Optical Components

D K Kothari
Directional Couplers
• Use to combine and split optical signals in a network
• 2x2 DC consists of two input and two output ports
• Most commonly used couplers are made by fusing two fibers
together in the middle – called as fused fiber couplers
• A 2x2 coupler takes a fraction alpha of power from i/p 1 and places it
on o/p 1 and remaining power (1-alpha) on o/p 2
• Similarly fraction (1-alpha) of power from i/p 2 is distributed on o/p 1
and remaining on o/p 2.
• Alpha is called as coupling ratio
Directional Couplers
Directional Couplers
• Can be designed to be either wavelength selective or wavelength
independent (wavelength flat) over a wide range
• In wavelength selective device, alpha is depends on wavelength of the signal
while it is independent on wavelength in later case.
• Application – Very versatile device used as a building block for several other
optical devices
• 3 dB 2x2 coupler – equally divides the power from each input port on to
each of the output port by adjusting the coupling length
• An nxn coupler is a natural generalization of a 3dB 2x2 coupler – divides
power from n incoming ports equally amongst n o/p port
• Other applications – to tap off small fraction of power from a light stream
for monitoring purposes (alpha = 0.9 to 0.95), Modulators, switches etc
Coupler - Principle of operation
Ei2

A quantitative analysis of the coupling phenomenon is made by using

Coupled Mode Theory

The result of this analysis – the electric fields Eo1 and Eo2 at the o/ps of a
directional coupler can be expressed in terms of electric field at the i/ps Ei1
and Ei2 as follows:

Where l = Coupling length;

β = propagation constant of each of the two waveguides;
k (keta) = coupling coefficient, f(width of the waveguides, RI of wave guiding
region and the substrate and the proximity of the two waveguides
Directional coupler is often used with only one active i/p, say i/p 1
The power transfer function in this case can be derived as eq 2

Here Tij(f) = PTF from i/p i to o/p j and is defined by

Note from eq 2 that for a 3 dB coupler the coupling length must be chosen to
satisfy

Where k = a non-negative integer

Assuming the directional coupler is lossless, and i/p and o/p electric
fields are related by a general equation of the form

The matrix

is the transfer function of the device relating i/p and o/p electric fields
and is called the Scattering Matrix
We denote
or compactly can be written as
The sum of powers of the i/p fields is proportional to
Similarly the sum of powers of the o/p fields is proportional to

If the directional coupler is lossless, the power in the o/p fields must equal
the power in the i/p fields so that

Since the relationship must hold arbitrary Ei, we must have

where I =Identity matrix
Star Coupler – 8x8

Made using 3 dB couplers which

split power from each i/p equally
among all outputs
Isolators
• It a nonreciprocal passive optical device unlike couplers and most other
optical devices
• Its main function is to allow transmission in one direction through it but
block all transmission in other direction
• It is used in systems at the output of optical amplifiers and lasers
primarily to prevent reflections to enter these devices which would
otherwise degrade their performance
• Two key parameters - Insertion loss and Isolation
• Insertion loss is loss in forward direction and is as low as possible
• Isolation is the loss in the reverse direction and is as high as possible
Typically first is 1 dB and the later is 40 to 50 dB
Isolators….
State of Polarization (SOP)
For light propagating in a single mode fiber it refers to:
• The orientation of its electric field vector on a plane that is
orthogonal to its direction of propagation.
• At any time, electric field vector can be expressed as a linear
combination of two orthogonal linear polarizations supported by the
fiber
• These polarization modes are called as horizontal and vertical modes
 Isolators – Polarization dependent
I/p Polarizer – Passes the light energy
in one SOP and blocks the light energy
in other SOP e.g. passes vertical SOP –
Made up of dichroic crystals
Faraday Rotator – a non-reciprocal
device - rotates the SOP, say
clockwise, by 45 degrees, regardless
of direction of propagation. Made of a
crystal
O/p Polarizer – Passes 45 degrees
tilted SOP as it is to the o/p
Works only for a particular state of polarization of the input signal
As observed, the reflected light from
the other end is blocked at the i/p
end
 Isolators – Polarization Independent
(a) Propagation from left to right

Make use of special walk- Half-wave plate –

off polarizers at the inputs nonreciprocal device –
and outputs . On i/p side rotates SOP by 45 degrees
SWP splits the signal in in clockwise direction for
two orthogonally signals propagating from L-
polarized components R and in counter clockwise
and combine them at the direction for signals
output side propagating from R-L

(b) Propagation from right to left

 Circulators

Functional representation of Circulators (a) three-port and (b) 4-port

Arrows represent direction of signal flow
Circulators….
It is a nonreciprocal passive optical device
It is similar to isolator except that it has multiple ports typically 3 or 4
In a typical 3-port circulator i/p on port 1 is sent on out on port 2, an
i/p signal on port 2 is sent out on port 3 and that n port 3 is sent out
on port 1
It is useful to construct optical add/drop elements
It operates on the same principles as that of isolators
Optical Filters and Multiplexers
• Optical filters are essential component in transmission systems for
atleast two applications:
 To multiplex and de-multiplex wavelengths in WDM systems
 To provide equalization of gain and filtering of noise in Opt Amps

• Multiplexers and demultiplexers are used in WDM terminals as well

as large wavelength crossconnects and wavelength add/drop
multiplexers
Multiplexers and Filters
Static Wavelength Crossconnet using Mux and Demux
Key characteristics of Optical Filtering Technologies
Key Characteristics…
Characterization of some important spectral-shape
parameters of Optical Filters

λ0 denotes the center wavelength of the filter

and λ is the wavelength of the light signal
Filter fabrication technologies
• Integrated-optic waveguide technology – Semiconductor Chip Technology
 Waveguides can be made on substrates like Silica, InGaAs or Polymers
 Inherently polarization dependent

• All-fiber devices
 Amenable for mass production and
 Inherently polarization independent
Gratings
• Grating describe any device whose operation involves interference among
multiple optical signals originating from the same source but with the different
relative phase shifts
• An electromagnetic wave (light) of angular frequency ω propagating in z
direction has a dependence on z and t of the form cos (ωt – βz)
Here β is the propagation constant dependent on medium and the phase of the wave is ωt – βz

• Thus a relative phase shift between the two waves from the same source can
be achieved if they traverse two paths of different lengths
• Gratings are used in optics to separate light into constituent wavelengths
• In WDM – it is used for Demux and Mux
Different Types of Grating and its operation

Θi is the angle of incidence of light

The angle at ehich the signal is diffracted

Depends on wavelength (Θ1d for λ1
and Θ2d for λ2

Transmission Grating Refection Grating

Transmission and Reflection Gratings
• Multiple narrow slits are placed equally apart on a plane, called the Grating plane
• The spacing between two adjacent slits is called as the pitch of the grating
• Light incident from a source on one side of the grating is transmitted through
these gratings and spreads out in all directions through diffraction as the slits are
narrow
• Each slit is thus act as a source of light
• There is another plane parallel to grating plane on which transmitted light
interferes. This plane is called as Imaging plane
• At any point on this plane, for a wavelength for which individual interfering waves
are in phase, there is constructive interference and enhancement in light intensity
for these wavelengths
• For a large number of slits, the interference is not constructive at other
wavelengths and there is little intensity at this point from these wavelengths
• Since different wavelengths interfere constructively at different point on imaging
plane, a WDM signal can be effectively separated spatially into its constituent
wavelengths
• Optical fibers could be placed at different imaging points to collect light at
different wavelengths
Transmission and Reflection Gratings
• Since multiple transmission occur in the grating, this grating is called as
transmission gratings
• If the transmission slits are replaced by narrow reflecting surfaces, with rest of
the surface being non reflected, it is called as reflection ratings
• Operation of reflection grating is analogous to transmission grating
• Due to the ease of fabrication most of the gratings in practice are reflection
gratings
• Since diffraction is key to operation of these devices, they are also called as
diffraction gratings
• In addition to plane geometry, gratings are also constructed in concave geometry
in which slits are located on the arc of a circle
• Devices with concave geometry uses fewer auxiliary parts like lenses and mirrors
for overall device construction
• Stimax grating is a reflection grating commercially used in WDM transmission
system
• This is integrated with a concave mirror and the input and output fibers
Operating Principle

The path length difference between the rays diffracted

At angle θd from the adjacent slits is
 Operating Principle
Let the pitch of the grating (distance between adjacent slits) = a
Assumption = light src is far from the grating plane compared to ‘a’
Angle of incidence from light source to grating plane at each slit = θi
Angle of diffraction of light from the grating plane = θd
then
the path length difference between the rays traversing through the adjacent slits
= difference in lengths between the line segments AB^ and CD^
= a[sin(θi) - sin(θd)]} [Assignment – Derive this]
Assuming imaging plane is far from the grating plane compared to the pitch
Also slit is small compared to λ so that the phase change across a slit is negligible
Thus the constructive interference at wavelength λ occurs at the imaging plane among the
rays diffracted at angle θd if the following grating equation is satisfied
a[sin(θi) - sin(θd)] = mλ
for some integer m, called the order of the grating
Operating Principle…
• Grating effects the separation of individual wavelengths in WDM signal since the
grating eq is satisfied at different points in imaging plane at different wavelengths
• This is illustrated in figure where different wavelengths are shown being
diffracted at the angles at which the grating eq is satisfied for that wavelength
e.g., θd1 is the angle at which the grating eq is satisfied for λ1.
• Energy at a wavelength is distributed over all the discrete angles that satisfy the
grating eq at this wavelength
• When grating is used as demux, light is collected only from one of these angles
and the remaining energy in other orders is lost
• In fact most of the energy will be concentrated in the zeroth-order (m=0)
interference maximum, which occurs at θi = θd for all wavelengths
• The light energy in this zeroth-order interference maximum is wasted since WLs
are not separated
• Thus gratings must be designed so that the light energy is maximum at one of the
other interference maxima using a technique called Blazing
Blazed Grating

• In such grating the reflecting slits are at angle α to the grating plane ,
• This is called as blaze angle
• This maximizes the light energy in the interference maximum whose order
corresponds to the blazing angle
Bragg Gratings
• Widely used in fiber optic communication systems
• In general, any periodic perturbation in the propagating medium
serves as a Bragg grating
• This perturbation is usually a periodic variation of the RI of the
medium
• Laser use BG to achieve single frequency operation. In this case Bragg
gratings are “written” in waveguides
• Bragg Gratings written in fiber can be used to make a variety of
devices like filters, add/drop multiplexers and dispersion
compensators
 Bragg Grating…

• bragg grating with periodic variation in RI

• The incident wave is reflected back from each period of the grating
• The reflection add in phase when the path length in wavelength λ0 each period is equal to
half the incident wavelength λ0
• This equivalent to the Bragg condition
where ʌ is the period of the grating, ηeff = effective RI of the waveguide or the fiber
 Bragg Grating…

Reflection spectrum for a grating with a uniform RI pattern change Apodized Grating
across its length Undesirable side lobes are eliminated here
∆ = a measure of the bandwidth of the grating and is the wavelength by making the RI changes smaller towards
separation between the peak WL and the first reflection minimum the edges
and Is inversely proportional to the length of the grating
∆λ = detuning from the phase matching wavelength
 Fiber Gratings
• Being all fiber devices, their main advantages are their low cost, ease of coupling,
polarization insensitive low temp-coefficient and simple packaging
• Grating are written in fibers by making use of photosentivity of certain type of optical
fibers
• A conventional silica fiber doped with Germanium becomes extremely photosensitive
• Exposing this fiber to UV light causes changes in RI within the fiber core
• A grating can be written in such a fiber by exposing them to two interfering UV beams
• This causes the radiation intensity to vary periodically along the length of the fiber
• Where the intensity is high, RI increased. Where it is low , RI remain unchanged
• Changes needed in RI is very small - of the order of 10 to the power -4.
• Other technique is the phase mask
• It is a diffractive element which when illuminated by a light beam, splits the beam into
different diffractive orders, which then interfere with one another to write the grating
into the fiber
• Classified based on period of grating – Short and Long period Gratings
 Add/drop element using fiber bragg grating

a Drop element

b Add/Drop element
Fabry-Perot Filters
• Consists of cavity formed by two highly reflective
mirrors placed parallel to each other
• The input light beam to the filter enters the first mirror)
at right angles to its surface

• After one pass through the cavity, a part of the light leaves the cavity through the right
facet and a part is reflected
• For those wavelengths for which the cavity length is an Integral multiple of half
wavelength in the cavity – so that the round trip through the cavity is an integral
multiple of the wavelength
• All the waves transmitted through the right facet add in phase
• Such wavelengths are called as the resonant wavelengths of the cavity
Fabry Perot filter….
• Power transfer function of a filter is the fraction of the i/p light power
that is transmitted by the filter as a function of the frequency f, or
wavelength
• For Fabry-Perot filter it is given as

• In terms of optical free-space wavelength λ, it can be expressed as

Where A = absorption loss of each mirror; R = reflectivity of each mirror

τ = one-way propagation delay across the cavity = nl/c
l = length of the cavity and n = RI of the cavity
• Tfp (f) is periodic in f and the peaks, or the passbands of the PTF plotted for A = 0 and R = 0.75, 0.9 and 0.99
tx fn occur at frequencies f that satisfy fτ = k/2, for some positive
integer k

• Thus in a WDM system if the wavelengths are spaced suffciently far

apart compared to the width of each passband of the filter transfer
function several frequencies (or λs) may be transmitted by the filter
if they coincide with different passbands

• The spectral range between two successive passbands of the filter

is called as free spectral range (FSR)

• Measure of width of a passband => its full width at a point where

the trf is half of its maximum (FWHM)

• To minimize crosstalk in WDM sytem, two adjacent λs must be

separated at least by a FWHM

• Thus the ration FSR/FWHM is an approximate measure of no. of

wavelengths that can be accommodated by the system

• The ratio is called as Fitness f, of the filter and is given as

Fabry-Perot Filter…
Tunability
• FPF can be tuned to select different wavelengths by two methods
1. By changing the cavity length
2. By changing RI within the cavity

• Consider a WDM system all of whose WLs lie within one FSR of FPF
• The frequency 𝒇𝟎 that is selected by the filter satisfy 𝒇𝟎 τ = k/2 for some positive integer k
(τ = one way propagation time for the lightbeam)
• 𝒇𝟎 can be changed by changing τ and it is defined as τ = ln/c
(l =length of the cavity, n = RI of cavity and c = speed of light)

• Thus τ can be changed by by changing either l or n

• Length of the cavity can be changed in two ways

 By moving one of the mirror (mechanical method)and
 By filling the cavity by some piezoelectric material which undergoes compression on
application of voltage
Multilayer Dielectric Thin-film filters
• A thin-film resonant cavity filter (TFF) is a Fabry-Parot interferometer, or etalons
• The mirrors surrounding the cavity are realized using multiple reflective dielectric
thin-film layers
• This device acts as a bandpass filter, passing through a particular wavelength and
reflecting all the other wavelengths
• The wavelength that is passed is determined by the cavity length
• A thin-film resonant multicavity filter consists of two or more cavities separated
by reflective dielectric thin-film layers
• As more cavities are added, on the response of the filter, the top of the passband
becomes flatter and the skirts become sharper
( both desirable features)
 Multilayer Dielectric Thin-film filters….

3-cavity thin-film resonant dielectric thin-film filter Transfer function of one, two and three cavity
dielectric thin-fim filters
Multilayer Dielectric Thin-film filters….
To obtain a mux or demux, a no. of filters are
Cascaded

Each filter passes a different wavelength and

reflects all the others

In a demux, first filter in the cascade passes one

Wavelength and reflects all the other onto the
Second filter

Second filter passes another wavelength and

Reflects the remaining ones and so on

Salient features
• Can be made to have very flat top on the passband and very sharp skirts
• Extremely stable with regard to temperature variations
• Low loss and insensitive to the polarization of the signal
Mac-Zehnder Inerferometers (MZI)
• Principle – make use of two interfering paths of different lengths to
resolve different wavelengths
• MZI are useful for both wideband filters and (de)multiplexers
• For example used for separating 1.3 μm and 1.55 μm bands
• Narrow band MZI filters are fabricated by cascading no. of stages
• Have good crosstalk performance
• Useful for two input and two output Mux and Demux and tunable
filters
• Today MZI are typically constructed in integrated optics
• The substrate is usually silicon and the waveguide and cladding
regions are Silica (SiO2)
 2x2 MZI Consists of two 3-db couplers interconnected
through two paths of differing lengths

Initial 3-dB coupler splits the i/p signal equally

In Central section one of the waveguide is longer by

∆L to give a wavelength dependent phase shift
between the two arms

Another 3-dB coupler recombine the signals at the

o/p

By splitting i/p beam and introducing phase shift in

one of the paths – the recombined signal interfere
constructively at one of the o/p and destructively at
the other o/p

The signal then finally emerges from only one o/p

c. 4-stage MZI which uses different pat lengths in each
port
stage
Tuning of MZI filters
• Tuning in case of filter is achieved by varying temp of one of the arm
of the device, thereby causing RI of the arm to change which in turn
affects the phase relationship between the two arms and causes a
different wavelength to be coupled out
 Operation of MZI as a demultiplexer
• Only one I/P say 1 has a signal
• After first section, power of the signal gets equally divided between two
arms with the signal in the lower arm having phase shift of π/2 wrt to
upper arm
• Since there is a length difference of ∆L between the two arms there is
further phase lag of β∆L introduced in the lower arm
• In the second directional coupler, the signal from the lower arm undergoes
another phase delay of π/2 in going to the first o/p relative to the signal
from the upper arm
• Thus the total relative phase shift at the upper o/p between the two signals
is π/2 + β∆L + π/2
• At the o/p dc, in going to the second o/p, the signal from the upper arm
lags the signal from the lower armin phase by π/2
• Thus the o/p at the lower port has a total phase shift of π/2 + β∆L - π /2 =
β∆L
Operation of MZI as a demultiplexer
• If β∆L = k π and k is odd, the signal at the first o/p add in phase
whereas the signal at the second o/p add with opposite phases and
thus cancel each other
• Thus the wavelength passed from the first i/p to the first o/p are
those wavelengths for which β∆L = k π and k is odd
• The wavelengths passed from the first i/p to the second o/p are those
for which β∆L = k π and k is even
Arrayed Waveguide Grating (AWG)
• It is generalisation of MZI
• Consists of two multiport couplers interconnected by an array of
waveguides
• In AWG several copies of the same signal, but shifted in phase by
different amounts are added together
• Several uses – nx1 multiplexer and 1xn demultiplexer
• n i/p signals at different wavelengths are combined onto a single o/p
and inverse of this function respectively
• Compared to MZI realisation
it has low loss, flatter passband
and easy to realize on an
integrated-optic substrate
Acousto-optic Tunable filter (AOTF)
• It is versatile device that is capable of selecting several wavelengths
simultaneously
• This characteristic make it suitable to construct wavelength
crossconnect
• Its construction is based on the interaction of sound and light
• Basically an acoustic wave is used to create Bragg grating in a
waveguide, which is then used to perform wavelength selection
• Its operation is polarization dependent and can be made independent
of it using integrated optics
AOTF….

• It consists of a waveguide constructed from a birefringent material

• Supports only lower order TE and TM modes
• Assume the input energy is entirely in the TE mode
• A polarizer, which selects only the light energy in TM modes, is placed
at the other end of the channel waveguide
• The light energy in a narrow spectral range around the wavelengths to
be selected is converted to TM mode by launching an acoustic wave
along, or opposite, to the direction of propagation of the light wave
• As a result of propagation of acoustic wave the density of the medium
varies in periodic manner
AOTF…
• The period of this density variation is equal to the wavelength of the acoustic
wave
• This periodic variation acts as a Bragg grating
• The light couples from one mode to the other if the refractive indices 𝑛 𝑇𝐸 and
𝑛 𝑇𝑀 of TE and TM modes satisfy the Bragg condition 𝑛 𝑇𝑀 /λ = (𝑛 𝑇𝐸 / λ) ± 1/ʌ
• Thus the light energy in the narrow spectral range around the wavelength λ that
satisfies Bragg condition undergoes TE to TM mode conversion
• The device, thus, acts as a narrow bandwidth filter when light energy in the TE
mode is input and only the light energy in TM mode is selected at the output
• In LiNbO3, 𝑛 𝑇𝐸 and 𝑛 𝑇𝑀 differs by 0.07
• So if RI difference is denoted by ∆n, then Bragg condition becomes λ = ʌ (∆n)
• A wavelength that undergoes the mode conversion and thus lies in the passband
of AOTF can be selected, or tuned, by suitably choosing the acoustic wavelength ʌ
AOTF
• In order to select a wavelength of 1.55 μm (corresponding to ∆n =
0.07), the acoustic WL is 22 μm = RF of 170 MHz as velocity of light in
LiNbO3 is 3.75 km/s
• Since the RF frequency is easily tuned, the wavelength selected by the
filter can easily be tuned
Polarization Independent AOTF

• A polarizer splits the input signal into its constituent TE and TM polarization mode
and sending them separately through the AOTF
• Each mode is converted in two separate arms and then recombined at the output
• It is realized in integrated-optics
Optical Switches
• Used in ONs for variety of applications and having different switching
times and number of ports
• Switching time and port requirements for different applications are as
follows
Other important parameters of Optical Switches
• Extinction ratio:
Ratio of o/p power in the OFF state to the o/p power in the ON state
A high value in for external modulator application is expected
Simple mechanical switches – 40 - 50 dB
External Modulators – 10 -25 dB
• Insertion Loss:
Fraction of power that is lost because of presence of the switch
Low as much as possible
Some switches have different losses for different input-output connections
which is undesirable as it increases dynamic range of the signals in the N/W
Loss Uniformity is important factor and is dependent on switching architecture
(i.e., independent of fabrication technology)
Important parameters…
• Cross talk:
The ratio of power at the o/p from the desired i/p to the power o/p from
all the other i/ps
A high value of this ration is desirable
Usually defined as the worst-case crosstalk over all outputs and
interconnection patterns

• Polarization dependent loss (PDL):

Should be as low as possible
In external modulators, polarization dependence can be tolerated as it is
used immediately following laser whose SOP can be controlled by using
special polarising-preserving fiber to couple the light from the laser into
the external modulator
Important Parameters….
• Latching:
Such switch maintains its switch state even if the power to the switch is turned
off
It enables the traffic to be passed through the switch even in the event of power
failures
So it is a desirable feature

• Read-out capability:
Allows current state to be monitored
Used for verifying that the right connections are made through the switch

• Reliability:
To ensure this the switch is cycled through its various states a few million times
In provisioning and protection applications, switch remain in a single state over a
very long period of time and reliability issue here is that whether the switching
operation will be done correctly even after remaining untouched for such a long
period
Consideration for building Large Scale Switches
• No of switch elements required:
Large switches are made using multiple switch elements in some form or the other
Cost and complexity depends on no. of switch elements required in addition to other
factors
• Loss Uniformity:
A measure of this is the minimum and maximum no. of switch elements in the optical path for
different combinations of input and output
• No of crossovers:
 Switches are fabricated by integrating multiple switches on a single substrate and on a single
layer interconnected by waveguides
 When such waveguides paths are crossed, it gives rise to the power loss and cross talk
 For acceptable loss and crosstalk performance, no. of such crossovers must be minimized or
eliminated
Consideration for building Large Scale Switches….
• Blocking Characteristics – Switching function achievable are
Non Blocking: If an unused input port can be connected to any unused output port
 Thus is capable of realizing every interconnection pattern between input and output

Blocking: If some interconnection pattern(s) can not be realized then it is said to be blocking
 Most applications require nonblocking switches

Nonblocking switches can further be distinguished as:

Wide sense nonblocking: When any interconnection pattern can be realized without requiring
rerouting of any of the existing connection
 uses specific routing algorithms to route connections so that future connections will not be
blocked
Strict-sense nonbloking: Any interconnection patterns are realised regardless of their previous
connections through the switch
Considerations for large optical switches…
• Rearrangeably nonblocking switch: a nonblocking switch that require
rerouting of connections to achieve nonblocking property is said to be
rearrangeably nonblocking
• Adv of rearrangeably nonbloking switch architecture is that they need
fewer small switches to build large switch of a given size
• Such switches needed more complex control algorithms to set up the
connections
• Such algorithms in toda’s switches are implemented using
microprocessors
Types of Switching Architectures
4 x 4 Cross-Bar Switch
To set path from input I to output j, the path taken traverses
Through 2x2 switches in row I till it reaches column j and then
Traverses the switches in column j till it reaches output j

2x2 Switches in row i and column j must be set appropriately in

this path for this connection to be made

Total 2x2 switching elements required = 𝒏𝟐

Minimum path length = 1
Maximum path length = 2n - 1
 1024 x 1024 Clos Switch
Uses 3 parameters m,k and p
Let n = mk

No. of switches in I and III stage = k(m x p)

No.of switches in Middle stage = p(k x k)

Each of the k switches in the first stage is connected

to the all switches in the middle stage
Ie each of the p outputs in first stage is connected to
the input of a different switch in the middle stage

Likewise each of the k switches in the third stage is

connected to all the stiches in the middle stage

A strict sense nonblocking switch realized using 32 x 64 and If p > or = 2m – 1, the switch is strictly nonblocking
32 x 32 switches interconnected
in a three stage Close architecture
Clos Switch….
• Total No. of switching elements required =
• No of 2x2 switching elements required in minimum cost
configuration =
• Significantly lower than 𝒏𝟐
• Loss uniformity is better than Cross bar architecture
n x n Spanke Architecture

• A very popular architecture for building large switches

• An n x n switch is made by combining n 1xn switches along with n nx1 switches to obtain a
strict-sense nonblocking architecture
• It is built using single 1x n switch element so do not need 1x2 or 2x2 switch elements
• Built using MEMS analog beam steering mirror technology
• Requires 2n such switch elements to build an n x n switch
Spanke Architecture….
• Switch cost linearly scale with n which is significantly better than other switch
architectures

• Have much lower insertion loss as each connection passes through only two
switch elements (significantly lower than the same in the path for other
multistage designs)

• Optical path length for all input-output combinations can be made essentially the
same, so the loss is same regardless of the specific input-output combinations
 • It is a rearrangeably nonblocking architecture
Benes᷉ Architecture • Efficient architecture in terms of 2x2 switches used
to build larger switches
• An 8 x 8 switch requires only 20 2 x 2 switches
• In general for a n x n architecture, 2 x 2 switching
elements it requires can be given as

n being a power of two

• Each path passes through 2 x 2 switching
elements
So path loss is same through every path in the switch
• Disadvantages
– Not wide-sense nonblocking and
- More no. of waveguide crossovers are required
Spanke-Benes᷉ Architecture
• It is rearrangeably nonblocking architecture
• Good compromise between Crossbar and Benes
architectures
• A 8 x 8 switch can be constructed using 28 2 x 2 switches
• It is called as n-stage planner architecture since it requires
n stages (columns) to realize an n x n switch
• It requires n(n-1)/2 switches
• The shortest path length is n/2 and the longest path length
is n
Comparison between different switching
architectures
Optical Switch Technologies
A. Bulk Mechanical Switches
B. Micro-Electro-Mechanical System (MEMS) Switches
C. Liquid crystal switches
D. Electro-optic switches
E. Thermo-optic switches
F. Semiconductor optical amplifier switches
 A. Bulk Mechanical Switches
• In this switch switching is done by mechanical means
i. Mirror arrangement technique - switching state is controlled by moving mirror in and
out of the optical path
ii. Directional coupler technique – bending or stretching the fiber in the interaction region
changes the coupling ratio of the coupler and can be used to switch light from an input
port between different output ports
Features and applications
• Low insertion, low PDL, low crosstalk and are relatively inexpensive
• Available mostly in crossbar configuration
• Switching speed is of the order of few milliseconds
• No of ports are fairly small, 8 or 16
• Used in small size wavelength cross connects for provisioning and protection switching
• Being mechanical, reliability is an issue with this kind of switches
B. Micro-Electro-Mechanical System (MEMS) Switches
• In context of optical switches it refers to miniature movable mirrors fabricated in
silicon
• Dimensions ranging from few hundred micrometres to few millimetres
• A single silicon wafer yields large number of mirrors which means that these
mirrors can be manufactured and packaged as arrays
• Mirrors can be fabricated using fairly standard semiconductor manufacturing
processes
• Theses mirrors can be deflected from one position to another using a variety of
actuation techniques such as electromagnetic, electrostatic, or piezoelectric
methods – hence the name MEMS
• Electrostatic deflection is particularly power efficient but is relatively hard to
control over a wide deflection range
Micro-Electro-Mechanical System (MEMS) Switches…
Simplest possible structure is so-called two state pop-up mirror or 2D mirror

In one state, the mirror is flat in line with the substrate so the light beam is not deflected
In other state mirror pop-up in the vertical position so the lightbeam, if present, is
deflected
Such a mirror can be used in a crossbar arrangement to realize n x n switch
Mirrors can be controlled by digital means as only two positions need to be supported
Size is limited to 32 x 32 switch due to processing constraints
Micro-Electro-Mechanical System (MEMS) Switches…

• Analog beam steering mirror

Micro-Electro-Mechanical System (MEMS) Switches…
• The mirror is connected through flexures to an inner frame, which in turn is
connected through another set of flexures to an outer frame
• The flexures allow the mirror to be rotated freely on two different axes
• Mirror is controlled in an analog fashion to realize a continuous range of
angular deflections
• This type of mirror is also known as analog beam steering mirror, a
gambled mirror, or a 3D mirror
• Used to realize 1 x n switch
• Sophisticated servo controlled mechanism is required to deflect the mirrors
to their correct position and hold them there
n x n switch built using analog beam steering MEMs
mirrors
C. Liquid crystal switches • Make use of polarization effects to perform switch
Function

• By applying a voltage to a suitably designed liquid

crystal cell, polarization of the light passing through it can
Be caused either to be rotated or not

• This can then be combined with passive polarization

beam splitters and combiners to yield a polarization-
independent switch

• Typically, passive polarization beam splitter, combiner

And an active switch element all can be realized using an
Array of liquid crystal cells

• The polarization rotation can be controlled in an analog

fashon by controlling voltage

• Switching time of the order of few milliseconds can be

obtained
D. Electro-optic switches
• A 2 x 2 switch can be realized using external modulator configurations
• A commonly used material is lithium neobate (LiNbO3)
• In the directional coupler configuration, the coupling ratio is varied by
changing the voltage and thus the refractive index of the material in the
coupling region
• In the Mach-Zehnder configuration, the relative path lengths between the
two arms of the Mach-Zehnder is varied
• An electro-optic switch is capable of changing its state extremely rapidly –
less than 1 ns
• This switching time limit is determined by the capacitance of electrode
capacitance
• Allows modest level of integration compared to the mechanical switches
• Large size switches can be made integrating several 2 x 2 switches but tend
to have relatively high loss and PDL and are more expensive than
mechanical switches
E. Thermo-optic switches
• Essentially a 2 x 2 intrgrated-optic Mach-Zehnder interferometers,
constructed on waveguide material whose RI is a function of the
temperature
• By varying the RI in one arm of the interferometer, the relative phase
difference between the two arms can be changed
• Thereby switching of input signal from one output port to another
can be achieved
• Such devices are made on silica or polymer substrate
• Have poor crosstalk, quite slow, and switching speeds of the order of
few milliseconds can be obtained.
F. Semiconductor optical amplifier switches
• SOA can be used as a on-off switch by varying the bias voltage to the device
• If bias voltage is reduced no, population inversion can be achieved and the
device absorbs the input signals
• And if the bias voltage is present, it amplifies the input signals
• The combination of amplification in the on state and absorption in the off
state makes this device of capable of achieving very large extinction ratios
• The switching speed of the order of 1 ns
• Large switches can be fabricated by integrating SOAs with passive couplers
• However, this is an expensive component and is difficult to make
polarization independent