HAGAN et al.: NOVEL WAVEGUIDE-BASED RF/MW EXPOSURE SYSTEM FOR STUDYING NONTHERMAL EFFECTS

1671

TABLE I

CELL PERFUSION CHAMBER PARTS AND THEIR DIELECTRIC PROPERTIES

exposure system, the probe launch in the FDTD model was ex-

simulation ran for 9550 time steps. For the sake of brevity, the

cited by a discrete sinusoidal 50- RF/MW source. This source

results are only presented in this paper at a frequency of 1 GHz.

was placed at the base of the probe launch.

For the same net input power level (forward minus reflected

All parts of the cell perfusion apparatus, except for the fil-

power) into the waveguide, the XFDTD model provided similar

ters, were represented accurately by the Yee cell size used in the

results for homogeneity and SAR distributions at other frequen-

subgrid. Because the lattice dimension of the filters is smaller

cies in the range of interest.

than the minimum size of a subgrid cell (0.3048 mm), the lat-

tice can be represented only approximately. The glass fiber filter

D. Dielectric Properties of Materials Used in the Exposure

has a very fine mesh and was approximated as a solid dielec-

System

tric material having a radius (12 mm) and thickness (0.26 mm)

equal to that of the actual filter when it was saturated with BSS.

The nylon filters sandwiching the glass fiber filter are comprised

The accuracy of the results of the XFDTD model depends

of a much coarser weave of nylon thread that forms relatively

on the accuracy of both the representation of the geometry and

large rectangular hollow regions in which BSS can accumulate.

the dielectric properties of the materials used. The relative di-

The dimensions of the rectangular hollow regions and the diam-

electric constant and electrical conductivity (Table I) of most of

eter of the nylon threads are close to the size of a subgrid cell

the materials comprising the cell perfusion apparatus assembly

(0.3048 mm) and are approximated as such.

were obtained from well-established literature values [16]. For

For all FDTD simulations, the default time step was chosen

those materials for which published values of dielectric prop-

in both the main grid (2.0964 ps) and subgrid (0.420 ps). Each

erties were not easily available (dental putty, Plexiglas, BSS,