Session N46: Silicon Qubits III - Control and Readout

A 6-Qubit Electron Spin Quantum Computer in the cloud

Quantum Inspire (www.quantum-inspire.com) is a cloud platform giving access to quantum computers based on superconducting and semiconducting qubits (electrons in SiGe heterostructures) and an emulator. In 2020 it was launched with a 2-qubit Si QPU. We have incorporated a new 6-qubit chip, developed at QuTech, based on isotopically purified ²⁸Si to maximize spin coherence (Philips, S.G.J., Mądzik, M.T., Amitonov, S.V. et al. Universal control of a six-qubit quantum processor in silicon. Nature 609, 919–924 (2022)) in a full stack, cloud accessible quantum computer. Rabi frequencies are on the order of few MHz. Qubit initialization and measurement is done using Pauli spin blockade, combined with active feedback and post-selection. The stack includes a full control system, automated tuning and calibration to provide continuous availability, compilers, and account and data management facilities. It can be used in tandem with a classical runtime to execute hybrid quantum-classical algorithms and is provided with an software development kit for programming in Python plus a web GUI for users less familiar with quantum programming. We will present the latest results on the functionality and performance of the system, which is freely accessible to everyone. We will also present a preview of our roadmap towards a 200-qubit chip in the European flagship programs.   

Automated Calibrations in 6-Qubit Quantum Dot Spin Devices in Silicon

Quantum Inspire is a cloud platform giving access to quantum computers based on superconducting and semiconducting qubits. Structured tune-up and automated calibration techniques are vital for keeping these systems available online. In this talk, we will discuss the tune-up and automated calibration techniques utilized in our new six-qubit device [1] based on electron spins in silicon. The qubits are hosted in a linear array of six quantum dots in a SiGe/²⁸Si/SiGe heterostructure grown inhouse at QuTech [2]. Readout is done using Pauli Spin Blockade (PSB), single qubit operation through EDSR, and two-qubit gates are symmetric CZ gates.

Automated calibrations are performed using a directed acyclic graph to keep our qubits up and running. Nodes in the graph represent different calibration steps starting from tuning the charge sensor up to compensating for qubit crosstalk. We will go through the different calibration steps needed to keep a spin qubit device online. Finally, we will show some of the latest performance metrics on single and two-qubit gates reached by automated calibration and elaborate on our future plans.

References:

[1] S. G. J. Philips, M. T. Mądzik, S. V. Amitonov, S. L. de Snoo, M. Russ, N. Kalhor, C. Volk, W. I. L. Lawrie, D. Brousse, L. Tryputen, B. Paquelet Wuetz, A. Sammak, M. Veldhorst, G. Scappucci and L. M. K. Vandersypen, "Universal control of a six-qubit quantum processor in silicon," Nature 609, 919 (2022).

[2] D. D. Esposti, B. P. Wuetz, V. Fezzi, M. Lodari, A. Sammak and G. Scappucci, "Wafer-scale low-disorder 2DEG in 28Si/SiGe without an epitaxial Si cap," Applied Physics Letters, vol. 120, no. 18, p. 184003, 2022.   

The Dispersive Signature of Pauli Spin Blockade

Pauli Spin Blockade (PSB) serves as an efficient means of mapping the spin state of an electron or hole to the charge state of a quantum dot network. This spin-to-charge conversion is a key ingredient for quantum computation based on semiconductor spins, as it allows for efficient spin readout and initialisation with high-fidelity even at elevated temperatures [Camenzind et al. Nat. Electron. 5 (2022)]. Different readout schemes have been developed, including direct-current measurements, capacitively coupled dedicated sensors like single-electron transistors and gate-dispersive charge sensing. The gate-dispersive approach holds great promise due to the relaxed requirements for additional on-chip nanostructures, enabling single-shot spin readout in dense qubit arrays and 1D systems whose architecture prevents the implementation of nearby sensor dots. PSB manifests in different ways, depending on the particular readout scheme chosen and can be tuned to implement singlet/triplet or parity readout [Seedhouse et al. PRX Quantum 2 (2021)].

Here, we report on the direct observation of PSB in a gate-dispersively sensed hole double quantum dot [Eggli et al, arXiv, 2303.029339]. We investigate the dependence of the dispersive features on externally applied magnetic fields and theoretically model the qubit system. Our results present a route towards high-fidelity gate-dispersive spin readout.   

Semiconductor qubits on the move

In recent years, planar germanium qubits in Ge/SiGe heterostructures emerged as a compelling platform for quantum computation[1]. Their favourable properties enabled to demonstrate a four-qubit quantum processor [2] and the implementation of scalable control strategies [3]. However, to demonstrate a quantum advantage with semiconductor qubits, larger quantum dot systems need to be developed meeting stringent requirements in device quality and performance.

To this end, we develop an extended system comprising 10 quantum dot qubits, where we have control over the interdot tunnel coupling and the detuning of each of the individual quantum dots. We operate the array at low magnetic fields of tens of mT. We develop a method that can statistically characterize the full system, by shuttling coherently through the array and only requiring to establish readout of a single spin. [4]. Variations in the spin quantization axes give rise to spin rotations, which allow to implement low-power single-qubit gates and to rapidly assess the 10 g-factors and dephasing times. By characterising multiple devices, we can interpret our results in terms of shape and quantum dot confinement, paving the way towards g-factor engineering for better qubit addressability in large architectures.

References:

1. G. Scappucci et al., Nature Rev. Mat. (2021).

2. N.W. Hendrickx et al., Nature 591, 580–585 (2021)

3. F. Borsoi et al., Nature Nano. 1-7 (2023)

4. F. van Riggelen-Doelman, arXiv:2308.02406 (2023)   

Two-qubit logic between distant spins in silicon

Semiconductor spin qubits hold promise for quantum computation due to their long coherence times and potential for scaling. So far, interactions between spin qubits are limited to spins a few hundreds of nanometers apart. A distributed architecture with local registers and long-range couplers will be needed to scale up to millions of qubits. Circuit quantum electrodynamics can provide a pathway to realize interactions between distant spins. Here, we report long-range two-qubit logic using an on-chip superconducting resonator.

We first show universal single-qubit control over two distant spin qubits separated by 250 µm and characterize their coherence times. Next, we couple the two spins via virtual photons in a superconducting resonator, and show two-qubit iSWAP logic between the distant qubits with an interaction strength of 21 MHz. This frequency is consistent with spectroscopic measurements. Furthermore, we demonstrate that the coupling strength as well as the coherence times of the qubits can be tuned by the inter-dot tunnel coupling and the spin-cavity detuning. In future work we intend to implement single-shot readout and improve the spin lifetimes while dispersively coupled to the resonator. These results pave the way for scalable networks of spin qubits on a chip.   

High-fidelity operation of encoded spin qubits on Intel Tunnel Falls

Quantum computers executing complex algorithms with stable logical qubits will require large numbers of highly coherent physical qubits in concert with precise pulse control. Silicon quantum dot based qubits provide both the foundation for high-fidelity qubits, and the ability to leverage the proven scalability and reliability of semiconductor manufacturing over the past decades. In fact, multiple spin-qubit encodings are available as promising candidates to form the cornerstone of scalable quantum computation in Silicon. High throughput characterizations and advanced EUV fabrication on the Intel 300mm fabrication line used for next generation processors have led to the development of the newly released Intel Tunnel Falls quantum chip available to research groups in the community, which lends itself to exploring different qubit encodings on devices with up to 12 quantum dots. The Exchange-Only (EO) encoding is one of them, allowing for complete qubit control with only baseband pulses. Here, we present our first results on high-fidelity spin qubit control based on the EO encoding, as well as continued progress in high-fidelity operation of single-spin qubits.   

Single Electron Router in Silicon Quantum Dot Array

For the realization of large-scale quantum computers based on two-dimensional quantum dot arrays in solid-state devices, technologies for the free transport of electrons on the arrays are essential. In particular, controlling the electron routes in branching paths in two-dimensional quantum dot arrays is necessary. Previously, we developed an electron transport technology using a single-electron pump built on a silicon quantum dot array [1]. Here, we develop a single-electron router [2]. It routes electrons periodically emitted from the single-electron pump to the right or left in a branching path of electrons. We showed that accurate routing operation at 100 MHz can be achieved by applying a sinusoidal signal synchronized with the single-electron pump to the switching gate pair in the branching path and to the assist gate in front of it and adjusting their phases. The behavior of this single-electron router is analyzed by a model based on the Wigner representation in energy-time space. We show this model can fit experimental results and estimate the minimum error. Our silicon single-electron router is valuable not only for quantum computing but also for stable quantum metrology with single-electron pumps and as a component of electron quantum optics experiments.

References

[1] T. Utsugi et al., Jpn. J. Appl. Phys. 62, 1020 (2023)

[2] arXiv:2308.05271   

Oral: Numerical modeling of decoherence of entangled spin qubits during shuttling

Creation of a coherent link between quantum registers is among most critical problems for many quantum information processing architectures, including the prospective large-scale semiconductor-based quantum computing devices. A promising technique, based on shuttling the spin qubits between different quantum dot registers, has attracted much attention recently, and has already been demonstrated on several systems [1-4]. 

However, realistic modeling of decoherence of several entangled qubits during the shuttling process remains an outstanding problem, since the noise affecting this system explicitly includes both time and space correlations, and generally cannot be reduced to a standard mathematical model of random process. We propose a new approach for describing the shuttled entangled qubits, employing the concept of random Ornstein-Uhlenbeck sheet, a Gaussian stochastic field explicitly denependent on time and space coordinates. We provide several numerical approaches to modeling decoherence under such noise, justify their validity, and assess their performance. We present the results of representative simulations for shuttling of few-qubit entangled states, and discuss extension of these methods to other decoherence sources, such as the noise hot spots and 1/f noise. 

[1] A. Noiri et al., Nature Comm. 13, 5740 (2022)

[2] F. van Riggelen-Doelman et al., arXiv:2308.02406 (2023)

[3] I. Seidler et al., npj Qu. Inf. 8, 100 (2022)

[4] A. M. J. Zwerver et al., PRX Quantum 4, 030303 (2023)   

Resistive silicon spin qubit interconnects as platforms for mesoscopic physics

Large-scale quantum computers built from silicon spin qubits will require medium- and long-range interconnects. We consider an interconnect between two dots consisting of a quasi-1D channel created by a resistive topgate. In the absence of interactions, a single electron moves independently with its spin through the channel. If the channel hosts a finite density of electrons, however, Coulomb interactions between those electrons can dramatically change the nature of the ground state, e.g., to form a Luttinger-liquid state. We investigate how this physics is changed by the details of such a channel in an Si/SiGe quantum well. We consider the effect of phenomenological disorder, valley-splitting disorder due to Ge alloy disorder at the Si/SiGe interface, screening by the resistive topgate. We also consider coupling to quantum dots on each end of the channel, as opposed to the non-interacting (or Fermi liquid) leads commonly studied in mesoscopic physics.   

Shuttling electrons between Si/SiGe quantum dots using a resistive topgate

With recent demonstrations of operation fidelity exceeding 99% in few qubit Si/SiGe quantum dot processors [1,2], efforts are turning to scaling up systems. A key element to large-scale quantum computers is the spin-shuttle, which can couple spatially separated qubits. Recent results based on interdigitated gates akin to a CCD have demonstrated proof-of-principle shuttling in Si/SiGe with high fidelity [3]. In this talk we will propose an alternative architecture based on resistive top gates defined in a single layer of lithography. We will discuss the shuttle approach and report preliminary results for double quantum dot devices coupled via a resistive shuttle.

[1] Mills et al., Sci. Adv. 8, eabn5130 (2022)

[2] Philips, S. G. J. et al., Nature. 609, 919 (2022)

[3] Xue et al., arXiv.2306.16375   

Valley splitting and spin shuttling in Si/SiGe heterostructures

Coherent coupling between distant qubits is needed for any scalable quantum computing scheme. In quantum dot systems, one proposal to achieve long distance coupling is the coherent transfer of electron spins across a heterostructure, called spin shuttling [1,2,3,4]. In this talk, we examine how the valley degree of freedom poses challenges for spin shuttling in Si/SiGe heterostructures. We show that, for most known devices, valley splitting is dominated by alloy disorder. In such devices, pockets of low valley splitting are distributed throughout the heterostructure. While such pockets may be small in size, an electron is likely to encounter one on a long enough shuttling path. At these spots, inter-valley tunneling leads to dephasing of the spin wavefunction, substantially deteriorating the shuttling fidelity. We demonstrate how heterostructure modifications and fine-tuning strategies can be used to mitigate this problem. In particular, we consider varying the heterostructure composition, modulating the vertical electric field, tuning the dot position within the shuttling path, modulating the shuttling velocity, and modulating the orbital energy of the dot. We show that combinations of these strategies can improve shuttling fidelity by several orders of magnitude, putting shuttling fidelities within the error correction threshold.

[1] Xue et al., arXiv:2306.16375 (2023)

[2] Langrock and Krzywda et al., PRX Quantum 4 (2023)

[3] Künne and Willmes et al., arXiv:2306.16348 (2023)

[4] Struck et al., arXiv:2307.04897 (2023)   

Coherent electron spin shuttling in an industrially-manufactured Si/SiGe triple quantum dot device

Electron spin-based qubits in silicon quantum dots (QDs) are a promising platform for quantum information processing thanks to their long coherence times and the availability of mature silicon microelectronics fabrication techniques. An open challenge in the field is to determine the best mechanism for transporting quantum information over long length scales, a feature that will likely be needed in future implementations to facilitate coherent links between distant qubits. One potential path to achieving a stable long-ranged interaction is to use coherent spin shuttling through an array of quantum dots. Here, we demonstrate coherent shuttling of an electron spin in an isotopically enriched, industrially-manufactured Si/SiGe triple-QD device. We operate an encoded two-electron singlet-triplet qubit to characterize the transferability of an electron between neighboring QD sites while maintaining the coherence of the entangled two-electron spin state. The results are encouraging steps towards achieving shuttling-based long-ranged entanglement in semiconductor spin qubit devices.