Molecular Dynamics Studies of Amorphous Silicon Systems

Molecular dynamics simulations have become a cornerstone for elucidating the atomic‐scale structure and behaviour of amorphous silicon. Early models treated the material as a continuous random network, capturing short‐range tetrahedral bonding but leaving medium‐range order poorly characterised. Classical molecular dynamics using empirical interatomic potentials then enabled exploration of thermal histories, defect evolution and mechanical response under varying conditions of quench rate and temperature. More recently, first‐principles and machine‐learning‐driven force fields have refined our understanding of local distortions, ring distributions and the onset of paracrystalline regions within an otherwise disordered matrix. These approaches have provided unified explanations for conflicting experimental observations, revealing how nano‐scale pockets of crystalline order emerge during rapid cooling and how they influence electronic properties such as the band gap and carrier mobility. Simulations of self‐diffusion and vacancy migration have delineated atomic rearrangement mechanisms, while advanced analysis of time correlation functions has shed light on vibrational spectra and ageing processes. Together, these studies inform the design of thin films and devices—from solar cells to flexible electronics—by linking processing parameters to atomic structure, defect populations and functional performance.

Research from Nature Portfolio

Recent studies have employed machine-learning-driven molecular dynamics to map the boundary between amorphisation and crystallisation. By training interatomic potentials on quantum-mechanical data, researchers have systematically sampled quenched silicon configurations and demonstrated that small clusters exhibiting local crystalline motifs are consistent with experimental scattering and energy distributions. This work reconciles long-standing debates between continuous random network and paracrystalline models by showing that both emerge naturally under realistic cooling conditions. Another line of inquiry has combined first-principles atomistic simulation with diffraction data inversion to yield high-fidelity computer models of amorphous silicon. By embedding experimental structure factors into an optimisation loop, these simulations produce realistic networks with correct ring statistics, radial and angular distribution functions, and defect densities below a few per cent. This hybrid method represents a significant step toward predictive computational design of non-crystalline materials using accurate interatomic interactions supplemented by experimental constraints.

Molecular Dynamics Studies of Amorphous Silicon Systems publication trend

The graph below shows the total number of articles in molecular dynamics studies of amorphous silicon systems across all publications each year (not limited to Nature Index journals).

Technical terms

Continuous random network (CRN): A model of amorphous structure in which atoms form a disordered tetrahedral network without long-range periodicity.

Paracrystallinity: The presence of small regions exhibiting local crystalline order embedded within an overall amorphous matrix.

Quenching: Rapid cooling of a liquid or high-temperature solid to freeze in a non-equilibrium structure before crystallisation can occur.

Time correlation function: A statistical measure of how fluctuations in a physical quantity at one time relate to their values at a later time, used to derive vibrational and transport properties.

Mean squared displacement (MSD): The average of the squared distances that atoms travel over a given time interval, serving as a quantitative metric for diffusion.

References

1. Signatures of paracrystallinity in amorphous silicon from machine-learning-driven molecular dynamics. Nature Communications (2025).
2. Fourier transforms of time correlation functions using Hermite functions. Computer Physics Communications (2025).
3. Inversion of diffraction data for amorphous materials. Scientific Reports (2016).
4. Atomic mechanisms of self-diffusion in amorphous silicon. AIP Advances (2022).