Search for: All records

The ability to make high-fidelity qubit measurements with minimal collateral disruption to the system is not only relevant to initialization and final read-out -- it is also essential to achieving quantum error correction on a universal quantum computation. Qubit state measurements in a neutral atom array are achieved by probing the array with light detuned from a cycling transition and capturing resulting fluorescence with a high quantum efficiency imaging device, producing a greyscale image of the neutral atom array. Conventionally, to achieve a fidelity above 99%, the typical probing period is several ms. This is a significant delay, given that the longest gate operation only takes several micros. In this poster, we demonstrate qubit state measurements assisted by a supervised convolutional neural network (CNN) in a neutral atom quantum processor. We present two CNN architectures for analyzing neutral atom qubit readout data: a compact 5-layer single-qubit CNN architecture and a 6-layer multi-qubit CNN architecture. We benchmark both architectures against a conventional Gaussian threshold analysis method. We demonstrate up to 56% reduction of measurement infidelity using the CNN compared to a conventional analysis method. This work presents a proof of concept for a CNN network to be implemented as a real-time readout processing method on a neutral atom quantum computer, enabling faster readout time and improved fidelity. 

Abstract Dielectric mirrors comprising thin‐film multilayers are widely used in optical experiments because they can achieve substantially higher reflectance compared to metal mirrors. Here, potential problems are investigated that can arise when dielectric mirrors are used at oblique incidence, in particular for focused beams. It is found that light beams reflected from dielectric mirrors can experience lateral beam shifts, beam‐shape distortion, and depolarization, and these effects have a strong dependence on wavelength, incident angle, and incident polarization. Because vendors of dielectric mirrors typically do not share the particular layer structure of their products, several dielectric‐mirror stacks are designed and simulated, and then the lateral beam shift from two commercial dielectric mirrors and one coated metal mirror is also measured. This paper brings awareness of the tradeoffs between dielectric mirrors and front‐surface metal mirrors in certain optics experiments, and it is suggested that vendors of dielectric mirrors provide information about beam shifts, distortion, and depolarization when their products are used at oblique incidence. 

Silica-encapsulated gold core@shell nanoparticles (Au@SiO 2 CSNPs) were synthesized via a tunable bottom-up procedure to catalyze the aerobic oxidation of benzyl alcohol. The nanoparticles exhibit a mesoporous shell which enhances selectivity by inhibiting the formation of larger species. Adding potassium carbonate to the reaction increased conversion from 17.3 to 60.4% while decreasing selectivity from 98.4 to 75.0%. A gold nanoparticle control catalyst with a similar gold surface area took 6 times as long to reach the same conversion, achieving only 49.4% selectivity. These results suggest that the pore size distribution within the inert silica shell of Au@SiO 2 CSNPs inhibits the formation of undesired products to facilitate the selective oxidation of benzaldehyde despite a basic environment. A smaller activation energy, mass transport analysis, and mesopore distribution together suggest the Au@SiO 2 CSNP catalyst demonstrates higher activity through beneficial in-pore orientation, promoting a lower activation energy mechanistic pathway. Taken together, this is a promising catalytic structure to optimize oxidation chemistries, without leveraging surface-interacting factors like chelating agents or active support surfaces. 

We present recent progress towards building a neutral atom quantum computer. We use a new design for a blue-detuned optical lattice to trap single Cs atoms. The lattice is created using a combination of diffractive elements and acousto-optic deflectors (AODs) which give a reconfigurable set of cross-hatched lines. By using AODs, we can vary the number of traps and size of the trapping regions as well as eliminate extraneous traps in Talbot planes. Since this trap uses blue-detuned light, it traps both ground state atoms and atoms excited to the Rydberg state; moreover, by tuning the size of the trapping region, we can make the traps “magic” for a selected Rydberg state. We use an optical tweezer beam for atom rearrangement. When loading atoms into the array, trap sites randomly contain zero or one atoms. Atoms are then moved between different trapping sites using a red-detuned optical tweezer. Optimal atom rearrangement is calculated using the “Hungarian Method”. These rearrangement techniques can be used to create defect-free sub-lattices. Lattice atoms can also be used as a reservoir for a set of selected sites. This allows quick replacement of atoms, and increased data rate, without reloading from a MOT. 

We present progress in demonstrating Rydberg interactions between a single Rb and a single Cs atom simultaneously trapped in a single 976 nm optical tweezer. Rydberg lev- els in heteronuclear systems have different quantum defects, as opposed to homonuclear systems, and can therefore be chosen to minimize the Forster defect and increase the Rydberg interaction strength beyond symmetric Rydberg pairs at comparable energy levels. Additionally, multi-species systems are distinguishable and can be frequency multi- plexed in a straightforward manner. Frequency multiplexing both the state preparation and state readout is used in characterizing elastic and inelastic collision rates between Rb and Cs, as well as enabling crosstalk free ancilla measurements for quantum error correction. 

We demonstrate a 2D lattice of blue-detuned optical traps which uses laser power efficiently, is tolerant to perturbations in beam alignment, and is insensitive to interferometric phases. Blue traps have several advantages over red traps despite requir- ing a more complicated beam geometry. Since atoms in a blue trap sit at an intensity minimum, Stark shift noise and site-to-site calibrations are minimized. However, constructing a blue lattice which efficiently con- verts laser power into trap depth, is challenging. For example, a lattice of bottle beams is inefficient because neighboring sites are separated by two walls, limiting the number of traps that can be formed. An array of tightly spaced Gaussian beams is a more efficient blue trap, but the trap potentials are susceptible to alignment perturbations. We demonstrate an array which uses diffractive optical elements to create a cross-hatched pattern of lines in the focal region where the atoms are trapped in up to 121 sites. This "line array" is almost twice as efficient as the Gaussian beam array and is more resilient to perturbations in beam alignment.