Course: Materials Simulation Practical | FAU Erlangen-Nürnberg
Tools: Python · ASE · EMT potential · NumPy · SciPy · Matplotlib
📄 Written report | 💻 Simulation code
Atomistic simulation of copper (Cu) using the Effective Medium Theory (EMT) interatomic potential. The project has three parts: determining the equilibrium lattice parameter across SC, BCC, and FCC crystal structures; extracting all three independent elastic constants via energy-strain analysis; and quantifying the strain range over which linear elasticity remains valid by comparison with finite-temperature molecular dynamics.
Energy per atom scanned as a function of lattice constant for SC, BCC, and FCC structures using a two-stage coarse/fine grid with quadratic fitting to locate each minimum precisely.
![]() |
| Energy per atom vs. lattice constant for SC, BCC, and FCC structures |
| Structure | ℓ₀ (Å) | W₀ (eV/atom) |
|---|---|---|
| SC | 2.4265 | +0.446 |
| BCC | 2.8622 | +0.024 |
| FCC | 3.5978 | −0.008 |
FCC is the ground-state structure. The computed lattice parameter deviates from the experimental value (3.615 Å) by −0.48%.
Three independent strain paths applied to the relaxed FCC reference cell: uniaxial (ε₁₁), hydrostatic (εᵢᵢ), and shear (γ). The curvature of each strain energy density curve gives a combination of elastic constants, from which C₁₁, C₁₂, and C₄₄ are extracted analytically.
![]() |
| Energy-strain curves and cubic fits for the three strain paths |
| Constant | Computed (10¹¹ N/m²) | Literature | Error |
|---|---|---|---|
| C₁₁ | 1.7166 | 1.7620 | −2.6% |
| C₁₂ | 1.1617 | 1.2494 | −7.0% |
| C₄₄ | 0.8684 | 0.8177 | +6.2% |
All three constants are within ~7% of experimental values, consistent with the known accuracy limits of the EMT potential.
The small-strain quadratic approximation (fitted at ±0.5%) is extrapolated to ±20% and benchmarked against NVT Langevin MD at 10 K. Validity is defined as ≤5% relative error between the quadratic model and MD.
| Deformation Mode | Linear elasticity valid up to |
|---|---|
| Uniaxial | ~11.5% |
| Hydrostatic | ~5.0% |
| Shear | ~16.5% |
![]() |
![]() |
| Energy-strain curve | Relative error vs. MD |
![]() |
![]() |
| Energy-strain curve | Relative error vs. MD |
![]() |
![]() |
| Energy-strain curve | Relative error vs. MD |
Hydrostatic loading breaks down earliest (~5%), reflecting strong volumetric nonlinearity in the EMT potential. Shear deformation remains within the linear regime to the largest strains (~16.5%), consistent with the relatively flat off-diagonal energy landscape in close-packed metals. The large relative errors visible near ε = 0 in the error plots are a numerical artifact of near-zero division — absolute errors at those strains are negligible.
pip install ase numpy matplotlib scipyOpen atomistic_simulation_Cu.ipynb in Jupyter and run cells sequentially. Tasks 3.5
and 3.6 depend on the results dictionary populated in Task 3.2. Task 3.6 runs
Langevin MD across 123 strain points — expect approximately 30 minutes on a
standard laptop.







