Page 285 - Phase Space Optics Fundamentals and Applications
P. 285

266   Chapter Eight


               dimensions, the corresponding field estimate then takes the form of
               an integral over the PSC

                              U(x, z) ≈  u(z,  )g  (z, ) (x, z,  ) d   (8.85)

               where

                                 k          2
                g   (x, z,  ) = exp  −  [x−X(z,  )] +ik{L(z,  )+P(z,  )[x−X(z,  )]}
                                 2
                                                                    (8.86)
               and u(0,  ) and  (0,  ) must be chosen so that the superposition
               matches the initial field at z = 0. Again, this method is valid as long
               as the parabasal approximation holds for each independent Gaussian
               beam.




          8.9 Stable Aggregates of Flexible Elements
               A different approach for building wave field estimates based on rays
               throughthesuperpositionofGaussiancontributionsisnowdiscussed.
               Like Eq. (8.85), this framework, referred to as stable aggregates of flexi-
               ble elements (SAFE), 39–43  takes the form of a continuous superposition
               of Gaussian components around the rays in the Lagrange manifold.
               However, in SAFE, the Gaussian contributions are not independently
               propagating parabasal Gaussian beams; rather, they are interrelated
               contributions, where   is not constrained to have a specific depen-
               dance on z or  :

                               U(x, z) =  u(z,  )g   (x, z,  ) d    (8.87)

               In what follows, it is assumed for simplicity that   is a real and positive
               constant, although the results remain valid if this parameter has an
               imaginary part and/or is allowed to vary slowly with z,  , or even x.


               8.9.1 Derivation of the Estimate
               To find the specific form of this estimate, we substitute Eq. (8.87) into
               Eq. (8.1). By following the same steps as in Sec. 8.8, we obtain a result
               that is identical to that in Eq. (8.81), except for the presence of an
               integral in  :

                   2
                                   2
                 [k (C 20 +C 21  +C 22   +· · ·)+k(C 10 +C 11  )+C 00 ] g   d  = 0  (8.88)
   280   281   282   283   284   285   286   287   288   289   290