Low-vibration cryostat with nanopositioners and vector magnet

We explore two approaches for laser frequency stabilization that make use of novel frequency references, and which require a cryostat with specifications as outlined below. Approach 1 uses a fiber ring resonator on a vibration-isolating spool, while approach 2 uses waveguide ring resonators in erbium-doped silicon whose performance is enhanced by a reduction of the light's group velocity due to spectral hole burning in the inhomogeneous erbium absorption line. To serve this purpose, the cryostat will have to fulfill the following specifications. We require a base temperature of < 1.8 K to ensure sufficient lifetime of the spectral holes, a temperature stability of <1 mK to minimize thermally induced frequency fluctuations in the frequency references and a cooling power of >= 2 mW to maintain low T under optical and microwave driving, which is required for controlling the nuclear spin bath to decouple magnetic noise and improve the spectral hole properties. Efficient experimental work relies on frequent and rapid sample exchanges, so the cooldown of the sample inset must take <5 h. Fully automated control of temperature and magnetic field, at all temperatures from base to room temperature, is needed for measurements at varying and well-defined conditions. The temperature and magnetic field control must be automated and available via ethernet. Since Helium is a very expensive and limited resource and we have no infrastructure for He recovery, we require a closed-cycle cryocooler that does not consume liquid Helium or other liquid gasses, which may further be a source of vibration detrimental to the stability of our frequency references. We do require, however, a He exchange gas environment around the sample for efficient thermalization of thermally weakly coupled samples. In order to fit in the constrained space in our laboratory, the outer dimensions of the cryostat must not exceed 1400 x 1400 x 1400 mm^3. If there is an external rack, e.g., for electronics, it must not be bigger than 1200 x 1200 x 1200 mm^3. In order to facilitate long-term measurements, the system must be able to keep its base temperature for more than three months. The cryostat's sample space needs to have a diameter of >49.5 mm to hold the vibration-isolated fiber spool without prohibitive fiber bend loss. For efficient evanescent coupling between fibers and silicon ring resonators, the samples must be mounted on closed-loop XYZ sample positioners with a range of > 3 mm in all directions, a positioning reproducibility of 100 nm and a load capacity of 100g to ensure reliable successful coupling. To this end, the vibrational stability between a rigid fiber mount (built by us) and the sample on the positioner must be lower than 1 nm in all spatial directions. The system must have at least 12 unoccupied, low-resistance DC lines wired to the sample stage for sample positioners, position feedback and electrical control signals. Optical imaging of the sample for efficient coupling, using an existing cryo lens, requires an optical access to the sample with a window of at least 20 mm diameter. For reproducible spectral hole burning, it is essential to freeze out paramagnetic impurities in the devices. A base temperature of 1.8 K corresponds to a thermal energy of about 40 GHz. To freeze out the electron spins at that temperature, where the spins experience a magnetic-field dependent Zeeman energy shift of 30 GHz/T, the system must come with a magnet that provides a field up to 9T. To align this field precisely along a specific direction in the crystal, additional magnets with a strong, but slightly weaker field are required, which can be implemented with a vector magnet that supplies fields of 9:1:1 T. On the other hand, measurements in the absence of magnetic fields are also required, so that the entire system must be constructed in a non-magnetic way. In order to test the full potential of our approaches for laser frequency stabilization, all noise sources originating from the cryostat must be minimized. Besides the temperature fluctuations mentioned above, another noise source to be considered are vibrations, which besides their detrimental effect on the frequency stability are critical to keep at a minimum for stable evanescent coupling between optical fibers and the silicon ring resonators. We distinguish between noise in the z direction (along the sample tube axis) and noise in the xy plane (perpendicular to the sample tube axis). For our purposes, the vibration level in z direction, which coincides with the axis of the vibration-isolating fiber spool, has to be <0.1 nm (rms, constant force at 4 K, 5 ms pixel time), while the vibration level in the xy plane has to be <1 nm (rms within 200 Hz, measured between sample and fiber). We will only consider quotes from suppliers that require a payment of less than 50% of the total price upon order. Any requested prepayments must be accompanied with a bank guarantee, whose cost is covered by the supplier.