- Ultracold gallery
- Physics Courses
- SMART program
- Group talks
- MURI project
- On the web
- Postdoc search
- Qubit server
Experimental work and progress
Our setup is built to allow for trapping, cooling and manipulation of both Li-6 and Cs-133.
An atomic beam of Li is obtained from an oven heated to 400C. These hot atoms are slowed using a Zeeman slower and captured in a Magneto-Optical Trap (MOT) in the main experiment chamber.
Cs vapor obtained by heating metallic getter sources in the intermediate chamber are transversely cooled using a Two-Dimensional MOT. These atoms travel down the Zeeman slower tube and get captured by a MOT in the main chamber.
Both Cs and Li atoms need to be transferred to an optical dipole trap for further cooling using evaporation. Currently, we are loading Cs atoms into a crossed optical dipole trap formed using light from a 1064 nm ( 200 Watt YAG, multi-mode) laser.
A quantum degenerate mixture of Cs and Li atoms provides interesting prospects to investigate few and many-body physics and to realize quantum information processing. Simultaneous evaporative cooling of bosonic Cs-133and fermionic Li-6 is challenging. In an optical dipole trap formed with a 1064nm laser, Cs (resonant transition at 852nm) experiences a deeper potential than Li (at 671nm). Since laser cooling allows a much lower initial temperature for 133Cs (10μK) than for 6Li (150μK), a wide, shallow trap optimizes Cs loading, but Li requires a deep trap. Efficient cooling of Li involves Feshbach resonance tuning near 800G, but 133Cs collision properties are untested at high fields. Furthermore, sympathetic cooling of Li by Cs can suffer from the large mass ratio causing imperfect overlap of clouds and slow collisional rethermalization. One idea we are considering is to exploit the difference in magnetic moments of Li and Cs and use a magnetic field gradient to differentially tilt the dipole trap in order to find the ideal trajectory for efficient cooling of both species.
Once both species have been cooled to quantum degeneracy, we plan to load Li and Cs into two independently controllable, but overlapping optical lattices. Holding one Li atom per lattice site, we will have control over both the external and internal degrees of freedom of the atoms. These will be used as qubits.
Using an independent lattice for Cs, which is controlled using an electro-optic-modulator array, we can translate the Cs atom to overlap with any given atom in the entire Li lattice. The contact interaction between these two atoms will be used to alter the internal state of the atoms, i.e. to perform gate operations. This messenger qubit (or a quantum read/write head) will be used for performing gates on individual qubits as well as to transport entanglement between distant qubits.
Having a commensurate lattice for Cs means multiple copies of the same operation can be performed simultaneously in different parts of the lattice. Furthermore, since all gate operations are actuated through contact interaction between atoms and no further site specific operations are required, our scheme is easy to scale up to a large number of qubits.
We have constructed two overlapping triangular lattices ( 1064 nm for trapping Cs and 680 nm for trapping Li) [Klinger et al. Rev. Sci. Instrum. 81, 013109 (2010) arXiv:0909.2475] . These are obtained by imaging a diffraction grating onto the position of the atoms- hence creating lattices with the spacing determined only by the grating pattern and the magnification of the imaging setup. We have been able to limit the lattice constant mismatch to less than 2%. The noise in the relative translation of the two lattices is less than 10 nanometers, which is smaller than the atomic wavefunction spread (50 nm) for times on the order of 10 seconds, which is much longer than the time required for gate operations (milliseconds).