Nanoscale Buckling of Ultrathin Low-k Dielectric Lines during Hard-Mask Patterning

Abstract

Commonly known in macroscale mechanics, buckling phenomena are now also encountered in the nanoscale world as revealed in today's cutting-edge fabrication of microelectronics. The description of nanoscale buckling requires precise dimensional and elastic moduli measurements, as well as a thorough understanding of the relationships between stresses in the system and the ensuing morphologies. Here, we analyze quantitatively the buckling mechanics of organosilicate fins that are capped with hard masks in the process of lithographic formation of deep interconnects. We propose an analytical model that quantitatively describes the morphologies of the buckled fins generated by residual stresses in the hard mask. Using measurements of mechanical properties and geometric characteristics, we have verified the predictions of the analytical model for structures with various degrees of buckling, thus putting forth a framework for guiding the design of future nanoscale interconnect architectures.

Stochastic behavior of nanoscale dielectric wall buckling

“…The ILD samples were used in a similar study 16 and correspond to samples C, A, and B therein. Each sample had different dimensions or materials and different buckling characteristics.…”

“…1(a)) consisted of five layers. 16 From bottom to top: Layer 1 was a 300 mm diameter (100) Si wafer. Layer 2, 100 nm SiO 2 , was deposited by chemical vapor deposition (CVD) directly on the Si wafer.…”

“…1, as well as wide regions (~300 nm) of unetched area that enabled measurement of the TiN modulus of each sample, Table I. 16 The trench walls areas exhibited buckling and were the focus of study. Etching depth (wall-height) is one of the main parameters that determines buckling: 13,15,16 Sample S2 (Figs.…”

“…13,15,16 Deterministic buckling models inevitably predict the formation of perfectly periodic deflections; however, measured buckling deflections are clearly pseudo-periodic, exhibiting both phase decorrelation and amplitude fluctuations (Figs. 1(b)–1(d)).…”

“…Recently, it was hypothesized that large deviations from deterministic models of trench-wall buckling might result from stochastic effects. 16 The importance of incorporating randomness in models used for prediction or analysis will vary from system to system and with the quantitative inferences to be drawn, but the importance will not be known a priori .…”