Diffusion plays a major role in describing how particles and heat move in any nonequilibrium system. The diffusion of quantum properties is not as well understood as the diffusion of classical properties, especially at low temperatures or high densities. Exploring the diffusion of the quantum property known as spin is beneficial for explaining quantum effects that arise at low temperatures or high densities, and this knowledge acquired could assist in the development of new ultra-low-power devices. This thesis examines spin diffusion at various temperatures and densities using the highly-tunable experimental platform of ultracold atoms. Around one million rubidium-87 atoms are cooled to nanokelvin temperatures to create an ultracold gas, where quantum interactions between atoms can significantly modify spin diffusion compared to classical diffusion. One-dimensional spin diffusion is observed for an initial two-domain spin profile. Remarkably, diffusion of this spin profile is slowed at temperatures above quantum degeneracy, where the ultracold gas is largely classical but with quantum collisions. We demonstrate that this slowing of spin diffusion is due to the presence of spin coherence between spin domains, and that the removal of coherence speeds up spin diffusion to classical timescales. Spin diffusion is further modified by applying a linear differential potential that can speed, slow, or stop spin diffusion of a two-domain spin profile. Differential potentials spatially alter the precession of spins, which then alters the spin-rotating quantum collisions that modify spin diffusion. For a linear differential potential with a specific sign and magnitude, stabilized spin domains in an ultracold gas are observed for 40 times longer than classical diffusion timescales. In addition to modifying spin diffusion with coherence and differential potentials, we demonstrate arbitrary control of one-dimensional spin diffusion using temporally varying differential potentials and three-domain spin profiles. These spin diffusion modifying techniques could be useful for manipulating spin in other nonequilibrium systems, and set the stage for simulating spin-based devices in ultracold atoms.
Diffusion plays a major role in describing how particles and heat move in any nonequilibrium system. The diffusion of quantum properties is not as well understood as the diffusion of classical properties, especially at low temperatures or high densities. Exploring the diffusion of the quantum property known as spin is beneficial for explaining quantum effects that arise at low temperatures or high densities, and this knowledge acquired could assist in the development of new ultra-low-power devices. This thesis examines spin diffusion at various temperatures and densities using the highly-tunable experimental platform of ultracold atoms. Around one million rubidium-87 atoms are cooled to nanokelvin temperatures to create an ultracold gas, where quantum interactions between atoms can significantly modify spin diffusion compared to classical diffusion. One-dimensional spin diffusion is observed for an initial two-domain spin profile. Remarkably, diffusion of this spin profile is slowed at temperatures above quantum degeneracy, where the ultracold gas is largely classical but with quantum collisions. We demonstrate that this slowing of spin diffusion is due to the presence of spin coherence between spin domains, and that the removal of coherence speeds up spin diffusion to classical timescales. Spin diffusion is further modified by applying a linear differential potential that can speed, slow, or stop spin diffusion of a two-domain spin profile. Differential potentials spatially alter the precession of spins, which then alters the spin-rotating quantum collisions that modify spin diffusion. For a linear differential potential with a specific sign and magnitude, stabilized spin domains in an ultracold gas are observed for 40 times longer than classical diffusion timescales. In addition to modifying spin diffusion with coherence and differential potentials, we demonstrate arbitrary control of one-dimensional spin diffusion using temporally varying differential potentials and three-domain spin profiles. These spin diffusion modifying techniques could be useful for manipulating spin in other nonequilibrium systems, and set the stage for simulating spin-based devices in ultracold atoms.
Ultra-cold trapped atoms, with their high degree of tunability, provide ideal model systems to study physical phenomena with applications in many different fields of research. This thesis describes studies on spin transport phenomenon in a trapped ultra-cold spin-polarized 87Rb gas at temperatures above quantum degeneracy. This work is focused on the less studied regime of cross-over between classical and quantum transport. Diffusion is a fundamental dissipative process that tends to relax any system towards a state of minimum inhomogeneity. In this work we study longitudinal spin diffusion as a special case of spin transport. The system studied here consists of two anti-parallel longitudinal spin domains separated by a helical domain-wall. We report that the diffusion process manifests a significant deviation from classical diffusion due to purely quantum mechanical modifications. The two-domain spin textures are prepared using optical and microwave pulse techniques, and the dynamics of the spin structure is studied as it relaxes towards the final equilibrium state. Generally, there is a wide range of parameter space that could be studied. In this work we focused our studies on the effects related to the degree of coherence in the domainwall as well as the effective magnetic field acting on the spins. By controllably tuning these experimental parameters, we studied in detail how the spatiotemporal behaviour of the diffusion dynamics is modified. Our results show that the longitudinal spin diffusion time scales depend sensitively on the domain-wall degree of coherence. External magnetic field gradients also alter the dynamics noticeably, manifesting a significant dependence on the sign of the applied field gradients.
The three experiments described in this thesis investigate fundamental properties of ultracold atoms. Using laser cooling and evaporative cooling, a dilute gas of sodium atoms is cooled to ~100 nK. Under these circumstances a Bose-Einstein condensate (BEC) forms, where millions of atoms collapse into the lowest energy state of the system and share a macroscopic wavefunction. The experiments are done in an ultrahigh vacuum and the atoms are manipulated remotely using laser beams, magnetic fields, and RF-fields. Using ultracold atoms allows unprecedented control over the internal and external degrees of freedom of the system, making it very suitable for precision measurements. Ultracold atoms also serve as a model system for more complicated or remote fields such as solid-state physics and cosmology. In the first experiment the refractive index of an ultracold bosonic gas is studied. The atomic clouds are analyzed for temperature and number of particles using nearly non-destructive phase-contrast imaging. After each pulse of probe light a small fraction of the atoms is lost, while the cloud is simultaneously slightly heated. This makes it possible to study the scattering rate as a function of temperature using only a single sample. It is observed that the scattering rate increases below the critical temperature for BEC by more than a factor of 3 compared to the classical value. This enhancement is in good agreement with the predictions by theoretical work that takes two-body correlation and the Lorentz-Lorenz correction into account. The method is also used to perform calorimetry on the Bose gas and directly measure the heat capacity of the BEC. The second chapter deals with magnetism in ultracold gases. When cooling an ultracold gas containing different spin components below the critical temperature for BEC, stable magnetic domains form. The appearance of these domains and the dynamics of the domain walls are studied. The spin mixture is prepared by performing an RF-sweep on ground state atoms in the presence of a magnetic field. Spin dynamics are induced by applying magnetic gradients. Spin drag, which is the transport coefficient responsible for damping spin currents, was measured in a two-component system. Furthermore a new method for spin resolved phase contrast imaging is presented, which is used to study time evolution in situ non-destructively. The last experiment focuses on spontaneous symmetry breaking. When passing a continuous phase transition, spontaneous symmetry breaking causes the development of domains with an independently chosen order parameter. The domains are separated by boundary defects. In this experiment the spontaneous creation of solitons was observed, which are boundary defects in the phase of the condensate, as well as magnetic domain walls, which are the boundary defects in the magnetization. Through the Kibble-Zurek mechanism it is possible to determine the critical exponents that characterize the phase transition. The critical exponents are predicted to be universal, i.e. do not depend on the microphysics of the system, so the results from our ultracold atoms lab can help answer universal questions for example from cosmology and from the early universe.
Domain walls are carriers of spin transport whose controlled manipulation underlies a wide range of spintronic and information-processing technologies. Here we demonstrate tunable domain-wall transport in a weakly interacting nondegenerate ultracold Bose gas. We initialize a three-domain pseudo-spin-1/2 texture and observe spontaneous propagation of long-lived domain walls driven by exchange-mediated spin currents. By varying the orientation of the spin domains, we control the balance of spin currents across the walls and thereby tune their trajectories, including reversals of the initial direction of motion. Measurements reveal a crossover from an exchange-stabilized regime, in which coherent spin-exchange collisions suppress wall motion, to a diffusion-dominated regime characterized by rapid transport at thermal velocities. Numerical solutions of a quantum Boltzmann equation reproduce the observed dynamics and identify transverse phase gradients as an important control parameter governing domain-wall propagation. These results establish coherence and phase engineering as tools for programming spin transport in ultracold gases and provide a route toward controllable domain-wall dynamics in atomtronic systems.
In this thesis, we develop, validate and apply a non-destructive, spin-sensitive probing method for ultracold gases of sodium. By exploiting the polarization dependence of the atom-light interaction, the method combines optical hardware with computational reconstruction to access spin-resolved density and magnetization distributions from a single, minimally destructive exposure. The thesis is structured in three parts; firstly, we describe the experimental platform for making, probing and manipulating sodium spinor Bose–Einstein condensates, including laser cooling, optical and magnetic trapping, and radio-frequency techniques for coherent spin preparation. In the second part, we develop the core measurement technique of spin-dependent off-axis holography is developed, and evaluate its robustness to technical noise and the magnetic field variations. The reconstruction and analysis pipeline is presented in detail, with special emphasis on the background noise suppression using an inpainting technique based on principal component analysis. In the final part, the method is applied to investigate dynamics in spin-1 ultracold atomic sodium. To that end, we introduce the theory of three-level spin systems and propose magnetization-preparation and precise Zeeman-sublevel readout schemes, which are experimentally verified. These tools are then used to investigate spin diffusion and spin mixing dynamics, and the formation and evolution of spin domains. This work establishes spin-dependent off-axis holography as a powerful tool for probing spinor quantum gases. The techniques developed here enable efficient and minimally destructive imaging of spin dynamics, and provide new opportunities for studying non-equilibrium spin phenomena and magnetism in ultracold atomic systems.
We study the stability of two-domain spin structures in an ultracold gas of magnetically trapped 87Rb atoms above quantum degeneracy. Adding a small effective magnetic field gradient stabilizes the domains via coherent collective spin rotation effects, despite negligibly perturbing the potential energy relative to the thermal energy. We demonstrate that domain stabilization is accomplished through decoupling the dynamics of longitudinal magnetization, which remains in time-independent domains, from transverse magnetization, which undergoes a purely transverse spin wave trapped within the domain wall. We explore the effect of temperature and density on the steady-state domains, and compare our results to a hydrodynamic solution to a quantum Boltzmann equation.
Ultracold atoms in optical lattices are among the most developed platforms of interest for building quantum devices suitable for quantum simulation and quantum computation. Ultracold trapped atoms are advantageous because they are fundamentally indistinguishable qubits that can be prepared with high fidelity in well-defined states and read-out with similarly high fidelities. However, an outstanding challenge for ultracold atoms in optical lattices is to engineer interesting interactions and control the effects of heating that couple the system to states that lie outside the Hilbert space we wish to engineer. In this thesis, I describe a series of experiments and theoretical proposals that address several critical issues facing ultracold atoms in optical lattices. First, I describe experiments where the tunneling behavior of atoms in the lattice is modified to make our fundamentally neutral particles behave as though they are charged particles in a magnetic field. We show how engineering this interaction creates intrinsic degeneracy in the single particle spectrum of the many-body system and how to introduce strong interactions in the system with the goal of producing exotic many-body states such as a bosonic fractional quantum Hall states. Then, I discuss how this technique can be easily generalized to include spin and higher spatial dimensions in order to access a rich variety of new physics phenomena. Next, I report on the realization of a spin-1 Heisenberg Hamiltonian which emerges as the low energy effective theory describing spin ordering in the doubly-occupied Mott insulator of two spin components. This integer spin Heisenberg model is qualitatively different from the half-integer spin model because it contains a gapped, spin-insulating ground state for small inter-spin interaction energies which we call the spin Mott. Using a spin-dependent lattice to control the inter-spin interactions, we demonstrate high-fidelity, reversible loading of the spin-Mott phase and develop a probe of local spin correlations in order to demonstrate a spin entropy below 0.2 kB per spin. Progress on adiabatically driving the quantum phase transition from the spin Mott to the xy-ferromagnetic is discussed along with the progress towards the creation of a quantum gas microscope for single atom detection and manipulation..