Professor Fan Shanhui, an alumnus of China University of science and technology, proposed a scalable optical quantum computer architecture

Photons, as the basis of quantum information processing, provide unique advantages, such as long coherence time, room temperature operation, suitable for quantum communication, but also bring basic scalability challenges. Non deterministic schemes will bring a lot of resource overhead, while deterministic schemes need a large number of the same quantum transmitters to realize large-scale quantum circuits.

Now, according to a paper published on optica on November 29, the research team led by fan Shanhui, Professor of Stanford University School of engineering, has proposed a scalable light quantum computer architecture, which uses only the least quantum resources to realize any quantum circuit: a single coherent controlled atom. The optical switch gives the time dimension of photon quantum state synthesis by modulating photon atom coupling. Quantum operations applied to atomic qubits can be teleported to photonic qubits by projection measurement, and any quantum circuit can be compiled into a sequence of these teleportation operations.

This design does not need to integrate many of the same quantum emitters into photonic circuits, and allows effective full connection between photonic qubits. The size of the proposed device is not affected by the depth of the quantum circuit, does not need a single photon detector, operates deterministically, and has high fidelity even in the presence of experimental defects.

The corresponding author of this paper, Professor Fan Shanhui, is a member of the American Physical Society, the international society of electronic and electrical engineers, the international society of optical engineering and the American Optical Society. Fan Shanhui middle school graduated from Guangdong Experimental Middle School. In 1988, he was admitted to the teaching reform pilot class (class zero) of the juvenile Department of the University of science and technology of China. In 1992, he entered the Massachusetts Institute of technology to study physics and obtained his doctorate in 1997. Fan Shanhui is currently a professor at the school of engineering, Stanford University.

Photons provide many advantages for quantum information processing. For example, optical qubits have very long coherence time, quantum states can run at room temperature, and they are the best choice for quantum communication.

At present, the main difficulty faced by all quantum computing architectures is scalability, especially for photonic systems. Optical qubits must propagate, so they must be processed by transmitting photons through sequential optical elements during flight.

Since the photon quantum gate is a physical object (opposite to the sequential laser pulse used for atomic qubits), the machine size changes with the depth of the circuit, making the complex quantum circuit unimaginable and impossible to realize even using compact integrated photonics.

Further limiting the scalability of light quantum computers is the difficulty of integrating many high fidelity multiphoton gates into the optical path. This is a problem for both the non deterministic gate scheme and the method based on deterministic scattering. The former brings huge resource overhead to the fault-tolerant operation due to the low probability of gate success. Although the scattering based two-photon gate can be realized separately with high fidelity, the realization of large-scale quantum circuits requires a large number of the same quantum emitters. Due to the uniform and non-uniform broadening, the indistinguishability of solid-state quantum emitters is poor, which becomes more serious. Therefore, a quantum computer architecture using only a single quantum emitter to realize all gates in a quantum circuit will greatly improve the scalability and experimental feasibility of light quantum computation based on scattering.

In this paper, researchers use the concept of synthetic dimension. Synthetic dimension has recently aroused great interest in exploring topological physics in photonics, but it has not been widely used in quantum photonic systems.

In order to form a synthetic dimension, people design the coupling between system states by reusing the usual geometric dimensions, such as space and time, or increasing these dimensions by using internal degrees of freedom, such as frequency, spin and orbital angular momentum.

Because the coupling between states in the synthesis dimension can be reconfigured dynamically and will not be fixed by the physical structure, the lattice with complex connectivity can be implemented extensibly. This allows a single quantum emitter to manipulate multiple photon qubits in the synthesis space without the need for spatially separated structures.

The design proposed in this paper includes a fiber ring coupled to a cavity containing a single coherent control atomic qubit. The optical switch gives the time dimension of the synthesis of these photonic states by allowing the coupling between reverse cyclic photonic states.

By scattering photons onto atoms, then rotating and projecting to measure atomic states, the team can teleport the operation state to photon qubits; These operations can be combined to deterministically construct any quantum circuit. By sequentially exchanging atomic states with each photon qubit, the readout of photon quantum states can be performed without the need for a single photon detector.

The light quantum computer architecture described in this paper. (a) Photon qubits propagate back through the optical fiber storage ring. The optical switch can selectively guide photons through the scattering unit and interact with atoms in the laser coherently controlled cavity. (b) The energy structure of the atom: ω 1 resonates with the cavity mode and photon carrier frequency, while ω 0 is far detuning.

The design in this paper includes two main parts: storage ring and scattering unit. The function of the storage ring is similar to the memory in an ordinary computer. This is an optical fiber ring containing multiple photons, which propagate around the storage ring. Each photon in the system represents a qubit. The propagation direction of photons around the storage ring determines the value of qubits, which can be 0 or 1. Since photons can exist in two states at the same time, a single photon can flow in both directions at the same time, which is expressed as the value of the combination of 0 and 1.

Description of photon qubit state on Bloch sphere.

Researchers can manipulate photons by guiding photons from the storage ring to the scattering unit, where photons propagate into a cavity containing a single atom. Photons interact with atoms to produce "entanglement". The photons then return to the storage ring and the laser changes the state of the atom. Because atoms and photons are entangled, manipulating atoms will also affect the state of their paired photons. By measuring the atomic state, the operation can be teleported to photons. Only a controllable atomic qubit can be used as a proxy to manipulate all other photon qubits indirectly.

The quantum gate sequence corresponding to the photon passing through the scattering unit once. Projection measurement will be applied to the rotational teleportation of atomic qubits to photonic qubits.

Because any quantum logic gate can be compiled into a series of operations performed on atoms, in principle, you can run quantum programs of any size with only one controllable atomic qubit.

In order to run a quantum program, we need to convert the code into a series of operations, introduce photons into the scattering unit and manipulate atomic qubits. You can control the way atoms and photons interact, and the same device can run many different quantum programs.

A conceptual diagram of compiling a quantum circuit into a sequence of instructions to be executed on a device. (a) Universal target quantum circuit. (b) Equivalent circuit decomposed into single qubit and C σ Z door. (c) The circuit is further decomposed into a series of scattering interactions. This sequence can be assembled into an instruction set on a classical computer, which contains six different primitives corresponding to physical actions. (d) The controllable elements of quantum devices are optical switches, cavity lasers and atomic state readout.

Researchers say that for many optical quantum computers, the gate is the physical structure through which photons pass. If you want to change the running program, you usually need to physically reconfigure the hardware. But in this scheme, you don't need to change the hardware, just give the machine a different set of instructions.

The research shows that this scheme has high fidelity even in the case of experimental defects, and has significant advantages over many existing light quantum computing paradigms in the required physical resources and experimental feasibility.

The scheme in this paper has several unique characteristics.
Most notably, the only controllable quantum resource is a single atomic qubit, which acts as an agent for indirectly manipulating photonic qubits. All quantum operations and measurements on photon qubits are completed by operations performed on this atom, which are teleported to photons invisibly. This reduces the main implementation challenge of preparing a single strongly coupled atom cavity system.
In addition, the synthesis time dimension allows a single atom to act as the nonlinearity of all quantum gates and can provide an effective full connection between photon qubits.

The programmable nature of the teleportation gate allows atoms to sequentially apply each required single photon gate and two-photon gate without complex photon routing. This negates the requirement of traditional light quantum computing schemes to integrate many identical quantum emitters into photonic circuits.

Finally, this design does not require a single photon detector - an important limitation of light quantum computation. On the contrary, quantum teleportation technology can be used to measure atomic states with nearly 100% efficiency, which greatly improves the scalability of this design.

Compared with other quantum computing methods, this scheme shows a way to realize scalable, deterministic, gate based quantum computing using photonics. The scheme also does not need a single photon detector.

Compared with other QC platforms where qubits are a single physical structure (such as superconducting circuits and ion trap systems), only one controllable qubit has significant advantages in scalability: in their design, adding more qubits only needs to extend the fiber ring, while adding more qubits to superconducting devices requires adding complex independently addressable components.

However, this scheme is not without disadvantages: the design requires high cavity synergy and low fiber attenuation, which is challenging, but it is feasible to implement, and it depends on optical switches with very low insertion loss. The latest progress of lithium niobate modulators may achieve this soon. In addition, although only one controllable qubit greatly simplifies the experimental setup, it prevents the parallel execution of double qubit gates and most single qubit gates in quantum circuits.

In the future, if we can generalize the photonic storage mechanism of their proposed devices and consider the synthesis dimension other than time reuse, it may further improve the scalability of this design.

Link:https://doi.org/10.1364/OPTICA.424258