Data-based modeling for estimating etch rate in plasma etching process

Madeline A. Walters^1,2^, Zhaoyan Fan^2^, and Burak Sencer^2^

^1^HP Inc.  
^2^School of Mechanical, Industrial, and Manufacturing Engineering, Oregon State University

 

Reactive ion etch tools are widely used in modern semiconductor manufacturing. These plasma-based tools utilize a complex combination of physical and chemical effects to selectively remove nanometer-scale metal and dielectric layers. The input variables of an etch tool, such as gas flowrates, RF power, chamber pressure, etch time, etc, co-determine the process conditions and consequentially, the etch output quantities such as etch rate and uniformity.  Due to the complexity of the etch process, conventional input variable selection relies on the human operators’ experience and involve a trial period to search for the proper combinations leading to qualified products. The continued demand for decreased microelectronic feature sizes and tight process control specifications, as well as the monetary and time expenses of process characterization, make predictive modeling a compelling option for plasma etch recipe development. The objective of a plasma etch model is to give relevant information about etch output quantities as a function of tool input variables.

 

Physics-based models for plasma etch process have yet to be completely developed. One major reason is the difficulty involved in deriving accurate velocity distribution functions or equations of state for all species present in a given plasma and tool configuration, which makes it challenging to model charged particle motion for both feature scale and reactor scale models. A second reason is that reactive ion etching involves tens to hundreds of simultaneous chemical reactions occurring on the surface of the etched substrate. Reaction mechanisms and rates are often unknown. Widely varying dimensions (nanometer-scale or smaller chemical interactions, micron-scale feature sizes, meter-scale process chambers) and timescales (femtoseconds for chemical reactions, minutes for multi-step etch processes) further complicate the physical model and add computational expense.  Additionally, the relationship between physical and chemical plasma characteristic and measurable etch outputs is often unclear. For example, knowledge about electron velocity distribution functions and chemical kinetics does not provide quantitative information about etch rate. 

 

Alternatively, this paper presents a data-based approach to model the plasma etch process. The work seeks to use an Artificial Neural Network (ANN) model to build a correlation between the controlled tool parameters and etch rate and uniformity for a blanket 1100 Å WSiN thin film using Cl~2~ and BCl~3~ chemistry. Experimental data was collected using a Lam 9600 PTX plasma metal etch chamber in an industrial cleanroom. The WSiN film was deposited over 3000 Å TEOS to ensure adhesion, with an 8-inch bare silicon wafer as the base layer. Controlled tool parameters were radio frequency (RF) upper electrode power, RF lower electrode power, Cl~2~ gas flowrate, BCl~3~ gas flowrate, and chamber pressure. The design of experiment method was used to select the combinations of experimental configurations. The ANN model was validated using additional data collected from the plasma etch tool.


Data-Based Modeling for Reactive Ion Etching: Effectiveness of an Artificial Neural Network Model for Estimating Tungsten Silicon Nitride Etch Rate

Division / Sponsors: 
Aerospace, Bioengineering, Applied Mechanics, Design Engineering Division, Advanced Energy Systems, Manufacturing Engineering Division, Noise Control and Acoustics, Nondestructive Evaluation, Diagnosis, and Prognosis (NDPD), Nuclear Engineering Division, Process Industries Division, Fluids Engineering, Heat Transfer, Materials, Microelectromechanical Systems, Safety Engineering and Risk Analysis, Computers & Information in Engineering


Paper Proceedings: https://asmedigitalcollection.asme.org/IMECE/IMECE2020/volume/84478

IMECE2020 - Virtual Conference

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.