Atomic layer deposition (ALD) has emerged as a powerful technique for synthesizing two-dimensional (2D) transition metal dichalcogenides (TMDCs) such as molybdenum disulfide (MoS2) and tungsten disulfide (WS2). Unlike conventional chemical vapor deposition (CVD), ALD offers precise control over film thickness at the atomic scale, enabling uniform and conformal coatings even on complex substrates. The self-limiting nature of ALD reactions ensures layer-by-layer growth, making it particularly suitable for applications requiring ultrathin, stoichiometric films. However, achieving high-quality 2D TMDCs via ALD presents unique challenges, including precursor selection, substrate compatibility, and maintaining stoichiometry during deposition.
**Substrate Compatibility and Surface Preparation**
The choice of substrate plays a critical role in ALD of 2D materials. Common substrates include silicon with native oxide (SiO2/Si), sapphire (Al2O3), and highly oriented pyrolytic graphite (HOPG). Each substrate influences nucleation density and film morphology due to differences in surface energy and lattice mismatch. For instance, SiO2/Si provides a hydrophilic surface that promotes uniform precursor adsorption, while HOPG’s inert basal plane often results in sparse nucleation. Pretreatment methods such as oxygen plasma or chemical functionalization can enhance nucleation density by introducing reactive sites.
Single-crystalline substrates like sapphire offer epitaxial growth opportunities due to their hexagonal symmetry, which aligns with the crystal structure of TMDCs. However, amorphous substrates require careful optimization of deposition parameters to minimize defects. Temperature is a key variable; typical ALD processes for MoS2 and WS2 operate between 200°C and 350°C. Lower temperatures may lead to incomplete precursor reactions, while excessive heat can induce sulfur deficiency or undesired phase transformations.
**Layer-by-Layer Growth Mechanism**
ALD growth of TMDCs follows a cyclical sequence of precursor pulsing and purging. For MoS2, a common approach involves alternating molybdenum precursors (e.g., MoCl5 or Mo(CO)6) with sulfur sources (e.g., H2S or organic sulfides). The process begins with the substrate exposure to the metal precursor, which chemisorbs onto surface sites. Excess precursor is purged with inert gas, followed by the introduction of the sulfur precursor to complete the reaction. Each cycle ideally adds a single monolayer, though growth per cycle (GPC) can vary depending on precursor reactivity and surface saturation.
Achieving self-limiting growth requires precise control over pulse durations and purge times. For example, MoCl5 and H2S reactions typically use pulse times of 0.5–2 seconds, with purging intervals of 5–20 seconds to prevent gas-phase reactions. Incomplete purging can lead to CVD-like growth, compromising thickness uniformity. The growth rate is also influenced by precursor volatility and substrate temperature. Metalorganic precursors like Mo(CO)6 offer lower decomposition temperatures but may introduce carbon contamination, while halide-based precursors like MoCl5 require higher temperatures for complete ligand removal.
**Challenges in Stoichiometric Film Formation**
Stoichiometry control is a major challenge in ALD of TMDCs. Sulfur deficiency is common due to the high volatility of sulfur species and preferential re-evaporation during growth. This can lead to non-stoichiometric phases (e.g., MoS2-x) or metallic clusters, degrading electronic properties. Strategies to mitigate sulfur loss include:
- Using excess sulfur precursors during the sulfurization step.
- Post-deposition annealing in sulfur-rich atmospheres to restore stoichiometry.
- Employing low-volatility sulfur sources like thiols or thiourea.
Another issue is the formation of undesired crystalline phases. MoS2 can exist in metallic (1T) or semiconducting (2H) phases, with the latter being desirable for most electronic applications. The 2H phase dominates at moderate temperatures (250–350°C), but higher temperatures or improper precursor ratios may stabilize the 1T phase. Oxygen incorporation is another concern, particularly when using oxide substrates or moisture-sensitive precursors. Even trace oxygen can form MoOx intermediates, disrupting the growth of pure MoS2.
**Advanced Process Modifications**
Recent advancements in ALD techniques address these challenges through process innovations. Plasma-enhanced ALD (PE-ALD) uses reactive plasma species (e.g., H2S plasma) to enhance precursor dissociation at lower temperatures, reducing sulfur loss and improving film quality. Spatial ALD, where substrates move between separated precursor zones, enables faster growth without vacuum purging, though it requires precise mechanical control.
Area-selective ALD is another promising approach, leveraging patterned inhibitors or substrate activation to grow TMDCs only on predefined regions. This is critical for fabricating integrated circuits with minimal post-processing. For example, self-assembled monolayers (SAMs) can block precursor adsorption on non-target areas, enabling direct synthesis of patterned MoS2 films.
**Conclusion**
ALD provides a scalable and controllable route for synthesizing 2D TMDCs with atomic-level precision. Substrate engineering, precursor chemistry, and process optimization are essential for achieving stoichiometric, high-quality films. While challenges such as sulfur deficiency and phase control persist, advancements in plasma assistance and selective growth techniques continue to expand the capabilities of ALD for 2D materials. Future developments may focus on lowering deposition temperatures for compatibility with flexible substrates and integrating ALD-grown TMDCs into functional devices with minimal post-processing.