Longitudinal fiber parameter measurements of

multi-core fiber using OTDR

Masaharu Ohashi,1,* Yuji Miyoshi,1 Hirokazu Kubota,1 Ryo Maruyama,2 and Nobuo
Kuwaki2
1
Graduate School of Engineering, Osaka Prefecture University, 1-1, Gakuen, Naka, Sakai, Osaka, 599-8531 Japan
2
Fujikura Ltd., 1440, Mutsuzaki, Sakura, Chiba, 285-8550 Japan
*
[email protected]

Abstract: A novel technique is proposed for measuring the longitudinal

fiber parameters of multi-core fiber (MCF). The mode field diameter
(MFD)of a fiber link composed of MCF is successfully estimated with a
modified optical time domain reflectometer (OTDR). The measurement
accuracy of the MFD distribution is revealed by simulation as a function of
the mode coupling coefficient. It is also shown that the relative-index
difference and chromatic dispersion of MCF can be estimated with the
present technique.
©2014 Optical Society of America
OCIS codes: (060.2270) Fiber characterization; (120.4825) Optical time domain reflectometry.

References and links

1. T. Morioka, “New generation optical infrastructure technologies: EXAT initiative: toward 2020 and beyond,” in
Proc. OECC2009, paper FT4(2009).
2. R. Kashyap and K. J. Blow, “Observation of catastrophic self-propelled self-focusing in optical fibres,” Electron.
Lett. 24(1), 47–49 (1988).
3. R.-J. Essiambre, G. Kramer, P. J. Winzer, G. J. Foschini, and B. Goebel, “Capacity limits of optical fiber
networks,” J. Lightwave Technol. 28(4), 662–701 (2010).
4. P. J. Winzer and G. J. Foschini, “MIMO capacities and outage probabilities in spatially multiplexed optical
transport systems,” Opt. Express 19(17), 16680–16696 (2011).
5. M. Koshiba, K. Saitoh, and Y. Kokubun, “Heterogeneous multi-core fibers: proposal and design principle,”
IEICE Electron. Express 6(2), 98–103 (2009).
6. ITU-T Recommendation G.650.1.
7. M. S. O’Sullivan and J. Ferner, “Interpretation of SM fiber OTDR signatures,” Proc. SPIE 86(661), 171–176
(1986).
8. M. Ohashi and M. Tateda, “Novel technique for measuring longitudinal chromatic dispersion distribution in
single-mode fibres,” Electron. Lett. 29(5), 426–427 (1993).
9. K. Nakajima, M. Ohashi, and M. Tateda, “Chromatic dispersion distribution measurement along a single-mode
optical fiber,” J. Lightwave Technol. 15(7), 1095–1101 (1997).
10. A. Rossaro, M. Schiano, T. Tambosso, and D. D’Alessandro, “Spatially resolved chromatic dispersion
measurement by a bidirectional OTDR technique,” IEEE J. Sel. Top. Quantum Electron. 7(3), 475–483 (2001).
11. M. Ohashi, “Novel technique for measuring relative-index difference of fiber links,” IEEE Photon. Technol.
Lett. 18(24), 2584–2586 (2006).
12. M. Wuilpart, G. Ravet, P. Megret, and M. Blondel, “Distributed measurement of Raman gain spectrum in
concatenations of optical fibres with OTDR,” Electron. Lett. 39(1), 88–89 (2003).
13. K. Toge, K. Hogari, and T. Horiguchi, “Raman gain efficiency distribution measurement in single-mode optical
fibers by using backscattering technique,” IEEE Photon. Technol. Lett. 17(8), 1704–1706 (2005).
14. Y. Tsutsumi and M. Ohashi, “Indirect technique for measuring Raman gain efficiency spectrum using OTDR,”
IEICE J95-B(2), 146–154 (2012) (in Japanese).
15. K. Takenaga, Y. Arakawa, S. Tanigawa, N. Guan, S. Matsuo, K. Saitoh, and M. Koshiba, “An investigation on
crosstalk in multi-core fibers by introducing random fluctuation along longitudinal direction,” IEICE Trans.
Commun., E-94-B(2), 409–416 (2011).
16. M. Ohashi, K. Kawazu, A. Nakamura, and Y. Miyoshi, “Simple backscattered power technique for measuring
crosstalk of multi-core fibers,” in Proc. OECC2012, 357–358 (2012).
17. N. Shibata, M. Kawachi, and T. Edahiro, “Optical loss characteristics of high-GeO2 content silica fibers,” IEICE
Trans. E63(12), 837–841 (1980).
18. C. Pask, “Physical interpretation of Petermann’s strange spot size for single-mode fibres,” Electron. Lett. 20(3),
144–145 (1984).
19. D. Marcuse, “Loss analysis of single-mode fiber splices,” Bell Syst. Tech. J. 56(5), 703–718 (1977).
20. M. Nakazawa, M. Yoshida, and T. Hirooka, “Nondestructive measurement of mode couplings along a multi-core
fiber using a synchronous multi-channel OTDR,” Opt. Express 20(11), 12530–12540 (2012).

#222453 - $15.00 USD Received 4 Sep 2014; revised 5 Nov 2014; accepted 5 Nov 2014; published 25 Nov 2014
(C) 2014 OSA 1 December 2014 | Vol. 22, No. 24 | DOI:10.1364/OE.22.030137 | OPTICS EXPRESS 30137
21. E. P. Patent, 0982577–A1, “ Device for measuring crosstalk between multicore optical fibres,” (1999).

1. Introduction
As the power launched into communications fiber has increased to cope with the rapid
increase in transmission capacity, optical nonlinear effects and the fiber fuse phenomenon
have become a problem [1, 2]. Because of these effects, the transmission capacity of
conventional fiber is approaching its fundamental limit [3]. Recently, the research and
development of novel optical fibers have been undertaken throughout the world with the goal
of overcoming the above limit [4, 5].
One innovative optical fiber is multi-core fiber (MCF) which accommodates a number of
single-mode fiber cores, each of which carries a different signal. In MCFs, the mode field
diameter (MFD), cutoff wavelength and optical loss are important parameters as well as the
crosstalk, namely the maximum power transferred between the cores. There have been many
reports on techniques for measuring the MFD, cutoff wavelength, optical loss and chromatic
dispersion. The MFD and cutoff wavelength are especially important and there are many
techniques for their measurements including the far-field pattern technique [6], the near field
pattern technique [6], and the optical time reflectometer (OTDR) technique [7–10].
In particular, the OTDR technique provides a very simple way of measuring the MFD and
it can measure its longitudinal distribution [9, 10]. Moreover, there have been many reports
on the use of OTDR techniques to estimate the relative-index difference [11], chromatic
dispersion [8–10], and Raman gain coefficients [12–14] in a fiber link. However, there have
been few reports on the use of OTDR to measure the fiber parameters of MCFs.
In this paper, we propose a simple technique for measuring the longitudinal fiber
parameters and the crosstalk of MCF based on an improvement we made to conventional
OTDR. We reveal the measurement accuracy of the MFD distribution as a function of the
mode coupling coefficient by simulation. We successfully estimate the MFD, relative-index
difference and chromatic dispersion distributions of a fiber link composed of MCF with our
proposed technique.
2. Principle of our proposed technique
2.1 Bi-directional OTDR technique
The experimental setup for measuring the longitudinal fiber parameters of MCF is shown in
Fig. 1.

Fig. 1. Experimental setup for measuring MFD.

We consider a homogeneous seven-core MCF as shown in Fig. 1. Thus, we assume that

the mode coupling coefficients h between center and outer cores are the same and the mode
coupling coefficients between outer cores are also as the same. Here, we consider the
longitudinal fiber parameter of the center core in MCF. When the input power P0(0) is
launched into the center core of MCF, the powers of center and outer cores can be obtained
by solving the coupled power equations with regards to seven cores. P1(z) and P2(z) denote
the powers of center core and one of outer cores with respect to fiber length z, respectively.
Both powers P1(z)and P2(z) can be obtained as [15]
1 + 6 exp( −7hz )  (1)
P1 ( z ) = P0 (0)  
 7

#222453 - $15.00 USD Received 4 Sep 2014; revised 5 Nov 2014; accepted 5 Nov 2014; published 25 Nov 2014
(C) 2014 OSA 1 December 2014 | Vol. 22, No. 24 | DOI:10.1364/OE.22.030137 | OPTICS EXPRESS 30138
 1 − exp( −7hz )  (2)
P2 ( z ) = P0 (0)  
 7
Two different kinds of reference fiber with a length of L0 as shown in Fig. 1 are employed
to determine the absolute value of the longitudinal fiber parameter before an MCF of length
L. By using the Eqs. (1) and (2) and reference [16], OTDR signals of the center core S1(z) and
S2(z) (in dB) launched from the opposite ends (subscripts 1 and 2) of a fiber of length L + L0
can be expressed as
 10 log P1 + 10 log(α s ( z )) + 10 log( B( z )) − 2αz10 log e ( 0 ≤ z ≤ L0 )
10 log P1 + 10 log(α s ( z )) + 10 log( B( z )) − 2αz10 log e (3)
S1 ( z ) =  ,
+ 10 log 1 + 6 exp[− 14h( z − L0 )] ( L0 + L ≥ z ≥ L0 )
 
 7

 10 log P2 + 10 log(α s ( z )) + 10 log( B( z )) − 2α (L + L0 − z )10 log e


 + 10 log 1 + 6 exp[− 14hL] (0 ≤ z ≤ L0 ) (4)
  7 
S2 ( z) =  ,
10 log P2 + 10 log(α s ( z )) + 10 log( B( z )) − 2α (L + L0 − z )10 log e

 1 + 6 exp[− 14h( L + L0 − z )]
+ 10 log   ( L0 + L ≥ z ≥ L0 )
  7

where P1 and P2 are the input powers, αs the local scattering coefficient, B the backscattered
capture fraction, and α the local attenuation coefficient. h denotes the mode coupling
coefficient between center core and outer cores in an MCF.
Equations (3) and (4) include the optical losses caused by optical power decay,
imperfection and mode coupling. When we estimate the longitudinal fiber parameter of the
core in an MCF using the bidirectional OTDR technique, we need to know the loss induced
by mode coupling between center core and outer cores. The imperfection loss contribution
U(z) [7] excluding the mode coupling loss can be obtained as
U ( z ) = 5 log P1P2 + 10 log(α s ( z ) B ( z )) − 2α ( L + L0 )(10 log e )
 S1 ( z ) + S 2 ( z ) 1 + 6 exp[− 14hL ] (5)
 − 5 log   ( 0 ≤ z ≤ L0 )
 2  7 
=  S1 ( z ) + S 2 ( z )  [1 + 6 exp[− 14h( z − L0 )]][1 + 6 exp[− 14h( L + L0 − z )]]  .
− 5 log  
 2  49 

 ( L0 ≤ z ≤ L0 + L)
The backscatter capture fraction B is expressed using the refractive index of the core n and
the MFD 2w(λ,z) as [7]
2
3  λ  (6)
B(λ , z ) = log  .
2  2πnw(λ , z ) 
The imperfection loss contribution Un(z) normalized by the value at the first reference
point z = z0 can be expressed as
 α ( z ) n 2 ( z0 )   2 w(λ , z0 )  (7)
U n ( z ) = U ( z ) − U ( z0 ) = 10 log s  + 20 log 2 w(λ , z ) .
α (
 s 0 z ) n 2
( z )   
With conventional single-mode fibers, the variation in the local scattering coefficient αs is
negligible compared with that in the MFD. The second reference point z = z1 satisfies Eq. (7).
Therefore, the longitudinal MFD distribution of the core in the MCF can be estimated by
using the normalized imperfection loss and MFDs at the two reference points z0 and z1 as [10]
U ( z ) −U ( z 0 )
 2w(λ , z0 ) U ( z1 ) −U ( z 0 ) (8)
2w(λ , z ) = 2w(λ , z0 )  .
 2w(λ , z1 ) 
To obtain the imperfection loss contribution U(z), we have to estimate the crosstalk
between cores. When the OTDR pulse is launched into center core of an MCF with a length

#222453 - $15.00 USD Received 4 Sep 2014; revised 5 Nov 2014; accepted 5 Nov 2014; published 25 Nov 2014
(C) 2014 OSA 1 December 2014 | Vol. 22, No. 24 | DOI:10.1364/OE.22.030137 | OPTICS EXPRESS 30139
L, the backscattered powers S3(z) ( = 10logP3(z)) of center core and S4(z) ( = 10logP4(z)) of
the outer core in dB from a given position z in an MCF are expressed as
1 + 6 exp( −14hz )  (9)
S3 ( z ) = 10 log P3 + 10 log(α s1 B1 ) − 2α1 z (10 log e) + 10 log ,
 7

1 − exp( −14hz )  (10)

S 4 ( z ) = 10 log P3 + 10 log(α s 2 B2 ) − 2α 2 z (10 log e) + 10 log ,
 7

where P3 is the input power, αsi (i = 1,2) the local scattering coefficient, and Bi (i = 1,2) the
backscattered capture fraction, and αi(i = 1,2) the local attenuation coefficient.
Here, if we assume that αs1 = αs2 = αs, α1 = α2 = α, and B1 = B2 = B, the crosstalk (XT)
between two cores in an MCF estimated using the OTDR is obtained with Eqs. (9) and (10) as
the following equation.
 P ( z)   1 − exp( −14 hz )  (11)
XT ( z ) = 10 log  4  = 10 log  .
P (
 3  z ) 1 + 6 exp( −14 hz ) 
Note that the crosstalk given by Eq. (11) estimated using the OTDR is that for a fiber
length of 2z. The average mode coupling coefficient h(z) over the fiber length z can be
estimated from Eq. (11). Thus, the imperfection loss contribution U(z) can be estimated by
taking the mode coupling coefficient h(z) into account. The MFD distribution of the core in
an MCF can be measured with our present technique.
Once the longitudinal MFD distribution is known, the relative-index difference Δ(z) (in
%) can be estimated as [11]
  2 w ( λ , z0 ) 
U ( λ , z ) −U ( λ , z 0 ) − 20 log   
1  2 w( λ , z ) 
 (12)
Δ( z ) = (1 + kΔ( z0 ) ) × 10 10
− 1,
k 
 

where k is a proportional constant of the Rayleigh scattering coefficient to the dopant

concentration of GeO2-doped core fiber.
Moreover, we can estimate the longitudinal chromatic dispersion D(z) by using the MFDs
and the relative-index difference Δ. The chromatic dispersion D is given as the sum of the
waveguide dispersion Dw and the material dispersion Dm. The material dispersion Dm and
waveguide dispersion Dw are expressed as [16]
D = Dm + D w , (13)

λ d 2n (14)
Dm = − ,
c dλ2
λ d λ  (15)
Dw =  ,
2π 2cn dλ  w2 

where c is the light velocity and n is the refractive index of the core.
The material dispersion Dm can be estimated from the dopant concentration in an optical
ðber by using Sellmeier’s relation [17]. By contrast, the waveguide dispersion Dw can be
estimated by determining the wavelength dependence of the MFD [18]. We approximated the
wavelength dependence of the MFD [19] as
w(λ ) = g 0 + g1λ1.5 . (16)
Substituting Eq. (16) into Eq. (15), the waveguide dispersion can be expressed as
 3g1λ1.5 
λ (17)
Dw = 1 − .
2π cnw 
2
w 
2

#222453 - $15.00 USD Received 4 Sep 2014; revised 5 Nov 2014; accepted 5 Nov 2014; published 25 Nov 2014
(C) 2014 OSA 1 December 2014 | Vol. 22, No. 24 | DOI:10.1364/OE.22.030137 | OPTICS EXPRESS 30140
 Here, when the MFDs are given at two wavelengths, the coefficient g1 is determined by
solving Eq. (16) with regards to the two wavelengths.
2.2 Measurement accuracy of MFD
We discuss the effect of the mode coupling between cores on the MFD estimation. The
imperfection loss I(z) including the mode coupling loss contribution can be obtained from Eq.
(3) as
 1 + 6 exp[− 14hL]
 U ( z ) + 5 log  7  (0 ≤ z ≤ L0 )
(18)
 
I ( z) =   [1 + 6 exp[− 14h( z − l0 )]] [1 + 6 exp[− 14h( L + l0 − z )]]  .
U ( z ) + 5 log  
  49 

 ( L0 ≤ z ≤ L0 + L)

Note that I(z) corresponds to U(z) when there is no mode coupling in the MCF.
We examined the MFD estimation with and without mode coupling as a function of the
mode coupling coefficient h for the fiber link. 2w*(λ, z) denotes the MFD that we estimated
from the imperfection loss I(z). Here, we assumed that the MFD can be estimated using Eq.
(8) accurately. Therefore, we consider the principle relative error Δε of the MFD estimation
caused by the mode coupling between cores as
F (λ , z,h)
2 w(λ , z ) − 2 w* (λ , z ) 2 w* ( λ , z )  2 w(λ , z0 )  U ( z1 )−U ( z0 )
Δε = =1− = 1 −   , (19)
2 w(λ , z ) 2 w(λ , z )  2 w(λ , z1 ) 
F (λ , z, h ) = 5 log
[1 + 6 exp( −14h(z − L0 )][1 + 6 exp( −14h(L + L0 − z )] .
7[1 + 6 exp( −4hL)]
The imperfection loss difference Un(z1) = U(z1)-U(z0) between two reference points can
be expressed in terms of relative-index difference Δ and MFD 2w as
 {1 + kΔ ( z1 )}{1 − 2 Δ ( z 0 )}  2 w( λ , z 0 )  (20)
U n ( z1 ) = U ( z1 ) − U ( z 0 ) = 10 log   + 20 log  2 w(λ , z ) 
 {1 + kΔ ( z 0 )}{1 − 2 Δ ( z1 )}  1 

We calculated the relative error using Eqs. (19) and (20). In calculations, we used the
parameters of the two reference fibers in Table 1.
Table 1. Parameters of Test MCF and Reference Fibers used in calculations
Parameters Ref #1 Ref #2
MFD at 1.55 μm (μm) 9.5 8.0
Relative-index difference Δ (%) 0.4 0.9

Figure 2 shows the relationship between fiber length and relative error Δε (in %) of MFD
as a function of mode coupling coefficient h. Here, the length of MCF was assumed to be
4km.

#222453 - $15.00 USD Received 4 Sep 2014; revised 5 Nov 2014; accepted 5 Nov 2014; published 25 Nov 2014
(C) 2014 OSA 1 December 2014 | Vol. 22, No. 24 | DOI:10.1364/OE.22.030137 | OPTICS EXPRESS 30141
 2
10
0
h=0.1 (1/km)
10 h=1 (1/km)

Relative error (%)

10
-2 h=1x10-2 (1/km)

-4
10 h=1x10-3 (1/km)

-6
10 h=1x10-4 (1/km)

-8
10
0 0.5 1 1.5 2 2.5 3 3.5 4
Fiber length (km)

Fig. 2. Relationship between fiber length and relative error Δε of the MFD as a function of
mode coupling coefficient h.

It is seen that the relative error is maximized at z = 2 km, which corresponds to the center
position of the MCF. This is because the function of F(λ,z,h) becomes minimum in the center
position of the MCF. It is also seen that the principle relative error decreases as h decreases.
The maximum relative errors for the fiber lengths L of 4km, 20km, and 40km were
calculated as a function of mode coupling coefficient as shown in Fig. 3.
2
10
Maximum relative error (%)

1
10

10
0 L/2=20km

-1
10 L/2=10 km
-2
10 L/2=2 km
-3
10
-4
10
-3 -2 -1 0
10 10 10 10
Mode coupling coefficient h (1/km)

Fig. 3. Relationship between the mode coupling coefficient h and maximum relative error
Δε of the MFD as a function of fiber length.
It is seen that the maximum relative error increases as the mode coupling coefficient h
(1/km) increases. It is also seen that the relative error increases with fiber length of MCF
when the h is given. As a result, it is clarified from Fig. 3 that MFD of 4km long MCF with a
mode coupling coefficient h of less than 0.02 (1/km) can be estimated with an accuracy of
better than 1% without taking the mode coupling between cores into account. Therefore, in
this case, we can estimate the longitudinal fiber parameters of an MCF by using the
conventional bidirectional OTDR technique [9, 10].
Next, we examine validity of the approximation when deriving Eq. (12) regarding
crosstalk. The crosstalk between center and outer cores is obtained from Eqs. (10) and (11) as
XT ( z ) = S 4 ( z ) − S 3 ( z )
(21)
α B  1 − exp( −14hz ) 
= 10 log( s 2 2 ) + 2(α1 − α 2 ) z (10 log e) + 10 log .
α s1 B1 1 + 6 exp( −14hz ) 
Equation (21) can be rewritten in terms of refractive index, relative-index difference Δ and
MFD as

#222453 - $15.00 USD Received 4 Sep 2014; revised 5 Nov 2014; accepted 5 Nov 2014; published 25 Nov 2014
(C) 2014 OSA 1 December 2014 | Vol. 22, No. 24 | DOI:10.1364/OE.22.030137 | OPTICS EXPRESS 30142
  1 − exp( −14hz ) 
XT ( z ) = g ( z, λ , Δ1 , Δ2 , w1 , w2 ) + 10 log   (22)
1 + 6 exp( −14hz ) 
 (1 + kΔ2 )   n12 w12  R0k
g ( z , λ , Δ1 , Δ2 , w1 , w2 ) = 10 log   + 10 log  2 2  + 4 ( Δ1 − Δ2 )2 z (10 log e).
 (1 + kΔ1 )   n2 w2  λ

where the function g (in dB) represents the ratio of Rayleigh scattering coefficient αs, the
backscatter capture fraction B and optical loss in the outer core to that in the center core.
n1and n2 denote the refractive indexes of the center and outer cores, respectively.
The relative error Δε between the exact hth value in Eq. (22) and approximated hap value in
Eq. (11) can be derived as
 XT − g XT

 1 − 10 10 1 + 6 ⋅ 10 10 
ln (23)
 XT XT − g

h − hap  1 − 10 10 1 + 6 ⋅ 10 10 
Δε = ex ==  .
hex  XT − g

1 − 10 10
ln  
 XT − g

1 + 6 ⋅ 10 10

Figure 4 shows the relationship between the relative error Δε (in %) and g (in dB) as a
function of XT (in dB).
50
λ=1.55 μm
Relative error of h (%)

0 XT(z)=-10 dB

-50
-20 dB

-60 dB
-100
-3 -2 -1 0 1 2 3
g (dB)

Fig. 4. Relationship between g value and relative error Δε of the h as a function of crosstalk
between center and outer cores.

It is seen that the absolute value of relative error increases as g value move away from
zero. It is also seen that h is underestimated as the g value is positive and h is overestimated
as the g value is negative. Moreover, it is also seen that when the g value of 3dB, the relative
error becomes 100% and the approximated hap value corresponds to twice exact hex value. For
example, we consider the 4km long MCF, which is composed of a center core of conventional
single-mode fiber (SMF) and six outer cores of dispersion-shifted fibers (DSFs). By using the
parameter values in Table 1, the g value was obtained to be 0.92 dB. Here, a k value of 0.62
was used. The relative error of h is about −23% from Fig. 4. Therefore, the assumption in Eq.
(11) approximately holds taking the relative error of MFD in Fig. 3 into account if the
average h is assumed to be 10−2 (1/km) for the fiber length of 4km.
3. Experiments
3.1 MFD distribution
To confirm the effectiveness of our present technique, we measured the MFD distribution
along a fiber link composed of two single core fibers (Ref #1, Ref #2) and a test MCF with
seven cores. The parameters of these fibers are listed in Table 2. Ref #2 was spliced with the
center core in the test MCF as shown in Fig. 1.

#222453 - $15.00 USD Received 4 Sep 2014; revised 5 Nov 2014; accepted 5 Nov 2014; published 25 Nov 2014
(C) 2014 OSA 1 December 2014 | Vol. 22, No. 24 | DOI:10.1364/OE.22.030137 | OPTICS EXPRESS 30143
 Table 2. Parameters of Test MCF and Reference Fibers
Parameters MCF Ref #1 Ref #2
MFD at 1.55 μm (μm) 10.4 7.9 10.3
at 1.31 μm (μm) — 6.4 9.2
Cutoff wavelength λc (μm) 1.21 1.11 1.23
Relative-index difference Δ (%) — 0.915 0.41
Core pitch Λ (μm) 40.46 — —
Outer diameter (μm) 179.4 125.1 125.0
Chromatic dispersion at 1.55 μm D (ps/km/nm) — 0.9 15.5
Crosstalk* at 1.55 μm (dB) −24.9 — —
*Average crosstalk between the center and outer cores.

Figure 5 shows the experimental setup for measuring the crosstalk in an MCF using a
modified OTDR [16, 20, 21]. The modified OTDR consists of a conventional OTDR
including a light source and a detector, a circulator, and a switch. The backscattered power of
the adjacent core is measured using the switch. The coupling between the conventional
single-mode fiber and each core of the multi-core fiber was realized with a space coupling
device as shown in Fig. 5.

Fig. 5. Experimental setup for measuring crosstalk using modified OTDR.

Figures 6(a) and 6(b) show the bidirectional backscattered powers of the fiber link at λ =
1.31 and 1.55 μm, respectively. The modified OTDR at wavelengths of both 1.31 and 1.55
μm was used to measure the backscattered signal powers for the link. In our measurements,
the OTDR pulse width was set at 1μs and the averaging time was 3 min.

Fig. 6. Bidirectional backscattered powers of fiber link at (a) 1.31 μm and (b) 1.55 μm.

Figure 7(a) and 7(b) show the backscattered powers of the center and outer cores of the
test MCF at λ = 1.31 and 1.55 μm obtained using the modified OTDR when the optical pulse
from the OTDR was launched into the center core of the MCF. The backscattered power from
the outer core was measured by switching the output port to the outer core as shown in Fig. 4.
In the OTDR measurements, the pulse width of the OTDR was 1 μs and the averaging time
was 3 min.

#222453 - $15.00 USD Received 4 Sep 2014; revised 5 Nov 2014; accepted 5 Nov 2014; published 25 Nov 2014
(C) 2014 OSA 1 December 2014 | Vol. 22, No. 24 | DOI:10.1364/OE.22.030137 | OPTICS EXPRESS 30144
 Fig. 7. Backscattered powers of center and outer cores in MCF when an optical pulse was
launched into the center core. (a) 1.31 μm, and (b) 1.55 μm.

Figure 8 shows the crosstalk between the cores for the test MCF at λ = 1.31 and 1.55 μm.
The crosstalk was estimated from the backscattered powers of the center and outer cores as
shown in Fig. 7. We found that the estimated crosstalk of L = 4km at 1.55 μm was in good
agreement with the value measured with the conventional power method. We also found that
the estimated crosstalk at 1.31 μm was about −55 dB at a distance of 4km. The difference
between the crosstalk values at 1.31 and 1.55 μm was about 30dB at L = 4 km.
-25
-30 λ=1.55 μm

-35
Crosstalk (dB)

-40
-45
-50
λ=1.31 μm
-55
-60
0 0.5 1 1.5 2 2.5 3 3.5 4
Fiber length (km)

Fig. 8. Crosstalk between cores at λ = 1.31 and 1.55 μm.

Figure 9 shows the average mode coupling coefficient h at λ = 1.55 μm as a function of

fiber length. The average mode coupling coefficient h over the fiber length can be estimated
by using Fig. 8 and Eq. (12).
-4
5 10
-4
4 10
Average h (1/km)

-4
3 10
-4
2 10
-4
1 10 λ=1.55 μm

0
0 0.5 1 1.5 2 2.5 3 3.5 4
Fiber length (km)

Fig. 9. Relationship between fiber length and average h at 1.55 μm.

The average mode coupling coefficient over entire fiber length of 4km was obtained to be
6.2x10−4 (1/km) from Fig. 9. The crosstalk of the MCF at 4km was about −26dB at 1.55 μm.

#222453 - $15.00 USD Received 4 Sep 2014; revised 5 Nov 2014; accepted 5 Nov 2014; published 25 Nov 2014
(C) 2014 OSA 1 December 2014 | Vol. 22, No. 24 | DOI:10.1364/OE.22.030137 | OPTICS EXPRESS 30145
This is because the seven cores were step-index profiles and the core pitch Λ was about 40
μm.
Figure 10 shows the imperfection loss contribution U(z) of the fiber link at 1.55 μm. The
imperfection loss was obtained from the bidirectional OTDR in Fig. 5 and the crosstalk
between the cores in Fig. 7.
80

Imperfection loss U(z) (dB)

78 λ=1.55 μm

76

74 Ref #1

72
Ref #2 MCF
70

68
0 1 2 3 4 5 6 7 8
Fiber length (km)
Fig. 10. Imperfection loss contribution U(z)of fiber link at 1.55 μm.

Figure 11 shows the MFD distribution 2w(z) of the fiber link at λ = 1.55 μm. It is seen that
the MFD of MCF decreases as the fiber length increase. We found that the experimental
result for the MFD of the test MCF agreed well with the value measured with the far field
pattern (FFP) technique as shown in Table 1. As a result, we confirmed that the MFD
distribution was successfully estimated by our proposed technique.
11
Mode field diameter 2w (mμ)

10 Ref #2 MCF

9
Ref #1
8

7
λ=1.55 μm
6
0 1 2 3 4 5 6 7 8
Fiber length (km)
Fig. 11. MFD distribution of test fiber link at λ = 1.55 μm.

3.2 Other longitudinal distributions of MCF

The longitudinal properties of the MCF were measured with the OTDR technique as
mentioned in Section 2.
Figure 12 shows the relative-index difference distribution Δ(z) of the fiber link. The
relative-index difference Δ was estimated by using Eq. (12). It is seen that the relative-index
difference of the test MCF is almost the same as that of the conventional single-mode fiber,
Ref #2. The Δ of Ref #2 was used to estimate the relative-index difference of the MCF.
Therefore, the Δ value of Ref #1 is slightly smaller than that shown in Table 2 [11]. The Δ of
MCF becomes large as the fiber length increase. This property of the MCF has an effect on
the MFD distribution as shown in Fig. 11.

#222453 - $15.00 USD Received 4 Sep 2014; revised 5 Nov 2014; accepted 5 Nov 2014; published 25 Nov 2014
(C) 2014 OSA 1 December 2014 | Vol. 22, No. 24 | DOI:10.1364/OE.22.030137 | OPTICS EXPRESS 30146
 1

Relative-index difference Δ (%)

0.8 Ref #1

0.6
Ref #2 MCF
0.4

λ=1.55 μm
0.2
0 1 2 3 4 5 6 7 8
Fiber length (km)

Fig. 12. Relative-index difference Δ (%) distribution of fiber link.

Figure 13 shows the chromatic dispersion distribution D(z) of the fiber link at a
wavelength of 1.55 μm. The chromatic dispersion D can be estimated by using the estimated
relative-index difference Δ and the wavelength dependence of the MFD. We used the
experimental MFD results that we obtained at 1.31 and 1.55 μm. Here, a k value of 0.62 was
used.
20
Chromatic dispersion D (ps/km/nm)

Ref #2 MCF
16

12

4
Ref #1
0 λ=1.55 μm

-4
0 1 2 3 4 5 6 7 8
Fiber length (km)

Fig. 13. Chromatic dispersion distribution of fiber link at 1.55 μm.

The longitudinal parameters such as MFD 2w, relative-index difference Δ and chromatic
dispersion D can be successfully estimated with the present OTDR technique. This technique
will provide a powerful way of measuring the longitudinal fiber parameter of MCF.
4. Conclusions
We proposed a novel technique for obtaining the longitudinal fiber parameter of MCF based
on the bidirectional OTDR technique. By employing a simulation, we clarified that the MFD
distribution of the core in MCF with a mode coupling coefficient h of less than 0.02 (1/km)
can be estimated within an accuracy of 1% without taking the mode coupling between cores
into account. We clarified experimentally that the technique can be applied to a link
consisting of MCF. The MFD distribution of the fiber link was successfully estimated by
using the crosstalk measured with the modified OTDR. We also showed that the relative-
index difference Δ and chromatic dispersion distributions of the fiber link can be estimated
with the present technique.
This technique will constitute a powerful approach for measuring the longitudinal fiber
parameter of MCF.
Acknowledgments
The authors would like to thank Mr. K. Kawazu for his cooperation in experiments. Part of
this research uses results from a study entitled ‘Research and development of innovative
optical fiber for practical use’, commissioned by the National Institute of Information and
Communications Technology (NICT).

#222453 - $15.00 USD Received 4 Sep 2014; revised 5 Nov 2014; accepted 5 Nov 2014; published 25 Nov 2014
(C) 2014 OSA 1 December 2014 | Vol. 22, No. 24 | DOI:10.1364/OE.22.030137 | OPTICS EXPRESS 30147