Processing, Microstructure, and Critical Current Density of Ag-Sheathed Bi₂Sr₂CaCu₂Oₓ Multifilamentary Round Wire

Ag-sheathed multifilamentary Bi2Sr2CaCu2Ox round wire is one of the leading hightemperature superconductors that can generate a magnetic field exceeding the maximum of ~23 T available in present Nb-based low-temperature superconducting magnet technology. However, the magnet fabrication of power-in-tube (PIT) Ag-Bi2Sr2CaCu2Ox multifilamentary round wire to develop critical current density Jc > 105 A/cm2 in magnetic fields up to 45 T is difficult, due to complicated material processing, as-yet incompletely understood microstructure, and the problem that Jc is sensitive to high-temperature reactions. This thesis analyzed the critical steps of melt processing PIT Bi2Sr2CaCu2Ox multifilamentary wires, systematically investigating the relationships between processing, microstructure, and conductor & magnet performance. The phase transformation and microstructure development during the melt processing of Bi2Sr2CaCu2Ox wires were thoroughly examined using a brine-quench technique that preserves the high-temperature microstructures. On heating to the maximum temperature (~890 °C), Bi2Sr2CaCu2Ox powder melts incongruently, producing a mixture of liquid and secondary solid phases. On subsequent cooling, the liquid reacts with the solid phases and Bi2Sr2CaCu2Ox reforms. The phase reaction to Bi2Sr2CaCu2Ox is often incomplete, leaving remnant nonsuperconducting phases from the melt and the Bi2Sr2CaCu2Ox phase and intergrowth in the superconducting matrix, all of which become current limiting mechanisms (CLMs) and block current flow. Moreover, the gas between precursor powder grains accumulates into large pores upon melting, which divide the filament into segments. The consequence of having large pores in the melt is that the pore regions may become bottlenecks for current flow in fully reacted wires. The high population of CLMs strongly indicates that the fraction of oxide filament area that is effectively used for carrying current is low and increasing the connectivity is the key to improving Jc of Bi2Sr2CaCu2Ox wires. The formation mechanisms of filament bridges that populate melt processed PIT multifilamentary wires were studied. Two types of filament bridges were found. Type-A bridges are single-grain Bi2Sr2CaCu2Ox that couple multiple filaments. Type-A bridges were suggested to enable an inter-filament current flow that may be important for increasing the superconducting cross-section for effectively carrying current. The Type-A bridges form because filaments can bond to adjacent filaments in the melt by Ag preferentially dissolving into liquid at Ag grain boundaries. This discovery of filament bonding and Ag&liquid transport has general application to the design and optimization of multifilamentary Bi2Sr2CaCu2Ox wires. Jc development through the melt processing was examined. Jc of wires doubled during the final cooling stage to room temperature. The fundamental cause of this Jc increase was identified as oxygen overdoping, which reduces the superconducting transition temperature, but increases the flux pinning and most importantly, improves the grain boundary current transport and connectivity. A critical limitation of Bi2Sr2CaCu2Ox for magnet fabrication is that melt processing yields an optimum Jc only within a narrow processing window (both maximum temperature Tmax and soaking time tmax need to be precisely controlled), which makes uniform heat treatment of large coils with large thermal mass difficult. The systematics of this temperature and time dependence were probed by examining the microstructure evolution and Jc of wires prepared at various tmax and Tmax using different melt processing schedules. The final Jc of wires was found to correlate weakly to Tmax or tmax, but it strongly correlates to tmelt, a hidden processing parameter that measures how long conductor spends in the melt state. This strong correlation between Jc and tmelt suggests that the Jc of wires is dominantly controlled by tmelt, not by Tmax or tmax, and careful control of tmelt creates a wider processing window for coils. Raising tmelt above an optimum time caused a decrease in connectivity and Jc. This Jc degradation was found to be associated with lower Bi2Sr2CaCu2Ox nucleation temperature and inhomogeneous Bi2Sr2CaCu2Ox nucleation. The fundamental cause of Jc decreasing with extended tmelt appears to be the Cu loss from the Bi2Sr2CaCu2Ox melt. There are three known ways of Cu loss in the literature; the new finding in the study is that Cu can be lost from wire by a fourth mechanism: Cu diffuses through the Ag from the filament to the surface of the wire where it evaporates as a copper oxide. Cu loss is a pervasive process during melt processing that can explain why nearly all wind-and-react Bi2Sr2CaCu2Ox coils have Jc values much lower than those of bare, short samples of Bi2Sr2CaCu2Ox wires melt processed at the same time. The study indicates that eliminating Cu loss would potentially raise the Jc of Bi2Sr2CaCu2Ox coils.