Historical research guide · 0450 · 2023
Demonstration of Quantum Energy Teleportation on Superconducting Quantum Hardware
Kazuki Ikeda
In one minute
Kazuki Ikeda, at Stony Brook University, ran a protocol on IBM’s public quantum computers that moves energy — not a description of energy, but a usable amount of it — using nothing but a phone call. Quantum energy teleportation, proposed by Masahiro Hotta about fifteen years earlier, works like this. Two qubits sit in the entangled ground state of a small system. Alice measures her qubit; the measuring device injects energy, and the local entanglement is broken. She then sends Bob one classical bit, heads or tails. Because the ground state’s correlations already reach across to Bob, that single bit tells him exactly which operation to perform on his own qubit, and performing it lets his device draw energy back out of the system. Ikeda built the two-qubit circuit for it, six gates deep, ran it on six IBM machines including one free to anyone in the world, and measured the local energy going below the ground-state level exactly as the theory says.
Connection
Chapter 6 is about drawing usable energy out of the structure of the vacuum, and this is the cleanest laboratory version of the idea: the energy comes from correlations already present in the ground state, released by a local operation that only a distant measurement result can specify. Chapter 2 gains a working demonstration that a ground state has more structure than its average energy suggests. The bookkeeping stays ordinary throughout — Alice’s device puts the energy in and Bob’s device takes it out — which is exactly why the result is checkable. Hotta’s review of the protocol is on this site at /library/stm-5aa6b8b006 and the distance-free version at /library/stm-2e579e12bc.
From the source
Quantum energy teleportation, proposed by Hotta about fifteen years before this paper and since studied for spin chains, an ion-trap system and a quantum Hall system, transfers energy using only local operations and classical communication: it is classical information rather than energy that travels the channel, and the intermediate subsystem between sender and receiver is not excited by the system’s energy carriers during the short duration of the protocol. It had been experimentally validated once before, in a nuclear magnetic resonance setup.Section I, opening paragraphs, citing references 5 to 12 and 20
The energy bookkeeping is ordinary and stated explicitly: any non-trivial local operation on the ground state of a many-body system raises its energy, and that increase is supplied by the measuring device. Because the pre-existing entanglement stores information about the injected energy at locations away from Alice, operations near Alice alone cannot recover it; Bob’s conditional operation, chosen using her result, is what allows his device to extract energy from the system, with conservation of energy holding throughout.Section I, third and fourth paragraphs; Equations 6 and 13
The minimal model runs on two qubits with a maximum circuit depth of six. The exact ground state is prepared with one rotation and one CNOT; the constants in the Hamiltonian are chosen so that this ground state reads zero energy for every local and global piece of it; Alice’s projective measurement of her Pauli X operator deposits a mean energy of h squared divided by the square root of h squared plus k squared; and Bob applies a rotation whose angle is fixed by h and k. Since the coupling term always has a negative expectation value and Bob’s local term always a positive one, it is enough for Bob to measure the coupling term alone to receive energy.Section II, subsections A to C, Equations 7 to 14; Figure 1
The central measurement is the observation of a local energy below the level the ground state defines as zero. Ikeda reports a negative coupling-term expectation value for every combination of the two model parameters on every machine used — six IBM devices, ibmq lima, ibmq jakarta, ibmq hanoi, ibm cairo, ibm auckland and ibmq montreal — with the closest approach to the exact value being minus 0.1079 at h of 1.5 and k of 1 on ibmq jakarta, about 76 percent of the analytic result. A simple measurement-error mitigation built from four calibration circuits improved the raw hardware values, in some cases enough to bring the negative expectation value into view.Section II D, Table I and Figure 2(B); Tables III and IV in the appendices
The transfer beats the system’s own dynamics. Alice must report her result within a time far shorter than the coupling timescale, and in this experiment that meant of order ten nanoseconds against a coupling time of order one hundred; on the hardware a single-qubit operation takes 20 to 40 nanoseconds while a two-qubit CNOT takes 200 to 500, so Bob extracts his energy faster than the natural unitary evolution of the system could carry energy from Alice to him.Section I, fifth paragraph; Section II D and Table III
The protocol is now a reproducible benchmark rather than a proposal: a two-qubit device six gates deep is enough, the circuits are published as open-access code, and the machine properties are public in real time, so anyone can repeat the run. The named next steps are a version with no limit of distance and a universal protocol for arbitrarily large quantum networks, with the milestone being quantum energy teleportation across a real network — the author points to the 158 kilometre Stony Brook to Brookhaven link on Long Island, and to quantum networks expected in practical use around the 2030s.Section I, second paragraph; Section III, first two paragraphs, citing references 21, 26 and 27
The way in
https://doi.org/10.1103/PhysRevApplied.20.024051How to cite it
Kazuki Ikeda (2023) Demonstration of Quantum Energy Teleportation on Superconducting Quantum Hardware doi:10.1103/PhysRevApplied.20.024051