On the road to metallic nanoparticles by rational design: bridging the gap between atomic-level theoretical modeling and reality by total scattering experiments

Abstract

The extent to which current theoretical modeling alone can reveal real-world metallic nanoparticles (NPs) at the atomic level was scrutinized and demonstrated to be insufficient and how it can be improved by using a pragmatic approach involving straightforward experiments is shown. In particular, 4 to 6 nm in size silica supported Au(100-x)Pd(x) (x = 30, 46 and 58) explored for catalytic applications is characterized structurally by total scattering experiments including high-energy synchrotron X-ray diffraction (XRD) coupled to atomic pair distribution function (PDF) analysis. Atomic-level models for the NPs are built by molecular dynamics simulations based on the archetypal for current theoretical modeling Sutton-Chen (SC) method. Models are matched against independent experimental data and are demonstrated to be inaccurate unless their theoretical foundation, i.e. the SC method, is supplemented with basic yet crucial information on the length and strength of metal-to-metal bonds and, when necessary, structural disorder in the actual NPs studied. An atomic PDF-based approach for accessing such information and implementing it in theoretical modeling is put forward. For completeness, the approach is concisely demonstrated on 15 nm in size water-dispersed Au particles explored for bio-medical applications and 16 nm in size hexane-dispersed Fe48Pd52 particles explored for magnetic applications as well. It is argued that when "tuned up" against experiments relevant to metals and alloys confined to nanoscale dimensions, such as total scattering coupled to atomic PDF analysis, rather than by mere intuition and/or against data for the respective solids, atomic-level theoretical modeling can provide a sound understanding of the synthesis-structure-property relationships in real-world metallic NPs. Ultimately this can help advance nanoscience and technology a step closer to producing metallic NPs by rational design.

Solving the Structure and Dynamics of Metal Nanoparticles by Combining X-Ray Absorption Fine Structure Spectroscopy and Atomistic Structure Simulations

“…For example, one can force the interatomic distances and bond angles to be within the physically reasonable ranges (87) or preserve a specific number of nearest neighbors around each atom (81). Even more information can be introduced by using the so-called hybrid-RMC method (17,27,88) where, in addition to determination of the discrepancy between experimentally observed and theoretically calculated properties for each model, the corresponding energy is also calculated. In this case, energetically unfavorable configurations are penalized by additional contribution to the cost function .…”

“…However, the design of such empirical potential models is a complex task. The advisability of transferring potentials even between similar systems is still an open question (17). Another problem is the development of potentials for heterogeneous systems, e.g., where metallic phases coexist with oxidized phases, which is a common situation in heterogeneous catalysis (33,50,101).…”

“…Recent years have, however, seen a rise in the importance and availability of atomic-level simulations of atomic and electronic structure and properties of NPs. In fact, theoretical modeling approaches are now used to fill the gaps in experimental data (6,17). For example, the use of ab initio simulations within the density functional theory (DFT) formalism allows one to model relaxation processes in metallic NPs, and it has been used to predict the properties and structure of nanosized catalysts (18)(19)(20).…”

“…For example, the use of ab initio simulations within the density functional theory (DFT) formalism allows one to model relaxation processes in metallic NPs, and it has been used to predict the properties and structure of nanosized catalysts (18)(19)(20). Molecular dynamics (MD) and Monte Carlo (MC) simulations using interatomic forces that are derived either from ab initio calculations (15,18) or from empirical force fields (17,(21)(22)(23) can be employed, in turn, to model thermal disorder and anharmonic effects. Thus, theoretical modeling can be a very powerful tool to advance our understanding of the local structure and dynamics in small NPs.…”

“…To summarize, atomic-scale simulation of nanomaterial structure, guided by XAS data, is a new and rapidly developing field. The importance of and growing interest in such studies are supported by a number of recent publications devoted to this topic (15,17,19,26,(28)(29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39)(40)(41)(42)(43)(44)(45). The approaches described in these articles often rely on similar ideas, but their implementation varies because such combined studies are still a relatively new tool for modeling NP structure.…”