Controlling Mixed Mo/MoS$_2$ Domains on Si by Molecular Beam Epitaxy for the Hydrogen Evolution Reaction

This study demonstrates that controlling sulfur stoichiometry and growth kinetics during molecular beam epitaxy on silicon substrates enables the creation of defect-engineered Mo/MoS$_2$ heterostructures with residual metallic Mo and sulfur vacancies, which significantly enhance hydrogen evolution reaction performance by activating inert basal planes and improving charge transfer compared to stoichiometric films.

The Gist

The Big Picture: Making Better "Hydrogen Factories"

Imagine you want to build a factory that produces Hydrogen, a super-clean fuel for the future. The best "machines" for this job are made of a material called Molybdenum Disulfide (MoS₂). Think of MoS₂ like a stack of ultra-thin, slippery sheets (like a deck of cards).

For a long time, scientists knew that the edges of these sheets were the "magic spots" where hydrogen is made. The flat middle of the sheet (the "basal plane") was considered useless, like a smooth table where nothing happens.

The Problem:
Making these sheets perfectly is tricky. If you make them too perfect, they become too smooth and lose their "magic edges." If you make them too messy, they stop conducting electricity. It's like trying to bake the perfect cake: if you mix too much, it's dense; if you mix too little, it falls apart.

The Solution:
This team of scientists used a high-tech oven called Molecular Beam Epitaxy (MBE). Think of this as a "3D printer for atoms" that can build these sheets directly onto silicon computer chips with incredible precision. They wanted to find the "Goldilocks zone"—not too perfect, not too messy, but just right.

---

The Three Experiments: Tuning the Recipe

The researchers tested three different "knobs" on their machine to see how they changed the performance of the hydrogen factory.

1. The Heat Knob (Annealing Temperature)

• The Analogy: Imagine baking cookies.
  • Low Heat (600°C): The cookies are a bit rough and crumbly. They have lots of jagged edges.
  • High Heat (800°C): The cookies bake into big, smooth, perfect circles. They look beautiful, but they have very few jagged edges.
• The Result: The "rough" cookies (lower heat) worked better! Even though they looked less perfect, they had more edges where the hydrogen could be made. The super-smooth cookies (high heat) were too organized and lost their active spots.

2. The Thickness Knob (Number of Layers)

• The Analogy: Think of a stack of paper.
  • Too Thin: A single sheet is flimsy and hard to handle.
  • Too Thick: A massive stack of 50 sheets. The electricity has a hard time traveling all the way through the stack to get to the bottom. It's like trying to shout through a thick wall; the signal gets lost.
  • Just Right: A stack of about 10 sheets.
• The Result: The medium stack (10 layers) was the winner. It was thick enough to be stable but thin enough to let electricity flow easily to the active spots.

3. The Sulfur Knob (Sulfur Supply)

• The Analogy: Imagine building a wall with bricks (Molybdenum) and mortar (Sulfur).
  • Too Much Mortar: You cover the bricks completely. The wall is solid, but the bricks can't "talk" to each other. The electricity gets stuck.
  • Too Little Mortar: The wall falls apart.
  • The "Secret Sauce" (Slightly Less Mortar): The researchers found that leaving a few gaps (missing sulfur) and letting some raw bricks (metallic Molybdenum) peek through actually helped!
• The Result: The "imperfect" walls with a few gaps and some exposed metal bricks worked the best. The metal acted like a highway for electricity, and the gaps acted like open doors for the reaction to happen.

---

The "Aha!" Moment: Why Imperfection Wins

The biggest discovery of this paper is that perfection is overrated.

In the past, scientists tried to make MoS₂ sheets that were 100% pure and perfectly ordered. This paper shows that the best catalysts are actually hybrids. They are a mix of:

1. The Sheet (MoS₂): Where the reaction happens.
2. The Metal (Mo): Which acts as a super-fast wire to deliver electricity.
3. The Gaps (Defects): Which turn the "useless" flat middle of the sheet into a place where reactions can happen.

By using their "atomic 3D printer," they created a material that is slightly messy on purpose. This "controlled mess" allows electricity to flow faster and gives the hydrogen reaction more places to happen.

The Bottom Line

This research is a breakthrough because:

1. It works on Silicon: They grew these materials directly on silicon chips, which is a huge step toward putting these fuel factories inside our computers and electronics.
2. It's Efficient: Their best sample produced hydrogen much faster and with less energy waste than previous versions.
3. It Changes the Rules: It proves that for this specific job, a little bit of disorder and a mix of materials is better than a perfectly pure crystal.

In short: They figured out how to build a better hydrogen factory by intentionally making it a little bit "imperfect" in the right ways, using a high-tech method that could eventually help power our clean energy future.

Technical Summary

1. Problem Statement

Molybdenum disulfide (MoS₂) is a promising catalyst for the Hydrogen Evolution Reaction (HER), but its performance is governed by a complex trade-off between crystallinity, defect density, and electronic conductivity.

• The Challenge: In conventional synthesis methods (e.g., Chemical Vapor Deposition or solution-based growth), it is difficult to independently control these parameters. Typically, high crystallinity leads to a loss of active edge sites and reduced conductivity, while defect-rich films often suffer from poor structural order.
• The Gap: There is a lack of systematic studies that can decouple the effects of growth kinetics, sulfur stoichiometry, and annealing on the structural evolution of MoS₂, particularly on semiconductor substrates (Si) where direct integration is desired for device applications. Furthermore, the deliberate formation of spatially coexisting metallic Mo and MoS₂ domains within a continuous thin film for HER has not been systematically reported.

2. Methodology

The authors utilized Molecular Beam Epitaxy (MBE) to grow MoS₂ thin films directly on Silicon (Si) substrates. MBE was chosen for its ability to provide sub-monolayer flux precision and wafer-scale uniformity.

• Experimental Design: Three systematic sample series were created to isolate specific variables while keeping all other parameters constant:
  1. Annealing Temperature Series: Varied from 600°C to 800°C (fixed deposition cycles and sulfur flux).
  2. Deposition Cycle Series: Varied from 5 to 50 cycles (fixed annealing and sulfur flux) to control film thickness.
  3. Sulfur Stoichiometry Series: Varied sulfur layer thickness (2.0 Å to 9.0 Å per Mo layer) to control the Mo/S ratio and phase purity.
• Characterization Techniques:
  • Structural: X-ray Diffraction (XRD), Reflection High-Energy Electron Diffraction (RHEED), Atomic Force Microscopy (AFM), Scanning Transmission Electron Microscopy (STEM).
  • Electronic/Local Structure: X-ray Absorption Near-Edge Structure (XANES) and Extended X-ray Absorption Fine Structure (EXAFS) at the Mo K-edge to distinguish between metallic Mo and 2H-MoS₂ phases.
  • Electrochemical: Linear Sweep Voltammetry (LSV), Tafel analysis, Electrochemical Impedance Spectroscopy (EIS), and Double-Layer Capacitance (Cdl) measurements to determine Overpotential ($\eta$), Charge-Transfer Resistance ($R_{ct}$), and Electrochemical Surface Area (ECSA).

3. Key Contributions

• Direct Epitaxial Growth on Si: Successfully demonstrated the growth of atomically uniform MoS₂ films directly on Si substrates, enabling strong interfacial coupling and semiconductor-integrated catalytic architectures.
• Decoupling Growth Parameters: Established a "one-variable-at-a-time" framework to systematically correlate specific growth conditions with structural and catalytic outcomes, overcoming the limitations of post-hoc correlations in CVD-grown samples.
• Discovery of Optimal Mixed-Phase Regime: Identified that non-stoichiometric growth conditions (specifically sulfur deficiency) that yield a coexistence of metallic Mo and semiconducting MoS₂ domains outperform fully stoichiometric, highly crystalline films.
• Defect Engineering Strategy: Demonstrated that controlled sulfur vacancies and residual metallic Mo act as conductive pathways and activate the otherwise inert basal planes of MoS₂.

4. Key Results

A. Effect of Annealing Temperature

• Structural: Higher temperatures (up to 800°C) improved crystallinity (sharper XRD peaks, larger crystallite sizes) and surface smoothness.
• Performance: Despite better crystallinity, HER activity decreased with higher annealing temperatures.
  • Reason: High temperatures reduced the density of catalytically active edge sites and increased bulk resistivity.
  • Best Result: MoS-T600 (600°C) showed the best activity among annealed samples ($\eta = -0.456$ V at -10 mA cm⁻²) due to a higher density of edge sites.

B. Effect of Deposition Cycle Number (Thickness)

• Structural: Increasing cycles increased film thickness and crystallite size but also increased resistivity due to suppressed vertical electron transport across van der Waals layers.
• Performance: An intermediate thickness was optimal.
  • Best Result: MoS-N10 (10 cycles) achieved the lowest overpotential (-0.33 V), the largest ECSA (8.0 cm²), and the highest mass-based turnover frequency (24.9 mmol H₂ g⁻¹ s⁻¹).
  • Trend: Too thin (N5) lacked sufficient active sites; too thick (N30–N50) suffered from high resistivity and reduced edge-to-area ratios.

C. Effect of Sulfur Stoichiometry

• Structural:
  • Sulfur-Deficient (M2.0–M6.0): Resulted in mixed phases containing residual metallic Mo and sulfur vacancies. EXAFS subtraction analysis confirmed the coexistence of Mo–Mo (metallic) and Mo–S (MoS₂) coordination shells.
  • Sulfur-Rich (M8.0–M9.0): Yielded pure 2H-MoS₂ but introduced stacking disorder and eliminated conductive metallic pathways.
• Performance: Sulfur-deficient films significantly outperformed stoichiometric ones.
  • Best Result: MoS-M6.0 (moderate sulfur deficiency) achieved an overpotential of -0.35 V and an ECSA of 9.2 cm².
  • Mechanism: The metallic Mo domains provided high electrical conductivity, while sulfur vacancies activated the basal planes. Sulfur-rich films suffered from high charge-transfer resistance ($R_{ct}$) and lower ECSA.

D. Overall Structure-Activity Correlation

• Optimal Regime: The highest HER performance was not found in perfectly crystalline or stoichiometric films. Instead, it emerged in an intermediate regime characterized by:
  1. Moderate crystallinity (preserving edge sites).
  2. Controlled sulfur deficiency (creating vacancies and metallic Mo domains).
  3. Intermediate thickness (balancing surface area and conductivity).
• Kinetics: The optimal samples exhibited lower Tafel slopes (e.g., 80 mV dec⁻¹ for N10) and lower $R_{ct}$, indicating facilitated Volmer steps and efficient interfacial charge transfer.

5. Significance

• Scientific Impact: This work challenges the conventional paradigm that "higher crystallinity equals better catalysis" for MoS₂. It establishes that controlled disorder (defects, mixed phases, and off-stoichiometry) is a superior strategy for enhancing HER activity by simultaneously optimizing conductivity and active site density.
• Technological Application: The ability to grow high-quality, defect-engineered MoS₂ directly on Si via MBE opens new avenues for integrating electrocatalysts with semiconductor electronics, potentially leading to on-chip hydrogen generation devices.
• Generalizability: The findings provide a generalizable design principle for layered transition-metal dichalcogenides (TMDs), suggesting that atomic-scale engineering of stoichiometry and growth kinetics can be used to tune the interplay between structural order and catalytic reactivity for next-generation energy materials.

Key Metrics of the Best Sample (MoS-N10):

• Overpotential: -0.33 V at -10 mA cm⁻²
• ECSA: 8.0 cm²
• Mass-based TOF: >23 mmol H₂ g⁻¹ s⁻¹ (more than double that of stoichiometric counterparts)
• Tafel Slope: 80 mV dec⁻¹