COLLEGE OF ENGINEERING KIDANGOOR

UN
((U DE
ND ER HE
R TTH CO
EC OP
O--O ER
PE VE
ATTIIV
RA AC
EA AD
CA EM
DE YO
MY PR
OFF P OFFE
RO ON
ESSSSIIO AL
NA ED
LE UC
DU CA ON
ATTIIO N)))
(CAPE)
Essttaabblliisshheedd bbyy G
E Keerraallaa
Goovvtt.. ooff K
DE
D PA
EP AR ME
R TT M EN O FF
N TT O
EL
E EC
LE RO
C TT R N II C
ON CO
& C
C SS & MM
OM UN
MU CA
N II C ON
A TT II O EN
N E NE
G II N
NG ER
EE NG
R II N G

THE MICRO OPTO ELECTRO MECHANICAL

SYSTEMS

SEMINAR PRESENTED BY

PIYA SUSAN PUNNEN

Department of Electronics & Communication Engineering

October 2004
 COLLEGE OF ENGINEERING KIDANGOOR
(UNDER THE CO-OPERATIVE ACADEMY OF PROFESSIONAL EDUCATION)
(CAPE)
ESTABLISHED BY GOVT. OF KERALA

DEPARTMENT OF
ELECTRONICS & COMMUNICATION ENGINEERING

Certificate
Date:

Certified that the paper entitled THE MICRO OPTO

ELECTRO MECHANICAL SYSTEMS is prepared &

presented by PIYA SUSAN PUNNEN of Final year B.Tech in
Electronics & Communication Engineering, towards the partial
fulfillment of the requirement for the award of the degree of Bachelor
of Technology, by the Cochin University of Science And
Technology.

GUIDE HOD CO-ORDINATOR

(DEEPAK V.K) (SEEMA A.K) (DEEPAK V.K)

 ACKNOWLEDGEMENT

First and Foremost I thank God Almighty for His divine grace and blessings
bestowed on us in completing this seminar in time.

I am grateful to our respected principal, Dr. K. Venkateswara Mallan, for

his guidance and support. I extend my deep gratitude to our internal guide,
Mr. Deepak V.K., Lecturer in Electronics & Communication, whose careful and
meticulous reviews improved the quality of this work.

I express my sincere thanks to Mrs. Seema A. K. the HOD of Electronics &

Communication for her support and guidance and thanks to the entire cast and crew
of Electronics & Communication for their co-operation.

Last, but not the least, I would like to thank all my friends and my parents
who supported me with their love and encouragement for the completion of this
seminar.

PIYA SUSAN PUNNEN

 ABSTRACT

MOEMS:-MICRO-OPTO-ELECTRO-MECHANICAL SYSTEMS

With the slogan that “size is not a barrier to advanced technology”,

surely MOEMS is here to stay. Imagine working with machines of the
size of sand which are batch fabricated thousands at a time costing only
a few pennies! This is the world of MEMS and its optically related sub
segment optical MEMS.

A MEMS device combines the computational functionality of the

integrated circuit with the sensory ability of a mechanical device. A
MEMS device can discern when something has happened and react to it.
And, like an integrated circuit, a MEMS device can also compute.

Components fabricated with the emerging technologies of micro-

electro-mechanical systems (MEMS) are being incorporated in an
increasing number of sensor and actuator applications. There is rapid
progress in optical systems for telecommunications to meet the needs for
increased bandwidth, for optical networks with terabit capacities per fiber
link, and for local area networks. Some applications require precise
features for optical alignment while others involve the precise movement
of small optical parts to achieve advanced functionality. In the MEMS
approach accurate, low-loss, optical connections are made between
different guided wave optical components including fibers, waveguides
and lasers. Mechanisms to allow motion and structures to provide
electrical actuation are combined in micro-opto- electro-mechanical
systems (MOEMS).

Today's optical networking equipment must switch from the

optical to the electronic domain.But with MOEMS switching is in the all-
optical domain and because it controls the flow of light rather than the
flow of electrons, it is ultimately faster, smaller and cheaper.
 The MOEMS switches and waveguides are made together on a
single crystal silicon wafer using widely available semiconductor
processing equipment. Such on-chip integration avoids the complex
alignment issues associated with manually connecting different and
larger components with optical fibers, and avoids the cost and space
associated with manufacturing, assembling and packaging the separate
components of Add/Drop Multiplexers.

Research has resulted in the fabrication of MOEMS that require

minimal interconnects, which increases reliability and also decreases
manufacturing cost - both important factors to the continued adoption of
MOEMS in biomedical, telecommunications, automotive and aerospace
industries.

The versatility, functionality, reliability and cost savings of

MOEMS technology will be a key advantage in many industries.
 INDEX
CHAPTER 1

INTRODUCTION Page no. 1

CHAPTER 2

FABRICATION & TECHNOLOGICAL CONVERGENCE 4

2.1 FABRICATION TECHNOLOGY FOR MOEMS

2.2 MEMS PACKAGING IN OPTOELECTRONICS 9

2.3 MEMS OPTICAL DEVICES 11

CHAPTER 3

ADVANTAGES OFFERED BY MOEMS 19

3.1 CURRENT CHALLENGES FACED BY MOEMS

CHAPTER 4

APPLICATIONS OF OPTICAL MEMS 24

4.1 SURVEY OF OPTICAL MEMS APPLICATION

4.2 OPTICAL SWITCHING 25

4.3 BIOPHOTONICS 30

4.4 ADAPTIVE OPTICS

4.5 FUTURE EXPANSION 31

CHAPTER 5

CONCLUSION 34

REFERENCE

APPENDIX
 CHAPTER 1

INTRODUCTION
Seminar Report - 2004 1

1. INTRODUCTION
1.1 MOEMS DEFINITION:

A MOEMS-Micro Opto Electro Mechanical System is a miniaturised

system combining optics and micro-mechanics, fabricated with collective
techniques derived from the microelectronic industry. These devices
generally offer reduced cost and improved reliability.

It was Physicist Richard Fenyman who theorized that size was not a
barrier to advanced technology, and conceived the idea of MEMS.
Microelectronic integrated circuits can be considered as the “brains”
of a system and MEMS augments this decision making capability with
“eyes” and “arms”,to allow Microsystems to sense and control the
environment.
MEMS is the integration of mechanical elements, sensors, actuators,
and electronics on a common silicon substrate through micro fabrication
technology. While electronics are fabricated using integrated circuit(IC)
process sequences, the micro mechanical components are fabricated using
micromachining processes that selectively etch away parts of silicon wafer
or add new structural layers to form mechanical and electromechanical
devices.MEMS technology has several advantages over current devices that

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 2

perform the same functions. They can be so small that hundreds can fit into
the same space as one macro device and perform the same function.
Cumbersome electrical components are not needed. In a MEMS device the
electronics can be placed on the same substrate, with the actuator and sensor.
This integration helps to reduce noise, thus improving the precision and
sensitivity of the sensors. Using IC processes thousands of devices can be
manufactured on a single wafer, this mass production greatly reduces the
price of individual devices.

MEMS brings together silicon based microelectronics with

micromachining technology, making possible the realization of a system-on-
a-chip. Hence allowing the development of smart products, augmenting the
computational ability of microelectronics with the perception and control
capabilities of micro sensors and microactuators. Sensors gather information
from the environment through measuring mechanical, thermal, biological,
optical, chemical phenomena. The electronics then process the information
derived from sensors and through some decision making capability direct the
actuators to respond by moving, positioning, regulating, pumping, filtering,
thereby controlling the environment for some desired outcome.

MEMS technology has advanced significantly in recent years and has

spawned a promising optically-related sub segment called
microoptoelectromechanical systems (MOEMS) or, commonly known as,
optical MEMS. These optical systems show great potential for use in several
industry segments, including displays, biomedicine, and
telecommunications.

Optical MEMS- the result of a technological convergence of computing

and information processing with photonics- are being increasingly used in
the development of lasers, and fiberoptics, which are providing the backbone
of telecommunications. Optical MEMS technologies are based on
micromachining- a process that is also used in IC (integrated circuit)
fabrication. Micromachining involves the manufacture of mechanical
structures in the micron to millimeter range.

In a nutshell, optical MEMS technology will be mainly driven by the

progress in optical communications. Photonic switches, scanners, displays,
and micromirrors will be the main devices that will enable optical MEMS to
penetrate large instrumentation and telecommunications markets. Various

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 3

MEMS components and sub-systems such as high-speed optical modulators,

reconfigurable wavelength-add/drop multiplexers, and optical cross connects
have been demonstrated for optical-fiber communications and are either
already commercialized or are on the verge of it.

The unique capability of optical MEMS to integrate optical,

mechanical, and electrical components on a single wafer allows for the
implementation of various key optical-network elements in a compact, low-
cost form.

However, there are also other promising areas for optical MEMS. As
these devices are essentially laboratories on a chip carrying out analysis on a
miniaturized scale, biomedicine seems to be an emerging area. The display
sector is also a prospective application segment, with use for optical MEMS
in projection displays, portable communications devices, other handheld
devices, and instrumentation displays.

Since cost and reliability are always key issues in the

commercialization of any technology, researchers throughout the world have
been attempting to make these devices without using any macroscopic
mechanical parts that might lead to higher costs and unreliability. New ways
to overcome transmission obstacles and provisioning and restoring network
traffic in units at roughly the wavelength level also have to be found. Once
solutions to such problems are found, the field of optical MEMS will grow
by leaps and bounds.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

 CHAPTER 2

FABRICATION & TECHNOLOGICAL CONVERGENCE

Seminar Report - 2004 4

2. FABRICATION AND TECHNOLOGICAL CONVERGENCE

Components fabricated with the emerging technologies of micro-

electro-mechanical systems (MEMS) are being incorporated in an increasing
number of sensor and actuator applications . There is rapid progress in
optical systems for telecommunications to meet the needs for increased
bandwidth, for optical networks with terabit capacities per fiber link, and for
local area networks. Some applications require precise features for optical
alignment while others involve the precise movement of small optical parts
to achieve advanced functionality. In the MEMS approach accurate, low-
loss, optical connections are made between different guided wave optical
components including fibers, waveguides and lasers. Mechanisms to allow
motion and structures to provide electrical actuation are combined in micro-
opto- electro-mechanical systems (MOEMS). Many MOEMS components
process light in miniature free-space optical beams, and the performance
scaling laws are often different from those for guided wave optical systems.
The silicon and other materials and technologies used to create MEMS
devices are discussed in detail, as are example applications in packaging
such as fixed and demountable connectors and methods for hybrid
integration.

2. FABRICATION TECHNOLOGIES FOR MOEMS

The sets cross-sectional diagrams shown in Figure 2.1 illustrate the two
main processes used for micromachining which are 1.Silicon Bulk
Micromachining 2.Polysilicon Surface Micromachining. The first
established process is silicon bulk micromachining using alkaline solutions
which have a much smaller etch rate of the (111) crystallographic plane in Si
compared with other planes. In this process, a mask is first patterned in an
etch-resistant surface layer such as thermally grown SiO2 or deposited Si3N4,
as shown in Row 1 of Fig.2.1. The silicon is then etched. Since the (111)
planes etch the slowest, V-shaped grooves are produced by etching standard
(100) oriented substrates to termination. V-grooves can be used for the
precise positioning of optic fibres as shown in Figure2.2. The features may
be hundreds of micrometers deep and precisely defined by the initial planar
lithographic process. Even grooves with diamond shape cross sections are
made, although the range of possible shapes is restricted by the

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 5

characteristics of the etch process. Encapsulated structures are made by

fusion bonding of glass to bulk micromachined wafers, and multilayer
structures are built up by bonding several silicon wafers together. Suspended
structures can also be made by undercutting of etch-resistant features.
Suspended structures can also be made by undercutting of etch-resistant
features.

Fig 2.1 Sequences of schematic cross sections showing bulk micromachining and
polysilicon micromachining for MEMS.

Polysilicon surface micromachining, exploits differences between

deposited polysilicon and silica layers to form three-dimensional features as
seen in Fig.2.1 row 2. The process is adapted from conventional silicon
integrated circuit technology, and the chemical vapour deposited (CVD)
polysilicon mechanical layer is typically 2 µm thick. The underlying silica
sacrificial layer is later wet etched away to produce a free-standing
polysilicon MEMS layer. The polysilicon layer can be incorporated into a
wide variety of sensors and actuators such as electrostatic comb drives. By
careful control of the CVD process conditions, the stress in the polysilicon
can be made reproducibly low. However, the thickness of the deposited
layers is limited to a few tens of micrometers by cost considerations, and by
the mechanical and electrical properties of deposited polysilicon which are
inferior to those of single crystal Si. The cycle of deposition, patterning and
etching of each material can be repeated several times to build up multilayer
structures, and feature shapes can be arbitrary. Foundries operate semi-
standard processes with several levels of polysilicon.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 6

Figure 2.3 shows more advanced micromachining processes which are under
development. When the fabricated structures must be thicker than is
achievable using polysilicon, an alternative surface micromachining process
is to use lithographic exposure of thick photoresist, followed by
electroplating, to form the mechanical parts as illustrated in Fig.2.3 row 1.
The original Lithographie, Galvanoformung, Abformung (LIGA) process
uses synchrotron radiation to expose the resist. The short X-ray wavelength
allows deep (up to 1 mm) resist layers to be exposed without significant
diffraction effects, and high aspect ratio structures are made. The released
metal layer can be used in a variety of MEMS applications including
packaging. Cheaper alternatives under development use ultraviolet mask
aligners to exposure special resist and achieve features several hundred
micrometers thick with aspect ratios of order 15. Parts are usually
electroplated in nickel and, after removal of the resist, they can be replicated
in other materials. Only low temperatures are needed, so LIGA can be used
as post processes to add microstructures to CMOS (Complementary Metal
Oxide Semiconductor technology).

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 7

Fig 2.2 Elastic fixtures for optical components: a) secondary electron micrograph
showing silicon nitride clips holding an optic fibre in place in a silicon V-shaped groove;
b) schematic of a proposed new approach to fixtures which avoids V grooves; c) silicon
nitride beams made by bulk silicon micromachining.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 8

A recently developed process for forming suspended single crystal

silicon structures is based upon the use of bonded silicon-on-insulator (SOI)
starting material, which is available as a by-product of the high-performance
silicon integrated circuit industry. As shown in Fig.2.3 row 2, the starting
material is a silicon wafer that is thermally bonded to an oxidised silicon
substrate. The bonded wafer is polished back to the desired thickness,
usually in the range 5 µm - 200 µm. The bonded layer is then structured by
deep reactive ion etching (DRIE) using an inductively coupled plasma etcher
and specific gas chemistry to obtain high etch anisotropy. Movable
structures may be made by removal of the buried oxide. There are many
MEMS applications including micromechanical shutters for optical
switching.

All of the above processes involve surface patterning and the resulting
structures are quasi three-dimensional. Fully 3-D microstructures are made
in polysilicon and in SOI by rotating surface micromachined parts out of the
wafer plane, and latching them into position as seen in Fig.2.3 row 3. The
parts are held by micromachined staple hinges. Assembly is usually manual,
but mass-parallel powered assembly has been demonstrated by differential
shrinkage of a polymer due to surface tension forces. Surface
micromachined engines are also used to push parts out of plane. The many
potential MEMS applications include components for free-space MOEMS.

2.2 MEMS PACKAGING IN OPTO ELECTRONICS

It is well known that passive alignment features for connecting single-

mode optical fibres can be made by anisotropically etching single crystal Si.
As shown in Fig.1, etching of (100) oriented Si through an appropriate mask
is used to make V-shaped grooves which act as kinematic mounts for optic
fibres. When two fibres are aligned in a groove and then butt- coupled
together, all degrees of freedom except uniaxial motion are fixed. The
assembly may then be epoxied. Alternatively, as shown in Fig.2.2, a fibre
may be held by flexible Si3N4 cantilevers . Single crystal silicon cantilevers
have also been demonstrated . Demountable connectors for ribbon optical
fibres requiring the simultaneous connection of many cores are made using
sets of etched V-grooves in Si substrates. On the male half of the connector,
two large grooves are used to locate a pair of precision steel pins, which
mate with corresponding grooves on the female half. The joint is made by
aligning the pins, and sliding the connector halves together. V-grooves can

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 9

also be used to construct simple subsystems known as opto-hybrids such as

the connection between an optical fibre and a photodiode. Systems with
detectors, lasers and high speed interconnects are used as transceivers.

Precision mechanical components are also used to connect different

types of waveguide. Particular attention has been given to the optic
interfaces between fibre and silica- on-silicon integrated optical components.
Silica-on-silicon has several major advantages over the earlier technologies
such as ion-exchanged glass, Ti : LiNbO3, GaAlAs and InGaAsP. Because
of this, silica-on-silicon components are likely to find widespread
application in low- cost systems such as fibre-to-the-home. A typical device
comprises a single-crystal Si substrate with fibre alignment grooves, a thick
buffer layer of silica to isolate the guided mode from the substrate, channel
guide cores formed from doped silica, and a thick over-cladding to bury the
cores. When combined with MEMS fabrication, devices may be constructed
with a large number of ports.

Fig 2 3 Cross sections showing metal microforming using thick resist; deep reactive ion
etching of silicon on insulator starting material; and three dimensional microassembly
processes for optical MEMS.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 10

Silica-on-silicon is also being used as a platform for the hybrid

integration of optical subsystems containing components such as lasers and
detectors that cannot be fabricated in the glass itself. Buried silicon terraces
or other precision features are used to locate the components accurately in
the vertical plane, and mesas are etched into the glass to provide lateral
location. For example, a hybrid-integrated 1.3 µm / 1.5 µm transceiver based
on a terraced silica-on-silicon platform is now available commercially. The
chip contains waveguide circuitry, an embedded dielectric filter for
demultiplexing, a photodiode receiver, and a laser diode source with monitor
photodiode.

Many new assembly techniques are under development. One proposed

geometry is shown schematically in Fig.2.2 where a substrate carries elastic
flexures on both sides to secure an inserted component in position . In this
arrangement the flexure length L is much greater than the clearance a and
the force F acts essentially at the point of initial contact. The precision with
which an optical component can be located is determined primarily by the
lithography and etching technology used for the deposited thin film flexures.
In contrast to V- groove technology, the precise geometry of the etched
substrate is of secondary importance. A possible candidate for the microclip
material is silicon nitride. A micrograph of chemical vapour deposited SiN
beams 3 µm thick is shown in the lower part of Fig 2.2 where the silicon
substrate has been wet etched to form a V-groove. Silicon carbide is another
possible material which has the required large Young's modulus and low
stress. Preliminary experiments on 5 µm thick CVD SiC carbide films and
etched Si substrates indicate that the material is suitable . In the longer term,
active flexures would be useful to allow for the adjustment of inserted
components.

An alternative alignment method exploits LIGA to construct complex

components such as multi-axis flexures in Ni metal. Several flexures
soldered to a breadboard provide individual mounts for components in an
optical component train. For example, Figure 2.4 shows a flexure for an
optical fibre and a collimating lens. After assembly, the optical throughput is
optimised by adjusting the position of each component in turn, using plastic
deformation of the flexure by an assembly robot. Complete optical systems
such as a channel monitor based on a tunable Fabry-Perot filter have been
assembled.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 11

Optical breadboards are also constructed using components that are

fabricated flat by surface micromachining and then rotated out-of plane.
Small, free-space optic beams propagate parallel to the surface of the chip
through fixed and movable lenses and gratings, and reflect off micro mirrors.
The structures are hand-assembled using precision micromanipulators. Other
systems have been assembled by using surface-micromachined engines to
push hinged devices out of the wafer plane. A typical assembly engine is a
scratch- drive actuator which acts as a controllable linear translator, allowing
dynamic component positioning. Fully automatic techniques are also being
developed for assembly of 3D MOEMS. One process based on surface
tension force simply requires the melting of small pads of material to rotate
parts out-of-plane,as shown in Figure 2.5. This method allows angular
positions to be set to an accuracy of minutes of arc. Demonstrated devices
again include fixed and movable mirrors, and lenses. Surface tension has
been used to position and fix the parts of the bonded silicon support frame in
Fig.2.5. The lenses are formed simultaneously by using reflow moulding of
photoresist which gives optical surfaces.

2.3 MEMS OPTICAL DEVICES

Applications for MOEMS in telecommunications were inspired by the

emergence of a successful device in a different field: the digital micromirror
projection display. On the display chip an array of more than a million
electrostatically actuated torsion mirrors rotate to-and-fro against a system of
end-stops to reject or accept individual pixels for the display. The
advantages of MOEMS are small size, low cost, and the ease with which
large-scale systems may be integrated. Compared with interferometric
guided wave devices, the advantages are low loss, low cross talk, scalability
and insensitivity to process variations. These advantages can outweigh
disadvantages such as the relatively slow speed of a mechanical switch or
tuning structure.

Most MOEMS devices have been constructed either in polysilicon or in

single-crystal silicon using bonded silicon-on-insulator, although III-V
MOEMS are now being developed. The advantages of polysilicon surface
micromachining are the existence of foundry-standard processes and the
ease with which complex multi-layer devices may be constructed. The
advantages of SOI are often improvements in structural thickness (and hence
the flatness of released surfaces) and low surface roughness.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 12

2.3.1 Optical switches

Switches can be based upon the electrostatic deflection of a waveguide

for integrated optics. 1 x 2 silica-on-silicon moving waveguide switches
have been demonstrated with the input channel guide fabricated on a
cantilever suspended over an etched cavity. Using surface electrodes, the
cantilever is deflected electrostatically from side-to-side, to allow connection
to either of the output waveguides. The switching time of the order of
milliseconds is long compared with the data rate, but is sufficient for
network reconfiguration.

Other switches are based on the insertion of small mirrors into the
nodes of a classical cross-point geometry. Suitable devices can be fabricated
by deep reactive ion etching silicon. DRIE can simultaneously form vertical
mirrors, fibre alignment grooves and a simple electrostatic drive to remove
or insert the mirrors into the regions where the optical axes intersect. A
typical structure is shown in Figure 2.6. Switching times are also of the order
of milliseconds, and optical isolation is excellent. However, the devices are
only scalable to small-size switch arrays.

2.3.2 Cross-Connects

Dynamically assembled 3D MOEMS are being used to construct larger

mirror insertion cross-connects. A commercially available polysilicon switch
uses an 8 x 8 array of small mirrors to form a 64-node cross-point. The
mirrors are driven between two states by an electrostatic drive acting
through a linkage. Switches of this type are now well developed and reliable,
despite their apparent mechanical complexity. Two-dimensional mirror
farms can suffer from increasing diffraction losses as the size of the array is
increased. Alternative mirror plane switches with a larger number of ports
are now under development. Instead of using deflection between a set of
single-axis mirrors, whose positions are effectively binary, the devices
operate by routing beams between two arrays of continuously adjustable,
dual- axis mirrors. Using surface micromachining Lucent Technologies
Inc./Bell Labs have recently reported a large (238 x 238) switch fabric with
a mean fiber-to-fiber insertion loss of 1.33 dB . The maximum insertion loss
for the 56644 possible connections was only 2.0 dB. Two fiber arrays with
collimating microlenses attached are used to project and receive beams from

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 13

two 2-dimensional arrays of micro-mirrors. Figure2.7 shows an

electrostatically actuated 2-axis tilt mirror similar to the mirrors used in the
large array .

Fig 2.4 Electron micrograph showing the hybrid integration of optical components using a
LIGA-type flexure (http://www.axsun.com/).

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 14

Fig 2.5 Scanning electron micrograph of a microlens array after surface tension self-
assembly .

2.3.3 Optical Attenuators

Mirror insertion devices are also being used as variable optical

attenuators, for example for channel equalisation in dense wavelength
division multiplexed systems involving erbium doped fibre amplifiers. A
small mirror on a cantilever is inserted into the optical path to block part of a
collimated beam, the reflected component being dumped into an absorber.
Alternatively, switches may be based on a microfluidic systems and the
controlled movement of gas bubbles in liquid channels.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 15

Fig2. 6 Electron micrograph showing a linear translation mirror switch fabricated by

deep reactive ion etching (http://www.sercalo.com/).

2.3.4 Add-Drop Multiplexers

Linear arrays of tilt mirrors have also been used to construct ADD-
DROP multiplexers. A lens is first used to collimate the output from a fibre
carrying a number of different wavelengths. A diffraction grating is used to
disperse the wavelengths into a fan of beams, which are then passed through
another lens to form spatially separated foci. The mirror array is placed at
the focal plane, so that each wavelength falls on a separate mirror. Each
mirror may then either retro reflect the light back through the multiplexer, or
redirect it to a second multiplexer and fibre. Consequently, the spectral
compositions of two output channels may be arbitrarily selected.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 16

Fig 2.7 Two axis beam steering micro-mirror built using surface micro-machined
technology [8], (Lucent Technologies Inc./Bell Labs).

When an electrostatically-deflectable membrane mirror is combined with a

second, fixed mirror, the result is a mechanically tunable Fabry-Perot cavity.
This structure forms the basis of the mechanically actuated anti-reflection
switch used as a modulator and filter. A more complex structure with a
segmented drive electrode is used as a deformable membrane. This device
has been used as a spectral plane filter, controlling the reflectivity as a
continuous function of wavelength to achieve spectral equalisation of the
gain of a fibre amplifier.

2.3.5 Tunable Lasers

MOEMS components have been used as the basis of tunable external

cavity lasers. Gain is provided by a conventional semiconductor optical
amplifier, frequency selection is achieved using a reflection grating, and
tuning is carried out by rotating an external mirror about a remote pivot. The
actuator is a comb drive structure formed by deep reactive ion etching of
bonded silicon-on-insulator material.

Highly integrated vertical cavity semiconductor lasers with an external

cavity formed by a small movable mirror are also under development. They
may be optically pumped with a membrane as the external mirror. Another
electrically pumped device is grown in one step and etched to create a

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 17

cantilevered tuning arm. Tuning is achieved by applying a small voltage to

he top mirror, which causes the cantilever to move up or down, the
cantilever motion altering the laser cavity length and changing the
wavelength of operation.

2.4 Stiction Problem

A major concern with MEMS is stiction, or the propensity of two

silicon surfaces to stick to each other if they touch. Stiction can be a problem
during the wafer fabrication, during subsequent assembly and test and in the
field. During wafer fabrication, stiction is an issue during the critical release
phase of the device when the mechanical structure is finally released and is
free to move. Release methods that use wet-etches have the problem that the
surface tension of drying droplets will pull two surfaces together as they
evaporate eventually bringing the surfaces into contact. This manifests itself
as a yield issue to the supplier and a delivery issue to the customer.
Manufacturers must develop specialized release processes to solve the
problem and successfully commercialize a component.

Once the device is released it will face a shock and vibration

environment during assembly, test and shipment. MEMS devices are very
strong and their low mass makes them more robust than larger mechanical
systems. Still much of the standard equipment used for die processing and
assembly is not optimized to address the shock and vibration environment
required by MEMS, as it is of little consequence to integrated circuits.
Normal handling can produce 100's or 1000's of g's of shock anytime the
MEMS package comes in contact with another hard surface and can cause
surfaces to touch. Therefore producers of MEMS must assume silicon
surface will touch and must develop 'Anti-Stiction' coatings that prevent the
two surfaces from sticking. These can be thought of as a passivation for
mechanical devices analogous to the use of oxynitride passivation in the IC
world. A robust anti-stiction coating is a requirement in producing a high
reliability MEMS device.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 18

2.5 Future Prognosis

The key to new developments in optical MEMS is to successfully apply

the underlying physics and develop cost-effective processing to make high-
precision components. New materials are being used in the silicon integrated
circuit industry where there is a huge on-going investment, and the much
maller MEMS activity can benefit from these developments. Examples are
(i) SOI technology as seen in Fig.2.3 whereby low stress mirrors and other
MEMS components are made with reproducible mechanical properties and
excellent control of planarity, and (ii) low stress Si3N4 which has a high
Young's modulus and has promising applications in packaging as seen in
Fig.2.2. The current MEMS product containing the largest number of
micromechanical components is the mircromirror projection display.

As integrated circuit lithography and reactive ion etching techniques

improve in accuracy, the necessary precision for optical MEMS can be
obtained in a manufacturable way. There is a strong incentive to stay with
silicon substrate technology for MEMS packaging to piggyback on this
further investment, but within this paradigm there is great scope for applied
physicists in both the invention of device structures and the development of
materials processing.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

 CHAPTER 3

ADVANTAGES OFFERED BY MOEMS

Seminar Report - 2004 19

3 ADVANTAGES OFFERED BY MOEMS

MOEMS offer all the commonly cited potentialities of Optics:

• Accuracy in distance measurements

• EMI insensitivity in harsh environments
• Security in exploding fields
• Selectivity in chemical detection

Together with the features of micro-system technology:

• Miniaturisation :Miniaturization of the system brings about

ruggedness.
• Reliability –Precision & hence improved performance.
• Mass and size reduction –Portability.
• Batch Fabrication -Mass production capability.
• Low manufacturing costs .
• Less Manufacturing time.
• Low Power consumption.
• Massively Deployed.
• Easily maintained by replacements.
• Cause little harm to environment.

3.1 CURRENT CHALLENGES FACED BY MOEMS

3.1.1 The promise of MOEMS

MOEMS offers the promise of building more than 1000 silicon mirrors
in an array using integrated circuit batch manufacturing techniques as
opposed to manual assembly. Silicon is a well-understood and strong

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 20

material that has ideal properties for mechanical devices. However, using
silicon in optics faces a number of challenges before these devices become
widely accepted. In addition to mechanical properties, optical MEMS
devices add requirements for optical properties. For example, MEMS
devices used in optical cross connects require optical quality surfaces with
highly reflective oatings, the ability to precisely control position of the
mirrors, and the capability to route a large number of MEMS control signals
on and off the chip.

MEMS must face these challenges while also meeting the 5-9's
reliability required in telecom systems; there are several challenges facing
the commercialization of MEMS in telecom systems. First, despite all of the
hype about MEMS there are actually only a few MEMS applications,
(pressure sensors, ink jet heads, projection displays and accelerometers), that
have been successfully commercialized, and the number actually shipping
into telecom systems are nil. Most of the expertise and commercial volume
for MEMS devices resides with a handful of companies that have made the
investment to solve the design, manufacturing, quality and reliability
problems inherent in producing MEMS. Second, optical MEMS components
will be manufactured in very low volumes when compared with the average
integrated circuit device or process. To make matters worse, each optical
device may require its own specialized custom process, further reducing the
volume produced. This means that the yield, quality and reliability issues
that tend to surface as material is manufactured in volume may lie
undiscovered until a problem crops up in the field.

The solution to solve most of the yield, quality and reliability issues of
MEMS requires having already built higher volume MEMS products and to
have optical MEMS fabrication processes with the flexibility and versatility
to allow many different types of optical MEMS devices to be designed and
built—enabling an the accumulation of volume needed for yield and quality
improvement on a single fabrication process.

3.1.2 Design Challenges Of Cross Connects

A free space, 3-D MEMS optical cross connect works by aligning an

array of collimated input fibers to a set of mirrors that can re-direct the light
by tilting the mirror in X and Y . A second array of collimated output fibers
is aligned to a 2nd set of mirrors. By precisely aiming the mirror on the input

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 21

fiber so that it projects to the mirror assigned to the desired output fiber a
light connection can be made. The connections must be made in a short
amount of time, (>10ms) to allow network protection functions. The mirror
must remain extremely stable over a long period of time and have the ability
to point with a high degree of resolution.

A practical MEMS mirror array requires the convergence of optical,

mechanical and electrical design spaces in order to work. Whereas the
traditional integrated circuit design must pay attention only to electrical and
process parameters, design of mirror array components adds additional
mechanical and optical constraints . It is, therefore, truly a system on a chip.

The specification requirements in the optical domain trade off against

specifications in the electrical and mechanical domains, requiring an
optimization of all three, leaving a narrow solution space to the system
designer. For example, tilt angle range, which determines how large of a
mirror plane can be scanned, directly trades off against actuation voltage and
the resonant frequency of the mirror. Power density trades off against optical
flatness. The number of interconnects required trades off against the number
of mirrors possible per unit area and thus with tilt angle. By constraining two
of the three related specifications the third often becomes impossible.
Nevertheless solutions do exist.

An array of 1000 mirrors requires a minimum of 5000 bond wires.

Typical industry practice can achieve wire bond pitches of 100um for
hermetic packages, and thus 5000 interconnects would require 50cm of
perimeter just for bondpads, defining a square area 12.5cm on a side. This is
an enormous package by IC standards and an enormous die if one were to
ever consider building the array on a single chip. In addition there is the
problem of routing signals from the center of a large array to the periphery,
which trades off interconnect against mirror fill factor.

On-board electronics allow signal conditioning to be done on chip, such

as on-board position sensing, by using the same capacitive techniques that
are used to decode accelerometer devices. The advantage of on -board
electronics is not limited to large cross connects. Having actuation and
control electronics on board simplifies the assembly and reduces the overall
sizes of the optical system. The ability to do precision capacitive position
measurement has applications for many optical components including fabry-

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 22

perot filters, tunable lasers, small switches and other devices where position
feedback from a moveable element is useful.

3.1.3 MOEMS Manufacturing, Quality and Reliability

MEMS devices, although built using semiconductor equipment, are not

integrated circuits and have a completely different set of failure mechanisms
and reliability concerns. MEMS reliability focuses on mechanical failure
modes rather than electrical ones. For example, in an electronic circuit we
may be concerned with the Ft of the transistors, the supply current or the
electrical gain. In a mechanical system we are interested in analogous but
different parameters such as the mechanical resonant frequency, spring
stiffness, and the voltage to translation transfer function. Performance
characteristics are not derived from dopant levels and profiles but by the
ability to accurately control the X, Y and Z dimension of the structure, the
surface roughness and surface charge.

A major concern with MEMS is stiction, or the propensity of two

silicon surfaces to stick to each other if they touch. New types of reliability
testing must be used to test MEMS devices. Mechanical tests can confirm
resistance to mechanical shock, stiction, and other MEMS specific failure
modes. For example, drop testing is one of the best tools to confirm good
mechanical design and stiction performance. Devices are typically dropped
from 1 Ft to 3 Ft onto hard surfaces in a variety of orientations as a test of
field robustness.

3.1.4 Packaging

Packaging is sometimes an overlooked detail, but in fact is one of the

most difficult and expensive aspects of MEMS. MEMS devices contain
exposed moving parts that can be made non-functional or unreliable by the
presence of liquid, vapor, gases, particles or other contaminants. Unlike a
standard IC, it is not possible to clean a MEMS device once it has been
released. For this reason the MEMS wafers must be singulated (cut up into
individual die) and assembled using very specialized techniques. For
example during the singulation process, the MEMS die must be protected
from particulates and contamination such as saw slurry, particles generated
by laser scribing or from scribe and break. Dust from machines or people

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 23

landing on the wrong place can impede movement of a MEMS device, or

affect the electrostatic fields that govern its motion.

Package cleanliness acceptable for a standard integrated circuit is a

reliability concern for a MEMS device, again because particles and
contamination that do not affect operation of an IC interacts with the micro-
mechanical device. The package environment, including such issues as out-
gassing of die attach, presence of particles, moisture levels, and chemical
interactions with anti-stiction coatings, assembly temperature and other
issues all must be evaluated, considered and addressed in the quality and
reliability of a MEMS device. Because this is so critical, it is important to
package the MEMS devices in a controlled, hermetic, particle free
environment.

Every step from die preparation to package seal must be performed in a

class 100 clean room environment until the device is safely sealed in a
hermetic particle free environment. Clean room techniques normally
reserved for wafer fabrication must be extended for use in probe, die prep
and assembly areas. Thus the packaging of the MEMS device is as
challenging as building the MEMS die itself. Customers who purchase a raw
unpackaged die from a MEMS vendor and package the device themselves
are likely underestimating the difficulty of the quality and reliability
challenges involved.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

 CHAPTER 4

APPLICATIONS OF OPTICAL MEMS

Seminar Report - 2004 24

4 APPLICATIONS OF OPTICAL MEMS

Optical MEMS is rightly called as the backbone of telecom

industry.The various MOEMS devices include optical switches,
crossconnects, attenuators,add drop multiplexers etc.

4.1 Survey of Optical MEMS Applications

4.1.1. Micro Optical Sensors; Actuators; and Related Technologies

1. Optical Sensor Technologies
4.1.2. Industiral Applications Of Optical Mems
1. MEMS Technology Enables Sensitive Low-Cost Gas Sensor
Chip
2. Automotive Alert Systems Help Avoid Rear-End Crashes
4.1.3. Medical and Biomedical Applications of Optical MEMS
1. New Horizons for Medicine
2. Robotic Surgery
3. Lasers
4. Drug Delivery
5. Nanobiotechnology
6. Tissue Engineering
7. Novel Prosthetic Lens
8. Synthesis of Maskless Arrays
9. University Research on MEMS for Medicine
4.1.4. Military and Aerospace Applications Of Optical MEMS
1. Picosatellites Liftoff
2. Battlefield-Vision Applications
3. Weapons Command and Control
4.1.5. Other Optical Devices Using Optical MEMS
1. The Grating Light Valve
2. Diffractive Optical MEMS Use Grating Light Valve
Technique
3. Introduction to the GLV Technology
4. Silicon Light Machines
5. Dynamic Gain Equalizers

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 25

4.2 OPTICAL SWITCHING

Data-centric long haul and regional networks are beginning to be

redesigned around large all-optical cross connect switches that will allow
BW management, real time provisioning and mesh protection. A key-
enabling component is the all-optical cross connect switch that allows the
connection of 1000's of input wavelengths to 1000's output wavelengths
under electronic control. Such a switch does not care about protocol,
bandwidth, or the data rate of the optical signal. The new network will set up
dynamically configurable highways of data in a nation-wide network. The
ability to change the connections on the fly allows mesh protection schemes
to replace the current point to point methods and allows carriers to
dynamically add or delete bandwidth through multiple paths and through
multiple switch points to make better overall utilization of their available
capacity.

Central to the promise of the new network architecture is the physical

switch fabric. Micro Electro-Mechanical Systems, or MEMS, is one of a
number of technologies being considered for implementing optical switches
and is the leading candidate for building large scale all-optical cross
connects. Companies have gravitated to so-called 3D switch fabrics that use
individual mirrors to direct beams of light from arrays input fibers to arrays
output fibers by physically steering the beams of light in an analog fashion
from one fiber to another. This contrasts to the more common cross bar
switch found in both the electrical and optical domain that has one switch
element for every possible input and output connection. Optical switches of
this kind are known as a 2D or N2 switches because the number of switching
elements increases as the square of the number of ports and because the light
remains in a plane of 2 dimensions. When switches of this type get much
larger than about 32 x 32 the number of switching elements and the size of
the die become unmanageable and uneconomical.

3D architectures transmit beams of light in free space and require

extremely high accuracy and stability to work. This approach to switching
does not have an analogous architecture in the electronic world, the closest
being the free space scanning electron gun used in television sets. While
MEMS is the most attractive technology for solving the problems of 3D
optical systems it is not without risks or challenges.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 26

In the electrostatic actuation scheme shown in Figure 4.2, the mirrors

are double gimbaled, torsional mirror devices capable of tilting in X and Y.
Electrodes below the mirror are used to set up a parallel plate capacitor for
electrostatic actuation. Any voltage between the mirror electrode and the
bottom electrodes creates an attractive force that is proportional to the area
of the plates, the voltage difference, and the distance between the plates as
follows (and ignoring physical constants):

F is proportional to: A*V2/d2 where A = area of plates, V = voltage

difference, d = distance between the plates (see Figure 3).

Thus the closer the plates are together, the more force that can be
generated per volt. However, the distance between the plates also sets the
limit on how far the mirror can tilt before touching or snapping down,
setting up a fundamental trade-off. The tilt ranges and mirror sizes needed
for cross connect applications typically require actuation voltages greater
than 100V. However, because the actuation plates are capacitors, once the
charge is put on the plates it stays there except for leakage current.
Therefore, power consumption can be very low, especially when compared
to schemes such as magnetic actuation. This is important in large arrays
where many mirrors must be put in a small space and power density
becomes a concern. For example, if one were to attempt to put 1000 mirrors
in a space of 10 sq. cm and each were to require 25mw per mirror (typical of
magnetic actuation schemes), then the power density would be 2.5W/sq. cm
creating a difficult heating issue that must be dealt with. Contrast this to an
electrostatic design that may require only 0.2mW per mirror or 20mW per
sq. cm.

Complicating the discussion is the problem of interconnects. A simple

open loop actuation scheme for a X-Y gimbaled mirror using electrostatic
actuation requires a minimum of 5 electrical connections per mirror. Thus an
array of 1000 mirrors requires a minimum of 5000 bond wires. Typical
industry practice can achieve wire bond pitches of 100um for hermetic
packages, and thus 5000 interconnects would require 50cm of perimeter just
for bondpads, defining a square area 12.5cm on a side. This is an enormous
package by IC standards and an enormous die if one were to ever consider
building the array on a single chip. In addition there is the problem of
routing signals from the center of a large array to the periphery, which trades
off interconnect against mirror fill factor.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 27

Figure4.1. Schematic diagram of a 3-D MEMS optical switch

Important Design Specification in the Optical, Mechanical and Electrical

Domain
Optical Mechanical Electrical/Packaging
Mirror Material Flatness Actuation Voltage and Current
- Optical Quality Required
- Reflectivity
- Flatness
Tilt Angle Range Resonant Frequency # Interconnects Required
Position Resolution and Power density (Watt/Sq cm)
Mirror Size
Repeatability
# of mirrors per unit
Tilt Angle Range
area

Figure 4.2. Top view of a micromirror showing Figure 4.3. Side view of mirror showing electrostatic plates
electrodes under mirror layer area, plate-to-plate distance, and applied voltage

The vast expansion in the use of the Internet over the past two years
resembles the upsurge in tourists traveling to an increasingly popular resort.
Traveling along the main highway to the resort at the start of a summer

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 28

weekend, the stream of tourists experiences a series of bottlenecks whenever

the six-lane highway narrows down to four lanes to cross bridges or pass
other obstructions. Internet users face similar kinds of bottlenecks. These
take the form of hub regions that must switch pulses of light moving rapidly
and efficiently along optical fibers onto new routes. At those hubs, the
system must convert the light pulses into electronic packages to permit them
to travel through the switch. Once rerouted, the packets are converted back
into light.

The conversion and reconversion slows down the entire process. Any
sort of electronic system at the end of an optical fiber will inevitably cause a
traffic jam because it lacks the capacity to deal with the fiber's bandwidth.
So far, electronic switches have managed to keep up with increasing
bandwidths; but as the amount of data sent over optical fibers is increasing
twice as fast as the capacity of electronic switching, that situation won't last
much longer.

The need for fast, reliable optical switches will only increase as
technologies such as dense wavelength division multiplexing (DWDM)
come into play. DWDM enables a large number of different data streams to
travel along a single optical fiber. Routing those streams in the correct
directions at hubs demands a technology more sophisticated than an optical-
electronic-optical system can provide.

The obvious solution is to avoid the transitions between optical and

electronic transmitters by using optical switches. Lucent and several other
companies are working hard to develop optical switching devices, largely
through the application of microelectromechanical and micro-
optoelectromechanical systems (MEMS and MOEMS). A strong demand
will be created for optical switching for optical networking,
telecommunications, and wireless technologies, as we enter the new
millennium. The solution is far from monolithic. Competitors in the race for
an effective optical switch have developed technologies that range from
microscopic mirrors to miniature bubbles to liquid crystals. Some switches,
notably those built by Lucent and Optical Micromachines Inc., have already
made their commercial debuts.

Mirror-based switches have drawn the most interest so far. Within the
past 18 months, at least three companies have developed optical switches

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 29

based on micromirror technology. Early last year, for example, Lucent's Bell
Labs built what it called "The world's first practical optical switching
technology using MEMS." The device is based on a tiny pivoting bar with a
gold-plated mirror at one end that fits in a tiny space between two hair-thin
optical fibers lined up end to end. When the switch is in the off position, the
mirror rests below the cores of the two fibers, permitting light signals to
travel across the gap from one core to the other. However, voltage applied to
the far end of the bar helps to lift the mirror between the fibers, where it
reflects incoming light rather than permits its passage. Another example of
Lucent's micromirror technology, its WaveStar LambdaRouter, uses 256
miniature mirrors, each no larger than the head of a pin, to steer light signals
from one optical fiber to another. This promises the ability to direct traffic
through optical networks 16 times faster than is now possible.

Early this year, Xros, Inc. announced "the world's highest capacity
optical cross-connect system for open optical networks," based on its own
micromirror technology. The system uses two facing 6 X 6-in. arrays, each
containing 1152 mirrors made out of pure silicon. Computer-controlled
electrical signals can tilt each mirror in several dimensions, allowing the
system as a whole to direct any incoming light beam (or light path) in more
than 1000 directions. That capability provides optical networks with extreme
redundancy, in case of breakages in the optical fibers.

Optical Micromachines Inc., meanwhile, has used its new line of optical
switches since February to route live traffic over an optical network in
Oakland, California. The company said, "This represents the first time
microelectromechanical subsystem-based switching systems have been
deployed to carry live data traffic." Like Xros's switches, Optical
Micromachines' are micromachined mirrors fabricated on silicon chips.

Other technologies rely less on MEMS and MOEMS. Agilent

Technologies, Inc., a company spawned by Hewlett Packard, uses bubbles to
shuffle light packets between tiny waveguides made of glass. Corning Inc.
and Chorum Technologies Inc. rely on liquid crystals to shift light beams
from one path to another. Nanovation Technologies, meanwhile, is
developing a hybrid technology that integrates MEMS-based optical
switches with silica-on-silicon waveguides on a single chip.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 30

One fact is common to the first generation of optical switch

technologies: Whatever current problems they solve, they will soon be
outdated. For while the switches effectively transfer packets of light from
one path to another, they can't figure out the directions that each incoming
light packet should take. That requires the ability to read rudimentary
messages that the light packets carry. At present, only microelectronic
systems have that ability.

Few MEMS researchers doubt that optical systems will eventually catch
up. By mimicking optical circuits, the researchers expect to devise
intelligent optical networks that integrate MEMS and waveguide
technologies. The ultimate goal of Internet transmission mediated entirely by
light is coming into sight.

4.3 BIOPHOTONICS

Another revolution visible is in biophotonics. It has been recently

demonstrated that light can be used not only for detection (microscopy,
fluorescence, or spectroscopy) but also for sample or macroparticle
manipulation. Classical microscopy will also soon be revolutionized as
several academic institutions are currently working on "on chip integrated
microscopy".Also several biophotonic capabilities, such as interferometry of
microfluidic cytometry, laser light scattering on cells with organelles, and
optical manipulation of cells and macromolecules (optical tweezers).

4.4 ADAPTIVE OPTICS

Optical elements typically mirrors that can adapt their optical

performance to compensate for poor optical quality of the light reflecting
from them come under adaptive optics.

DeformableMirrors:

Adaptive optics that have the ability to change the shape of a mirror
surface to optimize the reflected light for a specific application.

Adaptive optics and deformable mirrors have been around almost as

long as lasers. The use of adaptive optics has been limited by their
significant complexity, and correspondingly high cost. Adaptive optic

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 31

systems have typically cost over $100,000. These realities narrowed the
application of adaptive optics to only the highest priced lasers or imaging
systems. In the past, adaptive optics systems were hand built, and this lead to
very long lead times, significant expense, and unpredictable functionality.
Deformable Mirrors (DMs) provide sharper images and tighter laser beams,
even in a dynamic environment. For example, a long-range surveillance
camera with a telephoto lens will record a distorted, fuzzy image due to
wavefront aberrations caused by small variations in the atmospheric density
between the camera and the subject being viewed. An extreme example of
this effect can be observed above a blacktop road on a hot day. Adding a
deformable mirror to the lens system, these “wrinkles” can be measured and
removed to sharpen and clarify the image. The figure below illustrates this
application. We use a technique called bulk silicon micro machining to
fabricate its deformable mirrors. Adaptive optics can correct for optical
aberrations and imperfections in the human eye’s lens. Free-space optical
communications have a distinct advantage over radiofrequency
communications– wireless links cannot be intercepted, are un-jammable, and
low powered. However,they have two drawbacks beam divergence and
atmospheric distortion. Deformable mirrors can correct both problems in an
adaptive optics system. Applications include satellite communication space
to-space, space-to-aircraftand space to ground .

4.5 FUTURE EXPANSION

It’s no secret that tremendous growth of the Internet and burgeoning
telecommunications traffic are causing severe congestion in the overall
network structure. To address the ever-increasing need for bandwidth and
agility, networks will continue to grow in size and complexity.
As service providers eagerly look for new ways to generate revenue and
lower operating costs, many telecom companies are deploying WDM and
DWDM network topologies and transport technologies to boost the
efficiencies of their networks. Currently, much of today’s telecom
infrastructure is based on a foundation of optoelectronics equipment.
Because of the bottlenecks created by optical-to-electrical conversions in
today’s network, industry experts agree that most of the electronics will be
taken out of network signal processing and routing applications and replaced
with more efficient optical technologies that are protocol-independent and
have much higher bandwidth and wavelength agility.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 32

With the advent of fiber-optic networks-a major part of the

unprecedented growth in communications systems-the deployment of all-
optical networks is clearly the ultimate goal for the next generation of
telecom networks. But for very agile all-optical networks to become a
reality, increasing carrying capacity and wavelength tunability are critical
elements. That requires development of terabit routing and very flexible
wavelength routing devices such as optical crossconnects, optical
multiplexers, and wavelength translators.
Compliant MEMS technology
Recently, a new class of MEMS devices started to surface-components
made with highly compliant polymeric materials as a principal design
element. Typically in MEMS devices, electrostatic or piezo forces are used
to bend or deflect one of the MEMS layers to produce the desired
mechanical motion. Stiffer materials require higher voltages to achieve a
given mechanical deflection, and traditional silicon-based materials are all
extraordinarily stiff.
The compliant MEMS (CMEMS) technology platform differs from
conventional MEMS by adding a set of softer, more compliant polymeric
materials to the list of conventional rigid silicon-based materials used in
MEMS. Adding these compliant materials to the list of substances that can
be deposited, patterned, and etched greatly widens the design space.
The compliant materials used in this new class of MEMS devices are as
much as six orders of magnitude less stiff and can be easily tailored over a
range of three orders of magnitude. Additionally, they can be deposited in a
much broader range of layer thickness. This very wide range of flexibility
expands the design space for MEMS devices far beyond what is possible
with traditional silicon-based materials.
This CMEMS platform enables the development of components and
assemblies that demonstrate superior performance and lower costs compared
with using the silicon-based MEMS solutions currently available. Its unique
architecture is ideal for a broad range of high-performance tunable photonic
components demanded by the telecommunications industry, where it offers
several significant benefits over traditional MEMS-based designs
Optical performance monitoring
As noted earlier, the emergence of DWDM and all-optical networks has
created the need for a number of tunable devices. Key among these is the

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 33

optical performance monitor (OPM). As network providers move toward

higher data rates and denser channel spacing-allowing use of hundreds of
channels in a single fiber-performance monitoring becomes increasingly
critical. The more wavelengths crowded together, the greater the chance that
individual optical signals will be lost or degraded. For a service provider,
that can be a financial nightmare to the tune of up to $10,000 per minute per
channel!
Thus, a pressing need currently exists in DWDM networks to measure,
in the optical domain, the performance and health of all channels on a fiber.
The vital element of the OPM is the tunable filter that must select the desired
DWDM channel for quality measurements such as power, signal/noise, and
actual wavelength.
CMEMS technology is very robust and elegantly simple in its design,
making it both reliable and less expensive than the alternative solutions
available today. Its first application is in a new tunable Fabry-Perot filter-
currently being developed for the OPM market-that uses an elastomer to
suspend the movable mirror. This design offers better performance, higher
yields, and lower costs than those fabricated with conventional silicon-based
materials

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

 CHAPTER 5

CONCLUSION
Seminar Report - 2004 34

5.CONCLUSION

By mimicking optical circuits, the researchers expect to devise

intelligent optical networks that integrate MEMS and waveguide
technologies.The obvious solution to the bottlenecks faced by internet users
is to avoid the transitions between optical and electronic transmitters using
optical switches. The ultimate goal of Internet transmission mediated
entirely by light is coming into sight.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 35

REFERENCES

1. Optical MEMS(MOEMS) & Biophotonics.html

2. MEMS enables fast,reliableoptical switching.html
3. Lightwave-All optical MEMS continue to evolve.html
4. Optical MEMS abstract.html
5. Silicon Technology for MEMS –D.F MOORE& R.R.A SYMS
Europhysics News (2003) vol 34
6. Tech supporting MEMS tech in optical switching.html
7. Techonline _optical MEMSpromises BigBuck in Microspace.html
8. Optical Switching Technology Comparison:Optical MEMS vs.Other
technologies-IEEE Optical Communications-Nov -2003.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 36

APPENDIX

MEMS enables fast, reliable optical switching

by Peter Gwynne
The vast expansion in the use of the Internet over the
past two years resembles the upsurge in tourists
traveling to an increasingly popular resort. Traveling
along the main highway to the resort at the start of a
summer weekend, the stream of tourists experiences a
series of bottlenecks whenever the six-lane highway
narrows down to four lanes to cross bridges or pass
other obstructions. Internet users face similar kinds of
bottlenecks. These take the form of hub regions that
must switch pulses of light moving rapidly and
efficiently along optical fibers onto new routes. At At the heart of Lucent
those hubs, the system must convert the light pulses Technologies' WaveStartm
into electronic packages to permit them to travel
Lambda Router, an all-
through the switch. Once rerouted, the packets are
converted back into light. optical lightwave-routing
device, is an array of
microscopic mirrors, each
able to tilt in various
directions to steer light. The
micro-mirrors route
information in the form of
photons to and from any of
256 input/output optical
fibers.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 37

Optical MEMS (MOEMS) & Biophotonics

optical MEMS, and optical micromirrors in particular, have had staggering success in visual
displays, first demonstrated by Texas Instruments. This technology has been extended to
optical communication where fast switching of laser light is enabled with micromirrors,
microshutters, or microbubbles. The communication industry will soon see MEMS devices
as tunable lasers (VCSELs), resonators, interferometers, and multiplexers.

Another revolution will soon be visible in biophotonics. It has been recently demonstrated
that light can be used not only for detection (microscopy, fluorescence, or spectroscopy) but
also for sample or macroparticle manipulation. Classical microscopy will also soon be
revolutionized as several academic institutions (e.g. U.C. Berkeley BSAC) are currently
working on "on chip integrated microscopy".

As with other MEMS devices, computational multiphysics modeling will play a major role
in development and commercialization of this technology. CFDRC has the unique expertise
and research tools needed for the design of new generation MOEMS and Biophotonics.

CFD-ACE+ offers multiphysics modeling capabilities for both:

• MOEMS - mechanics, dynamics, vibration, air damping, electrostatics, electromagnetics,

full wave optics, optoelectronics, control circuitry
• Biophotonics - fluidics, heat transfer, biochemistry, cell transport and biology, wave
optics, physics of light absorption, scattering, and interference

CFD-ACE+ offers the EMAG module with full Maxwell equation solver for wave optics on
3D meshes and EM-LINK 3D wave solver for meshless free space optics (surface-to
surface). The figure below demonstrates an optical wave guide beam splitter and an optical
diffractive lens simulated with CFDRC tools.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 38

Europhysics News (2003) Vol. 34 No. 1

Silicon technology for optical MEMS

D.F. Moore1 and R.R.A. Syms2
1
Cambridge University Engineering Department, Cambridge, UK
2
Optical and Semiconductor Devices Group, Department of Electrical and Electronic
Engineering, Imperial College, London, UK

Micro-electro-mechanical systems (MEMS)

Components fabricated with the emerging technologies of micro-electro-mechanical
systems (MEMS) are being incorporated in an increasing number of sensor and actuator
applications [1,2]. There is rapid progress in optical systems for telecommunications to
meet the needs for increased bandwidth, for optical networks with terabit capacities per
fiber link, and for local area networks [3]. Some applications require precise features for
optical alignment while others involve the precise movement of small optical parts to
achieve advanced functionality. In the MEMS approach accurate, low-loss, optical
connections are made between different guided wave optical components including
fibres, waveguides and lasers. Mechanisms to allow motion and structures to provide
electrical actuation are combined in micro-opto- electro-mechanical systems (MOEMS).
Many MOEMS components process light in miniature free-space optical beams, and the
performance scaling laws are often different from those for guided wave optical systems.
This paper surveys MEMS with an emphasis on the outstanding challenges for Physicists
and Engineers. The silicon and other materials and technologies used to create MEMS
devices are discussed in detail, as are example applications in packaging such as fixed
and demountable connectors and methods for hybrid integration.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r

Seminar Report - 2004 39

Using MEMS Technology in Optical Switching

While MEMS is the most attractive technology for solving the problems of 3D optical
systems it is not without risks or challenges.

James Doscher, Analog Devices

Data-centric long haul and regional networks are beginning to be redesigned around
large all-optical cross connect switches that will allow BW management, real time
provisioning and mesh protection. A key-enabling component is the all-optical cross
connect switch that allows the connection of 1000's of input wavelengths to 1000's
output wavelengths under electronic control. Such a switch does not care about protocol,
bandwidth, or the data rate of the optical signal. The new network will set up
dynamically configurable highways of data in a nation-wide network. The ability to
change the connections on the fly allows mesh protection schemes to replace the current
point to point methods and allows carriers to dynamically add or delete bandwidth
through multiple paths and through multiple switch points to make better overall
utilization of their available capacity.

Central to the promise of the new network architecture is the physical switch fabric.
Micro Electro-Mechanical Systems, or MEMS, is one of a number of technologies being
considered for implementing optical switches and is the leading candidate for building
large scale all-optical cross connects. Companies have gravitated to so-called 3D switch
fabrics that use individual mirrors to direct beams of light from arrays input fibers to
arrays output fibers by physically steering the beams of light in an analog fashion from
one fiber to another. This contrasts to the more common cross bar switch found in both
the electrical and optical domain that has one switch element for every possible input
and output connection. Optical switches of this kind are known as a 2D or N2 switches
because the number of switching elements increases as the square of the number of ports
and because the light remains in a plane of 2 dimensions. When switches of this type get
much larger than about 32 x 32 the number of switching elements and the size of the die
become unmanageable and uneconomical.

3D architectures transmit beams of light in free space and require extremely high
accuracy and stability to work. This approach to switching does not have an analogous
architecture in the electronic world, the closest being the free space scanning electron
gun used in television sets. While MEMS is the most attractive technology for solving
the problems of 3D optical systems it is not without risks or challenges and these will be
discussed in the article.

Dept. of Electronics & Communication College of Engg. K i d a n g o o r