| I have designed, built and tested a cryo-cooled scanning SQUID microscope for imaging room-temperature samples, which uses a commercially available closed-cycle refrigerator to cool a high-Tc YBa2Cu 3O7-delta dc SQUID to 76 K. The system uses a custom dewar design for a minimum SQUID-sample separation of 50 mum. The flux white noise is 10.5 muphi0/Hz1/2 above 1 kHz. Below 1 kHz, 1/f noise is dominant, reaching 160 muphi0/HZ 1/2 at 10 Hz.;I used the microscope to perform non-destructive fault isolation on short circuits in modern microelectronic chips. Through a 120 mum thick silicon die, I isolated a power to ground plane short to within +/-100 mum at a SQUID-sample separation of 200 mum. In a multi-chip module package, I isolated a trace-to-trace short to within +/-18.5 mum at a SQUID-sample separation of 340 mum. These results helped lead to the successful commercialization of a SQUID microscope based upon my system. Using one of these commercial systems, I trained technical personnel and performed short-circuit fault isolation research at a major microelectronic manufacturer.;I also used the microscope for non-destructive evaluation of defects in low-Tc NbTi wires, comparing injected current vs. eddy current techniques. I found that the phase of the eddy currents was much less sensitive to geometry effects of both the sample and the defect when imaging fabricated test samples. I discovered that imaging injected currents with a SQUID oriented to detect the x-component of magnetic field produced a strong response to a defect in a NbTi wire, while the z-oriented SQUID produced no observable response to this defect.;I developed analysis techniques for calculating the magnetic pole density rho M, magnetic field H&ar; and magnetization M&ar; from images of Bz from in-plane magnetized samples. I used these techniques to screen sputtered magnetic combinatorial libraries of rare earth compounds. I found at least one interesting new compound, Fe11 Nd10Bx, which exhibited anisotropic characteristics and a remanent magnetization of at least 416 emu/cc. Finally, I developed a technique for making zero-applied-field non-remanent magnetization measurements, which would be useful to screen for the highest energy product (B*H) max of samples in magnetic combinatorial libraries. |
