The Application of Thermal Methods for Determining Chemical Composition of Carbonaceous Aerosols: a Review

Journal of Environmental Science and Health, Part A

ISSN: 1093-4529 (Print) 1532-4117 (Online) Journal homepage: http://www.tandfonline.com/loi/lesa20

The application of thermal methods for determining chemical composition of carbonaceous aerosols: A review

Judith C. Chow , Jian Zhen Yu , John G. Watson , Steven Sai Hang Ho , Theresa L. Bohannan , Michael D. Hays & Kochy K. Fung

To cite this article: Judith C. Chow , Jian Zhen Yu , John G. Watson , Steven Sai Hang Ho , Theresa L. Bohannan , Michael D. Hays & Kochy K. Fung (2007) The application of thermal methods for determining chemical composition of carbonaceous aerosols: A review, Journal of Environmental Science and Health, Part A, 42:11, 1521-1541, DOI: 10.1080/10934520701513365 To link to this article: http://dx.doi.org/10.1080/10934520701513365

Published online: 19 Oct 2007.

Submit your article to this journal

Article views: 345

View related articles

Citing articles: 64 View citing articles

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=lesa20

Download by: [NIH Library] Date: 28 August 2017, At: 06:07 Journal of Environmental Science and Health Part A (2007) 42, 1521–1541 Copyright C Taylor & Francis Group, LLC ISSN: 1093-4529 (Print); 1532-4117 (Online) DOI: 10.1080/10934520701513365 The application of thermal methods for determining chemical composition of carbonaceous aerosols: A review

JUDITH C. CHOW1,2, JIAN ZHEN YU3,JOHN G. WATSON1,2, STEVEN SAI HANG HO1, THERESA L. BOHANNAN1, MICHAEL D. HAYS4 and KOCHY K. FUNG5 1Desert Research Institute, Reno, Nevada USA 2Institute of Earth and Environment of the Chinese Academy of Sciences, Xian, China 3Hong Kong University of Science and Technology, Kowloon, Hong Kong 4U.S. Environmental Protection Agency, Research Triangle Park, North Carolina, USA 5Atmoslytic, Inc., Calabasas, California, USA

Thermal methods of various forms have been used to quantify carbonaceous materials. Thermal/optical carbon analysis provides measurements of organic and elemental carbon concentrations as well as fractions evolving at specific temperatures in ambient and source aerosols. Detection of thermally desorbed organic compounds with thermal desorption-gas chromatography/mass spectrometry (TD-GC/MS) identifies and quantifies over 100 individual organic compounds in particulate matter (PM) samples. The resulting mass spectra contain information that is consistent among, but different between, source emissions even in the absence of association with specific organic compounds. TD-GC/MS is a demonstrated alternative to solvent extraction for many organic compounds and can be applied to samples from existing networks. It is amenable to field-deployable instruments capable of measuring organic aerosol composition in near real-time. In this review, thermal stability of organic compounds is related to chemical structures, providing a basis for understanding thermochemical properties of carbonaceous aerosols. Recent advances in thermal methods applied to determine aerosol chemical compositions are summarized and their potential for uncovering aerosol chemistry are evaluated. Current limitations and future research needs of the thermal methods are included. Keywords: Carbonaceous aerosol, thermal desorption, organic carbon, organic compounds, organic markers, carbon speciation.

Introduction Large numbers of samples are needed to relate ambient carbonaceous materials to pollution sources and

Downloaded by [NIH Library] at 06:07 28 August 2017 Airborne particulate matter (PM) has adverse effects on to assess their environmental and health consequences. human morbidity and mortality,[1−4] visibility,[5,6] climate This demands fast, simple, and sensitive analytical meth- change,[7] and materials.[8] These effects have sparked inter- ods. Solvent extraction (SE) followed by gas chromatog- est in the chemical and physical properties of PM, with raphy/mass spectrometry (GC/MS) analysis has been the increasing interest in its organic matter (OM) composi- most widely used approach for PM collected on filters. De- tion. The OM fraction is complex, comprising thousands spite some documented advantages of using SE-GC/MS of chemical constituents from numerous organic com- methods (e.g., selectivity, feasibility), it is labor- and time- pound classes, such as aromatics (i.e., polycyclic aromatic consuming, which limits the number of samples that can hydrocarbons [PAHs]), alcohols alkanes (i.e., n-alkanes, be analyzed. It also requires large sample deposits (>1mg hopanes, and steranes), and carboxylic acids (i.e., alkanoic organic carbon)[24] that are not normally acquired in speci- acids).[9−12] Various compounds within these classes are ation networks. recognized as biologically toxic.[13−15] Many are also used Simple and fast thermal methods are used to determine as organic markers in source/receptor models for source thousands of organic and elemental carbon (OC/EC) in apportionment.[16−23] ambient and source aerosols collected on quartz-fiber filters each year.[16,25−30] Similar analytical simplicity can also be expected for analysis of individual organic compounds Address correspondence to: Judith C. Chow, Division of Atmo- using more specific detection methods for the thermally spheric Sciences, Desert Research Institute, 2215 Raggio Parkway, evolved organic compounds. Reno, NV 89512. E-mail: [email protected] The term thermal extraction (TE) refers to the direct Received May 21, 2007. removal of OM from the aerosol matrix by heating the 1522 Chow et al.

sample; it has been used interchangeably with the term ther- organic alcohols and acids have higher vH than alkanes mal desorption (TD). TD is a technique by which relatively with equivalent carbon numbers. volatile analytes trapped on a solid adsorbent (such as Apart from heating, vaporization occurs at reduced sys- Tenax) are released at high temperatures, and then ana- tem pressures. The boiling point of a substance is the tem- lyzed. TD has been used to measure volatile organic com- perature at which its vapor pressure equals the pressure pounds (VOCs) from soil, water, and air.[31−42] Since the of its environment.[50] When pressure is reduced, vaporiza- 1970s, TD has been applied to aerosol samples[43] for molec- tion takes place at a lower temperature. The boiling point ular composition.[35,44−47] TD has also been combined with of the substance decreases while the pressure decreases semi-continuous and continuous instrumentation that mea- such that: sures organic compounds in picograms collected over time 1 R P 1 frames of seconds.[48,49] = · ln + (1) T vH P T The current knowledge of thermal stability of organic B o o compounds as a function of chemical structures is re- where TB is the boiling point in K under pressure P, R is the viewed herein, providing a chemical basis for understanding ideal gas constant (8.3143 m3 × Pa = K × mole), and To is thermochemical properties of aerosol organic compounds the given temperature boiling point in K under the standard [50] and matrix effects during thermal analysis. Recent TD pressure Po. In a mixture, the vapor pressure of a solute advancements and applications are summarized, the po- is reduced, leading to boiling point elevation of the solute. tential for uncovering aerosol chemistry is evaluated, and According to Raoult’s Law, the vapor pressure reduction current limitations and future research needs of thermal is proportional to the mole fraction of the solute in the methods are addressed. This paper complements the re- mixture. Deviations from Raoult’s Law are often observed view of Hays and Lavrich,[24] which focused on molecular as the interaction between solute and solvent is not ideal. marker species identified and quantified by off-line TD- GC/MS techniques. Thermal stability of organic molecules The thermal stability of an organic molecule is a combina- Thermal desorption mechanisms tion of its thermodynamic and kinetic stability. The finite strength of chemical bonds puts an upper limit on the vi- Vaporization brational energy that molecules may possess without bond Vaporization is a conversion process by which a chemical rupture, i.e., decomposition. The Boltzmann energy distri- in solid or liquid form is changed to the gas state by heating bution law among molecules dictates that at any tempera- or reducing pressure. Most substances change their phys- ture a portion of the molecules will possess energy greater ical states in the order of solid to liquid to gas with in- than the bond energy. The thermal degradation rate (k)of creasing temperature. A few sublimate, or change directly a molecule is expressed as: from solid to gas. The molecules in solid or liquid states − / k = Ae E RT (2) bind with the intermolecular forces, which are fundamen- tally electrostatic interactions (ionic interactions, hydrogen A is the frequency factor related to a critical mode of

Downloaded by [NIH Library] at 06:07 28 August 2017 bond, dipole-dipole interactions) or electrodynamic inter- vibration in the molecules and has the same dimension as actions (Van der Waals force).[50] Vaporization is an en- k. For ﬁrst order kinetics, A is in units of 1/sec. E is the dothermic process in which molecules gain kinetic energy activation energy supplied to the molecules to cause de- as temperature increases. When molecular kinetic energy composition. For decomposition to occur, E needs to be exceeds the enthalpy change of vaporization (vH), which greater than the dissociation energy of the weakest bond is the energy required to overcome intermolecular forces in in the molecule; the weakest bond strength determines the the solid or liquid, individual molecules escape and enter upper limit of its thermal stability, if there are no other the gas state.[50] lower-energy decomposition pathways.[55] Simple hydrocar- This temperature is the boiling point at a constant pres- bons, such as alkanes, decompose through rupture of a C–C sure. Methane has a low standard enthalpy change of va- bond, with the formation of free radicals;[56,57] however, the porization (vHo = 8.19 kJ mol−1) because of weak Van functional groups of many organic molecules can lead to the der Waals forces between its molecules (Sears et al., 1982). rearrangement of pathways and lower the thermal decom- These forces result from formation of momentary dipoles position temperature.[58] The rearrangement pathways take in uncharged non-polar compounds.[51] Higher vH are place at temperatures far below the temperature required measured for formic acid and methanol (vHo=19.9 kJ for straightforward bond rupture. mol−1and 37.4 kJ mol−1,respectively) since the molecules Thermal stability of a molecule is primarily determined bind with strong hydrogen (H) bonds.[52,53] The bond is by its chemical structure.[58,59] According to Eq. 2, chemical an attractive interaction between an electronegative atom structures that minimize A and maximize E lower the rate (e.g., oxygen [O] and nitrogen [N]) and an H atom bonded of k, thereby increasing stability of the molecules at a given to another electronegative atom.[54] This explains why polar temperature. Polybonding, in which an atom is bonded to Thermal methods used to study carbonaceous aerosols 1523

more than one other atom in the molecule, has the effect Pyrolysis of decreasing A,ascollision energy can be dissipated over Pyrolysis converts an organic compound to one or more more than one path (e.g., the thermal stability of graphite). different compounds by thermal energy in an inert envi- The thermal stability of many inorganic solids is also related ronment (absence of oxygen or catalysts).[64] Carboniza- to this factor, rather than to their high bond strength. More , tion happens in extreme pyrolysis to form EC.[64 65] Above specifically, the structural features of compounds that favor a particular temperature, organic molecules gain sufficient high thermal stability include: (i) all bonds in the molecule vibrational energy to rupture their bonds and form free having high dissociation energies; (ii) no easy paths to de- radicals that subsequently reform other lower molecular- composition; (iii) resonance structure; and (iv) substituents weight compounds or combine with higher molecular- that enhance resonance stabilization. Resonance structure − weight compounds.[66 71] Saturated alkanes break down moieties have the effect of increasing E in Eq. 2, hence in- [58] C C and C H bonds to radicals that form smaller alka- creasing thermal stability. − nes, alkenes, and hydrogen.[72 74] Simple aromatic com- Aromatic compounds have high thermal stability be- pounds (e.g., benzene and toluene) produce aryl radicals, cause of their resonance structure. For example, naphtha- which attack other hydrocarbon molecules to give bi- and lene in the vapor phase (at 1 atm) is thermally stable up − polyaryls.[75 78] The reaction causes loss of H and X (any to 620◦C.[59] Nine PAHs (i.e., anthracene, phenanthrene, leaving group [e.g., F, NH , OH]) from a compound with fluoranthene, chrysene, pyrene, triphenylene, benz[a]- 2 a configuration of H-C-C-X and results in formation of a anthracene, benzo[a]pyrene, and dibenzo[a,h]anthracene) − C=C bond.[79 82] are thermally stable beyond 475◦C.[60] In comparison, alkyl- This “elimination” also occurs by loss of small molecules PAHs have lower thermal stability. This follows from the such as N ,CO,CO,orSO, leading to formation of re- introduction of lower dissociation energy bonds (i.e., alkyl 2 2 2 active intermediates such as arynes, diradicals, carbenes, or C–C bonds) than the bonds in the aromatic rings. Madi- , nitrenes.[80 82] Cleavage of particular C–C bonds in a car- son and Roberts[60] observed slight thermal decomposition bon chain or ring always happens in pyrolysis. This includes for 1-methyl naphthalene at 475◦C. For the same reason, transfer of hydrogen during fragmentation that occurs by alkanes in general have lower thermal stability than PAHs. − acyclic process.[83 85] Pyrolysis can involve isomerization Hydrocarbons containing alkyl chains have been found to without elimination or fragmentation. These processes oc- remain stable to about 350◦C.[55] Table1lists the chemical cur by way of intermediates such as di-radicals from peri- and physical properties of major organic classes and mark- − cyclic reactions.[86 88] ers in aerosols. Thermally stable molecules lack easy paths to [59] decomposition. Polar functional groups, such as –NH2, Thermochemical properties of carbonaceous aerosols –OH, –COOH, often provide lower energy decomposition in which bonds are broken and formed simultaneously Carbonaceous materials can be thermally released to the (e.g., the loss of HX), and therefore are less thermally gas phase either in chemically unaltered or altered forms. stable. For example, thermal decomposition of alcohols For carbonaceous aerosol, EC is thermally resistant, re- involves the unimolecular elimination of water to form quiring oxidation to evolve into the gas phase in the form

Downloaded by [NIH Library] at 06:07 28 August 2017 [61] [62] alkenes. Schwarz et al. showed that polyhydroxylated of CO or CO2.OCiscomprised of tens of thousands of compounds like sugars (e.g., mannitol, dulcitol, sorbitol, organic compounds with diverse chemical structures. De- arabitol, xylitol, and ribitol) are stable only up to 200– pending on the chemical nature, some organic compounds 250◦C. Thermal stability of esters is lower than hydrocar- are volatilized into the gas phase unaltered while others bons. Esters of alcohols with β-hydrogen (e.g., methyl esters are thermally labile, leading to degradation products re- of fatty alcohols) are stable up to only ∼285◦C.[55] leased to the gas phase. As discussed earlier, oxygenated As shown in Eq. 2, temperature (T) is the most impor- compounds are prone to thermal conversion during TD. tant external factor affecting thermal stability. In addition, n-Alkanols in a GC injection port at 250–350◦C yielded atagiven temperature, decomposition generally occurs un- n-aldehydes, and 1-alkenes;[89,90] levoglucosan partially de- der high, rather than low, vapor pressure conditions due to grades to levoglucosenone at 300◦C;[49] cholesterol spiked greater collision frequencies.[58] onto a quartz-fiber filter during TD-GC/MS analysis par- The presence of other molecules can lead to higher or tially converted to cholesta-3,5-diene. Thermal degrada- lower stability by preventing or opening thermal degradation products could be mistaken for compounds originally tion/reaction pathways, respectively. For example, in the present in suspended particles. This highlights the impor- petroleum industry, it has been known that the thermal tance of removing the thermally labile constituents or con- stability of alkanes can be improved in the presence of hy- verting them to less sensitive functional groups before ap- drogen donors, e.g., tetralin.[63] This is achieved by interrup- plying TD. tion of the radical chain reaction through the H-donating Thermal behaviors might be modified by other compo- action. nents coexisting in the particles. For example, n-alkanoic Downloaded by [NIH Library] at 06:07 28 August 2017 1524

Table 1. Chemical and physical properties of major organic classes and markers in aerosols[149] Melting Boiling Vapor Pressure Amenability Amenability Organic Molecular Point Point at 25◦C Description Thermal to Thermal to Solvent Classes Subgroups Useful Marker Compounds Weight (◦C) (◦C) (mmHg) (Bonds) Stability Desorption Extraction

Aromatic PAHs ﬂuoranthene (C16H10) 202.6 110.8 375 1.23E-08 aromatic ring, C H High Yes Yes chrysene (C18H12) 228.3 255.8 448 6.23E-09 benzo[e]pyrene (C20H12) 252.3 178.0 492 5.70E-09 benzo[a]pyrene (C20H12) 252.3 176.5 495 1.80E-14 indeno[1,2,3-cd]pyrene (C22H12) 276.3 162.5 536 1.00E-10 benzo[ghi]perylene (C22H12) 276.3 278.3 500 1.00E-10 picene (C22H14) 278.4 366–367 518–520 3.73E-09 coronene (C24H12) 300.4 438 525 3.29E-11

Alkanes n-Alkanes n-nonacosane (C29H60) 408.8 63.7 440.8 3.24E-14 C C, C H High Yes Yes n-triacontane (C30H62) 422.8 62.8 449.7 2.74E-15 n-hentriacontane (C31H64) 436.8 67.9 458 1.40E-16 n-dotriacontane (C32H66) 450.9 69 467 8.49E-18 b n-tritriacontane (C33H68) 464.9 71–73 485.3 4.02E-19 b n-tetratriacontane (C34H70) 478.9 72.6 496.9 1.67E-20 n-pentatriacontane (C35H72) 493 75 490 5.39E-22 b n-hexatriacontane (C36H74) 507 74–76 520.1 1.34E-23

a b c Hopanes 17α(H)-22,29,30-trisnorhopane (C27H46) 370.7 112.8 387.8 2.09E-06 cyclic ring, C C, High Yes Yes a b c 17α(H),21β(H)-30-norhopane (C29H50) 398.7 124.4 407.7 5.19E-07 C H a b c 17α(H),21β(H)-hopane (C30H52) 412.8 127.1 412.3 3.91E-07 a b c 17α(H),21β(H)-22R-homohopane (C31H54) 426.8 133.8 423.9 1.71E-07

a b c Steranes ααα 20R-cholestane (C27H48) 372.7 115.5 392.5 8.79E-06 cyclic ring, C C, High Yes Yes a b c ααα 20R 24R-methylcholestane (C28H50) 386.7 122.9 405.2 9.19E-07 C H a b c ααα 20R 24R-ethylcholestane (C29H52) 400.7 130.3 417.9 3.84E-07

Esters Phthalates dimethylphthalate (C10H10O4) 194.2 0-2 282 1.65E-03 aromatic ring with Fair Yes Yes diethyl phthalate (C12H14O4) 222.2 −40.5 295 1.65E-03 COO-R1,C C, C H di-n-butyl phthalate (C16H22O4) 278.3 −35 340 7.30E-05 (R1 could be CH2 butyl benzyl phthalate (C19H20O4) 312.4 <−35 370 8.25E-06 chain or aromatic bis(2-ethylhexyl)phthalate (C24H38O4) 390.6 −50 386.9 9.75E-06 ring) di-n-octyl phthalate (C24H38O4) 390.6 −25 220 2.60E-06 Downloaded by [NIH Library] at 06:07 28 August 2017

Alcohols Sugars glycerol (C3H8O3) 92.1 17.8 290 1.68E-04 OH, C C, C HLowNo Yes c xylitol (C5H12O5) 152.1 95-97 216 1.96E-03 b c levoglucosan (C6H10O5) 162.1 170-180 313.8 3.47E-07 b glucose (C6H12O6) 180.2 83 380.7 1.82E-08 a b fructose (C6H12O6) 180.2 85.4 341 3.11E-07 mannitol (C6H14O6) 182.2 167 290-295 1.31E-09 b sorbitol (C6H14O6) 182.2 93-97 389.7 4.92E-09 b sucrose (C12H22O11) 342.3 190-192 591.6 5.15E-17 cholesterol (C27H46O) 386.7 148.5 360 7.79E-10

Carboxylic Mono- and oxalic acid (C2H2O4)90189 190 2.34E-04 COOH, C C, C HLowNo Yes acids Di- malonic acid (C3H4O4) 104.1 135.6 140 1.50E-03 carboxylic succinic acid (C4H6O4) 118.1 183–185 235 7.86E-06 f acids adipic acid (C6H10O4) 146.1 152 337 7.30E-02 d pimelic acid (C7H12O4) 160.2 103–105 272 2.08E-05 e suberic acid (C8H14O4) 174.2 144 230 1.69E-08 e palmitic acid (C16H32O2) 256.4 62 271.4 1.20E-05 oleic acid (C18H34O2) 282.3 13–14 460 5.46E-07 stearic acid (C18H36O2) 284.5 69.3 361 7.22E-07

a Melting point was estimated with Gold and Ogle Method (using EPI SuiteTMdeveloped by the U.S. Office of Pollution Prevention Toxics and Syracuse Research Corporation [SRC]). bBoling point was estimated with Adapted Stein and Brown Method (using EPI SuiteTMdeveloped by the U.S. Office of Pollution Prevention Toxics and Syracuse Research Corporation [SRC]). cVapor pressure was estimated with Modified Grain Method (using EPI SuiteTMdeveloped by the U.S. Office of Pollution Prevention Toxics and Syracuse Research Corporation [SRC]. d Boiling point reported at the pressure of 100 mmHg. eBoiling point reported at the pressure of 15 mmHg. f Vapor pressure reported at the temperature of 18◦C. 1525 1526 Chow et al.

acid amides (C12-C20) and nitriles (C14-C32)were measured ForTD, thermal conditions are optimized for fast and in the TD-GC/MS analysis of ambient aerosols collected in complete vaporization of target compounds in their unal- Augsburg, Germany.[91] Five nitriles were also reported in tered forms, along with a sensitive detection system for indi- aerosols of a forest site in Germany.[92] These amides and vidual organic compounds (e.g., MS, GC/MS, GC/FID). nitriles are actually thermal reaction products from analo- Thermal conditions can also be selected to evolve non- gous acids and ammonium nitrate (NH4NO3) coexisting in volatile and insoluble organic compounds as their pyrolysis the aerosol samples.[93] Yu et al.[94] observed changes in the products, and the parent compounds identities are inferred thermal evolution pattern of levoglucosan in the presence from their thermal degradation products as in the applica- and absence of ammonium bisulfate (NH4HSO4). The pres- tion of pyrolysis GC for the determination of rubbers and ence of transition metal catalysts (e.g., Fe, Cu) in aerosols plastics. TD coupled with a suitable detection system has can also affect the conversion of OM during heating. the potential to become a simple and routine tool for the de- Matrix effects are little understood and may vary from termination of organic aerosol chemical composition. With sample to sample. On-line instruments using TD minimize perfection of the equipment, procedures, and data process- thermal decomposition due to reduced pressure and lower ing methods, it would be possible to obtain data on cur- aerosol loading in the analysis zone.[95,96] Laser desorp- rently acquired or archived samples with little additional tion has been employed for on-line MS-based instruments cost and effort than is currently expended on OC and EC (e.g.,[95,96]). However, it experiences fluctuations in desorp- measurement. tion/ionization conditions which lead to large pulse-to- pulse variation in ion signals that impede quantification. Off-line thermal desorption analyses of filters TD-MS Variations among thermal methods Schuetzle et al.[44,45] and Cronn et al.[97] first collected par- Figure 1 shows a conceptual diagram of various TD meth- ticles on glass-fiber filters and impaction plates, then in- ods applied to aerosol samples. Thermal conditions can troduced them directly into a high-resolution MS utilizing be adjusted to extract the desired analytical information, a temperature-programmed insertion tube (solid probe), ◦ ◦ −1 and differ among methods. In aerosol carbon analyzers, which was heated from 20 to 400 Cat20–30 C/min un- −9 the thermal conditions are targeted to evolve OC and EC der a vacuum of 10 atm. Compounds were selectively separately without preserving the original forms of organic removed from the aerosol matrix according to their volatility. The TD-MS-generated mass thermograms, i.e., the in- compounds. The detection of a single product (i.e., CO2 tensity of each ion current peak as a function of time or or CH4) after catalytic oxidation is sufficient to achieve the analytical objective.[22,25−29] Long-term databases have temperature. Compound identification was based on: (i) been compiled with the data (e.g., http://vista.cira. a major characteristic ion and the temperature at which colostate.edu/improve/, http://www.epa.gov/ttn/amtic/ this signal was detected; (ii) one or two additional ions at speciepg.html, http://www.atmospheric-research.com/ the same temperature; and (iii) the plausibility of the ex- studies/SEARCH/index.html, http://www.arb.ca.gov/ istence of the component in the aerosol by vapor pressure airways/CRPAQS/default.htm). calculation. Downloaded by [NIH Library] at 06:07 28 August 2017

Fig. 1. Conceptual depiction of various thermal methods as applied to aerosols. Thermal methods used to study carbonaceous aerosols 1527

Cronn et al.[97] identified alkanes, alkenes, aromatics The Curie point TD device uses radio frequencies to in- (i.e., alkyl benzenes, substituted naphthalenes, and PAHs), duce eddy currents in a ferromagnetic foil, in which filter amines (i.e., alkyl piperidines), alcohols (i.e., phenol), and strips are positioned, to heat the sample to the Curie point carboxylic acids. Schuetzle et al.[45] identified a number of temperature of the foil. At the Curie point, the ferromag- peroxides in aerosols collected in Pasadena, California, netism of the metal changes to paramagnetism so that the suggesting low probability of decomposition of organic foil remains at that constant temperature. The Curie point compounds with the TD under high vacuum conditions. devices offer rapid heating. This might reduce decomposi- Nevertheless, there were signs of matrix effects and pos- tion as a result of volatilization being kinetically favored [112] sible thermal decomposition. For example, C3-C7 dicar- relative to the competitive decomposition processes. boxylic acids were found while C2 dicarboxylic acid was External TD units require a transfer line and a connector not found, inconsistent with results from other studies that to the GC injector and consequently may lose organic com- showed oxalic acid is more abundant than the higher di- pounds when the material of the transfer line is not inert or carboxylic acids.[98,99] This failure to detect oxalic acid in the line has cold spots. A contaminated line of inert mate- southern California reflects thermal decomposition. The rial (due to deposition over time) can also become active. matrix effect was demonstrated by the observation that four The line has to be inert and heated uniformly to a tem- tested compounds (i.e., naphthalene, phenol, glutaric acid, perature above the desorption temperature. Residue (e.g., and phenanthrene) spiked onto aerosol loaded substrates polar substances) may be trapped between the TD unit and , required 44–152◦C higher temperatures to vaporize than the GC/MS, causing carryover.[46 103] TD performed in the those spiked onto a particle-free brass disk.[45] This is the GC injector is simple and eliminates the need for an exter- effect of boiling point elevation due to reduction of vapor nal oven and transfer line. It requires no modifications or pressure of the substances in a mixture. The effect was more only a minor rearrangement of carrier gas lines or an injec- prominent for two polar compounds (phenol and glutaric tor cap. This development not only reduces the carryover, acid) with hydrogen bonding. but also offers maximum transfer efficiency. Since TD takes place in a vacuum, organic compounds TD-GC/MS differs from SE-GC/MS primarily in the are vaporized at reduced temperatures. This in turn ex- principle of extracting individual organic compounds (i.e., tends analytical feasibility to compounds of lower ther- by heat or solvent) from the filter. GC/MS analyses of the mal stability (e.g., peroxides) and higher molecular weight. extracts are the same. Table 2 compares attributes of the Perry et al.[100] reported 12-ring hydrocarbons using TD- two methods. MS in soot as compared to a 7-ring PAH (coronene) us- The SE methods are better known for their selectivity, ing GC. However, organic compounds with similar ther- feasibility, and for sample extract derivatization, fraction- mal and mass fragmentation behaviors cannot be resolved ation, and clean-up. Solvent extracts generally pair bet- by TD-MS. ter with off-line analytical instrumentation (e.g., ion chromatography [IC] and high performance liquid chromatography [HPLC], both transfer samples with liquid mobile TD-GC/MS phases) than TD, which has resulted in an expanded list of With capillary GC, organic compounds in aerosol samples non-polar and polar organic compounds.[12,21,113,114] TD-

Downloaded by [NIH Library] at 06:07 28 August 2017 are separated according to their physical (boiling point) GC/MS applied to aerosols has been largely limited to non- and chemical (polarity) properties before quantification by polar PAHs and n-alkanes. Unlike SE in which only a small MS. Chromatographic separation prior to FID or MS de- portion of the extract (typically a few percent) is injected, tection minimizes interferences from coexisting species. As the entire sample mass from a filter portion is transferred to shown in Figure 2, two configurations exist for off-line TD- the TD-GC/MS. As a result, TD is at least an order of mag- GC/MS, with TD carried out either in an external oven or nitude more sensitive than SE and requires considerably less inside the GC injection port. A brief summary of these units sample mass. This attribute is particularly important when is given below; details can be found in Hays and Lavrich.[24] only micrograms of aerosols are available as on a 25 mm- External TD units use either conventional resistive to 47 mm-diameter filter from speciation samplers.[43,115] heating[101−103] or Curie point heating[104,105] to thermally The high sensitivity of TD improves temporal resolution extract organic compounds. The in-injection port TD also of organic compounds. Since TD requires a small frac- has several variants. Falkovich and Rudich[106] used a micro- tion (0.3–5.0 cm2)ofthe filter for analysis,[106,108,116] it vial to hold the filter strip inside a GC liner. Wauters can be applied to remnants of archived quartz-fiber fil- et al.[107] and Ho and Yu[108] adapted a conventional GC ters after ionic and carbon analyses without additional liner without any modifications. Helmig et al.[92] altered the sampling. carrier gas lines, while Blanchard and Hopper[109] modified the GC injector with a small T-connector, a three-way valve, TD with two-dimensional GC-time-of-flight/MS and a needle valve. Zimmermann and co-workers[91,110,111] (TD-GC×GC-TOF/MS) automated the insertion of a GC injection-liner, preloaded Since the greater part of organics remain unresolved by TD- with filter pieces, and the closing of the injector port. GC/MS, TD paired with comprehensive two-dimensional 1528 Chow et al. Downloaded by [NIH Library] at 06:07 28 August 2017

Fig. 2. Conceptual diagrams of different off-line thermal desorption-gas chromatography/mass spectrometry (TD-GC/MS) ap- proaches for: (a) external TD or Curie point pyrolysis-GC/MS: (i) desorbent refocusing with a cryogenic trap inside TD unit; (ii) desorbent refocusing onto GC column head; and (iii) desorbent refocusing with a cryogenic trap in front of GC column; (b) In-injection port TD-GC/MS: (iv) desorbent refocusing onto GC column head, and (v) desorbent refocusing with a cryogenic trap in front of GC column; and (c) Direct TD GCxGC-TOF/MS: desorbent refocusing with a cryogenic trap in front of the second GC column.

GC and time-of-flight (TOF)/MS (Fig. 2c) has been used stationary phase column for further separation accord- to provide additional resolution. The compounds are first ing to polarity. Using this approach, Welthagen et al.[110] separated by boiling points with a non-polar stationary detected ∼15,000 compounds for particles collected in phase column. As the compounds elute from this first Germany, including categories of: alkanes (i.e., n-alkanes, column, they are refocused onto a shorter (1 m) polar steranes, hopanes), alkenes (i.e., terpenes), aromatic (i.e., Thermal methods used to study carbonaceous aerosols 1529

Table 2. Attributes of thermal desorption and solvent extraction methods for the analysis of organic compounds in aerosols collected on ﬁlters.

Variables Thermal Desorption Solvent Extraction

Extraction Theory - Thermal behavior of analytes (e.g., boiling - Solubility of analytes in extraction solvent or point, volatility). solvent mixture (normally binary or tertiary). Sample Size and Loading - Small punches (0.3–5.0 cm2)offilter depending -Requires >50 ◦g1µg OC; on the design of thermal desorption unit and - Combining samples (>60 cm2) into composites sample loading; may be needed for low loading samples; - Can use remnants of archived quartz-fiber - High collection flow rates (e.g., 100–1,000 L samples from existing networks; min−1)are required to increase the sample -Requires >1 µgOCa; loading. - Medium collection flow rates (e.g., 10–30 L min−1)provides adequate sample loading for analysis. Sample Deposit - Essential since only a fraction of filter is used. - Not relevant if entire filter is used for Homogeneity extractions. Solvent Consumption - Low, only few microliters (µL) internal - High, ranging from 10 to 300 mL per sample; standard containing solvent used per sample. - Common solvents used: dichloromethane (DCM), methanol, 1-butanol, acetone, and n-hexane; - Generally >90% cannot be recycled. Sample Pretreatment - Spiking with internal standard, a few minutes - Complicated steps (i.e., single or multiple labor volume reductions, extract transfer, and sample filtration); -1to20hours labor per sample. Analysis Timeb -1to 2 labor hours per sample. - A few to tens of hours. Analytical Instrument - Mass spectrometer (MS); - GC/MS - Gas chromatography/flame ionization detector - GC/FID; (GC/FID); - High performance liquid chromatography - Gas chromatography/mass spectrometer (HPLC); (GC/MS). - Ion chromatography (IC); -Capillary electrophoresis (CE). Sample Introduction - 100%. - 0.5 to 25% of the solvent extract. Volume Sample Contamination - Low probability; High probability; -Potential carryover (usually negligible), -Possible contaminants from solvents and transfer loss, molecular arrangement, or complicated extraction procedures; breakdown as temperature ≥350◦C; - Loss of volatile compounds during the Downloaded by [NIH Library] at 06:07 28 August 2017 -Fragmentation of thermally labile compounds. extraction and pretreatment steps; -Possible carryover from injection port when dirty samples are introduced. Sensitivity and Minimum - High sensitivity; -Low sensitivity; Detection Limits - n-Alkanes: 0.41 to 4.36 ng/ sample−1c; - n-Alkanes: 37.7 to 125 ng/ sample−1c; (MDLs) -PAHs: 0.08 to 2.40 ng/ sample−1c; -PAHs: 10.3 to 47.9 ng/ sample−1c; Suitability of Filter Matrix - Filters that can remain intact at TD - Filters that do not dissolve in solvents, e.g., temperatures, e.g., quartz-fiber or quartz-fiber, glass-fiber, or TIGF filters. Teflon-impregnated glass-fiber (TIGF) filters.

a Calculation based on the detection limits of the organic compounds reported in the thermal desorption method.[103] bIncluding sample pre-treatment and analytical separation. cData from Ho and Yu.[108]

alkyl-benzenes, substituted aromatics, heterocyclic aromat- hexenal, furan) identiﬁed by Hamilton et al. were fairly ics, PAHs, oxygenated-PAHs, branched-PAHs), aldehydes, volatile (vapor pressure >10−2 atm at20◦C). The presence ketones (i.e., n-alkane-2-ones), carboxylic acids, nitriles (i.e, of these volatile species needs further evaluation. It is pos- alkyl nitriles), and esters.[91] These methods can also iden- sible that some of the oxygenated species detected are TD tify oxygenated compounds (e.g., aldehydes, ketones, and products of thermally labile compounds that were origi- carboxylic acids) ascribable to secondary organic aerosols nally present in the aerosols. Using GC×GC-TOF/MS to (SOA).[117,118] Many of the oxygenated species (e.g., 2- quantify organic compounds shows TD-potential.[91] 1530 Chow et al.

Pyrolysis-TD-GC/MS TD-GC/MS also provides a rapid screening tool, since no The two most common pyrolysis techniques are Curie point sample pretreatment is required. and controlled temperature programming pyrolysis. Pyrol- ysis is often used as a form of sample pretreatment for Parameters for aerosol analysis with TD-GC/MS mass spectrometric investigations of complex materials.[119] and pyrolysis-TD-GC/MS Pyrolysis-TD-GC/MS uses higher TD temperatures (of- Table3compares experimental parameters for 30 stud- ten exceeding 400◦C) than TD-GC/MS.[104,119,120] As a re- ies using TD-GC/MS and/or pyrolysis-TD-GC/MS. Im- sult, pyrolysis-TD-GC/MS produces many more evolved portant parameters include: (i) type of sample; (ii) sample products. Pyrolysis of resuspended soil samples at 740◦C preparation; (iii) TD unit; (iv) analytical instrument; (v) TD resolved 174 compounds,[119] five times more than were ob- temperature program; (vi) TD time; (vii) desorbent refocus- tained with TD for the same sample at 315◦C. ing pathway; (viii) GC column; (ix) GC initial temperature; During pyrolysis, vaporized compounds include both (x) GC oven temperature program; and (xi) total analysis volatile molecules (in the sample) and volatile fragments time. The quantitative significance of different parameter produced from cracking of non-volatilized molecules. The selections has not yet been determined. vaporized compounds are then collected, separated, and TD temperature ramping and holding times varied by identified by GC/MS. In pyrolysis, polar compounds, ei- sample type, experimental objective, analytical equipment ther indigenous or those formed during thermal degrada- and operation, which affected the analysis times. Both iso- tion, can readily undergo reactions such as decarboxyla- cratic and gradient heating were applied in resistively heated tion and dehydration, consequently losing the functional TD devices, whereas rapid induction heating (∼3s) was used group information. To overcome this, reactive pyrolysis in Curie point pyrolysis. TD of aerosols in off-line ovens and mixes a derivatizing reagent with the sample to preserve inside the GC injection port is completed in a matter of min- the polar functional group information and also to im- utes, or seconds in the case of Curie point pyrolysis. Isocratic prove their detection.[121] Tetramethylammonium hydrox- desorption temperatures of 250–320◦Chave been used in ide (TMAH) and hexamethyldisilazane (HMDS) are the afew studies (e.g.,[46,92,109,116,127−129]). For gradual heating, common methylation and silylation reagents, respectively, temperatures varied from ambient (∼25◦C)[103−105,130,131] to for thermally assisted derivatization.[121] 175◦C.[102] Temperature ramping in TD varied from 12◦C Pyrolysis-TD-GC/MS has characterized biopolymers min−1[103,130] to 200◦C min−1[106] without holding time, and (e.g., lignins), soil particles, sediments, and other matrices from 15◦C min−1[101] to 50◦C min−1[102] with 5 to 7.5 min containing macromolecules (e.g.,[122,123]). Subbalakshmi holding time. Maximum reported TD temperatures range et al.[124] identified aerosol pyrolysis products and linked from 275 to 315◦C[105,108] with holding times ranging from them to their parent molecules: (i) methoxyphenols origi- 0to10min. These temperatures and holding times were nating from lignin; (ii) furans, aldehydes, and ketones from optimized to ensure complete extraction and to minimize carbohydrates; (iii) pyrrole and indoles from proteins; and pyrolysis.[46] (iv) hydrocarbons from lipid structures. Fabbri et al.[121] ap- The desorption temperature is directly related to the ◦ plied conventional and reactive pyrolysis at 700 CtoPM10 volatility and thermal stability of the targeted analytes. from an industrial area of Bologna, Italy, providing infor- At desorption temperatures < 275◦C, the analytes are not

Downloaded by [NIH Library] at 06:07 28 August 2017 mation on precursors (e.g., carbohydrates, lipids, synthetic completely extracted, especially for less volatile compounds rubbers, and conifer resins). Blazso et al.[120] used thermally such as PAHs and n-alkanes.[108] Temperatures higher than assisted methylation and silylation in their pyrolysis-TD- 340◦Cmay cause thermally labile compounds to decom- GC/MS analysis of size-segregated aerosol samples from pose or pyrolyze.[132] aforest and semi-remote site in tropical Brazil, identi- Helium (He) is the typical carrier gas. Hays and fying fatty acids, alkandioic acids, hydroxybenzoic acids, Lavrich[24] suggested that the flow rate of He over the sam- levoglucosan, and other compounds of polysaccharide ple in the desorption unit may influence the desorption ef- origin. ficiency. Commercial TD units can be adjusted to different With pyrolysis-TD-GC/MS, the chemical nature of the flows and maintain a stable gas stream into the GC column sample is inferred from the molecular structure of its ther- and the detector. However, the relationship between He gas mal degradation products. Compounds detected can be py- flow and desorption efficiency has not been studied. rolysis or evaporation products from one or more parent Thermally desorbed organic compounds are usually molecules. Identification of parent compounds requires ex- cryo-focused, then heated and separated on an analyti- tensive mapping to their pyrolysis products. Quantitative cal column by their boiling points and polarities. Table 3 analysis is hindered by the presence of the large number shows that initial temperatures in the GC oven vary from of compounds detected and the effects of aerosol loading. −60◦C[46,127] to 70◦C,[118] with the most common selection However, pyrolysis-TD-GC/MS has the advantage of ob- at 30–50◦C. Higher initial temperatures may volatilize some taining structural information for polymeric and insoluble of the OM during sample loading. Studies using higher ini- organic compounds, which could not be otherwise obtained tial temperatures did not quantify VOCs adsorbed onto by conventional SE- or TD-GC/MS.[119,125,126] Pyrolysis- the particles. Ramping rates in the GC oven vary from Downloaded by [NIH Library] at 06:07 28 August 2017

Table 3. Summary of key parameters for TD-GC/MS and pyrolysis-GC/MS. TD Desorbent Initial GC Oven Total Sample Sample Analytical Temperature TD Refocusing GC Temperature Analysis Reference Type Preparation TD Unit Instrument Program Time Pathway GC Column Temperature Program Time 1. TD-GC/MS with Resistively Heated external oven Greaves Aerosol sample A 1.3 cm diameter piece of A cylindrical HP 5892A GC/MS Isocratic 15 min Cold trapping Ultra Performance ambient sample: ambient sample: ambient sample: 55.5 et al.a [46,127] and NIST SRM quartz filter was aluminum block in EI mode desorption at onto capillary fused silica capillary −50◦C or −60◦C −60◦Cto0◦Cat min NIST standard: Veltkamp 1649 punched and placed on containing a 254◦C. column head (25 m × 0.32 mm × NIST standard: 20◦C min−1, then 45.5 min et al.[128] the top of a glass frit of heating cartridge 0.52 µm); Ultra-2 20◦C to 300◦Cat8◦C a fritted glass sealing connected to a (25 m × 0.32 mm × min−1; NIST tube. The filter was thermocouple 0.17 µm) standard: 20◦Cto firmly held on the 180◦Cat20◦C position during min−1, then to sampling. The filter 300◦Cat8◦C after collection was then min−1 directly analyzed by TE. Waterman NIST SRM 1649a NIST urban dust was External oven HP 5890 GC/Fisons 175–300◦Cat 10 min Cryogenic Phenomenex ZB-5 (25 40◦C40◦Cfor 13 min, 90 min et al.[102] weighed, transferred mounted on the MD 800 MS, scan 50◦C min−1, refocusing at m × 0.25 mm × 25 increased to 300◦C into a glass-lined top of the range: 40–520 held at 300◦C −196◦C µm) at 5◦C min−1, held stainless steel GC liner GC/MS system amu for ∼7.5 min. during TE, at 300◦Cfor25 containing glass wool ramped to min.c and spiked with internal 300◦Cin20s. standards. Waterman NIST SRM 1649a. NIST urban dust standard Same as above HP 5890 GC/Fisons Isocratic 2 min Same as above Phenomenex ZB-5 (25 40◦C40◦Cfor 3 min, 90 min et al.[116] (1–5 mg) were weighed MD 800 MS, scan desorption at m × 0.25 mm × 25 increased to 300◦C and sealed into a glass range: m/z 40 to 300◦C µm) at 5◦C min−1, held tube filled with 120 520 at 300◦Cfor33 mesh glass beads. min.c Sidhu et al.[150] Aerosol collected Filter was sealed in a TD A stainless steel tube Two GCs and one 30–300◦Cat 13.5 min Cold trapping DB-5MS (30 m × 0.25 −60◦C Not available Ua on glass fiber tube. (0.635 cm O.D.) MS. The first GC 20◦C min−1 onto column µm) filters from placed in a GC is used as the TE head combustion of oven unit. The second alternative GC separates the diesel fuel. desorbent. Hays Aerosol collected A punch of filter was A glass tube placed Aglient 6890 25–300◦Cat ∼ 23min Cryogenic HP-5MS (30 m × 0.25 65◦C65◦Cfor 10 min, 99 min et al.[103,130]; from residential removed, spiked with in an external GC/5793 MSD, 12◦C min−1 focusing at mm × 0.25 µm) 60–300◦Cat10◦C Dong wood internal standards and oven (TDS2 scan range: 50 to -100◦Cinthe min−1, held at et al.[131] combustion, placed into a glass Gerstel Inc.) 500 amu inlet system, 300◦Cfor 41.5 min. residential oil desorption tubes. ramped to furnace and 300◦Cat fireplace 720◦C min−1. appliance. 2. Curie Point TD-GC/MS [105] 2 ◦ ◦ ◦ a Jeon et al. High-volume PM10 A 1.5 × 18 mm filter strip Curie point pyrolyzer HP 5890 GC/5792 From ambient to 10 s Cold trapping HP-5MS (25 m × 0.20 50 C50Cto320 Cat U ambient was cut and positioned MSD 315◦C onto capillary mm × 0.33 µm) 3–15◦C min−1 samples inside a glass reaction column head collected along tube lined with a the ferromagnetic foil U.S./Mexico border Neususs Ambient aerosol The filter strips were Curie point pyrolyzer Fisons Trio 1000 From ambient to 3.2 s Cold trapping Chrompack CP-Sil-5 50◦C50◦Cfor 2 min, 35 min et al.[151] collected during wrapped in a Pyrofoil GC/MS system 590◦C. onto capillary CB (30 m × 0.25 50–280◦Cat10◦C the 2nd Aerosol with a Curie point of column head mm × 0.1 µm) min−1, held at Characteriza- 590◦C and placed into 280◦Cfor10 min tion the pyrolyzer. Experiment (Continued on next page) 1531 Downloaded by [NIH Library] at 06:07 28 August 2017 1532

Table 3. Summary of key parameters for TD-GC/MS and pyrolysis-GC/MS. (Continued) TD Desorbent Initial GC Oven Total Sample Sample Analytical Temperature TD Refocusing GC Temperature Analysis Reference Type Preparation TD Unit Instrument Program Time Pathway GC Column Temp. Program Time 3. In-injection port TED-GC/MS Helmig et al.[92] Aerosol samples A small piece of sample GC injector port, Carlo Erba Mega Isocratic 15 min Cold trapping SE54 (15 m × 0.23 50◦C50◦Cto175◦Cat 47 min collected on filter was placed inside a with modified 5160 GC / VG desorption at onto capillary mm). 7.5◦C min−1, then glass-fiber filters glass liner septum cap. 250/70 SE MS, 320◦C column head to 295◦Cat10◦C ataforest site. scan range: min−1, held at 45–400 amu 295◦Cfor 5 min. Hall et al.[101] NIST SRM 1649 A glass reaction tube with Micro-scale sealed HP 5890 GC/Fisons 45-300◦Cat15◦C 22 min Cold trapping GC-5MS capillary 45◦C45◦Cfor 1 min, 82.5 min an internal volume of vessel placed MD800 MS, scan min−1 and onto capillary column (25 m × 45-300◦Cat6◦C 30–40 µLwas filled with inside the injector range: 40-500 amu held at 300◦C column head 0.25 mm ) min−1, held at 5mgNIST urban dust port. for 5 min 300◦Cfor 17 min. standard. The tube was sealed before TED. Blanchard and Aerosol samples A1cmdiameter section of AGCinjection port HP5890 GC/5972A Isocratic 15 min Cold trapping HP-5MS (30m × 0.25 30◦C30◦Cfor 0.75 min, 71 min Hopper[109]; collected on sample filter was wasadded with MS in EI mode desorption at onto capillary mm × 0.25µm) increased to 175◦C Blanchard quartz-and punched and transferred three minor 300◦C column head at 7.5◦C min,−1 et al.[129] glass-fiber filters into a GC liner spiked components, then to 295◦Cat in Ontario with internal standards. including a small 10◦C min, held at T-connector, 295◦Cfor 5 min. 3-way valve, and needle valve. Falkovich, and NIST SRM 1649a; NIST standard or small Direct Sample Varian Saturn 3400 120-350◦Cata 4.15 min Cold trapping DB-5MS column (30 m 40◦C40◦Cfor 4 min, 64.2 min Rudich[106]; urban aerosols piece of filter sample Introduction GC/MS 200◦C min−1 onto capillary × 0.25 mm × 0.25 40-300◦Cat5◦C Falkovich et collected with were loaded into a (DSI) device column head µm) min−1, held at al[152]; an 8 stage disposable microvial (ChromatoProbe, 300◦Cfor 4 min. Graham impactor in (1.9 mm ID × 12 mm Varian Co.). et al.[153] Tel-Aviv, Israel. long). The sample vials are placed by the vial holder in the GC liner. Ho and Yu[108]; Ambient aerosol Two pieces of 1.45 cm2 Conventional GC HP 5890 GC/5791 100-275◦C within 7.0 min Cold trapping HP-5MS (30 m × 0.25 30◦C30◦Cfor 2 min, 41.5 min Yang samples filter samples were injection port. No MSD, scan range: 7-8 min onto capillary mm × 0.25 µm) 30–120◦Cat20◦C et al.[154] collected on spiked with deuterated modification of 50-650 amu column head min−1, then to Teflon- internal standards, GC injector and 300◦Cat10◦C impregnated placed. in the GC inject liner. min−1, held at glass-fiber filters liner and held with glass 300◦Cfor 10 min. in Hong Kong wool. and on quartz-filters at Nanjing, China. 4. TD-GC × GC-MS Welthagen Ambient samples The filter strip was directly Injection port Optic Agilent 6890 50-350◦Cat 5 min Cold trapping BPX5 (50 m ×0.22 mm GCxGC 60◦Cfor 10 min, 175 min et al.[110]; in Augsburg, inserted into the direct III with GC/LECO 60◦C min−1 onto capillary × 0.25 µm) followed (2-D): 60◦C increased to 300◦C Schnelle- Germany thermal desorption autoloader Pegasus III TOF/ column head by BPX50 (1.5m × at 1.5◦C min−1. Kreis (DTD) liner and (ATAS-GL, MS with a LECO 0.10mm × 0.10 µm) The second column et al.[91] subsequently loaded Veldhoven, NL). Pegasus 4D was kept 5◦Cabove into a cold injection GCxGC the first column. port modulator Downloaded by [NIH Library] at 06:07 28 August 2017

[117] ◦ ◦ ◦ Hamilton et al. PM2.5 aerosol The filter strip (with 10µg Conventional GC The same as above, 40–300 Cata 13 min Cold trapping HP-5 (10 m × 0.18 mm 40 C40Cfor 5 min, 93.7 min collected in loading) was inserted injection port. scan range: 20 to 20◦C/ min−1 onto capillary × 0.18 µm) followed increased to 270◦C London. into the GC liner and 350 amu column head by DB17 (1.66 m × at 3.5◦C min−1, directly introduced into 0.1 mm × 0.1 µm) held at 270◦Cfor10 the GC injector. min. Hamilton et al.[118] Secondary organic The same as above. The same as above The same as above. The same as 12.5 min Cold trapping The same as above 70◦C70◦Cfor 2 min, 102.5 min aerosol formed above onto capillary 70–250◦Cat2.5◦C during the column head min−1, held at photo-oxidation 270◦Cfor 16 min. of toluene with OH radicals 5. In-situ semi-continuous and continuous TD systems William et al.[49] In-situ aerosol The aerosol s were Collection-TE cell Agilent 6890 50–300◦Cat 10 min Cold trapping Rtx-5MS (30 m × 0.25 45◦C45◦Cfor 12 min, 59 min samples collected into the with conventional GC/5793 MSD, 30–40◦C onto capillary mm × 0.25 µm) 45–300◦Cat8.5◦C collected in collection-TE cell and GC injection port. scan range: min−1 column head min1, held at 300◦C Berkely, CA thermally desorbed and 29–550 amu for ∼7 min. transferred into the GC injector. 6. Pyrolysis TD-GC/MS [126] ◦ ◦ ◦ ◦ Voorhees et al. PM0.6 and Filters were first extracted A tube furnace Extrel Simulscan 550 Cfor 10 min 10 min Cold trapping DB-5 (25 m × 0.25 35 C 35–220 Cat8.5 C 31.7 min −1 PM>0.45collected in sequence with directly interfaced GC/MS, scan onto capillary mm) min on quartz-fiber methanol, to an GC/MS. range: 35–450 column head in pristine dichloroethane, acetone, amu regions of and hexane. The Colorado resulting insoluble material was introduced to a tube furnace. Subbalakshmi Ambient aerosol Filter was introduced into Apyroinjector. Agilent6890 450◦Cfor 1 min 1 min Cold trapping BP-5 (30 m × 0.22 mm 40◦C 40–290◦Cat4◦C 63.5 min et al.[124] collected on a quartz capillary. GC/5973 MS, onto capillary × 1.0 µm) min−1 glass-fiber filters scan range: column head in Jakarta, 50–550 amu Indonesia [121] ◦ ◦ ◦ Fabbri et al. PM10 collected on Filter was scraped with Apyrolyzer directly Varian 3400 700 Cfor 10 s 10 s Cold trapping SPB-5 (30 m × 0.32 50 C50Cfor 2 min, 57 min glass-fiber filters tweezers and a weighted connected to the GC/Saturn II ion onto capillary mm × 0.25 µm) 50–300◦Cat5◦C in an industrial amount of the scraped GC injector port trap MS, scan column head min−1, held at area, Italy material was inserted through an range: 45 to 400 300◦Cfor 5 min. into a pre-pyrolyzed interface heated at amu quartz tube. 250◦C. [120] ◦ ◦ ◦ Blazso et al. PM2.6 colled on The collection substrates Apyrolyzer. Agilent 6890 400 Cfor 20 s 20 s Cold trapping HP-5 (30 m × 0.25 50 C50Cfor 1 min, 30.3 min quartz-fiber were spiked with GC/5973 MS, onto capillary mm) 50–300◦Cat10◦C filters and derivatization agents, scan range: column head min−1, held at size-segregated dried, and introduced 300◦Cfor 4 min. aerosol samples into a pyrolysis sample collected on Al holder. foils in Brazil ◦ ◦ ◦ ◦ Labban et al., 2006 PM10 of Filters were placed in Curie point HP5890GC/5972 740 Cfor 10 s 10 s Cold trapping Rtx-5MS (30 m × 0.25 50 C 50–220 Cat10 C 25.5 min resuspended soil pyrofoil. pyrolyzer. MS onto capillary mm × 0.25 µm) min−1, 220–320◦C collected on column head at 30◦C min1, held quartz-fiber at 320◦Cfor 5 min. filters a Total analysis time could not be determined because of insufficient experimental details. 1533 1534 Chow et al.

◦ 1.5◦C min−1[110] to 100◦C min−1,[91] with the most common is temperature-programmed at 100 Ctopurge water from ◦ in the range of 5–20◦C min−1. the system and then at 300 Ctoremove organic compounds followed by GC/MS-FID analysis. Desorbed analytes are Intercomparison between TD- and SE-GC/MS. focused onto the head of a GC column with the oven tem- ◦ Hays and Lavrich[24] tabulated compound classes from sev- perature set at 45 C. During the chromatographic separa- eral TD-GC/MS studies. These include alkanes (i.e., n- tion step and subsequent MS or FID detection, the system [49] alkanes, steranes, and hopanes), alkenes, aromatic (i.e., is switched back to the sampling mode. Williams et al. < 3 PAHs and furans), alcohols (i.e., phenols), amines, amides, concluded that a sampling time of 30 min ( 1m of air esters (i.e. phthalates), carboxylic acids, aldehydes, ketones, in several cases) in urban environments (i.e., Berkeley, CA) > and carboxylic acids. The quantification of PAHs by TD- provides sufficient sample to speciate 100 organic com- < GC/MS has been evaluated with Standard Reference Ma- pounds. With the current design, TAG provides 50% col- ∼ terial (SRM) 1649 (dusts collected in the urban Wash- lection efficiency for particles smaller than 65 nm, thereby ington, DC area in 1976–77 and sieved to <125 µm) underestimating ultrafine organic constituents. from the National Institute of Standards and Technology , , , , Thermal desorption-particle beam/mass spectrometer (NIST).[46 102 106 133 134] SE- and TD-GC/MS results were (TD-PB/MS) compared for PAHs and alkanes by Jeon et al.[105] and Ho and Yu.[108] TD-PB/MS identifies and quantifies organic species in These two methods agreed within a factor of 2.5 for the chamber-produced SOA and diesel engine nanoparti- [135,139,140] majority of PAHs and alkanes. Hansen et al.[90] character- cles. As shown in Figure 3b, particles are drawn −11 ized the TD recoveries of three alcohols (tridecanol, oc- into a high-vacuum chamber (6.6 × 10 atm), focused tadecanol, octacosanol), four carboxylic acids (nonanoic, onto a narrow, low-divergence particle beam using aero- dodecanoic, octadecanoic, and eicosanoic acid), two PAHs dynamic lenses followed by impaction onto a metal (Mo) (9-fluorenone and pyrene), and three source markers (nico- foil. The particles are either continuously vaporized by re- ◦ sistively heating the foil to a set temperature (e.g., 165◦C, tine, guaiacol, vanillin) at a TD temperature of 250 C. ◦ Only the two PAHs showed better than 75% recov- 250 C) for real-time analysis or heated using a linear tem- ◦ −1 eries with <55% for other compounds. Jeon et al.[105] perature ramp as low as 1 C/ min to achieve compound [135] showed good correlation between SE- and TD-GC/MS separation by volatility. In both temperature ramping for hopanes and steranes (R2= 0.998), PAHs (R2 = 0.978), methods, the vapors are continuously mass-analyzed using n-alkanes (R2 = 0.975), and phenols (R2= 0.940) with a quadrupole MS. Optimal separation of target compounds the exception of n-alkanoic acids (R2= 0.731). Neususs is achieved by adjusting the temperature ramping rate. With et al.[104] also demonstrated agreement within 130 ± 30% the linear-ramping temperature-programmed Thermal de- forPAHs between SE- and Curie point TD-GC/MS, except composition mode, separation was demonstrated for com- for the lower concentrations of indeno[1,2,3]pyrene and pounds having vapor pressures exceeding a factor of five. benzo[g,h,i]perylene by the TD method. Neususs et al.[104] Mass spectra of the individual components are inferred suggested that transfer losses of the less volatile compounds from the time-dependent spectra for compound identifi- might be the cause. Poor comparisons are probably due to cation. Size-dependent chemical composition can also be Downloaded by [NIH Library] at 06:07 28 August 2017 partial decomposition during the TD process. obtained by passing the incoming aerosol through a differ- ential mobility analyzer.[135] Continuous and semi-continuous TD aerosol TD-PB/MS provides higher time resolution and allows measurement systems study of aerosol chemistry dynamics than is possible with GC/MS analysis of filter samples. Sampling artifacts due TD has been applied in several first-generation semi- to vapor adsorption on or particle volatilization from the continuous and continuous instruments, typically coupled filter are much reduced.[140] Both laboratory experiments with MS.[48] Semi-continuous instruments integrate aerosol and theoretical calculations indicate that little evaporation sample collection with TD-GC/MS-FID analysis,[49] par- occurs in the ∼ 0.2 sec transit time from the TD-PB/MS , ticle beam/MS (PB/MS)[135 136] and chemical ioniza- inlet to the vaporizer for alkanes down to ∼C –C . Ana- , 15 17 tion/MS (CI/MS).[137 138] Figure 3 shows conceptual di- lytical artifacts caused by thermal decomposition are lower. agrams of these instruments. TD of unstable compounds, which can occur on a GC column, is reduced because vaporization during TD-PB/MS TD-GC/MS-FID occurs in a vacuum (∼1.3 × 10−5 atm).[139] Thermally la- Williams et al.[49] developed an in situ Thermal desorp- bile species such as α-alkoxylalkyl hydroperoxides and sul- [135,136,141] tion Aerosol GC (TAG)/MS-FID to achieve 30-min av- furic acid (H2SO4)are detectable by TD-PB/MS. erage organic speciation. As shown in Figure 3a, TAG im- Hydroperoxides would have dehydrated into correspond- pacts particles into a thermal desorption cell (TDC) after ing esters during GC analysis. A vaporization tempera- ◦ aPM2 cyclone and humidifier (to increase adhesion and ture of 165 Cwas demonstrated to completely evaporate minimize bounce). TD occurs under He in the TDC, which hexadecanoic acid, tridecanoic acid, and oleic acid.[139] In Thermal methods used to study carbonaceous aerosols 1535 Downloaded by [NIH Library] at 06:07 28 August 2017

Fig. 3. Conceptual diagrams of different in-situ continuous or semi-continuous thermal desorption (TD) instruments for: a) TD- GC/MS-FID or thermal desorption aerosol (TAG)/GC-MS-FID[49];b)TD-particle beam (PB)/MS[139]; and c) TD-chemical ionization (CI)/MS.[137]

comparison, TD with a GC inlet at 200◦C only removed Thermal desorption-chemical ionization/mass spectrometer 70% of octadecane in ambient aerosols.[108] (TD-CI/MS) TD-PB/MS captures ultrafine particles in the range of TD-CI/MS is designed for real-time chemical analysis of 20–500 nm in diameter and has quantified laboratory- ultrafine particles in the 6–20 nm in diameter using a high- generated SOA and nanoparticles from diesel engine ex- efficiency aerosol charger.[137,138] As shown in Figure 3c, haust. Quantification with an estimated uncertainty of 20% particles are charged and collected on a Ni/Cr wire. The col- is achieved using calibration techniques that account for lected particles are heated to 300◦Cinanionization cham- [139] particle shape and transport efficiency. The application ber for particle evaporation, followed by chemical ioniza- of TD-PB/MS to ambient aerosols is still untested. Nei- + − − tion by N2 buffer gas ions such as H3O ,O2 , and CO3 . The ther diesel particles nor chamber-generated SOA require buffer gas ions are produced by α-particles emitted from a prior knowledge of the types of products to be expected. radioactive material (i.e.,231Americium [Am]). The ions are Extraction of single compounds from the composite of the detected with a triple quadrupole MS. mass spectra of many compounds, in the case of ambient On-line TD-CI/MS has been used in selective ion moni- aerosols, is yet to be achieved. toring mode to size and identify inorganic ions (e.g., nitrate 1536 Chow et al.

− = + TD analysis, although their boiling points are lower than [NO3 ], sulfate [SO4 ], and ammonium [NH4 ]). The instrument has the potential to identify and detect OM in particle the desorption temperature. They also affect the quantifica- nuclei, but this has not yet been demonstrated. tion of some thermally stable compounds when their degradation products are targeted for quantification. For exam- Aerosol mass spectrometer (AMS) ple, n-alkanols degrade to their corresponding 1-alkenes, n-aldehydes, or possibly to n-alkane compounds.[89,93] The AMS integrates high-time resolution (5 sec) sampling and sizing with MS detection of aerosol components. Par- ticles passing through an aerodynamic lens are impacted ◦ − Future research directions onto a plate at 600 Cinavacuum (<1.3 × 10 6 atm). Flash volatilization converts PM constituents from con- On-filter derivatization techniques such as thermally as- densed phase to gaseous form, which are introduced into sisted alkylation using tetramethylammonium hydroxide, either a quadrupole- or TOF/MS detector. AMS also mon- tetrabutylammonium hydroxide, or silylation reagents are itors inorganic components that can be evaporated, in- − = + − likely to be applicable to TD-GC/MS with minor filter pre- cluding NO3 ,SO4 ,NH4 , chloride (Cl ), and potassium [120,121] + [142,143,144] treatment. On-line derivatization methods similar to (K ). Much progress has been made in interpret- [146] / those developed by Nimz and Morgan; Docherty and ing the ion fragments. For example, fragment ions of m z at Ziemann;[147] and Blokker et al.[148] appear to be feasible 57 and 44 are used to separate OM into hydrocarbon-like for measuring organic acids and alcoholic compounds. The organic compounds (HOCs) and oxygenated organic com- [145] derivatization reaction can proceed by filter spiking with pounds (OOCs). However, the de-convolution of single reagent or by spiking inert material in the GC injector with mass spectra from multiple organic compounds remains a reagent. Care must be taken with most derivatizing reagents daunting challenge. because of their hazardous nature and ability to damage analytical instruments. When TD methods are established for organic com- TD method limitations pounds, one must consider the possibility of the target compound undergoing thermal decomposition or being a de- Not all types of substrates are suitable for high temperature composition product of other coexisting constituents. This desorption.[92,131] Filters containing organic binders (e.g., can be achieved by evaluating recovery through spiking the Teflon with an operational temperature < ∼320◦C) would target compound onto aerosol matrices representative of decompose during the analysis. Sampling with quartz- or the real samples. Comparison with other methods (e.g., glass-fiber filters are the proper choices. SE-GC/MS) will also help to minimize mis-identification Compounds with similar volatilities limit the specificity of thermal decomposition products as organic compounds of TD-MS without chromatographic separation. In addi- originally present in the aerosols. Discrepancies between the tion, matrix effects associated with TD-MS analysis have independent methods need to be further examined. When been reported. Sakurai et al.[140] found that the desorption a compound identified in the TD approach is absent from profiles of diesel nanoparticles were influenced by loading the analysis using an alternative method for similar sam-

Downloaded by [NIH Library] at 06:07 28 August 2017 due to adsorption onto diesel soot particles that accumulate ples, there must be a reason to expect that the compound on the vaporizer. in question is actually present in the aerosol sample and Carryover, transfer loss, analyte conversion, and degra- not an artifact, such as thermolysis product arising from dation may interfere with some organic compounds. These the TD analytical procedure. Conversely, there should be analytical issues are inherently related to the presence of a chemical/physical reason why a compound was not de- OOCs. Polar functional groups (OOCs, e.g., carboxylic tected by one method or the other.[90] While method com- acids and alcohols) adsorb on surfaces inside the analytical parisons for an exhaustive list of organic compounds de- instruments (e.g., GC injection inlet), resulting in incom- tected by the TD methods is desired, priority should be plete desorption and subsequent carryover. These groups given to those species known to be source markers or of are also responsible for analyte conversion and degrada- toxicological importance. tion under thermal stress. Consistently low recoveries (less TD-MS has the potential to analyze unaltered thermally than 30%) of carboxylic acids and alkanols have been labile compounds because TD takes place in a vacuum. documented.[90,105,117] The presence of these compounds in This feature could be useful and should be explored for the aerosols is likely to impair quantification efforts and lead analysis of species such as organic peroxides that have low to compound misidentification. For example, even though thermal stability and are also sensitive to solvents. Williams et al.[49] identified polar levoglucosan, high mini- As mentioned, few TD-GC/MS experiments use the tem- mum detection limits (MDLs) and low regression values in perature programming option for separating and analyz- calibration raise uncertainties for its quantification. Highly ing organics. GC/MS analysis of PM evolves incremen- polar compounds such as dicarboxylic acids have not yet tally by decreasing GC column temperature and sample been quantified with acceptable analytical uncertainties in load, potentially increasing chromatographic resolution. In Thermal methods used to study carbonaceous aerosols 1537

addition, they might also support individual compound [3] Pope, C.A., III; Dockery, D.W.Health effects of fine particulate air identification with on-line MS measurements that em- pollution: Lines that connect. J. Air Waste Mgmt, Assoc 2006, 56 ploy specific temperature programming. This assumes that (6), 709–742. [4] Chow, J.C.; Watson, J.G.; Mauderly, J.L.; Costa, D.L.; Wyzga, nanogram or individual particle mass spectra can be de- R.E.; Vedal, S.; Hidy, G.M.; Altshuler, S.L.; Marrack, D.; Heuss, convoluted mathematically from temporally integrated ion J.M.; Wolff, G.T.; Pope, C.A., III; Dockery, D.W. 2006 critical re- chromatograms of the same sample. Nonetheless, this is sur- view discussion—Health effects of fine particulate air pollution: mised as a strong argument for the field deployment of both Lines that connect. J. Air Waste Mgmt. Assoc. 2006, 56 (10), 1368– real-time and time-integrated aerosol sampling instruments 1380. [5] Watson, J.G. Visibility: Science and regulation. J. Air Waste Mgmt. at ambient monitoring stations. Assoc. 2002, 52 (6), 628–713. Pyrolysis-TD-GC/MS can obtain polymeric and insol- [6] Chow, J.C.; Bachmann, J.D.; Wierman, S.S.G.; Mathai, C.V.; uble carbonaceous materials which could not be obtained Malm, W.C.; White, W.H.; Mueller, P.K.; Kumar, N.K.; Watson, with any SE-based methods. However, work is needed to J.G. Critical review discussion - Visibility: Science and regulation. understand the relationship between parent molecules and J. Air Waste Mgmt. Assoc. 2002, 52 (9), 973–999. [7] Jacobson, M.Z. Control of fossil-fuel particulate black carbon their pyrolysis products before this unique advantage can plus organic matter, possibly the most effective method of slow- be realized. ing global warming. J. Geophys. Research 2002 , 107 (D19), ACH 16-1-ACH 16–22. [8] Davidson, C.I.; Tang, W.; Finger, S.; Etyemezian, V.; Striegel, Summary M.F.; Sherwood, S.I. Soiling patterns on a tall limestone build- ing: Changes over 60 years. Environ. Sci. and Technol. 2000, 34 TD has been widely applied and proven for the determina- (4), 560–565. tion of aerosol OC and EC for decades. The basic princi- [9] Rogge, W.F.; Mazurek, M.A.; Hildemann, L.M.; Cass, G.R.; ples of thermal methods are now increasingly being applied Simoneit, B.R.T. Quantification of urban organic aerosols at a molecular level: Identification, abundance and seasonal variation. with advanced analytical detection of the evolved gases, Atmos. Environ. 1993, 27A (8), 1309–1330. furthering our understanding of aerosol chemistry. TD is a [10] Zheng, M.; Wan, T.S.M.; Fang, M.; Wang, F. Characterization viable and attractive alternative to liquid injection for off- of the non-volatile organic compounds in the aerosols in Hong line analysis of trace level organic compounds in aerosol Kong—Identification, abundance and origin. Atmos. Environ. samples. The most prominent advantages of the TD ap- 1997, 31 (2), 227–237. [11] Mazurek, M.A. Molecular identification of organic compounds proach are dramatic reductions of time and labor for sam- in atmospheric complex mixtures and relationship to atmospheric ple pre-treatment and improved sensitivity. Exploration of chemistry and sources. Environ. Health Perspect. 2002, 110(S6), TD methods to cover more organic species, especially po- 995–1003. lar organics, and overcoming the possible artifacts are nec- [12] Fine, P.M.;Chakrabarti, B.; Krudysz, M.; Schauer, J.J.; Sioutas, C. essary. While TD-MS has the potential to detect species Diurnal variations of individual organic compound constituents of ultrafine and accumulation mode particulate matter in the of low thermal stability and high sensitivity, pyrolysis-TD- Los Angeles basin. Environ. Sci. Technol. 2004, 38 (5), 1296– GC/MS has the potential to extend the analytical capability 1304. to insoluble carbonaceous materials. [13] Weyand, E.H.; Bevan, D.R. Benzo(a)pyrene disposition and metabolism in rats following intratracheal instillation. Cancer Res. 1986, 46, 5655–5661.

Downloaded by [NIH Library] at 06:07 28 August 2017 Acknowledgments [14] Bayona, J.M.; Casellas, M.; Fernandez, P.; Solanas, A.M.; Al- baiges, J. Sources and seasonal variability of mutagenic agents in This work was supported by the California Regional the Barcelona city aerosol. Chemosphere 1994, 29 (3), 441–450. [15] Hsiao, W.L.W.;Mo, Z.Y.;Fang, M.; Shi, X.M.; Wang, F. Cytotox- PM10/PM2.5 Air Quality Study (CRPAQS) Agency under icity of PM2.5 and PM2.5−10 ambient air pollutants assessed by the the management of the California Air Resources Board MTT and the Comet assays. Mutat. Res. Genetic Toxicol. Environ. and by the U.S. EPA under Contract #R-82805701 for Mutagen. 2000, 471 (1–2), 45–55. the Fresno Supersite. Additional support was provided by [16] Chow, J.C.; Watson, J.G.; Lowenthal, D.H.; Chen, L.W.A.; Zielin- the EPA STAR Grant RD-83108601-0 and the Electric ska, B.; Mazzoleni, C.R.; Mogliano, K.L. Evaluation of organic markers for chemical mass balance source apportionment at the PowerResearch Institute (EPRI) under contract number Fresno supersite. Atmospheric Chemistry and Physics, 2007, 7(7), EP-P19865/C9746. 1741–2754. [17] Simoneit, B.R.T. Application of molecular marker analysis to vehicular exhaust for source reconciliations. International Journal of References Environmental Analytical Chemistry 1985, 22, 203–233. [18] Rogge, W.F.; Hildemann, L.M.; Mazurek, M.A.; Cass, G.R.; Si- [1] Vedal, S. Critical review—Ambient particles and health: Lines that moneit, B.R.T.Sources of ﬁne organic aerosol—6. Cigarette smoke divide. J. Air Waste Mgmt. Assoc. 1997, 47 (5), 551–581. in the urban atmosphere. Environ. Sci. Technol. 1994, 28(7), 1375– [2] Watson, J.G.; Wyzga, R.E.; Burnett, R.T.; Vostal, J.J.; Romieu, I.; 1388. Chow,J.C.; Holcombe, J.T.;Lipfert, F.W.; Marrack, D.; Thompson, [19] Schauer, J.J.; Kleeman, M.J.; Cass, G.R.; Simoneit, B.R.T. Mea- R.J.;Vedal, S. Ambient particles and health: Lines that divide: 1997 surement of emissions from air pollution sources—2. C1 through critical review discussion. J. Air Waste Mgmt. Assoc. 1997, 47 (9), C30 organic compounds from medium duty diesel trucks. Environ. 995–1008. Sci. Technol. 1999, 33 (10), 1578–1587. 1538 Chow et al.

[20] Simoneit, B.R.T.A review of biomarker compounds as source indi- [39] Woolfenden, E. Monitoring VOCs in air using sorbent tubes fol- cators and tracers for air pollution. Environ. Sci. Pollut. Res. 1999, lowed by thermal desorption-capillary GC analysis: Summary of 6 (3), 159–169. data and practical guidelines. J. Air Waste Mgmt. Assoc. 1997, 47 [21] Schauer, J.J.; Cass, G.R. Source apportionment of wintertime gas- (1), 20–36. phase and particle-phase air pollutants using organic compounds [40] U.S. EPA. Compendium of methods for the determination of toxic as tracers. Environ. Sci. Technol. 2000, 34 (9), 1821–1832. organic compounds in ambient air, Method TO-17: Determination [22] Watson, J.G.; Chowe, J.C. Receptor Models. In Air Quality of volatile organic compounds in ambient air using active sampling Modeling-Theorics, Methodologies, Computational Techniques, onto sorbent tubes,U.S. EPA, Washinngton, DC; 1997. and Available Databases in Software. Vol II-Advanced Topics, P. [41] Lamanna, M.S.; Goldstain, A.H. In-situ measurements of C2-C10 Zanett., Ed.; Air and Waste Management Association and Environ VOCs above a Sierra Nevada ponderosa pine plantation. J. Geo- Comp Institute, Pittsburgh, PA, 2005, pp 455–501. phys. Res. 1999, 104, 21247–21262. [23] Watson, J.G.; Chow, J.C. Receptor models for source apportion- [42] Sunesson, A.L.; Sundgren, M.; Levin, J.O.; Eriksson, K.; Carlson, ment of suspended particles. In: Introduction to Environmental R. Evaluation of two adsorbents of diffusive sampling and thermal Forensics, 2nd Edition; Murphy, B.; Morrison, R., Eds.; Academic desorption-gas chromatographic analysis of monoterpenes in air. Press: New York, 2007; 279–316. J. Environ. Monitor. 1999, 1, 45–50. [24] Hays, M.D.; Lavrich, R.J. Developments in direct thermal [43] Chow, J.C. Critical review: Measurement methods to determine extraction-GC/MS of fine aerosols. Trends Anal. Chem. 2007, 26 compliance with ambient air quality standards for suspended par- (2), 88–102. ticles. J. Air Waste Mgmt. Assoc. 1995, 45 (5), 320–382. [25] Chow, J.C.; Watson, J.G.; Pritchett, L.C.; Pierson, W.R.; Frazier, [44] Schuetzle, D.; Crittenden, A.L.; Charlson, R.J.Application of com- C.A.; Purcell, R.G. The DRI Thermal/Optical Reflectance carbon puter controlled high resolution mass spectrometry to the analysis analysis system: Description, evaluation and applications in U.S. of air pollutants. J. Air Poll. Control Assoc. 1973, 23 (8), 704–709. air quality studies. Atmos. Environ. 1993, 27A (8), 1185–1201. [45] Schuetzle, D.; Cronn, D.; Crittenden, A.L.; Charlson, R.J. Molec- [26] Chow, J.C.; Watson, J.G.; Crow, D.; Lowenthal, D.H.; Merrifield, ular composition of secondary aerosol and its possible origin. T.M. Comparison of IMPROVE and NIOSH carbon measure- Enivron. Sci. Technol. 1975, 9 (9), 838–845. ments. Aerosol Sci. Technol. 2001, 34 (1), 23–34. [46] Greaves, R.C.; Barkley, R.M.; Sievers, R.E. Rapid sampling and [27] Chow, J.C.; Watson, J.G.; Chen, L.-W.A.; Arnott, W.P.; analysis of volatile constituents of airborne particulate matter. Moosmuller,¨ H.; Fung, K.K. Equivalence of elemental carbon Anal. Chem. 1985, 57 (14), 2807–2815. by Thermal/Optical Reflectance and Transmittance with differ- [47] Veltkamp, P.R.; Hansen, K.J.; Barkley, R.M.; Sievers, R.E. Princi- ent temperature protocols. Environ. Sci. Technol. 2004, 38 (16), pal component analysis of summertime organic aerosols at Niwot 4414–4422. Ridge, Colorado. J. Geophy. Res. 1996, 101 (D14), 19495–19504. [28] Chow, J.C.; Watson, J.G.; Chen, L.W.A.; Paredes-Miranda, G.; [48] Jayne, J.T.; Leard, D.L.; Zhang; X.; Davidovits, P.; Smith, K.A.; Chang, M.-C.O.; Trimble, D.; Fung, K.K.; Zhang, H.; Yu, J.Z. Kolb, C.E.; Worsnop, D.E. Development of an aerosol mass spec- Refining temperature measures in thermal/optical carbon analy- trometer for size and composition analysis of submicron particles. sis. Atmos. Chem. Phy. 2005, 5 (4), 2961–2972. Aerosol Sci. Technol. 2000, 33, 49–70. [29] Chow, J.C.; Watson, J.G.; Chen, L.-W.A.; Chang, M.C.O.; Robin- [49] Williams, B.J.; Goldstein, A.H.; Kreisberg, N.M.; Hering, S.V. son, N.F.; Trimble, D.L.; and Kohl, S.D. The IMPROVE A tem- An in-situ instrument for speciated organic composition of at- perature protocol for thermal/optical carbon analysis. Maintain- mospheric aerosols: Thermal Desorption Aerosol GC/MS-FID ing consistency with a long-term database. J. Air Waste Manage. (TAG). Aerosol Sci. Technol. 2006, 40 (8), 627–638. Assoc., 2007. [50] Atkins, P.W. Physical Chemistry, 7th; Oxford University Press, New [30] Watson, J.G.; Chow, J.C.; Chen, L.-W.A.Summary of organic and Yo rk, 2002. elemental carbon/black carbon analysis methods and intercom- [51] Seinfeld, J.H.; Pandis, S.N. Atmospheric Chemistry and Physics: parisons. Aerosol Air Qual. Res. 2005, 5 (1), 65–102. From Air Pollution to Climate Change,John Wiley & Sons: New [31] DeLuca, S.; Sarver, E.W.;de B.Harrington, P.; Voorhees, J.J. Direct Yo rk, 1998.

Downloaded by [NIH Library] at 06:07 28 August 2017 analysis of bacterial fatty acids by curie-point pyrolysis tandem [52] Sears, F.W.; Zermansky, M.W.; Young, H.D. University Physics, mass spectrometry. Anal. Chemi. 1990, 62, 1465–1472. 6th ed. ; Addison Wesley, Boston, MA, 1982. [32] Helmig, D.; Arey, J. Organic chemicals in the air at Whitaker’s [53] Stout, J.W.; Fisher, L.H. The entropy of formic acid. J. Chem. Phy. Forest Sierra Nevada Mountains, California. Sci. Total Environ. 1941, 9 (2), 163–168. 1992, 112, 233–250. [54] McMurry, J. Organic Chemistry, 5th ed.; Thomson Brooks/Cole: [33] Hirvonen, A.; Pasanen, P.; Tarhanen, J.; Ruuskanen, J. Thermal Belmont, CA, 1999. desorption of organic compounds associated with settled house- [55] Blake, E.S.; Hammann, W.C.; Edwards, J.W.; Reichard, T.E.; Ort, hold dust. Indoor Air 1994, 4, 255–264. M.R. Thermal stability as a function of chemical structure. J. [34] Marbot, R. The selection of pyrolysis temperatures for the analysis Chem. Eng. Data 1961, 6, 87–98. of humic substances and related materials 1. Cellulose and chitin. [56] Rice, F.O. The thermal decomposition of organic compounds from J. Anal. Appl. Pyrolysis 1997, 39, 97–104. the standpoint of free radicals I. Saturated hydrocarbons. J. Am. [35] Sigman, M.E.; Ma, C.Y. In-injection port thermal desorption for Chem. Soc. 1931, 53, 1959–1972. explosives vapor evidence analysis. Anal. Chem. 1999, 71 (19), [57] Rice, F.O.; Johnston, W.R.; Evering, B.L. The thermal decompo- 4119–4124. sition of organic compounds from the standpoint of free radi- [36] Tran, N.K.; Steinberg, S.M.; Johnson, B.J. Volatile aromatic hydro- cals II. Experimental evidence of the decomposition of organic carbons and dicarboxylic acid concentrations in air at an urban site compounds into free radicals. J. Am. Chem. Soc. 1932, 54, 3529– in the southwestern U.S. Atmos. Environ. 2000, 34 (11), 1845–1852. 3543. [37] Liu, N.A.; Fan, W.; Dobashi, R.; Huang, L. Kinetic modeling of [58] Johns, I.B.; McElhill, E.A.; Smith, J.O. Thermal stability of organic thermal decomposition of natural cellulosic materials in air atmo- compounds. I&EC Prod. Res. Devel. 1962, 1, 2–6. sphere. J. Anal. Appl. Pyrol. 2002, 63, 303–325. [59] Johns, I.B.; McElhill, E.A.; Smith, J.O. Thermal stability of some [38] Coleman, W.M., III; Chung, H.L. Pyrolysis GC-MS analysis of organic compounds. J. Chem. Eng. Data 1962, 7, 277–281. Amadori compounds derived from selected amino acids with glu- [60] Madison, J.J.; Roberts, R.M. Pyrolysis of aromatic and related het- cose and rhamnose. J. Anal. Appl. Pyrol. 2002, 63, 349–366. erocyclics. Ind. Eng. Chem. 1957, 50, 237–250. Thermal methods used to study carbonaceous aerosols 1539

[61] Johnson, W.D. Thermal decomposition of alcohols. 1,2 [83] Arnold, R.T.; Elmer, O.C.; Dodson, R.M. Thermal decarboxyla- methylbutan-2-ol (t-pentyl alcohol). Australian J.Chem. 1974, 27 tion unsaturated acids. J. Am. Chem. Soc. 1950, 72, 4359–4361. (5), 1047–1052. [84] Arnold, R.T.; Danzig, M.J. Thermal decarboxylation unsaturated [62] Schwarz, E.M.; Grundstein, V.V.; Ievins, A.F. Thermal investiga- acids. II. J. Am. Chem. Soc. 1957, 79, 892–894. tion of polyols. Therm. Anal. and Calor. 1972, 4, 331–337. [85] Jahngen, E.; Mallett, J.; O’Connor, R.; Fischer, S. Mechanism [63] Song, J.; Zhan, B.; Li, P. A study of SO2 in fogwater of Qingdao. of the decarboxylative rearrangement of a(carbonyl)cyclopropane Mar. Sci. 1994, 2, 66–69. carboxylic acids to 2-substituted-4,5-dihydrofurans. ARKIVOC [64] McNaught, A.D.; Wilkinson, A. IUPAC Gold Book Compendium 2007, ix, 135–149. of Chemical Terminology, 2nd ; Royal Society of Chemistry: Cam- [86] List, G.R.; Evans, C.D.; Selke, E.; Glass, C.A.; Hoffmann, R.L.; bridge, UK, 1997. McManis, G.E. Pyrolysis of some acetoxynonenes. Lipids 1971, 6 [65] Demirbas, A. Effects of temperature and particle size on bio-char (9), 635–640. yield from pyrolysis of agricultural residues. J. Anal. Appl. Pyrol. [87] Garnett, J.L.; Johnson, W.D.; Sherwood, J.E. Pyrolysis of 1- 2004, 72 (2), 243–248. methylcyclohexene and methylenecyclohexane. J. Chem. 1976, 29 [66] Napier, I.M.; Norrish, R.G.W. The photolysis and pyrolysis of ni- (3), 599–607. tromethane and methyl nitrite. Proc. Roy. Soc. Lond. A., Math. [88] Morgan, K.M. Reaction mechanisms Part (ii) Pericyclic reactions. Phy. Sci. 1967, 299, 317–336. Annu. Rep. Prog. Chem., Sect B 2000, 96, 321–345. [67] Yuan, W.; Rabinovitch, B.S.; Tosa, R. Vibrational energy transfer [89] Helmig, D.; Bauer, A.; Muller, J.; Klein, W.Discussion on analysis and pyrolysis of nitromethane by the variable encounter method. of particulate organics in a forest atmospher by thermodesorption J. Phys. Chem. 1982, 86, 2796–2799. GC/MS. Atmos. Environ. 1992, 26A, 517–519. [68] Merkulov, A.A.; Ovsyannikov, A.A.; Polak, L.S.; Popov, V.T.; [90] Hansen, K.J.; Hansen, B.N.; Cravens, E.; Sievers, R.E. Supercriti- Pustilnikov, V.Y. Initial stages of soot formation in thermal py- cal fluid extraction-gas chromatographic analysis of organic com- rolysis of acetylene I. Mechanism for homogeneous pyrolysis of pounds in atmospheric aerosols. Anal. Chem. 1995, 67 (19), 3541– acetylene. Plasma Chem. Plasma Proc. 1989, 9 (1), 95–104. 3549. [69] Buchanan, A.C., III; Britt, P.F.; Skeen, J.T.; Struss, J.A.; Elam, [91] Schnelle-Kreis, J.; Welthagen, W.; Sklorz, M.; Zimmermann, R. C.L. Pyrolysis of silica-mobilized benzyl phenyl ether: Competing Application of direct thermal desorption gas chromatography and radical rearrangement pathways under restricted diffusion. J. Org. comprehensive two-dimensional gas chromatography coupled to Chem. 1998, 63 (26), 9895–9903. time of flight mass spectrometry for analysis of organic compounds [70] Fulle, D.; Dib, A.; Kiefer, J.H.; Zhang, Q.; Yao, J.; Kern, R.D. Py- in ambient aerosol particles. J. Sep. Sci. 2005, 28, 1648–1657. rolysis of furan at low pressures: Vibrational relaxation, unimolec- [92] Helmig, D.; Bauer, A.; Muller, J.; Klein, W.Analysis of particulate ular dissociation, and incubation times. J. Phys. Chem. A 1998, 102 organics in a forest atmosphere by thermodesorption GC/MS. At- (38), 7480–7486. mos. Environ. 1990, 24A (1), 179–184. [71] Evans, C.S.; Dellinger, B. Formation of bromochlorodibenzo-p- [93] Simoneit, B.R.T.Discussion on analysis of particulate organics in a dioxins and dibenzofurans from the high-temperature oxidation forest atmosphere by thermodesorption GC/MS. Atmos. Environ. of a mixture of 2-chlorophenol and 2-bromophenol. Environ. Sci. 1992, 26A, 517–519. Technol. 2006, 40 (9), 3036–3042. [94] Yu, J.Z.; Xu, J.H.; Yang, H. Charring characteristics of atmo- [72] Bradley, J.N. A general mechanism for the high-temperature pyrol- spheric organic particulate matter in thermal analysis. Environ. ysis of alkanes. The pyrolysis of isobutane. Proc. Roy. Soc. Lond. Sci. Technol. 2002, 36 (4), 754–761. A., Math. Phys. Sci. 1974, 337, 199–216. [95] Murphy, D.M.; Thomson, D.S. Laser ionization mass spectroscopy [73] Rumyantsev, A.N.; Oganesova, E.Y. Mechanism of pyrolysis of of single aerosol particles. Aerosol Sci. Technol. 1995, 22, 237– n-alkanes 1. N-butane. Chem. Technol. Fuels Oils 1987, 23 (1), 249. 35–37. [96] Noble, C.A.; Prather, K.A. Real-time measurement of correlated [74] Watanabe, M.; Adschiri, T.; Arai, K. Overall rate constant of py- size and composition profiles of individual atmospheric aerosol rolysis of n-alkanes at a low conversion level. Ind. Eng. Chem. Res. particles. Enivron. Sci. Technol. 1996, 30 (9), 2667–2680.

Downloaded by [NIH Library] at 06:07 28 August 2017 2001, 40 (9), 2027–2036. [97] Cronn, D.R.; Charlson, R.J.; Knights, R.L.; Crittenden, A.L.; [75] Smith, R.D. A direct mass spectrometric study of the mechanism Appel, B.R. A survey of the molecular nature of primary and sec- of toluene pyrolysis at high temperatures. J. Phys. Chem. 1979, 83 ondary components of particles in urban air by high-resolution (12), 1553–1563. mass spectrometry. Atmos. Environ. 1977, 11, 929–937. [76] Brouwer, L.D.; Muller-Markgraf, W.;Troe, J. Thermal decomposi- [98] Kawamura, K.; Kaplan, I.R. Motor exhaust emissions as a primary tion of toluene: A comparison of thermal and laser photochemical source for dicarboxylic acids in Los Angeles ambient air. Enivron. activation experiments. J. Chem. Phys. 1988, 92, 4905–4914. Sci. Technol. 1987, 21, 105–110. [77] Hippler, H.; Reihs, C.; Troe, Z.J. Elementary steps in the pyrolysis [99] Poore, M.W.Oxalic acid in PM2.5 particulate matter in California. of toluene and benzy radicals. Phys. Chem. 1990, 167, 1–16. J. Air Waste Mgmt. Assoc. 2000, 50 (11), 1874–1875. [78] Damm, M.; Deckert, F.; Hippler, H.; Rink, G. Specific rate con- [100] Perry, R., Long, R., and Majer, J.R. The Use of Mass Spectrometry stants for the fragmentation of vibrationally excited benzyl radicals. in the Analysis of Air Pollutants 1971, 557–562. Phys. Chem. Chem. Phys. 1999, 1, 81–90. [101] Hall, P.A.; Watson, A.F.R.; Garner, G.V.; Hall, K.; Smith, S.; [79] Boyd, J.; Overton, K.H. 1,3-Elimination with rearrangement dur- Waterman, D.; Horsfield, B. An investigation of micro-scale sealed ing ester pyrolysis. A simple route to protoadamantene and 2,4- vessel thermal extraction-gas chromatography-mass spectrome- didehydroadamantane. J. Chem. Soc. Perkin Trans. 1972,1, 2533– try (MSSV-GC-MS) and micro-scale sealed vessel pyrolysis-gas 2539. chromatography-mass spectrometry applied to a standard refer- [80] Brown, R.F.C. Pyrolytic Methods in Organic Chemistry.Application ence material of an urban dust/organics. Sci. Total Environ. 1999, of Flow and Flash Vacuum Pyrolytic Techniques. Academic Press: 235, 269–276. New York, 1980. [102] Waterman, D.; Horsfield, B.; Leistner, F.;Hall, K.; Smith, S. Quan- [81] Dominguez, R.M.; Chuchani, G. The direction of elimination in tification of polycyclic aromatic hydrocarbons in the NIST stan- the pyrolysis kinetics of 3-bromobutyronitrile. Reaction Kinetics dard reference material (SRM1649a) urban dust using thermal des- Catal. Lett. 1992, 47 (2), 247–254. orption GC/MS. Anal. Chem. 2000, 72 (15), 3563–3567. [82] Moss, R.A.; Platz, M.S.; Jones, M., Jr. Reactive Intermediate Chem- [103] Hays, M.D.; Smith, N.D.; Kinsey, J.; Dong, Y.; Kariher, P. Poly- istry,John Wiley & Sons, Inc. Hoboken, NJ, 2005. cyclic aromatic hydrocarbon size distributions in aerosols from 1540 Chow et al.

appliances of residential wood combustion as determined by di- [120] Blazso, M.; Janitsek, S.; Gelencser, A.; Artaxo, P.; Graham, B.; An- rect thermal desorption—GC/MS. J. Aerosol Sci. 2003, 34 (8), dreae, M.O. Study of tropical organic aerosol by thermally assisted 1061–1084. alkylation-gas chromatography mass spectrometry. J. Analy. Appl. [104] Neususs, C.; Pelzing, M.; Plewka, A.; Herrmann, H. A new analyt- Pyrol. 2003, 68–9, 351–369. ical approach for size-resolved speciation of organic compounds [121] Fabbri, D.; Pratis, S.; Vassura, I. Molecular characterization of in atmospheric aerosol particles: Methods and first results. J. Geo- organic material in air fine particles (PM10) using conventional phys. Res. 2000, 105 (D4), 4513–4527. and reactive pyrolysis-gas chromatography-mass spectrometry. J. [105] Jeon, S.J.; Meuzelaar, H.L.C.; Sheya, S.A.N.; Lighty, J.S.; Jarman, Environ. Monitor. 2002, 4, 210–215. W.M.; Kasteler, C.; Sarofim, A.F.; Simoneit, B.R.T. Exploratory [122] Schulten, H.R.; Leinweber, P. Characterization of humic and soil studies of PM10 receptor and source profiling by GC/MS and particles by analytical pyrolysis and computer modeling. J. Anal. principal component analysis of temporally and spatially resolved Appl. Pyrol. 1996, 38, 1–53. ambient samples. J. Air Waste Mgmt. Assoc. 2001, 51 (5), 766– [123] Leinweber, P.; Schulten, H.R. Advances in analytical pyrolysis of 784. soil organic matter. J. Anal. Appl. Pyrol 1999, 49, 359–383. [106] Falkovich, A.H.; Rudich, Y.Analysis of semivolatile organic com- [124] Subbalakshmi, Y.; Patti, A.F.; Lee, G.S.H.; Hooper, M.A. Struc- pounds in atmospheric aerosols by direct sample introduction ther- tural characterization of macromolecular organic material in air mal desorption GC/MS. Environmental Science & Technology particulate matter using Py-GC-MS and solid state C-13NMR. J. 2001, 35(11), 2326–2333. Environ. Monitor. 2002, 2, 561–565. [107] Wauters, E.; Sandra, P.; Verzele, M. Qualitative and semi- [125] Voorhees, K.J.; Schulz, W.D.; Currie, L.A.; Klouda, G.A. An in- quantitative analysis of the non-polar organic fraction of air par- vestigation of the insoluble carbonaceous material in airborne particulate matter. J. Chromatog. 1979, 170, 125–131. ticulates from vehicular traffic. J. Anal. Appl. Pyrol. 1988, 14, [108] Ho, S.S.H.; Yu, J.Z. In-injection port thermal desorption and sub- 83–98. sequent gas chromatography-mass spectrometric analysis of poly- [126] Voorhees, K.J.;Schulz, W.D.; Kunen, S.M.; Hendricks, L.J.;Currie, cyclic aromatic hydrocarbons and n-alkanes in atmospheric aerosol L.A.; Klouda, G.A. Analysis of insoluble carbonaceous materials samples. J. Chromatogr. A 2004, 1059(1–2), 121–129. from airborne particles collected in pristine regions of Colorado. [109] Blanchard, P. and Hopper, J. F. Concentrations and distributions of J. Anal. Appl. Pyrol. 1991, 18 (3–4), 189–205. PAHs and n-alkanes in atmospheric aerosols at non-urban sites in [127] Greaves, R.C.; Barkley, R.M.; Sievers, R.E.; Meglen, R.R. Co- Ontario, Canada; 6th International Conference on Carbonaceous variations in the concentrations of organic compounds associated Particles in the Atmosphere, Austr. Soc. of Anal. Chem., Austr. with springtime atmospheric aerosols. Atmos. Environ. 1987, 21 Acad. of Sci.: Vienna, Austria, 1997. (12), 2549–2561. [110] Welthagen, W.; Schnelle-Kreis, J.; Zimmermann, R. Search [128] Veltkamp, P.R.; Hansen, K.J.; Barkley, R.M.; Sievers, R.E. Chro- criteria and rules for comprehensive two-dimensional gas matographic measurement of molecular markers of sources of at- chromatography-time-of-flight mass spectrometry analysis of air- mospheric aerosol particles. Environ. Geochem. Health 1996, 18, borne particulate matter. J. Chromatog. A 2003, 1019 (1–2), 233– 77–80. 249. [129] Blanchard, P.; Brook, J.R.; Brazal, P.Chemical characterization of [111] Schnelle-Kreis, E.; Sklorz, M.; Peters, A.; Cyrys, J.; Zimmer- the organic fraction of atmospheric aerosol at two sites in Ontario, mann, R. Analysis of particle-associated semi-volatile aromatic Canada. J. Geophys. Res. 2002, 107 (D21), 8348. and aliphatic hydrocarbons in urban particulate matter on a daily [130] Hays, M.D.; Smith, N.D.; Dong, Y.J.Nature of unresolved complex basis. Atmos. Environ. 2005, 39 (40), 7702–7714. mixture in size-distributed emissions from residential wood com- [112] Beuhler, R.J.; Flanlgan, E.; Greene, L.T.; Friedman, L.J. Proton bustion as measured by thermal desorption-gas chromatography- transfer mass spectrometry of peptides. A rapid heating technique mass spectrometry. J. Geophys. Res. 2004, 109 (D16). for underivatized peptides containing arginine. J. Am. Chem. Soc. [131] Dong, Y.J.; Hays, M.D.; Smith, N.D.; Kinsey, J.S. Inverting cascade 1974, 96, 3990–3999. impactor data for size-resolved characterization of fine particulate [113] Kiss, G.; Gelencser, A.; Krivacsy, Z.; Hlavay, J. Occurrence and source emissions. J. Aerosol Sci. 2004, 35(12), 1497–1512.

Downloaded by [NIH Library] at 06:07 28 August 2017 determination of organic pollutants in aerosol, precipitation, and [132] Cachier, H.; Bremond, M.P.; Buat-Menard,´ P. Thermal separa- sediment samples collected at Lake Balaton. J.Chromatogr.A 1997, tion of soot carbon. Aerosol Sci. Technol. 1989, 10(2), 358– 774, 349–361. 364. [114] Odabasi, M.; Ongan, O.; Cetin, E. Quantitative analysis of volatile [133] Waterman, D.; Horsfield, B.; Hall, K.; Smith, S. Application of organic compounds (VOCs) in atmospheric particles. Atmos. En- micro-scale sealed vessel thermal desorption-gas chromatography- viron. 2005, 39 (20), 3763–3770. mass spectrometry for the organic analysis of airborne particu- [115] Watson, J.G.; Chow, J.C. Ambient air sampling. In Aerosol Mea- late matter: Linearity, reproducibility and quantification. J. Chro- surement: Principles, Techniques, and Applications, 2nd Ed. Baron, matogr. A 2001, 912, 143–150. P. ;Willeke, K., Eds. John Wiley & Sons, New York, 2001; 821–844. [134] Sin, D.W.M.; Fung, W.H.; Lam, C.H. Measurement of car- [116] Waterman, D.; Smith, S.; Green, D.; Horsfield, B.; Hall, K. The bonaceous aerosols: Validation and comparison of a solvent application of a thermal desorption GCMS technique for the or- extraction—gas chromatographic method and a thermal optical ganic analysis of airborne particulate matter. Adv. Mass Spectrom transmittance method. Analyst 2002, 127(5), 614–622. 2001; 887–898. [135] Tobias, H.J.; Ziemann, P.J. Compound identification in organic [117] Hamilton, J.F.; Webb, P.J.; Lewis, A.C.; Hopkins, J.R.; Smith, S.; aerosols using temperature-programmed thermal desorption par- Davy, P. Partially oxidised organic components in urban aerosol ticle beam mass spectrometry. Anal. Chem. 1999, 71(16), 3428– using GCXGC-TOF/MS. Atmos. Chem. Phys. 2004, 4, 1279–1290. 3435. [118] Hamilton, J.F.; Webb, P.J.; Lewis, A.C.; Reviejo, M.M. Quantify- [136] Tobias, H.J.;Ziemann, P.J.Thermal desorption mass spectrometric ing small molecules in secondary organic aerosol formed during analysis of organic aerosol formed from reactions of 1-tetradecene the photo-oxidation of toluene with hydroxyl radicals. Atmos. En- and O3 in the presence of alcohols and carboxylic acids. Environ. viron. 2005, 39 (38), 7263–7275. Sci. Technol. 2000, 34(11), 2105–2115. [119] Labban, R.; Veranth, J.M.; Watson, J.G.; Chow, J.C. Feasibility of [137] Voisin, D.; Smith, J.N.; Sakurai, H.; McMurry, P.H.; Eisele, F.L. soil dust source apportionment by the pyrolysis-gas chromatogra- Thermal desorption chemical ionization mass spectrometer for ul- phy/mass spectrometry method. J. Air Waste Mgmt. Assoc. 2006, trafine particle chemical composition. Aerosol Sci. Technol. 2003, 56 (9), 1230–1242. 37, 471–475. Thermal methods used to study carbonaceous aerosols 1541

[138] Smith, J.N.; Moore, K.F.;McMurry,P.H.;Eisele, F.L.Atmospheric tion of hydrocarbon-like and oxygenated organic aerosols based measurements of sub-20 nm diameter particle chemical composi- on aerosol mass spectrometry. Environ. Sci. Technol. 2005, 39(13), tion by thermal desorption chemical ionization mass spectrometry. 4938–4952. Aerosol Sci. Technol. 2004, 38(2), 100–110. [146] Nimz, E.L.; Morgan, S.L. On-line derivatization for complex fatty [139] Tobias, H.J.; Kooiman, P.M.; Docherty, K.S.; Ziemann, P.J. acid mixtures by capillary gas chromatography/mass spectrometry. Real-time chemical analysis of organic aerosols using a thermal J. Chromatog. Sci. 1993, 31, 145–149. desorption particle beam mass spectrometer. Aerosol Sci. Tech- [147] Docherty, K.S.; Ziemann, P.J. On-line, inlet-based trimethylsilyl nol. 2000, 33(1–2), 170–190. derivatization for gas chromatography of mono-and dicarboxylic [140] Sakurai, H.; Tobias, H.J.; Park, K.; Zarling, D.; Docherty, K.S.; acids. J. Chromatogr. A 2001, 921, 265–275. Kittelson, D.B.; McMurry, P.H.; Ziemann, P.J. On-line measure- [148] Blokker, P.; Pel, R.; Akoto, L.; Brinkman, U.A.T.; Vreuls, R.J.J. ments of diesel nanoparticle composition and volatility. Atmos. At-line gas chromatographic-mass spectrometric analysis of fatty Environ. 2003, 37(9–10), 1199–1210. acid profiles of green microalgae using a direct thermal desorption [141] Tobias, H.J.; Ziemann, P.J. Kinetics of the gas-phase reactions of interface. J. Chromatogr. A 2002, 959, 191–201. alcohols, aldehydes, carboxylic acids, and water with the C13 stabi- [149] Howard, P.H.; Meylan, W.M. Handbook of Physical Properties of lized criegee intermediate formed from ozonolysis of 1-tetradecane. Organic Chemicals,CRC Press, Boca Raton, FL, 1997; 1–2112. J. Phys. Chem. A 2001, 105, 6129–6135. [150] Sidhu, S.; Graham, J.; Striebich, R. Semi-volatile and particulate [142] Allan, J.D.; Alfarra, M.R.; Bower, K.N.; Williams, P.I.; Gallagher, emissions from the combustion of alternative diesel fuels. Chemo- M.W.; Jimenez, J.L.; McDonald, A.G.; Nemitz, E.; Canagaratna, sphere 2001, 42(5–7), 681–690. M.R.; Jayne, J.T.; Coe, H.; Worsnop, D.R. Quantitative sampling [151] Neususs, C.; Weise, D.; Birmili, W.; Wex, H.; Wiedensohler, A.; using an Aerodyne aerosol mass spectrometer 2. Measurements of Covert, D.S. Size-segregated chemical, gravimetric and number fine particulate chemical composition in two U.K. cities. J. Geo- distribution-derived mass closure of the aerosol in Sagres, Portugal phys. Res. 2003, 108(D3), AAC2-1-AAC2-9. during ACE-2. Tellus Ser. B Chem. Phys. Meteorol. 2000, 52(2), [143] Allan, J.D.; Delia, A.E.; Coe, H.; Bower, K.N.; Alfarra, M.R.; 169–184. Jimenez, J.L.; Middlebrook, A.M.; Drewnick, F.; Onasch, T.B.; [152] Falkovich, A.H.; Schkolnik, G.; Ganor, E.; Rudich, Y. Adsorp- Canagaratna, M.R.; Jayne, J.T.; Worsnopf, D.R. A generalised tion of organic compounds pertinent to urban environments onto method for the extraction of chemically resolved mass spectra from mineral dust particles. J. Geophys. Res. 2004, 109(D2). aerodyne aerosol mass spectrometer data. J. Aerosol Sci. 2004, [153] Graham, B.; Falkovich, A.H.; Rudich, Y.; Maenhaut, W.; Guyon, 35(7), 900–922. P. ; Andreae, M.O. Local and regional contributions to the atmo- [144] Jimenez, J.L.; Jayne, J.T.; Shi, Q.; Kolb, C.E.; Worsnop, D.R.; spheric aerosol over Tel Aviv, Israel: A case study using elemental, Yourshaw, I.; Seinfeld, J.H.; Flagan, R.C.; Zhang, X.; Smith, ionic and organic tracers. Atmos. Environ. 2004, 38(11), 1593– K.A.; Morris, J.W.; Davidovits, P. Ambient aerosol sampling us- 1604. ing the aerodyne aerosol mass spectrometer. J. Geophy. Res. 2003, [154] Yang, H.; Yu, J.Z.; Ho, S.S.H.; Xu, J.H.; Wu, W.S.; Wan, C.H.; 108(D7), SOS13-1-SOS13-13. Wang, X.D.; Wang, X.R.; Wang, L.S. The chemical composition of [145] Zhang, Q.; Alfarra, M.R.; Worsnop, D.R.; Allan, J.D.; Coe, H.; inorganic and carbonaceous materials in PM2.5 in Nanjing, China. Canagaratna, M.R.; Jimenez, J.L. Deconvolution and quantifica- Atmos. Environ. 2005, 39(20), 3735–3749. Downloaded by [NIH Library] at 06:07 28 August 2017