Computational mechanistic investigation of the initial growth of Alumina ALD: Effect of substrate pretreatment on nucleation period reduction - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2018

Computational mechanistic investigation of the initial growth of Alumina ALD: Effect of substrate pretreatment on nucleation period reduction

Résumé

In this work, the surface mechanisms including reaction, adsorption and desorption steps during the ALD of Al2O3 from TMA and H2O on Si substrates are investigated. The analysis is performed using a surface chemistry model, coupled to a CFD model for a commercial reactor treating 20 cm Si wafers. [4] This model is used to feed a geometric model that simulates island growth on a surface. The model predictions are validated by comparing its results with experimental measurements and literature data. It is shown that for ALD exposures in the range of ms, no reactions between a perfect Si substrate and reactants take place due to the low activation energy of the desorption step. The effect of initial long duration TMA exposures before deposition is investigated, showing that they clearly lead to an increase of nucleation sites on the surface by allowing TMA adsorption and reaction to occur, thus producing methyl groups transformed in active hydroxyl bonds by the subsequent H2O exposure. This significantly reduces the nucleation period, hence decreasing the number of ALD cycles needed to obtain a conformal continuous film.
Fichier non déposé

Dates et versions

hal-04415188 , version 1 (24-01-2024)

Identifiants

  • HAL Id : hal-04415188 , version 1
  • OATAO : 21802

Citer

Giorgos Gakis, Hugues Vergnes, Emmanuel Scheid, Andreas G. Boudouvis, Constantin Vahlas, et al.. Computational mechanistic investigation of the initial growth of Alumina ALD: Effect of substrate pretreatment on nucleation period reduction. Rafald 2018, Nov 2018, Lyon (FRANCE), France. ⟨hal-04415188⟩
8 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More