Requirements on melt-textured Y-Ba-Cu-O for the use in magnetic bearings or electric motors

IEEE TRANSACTIONS ON APPLIED SUPERCONDUCTIVITY, VOL. 1 I, NO. 1, MARCH 2001 3501 Requirements on Melt-textured Y-Ba-Cu-0 for the Use in Magnetic Bearings or Electric Motors Tobias Habisreuther, Doris Litzkendorf, Oleksiy Surzhenko, Matthias Zeisberger, Robert Muller, James Riches, Stefan Schauroth, Jan Dellith, and Wolfgang Gawalek Abstract-Melt-textured Y-Ba-Cu-0 (YBCO) is prepared in a batch process with reproducible material properties. Functional models of magnetic bearings and electric motors were equipped with this material. From the tests of these models the requirements on further material development can be derived. The quality of a magnetic bearing can be improved by increasing the size of the superconducting domains. The current density should be in the order of 200 - 400 A/md. Reluctance motors with output powers up to 38 kW at 77K were constructed. Further improvements can be achieved using large superconducting domains with higher current density. Index Terms- Electric Motors, Magnetic Bearings, Melt- Textured YBCO, Preparation, Material-Properties. INTRODUCTION ELT-TEXTURED Y-Ba-Cu-0 (YBCO) is available both in the quality and in the quantity to use it in first demonstrators. From the tests we can learn which properties of the material have to be optimized or improved for the use in magnetic bearings or electric motors. In this article we want to discuss the use of melt-textured YBCO in these applications and we will show some directions for the further material development. M BATCH PROCESSING The base for technical use of monolithic superconductors is the preparation technology. The material must have reproducible properties of a high quality. Also it must be available in the quantity that is needed for the application. A high efficient modified MTG batch process [l] was developed. Here in one furnace several monoliths can grow at the same time. In our batch process we use box furnaces with a useable volume of 30*30*30 cm3. Convection prohibits the building of large and constant thermal gradients in the furnace. The absence of large thermal gradients requires a low cooling rate of the furnace and thus very stable control of the temperature distribution in the furnace. The temperature- distribution in the furnace has to be very homogeneous to guarantee identical growth conditions for all samples. The box furnaces we use fulfill these demands. They are six side heated and each side is temperature-controlled. Seeding is used to prohibit spontaneous homogeneous nucleation and the seeds define the final orientation of the samples and also the starting point of the growth on the Manuscript received September 18, 2000. This work was supported by the German BMBF under Grant No. 13N6854A3 and by the German BMWi under Grant No. 03273055. All Authors are with the Institut fur Physikalische Hochtechnologie (IPHT), Winzerlaer Str. 10, P.O.B. 100239, D 07702 Jena, Germany (telephone: +49 3641 206106, e-mail: habisreuther@ipht-jenade). surface of the monolith. Either single crystals of MgO or isomorphous Nd123 or Sm123 can be used. MgO is commercially available in reproducible quality. The other materials can be prepared by melt-texturing. In the batch process we use a standard composition of Y ~ S B ~ ~ C U ~ O ~ . ~ with an addition of 1 wt.% CeOz. Y203 is added to a commercial YBazC~30-i.~ - powder (Solvay Barium Strontium GmbH). This precursor powder is homogeneously mixed. The shape of the green body is given by uniaxial pressing. Depending on the application, cylinders with diameters up to 50" or plates with a quadratic base up to 60 mm are pressed (Fig. 1). The height is about 18 mm. In one batch about 2 kg of melt-textured YBCO can be processed. This illustrates the production rate. Differently shaped monoliths can be prepared in the same batch. The increased production rate leads to the demand for adapted characterization techniques. A cross structure on the surface of the monolith with the seed centered in the cross is strong evidence for single grain material. For microscopic methods, polarization microscopy, SEM, TEM, or EDX the textured monoliths have to be destroyed and can not be used in the demonstrators. These investigations have to be performed by accompanying random sampling. Fig. 1: Melt-textured monoliths prepared in a batch process with a size up to 55 mm after the texturing process. 105 1-8223/01$10.00 0 2001 IEEE 3502 Most important for applications are superconducting material properties. The fastest method to characterize melt- textured YBCO is to measure the levitation forces. The method is directly related to the situation in magnetic bearings. The absolute forces on the other hand are dependent on the experimental set-up. The forces are limited by the permanent magnet. To extract a "figure of merit" for the quality of the superconductor from levitation force measurements is not an easy task. It is hard to compare levitation forces measured in different laboratories. Field mapping, measuring the trapped flux of a superconductor, gives an easier interpretation of the material quality. According to the Bean model [2] each domain is reflected by a cone shaped flux distribution. The base of the cone gives the size of the magnetic domains. The current density can be evaluated from the slope of the cone when size effects are taken into account [3]. This method also is non- destructive and excellent for the material characterization. But it takes a longer time in comparison to the force measurements. Therefore it is less suitable for standard characterization in the moment. Intrinsic magnetic properties of small domains cut from the large blocks are determined using a VSM with superconducting magnet up to 12 T or a vector VSM up to 2 T. The material prepared in the standard process shows critical current densities calculated from the magnetization loops using the Bean model of about 2 - 5 lo4 Ncm2 at 77K and OT and lo4 Ncm2 at 77K and 2T [4]. Due to the destruction of the textured monoliths these investigations can be performed as accompanying random sampling. MAGNETIC BEARINGS Most of the magnetic bearings proposed consist of melt- textured superconductors and permanent magnet systems. The magnetic field has rotation symmetry. The low friction of the bearings is given by the homogeneity of the magnetic field in this symmetry direction. Due to pinning effects the superconductors stabilize the permanent magnets in the other 5 directions (lateral displacement, tilting). In principle the superconductors can be placed at any position of the permanent magnets stray-field. Also there is no absolute necessity that the superconductors are arranged symmetrically. Nevertheless to reduce losses in the bearing the superconductor should also be arranged symmetrically. An example is show in Fig. 2. Starting from 40 quadratic plates the monoliths were cut and joined so that they cover a copper plate that is used as a cold head. The single pieces were fixed mechanically to the copper. The performance of a magnetic bearing is given both by the design of the permanent magnet and the quality of the superconductors. The permanent magnet system defines the theoretical limit of the arrangement [5] . This means that the maximum forcesand the maximum stiffness is given by the permanent magnet. A real superconducting material will always lead to a reduction of the forces and the stiffness. The quality of melt-textured YBCO is influenced by the critical current density and by the size of the superconducting domains. Because the magnetic induction of the permanent magnets is smaller than 1.5 T the properties of melt-textured material Fig. 2: 40 melt-textured YBCO monoliths arranged on a cold head of a magnetic bearing. must be optimized for this field region. High current densities achieved due to fishtail effects at 2 T will have no direct benefit for a magnetic bearing. Using a SmCo permanent magnet levitation forces of about 60% of the theoretical limit were measured. An increase in the current density of a factor 10 will only lead to an increase in the forces of a factor 1.5. The benefit for a magnetic bearing will be in the increase of the stiffness. This increase leads to a reduction of losses. Also the size of the domains have to be adapted to the permanent magnet system. Using a permanent magnet system that consists of several magnetic poles enhances the stiffness of the bearing. The size of the superconducting domains should be similar to the size of the magnetic poles. Smaller domains will reduce the levitation forces and the stiffness. If the domains are much larger than the poles of the permanent magnet system the benefit will be small. Another important point is the reliability of the total system. Cooling cycles and the mechanical forces must not lead to degradation effects. The mechanical properties have to be studied. The material that is available today can be used for magnetic bearings. The main future tasks are to fabricate material that is tailored for the special design of the bearing and to guarantee the reliability of the material. Bulk superconductors are one component of a system. ELECTRIC MOTORS The YBCO parts in the rotors of electric motors are more complex shaped than the monoliths that were textured in the standard batch process. The design of the motor demands hollow cylinders, rings, plates or bars. These elements also are larger in size than the prepared monoliths. So the needed parts can not be prepared from a single textured monolith. To solve this problem a technique of cutting, bonding and grinding was developed [ 11. From the textured monoliths (cylinders, plates) special shaped blocks were cut. Diamond wire saw or blades were used. These monoliths were joined and specially bonded [I] . The bonding technique was specially developed because the bonded elements have to work under extreme conditions. The working temperature is 3503 2.0 cm Fig. 3: Plate consisting of 12 single domain melt-texurted YBCO-monoliths (upper part). Trapped field profile of the plate (lower part). 77K where normal glues become brittle. The rotors move with 3.000 rpm or even higher speeds. Therefore the mechanical stress on the bonded parts is very high. After bonding the parts get their final shape by grinding. A special computer controlled machine was used that was optimized for ceramic machining [6]. During thcse steps water was used to cool the tools and the monoliths. We did not observe negative influences of the water on the properties of the material after the mechanical treatment because the parts were dried immediately after each mechanical step. Finally a passivation, either by varnish or by paraffin was applied to avoid long term reactions of the final motor parts. For hystersis motors rings were prepared where the c-axis of each element was aligned in radial direction [7]. So the magnetic field of the stator was parallel to the c-axis. This artificial texture that could not be prepared by "pure" melt- texturing leads to a better use of the anisotropic material properties and thus to an increase of the output power or the efficiency of the electric motor. The second design is the reluctance motor. Here the rotor consists of combinations of soft ferromagnetic and superconducting materials. One possible rotor design is to stack plates of the two materials ("Zebra" design)[8]. Another design was developed at the Oswald Electromotor Company. Here bars joined from melt-textured YBCO were used ("Pilz" design). An output power of 38kW at 3.000 rpm was achieved at 77K [9]. In all designs the magnetic field is conducted in the ferromagnetic parts, the HTS bulk-elements act screening. We have rotors with a very high magnetic anisotropy. The magnetic field in motors is smaller than 2T. At higher fields the iron is saturated, Similar to the situation in magnetic bearings a high current density is needed up to 2 T. A very high output power with a high efficiency and a high power factor can be achieved if the superconducting material screens the magnetic field totally. Shielding of magnetic field with a hard type I1 superconductor can be achieved if a super-current can flow over a distance that is larger than the magnetic pole. The flow should not be interrupted by weak links, grain boundaries or cracks. If the current density is finite there is a penetration of flux into the superconductor according to the Bean model. This penetration is similar to an increase of the air gap in the motors and can reduce the efficiency the power factor and the output power. In reluctance motors the demands on a bulk superconductor are: the domain size should be as large as the magnetic pole in the motor and the current density should be as high as possible. Fig. 3 shows a plate fabricated from 12 single domain monoliths. The single parts were bonded, so no current can flow over the grain boundary. The lower part shows the trapped field profile of the plate. A motor equipped with "single domain" plates showed a higher efficiency and a higher power factor in comparison to a motor that was equipped with plates made from smaller domains [lo]. The first aim is to increase the domain size. Here the growth of melt-textured YBCO still has to be optimized. Another approach is to develop welding techniques [ 111, where a superconducting join between several single domain monoliths is achieved. The second aim has to be to increase the current density of the bulk material. This may be achieved by doping with special impurities [ 121, changing from the Y-Ba-Cu-0 system to other rare-earth 123 materials [13]. CONCLUSIONS From the investigation of the demonstrators we learned that the material properties of melt-textured YBCO have to be improved. Magnetic bearings work preferably in the field cooled mode. Magnetic flux has penetrated in the material. A high stiffness in magnetic bearings can be achieved by using a design with several magnetic poles. The size of the superconducting domains should be similar to the size of the poles. A higher current density will improve the performance of a bearing. In electric motors there are more demands for the improvement of the material properties. Again the domains of a superconductor should be similar to the magnetic poles of the electric motors stator. In comparison to magnetic bearings these poles are larger. New techniques to increase the domain size have to be developed. In addition the current density should be increased to about 1000 A/"*. By intrinsic pinning mechanisms this seems to be hard to achieve. Additional impurities [I21 or mixtures of rare-earth material are already tested. In both fields of applications, in magnetic bearings or in electric motors, bulk superconductors have to work in magnetic fields smaller than 2 T. Therefore the current density up to 2 T has to be high. Another possible field of application for this material is trapped field permanent magnets [13]. Here the critical current density has to be high 3504 even at higher magnetic fields. For this field of application the use of fish-tail [ 141 effects may be helpful. Replacing the Yttrium by other Rare Earth atoms or mixtures of Rare Earth atoms may lead to this increase of the current density with increasing magnetic field. With the knowledge from the demonstrators the material will be improved and used in the German "Dynastore" fly- wheel program and in electric motors up to an output power of 250 kW. ACKNOWLEDGMENT [5] [6] [7] M. Zeisberger, "Optimization of Levitation Forces", to be published in IEEE Trans. on Applied Superconductivily, 2001 Final precision machining performed by the BNM Jena, Germany T. Habisreuther, T. Strasser, D. Litzkendorf, M. Wu, M. Helbig, P. GBrnert, W. Gawalek, "Preparation of Rotors from Melt-Texured YBCO for Electromotors", Advances in Superconductivity X, Tokyo, L. K. Kovalev, K.V. Ilushin, S . M. - A. Koneev, K. L. Kovalev, V. T. Penkin, V.N. Poltavets, W. Gawalek, T. Habisreuther. B. Oswald, K. J. Best, "Hysteresis and Reluctance Electric Machines with HTS Bulk Rotor Elements", IEEE Trans. on Applied Superconductivify, Vol. 9, No.2, 1999, pp. 1261 - 1264. B. Oswald, M. Krone, M. Soll, T. Strakr, J. Oswald, K.-J. Best, "Optimization of Superconducting Motors with YBCO Bulk Material", 1998, pp. 937 - 940. [8] [9] The authors would like to thank P. Dittmann, M. Wotzel,, Ch. Schmidt, G, Bruchlos, M. for the technical support, K. V. Ilushin,, K. K. Kovalev, V. Penkin, L. K. Kovalev (MA1 Moscow), H. J. Gutt, V. Schlechter, A. Inst. Phvs. Conf Ser. No 167, Vol I. 2000. PP. 51 - 53. paper presented at the ICMC, Montreal, 1999. [lo] L. K. Kovalev, MA1 Moscov, Russia, personal communication, 2000. [ I l l H. Zheng. B. W. Veal, A. Paulikas, R. Nikolova, U. Welp, H. Claus, G. W. Crabbtree, "Top seeded growth and Joining of Bulk YBCO', , * Gruner(1EMAUni Stuttgart) and B. Oswald, K. J. Best, T. [I21 Z. H. He, 0. B. Surzhenko, T. HabisreLther, M. Zeisberger, W. Straljer (Oswald Elektromotoren GmbH) for constructing and Gawalek, "Preliminary Investigation on the Superconducting Properties of Textured YBaCu01YZBaCu05 Grown with the Addition of Nanoparticles", Inst. Phys. Con5 Ser. No 167, Vol 1, 2000, pp. 151 - testing the motors. 154. 1131 H. Ikuta, T. Hosokawa. H. Ishihara, K. Tazoe, M. Yoshikawa, Y. REFERENCES ~~ Yanagi, Y. Itoh, T. Oka, U. Mizutani, "Melt Processing of RE-Ba-Cu- 0 (RE = Sm, Nd) Bulk Superconductors and their Application to Quasi-Permanent Magnets", The 2000 International Workshop on Superconductivity, Shimane, Japan, 2000, pp. 170 - 173. [I41 M. Muralidhar, M. Murakami, "Flux Pinning in OCMG Processed ( N D O . ~ ~ E U O . ~ ~ G ~ O . ~ ~ ) B ~ ~ C U ~ O ~ Superconductors with Eu2BaCuO5 Addition", The 2000 International Workshop on Superconductivity, Shimane, Japan, 2000, pp. 130 - 132. [ I ] D. Litzkendorf, T. Habisreuther, M. Wu, T. Strasser, M. Zeisberger, W. Gawalek, M. Helbig, P. Gornert, Materials Science and Engineering B vol. 53 , 1998, p. 75 C.B. Bean, Phys. Rev. Lett. vol. 8, 1962, p. 250, Rev. Mod. Phys. vol. 36, 1964, p. 31 Th. Klupsch, T. Strasser, T. Habisreuther, W. Gawalek, S . Gruss, H. May, R. Palka, F. J . Mora, J. Appl. Phys. vol. 8 2 , 1997, p. 3035 M. Zeisberger, T. Habisreuther, D. Litzkendorf, 0. Surzhenko, W. [2] [3] 141 Gawalek, 'Inhomogeneous Magnetic Properties of Melt-Textured YBCO", Inst. Phys. Con5 Ser. No 167, Vol I , 2000, pp. 71 - 74.