Quantum computers have the potential to perform certain tasks that even the world’s most powerful supercomputers cannot. In the future, scientists anticipate using quantum computing to emulate material systems, simulate quantum chemistry and optimize difficult tasks, with implications ranging from finance to medicine. There is a nature.
But delivering on this promise requires resilient and scalable hardware. One of the challenges in building large-scale quantum computers is that researchers must find effective ways to interconnect quantum information nodes (small processing nodes isolated across computer chips). is. Quantum computers are fundamentally different from classical computers, so the traditional techniques used to communicate electronic information do not translate directly to quantum devices. However, one requirement is certain. That is, we need to send and receive the information being conveyed, whether through conventional or quantum interconnects.
To this end, MIT researchers have developed a quantum computing architecture that enables scalable, high-fidelity communication between superconducting quantum processors.in a work published in natural physics, MIT researchers demonstrate step 1, the deterministic emission of a single photon (information carrier) in a user-specified direction. Their method guarantees more than 96% of the time that quantum information flows in the right direction.
Linking several of these modules together will enable a larger network of interconnected quantum processors regardless of their physical separation on a computer chip.
“Quantum interconnection is an important step towards modular implementation of large-scale machines built from small individual components,” said Bharath Kannan PhD ’22, co-author of the research paper describing the technology. increase.
“The ability to communicate between small subsystems enables a modular architecture for quantum processors, which scales to larger system sizes compared to a brute-force approach using a single large and complex chip. It could be an easier way to do that,” adds Kannan.
Kannan wrote the paper with co-lead author Aziza Almanakly, a graduate student in electrical engineering and computer science in the Engineering Quantum Systems group at MIT’s Research Laboratory of Electronics (RLE). The senior author is William D. Oliver, Professor of Electrical Engineering, Computer Science, and Physics, MIT Lincoln Laboratory Fellow, Director of the Center for Quantum Engineering, and Associate His Director of the RLE.
Moving quantum information
In a traditional classical computer, different components perform different functions such as memory, computation, etc. Electronic information, encoded and stored as bits (which can take the values 1 or 0), travels between these components using interconnects, which are wires. It moves electrons on a computer processor.
But quantum information is more complicated. Instead of holding only 0 or 1 values, quantum information can also be both 0 and 1 at the same time (a phenomenon known as superposition). Quantum information can also be carried by particles of light called photons. These added complexities make quantum information fragile and cannot be simply transferred using traditional protocols.
Quantum networks link processing nodes using photons that travel through special interconnects called waveguides. A waveguide can be unidirectional, allowing photons to only move left or right, or it can be bidirectional.
Most existing architectures use unidirectional waveguides, which are easy to implement because the direction in which photons travel is easily established. However, since each waveguide only transports photons in one direction, more waveguides are required as the quantum network expands, making this approach difficult to scale. Additionally, unidirectional waveguides typically incorporate additional components to enhance directionality, resulting in communication errors.
“If we had a waveguide that could support propagation in both left and right directions, and a means of freely choosing the direction, we could eliminate these loss components. and is the first step towards higher fidelity two-way communication,” says Kannan.
Their architecture can be used to line up multiple processing modules along a single waveguide. A striking feature of the architectural design is that the same module can be used as both a transmitter and a receiver, he says. Also, photons can be transmitted and captured by any he two modules along a common waveguide.
“There is only one physical connection that allows you to connect any number of modules along the way. That is why it is scalable. We are working to do that,” adds Almanakly.
Utilization of quantum properties
To achieve this, the researchers built a module consisting of four qubits.
Qubits are the building blocks of quantum computers and are used to store and process quantum information. However, qubits can also be used as photon emitters. Adding energy to a qubit excites it, and when it is de-excited, the qubit emits energy in the form of photons.
However, just connecting one qubit to a waveguide does not guarantee directionality. A single qubit emits a photon, but it is completely random whether it moves left or right. To circumvent this problem, the researchers take advantage of his two qubits and a property known as quantum interference to ensure that emitted photons travel in the correct direction.
In this technique, we prepare two qubits in a single-excited entangled state called the Bell state. This quantum mechanical state consists of his two sides, the left qubit being excited and the right qubit being excited. Both aspects exist simultaneously, but it is unknown which qubit will be excited at any given time.
When the qubits are in this entangled Bell state, photons are effectively emitted into the waveguide at two qubit locations simultaneously, and these two ’emission paths’ interfere with each other. Depending on the relative phases within the Bell states, the resulting photon emission should move left or right. By preparing Bell states with the correct phase, researchers choose the direction in which photons travel through the waveguide.
This same technique can be used in reverse to receive photons with another module.
“Photons have a certain frequency, a certain energy, and by tuning the modules to the same frequency, you can prime the modules. If they’re not on the same frequency, the photons just pass by. It’s like tuning into a radio station: if you choose the right radio frequency, you can receive music transmitted on that frequency,” says Almanakly.
The researchers found that their technique achieved fidelity of over 96%. This means that if you try to emit a photon to the right, 96% of the time you went right.
Having used this technique to effectively emit photons in a specific direction, researchers hope to connect multiple modules and use this process to emit and absorb photons. This is a big step towards developing a modular His architecture that combines many smaller processors into one larger and more powerful His quantum processor.
