Condensation particle counting below 2 nm seed particle diameter and the transition from heterogeneous to homogeneous nucleation

Particle detection by condensation particle counters (CPCs) is ultimately limited by the onset of homogeneous nucleation. At vapour supersaturations around the homogeneous nucleation limit the diameter of critical clusters is typically about 2 nm. It is widely assumed that only particles larger than critical clusters can be activated by vapour condensation and the general detection limit of CPCs is therefore currently accepted to be around 2 nm particle diameter. Using an expansion type CPC with n-propanol as working fluid we investigated the transition from heterogeneous to homogeneous nucleation, clearly showing that particles are activated much before the onset of homogeneous nucleation, even at particle diameters as small as 1.4 nm. For particle diameters below 2 nm we have usually found condensation particle counting to be influenced by the simultaneous presence of ions as generated in a bipolar diffusion charger. In this paper we illustrate how the presence of ions influences particle number concentration measurement and how ions can be removed in order to obtain accurate seed particle number concentrations for particle diameters down to 1 nm.     

Effects of seed particle size and composition on heterogeneous nucleation of n-nonane

Heterogeneous nucleation of supersaturated n-nonane vapour on seed particles of different size and composition has been investigated using a fast expansion chamber. Monodisperse seed particle sizes were ranging from about 4 nm up to about 24 nm in diameter. By using different types of particle generators WO_x, Ag and (NH₄)₂SO₄ particles were generated. For direct comparison between different particle compositions overlapping sizes have been generated for WO_x and Ag at about 7 nm particle diameter as well as for Ag and (NH₄)₂SO₄ at about 15 nm. Nucleation temperature was kept constant at about 278 K. Experimental data were compared to Kelvin equation and Fletcher theory including the effect of line tension. It was found that heterogeneous nucleation of n-nonane seems to be independent of seed particle composition and starts well below the Kelvin curve. Good agreement was achieved with Fletcher theory including the effect of line tension.     

Charging state of atmospheric nanoparticles during the nucleation burst events

In this work, the charging state of atmospheric nanoparticles was estimated through simultaneous measurements of aerosol size distribution and air ions mobility distribution with the aim to elucidate the formation mechanisms of atmospheric aerosols. The measurements were performed as a part of the QUEST 2 campaign at a boreal forest station in Finland. The overlapping part of the measurement ranges of the particle size spectrometers and air ion mobility spectrometers in the mass diameter interval of 2.6–40 nm was used to assess the percentage of charged particles (charging probability). This parameter was obtained as the slope of the linear regression line on the scatterplot of the measured concentrations of total (neutral + charged) and charged particles for the same diameter interval. Charging probabilities as a function of particle diameter were calculated for different days and were compared with the steady state charging probabilities of the particles in the bipolar ion atmosphere. For the smallest particles detectable by the particle size spectrometers (2.6–5 nm), the high percentages of negatively charged particles were found during the nanometer particle concentration bursts. These values considerably exceeded the values for the steady charging state and it was concluded that negative cluster ions preferably act as condensation nuclei. This effect was found to be the highest in the case of comparatively weak nucleation bursts of nanoparticles, when the rate of the homogeneous nucleation and the concentration of freshly nucleated particles were low. The nucleation burst days were classified according to the concentration of the generated smallest detectable new particles (weak and strong bursts). Approximately the same classification was obtained based on the charge asymmetry on particles with respect to the charge sign (polarity). The probabilities of negative and positive charge on the particles with the diameter of 5–20 nm were found to be nearly equal and they approximately agree with the values corresponding to the steady state charge distribution for negative particles known from lab experiments. It means that the steady charging state was reached during the growing time of particles up to 5 nm. The natural charging state of particles with a diameter between 2.5 and 4.5 nm was estimated by means of a special DMPS setup. Results were found to be in good correlation with the data by the particle size spectrometers and air ion mobility spectrometers.     

Characteristic features of air ions at Mace Head on the west coast of Ireland

Coastal nucleation events and behavior of cluster ions were characterized through the measurements of air ion mobility distributions at the Mace Head research station on the west coast of Ireland in 2006. We measured concentrations of cluster ions and charged aerosol particles in the size range of 0.34–40 nm. These measurements allow us to characterize freshly nucleated charged particles with diameters smaller than 3 nm. The analysis shows that bursts of intermediate ions (1.6–7 nm) are a frequent phenomenon in the marine coastal environment. Intermediate ion concentrations were generally close to zero, but during some nucleation episodes the concentrations increased to several hundreds per cm³. Nucleation events occurred during most of the measurement days. We classified all days into one of seven classes according to the occurrence and type of new particle formation. Nucleation events were observed during 207 days in 2006, most prominently in the spring and summer months. Rain-induced events, in turn, were observed during 132 days. Particle formation and growth events mostly coincided with the presence of low tide. Also small cluster ions (0.34–1.6 nm) were characterized. Average concentrations of small ions were 440 cm^− 3 for the negative ions and 423 cm^− 3 for the positive ions. Average mean mobilities of small ions were 1.86 cm²V^− 1s^− 1 and 1.49 cm²V^− 1s^− 1 for the negative and positive polarities, respectively. Concentrations of small ions were observed to be strongly dependent on the variations of meteorological parameters including wind speed and direction.     

Effects of seed particle size and composition on heterogeneous nucleation of n-nonane

Heterogeneous nucleation of supersaturated n-nonane vapour on seed particles of different size and composition has been investigated using a fast expansion chamber. Monodisperse seed particle sizes were ranging from about 4 nm up to about 24 nm in diameter. By using different types of particle generators WO_x, Ag and (NH₄)₂SO₄ particles were generated. For direct comparison between different particle compositions overlapping sizes have been generated for WO_x and Ag at about 7 nm particle diameter as well as for Ag and (NH₄)₂SO₄ at about 15 nm. Nucleation temperature was kept constant at about 278 K. Experimental data were compared to Kelvin equation and Fletcher theory including the effect of line tension. It was found that heterogeneous nucleation of n-nonane seems to be independent of seed particle composition and starts well below the Kelvin curve. Good agreement was achieved with Fletcher theory including the effect of line tension.     

A closure study of sub-micrometer aerosol particle hygroscopic behaviour

The hygroscopic properties of sub-micrometer aerosol particles were studied in connection with a ground-based cloud experiment at Great Dun Fell, in northern England in 1995. Hygroscopic diameter growth factors were measured with a Tandem Differential Mobility Analyser (TDMA) for dry particle diameters between 35 and 265 nm at one of the sites upwind of the orographic cloud. An external mixture consisting of three groups of particles, each with different hygroscopic properties, was observed. These particle groups were denoted less-hygroscopic, more-hygroscopic and sea spray particles and had average diameter growth factors of 1.11–1.15, 1.38–1.69 and 2.08–2.21 respectively when taken from a dry state to a relative humidity of 90%. Average growth factors increased with dry particle size. A bimodal hygroscopic behaviour was observed for 74–87% of the cases depending on particle size. Parallel measurements of dry sub-micrometer particle number size distributions were performed with a Differential Mobility Particle Sizer (DMPS). The inorganic ion aerosol composition was determined by means of ion chromatography analysis of samples collected with Berner-type low pressure cascade impactors at ambient conditions. The number of ions collected on each impactor stage was predicted from the size distribution and hygroscopic growth data by means of a model of hygroscopic behaviour assuming that only the inorganic substances interacted with the ambient water vapour. The predicted ion number concentration was compared with the actual number of all positive and negative ions collected on the various impactor stages. For the impactor stage which collected particles with aerodynamic diameters between 0.17–0.53 μm at ambient relative humidity, and for which all pertinent data was available for the hygroscopic closure study, the predicted ion concentrations agreed with the measured values within the combined measurement and model uncertainties for all cases but one. For this impactor sampling occasion, the predicted ion concentration was significantly higher than the measured. The air mass in which this sample was taken had undergone extensive photochemical activity which had probably produced hygroscopically active material other than inorganic ions, such as organic oxygenated substances.     

Scavenging of atmospheric ions and aerosols by drifting snow in Antarctica

Measurements of the small-, intermediate-, and large-ion concentrations and the air–earth current density along with simultaneous measurements of the concentration and size distribution of aerosol particles in the size ranges 4.4–163 nm and 0.5–20 μm diameter are reported for a drifting snow period after the occurrence of a blizzard at a coastal station, Maitri, Antarctica. Ion concentrations of all categories and the air–earth current simultaneously decrease by approximately an order of magnitude as the wind speed increases from 5 to 10 ms^− 1. The rate of decrease is the highest for large ions, lowest for small ions and in-between the two for intermediate ions. Total aerosol number concentration decreases in the 4.4–163 nm size range but increases in the 0.5–20 μm size range with wind speed. The size distribution of the nanometer particles shows a dominant maximum at ~ 30 nm diameter throughout the period of observations and the height of the maximum decreases with wind speed. However, larger particles show a maximum at ~ 0.7 μm diameter but the height of the maximum increases with increasing wind speed. The results are explained in terms of scavenging of atmospheric ions and aerosols by the drifting snow particles.     

On the formation and growth of atmospheric nanoparticles

In this paper we summarize recent experimental, theoretical and observational results on the formation and growth of atmospheric nanoparticles. During the last years significant progress has occurred to explain atmospheric nucleation and initial steps of the growth. Due to climatic and health effects of fine and ultrafine particles the formation and growth of new aerosol particles is of growing interest. The question “How and under which conditions does the formation of new atmospheric aerosol particles take place?” has exercised the minds of scientists since the time of John Aitken, who in the late 1880s built the first apparatus to measure the number of dust and fog particles. However, only during the last 15–20 years has the measurement technology developed to such a level that size distributions of nanometer-size particles and concentrations of gases participating in particle formation can be measured in the atmosphere. Also from a theoretical point of view atmospheric nucleation mechanisms have not been resolved: several mechanisms such as ion-induced (or ion mediated) nucleation, ternary and kinetic (barrier-less) nucleation have been suggested. In the most recent theory, the activation of existing neutral and/or ion clusters has been suggested.     

Development of particle size and composition distributions with a novel aerosol dynamics model

An aerosol dynamics model, AEROFOR2, is developed in the context of the BIOFOR project focussing on boreal forest aerosol. It is the second version of a Lagrangian type box model AEROFOR for investigating the formation and growth of particles under clear sky atmospheric conditions. Particles can consist of soluble and insoluble material and the particle population can be externally or internally mixed. AEROFOR2 includes gas phase chemistry and aerosol dynamics, and calculates the number and composition distributions of particles as functions of time. Observed growth rates of the nucleation mode particles after a typical nucleation event are 2–3 nm/h. The model simulations predict that 3·10⁷ molecules cm⁻³ of insoluble organic vapour and less than 6·10⁶ molecules cm⁻³ of soluble vapour condensing onto particles are enough to make them grow in good agreement with the observed growth rates. Then the source rate of the organic vapour must be an order of 10⁵ molecules cm⁻³ s⁻¹, and its saturation vapour density should be below 10⁶ molecules cm⁻³. If the aerosol was initially an internal mixture of soluble (70%) and insoluble (30%) constituents it transformed to an externally mixed aerosol during the simulation. By applying the externally‐mixed aerosol based on measured soluble volume fractions, it was concluded that the modelled soluble fraction of the nucleation mode was too low in comparison with the measurements, and thus, a part of the condensable organic vapour must be water soluble.     

A review of global stratospheric aerosol: Measurements, importance, life cycle, and local stratospheric aerosol

Stratospheric aerosol, noted after large volcanic eruptions since at least the late 1800s, were first measured in the late 1950s, with the modern continuous record beginning in the 1970s. Stratospheric aerosol, both volcanic and non-volcanic are sulfuric acid droplets with radii (concentrations) on the order of 0.1–0.5 µm (0.5–0.005 cm^− 3), increasing by factors of 2–4 (10–10³) after large volcanic eruptions. The source of the sulfur for the aerosol is either through direct injection from sulfur-rich volcanic eruptions, or from tropical injection of tropospheric air containing OCS, SO₂, and sulfate particles. The life cycle of non-volcanic stratospheric aerosol, consisting of photo-dissociation and oxidation of sulfur source gases, nucleation/condensation in the tropics, transport pole-ward and downward in the global planetary wave driven tropical pump, leads to a quasi steady state relative maximum in particle number concentration at around 20 km in the mid latitudes. Stratospheric aerosol have significant impacts on the Earth's radiation balance for several years following volcanic eruptions. Away from large eruptions, the direct radiation impact is small and well characterized; however, these particles also may play a role in the nucleation of near tropopause cirrus, and thus indirectly affect radiation. Stratospheric aerosol play a larger role in the chemical, particularly ozone, balance of the stratosphere. In the mid latitudes they interact with both nitrous oxides and chlorine reservoirs, thus indirectly affecting ozone. In the polar regions they provide condensation sites for polar stratospheric clouds which then provide the surfaces necessary to convert inactive to active chlorine leading to polar ozone loss. Until the mid 1990s the modern record has been dominated by three large sulfur-rich eruptions: Fuego (1974), El Chichón (1982) and Pinatubo (1991), thus definitive conclusions concerning the trend of non-volcanic stratospheric aerosol could only recently be made. Although anthropogenic emissions of SO₂ have changed somewhat over the past 30 years, the measurements during volcanically quiescent periods indicate no long term trend in non-volcanic stratospheric aerosol.     

Hygroscopic growth of aerosol particles and its influence on nucleation scavenging in cloud: Experimental results from Kleiner Feldberg

The hygroscopic growth of individual aerosol particles has been measured with a Tandem Differential Mobility Analyser. The hygroscopic growth spectra were analysed in terms of diameter change with increasing RH from 20% to 85%. The measurements were carried out during the GCE cloud experiment at Kleiner Feldberg, Taunus, Germany in October and November 1990.Two groups of particles with different hygroscopic growth were observed. The less-hygroscopic group had average growth factors of 1.11, 1.04 and 1.02 for particle diameters of 50, 150 and 300 nm, respectively. The more-hygroscopic group had average growth factors of 1.34, 1.34, and 1.37 for the same particle diameters. The average fraction of less-hygroscopic particles was about 50%. Estimates of the soluble fractions of the particles belonging to the two groups are reported.Hygroscopic growth spectra for total aerosol, interstitial aerosol and cloud drop residuals were measured. A comparison of these hygroscopic growths of individual aerosol particles provides clear evidence for the importance of hygroscopic growth in nucleation scavenging. The measured scavenged fraction of particles as a function of diameter can be explained by the hygroscopic growth spectra.     

Vertical fluxes and micrometeorology during aerosol particle formation events

Fluxes of aerosol particles with sizes larger than 10 nm together with fluxes of momentum, sensible and latent heat and CO₂ were measured 10 m above a Scots pine forest with the eddy covariance method. During days when nucleation events were observed particle size distribution measurements showed particle growth from 3 nm sizes to the Aitken mode. Analysis of the experimental data showed systematic differences in fluxes during the days when new particle production was observed compared to other days. During the nucleation events the particle flux measurements showed downward aerosol particle transport, i.e., indicating an elevated source, with respect to the measurement level, of particles larger than 10 nm. Furthermore the turbulence intensity and the heat fluxes were observed to be significantly higher. Evidences of mesoscale circulation were observed in wind speed records as well as in turbulent fluxes on nucleation days. The measurement results show that micrometeorology, the synoptic scale conditions and the particle formation are closely related.     

Relevance of ion-induced nucleation of sulfuric acid and water in the lower troposphere over the boreal forest at northern latitudes

The relevance of ion-induced nucleation of sulfuric acid and water (IINSW) in the troposphere over the boreal forest at northern latitudes is investigated by combining two existing and previously published models (MALTE — model to predict new aerosol formation in the lower troposphere; PARNUC — a parameterized steady-state model of neutral and ion-induced nucleation of sulfuric acid and water for atmospheric conditions). Simulations were performed for 4 days with observed new particle formation at ground level by using input data from the SMEAR II station in Hyytiälä, Finland. The selected days were chosen to cover a wide range of values of the parameters most relevant for IINSW. The results showed that ion-induced nucleation of sulfuric acid and water can contribute up to 15% to the total amount of newly formed particles in the size range of 3–10 nm inside the mixed layer at the Hyytiälä site. The importance of IINSW seemed to increase in the free troposphere above the boundary layer, however, lack of measurements in the vertical structure of the input parameters suggest that the model results are burdened with high uncertainties.     

A study of aerosol particle sizes in the atmosphere of Athens, Greece, retrieved from solar spectral measurements

This work aims at determining the aerosol particle radii in the atmosphere of Athens. Such a work is carried out in Athens for the first time. For this purpose, solar spectral direct-beam irradiance measurements were used in the spectral range 310–575 nm. To estimate the particle radius from aerosol optical depth retrieval, a minimization technique was employed based on the golden-section search of the difference between experimental and theoretical values of the aerosol optical depth. The necessary Mie computations were performed based on the algorithm LVEC.In this study, the mean particle radius of a given distribution was calculated every 30 min during cloudless days in the period November 1996 to September 1997. The largest particles were observed in the summer and the smallest during winter. The result was verified by the increased values of the aerosol optical depth and the turbidity factors calculated in the summer. The differences in the diurnal variation from season to season are attributed to the prevailing wind regime, pollutant emission and sink rates in the atmosphere of Athens.     

Particulate contribution to extinction of visible radiation: Pollution, haze, and fog

A data set acquired by eight particle-dedicated instruments set up on the SIRTA (Site Instrumental de Recherche par Télédétection Atmosphérique, which is French for Instrumented Site for Atmospheric Remote Sensing Research) during the ParisFog field campaign are exploited to document microphysical properties of particles contributing to extinction of visible radiation in variable situations. The study focuses on a 48-hour period when atmospheric conditions are highly variable: relative humidity changes between 50 and 100%, visibility ranges between 65 and 35 000 m, the site is either downwind the Paris area either under maritime influence. A dense and homogeneous fog formed during the night by radiative cooling. In 6 h, visibility decreased down from 30 000 m in the clear-sky regime to 65 m within the fog, because of advected urban pollution (factor 3 to 4 in visibility reduction), aerosol hydration (factor 20) and aerosol activation (factor 6). Computations of aerosol optical properties, based on Mie theory, show that extinction in clear-sky regime is due equally to the ultrafine modes and to the accumulation mode. Extinction by haze is due to hydrated aerosol particles distributed in the accumulation mode, defined by a geometric mean diameter of 0.6 μm and a geometric standard deviation of 1.4. These hydrated aerosol particles still contribute by 20 ± 10% to extinction in the fog. The complementary extinction is due to fog droplets distributed around the geometric mean diameter of 3.2 μm with a geometric standard deviation of 1.5 during the first fog development stage. The study also shows that the experimental set-up could not count all fog droplets during the second and third fog development stages.     

Hygroscopicity of pre-existing particle distribution and formation of cloud droplets: a model study

The effects of the hygroscopicity of a pre-existing particle distribution and condensation of nitric acid on cloud droplet formation were studied by using an air parcel and multicomponent condensation model. The pre-existing particle distribution used is a bimodal distribution in which the particles are assumed to be internally mixed, i.e. they are composed partly from ammonium nitrate salt and partly from some insoluble substance. The mean diameters of the distributions and the mass fraction of soluble salt were varied in the simulations. Generally, the number of activated cloud droplets was found to be increased, when the initial mass fraction of salt (i.e. the initial amount of salt) was increased. However, the effects of increased initial amount of salt on the cloud droplet formation were not straightforward in all cases studied. The effects of the condensing hygroscopic substance, with initial nitric acid concentrations of 0.1, 1.0 and 10.0 ppbv on the activation of cloud droplets were also studied. The number of activated droplets increased when the initial concentration of nitric acid was increased.     

Continuous scanning of the mobility and size distribution of charged clusters and nanometer particles in atmospheric air and the Balanced Scanning Mobility Analyzer BSMA

Measuring of charged nanometer particles in atmospheric air is a routine task in research on atmospheric electricity, where these particles are called the atmospheric ions. An aspiration condenser is the most popular instrument for measuring atmospheric ions. Continuous scanning of a mobility distribution is possible when the aspiration condenser is connected as an arm of a balanced bridge. Transfer function of an aspiration condenser is calculated according to the measurements of geometric dimensions, air flow rate, driving voltage, and electric current. The most complicated phase of the calibration is the estimation of the inlet loss of ions due to the Brownian deposition. The available models of ion deposition on the protective inlet screen and the inlet control electrofilter have the uncertainty of about 20%. To keep the uncertainty of measurements low the adsorption should not exceed a few tens of percent. The online conversion of the mobility distribution to the size distribution and a correct reduction of inlet losses are possible when air temperature and pressure are measured simultaneously with the mobility distribution. Two instruments called the Balanced Scanning Mobility Analyzers (BSMA) were manufactured and tested in routine atmospheric measurements. The concentration of atmospheric ions of the size of about a few nanometers is very low and a high air flow rate is required to collect enough of ion current. The air flow of 52 l/s exceeds the air flow in usual aerosol instruments by 2–3 orders of magnitude. The high flow rate reduces the time of ion passage to 60 ms and the heating of air in an analyzer to 0.2 K, which suppresses a possible transformation of ions inside the instrument. The mobility range of the BSMA of 0.032–3.2 cm² V^− 1 s^− 1 is logarithmically uniformly divided into 16 fractions. The size distribution is presented by 12 fractions in the diameter range of 0.4–7.5 nm. The measurement noise of a fraction concentration is typically about 5 cm^− 3 and the time resolution is about 10 min when measuring simultaneously both positive and negative ions in atmospheric air.     

Optical attenuation in fog at a wavelength of 1.55 micrometers

As part of a series of studies on laser propagation for terrestrial free space optical (FSO) telecommunications or laser telecommunications, an experiment was conducted to determine the relationship between visibility in fog and optical attenuation (dB/km) at a laser wavelength of 1.55 μm. In the telecommunications industry, a semi-empirical equation, called the Kruse formula [Kruse, P.W., McGlauchlin, L., McQuistan, R.B., 1962. Elements of infrared technology: generation, transmission, and detection. John Wiley and Sons, New York] is typically used to calculate expected attenuation for a given meteorological visibility. The Kruse formula, however, was developed to relate meteorological visibility to optical attenuation over wavelengths from the visible to the near infrared (IR), and for dust and small particle aerosols with dimensions much smaller than the wavelength. Typically, suspended small aerosols have diameters that average about 0.1 μm while fog droplets have diameters that range upward from 2.5 μm with mean diameters that exceed 10 μm in some fogs. Therefore, application of the Kruse formula to attenuation in fog is not appropriate since fogs consist mainly of particles much larger than the laser wavelength. As part of the experiment, a transmissometer with an 85-m baseline and a dynamic range of 60 dB operated for thirteen months in an area prone to radiation fog. A commercial visibility sensor, similar to those used at airports, was located near the middle of the optical path of the transmissometer and operated over the same period. The largest attenuation measured at this site was just over 300 dB/km, corresponding to a visibility of 32 m. The key finding of the study is that the generally accepted Kruse formula relating visibility and optical attenuation may be too pessimistic at low visibilities, and actual attenuation values for a given visibility may be more than 20% lower than previously thought. At visibilities exceeding about 650 m, the Kruse formula gives a good estimate of optical attenuation.     

Observation of regional new particle formation in the urban atmosphere

Long-term measurements of fine particle number-size distributions were carried out over 9.5 yr (May 1997–December 2006), in the urban background atmosphere of Helsinki. The total number of days was 3528 with about 91.9% valid data. A new particle formation event (NPF) is defined if a distinct nucleation mode of aerosol particles is observed below 25 nm for several hours, and it shows a growth pattern. We observed 185 NPF events, 111 d were clear non-events and most of the days (around 83.5%) were undefined. The observed events were regional because they were observed at Hyytiälä (250 km north of Helsinki). The events occurred most frequently during spring and autumn. The observed formation rate was maximum during the spring and summer (monthly median 2.87 cm⁻³ s⁻¹) and the modal growth rate was maximum during late summer and Autumn (monthly median 6.55 nm h⁻¹). The events were observed around noon, and the growth pattern often continued on the following day. The observation of weak NPF events was hindered due to pre-existing particles from both local sources. It is clear that regional NPF events have a clear influence on the dynamic behaviour of aerosol particles in the urban atmosphere.     

Relationship of heterogeneous nucleation and condensational growth on aerosol nanoparticles

Heterogeneous nucleation and condensation of dibutylphthalate, octadecane, octadecanol, and octadecanoic acid vapors at various pressures on insoluble AgCl and Ag nanoparticles in a turbulent mixing condensation nuclei counter (TMCNC) have been studied theoretically. A method to interpret the particle size distributions measured with a DMA and estimate the parameters for nucleation on single particles is proposed. Based on this semi-empirical method, the Gibbs free energy is calculated and a rate of heterogeneous nucleation on single particles is estimated directly from the experimental “condensation spectra” of inactive and active CN using the DMA data. In some cases, the dependence of the Gibbs nucleation energy on the vapor supersaturation had two maximums and one minimum, instead of one maximum as described by Gibbs' classical thermodynamics of phase transitions. This phenomenon, called “double barrier nucleation” (DBN) is caused by the surface heterogeneity of nano-CN; this is first experimental verification of DBN that had been previously predicted theoretically. Two types of heterogeneity may be present: topographic or energetic. Focusing on energetic heterogeneity, a theoretical model of DBN for spherical geometry is developed. The surface heterogeneity for insoluble nano-sized CN is shown to be critical to explaining the unusual transformation of a monomodal size distribution of inactive CN into a bimodal distribution of activated CN when coagulation is excluded. Future studies will be directed toward more data for further refining the theory and developing a model that simultaneously accounts for both types of surface heterogeneity of nano-CN.