Part of the work was dedicated to build the entire setup to perform the experiments. A confocal microscope, with single-photon detection capabilities, was built to optically excite the nitrogen-vacancy centres using green light and to to collect their fluorescence, in the red/near infrared spectrum. An atomic force microscope (AFM), employing diamond tips mounted on tuning forks which enable electrical readout of the tip-sample force, was also developed from scratch. Part of the effort was dedicated to successfully testing and deploying a different type of scanning modality. While conventional AFMs scan the sample horizontally, the 15mm x 15mm chip size prevented the tip from reaching its centre. To address this, new mounts were developed, enabling vertical plane scanning and allowing the tip to access the membrane’s centre. A python code, based on the open-source python infrastructure Qudi, was used to control the confocal scanning, synchronize it with the AFM scanning, and to simultaneously acquire photon counts, while also controlling other pieces of of equipment in parallel.
The spin of the nitrogen-vacancy centre has a resonance at ~2.9GHz. To address it efficiently, a broadband microwave antenna, with proven range from ~100MHz to >4GHz, was designed. The antenna geometry, a gold circuit evaporated on a silicon substrate, was chosen to be compatible with the final spin-mechanical assembly where the membrane chip is glued to the antenna chip, such that the antenna-membrane separation, and thus the antenna-spin separation, is kept small (<2um). The antenna was also capable of delivering fast pulses to flip the spin-a capability which is required to induce spin-mechanical interactions-, which could be as short as 25ns, limited only by the bandwidth of the experimental instrumentation.
At the same time, the code infrastructure required to perform coherent control of the spin state was implemented through python. The low-level programming of the AWG was made into a high-level interface, which allowed to generate sequence of pulses with configurable shape and phsae. The interface was successfully tested to generate spin control sequences such as Rabi oscillation, Ramsey interference, and spin-echo.
A key aspect of the project relied on nanomagnets to generate the spin-mechanical coupling. Importantly, the nanomagnet had to be precisely placed at a point on the membrane where the largest oscillations occur. This presented challenges from the fabrication point of view. The nanomagnets could not deposited before the fabrication of the membrane, since the strong base used to release the membrane from its silicon substrate would remove most of the materials deposited on it. On the other hand, standard growth techniques cannot be used after the release of the membrane, since they would severely degrade its mechanical properties in the best case scenario. The solution was found in a technique called focussed electron beam ion deposition, which is analogous to 3D printing in the nanoscale. In this way, nanomagnets were successfully deposited on the membrane and their magnetic field was imaged, opening the way to spin-mechcanical experiments.
