Broadband high birefringence and polarizing hollow core antiresonant fibers

Abstract

We systematically study different approaches to introduce high birefringence and high polarization extinction ratio in hollow core antiresonant fibers. Having shown the ineffectiveness of elliptical cores to induce large birefringence in hollow core fibers, we focus on designing and optimizing polarization maintaining Hollow Core Nested Antiresonant Nodeless Fibers (HC-NANF). In a first approach, we create and exploit anti-crossings with glass modes at different wavelengths for the two polarizations. We show that suitable low loss high birefringence regions can be obtained by appropriately modifying the thickness of tubes along one direction while leaving the tubes in the orthogonal direction unchanged and in antiresonance. Using this concept, we propose a new birefringent NANF design providing low loss (~40dB/km) and high birefringence (>10⁻⁴) over a record bandwidth of ~550nm, and discuss how bandwidth can be traded off to further reduce the loss to a few dB/km. Finally, we propose a polarization mode-stripping technique in the birefringent NANF. As a demonstration, we propose a polarizing birefringent NANF design that can achieve orthogonal polarization loss ratios as large as 30dB over the C-band while eliminating any undesirable polarization coupling effect thereby resulting in a single polarization output in a hollow core fiber regardless of the input polarization state.

Full text

Broadband high birefringence and polarizing
hollow core antiresonant fibers
SEYEDMOHAMMAD ABOKHAMIS MOUSAVI,* SEYED REZA SANDOGHCHI,
DAVID J. RICHARDSON, AND FRANCESCO POLETTI
Optoelectronics Research Centre University of Southampton, SO17 1BJ, UK
*sam1e12@soton.ac.uk
Abstract: We systematically study different approaches to introduce high birefringence and
high polarization extinction ratio in hollow core antiresonant fibers. Having shown the
ineffectiveness of elliptical cores to induce large birefringence in hollow core fibers, we focus
on designing and optimizing polarization maintaining Hollow Core Nested Antiresonant
Nodeless Fibers (HC-NANF). In a first approach, we create and exploit anti-crossings with
glass modes at different wavelengths for the two polarizations. We show that suitable low
loss high birefringence regions can be obtained by appropriately modifying the thickness of
tubes along one direction while leaving the tubes in the orthogonal direction unchanged and
in antiresonance. Using this concept, we propose a new birefringent NANF design providing
low loss (~40dB/km) and high birefringence (>10−4) over a record bandwidth of ~550nm, and
discuss how bandwidth can be traded off to further reduce the loss to a few dB/km. Finally,
we propose a polarization mode-stripping technique in the birefringent NANF. As a
demonstration, we propose a polarizing birefringent NANF design that can achieve
orthogonal polarization loss ratios as large as 30dB over the C-band while eliminating any
undesirable polarization coupling effect thereby resulting in a single polarization output in a
hollow core fiber regardless of the input polarization state.
©2016 Optical Society of America
OCIS codes: (060.2310) Fiber optics; (060.2400) Fiber properties; (060.4005) Microstructured fibers; (060.2420)
Fibers, polarization-maintaining.
References and links
1. M. A. Terrel, M. J. F. Digonnet, and S. Fan, “Resonant fiber optic gyroscope using an air-core fiber,” J.
Lightwave Technol. 30(7), 931–937 (2012).
2. G. Epple, K. S. Kleinbach, T. G. Euser, N. Y. Joly, T. Pfau, P. S. J. Russell, and R. Löw, “Rydberg atoms in
hollow-core photonic crystal fibres,” Nat. Commun. 5, 4132 (2014).
3. A. D. Slepkov, A. R. Bhagwat, V. Venkataraman, P. Londero, and A. L. Gaeta, “Spectroscopy of Rb atoms in
hollow-core fibers,” Phys. Rev. A 81(5), 053825 (2010).
4. X. Peng and L. Dong, “Fundamental-mode operation in polarization-maintaining ytterbium-doped fiber with an
effective area of 1400 µm2,” Opt. Lett. 32(4), 358–360 (2007).
5. T. Hosaka, K. Okamoto, T. Miya, Y. Sasaki, and T. Edahiro, “Low-loss single polarisation fibres with
asymmetrical strain birefringence,” Electron. Lett. 17(15), 530–531 (1981).
6. R. D. Birch, D. N. Payne, and M. P. Varnham, “Fabrication of polarisation-maintaining fibres using gas-phase
etching,” Electron. Lett. 18(24), 1036–1038 (1982).
7. R. B. Dyott, J. R. Cozens, and D. G. Morris, “Preservation of polarisation in optical-fibre waveguides with
elliptical cores,” Electron. Lett. 15(13), 380–382 (1979).
8. F. Couny, F. Benabid, P. J. Roberts, P. S. Light, and M. G. Raymer, “Generation and photonic guidance of
multi-octave optical-frequency combs,” Science 318(5853), 1118–1121 (2007).
9. P. S. J. Russell, P. Hölzer, W. Chang, A. Abdolvand, and J. C. Travers, “Hollow-core photonic crystal fibres for
gas-based nonlinear optics,” Nat. Photonics 8(4), 278–286 (2014).
10. F. Benabid and P. J. Roberts, “Linear and nonlinear optical properties of hollow core photonic crystal fiber,” J.
Mod. Opt. 58(2), 87–124 (2011).
11. O. H. Heckl, C. R. E. Baer, C. Kränkel, S. V. Marchese, F. Schapper, M. Holler, T. Südmeyer, J. S. Robinson, J.
W. G. Tisch, F. Couny, P. Light, F. Benabid, and U. Keller, “High harmonic generation in a gas-filled hollow-
core photonic crystal fiber,” Appl. Phys. B 97(2), 369–373 (2009).
12. J. C. Knight, J. Broeng, T. A. Birks, and P. S. J. Russell, “Photonic band gap guidance in optical fibers,” Science
282(5393), 1476–1478 (1998).
13. P. J. Roberts, F. Couny, H. Sabert, B. J. Mangan, D. P. Williams, L. Farr, M. W. Mason, A. Tomlinson, T. A.
Birks, J. C. Knight, and P. St. J. Russell, “Ultimate low loss of hollow-core photonic crystal fibres,” Opt.
Express 13(1), 236–244 (2005).
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22943
#269041
http://dx.doi.org/10.1364/OE.24.022943
Journal © 2016
Received 23 Jun 2016; revised 15 Sep 2016; accepted 21 Sep 2016; published 26 Sep 2016

14. M. A. Duguay, Y. Kokubun, T. L. Koch, and L. Pfeiffer, “Antiresonant reflecting optical waveguides in SiO2-Si
multilayer structures,” Appl. Phys. Lett. 49(1), 13 (1986).
15. N. M. Litchinitser, A. K. Abeeluck, C. Headley, and B. J. Eggleton, “Antiresonant reflecting photonic crystal
optical waveguides,” Opt. Lett. 27(18), 1592–1594 (2002).
16. Y. Y. Wang, N. V. Wheeler, F. Couny, P. J. Roberts, and F. Benabid, “Low loss broadband transmission in
hypocycloid-core Kagome hollow-core photonic crystal fiber,” Opt. Lett. 36(5), 669–671 (2011).
17. Y. Wang, F. Couny, P. J. Roberts, and F. Benabid, “Low loss broadband transmission in optimized core-shape
kagome hollow-core PCF,” in Conference on Lasers and Electro-Optics 2010, OSA Technical Digest (CD)
(Optical Society of America, 2010), paper CPDB4.
18. F. Poletti, M. N. Petrovich, and D. J. Richardson, “Hollow-core photonic bandgap fibers: technology and
applications,” Nanophotonics 2(5-6), 315 (2013).
19. F. Couny, F. Benabid, and P. S. Light, “Large-pitch kagome-structured hollow-core photonic crystal fiber,” Opt.
Lett. 31(24), 3574–3576 (2006).
20. J. R. Hayes, F. Poletti, M. S. Abokhamis, N. V. Wheeler, N. K. Baddela, and D. J. Richardson, “Anti-resonant
hexagram hollow core fibers,” Opt. Express 23(2), 1289–1299 (2015).
21. A. Hartung, J. Kobelke, A. Schwuchow, K. Wondraczek, J. Bierlich, J. Popp, T. Frosch, and M. A. Schmidt,
“Double antiresonant hollow core fiber - guidance in the deep ultraviolet by modified tunneling leaky modes,”
Opt. Express 22(16), 19131–19140 (2014).
22. A. D. Pryamikov, A. S. Biriukov, A. F. Kosolapov, V. G. Plotnichenko, S. L. Semjonov, and E. M. Dianov,
“Demonstration of a waveguide regime for a silica hollow - core microstructured optical fiber with a negative
curvature of the core boundary in the spectral region > 3.5 μm,” Opt. Express 19(2), 1441–1448 (2011).
23. A. N. Kolyadin, A. F. Kosolapov, A. D. Pryamikov, A. S. Biriukov, V. G. Plotnichenko, and E. M. Dianov,
“Light transmission in negative curvature hollow core fiber in extremely high material loss region,” Opt. Express
21(8), 9514–9519 (2013).
24. W. Belardi and J. C. Knight, “Hollow antiresonant fibers with reduced attenuation,” Opt. Lett. 39(7), 1853–1856
(2014).
25. W. Belardi, “Design and properties of hollow antiresonant fibers for the visible and near infrared spectral range,”
J. Lightwave Technol. 33(21), 4497–4503 (2015).
26. A. F. Kosolapov, G. K. Alagashev, A. N. Kolyadin, A. D. Pryamikov, A. S. Biryukov, I. A. Bufetov, and E. M.
Dianov, “Hollow-core revolver fibre with a double-capillary reflective cladding,” Quantum Electron. 46(3),
267–270 (2016).
27. W. Belardi and J. C. Knight, “Hollow antiresonant fibers with low bending loss,” Opt. Express 22(8), 10091–
10096 (2014).
28. B. Debord, M. Alharbi, L. Vincetti, A. Husakou, C. Fourcade-Dutin, C. Hoenninger, E. Mottay, F. Gérôme, and
F. Benabid, “Multi-meter fiber-delivery and pulse self-compression of milli-Joule femtosecond laser and fiber-
aided laser-micromachining,” Opt. Express 22(9), 10735–10746 (2014).
29. F. Emaury, C. J. Saraceno, B. Debord, D. Ghosh, A. Diebold, F. Gèrôme, T. Südmeyer, F. Benabid, and U.
Keller, “Efficient spectral broadening in the 100-W average power regime using gas-filled kagome HC-PCF and
pulse compression,” Opt. Lett. 39(24), 6843–6846 (2014).
30. F. Poletti, “Nested antiresonant nodeless hollow core fiber,” Opt. Express 22(20), 23807–23828 (2014).
31. J. M. Fini, J. W. Nicholson, B. Mangan, L. Meng, R. S. Windeler, E. M. Monberg, A. DeSantolo, F. V.
DiMarcello, and K. Mukasa, “Polarization maintaining single-mode low-loss hollow-core fibres,” Nat. Commun.
5, 5085 (2014).
32. S. A. Mousavi, D. J. Richardson, S. R. Sandoghchi, and F. Poletti, “First design of high birefringence and
polarising hollow core anti-resonant fibre,” in 2015 European Conference on Optical Communication (ECOC)
(2015), pp. 1–3.
33. W. Ding and Y.-Y. Wang, “Hybrid transmission bands and large birefringence in hollow-core anti-resonant
fibers,” Opt. Express 23(16), 21165–21174 (2015).
34. P. J. Roberts, “Birefringent hollow core fibers,” Proc. SPIE 6782, 67821R (2007).
35. F. Poletti, N. G. R. Broderick, D. Richardson, and T. Monro, “The effect of core asymmetries on the polarization
properties of hollow core photonic bandgap fibers,” Opt. Express 13(22), 9115–9124 (2005).
36. X.-L. Tang, B.-S. Sun, and Y.-W. Shi, “Design and optimization of low-loss high-birefringence hollow fiber at
terahertz frequency,” Opt. Express 19(25), 24967–24979 (2011).
37. P. J. Roberts, D. P. Williams, H. Sabert, B. J. Mangan, D. M. Bird, T. A. Birks, J. C. Knight, and P. S. J. Russell,
“Design of low-loss and highly birefringent hollow-core photonic crystal fiber,” Opt. Express 14(16), 7329–
7341 (2006).
38. J. Hayes, S. Sandoghchi, T. Bradley, Z. Liu, R. Slavík, M. A. Gouveia, N. Wheeler, G. Jasion, Y. Chen, E.
Numkam-Fokoua, M. Petrovich, D. Richardson, and F. Poletti, “Antiresonant hollow core fiber with octave
spanning bandwidth for short haul data communications,” in Optical Fiber Communication Conference
Postdeadline Papers (Optical Society of America, 2016), p. Th5A.3.
39. C. Wei, R. A. Kuis, F. Chenard, C. R. Menyuk, and J. Hu, “Higher-order mode suppression in chalcogenide
negative curvature fibers,” Opt. Express 23(12), 15824–15832 (2015).
40. M. Michieletto, J. K. Lyngsø, C. Jakobsen, J. Lægsgaard, O. Bang, and T. T. Alkeskjold, “Hollow-core fibers for
high power pulse delivery,” Opt. Express 24(7), 7103–7119 (2016).
41. L. Vincetti and V. Setti, “Waveguiding mechanism in tube lattice fibers,” Opt. Express 18(22), 23133–23146
(2010).
42. E. A. J. Marcatili and R. A. Schmeltzer, “Hollow metallic and dielectric waveguides for long distance optical
transmission and lasers,” Bell Syst. Tech. J. 43(4), 1783–1809 (1964).
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22944

1. Introduction
Besides loss, bandwidth and modal properties, the ability to control and maintain the
polarization state (e.g. by inducing a large birefringence or a differential loss between the two
orthogonally polarized modes) is a key feature for a wide range of optical fiber applications,
including precise interferometric sensing (e.g. fiber based gyroscopes [1]), quantum
computing [2], atom spectroscopy [3], and polarization maintaining optical amplifiers [4].
Although conventional polarization-maintaining (PM) fibers have been around for many
years, in which a high birefringence (Hi-Bi) is introduced by applying either stress (e.g.
PANDA or Bow-tie fibers [5, 6]) or form anisotropy (e.g. elliptic core [7]), they face a few
fundamental limitations associated with their solid core. For example, material absorption,
low damage thresholds and optical nonlinearity fundamentally restrict their application to
wavebands where the glass is transparent and to optical powers low enough not to cause
undesired nonlinear effects and/or dielectric damage. Some of these limitations can be
addressed by using hollow core optical fibers, which in addition can also enable interesting
new linear and nonlinear applications exploiting long path lengths for gas-light interactions
[8–11]. Despite many potential advantages, hollow core fibers are a less mature technology
than solid core versions, which still require physical understanding and new optical designs in
order to achieve their full potential. In general, there are two major hollow core fiber
categories in terms of transverse modal confinement mechanism: photonic bandgap fibers
(HC-PBGFs), which rely on the creation of a photonic band gap in the cladding [12, 13] and
anti-resonance fibers (HC-ARFs) where guidance occurs as a result of antiresonant effects in
glass membranes surrounding the fiber core [14, 15], and is strengthened by inhibited
coupling between core (air) and cladding (glass) modes [16, 17].
So far, HC-PBGFs have demonstrated lower loss than HC-ARFs [13, 18]. However, their
narrow bandwidth (i.e. 10-30% of the central operating wavelength) and larger field overlap
with the glass in comparison to HC-ARFs [9] make them less suitable for broadband
applications (e.g. broadband nonlinear processes) or for high power delivery. During the last
decade considerable research has focused on lowering the loss in HC-ARFs and many
different structures have been proposed based on: a Kagomé cladding [16, 19], a hexagram
structure [20], a double anti-resonance layer [21], tubular “negative curvature” structure [22,
23] and nested tubular structure [24–26]. These have resulted in significant improvement of
properties such as low bend-loss [27], ultra-large bandwidth [8] and high damage-threshold
[28, 29]. In addition to the above, the recent nested antiresonant nodeless hollow core fiber
(NANF) design, which incorporates nested glass rings, has the potential to achieve loss levels
as low as that of solid core fibers [30].
Generating a birefringence of the same order of magnitude as in commercial solid core
PM fibers (~10−4 for 1550nm operation) in a hollow core fiber is however difficult.
Conventional methods of using stress rods (and, as we shall see, elliptical shapes) are not
applicable or effective. Besides, structural requirements to minimize confinement loss in
hollow core fibers add additional limitations. Arguably, the best result so far was achieved in
a HC-PBGF with asymmetric rods in the core surround, presenting low loss (10dB/km at
1.530µm) and birefringence in the range of that of solid-core fibers, albeit over a narrow
operating bandwidth (<10nm) [31]. Achieving such a high birefringence in HC-ARFs is even
harder since the field overlaps with glass and the resulting sensitivity of the birefringence to
the core’s geometry is even lower. Only recently, two theoretical Hi-Bi HC-ARFs based on
modified NANF concepts were proposed which could address the problem by operating in the
normal [32] or hybrid band transmission regimes [33]. Both show promising routes to achieve
high birefringence (~10−4) in ARFs. However, the underpinning differences between the two
approaches require detailed study, and further reduction of the loss and widening the Hi-Bi
bandwidth are still possible and as targeted here.
In this paper, we report a detailed study of the birefringence of NANF designs that have
tubes with different membrane thicknesses along orthogonal axes (bi-thickness NANFs) and
propose a design procedure to maximize the bandwidth and minimize loss while maintaining
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22945

high birefringence operation. In the following Section, we demonstrate that unlike solid
fibers, an elliptical core shape cannot provide a large enough birefringence in a hollow core
fiber. In Section 3, we propose a way to circumvent this using a bi-thickness NANF with a
small core and 4 triple nested tubes. We also introduce an approach for analyzing the effect of
the design parameters of the fiber on its loss, bandwidth and regime of operation, using this
we propose a new HC-ARF design providing the lowest loss and the widest bandwidth high
birefringence operation ever reported, to the best of our knowledge. In Section 4, we propose
for the first time, the out-stripping of one selected fundamental polarization mode to provide
the hollow core fiber with an additional polarizing ability, in addition to a polarization
maintaining behavior. We finally summarize our work in Section 5.
2. Birefringence by asymmetry
While the PM property of HC-PBGFs has already been investigated theoretically in various
works [34, 35] and promising results have been achieved experimentally [31], extending these
results to anti-resonance fiber types is not straightforward, and only recently the first designs
have been proposed [32, 33]. In general, symmetrically arranged HC-ARFs can support
degenerate polarization modes and polarization coupling due to random perturbations is
inevitable. The idea of breaking the core symmetry and introducing ellipticity in the
antiresonant reflecting optical waveguide (ARROW)-guiding fibers has therefore been
investigated in some of the simplest structures for terahertz applications [36], following the
idea of exploiting form-birefringence that is used in solid core PM fibers. In this work, to
study the potential of generating polarization mode birefringence (PMB) through core
ellipticity in HC-ARFs, we start by considering an ideal elliptical tube as the simplest
possible HC-ARF structure. The thickness of the tube needs to satisfy the resonance equation
of ARROWs in order to support the leaky fundamental mode (FM) at frequencies f in the
desired (mth) anti-resonance band [30]:
22 22
12 1 12 1
.( 1) . ,1,2,....
22
cm cm
fm
tn n tn n
−<< =
−− (1)
where c is the speed of light in vacuum. n1, n2 and t1 are the refractive indices of core, clad
and the thickness of the tubes, respectively. Using Eq. (1) and choosing an operating
wavelength (λ0 = 1.55µm) in the second anti-resonance band, a thickness t1 = 1.172µm is
estimated to produce the lowest loss for a core radius of R = 10µm. In order to investigate the
effect of the tube’s ellipticity on its PMB, elliptical tubes with different major (b) to minor (a)
axis ratios have been simulated by using a commercial finite element (FEM) mode solver
(Comsol®) with optimized meshes and perfectly matched absorbing boundary layers [30].
Figure 1 shows the simulated confinement loss and PMB (NV/H = effective index of
vertical/horizontal polarization) of the FM for different tube core dimensions. The simulation
result shows that the PMB initially increases on increasing the ellipticity of the core, as shown
in Fig. 1(a). However, after a peak around b/a ~1.6, it decreases and settles to an almost
constant value for further increase in the ellipticity, regardless of the core size. This indicates
an optimum point for ellipticity. Additionally, it is seen that decreasing the core size to
achieve a discernible PMB is necessary. However, as shown in Fig. 1(c), reducing the core
size results in impractically large loss, which in practice limits the minimum core size one can
choose.
It is well known that loss in an antiresonant HC fiber can be reduced by designing the core
surrounding membranes with negative curvature [16, 17, 22]. Amongst various designs, the
NANF structure has shown the potential for very low optical loss, in principle comparable to
that of solid silica fibers [30]. This structure consists of a number of equally spaced non-
touching nested glass tubes, as shown in Fig. 2(a), which are attached to the inner wall of a
larger jacket tube. The nested tube arrangement provides the anti-resonant cladding of the
fiber, with the hole in the middle acting as the core. In this design, confinement loss is
reduced dramatically due to the absence of nodes in the core surround, which can induce
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22946

undesired coupling to lossy cladding elements (e.g. membranes) [20], and to the addition of
coherent radially arranged reflecting layers. One might therefore want to consider whether a
NANF with an elliptical cross section (and a small core) might provide a way to generate a
significant PMB and low loss. The simulated loss and PMB of an elliptical core NANF
obtained by rearranging and reducing the size of two opposite sets of nested tubes (see bottom
structure in Fig. 1(b)) for different ellipticity is shown with markers in Fig. 1, and compared
with the behavior of an elliptical tube. For both structures, a reduction in core size is helpful
in increasing the birefringence, but at all core sizes the birefringence peaks around an
optimum value of ellipticity for each core size. Larger ellipticities than that do not produce
any higher birefringence.
Fig. 1. The characteristic profile of a single tube for different minor axis (a = 10µm, 15µm,
20µm) vs major axis b with ts = 1.172µm at λ0 = 1.55µm. (a) Birefringence of the fibers for
different ellipticity, (b) tube structure (top) and NANF structure (bottom), (c) loss profile of
fibers for horizontal and vertical polarization for different ellipticity. Properties of elliptical
core NANF for a = 10µm, 15µm at optimum ellipticity are provided for comparison.
As the results suggest, although the NANF structure can reduce the loss by many orders of
magnitude in comparison to single tube, the resultant PMB of only 1-2 × 10−6 is too low for
PM operation. Additionally, to increase the PMB in NANF fibers by form-birefringence, we
have explored other geometrical arrangements of the nested tubes that provide a large core
ellipticity, as shown in Fig. 2(b). However, our simulations show that despite the fairly large
core ellipticity achieved in these fibers (b/a> 2), the PMB remains below ~10−5 in all cases.
Fig. 2. (a) 6 tube NANF design, (b) alternative NANF with elliptical core designs studied in
this work, which we found to be unable to provide birefringence >10−5.
In conclusion, the achievable PMB by making the air hole elliptical is not adequate to
guarantee the HC-ARF a PM behavior in the visible/infrared regions of the electromagnetic
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22947

spectrum. However, there are some examples of multilayer ARROW fibers with large enough
PMB factor at longer operating wavelengths (THz guidance) [36].
3. Birefringence in bi-thickness NANFs
In addition to elliptical core designs, another potential approach to induce significant
birefringence in ARFs has been proposed and has previously been commercially exploited in
HC-PBGFs [31] This relies on introducing a controlled anti-crossing between air-guided and
glass-guided (surface) modes. By using glass elements of different thickness along orthogonal
directions in the core surround, a significantly different effective index for the two
polarizations of the FM can be achieved at the edge of an anti-crossing point [37]. In this way
the birefringence is introduced by anti-crossing phenomena rather than by geometrical
asymmetry [34].
Whilst this has only been experimentally applied to PBGFs so far, a couple of theoretical
studies have explored the extent to which the concept can be applied to ARFs that guide light
through a different physical mechanism [32, 33]. The main idea is to generate birefringence
by introducing an asymmetry through modifying the thickness of selected core surrounding
membranes in ARFs. Horizontal and vertical glass membranes can operate in either the same
antiresonant window [32] or in a so called “Hybrid” regime where they operate in different
anti-resonance bands with opposite locked field phases at the outer boundaries [33]. In this
work, we undertake a systematic study of both operating regimes, highlighting their benefits
and drawbacks. Additionally, we show that by applying a suitable modification to the design
of a state-of-the-art NANF, not only it is possible to enhance its birefringence and therefore
its polarization maintaining behavior, but also to achieve a broad bandwidth that spans several
100s nm at a low enough loss for most device and power delivery applications.
3.1 The effect of thickness change on birefringent NANF
Following the abovementioned strategy (de-symmetrizing by an appropriate modification of
the thickness of some glass membranes), one could think of a 6-tube low-loss NANF [30] and
change the thickness of two oppositely located nested tubes with a different thickness than the
remaining four. By simulating ANFs with 3 to 8-tubes of two different thicknesses, however,
we have found that a structure with 4-tubes can achieve a considerably higher birefringence,
as shown in Fig. 3. We believe that the orthogonality of the core surrounding glass webs to
the polarization of the core field is key to explain these effects. The fiber with 4 tubes is the
only geometry able to guarantee membranes that are orthogonal to the mode polarization This
effect can also be explained by the tendency of each polarization of the FM at the edge of the
anti-crossing frequency to couple preferentially to cladding tube modes with a similar
polarization [33, 35].
Fig. 3. Birefringence (Nv-Nh) of 3-tube to 8-tube ANF design at λ0 = 1.55µm alongside their
structures with same core size (R = 7µm) and bi-thickness cladding tubes (thin tubes (blue) =
0.372µm and thick tubes (orange) = 0.633µm). The 4-tube structure shows maximum
birefringence, and the structures with odd number of tubes show lower birefringence due to
lack of 2-fold symmetry.
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22948

It is also noticeable that structures with odd number of tubes show lower birefringence,
which can be explained by their lack of 2-fold symmetry, resulting in worse orthogonality
between core boundaries and direction of polarization of the fundamental modes. We leave
this topic open for further investigation in future works.
Since the 4-tube fiber was found to offer higher birefringence, the rest of the paper will
focus on this structure and further optimize its performance.
Our starting point is a 4-tube NANF in which the tubes aligned along the vertical axis
have a thickness t1 which can be the same (normal operation) or different from that (t2) of the
tubes along the horizontal axis (i.e. a bi-thickness fiber). In order to minimize the loss while
maintaining a small enough core to provide large birefringence in air (see section 2), we
chose a design with two nested tubes inside a larger external one, as shown in Fig. 4(a), which
details all the relative design parameters. In order to optimize the confinement loss for a core
radius of R = 7µm and an azimuthal tube distance d = 1.5 µm, the distance between the inner
tubes (z) was calculated using the quarter wave condition which ensures the in-phase
reflection of each layer towards the core [20]:
0
0.65 .
2
zRR
u
π
≈≈ (2)
where u0 is the solution of the first zero of the zeroth order Bessel function. Depending on the
choice of t1 and t2, the fiber can operate in the following three regimes:
1. Non-birefringent guiding (standard operation, see Fig. 4(b)): t1 = t2 and the fiber shows
no birefringence.
2. Same band birefringent operation: t1 and t2 are slightly different so that anti-crossing
shifted birefringence is obtained whilst anti-resonance occurs in the same window
for both FM polarizations.
3. Different band birefringent operation: t1 and t2 are considerably different, such that the
anti-resonance in the orthogonal directions occurs in different bands and the FM
polarizations have opposing locked field-phase at the outer boundaries (hybrid
regime birefringence) [33].
As a starting point we begin with a uniform NANF (t1 = t2). Using Eq. (1), a thickness t1 =
t2 = 0.6µm is selected such that the fiber operates in the first anti-resonance band around near-
IR telecoms wavelengths (~1.55µm). Figure 4(b) shows the effective refractive index (Neff)
and loss profile of the designed fiber. The target operating wavelength of 1.55µm has been
placed within the first anti-resonance band but not at the lowest loss point, for the purpose of
examining different scenarios. The fiber, as expected, shows no birefringence. However, the
fiber’s polarization properties change dramatically when the thickness of all tubes along one
axis is modified. Figure 5 shows the evolution of the loss for each FM polarization and the
effective index difference at λ0 = 1.55µm when t2 is changed for a fixed t1 = 0.6µm. It worth
mentioning that increasing t2 causes the inner radii of the corresponding tubes to decrease,
while the other parameters remain unchanged.
At t2/t1 = 1 the fiber presents no birefringence and operates in the first anti-resonance
band. As the thickness of the horizontal tubes (t2) increases, the resonance region of these
tubes start to shift towards the wavelength λ0 = 1.55µm while the vertical tubes do not shift
and still work around the original anti-resonance condition. The induced birefringence under
this condition can be explained in various ways through:
a) the different behavior between parallel and perpendicular polarization of core modes at the
anti-crossing point [34];
b) the close relationship between the increase in Neff and the sharp rise in confinement loss
based on the Kramers-Kronig relation, which is caused by the reduction in anti-resonance
reflection of the glass membranes at the edge of the resonance frequency band [33];
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22949

c) the difference in confinement radius (not to be confused with the geometrical radius) of the
individual fundamental polarization modes for different thicknesses, which cause a
change in Neff that can be explained by the well-known connection between mode radius
and Neff in an anti-resonance structure [33].
Fig. 4. (a) Bi-thickness NANF and its design parameters, (b) simulated optical performance of
a uniform NANF with t1 = t2 = 0.6µm, R = 7µm, d = 1.5µm and z = 0.65R.
As t2 continues to increase, its associated resonance frequency reaches λ0, which induces a
large loss, marked as “resonance region” in Fig. 5. Beyond this point, the birefringence
repeats the previous pattern, with fairly large (>10−4) positive and negative values achievable.
However, in this case the horizontal tubes operate in a higher anti-resonance band (second
band when t2/t1 = 1.5-2.3, or third band when t2/t1 = 2.7-3.6). Moreover, a zero birefringence
crossing is obtained for certain thicknesses. Interestingly, as Fig. 5(a) shows, while the
vertical polarization (V-pol) experiences a similar loss as for t2/t1~1, the loss of the horizontal
polarization (H-pol) decreases by as much as two orders of magnitude; a feature that can be
exploited in a polarizing fiber design.
Fig. 5. Optical properties of the NANF in Fig. 4(a) (t1 = 0.6µm, R = 7µm, d = 1.5µm, z =
0.65R) at λ0 = 1.55µm as t2 / t1 varies between 1 and 4: (a) polarization mode loss, (b)
birefringence.
As shown above, sweeping t2 provides a good design tool to find the highest possible
birefringence at a single wavelength for a pre-determined t1. However, the effect of the initial
choice of t1 on loss, and the mutual effect of t1 and t2 on the bandwidth over which a large
birefringence is achieved are not fully explained by the above picture. In the next section, we
study the characteristic profile of different combinations of fibers in order to draw a clearer
picture of the role of each design parameter.
3.2 Bi-thickness NANF: a systematic study
In order to study the behavior of the fiber for different combinations of thicknesses (t1, t2) and
working regimes, we simulated the spectral characteristics (i.e. loss and birefringence vs
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22950

wavelength) for several combinations of thicknesses (t1, t2) from Fig. 5 as highlighted by the
dashed lines. Each combination case represents a different working regime in order to
highlight the mutual effect of t1 and t2 on loss, birefringence and bandwidth. Figure 6 shows
the results for fibers with t1 = 0.6µm, and t2 = 0.66, 1.05, 1.42 and 1.8 µm, alongside the non-
birefringent NANF case (Fig. 6 (a)). We defined a practical working condition for the fibers,
where the combination of birefringence and loss is reasonable. That is a loss lower than
1dB/m (practical loss for most short length applications) and a birefringence with a beat
length smaller than 15.5 mm (corresponding to a birefringence of 10−4 at 1.55µm - the same
order of magnitude as conventional solid-core PM fibers). These regions correspond to below
the red dashed-line in the loss curve and outside the red dashed-line box in the birefringence
plots.
Starting with the case t2 = 0.66µm, the loss shows only a slight difference for the
perpendicular FM polarization and some birefringence starts to appear, although this is still
rather modest in magnitude, as shown in Fig. 6(b). The fiber therefore does not meet the
defined loss and birefringence criteria.
Beyond a certain thicknesses, the resonance region of the horizontal tubes falls after λ0.
For t2 = 1.05µm, a large birefringence and acceptable loss can be observed as shown in Fig. 5.
Figure 6(c) shows the optical properties of the fiber in this case. Here, the second resonance
wavelength of t2 has become large enough to reach the first resonance wavelength of the
thinner tubes (t1) (mt2 = 2 in Fig. 6(c)). The fiber has an acceptable low loss window in the
Hi-Bi region between the two resonances mt2 = 1 and 2. In fact, as shown earlier, only one
polarization of the fiber has a lower loss as compared to the symmetric fiber. This behavior
can be explained by the fact that the loss associated with each polarization of a FM is mainly
controlled by the membranes perpendicular to the electric field of the mode, and operation in
the second window creates a lower leakage loss than in the first one. That is, in this case, the
losses of the horizontal/vertical polarization mode, which are perpendicular to the
thicker/thinner tubes (t2/t1), are determined mainly by the anti-resonance condition of the
perpendicular tube, respectively.
As a result, compared to the uniform NANF, the H-pol experiences a lower loss as the
thicker tubes (t2) are working in the second anti-resonance band, while the loss of the V-pol is
dominated by the thinner tubes (t1), which operate in the first anti-resonance band and shows
higher loss. In other words, the bi-thickness fiber operates on each individual polarization in
the same way as a uniform fiber with tubes as thick as the tubes perpendicular to the electric
field of that polarization.
Since the overlap between second and first anti-resonance bands of t2 and t1 respectively
provides a good window of operation, we studied the possibility to shift the first resonance of
t2 further towards longer wavelengths in order to widen its second anti-resonance band. To
examine this approach, we simulated a bi-thickness fiber with t2 = 1.42µm (see Fig. 5), with
the results shown in Fig. 6(d). At a first glance, the bandwidth of low loss operation has
increased. However, the birefringence of the fiber has significantly reduced. A closer look
reveals a zero birefringence crossing in this region. This is a characteristic feature of this
region, which essentially limits the bandwidth of high birefringence achievable. It is caused
by the fact that the second window is narrower than the first and hence has a steeper
dispersion, which inevitably crosses that of the first window.
To include other possibilities and to study a broader span of thicknesses, we have finally
considered t2 = 1.8um, chosen from the third region of Fig. 5 with high birefringence and low
loss. In this case, as it can be seen in Fig. 6(e), three anti-resonance bands of t2 overlap with
the first anti-resonance band of t1. While the first one (at λ>3um, not shown) has a very high
loss, the second and third bands show low loss operation. In this case, similar to the previous
case (t2 = 1.42µm), a zero-crossing occurs in the second band which limits its bandwidth
significantly. However, there is a small high birefringence window overlapping with a low
loss band, which provides a good operational window in the third anti-resonance band, as
shown by the green arrows in Fig. 6(e). Here, some features are noticeable in the
characteristic profile of this fiber. For instance, the third anti-resonance band shows almost
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22951

identical birefringence and loss profiles to the second band in the case with t2 = 1.05µm,
which demonstrates that it is possible to achieve Hi-Bi even in the same outer locked field
phase condition (third band over first band), see [29]. In other words, the existence of the
“Hybrid” regime [33] is not necessary to achieve Hi-Bi and low loss simultaneously.
However, as the bandwidth of the third anti-resonance band is intrinsically smaller than the
second one, the achievable birefringence bandwidth in this region is generally narrower.
Fig. 6. Simulated loss and birefringence profile of a bi-thickness 4 tube NANF with t1 =
0.6µm, R = 7µm, d = 1.5µm and t2 = 0.6, 0.66, 1.05, 1.42 and 1.8µm ((a) to (e) respectively).
The red dashed line in the loss profile shows 1 dB/m and the red dashed lines in the
birefringence profile indicates a beat length equal to 15.5 mm (roughly the birefringence of a
conventional solid-core PM fiber). The green arrows show the defined practical operating
window with loss lower than 1 dB/m and beat length larger than 15.5 mm.
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22952

Amongst all explored scenarios, just three practical regions can be used to simultaneously
achieve Hi-Bi and low loss, as shown by green arrows in Fig. 6. A) The region right after,
where the second resonance of t2 and the first resonance of t1 overlap (case t2 = 1.05µm); B &
C) The regions where the third and second anti-resonance band of t2 and the first anti-
resonance band of t1 overlap (case t2 = 1.8µm). Although the regions of interest show high
birefringence and low loss, their bandwidth is limited not only by the intrinsic bandwidth of
their anti-resonance band, but also by the abrupt loss increases near the resonance regions.
Additionally, we have shown that shifting the second resonance away from the first one has
some limitations due to the presence of a birefringence zero-crossing, which prevents ultra-
broad band operation (Fig. 6 (d)).
Figure 6 also shows another region of potential interest, the broad Hi-Bi region in the first
anti-resonance band of the fiber with t2 = 1.05µm (Fig. 6(c)). This spectral range has
intrinsically the largest possible bandwidth between all the cases as it is placed in the first
anti-resonance band of both t1 and t2, simultaneously. However, despite its Hi-Bi and
broadband feature, this region suffers from a high loss because it is close to the edge of the
resonance of t2 and operates in two different parts of the same anti-resonance band of t1 and
t2. The next section focuses on improving the loss of this region.
3.3 Modified bi-thickness NANFs
In order to achieve broadband and large birefringence that benefits from the potential of the
first anti-resonance band, we need to overcome its large loss through a modified design. Here,
we propose a structure that has reduced loss at the edge of an anti-crossing by eliminating the
in-phase resonance layers. The proposed structure exploits anti-resonant inner tubes operating
at the optimum point to minimize the loss in the first anti-resonance band, while the
birefringence is still induced by the change in the thickness of the outer tubes. Figure 7(a)
shows the proposed modified structure and its design parameters. Figures 7(b) and (c) show
the performance of the proposed fiber with similar parameters as those presented in Fig. 6(c),
alongside the properties of a normal bi-thickness fiber with similar parameters, shown in Fig.
7(d) (a duplicate of Fig. 5(c) for ease of comparison).
A significant loss reduction is achieved in the first anti-resonance band of our new design,
reducing the minimum loss below 1dB/m, while the loss increases in the second anti-
resonance band compared to the previous fiber, as shown in Fig. 7(c). Differently from the bi-
thickness structure of Fig. 4 (a), in this arrangement, similar thickness t1 of all the inner tubes
guaranties an improved in-phase field confinement with lower loss in the first anti-resonance
band, whilst a large birefringence through the outer tubes (t2) is achievable. In the proposed
structure, the dominant contribution to confinement is provided by t1 for both polarizations,
while in the original bi-thickness structure, t1 and t2 almost independently control the loss of
the perpendicular polarization modes. This feature in the new design not only gives the ability
to minimize the loss by optimizing only one thickness, but also increases the possible tuning
range of t2 to provide larger and broader birefringence outside its resonance by eliminating
the strong dependency of loss on this parameter.
Although the design in Fig. 7 shows improvements in terms of loss and birefringence in
the first antiresonant window, its parameters are not optimized for the lowest loss and largest
operating bandwidth. To obtain the optimum design at the operating wavelength λ0 = 1.55µm,
we started by considering a uniform NANF with single thickness t1. We used Eq. (1) to
calculate a starting value for t1. Then, we refined the thickness t1 using a parametric sweep
simulation to achieve the lowest loss at the first anti-resonance band, which resulted in t1 =
0.372µm. Then we introduced the thicker tubes with thickness t2 in the design. Using another
parametric sweep similar to Fig. 5, t2 was optimized to provide the highest birefringence. The
result is a bi-thickness 4-tube NANF with t1 = 0.372µm, t2 = 1.7t1, R = 7µm, d = 1.5 µm and z
= 0.65R. Figure 8 shows the simulated properties of the optimized Hi-Bi NANF with
improved loss in the first anti-resonance band around λ0 = 1.55µm.
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22953

Fig. 7. Proposed modified bi-thickness NANF: (a) structure; (b) birefringence and (c) loss of
the proposed NANF with the same parameters as in Fig. 6(c), (d) loss of standard bi-thickness
NANF (duplication of Fig. 6(c) for comparison). Anti-resonant field confinement by inner
tubes of thickness t1 provides a significant improvement in the loss of the first anti-resonance
region.
This design and method, not only reduces the loss for both polarizations over several
hundred of nm, but also increases the birefringence of the fiber simultaneously. Using this
design strategy, we achieved an anti-resonance hollow core fiber with ultra-broadband high
PMB and sub dB/m loss. The fiber has a birefringence of ~1.5 × 10−4 over a ~550nm
bandwidth (loss <1dB/m) with a minimum loss as low as 0.043dB/m at λ0. The proposed fiber
is therefore ideally suited for ultra-broadband applications requiring polarization maintenance
over almost all optical communication bands (E, S, C, L and U bands).
Fig. 8. (a) Loss and (b) birefringence of the optimized new proposed design for λ0 = 1.55µm
with R = 7µm, Z = 0.65R, d = 1.5µm, t1 = 0.372µm and t2 = 0.633µm. The fiber shows a very
broad bandwidth (~550nm) PMB = ~1.5 × 10−4 with a loss <1dB/m.
Figure 9 shows the mode profile of both FM polarizations at λ0. As can be seen, the outer
horizontal thicker tubes (t2) provide lower reflectivity for the horizontal mode, as they operate
at the edge of the resonance window, allowing some of the field to leak. However, the inner
rings, which operate in perfect anti-resonance conditions, reduce the leakage of the field. As a
result, the loss of the horizontal mode is only slightly higher than that of the vertical one
whereas a large birefringence is obtained.
In practice, the proposed tube thicknesses in our design are achievable and have already
been demonstrated in simpler single ring tubular fibers [38]. However, in order to relax the
requirement for thin tubes and to allow easier fabrication, one can design the fiber to operate
in the second anti-resonance window, if the full 550nm of low loss PM operation are not
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22954

strictly required. Figure 10 shows the optical properties of such an alternative fiber with
thicker tubes operating in the second anti-resonance band (t1 = 1.172 µm, t2 = 1.42µm).
Fig. 9. 3-dB contour plots of the normalized power of the fundamental mode with (a)
horizontal and (b) vertical polarizations (red arrows indicate the direction of the electric field)
at λ0 = 1.55µm for the optimum proposed design. (c) Cross-sectional profile of the normalized
power in each polarization direction. A larger field penetration through the outer horizontal
tubes (t2) than the vertical ones (t1) is clear while the horizontal inner tubes (t1) play a critical
rule in confining the field and reducing the loss.
This fiber shows an almost constant PMB value as large as ~1.45 × 10−4 with loss <1dB/m
over a ~300nm bandwidth covering the entire C, L and U optical communication bands. The
loss in this design reaches its lowest value of 0.006dB/m at λ = 1.57µm, and is around
0.01dB/m across the C + L bands. In comparison, based on our simulations, a fiber with a
similar core size in the hybrid regime (see Section 3.2 and [33]) has a loss value that is more
than 20 times larger due to the presence of leakier inner tubes and the close spectral proximity
of the resonant edge of t1 and t2. Although the design of Fig. 10 shows a narrower bandwidth
than the first band design (Fig. 9), it presents some additional benefits, such as a lower
minimum and polarization dependent loss, and likely greater mechanical robustness because
of its thicker tubes.
Fig. 10. (a) Loss and (b) birefringence of the optimized new proposed design with thicker
tubes at λ0 = 1.55µm with R = 7µm, Z = 0.65R, d = 1.5µm, t1 = t22 = 1.172µm and t2 = 1.42µm.
The fiber shows a large bandwidth (~300nm) PMB = ~1.45 × 10−4 with a loss < 1dB/m.
4. Single polarization designs
In addition to PMB, some applications like ultra-sensitive sensing or gyroscopes, or ultra-high
power single polarization laser delivery may require a large polarization extinction ratio. The
large birefringence in the fiber designs proposed so far decreases the beat length and reduces
the possibility of coupling between polarization modes, which allows the input polarization
state to be maintained. Here, we propose a modification to our polarization maintaining Hi-Bi
NANF to provide both PMB and a large polarization extinction ratio, which transforms it into
a polarizing fiber. The new structure not only has the favorable Hi-Bi characteristic but also
shows a very large polarization differential loss (PDL) that can remove the unwanted
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22955

polarization component of the input light, as well as limiting distributed inter-polarization
coupling due to surface roughness and micro-bends [34].
Our modification is based on a well-known method used to strip higher order modes from
a few moded fiber [39, 40], which we have adapted to achieve “polarization dependent mode
stripping”. In this method, we selectively couple a specific polarization of the fundamental
core mode into a lossy cladding mode, while the orthogonal polarization experiences almost
no change. Here, the two main structural features of the former proposed 4-tube NANF are
fundamental.
First, the induced large birefringence in the fiber creates an opportunity to clearly separate
the effective index of the two polarizations, of which only one is index matched and out-
coupled to lossier cladding (i.e. guided inside the tubes) modes. Based on the similarity of the
guidance mechanism in NANFs and in simpler hollow core fibers made of a borehole in
glass, it is possible to calculate the necessary geometrical specifications to match the index of
the core and cladding modes [41]. The effective index (Neff) of the FM in a single tube can be
calculated with the following exact equation [42]:
2
0
0
1.
.
eff
u
NRK

=−


(3)
where K0 = 2π/λ, R is the core radius and u0 is similar to Eq. (1). As it is clear from this
equation, the effective index is related to the size of the tube. Therefore, in order to have
index matching between the core mode and one of the peripheral tubes in a NANF, they must
have approximately the same size.
Here, the second useful feature of the 4-tube NANF is the very large size of the nested
tubes. This allows the described mode matching to be engineered in the design. In fibers with
more than four nested tubes, although the cladding tubes are large enough to strip the higher
order modes [30, 39, 40], they are too small to enable fundamental mode out-stripping.
Following this idea, we modified the radius of two opposite inner clad tubes in order to
introduce index matching to the polarization with the highest loss in the Hi-Bi NANF of Fig.
8.
Figure 11 shows the structure and mode profiles of both high and low loss polarizations of
the proposed polarizing NANF (P-NANF). The modified nested structure with a different
distance (Z1) between the two innermost rings in the vertical direction opens up a matching
window for vertical polarization to couple to the cladding mode. Consequently, vertically
polarized light leaks through the outer tube. This does not occur for the light on the
Horizontal polarization, Figs. 10(b) and (c).
Figure 12 shows the simulated PMB, the loss of each polarization and the loss ratio
between the two polarization for a fiber with R = 7µm, Z = 0.65R, Z1 = 1.74R, d = 1.5µm, t1
= 1.172µm, and t2 = 1.42µm that has been designed for operation around λ0 = 1.55µm. Here,
we plot the polarization loss ratio (PLR) factor instead of the polarization dependent loss
(PDL) as the ratio can provide a better measure of the effectiveness of the polarization
stripping effect.
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22956

Fig. 11. (a) The proposed structure for polarizing NANF (P-NANF) and its (b) low loss
polarization (horizontal), and its (c) high loss polarization (vertical). The coupled field to the
cladding tubes in the vertical polarization increases the loss significantly.
The modified design shows a relatively large bandwidth of ~100nm with an almost
constant PMB of ~1.5 × 10−4. The loss of the low loss polarization is less than 1dB/m and the
design exhibits a PLR of about 1000 at the wavelength of 1.55µm. A polarizing window of
~10nm with PLR>100 can completely eliminate any undesirable cross-coupled polarization
over a sufficiently long propagation distance. The fiber can therefore provide an almost pure
polarized output, regardless of the input polarization state, which may have applications in,
manufacturing and machining as well as in high precision gyroscopes that require modes with
relatively pure linear-polarization.
Fig. 12. (a) PMB, (b) loss and (c) PLR (High-loss/Low-loss) of the optimized P-NANF at λ0 =
1.55µm with R = 7µm, Z = 0.65R, Z1 = 1.74R, d = 1.5µm, t1 = 1.172µm and t2 = 1.42µm. The
fiber has a bandwidth of ~100nm with PMB~1.5 × 10−4 and loss of the low-loss polarization
<1dB/m. The PLR at λ0 = 1.55µm is 1000 which gives the fiber a strong polarizing property.
5. Conclusions and discussion
In summary, we have presented a detailed study of various ways to obtain broadband
polarization maintaining capability in antiresonant hollow core fibers, in particular for the
NANF structure. We showed that achieving birefringence through asymmetric core designs
adopted following the form-birefringence strategy in solid-core PM fibers is not effective in
ARFs. Next, we adopted a different strategy, used in birefringent PBG fibers, where we broke
the geometrical symmetry of a 4-tube NANF by introducing a bi-thickness design. To study
the induced birefringence in this design, we introduced a decomposition analysis through
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22957

which we showed the thickness of tubes along the horizontal and vertical directions control,
almost independently, the properties of the two respective polarizations. This approach
provides a better understanding of the role of the individual thickness parameters in this
structure as well as a method for practical fiber design.
Using the developed method, we identified a region with potential for ultra-broadband
high birefringence within the first antiresonant window of a bi-thickness NANF. By
proposing a radical design, we showed that the loss in this region can be mitigated, despite the
small core size necessary for large birefringence. Our design not only offers a high
birefringence ultra-broadband spectral window ~550nm (that is almost 6 times boarder than
ever reported in a birefringent HC-ARF [33]) with birefringence as large as 1.5 × 10−4 and a
loss of less than 1dB/m, but it also allows for losses as low as 0.04dB/m at a wavelength of
1.55μm. In addition, by sacrificing some bandwidth we proposed an alternative design with
thicker tubes operating in the second anti-resonance band that achieves an order of magnitude
lower loss (0.004dB/m at the wavelength of 1.57μm) and a large bandwidth (~300nm).
In order to add a polarizing effect into the design, we introduced a new polarization
dependent mode stripping mechanism into the proposed bi-thickness NANF structure. The P-
NANF structure offers a polarization dependent loss ratio of 30dB at the operating
wavelength of 1.55μm with a minimum loss of 0.076dB/m and a low loss bandwidth of
100nm. With such a high loss ratio, the fiber can act as a polarizer for any unpolarized input
over only a few meter of propagation, which would be of interest in applications requiring
extremely high polarization extinction ratio.
Although the proposed fibers have been designed and optimized to operate in the
telecommunication wavelength bands (i.e. around 1.55μm), the geometrical scalability of
ARFs with the operating wavelength allows a straightforward application of similar design
scenarios to other wavelengths more relevant to other applications, e.g. 1 µm for high power
laser delivery or visible wavelengths for biomedical nonlinear endoscopy.
Funding
UK EPSRC through grants EP/I01196X/1 (HYPER-HIGHWAY).
Acknowledgments
FP and DJR gratefully acknowledge financial support from the UK Royal Society. The data
for plots in this paper can be found at http://dx.doi.org/10.5258/SOTON/397073.
Vol. 24, No. 20 | 3 Oct 2016 | OPTICS EXPRESS 22958