Am. Melo et al., Optical time-division multiplexing using picosecond solitons in a terahertz optical asymmetric demultiplexer, OPT COMMUN, 186(1-3), 2000, pp. 87-97
The performance of a terahertz optical asymmetric demultiplexer (TOAD) oper
ating with an ordinary fiber and with a dispersion decreasing fiber (DDF) a
nd dispersion increasing fiber (DIF) configurations, for three length of fi
ber (xi = pi /2, 2 pi and 5 pi) and using soliton and quasi-soliton laser p
rofile for the control pulse, was studied. The numerical simulations shows
that the increase of the fiber length lead to the decrease of the power for
the first and second demultiplexed pulses and lead to a broadening of thes
e pulses, with exception to the TOAD operating with the DDF fiber. For the
TOAD operating with a basic telecommunication fiber one see that the increa
se of the power of the control power lead to a strong compression of the de
multiplexed pulse. One can also conclude that operating the TOAD with one s
oliton period (xi = pi /2) one has compression of the demultiplexed pulse.
For higher length (four and 10 soliton periods) one start having broadening
of the switched pulses. Operating the TOAD using a DDF fiber one can say t
hat the control power (P-CTL1) necessary to demultiplex the signal pulse is
always lower compared with the TOAD with the normal telecommunication fibe
r. This is a strong suggestion that the use of the DDF fiber will allow the
use of less control power. One can also say that P-CTL1 is always higher f
or the soliton profile compared with the quasi-soliton profile. It was also
observed that the compression factor for the demultiplexed pulse (C-1) is
between 2 and 3 for all the lengths of the TOAD and considering the soliton
and quasi-soliton profile for the control pulse. When operating the TOAD u
sing a DIF fiber it was observed that P-CTL1 for the DIF TOAD is always low
er compared with the TOAD with the normal telecommunication fiber (with exc
eption to the xi = 5 pi fiber in the soliton profile), however the power va
lues are higher when compared with the DDF TOAD. It was also observed that
the compression factor C-1 is showing compression (C-1 similar to 2.2) for
shorter length of fiber xi = pi /2 considering the soliton and quasisoliton
, profile for the control pulse. For higher lengths of fiber the compressio
n factor is showing broadening of the demultiplexed pulse (with exception t
o the xi = 5 pi fiber in the soliton profile). Our simulations considering
the TOAD operating with a DDF and DIF with a linear profile conclude that i
t is possible to operate the TOAD with lower control power using a DDF fibe
r setup. For this device the demultiplexed pulses will present a compressio
n on time duration and will be insensitive to the time profile of the contr
ol pulse. (C) 2000 Elsevier Science B.V. All rights reserved.