A structured sapphire-derived all-glass optical fiber with an aluminum content in the core of up to 50 mol% was used for fiber Bragg grating inscription. The fiber provided a parabolic refractive index profile. Fiber Bragg gratings were inscribed by means of femtosecond-laser pulses with a wavelength of 400 nm in combination with a two-beam phase mask interferometer. Heating experiments demonstrated the stability of the gratings for temperatures up to 950°C for more than 24 h without degradation in reflectivity.
© 2014 Optical Society of America
Fiber Bragg gratings (FBGs) are well applicable as sensors for harsh environments. Especially the measurement of high temperatures of 800°C and higher is a challenging field of application. FBGs which can resist such high temperatures represent an important field of current research [1–4]. There exist different approaches to achieve highly temperature- resistant FBGs.
One possibility is the inscription of Type-II gratings, which are based on the local destruction of the glass matrix especially at the core-cladding interface . The resulting stress inside the fiber creates a change in refractive index, one of the basic requirements for a FBG. Type-II gratings typically show broad and strongly structured spectra [5,6].
Another option to increase the temperature stability of FBGs in silica fibers are so-called regenerated FBGs (rFGB) [7–9]. They are observed if a high-reflecting seed grating undergoes a special thermal treatment. In this case the seed grating can be (partially) bleached out before a regeneration of a more stable grating occurs. These rFBGs typically have very smooth spectra and can resist temperatures up to their regeneration temperature. With these gratings, temperature measurements up to 1295°C have been demonstrated . The fundamental mechanisms of the regeneration process are still under discussion and research. However, such rFBGs require strong seed gratings, mostly with the presence of hydrogen in the fiber, and a special thermal treatment process, and they typically provide only low reflectivity.
The limit for the highest possible application temperature of a grating is given by the softening point of the fiber material. For silica the softening point is around 1200°C. Higher temperatures may be achieved with different fiber materials. Single crystalline sapphire fibers are a very attractive candidate and have been shown to be applicable as a grating sensor for up to 1850°C [11–14]. But such sapphire fibers have limitations, since they are large-core air-clad fibers with strong multimodal properties resulting in typical spectral grating widths of 8 nm [12,13]. In addition, due to the missing fiber cladding, the guiding properties are very sensitive to the environmental conditions and to surface defects.
Therefore, a structured sapphire-type fiber with smaller core size would be of great interest. In 2012 a sapphire-derived fiber was reported as a structured all-glass fiber with an aluminosilicate core . This fiber was originally investigated for its very small Brillion gain coefficient. However, such a fiber with an aluminosilicate core also offers the possibility of high-temperature-stable FBGs without a regeneration process in a fiber with a core-cladding structure. The inscription of gratings into this special kind of fiber with an aluminum oxide content of 30 mol% was reported, but with degradation already at 700°C . The lower temperature stability may be related to the lower aluminum content of the core or to different grating inscription conditions using a combination of phase mask inscription technique and 800 nm femtosecond pulses.
We have prepared such a fiber with an aluminosilicate core and have successfully inscribed temperature-stable FBGs. The inscription was performed with a femtosecond (fs)-laser system at 400 nm inscription wavelength in combination with a two-beam phase mask interferometer. In this report we describe the fabrication and characterization of such a fiber with inscribed FBGs and discuss the specific characteristics of FBGs with an analysis of their temperature properties.
2. Fabrication and characterization of the fiber
For fabrication of the fiber, a crystalline sapphire rod with a diameter of 2.8 mm was stacked into a pure fused-silica tube with an inner diameter of 3.0 mm and an outer diameter of 30.0 mm. The stacked preform was heated up to 2200°C and subsequently drawn to fibers with different diameters. For FBG inscription, a fiber with a core diameter of 21.0 µm and an outer diameter of 125 µm was used.
The chemical composition of the fiber cross section was measured with an electron microprobe analyzer. The aluminum content of the fiber shows a graded parabolic-like profile with a maximum aluminum content of 49.4 mol% in the center of the core (see Fig. 1). This profile can be explained by diffusion and convective mixing between the molten aluminum oxide core and silica cladding during the fiber drawing process. In this way, an aluminosilicate core was created with a continuous transition between pure silica in the cladding and a high aluminum content glass in the center of the core. Further investigations with an electron backscattering technique ruled out the existence of any crystalline phase inside the core.
With an estimation of 2.2x10−3 for refractive index change per mol% aluminum oxide [15,17], the maximum of the absolute refractive index change of the core can be estimated to be about 0.11 above the silica index (1.46). This empirical estimation was confirmed by a measurement of the fiber’s refractive index profile (see Fig. 1), resulting in a proper core center refractive index of 1.56. The theoretical numerical aperture according to gives a maximum NA = 0.54 for the center of the graded-index fiber under investigation. As a consequence, this high numerical aperture has a multimodal light guidance behavior, which has to be considered for the evaluation of the FBG reflection spectra.
The loss of the fiber was measured by the common cut-back method. Comparing the transmission spectra for fiber samples with lengths of 30, 10, 3 and 1 m, a fiber attenuation of 0.370 dB/m at 1550 nm was estimated, which is well applicable for short fiber sensor lengths (see Fig. 2). The additional absorption peak around 550 nm is due to impurities of the initial crystalline sapphire rod.
3. Inscription and characterization of fiber Bragg gratings
For inscription of the FBGs a flexible two-beam phase mask interferometer was applied . Within this interferometer, the incoming laser pulses are split by a phase mask into the first two diffraction orders. These split pulses are reflected individually at mirrors and overlap again creating an interference pattern, which is directly inscribed into the fiber. By turning the mirrors, the interference angle between the two split pulses can be changed, which affects the period of the interference pattern and therefore the resulting Bragg wavelength. Using this interferometer the gratings were inscribed with the help of a phase mask for C-band gratings in a standard fiber with a period of 1066 nm. Due to the high refractive index, the typical reflection wavelength for the aluminosilicate fiber would then be around 1662 nm, which was corrected by turning the interferometer mirrors for Bragg wavelengths around 1520 nm. An additional cylindrical lens with a focal length of 235 mm was used in front of the interferometer to vertically focus the incoming pulses. With this focusing setup, a horizontal focus line at the position of the fiber was achieved in order to maximize the local intensity.
As a light source for grating inscription, we used the frequency-doubled wave of a Ti:Sa-amplified femtosecond laser system with pulse lengths in the order of 300 fs. The pulses with a wavelength of 400 nm are provided with a repetition rate of 1 kHz. The optimum averaged power was found to be 560 mW, being a compromise between high power for high reflectivity gratings and keeping intensities away from the destruction threshold of the fiber. These parameters are very similar to the parameters reported for grating inscription into crystalline sapphire fibers .
For interrogation of the gratings, we used a multimodal interrogation setup like that used for crystalline sapphire FBG interrogation [12,13]. All components (like interrogator, circulator and supply fiber) use 50-micron graded index fibers to achieve a stable excitation of a large number of modes inside the fiber under test and to allow optimized collection of the gratings’ reflected intensity for measurement. The resulting spectrum at room temperature is shown in Fig. 3(a).
The spectrum shows several grating reflection peaks corresponding to the multimodal guiding behavior of the fiber. The first six peaks starting from the long wavelength side are indicated by P1 to P6. The spectral distance between the peaks is approximately constant with a value of about 2.5 nm, and the spectral peak width is about (0.5-1.5) nm. This wide spectral range can be explained by reflection of mode groups of similar effective refractive indices which overlap in the reflection spectrum. The peak wavelength positions of the Bragg wavelength can be derived according to the Bragg condition:
To describe the spectral behavior also theoretically, we calculated the guided modes and their effective refractive indices, using the measured refractive index profile of the fiber. Both the analytical scalar transfer method based on  and the numerical finite element method mode solver (COMSOL®) were applied. Both methods describe the mode behavior as shown in Fig. 3(b). Due to the parabolic index profile, mode groups are found with several modes of almost the same effective refractive index. These mode groups have an almost equidistant spectral position for the first groups. With increasing mode number the difference of refractive index between two consecutive modes decreases, resulting in a continuum-like behavior for higher modes above mode group 8.
Due to the facts that higher modes have higher losses, induced, e.g., by bending and scattering, and that lower-order modes are easier to excite and have better overlap with the grating, not all mode groups are observed in the spectrum. The lower mode groups are expected to produce peaks with higher intensities at the long wavelength side of the reflection signal due to their higher effective refractive indices.
Reflection peaks may be achieved, e.g., by coupling of modes of the same mode groups in forward and backward direction (e.g. 1-1 for P1, 2-2 for P3, 3-3 for P5, etc.). Additional peaks may also be created by coupling of modes originated from different mode groups (e.g. 1-2 and 2-1 for P2; 2-3 and 3-2 for P4; etc.). Such peaks may overlap if the sum of the effective indices of the modes originating from different mode groups is the same as for modes of a single mode group (e.g. 1-3, 3-1 and 2-2 all P3), thus increasing the reflection signal. This effect of higher reflection intensities for reflection peaks at lower wavelengths is offset by the higher attenuation and losses for higher mode groups. Therefore a maximum of reflected intensities is observed, e.g., around P3, P4 and P5 in Fig. 3(a).
For simulation of the Bragg spectrum, also a decreasing amplitude towards higher modes (therefore lower wavelength) was assumed as well as a decreasing coupling efficiency for higher mode group order distance. The reflection spectrum was then modelled as an additive composition of single Gaussian-shaped reflection profiles for each combination of individual modes. The simulation result shown in Fig. 3(a) is well in accordance with the measured reflection peaks.
4. Heating experiments
The fiber gratings were investigated with regard to their temperature behavior. The fibers with the gratings were put into a thin fused silica capillary to protect the fibers from mechanical impacts and to place them vertically in the furnace. The temperature of the furnace was controlled by means of a thermocouple.
For temperatures below 700°C, the spectral structure remains almost stable except for the temperature related shift of the reflection peaks (see Fig. 4). But during the heating process some peaks, especially the second peak, became weaker in reflectivity due to slight changes in the mode coupling conditions within the heated grating compared to the initial grating. Therefore, the second peak was not used for further evaluation of the temperature sensitivity.
The difference in amplitude for both spectra in Fig. 4 is related to the spectral intensity distribution of the light source used. No change in intensity was observed while the temperature of almost 700°C was held for six hours. During long term experiments an intensity fluctuation within 5% of the light source occurred, resulting in slight homogeneous fluctuations for all peaks. A significant increase of the background due to the blackbody radiation was not observed, because the influence of this effect is negligible for temperatures below 1400°C .
Afterwards the fiber was heated to 900°C, and this temperature was kept for 29 h. The Bragg wavelength shows a permanent shift of about 0.5 nm – 0.8 nm in the first four hours after reaching the 900°C (see Fig. 5(a)). This additional wavelength shift remains also after cooling the fiber back to room temperature. This permanent shift indicates a change in the refractive index profile of the optical fiber and may be caused by relaxation effects within the fiber. The reflected peak intensities remained almost constant within the whole measurement time frame of up to 29 h (see Fig. 5(b)).
If the FBG is heated stepwise from room temperature to 900°C, the characteristic Bragg wavelength shift can be observed (see Fig. 6). All evaluated peaks show a parabolic dependency for the Bragg wavelength from temperature. The single measurement error was always below 72 pm and typically below 35 pm. Within the experimental error margins, the temperature sensitivity was the same for all peaks and also for heating only to 700°C before stabilization of the grating at 900°C. The mean linear coefficient was found to be (12.4 ± 0.2) pm/K and the quadratic coefficient to be (3.2 ± 0.2)x10−3 pm/K2. These values are slightly higher than the coefficients for standard fibers (SMF-28) but different from typical parameters of crystalline sapphire fibers . For the evaluation of the repeatability a grating was cycled between 400°C and 500°C and the Bragg wavelength was determined for each heating and cooling step after thermal stabilization to the target temperature. The repeatability was best for P1 with 49 pm and increased with the peak number to maximal 370 pm for P5.
A further increase in temperature to 950°C showed a higher permanent Bragg wavelength shift of 0.9 nm up to 1.8 nm with no indication of saturation within the measurement time of 23 h (see Fig. 7(a)). The amplitude of the reflection signal shows no relevant change during the heating process (see Fig. 7(b)). Slight fluctuations are related to spectral instabilities of the light source and instabilities in mode coupling.
Heating a grating up to 1000°C leads to a permanent change in the grating spectrum (see Fig. 8(a)); also, a decrease in reflectivity is now observed (see Fig. 8(b)). Especially short wavelength peaks start to vanish and merge into broadband reflection peaks. Due to the strong change in the spectrum’s shape, the measurement of a permanent wavelength shift was not possible. With the fiber kept at a constant temperature of 1000°C, the amplitude of every individual peak was decreasing by 20% within 110 minutes.
Both - decrease in reflectivity and change in spectral shape - indicate that at 1000°C the fiber becomes chemically unstable, due to the large diffusion affinity between the aluminum and silicon ions . Diffusion processes have their largest effect at the core-to-cladding interface. Changes within this fiber region influence mostly higher order modes which have a greater field overlap within this area. Furthermore, a phase separation could be possible, which would have direct influences on the losses of the fiber under test.
We have shown that inscription of gratings within a structured aluminosilicate-core fiber with an aluminum content of 50 mol% is possible. Such a structured fiber provides great advantages compared to the single crystalline fibers without cladding. The reflection spectra of the FBG show a multiple-peak behavior with an almost constant spectral distance of 2.5 nm between mode groups due to the graded parabolic-like refractive index profile of the fiber. The structured fiber concept might be apt for further development for few-mode or even single-mode propagation properties.
Heating the fiber shows that all the individual peaks have the same parabolic wavelength dependence on temperature. The gratings are stable up to temperatures of 700°C and higher. Fiber heating to 900°C leads to a permanent wavelength shift of about 1 nm within the first four hours, and higher wavelength shifts are observed for temperatures beyond 900°C. At 1000°C the reflectivity starts to decrease, but 80% of the initial reflectivity is preserved within the first two hours. For a temperature sensor, either individual spectral peaks could be used or a correlation technique could be applied taking into account the full spectral structure and hence improve the measurement accuracy. Compared to regenerated fiber Bragg gratings in all-silica fibers, the aluminosilicate core allows considerably higher grating reflectivity and better stability at higher temperatures.
The authors would like to thank the fiber technology group (Jens Kobelke, Katrin Wondraczek, Anne Ludwig) for drawing the fiber, the passive fiber module group (Martin Becker, Alexander Hartung) for discussions, the fiber sensor group (Albrecht Graf) for preparing the fiber ferrules, and the spectroscopy and imaging group (Andy Scheffel) for measurements using the EMPA and EBSD. In particular, John Ballato is acknowledged for the helpful discussions.
References and links
1. A. Azhari, R. Liang, and E. Toyserkani, “A novel fibre Bragg grating sensor packaging design for ultra-high temperature sensing in harsh environment,” Meas. Sci. Technol. 25(7), 075104 (2014). [CrossRef]
2. S. Gao, J. Canning, and K. Cook, “Ultra-high temperature chirped fiber Bragg gratings produced by gradient stretching of viscoelastic silica,” Opt. Lett. 38(24), 5397–5400 (2013). [CrossRef] [PubMed]
3. G. Laffont, R. Cotillard, and P. Ferdinand, “Multiplexed regenerated fiber Bragg gratings for high-temperature measurement,” Meas. Sci. Technol. 24(9), 094010 (2013). [CrossRef]
4. S. S. Chong, W. Y. Chong, S. W. Harun, and H. Ahmad, “Regenerated fibre Bragg grating fabricated on high germanium concentration photosensitive fibre for sensing at high temperature,” Opt. Laser Technol. 44(4), 821–824 (2012). [CrossRef]
5. J.-L. Archambault, L. Reekie, and P. St. J. Russell, “100% reflectivity Bragg reflectors produced in optical fibres by single excimer laser pulses,” Electron. Lett. 29(5), 453–455 (1993). [CrossRef]
6. C. Chojetzki, M. Rothhardt, J. Ommer, S. Unger, K. Schuster, and H.-R. Mueller, “High-reflectivity draw-tower fiber Bragg gratings – arrays and single gratings of type II,” Opt. Eng. 44(6), 060503 (2005). [CrossRef]
7. S. Bandyopadhyay, J. Canning, M. Stevenson, and K. Cook, “Ultrahigh-temperature regenerated gratings in boron-codoped germanosilicate optical fiber using 193 nm,” Opt. Lett. 33(16), 1917–1919 (2008). [CrossRef] [PubMed]
8. E. Lindner, C. Chojetzki, S. Brückner, M. Becker, M. Rothhardt, and H. Bartelt, “Thermal regeneration of fiber Bragg gratings in photosensitive fibers,” Opt. Express 17(15), 12523–12531 (2009). [CrossRef] [PubMed]
9. S. Bandyopadhyay, J. Canning, P. Biswas, M. Stevenson, and K. Dasgupta, “A study of regenerated gratings produced in germanosilicate fibers by high temperature annealing,” Opt. Express 19(2), 1198–1206 (2011). [CrossRef] [PubMed]
10. J. Canning, M. Stevenson, S. Bandyopadhyay, and K. Cook, “Extreme silica optical fibre gratings,” Sensors (Basel Switzerland) 8(10), 6448–6452 (2008). [CrossRef]
11. D. Grobnic, S. Mihailov, C. Smelser, and H. Ding, “Sapphire fiber Bragg grating sensor made using femtosecond laser radiation for ultrahigh temperature applications,” IEEE Photon. Technol. Lett. 16(11), 2505–2507 (2004). [CrossRef]
12. T. Elsmann, T. Habisreuther, A. Graf, M. Rothhardt, and H. Bartelt, “Inscription of first-order sapphire Bragg gratings using 400 nm femtosecond laser radiation,” Opt. Express 21(4), 4591–4597 (2013). [CrossRef] [PubMed]
13. M. Busch, W. Ecke, I. Latka, D. Fischer, R. Willsch, and H. Bartelt, “Inscription and characterization of Bragg gratings in single-crystal sapphire optical fibres for high-temperature sensor applications,” Meas. Sci. Technol. 20(11), 115301 (2009). [CrossRef]
14. T. Elsmann, T. Habisreuther, M. W. Rothhardt, and H. Bartelt, “High temperature sensing with fiber Bragg gratings in sapphire fibers,” in Advanced Photonics, OSA Technical Digest Series (Optical Society of America, 2014), paper BTu5B.2.
15. P. Dragic, T. Hawkins, P. Foy, S. Morris, and J. Ballato, “Sapphire-derived all-glass optical fibres,” Nat. Photonics 6(9), 629–635 (2012). [CrossRef]
16. D. Grobnic, S. J. Mihailov, J. Ballato, and P. Dragic, “Bragg gratings made with IR femtosecond radiation in high alumina content aluminosilicate optical fibers,”, ” in Advanced Photonics, OSA Technical Digest Series (Optical Society of America, 2014), paper BW2D.4.
17. S. Unger, J. Dellith, A. Scheffel, and J. Kirchhof, “Diffusion in Yb2O3-Al2O3-SiO2 glass,” Phys.Cham. Glasses: Eur. J. Glass. Sci. Technol. B 52(2), 41–46 (2011).
18. M. Becker, J. Bergmann, S. Brückner, M. Franke, E. Lindner, M. W. Rothhardt, and H. Bartelt, “Fiber Bragg grating inscription combining DUV sub-picosecond laser pulses and two-beam interferometry,” Opt. Express 16(23), 19169–19178 (2008). [CrossRef] [PubMed]
19. T. Grujic, B. T. Kuhlmey, C. M. de Sterke, and C. G. Poulton, “Modelling of photonic crystal fiber based on layered inclusions,” J. Opt. Soc. Am. B 26(10), 1852–1861 (2009). [CrossRef]
20. P. D. Dragic, C. Kucera, J. Ballato, D. Litzkendorf, J. Dellith, and K. Schuster, “Brillouin scattering properties of lanthano-aluminosilicate optical fiber,” Appl. Opt. 53(25), 5660–5671 (2014). [CrossRef]