Ion induced nucleation based-generator of strictly singly charged (solid - liquid) polydisperse particles up to 1 μm

Highlights

• Polydisperse aerosol generator based on ion induced nucleation is presented and characterized in this paper.

• The generator produces particles from low melting points solids or low vapor pressure liquids.

• The generator is based on evaporation-condensation method used in the widely used Sinclair Lamer generators.

• Atmospheric ions produced by a corona source are used as seed residues.

• Singly charged spherical particles (solid or liquid) up to 1 μm are produced by the generator without neutralizer.

Abstract:

Calibration of optical counters, condensation particle counters, electrical impactors, filtration efficiency measurement, heterogeneous condensation (and many other laboratory applications in aerosol science and technology) are widely used with singly charged particles classified by a differential mobility analyzer with particles smaller than 1 μm. Nevertheless the method presents an ambiguity because of the presence of multi charged large particles of the same mobility than singly ones. This ambiguity results in uncertainty in the concentration regarded as a reference. The present paper reports a generation method of strictly singly (positive and negative) charged liquid and solid particles up to 1 μm.

The generation-method is based on ion-induced nucleation of singly charged atmospheric ions. The method uses evaporation and controlled condensation of liquids or solid on seed ions. The ions grow and form wide continuum polydisperse distributions of positive or negative liquid/solid polydisperse particles up to 1 μm. The ions are produced with a corona or radioactive or soft X ray or UV source in filtered dry stream air. The final formed droplets or spheres around the ions keep the positive or negative initial charge of the ions.

Vapors from low meting point solids or high boiling point and low vapor pressure liquids are used in this vaporizer-condenser generator. The paper focuses on particles with diameters up to 1 μm dedicated to classical differential mobility analyzers (DMA). The results with larger particles produced with the same generator design using heterogeneous condensation of vapors on sub 10 nm singly charged particles from melting NiCr (Kanthal) wires in air will be published in a second paper (manuscript in preparation). A companion paper on the saturation ratio profiles modeling with Comsol and comparison theory-experiments is in preparation.

Introduction

Liquid or solid aerosol particles (monodisperse or polydisperse) in the super and submicronic diameter range are needed in laboratories for filtration efficiency measurement, calibration of instruments, deposition-losses measurements in tubes or on surfaces, coating, evaporation/condensation studies, inhalation studies, charge states measurements and many others. Atomization/nebulization and evaporation-condensation are the most used simple and robust cheap methods of particles generation for laboratory applications. Atomization/electrospray of solutions (or suspensions) produces particles from 1 nm to 500 nm mostly. Evaporation-condensation of pure compound at high temperatures in an oven or from hot melting material (glowing wires, spark electrodes, lasers) produces particles from 1 to 30 nm single spheres non agglomerates particles.

Large particles (up to 10 μm) are produced by controlled condensation of liquid or low melting point solid compounds upon nuclei. Sinclair and La Mer (1949) have introduced the controlled condensation method in their well-known generator. Others have made investigations and improvements of the Sinclair-Lamer method to produce quasi monodisperse particles up to 7 μm (Muir &Davis 1967; Swift, 1967, Okada et al., 1969; Prodi (1972); Nicolaon et al. (1970); Liu and Lee (1975)) The other solution to produce monodisperse particles is to classify with a differential mobility analyzer particles produced by polydisperse distributions sources (Liu and Pui (1974)).

The differential mobility analyzer (DMA) is used to select particles in a narrow size range from existing polydisperse aerosol sources. The DMA is attractive and unique classifier-generator to study size dependency properties for many aerosol applications.

The main problem of the DMA as a classifier-generator especially for particles larger than 0.1 μm is the presence of multi charged particles mixed with the singly charged ones. Indeed most of the polydisperse aerosol sources produce singly charged particles mixed with multi charged ones. Neutralizers (mostly radioactive or soft X ray based) are used to bring the polydisperse generated aerosol population to a steady charge state with low concentration of multi charged particles.

The DMA classifies particles according to electrical mobility which depends on the size and charge of the particles. Particles of a given mobility can include singly, doubly and triply charged particles of different diameters. At the Boltzman charge equilibrium (for example) after the ‘neutralizer’ upstream of the DMA a non-negligible population of doubly and even triply charged particles are mixed with the singly charged ones depending on the size (Liu & Pui, 1974).

Efforts have been devoted by several teams to solve this problem. Romay-Novas and Pui (1988) have discussed a method to improve the monodispersity by using a single stage impactor downstream of the DMA to remove the larger multi charged particles of the same mobility. The cut-off of the impactor needs to be adjusted to the geometric mean diameter of the singly and doubly charged particles.

Gupta and McMurry (1988) developed a bipolar charger that avoids the production of multi charged particles in the 0.1–1 μm diameter range. Their charger consists of a 23 cm long, 2.1 cm ID cylindrical tube internally coated with 0.09 μCi ⁶³Ni radioactive source. Zero charged particles at the inlet of the charger are exposed to low ion concentrations for short time periods. Ion concentrations and residence time must be adjusted to ensure that the particles are mostly neutral or singly charged. For example 1.09 μm particles leaving their charger have far fewer multi charged particles (21%) whereas 72% in the classical ‘neutralizer. Vivas et al. (2010) applied pulse voltage of variable amplitude and duration to a corona charger to reduce the multi charging level similar to bipolar charger with an increase of the singly charging efficiency. Recently Chen et al. (2019) used a corona based unipolar charger where they control the Nt product to maximize the singly charged fraction of sub-micrometer particles. N is the concentration of ions per cm³, t is the residence time (Pui & Liu 1974). They managed to increase singly charged population of particles in the 20–200 nm versus the concentration given by a classical bipolar charger.

The other way or method is to use controlled condensation on singly charged particles produced with an atomizer or an oven. Uin et al. (2009) introduced a method to produce standard large aerosols from singly charged small size nuclei for producing standard aerosols which eliminates the problems arising from the presence of the multi charged particles. They used a tube furnace (~1050 °C) to produce silver nucleus (~18 nm), a neutralizer and evaporator-condenser Sinclair-Lamer generator. Their Sinclair Lamer generator uses heated di octyl phtatlate (DOP) between 60 and 225 °C. The use of DOP is not encouraged anymore because of its toxicity. Because DOP causes cancer and birth effects among others its use is regulated in EU, USA and many other countries. Di ethyl octyl sebacate nowadays is used thanks to its inhalation safety.

Yli-Ojanperä et al. (2010) use mainly the same set up to develop a standard generator of singly charged particles (traceable particle number concentration standard: singly charged aerosol reference, SCAR). They use an atomizer-dryer before the oven (as suggested by Bartz et al., 1987) before a radioactive source and a differential mobility analyzer to produce small NaCl particles as condensation nucleus. They preferred the atomizer (sodium chloride) thanks to the stability of the nuclei source. Indeed the atomization method of Bartz et al. (1987) is more stable than the evaporation in a tubular oven method given by Scheibel and Porshtendorfer (1983). The aerosol selected by the DMA is then led through a saturator filled with di-octyl sebacate (DOS). A vertical descendant evaporator-condenser similar to the instrument introduced by Liu and Lee (1975) is used by the Finn team to growth the particles. They were able to produce singly charged particles up to 500 nm with an error of 0.15% of multi charged particles.

These methods are based on the use of neutralized particles which cannot ensure 100% of singly charged particles before and after growth. To circumvent or solve this problem it's convenient to start with strictly singly charged nuclei before the evaporation-condensation process for the growth. Atmospheric ions are the solution since one can produce them bipolar in high concentrations (up to 10⁹ ions/cm³) with a home-made and cost effective corona source without any permission nor health risk. One have to keep in mind that the method produces NOx and O₃.

Evaporation-condensation method is a well-known (documented and used) method in aerosol science and technology for different purposes. For example Suh et al. (2005); Choi and Kim (2007), used ethylene glycol in a particle size magnifier (PSM) to enhance the charge of nanoparticles.

Kogan and Burnasheva (1960) developed a kind of mixing CPC called particle size magnifier where they mixed a room temperature aerosol in a conical nozzle with a very hot vapor of dibutylphtalate. Their concept was used later by many others in different designs for turbulent CPCs or PSMs. Kogan and Burnasheva (1960) managed to activate and enlarge sub 10 nm aerosol particles to stable droplets of around 1 μm. Okuyama et al. (1987) used the evaporation condensation in his PSM to study the homogeneous nucleation of DBP.

The key idea is to combine condensational growth of nanoparticles and unipolar charging is given by equation (1) (Hinds, 1999).

$$\frac{d\left(Dp\right)}{dt}=\frac{4MDv}{R{\rho }_{p}Dp}\left(\frac{P_{\infty }}{T}-\frac{P_d}{T}\right)\cong \frac{4MDv{P}_d}{R{\rho }_{p}TDp}\left(S-1\right)\propto \frac{1}{Dp}$$

Where M is the molecular weight of the condensing vapor; ρ_p is the density of the liquid (or solid) working fluid; Dv the diffusion coefficient of the vapor; R the gas constant; $P_{\infty }$the partial vapor of the vapor close to the particle; P_d and T are the vapor pressure and temperature on the surface of the growing particle; T the temperature of the surrounding vapor; S the saturation ratio of the vapor. The generator presented here uses the same evaporation condensation precept. The only (but significant) difference is the use of singly charged atmospheric ions as nuclei. This allows to simplify considerably the set up and give better results in the sense that there is no multi charged droplets or spheres in the final aerosol. The method no longer asks a neutralizer nor dryer in the case of atomization. The source of seed ions can be unipolar (corona) or any bipolar (radioactive or soft X ray source) based neutralizer. The method is not exactly evaporation condensation but ion induced nucleation through evaporation condensation. Ion induced nucleation has been widely studied theoretically and experimentally because of its importance in atmospheric science.

Previous designs have focused on the monodispersity generation (standard deviation between 1.2 and 1.4 in the case of the Liu and Lee (1975)) or quasi monodispersity (Sinclair–Lamer (1949); Ristovski et al., 1998)). The main goal of the present generator concept is to produce a broad continuum of polydisperse unipolar particles, as is the case of the atomizers. Nevertheless, there is a difference with the atomizers. The fundamental difference mays in the charge state of the generated droplets or spheres. Atomizers indeed produce bipolar multi charged particles.

The present generator is designed as singly charged polydisperse unipolar source to be used with a DMA. Geometric standard deviation of the final droplets diameters varies from 1.7 to 2.4, which is considered polydisperse (Chen, 1993).

Charles Thomson Rees Wilson (1896) have shown the ability of positive and negative atmospheric ions produced with Röntgen X ray source to form particle droplets in a clean and free atmosphere 124 years ago. He was the first to talk in terms of ion induced-nucleation.

During the last decades mixing and laminar condensation particle counters (CPCs) have shown their ability to activate and detect positive and negative atmospheric ions produced with a radioactive source neutralizer and or a corona source (Gamero (2000; 2002); Iida et al., 2011; Hering et al., 2016; Kangasluoma et al., 2016). Water, butanol, ethylene glycol, dibutyl phtalate and di ethylene glycol as working fluid in CPCs have been used to activate atmospheric ions (Kangasluoma & Attoui, 2019; Maisser & Hogan, 2017) with detection efficiencies between 30 and 70% (Attoui (2018); Hering et al. (2016); Vanhanen et al. (2011)). The present generator uses these findings to produce singly charged particles in a laminar device given in Fig. 1 or mixing device given in figure S3.

As in the case of laminar or turbulent mixing CPCs singly charged atmospheric ions are activated and grown in a condenser (growth tube). The carrier gas (air in the present study) is saturated with vapor from heated high boiling point and viscous liquids or low melting point solids.

The probability to produce doubly charged atmospheric ions with unipolar (corona) or bipolar (soft X-ray or radioactive material) device is equal to zero. As the consequence of that the droplets or particles produced by the final coated ion keep the same electrical charge that the original residue ion. NO_x and O₃ production presents the unique disadvantage when corona source is used to produce ions.

Tandem DMA experiments with a Kr85 neutralizer in the middle (Kim et al., 2005) to check if multi charged particles could appear by coagulation have been carried for glycerol and stearic acid as suggest by one of the reviewers. After 10 h experiments for each compound no multi charged particles have been detected.