High-Temperature Superconductivity: A Metallurgical Approach by Angelo L. Giorgi, Gregory R. Stewart, James L. Smith, and Bernd T. Matthias A mysterious 30-fold enhancement of the critical temperature in the yttrium-iridium system has been traced to the formation of a eutectic structure and to a dramatic decrease in the stiffness of the crystal lattice. S uperconductivity is the sudden complete disappearance of electrical resistance in some materials when they are cooled below a critical temperature. This phenomenon (Fig. 1) has intrigued solid-state scientists, metallurgists, and engineers ever since its discovery by Kammerlingh Onnes in 1911. Visions of many possible applications of superconductivity to electrical power generation and distribution, fusion reactors, high-energy particle accelerators, and propulsion systems, and of the many new superconducting devices have provided a powerful stimulus to the search for superconducting materials with high superconducting critical temperature (T). Development of a microscopic theory on superconductivity (the BCS theory) in 1957 brought hope that the theory would prove useful in predicting new high T materials. Unfortunately, none of the predictions have been successful. Many of the new high T, superconductors shown in Table I, which form the basis for the present superconducting c technology, were discovered by applica- In 1964, when workers at the Bell c only 0.1 K, and neither yttrium nor YIr2 (the only known compound in the system) become superconductors even when they are cooled to 0.3 K. The immediate conclusion was that another compound, probably with a composition close to YIr4, must exist and that this hypothetical phase was the source of the superconductivity. Many samples, prepared with compositions varying between iridium and YIr2, were heat treated and examined by x-ray diffraction. All attempts to discover a new phase were unsuccessful. The source of the enhanced superconductivity Fig. 2. Empirical behavior of T as a function of the average number of valence electrons per atom (e/a) for pure elements and alloys. This empirical relationship was discovered by B. T. Matthias in 1957. Fig. 3. A typical phase diagram for a mixture of metals showing melting point (M.P.) minimum at the eutectic composition. Solution of B into A lowers the M.P. of A; similarly, solution of A into B lowers the M.P. of B. Solid a is a solid solution of B atoms dissolved in the crystal lattice of A. Solid ẞ is a solid solution of A atoms in the crystal lattice of B. remained a mystery. The investigation was finally abandoned and all but forgotten. This year, when interest in the Y-Ir system was revived, a much more thorough study was conducted at the Los Alamos Scientific Laboratory (LASL) in collaboration with the University of California (UC) at La Jolla, California. The careful characterization of the various compositions was extended to include low-temperature specific heat measurements, metallographic examination, and transmission electron microscopy as well as the usual x-ray diffraction and magnetic susceptibility measurements. From this study, we have learned that the source of the enhanced superconductivity in the Y-Ir system is a eutectic structure consisting of a mixture of iridium and the neighboring phase, YIr,. A eutectic is the unique mixture of two constituents, usually metals, that has the lowest melting point. Eutectics have been known since the days of the Roman Empire and have been in continued use since then. Bronze and solder are common examples. A typical phase diagram for a eutectic mixture is shown in Fig. 3. The melting point for the eutectic composition is considerably lower than for either constituent. The low-temperature specific heat measurements on the Y-Ir samples disclosed that, with the formation of the eutectic structure, the stiffness of the lattice decreases dramatically. The enhanced superconductivity probably is a result of the lattice softening. This intriguing solution to the mystery promises a new investigative approach to hightemperature superconductivity. Measurement Techniques Reviewing some of the techniques used to measure the superconducting critical temperature will help explain the problems associated with a study of this type. At present there are over a thousand known superconducting materials. Their critical temperatures range from minimum values of a few millidegrees above absolute zero to the present maximum value of 23.2 K. Most of these materials are intermetallic alloys; that is, they are mixtures, compounds, or solid solutions of two or more metals. The three methods used most commonly to determine superconductivity in such materials are based on three distinctive thermal and electromagnetic properties of the superconducting state: 1. The complete disappearance of electrical resistance. 2. The complete exclusion of magnetic fields, up to some critical value He, from the body of the superconductor. 3. A sharp increase in the electronic specific heat of the superconductor at Te owing to the marked decrease in the entropy. superconducting, the resistance and, therefore, the voltage suddenly drops to zero. The temperature at which the drop occurs is the T. of the material. This method suffers from a demonstrated weakness. It requires a continuous superconducting path across the material, and it gives no indication of how much of the material is superconducting. Quite often the bulk of a material is not superconducting but instead contains microscopic filaments in a superconducting phase. A resistive measurement on such a sample gives results similar to those for a true bulk superconductor because the filaments, having zero resistance, short-circuit the sample and cause the voltage to drop to zero. Using x-ray diffraction to determine the phases present in a sample also can be misleading. If, as is often the case, the filaments in the superconducting phase represent only a small fraction of the total sample, their concentration may lie below the detection limit, and the xray diffraction pattern will indicate that the nonsuperconducting bulk material is the only phase present. An investigation limited to these techniques can lead to the erroneous conclusion that the phase representing the bulk material is the superconductor. A much more widely used technique is the ac susceptibility method, based on the exclusion of magnetic fields (Fig. 5). A stable alternating current of less than 60 hertz is applied to the primary of a sensing coil. The output from two matched secondary coils, connected in opposition, is fed to the input of a frequency-locked amplifier. A small variable mutual inductance is connected between the primary and secondary circuits and adjusted to cancel out any residual imbalance in the secondary circuit. The sample, which can be a powder or solid, is placed in one of the secondary coils and cooled slowly. A sharp change in the sample's magnetic permeability at T, causes an imbalance in Fig. 5. Technique for measuring the change in magnetic permeability of a sample when it becomes superconducting. A rapid change in the electrical signal from the secondary coils caused by expulsion of the magnetic field from the sample indicates the onset of superconductivity. the secondary circuit, which results in a signal at the output of the amplifier. Because the signal is proportional to the change in magnetic permeability, it gives a semiquantitative value of the amount of superconductor in the sample. The ac susceptibility method is preferred over the resistive method because it detects superconductivity in the individual particles and does not depend on a continuous superconducting path. However, even this method can give misleading results. Occasionally, a superconducting phase is deposited as a thin film at the grain boundaries of a bulk material. When the film completely encloses the grains, it acts as a superconducting can; that is, it prevents any magnetic flux from penetrating into the canned bulk material. Under these conditions, the signal strength from the amplifier suggests that all the material is superconducting. The method generally accepted as giving a true indication of both the presence and amount of superconducting material is measurement of the variation of the low-temperature specific heat with temperature (Fig. 6). The specific heat (C) is the amount of energy (dQ) required to raise the temperature of a unit mass (usually a mole) of material by a small increment (dT). where y depends on the properties of the electrons and ẞ depends on the crystal lattice. As shown in Fig. 7, these parameters are determined by plotting C/T vs T2. The slope of the resulting straight-line-curve is equal to ẞand the intercept is y. The information about the normal-state properties contained in these two parameters is discussed later. When the metal becomes superconducting, the lattice contributions to the specific heat (BT3) remain the same, but the electronic contributions (YT) change dramatically. The specific heat (C) for the superconducting phase is given by allowable energy levels that the electrons -A/kT c (4) Thus the ratio of the electronic specific heats at T. can be used to estimate the amount of the sample that has become superconducting and to rule out ambiguities caused by either microscopic amounts of one-dimensional paths in the resistive method or two-dimensional grain boundary cans in the ac susceptibility method. Further, if x-ray diffraction indicates the presence of a second phase, the amount of second phase estimated from a comparison of the relative intensities of the diffraction patterns can also be compared to the ratio of the electronic contributions to the specific heat. c A much more accurate estimate of the amount of superconducting material is possible for materials with T, values at least four times greater than the lowest temperature to which the specific heat can be measured (for example, a T > 4.8 K and a lowest temperature of 1.2 K). The superconducting phase contributes significantly to the specific heat at Te, but its contribution dies away exponentially as the temperature drops below Te, when the only terms left in the specific heat behavior vs temperature expression are BT3 and any yT contribution from nonsuperconducting material. The ratio of this remnant yT to the value of YT above Te, where the sample is 100% nonsuperconducting, is the exact fraction of the material that is nonsuperconducting. Fig. 6. Technique for measuring low-temperature specific heat. The sample, mounted on a sapphire disc 25 μm thick, is suspended in the center of a metal frame by 75-μmdiameter gold wires. The gold wires isolate the sample and platform thermally while providing two separate electrical connections to the metal frame. One connection supplies heat to the sample and platform through a film heater (the dark strip across the platform). The other connection measures the change in temperature by monitoring the change in resistivity of a germanium thermometer. The heat is supplied to the sample and platform in fixed increments by pulses of electrical energy to the film heater. Since sapphire is an excellent thermal conductor, sample and platform equilibrate rapidly after each pulse. By measuring the changes in temperatures for fixed increments in energy, the specific heat for the platform and sample is measured as a function of temperature. The total specific heat minus the known specific heat of the sapphire platform is the specific heat of the sample. |