Scanning SQUID Microscopy
Scanning SQUID microscopy generates images or local measurements of the magnetic properties such as magnetisation, susceptibility and nonlinear response of a sample by moving a magnetic sensor over the surface of the sample.The sensors we use are superconducting quantum interference devices (SQUIDs) with a very high sensitivity and a spatial resolution on the order of 1 µm. We are constructing a SQUID microscope operating at temperatures down to 20 mK and optimized for high frequency measurements. Future uses will include the scanning spin resonance measurements (NMR and ESR) for high resolution imaging and the characterization of ultra-small sample volumes. We are also interested in the origin of flux noise in superconducting devices, molecular magnets, persistent currents and other quantum phenomena.
Our microscopeCopyright: © Pascal Cerfontaine
Our home-built scanning SQUID microscope is operating in a dilution refrigerator and has a scan range of about 150 µm. Good spatial resolution is achieved by placing the sensor in the same vacuum space as the sample, thus not requiring any shields between them. Attocube positioners with a travel range of 5 mm enable navigating over large samples. Several microwave connections to the sensor allow the use of SQUIDs with dispersive readout for high sensitivity and bandwidth and the application of microwave excitiations to the sample. The microscope incorporates a capacitive position readout and feedback system in order to compensate piezo drift.
SQUIDs with dispersive readout
The SQUID sensors most commonly used today are so called DC SQUIDs, which consist of a superconducting loop interrupted by two Josephson junctions. A magnetic flux threading the loop leads to a change of their I-V characteristics, which can be detected via DC measurements. In order to achieve a higher sensitivity and bandwith, we have developed a novel type of scanning SQUIDs with dispersive readout . The underlying principle of this readout scheme is that a flux threading the SQUID changes the Josephson inductance of the junctions and thus the total inductance of the SQUID. Placing a capacitor in parallel with the SQUID forms an LC resonator with a flux-dependent resonance frequency. The resonance frequency can be measured via the phase of a microwave signal reflected by the SQUID. First tests indicate that order of magnitude improvements in sensitivity and bandwidth compared to most state of the art scanning sensors can be achieved with this method. Furthermore, the back action on the sample, which can lead to heating and disturb fragile quantum phenomena, can be controlled more easily.
Spin resonance imaging
The high bandwidth of our sensors and the ability to apply AC magnetic fielts with frequencies up to 20 GHz to the sample make it well-suited for spin-resonance measurements on electron and nuclear spins. The resolution set by the sensor size would be significantly better than traditional NMR and ESR methods based on field gradients, and estimates indicate that much fewer spins - possibly down to individual electrons - can be detected. Another interesting feature is that rather than requiring an externally induced average polarizsation of the sample, which is often difficult to achieve, the small sample sizes favor the detection of the statisical fluctuations of the net spin polarization. Finally, the micron-scale field coils integrated into our sensor enable the application of broad-band exciation signals without the use of bandwidth-limiting resonators.
Flux noise in superconducting devices
Superconducting devices such as SQUIDs and qubits have been found to exhibit magnetic flux noise with a 1/f like spectrum and a surprisingly universal intensity. For qubits, this noise is a strong source of decoherence . It has become clear that this noise arises from fluctuating spins at or near surfaces, but the nature of these spins is unknown and the physics governing their dynamics is not understood. We have shown that the same spins can also be detected and characterized via scanning susceptibility measurements . Our scanning SQUID microscope will enable detailed studies of the open questions and the systematic comparison of different materials and sample treatments and can thus contribute to solving this problem.
According to classical physics, threading a magnetic flux through a metallic ring induces a transient electric current that decays due to the resistance of the conductor. Quantum mechanics predicts that due to the quantization of energy levels, this decay is incomplete and that sufficiently small and cold metallic rings exhibit a nonzero current in their thermodynamic ground state. This persistent current is periodic in the applied flux, even when the latter is static. Our earlier scanning SQUID measuremrents  and other works  have confirmed this prediction for the contribution from non-interacting electrons. The improved performance of our new instrument should enable us to also study contributions from electron-electron interactions. Furthermore, it will be interesting to consider similar effects in exotic materials.