Spin-orbit-enabled realization of arbitrary two-qubit gates on moving spins
D. Fernández-Fernández, Y. Matsumoto, L.M.K. Vandersypen, G. Platero, S. Bosco · 2025
Read Spin-orbit-enabled realization of arbitrary two-qubit gates on moving spins by D. Fernández-Fernández, Y. Matsumoto, L.M.K. Vandersypen, G. Platero, S. Bosco (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-08394.txt.
Introduction
Fernández-Fernández, Matsumoto, Vandersypen, Platero, and Bosco wrote this paper in two thousand twenty-five, with given names shown as initials in the byline. It shows how spin-orbit coupling, a foe when spin qubits are shuttled, can be turned to run any two-qubit gate in one move.
Two spin qubits in a double quantum dot are slid toward each other by gate voltages. By setting the slide speed and the waits, and riding the strong spin-orbit force, the pair takes on any sought two-qubit gate at high fidelity. Gates that need many steps on still dots land in one step here. The route cuts control load for scaled quantum machines.
Catalog record
| CREATOR | D. Fernández-Fernández, Y. Matsumoto, L.M.K. Vandersypen, G. Platero, S. Bosco |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 56 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-08394
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Spin-orbit-enabled realization of arbitrary two-qubit gates on moving spins D. Fernández-Fernández, Y. Matsumoto, L.M.K. Vandersypen, G. Platero, S. Bosco Abstract Shuttling spin qubits in systems with large spin-orbit interaction (SOI) can cause errors during motion. However, in this work, we demonstrate that SOI can be harnessed to implement an arbitrary high-fidelity two-qubit (2Q) gate. We consider two spin qubits defined in a semiconductor double quantum dot that are smoothly moved toward each other by gate voltages. We show that an arbitrary high-fidelity 2Q gate can be realized by controlling the shuttling speed and waiting times, and leveraging strong intrinsic or extrinsic SOI. Crucially, performing 2Q operations during qubit transport enables a one-step realization of a wide range of 2Q gates, which often involve several steps when implemented using static dots. Our findings establish a practical route toward direct implementation of any 2Q gate via spin shuttling, significantly reducing control overhead in scalable quantum computing architectures. Introduction.– Quantum computing promises advantages over classical computation Zhong et al. 2020 ; Wu et al. 2021 , but practical algorithms demand large, highly-connected quantum chips Bravyi et al. 2024 ; Goto 2024 . Semiconductor quantum dots (QDs) are leading candidates for scalable architectures Vandersypen and Eriksson 2019 ; György et al. 2022 ; Burkard et al. 2023 , offering long coherence times Hansen et al. 2022 ; Foster et al. 2025 ; de Fuentes et al. 2025 , high-fidelity gates Xue et al. 2022 ; Noiri et al. 2022 , and CMOS compatibility Maurand et al. 2016 ; Zwerver et al. 2022 ; Geyer et al. 2024 ; Huckemann et al. 2025 . A key challenge remains maintaining high fidelities in dense arrays. Modular architectures address this by connecting small, efficient units through quantum links Vandersypen et al. 2017 ; Künne et al. 2024 ; Ginzel et al. 2024 . Recent experiments have shown coherent spin-qubit shuttling across QD arrays Zwerver et al. 2023 ; van Riggelen-Doelman et al. 2024 ; Struck et al. 2024 ; De Smet et al. 2025 ; Foster et al. 2025 , enabling quantum buses for quantum information transfer. A key mechanism for spin-qubit control is the spin-orbit interaction (SOI), allowing all-electrical manipulation via electric dipole spin resonance Nowack et al. 2007 ; Fernández-Fernández et al. 2023 ; John et al. 2024 . However, SOI combined with noise can introduce decoherence and path-dependent dynamics, degrading shuttling fidelity. Recent studies show SOI can enhance shuttling by filtering low-frequency noise Bosco et al. 2024 and enabling qubit control en route Fernández-Fernández et al. 2024 ; Wang et al. 2024 . Figure 1: (a) Schematic of two spin qubits confined in a double QD potential V(x,t) . (b) Conveyor-mode potential V(x,t) , where the interdot distance is d(t) and the shuttling speed is v . (c) The shuttling protocol starts with two spin qubits separated by d_{\mathrm{max}} , and last for a total time \mathcal{T}=2t_{s}+t_{w} . (d) Khaneja-Glaser decomposition of a general two-qubit gate into one-qubit local gates (pink) A_{i} and B_{i} , and a single two-qubit nonlocal gate (green) U(\theta_{x},\theta_{y},\theta_{z}) . (e) Weyl chamber in the space spanned by \theta_{i} . Efficient two-qubit (2Q) gates are essential for scalability. SOI-enabled 2Q gates have been explored in static arrays Geyer et al. 2024 ; Spethmann et al. 2024 ; Qi et al. 2024 ; Bosco and Rimbach-Russ 2024 ; Rimbach-Russ et al. 2024 ; Nguyen et al. 2025 , but implementing them during shuttling, rather than through complex sequences of static operations, can reduce circuit depth and cumulative error. Direct access to 2Q gates in motion suppresses computation overhead. Although conveyor-mode Seidler et al. 2022 ; Langrock et al. 2023 ; Xue et al. 2024 ; Struck et al. 2024 ; De Smet et al. 2025 ; Nagai et al. 2025 spin shuttling has recently been used to implement high-fidelity 2Q gates Matsumoto et al. 2025 , the role of SOI in enabling 2Q gates during motion remains unresolved. In this Letter, we show that large SOI, intrinsic or engineered, enables an arbitrary high-fidelity 2Q gate during spin shuttling. By tuning only the shuttling speed and waiting time of the conveyor mode, we implement a broad class of entangling gates, not only the typical \mathrm{CPHASE} and \mathrm{SWAP} families, but also fermionic simulation Kyriienko and Elfving 2021 ; Nguyen et al. 2024 and Berkeley gates Zhang et al. 2004 , which are challenging to reach in a single step in current spin-based devices. Our general framework includes position-dependent magnetic fields, SOI, and anisotropic g -factors, and applies to electrons and holes in, e.g., Si, Ge, GaAs and InAs, in planar and nanowire geometries. Our results provide a realistic and scalable pathway toward long-range 2Q gates with minimal overhead for quantum simulation, distributed computing, and shuttling-based quantum error correction Siegel et al. 2025 . Model.– We consider two spin qubits confined in a double QD potential V(x,t) moving in a one-dimensional path along the x -axis, see Fig. 1 (a). The Hamiltonian Bosco et al. 2024 \begin{split}H=\sum_{i=1}^{2}&\left[\frac{p_{i}^{2}}{2m^{*}}+\left\{\vec{v}_{\mathrm{SOI}}(x_{i}),p_{i}\right\}\cdot\vec{\sigma}^{(i)}+\frac{\mu_{B}\overline{g}}{2}\vec{B}(x_{i})\cdot\vec{\sigma}^{(i)}\right.\\ &\;\left.\vphantom{\frac{p_{i}^{2}}{2m^{*}}}+V(x_{i},t)\right]+\frac{1}{4\pi\epsilon}\frac{e^{2}}{\absolutevalue{x_1 - x_2}},\end{split} (1) includes kinetic energy, SOI, magnetic field, confinement and Coulomb repulsion, respectively. We introduce the momentum acting on the i -th particle as p_{i}=-i\hbar\partial_{x_{i}} , the effective mass m^{*} , the SOI \vec{v}_{\mathrm{SOI}}(x_{i})=v_{\mathrm{SOI}}(x_{i})\hat{n}_{\mathrm{SOI}}(x_{i}) , with velocity v_{\mathrm{SOI}}=\hbar/m^{*}l_{\mathrm{SOI}} and l_{\mathrm{SOI}} being the SOI length, the Pauli matrices \vec{\sigma}^{(i)} , and the normalized anticommutator \{\cdot,\cdot\} . The Bohr magneton is \mu_{B} , \overline{g} is the effective g -tensor, \vec{B}(x_{i}) is the magnetic field, and \epsilon is the dielectric constant of the medium. For the sake of concreteness, we focus on Si devices with m^{*}=0.1m_{e} , \epsilon=11\epsilon_{0} , and isotropic tensor \overline{g}=g^{*}\delta_{ij}=2 , noting that materials with spacially-varying \overline{g} , e.g., holes in Ge, can be incorporated by rescaling \vec{B}(x) . The conveyor-mode potential is modeled as V(x,t)=\frac{1}{2}m^{*}\omega_{0}^{2}\frac{\left[x^{2}-4d(t)^{2}\right]^{2}}{16d(t)^{2}}, (2) where \omega_{0} is the confinement frequency and d(t) the time-dependent interdot distance. This form ensures constant confinement throughout shuttling, see Fig. 1 (b). The confinement length scale reads l_{0}=\sqrt{\hbar/m^{*}\omega_{0}} . Computing the time-evolution operator of Eq. ( 1 ) is challenging, even with split-operator methods Fillion-Gourdeau et al. 2012 ; Glowinski 2017 ; Choi and Vaníček 2019 ; Roulet and Vaníček 2021 . We thus derive a low-energy Hamiltonian acting on the computational basis \{\ket{\uparrow\uparrow},\ket{\uparrow\downarrow},\ket{\downarrow\uparrow},\ket{\downarrow\downarrow}\} Hetényi et al. 2020 ; Geyer et al. 2024 ; Saez-Mollejo et al. 2025 : \tilde{H}=\frac{1}{2}\left(\vec{\Delta}_{L}\cdot\vec{\sigma}_{L}+\vec{\Delta}_{R}\cdot\vec{\sigma}_{R}\right)+\frac{1}{4}\vec{\sigma}_{L}\cdot\overline{J}\vec{\sigma}_{R}, (3) with \vec{\Delta}_{i}=\mu_{B}\overline{g}_{i}\vec{B}(x_{i}) capturing local Zeeman fields. The kinetic energy, SOI, and Coulomb interaction contribute to the exchange matrix \overline{J} , which can be written as \overline{J}=J_{0}R(\theta,\phi,\alpha) , where R is a rotation matrix around a vector parametrized by polar angle \theta and azimuthal angle \phi , and rotation angle \alpha . These parameters vary with the interdot distance, and thus evolve during shuttling, unlike static 2Q gates. This effective model is accurate for linear-in-momentum SOI Hetényi et al. 2022 . Further numerical details on \overline{J}(t) are given in the Supplemental Material (SM) Sup . The shuttling 2Q gate protocol, as in Matsumoto et al. 2025 , consists of a linear ramp reducing d(t) from d_{\mathrm{max}} to d_{\mathrm{min}} at constant speed v , lasting a time t_{s}=(d_{\mathrm{max}}-d_{\mathrm{min}})/2v . After a waiting time t_{w} , the qubits are separated back at the same speed. The full protocol lasts