Title: Semiconductor Nanostructures for Photonic Systems of the Future
1Nanotechnology International Forum 2008
Semiconductor Nanostructuresfor Photonic Systems
of the Future
Prof. Dieter Bimberg Institute of Solid State
Physics and Center of Nanophotonics Technische
Universität Berlin
2Contents
- Change of Paradigms Quantum Dots as Giant Atoms
in a Dielectric Matrix - High Speed QD-Lasers and Amplifiers for the 100 G
Ethernet - QD VCSELs for the Terabus
- A Glimpse to the Future Security in
Communication Quantum Cryptography - Higher Brightness for Upconversion, Welding
- New Wavelengths, ...
3Introduction
- Matter is classified
- Gas discrete electronic states
- Liquid very dense discrete electronic states
- Solid energy bands
4Quantum Dots Definition
A Quantum Dot resembles electronically an atom in
a dielectric cage
- Conditions
- Lateral extensions in all three directions of
space smaller than the de Broglie wavelength of
the charge carriers - Not too small ? otherwise no bound state
- Not too large energy difference
betweensublevels gt kT at 300 K
5Pyramids classical ? quantum
Chefren Pyramide
PbSe quantum dots on asemiconductor surface
G. Springholz et al.,Science 282, 734 (1998)
- Size relations
- Base length 1010 times larger - volume 1030
times larger !
6Zero-dimensional Systems are Differentto 3-, 2-
and 1- Dimensional Systems
Impact of
7Contents
- Change of Paradigms Quantum Dots as Giant Atoms
in a Dielectric Matrix - High Speed QD-Lasers and Amplifiers for the 100 G
Ethernet - QD VCSELs for the Terabus
- A Glimpse to the Future Security in
Communication Quantum Cryptography - Higher Brightness for Upconversion, Welding
- New Wavelengths, ...
8Evolution of Network Traffic
x 30 in 5 years
2000
2006
- Increase of traffic from 2001 to 2006 by a factor
of 30? A factor of 1000 every ten years !
9Higher Speed Study Group (HSSG)
- For standardization of 100G Ethernet
- IEEE 802.3ba HSSG Objectives Criteria
- Provide Physical Layer specifications which
support 40 / 100 Gb/s operation - Bandwidth requirements for computing and
networking applications are growing at different
rates. These applications have different cost /
performance requirements, which necessitates two
distinct data rates, 40 Gb/s and 100 Gb/s. - HSSG is predicting that the standard will be
completed in 2010.
10Mode-Locked QD Lasers for 100 G Ethernet
- Advantages of QD gain medium
- Broadband gain medium (30 50 nm ground state,
gt80 nm incl. excited state) - Ultrafast carrier dynamics (10-100 fs components
of carrier relaxation) - Low noise gain medium (record low RIN -160 dB/Hz)
- High temperature stability (T0 inf. below 75
C) - Low threshold current density (lt 10 A/cm2 per dot
layer)
11Threshold Current Densities of Semiconductor
Lasers
after Ledentsov et al., IEEE J. Sel. Top. Quantum
El. 6, 439 (2000)
12Mode-Locked QD Lasers for 1000 GbE
- Advantages of QD gain medium
- Broadband gain medium (30 50 nm ground state,
gt80 nm incl. excited state) - Ultrafast carrier dynamics (10-100 fs components
of carrier relaxation) - Low noise gain medium (record low RIN -160 dB/Hz)
- High temperature stability (T0 inf. below 75
C) - Low threshold current density (lt 10 A/cm2 per dot
layer) - Translate into superior operation as mode-locked
devices - Ultra-short pulse generation (400 fs)
- Very high (80 GHz) / very low (5 GHz) monolithic
MLL repetition rate - Very low timing jitter (lt 200fs) for passive
mode-locking - Temperature stable mode-locking operation up to
70C
13Mode-Locked QD Lasers for 1000 GbE
- Advantages of QD gain medium
- Broadband gain medium (30 50 nm ground state,
gt80 nm incl. excited state) - Ultrafast carrier dynamics (10-100 fs components
of carrier relaxation) - Low noise gain medium (record low RIN -160 dB/Hz)
- High temperature stability (T0 inf. below 75
C) - Low threshold current density (lt 10 A/cm2 per dot
layer) - Translate into superior operation as mode-locked
devices - Ultra-short pulse generation (400 fs)
- Very high (80 GHz) / very low (5 GHz) monolithic
MLL repetition rate - Very low timing jitter (lt 200fs) for passive
mode-locking - Temperature stable mode-locking operation up to
70C - Making QD mode-locked devices ideal candidates
for - High-speed, ultra-short pulse, low noise optical
pulse sources for OTDM / WDM
14Short, high power pulses from QD MLL
InGaAs QD passive MLL at 1280 nm
InAs/InP QD passive MLL at 1550 nm
Thompson et al., SQDA Rennes, 2008
Lu et al., Opt. Expr. 16 (14), 2008
- QD MLL yield shorter pulses than any other
semiconductor gain medium due to their broad gain
spectrum - 2.25 W pulse peak power, 360 fs Fourier-limited
pulse width w/o compression - ? Ideal gain medium for high-speed time division
multiplexing!
1580 GHz QD MLL in QD SOA
Output after SOA 15 increase of pulse width
only
Input 80 GHz QD mode-locked laser
- 80 GHz mode-locking frequency little alteration
of the pulse input - Fits to 100 Gb/s Ethernet
16Double-pulse pump-probe measurement
Recovery ofground state
- Two 150 fs pulses with 5 ps delay
- Investigation of gain recovery after first pulse
DG1 and second pulse DG2 - Complete gain recovery of the ground state for
high current density - Implementation of QD SOAs in 100G Ethernet viable
S. Dommers, D. Bimberg et al., Appl. Phys.Lett.,
90 (2007)
17QD SOA-based Ultrafast Digital Signal Processing
- Bulk/QW-SOA interdependence of refr. index and
gain ? patterning effects - QD-SOA gain and refractive index are decoupled
in regime of maximum gain - phase changes are not accompanied by amplitude
changes - ? PEF-free amplification
- Condition ngtno ? gconst., trecltlt tpulseltlt Tp
- QD based SOAs enable ultrahigh bit rate
amplification (gt100 Gb/s) without patterning - ultrafast optical switching in nonlinear
interferometer with QD SOAs
A. Uskov, D. Bimberg, IEEE PTL 16, 1265 (2004)
18High-Gain Quantum Dot SOAs at 1.3 µm
- Self-assembled QD at 1.3 µm
- 15-fold stacked InGaAs QD layers for maximum gain
- Deep mesa waveguide, width 4 µm, length 4 mm
- Untilted, anti-reflection coated facets for
improved coupling efficiency
- Experimental chip gain up to 26 dB
- Theoretical unsaturated gain 40 dB
- Signal-to-ASE ratio lt 25 dB
- Comparison to simulation shows good
- agreement, influence of gain saturation
19Contents
- Change of Paradigms Quantum Dots as Giant Atoms
in a Dielectric Matrix - High Speed QD-Lasers and Amplifiers for the 100 G
Ethernet - QD VCSELs for the Terabus
- A Glimpse to the Future Security in
Communication Quantum Cryptography - Higher Brightness for Upconversion, Welding
- New Wavelengths, ...
20Optics for Interconnects
A. F. Benner et al., IBM
- Bandwidth is driven by IC scaling and increase in
computational power - Intel, IBM, Fujitsu, NEC, ... work on high-speed
VCSEL devices
21980 nm VCSELs for Short Distance Optical
Interconnects
A.F.J. Levi, Stanford, April 2005, www.usc.edu/ale
vi
Memory-Access Performance Gap
Optical interconnects can fill this gap
J. A. Kash et al, OFC06, OFA3, 2006
VCSEL ARRAY
Terabus, IBM 4x12 VCSEL _at_ 20 Gb/s, 985 nm
22Work on 20 Gbit/s VCSEL
next year
Mutig, ..., Bimberg,El. Lett. 2008
NEC 7 InGaAs/GaAsP strain-compensated QWs,
Proton implanted, oxide confined
Agilent 2 Flip-Chip, 4x12 array
TU Berlin 1 Sub-monolayer grown active region
UCSB 4 Deep oxidation, thick tapered aperture,
bottom emitting, SM
Finisar 6 Proton implanted, thick tapered
aperture
NEC 5 TJ-VCSEL
Cha 8 Two apertures, InGaAs
NEC 3 Proton implanted, oxide confined
1 F. Hopfer et al., ISLC 2006, WC3 2 Chao-Kun
Lin et al., JSTQE, 13 (5) 2007 3 K. Fukatsu et
al., IPRM 2007 4 Y.-C. Chang et al., EL, 43
(19) 2007
5 T. Anan et al., VCSEL symposium, Tokyo Dec.
2007 6 R. H. Johnson et al., CLEO/QELS 2008,
CPDB2 7 H. Hatakeyama et al., CLEO/QELS 2008,
CMW3 8 P. Westbergh et al., EL, 44 (15) and
ISLC 2008
23Novel Concept for gtgt40 Gb/s Modulation Single
Cavity EOM VCSEL
- Keep drive current constant
- Modulate the mirror transmission
N. Ledentsov et al., Proc. of the IEEE, Vol. 95,
No. 9, pp. 1741-1756 (2007)
24Contents
- Change of Paradigms Quantum Dots as Giant Atoms
in a Dielectric Matrix - High Speed QD-Lasers and Amplifiers for the 100 G
Ethernet - QD VCSELs for the Terabus
- A Glimpse to the Future Security in
Communication Quantum Cryptography - Higher Brightness for Upconversion, Welding
- New Wavelengths, ...
25Cyber Security Issue
- Annual U.S. businesses losses because of
computer-related crimes is 67.2 billion (FBI
estimation). - Risk of terrorist invasion in strategic military
and civil systems (nuclear and chemical plants,
access to military information and technologies) - 13,723 computer crimes were committed in Russia
in the calendar year 2004 (Information Center of
the Ministry of Internal Affairs of the Russian
Federation).
26Ensure Secure Data Transmission
Classical cryptographyKey based on
factorisation in prime numbers
27Ensure Secure Data Transmission
Today use of two data channels - fast channel
encrypted signal - slow channel key for
decryption
Quantum cryptography Key based on q-bits or
entangled states
28Single Photon Source (SPS)
Ideal SPS
The ideal light source for a quantum
key-distribution system emits one and only one
photon in each time interval T1/R (photon
gun).
Classical light source described by Poissonian
Statistics probability of producing n
photons (µ - average photon number per pulse)
A Poisson distribution of the number of photons
per pulse from strongly attenuating light pulse
shows no photons in most time slots and two
photons in a small number.
Non-classical (quantum) light source is needed!
29Basic Properties of Quantum Mechanics
- Position and momentum can not be determined
independently - Measurement perturbs the system
- Notion of trajectory no longer valid
- Quantum states Qbits - can not be duplicated
- (non-cloning theorem)
- Measurement of the polarization of a single
photon destroys the initial polarization
Signal transmission security benefits
from quantum mechanics!
30Single Photon Emitter
Novosibirsk
Berlin
The novelty
Oxide
aperture
A
Oxide aperture size
- Current confinement by oxide aperture
- Real single dot excitation
31Single Photon Emitter
Novosibirsk
Berlin
Emission of linearly polarized photons !
- Only exciton recombination is observed
- 870 pA (1.65 V) 0.2 photons / injected
electron-hole-pair
A. Lochmann, D. Bimberg et al., Proc. ICPS-28,
Vienna, late news session, Austria (2006) and
phys. stat. sol. (c) 4, No. 2, 547 (2007)
32Single Photon Emitter
Novosibirsk
Berlin
Hanbury-Brown Twiss Experiment Proof of
nondeterministic (Single Photon) Light Source
A. Lochmann, D. Bimberg et al., Proc. ICPS-28,
Vienna, late news session, Austria (2006) and
phys. stat. sol. (c) 4, No. 2, 547 (2007)
33Single Photon RC-LED
Complex Nanophotonics Device
- Double-DBR periodicity, photonic crystal to
prohibit tilted modes - Single photon source with gt90 efficiency
- Fast (lt0.1 ns)
- Double-DBR RC-LED first time demonstrated 11/2008
34Contents
- Change of Paradigms Quantum Dots as Giant Atoms
in a Dielectric Matrix - High Speed QD-Lasers and Amplifiers for the 100 G
Ethernet - QD VCSELs for the Terabus
- A Glimpse to the Future Security in
Communication Quantum Cryptography - Higher Brightness for Upconversion, Welding
- New Wavelengths, ...
35Background
What we have learned No photonic devices no
fast and secure data transmission
But! 34 of world-wide semiconductor-laser sales
in datacom only !
Diffusion into new fields of application
- Optical memories (blue ray, ...)
- Scanner
- Printer
- Pump laser
- Material treatment
- VCSEL mouse
- Automobile (MOST 3)
36Semiconductor Lasers of High Brilliance
What we have learned No photonic devices no
fast and secure data transmission
But! 34 of world-wide semiconductor-laser sales
in datacom only !
Diffusion into new fields of application
- Optical memories (blue ray, ...)
- Scanner
- Printer
- Pump laser
- Material treatment
- VCSEL mouse
- Automobile (MOST 3)
other3
material treatment6
military/aviation 6
printer10
pump laser 49
medicine 26
Source Laser Marketplace
37Semiconductor Lasers of High Brilliance
- Conventional
- angle of aperture typ. 45
Brilliance (number of photons per solid angle)
is decisive for optical power density at working
point
38Structure of German Center of ExcellenceSemicond
uctor Nanophotonics
- Mainstream nanophotonic materials research on
InP, GaAs, GaN - Advanced numerical modelling
- Advanced device design
- Device RD following mainstream system demands
- Goal Impact on the development of a
multi-billion market
39Gain Synergy
!!! RD in different fields of nano-photonic
applications is conducted by non-interacting
physicists and engineers (uncorrelated). !!!
There are no modular design sets for
nano-photonics like for electronics (logics,
memories, micro processors), ... ! Setting up
such sets is difficult as one has to include all
different kind of scales in time and space
40Our New Center of ExcellenceSemiconductor
Nanophotonics
internationalcooperations
41(No Transcript)
42Our New Center of ExcellenceSemiconductor
Nanophotonics
localcooperations