MASTER THESIS NORMAL ZONE PROPAGATION IN A YBCO SUPERCONDUCTOR AT 4.2 K AND ABOVE A.R. Hesselink FACULTY OF SCIENCE AND TECHNOLOGY ENERGY, MATERIALS AND SYSTEMS GROUP EXAMINTATION COMMITTEE Dr. M.M.J. Dhallé Prof. Dr. Ir. H.J.W. Zandvliet Prof. Dr. Ir. H.J.M. ter Brake 04-03-2015 Abstract The normal zone propagation velocity and minimum quench energy of a 2-mm wide YBCO su- perconductin ’coated conductor’ tape are investigated at temperatures between 4.2 and 29 K, at three di(cid:27)erent magnetic (cid:28)eld strengths and at varying operating currents. Con(cid:28)rming earlier observations on a wider tape at higher temperatures and in contrast to the simplest and most widely used theoretical model, it was found that the normal zone propagation velocity predom- inately depends on the current and hardly so on temperature or magnetic (cid:28)eld. In agreement with theoretical predictions, the minimum quench energy was found to depend on both tem- perature and current, while the collected data do not allow to make a reliable conclusion about its magnetic (cid:28)eld dependence. A more sophisticated analytical and a numerical model con(cid:28)rm the temperature independence of the normal zone propagation velocity for temperatures below 25 K. Comparison between the absolute value of the normal zone propagation velocity show a quantitative di(cid:27)erence of 50% between the 2-mm and 4-mm wide sample. Although several likely causes were investigated, this di(cid:27)erence remains as yet unexplained. Contents Table of Contents 1 1 Introduction 2 1.1 Superconductivity and the critical surface . . . . . . . . . . . . . . . . . . . . . 2 1.2 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.3 Minimum quench energy and stability . . . . . . . . . . . . . . . . . . . . . . . . 4 1.4 Normal zone propagation velocity . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.5 Previous work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.6 Assignment and layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2 Experimental Aspects 8 2.1 Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.1.1 Probe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.2 Changes in instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.3 Measurements Ic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.4 Measurements Vnzp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2.5 Signal analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 3 Results 20 3.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 3.2 Mimimum Quench Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.3 Normal zone propagation velocity . . . . . . . . . . . . . . . . . . . . . . . . . . 22 4 Analysis 26 4.1 Analytical model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 4.2 Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 4.3 Comparison of normal zone propagation in the 2 - and 4 mm wide tapes . . . . 32 4.3.1 Possible causes for the di(cid:27)erences between the 2-mm and 4-mm tape . . 32 5 Discussion and conclusion 35 Recommendations 37 Acknowledgements 38 Nomenclature 39 Bibliography 41 A Manual NZP experiment 42 B Ic-graphs 65 1 Chapter 1 Introduction This report describes the continuation of measurements performed on the thermal behavior of a second generation high temperature superconductor wire. Superconductors are used in magnets, for they can handle very high current densities, required to create strong magnetic (cid:28)elds. In the last decade the HTS is available as a practical conductor in the form of tapes. HTS can handle higher temperatures, currents and magnetic (cid:28)elds than LTS. Therefore they are the future for high (cid:28)eld magnet systems. But the properties of HTS are less understood. The understanding of the thermal behavior of HTS is important for its protection. When a part of a superconducting magnet transitions to a normal, resistive state during operation, a ’quench’, the magnet has to be shut down or it can be damaged or even destroyed. So a quench has to be detected as fast as possible. The occurrence of a quench depends on the minimum quench energy and its detection on normal zone propagation velocity. 1.1 Superconductivity and the critical surface Superconductivity has been around for more than a century now, though the phenomenon of losing electrical resistance, is still unfamiliar to many people. The discovery was made by Heike Kamerlingh Onnes [1] in 1911, who cooled mercury to a temperature of 4.2 Kelvin in liquid helium. At room temperature a bad conductor (for a metal), mercury became a supercon- ductor. Since then many superconducting materials have been discovered and a fundamental theory has been developed. Below a certain temperature the materials lose their resistance: the critical temperature (T ). The (cid:28)rst class of superconductors (LTS) were metals, metal-alloys c and compounds, with critical temperatures varying from below 4.2 K to 30 K. Niobium Ti- tanium (NbTi)and niobium tin (Nb Sn) are now used commercially. In 1986, a new family 3 of superconductors was found: ceramic copper oxide materials with unpredicted high critical temperatures [2]. Many of these copper oxides had a T above 77 K, the temperature of boiling c nitrogen. They were called high temperature superconductors (HTS). Practical HTS materials are bismuth strontium calcium copper oxide (BiSCCO) [3] and yttrium barium copper oxide [4] (YBCO, see (cid:28)gure 1.1). Temperature is not the only parameter which is of importance for a superconductor. When transporting a current through the conductor, it will induce a magnetic (cid:28)eld (self (cid:28)eld). Su- perconductors have a special interaction with a magnetic (cid:28)eld. At (cid:28)rst, the magnetic (cid:28)eld is expelled from the interior of the material by inducing surface currents. Beyond a (cid:28)rst critical (cid:28)eld point, B , the ’type I’-superconductors (mostly the pure elements) lose their supercon- c1 ductivity. ’Type II’-superconductors (including all HTS) allow the magnetic (cid:28)eld to penetrate the material, as vortices of supercurrent with a normal core, enclosing a single magnetic (cid:29)ux The picture on the titlepage shows the wire, unfolded at the YBCO layer. The two pieces come from the remainder of a sample after a dramatic quench. 2 quantum. There are Lorentz forces that work on these vortices as well as repelling forces be- tween them. When the current is increased, more vortices move into the conductor from the outside. But the movement of the vortices causes an electric resistance, like friction. This is problematic, but defects in the crystal lattice can pin the vortices and so hinder their move- ment. With the increasing current, the vortices are compressed in the superconductor until the Lorentz force are larger than the pinning forces at the critical current (I ). The second, upper c critical (cid:28)eld (B ) is reached when, without a transport current, the vortices are compressed in c2 the superconductor until their normal cores overlap and no superconducting part is left [5] [6]. These three critical parameters are interdependent and can be combined in a graph, creating the so called (material-speci(cid:28)c) critical surface. The critical surface of YBCO is shown in (cid:28)gure 1.2. The surface shows the interface between the superconducting state and the non- superconducting (normal) state. A transition will occur when one passes from a point ’below’ the surface to a point ’above’. The critical surface of HTS materials is much higher than that of LTS materials. Of course there is a downside to the newer superconductor. HTS materials are ceramic and therefore fragile. Even more troublesome is the anisotropy of the superconductor properties: it is orientation dependent. The material has to be monocrystalline (green surface in (cid:28)gure 1.2) or the crystals have to be aligned with respect to each other a very small angle. Otherwise the current can not cross the interfaces between the crystals and the e(cid:27)ectiveness of the conductor is compromised. Wires with YBCO or REBCO (RE stands for Rare Earth) as the HTS component are made commercially as tapes or ’coated conductors’. By using thin (cid:28)lm technology to grow the HTS epitaxially on meter-long substrates, the problem of anisotropy is overcome. Figure 1.1: The unit cell of YBCO 1.2 Applications Superconductors can transport current densities in the order of 100-100 A=mm2 without any loss. The main application for superconductors are electromagnets, to build them into magnets. Using superconductors instead of copper, smaller and more powerful magnets can be made. Modern medical Magnetic Resonance Imaging (MRI) would not be possible without the use of superconductors, norcouldthelargerparticleacceleratorsoperatewithonlynormalconductors. 3 Figure 1.2: The critical surface of YBCO. The green top surface is for a perfect sample. The red surface is the practical engineering current density. CERN in Geneva, Switzerland, is developing the successor of the Large Hadron Collider (LHC). To increase the resolution of these devices, stronger magnets are needed and HTS are the future conductor materials to build these. In the ITER project, superconductors are used to power and control the plasma in the world’s (cid:28)rst energy producing fusion plant. The potential in the energy sector for superconductors is big, but material costs and the requirement of cryogenic temperatures is slowing down the commercialization. HTS may operate with liquid nitrogen, which is much more practical than liquid helium, but the costs of HTS wires are still relatively high. Although HTS can work at relatively high temperature, often they will still operate at 4.2 K in hybrid systems together will LTS materials. 1.3 Minimum quench energy and stability Superconductors are often used submerged in a bath of liquid cryogen during operation. Nev- ertheless they may heat up locally under in(cid:29)uence of an external disturbance. Especially at temperatures below 100 K, there is a chance that this will happen, because the heat capacity of materials drops rapidly ((cid:28)gure 1.3). A minimal amount of energy is needed to heat up the material: the movement of a strand due to Lorentz forces, cracking of the impregnation of a magnet or the impact of high energetic particle in case of an accelerator magnet. When a small area of wire has a transition to the normal state, Ohmic heating occurs in this local zone. If the ’normal’ zone is small enough and the cooling is su(cid:30)cient, the current can sort of bypass the normal zone and the zone collapses. If the zone is big enough, a chain reaction commences, where the size of the normal zone increases, generating more heat. This is called a ’quench’. The least amount of energy required to cause a quench, is called the minimum quench energy (MQE), see equation 1.1. If no countermeasures are taken, the temperature of the wire will become too high, eventually destroying it. To raise the minimum quench energy, a stabilizer material is used, such as copper or aluminum. The stabilizer has a higher thermal and electri- cal conductivity when the superconductor turns normal. It can adsorb heat and redirect the current, preventing an immediate temperature rise. Z Tt MQE = ‘ C(T)dT (1.1) MPZ T0 4 The temperature of a minimum length of the normal zone (‘ ) with a certain heat capacity MPZ (C(T)), has to be raised from the operating temperature (T ) above the transition temperature 0 (T ). The minimum length of the normal zone to cause a quench is called the ’minimum t propagation zone’, see equation 1.2: (cid:26)2k(T (cid:0)T ) (cid:27)1=2 t 0 ‘ = (1.2) MPZ (cid:26)I2 with k the thermal conductivity and (cid:26) the resistivity of non-superconducting material. I is the operating current. The di(cid:27)erence between T and T is called the thermal margin. 0 t Figure 1.3: Speci(cid:28)c heat of copper. Below the 100 K, the speci(cid:28)c heat drops rapidly. 1.4 Normal zone propagation velocity To create a high current density magnet, only a limited amount of stabilizer can be used. So quenches cannot be prevented and therefore a magnet has to be designed to allow quenches. During a quench, the magnet has to be protected against energy buildups, which cause high temperatures. The current has to interrupted and the energy stored in the magnet has to be dumped. This energy dump can be done in external resistors or in the cold mass of the magnet, by (cid:28)ring quench heaters. The quench heater create arti(cid:28)cial normal zones and in this way the energy is smeared out. Before the quench protection can be initiated, it has to be detected. And within a very small time frame. A simple and fast method is using voltage taps connected to the conductor, indirectly measuring a resistance. Before the voltage taps can detect a quench, the normal zone has to reach them. The ’normal zone propagation velocity’ (V ) is nzp the speed with which the superconducting-to-normal-transition front travels. It is shown for several superconductors in (cid:28)gure 1.1 as U . Looking at the HTS materials, the propagation is l very slow. For the protection, this is problematic. Especially for the expensive HTS materials. 5 Table 1.1: Typical literature values for the normal zone propagation velocities (U ) in several super- l conductors at depicted circumstances.[7] 1.5 Previous work The V has not been investigated extensively at low temperatures. Therefore the setup made nzp by H. van Weeren [8] for measuring the V of MgB , was used by J. van Nugteren [9] for nzp 2 measurementsonREBCOtapeinthetemperaturerangeof45to25K.Hefoundanexponential relation for the V depending on the sample current only (see (cid:28)gure 1.4). Due to limitations nzp of the setup, the measurements could not be evaluated at lower temperatures. 1.6 Assignment and layout The goal of the research presented in this thesis work is to extend earlier normal zone prop- agation measurements on REBCO HTS tapes to temperatures lower than 25 Kelvin, ideally all the way to 4.2 Kelvin. The measurements are done on a 2 mm wide HTS tape at varying magnetic (cid:28)elds strengths and currents. Speci(cid:28)cally, the goal was to check whether the unex- pected power-law dependence of V on current holds also in this temperature window, and to nzp establish the in(cid:29)uence of temperature and magnetic (cid:28)eld in more detail. After this introductory chapter, the report continues with four more chapters. Chapter 2 describes the setup, sample preparation and measurement procedures. The results of the mea- surements are presented in chapter 3. Results from previous research and new simulations are analyzed in chapter 4. In the last chapter, the outcome from measurements and simulations will be discussed and concluded. 6 e h t n o y l n o d n e p e d e r a s e i t i c o l e v e h T . ] 9 [ n e r e t g u N n a v . J y b d n u o f s e i t i c o l e v n o i t a g a p o r p e n o z l a m r o n e h t f o h pr. ao grct u dd en alo cc se ch it m h hg tu rio ar gh ot Lg n 4:ssi 1.pa et rn ue grr iu F c 7 Chapter 2 Experimental Aspects In this chapter, various aspects of the measurements are clari(cid:28)ed. The setup for measuring the MQE and V , is explained in section 1, including several modi(cid:28)cations to the probe nzp that were needed for this assignment. The instrumentation hardware and software is discussed in section 2. Then the experimental procedure for measuring the critical current and normal zone propagation is explained. In the last section, the signal analysis and accuracy procedure is discussed. For more information, a manual for the "NZP experiment 2014" is included in Appendix A. 2.1 Setup The setup for measuring the MQE and V was designed and build in 2008 by H. van Weeren nzp [8] for the characterisation of Magnesium Diboride (MgB superconductors). J. van Nugteren 2 [9] modi(cid:28)ed the probe and reassembled the setup for measurements on REBCO tapes. Also a new software environment was written to control the experiments, with additional protection measures. The setup is a so-called "time-of-(cid:29)ight" experiment. The sample is placed in a controlled environment, in a stabilized temperature and magnetic (cid:28)eld, transporting a stable current. A resistor is used as "quench heater" to create a normal zone. The heat pulse signal is registered to (cid:28)nd the quench energy. In case of a quench, the normal zone expands and the resulting voltage signal is recorded with voltage taps soldered to the wire at known distances from the heater. The normal zone velocity is calculated from the time interval that it takes the normal zone to travel over the distance between two voltage pairs, i.e. the time taken to reach a matching voltage level over the next pair. Challenging,butimportantinthiskindofexperimentisthatitshouldbeperformedasadiabatic as possible. Ideally, all the heat generated in the sample should only be used to drive the propagating normal zone further. However, it is not possible to execute these experiments under fully adiabatic conditions. Heat will (cid:29)ow away from the zone to the environment. Several precautions are taken to keep the sample in a ’quasi-adiabatic’ condition. First, it is held in a vacuum. Second, the electrical wiring to the voltage taps, heaters and temperature sensor consists of manganin, a copper alloy with a low thermal conductance. Third, ridges in the embedded heater structure ensure lower heat loss to the support underneath the sample. Heat loss from thermal radiation cannot be prevented. Also, the time scale of the experiment is relatively small and therefore losses to the sample holder are correspondingly small. Sample description The earlier NZP experiment on REBCO used standard-width 4 mm wide tape, but measure- ments could not be continued because the niobium tin current leads in the set-up could not handle the high currents that the HTS are able to conduct at temperatures below 25 K. The 8