Page 226 - Introduction to Information Optics

4.2. All-Optical Switches .: i !

can only be obtained through numerical analysis. It should be noted that a
pulse contains a distribution of power, ranging from zero at the edge to peak
power at the center. Therefore, different parts of the pulse are switched by
different amounts, which may lead to pulse breakup. Several schemes have
been developed to minimize this problem. One approach is to use temporal
solitons, where the nonlinear effect for pulse breakup is balanced by dispersion.
However, solitons require special conditions for their existence. More details
on this topic as well as fundamentals of solitons can be found in [1].
NLDC has been realized using nonlinear waveguides using GaAs and other
semiconductors as well as dual-core optical fibers.

4.2,3.2. Controlled Switching
Controlled switching can also be implemented. In this case, a separate
control beam can be coupled into one of the nonlinear waveguides, and the
refractive index of this waveguide depends on the intensity of the control signal.
Since the input is an optical beam independent from the signal, the result is the
same as the linear case, except that output in waveguides A and B is controlled
by the nonlinear phase shift which affects A in Eq. (4.13). Assume that the two
nonlinear waveguides are identical, and the control beam was sent into one of
the waveguides. In this case, the phase mismatch (A) is proportional to
(27r//,)n 2/ c, where I c is the control intensity. Figure 4.9 shows the output power
in waveguides A and B as a function of A (normalized to K) for z = L c.
Note that this scheme for controlled switching has also been realized using
electro-optical media, such as LiNbO 3 and GaAs. In these cases the phase shift
is controlled by an external electric field [7], as discussed in Sec. 3.

4.2.4 NONLINEAR INTERFEROMETRIC SWITCHES

The most successful form of all-optical switches is the nonlinear interfer-
ometer. The Sagnac, Mach -Zehnder, and Michaelson configurations have all
been used for implementing optical switches. The basic operation is the same
for all interferometric configurations where the nonlinear effect creates an
additional intensity-dependent phase shift between the two arms. The signal to
be switched is split between the arms of the interferometer. The output
intensity of the interferometer depends on the total phase and, therefore, can
be controlled using an external control signal. If we have two beams (1 and 2)
with different optical paths, n,r, and n 2r 2

£\ = ,4 1 cos[(27i/A)« 1r 1 + </>]]
(4.18)
E 2 = A 2 cos[(27r//)n 2r 2 + </> 2],