Anal. Chem. 2006, 78, 5354-5361
Thermal Desorption/Pyrolysis Coupled with Photoionization Time-of-Flight Mass Spectrometry for the Analysis of Molecular Organic Compounds and Oligomeric and Polymeric Fractions in Urban Particulate Matter Thorsten Streibel,*,†,‡ Jochen Weh,‡ Stefan Mitschke,†,‡ and Ralf Zimmermann†,‡,§
Analytical Chemistry, Institute of Physics, University of Augsburg, D-86159 Augsburg, Germany, Institute of Ecological Chemistry, GSFsNational Research Centre for Environment and Health, D-85764 Neuherberg, Germany, and Environmental Chemistry, BIfAsBavarian Institute of Applied Environmental Research and Technology GmbH, D-86167 Augsburg, Germany
Atmospheric aerosols are subject to be responsible for human health effects. In this context, besides mass and number concentration of particles, their chemical composition has gained interest recently. However, knowledge about the organic content of particulate matter is still relatively scarce; i.e., only 10-40% of compounds present in the aerosol are as yet identified. By means of a newly developed measurement technique, thermal desorption/ photoionization time-of-flight mass spectrometry (TOFMS), organic species evolved from urban aerosol samples collected at Augsburg, Germany, are analyzed. Thereby, compounds desorbed according to a temperature protocol following procedures for OC/EC analysis (120, 250, and 340 °C as desorbing temperatures) are ionized by soft, fragmentationless resonance multiphoton ionization (REMPI) and single photon ionization (SPI), respectively. With REMPI-TOFMS, a large variety of PAH is detectable. A comprehensive analysis is enabled by adding SPI-TOFMS, which gives access to aliphatic and carbonylic hydrocarbons as well as alkanoic acids and esters. Analysis of the data showed a high abundance of phenol and guiacol as well as retene, which are known markers for wood combustion. Similar patterns were found with ash from spruce wood combustion. An increase of volatile substances at 340 °C gave rise to the suggestion that these compounds are re-formed by pyrolytic decomposition reactions from oligomeric, polymeric, and polyfunctional oxygenated species. This was corroborated by the investigation of the behavior of cellulose acetate, which exhibited a similar pattern in its SPI-TOFMS spectrum at 340 °C as the aerosol. More thorough investigations of urban aerosol and source material with respect to problems such as the mass closure of carbonaceous material, indications for source apportionment, and allotment of organic species on a molecular level to fractions of organic and 5354 Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
elemental carbon seem feasible with this measurement method. Besides its influence on atmosphere and climate, the effects of atmospheric aerosol on public health have focused attention to a more thorough investigation of particulate matter (PM) in recent years. Several epidemiological studies to date show a significant relationship between PM exposure and health effects.1 For instance, particles with diameters of less than 2.5 µm (referred to as PM2.5) are strongly associated with mortality and other consequences such as cardiopulmonary diseases.2-4 However, it has not been possible to establish a correlation between those health effects and distinct chemical components of ambient aerosol, although intensive research has been carried out in the past decade concerning these issues. Organic and elemental carbon is among the major chemical constituents of atmospheric aerosol.5 For PM2.5, the overall concentration amounts to 1-12 µg/m3, which makes up roughly up to 50% of the total particle mass.6,7 So far, several investigations suggest a significant contribution of organic components to the toxicity of particulate matter.8,9 For example, particle-associated oxygenated organic * Corresponding author. E-mail:
[email protected]. † University of Augsburg. ‡ National Research Centre for Environment and Health. § Bavarian Institute of Applied Environmental Research and Technology GmbH. (1) Health aspects of air pollution with particulate matter, ozone and nitrogen dioxide; WHO: Bonn, 2003. (2) Hoek, G.; Brunekreef, B.; Goldbohm, S.; Fischer, P.; van den Brandt, P. A. Lancet 2002, 360, 1203-1209. (3) Pope, C. A.; Burnett, R. T.; Thun, M. J.; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D. J. Am. Med. Assoc. 2002, 287, 1132-1141. (4) Janssen, N. N.; Schwartz, J.; Zanobetti, A.; Suh, H. H. Environ. Health Perspect. 2002, 110, 43-49. (5) Huntzicker, R. L., Johnson, R. L., Shah, J. J., Cary, R. A., Eds. Analysis of organic and elemental carbon in ambient aerosol by thermal-optical method; Plenum Press: New York, 1982. (6) Turpin, B. J.; Saxena, P.; Andrews, E. Atmos. Environ. 2000, 34, 29833013. (7) Tolocka, M. P.; Solomon, P. A.; Mitchell, W.; Norris, G. A.; Gemmill, D. B.; Wiener, R. W.; Vanderpool, R. W.; Homolya, J. B.; Rice, J. Aerosol Sci. Technol. 2001, 34, 88-96. 10.1021/ac060227y CCC: $33.50
© 2006 American Chemical Society Published on Web 06/27/2006
compounds and polycyclic aromatic hydrocarbons (PAH) are suspected to induce oxidative stress in lung cells, which can be associated with inflammatory processes and thus also with heart and lung diseases. Major compound classes building up the organic matter in urban aerosols besides these aromatic species are aliphatic hydrocarbons and organic acids.10 A reliable method for quantifying the organic content of particulate matter and making a coarse distinction of the respective allotment of organic and elemental carbon is thus of great value for an estimation of the carbonaceous portion of urban aerosol. This led to the introduction of sum parameters for description of the carbonaceous material in particulate matter by the National Institute for Occupational Safety and Health (NIOSH) such as the elemental carbon content (EC, in µg/m3), the sum content of organic compounds (OC, in µg/m3), and the total carbon content (TC, in µg/m3; TC ) OC + EC). Very common is the use of dimensionless ratios, such as the EC/TC value (i.e., the ratio of elemental carbon mass to total carbon mass) or the EC/OC value (i.e., the ratio of elemental carbon mass to organic carbon mass).11,12 Methods for the determination of EC/OC or EC/TC values are thermal desorption of semivolatile and low volatile hydrocarbons (OC) and combustion of the nonvolatile EC using a defined temperature protocol. All thermal desorption techniques use PMloaded quartz fiber filter samples for analysis. The evolving carbonaceous species are detected integrally as a function of the temperature. The definition of the separation line between OC and EC depends on the respective method, in particular on the applied temperature protocol and the changing point from the thermal desorption mode with inert gas (OC analysis) and the combustion mode with oxygen (EC analysis). An overview of different methods and their variations in determination of OC and EC values is given in ref 13. As an example, using the IMPROVE (interagency monitoring of protected visual environments) method,14 first organic carbon species are desorbed at different temperatures in a pure helium atmosphere. The evolving carbon compounds are oxidized to carbon dioxide by passing through a catalytic medium (MnO2). The generated CO2 is subsequently reduced to methane by means of a nickel catalyst. The methane is then quantified by a flame ionization detector. In a second step, the atmosphere is changed to a mixture of 2% oxygen in helium to oxidize and volatilize elemental carbon. Often pyrolytic reactions of low volatile organic material in the absence of oxygen evoke complex reactions including carbonization. The carbonized fraction of the low volatile organic material appears to be elemental carbon.14 The IMPROVE method corrects for this effect using a laser light transmission measurement during the oxidizing phase (8) Tolbert, P. E.; Klein, M.; Metzger, K. B.; Peel, J.; Flanders, W. D.; Todd, K.; Mulholland, J. A.; Ryan, P. B.; Frumkin, H. J. Exposure Anal. Environ. Epidemiol. 2000, 10, 446-460. (9) Monn, C.; Becker, S. Toxicol. Appl. Pharmacol. 1999, 155, 245-252. (10) Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R. Environ. Sci. Technol. 1993, 27, 636-651. (11) Dahmann, D.; Fricke, H.-H.; Bauer, H.-D. Occup. Hyg. 1996, 3, 255-262. (12) Cadle, S. H. Environ. Sci. Technol. 1999, 33, 2328-2339. (13) Schmid, H.; Laskus, L.; Abraham, H. J.; Baltensperger, U.; Lavanchy, V.; Bizjak, M.; Cachier, H.; Crow, D.; Chow, J.; Gnauk, T.; Even, A.; Brink, H. M. t.; Giesen, K.-P.; Hitzenberger, R.; Hueglin, C.; Maenhaut, W.; Pio, C.; Carvalho, A.; Putaud, J.-P.; Toom-Sauntry, D.; Puxbaum, H. Atmos. Environ. 2001, 35, 2111-2121. (14) Chow, J. C.; Watson, J. G.; Pritchett, L. C.; Pierson, W. R.; Frazier, C. A.; Purcell, R. G. Atmos. Environ. 1993, 27A, 1185-1201.
to distinguish between (colorless) OC and (black) EC components. With this thermal/optical method, several different fractions of OC and EC may be defined depending on the number of temperature steps in the method.13-15 In contrast, the thermocoulometric method16 volatilizes organic carbon in a nitrogen atmosphere and elemental carbon in oxygen. Oxidation of volatilized material takes place utilizing a CuO catalyst. The resulting CO2 is transferred into an electrolyte solution (e.g., Ba(OH)2), and the thus increasing pH is titrated back to the original value, yielding a quantitative measure of carbonaceous material. Recently, the practicability of single-particle aerosol mass spectrometry for OC/EC analysis has been investigated.17 However, the organic fraction of urban aerosols constitutes a complex mixture of a multitude of different compounds, which are present in a respectively small individual concentration range from 0.1 to 10 ng/m3.18 Up to now, only 10-40% of the organic matter has been identified on a molecular level.19,20 As a consequence, there is a significant missing mass fraction with respect to the organic content. This mass closure is often accounted for by oligomeric and polymeric species as well as polyfunctional oxygenated compounds.19-22 Model approaches showed the significance of possible reaction paths for the formation of such polymers, which resemble humic-like substances in their behavior.23 In this context, species formed by oxidation processes from primary aerosols after their release to the atmosphere, the socalled secondary organic aerosols, have gained special notice recently. With the aforementioned methods for OC and EC determination, no information on the presence of individual organic compounds or polymeric fractions in a given OC or EC fraction is possible. On the other hand, the identification and analysis of organic compounds using conventional methods such as gas chromatography/mass spectrometry (GC/MS)10,24 or high-pressure liquid chromatography25 require extensive sample preparation and clean up as well as a long analysis time in the laboratory and does not address the polymeric fractions. Other chromatographic methods such as comprehensive GC (GC×GC)26-28 are used for comprehensive characterization of the GC-detectable compounds, of which up to 10 000 different species could be detected. (15) Birch, M. E.; Cary, R. A. Aerosol Sci. Technol. 1996, 25, 221. (16) VDI/DIN-Handbuch Reinhaltung der Luft; Beuth Verlag GmbH: Berlin, 1996; Vol. 4. (17) Ferge, T.; Karg, E.; Schro¨ppel, A.; Coffee, K. R.; Tobias, H. J.; Frank, M.; Gard, E. E.; Zimmermann, R. Environ. Sci. Technol. 2006, 40 (10), 33273335. (18) Schnelle-Kreis, J.; Welthagen, W.; Sklorz, M.; Zimmermann, R. J. Sep. Sci. 2005, 28, 1648-1657. (19) Gelencse´r, A.; Hoffer, A.; Molna´r, A.; Kriva´csy, Z.; Kiss, G.; Me´sza´ros, E. Atmos. Environ. 2000, 34, 823-831. (20) Po ¨schl, U. Angew. Chem., Int. Ed. 2005, 44, 7520-7540. (21) Kalberer, M.; Paulsen, D.; Sax, M.; Steinbacher, M.; Dommen, J.; Prevot, A. S. H.; Fisseha, R.; Weingartner, E.; Frankevich, V.; Zenobi, R.; Baltensberger, U. Science 2004, 303, 1659-1662. (22) Kriva´csy, Z.; Kiss, G.; Varga, B.; Galambos, I.; Sa´rva´ri, Z.; Gelencse´r, A.; Molna´r, A.; Fuzzi, S.; Facchini, M. C.; Zappoli, S.; Andracchio, A.; Alsberg, T.; Hansson, H. C.; Persson, L. Atmos. Environ. 2000, 34, 4273-4281. (23) Tolocka, M. P.; Jang, M.; Ginter, J. M.; Cox, F. J.; Kamens, R. M.; Johnston, M. V. Environ. Sci. Technol. 2004, 38, 1428-1434. (24) Cass, G. R. Trends Anal. Chem. 1998, 17, 356-366. (25) Colombini, M. P.; Fuoco, R.; Giannarelli, S.; Termine, M.; Abete, C.; Vincentini, M.; Berti, S. Microchem. J. 1998, 59, 228-238. (26) Haglund, P.; Harju, M.; Danielsson, C.; Marriott, P. J. Chromatogr., A 2002, 962, 127-134.
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
5355
Additional methods are desirable, which circumvent the timeconsuming sample preparation and enable the researcher to easily analyze smaller sample quantities. In the recent past, laser desorption/ionization (LDI) as well as two-step laser desorption photoionization (LD-PI) mass spectrometry29-36 has proven to be a suitable candidate for such a purpose as it allows chemical analysis within minutes and without any sample preparation. LDI recently also allowed the identification of oligomeric and polymeric species as components of atmospheric aerosol.21 Pyrolysis as well as thermal analysis in combination with mass spectrometry or gas chromatography (Py-MS or Py-GC/MS) is a well established and reliable method for the analytical investigation of polymers,37-40 providing one of the classic approaches for degradation of large molecules and subsequent analysis of the pyrolytic products. Therefore, an approach that consists of the combination of thermal treatment of the analyte with subsequent mass spectrometric detection of the evolved products seems suitable for the investigation of oligomeric and polymeric content of particulate matter. Furthermore, referring to temperature protocols from OC/EC in the low-temperature range and investigating the evaporated gas phase that is produced in this process by means mass spectrometry seems reasonable. The resonance enhanced multiphoton ionization (REMPI) is a selective and soft means of ionization, offering the detection of only those molecules with appropriate electronic transitions, thus representing an ideal method for analysis of aromatic trace compounds in complex samples.35,41-47 However, REMPI, while providing high sensitivity for aromatic compounds, is not susceptible to aliphatic, carbonylic, and carboxylic compounds, which represent a significant portion of the (27) Xu, X.; van Stee, L. L. P.; Williams, J.; Beens, J.; Adahchour, M.; Vreuls, R. J. J.; Brinkman, U. A. T.; Lelieveld, J. Atmos. Chem. Phys. Discuss. 2003, 3, 1139-1181. (28) Welthagen, W.; Schnelle-Kreis, J.; Zimmermann, R. J. Chromatogr., A 2003, 1019, 233-249. (29) Haefliger, O. P.; Zenobi, R. Anal. Chem. 1998, 70, 2660-2665. (30) Haefliger, O. P.; Bucheli, T. D.; Zenobi, R. Environ. Sci. Technol. 2000, 34, 2178-2183. (31) Haefliger, O. P.; Bucheli, T. D.; Zenobi, R. Environ. Sci. Technol. 2000, 34, 2184-2189. (32) Kalberer, M.; Morrical, B. D.; Sax, M.; Zenobi, R. Anal. Chem. 2002, 74, 3492-3497. (33) Specht, A. A.; Blades, M. W. J. Am. Soc. Mass Spectrometry 2003, 14, 562570. (34) Elsila, J. E.; de Leon, N. P.; Zare, R. N. Anal. Chem. 2004, 76, 2430-2437. (35) Hauler, T. E.; Boesl, U.; Kaesdorf, S.; Zimmermann, R. J. Chromatogr., A 2004, 1058, 39-49. (36) Ferge, T.; Mu ¨ hlberger, F.; Zimemrmann, R. Anal. Chem. 2005, 77, 45284538. (37) Meuzelaar, H. L. C.; Windig, W.; Harper, A. M.; Huff, S. M.; McClennen, W. H.; Richards, J. M. Science 1984, 226, 268-274. (38) Blazso´, M. J. Anal. Appl. Pyrolysis 1997, 39, 1-25. (39) Wampler, T. P. J. Chromatogr., A 1999, 842, 207-220. (40) Wampler, T. P. J. Anal. Appl. Pyrolysis 2004, 71, 1-12. (41) Lubman, D. M., Ed. Lasers and Mass Spectrometry; Oxford University Press: New York, 1990. (42) Oser, H.; Thanner, R.; Grotheer, H.-H. Combust. Sci. Technol. 1996, 116117, 567-582. (43) Gittins, C. M.; Castaldi, M. J.; Senkan, S. M.; Rohlfing, E. A. Anal. Chem. 1997, 69, 286-293. (44) Boesl, U. J. Mass Spectrom. 2000, 35, 289-304. (45) Grotheer, H.-H.; Nomayo, M.; Pokorny, H.; Thanner, R.; Gullett, B. K. Trends Appl. Spectrosc. 2001, 3, 181-206. (46) Cao, L.; Mu ¨ hlberger, F.; Adam, T.; Streibel, T.; Wang, H. Z.; Kettrup, A.; Zimmermann, R. Anal. Chem. 2003, 75, 5639-5645. (47) Dorfner, R.; Ferge, T.; Yeretzian, C.; Kettrup, A.; Zimmermann, R. Anal. Chem. 2004, 76, 1368-1402.
5356
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
organic mass in ambient PM. An alternative to multiphoton ionization in this case is single photon ionization (SPI) with VUV photons.48-51 Due to medium selectivity by virtue of the ionization threshold, it is suitable for both aromatic and aliphatic organic compound classes. The typically used radiation (118 nm, 10.5 eV) is just above the ionization threshold of most organic compounds, making it a soft ionization method, which produces little or no fragmentation. SPI was employed in the aerosol field for real-time detection of molecules desorbed from individual particles as well as depth profiling of organic compounds in single particles.52 Also, characterization of thermodesorbed organic compounds from particles impacted on a heated probe has been reported.53 In this paper, a novel approach is presented, which consists of the utilization of thermal desorption of aerosol samples and subsequent analysis of the evolved gas phase by means of soft photoionization time-of-flight mass spectrometry (PI-TOFMS),54 including both REMPI and SPI as ionization techniques. Thereby, thermal desorption applying a temperature protocol similar to OC/ EC approaches in the low-temperature range coupled to a REMPI/ SPI mass spectrometer system for analysis of the in the process evolving organic compounds is presented. The measurement technique is described subsequently, and first measurements of urban aerosol samples utilizing this method are presented. EXPERIMENTAL SECTION Urban aerosol sampled with a sequential sampler (Partsol-Plus model 2025, Rupprecht & Patashnick, New York) on quartz fiber filters (T293, Munktell, Grycksbro, Sweden) at a flow of 1 m3/h has been investigated. Sampling period was 24 h in every case. Four ambient urban aerosol samples (PM2.5) from a monitoring station in Augsburg have been investigated in the framework of this study (samples 1-4). The monitoring station is located on one hand inside the city limits, but on the other hand at a rather remote location with respect to traffic and industrial emission sources, thus representing a typical inner city residential location. The samples have been taken during periods with rather cold and wet weather conditions. The loaded filters were stored at -18 °C until analysis. Prior to analysis, the filters were cut into stripes, each representing 1 m3 of sampled air. Three of these filter stripes were used in every measurement cycle. For comparison and validation purposes, a sample of ash from combustion of spruce wood and pure cellulose acetate has been measured using the same temperature protocol. Both samples resembled the urban aerosol in weight. The experimental setup for the thermal desorption/laser ionization technique presented here is shown in Figure 1. The filter stripes were put in a quartz glass liner, which is inserted into a GC injector heatable up to 350 °C. In principle, thermal desorption could be carried out at any given temperature below (48) Nir, E.; Hunziker, H. E.; Vries, M. S. Anal. Chem. 1999, 71, 1674-1678. (49) Butcher, D. J. Microchem. J. 1999, 62, 354-362. (50) Mu ¨ hlberger, F.; Zimmermann, R.; Kettrup, A. Anal. Chem. 2001, 73, 35903604. (51) Mu ¨ hlberger, F.; Wieser, J.; Morozov, A.; Ulrich, A.; Zimmermann, R. Anal. Chem. 2005, 77, 2218-2226. (52) Woods, E.; Smith, G. D.; Miller, R. E.; Baer, T. Anal. Chem. 2002, 74, 16421649. (53) Sykes, D. C.; Woods, E.; Smith, G. D.; Baer, T.; Miller, R. E. Anal. Chem. 2002, 74, 2048-2052. (54) Streibel, T.; Mitschke, S.; Welthagen, W.; Adam, T.; Zimmermann, R. Organohalogen Compd. 2005, 67, 2630.
Figure 1. Experimental setup of the newly developed thermal desorption/photoionization time-of-flight mass spectrometry instrument.
this maximum limit. Three distinct temperatures (120, 250, and 340 °C in accordance to OC/EC measurements methods13) have been applied. Temperature steps were adjusted by means of maximum possible heating rate (∼80 K/s). It took ∼1-2 s to move to the next temperature. Desorbed compounds are guided through a deactivated fused-silica capillary of 200-µm i.d., and helium is used as carrier gas. The capillary acts as the restrictor between the vacuum of the ion source and ambient pressure of the thermal desorption unit. Moreover, the capillary serves as transfer line for the evolved gas-phase species from the desorption process into the ion source of the reflectron time-of-flight mass spectrometer (custom-made device by Stefan Kaesdorf, Munich, Germany), which provides a mass resolution R50% of 1800 m/z measured at 92 m/z. The TOFMS is part of a mobile measurement device allowing (quasi-) simultaneous application of REMPI and SPI, respectively, in an alternating manner.46,50,55,56 Evolved compounds were detected until no SPI- or REMPI-TOFMS signal could be observed anymore, thus ensuring that all desorbed species have been detected before moving on to the next temperature step. All single mass spectra at every temperature step were added up subsequently. In brief, a frequency tripled Nd:YAG laser (Surelite-III, Continuum, Santa Clara, CA), which yields 3-5-ns pulses at 355 nm with an average energy of 225 mJ and a repetition rate of 10 Hz, is split in two partial beams. The first of these beams is used to pump an optical parametric oscillator (OPO; GWU Lasertechnik, Erftstadt, Germany). The OPO system equipped with a frequency doubling unit yields tunable UV laser pulses ranging from 220 to 355 nm for REMPI ionization. In this study, a REMPI wavelength of 275 nm has been applied, which is quite favorable for the (55) Adam, T.; Streibel, T.; Mitschke, S.; Mu ¨ hlberger, F.; Cao, L.; Baker, R. R.; Zimmermann, R. J. Anal. Appl. Pyrolysis 2005, 74, 454-464. (56) Mitschke, S.; Adam, T.; Streibel, T.; Baker, R. R.; Zimmermann, R. Anal. Chem. 2005, 77, 2288-2296.
detection of PAH. The line width of the laser is 0.07 nm, and the pulse energy ranges between 0.2 and 0.4 mJ (depending on the wavelength). The second beam (∼10% of the initial 355-nm beam) is focused by a quartz lens (f ) 100 mm) through a quartz window into a 180-mm-long stainless steel tube filled with xenon. Due to nonlinear polarization effects, the 355-nm beam is frequency tripled, yielding VUV photons with a wavelength of 118 nm for SPI ionization. Since only 0.001% of the initial 355-nm beam is converted in VUV photons, the 118-nm beam has to be separated from the UV laser beam. This is performed by an off-axis irradiation of the 355/118-nm beams onto a MgF2 lens, which separates the 355- and 118-nm focus spatially and angularly due to the difference in refractive index of MgF2 for these two wavelengths. The OPO laser as well as the tripling cell for generation of 118-nm VUV light needs the 355-nm beam provided by the Nd: YAG pump laser. Therefore, the VUV light for the SPI process and the UV light for REMPI cannot be used at the same time. The 355-nm beam used for the generation of VUV beams and the resulting OPO UV laser beam are alternatively blocked by beam dumps operating with a frequency of 5 Hz. With this setup, the gas mixture, which has to be analyzed, can be ionized with both methods in a quasi-parallel mode without instabilities in the beam paths. The sample inlet system is similar to a capillary-based inlet system described in detail in the literature.50,57-60 Briefly, it consists of a heated, hollow, stainless steel needle reaching into the center of the TOFMS ion source. Within this needle runs the aforementioned capillary. The outlet of the capillary is aligned with the tip of the needle and is located ∼2 mm above the center of the ion source. Behind the orifice of the capillary, an effusive molecular beam is formed. Both the UV laser beam for REMPI and the VUV laser beam for SPI are focused underneath the needle, crossing the molecular beam. Once molecular ions are formed by either REMPI or SPI, they are analyzed with the time-of-flight mass spectrometer using a multichannel plate detector (Chevron S304-10-D, Burle, Lancaster, PA). The acquisition of the laser ionization TOF mass spectra is performed by a 250-MHz, 1 GS/s, 12 k transient recorder PC card (model DP 110, Acquiris) at a repetition rate of 10 Hz. The spectra are stored in real-time on the hard disk. Storage and data processing are carried out by means of a home-written software package (developed with LabView, National Instruments, Austin, TX). Limits of detection down to 10 ppb are reached for aliphatic, aromatic, and carbonylic hydrocarbons (34 ppb, average 100 (SPI), average 10 (REMPI), S/N > 2).56 The linearity of the instrument for SPI-TOFMS is larger than 3 orders of magnitude.56 For the REMPI-TOFMS method, linearity in the range of 6 orders of magnitude has been shown.61 (57) Waileong, T.; Weicheng, S.; Jingang, Z.; Kopin, L. Chin. J. Chem. Phys. 2002, 15, 218-223. (58) Wiley, W. C.; McLaren, I. H. Rev. Sci. Instrum. 1955, 26, 1150-1157. (59) Harris, S. Proc. IEEE 1969, 57, 2096-2113. (60) Heger, H. J.; Zimmermann, R.; Dorfner, R.; Beckmann, M.; Griebel, H.; Kettrup, A.; Boesl, U. Anal. Chem. 1999, 71, 46-57.
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
5357
Figure 2. (A) Total ion current from TD/REMPI-TOFMS (above) and TD/SPI-TOFMS (below) of urban aerosol sample 1. All signals recorded at a given temperature are summed up over the whole holding time of that temperature. (B) OC fraction signals as found, for example, by Chow et al.14
RESULTS AND DISCUSSION In general, the recorded mass spectra of the different samples looked similar to each other in the applied qualitative approach. Therefore, only exemplary spectra are presented to demonstrate the applicability and practicability of the new technique for analyzing organic matter from urban aerosol. Figure 2A shows SPI-TOFMS and REMPI-TOFMS total ion current profiles of urban aerosol sample 1, integrated over the whole accessible mass range (up to 500 m/z). This procedure yields results comparable qualitatively to those obtained by the conventional OC/EC analysis, which is demonstrated in Figure 2B, where signals from OC fractions in dependence on the work of Chow et al.14 are (61) Oser, H.; Thanner, R.; Grotheer, H.-H.; Richters, U.; Walter, R.; Merz, A. In Combustion Diagnostics; Tacke, M., Stricker, W., Eds.; 1997; Vol. 3108, pp 21-29.
5358
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
depicted. Therefore, it has to be noted that Chow et al. determined the OC3 fraction at 450 °C, a temperature too high to achieve with the current experimental setup. The general shape of the curves are similar to those obtained with the IMPROVE method. However, one has to keep in mind that only these compounds are included, which can positively be ionized with either SPI at 118 nm or REMPI at 275 nm. Nevertheless, the data in Figure 2A reflect approximately the whole organic carbon content of the investigated fractions. REMPI-TOFMS is a selective measurement technique for the detection of PAH and their derivatives. Thus, the REMPI sum signals in Figure 2A could serve as an estimated measure for a sum value of the aromatic fraction of aerosol, which may be a suitable parameter for epidemiological studies on human health effects of particulate matter. In the same vein, the corresponding SPI signal may serve as a sum parameter for the total organic fraction, i.e., the OC fraction value out of conventional analysis. A closer look on the graphs in Figure 2A reveals that at 120 °C relatively few detectable compounds have been evolved from the aerosol, mostly the high volatile inorganic species ammonia and nitric oxide, which are also accessible with SPI, among some small organic molecules exhibiting very low concentrations. As described below in more detail, this reflects the onset of pyrolytic decomposition from higher molecular structures. The peaks at 250 °C are relatively sharp compared to the peaks observed at 340 °C, for both SPI and REMPI. This shows that a prominent fraction of the evolved volatile and semivolatile material at 250 °C is released suddenly after reaching the temperature step. For aromatic species, this suggests that those compounds are desorbed which boiling points or vapor pressures exhibit appropriate values. The more tailed peaks at 340 °C give a first hint that possibly secondary reactions such as pyrolytic processes of larger species take place at the higher temperature, which prolong the desorbing process. Due to the very low concentration and signal level for the measurements at 120 °C, which was also observed with the other three investigated samples, in the following, only spectra recorded at the two elevated temperatures are presented. One task of this study consists of the evaluation of the potential of the soft ionization mass spectrometry for detection of organic compounds on a molecular level present in the different OC fractions. SPI- and REMPI-TOFMS spectra summed up over the time of a given temperature step reveal the respective molecular pattern coming along with each temperature. Concerning the peak assignment, it has to be noted that isomers with the same m/z value cannot be separated with the used experimental setup. In some cases, peaks can be assigned primarily to one distinct compound based on knowledge obtained from previous experiments (e.g., refs 60 and 62) or results from conventional measurements (e.g., refs 63-66). Furthermore, some compounds may be excluded due to inappropriate electronic transitions or ionization (62) Mu ¨ hlberger, F.; Hafner, K.; Kaesdorf, S.; Ferge, T.; Zimmermann, R. Anal. Chem. 2004, 76, 6753-6764. (63) Zimmermann, R.; Blumenstock, M.; Heger, H. J.; Schramm, K.-W.; Kettrup, A. Environ. Sci. Technol. 2001, 35, 1019-1030. (64) Schnelle-Kreis, J.; Sklorz, M.; Peters, A.; Cyrys, J.; Zimmermann, R. Atmos. Environ. 2005, 39, 7702-7717. (65) Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T. Environ. Sci. Technol. 2001, 35, 1716-1728. (66) Fine, P. M.; Cass, G. R.; Simoneit, B. R. T. Environ. Sci. Technol. 2002, 36, 1442.
Figure 3. TD/REMPI at 275-nm TOFMS spectra of urban aerosol sample 2 at 250 (above) and 340 °C (below). Depiction of compounds exemplifies a likely representative isomer, if ambiguity occurs.
potentials. Despite the conclusions drawn from such considerations, it is often not possible to narrow down the peak assignment to a single chemical species, especially for larger values of m/z; thus, a likely representative compound is given in such cases to exemplify the peak assignment. Figure 3 shows REMPI at 275nm TOFMS spectra from the aerosol sample 2 recorded at 250 and 340 °C, respectively. At 250 °C, a large variety of PAH is visible in the mass spectrum. Typical homologue rows starting with phenanthrene/anthracene (178 m/z), pyrene/fluoranthene (202 m/z), and chrysene (228 m/z) can clearly be distinguished. The utilized wavelength is especially well suited for the detection of chrysene, because this molecule exhibits strong electronic transitions in the wavelength range around 275 nm, thus causing a strong resonant absorption/ionization process. Besides the multitude of PAH molecules, the abundance of phenol (94 m/z) and guiacol (124 m/z) is striking in this spectrum. Moving on to the mass spectrum recorded at 340 °C, the overall REMPI signal is considerably reduced. However, especially for compounds with lower molecular masses such as phenol and indole (117 m/z), larger peak heights are observed, whereas the signal intensity for the larger PAH has been decreased by a factor up to 10 (for chrysene). On the other hand, five- to seven-ring PAH such as perylene (252 m/z) and its methylated derivatives or coronene (300 m/z) are now more clearly visible at this temperature compared to 250 °C. These findings seem to suggest that at the lower temperature most PAH molecules with less than five rings are desorbed from the aerosol, whereas a significant portion of the larger, less volatile PAH remain on the particle phase and
Figure 4. TD/REMPI at 275-nm TOFMS spectra at 250 °C of Augsburg urban aerosol (from Figure 3) (above) and a ash sample from spruce wood combustion (below). Depiction of compounds exemplifies a likely representative isomer, if ambiguity occurs.
evolve subsequently at the higher temperature. Again, the relatively high abundance of volatile substances at 340 °C gives an indication of their origin from pyrolytic reactions of larger oligomeric and polymeric structures, because species such as phenol and indole should already be desorbed almost completely at lower temperatures. Figure 4 allows a closer look on some species abundant in the REMPI mass spectrum of the urban aerosol sample 2. In this figure, the REMPI at 275-nm TOFMS spectrum at 250 °C is compared with the corresponding mass spectrum of a ash sample from spruce wood combustion. Generally, the overall structure of both spectra looks very similar. There is a large variety of different PAH to be noticed along with the distinct peaks of phenol and guiacol. However, a closer look reveals that in both spectra different homologue rows are predominant. In urban aerosol especially, the occurrence of chrysene, and to a lesser extent phenanthrene and pyrene with their alkylated derivatives, are striking, whereas with the ash sample the homologue row of acenaphthene is salient in the spectrum along with the peak at 234 m/z, the latter representing, for example, retene. Phenol, guiacol, and retene are well-known marker compounds associated with the emissions from wood combustion.65-67 Thereby, phenol and guiacol can be related to lignin, an integral part of the cell wall of plants, which is a macromolecule consisting of oxygenated aromatic rings. Retene derives from pyrolytic processes out of resin acids such as abietic acid. Since phenol, guiacol, and 234 m/z (retene among other isomers) also give relatively intense (67) Schauer, J. J.; Cass, G. R. Environ. Sci. Technol. 2000, 34, 1821-1832.
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
5359
Figure 6. TD/SPI at 118-nm TOFMS spectra at 340 °C of urban aerosol sample 3 (from Figure 5) (above) and pure cellulose acetate (below).
Figure 5. TD/SPI at 118-nm TOFMS spectra from urban aerosol sample 3 at 250 (above) and 340 °C (below). Depiction of compounds exemplifies a likely representative isomer, if ambiguity occurs.
signals in the mass spectrum of the aerosol sample, it can be concluded that part of the particulate matter in this sample derives from emissions from wood combustion processes. This is corroborated by the occurrence of a peak at 314 m/z in both spectra in Figure 4, which in GC/MS measurements of the ash from spruce wood combustion could be assigned to dehydroabietic acid methyl ester, a further derivative of abietic acid.68 This compound has also been found in particles emitted from fireplace combustion of woods in the United States.66 In general, high abundance of PAH is an indication for combustion-related aerosol, and due to the fact that the sample was collected in December, where domestic heating is a major factor in anthropogenic emissions, this is reflected in the mass spectrum of the aerosol sample. This goes along with recent findings that Augsburg urban aerosol in winter is dominated by compounds emitted during wood combustion.69 The REMPI technique is well suited for the detection of PAH molecules; however, it is limited in terms of aliphatic and carbonylic hydrocarbons. This drawback is met by the application of SPI. Figure 5 shows SPI-TOFMS spectra of urban aerosol sample 3 at 250 and 340 °C, respectively. At 250 °C and lower mass range, most abundant peaks consist of ammonia (17 m/z), nitric oxide (30 m/z), and hydrogen sulfide (34 m/z). These inorganic species derive from the decomposition of ammonia (68) Sklorz, M. Personal communication 2005. (69) Schnelle-Kreis, J.; Sklorz, M.; Orasche, J.; Peters, A.; Zimmermann, R. Environ. Sci. Technol. In preparation.
5360 Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
nitrate and ammonia sulfate, the main components of particulate matter. Ammonia and nitric oxide are already present in the evolving gas at 120 °C, as stated above. Moving on to larger masses, peaks at 228, 242, 256, 270, 284, and 410 m/z are striking. These peaks can be assigned to alkanoic acids (tetra- to octadecanoic acid) or esters thereof and to cerotic acid methyl ester (for 410 m/z), the latter a derivative of cerotic acid, a known component of beeswax and pollen. Alkanoic acids are well known to occur in urban aerosol.67 Between these most intense peaks of inorganic material and carboxylic species, there is a variety of rather less intense signals, which seem to be coming along in small distinct groups of approximately 8-10 single peaks at a time. These peaks can be assigned to alkanes, alkenes, aldehydes, and ketones. The spectrum at 340 °C exhibits a pattern similar to that seen at the lower temperature. Again it is striking that high volatile organic substances such as acetaldehyde (44 m/z) and acetone (58 m/z) as well as the aliphatic and carbonylic hydrocarbons are abundant at this temperature. This shows that higher mass molecular structures such as oligomeric and polymeric compounds or polyfunctional oxygenated species are present in the aerosol, which produce the small compounds via pyrolytic processes at this elevated temperature. There is no possibility that such volatile species originate directly from the particulate matter in this temperature range. For validation purposes, a pure polymeric substance, cellulose acetate, was thermally treated using the same temperature protocol, and its SPI-TOFMS spectrum at 340 °C has been compared to the corresponding one from Figure 5. The result is shown in Figure 6. In the mass spectrum from cellulose acetate, the same groups of hydrocarbons and oxygenated hydrocarbons can be observed as in the SPI-TOFMS spectrum of urban aerosol. In the case of the polymer, these compounds have to be produced
Figure 7. Difference spectrum of urban aerosol sample 4 between TD/REMPI at 275-nm TOFMS signals at 250 and 340 °C.
by pyrolytic decomposition reactions. On the other hand, the higher mass molecular alkanoic acids are not visible with cellulose acetate. This finding gives a good indication for the hypothesis that the aliphatic and carbonylic matter at lower masses from the urban aerosol is generated by pyrolytic reactions from higher molecular species, e.g., oligomers and polymers or polyfunctional oxygenated compounds such as levoglucosan.20 The alkanoic acids, however, seem to originate from direct desorption from the particulate matter. REMPI data from sample 4 corroborates the above findings. Figure 7 shows a difference spectrum between the respective REMPI-TOFMS spectra recorded at 250 and 340 °C. From the figure it becomes obvious that volatile compounds such as phenol and indole predominate at the higher temperature, whereas the larger PAH are more abundant at 250 °C. This argues for the thermal desorption of virtually all aromatic species at 250 °C and a re-forming of volatile species at 340 °C from decomposition reactions of larger molecules present in the particulate matter.
CONCLUSION The newly developed method of thermal desorption combined with photoionization time-of-flight mass spectrometry seems to be a promising and reliable technique for the fast analysis of organic content of urban aerosol and source particles. Allotment of organic species on a molecular level to fractions of organic carbon is possible with this method. However, for further improvement of the experimental technique to address more advanced research questions in the field of aerosol science, more measurements are required in the future. In this context, expansion of the method to higher temperatures is of importance. Especially the topic of the polymeric and polyfunctional oxygenated portion of aerosol mass and the problem of mass closure for carbonaceous material of particulate matter need thorough investigation to substantiate the results, which could only be described in this study. For instance, moving on to higher temperatures and summing up of the total ion signal of low molecular pyrolytic produced species could yield a further sum parameter for estimation of the polymeric fraction of the aerosol. Furthermore, future applications could deal with the analysis of bioaerosols with respect to their higher molecular proteomic and oxygenated content. ACKNOWLEDGMENT This work was carried out in cooperation with the GSF-FocusNetwork “Aerosols and Health”, which coordinates aerosol-related research within the GSF Research Center. Special thanks are due to Dr. Martin Sklorz, Dr. Ju¨rgen Schnelle-Kreis, and Thomas Gro¨ger for providing the samples and thorough discussion of the results.
Received for review February 3, 2006. Accepted May 12, 2006. AC060227Y
Analytical Chemistry, Vol. 78, No. 15, August 1, 2006
5361