Big-data for nano-electronics

对高密度、集成电子产品的需求已成为现代技术的一个决定性特征。在其极限下，纳米技术可以实现低成本和高度可扩展的传感器、计算元件和照明。工业上的好处是显而易见的--尤其是自下而上的制造允许以低成本实现高水平的功能和巨大的生产规模。随着这种生产技术从实验室走向工业，成品率、异质性和功能参数扩散等问题已经成为先进纳米材料功效建立的关键方面。到目前为止，还没有框架来研究功能纳米电子学中的不均质性。我将把高尺度测量与尖端数据技术相结合，为功能性纳米技术开发建立黄金标准方法，使工业能够被接受。这将建立在我最近演示的实验方法的基础上，包括纳米材料的机器视觉识别和自动化电子和光学光谱分析，以及快速和独立于技术的单一纳米粒子重新识别的计算方法。我将实施分析，利用现有的人口研究方法，如线性和多变量相关；该项目的一个具体目标是转化来自不同领域的先进技术，包括天体物理学和健康研究，特别是在适当的情况下应用贝叶斯分析进行模型识别和增强智能(包括机器学习方法)。开发这些方法是为了研究功能纳米材料的前沿挑战；首先是开发用于芯片间通信的激光，以及生产用于单粒子纳米技术表征的与工业相关的能力。通过将这一方法与项目合作伙伴的挑选和放置能力结合在一起，该项目将能够展示基于新型纳米技术的极低产量但具有变革性的设备，用于传感、电信或量子设备。

Demand for high density, integrated electronics has become a defining feature of modern technology. At its ultimate limit, nanotechnology can enable low-cost and highly scalable sensors, computing elements, and lighting. The industrial benefits are clear - in particular bottom-up fabrication allows for high-level functionality and huge production scale at low cost. As this production technique emerges from the laboratory and into industry, issues such as yield, heterogeneity, and functional parameter spread have emerged as a critical aspect for efficacy to be established in advanced nanomaterials.To date, no framework exists for studying inhomogeneity in functional nano-electronics. I will combine highly-scaled measurements with cutting-edge data techniques to establish a gold-standard methodology for functional nanotechnology development, enabling industrial take-up. This will build on experimental approaches that I have recently demonstrated, including machine-vision identification of nanomaterials and automated electronic and optical spectroscopy, alongside computational approaches for rapid and technique-independent re-identification of single nanoparticles. I will implement analytics which draw on existing population-study methods such as linear and multivariate correlation; a specific goal of this project is to translate advanced techniques from diverse fields including astrophysics and health research, and in particular apply Bayesian analysis for model identification and augmented intelligence (including machine learning methods) where appropriate. These methodologies will be developed to study cutting edge challenges in functional nanomaterials; starting with the development of lasers for chip-to-chip communication, and the production of an industrially relevant capability for single-particle nanotechnology characterisation. By bringing this methodology together with pick-and-place capability through project partners, this project will enable demonstration of extremely low-yield yet transformative devices based on novel nanotechnology, for sensing, telecommunication or quantum devices.