Intrinsic apodization of Bragg gratings written using UV-pulse interferometry

We present a new interferometric technique to write apodized Bragg gratings. Along the fiber, two UV-pulses interfere with a variable time delay creating a gaussian apodization profile which strongly reduces the reflectivity side-lobes. This method allows writing of truly apodized gratings by single exposure of a uniform phase-mask. Introduction: In optical communication systems, applications of fibre Bragg gratings include, for example, channel filtering, chromatic dispersion compensators, optical gain equalizers, pulse compression and soliton sources. In this context, the fabrication of apodized Bragg gratings has raised much interest because of its reduced reflectivity sidelobes. It therefore increases the quality of optical filters [1] and improves the P.-Y. Cortès et al., “Intrisic apodization of Bragg gratings... ”, Electron. Lett., vol. 34, pp. 396-397 (1998). 2 dispersion compensation by simultaneously reducing the group-delay ripples [2]. Several techniques have been proposed to write apodized Bragg gratings. The moving fibre/phase mask-scanning beam approach can be used with uniform phase masks [3]. It is also possible to vary the exposure time along the grating length and to subsequently compensate the variation of the average index by a second exposure without the phase mask [4]. An alternative method is to use a phase mask with a variable diffraction efficiency [5]. In the following, we present a simple interferometric technique in which the variation in the overlap of two interfering UV-pulses along a fibre results in intrinsically apodized Bragg grating. This approach achieves a uniform average index change over the grating length using a single-exposure and a uniform phase mask. Experiment and results: The experimental setup, displayed in Fig. 1, basically consists of a Sagnac interferometer where the phase mask acts as a beam splitter. The writing beam consists of short UV pulses from a mode-locked laser. The mirror (M) directs the UV-beam to the phase mask. This mirror can move along the x-axis to scan the phase mask length. The optical fiber is placed at the location of the phase mask but slightly above the incoming UV-beam. The phase mask splits the UV-beam into the ±1 orders. The two diffracted beams are then reflected off the mirrors M1 and M2. A slight tilt with respect to the x-y plane is introduced in these mirrors so that the beams are recombined to form an interference pattern on the optical fibre. A cylindrical lens is introduced between the two mirrors to focus the beams on the fibre core. When the UVlight is incident on the right side of the phase mask, the interference occurs on the left portion of the fibre and vice versa. The diffraction angle θ is determined by the phase P.-Y. Cortès et al., “Intrisic apodization of Bragg gratings... ”, Electron. Lett., vol. 34, pp. 396-397 (1998). 3 mask period Λm and the writing wavelength. The Bragg wavelength (λB) of the photoinduced grating can be varied by changing the angle Φ of the mirrors as will be reported elsewhere. The position of the interference pattern and therefore of the fiber is also displaced along the y-axis as Φ is varied. When the interference occurs directly behind the phase mask, λB is the same as in other phase-mask writing method, i.e λB = n Λm /2, where n is the fiber refractive index. Inthis case, the angle Φ is given by (π/4)-(θ/2). From Figure 1, it can be seen that the optical path difference (∆δx) between the two counter-propagating beams is zero when the UV beam is incident on the center of the phase mask. However, ∆δx increases as the UV beam moves to the edges of the mask. From geometrical consideration, when Φ =(π/4)-(θ/2), it is easy to demonstrate that ∆δx depends linearly on the position x of the UV-light on the phase mask as given by the relation: ) sin 2 1 ( 2 θ δ − = ∆ x x where x=0 at the center of the phase mask. The two UVpulses coming from the -1 and +1 order of the mask will therefore arrive with different delays on the fibre. At x=0, the pulses arriving at the fiber completely overlap in time. For all other positions x, the pulses will be only partially overlapping. The strength of the interference pattern being related to the product of the field amplitudes in both arms of the interferometer, it can be concluded that the visibility of the interference fringes will degrade towards the fiber ends. In fact, assuming gaussian pulses, this convolution of the two pulses will produce a gaussian apodization profile (Figure 2). The dark area of the displayed pulses contributes to create the grating and the remaining energy of the two pulses contributes to increase the average refractive index. Moreover, in this experiment, the total UV-exposure is constant over the grating length whatever the P.-Y. Cortès et al., “Intrisic apodization of Bragg gratings... ”, Electron. Lett., vol. 34, pp. 396-397 (1998). 4 overlap of the pulses. In a linear regime, the average index change is therefore intrinsically uniform along the apodized grating which is another advantage of this technique. However, it should be noted that the temporal characteristics of the UVpulses completely determines the apodization profile and the grating length. Despite this limitation, this experimental configuration is attractive since it combines the stability of a Sagnac interferometer, the flexibility in the choice of the Bragg wavelength offered by an interferometric set-up, and the intrinsic apodization of Bragg gratings written with single exposure of a uniform phase-mask To write the gratings, we used a frequency quadrupled CW-mode locked YAG producing an average UV-power of approximately 70 mW. The UV-pulses are modeled as gaussian function of 60 ps width (Full Width Half Maximum) corresponding to spatial width of 1.8 cm. The diameter of the mirrors M1 and M2 and the phase mask length are 5 cm. However, because of experimental constraints only 3.8 cm of the mask is used and therefore the resulting apodization profile is truncated. The apodization function along the fibre, as given by the convolution of two gaussian pulses, is shown in Figure 2. With sin(θ) = 0.248, the OPD is ∆δx ≈ x, and the convolution of the two incoming pulses directly translates in the apodization curve. Figure 3 shows the experimental reflection spectrum of an apodized grating along with its calculated reflection transmission. The calculated spectral reflection of a uniform grating is also displayed in the insert. The spectral resolution of the measurement is 0.01 nm. The large 3 dB bandwidth of the grating, 0.26 nm, results from the strong index modulation. The measured relfection spectrum displays a smooth profile although some residual sidelobes can be seen. These small peaks are also present on the theoretical curve and P.-Y. Cortès et al., “Intrisic apodization of Bragg gratings... ”, Electron. Lett., vol. 34, pp. 396-397 (1998). 5 result from the truncated apodization profile. This problem could easily be solved with larger diameter mirrors. Nonetheless, the comparison with uniform grating transmission clearly shows the strong attenuation of the reflectivity sidelobes. Conclusion: We have demonstrated a new and very simple interferometric technique to write intrinsically apodized Bragg grating using single-exposure of a uniform phasemask. This technique is based on the time delay between two counter-propagating pulses in a Sagnac interferometer. A uniform average index change results from the constant UV-exposure. Moreover, this technique easily allows tuning of the Bragg grating wavelength by changing the mirror angle. Acknowledgements: This work was supported by NSERC, QuébecTel and Bell Canada. References1. Komukai T., Tamura K., and Nakazawa M.: “An efficient 0.04 nm apodized fibreBragg Grating and its application to narrow-band spectral filtering”, IEEE Photon.Technol. Lett., 1997, vol. 9, no 7, pp. 934-936.2. Loh W. H., Laming R. I., Ellis A. D., and Atkinson D.: “10 Gb/s transmission over700 km of standard single-mode fiber with 10-cm chirped fiber grating compensatorand duobinary transmitter”, IEEE Photon. Technol. Lett., 1996, vol. 8, no 9, pp. 1258-1260.3. Loh W. H., Cole M. J., Zervas M. N., Barcelos S., and Laming R. I.: “Complexgrating structures with uniform phase masks based on the moving fibre-scanning beamtechnique”, Opt. Lett., 1995, vol. 20, no 20, pp. 2051-2053. P.-Y. Cortès et al., “Intrisic apodization of Bragg gratings... ”, Electron. Lett., vol. 34, pp. 396-397 (1998). 64. Malo B., Thériault S., Johnson D. C., Bilodeau F., Albert J. and Hill K. O.: “Apodisedin-fibre Bragg grating reflectors photoimprinted using a phase mask”, Electron. Lett.,1995, 31, pp. 223-225.5. Albert J., Hill K. O., Malo B., Thériault S., Bilodeau F., Johnson D.C. and Erickson L.E.: “Apodisation of the spectral response of fibre Bragg gratings using a phase maskwith variable diffraction efficiency”, Electron. Lett., 1995, 31, pp. 222-223. P.-Y. Cortès et al., “Intrisic apodization of Bragg gratings... ”, Electron. Lett., vol. 34, pp. 396-397 (1998). 7Figure captions Fig. 1: Experimental setup Fig. 2: Normalized apodization profile corresponding to the convolution of the two UV-pulses as a function of the position of the interfering beams on the optical fiber. Fig. 3: Reflection spectra of apodized Bragg gratings: experimental (dark curve),calculated apodized grating (grey curve), and calculated uniform grating (insert). P.-Y. Cortès et al., “Intrisic apodization of Bragg gratings... ”, Electron. Lett., vol. 34, pp. 396-397 (1998). xθLφM1φ UV-beamPhase maskFiberLens