3He-4He-Entmischungskryostat (cryo-free)
Final Report Abstract
We received our cryo-free dilution refrigerators early December 2012. The first six months were dedicated to cabling design, low-noise electronic set-up design and various tests. All inner cryostat parts have been designed to be “plug’n play” style and fully removable and interchangeable (filters, cables, sample holders). Our ultra-low temperature filters (RC filters and “powder filters”), sample holders and specially designed high-frequency systems (bias-T box and matching impedance circuits) have been designed PCBs within this view. The first real sample measurements were performed for three months on mid-2013. The results obtained from these experiments showed very high quality data demonstrating the excellent thermalization and filtering of our dc-lines. We benefited from already many electrical dc connections (48), we were able to measure two devices at the same time without any cross-talk, thanks to the well shielded set of connections and the individual connectors. In addition to the 12 T magnet (which was tested first at the end of 2013), we can use a smaller superconducting magnet (200 mT) to gain time in cooldown if no high magnetic field is required for our experiments. During the year 2014, the cryostat was implemented with 24 dc lines at 4K to provide power to cryo-amplifiers and RF-switches in order to perform high frequency shot noise detection and quantum limited charge detection at millikelvin temperature. We needed a lot of time to optimize our detection scheme and tests all circuitry (low temperature coaxial cabling and 50Ω impedance circuits on PCB) as well as RF components that are in general not design for low temperature measurements (amplifiers, switches, attenuators, dc-blockers, directional couplers, circulators, ). In 2015, an additional two times 24 dc lines at 4K and mixing chamber stage were added to perform additional experiments within one cool down as well as implementing the setup with a piezo-rotator which needs low impedance connection. The entire set-up needed many extra mechanical parts, all design by our team and made of gold platted copper for optimized thermal anchoring. The cryo-free dilution refrigerator is installed in a microwave-proof Faraday cage for high frequency noise measurements (100 dB attenuation at 1GHz). The experimental set-up is fully control by homemade Python programs. From this first dc measurements one article has been published already (the persistent hysteretic behavior down to mK temperatures of mica-graphene heterostructures), several others are in preparation (we demonstrated the tunability of the supercurrent by band gap engineering in hBN- bilayer graphene weak link, we studied the interplay between superconductivity and anti-Klein tunneling in hBN-bilayer graphene weak link, we observed a Berry phase transition in hBN-bilayer graphene heterostructures, we probed the snake states in hBN-graphene single layer heterostructures and we detected a shot noise suppression at multiple Andreev reflection in hBN- graphene single layer heterostructures). We have developed sample fabrication techniques which allow us to elaborate graphene based heterostructure of very high quality. We have designed mica-graphene-hexagonal boron nitride (hBN) heterostructure with double gate (back and top). While tuning the back gate which use mica as dielectric the resistance versus gate voltage characteristics show a strong hysteretic behavior persistent down to millikelvin while quasi absence of hysteresis was observed by sweeping the top gate the hBN as dielectric. While being both atomically flat and therefore might be viewed as ideal substrate for graphene, mica structure and high hydrophilicity makes it not appropriate to be used as gate dielectric.
Publications
- Persistent gate hysteresis in mica-graphene van der Waals heterostructures. Nanotechnology 26, 015202 (2015)
J. Mohrmann, K. Watanabe, T. Taniguchi and R. Danneau
(See online at https://doi.org/10.1088/0957-4484/26/1/015202)
