Page 516 - Engineering Electromagnetics, 8th Edition
P. 516

498                ENGINEERING ELECTROMAGNETICS

                                        The main result of the weak-guidance condition is that a set of modes appears in
                                     which each mode is linearly polarized. This means that light having x-polarization,
                                     for example, will enter the fiber and establish itself in a mode or in a set of modes
                                     that preserve the x-polarization. Magnetic field is essentially orthogonal to E, and so
                                     it would in that case lie in the y direction. The z components of both fields, although
                                     present, are too weak to be of significance; the nearly equal core and cladding indices
                                     lead to ray paths that are essentially parallel to the guide axis—deviating only slightly.
                                                                       .
                                     In fact, we may write for a given mode, E x = ηH y , when η is approximated as the in-
                                     trinsicimpedanceofthecladding.Therefore,intheweak-guidanceapproximation,the
                                     fiber mode fields are treated as plane waves (nonuniform, of course). The designation
                                     for these modes is LP  m , meaning linearly polarized, with integer order parameters
                                       and m. The latter express the numbers of variations over the two dimensions in the
                                     circular transverse plane. Specifically,  , the azimuthal mode number, is one-half the
                                     number of power density maxima (or minima) that occur at a given radius as φ varies
                                     from 0 to 2π. The radial mode number, m,expresses the number of maxima that occur
                                     along a radial line (at constant φ) that extends from zero to infinity.
                                        Although we may assume a linearly polarized field in a rectangular coordinate
                                     system, we are obliged to work in cylindrical coordinates for obvious reasons. In
                                     a manner that reminds us of the rectangular waveguide, it is possible to write the
                                     x-polarized phasor electric field within a weakly guiding cylindrical fiber as a product
                                     of three functions, each of which varies with one of the coordinate variables, ρ, φ,
                                     and z:


                                                     E xs (ρ, φ, z) =  R i (ρ)  i (φ)exp(− jβ i z)  (144)
                                                                   i
                                     Each term in the summation is an individual mode of the fiber. Note that the z function
                                     is just the propagation term, e − jβz , since we are assuming an infinitely long lossless
                                     fiber.
                                        The wave equation is Eq. (58), which we may write for the assumed x component
                                     of E s ,but in which the Laplacian operator is written in cylindrical coordinates:

                                                           2
                                                                      2
                                                  1 ∂     ∂ E xs     1 ∂ E xs  2   2
                                                        ρ       +         + (k − β )E xs = 0        (145)
                                                  ρ ∂ρ     ∂ρ      ρ 2  ∂φ 2
                                                                  2
                                                              2
                                     where we recognize that the ∂ /∂z operation, when applied to (144), leads to a
                                     factor of −β .Wenow substitute a single term of (144) into (145) [since each term
                                               2
                                     in (144) should alone satisfy the wave equation]. Dropping the subscript i,expanding
                                     the radial derivative, and rearranging terms, we obtain:
                                                                                      2
                                                      2
                                                        2
                                                     ρ d R   ρ dR     2  2   2     1 d              (146)
                                                     R dρ 2  +  R dρ  + ρ (k − β ) =−   dφ 2

                                                                   2                  2
                                     We note that the left-hand side of (146) varies only with ρ, whereas the right-hand side
                                     varies only with φ. Since the two variables are independent, it must follow that each
                                                                                              2
                                     side of the equation must be equal to a constant. Calling this constant   ,as shown,
   511   512   513   514   515   516   517   518   519   520   521