Effect of phase mask stitching errors on the spectral response of uniform and apodized fiber Bragg gratings

Stitching errors can be divided in two types, random and systematic, and each one modifies in its own way the spectral response of the Bragg grating. For example, we compare the calculated spectral responses of a perfectly coherent Bragg grating with another one containing both types of errors. When the stitching error is systematic, with equal values at every writing-field border, Fabry-Perot effects occur inside the grating. This explains the strong sidebands appearing at regular intervals on either side of the resonant lobe. The spacing of the Fabry-Perot sidebands is related to the length of the writing-field. On the other hand, the random stitching errors vary from field to field and induce sidebands close to the main lobe. For coherent gratings, sidelobe levels can be reduced, or “apodized”, by modulating the coupling coefficient along the grating. For stitched gratings, the effect of apodization has been calculated. We see that apodization has no effect on the Fabry-Perot sidebands generated by the systematic errors. However, the magnitude of the sidebands resulting from random stitching errors is reduced by more than 12 dB. Experimental measurements of the spectral responses of two Bragg gratings written with stitched masks are given. The level of the sidebands close to the resonant lobe has been reduced by more than 14 dB. The dissimilarity between the spectra is explained by the fact that the gratings have been written with different masks and thus different random patterns