Fundamental Research in Quantum Science
The DeMarco group carries out fundamental research on one of the frontiers of 21st century science: interacting many-particle quantum mechanics. Many-particle quantum mechanics is the physics behind the behavior of systems as diverse as metals and neutron stars. We focus on regimes where our conceptual understanding and ability to make predictions are poor, such as the transformation of strongly interacting metals into insulators by disorder. It is precisely this type of fundamental research that leads to new never-before-imagined applications. Hertz's discovery of radio waves is a terrific example of the process by which fundamental research leads to unforeseen transformational outcomes. It is hard to imagine in this wireless age how Hertz could have thought “It's of no use whatsoever. This is just an experiment that proves Maestro Maxwell was right, we just have these mysterious electromagnetic waves that we cannot see with the naked eye. But they are there...” about his discovery.
Hubbard models are our simplest paradigms for electronic solids in which interactions between electrons play an important role. In a Hubbard model, the electrons can tunnel between adjacent latttice sites, and two electrons on the same site interact with each other. Despite this apparent simplicity, we do not understand some of the most basic features of Hubbard models. Outstanding questions include whether or not Hubbard models give rise to d-wave superconductivity and the impact of disorder on properties such as conductivity. The difficulty in understanding Hubbard models arises because the particles are strongly correlated: the interaction and kinetic energy scales are similar, and thus no single-particle theory (including mean-field theories) can be applied. Furthermore, we can only exactly simulate tens of particles on our most powerful supercomputers, and advances in computing consistent with Moore's Law enable us to add just one particle every decade.
Our tool for studying Hubbard models are optical lattices. In our experiments, we cool atoms to temperatures just a billionth of a degree above absolute zero. We trap those ultracold atoms in an optical lattice: a period potential formed from interfering laser beams. The atoms play the role of either electrons or Cooper pairs, and the lattice that of the ionically or covalently bonded crystalline matrix. Atoms in the lattice are exactly described by a Hubbard model. By adjusting the laser intensity we can continously tune the ratio of the Hubbard tunneling to interaction energy and explore all possible parameter regimes.
We have two apparatuses for studying Hubbard models using optical lattices. In one, we trap and cool 87Rb atoms, which are bosons. In the other, we use 40K atoms, which are fermions. Every experimental cycle takes about 90 seconds. In one run of the experiment, we cool atoms starting from a room-temperature vapor, transfer them to a region evaucated to about 10-12 torr, and then trap them in an optical lattice. We manipulate the atoms in some way (e.g., push on them using a magnetic field gradient), release the gas from the lattice, and then image using resonant light and a CCD camera. In contrast with experiments on solids that may involve years of material development, we can change the analogue of the material parameters during every 90-second run.
Our group focuses primarily on understanding the impact of disorder on Hubbard models and the dynamics and out-of-equilibrium properties of strongly interacting quantum systems.
Disorder is ubiquitous in solids and can play a central role in determining their properties. While non-interacting disordered quantum systems are well understood, the combination of inter-particle interactions and diosrder has foiled researchers for decades. We add disorder to our naturally clean lattices using optical speckle, which is produced by focusing 532 nm laser light that has passed through a holographic diffuser. The atoms experience a potential proportional to the speckle intensity, which varies randomly in space.
The way we produce disorder is a powerful tool, because we can control the disorder strength by changing the speckle laser intensity, from completely removing it to making it the largest energy scale present. The disorder can also be completely characterized using optical microscopy, and hence the disordered Hubbard parameters are precisely known.
Recent Results and Ongoing Work
Recent results include the first evidence for many-body localization, which was obtained by measuring the response of a Fermi gas trapped in the metallic regime of a lattice to an applied force. We were able to observe that sufficient disorder localized the atoms, resulting in an insulating state. Futhermore, that insulating state persisted over a range of thermal energies.
You can read a preprint about this work here, and view a talk by clicking the image below.
Bose-glass in disorder quenches
We have also recently investigated quantum quenches of disorder for bosonic atoms trapped in a lattice. Vortices and other excitations were observed when the disorder was above a critical value. By comparing to state-of-the-art quantum Monte Carlo simulations from David Ceperley's group, we showed that the excitations resulted from the presence of a Bose-glass phase.
View a poster about this result by clicking the image below.
Even though the equilibrium properties of strongly interacting quantum particles can be understood in limited scenarios, many open questions persist about the dynamics of those systems. Dynamical properties of quantum systems are critical to applications, such as quantum information processing and energy transmission. Ultracold atom experiments are ideal for exploring these phenomena, because all timescales are experimentally resolvable and the system can be removed far from equilibrium.
Recent Results and Ongoing Work
Metastable BECs in optical lattices
The processes by which closed quantum systems equilibriate are poorly understood. We probed equilibration of a bosonic gas in a clean optical lattice by measuring peak fraction. We found that out-of-equilibrium gases can be produced for even slow turn-on times of the lattice through comparison to state-of-the-art quantum Monte Carlo simulations from David Ceperley's group. At high lattice potential depths, we discovered that higher peak and condensate fraction was measured than allowed by the second law of thermodynamics. In collaboration with Stefan Natu, we showed that a suppression of Landau relaxation may give rise to this surprising behavior.
Read a preprint about this work here.
Thermalization in optical lattices
Very little is known about thermalization in Hubbard models and optical lattices. We are investigating thermalization and momentum relaxation for bosonic and fermionic gases in optical lattices. By making rapid changes in the momentum distribution of a thermal Bose gas, we measure the relaxation time to equilibrium. Which find that relaxation occurs extraordinarily rapidly, which has enabled us to realize a new scheme for cooling quasimomentum distributions. For fermions, we use stimulated Raman transitions to excite motion of one spin component. We measure the time for that motion to damp and its dependence on interaction strength and temperature.
View a talk about the cooling measurements by clicking the image below.