Nanophotonics in Silicon-On-Insulator Wim Bogaerts

nanophotonics in silicon-on-insulator Wim Bogaerts

IMEC, 9 June 2006

Photonics Research Group http://photonics.intec.ugent.be “the art of playing with light in all its forms” nanophotonics in silicon-on-insulator Wim Bogaerts

IMEC, 9 June 2006

“a trendy prefix indicating the need for nanometer-scale accuracy” “a wafer of silicon buttered with a slice of oxide (insulator) and a delicious layer of silicon”

Photonics Research Group http://photonics.intec.ugent.be Photonics Research Group

• research group of Ghent University • within Engineering Faculty • within Dept. of Information Technology (INTEC) • associated with IMEC • permanent staff: Roel Baets, Peter Bienstman, Geert Morthier, Dries Van Thourhout, Steven Verstuyft • 30-35 researchers (7 post-docs) • research domains: microphotonics and nanophotonics • http://photonics.intec.ugent.be

z 3 decades of photonics research

z Cleanroom facilities for III-V processing

z Photonics modelling

z Characterisation z Associated Lab of IMEC 6 years of collaboration on Silicon Photonics IMEC:

z Microelectronics Research

z Advanced Silicon Processing

z “Nanoelectronics”

z What’s the use?

z How does a waveguide work?

z Photonic Crystals

z SOI Nanophotonics UGent - UGent - IMEC achievements

z What can we do now?

z Some key results Worldwide State-of-the-art Conclusion

z What’s the use?

z How does a waveguide work? Whatz Photonic will Crystalswe use photonics for? z SOI Nanophotonics UGent - IMEC achievements Long-haulz What can we Telecom do now? and FTTH Datacomz Some key (chip, results board, metro) SensorsWorldwide & State-of-the-art spectroscopy Conclusion

medium

radio tower radio set

radio waves over air

TV set TV studio radio waves over coaxial cable

laser photodetector

light packets over an optical fiber (= electromagnetic wave)

radio tower radio set

radio waves over air

TV set TV studio radio waves over coaxial cable

radio tower radio set

radio waves over air

Transmission of signal on carrier wave z Carrier wave has a frequency/wavelength z All carriers travel independent over medium z Receiver can tune in on one carrier

laser photodetector

light packets over an optical fiber (= electromagnetic wave) z Use light of different wavelengths (‘colors’) z Modulate each carrier independently z combine/split carriers at the end of fibers

laser detector

laser detector NCE ISTA H laser NG D WIDT detector LO BAND RGE laser LA detector

Cabinet on the curbside submarine cable

Required: z optical fibres z components in between the fibres z electro-optic conversion at end points LARGE QUANTITIES AND CHEAP

© intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Example: FTTH Triplexer 3 wavelength channels z digital downstream digital downstream z digital upstream (internet, VoIP, digital TV) z analog downstream (CATV)

detector

laser fiber triplexer: split/combine digital upstream 3 wavelength channels (internet, ...) signals on different detector carrier wavelengths

analog downstream (analog TV channels)

digital downstream (internet, VoIP, digital TV)

detector

fiber digital upstream laser dichroic mirrors (internet, ...)

detector

analog downstream (analog TV channels) Align 6 optical elements!!! box = 20x10x10 mm

z Electronics: photonic IC transistors intel 4004 metal wires for 1971 electrical connections between components

z Photonics: waveguides to transport light pentium 4 between components 2002 wavelength filters sources and detectors

nanophotonic IC

z measure reflection of fiber grating

z requires “long” fiber

z complex mounting

Ring resonator

z resonance peak depends on circumference

z peak shifts when strained

Today’s Commercial biosensors: Integrated photonic biosensor:

Elisa test for protein binding 3µm

75 µm

DNA microarray

Æ Complex and slow label-processes Æ No labeling Æ Complex interpretation of results Æ Quantitative and fast results Æ Big amounts of analyte Æ Very small amounts of analyte

z large bandwidth

z no ohmic losses and heating in waveguide

III-V material microlaser SOI waveguide microdetector

Optical Interconnect layer

Electrical Interconnect layer

Silicon transistor layer

a photonic chip

sources

modulators

optical I/O

functional elements

waveguides electrical I/O detectors

z What’s the use?

z How does a waveguide work?

z Photonic Crystals

z SOI Nanophotonics WhatUGent is - I MEClight? achievements Howz What can can we we do guide now? light? Whatz Some is keya goodresults waveguide? Worldwide State-of-the-art Conclusion

© intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Light = Electromagnetic Wave Ray of light ≈ Electromagnetic wave z Propagates at speed of light c z Electrical oscillation E z Magnetic oscillation H z Oscillation frequency f f ×λ= c z with a wavelength λ

H c

wavelength λ

-12 10 cosmic radiation 1020 Visible light: wavelength γ -radiation 380 - 760 nm (nm) -10 röntgen 10 18 (X-rays) 10 violet 400 λ -8 blue 10 ultraviolet 1016 green 500

-6 yellow 10 1014 orange 600 infrared -4 red 10 1012 700

-2 10 Radar 1010 1300

1 TV - radio 108 1550 Light for optical fibre communication

In material: light propagates low refractive index n times slower

n = refractive index

wavelength wavelength becomes n high refractive index times shorter for the same frequency

z Light is partially reflected

Effect is more pronounced with a stronger contrast in refractive index

high refractive index

= Total internal reflection

The critical angle with the surface is larger for a stronger contrast in refractive index (less oblique rays are also reflected)

high refractive index

Light is guided by total internal reflection in a core of high low refractive index refractive index surrounded by a CLADDING cladding of low refractive index CORE

high refractive index CLADDING

low refractive index

Some rays can escape from the waveguide

z Better confinement if the contrast in refractive index is adequately large

z Less loss if the bends are made sufficiently wide Sharp bends possible with large refractive index contrast

© intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Mode of a waveguide Thin core: Rays are an inaccurate model Light is located in a smeared-out ‘blob’ in and around the waveguide core

= a mode

z a mode propagates as low refractive index a single entity

z Guided modes: remain localised around the core

high refractive index

low refractive index

n tio a g a p ro p f o s xi A Core

Cladding Carrier substrate Carrier substrate

Mode 0 (ground mode) is the most useful

z best confinement: Smallest cross section

z most elegant distribution (no zeroes)

012 We’d like a waveguide that only supports a ground mode (= single-mode waveguide)

For telecommunication: Waveguides should guide only a single mode: Core must be sufficiently small

z Optical fibres (low refractive index contrast): Core diameter ~ 10µm

Larger refractive index contrast smaller core

z Waveguides in semiconductor and air (high index contrast): Core ~ 0.2 x 0.5µm.

Reduce bend radius: increase refractive index contrast From 1.46-to-1.44 to 3.45-to-1 Silicon core Silica SEMICONDUCTORS AND AIR buffer

Keep only one guided mode: Reduce dimensions From 10µm to 0.5µm Si substrate ‘PHOTONIC WIRES’

Silica-on-Silicon Silicon-on-Insulator Contrast: 1.46 to 1.44 Contrast: 3.45 to 1 Bend radius = 2cm Bend radius = 3-5µm

z What’s the use?

z How does a waveguide work?

z Photonic Crystals

z SOI Nanophotonics UGent - IMEC achievements z WhatWhat can we do is now? a photonic crystal? z SomeWhat key results can we use it for? Worldwide State-of-the-art Conclusion

source detector

transmission 100%

wavelength

source detector

transmission 100%

100%-R

wavelength

source detector

a transmission 100%

Photonic Band Gap (PBG)

~2a wavelength

source detector

a transmission 100%

Photonic Band Gap (PBG)

~2a wavelength

source detector

a transmission 100%

Photonic Band Gap (PBG)

~2a wavelength

A A

B B Full Photonic Band Gap transmission transmission

B A B A

wavelength wavelength

Periodic structures for light = photonic crystals

2-D 1-D

3-D

High refractive index contrast (larger than 2-to-1) needed for Full photonic band gap

z Only a photonic bandgap for z Only a photonic bandgap for light with the electric field light with the electric field parallel to the pillar axis perpendicular to the pillar (= TM-polarisation) axis (= TE-polarisation)

E E TM TM H H

H H TE TE E E

High refractive index contrast needed

z No light can exist there with a wavelength in the photonic band gap Defect: change holes locally

z Around the defect light can exist with wavelengths in the PBG

z The light cannot propagate away because of the photonic crystal

z e.g. in a line defect light has to follow the defect = a waveguide light cannot ‘miss the bend’

z remove one row of holes = waveguide

z Light is confined by the crystal in the horizontal direction

z Light can spread out in the vertical direction How do we confine the light vertically?

z Light is confined vertically by total internal reflection

z or more correct: a guided mode

photonic crystal waveguide defect holes y x z

slab waveguide

Vertical confinement by Horizontal confinement by a layered waveguide photonic crystal: high index contrast required

z What’s the use?

z How does a waveguide work?

z Photonic Crystals

z SOI Nanophotonics UGent - IMECWhich achievements one to choose? z What canWhy we do now?Silicon-on-Insulator? z Some key results Worldwide State-of-the-art Conclusion

Photonic Crystals: Photonic Wires: z In-plane: Guiding by the z In-plane: Guiding by Total photonic band gap internal reflection z Vertical: Total internal reflection z Vertical: total internal reflection

Both cases: • Details : a few 100 nm • Required precision: <10 nm NANOPHOTONIC waveguides

Look Zog!. This will be the next breakthrough in telecommunications

Solution: Optimise the bend geometry (difficult heavy number crunching)

z roughness

z positioning

z hole size

z ... Etch geometry

z slanted sidewalls

z footing

z roughness

rough sidewalls

Light is being scattered out of the waveguide

More sensitive to scattering at roughness

z change radiation patterns

z change light-matter interaction Very tight confinement

z cubic wavelength cacvities Strong dispersion

z Wavelength-dependent behaviour

z Slow light (IBM: c/300)

Compact functional elements

Photonic crystals: Photonic Wires: z Many possibilities z Simple z Hard to design z Less loss (given good fabrication technology) z Losses

Use for compact Use for waveguides functional elements (connections between elements) Good fabrication technology needed

© intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Silicon-on-Insulator Why this material system? z Transparent at telecom wavelengths (1550nm en 1300nm) z High refractive index contrast in-plane: 3.45 (Silicon) to 1.0 (air holes) out-of-plane: 3.45 (Silicon) to 1.45 (silica)

m 0n 50

m µ Layer structure: 1 • 220nm Si Silica Silicon • 1000nm SiO2

z Transparent at telecom

z High index contrast Guiding of light z Native oxide Splitting, combining light

z Mature technology Wavelength selective functions

z Electronics in same material

But

z No generation of light Sources, detectors z No detection of light Amplicifation z Weak electrooptic effects Modulation, switching : al systems ible) materi f (incompat Range o , ... doped SiO2 LiNbO3, GaAs, InP,

a photonic chip

sources

modulators

optical I/O

functional elements

waveguides electrical I/O detectors

CMOS Technology happy marriage? photonic wires

z Layered structures: z Planar structures each layer one type of All types (photonic structures crystals, dense/isolated wires, ...) together separate litho + etch step One litho for all per layer

z CD tolerances ~ 10% z CD tolerance ~ 10 → 1nm

z Alignment tolerances ~ 50nm z Alignment tolerances ~ 5nm (due to good design) (mode mismatch)

z No problems with roughness z Severe problems with (for lines > 90nm) roughness from 600nm down

z Bag of tricks optimized for z Use best ‘common each type of structure denominator’ process

z SOI: evolution to thin oxide for z SOI: thick oxide needed to thermal dissipation prevent leaking of light

z What’s the use?

z How does a waveguide work?

z Photonic Crystals

z SOI Nanophotonics UGent - IMECHow achievements do we make it? z What canWhat we do now? are the problems? z Some key results

Worldwide State-of-the-art Conclusion

z 3 decades of photonics research

z Cleanroom facilities for III-V processing

z Photonics modelling

z Characterisation z Associated Lab of IMEC 6 years of collaboration on Silicon Photonics IMEC:

z Microelectronics Research Nanophotonic z Advanced Silicon Processing components fabricated z “Nanoelectronics” with CMOS technology

Do you really YES need such high- E-beam lithography resolution lithography? (slow) DUV lithography (in IMEC)

z 220nm Silicon (slightly p-doped)

z 1000nm Silica buffer

220nm Silicon 1000nm Silica

Silicon substrate

z Shipley UV3

z 800nm thick layer

photosensitive resist

z to avoid reflections at the air-photoresist interface

AR-coating

Deep UV Lithography

z ASML PAS5500/750

z KrF, 248nm

z 0.63 NA, 0.4σ

z Dose = 10-40 mJ/cm2

Unexposed areas become solid Exposed areas are dissolved

z reduce roughness

z compensate for litho-etch bias

But

z “Consumes” litho process window

z LAM-POLY

Spot-size Converters Photonic Crystals Gratings

“Silicon Art” Photonic Wires

Problem: photoresist pattern Holes near edges differ from holes in the bulk (while they should be identical!)

hole in the bulk = 420nm

Hole on the edge = 380nm 1um

Hole on the corner = 350nm

z Design: Elliptical holes

z E-beam lithography: holes are elliptical

z Deep UV lithography: holes end up round

Deep etching: Silicon + Oxide Silicon Only etching

T OO SM ROU M OOT G UCH H HE NE R? SS

© intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Still quite rough z still sidewall roughness z sidewall slope 8° (Wtop –Wbottom > 60nm) z (unwanted) etch in the oxide (~20nm) z damage at top edges (resist breakthrough?)

z What’s the use?

z How does a waveguide work?

z Photonic Crystals z SOI NanophotonicsPhotonic Crystals UGent - IMEC achievementsPhotonic wires z What can we doWavelength now? filters z Some key results Worldwide State-of-the-artFibre coupling Conclusion

-5 TE PBG guided mode 7.5 dB/mm -2

[dB/mm] -10 phc α propagation loss -4 -15

-20 -6

Propagation loss -8

other cavity losses Fixed internal cavity losses [dB] -10 1520 1530 1540 1550 1560 1570 1580 1590 1600 Wavelength [nm]

w Propagation losses 400nm 33.8 ± 1.7 dB/cm 440nm 9.4 ± 1.8 dB/cm 450nm 7.4 ± 0.9 dB/cm 500nm 2.4 ± 1.6 dB/cm 40

25 w 20 Si 220nm 15 SiO 1µm Losses (dB/cm) 2 10

5 Si substrate

0 300 350 400 450 500 550 Wire width (nm)

-10.00 Ring resonator demux z 4 rings in series -20.00 z Linearly increasing radius -30.00 z λc does not increase linearly as expected !! -40.00 z Fabrication problem: mask -50.00 discretisation

-60.00 1520 1540 1560 1580

z -3dB insertion loss

z -15dB to -20dB crosstalk 0 FSR=25.3nm

[dB] -10

ssion 1234 8 161

-20 alized tansmi Norm

-30 1.52 1.53 1.54 1.55 1.56 wavelength [µm] © intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Deep and shallow etch Star coupler:

z shallow etch for less contrast

70nm 220nm

500nm

15µm 2µm 5µm oxide

silicon

a photonic chip

sources

modulators

optical I/O

functional elements photonic crystals waveguides electrical I/O AWG, resonators detectors

CMOS Technology photonic wires

The problem 1µm z efficient coupling between a submicrometre waveguide and a fiber

z spot-size converter needed : in plane SOI wire (horizontal) out-of-plane (vertical) : more difficult

z the polarization problem Single-mode fiber

single-mode 1-D grating 1-D grating fiber core z Butt-coupled

z Period ~ 600nm

z 20 periods taper z Etch depth = 45nm

z Optimized design: 31% coupling

air air

airSi

SiOair2 Si

shallow fibre coupler

deep trench

Polarisation maintaining single-mode fibres under 10 degrees angle for incoupling and outcoupling

500nm wide photonic wire acts as a mode filter

2mm long linear taper linear taper

10µm ridge waveguide to Oxide bottom cladding match width of fibre mode Shallow-etched fibre coupler grating for coupling to single-mode fibre © intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Fiber coupler measurements Measurement setup

z no facets needed

z wafer-scale testing becomes possible

0.4µm

0.2µm

80nm z Broad wavelength range z Single mode polished 500 µm facet z Easy to fabricate (if you can do the tips) z Low facet reflections

Shoji et al. EL 38, p.1669 (2002) McNab et al. OpEx 11(22), P. 2927 (2003) Roelkens et al. PTL 17(12), p. 2613 (2005)

But: unwanted trimming of other structures

<100nm

220nm

a photonic chip

sources

modulators

optical I/O

functional fiber couplers elements photonic crystals waveguides electrical I/O AWG, resonators detectors nonlinear functions?

CMOS Technology photonic wires

sources

modulators

optical I/O

functional elements

waveguides electrical I/O detectors

Silicon waveguide structure Heterogeneous circuit Passive only Active + passive

InP InP

Silicon Silicon

InP InP

Polymer bonding layer (BCB)

InP thin film

Silicon substrate 2um

InP/InGaAsP thin film

Measured response of 4 detectors

bonded microlasers a photonic chip

sources

modulators

optical I/O

functional fiber couplers elements photonic crystals waveguides electrical I/O AWG, resonators detectors nonlinear functions?

CMOS Technology photonic wires bonded detectors © intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Applications What can/could we do with SOI nanophotonics?

z Wavelength selective elements for telecom

z Inter-chip and intra-chip communication

z Strain and pressure sensors

z Optical bio-sensors

z Lab-on-a-chip by combining SOI with

z III-V (InP, GaAs) With CMOS z overlay claddings Silicon With microphotonics fluidics z MEMS

z µ-fluidics With III-V

z ...

z Four inputs and four outputs in standard array connector

z 20dB power budget

z 10-15 dB crosstalk in

reference waveguides

© intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Liquid Crystal cladding z Liquid crystal: anistropic material z Surface forces: LC is oriented parallel to the surface z Apply voltage: LC molecules rotate: Change in refractive index

Unpredictable effects: • unknown geometry • bent field-lines 0V 3V 10V

Without voltage Transmission

Voltage 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15V 1500.0 1500.2 1500.4 1500.6 1500.8 1501.0 1501.2 1501.4 1501.6

0.05

0.04

0.03 ansmission r t 0.02

0.01

0 1545 1550 1555 1560 wavelength © intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be SOI microring sensor Measure salt concentration

z Fluid overcladding

z Refr. index ~ Salt concentration

z Response of ring ~ refr. index

z Q = 20000 → minimum ∆n ~ 5.10-5

0.016 2% NaCl 0.012 2,05% NaCl .] .u 2,15% NaCl dλ [a t 0.008 = u 86,6nm/RIU tp

u dn o 0.004 4% NaCl

0.000 1557.50 1557.60 1557.70 1557.80 1557.90 1558.00 wavelength [nm]

z 2 waveguides close together: light leaks

z coupling efficiency ~ waveguide spacing Freely Suspended directional coupler

z Oxide removal

z Vibrations change spacing

Iwijn de Vlaminck, IMEC

First SOI waveguides First AWG Compact with DUV Ring resonators Record-low AWG wire loss Slow light Laser on SOI First fabrication! LC on SOI Router

IST-PICCO NEDO IST-PICMOS ESA-ESTEC 2000 2001 2002 2003 2004 2005 2006 IST-ePIXnet epSOC IUAP-13 IAP-PHOTONnetwork

PICCO1 PICCO2PICCO3 PICMOS1 PICSOI7 PICCO4 PICMOS2 PICSOI8&9 Journals: 1 6(2) 4(1) Conferences: 1 5(1) 6(4) 11(3) 10(4) PhDs students: 3 7 8 11(2) 11(1)

DesignDesign • shared masks • design library

FabricationFabrication • DUV lithography • comparison with other techniques MeasurementsMeasurements • compare measurements • share facilities

z IMEC

z LETI Fabrication for SOI nanophotonics

z Only standard processes (e.g. waveguides)

z Standard modules (e.g. fiber couplers) Support the SOI nanophotonics community

z design libraries and software

z organise exchanges (e.g. for measurements)

z What’s the use?

z How does a waveguide work?

z Photonic Crystals

z SOI Nanophotonics UGent - IMEC achievements z WhatWho can we doare now? the players? z SomeAre key results we ahead or behind? Worldwide State-of-the-art Conclusion

Silicon Photonics

z 500 x 200nm

z 3dB/cm propagation losses

loss ≈ 3 dB/cm for Imbalance on distribution 500nm x 200nm <0.5 dB/cm R= 5 µm 0.04dB/90°

CEA-LETI z Fabricated by a 200mm line and 193nm lithography z integration of InP on SOI z integration of Ge detectors z SOI ridge waveguides (~0.4dB/cm) z Low-T Amorphous Si waveguides z InP processing z Coupling InP microlaser to SOI photonic wire

Yohohama z Fabrication with g-line/i-line z High propagation losses z AWGs, Crossings, Splitter trees

NTT

z E-beam lithography

z Low propagation losses: 6dB/cm

z Low-loss interface to fiber

IBM

z In-house CMOS processes

z e-beam lithography is the only out-of-the-line step

z Photonic crystals: low propagation losses

z Slow light in Photonic crystals (Nature, 3/11)

z Pitch = 420nm, diameter = 195nm

z No observable proximity effects

z Propagation loss: 14dB/cm

Settle et al, OpEx 14(6) - p2440 (March 2006)

z coupled to waveguide

100 Plotted: Speed evolution in recent years 10

1 Colace, Massini & al., Univ. Rome ® Oh & al.,Univ.Texas z Dehlinger & al., Infineon and IBM 0,1 Jutzi & al., Univ. StuttgartSi waveguide Fréquency (GHz) T Dosonmu & al., Univ. Boston and MIT S CEA-LETI & IEF 0,01 1998 2000 2002 2003 2005 Year © intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Silicon sources and modulators INTEL Waveguides integrated in lateral p-(i)-n diode

z Optical modulator @ 10GHz: modulate using current injection

z Silicon CW Raman laser: extract free carriers with reverse bias

SiO2 passivation

Al contact Al contact

Si rib waveguide

h H p-region W n-region

Buried oxide

Si substrate

H = 1.55 µm h = 0.7 µm W = 1.5 µm

signal groundGround High-speed electrical signal groundGround CW light in Bus Waveguide

Ring Waveguide

High-speed modulation region waveguide(pn diode )in p-i-n

z centrosymmetric crystal structure

z no electro-optic effect

Apply Si3N4 strain layer z Deform crystal structure

z Induce electro-optic effect: χ(2) ~15pm V-1

Jacobson et al., nature 04706 (May 2006)

z Fabless Silicon Photonics (Fabrication by Freescale)

z Integration of CMOS and photonic circuits: Waveguides are defined together with transistor gates

z Low-loss rib waveguide

z Grating fiber couplers

Silicon Optical Filters - DWDM Silicon 10G Modulators Flip-chip bonded lasers electrically tunable driven with on-chip circuitry wavelength 1550nm highest quality signal passive alignment integrated w/ control circuitry low loss, low power consumption non-modulated = low cost/reliable enables >100Gb in single mode fiber

Complete 10G Receive Path Ge photodetectors trans-impedance amplifiers output driver circuitry

The Toolkit is Complete 910Gb modulators and receivers Fiber cable plugs here 9Integration with CMOS electronics 9Cost effective, reliable light source 9Standard packaging technology Ceramic Package

Luxtera, Inc. 125 Approved for Public Release DARPA: EPIC Objectives z Si nanophotonics with CMOS processes z Application-specific EPIC z New photonic devices in Si (lasers, wavelength converters, amplifiers, ...) Partners z MIT z Luxtera z Freescale Recent results: z Low-loss waveguides in SOI, made in BAe fab

www.darpa.mil/mto/epic

Photonic IC (PIC) flip-chipped on Electronic IC (EIC)

Photonic layer embedded between metallisations

Combined front-end fabrication

Source: IBM Processor on SOI 0.12µm

raman laser, nanocrystals, ... a photonic chip

sources SiGe or doped Si modulators

optical I/O

functional fiber couplers elements photonic crystals waveguides electrical I/O AWG, resonators detectors nonlinear functions?

CMOS Technology photonic wires Germanium PD © intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Conclusion Silicon is great for nanophotonics z passive waveguide structures z photonic crystals z recently demontrated: modulators, detectors, sources

Many impressive results z good waveguides z wavelength-selective functions z based on CMOS technologies

A strong international interest z Universities: MIT, Kyoto, UCLA, DTU, Valencia, ... z Commercial: INTEL, IBM, NTT, Luxtera, AMO, ...

© intec 2006 - Photonics Research Group - http://photonics.intec.ugent.be Acknowledgements z The Photonic Research Group at Ghent University – IMEC z The European Union IST-PICCO IST-PICMOS IST-ePIXnet z The European Space Agency z The Belgian IAP-PHOTON network z The Flemish Institute for the industrial advancement of Scientific and Technological Research (IWT) z The Flemish Fund for Scientific Research (FWO-Vlaanderen) z The Silicon Process division at IMEC z The P-line at IMEC

I must admit that Silicon nanophotonics made quite an impression