Table Of Content~IEEE TRAN SACTI 0 NS ON
MICROWAVE THEORY
AND TECHNIQUES
OCTOBER 1993 VOLUME 41 NUMBER 10 IETMAB (ISSN 0018-9480)
A PUBLICATION OF THJE IEEE MICROWAVE THEORY AND TECHNIQUES SOCIETY
SPECIAL ISSUE ON QUASI-OPTICAL TECHNIQUES
Guest Editors' Overni w ................................................................... .J. W. Mink and D. B. Rutledge 1661
PAPERS
Overview
Focal Plane Imaging ystems for Millimeter Wavelength (Invited Paper) ................................................ .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P. F. Goldsmith, C.-T. Hsieh, G. R. Huguenin, J. Kapitzky, and E. L. Moore 1664
Gaussian Beams
Long Wave Optics ......................................................................... D. H. Martin and J. W. Bowen 1676
Gaussian Beam-Mode Analysis and Phase-Centers of Corrugated Feed Horns ............ R. J. Wylde and D. H. Martin 1691
Mode Conversion at Diffracting Apertures in Millimeter and Submillimeter Wave Optical Systems .................... .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . J. A. Mwphy, S. Withington, and A. Egan 1700
Slot-Fed Higher Order Mode Fabry-Perot Filters ................................... J. McCleary, M.-Y. Li, and K. Chang 1703
Gaussian-Beam en Resonator with Highly Reflective Circular-Coupling Regions ..................................... .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . T. Matsui, K. Araki, and M. Kiyokawa 1710
Quasi-Optical Antennas and Waveguides
Tapered Slotline Antennas at 802 GHz ..................................................................................... .
. . . . . . . P.R. Acharya, H. Ekstrom, S. S. Gearhart, S. Jacobsson, J. F. Johansson, E. L. Kallberg, and G. M. Rehei::. 1715
A Hybrid Dielectric Slab-Beam Waveguide for the Submillimeter Wave Region ......................................... .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .J. W. Mink and F. K. Schwering 1720
A Planar Wide-Band 80-200 GHz Subharmonic Receiver ................................................................ .
. . . . . . . . . . . , ..................... B. K. Kormanyos, P.H. Ostdiek, W. L. Bishop, T. W. Crowe, and G. M. Reheiz 1730
Double Slot Antennas on Extended Hemispherical and Elliptical Silicon Dielectric Lenses ............................. .
. . . . . . . . . . . . . . . . . . . . . . . . . . . : ................................ D. F. Filipovic, S. S. Gearhart, and G. M. Rebei::. 1738
An Improved Solution for Integrated Array Optics in Quasi-Optical Millimeter and Submillimeter Waves Receivers:
The Hybrid Antenna ................................................................................... T. H. Buttgenbach 1750
Active Grid Arrays
A 100-Element HBT Grid Amplifier ............................................................... M. Kim, E. A. Sovero,
........... J. B. Hacker, M. P. Delisio, J.-C. Chiao, S.-J. Li, D.R. Gagnon, J. J. Rosenberg. and D. B. Rutledge 1762
/{ 6.5 GHz-11.5 GHz Source Using a Grid Amplifier with a Twist Reflector ........................................... ..
. . . . . . . . . . . . . . . . . . . . . . . . . M. Kim, E. A. Sovero, J.B. Hacker, M. P. De Lisio, J. J. Rosenberg, and D. B. Rutledge 1772
Quasi-Optical VCO's .............................................................. T. Mader, S. Bundy, and Z. B. Popovic 1775
A Monolithic Diode Array Millimeter-Wave Beam Transmittance Controller ............................................ .
. . . . . L. 8. Sjogren, H.-X. Liu, F. Wang, T. Liu, X.-H. Qin, W. Wu, E. Chung, C. W. Domier, and N. C. Luhmann, Jr. 1782
Quasi-Optical Millimeter-Wave Hybrid and Monolithic PIN Diodes Switches ........................................... .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . K. D. Stephan, P.H. Spooner, and P. F. Goldsmith 1791
(Continued on back cover)
~IEEE
MICROWAVE THEORY AND TECHNIQUES SOCIETY
The Microwave Theory and Techniques Society is an organization, within the framework of the IEEE, of members with principal professional interest in the
field of microwave theory and techniques. All members of the IEEE are eligible for membership in the Society and will receive this TRANSACTIONS upon
payment of the annual Society membership fee of $20.00. Affiliate membership is available upon payment of the annual affiliate fee of $33.00, plus the
Society fee of $20.00. For information on joming write to the IEEE at the address below. Member copies of Transactions/Journals are/or personal use only.
ADMINISTRATIVE COMMITIEE
P. W. STAECKER, President E. J. CRESCENZI, JR., Vice President M. J. SCHINDLER, Secretary R. E. BRYAN, Treasurer
R.E.BRYAN R. H. JANSEN B. S. PERLMAN P. W. STAECKER J. W. WASSEL
E.D.COHEN S. A. MAAS R. POLLARD R. SUDBURY E. YAMASHITA
E. J. CRESCENZI, JR. M. A. MAURY, JR. J. E. RAUE D. G. SWANSON
D. HORNBUCKLE R. A. MOORE E. REZAK G.THOREN
Honorary Life Members Distinguished Lecturers Past Presidents
A. C. BECK T. S. SAAD W. CURTICE R. KAGIWADA (1992)
S. B. COHN K. TOMIY ASU P. GOLDSMITH F. !VANEK (1991)
A. A. 0LINER L. YOUNG F. !VANEK T. ITOH (1990)
V. RIZZOLI
J. R. WHINNERY
S-MTI Chapter Chairmen
Albuquerque: M. A. DINALLO Israel: E. LEVINE Schenectady: R. J. GUTMANN
Atlanta: A. J. GASIEWSKI Ithaca: L. PALMATEER Seattle: D. T. HARVEY
Baltimore: S. L. ANTHONISEN Kitchner-Waterloo: c. R. SELVAKUMAR Seoul: J. s. MYUNG
Beijing: Y.-R. ZHONG Y. L. CHOW Singapore: L. M. SENG
Benelux: K. VAN T. KLOOSTER Los Angeles: H. J. DE Los SANTOS South Africa: A. J. BOOYSEN
Buffalo: M. R. GILLETTE Milwaukee: J. RICHIE South Australia: B. D. BATES
Central Iowa/Cedar Rapids: D. OLIVER Montreal: G. L. YIP South Brazil: D. CONSONI
Central & South Italy: P. LAMPARIELLO New Hampshire: T. 0. PERKINS Southeastern Michigan: E. J. ROTHWELL
Central New England/Boston: R. T. WEBSTER New Jersey Coast: T. FITCH Spain: M. P. SIERRA
Chicago: T. G. JURGENS New South Wales: C. D. WILSON Springfield: K. M. LAU
Cleveland: D. A. FORCE New York/Long Island: F. J. WINTER St. Louis: R. INDECK
Columbus: T.-H. LEE North Italy: C. N. NALDI Sweden: P. T. LEWIN
Dallas: G. A. TRUITT North Jersey: R. v. SNYDER Switzerland: L. PROST
Dayton: G. R. SIMPSON Oregon: K. SHENA! Syracuse: D. w. FERGUSON
Denver-Boulder: R. VJDANO Orlando: D. L. DUNN Taipei: C. C. TZUANG
Egypt: 0. L. EL-SA YEO Ottawa: D. H. REEKIE Tokyo: T. YONEYAMA
Finland: E. T. KEMFPINEN Philadelphia: A. S. DARYOUSH Tucson: H. c. KOHLBACHER
Florida West Coast: R. HENNING Phoenix: G. MARACAS Turkey: A. BUYUKAKSOY
France: V. F. HANNA Poland: J. W. MODELSKI Twin Cities: A. GOPINATH
Germany: F. ARNDT Portland: D. SHENA! United Kingdom/Ireland: T. H. OXLEY
Greece: N. K. UZUNOGLU Princeton: A. KATZ Utah/Salt Lake City: M. F. ISKANDER
Houston: J. T. WILLIAMS Rio de Janeiro: J. R. BERGMANN Venezuela: A. N. BIANCHI
Hungary: I. FRIGYES San Diego: A. CLAVIN Virginia Mountain: I. J. BAHL
Huntsville: H. L. BENNETT San Fernando Valier,: A. M. MADNI Washington/Northern Virginia: G. TAIT
India: S. K. KOUL Santa Clara Valley /San Francisco: W. Ou Winnipeg: G. BRIDGES
Indonesia: S. NATANAGARA Yugoslavia: B. JOCANOVIC
IEEE TRANSACTIONS® ON MIC ROWA V E THEORY AND TECHNIQUES
Editor Associate Editors
D. MASS~ R. SORRENTINO J.B. HORTON P. W. STAECKER
(Tutorial/Review Papers) (Systems and Applications) (Index Issues)
THE INSTITUTE OF ELECTRICAL AND ELECTRONICS ENGINEERS, INC.
Officers
MARTHA SLOAN, President CHARLES K. ALEXANDER, Vice President, Professional Activities
H. TROY NAGLE, President-Elect HELEN M. WOOD, Vice President, Publication Activities
SOUGUIL J. M. ANN, Secretary LUIS T. GANDIA, Vice President, Regional Activities
THEODORE W. HISSEY, JR., Treasurer WALLACE S. READ, Vice President, Standards Activities
EDWARD A. PARRISH, Vice President, Educational Activities DONALD M. BOLLE, Vice President, Technical Activities
W. KENNETH DAWSON, Director, Division IV-Electromagnetics and Radiation
Headquarters Staff
JOHN H. POWERS, Executive Director and General Manager
PHYLLIS HALL, Staff Executive, Publications
FRANK R. MOORE, Staff Executive, Volunteer Activities
RICHARD D. SCHWARTZ, Staff Executive, Business Administration
IRVING ENGELSON, Staff Directqr, Corporate Activities EDWARD ROSENBERG, Control/er
W.R. HABINGREITHER, Staff Director, Customer Service Center ANDREW G. SALEM, Staff Director, Standards Activities
PETER A. LEWIS, Staff Director, Educational Activities W. THOMAS SUTTLE, Staff Director, Professional Activities
MELVIN I. 0LKEN, Staff Director, Regional Activities ROBERT T. WANGEMANN, Staff Director, Technical Activities
Publications Department
Director, Transactions/Journals: PATRICIA WALKER
Publications Manager: GAILS. FERENC
Electronic Production Manager: JERI L. Uzzo
Managing Editor: VALERIE CAMMARATA
Associate Editor: JOSEPH PAGNOTTA
IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES is published monthly by The Institute of Electrical and Electronics Engineers, Inc.
Responsibility for the contents rests upon the authors and not upon the IEEE, the Society/Council, or its members. IEEE Headquarters: 345 East 47 Street,
New York, NY 10017-2394. NY Telephone: 212-705 +extension: Information-7900; C1eneral Mana.s.er-7910; Public Information-7867; Spectrum-7556.
Telecopiers: NY (Headquarters) 212-752-4929, NY (Publications) 212-705-7682. IEEE Service Center (tor orders, subscriptions, address changes, Education
al Activities, Publishing Services, Region/Section/Student Services, Standards): 445 Hoes Lane, P.O. Box 1331, Piscataway, NJ 08855-1331. NJ Telephone:
Information: 908-981-0060; 908-562 +extension: Controller -5365; Technical Activities -3900. IEEE Wasbin,;on Office (for U.S. professional activities):
I 828 L Street, NW, Suite 1202, Washington, DC 20036-5104. Washington Telephone: 202-785-0017. Price/Publication Information: Individual copies: IEEE
members $10.00 (first copy only), nonmembers $20.00 per copy. (Note: Add $4.00 postage and handling charge to any order from $1.00 to $50.00, including
prepaid orders.) Member and nonmember subscription prices available on request. Available in microfiche and microfilm. Copyright and Reprint Permissions:
Abstracting is permitted with credit to the source. Libraries are permitted to photocopy beyond the limits of the U.S. Copyright Law for private use of patrons:
I) those post-1977 articles that carry a code at the bottom of the first page, provided the per-copy fee indicated in the code is paid through the Copyright
Clearance Center, 29 Congress Street, Salem, MA 01970; 2) pre-1978 articles without fee. For all other copying, reprint, or republication permission, write to
Copyrights and Permissions Department, IEEE Publishing Services, 445 Hoes Lane, P.O. Box 1331, Piscataway, NJ 08855-1331. Copyright© 1993 by The
Institute of Electrical and Electronics Engineers, Inc. All rights reserved. Second-class postage paid at New York, NY, and at additional mailing offices.
Postmaster: Send address changes to IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, IEEE, 445 Hoes Lane, P.O. Box 1331,
Piscataway, NJ 08855-1331. GST Registration No. 125634188. Printed in U.S.A.
IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 41, NO. 10, OCTOBER 1993 1661
Guest Editor's Overview
M ICROWAVE and millimeter wave systems are in com importance of these developments it appeared timely to devote
mon use today and their range of applications is ex a Special Issue of the MTT Transactions to this subject. The
panding. It is safe to predict that these systems will be utilized issue is focused on quasi-optical techniques that will allow
in the future across the spectrum of endeavors from com achieving reliable, easily manufactured, low-cost devices and
munications, radar, to transportation, industrial and scientific systems for the mm-wave and sub-mm-wave bands.
applications. Satisfying this expanding demand mandates the The call for papers for the Special issue found a strong
utilization of previously unused, or little used, mm-wave and response as evidenced by the volume of this issue. The Special
sub-mm-wave bands. i.e. the use of the frequency band from Issue has been organized in five major sections. An Invited
30 GHz to 3000 GHz, in accordance with a long term trend Overview that focuses upon imaging systems employing quasi
toward systems operating at higher and higher frequencies. The optics is provided in the paper entitled "Focal Plane Imaging
necessary technology, however, is not very well developed Systems for Millimeter Wavelengths". The overview is fol
at the present time, which holds in particular for the sub lowed by four sections. The first two sections address passive
mm-wave band above 300 GHz. In addition, this technology quasi-optical components and systems. The final two sections
(as far as it is available) suffers from high fabrication cost address quasi-optical systems of active components.
and lack of convenient power sources. Component costs I. The first, addressing Gaussian Beams containing five
have been driven by the small size and tight tolerances papers: "Long-Wave Optics". Gaussian Beam-Mode Analysis
associated with the 100- 3000 GHz band and in the case of and Phase-Centers of Corrugated Feed Horns", "Mode Conver
conventional waveguide components, by the need for hand sion at Diffracting Apertures in Millimetre and Submillimetre
assembly. Power sources were largely limited in the past Wave Optics", "Slot-Fed Higher Order Mode Fabry-Perot
to vacuum tubes requiring high primary power, and, again, Filters", and "Guassian-Beam Open Resonator with Highly
restrictive tolerances. Moreover, these sources are liable to Reflective Circular Coupling Regions".
catastrophic failure. Solid state sources are more reliable but II. The next section addresses Quasi-Optical Antennas and
their output power tends to be very low at frequencies above Waveguides also containing five papers: "Tapered Slotline
100 GHz due to the small physical size of the active region, Antennas at 802 GHz'', "A Hybrid Dielectric Slab-Beam
resulting in the well-known 1/f2 fall-off of available power. Waveguide for the Sub-Millimeter Wave Region", "A Planar
Hence a need exists to combine the outputs of many individual Wideband 80-200 GHz Subharmonic Receiver", "Double-Slot
elements to satisfy the system power requirements. Antennas on Entended Hemispherical and Elliptical Dielectric
Many of the problems stated above may be resolved through Lenses", and "An Improved Solution for Integrated Array
the use of quasi-optical techniques. Quasi-optical devices Optics in Quasi-Optical MM and SubMM Receivers: the
typically have cross sectional dimensions in the order of 10 to Hybrid Antenna".
100 wavelengths and are relatively easy to fabricate. Tolerance III. This section focuses upon Active Grid arrays and again
requirements are greatly relaxed since boundary surfaces along contains five papers: "A 100-Element HBT Grid Ampli
the propagation directions of the guiding structure are not fier", "A 6.5 GHz to 11.5 GHz Source using a Grid Am
critical for mode selection and maintenance of mode purity. plifier with a Twist Reflector", "Quasi-Optical VCO's", "A
Rather, easily manufactured lenses or reflectors, and their Monolithic Millimeter-Wave Diode Array Beam Transmit
spacing between them, establish the mode parameters. In tance Controller", and "Quasioptical Millimeter Wave Hybrid
addition, the rather large transverse dimensions of quasi and Monolithic PIN Diode Switches".
optical structures allow one the freedom to include numerous IV. The final section containing eight papers addresses
solid state sources or control elements to achieve the desired Quasi-Optical Coupled Oscillators: "Nonlinear Analysis of
output power or control function. During the past several years, Phase Relationships in Quasi-Optical Oscillator Arrays", "A
significant progress has been made in the area of quasi-optical New Phase-Shifterless Beam-Scanning Technique Using Ar
techniques. New passive components such as waveguides and rays of Coupled Oscillators", "Impedance Matrix of an An
antennas have been suggested and experimental models of new tenna Array in a Quasi-Optical Resonator", "Mode Analysis
active devices have been demonstrated including distributed and Stabilization of a Spatial Power Combining Array with
mixers, frequency multipliers, phase shifters, amplifiers and Strongly Coupled Oscillators'', "Quasi-Optical Planar Arrays
power combiners. Typically these active devices include a of FETs and Slots'', "A 60 GHz IMPATT Oscillator Array
planar diode or transistor array containing many solid state with Pulsed Operation", "Millimeter and Submillimeter Wave
devices whose functions are combined quasi-optically. Com Quasi-Optical Oscillator with Gunn Diodes", and "Active
plementing these experimental accomplishments, a variety of Inverted Stripline Circular Patch Antenna for Spatial Power
theoretical approaches has been developed for the design Combining".
and characterization of quasi-optical components and systems. We hope that this issue will provide not only state-of-the-art
Much of this pioneering work has been presented at workshops information but a perspective on a fast developing and diverse
and professional society meetings. Because of the technical field.
0018-9480/93$03.00 © 1993 IEEE
1662 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 41, NO. 10, OCTOBER 1993
The editors want to thank the reviewers for their participa
tion in the paper evaluation process and for their constructive
comments, which many authors acknowledged as a valuable
help in the preparation of their final manuscripts. We sincerely
appreciate the reviewers' efforts.
JAMES W. MINK
DAVID B. RUTLEDGE
Guest Editors
Dr. James W. Mink (S'59-M'65-SM'81-F'91) joined the U.S. Army Research Office
in 1976, where he currently serves as Director of the Electronics Division. In addition, he
continues to direct an extramural research program in electromagnetic theory and millimeter
wave integrated circuits and devices and is the principle Army representative of the Joint
Services Electronics Program (JSEP). From 1984-1990, he served as the Associate Director
of the Electronics Division. During 1982, he served as an intern in the Office of the Army
Deputy Chief of Staff for Research, Development and Acquisition. Since 1979 he has been
an Adjunct Associate Professor at North Carolina State University, teaching electromagnetics
and microwave theory, and conducting research on millimeter wave devices and antennas.
He continues to serve as committee chairman, session chairman, and panelists for numerous
conferences and workshops and is a Fellow of the Institute of Electrical and Electronics
Engineers. He serves as an evaluator of University Departments of Electrical Engineering
for the Accreditation Board for Engineering and Technology. From 1964 through 1975, he was
engaged in research at the U.S. Army Electronics Command, Fort Monmouth, NJ. During this time, he performed basic research
on free space and guided propagation of electromagnetic waves on electrically small antennas. Dr. Mink received the B.S., M.S.,
and Ph.D. degrees in electrical engineering in 1961, 1962, and 1964, respectively, from the University of Wisconsin, Madison.
David B. Rutledge (M'75-SM'89-F'93) received the B.A. degree in mathematics from
Williams College, Williamstown, MA, in 1973, the M.A. degree in electrical sciences from
Cambridge University, Cambridge, England, in 1975, and the Ph.D. degree in electrical
engineering from the University of California at Berkeley in 1980. In 1980 he joined the faculty
at the California Institute of Technology, Pasadena, CA, where he is now Professor of Electrical
Engineering. Previously he designed microwave datalink systems as an Aerosystems Engineer
at General Dynamics Corporation, Fort Worth, Texas, from 1975 to 1976. He was a visiting
scientist at CSIRO, New South Wales, Australia, in the summer of 1985, and at the Research
Institute for Electrical Communication, Tohoku University, Sendai, Japan, in the spring and
summer of 1988. His research is in developing microwave and millimeter-wave integrated
circuits and applications, and in software for computer-aided design and measurement. He is
co-author with Scott Wedge and Richard Compton of the software CAD program, Puff, which
has over 10 000 users worldwide. He is a winner of the NSF Presidential Young Investigator Award and the Japan Society for
the Promotion of Science Fellowship. He has been a Distinguished Lecturer for the Antennas and Propagation Society, and
is a winner of the 1993 Microwave Prize.
REVIEWERS FOR THIS SPECIAL ISSUE
Constantine Balanis Derek Martin Lance Sjogren
Kai Chang William McGrath Karl Stephan
Richard Compton Koji Mizuno Michael Steer
Michael De Lisio Raj Mittra Philip Stimson
Neal Erickson Ellen Moore Kiyo Tomiyasu
Paul Goldsmith J. Anthony Murphy Robert Weikle
Jon Hacker Dean Neikirk Michael Wengler
Tatsuo Itoh Herbert Pickett James Wiltse
Christina Jou Zorana Popovic Stafford Withington
Linda Katehi David Pozar Richard Wylde
Anthony Kerr Gabriel Rebeiz Sigfrid Yngvesson
Wayne Lam Bernard Robert Robert York
IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 41, NO. 10, OCTOBER 1993 1663
James Lamb Felix Schwering Jonas Zmuidzinas
Shijie Li Arthur Sheiman
1664 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 41, NO. 10, OCTOBER 1993
Focal Plane Imaging Systems for
Millimeter Wavelengths
P. F. Goldsmith, Fellow, IEEE, C.-T. Hsieh, Member, IEEE, G. R. Huguenin,
Senior Member, IEEE, J. Kapitzky, and E. L. Moore, Senior Member, IEEE
Abstract-We discuss critical aspects of imaging system design The millimeter wave region has several important advan
and describe several different imaging systems employing focal tages in terms of system design, but these very much depend
plane array receivers operating in the 3mm-2mm wavelength
on the precise application. In a general sense, millimeter
range. Recent progress in millimeter-wavelength optics, antennas,
wavelength propagation is superior to that found in the infrared
receivers and other components permits greatly enhanced system
performance in a wide range of applications. We discuss a and visible spectral ranges in poor weather conditions, and
radiometric camera for all-weather autonomous aircraft landing is thus preferable for systems that must operate independent
capability and a high sensitivity cryogenically cooled array for of environmental conditions. Millimeter wavelengths clearly
use in radio astronomical spectroscopy. A near-focus system for
offer better angular resolution for a given antenna diameter
identification of plastic materials concealed underneath clothing
than do microwaves, and thus have a great potential advantage,
employs a two element lens, and has been demonstrated in active
(transmitting) and passive (radiometric) modes. A dual mode particularly for commercial systems which must fit into re
imaging system for plasma diagnostics utilizes both active and stricted envelopes. The particular properties of the interaction
passive modes at its '.::::'.140 GHz operating frequency to study of millimeter wavelength energy with materials can offer
small-scale structure. The radiometric imaging systems employ
advantages for certain applications, such as remote sensing
between 15 and 256 Schottky barrier diode mixers while the
of trace gases, and this has been used to very great effect
imaging receivers for the active systems include 64 element video
detector arrays. recently in studies of the Earth's atmosphere [1).
The millimeter wavelength region of the electromagnetic
spectrum is situated between microwave frequencies, in which
l. INTRODUCTION range coherent signal processing techniques are well devel
I MAGING can be considered to be the process of measuring oped, and the infrared region, in which incoherent technology
has predominated. It is indicative of the position of the
the radiation arriving from different directions. Our experi
technological frontier to compare the lack of imaging arrays at
ence with imaging derives most directly from our experience
millimeter wavelengths with infrared cameras which already
with the eye, an optical system employing an array of inde
employ focal plane arrays with thousands of elements [2), and
pendent detector elements in the focal surface of an imaging
highly sophisticated microwave phased array systems [3).
lens. A focal plane imaging array is only one architecture
Phased arrays have, however, been limited in their success
for imaging systems. Obtaining an image of a scene can be
at millimeter wavelengths due to difficulties designing efficient
carried out in many different ways, several of which are shown
schematically in Fig. 1. It is possible to build up an image of radiating elements as well as relatively large feed system loss.
a scene by scanning a single pixel receiver, either by moving Interferometers have been restricted to research applications
a detector in the focal plane of an antenna or mechanically as a consequence of their cost and complexity. Mechanically
or electrically scanning the beam direction of the system. This scanned systems have been employed for different imaging
technique, while slow and cumbersome, has been the only one applications in the millimeter range, but suffer the very ma
available at many wavelengths until very recently. Another jor handicaps of expensive, relatively unreliable mechanical
type of imaging system is the phased array, in which the systems together with low data rate from the single receivers
signals from a number of independent radiating elements can utilized. Focal plane arrays have to date been little developed
be processed in order to synthesize beams which effectively due to problems with developing an effective type of feed for
are sensitive to radiation arriving from different directions. A coupling between free space and single mode detectors, as
close relative is the interferometric array widely used in radio well as the significant issues of cost and complexity inherent
astronomy, in which the signals from different antennas are in systems employing large numbers of pixels.
correlated and an image obtained by carrying out an off-line Extensive work during the last decade or so on overcoming
Fourier transformation of the data. these problems has resulted in much improved feed elements
for focal plane imaging systems, and the potential for very
low cost individual radiometers to use in large scale focal
Manuscript received October 1, 1992; revised April 15, 1992.
P. F. Goldsmith is with Millitech Corporation and the National Astronomy plane arrays. As a result of these developments, we have
and Ionosphere Center, Cornell University, Ithaca, NY 14853. adopted focal plane array technology as our design approach
C.-T. Hsieh, G. R. Huguenin, J. Kapitzky and E. L. Moore are with Millitech
for imaging systems in the 2 mm to 3 mm wavelength range.
Corporation, P.O. Box 109, South Deerfield, MA 01373.
IEEE Log Number 9211925. In this paper we indicate some of the design considerations
0018-9480/93$03.00 © 1993 IEEE
GOLDSMITH et al.: FOCAL PLANE IMAGING SYSTEMS 1665
while its importance for active systems is very much dependent
MECHANICALLY on the configuration. The basic radiometric equation for the
SCANNED
SINGLE-ELEMENT rms uncertainty of the scene temperature that can be measured
SINGLE with a coherent detector system (heterodyne or amplifier) is
~~ BEAM
\'----_
MOTIONS (1)
(a)
where /3 is a factor, generally between 1 and 2, which
depends on the type of radiometer and how the measurement is
performed; E is the coupling efficiency between the radiometer
and the scene being measured, which includes both antenna
source coupling effects and losses in the input optical system;
Ts is the system noise temperature; 8 f is the predetection
1 BEAM PER bandwidth; and T is the integration time. All types of systems
ELEMENT
000 benefit from efficient coupling and low noise temperature.
000 Systems which are measuring thermal sources can use a
000 relatively large bandwidth to minimize !J.T. Active systems for
which the signal is dominated by the return of a transmitted
(b)
signal are better characterized by their minimum detectable
power, given by
ANTENNAS RF SYSTEMS CORRELA TORS
11 (2)
1•2
Bl
l
! -N 2(N-1-) where k is Boltzmann's constant. We see that the minimum
2
INTERFEROMETER CORRELA TOR detectable power increases as the bandwidth is increased, so
BASELINES 2,3 OUTPUTS
B2 EACH SENSITIVE that in this case, we want the bandwidth to be large enough
TO SPECIFIC
13 SPATIAL FREQUENCY only to allow the return signal to be processed. If the signal
is sufficiently strong, it can be directly detected without any
1•3
RF processing. In such video detector systems the sensitivity
MUL TIELEMENT
INTERFEROMETER is set by noise fluctuations in the detector element and the
predetection bandwidth is generally not a critical parameter.
(c)
The sensitivity limits set by the radiometer equation given
above, or by video detector sensitivity, are lower limits. A
ANTENNA BEAM FORMING
ELEMENTS NETWORK variety of effects can degrade performance, including both
instabilities in the radiometer and characteristics of the scene
OUTPUTS FOR being imaged. Gain instabilities are well known to radiometer
INDIVIDUAL
BEAMS designers and can be countered by employing a rapid calibra
EACH CORRESPONDING
TO DIFFERENT tion cycle or switching between the scene and a comparison
DIRECTION
load (Dicke switching). Single-pixel imaging systems have
I I relied on mechanical scanning of the beam across the scene
PHASED ARRA y
[4], with calibration performed by looking at thermal loads
(d) [5] or by weak coupling of a noise source [6]. For an
Fig. 1. Schematic of different types of imaging systems (a) mechan imaging system with many pixels, a quasioptical input switch
ically-scanned single element; (b) focal plane array; (c) multielement is essential for load switching as well as for calibration.
interferometer; (d) phased array.
The Autonomous Aircraft Landing System discussed below
employed a mechanical quasioptical switch, but development
which are of critical importance in the development of systems of an electronic replacement using PIN diodes has already
for commercial and research applications. We describe four yielded extremely promising results [7].
different systems and present some images that have been In some situations, small-scale variations of the scene tem
obtained to indicate the degree to which this technology has perature which are not of interest set an effective lower limit
matured. to the useful sensitivity for studying variations in the scene.
As an example, the scenes studied in [6 ] were characterized
by a "clutter" of 2-3 K. With RF/IF bandwidths of several
II. GENERAL CONSIDERATIONS FOR IMAGING GHz which are now feasible, the noise temperature required
SYSYEM SENSITIVITY to have a radiometric !J.T comparable to or less than this
The relatively low intensity of thermal emission makes sen value is quite reasonable, being 104 K for 8t = 3 GHz and
sitivity an important concern for passive radiometric systems, T = 1/30 second.
1666 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 41, NO. 10, OCTOBER 1993
III. EFFICIENCY, PACKING, AND COST illumination pattern appropriate to the antenna focal ratio. A
OF ARRAY FEED ELEMENTS close-packed array of such horns cannot fully sample the focal
plane because the electric field distribution within the horns is
These considerations are central to the successful imple
generally quite tapered in order to achieve a radiation pattern
mentation of any focal plane imaging system and have been
with low sidelobes. For scalar feedhorns, for example, close
discussed in some generality (8], (9], (10]. The feed elements
packing of the horns produces only approximately every-other
have been the subject of very intense development during
beamwidth sampling of the focal plane (10]. Reducing the size
recent years. Traditional rectangular horn elements have been
of the feed horns results in a broader beam and much reduced
compared (11] and developed in monolithic form (12]. A
antenna efficiency. For travelling wave antennas which derive
variety of different types of travelling wave slot antennas
their gain in part from their length, the situation is more
have been proposed and studied including constant width
complex, but placing the elements too close together results
slots [13] and linearly tapered slots (14]. Other types of feed
in interaction via fields and/or currents extending away from
elements include double dipole (15], Yagi [16] and dielectric
the feed elements themselves. In an experimental study, it was
rod antennas [17], among others. The general criteria that these found that a L\ spacing of feeds appropriate for a f / D = 1
designs seek to satisfy include:
system was about the minimum that gave acceptable results
1) efficient illumination of antenna
(17]. Since the distance in the focal plane corresponding to
2) low loss one beamwidth is ~x = f >./ D, this corresponds to about 1
3) effective coupling to active devices
beamwidth spacing.
4) good packing efficiency.
From a more general viewpoint, full sampling of the focal
In addition, qualities such as high polarization purity and
plane for incoherent illumination of the scene and measure
reasonable angular divergence end up being of importance in
ment of the intensity in the focal plane requires an element
many applications, while low cost is a factor that must be spacing ~x = 1/2 · (f >./ D) (19]; in f / D = 1 system we
borne in mind particularly for commercial applications.
thus require an element spacing not much greater than >./2.
Achieving good illumination efficiency demands low side
This is far closer to the limiting spacing found theoretically
lobes and backlobes, which have been persistent problems for
for a variety of feed elements (20]. These authors found that
antennas on substrates. This is one of the motivations behind
the aperture efficiency for an antenna with slot element feeds,
the use of lens-coupled printed circuit antennas, which exploit for example, is ~0.5 for ~x ~ 1.15>. in a f / D = 1 system.
the tendency to radiate into the dielectric and combine this with
For a corrugated feedhorn, by contrast, the aperture efficiency
collimation of the beam by the dielectric lens (18]. Low loss
has dropped to 0.17 for a spacing just slightly less than this. In
is an issue with feed elements employing dielectric materials. the focal plane arrays described below, the constant width slot
For commercial systems operating at ambient temperature, the antennas used with f / D = 1.1 optics are generally spaced by
effect is also more severe than for radio astronomical systems 1.35>., which is about as close as possible without excessive
in which the feed elements are cooled to low temperature, interaction between elements. These systems, while achieving
which greatly reduces their added noise as well as reducing good illumination efficiencies, are clearly far from the desired
their dielectric loss. Most types of planar designs, including spacing for full sampling, and from the theoretical limit for
travelling wave slot and dipole antennas, offer good coupling reasonable overall system efficiency (21].
to active devices, either mixers or amplifiers; coupling is,
however, a more severe problem for dielectric rod and some
other types of antennas.
The question of packing efficiency is an interesting one, as
well as being very important for development of effective focal IV. IMAGING ABILITY OF FOCUSING SYSTEMS
plane arrays. Radio astronomical systems observe essentially Analysis of the variation of performance of lens and reflec
non-changing scenes and mosaicking an image by repointing tor antennas is a major branch of electromagnetics and has
the antenna to fill in points in the focal plane not sampled interesting parallels and differences from imaging in optical
by the array is, in principle allowed. However, it is generally systems at visible wavelengths. Although work on scanning
inefficient to have the array elements too widely separated, systems started even earlier, the now classic papers on imaging
since some radio sources will not be very extended compared properties of reflectors (22] and lenses (23] have burgeoned
to the footprint of the array on the sky, and array elements are into a literature so vast that is not possible even to summarize
thus not being used effectively. For commercial applications, the results, but rather only to give a sampling of refer
where near real-time updating of the radiometric image is ences (24].
required, minimum element spacing is of vital importance to The issue has, in fact, become rather more acute with recent
achieve a reasonably well-sampled image without any motion progress in focal plane array technology-the problem has
of the antenna. shifted from one only of academic interest, to a critical aspect
The ability to sample the image in the focal plane of an of developing practical array systems. The critical parameters
antenna is intimately related to proper illumination of the include maintaining antenna gain, beam size, and beam quality
antenna itself by the feed element. We assume that an indi over the range of angles scanned; achieving these goals is
vidual feed is used for each pixel. If it is an aperture-type feed made difficult by a combination of factors including the
(e.g. horn) then it has a characteristic lateral size to prqduce the following:
GOLDSMITH et al.: FOCAL PLANE IMAGING SYSTEMS 1667
1) Focal plane arrays with ::::::104 pixels are being seriously (as discussed below). The full three-dimensional geometry is,
considered, with the result that scan angles ::::::100 beam widths in fact, extremely important in making a system which not
off boresight must be considered. only can contain the required components, but which can be
2) System aperture diameters are relatively small, typically assembled and serviced effectively.
only a few hundred wavelengths for commercial systems, and Power dissipation is already a non-negligible issue for exist
often less. Blockage loss for symmetric reflector systems of ing arrays. The ambient temperature heterodyne systems dis
such limited size is generally excessive. cussed below for radiometric imaging consume approximately
3) The limited volume available restricts f / D for lens 200 mW per pixel for the IF amplifiers, or approximately
systems to values :::; 1.25, which seriously restricts imaging 51 W of power for the 256 element array. Even with the
capability. For this same reason, off-axis optical systems are relatively large volume available, dealing with this thermal
generally not acceptable. dissipation is a serious issue. Millimeter wavelength transistor
Meeting the requirements in the face of these restrictions amplifiers are currently undergoing rapid development and
is often a very difficult challenge. While it is true that there are receiving consideration for imaging arrays [27], [28].
have been no microwave or millimeter focal plane array These components, while very attractive in terms of low noise
systems with the number of pixels approaching 104 actually performance, are presently problematic in regard to power
implemented, several of the systems discussed below have consumption. A 91-95 GHz amplifier employing 8 stages to
been designed to be extended to this level, and the imaging achieve the 50 dB gain necessary for direct detection dissipates
properties of the optical systems analyzed accordingly. Imag 560 mW, or almost three times more than the heterodyne
ing lens design studies have primarily been based on analytical system [29]. Clearly, further progress will be required to make
treatments, which try to eliminate certain classes of phase error RF amplifier systems usable in large-scale integrated arrays
in the aperture plane which are called aberrations. However, and further reduction of power consumption of mixer-based
the resulting systems are not fully optimized for practical use, systems is definitely desirable.
given that arrays are modular and the locations of individual Cost is always an issue in systems for commercial appli
feed elements cannot exactly follow a specified focal curve. cations, but in particular for focal plane arrays with large
Thus it is often necessary to resort to simulations rather quantities of millimeter-wavelength circuitry, a small change
than analysis. Full-scale diffraction calculations would be in the cost per pixel can translate into an impressively large
excessively time consuming, so that a hybrid approach that variation in the total cost of a system. Again, technology is
uses Gaussian beam propagation between feeds and optical changing rapidly, but we have found the use of circuit cards
elements, ray tracing through the lens systems, and finally fabricated on "soft" dielectric materials with bonding of dis
a single diffraction calculation of propagation to the far crete and hybrid components to be reliable and cost effective.
field or to a specified plane where the properties of the Radio astronomical and other research-oriented systems with
beam can be examined is very attractive [25], [26]. With the relatively fewer high performing pixels will probably continue
ever-increasing capability of affordable computers, it should to employ waveguide and rigid substrate technology to achieve
soon be possible to connect this type of imaging analysis lowest loss and system noise.
program with an optimizing function to obtain best imaging
performance over some fixed scan range with appropriate
VI. DESCRIPTION OF SOME SPECIFIC SYSTEMS AND RESULTS
constraints on feed locations. This capability, already available
In the following we describe a number of millimeter
for geometrical optics systems with analysis based purely on
wavelength imaging systems developed for a variety of ap
ray tracing, should significantly enhance the design capability
plications. Along with basic parameters of the systems, we
for diffraction-limited imaging systems.
present some imaging results that have been obtained.
V. OTHER CONSIDERATIONS A. 140 GHz Plasma Imaging Camera
Other considerations in the design of imaging focal plane Diagnostics of conditions in the ionized gas component of a
arrays, which are less fundamental than those discussed above Tokamak fusion reactor is vital for improving the containment
but are nevertheless important, are power dissipation and cost. of the plasma and understanding energy loss processes. The
Even with the current early state of the development of focal camera described in this section was designed to determine
plane arrays, it is not too early for these issues to be addressed, from a single "event" or ignition whether small scale fluctuat
especially when extrapolating to larger systems. The basic ar ing plasma modes in the interior of the plasma are responsible
chitecture and geometry of the focal plane array are intimately for plasma transport. Test bed plasma fusion reactors to
tied to these questions. For example, while purely planar (e.g. date have been pulsed systems, with measurements being
broadside radiating) arrays may be desirable from the point carried out at a few discrete locations. Although this approach
of view of entirely monolithic fabrication, they suffer from has adequate time resolution, it lacks the spatial resolution
problems of inadequate space for complete IF circuits, and necessary to identify the modes which fluctuate over periods
potentially from difficulties in thermal dissipation. We have of tens to hundreds of microseconds and have scale lengths
found that the endfire slot antenna geometry is effective in of centimeters. The configuration and properties of the plasma
this regard, even though obtaining a two dimensional array can vary from one run to the next. Thus, "single shot" imaging
requires combining "cards" each with a one dimensional array capability is essential to understand what is happening.
1668 IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 41, NO. 10, OCTOBER 1993
Measurements of a number of materials at 95 GHz (see
Table I) have shown that substances such as polycarbonate,
polypropylene, rexolite and black delrin have reflectivities in
the range - 8 to -12 db. The reflectivity of various portions of
the human body range from -11 db (calf) to -15 db Qowl).
Furthermore, a wide variety of tested clothing items show
reflectivities in the range of -15 to - 25 db with transmission
losses of less than 0.5 db. It appears, then, that dielectric
materials of interest have reflectivities 2 to 4 times that of
human skin, while clothing provides little in the way of
absorption or extended reflectivity.
With imaging spot size being proportional to the product of
the imaging lens f / D ratio and the wavelength of operation, it
is clear that improved resolution can be obtained by operating
at higher frequencies. However, at submillimeter and shorter
wavelengths, materials are too lossy and the technology for
Fig. 2. Photograph of 140 GHz dual mode plasma imaging camera. The
cost-effective imaging systems is not available. At longer
transmitter feedhorn is at the lower left, and above it is the 90° reflector to
couple its beam to the scene. The focal plane array, local oscillator injection (microwave) wavelengths, the technology is well developed,
feed, and diplexer, are on the right. but the spatial resolution is inadequate. If the system optics are
designed such that f / D is approximately 1, then a wavelength
of 3 mm gives a spot size adequate for the recognition of
Millimeter wavelength radiation can play several valuable
potentially harmful hidden plastic objects. To achieve high
roles in plasma diagnostics. The intensity of emitted radia
throughput of subjects, an array of detectors sampled at video
tion allows determination of the plasma temperature, while
rates (30 Hz) is an attractive approach. The necessary optics
the reflected signal can yield information about the density
+/ -
must be capable of a large field of view (e.g. 30 degrees)
structure in the ionized gas [30]. A dual mode plasma imaging
with diffraction-limited performance.
camera has been developed for the Princeton Plasma Physics
Reflective imaging makes use of a relatively intense mil
Laboratory to take advantage of both of these capabilities,
limeter wave source to illuminate the subject. Higher reflec
with an emphasis on studying small-scale fluctuating plasma
tivity portions of the subject result in a higher output from the
modes. The imaging system, shown in Fig. 2, includes a 4
system. An extended array of slot antennas can then produce
by 16 element focal plane array operating at 140 GHz. This
a video image or a smaller array can be mechanically scanned
frequency is dictated by the properties of the plasma in the
about the subject (albeit at the cost of reduced throughput).
Tokamak Fusion Test Reactor.
The major problem with this approach is the presence of
The feed elements are constant-width slot antennas, and
"glints" in the millimeter wave image due to reflections of the
the down converters are second harmonic mixers pumped by
illumination sources from boundaries in the subject's clothing.
a phase locked 70 GHz Gunn oscillator. In its radiometric
This effect is well known from active millimeter-wavelength
mode, a 2 GHz bandwidth IF is amplified and detected. For
imaging of tactical targets [31].
reflectometry, a 30 mW transmitter at approximately 140 GHz
Investigation is proceeding on two approaches to solving
illuminates an area of the plasma about 30 cm in diameter.
this problem: 1) using a number of lower level sources to
Each beam from the focal plane array is focused by the 1 m
illuminate the subject from a wide range of angles, and 2)
diameter lens to an area about 1.5 cm diameter at the plasma,
investigating the behavior of a subject rotating in the field of
which is approximately 2 m away from the camera. The return
view of the imaging device. It appears that the "glints" are
signal is processed to obtain in-phase and quadrature signals
highly dependent on the source-subject-detector geometry and
from which the phase of the reflected signal can be obtained.
change rapidly as this geometry varies, whereas reflections
from real objects have longer duration.
B. Detection of Concealed Weapons and Explosives
To study the performance of concealed object detection
In recent years the ability to detect plastic weapons and systems, a mannequin was covered with a millimeter wave
explosives concealed beneath the clothing of airline passen absorptive material that approximates the absorption charac
gers has received increasing interest by law enforcement and teristics of human skin. The mannequin was then clothed and
security agencies. To be effective, such detection devices must objects concealed on the torso. A 64 element focal plane array
1) have a low false alarm rate, 2) be non-invasive, and 3) imaging array was used, employing endfire slot antennas each
have high throughput. Both active (reflecting) and passive of which was coupled to a detector diode. The illuminating
(radiometric) millimeter wavelength imaging systems have source produces levels of radiation well below current safety
been investigated for these purposes. Frequencies between 30 thresholds. It consists of a number of Gunn oscillators which
and 300 GHz provide a number of advantages including 1) are frequency-modulated over a few hundred MHz range,
low absorption by most dry dielectric materials, 2) moderately and also amplitude-modulated. An appropriate synchronous
high spatial resolution, and 3) good imaging capability via detector circuit at the output of the slot antennas converts
compact optics. the reflected return into a de voltage. The imaging optics
