Abstract
The optical properties of metallic nanoparticles with nanometre dimensions exhibit features that cannot be described by classical electrodynamics. In this quantum size regime, the nearfield properties are significantly modified and depend strongly on the geometric arrangements. However, simulating realistically sized systems while retaining the atomistic description remains computationally intractable for fully quantum mechanical approaches. Here we introduce an atomistic electrodynamics model where the traditional description of nanoparticles in terms of a macroscopic homogenous dielectric constant is replaced by an atomic representation with dielectric properties that depend on the local chemical environment. This model provides a unified description of bare and ligandcoated nanoparticles, as well as strongly interacting nanoparticle dimer systems. The nonlocal screening owing to an inhomogeneous ligand layer is shown to drastically modify the nearfield properties. This will be important to consider in optimization of plasmonic nanostructures for nearfield spectroscopy and sensing applications.
Similar content being viewed by others
Introduction
The ability to localize and manipulate light at the nanoscale well below the diffraction limit makes plasmonic nanoparticles unique. Hence, plasmonic nanoparticles offer advantages in a large array of applications such as energy harvesting^{1}, plasmonassisted chemistry^{2,3}, chemical sensing^{4} and using the plasmonic nearfield to enhance molecular spectroscopy^{5}. While classical electrodynamics simulations are typically used to explain the optical properties of the plasmonic nanoparticles, recent works^{6,7,8,9,10} have clearly demonstrated that this description fails for small nanoparticles and aggregates with small separations.
For individual nanoparticles in this quantum size regime (diameter<10â€‰nm), a significant blue shift of the plasmon resonance is observed in contrast to classical predictions^{7,11,12}. This blue shift arises from the smearing of the electronic charge distribution over the particle surface and is accounted for both in quantum mechanical^{13} and nonlocal electrodynamics^{14} models. Nanoparticle dimers with separations less than their radius are of particular interest owing to a plasmon shift and a significantly enhanced electromagnetic field in the junction arising from strong coupling. The large field is utilized in surfaceenhanced spectroscopy. For nanoparticle dimers, the quantum size regime occurs when the separation becomes less than âˆ¼1â€‰nm (ref. 6). At these small distances the electronic charge distribution between the two nanoparticles overlaps, leading to a reduction of the field in the junction and a smooth transition between touching and nontouching particles^{10}. This puts a fundamental limit on how large the field enhancements in the junction can be^{8,9,15,16}. These effects are correctly described using timedependent densityfunctional theory (TDDFT) simulations and are interpreted as the effect of electron tunnelling between the two particles. A recently developed quantumcorrected electrodynamics (QCE) model that incorporates the gap tunnelling into classical simulations has been shown excellent agreement with the TDDFT simulations^{10}. The generalized nonlocal optical response (GNOR) model correctly produces the smooth transition in the contact region by accounting for the electron diffusion at the nanoparticleâ€™s surface^{14}.
Much effort has been focused on understanding the quantum size effects of bare particles. However, recent work has shown that molecules attached to the surface significantly alter the plasmon in the quantum regime. A reversal of the expected blue shift was observed for small ligandcovered particles^{17}. Furthermore, it has been shown that molecules placed in the junction between two nanoparticles can tune the plasmon by modulating the tunnelling between the particles^{18}. Clearly it is important to understand the optical response of the particles in the quantum regime accounting for the interactions with the ligands. Quantum mechanical simulations can describe ligandcovered particles as large as 2â€‰nm (sufficiently large enough to support plasmon excitations), but remain computationally intractable for larger systems^{19}. Both the QCE and GNOR models treat the ligand layer as a homogeneous environment described by a local dielectric constant and thus cannot account for the specific interactions with the metal surface.
In this work we introduce an atomistic electrodynamics model for the simulations of bare and ligandcoated nanoparticles in the quantum size regime that naturally incorporates the effects of the local chemical environment through atomic coordination dependence. In this coordinationdependent discrete interaction model (cdDIM), the nanoparticle is described by a collection of atoms that are assigned an atomic polarizability representing a spherical charge distribution with properties that depend on its local environment. Thus, cdDIM can be considered an atomistic variant of the popular discrete dipole approximation (DDA) which is commonly used to simulate plasmonic properties of large metallic nanoparticles^{20,21}. Using cdDIM, we present a unified description of both bare and ligandcoated nanoparticles in the quantum size regime. We show that this model correctly describes the strongly interacting nanoparticle dimer system, including the reduction of the nearfield in the junction, in good agreement with recent quantum mechanical simulations. This opens up the ability to perform atomistic simulations of the optical properties of realistic inhomogeneous plasmonic systems. This will be important for further optimization of plasmonic nanostructures for nearfield spectroscopy and sensing applications.
Results
The coordinationdependent discrete interaction model
The central idea of cdDIM is illustrated in Fig. 1, where the traditional description of nanoparticles in terms of a macroscopic homogenous dielectric constant is replaced by an atomic representation with dielectric properties that depend on the local chemical environment. In this way, we can differentiate between surface and bulk atoms as well as local chemical interactions with the ligands, here oleylamine (OAm), which binds through the nitrogen atom. In cdDIM, the total energy of a set of interacting fluctuating dipoles is given by
where the first term is ascribed to the selfenergy, the second term is resposible for the dipoleâ€“dipole interaction energy and the last term arises from the interactions with the external field. Here Î±_{I,Î±Î²} represents a component of the atomic polarizability tensors of atom I and is the secondorder interaction tensor describing the interactions between dipole I and J. In cdDIM, Gaussian charge distribution is used to describe the polarizable sphere. The interaction tensor is therefore renormalized which effectively screens the interactions at short distances^{22,23}. In contrast, the DDA model treats the atoms as point objects and thus the bare unscreened interaction tensor is used. Retardation effects are not included because of the small size of the nanoparticles studied; however, this could be straightforwardly included using the fully retarded interactions tensor as is common in the DDA method^{20,21}. The total polarizability can then be found by minimizing the total energy with respect to the induced atomic dipoles which leads to the following set of linear response equations
which are solved selfconsistently using an iterative solver combined with a fast multilevel cellâ€”multipole matrixâ€”vector multiplication scheme^{24}. The total polarizability is then found as
from which the optical absorption across section at a frequency Ï‰ can be obtained as
where n is the refractive index of surrounding medium and is the isotropic polarizability of the whole system. The electric field around the nanoparticle can be obtained directly from the atomic dipoles using either the renormalized interaction tensor with a probe radius of 1.58â€‰Ã… for cdDIM or the bare interaction tensor for DDA. The atomic polarizability is taken to be isotropic and obtained from a Clausiusâ€“Mossotti relation as
where R_{I}(X) is the coordinationdependent radii of the atom, Îµ(X) is the coordinationdependent dielectric constant of the material and Îµ_{0} is the dielectric constant of the environment. The specific parametrization of the coordinationdependent polarizability and dielectric constant is given in the Methods section.
Plasmonic response of bare Ag nanoparticles
Quantifying the optical properties of metallic nanoparticles with diameters <10â€‰nm is difficult because of the strong dependence on size, shape and local environment for such small nanoparticles. Recent works on bare nanoparticles using scanning transmission electron microscopy electron energyloss spectroscopy have demonstrated a 0.5â€‰eV blue shift of the plasmon excitation as the size of the nanoparticle decreases from 10 to 2â€‰nm (refs 7, 12). This blue shift is consistent with that found in small, molecularlike Ag clusters^{11}. Simulations based on quantum mechanical^{25} and nonlocal electrodynamics^{26} models have shown that the origin of this blue shift can be ascribed to a smearing of the charge density at the surface with the charge centroid pushed towards the particle centre due to the screening from the bound delectrons^{26}.
By differentiating between surface and bulk atoms, we show that cdDIM correctly describes the blue shift of the plasmon excitation as the size decreases. In Fig. 2a, the absorption spectra of icosahedral nanoparticles simulated using cdDIM are compared with the experimental data obtained for small spherical particles using electron energyloss spectroscopy^{7}. Good agreement with the experimental results is found, especially for diameters <6â€‰nm. For the larger diameters, we see that the plasmon excitation is slightly larger than the experimental results. The observed blue shift is the result of a combination of spillout effects and quantum size effects incorporated into the coordinationdependent dielectric constants. In cdDIM, each atom is described by a spherical charge distribution with a radius that depends on the local coordination environment. Atoms at the surface are assigned a larger radius than atoms in the bulk to account for spillout effects at the surface. The dipoleâ€“dipole interactions between the atoms are renormalized owing to the smearing of the charge distribution which leads to a blue shift of the plasmon excitation. Furthermore, in cdDIM the surface atoms are also described by a modified dielectric constant, where the plasmon frequency is pushed to higher energies to reflect the less freeelectron behaviour at the surface. This also contributes to a blue shift of the plasmon excitation. In contrast, when the atoms are treated as pointlike objects described by a local bulk dielectric constant, the expected sizeindependent plasmon excitation is found as illustrated in Fig. 2b. For comparison, results from the quasistatic limit of the GNOR model which incorporated nonlocal effects are shown in Fig. 2c which are in good agreement with the experimental data over the whole size range. Both the results from the GNOR model and cdDIM find a somewhat smaller blue shift than experiments find. Substrate effects have been ruled out as the origin for this discrepancy^{27}. Finally, the cdDIM results for the larger nanoparticles are a little higher in energy than both the GNOR results and what is expected from classical electrodynamics. This is likely because of the parametrization of the radii in terms of the static polarizabilities against TDDFT data for small metal clusters, see Supplementary Fig. 1 and Supplementary Methods for details.
Plasmonic response of bare Ag nanoparticle dimers
Two strongly coupled plasmonic dimers separated by a small distance is another example where classical electrodynamics is known to fail in describing the optical properties. Here we show that cdDIM correctly describes the physics of this system even for very short separations owing to the inclusion of nonlocal screening. The absorption spectra of two icosahedral Ag nanoparticles with a diameter of 4.6â€‰nm are shown in Fig. 3. The two nanoparticles are oriented such that their vertices are aligned along the axis of separation, and the distance refers to the separation between the two Ag vertex atoms, as demonstrated in Fig. 3a. For separations >5â€‰Ã…, we are in the noncontact regime where classical electrodynamics is valid, whereas for distances smaller than that we are in the bonding regime. For zero separation, the nanoparticles share a common vertex atom and start to form a single unified nanoparticle for which a classical description again becomes valid. It is in the intermediate regime where nonclassical behaviour is observed and where we see the largest difference between the cdDIM and DDA results. The results obtained using cdDIM show a gradual redshifting and weakening of the main dipolar plasmon around 3.7â€‰eV, a smooth transition between the quadrupolar plasmon and the highenergy chargetransfer plasmon at short distances, and the emergence of a lowenergy chargetransfer plasmon around 2.5â€‰eV. This result is in good agreement with the recent TDDFT, QCE and GNOR predictions. In contrast to this, the DDA results show a more discontinuous transition in the bonding regime with the lowest chargetransfer plasmon only emerging when the nanoparticles overlap significantly. In cdDIM the emergence of the chargetransfer plasmon in the contact region is captured through the coordination dependence of the atomic polarizabilities. As the distance between the two nanoparticles in the dimer gets smaller the nature of the vertex atom changes continuously from being surfacelike to being bulklike. This change in character of the atom leads to both a change in the spillout effect through a change in the atomic radii and a change in its dielectric constant to make it more freeelectron like. These effects combined is what leads to this improved description at small distances and the correct description of the chargetransfer plasmon.
In addition, the magnitude of the electric field enhancement in the centre of the gap depends strongly on the separation as shown in Fig. 3d,e for cdDIM and DDA, respectively. For cdDIM, the largest field enhancements are found around 5â€‰Ã… in good agreement with predictions from TDDFT and QCE. The field in the junction is completely screened for the shortest separations due to the smearing of the charge distribution in the cdDIM model. The results obtained using DDA show a completely different picture with the strongest enhancements at the shortest distances. Furthermore, in DDA the enhancements track the dipole plasmon leading to the largest fields around 2â€‰eV, and the enhancements are much larger than those predicted from cdDIM. While the GNOR model correctly predicts a screening of the electromagnetic field in the junction, the field enhancements still track the dipolar plasmon and becomes the strongest at the shortest distances, see Supplementary Fig. 2. This is owing to the neglect of the overlap of the spillout charge distribution in the GNOR model which becomes important for distance below 5â€‰Ã… (ref. 28). Overall, the cdDIM succeeds in describing the optical properties of strongly coupled metallic dimers with subnanometre gaps in good agreement with the accurate predictions based on TDDFT.
Plasmonic response of ligandcoated Ag nanoparticles
In contrast to the large amount of simulations carried out for bare nanoparticles, much less has been done for the ligandcoated nanoparticles. The ligand layer is typically described as a homogenous layer with a small real refractive index. However, recent work has shown that a OAm ligand layer can cause a reversal of the blue shift in the small bare Ag nanoparticles^{17}. This reversal was attributed to a reduction in the conductivity of the outer layer of metal atoms caused by the binding ligands. Simulations based on a multilayer Mie theory where the outer layer was described using a modified dielectric constant with reduced conductivity were shown to agree with the experiments. So far simulations that provide a realistic description of the ligand layer have not been pursued beyond TDDFT simulations of small monolayerprotected metal clusters^{19}.
In cdDIM the interactions between the ligands and Ag atoms on the surface of the metallic nanoparticles cause a local modification of the atomic polarizability which leads to the redshifting of the plasmon response for small nanoparticles. The interactions with the ligands cause a reduced conductivity due to the formation of the chemical bond^{17}. This is illustrated in Fig. 4a where we plot the plasmonic absorption spectra of various coated Ag nanoparticles with diameters ranging from 2.3 to 12.1â€‰nm obtained from cdDIM calculations. For comparison, the experimental results are shown in Fig. 4b. Overall we see good agreement between the cdDIM results and the experimental observations in spite of a few slight differences. The simulations predict a 0.42â€‰eV shift, whereas a slightly smaller shift of 0.3â€‰eV is found in the experiments. Also, we see that the experimental spectra are broader compared with the simulations especially for the smaller particles, see Supplementary Fig. 3. This is likely caused by experimental deviations from a perfect icosahedral geometry. We can eliminate the discrepancy by applying a slightly increased sizedependent damping parameter, however, we did not choose this to preserve the weaker bands on either side of the main peak due to the perfect icosahedral structure used in the simulation.
The main advantage of the cdDIM model is that it retains the atomistic description of the nanoparticle and the ligand layer. This allows us to understand how the ligand layer perturbs the electric field around the nanoparticle under the nonlocal screening effects. Typically the ligand layer is treated as a homogenous medium described by a local dielectric constant where the inhomogeneous nature of the ligand layer and the nonlocal screening have not been considered. Here we examined this by considering the electric field around a small nanoparticle as a function of the ligand coverage ranging from 0.8 to 7.2 molecules per nm^{2}. The results in Fig. 5 clearly show that as the ligand coverage increases, the local field around the nanoparticle is screened in the vicinity of the ligands. Once a monolayer of ligands is formed, the field is significantly perturbed and looks nothing like the initial field around the bare particle. Since the nanoparticle is small, the dimensions of the ligand layer are comparable with the nanoparticleâ€™s size. This causes the ligand layer to be highly disordered which is then reflected in the field distribution around the nanoparticle.
For larger nanoparticles one could envision that the field would be less perturbed by the presence of the ligand layer due to the larger relative dimensions of the nanoparticle and that the ligand layer is more homogeneous. In Fig. 6 we examined this by considering the nearfield around a bare and ligandcoated icosahedral nanoparticle with a diameter of 12.1â€‰nm with the incident electric field polarized along the long axis. Even though the nearfield now extends beyond the ligand layer, it is still significantly modified in the vicinity of the surface. It is especially noteworthy that the nearfield is significantly decreased around the nanoparticle tips. Consequently, the regularity of the electric field enhancement found for bare icosahedral Ag nanoparticle is not preserved in coated nanoparticle which is likely to affect the nearfield coupling between nanoparticles that are used in sensing applications.
Discussion
In summary, we have presented an atomistic electrodynamics model that provides a unified description of bare and ligandcoated nanoparticles, as well as strongly interacting nanoparticle dimer system in the quantum size regime where traditional classical models fail. We have shown that the model describes the blue shift of the plasmon in small Ag nanoparticles, the reversal of this effect due to interactions with ligands, and the reduction of the nearfield in strongly coupled nanoparticle dimer system. Furthermore, we have demonstrated that nonlocal screening due to an inhomogeneous ligand layer around the nanoparticle leads to a significant modification of the nearfield. This will be important to consider in applications that rely on the strong interactions between nanoparticles coated by inhomogeneous ligand layers such as plasmonic rulers^{29} and the recent demonstration of molecular tuning of quantum plasmon resonances^{18,30}. Furthermore, we expect that the atomistic description combined with the nonlocal screening will be necessary for understanding surfaceenhanced spectroscopies as they push towards subnm resolution^{31}.
Methods
Paremeterization of cdDIM
To account for the different environments of the Ag atoms in the nanoparticle (that is, surface, edge, vertex and core), the effective coordinationdependent radius is introduced^{32}
where
where CN_{max} is the maximum coordination numbers for the atom type (CN_{max}=12 for Ag), and CN_{m} is the effective coordination number for DIM atom m. The effective coordination number is calculated from^{32}
where the cutoff function is given by
which smoothly connects atoms inside and outside of the coordination sphere. The advantage of using an effective coordination number compared with a standard coordination number (that is, integer value) is that the effective coordination number can account for defects in the nanoparticle, or the static case we use Îµ=âˆž and thus the polarizability reduces to the volume of the atom as expected for a perfect conductor. The following atomic radii were used, R_{Ag,bulk}=1.56â€‰Ã…, R_{Ag,surf}=1.65â€‰Ã…, R_{C}=0.82â€‰Ã…, R_{N}=0.76â€‰Ã… and R_{H}=0.63â€‰Ã…. They were obtained by parameterizing the static polarizability as described in the Supplementary Method and Supplementary Fig. 4.
Within the cdDIM framework, the coordinationdependent dielectric constant is represented as
where Îµ_{exp} is taken from Johnson and Christy^{33}. In equation 10, the Drude function is given by
and
where a is the radius of the particle, and
with Ï‰_{p,surf}=22.0â€‰eV, and Ï‰_{p,bulk}=9.6â€‰eV. The g factor accounts for the lowering of the optical conductivity owing to the interactions with the ligand layer^{17} and is given by
where CN_{Mâˆ’L} represents the coordination number of metal atoms binding to the atoms in the ligands. In this work, the parameter f is given as 0.7, leading to gâˆ¼0.8 for a monolayer coverage in agreement with the previous work based on a multilayer Mie model^{17}. A sizedependent damping correction is adopted where Î½_{F} represents the Fermi velocity and
with Î³_{surf}=2.0â€‰eV and Î³_{bulk}=0.0228, eV.
GNOR simulations
The nonlocal polarizabilities for spheres of various radii were calculated within the quasistatic limit of the GNOR model as^{14}
where the nonlocal correction factor is given by
with the nonlocal factor given by
where the hydrodynamic parameter is given by , and the diffusion parameter D was fitted such that the full width at half maximum (FWHM) of a single particle using GNOR matched the FWHM using the localresponse approximation with a sizedependent Drude damping parameter^{34}. The bound dielectric term (É›_{âˆž}) was described using a frequencydependent polynomial form found in ref. 12 but rescaled by 0.948 to better match the Johnson and Christy^{33} dielectric for Ag. The background dielectric (É›_{b}) was 1.0, and the Fermi velocity of Ag (v_{F}) was 1.39 Ã— 10^{6}â€‰mâ€‰s^{âˆ’1} (ref. 35).
Molecular dynamics simulations
We performed molecular dynamics (MD) simulations using the Largescale Atomic/molecular Massively Parallel Simulator (LAMMPS, http://lammps.sandia.gov.)^{36} to obtain the structures of the Ag nanoparticle coated by the OAm ligands. The initial structure of the OAm ligands were chosen to be perpendicular to the Ag surface forming a compact monolayer, and then the coated nanoparticle was immersed into a cubic box filled with OAm molecules. The initial surface coverage was set to be 5.5 molecules per nm^{2} (ref. 37). For the MD forcefield we used the General AMBER force field^{38,39} to describe the intramolecular interactions of OAm ligands. To describe the interactions between the ligands and the nanoparticle we adopted a modified AgOAm embeddedatommethod potential as described in the Supplementary Methods, Supplementary Figs 5â€“6 and Supplementary Table 1. After the energy minimization, the whole system was slowly heated up and equilibrated at 100â€‰K in order to yield a condensed ligand zone around the Ag cluster. Then, the temperature was increased from 100 to 800â€‰K and annealed to 300â€‰K gradually leading to the desorption of the excess ligands away from the Ag nanoparticle surface. Finally, the whole system was allowed to equilibrate at 300â€‰K for 1â€‰ns with an integration time step of 1â€‰fs. The Ag nanoparticle was kept frozen during the MD simulations to reduce computational cost.
Additional information
How to cite this article: Chen, X. et al. Atomistic electrodynamics simulations of bare and ligandcoated nanoparticles in the quantum size regime. Nat. Commun. 6:8921 doi: 10.1038/ncomms9921 (2015).
References
Atwater, H. A. & Polman, A. Plasmonics for improved photovoltaic devices. Nat. Mater. 9, 205â€“213 (2010).
Christopher, P., Xin, H. & Linic, S. Visiblelightenhanced catalytic oxidation reactions on plasmonic silver nanostructures. Nat. Chem. 3, 467â€“472 (2011).
Mukherjee, S. et al. Hot electrons do the impossible: plasmoninduced dissociation of H2 on Au. Nano Lett. 13, 240â€“247 (2013).
Anker, J. N. et al. Biosensing with plasmonic nanosensors. Nat. Mater. 7, 442â€“453 (2008).
Willets, K. A. & Van Duyne, R. P. Localized surface plasmon resonance spectroscopy and sensing. Ann. Rev. Phys. Chem. 58, 267â€“297 (2007).
Halas, N. J., Lal, S., Chang, W.S., Link, S. & Nordlander, P. Plasmons in strongly coupled metallic nanostructures. Chem. Rev. 111, 3913â€“3961 (2011).
Scholl, J. A., Koh, A. L. & Dionne, J. A. Quantum plasmon resonances of individual metallic nanoparticles. Nature 483, 421â€“427 (2012).
Zhu, W. & Crozier, K. B. Quantum mechanical limit to plasmonic enhancement as observed by surfaceenhanced Raman scattering. Nat. Commun. 5, 5228 (2014).
Savage, K. J. et al. Revealing the quantum regime in tunnelling plasmonics. Nature 491, 574â€“577 (2012).
Esteban, R., Borisov, A. G., Nordlander, P. & Aizpurua, J. Bridging quantum and classical plasmonics with a quantumcorrected model. Nat. Commun. 3, 825 (2012).
TiggesbÃ¤umker, J., KÃ¶ller, L., MeiwesBroer, K.H. & Liebsch, A. Blue shift of the mie plasma frequency in ag clusters and particles. Phys. Rev. A 48, R1749â€“R1752 (1993).
Raza, S. et al. Blueshift of the surface plasmon resonance in silver nanoparticles studied with eels. Nanophotonics 2, 131â€“138 (2013).
Liebsch, A. Surfaceplasmon dispersion and size dependence of mie resonance: silver versus simple metals. Phys. Rev. B 48, 11317â€“11328 (1993).
Mortensen, N. A., Raza, S., Wubs, M., SÃ¸ndergaard, T. & Bozhevolnyi, S. I. A generalized nonlocal optical response theory for plasmonic nanostructures. Nat. Commun. 5, 3809 (2014).
Hajisalem, G., Nezami, M. S. & Gordon, R. Probing the quantum tunneling limit of plasmonic enhancement by third harmonic generation. Nano Lett. 14, 6651â€“6654 (2014).
CiracÃ, C. et al. Probing the ultimate limits of plasmonic enhancement. Science 337, 1072â€“1074 (2012).
Peng, S., McMahon, J. M., Schatz, G. C., Gray, S. K. & Sun, Y. Reversing the sizedependence of surface plasmon resonances. Proc. Natl Acad. Sci. USA 107, 14530â€“14534 (2010).
Tan, S. F. et al. Quantum plasmon resonances controlled by molecular tunnel junctions. Science 343, 1496â€“1499 (2014).
Malola, S., Lehtovaara, L., Enkovaara, J. & HÃ¤kkinen, H. Birth of the localized surface plasmon resonance in monolayerprotected gold nanoclusters. ACS Nano 7, 10263â€“10270 (2013).
Purcell, E. M. & Pennypacker, C. R. Scattering and absorption of light by nonspherical dielectric grains. Astrophys. J. 186, 705â€“714 (1973).
Draine, B. T. & Flatau, P. J. Discretedipole approximation for scattering calculations. J. Opt. Soc. Am. A 11, 1491â€“1499 (1994).
Mayer, A. Formulation in terms of normalized propagators of a chargedipole model enabling the calculation of the polarization properties of fullerenes and carbon nanotubes. Phys. Rev. B 75, 045407 (2007).
Jensen, L. L. & Jensen, L. Atomistic electrodynamics model for optical properties of silver nanoclusters. J. Phys. Chem. C 113, 15182â€“15190 (2009).
Payton, J. L., Morton, S. M., Moore, J. E. & Jensen, L. A hybrid atomistic electrodynamicsquantum mechanical approach for simulating surfaceenhanced raman scattering. Acc. Chem. Res. 47, 88â€“99 (2014).
Kuisma, M. et al. Localized surface plasmon resonance in silver nanoparticles: Atomistic firstprinciples timedependent densityfunctional theory calculations. Phys. Rev. B 91, 115431 (2015).
Toscano, G. et al. Resonance shifts and spillout effects in selfconsistent hydrodynamic nanoplasmonics. Nat. Commun. 6, 7132 (2015).
Raza, S., Yan, W., Stenger, N., Wubs, M. & Mortensen, N. A. Blueshift of the surface plasmon resonance in silver nanoparticle: substrate effects. Opt. Express 21, 27344 (2013).
Raza, S., Bozhevolnyi, S. I., Wubs, M. & Mortensen, N. A. Nonlocal optical response in metallic nanostructures. J. Phys. Condens. Matter 27, 183204 (2015).
SÃ¶nnichsen, C., Reinhard, B. M., Liphardt, J. & Alivisatos, A. P. A molecular ruler based on plasmon coupling of single gold and silver nanoparticles. Nat. Biotechnol. 23, 741â€“745 (2005).
Nordlander, P. Molecular tuning of quantum plasmon resonances. Science 343, 1444â€“1445 (2014).
Zhang, R. et al. Chemical mapping of a single molecule by plasmonenhanced raman scattering. Nature 498, 82â€“86 (2013).
Legenski, N. et al. Force fields for metallic clusters and nanoparticles. J. Comput. Chem. 32, 1711â€“1720 (2011).
Johnson, P. B. & Christy, R. W. Optical constants of the noble metals. Phys. Rev. B 6, 4370â€“4379 (1972).
SÃ¸ren, R., Martijn, W., Sergey, I. B. & Mortensen, N. A. Nonlocal study of ultimate plasmon hybridization. Opt. Lett. 40, 839â€“842 (2015).
Ashcroft, N. W. & Mermin, N. D. Solid State Physics Brooks Cole (1976).
Plimpton, S. Fast parallel algorithms for shortrange molecular dynamics. J. Comput. Phys. 117, 1â€“19 (1995).
Hinterwirth, H. et al. Quantifying thiol ligand density of selfassembled monolayers on gold nanoparticle by inductively coupled plasmamass spectrometry. ACS Nano 7, 1129â€“1136 (2013).
Wang, J., Wolf, R. M., Caldwell, J. W., Kollman, P. A. & Case, D. A. Development and testing of a general amber force field. J. Comput. Chem. 25, 1157â€“1174 (2004).
Wang, J., Wang, W., Kollman, P. A. & Case, D. A. Automatic atom type and bond type perception in molecular mechanical calculations. J. Mol. Graph. Mod. 25, 247â€“260 (2006).
Acknowledgements
This work was supported by the NSF Center for Chemical Innovation dedicated to Chemistry at the SpaceTime Limit (CHE1414466). J.M. acknowledge support from the NSF award CHE1362825, and M.Z. acknowledges support from a 3M fellowship. We acknowledge support received from Research Computing and Cyberinfrastructure, a unit of Information Technology Services at Penn State.
Author information
Authors and Affiliations
Contributions
L.J. conceived the basic idea, X.C., J.M. and M.Z. did the implementation and ran the simulations. L.J., X.C. and J.M. analysed the results and wrote the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing financial interests.
Supplementary information
Supplementary Information
Supplementary Figures 16, Supplementary Table 1, Supplementary Discussion, Supplementary Methods and Supplementary References. (PDF 1301 kb)
Rights and permissions
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articleâ€™s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
About this article
Cite this article
Chen, X., Moore, J., Zekarias, M. et al. Atomistic electrodynamics simulations of bare and ligandcoated nanoparticles in the quantum size regime. Nat Commun 6, 8921 (2015). https://doi.org/10.1038/ncomms9921
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/ncomms9921
This article is cited by

Highresolution tipenhanced Raman scattering probes submolecular density changes
Nature Communications (2019)

Multiscale modelling of photoinduced processes in composite systems
Nature Reviews Chemistry (2019)

Nonempirical atomistic dipoleinteractionmodel for quantum plasmon simulation of nanoparticles
Scientific Reports (2017)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.