Spotting Local Environments in Self-Assembled Monolayer-Protected Gold Nanoparticles

Abstract

Organic-inorganic (O-I) nanomaterials are versatile platforms for an incredible high number of applications, ranging from heterogeneous catalysis to molecular sensing, cell targeting, imaging, and cancer diagnosis and therapy, just to name a few. Much of their potential stems from the unique control of organic environments around inorganic sites within a single O-I nanomaterial, which allows for new properties that were inaccessible using purely organic or inorganic materials. Structural and mechanistic characterization plays a key role in understanding and rationally designing such hybrid nanoconstructs. Here, we introduce a general methodology to identify and classify local (supra)molecular environments in an archetypal class of O-I nanomaterials, i.e., self-assembled monolayer-protected gold nanoparticles (SAM-AuNPs). By using an atomistic machine-learning guided workflow based on the Smooth Overlap of Atomic Positions (SOAP) descriptor, we analyze a collection of chemically different SAM-AuNPs and detect and compare local environments in a way that is agnostic and automated, i.e., with no need of a priori information and minimal user intervention. In addition, the computational results coupled with experimental electron spin resonance measurements prove that is possible to have more than one local environment inside SAMs, being the thickness of the organic shell and solvation primary factors in the determining number and nature of multiple coexisting environments. These indications are extended to complex mixed hydrophilic-hydrophobic SAMs. This work demonstrates that it is possible to spot and compare local molecular environments in SAM-AuNPs exploiting atomistic machine-learning approaches, establishes ground rules to control them, and holds the potential for the rational design of O-I nanomaterials instructed from data.

Keywords: ESR; SOAP; fluorinated nanoparticles; machine learning; mixed monolayers; multiscale modeling; nanoconfinement.

Figure 1

Exemplification of the concept of local (supra)molecular environment (highlighted in blue) in SAM-AuNPs and its exploitation. (a) If ligands contain a catalytic group and the surrounding molecules adopt specific cooperative conformation and order, 3D binding sites similar to those in enzymes may arise with enhanced catalytic properties. (b) The end group on the surface switches on/off the access to a catalytic center and grants selective diffusion to the organic layer, causing different local structural features and reagent concentration. (c) Heteroligand monolayers of two immiscible ligands lead to surface anisotropy with implications for surface related biological processes and sensing of biomolecules, biomarkers, and drugs.

Figure 2

Structure of the thiolates 1–6 for the AuNP coating. Radical probe 7 for ESR investigation. Ligand 6 is used in mixed monolayers with 1–5. Thiolates differ in nature and charge of the terminal group (1 and 2, a positively charged quaternary ammonium ion; 3 and 4, a negatively charged sulfonate ion; 5, a zwitterionic group, composed by a trimethylammonium and a phosphate group) as well as in length of the alkyl chain (C₁₂ in 1, 3, 5; C₁₆ in 2, 4).

Figure 3

Representative molecular structures of homoligand NP1, NP3, and NP5 AuNP and its heteroligand NP1/6, NP3/6, and NP5/6 counterpart from molecular dynamics simulations in explicit solvent (water). For clarity, water and counterions are not shown. Color representation of atoms: C, gray; O, red; S, yellow; P, orange; N, blue; F, green; H, white.

Figure 4

(a) Normalized water distribution at increasing distance from the gold surface for NP1/6, NP3/6, and NP5/6. The graphs plot the distribution of the atom (oxygen of water or carbon of thiolates) closest to gold surface (centered on the gold core and placed at increasing distances from its surface) shown as a two-dimensional projection of the sphere surface (x-axis, the azimuthal angle φ; y-axis, the cosine of the polar angle θ). A value of 1 indicates that an oxygen atom of a water molecule is always the closest; if it is equal to 0, it indicates that a carbon/fluorine atom of a chain is always the closest. Simplifying, red to salmon areas represent poorly hydrated zones, while blue areas stand for highly hydrated parts of the monolayer (at a certain distance from the gold surface). At distances lower than those considered, the microenvironment is almost hydrophobic, while at higher distances, it is fully hydrated and no major difference between the monolayers could then be detected. Maps for NP2/6 and NP4/6 can be found in the SI (Figures S5 and S6). (b) Examples of possible different hydration states within SAMs.

Figure 5

Conceptual diagram of the workflow used for the detection and comparison of local molecular environments within self-assembled monolayers (SAMs) using the Smooth Overlap of Atomic Positions (SOAP)-based structural analysis. Molecular dynamics calculations of the SAM-AuNP and reporter 7 are conducted in explicit solvent. The SOAP descriptor vector is constructed taking the reporter atoms (here the nitrogen atom) as the center of the structural environment up to a given cutoff radius r₁ (medium-range description) and employed for the identification of molecular fingerprints assigned by an unsupervised clustering algorithm (step 1). The short-range SOAP descriptor is built considering only solvent molecules up to a range of r₂ (<r₁), and a linear kernel between SOAP vectors is used to measure the similarity between the environments (step 2) and interpreted by correlating the location of the data with the MD evidence. For more details on each step, see Figure S8 and Section S3 in the SI.

Figure 6

First two principal components (PCA1 and PCA2) obtained from dimensionality reduction of the medium-range SOAP feature space of the probe 7 in thicker homoligand NP4 (a) and NP2 (b). Dots are colored according to the clusterization obtained by the GMM analysis. For each cluster, the inset shows the molecular environment centered on the probe 7, extracted from the corresponding MD frames. Color legend: probe, same color of the cluster; ligands 4 and 2 in gray; solvent not shown for clarity. (c, d) Example of the molecular view of the local environments NP4₁ and NP4₂ including all atoms within the cutoff r₁. The reporter is colored according to the cluster assigned as a sphere; water is shown in the same color of the probe but as a transparent surface, and the ligands belonging to the environment are highlighted as white spheres. The remaining ligands are left as a background gray surface. (e, f) Free energy surface (FES) (kcal/mol) calculated from the state’s probability distribution in (a) and (b), respectively. Dots identify the minima on the FES and are colored based on the microstate (cluster) they refer to.

Figure 7

First two principal components (PCA1 and PCA2) obtained from dimensionality reduction of the medium-range SOAP feature space of the probe 7 in heteroligand bundled NP3/6 and NP4/6 (a) and isotropic NP1/6, NP2/6, and NP5/6 (c) monolayers. Dots are colored according to the clusterization obtained by the GMM analysis. For each cluster, the inset shows the molecular environment centered on the probe 7, as extracted from the corresponding MD frames. Color legend: probe, same color of the cluster; ligands 1–5 colored in gray; ligand 6 colored in dark gray; solvent not shown for clarity. (b) Free energy surface (FES) (kcal/mol) calculated from the state’s probability distribution for NP3/6 and NP4/6 (b) and NP1/6, NP2/6, and NP5/6 (d). Dots identified the minima on the FES and are colored based on the microstate (cluster) they refer to. The arrows indicate the transition probabilities between the states from the minimum.

Figure 8

Similarity matrix for all local (most visited) environments generated by calculating the pairwise SOAP kernels K_SOAP between all the reduced short-range SOAP feature vectors. Dark blue color indicates high similarity between the environments.

Figure 9

ESR spectra of the radical probe 7 recorded in water in the presence of NP3/6 (a) and NP4/6 (b) at 300 K. In red are reported the corresponding theoretical simulations obtained by employing the spectroscopic parameters reported in Table 1.