I know this is a very controversial question, but is there a safest value of σ (sigma) and ϵ (epsilon) that can be used for the interaction between different atoms types such as Al-Fe (especially metals)?
- There is no such thing. Suitable potential parameters always depend on the conditions and composition.
- There are much better models for metals than the LJ potential, especially for alloys.
- What are good models and have been applied successfully can be learned from a proper and thorough survey of the published literature. The chances to get a meaningful response here are very slim, as your question is not a question about LAMMPS but a question about your research.