通过工程化表面分离壁抑制纳米颗粒聚集：一项机器学习驱动的分子动力学与实验研究

《Journal of Catalysis》：Suppressing nanoparticle aggregation via engineered surface separation walls: a machine learning–driven molecular dynamics and experimental study

　　

This paper presents a systematic study on using **surface separation walls** to enhance the thermal stability of Fe nanoparticles supported on Al?O?, with applications to carbon nanotube (CNT) growth.

## Key Elements:

**Problem:**
At CNT growth temperatures (500–1000°C), metal nanoparticles undergo deformation, migration, and aggregation via Ostwald ripening and particle migration coalescence, compromising CNT quality.

**Approach:**
- Developed a **deep neural network-based machine learning potential** for the Fe/Al?O? system, trained on high-precision DFT data
- Combined MD simulations with experimental validation
- Investigated how wall geometry (height, width, length) affects nanoparticle aggregation

**Main Findings:**

1. **Infinite-length walls:** Increasing height or width consistently suppresses aggregation. The critical **height-to-diameter ratio** determines the onset of aggregation, and this ratio decreases with increasing wall width.

2. **Finite-length walls:** Simply increasing height is insufficient—nanoparticles can migrate around wall ends, a limitation not present in the previously studied Pt/TiO? system.

3. **Synergistic mechanism:** Two stabilization pathways act together:
- **Physical/geometric confinement** from separation walls
- **Chemical anchoring** via defect-mediated metal-support interactions (undercoordinated surface atoms at defect sites provide enhanced local binding)

**System-Specific Distinction:**
The authors emphasize this is not merely a transfer of prior frameworks. Per the Sabatier principle, Fe (E_c = ?4.28 eV) requires weaker MSI (?2.27 eV) than Pt (?5.84 eV, requiring ?3.10 eV) to reach peak stability. This gives Fe/Al?O? distinctive mechanistic characteristics directly relevant to CNT synthesis, where CNT diameter and areal density are determined by Fe nanoparticle size.

**Design Guidance:**
The derived critical height-to-diameter ratio is intended as a **qualitative design guideline** rather than a universal quantitative threshold.

**Significance:**
This work bridges atomic-scale simulations → mesoscale nanoparticle stability (experiments) → macroscopic CNT growth quality, establishing an integrated "simulation–experiment–application" chain not previously demonstrated in separation-wall studies.