A detailed computational investigation of 55-atom Pd \(_{\varvec{N}}\) Ag \(_{\varvec{42-N}}\) Ni \(_{\varvec{13}}\) Mackay-icosahedral nanoalloys is presented, combining Gupta-potential Basin-Hopping, spin-polarized DFT, local virial stress mapping, and finite-temperature molecular dynamics (MD). Chemical-ordering optimizations reveal that Ag preferentially occupies vertex sites while Pd populates edges and subsurfaces, yielding composition-dependent mixing energies that remain unfavorable up to \(\varvec{N\approx 8}\) at the DFT level. Local pressure maps show anomalous surface compression and tensile subshells in Ag-rich clusters (Ni–Ag mismatch \(\varvec{\approx 16\%}\) ) and conventional compressive-core/tensile-shell distributions in Pd-rich clusters (Ni–Pd mismatch \(\varvec{\approx 10\%}\) ). Melting-dynamics analyses indicate sharp transitions with no pre-melting surface rearrangements. The anomalous pressure distribution in Ag-rich compositions lowers their thermal resistance, resulting in melting at significantly lower temperatures than Pd-rich clusters, which maintain conventional stress profiles and higher thermal stability. Spin-polarized DFT confirms Ni atoms carry the dominant local magnetic moments, governed by coordination rather than local stress. Alternative Bergman-shell variants were also tested, demonstrating that a 32-atom shell restores the expected compressive-core/tensile-shell stress pattern.
