Explicit-atom molecular dynamics simulations have revealed that tiny nanoparticles measuring between 0.5 and 5 nanometers in diameter diffuse through gases at rates up to an order of magnitude slower than predicted by standard aerosol equations. In both gas-phase materials processing and atmospheric physics, transport behavior at this extreme lower boundary of the nanoscale is governed by incessant collisions between host gas molecules and diffusing clusters. While modern industrial operations routinely manufacture larger nanoparticles such as fumed oxides at commercial rates exceeding several tons per hour, process design for smaller particles has long been hindered by widespread theoretical uncertainty. The newly reported computational findings, corroborated by thirty years of published experimental diffusion measurements across multiple independent laboratories, demonstrate that the classical Stokes-Einstein-Cunningham and Epstein formulations substantially overestimate particle mobility because they treat carrier gas molecules and clusters as idealized hard spheres. By explicitly accounting for both attractive and repulsive intermolecular potentials alongside the specific geometric shapes of diatomic gas molecules and nanoparticles, researchers have established that the effective mean free path of air is nearly half of the value traditionally cited in standard textbooks, fundamentally altering the quantitative picture of sub-five-nanometer aerosol dynamics.
Historically, researchers relied on classical kinetic theory to model aerosol transport, assuming that collisions between carrier gas molecules and tiny particles are purely elastic and specular. However, as the physical diameter of a nanoparticle approaches the molecular dimensions of the ambient medium, this hard-sphere approximation breaks down completely. The new molecular dynamics simulations show that realistic intermolecular forces create a diverse spectrum of complex collision patterns rather than simple ballistic rebounds. The attractive component of the intermolecular potential bends molecular trajectories, causing grazing deflections and capturing carrier gas molecules into multi-stage orbiting interactions. Consequently, the resistance experienced by tiny nanoparticles as they move through a carrier gas is substantially higher than previously assumed, which directly explains why observed diffusion rates consistently fall far below the predictions generated by textbook continuum and free-molecule transport equations.
To uncover the microscopic mechanisms that govern this reduced mobility, the researchers examined detailed atomistic collision trajectories between diatomic nitrogen molecules and tiny silica nanoparticles with a mass diameter of 1.5 nanometers. Under relative coordinate systems where the nanoparticle center is held stationary, the interaction trajectories departed dramatically from classical elastic scattering models. In classical kinetic formulations based on Lennard-Jones potentials, molecular collisions are assumed to occur strictly when center-to-center separations decrease below two to the one-sixth power times the gas atom diameter, a threshold defined by mutual repulsive forces. When the full atomistic structure, molecular orientation, and force fields are incorporated, gas-particle impacts become predominantly inelastic. Nitrogen molecules engaged in head-on ballistic collisions experience trajectory bending toward the nanoparticle surface due to attractive forces prior to rebound. Furthermore, the simulations revealed distinct orbiting patterns wherein an incoming nitrogen molecule undergoes as many as three successive inelastic impacts with the same nanoparticle before acquiring sufficient relative velocity to escape, all occurring without any third-body molecular intervention.
The detailed trajectory analyses also documented grazing collisions and prolonged surface residency phenomena that cannot occur within classical hard-sphere frameworks. In grazing interactions, fast-moving nitrogen molecules undergo gradual trajectory deflections induced solely by the attractive portion of the intermolecular potential, while repulsive forces exert negligible influence on the encounter path. Extending the molecular dynamics simulations of a 1.5-nanometer silica nanoparticle diffusing in ambient air over a continuous duration of 1.42 nanoseconds revealed frequent multi-body collisions, wherein two or more gas molecules strike the nanoparticle surface at the exact same moment, observed at discrete simulation timestamps of 0.72, 1.22, and 1.42 nanoseconds. Furthermore, the attractive intermolecular force field caused gas molecules to become transiently adsorbed onto the nanoparticle surface for durations spanning several hundreds of picoseconds before desorbing back into the bulk gas phase, exemplified by an air molecule remaining bound from 1.22 to 1.42 nanoseconds.
Translational diffusion coefficients were extracted from the molecular dynamics trajectories by tracking the center-of-mass mean square displacement over time. Because individual simulation cells contained only a single nanoparticle, raw displacement trajectories exhibited considerable stochastic noise, making direct regression on single runs statistically unreliable. To resolve this statistical limitation, the investigators conducted thirty independent simulations for each particle configuration, initializing each trajectory with identical composition but distinct starting velocities and coordinates. Averaging across these tens of independent trajectories eliminated transient noise and revealed classic Fickian behavior, characterized by a smooth, strictly linear scaling between mean square displacement and elapsed simulation time. Linear regression applied to these ensemble-averaged displacement profiles yielded precise diffusion coefficients with minimal statistical uncertainty. By applying this rigorous sampling protocol, the researchers mapped out the diffusion coefficients for a broad suite of tiny nanoparticle chemistries, including silicon dioxide, titanium dioxide, and elemental zinc, across a systematic range of particle sizes under standard ambient conditions.
A systematic comparison of the molecular dynamics results with established analytical equations revealed that particle composition exerts an undeniable influence on mass transport at the lower end of the nanoscale. For nanoparticles larger than approximately 3 nanometers in diameter, the calculated diffusivities converged cleanly toward the classical Epstein and Stokes-Cunningham equations. Conversely, for particles smaller than 3 nanometers, the computed diffusivities fell well below the classical curves, exhibiting a wide compositional scatter distributed between the Epstein predictions and the empirical formulation developed by Fuller and coworkers. Fullerenes smaller than 3 nanometers aligned closely with the Fuller equation, whereas inorganic silica, titania, and zinc nanoparticles exhibited distinct, intermediate diffusion rates dictated by their unique atomistic structures and force fields. While historical aerosol literature previously suggested that clusters below 1.4 nanometers behave as rigid molecules experiencing purely elastic collisions, the molecular dynamics findings prove that tiny nanoparticles experience largely inelastic gas-surface collisions, demonstrating that accurate transport models must explicitly incorporate both the specific chemical composition of the diffusing particle and the molecular characteristics of the host gas.
The accuracy of the explicit-atom molecular dynamics methodology was validated against extensive historical experimental datasets, beginning with mobility measurements of charged fullerenes diffusing through helium gas. Previous laboratory investigations conducted by Gotts, Dugourd, Mesleh, and their respective colleagues had examined the mobility of positively and negatively charged fullerenes, including C34, C60, and C240 clusters. The published experimental data demonstrated complete overlap between positively and negatively charged species, confirming that the net sign of the electrical charge does not alter the measured diffusion coefficient in helium. The molecular dynamics calculations reproduced these empirical values across all fullerene sizes with remarkable fidelity, with statistical error margins smaller than the plotted symbol dimensions. In sharp contrast, the classical Epstein and Stokes-Cunningham equations systematically overpredicted experimental helium diffusivities by a factor of two to three, with the magnitude of theoretical overestimation growing progressively more severe as the fullerene diameter approached the molecular dimensions of the carrier gas.
Evaluating fullerene diffusion in helium also exposed the operational boundaries of empirical semi-molecular equations, most notably the correlation formulated by Fuller and coworkers. While the Fuller equation matched fullerene transport in ambient air with high precision, it systematically underpredicted experimental fullerene diffusivities in helium by 20 to 30 percent. The researchers traced this discrepancy directly to the macroscopic and microscopic thermodynamic properties of the carrier gases. Because the Fuller equation was originally parameterized using diffusion data measured exclusively in air, it inherently reflects the transport characteristics of a multicomponent gas mixture composed of diatomic nitrogen and oxygen. Air and helium possess radically different gas-particle interaction potentials, densities, viscosities, and mean free paths. While atomistic models that correctly match macroscopic air properties naturally yield accurate microscopic molecular dynamics in air, applying an air-calibrated correlation to a light, monoatomic carrier gas like helium introduces systematic errors, reinforcing the necessity of using explicit molecular force fields rather than simplified empirical adjustments.
To evaluate transport across different carrier gas chemistries, the researchers compared molecular dynamics calculations against molecular beam measurements of fullerene diffusion in pure nitrogen gas reported by Weis and coworkers. The experimental beam studies, originally performed at ambient conditions to extract collision cross sections and calibrate trajectory models, showed exceptional agreement with the explicit-atom molecular dynamics results. Furthermore, predictions from the Fuller equation closely tracked both the experimental measurements and the simulation points in nitrogen. For a C60 fullerene, the calculated diffusivity in pure nitrogen was approximately 6.5 percent lower than its corresponding diffusivity in air, a minor shift that fell entirely within the simulation’s statistical uncertainty bounds. This close alignment is physically consistent with the shared diatomic structures and nearly identical ambient transport properties of air and pure nitrogen, which differ by less than 3.5 percent across bulk viscosity, density, and gas diffusivity, contrasting sharply with the classical Epstein equation, which again overestimated experimental nitrogen diffusivities by a factor of two to three.
The validity of the molecular dynamics framework was further demonstrated across elevated temperatures and reduced pressures using aerosol mobility data reported by Tanaka and Takeuchi. That experimental study measured the diffusion of C60 and C70 fullerenes in nitrogen at a sub-atmospheric pressure of 0.267 atmospheres and temperatures of 673 and 723 Kelvin to develop standardized mobility benchmarks for sub-nanometer aerosols. In agreement with fundamental kinetic principles, both measurements and atomistic simulations documented substantial increases in diffusion coefficients with rising thermal energy. At 723 Kelvin, the molecular dynamics diffusivities matched the experimental data points perfectly, while at 673 Kelvin the computed values remained in exceptionally close agreement with the measured points. The Fuller equation likewise maintained respectable consistency with the high-temperature nitrogen data, whereas the classical Epstein and Stokes-Cunningham expressions failed completely, overpredicting the experimental mobility of both fullerenes by broad margins across every tested thermal condition.
Thermal robustness was tested across an even wider temperature envelope by benchmarking simulated C60 transport against cryogenic and room-temperature molecular beam measurements in helium published by Mesleh and colleagues. Spanning an operational temperature range from 70 Kelvin up to 370 Kelvin, the experimental measurements aimed to characterize fullerene structure and calibrate intermolecular interaction potentials. The molecular dynamics simulations tracked the experimental temperature dependence with outstanding accuracy across the entire 300-Kelvin domain, confirming the stability and physical realism of the atomistic force fields. While the Fuller equation exhibited a consistent 5 to 30 percent underprediction due to carrier gas differences, the classical Epstein and Stokes-Cunningham equations displayed severe, temperature-dependent divergence.
By unifying three decades of experimental mobility measurements with explicit-atom molecular dynamics, the study provides a rigorous physical foundation for analyzing and engineering tiny aerosol particles. Rather than behaving as inert elastic spheres or rigid frozen clusters, sub-five-nanometer nanoparticles undergo frequent inelastic, orbiting, and multi-body collisions with carrier gas molecules, resulting in diffusion coefficients that can be up to tenfold lower than classical aerosol textbooks claim. Furthermore, the demonstrated dependence of diffusion on particle chemical composition establishes that accurate engineering models must account for both particle and gas force fields. This validated atomistic approach provides the quantitative precision needed to model critical nucleation and early-stage growth mechanisms in materials synthesis reactors, bringing the rational process design of tiny functional nanoparticles to the same high level of scientific maturity currently enjoyed by industrial microscale and large-nanoscale manufacturing.
Subject of Research: Gas-phase diffusion dynamics and collision mechanisms of tiny sub-5-nanometer nanoparticles
Article Title: Pushing the limit of aerosol nanoscale engineering: tiny nanoparticle diffusion
Article References: Tsalikis, D., & Pratsinis, S. E. (2026). Pushing the limit of aerosol nanoscale engineering: tiny nanoparticle diffusion. Journal of Nanoparticle Research, 28(12), Article 307. https://doi.org/10.1007/s11051-026-06810-y
Image Credits: AI Generated
DOI: 10.1007/s11051-026-06810-y
Keywords: nanoparticle diffusion, tiny nanoparticles, molecular dynamics, aerosol engineering, Stokes-Cunningham equation, Epstein equation, Fuller equation, mean free path, inelastic collisions, orbiting collisions, silica nanoparticles, fullerenes, mean square displacement, Brownian motion, gas-phase synthesis, collision cross section
Cite Scienmag News
Denise Maddox. (November 21, 2026). Tiny Nanoparticles Diffuse Far Slower in Gases Than Classical Theory Predicts. Scienmag. https://scienmag.com/tiny-nanoparticles-diffuse-far-slower-in-gases-than-classical-theory-predicts/
Denise Maddox. "Tiny Nanoparticles Diffuse Far Slower in Gases Than Classical Theory Predicts." Scienmag, 21 November 2026, https://scienmag.com/tiny-nanoparticles-diffuse-far-slower-in-gases-than-classical-theory-predicts/. Accessed 21 November 2026.
Denise Maddox. "Tiny Nanoparticles Diffuse Far Slower in Gases Than Classical Theory Predicts." Scienmag. November 21, 2026. https://scienmag.com/tiny-nanoparticles-diffuse-far-slower-in-gases-than-classical-theory-predicts/








