We present the fabrication of multimaterial polarization maintaining optical fibers. We exploit the flexibility of the powder-in-tube process for fabricating silica-based optical fibers composed of two rods of glass material on the sides of the core. We demonstrate the capability of this process to use glass material with properties sorely different from the ones of silica for developing high-birefringence optical fiber. This proof-of-concept paves the way for the use of different materials with specific properties for improving the performances of polarization maintaining optical fibers.
© 2017 Optical Society of America
Polarization Maintaining (PM) optical fibers are of great interest for many applications in fiber lasers, non-linear optics, telecommunications and sensing. A PM fiber is characterized by its value of birefringence that is introduced either by breaking the hexagonal symmetry of the fiber structure [1–5] or by internal stress inducing an anisotropy of the refractive index in the core region [6–8].
The first approach mainly based on photonic crystal fiber technology leads to large birefringence, typically exceeding 10−3 at 1.55 µm . However, large birefringence is obtained in a photonic crystal cladding with a pitch close to wavelength size, limiting this approach to PM fibers with small core size. In contrast, PM fibers with larger core size are possible with the second approach since the birefringence is introduced by two stress-applying parts (SAP) placed opposite to each other on the side of the core [6, 9]. The SAP consists of a material with a coefficient of thermal expansion (CTE) different than the one of silica which gives rise to stress fields in the fiber. The birefringence induced by the elasto-optic effect depends only on the level of stress imposed by the material, the shape and the positions of the SAP. Boron-doped silica is mostly used for making the SAP. It has a CTE almost two times larger than silica one, with thermo-mechanical properties compatible with fabrication processes of optical fibers. Its refractive index, lower than silica one, allows positioning the SAP close to the core for enhancing the birefringence. However, the birefringence is limited to typically only a few times of 10−4, depending on the shape and position of the SAP [10–12].
In the present work we propose to exploit the flexibility of the powder-in-tube fabrication process for replacing the boron-doped silica in the SAP with another glass having a larger CTE. Indeed, the powder-in-tube technique is very suitable for the fabrication of multimaterial optical fibers composed of glass-ceramics , metallic cores [14-15], or with embedded metallic particles in the cladding . The aim of this work is to propose a method that gives more flexibility for fabricating PM fibers with SAP composed of a material with larger CTE. This would yield to higher birefringence and may confer new properties to PM fibers allowing new ranges of performances in the fields of fiber lasers, non-linear optics, telecommunication and sensing domains.
As a proof-of-concept, we use a glass system based on SiO2-Al2O3-La2O3 . This glass is sorely different than boron-doped silica. It has a CTE almost ten times larger than silica one, a much lower glass transition temperature and a larger refractive index. Furthermore, we have doped it with CuO in order to deteriorate its transparency and thus proving the possibility to use materials with large absorption coefficient.
2. Glass fabrication and characterizations
2.1 Glass fabrications
The glass compositions 70SiO2-20Al2O3-10La2O3 mol.% (SAL) and 70SiO2-20Al2O3-10La2O3 mol.% doped with 1% of CuO (SALC) were fabricated by the conventional melt-quenching technique. Commercial powders of SiO2 (Acros Organics, 99.995%), Al(OH)3 (Acros Organics, extra pure), La2O3 (Acros Organics, 99.999%) and CuO (Prolabo, 99%) were weighed in appropriate amounts and intimately mixed in a planetary ball-milling equipment (at 450 rev/min for 1 hour) to ensure good homogenization. The mixtures were melted in a covered platinum crucible in air in two steps: 1 hour at 1550 °C and then 2 hours at 1700 °C. The liquid melts were poured in a heat-resistant stainless steel mould to quench it as a bulk.
The glass powders were prepared by crushing bulk glasses in an agate mortar. Then the small pieces were introduced in a planetary ball-milling apparatus to reduce the particle size. The particle size distribution was measured by a laser particle size analyzer (Malvern Mastersizer) and the median diameter (D50) is around 15 µm.
2.2 Glass characterizations
Glass powders were characterized by thermogravimetric analysis and differential thermal analysis (TGA/DTA) measurements under air atmosphere on a Netzsch STA 449 F3 Jupiter (with a heating rate of 10 °C/min). The glass transition temperature was determined from a glass powder sample (50 mg) introduced in a platinum crucible. The glass transition temperatures (Tg) of SAL and SALC glasses were 877 °C and 850 °C respectively. Since the glass transition temperature of silica is around 1200 °C , SAL and SALC powders in the preform will melt during the drawing step.
Bulk glasses were polished in 5 x 5 x 5 mm samples. Their CTE were measured under air atmosphere on a Netzsch DIL 402 C equipment, with a heating rate of 5 °C/min in the range 30-650 °C. The measured CTE of SAL and SALC glasses are 5.32·10−6 and 6.46·10−6 K−1 respectively. The CTE of SAL glass is 10 and 5 times higher than silica (0.54·10−6 K−1 ) and boron-doped silica (1·10−6 K−1 ), respectively.
Bulk glasses were cut and mechanically polished down to 1.14 mm thick slices for measuring their transmission spectrum with a spectrometer (Varian Cary 5000 UV-Vis-NIR). SAL glass exhibits high transparency in the range 300-2000 nm (Fig. 1) while SALC glass has two absorption bands located around 300 and 800 nm which correspond to absorption bands of Cu+  and Cu2+  ions respectively.
3. Preform and optical fiber fabrications
3.1 Preform fabrications
For this proof-of-concept we have chosen to fabricate PM fibers composed of two circular SAP on the side of a germanium doped silica core, similar to classical Panda PM fibers. Panda PM fibers are typically fabricated by inserting boron-doped silica rods into two drilled holes of a standard single mode fiber perform (usually fabricated by Chemical Vapor Deposition (CVD) method) . In order to gain in flexibility in the choice of SAP material and shape, the SAP were created by simply filling the preform with powdered material (SAL or SALC glasses).
In contrast with CVD method, the powder-in-tube process is more suitable for managing fabrication constraints induced by different materials. Furthermore, this process does not require shaping step of the materials, in contrast with rod-in-tube process, allowing more flexibility in the SAP shape and also longer preform length. The preform of the fiber was realized by stacking into a large silica tube, silica rods around a germanium-doped silica rod that was drawn from a MCVD preform of standard single mode fiber. Two tubes filled with SAL (or SALC) powder are also stacked on the side the germanium-doped silica rod for forming the SAP (Fig. 2(a)). Silica rods are added in the stack to maintain the design of the fiber. It is worth notifying that as for Panda PM fibers, we could replace the stack by a MCVD preform with two air holes drilled at specific positions to respect the required distance between the SAP and the core.
In contrast with boron-doped silica, the refractive index of SAL glass is higher than silica one, which makes it necessary to space the SAP from the core in order to avoid parasitic optical couplings between the core and the SAP. The positions of the SAP and consequently of the two tubes in the preform were calculated with a finite element method based software (Comsol Multiphysics). The simulated Panda fiber was composed of a core diameter of 8 µm with a core/cladding index difference of + 5·10−3 and two SAP of 24 µm diameter. The refractive index of SAP was set to silica + 50·10−3, and an imaginary par of -i·10−6 was added for simulating a SAP material with absorption coefficient. The properties of the fundamental mode (LP01) have been calculated for different positions of the SAP (distance between the core and the SAP (Δ1)). As expected, larger Δ1 leads to smaller fraction of power in both SAP and attenuation coefficient of LP01 mode, as shown in Fig. 2(b).
In order to ensure a large optical isolation between the core and SAP we have chosen to place the SAP at 22 µm to the core. For this value of Δ1 the fraction of power in both SAP is 3.9·10−8 leading to a ratio attenuation coefficient of LP01 versus the absorption of coefficient of SAP material, of 4·10−8, which allow us to use absorbent material such as SALC.
3.2 Optical fibers fabrication
The preforms were drawn down in more than 100 meters of fiber using a drawing tower. The SEM pictures of the cross section of three different fibers are shown in Figs. 3(a)-3(c). The first and second fiber, namely “SAL fiber” and “SALC fiber”, were fabricated by filling both tubes with SAL and SALC powders, respectively. The third fiber, namely “Air fiber”, was drawn with both tubes empty. The depression applied in the preform was not sufficient to close, during the fiber drawing, the interstices in the outer cladding. This was corrected during the fabrication of “SAL fiber” and “SALC fiber”. Nevertheless these small holes do not affect the optical properties (attenuation or birefringence) of Air fiber. This fiber will be used as a reference for evaluating the effect of the SAL and SALC glasses in the SAP.
The refractive index profiles of the fibers were measured at 667.94 nm with a refractive index profile fiber analyzer (EXFO NR-9200) based on the refractive near-field technique. They have the same core/cladding index difference of + 5.9·10−3. For the SAL fiber, the index difference between the SAP and the cladding is + 47.5·10−3 (Fig. 4(a)). It rises to + 91.4·10−3 in the SALC fiber, which is 15 times higher than the core/cladding index difference. This contrasts with the negative value of −12·10−3 for the SAP/cladding index difference in a commercial PM fiber with Panda topology (Thorlabs PM-1550-XP), as shown in Fig. 4(b).
Parameters of the optical fibers including the commercial fiber are summarized in Table 1. All fibers have similar core diameters. For manufactured fibers the distance between the SAP and the core is 5 times larger compared to commercial fiber. This distance was taken in order to avoid optical couplings between the core and the SAP. It is noteworthy that the surface of the SAP is 2 times smaller than the one of the commercial fiber. The size of SAP could be increased by using different silica tubes during the fiber preform preparation.
4. Optical characterizations of the polarization maintaining fibers
4.1 Optical characterizations
The optical attenuations of the fabricated fibers were measured by the standard cut-back method with a broadband light source (supercontinnum based source Leukos SM-20) and an optical spectrum analyser (ANDO AQ-6315A). Three loops of 2 cm radius were applied on the fibers for removing cladding modes. The attenuation spectrums of the SAL and SALC fibers are presented in Fig. 5(a) for cut-back lengths of 27 and 17 m, respectively. Spectral features corresponding to optical couplings between core and SAP (such as oscillations or higher attenuation for larger wavelengths) are not visible in both spectra indicating that the core is optically isolated from the SAP.
The optical attenuation coefficients at 1550 nm are around 0.06 and 0.13 dB/m for the SAL and SALC fibers, respectively. These attenuation coefficients could be largely reduced by applying more drastic cleaning processes during the fabrication steps. It is worth notifying that the insertion of copper (1% mol) in the SAP has a moderate effect on the increase of the attenuation coefficient. Copper is well known to introduce excessive absorption losses with absorption peak of Cu2+ ions around 800 nm .
The single mode guidance regime of the fabricated fibers was confirmed by measuring the near-field pattern at the fiber outputs with an InGaAs camera (FLIR Indigo Alpha NIR camera 900-1700 nm) and a band-pass filter centered at 1550 nm ± 12 nm. As shown in Fig. 5(b), light is confined in the core with a Gaussian-like shape intensity distribution. Light-coupling between the core and the SAP was not observed regardless of the fiber length.
4.2 Birefringence measurements
The birefringence measurement of the fibers was realized by using the high-birefringence fiber loop mirrors (FLM) configuration [23-24] shown in the Fig. 7(a). The PM fiber was spliced with a 3-dB fiber coupler in order to split the input light into two counter propagating light waves that are recombined at the output port of the coupler. The birefringence of the fiber induces a phase difference between the two orthogonal modes resulting in an interference spectrum. In this configuration, the wavelength spacing (Δλ) between interference dips depends on the birefringence (B) and the length of the fiber (L), which leads to this simple equation for calculating the fiber birefringence:
The FLM configuration is simple to implement because the spectral characteristics are independent of the polarization state of the input light and are not so much sensitive to noise disturbances. This configuration is very soft and flexible. The performance of FLM is limited only by the losses of the splice between the fiber and the coupler.
The splicing of PM fibers with the SMF (single mode fiber) of the coupler was realized using a standard program of an arc fusion splicer (Fujikura FSM-100P). The splicing losses are lower than 0.5 dB. The quality of the splicing was observed using an optical microscope (Figs. 6(a) and 6(b)). For SAL fiber, the presence of a small air bubble in the SAP was observed (Fig. 6(b)), but no swelling or deformation of core or outer cladding.
The optical spectrum of the source after the FLM composed of the SAL and Air fibers are plotted in Fig. 7(b), with the output spectrum of the source. In contrast with the Air fiber, the insertion of the SAL fiber in the FLM yields to the apparition of interference dips induced by the birefringence of the fiber. Since the same fiber with empty SAP (Air fiber) has no birefringence, the birefringence of the SAL fiber can be attributed to the SAP, and not to an extrinsic effect introduced during the fiber fabrication.
The mean birefringence of the SAL fiber is 3.19·10−4 ± 1·10−6. It was obtained by calculating (Cf. Equation (1)) the mean birefringence of different dip couples measured on a FLM spectrum. Six different lengths (from 10 to 120 cm) of SAL fiber were used to measure more accurately the birefringence of the fiber. As shown in Fig. 7(c), the wavelength spacing between the dips is inversely proportional to the length of PM fiber. The birefringence of the SAL fiber is slightly smaller than the one of the commercial fiber (4.17·10−4 ± 1·10−6). Knowing that the SAP are placed almost 5 times away from the core compared to commercial fiber and are 2.3 time smaller than the ones in the commercial fiber, this result demonstrates the large elasto-optic effect induced by SAL glass in the SAP. Thereby larger birefringence values might be obtained by a simple increase of the size of the SAP as presented in the following section.
The birefringence of the SALC fiber was measured from FLM spectra shown in Fig. 7(d) with five different lengths (from 40 to 120 cm). The mean birefringence is 3.39·10−4 ± 1·10−6, demonstrating the interest of doping the SAL glass with copper for increasing the CTE of the glass (from 5.32·10−6 and 6.46·10−6 K−1 respectively). Furthermore, this result paves the way to the use of materials with large absorption coefficient and targeted properties (ex. large CTE). This leads to new possibilities for developing very high birefringence fibers and high sensitivity sensors based on high-birefringence FLM configuration.
5. Modeling of SAP size influence on the birefringence
In order to demonstrate further the interest of this proof-of-concept, we have studied the evolution of the birefringence in function of the size of SAP composed of SAL or boron-doped silica, with the parameters of the SAL fiber or commercial fiber, respectively. In the simulation, only the external diameter of the fiber increases with the SAP size. The distances between the core and the SAP (Δ1) and between SAP and the fiber outer diameter (Δ2) are fixed. The properties of silica, boron-doped silica and SAL are summarized in Table 2. They are extracted from reported works on PM fiber simulations [9, 18, 25–27]. We have measured the density on bulk SAL glass by Archimedes method, and the Young’s modulus and Poisson’s ratio with the standard ultrasound technique. The stress distribution during the cooling and its elasto-optical effect on the perpendicular polarization modes and  were simulated with finite element method based software (Comsol Multiphysics). The stress-optical coefficients are 4.2·10−12 and 6.5·10−13 Pa−1  and the temperature values used ranged from 1000°C to 20°C .
The simulation results are plotted in Fig. 8 with the measured birefringence of SAL and commercial fibers (red and black square, respectively). As expected, the birefringence increases with the size of the SAP. The birefringence of the commercial fiber tends to saturate to 6·10−4, while the birefringence of the SAL fiber rise almost linearly with the SAP diameter. Furthermore, the birefringence of the SAL fiber is higher than the one of the commercial fiber for a SAP diameter above 30 µm. These results demonstrate that this process could lead to large core size PM fiber with birefringence larger than 10−3 which is unattainable with boron-doped silica in SAP.
We demonstrated a novel fabrication process based on the powder-in-tube method for realizing high-birefringence optical fibers. This process enables the use of materials in the SAP with properties sorely different from the ones corresponding to boron-doped silica based compositions. We realized a PM fibers with a Panda topology composed of two SAP filled with SAL glass. The birefringence of this fiber (3.19·10−4 ± 1·10−6) is close to the one of a commercial fiber (4.17·10−4 ± 1·10−6) whereas the SAP are almost 5 times away from the core and 2.3 times smaller than in the commercial Panda fiber. This concept can be translated to fibers with more complex designs such as PM large mode area photonic crystal fibers , where the SAP are located far away from the core, which limits their birefringence. The process might allow the fabrication of PM fibers with a birefringence larger than 10−3, which is not accessible for large core fiber (to the best of our knowledge). We extended our study by using copper-doped SAL glass in the SAP in order to increase the birefringence and to demonstrate further the flexibility of this process for fabricating PM fibers with various materials, including materials with large absorption coefficient. This proof-of-concept paves the way to the use of new materials with specific properties for constituting the SAP and improving the performances of the PM fibers. For example, using Al2O3 (high refractive index n ~1.76, CTE ~8·10−6 K−1 and rigidity E ~300 GPa) in the SAP may increase the level of stress on the core and therefore the birefringence of the fiber.
The authors would like to thank Dr. K. Schuster and D. Litzkendorf from Institut für Photonische Technologien (IPHT), Jena, Germany, for their advices in the fabrication of the SAL glass. Also, the authors would like to thank A. Passelergue, Dr. A. Baz, Dr. R. Dauliat, Dr. G. Granger and Dr. D. Pomarede from XLIM Research Institute, Limoges, France.
References and links
1. A. Ortigosa-Blanch, J. C. Knight, W. J. Wadsworth, J. Arriaga, B. J. Mangan, T. A. Birks, and P. St. J. Russell, “Highly birefringent photonic crystal fibers,” Opt. Lett. 25(18), 1325–1327 (2000). [CrossRef] [PubMed]
2. K. Suzuki, H. Kubota, S. Kawanishi, M. Tanaka, and M. Fujita, “Optical properties of a low-loss polarization-maintaining photonic crystal fiber,” Opt. Express 9(13), 676–680 (2001). [CrossRef] [PubMed]
3. T. P. Hansen, J. Broeng, S. E. B. Libori, E. Knudsen, A. Bjarklev, J. R. Jensen, and H. Simonsen, “Highly birefringent index guiding photonic crystal fibers,” IEEE Photonics Technol. Lett. 13(6), 588–590 (2001). [CrossRef]
4. M. J. Steel and R. M. Osgood, “Polarization and dispersive properties of elliptical-hole photonic crystal fibers,” J. Lightwave Technol. 19(4), 495–503 (2001). [CrossRef]
5. M. Szpulak, G. Statkiewicz, J. Olszewski, T. Martynkien, W. Urbańczyk, J. Wójcik, M. Makara, J. Klimek, T. Nasilowski, F. Berghmans, and H. Thienpont, “Experimental and theoretical investigations of birefringent holey fibers with a triple defect,” Appl. Opt. 44(13), 2652–2658 (2005). [CrossRef] [PubMed]
6. J. Noda, K. Okamoto, and Y. Sasaki, “Polarization-Maintaining Fibers and Their Applications,” J. Lightwave Technol. 4(8), 1071–1089 (1986). [CrossRef]
9. P. L. Chu and R. A. Sammut, “Analytical method for calculation of stresses and material birefringence in polarization-maintaining optical fiber,” J. Lightwave Technol. 2(5), 650–662 (1984).
10. N. Shibata, Y. Sasaki, K. Okamoto, and T. Hosaka, “Fabrication of polarization-maintainig and absorption-reducing fibers,” J. Lightwave Technol. 1(1), 38–43 (1983). [CrossRef]
12. M. P. Varnham, D. N. Payne, A. J. Barlow, and R. D. Birch, “Analytic solution for the birefringence produced by thermal stress in polarization-maintaining optical fibers,” J. Lightwave Technol. LT-1(2), 332–339 (1983). [CrossRef]
13. Y.-C. Huang, C.-N. Liu, Y.-S. Lin, J.-S. Wang, W.-L. Wang, F.-Y. Lo, T.-L. Chou, S.-L. Huang, and W.-H. Cheng, “Fluorescence enhancement in broadband Cr-doped fibers fabricated by drawing tower,” Opt. Express 21(4), 4790–4795 (2013). [CrossRef] [PubMed]
14. S. Morris, T. Hawkins, P. Foy, C. McMillen, J. Fan, L. Zhu, R. Stolen, R. Rice, and J. Ballato, “Reactive molten core fabrication of silicon optical fiber,” Opt. Mater. Express 1(6), 1141–1149 (2011). [CrossRef]
15. J.-L. Auguste, G. Humbert, S. Leparmentier, M. Kudinova, P. O. Martin, G. Delaizir, K. Schuster, and D. Litzkendorf, “Modified Powder-in-Tube Technique based on the consolidation processing of powder materials for fabricating specialty optical fibers,” Materials (Basel) 7(8), 6045–6063 (2014). [CrossRef] [PubMed]
16. S. Leparmentier, J.-L. Auguste, G. Humbert, G. Delaizir, and S. Delepine-Lesoille, “Fabrication of optical fibers with palladium metallic particles embedded into the silica cladding,” Opt. Mater. Express 5(11), 2578–2586 (2015). [CrossRef]
17. D. Litzkendorf, S. Grimm, K. Schuster, J. Kobelke, A. Schwuchow, A. Ludwig, J. Kirchhof, M. Leich, A. Schwuchow, and J. Dellith, “Study of lanthanum aluminum silicate glasses for passive and active optical fibers,” Int. J. Appl. Glass Sci. 3(4), 321–331 (2012). [CrossRef]
18. M. J. Weber, Handbook of optical materials (CRC Press, 2002).
19. T. Schreiber, H. Schultz, O. Schmidt, F. Röser, J. Limpert, and A. Tünnermann, “Stress-induced birefringence in large-mode-area micro-structured optical fibers,” Opt. Express 13(10), 3637–3646 (2005). [CrossRef] [PubMed]
20. A. Chahadih, H. El Hamzaoui, O. Cristini, L. Bigot, R. Bernard, C. Kinowski, M. Bouazaoui, and B. Capoen, “H2-induced copper and silver nanoparticle precipitation inside sol-gel silica optical fiber preforms,” Nanoscale Res. Lett. 7(1), 487 (2012). [CrossRef] [PubMed]
22. P. C. Schultz, “Optical Absorption of the transition elements in vitreous silica,” J. Am. Ceram. Soc. 57(7), 309–313 (1974). [CrossRef]
23. E. De la Rosa, L. A. Zenteno, A. N. Starodumov, and D. Monzon, “All-fiber absolute temperature sensor using an unbalanced high-birefringence Sagnac loop,” Opt. Lett. 22(7), 481–483 (1997). [CrossRef] [PubMed]
24. Y. Liu, B. Liu, X. Feng, W. Zhang, G. Zhou, S. Yuan, G. Kai, and X. Dong, “High-birefringence fiber loop mirrors and their applications as sensors,” Appl. Opt. 44(12), 2382–2390 (2005). [CrossRef] [PubMed]
25. K. Okamoto, T. Hosaka, and T. Edahiro, “Stress analysis of optical fibers by a finite element method,” IEEE J. Quantum Electron. 17(10), 2123–2129 (1981). [CrossRef]
28. W. Primak and D. Post, “Photoelastic constants of vitreous silica and its elastic coefficient of refractive index,” J. Appl. Phys. 30(5), 779–788 (1959). [CrossRef]