1. Introduction

Photonic bandgap fibers (PBGFs) have attracted much fundamental and technical interest because of their distinct characteristics such as unusual light guiding mechanism, single material composition, and photonic bandgap effect [1]. Large degree of design flexibility and strong wavelength dependency of the effective refractive index provide a wide range of applications such as wide band transmission, high power delivery, and novel optical devices [2–3]. It was suggested that the birefringence of the PBGF can be increased to be order of 10^-3, which is much higher than that of the conventional polarization maintaining fibers (PMFs) with asymmetric core designs [4]. The first PBGF with a very high group birefringence of 0.025 was experimentally fabricated and characterized by X. Chen et al. [5]. There have been also many theoretical and experimental works to study and demonstrate high birefirngent PBGFs [6–12].

In this paper, we report the results of measurements of birefringence and its sensitivity to strain and temperature in the fabricated EC-PBGF. Using a Sagnac interferometric method, we measured the birefringence of the EC-PBGF with an ellipticity of 1.11 which was fabricated by inducing the asymmetric temperature distribution along the hollow core region in the PBGF. The fabricated EC-PBGF has the wavelength-dependent birefringence within the full bandgap. We also investigated strain and temperature sensitivities of the EC-PBGF-based Sagnac interferometer. The results show that temperature sensitivity of the investigated fiber is positive and decreases with wavelength in the longer wavelengths, becomes zero at zero birefringence wavelength (~1413 nm), and vice versa for shorter wavelengths. Meanwhile, the strain sensitivity is very much dependent upon the wavelength and is negative over the bandgap region.

2. Fabrication of elliptical-core photonic bandgap fiber

The EC-PBGF was fabricated via the well-known stack-and-draw technique. A number of 1mm capillary tubes were stacked to form the designed hexagonal structure and the hollow core was constructed by removing seven central capillaries. It was drawn to make the intermediate preform with the shape of a cane, which has a diameter of ~3 mm. To prevent capillaries from over-inflating or melting together, the fiber tension was increased by reducing the operating temperature to be ~1850°C [13]. This intermediate preform was inserted into an out-jacketing tube prior to drawing into the fiber itself. In this process, we placed the second out-jacketed cane at an off-center position in the furnace, as shown in Fig. 1. Thus, we could induce asymmetric temperature distribution along the hollow core region, i.e., inducing lateral tension, resulting in the elliptical hollow core. Negative pressure was applied to the region between the intermediate preform and second out-jacket tube, whilst applying positive pressure at the top of the intermediate cane region, to ensure a thinner web structure. By regulating negative pressure in the space between the preform cane and jacket, we could change the aspect ratio of the hollow core [14].

 

Fig. 1. Scheme for the fabrication technique of the EC-PBGF; the second out-jacketed cane placed at an off-center position in the furnace

Download Full Size | PDF

 

Fig. 2. SEM image of the fabricated EC-PBGF

Download Full Size | PDF

Figure 2 shows a scanning electron micrograph (SEM) image of the cross section of the EC-PBGF. The core has a long axis of 11.56 μm and a short axis of 10.44 μm and the ellipticity was ~1.11. The photonic crystal structure surrounding the hollow core has a pitch (hole-to-hole distance) of ~4 μm. There were 8 rows of air holes surrounding the hollow core. The air filling fraction of PBGFs in the cladding was over 90%. The fiber diameter was ~125 μm.

3. Birefringent properties of EC-PBGF

 

Fig. 3. Transmission spectrum of the fabricated EC-PBGF.

Download Full Size | PDF

To measure the transmission spectrum of the fabricated EC-PBGF, we spliced two ends of the EC-PBGF to the two single mode fibers (SMFs). When splicing the EC-PBGF to SMF, we adjusted the splicing parameters of an Ericsson FSU-975 splicer to reduce the splice loss between the EC-PBGF and SMF. Figure 3 shows the transmission spectrum of the fabricated EC-PBGF, which was measured by using an optical spectrum analyzer (OSA, Ando AQ 6315B) with a white light source (Ando AQ 4303B). The length of the fabricated EC-PBGF for the transmission spectrum was about ~3 m. Transmission spectrum shows the photonic bandgap extends from 1210 nm to 1610 nm with several dips around 1290, 1480 and 1520 nm. These dips are attributed to the mode coupling between the core and the surface modes. It is already reported that these surface modes which are supported at the core wall (i.e., at the core-cladding interface) lie within the bandgap and lead to a loss in PBGF [15]. The loss of the EC-PBGF in the transmission window was measured to be around 1 dB/m. At this loss levels, the surface modes may contribute to loss over the entire transmission spectrum. An appropriate fiber design may lead to lower loss by removing these surface modes within the bandgap [16,17].

Also we measured the birefringence of the EC-PBGF by building a Sagnac loop interferometer. The relationship between the group birefringence (∆N) and the wavelength spacing of the peaks (∆λ) in the Sagnac interferometer can be given by [18]

where λ is the operating wavelength and L is the length of the EC-PBGF. As shown in Fig. 4, the measured wavelength spacing between all two adjacent peaks at the bandgap wavelength was ranged from ~1 nm to ~44.8 nm.

Using Eq. (1), we calculated the group modal birefringence as a function of wavelength from the transmission spectrum of the EC-PBGF based Sagnac interferometer as shown in Fig. 4(b). Our results show that for the investigated EC-PBGF, the group modal birefringence is relatively high values near the bandgap boundaries, respectively, ~3.5× 10^-3 at 1325 nm and 3.0× 10^-3 at 1480 nm, and probably changes its sign around 1413 nm. The reason for this could be related to the fact that high birefringence is obtained when one polarization mode is close to anti-crossing with a mode associated with the core surround whilst the other remains free of such an interaction as reported in ref. [8]. Similar results were also obtained for different PBGFs [7, 8, 12].

 

Fig. 4. (a) Transmission spectrum of the EC-PBGF-based Sagnac interferometer with a length of 50 cm and (b) the calculated group modal birefringence as a function of wavelength from the transmission spectrum.

Download Full Size | PDF

We also measured the strain and temperature sensitivities of the EC-PBGF-based Sagnac interferometer. The phase shift (∆ϕ) between two polarization modes in the EC-PBGF-based Sagnac interferometer can be written as [11, 19]

where B is the phase modal birefringence and L is the length of the PM-PBG exposed to the external perturbation (x). ∆N is the group modal birefringence. From Eq. (2), polarimetric sensitivity (K_x) to the external perturbation like temperature or strain can be derived as [11, 19]

where L_x is a relative fiber elongation due to the external perturbation. To measure the strain sensitivity (dλ/dℰ), we fixed one end of the investigated fiber and stretched the other end using a translation stage. Fig. 5(a) shows several measured transmission spectra of the EC-PBGF-based Sagnac interferometer for different strains. When the applied strain was changed from 0 to 1000 μℰ, the transmission peak shifted to a shorter wavelength as seen in Fig. 5(b). The strain sensitivity was measured to be -0.81 pm/μℰ for wavelength 1474.5 nm. Here, it is worth mentioning that the strain sensitivity is very much dependent upon the wavelength and is negative. It is related to the fact that axially applied tensional strain causes reduction of the fiber transversal dimensions and thus shifts the bandgap towards shorter wavelengths [11].

 

Fig. 5 (a) Transmission spectra of the EC-PBGF-based Sagnac interferometer with the applied strain change and (b) wavelength shift of transmission peak versus strains.

Download Full Size | PDF

We also investigated the polarimetric sensitivity to temperature. Figure 6 shows the transmission spectra of the EC-PBGF-based Sagnac interferometer at various temperatures. The peak wavelength shifted to a longer wavelength and the temperature sensitivity was estimated to be 3.97 pm/°C for wavelength 1474.5 nm. Note that temperature sensitivity of the sampled fiber is positive and decreases with wavelength in the longer wavelengths, crosses zero at 1413 nm, and vice versa for shorter wavelengths. It means that the group birefringence sensitivity to temperature (d∆N/dT) is positive in our EC-PBGF.

 

Fig. 6. (a) Transmission spectra of the EC-PBGF-based Sagnac interferometer with the applied temperature change and (b) wavelength shift of transmission peak as a function of applied temperature.

Download Full Size | PDF

4. Conclusion

We discussed the fabrication technique of an EC-PBGF by controlling asymmetric temperature characteristics in the hollow core region and also experimentally investigated the polarization properties of the EC-PBGF. It has a relatively large group modal birefringence at the wavelengths of both band edges (~3.5× 10^-3 at 1325 nm and 3.0× 10^-3 at 1480 nm). We also measured strain and temperature sensitivities of the EC-PBGF-based Sagnac interferometer. The results show that temperature sensitivity of the investigated fiber is positive and decreases with wavelength in the longer wavelengths, becomes zero at zero birefringence wavelength (~1413 nm), and vice versa for shorter wavelengths. Meanwhile, the strain sensitivity is very much dependent upon the wavelength and is negative over the bandgap region.

To the best of our knowledge, this is the first experimental demonstration of EC-PBGF based on Sagnac interferometer to sensor of physical parameters such as strain and temperature. We believe that the proposed EC-PBGF is very useful for applications to fiberoptic sensors and fiber-based devices.