Literature Review of Indoor Ultrafine Particles and Residential Gas Appliances

Transcription

1 Literature Review of Indoor Ultrafine Particles and Residential Gas Appliances Prepared For American Gas Foundation 400 N. Capitol St., NW, Suite 450 Washington, D.C July 23, 2008 Prepared By A.L. Wilson, QEP and Alexander Karpukhin Wilson Environmental Associates 41 Valente, Irvine, California

2 Table of Contents Section Title Page List of Tables List of Figures iii iv Section 1 Executive Summary 1 Section 2 Introduction 5 Section 3 Overview of Particles in Air 7 Section 4 Indoor Ultrafine Particles and Residential Natural Gas Appliances 19 Section 5 Section 6 Summary of Health Effects Studies Related to Particulate Matter and Ultrafine Particles Instruments and Techniques for Physical and Chemical Particle Characterization Section 7 References 57 Appendix A. Glossary B. Acronyms and Abbreviations C. Abstracts Wilson Environmental Associates Page ii

3 List of Tables Table Description Page 1 Urban Ultrafine Particle Concentrations 18 2 Peak Number of Ultrafine Particles 29 3 Instruments Used in Recent Ultrafine Particle Studies 50 4 Particle Size Instruments Used in Recent Studies 54 Wilson Environmental Associates Page iii

4 List of Figures Figure Description Page 1 Example of Particle Sizes 5 2 General Size Range of Common Particles 8 3 Particle Formation 9 4 Typical Fine Particle Number and Mass Distributions 11 5 Typical Monthly Ultrafine Particle Number Variation - Upland, CA 17 6 Typical Summer and Winter Size Distributions Long Beach, CA 18 7 Estimated Penetration Factor and Deposition Rate Indoors of Outdoor Generated Particles 23 8 Maximum Chamber Concentration 31 9 Estimated Source Strength Nose Breathing 34 Wilson Environmental Associates Page iv

5 Section 1 Executive Summary A literature review of indoor ultrafine particles and the reported impact of residential gas appliances and other indoor sources was commissioned by the American Gas Foundation. An exhaustive search was conducted in the English language peer reviewed literature for any publications regarding ultrafine particles and indoor sources. Conference proceedings, agency websites, research institution websites and professional association websites were also included in the search. As with most pollutants, the indoor concentrations are strongly impacted by outdoor sources and, therefore, articles concerning outdoor ultrafine particles were also reviewed. Recent health effects studies were identified and articles related to measurement techniques and instrumentation were reviewed. Ultrafine particles have been defined as those which are smaller than 100 nm (0.1 µm). However, the division between ultrafine and larger particles is somewhat arbitrary. There are no obvious boundaries created by nature between these size classes. Nearly all indoor and outdoor sources generate particles within a certain range of diameters but there is considerable size overlap between sources. General Findings Outdoor particles in the ultrafine, and more generally, submicron ranges, are generated mainly from combustion, gas to particle conversion, nucleation processes or photochemical processes, with some of them being emitted directly by the source and some formed in the air from the precursors emitted by the sources. Indoor particles in the ultrafine size range are generated from the same general source categories of outdoor particles, additionally, electric appliances including electric heaters and electric cooking devices are important indoor sources. In terms of numbers, the vast majority of indoor or outdoor airborne particles are in the ultrafine range. The total mass of the ultrafine particles is, however, often insignificant in comparison with the mass of the larger particles. Chemical composition of indoor and outdoor particles is multi-factorial and depends on particle source as well as post-formation processes. The most important chemical Wilson Environmental Associates Page 1

6 properties of outdoor particles include elemental composition, inorganic ions and carbonaceous compounds (organic and elemental carbon). Indoor ultrafine particles are most commonly found to be organic carbon compounds. Primary particles generated from combustion processes consist of soot, which is formed from hydrocarbons burning under fuel-rich conditions, trace minerals and nucleated semi-volatile organic compounds. The main chemical constituents of secondary particulate matter in urban outdoor locations commonly include sulfuric acid and ammonium sulfate, ammonium and other nitrates and organic compounds. There is also a whole suite of trace metals associated with ultrafine particles. Chemical composition of particles differs significantly and depends on the type of the local indoor and outdoor sources. Since ultrafine particles reach high concentrations in terms of their numbers but their mass is often very small, measurements of particles in the submicron ranges are more commonly based on particle number rather than mass concentration. Particle number concentration is usually measured in real time, while particle mass concentration, mass size distribution and morphology, require that samples are first collected, and then the properties investigated under laboratory conditions, using appropriate instrumentation. Particle size distributions can be estimated in real time; however, the results are less accurate and less repeatable than the real time particle number measurements. In general, the instrumentation used for particle number concentration and size distribution measurements is complicated and expensive, as the particles which they investigate, can range down to molecular sizes. Analysis of particle chemical composition is almost entirely conducted using sophisticated laboratory instrumentation, which again, requires that a representative sample be collected. The degree of correlation between particle number and mass depends on specific local conditions, of which the degree of contribution from different sources is of key importance. In general, from the measurements of particle mass, only limited information, or no information at all can be obtained about particle number and vice versa. Past studies of particle mass (PM 2.5 or PM 10 ) offer little insight regarding indoor or outdoor ultrafine particles concentrations. Particle number concentration levels in very clean environments are usually of the order of a few hundred particles/cc. In urban environments, background outdoor particle number concentrations range from a few thousand to about twenty thousand Wilson Environmental Associates Page 2

7 particles/cc. Background concentrations mean the concentrations measured at monitoring stations, which are not influenced by a nearby emission source. Near roads or inside vehicles, particle number concentrations can be more than ten times higher than the background. Outdoor particle number concentration, like the concentration of carbon monoxide, decreases exponentially with distance from the road. The outdoor concentration decreases to the urban background levels at a distance usually not greater than about 300 m from the major roadways. Indoor concentrations within 300 m of a major roadway show a significant association with traffic volume. Indoor Ultrafine Particles A limited number of studies have measured residential indoor ultrafine concentrations, normally 1 to 10 homes were monitored in each study. Smoking, cooking and candles were usually identified as indoor ultrafine sources. Indoor ultrafine particle concentration, in the absence of significant indoor sources, is normally about half of the outdoor concentrations. Outdoor particle size distributions and air exchange rates are primary variables than influence how ultrafine particles penetrate the building envelop; many other factors influence how fast the particles decay once they travel indoors. The most studied residence was probably Lance Wallace s (USEPA) townhome in Virginia; it was equipment with gas fired appliances. He found that his gas cooking appliance and gas his clothes dryer produced significant amounts of ultrafine particles. He also identified the use of candles and the use of a toaster as significant indoor ultrafine sources. The largest number of homes monitored was in a Canadian study of 36 homes with electric cooking and a variety of heating appliances. The objective was to examine the contribution of home heating systems (electric baseboard heaters, wood stoves, forced-air oil/natural gas furnace) to indoor ultrafine particle exposures. Overnight baseline ultrafine particle exposures were significantly greater in homes with electric baseboard heaters as compared to homes using forced-air oil or natural gas furnaces. Homes using wood stoves had significantly greater overnight baseline ultrafine particle exposures than homes using forced-air natural gas furnaces. In Wilson Environmental Associates Page 3

8 general, the findings suggest that home heating systems (regardless of fuel type) were not important determinants of indoor ultrafine particle exposures. Chamber Studies Two chamber studies were identified that investigated typical indoor sources. The first compared electric versus gas cooking appliances. The highest measured ultrafine particle concentration was during frying bacon on a gas range, the second highest was frying bacon on an electric grill. Burning natural gas without any pan or food produced ultrafine particles but heating the electric range tops or grill without food produced about the same number of ultrafine particles in the chamber as the gas stove. The second chamber study investigated a number of indoor sources but no natural gas appliances. It did include a propane gas camping stove. The number of particles generated per minute was estimated. The tests indicated that an electric radiator, an electric stove with no food being cooked, and several other sources produced more particles per minute of usage than the propane gas stove. There are five national particle centers funded by the USEPA that are currently conducting ultrafine particle research, primarily for outdoor air. Two large scale field studies are currently being conducted in California regarding children s asthma. They have both begun to take measurements of indoor and outdoor ultrafine particles. Wilson Environmental Associates Page 4

9 Section 2 Introduction Ultrafine particles are those smaller than 100 nm. They are more than 10,000 times smaller than the head of a pin and in the same size range as viruses (see Figure 1 below). Current federal and state regulations set health standards based on the mass of particles smaller than 10,000 nm (PM 10 ) and for those smaller than 2,500 nm (PM 2.5 ). Regulatory initiatives aimed at reducing exposure to ultrafine particles are currently under investigation and may affect the gas industry in a few years. Figure 1. Example of Particle Sizes Epidemiology studies have documented the adverse health effects of the relatively large particles (PM 10 and PM 2.5 ), including increased asthma and other respiratory diseases in children, decreased lung development in children, low birth weight of infants and premature births, cardiovascular disease including atherosclerosis and premature deaths. A small number of clinical and epidemiological studies have focused solely on ultrafine particles and have found the potential for health effects. Due to their minute size, some studies indicate that ultrafine particles can migrate from the airways to the central nervous system and organs throughout the body including the heart and brain. Research at UCLA has shown that ultrafine particles can penetrate cells and damage mitochondria. Most outdoor ultrafine particles are thought to be formed as engine exhaust gases exit the tailpipe and become diluted in air. Outdoor concentrations of ultrafine particles tend to be highest in the urban areas, where most vehicle traffic is concentrated. Levels near the road can be very high but drop off rapidly as distance from the roadway increases. Combustion of Wilson Environmental Associates Page 5

10 natural gas and other fuels from stationary sources is also estimated to be a large source of outdoor ultrafine particles. As with most other outdoor pollutants, ultrafine particles penetrate indoor and is the dominant source of the average indoor exposures. There are several indoor sources of ultrafine particles: mostly combustion related. Few studies have specifically investigated indoor sources of ultrafine particles and there is much speculation regarding the role of all sources of combustion indoors to exposures of ultrafine particles. The purpose of this report was to conduct a literature review of technical research articles that discuss issues related to indoor ultrafine particles concentrations and sources, with emphasis on those that mention natural gas appliances. Section 3 presents a general discussion of particles in air, their formation and fate in air. Section 4 is the main body of the literature search, and the most relevant studies are summarized. Section 5 discusses the health effects literature and Section 6 discusses the measurement methodologies. Section 7 is the list of references cited in the earlier sections. The appendix includes a glossary, an acronym listing and over 100 abstracts of the articles reviewed with links to the full text of each article. Wilson Environmental Associates Page 6

11 Section 3 Overview of Particles in Air Indoor particulate matter is a complex mixture of particles (solid or liquid) ranging from the size of large molecules to those large enough to be seen with the naked eye. Common household dust is a solid particle with which we re all familiar. However, particles may also be very small liquid droplets suspended in air. Different particles may have substantially different shapes and chemical properties. Thus the health effects they may cause are highly variable. This Section provides a brief overview of particles and how they are classified for air pollution studies. A brief discussion of their sources, formation and transformation is provided. The current state of knowledge on ultrafine particles related to indoor air and natural gas appliances is summarized in Section 4. Section 5 discusses the general health concerns of ultrafine particles and Section 6 discusses measurement techniques and instrumentation. Particle Terminology Discussion of particles in air involves unique and sometimes confusing terminology that is related to size, mass and how the particles were formed. Particles are also commonly referred to as aerosols in many air pollution publications. However, aerosol can also refer to both the particle and the gas in which it is suspended. Various classifications and terminologies have been used to define particle size. Four air pollution terms most commonly used to discuss particle size are coarse particles, fine particles, ultrafine particles and nanoparticles. Course particles are those larger than 2500 nm (2.5 µm), fine particles are those between 100 and 2500 nm, and ultrafine particles are those smaller than 100 nm (0.1 µm). Nanoparticles generally refer to those that are smaller than 50 nm. Figure 2 is a sample of common substances and their approximate size range. Wilson Environmental Associates Page 7

12 Figure 2 General Size Range of Common Particles µm Ultrafine Particles Molecules Viruses Bacteria Tobacco Smoke Particles Suspended in Air Particle Diameter (nm) The most common terminology that is currently used in ambient air quality standards are PM 2.5 and PM 10. PM 2.5 (or fine particle mass) is the mass concentration of particles with aerodynamic diameters smaller than 2.5 µm (2,500 nm). PM 10 is the mass concentration of particles with aerodynamic diameters smaller than 10 µm (10,000 nm). Measurements of PM 2.5 and PM 10 include the mass contributed by ultrafine particles, however, that mass is typically a very small fraction of the total. Two common particle size terms related to the ultrafine particle formation process are: Nucleation mode: particles in this mode are formed by nucleation of gases in a supersaturated atmosphere and are the smallest of the ultrafine particles. Accumulation mode: particles in this mode originate from primary emissions as well as through gas to particle conversion, chemical reactions, condensation and coagulation. A large fraction of these particles are typically larger 100 nm. A primary particle is a particle introduced from the source into the air in solid or liquid form, while a secondary particle is formed in the air by gas to particle conversion. Particles in the Wilson Environmental Associates Page 8

13 ultrafine range outdoors are both primary and secondary particles generated mainly from combustion, gas to particle conversion, nucleation or photochemical processes. Ultrafine particles in indoor air are mixtures generated by a large number of sources: outside air (mobile sources, photochemical reactions), cigarette smoking, cooking, electrically heating and cooking appliances, candles, chemical reactions, etc. Figure 3 shows a diagram of particle formation from gaseous molecules. Ultrafine particles are volatile, they do not maintain their size or composition and smaller particles tend to evaporate. Small particles grow to larger ones and large particles can become smaller particles or simply evaporate. Combustion Molecules Ultrafine particles are volatile. They do not maintain their size or composition. Smaller particles tend to evaporate. Neutral or Ion Clusters (~ 1 nm) Detectable Particles (~ 3 nm) Loss By Condensation To Pre-Existing Surfaces Figure 3. Particle Formation Physical Properties of Particles The most important physical properties of indoor and outdoor particles include: number concentration (particles per cc), number size distribution, mass concentration, mass size Wilson Environmental Associates Page 9

14 distribution, surface area, shape and electrical charge. To a large extent these are the physical properties of the particles that underlie particle behavior in the air and ultimately removal from the air. The health effects of particles are strongly linked to particle size. It is their size that is a predictor of the region in the lung where the particles will deposit and of the outdoor and indoor locations to which the particles are able to penetrate or be transported. Sampling of particles and choice of appropriate instrumentation and methodology is primarily based on particle physical properties. Particles suspended in the air range in size from about 1 nm to about 100,000 nm (Baron and Willeke, 2001): the former is a large molecular size and the latter is the size above which particles settle rapidly due to gravitational force and are thus removed from the air. Indoor and outdoor sources generate particles with a wide range of sizes, rather than particles of a single size. A single pollution source operating under steady state conditions, is likely to have a particle size distribution with one distinctive peak and sometimes additional, usually much smaller peaks. Those peaks are called modes of the distribution. Different emission sources are characterized by different size distributions. However, these distributions are not unique to these particle sources alone. For such mixtures of particles of different size distributions, the measured distribution may or may not display individual peaks from the contributing sources, and thus may or may not be used for source identification (source signature). Particle size distributions are most commonly presented either in terms of number or mass. Sometimes surface area or volumes are calculated and shown as size distributions. In terms of number, the vast majority of outdoor or indoor particles are in the ultrafine range. The total mass of the ultrafine particles is, however, often insignificant in comparison with the mass of a small number of larger particles with which most of the mass of airborne particles is associated. Therefore the peak in the number distribution spectrum appears in the area where there is almost no mass in the mass distribution spectrum and vice versa. The peak in the mass distribution spectrum is where the particle number is very low. Particle surface area, in turn, is the largest for particles somewhat above the ultrafine size range. Figure 4 shows a typical particle number and particle mass distribution. Notice that most of the mass is comprised of a relative small number of particles. Wilson Environmental Associates Page 10

15 Particles from combustion sources, including vehicles and gas appliances, are generally very small and are formed by a complex process within the flame (Bockhorn, 2000). Almost all of the particles from natural gas combustion are in the ultrafine range resulting from nucleation and accumulation, typically nm (Chang, 2004). Particles from CNG vehicles are smaller than from diesel or gasoline vehicles and range from nm, with a majority being between 20 and 60 nm (Ristovski, 2000). Gasoline vehicle combustion particles are mostly from nm. A significant proportion of diesel vehicle combustion emission particles have diameters smaller than 100 nm (Morawska, 1998; Ristovski, 1998). The majority of particles emitted from biomass burning, which includes controlled burning and uncontrolled fires, are ultrafine, with only a small fraction in the larger size range (WHO, 1999). Figure 4 Typical Fine Particle Number and Mass Distributions 100 Number Distribution Relative Value Mass Distribution Particle Diameter (nm) Ultrafine, fine and coarse particles usually result from different generation processes and only occasionally will the same source generate particles with broad size distributions covering all three ranges. Forest fires are an example of a source of airborne combustion products of the fire as well as large diameter particles that are entrained into the smoke column. Thus different sources contribute to generation of particles in the submicron range which is predominant in particle number, and different sources contribute to larger particles, Wilson Environmental Associates Page 11

16 which predominate in mass. Therefore, it is rare that there is a correlation between fine and coarse airborne particles, or a correlation between particle number and mass. Chemical Composition of Particles The chemical composition of particles is multi-factorial and depends on particle sources as well as post-formation processes. The most important chemical properties of particles include: Elemental composition Inorganic ions Carbonaceous compounds (organic and elemental carbon) Interest in the elemental composition, in general, derives from the potential health effects of heavy elements like lead, arsenic, mercury and cadmium, and the possibility of using the elements as source tracers. Water-soluble ions such as potassium, sodium, calcium, phosphates, sulfates, ammonium and nitrate associate themselves with water in the indoor environments and can also be used for source apportionment. Carbonaceous compounds are composed of organic and elemental carbon. The former can contain a wide range of compounds such as polycyclic aromatic hydrocarbons, pesticides, phthalates, flame retardants and carboxylic acids some of which are tracers for certain sources while the latter is sometimes termed soot, black carbon and graphitic carbon. Particles arising through different formation mechanisms display significantly different properties both physical (as discussed above) and chemical. In particular, there are substantial differences in the chemistry of primary and secondary particles. Primary particles generated from combustion processes are mainly soot, which is formed from hydrocarbons burning under fuel-rich conditions. Soot formed under such conditions appears most commonly as an ensemble of ultrafine particles, but the size of the agglomerates can extend up to a few hundred nanometers (Bockhorn, 2000). The formation of soot, which is the conversion of a hydrocarbon fuel molecule containing few carbon atoms into carbonaceous agglomerate containing some millions of carbon atoms, is an extremely complicated process. It is a gaseous-solid phase transition where the solid phase exhibits no unique chemical or physical structure. Therefore, soot formation encompasses Wilson Environmental Associates Page 12

17 chemically and physically different processes, including formation and growth of large aromatic hydrocarbons and their transition to particles, the coagulation of primary particles to larger aggregates, and the growth of solid particles by picking up growth components from the gas phase. Natural gas combustion has been shown to produce much lower numbers of primary particles than fuel oil or coal (Chang, 2004) and the subsequent particle formation in the plume is very different from other fuels. Secondary particles are generated through two basic processes (Baron, 2001; Bockhorn, 2000): (i) gas-phase chemical reactions - gases produce low-volatility products which are capable of homogenous nucleation to form new particles which are of molecular sizes that can then increase in size by coagulation and are captured by preexisting ambient particles; (ii) low-volatility gas-phase reaction products condense onto pre-existing ambient particles - the so-called heterogeneous nucleation process. While homogenous nucleation may potentially increase both the number and the mass of particles per unit volume in air, heterogeneous nucleation can only increase the mass. The main chemical constituents of secondary particulate matter found in outdoor and indoor air commonly include sulfuric acid and ammonium sulfate, ammonium and other nitrates and organic compounds (Derwent, 2000). Measurements of ultrafine particle mass concentration made in seven Southern Californian cities (Cass, 2000) showed that outdoor ultrafine particle concentrations averaged µg/m 3. The chemical composition of the particles averaged 50% by mass organic compounds, 14% trace metal oxides, 8.7% elemental carbon, 8.2% sulfate, 6.8% nitrate, 3.7% ammonium ion, 0.6% sodium and 0.5% chloride. The most abundant catalytic metals identified in the ultrafine region included: Fe, Ti, Cr, Zn, with Ce also present. The ranges of concentrations are quite large and depend on the local composition of the emission sources. Most indoor ultrafine particles are organic compounds. Wilson Environmental Associates Page 13

18 Combustion Emissions Under ideal conditions, complete combustion of natural gas or other fuel would result only in the generation of carbon dioxide (CO 2 ) and water. Both are greenhouse gases. Any products other than CO 2 and water are often called products of incomplete combustion and include particles and gases. The most common gas associated with incomplete combustion of natural gas is carbon monoxide (CO). Nitrogen oxides (primarily NO and NO 2 ) are emitted due to the nitrogen content of the combustion air and the nitrogen entrained in the fuel. Combustion of different types of fuels or other materials results in emissions of various trace elements, which are present in the fuel or material. In most cases, there is not just one specific element that is related to the combustion, but an entire source profile of elements. All combustion sources generate large amounts of volatile and semi-volatile organic compounds. Semi-volatile organic compounds can be present in the air either in the vapor or in particle from (solid or liquid). Exposure to many of the organic compounds emitted to the air has been associated with various types of health effects. Polynuclear aromatic hydrocarbons (PAH), some of which are classified as carcinogenic, are one class of compounds contained in the organic fraction of the fine particulate matter. PAH compounds are synthesized from carbon fragments into large molecular structures in low-oxygen environments, such as occurs inside the flame envelope in the fuel-rich region of the flame structure. If the temperature is not adequate to decompose compounds upon exiting from the flame zone, then they are released into the free atmosphere and condense or are adsorbed onto the surface of particles. Many different combustion sources, including natural gas, are known to produce PAH compounds. The most studied PAH is benzo[a]pyrene, which is a physiologically active substance that can contribute to the development of cancer in human cells. From the point of view of the effect on human health, the specific physical form of the semivolatile compounds when they are inhaled could be of significance. They could be either in vapor form, or could be associated with particles of specific sizes. Wilson Environmental Associates Page 14

19 Residence Time in the Air Following emission, pollutants undergo dilution with air, and then undergo various types of changes and transformations. Larger particles are gravitationally deposited soon after being emitted, while smaller particles remain suspended in the air for hours and days. Processes such as coagulation, diffusion, and convection transport, govern the behavior and fate of ultrafine particles in air. Outside particles can travel indoors, penetrating the structure through the infiltration and ventilation air. The physical process of penetration removes some of the particles and is dependent on particle size and composition. Indoor particles experience the same chemical and physical removal processes as outdoor particles but removal rates can be accelerated due in part to the large surface areas of interior walls, flooring and furniture; and, being inhaled by the occupants. Natural Gas Combustion and Ultrafine Particles Particle formation in a plume is a complex mixture of nucleation, coagulation, and condensational growth. The concentrations of condensable organic compounds and solid or liquid particles in combustion exhausts are the primary determinants of particle formation while the plume mixes with the relatively cold ambient air. A study, partially funded by GRI, investigated particle formation in combustion exhaust plumes from burning natural gas, fuel oil and coal (Chang, 2004). The found that ultrafine particle generation was highly variable within the first 10 seconds of dilution with ambient air but had fairly stabilized after about one minute. Particles in combustion exhaust derive from mineral matter and other impurities in the fuel, carbonaceous particles formed during combustion, condensation of inorganic and organic vapors, and chemical reactions. The hot exhaust when mixed with ambient air is rapidly cooled, resulting in vapor compounds nucleating homogeneously and heterogeneously, or condensing on pre-existing particles. Condensational growth of particles in a diluted plume depends on temperature, relative humidity, aging time, mixing rate, and partitioning of species between the gaseous and solid phases. Similar processes should occur within the indoor environment as a plume from a gas range mixes with the other cooking fumes and gases as well as with the room air. Wilson Environmental Associates Page 15

20 Chang, et al, attempted to burn the three fuels in a similar manner using the same combustion equipment and test methodologies in an effort to develop a common test procedure for stationary combustion sources. Their results are illustrative of the size distribution from natural gas combustion in a burner and allows a relative comparison with coal and fuel oil combustion. The tests showed that combustion of natural gas produced particles generally much smaller than with fuel oil or coal. The particle size mode for natural gas was 15 to 25 nm, coal was 40 to 50 nm and fuel oil was 70 to 100 nm. After the same stable plume travel time and dilution, the total ultrafine particle count per volume of exhaust was similar for all three fuels, normally within a factor of 2 or 3 between the fuels. This is interesting given the huge difference in total particulate mass emissions expected from the three fuels. An earlier test program (Hildemann, 1991) also used a dilution sampling system to collect primary particle emissions from numerous sources in the Los Angeles Basin including a residential natural gas water heater. Particle size was not measured but their findings may be relevant since most of the mass was probably in the ultrafine size range for the gas appliances. Particulate emissions from residential natural gas combustion when tested by dilution sampling were found to be extremely low, ranging from 23 to 72 ng/kj of fuel burned. By comparison, an another earlier study (Muhlbaier, 1982) measured a total particulate emission rate of 90 ng/kj from combustion of 93% methane in a residential-sized furnace. In other earlier studies, the measured fine particulate emission rates were found to be 410 ± 190 ng/kj from a stovetop burner (Traynor, 1982), and ng/kj from several gas-fired space heaters (Girman, 1982). The dilution sampling resulted in lower values compared to more common test protocols. Organic carbon accounted for between 83 and 91% of the measured fine particulate mass emitted during the Hildemann, et al, home appliance tests. Conversion of the carbon emissions to an estimate of the total organic mass emitted showed that 99% of the total fine particulate mass can be attributed to organic material. Previous studies of carbonaceous aerosol emission from residential natural gas combustion have been published (Muhlbaier, 1982; Traynor,1982). Two earlier studies also measured detectable amounts of particulate sulfur compounds from natural gas home appliances (Sexton,1968; Traynor, 1982), which can be expected to vary depending on the sulfur content of the natural gas burned. Wilson Environmental Associates Page 16

21 Temporal And Spatial Variation of Outdoor Ultrafine Particles Several studies have shown that urban outdoor ultrafine particles are mostly generated from mobile sources, similar to carbon monoxide (CO) and nitrogen oxides (NO and NO 2 ). Near the road, high correlations have been found for ultrafine particle number, CO and NO concentrations. Ultrafine particles decrease rapidly with distance from the road. Typically, the ultrafine particle concentration 300 m from major roadways is nearly the same as the urban background. Figure 5 shows the typical monthly particle number concentration variation for Upland, Ca. Other locations will have slightly different patterns, but this one is thought to be representative of the general pattern. Particle counts at this site were nearly twice as high in the winter compared to the summer months. This is similar to the distribution of CO but opposite of photochemical pollutants such as ozone. Figure 5 Typical Monthly Ultrafine Particle Number Variation - Upland, CA Particle Number per cc Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec The ultrafine particle counts change with the seasons and the particle sizes of the ultrafines also change with the seasons. Figure 6 shows a typical size distribution shift for winter and summer (Long Beach, Ca). Other locations will show different patterns but Long Beach is thought to be representative of an upwind urban area with a mixture of point, area and mobile sources. Wilson Environmental Associates Page 17

22 Figure 6 Typical Summer and Winter Size Distributions - Long Beach, CA Particle Number per cc Particle Diameter Summer Winter Human Exposure to Ultrafine Particles In the absence of unusually high indoor sources (heavy smoking, extensive cooking, long term candle burning, malfunctioning appliances, etc) the indoor ultrafine particles are expected to be lower than outside and especially lower than on roadways. The following are order of magnitude estimates of average exposures to ultrafine particles and are presented as an illustration of likely exposures: Table 1. Urban Ultrafine Particle Concentrations Approximate Typical Urban Ultrafine Microenvironment Particle Concentration (number per cc) Suburban Residential (sleeping) 2000 Suburban Residential (evening hours) 5000 Urban Office 5000 Urban Outdoor 20,000 In-vehicle (arterial road) 50,000 In-vehicle (freeway) 150,000 The amount of time spent on major roadways is a major determinant of the expected exposures to ultrafine particles. Wilson Environmental Associates Page 18

23 Section 4 Indoor Ultrafine Particles and Residential Natural Gas Appliances This chapter presents a summary of selected studies regarding residential indoor ultrafine particle measurements with an emphasis on natural gas appliances. The appendix contains the abstracts of over 100 articles that are related to the ultrafine particle issue. Section 6 presents a discussion of the measurements techniques and the appendix contains literature for selected instruments. As with most other indoor pollutants, the outside air is a large contributor to indoor ultrafine particles. In fact, indoor air usually contains less ultrafine particles than outdoor air. Smoking, cooking, candles and other sources can, of course, cause indoor levels to temporarily exceed outdoor concentrations. A few recent outdoor air studies that don t include indoor measurements are discussed to provide a general understanding of the sources and the spatial variability of ultrafine particles. Most of the basic research regarding particle formation and potential health effects are derived from outdoor measurements. Instrumentation and measurement techniques have been involving over the past 10 years. However, there is no standard or reference method that is used for indoor or outdoor measurements of ultrafine particles. Probably the most accurate and reproducible measurement of ultrafine particles is total particle count. However, as discussed in Section 6, great variability can exist from one instrument to another depending upon manufacturer or model. Care should be used when comparing absolute values of ultrafine particle measurements between the different studies. The selected studies are presented by publication date. Most of the summaries are derived from the abstracts. The appendix contains the complete abstract and links to the full article. Many articles are free to download but most will cost $15 to $40 for the full text. Wilson Environmental Associates Page 19

24 Residential Indoor Measurements Real-Time Monitoring of Particles, PAH, and CO in an Occupied Townhouse (Wallace, 2000) Wallace has published several papers concerning indoor particles and is recognized for his pioneering efforts regarding indoor air quality. His residence has been a virtual test house for indoor measurements for several years. His home has natural gas cooking, gas water heater and a gas forced air furnace. All of his ultrafine particle publications are included in this chapter in order of the publication year. In this first paper, he discusses particles, PAH, and CO results from the first 16 months of monitoring. The monitoring indicated that neighborhood wood burning and morning rush hour traffic were the most important sources of PAH and black carbon outdoors at his residence. Also, candles, matches, incense, and frying, sautéing, broiling, deep-frying, and stir-frying were important indoor sources of particles. One citronella candle was an extremely powerful PAH source. Frying, grilling, and sautéing were extremely strong indoor sources, together with combustion events such as use of matches and candles. Physical movement was an important source of coarse but not fine particles. Use of the gas stove for extended periods of time led to increased CO concentrations--vehicles and wood burning were relatively minor sources in comparison. The gas oven, gas burners, and electric toaster oven were important sources of ultrafine particles. Characterization of Indoor Particle Sources Using Continuous Mass and Size Monitors (Long, 2000) A comprehensive indoor particle characterization study was conducted in nine Boston-area homes in 1998 in order to characterize sources of particles in indoor environments. State-of-the-art sampling methodologies were used to obtain continuous PM 2.5 concentration and size distribution particulate data for both indoor and outdoor air. Study homes, five of which were sampled during two seasons, were monitored over week-long periods. Among other data collected during the monitoring efforts were 24-hr elemental/organic carbon particulate mass data as well as semicontinuous air exchange rates and time-activity information. The data set showed that indoor particle events tend to be brief, intermittent, and highly variable, thus requiring the use of continuous instrumentation for their characterization. Source event data demonstrated that the impacts of indoor activities were especially pronounced in both ultrafine and coarse modes. Among the sources of ultrafine particles characterized in this study were indoor ozone/terpene reactions. The data suggest that organic carbon was a major constituent of particles emitted during indoor source events. Characterization of Indoor Particle Sources: A Study Conducted in the Metropolitan Boston Area (Abt, 2000) An intensive particle monitoring study was conducted in homes in the Boston, Massachusetts, area during the winter and summer of 1996 in an effort to characterize sources of indoor particles. As part of this study, continuous particle size and mass concentration data were collected in four single-family homes, with each home monitored for one or two 6-day periods. Additionally, housing activity and Wilson Environmental Associates Page 20

25 air exchange rate data were collected. Cooking, cleaning, and the movement of people were identified as the most important indoor particle sources in these homes. Continuous Monitoring of Ultrafine, Fine, and Coarse Particles in a Residence for 18 Months in (Wallace, 2002) Continuous monitors were employed for 18 months in an occupied townhouse to measure ultrafine, fine, and coarse particles; air change rates; wind speed and direction; temperature; and relative humidity (RH). A main objective was to document short-term and long-term variation in indoor air concentrations of size-resolved particles caused by (1) diurnal and seasonal variation of outdoor air concentrations and meteorological variables, (2) indoor sources such as cooking and using candles, and (3) activities affecting air change rates such as opening windows and using fans. A second objective was to test and compare available instruments for their suitability in providing real-time estimates of particle levels and ancillary variables. Despite different measuring principles, the instruments employed in this study agreed reasonably well for particles less than 10,000 nm in diameter. When an indoor source was operating, particle concentrations for different sizes ranged from 2 to 20 times the average concentrations when no indoor source was apparent. Indoor sources, such as cooking with natural gas (no electric stove was tested), and simple physical activities, such as walking, accounted for a majority (50-90%) of the ultrafine and coarse particle concentrations, whereas outdoor sources were more important for accumulation-mode particles between 100 and 1000 nm in diameter. Averaged for the entire year and including no periods when indoor sources were apparent, the number distribution was bimodal, with a peak at approximately 10 nm (possibly smaller), a shallow minimum at approximately 14 nm, and a second broad peak at approximately 68 nm. Source Strengths Of Ultrafine And Fine Particles Due To Cooking With A Gas Stove (Wallace, 2004) Semi-continuous instruments with fine size discriminating ability were used to calculate particle counts in 124 size bins from 10 to 2500 nm. Data were collected at 5 min intervals for 18 months in an occupied house. Tracer gas measurements were made every 10 min in each of 10 rooms of the house to establish air change rates. Cooking episodes were selected meeting certain criteria, and the number and volume of particles produced were determined for each size category. The selected cooking episodes (mostly frying) were capable of producing about particles over the length of the cooking period (about 15 min), more than 90% of them in the ultrafine range. Frying produced peak numbers of particles at about 60 nm, with a secondary peak at 10 nm. The peak volume occurred at a diameter of about 160 nm. Since the cooking episodes selected were biased toward higher concentrations, the particle concentrations measured during about 600 hours of morning and evening cooking over a full year were compared to concentrations measured during noncooking periods at the same times. Cooking was capable of producing more than 10 times the ultrafine particle number observed during non-cooking periods. Levels of PM 2.5 were increased during cooking by a factor of 3. Breakfast cooking (mainly heating water for coffee and using an electric toaster) produced concentrations about half those produced from more complex dinnertime cooking. Wilson Environmental Associates Page 21

26 Ultrafine Particles From A Vented Gas Clothes Dryer (Wallace, 2005) Ultrafine particles were measured continuously for 18 months in an occupied townhouse. A major source was determined to be the gas clothes dryer. Although the dryer was vented to the outdoors it consistently produced an order of magnitude increase in the ultrafine concentrations compared to times with no indoor sources. Short-term peak number concentrations exceeded 100,000/cm 3 on a number of occasions. The source strength was conservatively estimated at about 6x10 12 ultrafine particles produced per drying episode. These values are underestimates, since the part of the peak below 9.8 nm was not measured. Averaged over 150 hours of operation, the number concentration showed a major peak at the smallest size measured (9.8 nm) and a secondary peak at 30 nm. Loss rates of the ultrafines due to diffusion, deposition, and particle growth were high compared to losses due to air exchange. Penetration of Freeway Ultrafine Particles Into Indoor Environments (Zhu, 2005) Four two-bedroom apartments within 60 m from the center of the 405 Freeway in Los Angeles, CA were used for this study. Indoor and outdoor ultrafine particle size distributions in the size range of nm were measured concurrently under different ventilation conditions without indoor particle generation sources. The size distributions of indoor particles showed less variability than the adjacent outdoor particles. Indoor to outdoor ratios for ultrafine particle number concentrations depended strongly on particle size. Indoor/outdoor (I/O) ratios also showed dependence on the nature of indoor ventilation mechanisms. Under infiltration conditions with air exchange rates ranging from 0.31 to 1.11 per hour, the highest I/O ratios were usually found for larger ultrafine particles, while the lowest I/O ratios were observed for particulate matter of nm. This research effort is an excellent source of information concerning the impact of outdoor ultrafine partice sources on indoor air quality. Figure 7 shows the estimated penetration factor of outdoor particles generated by a nearby freeway and the deposition rate of those particles after they entered the residence. Notice that the smallest particles had low penetration (less were able to enter the structure) but once they were inside, the smaller particles remained in the air longer than the larger particle. Wilson Environmental Associates Page 22

27 Figure 7 Estimated Penetration Factor and Deposition Rate Indoors of Outdoor Generated Particles Particle Diameter (nm) Penetration Factor (I/O) Deposition Rate (per Hr) Volatility of Indoor And Outdoor Ultrafine Particulate Matter Near A Freeway (Kuhnnext, 2005) This study examined the volatility of penetrating ultrafine outdoor particles, predominantly from freeway emissions, into indoor environments where other particle sources were minimized and no cooking activities took place. A tandem differential mobility analyzer (TDMA) system was used to study particle volatility at two apartments, 15 and 40 m downwind of the I-405 Freeway in Los Angeles, CA. The first differential mobility analyzer (DMA) selected particles of a certain diameter and subsequent heating of this monodisperse aerosol allowed for detection of changes in particle diameters by measuring the resulting size distribution with a second DMA. Particle volatility was examined by measuring changes in particle diameters as well as volume and number concentrations. Results suggest that outdoor particles are more volatile than indoor aerosols in those homes with no cooking activities. Evaluation of outdoor particle volatility as a function of distance to the freeway revealed that outdoor particle volatility decreased with distance from the source. Indoor and Outdoor Concentrations of Ultrafine Particles in Some Scandinavian Rural and Urban Areas (Matson, 2005) The concentration of ultrafine particles was measured in some rural and urban areas of Sweden and Denmark. The instruments used were handheld real-time condensation particle counters, models CPC 3007 and P-Trak 8525, both manufactured by TSI. Field measurements in Sweden were conducted in a few residential and office buildings, while in Denmark the measurement sites comprised two office buildings, one of them located in a rural area. The concentration of UFPs was measured simultaneously indoors and outdoors with condensation particle counters. The results revealed that the outdoor-generated particle levels were major contributors to the indoor particle number concentration in the studied buildings when Wilson Environmental Associates Page 23

28 no strong internal source was present. The results showed that in office buildings, the UFP concentrations indoors were typically lower and correlated fairly well to the number concentration outdoors. The determined indoor-outdoor ratios varied between 0.5 and 0.8. The indoor levels of UFPs in offices where smoking was allowed was sometimes recorded higher than outdoor levels, as in one of the Danish offices. In residential buildings, the indoor number concentration was strongly influenced by several indoor activities, e.g., cooking and candle burning. In the presence of significant indoor sources, the indoor/outdoor (IO) ratio exceeded unity. The magnitude of UFP concentrations was greater in the large city of Copenhagen compared to the medium-size city of Gothenburg and lowest at more rural sites. Indoor Ultrafine Particle Exposures and Home Heating Systems: A Cross-Sectional Survey of Canadian Homes During the Winter Months (Weichenthal, 2006) The objective was to examine the contribution of home heating systems (electric baseboard heaters, wood stoves, forced-air oil/natural gas furnace) to indoor ultrafine particle exposures. A cross-sectional survey was conducted in 36 homes with electric cooking in the cities of Montréal, Québec, and Pembroke, Ontario. Realtime measures of indoor UFP concentrations were collected in each home for approximately 14 hours, and an outdoor ultrafine particle measurement was collected outside each home before indoor sampling. A home-characteristic questionnaire was also administered, and air exchange rates were estimated using carbon dioxide as a tracer gas. Average ultrafine particle exposures of 21,594/cc and 6660/cc were observed for the evening and overnight hours, respectively. In an unadjusted comparison, overnight baseline UFP exposures were significantly greater in homes with electric baseboard heaters as compared to homes using forced-air oil or natural gas furnaces, and homes using wood stoves had significantly greater overnight baseline ultrafine particle exposures than homes using forced-air natural gas furnaces. However, in multivariate models, electric oven use, indoor relative humidity, and indoor smoking were the only significant determinants of mean indoor ultrafine particle exposure, whereas air exchange rate and outdoor ultrafine particle s were the only significant determinants of overnight baseline ultrafine particle exposures. In general, the findings suggest that home heating systems were not important determinants of indoor ultrafine particle exposures. The measurements did not include particles lower than 20 nm and may have underestimated the impact of natural gas appliances or other sources of particles below that size. Kinetic Analysis of Competition Between Aerosol Particle Removal and Generation by Ionization Air Purifiers (Alshawa, 2007) A simple kinetic model was used to analyze the competition between the removal and generation of particulate matter by ionization air purifiers under conditions of a typical residential building. The model predicted that certain widely used ionization air purifiers may actually increase the mass concentration of fine and ultrafine particulates in the presence of common unsaturated VOC, such as limonene contained in many household cleaning products. The prediction was supported by an explicit observation of ultrafine particle nucleation events caused by the addition of D-limonene to a ventilated office room equipped with a common ionization air purifier. Wilson Environmental Associates Page 24

29 Levels of Ultrafine Particles in Different Microenvironments Implications to Children Exposure (Diapouli, 2007) Indoor and outdoor ultrafine particles concentration levels were examined in the area of Athens during cold a period of 2003 and Seven primary schools, located in areas with different characteristics of urbanization and traffic density, as well as a typical suburban residence, were monitored. Moreover, in-vehicle concentration levels, while driving along major avenues and in the heavy-trafficked centre of Athens, were measured. UFPs number concentration was monitored by condensation particle counter. The highest mean indoor concentrations were observed in a small carpet-covered library and a teachers' office at the same school. The highest outdoor concentrations were measured at two schools, both affected by heavy traffic. Finally, the highest in-vehicle concentrations were measured in central commercial areas of Athens during, on average, 55 min drives. Indoor-to-outdoor concentration (I/O) ratios were below 1.00 at all sites. The largest ratio (0.88) was observed in the residence, during a day when there was cleaning activity in the room monitored. Outdoor concentrations diurnal cycles, both outside the schools and the residence, were closely related to traffic. Indoor concentrations inside schools were relatively stable in classrooms. Nevertheless, number concentrations exhibited variability when there were significant changes in room occupancy. Diurnal variation of indoor concentrations at the residence followed the respective outdoor one with a delay of 1 hour or less, in the absence of strong indoor sources, indicating the major contribution of outdoor UFPs to the indoor concentration levels. Wilson Environmental Associates Page 25

30 Characteristics of Indoor Aerosols in Residential Homes in Urban Locations: A Case Study in Singapore (Balasubramanian, 2007) As part of a major study to investigate the indoor air quality in residential houses in Singapore, intensive particle measurements were made in an apartment in a multistory building for several consecutive days in The purpose of this work was to identify the major indoor sources of fine airborne particles and to assess their impact on indoor air quality for a typical residential home in an urban area in a densely populated country. Particle number and mass concentrations were measured in three rooms of the home using a real-time particle counter and a lowvolume particulate sampler, respectively. Particle number concentrations were found to be elevated on several occasions during the measurements. All of the events of elevated particle concentrations were linked to indoor activities based on house occupant log entries. This enabled identification of the indoor sources that contributed to indoor particle concentrations. Activities such as cooking elevated particle number concentrations. The fine particles collected on Teflon filter substrates were analyzed for selected ions, trace elements, and metals, as well as elemental and organic carbon (OC) contents. To compare the quality of air between the indoors of the home and the outdoors, measurements were also made outside the home to obtain outdoor samples. The chemical composition of both outdoor and indoor particles was determined. Indoor/outdoor (I/O) ratios suggest that certain chemical constituents of indoor particles, such as chloride, sodium, aluminum, cobalt, copper, iron, manganese, titanium, vanadium, zinc, and elemental carbon, were derived through migration of outdoor particles, whereas the levels of others, such as nitrite, nitrate, sulfate, ammonium, cadmium, chromium, nickel, lead, and OC, were largely influenced by the presence of indoor sources. Wilson Environmental Associates Page 26

31 Outdoor Measurements Physical and Chemical Characterization of Atmospheric Ultrafine Particles in the Los Angeles Area (Hughs, 1998) The number concentration, size distribution, and chemical composition of atmospheric ultrafine particles was determined under wintertime conditions in Pasadena, CA. These experiments were conducted using a scanning differential mobility analyzer, laser optical counter, and two micro-orifice impactors. Samples were analyzed to create a material balance on the chemical composition of the ultrafine particles. The number concentration of ultrafine particles analyzed over 24-h periods, was found to be consistently in the range of 13,000 particles/cc. Ultrafine particle mass concentrations were in the range µg/m 3. Organic compounds were the largest contributors to the outdoor ultrafine particle mass concentration. The Chemical Composition of Atmospheric Ultrafine Particles (Cass, 2000) Measurements of ultrafine particle mass concentration made in seven Southern California cities showed that ultrafine particle concentrations in the size range nm aerodynamic diameter average 0.55 to 1.16 µg/m 3. The chemical composition of these ultrafine particle samples averaged 50% organic compounds, 14% trace metal oxides, 8.7% elemental carbon, 8.2% sulphate, 6.8% nitrate, 3.7% ammonium ion, 0.6% sodium and 0.5% chloride. The most abundant catalytic metals measured in the ultrafine particles were Fe, Ti, Cr, Zn, with Ce also present. Ultrafine particle primary emissions were estimated to arise principally from mobile and stationary fuel combustion sources and were estimated to consist of 65% organic compounds, 7% elemental carbon, 7% sulphate, 4% trace elements, with very small quantities of sodium, chloride and nitrate. Size-Fractionated Measurements of Ambient Ultrafine Particle Chemical Composition in Los Angeles Using The NanoMOUDI (Sardar, 2004) Reported the size-fractionated ultrafine chemical composition at urban source sites and inland receptor sites in the Los Angeles basin over three different seasons. Sizefractionated ultrafine particles were collected by a NanoMOUDI over a period of 2 weeks at each site. Measurements of ultrafine mass concentrations varied from 0.86 to 3.5 µg/m 3 with the highest concentrations observed in the fall. The chemical composition of ultrafine particles ranged from 32 to 69% for organic carbon (OC), 1-34% for elemental carbon (EC), 0-24% for sulfate, and 0-4% for nitrate. A distinct OC mode was observed between 18 and 56 nm in the summer, possibly indicating photochemical secondary organic aerosol formation. Nitrate and sulfate were measurable only in the larger size ranges of ultrafine particles. Size and Composition Distribution of Airborne Particulate Matter in Northern California: I - Particulate Mass, Carbon, and Water-Soluble Ions (Herner, 2005) Outdoor measurements of particulate mass for several size ranges were made during the period December 16, February 3, 2001, at six locations near or within the San Joaquin Valley (SJV) of California. Peak ultrafine particle mass concentrations (8-12 hr average) were ~2.4 ug/m 3 measured at night in Sacramento and Bakersfield. Ultrafine particle concentrations were distinctly diurnal, with daytime Wilson Environmental Associates Page 27

32 concentrations ~50% lower than nighttime concentrations. Ultrafine particulate mass concentrations did not accumulate during the multi-week stagnation period; rather, ultrafine particle mass decreased at Bakersfield as PM 1.8 mass was increasing. The majority of the ultrafine particle mass was associated with carbonaceous material. Ultrafine particle mass concentrations in the SJV did not increase during a multiweek stagnation episode, even though PM 2.5 mass concentrations during the same time period increased by a factor of 7. Outdoor nighttime mass concentrations of ultrafine particles were twice as large as outdoor daytime mass concentrations. Mobile Platform Measurements of Ultrafine Particles and Associated Pollutant Concentrations on Freeways and Residential Streets in Los Angeles (Westerdahla, 2005) This paper describes the integration of multiple monitoring technologies on a mobile platform designed to characterize ultrafine particle and associated pollutants in Los Angeles traffic. Monitoring technologies included two condensation particle counters and scanning mobility particle sizers for ultrafine particle. Real-time measurements made of NOx, black carbon, particulate matter-phase PAH, and particle length showed high correlations with ultrafine particle counts, (r 2 =0.78 for NO, 0.76 for BC, 0.69 for PAH, and 0.88 for particle length). Average concentrations of ultrafine particle and related pollutants varied strongly by location, road type, and truck traffic volumes, suggesting a relationship between these concentrations and truck traffic density. Wilson Environmental Associates Page 28

33 Chamber Studies Ultrafine Particles and Nitrogen Oxides Generated by Gas and Electric Cooking (Dennekamp, 2001) Laboratory experiments with gas and electric range top burners, range top grills, and ovens were used to compare different cooking procedures in Aberdeen, Scotland. Nitrogen oxides were measured by a chemiluminescent ML9841A NOx analyzer. A TSI 3934 scanning mobility particle sizer was used to measure average number concentration and size distribution of particles in the size range nm. High concentrations of particles were generated by gas combustion, by frying, and by cooking of fatty foods. Electric range top burners and grills may also generate particles from their surfaces. In the two experiments where electric range tops was the only source of particles (no cooking), the particle size at peak concentration was 32 and 37 nm. In the two experiments where gas burning on the range top was the only source of particles, the particle size at peak concentration was 16 and 26 nm. When no food was on the range top grill, the electric unit particle size at peak concentration was 20 nm. Similarly, the gas unit particle size at peak concentration was 24 nm with no food. When bacon was fried on the gas or electric top burners the particles were of larger diameter, in the size range nm. The smaller particles generated during experiments grew in size with time because of coagulation. Substantial concentrations of NOx were generated during cooking on gas; four range top burners for 15 minutes produced 5 minute peaks of about 1000 ppb nitrogen dioxide and about 2000 ppb nitric oxide. The following table summarizes the available Dennekamp, et al, chamber tests for ultrafine particle concentration of similar cooking events while using gas and electric stoves. It is interesting to note that the electric stove range top units produced about the same amount of particle concentration as gas burners when no food was being prepared. Table 2. Peak Number of Ultrafine Particles Chamber Cooking Experiment Peak Number of Ultrafine Particles /cc (x10,000) Gas Cooking Electric Cooking 1 Burner No Food Burners No Food Ring Stir Fry 14 1 Ring Fry Bacon Bake Cake (oven) 10 3 Roast Meat (oven) 12 2 Bake Potatoes (oven) 13 2 Grill No Food 10 8 Grill Toast Grill Bacon Wilson Environmental Associates Page 29

34 Characterization of Indoor Sources of Fine and Ultrafine Particles: A Study Conducted in a Full-Scale Chamber (Afshari, 2005) In this study, 13 different typical indoor particle sources were for the first time examined in a full-scale chamber (no other chamber studies were found during this literature search). Two different instruments, a condensation particle counter (CPC) and an optical particle counter (OPC) were used to quantitatively determine ultrafine and fine particle emissions, respectively. The CPC measured particles from 20 nm to larger than 1000 nm. The OPC was adjusted to measure particle concentrations in eight fractions between 300 nm and 1000 nm. The sources tested were cigarette side-stream smoke, pure wax candles, scented candles, a vacuum cleaner, an airfreshener spray, a flat iron (with and without steam) on a cotton sheet, electric radiators, an electric stove, a propane gas stove, and frying meat on the electric stove. The cigarette burning, frying meat, air freshener spray and propane gas stove showed a particle size distribution that changed over time towards larger particles. In most of the experiments the maximum concentration was reached within a few minutes. Typically, the increase of the particle concentration immediately after activation of the source was more rapid than the decay of the concentration observed after deactivation of the source. The highest observed chamber concentration of ultrafine particles was approximately 241,000 particles/cm 3 and originated from the combustion of pure wax candles. The lowest observed chamber concentration of ultrafine particles was approximately 550 particles/cm 3 was observed when ironing without steam on a cotton sheet. The highest generation rate of particles was observed in the electric radiator test. Figures 8 and 9 show the Afshari chamber data related to ultrafine particle concentration and emission rate for all of the reported tests. Electric appliances were found to generate large amounts of ultrafine particles. The electric radiator, electric stove and electric air heater were all higher emitters than the propane gas camping stove. Wilson Environmental Associates Page 30

35 Figure 8 Maximum Chamber Concentration (ultrafine particles/cc) Flat iron (w ithout steam) on a cotton sheet Flat iron (w ith steam) on a cotton sheet Vacuum cleaner w ith full bag (50 min) Air-freshener spray (20g for 20 sec) Vacuum cleaner (motor) w ithout bag (200 min) Scented candles (2 for 2 hrs) Propane gas camping stove (no food, no pot) Electric stove (no food, no pot) Electric air heater Electric stove frying meat (15 hamburgers, 3 min each) Cigarettes(3 for 10 min each) Electric radiator Pure w ax candle (2 for 1 hr) Figure 9 Estimated Source Strength (ultrafine particles/min) Flat iron (w ithout steam) on a cotton sheet Flat iron (w ith steam) on a cotton sheet Vacuum cleaner w ith full bag (50 min) Vacuum cleaner (motor) w ithout bag (200 min) Scented candles (2 for 2 hrs) Propane gas camping stove (no food, no pot) Air-freshener spray (20g for 20 sec) Pure w ax candle (2 for 1 hr) Cigarettes(3 for 10 min each) Electric air heater Electric stove (no food, no pot) Electric stove frying meat (15 hamburgers, 3 min each) Electric radiator 0.00E E E E E E+12 Wilson Environmental Associates Page 31

36 Section 5 Summary of Health Effects Studies Related to Particulate Matter and Ultrafine Particles The objective of this Section is to give a brief summary of health effect studies related to particulate matter (PM) with an emphasis on recent ones that included ultrafine particles. Three recent publications provided excellent resources for this Section: US Environmental Protection Agency (USEPA 2003) Fourth External Review Draft of Air Quality Criteria for Particulate Matter, Volumes 1 and 2, World Health Organization publication (WHO, 2002), Guidelines for Concentration and Exposure-Response Measurement of Fine and Ultra Fine Particulate Matter for use in Epidemiological Studies, Australian Department of the Environment and Heritage (Morawska 2004) Health Impacts of Ultrafine Particles. In 2006, the Southern California Air Quality Management District (SCAQMD) hosted a conference titled Ultrafine Particles: The Science, Technology, and Policy Issues. Several interesting and useful presentations are available through the District s website ( An excellent summary titled Ultrafine Particle Health Effects by John R. Froines of the Southern California Particle Center provides a discussion of the issues for a general audience. The Froines presentation can be viewed at: Ultrafine Particle Issue Within most established ambient monitoring networks, particles are measured by PM 10 and PM 2.5 mass measurement protocols. Therefore, most of the evidence of health effects of outdoor particles is based on these measures. Total suspended particulate mass is almost entirely dominated by particles generated by mechanical processes such as dust resuspension, grinding, and mining. PM 10 is sometimes dominated by particles generated by mechanical processes and sometimes by accumulation mode particles, which is an aged fraction of the outdoor aerosol emitted by many sources such as industrial emissions, Wilson Environmental Associates Page 32

37 vehicle emissions, and power generation. In addition, a portion of the particles is of biogenic origin and generated from secondary processes such as photochemical processes. PM 2.5 is more likely to be dominated by accumulation mode particles, but can also include a significant fraction of mechanically generate particles. The relative contributions from different types of sources are very site specific. Health effects studies since the 1980 s have provided evidence for associations between exposure to particulate matter and various types of health effects (Weichenthal 2007; Franck 2006; Sioutas. 2005; Pope 2000; Vinzents 2005). Particulate matter was characterized in those studies in different ways, including total mass concentration or its fractions such as PM 10 or PM 2.5 or PM 0.1 or its elemental or chemical composition. Increases in all-cause mortality have been reported in studies of the acute and chronic effects of particulate matter. The increase in all-cause mortality was attributed to increases in respiratory disease and cardiovascular disease mortality (Barregard 2006; Cohen 2000). Consistent with the reported increases in acute mortality, there was also a reported increase in hospital admissions for respiratory diseases and cardiovascular diseases in association with particulate air pollution (Morawska 2007). Subpopulations consisting of susceptible individuals were reported to show increases in symptoms, medication use and physiological parameters consistent with an exacerbation of pre-existing respiratory or cardiovascular diseases. Recent animal and human clinical studies are examining associations between exposure expressed in terms of particle mass and number, particularly of fine and ultrafine particles, and adverse effects. Toxicological studies have indicated that measures such as particle number or particle surface might be better indicators for ultrafine particles and that these particles potentially cause health effects (Lighty et al. 2000; Oberdörster et al. 2002; Fromme 2006). The evidence for health effects due to exposures characterized in terms of particle number, particle surface, or ultrafine particles from epidemiological studies is limited (MacNee and Donaldson 1998; Cyrys 2004). Two studies on adult asthmatics conducted (Cohen et al. 2004; Penttinen et al. 2001) suggested that decreases in peak expiratory flow rates were more strongly associated with ultrafine than with fine particles. In addition, a recent mortality study suggested independent associations of ultrafine and fine particles with acute mortality (Cyrys et al 2004). Wilson Environmental Associates Page 33

38 Most particles used in laboratory animal toxicology studies were in the fine particle mode. However, the great number and huge surface area of ultrafine particles in some outdoor PM atmospheres necessitated considering the size of the particle in assessing biological responses. Some studies have examined differences in toxicity between fine and ultrafine particles, with the general finding that the ultrafine particles show a significantly greater response at similar mass doses, as well as some differences in rates and pathways of clearance from the respiratory tract (MacNee 1998). Furthermore, the amplitude of any biological response due to PM exposure may be dependent upon the total surface area, which is in contact with tissue, indicating that surface chemistry phenomena are also likely involved in PM-induced effects (Oberdörster 2004). Thus, size, as well as other physical properties of inhaled PM, may play a significant role in changing a biological response in some cases, perhaps for some responses, and this may be independent of the specific chemical composition of the particles being evaluated. However, it is clear that all particles do not produce the same response, and particle composition likely does play a role in some of the adverse health effects associated with exposure to indoor and outdoor PM (Dreher 2000). Figure 10. Nose Breathing Wilson Environmental Associates Page 34

39 The contribution of particle size to toxicity remains a key question in the study of particulate matter health effects. Although ultrafine particles contribute little to mass, they may be important in producing biological responses because of their high number concentration and surface area. As shown in the figure above (ICRP Model,1994; Nose-breathing) the ultrafine particles can travel to the lowest parts of the lungs. Ultrafine particles may be important inducers of pulmonary and vascular effects because of their ability to evade cell macrophage phagocytosis (removal of pathogens), and to enter alveolar epithelial cells, the lung interstitium, and even pulmonary capillary blood (Frampton et al. 2005). These studies suggest that exposure to ultrafine particles at very low mass concentrations may cause effects on circulating leukocytes. Because ultrafines may rapidly enter the circulation, the effects of ultrafine particles on blood leukocytes and cardiovascular function remain an important area of investigation. Both fine and ultrafine particles appear to affect health outcomes such as mortality and respiratory and cardiovascular morbidity and may do so independently of each other. Past Approaches to Estimate Health Effects of Particulate Matter Studies of acute health effects evaluate the impact of day-to-day or hour-to-hour variation in outdoor pollution on health. Studied outcomes span the range from mortality counts of populations to changes in physiological parameters or biomarkers of individuals. Associations between particulate matter and increases in mortality have been shown in more than 60 studies conducted in at least 35 cities throughout the world as recently reviewed by Pope (2000a). These studies evaluated the effects of particulate matter in locations with different climatic conditions and different sources of particulate matter. In addition to mortality data, many studies and reviews have also evaluated the association between the prevalence of symptoms, diseases and hospitalization and PM air pollution (Pekkanen 2004). Hospital admission data provides a powerful tool to study the effects of particle air pollution on specific diseases or a combination of conditions such as respiratory and cardiovascular diseases. Panel studies provide a powerful tool to study the day-to-day changes in health status in Wilson Environmental Associates Page 35

40 potentially susceptible subpopulations. Their aim is to provide evidence that early physiological changes or changes in biomarkers occur, which are consistent with disease exacerbation or mortality. For example, elevated levels of particulate air pollution have been associated with short-term responses such as decreased lung function, and increased respiratory symptoms such as cough, shortness of breath, wheezing, and asthma attacks (Dominici et al. 2002). Recent panel studies have also focused on cardiac endpoints and found air pollution to be associated with increased heart rate, decreased heart rate variability, and detection of arrhythmias by implantable defibrillators (Pope 2000). No associations between PM and blood oxygenation have been observed. Often these panel studies are combined with intermittent clinical examinations. Studies of chronic health effects evaluate the impact of particle exposures over several years. These studies provide evidence on the impact of outdoor particulate air pollution on life expectancy as well as the development of chronic diseases or impaired physiological parameters. In one of the first major prospective cohort study on the relationship between annual average pollution levels and adjusted mortality-rate ratios, a cohort of over 8000 adults in six cities followed over years, Dockery et al. (1993) found that although many pollutants were associated with increasing mortality, the association was strongest for PM 2.5. Studies based on cohorts have been a powerful tool to assess the chronic health effects of air pollution. A cohort is recruited and characterized with respect to major known risk factors for the outcome. Air pollution as well as health outcomes are monitored thereafter during the follow-up. In order to study air pollution, exposure gradients have to be created by selecting appropriate geographical regions. Geographical variation in exposures within a city or district is generally thought to be insufficient to allow study of long-term effects. Ideally the regions would differ only with respect to air pollution concentrations, but would otherwise be similar with respect to risk factor profiles. Toxicological studies, which involve controlled exposures of either humans or animals in laboratory settings, have been employed in attempts to evaluate the physicochemical properties of particulate matter which may be responsible for the health outcomes observed in epidemiological studies and to understand the underlying biological mechanisms for these outcomes. Wilson Environmental Associates Page 36

41 Early studies examining the toxicology of inhaled PM were largely focused on a limited number of chemical species. One such group, which received extensive attention was acidic sulphates, namely sulphuric acid and its partially neutralized products of reaction with ambient ammonia (Fromme 2006). Short-term exposures of healthy animals (with the exception of the guinea pig) to sulphuric acid at high concentrations, well above those found in outdoor air, did not alter standard tests of pulmonary function, nor did it result in pulmonary inflammation. Another chemical group examined extensively in early studies was the metals. These studies noted that acute or chronic exposures to pure metallic particles, e.g., arsenic, cadmium, copper, vanadium, iron, or zinc, could have various effects on the respiratory tract (Oberdörster et al 2004), although concentrations employed were much greater than those which would occur in the outdoor atmosphere. Some other types of particulate matter commonly used in earlier toxicological evaluations of PM included fly ash, volcanic ash, coal dust, carbon black, and titanium dioxide. Some of these particles were models of nuisance or inert dusts, i.e., those having low intrinsic toxicity, and were used to evaluate the occurrence of non- specific particle effects, i.e., responses which may be independent of the actual chemical nature of the material used. In general, only some inflammatory responses and mild pulmonary functional changes were noted following exposure to these particles, and this occurred only at high concentrations. More current toxicological studies have focused on the use of actual outdoor air PM and on various types of specific combustion-generated particles, and select constituents of these, increasing the relevance to actual outdoor inhalation exposures. In addition, the role of physical properties of PM, especially particle size, in modulating biological response has received much interest. A number of types of combustion-generated particles have received interest in toxicological studies. One type is diesel exhaust particles. These have been evaluated in terms of the relationship between exposure and alterations in aspects of immune function, especially the potential to modulate the induction or exacerbation of allergic airway disease (Cohen et al. 2004). The inhalation or instillation, albeit at high concentrations, of diesel exhaust particles enhanced the production of antigen-specific antibodies, and also induced non-specific airway hyper- responsiveness. Wilson Environmental Associates Page 37

42 Another class of particulate matter which has received attention in recent toxicological studies is fly ash, generally coal or residual oil fly ash (ROFA) (Dreher 2000). Most such studies involved ROFA, which has a high content of water-soluble sulphate and metals. Studies with this material, rather than with pure metal particles as was done in the past, have allowed for the examination of the role of metals in biological responses to mixed chemical PM. The instillation of high doses of ROFA generally produced pulmonary inflammation, which was most likely due to aqueous leachable chemical constituents of the particles, generally metals. Furthermore, different biological effects were found to be due to different metal constituents, and were also influenced by the interaction between the metals and the acidity of the particles. Current Health Effect Research Activities That Include Ultrafine Particles There are several ongoing air pollution research efforts that have begun to include ultrafine particles in their evaluations. The following is a discussion of current activities that were identified in the published literature. Much of the future research efforts in the USA is expected to occur through the efforts of five research centers funded mostly by the USEPA: The Bloomberg School of Public Health at Johns Hopkins University, Baltimore, Md. The Center will map health risks of PM across the US based on analyses of national databases on air pollution, mortality, and hospitalization, and then use the maps to guide detailed monitoring and collection of PM samples for physical, chemical, and biological characterization in assays relevant to pulmonary and cardiovascular outcomes. Harvard University, Boston, Mass. The fundamental objective of this Center is to understand how specific PM characteristics and sources impact inflammation, autonomic responses, and vascular dysfunction. University of California, Los Angeles, Calif. -- Center researchers will investigate the underlying mechanisms that produce the health effects associated with Wilson Environmental Associates Page 38

43 exposure to particulate matter, and attempt to understand how toxic mechanisms and resulting health effects vary with the source, chemical composition and physical characteristics of particulate matter. University of California, Davis, Calif. Researchers will investigate the properties of particles that are responsible for human health effects, the metabolism that underlies these effects, and the consequences of chronic exposures, especially during childhood, that make individuals more susceptible to adverse effects. [more information on this center] University of Rochester, Rochester, N.Y. Researchers will investigate the mechanisms by which fine and ultrafine particles from specific sources cause adverse cardiovascular effects, particularly in susceptible groups such as diabetics and those with cardiovascular disease. California Air Resources Board (ARB) is sponsoring research to investigate issues related to ultrafine particles. One research effort investigates emissions, including studying the characteristics of ultrafine particles emitted from compressed natural gas engines compared to those from diesel engines, including the impact of control technologies. Exposure-related ultrafine PM measurements are being conducted under ARB contracts supporting the Southern California Particle Center exposure facility, a 12-station ultrafine monitoring network, in-vehicle and in-home measurement studies, and an upcoming mobile monitoring study. Health-related research on ultrafine PM is being performed under contracts supporting studies of the mechanisms of PM toxicity. More importantly, the ARB is supporting studies to evaluate health impacts from exposure to ultrafine PM as part of the Children's Health Study and the Fresno Asthmatic Children's Environment Study. A new ARB program with the University of California is an epidemiology study focusing on the potential health effects to elderly individuals exposed to ultrafine PM. US Department of Health and Human Services, National Institutes of Health (NIH) is funding several health studies that include ultrafine particles in the analysis. Wilson Environmental Associates Page 39

44 The Mickey Leland National Urban Air Toxics Research Center (NUATRC) and the Health Effects Institute (HEI) are also currently funding ultrafine particle research activities including Relationships of Indoor, Outdoor, and Personal Air (RIOPA). European Union is funding Health Effects of Air Pollution on Susceptible Subpopulations (HEAPSS) project that studies the risk of hospitalization and death due to air pollution, in particular airborne ultrafine particles. European Respiratory Society (ERS) is sponsoring several health effects studies related to ultrafine particles. Wilson Environmental Associates Page 40

45 Section 6 Instruments and Techniques for Physical and Chemical Particle Characterization The objective of this Section is to give an overview of methods of physical and chemical characterization of airborne particulate matter with an emphasis on ultrafine particles. A recent World Health Organization publication (WHO, 2002), Guidelines for Concentration and Exposure-Response Measurement of Fine and Ultra Fine Particulate Matter for use in Epidemiological Studies, provided an excellent resource for this Section. Mass Measurement by Filtration Filtration is the most commonly used method for collection of airborne particles. The conventional procedure involves drawing air through a filter followed by deposition of the particles onto the filter. Interferences to the particle sample encountered during the filtration stage include: loss of particles passing through the pores of the filters loss of particles from incorrect seals employed in the filter holder around the filter loss or contamination of particles from the filter holder particle transformations that may occur on the filter. A detailed discussion of sampling issues and techniques is given by the EPA (US EPA 1996; US EPA 1999; US EPA 2001). These interferences and the requirements of filter analysis methods are discussed in detail in Lehtimaki and Willeke (1993). The filter should be made of materials that do not add the concentration of elements under study. Berg et al (1993) analyzed 19 filters from various manufacturers for 30 elements by ICP-MS to determine the most appropriate filter for ultratrace metal analysis. The authors found that PTFE filters were the cleanest, whereas glass fiber filters contained a number of elements contributing to high blanks. Similar studies have been performed solely on glass fiber filters to evaluate the analytical blanks obtained as a result of the dissolution of the filter matrix (Wang et al. 1989). The inlet used for particle sampling should follow certain design guidelines so that comparable samples are obtained. The purpose of the inlet system is to collect a sample Wilson Environmental Associates Page 41

46 that is representative of the air and then deliver it to the samplers. The US EPA requires, for measurement of outdoor particulate matter mass by the US Federal Reference Method, that the filter be maintained at a temperature near the outdoor temperature. Similar requirements are also made in the European reference method (CEN 1998). However, for other measurements of indoor and outdoor particle properties, sampling equipment may need to be housed in a shelter that provides a controlled laboratory environment (temperature C). However, consideration should be given to changes in particle size and composition due to changes in temperature and relative humidity relative to the indoor or outdoor conditions. An omni-directional inlet, is required. This can be achieved with a vertical configuration. The size of the entrance configuration must be well designed to provide a high inlet sampling efficiency for particles over a wide range of wind speeds. For indoor measurements, precautions related to wind speed and precipitation need not be met. Special care should be given to air splitting, bends and horizontal lines to avoid losses of large particles. Mass concentration is the most commonly made measurement on filtered particulate samples. It is used to determine compliance with ambient standards (PM 2.5, PM 10, etc). Gravimetric analysis is used almost exclusively to obtain mass measurements of filters in a laboratory environment. Detailed standardized procedures for mass analyses have been published in the European Standard EN and by the United States Environmental Protection Agency (US EPA 1987; US EPA 1999; CEN 1998). Gravimetric measures the net mass on a filter by weighing the filter before and after sampling with a balance in a temperature- and relative humidity-controlled environment. Beta Attenuation Mass Measurement In the beta attenuation mass monitor (Marcias et al. 1976; Willeke and Baron 1993) particulate material is continuously collected on a foil, and the attenuation of beta radiation by the collected material is recorded. The collection process is repeated in fixed intervals, which leads to a quasi- continuous measurement of particulate mass vs. time. In contrast to optical absorption, the absorption of beta radiation is closely proportional to mass, and the proportionality constant is only weakly material dependent. Thus, although the physical quantity of mass is not directly determined, beta absorption can be calibrated for mass, with rather small errors arising from the variability of the chemical species present. With the correct sampling head, the instrument can measure selected size ranges. Wilson Environmental Associates Page 42

47 Vibrational Microbalance Mass Measurement A piezoelectric quartz crystal as used in wristwatches typically has a natural vibration frequency (first harmonic) in the range of several MHz, which can be excited electrically. If aerosol particles are deposited on such a crystal, the frequency changes according to the respective mass. The sampling time required to collect a sufficient amount of material on the quartz surface is typically minutes. A serious disadvantage of this method is the change in sensitivity with loading, when more than one monolayer of particles has been collected. The measurement time before necessary cleaning is thus rather short. Strongly agglomerated particles are not measured correctly, because they do not firmly attach to the crystal. These limitations are eliminated in the Tapered Element Oscillating Microbalance Mass Detector (TEOM) (Patashnik and Rupprecht 1980), which oscillates at much lower frequencies than the piezoquartz. The particle sample is collected in a filter mounted on the thin end of a tapered oscillating hollow element, which is fixed to a surface at its thick end. This element is electrically excited to oscillate in its natural frequency, which changes with the mass loading of the filter. The instrument can be equipped with various size sampling inlets. The inlet air is either heated to 50 C to keep moisture in the vapor phase or dried with a diffusion dryer. In the latter case it only requires heating to 30 C. Optical Scattering Mass Measurement Instruments are on the market that measure an integrated light scattering signal of all the particles drawn into the device and then convert the signal into an approximate particle mass. For this conversion, assumptions have to be made regarding the particle optical properties, and especially the size distribution of the particles. The reading of the instrument can be orders of magnitude wrong, if these assumptions are badly met. Using such instruments for scientific studies is therefore generally not recommended. In special cases, where continuous recording of the mass is required, and where the reading is periodically compared to a reference method indicating the actual mass, the response can be used to interpolate between the points of reference. Wilson Environmental Associates Page 43

48 Determination of Chemical Composition In general, no single method can measure all chemical species, comprehensive aerosol characterization programs may use a combination of methods to address complex needs. This allows each method to be optimized for its objective, rather than be compromised to achieve goals unsuitable to the technique. Such programs also greatly aid quality assurance objectives, since confidence may be placed in the accuracy of a result when it is obtained by two or more methods on different substrates and independent samplers. In general terms, methods for characterization of chemical composition of particles can be divided into two classes: Methods that require sample collection followed by laboratory analyses; and Near real time methods. The most important features of both these classes of chemical analysis are summarized below. More details of instrumentation used for these analyses are provided in the cited literature. It should be stressed that this summary is designed to be illustrative rather than exhaustive, since new methods are constantly appearing as old methods are being improved. The more commonly used methods for chemical analysis requiring sample collection can be divided into four categories: Elements Water-soluble ions Organic compounds, and Elemental carbon. Elements, water soluble ions, and organic and elemental carbon typically explain 65 to 85% of the measured particulate mass and are adequate to characterize the chemical composition of measured mass for filter samples collected in most urban and non-urban areas. Some chemical analysis methods are non-destructive. These are preferred because they preserve the filter for other uses. Methods, which require destruction of the filter, are Wilson Environmental Associates Page 44

49 best performed on a section of the filter to save a portion of the filter of other analyses or as a quality control check on the same analysis method. The most common interest in elemental composition derives from concerns about health effects and the utility of these elements to trace the sources of suspended particles. Instrumental neutron activation analysis (INAA), photon-induced x-ray fluorescence (XRF), particle-induced x-ray emission (PIXE), atomic absorption spectrophotometry (AAS), inductively-coupled plasma with atomic emission spectroscopy (ICP/AES), inductivelycoupled plasma with mass spectroscopy (ICP/MS), and scanning electron microscopy with x-ray fluorescence (SEM/XRF) have all been applied to elemental measurements of particle samples. AAS and ICP/AES are also appropriate for ion measurements when the particles are extracted in deionized-distilled water (DDW). Since air filters contain very small particle deposits, preference is given to methods that can accommodate small sample sizes. XRF and PIXE leave the sample intact after analysis so that it can be submitted to additional examinations by other methods. Excellent agreement was found for the inter-comparison of elements acquired from the XRF and PIXE analyses (Cahill 1989). The analysis for water-soluble ions can be divided into two main classes: ion chromatographic and automated colorimetric analysis. Chromatographic methods make use of a stationary phase and a mobile phase. Components of a mixture are carried through the stationary phase by the flow of the mobile one. In ion-exchange chromatography, ionic components of the sample are separated by selective exchange with counter ions of the stationary phase. The use of ion exchange resins as the stationary phase brings about a classification based largely on geometry and size. Colorimetric methods involve the passing of radiation through an unknown sample solution and measurement of the absorbency of the solution for comparison with a set of standards. In the simplest form, daylight may commonly serve as the radiation source and the human eye as the detector. Automated colorimetric analysis (AC) applies different colorimetric analyses to small sample volumes with automatic sample throughput. Since IC provides multi-species analysis for the anions, AC most commonly measures ammonium. Organic compounds comprise a major portion of airborne particles in the atmosphere. Specific groups of organic compounds (e.g., polycyclic aromatic hydrocarbons, PAHs) have Wilson Environmental Associates Page 45

50 also been implicated in human health effects. However, due to the very complex composition of the organic fraction of atmospheric aerosols, the detailed composition and atmospheric distributions of organic aerosol constituents are still not well understood. For organic analysis, particulate matter is most frequently collected on quartz-fiber filters that have been specially treated to achieve low "carbon blanks". Outdoor and indoor organic particulate matter has also been collected on a variety of particle sizing devices, such as low pressure impactors and Micro Orifice Uniforms Deposit Impactors ("MOUDI"). However, the task of sampling organic compounds in airborne particles is complicated by the fact that many of these compounds have equilibrium vapor pressures so that the concentration in the gas phase is of the same order of magnitude as the concentration in the condensed phase. This implies a temperature- and concentration-dependent distribution of such organic compounds between particulate and vapor phases. It also suggests that artifacts may occur during the sampling process. Volatilization would cause the under- estimation of the particlephase concentrations of organic compounds. Conversely, the adsorption of gaseous substances on deposited particles or on the filter material itself would lead to overestimation of the particle-phase fraction. In addition, chemical degradation of some organic compounds may occur during the sampling procedure. Total carbon in aerosol particles (TC) can be expressed as the sum of organic carbon (OC), elemental carbon (EC), and carbonate carbon (CC), with the contribution of CC to TC usually on the order of 5% or less, for particulate samples collected in urban areas (Appel 1993). In thermal separation methods, thermally evolved OC and EC are oxidized to CO 2 and quantified either by non-dispersive infrared detection or electrochemically, or the CO 2 can be reduced to CH 4 and quantified via flame ionization detection (FID). The various methods give similar results for TC, but not for EC or OC. In a methods comparison study (Countess 1990), it was shown that it is necessary to minimize or correct for pyrolytically generated EC ( char ), and that CC found in wood smoke and automobile exhaust samples may interfere with some of the thermal methods. Pyrolytic char is corrected for in Thermal/Optical Reflectance (TOR) and Thermal/Optical Transmission (TOT) methods for the quantification of OC and EC on quartz fiber filter deposits. In thermal-optical carbon analysis (Birch and Cary 1996; Chow et al. 1993) punches from a quartz sampling filter are inserted into the carbon analyzer and heated in a Wilson Environmental Associates Page 46

51 helium atmosphere to volatilize organic carbon. Then, the temperature is reduced, and oxygen is added to the carrier gas, so that desorbed compounds are then oxidized to CO 2, reduced to methane, and measured in a flame ionization detector. In order to account for the portion of the OC that is pyrolyzed, a He-Ne laser monitors the sample reflectance (or transmittance). As the pyrolysis occurs, the sample gets darker, and the reflectance decreases. Then, as elemental carbon is removed, the filter lightens, and the reflectance increases until all carbon has been removed from the filter. The split between organic and elemental carbon is considered to be the point at which the reflectance regains its prepyrolysis value, with material removed prior to this point being considered organic, and that after, elemental. In the TOR method the reflectance is monitored throughout the analysis and the TOR detection limit for EC and OC is 100 ng/m3. Continuous Measurement of Chemical Components of Particulate Matter Recently, several researchers have developed instruments for real-time in situ analysis of single particles (e.g., Noble and Prather 1996; Gard et al. 1997; Johnson and Wexler 1995). Although the technique varies from one laboratory to another, the underlying principle is to fragment each particle into ions using either a high power laser or a heated surface and to then use a time-of-flight mass spectrometer (TOFMS) to measure the ion fragments in a vacuum. Each particle is analyzed in a suspended state in the air stream (i.e, without collection), avoiding sampling artifacts associated with impactors and filters. By measuring both positive and negative ions from the same particle, information can be obtained about the chemical composition, not just the elemental composition, of individual particles of know aerodynamic diameter. This information is especially useful in determining sources of particles. Since particles are analyzed individually, biases in particle sampling (the efficiency of particle transmission into the sensor chamber as a function of size; particle size measurement, and detection of particles prior to fragmentation) represent a major challenge for these instruments. Moreover, the mass spectrometer has a relatively large variability in ion yields (i.e., identical samples would yield relatively large differences in MS signals Wilson Environmental Associates Page 47

52 [Thomson and Murphy 1994]); therefore, quantification is inherently difficult (Murphy et al. 1997). Quantification is even more challenging for complex organic mixtures. An integrated collection and vaporization cell was developed by Stolzenburg and Hering (2000) that provides automated, 10-min resolution monitoring of fine particulate nitrate. In this system, particles are collected by humidified impaction process and analyzed in place by flash vaporization and chemiluminescent detection of the evolved nitrogen oxides. Several instruments have been developed that collect and analyze atmospheric organic particulate matter with better than 2 hour time resolution. An automated carbon analyzer with 15-min to 1-h time resolution is now commercially available (Rupprecht et al. 1995) and has been operated in several locations. It collects samples on a 100 nm impactor downstream of an inlet with a 2500 nm cut point. Use of an impactor eliminates gas adsorption that must be addressed when filter collection is used. However, this collection system may experience substantial particle bounce, and a sizeable fraction of EC is in particles 200 nm. In the analysis step carbonaceous compounds are removed by heating in filtered outdoor air. Carbonaceous material removed below 340 EC is reported as organic carbon, material removed between 340 and 750 EC is reported as elemental carbon. Turpin et al. (2000) comment that it would be more appropriate to report carbon values obtained by this method as low- and high-temperature carbon, since some organics are known to evolve at temperatures greater than 340 EC (e.g., organics from wood smoke). Hering (unpublished) has modified this system. Higher collection efficiencies are obtained for smaller particles by growing the particles by humidification prior to impaction. Number Concentration Measurement Most of the recent ultrafine particle field studies report particle number concentration (particles per cc) as their primary measurement. The vast number of indoor or outdoor particles are in the ultrafine mode. The most direct way of measuring particle number concentration is by counting particles in a given volume. In optical particle counters (OPCs) the light scattered by single particles is recorded. This is done by directing a constant aerosol flow through the focus of a light beam and by collecting a portion of the scattered light into a detector. Each particle produces a pulse of the detector Wilson Environmental Associates Page 48

53 output and the number of pulses during fixed time intervals yields the particle concentration. This technique only detects particles large enough to deliver a scattering signal in the measurable range. The cut-off size for most OPCs is above 100 nm (upper end of ultrafine particles). In order to register all particles in the relevant size range, condensation particle counters (CPCs) are used. A CPC contains a particle magnifier, in which a liquid (normally butanol or water) condenses on the particles, growing them to a size that is detectable by an OPC, which follows in a continuous flow arrangement. Most CPCs detect particles down to a diameter of 10 nm, some units can detect particles down to 3 nm. The lower cut-off size of particle counters is of great importance, especially if a significant amount of the particles are smaller than the cut-off. This is the case, even for a cut-off of 10 nm in environments where fresh aerosols from combustion are present, e.g. in cities, where automobiles are a major source or indoors where natural gas may be an important source. In order for the response of particle counters to be comparable, the counting efficiencies must be precisely known as a function of particle size. For most instruments the counting efficiencies are near unity for most of the stated size range and steeply fall to zero, when the particle size is reduced to the minimum size certified. The critical size defining this cut-off curve may vary considerably from instrument to instrument, even within the same series. Comparison and inter-calibration of the counting instruments used within a large field study is essential. A calibration procedure is described by Wiedensohler et al (1997). There is presently no alternative to CPCs for particle concentration measurement. All other existing methods like those measuring collective extinction or scattering from an aerosol probe require information on particle size, shape and composition to derive the true number concentration. Section 4 discusses some of the major indoor and outdoor ultrafine particle studies. The following table lists the CPC instruments used in those studies and the lower limit of detection for the particles (for detailed specification see TSI official web-site Wilson Environmental Associates Page 49

54 Table 3. Instruments Used in Recent Ultrafine Particle Studies. Brand Model Lowest Study Reference Particle Size TSI TSI nm Diapouli et al. 2007; Vinzents et al. 2005; Matson 2005; Balasubramanian 2007; Westerdahl et al. 2005; Morawska 2007 TSI TSI nm Wallace 2002; Wallace 2004; Long et al TSI TSI 3022A 7 nm Zhu et al. 2002; Zhu et al. 2005; Westerdahl et al.2005; Abt et al. 2000; Hussein et al. 2004; Pekkanen 2004; Geller et al. 2005; Kuhnnext et al TSI TSI nm Afshari et al. 2005; Zhu et al. 2005; Matson 2005; Valente et al. 2005; Levy et al. 2002; The most common unit used was the TSI model 3022A as shown in the photo below. Wilson Environmental Associates Page 50

55 TSI considers TSI 3022A model no longer represent the latest technology, and new line (377X) of CPCs are manufactured based on this model. For example, the Model 3776 Ultrafine Condensation Particle Counter (UCPC) is designed primarily for researchers interested in airborne particles smaller than 20 nm. It is a continuous-flow instrument that detects particles down to 2.5 nm (photo below). Size Measurement Impactors are devices primarily used for determination of size dependent particle composition. Multistage impactors (Willeke and Baron 1993) consist of several stages, each composed of an orifice and an impaction plate opposite to it. Between orifice and plate the flow performs a 90 bend. Due to their inertia, large particles follow the flow less well than small ones. Particles above a certain aerodynamic size impinge on the plate, where they stick. The nozzle-plate geometry and the gas pressure at each particular stage define the cut-off diameter. Thus, particles are separated according to their aerodynamic diameters. The smallest diameters separable with commercial impactors are some tens of nanometers. In order to obtain the particle size distribution, the masses of the deposits of each stage must be determined gravimetrically. Impactors have been developed, which use quartz microbalances to give a direct measure of the deposited mass on each stage, and thus deliver a discrete approximation of the aerodynamic size distribution directly (Chuan 1976). These devices suffer from the restrictions concerning mass loading and poor response to agglomerated particles. Wilson Environmental Associates Page 51

56 Besides impactors, so-called aerodynamic sizing instruments or aerodynamic particle counters (APC) yield the distribution of particle sizes. In these instruments, the aerosol sample is accelerated in the flow through an accelerating orifice. The aerodynamic size of a particle determines its rate of acceleration, with larger particles accelerating more slowly due to increased inertia. As particles exit the nozzle, they cross through two partially overlapping laser beams in the detection area. Light is scattered as each particle crosses through the overlapping beams. Part of the scattered light is collected into a photo detector, which converts the light pulses into electrical pulses. The time between these is related to their velocity and the aerodynamic diameter is calculated for each particle detected. The smallest particle diameter to be measured this way is typically around 500 nm. The TSI Model 3320 is an example of such an instrument (TSI 2000a). It also measures the scattering peak height, which provides additional information related to the composition of the particle. The mobility equivalent diameter distribution is measured in a so-called differential mobility analyzer (DMA). The particles must be electrically charged to apply this technique. The aerosol is subjected to an electric field, usually perpendicular to the laminar flow. Due to the resulting electrostatic force on the particles, they are deflected. The deflection depends on their size, and particles of a certain size land at a slit, where they are separated from the rest of the aerosol. The particle size extracted depends on the electric field. Scanning the field and measuring the number concentration of the extracted particles can determine the size distribution of an outdoor aerosol determined quasi continuously. A variety of DMA models are on the market. Among others, they are sold as measurement systems called the Differential Mobility Particle Sizers (DMPS) and the Scanning Mobility Particle Sizers (SMPS) (TSI 2000a). They contain a diffusion charger, a DMA and condensation particle counter. The recorded function is automatically converted to the particle size distribution by a suitable algorithm. The difference between the DMPS and SMPS systems consists of the time required for recording of a mobility distribution. This takes less than a minute with an SMPS system and several minutes with a DMPS system, which provides more precise size distribution data. The inlets are usually equipped with impactors with a 1000 nm cut-off. This defines a well-defined maximum size, which is required for data reduction. Different data reduction algorithms have been applied in these systems, leading to significantly different results. The size distributions obtained with them Wilson Environmental Associates Page 52

57 must therefore still be regarded as approximations. When using several sizing instruments in a field study, these must be inter-calibrated. The largest particle size measured by commercial systems is 1000 nm. The diameters obtained by mobility analysis are called mobility equivalent diameters. Optical sizing instruments use the fact that the light scattered by a particle under a given radiation depends on its size. The smallest particle sizes measurable are around 100 nm. While the mobility equivalent diameter only depends on the particle geometry, and the aerodynamic diameter depends on geometry plus density, any optical diameter depends very much on the composition of the particle in addition to its geometry. Forward scattering instruments minimize this dependence, but their sensitive size range is restricted to particles larger than 1000 nm. Mobility analysis is restricted to particles smaller than 1000 nm with the charging techniques applied today. For the large particle end of the spectrum, aerodynamic methods are used. The aerodynamic particle sizer measures from 500 nm upwards and thus has an overlap with the mobility analyzers on the market. Optical sizing is used as an alternative to cover the range from approximately 300 nm and up.. There are presently no national or international standards for particle sizing instruments. Absolute calibration is usually based on comparison with electron micrographs of deposited particles. Absolute precision of the particle size is of minor importance in epidemiological studies, whereas the agreement of the performance of the instruments is crucial. Intercomparison studies are therefore essential. Section 4 discusses some of the major indoor and outdoor ultrafine particle studies. The following table lists the APC, DMA, SMPS instruments used in those studies and the lower limit of detection for the particles (for detailed specification see TSI official web-site Wilson Environmental Associates Page 53

58 Table 4. Particle Size Instruments Used in Recent Studies. Instrument Make Model Lower Detection Size Study Reference APC TSI TSI nm Wallace 2002; Wallace 2004; Long et al. 2000; Abt et al. 2000; DMA TSI TSI nm Spencer et al TSI TSI nm Geller et al. 2005; Kuhnnext et al SMPS TSI TSI nm Wallace 2002; Cass et al. 2000; Long et al. 2000; Brouwer 2004; Derwent 2000 TSI TSI nm Long et al. 2000; Abt et al. 2000; Dennekamp et al MIE PDR nm Wallace et al. 2003; Olson 2006; Dasgupta et al TSI TSI nm Zhu et al. 2002; Zhu et al. 2005; Kuhnnext et al. 2005; Herner et a l Wilson Environmental Associates Page 54

59 The most common units used were the APC TSI model 3220, DMA TSI model 3080 and SMPS TSI model 3936 as respectively shown in the photos below: Particle Surface Related Measures The classical way of measuring the joint surface of particles is via gas adsorption, which is usually referred to as the BET method. The amount of inert gas (usually nitrogen) adsorbed to a cooled sample is measured as a surface proportional quantity. A typical particle mass of 1 g as a powder is required for these methods. If the particles are sampled from the air with a high volume sampler at 1.5 m 3 /min, it would require 50 days to obtain 1 g, if the air contains 10 µg/m3 of particulate matter. This example demonstrates that conventional gas adsorption methods are far from capable of delivering data with a time resolution as required for most indoor or outdoor studies. Woo et al. (2001) have demonstrated that from measurements of number and mass, an estimate of the total surface can be derived in addition to mass and number, as well as an approximate size distribution. It must be kept in mind that the value of this total surface is an equivalent one based on unrealistic assumptions such as spherical particle shape. Particle-Bound Water Interference A major fraction of the particles in outdoor and indoor air are hygroscopic, and their size can be strongly dependent on the relative humidity (RH). If samples are needed to be size- Wilson Environmental Associates Page 55