Collaborative Research: DMREF: Deep learning guided twistronics for self-assembled quantum optoelectronics

非技术描述：原子薄的二维（2D）材料可以承载其块状材料中没有的有趣量子特性。此外，通过控制相邻层之间的扭转角来堆叠二维材料，为获得具有前所未有性能的新型量子材料提供了一种通用的方法。这种“双旋电子”材料可以应用于电子学、光子学和量子信息科学与技术。然而，在新的自由度下，材料的设计参数空间变得非常大，这对材料的可预测性设计和精确制造提出了重大挑战，以实现这种独特的性能。在这个DMREF项目中，来自宾夕法尼亚大学、威斯康星大学麦迪逊分校和东北大学的合作团队将使用计算机辅助深度学习模型和理论工具来预测在特定状态下制备的设计师双涡旋材料，并指导独特的自组装晶体生长来设计不同二维材料的扭曲角度。该团队将执行性能测量来表征这些系统，并将想法扩展到量子光子学，以组装芯片上的设备。合成、表征和器件测量的结果将反馈到理论模型中，以建立自一致和紧密集成的研究，进一步发现具有精确控制响应的新型设计双创材料，这可以为量子材料研究提供新的范例，并应用于计算、通信、成像和传感。跨学科研究活动将与教育和推广活动相结合，让来自不同背景的各级学生参与合作研究项目，重点是量子材料和光子学。技术描述：现代量子材料通常是通过在原子和晶格长度尺度上结合强自旋轨道耦合的工程对称性来设计的。在具有手性对称性的二维（2D）材料中，通过层间耦合等多体相互作用，控制层间扭转角提供了一种有前途的策略来实现新的量子特性，如平带、拓扑相和大非线性光学响应。然而，两大挑战阻碍了“双旋电子”材料的进展：1)系统自由度的急剧增加使得预测材料成分、晶体相和实现特定量子相所需的层间扭曲变得非常困难；2)目前的材料制造方法包括剥离和重组二维材料层，并手动控制层间扭转角，这是一个费力的过程，产量低。在这个DMREF项目中，一个高度跨学科的团队将通过基于深度学习的材料和超材料的对称性和拓扑工程来打破设计涡旋材料的基本限制。从量子范式出发，二维材料的原子尺度对称性和拓扑结构将针对目标手性响应进行优化。在理论指导下，多层扭曲二维材料将通过合理控制层间扭曲角、成分和晶体相来实现新颖和可预测的量子特性。将产生新的知识，使具有高预测能力的量子扭转电子材料的合理设计能够展示新的手性光电响应，这也将扩展到量子光子系统。这些进步可以实现下一代电子和光学器件，如片上相干手性发射器，纠缠光子发射和具有精确控制响应的检测。这个跨学科项目将为研究生、本科生和K-12学生提供一个很好的教育机会，让他们了解几何、晶体结构和量子物理的重要概念，重点是增加代表性不足群体的参与。该奖项的资金来自数学和物理科学（MPS）材料研究（DMR）和化学（CHE）部门，通过设计材料来革新和工程我们的未来（DMREF）计划。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Non-technical Description: Atomically thin two-dimensional (2D) materials can host intriguing quantum properties not found in their bulk counterparts. Furthermore, stacking 2D materials with control over the twist angles between adjacent layers provides a versatile way to obtain novel quantum materials with unprecedented properties. Such “twistronic” materials can have applications in electronics, photonics and quantum information science and technologies. However, with the new degrees of freedom, the materials design parameter space becomes exceedingly large, posing a significant challenge to predictably design and precisely make materials to enable such unique properties. In this DMREF project, the collaborative team from University of Pennsylvania, University of Wisconsin-Madison, and Northeastern University will use computer aided deep learning models and theoretical tools to predict designer twistronic materials prepared in specific states and guide the unique self-assembled crystal growth to engineer twist angles in different 2D materials. The team will perform property measurements to characterize these systems and also extend the ideas to quantum photonics to assemble on-chip devices. Results from synthesis, characterization and device measurements will be fed back to the theoretical models for establishing a self-consistent and tightly integrated research for further discovery of new designer twistronic materials with precisely controlled responses that can enable a new paradigm for quantum materials research with applications in computing, communications, imaging and sensing. Interdisciplinary research activities will be integrated with educational and outreach initiatives by involving students at all levels from diverse backgrounds in the collaborative research project with emphasis on quantum materials and photonics. Technical Description: Modern quantum materials are typically designed by engineering symmetries combined with strong spin-orbit coupling at the atomic and lattice length scales. In two-dimensional (2D) materials with chiral symmetry complemented by many-body interactions such as interlayer coupling, controlling the interlayer twist angle offers a promising strategy to achieve novel quantum properties such as flat bands, topological phases, and large nonlinear optical responses. However, two major challenges impede the progress in “twistronic” materials: 1) the dramatic increase in the degrees of freedom of the systems makes it prohibitively difficult to predict the material compositions, crystal phases and interlayer twists needed to achieve a particular quantum phase; and 2) the current material fabrication method consisting of exfoliating and reassembling 2D material layers with manual control over the interlayer twist angles is a laborious process with low yields. In this DMREF project, a highly interdisciplinary team will break the fundamental limitation of designing twistronic materials via deep learning-based symmetry and topological engineering of materials and metamaterials. Starting from a quantum paradigm, the atomic scale symmetry and topology in 2D materials will be optimized for targeted chiral responses. Guided by theory, multilayer twisted 2D materials will be synthesized with rational control over interlayer twist angles, compositions, and crystal phases to realize novel and predictable quantum properties. New knowledge will be generated to enable the rational design of quantum twistronic materials with highly predictive power to demonstrate novel chiral optoelectronic responses, which will also be extended to quantum photonic systems. These advances can enable the next generation of electronics and optical devices such as on-chip coherent chiral emitters, entangled photon emission and detection with precisely controlled responses. The interdisciplinary project will provide an excellent educational opportunity for training graduate, undergraduate and K-12 students on the important concepts of geometry, crystal structures and quantum physics with an emphasis on increasing the participation of underrepresented groups. Funding for the award is from the Mathematical and Physical Sciences (MPS) Divisions of Materials Research (DMR) and Chemistry (CHE) through the Designing Materials to Revolutionize and Engineer our Future (DMREF) program.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.