Area-Selective Deposition of WS2 via Sacrificial ZnS Conversion at Low Temperatures

Le résumé fourni par la source

2D transition metal dichalcogenides (TMDCs) offer a promising path forward as standard silicon technology approaches its physical scaling limits. However, integrating TMDCs into semiconductor manufacturing requires precise, back-end-of-line (BEOL)-compatible deposition methods. Area-selective deposition (ASD) enables direct bottom-up patterning, yet achieving surface-selective growth of tungsten disulfide (WS 2 ) under low thermal budgets remains a significant challenge. Here, we demonstrate a thermally driven ASD route for ultra-thin WS 2 films at 400°C, using tungsten(VI) sulfide tetrachloride (WSCl 4 ) and H 2 S as precursors and leveraging the unique reactivity of the halogenated, sulfur-containing metal precursor to drive selective ligand exchange and conversion. This approach uses a sacrificial, ultra-thin ZnS template layer to guide spatial selectivity. The process selectively converts pre-deposited ZnS into high-quality WS 2 at 400°C while completely suppressing growth on adjacent dielectrics. Applied to 200 mm undoped silicate glass (USG) / Si(100) wafers, this precursor combination yields pure, smooth, ultra-thin (< 10 nm) crystalline WS 2 films with predominantly horizontally oriented layers. The surface conversion mechanism shows remarkable chemical contrast, with high selectivity (S > 0.994) over 1200 deposition cycles, with USG as a non-growth area (NGA). Electrical measurements on tungsten-contacted transmission-line structures show functional behavior, yielding a layer resistivity of 2.9 × 10 7 µΩ·cm and a specific contact resistivity of 5.4×10 -3 Ω·cm 2 for ~7.5 nm WS 2 films. By leveraging sacrificial template conversion to enable highly selective, low-temperature thermal ASD of WS 2 , this work provides a scalable, chemistry-driven, direct, bottom-up integration of 2D materials into future advanced semiconductor device architectures.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.