Acidic sulfate aerosols: characterization and exposure.

Exposures to acidic aerosol in the atmosphere are calculated from data reported in the scientific literature. The majority of date was not derived from studies necessarily designed to examine human exposures. Most of the studies were designed to investigate the characteristics of the atmosphere. However, the measurements were useful in defining two potential exposure situations: regional stagnation and transport conditions and local plume impacts. Levels of acidic aerosol in excess of 20 to 40 micrograms/m3 (as H2SO4) have been observed for time durations ranging from 1 to 12 hr. These were associated with high, but not necessarily the highest, atmospheric SO4(2)- levels. Exposures of 100 to 900 micrograms/m3/hr were calculated for the acid events that were monitored. In contrast, earlier London studies indicated that apparent acidity in excess of 100 micrograms/m3 (as H2SO4) was present in the atmosphere, and exposures less than 2000 micrograms/m3/hr were possible. Our present knowledge about the frequency, magnitude, and duration of acidic sulfate aerosol events and episodes is insufficient. Efforts must be made to gather more data, but these should be done in such a way that evaluation of human exposure is the focus of the research. In addition, further data are required on the mechanisms of formation of H2SO4 and on what factors can be used to predict acidic sulfate episodes.


Introduction
Animal and human exposure studies in the laboratory indicate that acidic sulfate particles will produce functional changes in the respiratory tract at levels well below lethal ones (1). Although the bulk of toxicological data on acidic aerosols involves sulfuric acid (H2SO4) or bisulfate (HSOj) particles, the available evidence indicates that the observed irritant responses are likely due to hydrogen ion (H+) rather than to sulfate (2).
The information available on the concentration patterns and human exposure to atmospheric acidic aerosol is meager. Many of the available studies were not designed to measure human exposure but rather were part of a program of research directed at characterizing the ambient atmosphere. Nonetheless, the information is useful for pointing out directions for future work, as well as for estimating the exposure to acidic species for specific conditions. This paper attempts to summarize the data available on acidic aerosol species with special emphasis on acid events. It also includes an analysis of measurement methods used in the reported studies. tion, we have included discussions of these factors. In addition, London and California fog data are discussed to provide historical and regional contexts for the data base of largely eastern United States acidic sulfate aerosol measurements.
Despite the toxicological data, there is no epidemiological evidence clearly linking acidic aerosols acting alone with adverse health effects in humans. The body of available research suggests that there is a likely association, but studies aimed at ascertaining this association have been strapped by limited field data. Moreover, more research is needed to satisfactorily resolve the extent that community health is affected by episodic and chronic acidic aerosol exposures.

Methods for Acidic Aerosol Measurement
Because there is no reference method for acidic sulfate species, a major area of concern is the multitude of techniques that have been used to detect various acidic species in the environment. Measurements are generally made by aerosol collection on filters with subsequent analysis or continuous analysis with a real-time detector. From an exposure assessment standpoint, continuous data offer few advantages to short duration (1-8 hr) filter intervals, although the continuous data allow assembly of any desired integrated-sample schedule. The detection of acidic sulfur species has usually resulted in data being available in one of two forms: the compound H2SO4 or the strong acid (free or titratable), H+. There are no techniques available to exclusively detect the interinediate acid species, ammonium bisulfate, although there have been a few attempts to calculate the ammonium bisulfate concentration in the field (3)(4)(5).
Filter Collection Thermal Volatilization. Thermal volatilization schemes were popular for several years for speciation of acidic sulfate compounds in aerosols. Filter samples were heated and H2SO4 collected by microdiffusion (6,7) or determined directly by flame photometry after volatilization from Teflon filters (8,9). In one method, H2SO4 was distinguished from other volatile sulfates [e.g., ammonium bisulfate (NH4HSO4) and ammonium sulfate ((NH4).)S04)] and nonvolatile sulfates [e.g., sodium sulfate (Na2SO4) by heating consecutively at two different temperatures (10). In another technique, 2-perimidinylammonium sulfate is formed from acidic sulfates and thermally decomposed to SO2 for West-Gaeke analysis (11). Due to recovery problems and limited success in distinguishing NH4HSO4 and (NH4)2SO4 from each other, thermal volatilization from filter samples have fallen into disfavor in recent years. Nonetheless, coupled with a heated denuder system, it is the basis of the continuous H2SO4 measurements using real-time flame photometric detectors (see next section).
Filter Extraction with pH Measurement or Titration. Filter samples may be analyzed for strong acid content by extraction into water or dilute mineral acid. The extracted free acid content may be determined by simple measurement of pH and, in the absence of weak acids, equated to the amount of strong acid originally present in the sample. However, this determination can be in error because of the potential presence of buffering agents such as weak carboxylic acids or hydrated forms of heavy metal ions, e.g., Fe(III) and Al(III) (12,13). A titration procedure for strong acid based on the Gran titration (13) and perfected by Brosset and Ferm (14) has been used widely (15). Coulometric generation of strong base for Gran titrations has also been used [e.g., Tanner et al. (16)1. Dissolution of filter samples in a 0.05 to 0.10 pM mineral acid solution followed by Gran titration with correction for blank allows for titration of micromolar levels of strong acid with precision and accuracy better than ± 10% (16,17). All samples must be protected against ammonia.
Filter Extraction with Specific Analysis of Sulfuric Acid. Most efforts in specific extraction of atmospheric acids has been related to H2SO4. Benzaldehyde has been shown to be specific for H2SO4 in dried acidic aerosol sulfate/nitrate samples with analysis for sulfate in aqueous back-extracts (10). Several methods for derivativization of collected H2SO4 aerosol have been proposed (18,19) but these suffer from interference by sulfate salts. The use of specific extractant methodologies has decreased in recent years in favor of generic strong acid determinations.

Continuous Analysis
H2SO4 may be determined using a continuous flame photometric detector (FPD) with a diffusion denuder tube for SO2 and H2SO4 removal attached before the FPD. The temperature of the denuder tube or a zone just upstream from it is cycled between about 200C and 1200C. At the lower temperature, H2SO4 remains in the aerosol phase, but at 1200C it is volatilized and removed as the air flows through the denuder tube. The difference in response at the two temperatures represents ambient H2S 04 levels (20)(21)(22). The minimum cycle time and, hence, time resolution for the technique, is about 6 to 8 min (23,24). Sensitivity is found to be enhanced by using sulfur hexafluoridedoped hydrogen gas (25).
Acidic sulfates including H2SO4 may also be continuously analyzed using the humidograph techniques of Charlson et al. (26) or thermidograph variations developed more recently (27,28). The latter technique involves heating the aerosol-containing air stream progressively from 200C to 3800C in 5-min cycles, rapidly cooling it to the dry bulb temperature and measuring the light scattering at 65 to 70% relative humidity with a nephelometer. By comparing this output with the thermidograms of test aerosols, the fractional acidity can be measured and approximate level of H2SO4 determined.

Sampling Anomalies
Sampling anomalies have plagued the measurement technologies described previously and generally fall into two classes: reversible or irreversible sorption losses onto filter materials or sampling lines; and equilibrium-driven loss or gain of species due to nonsteady-state conditions in the sampled atmosphere over the time period of the measurement. Sorption Losses. Studies of sorption losses on filters have centered on three major areas. One area is the loss of strong acid in aerosol particles by reaction with basic sites in the filter matrix used to collect the particles (29). Depth-filter matrices used for highvolume samples, in particular glass-fiber filters, are unsuitable for collection of acidic aerosol particles (16,30,31). This is true even if glass fiber filters are pretreated with acid and fired to a high temperature, as subsequent rinsing produces additional free-basis sites in the glass (16). However, high-purity quartz filters can be pretreated to remove basic sites for high-volume samples. Treated quartz and Teflon filter media are generally inert with respect to acidic particles and have replaced glass fiber and cellulose filters for sampling acidic aerosols by high-volume and low-volume techniques, respectively.
Avoiding the use of glass-fiber filters also eliminates a positive source of error in sulfate measurements: The artifact sulfate formed by base-catalyzed oxidation of SO2 sorbed on the filter surface to form H2SO4 (32). While the acid formed on reactive filter surfaces is neutralized, the residual sulfate remains and can lead to a substantial measurement bias.
A third problem is the loss of strong acid contents by reactions between co-collected basic and acidic particles on the filter surface. This most frequently occurs as the result of coarse ( > 2.5 rim), alkaline, soil-derived particles interacting with fine (< 2.5 Mm), acidic sulfate particles (33,34). This problem can be eliminated by using a coarse particle preseparator (e.g., a cyclone or impactor inlet). Alternatively, sampling for shorter time durations, in most ambient environments for 4 hr or less, results in low enough surface coverage to prevent topochemical reactions. However, this does not prevent subsequent neutralization during extraction procedure.
Equilibria-Driven Losses. Equilibria involving gas-and particulate-phase species may lead to errors in sampling strong acids', as well as other species in the atmosphere, both under steady-state and nonsteadystate conditions. Thermodynamic considerations suggest that aerosol sulfate/H2SO4 mixtures should be in equilibrium with atmospheric ammonia (NH3) (35). NH3 is the principal neutralizing agent H2SO4, and the equilibrium level of NH3 for even slightly acidic sulfate is much below average ambient NH3 levels (36). Hence, acidic aerosol on filter membranes are subject to neutralization if NH3 ammonia-laden air is sampled later in the same interval. Another equilibrium-driven filter sampling artifact is the coupled volatility of NH3 and nitric acid (HNO3) when ammonium nitrate (NH4NO3) is collected with acidic sulfate aerosol (37). Denuders are used to scavenge neutralizing species.

Regional Patterns of Precursor Emissions and Aerosol Concentrations
Oxides of Sulfur and Nitrogen For the eastern United States (EUS), the distribution of SO2 emissions are highest in the area within the Ohio Valley from the Mississippi River through southwestern Pennsylvania (Fig. 1A). Other locales with significant emissions that surround this core area include rural and urban centers. In contrast, the NO, emissions are centered in the major urban areas of metropolitan New York-New Jersey-Connecticut, Chicago, and Detroit (Fig. 1B). However, it can be seen that the distribution of NO, is more ubiquitous throughout the EUS. In the western half of the nation, SO2 emissions are centered around more isolated, large stationary sources such as smelters. The NO, emission in the west are distributed much more broadly. The actual emission densities tend to be much lower than in the east, except in the southern and central California urban centers.

Sulfate
While the toxicological data suggest that the SO2is not the biologically active portion of particulate sulfate species (38), it has been used as an indicator pollutant in many health studies. In addition, SO4 is relatively easy to measure and is a major constituent in the ambient aerosol. Therefore, many studies and monitoring programs have routinely provided data on the SO2ion, and a large data base exists.
However, the presence of high sulfate does not necessarily indicate the presence of a high acidic component in the sulfate aerosol. Knowledge of its distribution, nonetheless, gives some indication of the possible spatial extent of acidic sulfate in the worst possible case (i.e., little neutraiization). Figure 2 shows the monthly average SO 2 concentrations measured in the EUS during the sulfate regional experiment (SURE) project (39). The peak values recorded in SURE reached in excess of 25 ,ug/m3 during the summer. Lesser peak values were recorded during the winter with the seasonally highest concentrations shifting to the southeast. The data also indicate that a substantial portion of the EUS experienced average sulfate above 8 jig/rn3. results for each season are shown in Figure 3. The findings indicate that the ratios for summer were much higher than for the winter. The ratios were characteristically lower in the northern part of the SURE study area then in the southern and the coastal sections. These differences were consistent with the expectation of both greater rates of SO2 oxidation and greater losses by dry deposition of SO2 during the warmer months and in warmer regions.

Formation of Acidic Sulfates
The mechanisms involved with the transformation of SO2 were reviewed in the Air Quality Criteria for Fi(1tJRI. 3. Geographical distribution of the ratio of sulfate sulfur to total airborne sulfur for different seasonal periods (%). From Mueller et al. (39).
Particulate Matter and Sulfur Oxides (40) and are summarized in Table 1. Included are both heterogeneous and homogeneous reactions that produce H2SO4 in the gas phase or within the aqueous phase. Additional potential mechanisms involve metal catalysts, carbon, and surfaces. The predicted atmospheric conversion rates range from < 0.5% to > 100% per hour. These reactions do not necessarily occur at the same time. Each will be associated with particular environmental conditions such as near a source, in a plume, or in the general atmosphere. Some of these conditions are mutually exclusive. The importance of particular reaction schemes will be dependent upon insulation and humidity, hence, the season, geographic location, the time of day (night or daylight), and the local or prevailing meteorology.

Ammonia
A key factor in determining the persistence of atmosphere acidic sulfate species is the potential for atmospheric ammonia neutralization. This will be controlled by the reaction of NH3 with H2SO4 and the formation of ammonium salts [NH4H S O4 and (NH4)2SO4]. The major sources of environmental ammonia appear to be animals and humans, although significant contributions can come from manufacture of fertilizer. For both animals and humans the emissions are a result of excretion through the skin or from urea. Animals are probably a much larger source of ammonia due to the sewage treatment of human excrement. A regional mapping of domestic animal (cattle and pigs) and human ammonium emissions by Husar (41) yielded quite different distributions. Animal ammonium emissions centered in the midwest, and human emissions were keyed to the major population centers. Husar and Holloway (42) also added the emissions inventory for fertilizer production and a combined emissions estimate for ammonium (NH+) is shown in Figure 4. These estimated emissions suggested that ammonia neutralization will be dependent upon the land areas over which a sulfuric acid-ladened air mass travels and the contact time of air mass with the surface.
There are very limited data for ambient ammonia concentrations in the northeastern U.S. An extensive bComments: a: typical range for daytime at northern midlatitudes during the summer; b: this reaction rate is not well established; c: assumed that liquid water volume of aerosol = 50 x 10-12m3/m3, {SO2 } = 10 ppb; d: assumed that Mn(II) mass concentration = 20 ng/m3, also, the Mn(II) is assumed to be uniformly dissolved in the liquid water of the aerosol [Mn(II) = 8.9 x 10 -3M]. Rate calculation used the expression of Neytzell-de Wilde and Taverner (78); e: assumed that Fe(III) mass concentration = 2 1g/m3, also, the Fe(III) is assumed to be uniformly dissolved in the liquid water of the aerosol [Fe(III) = 0.9 Ml. Rate calculation used the expression of Neytzell-de Wilde and Taverner (78);f: assumed that C mass concentration = 10 Ag/iM3 and behaves as the soots studied by Chang et al. (79), whose expression was used for this calculation; g: rate calculation was based on Eqs.   (40); h: rate calculation was based on Eqs.   (40); i: influence of inhibitors has been ignored, but they are likely to suppress the rate by orders of magnitude. CE denotes exponent to the base 10 (e.g., 3E -1 = 3 x 10 20 monitoring program conducted by the Dutch (43) showed that the range in monthly averages for NH3 reached from 0.2 to 35 Pg/m3. These concentrations are sufficient to fully neutralize from 0.6 to 100 jig/m3 of H2SO4 aerosol. This represents an upper range to be expected in the U.S., since the density of livestock breeding is far greater in the Netherlands.
Data on aerosol NH' are generally linked to measurements of sulfate. The SURE study included aerosol measurements of the two species for nine sites in the EUS over a 2-year period (39). The ratios of SO2f /NH4+ in total suspended particulate (TSP) samples were calculated for all sites and indicated that 40% of the samples were neutralized ammonium salts. The remaining 60% of the samples could have contained unneutralized acidic sulfate. A scatter diagram of the ratios for the SURE data is shown in Figure 5. usually measured during the intense portions of a photochemical smog episode. It is likely that the fraction of acidic-to-total sulfate is lower because the air parcels associated with the episodes have greater opportunity to come in contact with local ammonia emissions. However, these data are likely to underestimate the potential for acidic sulfate. The caveats regarding neutralizing capacity of the coarse fraction described in the methods section applies here to the TSP samples of the SURE study. Nonetheless, other studies have reported similar findings, i.e., the highest H+/SO2ratios do not occur at the highest sulfate concentrations (3,(44)(45)(46). Morandi et al. (3) examined the distribution of acidic species for a regional smog episode that occurred in the summer of 1980. For measurements made at Tuxedo, NY, the most acidic portion of the smog episode they examined occurred at its beginning, and the acidity decreased with time. This is in contrast with the reported results of a study at Fairview Lake, NJ, in  (4). It was observed during a 3-day episode that the acidic aerosol was primarily partially neutralized sulfate (i.e., NH4HSO4) with no H2S 04 measured. However, during the last day of that period, H2SO4 increased from approximately 0 to 4 ,ug/m3 for 4 hr in the afternoon. These different phenomena must be examined more fully to understand the dynamics of the atmospheric environment when acidic species could be present.
Particle Size of Acidic Sulfate Aerosol The sulfate particles are generally found in the fine particle size range ( > 2.5 Mm). The SURE results in the EUS showed that on the average, 80% of the S02is in the fine fraction with a range of 50 to 100% (39).
Size distribution data have been acquired for So4and H+ by Pierson et al. (44) and Keeler (45) (47). The size distribution in Figure 8A shows most of the mass centered around 0.5 gim; Friedlander has classified this as a coastal sulfur aerosol, and it reflects the influence of coastal humidity. In Figure 8B, the sulfur mass mode is between 0.1 and 0.2 ,um. This has been classified as desert-type sulfur aerosol, which appeared to be formed from homogeneous reaction mechanisms and does not grow appreciably in the arid atmosphere. Recently, measurements of the distribution of H+ and S02as a function of particle size were made in Claremont, CA, where H+ was found only in the range < 1.5 Mm (48).

Record of Acidic Aerosol Concentrations and Exposures
As stated previously, the available data are based on direct measurements of H2SO4 or determinations of free or titratable acidity. To place the results from these studies in context, the concentrations are expressed as jig/M3 (as H2SO4), even when only H+ measurements were made. These values are probably an overestimate of the actual H2SO4 concentration since in both urban and rural areas acidic sulfate particles are usually in a partially neutralized form. However, the data are readily converted to the same scale of strong acid (e.g., nano equivalents/m3).

London Historical Data
A relationship between air pollution and mortality/morbidity was recognized in England, especially after the severe London fog episode of December 5-8, 1952. Although routine air pollution monitoring conducted by local authorities and other regulatory bodies did not have data for H2SO4, it was conjectured that H2SO4 was considered to be one of the pollutants potentially responsible for the increased mortality and morbidity (49). Apparent H2SO4 measurements were

Los Angeles Historical Data
In 1949, a study was conducted in Los Angeles, CA to determine the amount of free H2SO4 in the atmosphere during periods of intense fog and clear days (52). There were 2 days with levels above 150 ,ug/m3, and each occurred on days with high relative humidity. The extreme in acidity was corroborated by an independent observation of lesions on leaves during the same period (53). A short-term study conducted in Los Angeles in the winter of 1961 recorded a peak concentration of 60 pg/m3 (54).

Acid Fog
Fog and clouds are a special type of atmosphere aerosol (55). While aerosols are generally comprised of crystalline, aqueted salt, or other solid particles, fogs and clouds are the suspension of liquid water droplets in the air. From condensation nuclei of 0.1 to 1.0 pm, droplets grow to diameters between 2 to 100 Im, although the majority of droplet mass occurs in the range of 5 to 30 gim. The presence of nuclei, composed of both soluble and nonsoluble material, is essential to the formation of atmospheric water droplets.
Fog droplets are highly effective at scavenging pollutant materials present in the air. The overall fraction incorporated into fog droplets depends upon two processes: nucleation scavenging (i.e., activation) of aerosol and gas dissolution. The speciation of pollutant components present prior to fog formation is therefore important. The pollutant species of concern are often present as hygroscopic aerosol (e.g., sulfates and nitrates). Furthermore, in situ chemical transformations, such as oxidation of SO2 to H2S 04, alter this speciation and the effectiveness of fog scavenging while the droplet phase is present (56).
The study of fog has traditionally remained in the domain of atmospheric physicists, who were principally concerned with its effect on visibility or the mechanisms of cloud formation. Yet, the data of early investigations of fog demonstrate that fogwater can be highly concentrated with respect to a variety of chemical components ( Table 2). The high solute concentrations and extremes in acidity found in fogwater are reason for concern. There has been a historical correlation of fog with the most severe air pollution episodes (57). Identification of a link between urban fog events and human injury was made even before detailed measurements of fog composition (58).   (92) aFog type: I, intercepted stratiform cloud; M, marine or coastal fog; R, radiation fog. bNumber of samples or events: (a) average; (m) median; (vw) volume-weighted mean of n samples; (av) average and/or range of r events. CNonprecipitating stratiforn cloud data only.

Atmospheric Acidic Sulfate Studies from 1974-1986
Measurement studies of acidic sulfate species in the U.S. and Canada since 1974 have been tabulated with references (Table 3). A variety of sampling strategies and analytical methodologies were used, and the individual sampling times ranged from 1 to 24 hr, although none used glass fiber filters. Most of the studies were conducted during the summer when large scale and prolonged regional 03 and sulfate smog episodes can occur in the eastern U.S. and Canada (59). In addition, photochemical smog episodes are most frequent during the summer in Los Angeles, CA (60), and Houston, TX (61). The magnitude of the wintertime levels of acidic sulfate species were derived from a more limited data base.
The range of concentrations recorded in the studies outlined in Table 3 is shown in Table 4. It is apparent that a wide range of SO4 and acid sulfate levels were encountered. The peak H2SO4 value, measured with a flame photometric detector (FPD), was 41 jg/m3 (1-hr average)  An H+ concentration of 39 ,ug/m3 (as H2S 04) was  (64). Important features of the Glasgow data were: (a) that the H+ was measured in the submicrometer size range; (b) the measurements were 12-hr duration samples; and (c) the visibility maps indicated the development of an urban plume from St. Louis. The measured levels of aerosol NH+ on July 29, 1975, indicate that the aerosol in Glasgow, IL, contained both NH4HSO4 and H2SO4. A number of other studies also indicated the presence of both species (3,64,65).
Morandi et al. (3) included an attempt to actually infer the distribution of acidic species from coincident instrumental FPD measurements of H2SO4 and filter analyses for H+. From their data, there was no evidence of the presence of strong acids (pKa < 2) other than NH4HSO4, or H2SO4. Their results also showed that the species H2SO4, NH4HSO4, and (NH4)2SO4 can occur simultaneously. Thurston and Waldman (5) showed that although the sulfate aerosol measured in Toronto was as high as 30% acidic at times, the measured H2SO4 was less than 2% of the total sulfate composition.
For most of the summertime studies, peaks of H2SO4 and/or H+ were associated with the presence of a slowmoving high pressure system. For example, a major acidic sulfate event occurred in the EUS during early August 1977 (66). The daily variation of the 6-hour particulate sulfate and H+ concentrations and the daily maximum 03 concentrations are shown in Figure 10. Samples for this period measured at High Point, NJ, and Brookhaven, NY (separated by 160 km), showed peak excursions of H+ at the two sites with the passage of pollutant-laden air masses (66).
Samples collected in New York City for the same period were reported by Tanner et al. (34). The trajectory analysis indicated that air parcels passed from High Point to New York City or Brookhaven (i.e., southeasterly). However, no aerosol acidity was observed in New York City on any of the sampling days, even when acidity was recorded in High Point.
This would indicate that the flux of NH3 emanating from the metropolitan area was sufficient to neutralize the acid-laden air parcels. Values of H+ at Brookhaven (80 km to the east of New York City) were consistently lower than the values reported at High Point. This suggests that there was partial neutralization and/or fresh acid production over the metropolitan area. In 1984, Spengler et al. (67) indicated H2SO4 studies at two of the Harvard Six-Cities Health Study sites: Harriman, TN (coal-burning area), and Watertown, MA (suburb of Boston). Using the continuous FPD system for H2SO4 measurements, they classified acid events. Those that occurred in the coal burning area of Harriman, TN, were of longer duration and higher in concentrations than observed at Watertown, MA. In another investigation by Harvard, an H2SO4 concentration of 41 ,ug/m3 was observed in St. Louis (62). In this case wintertime conditions were studied, and the results suggested an association with a power plant plume.
In a July 1979 study at Lennox, CA, a 4-hr maximum of 11 ,ug/m3 of H2SO4 was observed (68). This was apparently associated with the daytime oxidation of local SO2 emissions. Only three other acid events were measured, although elevated levels of sulfate were measured throughout the week-long study. A later study in Los Angeles found lower concentrations of acidity (69). Of the two sites monitored, the downwind location, east of Los Angeles, recorded the higher acid concentrations. showed the presence of acidic species in nonurban areas of the EUS. Cadle (72) reported the results of H+ made at a suburban site in Warren, MI, from June 1981 through June 1982. The most frequent excursions occurred during the summer, with six 24-hr samples exceeding 4 gg/m3. However, the highest H2S O4 concentration Cadle reported occurred during the wintertime. Rural investigations, one in Whiteface Mt., NY (73), and the other in Nova Scotia, Canada (74), focused on acidic sulfate aerosolfi In each case there was at least one period during which the acid concentration exceeded 9 ,ug/m3.

Studies in the
To date, there have been only two studies in which the acidic sulfate species were measured simultaneously at several nearby (less than 50 km) sites (46,63). A study in Toronto, Ontario, was conducted during the summer of 1986 using three sites within the metropolitan area (46). Separate filters were run during the day (9:00 A.M.. to 5:00 P.M.) and night (5:00 P.M. to 7:00 A.M.). During several episodes, the nighttime period was broken further by sampling from 7:00 P.M. to 9:00 P.M. (evening) and from 9:00 P.M. to 9:00 A.M. The peak concentrations for the major species (H+, NH+, and SO2-) occurred on the July 25-26 sampling periods.
While the 24-hr values for sulfate were quite uniform at the three sites (34, 34, and 35 .g/m3), H+ concentrations were 9.4, 8.3, and 5.9 ,ug/m3 H+ (as H2SO4) (Fig. 11), respectively. The highest concentrations occurred during the brief evening period on July 25 (broken out of the nighttime period at sites 1 and 2 only); these were 48 and 75 jg/m3 so42 and 14.4 and 19.4 ,ug/m3 H+ (as H2SO4). During this episode, the degree of sulfate neutralization was between 70 and 75% at sites 1 and 2, which were located north 30 km northeast and northwest of downtown. However, for samples collected at the third site, located in the urban downtown area, the sulfate aerosol samples were 80 to 85% neutralized.

Acidic Sulfate Exposure and Events
The ambient atmosphere studies measured a wide range of acidic sulfate concentrations in the atmosphere. Most of the H2SO4 or H+ (as H2SO4) values were below 5 ,ug/m3, although events with higher values occurred occasionally (at least once) over the course of each study. It is not appropriate, based on the available data, to equate an acid event with a sulfate episode, since periods of high acid concentrations do not necessarily coincide with periods with the highest So2-, e.g., photochemical haze or smog. Therefore, the definition of an acidic sulfate event is somewhat arbitrary. For this discussion, a pollution event for acidic sulfate will be defined as an event when the measurement of H2SO4 or H+ (as H2SO4) reached levels 5 jg/m3 for at least 1 hr.
With the preceding criterion, some of the studies listed in Table 4  may be of some biological significance since it has been shown in a controlled human study that the effects of exposures as low as 100 jig/m3 H2SO4 for 2 hr enhance the effects of mucociliary clearance observed for 1-hr exposures (75). For the purposes of the following discussion, however, the exposure values (,ug/m3)-hr should be interpreted as the maximum potential exposure during an event, and not as the biologically effective dose to an individual. It should be emphasized that the exposure calculation (concentration x time) will only be applied for determination of exposures that can occur during an event., For example, annual average calculations of exposure are to be avoided.
The exposure and event results shown in Table 5 and are divided into representative studies with sample collection times of 24 hr, 12 hr, 6 hr, 1 hr, and a combined sampling time. This was done because the length of reported events is different for the various collection times. For the 24-hr studies (Table 5), the events lasted a maximum of 24 hr, which was the sample duration. This may be an underestimate of the length of the event since the portion of the next or preceding 24-hr period may have been above 5 jg/m3 and could not have been detected. The exposures for the 24-hr episodes ranged from 120 to 336 (.ug/m3)-hr and were isolated instances of high acidic sulfate.
The acidic sulfate events for the 12-hr samples reached 36 hr in duration and recorded peak exposures of 510 and 925 (jig/m3)-hr. An interesting feature of the 12-hr values ( Table 6) was that for studies conducted for a period of 1 month or less, each had at least one event with acidic sulfate exposure above 100 (gg/m3)-hr. Obviously, the frequency of acid events in each location cannot be described since longer duration studies would be required.
The two examples of 6-hr samples (  (66). In the period from August 5 through 12, there were almost daily exposures to acidic sulfate above 100 (,ug/m3)-hr. All studies reporting 1-hr sample intervals used direct measurements of H2SO4 by FPD, and the exposure results are found in Table 8. The duration of episodes ranged from 1 to 20 hr. For St. Louis, there were only two instances during which exposures to H2SO4 were above 100 (,ug/m3)-hr and three above 80 (,ug/m3)-hr. This is puzzling since the sampling included an entire year, and there are major sources of SO2 in the area (76). However, the work of Ellestad (77) and the previously cited work of Tanner and Marlow (64) may explain the lack of frequent high excursions in H2S 04. Particle light scattering was measured during the 1975 MISST study at Glasgow, IL, 17 km downwind of St. Louis urban and power plant plumes (77). Those data indicated aerosol production during the summer. H+ measurements made in summer 1975 at the same location indicated very high exposures [> 900 (,ig/m3)-hr] (64). Thus, the main acid impact from the urban plumes are likely to be more sub-  stantial at downwind locations. In the future, studies literature. The majority of data was not derived from will require careful consideration of topographical and studies necessarily designed to examine human expometeorological data to determine where major plume sures. Most of the studies were designed to investigate impacts will occur and their potential for producing the characteristics of the atmosphere. However, the high acid concentrations and exposures.
measurements were useful in defining two potential In the Toronto study (46), the exposure analyses exposure situations: regional stagnation and transport completed for the three sites showed that there was one conditions and local plume impacts. period during which exposures at two of the three sites Levels of acidic aerosol in excess of 20 to 40 jig/m3 (as were above 100 (,ug/m3)-hr (Table 9). At the third site, H2SO4) have been observed for time durations rangthe acid levels were much lower, which was probably ing from 1 to 12 hr. These were associated with high, due to local neutralization. Thus, across this metro-but not necessarily the highest, atmospheric SO2politan area of 2.5 million people (and probably other levels. Exposures of 100 to 900 (jg/m3)-hr were areas, too), the potential for local differences in acidic calculated for the acid events that were monitored. In sulfate must be examined in order to define the contrast earlier London studies indicated that acidity locations of maximum potential exposures. in excess of 100 jIg/m3 (as H2SO4) were present in the atmosphere, and exposures > 2000 (,ug/m3)-hr were Su m mary possible.
It is apparent that at least two types of conditions High exposures to acidic aerosol in the atmosphere mentioned can lead to a number of possible situations were calculated from data reported in the scientific for epidemiological and human exposure studies. The   aOnly periods where concentrations exceeded 5 pg/M3 H2SO4 were considered episodes. first would be studies conducted during major summertime haze episodes, where large populations in the eastern U.S. could be affected periodically by high acidic sulfate levels. The exposure could be manifested by high H2SO4 and/or NH4HSO4 accumulated over periods of 1 hr or more throughout a day or sequence of days. Because these episodes occur in the summer, large segments of the population will be participating in outdoor activities in the rural and suburban areas. These individuals would be most at risk during regional episode and would be the focus of opportunities for epidemiological studies.
The second type of exposure would be confined to places downwind of a power plant plume or urban plume during any season of the year. Obviously, the greatest potential for population exposure would still be the summer; however, exposures could still occur during all of the other seasons. The latter would be confounded by variable levels of outdoor activity and degree of penetration of the acidic sulfate into the home. It should be noted that the highest acid concentrations observed in the cited studies appear to be associated with direct plume impacts. Therefore, these types of locations warrant considerable attention in future epidemiological studies.
Our present knowledge about the frequency, magnitude, and duration of acidic sulfate aerosol events and episodes is insufficient. Efforts must be made to gather more data, but these should be done in such a way that evaluation of human exposure is the focus of the research not just to establish a routine monitoring network. In addition, further data are required on the mechanisms of formation of H2S 04 and on what factors can be used to predict acidic sulfate episodes.
We are very grateful for the exhaustive literature searches conducted by Timothy Buckley, a doctoral student in the Exposure Assessment joint curriculum of Rutgers University Department of