Photonic chips for fault-tolerant quantum computing


Article by: Maurizio Di Paolo Emilio

The collaboration between Xanadu and Imec includes the manufacture of low-loss silicon nitride circuits that can correct qubit errors and increase capacitance.

Xanadu and Imec have teamed up to develop photonic chips for fault-tolerant quantum computing. Their common goal is to develop a machine based on quantum theory that is able to execute any algorithm, detect and correct any error that could affect the calculation, and thus be able to accommodate a large number of qubits.

Xanadu and Imec said they want to use photonics to build quantum computers with millions of qubits that are immediately available to customers. Their collaboration includes making low-loss silicon nitride circuits that can correct qubit errors and increase capacitance.

In an interview with EE Times Europe, Amin Abbasi, Business Development Manager at Imec, and Zachary Vernon, Head of Hardware Xanadu, have shown the prospects of quantum computers for future applications. “There are some applications that may not require full fault tolerance to get any benefit. There are examples in machine learning and optimization. Xanadu has already made some chips available to users via its cloud platform that can have an impact in these areas. However, the greatest value is likely to be in chemical applications that cannot be addressed without a device with millions of physical qubits and error correction implemented, ”Vernon said.

[Webinar] How to overcome the EMI testing challenge

He added, “Demonstrating robust error correction and achieving error tolerance is the biggest hurdle quantum computing must overcome before it can address its most valuable applications. Good material production and high-quality lithography and chip packaging are crucial. “


Quantum computing potentially offers exponential acceleration compared to classic computing for certain tasks. A central and outstanding challenge in making quantum computing practicable is achieving fault tolerance, which means that calculations of any length and size can be carried out in the presence of noise, provided that it does not exceed a certain threshold.

At its most basic level, a quantum computer is a machine that creates qubits in certain states, transforms them using quantum gates, and then measures them. Quantum gates that act on states use the entanglement that merges qubits in such a way that a description of the state of each qubit is not possible.

In addition to this gate model, a quantum computer can be viewed as a cluster state and represented as pearls on a wire. As Xanadu points out on one of his blogs, the pearls here are qubits, and the wire represents the entanglement between them. The goal here is to measure the qubits. The measurement is not just a reading of the calculation; it is the real calculation. Another measurement setup leads to a different gate that is applied to the qubit.

For the computer to be universal, the cluster state must be at least two-dimensional; In order for it to be fault tolerant, it has to be three-dimensional. In standard computers we can protect every bit of information from errors through repetition or redundancy. With quantum computers, redundancy is forbidden. Therefore, 3D clusters are used, and their size and arrangement enable quantum discrete variable (DV) topological error correction.

Since they have built-in but sophisticated error correction functions, GKP states are ideal qubits. The Gottesman-Kitaev-Preskill (GKP) code is a potential approach for fault-tolerant quantum computation, as it encodes logical qubits into lattice states of harmonic oscillators. The quality of the network states, on the other hand, must be exceptionally high so that the code is fault-tolerant.

Silicon nitride enables the creation of compressed states that can be used to synthesize qubits instead of individual photons. The compressed states are generated deterministically and can be used to distill fault tolerant qubits (GKP states).

“The leading approaches are based on trapped ions, superconducting circuits and photonics. Photonics is much easier to scale because chips can be networked via fiber optics. However, the number of individual components required per qubit is also greater. Fortunately, chip integration and mass production of semiconductor processes help, ”said Vernon.

Xanadu’s photonic architecture essentially consists of four blocks: the state preparation factory, the multiplexer, the computation module and the photonic quantum processing unit (QPU). The multiplexer executes many generations of states in parallel to increase the likelihood of generating a GKP state. The QPU takes the measurements necessary to implement each quantum algorithm and correct errors.

Figure 1: GKP countries (source: Xanadu)


The photonic chip transports light instead of electrons. Xanadu’s approach offers several advantages such as scalability up to a million qubits over the optical network, room temperature computing, and the natural ability to run R&D centers like Imec, a semiconductor R&D center for advanced technologies on advanced 200mm and 300mm mm lines, as well as mass production on their 200 mm line.

“To achieve its full potential, quantum computing must include error correction,” Vernon said. “To do this, existing prototypes for qubit chips have to be scaled up to thousands of identical modules. Achieving this scaling with high performance at the same time – low losses and good uniformity – and the ability to connect these chips together is a great challenge and something for which the photonic approach is very well suited. “

“Imec’s role is to provide its industrial and academic customers with an extremely low-loss, integrated platform on a wafer scale,” said Abassi. “Quantum computing based on integrated photonics requires low-loss, highly scalable and powerful circuits. at Imec, we are working on these requirements in order to deliver the state-of-the-art photonic wafers to our customers. “

The modular architecture of photonic chips simplifies the design, manufacture and integration of quantum and other optoelectronic solutions. The cryogenic demand is reduced and the homodyne detectors set the device clock rate in the QPU, which works efficiently and quickly. The modularity, speed and operation at room temperature of the upcoming architectures will enable photonics to contribute to the rapid construction of quantum computers.

This article was originally published on EE Times Europe.

Maurizio Di Paolo Emilio has a Ph.D. in physics and is a telecommunications engineer and journalist. He has worked on various international projects in the field of gravitational wave research. He works with research institutions to develop data acquisition and control systems for space applications. He is the author of several books published by Springer as well as numerous scientific and technical publications on electronics design.

Congratulations to the first round winners! Round 2 starts now!


Comments are closed.