Hole spin qubits in silicon quantum dots fabricated in planar 300 mm manufacturing technology: from detailed interface physics to fast, high fidelity multi-qubit control and readout

Electron spin based qubits in quantum dots have been among the leading candidates for a potential scaled, fault tolerant quantum computer for the past 25 years [1]. The low spin orbit interaction of electrons results in naturally well protected qubits, with long coherence times. However, electron spins also have downsides: this same lack of natural spin-orbit coupling results in relatively slow coherent spin manipulation, either via ESR or EDSR – the latter using magnetic gradients that induce artificial spin-orbit effects. In addition, strong contact hyperfine interactions with nuclear spins can limit spin coherence. Some 20 years, when III-V quantum dots were still largely en vogue, several proposals therefore pointed to hole based qubits for faster manipulation and lower nuclear spin dephasing [2]. Despite initial results [3,4], it became soon clear that the intricacies and degeneracies of the valence band in covalent semiconductors made exact control over the hole wave function in quantum dots a strong function of strain, electric fields and other control knobs – further complicating progress. Fast forward 20 years later: with well-controlled, low-strain silicon quantum dots now available in advanced silicon manufacturing facilities such as the one at imec, the

question re-emerges whether hole qubits could complement or even replace electron doped quantum dots towards a scaled quantum processor, despite (or even thanks to) the complicated band structure. We will show recent data of the effect of ultra-clean, low disorder SiMOS interfaces - as previously evaluated for electron doped quantum dots [5] – on holes, and demonstrate record high mobilities that are nevertheless an order of magnitude lower than for electrons: a consequence of the complex band structure of holes [6]. Using coupled quantum dot hole qubits, and without a need for ‘sweet spot’ operation, we nevertheless show for the first time high fidelity initialization and readout, fast and high fidelity single qubit control, and two-qubit control on one and the same device, that is in principle scalable to a large array [7]. While holes display significantly more complex physics, our results show that holes in Si MOS could indeed complement electron qubits in a future silicon quantum dot processor.

[1] D. Loss and D. DiVincenzo, Phys. Rev. A 57, 120 (1998)

[2] D. Bulaev and D. Loss, PRL 95, 076805 (2005)

[3] D. Brunner et al., Science 325, 70 (2009)

[4] K. De Greve et al., Nat. Phys. 7, 872 (2011)

[5] A. Elsayed et al., NPJ Quantum Information 10, 70 (2024)

[6] J. P. Wendoloski et al., arXiv:2502.21173 (2025)

[7] I. Vorreiter et al, unpublished (2025)