Page 115 - Optofluidics Fundamentals, Devices, and Applications
P. 115

96     Cha pte r  F i v e


               particle moving along a waveguide, we shift reference frames by
               keeping the particle still and applying boundary conditions that sim-
               ulate the flow moving by it. In the low-Reynolds number regime the
               drag force is linearly proportional to the flow velocity, so the simplest
               method is to apply a “slip” velocity boundary condition on all sur-
               faces of some low flow speed (say 1 μm/s) opposite the direction one
               expects the particle to move. By integrating Eq. (5-2) over the particle
               surface a drag force is obtained, which can be scaled to the particle
               speed by simply dividing it by the imposed slip velocity. For cases
               where an imposed flow is incident on the particle, such as in [Ref. 69],
               it is appropriate to use standard microfluidic boundary conditions
               (see Erickson [71] for details).
                  Figure 5-7 shows the results of finite element-based computations
               performed on dielectric spheres of various sizes.
                  From Fig. 5-7b it is apparent that the propulsive force follows an
               approximate squared relation with particle size (F  ∝ a ) and from
                                                               2
                                                          scat
               the experiments described earlier we know that F  is also propor-
                                                          EM
               tional to the optical intensity. As such it is reasonable to qualitatively
               approximate F    = C a I , where C  is a constant and an unknown
                                    2
                           EM, scat  1  o     1
               function of the physical parameters of the system and I  is the optical
                                                             o
               intensity. Using Faxen’s law approximation  F  = 6πηaU /g(a/h),
                                                        flow      o
               where g(a/h) is the denominator in Eq. (5-4). If the distance between
               the bottom of the particle and the top of the waveguide is small com-
               pared to the particle diameter, a/h ~1 then g(a/h) is a constant and we
               can derive the following approximate equation descriptive of the
               transport velocity:
                                          2
                                       Ca I        aI
                               U =      1      =  C  o              (5-10)
                                 o                2  η
                                    6πη /( /
                                       ag a h)
               where C  is a different constant comprising of the same system phys-
                      2
               ical parameters as C . We note from Eq. (5-10) that (based on these
                                 1
               computations) the transport velocity in a ≥ λ appears to contain only
               a linear dependence on particle size (as opposed to the much stronger
               dependence in the Rayleigh regime).
               5-4-5  Comments on the Influence of Brownian Motion
                       and Trapping Stability
               The transport path taken by a particle with a diameter greater than
               approximately 1 μm is largely deterministic and well described by
               Eq. (5-10). Below 1 μm, however, they begin to be significantly affected
               by Brownian motion and the hydrodynamic solution becomes pre-
               dictive of the average motion of the transported particle (about
               which it will diffuse), rather than its actual pathline. The underlying
               cause of Brownian motion is the random fluctuating collisions of sol-
               vent molecules impacting larger microscopic particles. Brownian motion
   110   111   112   113   114   115   116   117   118   119   120