1/20

Multi-Core Fibers and

Their Technical Feasibility

Uncoupled core MCF

Takashi SASAKI
Sumitomo Electric Industries, LTD.

2/20

Outline
1. Background
2. Challenges of multi-core fiber(MCF)
3. Scalability for longer transmission
~ potential of MCF as for crosstalk ~
4. Summary

3/20

1. Background
Limits of the current
transmission system?
Status of telecom
Challenge?
-BW shortage: C+L+
May not be scalable for
single core?
-Fiber fuse
Pth~1.3W (=1.48m)
R. Essiambre et al., Proc OFC'09 OThL1

Candidates of breakthrough technologies for the limits

Multi-level modulation
Multi-mode control
Multi-core fiber (MCF)

3M

4/20

Why MCF is a promising measure?

Spectral efficiency/fiber [bit/s/Hz]

Merit of improving SE using SDM (Shannon limit)

80

7-core
Constant: power / core

70

SE=7log2(1+SNR)

90

60

7-core
Constant: power/ fiber

50
40

SE=7log2(1+SNR/7)

30
20
10

Single-core

SE=log2(1+SNR)
SNR+3dB
SE+1

10

15
20
25
30
Single core SNR (dB)

35

MCF is a promising candidate

to enhance OSNR and avoid Fiber fuse

5/20

2. Challenges of MCF

Densely arranged cores

Single core fiber

7-core fiber

1.

6/20

More cores

Items to be considered: Should design carefully

considering crosstalk
Fiber structure
1) Core pitch :
2) Cladding diameter : D
3) Core- outer cladding distance : d
4) Tight confinement into core

Condition of the evaluation

Fiber bend & twist

Other optical property?

7/20

2.

Fiber loss

OSNR = 58 + Pin NF span (attenuation/Splice loss) 10log10Namp

Enhancement of OSNR is always a MUST for any fiber

Tx
Amp.

Rx
Amp.

Low Loss & Low Nonlinear Fiber

Enhance Linear Transmission

0.330

80km span

-10dB

0.290
-5dB

0.250
0.210

0dB

Aeff [m2]

0.170
0.130

150

130

110

+5dB

90

70

50

SSMF
30

-Low Nonlinearity
-Larger Aeff
-Low Loss
- Fiber attenuation
- connection loss
- Bending loss

Attenuation [dB/km]

Pin :Launch power

span:Span attenuation
NF :noise figure
Namp:Number of amp.

8/20

3.

Inter-core crosstalk

5dB

-20
10dB

-30

15dB
20dB

-40

25dB

Leakage of light

30dB

noise for center core

-50

10 15 20 25 30 35
Center core OSNR [dB]

Crosstalk severely affects OSNR

Crosstalk of two core [dB]

Substantial OSNR of center core

9/20

Options for low-crosstalk design (1)

Decreasing mode coupling coefficient ()
1) Increase power confinement into core
High /small core

Hole/trench-assisted core profile

Fiber A Fiber B
Core

0.7%

0.7%

Core dia. 6.1 m 5.3 m

K. Takenaga et al., IEICE Trans.
Commun., E94-B(2), 409 (2011).

(Left)
K. Saitoh, et al., OECC2010, 7C2-1.
(Center) K. Takenaga et al., OFC2011, OWJ4.
(Right)
T. Hayashi et al., Opt. Express, 19(17), 16576 (2011).

2) Enlarge core pitch

Most of the reported MCFs have core pitch larger than 35 m.

10/20

Options for low-crosstalk design (2)

Apply phase-mismatch () between cores
Heterogeneous MCF

Designed to induce

M. Koshiba et al., IEICE ELEX, 6(2), 98-103 (2009).

K. Imamura et al., Opt. Express, 19(11), 10595 (2011).

Quasi-homogeneous MCF
Structure fluctuation
to induce .
K. Takenaga et al., IEICE Trans. Commun., E94-B(2), 409 (2011).

Homogeneous MCF utilizing fiber bend

Fiber bend to
induce .
T. Hayashi et al., Opt. Express, 19(17), 16576 (2011).

Homogeneous MCF using fiber bend

for low-loss and crosstalk
Designed crosstalk

XT (L=100km) [dB]

Core design
Pure Silica was applied
0.4
[%]

0.2
0
-0.2
-0.4
-0.6
0

10
Radius [m]

Trench-assisted core

11/20

20

-20
-30
-40

R=200mm

=1625nm

XT of
Center
core

XT
-50 between
neighboring cores
-60
30
40
Core pitch [m]

50

Core pitch D was 45 m

12/20

Optical characteristics of fabricated MCF

7
1

6
5

3
4

Optical properties
cc
[nm]

MFD
[m]

Aeff
[m2]

[nm]
Min.
Max.

Loss
Attenuation
Loss [dB/km]

Crossection (OD=150m)
Marker

0.35
0.3
0.25

0.1920.202 dB/km
@ 1625 nm
0.1750.181 dB/km
@ 1550 nm

0.2

0.15
1450 1500 1550 1600 1650 1700
Wavelength [nm]

CD
D. slope
[ps/nm/km] [ps/nm2/km]
1550

1483
1509

9.72
9.88

78.2
81.3

Core 1
Core 2
Core 3
Core 4
Core 5
Core 6
Core 7

C+L band

22.1
22.2

0.062

PMD
[ps/km]

Bending loss
(R = 5 mm)
[dB/turn]

C+L

1625

0.044
0.205

0.019
0.022

Achieved low loss and homogeneous characteristics

13/20
Crosstalk evolves longitudinally by
perturbation and shows statistical distribution

Crosstalk [10-3]

Equivalent
effective
index

Evolution of XT in power

Evolution of XT in I-Q plane

Q

I
I

1.2
0.8
0.4
0.0

0.5
1
1.5
Propagation distance [m]

Direction of the changes

can be varied significantly
by slight perturbation of
the propagation constants

The propagation constant: can be varied by..

Changes of fiber bend & twist
Change of temperature
Change of wavelength

14/20

Statistical measurement of crosstalk

Measured XT spectrum
XT probability distribution
(proposed - scan method)
Crosstalk [dB]
Crosstalk [dB]

8e-7

8e+6

6e-7

-70

-67

-65.2 -64

-63

-62.2

6e+6

4e-7

4e+6

2e-7
0e+0
1620

1625
Wavelength [nm]

1630

1 nm

2e+6
0e+0
0e0 1e-7 2e-7 3e-7 4e-7 5e-7 6e-7
Crosstalk

Mean: XT = 2
*Detailed in
T. Hayashi et al., Opt. Express, 19(17), 16576 (2011).

2 R

15/20

Measured crosstalk by -scan method

Crosstalk [dB]

-65

L=17.4 km, R=140 mm

1625 nm

-70
-75

1550 nm

1
2

Calculated

-80
-85
1

3
4
5
Output core #

Input
core
#

3
4
5
6
7

1550 nm

1625 nm

Average value of two neighboring core

-79.5 dB

-69.8 dB

Crosstalk of center core

-72.3 dB

-62.1 dB

16/20

3. Feasibility for longer transmission

~ potential of MCF as for crosstalk ~

17/20

Crosstalk varies by bend and Length

Bending radius: R

L=17.4 km
=1625 nm

Crosstalk

1.6e-7

Length: L

#1-#4
#1-#5

1.2e-7

#4-#1
8.0e-8

#4-#5
#5-#1

4.0e-8

#5-#4
0.0e+0
0

50
100
150
Bending radius R [mm]
K. Takenaga et al., IEICE Trans. Commun., E94-B(2), 409 (2011).

XT R

L 10 XT +10 dB
XT L

Crosstalk is affected by the length and bend

18/20

XT calculations for longer reach

&

XT-62.1 dB

= 1625 nm

dB

dB

dB

dB

dB

dB

0
-5

dB

When R ~1 m
XT < -30 dB
up to ~3,900 km.

1,000

5
-5
dB
0
-6

XT < -30 dB
when R ~40 cm
after 10,000 km.

dB

Bending radius R [mm]

0
-2

5
-2

0
-3

5
-3

0
-4

5
-4

10,000

2 R
L
XT = 2

100
10

100
1,000
Fiber length L [km]

10,000

19/20

4. Summary
Low-XT is necessary for present transmission technology
Mechanism of crosstalk is investigated and proved lowest
XT MCF, feasible of -30dB after 10,000km transmission
Crosstalk change by microbending must be considered
for the cable manufacturing (Mo.1.LeCervin.3 15:00 )
In near future, more dense MCF expected to be proposed
Expansion of peripheral technologies may open the
real SDM systems while keeping the merit of SDM
transponderconnector amplifiers attenuator monitor?

Multi-Tx
MCF

Multi-Rx

Multi-Amp.

Establishing standard must be the future issue

[ XT / c / outer diameter] [XT meas. method] etc..

Acknowledgement

20/20

This research is supported by the National

Institute of Information and
Communications Technology (NICT), Japan
under "Research on Innovative Optical Fiber
Technology"