Source-linked AI summary
Demonstration of the trapped-ion quantum-CCD computer architecture
J. M. Pino, J. M. Dreiling, C. Figgatt, J. P. Gaebler, S. A. Moses, M. S. Allman, C. H. Baldwin, M. Foss-Feig, D. Hayes, K. Mayer, C. Ryan-Anderson, B. Neyenhuis
TL;DR
Small ion-crystal computers face scaling limits, while combining the QCCD architecture’s requirements in one machine remains challenging. The authors integrate those capabilities in a programmable processor using up to six qubits, achieving the maximal quantum volume allowed by six qubits and demonstrating a teleported CNOT. The results show that QCCD capabilities can operate together in this processor, though scaling to thousands or millions of qubits remains unclear.
Problem
A single ion crystal is unlikely to scale beyond one hundred qubits, while combining the transport, control, cooling, and parallel-operation requirements of QCCD in one machine is challenging.
Source: p. 1Method
The authors demonstrate a QCCD processor that integrates the architecture’s required capabilities and operates with up to six qubits.
Source: p. 1Results
The processor achieves the maximal quantum volume allowed by six qubits and demonstrates a teleported CNOT with Favg ≥0.899(6).
Source: pp. 1, 4Takeaways & Limitations
The results show that low-error operations can be integrated with QCCD capabilities in a processor operating with up to six qubits.
Source: p. 1Scaling to thousands or millions of qubits remains unclear; linear end-to-end scaling would increase transport time and constrain connectivity under finite coherence without fault-tolerant methods.
Source: p. 5
Abstract
from arXiv · showhide
The trapped-ion QCCD (quantum charge-coupled device) architecture proposal lays out a blueprint for a universal quantum computer. The design begins with electrodes patterned on a two-dimensional surface configured to trap multiple arrays of ions (or ion crystals). Communication within the ion crystal network allows for the machine to be scaled while keeping the number of ions in each crystal to a small number, thereby preserving the low error rates demonstrated in trapped-ion experiments. By proposing to communicate quantum information by moving the ions through space to interact with other distant ions, the architecture creates a quantum computer endowed with full-connectivity. However, engineering this fully-connected computer introduces a host of difficulties that have precluded the architecture from being fully realized in the twenty years since its proposal. Using a Honeywell cryogenic surface trap, we report on the integration of all necessary ingredients of the QCCD architecture into a programmable trapped-ion quantum computer. Using four and six qubit circuits, the system level performance of the processor is quantified by the fidelity of a teleported CNOT gate utilizing mid-circuit measurement and a quantum volume measurement of $2^6=64$. By demonstrating that the low error rates achievable in small ion crystals can be successfully integrated with a scalable trap design, parallel optical delivery, and fast ion transport, the QCCD architecture is shown to be a viable path toward large quantum computers. Atomic ions provide perfectly identical, high-fidelity qubits. Our work shows that the QCCD architecture built around these qubits will provide high performance quantum computers, likely enabling important near-term demonstrations such as quantum error correction and quantum advantage.
Methods
The Methods section describes the processor’s transport and gate operations and the procedures used to assess performance, including randomized benchmarking, a teleported CNOT, and quantum-volume tests.
Methods
- Transport waveforms are built from simulated electrode fields, calibrated shim voltages, and interpolated voltage sequences that preserve valid trapping potentials.The transport library specifies linear shifts, swaps, and split-combine operations through constraints on potential minima, axes, confinement, and potential terms.Source: Methods · pp. 7–8
- The processor characterizes single- and two-qubit gates with randomized benchmarking, including simultaneous TQ benchmarking to test for crosstalk between gate zones.The simultaneous method compares separate-zone and joint-operation decay rates and assesses correlated errors.Source: Methods · p. 9
- Measured addressability errors are below 10^-3 and within one standard deviation of zero, while correlation measurements are also consistent with zero.The correlation results have a minimum uncertainty of 4.1×10^-4.Source: Methods · p. 9
- Favg ≥0.899(6) bounds the teleported CNOT fidelity, using three entangling operations, two mid-circuit measurements, and two measurement-conditioned single-qubit operations.The four-qubit circuit implements a CNOT between qubits that never directly interact.Source: Methods · p. 10
- The quantum-volume tests used N=2,3,4 with four qubit ions and N=6 with six; optimization reduced the N=6 circuits’ average TQ gates from 54 to 45.The N=2,3,4 circuits were run without optimization or approximations.Source: Methods · p. 10