Development of OES-based Non-Maxwellian EEDF Diagnostic Method for Narrow-Gap VHF-CCP Heating Characterization

To achieve high-density plasma conditions required for advanced semiconductor processes, narrow gap VHF-CCP are widely utilized. In such systems, non-invasive plasma diagnostics are essential for real-time process monitoring, and OES remains a commonly adopted method. Conventional OES techniques typically assume maxwellian EEDF. However, this assumption limits the ability to accurately capture electron heating behaviors and to predict reactive species generation under thermally non-equilibrium conditions.

In this study, a novel diagnostic framework is proposed in which non-Maxwellian EEDFs ((a) generalized Maxwellian (b) bi-maxwellian (c) beam shifted maxwellian) are arbitrarily selected and incorporated into CRM-based line-ratio calculations. These computed emission intensities are quantitatively compared with experimental OES data to extract best EEDF represents the actual heating state of the target plasma. The methodology was evaluated by CCP-A and CCP-B with 60 MHz and 100 MHz power. The beam-shifted maxwellian showed to be the most suitable EEDFs among the three models. The beam electron energy and fraction by the OES method were (27.5 eV, 2.99 %) and (34.09 eV, 2.36 %) for the most enhanced case, respectively.

This method enables more precise analysis of electron heating dynamics (named PI: plasma information) and it allows reconstruction of electron-driven reactions in plasmas. In etch process, this approach can significantly improve the utilization of reactive species information and the applicability of EPD sensors with high accuracy by embedding physical plasma heating models, ultimately facilitating more robust and flexible process control strategies. The integration of domain knowledge in plasma physics with spectral signal interpretation is expected to drive the evolution of next-generation plasma sensing technologies for nanoscale etching environments.