Machine Learning-Assisted Precision Manufacturing of Atom Qubits in Silicon.

Donor-based qubits in silicon, manufactured using scanning tunneling microscope (STM) lithography, provide a promising route to realizing full-scale quantum computing architectures. This is due to the precision of donor placement, long coherence times, and scalability of the silicon material platform. The properties of multiatom quantum dot qubits, however, depend on the exact number and location of the donor atoms within the quantum dots. In this work, we develop machine learning techniques that allow accurate and real-time prediction of the donor number at the qubit site during STM patterning. Machine learning image recognition is used to determine the probability distribution of donor numbers at the qubit site directly from STM images during device manufacturing. Models in excess of 90% accuracy are found to be consistently achieved by mitigating overfitting through reduced model complexity, image preprocessing, data augmentation, and examination of the intermediate layers of the convolutional neural networks. The results presented in this paper constitute an important milestone in automating the manufacture of atom-based qubits for computation and sensing applications.

Engineering Spin-Orbit Interactions in Silicon Qubits at the Atomic-Scale.

Spin-orbit interactions arise whenever the bulk inversion symmetry and/or structural inversion symmetry of a crystal is broken providing a bridge between a qubit's spin and orbital degree of freedom. While strong interactions can facilitate fast qubit operations by all-electrical control, they also provide a mechanism to couple charge noise thereby limiting qubit lifetimes. Previously believed to be negligible in bulk silicon, recent silicon nano-electronic devices have shown larger than bulk spin-orbit coupling strengths from Dresselhaus and Rashba couplings. Here, it is shown that with precision placement of phosphorus atoms in silicon along the [110] direction (without inversion symmetry) or [111] direction (with inversion symmetry), a wide range of Dresselhaus and Rashba coupling strength can be achieved from zero to 1113 × 10-13eV-cm. It is shown that with precision placement of phosphorus atoms, the local symmetry (C2v, D2d, and D3d) can be changed to engineer spin-orbit interactions. Since spin-orbit interactions affect both qubit operation and lifetimes, understanding their impact is essential for quantum processor design.

Atomic Engineering of Molecular Qubits for High-Speed, High-Fidelity Single Qubit Gates.

Universal quantum computing requires fast single- and two-qubit gates with individual qubit addressability to minimize decoherence errors during processor operation. Electron spin qubits using individual phosphorus donor atoms in silicon have demonstrated long coherence times with high fidelities, providing an attractive platform for scalable quantum computing. While individual qubit addressability has been demonstrated by controlling the hyperfine interaction between the electron and nuclear wave function in a global magnetic field, the small hyperfine Stark coefficient of 0.34 MHz/MV m-1 achieved to date has limited the speed of single quantum gates to ∼42 µs to avoid rotating neighboring qubits due to power broadening from the antenna. The use of molecular 2P qubits with more than one donor atom has not only demonstrated fast (0.8 ns) two-qubit SWAP gates and long spin relaxation times of ∼30 s but provides an alternate way to achieve high selectivity of the qubit resonance frequency. Here, we show in two different devices that by placing the donors with comparable interatomic spacings (∼0.8 nm) but along different crystallographic axes, either the [110] or [310] orientations using STM lithography, we can engineer the hyperfine Stark shift from 1 MHz/MV m-1 to 11.2 MHz/MV m-1, respectively, a factor of 10 difference. NEMO atomistic calculations show that larger hyperfine Stark coefficients of up to ∼70 MHz/MV m-1 can be achieved within 2P molecules by placing the donors ≥5 nm apart. When combined with Gaussian pulse shaping, we show that fast single qubit gates with 2π rotation times of 10 ns and ∼99% fidelity single qubit operations are feasible without affecting neighboring qubits. By increasing the single qubit gate time to ∼550 ns, two orders of magnitude faster than previously measured, our simulations confirm that >99.99% single qubit control fidelities are achievable.

The Use of Exchange Coupled Atom Qubits as Atomic-Scale Magnetic Field Sensors.

Phosphorus atoms in silicon offer a rich quantum computing platform where both nuclear and electron spins can be used to store and process quantum information. While individual control of electron and nuclear spins has been demonstrated, the interplay between them during qubit operations has been largely unexplored. This study investigates the use of exchange-based operation between donor bound electron spins to probe the local magnetic fields experienced by the qubits with exquisite precision at the atomic scale. To achieve this, coherent exchange oscillations are performed between two electron spin qubits, where the left and right qubits are hosted by three and two phosphorus donors, respectively. The frequency spectrum of exchange oscillations shows quantized changes in the local magnetic fields at the qubit sites, corresponding to the different hyperfine coupling between the electron and each of the qubit-hosting nuclear spins. This ability to sense the hyperfine fields of individual nuclear spins using the exchange interaction constitutes a unique metrology technique, which reveals the exact crystallographic arrangements of the phosphorus atoms in the silicon crystal for each qubit. The detailed knowledge obtained of the local magnetic environment can then be used to engineer hyperfine fields in multi-donor qubits for high-fidelity two-qubit gates.

Ramped measurement technique for robust high-fidelity spin qubit readout.

State preparation and measurement of single-electron spin qubits typically rely on spin-to-charge conversion where a spin-dependent charge transition of the electron is detected by a coupled charge sensor. For high-fidelity, fast readout, this process requires that the qubit energy is much larger than the temperature of the system limiting the temperature range for measurements. Here, we demonstrate an initialization and measurement technique that involves voltage ramps rather than static voltages allowing us to achieve state-to-charge readout fidelities above 99% for qubit energies almost half that required by traditional methods. This previously unidentified measurement technique is highly relevant for achieving high-fidelity electron spin readout at higher temperature operation and offers a number of pragmatic benefits compared to traditional energy-selective readout such as real-time dynamic feedback and minimal alignment procedures.

Coherent control of a donor-molecule electron spin qubit in silicon.

Donor spins in silicon provide a promising material platform for large scale quantum computing. Excellent electron spin coherence times of [Formula: see text] µs with fidelities of 99.9% have been demonstrated for isolated phosphorus donors in isotopically pure 28Si, where donors are local-area-implanted in a nanoscale MOS device. Despite robust single qubit gates, realising two-qubit exchange gates using this technique is challenging due to the statistical nature of the dopant implant and placement process. In parallel a precision scanning probe lithography route has been developed to place single donors and donor molecules on one atomic plane of silicon with high accuracy aligned to heavily phosphorus doped silicon in-plane gates. Recent results using this technique have demonstrated a fast (0.8 ns) two-qubit gate with two P donor molecules placed 13 nm apart in natSi. In this paper we demonstrate a single qubit gate with coherent oscillations of the electron spin on a P donor molecule in natSi patterned by scanning tunneling microscope (STM) lithography. The electron spin exhibits excellent coherence properties, with a [Formula: see text] decoherence time of 298 ± 30 µs, and [Formula: see text] dephasing time of 295 ± 23 ns.

Exploiting a Single-Crystal Environment to Minimize the Charge Noise on Qubits in Silicon.

Electron spins in silicon offer a competitive, scalable quantum-computing platform with excellent single-qubit properties. However, the two-qubit gate fidelities achieved so far have fallen short of the 99% threshold required for large-scale error-corrected quantum computing architectures. In the past few years, there has been a growing realization that the critical obstacle in meeting this threshold in semiconductor qubits is charge noise arising from the qubit environment. In this work, a notably low level of charge noise of S0  = 0.0088 ± 0.0004 µeV2 Hz-1 is demonstrated using atom qubits in crystalline silicon, achieved by separating the qubits from surfaces and interface states. The charge noise is measured using both a single electron transistor and an exchange-coupled qubit pair that collectively provide a consistent charge noise spectrum over four frequency decades, with the noise level S0 being an order of magnitude lower than previously reported. Low-frequency detuning noise, set by the total measurement time, is shown to be the dominant dephasing source of two-qubit exchange oscillations. With recent advances in fast (≈µs) single-shot readout, it is shown that by reducing the total measurement time to ≈1 s, 99.99% two-qubit S W A P gate fidelities can be achieved in single-crystal atom qubits in silicon.