Analysis of the Unresolved Organic Fraction in Atmospheric Aerosols

Sep 30, 2010 - Division, Atmospheric Science and Technology Directorate, Environment Canada, 335 River Road, Ottawa,. ON K1A 0H3, Canada, and ...
0 downloads 0 Views 5MB Size
Anal. Chem. 2010, 82, 8017–8026

Analysis of the Unresolved Organic Fraction in Atmospheric Aerosols with Ultrahigh-Resolution Mass Spectrometry and Nuclear Magnetic Resonance Spectroscopy: Organosulfates As Photochemical Smog Constituents† Philippe Schmitt-Kopplin,*,‡,§ Andras Gelencse´r,| Ewa Dabek-Zlotorzynska,⊥ Gyula Kiss,| Norbert Hertkorn,‡ Mourad Harir,‡ Yang Hong,X and Istvan Gebefu¨gi‡ Department of BioGeoChemistry and Analytics, Institute of Ecological Chemistry, Helmholtz Zentrum Mu¨nchen, German Research Center for Environmental Health, Ingolstaedter Landstrasse 1, 85764 Neuherberg, Germany, Lehrstuhl fu¨r Chemische-Technische Analyse und Chemische Lebensmitteltechnologie, Technische Universita¨t Muenchen, Weihenstephaner Steig 23, 85354 Freising-Weihenstephan, Germany, Air Chemistry Group of the Hungarian Academy of Sciences, Veszpre´m, Hungary, Analysis and Air Quality Section, Air Quality Research Division, Atmospheric Science and Technology Directorate, Environment Canada, 335 River Road, Ottawa, ON K1A 0H3, Canada, and Biogeochemistry Department, Max Planck Institute for Chemistry, 55128 Mainz, Germany Complementary molecular and atomic signatures obtained from Fourier transform ion cyclotron resonance (FTICR) mass spectra and NMR spectra provided unequivocal attribution of CHO, CHNO, CHOS, and CHNOS molecular series in secondary organic aerosols (SOA) and highresolution definition of carbon chemical environments. Sulfate esters were confirmed as major players in SOA formation and as major constituents of its water-soluble fraction (WSOC). Elevated concentrations of SO2, sulfate, and photochemical activity were shown to increase the proportion of SOA sulfur-containing compounds. Sulfonation of CHO precursors by means of heterogeneous reactions between carbonyl derivatives and sulfuric acid in gas-phase photoreactions was proposed as a likely formation mechanism of CHOS molecules. In addition, photochemistry induced oligomerization processes of CHOS molecules. Methylesters found in methanolic extracts of a SOA subjected to strong photochemical exposure were considered secondary products derived from sulfate esters by methanolysis. The relative abundance of nitrogen-containing compounds (CHNO and CHNOS series) appeared rather dependent on local effects such as biomass burning. Extensive aliphatic branching and disruption of extended NMR spin-systems by carbonyl derivatives and other heteroatoms were the most significant structural motifs in SOA. The presence of heteroatoms in elevated oxidation states suggests a † Part of the special issue “Atmospheric Analysis as Related to Climate Change”. * To whom correspondence should be addressed. Phone: (+49) 089 3187 3246. Fax: (0049) 089 3187 3358. E-mail: schmitt-kopplin@helmholtz-muenchen. de. ‡ German Research Center for Environmental Health. § Technische Universita¨t Muenchen. | Air Chemistry Group of the Hungarian Academy of Sciences. ⊥ Environment Canada. X Max Planck Institute for Chemistry.

10.1021/ac101444r  2010 American Chemical Society Published on Web 09/30/2010

clearly different SOA formation trajectory in comparison with established terrestrial and aqueous natural organic matter. Atmospheric aerosol particles undoubtedly contribute strongly to climate change as recently published by the Intergovernmental Panel on Climate Change.1 Organic compounds are major constituents of fine continental aerosols (up to 30-50 m/m %), and the understanding of their chemical composition, their properties, and reactivity are becoming even more important for assessing aerosol effects both on global change and human health. Among organic aerosols, secondary organic aerosols (SOA) are getting predominant over most parts of the continents,2 especially in the summer due to more intense photochemistry. A recent source apportionment study backed up by radiocarbon measurements for continental sites in Europe concluded that in the summer SOA from biogenic precursors are clearly predominant (63-76% of total carbon, TC).3 An earlier study on SOA in rural aerosols in Portugal arrived at very similar conclusions (68-78% of organic carbon, OC).4 It was also pointed out that even if most of the SOA form from biogenic precursors, their generation (1) Forster, P.; Ramaswamy, V.; Artaxo, P.; Berntsen, T.; Betts, R.; Fahey, D. W.; Haywood, J.; Lean, J.; Lowe, D. C.; Myhre, G.; Nganga, J.; Prinn, R.; Raga, G.; Schulz, M.; Van Dorland, R. Changes in Atmospheric Constituents and in Radiative Forcing. In Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K. B., Tignor, M., Miller, H. L., Eds.; Cambridge University Press: Cambridge, U.K., 2007. (2) Hallquist, M.; Wenger, J. C.; Baltensperger, U.; Rudich, Y.; Simpson, D.; Claeys, M.; Dommen, J.; Donahue, N. M.; George, C.; Goldstein, A. H.; Hamilton, J. F.; Herrmann, H.; Hoffmann, T.; Iinuma, Y.; Jang, M.; Jenkin, M. E.; Jimenez, J. L.; Kiendler-Scharr, A.; Maenhaut, W.; McFiggans, G.; Mentel, T. F.; Monod, A.; Prevot, A. S. H.; Seinfeld, J. H.; Surratt, J. D.; Szmigielski, R.; Wildt, J. Atmos. Chem. Phys. 2009, 9, 5155–5236. (3) Gelencse´r, A.; May, B.; Simpson, D.; Sanchez-Ochoa, A.; Kasper-Giebl, A.; Puxbaum, H.; Caseiro, A.; Pio, C.; Legrand, M. J. Geophys. Res. Atmos. 2007, 112, D23S04.

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

8017

involves photooxidants whose levels have been strongly enhanced by human activities. Therefore, strictly speaking, most of the biogenic SOA should be considered anthropogenic by origin.5,6 Globally tropospheric oxidation of biogenic and anthropogenic volatile organic compounds represents a highly uncertain source strength for SOA of 30-270 Tg year-1.7 Aerosols undergo source independent aging processes, i.e., chemical reactions and mixing processes reduce the initially large chemical diversity and eventually converge toward rather common SOA structural and bulk properties.8,9 Despite the increasing importance of SOA, most atmospheric chemistry models still severely underestimate SOA mass concentrations observed in most parts of the troposphere. Albeit the importance of heterogeneous and multiphase reactions in SOA formation was recognized earlier,10 our understanding on the extent, mechanisms, and significance of these processes in the troposphere is still in its infancy mainly due to the lack in chemical and structural information on the reaction products. Very recently Fourier transform ion cyclotron resonance (FTICR) and Orbitrap mass spectrometry enabled the assignment of exact elemental formulas from highly accurate experimental masses and revealed the presence of large numbers of sulfur containing molecules in atmospheric aerosols,11-13 rainwater,14 and fog15 or from samples in targeted smog chamber studies and laboratory experiments involving known precursors.16-21 Nuclear magnetic resonance spectroscopy (NMR) offered complementary structural characterization of organic aerosols involving proton NMR (1H NMR) analysis directly from filter extracts22 or after (4) Castro, L. M.; Pio, C. A.; Harrison, R. M.; Smith, D. J. T. Atmos. Environ. 1999, 33, 2771–2781. (5) Kanakidou, M.; Tsigaridis, K.; Dentener, F. J.; Crutzen, P. J. J. Geophys. Res. 2000, 105, 9243–9254. (6) Kroll, J. H.; Seinfeld, J. H. Atmos. Environ. 2008, 42, 3593–3624. (7) Andreae, M. O.; Crutzen, P. J. Science 1997, 276, 1052–1058. (8) Andreae, M. O.; Crutzen, P. J. Science 2010, 276, 1052–1058. (9) Heald, C. L.; Kroll, J. H.; Jimenez, J. L.; Docherty, K. S.; DeCarlo, P. F.; Aiken, A. C.; Chen, Q.; Martin, S. T.; Farmer, D. K.; Artaxo, P. Geophys. Res. Lett. 2010, 37, L08803. (10) Ravishankara, A.R. Science 1997, 276, 1058–1065. (11) Reemtsma, T.; These, A.; Venkatachari, P.; Xia, X.; Hopke, P. K.; Springer, A.; Linscheid, M. Anal. Chem. 2006, 78, 8299–8304. (12) Wozniak, A. S.; Bauer, J. E.; Sleighter, R. L.; Dickhut, R. M.; Hatcher, P. G. Atmos.Chem. Phys. Discuss. 2008, 8, 6539–6569. (13) De Haan, D. O.; Corrigan, A. L.; Smith, K. W.; Stroik, D. R.; Turley, J. J.; Lee, F. E.; Tolbert, M. A.; Jimenez, J. L.; Cordova, K. E.; Ferrell, G. R. Environ. Sci. Technol. 2009, 43, 2818–2824. (14) Altieri, K. E.; Turpin, B. J.; Seitzinger, S. P. Atmos. Chem. Phys. 2009, 9, 2533–2542. (15) Mazzoleni, L. R.; Ehrmann, B. M.; Shen, X. H.; Marshall, A. G.; Collett, J. L. Environ. Sci. Technol. 2010, 44, 3690–3697. (16) Reinhardt, A.; Emmenegger, C.; Gerrits, B.; Panse, C.; Dommen, J.; Baltensperger, U.; Zenobi, R.; Kalberer, M. Anal. Chem. 2007, 79, 4074– 4082. (17) Altieri, K. E.; Seitzinger, S. P.; Carlton, A. G.; Turpin, B. J.; Klein, G. C.; Marshall, A. G. Atmos. Environ. 2008, 42, 1476–1490. (18) Nguyen, T. B.; Bateman, A. P.; Bones, D. L.; Nizkorodov, S. A.; Laskin, J.; Laskin, A. Atmos. Environ. 2010, 44, 1032–1042. (19) Mu ¨ ller, L.; Reinnig, M. C.; Hayen, H.; Hoffmann, T. Rapid Commun. Mass Spectrom. 2009, 23, 971–979. (20) Bateman, A. P.; Walser, M. L.; Desyaterik, Y.; Laskin, J.; Laskin, A.; Nizkorodov, S. A. Environ. Sci. Technol. 2008, 42, 7341–7346. (21) Perri, M. J.; Seitzinger, S.; Turpin, B. J. Atmos. Environ. 2009, 43, 1487– 1497. (22) Decesari, S.; Mircea, M.; Cavalli, F.; Fuzzi, S.; Moretti, F.; Tagliavini, E.; Facchini, M. C. Environ. Sci. Technol. 2007, 41, 2479–2484.

8018

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

derivatization.23 Two-dimensional NMR spectroscopy allows better resolved structural elucidation of these very complex mixtures obtained from field sampling24 or laboratory experiments.25 The combination of FTICR mass spectrometry and NMR spectroscopy is ideally suited for characterizing very complex mixtures like terrestrial and extraterrestrial organic matter26,28,29 with molecular resolution. Photochemically formed SOA with a high ratio of heteroatoms appears to be ideally suited for a comprehensive characterization in which parallel determination of functional groups and carbon chemical environment is desirable. Analogously, high-resolution organic structural spectroscopy was used in this study to evaluate the ubiquitous occurrence of organosulfates in various atmospheric samples. This was achieved by analyzing filter extracts of organic aerosols from diverse rural sources that were collected during various weather conditions and by analyzing rain and hail samples. The aim of the study was to investigate the importance of sulfuric acid in the formation of organosulfur compounds, as followed with FTICR-MS and NMR on a molecular level. This study demonstrates the necessity of high magnetic field (12 T) ultrahigh resolution Fourier transformation ion cyclotron resonance mass spectrometry for the unequivocal elemental composition assignment of a CHO, CHNO, CHOS, and CHNOS series within such complex organic matrixes. In addition, 2D-NMR techniques complement the FTICR-MS characterization of the water-soluble fraction of OC. Potential SOA formation mechanisms under natural conditions (i.e., presence of sulfuric acid in the condensed phase) and the evolution of sulfur diesters are proposed. The elemental composition and molecular mass distributions of organosulfur compounds as well as their inferred precursors as monomers and dimers were evaluated. EXPERIMENTAL METHODS Sampling. Fine aerosol samples were collected in different seasons and weather conditions at rural K-puszta sampling sites on the Great Hungarian Plain in Hungary (2004 and 2005) and at Canadian rural sites (Saint Anicet, Quebec, and Canterbury, New Brunswick) in 2005 and 2007.30 Rain samples followed by hail samples were sampled on May 30, 2008 in Neuherberg/Oberschleissheim, Bavaria, Germany; a heavy hail was followed by heavy rain and sampling took place within the same 30 min. Further details on the samples can be found in the Supporting Information. (23) Moretti, F.; Tagliavini, E.; Decesari, S.; Facchini, M. C.; Rinaldi, M.; Fuzzi, S. Environ. Sci. Technol. 2008, 42, 4844–4849. (24) Sannigrahi, P.; Sullivan, A. P.; Weber, R. J.; Ingall, E. D. Environ. Sci. Technol. 2006, 40, 666–672. (25) Maksymiuk, C. S.; Gayahtri, C.; Gil, R. R.; Donahue, N. M. Phys. Chem. Chem. Phys. 2009, 36, 7810–7818. (26) Hertkorn, N.; Meringer, M.; Gugisch, R.; Ruecker, C.; Frommberger, M.; Perdue, E. M.; Witt, M.; Schmitt-Kopplin, Ph. Anal. Bioanal. Chem. 2007, 389, 1311–1327. (27) Hertkorn, N.; Frommberger, M.; Schmitt-Kopplin, Ph.; Witt, M.; Koch, B.; Perdue, E. M. Anal. Chem. 2008, 80, 8908–8919. (28) Einsiedl, F.; Hertkorn, N.; Wolf, M.; Frommberger, M.; Schmitt-Kopplin, Ph.; Koch, B. Geochim. Cosmochim. Acta 2007, 71, 5474–5482. (29) Schmitt-Kopplin, Ph.; Gabelica, Z.; Gougeon, R. D.; Fekete, A.; Kanawati, B.; Harir, M.; Gebefuegi, I.; Eckel, G.; Hertkorn, N. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 2763–2768. (30) Yassine, M.; Dabek-Zlotorzynska, E.; Schmitt-Kopplin, Ph. Electrophoresis 2009, 30, 1756–1765.

Figure 1. (a) ESI(-)-FTICR MS spectra of K-Puszta samples from 2005 with a biomass burning event (KPBMB 2005) showing the regular shaped signal distribution over the whole mass range (m/z 150-1000) as well as in each nominal mass; (b) elemental composition distribution in nominal mass 411 with the assigned CHO, CHNO, and CHOS molecular series. (b) Advantage of ultrahigh resolution, especially at higher masses for the analysis of complex organic samples; details of nominal mass 221 (I), 411 (II), and 621 (III) at different computed resolutions (b-d) from experimental data (a). The elemental composition assignments of (IIa) are given in Figure 1a, and the closely spaced doublet signals correspond to CHO (left) and CHOS (right) compounds. At a resolution of 250 000, CHO can be distinguished from CHOS with a m/z difference of 2.4 millimasses; however, lower mass resolution as shown in IId, for example, does not allow any reliable elemental composition assignment out of these complex SOA mixtures.

Sample Preparation. Samples were extracted in aqueous solutions and desalted on solid phase extraction columns (HLB and C18 Extraction). For extraction details, please refer to the Supporting Information. Although the technical details of the procedures used are slightly different, the concept is that acidic and neutral organic compounds with considerable nonpolar parts can be widely separated from the inorganic ions present in the aqueous (alkaline) extract which provides superior reproducibility in electrospray ionization. FTICR Mass Spectrometry. Ultrahigh-resolution mass spectra were acquired on a Bruker (Bremen, Germany) APEX 12 Qe Fourier transform ion cyclotron resonance mass spectrometer equipped with a 12 T superconducting magnet and a APOLLO II electrospray source from methanolic SOA solutions as described in the Supporting Information. NMR Spectroscopy. The NMR spectra of SOA KP2004 (approximately 50 µg) and K2005 (approximately 250 µg) were acquired with a Bruker DMX 500 spectrometer (B0 ) 11.7 T) at 278 K (KP2004) or 283 K (KP2005) from redissolved solids in CD3OD (99.95% 2H; Merck) with Bruker standard pulse sequences and a 5 mm TXI cryogenic probehead as described

in the Supporting Information. Most of the analysis were done already between 2005 and 2006 only shortly after their collection on filters and that were always conserved at -20 °C; some later analysis of the samples in the same conditions did not show any alteration of the sample.

RESULTS AND DISCUSSION FTICR-MS of Organic Aerosols. The FTICR-MS spectra of all samples showed a typical distribution of negative ion signals, very similar at first glance for all the samples and exemplified in Figure 1Aa with the K-Puszta sample KPBMB2005, covering the m/z-range from 150 to 800 Da in a near continuous distribution and with additional single intense peaks between m/z 250 and 450. Fine details of the mass spectra differed from sample to sample. The K-Puszta sample (KPBMB2005) was collected in 2005 during a biomass burning event (field burning/straw) and shows mass distribution as a “dragon tail” up to m/z 900. Figure 1a shows a typical detail of nominal mass 411 with the respective annotation of the exact masses in elemental compositions. The mass error was always found lower than 200 ppb and allowed a credible Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

8019

Table 1. Counts of CHO and CHOS Components in PM2.5 Samples Collected at Different Locations and under Various Meteorological Conditions elemental compositions origin

SPE resin

CHO

CHOS

CHNO

CHNOS

CHOS/CHO

KP2004 KP2005 KPBMB2005 KPBMB2005 + H2SO4 SRW SRW + H2SO4 rain water hail ice water

Oasis HLB C18 Oasis HLB C18 Oasis HLB C18 Oasis HLB C18 Oasis HLB C18 Oasis HLB Oasis HLB C18 C18 no SPE no SPE C18 C18

596 413 740 776 1113 992 878 925 880 924 584 293 2184 1714 1756 1761 2346 447

107 223 167 394 145 290 210 368 235 455 284 1227 558 2267 30 234 1158 89

87 134 90 183 143 63 96 178 91 235 132 132 2149 1515 27 131 1088 47

11 18 11 86 4 38 56 96 46 105 42 523 44 2851 0 0 230 18

0.179 0.539 0.225 0.507 0.130 0.292 0.239 0.397 0.267 0.492 0.486 4.187 0.255 1.322 0.017 0.132 0.493 0.199

ANI294 ANI292 ANI514 CAN124 CAN110

conversion of experimental mass peaks into elemental formulas involving all isotopologues with high confidence (cf. Figure SI-2 in the Supporting Information). The ultrahigh-resolution of the FTICR-mass spectrometer enables differentiating m/z signals within the millimass range, and the mass accuracy in the subparts per million range (lower 200 ppb) allows the determination of molecular formulas up to higher masses according to the element allowed. This certainly is the key to convert all experimental signals into an elementary formula involving all isotopologues with high confidence; this possibility is illustrated with mass 294 in Figure SI-2 in the Supporting Information. The resolution at this mass is between 450 000 and 600 000, enabling the required “base line separation” of the various possible CHNOS-species based on their elemental compositions (31 compositions were already found within m/z 411). The signals show a continuous intensity distribution reflecting the lower probability in the number of feasible isomers in the lowest (highest oxygen content) and highest mass defect range (higher hydrogen content).26 This regular peak shape within the nominal masses is typical of all analyzed samples and reflects the abiotic origin29 in their chemical formation mechanisms typical of atmospheric chemistry (photochemistry, hydrolysis, catalysis, radical reactions). Different series of elemental compositions can be observed within nominal mass 411 with identical mass defects corresponding to a formal compositional exchange of CH4 versus O (mass difference 36.4 mDa); we can distinguish two distinct CHO-series, two distinct CHOS series, and a CHN2O-series (Figure 1b). The key in being able to differentiate the sulfur containing compounds in these samples is the ultrahigh-resolution characteristic of high-field FTICR-MS as illustrated in Figure 1b. The experimental spectrum is shown in part a at three different m/z values: 221, 411, and 621 (respectively, I, II, III). Figure 1b(b),(c),(d) were obtained by smoothing the raw data with a Gaussian function to obtain line broadening, simulating a lower resolution equivalent to lower resolution mass spectrometry. Figure 1b clearly shows the limits of lower resolution approaches because a minimum of 150 000 in resolution is needed to distinguish the CHO from the CHOS series (peak doublets, see annotation of m/z 411 in Figure 1b). 8020

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

average meteorological condition temp [°C]

RH [%]

18.3 (fog/rain) 23.3 (clear) 27 (clear) 18.4 (rain) 21.1 (clear) 25.3 (clear) 26.1 (clear)

91

28.3 28.3

100 100

78 50 87 59 60 67

The visualization of these complex data sets by means of various projections derived from Kendrick or van Krevelen diagrams enabled to extract maximal structural information.26,27,29 Herein we chose the van Krevelen diagrams as a projection of the data in the H/C versus O/C space. Figure SI-1 in the Supporting Information represents the positioning of many targeted compounds in the van Krevelen diagram and describes possible SOA components from the literature on organic aerosols. Sections typical of SOA constituents and precursors can be distinguished directly according to the number of double bond equivalents and their degree of oxidation. Secondary organic aerosols are typically located in the upper right part of the diagram corresponding to highly aliphatic and oxygenated SOA species (see the real samples in this study). The successful annotation of the elemental formulas revealed a considerable contribution of sulfur containing organic compounds to the number of organic compounds in the isolated fractions (Table 1). Within all samples analyzed so far, collected under different atmospheric conditions and climatic regions, a tendency toward a higher amount of S-containing compounds with increasing sulfate, SO2, and photochemical reactivity could be observed. In the samples collected at Canterbury on a clear sunny day, the amount of S-containing compounds increased relative to rainy days; the increase was also observed for samples collected at Saint Anicet and K-puszta. This has been found valid for C18- and HLB-SPE extractions (Table 1). However, C18-SPE generally retained a larger number of Scontaining compounds in comparison with HLB-SPE. The most pronounced difference was observed in a sample collected at K-puszta and extracted with OASIS SPE-columns: the CHOS/CHO ratio reached 418% in sample KP2005 collected under photochemical smog conditions in 2005 while it was only 48% under moderate photochemical reactivity conditions in the year 2004 (KP2004). The percentage of S-containing compounds in two samples collected in rural environments in the U.S.12 was similar to those obtained for the Canadian samples in this study. The high number of S-containing compounds in all the samples indicates the importance of these compounds among the water-soluble organic

Figure 2. (a) Mass resolved atomic H/C ratio and van Krevelen diagrams of a smoke sample from wheat straw burning and (b) reflects the data visualization of the K-Puszta 2005 burning event (KPBMB 2005). The differences in relative amounts in CHO, CHOS, CHNO, and CHNOS, respectively (on the right side), clearly show that the high amounts of CHNO compounds originated from the burning events and the increase of the S-containing molecules (CHOS and CHNOS) over all the m/z range.

aerosol constituents. It is important to note, however, that quantitative conclusions cannot be drawn merely from MS spectra because the ionization efficiency of different compounds may differ by orders of magnitude. Nevertheless, quantification of S-containing organic compounds in a recent study31 supports the assumption that S-containing components contribute to a considerable extent (up to 30%) to the water-soluble organic fraction as well as to the sulfur content of fine rural aerosol. In contrast, the relative abundance of N-containing compounds appears rather dependent on local effects such as, for example, biomass burning as shown with sample KPBMB2005. With a signal-to-noise ratio of 3, 16.933 signals could be extracted as raw data from sample KPBMB2005 and converted into CHO (2.184), CHOS (558), CHNO (2.149), and CHNOS (44) species (29.1% coverage). The corresponding van Krevelen diagram (Figure 2a) reveals a different positioning of the respective molecular series. The sulfur species (CHOS + CHNOS) accounted for more than 10% of all assigned compounds (under these ESI (-) conditions) and CHNO molecules nearly as much as CHO species (Figure 2a). The extensive occurrence of nitrogen-containing organic species likely results from the fact that this sample has been collected during a biomass burning event at K-Puszta. The same masses corresponding to analogous CHO and CHNO series resulted in similar van Krevelen diagrams with CHO and CHNO signals in the aromatic region (see Figure SI-1 in the Supporting Information) and were obtained in a wheat straw burning sample obtained under controlled burning conditions. This congruence demonstrates that a major part of the compounds found in KPBMB2005 had originated from the biomass burning event

(Figure 2b). Note that wheat straw exhibits a C/N ratio of approximately 120 and wood a C/N ratio of 400-800. Accordingly, wheat straw burning releases a significantly elevated number of nitrogen containing compounds in comparison with wood burning events. The K-Puszta samples obtained under pristine conditions did not exhibit aromatic signatures and only showed low amounts of nitrogen containing species (Table 1 and Figure SI-3 in the Supporting Information). The mass distribution of S-containing species are shown together with their H/C ratio for the samples collected at K-Puszta 2004 and 2005 (Figure 3). Similarly, the Canadian SOA samples showed signals toward higher m/z in sunny compared with cloudy weather conditions. These results indicate a shift toward higher m/z in the samples collected in sunny conditions (Canadian samples) or in a period with high photochemical reactivity. This phenomenon can be explained by oligomerization processes which have been observed in smog chamber studies32 but also by other oxidation or condensation reactions as detailed below, or the combination of both. The elemental composition of the S-containing compounds in KP2005 revealed the absence of assignable structures with two and more sulfur indicating oligomerization (mainly dimerization) reactions in SOA to explain the observed increase in molecular weight. A detailed analysis of the m/z distributions of CHO and CHOScontaining compounds revealed a decrease of monomers and an increase of dimer contents and S-containing species with increasing photochemical reactivity. Removal of three oxygen and one sulfur atoms (SO3) from the formula of the S-bearing components can yield the SOA

(31) Luka´cs, H.; Gelencse´r, A.; Hoffer, A.; Kiss, G.; Horva´th, A.; Hartya´ni, Z. Atmos. Chem. Phys. Discuss. 2008, 8, 6825–6843.

(32) Kalberer, M.; Paulsen, D.; Sax, M.; Steinbacher, M.; Dommen, J.; Fisseha, R.; Prevot, A.; Frankevich, V.; Zenobi, R.; Baltensperger, U. Science 2004, 1659–1662.

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

8021

Figure 3. m/z-resolved atomic H/C ratios (top) and corresponding reconstructed mass spectra (middle) of the assigned signals as CHO (blue) and CHOS (green) in KP2004 and KP2005. The mass spectra show a bimodal distribution in K-Puszta 2005 with a decrease of monomers and increase of sulfonated species (monomers and dimers/oligomers). The bottom line shows the respective van Krevelen Diagrams and a high aliphaticity and oxygen richness of the components. The computed CHO components that could form the experimentally found CHOS molecules are depicted in red color.

precursor components that might have been sulfonated; in all cases these resulting components correspond to CHO molecules of SOA type in the van Krevelen diagram (bottom Figure 3), indicating that sulfur entered the molecules together with oxygen. This is in an agreement with recent smog chamber studies of terpenes that proposed the formation of sulfate esters.33,34 Furthermore, such compounds have also been suggested in atmospheric samples.35 A compound with nominal mass 295 reported to produce the most intense individual (-)-ESI ion with m/z 294 in rural samples35,36 is illustrated in Figure SI-2 in the Supporting Information. This ion was attributed to a nitrogen and sulfur containing species with an elemental composition of C10H17NO7S and being identified with liquid chromatography/ tandem mass spectrometry (LC/MS2) as a sulfate ester photometabolite of R-pinene produced in the presence of NOx and (33) Iinuma, Y.; Mu ¨ ller, C.; Berndt, T.; Bo¨ge, O.; Claeys, M.; Herrmann, H. Environ. Sci. Technol. 2007, 41, 6678–6683. (34) Surratt, J. D.; Lewandowski, M.; Offenberg, J. H.; Jaoui, M.; Kleindienst, T. E.; Edney, E., O.; Seinfeld, J. H. Environ. Sci. Technol. 2007, 41, 5363– 5369. (35) Gao, S.; Surratt, J. D.; Knipping, J. D.; Edgerton, E. M.; Shahgholi, E. S.; Seinfeld, M. J. Geophys. Res. 2006, 111, D14314. (36) Kiss, G.; Tomba´cz, E.; Varga, B.; Alsberg, T.; Persson, L. Atmos. Environ. 2003, 37, 3783–3794.

8022

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

acidic seeds.35,37 In the KP2004 sample, this ion (294.0652) showed the highest intensity while in the KP2005 sample its compositional analogues were found at m/z 296.0611 (C9H15NO8S) and m/z 298.0683 (C9H17NO8S) but with lower intensity. Within the Canadian samples, CAN110 (clear sunny) showed the highest intensity in this (-)-ESI ion m/z 294. Figure S-2 in the Supporting Information shows that this ion represents the highest intensity in ESI(-) FTICR MS in the 2004 sample KP2004 with additional detailed spectra of its higher mass isotopologues, confirming the structural assignment. In addition, hundreds of other sulfate esters can be identified from the mass spectra in contrast to previous studies in which only a few sulfate ester compounds were identified because of the lack of mass accuracy and resolution at higher masses. NMR Description. A comprehensive proton NMR characterization of SOA samples KP2004 (2004) and KP2005 (2005) complemented the FT mass spectrometric distinction of the various molecular series (CHO, CHNO, CHOS, CHNOS) with an extensive evaluation of carbon and hydrogen chemical environ(37) Surratt, J. D.; Murphy, S. M.; Kroll, J. H.; Ng, N. L.; Hildebrandt, L.; Sorooshian, A.; Szmigielski, R.; Vermeylen, R.; Maenhaut, W.; Claeys, M.; Flagan, R. C.; Seinfeld, J. H. J. Phys. Chem. A 2006, 110, 9665–9690.

ments (low sample amounts precluded acquisition of meaningful one-dimensional 13C NMR spectra). The minimum confirmed variance of C,O-chemical environments as derived from the assigned FTICR mass peaks in both KP2004 and KP2005 ranged in the ten-thousands and one-dimensional NMR spectra of complex SOA showed the expected severe NMR resonance overlap. Similarity in such resolution-limited spectra does not necessarily imply identical molecules present but rather extensive conformity of principal chemical substructures. 1H NMR spectra of 2004 and 2005 SAO dissolved in CD3OD showed excellent resolution of recognizable NMR resonances, indicating preponderance of small molecules with long transverse relaxation time (favorable for detecting highly resolved 2D NMR spectra), in accordance with the molecular formulas derived from FTICR mass spectra. The envelopes of onedimensional 1H NMR spectra of SOA KP2004 and KP2005 showed remarkable similarity (Figure 4A,B and Figure SI-8 in the Supporting Information) with one striking difference, substantial amounts of methoxy groups observed in SOA KP2005 (10% of 1H NMR integral) contrasted with about 3% occurrence of those units in SOA KP2004 (Table 2). KP2004 with its larger aliphatic proton content appeared less oxidized than KP2005. KP2004 showed a larger abundance of longer chain aliphatics including oligomethylene units. With more of KP2005 available for study, we have acquired the high-resolution 2D NMR spectra from this sample than from KP2004. However, superposition of correlation spectroscopy (COSY) and heteronuclear single quantum coherence (HSQC) NMR spectra of SOA KP2004 and KP2005 (data not shown) showed extensive congruence of cross peaks. Connectivity information obtained from complementary homonuclear (COSY, TOCSY, JRES) and heteronuclear (HSQC, heteronuclear multiple bond correlation (HMBC), DEPT-HSQC) 2D NMR spectra were employed to improve the NMR resonance assignments in SOA KP2005. These findings are discussed in detail in the Supporting Information (cf. Figures 4 and 5 and Figures SI-8-SI-16 in the Supporting Information). The very abundant 2D NMR cross peaks across large chemical shift areas confirmed at first the very rich and diverse chemistry of SOA KP2005 (cf. Figures 4 and 5 and Figures SI-8-SI-15 in the Supporting Information). In contrast to terrestrial and aqueous natural organic matter, KP2005 showed lesser amounts of open chain branched methyl groups (δH < 1.3 ppm). Instead, a very sizable fraction of isolated methyl (i.e., bound to quaternary carbon) was recognized from 1H,13C DEPT-HSQC NMR spectra (Figure 5) and conformed to previously suggested fused alicyclic ring structures with heteroatoms three and more bonds away. Methoxy occurred in the form of aromatic and aliphatic methyl esters and ethers in descending20 order RalCOOCH3 . RalOCH3 > RarOCH3 > RarCOCH3 (Figure SI-12 in the Supporting Information). We considered the abundance of methyl esters found in KP2005 extracts as an indirect confirmation of sulfate esters present in original SOA, which underwent methanolysis prior to NMR measurement during sample preparation. Analogous reactions were confirmed for sulfated Suwannee river fulvic acid subjected to methanolysis (cf. Figure SI-16 in the Supporting Information). Aliphatic spin systems often extended through 3-5 bonds (Figures SI-9-SI-11, SI-14, and SI-15 in the Supporting Information) reflecting the disruption by abundant carbonyl derivative (COX) units and quaternary carbon (with carbon

Figure 4. NMR spectra of methanolic SOA extract KP2005, aliphatic section between δH ) 0.75-3.0 ppm; (A) 1H NMR spectrum of year 2004 SOA KP2004 (orange); (B) 1H NMR spectrum of year 2005 SOA KP2005 (purple); (C) 1H,1H J-resolved (JRES) NMR spectrum of KP2005; (D) overlay of 1H,1H COSY (red), and 1H,1H total correlation spectroscopy (TOCSY) NMR spectra (green) of KP2005. For assignment of key substructures and detail description, please refer to Figures SI-8-SI-11 in the Supporting Information. Table 2. 1H NMR Section Integrals (% of Total 1H NMR Integral) of KP2004 and KP2005a δ(1H) [ppm]

SOA KP2004 [%]

SOA KP2005 [%]

attribution

6.5-10 5.5-6.5 3.1-5.5 1.95-3.1 0-1.95 SUM OCH3

1.8 1.1 24.2 28.6 44.4 100.0 3.3

4.7 1.4 28.2 27.5 38.2 100.0 9.9

Har double bonds OCH-OH-CO H-CC OCH3

OCH3 denotes NMR section integral ranging from δH ) 3.65-3.90 ppm (included in OCH-O section integral δH ) 3.1-5.5 ppm), which is attributed to aliphatic and some aromatic methyl esters (cf. text). a

only environments, abundant single heteroatom and also double heteroatom substitution). The typical aliphatic chemical environments within SOA were heteroatom-substituted functional groups adjacent Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

8023

Figure 5. (A) Overlay of HSQC (yellow/gray) and HMBC (purple/blue) NMR spectra of SOA KP2005, with chemical shift range δC ) 120-160 ppm truncated because of the near absence of cross peaks (cf. Figure SI-14B in the Supporting Information). Extensive aliphatic branching is indicated by both HSQC and HMBC cross peaks; quaternary carbons with δC ) 80-90 do not exhibit HSQC cross peaks and likely indicate heteroatom (O, S, N) substituents. The yellow area corresponds to panels B and C aliphatic area; the carbon downfield region is highlighted in green for improved discrimination; (C) overlay of 1H,13H HSQC (gray) and 1H,13H HMBC NMR spectra (purple) of KP2005, aliphatic section; (B) overlay of edited 1H,13C DEPT-HSQC-CH3 (red) and 1H,13C DEPT-HSQC-CH2 (blue) of KP2005, aliphatic section. For assignment of key substructures and detailed description, please refer to Figures SI-12-SI-15 in the Supporting Information.

to highly branched aliphatics, likely in the form of strongly coupled fused alicyclic ring spin systems as deduced from proton chemical shifts in the homonuclear 2D NMR spectra (cf. Figure 4 and Figures SI-9-SI-11 in the Supporting Information) and strongly corroborated by heteronuclear 2D NMR spectra (cf. Figure 5 and Figures SI-12-SI15 in the Supporting Information. Aromatics were typically found to be highly substituted, and electron withdrawing COX and (O)NOx substitution was considerably more common than the presence of electron donating oxygen-containing functional groups (OH, OR, OSOnR) and neutral substitution (aliphatic carbon), which contributed less than 10% of aromatic substitution. 8024

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

In conclusion, the NMR properties of SOA KP2005 confirmed in many aspects the terpenoid- and highly oxidized polyfunctional molecules proposed earlier39,54 as a consequence of successive evaporation, (photo)oxidation, and recondensation cycles. In nature, CHO (and low volatility CHOS, CHNO, and CHNOS) precursor molecules larger than those typically employed in model SOA formation studies could have produced the increased proportion of alkyl carbon observed in the NMR spectra of KP2005. The large proportion of carbonyl derivatives present in SOA KP2005 (ketones and COX derivatives; cf. Figure 5 and

Figures SI-14 and SI-15 in the Supporting Information) could have been easily introduced into analogous intermediates by the rich photo and aqueous chemistry available under typically vigorous ambient conditions of SOA formation. High-resolution NMR derived definition of carbon and hydrogen chemical environments within SOA as shown here will offer improved assessment of SOA source attribution and chemical reaction pathways leading to the complex mixture of products observed. SOA exhibit strong reactivity with biomolecules, and the knowledge of the rich and diverse SOA chemistry will contribute to a better understanding of health aspects of atmospheric particulate organic matter.20,55,56 Possible Formation Mechanism of Sulfate Esters. Previously, bulk laboratory experiments showed that concentrated sulfuric acid can transform isoprene into nonvolatile higher molecular weight species.38 Direct uptake of isoprene and R-pinene on highly acidic particles was also proven in smog chamber experiments.39 Acidic seed aerosols have also been shown to enhance secondary organic aerosol (SOA) yields in controlled photooxidation experiments in smog chambers from R-pinene,40-43 β-pinene,,33 or isoprene.39,44,45 Similar results have been obtained for semivolatile photooxidation products like pinonaldehyde44,45 or glyoxal.46 It has been shown very recently that in the absence of acidic seeds, the very same effect can be induced by simultaneous photooxidation of aerosol precursors and SO2.34 Recent mass spectrometric evidence suggests that formation of organosulfates (sulfate esters and their derivatives) formation is involved in enhanced SOA yields.34,35 Some compounds have been identified by their abundant [M - H]- ions in ESI(-)-MS mass spectra, by their fragment ions 96 (SO4-) or 97 (HSO4-) and 80 (SO3-) with typical isotope peaks upon collisional activation of these ions, similar to authentic sulfate ester standards.47 The number and abundance of observed sulfate esters have been found to increase with increasing acidity of the seed aerosol. The possibility of these sulfate esters being formed as artifacts in the electrospray or in any part of the mass spectrometer has been systematically ruled out by control experiments. Such compounds were also detected in limonene photooxidation experiments in which no apparent increase in SOA yield was observed with increasing seed acidity.42 High NOx mixing ratios have been shown to strongly suppress organosulfate formation possibly because organic acid formation is favored at the expense of esterification reactions.34,47 In addition to detailed smog-chamber studies, a few ambient filter measurements indicated the presence (38) Limbeck, A.; Kulmala, M.; Puxbaum, H. Geophys. Res. Lett. 2003, 30, No1996. (39) Liggio, J.; Li, S. M.; Brooks, J. R.; Mihele, C. Geophys. Res. Lett. 2007, 34, L05814. (40) Gao, S.; Keywood, M.; Ng, N. L.; Surratt, J. D.; Varutbangkul, V.; Bahreini, R.; Flagan, R. C.; Seinfeld, J. H. J. Phys. Chem. 2004, 108, 10147–10164. (41) Gao, S.; Ng, N. L.; Keywood, M.; Varutbangkul, V.; Bahreini, R.; Nenes, A.; He, J.; Yoo, K. Y.; Beauchamp, J. L.; Hodyss, R. P.; Flagan, R. C.; Seinfeld, J. H. Environ. Sci. Technol. 2004, 38, 6582–6589. (42) Iinuma, Y.; Boge, O.; Gnauk, T.; Herrmann, H. Atmos. Environ. 2004, 38, 761–773. (43) Kleindienst, T. E.; Edney, E. O.; Lewandowski, M.; Offenberg, J. H.; Jaoui, M. Environ. Sci. Technol. 2006, 40, 3807–3812. (44) Liggio, J.; Li, S. M. Geophys. Res. Lett. 2006, 33, L13808. (45) Liggio, J.; Li, S. M. J. Geophys. Res. 2006, 111, D24303. (46) Kroll, J. H.; Ng, N. L.; Murphy, S. M.; Varutbangkul, V.; Flagan, R. C.; Seinfeld, J. H. J. Geophys. Res. 2005, 110, D23207. (47) Surratt, J. D.; Gomez-Gonzalez, Y.; Chan, A. W. H.; Vermeylen, R.; Shahgholi, M.; Kleindienst, T. E.; Edney, E. O.; Offenberg, J. H.; Lewandowski, M.; Jaoui, M.; Maenhaut, W.; Claeys, M.; Flagan, R. C.; Seinfeld, J. H. J. J. Phys. Chem. A 2008, 112, 8345–8378.

of a few specific sulfate esters using LC-ESI-MS35,34,42 as well as the presence of organosulfur functional groups by FT-IR.48,49 Organosulfates have also been found in the water-soluble fraction of urban and rural aerosols.50 This finding implies a possible formation mechanism for organosulfates that was also suggested by size-distribution measurements in a recent quantitative study:31 organosulfates likely form in heterogeneous reactions between partitioned semivolatile carbonyl compounds and sulfuric acid formed in gas-phase photooxidation reactions. Under the conditions of photochemical smog at large preexisting organic aerosol mass concentrations, this process is favored and even more volatile species such as pinonaldehyde can measurably partition into the aerosol phase.51,52 It was shown that intense photochemistry which is conducive to producing high concentrations of SOA also triggers high rates of gas-phase oxidation of SO2 to sulfuric acid which condenses immediately onto preexisting particles.53 Given that reactive organic species are present on the particle surfaces in relatively large abundance, these esterification reactions may proceed quite readily in the aerosol phase and successfully compete with partial neutralization reactions with gaseous ammonia depending on the SO2 pollution rate. The reactive organic species consumed in these heterogeneous reactions are replenished from the gas phase in an attempt to reestablish the partitioning equilibrium, which is likely the ratelimiting step to produce SOA. This mechanism is implicitly supported by a recent smog chamber study in which seed aerosols were not applied but just SO2 was added and sulfate ester formation has been observed.34 To verify this hypothesis, sulfuric acid was first added to a natural organic matter of low nitrogen content (Suwannee river fulvic acid; SuwFA), and the FTICR-MS spectra were compared to the nontreated sample (see Figure SI-4a,b in the Supporting Information). The use of C18 solid phase extraction did not alter the amount of detected CHOS compounds (Figure SI-4c in the Supporting Information). The CHOS compounds produced especially originated from low oxygenated and aliphatic types of compounds (Figure SI-4d in the Supporting Information). The nitrogen rich sample KPBMB2005 aerosol was also treated with sulfuric acid, and the obtained FTICR-mass spectra were compared to the untreated sample described above (Figure 3b,c). Similar to the sulfonation of SRW, sulfuric acid reacted intensively with the CHO and CHNO compounds of SOA to form the corresponding CHOS and CHNOS compounds, respectively (48) Blando, J. D.; Porcja, R. J.; Li, T.-H.; Bowman, D.; Lioy, P. J.; Turpin, B. J. Environ. Sci. Technol. 1998, 32, 604–613. (49) Gilardoni, S.; Russell, L. M.; Sorooshian, A.; Flagan, R. C.; Seinfeld, J. H.; Bates, T. S.; Quinn, P. K.; Allan, J. D.; Williams, B.; Goldstein, A. H.; Onasch, T. B.; Worsnop, D. R. J. Geophys. Res. 2007, 112, D10S27. (50) Romero, F.; Oehme, M. J. Atmos. Chem. 2005, 52, 283–294. (51) Odum, J. R.; Hoffmann, T.; Bowman, F.; Collins, D.; Flagan, R. C.; Seinfeld, J. H. Environ. Sci. Technol. 1996, 30, 2580–2585. (52) Andersson-Sko ¨ld, Y.; Simpson, D. J. Geophys. Res. 2001, 106, 7357–7374. (53) Alkezweeny, A. J.; Laulainen, N. S. J. Appl. Meteorol. 1981, 20, 209–212. (54) Surratt, J. D.; Kroll, J. H.; Kleindienst, T. E.; Edney, E. O.; Claeys, M.; Sorooshian, A.; Ng, N. L.; Nga, L.; Offenberg, J. H.; Lewandowski, M.; Jaoui, M.; Flagan, R. C.; Seinfeld, J. H. Environ. Sci. Technol. 2007, 41, 517–527. (55) Mang, S. A.; Henricksen, D. K.; Bateman, A. P.; Andersen, M. P. S.; Blake, D. R.; Nizkorodov, S. A. J. Phys. Chem. 2008, 112, 8337–8344. (56) De Haan, D. O.; Corrigan, A. L.; Smith, K. W.; Stroik, D. R.; Turley, J. J.; Lee, F. E.; Tobert, M. A.; Jimenez, J. L.; Cordova, K. E.; Ferrell, G. R. Environ. Sci. Technol. 2009, 43, 2818–2824.

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

8025

(Figure SI-5 in the Supporting Information and Table 1). A detailed description of the influence of sulfuric acid on the CHO, CHNO, CHOS, and CHNOS molecular distributions (Figure SI-6 in the Supporting Information) confirmed the sulfonation process of the CHO and CHNO compounds in sample KPBMB2005. Molecules of lower masses and fewer double bond equivalents were preferentially involved in the formation of the sulfonated SOA that forms an extended and very significant class of compounds with particular chemistry, which has only been partially described by means of targeted analysis in the previous literature. The hypothesis of SOA reactions in the condensed phase is further corroborated by the comparative study of a consecutive hail rainwater event. Washout of the particles and dissolution of the SOA in rainwater can take place during extended residence times and leads to very complex van Krevelen diagrams with a large number of CHOS molecules. Hail samples, in which a cryosampling of the semivolatiles and the particle phase takes place without an opportunity for sulfuric acid formation or of sulfonation reactions, showed a rather weak abundance of sulfate esters (Figure SI-7 in the Supporting Information). To summarize, sulfuric acid and SO2, which are produced by many anthropogenic or natural sources, such as industrial as well as ship emissions and, e. g., volcanic eruptions that may contribute significantly to the increase of sulfate esters and thus affect the (CHO + CHNO)/(CHOS + CHNOS) balance. Sulfonations will change the polarity of the SOA molecules, providing them with modified biological properties, especially through an increase of surface active properties. Studies are ongoing in this important field of environmental health. Here, comprehensive analysis by means of NMR and FTICR mass spectrometry will enable thorough characterization of carbon and hydrogen as well as

8026

Analytical Chemistry, Vol. 82, No. 19, October 1, 2010

heteroatom chemical environments in SOA, enabling improved options for SOA source assignment, evolution of reaction mechanisms, and assessment of SOA adverse effects on human health. In addition, improved comprehension of SOA impact upon regional and worldwide climate patterns might derive from a timely acknowledgment of SOA chemistry. ACKNOWLEDGMENT This work on Hungarian samples was enabled through an EU FP6 ACCENT Access to Infrastructures Project 2004/2005 on “Molecular level characterization of aerosol extracts” providing access to the samples and the mobility exchange grants for P.S.K. and I.G. This work was also supported in part by the Canadian Federal Program on Energy Research and Development under PERD 2.1.1. Project “Support the Development of Technological and Other Measures to Control and Reduce Emissions of Particulate Matter”. Environment Canada is thanked for the mobility exchange grants for P.S.-K. and E.D.-Z., respectively. H.Y. is supported by Max Planck Institute for Chemistry and is grateful to Y. Liu and M. Shao of Peking University for providing the wheat straw combustion sample. SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review June 21, 2010. Accepted September 14, 2010. AC101444R