• Innovation Thoughts
  • Blog Post

LOW-K INTERMETAL DIELECTRICS

Publish Date

02 NOV 2021

Author

Robert Ridgeway

Overview

Low-k Intermetal Dielectrics (ILD) have evolved from dense, single precursor based oxides to porous, structural and porogen based Low-k films. This blog discusses about the needs, challenges and our new precursor innovation in developing a successful ILD process for next-generation devices.

Low-k Intermetal Dielectrics (ILD) have been used in back-end integration for well over a decade. These films have evolved from dense, single precursor based carbon doped oxides to porous, structure former and porogen based carbon doped oxides, or porous Low-k films. The addition of porosity was necessary to reduce k between 2.5–2.6, which is the typical porous Low-k films currently in HVM.                                                                                 

As film pitches have shrunk with technology node advancements, the demands put upon Low-k films have evolved. Depending on an Integrated Device Manufacturer’s integration scheme, mechanical properties can become very critical with higher elastic modulus and hardness is required to reduce mechanical defects in very thin ILD layers. Another consideration for thin ILD layers is Plasma Induced Damage (PID) of the ILD film through side wall exposure to the etching and ashing steps during integration, which can impact the electrical performance of the films. PID occurs when reactive radicals strip carbon from the film leaving reactive sites that ultimately get converted to hydroxyl groups and increase the dielectric constant.

In current Low-k process chemistries, these two principals, mechanical strength and PID with subsequent k loss, a trend in opposite directions. For a given k value, increasing carbon content of Low-k film reduces the depth of PID by reducing the penetration depth of reactive radicals formed in the plasma. However, higher carbon content films have a lower silicon-oxygen bond density resulting in lower mechanical strength. Conversely, reducing the carbon content to increase mechanical strength requires a higher degree of porosity to maintain the dielectric constant, thereby increasing the depth of the plasma damage.

Balancing these needs can be critical to developing a successful ILD process for next-generation devices. We are dveloping solutions that can bridge the gap between strong mechanical properties and reduced PID. By controlling the type and amount of carbon deposited into the film, a balance between higher mechanical strength and reduced PID can be achieved. In addition to working with OEMs, we can work with individual IDMs to tailor an ILD solution to meet their specific integration needs. Our applications lab is equipped with a state-of-the-art 300mm PECVD process tool that is used to develop Best Known Methods (BKM) that can be readily transferred to HVM process tools.

We have designed new structure forming precursors for use with existing porogens, or used alone for very high mechanical strength films, to help meet the evolving film requirements of advanced technology nodes at 10 nm, 7 nm, 5 nm and beyond. A fundamental understanding of how precursor structure impacts film properties in a PECVD process combined with a statistical Design of Experiments (DOE) approach is used to map process conditions such as plasma power, precursor and oxidant flow rates, temperature and pressure. This allows prediction of process parameters required for a precursor to obtaining specific film properties. State-of-the-art metrology for measuring film properties including thickness, refractive index, porosity, composition, chemical bonding, dielectric constant and mechanical properties are used.  These metrology tools have been benchmarked against external metrology tools including those at OEMs, IDMs and research institutes.

We welcome the opportunity to work with you to discuss and help develop your specific ILD film requirements for next-generation devices. Our extensive experience combined with a robust process database can provide a starting point to develop a tailored solution to meet your ILD needs for single precursor dense or porous Low-k ILD.                                             

References:                                                                                                                  O’Neil M. L. et al. Optimized Materials Properties for Organosillicate Glasses Produced from Plasma Enhanced Chemical Vapor Deposition, MRS Proceedings, Vol 766, Cambridge Univ Press 2003                                                                                                                  Ridgeway R. G. et al. Development of Porous Low-k Precursor to Provide Enhanced Mechanical Properties without Sacrificing Carbon Content, Proceedings Semicon Taiwan, Sep 2016