Environmental Science and Pollution Research

, Volume 24, Issue 5, pp 4480–4493 | Cite as

Novel aerosol analysis approach for characterization of nanoparticulate matter in snow

  • Yevgen Nazarenko
  • Rodrigo B. Rangel-Alvarado
  • Gregor Kos
  • Uday Kurien
  • Parisa A. Ariya
Research Article

Abstract

Tropospheric aerosols are involved in several key atmospheric processes: from ice nucleation, cloud formation, and precipitation to weather and climate. The impact of aerosols on these atmospheric processes depends on the chemical and physical characteristics of aerosol particles, and these characteristics are still largely uncertain. In this study, we developed a system for processing and aerosolization of melted snow in particle-free air, coupled with a real-time measurement of aerosol size distributions. The newly developed technique involves bringing snow-borne particles into an airborne state, which enables application of high-resolution aerosol analysis and sampling techniques. This novel analytical approach was compared to a variety of complementary existing analytical methods as applied for characterization of snow samples from remote sites in Alert (Canada) and Barrow (USA), as well as urban Montreal (Canada). The dry aerosol measurements indicated a higher abundance of particles of all sizes, and the 30 nm size dominated in aerosol size distributions for the Montreal samples, closely followed by Barrow, with about 30% fewer 30 nm particles, and about four times lower 30 nm particle abundance in Alert samples, where 15 nm particles were most abundant instead. The aerosolization technique, used together with nanoparticle tracking analysis and electron microscopy, allowed measurement of a wide size range of snow-borne particles in various environmental snow samples. Here, we discuss the application of the new technique to achieve better physicochemical understanding of atmospheric and snow processes. The results showed high sensitivity and reduction of particle aggregation, as well as the ability to measure a high-resolution snow-borne particle size distribution, including nanoparticulate matter in the range of 10 to 100 nm.

Keywords

Particles in snow Nanoparticulate Atmospheric aerosols Aerosol Ice nucleation Ice nuclei Nanoparticles Ultrafine particles Aerosolization Cloud condensation nuclei 

Introduction

Cloud formation processes play a key role in climate change because water vapor and cloud droplets absorb and emit infrared light, thereby influencing the global radiation budget and water cycle (Archer 2011). Consequently, nucleation of water droplets and ice crystals in the atmosphere is a focus of a large body of research (Kärcher 2012; Murray et al. 2012; Yakobi-Hancock et al. 2013). Pure water droplets that contain no particulate matter do not form ice crystals at temperatures above −38 °C. Without ice nucleating particles in water droplets, temperature has to be lower than −38 °C for water to freeze by homogeneous ice nucleation (Koop et al. 2000), which occurs when water starts crystallizing on its own by self-assembly of water molecules (Khvorostyanov and Curry 2014). For ice to form in the atmosphere at temperatures warmer than −38 °C, airborne particles must be present to initiate heterogeneous ice nucleation, during which ice crystallization starts on a surface. The airborne particles providing their surfaces for initiation of ice crystallization are called ice nuclei (IN) and can be of various sizes and composed of diverse materials (Barkan et al. 2005; Xuan et al. 2004). Similarly, water droplets rarely form at the conditions found in the atmosphere. This process requires water vapor supersaturation of several hundred percent. Instead, atmospheric water droplets usually form heterogeneously in the presence of particles called cloud condensation nuclei (CCN) (Ariya et al. 2009). One of the most common IN and CCN in the atmosphere is mineral dust (Archuleta et al. 2005), which originates in locations such as the Sahara Desert. Mineral dust is considered one of the most efficient ice forming nuclei, partly due to its abundance in global atmospheric aerosols. Certain marine-, vegetation-, and soil-derived aerosols, bacteria, fungi, and their fragments also efficiently nucleate ice (Bigg and Leck 2001; Fröhlich-Nowoisky et al. 2014; Georgakopoulos et al. 2009; Mortazavi et al. 2014a, b, 2015; Mortazavi et al. 2008; Schnell and Vali 1976; Vali et al. 1976). Recently, research pointed to ice nucleating activity of nanosized organic compounds, possibly peptides, as well as nanosized inorganic compounds such as metal oxides (Rangel-Alvarado et al. 2015). In addition to the above categories of ice nuclei incorporated in the snow crystals, snowpack contains pollution particles (Ariya et al. 2014; Ariya et al. 2011; Kos and Ariya 2010; Nemirovskaya and Kravchishina 2015; Radke et al. 1980; Skeie et al. 2011). These include aerosols from engine exhaust (Kuoppamäki et al. 2014; Nazarenko et al. 2015), power plants (Barrie 1980), mining activities (Williamson et al. 2004), and other pollution sources. Characterization of particles in snow is a crucial element of research to understand the sources and properties of particles in the atmosphere and, consequently, their effects on weather, climate, and pollution (Ariya et al. 2014; Rosenfeld 2000). The physicochemical characterization and quantification of snow-borne particles, especially nanometric material, are technically challenging due to instrumental limitations of traditionally used analytical techniques for counting, sizing, and physicochemical analysis. Specifically, electron microscopy techniques are expensive and time-consuming and require staining with carcinogenic uranium-based compounds to detect most organic particles, which are otherwise nearly electron-transparent. Organic particles are also unstable under the electron beam (Egerton et al. 2004). Nanoparticle tracking analysis (NTA) is expected to have a progressively reducing efficiency of detection of particles, particularly smaller than 50–100 nm, as light scattering efficiency of smaller particles declines (Jacobson 2005). Particle detection efficiency also fluctuates depending on their physicochemical nature, which influences how efficiently they scatter light. Dynamic light scattering is biased towards larger sizes and cannot analyze melted snow samples in original state, requiring pre-concentration. This leads to a challenge in determining the role of different types and sizes of particles found in snow matrices using the traditional particle analysis techniques. Another problem that accompanies sample preparation for traditional analytical techniques that require dry or pre-concentrated samples is agglomeration of particulate matter. Pre-concentration, if needed to improve detection (Domingos et al. 2009), can itself alter the original particles that are being analyzed.

This study presents a first-of-its-kind systematic characterization of nanoparticles in snow, which involved the development of a new aerosol-mediated characterization technique. The purpose of this technique is to analyze particles in snow independent of their ice-nucleating abilities. We also tested the applicability of existing complementary analytical techniques to environmental snow samples obtained from High Arctic and urban locations. New types of characterization data were obtained for snow samples from Alert, Nunavut in Canada; Barrow, Alaska in the USA; and from the urban site of Montreal, Quebec, Canada.

Materials and methods

Sampling and storage

We collected samples at several sites in Alaska, USA (2009) and Nunavut (2006) and Montreal, Quebec (in 2006 and 2014), Canada. Samples from Alaska were collected in Barrow (71.31° N, 156.6° W), approximately 400 m to the southeast of the Barrow Arctic Research Center (BARC). Snow originated from a relatively pristine snow region in a designated clean air sector, as part of the 2009 international Ocean-Atmosphere-Sea Ice-Snowpack (OASIS) campaign.

Samples obtained from Alert, Nunavut (74.70° N, 95.05° W) were collected at the northern tip of Ellesmere Island at a pristine site in the vicinity of the Environment Canada Global Atmosphere Watch (GAW) station in the High Arctic in May and June 2006. Local contamination from a military base was a generator, an incinerator, and vehicle traffic. Sampling was performed upwind approximately 6.5 km away from the base on a flat plateau with restricted vehicle traffic.

Arctic snow collection and storage have previously been described in detail elsewhere (Ariya et al. 2014; Kos and Ariya 2006a; Kos and Ariya 2006b; Kos and Ariya 2010; Mortazavi et al. 2006; Poulain et al. 2007). Arctic snow from the top 3 cm of the snowpack was collected into 950-mL amber glass jars (Daniels Scientific, Charleston, SC, USA), with PTFE-lined lids, pre-cleaned as previously described (Rangel-Alvarado et al. 2015). A sterile HDPE spoon was used to transfer snow (Bel-Art Products, Inc., Wayne, NJ, USA). The Barrow (Alaska) and Alert (Nunavut) samples were kept in a −20 °C walk-in freezer until shipped to Montreal by plane in commercial coolers filled with ice packs (Coleman Company, Inc., Wichita, KS, USA) and snow. The snow samples did not warm up above −5 °C before reaching the laboratory at McGill University in Montreal, QC, Canada where they were immediately transferred to a −20 °C freezer.

Fresh snow samples from Montreal were collected at two sites. The first location was near McGill University downtown campus in Mont-Royal Park, away from any trails and vehicular or pedestrian traffic at −11 °C in February of 2006. Snow collection occurred in a wooded area of the abovementioned park (45° 35′ N, 70° 52′ W) after a heavy snowfall. Montreal receives regular fresh snow precipitation approximately 5 months a year. The second set of snow samples was obtained during a light snowfall, also in Mont-Royal Park (45° 30′ 45″ N, 73° 35′ 15″ W) immediately following a heavy nightly snowfall. Collection time was between 1 and 3 p.m. on March 23, 2014. Meteorological conditions were as follows: air temperature between 0 and −1 °C, light snow, with wind NNE ∼43 km/h, UV index of 2 (low), RH 87%, atmospheric air pressure 1014.4 mb, and visibility of 4 km (The Weather Channel, LLC 2014). The collection site was not subject to human or pet traffic. We also avoided collection in the vicinity of freshly fallen vegetation and animal tracks or droppings. Following collection of snow, we brought the snow samples in the pre-cleaned jars, certified according to EPA standards (U.S.EPA 1992), to the laboratory in a wheeled cooler. The distance between the sampling location and the laboratory was around 2 km. Upon delivery to the laboratory, the jars were immediately placed in a freezer. The 2006 samples were kept in a −20 °C freezer and the recent 2014 samples in a −39 °C freezer. Our above-referenced studies confirmed that these storage conditions maintain snow in near-pristine condition.

Processing of melted snow

After transfer from the original storage containers, snow samples were stored in sterile Falcon™ 50-mL Conical Centrifuge Tubes (Corning Life Sciences, Inc.) at −39 °C until processed. Processing of melted snow by filtration and/or dialysis was intended to determine contribution of different particle size fractions to chemical composition of particulate matter in snow (total organic carbon) and explore influence of dissolved matter on measurement results. The samples were allowed to melt at room temperature (23 °C) and then a fraction of each sample filtered through a Millex syringe filter with a hydrophilic polyvinylidene fluoride (PVDF) membrane, 0.22 μm pore size, part no. SLVV033RS (EMD Millipore Corp., Billerica, MA, USA). Another fraction of each sample was dialyzed using a Standard Grade Regenerated Cellulose membrane (Spectrum Laboratories, Inc., Rancho Dominguez, CA, USA) with a molecular weight cutoff of 5–8 kDa. The filled membrane tubes were dialyzed at room temperature (23 °C) in a bath of ultrapure water (EMD Millipore Corp.) with stirring for 5 days. The water bath was replaced once daily.

Chemical analysis of melted snow

Unprocessed and filtered melted snow samples were analyzed for total organic carbon (TOC) and by ion chromatography (IC) using standard automatic procedures on an Aurora 1030W TOC Analyzer (O.I. Corp., College Station, TX) and a Dionex ICS-500 ion chromatograph. The methodologies of these analyses are described in detail in Supplementary Information.

Nanoparticle tracking analysis of melted snow

Unprocessed, filtered, and/or dialyzed snow samples were analyzed on a NanoSight NS500 instrument (Malvern Instruments, Ltd., Malvern, Worcestershire, UK). The principle of operation of the instrument lies in using a microscope equipped with a CCD camera to digitally film scattered light from individual particles in liquid samples illuminated by a 532 nm laser beam. A ×20 magnification was used for the microscope. The camera was set to take 1 min videos operating at 30 frames per second. The NanoSight NTA 3.0 Software was then used to process the videos using the Stokes-Einstein equation for calculating the hydrodynamic size of each particle individually. The instrument’s specifications mention the theoretical range of detectable particle size from 10 to 2000 nm. However, the particle and liquid properties likely limited this range in our experiments on both ends of the operating particle size range. Two samples from each location were analyzed. Each sample was analyzed 10 times using a new subsample and these measurements were averaged. The NTA technique has gained popularity in the last few years (Domingos et al. 2009; Filipe et al. 2010) but has not been used or optimized for snow samples.

Generation of melted snow aerosol

The central novelty of this work is the aerosol-mediated snow analysis technique. The system for aerosolization of melted snow we built is diagramed in Fig. 1.
Fig. 1

Setup for collection of samples for electron microscopy through the aerosol phase. HEPA Filter high-efficiency particulate air filter, SMPS scanning mobility particle sizer, OPS optical particle sizer

Unprocessed, filtered, and/or dialyzed snow samples were aerosolized in the system. The setup consisted of three modules for (1) primary liquid droplet aerosol generation, (2) aerosol drying, and (3) dry aerosol stream analysis and collection. All the elements where aerosol transport occurred were positioned in a straight horizontal line to minimize aerosol particle losses due to impaction. Note, however, that the diagram shows these elements compactly. The elements of the setup were connected with electrically conductive silicone tubing (TSI, Inc., Shoreview, MN) to minimize electrostatic aerosol particle losses.

The melted snow samples were drawn into an autoclaved all-glass syringe that was installed into a syringe pump (GenieTouch™, Kent Scientific Corp., Torrington, CT, USA). The syringe pump quantitatively fed the melted snow into an autoclaved C-Flow 700d PFA Nebulizer (Savillex Corporation, Minnetonka, MN, USA). The nebulizer was connected to the syringe with autoclaved low-volatile grade, platinum-cured silicone tubing. The liquid feed rate to the nebulizer was set at 500 μL/min. Dry air from a cylinder was passed to the nebulizer at a pressure of 60 psi through a large high-efficiency particulate air (HEPA) filter, Whatman™ HEPA-Cap 150 (Little Chalfont, Buckinghamshire, UK). This resulted in the output air flow rate from the nebulizer of 0.9 SL/min (SL: liters at standard temperature and pressure, 273.15 K, 100 kPa). We checked the output airflow rate with a mass flow meter, model 4040 (TSI Inc.). The mass flow meter was equipped with another HEPA filter to prevent instrument contamination. Operation of the large HEPA filter in the nebulizer air supply line, at a flow rate substantially below nominal level, allowed obtaining clean air stream that was free of detectable particles. This step eliminated contamination with aerosol particles found in the air from the compressed air cylinder.

The aerosol stream containing aqueous droplets of melted snow particles was introduced into an in-house-built mixing element. The mixing element was a steel box with internal dimensions of 5 × 4 × 4 cm and three ports in its three facets. The mixing element too had a HEPA filter connected to it. This HEPA filter was open to the atmosphere, which allowed makeup air to be drawn into the mixing element for dilution of the sample stream to reach the total flow rate, drawn by the measurement instruments (1.75 L/min). The diluted aerosol then passed to an in-house-made double diffusion dryer. In a typical diffusion dryer, the aerosol stream travels through a tubular diffusion screen (wire mesh tube) surrounded by a desiccant (silica gel). Relative humidity (RH) inside the diffusion screen tubes is reduced due to diffusion of water molecules into the desiccant-filled space, leading to water evaporation from the aerosol droplets. The process of aerosol drying in the diffusion dryer removes water from the droplets generated by the atomizer, leaving any non-volatile particles dry and solutes precipitated. Each of the two diffusion dryers consisted of two transparent polycarbonate cylindrical silica gel-filled capsules, 16 in. long and 3 in. in diameter. The tubular diffusion screens inside the diffusion dryers were 0.5 in. in diameter.

Temperature and RH was checked in the output aerosol stream at the exit from the double diffusion dryer. For these measurements, we used an RH/temperature (RH/T) probe, model PROBE-SHT2X (Aginova, Inc., Freehold, NJ, USA) with a data transmitter/logger, and model iCelsius Wireless IPROBE-3000-0001 (Aginova, Inc.). During snow aerosolization, the RH and temperature of the aerosol stream just before electrostatic sampling were routinely measured at 2.2–2.7% and 22–23 °C, respectively.

Collection of samples for electron microscopy

Another novel aspect of this work is aerosol-mediated preparation of snow samples for electron microscopy. Several alternative methods of sample preparation, technically applicable to analysis of snow, such as drying a drop of liquid sample directly on an electron microscopy grid, have been described by Domingos et al. (2009). In our approach, the dry aerosol stream generated from processed or unprocessed melted snow was sampled directly into an electrostatic precipitator, ESPnano Model 100 (ESPnano, Spokane, WA, USA), at the exit from the double diffusion dryer at an aspiration inlet flow rate of 0.1 L/min. The excessive aerosol was released into the atmosphere without restriction. The electrostatic precipitator uses a corona discharge to simultaneously charge and collect aerosol particles onto media such as EM-grids. We operated the instrument at 5 kV and 100 s sampling duration with a 15 s purge before sampling. The aerosolization module was operated for at least 30 s before the start of each cycle of sampling with the electrostatic precipitator to ensure a stable particle size distribution throughout sampling.

For electrostatic sample collection, we used FCF200-Cu-EB grids (Formvar carbon film on 200-mesh copper extra thick option B grids, 25–50 nm Formvar and 3–4 nm carbon). This choice of grids was made to ensure a mechanically robust conductive surface for collection of fast-moving charged particles. The grids were loaded and unloaded from the electrostatic precipitator-compatible sample holder keys inside a Class II Biosafety Cabinet (NuAire, Inc., Plymouth, MN, USA) to prevent potential contamination with airborne particles found in ambient air.

Transmission electron microscopy

Analysis of the EM-grids with samples collected using electrostatic precipitation was conducted using a Tecnai G2F20 S/TEM with a field emission gun (FEI Comp., Hillsboro, OR, USA). The microscope was operated at 200 kV. The images were acquired with an UltraScan™ 1000 2 k × 2 k CCD Camera System, Model 895 (Gatan, Inc., Pleasanton, CA, USA). Energy-dispersive X-ray spectroscopy (EDS) was performed with a Genesis EDS Analysis System (EDAX, Inc., Mahwah, NJ, USA).

We also investigated the samples after staining them with 0.75% uranyl formate. The staining solution was prepared by dissolving 37.5 mg of uranyl formate powder (Electron Microscopy Sciences, Inc., Hatfield, PA, USA) in boiling hot ultrapure water. The test tube was wrapped in aluminum foil to reduce exposure to light. After complete dissolution, we added 150 μL of 5 M solution of sodium hydroxide. The mixture was then shaken in a test tube shaker for 5 min. The resulting solution was cooled to near-room temperature and filtered using the same type of Millex syringe filter with a 0.22 μm pore size that we used for melted snow filtration. The resulting staining solution was kept in a refrigerator at 4 °C wrapped in aluminum foil and used within 5 days. For safety, all the above procedures to prepare the staining solution were performed in a chemical fume hood to prevent any inhalation exposure to aerosols potentially generated during the handling and processing of uranyl formate, and all the liquid and solid waste was disposed of according to the local regulations. Alternatively, a 2% aqueous solution of uranyl acetate (SPI Supplies, Inc., West Chester, PA, USA) was used for staining the samples followed by air-drying. Imaging of these samples was done using a FEI Tecnai 12 Biotwin TEM microscope (FEI Comp.), equipped with a tungsten filament, at 120 kV.

Aerosol size distribution analysis

Aerosol particle size distributions in the output aerosol stream exiting from the double diffusion dryer were measured with a NanoScan™ scanning mobility particle sizer (SMPS), Model 3910 (TSI, Inc.), and an optical particle sizer (OPS), model 3330 (TSI, Inc.). The NanoScan™ SMPS is a particle electrical mobility-based instrument with a measurement size range of 10 nm to 420 nm and measurement time of 60 s for size distribution measurements. Its inlet sampling flow rate is 0.75 L/min. The OPS is a single-particle optical counting instrument with a measurement size range of 0.3–10 μm in 16 aerosol particle size channels. The OPS’s inlet sampling flow rate is 1 L/min.

Blank samples were obtained before each sampling session by operating the nebulizer without any liquid feed and then with ultrapure water that was either unfiltered or filtered through PVDF membrane filters with 0.22 μm pore size, the same type as those used for filtration of snow samples. Two samples from each location were analyzed. Each sample was nebulized and analyzed five times, and these measurements were averaged.

Results and discussion

The work encompassed the development of the aerosol-mediated snow analysis technique in parallel with application of alternative techniques: electron microscopy and NTA, as well as a complementary characterization of total organic carbon. The samples for electron microscopy were prepared by particle collection from aerosol phase, and this way of collecting particles, as we show below, has the advantage of preventing agglomeration of particles. Figure 2 depicts a flowchart showing various snow processing procedures and analytical techniques outlined in the “Materials and methods” section in the sequence they were performed.
Fig. 2

Snow processing procedures and analyses

Direct characterization of liquid melted snow samples

The first step of the work was to characterize snow samples by NTA—an established technique, yet only recently applied to analysis of snow samples. We snow samples from different locations and sampling campaigns in original state and processed by filtration and/or dialysis, as outlined in Fig. 2. Filtration was required to determine the potential or absence thereof of the larger particles to obscure detection of smaller particles and to determine the fraction of total organic carbon in the large particles. Dialysis was aimed at reducing concentration of dissolved substances (salts and organics) to determine their potential to affect size distribution measurements. Application of dialysis was a novel approach for snow samples. NTA found a higher abundance and larger sizes of particles in snow samples from Montreal, compared to samples from Alert and Barrow, as seen in Fig. 3. Expectedly, the unfiltered samples had a higher abundance of larger particles compared to the filtered samples. Most unprocessed samples’ hydrodynamic diameter distributions (NTA distributions) showed several modes (peaks); however, all distributions had the maximum peak between 100 and 200 nm. Filtration (0.22 μm pore size) dramatically reduced concentrations of particles with hydrodynamic diameter larger than a little over 200 nm. We note that hydrodynamic diameter, measured by NTA, includes the double electric layer. All filtered samples from all three locations (Alert, Barrow and Montreal) seemed to produce a generally similar shape of the hydrodynamic particle size distribution with the highest mode ranging from 100 to 200 nm. At this mode, the highest absolute particle concentration per 1 mL of melted snow of about (1.2 ± 0.2) × 106 cm−3 was observed in melted 2006 Montreal snow sample, followed by Alert, with about (8.2 ± 1) × 105 cm−3. The lowest concentration of 100–200 nm particles was measured in the Barrow sample: about (5 ± 0.5) × 105 cm−3. Barrow, however, had another peak at 300 nm, where particle concentration was (2 ± 1) × 105 cm−3.
Fig. 3

Hydrodynamic particle size distributions of snow samples from (a, b) Montreal 2006, (c, d) Alert, and (e, f) Barrow. All samples were analyzed both unprocessed and filtered using a syringe filter with pores of 0.22 μm in diameter. a, c, e Adapted with permission from Supporting Information for (Rangel-Alvarado et al. 2015). Error bars indicate ±1 standard error of the mean

Unprocessed samples from downtown Montreal (2006) showed a distribution with few larger particles, up to 900 nm, whereas particle size distributions of melted snow from Alert and Barrow showed particles up to around 400 nm. As the urban location of Montreal has multiple potential sources of large aerosol particles compared to the remote locations of Barrow and Alert lacking vegetation and with substantially lower anthropogenic activity, the observed higher abundance of larger particles was expected.

After filtration aimed at removing particles larger than about 200 nm, the distributions expectedly tended to become mono-modal, with a maximum peak still around 100 and 200 nm, but about 50% lower absolute concentration of these most abundant particles, with the exception of Barrow, where it increased. The trend was even more pronounced for the dialyzed Alert sample (Fig. S1) where dialysis also seemed to remove or lower the concentration of particles in the size range of 600 to 800 nm.

Soluble particles originally present in the snowpack, including soluble ice nuclei, dissolve when snow is melted. By nature, dialysis reduces soluble material, which may contribute to particle coatings, so reduction of soluble substances in melted snow was expected to lead to a reduction of both hydrodynamic and aerosol diameters of particles. Therefore, dialysis was indeed important to measure primary particle size in melted snow more accurately.

Soluble materials are of high importance due to their effects on ice nucleating properties of atmospheric particles. It has been reported that insoluble particles coated by soluble components, such as ammonium sulfate, may have altered ice nucleating properties and possibly other characteristics compared with the core insoluble particles themselves (Baustian et al. 2010; Wise et al. 2010).

To investigate the ionic content in the melted snow (soluble salts) and the effects of dialysis on it, we performed ion chromatography, which showed a substantial reduction of ionic content in all samples after performing dialysis (Table 1). This decrease was expected since most of the ions found in the unprocessed samples are soluble in water and can diffuse through the dialysis membrane effectively. However, we also observed that the concentrations of sulfate and calcium ions did not reduce as drastically as the concentrations of the other measured ions. For some samples, the concentration of calcium increased instead by a bit, and it was a consistent trend among those samples. The manufacturer (Spectrum Laboratories, Inc.) of the dialysis membrane we used indicated that the membrane could be a source of up to 0.5 mg/L of sulfur, which is comparable to the concentrations we observed in the original and dialyzed samples. However, the manufacturer did not find the membrane to leach any detectable concentration of calcium ions. Consistently, previous research showed that calcium ions bind to cellulose dialysis membranes (Reed 1973). A possible explanation for increasing calcium ion concentration may be that calcium ions are bound to colloidal matter in melted snow that does not pass through the dialysis membrane and may be released when ionic strength decreases in the process of dialysis. Further research, which is beyond the scope of this publication, is recommended to clarify the situation.
Table 1

Comparison of ion content in the original and dialyzed snow

Sample

Ions and their concentration, mg/L

Sulfate

Phosphate

Chloride

Nitrate

Sodium

Potassium

Magnesium

Calcium

Ammonium

Alert 1

0.39

<0.1

12.76

<0.1

7.12

0.23

0.87

1.01

<0.1

s

0.06

<0.1

0.26

<0.1

0.22

0.02

0.04

0.07

<0.1

Alert 1, dialyzed

0.65

<0.1

0.82

<0.1

0.38

0.15

0.07

1.36

<0.1

s

0.05

<0.1

0.06

<0.1

0.15

0.03

0.02

0.16

<0.1

Alert 2

0.43

<0.1

16.78

<0.1

7.13

0.26

0.98

0.74

<0.1

s

0.06

<0.1

0.14

<0.1

0.31

0.02

0.06

0.08

<0.1

Alert 2, dialyzed

0.36

<0.1

0.41

<0.1

0.26

0.05

0.04

0.81

<0.1

s

0.04

<0.1

0.08

<0.1

0.15

0.02

0.01

0.11

<0.1

Barrow 1

0.68

<0.1

13.59

<0.1

5.24

0.21

0.66

1.06

<0.1

s

0.02

<0.1

11.05

<0.1

4.13

0.03

0.50

0.35

<0.1

Barrow 1, dialyzed

0.55

<0.1

0.68

<0.1

0.30

0.05

0.08

1.08

<0.1

s

0.02

<0.1

0.03

<0.1

0.13

0.00

0.00

0.01

<0.1

Barrow 2

0.53

<0.1

7.91

<0.1

3.70

0.19

0.51

0.58

<0.1

s

0.04

<0.1

0.07

<0.1

0.14

0.02

0.02

0.05

<0.1

Barrow 2, dialyzed

0.52

n.d.

3.45

n.d.

0.43

2.38

n.d.

1.19

n.d.

s

0.01

n.d.

0.03

n.d.

0.13

0.03

n.d.

0.09

n.d.

Montreal 2006

0.82

n.d.

0.29

n.d.

0.48

0.09

n.d.

0.86

n.d.

s

0.02

n.d.

0.01

n.d.

0.10

0.02

n.d.

0.04

n.d.

Montreal 2006, dialyzed

0.46

n.d.

0.53

n.d.

0.27

n.d.

n.d.

0.71

n.d.

s

0.02

n.d.

0.57

n.d.

0.14

n.d.

n.d.

0.10

n.d.

Montreal 2014

1.47

n.d.

11.40

n.d.

4.74

n.d.

0.59

5.16

n.d.

s

0.01

n.d.

0.06

n.d.

0.16

n.d.

0.02

0.06

n.d.

Montreal 2014, dialyzed

0.20

n.d.

0.07

n.d.

0.28

n.d.

n.d.

0.67

n.d.

s

0.00

n.d.

0.01

n.d.

0.14

n.d.

n.d.

0.13

n.d.

s standard deviation

We did not observe a clear correlation between the total concentrations of different ions measured by IC and concentrations of aerosol particles, which mean that the influence of particles formed from dissolved substances had no significant influence on the aerosol size distributions, similar to observations by other researchers (Lee et al. 2011). The concentrations of various ions, as measured by IC, even in undialyzed samples, were also below the concentrations used for generation of salt crystal aerosols from saline solutions with particle size in the region of interest for melted snow (TSI, Inc. 2005). We note that generation of salt crystal aerosols from saline solutions is a well-established technique in aerosol science (Borrego and Brebbia 2007). Moreover, even undialyzed samples had salt concentrations (almost exclusively Na+ and Cl, as other ions were generally <1 mg/L) lower or on the borderline with those required to produce any interfering salt aerosol. For example, it can be seen that higher concentrations than what we had in dialyzed and even undialyzed (Table 1) samples of salts are needed to produce dry aerosol particles in the size range of interest: 120 mg/L of NaCl for 50 nm mode diameter, 12 mg/L for 35 nm mode diameter, and 1 mg/L for ∼30 nm mode diameter. In the latter case, no distinct mode is observed when salt concentration is only 1 mg/L. In our distributions, we observed a more distinct mode at 15 nm for Alert and secondary modes at this size for Montreal and Barrow. Moreover, we had the same mode at 15 nm for both undialyzed and dialyzed samples. The effects of filtration on total organic carbon concentration differed substantially between snow samples. The concentration of organic carbon in the most recent Montreal snow sample (2014) changed little after filtering out particles larger than 200 nm (Table 2).
Table 2

Total organic carbon in snow samples. s is standard deviation

Sample

Average, mg/L

s

%CV

Arctic snow sample (Alert, Nunavut)

0.723

0.047

6.530

Filtered Arctic snow sample (Alert, Nunavut)

0.496

0.040

8.184

Montreal snow sample (2014)

1.484

0.018

1.239

Filtered Montreal snow sample (2014)

1.410

0.090

6.367

Montreal snow sample (2006)

2.803

0.045

1.589

Filtered snow sample (2006)

0.980

0.049

5.051

Therefore, most particulate organic matter in it was in the size fraction <200 nm and in the molecularly dissolved form that passed through the filter. On the contrary, for the 2006 Montreal snow sample, removal of particles with diameters greater than 200 nm by filtration led to an almost threefold drop of total organic carbon (TOC), meaning that most particulate organic matter by mass in the 2006 Montreal snow sample was in the size fraction larger than 200 nm. The substantially higher TOC in unfiltered 2006 Montreal sample compared to the 2014 Montreal sample points to the presence of an initially higher abundance of large (>200 nm) organic matter-containing particles in the 2006 Montreal snow sample, relative to the combined contribution of small particles (<200 nm) and dissolved organic carbon. The sources of these particles are both anthropogenic and biogenic. Approximately 20 to 50% of the total organic fine aerosol mass at the continental mid-latitudes originate from fossil fuel burning, domestic and biomass burning, and naturally from viruses, bacteria, fungal spores, and plant debris (Kanakidou et al. 2005). We note that microorganisms are mostly larger than the filter pore size we used (Kokhanovsky 2008; Talaro 2008), so their removal is expected to also contribute to the observed lowering of TOC after filtration. In the Montreal snow samples, anthropogenic particles likely represent a larger percentage of the total aerosol, compared to snow at Alert due to absence of considerable immediate sources of organic matter emission in Alert and presence of considerable emission sources in the Montreal area. Indeed, we observed the lowest TOC among all analyzed samples in the samples from Alert, where it was approximately half the level in the lowest TOC Montreal sample from 2014. With respect to the content of dissolved organic matter, small (<200 nm) and large (>200 nm) organic particles: about a third of TOC in the sample from Alert constituted large particles corresponding to TOC removed by filtration, and the remaining two thirds of the TOC were small particles and dissolved organic matter.

When we examined particulate matter, collected by electrostatic precipitation from the aerosolized melted snow, by TEM, we observed a high on-grid concentration of particles of various electron contrast and crystalline, as well as amorphous structures (Fig. 4a–f). However, dialyzed versions of the same samples were dominated by smaller particles (20–100 nm) with low electron contrast. We note a very low occurrence of agglomerates, not excluding a possible presence of electron-transparent agglomerates that transmission electron microscopy (TEM) cannot detect. We then “washed” a portion of the samples directly on EM-grids to remove soluble matter and stained them, along with original unwashed samples, with uranyl acetate and uranyl formate to view organic particles as well. Uranyl formate and uranyl acetate are negative stains—they leave organic particles unstained and near-transparent to the electron beam while staining the background. Stained samples presented a high concentration of particles showing as light dots at the dark background of other high-electron-contrast particles and the stained substrate, as well as particles with low electron contrast (Fig. 4g–i).
Fig. 4

Representative high- and mid-resolution TEM images of aerosolized particles. ac With prior dialysis and di without any prior processing. af Stained with uranyl formate and gi stained with uranyl acetate

Most of the high-electron-contrast particles exhibited what looked like a coating surrounding the particles, similar to recent observations of such features in sea spray aerosol particles (Prather 2015). This coating had lower electron contrast than the core particles. The electron microscopy images of dialyzed samples did not show residue from precipitation of dissolved matter and formation of salt crystals, supporting the abovementioned evidence for effectiveness of dialysis in reducing effects of dissolved substances. TEM of any blank samples (not shown) did not reveal electron-contrast particles beyond very low contamination with metallic copper particles in the very first background sample, where these particles likely originated from the electrostatic precipitator itself, brand new and used for the first time since acquisition. This contamination was not observed in subsequent samples under our operating conditions.

The particles visible by TEM in samples from all three locations were found by EDS to contain carbon and silicon in detectable quantities (Figs. S2–S4), silicon likely originating from mineral dust and carbon from either organic or carbonaceous matter. Only the sample from Montreal contained sodium, potassium, calcium, magnesium, and chlorine within electron contrast particles, and these elements were probably detected due to their elevated concentration in the urban snow where salt is applied to deice roadways and sidewalks and could be carried by wind as aerosol to areas of snow sampling. The sample from Alert contained the lowest number of detected elements: besides carbon, oxygen, and silicon, EDS found tellurium in it. The sample from Barrow, besides carbon and silicon, contained chromium, nickel, and iron—likely from mineral dust. It must be noted that the copper detected by EDS is in the region of spectra that indicates it came from the TEM-grids, made out of copper, and it was feasible to only investigate select particles by EDS due to the high cost of the technique, so the presented data may not be exhaustive with respect to all particles present in the samples. Ions of iron, chromium, and nickel that were present in the EDS spectra (Figs. S2–S4) were likely mostly in an insoluble form within colloidal particles. The EDS data are complementary to the IC results, which analyzed ions in the entire volume of melted snow samples.

Aerosol-mediated characterization of melted snow samples

The aerosol measurements indicated a higher abundance of particles of all sizes in aerosol size distributions for the Montreal samples, closely followed by Barrow, as seen in Fig. 5 and Table 3. We saw much fewer large particles (0.3–10 μm) in the dialyzed filtered samples, which indicates that filtration effectively removes >0.3 μm particles, but the effect may be clearly seen only when the sample is also dialyzed, again pointing to the necessity of dialysis in terms of increased definition of the aerosol size distributions. Dialysis also reduced the concentration of larger particles (Fig. 5) and somewhat reduced small particles as well (Fig. 6).
Fig. 5

Aerosol particle size distributions of the aerosolized snow samples measured using the NanoScan scanning mobility particle sizer (SMPS) and the optical particle sizer (OPS)

Table 3

Comparison of the mode diameters and concentrations for different locations of snow sampling. The concentrations are at corresponding primary and secondary mode particle diameter and expressed as dN/dlogDp, where dN is the number-based concentration of particles in the particle size range delimited by the upper and lower particle size boundaties of the corresponding measurement bin. Dp is the difference between these upper and lower instument measurement bin boundaries

Sample

Primary mode diameter (nm)

Concentration × 105, cm−3

Secondary mode diameter (nm)

Concentration × 105, cm−3

Alert snow sample

10

20

30

15

Dialyzed alert snow sample

10

7

30

2

Barrow snow sample

30

40

10

20

Dialyzed Barrow snow sample

10

2

30

1

Montreal snow sample

30

40

10

10

Dialyzed Montreal snow sample

10

2

30

1

Fig. 6

High-resolution aerosol nanoparticle size distributions of the aerosolized snow samples measured with the NanoScan scanning mobility particle sizer (SMPS)

However, after dialysis, despite substantial differences in concentrations of large particles between filtered and unfiltered samples, no change of mode diameters occurred. We note that some particles may have been lost to the walls of the dialysis tubing. We did not observe as dramatic a reduction of aerosol particles after filtration of the undialyzed samples. It was expected that filtration would remove large particles. We, therefore, conclude that soluble substances, originally present in undialyzed melted snow, must have formed a substantial fraction of large aerosol particles, and dialysis remediated the masking problem of particles derived from soluble substances revealing a size distribution of larger aerosol particles, including larger ice nuclei, more representative of that in the melted snow samples. Therefore, desalination by dialysis of melted snow samples was found to be an important additional experimental step, which allowed improving definition of aerosol size distributions of snow-borne particles by reducing interference from salt particles and other soluble matter that may otherwise condense after aerosol drying.

When the nebulizer was operated without any liquid feed (blank sample group 1), the NanoScan™ and the OPS registered only occasional single particles, indicating a very clean background. The ultrapure water aerosol size distributions (blank sample group 2) were substantially below the size distributions of aerosolized melted snow samples and are, therefore, not shown in Figs. 5 and 6.

Verification of the aerosol-mediated characterization technique

To rule out any significant effect of particles formed from soluble substances that may still remain after dialysis in the melted snow samples, including any residual salts, which may precipitate or crystallize during drying of aerosol, we prepared and analyzed a number of 2-, 3-, and 4-fold dilutions of select filtered and unfiltered dialyzed samples by mixing the melted snow samples with ultrapure water. This standard approach was successfully used before to rule out effects of soluble substances on aerosol size distribution (Lee et al. 2011). The aerosol size distributions of these original and diluted samples did not reveal any changes of the aerosol size distributions for any samples (Fig. S5). At the same time, aerosol concentrations in different aerosol size channels changed proportionally to the degree of dilution of the original melted snow samples, which was expected and confirms that these particles originated from insoluble particles, originally present in the melted snow. We, therefore, conclude that the concentration of particles formed due to precipitation and crystallization of any soluble matter in dialyzed snow samples was insignificant and did not distort the aerosol size distributions.

As an additional test of the robustness of the aerosolization technique coupled with dialysis, we tested diluted standard TiO2 suspension (rutile, 20%wt in water, 30–50 nm, US Research Nanomaterials, Inc., Houston, TX, USA) with a soluble salt (NaCl) and a soluble organic material (oxalic acid), both reagents from Sigma Aldrich, Inc. (St. Louis, MI, USA). We used two different concentrations of NaCl. The first was 28 mg/L of NaCl—equivalent to 17 mg/L of chloride, corresponding to the highest concentration of chloride measured in undialyzed snow samples. The second was 1 mg/L of NaCl (typical concentration after dialysis of snow samples). We chose the concentration of oxalic acid based on the measurements of this organic substance in cloud water at Mount Lu and Mount Heng as reported by Sun et al. (2016). Note that the concentration in Sun et al. (2016) is reported in microequivalents per liter. Therefore, the concentration of 630 μg/L we used here corresponds to 10 μeq/L of oxalic acid (H2C2O4·2H2O) as reported in Sun et al. (2016). We note that dialysis is expected to greatly reduce concentrations of soluble organic substances too (<5–8 kDa), including oxalic acid, so the original 630 μg/L is a high estimate. As seen in Fig. 7, soluble materials at low concentration, typical of post-dialysis samples (1 mg/L NaCl and 630 μg/L oxalic acid) do not have an effect on aerosol particle concentrations or the particle size distribution of TiO2, whereas soluble material at high concentration (as found in melted snow prior to dialysis) do modify the aerosol concentration and particle size distribution. It is clear that dialysis is an effective and crucial step when using the aerosol-mediated technique we developed for analysis of samples with high concentrations of soluble materials.
Fig. 7

Aerosol particle size distributions measured using the NanoScan scanning mobility particle sizer (SMPS) and the optical particle sizer (OPS) of the aerosolized TiO2 colloid solutions with or without added solutes: oxalic acid and different concentrations of NaCl. Oxalic acid concentration is 630 μg/L

The NTA’s ability to detect particles by the light they scatter sharply reduces below a certain size, which, depending on optical properties of particles, can be between 10 and 50 nm or sometimes even larger, according to the manufacturer of the instrument. Therefore, the aerosolization technique, involving real-time aerosol measurements by the SMPS and OPS systems, has a superior particle size measurement range—down to 10 nm with the instrument we used. The NTA size distributions predictably contrast in their not showing a high number of small nanoparticles. The aerosolization-based SMPS measurement showed a high abundance (∼106 cm−3) of aerosol particles down to the smallest measurable size of 10 nm for all snow samples (Fig. 6).

In summary, we developed a new technique, at the core of which is aerosolization of melted snow with subsequent drying of aqueous droplets in airborne state. The melted snow is processed before aerosolization by dialysis and optionally by filtration, which allows removing larger particles (>200 or >100 μm, depending on the filter used) and soluble substances, including salts, such as chlorides, sulfates, etc. Thus, when processed melted snow is aerosolized, and the minute droplets are dried, an aerosol of core snow-borne (nano)particles is produced. These particles are then analyzed by real-time aerosol analyzers to determine particle size distributions. The particles are also collected on TEM-grids using electrostatic precipitation for electron microscopy and EDS to determine their elemental composition. In this study, the use of this new technique was complemented with analysis by NTA, providing a more comprehensive set of information about snow-borne (nano)particles. Compared to the traditionally used direct liquid-to-grid sample transfer, the aerosol-mediated approach allows amplification of deposited particle density on EM-grids while minimizing particle agglomeration. Another advantage of the new approach is the higher particle size resolution and a comparatively more extensive measurement size range of provided by the aerosol analyzers compared to traditional techniques for particle measurement in liquid phase, such as nanoparticle tracking analysis.

Concluding remarks and future work

This study presents a novel technique based on aerosolization of processed melted snow. The processing by dialysis of melted snow samples before aerosolization is an important step to reduce interference effects of soluble substances on the measured size distributions. Dialyzing melted snow that we suggest here could also become an important method in research of effects of soluble substances, such as biogenic chemicals and anthropogenic pollutants, on ice nucleating properties of atmospheric aerosols.

The developed technique combines sizing and counting of individual dried aerosol particles with their collection by electrostatic precipitation. Consequently, acquisition of particle size distributions of snow-borne particles is complemented with their simultaneous characterization by electron microscopy and energy-dispersive X-ray spectroscopy. The samples prepared through aerosolization and subsequent collection of dispersed particles are suitable for analysis by other techniques in future research.

In this study, a variety of snow samples were characterized, providing new information about the similarities and differences in composition and physicochemical features of particulate matter in snow while demonstrating broad applicability and usefulness of the new aerosolization technique.

The developed technique has a high sensitivity and minimizes particle aggregation. The results demonstrate the added value of the new technique in its ability to measure a high-resolution snow-borne particle size distribution, including between 10 and 100 nm, where other techniques are limited. Future research should focus on segregating size distributions of different categories of ice nucleating particles found in snow from non-ice nuclei.

Notes

Acknowledgements

We thank Mr. Jean-Philippe Guay for building the diffusion dryers and the mixing elements, Ms. Katherine Velghe for conducting TOC, and Ms. Monique Riendeau for conducting IC analyses of the melted snow. We would like to express special thanks to TSI, Inc. and personally to Ms. Sherrie Elzie and Ms. Sarah Sakamoto for providing the NanoScan™ and the OPS instruments. We also kindly thank the anonymous reviewers for their valuable feedback that helped improve the final manuscript. The study is jointly funded by Natural Science and Engineering Research Council of Canada, Environment Canada, Canadian Foundation for Innovation, and FRQNT. Dr. Yevgen Nazarenko is supported by Fonds de recherche du Québec—Nature et technologies. The views expressed in the manuscript are solely of the authors and do not necessarily reflect those of the funding agencies.

Supplementary material

11356_2016_8199_MOESM1_ESM.pdf (448 kb)
ESM 1(PDF 448 kb)

References

  1. Archer D (2011) Global warming: understanding the forecast. Wiley, Hoboken, pp 203Google Scholar
  2. Archuleta CM, DeMott P, Kreidenweis S (2005) Ice nucleation by surrogates for atmospheric mineral dust and mineral dust/sulfate particles at cirrus temperatures. Atmos Chem Phys 5:2617–2634CrossRefGoogle Scholar
  3. Ariya PA, Sun J, Eltouny N, Hudson E, Hayes C, Kos G (2009) Physical and chemical characterization of bioaerosols—implications for nucleation processes. Int Rev Phys Chem 28:1–32CrossRefGoogle Scholar
  4. Ariya PA, Domine F, Kos G, Amyot M, Côté V, Vali H, Lauzier T, Kuhs WF, Techmer K, Heinrichs T, Mortazavi R (2011) Snow—a photobiochemical exchange platform for volatile and semi-volatile organic compounds with the atmosphere. Environ Chem 8:62–73CrossRefGoogle Scholar
  5. Ariya PA, Kos G, Mortazavi R, Hudson E, Kanthasamy V, Eltouny N, Sun J, Wilde C (2014) Bio-organic materials in the atmosphere and snow: measurement and characterization, atmospheric and aerosol chemistry. Springer, Berlin Heidelberg, pp 145–199Google Scholar
  6. Barkan J, Alpert P, Kutiel H, Kishcha P (2005) Synoptics of dust transportation days from Africa toward Italy and central Europe. J Geophys Res 110:1–14CrossRefGoogle Scholar
  7. Barrie LA (1980) The fate of particulate emissions from an isolated power plant in the oil sands area of western Canada. Ann N Y Acad Sci 338:434–452CrossRefGoogle Scholar
  8. Baustian KJ, Wise ME, Tolbert MA (2010) Depositional ice nucleation on solid ammonium sulfate and glutaric acid particles. Atmos Chem Phys 10:2307–2317CrossRefGoogle Scholar
  9. Bigg EK, Leck C (2001) Cloud-active particles over the central Arctic Ocean. Journal of Geophysical Research: Atmospheres 106:32155–32166CrossRefGoogle Scholar
  10. Borrego C, Brebbia CA (2007) Air pollution XV. WIT transactions on ecology and the environment. WIT Press, Billerica, pp 624Google Scholar
  11. Domingos RF, Baalousha MA, Ju-Nam Y, Reid MM, Tufenkji N, Lead JR, Leppard GG, Wilkinson KJ (2009) Characterizing manufactured nanoparticles in the environment: multimethod determination of particle sizes. Environ Sci Technol 43:7277–7284CrossRefGoogle Scholar
  12. Egerton RF, Li P, Malac M (2004) Radiation damage in the TEM and SEM. Micron 35:399–409CrossRefGoogle Scholar
  13. Filipe V, Hawe A, Jiskoot W (2010) Critical evaluation of nanoparticle tracking analysis (NTA) by NanoSight for the measurement of nanoparticles and protein aggregates. Pharm Res 27:796–810CrossRefGoogle Scholar
  14. Fröhlich-Nowoisky J, Ruzene Nespoli C, Pickersgill DA, Galand PE, Müller-Germann I, Nunes T, Gomes Cardoso J, Marta Almeida S, Pio C, Andreae MO, Conrad R, Pöschl U, Després VR (2014) Diversity and seasonal dynamics of airborne archaea. Biogeosciences Discuss 11:6067–6079CrossRefGoogle Scholar
  15. Georgakopoulos DG, Despres V, Fröhlich-Nowoisky J, Psenner R, Ariya PA, Posfai M, Ahern HE, Moffett BF, Hill TCJ (2009) Microbiology and atmospheric processes: biological, physical and chemical characterization of aerosol particles. Biogeosciences 6:721–737CrossRefGoogle Scholar
  16. Jacobson MZ (2005) Fundamentals of atmospheric modeling. Cambridge University Press, New York, pp 813Google Scholar
  17. Kanakidou M, Seinfeld J, Pandis S, Barnes I, Dentener F, Facchini M, Dingenen RV, Ervens B, Nenes A, Nielsen C (2005) Organic aerosol and global climate modelling: a review. Atmos Chem Phys 5:1053–1123CrossRefGoogle Scholar
  18. Kärcher B (2012) Atmospheric ice formation processes, atmospheric physics. Research topics in aerospace. Springer, Berlin Heidelberg, pp. 151–167Google Scholar
  19. Khvorostyanov VI, Curry JA (2014) Thermodynamics, kinetics and microphysics of clouds. Cambridge University Press, New York, pp 800Google Scholar
  20. Kokhanovsky AA (2008) Light scattering reviews 3: light scattering and reflection. Environmental Sciences. Springer Berlin Heidelberg, Chichester, UK, 408 ppGoogle Scholar
  21. Koop T, Luo B, Tsias A, Peter T (2000) Water activity as the determinant for homogeneous ice nucleation in aqueous solutions. Nature 406:611–614Google Scholar
  22. Kos G, Ariya PA (2006a) Determination of volatile organic compounds in snow and air in alert, Nunavut. AGU Fall Meeting 2006, Abstract #C41E-08Google Scholar
  23. Kos G, Ariya PA (2006b) Determination of a wide range of volatile and semivolatile organic compounds in snow by use of solid-phase micro-extraction (SPME). Anal Bioanal Chem 385:57–66CrossRefGoogle Scholar
  24. Kos G, Ariya PA (2010) Volatile organic compounds in snow in the Quebec-Windsor Corridor. Journal of Geophysical Research: Atmospheres 115:2156–2202CrossRefGoogle Scholar
  25. Kuoppamäki K, Setälä H, Rantalainen A-L, Kotze DJ (2014) Urban snow indicates pollution originating from road traffic. Environ Pollut 195:56–63CrossRefGoogle Scholar
  26. Lee I-S, Kim HJ, Lee DH, Hwang GB, Jung JH, Lee M, Lim J, Lee BU (2011) Aerosol particle size distribution and genetic characteristics of aerosolized influenza A H1N1 virus vaccine particles. Aerosol and Air Quality Resarch 11:230–237Google Scholar
  27. Mortazavi R, Kos G, Ariya PA (2006) A comparison study of volatile organic compound species and concentrations in snow samples from rural sites in south-western Quebec and at Alert, Nunavut. AGU Fall Meeting 2006, Abstract #A11A-0833Google Scholar
  28. Mortazavi R, Hayes CT, Ariya PA (2008) Ice nucleation activity of bacteria isolated from snow compared with organic and inorganic substrates. Environ Chem 5:373–381CrossRefGoogle Scholar
  29. Mortazavi R, Attiya S, Ariya PA (2014a) Microbial analysis of arctic snow and frost flowers: what next generation sequencing method can reveal. AGU Fall Meeting 2014, Abstract #C31C-0316Google Scholar
  30. Mortazavi R, Attiya S, Ariya PA (2014b) A next generation sequencing of Arctic bacteria in snow and frost flowers: identification, abundance and freezing nucleation. Atmos Chem Phys Discuss 14:32093–32131CrossRefGoogle Scholar
  31. Mortazavi R, Attiya S, Ariya PA (2015) Arctic microbial and next-generation sequencing approach for bacteria in snow and frost flowers: selected identification, abundance and freezing nucleation. Atmos Chem Phys 15:6183–6204CrossRefGoogle Scholar
  32. Murray B, O’Sullivan D, Atkinson J, Webb M (2012) Ice nucleation by particles immersed in supercooled cloud droplets. Chem Soc Rev 41:6519–6554CrossRefGoogle Scholar
  33. Nazarenko Y, Kurien U, Nepotchatykh O, Rangel-Alvarado RB, Ariya PA (2015) Role of snow and cold environment in the fate and effects of nanoparticles and select organic pollutants from gasoline engine exhaust. Environ Sci Process Impacts 18:190–199CrossRefGoogle Scholar
  34. Nemirovskaya IA, Kravchishina MD (2015) Biogeochemical features of the distribution of organic compounds and particulate matter in the snow-ice cover in East Antarctica. Geochem Int 53:430–440CrossRefGoogle Scholar
  35. Poulain AJ, Garcia E, Amyot M, Campbell PG, Ariya PA (2007) Mercury distribution, partitioning and speciation in coastal vs. inland High Arctic snow. Geochim Cosmochim Acta 71:3419–3431CrossRefGoogle Scholar
  36. Prather KA (2015) Influence of sea spray aerosols on cloud and climate. 250th American Chemical Society National Meeting & Exposition, Boston, p 155Google Scholar
  37. Radke LF, Hobbs PV, Eltgroth MW (1980) Scavenging of aerosol particles by precipitation. J Appl Meteorol 19:715–722CrossRefGoogle Scholar
  38. Rangel-Alvarado RB, Nazarenko Y, Ariya PA (2015) Snow-borne nanosized particles: abundance, distribution, composition, and significance in ice nucleation processes. J Geophys Res Atmos 120:11760–11774Google Scholar
  39. Reed K (1973) Binding of calcium by cellulose membranes and Sephadex. Biochem Biophys Res Commun 50:1136–1142CrossRefGoogle Scholar
  40. Rosenfeld D (2000) Suppression of rain and snow by urban and industrial air pollution. Science 287:1793–1796CrossRefGoogle Scholar
  41. Schnell R, Vali G (1976) Biogenic ice nuclei: part I. Terrestrial and marine sources. J Atmos Sci 33:1554–1564CrossRefGoogle Scholar
  42. Skeie RB, Berntsen T, Myhre G, Pedersen CA, Ström J, Gerland S, Ogren JA (2011) Black carbon in the atmosphere and snow, from pre-industrial times until present. Atmos Chem Phys 11:6809–6836CrossRefGoogle Scholar
  43. Sun X, Wang Y, Li H, Yang X, Sun L, Wang X, Wang T, Wang W (2016) Organic acids in cloud water and rainwater at a mountain site in acid rain areas of South China. Environ Sci Pollut Res Int 23:9529–9539CrossRefGoogle Scholar
  44. Talaro KP (2008) Foundations in microbiology, basic principles. McGraw-Hill, New York, pp 534Google Scholar
  45. The Weather Channel, LLC (2014) Weather.com. The Weather Company, LLC weather.com®
  46. TSI, Inc (2005) Model 3076 constant output atomizer instruction manual. TSI, Inc., Shoreview, pp 63Google Scholar
  47. U.S.EPA (1992) Specifications and guidance for contaminant-free sample containers. Office of Solid Waste and Emergency Response, Washington, DCGoogle Scholar
  48. Vali G, Christensen M, Fresh R, Galyan E, Maki L, Schnell R (1976) Biogenic ice nuclei. Part II: Bacterial sources. Journal of the Atmospheric Sciences 33:1565–1570CrossRefGoogle Scholar
  49. Williamson BJ, Udachin V, Purvis OW, Spiro B, Cressey G, Jones GC (2004) Characterisation of airborne particulate pollution in the Cu smelter and former mining town of Karabash, south Ural Mountains of Russia. Environ Monit Assess 98:235–259CrossRefGoogle Scholar
  50. Wise ME, Baustian KJ, Tolbert MA (2010) Internally mixed sulfate and organic particles as potential ice nuclei in the tropical tropopause region. Proc Natl Acad Sci USA 107:6693–6698CrossRefGoogle Scholar
  51. Xuan J, Sokolik IN, Hao J, Guo F, Mao H, Yang G (2004) Identification and characterization of sources of atmospheric mineral dust in East Asia. Atmos Environ 38:6239–6252CrossRefGoogle Scholar
  52. Yakobi-Hancock JD, Ladino LA, Abbatt JP (2013) Review of recent developments and shortcomings in the characterization of potential atmospheric ice nuclei: focus on the tropics. Revista de Ciencias 17:15–34Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2016

Authors and Affiliations

  • Yevgen Nazarenko
    • 1
  • Rodrigo B. Rangel-Alvarado
    • 2
  • Gregor Kos
    • 1
  • Uday Kurien
    • 1
  • Parisa A. Ariya
    • 1
    • 2
  1. 1.Department of Atmospheric and Oceanic SciencesMcGill UniversityMontrealCanada
  2. 2.Department of ChemistryMcGill UniversityMontrealCanada

Personalised recommendations