Measurements of Atmospheric Carboxylic Acids and Carbonyl

Jun 22, 2001 - Air quality impacts of increased use of ethanol under the United States' Energy Independence and Security Act. Rich Cook , Sharon Phill...
1 downloads 12 Views 189KB Size
Environ. Sci. Technol. 2001, 35, 3071-3081

Measurements of Atmospheric Carboxylic Acids and Carbonyl Compounds in Sa˜o Paulo City, Brazil L. MONTERO,† P. C. VASCONCELLOS,‡ S. R. SOUZA,† M. A. F. PIRES,‡ O. R. SA ´ NCHEZ-CCOYLLO,§ M. F. ANDRADE,§ AND L . R . F . C A R V A L H O * ,† Instituto de Quı´mica, Universidade de Sa˜o Paulo, Cx.P. 26077, 05599-970 Sa˜o Paulo, SP, Brazil, Instituto de Pesquisas Energe´ticas e Nucleares, Sa˜o Paulo, SP, Brazil, and Instituto Astronoˆmico e Geofı´sico, Universidade de Sa˜o Paulo, Sa˜o Paulo, SP, Brazil

Winter atmospheric measurements of gaseous lower carbonyl and carboxylic acids were carried out simultaneously (in 1999) at two distinct urban sites located in the city of Sa˜o Paulo, Brazil. The greater metropolitan area of Sa˜o Paulo is the largest industrialized region of Latin America and has a highly polluted atmosphere. It has an unconventional mix of vehicle types in that a variety of gasoline blends, including oxygenated ones, are used. Mixing ratios of formic and acetic acids ranged, respectively, from 0.6 to 19.4 and from 0.1 to 10.6 ppbv in one of the sites studied and from 1.4 to 18.4 and from 0.4 to 6.7 ppbv in the other site. High values of formic to acetic ratios were found, especially in the latter site (average ) 4.3), suggesting that photochemical production was the predominant source of the formic and acetic acid during the afternoon. Differing from the acids, levels of carbonyls were similar at both sites. Higher average mixing ratios of acetaldehyde and formaldehyde were found in the morning (18.9 and 17.2 ppbv) and gradually decreased from midday (9.5 and 11.8 ppbv) to evening (7.2 and 10.2 ppbv). In the morning, vehicular direct emission seemed to be the main primary source of formaldehyde and acetaldehyde, whereas at midday and evening these compounds appeared to be mainly formed by photochemistry. Secondary photochemical production of organic acids and aldehydes (rather than primary emissions from vehicles) was shown to be more important in Sa˜o Paulo’s atmosphere from midday to evening, particularly on days with strong solar radiation.

Introduction Organic acids and aldehydes are important constituents of the troposphere, for they contribute to a large fraction of the non-methane hydrocarbon mixture (1, 2). These compounds may be emitted directly into the atmosphere by mobile and stationary sources and may also be produced in situ by the photo-oxidation of gas-phase hydrocarbons. As sources of free radicals in the atmosphere and precursors in the * Corresponding author fax: 11-3818-3837; e-mail: lrfdcarv@ iq.usp.br. † Instituto de Quı ´mica. ‡ Instituto de Pesquisas Energe ´ ticas e Nucleares. § Instituto Astrono ˆ mico e Geofı´sico. 10.1021/es001875g CCC: $20.00 Published on Web 06/22/2001

 2001 American Chemical Society

formation of organic aerosols, they are major contributors to urban photochemical smog (3). Low molecular weight carbonyls and carboxylic acids play important roles in photochemical processes that occur in the atmosphere. Ozone-olefin reactions have been shown to produce such organic compounds (4, 5). Aldehydes photolyze and, consequently, are significant sources of OH radicals in the atmosphere (6). Furthermore, the photooxidation of hydrocarbons, either by OH radicals or by addition of ozone to unsaturated hydrocarbons of anthropogenic or biogenic origin, also leads to carbonyl formation (7, 8). The carbonyls may react with free radicals, forming photochemical products, such as CO, CO2, peroxyacetylnitrates, or carboxylic acids (produced either in the gas phase or in water droplets). Like carbonyls, carboxylic acids can also be generated by the ozonolysis of anthropogenic or natural unsaturated hydrocarbons (9, 10). The dominant removal processes for carbonyls are chemical reactions and photolysis, whereas for carboxylic acids they are dry and wet depositions. Due to their low reactivity, gaseous carboxylic acids are slowly removed from the atmosphere by chemical reactions (11). Carboxylic acids are found in the atmosphere in a variety of phases. In particular, gaseous formic and acetic acids are present in significant amounts in the global troposphere, as they are important contributors to the precipitation acidity (11-14). Possible sources for organic acids in the atmosphere and several mechanisms for their production have been supposed, but there is still considerable uncertainty about them (5, 10, 15-17). It is well-known that combustion of fuels and biomass is an important source of organic acids, for example, vehicle exhausts, smoke stacks, and the burning of organic materials (such as leaves, hay, and wood) release these compoundssin particular, acetic acid (18). Biogenic emissions from vegetative sources have been suggested as important natural generators of formic and acetic acids (14, 19). On the other hand, formic and acetic acids are formed in the urban atmosphere by ozone-olefin reactions, but the amount of acetic acid formed is believed to be smaller than that of formic acid (7, 16, 20). Among the atmospheric carbonyls, formaldehyde and acetaldehyde are the most abundant in urban air. In addition to photochemical production, they can be emitted from several mobile and stationary sources (2, 7, 21-23). In most of the measurements carried out in urban areas, formaldehyde levels are generally higher than acetaldehyde levels. However, data from the aldehyde distribution at different urban areas have revealed a considerable variability (0.314.3) (24). Differences in fuel utilization, use of catalytic converters, and so on for each urban area are probably responsible for this large variation. Particularly, HCHO/ CH3CHO values found in Brazilian cities are relatively low due to the use of ethanol-containing fuels (18-20). The combustion of oxygenated fuels in spark-ignition engines results in increased emissions of primary acetaldehyde and formaldehyde, and the reaction of acetaldehyde with OH radical leads to the formation of peroxyacetyl radical, which can react with NO2 to form peroxyacetylnitrate (PAN) (28). A comprehensive survey of atmospheric contaminants from a variety of sources is necessary to help identify the major emission sources responsible for urban air pollution. Although hydrated ethanol has been used in Brazil as an alternative fuel to improve urban air quality, its impact on the urban atmosphere is still not well-known (28, 29). Few studies about the chemical atmosphere of Sa˜o Paulo, a VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3071

FIGURE 1. Evolution of the profile of vehicle types in the metropolitan area of Sa˜o Paulo, Brazil.

FIGURE 2. Topography of the metropolitan area of Sa˜o Paulo. megacity with serious air pollution problems (mainly in winter), have been conducted (14, 30, 31). This paper reports winter measurements of gaseous carboxylic acids and carbonyl compounds carried out in two different urban areas inside the city of Sa˜o Paulo. Transport of these photochemical products and the emission sources are also analyzed. This work is part of a study on photochemical air pollution in Sa˜o Paulo that emphasizes the composition of urban air and the chemical processes among these species that lead to urban ozone formation. The results of this paper, in conjunction with other meteorological and chemical data, will be applied in air quality mathematical 3072

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001

models to mimic the urban chemical environment in the greater Sa˜o Paulo metropolitan area.

Experimental Section Sampling. The greater metropolitan area of the city of Sa˜o Paulo is the largest industrialized region in Latin America. It has an area of 8051 km2 with a population of 16.3 million inhabitants. Currently, there are ∼6.5 million automotive vehicles: 390000 heavy-duty diesels and 5.5 million lightduty (LD) vehicles. Approximately 4.2 million of the lightduty cars are fueled with a mixture containing 78-80% (v/v)

FIGURE 3. Meteorological parameters in Sa˜o Paulo area: (a) the planetary boundary layer (PBL) inversion height (m); (b) profiles of temperature (°C) and wind speed (m/s). gasoline and 20-22% ethanol, which is referred to as gasohol, and 1.1 million are fueled with hydrated ethanol (32). The ratio between the number of ethanol and gasoline cars has changed substantially in the past few years. In 1995, for instance, 97% of the cars produced in Brazil were fueled with gasohol, revealing a substantial reduction in the production of cars fueled with hydrated ethanol, which had reached 76% of national LD vehicle production in 1986. In 1989, nearly half of the LD cars were gasohol-fueled and the other half ethanol-fueled (33). From 1996 to 1999 (see Figure 1), there was a notable increase in the number of cars fueled with gasohol (25%), whereas the number of cars fueled with hydrated ethanol remained at the previous level (32). In Figure 2, it is possible to observe the complex topography of the metropolitan area of Sa˜o Paulo (MASP). Because the MASP is surrounded by mountains ranging from 650 to 1200 m high, dust and atmospheric pollutants are

often trapped in inversion layers, providing ideal conditions for photochemical reactions (33). In this work, two urban sites within the MASP, ∼15 km distant from each other, were chosen on the basis of local differences in the type, distribution, and proximity of emission sources, as well as differences in the wind direction frequencies. The AÄ gua Funda (AF) site, located on the southeastern side of the city, is a large area with abundant vegetation. It is one of the last remaining parts of the Mata Atlaˆntica forest. The area receives minimal impact from local anthropogenic sources. There are no industrial and commercial operations in the immediate vicinity of the site, but at ∼20 km southeast is the largest Latin-America industrial park with several emission sourcessincluding petrochemical and oil refinery areas. The atmospheric air collections in the AF site were performed from a tower at ∼30 m above ground level. The VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3073

circulation may explain the transport of pollutants from the southeast to the northwest region of the MASP during these events. Thermal inversions can frequently occur in relation to a polar mass stagnation over the MASP. Before the onset of frontal systems, the wind blows from the northwest, bringing dry and warmer air from continental areas. This process characterizes the air mass transport from these regions.

TABLE 1. Daily Integration of Total Solar Irradiance day (August)

irradiance (MJ/m2)

day (August)

irradiance (MJ/m2)

01 02 03 04 05 06 07

11.08 12.41 18.25 18.41 17.10 17.93 18.31

08 09 10 11 12 13 14

6.18 15.75 18.36 17.34 16.83 16.63 3.38

Thermal inversion episodes and sunshine conditions were recorded during the sampling period. The mixing layer average height was 1000 m in daytime and 300 m at night. The average variation of the planetary boundary layer (PBL) inversion height obtained from SODAR data at three different sampling periods (morning, midday, and evening) is presented in Figure 3a. From 8:00 to 10:00 a.m., the PBL height is low and photochemical process is not expressive, from 12:00 to 2:00 p.m., pollutant dispersion and photochemical reactions are the dominant processes, and from 4:00 to 6:00 p.m., a decrease in both PBL height and photochemical activity is expected. Meteorological parameters, such as temperature (average ) 20 °C for both sites) and relative humidity (average ) 54% for the AF site and 40% for the CID site) were recorded during the sampling. Wind velocity and temperature during the sampling period are shown in Figure 3b. To evaluate the photochemical activity, the daily integration of the total solar irradiance measured by a Bimetalic Actinograph Fuess 58D during the sampling period is presented in Table 1. Among the measurements recorded, only two of them (August 2 and 8) presented a very low penetration of solar radiation due to the presence of cold fronts, but measurements taken on other dates (showing high levels of solar radiation reaching the surface on clear sky days) characterized this period.

sampling devices were installed in an open area on the tower roof. The Cidade Universita´ria (CID) site, located in the southwestern area of Sa˜o Paulo, can be considered potentially impacted by many different types of sources. The sampling site is ∼2 km far from a major highway with heavy vehicle traffic fueled by gasohol, diesel, and ethanol. At this site, samplers were placed in an open area on the roof of the Department of Atmospheric Sciences, located on the main campus of the University of Sa˜o Paulo, ∼20 m above ground level. Samples were collected simultaneously at both sites during winter 1999. Sampling was performed on consecutive days, from August 2 to 13, with a total of 37 samples and 81% detection. Measurements were carried out in three diurnal periods: morning, 8:00-10:00 a.m. (mo); midday, 12:002:00 p.m. (md); and evening, 4:00-6:00 p.m. (ev). Meteorological Conditions. Sa˜o Paulo presents an upland tropical climate with its dry season during winter. In summer, monthly average temperatures reach 23 °C (from December to February), and in winter, monthly temperatures are ∼16 °C (from June to August). The rainy season normally begins in September and ends in March, with an annual precipitation of ∼1200 mm. The local circulation is given by winds from southeast and northeast, mainly associated with the Atlantic Ocean breeze circulation. During winter, there are often polar mass arrivals associated with cold front systems that can intensify the circulation coming from the southeast. This

Analytical Methods. Gaseous carboxylic acids were collected using a gas diffusion denuder, 6 mm i.d. × 50 cm, coated with a solution of 5% m/v Na2CO3 at an air flow rate of 2.0 L/min. Carbonyls were collected at an air flow rate of 2.0 L/min on a silica gel cartridge (Sep-Pak, Waters) previously cleaned with methanol and acetonitrile and coated with 2,4-

TABLE 2. Performance Characteristics of Measurements for Species Studied

formaldehyde acetaldehyde acetic acid formic acid a

LOD,a µg/L (standard)

blanks, ppbv (SD)b

LOD, ppbv

collection efficiency, % (SD)

precision of measurement, %

overall uncertainty, %

85.3 64.5 18.4 142.5

0.193 (0.0352) 0.149 (0.0171) 0.124 (0.0068) 0.395 (0.0142)

0.185 0.128 0.034 0.356

98 (3.4) 97 (3.9) 97 (2.5) 113 (2.9)

2.1 1.8 1.2 1.9

10.1 9.7 10.7 9.5

LOD, limit of detection.

b

SD, standard deviation.

TABLE 3. Mixing Ratios of Gaseous Carboxylic Acids and Aldehydes during Three Collection Periods: Morning (8:00-10:00 a.m.), Midday (12:00-2:00 p.m.), and Evening (4:00-6:00 p.m.), Sa˜o Paulo, AF (a) and CID (b) Sites, Winter 1999 mixing ratio, ppbv morning range

midday av ( SD

range

acetic acid formic acid acetaldehyde formaldehyde

10.6-1.9 19.0-0.6 50.9-4.2 46.3-1.0

3.1 ( 3.1 5.9 ( 5.4 18.1( 16.3 16.4 ( 14.1

(a) AF Site 5.4-0.1 12.1-1.4 28.4-1.9 28.3-1.2

acetic acid formic acid acetaldehyde formaldehyde

6.4-0.5 18.4-3.1 56.6-2.2 45.6 - 3.6

2.7 ( 2.3 10.1 ( 5.7 19.7 ( 20.0 18.0 ( 17.0

(b) CID Site 4.5-1.0 14.9-5.0 27.5-1.2 28.3-3.3

3074

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001

evening av ( SD

range

av ( SD

3.0 ( 2.0 6.4 ( 5.1 8.7 ( 7.7 10.7 ( 7.5

6.1-0.2 15.1-0.6 22.2-1.4 22.7-2.7

3.1 ( 2.6 6.4 ( 5.9 7.1 ( 7.1 9.8 ( 7.7

2.6 ( 1.3 10.0 ( 4.3 10.5 ( 9.3 13.1 ( 9.7

6.0-0.4 13.4-4.9 37.2-1.6 38.1-3.7

3.2 ( 1.7 10.4 ( 4.3 7.3 ( 9.5 10.7 ( 9.3

FIGURE 4. Mixing ratios (ppbv) of acetic and formic acids in the gas-phase samples: (a) AF site and (b) CID site (Sa˜o Paulo, 1999). dinitrophenylhydrazine (DNPH). An ozone trap, consisting of a KI cartridge, was connected to the upstream end of the cartridge to avoid ozone sampling artifact (34). Gaseous carboxylic acids retained in the denuders were extracted with 20 mL of deionized water. The aqueous extracts were filtered in a Millipore HAWP 0.45 µm membrane and stored in polyethylene flasks in a freezer until analysis (35). For every six samples collected, one blank was analyzed. The analysis was done using a Dionex model 4000i ion chromatographic system (14). Quantification was obtained through external standard comparison. Gaseous aldehydes retained in the DNPH-silica cartridges were eluted with 5 mL of acetonitrile, and the extracts were stored in amber glass vials in a freezer until analysis by HPLC, on a Shimadzu model LC-9A, with UV detection at 360 nm (34). For every six samples, one blank was analyzed. The concentration of carbonyls in air samples was calculated using the external calibration data for carbonyl-DNPH standards. Performance

characteristics of the carboxylic acids and aldehydes measurements are presented in Table 2. Air Parcel Trajectories. Backward trajectories were calculated by the University of Sa˜o Paulo Trajectory Model (USPTM). It is a sigma-z kinematic three-dimensional model coupled to the Regional Atmospheric Modeling System (RAMS), which uses the three wind components (U, V, W) for determination of air parcel trajectories (36). Every 6 min, the latitude, longitude, and altitude coordinates of the air parcels were output, but only trajectory locations every 6 h are presented in this study. A fixed value of the 200 m of the surface was chosen as the arrival level. The USPTM is based on the trajectory equation

db(t)/dt r )V B(b, r t)

(1)

with an initial value of b(t r 0) ) b r0, where b(t) r denotes the positions, in time, and V B the velocity field. VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3075

FIGURE 5. Formic to acetic acid ratios at both sites (Sa˜o Paulo, 1999).

FIGURE 6. Backward trajectories, from 200 m of the surface, on August 6 (a) and 11 (b) at 3:00 p.m. local time (SP ) Sa˜o Paulo City). Heights of the air parcels above mean sea level are shown in the trajectories.

Results and Discussion Carboxylic Acids. The mixing ratio ranges of formic and acetic acids were, respectively, 0.6-19.4 and 0.1-10.6 ppbv at the AF site and 1.4-18.4 and 0.4-6.7 ppbv at the CID site. Formic and acetic acids were found in all samples at concentrations close to those observed previously in other urban regions (7, 20). Table 3 shows the average mixing ratios of the carboxylic acids at three different collection periods for the AF and CID sites. It is interesting to note that the organic acids average level is maintained throughout the day. Although there were in situ formation processes and direct emissions, the mixing ratio of the carboxylic acids was practically constant during the day because of the low chemical reactivity and the slow 3076

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001

dry deposition of these species in the atmosphere. For ∼60% of samples, formic acid (which has average mixing ratios of 10.1 ppbv for the CID samples and 6.2 ppbv for the AF samples) was the most abundant carboxylic acid. In Figure 4, daily profiles of the formic and acetic acids are presented. Formic and acetic acid distributions were very similar, but the formic acid levels were much higher than the acetic acid levels. Formic and acetic acids were strongly correlated at the AF site (morning, r ) 0.92; midday, r ) 0.72; and evening, r ) 0.91), suggesting that these acids are linked together in the vapor phase and their sources are closely related, if not identical (37). At the CID site, a significant good correlation between formic and acetic acids was also observed in the morning (r ) 0.93) and at midday (r ) 0.76),

FIGURE 7. Average mixing ratios of carboxylic acids and carbonyls at both sites: August 6 (a) and 11 (b) (Sa˜o Paulo, 1999). indicating that these species probably arise from similar processes (38). Production of formic acid has been seen to be dominated by in situ formation and the reverse has been true for acetic acidsfor which direct emissions predominate (16). In several studies on measurements of gaseous carboxylic acids, the formic to acetic acid ratio has been used to suggest their sources. The resulting ratio may be taken as an indicator of direct emission (low ratio, 1) (14, 20, 39, 40). It is known that direct emissions from vehicular motor exhaust lead to increased acetic acid ambient levels, whereas direct emissions from vegetation may provide higher levels of formic acid than acetic acid (14, 18, 41). On the other hand, atmospheric ozone can react with olefins from organic

emissions to produce formic acid; high ambient temperature and solar irradiation contribute to the occurrence of this photochemical mechanism (3). Our previous experiments carried out in a tunnel located in the city of Sa˜o Paulo, in which emissions from gasohol vehicles and alcohol motor exhaust were the dominant sources, have shown a formic to acetic acid ratio of 1:4 (35). A ratio of 1:2 was obtained in tunnel studies described by Grosjean et al. (41), indicating a vehicular emission with a different ratio of acetic and formic acids than that observed in the tunnel in Sa˜o Paulo. Variables, such as the numbers and types of vehiculars, fuel chemical composition, and vehicular motor operating condition, should influence the formic/acetic ratios. In this sense, direct emissions from vehicles would be defined as the principal source at an urban VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3077

TABLE 4. Winter Measurements of Formaldehyde and Acetaldehyde in Urban Sites of Brazil

a

sampling site

HCHO (ppbv)

CH3CHO (ppbv)

HCHO/CH3CHO

ref

Rio de Janeiro, RJ Rio de Janeiro, RJ Rio de Janeiro, RJ Sa˜o Paulo, SP Salvador, BA Salvador, BA Sa˜o Paulo, SP, 1993 Sa˜o Paulo, SP, 1997 Sa˜o Paulo, SP Porto Alegre, RS Sa˜o Paulo, SP

10.1-15.0 70a 0.7-6.5 2.3-19.3 9.62-27.8 73.7-88.0a 43-50a 28-34a 4.2b 13.9 1.05-46.7

13.3-17.3 241a 2.6-6.6 0.9-18.0 1.56-12.20 54.8-93.0a 49-60a 31-37a 9.2b 5.7 1.21-56.6

0.80 0.29 0.88 0.29-1.25

25 24 24 24 26 26 46 46 31 27 this work

Tunnel sample.

b

9

0.79-2.72

Average mixing ratio.

site if the formic/acetic ratios were similar for both the tunnel and urban sites. In contrast to the tunnel data, the formic to acetic acid ratios obtained for both sites investigated in this work were 2:1 for the AF samples and 4:1 for the CID samples. Average formic/acetic ratios of morning, midday, and evening were, respectively, 1.8, 2.8, and 2.6 for the AF samples and 4.1, 4.4, and 4.4 for the CID samples. The formic/acetic ratios showed a large variability, considering both sites (Figure 5), but much higher values were found of the formic/acetic ratio at the CID site. In addition to the meteorological variables, direct emissions, in situ formation, and dry deposition removal rates are dissimilar at the AF and CID sites. On the basis of our tunnel (formic/acetic ) 0.25) and urban site (formic/acetic average ) 2.4 for the AF samples and 4.3 for the CID samples) results, photochemical formation seems to be the main source of the acetic and formic acids present at both sites studied (16). Several thermal inversion episodes (high ambient temperature and solar irradiation) that occurred during the sampling period may justify the in situ formation of photochemical products. Compared to other urban sites, the formic/acetic ratios found at the CID site are higher than those seen in Los Angeles (41) and Brussels (7). It is worth mentioning that, in our first study on winter measurements (1996) of carboxylic acids (carried also out at the CID site of Sa˜o Paulo), the average ratio was quite lower (1.2) than the one found in this work (4.3) (14). This might be attributed to a change in the types of vehicles and the fuel chemical composition. From 1996 to 1999 (see Figure 1), there was a significant increase in the number of vehicles fueled with gasohol (25%) in the MASP, whereas the number of vehicles fueled with hydrated ethanol was maintained (32). Carbonyl Compounds. Gaseous formaldehyde and acetaldehyde, the most abundant aldehydes found in urban areas, were detected in all samples. The ambient carbonyl mixing ratios recorded were comparable to those observed previously in other urban areas. They were relatively high, ranging from 1.0 to 46.7 (formaldehyde) with 1.4-50.9 ppbv (acetaldehyde) for the CID site, and from 2.5 to 45.6 (formaldehyde) with 1.2-56.6 ppbv (acetaldehyde) for the AF site. In Rome, the most abundant carbonyl observed was formaldehyde (up to 27 ppbv), followed by acetaldehyde (