Los Angeles summer midday particulate carbon ... - ACS Publications


Los Angeles summer midday particulate carbon...

0 downloads 21 Views 761KB Size

Environ. Sci. Technol. 1991, 25, 1788-1793

Los Angeles Summer Midday Particulate Carbon: Primary and Secondary Aerosol Barbara J. Turpln**t and James J. Huntzicker

Department of Environmental Science and Engineering, Oregon Graduate Institute of Science and Technology,* 19600 NW Von Neumann Drive, Beaverton, Oregon 97006 Susan M. Larsod and Glen R. Cass

Environmental Engineering Science Department and Environmental Quality Laboratory, 138-78California Institute of Technology, Pasadena, California 91 125 Aerosol sampling during photochemically active times across the Los Angeles Basin has provided evidence of secondary formation of organic aerosol from gas-phase precursors a t midday. Ambient organic carbon/elemental carbon ratios exceeded the estimated ratio of organic carbon/elemental carbon in primary source emissions on most sampling days a t all sites. The concentration of secondary organic aerosol was calculated by using ambient data and estimates of the organic carbon/elemental carbon ratio in primary source emissions. Nonparametric sign correlations comparing calculated secondary organic carbon concentrations with tracers of both primary and secondary aerosols supported the method used to quantify secondary organic carbon. Secondary organic aerosol appears to have contributed roughly half of the organic aerosol in Pasadena during midday summer conditions.

Introduction Carbonaceous materials account for a significant fraction of urban and rural aerosols ( I , 2). They include carcinogenic compounds such as polycyclic aromatic hydrocarbons (3, 4 ) and contribute to visibility reduction ( 5 ) . The complexity of carbonaceous aerosols has prompted many investigators to categorize them into organic and elemental classes. Elemental carbon (EC), also called black or graphitic carbon, is predominantly formed through combustion processes and is emitted into the atmosphere in particulate form. For this reason it is a good tracer for primary carbonaceous aerosol of combustion origin. Organic aerosol can be emitted directly in particulate form (primary aerosol) or formed in the atmosphere from products of gas-phase photochemical reactions (secondary aerosol). The formation of organic aerosols through solar irradiation of gaseous mixtures containing specific organic compounds has been observed in smog chamber experiments (6-10). However, the extent of secondary organic aerosol formation in the atmosphere is not well understood and is highly disputed (9, 11-14). It is important to understand the contributions of primary and secondary sources to the ambient aerosol in order to develop effective air quality control strategies. Several investigators have used the ratio of the ambient concentrations of particulate organic carbon (OC) to elemental carbon (OC/EC) or a related ratio to investigate the extent of secondary formation (2,11,15). In such an approach ambient OC/EC ratios greater than those observed for the primary aerosol are considered indicative Current address: Department of Mechanical Engineering, University of Minnesota, 111 Church St. SE, Minneapolis, MN 55455. *Formerly Oregon Graduate Center. 8 Current address: Department of Civil Engineering, University of Illinois, 3230 Newmark Civil Engineering Laboratory, 205 North Mathews Ave., Urbana, IL 61801. 1788

Environ. Sci. Technol., Vol. 25, No. 10, 1991

of secondary formation. However, an estimate of the primary OC/EC ratio is needed to proceed with an analysis of this type. OC/EC emissions ratios vary considerably from source to source, and therefore the primary OC/EC ratio will be influenced by meteorology, diurnal and seasonal fluctuations in emissions, and the influence of local sources. For this reason it is only possible a t this point to determine the range in which the primary ratio is likely to fall, and this range will be selected by use of data from previous studies. Atmospheric sampling conducted in downtown Los Angeles, CA, during winter, peak morning traffic periods (16) yielded a mean OC/EC ratio of 1.7. Because these samples were collected in a source region a t a time when the residence time of the particulate material in the atmosphere was short and little photochemical activity occurred, this OC/EC ratio represents a primary ratio for those source conditions. A comprehensive inventory of primary emissions in the Los Angeles Basin (17) provided a second estimate, yielding a mean OC/EC ratio of 2.4. A monitoring program in which 24-h ambient samples were collected over the course of 1 year in the Los Angeles Basin resulted in mean OC/EC ratios of 1.4 for Lennox, 1.6 for downtown Los Angeles, 1.8 for Pasadena, and 1.8 for Upland (18). Little seasonal dependence was seen in the ratios observed a t these sites, suggesting that primary organic aerosol predominated in the 24-h average OC/EC ratios. Thus, ambient data and emissions information indicate that the primary OC/EC ratio in the Los Angeles Basin should fall in the range of 1.4-2.4. This range of values will be used to characterize primary emissions in the present study.

Experimental Section The data used in this investigation were taken during a sampling program designed to investigate summer, midday low-visibility events in the Los Angeles area (19). This sampling program examined the Los Angeles Basin aerosol a t times when photochemical activity was a t a maximum. Four-hour filter samples were collected at five Southern California locations (Lennox, Pasadena, Azusa, Upland, and San Bernardino) from 1000 to 1400 hours Pacific standard time (PST) every 6 days from July 5 to September 27, 1984. Lennox is near the ocean on the western edge of the basin. Pasadena, Azusa, Upland, and San Bernardino are along the northern edge of the basin from west to east as shown in Figure 1. A cyclone separator a t each sampler inlet removed particles with aerodynamic diameters larger than 2.0 pm, and fine aerosol particles were collected on quartz fiber filters. The system operated a t a flow rate of 10.1 L/min and a filter face velocity of 17 cm/s. Quartz fiber filters were heat treated for at least 1h a t 600 "C prior to use. Following collection, the filters were placed in air-tight Petri dishes lined with heat-cleaned aluminum foil and stored a t approximately -10 "C until analysis. Nuclepore and Teflon filter samples

0013-936X/91/0925-1788$02.50/0

0 1991 American Chemical Society

N

Table I. Site-Specific Ozone and Secondary Organic Carbon D a t a

Sen Gabriel Mountains

E+.

S

Pasadena

e

Sanla Monica

Mountains

8

San Bernardino Upland

70of days secondary OC higher than intersite a?

time of peak ozone (PST)

8 89 50 50 5

1230 1220 1320 1420 1500

Los Angeles

7

Y' ,

8

L.ennox

PACIFIC OCEAN

Lennox Pasadena Azusa Upland San Bernardino

a Percentage of sampling days where the secondary OC estimate a t that site is higher than the average secondary OC estimate computed over all five sampling sites.

Figure 1. Schematic of the Los Angeies Basin showing 1984 field study sites.

were taken concurrently and analyzed for sulfates and nitrates by ion chromatography (Nuclepore filters) and for trace elements by X-ray fluorescence (Teflon filters). Quartz fiber filter samples were analyzed for organic and elemental carbon by use of the Oregon Graduate Institute (OGI) thermal-optical carbon analyzer. The OGI carbon analyzer uses a thermal-optical analysis technique (20-22) that explicitly corrects for the pyrolytic conversion of organic to elemental carbon (i.e., charring) that occurs during the organic analysis. Instrument blanks, field blanks, and external sucrose standards were analyzed daily during sample analysis. The blanks yielded detection limits (3a) of 0.4 and 0.2 pg of C/m3 for OC and EC, respectively. Replicate analysis of 17% of the sample set provided the data for the determination of uncertainties. The 95% confidence interval was 0.4 pg of C/m3 for EC and 0.8 pg of C/m3 for OC. Because this experiment was conducted before the significance of the organic vapor adsorption artifact (22-24) was understood, the necessary measurements to correct for this artifact were not made. Thus, the particulate organic measurements in this paper do not include such a correction. As a result of this artifact, filter-collected organic carbon includes both particulate organic carbon and organic vapors adsorbed onto the quartz fiber filter substrate. The amount of adsorbed organic vapor on the filter (as pg of C/cm2) depends on the filter face velocity, the concentration of adsorbable vapors, and the sampling duration. From results obtained in the 1986 Carbon Species Methods Comparison Study (22, 25) in the Los Angeles Basin, it can be estimated that on average as much as 60% of the organic carbon measured on the filter samples was organic vapor adsorbed onto the filter substrate. Because ambient OC/EC ratios as measured in this research are compared with primary OC/EC ratios determined under different conditions, it is important to assure consistency between the different sets of measurements. T o accomplish this, the results of Gray et al. ( I @ , which were used t o estimate primary OC/EC ratio, were adjusted to the conditions of the current measurements by use of the results of the 1986 Carbon Species Methods Comparison Study (22,25). This increased the OC/EC ratios of Gray et al. as follows: Lennox, 2.1; downtown Los Angeles, 2.4; Pasadena, 2.7; Upland, 2.7. The OC/EC ratios for Lennox and Los Angeles, where primary organic aerosol is expected to dominate, are still within the 1.4-2.4 range. The ratios for Pasadena and Upland, which are receptors of both primary and secondary components from local and upwind sources, are just slightly outside this range. Thus, although the effect of the vapor adsorption artifact cannot be precisely specified, the use of the 1.4-2.4 range for the primary OC/EC ratio

should be qualitatively correct. More detailed, time-resolved results obtained during the Southern California Air Quality Study (SCAQS) in 1987 support the use of this range (26).

Results In the sampling program of Gray et al. (18) 24-h samples were taken a t Lennox. A very high correlation was observed between OC and EC ( R = 0.99), and the ratio of total carbon to elemental carbon showed little seasonal dependence. This indicated that secondary formation did not have a significant effect on the Lennox aerosol when averaged over 24-h periods. In contrast, the Pasadena 4-h midday samples analyzed in the present study showed a much lower correlation between OC and EC ( R = 0.78), and OC correlated nearly as well with ozone ( R = 0.76). The secondary component of the organic aerosol can be estimated from eqs 1 and 2, where OC,, is secondary OC, oc,,, = oc,, - OC,,i (1) OC,,i = EC(OC/EC),,i

(2)

OCw is measured OC, OC, is primary OC, EC is measured EC, and (OC/EC)p,i is the primary ratio estimate. Table I shows the average time of peak ozone concentration for each site over the period of this study. The ozone peak for Pasadena was near the middle of the sampling period. Because filter samples were only taken from 1000 to 1400 PST, it is not possible to determine the exact time when secondary OC peaks occurred. However, insight into the temporal relationship between ozone and secondary OC maxima is provided by the following reasoning. Midday secondary OC concentrations in Pasadena exceeded the average midday secondary OC concentration computed over all five sampling sites on 89% of the sampling days. Moving downwind through Azusa, Upland, and San Bernardino the number of above-average secondary OC concentrations decreased, and the time of peak ozone concentration increased from 1220 to 1500 PST. Data taken during the Southern California Air Quality Study suggest a lag time of 1.5 f 1.0 h between the ozone maximum and the secondary organic aerosol maximum (22,26,27). This suggests that the secondary OC peak had not yet arrived during much of the temporal sampling window a t the most eastern sites. Within the 1000 t o 1400 PST window, the highest ozone concentrations were found in Azusa whereas the highest secondary OC concentrations were found further to the west in Pasadena. Therefore, the Pasadena site will be the focus of this study because both ozone and secondary OC peaks should be included within the sampling period. The use of eqs 1and 2 would be supported if episodes identified with secondary organic aerosol formation corresponded with elevated concentrations of other photoEnviron. Sci. Technoi., Voi. 25, No. 10, 1991

1789

e

I

0

UJ

e

e

e

0.

3 ' -3

1

I

-2

-1

0

1

2

3

-3

-2

-1

LEAD (SNV) 3

0

1

2

3

LEAD (SNV)

I

-3

-3

-2

-1

0

1

2

3

NITRATE (SNV)

-3

-2

-1

0

1

OZONE (SNV)

2

3

3

2

-

1

0

1

2

3

OZONE (SNV)

Flgure 2. Comparison of standard normal variates (SNVs) of (a) secondary organic carbon (OC) and elemental carbon (EC) concentrations, (b) secondary OC and lead concentrations, (c) EC and lead concentrations, (d) secondary OC and nitrate concentrations, (e) secondary OC and ozone concentrations, and (f) total OC and ozone concentrations at Pasadena, CA, July 5-September 27, 1984. Standard normal varlates in (a) and (b) are not expected to be positively correlated. A data set presented in the form of standard normal variates has zero mean and a standard deviation of 1.

chemical reaction products. Nitrate, like secondary OC, is a photochemically generated secondary aerosol component, and ozone is a gas-phase product of photochemical activity. The chemical and dynamical processes involved in secondary aerosol formation are quite complex, and it is unlikely that linear expressions can adequately describe the relationships between such photochemical products as ozone, nitrate, and secondary OC. Therefore, a simple sign test might be more appropriate than linear regression techniques to test for relationships between primary and secondary OC and other photochemically generated pollutants. In this test variables are expressed in terms of their standard normal variates ( S N V ) ,and correlations are sought between the signs of those values. The standard normal variate is defined as SNV = (X- X , , ) / U (3) where X is the concentration (pg/m3), X,, is the average concentration a t that site, and cr is the standard deviation of the concentrations at that site. A data set presented in the form of standard normal variates has a mean of zero and a standard deviation of 1. The purpose of a nonparametric sign test (28) is to determine whether or not the signs of two variables are positively or negatively correlated a t a certain level of significance. If (al, a2,..., a,) and (bl, b2, ..., b,) represent two sample sets, A and B, between which there is no sign correlation, then roughly half of the products of the SNVs of ai and bi (SNV(ai) X SNV(bi))will be positive. More precisely, the fraction of products that is positive will follow a binomial distribution. The level of significance of an observed number of positive products is the probability of observing a t least that many positive products in a sample set that follows a binomial distribution. If A and B are positively correlated, then ai and bi are usually either both greater than average or both 1790

Environ. Sci. Technoi., Voi. 25, No. 10, 1991

less than average. If A and B are negatively correlated, then an ai that is greater than average is usually accompanied by a bj that is less than average and vice versa. The concentrations of all compounds can be expected to vary together because of the common affects of local meteorology such as mixing depth, and compounds emitted or formed in the same vicinity will correlate due to common transport. In addition to the correlation resulting from meteorological factors, primary aerosol components emitted from the same source will correlate due to common emission. Secondary components should also correlate because of the general commonality of the chemical processes leading to their formation. Because the later two sources of correlation are of interest here, it is useful to compare the sign correlations of primary OC and primary and secondary tracers with the sign correlations of secondary OC and primary and secondary tracers. Although the primary OC/EC ratio undoubtedly varied to some extent from day to day, primary and secondary OC were determined by using a primary OC/EC ratio of 1.9, which is the midpoint of the primary OC/EC range. Table I1 shows the fraction of products (SNV(ai) X S N V ( b J )that is positive for several pairs of variables. The sign correlation between primary OC and lead, a primary aerosol component, was positive and significant at the 98% confidence level. Parts a-c of Figure 2 show the standard normal variates of Pasadena secondary OC and EC, secondary OC and lead, and EC and lead, respectively. At the 91% confidence level a negative sign correlation was observed between secondary OC and lead, while secondary OC and EC (and thus primary OC) were uncorrelated. A positive sign correlation significant at the 98% confidence level was observed between SNVs of the two primary tracers, lead and EC. A strong correlation between lead and elemental carbon is expected in an area like Pasadena

5

17

11

JULY

+ -

23

29

4

+ -

10

16

-

22

AUGUST

28

3

9

15

-SEPTEMBER

21

27

Flgure 3. Ambient ratios of organic carbon/elemental carbon (OC/EC) for particles under 2.0 pm in diameter at Pasadena, CA, July &September 27, 1984. Shown with estimates of the primary OClEC ratio (1.4 and 2.4). Error bars are 95% confidence limits.

Table 11. Fraction of Positive Products of SNVs”

OC,, OCpri

OCwc 03

NO3Pb

EC

OC,,

0Cpri OC,,,

1.00

0.87 1.00

0.60 0.47 1.00

O3

NO3-

Pb

EC

0.87 0.80 0.67 1.00

0.67 0.67 0.80 0.73 1.00

0.73 0.73 0.33 0.67 0.40 1.00

0.87 1.00

0.47 0.80 0.67 0.73 1.00

aEach entry shows the fraction of the products, SNV(ai) X SNV(bi),that is positive where aiand bj are values of the variables in the table. No correlation exists when the fraction of products that is positive is 0.5. The level of significance of a positive sign correlation is the probability of observing a greater or equal fraction of positive products in a sample set that follows a binomial distribution. Fractions greater than 0.71 are significant a t 95% confidence intervals.

where both lead and EC are primarily from motor vehicles. A positive sign correlation significant a t the 99.690 confidence level was observed between secondary OC and nitrate (Figure 2d). Nitrate and secondary OC are both secondary aerosol components formed through photochemical processes. In contrast, the total OC-nitrate sign correlation was significant only a t a 915% confidence level. Taken together, these observations support the usefulness of eqs 1 and 2 as a means of estimating primary and secondary OC. In particular, (1)a high sign correlation exists between secondary OC and a photochemically generated secondary aerosol component (nitrate), (2) a high sign correlation exists between primary OC and a primary aerosol tracer (lead), and (3) there are poor sign correlations between secondary OC and primary tracers. Statistically similar results were obtained by use of Pearson product correlations with 95% confidence levels. Despite the evidence supporting the use of eqs 1 and 2 for quantifying secondary OC, good correlations between secondary OC and ozone were not observed. Figure 2e shows the SNVs of secondary OC and ozone, and Figure 2f shows the SNVs of total OC and ozone. Elevated secondary OC concentrations were accompanied by elevated

ozone concentrations on all days but one, August 28. However, the sign correlation between secondary OC and ozone was less significant (91%) than the sign correlation between total OC and ozone (99.9%). Some insight as to why a good secondary OC-ozone correlation may not be observed is provided in the following discussion. The midday samples taken in this study represented average concentrations present in an air mass transported through Pasadena between 1000 and 1400 PST. If an observer traveling with such an air mass for a full day were to take measurements of pollutant evolution within the air parcel, concentrations of reactive organic compounds and concentrations of ozone would rise and fall, but the concentration of secondary aerosol would continue to accumulate, the only major “losses” being deposition and dilution. The rate of secondary organic aerosol formation in smog chamber irradiations of olefin-NO, mixtures has been observed to be approximately proportional to the product of ozone and olefin concentrations (29). Likewise the rate of secondary formation in the air mass might be expected to correlate with ozone concentrations, but the concentration of secondary organic aerosol would depend on the O,, NO,, OH, and organic precursor concentrations integrated over the history of the air mass. This history could not be adequately represented by midday Pasadena ozone concentrations unless emissions and transport of the air mass were similar on all sampling days. Thus, a good correlation between ozone concentrations and the rate of secondary organic aerosol formation a t a particular time is somewhat likely, but a good correlation between ozone and secondary organic aerosol concentrations is not to be expected. The higher correlation between ozone and total OC, which has a significant primary component, is probably due to the common transport experienced by primary OC and ozone and its precursors. The importance of secondary OC during the midday period is demonstrated by the fact that OC/EC ratios were significantly greater than the primary ratio a t 95% confidence levels on most sampling days a t all sites. Figure 3 shows the OC/EC ratio for Pasadena; the range in which the primary ratio falls (1.4-2.4) is shown for comparison. Environ. Sci. Technol., Vol. 25, No. 10, 1991

1791

a E

:

TOTAL OC PRIMARY OC (2.4) PRIMARY OC (1.4)

M

A M P -

=L

v

z

s!-

'0

a a Iz w

0 z

0

5

0

0

- -5

11

17

23

29

4

JULY

10

16

--

22

AUGUST

28

3

9

15

21

SEPTEMBER

27

Figure 4. Total organic carbon (OC) concentrations and primary OC concentrations (pg of C/m3) for particles under 2.0 pm in diameter at Pasadena, CA, July 5-September 27, 1984. Primary OC (2.4) and primary OC (1.4) were estimated by using primary organic carbon/elemental carbon (OC/EC) ratios of 2.4 and 1.4, respectively. Error bars are 95% confidence limits (f0.8,0.9, and 0.5 pg of C/m3 for total OC, primary OC (2.4), and primary OC (1.4), respectively).

The conclusion that secondary OC is important in Pasadena is relatively insensitive to the primary OC/EC ratio. For example, on 9 of the 15 days studied the OC/EC ratio was greater than 4.0 a t a 95% confidence level. Plots of OC/EC for Lennox, Azusa, Upland, and San Bernardino are presented as supplementary material. OC/EC ratios a t these sites ranged from 1.5 to 13.2, with the largest values observed a t the eastern sites of Upland and San Bernardino. Each site has days in which secondary organic aerosol is important during the 1000-1400 (PDT) period. Total organic carbon concentrations for Pasadena are shown in Figure 4 with primary organic carbon concentrations calculated by use of primary ratios of 1.4 and 2.4. (Note that neither the total nor the primary organic carbon concentrations have been corrected for the vapor adsorption artifact. As noted above, this artifact is estimated to comprise on average 60% of the total organic carbon in Pasadena.) Secondary organic carbon concentrations can be obtained by subtracting the primary OC concentration from the total OC concentration. The identical plots for the other four sites are presented as supplementary material in the microfiche edition of this journal. These results indicate that a significant amount of secondary organic aerosol formation took place on most days, particularly in Pasadena, Upland, and San Bernardino. Site averages are presented in Table 111. Each entry is presented as a range of values to take into account the uncertainty in the primary OC/EC ratio. In a previous study Larson e t al. (5) accounted for background OC and EC levels estimated from 1982 San Nicolas Island samples (5, 18) in their determination of secondary OC in the Los Angeles Basin. A second set of estimates, which includes this background subtraction (background OC = 1.4 pg of C/m3 and EC = 0.16 pg of C/m3), has been included in Table 111. By either method, secondary organic aerosol concentrations are a substantial contributor to total organic aerosol concentrations. Large secondary contributions were seen in Pasadena on August 28 and San Bernardino, the easternmost site, on Septem1782

Environ. Sci. Technol., Vol. 25, No. 10, 1991

Table 111. Secondary Organic Carbon Contributions

Lennox Pasadena Azusa Upland San Bernardino

6.3 12.1 13.5 10.8 8.0

1.7 2.5 4.1 2.8 1.4

32-63 50-74 30-59 37-65 62-75

16-48 41-58 22-52 28-56 50-62

O1 Values are averaged over all sampling days for the time period 1000-1400 PST. Calculations assume a primary OC/EC ratio of 1.4-2.4. No corrections have been made for the vapor adsorption artifact. Secondary organic carbon contributions after backaround estimated by Larson et al. (5) has been subtracted.

ber 15. The contributions of primary and secondary sources vary considerably from day to day. August 10 in Lennox was a day when apparently all the organic aerosol observed during the midday period was of primary origin. Under midday, summer conditions, secondary formation appears to be responsible for roughly half of the organic aerosol in Pasadena, compared to less than 29% of the organic aerosol observed on an annual average basis (2).

Summary Aerosol sampling during photochemically active times across the Los Angeles Basin has provided evidence of secondary formation of organic aerosol in the ambient atmosphere. Ratios of ambient OC/EC exceeded the primary ratio estimates on most sampling days at all sites, and it was possible to estimate the secondary OC component on the basis of ambient data and primary OC/EC ratios. The performance of calculated secondary OC concentrations in nonparametric sign correlations with tracers of both primary and secondary aerosols supported the means for quantifying secondary OC adopted in this paper (eqs 1 and 2). Secondary organic aerosol so defined appears to have contributed roughly half of the organic aerosol in Pasadena

during midday summer conditions. A contribution of this magnitude to midday aerosol concentrations can be expected to have a significant impact on midday visibility and human exposure. Additional research is needed to evaluate the variability of the primary OC/EC ratio and to explore the time dependence of carbonaceous aerosol concentrations. Acknowledgm#ents

Field sites were located a t South Coast Air Quality Management District monitoring stations. The advice and assistance of Dr. Stephen McDow is greatly appreciated. Supplementary Material Available

Plots of the 1984 summer midday OC/EC ratios and total organic aerosol concentrations indicating the concentrations that can be attributed to primary aerosol based on primary OC/EC ratio estimates of 1.4 and 2.4 for the Lennox, Azusa, Upland, and San Bernardino sites (8 pages) will appear following these pages in the microfilm edition of this volume of the journal. Photocopies of the supplementary material from this paper or microfiche (105 X 148 mm, 24X reduction, negatives) may be obtained from Microforms Office, American Chemical Society, 1155 16th St., N.W., Washington, DC 20036. Full bibliographiccitation Gournal, title of article, authors’ names, inclusive pagination, volume number, and issue number) and prepayment, check or money order for $17.50 for photocopy ($19.50 foreign) or $10.00 for microfiche ($11.00 foreign), are required. Registry No. OC, 7440-44-0; Os, 10028-15-6; Pb, 7439-92-1. Literature Cited (1) Shah, J. J.; Johnson, R. L.; Heyerdahl, E. K.; Huntzicker, J. J. J . Air Pollut. Control Assoc. 1986, 36, 254. (2) Gray, H. A; Caw, G. R.; Huntzicker, J. J.; Heyerdahl, E. K.; Rau, J. A. Sei. Total Environ. 1984, 36, 17. (3) Hoffman, D.; Wynder, E. L. In Air Pollution; Volume IZ Analysis, Monitoring and Surveying; Stern, A. C., Ed.; (4) (5)

Academic Press: New York, 1968; pp 187-247. Daisy, J. M. Ann. N . Y . Acad. Sci. 1980, 388, 50. Larson, S. M.; Cas, G. R.; Gray, H. A. Aerosol Sci. Technol.

(6)

Stern, J. E.; Flagan, R. C.; Grosjean, D.; Seinfeld, J. H.

(7)

Environ. Sci. Technol. 1987, 21, 1224. McMurry, P. H.; Grosjean, D. Atmos. Environ. 1985,19,

1989, 10, 118.

1445.

( 8 ) Heisler, S. L.; Friedlander, S. K. Atmos. Environ. 1977, 11, 157. (9)

O’Brien, R. J.; Holmes, J. R.; Bockian, A. H. Enuiron. Sci.

(12)

Grosjean, D.; Friedlander, S. K. J. Air Pollut. Control Assoc. 1975,25, 1038.

Knights, R. L.; Cronn, D. R.; Crittenden, A. L. Presented at the 1975 Pittsburgh Conferenceon Analytical Chemistry and Applied Spectroscopy, Cleveland, OH, March 3, 1975; Paper No. 3. (14) Chu, L.-C.; Macias, E. S. In Atmospheric Aerosol: SourcelAir Quality Relationships; Macias, E. S., Hopke, P. K., Eds.; ACS Symposium Series 167; American Chemical Society: Washington, DC, 1981; 251. (15) Wolff, G. T.; Countess, R. J.; Groblicki, P. J.; Ferman, M. A.; Cadle, S. H.; Muhlbaier, J. L. Atmos. Enuiron. 1981, (13)

15, 2485. (16)

Conklin, M. H.; Cass, G. R.; Chu, L.-C.; Macias, E. S. In Atmospheric Aerosol: Source/ Air Quality Relationships; Macias, E. S., Hopke, P. K., Eds.; ACS Symposium Series

167; American Chemical Society: Washington, DC, 1981; pp 235-250. (17) Gray, H. A. EQL Report No. 23; Environmental Quality Laboratory, California Institute of Technology, Pasadena, CA, 1986; pp 103-108. (18) Gray, H. A.; Cass, G. R.; Huntzicker, J. J.; Heyerdahl, E. K.; Rau, J. A. Environ. Sci. Technol. 1986, 20, 580. (19) Larson, S. M.; Cass, G. R. Enuiron. Sci. Technol. 1989,23, 281.

Johnson, R. L.; Shah, J. J.; Cary, R. A.; Huntzicker, J. J. In Atmospheric Aerosol: SourcelAir Quality Relationships; Macias, E. s.,Hopke, P. K., Eds.; ACS Symposium Series 167; American Chemical Society: Washington, DC, 1981; pp 223-233. (21) Huntzicker, J. J.; Johnson, R. L.; Shah, J. J.; Cary, R. A. In Particulate Carbon: Atmospheric Life Cycle;Wolff, G. T., Klimisch, R. L., Eds.; Plenum: New York, 1982; pp (20)

79-88.

Turpin, B. J. Ph.D. Dissertation, Oregon Graduate Center, Beaverton, OR, 1989. (23) McDow, S. R.; Huntzicker, J. J. Atmos. Environ. 1990,24A, (22)

2563.

McDow, S. R. Ph.D. Dissertation, Oregon Graduate Center, Beaverton, OR, 1986. (25) Turpin, B. J.; Huntzicker, J. J. Sampling and Analysis of Organic Aerosol. Final report to the California Air Resources Board, Contract A5-149-32, 1988. (26) Turpin, B. J.;Huntzicker, J. J. Secondary Organic Aerosol in the Los Angeles Basin. Final report to the California Air Resources Board, Contract A-8 32-129, 1991. (27) Turpin, B. J.; Huntzicker, J. J. Atmos. Environ. 1991,25A, (24)

207.

Chatfield, C. Statistics for Technology;Chapman and Halk New York, 1983; pp 157-158. (29) Heisler, S. L. Ph.D. Dissertation, California Institute of Technology, Pasadena, CA, 1976. (28)

Technol. 1975, 9, 568.

Groblicki, P. J.; Nebel, G. J. In Chemical Reactions in Urban Atmospheres; Tuesday, C. S., Ed.; American Elsevier Publishing Co.: New York, 1971; pp 241-263. (11) Novakov, T. In Particulate Carbon: Atmospheric Life Cycle; Wolff, G. T., Klimisch, R. L., Eds.; Plenum Press: New York, 1982; pp 19-41. (10)

Received for review April 22, 1991. Accepted June 6, 1991. Support for this project was provided by the National Aeronautics and Space Administration and the California Air Resources Board, by a grant from the Hewlett Foundation, and by gifts to the Environmental Quality Laboratory at the California Institute of Technology.

Environ. Sci. Technol., Vol. 25, No. 10, 1991

1793