Etching Bragg gratings in Panda fibers for the temperature-independent refractive index sensing

We demonstrate the evolution of the Bragg gratings inscribed in Panda fibers with chemical etching. The resonance wavelengths can blueshift with cladding reduction similar to the conventional counterparts. But the wavelength separation between the two polarizations is codetermined by the stress and the asymmetric shape effects. The fast and slow axes of the fiber can be reversed with each other and zero birefringence can be achieved by chemical etching the structure. When the stress-applying parts of the fiber are removed, the finalizing grating can be exploited for the temperature-independent refractive index sensing, since the modes corresponding to the two polarizations exhibit the dissimilar responses to the external refractive index change but the same response to temperature. Our device is featured with easy achievement, spectral controllability, and relative robustness. ©2014 Optical Society of America OCIS codes: (050.2770) Gratings; (060.2370) Fiber optics sensors; (060.2270) Fiber characterization; (260.5430) Polarization; (230.1150) All-optical devices. References and links 1. M. G. Xu, The Measurement of Physical Fields Using Optical Fibres and Bragg Gratings (University of Southampton, 1995). 2. A. Teixeira, R. Nogueira, P. Andre, M. Lima, J. Pinto, and J. da Rocha, “Applications of Highly Birefringent Fibre Bragg Gratings,” in Proceedings of International Conference on Transparent Optical Networks (IEEE, 2004), pp. 69 – 72. 3. M. Wuilpart, C. Caucheteur, S. Bette, P. Megret, and M. Blondel, “Polarization properties of uniform fiber Bragg gratings written in highly birefringent fibers,” Opt. Commun. 247(4-6), 239–245 (2005). 4. E. Udd, “Review of multi-parameter fiber grating sensors,” Proc. SPIE 6770, 677002 (2007). 5. T. Tenderenda, M. Murawski, M. Szymanski, L. Szostkiewicz, M. Becker, M. Rothhardt, H. Bartelt, P. Mergo, K. Skorupski, P. Marc, L. R. Jaroszewicz, and T. Nasilowski, “Fiber Bragg grating inscription in few-mode highly birefringent microstructured fiber,” Opt. Lett. 38(13), 2224–2226 (2013). 6. K. O. Hill, B. Malo, F. Bilodeau, D. C. Johnson, and J. Albert, “Bragg gratings fabricated in monomode photosensitive optical fiber by UV exposure through a phase mask,” Appl. Phys. Lett. 62(10), 1035–1037 (1993). 7. O. Frazão, T. Martynkien, J. M. Baptista, J. L. Santos, W. Urbanczyk, and J. Wojcik, “Optical refractometer based on a birefringent Bragg grating written in an H-shaped fiber,” Opt. Lett. 34(1), 76–78 (2009). 8. Y. Ran, L. Jin, L. P. Sun, J. Li, and B. O. Guan, “Bragg gratings in rectangular microfiber for temperature independent refractive index sensing,” Opt. Lett. 37(13), 2649–2651 (2012). 9. K. Chah, D. Kinet, M. Wuilpart, P. Mégret, and C. Caucheteur, “Femtosecond-laser-induced highly birefringent Bragg gratings in standard optical fiber,” Opt. Lett. 38(4), 594–596 (2013). 10. T. Erdogan and V. Mizrahi, “Characterization of UV-induced birefringence in photosensitive Ge-doped silica optical fibers,” J. Opt. Soc. Am. B 11(10), 2100–2105 (1994). 11. R. Kashyap, Fiber Bragg Gratings (Elsevier Academic, 2010). 12. L. M. Tong, R. R. Gattass, J. B. Ashcom, S. L. He, J. Y. Lou, M. Y. Shen, I. Maxwell, and E. Mazur, “Subwavelength-diameter silica wires for low-loss optical wave guiding,” Nature 426(6968), 816–819 (2003). 13. K. Zhou, X. Chen, L. Zhang, and I. Bennion, “High-sensitivity optical chemsensor based on etched D-fibre Bragg gratings,” Electron. Lett. 40(4), 232–234 (2004). 14. L. Dong, J. L. Cruz, L. Reekie, and J. L. Archambault, “Tuning and chirping fiber Bragg gratings by deep etching,” IEEE Photon. Technol. Lett. 7(12), 1433–1435 (1995). 15. A. N. Chryssis, S. M. Lee, S. B. Lee, S. S. Saini, and M. Dagenais, “High sensitivity evanescent field fiber Bragg grating sensor,” IEEE Photon. Technol. Lett. 17(6), 1253–1255 (2005). #224235 $15.00 USD Received 1 Oct 2014; revised 24 Nov 2014; accepted 29 Nov 2014; published 17 Dec 2014 (C) 2014 OSA 29 Dec 2014 | Vol. 22, No. 26 | DOI:10.1364/OE.22.031917 | OPTICS EXPRESS 31917 16. B. N. Shivananju, M. Renilkumar, G. R. Prashanth, S. Asokan, and M. M. Varma, “Detection limit of etched fiber Bragg grating sensors,” J. Lightwave Technol. 31(14), 2441–2447 (2013). 17. G. Tsigaridas, D. Polyzos, A. Ioannou, M. Fakis, and P. Persephonis, “Theoretical and experimental study of refractive index sensors based on etched fiber Bragg gratings,” Sens. Actuators A Phys. 209, 9–15 (2014). 18. N. Chen, B. Yun, Y. Wang, and Y. Cui, “Theoretical and experimental study on etched fiber Bragg grating cladding mode resonances for ambient refractive index sensing,” J. Opt. Soc. Am. B 24(3), 439–445 (2007). 19. W. Liang, Y. Huang, Y. Xu, R. K. Lee, and A. Yariv, “Highly sensitive fiber Bragg grating refractive index sensors,” Appl. Phys. Lett. 86(15), 151122 (2005). 20. X. F. Huang, Z. M. Chen, L. Y. Shao, K. F. Cen, D. R. Sheng, J. Chen, and H. Zhou, “Design and characteristics of refractive index sensor based on thinned and microstructure fiber Bragg grating,” Appl. Opt. 47(4), 504–511 (2008). 21. J. X. Dai, M. H. Yang, X. B. Li, H. L. Liu, and X. L. Tong, “Magnetic field sensor based on magnetic fluid clad etched fiber Bragg grating,” Opt. Commun. 17(3), 210–213 (2011). 22. S. S. Saini, C. Stanford, S. M. Lee, J. Park, P. DeShong, W. E. Bentley, and M. Dagenais, “Monolayer detection of biochemical agents using etched-core fiber Bragg grating sensors,” IEEE Photon. Technol. Lett. 19(18), 1341–1343 (2007). 23. A. N. Chryssis, S. S. Saini, S. M. Lee, Y. Hyunmin, W. E. Bentley, and M. Dagenais, “Detecting hybridization of DNA by highly sensitive evanescent field etched core fiber Bragg grating sensors,” J. Select. Topics Quantum Electron. 11(4), 864–872 (2005). 24. J. Noda, K. Okamoto, and Y. Sasaki, “Polarization-maintaining fibers and their applications,” J. Lightwave Technol. 4(8), 1071–1089 (1986). 25. J. Kumar, R. Mahakud, O. Prakash, and S. K. Dixit, “Study on hydrofluoric acid-based clad etching and chemical sensing characteristics of fiber Bragg gratings of different reflectivity fabricated under different UV exposure times,” Opt. Eng. 52(5), 054402 (2013). 26. L. Rindorf and O. Bang, “Sensitivity of photonic crystal fiber grating sensors: biosensing, refractive index, strain, and temperature sensing,” J. Opt. Soc. Am. B 25(3), 310–324 (2008). 27. Y. Jung, S. Kim, D. Lee, and K. Oh, “Compact three segmented multimode fibre modal interferometer for high sensitivity refractive-index measurement,” Meas. Sci. Technol. 17(5), 1129–1133 (2006).

In particular, FBG structures have attracted considerable research attention in detection of refractive index (RI) for the chemical and biological applications, due to their advantages such as high measurement accuracy, multiplexing capability, and etc. However the modes are mainly confined in the fiber core but screened from the ambient medium.To improve the responsivity to external RI influence, one can enhance the evanescent field interaction of the fiber by heating-and-tapering [12], side-polishing [13], or chemical etching [14][15][16][17][18][19][20][21][22][23] the structure.Comparatively, chemical etching is appealing for producing relatively robust and stable fiber devices with reduced size in a convenient way, without relying on precision machines.A lot of works have been done to achieve highly RI sensitive FBGs with the chemical etching technique.For examples, Dong et al realized a chirped FBG using the influence of external RI on the etched fiber [14].Chryssis et al. measured a high RI sensitivity with the etched FBG [15].Shivananju et al. studied the limit of detection of the etched FBG sensor based on nominal measurement noise levels [16].Tsigaridas et al obtained analytic functions of RI sensitivity in cladding-thinned fibers [17].Chen et al studied the backward cladding-mode resonances in etched FBGs [18].Liang et al. fabricated RI sensitivity enhanced grating structures [19] and Huang et al. obtained a phase-shift grating with the chemical etching [20].Etched FBG sensors have been explored in the detection of external RIs [14][15][16][17][18][19][20], magnetic fields [21], bio-chemical agents [22], DNAs [23], and etc.However, those previous works are mainly focused on studying of conventional FBG counterparts with no special stress.In this paper, we demonstrate the evolution of FBGs inscribed in Panda fibers by chemical etching the structure.It is found that the fast and slow axes can be reversed and zero birefringence can be achieved, suggesting a precision control of fiber birefringence for polarization maintaining, polarization interference, and etc. in the optical communication and sensing areas.Our finalized structure can be exploited for the temperature-independent RI sensing, since the modes corresponding to the two polarizations exhibit the dissimilar responses to the external RI change but the identical response to temperature.

Evolution of transmission spectra with the chemical etching
Figure 1 illustrates the experimental setup for chemical etching FBGs inscribed in Panda fibers.As shown in the figure, fibers A and B are polarization maintaining but the fiber C can be not, as demonstrated later.In our experiment, the FBG is first fabricated and subsequently placed into an aqueous solution of HF for the cladding etching [14].As shown in Fig. 1, after a beam of light from a broad-band source (BBS) is injected into the etched FBG, the light power is split into the two beams corresponding to the orthogonal polarizations at a polarization beam splitter (PBS) and subsequently, they are received separately in an optical spectrum analyzer (OSA).The polarization controller (PC) is used to modify the polarization states of light.In our experiment, we have adopted the Panda fiber (manufactured by the Yangtze Company, China) that contains a germanium-doped core with a diameter of 6.0μm and two boron-doped stress-applying parts (SAPs) with the diameter of ~33μm, respectively, as indicated by the insets shown in Fig. 1.Owing to the higher thermal-expansion coefficient of the SAPs compared with the cladding material, an obvious birefringence with the order of 10 −4 can be achieved after the fiber's fabrication [24].We fabricate a Bragg grating along the core with the assistance of the conventional 193-nm excimer laser scanning and phase mask inscription technique [6].The laser output energy is ~3mJ/pulse, the repetition rate is ~200Hz, and the grating period is set as 534.42nm, respectively.For a general birefringent FBG, two resonance wavelengths corresponding to the different polarizations are generated [1,11].The grating resonance aspects are mainly governed by the phase-matching condition [11], i.e., 2n eff i Λ = λ res i , where Λ is the grating period, λ res i is the resonance wavelength, n eff i = β i λ i /(2π) is the effective index, β i is the mode propagation constant, and i = x or y stands for the x or y polarization, respectively.By taking a small variation of the effective index from the above condition, after some mathematic treatments, the wavelength separation Δλ between the differential polarizations can be expressed as  )/2 is the average mode index between the polarizations, respectively.From Eq. ( 1), it is clear that the magnitude of Δλ is intimately dependent on birefringence B. The higher B value normally generates the larger separation Δλ.The mode along the fast axis has a shorter resonance wavelength than that along the slow axis.
The fiber structure can be etched within an HF acid via the chemical reaction between the HF and the fiber material [25].For the stress-applied Panda fiber, the SAPs correspond to a much faster etching speed due to the lower binding energy compared with the silica cladding, leading to the formation of asymmetric fiber shapes as shown in Fig. 1.The etched structure is characterized with parameters a and b, which are defined as the fiber scales along the x and the y directions, respectively.For the fiber with cladding partly etched and SAPs unetched, as shown in fiber A in Fig. 1, the birefringence is principally due to the existence of inner stress as previously demonstrated.But as SAPs are gradually etched, as shown in fibers B and C in Fig. 1, the circular structural symmetry of the fiber is broken and the birefringence may be generated due to the asymmetric geometric shape.It is necessary to measure how the grating spectra evolve with the cladding etching.Figure 2(a) plots the evolution of the transmission spectra of the grating and Fig. 2(b) details the resonance wavelengths and the separation Δλ between the resonance wavelengths corresponding to the two polarizations with respect to the fiber sizes.The used HF concentration is ~15%.It is shown that the grating resonance wavelengths can normally blueshift with the cladding reduction mainly due to the decrease of mode effective index, similar to the behavior of conventional counterparts [14].The mode along the y polarization has a faster wavelength shift compared with that along the x polarization, since the SAPs are mainly distributed in the y direction from Fig. 1.The small fluctuation in the depths of resonant dips may be attributed to the variation of the coupling strength between resonant modes [11].As shown in Fig. 2 Within region A, where the birefringence is mainly induced by inner stress, the absolute value of Δλ decreases gradually with the cladding reduction with an average rate of around 0.008nm/μm, due to the decrease of the stress in the core.The measured maximal birefringence is |B| = 5.67 × 10 −4 , which occurs at a = b = 125μm.Within region B, the fiber size along the y direction has a much faster decreasing speed with chemical etching compared to that along the x direction.As the SAPs are partially etched away, the initial increase in resonance wavelengths may be due to the change of stress in the fiber core or the generation of heat caused by chemical reaction of HF with the fiber material [25].We can show that the fast axis with relation to the stress is along the y direction, but that with relation to the asymmetric shape is along the x direction, respectively.With the thinning of the cladding, the stress decreases but the structural asymmetry increase gradually.As a result, the value of |Δλ| may tend to zero at a certain etched size.From Fig. 2(b), zero birefringence occurs around a ~84.2µm and b~30.2µm.Within region C, where the SAPs are completely removed, the stress-induced birefringence ceases but the asymmetric shape-induced birefringence gets larger continuously with etching.As shown in Fig. 2(b), the resonance wavelengths blueshift dramatically due to the lowering of mode effective index.The wavelength separation Δλ between the two polarizations becomes wider with an average rate of around −0.064nm/μm, due to the increase of fiber asymmetry.The birefringence direction has been rotated by 90 degrees compared with that in region A. From Fig. 2(b), the estimated maximal birefringence is around B = 5.56 × 10 −4 at a~74.5μm and b~7.5μm, which is comparable to the maximal stress-applied one in region A. Further study shows that too thin structure size may increase the radiation loss of light and make the transmission spectra become chirped.

Temperature-independent refractive index sensing
After the SAPs are completely removed, a large fraction of mode energy may extend outside of the fiber and interacts with the liquid localized in the vicinity of fiber core, leading to improvement of RI sensitivity [14][15][16][17][18][19][20].Note that for a general FBG, the resonance wavelengths should be changed by temperature, which inevitably induces a large influence to RI measurement.Conventionally, such a problem is solved by introducing another waveguide that is only sensitive to temperature to compensate the temperature influence, which, nevertheless, complicates the system.For our present device, the temperature cross-sensitivity may be eliminated by monitoring the spectral separation between the two polarizations.In the following experiment, the sensing characteristics are investigated using an etched FBG with parameters of a ~74.0μm and b~7.0μm.The microscope image for the fiber cross section has been provided as shown in the inset of Fig. 1.The RI responsivity is measured by immersing our device into an aqueous solution of sucrose at the room temperature (25°C).The solution's RI is modified continuously from 1.333 to 1.443 by tuning the sucrose concentration.3(b) plots the wavelength shifts as functions of the external RI for the x and y polarizations, respectively.The increase of RI induces the blueshift of resonance wavelengths thanks to the improvement of mode effective index as similarity to the behavior in conventional counterparts [13].From Fig. 3(a), while the change of external RI has a small influence to the spectral shape, the power level decreases with an increase of RI due to the increasing of mode power loss.The spectral dependencies on RI for the two polarizations are obviously different.From Figs. 3(a) and (b), the y polarization corresponds to a higher RI sensitivity since the mode has a larger fraction of energy outside of the fiber, compared with that of the x polarization.Inset of Fig. 3(b) gives the separation between the resonance wavelengths corresponding to the two polarizations versus RI.Investigation shows that the Δλ value decreases almost linearly with an increase in RI, due to the decreasing of birefringence.By performing a linear fitting to the experimental data, we obtain an expression with a coefficient of determination of R 2 = 98.5%:Δλ = −1.492⋅RI+ 2.202.Thus the RI sensitivity is given by −1.492nm/RI-unit, that corresponds to a measurement precision of ~4 × 10 −3 RI-unit considering a resolution of 0.06nm for our OSA.
The temperature coefficient is characterized by placing our etched FBG into a tube oven.A thermocouple is placed close to the grating to calibrate the temperature.Figure 4(a) displays transmission spectra at different temperatures and Fig. 4(b) gives the wavelength shifts as functions of temperature for the x and y polarizations, respectively.The temperature dependencies for the two resonance wavelengths are measured to be ~10.7pm/°C,which is analogous to the FBG in the conventional single-mode fiber.The difference of dependencies between the two polarizations is smaller than ~0.1%.Since most mode energy resides in the fiber, the thermo-optic effect dominates the temperature coefficient and thus the resonance wavelengths exhibit the similar response.Inset of Fig. 4(b) presents the separation between the resonance wavelengths corresponding to the two polarizations versus temperature.Analysis shows that the dependency of Δλ on temperature is smaller than 0.012pm/°C.Within the temperature region of Fig. 4(b), the wavelength fluctuation due to the temperature crosssensitivity is much smaller than the resolution of 0.06nm for our OSA.Compared to the latest techniques, such as photonic crystal fiber [25] or multimode interference [26] techniques, our achieved temperature coefficient has been improved by two or more orders of magnitude.Note that the index of liquid may also be influenced by temperature in practice.However, to study how many part of the index change for the liquid is due to the temperature variation is out of the scope of this article.Our present structure has the advantages of easy achievement, spectral controllability, and robustness in contrast to the counterparts reported in the literature such as [7,8].

Conclusion
We demonstrate the evolution of the FBGs inscribed in Panda fibers with chemical etching.The resonance wavelengths can normally blueshift due to the cladding reduction similar to the behavior of conventional counterparts.But the wavelength separation between the two polarizations may decrease or increase, which is dependent on the stress and the asymmetric shape effects in the fiber.It is found that both the fast and slow axes can be reversed with each other and zero birefringence can be achieved by chemical etching the structure, suggesting a precision control of fiber birefringence for polarization maintaining, polarization interference, and etc. in the optical communication and sensing fields.The maximum asymmetric shape-induced birefringence up to B = 5.56 × 10 −4 is obtained.After the two SAPs are etched away, our finalizing grating can be used as a temperature-independent refractive index sensor, because the modes corresponding to the two polarizations exhibit the dissimilar responses to the external RI change but the same response to temperature.This device is featured with easy achievement, spectral controllability, and relative robustness.

Fig. 1 .
Fig. 1.Experimental setup for chemical etching the FBGs inscribed in Panda fibers, corresponding to the structures with cladding partly etched and SAPs unetched (fiber A), partially-etched (fiber B), and completely-removed (fiber C), respectively.Several microscope pictures of fiber cross sections are provided as insets.
Figure2(a) plots the evolution of the transmission spectra of the grating and Fig.2(b) details the resonance wavelengths and the separation Δλ between the resonance wavelengths corresponding to the two polarizations with respect to the fiber sizes.The used HF concentration is ~15%.It is shown that the grating resonance wavelengths can normally blueshift with the cladding reduction mainly due to the decrease of mode effective index, similar to the behavior of conventional counterparts[14].The mode along the y polarization has a faster wavelength shift compared with that along the x polarization, since the SAPs are mainly distributed in the y direction from Fig.1.The small fluctuation in the depths of resonant dips may be attributed to the variation of the coupling strength between resonant modes[11].As shown in Fig.2(b), the regions corresponding to the configurations A, B, and C shown in Fig. 1 are marked and both the sizes a and b in different regions are indicated.

Fig. 3 .
Fig. 3. (a) Transmission spectra at different RIs.(b) Wavelength shifts as functions of external RI.Inset is the separation between the resonance wavelengths corresponding to the two polarizations versus RI.

Figure 3 (
Figure3(a) plots transmission spectra at different RIs and Fig.3(b) plots the wavelength shifts as functions of the external RI for the x and y polarizations, respectively.The increase of RI induces the blueshift of resonance wavelengths thanks to the improvement of mode effective index as similarity to the behavior in conventional counterparts[13].From Fig.3(a), while the change of external RI has a small influence to the spectral shape, the power level decreases with an increase of RI due to the increasing of mode power loss.The spectral dependencies on RI for the two polarizations are obviously different.From Figs.3(a) and (b), the y polarization corresponds to a higher RI sensitivity since the mode has a larger fraction of energy outside of the fiber, compared with that of the x polarization.Inset of Fig.3(b)gives the separation between the resonance wavelengths corresponding to the two polarizations versus RI.Investigation shows that the Δλ value decreases almost linearly with an increase in RI, due to the decreasing of birefringence.By performing a linear fitting to the experimental data, we obtain an expression with a coefficient of determination of R 2 = 98.5%:Δλ = −1.492⋅RI+ 2.202.Thus the RI sensitivity is given by −1.492nm/RI-unit, that corresponds to a measurement precision of ~4 × 10 −3 RI-unit considering a resolution of 0.06nm for our OSA.The temperature coefficient is characterized by placing our etched FBG into a tube oven.A thermocouple is placed close to the grating to calibrate the temperature.Figure4(a)displays transmission spectra at different temperatures and Fig.4(b) gives the wavelength shifts as functions of temperature for the x and y polarizations, respectively.The temperature dependencies for the two resonance wavelengths are measured to be ~10.7pm/°C,which is analogous to the FBG in the conventional single-mode fiber.The difference of dependencies between the two polarizations is smaller than ~0.1%.Since most mode energy resides in the fiber, the thermo-optic effect dominates the temperature coefficient and thus the resonance wavelengths exhibit the similar response.Inset of Fig.4(b) presents the separation between the resonance wavelengths corresponding to the two polarizations versus temperature.Analysis shows that the dependency of Δλ on temperature is smaller than 0.012pm/°C.Within the temperature region of Fig.4(b), the wavelength fluctuation due to the temperature crosssensitivity is much smaller than the resolution of 0.06nm for our OSA.Compared to the latest techniques, such as photonic crystal fiber[25] or multimode interference[26] techniques, our achieved temperature coefficient has been improved by two or more orders of magnitude.Note that the index of liquid may also be influenced by temperature in practice.However, to study how many part of the index change for the liquid is due to the temperature variation is out of the scope of this article.Our present structure has the advantages of easy achievement, spectral controllability, and robustness in contrast to the counterparts reported in the literature such as[7,8].

Fig. 4 .
Fig. 4. (a) Transmission spectra at different temperatures.(b) Wavelength shifts as functions of temperature.Inset is the separation between the resonance wavelengths corresponding to the two polarizations versus temperature.