Fabrication and evaluation of superconducting devices.
A recent article  provided an overview of the impact superconducting materials have on microwave electronics and offered some examples of both low [T.sub.c], low superconducting critical temperature, and high [T.sub.c] devices. This article provides information on the design, fabrication and testing of high [T.sub.c] microwave devices. It also gives an overview of recent superconducting microwave device research.
Two areas of interest for using high temperature superconducting (HTS) microwave devices are space applications and systems with infrared capabilities, which have cryogenic cooling.
The US Army Electronics Technology and Devices Laboratory (ETDL) is in the process of developing and evaluating HTS films for microwave applications. Efforts include the deposition of HTS films, fabrication of microwave superconducting microwave devices and insertion into subsystems. HTS microwave devices are tested in a cryogenic chamber cooled by liquid nitrogen, thus these devices are only evaluated at 77[degrees]K. However, film evaluations are done over a broader temperature range. Currently, any further improvement of device performance that may occur at lower temperatures cannot be determined. Also, film size of 0.5 [inch.sup.2] limits the device designs that can be fabricated.
Deposition and Characterization
of Superconducting Thin Films
Several factors, including deposition technique, substrate material and criteria for material evaluation, must be considered when developing high [T.sub.c] superconductors for microwave applications. Various deposition techniques, such as off-axis sputtering, E-beam evaporation and laser ablation, are capable of producing high quality films of the new copper-oxide superconductors. Laser ablation has been the most successful thin-film deposition method for high [T.sub.c] superconductors and for other materials. In the typical laser ablation system, a pulsed excimer laser operating at 248 nm is focused onto a target of the superconductor so that the single pulse energy density is 2 to 5 Joules/[cm.sup.2]. The power of the short pulse, typically 20 ns, is sufficient to form a plasma of the target material, which then is deposited onto a substrate. One of the advantages of laser ablation is that the thin-film stoichiometry is essentially that of the target material. The most commonly deposited material at ETDL is high [T.sub.c] [YBa.sub.2.Cu.sub.3.O.sub.7]. In order to form the correct phase of the superconductor, the deposition chamber is filled with 200 mTorr of oxygen and the substrate is heated to 700 to 750[degrees]C.
Substrate material is a critical parameter in the successful development of high [T.sub.c] films for microwave devices. Acceptable substrate materials must fulfill two major requirements. They must promote crystalline growth of the high [T.sub.c] films and they must have low loss tangents. The most common single crystal substrates currently in use are magnesium oxide (MgO) and lanthanum aluminate ([LaAIO.sub.3]). At 10 GHz and 77[degrees]K, the dielectric constant for MgO is 9.6 and the loss tangent is less than 6 X [10.sup.-5]. The dielectric constant and loss tangent for [LaAIO.sub.3] at 6.2 GHz and 77[degrees]K are 23.8 and 6 X [10.sup.-5], respectively.  However, there are some drawbacks to these materials because MgO is hydrophyllic and will become more lossy when exposed to moisture; and [LaAlO.sub.3] exhibits an inhomogeneity in dielectric constant by several percent over a typical 1 cm X 1 cm substrate. Other substrate materials being studied include [NdGaO.sub.3] and [LaSrAlO.sub.4].  Sapphire is also of interest because of its extremely low loss, but must be used with a buffer layer to prevent film degradation.
The electrical transport properties of superconductors typically are determined by measuring the DC resistivity, AC susceptibility and critical current density. However, the surface resistivity ([R.sub.s]) is the most important criterion for a superconductor when used in a passive microwave device. Measurements of [R.sub.s] were made at 35 GHz over a temperature range from 10[degrees]K to room temperature. A cylindrical [TE.sub.011] mode cavity was used in which the sample serves as one end wall.  By comparing the temperature dependent Q and resonant frequency data for a superconducting sample with the data for a copper standard, the [R.sub.s] for the superconductor as a function of temperature can be calculated, as shown in Figure 1. One drawback of this technique is that the experimental uncertainty that results from using the difference between two similar Q values limits the resolution, in this case, 8 m[Omega]. From Figure 1, one can see that [R.sub.s] falls below the minimum resolution for temperatures lower than 80[degrees]K. These results are consistent with other reported values.  The surface resistivity of a superconductor can be calculated for a particular frequency of interest from the [R.sub.s] value measured at some frequency by assuming that the surface resistivity scales as the square of the frequency. The quadratic dependence of [R.sub.s] is expected as a consequence of the two-fluid model of superconductivity,  and is supported by experimental results.  If a value of 1 m[Omega] is given for [R.sub.s] at 10 GHz, the [f.sup.2] scaling provides a conservative approximation to the current state-of-the-art surface resistivity values at 77[degrees]K. The materials technology is still improving and [R.sub.s] values will most likely decrease further.
Device and Circuit Fabrication
An important issue in the fabrication of an operational HTS circuit is the post processing of the laser ablated film. Since the majority of circuits are in a microstrip configuration, the application of a ground plane to the substrate and a pattern etch process is essential. This initially proved to be a difficult task, since during laser ablation a thermally conductive and adhesive silver paste was used to hold the substrate to the heater, which is still adhered to the back of the substrate after the ablation process. A complete cleaning process for the removal of the residual silver paste was developed. This process includes a plasma cleaning followed by a silver etch. Once the substrate is clean, a 1 [micrometer] silver ground plane is evaporated onto the back of the substrate. The HTS films then are coated with photoresist and baked at 90[degrees]C. These films then are patterned and etched using a saturated solution of ETDA.
After the circuits have been fabricated, they are mounted in a brass housing with RF connections made via K connectors. Adhesion between the HTS ceramic and K connectors is made with silver epoxy. The ground plane contact between the package and circuit is made either with a spring loaded pressure contact or silver paste. Good ground plane is essential for optimum performance.
The cryogenic environment required for HTS materials poses certain problems not encountered with other measurements. Before HTS can be applied to microwave circuits, a method for testing had to be developed. The testing method must involve a calibration routine so meaningful results are obtained. One difficulty is the loss associated with the extra lengths of coaxial cable typically required to connect the ANA to the circuit inside the cryogenic chamber. The cable loss cannot be corrected since the temperature gradient down the cable is unknown. The transmission characteristics change considerably with temperature and will change with time during a measurement procedure, nullifying any correction factors. This problem can be corrected by devising a method whereby a fixed length of cable is always cooled to the same temperature, thus producing repeatable losses.
Figure 2 shows the used cryogenic system. A stainless steel chamber was fitted with foam insulation. A copper liner, which contains liquid nitrogen (LIN), was made to fit inside the foam. The copper liner reduces any temperature gradient inside the chamber. Under the top of the chamber, a rectangular copper container was mounted that also contains LIN. Semi-rigid coaxial cable runs from the outside lid through the upper container, ensuring minimum heat transfer from the room into the device under test (DUT). The upper container is filled through a fitting in the top plate and automatically fills the lower container through a fixed overflow downspout. A level indicator is mounted on the chamber wall to control the LIN level. Heating coils mounted on the SMA bulkhead connectors maintain constant temperature on the test cables. The chamber can maintain 77[degrees]K temperature for days depending on the supply of liquid nitrogen. An external tap provides vacuum to the inner DUT fixture, if needed.
A standard two-port calibration can be run for this testing arrangement. Before this could be done, the impedance of several 50 [Omega] terminations was measured over the temperature range using a digitizing oscilloscope in TDR mode. Only two terminations were found to maintain 50 [Omega] impedance while at 77[degrees]K. These terminations were used to calibrate the cryogenic chamber.
HTS Circuit Design
HTS filters and resonators can produce great improvements in system noise temperature and stability. Because of the extremely high Qs attainable with these materials, current devices, such as cavity filters, may be replaced by HTS filters with equal or better performance.
Generally, microwave filters, which may be realized in some sort of planar structure, can be fabricated using HTS thin films. No special techniques or methods are needed when average dimension substrates are used. Neglecting the surface resistivity [R.sub.s], the equation for superconducting microstrip phase velocity [V.sub.ps] is,
[Mathematical Expression Omitted]
[Mathematical Expression Omitted]
d = the substrate thickness
[lambda.sub.0] = the zero temperature magnetic penetration depth  (on the order of 1700 [angstroms] for YBCO thin films) 
This shows that for d >> [lambda], [V.sub.ps] is equivalent to its normal conductor counterpart. This allows a circuit designer using HT superconductors to use conventional methods employed for normal conductors when d is large compared to [lambda].
Lowpass and bandpass filters, as well as ring resonators, were constructed employing various configurations. A distributed element bandpass filter is shown in Figure 3. The passband is centered at 10.5 GHz with a 20 percent bandwidth and a maximally flat response over frequency. This configuration was chosen since the high Q achievable with HTS will produce steeper skirts than the conventional maximally flat response while maintaining a flat passband. The circuit was actually that of a distributed element lowpass filter, and thus had a periodic passband that was used as a bandpass filter. The filter design was optimized using an MDS CAD system. It then was patterned in YBCO on an MgO substrate of 20 mil thickness and tested in the cryochamber with an ANA. The transmission response is shown in Figure 4. This filter also was inserted into a simple communications system and operated for several weeks and temperature cycled many times.
A second bandpass filter was designed in a six-pole coupled microstrip line configuration. This filter, shown in Figure 5, is centered at 10.5 GHz and has a 10 percent bandwidth and a Chebyshev type response. The limited size of the substrates available required a 100[degrees] bend in an intermediate line to reduce the overall length. This bend was modeled using the Sonnet em electromagnetic simulator and modified to correct for stray coupling. The filter was patterned on a 20 mil [LaAlO.sub.3] substrate with a gold groundplane. The measured vs. modeled response is shown in Figure 6. The difficulty in producing double-sided HTS substrates has created interest in structures requiring a single conducting plane, as well as inverted and suspended microstrip, which also does not require double-sided substrates. A hybrid coplanar (CPW)/slotline approach was taken where a distributed element filter was realized with CPW series elements and slotline shunt elements, as shown in Figure 7.
It is a natural question to ask where in a practical system can HTS devices be inserted, and what are the overall system benefits. In the case of the superconducting microwave filters, a potential application is in the front end of a channelized receiver or any receiver application. A noise figure analysis of a typical (two-element) block configuration, as shown in Figure 8, was performed.
A typical noise figure of 1.5 dB was chosen for the low noise amplifier (LNA). Figure 9 is a graph of noise figure vs. insertion loss of the filter generated using the Ferris equation,
[Mathematical Expression Omitted]
[F.sub.1] = noise figure of filter
[F.sub.2] = noise figure of LNA
[G.sub.1] = insertion loss of the filter
[G.sub.2] = gain of the LNA
In reality, a system designer would not use a filter with an insertion loss of 0.6 dB for a channelized receiver, and an insertion loss of 0 dB for a superconducting microstrip filter is unachievable due to dielectric loss and radiation effects. The advantage of the superconducting filter is its ability to match the performance of a waveguide filter combined with the advantage of size and weight reduction of a microstrip configuration. The constructed filters were 0.5" X 0.5" on an MgO 20 mil thick substrate mounted in a brass package.
One of the HTS bandpass filters was tested in a microwave digital communications system. The filter element was in the receiver portion of the system and the peak power of the RF pulses in band was 17 dBm with 50 percent duty cycle at a rate of 10 KHz. The filter was cooled in the cryogenic chamber and was cycled from room temperature to 77[degrees]K each day for two weeks. No special precautions were made to protect the filter from the environment. After two weeks of field tests, the filter was returned to ETDL and remeasured under controlled conditions; it was discovered that there was minor degradation in insertion loss. This observation concurs with current literature claiming that if the produced films are of a high quality and uniformity, then the material is less susceptible to moisture damage.
Techniques using the low loss superconducting materials to stabilize microwave oscillators also are being pursued. Two different design approaches have been taken. One design uses a superconducting microstrip ring that has a fundamental resonant frequency of 10.5 GHz. For coupling purposes, this ring was placed directly next to the drain transmission line of a MESFET. All DC bias leads and impedance matching are done in gold microstrip on 25 mil aluminum. The MESFET was mounted in a common source drain stabilized oscillator configuration. The gate length has been turned so that the FET exhibits negative resistance at 10.5 GHz, and the superconducting ring is used as a high Q element to lock the oscillator to the desired frequency. This configuration is used commonly with DROs and the design equations are readily available.  The entire circuit is mounted in a brass package with a silver epoxy, sealed and then submersed in liquid nitrogen where it reaches an ambient temperature of 77[degrees]K, as shown in Figure 10. This design was first tested using a gold ring on aluminum, where the entire package was cooled to 77[degrees]K to test the MESFET operation.
The second oscillator also operated at X-band. The design consists of two wideband 6 to 18 GHz distributed MMIC amplifiers, an uneven power splitter, a superconducting resonator and a phase shifter, as shown in Figure 11. In this design, only the resonator is cooled and connected in a feedback loop via 50 [omega] stainless steel coax cables. The phase shifter, amplifiers and power splitter are all contained in separate packages at room temperature. Two MMIC amplifiers each with 6 dB of broadband gain are used to compensate for 10 dB of the resonator's insertion loss and to provide the loop gain. The phase shifter is used to ensure that positive feedback for instability is achieved. The advantage of this design, although not yet practical for compact systems, is that the resonant element is easily changed with resonators of different configurations at frequencies within the 6 to 18 GHz band.
 W.G. Lyons and R.W. Withers, "Passive Microwave Device Applications of High [T.sub.c] superconducting Thin Films," Microwave Journal, Vol. 33, November 1990, pp. 85-102.
 Comprehensive Coverage of the Most Recent Research Efforts in Superconductivity Can Be Found if the Proceedings of the 1990 Applied Superconductivity Conference, to be Published in the March 1991 IEEE Transactions on Magnetics.
 S. Kaur, A. Fathy, J. Matey and R. Brown presented at the International Conference on Electronic Materials, Newark, NJ, September, 1990 to be published in the Conference Proceedings.
 R. Brown, V. Pendrick, D. Kalokitis and B.H.T. Chai, Appl. Phys. Lett., Vol. 57, No. 13, 1990, p. 1351.
 See, for example, J.S. Martens, J.B. Beyer and D.S. Ginley, Appl. Phys. Lett. Vol 52, No. 21, 1988, p. 1822.
 A. Inam, X.D. Wu, L. Nazar, M.S. Hedge, C.T. Rogers, T. Venkatesan, R.W. Simon, K. Daly, H. Padamsee, J. Kirchgessner, K. Moffat, D. Rubin, Q.S. Shu, D. Kalokitis, A. Fathy, V. Pendrick, R. Brown, B. Brycki, E. Belohoubek, L. Drabeck, G. Gruner, R. Hammond, F. Gamble, B.M. Lairson and J.C. Bravman, Appl. Phys. Lett. Vol. 56, No.12, 1990 p. 1178.
 T. Van Duzer and C.W. Turner, Principles of Superconductive Devices and Circuits, Elsevier, 1981.
 G. Ghione and C. Naldi, "Coplanar Waveguides for MMIC Applications," IEEE Trans. on Microwave Theory and Techniques, Vol. MTT-35, March 1987, p.260.
 J. Halbritter, "RF Residual Losses, Surface Impedance and Granularity in Superconducting Cuprates," Journal of Applied Physics, Vol. 68, No. 23, December 15, 1990.
 G. Gonzales, Microwave Transistor Amplifier Design, Prentice Hall, 1984.
Richard W. Babbitt received his BS degree in engineering physics from Lehigh University in 1958. Currently, he is a senior project engineer in the Microwave and Signal Processing Devices Division of the US Army Electronics Technology and Devices Laboratory, LABCOM. He is a member of the IEEE, MIT Society. Babbitt has numerous patents in the areas of microwave and magnetic devices. Currently, he is involved with superconducting microwave devices and electronic scanning techniques.
Erik H. Lenzing received his Eng Tech degree from ITT Technical Institute and his BS in applied mathematics from Monmouth College. Currently, he is pursuing his MS in applied mathematics at Stevens Institute of Technology. Lenzing has spent two years with the US Army's ET&D Laboratoty with the microwave/mm-wave team and is actively involved in CAD/CAE and fabrication of passive microwave devices. Current interests are in methods of testing of [HT.sub.c] superconducting microwave structures.
Adam Rachlin received his BSEE degree from Rutgers University in 1989, and is currently pursuing a MSEE degree. He has been with the Microwave and Signal Processing Devices Division of the US Army Electronics Technology and Devices Laboratory, LABCOM for two years as a research engineer. His primary area of concentration is the application and insertion of high temperature superconducting microwave devices into military radar and communication systems with an emphasis on frequency control.
William D. Wilber received his PhD in physics in 1987 from Colorado State University. In 1984, he joined the US Army Electronics Technology and Devices Laboratory as a research physical scientist. From 1984 to 1987, he was a member of the magnetic materials team, where he and his colleagues studied the mechanisms for microwave power loss in hexagonal ferrites at mm-wave frequencies. Since 1987, he has been a member of the device electronics team, which is responsible for the development of high [T.sub.c] materials. His main involvement is in laser-assisted film deposition and in the basic materials characterization.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||microwave electronic devices|
|Author:||Babbitt, Richard W.; Lenzing, Erik H.; Rachlin, Adam; Wilber, William D.|
|Date:||Apr 1, 1991|
|Previous Article:||Patents and copyrights: protecting intellectual property.|
|Next Article:||MDHL: bridging the design and communication gap.|