In this thesis, I describe a proposal for experiments with dense ensembles of ultracold Rubidium atoms optically excited to highlying electronic states known as Rydberg states. The proposed experiment is designed to implement a model of quantum computation known as deterministic quantum computation with one qubit (DQC1). My proposal is novel among other proposals for Rydberg atoms because it employs mixed states and could provide insights into the source of the enhancements of quantum technology over classical technology. I demonstrate that ultracold atoms and Rydberg interactions offer new perspectives for exploring the DQC1 model. This is mainly due to the large strength, long range and controllability of Rydberg interactions. My specific contributions include: a proposal that solves one of the main issues with implementing quantum logic gates between ensembles of atoms; numerical simulations of the proposed cold atom implementation of a DQC1 protocol; and the design and laboratory implementation of a high-performance optical system for imaging ultracold atoms. These contributions bring the experimental exploration of the DQC1 model closer to realisation.
In this thesis, I describe a proposal for experiments with dense ensembles of ultracold Rubidium atoms optically excited to highlying electronic states known as Rydberg states. The proposed experiment is designed to implement a model of quantum computation known as deterministic quantum computation with one qubit (DQC1). My proposal is novel among other proposals for Rydberg atoms because it employs mixed states and could provide insights into the source of the enhancements of quantum technology over classical technology. I demonstrate that ultracold atoms and Rydberg interactions offer new perspectives for exploring the DQC1 model. This is mainly due to the large strength, long range and controllability of Rydberg interactions. My specific contributions include: a proposal that solves one of the main issues with implementing quantum logic gates between ensembles of atoms; numerical simulations of the proposed cold atom implementation of a DQC1 protocol; and the design and laboratory implementation of a high-performance optical system for imaging ultracold atoms. These contributions bring the experimental exploration of the DQC1 model closer to realisation.
In this work we report on the development progress of a cold atoms platform to demonstrate a model of quantum computation known as deterministic Quantum computing with One Clean Qubit (DQC1). This is a novel class of algorithms where interactions between qubits enable nonclassical correlations other than entanglement, namely discord. We aim to demonstrate the DQC1 protocol using a scheme in which the information processing is realized on cold neutral rubidium atoms through the cNOT gate based on a Rydberg blockade. The implementation of this protocol imposes very stringent requirements on the cloud of atoms: to realise the cNOT gate with high fidelity it is required that a full Rydberg blockade is operating throughout the extent of the ensemble. This means that a cloud of atoms needs to have a radius below approximately 5 μm (for Rydberg states of atoms n = 43). A dipole trap system was designed and implemented during the course of this PhD with the aim of satisfying these experimental constraints. In our setup, a vapour of 87Rb is first trapped in a Magneto-Optical Trap (MOT) and cooled to sub-Doppler temperatures. The atoms are then subsequently trapped in the dipole trap. Our dipole trapping setup is based on a high numerical aperture lens (NA=0.53) to achieve a sub-micron trapping potential and simultaneously observe trapped atoms with a micrometre resolution. According to the design and the optical tests, the system is capable of creating the trapping potential with about 1 μm waist and 4 μm Rayleigh length. The aim of this work was to examine and optimise experimental conditions under which the dipole trap should be operated, so that the sufficiently small trap for the DQC1 algorithm can be obtained in the future. For characterisation, a trap of a bigger volume (about 6 μm waist and 28 μm length) was setup instead of using the full capacity of the system. This enhanced the atom loading procedure and provided more convenient conditions for experiments. The temperature of atoms was typically around 250 μK, and the trap lifetime (2.5 s) was limited by the background collisions, which indicates that the vacuum was maintained on the expected level. The key findings of the thesis are that the properties of the atoms reservoir (MOT temperature and density), and the loading processes (intensity of MOT beams), need to be optimised to be able to load a micrometre size trap. These findings are instrumental for the future implementation of the DQC1 protocol and also have implications for other protocols founded on the use of a Rydberg blockade.
We propose the use of Rydberg interactions and ensembles of cold atoms in mixed state for the implementation of a protocol for deterministic quantum computation with one quantum bit that can be readily operated in high dimensional Hilbert spaces. We propose an experimental test for the scalability of the protocol and to study the physics of discord. Furthermore, we explore the possibility of extending to non-trivial unitaries, such as those associated to many-body physics. Finally develop a scheme to add control to cold atom unitaries in order to facilitate their implementation in our proposal.
In this thesis, the development of an experimental system for microscopic dipole trapping of ultracold neutral rubidium atoms is presented. The purpose of this system is to advance towards the experimental realisation of a quantum computational protocol utilising neutral atoms as qubits. It is intended that the quantum gate operations between qubits will be implemented by a scheme using Rydberg blockade, imposing a restriction on the maximum size of the dipole-trapped atom cloud; the spatial extent of the atomic ensemble contained in this trap must be smaller than the blockade radius to ensure that one single collective Rydberg state per qubit can be achieved. Therefore the experiment was designed with the intent of fulfilling these challenging requirements. This project involved the design and construction of an improved ultra-high vacuum chamber containing the optical setup for the experiment, successfully achieving pressures below 5 x 10-10 mbar. A magneto-optical trap was produced to act as a background reservoir of atoms from which to load the dipole trap. Numerous experimental measurements were done to characterise the physical properties of the trapped atoms, including the number, density and temperature of atoms, as well as the lifetime of the trap. The results of these measurements led to the conclusion that a suitable reservoir for loading the dipole trap had been produced. Significant work was carried out to set up and obtain the dipole trap in the laboratory. Measurements of the characteristic properties of the trap and the atoms confined in the trap were carried out to investigate the behaviour of the atoms and to validate our design. Ultimately a trap containing tens of atoms was achieved, with an atom cloud diameter of ~1.2 µm in two dimensions, being well within the estimated Rydberg blockade radius of ~4.4 µm for n ~ 60 as intended. The two-photon excitation laser system for the probing of Rydberg states, for future applications in Rydberg blockade-based quantum gate operations, was also developed during the course of this work. Different Rydberg states were detected experimentally by the observation of Autler-Townes splitting in a three-level atom scheme. Overall, the work presented in this thesis provides a strong groundwork for the advancement towards neutral atom-based quantum gates, including the development of the experimental system and the production of standard procedures to carry out characterisation measurements of the traps efficiently in the future. The main achievements of this work are the establishment of the experimental apparatus, the achievement of a microscopic dipole trap which conforms to the requirements of an atomic qubit, and the significant growth in the knowledge of atom trapping specific to our system.
I hereby declare that I am the sole author of this thesis. This is a true copy of the thesis, including any required final revisions, as accepted by my examiners. I understand that my thesis may be made electronically available to the public. G. Passante ii Quantum information processors have the ability to drastically change our world. By manipulating bits of information ruled by the laws of quantum mechanics, computational devices can perform some computations that are classically intractable. Most quantum algorithms rely on pure qubits as inputs and require entanglement throughout the compu-tation. In this thesis, we explore a model of computation that uses mixed qubits without en-tanglement known as DQC1 (deterministic quantum computation with one quantum bit), using the physical system of liquid-state Nuclear Magnetic Resonance (NMR). Throughout our research, we experimentally implement an algorithm that completely encapsulates the DQC1 model, and take a close look at the quantum nature of DQC1-states as given by the
High optical nonlinearities can be achieved at the single photon level by coupling the photon states to strongly interacting Rydberg excitations under the conditions of electromagnetically induced transparency. The nonlinear response in Rydberg quantum optics comes as a direct result of the long range dipole-dipole interactions via Rydberg blockade and interaction induced dephasing. In this thesis, Rydberg quantum optics experiments are performed with two spatially separated mediums with a nonlinear response at the single photon level. Two ultracold rubidium atomic clouds are tightly confined by a pair of in-vacuo aspheric lenses such that only a few Rydberg excitations can exist in each one of the clouds simultaneously. The long range character of the dipole-dipole interactions leads to the generation of quantum states of light. The effective photon-photon interactions are directly observed as an anti-correlation in the simultaneous photon retrieval as a result of the non resonant van der Waals interactions. The collective Rydberg excitations, stored in the non-overlapping mediums, experience an additional spatial non-uniform phase. In addition, we experimentally characterize the cooperative optical response of a cold atomic medium at a single photon level. This aims to exploit the enhancement of the spontaneous emissions decay rate (superradiance) as a dependence of the number of atoms involved in the ensemble. The observed atomic response in two different regimes gives a further understanding of the system dynamics and unlocks a promising potential route towards the implementation in scalable, multichannel quantum optical devices.
In any physical platform, two ingredients are essential for quantum information processing: single-qubit control, and entangling interactions between qubits. Neutral atoms can be individually controlled with high fidelity and are resilient to environmental noise, making them attractive candidates for implementing quantum information protocols. However, achieving strong interactions remains a major obstacle. One way to increase the interaction strength between neutral atoms is to excite them into high-lying Rydberg states, which exhibit large electric dipole moments (and by extension, strong electric dipole-dipole interactions). By slowly ramping up the Rydberg level coupling in a system, one can "dress'' the atomic ground states with some Rydberg character; this maps the Rydberg dipole interaction to an effective interaction between ground states. Such Rydberg-dressed interaction is the focus of this dissertation. After describing the physics of the Rydberg-dressed interaction, we propose three protocols that demonstrate its versatility and provide a framework for considering some of the details of realistic implementation. In all three cases, Rydberg dressing --- along with some form of single-atom control --- is used to generate highly entangled states of interest. Our first proposal relates to the adiabatic model of quantum computing, in which solutions to problems are encoded in the ground states of carefully engineered Hamiltonians. The Rydberg-dressed interaction can provide nonlinear Hamiltonian terms, allowing us to encode NP-hard and other interesting problems. We model this protocol in the presence of decoherence, and find that computational fidelities of ~0.98 for four atoms should be possible with currently realistic experimental parameters. Our second proposal is also related to quantum computing, this time in the circuit model. The Rydberg-dressed interaction can be used to generate a controlled-NOT logic gate which, when interwoven with single-qubit gates, can perform universal quantum computation. Experimentally, noise due to atomic thermal motion has been a primary limitation on the fidelities of these gates. We show that a Doppler-free setup, with counterpropagating lasers, effectively suppresses this type of noise, allowing simulated fidelities of up to ~0.998 per gate. Such strong suppression is only possible because the Doppler-free configuration can harness the natural robustness of adiabatic dressing; other gate schemes using, e.g., resonant pulses, do not exhibit the same degree of improvement. Finally, we consider exploiting the many-body character of the Rydberg-dressed interaction to generate collective entanglement in mesoscopic ensembles of neutral atoms. An atomic ensemble uniformly illuminated by a single Rydberg-exciting laser is isomorphic to the well-known Jaynes-Cummings model. In addition to adapting generic Jaynes-Cummings entanglement protocols developed in other platforms, one can apply microwaves to drive entanglement in a way that is unique to the atomic platform. We prove that by allowing the microwave phase to vary in time, one can generate arbitrary symmetric states of the ensemble. While this method compares favorably with other entanglement protocols in many ways, the required frequency of phase switching presents a fundamental limitation on its effectiveness. To mitigate this, we propose a variant scheme in which parameters are chosen to only allow excitations within the system's dressed-ground subspace; this effectively cuts phase switching demands in half. All three protocols serve to illustrate the power of the Rydberg-dressed interaction and suggest directions for future study.