Академический Документы
Профессиональный Документы
Культура Документы
1990
Boaget al.
A novel solution is presented for the problem of three-dimensional electromagnetic scattering of a plane wavefrom a
doubly periodic infinite array of perfectly conducting bodies. A set of fictitious spatially periodic and properly
modulated patches of magnetic current is used to simulate the scattered field. These patch currents are of dual
polarization and have complex amplitudes. The electromagnetic field radiated by each of the periodic patch
currents is expressed as a double series of Floquet modes. The complex amplitudes of the fictitious patch currents
are adjusted to render the tangential electric field zero at a selected set of points on the surface of any of the
scatterers. The procedure is simple to implement and is applicable to arrays composed of smooth but otherwise
arbitrary perfectly conducting scatterers. Results are given and compared with an analytic approximation.
1.
INTRODUCTION
The study of diffraction of a plane wave from periodic structures is long standing. It has been motivated by academic
curiosity as well as many engineering applications. It is of
practical importance in designing reflection and transmission gratings often used as filters, broadband absorbers,
polarizers, and frequency scanned reflectors. While singly
periodic gratings have been treated extensively, doubly periodic gratings, being in general more difficult not only to
analyze but also to fabricate, have received considerably less
attention. One type of doubly periodic structure that has
been investigated by many researchers comprises infinitesimally thin planar doubly periodic screens in various configurations, often referred to as frequency-selective surfaces.1-3
Perfectly conducting screens of finite thickness consisting of
doubly periodic arrays of apertures, known as inductive
grids, have also been extensively studied.4 5 While the
structures in Refs. 1-5 may be different, a common ingredient of essentially all of them is that the fields in the various
regions can be represented by modal expansions relative to
the axis normal to the grating. These modal representations are then matched by using appropriate boundary conditions, and the unknown modal coefficients are readily determined. In contrast to the above discussion, rigorous
studies of diffraction from doubly periodic arrays of finitesized perfectly conducting scatterers, referred to as capacitive grids, have not been reported in the literature. In this
latter case the fields in the space gap between the scatterers
forming the array cannot be represented in terms of analytically known modal functions.5 Therefore even the method
devised in Ref. 6 for the analysis of diffraction by doubly
periodic surfaces falls short, while differential method procedures might be infeasible with present-time computer
storage and speed limitations.
In this paper we present a new method for analyzing threedimensional electromagnetic scattering from doubly periodic arrays. The technique is applicable to arrays composed of
perfectly conducting bodies of smooth, but otherwise arbitrary, shapes. An example of an array is depicted in Fig. 1.
0740-3232/90/091712-07$02.00
(EH")
\k
Boag et al.
Unbounded Space (,
E)
kin'
e are
allowed to be complex.
r),
Einc(r) = EinCexp(-jkinc
Fig. 1.
1713
(1)
with harmonic exp(jwt) time dependence assumed and suppressed, is incident on the grating. Here, kincand E 0c denote, respectively, the wave vector and the amplitude of the
incident field. Our objective is to determine the field scattered by the grating (Es, Hs) (i.e., the actual field minus the
incident field). The field should be a source-free solution of
the Maxwell equations and obey the Floquet periodicity
conditions
p = 1, 2.
(2)
FORMULATION
A.
Hs) is simulated by a field of a set of doubly periodic fictitious patches of magnetic current Mqi, q = 1, 2, i = 1, 2, ... ,
N. These sources are located in the region occupied by the
scatterers in the original situation and are treated as sources
Unbounded Homogeneous Space
(Einc ,Hinc)
(gia)
kinc
(Es+Einc Hs+Hinc)
Mathematical Boundary C
1714
1990
Boag et al.
Es(r) =
Z
y KqjEqj(r),
(4)
q=1 i=1
es produce fields satisfying the Floquet periodicity conditions, the simulated scattered field [Eq. (4)] also satisfies
them.
It is important to note that the location of the sources in
the simulated equivalence has not been specified yet. As far
as the formulation is concerned, their location can be arbitrary. The question of selecting source locations that are
suitable for a numerical solution is an important one. From
the numerous geometries considered in our earlier research
with perfectly conducting and penetrable scatterers,7 -'0 we
have concluded that the sources should be placed on surfaces
of a shape similar to that of the actual boundary. We will
give this issue further attention in Section 4.
B. Evaluation of the Unknown Amplitudes Kqj}
By the construction, the simulated scattered field Es satisfies the radiation and the periodicity conditions. Evidently,
if a set of periodic patch currents Mqijcould be found such
that the boundary condition [Eq. (3)] was strictly satisfied,
then Es would be the exact field scattered by the grating. To
obtain an approximate solution, the boundary condition is
imposed at M selected points on the boundary S. This
reduces the functional relation [Eq. (3)] to the matrix equation
[Z]K = V,
(9)
(10)
J
2
Mqj = tqjKqj(z
z)exp[j
k'c
(r-r)]
fPSP)
p=1 n=-'
(5)
(11)
where
with {ipn
[Z1
ZI
1
(6)
6
22
[]
(7)
V2]
(8)
tpm
component of the
(p = 1, 2) denote M-element
(12)
Boag et al.
(16)
f(X)
1.0
1715
and
2
kZmn = (k - kTmn kTmn)'
0.8
12
(17)
0.6
0.4
0.2
-0.50
0.50
0.25
0.00
-0.25
-kZmn are, respectively, the z components of the wave vectors of the z and -z traveling mnth Floquet modes. Thus, in
are used depending on whether z > zi or z
or k
Eq. (14), k
(a)
/2
fps
I
ap1 =
J-s,,/2
p = 1, 2,1 e ZZ
Vsp)exp(1bcp)d,
0-
(18)
-30
-60
-90
converge faster. However, the ability of the sources to simulate rapid field variations will obviously be less effective.
-120
-150
-150
-50
-100
100
50
150
which is valid for observation points r in the z < -b halfspace, results by substituting Eq. (14) into Eq. (13) and Eq.
(13) into Eq. (4), by using the inequality
k
(b)
Fig. 3.
(19)
i,
Z < Zi
and subsequently by interchanging the order of the summations involved. One obtains
Fig. 3(b), the Fourier transform of the Blackman-Harris
window is characterized
dB).
Eqi(r) = -V X Fqi(r),
where Fqi is the electric vector potential represented
series of Floquet modes:
= aqi
>
2 A
m=-
n=-O
exp[-jk'
as a
(r - ri)]. (14)
1 m
with
kZmn,
r),
(15)
(20)
alma2n 2
=
2k-
Zmn
q1
Kqjk-n
n
X
X
(21)
representation for the scattered field Es in the z < -b halfspace in terms of an infinite discrete set of Floquet modes.
We refer to this field as the reflected field.
Similarly, in the half-space below the lowest point of the
scatterers an alternative representation in terms of Floquet
modes for the total field E, valid for observation points r in
the z > b half-space, is obtained by using the inequality
(22)
Z > Zi V i.
Z Z Em.exp(-jkn
m=-X n=--
material that is identical to that composing the region surrounding the bodies. The electric field Eqiat an observation
Fqj(r)
E- =
The result is
E+ =
m=-= n=-X
Emexp(-jk
*r),
(23)
1716
Boaget al.
where
V
p
=
mObnOEic
alma2n 2
2kZ n n
=1
Kq=k1
exp(jkm
- ri). (24)
As a numerical study case, we consider the problem of diffraction from a doubly periodic orthogonal array comprising
perfectly conducting spheres. Some representative computations obtained with a computer program developed by
using the preceding formulation are given and compared
with an analytic approximation. Unfortunately, we were
not able to find appropriate data in the literature for comparison with our results.
As a first step, we define the accuracy criteria that will aid
us in evaluating the numerical results. In our solution we
force the boundary condition to be obeyed only at a finite
number of selected points on the boundary. Naturally, one
can question the behavior of the fields on the boundary
between the matching points because it can, in general, be
quite different from what is required by the boundary condition, thereby rendering the results inaccurate. Toward this
end, we carry out a study of the convergence of the boundary
condition error AEb, defined by
+ Einc)I
AEb = 1bc=
hx
X (Es
Ei sI
nc l
on S.
Pmn
mn
=p
ZP
E
kvj
ffic 2)
-20
CS
0
Cs
0.1
0.3
0.5
0.7
r r
Fig. 4. Plots of max(AEb,) and APversus r/r for the case of a plane
wave normally incident (inc = inc= 0) upon an orthogonal array
with periods d = d 2 = 0.8X composed of perfectly conducting
spheres of radius r = 0.2Xobtained with N = 44 patch-current
sources per sphere.
-20
-40..
(26)
,
inc5 + cos
Zfinc
cos Zinck + sin Zincsin
mn
pine
(28)
are the power flowsper unit area of the mnth Floquet modes
in the -z direction for reflected modes and in the +z direction for transmitted modes. Here, q is the intrinsic impedance of the medium. The power-conservation check by no
means totally validates the solution. However, we surmise
that a solution of a relatively large number of unknowns,
which satisfies not only the boundary conditions but also the
12
1E
(25)
Wnlnax
pinc
(29)
NUMERICAL RESULTS
_ kzmn
mn
(27)
'AP
xr
-60
max(^EC)
-80
10
.... 11.111-1.1.1
30
50
70
IT
1717
Opt. Soc. Am. A
~~~~~~~~~~~~~~Vol.
etal.7, No. 9/September 1990/J.
Boag
Boag et al.
/pinc,
p: 1 0 /pinc ()()
p~o/pdnc
pt 1 0 /pinc
150
300
5.47
43.14
20.55
30.84
0.3 X 10-2
0.08
59.80
20.12
14.85
0.9 X 10-3
0.08
00
0 inc
85.6145
%)-
14.3855
0.8 X i0-5
0.02
P-/pinc (%
AP' (%
MaX(AEbN) (%
450
5.35
5.23
65.28
19.82
9.55
0.1 X 10-3
0.05
a Plane waveof wavelengthXincident in the xz plane ((pin, = 00 at various angles oin upon a grating with periods di
Spheres of radius r = 0.2X obtained with N = 54 patch-current sources situated on spheres of radius rt = 0.2r.
7 0
mentioned above, previous studies -1 have shown that it is
advantageous to place the sources on surfaces similar in
8 5
sources Mqi are placed on a concentric sphere of radius r 1,r
<r. We use 2N patches of dimensionss S~ 2 =0. 135X and
exponential decay of
angles.
Also
shown are the power conservation error AP' and the maximum-boundary-condition error max(AEb,).
It should be emphasized that, for any calculated quantity
of interest, one should examine the numerical convergence
by comparing the results for an increasing number of sources
and match points. If the computed results are sufficiently
close,it can be assumed that a satisfactory accuracy has been
achieved. To validate the solution, one should at the same
time check the behavior of the error in the boundary condition between the matching points and examine the decrease
in the power-conservation error. These checks are easily
executed with summations of analytic terms. As explained
earlier, neither of these checks totally validates the result.
However, they can undoubtedly
d2
4.67
4.55
71.14
21.75
2.44
0.2 X 10-2
0.04
63.47
23.90
8.08
0.4 X 10-2
0.03
source location, the solution converges. The rate of convergence might be sensitive to the location of the sources,
though usually not heavily.
Figure 6 shows plots of the 0 component of the surfaceinduced current J givenby
= At X
(Hinc+ H )
on S
(31)
X d21)
pinco Oiclcs O0 I~n(kldl
2
Figure 7 shows a plot of ra(O, 0)/[Icos 0I(kdld 2) ] obtained by
1J.1/IHncI
2.5
2.0
1.5
(30)
d,1 ,=0.8X
be added that, even for choices that are less than optimal in
from a
750
----d12=1.52.
1.0
. . . dl,2 =-
0.5
0.0
0
45
90
135
180
0 (eg)
Fig. 6. Magnitude of induced electric surface current Jo versus 0
for gratings of various periods and for the single-scatterer case.
1718
Boag etal.
1990
0.15
0.10:
Y. Leviatan's present address is the Department of Electrical Engineering and Computer Science, George Washington University, Washington, D.C. 20052.
0.05-
REFERENCES
0.00
I
|........
45
90
135
180
0 (deg)
* n*c *
Va lues of reflection and transmission efficiencies p/pnc
at discrete mode angles (asterisks) and the corresponding plot of
Fig. 7.
7ra(O,OVO[cI
os 01(kdid2 )2] versus 0 for P= 0 (solid curve) for the case
of normal ii aicidence(inc = inc = 0) upon a grating with periods di
0.2X.
symmetry,
it is sufficient
from 0 to 180
the left-ha
that the continuous curve nearly coincides with the calculated discrete values,
5.
DISCI USSION
The
numerical
convergence
and accuracy
1405 (1987).
4. C. C. Chen, "Transmission of microwavethrough perforated flat
plates of finite thickness," IEEE Trans. Microwave Theory
Theory,