Page 227 - Chalcogenide Glasses for Infrared Optics

IR Imaging Bundles Made fr om Chalcogenide Glass Fibers 203

inside the channel with any accuracy. We requested a modification of
the software by Tom Loretz and Paul Modlin that was accomplished
in February 1999. In our first attempt, we chose to use large-diameter
fiber to minimize breakage. Core diameter was 175 µm with center-
to-center spacing of 210 µm. The bundle was formed from 24 ribbons,
33 count with a finished size of 7 mm × 4.7 mm which yielded a cal-
culated active area of 57.3 percent. The bundle was coarse, with voids,
and had poor resolution, about 2 lp/mm.
Navy Bundle 10-M-2
The second attempt was formed from 42 ribbons, 69 count with
100-µm core diameter. The finished bundle was 7.4 mm × 5.0 mm
with a calculated active area of 61.5 percent. The bundle contained
more fibers in number, 2898, and more fiber in meters, 28,980, than
any bundle we had made at that time. The fused area was large and
the fibers were closely packed. The numbers and bar pattern from the
−1 group, one step larger than the O group, were transmitted. Limiting
resolution was poor, about 1 lp/mm. The bundle showed low con-
trast, which makes image recording difficult. In the camera image,
one can resolve the “O” pattern and the 1 pattern for a resolution
limit of about 3.5 lp/mm. We were able to show the low contrast was
due to a higher than normal water content in the glass used for the
bundle. The fundamental absorption occurs in the glass at 2.92 µm
with an overtone at 1.45 µm, close to the wavelength of maximum
sensitivity of the NIR camera. The overtone absorption becomes
appreciable for a 10-m optical path.

Navy Bundle 10-M-3
The third attempt was designed to increase the active area of the
bundle by using a slow-drying epoxy on the layers, which should
have allowed the fibers to settle down (we thought) between the
fibers of the previous layer. In this way the packing density should
increase, which in turn would yield a higher active area. As very often
happens, our result was slightly the opposite. We wound 42 layers of
69 count ribbons for a total of 2899 fibers in a bundle 7.7 mm × 5.7 mm
with a calculated area of 58 percent, down slightly from 62 percent.
The appearance was also no better than 10-M-2 plus there were dark
regions in the bundle when viewed with the NIR camera. The dark
regions disappeared when viewed with the Agema 210, again sup-
porting the idea that the dark regions were caused by water concen-
tration variation in the fiber.

Navy Bundle 10-M-4
The last 10 m fabricated for the program was by far the best. First, we
had recognized that the plastic channel as a complete container for
the bundle was not flexible in all directions, leading to crimping and
fiber breakage. We restricted our use of the channel to only a small