NER: Debonding and Fracture of Ultra-Thin Films

摘要沉积在衬底(如硅)上的超薄(2~20 nm)薄膜的脱键和断裂是微电子元器件制造中的一个重要问题。由于在薄膜与衬底的界面处存在分子失配，因此在薄膜沉积过程中会产生内应力。这些应力既可能导致脱粘(薄膜与衬底分离)，也可能导致断裂。传统的应力-变形行为研究涉及大的材料样本。由此产生的对行为的描述假设材料在所有方向上都延伸到“无穷无尽”(可能是50纳米)。但在相界面附近，材料受到来自相邻相的长程分子间力的作用，观察到的材料行为是不同的。本文提出了一种新的将连续介质力学扩展到近分子尺度来处理纳米尺度的断裂和脱粘现象。这一理论采用了界面区域外已知的主体材料行为，并通过引入体力(互作力)来修正界面区域内的主体材料描述，体力(互作力)解释了当部分材料被相邻相取代时分子间力的差异。可以推测，该理论可以提供一种方法，将裂纹尖端的纳米尺度模型结合到宏观断裂问题中，以便在不需要(或最小)经验参数拟合的情况下，从第一原理预测裂纹的萌生和扩展。

AbstractDebonding and fracture of ultra-thin (2 to 20 nm) films deposited on substrates (such as silicon) is an important problem in the manufacture of micro-electronic components. Internal stresses are developed in the thin films during their deposition, because there is a molecular mismatch at the interface between the thin film and its substrate. These stresses can result in both debonding (separation of the thin film from the substrate) and fracture.Traditional studies of stress-deformation behavior involve large material samples. The resulting descriptions of behavior assume that the material extends to ``infinity'' (perhaps 50 nm) in all directions. But within the immediate neighborhood of a phase interface, the material is subjected to long-range intermolecular forces from the adjoining phase, and the observed material behavior is different.Proposed herein is a novel extension of continuum mechanics to the near molecular scale to treat nanoscale fracturing and debonding phenomena. This theory employs known bulk material behavior outside an interfacial region and corrects this bulk material description within the interfacial region through the introduction of a body force (mutual force) that accounts for the difference in intermolecular forces when a portion of the material is replaced by an adjoinging phase. It is conjectured that this theory can provide a means to incorporate nanoscale modeling at a crack tip into macroscale fracture problems in order to predict fracture initiation and propagation from first priciples without (or with minimal) empirical parameter fitting.