Quantum States Transferred with Improved Accuracy and Speed

Scientists are continually seeking methods to optimise control over complex quantum systems, and Christian Ventura-Meinersen, Edmondo Valvo, Stefano Bosco, and Maximilian Rimbach-Russ, all from QuTech and the Kavli Institute of Nanoscience at Delft University of Technology, have developed a novel geometric framework for achieving high-fidelity state transfer in multi-level quantum systems. Their research introduces a technique that smoothly interpolates between adiabatic and diabatic dynamics, minimising unwanted excitations and maximising desired transitions, even within intricate multi-level structures. Significantly, this optimisation simplifies to solving a first-order ordinary differential equation for single-parameter pulse control, offering a computationally efficient pathway to advance quantum information processing, as demonstrated through examples of state initialisation and transfer.

Scientists have devised a new technique for manipulating quantum systems with greater precision and speed. The method allows for reliable transfer of information between quantum states, even when operating beyond previously established limits, promising to simplify the complex task of building practical quantum technologies. Researchers have developed a new geometric framework for controlling multi-level quantum systems, achieving high-fidelity state transfer even beyond the limitations of adiabatic processes. This work addresses a longstanding challenge in quantum information processing, efficiently navigating complex quantum states through dense energy spectra common in platforms like superconducting circuits and trapped ions. The research introduces a method that smoothly blends adiabatic and diabatic dynamics, minimising unwanted transitions and maximising the probability of reaching a desired quantum state. Optimising the control pulse for a single parameter simplifies to solving a first-order ordinary differential equation, greatly easing experimental implementation. This approach offers enhanced experimental flexibility, allowing researchers to tailor pulse smoothness to match the capabilities of their specific hardware by adjusting the underlying Riemannian geometry. This adaptability is particularly valuable as quantum systems grow in complexity, demanding control methods that are both precise and practical. The resulting diabatic-adiabatic (di-ad) protocols represent a system-agnostic approach to quantum state control, applicable to a wide range of quantum systems. Initial demonstrations focus on spin-based quantum information processing, specifically quantum state initialisation and qubit state transfer, showcasing the versatility of the new framework. At the heart of this work lies a geometric formulation using the quantum metric tensor, a tool for describing the geometry of quantum states. By framing the problem as geodesic evolution on the space of experimentally accessible parameters, the team found a way to minimise state overlap and optimise transitions. The quantum metric depends on both the overlap of gradients of the system’s Hamiltonian and the energy differences between quantum states. Purely adiabatic transfer, while effective, struggles with rapid transitions. To overcome this, the researchers introduced a generalised quantum metric tensor, the di-ad tensor, which interpolates between adiabatic and diabatic regimes. This di-ad tensor incorporates a transition matrix allowing selective operation between adiabatic and diabatic dynamics. The parameters governing this tensor determine the degree of adiabaticity and diabaticity, offering a continuous pathway for tuning the control pulse. By locally maximising components of this tensor, the team achieved diabatic state transfer, even when initial and final states are not adiabatically connected. A simplified model using the Landau-Zener model, describing transitions at avoided crossings, demonstrated the effectiveness of this approach, showing coherent state transfer with significantly reduced infidelity. Diabatic-adiabatic interpolation streamlines multi-level quantum state control A geometric framework underpinned this work, designed for efficient optimisation of few-parameter pulses in multi-level quantum systems. This approach enables high-fidelity state transfer, exceeding the limitations of purely adiabatic dynamics. The research team focused on interpolating between adiabatic and diabatic dynamics to minimise unwanted excitations and maximise desired transitions within multi-level structures. For single-parameter pulse control, the optimisation process was elegantly reduced to solving a first-order ordinary differential equation, simplifying the computational burden considerably. The versatility of these diabatic-adiabatic protocols was demonstrated through applications in spin-based quantum information processing, specifically state initialisation and qubit state transfer. The core of the methodology involves framing quantum information transfer as a dynamical problem, mapping an initial state to a target state through engineered time evolution. This was achieved using the quantum metric tensor, a mathematical tool describing the geometry of quantum states and their evolution under parameter changes. By defining the task as geodesic evolution on the space of experimentally accessible parameters, the team could precisely control the system’s trajectory. Once established, the infinitesimal state overlap of the evolved state was expressed using the quantum metric tensor, allowing for a detailed analysis of state changes. In the adiabatic limit, this tensor was directly related to energy fluctuations, providing a clear connection between theoretical calculations and measurable quantities.

The team calculated gμν, representing the metric tensor, using the Hamiltonian, its eigenvalues, and eigenvectors, offering a convenient way to quantify the system’s response to control parameters. The method extends beyond the adiabatic limit, allowing for flexible pulse shaping tailored to experimental hardware constraints. At the heart of this work lies a continuous family of Riemannian geometries, defining the diabatic-adiabatic protocols. These geometries permit adjustment of pulse smoothness, accommodating the bandwidth limitations of real-world experimental setups. This adaptability represents a significant advantage over previous methods, which often require approximations or constraints on the energy spectrum. The research provides a system-agnostic approach to quantum state control, applicable to diverse multi-level quantum systems. High-fidelity spin qubit initialisation using optimised diabatic-adiabatic and adiabatic charge transfer Initialising a spin qubit via spin-to-charge conversion achieved fidelities exceeding 99 percent, observed alongside short simulation times, demonstrating the efficiency of the developed diabatic-adiabatic protocols. Specifically, the research focused on transferring charge between quantum dots, starting from a state with two charges in one dot and moving to a state with one charge in each dot. This process is susceptible to unwanted spin-flip events due to a finite spin-flipping term, denoted as ∆EX, which causes a small anticrossing between energy states. Two distinct initialisation schemes were explored: adiabatic evolution, transitioning from |↑↓, ·⟩ to |↓, ↓⟩, and diabatic-adiabatic evolution, transitioning from |↑↓, ·⟩ to |↑, ↓⟩. For the adiabatic case, optimisation involved sweeping adiabatic parameters (α, β) to suppress excitations to any of the four possible excited states. Conversely, the diabatic-adiabatic approach maintained fixed adiabatic parameters at (α, β) = (2, 2) while varying diabatic parameters. Remarkably, both methods attained fidelities above 99 percent, indicating a high degree of control over the quantum state. Further enhancements to the diabatic-adiabatic protocol were achieved through machine learning sampling, reducing the number of simulations needed to reach the desired parameter configuration. Beyond initialisation, the study examined state transfer during shuttling, a technique for physically moving qubits on a chip. Simulations revealed the ability to excite the ground state to the first excited state while minimising leakage to undesired states. Analysis using the Horodecki identity showed that the average gate fidelity, Fgate, for the computational subspace varied with pulse time ∆L tf, dependent on the chosen parameters. These results demonstrate the potential for on-the-go operations and flexible qubit configurations in quantum circuits. Geometric control simplifies qubit manipulation and scaling potential Scientists have devised a new technique for controlling quantum systems with remarkable precision. Rather than painstakingly crafting individual pulses for each quantum operation, this work introduces a geometric method that streamlines the optimisation process. It’s a subtle but important shift, moving away from bespoke solutions towards a more general, adaptable approach to manipulating qubits. Building practical quantum computers demands more than faster operations. For years, the challenge has been scaling up these systems while maintaining control, a task complicated by the inherent fragility of quantum states. This geometric framework offers a potential pathway to address this, because it reduces the number of parameters needing fine-tuning. Once established, the method appears to hold across different qubit types and system architectures, suggesting broad applicability. Extending this to complex, multi-qubit interactions will undoubtedly present new hurdles. Automated pulse design is essential for any quantum device intended for real-world use, and this method could accelerate the development of quantum sensors and materials. The reliance on specific initial conditions and the potential for unforeseen interactions in larger systems remain open questions. We can anticipate further exploration of geometric control methods, perhaps combined with machine learning techniques to adapt to the unique characteristics of individual quantum devices. Ultimately, the goal is not just to control qubits, but to build systems that can correct their own errors and operate reliably for extended periods. 👉 More information 🗞 Multi-level spectral navigation with geometric diabatic-adiabatic control 🧠 ArXiv: https://arxiv.org/abs/2602.14756 Tags: