TRANSACTIONS
ON MICROWAVE
THEORY
AND
TECHNIQUES,
VOL.
1609
IEEE
Trans. Znstrum.
<The measurement
anisotropy
[43] [44]
[45]
microwave open resonator J. Phys. D. Appl. Phys., vol. 9, pp. 819-827, 1976. K. H. Breeden and J. B. Langley, Fabry-Perot cavity for dielectric measurements: Reu. Sci. Inst., vol. 40, pp. 11621163, 1969. dielectric measurement R. J. Cook and R. G. Jones, Precise techniques for the frequency range 10 GHz to 150 GHz7 in Proc. 8th Eur. Microwave Conf. Paris (Sevenoaks: Microwave Exhibitions and Publishers), 1978, pp. 528532. R. H. Stolen, Far infrared absorption in high resistivity GrrAs~ Appi. Phys. Letts., VO1. 15, pp. 74-75, h.dy 1969.
ods such as Fourier-transform spectroscopy and laser spectroscopy. He has rdso made theoretical studies of various systems such as water, alcohol, and hydrogen bonded molecules, criticaf mixtures, polymers, semiconductors, and common gases and pollutants. He was also responsible for the European Economic Community (EEC) funded Inter-European project on standard reference materiaf. He was awarded the 1977 Duddell Premium (prize) by the Institution of Electncrd Engineers, London (IEE), for one of his outstrmding publications. In 1978, he joined the Massachusetts Institute of Technology (MIT), ,Cambridge, where he is a Senior Scientist and Pnnciprd Investigator of severaf research projects. His main line of research at the MIT Nationaf Magnet Laboratory is millimeterand submillimeter-wave techniques and precision measurements and the theoretical study of complex refractive index and dielectric permittivity, and magnetic permeability of condensed materials and window design for high-power gyrotron ~ource and rnagnetoopticaf study and impurity characterization of gallium arsenide, iridium phosphide, and ternary and quaternary compound seryiconductors. He is one of the organizers of the annuaf Intemationaf Conference on Infrared and Millimeter Waves. In September 1984, he took a new position as full Professor, Department of Electncaf Engineering, at the City College of the City University of New York and Graduate School of the City University of New York. He has published over one hundred papers, including severrd book chapters, and has established himself as a world expert in the microwave, millimeterwave, micro-electronics fields. He routinely presents invited talks at many intematiorud conferences.. Dr. Afsar is a Chartered Engineer, England, and a Member of the Institution of Electrical Engineers, London (IEE), and a Member of the Institute of Physics, London, and a member of the American Physical Society.
London, his research field wa+s on tropospheric propagation. From 1972 until 1978, he worked at the prestigious Nationaf Physical Laboratory (NPL) of England, in the Division of Electncaf Science, where he was responsible for the development of millimeterand
,.
for
ALLEN
W. GLISSON,
MEMBER, IEEE
MEMBER,
IEEE,
JOSEPH
JAMES,
STUDENT
Abstract resonant
Electric-
for five
we present results for the computed modes of an isolated by applying Numerical this formulation. procedure of implementation
are dkplayed
of several obtained
resonator
is discussed,
is described.
is described.
instabilities
I.
INTRODUCTION
higher order inodes am studied if the moment matrix is not adequately normalized. It is shown that a straightforward normalization can significantly number of the matrix [2], and reduce the numerical Traditional difficulties. of the method of moments to applications improve the condition consequently remove or
HE resonant frequencies and the Q factors of various modes in isolated dielectric resonators can be accucomputed by using the surface integral equation in [1]. In for bodies of revolution as described
rately
formulation
Manuscript received May 3, 1984; revised July 25, 1984. This materiat is based upon work supported by the Nationaf Science Foundation under Grant ECS-8304442. D. Kajfez and A. W. Glisson are with the Department of Electrical Engineering, University of Mississippi, University, MS 38677. J. James is with South Carolina State College, Orangeburg, SC 29117.
problems involting perfectly conducting bodies have been either an ~-field or an H-field integral equation [3], [4]. In either case, all the components of the unknown column vector are automatically expressed in the same physical units for most structures so that the normalization pf the. matrix is not an issue. The integral equation utilized here, @198~ IEEE
0018-9480/84/1200-1609$01.00
1610
however,
is of the combined-field
both the electric and the magnetic equivalent currents appear as unknowns. In such situations, and when different types of basis functions nents of current, sides A of the units. computational procedure for evaluating the and magnetic-field distributions in and around the matrix way that all the components matrix physical electricare used for different should be normalized of the column are reduced to compoin such a the same
vectors on both
equation
convenient
dielectric resonator is described in the second section. It is shown that the components of the modal field vectors can be computed using a slight modification of the existing algorithm for computation of the matrix elements field in the
Fig. L Orientation
!T
J+ J+
Jt
J+
-1
method-of-moments In the last part butions played. electric for A various detailed
procedure. of the paper, resonant knowledge the computed modes distridisof the should are graphically the resonator
of equivalent
currents nator.
on surface
of dielectric
reso-
of the orientation
vector
equivalent
currents
(and Ilt),
field around
densities)
in four column
vectors
be very useful in designing the coupling circuits for particular modes, as well as in designing the traps for undesirable modes. symmetric Only the field distributions on dielectric of the rotationally [7], [8]. TEOla and TM018 modes have been previously resonators
lKt), and Pf+). The homogeneous yields the numerical the complex However, method, when
solution of the matrix equation values of the natural frequencies of points numerical N is gradually convergence instabilities
plane which belong to various resonant the number to study the numerical currents the matrix
creased in order MATRIX NORMALIZATION integral equations procedure which utilized serve as a here are puting This
point
in the solution
indicates
ill-condi-
obtained from the boundary condition that both the tangential electric and the tangential magnetic fields are continuous at the surface of the dielectric resonator. When the unknown equivalent surface currents are expanded in pulse basis functions and the integral equations are tested in the manner described in [1], a matrix equation of the following form is obtained:
tioned for large N. The reason for the ill-conditioning can be traced to the fact that the individual components of both column vectors appearing in (1) are expressed in mixed physical units. For instance, lr is given in amperes, J@ in amperes per meter, K, in volts, this situation, impedance and M@ in volts per meter. In order to correct If and J+ are both multiplied q, of the dielectric qr= material 120n 6 In addition, resonator. primed It is divided are then by 2wa, the circumference column of the by (2) by the intrinsic
II*)
IJJ
IK,)
. (1)
The new normalized
P%)
Subscripts azimuthal @ and t in (1) denote vector components direction and in the direction along in the the gener-
vectors, denoted
symbols,
IJ,) =
&t)
(3) (4)
ating curve for the body of revolution, respectively. Fig. 1 illustrates the orientation of the components of equivalent electric surf ace currents J1 and J+ on various parts of the dielectric resonator of cylindrical shape. The integrals appearing in the evaluation of elements of the moment matrix are simplified if one multiplies the ~ and Mt components of the unknown equivalent currents by the factor 2np, where p is the radial distance to the current source in a cylindrical system of coordinates, and then treats this product as the unknown quantity in the The new primed physical tioned units, matrix ables requires equation. By doing this, the physical dimension of these two variables is changed from current density to current. Thus, the t-components of the electric and magnetic currents appearing in (1) are denoted It and Kr. The column
IJJ) = TrlJ+)-
We note that the new vector IJ,) has units of current density, but it is not the actual equivalent surface current density because of the factor p/a. The vector IM+) is left unchanged, and the vector ]Kt ) is normalized by dividing by 2wa
P-G) +-#). =
namely that volts/meter. This change
(5)
symbols are now all expressed in the same of varithe corresponding parts of the partiby the same
in (1) be divided
(or multiplied)
KAJFEZ
et al.: COMPUTED
MODAL
FIELD
DISTRIBUTIONS
FOR RESONATORS
1611
fnmun
(a) Fig. 2.
Matrix areas affected by normalization: (a) division by q,, (b) multiplication by 2 na, and (c) multiplication by q..
Therefore,
the first
two block-columns
of the ma-
by q, as indicated
block-columns
by 2~a as indicated in Fig. 2(b). To bring the right-hand side of (I)tothe same physical units, all the magnetic-field quantities are multiplied with the free-space intrinsic impedance qo. Consequently, the third and the fourth block-rows of the matrix must be multiplied bythesame factor as shown in Fig. 2(c). The condition been computed number before based on the infinite and after with norm [2] has indion the resoand the normalization 27 points the matrix
++-=++-$+
(a) $.
cated above. For the mode HEMIZ$ body (resulting number nator dimensions in a 102 x102 matrix), was reduced by a factor
condition
the relative dielectric constant is t,= 38. The effect of bringing all the elements vector figure, to the same units the values of the individual in the column values plotted. currents
computed without matrix normalization and in Fig. 3(b), the currents computed are plotted. One can clearly see that
currents are so dissimilar in magnitude that, the equivalent current Z, is not even visible in After ,normalization, all the equivalent are of the same order of magnitude, as
Fig. 3.
t-J-t-JPt-+t--H
(b) Magnitude of modal equivalent currents for HEM12$ mode: for unnormalized matriz, and (b) for normalized matrix. (a)
can be seen in Fig. 3(b). III. Once a resonant been determined, can be computed A convenient COMPUTATION OF FIELDS frequency the modal of the dielectric equivalent resonator has
form [ES(.I, currents constant. where (E, [HS(J, H) M)M)E(H(J, -M)]tm=E:: J, -kf)]tw=H& are the fields evaluated (6) (7) just
surface
and (E, H)
numerical
this is Gaus-
sian elimination. The resulting electric and magnetic equivalent surface currents on the elementary bands on the resonator surface can be utilized to compute the electric and magnetic fields everywhere in space. However, the vector summation of the field contributions for various observation points inside and around the resonator requires a considerable additional programming effort. This work can be circumvented by utilizing the information already available in the subroutine which computes the elements of the matrix on the left-hand side of (l). A convenient procedure can be developed for computing the fields if one notes that the surface integral equation approach comprises a difference of the two fields of interest at the resonator surface since the equations are of the
outside and just inside the resonator surface, respectively. In particular, a single element of the moment matrix represents the difference in two field components at some point in space due to two different unit current sources (e.g., from basis functions for J and J) located at another poirit media. in space, but which radiate in different The two radiated individually, fields appearing homogeneous in (6) and (7) are
computed
region can be computed by retaining only the appropriate terms in the moment-matrix calculation. This is easily accomplished by retaining only the set of potentials where the medium parameters are those of the medium in which the field is desired. To compute the fields, therefore, we first specify the generating arc for a phantom surface on which we wish to
1612
IEEE
TRANSACTIONS
ON MICROWAVE
THEORY
AND
TECHNIQUES,
VOL.
Mrr-32,
1984
TABLE
Mode
(GHz)
Q
45. s 76.8 30,7 52.1 327.1
iii
~ol
01 & 6
21
tric constant:
fields. The fields tangential to this sources on the resonator surface are
In the above, we use tiW.P to represent only the imaginary part of the complex natural frequency of the mode (nz, n, p). A computer-generated to show the field orientation graphical at equidistant display points is used as well
next computed using a modified version of the momentmatrix routine which includes only the potentials involving the Greens be evaluated. previously tangential function The computed fields of the medium resulting modal matrix current in which the field is to ii multiplied by the the solution to obtain
as to provide some relative amplitude information. The plane containing the z axis (axis of rotation) is referred to as the meridian passing through as the equatorial resonator having mm, and which resonant plane, plane. and the plane All the plots perpendicular are computed to it, to for a the center of the resonator, is referred
on the generating
surface. The fields so computed are actually weighted averages, since (6) and (7) were tested as described in [1] to obtain expressions for the moment-matrix elements. The appropriate values for the fields are easily obtained, however, by dividing subdomain the phantom propriate by the length of the phantom Fields by including where obtained surface on the apm is the via this distriat the observation surface cos(m+) or sin(nz$) point. anywhere
the dimensions a = 5.25 mm and h = 4.6 is made of material with c,= 38. The and the Q factors (due to radiation) the results frequencies differ by as
frequencies
are listed in Table I. When compared with presented in [1], it is noticed that the resonant are almost much factors because identical, but some of the Q factors The results presented to be more accurate of points as 12 percent. are believed a larger
are obtained
behavior,
azimuthal mode number. For the mode TE018, field distributions procedure have been compared with bution for a dielectric-rod waveguide tion is available in terms of Bessel observes in Fig. 4 that the agreement the resonator. The field of the isolated decays section more slowly than of the dielectric-rod waveguide
number
the theoretical
for which the solufunctions [9]. One is quite good inside resonator, however, for a resonant by two terminated
model the generating arc of the resonator. Fig. 5 shows the electric field of the mode TE018 in the equatorial plane at the moment a~HP t = O. In this and in subsequent illustrations, double arrows indicate the area within which the field is less than 3 dB below the maximum, while the longer lines indicate a level between 3 and 10 dB below the maximum, and the shorter lines indicate a level between 10 and 20 dB below the maximum. When the transverse field is more than 20 dB below the maximum value of the field, the points are left blank. The magnetic field of the mode TE013 in the meridian plane is shown in Fig. 6. The magnetic field is perpendicular to the field shown period patterns in Fig. in 5, and time. its maximum In general, all with occurs one-quarter later the magnetic-field all the electric-field indicated
parallel magnetic walls. This behavior is to be expected since the magnetic walls form a cutoff waveguide in the radial direction IV.
In this section,
present for
catalog
dielectric
computed
are in time
quadrature
F (electric
or magnetic decaying
patterns. Therefore, this fact will in the remaining illustrations. The magnetic and electric field
not be further
field
is ignored,
spatial F.
distribution we
magnitude instanta-
Therefore,
display
be seen in Figs. 7 and 8 in the equatorial and meridian planes, respectively. The magnetic field of this mode is well
KAJFEZ
et al.:
COMPUTED
MODAL
FIELD
DISTRIBUTIONS
FOR RESONATORS
1613
Fig.
7.
TM018 mode,
H-field
in equatorial
plane,
TE018
\\\\ -.\\ -.\<\ -~., / / I \
\\\ll \\\\ \\ll \\\\l \\\ll .\\\ \\\\ \\\llj(l(llJu \lj$ lJu#(l \ll
Fig.
6.
TE018 mode,
H-field
inmeridian
plane,
~t=r/2,
\\\\ \\\\ \\\lll/ ////// TM018 E s\\\ \\\tll/ !l ////,,- ////// -..\\\\\\ \l ///,,, ,~-..\\\\\ ---, ,, ,,0 , ,..- ---..\\/ ,---~-. --+ ///l /--\\ \I// /O%\\\\ ///1//\\\\nl///\ l\\\/ \ \\\ \\\l\\///lll / 111 \\\\ \--///j\\ ,---~-. /// , .-- #///\ \--/ \.---, a,, \\ -4,,,,///1 \\ \\\\..-\---.////// /l\\\\\\ >.//// ///l\l\ \\\\>\ ///////n\\ \\\\\\ //////ill\\\ \\\
\\ll l//// //
Fig.
8.
TM018 mode,
E-field
in meridian
plane.
the resonator, whereas the outside electric strong near the top and bottom faces of If strong coupling to an external circuit is probe, directed along theaxisof this mode. here are
\\\l \\\\ \l
a short capacitive
for couplingto
HEM118
\ll/ l/// l/////
three resonant
modes displayed
modes (HEM)
acquire either a cos(m+) or a sin(m+) dependence. For this reason, the angular reference +=0 is indicated to specify the orientation of various field patterns with respect to each other. The hybrid HEMlla mode with the lowest resonant frequency is in
/////n\\\\\
@
-klllllllt{ . .. .. .. lo \\\-\\\\\ ~J\\
,.. ,.
12. As mentioned
in [1],
-.-
the free-space simulated environment. Table I shows that the Q factor of this mode is, lowest of all the modes, which makes it very difficult to achieve sufficient coupling to the coaxial cable leading to the observation instrument. This mode has been recently utilized by Long et al. [11] in the so-called resonant cylindrical dielectric cavity antenna.
/ill
/n\
l\\\
Fig.
9. HEM118
mode,
H-field
in equatorial
plane
1614
TECHNIQUES,
VOL.
1984
HEM118
E
I
I /
I I
I
I \ \
I
I \ \
I
I I \
/
I I 1
/
/ I I
///<.--.\\\\
////.--.\\\\ / I ///.--.\\\ /// \\\lllillt// \
\
\ I
\
I I
\
I 1
I
J /
I
I // f
11/////// \\\\\\\\
l//\\l
\\\\...
I!fl // / f I / I I n
\\\
Fig.
10.
HEMIIO
to and
offset
from
Fig.
13.
HEM128
mode,
E-field
inequatorial
plane.
,/ HEM118/ //0------ \\\\\\\\ ~// / ////////--<% N<\\\\\\ E llllli l/////O- >\\\\\\j \\~\\\ //////f/////-=<*%\\ \\\\\\\\\\ d@I 1/1I I I / ./eeeG-=e\. \ 1I I Ill 1 1111111 / I I I Ill
///0/----..\\\\ \\\\
..
I
-
I I I I I I I I I I I I I I 1111111 1 I 1 I I I i
/////===-*-\\\\
j / / ////-..\\\ \ \
l\\\\\---
/////
, ////////// //fllj\[jj / // .--##,,///////
, , , , / , , , ,.>,
___
_ -
-=>=.=
=__
__
---
.-_>~=,==___ -
-----.%.
--.
..-
-_____ / -----
-,
.../
,,,
/ /
/
/
I
/
/
,~-.,
,/-., ,
N...... \ \ ,,,,,,.
, ,,,,,
,/,,/
/ /
/ /
/ /
I /
I /
I i / I II
///////ff/ ///// /
\\\\N
1 /
\N\\\
////11
Fig. 14. HEMIz~ mode, E-field
\\\\\\
in meridian plane +=0.
Fig.
11.
HEM118
\
---
.--,
HEMII$
mode,
E-field
inmeridianplane+=O
//
//
--/
II \ / / I \ / j I / I
\ +
\\
\ \ \ \
Fig.
12.
HEMl10
mode,
H-fieldinmeridianplane@=w/2.
Fig.
15.
HEM12&
mode,
H-field
inmendianplane
+=r/2.
The magnetic field in the equatorial plane can be seen in Fig. 9. The corresponding equatorial plane componentsof the electric field are zero, notwithstanding the time quadrature. This occurs because the E-field pattern has an odd symmetry E-field parallel distance with shown with respect to the equatorial in Fig. the equatorial plane, but plane. Thus, by the a 10 has been computed in a plane
face, but still inside the resonator). The meridian which contains the maximum electric-field intensity
plane is the
@= O plane (Fig. 11), whereas the maximum magnetic-field intensity occurs in the meridian plane @=7r/2 (Fig. 12). Note that the magnetic field is very weak outside each resonator face, while the electric field is strongest while there. a This property should be kept in mind coupling device for this mode. designing
displaced
2.15 mm from
RAJFEZ
et (il.:
COMPUTED
MODAL
FIELD
DISTRIBUTIONS
FOR RESONATORS
1615
HEM218
\\\ \\\\\ ..\
H
I / // \ I /
HEM218
\\
-..
I \
.u.=ee
--m== --------
. - . . .
\
---
**=--/.-
./
~-----
//
//
0--
.,,,
\\
Fig.
/ I \ \\--/ / I n,\ / I / I Ill \ \ \\\ \ \
18.
HEMJM
mode,
H-field
in meridian
plane + = m/4.
Fig.
16.
HEMz10
mode,
H-field
inequatonalpkme.
. .,,,,\\\\ HEM218 //////0E .%NN \\\\\\ 11/////0 \\\ 1111/11//// ~0- -. s\\\\\!\\ -.x\\\\\\\\\ \\ llllli l\\\\\l 1/////1I l////~/- /---\ I I / I / I ///-==-----=+\\
/////OH ., \\\\\\\ llll\ll
-%\%N$ *\\N~y\llllll 0//llfq -0/##it/111111 L-&.#// -./////\\\\ #.,,//f[~l a///////// -//,,,//// //8//////
D
q
\. .-./
-.
I/#@@/f \\tiah\\ .-ee~, / \ .**s-. .N=ew.,,,eg?e-. \\\k\\\ //fl@#// \\\\\ <%\l//-z, , / / \\\\.\\\\\ /////
+=cJ
field H-field
Fig.
19.
HEM21$
mode,
E-field
in meridian
plane o = O.
differ
by only
3 percent.
Therefore, of interest
m~de is mainly
an effective
pattern respect
consisting
cient radiator, and consequently, the Q factor of this mode is much higher than of any other mode listed in Table I. The electric-field distribution again has an odd symmetry
Fig. 17. HEM216 mode, E-field inplane paralel toandoffset equatorial plane by 2.15 mm. from the
about
the equatorial
plane. Therefore,
the E-field
in Fig. 17 because, at
is shown Figs. 13 through a resonant HEM118 frequency (see Table 15 depict themode only 5 percent I), Fig. 13 depicts HEM128, higher theelectric which has field in
the equatorial plane, the transverse E-field is zero. The H-field in the meridian plane is maximum at + = 45. Fig. 18 shows that the magnetic field is strongest plane near the equator. The E-field in Fig. 19. in the meridian rp = O is shown
the equatorial plane. In Fig. 14, which shows the E-field in the meridian plane, one notices the strong localized field at the four corners of the resonator. One can expect that, when the resonator is placed face down on a microstrip substrate, citive this mode will be strongly field coupled to the adjaaction (capain the cent microstrip coupling). line through the electric-field
V.
The electric qualitative dielectric numerical bodies data of of solution
CONCLUSIONS of the integral equation quantitative practical symmetric design resonant for diand of
The magnetic
meridian plane + = 7r/2 is shown in Fig. 15. The last four illustrations (Figs. 16 through This mode has resonant mode HEMZ18.
resonators. (m
circularly
modes
= O) are classified
as TE0.8
and TMO.O.
The
frequency close to the resonance of the TM018 mode, and if the TMOl~ is the desired mode of operation, HEMZ18 creates a spurious resonance nearby. From Table I one can see that, in the example chosen, the resonant frequencies of
other resonant modes (m > 1) are all of the hybrid nature, The resonant frequencies of various denoted HEId~.a. modes are sometimes located very closely together. In order to design the coupling circuits which will enhance the desired resonant mode and suppress the undesirable ones,
1616
TECHNIQUES,
VOL.
Mn-32,
distribution It will
.. --~
.~
is hoped provide
are presented
[1]
[2]
[3] [4]
A. W. Glisson, D. Kajfez, and J. James, Evaluation of modes in dielectric resonators using a surface integral equation formulation, IEEE Trans. Microwave Theo~ Tech., vol. MTT-31, pp. 1023-1029, Dec. 1983. C. Klein and R. Mittra, Stability of matrix equations arising in electromagnetic< IEEE Trans. Antennas Propagat., vol. AP-21, pp. 902905, NOV. 1973. R. F. Barrington, Field Computation by Moment Methods. New York: Macmillan, 1968. Integral equation solutions of A. J. Poggio and E. K, Miller, three-dimensional scattering problems, in Computer Techniques for Electromagnetic, R. Mittra, Ed. New York: Pergamon, 1973, ch. 3. H-field, E-field and combinedJ. R. Mautz and R. F. Barrington, field solutions for conducting bodies of revolution, Arch. Elek. Ubertragung., vol. 32, pp. 157-164, Apr. 1978. Electromagnetic scattering J. R. Mautz and R. F. Barrington, ffom a homogeneous material body of revolution, ,4 rch. Elek. Ubertramnz.. vol. 33. VD. 7180. Feb. 1979. M. Jaw~rs~i and M. W: Pospieszalski, An accurate solution of the cylindrical dielectric resonator problem, IEEE Trans. Microwaue Theo~ Tech., vol. MTT-27, pp. 639-643, July 1979. P. Guillon, J. P. Balabaud, and Y. Garault. TM.,. tubular and cylindrical dielectric resonator mode, in IEEE MT?~S Int. Microwave Symp. Dig., June 1981, pp. 163166. C. C. Johnson, Field and Wave Electrodynamics. New York: McGraw-Hill, 1965. D. Kajfez, Basic principles give understanding of dielectric waveguides and resonators,; Mi;rowave Systems News, vol. 13, pp. 152-161, May 1983. S. A. Long, M. McAllister, and L. C. Shen, The resonant cylinIEEE Trans. Antennas Propagat., drical dielectric cavity antenna, vol. AP-31, no. 3, pp. 406412, May 1983.
panies Iskra and Rudi Cajavec in Yugoslavia, primarily in microwave radios and radars. From 1963 to 1966, he was a Research Assistant at the Electronics Research Lab., University of California, Berkeley. In 1967, he joined the University of Mississippi, where he is presently Professor of Electrical Engineering. His research and teaching interests are in electromagnetic theory and its applications to microwave circuits and antennas.
[5]
[6]
[7]
Allen W. Glissorr (S71-M78) was born in Meridian, MS, on June 26, 1951. He received the B. S., M. S., and Ph.D. degrees in electrical engineering from the University of Mississippi, in 1973, 1975, and 1978, respectively. From 1973 to 1978, he served as a Research Assistant in the Electrical Engineering Department of the University of Mississippi. In 1978, he joined the faculty of the University of Mississippi, where he is currently Associate Professor of Electrical Engineering. His current research interests include the development and application of numerical techniques for treating electromagnetic radiation and scattering problems. Dr. Glisson is a member of Sigma Xi, Tau Beta Pi, Phi Kappa Phi, Eta Kappa Science. Nu, and Commission B of the International Union of Radio
[8]
[9] [10]
[11]
*
Darko Kajfez (SM67) He received the Dipl. University of Ljubljana, was born in Delnice, Yugoslavia, in 1928. Ing. degree in electncaf engineering from the Yugoslavia, in 1953, and the Ph.D. degree from
Joseph James (S84) was born in Chenappady, India, on May 25, 1956. He received the B.S. degree in electronics and communication engineering from the University of Kerala, India, in 1977, and the M.S, degree in electrical engineering from the University of Mississippi in 1984. From 1978 to 1981, he worked as a Research Engineer with the Electronics Corporation of India Limited, Hyderabad. He is currently with South Carolina State College, Orangeburg, SC.