Optical Model Fitting for CMP Endpoint Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Chemical mechanical polishing (CMP) processes face challenges in determining the polishing endpoint due to variations in substrate layer thickness, slurry composition, polishing pad condition, and load, making it difficult to accurately assess when a desired flatness or thickness is achieved, especially on patterned substrates where existing optical monitoring techniques are unreliable.
Innovation Solution
The method involves fitting an optical model to in-situ measured spectra during polishing, accounting for diffraction effects using Rigorous Coupled Waveform Analysis, to determine the polishing endpoint and adjust pressure, by calculating parameters such as layer thickness, line pitch, and material composition, which improves endpoint detection reliability and reduces within-wafer and wafer-to-wafer thickness non-uniformity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing optical monitoring techniques are used to determine polishing endpoint, then the process is simple and fast, but the endpoint determination is unreliable on patterned substrates due to inability to account for diffraction effects
Solution Approach 1:
The patent introduces an optical model as an intermediary between the measured spectrum and the endpoint determination. This model accounts for diffraction effects by incorporating parameters such as line pitch, linewidth, and layer thickness, allowing reliable endpoint detection on patterned substrates without directly measuring these complex physical quantities
Solution Approach 2:
The patent transforms the endpoint determination problem by changing the parameters used in the optical model. By fitting the model to measured spectra and extracting parameters like line pitch and linewidth, the system achieves reliable endpoint detection that accounts for the actual physical dimensions of the patterned substrate features
2Measurement precision
If a simple optical model is used without diffraction effects, then the calculation is fast and simple, but the endpoint determination is inaccurate for patterned substrates
Solution Approach 1:
The patent replaces direct physical measurement of diffraction effects with an optical modeling approach. Instead of mechanically measuring line pitch and linewidth, the system uses spectral fitting to infer these parameters, substituting a computational model for physical measurement and achieving high precision without complex measurement hardware
Solution Approach 2:
The patent creates an optical copy or model of the patterned substrate's physical structure. By replicating the layer stack geometry and optical properties in the model, the system can predict spectral responses and accurately determine endpoint conditions without directly measuring the physical dimensions of the substrate features
3Ease of operation
If polishing endpoint is determined solely as a function of polishing time, then the process is simple to control, but variations in material removal rate cause inaccurate endpoint determination
Solution Approach 1:
The patent implements real-time feedback by continuously monitoring the reflected spectrum during polishing and comparing it against the optical model predictions. This feedback loop allows dynamic adjustment of the polishing process and accurate endpoint detection that compensates for variations in material removal rate, maintaining precision while keeping the control system relatively simple
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the reliability of endpoint detection and reduces thickness non-uniformity by accurately modeling the patterned substrate and diffraction effects, ensuring consistent polishing results.
Implementation Method 1
The fitting includes calculating the output spectrum using diffraction effects of the repeating structure
Data Source
AI summary
A method of controlling a polishing operation includes obtaining a sequence over time of measured spectra with an in-situ optical monitoring system during polishing. For each measured spectrum from the sequence an optical model is fit. The optical model includes dimensions of a repeating structure and the fitting includes calculating a output spectrum using diffraction effects of the repeating structure, and parameters of the optical model include an endpoint parameter and a parameter of the repeating structure. The fitting generates the sequence of fitted endpoint parameter values, and at least one of a polishing endpoint or an adjustment of a pressure to the substrate is determined from the sequence of fitted endpoint parameter values.