Effect of Ag addition on the thermal characteristics and structural evolution of Ag-Cu-Ni ternary alloy nanoclusters: Atomistic simulation study

Atomic-scale compositional variation in Ag contents across Ag-Cu-Ni alloy upon being subjected to repeated annealing cycles is shown to result in significant differences in the structure and the thermal stability of ternary alloy nanoclusters. Molecular dynamics (MD) simulations employing quantum Sutton–Chen potentials were used to investigate the effect of Ag addition on the thermal characteristics of Ag-Cu-Ni ternary alloy nanoclusters of 4-nm diameter. The initial configurations were generated using Monte Carlo simulations and comprise surface-segregated structures with the lowest surface energy component, Ag, occupying low coordination sites such as corners, edges, and faces. A compositional oscillation between the Cu and Ni atoms was observed for layers beneath the surface which transitions into a bulk alloy composition at the core. We find that the Cu-Ni binary alloys on being subjected to annealing schedules demonstrated an increase in thermal stability, as indicated by the increase in melting points. The annealed configurations of the Ag-Cu-Ni ternary alloy, on the other hand, showed a nonmonotonic behavior. For Ag compositions less than 20%, we observe an initial increase in melting point followed by a decrease in the third cycle. For higher Ag compositions (>20%), we observe a decrease in melting point with annealing; the rate of decrease is strongly correlated to the Ag composition in the alloy. Cu-Ni nanoclusters having 50% Cu showed a transition from an initial icosahedral to a cuboctahedron-like structure whereas Ag-rich Ag-Cu-Ni ternary alloys showed a transition from icosahedral to an amorphous structure. Compositional analysis based on radial distribution functions and density profiles indicate that these transitions were dependent on the distribution of the alloying elements in the nanocluster. Calculated root-mean-square displacements and diffusion coefficients indicate that the rate of mixing of Ag increases with Ag content in the Ag-Cu-Ni ternary alloy. Ternary alloys show heterogeneous melting during the first heating cycle followed by a bulk-like melting during the subsequent annealing cycles. The simulation results are consistent with available experimental studies.

Figure 1

Sample energy variation during MC simulations of binary and ternary alloy nanoclusters as a function of the MC iteration. The system energy converges after ∼150 000 iterations for 34%Ag-33%Cu-33%Ni shown here.

Figure 2

Compositional profile for a Ag${}_{0.50}$-Ni${}_{0.25}$-Cu${}_{0.25}$ ternary alloy nanocluster: (a) initial configuration of the ternary alloy, (b) final equilibrated configuration of the ternary alloy showing the surface segregation of Ag.

Figure 3

Variations of potential energy with temperature for the various compositions of Ag-Cu-Ni ternary alloy nanoclusters: (a) 0%Ag-50%Cu-50%Ni, (b) 10%Ag-45%Cu-45%Ni, (c) 20%Ag-40%Cu-40%Ni, (d) 34%Ag-33%Cu-33%Ni, (e) 50%Ag-25%Cu-25%Ni. The melting point can be estimated based on the jump in potential energy curve and is shown for a series of three annealing cycles for each of the simulated alloy compositions.

Figure 4

The effect of Ag addition on the melting-point variation in a ternary alloy nanocluster for a series of three annealing cycles. The Ag composition is systematically varied from 0% to 50% Ag in a Ag-Cu-Ni ternary alloy nanocluster.

Figure 5

Variations in specific heat capacity with temperature for the various compositions of Ag-Cu-Ni ternary alloy nanoclusters. The Ag composition is systematically varied from 0% to 50% Ag in a Ag-Cu-Ni ternary alloy nanocluster: (a) 0%Ag-50%Cu-50%Ni, (b) 10%Ag-45%Cu-45%Ni, (c) 20%Ag-40%Cu-40%Ni, (d) 34%Ag-33%Cu-33%Ni, (e) 50%Ag-25%Cu-25%Ni. The melting point can be estimated based on the peaks in the specific heat capacity curves and is shown for a series of three annealing cycles for the various alloy compositions.

Figure 6

Evolution of bond orientational order parameter ($Q_6$) as a function of the simulation temperature for the various Ag-Cu-Ni ternary alloy nanocluster compositions. The dash-doted line gives variation of temperature as function of the simulation step and shows the sequence of heating-cooling cycles. The corresponding $Q_6$ order parameters for the various alloy compositions are represented by different markers.

Figure 7

Snapshots showing the structural evolution of the 50%Cu-50%Ni nanocluster during the first annealing cycle. The snapshots shown correspond to (a) 300 K, (b) 600 K, (c) 1100 K (d) 700 K, (e) 300 K temperatures. The crystal structure transforms from an initial icosahedral structure in (a) to a liquid phase in (c) at the melting point. Upon cooling, the structure shown in (c) recrystallizes to the cuboctahedron shown in (e). Cu atoms are represented in light blue and Ni atoms are shown as dark blue spheres.

Figure 8

Snapshots showing the structural evolution of the 50%Ag-25%Cu-25%Ni nanocluster during the first annealing cycle. The snapshots shown correspond to (a) 300 K, (b) 600 K, (c) 1000 K, (d) 500 K, (e) 300 K temperatures. The crystal structure transforms from an initial icosahedral structure in (a) to a liquid phase in (c) at the melting point. Upon cooling, the structure shown in (c) recrystallizes to the amorphous structure shown in (e), which is retained during subsequent annealing cycles. Ag atoms are shown in green color, Cu atoms are represented in light blue and Ni atoms are shown as dark blue spheres.

Figure 9

Calculated RDFs for the various annealed binary and ternary alloy nanoclusters. The RDF is calculated at the end of the third annealing cycle. The absence of long-range order in the calculated RDF for 34%Ag and 50%Ag suggests that the addition of Ag to Cu-Ni alloy results in annealed structures which are amorphous in nature.

Figure 10

(a) Calculated PDFs for the various metal-metal interactions in a 50%Cu-50%Ni binary alloy nanocluster. The solid and dashed lines refer to PDFs obtained before and after the annealing cycles. The PDFs for the various metal interactions in a 50%Cu-50%Ni binary alloy nanocluster show sharp peaks at distances corresponding to near neighbor distances. The presence of long-order in the annealed Cu-Ni structures is indicative of a crystalline order being retained after the heating-cooling schedules. (b),(c) Calculated PDFs for the various metal-metal interactions in a 50%Ag-25%Cu-25%Ni ternary alloy nanocluster. (b) represents RDFs for Cu-Cu, Cu-Ni, and Ni-Ni correlations in the ternary alloy, whereas (c) represents Ag-Ag, Ag-Cu, and Ag-Ni correlations. The solid and dashed lines refer to PDFs obtained before and after the annealing cycles, respectively.

Figure 11

Compositional variations across the nanocluster radius for 50%Cu-50%Ni, 10%Ag-45%Cu-45%Ni, and 20%Ag-30%Cu-30%Ni alloys. The solid and dashed lines represent compositions before and after annealing, respectively.

Figure 12

(a) Variations of the self-diffusion coefficient with temperature for the various compositions of the ternary alloys. Ag composition is increased from 0 to 50%. Solid lines indicate diffusion coefficients at various temperatures before annealing whereas dashed lines with markers are used to indicate the diffusion coefficients after annealing. (b)–(e) Root MSD calculated for atoms in various shells for 50%Cu-50%Ni and 50%Ag-25%Cu-25%Ni as a function of distance from the center of cluster. (b) and (c) represent the shell-based root MSD calculated for preannealed and annealed configurations, respectively, for the 50%Cu-50%Ni alloy. (d) and (e) represent the shell-based root MSD calculated for preannealed and annealed configurations, respectively, for the 50%Ag-25%Cu-25%Ni ternary alloy.