. No special functional materials are required. Other sensing
methods, such as piezoresistive and piezoelectric sensing, and other actuation methods, such
as piezoelectric actuation, require deposition, patterning, and integration of special
piezoresistive and piezoelectric materials.
2.Low power. Electrostatic actuation relies on differential voltage rather than current.The
method is generally considered energy efficient for low-frequency applications. This is
especially true under static conditions, when no current is involved. At high frequencies, a
time- and frequency-dependant displacement current, i(t), will develop in response to a time-
varying bias voltage V(t). The magnitude of the current is C=dV(t)/dt. The instantaneous
power delivered to a capacitor is p(t) = i(t) x V(t).
3.Fast response. Electrostatic sensing and actuation offers high dynamic response speed, as
the transition speed is governed by the charging and discharging time constants that are
typically small for good conductors. For example, the switching time of the mirrors in the

26
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

digital micromirror display (DMD) array is smaller than fast enough to support 8-bit gray
scale display.
Relative disadvantages of electrostatic actuation and sensing should be recognized as well.
High voltage required for static actuator operation is considered a drawback. The DMD
mirrors are switched by voltage on the order of 25 V to achieve plus or minus 7.5 degree
tilting.A monolithically integrated optical mirror with 90 tilt range requires 150 V bias
voltage. Linear electrostatic actuators requiring hundreds of volts to move a microstructure
by several tens of micrometers are rather common. High voltage introduces electronics
complexity (for providing high voltage supply) and material compatibility issues. Electrodes
that are mechanically connected with insulators tend to accumulate charges, especially under
DC operation modes. The trapped charges will change the operation characteristics.

There are two major categories of capacitive electrode geometries: parallel-plate capacitor
and interdigitated finger (comb-drive) capacitors. The static capacitance values of parallel
plates and comb drives are generally very small, being on the order of pF or smaller. For
capacitive sensors, careful design of electrical circuits is needed to register capacitance
changes on the order of fF or lower, in the presence of noise and interference sources.
Capacitive actuators often take advantage of electrostatic attraction between two oppositely
charged surfaces. Repulsive electrostatic forces, which are used less often, can be developed
between two surfaces with same charge polarities. For example, vertical levitation of
interdigitated fingers up to 0.5 µm has been demonstrated under applied voltage of 20 V for
an 18-finger device.

PARALLEL-PLATE CAPACITOR
Capacitance of Parallel Plates
A parallel-plate capacitor is the most fundamental configurations of electrostatic sensors and
actuators. Such a capacitor, as its name suggests, consists of two conducting plates with their
broad sides parallel to each other. In a broader definition, these two plates are not required to
be exactly parallel at all time nor are they required to be planar.

A representative application using the parallel-plate electrostatic actuation principle is the

DMD chip made by Texas Instruments Inc.. A DMD chip consists of a large array of
micromirrors, or binary light switches. Each mirror is made of a reflective plate capable of
rotating about a torsional support hinge.Two electrodes are embedded below the reflective
plate, each covering one half of the plate area on each side of the torsional hinge.Two sets of
parallel capacitors are thus formed, between each of the underlying electrodes and the mirror
surface above. By applying voltages to either bottom electrode, the mirror device would tilt
under electrostatic attractive force by plus r minus 7.5 degrees, to deflect incoming light.

27
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Let us examine a simple parallel-plate capacitor depicted in Figure . These two plates have an
overlapping area of A and a spacing of d.The dielectric constant, or relative electrical
permitivity, of the media between the two plates is denoted . The permitivity of the media
is .
The value of the capacitance, C, between the two plates is defined as C=Q/V.
where Q is the amount of stored charge and V the electrostatic potential. The electric energy
stored by a given capacitor, U, is expressed as U =1/2 CV2. For a parallel-plate capacitor,
electric field lines are parallel to each other and perpendicular to the plate surfaces in the
overlapped region. Fringe electric fields reside outside the boundary of electrode plates.The
fringe field lines are three-dimensional in nature and should be considered in rigorous design
analysis.
According to the Gauss’s law, the magnitude of the primary electric field,E, is related to Q by
E=Q/ᶓA. The magnitude of the voltage is the electric field times the distance between two
plates (d).The capacitance of a parallel-plate capacitor is

The capacitance value is proportional to A and inversely proportional to d. It is a function of

the electric permitivity, Two parallel plates can move with respect to each other in two
ways—normal displacement or parallel sliding displacement. In this textbook, we are mainly
concerned with the cases where two parallel plates move along their normal axis. The
parallel-plate capacitor is a versatile platform for physical, chemical, and biological sensors.
By measuring the capacitance value of a parallel capacitor, one can sense the changes of
permitivity,A, or d.The permitivity can be changed by temperature and humidity [11] of the
capacitor media. The capacitance change via the permitivity route can be used to characterize
liquid, air, or even biological particles in the gap. For example, the DNA content of
eukaryotic cells can be observed through a linear relation between the DNA contents of cells
and the change in capacitance that is evokes by the passage of individual cells across a 1-kHz
electric field.The overlapped area of and distance between plates can be changed by contact
force, static pressure, dynamic pressure (acoustics), and acceleration. The capacitor can be
used as an actuator to generate force or displacement (if at least one of the capacitor plate is
suspended or deformable). Let us examine a pair of electrode plates, with one plate firmly
anchored and another one suspended by a mechanical spring. As a differential voltage is
applied between the two parallel plates, an electrostatic attraction force will develop. The
magnitude of the forces equals the gradient of the stored electric energy, U, with respect to
the dimensional variable of interest.The expression for the magnitude of the force is

where x is the dimensional variable of interest. If the plates move along their normal axis, the
gap between electrodes changes.The magnitude of the force can be rewritten as

by replacing the general dimensional coordinate x in

28
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Under a constant biasing voltage, the magnitude of the electrostatic force decreases
drastically with increasing gap, d. The electrostatic force is considered a short-range force,
most effective when the gap is on the order of a few micrometers. The upper limit of the
applied voltage for actuation is the breakdown voltage of dielectric media. If the media is air,
the breakdown voltage can be predicted following the Paschen curve, which predicts that the
breakdown voltage increases for decreasing gap sizes.This is a favourable scaling law for
MEMS capacitive devices.

Equilibrium Position of Electrostatic Actuator Under Bias

Many electrostatic sensors and actuators involve at least one deformable plate supported by
springs. An important design aspect for such sensors and actuators is to determine the amount
of static displacement under a certain biased voltage. In this section, I will discuss the
procedures for calculating the equilibrium displacement under static (DC) and quasi-static
(low frequency) biasing conditions. A parallel-plate capacitor with one movable plate
supported by a mechanical spring is diagramed in Figure below. The top plate is supported by
a spring with the force constant being km. At rest, the applied voltage, displacement, and the
mechanical restoring forces are zero. Gravity does not play an important role in static
analysis of micro devices, because the mass of plates is generally very small and the
gravitational force would not cause appreciable static displacement. When a voltage is
applied, an electrostatic force Felectricwill be developed.The magnitude of Felectric when the
movable plate is at its starting position is given by

This force will tend to decrease the gap, which gives rise to displacements and mechanical
restoring force. Under static equilibrium, the mechanical restoring force has equal magnitude
but opposite direction as the electrostatic force.

The case of an electrostatic actuator is very intriguing. The magnitude of electrostatic force
itself is a function of the displacement. To make things more complex, the electrostatic force
modifies the spring constant as well.

The spatial gradient of the electric force is defined as an electrical spring constant,

29
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

As seen, the magnitude of electrical spring constant changes with position (d) and the biasing
voltage (V). On the other hand, the mechanical spring constant Km stays unchanged for small
displacements.
The effective spring constant of a structure is the mechanical spring constant minus the
electrical spring constant. Let us derive the equilibrium displacement of a spring-supported
electrode plate under a bias voltage V. Suppose the resulted equilibrium displacement is x,
with the x-axis pointing in the direction of increasing gap. With the displacement x, the gap
between electrodes becomes d + x (d + x <d). The electrostatic force at the equilibrium
position is

The magnitude of the mechanical restoring force is

Equating the magnitudes of Fmechanical and Fmechanical at x and rearranging terms, we obtain

The equilibrium distance between two capacitor plates can be calculated by solving the
quadratic equation above with respect to x.

30
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The analysis of equilibrium position can be visualized graphically as well (Figure ).We first
setup a coordinate system where the horizontal axis represents the space between the two
plates and the vertical axis is the amplitude of the mechanical or electrostatic forces,
irrespective of their directions. A plate located at the origin of the x-axis is rigid, whereas the
plate originally placed at a distance xo away from the origin is movable. Two curves, one
representing the amplitude of the mechanical restoring force and one representing that of the
electrostatic force, are plotted as a function of electrode positions. The mechanical restoring
force (Fmecahnical) changes linearly with the position for a fixed bias voltage V.The
electrical force (Felectrical) increases with x in a nonlinear fashion, following the expression
of Equation 4.9. In the diagram above, there are multiple interception points for the two
curves, corresponding to the solutions of Equation 1. At each interception point, the
magnitude of the electrical force and the mechanical force are identical. The horizontal
coordinates of the interception points therefore indicate the equilibrium positions of the
movable plate. It is noteworthy that although several interception points are possible, only
one of them will be achieved in reality. Specifically, the solution that is closest to the rest
position is realized first and is generally the realistic solution. The graphical method can be
used to track the equilibrium position as the bias voltage is changed. The equilibrium
positions of the actuator under three representative bias voltages V1,V2,V3, and can be found
graphically (Figure ). As the voltage increases, the family of curves corresponding to electric
forces shifts upwards. The x coordinates of the interception points are moved further away
from the rest position.

31
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Pull-In Effect of Parallel-Plate Actuators

At a particular bias voltage, the two curves representing the mechanical restoring force and
the electrostatic force intercept at one point tangentially. At the interception point, the
electrostatic and mechanical restoring forces balance each other. Moreover, the magnitude of
the electric force constant (given by the gradient at the intercept point) equals the mechanical
force constant. The effective force constant of the spring is zero (i.e., extremely soft). This is
a special condition and should be dealt with carefully in practice. The bias voltage that
invokes such a condition is called the pull-in voltage, or Vp. If the bias voltage is further
increased beyond Vp, the two curves will not have a common interception, thus the
equilibrium solution disappears. In reality, the electrostatic force will continue to grow while
the mechanical restoring force, increasing only linearly, is unable to catch up with it again.
The two plates will be pulled against each other rapidly until they made a contact, at which
point the mechanical contact force will finally balance the electric one. This condition is
called pull in, or snap in.

The voltage and displacement necessary to induce the pull-in condition is therefore quite
important for designing electrostatic actuators. Although the pull-in effect can be explained
quite easily using graphics methods, an analytical model is needed to yield the exact values of

32
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

voltage and positions.We will review the pull-in condition for a simple spring-loaded
electrode system depicted in Figure .

At the pull in voltage, there is one tangential intersection point between Felectrical and
Fmechanical curves.The magnitudes of the electrical force and the mechanical balance forces
are the same. By equating these two forces and rearrange terms, we obtain

The value of x is negative when the spacing between two electrodes decreases. In addition,
the gradients of these two curves at the intersection point are identical, namely,
[Ke] = [Km]
The expression for the electric force constant,

can be rewritten by plugging in the expression for V2. This procedure results in a new
expression for the electric force constant,

33
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The only solution for x under which Equation 3 can be satisfied is

The above equation states that the relative displacement of the parallel plate from its rest
position is exactly one third of the original spacing at the critical pull-in voltage.This critical
displacement is true irrespective of the actual mechanical force constant and the actual pullin
voltage.
The voltage at pull in can be found easily by plugging in the transition distance to
Equation.This yields,

Consequently, the pull-in voltage is found to be

In reality, the pull-in voltage and threshold distance will deviate from the calculation obtained
using the idealized case.There are two sources of deviation from the ideal model. First, the
fringe capacitance will alter the expression of the electrostatic force. Second, the restoring
force provided by the mechanical springs will differ from that predicted by the linear model if
the displacement is large. Understanding about the pull-in effect and mathematical tools for
estimating the pullin effects have been advanced steadily. Comprehensive analysis of the
pull-in phenomenon can be carried out for systems with multiple degrees of freedom and
hysteresis. More complex cases, such as pull-in effect of rotational electrostatic actuators
based on torsional support and that of electrodes with complicated profiles (e.g., cantilevers
and membranes) have been studied in the past. Many electrostatic actuators and sensors are
operated at high frequencies. The pull-in effects of such dynamic systems are important for
the system performance and have been studies numerically and experimentally.

INTERDIGITATED FINGER CAPACITORS

While parallel-plate capacitors generate sensing and actuation across planar electrodes facing
each other, a different class of capacitors take advantage of capacitance generated from
sidewalls of electrodes. Such capacitors provide alternative fabrication and operation modes
compared with parallel-plate capacitors.They involve interdigitated fingers (IDT) to
increase the edge coupling length.Two sets of electrodes are placed in the same plane parallel
to the substrate. Generally, one set of finger-like electrode is fixed on chip while a second set
is suspended and free to move in one or more axes. Since the interdigitated fingers are shaped
like tooth of combs, such configuration is commonly referred to as the comb-drive device.

34
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

In a generic configuration (), two sets of fingers are in the same plane and their comb fingers
are engaged with an overlapping distance of l0. The length of the fingers is denoted Lc. For
simplicity, I assume the fingers in both sets have identical thickness (t) and width (wc). The
distance between a fixed comb finger and an immediately neighbouring movable finger is
d.The thickness of fingers corresponds to that of the conductive thin film. The capacitance
between a pair of electrode fingers is contributed by vertical surfaces of the fingers in the
overlapped region, as well as by fringe capacitance fields. Capacitances derived from
multiple finger pairs are connected in parallel. Hence the total capacitance is a summation of
capacitance contributed by neighbouring fingers. Opposite walls of comb fingers in the
overlapped region form a parallel-plate capacitor and contribute a capacitance C. The
magnitude of C between two immediate neighbouring fingers is

When designing a comb-drive capacitor, the thickness t and the distance d are highly process
relevant. The greater the thickness, the larger the capacitance effects. However, there are
different preferred methods for realizing desired thickness. The distance between gaps should
be as small as possible. However, it is determined by the ultimate lithography resolution.
The processing and design terms interact with each other. Since the thickness of the beam is
likely the same as the thickness of the mass, a thicker beam translate into a greater mass.
Often thick electrodes are made by using dry silicon etching, which is inevitably associated
with undercuts. Hence there is inherent conflict when trying to reach deep trench (large t) and
small gap (low d) at the same time. The fringe capacitance Cf, is difficult to estimate
analytically.The most accurate way to estimate the fringe capacitance is by using finite
element method (FEM). A simplistic way to estimate the fringe capacitance is to assume it is
a fixed fraction of the capacitance developed in the overlapped regions; however, this is not
entirely accurate and only suitable for first-order analysis.When dealing with homework
problems in this text, the fringe capacitance should be ignored. Two types of comb-drive

35
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

devices are commonly encountered, depending on the relative movement of two sets of comb
fingers allowed by their mechanical suspensions. The first is a transverse comb-drive device,
as shown in Figure below.The set of free fingers moves in a direction perpendicular to the
longitudinal axis of comb fingers. Accelerometers for automotive airbag deployment made by
Analog Devices Corp. Let us focus on a single fixed finger and its two neighbouring moving
fingers. According to the analysis earlier, there are two major capacitances associated with
each finger pair, one to the left-hand side of the finger, called Csl and one to the right-hand
side, called Csr. At rest, the values of these two capacitances are

When the free finger moves by a distance of x, the capacitance values of these two capacitors
become

The total value of capacitance is Ctot = Csl + Csr + Cf .

When the transverse comb is used as a sensor, the displacement sensitivity Sx can be
obtained

by taking the derivative of Ctot with respect to x, namely

If the transverse comb drive is used as an actuator, the magnitude of the force can be
calculated by taking a derivative of the total stored energy with respect to x,

A second type of comb-drive device is called the longitudinal comb drive. The direction of
relative movement is along the longitudinal axis of the fingers, allowable by the
suspension.With a lateral movement (y), the capacitances associated with a single finger
changes to

36
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The displacement sensitivity can be obtained by taking the derivative of C tot with respect to y,

There are many variations to the basic comb-drive design.The sensing and actuation
characteristics are strongly influenced by mechanical supports connected to the set of fingers.
For example, the design of flexural suspensions can be optimized to generate large
deformation while minimizing stress concentration. Coplanar transverse and longitudinal
comb drives are prevalent in MEMS. However, there are many different configurations and
geometries of comb-drive capacitors that deviate from these two major configurations. Some
comb drives are designed to produce vertical displacement, even when they initially reside in
the same plane. This is achieved by taking advantage of the fringe capacitance fields.This
levitation force is rather small but can be useful for applications such as optical phase tuning.
Reflective surfaces only need to be moved by up to one wavelength. Comb fingers that do not
reside in a common plane initially can be used to create out of plane forces or moments. The
out-of-plane arrangement of comb fingers can be achieved through novel fabrication
sequences or by using intrinsic-stress induced bending to lift one set of in-plane fingers
angularly. For example, an optical modulator capable of displacement has been made with 15
V driving voltage. Nearly all existing comb fingers have large length-to-width aspect ratio
and appear rectangular viewed from the top or side. Alternative profiles are possible. Comb-
drive fingers with curved profiles can provide important benefits such as maximum force,
increased linearity, or tailored force-displacement profiles.

Thermal Actuators
Actuation of microscale devices and structures can be achieved by injecting or removing
heat. Changes in temperature profile in turn result in mechanical displacement or force
output, through thermal expansion, contraction, or phase change.The temperature of a
microstructure can be raised by absorption of electromagnetic waves (including light), ohmic
heating (joule heating), conduction, and convection heating. Cooling can be achieved via
conduction dissipation, convection dissipation, radiation dissipation, and active
thermoelectric cooling. Thermal actuation is used in commercial MEMS products. Many ink
jet printers today eject ink droplets using thermal expansion of liquid inks.A schematic
diagram of a single thermal ink jet printer nozzle is shown in Figure below. It consists of a
micromachined fluid chamber with an opening.A micro heater is embedded on the substrate
within the cavity.A pulse of electrical current raises the temperature of the heater and
produces a vapor bubble that squeezes the liquid ink and ejects a droplet. Upon cooling, the
vapour bubble collapses and the cavity is refilled with ink again for next firing. Due to the
use of microfabrication method, the volume of ink and the thermal mass associated with the
ink are small. The heating and cooling of the fluid can be performed at high speed. For many
commercial ink cartridges, it takes 1µs or less for the vapor bubble to initiate, and 15 µs for
the ink to be ejected.The refill of the cavity takes approximately 24µs. The successful design
of a reliable ink jet nozzle involves electrical, mechanical, and thermal aspects. For example,
the peak pressure within the cavity can reach 14 ATM, with the peak temperature at the
surface around 3300. Both the positive and negative (cavitation) pressures are so intense that
they can cause cracking of material layers over time. This warrants special design
considerations. In order to eliminate cracking stemming form repeated thermal expansion and
contraction, the heater is engineered from a proprietary metal oxide material that has a
thermal expansion coefficient of nearly zero.

37
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Fundamentals of Thermal Transfer

Temperature is manifested at the microscale by the vigorousness of atomic vibration. Heat
transfer occurs whenever a temperature gradient is present in a material. Successful design of
thermal actuators and sensors require familiarity with heat transfer processes.

There are four possible mechanisms for heat to move from one point to another:
1. Thermal conduction is the transfer of heat through a solid media in the presence of a
temperature gradient.
2. Natural (passive) thermal convection is the transfer of heat from a surface into a
stationary body of fluid. A temperature gradient in a fluid induces local flow movement
through buoyancy.The movement of fluid mass facilitates heat transfer.
3. Forced thermal convection is the transfer of heat to a body of moving fluid.The bulk fluid
movement provides enhanced heat transfer compared with that of natural convection.
4. Radiation represents the loss or gain of heat through electromagnetic radiation
propagating in vacuum or air.
The governing equations relating the heat transfer rate and the temperature gradient for the
four heat transfer mechanisms are

Let’s examine heat flow pathways associated with the case of boiling a pot of water on an
electric stove. Heat is generated at the heating coil by passing current and is eventually lost to
the ambient background, which is assumed to be at a constant room temperature. Major heat
transfer pathways and the direction of heat flux are identified by arrows. The heat produced
by the heating coil first travels through the walls of the pot to reach the body of water inside.
Heat transfer occurs within the thickness of the wall via thermal conduction. Once the heat
reaches the interior wall of the pot, heat transfer to the liquid within begins. The liquid mass
closest to the wall warms up and begins to rise, setting up a naturalconvection, which brings
heat away from the pot wall to the interior of the body of water. If the liquid in the pot were
stirred, the heat transfer from the inside wall of the pot to the liquid could become stronger as
naturally convection is replaced by forced convection.
The body of water is exposed to air at the top. If the air outside is still, heat will transfer from
the water to the air through natural convection. On the other hand, if the air were moving
(e.g., stirred by a fan), heat would travel from the hot liquid to the air by forced convection.
Meanwhile, a person standing nearby feels heat wave coming from the heating coil. Heat is
said to move through air by radiation. Assuming the heating coil is at a significantly high
temperature, the radiation heat transfer can be quite strong. Certainly, the radiative
dissipation reduces the energy efficiency of heating the water. Even for such a simple
example, the heat transfer pathways are quite complex.The thermal transfer pathway

38
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

illustrated in Figure is simplified and only serves the purpose of illustrating the difference
between various heat transfer principles. In fact, there are many secondary heat transfer
pathways pertaining to this case. For example, heat is lost to the air from the outer pot wall
through radiation and convection.Temperatures of objects are considered spatially discrete
here; in fact, the temperature profile varies continuously. A heat flow will result between two
points of different temperatures. The ability of a media or an object to transfer heat between
two points is quantified by its thermal resistance.

The greater the thermal resistance between two points, the better the thermal isolation and the
smaller the heat transfer rate under a given temperature difference (thermal driving force).
The concept of thermal resistance can be better understood and appreciated by examining the
analogy between heat transfer and electrical current flow. Temperature difference is the
driving force for heat flow and transfer.The difference of temperature between two points sets
up a heat flow, much like an electrical voltage gives rise to organized movement of charges
(electric current). The electrical resistance is defined as the ratio of voltage and current.
Likewise, the thermal resistance can be defined as the ratio of temperature difference and the
heat flow.
Thermal Bimorph Principle
The thermal bimetallic effect is a very commonly used method for sensing and actuation.
This mechanism allows temperature variation in microstructures to be exhibited as transverse
displacement of mechanical beams. The thermal bimorph consists of two materials joined
along their longitudinal axis serving as a single mechanical element [8]. (Oftentimes, a
thermal bimetallic actuator may consist of more than two layers of materials. This book
focuses on the analysis of two-layered architectures only.) Figure 5.3 shows a composite
beam with two layers, made of material 1 and 2, having same length (L) but different
coefficients of thermal expansion (CTE) The subscript refers to the material layer. Likewise,
the Young’s modulus, width, and thickness of the two layers are denoted

With a uniform temperature rise of the length of two sections changes unequally. Because the
two-layered materials are tightly joined at the interface, the beam must curve toward the layer
made of the material with a lower CTE value.A transverse beam bending is therefore

39
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

produced. Next we analyze the formula for calculating the displacement of a bimetallic beam.
Under a uniform temperature change of ÀT, the beam curves and assumes the shape of a
section of an arc with the length of the arc being L. The radius of curvature of the arc, r, can
be calculated using this formula:

The arc is a section of a circle with the radius of curvature being denoted r, spanning an arc
Angle ʘ. The value of ʘ is determined by

ʘ =l/r

Once the radius of curvature is found, the vertical displacement of the free end of the beam
can be determined by trigonometry according to Figure below.

The vertical displacement at the free-end of the cantilever is therefore

40
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

If the overall bending angle is small, the magnitude of the vertical displacement can be
estimated by replacing Cosʘ with first two terms in its Taylor series expansion:

THERMAL COUPLES
In 1823, a German physicist Seebeck discovered that a voltage was developed in a conductor
loop containing two dissimilar materials, when the temperature at the joint between two
materials was increased above the room temperature. The voltage is capable of maintaining a
current if the loop is closed. (In fact, the voltage was noticed only when the current was able
to deflect a magnetic compass needle nearby.) Two wires made of dissimilar materials joined
at one point constitute a thermal couple. A thermal couple is most commonly used to
measure the temperature difference between the joined sensing junction and a reference one.
It is also used for generating electricity from a temperature gradient between their two
junctions. Seebeck tested a number of materials, including naturally found semiconductors
ZnSb and PbS.When a temperature difference, is applied, it will be accompanied by an
electric voltage, He found that the open circuit voltage is linearly proportional to the
difference of temperature.The ratio between the developed open circuit voltage and the
temperature difference is the Seebeck coefficient:

The above term is the Seebeck coefficient of a thermal couple, unique to the combination of
two materials. The Seebeck coefficient is alternatively called the thermoelectric power, or
just thermopower.

Why must thermal couples involve two different materials?

Although a single piece of metal is theoretically capable of exhibiting a voltage difference

when a temperature difference is present between its two ends (junctions), this voltage can
not be easily measured or used, because another piece of conducting material must be present
to interrogate or communicate this voltage. The conducting wire inevitably exhibits Seebeck
effect as well. If the conducting material is the same as the conducting material under test, no
voltage difference would be detected at the referecne junction because the Seebeck voltage of
the two pieces cancel.
As a result, two pieces of dissimilar metals must be used to provide an open circuit
voltage at the reference junttion. Seebeck coefficients are in fact associated with individual
metal elements. If the Seebeck coefficient of the two constitutional materials (α a and αb) in
the thermal couple in Figure 0 are denoted and respectively, the Seebeck coefficient of the
thermal couple is defined as following:

41
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The resistance value of a resistor is a function of the resistivity and its dimensions, including
length l and the cross-sectional area A. It is given as

Both the resistivity and the dimensions are functions of temperature. As a result, the
resistance
value is sensitive to temperature. A thermal resistor is an electrical resistor with appreciable
temperature sensitivity.

Both the resistivity and the dimensions are functions of temperature. As a result, the
resistance value is sensitive to temperature. A thermal resistor is an electrical resistor with
appreciable temperature sensitivity. The resistance of a thermal resistor, R, is related to the
ambient temperature in a relationship shown below:

Where RT and R0 are the resistance at temperature T and T0 respectively.The term αb is called
the temperature coefficient of resistance (TCR). This equation is valid for moderate
temperature excursion. If the temperature difference is greater, nonlinear terms may be
needed to yield accurate expression of RT.
Thermal resistors can be made of metal or semiconductors. In both cases, the dimensions of a
resistor change with temperature.The electrical resistivity of both metals and semiconductors
varies with temperature, but the principles of such change is quite different for metals and
semiconductors. For metal resistors, temperature rises introduce enhanced lattice vibration,

42
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

which tends to impede the movement of charge carriers. In the case of semiconductors,
temperature affects the lattice spacing and, in turn, the effective mass and mobility of charge
carriers. Forsemiconductor thermal resistors, the resistivity-temperature relation is influenced
by doping concentration and dopant types.
The temperature coefficient of resistance can be measured simply by heating up a resistor
using a temperature-controlled stage and monitoring the resistance value. The temperature of
the stage should be increased slowly in small increments, to allow sufficient time interval in
between each temperature rise step to ensure thermal equilibrium. The bias voltage and
current must be kept low in this experiment, to minimize contribution of the electric heating
power, which is the product of voltage and current. The current and voltage used to
interrogate the resistance value of a thermoresistor may introduce heat to it. This
phenomenon is called self-heating. The heating power of a resistor under current I is

P=I2R.

Self-heating of resistors by the interrogation currents may change its temperature and the
resistance value.

The temperature and hence resistance value of a resistor under self-heating condition is
dependant on the rate of heat input and dissipation. The ohmic heating energy may be
dissipated through conduction and convection. If the convective heat loss is dominant, a self-
heated resistor can be used to measure the convective heat loss rate and flow rates.This forms
the basis of hot-wire anemometry. Under zero flow movement, the temperature of the heated
resistive wire reaches a steady-state value.When the resistive wire is subjected to a moving
fluid, heat is convectively moved from the element, reducing its temperature. The changes of
the temperature of the sensing element can be inferred by measuring the real-time resistance
value. Similarly, the self-heating effect can be used for measuring flow shear stress in the
boundary layer. The value of thermal resistance associated with a thermal resistor can be
easily obtained from the current-voltage characteristics (I–V). According to the Ohm’s law,
the current in a resistor will increase linearly with respect to the input voltage. A linear
current-voltage plot is often encountered, with the slope of the I–V characteristic curve
representing the inverse of the resistance of the resistor.

MEMS MAGNETIC ACTUATOR

A MEMS magnetic actuator is a device that uses the microelectromechanical
systems (MEMS) to convert an electric current into a mechanical output by employing the
well-known Lorentz Force Equation or the theory of Magnetism. Micro-Electro-Mechanical
System (MEMS) technology is a process technology in which mechanical and electro-
mechanical devices or structures are constructed using special micro-fabrication techniques.

43
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

These techniques include: bulk micro-machining, surface micro-machining, LIGA, wafer

bonding, etc.
A device is considered to be a MEMS device if it satisfies the following:

 If its feature size is between 0.1 µm and hundreds of micrometers. (below this range, it
becomes a nano device and above the range, it is considered a mesosystem)
 If it has some electrical functionality in its operation. This could include the generation of
voltage by electromagnetic induction, by changing the gap between 2 electrodes or by a
piezoelectric material.
 If the device has some mechanical functionality such as the deformation of a beam or
diaphragm due to stress or strain.
 If it has a system-like functionality. The device must be integrable to other circuitries to
form a system. This would be the interfacing circuitry and packaging for the device to
become useful.
For the analysis of every MEMS device, the Lumped assumption is made: that if the size of
the device is far less than the characteristic length scale of the phenomenon (wave or
diffusion), then there would be no spatial variations across the entire device. Modelling
becomes easy under this assumption.

Magnetic actuation
The principle of magnetic actuation is based on the Lorentz Force Equation.When a current-
carrying conductor is placed in a static magnetic field, the field produced around the
conductor interacts with the static field to produce a force. This force can be used to cause the
displacement of a mechanical structure.

Magnetic actuation

Governing equations and parameters

A typical MEMS actuator is shown on the right. For a single turn of circular coil, the
equations that govern its operation are:

 The H-field from a circular conductor:

44
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

 The force produced by the interaction of the flux densities:

The deflection of a mechanical structure for actuation depends on certain parameters of the
device. For actuation, there has to be an applied force and a restoring force. The applied force
is the force represented by the equation above, while the restoring force is fixed by the spring
constant of the moving structure.
The applied force depends on both the field from the coils and the magnet. The remanence
value of the magnet its volume and position from the coils all contribute to its effect on the
applied Force. Whereas the number of turns of coil, its size (radius) and the amount of current
passing through it determines its effect on the Applied Force. The spring constant depends on
the Young's Modulus of the moving structure, and its length, width and thickness.

Magnetostrictive actuators
Magnetic actuation is not limited to the use of Lorentz force to cause a mechanical
displacement. Magnetostrictive actuators can also use the theory of magnetism to bring about
displacement. Materials that change their shapes when exposed to magnetic fields can now be
used to drive high-reliability linear motors and actuators. An example is a nickel rod that
tends to deform when it is placed in an external magnetic field. Another example is wrapping
a series of electromagnetic induction coils around a metal tube in which a Terfenol-D
material is placed. The coils generate a moving magnetic field that courses wavelike down
the successive windings along the stator tube. As the traveling magnetic field causes each
succeeding cross section of Terfenol-D to elongate, then contract when the field is removed,
the rod will actually "crawl" down the stator tube like an inchworm. Repeated propagating
waves of magnetic flux will translate the rod down the tube's length, producing a useful
stroke and force output. The amount of motion generated by the material is proportional to
the magnetic field provided by the coil system, which is a function of the electric current.
This type of motive device, which features a single moving part, is called an elastic-wave or
peristaltic linear motor. (view: Video of a Magnetostrictive micro walker)

Advantages of magnetic actuators

 High actuation force and stroke (displacement)

 Direct, fully linear transduction (in the case of electrodynamic actuation)
 Bi-directional actuation
 Contactless remote actuation
 Low-voltage actuation
 A figure of merit for actuators is the density of field energy that can be stored in the gap
between the rotor and stator. Magnetic actuation has a potentially high energy density[

Magnet material selection for static B-Field

The operation of the magnetic actuator depends on the interaction between the field from an
electromagnet and a static field. To produce this static field, it is important to use the right

45
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

material. In MEMS, permanent magnets have become the favorite because they have a very
good scaling factor and they retain their magnetization even when there is no external field...
meaning that they need not be continuously magnetized when they are in use.

Integrating the magnet into the MEMS device

As earlier discussed, MEMS devices are designed and fabricated using special micro-
fabrication techniques. The major challenge however for magnetic MEMS is the integration
of the magnet into the MEMS device. Recent research has suggested solutions to this
challenge.
Fabrication (or molding) of the magnet
There are several ways by which the magnet could be fabricated on a MEMS structure:

Sputtering

 Sputtering: Argon ion bombardment of the material release particles of the material.
Mainly for depositing rare earth magnets. Deposition rate and film surface area depend
on sputtering tool and target size.
 Pulsed Layer Deposition: a high-power pulsed laser beam is focused inside a vacuum
chamber to strike a target of the material that is to be deposited.
 Electroplating
 Screen printing
 Wax/parylene bonding

Issues with magnetic actuation

 High-power dissipation. This is a major problem for magnetic MEMS, but work is
underway to circumvent this.
 Fabrication of the coil
 Integration of the micromagnet into the MEMS device
 Process-material compatibility
 Integratability into the overall microfabrication process (maintain cost and throughput)
 So that preexisting processes in the fabrication of the MEMS device will not be tampered
with, deposition temperatures and post-deposition treatment/conditions must be tolerable.
Also, the micromagnet must be able to withstand any chemical treatment that will come
after its deposition
 Issues with magnetization (One may want to have more than one direction of
magnetization; this creates a problem)

46
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Each of these challenges can be mitigated or lessened by the right choice of material, choice
of molding or fabrication method, and the type of device that is to be constructed.
Applications of the magnetic actuator include: the synthetic jet actuator, micro-pumps and
micro-relays.
Application – Hot wire Anemometer
Hot-wire anemometry (HWA) is a well-studied technique for measuring the velocity of fluid
flow. It utilizes a thermal element that serves as both a resistive heater and a temperature
sensor. The temperature and resistance of the thermal resistor, biased in the selfheating
regime, changes with the speed of flow movement. The HWA is noted for its low cost, fast
response (in the kilohertz range), small sizes, and low noise. Conventional HWA sensors are
assembled individually by mounting a thin wire made of platinum or tungsten onto support
prongs.The wires may be thinned (e.g., by etching in acidic solutions) until the desired
dimensions are reached (typically a few mm long and a few micrometers in diameter).This
active portion of the sensor is then mounted on a long probe with electrical connection for
ease of handling.The schematic diagram of a typical finished device is shown in Figure 4.
Conventional HWAs suffer from two major shortcomings, however. First, the fabrication and
assembly process is delicate and does not guarantee uniformity of performance. Second, it is
prohibitively difficult to form large arrays of HWA for measuring flow field distribution.The
research was motivated by the needs to reduce the cost of HWA, to fabricate and manufacture
hot wire anemometers efficiently, and to produce arrayed HWAs on potentially flexible
substrates.

Two strategic approaches are taken to achieve these goals. First, HWA sensors are made by
using surface micromachining in conjunction with three-dimensional assembly methods. This
circumvents the need to use bulk micromachining method, which requires relatively long
etching time and complex processing. Bulk etching using anisotropic wet etchants frequently
pose concerns of materials compatibility as all materials on a given substrate are required to
sustain wet etching for long periods (several hours to etch through typical silicon
wafers).This method reported enables more efficient assembly and allows formation of large
arrays of HWAs. Second, hot wires were made with thin film metal instead of polycrystalline
silicon. This could reduce costs as the use of silicon bulk (as substrate) or thin film (as hot
wire) is not required. By eliminating silicon doping and bulk etching steps, the fabrication
process can be realized in more efficient manner. The schematic diagram of the new out-of-
plane anemometer is shown in Figure 5. A thermal element is elevated from the substrate to a
predetermined height that corresponds to the length of the support prongs. By elevating the
thermal element away from the bottom of the velocity boundary layer, the thermal element

47
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

experiences greater fluid flow velocity and exhibits greater sensitivity.The thermal element is
electrically connected to the substrate through the support prongs as well.
The hot wire is made of temperature-sensitive metal thin films. The polyimide support is used
because it provides the hot wire with needed structural rigidity without increasing cross-
sectional area and thermal conductivity. The thermal conductivity of polyimide is low, almost
two orders of magnitude lower than that of a metal, e.g., nickel.

In this work, the thickness of the polyimide thin film is roughly 2.7µm. If the thickness is
much lower than this value, the mechanical rigidity will likely be degraded. On the other
hand, if the thickness is much greater, there is concern that the polyimide support will
decrease the frequency response of the HWA due to added thermal mass. The fabrication
process utilizes an efficient 3-D assembly method called the Plastic Deformation Magnetic
Assembly (PDMA).

A brief discussion of this method is provided in the following. The PDMA process utilizes
surface micromachined structures that are anchored to substrates with cantilever beams made

48
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

of ductile metal materials (e.g., gold and aluminum). The microstructure is attached to pieces
of electroplated ferromagnetic material (e.g., Permalloy). By applying an external magnetic
field, the ferromagnetic material is magnetized and interacts with the field to bend the
microstructure out of plane. If the amount of bending is significant, the cantilever support
hinges will be plastically deformed, resulting in permanently bent microstructures even after
the magnetic field is removed. The process is very efficient and can be realized in parallel on
the wafer scale. The overall fabrication process is shown in .The starting wafer is silicon.
However, the process can be performed on glass or polymer substrates as the overall
temperature of the process is intentionally kept low. First, a chrome/copper/titanium metal
stack is evaporated and patterned as the sacrificial layer (a). A 10-nm-thick chrome film
serves as the adhesion layer.A 250-Å-thick titanium thin film reduces the in-process
oxidation of the 2500-Å-thick copper film. A 27µm -thick photo-definable polyimide (HD-
4000) is spun-on, patterned via lithography, and cured at 3500C. (b) for 2 hrs. This polyimide
layer forms the support prong and part of the hot wire.A Cr/Pt/Ni/Pt film is then evaporated
and patterned to form the thermal element (c).The thickness of the Cr thin film, an adhesion
layer, is 200 Å. A 800-Å-thick Ni resistor is sandwiched between two 200-Å-thick Pt films,
which are used to reduce possible oxidation of Ni while in operation because Pt is relatively
inert at high temperatures.The authors then evaporate and pattern a 5000-Å-thick Cr/Au film
(6d) to serve as a mechanical bending element as well as electrical leads of the hot-wire
filament.They electroplate a -thick Permalloy thin film on portions of the cantilever support
prongs (e). Sacrificial layer release is performed by using a solution containing acetic acid
and hydrogen peroxide to selectively remove the copper thin film. PDMA assembly is carried
out to lift the entire sensor out of plane (i) by placing a permanent magnet (field strength 800
Gauss) at the bottom of the substrate.To finish the process, the device chip is then rinsed in
de-ionized water and dried. The adhesion between the Au layer and the polyimide one is very
important for device integrity.Without adequate adhesion, the polyimide and the gold film
would separate during the PDMA assembly. One of the ways to improve adhesion was to use
Cr as an adhesion layer and treat the polyimide layer by using O 2 reactive ion etching (RIE)
before the metal deposition. Cr seems to be the adhesion layer of choice, and the RIE
treatment creates a hydrophilic structure on the polyimide surface that enhances adhesion. Ti
was initially used as the adhesion material for both the Au and the hot-wire element due to its
good stability in chemical etchants and higher electrical resistivity. However, it failed to
provide sufficient adhesion. For certain applications, it may be advantageous to be able to
strengthen the bent hinge such that the HWA can operate at high flow speed. The mechanical
rigidity can be reinforced by electroplating metals (g). All values are cited from except for the
thermal conductivity of polyimide, which is cited from product guideline of HD
Microsystems. The steady state response of the sensor to air velocity has been experimentally
obtained up to 20 m/s under both constant current and constant temperature modes.
Frequency response up to 10 kHz has been demonstrated with the small thermal mass.

Ink Jet Droplet Injector

• Bubble formation time: 1 ms.
• Ink ejection time: 15 ms.
• Peak pressure: 14 ATM
• Upon removal of heat, vapor cools and the bubble retreats.
• Refill at 24 ms, lasts about 25 ms.
• Surface temperature: 90% of critical temperature (vaporization temperature) which is
330 oC.

49
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

• Homogeneous boiling across the surface of the heater, made of tantalum-aluminum

(Ta-Al) alloy.
• The heater has near zero TCR, so zero thermal expansion.
• HP Ink Jet Printer - Single Drop

• Thermal actuation
– Relatively high power: due to current operation.
– Lower response speed due to thermal time constant (dissipation and thermal
charging)
– Construction and fabrication: more complex due to material compatibility
considerations.
– Range of motion: relatively large.

Ink Jet Droplet Injector

50
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

UNIT III
SENSORS AND ACTUATORS-II
Piezoresistive sensors – Piezoresistive sensor materials - Stress analysis of mechanical
elements –Applications to Inertia, Pressure, Tactile and Flow sensors – Piezoelectric sensors
and actuators –piezoelectric effects – piezoelectric materials – Applications to Inertia,
Acoustic, Tactile and Flow sensors.

ORIGIN AND EXPRESSION OF PIEZORESISTIVITY

First discovered by Lord Kelvin in 1856, the piezoresistive effect is a widely used sensor
principle. Simply put, an electrical resistor may change its resistance when it experiences a
strain and deformation.This effect provides an easy and direct energy/signal transduction
mechanism between the mechanical and the electrical domains.Today, it is used in the
MEMS field for a wide variety of sensing applications, including accelerometers, pressure
sensors, gyro rotation rate sensors, tactile sensors, flow sensors, sensors for monitoring
structural integrity of mechanical elements, and chemical/biological sensors. The resistance
value of a resistor with the length being l and the cross-sectional area A is given by

The resistance value is determined by both the bulk resistivity and dimensions. Consequently,
there are two important ways by which the resistance value can change with applied strain.
First, the dimensions, including length and cross section, will change with strain. This is easy
to understand though the relative change in dimensions is generally small. Note that
transverse strains may be developed in response to longitudinal loading. For example, if the
length of a resistor is increased, the cross section will likely decrease under finite Poisson’s
ratios (Figure ). Secondly, the resistivity of certain materials may change as a function of
strain.The
magnitude of resistance change stemming from this principle is much greater than what is
achievable from the first one. By strict definition, piezoresistors refer to resistors whose
resistivity changes with applied strain. Metal resistors change their resistance in response to
strain mainly due to the shape deformation mechanism. Such resistors are technically called
strain gauges. The resistivity of semiconductor silicon changes as a function of strain. Silicon
is therefore a true piezoresistor. In this chapter, both semiconductor piezoresistors and metal
strain gauges are discussed. The fact that the resistivity of semiconductor silicon may change
under applied strain is fascinating. The reason for the strain dependence of resistivity is
explained as follows. The resistivity of a semiconductor material depends on the mobility of
charge carriers.The formula for the mobility is

where q is the charge per unit charge carrier, the mean free time between carrier collision
events, and m* the effective mass of a carrier in the crystal lattice. Both the mean free time
and the effective mass are related to the average atomic spacing in a semiconductor lattice,
which is subject to changes under applied physical strain and deformation.
Let us now focus on the macroscopic description of the behavior of a piezoresistor under a
normal strain. The change in resistance is linearly related to the applied strain, according to

51
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The proportional constant G in the above equation is called the gauge factor of a
piezoresistor.We can rearrange the terms in this equation to arrive at an explicit expression
for G,

The resistance of a resistor is customarily measured along its longitudinal axis. Externally
applied strain, however, may contain three primary vector components—one along the
longitudinal axis of a resistor and two arranged 90_ to the longitudinal axis and each other.

A piezoresistive element behaves differently towards longitudinal and transverse strain

Components.
The change of measured resistance under the longitudinal stress component is called
longitudinal piezoresistivity. The relative change of measured resistance to the longitudinal
strain is called the longitudinal gauge factor. On the other hand, the change of resistance
under transverse strain components is called transverse piezoresistivity. The relative change
of measured resistance to the transverse strain is called the transverse gauge factor.
Research has been conducted on single crystal silicon doped piezoresistors and
polycrystalline silicon doped piezoresistors, about the effects of doping types (n or p type),
doping concentration, and ambient temperature. It is important to know that the gauge factor
is a function of doping concentrations. Further, the resistance changes with respect to the
temperature, and even the gauge factor changes as a function of temperature. For any given
piezoresistive material, the longitudinal and transverse gauge factors could be different. It is
important to realize that longitudinal and transverse strains are often present at the same time
though one of them may play a clearly dominating role.The total resistance change is the
summation of changes under longitudinal and transverse stress components, namely

Three cases of piezoresistive force sensors are schematically illustrated below to exemplify
longitudinaland transverse piezoresistor configurations. Strain gauges, represented by the
resistor symbol, are bonded to the outer surfaces of rods which are subjected to external
loading forces. Different resistor orientations and external force loading directions are
presented (Figure ). In the case illustrated in Figure a, the longitudinal piezoresistance
dominates. In the cases illustrated in Figure b and c, transverse piezoresistances dominate.
Resistance changes are often read using Wheatstone bridge circuit configuration. A
basicWheatstone bridge consists of four resistors connected in a loop. An input voltage is
applied
across two junctions that are separated by two resistors. Voltage drop across other two
junctions forms the output. One or more resistors in the loop may be sensing resistors, whose
resistances change with the intended variables. In the bridge shown in Figure a, one resistor is
variable by strain. The other resistors— and are made insensitive to strains by
locating them in regions where mechanical strain is zero, such as on rigid substrates.

52
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

PIEZORESISTIVE SENSOR MATERIALS

Metal Strain Gauges
Metal strain gauges are commercially available, often in the form of metal-clad plastic
patches that can be glued to surfaces of mechanical members of interest. Resistors are etched
into the metal cladding layer. Typical strain gauge patterns are shown in Fiugre . A zigzagged
conductor path is commonly used to effectively increase the length of the resistor and the
amount of total resistance under a given area. Some of the criteria that are applied when
selecting a metal strain gauge include: accuracy;long-term stability; cyclic endurance; range
of operational temperature; ease of installation; tolerable amount of elongation; and stability
in harsh environment.To satisfy these requirements, commercial metal strain gauges are often
not made of pure metal thin films but of tailored metal alloys.

For micromachined sensors, the sizes of devices are very small. It is impractical to bond or
attach discrete strain gauges to devices. Instead, strain gauges are fabricated on mechanical
beams and membranes using monolithic integration processes. Metal resistors are generally
deposited (by evaporation or sputtering) and patterned. Elemental metal thin films can be
used as strain gauges in MEMS, with their gauge factors ranging from 0.8 to 3.0. The sources
and procedures for depositing elemental metal films are readily available. Strain gauges made
of thin film metals do not compare favorably with semiconductor strain gauges in terms of
piezoresistive gauge factors. However, metal can generally sustain much greater elongation
before fracture.As such, metal resistors can be placed on polymer materials for polymer
MEMS devices (e.g., tactile sensors) and provide improved mechanical robustness compared
with silicon counterparts.

Single Crystal Silicon

Semiconductor strain gauges are made by selectively doping silicon.The piezoresistive
coefficients of a doped single-crystal silicon piezoresistor are influenced by its relative

53
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

orientation to crystallographic directions. If we consider a rectangular coordinate system

having arbitrary orientation with respect to the crystallographic axes of a homogeneous
semiconductor, the electric field components and current density components are related by a
symmetric resistivity matrix as in

This is the familiar Ohm’s law expressed in a matrix form. From routine applications of
Ohm’s law, we are used to thinking of the relation between potential differences and current
as involving only a scalar constant of proportionality. This is obvious for a one-dimensional
conductor such as a link in a circuit. However, for electrical conduction in a three-dime sional
singlecrystalline media, the current density and the potential gradient will not in general have
the same direction.

The effective piezoresistive gauge factors attributed to each case in Table are determined
by multiplying the piezoresistive coefficient with the Young’s modulus in the direction ofthe
applied strain. Recall that the Young’s modulus of silicon is a function of the crystal
directions as well. For high-precision sensor applications, it is important to keep in mind that
piezoresistive sensitivity is not exactly constant.

APPLICATIONS OF PIEZORESISTIVE SENSORS

Over the years, piezoresistive sensing has been used for many categories of sensor
applications. A few representative examples are discussed in the following, to illustrate
unique device designs and fabrication processes, and performance specifications achievable.

PIEZOELECTRIC SENSING AND ACTUATION

The phenomenon of piezoelectricity was discovered in the late nineteenth century. It was
observed that certain materials generate an electric charge (or voltage) when it is under a
mechanical stress.This is known as the direct effect of piezoelectricity. Alternately, the
same material would be able to produce a mechanical deformation (or force) when an electric
field is applied to it. This is called the inverse effect of piezoelectricity. (Some literatures
refer to it as the converse effect of piezoelectricity.) Piezoelectric crystals can be considered
as a mass of minute crystallites (domains).The macroscopic behavior of the crystal differs
from that of individual crystallites, due to orientation of such crystallites. The direction of
polarization between neighboring crystal domains can differ by 900 or 1800 . Owning to
random distribution of domains throughout the material, no overall polarization or
piezoelectric effect is exhibited. A crystal can be made piezoelectric in any chosen direction
by poling, which involves exposing it to a strong electric field at an elevated temperature.
Under the action of this field, domains most nearly aligned with the field will grow at the
expense of others. The material will also lengthen in the direction of the field.When the field
is removed, the dipoles remain locked in approximate alignment, giving the crystal a remnant
polarization and a permanent deformation (albeit small).
The poling treatment is usually the final step of crystal manufacturing. Care must be taken in
all subsequent handling and use to ensure that the crystal is not depolarized, since this will
result in partial or even total loss of its piezoelectric effect.A crystal may be depolarized
mechanically, electrically, or thermally. Mechanisms for depolarization are further explained
in the following.

54
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Exposure to a strong electric field of opposite polarity to the poling field will depolarize a
piezoelectric element. The field strength required for marked depolarization depends on the
material grade, the time the material is subjected to the depolarization field, and the
temperature.
For static fields, the threshold is typically between 200-500 V/mm.An alternating field will
also have a depolarizing effect during the half cycles that it opposes the poling field.
Mechanical depolarization occurs when mechanical stress on a piezoelectric element
becomes high enough to disturb the orientation of the domains and hence destroy the
alignment of the dipoles.The safety limits for mechanical stress vary considerably with
material grade.
If a piezoelectric element is heated to a certain threshold temperature, the crystal vibration
may be so strong that domains become disordered and the element becomes completely
depolarized.This critical temperature is called the Curie point or the Curie temperature.A
safe operating temperature would normally be half way between 00C and the Curie point.The
properties of piezoelectric elements are time dependent.The stability as a function oftime is of
particular interest. Material characteristics may be degraded through aging effects
due to the intrinsic process of spontaneous energy reduction. The speed of aging can be
controlled through the addition of composite elements or through accelerated aging. Many
piezoelectric materials suffer from finite ion mobility. In other words, they do not provide
long-term static holding power when used in actuators. Design of piezoelectric actuators
operating in DC conditions must take into the account of electric leakage.

Mathematical Description of Piezoelectric Effects

Piezoelectric materials are crystals. The microscopic origin of piezoelectricity is the
displacementof ionic charges within a crystal, leading to polarization and hence electric field.
A stress (tensile or compressive) applied to a piezoelectric crystal will alter the spacing
between centers of positive and negative charge sites in each domain cell, leading to a net
polarization manifested as open circuit voltages measurable at the crystal surface.
Compressive and tensile stresses will generate electric fields and hence voltages of opposite
polarity. Inversely, an external electric field will exert a force between the centers of positive
and negative charges, leading to an elastic strain and changes of dimensions depending on the
field polarity. Not all naturally occurring or synthesized crystals exhibit piezoelectricity.
Crystals can be classified into 32 groups according to crystal symmetry. Centrosymmetric
crystal structures are crystals that are symmetric along all axes through the center of the
crystal. These crystals occupy 11 out of 32 possible groups and are non-piezoelectric
materials because the positiveand negative charge sites will not be spatially separated under
stress. Out of 21 noncentrosymmetricgroups, 20 are piezoelectric crystals. Piezoelectric
effects are strongly orientation dependent.The notation conventions for crystal orientations in
the context of piezoelectric polarization are discussed first. A piezoelectric material needs to
be poled in a particular direction to provide a strong piezoelectric effect, although some
materials exhibit natural or spontaneous polarization.The direction of positive polarization is
customarily chosen to coincide with the Z-axis of a rectangular system of crystallographic
axes X, Y, and Z. Alternatively, the normal stress components along axes X, Y, and Z are
denoted by subscripts 1, 2, and 3, respectively.As such, the poling axis always coincides with
axis 3. Shear stress and strain components about these axes are denoted by subscripts 4, 5,
and 6, respectively in below figure.

55
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

In a piezoelectric crystal, the constitutive equation that relates electrical polarization (D)
and applied mechanical stress (T) is

where d is the piezoelectric coefficient matrix, the electrical permitivity matrix, and E the
electrical field. Here, an electric field is applied in conjunction with the mechanical stress to
provide more generality.The electrical polarization is contributed by two parts—one
stemming from electrical biasing and one from mechanical loading. If no electric field is
present (i.e. E=0, ), then the second term on the right hand side of above equation can be
eliminated.

The inverse effect of piezoelectricity can be similarly described by a matrix-form constitutive

equation. In this case, the total strain is related to both the applied electric field and any
mechanical stress, according to

where s is the strain vector and S the compliance matrix. the above equation can be expanded
to a full matrix form:

If there were no mechanical stress present the strain is related to the electric field by

The electromechanical coupling coefficient k is a measure of how much energy is

transferred from electrical to mechanical, or vice versa, during the actuation process.

56
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

This relation holds true for both mechanical-to-electrical and electrical-to-mechanical energy
conversion.The magnitude of k is a function of not only the material, but also the geometries
of the sample and its oscillation mode.

Cantilever Piezoelectric Actuator Model

Piezoelectric actuators are often used in conjunction with cantilevers or membranes for
sensing and actuation purposes. General models for such piezoelectric actuators are rather
complex. Accurate analysis often involves finite element modeling. For limited cases, such as
cantilever actuator with two layers, analytical models have been successfully developed. we
will focus on the analysis of cantilevers with two layers of materials, at least one of them
being a piezoelectric layer.

The deflection of a two-layer piezoelectric structure can be described by compact formula.

Consider a cantilever with two layers, one elastic and one piezoelectric, joined along one
side.These two layers share the same length. A compact model for calculating the curvature
of bending has been made under the following assumptions:

1. The induced stress and strain are along axis-1, or the longitudinal axis of the cantilever;
2. Cross sections of the beam originally plane and perpendicular to the beam axis remain
plane and perpendicular to the resulting curved axis;
3. The beam maintains a constant curvature throughout the beam;
4. Shear effects are negligible;
5. Beam curvature due to intrinsic stress may be ignored;
6. The beam thickness is much less than the piezoelectric-induced curvature;
7. Second order effects such as the influence of d33 and electrostriction are ignored;
8. Poisson’s ratio is isotropic for all films.

The amount of force achievable at the free end of a piezoelectric bimorph actuator equals to
the force required to restore the tip of the actuator to its initial undeformed state. Since the
displacement is linearly related to force according to

the expression of the force is

57
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Piezoelectric sensors and actuators with more than two layers are commonly encountered
Several general techniques can be found in both under simple or arbitrary loading.

ZnO Piezoelectric Force Sensor

A patch of ZnO thin film is located near the base of a cantilever beam, as shown in Figure
below. The ZnO film is vertically sandwiched between two conducting films. The length of
the entire beam is l. It consists of two segments—A and B. Segment A is overlapped with the
piezoelectric material while segment B is not.The length of segments A and B are and
respectively. If the device is used as a force sensor, find the relationship between applied
force F and the induced voltage.
Solution. The c-axis (axis 3) of deposited ZnO is generally normal to the front surface of the
substrate it is deposited on, in this case the beam. A transverse force would produce a
longitudinaltensile stress in the piezoelectric element (along axis 1), which in turn produces
an electric field and output voltage along the c-axis. The shear stress components due to the
force are ignored.

The stress along the length of the piezoresistor is actually not uniform and changes with
position.For simplicity, we assume the longitudinal stress is constant and equals the
maximum stress value at the base. The maximum stress induced along the longitudinal
direction of the cantilever.

APPLICATIONS
Inertia Sensors
Cantilever Piezoelectric Accelerometer
One of the applications of piezoelectric sensors is micromachined accelerometers.An
exemplary device has been reported by a research group at the University of California at
Berkeley. The schematic diagram of the sensor is shown in Figure. A proof mass is attached
at the end of a fixed-free cantilever beam.The cantilever beam consists of multiple layers of
materials, with ZnO being the functional piezoelectric material. ZnO is used instead of PZT
because, although the PZT material offers greater piezoelectric coefficient, it does so at the
expense of greater dielectric constant and hence larger capacitance.

58
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The piezoelectric layer is sandwiched between a top conducting layer (platinum) and a
bottom conducting layer (polysilicon), similar to Example . A vertical acceleration will
deflect the cantilever, producing a longitudinal stress in axis 1, along the length of the
cantilever. (The ZnO film is polarized in the direction normal to the substrate.) Simple
analysis in Example does not apply here, however. The analysis is relatively complex
because the proof-mass weight and the piezoelectric layer are distributed in length and
thickness.The distribution of stress is nonuniform along the length. However, it can be said
that the output voltage is proportional to since the stress is applied in axis-1
and the electric polarization is measured in axis-3.The value of is 2.3 pC/N in this case. The
fabrication process is briefly discussed below. A silicon oxide and a silicon nitride layer are
deposited over the silicon wafer to serve as insulating layers. A phosphorous- doped
polycrystalline silicon film is deposited and patterned via reactive ion etching to define
electrical contacts to the bottom electrodes of the accelerometer. Here, the authors used
polysilicon instead of metal because polysilicon can withstand much higher processing
temperature and provide greater flexibility for down-stream process steps. Next, a -thick layer
of phosphosilicate glass is deposited by LPCVD, and patterned to define regions under the
suspended cantilever.A second layer of -thick silicon is deposited, covering bare silicon
surfaces, the first polysilicon trace, and the PSG sacrificial layer.This layer is patterned by
using reactive ion etching (RIE) process with photoresist as the mask. The RIE etch rate on
silicon nitride and oxide (including PSG) is smaller, reducing damages of over-time etch.
However, care must be exercised to prevent or minimize overetching the first polycrystalline
layer.The authors then removed the photoresist layer with acetone. A thin LPCVD silicon
nitride layer is deposited over the wafer. This film serves as a stress-compensation layer for
balancing a highly compressive stress in the ZnO film. The exact thickness of this layer
depends on the actual stress and thickness of the ZnO layer.
Next a ZnO layer on the order of 0.5 micometer thick is deposited by RF-magnetron
sputtering from a lithium-doped ZnO target. Finally, 0.25 micrometer -thick Pt thin film is
sputtered. The stress of the ZnO film is reduced by a rapid thermal annealing step.
Afterwards, the three layers area defined using ion milling to produce precisely defined
patterns. The underlying sacrificial layer is removed using HF solutions while the patterned
ZnO patch is protected by photoresist. The ZnO protection is necessary because, although the
film is covered by Pt on top, it is exposed on the side and through possible pin-holes on the Pt
films. The device exhibited a sensitivity of 0.95 fC/g and a resonant frequency of 3.3 KHz.

Acoustic Sensors
PZT Piezoelectric Acoustic Sensor
Bernstein and colleagues described the use of an array of piezoelectric transducers as an
underwater acoustic imager. The imager is akin to a CCD imager for optics. It consists of an
8 by 8 array of acoustic imaging sites.The cross-sectional view of each site is shown in
Figure . A layer of piezoelectric material—sol-gel deposited lead zirconate titanate (PZT),
lies on tops of a silicon micromachined diaphragm. The size of each membrane varies from
0.2 to 2 mm. Two electrodes sandwich a PZT thin film. The fabriation process is noteworthy
because it involves the use of sol-gel deposition of piezoelectric materials. After oxidizing the
silicon wafers (Figure a) and patterning the oxide, a heavy boron dose was diffused in all but
the oxide protected areas where through-wafer cavities would later be deposited (Figure b).
All remaining oxide is removed. A layer of low temperature oxide (LTO) is deposited using
the LPCVD method. A 50-nm-thick Ti and 300-nm-thick Pt layer are deposited as the bottom
electrode, with the Ti serving to increase the adhesion between the Pt and oxide (Figure c).
This was followed by deposition of PZT by spinning on a sol-gel mixture of lead acetate
trihydrate, zirconium n-propoxide and titanium isoproposide in a glacial acetic acid

59
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

solvent.The sol-gel material is spun on, dried at 150 degree to remove the solvent, heated at
400_C to remove residual organics, and preannealed at 600_C to densify the layer and
prevent further shrinkage (Figure d). Detailed receipe for making the sol-gel solutoin can be
found in the paper.The PZT material is poled at room temperature at 36 V DC bias for 2
minutes, to yield a resulting PZT film with a relative permitivity of 1400 and d 33 of 246
pC/N.
The PZT film is patterned by wet etching PZT in a solutoin containing buffered hydrogen
fluoride (BHF) and hydrochloric acid (HCl) (Figure e).The BHF is a mixture of aqueous
ammonium fluoride and acqueous hydrogen fluoride (HF). The top electrode is separated
from the ZnO by a 2.5 micrometer -thick polyimide dielectric layer.This is followed by the
deposition of a top electrode and its patterning (Figure f). The backside cavity is produced in
anisotropic etchant, with the frontside tempoerarily protected by a silicone elastomer cover
(Figure g). Prior to silicon etching, the LTO oxide on the backside is first removed.

Tactile Sensors
Polymer Piezoelectric Tactile Sensor
To reduce electrical noise and impedance mismatch effects, a two-dimensional matrix of high
input impedance metal-oxide-semiconductor field effect transistor (MOSFET) amplifiers
have been directly gate-contact coupled to the lower surface of a piezoelectric PVDF polymer
film. This MOSFET amplifier arrangement provides a separate, but identical high input
impedance voltage measurement capability for each taxel (tactile pixel). To realize a tactile
sensor no larger than an adult’s finger tip, a silicon IC with peripheral dimensions of 9.2 mm
by 7.9 mm is designed and fabricated.The prototyping was performed by MOSIS (Metal-
oxide-semiconductor Implementation System) foundry service.A portion of the IC’s area was

60
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

reserved for the MOSFET amplification and output interface circuitry.The 8 by 8 taxel matrix
was allocated to an area of 5.3 mm by 5.3 mm.

A schematic diagram of the sensor array with cross section revealed is shown in Figure. A
continuous, poled PVDF film is attached to the front surface of the silicon chip.A –thick
urathane conformal coating layer was deposited on the PVDF film by spin coating. Individual
taxel electrodes are separated from their nearest neighbors. Although this work focuses on
measuring the normal stress component, tactile sensors with polymer piezoelectric materials
have been developed with component-selective response characteristics.

61
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

UNIT IV MICROMACHINING

Silicon Anisotropic Etching – Anisotrophic Wet Etching – Dry Etching of Silicon – Plasma Etching –
Deep Reaction Ion Etching (DRIE) – Isotropic Wet Etching – Gas Phase Etchants – Case studies -
Basic surface micro machining processes – Structural and Sacrificial Materials – Acceleration of
sacrificial Etch – Striction and Antistriction methods – LIGA Process - Assembly of 3D MEMS –
Foundry process.

Silicon Anisotropic Etching

Bulk micromachining is an important class of MEMS process. In bulk micromachining
processes, a portion of the substrate (bulk) is removed in order to create freestanding
mechanical structures (such as beams and membranes) or unique three-dimensional features
(such as cavities, through-wafer holes, and mesas). Bulk micromachining can be applied to
silicon, glass, gallium arsenide and other materials of interests. In this chapter, we focus on
discussing bulk micromachining of silicon substrates. There are two major categories of
processes for bulk silicon etching according to the medium of the etchant: wet etching and
dry etching.Wet silicon etching processes use liquid chemical solutions in contact with
silicon. Dry etching processes use plasma (high-energy gas containing ionized radicals) or
vapor-phase etchants to remove materials. Bulk silicon etching can be classified according to
the three-dimensional distribution of etch rates and profiles of resultant microstructures. The
etch rate of wet bulk etching may depend on crystal orientation. The etch rate of dry bulk
etching may depend on directions in a wafer. If the etch rates in all directions are identical,
the etching is said to be isotropic. If the etch rate is orientation dependent, the etching is said
to be anisotropic. Representative crosssectional profiles of cavities created by isotropic and
anisotropic bulk etching are shown in Figure. As shown in Figure, etching action is not
confined to the region exposed through opening areas of a mask. Materials under the mask
may be undercut. Undercut is desired in order to create suspended mechanical structures. For
certain applications, undercut may be undesirable and should be minimized by proper mask
design and careful process control.

ANISOTROPIC WET ETCHING

Rules of Anisotropic Etching—Simplest Case
Let us first examine the simplest case—the profile of an etched pit in a <100> oriented
siliconsubstrate when the mask contains a rectangle or a square open window, with edges
aligned tothe <110> direction.

62
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

When a silicon wafer is immersed in a wet silicon etching solution, the atomic layer that is
exposed through the open window is etched first. Wet anisotropic etchants exhibit drastically
different etch rates along different crystal planes. In general, the etch rate along direction
<111>
is the slowest among all crystal orientations. The reason for crystalline-dependant etch rate is
not yet fully understood though conjectures exist. It is believed that, since wet anisotropic
etchants remove silicon by first oxidizing a silicon surface (using the oxidizing constituents
within the solution) and then removing the oxide (using oxide-etching constituents), the
difference of etch rate is attributed to the difference of oxidation rates. However, there seems
no reported direct evidence correlating experimentally data and reaction rate analysis.
Microscopically, the etch rate difference is attributed to atomic bonding energy of silicon
atoms on various surfaces.Atoms on different crystalline surfaces are associated with
different number of neighbors, bonding energy, and degrees of difficulty for removal.

In the case of a square window, the cavity would end in a perfect point. In the case of a
rectangular opening, a cavity would end in a knife-edge at the bottom. In practice, due to
nonideality of masks, a perfect point is almost never encountered.
Rules of Anisotropic Etching—Complex Structures
In the previous section we reviewed a very simple and commonly encountered case: masks
with rectangular or square openings with edges aligned to <110> directions, lying in <100>
silicon surface. We now build up the complexity of analysis by relaxing limitation on mask
shapes. First, we eliminate the restriction that the edges of a window must be a
noninterrupted straight line. Next, we eliminate the restriction that the edges must be parallel
to <110> directions.

63
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Instead of reviewing the text of rules first, let us observe illustrated examples of actualetching
cases as predicted by a computer simulation software.The progression of etch profiles from
two independent mask patterns—labeled A and B—is illustrated in Figure. These two
window openings in the mask are very different. However, after a sufficiently long etch time
(e.g., 170 min), they result in very similar three-dimensional profiles—blind STP-type
cavities with edges aligned to <110>(Figure h). Note that etching occurs underneath certain
regions that are originally masked.For example, the material underneath the extended beam in
B is gradually removed starting at its distal end.

64
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The etching rules can be discussed easily by tracking how corners of mask patterns are
transformed. There are two types of corners—convex corners and concave corners. At a
convex corner, the solid angle of masked region is less than 1800 At a concave corner, the
solid angle of masked region is greater than 1800Solid arrows in Figure a point to concave
corners, while convex corners are indicated by arrows with dashed lines. For convex corners,
the fastest etching planes dominate the three-dimensional shape evolving underneath the
masked region. For concave corners, the slowest etching planes dominate the three-
dimensional shape. In other words, convex corners tend to be undercut rapidly, exposing fast
etching planes such as <211>and <411>. At concave corners, slow etching planes such as
<111>tend to develop and eventually prevail. The undercut can be used to create suspended
micro mechanical structures. For example, the mask opening B will result in a suspended
cantilever, provided that the mask is chemically and mechanically sturdy enough to survive
the entire wet etching sequence.The material of the cantilever is that of the mask layer.The
cantilever can be made of a variety of materials, including silicon nitride, silicon oxide, and
heavily doped silicon. Other materials are possible as long as they exhibit good etching
selectivity.A comprehensive survey of etch rates by various etchants on silicon substrate and
important mask materials can be found in the literatures.

Chemicals for Wet Anisotropic Etching

EDP (ethylene diamine pyrocatechol) is a frequently used anisotropic etchant. It is sometimes
referredto as EPW (ethylenediamine pyrocatechol and water). Etch rate ratio for <110> and
<111>directions reaches 35:1 or higher.The etch rate on <110>silicon orientation is between
0.5 to 1.5 m/min. Etch rates for silicon nitride (LPCVD) and silicon oxide are almost
negligible. The etch rate on silicon dioxide is 1–2 Å/min and on silicon nitride is Å/min. On
one hand, it is highly selective, allowing thermally grown silicon oxide as mask. On the other
hand, this means that even very thin natively grown oxide becomes barrier to etching if it is
not properly removed.
Etching is often conducted in a chemical fume hood because of the toxicity of the EDP vapor.
The chemical solutions are often heated to 90 to 1000 C during the etching process. Escape of
vapor may alter the concentration of etching chemicals and the etch rate. To maintain a
constant concentration, the escaped water vapor is condensed on a water-cooled top and
65
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

allowed to drip back into the flask.This system, called a reflux system, is illustrated in Figure
. KOH (potassium hydroxide) is a low-cost alternative to EDP. The KOH etchant can be
easily made in a laboratory by mixing solid KOH with water in a slightly exothermic
reaction. The etching characteristics depend on the concentration of KOH and the
temperature of solutions. The most commonly used concentration is 20–40 wt%.The etch rate
ratio for <100> and <111> is also very high (even higher than that of EDP). The etch rate on
silicon nitride (LPCVD) is negligible. However, the etch rate on silicon oxide (thermally
grown) is 14 angstrom/min, much greater than the etch rate on silicon oxide by EDP. For
both EDP and KOH, the etch rate for silicon is greatly reduced when the silicon is heavily
doped with boron or phosphors. The degrees of etch rate reduction as a function of doping
concentration is different for these two chemicals. For EDP, greater than 7x1018 cm-3 boron
reduces etch rate by 50 times.Therefore, it is possible to use heavily doped silicon as an etch
stop layer against over time etch. Normally, the wet etching is conducted without electrical
bias to the wafer and under normal ambient light conditions. Active electrical and optical
activation, however, can be used to introduce new etching behavior. For example, pulsed
potential annodization exploits the etching behavior of a p-n junction. Photovoltaic
electrochemical etch stop technique (PHET) is based on the electrochemical growth of a
passivation silicon oxide. This method can be used to provide external, flexible control of
etch stop and does not require doping process.
Tetramethyl ammonium hydroxide (TMAH) is an alternative silicon etchant. It was
discovered after EDP and KOH. It has very slow etch rate on aluminum, a common IC
metallization material. As such, TMAH can be used to etch wafers with aluminum wire leads.
Etch rate on aluminum can be further reduced if silicon is predissolved into the solution;
however, the dissolved silicon is known to create more surface roughness. TMAH also
exhibits reduced etch rate for heavily doped silicon.

DRY ETCHING AND DEEP REACTIVE ION ETCHING

Gas phase plasma etching is a complementary alternative to wet silicon anisotropic etching.
Plasma etch does not require wafers to be in contact with wet chemical solutions, thus
simplifying cleaning steps. Plasma etch offers different options of masking material. The
process temperature is also lower. Generally speaking, wafers are not heated although
prolonged plasma etching may cause the wafer temperatures to increase. The sidewall profile
is different from what is achievable with web anisotropic etching. In fact, the anisotropy of
the vertical wall can be controlled, within a certain range, by adjusting several process
parameters, including the process pressure, DC bias voltage, power input, and the chemical
species used.
Deep reactive ion etching (DRIE) is a special kind of RIE process. It is capable of producing
deep and high-aspect ratio features with near vertical sidewalls. They are based onpatents
currently owned by Robert Bosch GmbH and Texas Instruments Corporation. The key to
achieving continuous deep etching with high aspect ratio is to conduct the etching process in
small depth increments. This approach utilizes repeated cycles each containing an etching
step followed by a passivation step.A layer of inhibition film is deposited during the
inhibition step on all walls of a cavity. During the next etching step, the inhibition film is
preferentially removed from the bottom of the trenches due to ion bombardment, while
preventing the etching of the sidewalls. Process parameters play important roles in
determining surface morphology and mechanical performances of devices. DRIE silicon
etching has gained popularity despite the high cost of processing equipment. The DRIE
process offers fast etch rate, vertical sidewalls, room temperature process capability, and the
ability to easily correlate the mask with the resultant three-dimensional structures.The etch
selectivity is a major benefit.The DRIE process can use photoresist, silicon oxide, and metal

66
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

as mask. In addition, the etching uniformity on a wafer scale is generally much improved
over those of regular RIE etch. Though powerful, the DRIE technique alone is capable of
creating nothing more than passive three-dimensional parts. For this reason, deep reactive ion
etching must be combined with other processing styles and steps to increase the electrical and
mechanical complexity of micro structures. The DRIE technique can be combined with many
processing styles, including anisotropic etching, to create a rich variety of innovative
geometries. Deep reactive ion etching of materials other than silicon has also been
demonstrated, including piezoelectric materials (i.e., quartz, PZT), Pyrex glass, and
commercial PMMA.

ISOTROPIC WET ETCHING

The most commonly used isotropic silicon etchant is HNA—a mixed chemical solution
consisting of at least three constituent acids—the letter H stands for hydrofluoric acid, N for
nitric acid, and A for acetic acid [31].Various mixing ratios of three major ingredients are
possible, resulting in different etch rates on silicon and etch selectivity against masking
layers. Because the solution contains HF, the etch rate on oxide is relatively quick, at
approximately 300 Å/min.
GAS PHASE ETCHANTS
Gas phase etching of silicon using XeF2 or BrF3 is a complementary alternative to wet
etching and plasma etching. Under room temperatures, XeF2 exist in solid phase and BrF3
exists in liquid phase. XeF2 sublimates and BrF3 vaporizes to form active gas when
subjected to pressure below 100 mTorr. These gases react with silicon (in single crystal or
polycrystalline forms) strongly, even under room temperature, resulting in isotropic etch
profiles. On the other hand, the gas phase etchants have negligible etch rate on common mask
materials, including silicon dioxide, metal, and even photoresist. Care must be taken because
the reaction byproducts contain HF vapor that, if released into air accidentally, may become
absorbed into water vapor and constitute air-borne drops of highly concentrated HF acid.
Bulk silicon etch rate is very high.The etch rate is generally not characterized as etching
depth per unit time. Rather, since etching is administered by injecting fixed amount of
etching agents in burst cycles, the etch rate is given as
m/cycle.

NATIVE OXIDE

A thin layer of silicon oxide would inevitably form on silicon naturally when the silicon
isplaced or stored under room temperature and environmental oxygen and moisture
levels.This layer of material is very thin as the oxide growth temperature is low. However,
this oxide may present a significant etching barrier if the silicon etchant has small etch rate on
silicon dioxide. It is a common practice to remove native oxide before wet anisotropic etch or
plasma etch steps to ensure that the reaction with silicon starts predictably. The thickness of
the native oxide is so small, it often eludes detection by thin film thickness measurement
instruments.A practical laboratory technique to determine whether a native oxide layer is
present is the water-beading test. A pure silicon surface is hydrophobic while an oxidized
silicon surface is hydrophilic.The hydrophobic/hydrophilic nature of a surface can be quickly
and inexpensively determined by dispensing drops of water on the substrate surface and
observe the shape of the water drop (Figure ).Water beads easily on hydrophobic silicon
surfaces and rolls off the substrate when the substrate is tilted. On the other hand, water
would spread uniformly on a hydrophilic oxide surface. The native oxide layer can be
removed easily by dipping the wafer in low-concentration hydrofluoric acid solutions (e.g.,
5%) for a few seconds.

67
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

SPECIAL WAFERS AND TECHNIQUES

Infinite selectivity between target and masking materials is always desirable but rarely a
reality. As the number of layers increase, high etch selectivity becomes difficult to
accomplish. This tends to reduce the yield of MEMS processes. Bulk etching involving long
etching time is especially vulnerable to etch rate nonuniformity and variability.Wafers with
built-in etch stop layers are important for increasing process yield.Two types of special
techniques are discussed here.
SOI Wafer SOI stands for silicon on insulator. An SOI wafer is a specialty wafer consisting
of two single crystal silicon wafer sandwiching an insulator layer (typically silicon dioxide).
The SOI wafer can be formed in a number of ways. The thickness of each layer can be
independently tailored. The cost of SOI is greater than that of a regular silicon wafer, since it
must undergo many extra steps and must pass stringent process control. However, there are
many important benefits:
1. The insulator wafer in the SOI process can serve as an important etch stop layer for both
silicon dry etching and wet etching;
2. the single crystal silicon layer can host high-quality piezoresistors and doped electrical
layers and circuitry;
3. the single crystal silicon layer is free from stress.
Electrochemical Etch Stop Layer Heavily doped silicon can reduce etch rate by wet silicon
etchants. Alternatively, doped layers can be biased during the etching process to realize
electrochemical etch stopping. Such layers are important for early MEMS commercial
products
such as the blood pressure sensor.
BASIC SURFACE MICROMACHINING PROCESSES
Sacrificial Etching Process
A simple, two-layer sacrificial etching process for forming a suspended cantilever beam is
illustratedin Figure .
Here, the desired freestanding structure is a fixed-free cantilever anchored at one end to the
substrate.The fabrication process starts with the deposition of a sacrificial material on the
silicon wafer (a).The sacrificial material is defined by photolithography and patterned (b),
followed by the deposition and patterning of a structural thin film (c and d). The sacrificial
material should ideally provide a mechanically rigid and chemically reliable support for the
structural layer. In other words, the sacrificial layer needs to behave as a reliable placeholder
during the process.The sacrificial material is later removed selectively to free the overlying
structural layer (e). If the sacrificial layer etching is conducted in a wet chemical solution, the
liquid must then be removed to produce the final structure (f).

68
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Micro Motor Fabrication Process—A First Pass

The basic motor fabrication process is shown in Figure. The process begins with asilicon
wafer. Since no bulk machining is involved, the orientation of the wafer is not relevant. The
wafer is first covered by a sacrificial thin film, such as silicon oxide deposited by using the
LPCVD method. (In this case, the thin film covers both sides of the wafer. However, only the
materials on the front side are of relevance for the outcome of fabrication.) A structural
material, polycrystalline silicon, is deposited on the wafer next (Figure b). In this case, the
structural layer forms the rotor. In Figure c, a photoresist thin film covers the front surface of
the wafer. The photoresist, after development and curing, serves as a mask in a subsequent
reactive ion etching step, which transfers the pattern in photoresist to the polysilicon
structural layer. Thephotoresist is then removed, either using oxygen plasma (dry etching) or
organic solvents (wet etching). The wafer is covered with yet another layer of oxide
sacrificial material (Figure d). The deposited film conformally covers both horizontal
surfaces and vertical sidewalls. The material of the second sacrificial layer may be different
from the previous one, although in thiscase, LPCVD silicon oxide is a convenient choice. It
offers satisfactory step coverage and temperaturecompatibility.
A center hub that restricts the lateral translation of the rotor is made next. The hub also
prevents the rotor from escaping from the substrate. In order to produce the hub, an anchor
window reaching to the substrate is made (Figure e). It is accomplished by photolithography
and wet chemical etching through two sacrificial layers. A second structural layer (Figure f)
is deposited. It is firmly attached to the substrate via the open window. A photoresist film
covers the second structural layer and then undergoes patterning and development. It is used
to define the second structural layer (Figure g). Finally, both sacrificial layers are removed by
immersing the wafer into HF etch solutions. Though the rotor is shown to be at an elevated
position in the drawing of the finished device (Figure h), it could easily drop to the substrate
under gravity and form contact on its broad side.The rotor may also be stuck to the substrate
if the wafer is not properly dried.

69
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Micro Motor Fabrication Process—A Second Pass

The fabrication process discussed earlier is capable of generating a basic micro motor.
However,
its structure and performance is far from optimal. One major drawback associated with the
design shown in Figure is the lack of friction control. One way for mitigating the friction
is to reduce the contact area between the rotor and the substrate.A process for realizing such a
feature is discussed in the following and diagrammed in Figure. Starting with a silicon wafer,
a dielectric insulating thin film is deposited first (Figure a).
Notice that only the deposited material on the front side is shown for brevity.
In Figure b, a sacrificial layer is deposited. Two sets of photolithography and etching steps
are performed to produce the cross-sectional profile illustrated in Figure e. The first step
involves coating the sacrificial layer with a photoresist, patterning and developing the
photoresist, and using the photoresist as a mask to remove a portion of the sacrificial layer
thickness (Figure d). The photoresist layer is stripped and then replaced by a new one. A new
round of processing steps including photoresist coating, patterning, and development of
photoresist is performed again. This time, the exposed sacrificial layer is etched through its
entire thickness (Figure e). Process steps from Figure d through Figure k parallel the process
discussed in Section .2.The only difference is that the second structural layer (for the hub) is
silicon nitride instead of polycrystalline silicon. In step k, the sacrificial layer is completely
removed to free the rotor. Notice that the rotor may contact the substrate only at tiny bumps
defined in Figure c. The effective contact area is much reduced compared to the previous
case—this feature reduces friction and the chance of the rotor being stuck to the substrate.The
rotor and the hub are not in tight contact at all times.The rotor and the hub must beseparated
by an in-plane gap whose size is determined by the thickness of the second sacrificiallayer
applied. During high-speed rotation, the rotor may wobble within the constraint of the huband
create alternating contact points, much like a hula ring moves around the waist of a dancer.

70
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

micromotor fabrication

Micro Motor Fabrication Process—Third Pass

The contact between the rotor and the hub creates additional friction and wear. One
solutionfor this problem is to reduce the friction coefficient between the rotor and the hub by
making contact surfaces out of silicon nitride. A process for integrating a sidewall silicon
nitride bearing structure is illustrated in Figure. A silicon wafer is first coated with a
dielectric barrier consisting of silicon oxide and silicon nitride (a). A polysilicon grounding
plane is deposited (b), followed by the deposition and patterning of a first sacrificial layer (c).
A second polysilicon structural layer (d) is conformally deposited. The oxide sacrificial layer
and the polysilicon layer are patterned photolithographically and etched. A layer of silicon
nitride is deposited conformally, covering horizontal and vertical surfaces (f). A globally
reactive ion etching follows. The process parameter is designed such that theetch rate in the
vertical direction is greater than that in the horizontal directions due to physicaletching.While
silicon nitride on front-facing surfaces (such as the hub bottom and rotor front surface) is
removed, silicon nitride on vertical surfaces is retained (g). A timed wet etch is performed to
create controlled lateral undercut in the first sacrificial layer (h). Following this step, a layer
of sacrificial material is deposited conformally. The sacrificial layer covers all surfaces,
including those below the overhanging structures (i). A photoresist layer is spin-coated and
patterned, exposing the bottom of the hub region. Sacrificial material in the open window is
etched to gain access to the underlying polycrystalline silicon (j). After deposition and
patterning of another layer of silicon nitride as the hub, (k), a global sacrificial etch is
performed to release the rotor. The finished motor, with cross section shown in (l), contains

71
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

friction-control films on the sidewall of the hub. Silicon nitride deposited in the undercut
region created in (h) prevents the rotor from dropping to the substrate surface.

STRUCTURAL AND SACRIFICIAL MATERIALS

There is a large selection of candidate structural and sacrificial materials.The choice of
materialsdepends on specific needs of each individual application. In this section, we will
first evaluate general criteria for identifying successful candidate materials for structural and
sacrificial layers.The major focus is on thin films of polysilicon, oxide, and silicon nitride.
Material Selection Criteria for a Two-Layer Process
Essential prerequisites for selecting structural and sacrificial layer materials in a basic
twolayer process is modified based on the Ideal Process Rule:
1. The deposition of the structural material on top of the sacrificial material must not cause
the sacrificial layer to melt, dissolve, crack, disintegrate, or becomes unstable or destroyed in
any way;
2. The method used for pattering the structural layer must not attack the sacrificial layer and
any existing layers on a substrate;3.The method used for removing the sacrificial layer must
not attack, dissolve, or destroy the structural layer or the substrate.
The most established material pair is the polycrystalline silicon (as structural layer) and
thephosphosilicate glass (or PSG, as the sacrificial layer).This material system satisfies the
three criteria mentioned earlier.The PSG can withstand the deposition of the polysilicon
layer. The polysilicon can be patterned by using plasma etching, which etches PSG and oxide
at much reduced rates. The PSG and oxide can be removed by using wet HF acid solution,
which exhibits very high undercut etch rate on PSG but minimal etch rate on silicon. Certain
processes may require two or more layers or structural layers and/or sacrificial layers. The
processes for micromotors generally involve multiple structural and sacrificial layers. In such
cases, the criteria of successful process design becomes more complex. Now must be
expanded in the followingmanner:
1. Any layer that is deposited must not damage the underlying layers;
2. Any sacrificial etching process must not damaging any layers on the wafer;
3. Any process to pattern any sacrificial or structural layer must not etch all layers present
on the wafer.
In many practical processes, the device application may dictate the use of certain types of
substrate, sacrificial, structural, or interfacial materials. In such cases, it is impossible to limit

72
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

the material choices to the LPCVD materials mentioned earlier. When other materials are
introduced, it may be difficult to satisfy the three criteria listed above. One may have to pick
a compromise choice of structural-sacrificial layer combination. Under these circumstances, a
number of general factors should be considered when selecting the structural and sacrificial
layers. These include: (1) etch rates and etching selectivity; (2) achievable layer thickness; (3)
temperature of material deposition; (4) intrinsic stress of the structural layer; (5) surface
smoothness; and (6) cost of materials and processes.
During the sacrificial etching, both the structural layer and the sacrificial layer are exposedto
etchants.The etch rates of the etchant on the sacrificial layer and the structural layer are
labeled r sa and rst respectively.The etch selectivity, defined as the ratio of and should be as
high as possible. For devices involving multiple structural layers and sacrificial layers,
successful choice of materials, etchants, and etching conditions must strive to maintain high
etch selectivity at each
and every step of the process. It takes time and effort to build a comprehensive and reliable
etch rate table. A good starting point for beginners is published etch rate data. Papers
published by Williams et al. archived the cross reactivity between 53 materials and 35
etching methods. However, it should be noted that etch rates can be influenced by a large
number of factors. Information provided in the two papers could be overwhelming for
beginning readers. To ease the learning curve, three tables are provided to capture essential
material- and processrelated information.
Ideally, a high sacrificial etch rate r sa is also desired to minimize the time required to
complete
the sacrificial etching step. If two competing process options have identical selectivity, the
one
with higher rsa is preferred. If a process has lower etch rate on the sacrificial material but
higher
etch selectivity over another, the choice should be made by considering other additional
factors.

Thin Films by Low Pressure Chemical Vapor Deposition

A widely practiced method for depositing structural and sacrificial materials is the chemical
vapor deposition (CVD). In a CVD deposition chamber, solid thin films are formed on
wafersby condensation of vapor or adhesion of solid-phase reaction byproducts.The material
growthis typically conducted in a sealed chamber to prevent introduction of particles and
leakage ofreaction gases.
The reaction energy is provided by heat or by plasma power. If the energy is provided byheat
alone and conducted under low processing pressure (say a few hundred millitorr), theprocess
is called low-pressure chemical vapor deposition (LPCVD). If the energy is provided by
plasma power, the process is called plasma enhanced chemical vapor deposition
(PECVD). In an LPCVD process, wafers are placed in the chamber filled with flowing
process gases.The chamber is typically made of quartz to sustain high deposition
temperatures. The typical range of pressure is 200–400 mTorr (of 1 atmosphere pressure).
There are three major LPCVD materials used in MEMS: polycrystalline silicon, silicon
nitride, and silicon dioxide. LPCVD polysilicon and oxide materials were established in the
integrated circuit industry. Naturally, these two materials were the first to be used in MEMS.
The LPCVD polycrystalline silicon (polysilicon) material is used extensively in the
integrated circuits manufacturing. Doped polysilicon typically serves as the transistor gate,
conductors, and resistors. Undoped LPCVD oxide is often deposited on a finished chip as an
encapsulation.

73
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

LPCVD Polycrystalline Silicon

Polysilicon is deposited within a temperature range of 580-6200C by decomposing silane gas.
The reaction formula is

Process parameters such as pressure, gas flow rates, and temperature determine the
microstructure and hence electrical and mechanical properties of the polysilicon films. For
example, it is known that at a deposition temperature below 580 0C amorphous silicon will
form in a horizontal LPCVD system. Doping affects not just electrical characteristics, but
also mechanical ones as well (e.g., stress of polysilicon). Doping of polysilicon can be
achieved using two methods. First, the dopant atoms can be incorporated into undoped
polysilicon by diffusion or ion implantation. Second, the dopant atoms can be incorporated in
situ during the LPCVD growth by introducing additive gas components containing dopant
atoms. For polycrystal silicon used in MEMS, the following properties are important:
conformality, stress, and etch rate. Conformal coating is desired to cover three-dimensional
topography. For clamp-clamped structures, internal compressive stress can cause buckling.
For cantilevers, gradient of stress through the thickness can cause deformation. The
polycrystal silicon used for CMOS circuitry and for MEMS are actually different. For
MEMS, the need to have high growth rate (since the film is typically thicker than used for
circuit applications), low stress, and low etch rate by hydrofluic acid sacrificial etchant
dictates the deposition conditions.

ACCELERATION OF SACRIFICIAL ETCH

Sacrificial release of large-area plates or removal of sacrificial material inside long or
blindchannels are generally very time consuming. Due to small characteristics dimensions
(thickness
in the case of large area plate and cross section in the case of channels), the transport rate of
fresh etchants and byproducts decreases as the etching progresses. The chemical transport
characteristic for wet chemical etching of sacrificial materials in long blind channels has been
studied for PSG (with silicon nitride as sacrificial layer) and photoresist (with Parylene as
sacrificial layer). One strategy for decreasing the overall time for sacrificially undercutting
large-area plates is to deploy small openings, called etch holes.A scanning electron
micrograph of suspended micromachined flaps, each supported by two cantilever beams.The
area of the flap is 200 m by 200 m. Four etch holes on each flap initiate sacrificial
undercut from within the interior of the flap, as well as from the edges. It can significantly
shorten the required time of etching. A micrograph of an etch hole in polysilicon. In most
cases, the etch holes are small so as not to influence performances adversely. In certain
applications, however, etch holes may exert finite influence on the performance of devices.
For example, etch holes on an optical reflector causes diffraction as well as reduction of the
reflectance. In cases where etch holes are not preferred or allowed, alternative
etchingmethods and materials become necessary. A number of possible techniques have been
demonstrated in the past. The following materials have been reported as having extremely
fast etch rate and high selectivity: (1) Dendritic polymer such as hyper branched polymers
(HBPs), (2) ZnO thin films.The etch rate on ZnO by 2% HCl exceeds 1000 Å/s with no
bubble formation. If the sacrificial material is metal, the undercut may be accelerated by
electrolytic dissolution with an active electrical bias. Furthermore, self-assembled
monolayers can be used to release large-area micro devices without wet chemical etching.
Researchers have also developed porous silicon (with through-film pores measuring 10–50
nm in diameter) for use as structural layers. The pores allow underlying sacrificial layer to be

74
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

removed rapidly without resorting to macroscopically defined etch holes. Because the holes
are nanoscopic in sizes, they have minimal impact on performances.

STICTION AND ANTI-STICTION METHODS

Sacrificial material removal is often accomplished using wet chemical solutions, because of
thehigh speed of etching, simplicity of setup, and generally good selectivity. It is necessary to
dry the wafer and chips afterwards, by natural or forced evaporation. However, the drying
process is not always straightforward.As liquid is gradually removed through evaporation, the
top surface of a microstructure is exposed to air first. Liquid trapped underneath the
suspended micro structure will take much longer time to escape.A surface tension force
develops at the interface of trapped liquid and air acting in the direction tangential to the
liquid–air interface. For macroscopic devices, the surface tension force is negligible and does
not cause any significant deformation. However, since microscale devices often use
compliant structures and involve small gap spacing, the surface tension forces can cause
appreciable deformation of the surface microstructures. Often, contacts with the substrate are
made. Studies have been conducted in detail to characterize surface energies of bonding of
selective material systems.The contact between the suspended structure and the substrate
can lead to irreversible damages. Upon contact, strong molecular forces (e.g., van der Waals
force) are incurred to reinforce the attraction. Further, solid bridging is very likely due to the
presence of fresh reaction byproducts.
This failure mode of microstructures is referred to as stiction, a hybrid word combining
sticking and friction.

75
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Many practical methods have been developed to address the stiction issue.These methods
take one of four approaches:
1. eliminate the capillary attractive force by modifying the solid–liquid interface chemistry;
2. prevent excessive bonding force from occurring, for example, by elevating solution
temperature or reducing surface contact areas;
3. release structures that are stuck to the substrate using various energetic input forms,
locally or globally;
4. provide counteractive force to mechanical structures to prevent contact, for example by
taking advantage of bending stemming from intrinsic stress.

Two representative methods under these approaches are discussed in more detail in
thefollowing. Adverse surface tension associated with the liquid-air interface can be
circumvented by using the supercritical fluid drying method. This technique was initially
developed in the biology community to dehydrate delicate biological tissues for morphology
studies while avoiding structure collapses and deformation. Let’s examine the process by
observing a generic phase diagram of a solvent material . Three phases—solid, liquid, and
air—are the most familiar to readers. In addition, a supercritical phase occurs at high
pressure and temperature. A typical drying processing utilizing the supercritical phase is
described below. The chip containing released microstructures is immersed in a liquid and
placed under a moderate pressure (instead of the atmosphere pressure) and room
temperature.This starting condition is indicated by Point 1 on the phase diagram. The
temperature of the liquid is increased while maintaining the pressure constant.The solvent
makes a transition from the liquid phase to the supercritical phase (Point 2).The pressure of
the supercritical fluid is then dropped, causing the supercritical fluid to turn into a vapor
(Point 3). The transformation between liquid to supercritical fluid and then from supercritical
fluid to vapor involve virtually zero surface tension. The stiction can be reduced by using
hydrophobic coatings on micro structures and the
substrate to reduce bonding energy. Surfaces can be terminated with long-chain molecules
that

76
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

self-assemble to form a hydrophobic self-assembled monolayer (SAM). Alternatively,

flurocarbon materials (Teflon like) deposited by plasma methods can be used as well.
Hydrophobic surface treatment finds applications beyond antistiction coating.For example,
the use of hydrophobic patterned regions reduces surface bonding energy and allows lifting
of wafer scale devices from a mold without using sacrificial wet etching. Local
hydrophobicity modification also allows automated assembly of small parts in an self-
organized manner.

77
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

UNIT V POLYMER AND OPTICAL MEMS

Polymers in MEMS– Polimide - SU-8 - Liquid Crystal Polymer (LCP) – PDMS – PMMA –
Parylene –Fluorocarbon - Application to Acceleration, Pressure, Flow and Tactile sensors-
Optical MEMS –Lenses and Mirrors – Actuators for Active Optical MEMS.

INTRODUCTION
Polymers are large, usually chainlike molecules that are built from small molecules. Long
chain polymers are composed of structural entities called mer units, which are successively
repeated along the chain. A bulk polymer is made of many polymer chains. The physical
characteristics of a polymer material depend not only on its molecular weight and make up of
polymer chains, but also on the ways the chains are arranged. Polymers can be classified into
three major classes: fibers, plastics, and elastomers (rubbers). The largest number of different
polymeric materials comes under the plastics classification. Polyethylene, polypropylene,
polyvinyl chloride (PVC), polystyrene, fluorocarbons, epoxies, phenolics, and polyesters are
all classified as plastics. Many plastic materials are manufactured by different vendors and
carry different trade (common) names. For example, acrylics (polymethylmethacrylate,
PMMA) is known as Acrylite, Diakon, Lucite, and Plexiglass in trade.Vendors may
incorporate additive substances into polymers to adjust their physical, chemical, electrical,
and thermal characteristics and to change their appearances. According to their origins,
polymers can be categorized into two groups: naturally occurring polymers and synthetic
polymers. Naturally occurring polymers—those derived from plants and animals, include
wood, rubber, cotton, wool, leather, and silk. Synthetic polymers are derived from petroleum
products.

Polymers can be classified by its response to temperature. Thermal plastic polymers

(thermalplasts) can be remelted and reshaped repeatedly whereas thermal setting polymers
(thermalsets) take on a permanent shape after being melt-processed once.The melting of a
polymer crystal corresponds to the transformation of a solid material,from an ordered
structure of aligned molecular chains, to a viscous liquid in which the structureis highly
random.This phenomenon occurs upon heat at the melting temperature,The temperature at
which a polymer experiences a transition from rubbery to rigid stateis termed the glass
transition temperature, Glass transition occurs in amorphous and semicrystallinepolymers,
and is due to a reduction in motion of large segments of molecular chainswith decreasing
temperature.The mechanical properties of polymers differ from those of metals and
semiconductors inseveral major aspects. Some notable facts are summarized below:
1. Polymer materials cover a wide range of Young’s modulus.The modulus of elasticity may
be as low as several MPa for highly elastic polymeric materials, but may run as high as 4 GPa
for some of the very stiff polymers.
2. Maximum tensile strengths for polymers are on the order of 100 MPa, much lower than
that of metal and semiconductor materials.
3. Many polymers exhibit viscoelastic behavior.When a force is applied to it, an
instantaneous elastic deformation may occur, followed by viscous, time dependent strain
changes. As a result, many polymeric materials are susceptible to time-dependent
deformation under a constantly maintained stress. Such deformation is called viscoelastic
creep.
4. The mechanical properties are influenced by temperature, molecular weight, additives
(many proprietary), degree of crystallinity, and heat treatment history. Mechanical properties
of certain polymers can change dramatically over narrow temperature range. For example,
PMMA (acrylics) is totally brittle at but extremely ductile at the stress-strain relation and the

78
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

viscoelastic behaviors are both influenced strongly by the temperature.Many organic

polymers are dielectric insulators. However, certain polymers exhibit interesting conducting
behaviors. In recent years, conducting polymer materials are being actively pursued for
making transistors, organic thin film displays, and memory. Such conductive polymers
include polypyrrole, polyaniline, and polyphenylene sulfide, to name a few. Mobility of
charge carriers in polymers is still orders of magnitude lower than that of silicon and
compound semiconductor materials. At this stage, polymer electronics devices are not able to
compete with semiconductor electronics in terms of performance. Polymers can be processed
using a large number of techniques, including injection molding, extrusion, thermoforming,
blow molding, machining, casting, compression molding, rotational molding, powder
metallurgy, sintering, dispersion coating, fluidized-bed coating, electrostatic coating,
calendaring, hot forming, cold forming, vacuum forming, and vapor deposition. Many
techniques can be combined with microfabrication.

POLYMERS IN MEMS
Micromachining technology for MEMS was derived from integrated circuit fabrication.
Naturally, silicon has been the predominant material choice. In recent years, polymers have
emerged as an important new class of materials for use in MEMS applications. There are a
number of unique merits associated with polymer materials. First, the cost of the material is
much lower than that of single-crystal silicon. Second, many polymer materials allow unique
low-cost, batch-style fabrication and packaging techniques such as thermal micromolding,
thermal embossing and injection molding. Instead of processing on one wafer at a time,
polymer substrates can potentially be processed in a high-throughput, roll-to-roll fashion.
Third, certain polymers offer unique electrical, physical and chemical properties that are not
available in silicon and silicon-derived materials. Examples of such properties include
mechanical shock tolerance,
biocompatibility, and biodegradability. There are barriers for using polymers in MEMS. The
viscoelastic behavior of polymers is undesirable in certain applications. Many polymer
materials have lower glass-transition and melting temperatures. The low thermal stability
limits fabrication methods and application potentials. A significant number of polymer
materials have been introduced to MEMS applications in recent years. These materials find
applications beyond being handle wafers or adhesive layers, but rather are used as structural
mechanical elements including cantilevers and membranes.
The list below includes polymers that have been used successfully and widely for MEMS
applications. Some items in the list represent a family of polymers while others represent a
specific product:
1. Polyimide;
2. SU-8;
3. Liquid crystal polymer;
4. Polydimethylsiloxane;
5. Polymethyl methacrylate (also known as acrylics. plexiglass, or PMMA);
6. Parylene (polyparaxylylene);
7. polytetrafluoroethylene (Teflon) and Cytop.

Polyimide
Polyimides represent a family of polymers that exhibit outstanding mechanical, chemical, and
thermal properties as a result of their cyclic chain-bonding structure. Bulk processed
polyimide parts are used widely, from cars (struts and chassis in some cars) to microwave
cookware. It is widely used in microelectronics industry as an insulating material as well.
Polyimides are formed from the dehydrocyclization of polyamic precursors into cyclic

79
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

polymers by incorporating aromatic groups R and R’.These aromatic groups are chosen to
affect the properties of the final polyimide. For example, by chemically altering the polyamic
acid precursor to include R” groups sensitive to UV light, photo-patternable precursors can be
made to crosslink where exposed to UV light. The mechanical properties of polyimide films
have been studied. Cured polyimide films exhibit intrinsic stress on the order of 4x106 Pa to
4x107 Pa as measured using suspended microfabricated polyimide strings. Further,
mechanical and electrical properties of polyimide may exhibit direction-dependent behavior.
Many properties such as index of refraction, dielectric constant,Young’s modulus, thermal
expansion coefficient and thermal conductivity vary with processing conditions.
Polyimide is commercially available as cured sheets, semi-cured sheets, or viscosity solutions
for spin-coating.The structure of a typical commercial polyimide—HN-type Kapton. In
MEMS, polyimide is used as insulating films, substrates, mechanical elements (membranes
and cantilevers), flexible joints and links, adhesive films, sensors, scanning probes, and
stress-relief layers. Polyimide materials offer many favorable characteristics in these roles,
including (1) chemical stability; (2) thermal stability up to around 400 0 C(3) superior
dielectric properties; (4) mechanical robustness and durability; and (5) low cost of materials
and processing equipment. Polyimide can be used as structural elements for sensors and
actuators. Unfortunately, polyimide is neither conductive nor strain sensitive. Functional
materials such as conductors or strain gauges need to be integrated externally. Thin film
metal strain gauges have been integrated with polyimide, exhibiting an effective gauge factor
on the order of 2 to 6. An alternative is to modify the polyimide material for sensing
purposes. For example, piezoresistive composite of polyimide and carbon particles with
effective gauge factor on the order of 2 to 13 has been demonstrated.

SU-8
The SU-8 is a negative tone, near UV photo-resist first invented by IBM in the late 1980s,
with the main purpose of allowing high aspect ratio features (>15) to be made in thick
photosensitive polymers. The photoresist consists of EPON® Resin SU-8 (from Shell
Chemical) as a main component. The EPON resin is dissolved in an organic solvent (GBL,
gammabutyrolacton), with the quantity of solvent determining the viscosity and the range of
achievable thickness. Processed layers as thick as 100 m can be achieved, offering
tremendous new capabilities for masking, molding, and building high aspect ratio structures
at low cost. The cost of SU-8 lithography is considerably lower than that of other techniques
for realizing high aspect ratio microstrucures, notably LIGA process and the deep reactive ion
etching. SU-8 has been integrated in a number of micro devices, including microfluid
devices, SPM probe, and micro needles. It can also serve as a thick sacrificial layer for
surface micromachining.

Liquid Crystal Polymer (LCP)

The liquid crystal polymer is a thermoplast with unique structural and physical properties.
LCP are available commercially in sheets of various thickness.When flowing in the liquid
crystal state during processing, the rigid segments of the molecules align next to one another
in the direction of shear flow. Once this orientation is formed, their direction and structure

80
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

persist, even when LCP is cooled below its melting temperature.This characteristic
differentiate LCP from most thermoplastic polymers (e.g., Kapton), whose molecule chains
are randomly oriented in the solid state. Owning to its unique structure, LCP offers a
combination of electrical, thermal, mechanical and chemical properties unmatched by other
engineering polymers. One of the earliest LCP films used in MEMS is a Vectra® A-950
aromatic liquid crystal polymer, produced by Hoechst Celanese Corporation.The reported
melting temperature of Vectra A-950 is 2800C. The specific gravity ranges from 1.37 to 1.42
kg/m3 and the molecular weight is greater than 20,000 g/mol.The compatibility of this LCP
with commonly used chemicals in micromachining was first investigated in. LCP is virtually
unaffected by most acids, bases and solvents for a considerably long time and over a broad
temperature range. Extensive tests showed that LCP was not attacked or dissolved by at least
the following chemicals common in microfabrication: (1) organic solvents including acetone
and alcohol, (2) metal etchants for Al,Au and Cr, (3) oxide etchants (49% HF and buffered
HF), and (4) developers for common photoresist and SU-8 resist. LCP films have excellent
stability. It has very low moisture absorption and low moisture permeability, which are better
than PMMA and comparable to that of glass. For other gases, including oxygen, carbon
dioxide, nitrogen, argon, hydrogen and helium, LCP also exhibits above-average barrier
performance. Further, the permeation of gases through LCP is not affected by humidity, even
under elevated temperatures.The thermal expansion coefficient of the LCP material can be
controlled during the fabrication process to be both small and predictable.The LCP film also
shows excellent chemical resistance. LCP was originally used as a high-performance
substrate material for high-density printed circuit board (PCB) [23]. A number of unique
processing methods have been developed for LCP, including laser drilling and via filling (for
low-resistance electrical through interconnects). LCP films are used as substrates for space
and military electronics systems, both for its performance and its stability. For instance, it has
been explored as a high-performance carrier of radio frequency electromagnetic elements
such as antennas. Results from high-frequency tests show that LCP has a fairly uniform
relative dielectric constant of 3 in the range 0.5 to 40 GHz and an extremely low loss factor.
For LCP film with uniaxial molecule orientation, its mechanical properties are anisotropic
and dependent on the polymer orientation. For instance, the uniaxial LCP film can withstand
less load in the transverse direction (i.e., the direction orthogonal to the orientation of its
molecules) than in the longitudinal direction (i.e., the direction along the orientation of its
molecules). If necessary, biaxially oriented film with equally good transverse and
longitudinal direction properties can be made to correct the anisotropy behavior.A uniaxial
film can be formed by bonding multiple layers of anisotropic films with angular offsets of
crystal orientations between layers. The orientation of LCP molecules varies through the
thickness of the film, while at the two sides of the film molecules are oriented at opposite
angles. If the angles are +450 and -450 at either side, the mechanical properties, such as
coefficient of thermal expansion, tensile strength and modulus are nearly isotropic.
Commercial LCP material is supplied in sheet format. The thickness of LCP film could vary
from several microns to several millimeters. Some can be provided with copper clad layers
on one or both sides. The optional copper clad layer is normally 15 to 20 m thick. This
copper layer is laminated in a vacuum press at a temperature around the melting point of
LCP.

PDMS
Elastomers are materials that can sustain large degree of deformation and recover their shape
after a deforming force. Poly (dimethylsiloxane) (PDMS), an elastomer material belonging
to the room temperature vulcanized (RTV) silicone elastomer family, offers many advantages
for general MEMS applications. It is optically transparent, electrically insulating,

81
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

mechanically elastic, gas permeable, and biocompatible.The biological and medical

compatibility of the material is reviewed in. PDMS is widely used in microfluidics. The
primary processing method is molding, which is straightforward and allows fast, lowcost
prototyping.A number of unique process characteristics of PDMS are worth noting:
1. The volume of PDMS shrinks during the curing step. Compensation of dimensions at the
design level should be incorporated to yield desired dimensions.
2. Due to volume shrinking and flexibility, deposited metal thin films on cured PDMS tends
to develop cracks, affecting the electrical conductivity.
3. The surface chemical properties (such as adhesion energy) can be varied by altering the
mixing ratio and through surface chemical or electrical treatment. PDMS is commercially
supplied as a viscous liquid—it can be cast or spin coated on substrates. Unfortunately, the
PDMS material is not photo definable. It therefore cannot be simply spin-coated and
patterned like photosensitive resists. Though UV curable PDMS is being developed, the
technology is not yet mature. It is possible to use plasma etching to pattern PDMS thin films.
However, the etch rate is rather slow.The measured etch rate is approximately 7 nm/min at
800 W power and 100 V bias. Etching of PDMS with O2 plasma leaves the surface and line
edges rough.

PMMA
PMMA is supplied in many different forms, including bulk, sheets, and solutions for
spincoating. PMMA bulk, most commonly known by its trade name acrylics, has been used
in making microfluidic devices. The photodefinable PMMA thin film is a widely used e-
beam and X-ray lithography resist. Spin coated PMMA has been used as a sacrificial layer as
well. Deep reactive ion etching processes for PMMA thin films has been demonstrated.

Parylene
Parylene is a thermalset polymer. It is the only plastic material that is deposited using
chemical vapor deposition (CVD) process.The deposition process is conducted under room
temperature. A Parylene deposition system consists of a source chamber connected to a
vacuum deposition chamber.A dimer (di-para-xylene) is heated inside the source chamber to
approximately It sublimates into a gaseous monomer, which then enters the vacuum chamber
and coats on objects within. Three parylene dimer variations are available from commercial
vendors, including Parylene C (widely used), Parylene N (for better dielectric strength and
penetration), and Parylene D (for extended temperature performance). The Parylene film
offers very useful properties for MEMS applications, including very low intrinsic stress,
room temperature deposition, conformal coating, chemical inertness, and etch selectivity.
Parylene coating is deal for electrical isolation, chemical isolation, preservation, and sealing.
Parylene has been used for microfluidics channels, valves, retina prosthesis, sensors
(acceleration sensors, pressure sensors, microphones, and shear stress sensors ). The
thickness of Parylene coating is generally controlled by the amount of dimer loaded.
Thickness monitors and end-of-point detectors for in-situ Parylene thickness monitoring have
been developed, for example based on thermal transfer principles.

Fluorocarbon
Fluoropolymers such as Teflon and Cytop provide excellent chemical inertness, thermal
stability, and nonflammability due to the strong C-F bond.They can be used as a surface
coating, insulation, antireflection coating, or as adhesion agent. Cytop is a trademarked
material (by Asashi Glass Company of Japan). It exhibits many good properties as Teflon but
offers high optical transparency and good solubility in specific fluorinated solvents.
Fluoropolymer films can be spin coated or deposited by PECVD method. In MEMS,Teflon

82
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

and Cytop films have been used for electrical insulation, adhesive bonding, and friction
reduction.

Applications
Acceleration Sensors
Acceleration sensors can be made entirely or partially out of polymer materials using a
variety of transduction principles.These generally involve depositing functional thin films on
polymer substrates or microstructures.

Silicon Accelerometer with Parylene Beams

Here we discuss a microfabricated acceleration sensor using polymer support beams. The
accelerometer incorporates a silicon proof mass and high aspect ratio Parylene beams. The
polymer beam increases the shock resistance, enabling large deformation without failure.
Because Parylene has a small Young’s modulus, the spring constant is low than if they were
replaced by silicon. A low spring constant translates into increased sensitivity but somewhat
reduced resonant frequency. In this design, Parylene beams are 10–40 m wide and have
aspect ratios (height over width) of 10–30. However, it is impractical to grow Parylene films
with thickness of hundreds of micrometers. In addition, there is no high aspect ratio reactive
ion etching process that can produce vertical etching. An alternative process for realizing
high aspect ratio Parylene structures is developed. It involves first creating high aspect ratio
trenches ( 400 m deep) as molds in a 500-mm-thick silicon substrate.The wafer is oxidized
by reacting with oxygen at a high temperature . The oxidized wafer (with conformal –thick
oxide coating) is then placed inside a Parylene deposition chamber. Parylene thin films with
thickness of 10 to 20- fill the trenches entirely. A global plasma etch is performed to remove
the Parylene on the open front surface. Parylene films in the trenches are preserved because
the effective thickness is much greater (Figure 13.4e). The wafer is turned over to pattern a
backside mask layer, which is used to define the wafer with deep reactive ion etching
(DRIE). The DRIE process has very high selectivity between silicon and silicon oxide
(Figure 13.4g) and stops when it reaches the oxide layer. At the end, the oxide is removed by
HF solutions to free the Parylene beams. Since the Parylene film cannot survive overtime
DRIE etching, the oxide layer effectively buffers the Parylene film. Since no active sensing
layers are incorporated, the displacement of the proof mass in response to acceleration is
detected using optical means.The proof mass has an area of 1.75 mm by 1.75 mm. The
resonant frequency was measured to be 37 Hz.

83
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

FIGURE 13.4 Parylene accelerometer.

Pressure Sensors
Parylene Surface Micromachined Pressure Sensor
The basic design of a surface micromachined Parylene membrane with integrated resistors is
shown in Figure 13.5. The membrane, circular as shown, is elevated from the substrate
surface by a distance of 0.5 to 30 m Strain gauge resistors for sensing membrane
displacement are typically placed along the periphery of the membrane, as depicted in Figure
13.5. Metal films can serve as piezoresistors in place of doped polycrystalline silicon.
However, one disadvantage lies in the fact that the resistivity of thin film metal is much
smaller compared with that of polycrystalline silicon. In order to achieve appreciable
magnitude of resistance (e.g., greater than 40 ohm), these resistors are zigzagged, consisting
of alternating radial segments and tangential ones.The redial segments are primarily
responsible for the displacement sensing. When a vertical force or pressure is applied on the
membrane, the membrane will be deformed to induce in-plane stress in the radial direction,
which is sensed by radial segments of the strain gauge resistors.
Major design variables of a membrane device include the diameter and thickness of the
membrane, the height of the underlying cavity, and the resistance of thin film resistors. A
successful design must take into consideration of processing and performance needs
simultaneously. For example, to avoid unwanted membrane collapse and sticking to the
substrate, it is generally desirable for a membrane to be smaller, thicker, and with greater
cavity heights. However, if a membrane is overly small, it may not have enough area to
accommodate appreciable resistance from embedded metal resistors.
Increasing the cavity height (and membrane clearance) will generally cause difficulties with
electrical continuity from the top of the membrane to the substrate level; there are also
practical difficulties with building thick sacrificial layers (e.g., greater than 20m). The
overall process, diagramed in Figure and Figure 13.7, can be achieved under relatively low
overall temperature (i.e., less than 1200 C). As a result, the process can be realized on a
variety of substrate materials, including silicon, glass, and even polymers. A layer of
photoresist is spin coated on the front surface of a substrate and patterned

84
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

photolithographically (Figure a). The spin-on photo resist is cured in a convection oven, first
at 600 for 5 min (to remove edge beads) and then at 1100 for 1 min.The patterned photoresist
will reflow slightly during post-development bake (1100 C for 2 min), rounding the edges of
features to create a sloped edge. Optionally, the photoresist can be selectively thinned (to a
target height of 2.5m) near etch/sealing holes (Figure b).The authors achieved this by
additional exposure near the etch hole regions using a separate mask.This reduces the amount
of Parylene needed to seal the cavity in optional step m. Major reasons for selecting
photoresist as the sacrificial material (as opposed to metal or silicon dioxide) include: (1) the
thickness of the sacrificial layer can reach 10–20m range relatively easily and quickly; (2)
the edge of photoresist sacrificial layer can be smoothed to realize gentle slopes. A 1 m
thick Parylene thin film is then deposited on top of the wafer surface (Figure c).The Parylene
is subsequently coated with a 150-nm-thick Al thin film, which is then patterned
photolithographically (Figure d). Oxygen plasma etch is used, with the thin film Al as the
mask, to remove the exposed Parylene (Figure e) and to reach underlying hard substrate. An
oxygen plasma etch step creates a rather steep transition between the substrate and the top
Parylene surface. This will pose problems of electrical continuity in the future when thin film
metal wiring traverses between the substrate and the Parylene top surface.
Therefore, after removing the metal etch mask, the authors made efforts to smooth the edge
of the Parylene using a process described below. A layer of photoresist is spin coated (2500
rpm for 10 sec) following the removal of the thin film metal.The top profile of the photo
resist after curing and reflow (5 min at 640C followed by 1 min at 1100C ) is much smoother
than the slope in Parylene (Figure f).A global oxygen plasma etch is performed to etch the
photoresist (Figure g). Since the oxygen plasma etch rate on photoresist and on Parylene is
roughly identical (at 300 mTorr of pressure and 350 W of power), the smooth, reflown edge
of the photoresist is transferred onto the Parylene after the photoresist on planar surfaces is
removed.

85
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Thin film metal resistors embedded in the polymer membrane must be located near the
surface, off the neutral axis of the membrane. After the edge profile is adjusted, a layer of
200-nm-thick Au (with a 5-nm-thick Cr thin film underneath for adhesion) is deposited and
patterned on top of the Parylene thin film (Figure h). Gold offers better conformal coverage
than others (e.g., Ni, NiCr, and Al), allowing electrical continuity across edges of membrane.

86
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The underlying Parylene shields the metal resistor from direct contact—and shorting, with
the bottom of the cavity if the membrane is fully displaced downward. Another layer of
Parylene encapsulates the metal resistor, preventing delamination and accidental electric
shorting. The thickness of short tangential segments are made thicker by depositing an
additional layer of metal using the lift-off process (Figure i). The entire device is then coated
with another 8 m thick (nominal) Parylene layer (Figure j). The Parylene is patterned using
a thin film metal (300-nm-thick Al) as a mask in oxygen plasma etching.A global plasma etch
is performed to pattern the newly deposited Parylene film. This exposes end regions of etch
hole and also reopens the contact pads (Figure k).The photoresist sacrificial layer inside the
cavity is removed using acetone (Figure l). This is conducted under room temperature for 3 h
(for a 400 m diameter membrane) or more (for larger membranes).The wafer is dried under
an infrared lamp for 10 min. It is impractical to remove the wet chemicals from the cavity
using spin-drying because the membranes are found to collapse and adhere to the substrates.
Optionally, the authors reported hermetically sealing the cavity by depositing another thin
film of Parylene (approximate thickness being 2m ). At the opening of each etch hole, two
fronts of Parylene will grow from opposite surfaces and eventually meet to seal the cavity
interior (Figure m).The pressure inside the cavity is kept at the processing pressure of
Parylene deposition (approximately 40 mTorr). The Parylene film is also deposited on the
bonding pad region, unfortunately. The authors performed another sequence of masking (with
metal film) and etching (using oxygen plasma) to reopen the bonding pads regions (Figure n).
An optical micrograph of a fabricated device is shown in Figure below.The device has
successfully performed for measuring contact pressure. Further characterization and
improvement of sensitivity is needed in order to demonstrate measurement of air or liquid
pressure changes.

LCP Piezoresistive Flow Sensor

A flow sensor consisting of a polymer cantilever beam has been made. As shown in Figure
below, flow imparts momentum on the cantilever and causes it to bend, inducing strain at the
base of the cantilever.The strain is transduced into an electrical signal using a piezoresistive
sensor made of thin film metal.While the gauge factor of doped silicon can reach 10–20, the
gauge factor of thin film metal is much lower, typically ranging from 1 to 5. However, the
increased thickness and compliance of these polymer film devices has been shown to offset
the reduced gauge factor of metal film strain gauges resulting in sensitivity comparable to

87
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

silicon-based devices. The flow sensor uses nickel-chrome (NiCr) strain gauges on an LCP
cantilever that is 1000m wide and 3000 m long. Figure a show a micrograph of the
completed device. Wind tunnel testing with flow rates from 0 to 20 m/s showed a velocity-
squared relationship as expected, as seen in the quadratic trend line in Figure below.

OPTICAL MEMS

Micro-optical systems presently have application in the following areas: Optical Scanning -
Medicine (minimally invasive surgery) - Small Space Inspection Optical Alignment Optical
Communication - Switching - Optical Filtering - Spectrometry (Wavelength Separation) -
Adaptive Optics and Free Space Communication.

Advantages:

Easy to manipulate light - Optical systems involve photon re-direction, therefore, since
photons have very little momentum, micro-actuators are easily able to manipulate photons.
Simplified packaging - Optical MEMS can be sealed in packages with transparent housings,
which allows light to pass through a glass window. - This allows the optical device to be

88
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

protected from dust and harsh environmental conditions, unlike flow sensors, or micro tactile
sensors, or other MEMS where the source of signal/phenomena must be in direct contact with
the chip surface. Device size - Able to pack a large number of devices into a relatively small
area/volume.

Application – DLP digital light processing

The DLP micro-optical projection system was developed by Texas Instruments. It uses a
matrix of micro-mirrors to selectively switch reflected light, to form a projected matrix of
pixels. Each mirror is 11 x 11 um in size, and a matrix may consist of 1024 x 768 mirrors, or
more. By selectively switching the mirrors on and off in the presence of a light source, the
mirrors will reflect the light source ‘toward’ or ‘away’ the optical path that forms the
projected image. To create the complete projection system, the DMD (Digital Micromirror
Device) is used with a light source, optics, a color filter and a projection lens.

89
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

90
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

MICRO LENSES
Lenses refract light for various purposes in optical systems. An ideal lens shape is shown in
the Figure below in (a). The ideal lens is difficult to fabricate directly using conventional
micromachining, so a number of variations are shown below, (b-e):

BINARY LENSES
The principle of operation of a Binary lens is as follows: Light passing through the ‘bulk
material’ essentially moves in a straight line, therefore, if the ‘bulk’ is removed, the operation
is essentially the same.

91
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Variable-focus micro-lens using an ‘expandable’ polymer membrane. As the ‘chamber’ is

pressurized, it fills with fluid and expands the membrane. The different expansion volumes
have different radius of curvature, and therefore different optical NA (numerical aperture).

Micro-Mirrors
Micro-Mirrors are a critical part of micro-optical systems, as they facilitate the re-direction of
light beams. A ‘Good’ micro-mirror has a number of important characteristics, such as: -
High reflectivity (low loss of light signal) - High flatness (low loss of light signal) - High
stiffness (stays flat during acceleration and deceleration) - Low mass (allows rapid
acceleration and deceleration) Micro-mirrors are used in: - DLP systems (Digital Light
Projection systems) - Optical Switching (fiber optic communication systems) - Scanning
(minimally invasive tools with optical scan tips).

Micro-actuators for Micro-Mirror Systems

Electrostatic Based Micro-Mirror Actuators:

92
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

A MEMS device having a deformable mirror. In a representative embodiment, the MEMS

device includes a deformable membrane supporting a plurality of light-reflecting segments
that form the deformable mirror. One or more actuators, at least one of which is configured to
apply torque to a side of the membrane, are used to deform the membrane. Membrane
deformation causes the segments to change orientation and thereby change the shape of the
minor. A representative MEMS device of the invention enables segment displacements in
two directions and thereby realizes effective mirror curvature values in the range from about
−2 mm−1 to about +2 mm−1. A typical deformable mirror of the prior art is a deformable
membrane, the shape of which is controlled by an array of underlying electrostatic actuators,
e.g., electrodes. However, the deformation induced by such actuators is typically unipolar.
That is, the electrodes can flex the membrane in one direction (i.e., toward these electrodes)
only. As a result, the types of distortions correctable with prior art membrane mirrors are
limited. In addition, for a membrane mirror having a lateral dimension of about 100 μm, the
maximum deformation amplitude is on the order of 1 μm, which puts a corresponding limit
on the range of distortions that can be corrected. Although fabrication of MEMS devices of
the invention has been described in the context of using silicon/silicon oxide SOI wafers,
other suitable materials, such as germanium-compensated silicon, may similarly be used. The
materials may be appropriately doped as known in the art. Various surfaces may be modified,
e.g., by metal deposition for enhanced reflectivity and/or electrical conductivity or by ion
implantation for enhanced mechanical strength. Differently shaped membranes, segments,
actuators, rods, members, and/or electrodes may be implemented without departing from the
scope and principle of the invention. Springs may have different shapes and sizes, where the
term “spring” refers in general to any suitable elastic structure that can recover its original
shape after being distorted. A different number and/or type of actuators may be used to
deform membranes and/or springs. Various MEMS devices of the invention may be arrayed
as necessary and/or apparent to a person skilled in the art.

93
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

UNIT I
INTRODUCTION
PART – A
1. Write the relation between resonant frequency and Q-factor?
At or near resonance frequency the mechanical vibration amplitude sharplyincreases. The
sharpness of resonant peak is characterized by a term called qualityfactor (Q). The sharper
the resonance peak, the higher the quality factor.
2. Name any two materials used in the fabrication of MEMS.
i) Silicon ii) Polymers
3. What do you mean by in situ doping?
The intentional introduction of impurities, called doping, would turn as intrinsic material
into an extrinsic semiconductor material. Impurities can be introduced into number of ways,
most notably through diffusion and ion implantation. They can also be incorporated into
semiconductor lattice during the growth of material as well. This process is called in situ
doping.
4. Define mean free path.
The sheet resistivity of an emitter layer is typically measured with a four-point-probe.
5. Define elastic modulus.
An elastic modulus, or modulus of elasticity, is the mathematical description of an object or
substance's tendency to be deformed elastically (i.e., non-permanently) when a force is
applied to it.
6. Define beam.
A beam is structure member subjected to lateral roads, that is, forces or moments having
their vectors perpendicular to the longitudinal axis.
7. What do you mean by cantilever?
A beam fixed at one end and free at another is conveniently referred to as fixed free beam,
commonly called a cantilever.
8. Name some commonly encountered beam structures in MEMS.
Double - clamped suspension structures and single-clamped cantilevers.
9. What are the three preparatory steps required to calculate the curvature of the beam.
Three preparatory steps:
1. Find the moment of inertia with respect to the neutral axis
2. Find the state of force and torque along the length of beam
3. Identify boundary conditions. Two boundary conditions are necessary to
deterministically find a solution.
10. Why intrinsic stress is important for MEMS devices.
Intrinsic stress is important for MEMS devices because it can cause deformation damages in
excessive cases, affect surface planarity, or change the stiffness of mechanical element.
11. Name the material having zero stress.
Single crystal bulk silicon.
12. What are the strategies for minimizing undesirable intrinsic bending?
The strategies are:
a. Use materials that inherently have zero or very low intrinsic stress.
b. For materials whose intrinsic stress depends on material processing parameter,
fine tune the stress by calibrating and controlling deposition conditions.
c. Use multiple layered structures to compensate stress-induced bending.
13. What do you mean by quality factor (Q)?
The sharpness of resonant peak is characterized by a term called the quality factor. The
sharper the resonance peak, the higher the quality factor.
14. Define stress and explain tensile and compressive stress.
Stress is defined as the external force applied per unit area. When body is subjected to equal
and opposite pulls, it elongates. The resistance offered to this elongation is tensile stress. If
body is subjected to equal and opposite pushes, it contracts and the resistance offered to this
contraction is compressive stress.

94
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

15. Define strain and explain tensile and compressive strain.

It is the ratio of change in dimension to the original dimension. The strain corresponding to
the tensile stress is called tensile strain and it is defined as the ratio of the increase in the
length to the original length. The strain corresponding to the compressive stress is called as
compressive strain. It is defined as the ratio of decrease in length to the original length.
16. Define shear stress and shear strain.
The two equal and opposite force act tangentially on any cross sectional plane of the body
tending to slide one part of the body over the other part. The stress induced is called shear
stress and the corresponding strain is known as shear strain. Or Shear stress is the resistance
offered to the force applied. It is the ratio of shear force to shear area. The corresponding
strain is called as shear strain.
17. State Hook’s law and explain elastic limit.
It states that within elastic limit, the ratio of the stress and strain is a constant.Elastic limit is
a point on the stress strain diagram below which the body regains its original shape when
deformed, not necessarily obeying the hook’s law.
18. What you mean by thermal stresses?
If the body is allowed to expand or contract freely, with the rise or fall of temperature no
stress is developed but if free expansion is prevented the stress developed is called
temperature or thermal stress and the corresponding strain is temperature strain.
Thermal stress = α. ΔT. E
Thermal strain = α. ΔT
A= the coefficient of thermal expansion of the material,
ΔT = change in temperature
19. What is bending moment?
The algebraic sum of the bending moment due to all the individual forces to the right or to
the right of the section.
20. What are the types of the beam?
The different types of the beam are
a. Fixed beam b. Cantilever beam
c. Simply supported beam d. Over hanging beam
e. Continuous beam
21. What is neutral axis?
As a result of bending moment, a length of the beam will take up a curved shape. It follows
that outer radii of the material will be in tension and at the inner radii in compression, and at
some radius there will be no stress. This layer of the material is neutral axis.
22. What are the assumptions made in the theory of torsion?
 The material of the shaft is uniform throughout.
 The twist along the shaft is uniform.
 Normal cross sections of the shaft which were plane and circular before twist must
remain plane and circular after twist.
 The diameter of the cross sections which were straight before twist must remain
straight without any change in the magnitude.
23. Define Torsion.
When equal and opposite torque are applied at ends, the shaft is said to be in torsion.
24. Define Torsional rigidity.
We know, θ = T ℓ / GJ,
The quantity GJ is called torsional rigidity. Product of rigidity modulus and polar moment
of inertia is called torsional rigidity.
25. Define Stiffness of spring.
Stiffness of spring (K) is the force per unit deflection. Its unit is N / mm in SI units.
26. What are the assumptions of simple bending theory?
Assumptions made in the simple bending theory are
 The material of the beam is perfectly homogeneous throughout.
 The stress induced is directly proportional to strain.

95
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

 The value of modulus of elasticity is the same, for the fibres of the beam under
compression or tension.
 The transverse of the beam, remains plane before and after bending.
 There is no resultant pull or push on the cross section of the beam.
 The loads are applied in the plane of bending.
27. What is mean free path.

Mean free path is the average distance travelled by a moving particle (such as an atom,
a molecule, a photon) between successive impacts (collisions), which modify its direction
or energy or other particle properties.
28. Define sheet resistivity.
For a uniformly doped layer, the sheet resistivity is defined as:

o 
t
The sheet resistivity is normally expressed as ohms/square or Ω/□
For non-uniformly doped n-type layers, ie., if ρ is non-uniform:
1
 t
1
0 (x)dx
The sheet resistivity of an emitter layer is typically measured with a four-point-probe.
29. List some military application of MEMS. (Apr 2017)
Optical Line of Sight Communication
RF communication
Power Generation, Storage, and Scavenging
Micro air vehicles
30. What is Micromachining? (Apr 2017)
The technique for fabrication of 3D and 2D structures on the micrometer scale.
Superfinishing, a metalworking process for producing very fine surface finishes.
Various microelectromechanical systems
Bulk micromachining
Surface micromachining
High-aspect-ratio microstructure technologies

31. What is drive in in the process of doping? (Dec 2017)

Deposition – dopant atoms are introduced at the wafer surface.
Drive-in – the dopant atoms then diffuse into the wafer to create the required concentration
gradient.
32. List the Intrinsic characteristics of MEMS(MAY 2018)
Miniaturization: dimensions of MEMS structures are much larger than in VLSI ICs (µm).
Further scaling leads to NEMS (nano) that are comparable/smaller than ICs (1-100 nm).
Scaling laws describe how properties/behavior change with dimensions
Scaling Law of Area-to-Volume Ratio
Microelectronics Integration - the most widely used is that with CMOS
33. What are the common methods of IC and MEMS Fabrication(MAY 2018)
Bulk Micromachining
Surface micromachining
Wafer Bonding
Deep reactive ion Etching of silicon
PART B
1 With neat sketch explain about the photolithography process for micro-machined pressure
sensor.

96
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Photolithography
The goal of photolithography is to produce fine features on wafer surfaces.A most common
lithography process involves depositing photo-sensitive chemicals (called photoresists, or simply
resists) on a silicon wafer, exposing it with light through a mask, and removing (develop)
photoresist material that has been modified by light.The starting point of a lithography process is
to coat a wafer with photoresist through spin coating (Figure ). A wafer is held on a rotating
stage. Photoresist is applied to the center of the wafer at rest position.The wafer is then spun at
high speed, causing the photoresist to move towards the edge of the wafer under centrifugal
forces. After the wafer spinning is stopped, a uniform thin layer of photoresist is coated on the
front surface of a wafer. Process variables include the wafer spinning speed, the viscosity of the
resist, and the types of resists (e.g., target wavelength, sensitivity).Typical thickness of
photoresist is generally 1–10 mm. A lithography patterning procedure involves multiple steps
(Figure ). A wafer is first covered with a uniform thin layer of resist (step a).A mask, consisting
of a transparent substrate (e.g., glass or quartz) with opaque features, are brought close to the
resist-coated wafer (step b). High energy, collimated light rays strikes the mask-wafer assembly.
Resist regions that are not covered by opaque features are exposed, changing the chemical
composition of the resist. For positive resist, the exposure by light causes the resist to be more
soluble in a wet chemical developer (step c). This allows the opaque features on the mask to be
faithfully transferred to the wafer (step d). A pattern in photoresist can be further transferred to
an underlying layer, using the photoresist as a mask layer.The process is shown in Figure . It
starts with a wafer with a thin film coating (step a).The wafer is covered with a spin-coated layer
of photoresist (step b), which is photolithographically patterned similar to the method discussed
in Figure .The wafer is then immersed in a chemical solution that preferably etch the thin film
but not the photoresist (step c). (Alternatively, the thin film may be etched with dry etch
methods.) With proper timing control, the thin film covered by the photoresist would stay intact,
whereas the thin film in areas not covered by the photoresist would be removed (step d). The
photoresist is then preferentially removed, leaving the thin film behind as patterned (step e).

Thin Film Deposition

Functional materials, conductors and insulators can be incorporated on a wafer through additive
deposition process. One such deposition process is direct transfer of the material from a source to
the wafer surface in an atom-by-atom, layer-by-layer fashion (Figure ). Examples include metal
evaporation and metal sputtering.The process is generally conducted in a lowpressure
environment so that atoms may travel from the source to the wafer surface without interruptions
caused by air molecules. One such system, an evaporator, is diagrammed in . A wafer and a
metal source are both placed inside a vacuum jar.The metal can be transferred either by heating it
(evaporation) or by bombarding it with high-energy ions (sputtering). The achieved thickness is
proportional to the power and time. In practice, the routine thickness of metal thin films ranges
from 1 nm to 2 mm. A second method for placing thin film materials on a wafer surface is
chemical vapor deposition (Figure 1). Two or more active species arrive at the vicinity of the
wafer surface (step a). They react under favorable conditions (with energy provided by heating or

97
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

plasma). The reaction of these species produces a solid phase, which is absorbed onto the nearby
wafer surface (step b). The byproducts of the reaction (if any) may be removed by the
surrounding media. Continuous reaction causes a layer of material to be built on the wafer
surface (step c). Typically the average thickness of thin film deposited by CVD, evaporation or
sputtering is below 1 mm.To deposit films of greater thickness is typically too time consuming or
impractical.

2 Discuss in detail about the calculation of charge carrier concentration.

The electrical conductivity of a semiconductor material is determined by the number of free
charge particles in a bulk and their agility under the influence of electric field. In this section, we
will discuss basic formulation for determining the number (or volumetric concentration) of free
charge carriers. There are two types of free charge carriers—electrons and holes. Electrons that
are freed by breaking covalent bonds are not the only participants in bulk current conduction.
The bond vacancy left by an escaped electron, called a hole, can facilitate bulk current
movement as a site for successive electron hopping.A hole carries a positive charge. In a given
material, the abilities of the electrons and holes to conduct electricity are different, but they are
often in the same order of magnitude.
The electron concentration in a semiconductor is denoted as n, while the hole concentration is
denoted as p. The concentrations of electrons and holes under steady state, thermal equilibrium
conditions (i.e., no external current and no ambient light) are commonly referred as no and po
respectively. The subscript 0 indicates values under thermal equilibrium.
A perfect semiconductor crystal has no impurities and lattice defects in its crystal structure. It is

98
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

called an intrinsic semiconductor material. In this type of material, electrons and holes are
created through thermal or optical excitation.When a valence electron receives enough energy,
either thermally or optically, it is freed from the silicon atom and leaves a hole behind. This
event is called electron–hole pair generation.
The number and concentrations of electrons and holes, however, do not increase without limit
over time. Free electrons and holes can recombine and give up energy along the way.This
process, called recombination, competes with the generation process. Under steady state
conditions, the generation and recombination rates are identical. Since the electrons and holes are
created in pairs, the electron concentration (n) is equal to the concentration of holes (p). For
intrinsic materials, their common value is further denoted as ni with the subscript istanding for
the word intrinsic.The relation is summarized as n=p=ni.

Conductivity and Resistivity

The conductivity of a bulk semiconductor is a measure of its ability to conduct electric current.
In this section, I will discuss formula for calculating electrical conductivity when carrier
concentrations of both types ( electrons and holes) are known. The overall conductivity of a
semiconductor is the sum of conductivities contributed by these two types individually. Free
charge carriers move under the influence of an electric field. This mode of carrier transport is
called drift. The agility of charge carriers drifting under the influence of a given field affects the
conductivity of the bulk. How fast can a free charge carrier move then? A free charge carrier in a
crystal lattice does not reach arbitrarily high speed over time when it is placed in an uniform and
constant electric field. Rather, it is frequently slowed or halted when it collides with lattice atoms
and other free charge carriers.A charge carrier reaches a statistical average velocity ( ) between
collision events.The magnitude of the velocity is the mathematical product of the magnitude of
the local electric field (E) and the agility of the carrier, usually represented by a term called the
carrier mobility µ, The mobility is defined as

The values of the mobility are influenced by the doping concentration, temperature, and crystal-
orientation, in a complex manner. Certain materials (such as GaAs) offer higher electron and
hole mobilities than silicon and are used in high-speed electronic circuits. The statistical average
distance a charge carrier travels between two successive collision events is called its mean free
path, .The average time between two successive collision events is the mean free time, .The
average velocity, mean free path and mean free time are linked by the relation

Armed with knowledge about carrier concentration and their speed, we are set to derive the
expression for the conductivity of a bulk semiconductor. We start from the familiar Ohm’s law,
which states that the bulk resistivity ( ) associated with a material is the proportionality constant
between an applied electric field and the resultant current density J,

The current density equals current divided by the cross-sectional area.The conductivity is the
reciprocal of the resistivity The relationship between the current density and the applied electric
field can be rewritten as

The conductivity ơ is explicitly defined as

3 Derive for conductivity and resistivity. Explain about this in detail.

The conductivity of a bulk semiconductor is a measure of its ability to conduct electric current.
In this section, I will discuss formula for calculating electrical conductivity when carrier
concentrations of both types ( electrons and holes) are known. The overall conductivity of a

99
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

semiconductor is the sum of conductivities contributed by these two types individually. Free
charge carriers move under the influence of an electric field. This mode of carrier transport is
called drift. The agility of charge carriers drifting under the influence of a given field affects the
conductivity of the bulk. How fast can a free charge carrier move then? A free charge carrier in a
crystal lattice does not reach arbitrarily high speed over time when it is placed in an uniform and
constant electric field. Rather, it is frequently slowed or halted when it collides with lattice atoms
and other free charge carriers.A charge carrier reaches a statistical average velocity ( ) between
collision events.The magnitude of the velocity is the mathematical product of the magnitude of
the local electric field (E) and the agility of the carrier, usually represented by a term called the
carrier mobility µ, The mobility is defined as

The values of the mobility are influenced by the doping concentration, temperature, and crystal-
orientation, in a complex manner. Certain materials (such as GaAs) offer higher electron and
hole mobilities than silicon and are used in high-speed electronic circuits. The statistical average
distance a charge carrier travels between two successive collision events is called its mean free
path, .The average time between two successive collision events is the mean free time, .The
average velocity, mean free path and mean free time are linked by the relation

Armed with knowledge about carrier concentration and their speed, we are set to derive the
expression for the conductivity of a bulk semiconductor. We start from the familiar Ohm’s law,
which states that the bulk resistivity ( ) associated with a material is the proportionality constant
between an applied electric field and the resultant current density J,

The current density equals current divided by the cross-sectional area.The conductivity is the
reciprocal of the resistivity The relationship between the current density and the applied electric
field can be rewritten as

The conductivity ơ is explicitly defined as

4 Discuss about crystal planes and orientations.

Silicon atoms in a crystal lattice are regularly arranged in a lattice structure. Materials properties
(such as the Young’s modulus of elasticity, mobility, and piezoresistivity) and chemical etch
rates of silicon bulk often exhibit orientation dependency.The cross-sectional views of the silicon
crystal lattice from several distinct orientations are shown in .3. Obviously, the atom packing
density is different according to different planes, giving rise to crystal anisotropy of electrical
and mechanical properties and etching characteristics.
A set of common notations, called Miller Indexes, has been developed for identifying and
visualizing planes and directions in a crystal lattice. Let us first discuss the procedure for naming
planes.A crystal plane may be defined by considering how the plane intersects the main
crystallographic axes of the solid. Silicon lattice belongs to the cubic lattice family. In a cubic
lattice, materials properties exhibit rotational symmetry. Hence (010) and (001) planes in the
lattice (.6) are equivalent to (100) plane in terms of material properties. To represent a family of
such equivalent planes, a set of integers are enclosed in braces { } instead of parentheses ( ). For
example, crystal planes (100), (010), and (001) are said to belong to the same {100} family (.6).
The Miller Index is also used to denote directions in a crystal lattice.The Miller Index of the
direction of a vector consists of a set of three integers as well, determined by following a two-
step process outlined below:
Step 1. In many cases a vector of interest does not intercept with the origin of the coordinate
system. If this is the case, find a parallel vector that start at the origin. The Miller Indexes of
parallel vectors are identical.

100
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Step 2. The three coordinate components of the vector intercepting with the origin are reduced to
the smallest set of integers while retaining the relationship among them. For example, the body
diagonal of the cube in .4 extending from point (0,0,0) to point (1a, 1a, 1a) consists of three X,
Y, and Z components, all being 1a. Therefore, the Miller Index of the body diagonal consists of
three integers, 1, 1, and 1, enclosed in a bracket [111]. In a cubic lattice a vector with Miller
Index [hkl] is always perpendicular to a plane (hkl). Vectors that are rotationally symmetric
belong to a family of vectors. The notation of a family of vectors consists of three integers
enclosed in <>.

101
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

5 In general discuss about intrinsic stress in MEMS.

When thin film material are under room temperature and zero external loading conditions, this
phenomenon is called intrinsic stress. It can cause deformation damages in excessive cases,
affect surface planarity or change the stiffness of a mechanical element and it may affect the
mechanical behaviour of membranes as well. The flatness of a membrane is guaranteed when the
membrane material is under tensile stress. Excessive tensile stress in a clamped membrane can
cause the membrane to fracture. On the other hand, a film would buckle if compressive stress
where present it can also result from micro structure of the deposited film. The incorporation of
oxygen atoms into silicon lattice during thermal oxidation process. Apart from this other
mechanisms possible are phase change of material and incorporation of impurity atoms. In
certain cases bending is desired and intentional

6 In brief explain about dynamic system, resonant frequency, and quality factor.

QUALITY FACTOR

In physics and engineering the quality factor or Q factor is a dimensionless parameter that
describes how underdamped an oscillator or resonator is, and characterizes a resonator's
bandwidth relative to its center frequency. Higher Q indicates a lower rate of energy loss relative
to the stored energy of the resonator; the oscillations die out more slowly. A pendulum
suspended from a high-quality bearing, oscillating in air, has a high Q, while a pendulum
immersed in oil has a low one. Resonators with high quality factors have low damping, so that
they ring or vibrate longer. factor is a parameter that describes the resonance behavior of an
underdamped harmonic oscillator (resonator). Sinusoidally driven resonators having higher Q

102
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

factors resonate with greater amplitudes (at the resonant frequency) but have a smaller range of
frequencies around that frequency for which they resonate; the range of frequencies for which the
oscillator resonates is called the bandwidth. Thus, a high-Qtuned circuit in a radio receiver would
be more difficult to tune, but would have more selectivity; it would do a better job of filtering out
signals from other stations that lie nearby on the spectrum. High-Q oscillators oscillate with a
smaller range of frequencies and are more stable.

RESONANT FREQUENCY

Resonance is a phenomenon in which a vibrating system or external force drives another system
to oscillate with greater amplitude at specific frequencies.
Frequencies at which the response amplitude is a relative maximum are known as the system's
resonant frequencies or resonance frequencies. At resonant frequencies, small periodic driving
forces have the ability to produce large amplitude oscillations, due to the storage of vibrational
energy.

DYNAMIC SYSTEMS

In physics, a dynamical system is described as a "particle or ensemble of particles whose state

varies over time and thus obeys differential equations involving time derivatives.” In order to
make a prediction about the system’s future behavior, an analytical solution of such equations or
their integration over time through computer simulation is realized.
The study of dynamical systems is the focus of dynamical systems theory, which has applications
to a wide variety of fields such as mathematics, physics, biology, chemistry, engineering,
economics, and medicine. Dynamical systems are a fundamental part of chaos theory, logistic
map dynamics, bifurcation theory, the self-assembly process, and the edge of chaos concept.

7 i)Illustrate the piezoelectric effect on a crystalline plate. (Apr 2017)

Let us first examine the simplest case—the profile of an etched pit in a <100> oriented silicon
substrate when the mask contains a rectangle or a square open window, with edges aligned to the
<110> direction (Figure 10.4a).

103
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

When a silicon wafer is immersed in a wet silicon etching solution, the atomic layer that is
exposed through the open window is etched first. Wet anisotropic etchants exhibit drastically
different etch rates along different crystal planes. In general, the etch rate along direction <111>
is the slowest among all crystal orientations. The reason for crystalline-dependant etch rate is not
yet fully understood though conjectures exist. It is believed that, since wet anisotropic etchants
remove silicon by first oxidizing a silicon surface (using the oxidizing constituents within the
solution) and then removing the oxide (using oxide-etching constituents), the difference of etch
rate is attributed to the difference of oxidation rates. However, there seems no reported direct
evidence correlating experimentally data and reaction rate analysis. Microscopically, the etch
rate difference is attributed to atomic bonding energy of silicon atoms on various surfaces.
Atoms on different crystalline surfaces are associated with different number of neighbors,
bonding energy, and degrees of difficulty for removal.

In the case of a square window, the cavity would end in a perfect point. In the case of a
rectangular opening, a cavity would end in a knife-edge at the bottom. In practice, due to
nonideality of masks, a perfect point is almost never encountered.

8 i)Explain Czochralski Growth Process in single crystal substrates. (Dec 2017)

The Czochralski process is a method of crystal growth used to obtain single crystals of
semiconductors (e.g. silicon, germanium and gallium arsenide), metals (e.g. palladium, platinum,
silver, gold), salts and synthetic gemstones. The process is named after Polish scientist Jan
Czochralski,[1] who invented the method in 1915 while investigating the crystallization rates of
metals.[2] He made this discovery by accident, while studying the crystallization rate of metals
when, instead of dipping his pen into the ink, he did so in molten tin and drew a tin filament, that
later proved to be a single crystal.[3]

104
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The most important application may be the growth of large cylindrical ingots, or boules, of single
crystal silicon used in the electronics industry to make semiconductor devices like integrated
circuits. Other semiconductors, such as gallium arsenide, can also be grown by this method,
although lower defect densities in this case can be obtained using variants of the Bridgman-
Stockbarger technique.

A phase change from solid, liquid, or gas phases or crystalline solid phase occurs in growing
crystals. Czochralski growth is the process used to grow most of the crystals from which silicon
wafers are produced. The silicon crystal growth is a liquid-solid mono component growth
system. The growth of a Czochralski (CZ) crystal, to be discussed in the next section, involves
the solidification of atoms from a liquid phase at an interface. The speed of the growth is
determined by the number of sites on the face of the crystal and the specifics of theheat transfer
at the interface. Some of the most important growth parameters of the crystal are the growth rate
or growth velocity of the crystal, and the instantaneous solidification rate. We should note that
pull rate is the macroscopic indication of net solidification rate. The temperature fluctuations
cause a huge difference in the rates near the interface. The growth rate can be greater than or
even less than the pull rate at a given time. When the growth rate is lesser than the pull rate, re-
melting occurs. Through re-melting, the crystal dissolves back into the melt. This problem has to
be solved so as to remove the crystal defects.
9Derive torsional moment of inertia of cylinder.
TORSIONAL DEFLECTIONS
Whereas beams, especially cantilevers, are frequently used in MEMS for producing linear
displacements and small angular rotations, torsional beams are often used to create large angular
displacement.A case in mind is the digital micro mirror device.The mirror plate is supported by
torsional bars to facilitate rotation. We begin our discussion of torsion by considering a prismatic
bar of circular cross section twisted by torques T acting at the ends. Since every cross section of
the bar is identical, and since every cross section is subjected to the same internal torque T, the
bar is said to be in pure torsion. It may be proved that cross sections of the bar do not change
shape as they rotate about the longitudinal axis. In other words, all cross sections remain plane
and circular and all radii remain straight. Furthermore, if the angle of rotation between one end of
the bar and the other is small, neither the length of the bar nor its radius will change. To aid in
visualizing the deformation of the bar, imagine that the left-hand end of the bar is fixed in
position. Then, under the action of a torque T, the right-hand end will rotate through a small
angle Ǿ known as the angle of twist.
Because of this rotation, a straight longitudinal line on the surface of the bar will become a
helical curve. The angle of twist changes along the axis of the bar, and at intermediate cross
sections it will have a value Ǿ(x) that is between zero at the left-hand end Ǿ and at the righthand
end. The angle Ǿ(x) will vary linearly between the ends. Point a on the right-hand side will
move by a distance d to a new location a’. The torsion induces shear stress throughout the bar.
Stress distribution shows radial symmetry. The magnitude of the shear stress is zero in the center

105
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

of the cross section and reaches maximum at the outer surface of the bar. The maximum stress in
the bar is denoted The expression of the maximum shear strain is

Further, the magnitude of the stress is linearly proportional to the radial distance to the center.
The distribution of shear stress along the radial direction is superimposed on the crosssectional
view of the bar in .22.

The relationship between a torque and the maximum shear stress is found by torque balance at
any given section.

10 i)Write in detail about the flexural beam bending analysis. (Apr/Dec 2017)
Flexural beams are commonly encountered in MEMS as spring support elements. Essential skills
and common practices for a MEMS researcher include calculating the bending of a beam under
simple loading conditions, analyzing induced internal stress, and determining the resonant
frequency associated with the element.
Types of Beams
A beam is a structure member subjected to lateral loads, that is, forces or moments having their
vectors perpendicular to the longitudinal axis. In this textbook, we focus on planar structures—
beams that lie in single planes. In addition, all loads act in that same plane and all deflections
occur in that plane. Beams are usually described by the manner in which they are supported.
Boundary conditions pertain to the deflections and slopes at the supports of a beam. Consider a
two-dimensional beam with movement confined in one plane. Each point along the length of the
beam can have a maximum of two linear degrees of freedom (DOF) and one rotational degree of
freedom.Three possible boundary conditions are summarized below according to their
restrictions on DOFs:
1. The fixed boundary condition restricts both linear DOFs and the rotational DOF. No
movement is allowed at the support. At the fixed support, a beam can neither translate nor rotate.
Representative examples include the anchored end of a diving board or the ground end of a
flagpole.
2. The guided boundary conditions allow two linear DOFs but restrict the rotational DOF.
3. The free boundary conditions provide for both linear DOFS and rotation. At a free end, a point
on beam may translate and rotate.A representative example is the free end of a diving board.
Flexural beams can be classified according to the combination of two mechanical boundary
conditions associated with it. For example, a beam fixed at one end and free at another is
conveniently referred to as a fixed-free beam, commonly called a cantilever. In MEMS research,
the following types of beams are the most frequently encountered: fixed-free (cantilevers), fixed-
fixed (bridges), and fixed-guided. It is important to correctly identify the boundary conditions
associated with a beam.The best way to learn and recognize boundary conditions is through
examples.

106
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The boundary conditions for these beams are summarized below.

(a) A fixed-free cantilever parallel to the plane of the substrate, with the free tip capable of
moving in a direction normal to the substrate. Lateral, in-plane movement of the free end would
encounter much significant resistance.
(b) A fixed-fixed beam (bridge) parallel to the substrate plane.
(c) There are actually two ways to classify this beam. It can be considered a fixed-fixed
cantilever parallel to the substrate plane with a thick and stiff part in the middle.Alternatively,
this beam can also be considered as two fixed-guided beams jointed in parallel to support a rigid
part.
(d) A fixed-free cantilever with the free end capable of movement perpendicular to the substrate.
Its behavior is similar to that of (a).
(e) A fixed-free cantilever with the free end capable of movement within the substrate
plane.With its thickness greater than its width, the movement of the free end in a direction
perpendicular to the substrate plane would encounter much greater resistance.
(f) A fixed-fixed beam (bridge).
(g) A fixed-free cantilever carrying a stiff object at the end.The stiff object does not undergo
flexural bending due to increased thickness.
(h) This beam is very similar to case (c) except for the fabrication method.
(i) A fixed-free cantilever with folded length. It consists of several fixed-free beam segments
connected in series.The free end of the folded cantilever is capable of movement in a direction
parallel to the substrate surface. Movement of the free end perpendicular to the substrate would
encounter much greater resistance.
(j) Two fixed-free cantilever connected in parallel.The combined spring is stiffer than any single
arm.
(k) Four fixed-guided beam connect to a rigid shuttle, which is allowed to move in the substrate
plane but with restricted out-of-plane translational movement.

UNIT II
SENSORS AND ACTUATORS-I
PART – A
1. State the principle of operation of electrostatic sensors.
A capacitor is broadly defined as two conductors that can hold opposite charges. It can be used as
either a sensor or an actuator. If the distance and relative position between two conductors changes
as a result of applied stimulus, the capacitance value would be changed. This forms the basis of
Capacitive ( electrostatic sensing).
2. Name the materials used in fabrication of parallel plate sensors.

107
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

i) Silicon ii) Germanium

3. What are the major advantages of electrostatic sensing and actuation?
Simplicity: The sensing and actuation principles are relatively easy to implement, requiring only
two conducting surfaces.
Low power: Electrostatic actuation relies on differential voltage rather than current. The method is
generally considered energy efficient for low – frequency applications. This is especially true under
static conditions, when no current is involved.
Fast response: Electrostatic sensing and actuation offers high dynamic response speed, as the
transition speed is governed by the charging and discharging time constants that are typically small
for good conductors.
4. What do you understand about comb drive device?
Generally, one set of finger like electrode is fixed on chip while second set is suspended and free to
move in one or more axis. Since the interdigitated fingers are sharped like tooth of combs, such
configuration is commonly referred to as comb drive device.
5. What are the two configurations of electrostatic transducer?
Parallel plate and interdigitated comb drive.
6. What are the two different movements by which parallel plate capacitor is moved?
• Normal displacement
• Parallel sliding displacement
7. What are the advantages of capacitive pressure sensor?
• Greater pressure sensitivity
• Lower temperature sensitivity
• Reduced power consumption
8. What is the principle of capacitive pressure sensor?
Capacitive pressure sensors are typically based upon a parallel plate arrangement whereby one
electrode is fixed and the other flexible. As the flexible electrode deflects under applied pressure,
the gap between electrodes decreases and the capacitance increases.
9. What is the main drawback associated with capacitive approach of sensor?
The main drawbacks associated with the capacitive approach are the inherently nonlinear output of
the sensor and the complexity of electronics (compared with the resistive bridge).
10. How nonlinearity can be avoided in capacitive pressure sensor?
The use of bossed diaphragms will mitigate this effect to some degree. Another linearizing
approach is to pattern the electrodes such that the sensing capacitance is measured from a particular
part of the diaphragm. Maximum deflection occurs at the diaphragm center but this is also the
location of maximum nonlinearity. By sensing the capacitance at an annulus removed a short
distance from the diaphragm center, non-linearity is reduced but at the expense of sensitivity.
11. How can be the parasitic capacitance avoided.
In order to reduce the effects of parasitic capacitance and achieve higher performance devices, the
pressure sensor should ideally be integrated with electronics. This is achieved by combining a bulk-
etched device with basic CMOS circuitry, but the more common solution is to employ surface
micro-machining.
12. What are the various advantages Comb-drive structures exhibit?
They have no sliding surfaces making them virtually showing no wear. The force of the drives is
independent of the position of the combs when one uses (the preferred) voltage control.
13. What are the major performance objectives in micro motor fabrication process?
Friction reduction and wear resistance.
14. How the sacrificial layers are removed in the process of micromotor fabrication?
Sacrificial layers are removed by immersing the wafer into HF etch solutions.
15. What are microactuators?
Microactuators are based on three-dimensional mechanical structures with very small dimensions
which are produced with the help of lithographic procedures and non-isotropic etching techniques.
For an actuator-like displacement the most different principles of force generation are used, such as
the bimetal effect, piezo effect, shape memory effect and electrostatic forces.
16. What is the use of micro pump?
Micropumps are attractive means for drug dosing in medicine, reagent delivery on chemical

108
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

applications or adhesive dispensing in the semiconductor industry.

17. Name any three thermal sensors.
i) Thermal bimorph sensors
ii) Thermal couples
iii) Thermal resistive sensors
18. State the principle of operation of thermocouple.
When any conductor is subjected to a thermal gradient, it will generate a voltage. This is known as
the thermoelectric effect or Seebeck effect.
19. What is thermal bimorph principle?
Two different materials, with different coefficients of thermal expansion, are joined together in
such a way, that a temperature change will cause the entire structure to deform in a desired way.
20. What are the two gripping methods used in micro-gripper?
Two different materials, with different coefficients of thermal expansion, are joined together in
such a way, that a temperature change will cause the entire structure to deform in a desired way.
21. What are the advantages of thermal bimetallic actuation?
• Relatively large range of movement can be achieved.
• Small actuator footprint for comparable displacement.
22. What are the disadvantages of thermal bimetallic actuation?
• Moderate to high power operation as current is used to generate ohmic heating.
• Lower response speed as the time constant is governed by thermal heating and
dissipation.
23. What is the term thermistor referred to?
The term thermistor is generally used to refer to semiconducting thermoresistors.
A thermistor is a type of resistor whose resistance varies significantly with temperature, more so
than in standard resistors. The word is portmanteau of thermal and resistor. Thermistors are widely
used as inrush current limiters, temperature sensors, self-resetting over current protectors, and self-
regulating heating elements.
24. What is bolometer?
A bolometer is a device for measuring the power of incident electromagnetic radiation via the
heating of a material with a temperature-dependent electrical resistance
25. What do you mean by pull-in voltage? (Apr 2017)
At particular bias voltage, the two curves representing the mechanical restoring force and the
electrostatic force intercept at one point tangentially. At the interception point, the electrostatic and
mechanical restoring forces balance each other. Moreover, the magnitude of the electric force
constant equals the mechanical force constant. The effective force constant of spring is zero. This is
a special condition and should be dealt with carefully in practice. The bias voltage that invokes
such a condition is called the pull – in – voltage, or Vp.
26. Write about thermal bi-morph principle. (Apr 2017)
The thermal bimetallic effect is a very commonly used method for sensing and actuation.
This mechanism allows temperature variation in microstructures to be exhibited as
transverse displacement of mechanical beams. The thermal bimorph consists of 2 materials
joined along their longitudinal axisserving as a single mechanical element.
27. What are the types of micro actuators? (Dec 2017)
Electrostatic • Electro‐magnet based • Thermal • Chemical • Piezo‐electric • Shape memory alloy
(SMA) • Smart material‐based • Light‐induced • Biological
28. What is the need of microgrippers? (Dec 2017)
Micro grippers are needed for manipulation of microscopic samples e.g. in the field of biology or
micro assembly. Most micro manipulators rely on the principle of micro tweezers which grip the
object, where the required force is delivered by micro actuators.
29. Define the role of actuators and sensors in the context of MEMS(MAY 2018)
The role of the sensor/actuator is to provide an interface between an electric world and the
physical world you interact with: a transducer.
30. What are the functions of interdigited finger capacitor?(MAY 2018)
The interdigital (or interdigitated) capacitor is an element for producing a capacitor-like, high pass

109
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

characteristic using microstrip lines.The shape of conductors is defined by the parameters that the
long conductors or “fingers” provide coupling between the input and output ports across the gaps.
PART B
1. Discuss in detail about parallel plate capacitive accelerometer.
The sensor consists of a metal ported oxide with 0.35µg-thick electro plated gold patch at its
distal end serving as proof mass. The length , width & thickness of cantilever are 108µm ,
25µm & 0.46µm.The counter electrode is made of heavily doped p type silicon.The capacitor
gap ( CB) is defined by epitaxy silicon layer grown on a silicon surface.
A surface micro machining process was developed using thermal oxide as cantilever structural
material &epitaxially grown silicon as a sacrificial layer.The process starts an n-type(100
silicon wafer). A heavily boron doped region is made using an oxide layer as the doping
barrier. An epitaxial silicon layer with resistivity of 0.5Ωcm is grown to a thickness of 5µm
over the entire wafer. Another layer of oxide is deposited &patterned , serving as a mask for
etching via hole & then as a barrier for doping. The oxide barrier is then removed followed by
growth of another layer of thick oxide which serves as the dielectric insulator, cantilever &
etching barrier other than gate. A layer of metal is deposited & patterned which provides
electrical interconnects to the bottom p + electrode, electrode on top of oxide cantilever & gate
of FET.Finally wet silicon h is performed to undercut epitaxial silicon beneath the oxide
cantilever.
𝐿
CB=∫0 𝐶 xdx =
𝐿
Ɛ0
∫ 𝑏 𝑑𝑥
0 𝑑 + 𝛿𝑥

2. What do you understand by membrane parallel plate pressure sensor?

A membrane pressure sensor can deduct pressure differential across the membrane.Two
pressure ports are required & to simplify design &use , absolute pressure sensors are desirable.
The use of vacuum avoids expansion of trapped air & increases the band width by eliminating
air damping inside the cavity.A membrane made of doped silicon serves as pressure sensing
element & 1 electrode. The counter electrode consists of patterned metal , thin film on the
bottom substrate . The process begins with (100) silicon wafer.An oxide mask is deposited &
patterned serving as chemical barrier during wet anisotropic etching of silicon. The oxide is
then photolithographically patterned further photo exposure action reduces line with resolution
& a boron diffusion step at 117.Celcius is conducted followed by stripping of oxide & growth
& pattering of another oxide layer.A layer of silicon layer is deposited & patterned to form
dielectric instillation & further provide electrical contact with the membrane later.
3. What are the applications of comb drive devices?
Comb drive actuator is one of the most common electrostatic actuator used in MEMS
applications. It uses both electrostatic energy from a DC voltage applied between the moving
& fixed comb drive structures, and the mechanical restoring force provided by the spring
structure . Comb drive actuators have been used as resonators, electromechanical filters,
optical shutters, micro grippers and voltmeters. These have also been used as the driving
element in vibromotors and micromechanical gears. It is desirable in comb drive to achieve
large displacements at low actuation voltages. The well known electrostatic micro-actuators
include side drive silicon micromotor, wobble micro-motor, comb drive microactuator and out
of plane diaphragm microactuator. Comb Drive actuators consist of two interdigitated finger
structures, where one comb is fixed and the other is connected to a compliant suspension. The
driving voltage between the comb structures causes the displacement of the movable fingers
towards the fixed fingers by an attractive electrostatic force. The position of the movable
finger structure is controlled by a balance between the electrostatic force and the mechanical
restoring force of the compliant suspension. Mechanical forces are generated through spring
structures. So besides electrostatic forces, mechanical forces also play a very important role.
Mechanical forces are directly depending upon the stiffness of the flexures. By changing these
flexures, mechanical forces also changes. It is very important to find the flexure compatibility
for large deflections at low actuation voltage. In the present work, different flexures of

110
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

electrostatic comb drives are simulated for large displacements.

4. Write in detail about the fabrication of thermal sensors.
Consider any of the thermal sensors-for eg: Hotwire anemometer- It utilizes PDMA(Plastic
Deformation Magnetic Assembly). Surface micro machined structures are anchored to
substrates with cantilever beam made of ductile material(Gold or Aluminum). The
microstructure is attached to pieces of electroplated ferromagnetic structures. When a
magnetic field is applied the ferromagnetic material interacts with the field and bend the
microstructure out of the plane. If the bending is significant the cantilever support hinges will
be plastically deformed resulting in permanent bent of microstructures. In fabrication Silicon
or glass or polymer is used as substrate. Above that a chrome/copper/titanium metal stack is
evaporated and patterned a sacrificial layer which will act as an adhesion layer. A thick layer
of titanium film reduces the in-process oxidation of copper film. A thick Photo definable
polymide is spun on patterned via lithography and cured. The polymide layer forms the
support prong and part of hotwire. A Cr/Pt/Ni/Pt film is then evaporated and patterned to form
a thermal element. A thick Ni resistor is sandwiched between two Pt films, which are used to
reduce possible oxidation of Ni while in operation process Pt is relatively inert at high
temperature. The authors then evaporate and pattern a thick Cr/Au film to serve as mechanical
bending element as well as electrical leads of hot-wire filament. They electroplate a thick
permalloy thin film on portions of cantilever support prongs. Sacrificial layer release is
performed by using a solution containing acetic acid and hydrogen peroxide to remove
selectively the copper thin film.PMDA assembly is carried out to lift the entire sensor out of
plane by placing a permanent magnet at the bottom of substrate. To finish the process the
device chip is then rinsed in de-ionized water and dried.
5. Explain briefly about heat transfer processes associated with pot heating.
Temperature is manifested at the micro scale by the vigorousness of atomic vibration. Heat
transfer occurs whenever a temperature gradient is present in a material. Successful design of
thermal actuators & sensors require familiarity with heat transfer processes.

There are 4 possible mechanisms for heat to move from one point to another:
1.Thermal
2.Natural
3.Forced thermal convection
4.Radiation
The convective heat transfer coefficient is influenced by surface geometrics, fluid velocity,
viscosity, & thermal diffusivity.
Heat is generated at the heating coil by passing current & is eventually lost to the ambient
background which is assumed to be a constant room temperature. Major heat transfer
pathways & the direction of heat flux are identified by arrows.
The heat produced by the heating coil first travels through the walls of the pot to reach the
body of water inside. Heat transfer occurs within the thickness of the wall via thermal
conduction. Once the heat reaches the interior wall of the pot, heat transfer to the liquid within
begins.The liquid mass closest to the wall warms up & begins to rise, setting up a natural
convection, which brings heat transfer from the pot wall to the interior of the body of water.If
the liquid in the pot were stirred, the heat transfers from inside wall of the pot to liquid could
become stronger as naturally convection is replaced by forced convection.
The body of water is exposed to air at the top.If the air outside is still, heat will transfer from
the water to the air through natural convection. On the other hand , if the air were moving, heat
would travel from the hot liquid to the air by forced convection.
Meanwhile, a person standing nearby feels heat wave coming from the heating coil.
Heat is said to move through air by radiation. Assuming the heating coil is at a significantly
high temperature, the radiation heat transfer can be quite strong.Certainly, the radiative
dissipation reduces the energy efficiently of heating the water. Even for such a simple
example, the heat transfer pathways are quite complex.
A heat flow will result between 2 points of different temperatures.The ability of a media or an
object to transfer heat between is quantified by its thermal resistance.

111
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

For two dimension thermal conductors & three dimensional ones , the effective
thermal resistance is much more difficult to estimate.
6. With neat diagram explain about thermal bimorph principle.
The thermal bimetallic effect is a very commonly used method for sensing and actuation. This
mechanism allows temperature variation in microstructures to be exhibited as transverse
displacement of mechanical beams. The thermal bimorph consists of 2 materials joined along
their longitudinal axisserving as a single mechanical element.
Principle of actuation is by increase in temperature heats the bimorph, as aluminum expands
more beam bends and results in an angular displacement. A thermal bimorph beam is
composed of two material films, with different Coefficients of thermal expansion CTE,
bonded at an interface. Typically thermal bimorphs are made of one material with a low CTE,
such as a dielectric like Polysilicon, and another material with a high CTE, such as metal like
aluminum. When the temperature of the bimorph is raised, the high CTE material will expand
more than the low CTE material. Since both materials are bonded together, stress develops in
both material layers due to the bonded interface constraint.. The high CTE material exhibits a
compressive stress because it is stretched below its equilibrium length, and the low CTE
material exhibits a tensile stress because it is stretched past its equilibrium length. The stresses
that developed upon an increase in temperature will cause the bimorph to curl towards the
Coefficient of thermal expansion of aluminum with α1=23e- 6 [1/K] where, the Coefficient of
thermal expansion of Polysilicon is α2 =2.33e-6[1/K], low CTE material to minimize the
internal energy stored by the stress. However thermal expansion co-efficient for most
materials are very small hence amount of displacement would be small. The temperature at
which the bimorph materials are deposited, assuming no residual stress is present from
processing conditions, both materials will be at their equilibrium lengths and the bimorph will
be flat. The opposite effect happens when the temperature of a bimorph is lowered. The high
CTE material will contract more than the low CTE material, and will develop tensile stress.
The lower CTE material will be contracted to a length shorter than its equilibrium length and
will develop compressive stress. When the temperature of a bimorph is lowered, the bimorph
will bend towards the high CTE material to minimize the internal energy stored by the stress.
The effects of an increase in temperature to the curl of a cantilevered bimorph are shown in
Fig 1. Figure 1 Thermal bimetallic bending (α1 > α2) Figure 2 Geometry of thermal bimorphs
actuator.

7. Discuss in detail about thermal resistors.

The resistance value of a resistor is a function of the resistivity p & its dimensions, including
length l & the cross-sectional area A. It is given as :
R = p l/A
Both the resistivity & the dimensions are functions of temperature. As a result , the resistance
value is sensitive to temperature. A thermal resistor is an electrical resistor with appreciable
temperature sensitivity.The resistance of a thermal resistor . R is related to the ambient
temperature in a relationship shown below:

112
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

RT = R0(1 + αR( T – T0) ),

Where RT &R0 are the resistance at temperature T & T0 respectively. The term αR is called the
temperature coefficient of resistance.Thermal resistors can be made of metal or
semiconductors. In both cases , the dimensions of a resistor change with temperature. The
electrical resistivity of both metals & semiconductors varies with temperature, but the
principles of such change is quite different for metals &semiconductors.The temperature
coefficient of resistance can be measured simply by heating up a resistor using a temperature
controlled stage & monitoring the resistance value. The temperature of the stage should be
increased slowly in small increments, to allow sufficient time interval in between each
temperature rise step to ensure thermal equilibrium. The bias voltage & current must be kept
low in this experiment, to minimize contribution of the electric heating power, which is the
product of voltage &current.The current & voltage used to interrogate the resistance value of a
thermoresistor may introduce heat to it. This phenomenon is called self-heating. The heating
power of a resistor under current l is:
P = I2R.
If self heatingoccurs , the resistance of the resistor is changed. Hence the slope of the I-Curve
will change. If the TCR value is positive, the resistance will increase at elevated power input
level. A representative I-V curve of a thermal resistor with TCR>0 is shown. Slope of I-V
curve decreases with the self-heating at high power input levels. On the other hand, if the TCR
value of the resistor is negative, the slope of the I-V curve would increase at high power input
levels.

8. Explain the working of interdigitated comb capacitor. (Apr 2017)

While parallel-plate capacitors generate sensing and actuation across planar electrodes facing
each other, a different class of capacitors take advantage of capacitance generated from
sidewalls of electrodes. Such capacitors provide alternative fabrication and operation modes
compared with parallel-plate capacitors.They involve interdigitated fingers (IDT) to increase
the edge coupling length () [45].Two sets of electrodes are placed in the same plane parallel to
the substrate. Generally, one set of finger-like electrode is fixed on chip while a second set is
suspended and free to move in one or more axes. Since the interdigitated fingers are shaped
like tooth of combs, such configuration is commonly referred to as the comb-drive device.

In a generic configuration (), two sets of fingers are in the same plane and their comb fingers
are engaged with an overlapping distance of l0. The length of the fingers is denoted Lc. For
simplicity, I assume the fingers in both sets have identical thickness (t) and width (wc). The
distance between a fixed comb finger and an immediately neighboring movable finger is d.The
thickness of fingers corresponds to that of the conductive thin film. The capacitance between a
pair of electrode fingers is contributed by vertical surfaces of the fingers in the overlapped
region, as well as by fringe capacitance fields. Capacitances derived from multiple finger pairs
are connected in parallel. Hence the total capacitance is a summation of capacitance
contributed by neighboring fingers. Opposite walls of comb fingers in the overlapped region
form a parallel-plate capacitor and contribute a capacitance C (). The magnitude of C between
two immediate neighboring fingers is

113
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

When designing a comb-drive capacitor, the thickness t and the distance d are highly process
relevant. The greater the thickness, the larger the capacitance effects. However, there are
different preferred methods for realizing desired thickness. The distance between gaps should
be as small as possible. However, it is determined by the ultimate lithography resolution.
The processing and design terms interact with each other. Since the thickness of the beam is
likely the same as the thickness of the mass, a thicker beam translate into a greater mass. Often
thick electrodes are made by using dry silicon etching, which is inevitably associated with
undercuts. Hence there is inherent conflict when trying to reach deep trench (large t) and small
gap (low d) at the same time. The fringe capacitance C f, is difficult to estimate
analytically.The most accurate way to estimate the fringe capacitance is by using finite
element method (FEM) (Figure 7). A simplistic way to estimate the fringe capacitance is to
assume it is a fixed fraction of the capacitance developed in the overlapped regions [48];
however, this is not entirely accurate and only suitable for first-order analysis.When dealing
with homework problems in this text, the fringe capacitance should be ignored. Two types of
comb-drive devices are commonly encountered, depending on the relative movement of two
sets of comb fingers allowed by their mechanical suspensions. The first is a transverse comb-
drive device, as shown in Figure 8.The set of free fingers moves in a direction perpendicular to
the longitudinal axis of comb fingers. Accelerometers for automotive airbag deployment made
by Analog Devices Corp. use transverse comb-drive configurations. Let us focus on a single
fixed finger and its two neighboring moving fingers. According to the analysis earlier, there
are two major capacitances associated with each finger pair, one to the left-hand side of the
finger, called Csl and one to the right-hand side, called Csr. At rest, the values of these two
capacitances are

When the free finger moves by a distance of x, the capacitance values of these two capacitors
become

The total value of capacitance isCtot = Csl+ Csr+ Cf.

When the transverse comb is used as a sensor, the displacement sensitivity Sx can be obtained

by taking the derivative of Ctot with respect to x, namely

114
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

If the transverse comb drive is used as an actuator, the magnitude of the force can be
calculated by taking a derivative of the total stored energy with respect to x,

A second type of comb-drive device is called the longitudinal comb drive (Figure 8c). The
direction of relative movement is along the longitudinal axis of the fingers, allowable by the
suspension.With a lateral movement (y), the capacitances associated with a single finger
changes to

The displacement sensitivity can be obtained by taking the derivative of Ctotwith respect to y,

There are many variations to the basic comb-drive design.The sensing and actuation
characteristics are strongly influenced by mechanical supports connected to the set of fingers.
For example, the design of flexural suspensions can be optimized to generate large
deformation while minimizing stress concentration. Coplanar transverse and longitudinal
comb drives are prevalent in MEMS. However, there are many different configurations and
geometries of comb-drive capacitors that deviate from these two major configurations. Some
comb drives are designed to produce vertical displacement, even when they initially reside in
the same plane. This is achieved by taking advantage of the fringe capacitance fields.This
levitation force is rather small but can be useful for applications such as optical phase tuning.
Reflective surfaces only need to be moved by up to one wavelength. Comb fingers that do not
reside in a common plane initially can be used to create out of plane forces or moments. The
out-of-plane arrangement of comb fingers can be achieved through novel fabrication
sequences or by using intrinsic-stress induced bending to lift one set of in-plane fingers
angularly. For example, an optical modulator capable of displacement has been made with 15
V driving voltage. Nearly all existing comb fingers have large length-to-width aspect ratio and
appear rectangular viewed from the top or side. Alternative profiles are possible. Comb-drive
fingers with curved profiles can provide important benefits such as maximum force, increased
linearity, or tailored force-displacement profiles
9. Explain the working of any one thermal actuator with neat diagram. (Dec 2017)
Ink Jet Droplet Injector
• Bubble formation time: 1 ms.
• Ink ejection time: 15 ms.
• Peak pressure: 14 ATM
• Upon removal of heat, vapor cools and the bubble retreats.
• Refill at 24 ms, lasts about 25 ms.
• Surface temperature: 90% of critical temperature (vaporization temperature) which is
330 oC.
• Homogeneous boiling across the surface of the heater, made of tantalum-aluminum
(Ta-Al) alloy.
• The heater has near zero TCR, so zero thermal expansion.
• HP Ink Jet Printer - Single Drop

• Thermal actuation
– Relatively high power: due to current operation.

115
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

– Lower response speed due to thermal time constant (dissipation and thermal
charging)
– Construction and fabrication: more complex due to material compatibility
considerations.
– Range of motion: relatively large.

Ink Jet Droplet Injector

10. 1)Explain the working of electrostatic rotary micromotors with neat diagram. (Apr/Dec 2017)
Electrostatic forces are not often used for driving macroscopic machinery. However,
micro devices have large surface-area-to-volume ratios and their masses are generally very
small, thus making electrostatic force, which is a surface force, an attractive candidate for
micro actuation.

The electrostatically driven micro motor was one of the earliest MEMS actuators . The motor,
schematically diagrammed in Figure , consists of a rotor that is attached to the substrate with a
hub, and a set of fixed electrodes on the periphery, called stators.The stators are grouped
together such that each group of four electrodes is electrically biased simultaneously.Three
such groups are present for the motor shown in Figure and are identified by different fill
patterns. To explain the operation principle of the electrostatic motor, let’s begin with the rotor
at an arbitrary angular rest position (Figure a). One group of stator electrode is first biased.
(The electrodes that are biased at a particular stage are identified by arrows placed next to
them.) An in-plane electric field develops between any given stator electrode in this group and
the closest rotor tooth next to it. This generates an electrostatic attractive force that aligns the
said tooth with the said stator electrode.The torque values are on the order of pico Nm for
voltages of the order of 100 V, large enough to overcome static friction. A small angular
movement of the rotor is made to reach a new configuration diagramed in Figure b.The

116
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

electric bias is shifted to the next group of stator electrodes (Figure c), resulting in another
small angular movement in the same direction. Continuous motion of the rotor can be
achieved by activating the stator electrodes by groups in succession. Later works in the area of
electrostatic actuators have provided new designs for greater power and torque output. In this
chapter, I will discuss the principles and governing equations of capacitive sensors and
actuators. Major advantages of electrostatic sensing and actuation are summarized in the
following.
1.Simplicity. The sensing and actuation principles are relatively easy to implement, requiring
only two conducting surfaces. No special functional materials are required. Other sensing
methods, such as piezoresistive and piezoelectric sensing, and other actuation methods, such
as piezoelectric actuation, require deposition, patterning, and integration of special
piezoresistive and piezoelectric materials.
2.Low power. Electrostatic actuation relies on differential voltage rather than current.The
method is generally considered energy efficient for low-frequency applications. This is
especially true under static conditions, when no current is involved. At high frequencies, a
time- and frequency-dependant displacement current, i(t), will develop in response to a time-
varying bias voltage V(t). The magnitude of the current is C=dV(t)/dt.The instantaneous power
delivered to a capacitor is p(t) = i(t) x V(t).
3.Fast response. Electrostatic sensing and actuation offers high dynamic response speed, as the
transition speed is governed by the charging and discharging time constants that are typically
small for good conductors. For example, the switching time of the mirrors in the digital
micromirror display (DMD) array is smaller than fast enough to support 8-bit gray scale
display.
Relative disadvantages of electrostatic actuation and sensing should be recognized as well.
High voltage required for static actuator operation is considered a drawback. The DMD
mirrors are switched by voltage on the order of 25 V to achieve plus or minus 7.5 degree
tilting.A monolithically integrated optical mirror with 90 tilt range requires 150 V bias voltage.
Linear electrostatic actuators requiring hundreds of volts to move a microstructure by several
tens of micrometers are rather common. High voltage introduces electronics complexity (for
providing high voltage supply) and material compatibility issues. Electrodes that are
mechanically connected with insulators tend to accumulate charges, especially under DC
operation modes. The trapped charges will change the operation characteristics.

There are two major categories of capacitive electrode geometries: parallel-plate capacitor and
interdigitated finger (comb-drive) capacitors. The static capacitance values of parallel plates
and comb drives are generally very small, being on the order of pF or smaller. For capacitive
sensors, careful design of electrical circuits is needed to register capacitance changes on the
order of fF or lower [4], in the presence of noise and interference sources. Capacitive actuators
often take advantage of electrostatic attraction between two oppositely charged surfaces.
Repulsive electrostatic forces, which are used less often, can be developed between two
surfaces with same charge polarities [5, 6]. For example, vertical levitation of interdigitated
fingers up to 0.5 µm has been demonstrated under applied voltage of 20 V for an 18-finger
device.
2)Explain working of magnetic actuator with neat diagram. (Apr/Dec 2017)
A MEMS magnetic actuator is a device that uses the microelectromechanical
systems (MEMS) to convert an electric current into a mechanical output by employing the
well-known Lorentz Force Equation or the theory of Magnetism. Micro-Electro-Mechanical
System (MEMS) technology is a process technology in which mechanical and electro-
mechanical devices or structures are constructed using special micro-fabrication techniques.
These techniques include: bulk micro-machining, surface micro-machining, LIGA, wafer
bonding, etc.
A device is considered to be a MEMS device if it satisfies the following:

 If its feature size is between 0.1 µm and hundreds of micrometers. (below this range, it

117
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

becomes a nano device and above the range, it is considered a mesosystem)

 If it has some electrical functionality in its operation. This could include the generation of
voltage by electromagnetic induction, by changing the gap between 2 electrodes or by a
piezoelectric material.
 If the device has some mechanical functionality such as the deformation of a beam or
diaphragm due to stress or strain.
 If it has a system-like functionality. The device must be integrable to other circuitries to
form a system. This would be the interfacing circuitry and packaging for the device to
become useful.
For the analysis of every MEMS device, the Lumped assumption is made: that if the size of
the device is far less than the characteristic length scale of the phenomenon (wave or
diffusion), then there would be no spatial variations across the entire device. Modelling
becomes easy under this assumption.

Magnetic actuation
The principle of magnetic actuation is based on the Lorentz Force Equation. When a current-
carrying conductor is placed in a static magnetic field, the field produced around the conductor
interacts with the static field to produce a force. This force can be used to cause the
displacement of a mechanical structure.

Magnetic actuation
Governing equations and parameters
A typical MEMS actuator is shown on the right. For a single turn of circular coil, the
equations that govern its operation are:

 The H-field from a circular conductor:

 The force produced by the interaction of the flux densities:

The deflection of a mechanical structure for actuation depends on certain parameters of the
device. For actuation, there has to be an applied force and a restoring force. The applied force
is the force represented by the equation above, while the restoring force is fixed by the spring

118
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

constant of the moving structure.

The applied force depends on both the field from the coils and the magnet. The remanence
value of the magnet its volume and position from the coils all contribute to its effect on the
applied Force. Whereas the number of turns of coil, its size (radius) and the amount of current
passing through it determines its effect on the Applied Force. The spring constant depends on
the Young's Modulus of the moving structure, and its length, width and thickness.

UNIT III
SENSORS AND ACTUATORS-II
PART-A
1. What is cantilever?
A cantilever is a beam anchored at only one end. The beam carries the load to support where
it is forced against by moment and shear stress. Cantilever construction allows for overhanging
structures without external bracing. Cantilevers can also be constructed with trusses or slabs.
2. What is dielectric effect of piezoelectricity?
It was observed that certain materials generate an electric charge when it is under a mechanical
stress. This is known as dielectric effect of piezoelectricity.
3. What is inverse effect of piezoelectricity?
The material would be able to produce a mechanical deformation when an electric field is
applied to it.
4. Define Curie point.
If a piezoelectric element is heated to certain threshold temperature, the crystal vibration may be
so strong that domains become disordered and the element becomes completely depolarized.
This critical temperature is called Curie point.
5. What do you mean by electromechanical coupling coefficient?(MAY 2018)
Electromechanical coupling coefficient k is a measure of how much energy is transferred from
electrical to mechanical or vice versa during the actuation process.
6. What is the most familiar application of quartz crystal?
The most familiar use of quartz crystal, a natural piezoelectric material is resonator in watches.
7. Which unique property of quartz enables it to be used in quartz crystal oscillator?
The property of its oscillation frequency is quite insensitive to temperature changes make it to
be used in quartz crystal oscillator.
8. Define Sputtering.
Sputtering is a form of Physical Vapor Deposition. It is used to deposit thin metal films in the
order of 100 A (1A = 10-10 m) onto the substrate surface. Sputtering process is carried out with
plasmas under very low pressure in high vacuum up to 5x10-7 torr and at room temperature. No
chemical reaction is involved in the deposition process.
9. What is the primary application of acoustic wave sensor?
Primary application of these sensors is to act like “band filters” in mobile telephones and base
stations.
10. What are the secondary applications of acoustic wave sensor?
• Sensing of torques and tire pressures
• Sensing biological and chemical substances
• Sensing vapors, humidity and temperature
• Monitor fluid flow in microfluidics
11. What are the two common types of micro pressure sensor?
Sensors using piezoresistors, Sensors using capacitances
12. What are the characteristics of sensor using piezoresistors?
• Small in size • Linear I/O relation • Temperature sensitive
13. What is PVDF?
The polyvinylidenfluoride is a synthetic fluoropolymer with monomer chains of (-CH2-CF2-)n.
It exhibits piezoelectric, pyroelectric, and ferroelectric properties, excellent stability to
chemicals, mechanical flexibility, and biocompatibility.
14. What are the assumptions to be made for calculating curvature of bending of a compact

119
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

model?
• The induced stress and strain should be along axis 1 or the longitudinal axis of
the cantilever
• The beam maintains constant curvature throughout the beam
• Shear effects are negligible
• Beam curvature due to intrinsic stress may be ignored
• Poisson’s ratio is isotropic for all films
15. What do you understand by PZT?
The lead zirconatetitanate (PZT) system is widely used in polycrystalline form with very high
piezoelectric coupling. The name PZT actually represents a family of piezoelectric materials.
Depending on the formula of preparation, PZT materials may have different forms and
properties.
16. What are the different methods of producing PZT materials?
Different methods of producing PZT materials are sputtering, laser ablation, jet molding and
electrostatic spray deposition.
17. Name the material used as bonding agent in preparing PZT materials?
Lithium carbonate and bismuth oxide are used as bonding agent in preparing PZT materials.
18. What are the applications of piezoelectric materials?
 Cantilever piezoelectric accelerometer
 Membrane piezoelectric accelerometer
 PZT piezoelectric acoustic sensor
 PZT piezoelectric Microphone
 Polymer piezoelectric tactile sensor
19. Why ZnO is used instead of PZT?
ZnO is used instead of PZT, because, although the PZT material offers greater piezoelectric
coefficient, it does so at the expense of greater dielectric constant and hence larger capacitance.
20. How are the electrical noise and impedance mismatch effects reduced in polymer
piezoelectric tactile sensor?
To reduce electrical noise and impedance mismatch effects, a two dimensional matrix of high
input impedance metal oxide semiconductor field effect transistor amplifiers have been directly
gate contact coupled to the lower surface of a piezoelectric PVDF polymer film.
21. What are the two most commonly encountered elastic wave?
Surface acoustic wave (SAW) and flexural plate wave (or Lamb wave). Saw occurs on samples
appreciable depth, whereas Lamb wave occurs in thin plates of materials.
22. What are the functions of SAW?
• Chemical sensing
• Environmental monitoring
• Electrical circuitry
• Transportation of fluid in contact with surface
23. What is the primary application of acoustic wave sensor?
Primary application of these sensors is to act like “band filters” in mobile telephones and base
stations.
24. What are the secondary applications of acoustic wave sensor?
• Sensing of torques and tire pressures
• Sensing biological and chemical substances
• Sensing vapors, humidity and temperature
• Monitor fluid flow in microfluidics
25. What is the major problem in pressure sensor?
Major problems in pressure sensors are in the system packaging and protection of thediaphragm
from the contacting pressurized media, which are often corrosive, erosive, and at high
temperatures
26. What is the function of Pirani Gauage? (Apr 2017)
The Pirani gauge is a robust thermal conductivity gauge used for the measurement of
the pressures in vacuum systems. It was invented in 1906 by Marcello Pirani.

120
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

27. List some Piezoelectric materials. (Apr 2017)

 Quartz
 Berlinite (AlPO4), a rare phosphate mineral that is structurally identical to quartz
 Sucrose (table sugar)
 Rochelle salt
 Topaz
 Tourmaline-group minerals
 Lead titanate
28. What is meant by piezoelectric effect? (Dec 2017)
Piezoelectric Effect is the ability of certain materials to generate an electric charge in response
to applied mechanical stress.
29. List some applications of Piezoelectric materials. (Dec 2017)
Engine Knock Sensors,Pressure Sensors, Sonar Equipment,Diesel Fuel Injectors,Fast
Response Solenoids,Ultrasonic Cleaning,Ultrasonic Welding,Piezoelectric Motors.
30. What is tactile sensor(MAY 2018)
It measures the coming information in response to the physical interaction with the
environment. The sense of touch in humans is generally modeled, i.e. cutaneous sense and the
kinesthetic sense
PART B
1. Write in detail about the principle of operation of piezoelectric sensors. Also discuss about the
properties of materials used in the fabrication of such sensors.
The phenomenon of piezoelectricity was discovered in the late nineteenth century. It was observed
that certain materials generate an electric charge (or voltage) when it is under a mechanical
stress.This is known as the direct effect of piezoelectricity. Alternately, the same material would
be able to produce a mechanical deformation (or force) when an electric field is applied to it. This
is called the inverse effect of piezoelectricity. (Some literatures refer to it as the converse effect of
piezoelectricity.) Piezoelectric crystals can be considered as a mass of minute crystallites
(domains).The macroscopic behavior of the crystal differs from that of individual crystallites, due
to orientation of such crystallites. The direction of polarization between neighboring crystal
domains can differ by 900 or 1800 . Owning to random distribution of domains throughout the
material, no overall polarization or piezoelectric effect is exhibited. A crystal can be made
piezoelectric in any chosen direction by poling, which involves exposing it to a strong electric field
at an elevated temperature. Under the action of this field, domains most nearly aligned with the
field will grow at the expense of others. The material will also lengthen in the direction of the
field.When the field is removed, the dipoles remain locked in approximate alignment, giving the
crystal a remnant polarization and a permanent deformation (albeit small).
The poling treatment is usually the final step of crystal manufacturing. Care must be taken in
all subsequent handling and use to ensure that the crystal is not depolarized, since this will result in
partial or even total loss of its piezoelectric effect.A crystal may be depolarized mechanically,
electrically, or thermally. Mechanisms for depolarization are further explained in the following.
Exposure to a strong electric field of opposite polarity to the poling field will depolarize a
piezoelectric element. The field strength required for marked depolarization depends on the
material grade, the time the material is subjected to the depolarization field, and the temperature.
For static fields, the threshold is typically between 200-500 V/mm.An alternating field will
also have a depolarizing effect during the half cycles that it opposes the poling field. Mechanical
depolarization occurs when mechanical stress on a piezoelectric element becomes high enough to
disturb the orientation of the domains and hence destroy the alignment of the dipoles.The safety
limits for mechanical stress vary considerably with material grade.
If a piezoelectric element is heated to a certain threshold temperature, the crystal vibration
may be so strong that domains become disordered and the element becomes completely
depolarized.This critical temperature is called the Curie point or the Curie temperature.A safe
operating temperature would normally be half way between 0_C and the Curie point. The
properties of piezoelectric elements are time dependent.The stability as a function of time is of

121
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

particular interest. Material characteristics may be degraded through aging effects

due to the intrinsic process of spontaneous energy reduction. The speed of aging can be controlled
through the addition of composite elements or through accelerated aging. Many piezoelectric
materials suffer from finite ion mobility. In other words, they do not provide long-term static
holding power when used in actuators. Design of piezoelectric actuators operating in DC
conditions must take into the account of electric leakage.
Properties of piezoelectric crystals- Since semi-conductor materials are often used in making
circuits and MEMS , it’s of interest to discuss piezoelectricity of important semiconductor
materials. Elemental semi-conductors such as silicon and germanium show centrosymmetric crystal
structure and do not exhibit piezoelectric behavior. III-V compound and II-VI compounds such as
Ga-AS and CdS on the other hand are piezo electric. Write in detail about Quartz, PZT,PVDF, ZnO

2. Discuss on piezoelectric actuator model. Explain any two applications of piezoelectric actuators.

Piezoelectric actuators are often used in conjunction with cantilevers or membranes for sensing and
actuation purposes. General models for such piezoelectric actuators are rather complex. Accurate
analysis often involves finite element modeling. For limited cases, such as cantilever actuator with
two layers, analytical models have been successfully developed. we will focus on the analysis of
cantilevers with two layers of materials, at least one of them being a piezoelectric layer.

The deflection of a two-layer piezoelectric structure can be described by compact formula.

Consider a cantilever with two layers, one elastic and one piezoelectric, joined along one
side.These two layers share the same length. A compact model for calculating the curvature of
bending has been made under the following assumptions:

1. The induced stress and strain are along axis-1, or the longitudinal axis of the cantilever;
2. Cross sections of the beam originally plane and perpendicular to the beam axis remain
plane and perpendicular to the resulting curved axis;
3. The beam maintains a constant curvature throughout the beam;
4. Shear effects are negligible;
5. Beam curvature due to intrinsic stress may be ignored;
6. The beam thickness is much less than the piezoelectric-induced curvature;
7. Second order effects such as the influence of d33 and electrostriction are ignored;
8. Poisson’s ratio is isotropic for all films.

The amount of force achievable at the free end of a piezoelectric bimorph actuator equals to the
force required to restore the tip of the actuator to its initial undeformed state. Since the
displacement is linearly related to force according to

122
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

the expression of the force is

Piezoelectric sensors and actuators with more than two layers are commonly encountered Several
general techniques can be found in both under simple or arbitrary loading.

ZnO Piezoelectric Force Sensor

A patch of ZnO thin film is located near the base of a cantilever beam, as shown in Figure below.
The ZnO film is vertically sandwiched between two conducting films. The length of the entire
beam is l. It consists of two segments—A and B. Segment A is overlapped with the piezoelectric
material while segment B is not.The length of segments A and B are and respectively. If the device
is used as a force sensor, find the relationship between applied force F and the induced voltage.
Solution. The c-axis (axis 3) of deposited ZnO is generally normal to the front surface of the
substrate it is deposited on, in this case the beam. A transverse force would produce a longitudinal
tensile stress in the piezoelectric element (along axis 1), which in turn produces an electric field
and output voltage along the c-axis. The shear stress components due to the force are ignored.

The stress along the length of the piezoresistor is actually not uniform and changes with position.
For simplicity, we assume the longitudinal stress is constant and equals the maximum stress value
at the base. The maximum stress induced along the longitudinal direction of the cantilever.

APPLICATIONS
Inertia Sensors
Cantilever Piezoelectric Accelerometer
One of the applications of piezoelectric sensors is micromachinedaccelerometers.An exemplary
device has been reported by a research group at the University of California at Berkeley [7]. The
schematic diagram of the sensor is shown in Figure . A proof mass is attached at the end of a fixed-
free cantilever beam.The cantilever beam consists of multiple layers of materials, with ZnO being
the functional piezoelectric material. ZnO is used instead of PZT because, although the PZT
material offers greater piezoelectric coefficient, it does so at the expense of greater dielectric
constant and hence larger capacitance.

The piezoelectric layer is sandwiched between a top conducting layer (platinum) and a bottom

123
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

conducting layer (polysilicon), similar to Example . A vertical acceleration will deflect the
cantilever, producing a longitudinal stress in axis 1, along the length of the cantilever. (The ZnO
film is polarized in the direction normal to the substrate.) Simple analysis in Example does not
apply here, however. The analysis is relatively complex because the proof-mass weight and the
piezoelectric layer are distributed in length and thickness.The distribution of stress is nonuniform
along the length. However, it can be said that the output voltage is proportional to since the stress is
applied in axis-1
and the electric polarization is measured in axis-3.The value of is 2.3 pC/N in this case. The
fabrication process is briefly discussed below. A silicon oxide and a silicon nitride layer are
deposited over the silicon wafer to serve as insulating layers. A phosphorous- doped polycrystalline
silicon film is deposited and patterned via reactive ion etching to define electrical contacts to the
bottom electrodes of the accelerometer. Here, the authors used polysilicon instead of metal because
polysilicon can withstand much higher processing temperature and provide greater flexibility for
down-stream process steps. Next, a -thick layer of phosphosilicate glass is deposited by LPCVD,
and patterned to define regions under the suspended cantilever.A second layer of -thick silicon is
deposited, covering bare silicon surfaces, the first polysilicon trace, and the PSG sacrificial
layer.This layer is patterned by using reactive ion etching (RIE) process with photoresist as the
mask. The RIE etch rate on silicon nitride and oxide (including PSG) is smaller, reducing damages
of over-time etch. However, care must be exercised to prevent or minimize overetching the first
polycrystalline layer.The authors then removed the photoresist layer with acetone. A thin LPCVD
silicon nitride layer is deposited over the wafer. This film serves as a stress-compensation layer for
balancing a highly compressive stress in the ZnO film. The exact thickness of this layer depends on
the actual stress and thickness of the ZnO layer.
Next a ZnO layer on the order of 0.5 micometer thick is deposited by RF-magnetron sputtering
from a lithium-doped ZnO target. Finally, 0.25 micrometer -thick Pt thin film is sputtered. The
stress of the ZnO film is reduced by a rapid thermal annealing step. Afterwards, the three layers
area defined using ion milling to produce precisely defined patterns. The underlying sacrificial
layer is removed using HF solutions while the patterned ZnO patch is protected by photoresist. The
ZnO protection is necessary because, although the film is covered by Pt on top, it is exposed on the
side and through possible pin-holes on the Pt films. The device exhibited a sensitivity of 0.95 fC/g
and a resonant frequency of 3.3 KHz.
Acoustic Sensors
PZT Piezoelectric Acoustic Sensor
Bernstein and colleagues described the use of an array of piezoelectric transducers as an
underwater acoustic imager [3]. The imager is akin to a CCD imager for optics. It consists of an 8
by 8 array of acoustic imaging sites.The cross-sectional view of each site is shown in Figure . A
layer of piezoelectric material—sol-gel deposited lead zirconatetitanate (PZT), lies on tops of a
silicon micromachined diaphragm. The size of each membrane varies from 0.2 to 2 mm. Two
electrodes sandwich a PZT thin film. The fabriation process is noteworthy because it involves the
use of sol-gel deposition of piezoelectric materials. After oxidizing the silicon wafers (Figure a)
and patterning the oxide, a heavy boron dose was diffused in all but the oxide protected areas
where through-wafer cavities would later be deposited (Figure b). All remaining oxide is removed.
A layer of low temperature oxide (LTO) is deposited using the LPCVD method. A 50-nm-thick Ti
and 300-nm-thick Pt layer are deposited as the bottom electrode, with the Ti serving to increase the
adhesion between the Pt and oxide (Figure c). This was followed by deposition of PZT by spinning
on a sol-gel mixture of lead acetate trihydrate, zirconium n-propoxide and titanium isoproposide in
a glacial acetic acid solvent.The sol-gel material is spun on, dried at 150 degree to remove the
solvent, heated at 400_C to remove residual organics, and preannealed at 600_C to densify the
layer and prevent further shrinkage (Figure d). Detailed receipe for making the sol-gel solutoin can
be found in the paper.The PZT material is poled at room temperature at 36 V DC bias for 2
minutes, to yield a resulting PZT film with a relative permitivity of 1400 and d 33of 246 pC/N.
The PZT film is patterned by wet etching PZT in a solutoin containing buffered hydrogen fluoride
(BHF) and hydrochloric acid (HCl) (Figure e).The BHF is a mixture of aqueous ammonium
fluoride and acqueous hydrogen fluoride (HF) [31]. The top electrode is separated from the ZnO by
a 2.5 micrometer -thick polyimide dielectric layer.This is followed by the deposition of a top

124
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

electrode and its patterning (Figure f). The backside cavity is produced in anisotropic etchant, with
the frontsidetempoerarily protected by a silicone elastomer cover (Figure g). Prior to silicon
etching, the LTO oxide on the backside is first removed.

3. Discuss about mathematical description of piezoelectric effects.

Mathematical Description of Piezoelectric Effects
Piezoelectric materials are crystals. The microscopic origin of piezoelectricity is the displacement
of ionic charges within a crystal, leading to polarization and hence electric field. A stress (tensile or
compressive) applied to a piezoelectric crystal will alter the spacing between centers of positive
and negative charge sites in each domain cell, leading to a net polarization manifested as open
circuit voltages measurable at the crystal surface. Compressive and tensile stresses will generate
electric fields and hence voltages of opposite polarity. Inversely, an external electric field will exert
a force between the centers of positive and negative charges, leading to an elastic strain and
changes of dimensions depending on the field polarity. Not all naturally occurring or synthesized
crystals exhibit piezoelectricity. Crystals can be classified into 32 groups according to crystal
symmetry. Centrosymmetric crystal structures are crystals that are symmetric along all axes
through the center of the crystal. These crystals occupy 11 out of 32 possible groups and are non-
piezoelectric materials because the positive and negative charge sites will not be spatially separated
under stress. Out of 21 noncentrosymmetric
groups, 20 are piezoelectric crystals. Piezoelectric effects are strongly orientation dependent.The
notation conventions for crystal orientations in the context of piezoelectric polarization are
discussed first.A piezoelectric material needs to be poled in a particular direction to provide a
strong piezoelectric effect, although some materials exhibit natural or spontaneous polarization.The
direction of positive polarization is customarily chosen to coincide with the Z-axis of a rectangular
system of crystallographic axes X, Y, and Z. Alternatively, the normal stress components along axes
X, Y, and Z are denoted by subscripts 1, 2, and 3, respectively.As such, the poling axis always
coincides with axis 3. Shear stress and strain components about these axes are denoted by
subscripts 4, 5, and 6, respectively in below figure.

125
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

In a piezoelectric crystal, the constitutive equation that relates electrical polarization (D)
and applied mechanical stress (T) is

whered is the piezoelectric coefficient matrix, the electrical permitivity matrix, and E the
electrical field. Here, an electric field is applied in conjunction with the mechanical stress to
provide more generality.The electrical polarization is contributed by two parts—one stemming
from electrical biasing and one from mechanical loading. If no electric field is present (i.e. E=0, ),
then the second term on the right hand side of above equation can be eliminated.

The inverse effect of piezoelectricity can be similarly described by a matrix-form constitutive

equation. In this case, the total strain is related to both the applied electric field and any mechanical
stress, according to

wheres is the strain vector and S the compliance matrix. the above equation can be expanded to a
full matrix form:

If there were no mechanical stress present the strain is related to the electric field by

The electromechanical coupling coefficient k is a measure of how much energy is transferred

from electrical to mechanical, or vice versa, during the actuation process.

This relation holds true for both mechanical-to-electrical and electrical-to-mechanical energy

126
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

conversion.The magnitude of k is a function of not only the material, but also the geometries of the
sample and its oscillation mode.

4. In brief discuss about properties of representive PZT materials.

Since semiconductor materials are often used in making circuits & MEMS, it is of interest to
discuss piezoelectric of important semiconductor materials. Elemental semiconductors such as
silicon & germanium show centrosymmetric crystal structure & do not exhibit Piezoelectric
behaviour. III- V compound & II – VI compounds such as GaAs &CdS on the other hand, are held
together by covalent & ionic bonding. The properties of thin film materials may be different from
their bulk counterparts. The progress of developing new Piezoelectric materials is fast-paced in
recent years.
Quartz:
The most familiar use of quartz crystal , a natural Piezoelectric material, is resonator in watches. In
a quartz-crystal oscillator, a small plate of quartz is provided with metal.

5. What do you know about PVDF and ZnO. Explain in detail about it.
PVDF:
The polyvinylidenfluoride (PVDF) is a synthetic fluoropolymer with monomer chains of ( - CH2 -
CH2-). It exhibits Piezoelectric , pyroelectric , & ferroelectric properties , excellent stability to
chemicals, mechanical flexibility , & biocompatibility. The Piezoelectric effect of PVDF has been
investigated & modelled. The stretched PVDF films are flexible & easy to handle as ultrasonic
transducers. The material is carbon based, usually deposited as a spin cast film from a dilute
solution in which PVDF powder has been dissolved. As for most Piezoelectric materials , process
steps after deposition greatly affect the behaviour of the film.

6. Discuss about ZnO piezoelectric actuator.

We have fabricated and operated two cantilevers in parallel in a new mode for imaging with the
atomic force microscope(AFM). The cantilevers contain both anintegrated piezoresistivesilicon
sensor and an integrated piezoelectric zinc oxide (ZnO) actuator. The integration of sensor and
actuator on a single cantilever allows us to simultaneously record two independent AFMimages in
the constant force mode. The ZnO actuator provides over 4 μm of deflection at low frequencies
(dc) and over 30 μm deflection at the first resonant frequency. The piezoresistive element is used to
detect the strain and provide the feedback signal for the ZnO actuator. Fabrication and
characterization of ZnO thin film piezoelectric sensors on a polyimide membrane is presented in
this paper. As a test device a differential pressure liquid flow sensor has been fabricated. We
discuss new material combinations for the fabrication of the test devices. The pressure sensor
membrane is a thin polyimide sheet bonded to a silicon wafer and the sensing material is sputtered
ZnO piezoelectric thin film. The fabricated liquid flow sensor has been tested with a piezoelectric
micropump for flow rates from 30 µl h−1 to 300 µl h−1. Stroke volumes of 1 to 10 nl have been
measured. The strain in the sensing layer has been modeled and a transverse piezoelectric
coefficient of e31,f = −0.294 C m−2 has been extracted.

7. With neat sketch explain about cantilever piezoelectric accelerometer.

The piezoelectric layer is sandwiched between a top conducting layer (platinum) & a bottom
conducting layer. A vertical acceleration will deflect the cantilever producing a longitudinal stress
in axis 1, along the length of the cantilever. The fabrication process is briefly discussed . A silicon
oxide & a silicon nitride layer are deposited over the silicon wafer to serve as insulating layers. A
phosphorous-doped polycrystalline silicon film is deposited & patterned via reactive ion etching to
defines electrical contacts to the bottom electrodes of the accelerometer.
A second layer of 2 μm-thick layer of phosphosilicate glass is deposited by LPCVD and patterned
to define regions under the suspended cantilever. The RIE etch rate on silicon nitride & oxide is
smaller, reducing damages of over-time etch. However care must be exercised to prevent or
minimize over-etching the first polycrystalline layer.The authors then removed the photoresist
layer with acetone. Next a ZnO layer on the order of 0.5 μm thick is deposited by RF-magnetron

127
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

sputtering from a lithium doped ZnO target. The ZnO protection is necessary because , although
the film is covered by Pt on top, it is exposed on the side & through possible pin-holes on the Pt
films.

8. Discuss about membrane piezoelectric accelerometer with neat diagram.

Cantilever Piezoelectric Accelerometer
One of the applications of piezoelectric sensors is micromachinedaccelerometers.An exemplary
device has been reported by a research group at the University of California at Berkeley [7]. The
schematic diagram of the sensor is shown in Figure . A proof mass is attached at the end of a fixed-
free cantilever beam.The cantilever beam consists of multiple layers of materials, with ZnO being
the functional piezoelectric material. ZnO is used instead of PZT because, although the PZT
material offers greater piezoelectric coefficient, it does so at the expense of greater dielectric
constant and hence larger capacitance.

The piezoelectric layer is sandwiched between a top conducting layer (platinum) and a bottom
conducting layer (polysilicon), similar to Example . A vertical acceleration will deflect the
cantilever, producing a longitudinal stress in axis 1, along the length of the cantilever. (The ZnO
film is polarized in the direction normal to the substrate.) Simple analysis in Example does not
apply here, however. The analysis is relatively complex because the proof-mass weight and the
piezoelectric layer are distributed in length and thickness.The distribution of stress is nonuniform
along the length. However, it can be said that the output voltage is proportional to since the stress is
applied in axis-1
and the electric polarization is measured in axis-3.The value of is 2.3 pC/N in this case. The
fabrication process is briefly discussed below. A silicon oxide and a silicon nitride layer are
deposited over the silicon wafer to serve as insulating layers. A phosphorous- doped polycrystalline
silicon film is deposited and patterned via reactive ion etching to define electrical contacts to the
bottom electrodes of the accelerometer. Here, the authors used polysilicon instead of metal because
polysilicon can withstand much higher processing temperature and provide greater flexibility for
down-stream process steps. Next, a -thick layer of phosphosilicate glass is deposited by LPCVD,
and patterned to define regions under the suspended cantilever.A second layer of -thick silicon is
deposited, covering bare silicon surfaces, the first polysilicon trace, and the PSG sacrificial
layer.This layer is patterned by using reactive ion etching (RIE) process with photoresist as the
mask. The RIE etch rate on silicon nitride and oxide (including PSG) is smaller, reducing damages
of over-time etch. However, care must be exercised to prevent or minimize overetching the first
polycrystalline layer.The authors then removed the photoresist layer with acetone. A thin LPCVD
silicon nitride layer is deposited over the wafer. This film serves as a stress-compensation layer for
balancing a highly compressive stress in the ZnO film. The exact thickness of this layer depends on
the actual stress and thickness of the ZnO layer.
Next a ZnO layer on the order of 0.5 micometer thick is deposited by RF-magnetron sputtering
from a lithium-doped ZnO target. Finally, 0.25 micrometer -thick Pt thin film is sputtered. The
stress of the ZnO film is reduced by a rapid thermal annealing step. Afterwards, the three layers
area defined using ion milling to produce precisely defined patterns. The underlying sacrificial
layer is removed using HF solutions while the patterned ZnO patch is protected by photoresist. The

128
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

ZnO protection is necessary because, although the film is covered by Pt on top, it is exposed on the
side and through possible pin-holes on the Pt films. The device exhibited a sensitivity of 0.95 fC/g
and a resonant frequency of 3.3 KHz.

9. In brief explain about PZT piezoelectric Microphone further shrinkage.

Cantilever microphones are chosen here because of the compliance. The cantilever is also free from
any residue-stress induced on a membrane. The authors stated that the use of a cantilever actually
created a microphone with great sensitivity (20Mv/μbar at 890 Hz, the resonant frequency).
Conversely, acoustic output can be generated when the device is used as an actuator , with a sound
pressure level of 75dB at 890 Hz with 4V drive. The size of the cantilever is 2 mm by 2 mm, with
the overall thickness being 4.5 μm. A ZnO thin film is located on the cantilever.

The fabrication begins with a <100> -oriented silicon wafer.A 0.2 -μm thick oxide is grown by
thermal oxidation, followed by the LPCVD deposition of a 0.5- μm.-thick silicon nitride. The
nitride is deposited at 835.C & 300 m Torr deposition pressures from a vapour with 6:1 ratio of
dichlorosilane (DCS, SiH2Cl2) to ammonia. A 0.2- μm.-thick LPCVD polysilicon electrode is
deposited on the wafer. The front side of the wafer is coated with spin-on photoresist electrode &
patterned. Because the etch rate on silicon nitride is finite, care must be taken to prevent excessive
over etch of the silicon nitride. The front side is then coated with a LPCV low temperature oxide
(LTO), an insulating layer.The wafer is annealed for 25min at 950 .C to reduce stress to relieve
stress and activate dopants in the polysilicon. A 05- μm- thick layer of ZnO is deposited using RF-
magnetron sputtering.

10. 1)Describe the working principle and circuit diagram of Tactile sensor. (Apr/Dec 2017)
reduce electrical noise and impedance mismatch effects, a two-dimensional matrix of high input
impedance metal-oxide-semiconductor field effect transistor (MOSFET) amplifiers have been
directly gate-contact coupled to the lower surface of a piezoelectric PVDF polymer film. This
MOSFET amplifier arrangement provides a separate, but identical high input impedance voltage
measurement capability for each taxel (tactile pixel). To realize a tactile sensor no larger than an
adult’s finger tip, a silicon IC with peripheral dimensions of 9.2 mm by 7.9 mm is designed and
fabricated.The prototyping was performed by MOSIS (Metal-oxide-semiconductor Implementation
System) foundry service.A portion of the IC’s area was reserved for the MOSFET amplification
and output interface circuitry.The 8 by 8 taxel matrix was allocated to an area of 5.3 mm by 5.3
mm.

A schematic diagram of the sensor array with cross section revealed is shown in Figure 0. A
continuous, poled PVDF film is attached to the front surface of the silicon chip.A –thick urathane
conformal coating layer was deposited on the PVDF film by spin coating. Individual taxel
electrodes are separated from their nearest neighbors. Although this work focuses on measuring the
normal stress component, tactile sensors with polymer piezoelectric materials have been developed

129
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

with component-selective response characteristics.

11. Determine the minimum thickness of the circular diaphragm a micro pressure sensor made of
silicon having a diameter of 600 μm and its edge is rigidly fixed to silicon die. The diaphragm is
designed to with stand a pressure of 20 MPa without exceeding the plastic yielding strength of
7000 MPa. The silicon diaphragm has a Young’s modulus E=190,000 MPa and Poisson’s ratio
ϑ=0.25.

12. Explain: (Apr 2017)

1)Piezoresistance in single crystal silicon.
Semiconductor strain gauges are made by selectively doping silicon [1, 2, 12].The piezoresistive
coefficients of a doped single-crystal silicon piezoresistor are influenced by its relative orientation
to crystallographic directions. If we consider a rectangular coordinate system having arbitrary
orientation with respect to the crystallographic axes of a homogeneous semiconductor, the electric
field components and current density components are related by a symmetric resistivity matrix as
in

This is the familiar Ohm’s law expressed in a matrix form. From routine applications of Ohm’s
law, we are used to thinking of the relation between potential differences and current as involving

130
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

only a scalar constant of proportionality. This is obvious for a one-dimensional conductor such as a
link in a circuit. However, for electrical conduction in a three-dime sionalsinglecrystalline media,
the current density and the potential gradient will not in general have the same direction.

The effective piezoresistive gauge factors attributed to each case in Table are determined
by multiplying the piezoresistive coefficient with the Young’s modulus in the direction of the
applied strain. Recall that the Young’s modulus of silicon is a function of the crystal directions as
well. For high-precision sensor applications, it is important to keep in mind that piezoresistive
sensitivity is not exactly constant.

2)Piezoresistance of polycrystalline silicon.

Polysilicon piezoresistors behave differently from ones made in single-crystal silicon. As MEMS
piezoresistors,polysilicon offers a number of advantages over single crystalline silicon, including
the ability to be deposited on a wide range of substrates [9, 10].The polycrystalline silicon also
exhibits piezoresistivity but the gauge factor is much smaller compared with that of single
crystalline one. The gauge factor is not dependent on the orientation of the resistor within the
substrate plane.However,it is influenced by growth and annealing conditions. Polysilicon used in
the MEMS field are required to have low stress and good conformality.There are subtle but
important distinctions from polysilicon used in electronic devices in terms of material
microstructures and processes

13. 1)Discuss on piezoelectric materials. (Dec 2017)

2)Explain the operation principle of piezoelectric sensor.

Let us now focus on the macroscopic description of the behavior of a piezoresistor under a normal
strain. The change in resistance is linearly related to the applied strain, according to

The proportional constant G in the above equation is called the gauge factor of a piezoresistor. We
can rearrange the terms in this equation to arrive at an explicit expression for G,

The resistance of a resistor is customarily measured along its longitudinal axis. Externally
applied strain, however, may contain three primary vector components—one along the
longitudinal axis of a resistor and two arranged 90_ to the longitudinal axis and each other.
A piezoresistive element behaves differently towards longitudinal and transverse strain
Components.
The change of measured resistance under the longitudinal stress component is called longitudinal
piezoresistivity. The relative change of measured resistance to the longitudinal strain is called the
longitudinal gauge factor. On the other hand, the change of resistance under transverse strain
components is called transverse piezoresistivity. The relative change of measured resistance to the
transverse strain is called the transverse gauge factor. Research has been conducted on single
crystal silicon doped piezoresistors and polycrystalline silicon doped piezoresistors, about the
effects of doping types (n or p type), doping concentration, and ambient temperature. It is
important to know that the gauge factor is a function of doping concentrations. Further, the
resistance changes with respect to the temperature, and even the gauge factor changes as a function
of temperature. For any given piezoresistive material, the longitudinal and transverse gauge factors
could be different. It is important to realize that longitudinal and transverse strains are often present
at the same time though one of them may play a clearly dominating role.The total resistance change
is the summation of changes under longitudinal and transverse stress components, namely

131
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Three cases of piezoresistive force sensors are schematically illustrated below to exemplify
longitudinal and transverse piezoresistor configurations. Strain gauges, represented by the resistor
symbol, are bonded to the outer surfaces of rods which are subjected to external loading forces.
Different resistor orientations and external force loading directions are presented (Figure ). In the
case illustrated in Figure a, the longitudinal piezoresistance dominates. In the cases illustrated in
Figure b and c, transverse piezoresistances dominate.
Resistance changes are often read using Wheatstone bridge circuit configuration. A basic
Wheatstone bridge consists of four resistors connected in a loop. An input voltage is applied
across two junctions that are separated by two resistors. Voltage drop across other two junctions
forms the output. One or more resistors in the loop may be sensing resistors, whose resistances
change with the intended variables. In the bridge shown in Figure a, one resistor is variable by
strain. The other resistors— and are made insensitive to strains by
locating them in regions where mechanical strain is zero, such as on rigid substrates.

UNIT IV
MICROMACHINING
PART – A
1. Define isotropic.
If the etch rates in all directions are identical, the etching is said to be isotropic.
2. Define anisotropic.
If the etch rate is orientation dependent, the etching is said to be anisotropic..
3. What is SLSP.
A three dimensional profile bound by slow etching <111> planes is called self limiting stable
profile (SLSP).
4. What is transition profile.
Before the self limited profile is reached, the profile is called transition profile.
5. Classify Transition profile.
Stable transition profile
Unstable transition profile
6. Define UTP.
A three dimensional etching profile that consists of fast etching, high indes planes [7],[8] are
called unstable transition profile.
7. What is STP.
A three dimensional etching profile that consists of only [6], [9], and [5] planes.
8. What is SLSP.
A three dimensional etching profile that consists of only [5] planes.
9. What is convex corner.
At a convex corner, the solid angle of masked region is less than 180 degree.
10. Define FSLC.

132
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

A footprint of self-limited cavity, is the intersection of SLSP cavity with front surface of wafer.
11. What is concave corner.
At a concave corner, the solid angle of masked region is greater than 180 degree.
12. Define DRIE.(MAY 2018)
Deep reactive ion etching is a special kind of RIE process. It is capable of producing deep and
high aspect ratio features with near vertical side walls.
13. What are the advantages of DRIE.
It offers fast fetch rate, vertical sidewalls, room temperature process capability, and the ability
to easily correlate the mask with the resultant three dimensional structures.
14. Which is the most commonly used isotropic silicon etchant.
The most commonly used isotropic silicon etchant is HNA – a mixed chemical solution
consisting of at least three constituent acids – the letter H stands for hydrofluoric acid, N for
nitric acid, and A for acetic acid.
15. What is SOI Wafer.
SOI stands for silicon on insulator. An SOI wafer is a specially wafer consisting of two single
crystal silicon wafer sandwiching, an insulator layer.
16. Define electrochemical etch stop layer.
Heavily doped silicon can reduce etch rate by wet silicon etchants. Alternatively, doped layers
can be biased during etching process to realize electrochemical tech stopping.
17. What is CVD.
In a chemical vapour deposition chamber, solid thin films are formed on wafers by
condensation of vapor or adhesion of solid phase reaction by products.
18. What is LPCVD.
If the energy is provided by heat alone and conducted under low processing pressure, the
process is called low pressure chemical vapour deposition.
19. What is PEVCD.
If the energy is provided by plasma power, the process is called plasma enhanced chemical
vapor deposition.
20. What are etch holes.
One strategy for decreasing the over all time for sacrificially under cutting large area plates is to
deploy small openings called etch holes.
21. What are the approaches taken for stiction issue.
Eliminate the capillary attractive force
Prevent excessive bonding force
Release structures that are stuck to the substrate
Provide counteractive force
22. Define stiction.
The failure mode of microstructures is referred to as stiction, a hybrid word combining sticking
and friction.
23. What is PSG.
The specific form of P-doped oxide is called phosphosilicate glass.
24. What is cantilever.
A beam that is fixed on one end and free on another.
25. Define suspension.
A beam that is fixed on both ends. The beam may contain plates in the middle. It is also referred
to as clamped- clamped beams.
26. Define Epitaxi. (Apr 2017)
Epitaxy refers to the deposition of a crystalline overlayer on a crystalline substrate.
The overlayer is called an epitaxial film or epitaxial layer. The term epitaxy comes from
the Greek roots epi (ἐπί), meaning “above”, and taxis (τάξις), meaning “an ordered manner”. It
can be translated as “arranging upon”. For most technological applications, it is desired that the
deposited material form a crystalline overlayer that has one well-defined orientation with
respect to the substrate crystal structure (single-domain epitaxy).
27. What is LIGA? (Apr/Dec 2017)
LIGA is a German acronym for Lithographite, Galvanoformung, Abformung (Lithography,

133
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Elctroplating, and moulding) that describes a fabrication technology used to create high aspect
ratio microstructures.
28. List the advantages of plasma etching than wet etching. (Dec 2017)
Some advantages of dry etching are its capability of automation and reduced material
consumption. Dry etching (e.g., plasma etching) costs less to dispose of the products compared
to wet etching. An example of purely chemical dry etching is plasma etching.
29. Compare bulk and surface machining
Bulk and surface micromachining are processes used to create microstructures on
microelectromechanical MEMS devices. While both wet and dry etching techniques are
available to both bulk and surface micromachining, bulk micromachining typically uses wet
etching techniques while surface micromachining primarily uses dry etching techniques.
PART B
1. Give the rules of anisotropic etching. Explain in detail with simplest case.
ANISOTROPIC WET ETCHING
Rules of Anisotropic Etching—Simplest Case
Let us first examine the simplest case—the profile of an etched pit in a <100> oriented silicon
substrate when the mask contains a rectangle or a square open window, with edges aligned to the
<110> direction (Figure 10.4a).

When a silicon wafer is immersed in a wet silicon etching solution, the atomic layer that is exposed
through the open window is etched first. Wet anisotropic etchants exhibit drastically different etch
rates along different crystal planes. In general, the etch rate along direction <111>
is the slowest among all crystal orientations. The reason for crystalline-dependant etch rate is not
yet fully understood though conjectures exist. It is believed that, since wet anisotropic etchants
remove silicon by first oxidizing a silicon surface (using the oxidizing constituents within the

134
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

solution) and then removing the oxide (using oxide-etching constituents), the difference of etch rate
is attributed to the difference of oxidation rates. However, there seems no reported direct evidence
correlating experimentally data and reaction rate analysis. Microscopically, the etch rate difference
is attributed to atomic bonding energy of silicon atoms on various surfaces. Atoms on different
crystalline surfaces are associated with different number of neighbors, bonding energy, and degrees
of difficulty for removal.

In the case of a square window, the cavity would end in a perfect point. In the case of a rectangular
opening, a cavity would end in a knife-edge at the bottom. In practice, due to nonideality of masks,
a perfect point is almost never encountered.

2. With an example explain about the rules of anisotropic etching (complex).

Rules of Anisotropic Etching—Complex Structures
In the previous section we reviewed a very simple and commonly encountered case: masks with
rectangular or square openings with edges aligned to <110> directions, lying in <100> silicon
surface. We now build up the complexity of analysis by relaxing limitation on mask shapes. First,
we eliminate the restriction that the edges of a window must be a noninterrupted straight line. Next,
we eliminate the restriction that the edges must be parallel to <110> directions.

Instead of reviewing the text of rules first, let us observe illustrated examples of actual etching
cases as predicted by a computer simulation software [10].The progression of etch profiles from
two independent mask patterns—labeled A and B—is illustrated in Figure . These two window
openings in the mask are very different. However, after a sufficiently long etch time (e.g., 170
min), they result in very similar three-dimensional profiles—blind STP-type cavities with edges
aligned to <110> (Figure h). Note that etching occurs underneath certain regions that are originally
masked.For example, the material underneath the extended beam in B is gradually removed
starting at its distal end.

135
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

The etching rules can be discussed easily by tracking how corners of mask patterns are
transformed. There are two types of corners—convex corners and concave corners [11]. At a
convex corner, the solid angle of masked region is less than 1800At a concave corner, the solid
angle of masked region is greater than 1800 Solid arrows in Figure a point to concave corners, while
convex corners are indicated by arrows with dashed lines. For convex corners, the fastest etching
planes dominate the three-dimensional shape evolving underneath the masked region. For concave
corners, the slowest etching planes dominate the three-dimensional shape. In other words, convex
corners tend to be undercut rapidly, exposing fast etching planes such as <211> and <411>. At
concave corners, slow etching planes such as <111> tend to develop and eventually prevail. The
undercut can be used to create suspended micro mechanical structures. For example, the mask
opening B will result in a suspended cantilever (Figure 10.9), provided that the mask is chemically
and mechanically sturdy enough to survive the entire wet etching sequence. The material of the
cantilever is that of the mask layer.The cantilever can be made of a variety of materials, including
silicon nitride, silicon oxide, and heavily doped silicon. Other materials are possible as long as they
exhibit good etching selectivity.A comprehensive survey of etch rates by various etchants on
silicon substrate and important mask materials can be found in the literatures.

136
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

3. Discuss in detail about dry etching of silicon.

DRY ETCHING AND DEEP REACTIVE ION ETCHING
Gas phase plasma etching is a complementary alternative to wet silicon anisotropic etching. Plasma
etch does not require wafers to be in contact with wet chemical solutions, thus simplifying cleaning
steps. Plasma etch offers different options of masking material. The process temperature is also
lower. Generally speaking, wafers are not heated although prolonged plasma etching may cause the
wafer temperatures to increase. The sidewall profile is different from what is achievable with web
anisotropic etching. In fact, the anisotropy of the vertical wall can be controlled, within a certain
range, by adjusting several process parameters, including the process pressure, DC bias voltage,
power input, and the chemical species used.
Deep reactive ion etching (DRIE) is a special kind of RIE process. It is capable of producing deep
and high-aspect ratio features with near vertical sidewalls. They are based on patents currently
owned by Robert Bosch GmbH and Texas Instruments Corporation. The key to achieving
continuous deep etching with high aspect ratio is to conduct the etching process in small depth
increments. This approach utilizes repeated cycles each containing an etching step followed by a
passivation step. A layer of inhibition film is deposited during the inhibition step on all walls of a
cavity. During the next etching step, the inhibition film is preferentially removed from the bottom
of the trenches due to ion bombardment, while preventing the etching of the sidewalls. Process
parameters play important roles in determining surface morphology and mechanical performances
of devices. DRIE silicon etching has gained popularity despite the high cost of processing
equipment. The DRIE process offers fast etch rate, vertical sidewalls, room temperature process
capability, and the ability to easily correlate the mask with the resultant three-dimensional
structures. The etch selectivity is a major benefit. The DRIE process can use photoresist, silicon
oxide, and metal as mask. In addition, the etching uniformity on a wafer scale is generally much
improved over those of regular RIE etch. Though powerful, the DRIE technique alone is capable of
creating nothing more than passive three-dimensional parts. For this reason, deep reactive ion
etching must be combined with other processing styles and steps to increase the electrical and
mechanical complexity of micro structures. The DRIE technique can be combined with many
processing styles, including anisotropic etching, to create a rich variety of innovative geometries.
Deep reactive ion etching of materials other than silicon has also been demonstrated, including
piezoelectric materials (i.e., quartz, PZT), Pyrex glass, and commercial PMMA.

4. Explain in detail about plasma etching.

Plasma etching is a very prominent method of removing materials from a wafer surface.Since the
process does not involve wet chemicals,it is often referred to as dry etching.The etching is carried
out in a specialized process equipment called the plasma etcher.A chamber with two opposing
electrodes is filled with a chemically active gas species.The process pressure is typically rather low
(perhaps 1/2000th of atmosphere pressure).Inside the plasma etcher,gas species are broken up by
the electric field,creating active gaseous radicals that are electrically charged.The radicals may
react with the wafer chemically.Meanwhile,the charged radicals may be accelerated in the electrical
field to high speed and interact with materials on the wafer physically
(bombardment,sputtering).Both physical and chemical processes may be present at the same time.
In general, physical etching is more directional and anisotropic,whereas chemical etching is
isotropic and material selective.If the electrode holding the wafer is grounded,the etching is called
plasma etching.If the wafer is fixed to an electrode on which AC bias is applied,the etching is
called reactive ion etching. Compared with plasma etching,reactive ion etching is more physical in
nature and its etch rate distribution is more anisotropic.A deep reactive ion etching (DRIE) process
is a special class of reactive ion etching,using special processing
equipment,gasmixture,andrecipe.The process is aimed at producing deep trenches with
vertical,smooth sidewalls in silicon.

5. Explain about isotropic wet etching.

The most commonly used isotropic silicon etchant is HNA—a mixed chemical solution consisting

137
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

of at least three constituent acids—the letter H stands for hydrofluoric acid, N for nitric acid, and A
for acetic acid [31].Various mixing ratios of three major ingredients are possible, resulting in
different etch rates on silicon and etch selectivity against masking layers. Because the solution
contains HF, the etch rate on oxide is relatively quick, at approximately 300 Å/min.

Gas phase etching of silicon using XeF2 or BrF3 is a complementary alternative to wet etching and
plasma etching. Under room temperatures, XeF2 exist in solid phase and BrF3 exists in liquid
phase. XeF2 sublimates and BrF3 vaporizes to form active gas when subjected to pressure below
100 mTorr. These gases react with silicon (in single crystal or polycrystalline forms) strongly, even
under room temperature, resulting in isotropic etch profiles. On the other hand, the gas phase
etchants have negligible etch rate on common mask materials, including silicon dioxide, metal, and
even photoresist. Care must be taken because the reaction byproducts contain HF vapor that, if
released into air accidentally, may become absorbed into water vapor and constitute air-borne drops
of highly concentrated HF acid. Bulk silicon etch rate is very high.The etch rate is generally not
characterized as etching depth per unit time. Rather, since etching is administered by injecting fixed
amount of etching agents in burst cycles, the etch rate is given as
m/cycle.
NATIVE OXIDE
A thin layer of silicon oxide would inevitably form on silicon naturally when the silicon is placed
or stored under room temperature and environmental oxygen and moisture levels.This layer of
material is very thin as the oxide growth temperature is low. However, this oxide may present a
significant etching barrier if the silicon etchant has small etch rate on silicon dioxide. It is a
common practice to remove native oxide before wet anisotropic etch or plasma etch steps to ensure
that the reaction with silicon starts predictably. The thickness of the native oxide is so small, it
often eludes detection by thin film thickness measurement instruments.A practical laboratory
technique to determine whether a native oxide layer is present is the water-beading test. A pure
silicon surface is hydrophobic while an oxidized silicon surface is hydrophilic.The
hydrophobic/hydrophilic nature of a surface can be quickly and inexpensively determined by
dispensing drops of water on the substrate surface and observe the shape of the water drop (Figure
).Water beads easily on hydrophobic silicon surfaces and rolls off the substrate when the substrate
is tilted. On the other hand, water would spread uniformly on a hydrophilic oxide surface. The
native oxide layer can be removed easily by dipping the wafer in low-concentration hydrofluoric
acid solutions (e.g., 5%) for a few seconds.

6. Discuss about LIGA process.(May 2018)

The LIGA consists of three main processing steps; lithography, electroplating and molding. There
are two main LIGA-fabrication technologies, X-Ray LIGA, which uses X-rays produced by a
synchrotron to create high-aspect ratio structures, and UV LIGA, a more accessible method which
uses ultraviolet light to create structures with relatively low aspect ratios.
The notable characteristics of X-ray LIGA-fabricated structures include:

138
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

 high aspect ratios on the order of 100:1

 parallel side walls with a flank angle on the order of 89.95°

 smooth side walls with R a {\displaystyle R_{a}} = 10 nm, suitable for optical mirrors
 structural heights from tens of micrometers to several millimeters
 structural details on the order of micrometers over distances of centimeters
Mask
X-ray masks are composed of a transparent, low-Z carrier, a patterned high-Z absorber, and a
metallic ring for alignment and heat removal. Due to extreme temperature variations induced by the
X-ray exposure, carriers are fabricated from materials with high thermal conductivity to reduce
thermal gradients. Currently, vitreous carbon and graphite are considered the best material, as their
use significantly reduces side-wall roughness. Silicon, silicon nitride, titanium, and diamond are
also in use as carrier substrates but not preferred, as the required thin membranes are comparatively
fragile and titanium masks tend to round sharp features due to edge fluorescence. Absorbers are
gold, nickel, copper, tin, lead, and other X-ray absorbing metals.

Substrate
The starting material is a flat substrate, such as a silicon wafer or a polished disc of beryllium,
copper, titanium, or other material. The substrate, if not already electrically conductive, is covered
with a conductive plating base, typically through sputtering or evaporation.

Exposure
A key enabling technology of LIGA is the synchrotron, capable of emitting high-power, highly
collimated X-rays. This high collimation permits relatively large distances between the mask and
the substrate without the penumbral blurring that occurs from other X-ray sources. In the electron
storage ring or synchrotron, a magnetic field constrains electrons to follow a circular path and the
radial acceleration of the electrons causes electromagnetic radiation to be emitted forward. The
radiation is thus strongly collimated in the forward direction and can be assumed to be parallel for
lithographic purposes. Because of the much higher flux of usable collimated X-rays, shorter
exposure times become possible. Photon energies for a LIGA exposure are approximately
distributed between 2.5 and 15 keV.

Development
For high-aspect-ratio structures the resist-developer system is required to have a ratio of dissolution
rates in the exposed and unexposed areas of 1000:1. The standard, empirically optimized developer
is a mixture of tetrahydro-1,4-oxazine (20 %), 2-aminoethanol-1 (5 %), 2-(2-butoxyethoxy)ethanol
(60 %), and water (15 %). This developer provides the required ratio of dissolution rates and
reduces stress-related cracking from swelling in comparison to conventional PMMA developers.
After development, the substrate is rinsed with deionized water and dried either in a vacuum or by
spinning. At this stage, the PMMA structures can be released as the final product (e.g., optical
components) or can be used as molds for subsequent metal deposition.

139
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Electroplating
In the electroplating step, nickel, copper, or gold is plated upward from the metalized substrate into
the voids left by the removed photoresist. Taking place in an electrolytic cell, the current density,
temperature, and solution are carefully controlled to ensure proper plating. In the case of nickel
deposition from NiCl2 in a KCl solution, Ni is deposited on the cathode (metalized substrate) and
Cl2 evolves at the anode. Difficulties associated with plating into PMMA molds include voids,
where hydrogen bubbles nucleate on contaminates; chemical incompatibility, where the plating
solution attacks the photoresist; and mechanical incompatibility, where film stress causes the plated
layer to lose adhesion. These difficulties can be overcome through the empirical optimization of the
plating chemistry and environment for a given layout.

Stripping
After exposure, development, and electroplating, the resist is stripped. One method for removing
the remaining PMMA is to flood expose the substrate and use the developing solution to cleanly
remove the resist. Alternatively, chemical solvents can be used. Stripping of a thick resist
chemically is a lengthy process, taking two to three hours in acetone at room temperature. In
multilayer structures, it is common practice to protect metal layers against corrosion by backfilling
the structure with a polymer-based encapsulant. At this stage, metal structures can be left on the
substrate (e.g., microwave circuitry) or released as the final product (e.g., gears).

Replication
After stripping, the released metallic components can be used for mass replication through standard
means of replication such as stamping or injection molding.

7. Discuss in detail about material selection criteria for a two layer process.
Material Selection Criteria for a Two-Layer Process
Essential prerequisites for selecting structural and sacrificial layer materials in a basic twolayer
process is modified based on the Ideal Process Rule:
1. The deposition of the structural material on top of the sacrificial material must not cause the
sacrificial layer to melt, dissolve, crack, disintegrate, or becomes unstable or destroyed in any way;
2. The method used for pattering the structural layer must not attack the sacrificial layer and any
existing layers on a substrate; 3. The method used for removing the sacrificial layer must not attack,
dissolve, or destroy the structural layer or the substrate.
The most established material pair is the polycrystalline silicon (as structural layer) and the
phosphosilicate glass (or PSG, as the sacrificial layer).This material system satisfies the three
criteria mentioned earlier.The PSG can withstand the deposition of the polysilicon layer.The
polysilicon can be patterned by using plasma etching, which etches PSG and oxide at much reduced
rates. The PSG and oxide can be removed by using wet HF acid solution, which exhibits very high
undercut etch rate on PSG but minimal etch rate on silicon. Certain processes may require two or
more layers or structural layers and/or sacrificial layers. The processes for micromotors generally
involve multiple structural and sacrificial layers. In such cases, the criteria of successful process
design becomes more complex. The rules that are discussed at the beginning of Section .1. now
must be expanded in the following manner:
1. Any layer that is deposited must not damage the underlying layers;
2. Any sacrificial etching process must not damaging any layers on the wafer;

140
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

3. Any process to pattern any sacrificial or structural layer must not etch all layers present
on the wafer.
In many practical processes, the device application may dictate the use of certain types of substrate,
sacrificial, structural, or interfacial materials. In such cases, it is impossible to limit the material
choices to the LPCVD materials mentioned earlier.When other materials are introduced, it may be
difficult to satisfy the three criteria listed above. One may have to pick a compromise choice of
structural-sacrificial layer combination. Under these circumstances, a number of general factors
should be considered when selecting the structural and sacrificial layers. These include: (1) etch
rates and etching selectivity; (2) achievable layer thickness; (3) temperature of material deposition;
(4) intrinsic stress of the structural layer; (5) surface smoothness; and (6) cost of materials and
processes.
During the sacrificial etching, both the structural layer and the sacrificial layer are exposed to
etchants.The etch rates of the etchant on the sacrificial layer and the structural layer are labeled r sa
and rstrespectively.The etch selectivity, defined as the ratio of and should be as high as possible. For
devices involving multiple structural layers and sacrificial layers, successful choice of materials,
etchants, and etching conditions must strive to maintain high etch selectivity at each
and every step of the process. It takes time and effort to build a comprehensive and reliable etch
rate table. A good starting point for beginners is published etch rate data. Papers published by
Williams et al. archived the cross reactivity between 53 materials and 35 etching methods.
However, it should be noted that etch rates can be influenced by a large number of factors.
Information provided in the two papers could be overwhelming for beginning readers. To ease the
learning curve, three tables are provided to capture essential material- and process related
information.
Ideally, a high sacrificial etch rate r sa is also desired to minimize the time required to complete
the sacrificial etching step. If two competing process options have identical selectivity, the one
with higher rsa is preferred. If a process has lower etch rate on the sacrificial material but higher
etch selectivity over another, the choice should be made by considering other additional factors
8. Discuss about striction and antistriction methods.
STICTION AND ANTI-STICTION METHODS
Sacrificial material removal is often accomplished using wet chemical solutions, because of the
high speed of etching, simplicity of setup, and generally good selectivity. It is necessary to dry the
wafer and chips afterwards, by natural or forced evaporation. However, the drying process is not
always straightforward.The situation is explained in Figure 11.9. As liquid is gradually removed
through evaporation, the top surface of a microstructure is exposed to air first. Liquid trapped
underneath the suspended micro structure will take much longer time to escape.A surface tension
force develops at the interface of trapped liquid and air acting in the direction tangential to the
liquid–air interface. For macroscopic devices, the surface tension force is negligible and does not
cause any significant deformation. However, since microscale devices often use compliant
structures and involves small gap spacing, the surface tension forces can cause appreciable
deformation of the surface microstructures. Often, contacts with the substrate are made. Studies
have been conducted in detail to characterize surface energies of bonding of selective material
systems.The contact between the suspended structure and the substrate
can lead to irreversible damages. Upon contact, strong molecular forces (e.g., van der Waals force)
are incurred to reinforce the attraction. Further, solid bridging is very likely due to the presence of
fresh reaction byproducts.
This failure mode of microstructures is referred to as stiction, a hybrid word combining sticking
and friction.

141
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Many practical methods have been developed to address the stictionissue.These methods take one
of four approaches:
1. eliminate the capillary attractive force by modifying the solid–liquid interface chemistry;
2. prevent excessive bonding force from occurring, for example, by elevating solution temperature
or reducing surface contact areas;
3. release structures that are stuck to the substrate using various energetic input forms, locally or
globally;
4. provide counteractive force to mechanical structures to prevent contact, for example by taking
advantage of bending stemming from intrinsic stress.

Two representative methods under these approaches are discussed in more detail in the following.
Adverse surface tension associated with the liquid-air interface can be circumvented by using the
supercritical fluid drying method. This technique was initially developed in the biology community
to dehydrate delicate biological tissues for morphology studies while avoiding structure collapses
and deformation. Let’s examine the process by observing a generic phase diagram of a solvent
material (Figure 0). Three phases—solid, liquid, and air—are the most familiar to readers. In
addition, a supercritical phase occurs at high pressure and temperature. A typical drying processing
utilizing the supercritical phase is described below. The chip containing released microstructures is
immersed in a liquid and placed under a moderate pressure (instead of the atmosphere pressure) and
room temperature.This starting condition is indicated by Point 1 on the phase diagram. The
temperature of the liquid is increased while maintaining the pressure constant.The solvent makes a
transition from the liquid phase to the supercritical phase (Point 2).The pressure of the supercritical
fluid is then dropped, causing the supercritical fluid to turn into a vapor (Point 3). The
transformation between liquid to supercritical fluid and then from supercritical fluid to vapor
involve virtually zero surface tension. The stiction can be reduced by using hydrophobic coatings
on micro structures and the

142
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

substrate to reduce bonding energy. Surfaces can be terminated with long-chain molecules that
self-assemble to form a hydrophobic self-assembled monolayer (SAM). Alternatively, flurocarbon
materials (Teflon like) deposited by plasma methods can be used as well. Hydrophobic surface
treatment finds applications beyond antistictioncoating.Forexample, the use of hydrophobic
patterned regions reduces surface bonding energy and allows lifting of wafer scale devices from a
mold without using sacrificial wet etching. Local hydrophobicity modification also allows
automated assembly of small parts in an self-organized manner.

9. Brief about acceleration of sacrificial etch.

ACCELERATION OF SACRIFICIAL ETCH
Sacrificial release of large-area plates or removal of sacrificial material inside long or blind
channels are generally very time consuming. Due to small characteristics dimensions (thickness
in the case of large area plate and cross section in the case of channels), the transport rate of
fresh etchants and byproducts decreases as the etching progresses. The chemical transport
characteristic for wet chemical etching of sacrificial materials in long blind channels has been
studied for PSG (with silicon nitride as sacrificial layer) and photoresist (with Parylene as
sacrificial layer). One strategy for decreasing the overall time for sacrificially undercutting large-
area plates is to deploy small openings, called etch holes.A scanning electron micrograph of
suspended micromachined flaps, each supported by two cantilever beams, is shown in Figure
11.7.The area of the flap is 200 m by 200 m. Four etch holes on each flap initiate sacrificial
undercut from within the interior of the flap, as well as from the edges. It can significantly shorten
the required time of etching. A micrograph of an etch hole in polysilicon is shown in Figure 11.8.
In most cases, the etch holes are small so as not to influence performances adversely. In certain
applications, however, etch holes may exert finite influence on the performance of devices. For
example, etch holes on an optical reflector causes diffraction as well as reduction of the reflectance.
In cases where etch holes are not preferred or allowed, alternative etching methods and materials
become necessary. A number of possible techniques have been demonstrated in the past. The
following materials have been reported as having extremely fast etch rate and high selectivity: (1)
Dendritic polymer such as hyper branched polymers (HBPs), (2) ZnO thin films.The etch rate on
ZnO by 2% HCl exceeds 1000 Å/s with no bubble formation. If the sacrificial material is metal, the
undercut may be accelerated by electrolytic dissolution with an active electrical bias. Furthermore,
self-assembled monolayers can be used to release large-area micro devices without wet chemical
etching. Researchers have also developed porous silicon (with through-film pores measuring 10–50
nm in diameter) for use as structural layers. The pores allow underlying sacrificial layer to be
removed rapidly without resorting to macroscopically defined etch holes. Because the holes are
nanoscopic in sizes, they have minimal impact on performances.

10. 1)Illustrate the anisotropic wet etching of cavities in {100} oriented silicon. (Apr/Dec 2017)
Let us first examine the simplest case—the profile of an etched pit in a <100> oriented silicon
substrate when the mask contains a rectangle or a square open window, with edges aligned to the
<110> direction (Figure 10.4a).

143
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

When a silicon wafer is immersed in a wet silicon etching solution, the atomic layer that is exposed
through the open window is etched first. Wet anisotropic etchants exhibit drastically different etch
rates along different crystal planes. In general, the etch rate along direction <111>
is the slowest among all crystal orientations. The reason for crystalline-dependant etch rate is not
yet fully understood though conjectures exist. It is believed that, since wet anisotropic etchants
remove silicon by first oxidizing a silicon surface (using the oxidizing constituents within the
solution) and then removing the oxide (using oxide-etching constituents), the difference of etch rate
is attributed to the difference of oxidation rates. However, there seems no reported direct evidence
correlating experimentally data and reaction rate analysis. Microscopically, the etch rate difference
is attributed to atomic bonding energy of silicon atoms on various surfaces. Atoms on different
crystalline surfaces are associated with different number of neighbors, bonding energy, and degrees
of difficulty for removal.

In the case of a square window, the cavity would end in a perfect point. In the case of a rectangular
opening, a cavity would end in a knife-edge at the bottom. In practice, due to nonideality of masks,
a perfect point is almost never encountered.

2)Describe applications of plasma etching.

Plasma etching is currently used to process semiconducting materials for their use in the fabrication
of electronics. Small features can be etched into the surface of the semiconducting material in order
to be more efficient or enhance certain properties when used in electronic devices. [3] For example,
plasma etching can be used to create deep trenches on the surface of silicon for uses in
microelectromechanical systems. This application suggests that plasma etching also has the
potential to play a major role in the production of microelectronics. [3] Similarly, research is

144
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

currently being done on how the process can be adjusted to the nanometer scale. [3]

Hydrogen plasma etching, in particular, has other interesting applications. When used in the
process of etching semiconductors, hydrogen plasma etching has been shown to be effective in
removing portions of native oxides found on the surface. [5] Hydrogen plasma etching also tends to
leave a clean and chemically balanced surface, which is ideal for a number of application. Chemical
plasma etching
Process gases are used whose molecules in plasma are mainly split into radicals. The etching effect
is mainly based on the reaction of these radicals with atoms or molecules of the substrate,
converting them into gaseous breakdown products.
Major applications:

 Removal of oxide layers

 Removing photoresist ("stripping")
 Ashing of matrices for analysis
 Etching of PTFE
 Structuring and microstructuring of semiconductors

Plasma is very selective, i.e. the process gases and substrates must be very well matched. The
etching is isotropic, i.e. it acts equally on all sides.
Reactive ion etching
Molecular gases form radicals and positively charged ions in the plasma. The reactive effect of the
radicals can be used for the etching process, as well as the kinetic energy of the ions. When the
plasma excitation is performed in this way, the ions are accelerated in the electric field and are fired
onto the substrate.
Reactive ion etching combines the effects of ion etching and plasma etching: A certain amount of
anisotropy is created and materials which do not chemically react with the radicals can also be
etched by this plasma. Above all, the etching rate is significantly increased. The substrate
molecules are excited by the ion bombardment and are thus much more reactive.
Application:

 Especially during the etching of semiconductors

11. Describe the wet etching of crystalline silicon with necessary diagram. (Dec 2017)
Removing materials by wet chemical reaction is common. It is used for removing metal,
dielectrics,semiconductors,polymers,and functional materials.The selectivity of etching again
masking materials, substrates and the target material is a crucial issue for MEMS process
design.Important performance characteristics include etch rate,temperature,and uniformity. A thin
film with a starting thickness of tf0 is covered by a masking material with an initial thickness of
tm0.The size of openings may be different, hence we include two representative windows open on
the front,one large (window A) and one smaller (window B).The wafer is exposed to an etchant.At
a given intermediate time t,the thickness of the film becomes t f,and the thickness of the mask
becomes tm.Here
and
tm0 - tm = t * etch-rate-on-mask.
tf0 - tf= t * etch-rate-on-film

145
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Ideally,the etch rate on the film should be much greater than that on the mask.The etch rate on film
in windows of different sizes may differ,due to the so called “loading effects”.At the end point of
the process (te),the films in both windows should be completely removed,whereas the thickness of
the mask may be reduced to
tm= tm0 - te* etch-rate-on-mask.
While the etch rate is defined in the vertical direction, an etchant may well attack the lateral walls
of the thin film.The extent of the lateral etch during time te is called undercut.Obviously, the
amount of undercut affects the precision of the pattern transfer process.

12. Describe in detail about DRIE. (Apr/Dec 2017)

Deep reactive ion etching (DRIE) is a special kind of RIE process. It is capable of producing deep
and high-aspect ratio features with near vertical sidewalls. They are based on patents currently
owned by Robert Bosch GmbH and Texas Instruments Corporation. The key to achieving
continuous deep etching with high aspect ratio is to conduct the etching process in small depth
increments. This approach utilizes repeated cycles each containing an etching step followed by a
passivation step. A layer of inhibition film is deposited during the inhibition step on all walls of a
cavity. During the next etching step, the inhibition film is preferentially removed from the bottom
of the trenches due to ion bombardment, while preventing the etching of the sidewalls. Process
parameters play important roles in determining surface morphology and mechanical performances
of devices. DRIE silicon etching has gained popularity despite the high cost of processing
equipment. The DRIE process offers fast etch rate, vertical sidewalls, room temperature process
capability, and the ability to easily correlate the mask with the resultant three-dimensional
structures. The etch selectivity is a major benefit. The DRIE process can use photoresist, silicon
oxide, and metal as mask. In addition, the etching uniformity on a wafer scale is generally much
improved over those of regular RIE etch. Though powerful, the DRIE technique alone is capable of
creating nothing more than passive three-dimensional parts. For this reason, deep reactive ion
etching must be combined with other processing styles and steps to increase the electrical and
mechanical complexity of micro structures. The DRIE technique can be combined with many
processing styles, including anisotropic etching, to create a rich variety of innovative geometries.
Deep reactive ion etching of materials other than silicon has also been demonstrated, including
piezoelectric materials (i.e., quartz, PZT), Pyrex glass, and commercial PMMA.

UNIT V
POLYMER AND OPTICAL MEMS
PART-A
1. Give any two applications of MEMS in medical field.
i) MEMS Pressure Sensors ii) MEMS Hearing-Aid Transducer
2. State the principle behind magnetic actuation of MEMS.
The principle of magnetic actuation is based on the Lorentz Force Equation.
 
F mag  q v xB
When a current-carrying conductor is placed in a static magnetic field, the field produced
around the conductor interacts with the static field to produce a force. This force can be used to
cause the displacement of a mechanical structure.
3. What are the classifications of polymers?
Thermal plastic polymer,Thermal setting polymer
4. Define viscoelastic creep.
When a force is applied to polymer, an instantaneous elastic deformation may occur, followed
by viscous, time dependent strain changes. As a result, many polymeric materials are
susceptible to time dependent deformation under constantly maintained stress. Such
deformation is called viscoelastic creep.
5. Define polyimide.
Polyimide represents a family of polymers that exhibit outstanding mechanical, chemical and

146
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

thermal properties a s a result of their cyclic chain bonding structure.

6. What is SU-8.
The SU-8 is a negative tone, near UV photo-resist first invented by IBM in late 1980’s with
main purpose of allowing high aspect ratio features (>15) to be made in thick photosensitive
polymers.
7. What is LCP.
Liquid crystal polymer is a thermoplast with unique structural and physical properties. LCP are
available commercially in sheets of various thicknesses.
8. What is PMMA.
The photodefinable PMMA thin film is a widely used e-beam and X-ray lithography resist.
9. Define Parylene.
Parylene is a thermalset polymer. It is the only plastic material that is deposited using chemical
vapor deposition process. The deposition process is conducted under room temperature
10. What is PDMS.
Poly dimethylsiloxane, an elastometer material belonging to the room temperature vulcanized
silicone alstometer family, offers many advantages for general MEMS applications.
11. What are fluoropolymers.
Fluoropolymers such as Teflon and cytop provide excellent chemical inertness, thermal stability
and nonflammability due to strong C-F bond.
12. What do you understand about cells?
Cells are basic functional units of life. The function of cell is determined by the genetic
sequence it carries. A basic human cell stores genetic codes, reproduces such codes upon cell
division, and manufactures protein molecules based on such codes.
13. What are the most commonly exploited biological binding protocols?
 Binding between antibody and antigens
 Binding between biotin and streptavidin molecules
 DNA complementary binding
14. What is electric double layer?
The high capacitance charged region of ions at the interface of liquid and wall is referred to as
the electric double layer.
15. What are the factors to be considered when selecting or developing a micro-machined
valve?
The reliability of valve operation
The repeatability of valve operation
The ability to withstand large pressure
The simplicity of valve construction
The simplicity of valve operation and control
16. How are the valves classified according to the mode of operations?
Cyclic valves, One-time valves and Constant off valves.
17. How is the valve structures categorized?
• Hard-membrane valves
• Soft-membrane valves
• Plug valves
• Threshold valves
18. Why channels made of PDMS popular?
PDMS is popular because of easy accessibility of material, rapid fabrication.
19. What are the major design variables in Nova sensor?
• Die size
• Membrane size
• Membrane thickness
• Resistor thickness
• Resistor doping level
• Resistor dimensions
20. What are the commercial applications of MEMS?

147
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

• Invasive and noninvasive biomedical sensors

• Miniature biochemical analytical instruments
• Cardiac management systems (e.g., pacemakers, catheters)
• Drug delivery systems (e.g., insulin, analgesics)
• Neurological disorders (e.g., neurostimulation)
• Engine and propulsion control
• Automotive safety, braking, and suspension systems
• Telecommunication optical fiber components and switches
• Mass data storage systems
• RF and wireless electronics
• Distributed sensors for condition-based maintenance and monitoring structural
health
• Distributed control of aerodynamic and hydrodynamic systems
21. Write the applications of MEMS in telecommunication industries?
• Optical switching and fiber optic couplings
• RF relays and switches
• Tunable resonators
22. Name few applications of MEMS in aerospace industry?
• Cockpit instrumentation
• Wind tunnel instrumentation
• Microsattelites
• Command and control systems with MEMtronics
• Propulsion systems with micro pressure sensor, chemical sensor for leak
detection
23. What do you mean by flow resistance of a channel?
The ratio between the pressure difference and the volumetric flow rate is called flow resistance
of a channel.
24. What are the applications of optical sensors? (Apr 2017)
These sensors are used for quality and process control, medico technologies, metrology,
imaging, and remote sensing.
25. Compare PDMS and PMMA. (Apr 2017)
PDMS- Poly dimethylsiloxane, an elastometer material belonging to the room temperature
vulcanized silicone alstometer family, offers many advantages for general MEMS applications.
PMMA- The photodefinable PMMA thin film is a widely used e-beam and X-ray lithography
resist.
26. List out some polymer MEMS materials. (Dec 2017)(MAY 2018)
Polyimide,SU-8,Liquid crystal polymer, Polydimethylsiloxane
27. What are the uses of PMMA?(Dec 2017)
Poly(methyl methacrylate) (PMMA), also known as acrylic or acrylic glass as well as by the
trade names Crylux, Plexiglas, Acrylite, Lucite, and Perspex among several others is
a transparent thermoplastic often used in sheet form as a lightweight or shatter-resistant
alternative to glass. The same material can be used as a casting resin, in inks and coatings.
28. Relate the factors involved in the performance of optical MEMS
The primary application domains, ordered by approximate precedence in
embracing MEMS tech- nology, have been: projection displays, components for optical fiber
communications (sources, switches, cross-connects, routers, etc.) and optical sensing and
imaging.
PART B
1. Write in detail about polyimide and SU-8.
Polyimide
Polyimides represent a family of polymers that exhibit outstanding mechanical, chemical, and thermal
properties as a result of their cyclic chain-bonding structure. Bulk processed polyimide parts are used
widely, from cars (struts and chassis in some cars) to microwave cookware. It is widely used in
microelectronics industry as an insulating material as well. Polyimides are formed from the
dehydrocyclization of polyamic precursors (Figure 13.1a) into cyclic polymers by incorporating

148
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

aromatic groups R and R’ (Figure 13.1b).These aromatic groups are chosen to affect the properties of
the final polyimide. For example, by chemically altering the polyamic acid precursor to include R”
groups sensitive to UV light as shown in Figure 13.1c, photo-patternable precursors can be made to
crosslink where exposed to UV light. The mechanical properties of polyimide films have been studied.
Cured polyimide films exhibit intrinsic stress on the order of 4x10 6 Pa to 4x107 Pa as measured using
suspended microfabricated polyimide strings. Further, mechanical and electrical properties of polyimide
may exhibit direction-dependent behavior. Many properties such as index of refraction, dielectric
constant,Young’s modulus, thermal expansion coefficient and thermal conductivity vary with
processing conditions.
Polyimide is commercially available as cured sheets, semi-cured sheets, or viscosity solutions for spin-
coating.The structure of a typical commercial polyimide—HN-type Kapton, is shown in Figure 13.1d.
In MEMS, polyimide is used as insulating films, substrates, mechanical elements (membranes and
cantilevers), flexible joints and links, adhesive films, sensors, scanning probes, and stress-relief layers.
Polyimide materials offer many favorable characteristics in these roles, including (1) chemical stability;
(2) thermal stability up to around 4000 C(3) superior dielectric properties; (4) mechanical robustness and
durability; and (5) low cost of materials and processing equipment. Polyimide can be used as structural
elements for sensors and actuators. Unfortunately, polyimide is neither conductive nor strain sensitive.
Functional materials such as conductors or strain gauges need to be integrated externally. Thin film
metal strain gauges have been integrated with polyimide, exhibiting an effective gauge factor on the
order of 2 to 6. An alternative is to modify the polyimide material for sensing purposes. For example,
piezoresistive composite of polyimide and carbon particles with effective gauge factor on the order of 2
to 13 has been demonstrated.

SU-8
The SU-8 is a negative tone, near UV photo-resist first invented by IBM in the late 1980s, with the
main purpose of allowing high aspect ratio features (>15) to be made in thick photosensitive polymers.
The photoresist consists of EPON® Resin SU-8 (from Shell Chemical) as a main component. The
EPON resin is dissolved in an organic solvent (GBL, gammabutyrolacton), with the quantity of solvent
determining the viscosity and the range of achievable thickness. Processed layers as thick as 100 m
can be achieved, offering tremendous new capabilities for masking, molding, and building high aspect
ratio structures at low cost. The cost of SU-8 lithography is considerably lower than that of other
techniques for realizing high aspect ratio microstrucures, notably LIGA process and the deep reactive
ion etching. SU-8 has been integrated in a number of micro devices, including microfluid devices, SPM
probe, and micro needles. It can also serve as a thick sacrificial layer for surface micromachining.

2. Explain in detail about LCP.

The liquid crystal polymer is a thermoplast with unique structural and physical properties. LCP are
available commercially in sheets of various thickness.When flowing in the liquid crystal state during
processing, the rigid segments of the molecules align next to one another in the direction of shear flow.
Once this orientation is formed, their direction and structure persist, even when LCP is cooled below its
melting temperature.This characteristic differentiate LCP from most thermoplastic polymers (e.g.,
Kapton), whose molecule chains are randomly oriented in the solid state. Owning to its unique
structure, LCP offers a combination of electrical, thermal, mechanical and chemical properties
unmatched by other engineering polymers. One of the earliest LCP films used in MEMS is a Vectra®
A-950 aromatic liquid crystal polymer, produced by Hoechst Celanese Corporation [23].The reported
melting temperature of Vectra A-950 is 2800C. The specific gravity ranges from 1.37 to 1.42 kg/m3 and

149
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

the molecular weight is greater than 20,000 g/mol.The compatibility of this LCP with commonly used
chemicals in micromachining was first investigated in [23]. LCP is virtually unaffected by most acids,
bases and solvents for a considerably long time and over a broad temperature range. Extensive tests
showed that LCP was not attacked or dissolved by at least the following chemicals common in
microfabrication: (1) organic solvents including acetone and alcohol, (2) metal etchants for Al,Au and
Cr, (3) oxide etchants (49% HF and buffered HF), and (4) developers for common photoresist and SU-8
resist. LCP films have excellent stability. It has very low moisture absorption and low moisture
permeability, which are better than PMMA and comparable to that of glass. For other gases, including
oxygen, carbon dioxide, nitrogen, argon, hydrogen and helium, LCP also exhibits above-average barrier
performance. Further, the permeation of gases through LCP is not affected by humidity, even under
elevated temperatures.The thermal expansion coefficient of the LCP material can be controlled during
the fabrication process to be both small and predictable.The LCP film also shows excellent chemical
resistance. LCP was originally used as a high-performance substrate material for high-density printed
circuit board (PCB) [23]. A number of unique processing methods have been developed for LCP,
including laser drilling and via filling (for low-resistance electrical through interconnects). LCP films
are used as substrates for space and military electronics systems, both for its performance and its
stability. For instance, it has been explored as a high-performance carrier of radio frequency
electromagnetic elements such as antennas. Results from high-frequency tests show that LCP has a
fairly uniform relative dielectric constant of 3 in the range 0.5 to 40 GHz and an extremely low loss
factor.
For LCP film with uniaxial molecule orientation, its mechanical properties are anisotropic and
dependent on the polymer orientation. For instance, the uniaxial LCP film can withstand less load in the
transverse direction (i.e., the direction orthogonal to the orientation of its molecules) than in the
longitudinal direction (i.e., the direction along the orientation of its molecules). If necessary, biaxially
oriented film with equally good transverse and longitudinal direction properties can be made to correct
the anisotropy behavior.A uniaxial film can be formed by bonding multiple layers of anisotropic films
with angular offsets of crystal orientations between layers. The orientation of LCP molecules varies
through the thickness of the film, while at the two sides of the film molecules are oriented at opposite
angles. If the angles are +450 and -450 at either side, the mechanical properties, such as coefficient of
thermal expansion, tensile strength and modulus are nearly isotropic. Commercial LCP material is
supplied in sheet format. The thickness of LCP film could vary from several microns to several
millimeters. Some can be provided with copper clad layers on one or both sides. The optional copper
clad layer is normally 15 to 20 m thick. This copper layer is laminated in a vacuum press at a
temperature around the melting point of LCP.

3. Dicuss about Parylene, Flurocarbon.

Parylene
Parylene is a thermalset polymer. It is the only plastic material that is deposited using chemical vapor
deposition (CVD) process.The deposition process is conducted under room temperature. A Parylene
deposition system consists of a source chamber connected to a vacuum deposition chamber.A dimer (di-
para-xylene) is heated inside the source chamber to approximately It sublimates into a gaseous
monomer, which then enters the vacuum chamber and coats on objects within. Three parylene dimer
variations are available from commercial vendors, including Parylene C (widely used), Parylene N (for
better dielectric strength and penetration), and Parylene D (for extended temperature performance). The
Parylene film offers very useful properties for MEMS applications, including very low intrinsic stress,
room temperature deposition, conformal coating, chemical inertness, and etch selectivity. Parylene
coating is deal for electrical isolation, chemical isolation, preservation, and sealing. Parylene has been
used for microfluidics channels, valves, retina prosthesis, sensors (acceleration sensors, pressure

150
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

sensors, microphones, and shear stress sensors ). The thickness of Parylene coating is generally
controlled by the amount of dimer loaded. Thickness monitors and end-of-point detectors for in-situ
Parylene thickness monitoring have been developed, for example based on thermal transfer principles.

Fluorocarbon
Fluoropolymers such as Teflon and Cytop provide excellent chemical inertness, thermal stability, and
nonflammability due to the strong C-F bond.They can be used as a surface coating, insulation,
antireflection coating, or as adhesion agent. Cytop is a trademarked material (by Asashi Glass Company
of Japan). It exhibits many good properties as Teflon but offers high optical transparency and good
solubility in specific fluorinated solvents. Fluoropolymer films can be spin coated or deposited by
PECVD method. In MEMS,Teflon and Cytop films have been used for electrical insulation, adhesive
bonding, and friction reduction.

4. Describe about parylene surface micromachined pressure sensor.

Parylene Surface Micromachined Pressure Sensor
The basic design of a surface micromachinedParylene membrane with integrated resistors is shown in
Figure 13.5. The membrane, circular as shown, is elevated from the substrate surface by a distance of
0.5 to 30 m Strain gauge resistors for sensing membrane displacement are typically placed along the
periphery of the membrane, as depicted in Figure 13.5. Metal films can serve as piezoresistors in place
of doped polycrystalline silicon. However, one disadvantage lies in the fact that the resistivity of thin
film metal is much smaller compared with that of polycrystalline silicon. In order to achieve appreciable
magnitude of resistance (e.g., greater than 40 ohm), these resistors are zigzagged, consisting of
alternating radial segments and tangential ones.The redial segments are primarily responsible for the
displacement sensing. When a vertical force or pressure is applied on the membrane, the membrane will
be deformed to induce in-plane stress in the radial direction, which is sensed by radial segments of the
strain gauge resistors.
Major design variables of a membrane device include the diameter and thickness of the membrane, the
height of the underlying cavity, and the resistance of thin film resistors. A successful design must take
into consideration of processing and performance needs simultaneously. For example, to avoid
unwanted membrane collapse and sticking to the substrate, it is generally desirable for a membrane to
be smaller, thicker, and with greater
cavity heights. However, if a membrane is overly small, it may not have enough area to accommodate
appreciable resistance from embedded metal resistors.
Increasing the cavity height (and membrane clearance) will generally cause difficulties with electrical
continuity from the top of the membrane to the substrate level; there are also practical difficulties with
building thick sacrificial layers (e.g., greater than 20m). The overall process, diagramed in Figure and
Figure 13.7, can be achieved under relatively low overall temperature (i.e., less than 1200 C). As a
result, the process can be realized on a variety of substrate materials, including silicon, glass, and even
polymers. A layer of photoresist is spin coated on the front surface of a substrate and patterned
photolithographically (Figure a). The spin-on photo resist is cured in a convection oven, first at 600 for 5
min (to remove edge beads) and then at 1100 for 1 min.The patterned photoresist will reflow slightly
during post-development bake (1100 C for 2 min), rounding the edges of features to create a sloped
edge. Optionally, the photoresist can be selectively thinned (to a target height of 2.5m) near
etch/sealing holes (Figure b).The authors achieved this by additional exposure near the etch hole regions
using a separate mask.This reduces the amount of Parylene needed to seal the cavity in optional step m.
Major reasons for selecting photoresist as the sacrificial material (as opposed to metal or silicon
dioxide) include: (1) the thickness of the sacrificial layer can reach 10–20m range relatively easily and
quickly; (2) the edge of photoresist sacrificial layer can be smoothed to realize gentle slopes. A 1 m
thick Parylene thin film is then deposited on top of the wafer surface (Figure c).The Parylene is
subsequently coated with a 150-nm-thick Al thin film, which is then patterned photolithographically
(Figure d). Oxygen plasma etch is used, with the thin film Al as the mask, to remove the exposed
Parylene (Figure e) and to reach underlying hard substrate. An oxygen plasma etch step creates a rather
steep transition between the substrate and the top Parylene surface. This will pose problems of electrical

151
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

continuity in the future when thin film metal wiring traverses between the substrate and the Parylene top
surface.
Therefore, after removing the metal etch mask, the authors made efforts to smooth the edge of the
Parylene using a process described below. A layer of photoresist is spin coated (2500 rpm for 10 sec)
following the removal of the thin film metal.The top profile of the photo resist after curing and reflow
(5 min at 640C followed by 1 min at 1100C ) is much smoother than the slope in Parylene (Figure f).A
global oxygen plasma etch is performed to etch the photoresist (Figure g). Since the oxygen plasma etch
rate on photoresist and on Parylene is roughly identical (at 300 mTorr of pressure and 350 W of power),
the smooth, reflown edge of the photoresist is transferred onto the Parylene after the photoresist on
planar surfaces is removed.

152
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Thin film metal resistors embedded in the polymer membrane must be located near the surface, off the
neutral axis of the membrane. After the edge profile is adjusted, a layer of 200-nm-thick Au (with a 5-
nm-thick Cr thin film underneath for adhesion) is deposited and patterned on top of the Parylene thin
film (Figure h). Gold offers better conformal coverage than others (e.g., Ni, NiCr, and Al), allowing
electrical continuity across edges of membrane. The underlying Parylene shields the metal resistor from
direct contact—and shorting, with the bottom of the cavity if the membrane is fully displaced
downward. Another layer of Parylene encapsulates the metal resistor, preventing delamination and
accidental electric shorting. The thickness of short tangential segments are made thicker by depositing
an additional layer of metal using the lift-off process (Figure 13.7i). The entire device is then coated
with another 8 m thick (nominal) Parylene layer (Figure 13.7j). The Parylene is patterned using a thin
film metal (300-nm-thick Al) as a mask in oxygen plasma etching.A global plasma etch is performed to
pattern the newly deposited Parylene film. This exposes end regions of etch hole and also reopens the
contact pads (Figure 13.7k).The photoresist sacrificial layer inside the cavity is removed using acetone
(Figure 13.7l). This is conducted under room temperature for 3 h (for a 400 m diameter membrane) or
more (for larger membranes).The wafer is dried under an infrared lamp for 10 min. It is impractical to
remove the wet chemicals from the cavity using spin-drying because the membranes are found to
collapse and adhere to the substrates. Optionally, the authors reported hermetically sealing the cavity by
depositing another thin film of Parylene (approximate thickness being 2m ). At the opening of each
etch hole, two fronts of Parylene will grow from opposite surfaces and eventually meet to seal the
cavity interior (Figure 13.7m).The pressure inside the cavity is kept at the processing pressure of
Parylene deposition (approximately 40 mTorr). The Parylene film is also deposited on the bonding pad
region, unfortunately. The authors performed another sequence of masking (with metal film) and
etching (using oxygen plasma) to reopen the bonding pads regions (Figure 13.5n). An optical
micrograph of a fabricated device is shown in Figure 13.8.The device has successfully performed for
measuring contact pressure. Further characterization and improvement of sensitivity is needed in order
to demonstrate measurement of air or liquid pressure changes.

5. Discuss about LCP piezoresistive flow sensor.

LCP Piezoresistive Flow Sensor
A flow sensor consisting of a polymer cantilever beam has been made. As shown in Figure below, flow
imparts momentum on the cantilever and causes it to bend, inducing strain at the base of the

153
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

cantilever.The strain is transduced into an electrical signal using a piezoresistive sensor made of thin
film metal.While the gauge factor of doped silicon can reach 10–20, the gauge factor of thin film metal
is much lower, typically ranging from 1 to 5. However, the increased thickness and compliance of these
polymer film devices has been shown to offset the reduced gauge factor of metal film strain gauges
resulting in sensitivity comparable to silicon-based devices. The flow sensor uses nickel-chrome (NiCr)
strain gauges on an LCP cantilever that is 1000m wide and 3000 m long. Figure a show a micrograph
of the completed device. Wind tunnel testing with flow rates from 0 to 20 m/s showed a velocity-
squared relationship as expected, as seen in the quadratic trend line in Figure below.

6. Describe the molecular structure and properties of LCP. (Apr 2017)

The liquid crystal polymer is a thermoplast with unique structural and physical properties. LCP are
available commercially in sheets of various thickness.When flowing in the liquid crystal state during
processing, the rigid segments of the molecules align next to one another in the direction of shear flow.
Once this orientation is formed, their direction and structure persist, even when LCP is cooled below its
melting temperature.This characteristic differentiate LCP from most thermoplastic polymers (e.g.,
Kapton), whose molecule chains are randomly oriented in the solid state. Owning to its unique
structure, LCP offers a combination of electrical, thermal, mechanical and chemical properties
unmatched by other engineering polymers. One of the earliest LCP films used in MEMS is a Vectra®
A-950 aromatic liquid crystal polymer, produced by Hoechst Celanese Corporation [23].The reported
melting temperature of Vectra A-950 is 2800C. The specific gravity ranges from 1.37 to 1.42 kg/m3 and

154
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

the molecular weight is greater than 20,000 g/mol.The compatibility of this LCP with commonly used
chemicals in micromachining was first investigated in [23]. LCP is virtually unaffected by most acids,
bases and solvents for a considerably long time and over a broad temperature range. Extensive tests
showed that LCP was not attacked or dissolved by at least the following chemicals common in
microfabrication: (1) organic solvents including acetone and alcohol, (2) metal etchants for Al,Au and
Cr, (3) oxide etchants (49% HF and buffered HF), and (4) developers for common photoresist and SU-8
resist. LCP films have excellent stability. It has very low moisture absorption and low moisture
permeability, which are better than PMMA and comparable to that of glass. For other gases, including
oxygen, carbon dioxide, nitrogen, argon, hydrogen and helium, LCP also exhibits above-average barrier
performance. Further, the permeation of gases through LCP is not affected by humidity, even under
elevated temperatures.The thermal expansion coefficient of the LCP material can be controlled during
the fabrication process to be both small and predictable.The LCP film also shows excellent chemical
resistance. LCP was originally used as a high-performance substrate material for high-density printed
circuit board (PCB) [23]. A number of unique processing methods have been developed for LCP,
including laser drilling and via filling (for low-resistance electrical through interconnects). LCP films
are used as substrates for space and military electronics systems, both for its performance and its
stability. For instance, it has been explored as a high-performance carrier of radio frequency
electromagnetic elements such as antennas. Results from high-frequency tests show that LCP has a
fairly uniform relative dielectric constant of 3 in the range 0.5 to 40 GHz and an extremely low loss
factor.
For LCP film with uniaxial molecule orientation, its mechanical properties are anisotropic and
dependent on the polymer orientation. For instance, the uniaxial LCP film can withstand less load in the
transverse direction (i.e., the direction orthogonal to the orientation of its molecules) than in the
longitudinal direction (i.e., the direction along the orientation of its molecules). If necessary, biaxially
oriented film with equally good transverse and longitudinal direction properties can be made to correct
the anisotropy behavior.A uniaxial film can be formed by bonding multiple layers of anisotropic films
with angular offsets of crystal orientations between layers. The orientation of LCP molecules varies
through the thickness of the film, while at the two sides of the film molecules are oriented at opposite
angles. If the angles are +450 and -450 at either side, the mechanical properties, such as coefficient of
thermal expansion, tensile strength and modulus are nearly isotropic. Commercial LCP material is
supplied in sheet format. The thickness of LCP film could vary from several microns to several
millimeters. Some can be provided with copper clad layers on one or both sides. The optional copper
clad layer is normally 15 to 20 m thick. This copper layer is laminated in a vacuum press at a
temperature around the melting point of LCP.

7. Describe the molecular structure and properties of PMMA. (Apr 2017)

PMMA
PMMA is supplied in many different forms, including bulk, sheets, and solutions for spincoating.
PMMA bulk, most commonly known by its trade name acrylics, has been used in making microfluidic
devices. The photodefinable PMMA thin film is a widely used e-beam and X-ray lithography resist.
Spin coated PMMA has been used as a sacrificial layer as well. Deep reactive ion etching processes for
PMMA thin films has been demonstrated.

155
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

8. Explain the principle of actuators for active optical MEMS with neat diagrams. (Apr 2017)
OPTICAL MEMS
Micro-optical systems presently have application in the following areas: Optical Scanning - Medicine
(minimally invasive surgery) - Small Space Inspection Optical Alignment Optical Communication -
Switching - Optical Filtering - Spectrometry (Wavelength Separation) -Adaptive Optics and Free Space
Communication.
Advantages:
Easy to manipulate light - Optical systems involve photon re-direction, therefore, since photons have
very little momentum, micro-actuators are easily able to manipulate photons. Simplified packaging -
Optical MEMS can be sealed in packages with transparent housings, which allows light to pass through
a glass window. - This allows the optical device to be protected from dust and harsh environmental
conditions, unlike flow sensors, or micro tactile sensors, or other MEMS where the source of
signal/phenomena must be in direct contact with the chip surface. Device size - Able to pack a large
number of devices into a relatively small area/volume.
Application – DLP digital light processing
The DLP micro-optical projection system was developed by Texas Instruments. It uses a matrix of
micro-mirrors to selectively switch reflected light, to form a projected matrix of pixels. Each mirror is
11 x 11 um in size, and a matrix may consist of 1024 x 768 mirrors, or more. By selectively switching
the mirrors on and off in the presence of a light source, the mirrors will reflect the light source ‘toward’
or ‘away’ the optical path that forms the projected image. To create the complete projection system, the
DMD (Digital Micromirror Device) is used with a light source, optics, a color filter and a projection
lens.

156
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

157
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

9. Discuss about optical lenses and mirrors. (Dec 2017)

MICRO LENSES
Lenses refract light for various purposes in optical systems. An ideal lens shape is shown in the Figure
below in (a). The ideal lens is difficult to fabricate directly using conventional micromachining, so a
number of variations are shown below, (b-e):

BINARY LENSES
The principle of operation of a Binary lens is as follows: Light passing through the ‘bulk material’
essentially moves in a straight line, therefore, if the ‘bulk’ is removed, the operation is essentially the
same.

158
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

Variable-focus micro-lens using an ‘expandable’ polymer membrane. As the ‘chamber’ is pressurized,

it fills with fluid and expands the membrane. The different expansion volumes have different radius of
curvature, and therefore different optical NA (numerical aperture).

Micro-Mirrors
Micro-Mirrors are a critical part of micro-optical systems, as they facilitate the re-direction of light
beams. A ‘Good’ micro-mirror has a number of important characteristics, such as: - High reflectivity
(low loss of light signal) - High flatness (low loss of light signal) - High stiffness (stays flat during
acceleration and deceleration) - Low mass (allows rapid acceleration and deceleration) Micro-mirrors
are used in: - DLP systems (Digital Light Projection systems) - Optical Switching (fiber optic
communication systems) - Scanning (minimally invasive tools with optical scan tips).
Micro-actuators for Micro-Mirror Systems
Electrostatic Based Micro-Mirror Actuators:

A MEMS device having a deformable mirror. In a representative embodiment, the MEMS device

159
Compiled by:Dr.P.Elamurugan,AP/EEE
Dept. of EEE MICRO ELECTRO MECHANICAL SYSTEMS EE6007

includes a deformable membrane supporting a plurality of light-reflecting segments that form the
deformable mirror. One or more actuators, at least one of which is configured to apply torque to a side
of the membrane, are used to deform the membrane. Membrane deformation causes the segments to
change orientation and thereby change the shape of the minor. A representative MEMS device of the
invention enables segment displacements in two directions and thereby realizes effective mirror
curvature values in the range from about −2 mm−1 to about +2 mm−1. A typical deformable mirror of the
prior art is a deformable membrane, the shape of which is controlled by an array of underlying
electrostatic actuators, e.g., electrodes. However, the deformation induced by such actuators is typically
unipolar. That is, the electrodes can flex the membrane in one direction (i.e., toward these electrodes)
only. As a result, the types of distortions correctable with prior art membrane mirrors are limited. In
addition, for a membrane mirror having a lateral dimension of about 100 μm, the maximum deformation
amplitude is on the order of 1 μm, which puts a corresponding limit on the range of distortions that can
be corrected. Although fabrication of MEMS devices of the invention has been described in the context
of using silicon/silicon oxide SOI wafers, other suitable materials, such as germanium-compensated
silicon, may similarly be used. The materials may be appropriately doped as known in the art. Various
surfaces may be modified, e.g., by metal deposition for enhanced reflectivity and/or electrical
conductivity or by ion implantation for enhanced mechanical strength. Differently shaped membranes,
segments, actuators, rods, members, and/or electrodes may be implemented without departing from the
scope and principle of the invention. Springs may have different shapes and sizes, where the term
“spring” refers in general to any suitable elastic structure that can recover its original shape after being
distorted. A different number and/or type of actuators may be used to deform membranes and/or
springs. Various MEMS devices of the invention may be arrayed as necessary and/or apparent to a
person skilled in the art.

10. Write short notes on PDMS. (Dec 2017)

PDMS
Elastomers are materials that can sustain large degree of deformation and recover their shape after a
deforming force. Poly (dimethylsiloxane) (PDMS), an elastomer material belonging to the room
temperature vulcanized (RTV) silicone elastomer family, offers many advantages for general MEMS
applications. It is optically transparent, electrically insulating, mechanically elastic, gas permeable, and
biocompatible.The biological and medical compatibility of the material is reviewed in. PDMS is widely
used in microfluidics. The primary processing method is molding, which is straightforward and allows
fast, lowcostprototyping.A number of unique process characteristics of PDMS are worth noting:
1. The volume of PDMS shrinks during the curing step. Compensation of dimensions at the design level
should be incorporated to yield desired dimensions.
2. Due to volume shrinking and flexibility, deposited metal thin films on cured PDMS tends to develop
cracks, affecting the electrical conductivity.
3. The surface chemical properties (such as adhesion energy) can be varied by altering the mixing ratio
and through surface chemical or electrical treatment. PDMS is commercially supplied as a viscous
liquid—it can be cast or spin coated on substrates. Unfortunately, the PDMS material is not photo
definable. It therefore cannot be simply spin-coated and patterned like photosensitive resists. Though
UV curable PDMS is being developed, the technology is not yet mature. It is possible to use plasma
etching to pattern PDMS thin films. However, the etch rate is rather slow.The measured etch rate is
approximately 7 nm/min at 800 W power and 100 V bias. Etching of PDMS with O 2 plasma leaves the
surface and line edges rough.

160
Compiled by:Dr.P.Elamurugan,AP/EEE