Sequential Writing
4.2 Nonlinear Photosensitivity and Pre UV Treatment
linear index response to pre UV treatment in the matching of the target grating index profile has not been studied. In our work, the side-diffraction method is adopted to determine the refractive index profiles of fabricated FBGs more precisely. The linear grating index response is first examined using a single Gaussian UV shot with pre UV treatment. Our work demonstrates how pre UV treatment helps to achieve a linear index change with UV flux, even though two writing scans are required. An improved method of unequal two-beam interference is then proposed to generate the required AC and DC amounts of UV flux in a single writing scan. This unequal interference setup provides greater stability in writing weak gratings in the linear regime.
4.2 Nonlinear Photosensitivity and Pre UV Treatment
This work begins by calibrating more carefully than before the relationship between the induced change in the index of a photosensitive optical fiber and the UV flux to which it is exposed. The pre UV treatment method is then performed to achieve a linear index change response. Unlike in previous works [4.6,4.7], which assumed that the grating index profile was sinusoidal or uniform to fit the reflectivity, the side-diffraction method [4.8-4.12] is employed herein to scan the fiber grating and thus determine accurately the grating index modulation profile. An iterative procedure is then utilized to fit the reflection/transmission spectra, based on the transfer matrix method for determining the actual grating index modulation. The measured results reveal that the grating shape is deformed for simple writing schemes when the UV intensity is low. The relationship between the induced refractive index change and the exposed UV flux is nonlinear at low UV flux, such that the written grating index profile is deformed. As stated above, the pre UV treatment can be performed in advance to prevent the nonlinear photosensitivity in the writing of a weak grating. The
45
working principle of pre UV treatment is to eliminate the nonlinear region using a DC bias flux so that the index change depends linearly on the UV flux, and the induced index profiles are similar to the envelope of the holographic writing UV beam.
The experiments are performed on Fibercore (PS1500) photosensitive fibers, which were hydrogenated at a pressure of 1,500 psi at room temperature for seven days.
The UV source is a frequency-doubled CW argon-ion laser with an output wavelength of 244 nm. A Gaussian-shaped UV fringe with its 1/e2 width about 6.5 mm was used to imprint holographically the fiber Bragg grating in the fiber core. The induced refractive index profile was determined using the side-diffraction method [4.12].
Furthermore, an ASE light source and an optical spectrum analyzer are employed to measure the spectral response. The normalized refractive index profile is used directly to calculate the reflection and transmission spectra using the transfer matrix method.
An iterative procedure is then applied to fit the measured spectra to determine the peak index modulation of the entire FBG.
Figure 4.1(a) depicts the refractive index profiles that measured by the side-diffraction method for exposure to specified amounts of UV flux. The scan step of the translation stage is set to 100 μm. Figure 4.1(b) plots the fitted peak refractive index modulation as a function of the UV flux at the center of grating. As expected, the change in the refractive index of the photosensitive fiber is nonlinear at low UV flux and linear thereafter (No saturation is observed at the UV flux of interest). The lowest UV flux case (15 J/cm2) in Fig. 4.1(a) corresponds to the rapidly changing exponential-shaped relationship in Fig. 4.1(b), which makes the grating index profile flat-topped. In previous works of mechanism studies [4.13], the photosensitization of optical fibers is modeled by a two-step process. The refractive index change is determined mainly by the concentration of Ge defect centers and defect sites for
46
hydrogen reaction. The model indicates that the change in the refractive index curve at low UV flux has an exponential form, which result agrees with the measurements in Fig. 4.1(b). Such nonlinear behavior causes the fiber grating index modulation profile to differ from the writing UV periodic intensity envelope. The shapes deform mostly at the tail edges of the Gaussian laser beam, where the UV intensity is low.
The pre UV treatment is conduced before the grating is made by pre-exposing the fiber uniformly to the UV radiation to prevent deformation of the grating index profile.
Uniform UV pre-exposure is achieved by translating the stage step by step so that the 244 nm Gaussian-shaped UV beam partially overlaps to write on a large length of the photosensitive fiber, forming a uniform DC index change in advance. The pre UV flux is estimated to be around 59.5 J/cm2. The pre UV flux is applied at the point between the linear region and nonlinear region in Fig 4.2(b).The middle of the pre-exposed region is then exposed to holographic Gaussian UV beam to write the FBG. Figure 4.2(a) displays the refractive index modulation profiles for specific exposed UV flux following pre UV treatment. Figure 4.2(b) plots the peak refractive index modulation as a function of the UV exposure flux at the center of the grating following pre UV treatment. Notably, the index modulation now depends linearly on the UV exposure flux and the induced grating index modulation profiles seem to be Gaussian-like. The 1/e2–intensity half widths of the one-shot UV used to generate grating index profile in Fig. 4.2(a) were about 6.55 mm, 6.35 mm, and 6.75 mm, similar to the UV beam width of 6.5 mm at a radius of 1/e2. This result establishes the linear refractive index response to UV flux after pre UV treatment. The pre UV treatment causes the operating point to jump away from the initial nonlinear regime such that the following grating writing process is entirely in the linear regime. The slopes in the linear region of Fig. 4.1(b) and in Fig. 4.2(b) are 4.15×10-7 and 8.22×10-7, respectively, indicating
47
that pre UV treatment enhanced the photosensitivity. This result is consistent with the results of previous studies which have demonstrated that photo-hypersensitivity increases the photosensitivity of optical fibers [4.7]. The pre UV treatment method is established herein to be a practical method of writing a weak fiber Bragg grating with the target index profile. The narrow-band FBGs were fabricated by using the strongly overlapping step-scan exposure scheme and the real-time interferometric side-diffraction position monitoring method [4.14] following pre UV treatment. The side-diffraction interferometric method is developed to connect the sections of gratings precisely. Our experience is that obtaining a low noise suppression ratio in the reflection spectrum of under 10 dB without pre-UV treatment is difficult because the nonlinear UV shots overlap process generates imperfections in the apodization index profile.