Transcript Nanophotonics - Worcester Polytechnic Institute

Nanophotonics The Emergence of a New Paradigm
Richard S. Quimby
Department of Physics
Worcester Polytechnic Institute
Outline
1. Overview: Photonics vs. Electronics
2. Fiber Optics: transmitting information
3. Integrated Optics: processing information
4. Photonic Crystals: the new paradigm
5. Implications for Education
Electronics
Photonics
1970’s
Tubes & transistors
1960’s
Integrated circuits
1980’s
VLSI
2000’s
Fiber optics
discreet components
1970’s
Planar optical
waveguides
1980’s
Integrated optical
circuits
1990’s
Molecular electronics
Photonic crystals
Electronics
Photonics
fiber
wire
10
f ~ 10
15
Hz
f ~ 10
Hz
sig in
sig out
control beam
v
5
~ 10 m/s
elec
Strong elec-elec interaction
v
8
~ 10 m/s
phot
Weak phot-phot interaction
Advantages of Fiber Optic Communications
* Immunity to electrical interference
-- aircraft, military, security
* Cable is lightweight, flexible, robust
-- efficient use of space in conduits
* Higher data rates over longer distances
-- more “bandwidth” for internet traffic
Erbium Doped Fiber Amplifiers
Advantages:
* Compatible with transmission fibers
* No polarization dependence
* Little cross-talk between channels
* Bit-rate and format transparent
* Allows wavelength multiplexing (WDM)
Disadvantages:
* Limited wavelength range for amplification
Erbium doped
glass
After Miniscalco, in Rare Earth Doped Fiber Lasers and
Amplifiers, M. Digonnet ed.,( Marcel Dekker 1993)
fiber attenuation
after Jeff Hecht,
Understanding Fiber
Optics, (PrenticeHall, 1999)
wavelength
Raman fiber amplifier
hn
scattered
hn
pump
hf
vibration
Signal in
Signal out
* amplification by stimulated scattering
* nonlinear process: requires high pump power
Raman amplifier
gain spectrum
• Can choose pump  for
desired spectral gain region
• typical gain bandwidth is
30-40 nm (~5 THz)
• gain efficiency is quite low
(~0.027 dB/mW)
• compare gain efficiency of
EDFA (~5 dB/mW)
• need high pump power (~1
W in single-mode fiber)
• need long interaction
lengths: distributed
amplification
Wavelength Division Multiplexing
Information capacity of fiber
Spectral efficiency = (bit rate)/(channel spacing)
= (BR)/(10 BR) = 0.1 bps/Hz [conservative]
In C-band (1530 <  < 1560 nm),
f ~ 3800 GHz
Compare: for all radio, TV, microwave, f  1 GHz
Max data rate in fiber = (0.1)(3800 GHz) = 380 Gbs
# phone calls = (380 Gb/s) / (64 kbs/call) ~ 6 million calls
Spectral efficiency can be as high as 0.8 bps/Hz
L-band and S-band increase capacity further
Fiber Bragg Gratings
Periodic index of refraction modulation inside core of optical
fiber:
Strong reflection when  = m(/2)
Applications:
• WDM add/drop
• mirrors for fiber laser
• wavelength stabilization/control for
diode and fiber lasers
How to make fiber gratings:
or:
Using fiber Bragg gratings for WDM
Other ways to separate wavelengths for WDM
Or, can use a blazed
diffraction grating to
spatially disperse the light:
The increasing importance of integrated optics
t/(18 mo.)
* Electronic processing speed ~ 2
(Moore’s Law)
t/(10 mo.)
* Optical fiber bit rate capacity ~ 2
t/(12 mo.)
* Electronic memory access speed ~ (1.05)
Soon our capacity to send information over
optical fibers will outstrip our ability to switch,
process, or otherwise control that information.
Advantages of Integrated-Optic Circuits:
• Small size, low power consumption
• Efficiency and reliability of batch fabrication
• Higher speed possible (not limited by inductance,
capacitance)
• parallel optical processing possible (WDM)
Substrate platform type:
• Hybrid -- (near term, use existing technology)
• Monolithic -- (long term, ultimately cheaper, more
reliable)
• quartz, LiNbO , Si, GaAs, other III-V
semiconductors
Challenges for all-optical circuits
• High propagation loss (~1 dB/cm, compared with ~1
dB/km for optical fiber)
• coupling losses going from fiber to waveguide
• photons interact weakly with other photons -- need large
(cm scale) interaction lengths
• difficult to direct light around sharp bends (using
conventional waveguiding methods)
• electronics-based processing is a moving target
Recent progress toward monolithic platform
GaAs devices
Strontium titanate layer
Silicon monolithic platform
• Recently developed by Motorola (2001)
• strontium titanate layer relieves strain from 4.1% lattice
mismatch between Si and GaAs
• good platform for active devices (diode lasers, amps)
Light modulation in lithium
niobate integrated optic circuit
From Jeff Hecht, Understanding Fiber Optics (Prentice Hall 1999)
Arrayed Waveguide Grating for WDM
* Optical path length difference depends on wavelength
* silica-on-silicon waveguide platform
* good coupling between silica waveguide and silica fiber
after Jeff Hecht,
Understanding Fiber
Optics (Prentice Hall
1999)
Echelle gratings as alternative for WDM
* advances in reactive-ion etching (vertical etched facets)
* use silica-on-silicon platform
* smaller size than arrayed-waveguide grating
* allows more functionality on chip
after Jeff Hecht, Understanding Fiber Optics (Prentice Hall 1999)
Confinement of light by index guiding
• need high index difference for
confinement around tight bends
lower index
cladding
lower index
cladding
• index difference is limited in
traditional waveguides
• limited bending radius achieved
in practice
Examples for Lithium Niobate:
-- thermal diffusion of Ti (n~ 0.025)
higher index
core
-- ion exchange (p for Li) (n~ 0.15)
-- ion implantation
(n~ 0.02)
Photonic crystals: the new paradigm
• light confinement by
photonic band-gap (PBG)
• no light propagation in
PBG “cladding” material
• index of “core” can be
lower than that of
“cladding”
• light transmitted through
“core” with high
efficiency even around
tight bends
Modified spontaneous emission
• First discussed by Purcell (1946)
for radiating atoms in microwave
cavities
• decay rate  #modes/(vol•f)
• if there are no available photon
modes, spontaneous emission is
“turned off”
• more efficient LED’s,
“no-threshold” lasers
• modify angular distribution of
emitted light
Photonic Bandgap (PBG) Concept
Electron moving through
array of atoms in a solid
Photon moving through array
of dielectric objects in a solid
energy
e
bandgap
Early history of photonic bandgaps
• Proposed independently by Yablonovitch (1987) and John
(1987)
• trial-and-error approach yielded “pseudo-PBG” in FCC
lattice
• Iowa State Univ. group (Ho) showed theoretically that
diamond structure (tetrahedral) should exhibit full PBG
• first PBG structure demonstrated experimentally by
Yablonovitch (1991) [holes drilled in dielectric: known
now as “yablonovite”]
• RPI group (Haus, 1992) showed that FCC lattice does
give full PBG, but at higher photon energy
Intuitive picture of PBG
After Yablonovitch, Scientific American Dec. 2001
First PBG material: yablonovite
require n > 1.87
After Yablonivitch, www.ee.ucla.edu/~pbmuri/
Possible PBG structures
after Yablonovitch, Scientific American Dec. 2001
Prospects for 3-D PBG structures
• Difficult to make (theory ahead of experiment)
 top down approach: controllable, not easily scaleable
 bottom up approach (self-assembly): not as
controllable, but easily scaleable
• Naturally occuring photonic crystals (but not full PBG)
 butterfly wings
 hairs of sea mouse
 opals (also can be synthesized)
Photonic bandgap in 2-D
• Fan and Joannopoulos (MIT), 1997
 planar waveguide geometry
 can use same thin-film technology that is currently
used for integrated circuits
 theoretical calculations only so far
• Knight, Birks, and Russell (Univ. of Bath, UK), 1999
 optical fiber geometry
 use well-developed technology for silica-based
optical fibers
 experimental demonstrations
2-D Photonic Crystals
After
Joannopuolos,
Photonic Crystals:
Molding the flow of
light, (Princeton
Univ. Press, 1995)
Propagation along line defect
• defect: remove dielectric
material
light out
• analogous to line of
F-centers (atom vacancies)
for electronic defect
• E field confined to region
of defect, cannot propagate
in rest of material
• high transmission, even
around 90 degree bend
light in
• light confined to plane by
usual index waveguiding
Optical confinement at point defect
• defect: remove single
dielectric unit
• analogous to single F-center
(atom vacancy) for electronic
defect
• very high-Q cavity resonance
after Joannopoulos, jdj.mit.edu/
• strongly modifies emission
from atoms inside cavity
• potential for low-threshold
lasers
Photonic Crystal Fibers
• “holey” fiber
• stack rods & tubes, draw
down into fiber
• variety of patterns, hole
width/spacing ratio
• guiding by:
- effective index
after Birks, Opt. Lett. 22, 961 (1997)
- PBG
Small-core holey fiber
after Knight, Optics & Photonics News, March 2002
• effective index of “cladding” is close to that of air (n=1)
• anomalous dispersion (D>0) over wide  range,
including visible (enables soliton transmission)
• can taylor zero-dispersion  for phase-matching in nonlinear optical processes (ultrabroad supercontinuum)
Large-core holey fiber
d
after Knight, Optics & Photonics News, March 2002

V=
•
•
•
•
2
2
a  ncore
- n2clad

effective index of “cladding” increases at shorter 
results in V value which becomes nearly independent of 
single mode requires V<2.405 (“endlessly single-mode”)
single-mode for wide range of core sizes
Holey fiber with hollow core
• air core: the “holey” grail
• confinement by PBG
• first demonstrated in
honeycomb structure
after Knight, Science 282, 1476 (1998)
• only certain wavelengths
confined by PBG
• propagating mode takes on
symmetry of photonic crystal
Holey fiber with large hollow core
• high power transmission
without nonlinear optical
effects (light mostly in air)
• losses now ~1 dB/m (can be
lower than index-guiding
fiber, in principle)
after Knight, Optics & Photonics
News, March 2002
• small material dispersion
Special applications:
• guiding atoms in fiber by optical confinement
• nonlinear interactions in gas-filled air holes
Implications for education
• fundamentals are important
• physics is good background for adapting to new technology
• photonics is blurring boundaries of traditional disciplines
At WPI:
- new courses in photonics, lasers, nanotechnology
- new IPG Photonics Laboratory (Olin Hall 205)
 integrate into existing courses
 developing new laboratory course
Prospects for nanophotonics
after Dowling,
home.earthlink.net/~jpdowling/pbgbib.html
after Joannopoulos, jdj.mit.edu/