Secondary aerosol particles, which are formed by nucleation processes in the atmosphere, play an important role in atmospheric chemistry and in the Earth's climate system. They affect the Earth's radiation balance by scattering solar radiation back to space and can also act as cloud condensation nuclei (CCN) and thereby affect the amount of cloud and its radiative properties. Clouds have a net cooling effect on the Earth's radiation budget of about -27.7 W m-2 . Thus, a small change in cloud properties can have significant effect on the climate system. Results by and have shown that a significant fraction (ranging between 31 and 70 %) of cloud-forming aerosol particles in the atmosphere are secondary particles that originate from nucleation. Therefore, understanding nucleation is crucial in order to fully understand the atmospheric and climatic effects of aerosols.

Sulfuric acid (H2SO4) is the primary ingredient in the production of secondary aerosols because of its low vapour pressure and its ability to bond with water, which is ubiquitous in the atmosphere . H2SO4 is primarily produced in the atmosphere from sulfur dioxide (SO2) via oxidation by the OH radical, produced photochemically with ultraviolet light coming from the Sun. When H2SO4 collides with other molecules, it starts forming small clusters of molecules that can grow into new stable aerosols. If only H2O and H2SO4 take part, the process is termed binary homogeneous nucleation. Nucleation can be significantly enhanced by other substances, the dominant ones being ammonia (NH3) and organic molecules . These processes are termed ternary and organic-mediated nucleation, respectively. Recent results show that in low H2SO4 environments nucleation also happens by condensation of highly oxygenated organic molecules alone . Further, ions enhance the nucleation process by stabilizing the molecular clusters. This process is termed ion-induced nucleation. The fraction of ion-induced nucleation of total particle formation was observed in various environments by . This study found that the fraction was in the range 1–30 % being highest in environments with generally low nucleation rates.

The typical concentration range of gas-phase H2SO4 in the atmosphere is 106–107 cm-3. The concentrations vary with location, time of day and emission of SO2, which can be both anthropogenic and natural. Ions are ubiquitous in the lower atmosphere and are mainly produced by galactic cosmic rays (GCRs), forming 1–40 ion pairs cm-3 s-1. The formation rate depends on factors such as altitude, latitude, and the solar cycle. Ionization is higher above land than above ocean due to natural radioactivity from soils, and the maximum ionization is at altitudes of ∼ 13 km . In addition to the natural variations in ionization, an event such as a nearby supernova would significantly increase the atmospheric ionization in the time following the event. There exists strong indications of a supernova at a relatively close distance of ∼ 50 pc from the solar system ∼ 2.2 million years ago . According to the increase in GCR from such an event would cause an increase in tropospheric ionization of up to a factor of 50 during the first few hundred years following the event.

Few measurements exist that quantify parameters affecting and assisting nucleation e.g.. Recent laboratory measurements made in the European Organization for Nuclear Research CLOUD (Cosmics Leaving Outdoor Droplets) chamber were presented in and showed the dependence on temperature, trace gas and ion concentrations. Based on the measurements a parameterization that can be incorporated into climate models was developed and this parameterization was improved by . These and other measurements, e.g. have verified that ionization helps the nucleation process. In this work we expand on these results by measuring nucleation at ion production rates (q), ranging from background levels to 560 cm-3 s-1 and corresponding to those following a nearby supernova, and atmospherically relevant H2SO4 concentrations (4×106–3×107 cm-3).

The measurements presented in this work were performed in an 8 m3 reaction chamber (SKY2). The set-up is shown schematically in Fig. . The chamber is made of electro-polished stainless steel and has one side fitted with a Teflon foil to allow UV light (253.7 nm) to illuminate the chamber and start the photochemical reaction to generate H2SO4. Dry purified air (20 L min-1) from a compressor with an active charcoal, citric acid, and particle filter was passed through a humidifier and added to the chamber to reach a relative humidity of 38 %. 5 L min-1 of dry air from the same compressor went through an ozone generator where O2 is photolysed by a UV lamp to produce O3. Sulfur dioxide (3.5 mL min-1) was added from a pressurized bottle (5 ppm SO2 in air, AGA). The resulting concentrations of O3 (20–30 ppb) and SO2 (0.6–0.9 ppb) were measured by a Teledyne T400 analyser and with a Thermo 43 CTL analyser, respectively. The H2SO4 concentration was measured with a chemical ionization atmospheric pressure interface time-of-flight (CI-API-ToF) mass spectrometer . The chamber is also equipped with instruments to measure temperature, differential and absolute pressure, humidity, and UV intensity. The pressure was held at a standard pressure of ∼ 1 bar with a slight (0.1 mbar) overpressure relative to the surroundings, the temperature was at 295 K, and the UV intensity was varied as part of the experiments.

Two different condensation particle counters (CPCs) and a particle size magnifier (PSM) were used to count the aerosols formed in the experiments. A TSI model 3775 CPC was used to determine the aerosol particle concentration above a cut-off diameter of 4 nm (dp,cut-off=4 nm). A TSI 3776 CPC (dp,cut-off=2.5 nm) was used in series with the PSM Airmodus A10, developed and described by to detect particles above a cut-off diameter of ∼ 1.4 nm. The cut-off diameter is defined as the mobility diameter of particles, of which 50 % are counted, and it depends on the saturator flow rate and the chemical composition of the particles. For the PSM, the saturator flow rate was set to 1.3 L min-1, which corresponds to a cut-off diameter of 1.4 nm for tungsten oxide particles. The cut-off for H2SO4 aerosols is not known exactly. The cut-off diameter of the PSM is very close to the critical size of ∼ 1.5 nm , which allows for direct measurements of nucleation rate, thereby avoiding extrapolations of the nucleation rate from larger sizes . Both instruments (PSM and CPC) sampled from the same line and had identical sampling pathways as illustrated in Fig. . The CPC with the larger cut-off diameter was used on its own to achieve a larger size span between the instruments, which enables the determination of the particle growth rate (GR).

Figure  shows an example of a run sequence (for 22 % UV and N=8400 [cm-3]) as a function of time. The UV lamps were turned on for 20 min from 02:26:10 to 02:46:12. The top panel shows the temperature in the chamber during the experiment. It shows that the temperature increased by ∼ 0.1 K when the UV lamps were turned on. When the UV was on the highest setting (45 %) the temperature increased by 0.2 K. This slight increase in temperature is negligible with regards to the nucleation rate (5 % change for a 0.2 K increase at the highest [H2SO4] based on ). Therefore, a constant temperature of 295.4 K was used in the further analysis. The second panel shows the H2SO4 concentration in units of 107 [cm-3]. The red line is the 50-point boxcar moving average. Immediately after the UV lamps were turned on, the H2SO4 concentration started to increase. When the UV was turned off, the H2SO4 was lost to scavenging by aerosol particles and to the chamber walls. The third panel shows the aerosol particle concentration measured with the PSM and CPC without any corrections for the wall losses. The red lines represent the 50-point boxcar moving average, which is used for further data analysis to avoid artefacts from noise. Corrections for particle loss to chamber walls are presented further down and the data analysis was performed on the corrected version of the moving average.

As the measurements presented here are an extension of the measurements presented in , at the given conditions, it is natural to compare the two. Therefore, the results shown in Fig.  are compared to , which presents the parameterization of the CLOUD experiments to the highest precision. In their work nucleation is represented as a sum of binary (b), ternary (t), neutral (n), ion-induced (i), and organic nucleation (org). The term representing the organic nucleation rate is not used in the following, as our study does not intentionally add or measure organic molecules. However, as deduced from the high GR, there might be traces of organic species that can also contribute to the nucleation rate. The concentration of organics is considered constant and is included in the ternary nucleation rate. Thus, a nucleation rate given by the sum of the four contributions is considered: J=Jb,n+Jt,n+Jb,i+Jt,i.

At the temperatures and gas concentrations used in this study, ternary nucleation is expected to dominate, as binary clusters are unstable . The model of binary nucleation presented in shows good agreement with measurements performed in the European Organization for Nuclear Research CLOUD (Cosmics Leaving Outdoor Droplets) chamber at lower temperatures (< 273 K). However at tropospheric temperatures (> 273 K) the binary nucleation rate cannot explain the nucleation rates that are observed in either of these studies at the given sulfuric acid concentrations. Since the sulfuric acid concentration is lower in our study, this is particularly important here. attributes the differences to contamination which is more important in providing stabilization for the pre-nucleating clusters when the temperature (and thereby the evaporation) is high. Nevertheless, we have included the binary term for the sake of completeness.

Ammonia (NH3) is filtered from the air that enters the chamber with a citric acid filter. However, trace amounts (which can originate from incomplete filtering or introduction through the humidifier or the bottled SO2 in air) are still present in the chamber and contribute to the production of stable clusters together with organic molecules. As the concentrations of neither NH3 nor organic molecules are measured, an ammonia equivalent concentration [NH3 + org] that represents both species is estimated by comparing the results from the two studies under the same conditions (including T=295 K, RH = 38 %, and N=1700 cm-3).

Figure  shows the parameterization from (dashed lines) on top of the data from our experiments with the [NH3] concentration set to 5.5×1010 cm-3 (2.2 ppbv) for N=1700 cm-3 (left) and all ionization levels (right). Since [NH3]=5.5×1010 cm-3 is the value that gives the best match between the data and the extrapolated parameterization, it is assumed to represent the concentration of NH3 and organic species, [NH3 + org]. Atmospherically observed NH3 concentrations are typically at the sub-ppbv and ppbv levels . Since the air is filtered before entering the chamber, we would expect a concentration that is lower than in the atmosphere. Once again, this is an indication of the presence of other nucleation-enhancing species in our chamber.

We note that nucleation rates were reported for a mobility diameter of 1.7 nm in , meaning that the rates measured in this study should be slightly overestimated since we measure at a mobility diameter of ∼ 1.4 nm (see Sect. 2).

As seen from Fig.  the parameterization presented in matches the nucleation rates from this study, when extrapolated to the same region, especially at [H2SO4]<2×107 cm-3. At higher values of [H2SO4] the nucleation rates from this study are higher than expected from the parameterization. The resulting parameterization shows that at an atmospherically relevant H2SO4 concentration of 1×107 cm-3, the increase in ions from background levels to the highest measured levels causes an increase in nucleation rate of around an order of magnitude.

The disagreement between the data and the expected parameterization could be caused by the narrow range of [H2SO4] in this study, which are all within 1 order of magnitude. Another explanation could be that the detection efficiency of the PSM is ∼ 50 % for particles close to the critical size of 1.4 nm. Since we use a percentage region instead of a fixed time interval when calculating the nucleation rate (see Sect. 3.3), it is possible that the dependence on [H2SO4] was overestimated due to the lower detection efficiency of PSM for particles smaller than 2 nm. At higher [H2SO4], more particles could grow into sizes that are detected more efficiently by the PSM compared to at lower [H2SO4]. This was taken into account by verifying that the regions between 20 and 80 % of each peak of particle concentration were linear. If the detection efficiency was dependent on [H2SO4], these regions would not be linear but the gradient would increase with time for a given peak. However, we still note that the detection efficiency of the PMS could have affected the results in another way. Likewise, it is worth noting that the effect of ions on the detection efficiency of the PSM is unknown, but ions may be more efficiently detected .

It can be complicated to compare different experimental studies, even under similar conditions, because it is unclear how experimental techniques and parameters affect the results. Four studies , all performed in the CLOUD chamber, are most relevant for intercomparison, because they were made in a reaction chamber analogous to this study. The experiments presented in were performed in a flow tube. Yet, we include these in the comparison, because in it nucleation was measured directly at the critical cluster size with a PSM instrument. Table  gives an overview of the studies and conditions that were compared. Some studies consist of several experiments using varying parameters. Table  only shows the measurements made under the conditions that were closest to the parameters of this study. The experiments from this study with lowest ionization are used for comparison, because this ionization level (4.4 cm-3 s-1) is close to the cosmic ray background ionization (GCR ∼ 3 cm-3 s-1).

From Table  it is clear that experiments conducted under the exact same conditions as in this study do not exist. Nevertheless, the nucleation rates in this study lie slightly below or within the range of the nucleation rates obtained in the experiments performed in the CLOUD chamber (studies 2–5). Except for study 3, which was made under lower temperatures, these studies have sulfuric acid concentrations that are 1 to 2 orders of magnitude higher than is the case for this study. As with the parameterization this indicates the existence of other nucleating species within our chamber. By comparing studies 2 and 5 made under almost identical conditions the effect on nucleation of an increase in NH3 concentration is evident (e.g. at [H2SO4]=2×108 cm-3, which is the lower limit for study 2 and the upper limit for study 5).

The temperature used in this paper is only relevant to the boundary layer of the troposphere. At this high temperature evaporation of pre-nucleation clusters is very important and the stabilization provided by ions can have a strong effect , as is also seen in this study. Higher in the troposphere where temperatures are lower the importance of evaporation decreases. However ions can still have a strong effect on the nucleation rates. showed that ions can affect pure binary nucleation rates at mid-troposphere conditions (∼ 250 K). An even higher increase in ionization, as used in this work, would increase the nucleation rates even more – by about 1 order of magnitude according to the parameterization used here. The concentrations of ternary gases are also expected to be lower in the free troposphere, which increases the effect of the ions.

In order to fully account for the variables in nucleation processes observed in this study, direct measurement of NH3 and organic substances would have been preferred. Nonetheless, the results show that the nucleation increases linearly with ion concentration, even at the highest ionization. Also, consistency with the results from and is shown.

The nucleation of H2SO4/H2O aerosols was studied under near-atmospheric conditions in an 8 m3 reaction chamber. Sulfuric acid was produced in situ in the range [H2SO4]=4×106–3×107 cm-3 and the ionization of the air in the chamber was increased from background levels of ∼ 4 up to 560 cm-3 s-1 (ion concentrations =1700-19000 cm-3) using gamma sources. Such levels of ionization are relevant for a nearby (∼ 50 pc) supernova which is thought to have occurred ∼ 2.2 million years ago. The experiments were performed at T=295 K and RH = 38 %. The study shows that nucleation increases linearly with ion concentration over the full range of ion concentrations. We find that nucleation increases by an order of magnitude, when the ion concentration is increased from background to maximum levels. We have not measured the concentration of nucleating species other than sulfuric acid, so the nucleation pathways are unclear. Based on comparisons with other studies we conclude that ternary nucleation involving ammonia or organics is required to explain the observed nucleation rates. Still, this study is a novel contribution to the experimental studies of nucleation rates for the ammonia/organic-mediated H2SO4/H2O system because of the direct measurements of nucleation rates at sizes close to the critical cluster size at high ion concentrations. This work expands the measurements presented in for [NH3+org]=2.2 ppb, RH = 38 % and T=295 K. Based on the presented experiments we find it possible to expand the parameterization from to higher ion concentrations.