We advanced ²⁸Si/SiGe and Ge/SiGe heterostructures for the next generation of spin-qubit processors. In silicon we obtained two-dimensional electron gas with enhanced and more uniform transport properties across a 100 mm wafer (Appl. Phys. Lett.). We also elucidated and statistically predicted the valley splitting by the holistic integration of 3D atomic-level properties, theory and transport (Nat. Comm.). In germanium we obtained mobilities in excess of one million cm²/Vs (Appl. Phys. Lett.,). We also pioneered germanium bilayers towards 3D quantum devices (Adv. Quantum Tech., front cover).

In germanium, we programmed the first rudimentary error correction circuit by realizing the phase-flip code on a four-qubit germanium quantum processor (NPJ Quantum Information).

We reported a two-qubit gate fidelity between two electron spins in Si/SiGe quantum dots of 99.65% (Nature). This constitutes a much-anticipated achievement that brings the spin qubit field in a position to achieve fault-tolerance.

We realized a linear array of 6 electron spin qubits in a Si/SiGe quantum dot array (Nature). This experiment not only pushes the envelope of the number of qubits in silicon, it also achieves high initialization fidelities, high readout fidelities, high single-qubit gate fidelities, and high two-qubit state fidelities - all at the same time.

Working closely together with engineers from Intel, we measured the first qubits made in the very same industrial manufacturing facilities that are used to mass-produce conventional computer chips (Nature Electronics). This advance has been a long-standing goal due to its promise of scalability and was featured on the cover of Nature Electronics.

Our Starmon-5 Quantum Computer available online via Quantum Inspire has undergone key upgrades, including faster two-qubit gates (from 60 to 40 ns), faster qubit readout (from 3 to 1 ms), and most importantly the ability to perform multiple measurements within a circuit, thus enabling mid-circuit measurement, as required in quantum teleportation and quantum error correction.

We joined the OpenSuperQPlus consortium of the European Quantum Technology Flagship, starting in March 2023, to build a 100-qubit computer available in the cloud within 3.5 years, in collaboration with TNO, Orange Quantum Systems, Qblox, QuantWare, and Delft Circuits.

We demonstrated a cryo-CMOS digital-to-analog converter for multiplexed spin-qubit biasing, able to bias up to 5000 spin qubits while keeping an extremely low power consumption compatible with sub-K refrigeration stages (VLSI Techn. and Circuits).

To provide the frequency reference for the qubit control and readout circuits, we developed a calibration technique to minimize the phase noise of cryo-CMOS voltage-controlled oscillators (TCAS-I), and we demonstrated the first cryogenic phase-locked loop (PLL) operating at 4.2 K, able to synthesize 9.4 to 11.6 GHz tones while maintaining high qubit fidelity (JSSC).

We analysed the trade-off in the development of real-time hardware neural-network decoders for surface-code quantum-error-correction schemes, laying the basis for cryo-CMOS implementation of quantum error correction decoders. (TQE).

We investigated the need for tight co-design of different layers of a full-stack quantum computing system and emphasized the development of hardware- and algorithm-driven compilation techniques (DATE).

We continued our efforts towards modular quantum computer architectures by investigating the mapping of quantum algorithms to multi-chip quantum computers as well as characterizing the spatio-temporal traffic overhead between different chips (NANOCOM).

We proposed a two-qubit gate between a fluxonium and a transmon qubit for which we show high fidelity by numerical simulation, and which could function as an essential component in a hybrid transmon-fluxonium qubit code architecture. We show how the gate would avoid the problem of frequency collision in all-transmon, all-microwave architectures (Phys. Rev. Research 2022).

We added functionality to perform mid-circuit measurements on backends that support this feature. This enables the execution of dynamic quantum circuits, where the flow is determined by the results of intermediate measurements.

We performed a huge software overhaul to support hybrid computing using both quantum resources and classical resources. This feature will become fully available in the summer of 2023 for all backends that support this feature.

We improved the duration of gates and fidelities for qubit control on the Starmon backend, increasing the length of quantum circuits that can be implemented on that hardware.