Journal of Superconductivity, Vol. 5, No. 4, 1992
Advanced High-Temperature Superconducting Components for Microwave Systems E. Belohoubek, 1'2 E. Denlinger, 1 D. Kalokitis, 1 A. Fathy, 1 R. Paglione, t V. Pendrik, ~ J. Brown, ~ A. Piqu6, 2 X. D. Wu, 2~ S. M. Green, 2 S. Mathews, 2 R. Edwards, 2 M. Mathur, 2 and T. Venkatesan 2
Received 3 April 1992
The application of high-temperature superconducting thin films to microwave systems is expected to lead to major volume and weight savings as well as improved performance. To take full advantage of the properties that the new materials have to offer and justify the additional cooling equipment that accompanies the introduction of superconductivity, many HTS components will have to be integrated. In this paper some of the key microwave circuits that show great promise in this respect, such as multiplexers, circulators, and very narrowband filters, will be discussed and experimental results presented. KEY WORDS: High-temperature superconductivity; multiplexer; narrow-band filter; circulator; high-Q spiral resonator.
1. INTRODUCTION
delay lines) with HTS counterparts, but rather from a comprehensive systems approach which combines a number of HTS components with other devices, such as circulators, switches, and possibly active devices that also may benefit from lower operating temperatures, and integrates them into a cooled subsystem. The earliest applications of HTS technology in practical microwave systems most likely will be found in satellites where partial passive cooling may be feasible or where the weight and volume advantages of the superconducting components make up for the additionally necessary cooling equipment and power consumption. With the rapid progress currently being made in the efficiency and size of cryogenic coolers, HTS applications should also become feasible in a variety of airborne and even land-mobile communications systems. The desire to reduce the overall volume and weight of advanced microwave systems has dominated technology developments for decades. The historic trend of trying to reduce the size of microwave resonators led, over the years, from bulky waveguide filters to coaxial filters, followed by printed-circuit
Today's high-temperature superconducting (HTS) thin-film technology has reached the stage where a variety of passive microwave components can be realized that have significant size and weight advantages as well as better performance characteristics than comparable state-of-the-art conventional components. The Navy's HTSSE Phase I program [1] led to the demonstration of a large number o f breadboard components that are expected to be flown on an experimental satellite in 1993. Phase II of the HTSSE program, now getting underway, should lead to more advanced and sophisticated components that could eventually see actual use in various defense systems. It is quite apparent by now that truly enabling benefits will not likely be derived from simple replacements of existing components (for example, filters and
~David Sarnoff Research Center, Princeton, New Jersey 08543. 2Neocera, Inc. College Park, Maryland 20742. 3Visiting Scientist from ERDC, Los Alamos National Laboratory, Los Alamos, New Mexico 87545.
423 0896-1107/92/0800-0423506.50/09 1992PlenumPublishingCorporation
424
Belohoubek et al.
components, first in distributed and later in lumpedelement form. These miniaturization efforts, however, resulted in progressively lower-Q elements since the unloaded Q is roughly proportional to the cube root of the volume. Thus, lumped-element components, which are predominantly used in monolithic circuits, occupy the smallest space but have also the lowest Q. It is in the areas of miniaturization and improved performance where superconductors can make major contributions to the state-of-the-art microwave circuits. In the following, some of the more advanced superconducting microwave components that are expected to become important building blocks in the realization of HTS subsystems will be discussed.
2. MULTIPLEXERS AND CHANNELIZERS ,The major contribution superconductivity can make to conventional microwave components is a drastic reduction in size and the achievement of much higher Q values than available with conventional technology. Narrow-band channelization based on stateof-the-art technology is inherently limited by the Q of the resonant elements and leads to rather large, bulky structures. This is one reason such devices have not found widespread application in present systems, especially not in air- or space-borne applications where size and weight is of primary concern. Multiplexers can take a variety of forms. The two configurations shown in Figs. la and lb permit the well-matched cascading of many channels without interactions between individual filters. The disadvantage of this approach in conventional technology is a rather high and varying insertion loss between channels, especially if a large number of channels is being f]' f2""fn
'L fl
Filters fn
f. f2 Fig. la. Hybrid coupled multiplexer.
B
Filter I fl f2
tn
Fig. lb. Circulator coupled multiplexer. cascaded. The hybrid-coupled multiplexer in addition suffers from a very high parts count, since it uses two identical filters for each channel. The configuration shown in Fig. 2 uses a central manifold with individual T-junctions for each channel. Interaction between channels is difficult to control and even more so in contiguous multiplexers. Rather extensive computer optimization programs have to be used to adjust filters for the reflections from adjacent filters. Channelizers of this type have been successfully realized with moderate numbers of channels, but the complexity and difficulty of the fine-tuning process make this a less preferred approach. The directional filter configuration shown in Fig. 3 is also not widely used, mainly because it produces a high insertion loss that cannot be eliminated even in superconducting form. In comparing the above multiplexer configurations it appears that the circulator-coupled version would be clearly superior, provided the size and the losses of the circulator could be substantially reduced and the filters directly integrated with the circulators on the same substrate. Superconductivity provides the ideal solution to this challenge. In a circulator-coupled multiplexer, the individual filters can be designed, aligned, and tested essentially free of interference from adjacent filters. A typical HTS filter configuration for a narrowband, X-band filter and its performance is shown in Fig. 4. This HTS quasi-interdigitated coupled-line filter was fabricated in stripline form to facilitate the design with a high dielectric constant material (LaAIO3), and to keep losses from the copper ground planes at a minimum [2]. Figure 4 also shows for comparison the performance of a copper version of this three-pole filter. The comparison filter is of identical design but uses copper conductors throughout and had an insertion loss of 18 dB at 77 K, which clearly illustrates the advantages of the superconductive implementation. In the future, a garnet substrate instead of LaAIO3 may be used to permit integration
HTS Components for Microwave Systems
425
fl, f2-'fn INPUT
f2
Fig. 2. T-junction manifold multiplexer.
Fig. 3. Directional coupler multiplexer.
Ll''"b" ~0' 4 " - L2- ' - ~
INPUT
Fig. 4. Photograph and performance of 3-pole band filter in copper and HTS implementation.
I" fo
426
Belohoubek et al. Thick Ferrite Discs ~ , ~ ~ _ _ _ ~ M a g n e t HTS Line on Thin
\~Fe [~Fe
of a nonreciprocal element. In addition, multiport circulators are valuable components in transmit/receive modules, modulators, comparators, and other critical-performance circuits. Whenever complex microwave systems are not operating properly, the cure very often is a liberal sprinkling of circulators between components. Recent experiments conducted under an SDIO SBIR program [3] demonstrated the feasibility of nonreciprocal interaction in an HTS quasi-stripline circulator. The device consists of a superconducting YBCO film deposited on a 150-#m-thick sapphire substrate, 12mm in diameter, that is sandwiched between two ferrite disks as shown in Fig. 5. A very thin sapphire substrate was chosen as the carrier for the HTS film to keep the perturbation of the ferrite cavity, formed by the two ferrite disks, at a minimum. The YBCO film was laser-deposited over a thin CeO2 buffer layer. The surface resistance for such films is typically an order of magnitude lower than that of copper. A cross section of the experimental X-band circulator is shown in Fig. 6. The ferrite disks cover the circulator as well as the three Z/4-matching transformers. The top ferrite disk radius is cut back by 250 #m in the areas where the three connectors are attached to the transformer sections via gold ribbons.
/
~,\\\\\'~'~-- Magnet
]Fig.5. Quasi striplinecirculatorusing HTS filmon thin sapphire substrate. directly with the circulators. Such a substrate has a lower dielectric constant ( s t " 14) that permits microstrip construction and thus a further reduction in the parts count, and is also more amenable to state-ofthe-art CAD design. 3. SUPERCONDUCTING STRIPLINE CIRCULATOR The need for a superconducting circulator that can be integrated with other HTS components is clearly apparent from the above section. Circulators or isolators are widely used in most systems where the emphasis is on tight amplitude, phase, and groupdelay specifications. Direct integration in close proximity with the sensitive components is important to derive the maximum benefit from the introduction
27.6 mm.dia.
ASSEHBLY
-
TOP VIEW
K-CONNECTOR ASSEHBLYB P ~
~CIRCULATOR
,AG,ET
. ~ 2 1 5,2.0 ~ 1 7 60,o.,
", . . ~ .X..~:~",~!~~ / / ~
~-)~'i;';';'~';'~-~-6.1"///i
SAPPHIRE OlSK {0.152 mm.X 12.0 mm. alia.}
"A
10.635mm.X12.0 ram.dia.)
/////I
.I Fig. 6. Assemblyviewof HTS circulator. I_
LID
26,0
HTS Components for Microwave Systems
427
A photograph of the assembled circulator with its top removed is shown in Fig. 7. This design proved to be mechanically stable, provided reproducible microwave measurement results, and could be repeatedly cycled to low temperatures without breakage. The breadboard circulator exhibited an insertion loss that dropped from 30 dB at room temperature to 0.25 dB at 77 K. The isolation was better than 20 dB over a 10% bandwidth with a peak value of 34 dB. The performance of the circulator as a function of temperature is shown in Fig. 8. The high transition temperature and sharpness of the transition is a clear indication that excellent superconducting performance can be obtained in the presence of high magnetic transverse fields (~1700 G). The insertion loss in the present experimental circulator at 77 K and below is dominated by the magnetic loss of the ferrite substrate material and the fact that the ground planes are made
of copper. The particular ferrite chosen for these early feasibility experiments, Trans-Tech TT2-125, was one readily available from previous circulator development efforts. This material is well suited for highpower applications but has fairly high magnetic losses which, however, in room temperature circulators are not dominant. For superconducting circulators where extremely low insertion loss is important, other materials such as YIG or calcium vanadium garnet may be a better choice. With such materials and superconducting ground planes the insertion loss can be expected to go well below 0.1 dB. The present configuration was chosen mainly to provide an initial feasibility demonstration. For practical systems applications, the preferred approach would be to use a microstrip version in which the HTS film is applied directly to a garnet material. The garnet, because of its low-loss characteristics, could
Fig. 7. Assembledcirculator.
428
Belohoubek et al. HTS Circulator 2 9.35 GHz lOO
. . . .
,
. . . .
=
. . . .
,
. . . .
,
.....
6 mm'---" ,
. . . .
,
. . . .
113 "o
25 TURNS 50 rtm LINE WIDTH 50 Ilm SPACING
10 m m 10
= o .1 ~r
x
-
10 m m
Fig. 9. Self-resonant spiral resonator (f0 = 140 MHz). . . . .
i 50
. . . .
i 100
. . . .
i 150
. . . .
,
. . . .
200
Temperature
i 250
;
,
i
, i . . . . 300
350
(K)
Fig. 8. Insertion loss vs. temperature of ,an HTS circulator (including 0.25 dB of connector losses).
also act as the common substrate for other passive microwave components. Thus, very low-loss circulators could be directly integrated with filters to form high-performance multiplexers in miniature form, or with delay lines, couplers, switches, and even active devices which also could benefit from cryogenic operating temperatures.
4. SUPERCONDUCTING HF AND UHF FILTERS One of the most promising applications for HTS thin films may develop in the form of miniature lumped resonators at relatively low frequencies, i.e., in the range from tens to hundreds of MHz. Small, self-resonant HTS spirals have the potential for extremely high-Q resonators in device volumes not achievable with any current state-of-the-art technology. To explore this very important and promising new area, we started to investigate the low-frequency capabilities of YBCO films. One of the earlier HTS films, deposited on LaAIO3 and belonging to a group that had good surface resistance values at X-band, was patterned into a spiral having a diameter of only 6 mm. The relevant dimensions of the spiral are shown in Fig. 9. A spiral of this size with a copper layer thickness of 25 pm would have an unloaded Q of less than 50 at the fundamental frequency of 140 MHz. The measured Q's for the HTS spiral are shown in Fig. 10 for three configurations: microstrip with Au ground plane,
microstrip with superconducting ground plane, and suspended substrate. Since the LaAIO3 substrate is 0.5 mm thick and has a high dielectric constant, one expects the electromagnetic fields to be closely bound to the dielectric at microwave frequencies. This is borne out by the fact that the Q's above 1 GHz for all three configurations are very similar. At low frequencies close proximity of a Au ground plane leads to a strong deterioration of the Q0 as indicated by the curve marked by open circles. A superconducting ground plane brings the Q0 of the fundamental resonance (,-~140 MHz) up close to that for the suspended substrate. The somewhat lower Q at the fundamental compared to the 2nd harmonic could probably be increased by using a second superconducting ground plane above the spiral. The major significance of these early measurements lies in the demonstration of very high Q values, in the range from 104 to 105, within a volume of less than 0.1 cm 3. Although the present results are preliminary, they point out the possibility of achieving very high-Q resonators in miniature form at U H F and HF
100000
o
n
10000
9
O
O
~
n 9
[]
D
9
[]
a
.
9
r
O O [ i
Microstrip,
HTS Ground
Plane
SuspendedS u b s t r a t e 1000
Microstrip,
Au Ground
Plane
O
Temp 100
. 0.0
i 0.2
.
, 0.4
.
, 0.6
.
, 0.8
,
, 1.0
.
Frequency
, 1.2
9
, 1.4
-
J 1.6
= 77K .
, 1.8
2.0
IGHz]
Fig. 1O. Unloaded Q of spiral resonator as function of frequency.
HTS Components for Microwave Systems
429
frequencies that could lead to the eventual development of highly compact, high-quality, low-loss filter structures that could not be realized with any other technology known today. Figure 11 shows a conceptual diagram of a compact, highly selective filter at 100 MHz using self-resonant spirals as resonator elements. The technology required for the fabrication of such a filter is well within the present state-of-the-art; a planar film deposition and patterning process can be used. Since the tap point for the input and output coupling is at the 50 f~ level, regular gold cross-overs should be sufficient for low loss performance.
5. CONCLUSION In the general trend toward the miniaturization of microwave components and subsystems, high-temperature superconducting implementations promise to play an important role. Recent results obtained on some of the HTS components described here, narrowband filters, multiplexers, and circulators show promise for the development of integrated HTS subsystems in the near future. By combining a number of components that all benefit from cryogenic operation and integrating them into a small volume, the additional cooling requirements may indeed become justifiable. ACKNOWLEDGMENT
fffjjJ/ii/f/ffjjffjfJJfJiffi'/
OooO 2trOt ((~
/ /
/
I/t-
.~
f e-/// 3.2 era
/ / / / / /
/
Portions of the work reported here were supported by the Strategic Defense Initiative under an SBIR program managed by Rome Laboratory, Contract F19628-91-C-0126, by a DARPA program under ONR sponsorship, Contract N00014-91-C0216, and by the NRL HTSSE program under Contract N00014-91-C-2017.
/////,'~
o scm
I..-
REFERENCES
o
Fig. 11. Superconducting spiral resonator filter (f= 100 MHz; Af= 50 KHz).
1. M. Nisenoff, D. U. Gubser, S. A. Wolf, J. C. Ritter, and G. Price, Supercond. Sci. Technol. 4, 449 (1991). 2. D. Kalokitis, A. Fathy, V. Pendrik, E. Belohoubek, A. Findikoglu, A. Inam, X. X. Xi, T. Venkatesan, and J. B. Barner, Appl. Phys. Lett. 58, 537-539 (1991). 3. E. Belohoubek, E. Denlinger, and A. Piqu6, Final Report SDIO SBIR Contract F19628-91-C-0126, February 1992.