Multi-channel Scanning SQUID Microscopy

Abstract

I designed, fabricated, assembled, and tested an 8-channel high-Tc scanning SQUID system. I started by modifying an existing single-channel 77 K high-Tc scanning SQUID microscope into a multi-channel system with the goal of reducing the scanning time and improving the spatial resolution by increasing the signal-to-noise ratio S/N. I modified the window assembly, SQUID chip assembly, cold-finger, and vacuum connector. The main concerns for the multi-channel system design were to reduce interaction between channels, to optimize the use of the inside space of the dewar for more than 50 shielded wires, and to achieve good spatial resolution.

In the completed system, I obtained the transfer function and the dynamic range (Fmax ~ 11F0) for each SQUID. At 1kHz, the slew rate is about 3000 F0/s. I also found that the white noise level varies from 5 mF0 /Hz1/2 to 20 mF0 /Hz1/2 depending on the SQUID. A new data acquisition program was written that triggered on position and collects data from up to eight SQUIDs. To generate a single image from the multi-channel system, I calibrated the tilt of the xy-stage and z-stage manually, rearranged the scanned data by cutting overlapping parts, and determined the applied field by multiplying by the mutual inductance matrix. I found that I could reduce scanning time and improve the image quality by doing so.

In addition, I have analyzed and observed the effect of position noise on magnetic field images and used these results to find the position noise in my scanning SQUID microscope. My analysis reveals the relationship between spatial resolution and position noise and that my system was dominated by position noise under typical operating conditions. I found that the smaller the sensor-sample separation, the greater the effect of position noise is on the total effective magnetic field noise and on spatial resolution. By averaging several scans, I found that I could reduce position noise and that the spatial resolution can be improved somewhat.

Using a current injection technique with an x-SQUID, and (i) subtracting high-frequency data from low-frequency data, or (ii) taking the derivative of magnetic field Bx with respect to x, I show that I can find defects in superconducting MRI wires.