The writing of ultrabroadband Fiber Bragg Gratings (FBGs) is demonstrated in both hydrogen-free and hydrogen-loaded standard telecom fibers by the use of IR femtosecond pulses and a highly chirped first-order phase-mask. A high reflectivity filter providing a wavelength coverage of five telecom bands (E+S+C+L+U) is demonstrated over a single 35mm long grating inscribed in only 30s in H2-loaded SMF-28 fiber. Refractive index modulation of about 2.5×10-3 and 5×10-3 are obtained after a few second exposure time in both hydrogen-free and hydrogen-loaded SMF28 fibers. This report paves the way to the development of new broadband fiber-based optical components such as multi-wavelengths filters and sources.
© 2009 Optical Society of America
The use of ultrashort pulse sources in conjunction with phase masks has been recently shown as a very promising alternative to the defect-resonant based UV-writing of fiber Bragg gratings (FBGs) in silica fibers [1,2]. The fs-induced FBG of good spectral quality with a reduced cladding mode loss are stable at elevated temperatures, where standard UV-written FBGs would be erased. The use of IR fs pulses with the first-order phase-mask was also shown as the only way until now to write efficient FBGs in both doped and undoped fluoride fibers . Such gratings have been used as reflectors for fiber lasers in both silica  and fluoride fibers . Linearly chirped FBGs are widely used for complex dispersion management , multiwavelength fiber lasers  and for all-fiber chirp pulse amplification scheme . Recently, a first successful inscription of a continuously chirped FBG into a nonphotosensitive fiber was reported by Thomas et al using IR fs pulses with a constant pitch phase mask and a bent fiber . The authors reported a maximum reflectivity of 50% with a bandwidth of 6 nm for a 20 mm long FBG.
In this letter, we report on a significant improvement in the fabrication of highly reflective and ultrabroadband FBGs based on the use of IR fs pulses and a strongly chirped first-order phase mask. Bandwidths (FWHM) of 85 nm and 310 nm with maximum reflectivity of 75% and 98.5% were obtained in the H2-free and H2-loaded SMF-28 fibers of 25 and 35 mm length, respectively, after only a few seconds of exposure time. The resulting highly chirped gratings (95 nm/cm) are precisely characterized over the C-band for their spectral and temporal responses using an optical vector analyzer.
A Ti-sapphire regenerative amplifier system (Coherent Legend-HE) that produced fs-laser pulses of 3.5 mJ per pulse at 1 kHz repetition rate with the central wavelength of λ=806 nm was used to write the FBGs. The temporal width of the Fourier-transform limited pulses was measured to be ~35 fs. The pulse width was enlarged to ~60 fs due to the group velocity dispersion when passing through the optical components. The laser beam was enlarged in one direction to ~8.5 mm × 43 mm size (at 1/e2). The beam was focused in the plane perpendicular to the fiber core by a cylindrical lens through a highly chirped silica phase mask. The width of the focal line is estimated by Gaussian beam optics as 1.27fλ/D ~14 μm, where f=112 mm is the focal length and D is the beam size at the focusing lens. The fs-laser exposure resulted in formation of a narrow filament of refractive index change of ~1 micron width . In order to ensure a uniform index modulation over the fiber core the focusing lens was vibrated using a piezoelectric stage in the direction perpendicular to the fiber, such that the focal line scanned over the fiber across a 20 μm area surrounding the fiber core with a frequency of 0.05Hz.
The linearly chirped phase mask used in the experiment had a high chirp rate of 65 nm/cm over a length of 4.8 cm. The phase-mask pitch ranges from 855 nm to 1165 nm which corresponds to a FBG spectrum range extending from 1240 nm to 1685 nm, assuming an effective index neff ~1.448. The phase mask was fabricated by holographic lithography on a UV-grade fused silica substrate. The zero order of diffraction at 800 nm for polarization parallel to the grooves is in the range of 15-35 %, depending on the position of the beam. However, since the zero order beam does not spatially overlap (inside the fiber) with the first order beams because of the group velocity walk-off effect  zero-order nullification is irrelevant except for the corresponding loss in energy. A pure two beam interference pattern can be obtained after a propagation of about 50–75 μm from the phase-mask. We put the fiber at 125 μm from the phase mask to ensure that the walk-off condition is respected over the entire pitch range of phase-mask.
Transmission and reflection spectra of the FBGs were measured using a fiber-based super-continuum white light source, a broadband (1400–1600 nm) optical circulator and an optical spectrum analyzer (AQ6375). The gray losses (GL) were measured using a cut-back method at 1700 nm, which was chosen to be not overlapped with the reflection band. The FBGs were also characterized by their Insertion loss (IL), group delay (GD) and polarization dependent losses (PDL) over the entire C band (1525–1570 nm) using a high resolution optical vector analyzer (Luna, OVA-CTe) of 3.2 pm resolution. A Gaussian filter of 100 pm width was used to smooth the data, remove the measurement noise and reveal the group delay ripples (GDR).
3. Results and discussion
We first exposed for 130 s a H2-free SMF-28 fiber over the grating length of 25 mm. The broadband reflection and transmission spectra are as shown in Fig. 1.
A maximum IL of -6 dB corresponding to the reflectivity of 75% is obtained in a range from 1535 nm to 1570 nm with the full-width half-maximum (FWHM) of 85 nm. The GLs were measured using a cut-back method to be smaller than 0.1dB. Figure 2 shows the reflection responses when measured from the short-pitch and the long-pitch sides of the grating, respectively, along with the GD measured from the short-pitch low-loss reflective side.
The cladding mode losses (CML) of the grating were obtained by comparing the reflection responses measured from the long-pitch and short-pitch sides of the FBG . The CML is relatively flat over the entire C band and is less than 0.2 dB as shown in Fig. 1(b). The reflectivity ripple is also measured to be less than ±0.3 dB. The linear GD response gives a constant group velocity dispersion (GVD) of 1.02 ps/nm, and GDR smaller than ±0.5 ps, which approaches the limit of the optical vector analyzer, which is ±0.1 ps. The PDL was also measured to have a mean value of 0.30 dB over the C band.
We exposed a H2-loaded SMF-28 fiber over a grating length of 25 mm for 20 s in the same setup. The broadband reflection and transmission spectra of the grating are shown in Fig. 3.
A maximum IL of -21dB was obtained at ~1550 nm. The FWHM of the spectrum was measured to be 206 nm covering the entire S+C+L telecom bands with a grating length of only 25 mm. The GLs were measured to be 0.4 dB using a cut-back method at 1700 nm. Figure 4 shows the reflection responses of the grating measured from the short and long-pitch sides, respectively along with the GD response measured for the short-pitch side.
The corresponding CML is relatively flat over the entire C band with a value of ~2.5 dB, which is significantly higher than that in the H2-free fiber. A maximum reflectivity of 98.5% was obtained at ~1550 nm with ripples smaller than ±0.15 dB. The linear GD response gives a constant GVD of 1.02 ps/nm with the GDR smaller than ±2.0 ps. The PDL was also measured to have a mean value of 0,09dB over the entire C band.
Finally, we exposed a H2-loaded SMF-28 fiber over a grating length of 35 mm for 30s using the same setup in order to obtain an even larger bandwidth. Figure 5 shows the reflection spectrum of the resulting grating along with the corresponding telecom bands coverage.
This grating has a reflectivity greater than 95% over a band of 290 nm width from 1385 nm up to 1675 nm with a FWHM of 310 nm. The wavelength coverage includes the E+S+C+L+U telecom bands with a grating length of only 35mm. The GLs were measured to be 0.6 dB using a cut-back method at 1700 nm.
It is well known that the FBG of a high chirp rate must have a high index modulation and it is demonstrated by experiments  that the transmission loss T(z) in dB of a linearly chirped FBG is expressed as
where IR and I0 are reflected and incident power, respectively, η(z) is the fraction of modal power that overlaps the grating, Δn(z) is the index modulation in 10-4 and α(z)=95 nm/cm is the chirp rate calculated using α(z) = neffC(z) with the phase-mask chirp rate C(z)=65nm/cm. The measured maximum effective transmission losses T(z) are -6 dB and -18 dB for H2-free and H2-loaded fiber, respectively (as shown in Fig. 1 and 3). Thus, a straight calculation gives a maximum refractive index modulation of Δn(z)~20–25×10-4 in H2-free and 35–45×10-4 in H2-loaded fiber, respectively, where we take η(z) in the range 0.74–1.00 when the FBG is overlapping partially or totally the propagating mode. A numerical simulation using IFO gratingTM confirms these values of refractive index modulation.
As the H2-loaded fiber is much more photosensitive than the H2-free fiber, we obtained about twice of the refractive index modulation with an exposure time 6.5 times shorter (30s for H2-loaded fiber instead of 130s for H2-free fiber). Both the densification and color-centers would occur simultaneously in the H2-loaded fiber, while the densification mainly occurs in H2-free case. This assumption can be better appreciated by comparing the cladding mode losses in the two cases. The CML of 2.5dB is about 10 times greater in the H2-loaded fiber than that of 0.2dB in the H2-free fiber, where very low CML are measured because the densification would occur in both core and cladding of the fiber, resulting in a good overlap between the propagating mode and the grating. This result is in good agreement with  where CML suppression is observed in a H2-free fiber. In the case of H2-loaded fiber, the color-centers are created only in the germanium-doped core leading to a high mismatch between the propagating mode and the grating. Such a mismatch is the main cause of the CML. Note that this high CML in the H2-loaded fiber can be suppressed using a special fiber design  with both core and cladding being photosensitive. The GLs were also measured to be very low in the H2-free fibers with a value lower than 0,1dB for the entire 25 mm of exposed fiber corresponding to a GL lower than 0.04 dB/cm of the exposed fiber. In the H2-loaded fibers, the GLs were measured to be ranging between 0.15 and 0.20 dB/cm of the exposed fiber.
In summary, ultrabroadband fiber Bragg gratings were successfully inscribed in both H2-free and H2-loaded SMF-28 fibers by using the IR femtosecond pulses and highly chirped first-order phase mask. Bandwidths (FWHM) of 85 nm and 310 nm with maximum reflectivity of 75% and 98.5% were obtained in H2-free and H2-loaded SMF-28 fibers respectively after only a few seconds exposure time. The resulting gratings show good spectral and temporal performance by precise characterization over the entire C band. We believe that this report will pave the way to the development of new broadband optical fiber-based components with the unique features for large bandwidths and low losses.
This research was supported by the Canadian Institute for Photonic Innovations (CIPI), the Fonds Quebecois de Recherche Nature et Technologies (FQNRT), the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Canadian Foundation for Innovation (CFI)
References and links
1. S. A. Slattery, D. N. Nikogosyan, and G. Brambilla, “Fiber Bragg grating inscription by high-intensity femtosecond UV laser light: comparison with other existing methods of fabrication,” J. Opt. Soc. Am. B 22, 354–361 (2005). [CrossRef]
2. S. J. Mihailov, D. Grobnic, C. W. Smelser, P. Lu, R. B. Walker, and H. Ding, “Induced Bragg Gratings in Optical Fibers and Waveguides Using an Ultrafast Infrared Laser and a Phase Mask,” Laser Chem. vol. 2008, Article ID 416251, 20 pages (2008).
3. M. Bernier, D. Faucher, R. Vallée, A. Saliminia, G. Androz, Y. Sheng, and S. L. Chin, “Bragg gratings photoinduced in ZBLAN fibers by femtosecond pulses at 800 nm,” Opt. Lett. 32, 454–456 (2007). [CrossRef] [PubMed]
4. E. Wikszak, J. Thomas, J. Burghoff, B. Ortaç, J. Limpert, S. Nolte, U. Fuchs, and A. Tünnermann, “Erbium fiber laser based on intracore femtosecond-written fiber Bragg grating,” Opt. Lett. 31, 2390–2392 (2006). [CrossRef] [PubMed]
6. Y. Dai, X. Chen, J. Sun, and S. Xie, “Wideband multichannel dispersion compensation based on a strongly chirped sampled Bragg grating and phase shifts,” Opt. Lett. 31, 311–313 (2006). [CrossRef] [PubMed]
7. G. Brochu, S. LaRochelle, and R. Slavick, “Modeling and experimental demonstration of ultracompact multiwavelength distributed Fabry-Pérot fiber lasers,” J. Lightwave Technol. 22, 44–53 (2005). [CrossRef]
8. A. Galvanauskas, M. E. Fermann, D. Harter, K. Sugden, and I. Bennion, “All-fiber femtosecond pulse amplification circuit using chirped Bragg gratings,” Appl. Phys. Lett. 66, 1053–1055 (1995). [CrossRef]
9. J. Thomas, C. Voigtländer, D. Schimpf, F. Stutzki, E. Wikszak, J. Limpert, S. Nolte, and A. Tünnermann, “Continuously chirped fiber Bragg gratings by femtosecond laser structuring,” Opt. Lett. 33, 1560–1562 (2008). [CrossRef] [PubMed]
10. R. Vallée, M. Bernier, A. Saliminia, and S. L. Chin, “Fiber Bragg Gratings Based on 1D Filamentation of Femtosecond Pulses,” in Bragg Gratings, Photosensitivity, and Poling in Glass Waveguides (2007), paper BWB3.
11. D. J. Kitcher, A. Nand, S. A. Wade, R. Jones, G. W. Baxter, and S. F. Collins, “Directional dependence of spectra of fiber Bragg gratings due to excess loss,” J. Opt. Soc. Am. A 23, 2906–2911 (2006). [CrossRef]
12. M. Ibsen, M. Durkin, M. Zervas, A. Grudinin, and R. Laming, “Custom design of long chirped Bragg gratings: application to gain-flattening filter with incorporated dispersion compensation,” IEEE Photon. Technol. Lett. 12, 498–500 (2000). [CrossRef]
13. L. Dong, G. Qi, M. Marro, V. Bhatia, L. Hepbrunt, M. Swan, A. Collier, and D. Weidman, “Suppression of cladding mode coupling loss in fiber Bragg gratings,” J. Lightwave Technol. 18, 1583–1590 (2000). [CrossRef]