Design and Comparison of Sf57 Over Sio2 On Same Structured PCF For Residual Dispersion Compensation
Design and Comparison of Sf57 Over Sio2 On Same Structured PCF For Residual Dispersion Compensation
Design and Comparison of Sf57 Over Sio2 On Same Structured PCF For Residual Dispersion Compensation
net/publication/311244005
Design and Comparison of SF57 over SiO2 on Same Structured PCF for
Residual Dispersion Compensation
CITATIONS READS
2 191
3 authors:
S M Abdur Razzak
Rajshahi University of Engineering & Technology
142 PUBLICATIONS 2,000 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Russel Reza Mahmud on 22 April 2018.
DOI: 10.1109/JPHOT.2016.2628802
1943-0655 © 2016 IEEE
IEEE Photonics Journal Design and Comparison of SF57
DOI:10.1109/JPHOT.2016.2628802
1943-0655 C 2016 IEEE. Translations and content mining are permitted for academic research only.
Manuscript received July 26, 2016; revised October 24, 2016; accepted November 8, 2016. Date of
publication November 17, 2016; date of current version December 12, 2016. Corresponding author:
Russel Reza Mahmud (e-mail: r.r.mahmud@gmail.com).
Abstract: Two square photonic crystal fibers (SPCFs), one made of silica (SiO2 ) and
the other made of high index lead silicate soft glass (SF57), are numerically analyzed,
compared, and proposed in this paper. Both designs of SPCFs show very high negative
and ultraflattened chromatic dispersion with very low variation of dispersion (D). P-1
offers −650(±6) ps/(nm.km) for the wavelength range of 1.46 → 1.63 μm and band-
width (BW) = 170 nm and P-2 offers −693(±4) ps/(nm.km) for the wavelength range of
1.45 → 1.68 μm (BW = 230 nm) with average high birefringence of the order of 10–2 . The
proposed designs can be suitable for residual dispersion compensation while maintaining
polarization due to having the properties of high negative flattened chromatic dispersion with
high birefringence. The influence of changing the structural parameters up to ±0.02μm are
also examined for ensuring the flatness of the property of chromatic dispersion within the
acceptable level.
Index Terms: Square photonic crystal fiber, polarization maintaining residual dispersion
compensation, negative flattened chromatic dispersion, high birefringence.
1. Introduction
Photonic crystal fibers (PCFs) possess several expedient characteristics for which they have de-
picted immense recognition in the field of optical fiber. A PCF for residual dispersion compen-
sation (R-DC) with maintaining polarization is one of the most crucial aspects for application [2],
[4]–[6] in optical data transmission system (O-DTS). Generally, a standard single mode optical
fiber (SSMOF) used in conventional O-DTS exhibits positive chromatic dispersion (C-D) of 12 to
22 ps/(nm.km) [2], [4]. This C-D reaches to a huge positive margin with a high peak large magni-
tude when an optical signal travels a very long distance in a fiber [2]. Since this positive dispersion
deteriorates the performance of the optical pulse signal, it should be diminished. A dispersion
compensation fiber (with very high sloping negative dispersion) needs to link with the conventional
SSMOF to minimize the huge positive dispersion. Unfortunately, after every dispersion compen-
sation, the residual dispersion always remains in a BW. Therefore, the system requires additional
compensation to nullify the extra residual dispersion that can be done by linking a residual disper-
sion compensation (R-DC) fiber. This R-DC fiber provides flattened negative chromatic dispersion
(FNC-D). A PCF should be tailored as much as the high negative and ultra-flattened chromatic dis-
persion property to reduce costs and all losses [6]. In O-DTS, for polarization maintaining purposes,
the property of high birefringence (B) in PCF is required. Therefore, a PCF with the dual charac-
teristics (FNC-D and B) will be a perfect applicant as a fiber for the use of polarization maintaining
residual dispersion compensation [4], [6].
Considering these theories, several researchers demonstrated PCFs on high FNC-D for ad-
dressing the issue of R-DC over the course of time [3]. For instance, some artificially created
elliptical cores in hybrid octagonal structures were reported in [2] and [4], which showed FNC-D
of −465.5(±5.25) and −562(±6) ps/(nm.km) for the wavelength (hereinafter represented as WL)
range of, respectively, 1.46 → 1.675 μm and 1.46 → 1.7 μm. Nevertheless, these two reported
designs [2], [4] are same in structure and are not purely octagonal. However, all the construction of
the PCFs make the models impractical to set up the air holes near the core region with specified
shapes and dimension for formation. Again, a hybrid decagonal [3] PCF that confirms FNC-D of
−558.96(±4.85) ps/(nm.km) for the WL range of 1.46 → 1.7 μm. This design is also complex to
form using conventional stack and draw method (CSDM) for being hybrid decagonal structure. A
conventional common pattern called hexagonal in the [5] exhibited FNC-D of −456(±6) ps/(nm.km)
within WL range of 1.37 → 1.7 μm. The prime disadvantage of [5] is a very small hexagonal PCF
is designed into another hexagonal PCF in a solo structure which is impractical to form employ-
ing CSDM. An uncommon pentagonal structure (almost impossible to form through CSDM) in ref.
no. 6 ensured FNC-D of −611.9(±6.1) ps/(nm.km) between the WL range of 1.46 → 1.625 μm.
Futhermore, equiangular decagonal spiral PCFs in [6] and [7] discovered FNC-D of −337.5(±7.5)
ps/(nm.km) and −453(±7) ps/(nm.km) within the WL range of, respectively, 1.35 → 1.65 μm and
1.15 → 1.75 μm. Owing to the pattern of spiral, the designs in [7] and [8] are complicated to form
through the CSDM. Besides, the structure of 7 is more difficult for placing an elliptical air hole in
the core region that makes the design as a slotted pattern. A conventional octagonal structure [9]
made of SiO2 reported FNC-D of −708 ps/(nm.km) with dispersion variation (D) of 20 ps/(nm.km)
for the WL range of 1.46 → 1.67 μm. The design has high negative dispersion. Neverthless, for
having D = 20 ps/(nm.km), it is not flattened well to work as an R-DC fiber properly. Very recently
a square PCF offers high negative slop dispersion of −2500 ps/(nm.km) at 1.55 μm WL which
can be used for dispersion compensation purposes [17]. In 2004, a square PCF offered chromatic
dispersion of 50 to −200 ps/(nm.km) for the WL range of 1.2 → 1.6 μm [18]. Moreover, dispersion
of 35 to −180 ps/(nm.km) for 1.2 → 2.0 μm WL was proposed by another square PCF in 2009
[13]. So far as we know, there is no PCF on square structure reported the property of FNC-D for
the application of residual dispersion compensation.
Hence, ultra-flattened very high negative dispersion of two new square (S) structure PCF de-
signs with high birefringence are proposed and compared in this monograph. The SiO2 made
SPCF (hereinafter represented as P-1) offers FNC-D of −644 to −656 ps/(nm.km) with D of
12 ps/(nm.km) for 1.46 → 1.63 μm (BW = 170 nm), whereas high index lead silicate soft glass
(SF57) made SPCF (also represented as P-2) performs −689 to −697 ps/(nm.km) with D of
8 ps/(nm.km) for 1.45 → 1.68 μm (BW = 230 nm) with average high B of the order 10−2 for that
offered bandwidth. The BW of P-1 involves S, C, L, and P-2 involves the S, C, L, and U bands
for the optical third window into the infrared region. In O-DTS, residual dispersion compensation
with maintaining polarization is the applicable field for any of the P-1 and P-2 SPCFs having their
properties.
Fig. 1. Cross sectional view of the P-1 SPCF (a) and P-2 SPCF (b) with symbols and different colors of
the diameters of air holes for each straight line, angle between air holes, pitch, and omitted air holes in
details.
air holes’ diameters have the same value in the longitudinal x-axis. The only dependent variable of
the proposed design is pitch that is denoted as for P-1 and 1 for P-2. This pitch is a distance
between any two consecutive air holes on any axis within the same or different square pattern
and it controls all the variables. The angle between two air holes is 90° and that is why any four
air holes can generate a square. We intentionally omit three air holes in the core location from
the first and second straight lines along the up and down position of the y-axis for obtaining high
birefringence which are reported circular dotted lines in Fig. 1. Four types (r1 , r2 , r3 , and r4 ) and
three types (r 1 , r 2 , and r 3 ) of circular air holes are used, respectively, for P-1 and P-2. In P-1, the
air holes’ diameters in the first, second, third, and fourth to seventh straight lines along the upper
and lower portion of the y-axis are denoted respectively as r1 , r2 , r3 and r4 . For P-2, r 1 , r 3 and r 2
depict the circular air holes’ diameters, respectively, in the first, third, and the rest of the straight
lines along the upper and lower portions of the vertical y-axis. For optimization, both the structures
of P-1 and P-2 have equal value of pitch that is equal to = 1 = 0.49 μm. The ODP of the
air holes in P-1 are r1 = 0.72 × , r2 = 0.78 × , r3 = 0.87 × and r4 = 0.75 × . Likewise,
in P-2, the ODP are r 1 = 0.72 × 1 , r 2 = 0.85 × 1 and r 3 = 0.95 × 1 . In addition, the P-1
and P-2 have, respectively, five and four variables, including pitch and diameters of air holes into
the whole structure. Therefore, in comparison between these two structures regarding parameters,
P-2 is more simple than P-1 for having less variables. In optimization procedure, we observe the
influence for changing one parameter when others are fixed, which are analyzed and reported in
the figures.
Fig. 2. Field pattern in optical form of the fundamental x-polarization mode (HEx 11 ) and y-polarization
mode (HEy 11 ) at the WL of 1.55 μm for the P-1 and P-2 SPCF designs. (a), (c) HEx 11 mode (b), (d)
HEy 11 mode.
of the SiO2 can be obtained by the Sellmeier equation (3), shown below, where the value of Sell-
meier constants are B 1 = 0.6961663, B 2 = 0.4079426, B 3 = 0.8974794, C 1 = 0.004679148,
C 2 = 0.01351206307 and C 3 = 97.93400254. Likewise, to define n of SF57, the sellmeier con-
stant are B 1 = 1.81651371, B 2 = 0.428893641, B 3 = 1.07186278, C 1 = 0.0143704198, C 2 =
0.0592801172, and C 3 = 121.419942. After getting n of SiO2 and SF57, the material dispersion
(D m ) can be calculated by (4), shown below. Total chromatic dispersion (D t ) can be evaluated by
(5), shown below. Birefringence is an important property which is defined as B, where, n x ef f and
n y ef f are real part of the mode indices of the two fundamental x-y polarization modes represent as
respectively HEx 11 and HEy 11 . L C can be obtained in dB/m by (7), shown below, where imaginary
part of n ef f needs to put into equation. Besides, the effective area (A ef f ) has the unit of μm2 is cal-
culated by (8), shown below, where electric field is need to put into Maxwell’s equation. Nonlinear
coefficient (γ) is obtained from (9), shown below, where the Kerr constant (n 2 ) for SiO2 is 2.1 ×
10−20 m2 W−1 [9] and for SF57 is 4.1 × 10−19 m2 W−1 [1]. The unit of γ is W−1 km−1 . The effective
V parameter (V ef f ) is calculated by the (10), shown below, where is the pitch, n ef f represents
effective refractive index, and n a represents refractive index of the air holes. The free space wave
number K 0 = (2π/λ). In that equation, air filling fraction is represented as F = A h ole /A cell . Here,
A h ole and A cell are the area of holes and PCF, respectively, [10].
∇ × (μ−1
r ∇ × E ) − k 0 [εr ]E = 0 (1)
D w = − (λ/C ) d 2 Re[n ef f ] /d λ2 (2)
B 1 λ2 B 2 λ2 B 3 λ2
n = 1+ 2 + 2 + 2 (3)
λ − C1 λ − C2 λ − C3
2
D m = − (λ/C ) d n /d λ2 (4)
Dt = Dw +Dm (5)
B = Re(n efX f ) − Re(n efY f ) (6)
L C = 8.686 × k O Im[n ef f ] (7)
∞ ∞
2
∞ ∞
A ef f (λ) = E (x , y )2 dxdy E (x , y )4 dxdy (8)
−∞ −∞ −∞ −∞
γ = (2π/λ)(n 2 /A ef f ) (9)
V ef f = (k 0 F 1/2 )(n ef f 2 − n a 2 )1/2 . (10)
The optical field distributions for the WL at 1.55 μm of the fundamental x- polarized mode (HEx 11 )
and y-polarization mode (HEy 11 ) are given in the Fig. 2. It can be noticed from the simulation in
Fig. 2 that at the central core region the light is strongly confined since the cladding part has lower
refractive index than the core. Since SF57 has higher n than SiO2 , the strong mode confinement
Fig. 3. D t with respect to WL. (a) P-1 covers S + C + L and P-2 covers S + C + L + U bands with
ODP. (b) shows flatness at negative level of P-1 and P-2 SPCF with some recent results of the reported
designs in references.
Fig. 4. (a) L C with respect to WL for HEx 11 and HEy 11 modes with ODP for P-1 (with emblems) and P-2
(without emblems). (b) B with respect to WL with ODP (without emblems solid line for P-2 and dashed
line for P-1 SPCF).
is more easily possible in the core of P-2 than P-1. Artificially created asymmetrical core pattern
offers higher n x ef f for HEx 11 than n y ef f for HEy 11 . HEy 11 mode sits in a large core area (see Fig. 2(b)
and (d)), and hence, it is well confined while the power of HEx 11 mode leaks into the adjacent air
holes (see Fig. 2(a) and (c)). That is why L C of HEy 11 is higher than HEx 11 , which is shown in Fig.
4(a). From the simulation, it has been found that total power have passed into the solid material
and cladding air holes for HEx 11 is about 73.4% and 26.6% and for HEy 11 is about 68% and 32%.
Since power fraction into cladding air holes for HEx 11 is comparatively lower than HEy 11 , therefore,
HEx 11 exhibits higher n ef f than HEy 11 . It is also observed from the results of simulation that the n ef f
of the HEx 11 and HEy 11 are decreased with increasing of the WL.
A true R-DC fiber should have high FNC-D as well as very shot in length for minimizing costs
and losses. Fig. 3(a) shows very high and ultra FNC-D for both P-1 and P-2 are, respectively,
−644 to −656 and −689 to −697 ps/(nm.km) with absolute D of 12 and 8 ps/(nm.km) within the
WL range of 1.46 → 1.63 μm and 1.45 → 1.68 μm for HEy 11 . At best, from our literature review,
chromatic dispersion is considered to be well flattened if D is within the limit of 12 ps/(nm.km)
[2]–[5], [8]. Fig. 3(b) views comparison at a glance between our proposed P-1 and P-2 and different
structures reported by researchers [2]–[8] regarding D t with the function of the WL. Therefore, we
do not consider the designs of the [9] for a comparison at a glance in Fig. 3(b) which has D
Fig. 5. D t effects with respect to WL for alteration of only pitch ( for P-1 and 1 for P-2) when
other parameters are stayed fixed. (Without emblem ODP is depicted by solid and dashed lines for,
respectively, P-2 and P-1. In addition, with emblem increment and decrement of pitch for P-2 and P-1
are indicated by solid and dashed lines).
equal to or greater than 20 ps/(nm.km) within a band [9] since is not perfect enough for RDC. The
proposed P-1 and P-2 provide dispersion variation of 12 and 8 ps/(nm.km), respectively, involving
S + C + L and S + C + L + U bands. It is evident from Fig. 3(b) that both of our P-1 and P-2
designs confirm ultra-flattened chromatic dispersion with highest amplitude of all reported designs
in references 2 to 8. Besides, the P-2 SPCF shows ultra-flattened characteristics of D t than the
references 3 to 8. A good R-DC fiber should have low L C for a WL boundary. Fig. 4(a) confirms that
L C with respect to the WL for HEx 11 and HEy 11 of both P-1 and P-2 is very low compared to the [2]
to [8]. For P-1, L C provides the value of 10−3 to 10−2.5 and 10−3.25 to 10−2.75 dB/m, respectively,
for the HEy 11 and HEx 11 . In the case of P-2, a much lower L C of 10−3.6 to 10−2.75 and 10−4.3 to
10−3.4 dB/m respectively for HEy 11 and HEx 11 . It is seen from Fig. 4(b) that B of 0.015 to 0.01 and
0.0142 to 0.0115 of HEy 11 can be obtained for the WL range of, respectively, 1.46 → 1.63 μm and
1.45 → 1.68 μm for P-1 and P-2 with ODP. This high birefringence can maintain polarization.
For proposing a new PCF design, we need to evaluate the sensitivity of the PCF properties
(mainly chromatic dispersion ensuring its flatness) for the alteration of ±(0.01 to 0.02) μm in air
holes’ diameters and pitch which may vary during formation process from the ODP [2]–[9]. The
responses of pitch variation are shown in Fig. 5, where D t changes slightly. Dispersion varies ±20
to ±10 ps/(nm.km) at the edge of lower to higher WL for both P-1 and P-2 because of changing pitch
by ±0.01 μm when other parameters are kept unchanged. Again, we found alteration of D t of ±50
to ±20 ps/(nm.km) at the edge of 1.46 → 1.63 μm for P-1 and ±45 to ±25 ps/(nm.km) at the edge
of 1.45 → 1.68 μm for P-2 for the variation of pitch at ±0.02 μm. Dispersion varies more at the edge
of 1.46 μm than 1.63 μm for P-1 and 1.45 than 1.68 μm for P-2. Fig. 6 explores that for both P-1 and
P-2, D t varies by only ±10 to ±20 ps/(nm.km) for variation of r1 and r 1 of ±(0.01 to 0.02) μm in the
first straight line within their specified WL range. Fig. 7 shows that in the case of P-1, D t changes
±30 to ±60 ps/(nm.km) for alteration of r2 only by ±(0.01 to 0.02) μm within the BW, but variation
of r2 degrades the flatness of D t . On the other side, in the case of P-2, D t changes only ±5 to
±10 ps/(nm.km) for changing of r 2 by ±(0.01 to 0.02) μm. For both P-1 and P-2, due to the variation
of r3 and r 3 , variation of D t changes very little amount and that is shown in the Fig. 8. In case of
P-1 only, the alteration of r4 when all others ODP are kept fixed. D t varies ±30 to ±60 ps/(nm.km)
from the ODP within the specified BW of WL. Finally, it is remarkable that for the variation of both
air holes’ diameters and pitch, both the P-1 and P-2 maintain the characteristics of the FNC-D and
that satisfies the formation tolerance. Compared to P-1, P-2 performs the dispersion flatness more
Fig. 6. D t effects with respect to WL for alteration of only r1 for P-1 and r 1 for P-2. (Without emblem
ODP is explored by solid and dashed line for P-2 and P-1. Furthermore, with emblem increment and
decrement of r 1 and r1 for P-2 and P-1 are displayed by other solid and dashed lines).
Fig. 7. D t effects with respect to WL for alteration of only r2 for P-1 and r 2 for P-2. (Without emblem
ODP for P-2 and P-1 is explored by solid and dashed lines, and with emblem increment and decrement
of r 2 and r2 for P-2 and P-1 are shown by solid and dashed lines).
accurately and confirms higher amplitude of negative flattened dispersion with small dispersion
variation. Therefore, P-2 performs better than P-1 considering D t .
From the data of the results of simulation, the effective area (A ef f ) is found 1.11 to 1.36 μm2
for P-1 and 0.95 to 1.22 μm2 for P-2, respectively, within the WL range of 1.46 → 1.63 μm and
1.45 → 1.68 μm with ODP. When the WL is increased, the A ef f is also increased. Since, the A ef f
of the fiber is very small and the HEx 11 and HEy 11 are tightly confined, the nonlinear phenomenon
such as stimulated brillouin scattering and four wave mixing may be raised noises into the fiber
material when high intensity of light propagates. This may limit such fiber for using high power
pulse delivery application. The RD-C fiber generally used in O-DTS to compensate the residual
dispersion, where high FNC-D leads small length of a fiber to compensate the residual dispersion.
The proposed fiber provides very high FNC-D and high B which can work perfectly in O-DTS using
a short fiber length [16]. Besides, this high FNC-D can suppress the nonlinear parametric process
of a fiber [4]. The splice loss (L s ) is generally high since for being smaller A ef f for the proposed
Fig. 8. D t effects with respect to WL for alteration of only r3 for P-1 and r 3 for P-2 when other parameters
are kept constant. (Here, without emblem ODP for P-2 and P-1 are reported by solid and dashed lines.
Nevertheless, with emblem decrement and increment of r 3 and r3 are indicated by solid lines for P-2
and dashed lines for P-1).
Fig. 9. γ and V eff parameter with respect to the WL for P-1 (a) and P-2 (b) of the SPCF designs.
P-1 and P-2. Recently, special techniques have been proposed by many researchers to minimize
L s in the coupling between conventional SSMOF and PCF. Fusion splicing [12] (invented by Xiao
et al.) is a technique that can reduce L s . A new splice-free coupling technique is proposed by L.
Saval et al. [11] which can reduce L s for coupling purpose between conventional SSMOF and any
structure of PCF.
For the analysis of mode test, the investigation is studied carefully that both the proposed P-1
y
and P-2 report a stronger fundamental mode (HE11 ) than the second mode (3968 and 9707 times
stronger for P-1 and P-2, respectively) at 1.55 μm WL. In addition, the L C of the second mode is
y
higher than (HE11 ) and that is more than 46 dB/m [2], [4], [8], [10]. Therefore, both the proposed
P-1 and P-2 SPCF will successfully operate as a single-mode fiber [5]–[7], [9]. Moreover, for the
single mode analysis by applying another method, the effective V parameter (V ef f ) is calculated
by equation (10) [10]. Fig. 9 reports that both the P-1 and P-2 can be performed as a single
mode fiber for the WL range of 1.46 → 1.63 μm and 1.45 → 1.68 μm since V ef f is less than the
numerical value of 1.5 and 2.4 or π within that bandwidth. Fig. 9 also exhibits γ of 70 and 1599.8
W−1 km−1 for P-1 and P-2 at 1.55 μm WL. The property of high γ is necessary for supercontinuum
generation, optical parametric amplification, and in nonlinear optics application. Now considering
the formation issue, both the proposed P-1 and P-2 square SPCF designs with circular air holes can
TABLE 1
Comparison the Properties at 1.55 μm WL Between the Proposed (P-1 and P-2) With Other
PCF Designs
[2] 1.46 → 1.675 −458.5 13 10−2 – Elliptical core in hybrid octagonal structure
[3] 1.46 → 1.7 −558.5 9.7 10−2 – Elliptical-circular hybrid decagonal
[4] 1.46 → 1.7 −562 12 10−2 50 Elliptical core in hybrid octagonal structure
[5] 1.37 → 1.7 −456 12 10−7 12 PCF in PCF in hexagonal
[6] 1.46 → 1.625 −611.6 12.2 10−2 45 Pentagonal
[7] 1.35 → 1.65 −337.5 15 10−3 – Elliptical-circular slotted hybrid decagonal spiral
[8] 1.15 → 1.75 −453 14 10−2 – Decagonal spiral
P-1 1.46 → 1.63 -650 12 10−2 70 SiO2 made square
P-2 1.45 → 1.68 -693 8 10−2 1599.8 SF57 made square
be formed by using the stack and draw procedure, which is explained in detail in [13]. In addition,
a new ultrasonic drilling machine can be used to form any complex holey structure [14]. Besides,
the sol-gel technique can be applied [15], providing flexibility to form any PCF of various irregular
structures. Using this method [15], any complex structure has been already formed. Table 1 shows
the comparison between the proposed (P-1 and P-2) and other PCFs’ design with data, results,
and analysis.
4. Conclusion
Two SPCFs (P-1 and P-2) have been proposed, and they offer very high negative flattened dis-
persion of −650 (D = 12) ps/(nm.km) within the wavelength range of 1.46 → 1.63 μm (BW =
170 nm) for P-1 and −693 (D = 8) ps/(nm.km) within 1.45 → 1.68 μm (BW = 230 nm) wave-
length range for P-2 with high birefringence of the order 10−2 . In addition, it demonstrates very low
confinement loss of the order of 10−3 to 10−2.5 dB/m for P-1 and 10−3.6 to 10−2.75 dB/m for P-2
within the bandwidth of interest. Among these two proposed designs with the same structure, P-2
(SF57 made SPCF) confirms better results than P-1 (SiO2 made SPCF) in all respects. Owing to
the excellent properties, both the proposed (P-1 and P-2) SPCFs will be promising applicants for
residual dispersion compensation with polarization maintaining application in optical systems.
References
[1] A. Agrawal, N. Kejalakshmy, B. M. A. Rahman, and K. T. V. Grattan, “Soft glass equiangular spiral photonic crystal
fiber for super continuum generation,” IEEE Photon. Technol. Lett., vol. 21, no. 22, pp. 1722–1724, Nov. 2009.
[2] M. Samiul Habib, R. Ahmad, M. Selim Habib, and M. I. Hasan, “Residual dispersion compensation over S + C + L + U
wavelength bands using highly birefringent octagonal photonic crystal fiber,” Appl. Opt., vol. 53, no. 14, pp. 3057–3062,
May 2014.
[3] M. S. Habib and E. Khandker, “Highly birefringent photonic crystal fiber with ultra-flattened negative dispersion over S
+ C + L + U bands,” Appl. Opt., vol. 54, no. 10, pp. 2786–2789, 2015.
[4] M. I. Hasan, S. M. A. Razzak, and M. S. Habib, “Design and characterization of highly birefringent residual dispersion
compensating photonic crystal fiber,” J. Lightw. Technol., vol. 32, no. 23, pp. 3976–3982, Dec. 2014.
[5] D. C. Tee, M. H. A. Bakar, N. Tamchek, and F. R. M. Adikan, “Photonic crystal fiber in photonic crystal fiber for residual
dispersion compensation over E + S + C + L + U wavelength bands,” IEEE Photon. J., vol. 5, no. 3, Jun. 2013, Art.
no. 7200607.
[6] X. Li, P. Liu, Z. Xu, and Z. Zhang, “Design of a pentagonal photonic crystal fiber with high birefringence and large
flattened negative dispersion,” Appl. Opt., vol. 54, no. 24, pp. 7350–7357, Aug. 2015.
[7] M. A. Islam and M. S. Alam., “Design optimization of equiangular spiral photonic crystal fiber for large negative flat
dispersion and high birefringence,” J. Lightw. Technol., vol. 30, no. 22, pp. 3545–3551, Nov. 2012.
[8] R. R. Mahmud, S. M. A. Razzak, M. I. Hasan, and G. K. M. Hasanuzzaman, “Ultraflattened high negative chromatic
dispersion over O + E + S + C + L + U bands of a microstructured optical fiber,” Opt. Eng., vol. 54, no. 9, pp.
0971051–0971057, Sep. 2015.
[9] R. R. Mahmud, M. A. G. Khan, and S. M. A. Razzak, “Management of residual dispersion of an optical transmission
system using octagonal photonic crystal fiber,” Opt. Eng., vol. 55, no. 4, pp. 0471071–0471077, Apr. 2016.
[10] S. M. A. Razzak and Y. Namihira, “Proposal for highly nonlinear dispersion-flattened octagonal photonic crystal fibers,”
IEEE Photon. Technol. Lett., vol. 20, no. 4, pp. 249–251, Feb. 2008.
[11] S. G. L. Saval et al., “Splice-free interfacing of photonic crystal fibers,” Opt. Lett., vol. 30, no. 13, pp. 629–1631,
Jul. 2005.
[12] L. Xiao, W. Jin, and M. S. Demokan, “Fusion splicing small core photonic crystal fibers and single-mode fibers by
repeated arc discharges,” Opt. Lett., vol. 32, no. 2, pp. 115–117, Jan. 2007.
[13] S. Kim, C. S. Kee, and C. G. Lee, “Modified rectangular lattice photonic crystal fibers with high birefringence and
negative dispersion,” Opt. Exp., vol. 17, no. 10, pp. 795–7957, May 2009.
[14] X. Feng, A. K. Mairaj, D. W. Hewak, and T. M. Monro, “Nonsilica glasses for holey fibers,” J. Lightw. Technol., vol. 23,
no. 6, pp. 2046–2054, Jun. 2005.
[15] R. T. Bise and D. J. Trevor, “Sol-gel derived microstructured fibers: Fabrication and characterization,” in Proc. Int. Conf.
Opt. Fiber Commun., Washington, DC, USA, Mar. 2005, Art. no. OWL6.
[16] L. Zhang, T. Luo, Y. Yue, and A. E. Willner, “High group birefringence in photonic crystal fibers with both positive and
negative phase birefringence,” J. Opt. A Pure Appl. Opt., vol. 10, no. 3, Feb. 2008, Art. no. 035004.
[17] P. S. Maji and P. R. Chaudhuri, “Designing broad band dispersion compensation with square lattice PCF and applications
to ASE suppression with ultra negative dispersion,” Optik, vol. 127, no. 5, pp. 2603–2607, Mar. 2016.
[18] A. H. Bouk, A. Cucinotta, F. Poli, and S. Selleri, “Dispersion properties of square-lattice photonic crystal fibers,” Opt.
Exp., vol. 12, no. 5, pp. 941–946, 2004.