Quantum computing is set to revolutionize the area of intensive computing and open up new possibilities for solving certain specific problems, deemed intractable, within a reasonable timeframe.
The construction of a quantum computer does however come up against several obstacles. The largest of these being decoherence, that is to say the fact that the system loses its quantum properties as a result of its interactions with its environment (with light for example).
This problem, which amplifies with the addition of qubits (the building blocks of quantum computers), severely limits the size of quantum machines and thus their performances and advantages compared to their “traditional” counterparts.
To remove the barriers, several avenues are being explored, in particular around the production of qubits. Among these, atoms that have been cooled by laser appear to be promising candidates to assemble devices with a high number of qubits and a high degree of control.
The potential of cold neutral atoms
Founded by Georges-Olivier Reymond in 2019, Pasqal emerged from the work carried out in the 2010s by Antoine Browaeys and Thierry Lahaye at the Charles Fabry laboratory (Institut d’Optique Graduate School, CNRS, Université Paris-Saclay) on the simulation of n-body problems thanks to a programmable quantum simulator composed of a network of non-ionized (“neutral”) atoms.
Rubidium atoms are cooled to several microkelvins by lasers and trapped inside a mechanism called a “magneto-optical trap”. They are then handled individually with optical tweezers to create atom matrices that are freely reconfigurable in 2D and even 3D.
Pasqal claims that neutral atom device architectures are unique, not only in comparison with classical devices but also with respect to its quantum counterparts (superconducting circuits, used by IBM or Google, are currently one of the most common approaches in quantum computing).
In effect, these atoms have several properties of interest. They are intrinsically identical, which is a strong asset when they are used as qubits to obtain low error rates during computing.
Trapping them makes it possible to isolate them from the outside world and thus helps preserve their quantum behavior. What’s more, thanks to recent breakthroughs in the areas of atomic physics and laser technology, they can be handled highly precisely.
Neutral atom platforms thus make it possible to obtain quantum registers (the processor’s memory location) that are bigger (meaning they are made up of a greater number of qubits) than with other quantum platforms.
The two main applications envisaged are quantum simulation and solving difficult mathematical problems. These are described in this presentation article published in 2020.
Stimulating the discovery of new drugs and materials
Quantum simulation makes it possible to explore complex physical phenomena in many scientific areas, such as condensed matter physics (the most active field of contemporary physics, including a wide range of specializations such as semiconductor physics) or quantum chemistry.
Programmable processors such as Pasqal could thus be used to help basic research to progress, with highly reduced energy consumption and computation costs, whilst also being of benefit to industry with many use cases in molecular and materials engineering.
It is in the latter area that Pasqal has launched the QUACHA (“Quantum Chemistry with Atoms Arrays”) project in partnership with startup Rahko.
The aim of this research and development project, jointly funded by the Île-de-France region and the French State, is to “progress in the understanding of interactions between certain classes of enzymes and their substrates, with applications for the development of drugs, but also for the search for better-performing catalysts for industry”.
The MIS problem, HPC computing, and smart charging
Beyond scientific simulation, neutral atom processors can be used to solve difficult mathematical problems.
Applications include the search for approximate solutions to complex combinatorial optimization problems, solving nonlinear partial differential equations, or improving the performances of machine learning procedures.
“This problem [of optimization], which has various direct applications in network design or finance, becomes hard to solve on a classical computer when the size of the graph grows”, the researchers explain.
“Solving efficiently the MIS problem would provide a way to solve any interval scheduling problem, with applications in many fields (telecommunication, computing tasks allocation in HPC [High Performance Computing], or even satellite photography to name a few).”
It is in the area of smart charging, where French electric utility company EDF faces scheduling problems, that Pasqal found one of its first industrial partners.
With demand on the increase, the company is turning to quantum computing to optimize the management of its power generation facilities as well as the supply to its EV charging stations.
The complexity of smart charging mathematical problems (based on the modelling of the whole set of stations and the demand on each one of these) is such that they cannot be solved by traditional computing.
EDF therefore wishes to use a Quantum Approximate Optimization Algorithm (QAOA) and the aim of the partnership with Pasqal is to implement this algorithm on the processor developed by the startup.