We report about a thermal regeneration of fiber Bragg gratings written in photosensitive fibers with nanosecond laser pulses. We observe a regenerative process in a highly photosensitive fiber without hydrogen loading which indicates a secondary grating growth in an optical fiber by thermal activation. This process is more temperature stable than the commonly known gratings produced by color center modifications. The writing conditions of such new type of gratings are investigated and the temperature behavior of these regenerated fiber Bragg gratings is analyzed. The application possibilities are in the field of high temperature sensor systems by making use of the combination of good spectral shape of a Type I grating with a Type II like temperature stability.
©2009 Optical Society of America
Fiber Bragg gratings are well known as cost effective and easily integratable optical components for telecommunication, laser and sensor technology. In sensor systems they are well established as temperature and strain sensor elements. The advantages of their small size, easy multiplexing and no electrical interference make them a strong competitor to conventional electrical and electronic sensor systems. For the application as a temperature sensor or their use in a high temperature environment the choice of the grating type and the type of refractive index modification technique is of great relevance. Commonly known Type I gratings result from color center changes and the related modification of the absorption spectrum as described by the Kramers Kronig relation [1,2] and also from compaction and densification [3,4]. The temperature behavior of this type of Bragg gratings was analyzed by Erdogan et. al. , who derived a model for calculating the thermal activated decrease of the index change. Gratings according to this model are only applicable up to maximum operating temperatures of 450°C. The more temperature stable Type II gratings (up to 1000°C ) are generated by using intensities near the damage threshold of the doped glass. This type of refractive index modification is also caused by stresses disappears near the transition temperature of the core glass. The disadvantage of a Type II grating is a bad spectral response of the reflected light which makes the temperature dependent analysis for sensor applications difficult. Spectral widths of more than 0.6 nm are typical and the spectrum is often very asymmetrical. Additionally, Type II gratings show a high loss in the lower wavelength range which makes them not very suitable for multiplexed sensor arrays . Another type of gratings called Type IIA are in between Type I and Type II gratings with respect to their temperature stability. Grating stabilities under temperatures up to more than 500 °C are reported  and these gratings are always connected with a negative refractive index change . This negative index change leads always to decay of grating reflectivity and also of the Bragg wavelength which could be observed during the writing process. Another method for generating temperature stable gratings was reported by Fokine  who uses the diffusion of dopants after UV exposure and thermal processing. This results in temperature stabilities up to more than 1000 °C. Recently, Canning et. al.  described another new type of refractive index modification which leads also to high temperature stable Bragg gratings by using 193 nm laser radiation for grating inscription and a secondary tempering process. The initial seed gratings were of Type I which were then processed at temperatures of approximately 950 °C. A secondary grating growth was observed which resulted in highly temperature stable gratings up to 1100°C and which were named as regenerated gratings. For such gratings boron-co doped germanosilicate fibers were used together with hydrogen loading. Without hydrogen loading there was no effect reported.
We have observed a similar type of regenerated index change, however by using nanosecond laser pulses at 248 nm for the primary grating in highly photosensitive germanium doped fibers without boron codoping and without the need of hydrogen loading for the regeneration process. The writing conditions and the fabrication processes of such types of gratings are being investigated and their properties are analyzed with respect to their potential for applications in the field of high temperature sensing.
2. Writing setup
For writing Bragg gratings several fabrication techniques giving different properties have been in use. We use a modified Talbot interferometer, first introduced by Dockney et. al.  in combination with an excimer KrF laser at 248 nm with pulse durations of 20 ns. This setup was used in combination with different writing laser sources [13,14] and in different Bragg wavelength ranges demonstrating a high flexibility in the writing parameter. The resulting Bragg wavelength λB can be adjusted continuously by varying the angle of the interferometer mirrors (Fig. 1a ).
The energy density of the writing laser at the position of the fiber can be adjusted by a movable cylindrical lens. The writing experiments are made with a highly germanium doped photosensitive fiber having an outer diameter of 125μm. The germanium content of the preform and of the final fiber core is calculated from the measured refractive index profile as 18 mol%. The fiber has a numerical aperture of 0.26 and a mode field diameter of (5.0 ± 0.5) μm at 1550nm. The coating was stripped with a clamp before writing a grating and no hydrogen loading was applied.
3. Grating inscription
The gratings have been inscribed with different energy densities in order to investigate the reflection properties of the regenerated gratings in dependence on the inscription conditions. With a variation of the distance between the lens and the fiber the energy density in the grating recording plane can be changed. By considering a cylindrical lens with a focal length f and focusing in z-y direction without changing the Bragg period (Fig. 1) one can approximate the energy density Ed at the position of the fiber by using simple trigonometric equations. The laser beam profile is assumed as a top-hat distribution with the dimensions of 20x7.5 mm2. A smaller section of 6x5 mm2 is selected for a more homogeneous energy density. This profile is modified after passing the lens according to the geometry of the interferometer. The path length distance for one arm of the interferometer between lens and fiber ages is then described by (Fig. 1):
For positions outside the focus area, we can use a ray optics approximation. The beam diameter as a function of the distance between lens and fiber d(ages) can be calculated from the beam diameter before the lens d0, the distance of the interferometer mirrors b, the diffraction angle of the used phase mask beam splitter α and the resulting Bragg angle ϑFBG) according to:
On this basis the energy density can be adjusted in the range of 60 mJ/cm2 up to more than 1000 mJ/cm2 by moving the cylindrical lens. We only consider here the sum of the calculated energy densities in both interferometer paths as an average value and disregard the interferometric effect giving a modulation in local intensity. We also disregard the additional refraction effect from the geometry of the fiber because since it is nearly constant over the whole range of focusing variations. The gratings are then characterized in reflection or transmission to determine the grating reflectivity R. With the measured reflectivity one can calculate the grating strength η by using the coupled mode theory which is proportional to the UV induced refractive index change :
4. Regenerative gratings process
Regenerative gratings are obtained by annealing the initially recorded gratings at defined high temperatures. Our annealing experiments were done in a temperature range of 700-750 °C in a Carbolite ELF/11/6B oven with an additional temperature control near the position of the fiber. The initial grating was typically generated in the highly Ge-doped fiber using 3000 laser pulses with an energy density of 330 mJ/cm2 and has a length of 5 mm. In Fig. 2 the evolution of the grating reflectivity and of the Bragg reflection wavelength during the writing process is shown. Because of the high photosensitivity the grating grows very fast to a reflectivity above 90% and is saturated already after about 2 min of exposure at a reflectivity above 99%. The grating is therefore strongly overexposed after an exposure time of about 5 minutes. The Bragg wavelength is continuously shifting to the red during the writing which indicates an increase of the average index change and is normal Type I behavior without any indication of negative index change like as observed in Type IIA gratings. The generated grating for the following discussion had a transmission loss of −23.5 dB, which corresponds to a reflectivity of 99.55%. The spectral width was measured to be 0.489 nm (FWHM) and its Bragg wavelength at room temperature was 1550.672 nm. This grating was then annealed at 700 °C.
The changes of the optical properties during the annealing process are shown in Fig. 3 . On the left axis with the black circles the grating strength (according to Eq. (5) and on the right axis with the red triangles the grating reflectivity is presented in dependence on the annealing time. The blue circles in Fig. 3 show for comparison a typical conventional grating strength behavior during annealing of a Type I grating by using the mentioned model by Erdogan and by modeling the decrease of a grating with the same starting properties.
For the specific grating under test we can observe at first a strong initial decrease of both parameters normalized index change and reflectivity as known from Type I gratings. Like in the model presented by Erdogan et. al.  the decay of UV induced refractive index can be approximated by a reciprocal power function (approaching zero for long annealing times). After the strong initial decrease, however, the grating strength increases again after about 10 minutes of the annealing process to reach a secondary maximum after about 50 minutes. This secondary growth of induced index leads to an increase of grating reflectivity by about 5% as shown in Fig. 3 (red triangles). The associated grating spectrum decreases at the same time in FWHM from 0.489 nm to approximately 0.200 nm. In Fig. 4 one can see the measurement of the spectral width and the Bragg wavelength during the annealing at 700 °C. The spectral width in Fig. 4 shows the same behavior like the grating reflectivity Fig. 3 which leads to a real increase of the index modulation. The big variation of the Bragg wavelength results from the variation of the used furnace. One can only use the envelop of the graph to make a qualitative analysis. The Bragg wavelength, which is proportional to the effective index in the grating region (Eq. (1), increases in the first minutes because of changing temperature from approximately 20 °C to 700 °C. After the increase a decreasing process starts which is comparable with the decreasing index modulation (Fig. 3) in the first 5 minutes. After the maximum of grating regeneration in Fig. 3 the grating reflectivity and so the index modulation decreases. Simultaneously, one can observe an increase of the average index which could not be explained by the Erdogan model. One can assume that the decrease of the index modulation after the maximum could explained by an increase of the average index change in the unexposed areas in the interference pattern. This leads to a decrease of index modulation and to an increase of average index. The reasons for such a refractive index change are currently unknown. On assumption could be a crystallization or diffusion process in the glass material. To analyze the effect more in detail we investigate the depending writing parameters in the next section.
5. Discussion of relevant grating inscription parameters
In order to understand conditions for achieving the regenerated grating type we may vary the parameters energy density and number of pulses during the initial fiber writing process. For analysis we investigate two grating parameters of the annealed gratings. The first parameter is a regeneration factor which corresponds to the difference between the second reflection maximum after regeneration and the first minimum of normalized reflectivity. The regeneration factor indicates the order of regeneration of the grating reflectivity. This factor is so proportional to the relative regenerated index modulation. The second parameter we derive from the measurements is the time for reaching the secondary maximum of the grating reflectivity. This gives us an understanding of the temporal behavior of the process.
In Fig. 5 we can observe a mostly linear dependence of the two parameters with energy density. All gratings were written using 3000 laser pulses at every specific energy density value. The initial absolute reflectivity of all gratings was similar and higher than 95% before annealing. In the range of higher energy densities above 800 mJ/cm2 the measurement values of time and regeneration factor spread more because the writing process was made near the damage threshold of the glass system. An indication for this limit is a bad spectral response and a Type II temperature behavior of gratings written around 1000 mJ/cm2. In the inset of Fig. 5 a typical annealing behavior of a grating written around 750 mJ/cm2 is shown.
The time to the secondary maximum is strongly depending on the temperature. Experiments at higher temperatures accelerated the regeneration of the grating while lower temperatures slow down the process.
In Fig. 6 the importance of the fluence (i.e. product of energy and number of pulses) for the initial grating inscription is demonstrated. For this investigation 4 gratings were generated with the same energy density (of 700 mJ/cm2) using different pulse numbers. These 4 gratings were also annealed at 700 °C and the regeneration factor is determined. Other gratings were inscribed with a variation of the energy density and keeping the pulse number constant. In the graph one can see clearly that the regeneration factor depends directly on the applied fluence. The grating with a fluence of 4200 J/cm2 gave a very strong regeneration factor achieving up to 80% of the initial reflectivity. We observed no regeneration process for fluences lower than 900 J/cm2.
6. Temperature stability
In this section we want to investigate the temperature stability of the regenerated gratings. For determining the temperature stability, several gratings were written again at an energy density of 700 mJ/cm2 with 3000 laser pulses and annealed at 700°C until reaching the maximum regeneration factor. The temperature experiments were made with a high stable temperature calibrator (stability ± 0.01 °C). These gratings were then processed in temperature cycles. In the first calibration cycle the gratings were heated from 50 °C to 500 °C and backwards in steps of 50 °C with measurement periods of 30 min at each step. After the first calibration cycle the gratings were heated at 550 °C for 72 h before doing a second calibration under the same conditions. In Fig. 7 the Bragg wavelength in dependence on the temperature for the first and second calibration cycle of the grating is shown.
The red triangles mark the first calibration cycle of the grating and in the inset sections at 250 °C and 500 °C of the measured values are shown with higher accuracy. The limit of the measurement accuracy is 0.001 nm due to the used spectrum analyzer. After 72 h of annealing at 550 °C no significant drift of the wavelength was observed and the second calibration (blue squares) was made under the same conditions. In comparison both curves show a very good agreement without hysteresis indicating a high temperature stability of the gratings. The fit of the curve in Fig. 6 shows normal behavior of the shift of the Bragg wavelength with temperature as known from typical gratings. In the lower temperature range from 50 °C to 200 °C one can approximate the temperature dependent wavelength shift by 11.5 pm/°C and from 200 °C up to 500 °C by approx. 15 pm/°C. This result shows that such type of grating is very well suitable for sensing in a temperature range up at least 550 °C and with good spectral reflection characteristics.
We have discussed thermal activated regeneration of fiber Bragg gratings written in non hydrogen loaded but highly photosensitive fibers. The processes responsible for the regeneration effect are not yet fully understood. Crystallization or diffusion effects, for example, could be expected to occur during high temperature annealing. The grating regeneration leads to a strong increase of the grating reflectivity after initial bleaching and gives more temperature stable gratings than the known Type I modifications. With their high temperature stability and their small and well-defined reflection spectrum they are perfect candidates for multiplexed sensor grating arrays up to at least more than 550 °C showing no drift or hysteresis.
Funding by the Thuringian Ministry of Education and Cultural Affairs is gratefully acknowledged. The authors would like to thank Dr. Kirchhof for fruitful discussions.
References and Links
2. P. St. J. Russell, L. J. Poyntz-Wright, and D. P. Hand, “Frequency doubling, absorption, and grating formation in glass fibers: effective defects or defective effects?” Proc. SPIE 1373, 126 (1991). [CrossRef]
3. H. G. Limberger, P. Y. Fonjallaz, R. P. Salathe, and F. Cochet, “Compaction- and photoelastic-induced index changes in fiber Bragg gratings,” Appl. Phys. Lett. 68(22), 3069 (1996). [CrossRef]
4. M. Douay, W. X. Xie, T. Taunay, P. Bernage, P. Niay, P. Cordier, B. Poumellec, L. Dong, J. F. Bayon, H. Poignant, and E. Delevaque, “Densification involved in the UV-based photosensitivity of silicaglasses and optical fibers,” J. Lightwave Technol. 15(8), 1329–1342 (1997). [CrossRef]
5. T. Erdogan, V. Mizrahi, P. J. Lemaire, and D. Monroe, “Decay of ultraviolet-induced fiber Bragg gratings,” J. Appl. Phys. 76(1), 73 (1994). [CrossRef]
6. 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]
7. C. Chojetzki, M. Rothhardt, S. Ommer, S. Unger, K. Schuster, and H. R. Mueller, “High-reflectivity draw-tower fiber Bragg grating arrays and single gratings of type II,” Opt. Eng. 44(6), 060503 (2005). [CrossRef]
9. L. Dong and W. F. Liu, “Thermal decay of fiber Bragg gratings of positive and negative index changes formed at 193 nm in a boron-codoped germanosilicate fiber,” Appl. Opt. 36(31), 8222–8226 (1997). [CrossRef]
10. M. Fokine, “Thermal stability of chemical composition gratings in fluorine-germanium-doped silica fibers,” Opt. Lett. 27(12), 1016–1018 (2002). [CrossRef]
11. 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]
12. M. L. Dockney, S. W. James, and R. P. Tatam, “Fibre Bragg gratings fabricated using a wavelength tuneable laser source and a phase mask based interferometer,” Meas. Sci. Technol. 7(4), 445–448 (1996). [CrossRef]
13. E. Lindner, M. Becker, M. Rothhardt, and H. Bartelt, “Generation and Characterization of First Order Fiber Bragg Gratings with Bragg Wavelengths in the Visible Spectral Range,” Opt. Commun. 281(18), 4612–4615 (2008). [CrossRef]
14. 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]