Vincenzo Lordi (15-ERD-051)
Proposed quantum computers will use the ability of quantum bits (qubits), such as a collection of atoms, to be in a large number of states simultaneously. Whereas conventional computer bits come either in ones or twos, a qubit can be the equivalent of both a one and a two at the same time. In theory, this capability will allow the quantum computer to perform many different computations simultaneously. Superconducting qubits and ion-trap qubits are two leading candidates for scalable quantum computer architectures because they both leverage well-developed microfabrication technologies. However, there are microscopic sources of noise associated with the surfaces of some materials used to fabricate specific superconducting qubits (see figure). Current research elsewhere does not treat noise in these systems realistically, nor integrate the noise models with device-level models. We intend to build an integrated modeling capability for the realistic treatment of noise (as a function of time and space) in engineered qubit systems by coupling atomic-scale models of materials-related noise sources to device-level circuit simulations. This approach treats noise realistically with a comprehensive physics treatment of its sources and long-range interactions, which will be far beyond the current idealized state of the art. The results will enable optimization of fabrication processes and device layouts so that practical qubits may be designed for a quantum computer. The main tools used will include density functional theory and full-wave time-domain electromagnetic device simulations. Validation of noise models on experimental quantum simulators, such as the D-Wave quantum processor, will be pursued as well.
We expect to produce novel physics-based, materials-related noise models for a variety of qubit implementations, including superconducting qubits with advanced materials and ion-trap devices. The noise models will be integrated directly into geometrically defined, device-level circuit quantum electrodynamics models that will result in simulations of noise correlation and cross talk among realistic qubits, enabling optimal engineered design of such devices. Ultimately, the outputs of this project could feed into system-level models of error propagation within a given quantum error-correction coding scheme to determine error thresholds for computing. As such, this could enable computer-aided design of quantum computers in the future.
This research project will help advance the state of the art toward realizing a quantum computer, which has significant relevance for the cyber security, space, and intelligence strategic focus area at LLNL because of its powerful capabilities for national security applications such as encrypting or decrypting information or its near-invincibility to cyber attacks. We will utilize and enhance Livermore core competencies in advanced materials and manufacturing as well as high-performance computing, simulation, and data science for which quantum computers may offer a significant future computing platform.
FY16 Accomplishments and Results
In FY16 we (1) generated several finite-element electromagnetic models of both superconducting and ion-trap qubit devices; (2) used the superconducting device models, which included simple resonator and transmission line models, to demonstrate the classical and quantum model couplings, in addition to full qubit circuits; (3) used the ion-trap device models, which included both a rod-type Paul trap (a quadrupole ion trap that uses dynamic electric fields to trap charged particles) and surface electrode devices, to explore trapping stability; (4) computed spin-exchange interactions among oxygen molecules adsorbed on differently terminated alumina surfaces using atomistic simulations, which formed the basis for estimating a source magnetic noise; and (5) performed measurements of the magnetic susceptibility of sapphire plates under different conditions to confirm experimentally the role of surface impurities as magnetic noise sources.
Publications and Presentations
- Lordi, V., et al., Materials models for quantum computing: Physics-based noise modeling from materials to devices. IBM ThinkQ 2015, Yorktown Heights, NY, Dec. 2–4, 2015. LLNL-POST-657112.
- Materise, N., et al., Simulating noisy superconducting qubits. IBM ThinkQ 2015, Yorktown Heights, NY, Dec. 2–4, 2015. LLNL-POST-679779.
- Ray, K. G., J. Dubois, and V. Lordi, Paramagnetic spins on Al2O3 with varied surface termination. American Physical Society March Mtg., Baltimore, MD, Mar. 14–18, 2016. LLNL-PRES-686344.