Low-Dimensional-Material-Based Integrated Systems for Energy Storage, Conversion and Delivery

能量存储、传输和转换电路的小型化和片上集成是开启从无电池无线传感器到低成本光伏等新一代电子应用的关键。事实上，单是自供电芯片在生物传感和生物医学方面的影响就令人震惊。然而，鉴于前端硅CMOS和后端金属的传统几何比例即将结束，这种电路的小型化仍然是不可能的。这一提议旨在通过使用低维材料(LDM)，如一维(1D)碳纳米管(CNT)、二维(2D)石墨烯和2D过渡金属双硫化物(TMDs)薄片来突破规模收益递减的问题。具有直径&lt；2 nm(碳纳米管)或厚度&lt；1 nm(2D LDM)的LDM可提供尽可能最小的材料尺寸。与传统材料相比，这使LDM具有更好的电学和光学性能，这正是节能电路小型化所需的。具体地说，该计划分两个阶段开发基于LDM的能量存储和转换系统，第一阶段(第1-3年)专注于三个基本技术级挑战的基本物理，而第二阶段(第4-5年)专注于集成电路级别的解决方案。阶段I：第一阶段：第一个技术挑战是将无源设备小型化，以便它们可以集成到芯片上。目前这是不可能的，因为大值电容器(&gt；1 nF)和电感(&gt；100 nH)太大了，必须安装在芯片外。LDM具有克服这一问题的潜力，具有高潜力的纳米结构，包括碳纳米管束/石墨烯螺旋结构和高密度碳纳米管森林/双层石墨烯/多层石墨烯-h-BN-石墨烯结构。然而，为了使LDM发挥其潜力，必须解决诸如LDM/金属接触不良导致的质量因数低得令人无法接受的问题。第二个技术挑战是逐步改变有源设备的大小和性能。随着传统的硅芯片尺寸越来越多地受到漏和寄生的限制，2DTMD由于其薄的体层和合理的禁带宽度而具有很高的潜力。然而，用于快速电路模拟和芯片设计的紧凑型模型尚不存在，许多关键的TMD特性(如频率相关噪声特性和传输线性度)仍然未知。第三个技术挑战是解决显著的寄生电容，这些寄生电容可能严重降低高度小型化电路的布图后性能。这可以通过采用“寄生感知”电路设计技术来实现，这种技术可以缩小布图前设计和布图后实现之间的差距，并有可能将寄生的“废物”转化为有用的无源元件。阶段II：一旦解决了实用的基于LDM的无源和有源器件的潜在技术挑战，第二阶段将整合学习并产生电路级模型和设计。这将包括器件结构和电路拓扑的共同优化，以确保充分利用LDM的独特特性，并针对目标应用优化器件结构。最终的结果将是物理原型，证明LDMS实际上可以将能量存储和转换电路微型化，适用于各种应用。此外，通过集成一个全面的主动和被动技术模型和联合仿真工具库，该计划将为未来的特定应用研究提供基础。

Miniaturization and on-chip integration of energy storage, delivery and conversion circuitry is the key to unlocking a new generation of electronic applications, from battery-less wireless sensors to low-cost photovoltaics. Indeed, the bio-sensing and bio-medical implications of self-powered chips alone are staggering. However, given that conventional geometric scaling of front-end Si CMOS and back-end metals are coming to an end, miniaturization of such circuitry remains will be impossible. This proposal aims to breakthrough diminishing scaling returns by using low-dimensional materials (LDMs) such as 1-dimensional (1D) carbon nanotubes (CNTs), 2-dimensional (2D) graphene and 2D transition metal di-chalcogenides (TMDs) sheets. With diameter <2nm (CNTs) or thickness <1nm (2D LDMs), LDMs offer the smallest possible material dimensions. This gives LDMs superior electrical and optical properties versus traditional materials, which is exactly what is needed for miniaturization of energy-efficient circuits. Specifically, this proposal develops LDM-based energy storage and conversion systems across two phases, with Phase I (Years 1-3) focused on the underlying physics of three fundamental technology-level challenges and Phase 2 (Years 4-5) focused on integrated circuit-level solutions.PHASE I:The first technology challenge is to miniaturize passive devices so they can be integrated on-chip. This is not currently possible because large value capacitors (>1 nF) and inductors (>100 nH) are so large that they must be located off-chip. LDMs have potential to overcome this, with high-potential nanostructures including CNT bundles/grapheme spirals and high-density CNT forests/bilayer grapheme/multi-layer graphene-h-BN-graphene structures. However, for LDMs to achieve their potential, problems like the unacceptably low quality factor caused by poor LDM/metal contact must be addressed. The second technology challenge is to step-change the size and performance of active devices. With conventional Si CMOS scaling increasingly limited by leakage and parasitics, 2D TMDs are high-potential due to their thin body and reasonable bandgap. However, compact models that capture sufficient device information to be used for rapid circuit simulation and chip design do not yet exist, and many key TMD properties are still unknown (e.g. frequency-dependent noise characteristics and transport linearity).The third technology challenge is to address the significant parasitic capacitances that can severely degrade post-layout performance of highly miniaturized circuits. This can be done by adopting “parasitic-aware” circuit design techniques that close the gap between pre-layout design and post-layout implementation and potentially convert parasitic “waste” into useful passive components.PHASE II:Once the underlying technology challenges of practical LDM-based passive and active devices are addressed, Phase II will integrate the learning and produce circuit-level models and designs. This will include co-optimization of device structures and circuit topologies to ensure that the unique properties of LDMs are fully utilized and device structures are optimized for target applications. The end result will be physical prototypes that prove LDMs can practically miniaturize energy storage and conversion circuitry for a myriad of applications. Furthermore, by integrating a comprehensive library of active and passive technology models and co-simulation tools, this program will be provide a foundation that future application-specific research can build from.