Bridging molecular and nanometer scales in atmospheric new particle formation with quantum chemical modeling
8 affiliations
- 1Department of Chemistry, Aarhus University, Langelandsgade 140, Aarhus, Denmark, 8000
- 2Department of Chemical Engineering, Columbia University, 500 West 120th Street, New York, New York, 10027
- 3Department of Chemical and Biochemical Engineering
- 4Technical University of Denmark
- 5Swedish Meteorological and Hydrological Institute, Folkborgsvägen 17, Norrköping, Sweden, SE-60176
- 6Department of Physics, Lund University, Professorsgatan 1, Lund, Sweden, SE-22363
- 7Department of Chemical Engineering, Carnegie Mellon University, Doherty Hall, Pittsburgh, PA, USA, 15213
- 8Center for Atmospheric Particle Studies, Carnegie Mellon University, Doherty Hall 2115, Pittsburgh, PA, USA, 15213
Abstract
New particle formation (NPF) is important for our global climate and is assessed to contribute to up to half of the cloud condensation nuclei. However, there has been a knowledge gap in the size ranges that can be described accurately by quantum chemical methods and routinely measured by experimental techniques in the field. Based on quantum chemical calculations, boosted by machine learning (ML) techniques, we here present a novel framework for accurate modeling of the cluster formation dynamics directly from single molecules up to measurable nanoparticle sizes. Using sulfuric acid–base clustering as a proof-of-concept, we obtained temperaturedependent thermochemistry providing a consistent molecular description of nucleation and early growth up to approximately 2 nm, within the cluster-to-particle transition regime. The newly calculated data was used in the Atmospheric Cluster Dynamics Code (ACDC) to compare formation rates measured in CLOUD chamber, flow reactor experiments and in the field. Time-resolved simulations also allow direct comparison with 1.6 nm particle concentrations measured by a CPC FARM, in Pittsburgh, USA. This is the first time that the measured nanoparticles are explicitly modeled at molecular resolution without artificial splitting of “nucleation” and “growth”. Furthermore, instead of relying on nucleation potentials, the temporal evolution of chemical composition in FNPs can be tracked by timeseries simulation of actual nanoparticle concentrations. We couple our quantum chemical data to the aerosol transport and gas phase chemistry model (ADCHEM) and further demonstrate that we can reproduce trends of the nanoparticle concentrations observed in Hyltemossa, Sweden and verified the inconsistency caused by previous QC data, which contains only small clusters. The proposed methodology can be extended to other environments and mechanisms, provided that the necessary quantum chemical data are available or can be computed.
Related
- Phase transitions in multicomponent systems relevant to atmospheric new particle formation
- The representativeness of observed Δt in calculating formation and growth rates of newly formed atmospheric particles – hypothesis and examination
- Single-Particle Microplastics Analysis via Laser-Assisted MicroPyrolysis Flowing Atmospheric-Pressure Afterglow Mass Spectrometry Imaging.
- Comparable Trifluoroacetic Acid Deposition in Guyana and Canada: Evidence of Atmospheric Formation and Remobilization
- On the Timescales of Atmospheric Adjustments to Contrail Cirrus Formation
BibTeX
@misc{wu2026bridging,
title = {{Bridging molecular and nanometer scales in atmospheric new particle formation with quantum chemical modeling}},
author = {Haide Wu and Morten Engsvang and Darren Cheng and Yosef Knattrup and Jakub Kubečka and Tinja Olenius and Pontus Roldin and Coty N. Jen and Jonas Elm},
howpublished = {ChemRxiv},
year = {2026},
month = oct,
doi = {10.26434/chemrxiv.15010418/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15010418/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Bridging molecular and nanometer scales in atmospheric new particle formation with quantum chemical modeling
AU - Haide Wu
AU - Morten Engsvang
AU - Darren Cheng
AU - Yosef Knattrup
AU - Jakub Kubečka
AU - Tinja Olenius
AU - Pontus Roldin
AU - Coty N. Jen
AU - Jonas Elm
PY - 2026
DA - 2026/10/13
PB - ChemRxiv
DO - 10.26434/chemrxiv.15010418/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15010418/v1
AB - New particle formation (NPF) is important for our global climate and is assessed to contribute to up to half of the cloud condensation nuclei. However, there has been a knowledge gap in the size ranges that can be described accurately by quantum chemical methods and routinely measured by experimental techniques in the field. Based on quantum chemical calculations, boosted by machine learning (ML) techniques, we here present a novel framework for accurate modeling of the cluster formation dynamics directly from single molecules up to measurable nanoparticle sizes. Using sulfuric acid–base clustering as a proof-of-concept, we obtained temperaturedependent thermochemistry providing a consistent molecular description of nucleation and early growth up to approximately 2 nm, within the cluster-to-particle transition regime. The newly calculated data was used in the Atmospheric Cluster Dynamics Code (ACDC) to compare formation rates measured in CLOUD chamber, flow reactor experiments and in the field. Time-resolved simulations also allow direct comparison with 1.6 nm particle concentrations measured by a CPC FARM, in Pittsburgh, USA. This is the first time that the measured nanoparticles are explicitly modeled at molecular resolution without artificial splitting of “nucleation” and “growth”. Furthermore, instead of relying on nucleation potentials, the temporal evolution of chemical composition in FNPs can be tracked by timeseries simulation of actual nanoparticle concentrations. We couple our quantum chemical data to the aerosol transport and gas phase chemistry model (ADCHEM) and further demonstrate that we can reproduce trends of the nanoparticle concentrations observed in Hyltemossa, Sweden and verified the inconsistency caused by previous QC data, which contains only small clusters. The proposed methodology can be extended to other environments and mechanisms, provided that the necessary quantum chemical data are available or can be computed.
ER -