Optical Communication Applications of

1
Optical Communication Applications of Fiber
Optical Parametric Amplifiers M. E. Marhic, G.
Kalogerakis and L. G. Kazovsky Photonics
Networking Research Laboratory Department of
Electrical Engineering Stanford University
USA marhic_at_wdm.stanford.edu Sponsors NSF,
OIDA
2
Outline

• Review of fiber OPAs
• Key features
• Possible applications
• Recent advances
• Desirable developments
• Conclusion

3
OPA Review

• Based on fiber third-order susceptibility ?(3)
• Two versions 1 or 2 pumps
• Involves four-wave mixing (FWM), aided by pump
  cross- self-phase modulation (XPM SPM)
• Energy transferred from pump(s) to signal and
  idler.

4
OPA Review
2-pump
1-pump
?p1
?p2
?s
?i
?c
?s ?i 2?p 2?c

• ?s ?i ?p1 ?p2 2?c

Signal idler are symmetric w.r.t. center
frequency ?c
5
OPA Review
?s ?i ?p1 ?p2
Let ? 2pn, where n optical frequency h
Plancks constant
hns hn i hnp1 hnp2
But hn energy of a photon of frequency n

• Conservation of energy (Manley-Rowe relations)
• Each pump releases a photon,
• and signal and idler each gain one,
  simultaneously
• Source of correlated photon pairs

6
OPA Review

• ??? fiber nonlinearity coefficient
• P0 total pump power
• L fiber length
• ??????s ??i -??p1 -??p2
• ???????propagation constant mismatch
• ?(2) ??s ???c ?2 ? ?(4) ??s ???c ?4 /12
  ???
• where ?(m) dm ?/d?m at ??c
• Need ?????0 for high gain gt ?(2) 0
• gt ?c must be near the zero-dispersion frequency,
  ?0

7
OPA Review

• Typical signal gain spectrum

8
Features of Gain Spectrum

• ? Gain bandwidth
• ?????(?P0/ ?(4) ) 1/4
• can be 10s to 100s of nm
• Max gain
• in dBs Gs,max 8.6 ?P0L - 6

9
Key Features of Fiber OPAs

• Fiber medium
• Arbitrary center wavelength
• Large bandwidth, adjustable
• Large signal gain
• Unidirectional (no DRS)
• Low noise figure (amp or converter)
• Fast nonlinearity (fs)
• High cw or average powers (gt1 W)
• Idler
• Large conversion gain (signal gt idler)
• High conversion efficiency (pump gt signal
  idler)
• Phase conjugation/spectral inversion

10
Potential Applications of Fiber OPAs (1)

• Amplifiers
• Broadband, or tunable narrowband
• Novel ranges (1300 nm, S- or U-band, etc.)
• ? conversion
• C to L L to U C to S etc.
• Format independent
• Spectral inversion, phase conjugation
• To mitigate dispersion, nonlinear effects
• High-speed modulation of pump
• Regeneration (2R, 3R)
• Demultiplexing
• Optical sampling
• Amplitude noise reduction

11
Potential Applications of Fiber OPAs (2)

• Quantum effects
• Squeezing
• Correlated photons (entanglement)
• Key distribution
• Teleportation
• Oscillators (OPOs)
• Pulsed or CW
• High-power sources
• High-power pumps ?-conversion
• Multi-Watt sources in novel ranges?

12
Components Used for Telecom Work

• High-power pumps
• Narrow-linewidth, tunable LDs
• High-power EDFAs (several Watts)
• Limited to C and L bands
• Fibers
• DSF ?? near 1550 nm ? 2
  W-1 km-1
• HNL-DSF ?? near 1550 nm ? 20 W-1
  km-1
• (HNL highly nonlinear)
• Limitations ?? varies with distance
  birefringence

13
Typical experimental setup
Fiber parameters ? 2.4 km1W1 ?0 1560 nm ?
0.26 dB/km
14
Stimulated Brillouin Scattering (SBS)

• Caused by sound waves excited by the pump
• Can reflect the pump very efficiently
• Threshold lower than typical OPA pump levels
• SBS must be suppressed for OPAs
• Suppression techniques
• Pump spectrum broadening by FM or PM (Most
  common)
• Isolators
• Air gaps
• Stress distribution
• Dopant distribution

15
Recent Advances

• Wideband amplification
• 1.1. 200-nm gain bandwidth (M.C. Ho et al,
  JLT, p.977, 2001)

• Pulsed pump
• 9 W peak power
• HNL-DSF

16
Recent Advances
2. Narrowband Tunable OPA

• Principle
• ?(2) ??c? ?(3) ??c ???0 ?
• gt By tuning the pump near l0, we can alter the
• shape of Db(w)? ?(2) ??s ???c ?2 ? ?(4) ??s ???c
  ?4 ????????,
• so that Db 0 occurs far from the pump

• Result
• Gain width decreases as distance from pump
  increases

17
Recent Advances
Graphical construction
0
(a)
(b)
(c)
(d)
18
Recent Advances
2.1. 400-nm gain bandwidth, or tunable
spectrum (M. Marhic et al, JSTQE, 2004)

• Pulsed pump
• 12 W peak power
• HNL-DSF

19
Recent Advances
2.2. Tunable narrowband spectrum (M. Marhic et
al, JSTQE, 2004)

• Pulsed pump
• 10 W peak power
• DSF

20
Recent Advances
Magnified Spectra
21
Recent Advances

• 3. Large cw gain
• 49 dB (J. Hansryd et al, PTL, p. 194, 2001)
• 60 dB (K. K. Y. Wong et al, PTL, p. 1707, 2003)

• 1.2 W CW pump
• Isolator in the middle
• (to reduce SBS)

22
Recent Advances
4. Low noise figure
(a) S. French J. Blows, Opt. Lett., p. 491,
2002.. (b) K. K. Y. Wong et al, Opt. Lett., p.
692, 2003. (c) L. Marazzi, PTL, p. 78, 2004. (d)
P. Voss P. Kumar, Opt. Lett, p. 549, 2003.
23
Recent Advances

• 5. Polarization-independent operation
• 1-pump OPA (K. K. Y. Wong et al., PTL., p.1506,
  2002)
• Polarization diversity Counterpropagating
  pumps in a loop

Gain deviation, dB
0
-2
-4
-6
-8
polarization diversity
no polarization diversity
-10
90
60
30
0
Angle of signal SOP, degree
24
Recent Advances
(B) 2-pump OPA (K. K. Y. Wong et al., PTL., p.
911, 2002) Orthogonal pumps copropagating waves
16
14
12
10
Gain, dB
8
6
4
Orthogonal (simulation)
Parallel (simulation)
Orthogonal (experiment)
2
Parallel (experiment)
0
0
20
40
60
80
100
HWP angle, deg
25
(No Transcript)
26
Recent Advances
(C) 2-pump OPA (G. Kalogerakis et al., OFC,
2006) Polarization diversity Counterpropagating
pumps in a loop
27
Recent Advances
6. Large pump depletion (A) 92 in 1-pump
OPA (M. E. Marhic et al., Opt. Lett.., p.620,
2001)
Pump output power, dBm
20
18

• 11 km DSF
• 100 mW pump

16
14
Analytical
Experimental
12
Simulation
10
8
-20
-15
-10
-5
0
5
10
15
20
Signal input power, dBm
28
Recent Advances
(B) 88 in 2-pump OPA (J. M. Chavez Boggio et
al., PTL, p.620, 2003)

• 15 km DSF
• 280 mW pump

29
Recent Advances
7. Pulsed-pump OPAs

• 160 Gb/s transmitter
• (T. Torounidis et al., PTL.,
  p.312, 2005.)
• 640 Gb/s optical sampling
• (S. Watanabe et al., ECOC 2004,
  Th.4.1.6..)
• OTDM DEMUX, 10-40 Gb/s
• (J. Hansryd et al., PTL, p.584, 2001.)
• Regeneration
• (S. Radic et al., PTL, p. 957, 2003.)

30
Recent Advances

• 8. Pulsed pump optical filtering (K. Shimizu et
  al., OFC03, ThT5)
• Rep. Rate gt Nyquist limit
• Benefits of high pump power, CW gain
• Trade-off larger channel spacing

Frequency domain
Time domain
31
Recent Advances
Experimental gain spectra
32
Recent Advances
(G. Kalogerakis et al., J. Lightwave Technol.,
Oct 2005)
Multiple replicas of input signal can provide
multicasting
33
Recent Advances
Effect of Optical Filter
Sampled signal
After optical filter
Pump rep rate 21.234456GHz 2 x signal bit
rate
34
Recent Advances
BER plots for signal replicas
35
Performance for all signal replicas
36
Recent Advances
9. Mid-span spectral inversion
(S. Radic et al., OFC03, PD12).
37
Recent Advances
10. XGM and FWM in WDM systems
(T. Torounidis et al., PTL p. 1061, 2003.)

• XGM cross-gain modulation
• amplified signal instantaneously depletes the
  pump
• pump modulation modulates all signals
• all signals modulate each other via the pump
• FWM four-wave mixing
• 2 or 3 signals mix gt new frequency
• even spacings gt FWM products may fall on
  signals
• uneven spacings gt can avoid this
• (but signal depletion by FWM remains!)

38
Recent Advances
39
Recent Advances
11. Pump-to-signal RIN transfer
(G. Kalogerakis et al., CLEO04, CFA5.)
?? ?RIN magnification
Pump
40
Recent Advances
12. Pump FM to signal IM conversion
(G. Kalogerakis et al., OECC04, 13P-87. A.
Mussot et al., PTL p. 1289, 2004)

• Pump FM needed to suppress SBS
• Typical excursion several GHz
• lp variations modulate gain spectrum
• Linearized theory assuming
• small IM
• constant l0

41
Recent Advances
Theory
Experiments
Pump

• 10-30 IM (combined with IM trannsfer)
• Qualitative agreement w. theory
• l0 variations modify theory
• Improve by 2 pumps with opposite FM

42
Recent Advances

• 13. Random fiber birefringence
• Averaging OPA equations over SOP distributions
• Uniform over Poincare sphere g? gt (8/9) g
• Uniform over great circle g gt (2/3)
  g
• (M. Marhic et al.,
  OFC04, TuC2)
• PMD can lead to
• Gain reductions
• Distortion of gain spectrum
• (E. Yaman, PTL, p. 431, 2004,
• Q. Lin, Opt. Lett., p. 1114, 2004.)

43
Recent Advances
(Q. Lin, Opt. Lett., p. 1114, 2004.)
44
Recent Advances

• 14. Random longitudinal variations of l0
• Can lead to spectrum distortions
  

(E. Yaman et al. PTL, p.1292, 2004)
45
Recent Advances
15. Distributed parametric amplification
Motivation

• Distributed Raman amplification (DRA) is used in
  transmission fibers
• Raman and parametric gain are closely related
• Can distributed parametric amplification (DPA) be
  used for communication signals?

(G. Kalogerakis et al., J. Lightwave Technol.,
Oct 2005.)
46
Recent Advances

• Advantages
• Need only 1/3 of pump power for Raman
  amplification
• No double Rayleigh scattering (DRS)
• Bandwith about 10 nm near l0
• Use near 1310 nm in SSMF?
• Reduction of amp. and link nonlinearities
• Disadvantages
• No backward pumping
• Works best nearl0
• Nonlinear effects in WDM systems?
• Will work best for a few channels

47
Recent Advances
DPA Theoretical Gain Spectra

• 75 km of DSF
• Pump lt 100 mW

48
Recent Advances

• 75 km DSF 15.5 dB total loss
• cancel loss with pump lt 100 mW

49
Recent Advances
Output Spectra
50
Recent Advances
(a) Eye patterns
(b) BER plots for received signal at 1549.14 nm
51
Recent Advances
DPA vs DRA Comparison
DRA is backward-pumped
52
Desirable Developments

• Effective, low-cost SBS suppression
• Improved fibers
• Cost-effective pumps
• Reduction of unwanted nonlinear effects in WDM
  systems

53
SBS Suppression

• Currently pump frequency dithering ( gt 3 GHz)
• problem for ?-conversion
• Possible alternatives
• isolators
• air gaps
• stress
• heating
• special fibers?

54
Pumps

• Laser diodes
• Single frequency
• High power ( gt 100 mW)
• Anywhere in 1200-1700 nm range
• Some tunability
• MOPA configurations ( gt 1 W?)
• Raman-shifted fiber lasers
• Double-clad Yb laser near 1 ?m
• Nested FBG Raman cavities
• Seed with LD gt narrow linewidth?

55
Fibers

• Need
• Better ?? uniformity
• Higher nonlinearity coefficient ?? n? /Aeff
• Reduction in ??
• ??Reduction in PMD
• Increasing ?
• Increasing n? non-silica glasses
• Reducing Aeff holey fibers (microstructured,
  etc.)

56
Fibers
Example New HNLF from Sumitomo

• M. Hirano et al., ECOC 05, Oct. 2005, Glasgow
• post-deadline paper
• Improvements
• ?? uniformity /- 0.1 nm/ 100m
• ????? 25 W-1 km -1 (XPM method)
• ?????????????? -56 s 4 m -1
• PMD 0.05 ps km -1/2

57
l0 Fluctuation of Fiber Samples
1575
Measured by FWM Method (Mollenauer et.
al.)
1570
1565
Zero-Disp. Wavelength l0 nm
Fiber-A
1560
Fiber-B
1555
0
500
1000
1500
2000
Length m
8/12
58
Wavelength Conversion Bandwidths
Normalized Conversion Efficiency dB
0
-4
-8
1450
1500
1550
1600
1650
1700
Probe Wavelength lS nm
10/12
59
Conclusion

• Many major features demonstrated
• Large bandwidth
• High gain
• Low noise figure
• Pulsed operation
• etc.
• Recent emphasis on second-order effects
• Undesired nonlinear interactions
• Fiber imperfections

60
Conclusion

• Progress made possible due to advances in
• Pump sources (high power EDFAs)
• Highly-nonlinear fibers
• Additional progress in
• Pump sources
• Highly-nonlinear fibers
• SBS suppression
• Reduction of higher-order effects
• could lead to practical fiber OPAs