The properties of three sensitization processes: UV hypersensitization, OH-flooding and H2-loading have been investigated through Bragg grating (BG) inscription within standard germanosilicate fibers. More specifically, the stability of the sensitization processes and that of the UV-induced index changes have been investigated through isochronal annealing experiments. Moreover, the level of excess loss induced near 1.4 µm by both the sensitization process and the BG inscription has been estimated by means of in fiber absorption spectroscopy. The level of loss proves to be higher in the hypersensitized or OH-flooded fiber than in the H2-loaded counterpart when pulsed 248 nm light was used.
© 2005 Optical Society of America
Hydrogen loading at high pressure (>100 bars) and room temperature (≤ 50°C) is now routinely used to enhance the photosensitivity of silica-based optical waveguides and fibers . On the other hand, UV hypersensitization of H2-loaded silica glasses and OH-flooding treatments are efficient means of enhancing the photosensitivity without incurring some disadvantages associated with the H2-loading technique [2–6]. Indeed, on the one hand, hypersensitization by use of a short initial uniform UV pre-exposure on a H2-loaded germanosilicate fiber and subsequent out-diffusion of the hydrogen proved to be efficient for locking the fiber photosensitivity . Thus, strong BGs with a stable reflectivity can be written in UV hypersensitized glass without the need for the presence of molecular hydrogen at the time of the inscription [2–5]. Furthermore, the inscription of BGs in these hypersensitized fibers involves a reduced formation of hydrogen-related species when cw 244 nm was used . On the other hand, OH-flooding is a thermal sensitization process in which the temperature of hydrogenated fibers is elevated up to ≈1000 °C for a short time (a few seconds) . For heating time longer than ≈1s, the photosensitivity enhancement proves to be less efficient . Permanent sensitization is still effective after complete hydrogen out-diffusion. Interestingly, the technique proved its efficiency to enhance the photosensitivity of pure silica fiber  and planar germanosilicate waveguide . However some properties of the process (its stability for example) are not yet well documented. Furthermore, the interaction between H2 molecules and the glass at a high temperature results in the massive formation of hydroxyl species that absorb near 1.4 µm . Yet, no spectroscopic data dealing with the UV-induced evolution of these species at the time of the BG writing is currently available.
As these sensitization processes look attractive, directly comparing the three methods seems relevant for practical applications. Accordingly, the paper has two main purposes, firstly to compare the properties of gratings written in either UV hypersensitized, OH-flooded or H2-loaded fibers and secondly to compare the stability of the processes. To this end, the kinetics of grating growth have been recorded in the three types of sensitized fibers, and the stability of the reflectivity and the Bragg wavelength of the corresponding BGs have been investigated by means of step isochronal annealing. Hypersensitization, OH-flooding processes and BG inscription in sensitized fibers are known to induce excess loss ascribed to the formation of T-OH species (T=Ge or Si). Thus, in fiber absorption spectroscopy near 1.4 µm has been used with a view to estimating the level of excess loss that results from both the sensitization process and the BG inscription. Isochronal annealing of either hypersensitized or OH-flooded fibers followed by BG inscription has been carried out to determine how stable are the sensitization processes.
2.1 Fiber sensitization and Bragg grating inscription
Firstly, standard telecommunication optical fibers (Corning SMF 28) were H2-loaded (140 atm at room temperature for 1 month). Secondly, 20 mm long parts of the H2-loaded optical fibers were hypersensitized by means of an uniform exposure to a burst of Npre=40 000 UV pulses at 248 nm (Npre=number of UV pulses used to perform the UV hypersensitization process). Thirdly, 50 mm long parts of the H2-loaded optical fibers were OH-flooded by means of a short annealing at 950°C. For the photosensitivity enhancement to be at maximum, the annealing time was fixed to ≈1s. H2 out-gassing of the hypersensitized or OH-flooded fibers was then achieved at 110°C for 3 days.
Uniform BGs were written in the sensitized fibers by exposing a phase-mask (Lasiris, pitch=1061 nm) to UV pulses from a KrF laser at 248 nm. All the exposures (blanket exposure or exposure to a UV fringe pattern) were carried out at a mean fluence per pulse of ≈160mJ/cm2 and a frequency rate of ≈20 Hz. The refractive index modulation data were extracted from the BG reflectivity through an iterative method that allows for the variation of η (the fraction of the total optical power propagating along the core) with treatment and exposure time. The data corresponding to all the BGs are displayed in table 1.
2.2 BG stability
The values of Ni were chosen so that the UV-induced refractive index modulations (≈10-3) of all the gratings were nearly equal. The grating written in the H2-loaded fiber has then been stored at room temperature for 1 month to ensure complete out-diffusion of the hydrogen. Finally, the BGs were annealed through 30 min isochronal heating treatments. The investigated range of annealing temperature T spanned from 273 K up to 1273 K by step of 50 K. After the step during which the fiber was kept at T for 30 min in a furnace, the temperature of the fiber was rapidly reduced at room temperature to be in position to record the transmission grating spectra.
2.3 Excess loss near 1.4µm
After each step of the sensitization processes or after a BG inscription, the fiber absorption spectrum was recorded in the spectral range [1300nm–1600nm] using a white light source and an optical spectrum analyzer. The change in the fiber T-OH species-related absorption coefficient was estimated by measuring the evolution of the overtone absorption band at the peak near 1.4 µm that results from the sensitization process or the BG inscription.
2.4 Process stability
Parts of the stripped H2 loaded fibers were uniformly sensitized (UV hypersensitization or OH-flooding). After accelerated H2 out-gassing, the fibers were set in a tubular furnace to be annealed for 3h in air at increasing temperatures by step of 100°C up to 800°C. Then Bragg gratings (Fp=160±20 mJ/cm2) were written in the annealed fibers and the kinetics of BG growth could be compared to that before annealing.
Figure 1 shows the typical kinetics of BG growth in H2-loaded, UV hypersensitized, OH-flooded and pristine fibers. The method used to carry out the sensitization process is the parameter of the experiment. In this figure, the symbols are experimental data and the lines are a guide for eye. Figure 1 allows the efficiencies of the three methods for enhancing the photosensitivity of a standard fiber to light at 248 nm to be compared. Regardless of the sensitization process, the grating growths are monotonous as a function of the number Ni of pulses impinging onto the fiber. Yet, the initial growth of the BG written in the OH-flooded fiber proves to be faster when compared to the two other sensitization processes. Moreover, it is also interesting to compare the levels of index changes for longer UV exposure time. After exposure to a cumulated fluence of ≈ 8kJ/cm2, the H2 loading technique leads to a gain in the modulation by a factor of ≈2, when compared to the other methods.
Figure 2 shows the evolutions of the normalized refractive index modulation as a function of the temperature T at which the three gratings were elevated. The full squares are for the grating (G1) written in the H2-loaded fiber, the crosses are for the G2 grating (UV hypersensitized fiber) while the full circles (OH-flooded fiber) are for the G3 grating. As already reported in the literature [5, 10], Fig 2 shows that the stability of the modulation for BGs written in UV hypersensitized fiber is significantly higher than that for a BG written in the H2-loaded counterpart. For example, a normalized refractive index modulation above 0.1 can even be detected after the step of annealing of G2 at 1223 K whereas at this step NImod for G1 is 0.05. It is worth noticing that the thermal evolution of the quantity NImod(30min,T) corresponding to the G2 grating matches that for G3. As a result, the above-mentioned conclusion about the lower stability of BGs in the H2-loaded fiber proves to be also relevant for the gratings written in the OH-flooded fiber. After the step of annealing at T=973 K, the rates of decay follow similar trends for all the gratings.
As shown in Fig. 3, the irreversible shift experienced by the BG wavelengths (λB) as the temperature is made to increase, differs according to the method used to sensitize the fiber. In this figure, the symbols have the same meaning as those in Fig 2. Yet, the conclusions about the Bragg wavelength stability versus the sensitization process are quite different from those drawn for the stability of the reflectivity. Indeed, the larger shifts are observed for the OH-flooded fiber, followed by the hypersensitized fiber , the smaller shifts being measured in the H2-loaded fiber. This observation indicates that the sensitization (uniform pre-exposure or OH-flooding treatment) process-induced improvement in the NImod thermal stability was made at the expense of lower λB stability. Indeed, as the sensitization process leads to a large increase in the mean refractive index, the refractive index change at the gratings contain a larger mean dc-term (typically Δnmean=2.10-3 in UV hypersensitized fiber) than in H2-loaded sample. Consequently the annealing-induced shifts in the Bragg wavelength remain high. It is worth noticing that after the step at which the gratings were kept at 1223 K for 30 min, the annealing has nearly completely bleached the shift in the λB induced by the sensitization and the inscription (see Table 1 in which the initial shifts are displayed).
The evolutions of the stability of the gain in modulation that results from the use of either the hypersensitization or the OH-flooding process are shown in Fig. 4 as a function of the temperature at which the fiber was annealed (annealing performed before any BG inscription). The gain in modulation is obtained as the difference between the amplitude of the modulation that corresponds to a grating written in the pristine fiber from that in the hypersensitized (or OH-flooded) fiber (the number Ni of pulses used for writing the grating is fixed; Ni=30 000). The normalization factors are the gains in modulation measured before annealing in a sensitized fiber fiber. More specifically, the full squares correspond to the residual enhancements in modulation for the hypersensitization process while the full circles are for the gain that results from the OH-flooding technique. The evolution of these normalized gains in Δnmod looks similar although the stability of the sensitization by use of the OH-flooding process seems better than that of the hypersensitization. It is worth noticing that after the step of annealing at 1073 K a normalized gain in photosensitivity between 0.1 and 0.2 still could still be measured. This in turn implies that after this step, the photosensitivity of the sensitized fibers remains higher than that of the pristine fiber.
The evolutions of the total excess losses at peak near 1.4µm recorded in the course of a BG inscription are displayed in Fig. 5 as a function of the refractive index modulation. Firstly, the sensitization process (UV hypersensitized or OH-flooded fibers) leads to the formation of the OH species related band centered at 1400 nm. The intensity of this band rises noticeably with the grating writing time although no molecular hydrogen remains present. Secondly, these BG inscription-induced excess losses are lower in UV hypersensitized or OH-flooded fiber than in the H2-loaded fiber. In contrast, in the H2-loaded fibers, only the writing step leads to excess loss while in the others methods, the origin of excess loss are twofold: the sensitization process and the BG writing step. For example, the reference spectrum from a UV hypersensitized fiber has been recorded after the blanket exposure of the fiber to Npre=20000 pulses of UV light and an accelerated out-gassing (110 °C for 3 days). The UV-induced excess loss at the peak near 1.39µm was, at this time, 1.65dB/cm whereas it reached 2 dB/cm at the end of the BG inscription. In conclusion, Fig. 5 shows that for a fixed modulation index value (Δnmod<10-3) the overall level of excess loss near 1.4µm is lower in the H2-loaded fiber. This conclusion proves to be different from that reported in Ref 6 when cw 244nm laser light was used.
In conclusion, we have compared the behaviors of the spectral characteristics of gratings written in sensitized fibers. Although the growth of the BG in the OH-flooded fiber is faster at the beginning of the inscription than that in the hypersensitized fiber, the modulations in both fibers show a similar trend towards saturation after 40000 pulses. Eventually, the H2-loading technique leads to a gain in the photosensitivity by a factor ≈2, compared to the two other methods of sensitization. We have shown that the improvement in the stability of NImod observed below 1073 K observed for gratings written in UV hypersensitized or OH-flooded fibers is made at the expense of a lower stability of the Bragg wavelength. This observation can be explained from the large mean index created at the time of the fringeless pre-exposure or when elevating the temperature of the fiber to get the OH-flooding. This large mean index is at the root of higher annealing-induced shifts in the λB for sensitized fibers than those for H2-loaded fiber. We have also compared the thermal stability of the two sensitization processes. We have demonstrated that after 1073 K, the enhancement in photosensitivity induced by the OH-flooding technique (or) although residual remain higher than that induced by the UV hypersensitization process. Moreover, we have shown that at fixed modulation level (Δnmod<10-3), the total excess loss near 1.4 µm is lower for the grating written in the H2-loaded fiber.
From a practical point of view, our results show that using H2-loaded fibers remains the best technological solution for writing low loss, strong short period gratings with a superior Bragg wavelength stability. The stability of the modulation in these fibers can be similar to that for BG written in sensitized fibers provided that the BG in H2-loaded fiber can be post-annealed to remove the unstable part of the UV-induced index change. Nevertheless, the two sensitization processes prove to be useful each time locking the photosensitivity is strictly necessary.
This project was supported financially by the EEC PLATON contract (IST-2002-381668).
References and links
1. R. M. Atkins, P.J. Lemaire, T. Erdogan, and V. Mizrahi, “Mechanisms of enhanced UV photosensitivity via hydrogen loading in germanosilicate glasses,” Electron. Lett. 29, 1234–1235 (1993) [CrossRef]
2. J. Canning, “Photosensitisation and photostabilisation of laser induced index changes in optical fibres,” Opt. Fib. Tech. 6, 275–289 (2000) [CrossRef]
3. G.E. Kohnke, D. W. Nightingale, P. G. Wigley, and C. R. Pollock, “Photosensitization of optical fiber by UV exposure of hydrogen loaded fiber,” Optical Fiber Communication Conference (OFC’99), paper PD 20 (1999)
4. M. Äslund, J. Canning, and G. Yoffe, “Locking in photosensitivity in optical fibres and waveguides,” Opt. Lett. 24, 1826–1828 (1999)
5. M. Äslund and J. Canning, “Annealing properties of gratings written into UV-presensitised hydrogen out-diffused optical fibres,” Opt. Lett. 25, 692–694 (2000) [CrossRef]
6. J. Canning, M. Äslund, and P.F. Hu, “UV-induced absorption losses in hydrogen-loaded optical fibres and in presensitised optical fibres,” Opt. Lett. 25, 1621–1623 (2000) [CrossRef]
7. M. Fokine and W. Margulis, “Large increase in photosensitivity through massive hydroxyl formation,” Opt. Lett. 25, 302 (2000) [CrossRef]
8. J. Albert, M. Fokine, and W. Margulis, “Grating formation in pure silica-core fibres,” Opt. Lett. 27, 809 (2002) [CrossRef]
9. C. Riziotis, A. Fu, S. Watts, R. Williams, and P. G. R. Smith, “Rapid heat treatment for photosensitivity locking in deuterium-loaded planar optical waveguides,” Proceedings of Bragg Gratings, Photosensitivity and Poling in glass waveguides, Stresa, Italy, paper BThC31 (2001)
10. B. O. Guan, H. Y. Tam, X. M. Tao, and X. Y. Dong, “Highly stable fiber Bragg gratings written in hydrogen-loaded fiber,” IEEE Photon. Technol. Lett. 12, 1349–1351 (2000) [CrossRef]