A 300 mm foundry silicon spin qubit unit cell exceeding 99% fidelity in all operations

TL;DR

This work demonstrates a silicon spin-qubit unit cell fabricated in a 300 mm CMOS foundry that achieves high-fidelity universal control, including a CZ entangling gate, with all gate fidelities exceeding 99% and SPAM above 99.9%. Using gate-set tomography on a full gate set, the authors reveal that the dominant residual errors arise from coupling to residual $^{29}$Si nuclear spins rather than charge noise, with isotopic purification proposed to push fidelities further. The device exhibits long $T_1$ and $T_2$ times (up to $T_1$ of several seconds and $T_2^{\mathrm{Hahn}}$ up to hundreds of microseconds) and operates within an industrial fabrication framework, addressing a key scalability question for wafer-scale quantum processors. Overall, the results confirm that high-quality spin qubits can be realized in commercial CMOS processes, providing a viable path toward scalable, fault-tolerant quantum computation with millions of qubits.

Abstract

Fabrication of quantum processors in advanced 300 mm wafer-scale complementary metal-oxide-semiconductor (CMOS) foundries provides a unique scaling pathway towards commercially viable quantum computing with potentially millions of qubits on a single chip. Here, we show precise qubit operation of a silicon two-qubit device made in a 300 mm semiconductor processing line. The key metrics including single- and two-qubit control fidelities exceed 99% and state preparation and measurement fidelity exceeds 99.9%, as evidenced by gate set tomography (GST). We report coherence and lifetimes up to $T_\mathrm{2}^{\mathrm{*}} = 30.4$ $μ$s, $T_\mathrm{2}^{\mathrm{Hahn}} = 803$ $μ$s, and $T_1 = 6.3$ s. Crucially, the dominant operational errors originate from residual nuclear spin carrying isotopes, solvable with further isotopic purification, rather than charge noise arising from the dielectric environment. Our results answer the longstanding question whether the favourable properties including high-fidelity operation and long coherence times can be preserved when transitioning from a tailored academic to an industrial semiconductor fabrication technology.

Figures (8)

• Figure 1: Two-qubit device and operation points.a, Schematic of a Diraq two-qubit device fabricated on a 300mm wafer, showing the full wafer, single die, and single device level. b, Charge stability diagram as a function of plunger (P1, P2) voltage detuning $\Delta V_\mathrm{P} = \Delta V_\mathrm{P1} = -\Delta V_\mathrm{P2}$ and exchange gate (J) voltage $V_\mathrm{J}$, showing four isolated electrons in the double-dot system. Voltage points for single-qubit operation ($J_{\rm off}$), two-qubit operation ($J_{\rm on}$) and readout are labelled as triangle ($\blacktriangle$), triangle down ($\blacktriangledown$), and star ($\bigstar$), respectively. Inset: Cross section of the double quantum dot device indicating electron numbers under the plunger gates (P1,P2), exchange gate (J), and double potential well.
• Figure 1: Single qubit metricsa,b, Spin relaxation of qubit 1 (a) and 2 (b) spin up to the ground state $\ket{\downarrow\downarrow}$. Curves are fitted to an exponential decay resulting in a relaxation time of $T_\mathrm{1,Q1} = 2.4\pm 0.2s$ and $T_\mathrm{1,Q2} = 6.3\pm 0.6s$, respectively. c,d, Ramsey experiment for qubit 1 (c) and 2 (d) fitted to a sinusoidal exponential decay resulting in a coherence time of $T_\mathrm{2,Q1}^{*} = 30.4\pm 0.8µ s$ and $T_\mathrm{2,Q2}^{*} = 29.1\pm 0.6µ s$, respectively. The oscillation is induced by a phase shift corresponding to a 1M Hz detuning. Every data point is averaged for 1000 repeats, each integrating the readout signal for $t_\mathrm{int} = 100µ s$. We use real-time Larmor frequency tracking between repeats using the protocol described in Ref. dumoulin_stuyck_silicon_2024. e,f, Hahn echo experiment for qubit 1 (e) and 2 (f) fitted to an exponential decay resulting in a coherence time of $T_\mathrm{2,Q1}^{\mathrm{Hahn}} = 445\pm 6µ s$ and $T_\mathrm{2,Q2}^{\mathrm{Hahn}} = 803\pm 6µ s$, respectively. In this experiment, we measure all six single qubit projections to fit the state purity $\gamma_\mathrm{state}$. Error bars represent the 95% confidence level.
• Figure 2: Single qubit operation.$\textbf{a, b,}$ Rabi chevron for qubit 1 and 2, respectively. $\textbf{c, d,}$ Coherent Rabi oscillations at $f_\text{Larmor}$ for qubit 1 and 2, respectively. Real-time feedback is implemented to counteract Larmor frequency deviation dumoulin_stuyck_silicon_2024. The local oscillator frequency is set to $f_{\mathrm{LO}} = 18.610G Hz$. The applied microwave power to Qubit 1 is approximately 7% larger resulting in matching Rabi frequencies of $f_\mathrm{Rabi} = 658.6 \pm 0.3k Hz$.
• Figure 2: Single qubit Larmor frequency tracking.a,b Single qubit Larmor frequency deviation as a function of lab time during the GST experiment discussed in the main text for qubit 1 and 2, respectively. Figures are plotted with the same y-axis range. Qubit 2 has a frequency standard deviation of $0.09MHz$ compared to $0.05MHz$ for qubit 1. Qubit 2 also shows signs of two discrete frequency levels in the histogram at $\approx 36.4MHz$ and $36.3MHz$. The local oscillator frequency is set to $f_{\mathrm{LO}} = 18.610G Hz$. Frequencies are tracked using the protocol described in Ref. dumoulin_stuyck_silicon_2024.
• Figure 3: Two qubit operation.a, b, Probability of detecting an even spin state, $P_\mathrm{even}$, after a microwave burst of fixed power and duration at different J gate voltages $V_\mathrm{J}$ when preparing a mixed odd state $\frac{1}{\sqrt{2}}(\ket{\downarrow\uparrow}+\ket{\uparrow\downarrow})$ (a), and a pure state $\ket{\downarrow\downarrow}$ (b), respectively. The power and duration of the microwave burst are roughly calibrated to a single-qubit $\pi$-rotation. The local oscillator frequency is set to $f_{\mathrm{LO}} = 18.610G Hz$. The following experiments are conducted with $\ket{\downarrow\downarrow}$ initialisation, unless otherwise specified. c Controlled phase (CZ) oscillation as a function of exchange time $t_\mathrm{exchange}$ and exchange gate voltage $V_\mathrm{J}$. We apply dynamical decoupling by a consecutive $\pi$-rotation on both qubits to filter out other phase accumulating effects such as AC Stark shift. d Decoupled exchange oscillation fingerprint for fixed exchange time $t_\mathrm{exchange} = 10µ s$ as a function of plunger voltage detuning $\Delta V_\mathrm{P}$ and exchange gate voltage $V_\mathrm{J}$. Triangle indicates the exchange gate voltage used for the CZ gate. e, Exchange strength as a function of exchange gate voltage extracted from fitting exchange oscillations in c. The controllability is fitted assuming $J_\mathrm{Exchange}\propto a\exp{bV_\mathrm{J}}$ resulting in $b = 18.4dec\per V$. f, g, Calibration of the CZ single qubit phase correction by preparing the target spin in superposition, applying a CPHASE followed by a virtual phase rotation for qubit 1 and 2, respectively xue_quantum_2022huang_high-fidelity_2024. Vertical lines indicate the phase values where the spin state of Q1 (Q2) is flipped if Q2 (Q1) is in the on/spin up state. Readout probability is unscaled in all data.