PsiQuantum’s goal is to build the world’s first useful quantum computer out of conventional silicon chips that process quantum information using individual photons.
After several years of joint development with our manufacturing partner GLOBALFOUNDRIES (GF), we have taken a major step towards that goal with our Q1 system: an integrated, silicon photonics-based quantum system. We are now producing Q1 system components and quantum chips, proving our ability to manufacture hundreds of thousands of components, such as single-photon sources and single-photon detectors, utilizing the 300 mm manufacturing process of GF’s world-leading semiconductor foundry.
A quantum computer will be useful when it can solve problems with a proven advantage over existing computing methods. These quantum algorithms require many billions of “gate operations”, and because quantum states are extremely fragile, we need to use quantum error correction to avoid the computation being ruined by random noise. This means that, regardless of the technology, the architecture for a useful quantum computer needs to be designed around quantum error correction. This comes with a very large overhead and is the reason that we need to aim towards millions of qubits. Our approach to solving this with photons is with a fusion-based quantum computing architecture, which was recently introduced in this technical paper.
The scale of a genuinely useful quantum computer is unlike any quantum information system we have seen to date. Because of this, developing a quantum technology that can scale to millions of physical qubits has been the single driver of our technology roadmap, which recognizes that the shortest path to a useful quantum computer is not one of incrementally increasing the number of qubits.
Our roadmap aims to progressively validate our approach, by resolving scaling, integration and manufacturing challenges. With our Q1 system, we have taken the first major step along this path.
Our manufacturing partnership is a critical part of the roadmap towards a million qubits. We are delighted to be partnering with GF, one of a select few world-leading semiconductor foundries, and in our opinion the very best in silicon photonics manufacturing. The quality of manufacturing at GF enables PsiQuantum to achieve the component performances we need, as well as the yield and process reliability necessary to build a full-scale quantum computer.
Our Q1 system represents the first system milestone in our roadmap to a full-scale quantum computer. It is designed to help us iterate, improve, and validate our manufacturing approach. It is not intended to achieve any form of quantum advantage. However, it is a crucial step in our roadmap as it allows us to resolve system and performance considerations for all the technologies involved.
The Q1 system demonstrates our ability to manufacture a scalable, fully integrated quantum system with photon sources, photon detectors, integrated control electronics, and other critical elements of a quantum computer, using the standard manufacturing processes of GF’s world-leading semiconductor fab.
This photograph shows a 300mm silicon wafer, designed by PsiQuantum and built within the GLOBALFOUNDRIES manufacturing process; it contains multiple copies of our unpackaged Q1 chip. There are 26 layers in the stack and more than 500 process steps involved in getting the wafer fabricated. Each of these wafers has tens of thousands of single photon sources and tens of thousands of single photon detectors on it, and we have now processed over a thousand wafers.
In the process of developing this system, we have gone through multiple design iterations and learning cycles, testing the performance of tens of thousands of individual components per wafer. This scale of testing relies on the high throughput testing tools developed by the microelectronics industry. In turn, this gives us access to large data sets, which has been crucial for the process and design co-optimization needed to engineer the highest performance and highest yield devices. This approach, which would have been impossible had we not partnered with GLOBALFOUNDRIES, is what gives us confidence in our ability to build quantum systems of the scale required to be genuinely useful quantum devices.
One of the most important technical achievements in the Q1 system is the introduction of manufacturable superconducting single-photon detectors into the process line at GF. While integrated superconducting single photon detectors have been built before, what we have developed in the Q1 system is a substantial step forward in terms of performance at this level of scale and integration. It’s the advanced tools and the rigorous process control, as well as the monitoring of the CMOS microelectronics industry that have enabled these high-performance devices. It is this technology that will allow us to scale to the millions of single photon detectors required for fault-tolerant quantum computing.
Q1 is far more than a quantum photonic chip with high performing components: it integrates new developments in control electronics, optical I/O, packaging, and the assembly that surrounds the photonic chip. With a million qubits as the goal, the technologies that will enable scalable manufacturing are just as critical as high qubit gate fidelities.
There will be additional important steps along our path to building the world’s first useful quantum computer with hundreds of logical qubits and billions of gates. As Jeremy discussed in a recent talk, we are confident that by the middle of this decade we will have completely stood up all the manufacturing lines and processes necessary to be able to begin assembling a fault-tolerant quantum computer.
Our mission of developing a full-scale million-qubit quantum computer continues to drive our work every day, and after a few quiet and very productive years, it’s very exciting to be able to share the significant progress we have made. Stay tuned for more to come.
Mercedes Gimeno-Segovia is Sr. Director of Quantum Architecture at PsiQuantum. She received her PhD from Imperial College London for her work on linear optical quantum computing architectures. After postdoctoral positions in Bristol (UK) and Calgary (Canada), she joined PsiQuantum in 2017, where she leads a team working on the design and development of an architecture for universal fault-tolerant quantum computing using silicon photonics.