Fujitsu unveils world's first prototype of diamond-spin quantum computer

Fujitsu has developed what it describes as the world's first working prototype of a diamond-spin quantum computer that incorporates tin-vacancy (SnV) centers into photonic integrated circuits. The prototype can operate at -271.6°C, a somewhat higher temperature than the -273.13°C typically required by superconducting quantum computers, and Fujitsu has demonstrated in a test environment that it can be operated through the Fujitsu Hybrid Quantum Computing Platform without requiring specialized quantum expertise. The development marks a step toward a modular quantum computing architecture, considered one of the more promising approaches to scaling up quantum computers because of its high fidelity, meaning low error rates, and efficient optical connections between modules.

The diamond-spin approach uses lattice defect structures known as color centers within diamond crystals as qubits, the basic units of quantum information. This approach can achieve high fidelity because diamond's inherent physical properties help keep quantum states stable, potentially allowing reliable logical qubits to be built from fewer physical qubits than approaches like superconducting quantum computing require. Additionally, light can be used to flexibly connect multiple quantum modules together, enabling a modular system architecture that can scale more efficiently.

Diamond-spin quantum computing typically relies on structures called nitrogen-vacancy (NV) centers, formed by nitrogen atom impurities within diamond crystals. For this prototype, however, Fujitsu instead used tin-vacancy (SnV) centers, which have a more symmetrical structure and are less sensitive to external noise than NV centers, making them attractive candidates for creating stable, high-brightness color centers.

The prototype builds on joint research that began in 2020 between Fujitsu, Delft University of Technology, and QuTech, a leading quantum technology research institute affiliated with TU Delft.

Fujitsu plans to develop a prototype multi-module diamond-spin quantum computer by 2027, and will also begin developing technology to integrate the diamond-spin approach with superconducting quantum computing, aiming to drive progress toward larger-scale quantum computers. This work is part of Fujitsu's broader quantum computing roadmap, which outlines the company's goal of achieving practical quantum computing by 2030.

Vivek Mahajan, corporate executive officer, corporate vice president, and CTO in charge of System Platform at Fujitsu, said the diamond-spin approach used in this prototype offers strong scalability on its own, but also has the potential to be integrated with superconducting quantum computers to further extend their capabilities, enabling more complex, larger-scale computations. He said that under Fujitsu's roadmap to reach a 250-logical-qubit system by fiscal year 2030 and a 1,000-logical-qubit system by fiscal year 2035, the company will continue advancing practical quantum computing across both software and hardware, leveraging the diamond-spin approach's key advantages in fidelity and optical connectivity.

Dr. Kees Eijkel, general director of QuTech at Delft University of Technology, said the organizations are pleased to announce this prototype as the result of collaborative research conducted since 2020 between Fujitsu, TU Delft, and QuTech, describing it as a major milestone in their partnership. He noted that demonstrating the scalability expected of diamond-spin quantum computing remains a long and challenging path, but said that by further strengthening the collaboration with Fujitsu, the institutions are committed to tackling that ambitious challenge and leading development of next-generation quantum technology.

The prototype relies on three technologies Fujitsu developed specifically for this work. The first is heterogeneous material bonding and thinning technology for building scalable quantum computing chips. To create quantum chips using SnV centers, Fujitsu developed a technique to bond high-quality diamond substrates, implanted with tin ions, onto alumina and silicon dioxide substrates. The company also developed thinning technology to reduce the thickness of these diamond substrates from several hundred micrometers down to several hundred nanometers, making them suitable for use in quantum computing chips.

The second is photonics-integrated circuit fabrication technology for SnV centers. Fujitsu developed a method to build photonic integrated circuits that combine nanometer-scale diamond crystals containing SnV centers with alumina optical waveguides, which are transparent to visible light, in order to extract the individual photons emitted by SnV centers during qubit readout. For the diamond processing itself, Fujitsu drew on joint research conducted with the University of Tokyo.

The third is quantum circuit conversion technology specific to the diamond-spin approach. Because this approach requires controlling qubits using a combination of light, microwaves, and radio frequency waves, Fujitsu developed a mechanism to convert quantum circuits, described using standard quantum logic gates, into the physical control sequences needed for these operations, enabling the system to be controlled through Fujitsu's hybrid quantum computing platform.

Part of this research was supported by the Advanced Research Infrastructure for Materials and Nanotechnology in Japan (ARIM) program under Japan's Ministry of Education, Culture, Sports, Science and Technology, and the broader research project is a collaboration between QuTech and Fujitsu, co-financed by Holland High Tech through funding for research and development in the Netherlands' high-tech systems and materials sector, along with research and innovation funding from the Dutch Ministry of Economic Affairs.