The role of organic condensation on ultrafine particle growth during nucleation events

A new aerosol dynamics model (DMANx) has been developed that simulates aerosol size/composition distribution and includes the condensation of organic vapors on nanoparticles through the implementation of the recently developed volatility basis set framework. Simulations were performed for Hyytiälä (Finland) and Finokalia (Greece), two locations with different organic sources where detailed measurements were available to constrain the new model. We investigate the effect of condensation of organics and chemical aging reactions of secondary organic aerosol (SOA) precursors on ultrafine particle growth and particle number concentration during a typical springtime nucleation event in both locations. This work highlights the importance of the pathways of oxidation of biogenic volatile organic compounds and the production of extremely low volatility organics. At Hyytiälä, organic condensation dominates the growth process of new particles. The low-volatility SOA contributes to particle growth during the early growth stage, but after a few hours most of the growth is due to semi-volatile SOA. At Finokalia, simulations show that organics have a complementary role in new particle growth, contributing 45 % to the total mass of new particles. Condensation of organics increases the number concentration of particles that can act as CCN (cloud condensation nuclei) (N100) by 13 % at Finokalia and 25 % at Hyytiälä during a typical spring day with nucleation. The sensitivity of our results to the surface tension used is discussed.


Introduction
New particles are introduced in the atmosphere by (i) direct emissions from a variety of (primary) sources and (ii) nucleation (in situ formation).Nucleation and subsequent growth of new particles are often observed in most areas of the globe (Kulmala et al., 2004) and represent an important source of ambient aerosol number concentration.Fresh particles formed by nucleation can either grow through condensation of vapors (e.g., sulfuric acid and ammonia, organics) to larger sizes or can be lost by coagulation with pre-existing larger particles.The newly formed particles that manage to survive coagulation can grow to larger sizes and become cloud condensation nuclei (CCN), affecting the cloud droplet number concentration (Adams and Seinfeld, 2002).Nucleation and subsequent growth by condensation can be an important source of CCN (Lihavainen et al., 2003;Kerminen et al., 2005;Laaksonen et al., 2005;Merikanto et al., 2009;Makkonen et al., 2009;Pierce and Adams, 2009b;Wang and Penner, 2009;Yu and Luo, 2009;Spracklen et al., 2010).Changes in CCN concentration affect cloud optical properties and cloud lifetime, perturbing the energy balance of the planet (Twomey 1974(Twomey , 1977(Twomey , 1991;;Albrecht, 1989).An increase in the number concentration of particles that can act as CCN results in higher cloud droplet number concentrations and brighter clouds with longer lifetimes.Measurements of CCN at a non-urban site in Germany suggested that CCN concentrations are mainly determined by the aerosol number size distribution, while the composition of aerosol plays a Published by Copernicus Publications on behalf of the European Geosciences Union.
secondary role (Dusek et al., 2006).Furthermore, nanoparticles can affect human health by deposition in human lungs or the neurological system.For small particles, the damage can be greater due to the larger surface area per unit mass (Peters et al., 1997;Donaldson et al., 1998Donaldson et al., , 2002)).The effects of aerosol composition on human health are still uncertain (Godleski et al., 2000).
Several mechanisms have been proposed to explain in situ particle formation.These include sulfuric acid-water binary nucleation (Nilsson and Kulmala, 1998;Vehkamaki et al., 2002), ternary nucleation (Coffman and Hegg, 1995;Korhonen et al., 1999;Kulmala et al., 2002;Napari et al., 2002), nucleation of organic vapors (Marti et al., 1997;Zhang et al., 2004), ion-induced nucleation (Laakso et al., 2002), and halogen oxide nucleation (Hoffmann et al., 2001).The binary nucleation mechanism has been the most commonly used in atmospheric models with the critical cluster assumed to be composed of H 2 SO 4 and H 2 O. Ternary nucleation theory usually includes ammonia (NH 3 ) as a third component.It is possible that other compounds (e.g., organics, amines) may play a similar role under certain conditions (Bonn et al., 2008;Kurtén et al., 2008;Metzger et al., 2010;Smith et al., 2010;Berndt et al., 2010;Zhao et al., 2011;Kirkby et al., 2011;Almeida et al., 2013;Riccobono et al., 2014).A strong correlation has been found between measured aerosol nucleation rate and the gas-phase sulfuric acid concentration (Weber et al., 1996;Sihto et al., 2006;Riipinen et al., 2007;Kuang et al., 2008;Nieminen et al., 2009;Paasonen et al., 2009Paasonen et al., , 2010) ) in various sites in Europe and the United States of America.Nucleation events observed in sulfur-rich regions like the northeastern USA appear to be initiated by the formation of gas-phase H 2 SO 4 (via SO 2 oxidation) but terminated by the exhaustion of gas-phase NH 3 or other bases (Jung et al., 2008).Ambient measurements and some laboratory studies (Sipila et al., 2010) have revealed a linear or squared correlation between new particle formation rate and concentration of sulfuric acid.
Significant uncertainties arise also from the incomplete understanding of the identity of the species involved in the growth of these nuclei (Kulmala et al., 2004).Field measurements (Eisele and McMurry, 1997;Weber et al., 1998Weber et al., , 1999;;Janson et al., 2001) and model simulations (Kerminen et al., 2001;Kulmala et al., 2000;Pirjola and Kulmala, 2001;Anttila and Kerminen, 2003) have indicated that the condensation of sulfuric acid alone is often not sufficient to explain the observed growth rates of fresh particles (Riipinen et al., 2011).The growth of fresh nuclei could be due to the condensation of organic species (Kerminen et al., 2000;Anttila and Kerminen, 2003), heterogeneous reactions (Zhang and Wexler, 2002), or ion-enhanced condensation (Laakso et al., 2002).Jung et al. (2010) developed a three-dimensional regional chemical transport model (CTM), PMCAMx-UF, with detailed aerosol microphysics (Gaydos et al., 2007;Karydis et al., 2007) that has been used for simulations over the USA and Europe (Fountoukis et al., 2012).In Europe the model predictions were compared against size distribution measurements from seven sites (Fountoukis et al., 2012).The model was found to reproduce more than 70 % of the hourly number concentrations of particles larger than 10 nm (N 10 ) within a factor of 2. For particles larger than 100 nm (N 100 , a proxy for the number of particles that can act as CCN) a systematic underprediction was seen.The growth rates were also underpredicted (with smaller errors in sites where the sulfate to organics mass ratio is high, e.g., Melpitz), possibly because of insufficient organic vapor condensation (Fountoukis et al., 2012) as the model did not explicitly include secondary organic aerosol (SOA) condensation on ultrafine particles.Yu (2011) and Riipinen et al. (2011) studied the condensation of organics on ultrafine particles using global CTMs.Yu (2011) estimated that the concentration of lowvolatility organics is a factor of 2-20 higher than the H 2 SO 4 concentration in many continental locations and can significantly enhance the growth rate of freshly nucleated particles.He compared predicted particle size distributions with field measurements in a boreal forest site (Hyytiälä, Finland) showing that the condensation of low-volatility organics can bring the simulation results closer to the observations.Riipinen et al. ( 2011) estimated that roughly half of the condensed organic mass needs to be distributed proportionally to the aerosol surface area to explain the observed aerosol growth.These organic compounds need to have both high yields and very low volatility, which is inconsistent with laboratory observations of the first-generation yields of SOA from biogenic precursors (Presto and Donahue, 2006;Pathak et al., 2007;Pandis et al., 2013).Pierce et al. (2011) estimated that the average effective saturation concentration (C * ) of condensing organics needs to be 10 −3 −10 −2 µg m −3 or less to enhance the growth of freshly nucleated particles.
SOA accounts for a significant mass fraction (20-90 %) of sub-micrometer particulate matter at many locations around the globe and is one of the most dominant particle components in the atmosphere (Jimenez et al., 2009).The sources and the chemical composition of OA (organic aerosol) are still uncertain due to the large number (tens of thousands) of different atmospheric organic compounds (Goldstein and Galbally, 2007).OA has been the subject of numerous studies during the last decade (Hallquist et al., 2009) but remains the least understood component of atmospheric aerosols.The organic aerosol composition continuously evolves with time due to various chemical reactions (Kanakidou et al., 2005).Gas-phase oxidation of volatile organic compounds (VOCs) produces semi-volatile products that can then condense to the particulate phase.Products with high vapor pressures can be oxidized to species with lower vapor pressures that can then condense on preexisting particles.The chemical aging (further oxidation) of semivolatile organic compounds is an important source of OA mass (Donahue et al., 2006).Until recently, most CTMs described SOA formation using two surrogate species per VOC (Odum et al., 1996).This ap-proach is computationally expensive due to the large number of products, while the use of only two products per VOC limits the concentration range and the accuracy of this approach (Murphy et al., 2009).SOA vapors may undergo further gasphase oxidation, and simulation of this aging process would require introduction of even more species (Ng et al., 2006).The volatility basis set (VBS) framework (Donahue et al., 2006) was proposed to address these problems, describing the complete volatility range of OA compounds using logarithmically spaced bins characterized by an effective saturation concentration, C * (in µg m −3 ).This framework has been shown to work well for simulations of aerosol mass distributions in 3-D CTMs (Murphy et al., 2009;Tsimpidi et al., 2010;Fountoukis et al., 2011).
The overall objective of this work is to examine the contribution of organic vapor condensation to the growth of fresh particles formed by nucleation and whether this condensation can explain the observed growth rate of new particles.We extend the Dynamic Model for Aerosol Nucleation (DMAN) of Jung et al. (2006), which originally assumed that particles can grow only by condensation (of sulfuric acid and ammonia) and coagulation.In this work, we develop an updated version of DMAN (DMANx) which includes the condensation of organic vapors on particles and the most recent version of the VBS framework.We estimate for the first time the effect of the chemical aging gas-phase reactions of SOA components on ultrafine particle growth.We examine the effects of condensation of organics, the gas-phase chemical aging reactions, and the Kelvin effect on the predicted particle number concentrations.We focus on the composition of fresh particles during nucleation events as well as during their subsequent growth, in two remote continental locations, Hyytiälä, Finland, and Finokalia, Greece, where there are sufficient measurements available to constrain the new model.This is the first step towards improving the predictions of the 3-D CTM, PMCAMx-UF.

Model description
DMAN simulates nucleation, coagulation, and condensation/evaporation for a multi-component atmospheric aerosol (Jung at el., 2006).It uses the Two-Moment Aerosol Sectional (TOMAS) algorithm of Adams and Seinfeld (2002), which is based on the sectional approach for the description of the aerosol size composition distribution.TOMAS is an adaptation of cloud microphysics algorithms (Tzivion et al., 1987(Tzivion et al., , 1989) ) to aerosol processes, is computationally efficient, and tracks both mass and number concentrations simultaneously.The aerosol size distribution is described with 41 size sections, with the lowest size bin corresponding to a 3.7 × 10 −25 kg dry aerosol mass per particle.That corresponds to 0.8 nm dry diameter assuming a density of 1.4 g cm −3 .Each successive section has double the mass of the previous one.The largest bin corresponds to a diameter of 10 µm.

Nucleation
DMAN has the option of using a number of nucleation parameterizations.In this work, the rate of nucleation is calculated using a scaled ternary nucleation parameterization based on the original expressions of Napari et al. (2002) if the NH 3 concentration exceeds 0.01 ppt, and the binary parameterization of Vehkamaki et al. (2002) if it is less than this threshold value.The original NH 3 -H 2 SO 4 -H 2 O parameterization has been successful in predicting the presence or lack of nucleation events (Gaydos et al., 2005) in sulfurrich environments.However, it overpredicts ultrafine number concentrations during nucleation events (Gaydos et al., 2005;Yu, 2006a, b;Jung et al., 2006Jung et al., , 2008;;Merikanto et al., 2007;Zhang et al., 2010) and thus a scaling factor of 10 −5 is applied to the nucleation rate following Jung et al. (2010).The critical nucleus is predicted to consist of roughly two molecules of sulfuric acid and two molecules of ammonia (Napari et al., 2002), so it is assumed here that the newly formed particles consist of ammonium bisulfate and their diameter is 1 nm.

Gas-phase chemistry
In this work, the simulation of gas-phase chemistry in DMAN is updated using the SAPRC99 chemical mechanism (Carter, 2000;Environ, 2003), which includes 211 reactions of 56 gases and 18 free radicals.It includes five lumped alkanes, two lumped olefins, two lumped aromatics, isoprene, a lumped monoterpene, and a lumped sesquiterpene species.Only the two highest molecular weight alkane species are considered as SOA precursors because the other three contain smaller hydrocarbons (Pandis et al., 1991).OLE1 contains all the terminal alkenes, while OLE2 consists of all the internal and cyclic alkenes.The major compounds for each VOC class used in SAPRC99 are listed in Table S1 (Tsimpidi et al., 2010).The nine lumped VOCs are considered as volatile SOA precursors, with three of them being biogenic and the rest anthropogenic.

Coagulation
Coagulation of particles in the atmosphere is an important sink of aerosol number but is also a mechanism by which freshly nucleated particles grow to larger sizes.The TOMAS algorithm is used for the simulation of coagulation.Following Adams and Seinfeld (2002), TOMAS assumes that the aerosol particles coagulate via Brownian diffusion and the effects of gravitational settling and turbulence are negligible.The calculation of the coagulation coefficients is based on the wet diameters of the particles.These wet diameters are calculated following the approach of Gaydos et al. (2005).For small particles (< 100 nm), we use the expression of Dah-neke et al. (1983) in order to correct for non-continuum effects.The coagulation algorithm uses an adaptive time step.The time step is limited so that the aerosol number or mass concentration in any size category does not increase by more than an order of magnitude or decrease by more than 25 %.

Condensation
Condensation of gas-phase species to existing aerosol particles is an important source of aerosol mass and a means by which small particles grow to CCN sizes.The TOMAS algorithm is used for the simulation of condensation/evaporation of sulfuric acid, ammonia, and organic vapors, using the wet diameters of the particles (Gaydos et al., 2005).The driving force for condensation of a vapor to an aerosol particle is the difference between its ambient vapor partial pressure and the equilibrium vapor pressure over the particles, or where p i is the condensational driving force of the organic vapor i (the difference between the partial pressure of condensing vapor and its equilibrium vapor pressure), p i is the ambient partial pressure, x i is the mole fraction of i, p * i is the effective saturation pressure over a flat surface, σ is the surface tension, M i is the molecular weight of i, R is the ideal gas constant, T is the temperature, ρ is the liquid-phase density, and D p is the diameter of the particle.The exponential term is known as the Kelvin effect due to the curvature of the particles.The exponential term is large for small particles and "prevents" the condensation of organic vapors on these.As a result, the Kelvin effect is important for the growth of newly formed particles.We use TOMAS with an adaptive time step to efficiently solve the equations for condensation.The time step is chosen so that individual particles in any size bin do not grow by more than 10 %, the partial pressure of the organic vapor does not fall below 25 % of its original value, and the time step is never longer than 15 min.
DMANx uses the pseudo-steady-state approximation (PSSA), in which the sulfuric acid concentration is calculated by assuming that its production rate (oxidation of sulfur dioxide) is equal to its consumption rate (condensation and nucleation).Pierce and Adams (2009a) showed that the PSSA for sulfuric acid increases the computational speed with a small loss in accuracy.The PSSA was tested for a variety of conditions ranging from highly polluted to extremely clean conditions.Its predictions for the sulfuric acid vapor concentration and the number of new particles formed during typical atmospheric nucleation events agreed well with the "benchmark model" (Pierce and Adams, 2009a).Jung et al. (2010) evaluated the performance of PSSA for sulfuric acid in DMAN against a fourth-order Runge-Kutta algorithm and showed that PSSA is accurate and computationally efficient.
Condensation of ammonia is simulated following the approach described by Jung et al. (2006).Ammonia conden-sation on the ultrafine particles ends when sulfate is fully neutralized to ammonium sulfate.The equilibrium ammonia vapor pressure is assumed to be zero when aerosols are acidic, i.e., when the molar ratio of NH + 4 to SO 2− 4 is < 2. If the amount of condensed ammonia exceeds the amount needed to neutralize particles during a time step, we limit the maximum amount of ammonia that can condense to avoid numerical problems.

Secondary organic aerosol formation
Gas-phase oxidation of volatile organic compounds (VOCs) produces semi-volatile products that can then condense to the particle phase.The VBS framework used in DMANx (Donahue et al., 2006) describes the complete volatility range of OA compounds using logarithmically spaced bins, characterized by an effective saturation concentration, C * (in µg m −3 ).SOA components partition between the aerosol and gas phases, and can be formed from anthropogenic SOA (aSOA) and biogenic SOA (bSOA) precursors.Each of these types is simulated here with 12 volatility bins (10 −5 -10 6 µg m −3 ).We assume an average molecular weight of 200 g mol −1 for both aSOA and bSOA, while the effective enthalpies of vaporization are 30 kJ mol −1 (Pathak et al., 2007;Stanier et al., 2007).The SOA yields used in the updated version of DMAN are based on the NO x -dependent stoichiometric yields of Murphy et al. (2009).The partitioning of OA between the gas and particulate phases is approximated using vapor-liquid equilibrium theory (Eq.1).
Semi-volatile and intermediate-volatility organics can be oxidized to species with lower volatility (Donahue et al., 2006) leading to SOA production.The gas-phase chemical aging of SOA precursors is modeled using a second-order gas-phase reaction with the hydroxyl radical.We assume that each chemical aging step reduces the volatility of the corresponding organic vapor by 1 order of magnitude (i.e., shifting organic material from a saturation concentration of, for example, 10 3 to 10 2 µg m −3 ), with a small net increase in mass (7.5 %) to account for the added oxygen.The chemical aging reactions for aSOA precursors are modeled with a rate constant k (298 K) = 1 × 10 −11 cm 3 molec s −1 (Murphy et al., 2009).
In the base case, the gas-phase chemical aging of bSOA precursors is assumed to have a negligible effect on OA concentration (Lane et al., 2008).The oxidation of biogenic VOCs produces semi-volatile organics with saturation concentrations of 1, 10, 10 2 , and 10 3 µg m −3 .An alternative hypothesis is that the condensation of very low volatility organics may explain the observed growth.We assume that a small fraction of organics, which are produced from the oxidation of biogenic VOCs, further reacts to form very low volatility organics with a saturation concentration of C * = 10 −3 µg m −3 (Pierce et al., 2011).The sensitivity of the model results to this assumption will be tested assuming a reaction converting the gas-phase surrogate species with C * = 1 µg m −3 to extremely low volatility SOA with C * = 10 −3 µg m −3 with a reasonable rate constant equal to 1 × 10 −11 cm 3 molec s −1 .The importance of extremely low volatility organic material for the growth of newly formed particles is explored in a subsequent section.This pathway is consistent with the recent observations of extremely low volatility organic compounds by Ehn et al. (2014), during ozonolysis reactions as well as reactions with the hydroxyl radical.

Model application
We simulated a "typical" spring day with nucleation at both Hyytiälä and Finokalia.First, we identified the days with observed particle formation and growth and then averaged the corresponding measurements during these days, generating in this way the meteorological and chemical characteristics of an "average" nucleation day for the specific periods in the two locations (April 2007 in Hyytiälä andMay 2008 in Finokalia).For the parameters for which measurements were not available, but were needed for the model input (e.g., OH concentration), we followed the same process using the predicted values from the 3-D chemical transport model PM-CAMx (Fountoukis et al., 2011).Results are compared to the corresponding average values observed during the specific periods.
The extended DMAN (DMANx) is first tested in Hyytiälä (Finland), an environment dominated by biogenic VOCs.Meteorological data, gas-phase concentrations, and aerosol number distributions are available from ground measurements at the SMEAR II station in Hyytiälä and used here as inputs for DMANx.SMEAR II (Station for Measuring Forest Ecosystem-Atmosphere Relations) is located in a rather homogenous Scots pine (Pinus sylvestris) stand on flat terrain at the Hyytiälä Forestry Field Station of the University of Helsinki (61 • 51 N, 24 • 17 E; 181 m a.s.l.).The biggest city near SMEAR II is Tampere, which has approximately 200 000 inhabitants and is located 60 km from the measurement site.Hari and Kulmala (2005) have described the station and its operation in detail.The main inputs of our simulations (Table 1) are meteorological data (temperature and relative humidity); SO 2 , NH 3 , OH, O 3 , and VOC concentrations; and the initial aerosol number distribution.The SO 2 and O 3 concentrations, T , and RH were measured continuously, while the OH concentration was based on the predictions of the 3-D CTM PMCAMx-2008 (Fountoukis et al., 2011).The concentrations of the lumped VOCs, TERP, ISOP, and ARO1 were estimated based on proton transfer reaction mass spectrometer (PTRMS) measurements.The rest of the lumped VOCs were taken from PMCAMx-2008.The concentration of NH 3 was based on the measurements during the QUEST IV campaign in Hyytiälä (Riipinen et al., 2007).The initial aerosol number distributions are available from DMPS (differential mobility particle sizer) measurements of ambient dry size distributions (Aalto et al., 2001).
DMANx was also tested in Finokalia, a remote area in the eastern Mediterranean region with high sulfate levels and relatively low VOC concentrations.Finokalia (35 • 24 N, 25 • 60 E) is a remote coastal station located in the southeast of the Mediterranean Sea on the island of Crete in Greece.The nearest large urban center is Heraklion, with 150 000 inhabitants, located 50 km west of Finokalia.The Finokalia station is located at the top of a hill at an elevation of 230 m, facing the sea.There is no notable human activity at a range of approximately 15 km (Kouvarakis et al., 2000).There are very few trees and little vegetation in the surrounding area.Most of the aerosol at the site is transported from the surrounding regions, such as Greece, Turkey, and northern Africa (Pikridas et al., 2010).The concentrations of NH 3 and SO 2 and aerosol number distributions were based on the measurements during the Finokalia Aerosol Measurement Experiment 2008 and 2009 (FAME-08 and FAME-09) (Pikridas et al., 2010(Pikridas et al., , 2012)).The other inputs (Table 1) were based on the predictions of PMCAMx.
The model simulates a full day, beginning at midnight.In each simulation, we assumed for the initial distribution that each size bin contains half organics and half ammonium sulfate.Our results are not especially sensitive to this assumed initial composition; the initial particle size distribution is a lot more critical.

Simulation without condensation of organic vapors
In the simulation neglecting the organic contribution to ultrafine particle growth in Hyytiälä, the new particles reach a diameter of 9 nm and the growth rate is only 1 nm h −1 (Fig. 1a).The typical observed growth rates in Hyytiälä are between 1 and 4.5 nm h −1 and the final diameters between 14 and 45 nm (Pierce et al., 2011;Yli-Juuti et al., 2011).A significant advantage of DMANx is that it can track the composition of fresh particles formed by nucleation.The growing nucleation mode consists of sulfate and ammonium without any organics (Fig. 1b).
In the Finokalia nucleation simulation, the model predicted that the fresh particles grew to 32 nm with a rate of 3 nm h −1 (Fig. S1a in the Supplement), which is less than the 5 nm h −1 reported by Pikridas et al. (2010Pikridas et al. ( , 2012)).The particles reached a diameter of 10 nm at 10:00 LST.These new particles consisted of sulfate and ammonium (Fig. S1b).At the start of the nucleation event, fresh particles consist of 93 % sulfate, which drops to 72 % by the end of the day.In the simulation with organic vapor condensation and σ = 0.025 N m −1 (Pierce et al., 2011) without gas-phase chemical aging of biogenic SOA (bSOA) precursors, the predicted growth rate in Hyytiälä remains low at 1.2 nm h −1 and the diameter of new particles at the end of the day is 12 nm (Fig. 2a).These results are low compared to typical measurements of growth rate in this area.The simulation showed that using a reasonable surface tension value practically prevents the condensation of organics on fresh particles until they can grow above a diameter of 5 nm.Surface tension has a major effect, as expected, on the composition of the new particles.At the beginning of the nucleation event (Fig. 2b), the new particles do not contain any organics and their initial growth is due to the condensation of sulfuric acid and ammonia.At 12:00 LST, when the new particles have reached a diameter of 5 nm, the effect of surface tension has decreased and organics start slowly condensing, contributing to growth.The mass fraction of organics in the new particles gradually in- creases, reaching 60 % at the end of the day, when these particles reach 10 nm (Fig. 2b).
The organic composition of the fresh particles in Hyytiälä is shown in Fig. S2.Components with lower volatility (C * = 10 −2 and 10 −3 µg m −3 ) contribute around 30 % of the organic mass in the initial stages of the growth (Fig. S2).Another 55 % is due to the C * = 0.1 and 1 µg m −3 components.As the day goes on, the contribution of the more volatile components increases, and at the end of the day 65 % of the new particle organic mass is semi-volatile material (C * of 1 and 10 µg m −3 ).We estimate that the fraction of the condensing organic mass that has gone to the ultrafine particles (diameter < 100 nm) reached a maximum value of 3.5×10 −5 at 12:00 LST.
In the case of σ = 0.025 N m −1 at Finokalia (without chemical aging of bSOA precursors), the predicted growth rate is 3.8 nm h −1 , which is still lower than the measured 5 nm h −1 rate.The mode diameter of the new particles reaches 42 nm (Fig. S3a).At 09:30 the newly formed particles consist of 90 % sulfate, 6 % ammonium, and 4 % organics (Fig. S3b).The organics condensation starts accelerating later, when these particles become larger than 3 nm (af- ter 10:00 LST).The predicted SOA mass fraction increases to 35 % by the end of the day.The surrogate OA species with C * = 1 and 10 µg m −3 were the major components of new particles in this case too, representing around 70 % of the OA during the day (Fig. S4).In this simulation a maximum of 3 % of condensing organic mass has gone to the ultrafine particles (at 12:00).

Contribution of gas-phase chemical aging of biogenic SOA precursors
The simulations described in the previous section (with condensation of organics and σ = 0.025 N m −1 ) showed that the addition of condensation of semi-volatile organics did not close the gap between predictions and field measurements of particle growth.Adding gas-phase chemical aging of bSOA precursors in the Hyytiälä simulation results in a growth rate of 2.2 nm h −1 and a final diameter of 23 nm.These are very similar to the values of growth rate (2.1 nm h −1 ) and final diameter (21 nm) observed in Hyytiälä on 10 April 2007 (Pierce et al., 2011) (Fig. 3).In the beginning of the simulation, the new particles consist mainly of ammonium sulfate and a small amount of organics (Fig. 4a).This small fraction of organics is the extremely low volatility organics (C * = 10 −3 µg m −3 ) produced from the aging reaction of bSOA precursors (Fig. 4b).During the day the new particles grow and the mass of organics increase.The lowvolatility material dominates the growth during the first few hours, when the diameter is less than 5 nm, while the semivolatile (C * = 1, 10 and 100 µg m −3 ) dominates the growth during the later stages.At 18:00 LST the new particles consist mostly (90 %) of organics (40 % of low volatility and 60 % semi-volatile SOA) (Fig. 4).The semi-volatile SOA contributes to growth when the particles pass the size of ∼ 10 nm where the Kelvin effect is small.The aging reaction affects new particles at the beginning of growth when the particles are very small.The maximum mass of condensing organics that has gone to the ultrafine particles is about 0.9 % at noon, which is larger than in the simulation without aging.The assumed gas-phase aging reaction of bSOA precursors helps new particles to grow to larger sizes in Finokalia too, and predictions are now consistent with field measurements (Fig. 5).The predicted growth rate is 4.7 nm h −1 , while the measured growth rate in Finokalia is around 5 nm h −1 (Pikridas et al., 2012).The revised model can reproduce the observed growth rate and the final particle size encouragingly well.At the start of the growth, the new particles consist of 85 % (NH 4 ) 2 SO 4 , while at the end of the day this drops to 55 % (Fig. 6a).The organic components initially comprise 12 % of the nucleated particle mass, and are mostly of low volatility (75 % is from the C * = 10 −3 µg m −3 volatility bin).During the day, the organics mass fraction increases, reaching a maximum of 45 % of the nucleated particle mass and consisting of 30 % low-volatility and 70 % semi-volatile organics (Fig. 6b).Four percent of the condensing organics at noon has gone to the ultrafine particles.

Organic condensation neglecting the Kelvin effect without aging of bSOA precursors
Condensation of organics neglecting the Kelvin effect (assuming σ = 0 N m −1 ) helps the newly formed particles to reach a diameter of 28 nm in Hyytiälä, while the average growth rate is 1.8 nm h −1 (Fig. S5a).At 08:00 LST the nu-  cleation mode particles have a size of ∼ 2 nm and consist of 55 % sulfate, 11 % ammonium, and 34 % organics.During the day, the new particles continue to grow, and the fraction of organics increases to 85 % at the end of the day (Fig. S5b).
The size of the new particles increases due to the condensation of organics.The surrogate OA components with effective saturation concentration of C * = 1 and 10 µg m −3 are predicted to be the major components of the fresh particles, contributing 80 % of the SOA, initially, and 70 % in the end (Fig. S6).
For the Finokalia case, the predicted growth rate of new particles is 3.8 nm h −1 and the diameter of the new fresh particles reaches 42 nm (Fig. S7a).The condensation of organics helps the particles to grow faster and to reach larger sizes compared to the case described in Sect.4.1.At 08:00, the particles consist of 35 % organics, 5 % ammonium, and 60 % sulfate (Fig. S7b).After 11:00, the new particles composition is relatively stable: 40 % organics, 42 % sulfate, and 18 % ammonium.For this case of zero surface tension, organics condense immediately on the newly formed particles, resulting in fast growth.In this case, about 6.5 % of condensing organics has gone to the ultrafines at 12:00.The SOA composition is similar to that of the σ = 0.025 N m −1 case (Fig. S8).

Sensitivity of particle number concentration to surface tension
So far the model has shown that organic condensation can cause an increase in the size of the ultrafine particles.However, large particles also grow by organic condensation, resulting in increased coagulation probability for the newly formed particles (Kuang et al., 2009;Westervelt et al., 2013Westervelt et al., , 2014)).This causes a significant reduction in the number concentration of small particles compared to the case without condensation of organics.The increase in surface tension leads to organic vapor condensation mostly on large particles and to a smaller extent on freshly nucleated particles, due to the increased Kelvin effect, which prevents the condensation of organics on new particles.The predicted daily mean number concentration of particles above 3 nm (N 3 ) in Hyytiälä is 3100 cm −3 for the simulation without organics, 5400 cm −3 for the zero surface tension case, 2500 cm −3 when the organic surface tension is 0.025 N m −1 , and 3000 cm −3 with the aging of bSOA precursors (Fig. 7a).The maximum value of N 3 is predicted at 13:00 in all cases.For the zero surface tension simulation, N 3,max is 13 500 cm −3 , while in the no-organics case N 3,max = 6300 cm −3 .When the organic surface tension is 0.025 N m −1 , N 3,max = 4600 cm −3 and increases to 6000 cm −3 with the addition of aging of bSOA precursors (Fig. 7a).
At Finokalia the average predicted daily concentration of N 3 is 3600 cm −3 without the condensation of organics and is predicted to be the same for the bSOA aging case.A similar N 3 is predicted for the σ = 0.025 N m −1 case (3550 cm −3 ) (Fig. S9a), while N 3 increases to 4300 cm −3 for zero surface tension.The maximum N 3 is predicted to be the largest for the zero surface tension case (Fig. S9a).
Neglecting the Kelvin effect in both locations allows rapid condensation of organics on the fresh particles, resulting in an increase in N 3 .The Kelvin effect suppresses the condensation of organics on new particles, resulting in a decrease in N 3 .The bSOA aging and the σ = 0.025 N m −1 simulations do not significantly affect N 3 compared to the simulation without organics.
The daily mean number concentration of particles above 100 nm (N 100 ) in Hyytiälä is 380 cm −3 for the no-organics case and 440 cm −3 for all the other simulations (Fig. 7b).In all the Hyytiälä simulations of condensing organics, N 100 increases after 08:00 due to organics, while in the simulation without organics a smaller increase is predicted after 12:00 due to sulfuric acid (Fig. 7b).Condensation of organics increased N 100 by 25 % at the end of the day.
A similar behavior is predicted at Finokalia.N 100 starts increasing after 08:00 for all the cases with organics condensation and after 09:00 for the no-organics case.The reason for the shorter delay before the onset of N 100 increase in the no-organics case compared to Hyytiälä is that the condensation of sulfuric acid-ammonium has a predominant role in Finokalia at increasing N 100 , while at Hyytiälä simulations organics dominate this growth process.Furthermore, at Finokalia the concentration of sulfuric acid is higher and the photochemistry is faster (much higher OH levels) than at Hyytiälä.The daily mean number concentration of particles that can act as CCN (N 100 ) is predicted to be 1000 cm −3 without organics and 1100 cm −3 for the other simulations (Fig. S9b).N 100 reaches a maximum (1070 cm −3 ) at 13:00 for the no-organics case and at 18:00 in the other simulations (1210 cm −3 ) (Fig. S9b).At the end of the day N 100 has increased by 13 % due to condensation of organics.The increase in surface tension and the addition of aging of bSOA precursors do not change N 100 .

Conclusions
We developed an updated version of DMAN (DMANx) which includes the condensation of organic vapors on ultrafine particles, using the volatility basis set framework.Simulations were performed for two locations with different organic sources, Hyytiälä and Finokalia, during a typical springtime day with aerosol nucleation and growth.
Using realistic values of surface tension we estimate that the semi-volatile organics condensation is not enough to grow the new particles to sizes comparable to those observed.Assuming that biogenic SOA precursors chemically age and produce extremely low volatility organics (i.e., with an effective saturation concentration of 10 −3 µg m −3 ) results in predicted growth rates similar to those measured.In the biogenic VOC-dominated environment of Hyytiälä, the very low volatility organics condense onto particles smaller than 3 nm.After this first-stage of growth for the new particles, the Kelvin effect becomes small and the semi-volatile organics are the major components controlling the subsequent growth of the nucleated particles.In an environment with more sulfuric acid and fewer biogenic VOCs (Finokalia), the condensation of organics plays a complementary role in the growth of nucleated particles, contributing 45 % of the total mass of new particles during a day representative of springtime nucleation.
The gas-phase chemical aging of biogenic SOA precursors contributes to the growth of the nucleated particles because of the extra mass added from the aging reactions.Laboratory experiments and direct field measurements (Ehn et al., 2014) support the importance of extremely low volatility VOCs produced in the gas phase from the oxidation of monoterpenes and other VOCs.
The condensation of organics with zero surface tension resulted in a predicted growth rate similar to the field measurements, but the zero value of surface tension is unrealistic.The condensation of organics with zero surface energy also affects the number concentration of particles.Increasing surface tension inhibits the growth of new fresh particles and thus results in a decrease of total particle number concentrations in both locations.Interestingly, when including aging of bSOA precursors, the new model predicts daily mean number concentrations similar to those for the no-organics simulation.The number concentration of particles that can act as CCN (N 100 ) increases (by 13 % at Finokalia and 25 % at Hyytiälä) during a typical spring day with nucleation compared to the case in which the condensation of organics is neglected.The increase in surface tension and the aging of bSOA precursors do not significantly affect N 100 compared to the zero surface energy case.
The Supplement related to this article is available online at doi:10.5194/acp-15-6337-2015-supplement.

Figure 1 .
Figure 1.(a) Predicted aerosol dry size distribution for a typical spring nucleation event at Hyytiälä without condensation of organics.Particle number concentration is plotted against local time of day (x axis) and particle diameter (y axis).(b) Predicted composition of new particles.

Figure 2 .
Figure 2. Simulation with surface tension σ = 0.025 N m −1 at Hyytiälä: (a) predicted particle size distribution with number concentration plotted against time of day (x axis) and particle diameter (y axis) and (b) the composition of new particles.

Figure 3 .
Figure 3.Comparison of (a) measured on 10 April 2007 and (b) predicted (with aging of bSOA precursors and σ = 0.025 N m −1 ) dry size distribution as a function of local time at Hyytiälä for a typical nucleation event day.

Figure 4 .
Figure 4. (a) Mass fraction of fresh particles and (b) mass fraction of SOA for the different volatility bins as a function of local time at Hyytiälä.Simulation includes bSOA aging and σ = 0.025 N m −1 .

Figure 5 .
Figure 5.Comparison of (a) measured on 19 March 2009 (Pikridas et al., 2012) and (b) predicted dry size aerosol distribution (with bSOA aging and σ = 0.025 N m −1 ) as a function of time at Finokalia.

Figure 6 .
Figure 6.(a) Mass fraction of fresh particles and (b) mass fraction of SOA for the different volatility bins as a function of local time at Finokalia.Simulation includes bSOA aging and σ = 0.025 N m −1 .

Figure 7 .
Figure 7. Predicted concentrations of (a) N 3 and (b) N 100 at Hyytiälä for the four simulated cases.Black line represents no condensation of organics, red is with condensation of organics with σ = 0.0 N m −1 , blue is with condensation of organics with σ = 0.025 N m −1 , and green is condensation of organics with aging reactions of bSOA precursors and σ = 0.025 N m −1 .