We study quantum control of the full hyperfine manifold in the ground-electronic state of alkali-metal atoms based on applied radio frequency and microwave fields. Such interactions should allow essentially decoherence-free dynamics and the application of techniques for robust control developed for NMR spectroscopy. We establish the conditions under which the system is controllable in the sense that one can generate an arbitrary unitary map on the system. We apply this to the case of
133Cs with its
d=16 dimensional Hilbert space of magnetic sublevels in the
6S1∕2 state, and design control wave forms that generate an arbitrary target state from an initial fiducial state. We develop a generalized Wigner function representation for this space consisting of the direct sum of two irreducible representations of SU(2), allowing us to visualize these states. The performance of different control scenarios is evaluated based on the ability to generate a high-fidelity operation in an allotted time with the available resources. We find good operating points commensurate with modest laboratory requirements.
This dissertation studies spin squeezing, entanglement and decoherence in large ensembles of cold, trapped alkali atoms with hyperfine spin f interacting with optical fields. Restricting the state of each atom to a qutrit embedded in the 2f+1 dimensional hyperfine spin enables us to efficiently model the coherent and dissipative dynamics of the ensemble. This formalism also allows us to explore the effects of local control on the internal hyperfine spins of the atoms. State preparation using such control increases the entangling power of the atom-light interface for f>1/2. Subsequent control of the internal spins converts entanglement into metrologically relevant spin squeezing. In the case of squeezing by quantum nondemolition measurement, we employ a numerical search to find state preparations that maximize spin squeezing in the presence of decoherence. Dissipative dynamics on our system include optical pumping due to spontaneous emission. While most works ignore optical pumping or treat it phenomenologically, we employ a master equation derived from first principles. This work is extended to the case of an atomic ensemble interacting with a non-homogeneous paraxial probe. The geometries of the ensemble and the probe are optimized to maximize both spatial mode matching and spin squeezing.
Ultracold neutral atoms in optical lattices are rich systems for the investigation of many-body physics as well as for the implementation of quantum information processing. While traditionally alkali atoms were used for this research, in recent years alkaline-earth-like atoms have attracted considerable interest. This is due to their more complex but tractable internal structure and easily accessible transitions. Furthermore, alkaline-earth-like atoms have extremely narrow
1S→3P intercombination transitions, which lend themselves for the implementation of next generation atomic clocks. In this dissertation, I show that exquisite control of alkaline-earth-like atoms can be reached with optical methods, and elucidate ways to use this controllability to further different aspects of research, mainly quantum information processing. Additionally, the control of alkaline-earth-like atoms is very interesting in many-body physics and the improvement of atomic clocks. Since heating usually degrades the performance of quantum gates, recooling of qubits is a necessity for the implementation of large scale quantum computers. Laser cooling has advantages over the usually used sympathetic cooling, given that it requires no additional atoms, which have to be controlled separately. However, for qubits stored in hyperfine states, as usually done in alkali atoms, laser cooling leads to optical pumping and therefore to loss of coherence. On the other hand, in the ground state, the nuclear spin of alkaline-earth-like atoms is decoupled from the electronic degrees of freedom. As I show in this dissertation, this allows for the storage of quantum information in the nuclear spin and laser cooling on the electronic degrees of freedom. The recooling protocol suggested here consists of two steps: resolved sideband cooling on the extremely narrow
1S0→3P0 clock transition and subsequent quenching on the much shorter lived
1P1 state. A magnetic field is used to overcome the hyperfine interaction in this excited state and thus ensures decoupling of the nuclear spin degrees of freedom during the quenching. The application of this magnetic field also allows for photon scattering on the
1P1 state, while preserving the nuclear spin, e. g. for electronic qubit detection. The manipulation of the scattering properties of neutral atoms is an important aspect of quantum control. In contrast to alkali atoms, whose broad linewidths cause large losses, this can be done with purely optical methods via the implementation of an optical Feshbach resonance for alkaline-earth-like atoms. Here, the scattering length resulting from the application of an optical Feshbach resonance on the
1S0→3P1 intercombination line, including hyperfine interaction and rotation is calculated for
171Yb. Due to their different parities, the p-wave scattering length can be controlled independently from the s-wave scattering length, thus allowing for unprecedented control over the scattering properties of neutral atoms. Furthermore, I also show how optical Feshbach resonances in alkaline-earth-like atoms can be used together with the underlying quantum symmetry to implement collisional gates between nuclear-spin qubits over comparatively long ranges.
Known as decoherence, the unavoidable interaction of a quantum system with its surrounding environment is usually considered to be detrimental for quantum information processing. In this thesis the coherent, open loop control of such open systems is studied. Concepts from quantum control theory and the theory of open quantum system are adopted in order to fight decoherence and implement quantum gates in a noiseless manner. In particular, Lie algebraic methods and numerical optimization tools are used to investigate the control properties of a single spin interacting with a spin environment. We show that, independent of the size of the environment, every unitary transformation can be implemented on the system spin through a single control field. We proceed by investigating dynamical decoupling, a method to suppress the interactions with the environment, for finite- and for in- finite dimensional systems. We prove that every finite dimensional system can be protected from decoherence, even if the environment is infinite dimensional, whereas for noise described by a Lindblad master equation dynamical decoupling will never succeed. This will lead to a new method to distinguish decoherence from intrinsic noise terms. We further prove that not every infinite dimensional system can be protected from decoherence through dynamical decoupling. Afterwards we investigate dynamical decoupling of systems that are described by quadratic Hamiltonians, showing that such interactions can always be suppressed with two simple operations. In the last part we investigate the coherent control of a Lindblad master equation. We show that a strong noise process exhibiting a decoherence free subspace can substantially increase the number of unitary operations that can be implemented, allowing us to fully control parts of the system. Afterwards we develop a scheme to make Hamiltonians and Lindbladians commutative by adding an auxiliary system. The old, possibly non-commutative dynamics, is recovered through a non-selective measurement.
This dissertation is a study of several aspects of discrete time quantum control. After reviewing the necessary background from quantum mechanics and control theory we construct a simple model and solve a control problem on it as a proof of concept for the discrete control formulation developed herein. We apply optimal control to solve a state preparation problem on a physically motivated time discretized single qubit system with simplified dynamics to enable the Bellman equation to be solved both numerically and approximately analytically. However for more than one qubit classical theories of control become problematic, a notable difficulty being the exponential increase in the size of the state space with the number of particles. To study workarounds for this problem we quantify how much information a quantum measurement can extract from the system and then use the results to show that a naive classical technique for controlling a multiparticle system performs poorly when the system has "quantum" properties. We then consider the problem of protecting the entanglement of a bipartite quantum system using discrete open loop control with a technique that seems to have no classical analog. The states of the bipartite system, called X-states, have a special structure which is generalized to states of systems with more than two qubits. We develop the properties of these generalized X-states which are a class of mixed states with interesting multipartite entanglement properties
Laboratory techniques to manipulate and observe ultracold atoms make these an attractive platform for testing new ideas in quantum control and measurement. Over the last decade we have revisited the tensor interaction between light fields and multilevel atoms, and have developed a theoretical framework suited for applications in quantum control and measurement (see [1] for a review). One important finding is that the combined action of a light shift and magnetic field on an atomic ground state can be used to implement a non-linear Hamiltonian for a hyperfine spin, and that its action can lead to nonlinear spin dynamics such as wavepacket collapse and revival. Using concepts from classical control theory it is straightforward to show that this Hamiltonian is sufficiently general for full control of an arbitrarily large spin. On this foundation we have developed a new protocol for quantum state reconstruction, based on continuous weak measurement of a spin observable during carefully designed coherent evolution [2]. We have further used our tools for control and state reconstruction to implement and verify protocols for optimal control of Cs hyperfine spns (Fig. 1), showing that an initial fiducial state can be transformed into any desired target state with a fidelity in the 80-90% range [3].