Observation of Entangled States of a Fully Controlled 20-Qubit System

Abstract

We generate and characterize entangled states of a register of 20 individually controlled qubits, where each qubit is encoded into the electronic state of a trapped atomic ion. Entanglement is generated amongst the qubits during the out-of-equilibrium dynamics of an Ising-type Hamiltonian, engineered via laser fields. Since the qubit-qubit interactions decay with distance, entanglement is generated at early times predominantly between neighboring groups of qubits. We characterize entanglement between these groups by designing and applying witnesses for genuine multipartite entanglement. Our results show that, during the dynamical evolution, all neighboring qubit pairs, triplets, most quadruplets, and some quintuplets simultaneously develop genuine multipartite entanglement. Witnessing genuine multipartite entanglement in larger groups of qubits in our system remains an open challenge.

Alternate abstract:

Plain Language Summary

At the heart of useful quantum technologies such as simulators or computers lies the ability to create quantum entanglement, a type of correlation among atomic particles that has no analog in everyday experience. One challenge is to establish and maintain this fragile property for large numbers of particles such as qubits, the quantum equivalent of digital bits. Another challenge is to develop methods to detect entanglement. Here, we report progress on both of these tasks. We present two methods for entanglement detection and use them to detect entanglement in a system of 20 qubits, the largest fully controllable entangled quantum system to date.

We experimentally study the dynamical evolution of the multipartite entanglement structure of 20 qubits encoded in a quantum simulator built from trapped ions. We track the buildup of entanglement between neighboring qubit pairs and triples, and we even verify genuine quantum correlations for up to 5 neighbors. Most importantly, our evaluation only requires a number of measurement settings proportional to the number of qubits—just 27 out of the more than 3 billion possible measurements. This is in contrast to full state reconstruction, where this number grows exponentially with the system size, which is not viable for scalable entanglement detection. Our two methods trade off computational effort and noise robustness. One method is easy to use at any system size, whereas the other performs better at the cost of a more involved numerical optimization.

Our work opens a path to further increase the size and complexity of entangled systems, allowing researchers to go beyond the realm of classical simulation.