Page 228 - Fiber Bragg Gratings

5.2 Basic principles and methodology 205

fiber, but also on the type of exposure, whether hydrogen loaded or not, use
of a pulsed or CW source, as well as the wavelength of the UV radiation. A
further complication may occur due to the effects of "incubated" grating
formation [11], in which nonlinear growth of a grating occurs. It is there-
fore expected that other methods may be easier to use routinely, requiring
less processing.

5.2.3 The variable diffraction efficiency phase mask
A phase mask with a variable diffraction efficiency has been used for the
fabrication of apodized gratings [12]. There are two methods of fabricating
such a mask. The diffraction efficiency into the +1 and -1 orders is
maximized for a 1:1 mark-space ratio of the grating and the zero order
minimized for a specific groove depth (see Chapter 3). Therefore, there
are two degrees of freedom to alter the diffraction efficiency. The
mark-space ratio of the grating etched in the phase mask or the groove
depth may be varied. In the technique reported, a variable diffraction
efficiency phase mask, was fabricated by direct exposure of a silica plate
to an ion beam of silicon. In this direct write method the ion beam was
focused to a spot diameter of 100 nm and scanned across the plate to
delineate the grooves. Wet etching in a l-mol% solution of hydrofluoric
acid in water was used to develop the mask. It was demonstrated that
the etching rate is dose dependent. Groove widths between 100 and 550
nm and depths from 7.5 to SOOnm could be achieved by varying the ion
14 2
dose from 0.5 to 4 X 10 ions/cm . The etch rate is faster for regions
exposed to higher doses. The diffraction efficiency into the ± 1 and 0 orders
was measured as a function of the dose; the diffracted orders were shown
14
2
to be a linear function up to a dose of 2.25 X 10 ions/cm . Thus, a variable
diffraction efficiency phase mask was fabricated, using a Gaussian profile
2
2
14
2
dose of (2.25 X 10 ions/cm ) exp(-x /(0.420) ), where x is measured in
mm from the center of the 1-mm-long grating, with a period of 1.075 /mi.
One of the difficulties of fabricating such a grating is the stepped move-
ment of the ion beam. As a result, the Gaussian phase mask profile can
only be approximated, and altering the dose in 40 dose steps did this.
Subsequently, apodized gratings were imprinted in standard telecom-
munications hydrogen loaded fiber and shown to have reduced the first
set of side lobes by approximately 14 dBs for a 10% reflectivity grating.
These results are not as good as those from the double exposure method
(see Section 5.2.2) in which a reduction in the side lobes of 20 dB was