MRFM combines the principles of magnetic resonance and atomic force microscopy.

working principle of MRFM

In magnetic resonance experiments, the core principle is based on the Larmor precession of spins in a magnetic field. This means that these spins start rotating with a resonance frequency that depends on the magnetic field that they experience. Atomic force microscopy works by scanning a thin needle (the so-called cantilever) over a sample and measuring the force between the cantilever tip and the sample.

In our lab, we use an MRFM setup with the magnet on tip geometry: It consists of an ultra-soft cantilever, a very floppy long needle and a micrometer sized magnet that we glue on the tip of the needle. We position the needle with the magnet above a sample of interest. The spins in the sample experience the field from the magnet and because the magnet is very small, there is a strong magnetic field gradient. This means the sample can be divided in very thin slices of constant magnetic field, the so-called resonance slices. All spins within one slice have the same Larmor precession frequency. If we send a microwave pulse with exactly this frequency, we can flip the spins only in this slice and all the other spins remain unaffected. Flipping the spins exerts a force on the cantilever, which is our measurement signal. We can probe different parts of the sample by varying the microwave pulse frequency and thereby create a 3D image.

Working at millikelvin temperatures requires a lot of adaptations in the setup. In MRFM, for example, we cannot read out the cantilever motion with a laser as that would heat up everything. Our cantilever detection is flux-based: we place a small pickup loop next to the sample and magnetically-tipped cantilever. Changes in cantilever motion change the magnetic flux threading the pickup loop, which we detect with a SQUID, a device that can measure the tiniest variations in magnetic flux.

This flux-based cantilever readout is unique and enables us to have the coldest MRFM set up in the world! It does come, however, at a price: due to the high flux sensitivity of the SQUID, fluctuating magnetic fields besides the one from the cantilever increase the noise level or even make measurements impossible. We can currently not perform any experiments requiring an external magnetic field as even the field from a normal electromagnet will produce too much noise, simply because fluctuations in the electrical current will lead to magnetic field fluctuations.

We are working on a solution for this: the persistent current switch. It consists of a superconducting coil in a closed loop. This coil can be fed with an electrical current and then decoupled from the normal conducting parts of the wiring. The superconducting current will then persist in the coil during the experiment, which can take days or even weeks. As there is no current source connected anymore, the persistent current produces an almost noise-free magnetic field.

Another requirement at low temperatures are low-dissipation motors to move the cantilever around. We use so-called Piezo knob slip stick motors for the “rough” positioning (taking micrometer-sized steps) and a home-built fine stage with shear Piezos for positioning with nanometer accuracy.

We also need to send the MRFM microwave pulses for spin manipulation producing as little heat as possible. At the moment, this is done by sending an AC current through a superconducting wire, which is dissipating much less than the commonly used microwave coils. There is, however, some dissipation left and therefore we are limited to spin manipulation protocols that require only moderate microwave fields. We are working on increasing our microwave power while keeping dissipation low.