SPICE-Compatible Degradation Modeling Framework for TDDB and LER Effects in Advanced Packaging BEOL Based on Ion...

Area of Science:

• Materials Science
• Electrical Engineering
• Semiconductor Physics

Background:

• Time-dependent dielectric breakdown (TDDB) is a critical failure mechanism in semiconductor devices and advanced packaging.
• Metal ion migration within inter-metal dielectric layers is a primary driver of TDDB.
• Accurate modeling is essential for predicting device lifetime and reliability.

Purpose of the Study:

• To develop a SPICE-compatible model for TDDB degradation caused by metal ion migration.
• To incorporate the effects of line-edge roughness (LER) into the degradation model.
• To enhance the accuracy of device lifetime predictions by accounting for process variations.

Main Methods:

• A SPICE-compatible model based on the physics of metal ion migration and conductive filament evolution was developed.
• The model was validated against numerical results for accuracy.
• A process variation model using power spectral density (PSD) was integrated to quantify LER effects.

Main Results:

• The developed model accurately predicts resistance changes and device lifetime under various operational conditions (voltage, temperature, structure).
• Incorporating LER through the PSD model significantly improves the accuracy of predicted lifetime curves.
• The model effectively captures reliability variations due to fabrication non-uniformities.

Conclusions:

• The proposed SPICE-compatible model provides a robust framework for understanding and predicting TDDB degradation.
• LER significantly impacts device reliability, and its inclusion in models is crucial for accurate lifetime prediction.
• This work offers a valuable tool for enhancing the reliability of semiconductor devices and advanced packaging.