versión impresa ISSN 0187-6236
Atmósfera v.20 n.4 México oct. 2007
Seasonal dependence of peroxyacetylnitrate (PAN) concentrations in downtown Santiago, Chile
M. A. RUBIO
Facultad de Química y Biología, Universidad de Santiago de Chile, Santiago, Chile Corresponding author; email: firstname.lastname@example.org
Facultad de Ciencias, Universidad de Santiago de Chile, Santiago, Chile
E. LISSI, G. VILLENA
Facultad de Química y Biología, Universidad de Santiago de Chile, Santiago, Chile
Received May 17, 2006; accepted January 18, 2007
El ozono y el nitrato de peroxiacetilo (PAN) son contaminantes secundarios de origen fotoquímico. Es por lo tanto esperable que alcancen sus máximas concentraciones en periodos de alta irradiación solar (primavera y verano). Sin embargo, en la ciudad de Santiago de Chile se presentan altas concentraciones de PAN en época de otoño. Los promedios del valor máximo diario son: 4.0 ± 1.7 ppb, en otoño (mayo y junio); 0.9 + 0.5 ppb en invierno (julio y agosto); 1.9 ± 1.0 ppb en primavera (octubre y noviembre) y 3.1 ± 1.4 ppb en verano (enero y marzo). En otoño, los valores integrados del perfil diario de concentración son también considerablemente mayores que en otros períodos del año (promedio del perfil diario: 38 ppbhora en mayo y 13 ppb hora en enero). Este comportamiento se relacionaría con la menor estabilidad térmica del PAN en verano. Este factor, junto a la alta estabilidad atmosférica y menores alturas de la capa de mezcla en los meses de otoño, explicaría los altos niveles de concentraciones medidos en períodos de baja irradiación solar.
Peroxyacetylnitrate (PAN) and ozone are considered typical photochemical secondary pollutants and, as such, are expected to reach higher concentrations in periods of high irradiance (summer). However, in Santiago, PAN highest values are reached in autumn. The average of daily maximums are: 4.0 ± 1.7 ppb in autumn (May and June); 0.9 ± 0.5 ppb in winter (July and August); 1.9 ± 1.0 ppb in spring (October and November); 3.1 ± 1.4 ppb in summer (January and March). In autumn the integrated daily values are also considerably higher than in other periods (averages of 38 ppb h in May and 13 ppb h in January). This behavior is related to a lower PAN's thermal stability in summer. This factor, together with an increased atmospheric stability and lower height of the mixing layer in autumn explains the high levels measured in periods of relatively low solar irradiance.
Keywords: PAN, downtown Santiago City, photochemical smog.
Peroxyacetylnitrates are photochemical products with important roles in local and regional atmospheric chemistry (Kourtidis et al., 1993; Williams and Grosjean, 1990; Nielsen et al., 1981; Rembges et al., 2001; Gaffney and Marley, 1993). Peroxyacetylnitrate (PAN) is the most abundant compound of the family, and concentrations as high as several ppb have been measured in polluted urban atmospheres (Rappenglück et al., 1993; Gaffney et al., 1999; Rappenglück et al., 2000). At these levels, PAN is very toxic to vegetation and humans, being a potent lachrymator (WHO, 1996; EnJang and MingHuei, 1995; Taylor, 1969). In addition, PAN can be transported, mostly during night time and/or in winter time due to the lower ambient temperatures, thus becoming a free radical source far away from emission areas (Crutzen, 1979; Derwent and Jenkin, 1991; Gaffney and Marley, 1993; Suppan et al., 1998; Aneja et al., 1999).
PAN has been considered a better indicator of photochemical smog episodes than ozone (Bottenheim et al., 1994; Rappenglück et al., 2003), due to its negligible natural background and the fact that its depletion by new nitrogen oxides (NO) emissions must be preceded by its thermal cleavage and can be considerably slower than that of ozone removal. Summer daily profiles of PAN have been determined in several highly polluted cities, and most studies indicate that daily maxima of ozone and PAN take place at hours of high irradiance. These maxima increase with temperature and are closely correlated (Wunderli and Gehring, 1991; Sillman and Samson, 1995; Rubio et al., 2004). The ratio of the maximum daily values (PANmax / Ozonemax) tends to increase when the maximum daily temperature increases.
Seasonal changes in PAN levels have been less studied. In remote sites, maximum PAN mixing ratios have been observed in winter time, a result related to the increased PAN's thermochemical stability (Gaffney and Marley, 1993). Also, relatively high PAN mixing ratios have been reported during autumn on windless clearsky days (Aneja et al., 1999; Pippin et al., 2001; Rembges et al., 2001). To explain these results it has been proposed that air masses can reach conditions of high NOx mixing ratios and high levels of volatile organic compounds (VOC), conditions that would favor PAN production. In order to assess the behavior of the PAN / ozone ratio under conditions of low solar irradiance and temperature (that would disfavor their production) and high atmospheric stability and low height of the mixing layer (that would favor the accumulation of primary pollutants) we have measured oxidant concentrations in Santiago de Chile atmosphere in different periods of the year. The data obtained are presented and discussed in the present communication.
PAN, toluene, ozone, nitrogen dioxide and UVA irradiance were measured at the Universidad de Santiago de Chile (USACH) campus, near downtown Santiago city.
Toluene (detection limit 1.3 ppb; precision 8%), ozone (detection limit 1.5 ppb: precision 4%) and nitrogen dioxide (detection limit 0.5 ppb: precision 6%) were measured at 7.5 m height with a DOAS automatic instrument (OPSIS) with 453 meters optical path length. The instrument provides a measurement every 10 minutes.
PAN was measured by gas chromatography employing a phenylmethyl polysiloxane/ dimethyl polysiloxane capillary column and electron capture detection (GCPAN from Meteorologieconsult GmbH, Germany) with 25 ppt detection limit (VolzThomas et al., 2002). Sampling was performed at 3 meters altitude above ground, and the equipment provides real time measurements, 24 hours a day, every 10 minutes. The calibration system is based on the photolysis of acetone in presence of NO in a flow reactor, with a reaction yield for PAN of 92 ± 2%. The estimated precision in PAN determinations was 9%. The technique has been validated in intercomparison experiments (Schrimpf et al., 1995).
Solar radiation (total UVA) was measured with a PMA 110 UVA detector from Solar Light Co. This detector has a spectral response in the 320 to 400 nm range, with a broad peak around 370 nm.
CO and NO concentration values were obtained from the local Government Monitoring Network. Data were taken from monitoring stations located at Parque O'Higgins and/or Cerrillos (henceforth denoted as POH and CER), ~0.2 km from the USACH campus. These data were obtained with a NOx/NO monitor Labs model 9811, limit of detection 0.5 ppb, and precision 1 ppb, and a CO monitor Labs model 9830, limit of detection 0.06 ppm, precision 0.1 ppm. Dews were collected and measured as previously described (Rubio et al., 2002)
3. Results and discussion
Measurements were carried out during October 2004 and November and December 2003 (spring), January and March 2004 (summer), April, May and June 2004 (autumn), and July, August and September 2004 (winter). Monthly averages of maximum daily PAN values are given in Table I. Rainy days have not been considered in these averages. The values given in this table are considerably smaller than those previously reported by Rappenglück et al. (2000), at a sampling site (Las Condes) located ~12 km east of downtown. This difference can be related to the different location of the sampling sites. This is evidenced when ozone levels reported by Rappenglück et al. (2000) in Las Condes and POH are compared. It is noticeable that in POH, a sampling site near to the USACH campus, ozone values are considerably smaller than those measured in Las Condes.
The data of Table I, as well as other relevant measurements, are summarized in Table II and in Figure 1. The results obtained for toluene, CO and NOx are those expected given the higher stability and lower mixing layer during cold time. However, it is noteworthy that maximum values are observed in late autumn and early winter (May and June). On the other hand, the behavior of ozone is that expected for a typical secondary pollutant generated in photochemical smog episodes, with maximum values in summertime and early autumn (November to April). As well, the behavior of PAN is intermediate between that of a primary pollutant and ozone, with maximum values during autumn. The tendency of PAN to accumulate in cold weather (relative to ozone) is further emphasized if the ratio PAN/ozone is considered (Fig. 2). This figure clearly shows a noticeable increase in PAN to ozone (maximum values) ratios during May and June, relative to summertime. Furthermore, the PAN diurnal peak is considerably wider in May than in summertime (Fig. 3 for one day example), leading to considerably larger integrated daily PAN values during late autumn. Averages of the integrated values, excluding rainy days, are given in Table III.
The high values of PAN observed in autumn are at first sight puzzling for a secondary pollutant if it is considered that during this time period UVA irradiance is minimum (Fig. 1). A plausible explanation for these high PAN levels is based on an increased amount of early morning precursors, as evidenced for the NOx , CO and toluene levels. These increased levels can be ascribed to the higher stability and lower mixing layer heights of the atmosphere during cold time (Rutlland and Garreaud, 1995). However, this factor should affect both ozone and PAN values and can not fully explain per se the increase in PAN to ozone ratios. A factor that could explain these results is the higher thermochemical stability of PAN during late autumn. In fact, average maximum temperatures are around 30 °C in January and 18 °C in May. This temperature difference implies a nearly five times slower rate of PAN thermal decomposition (FinlaysonPitts and Pitts, 2000). This slower decomposition rate can explain the higher PAN / ozone ratios and wider PAN peaks during cold weather. Furthermore, it could explain why during autumn and winter PAN nighttime levels are considerably larger than in summertime (Table III). It is interesting to note that, for ozone, the opposite trend is observed (Table IV). The low nighttime values of ozone can be understood in terms of an increased titration by emitted NO under conditions of high atmospheric stagnation.
Another factor that could favor PAN accumulation is the NO2/NO ratio, since its increase can increase both PAN production and its thermal stability. These values are also given in Table II.
PAN stability can be calculated if the temperature and NO2/NO ratios are known. As expected, the PAN stability is considerably higher in cold weather. Monthly average values of PAN lifetimes at the moment of its maximum concentrations were: May, 337 minutes; August, 276 minutes; October, 200 minutes; January, 93 minutes. However, differences in PAN stability per se can not explain the high PAN concentrations measured in autumn since, when all measurements are considered, it is not observed a significant correlation between PAN lifetimes and maximum daily values (data not shown). It can be then concluded that a combination of factors, including high atmospheric stability, relatively high NO2/NO ratios and increase PAN stability can explain the high PAN values and PAN /ozone ratios observed in autumn, particularly when daily integrated values are considered.
In order to assess the relevance of early morning precursors accumulation (measured by the morning maximum CO value), temperature and NO2/NO ratios, a multiparametric test comprising these three independent variables was applied to all the set of individual values.
In this equation "cte" is a parameter without a clear physical meaning that should take into account the average contribution of other factors to PAN values (cte = 1.5244 ppb). The three coefficients obtained, and the degree of confidence of their contribution, were:
showing that the three factors considered are important in determining PAN values. In particular, the dependence on the amounts of precursors could explain why maximum PAN values are observed in autumn, where maximum levels of CO and toluene have been determined (Table II and Fig. 1).
The agreement between calculated and experimentally determined maximum PAN values is given in Figure 4. The correlation between both parameters is:
PANcal = (1.46 ± 0.13) + (0.44 ± 0.04) PANexper
with r = 0.66 andp < 0.0001 forn = 132. However, the fact that the ordinate is significantly different from zero, and that the slope differs from 1.0 would indicate that a combination of variables and/ or other non considered variables are also important. A relevant parameter could be the radiation intensity. This factor was not considered in the multiparametric relationship given its dual effect. In fact, high irradiances increase the rate of hydroxyl radical production, increasing the level of photochemical oxidants and, among them, PAN levels. On the other hand, radiation intensity decreases NO2/NO ratios due to NO2 photolysis, increasing the rate of PAN decomposition. Furthermore, the fact that NO2 photolysis increases ozone values tends to favor the accumulation of this oxidant during periods of high irradiance (summer), leading to relatively low PAN/ozone ratios during summertime (Fig. 2).
During May, dew events in Santiago are frequent. The presence of the wet surface could modify PAN chemistry and/or increase its production rate by an increased input of hydroxyl radicals after sunrise (Rubio et al., 2002). In order to test this possibility, daily maximum PAN values were plotted as a function of the amount of collected dew water in April, May and June (Fig. 5). These data show only a very weak, not significant relationship, implying that PAN production and/or removal is almost unrelated to dew formation. This could reflect the low solubility of PAN in aqueous solutions (Gaffney et al., 1989; Frenzel et al., 2000).
High values of PAN concentrations measured in Santiago, Chile in autumn are higher than those observed in summer. These high values are due to increased levels of precursors, associated to the stability of the atmosphere, and a decreased rate of thermal decomposition.
Thanks are given to FONDECYT (project 1040032) for financial support.
Aneja V. P., B. E. Harstsell, DS. Kim and D. Grosjean, 1999. Peroxyacetyl nitrate in Atlanta, Georgia: comparison and analysis of ambient data for suburban and downtown locations. J. Air Waste Manage. 49, 2. [ Links ]
Bottenheim J. W., A. Sirios, K. A. Brice and A. J. Gallant, 1994. Five years of continuous observations of PAN and ozone at a rural location in eastern Canada. J. Geophys. Res. 99, 53335352. [ Links ]
Crutzen P. J., 1979. The role of NO and NO2 in the chemistry of the troposphere and stratosphere. Ann. Rev. Earth P. Sc. 7, 443472. [ Links ]
Derwent R. G. and M. E. Jenkin, 1991. Hydrocarbons and the long range transport of ozone and PAN across Europe. Atmos. Environ. 25A, 16611678. [ Links ]
EnJang S. and H. MingHuei, 1995. Detection of peroxyacetyl nitrate at phytotoxic level and its effects on vegetation in Taiwan. Atmos. Environ. 29, 28992904. [ Links ]
FinlaysonPitts B. and J. N. Pitts, 2000. Chemistry of the upper and lower atmosphere. Academic Press U.S. A. [ Links ]
Frenzel A., S. Kutzune, K. Tabuchi and T. Ibusuki, 2000. Solubility and reactivity of peroxi acetyl nitrate (PAN) in dilute aqueous salt solutions and sulphuric acid. Atmos. Environ. 34, 36413644. [ Links ]
Gaffney J. S., N. A. Marley and E.W. Prestbo, 1989. Peroxyacyl nitrates (PANs): their physical and chemical properties. In: Handbook of Environmental Chemistry, Volume 4/part B. (Air Pollution) (O. Hutzinger, Ed.), SpringerVerlag, Berlin, Germany, 138. [ Links ]
Gaffney J. S. and N. A. Marley, 1993. Measurements of peroxyacetyl nitrate at a remote site in the southwestern United States: tropospheric implications. Environ. Sci. Technol. 27, 19051910. [ Links ]
Gaffney J. S., N. A. Marley, M. M. Cunningham and P. V. Doskey, 1999. Measurements of peroxyacyl nitrates (PANs) in Mexico City: Implications for mega city air quality impacts on regional scales. Atmos. Environ. 33, 50035012. [ Links ]
Kourtidis K. A., P. Fabian, C. Zerefos and B. Rappenglück, 1993. Peroxyacetylnitrate (PAN), peroxypropionylnitrate (PPN) and PAN/ozone ratio measurements at three sites in Germany. Tellus 45B, 442457. [ Links ]
Nielsen T., U. Samuelsson, C. Grennfelt and E. L. Thomsen, 1981. Peroxyacetylnitrate in longrange transported polluted air. Nature 293, 553555. [ Links ]
Pippin M., B. Steve, T. Thornberry, M. Town, M. A. Corroll and S. Sillman, 2001. Seasonal variations of PAN, and O3 at the upper Midwest PROPHET site. J. Geophys. Res. 106, 2445124464. [ Links ]
Rappenglück B., K. Kourtidis and P. Fabian, 1993. Measurement of ozone and peroxyacetylnitrate (PAN) in Munich. Atmos. Environ. 27B, 293305. [ Links ]
Rappenglück B., D. Melas and P. Fabian, 2003. Evidence of the impact of urban plumes on remote sites in the Eastern Mediterranean. Atmos. Environ. 37, 18531864. [ Links ]
Rappenglück B., P. Oyola, I. Olaeta and P. Fabian, 2000. The evolution of photochemical smog in the Metropolitan Area of Santiago de Chile. J. Applied Meteorol. 39, 275290. [ Links ]
Rembges D., M. Duane, J. Hjorth, H. Bloem, W. Zaaiman and B. R. Larsen, 2001. PAN and air quality in northern Italy: diurnal variation and longterm trends. 8th European Symposium on PhysicoChemical Behavior of Atmospheric Pollutants, 1720 Sep. Torino, Italy. [ Links ]
Rubio M. A., P. Oyola, E. Gramsch, E. Lissi, J Pizarro and G. Villena, 2004. Ozone and peroxyacetylnitrate in downtown Santiago, Chile. Atmos. Environ. 38, 49314939. [ Links ]
Rubio M. A., E. Lissi, and G. Villena, 2002. Nitrite in rain and dew in Santiago city, Chile. Its possible impact on the early morning start of the photochemical smog. Atmos. Environ. 36, 293297. [ Links ]
Rutllant J. and R. Garreaud, 1995. Meteorological air pollution potential for Santiago, Chile: towards an objective episode forecasting. Environ. Monit. Assess. 34, 223244. [ Links ]
Schrimpf W., K. P. Müller, F. J. Johnen, K. Lienaerts and J. Rudolph, 1995. An optimized method for airborne peroxiacetyl nitrate (PAN) measurements. J. Atmos. Chem. 22, 303317. [ Links ]
Sillman S. and P. J. Samson, 1995. Impact of temperature on oxidant photochemistry in urban, polluted rural and remote environments. J. Geophys. Res. 100, 1149711508. [ Links ]
Suppan P., P. Fabian, L. Vyras and S. E. Gryning, 1998. The behavior of ozone and peroxyacetylnitrate concentrations for different wind regimes during the medcaphottrace campaign in the greater area of Athens, Greece. Atmos. Environ. 32, 20892102. [ Links ]
Taylor O. C., 1969. Importance of peroxyacetylnitrate (PAN) as a phytotoxic air pollutant. J. Air Poll. Control Assoc. 19, 347351. [ Links ]
VolzThomas A., I. Xueref and R. Schmitt, 2002. An automatic gas chromatograph and calibration system for ambient measurements of PAN and PPN. Environ. Sci. Pollut. 4, 7282. [ Links ]
WHO, 1996. Update and revision of the WHO air quality guideline for Europe. Classical air pollutants; ozone and other photochemical oxidants. European Center for Environment and Health, Bilthoven, Netherlands. [ Links ]
Williams E. L. II and D. Grosjean, 1990. Southern California air quality study: Peroxyacetyl nitrate. Atmos. Environ. 24A, 23692377. [ Links ]
Wunderli S. and R. Gehrig, 1991. Influence of temperature on formation and stability of surface PAN and ozone. A two year field study in Switzerland. Atmos. Environ. 25A, 15991608. [ Links ]