Search

Abstract: We have applied a multistep quantum chemistry method to study the formation energetics and binding patterns of sulfuric acid–peroxodisulfuric acid–water clusters, with special focus on the O–O bridge. The length of the O–O bridge correlates linearly with the average length of S–O bonds next to it. The clustering of peroxodisulfuric acid with sulfuric acid and water is thermodynamically favorable, as is the replacement by peroxodisulfuric acid of one (but only one) of the sulfuric acid molecules in a sulfuric acid–water cluster. However, the presence of H2S2O8 does not enhance the addition of sulfuric acid to the clusters. [Copyright &y& Elsevier]

Abstract: The binding of 1-propanol to neutral and singly charged tungsten oxide seed particles was studied using quantum chemical methods. Three different density functionals and three basis sets were employed, and the results were compared with each other as well as with results previously published by other groups. Our results implicate a positive sign preference for all studied tungsten oxide species. Molecular structures obtained for pure tungsten oxide show good agreement with previous results. [ABSTRACT FROM AUTHOR]

The formation of aerosol particles in the atmosphere is driven by the gas to particle conversion of extremely low volatile organic compounds (ELVOC), organic compounds with a particularly low saturation vapor pressure (p Sat). Identifying ELVOCs and their chemical structures is both experimentally and theoretically challenging: Measuring the very low p Sat of ELVOCs is extremely difficult, and computing p Sat for these often large molecules is computationally costly. Moreover, ELVOCs are underrepresented in available datasets of atmospheric organic species, which reduces the value of statistical models built on such data. We propose an active learning (AL) approach to efficiently identify ELVOCs in a data pool of atmospheric organic species with initially unknown p Sat. We assess the performance of our AL approach by comparing it to traditional machine learning regression methods, as well as ELVOC classification based on molecular properties. AL proves to be a highly efficient method for ELVOC identification with limitations on the type of ELVOC it can identify. We also show that traditional machine learning or molecular property-based methods can be adequate tools depending on the available data and desired degree of efficiency. [Display omitted] • Sparse ELVOC data can be identified with machine learning or empirical methods. • Active machine learning is very efficient in identifying ELVOC in molecular data. • Prediction accuracy of machine learning models on ELVOC remains subpar. • Published molecular properties of 3459 ELVOC computed with DFT and their conformers. [ABSTRACT FROM AUTHOR]

Insights into the effect of temperature (T) and relative humidity (RH) as well as structure and polarisation on ion mobility help the comparison and interpretation of mobility and mass-based data. We measured alkylammonium ions in air under different T (14 °C, 24 °C, 34 °C and 41 °C) and RH (0 %, 20 %, 40 %) conditions using two individual setups (in both cases a planar differential mobility analyser coupled with a time-of-flight mass spectrometer) and the results are in excellent agreement. Mobility increases with rising T and decreases with water vapour loading. When separating the measurement mobility by structures, clear mass dependence was observed. The measured mobilities exhibited large deviations from theoretically calculated results in dry conditions, which are possibly caused by adduct formation on the monomer ions via clustering (or reactions). This phenomenon seems to be unavoidably associated with light ions under atmospheric pressures, which is worth further exploration and bearing in mind when comparing measurements to calculations. Both methanol and oxygen (occasionally nitrogen or alkyl chain elongation) are possible candidates of the adduct. Under spherical assumption, we used the modified Mason–Schamp's approximation to link the measured mobility to the mobility equivalent diameter. The drag enhancement factor ξ and the effective gas-molecule collision diameter d g derived from our measurement data are comparable to literature values. Our data also exposed a non-linear dependence on the polarisation parameter ε *. Polarisation, ξ and d g were parameterised using linear models against ion structures, T , and RH for primary, secondary and tertiary alkylammonium ions with identical alkyl groups. Our model parametrisations predict mobilities within ±10 % deviation from the measured data. The model also has satisfying predicting power for alkylammonium ions with unidentical alkyl structures. • Mobility measured by two individual setups gave good agreement under different temperature and relative humidity conditions. • Mobility increases with rising temperature and decreases with water vapour loading. • Adduct formation seems to be unavoidable for light ions under atmospheric pressures. • The derived drag enhancement factor and the effective gas-molecule collision diameter are comparable to literature values. • Parameterization against ion structures, temperature and relative humidity led to good predicting power. [ABSTRACT FROM AUTHOR]

Abstract: The binding of sulfuric acid to a series of anions and cations of varying chemical complexity is studied using quantum chemical methods. Sulfuric acid is bound much more strongly to anions than cations, as expected from structural and general chemical considerations, and previous computational studies on charged sulfuric acid–water clusters. The results are likely to explain the sign effect observed in atmospheric nucleation phenomena, and indicate that the first steps of ion-induced nucleation are controlled by the specific chemical interactions between the core ion and the condensing molecules rather than by general physical (e.g. electrostatic) effects. [Copyright &y& Elsevier]

Abstract: We used quantum chemical calculations to attempt to explain why sulfuric acid produced in situ via the reaction of OH radicals with SO2 in the presence of water vapor nucleates much more effectively than sulfuric acid evaporated from a liquid reservoir. We investigated whether or not selected SO2 oxidation intermediates or alternative reaction products could take part in nucleation. We have calculated binding energies and formation free energies for H2SO4·H2SO4, HSO3·H2SO4, HSO5·H2SO4, H2SO5·H2SO4, H2S2O8·H2SO4, H2SO4·H2SO4·H2O, HSO3·H2SO4·H2O, HSO5·H2SO4·H2O, H2SO5·H2SO4·H2O, and H2S2O8·H2SO4·H2O clusters at the RI-MP2/QZVPP level. Our calculations indicate that the only dimer cluster that is more strongly bound than H2SO4·H2SO4 is H2S2O8·H2SO4. Hydration as modeled by the addition of one water molecule to the cluster does not change the situation. The binding of these intermediates, with the exception of H2S2O8 (peroxo-disulfuric acid), to sulfuric acid is therefore not likely to explain the observations. [Copyright &y& Elsevier]

A strong long-range transported (LRT) fine particle (PM) episode occurred from March 17–22, 2002 over large areas of Finland. Most of the LRT particle mass was in the submicrometre size fraction. The number of concentrations of 90–500nm particles increased by a factor of 5.6 during the episode, but the concentrations of particles smaller than 90nm decreased. This reduction of the smallest particles was caused by suppressed gas-to-particle conversion due to the vapour uptake of LRT particles. Individual particle analyses using SEM/EDX showed that the proportion of sulphur-rich particles rose strongly during the episode and that the relative weight percentage of potassium was unusually high in these particles. The median S/K ratios of S-rich particles were 2.1 at the beginning of the episode, 5.2 at the peak stage of the episode and 8.9 during the reference days. The high proportion of K is a clear indication of emissions from biomass burning, because K is a good tracer of biomass-burning aerosols. Trajectories and satellite detections of fire areas indicated that the main source of biomass-burning aerosols was large-scale agricultural field burning in the Baltic countries, Belarus, Ukraine and Russia. The higher S/K ratio of S-rich particles during the peak stage was obviously due to the increased proportion of fossil fuel-burning emissions in the LRT particle mass, since air masses arrived from the more polluted areas of Europe at that time. The concentrations of sulphate, total nitrate and total ammonium increased during the episode. Our results suggest that large-scale agricultural field burning may substantially affect PM concentrations under unfavourable meteorological conditions even at distances over 1000km from the burning areas. [Copyright &y& Elsevier]

For modeling or data analysis purposes, the concentration of atmospheric aerosol particles measured or modeled at a certain size is often extrapolated to a different size along the particle size axis. This is normally done using compact analytical formulae which are based on the assumption that particles of the same size move synchronously together in size space at the same rate, and thus a monodisperse population is not allowed to disperse. However, in reality the individual uncorrelated collision processes with vapor molecules introduce natural spreading to the particle distribution, which is fundamentally described by a set of discrete rate equations for each particle size. In the present work, predictions based on synchronous growth are compared to the results of the discrete treatment in the free molecular regime for particles of circa 1.5–3.3 nm in mass diameter in a situation where the population is affected by irreversible vapor condensation and scavenging losses. The comparison demonstrates that the results given by approaches that do not consider the discrete nature of the growth process may be distorted. In the studied cases, the differences to the discrete result are a factor of less than around ten, and the deviations are larger for faster particle formation events and often also for higher particle loss rates. Therefore, the discrete treatment should be used for obtaining reliable estimates on the size-dependent concentration of nanometer-sized particles. [ABSTRACT FROM AUTHOR]

We comment on a study by Nadykto et al. recently published in this journal. Earlier work from our group has been misrepresented in this study, and we feel that the claims made need to be amended. Also the analysis of Nadykto et al. concerning the implications of their own density functional calculations is incomplete. We present cluster formation simulations allowing more conclusions to be drawn from their data, and also compare them to recent experimental results not cited in their work. [ABSTRACT FROM AUTHOR]

Formation of secondary atmospheric aerosol particles starts with the formation and growth of small molecular clusters. The probability that freshly formed clusters reach larger sizes depends on the rate at which they grow with respect to the rate at which they are lost on pre-existing surfaces. At present, advances in condensation particle counter and mass spectrometer techniques enable the observation of cluster growth via time evolution of size resolved cluster concentrations, and recent studies have utilized measured concentrations to deduce growth rates from the appearance times of different cluster sizes. In this work, we use a dynamic model to simulate the time development of a population of clusters of up to ~2 nm in mass diameter, and examine the relation of the growth rates determined from the appearance times to the growth rates calculated from the molecular fluxes between the clusters. This study concentrates on a simple model substance where the clusters grow only by monomer additions and the growth involves a single free energy barrier. Each cluster size defined by the number of molecules in the cluster is explicitly treated instead of dividing the clusters into size classes. Effects of finite size resolution and cluster–cluster collisions will be discussed in future work. We find that the growth rates determined with the two different approaches may differ significantly, both quantitatively and qualitatively, for the smallest clusters with the highest evaporation rates. The relative difference decreases with increasing cluster size and decreasing evaporation rate. In addition to cluster size, the difference depends on ambient conditions including external losses and time profile of the monomer concentration. Thus a quantitative comparison requires information not only on the substance, but also on the external conditions. We also show that the size of a critical cluster, corresponding to the maximum of an energy barrier in cluster formation, cannot be inferred from the size-dependent growth rates in realistic conditions. [ABSTRACT FROM AUTHOR]

The first nucleation theorem is the most widely used method to assess atmospheric new-particle formation mechanisms from particle formation rate measurements. The theorem states that the slope ( ∂ log J ) / ( ∂ log C ) of the nucleation rate J versus the concentration C of a nucleating compound gives the number of molecules of that species in the critical cluster. In principle, the derivation of the theorem is solid, but it contains very restrictive assumptions, the validity of which is questionable in realistic situations. It applies only for systems where clusters grow by addition of single molecules, and there are no external losses. In addition, application of the theorem to experimental data requires that the nucleation rate can be determined from particle concentration observations. This work presents simulation results on particle formation rates in atmospherically relevant conditions. We show that the slope of the nucleation rate in realistic conditions differs from that in an ideal situation. The slope analysis can easily lead to erroneous conclusions on the critical cluster size, and should therefore not be used to interpret experimental data. [ABSTRACT FROM AUTHOR]

The frequency, strength and sources of long-range transport (LRT) episodes of fine particles (PM2.5) were studied in southern Finland using air quality monitoring results, backward air mass trajectories, remote sensing of fire hot spots, transport and dispersion modelling of smoke and chemical analysis of particle samples (black carbon, monosaccharide anhydrides, oxalate, succinate, malonate, SO4 2−, NO3 −, K+ and NH4 +). At an urban background site in Helsinki, the daily WHO guideline value (24-h PM2.5 mean 25 μg m−3) was exceeded during 1–7 LRT episodes per year in 1999–2007. The 24-h mean maximum concentrations varied between 25 and 49 μg m−3 during the episodes, which was 3–6 times higher than the local mean concentration (8.7 μg m−3) in 1999–2007. The highest particle concentrations (max. 1-h mean 163 μg m−3) and the longest episodes (max. 9 days) were mainly caused by the emissions from open biomass burning, especially during springs and late-summers in 2002 and 2006. During the period 2001–2007, the satellite remote sensing of active fire hot spots and transport and dispersion modelling of smoke indicated that approximately half of the episodes were caused partly by the emissions from wildfires and/or agricultural waste burning in fields in Eastern Europe, especially in Russia, Belarus and Ukraine. Other episodes were mainly caused by the LRT of ordinary anthropogenic pollutants, e.g. from energy production, traffic, industry and wood combustion. During those ‘other episodes’, air masses also arrived from Eastern Europe, including Poland. The highest concentrations of biomass-burning tracers, such as monosaccharide anhydrides (levoglucosan + mannosan + galactosan) and K+, were observed during open biomass... [Copyright &y& Elsevier]

In this study, we developed two methods to distinguish the long-range transport (LRT) episodes from local pollution (LP) episodes. The first method is based on particle number concentrations ratio between accumulation mode (diameter >90nm) and Aitken mode (diameter 25–90nm). The second method is based on a proxy variable (interpolated ion sum) for long-range transported PM2.5. The ion-sum is available from the measurements of sulphate, nitrate and ammonium at the nearest EMEP stations. We also utilised synoptic meteorological weather charts, locally measured meteorological data, and air mass back-trajectories to support the evaluation of these methods. We selected nine time periods (i.e. episodes) with daily average PM10>50μgm−3 in the Helsinki Metropolitan Area during year 2002. We characterized the episodes in terms of PM10 and PM2.5 concentrations and the fraction of fine particles in PM10 at an urban traffic and regional background air quality monitoring sites. Three of these episodes were clearly of local origin. They were characterized by a low average fraction of PM2.5 (<0.2) in PM10 at the urban traffic monitoring site, low ratio between PM10 concentrations at the regional background site and at the urban traffic site (<0.2), low average ion sums (1.5–2.5μgm−3) and low accumulation to Aitken mode ratios (0.13–0.26). Four of the episodes had distinct LRT characteristics: a high fraction of fine particles in PM10 (0.5–0.6) at the urban traffic site, a high ratio between PM10 concentrations at the regional background site and at the urban traffic site (0.7–0.8), high interpolated values for the ion sum (6.6–11.9μgm−3), and high accumulation to Aitken mode ratios (0.75–0.85). During the remaining two episodes there was significant contribution from both local sources and LRT. A detailed analysis of meteorological variables and air mass back-trajectories gave support to these findings. These characteristics can be utilised in a simple procedure to distinguish between LRT and LP episodes. Further quantitative investigations to these characteristics provide an indication to the episode strength. The quantitative results presented in the current study are applicable to the Helsinki Metropolitan Area and similar cities. Nevertheless, developing these methods for other cities require analyses of the meteorological conditions, behavior of the PM concentrations, and air-mass back trajectories for that specific city. [Copyright &y& Elsevier]

Abstract: The formation and growth of fresh atmospheric aerosol particles was investigated using a condensation particle counter battery (CPCB). This instrument is a matrix of four separate CPCs, which differ in the combination of both cut-off size and working liquid (water; n-butanol). In a first step, the CPC counting efficiencies and cut-off sizes were carefully characterised under laboratory conditions for different condensing vapours, temperature differences between condenser and saturator, and test aerosol types. In addition, the activation process was described theoretically, and modelled numerically for the given CPC configurations. These results confirmed that water-soluble and water-insoluble as well as butanol-soluble and butanol-insoluble aerosol particles may be discriminated in the CPCB through different activation diameters. Therefore, the CPCB represents a novel tool to infer information on the chemical composition of aerosol particles between 2 and 20nm. To test the applicability of the CPCB under field conditions, the CPCB was operated at a rural background station in Finland (Hyytiälä) in April and May 2005. The results indicate that growing nucleation mode particles were water-soluble both at 3 and 11nm. [Copyright &y& Elsevier]

Abstract: We compare two molecular Monte Carlo simulation methods, the discrete summation method and the growth/decay method, which calculate the vapor-liquid nucleation free energy barrier by simulating isolated clusters of fixed size without the surrounding vapor. The methods are applied to calculations of nucleation barriers of Lennard–Jones argon at 60K and 80K. Both of these methods are computationally efficient, as only isolated clusters without the surrounding vapor are simulated, and the methods can be applied with any given cluster definition. They give equivalent results to other methods where the vapor phase is also included. The discrete summation method is based on the calculation of the difference in free energies between two systems containing an n-cluster and an (n −1)-cluster plus one non-interacting (free) molecule. We show that the configurational space is not equivalent in the two systems. Hence, there has to be an additional term in the free energy calculation that accounts for several kT in magnitude. In contrast to previous studies we also show that it is not correct to prevent the overlap of the non-interacting molecule and another molecule by a zero or an arbitrarily small repulsive potential, but with a small excluded space around the free molecule. [Copyright &y& Elsevier]

Based on 6 years of outdoor measurements at a boreal forest site in Southern Finland, scavenging coefficients were calculated for aerosol particles having diameter between 10 and 510 nm. Median scavenging coefficients varied between 7×10−6 and 4×10−5 s−1 in this size-range. The dependence of scavenging coefficients on rain intensity was studied, and the scavenging coefficients were parameterized as a function of particle size for particle diameters of 10–500 nm and for rain intensities 0–20 mm h−1. [Copyright &y& Elsevier]

Molecular clusters are ubiquitous constituents of the ambient atmosphere, that can grow into larger sizes forming new aerosol particles. The formation and growth of small clusters into aerosol particles remain one of the largest uncertainties in global climate predictions. This has made the modeling of atmospheric molecular clustering into an active field of research, yielding direct molecular level information about the formation mechanism. We review the present state-of-the-art quantum chemical methods and cluster distribution dynamics models that are applied to study the formation and growth of atmospheric molecular clusters. We outline the current challenges in applying theoretical methods and the future directions to move the field forward. • Quantum chemical methods applied to cluster formation are reviewed. • A detailed overview of the progress in cluster formation studies. • Further prospects in the field. [ABSTRACT FROM AUTHOR]