QFlow Triangles: Quantum dot triangle plots data for machine learning

Arrays of QDs--interconnected islands of electrons confined in a semiconductor heterostructure with unique properties that allow them to act as artificial atoms--are a leading candidate for use as qubits, the fundamental information carriers in quantum computers. Scaling these systems to large arrays suitable for quantum computations is challenging. As the number of QDs grows, the number of gates needed to control them grows, making the manual tuning process unfeasible. Establishing a stable configuration of electrons in space is a non-trivial task achieved via electrostatic confinement, band-gap engineering, and dynamically adjusted voltages on nearby electrical gates. A key task is to determine a good set of control parameters (gate voltages) to achieve a desired charge configuration--in terms of number, location, and connectivity--for a successful experiment.

One sub-problem within the tuning procedure is determining the voltage placement of gates to allow the formation of isolated one-dimensional (1D) current channels inside the two-dimensional electron gas (2DEG) formed at the intersection of the Si and Si$_{x}$Ge$_{1-x}$ layers in the heterostructure. The 1D current channels are formed by selectively removing 2DEG from certain regions of the device. The formation of the channel manifests as a triangular-sloped region (the so-called triangle plots).

量子点（QDs）阵列——即局限于半导体异质结中的电子互联岛，其拥有可充当人工原子的独特性质——是作为量子计算机核心信息载体量子比特（qubits）的主流候选方案之一。将这类系统拓展至适配量子计算的大规模阵列颇具挑战：随着量子点数量增加，所需控制栅极的数量同步攀升，使得手动调谐过程变得不可行。通过静电约束、带隙工程，以及对邻近电控栅极施加动态调整的电压，方可构建出空间内电子的稳定构型，这并非易事。其中一项核心任务是确定一组最优控制参数（栅极电压），以达成实验所需的电荷构型——涵盖电子数量、空间位置与连通性层面。调谐流程中的一个子问题是确定栅极电压的配置方案，以在异质结中Si与SiₓGe₁₋ₓ层界面处形成的二维电子气（2DEG）内部，构建出孤立的一维（1D）电流通道。此类一维电流通道通过选择性移除器件特定区域内的二维电子气得以形成，其形成过程会表现为三角斜率区域（即所谓的三角图）。