Project Background

In July of 1997, the U.S. Environmental Protection Agency (EPA) revised the National Ambient Air Quality Standards (NAAQS) to address ambient air concentrations of particulate matter (PM) with an aerodynamic diameter of 2.5 micrometers or less (PM_2.5). Concern for the elevated concentrations of PM derives mainly from persistent evidence of human health effects associated with atmospheric particles. The epidemiological literature has more than 100 published papers, which, for the most part, support association of PM with increases in morbidity and/or mortality. However, no epidemiological study has identified particular PM chemical constituents as potential causative agents. The PM_2.5 standard promulgated by the EPA defines PM_2.5 mass measured according the Federal Reference Method (FRM PM_2.5) as the indicator for health effects. In 1997, the EPA also proposed regional haze regulations for pristine areas; fine particulate is the single greatest contributor to visibility impairment in these areas. Based on the limited existing data of ambient PM_2.5 concentrations, it appears likely that many areas of the country may exceed these new standards.

Large gaps exist in our understanding of the nature of, effects of, and control of ambient PM. These gaps are due to the fact that PM is a complex mixture of multicomponent particles whose size distribution, composition, and morphology can vary significantly in space and time. Atmospheric aerosol size ranges from a few nanometers to tens of micrometers. Major components include sulfate, nitrate, ammonium, and hydrogen ions; trace elements (including toxic and transition metals); organic material; elemental carbon (or soot); and crustal components. PM is emitted directly from sources such as diesel engines and is also formed in the atmosphere from gaseous precursors. The development of effective control strategies requires a better understanding of the link between PM_2.5 and the observed health effects. A number of aerosol characteristics (Table 1) have been proposed as the potential causes of the observed health effects.

Table 1. Possible PM Characteristics Resulting in Health Effects

PM Characteristics	Symbol (Units)
Total Number Concentration	N (cm-3)
Number Concentration of particles larger than x nm	PNx (cm-3)
Total Surface Area	S (µm2 cm-3)
Surface Area of particles between diameters x and y	Sxy (µm2 cm-3)
Particle mass less than x, 2.5, or 10 µm	PMx , PM2.5 , PM10 (µg m-3)
Metal PM 2.5 or PM10 Concentration (Fe, Mn, etc.)	MI,2.5 MI,10 (ng m-3)
Sulfate PM2.5 and PM10	Sulf2.5, Sulf10 (µg m-3)
Nitrate PM 2.5 and PM10	Nit2.5, Nit10 (µg m-3)
OC PM 2.5 and PM10	OC2.5, OC10 (µg C m-3)
EC PM 2.5 and PM10	EC2.5, EC10 (µg C m-3)
Acidity	H+ (ng m-3)
Bioaerosol number concentration	B (cm-3)
Polar, Non-polar organics	POC, NPOC (µg m-3)
Hydrogen peroxide, Organic peroxides	H2O2,RO2 (ppb)
Total soluble PM2.5 and PM10	SPM2.5, SPM10 (µg m-3)
Specific Sources (diesel or gasoline combustion, power plants, etc.)	Sourcei (µg m-3)
Gas-phase co-pollutants	CO, O3, NO, NOx, SO2 (ppb)

The link between ambient PM and health effects is further complicated because of the large uncertainty in the relationship between ambient PM concentrations and actual human exposure (NRC, 1998). Indoor measurements are critical for assessing exposures because individuals spend, on average, about 90% of the time indoors (70% in homes). Significant uncertainties exist in the penetration of outdoor aerosol to the indoor environment. If exposures are not adequately characterized, then causal relationships between ambient PM and health effects may be erroneously attributed. Understanding the relationship between the outdoor and indoor PM concentrations in a typical house in the area will be one of the objectives of the Pittsburgh Supersite Program.

Identifying the PM characteristics and/or sources that cause negative health effects requires a comprehensive data set of PM measurements and epidemiological data. Such a data set does not exist. Standard PM measurement techniques such as the Federal Reference Method for PM_2.5 mass (FRM PM_2.5) do not measure most of the aforementioned characteristics. Indicative of the difficulty of the PM measurements is the often-observed lack of agreement between the measured PM concentration and the sum of the concentrations of the individual PM components in the Eastern US. The unexplained mass can be as much as 30% of the total and has been speculated to be the result of errors in estimating the organic aerosol mass from the measured organic carbon, residual water, errors in estimating the dust contribution, or the unlikely existence of some unidentified major component (Andrews et al., 1999).

The difficulty and cost of PM measurements have impaired our ability to characterize their temporal and spatial variability, understand the processes that control their formation and removal, and quantify the exposure of populations to them. Overcoming these difficulties requires the evaluation of existing PM measurement methodologies and development of new technologies which will allow the cost-effective, accurate measurement of PM characteristics. The proposed Pittsburgh Supersite program has been designed to foster this measurement evolution process. Examples of emerging technologies that will be evaluated include semi-continuous metal measurements by GFAA and LIBS, single particle composition and size measurements by time of flight mass spectrometry, continuous nitrate, sulfate and carbon measurements by integrated collection and vaporization cell, continuous polar organic aerosol measurements by single particle mass spectrometry, semi-continuous OC and EC measurements, enhanced organic aerosol speciation by GC/MS, ultrafine aerosol size distribution measurements, semivolatile organic aerosol partitioning measurements, bioaerosols, continuous direct surface area measurements, etc.

The state-of-knowledge of organic PM illustrates the need for advanced instrumentation. Carbonaceous compounds comprise 20-70% of the dry fine particle mass in both urban and rural areas. Although more than 100 individual organic compounds have identified (Rogge et al. 1991, 1993a-e, 1994, 1997a,b, 1998, 1999a-d; Simoneit et al., 1993, 1998, 1999; Fraser et al., 1998; Schauer et al., 1999a,b), only about 20% of the organic fine particle mass has been identified on a molecular level. In addition, few measurements exist of the ambient concentrations of organic precursors that lead to secondary organic aerosol. Improved understanding of the organic PM phase requires further development of in-situ measurement techniques such as single particle mass spectrometry and improved analytical methods for examining filter samples. The proposed study using a number of new techniques will also address some of the most important questions regarding organic PM in the study area. What is the primary versus secondary organic PM fraction? What is the contribution of biogenic PM sources (primary aerosol or oxidation of terpenes and sesquiterpenes) to organic PM? How do the semivolatile organic PM components partition between the gas and aerosol phases?

Design of effective State Implementation Plans (SIPs) and other regulatory policies requires knowledge of source-receptor relationships that link ambient PM_2.5 levels with emissions. The design of these strategies is complicated by the importance of secondary aerosol to PM_2.5. Of particular concern are non-linearities between emissions and ambient PM levels. For example, reductions in SO₂ and sulfate can free the associated ammonia in aerosol ammonium sulfate. This additional ammonia can react with available nitric acid vapor forming aerosol ammonium nitrate (Seinfeld and Pandis, 1998). In extreme cases in specific areas in the NE US controls in SO₂ could result in a small increase in PM concentrations during the winter (West et al., 1999). Similar non-linearities are expected in the response of the PM concentrations to NO_x and ammonia emission changes. Recent developments in our understanding of the partitioning of semi-volatile organic aerosol components between the gas and aerosol phases suggest that the organic aerosol system is often non-linear (Odum et al., 1996). The proposed study will provide quantitative information about the possible non-linearity of the source-receptor relationships in the study area, especially for secondary particulate matter and also determine if there are any reactants controlling the secondary PM formation (e.g., ammonia).

Deterministic modeling with appropriate input data allows detailed examination of the contribution of different sources to secondary aerosol formation. Several of these models (including two models by the Carnegie Mellon and University of Delaware teams) are at various stages of development (see Seigneur et al., 1998 for a review). However, the evaluation and application of these models in the Eastern US is limited by the lack of suitable measurements. The Supersites together with the proposed EPA Speciation site network could provide the much needed information for the evaluation of these models.

Additional understanding of atmospheric aerosol chemistry and physics is needed to develop the next generation of deterministic air pollution models and establish the source-receptor relationships for secondary PM. For example, little is known about the relative importance of fine PM production and removal in fogs and low clouds in polluted urban areas of the eastern US such as Pittsburgh. The importance of nucleation as a source of ultrafine particles is unknown. Our lack of understanding of the interactions of ambient PM with water limits our ability to estimate their atmospheric lifetimes and transport distances. These gaps in understanding make the estimation of the local versus regional contributions to the PM levels in an urban area like Pittsburgh very uncertain. The proposed study includes a number of innovative approaches (e.g., single particle mass spectrometry coupled with Tandem Differential Mobility Analysis, continuous measurements) which will shed light on the above issues.