Atmospheric Nanoparticles: Formation and Physicochemical Properties Part 1 (Nanotechnology)

INTRODUCTION

Atmospheric nanoparticles, defined as particles with spherical equivalent diameters smaller than 50 nm, are either directly emitted from combustion sources, or are formed in the atmosphere by a process called nucleation. They then quickly grow by the condensation of gas monomers or clusters, or by coagulation with other particles, to become a critical participant in a number of important atmospheric processes such as heterogeneous chemistry, cloud formation, precipitation, and the scatter and absorption of solar radiation. An understanding of nanoparticles has increased dramatically in recent years because of significant advances in instrumentation to detect, size, and determine their chemical composition. Recent observations now show us that atmospheric nano-particles are ubiquitous: Measurements carried out in city centers, isolated islands and forests, and the remote troposphere have never failed to encounter periods characterized by concentrations of up to 106 nanoparticles/cm3. Theoretical advances in nucleation and growth, at present, have not made the same sort of progress as observation, but this is an area of rapid progress.

The goal of this article is to provide an overview of the role that nanoparticles play in the atmosphere, the observational and theoretical tools currently being employed in their study, and recent observations that continue to direct progress in understanding their formation and fate. Although much is known about these transient particles, researchers appear to be at the cusp of huge scientific advances that will make it possible to predict their formation and behavior in the atmosphere.

BACKGROUND

Excellent texts are available on the subject of atmospheric particulate matter,[1,2] which is commonly referred to as atmospheric aerosol. These texts are also adequate references on atmospheric nanoparticles; however, recently, a review chapter[3] and a special journal issue[4] have focused specifically on this area. The goal of this section is to provide a broad overview of the distribution, sources, and lifetimes of atmospheric aerosol. This is followed by a review of relevant atmospheric process involving nanoparticles.

Atmospheric Particle Size Distributions

Close to the Earth’s surface, continental air typically contains from 103 to 105 particles/cm3. These particles have spherical equivalent diameters dp ranging from 1 nm to 100 mm. Plots of particle number concentration, surface area, and volume vs. dp, such as those shown in Fig. 1 for typical rural settings, usually shows that these particles are distributed over three or more modes. Most particles are smaller than 100 nm in diameter (Fig. 1a), in a mode referred to as the Aitken mode. The term ultrafine aerosol is often used for those with diameters smaller than 100 nm. The other two predominant modes are the accumulation mode (0.1 <dp<2.5 mm) and the coarse mode (dp>2.5 mm). Within the Aitken mode, for dp< 50 nm, many properties of the condensed phase, most notably vapor pressure, become quite different from those of the bulk phase: Thus particles smaller than 50 nm have come to be known as nanoparticles. The implications of the distributions in Fig. 1 on atmospheric processes can be generalized as follows: For phenomena that depend on the particle number concentration, it is the Aitken mode that contributes the most; for those processes that are surface area dependent, the accumulation mode is most influential; for those that depend on total aerosol volume or mass, the coarse mode dominates.

Atmospheric Particle Sources and Sinks

The different modes in aerosol size distribution reflect differences in the atmospheric sources, transformations, and sinks (Fig. 2). Particles in the coarse mode are generated by mechanical processes such as wind and friction. The accumulation mode contains particles that have grown in the atmosphere by condensation of gases, or coagulation with other particles. The Aitken mode contains particles that are generated chemically, either directly emitted from a source (primary aerosol) or are created or grown from gas phase reactions in the atmosphere (secondary aerosol). The dominant loss mechanisms for Aitken mode particles are coagulation and condensational growth. Particles in the accumulation mode leave the atmosphere by adhering onto wetted surfaces, most commonly rain and cloud drops, through a process called wet deposition. Coarse mode particles eventually settle to the ground in a process called dry deposition. Fig. 3 shows the typical timescales for these removal processes. Dry deposition, coagulation, and condensational growth are relatively rapid processes, whereas wet deposition is the least efficient means of particle removal. The result is that particles in the accumulation mode are the longest-lived particles, having lifetimes that may span several days. Therefore these particles are commonly spatially homogeneous, and number distributions in this size range do not change rapidly with time. In contrast, Aitken mode particles have lifetimes of a few minutes to an hour. Therefore these particles commonly exist close to their source (e.g., near busy freeways) and their number distributions usually change rapidly with time.

Fig. 1 Particle (a) number, (b) surface area, and (c) volume distributions for typical rural conditions, generated using the parameterization of Jaenicke.[6]

Nanoparticles in the Atmosphere

Nanoparticles, as a subset of the Aitken mode, are generated chemically either as primary emissions or through gas phase reactions in the atmosphere. The latter process is, in fact, nucleation and will be the focus of much of the discussion that follows. It is also true that nanoparticles are some of the most transient in the atmosphere, with a corresponding heterogeneous spatial distribution that depends on source distribution and meteorological conditions. Additionally nanoparticles may frequently dominate an aerosol number distribution, but usually not distributions of surface area and almost never of volume (Fig. 1). In view of these observations, it is reasonable to ask if nanoparticles play an important role in atmospheric processes, outside of their crucial role as the primary source of new particles. A question of equal importance, although outside the scope of this article, is whether atmospheric nanoparticles play a deleterious role in human health through respiration. The answer to this question is a qualitative ”yes,” although the mechanism and magnitude of this effect are far from clear.[3,8,9]

The contribution of nanoparticles to scattering and absorption of solar radiation, which controls processes such as visibility reduction and climate, is understood to be negligible.[2] This is because of the low efficiency with which nanoparticles interact with solar radiation.[10] Perhaps the most important contribution of nanoparticles on climate is to influence the optical properties, abundance, and persistence of clouds.[11] An abundance of new particles could increase the number of cloud condensation nuclei (CCN), which are water-soluble aerosols on which water vapor condenses to form cloud droplets. The overall result of this so-called ”indirect effect of aerosols on climate” is an issue of active research and debate, but is generally thought to be a reduction in incoming solar radiation.[11]

At the time of this writing, very little is known about whether nanoparticles may play a unique role in the chemistry of the atmosphere. It is certainly true that chemical reactions occur on atmospheric nanoparticles, just as they occur on all aerosol;[1] however, in most circumstances, the minimal contribution of nanoparticles to overall aerosol surface area (e.g., Fig. 1b) decreases the likelihood that these particles may influence the chemistry of the gas phase. The exception to this may be during ”nucleation bursts,” which are periods of intense new particle production that can dominate the surface area distribution.1-12-1 Although it may be true that the concentrations of certain reactive compounds may be enhanced in nanoparticles through the processes of nucleation and condensational growth, a simple analysis has shown that, even during nucleation burst periods, this enhancement would need to be a factor of 106 greater than reactant concentrations in typical cloud droplets to have a similar effect.[3]

Fig. 2 Schematic of the distribution of particle surface area in the atmosphere. Modes, sources, and sinks are indicated. 

Fig. 3 Residence time of particles in the lower 1.5 km of the atmosphere as a function of particle radius, generated using data from Jaenicke.[19] Also shown are approximate size ranges for predominant removal mechanisms.

OBSERVATIONS OF ATMOSPHERIC NANOPARTICLE PHYSICOCHEMICAL PROPERTIES

Most of the important advances in our understanding of atmospheric nanoparticles have resulted from dramatic improvements in instrumentation for characterizing their size, number, and composition. As we will see later, these observations have challenged our theoretical understanding of both aerosol nucleation and growth. This section will provide an overview of these instruments, and some representative measurements.

Physical Characterization

Advances in the physical characterization of atmospheric nanoparticles now allow us to size and detect them when they are less than an hour old, corresponding to diameters of about 3 nm. It is now possible to characterize nanoparticle growth rates; thus we can start to make links between the presence of these nanoparticles and the variety of phenomena that they control. The current section briefly summarizes these physical characterization techniques.

The condensation nucleus counter (CNC)[13] can detect particles as small as 3 nm in diameter.[14] It accomplishes this by flowing the sample air first through a region that is saturated with a condensing vapor, usually n-butanol, and then through a cooled region. The gas then becomes supersaturated with the vapor, which causes the particles to grow to a size that can be detected by light scattering. The minimum detectable particle size is a function of vapor supersaturation; thus two CNCs with different supersaturations (or different transmission characteristics) can be operated in parallel and, by subtracting their readings, can be used to determine the concentration of particles within a certain size range. This subtraction technique is the most frequently used way of observing atmospheric nanoparticles.[15] Recent advances have been reported in the development of particle size mag-nifiers,[16] which are similar to CNCs but are being developed with the goal of detecting particles at diameters of 1 nm and below.

Size classification of nanoparticles is now possible using a number of techniques. The differential mobility analyzer (DMA)[17] is an ”electrical mobility classifier” device that can size-classify particles down to 3 nm in diameter.[18] It does this by charging particles and passing them through a region where a sheath gas and an electric field are combined in such a way that only particles of a certain surface area-to-charge ratio can pass through the exit aperture. These instruments are commonly combined with CNCs to obtain nanoparticle size distributions. Fig. 4a shows an example of a continuous record of particle size distributions from an urban site (St. Louis, MO), which features a new particle formation event that reached detectable levels at 8:00 a.m. Subsequent growth of the mean diameter of the spectrum is quite linear (Fig. 4b). This can be equated with aerosol growth if one assumes that the aerosol was homogeneous in a large-scale air mass.

Two additional methods for obtaining nanoparticle size distributions are the pulse height analysis (PHA) and the ion mobility spectrometer (IMS) techniques. PHA[20] operates on the principle that particles smaller than 10 nm grow by condensation in the CNC to reach a unique final diameter, which can be sized by optical techniques. The IMS can measure the size spectra of charged nano-particles down to diameters below 1 nm, by measuring their drift velocity in a constant electric field.[21] IMS measurements have shown that, under certain conditions, nanoparticle formation is correlated with bursts of atmospheric ion clusters.[21]

Fig. 4 (a) Particle size distribution vs. local time for a nucleation event in St. Louis, MO. (b) Number mean diameter vs. local time for distributions in (a), showing a linear nanoparticle growth rate.

Chemical Composition

Until recently, the chemical characterization of atmospheric nanoparticles has not seen the same sort of progress as physical characterization. However, currently, several techniques have been developed for obtaining chemical composition information by: 1) indirectly inferring it by measuring some other behavior; 2) using off-line collection and analysis techniques; 3) using online real-time mass spectrometry-based techniques. In this section, these various methods will be briefly described.

Some of the earliest measurements of the chemical properties of nanoparticles were performed by using a DMA to size-select particles, then exposing these particles to a controlled environment and acquiring a particle size distribution on the result to observe any changes in aerosol size.[22] Most of these instruments, called tandem differential mobility analyzers (TDMAs), expose size-selected particles to high humidity to investigate their hygroscopic properties, but others characterize volatility by applying high temperatures[23] or organic composition by using a gas saturated with various organic vapors.[24] In a related application, the growth characteristics of aerosol in a CNC have been investigated using the PHA technique, which was used to infer that nanoparticles in a boreal forest region were composed primarily of organic acids.[25]

Off-line sampling techniques involve collecting particles on a substrate, often in a size-segregated manner, for later chemical analysis. Such analysis might involve extraction of the constituents from the substrate and analysis of the integrated composition by techniques such as ion chromatography and inductively coupled plasma mass spectrometry (ICP-MS), illuminating the substrate with a light source to investigate optical absorption, or the study of isolated particles on the substrate by microan-alytical techniques such as secondary ion mass spectros-copy (SIMS) and Fourier transform infrared spectroscopy (FTIR).[26] The application of off-line techniques to nanoparticle composition is challenged by an inability to collect sufficient mass of these particles; nonetheless, most of what is known about the composition of the smallest particles in the atmosphere is derived from offline techniques. Low-pressure impactors are often employed for collecting size-segregated ultrafine parti-cles,[27-30] but these devices do not usually extend into the sub-50 nm diameter range. Observations of ultrafine urban aerosol from low-pressure impactors show particles to be composed of 50-70% (by weight) organic compounds and 6-14% (by weight) of each of the following: elemental carbon, sulfate, nitrate, and trace metals.[27,28] Recently, a particle concentrator that allows the collection of particles as small as 10 nm in diameter, requiring integration times of 3 hr or more, has been described.[31] That study found a distinct mode in the 36-50-nm diameter range, affected primarily by combustion processes; however, correlations between elemental and organic carbon compositions also suggested that these particles contained secondary organic compounds.[31] Dimethyla-mine has been identified by the chemical analysis of newly formed particles collected in a low-pressure im-pactor in a boreal forest.[30] One recent example of the use of off-line microanalytical techniques has been reported in the study of recently formed particles in a coastal setting. That study observed both iodine and sulfur in particles with diameters below 10 nm,[32] suggesting that biogenic iodine species emitted from seaweeds may be responsible for new particle formation or growth.

On-line chemical analysis techniques directly analyze particles in real time, usually by vaporizing particles through lasers or heated surfaces, ionizing the resulting gas, and injecting the ions into a mass spectrometer for analysis.[33] These techniques are highly desirable for the study of nanoparticle composition, as they have capabilities of both high sensitivity and short sampling times. Most current on-line techniques rely on improving the nanoparticle sampling efficiency of existing instruments (e.g., Phares et al.,[34]). This has been accomplished primarily by the use of aerodynamic focusing lens sys-tems[35] whose transmission efficiency degrades significantly for particle diameters smaller than 20 nm. Most of these instruments are very new—field measurements have only just started; therefore very little published data are available. One example of published results comes from the 1999 Atlanta Southern Oxidant Study, which found that particles as small as 14 nm are almost completely organic.[36] The authors point out the ability to analyze nanoparticles is based, to some degree, on particle composition. For example, sulfate was not identified in any particles with their instrument, whereas other measurements during the same field campaign identified sulfate as a major constituent.[37]

New instruments that have been designed specifically for the on-line characterization of nanoparticles have also been introduced. Most of these new devices are limited to obtaining atomic, rather than molecular, composition of the aerosol, and usually employ high-power lasers[38] and plasmas[39] to desorb and ionize aerosols smaller than 20 nm in diameter. The author has recently reported the development of the thermal desorption chemical ioniza-tion mass spectrometer (TDCIMS), an instrument capable of on-line measurements of the molecular composition of nanoparticles as small as 5 nm in diameter at time resolutions of ca. 20 min.[40,41] The TDCIMS operates by charging and then collecting nanoparticles on a metal filament, then resistively heating the filament and analyzing the desorbed gas by chemical ionization mass spectrometry (CIMS). Fig. 5 shows an example of continuous TDCIMS measurements of sub-20-nm-diam-eter aerosols performed outside our laboratories in Boulder, CO, in the spring of 2002. The major ion observed is ammonium (Fig. 5a); however, the integrated concentration of ions larger than 65 amu varies greatly during the day (Fig. 5b). The TDCIMS has also performed measurements of the composition of freshly nucleated particles at the 2002 Aerosol Nucleation and Real-time Characterization Experiment (ANARChE) in Atlanta, GA. Preliminary results from that experiment suggest that freshly nucleated aerosols in Atlanta are composed almost entirely of sulfate with variable degrees of neutralization by ammonium.[42]

Fig. 5 Measurements of the chemical composition of sub-20 nm diameter aerosol performed outside of the author’s laboratory in Boulder, CO, on June 6, 2002. (a) Ammonium ion concentration, normalized by collected aerosol volume (left axis) and as the ratio to the ion signal from an equivalent volume of ammonium sulfate aerosol. (b) Integrated volume-normalized concentration of ions with molecular weight greater than 65 amu.

ATMOSPHERIC NANOPARTICLE FORMATION AND GROWTH: MODELS AND OBSERVATIONS

Given the current understanding of the atmospheric impact of nanoparticles, it is clear that the ability to predict the physicochemical properties of nanoparticles requires equal effort placed on understanding the growth of nanoparticles and their formation. Because instruments are, thus far, unable to directly observe newly formed aerosol in the atmosphere, the interpretation of observations of nucleation also requires an understanding of the mechanism and dynamics of condensational growth. The goal of the present section is to review the current theoretical basis for nucleation and growth.

Fig. 6 shows a schematic of the formation and growth of atmospheric nanoparticles. The primary formation process is homogeneous nucleation, which is defined as the formation of thermodynamically stable particles from the condensation of gaseous precursors. Primary sources, such as diesel engines,[43,44] can be significant sources of nanoparticles in urban settings but will not be discussed in this article. Heterogeneous nucleation, defined as condensation of gases on foreign media such as gas phase ions, might also lead to the formation of atmospheric nanoparticles. It is important to note that, in the atmosphere, homogeneous nucleation is in competition with scavenging of the low-volatility gas by preexisting aerosol. Because of this, new particle formation in the atmosphere tends to occur in bursts (Fig. 4), either when the total aerosol surface area is suddenly dropped, or when a sudden meteorological or chemical change modifies the concentration of condensable vapor.

Condensational Growth

One of the most unique properties of nanoparticles involves their evaporation and condensation behavior. As the curvature of a particle surface decreases, the separation between adjacent surface molecules increases leading to an overall decrease in the attractive forces between them. The result of this is that, at equilibrium, the partial pressure surrounding the curved surface will exceed the saturation vapor pressure for the flat surface. The equation that describes this effect is known as the Kelvin, or Thomson-Gibbs, equation:

Here the vapor saturation ratio S is defined as the ratio of the vapor pressure surrounding the particle pd to the saturation vapor pressure of the flat surface ps. The surface tension, molecular weight, and density of the liquid are given by s, Mw, and p, respectively. R is the universal gas constant, T is temperature, and d* is the diameter of a particle that will neither grow nor evaporate at S. Fig. 7a shows the relationship between S and d*. The line given by Eq. 1 can be thought of as the boundary between condensational growth (region A in Fig. 7a) and evaporation (region B). If a particle’s diameter and saturation ratio place it on the line and S were suddenly lowered, then the droplet will completely evaporate; if, conversely, S were raised, then the particle will grow indefinitely. The Kelvin equation predicts that a pure 10-nm diameter water particle would require a minimum relative humidity of 125% to prevent evaporation.

Fig. 6 Schematic of nucleation and growth processes.