semiconductor lasers for optical...

1

C. Coriasso – Nov 2012

Semiconductor Lasers

for Optical Communication

Claudio Coriasso Manager

claudio.coriasso@avagotech.com

Turin Technology Centre

MQW

10Gb/s DFB Laser

2

C. Coriasso – Nov 2012

Outline

1) Background and Motivation

• Communication Traffic Growth

• Why Photonics?

• Photonics evolution

2) Semiconductor Laser Basics

• Active material

• Optical Feedback

3

C. Coriasso – Nov 2012

Communication Growth

Tim Berners-Lee computer at CERN:

World’s first Web Server

1991

Worldwide communication traffic is doubling every 18 months (2dB/year)

Communication has always been one of the main driving force for the

development of new technologies:

Telegraph, Telephone, Fiber Optic, Laser, …

The Internet (34.3% world population)

4

C. Coriasso – Nov 2012

Traffic structure and Energy Consumption

• Data Centers and the Internet consume about 4% of electricity

(8.7 1011 kWhr/year including PCs)

• By 2018 the energy utilized by IP traffic will exceed 10% of the total electrical power

generation in developed countries (1)

• Most of the traffic is between machines and much of the information created today cannot

even be stored (2)

(1) L.Kimerling, MIT

(2) R.Tkatch, Alcatel Lucent

(3) D. Lee, Facebook European Conference on Optical Communication 2010

FACEBOOK (3) :

• 1.01 billion active users worldwide (23M in Italy, 39.5% population)

• 584 million active users every day

• 10.5 billion minute/day

C. Coriasso – Nov 2012

Why Photonics?

• Short monochromatic optical pulses are easily produced with semiconductor lasers (ps range Gb/s to Tb/s)

• Photons do not interact each other

• Photons can be propagated in optical fiber with very low loss (0.2dB/km)

• Several data streams at different wavelength can be combined, propagated together in optical fibers and then split (high channel capacity)

6

C. Coriasso – Nov 2012

Photonics in Optical Communication

Photonics

Electronics

Today Photonic Network:

Storage Area Network Local Area Network

Wide Area Network

Metro Area Network

10km 100km 100m 10m 1km 1000km

Traffic volume

7

C. Coriasso – Nov 2012

Photonics

Emission,

Transmission,

Processing (modulation, switching, amplification, …)

Detection

Photonics:

Science and technology of light

Began with Laser (1960) and Fiber Optic (1966) inventions.

These inventions formed the basis for the telecommunications revolution

of the late 20th century and provided the infrastructure for the internet.

1st Laser demonstration : T. Maiman 1960 1st Low-Loss Fiber Optic Proposal: C. Kao 1966

8

C. Coriasso – Nov 2012

Fiber Optics 1966: First proposal of fiber optic for telecom. Basic design [Kao STC, Nobel Prize 2009] 1970: Production of first fiber optic [Corning] 1976-77: First fiber optic networks 1988: First transoceanic fiber-optic cable (3148 miles, 40000 simultaeous telephone calls)

C. K. Kao receiving

his Nobel Prize

Stockholm 2009

First operative optical cable in urban areas

Torino 1976 First optical system carrying live traffic

over public network Chicago 1977

9

C. Coriasso – Nov 2012

Semiconductor Laser Basics

10

C. Coriasso – Nov 2012

Semiconductor LASER history 1962: First Realization of Semiconductor Laser (GaAs @T = - 200 oC) [GEC, IBM, MIT] 1963: Proposal of Heterostructure Semiconductor Laser (H. Kroemer, Z. Alferov: Nobel Prize 2000) 1970: First Realization of Heterostructure Semiconductor Laser (Z. Alferov) 1972: Proposal of Distributed Feedback Laser (DFB) 1970: Room Temperature CW operation of 1.5m Laser 1977: Proposal of Vertical Cavity Surface Emitting Laser (VCSEL) 1984: First Realization of Strained MQW in semiconductor laser 1988: First Realization of VCSEL 2000: First Uncooled Telecom Lasers

Z. Alferov receiving

his Nobel Prize

Stockholm 2000

11

C. Coriasso – Nov 2012

Grating

(optical feedback)

Quantum wells

(optical gain) Waveguide

(optical confinement)

Electrodes

(current injection)

Laser requirements for optical communication

DFB MQW Laser

• High bandwidth ( 10Gb/s)

• Single mode operation

• Low consumption

• Uncooled operation (up to 80oC)

25Gb/s

State of Art of uncooled high-speed lasers

12

C. Coriasso – Nov 2012

High-speed laser key factors

High carrier density and high photon density in an active material

within a small-volume optical resonator

1. Electrical confinement (p-n heterostructure)

2. Optical confinement (single-mode waveguide)

3. Gain (quantum-confined material)

4. Feedback (distributed Bragg grating)

1 2 3

4

Laser rate equations:

0dt

d

13

C. Coriasso – Nov 2012

• Optical gain, light emission (direct band gap) ...

•... at wavelength of interest: = 1,3 m e 1,55 m

• compatibility with semiconductor substrates: Si, GaAs, InP

Semiconductor material basic requirements

for photonic devices

e

h

photon emission

through e-h

recombination

14

C. Coriasso – Nov 2012

III V

There are no single elements or binary compounds compatible with commercial substrates and emitting light at 1.3 m e 1.55 m. Semiconductor alloys of III-V elements are the best materials for photonic devices.

III-V semiconductor materials

15

C. Coriasso – Nov 2012

Quaternary alloy InGaAsP

T. P. Pearsall, , Wiley (1982)

In1-xGaxAsyP1-y alloy cover all the

spectral range required for optical

telecom

• High quality material

• Established

growth techniques

and material processing

• Suited for active devices (lasers,

amplifiers, modulators, ...) and

passive structures (waveguides,

couplers, ...)

In1-xGaxAsyP1-y

lattice matched to InP

(g=0.92 - 1.65m)

InGaAs

(g=1.65m)

InP (g=0.92m)

16

C. Coriasso – Nov 2012

Further alloy systems

1

1 In1-xGaxAsyP1-y on InP

2

Al1-x-yGaxInyAs on InP 2

3 Al1-xGaxAs on GaAs 3

17

C. Coriasso – Nov 2012

(Basic) Optical properties of semiconductors

lh hh

e

E

Eg

**

**9

elh

ehh

mm

mm

Three bands are involved in optical transitions:

- Electrons Conduction Band

- Heavy holes

- Light holes Valence Band

JDOS

Eg

a

gEJDOS

k

Joint density of states (available for optical

transitions) is a square root function of the

energy in excess of the energy gap

18

C. Coriasso – Nov 2012

Semiconductor Heterostructures

A

(Substrate)

A

B B

A

Double Heterostructure (DH)

n

p

e-h confinement

Eg(A)> Eg(B)

Combination of layers of different crystalline semiconductors.

H. Kroemer, Varian associates 1963

(Nobel Prize in Physics, 2000)

The idea was experimentally demonstrated using the Liquid Phase Epitaxy (LPE)

Separate Confinement

Heterostructure (SCH) ph > e

A

(Substrate)

A

B B

A

n

p

Photon confinement

C

C

Eg(A)> Eg(C) )> Eg(B) , n(A)< n(C) < n(B)

refractive index

Yphoton

19

C. Coriasso – Nov 2012

Quantum Wells

Quantum-Size Double Heterostructure (Quantum Well) is a planar waveguide for electrons

C. H. Henry, Bell Labs 1972

The idea was experimentally demonstrated in 1974 using the newly developed Molecular

Beam Epitaxy (MBE).

A

QUANTUM WELL

d = 3

z

y

x

d = 2 Lw ~ e

BULK

EgA Eg

B EgQW

1

2

2 1

B A B

Lw

QW is now a widely spread

quantum product

based in atomic-scale technology

B A B

Transmission Electron Microscope

view of QW

20

C. Coriasso – Nov 2012

)()()(222

0

2

2znzznk

dz

deff

neff

2

Yph(z)

n2(z)

Photon wave eqn. vs. Electron wave eqn.

Helmholtz equation (photon)

refractive index ridges confine photons (optical waveguides)

z Ye(z)

)()(2 zVzn

cladding barrier

core well

)()()(2 2

22

zEzzVdz

d

m

E

V(z)

Schroedinger equation (electron)

potential wells confine electrons (quantum wells)

21

C. Coriasso – Nov 2012

Eigenfunction/Eigenvalues Calculation

E0

E1 nTE0

nTE1

nTE0

nTE1

E0

E1 YTE0

YTE1

Y1

Y0

Optical Waveguide Quantum Well

)()(

)()(

Y

Y

YY

zz

zz

zz

TETE

TETE

)()(

1)(

)(

1

)()(

Y

Y

YY

zzzm

zzzm

zz

22

C. Coriasso – Nov 2012

Reduced dimensionality structures

A

QUANTUM WELL

d = 3 z

y

x

d = 2

Lz ~ e

BULK

PLANAR WAVEGUIDE

E.M.

A B

B

Q.M.

1975: R.Dingle and C.Henry

USA Patent Application

QUANTUM WIRE

d = 1

Lz ,Ly ~ e

A B

B

QUANTUM DOT

d = 0

Lz ,Ly ,Lx ~ e

CHANNEL WAVEGUIDE

OPTICAL RESONATOR A

B

B

23

C. Coriasso – Nov 2012

MQW heterostructures

Multi Quantum Wells are stacks of decoupled QWs

(with sufficiently thick barriers): enhancement of single QW effects

ENERGY

Z Y

X Z

WELLS

SUBSTRATE BARRIERS

VA

LE

NC

E B

AN

D

CO

ND

UC

TIO

N B

AN

D

24

C. Coriasso – Nov 2012

QW band structure

E

hh1

e1

e2

hh2

lh2

lh1

11 1’1

TE TM

22

z

2D JDOS related features:

• Sharp absorption edge (2D JDOS)

• High differential gain

• Wide gain bandwith

• High electroabsorption efficiency (QCSE)

• Strong optical nonlinearities

• . . .

JDOS

e1-h

h1

EgQW

Egbulk

e1-l

h1

e2-h

h2

One step for

every

allowed

transition

( ) ji

ji

jiEJDOS

2

bulk

gEE

25

C. Coriasso – Nov 2012

Excitons

e

h

e - h semi-bound states (t ~ 100 fs) which

produce sharp absorption peaks detuned from

the transition energies (absorption steps) by

their binding energy

Eb (~8 meV per InGaAsP)

E

a

h

26

C. Coriasso – Nov 2012

QW optical properties

Polarisation selection rules:

TE: 3/4 hh, 1/4 lh

TM: 0 hh, 1 lh

e1-h

h1

e1-l

h1

e1-h

h1

e1-l

h1

aTE

aTM

nTE

nTM

Strong dichroism

a

a

a

2

)()()(n̂

'

')'(1)()(

0

22

cin

dcnstateexcitonJDOS

27

C. Coriasso – Nov 2012

aL

aS

ma a

a

a lattice parameter of epitaxial layer

a lattice parameter of substrateL S

S

L

S

RST

Strain(1)

The epitaxial layer can be grown with a lattice parameter slightly different from the

substrate lattice parameter (lattice mismatch).

compressive strain m>0

tensile strain

aL

aS

m<0

, aL = lattice parameter of the epitaxial layer

aS = lattice parameter of the substrate

28

C. Coriasso – Nov 2012

T. P. Pearsall, , Wiley (1982)

InGaAs

InP

Strain(2)

InGaAs

InP

±1%

29

C. Coriasso – Nov 2012

hh1

e1

e2

lh1

11 1’1

k

E

lh hh

e

m=0

TM

TE

Strain effect on band structure:

lh1

k

E

lh hh

e

m>0

hh1

e1

e2

11 1’1

TM

TE

Low escape time

High T0 (low thermal dependence)

High-speed uncooled Lasers

compressive

strain

k

E

lh hh

e

m<0

hh1

e1

e2

lh1

11 1’1

TE, TM

Low dichroism

Polarization-independent devices

tensile

strain unstrained

30

C. Coriasso – Nov 2012

Optical gain in MQW

• high differential gain (dG/dN) step-like shape of density of states

• wider spectral bandwidth deep penetration of Fermi function

0 5 10 15 20 -15

-10

-5

0

5

10

Rel

. Am

plit

ud

e (d

B)

[GHz]

( )thdB IIdN

dG3

31

C. Coriasso – Nov 2012

Distributed Feedback

a+

a-

T

R

B

2.32

24.0

65 1

m

cm

aiikadz

da

ikaaidz

da

T

R

B

DF /4 a+

a-

T

R

B

( )

i

m

cm

4

1

104.32.32

24.0

65

Single- reflectivity:

Single mode laser

32

C. Coriasso – Nov 2012

ATOMS

SEMICONDUCTOR LASER

Quantum Wells:

Atomic-Controlled Artificial Structures

ACTIVE LAYER MQW

QUANTUM WELL

BARRIER

)()()(2 2

22

zEzzVdz

d

m

),,(),,( FNkiFNn

Control of Optical Properties through atomic-scale technology

Quantum Well requires sub-monolayer manufacturing control achievable with

Molecular Beam Epitaxy or Metal Organic Chemical Vapor Deposition.

33

C. Coriasso – Nov 2012

Thanks for your attention!

claudio.coriasso@avagotech.com