Structure elucidation of the water-soluble organic carbon fraction in atmospheric aerosols by mass spectrometry

Atmospheric aerosols can act as cloud condensation nuclei (CCN) and scatter or absorb light thus influencing many important processes in the atmosphere. The CCN activity of atmospheric aerosols depends mainly on the water solubility of their chemical components. Besides inorganic salts also polar and water-soluble organic compounds are responsible for the water solubility of atmospheric aerosols. Humic-like substances (HULIS) constitute a major part of the water-soluble organic carbon (WSOC) fraction of atmospheric aerosols and may thus be largely involved in cloud condensation processes. However, hitherto little is known about the chemical composition and structure of HULIS.
Structure elucidation was performed by different mass spectrometric techniques of the WSOC fraction and specially of HULIS from atmospheric aerosols collected in the city of Basel. HULIS were cleaved into smaller substructures by thermochemolysis using tetramethylammonium hydroxide (TMAH). The simultaneously methylated substructures were separated by gas chromatography and detected by mass spectrometry (GC-MS). Furthermore, composition and fragmentation behaviour of HULIS were studied by liquid chromatography coupled to ion trap multiple mass spectrometry (LC-MSn).
Structure elucidation of the WSOC fraction by thermochemolysis GC-MS
Solid phase extraction (SPE) was applied to separate the HULIS from the rest of the WSOC and inorganic compounds. Analysis of HULIS by thermochemolysis GC-MS revealed that aliphatic monocarboxylic acids (C9, C10, C12, C14, C16 and C18), aliphatic dicarboxylic acids (C4-10), mono-, di- and trihydroxylated benzoic acids, as well as benzenedicarboxylic and benzenetricarboxylic acids were the main substructures. However, HULIS could not be completely analysed by thermochemolysis GC-MS. A dark brownish residue persisted after thermochemolysis. Its chemical composition remained unknown, since no further structural information could be retrieved by pyrolysis.
Reference compounds such as fulvic acids, humic acids and lignin were analysed by thermochemolysis GC-MS as well. Their substructures were very similar to those of HULIS, which indicated fossil fuel burning or combustion of lignin containing biomass as possible sources. Moreover, thermochemolytical degradation of model compounds containing ester and ether groups was investigated. Aromatic esters, and esters with aliphatic and aromatic partial structures degraded easily. However, aromatic ethers and ethers with aliphatic and aromatic partial structures were less prone to cleavage. In conclusion, HULIS probably contain aromatic esters, and esters with aliphatic and aromatic partial structures rather than aromatic ethers and ethers with aliphatic and aromatic partial structures.
Principal component analysis was used to compare the fingerprint patterns in the thermochemograms of HULIS taken at different sampling dates. A seasonal variation of the composition was observed. HULIS composition of spring, summer and autumn samples seemed to be similar. However, HULIS from November, December, January and February were completely separated from the rest. Hydroxylated aromatic carboxylic acids were more abundant in the HULIS of these filter samples. Aliphatic dicarboxylic acids seemed to be typical for HULIS in spring, summer and autumn filter samples.
Structure elucidation of the WSOC fraction by LC-MS
WSOC was separated into five fractions using a HPLC column containing a 300 Å pore size reversed stationary phase. Mass spectrometry revealed that HULIS in fractions I-IV consisted of polar compounds of lower mass with a mass distribution between m/z 100-400 and a maximum at m/z 240. HULIS in fraction V were less polar substances of higher mass with a mass distribution between m/z 100-900 and a maximum at m/z 500.
MS2 and MS3 spectra showed that carboxylic and hydroxyl moieties were predominant functional groups of HULIS. Moreover, fragment m/z 97 was detected in most MS2 and MS3 spectra. TOF-MS and deuterium exchange experiments identified m/z 97 as HSO4-. These experiments supported the existence of sulphate covalently bound to
HULIS. Furthermore, a loss of -80 u (SO3) was observed in the fragment spectra. However, it could not be clearly associated to sulphonated or sulphated HULIS as both sulphonation and sulphatation of HULIS are possible in the atmosphere.
Quantification of sulphate covalently bound to HULIS was performed by source fragmentation of HULIS detecting HSO4- as m/z 97. Combined concentrations in fractions I-V were similar to other polar organic compounds common in atmospheric aerosols. A slight seasonal trend was observed with higher concentrations in winter and summer than in spring and autumn. However, the trend might be within the normal fluctuation of the concentrations. Concentrations of inorganic HSO4- did not correlate with those of organosulphates, which indicated that sulphatation reaction of HULIS does not only depend on the amount of sulphate in the atmospheric aerosol, but on other factors such as temperature, solar irradiation, acidity of other chemical components present in aerosols.
Combined structural information obtained by thermochemolysis GC-MS and LC-MS allowed to propose defined structures for lower mass HULIS based on 2-3 substructures as well as for higher mass HULIS with 3-4 substructures. In addition, sulphate is covalently bound. The mass distributions of the postulated substructures were in good agreement with the recorded full scan spectra of HULIS.