The importance of both charge exchange and proton transfer in the analysis of polycyclic aromatic compounds using atmospheric pressure chemical ionization mass spectrometry

Beata M. Kolakowski 0 1 2 J. Stuart Grossert 0 1 Louis Ramaley 0 1 0 Published online January 15, 2004 Address reprint requests to Dr. J. S. Grossert, Department of Chemistry, Dalhousie University , Halifax, Nova Scotia B3H 4J3, Canada 1 Department of Chemistry, Dalhousie University , Halifax, Nova Scotia, Canada 2 Current address: Ionalytics Corp., M-50 IPF, 1200 Montreal Road, ON K1A 0R6, Canada The response of atmospheric pressure chemical ionization (APCI) mass spectrometry to selected polycyclic aromatic compounds (PACs) was examined in a Micromass Quattro atmospheric pressure ion source as a function of both solvents and source gases. Typical PACs found in petroleum samples were represented by mixtures of naphthalene, fluorene, phenanthrene, pyrene, fluoranthene, chrysene, triphenylene, perylene, carbazole, dibenzothiophene, and 9-phenanthrol. A large range of different gases in the APCI source was studied, with emphasis on nitrogen, air, and carbon dioxide. Solvents used included water-acetonitrile, acetonitrile, dichloromethane, and hexanes. The signal responses were dependent on both the gases and solvents used, as was the ionization mechanism, as indicated by the degree of protonation with respect to the level of charge exchange. The combination of carbon dioxide in the nebulizer gas stream with nitrogen in the other streams gave a three- to fourfold better sensitivity than using nitrogen alone for both test mixtures and for complex samples. (J Am Soc Mass Spectrom 2004, 15, 301-310) 2004 American Society for Mass Spectrometry - Pin a wide variety of sources, both in endogenous olycyclic aromatic compounds (PACs) are found samples and in those arising from human activities. The PACs occurring in crude oil samples are known to compromise the activity of catalysts in oil refining operations. They affect both combustion processes in gasoline and diesel motors, as well as the nature of combustion byproducts. In addition, there are firm indications that many PACs have effects on the health of living organisms. Hence, the determination of PACs has been of great interest for many years, but complete, detailed analyses have been hampered by complex matrices, trace level concentrations, a large number of isomers, many similar compounds and a lack of reference standards. Special methods of analysis beyond classical procedures include multidimensional chromatographic separations with mass spectrometric detection, high resolution mass spectrometry and atmospheric pressure mass spectrometric methods [1 4]. Atmospheric pressure chemical ionization mass spectrometry (APCI-MS) is an important technique when interfacing mass spectrometry to high-performance liquid chromatography (HPLC) for the analysis of PACcontaining samples. The major drawback of APCI-MS in this regard for the nonpolar polycyclic aromatic hydrocarbon (PAH) portion of a PAC fraction is that the sensitivity of this ionization technique is often poor under conditions which are needed to accomplish satisfactory chromatographic separations [5]. In the preceding paper [6] we presented results from a study in which we examined the production of positive ions in a Micromass Quattro APCI ion source with a pepper pot counter electrode, using different source gases and solvents. These studies have led to some understanding of the conditions present in this type of source. As a result of the rich chemistry occurring in the source, both radical cations and protonated molecules were observed, with equilibrium conditions prevailing; the ions which were detected are those which are thermodynamically most stable. The case of PAHs is especially interesting since molecular ions can be expected to be produced by both direct ionization and charge exchange, whereas protonated molecules are produced by proton transfers, but for these to be effective, strong Brnsted acids must be present [7]. In the present paper we describe how the use of different source gases in combination with various solvents can influence the detection of PACs. These effects were most readily studied by using a mix of PACs designed to simulate those found in a typical lighter fraction refined from a crude oil mixture. The studies were undertaken to understand better the processes involved in APCI and to seek improvements in the analysis of PAC-containing samples. Materials and Solutions Naphthalene (Fisher, Nepean, Ontario, Canada), phenanthrene (Anachemia, Montreal, Canada), fluorene, chrysene, pyrene, fluoranthene, triphenylene, perylene, carbazole, dibenzothiophene and 9-phenanthrol (Aldrich, Oakville, Ontario, Canada) were used as received. A light gas oil sample was supplied by Syncrude Canada Ltd. Solvents (Edmonton, Canada) (water-acetonitrile [50:50 vol/vol], acetonitrile [MeCN], dichloromethane [DCM], and hexanes) and the compressed gases nitrogen, air, carbon dioxide, carbon monoxide, oxygen, methane, and hydrogen are described in the preceding paper [6]. Nitrogen obtained from a pressurized cylinder and nitrogen obtained by evaporation from a pressurized Dewar flask are referred to as T-nitrogen (or TN2) and D-nitrogen (or DN2), respectively. Air from both the building supply and from commercially supplied cylinders was studied, but, as described earlier [6], the air in the building supply was contaminated to such an extent that it did not provide satisfactory results. Thus all results reported below pertain to air obtained in commercial high-pressure cylinders. Standard Mixtures of PACs Structures of the PACs used in this study are shown in Scheme 1 with other details being given in Table 1. The PACs were divided into two groups to avoid overlaps in the m/z values. Group 1 consisted of naphthalene (1, Naph), fluorene (2, Fluo), phenanthrene (3, Phen), pyrene (4, Pyr), chrysene (6, Chry), dibenzothiophene (10, DBT), and 9-phenanthrol (11, 9-Phtr), and Group 2 consisted of 1, 3, fluoranthene (5, Flrn), triphenylene (7, Tri), perylene (8, Pery), and carbazole (9, Carb). Both groups contained 1 and 3 to provide a check on reproducibility of the experiments. Stock solutions were made up in acetonitrile, dichloromethane and hexanes except that 8 was too insoluble in hexanes for it to be included in this solvent. Each stock solution was diluted one hundredfold with the same solvent to yield analytical solutions with a concentration for each PAC of 10 M. The analytical solution with 50:50 (vol/vol) acetonitrile-water was made up by diluting the stock solution in acetonitrile with 50:50 (vol/vol) acetonitrilewater. Instrumentation and Procedures The mass spectrometric instrumentation, gas distribution system and liquid handling equipment were detailed in the preceding paper [6]. Regular mass spectrometric operating parameters (unless otherwise noted) were: corona voltage 4.0 kV (4.5 kV for DCM), optimized cone voltage 30 V, source temperature 120 C (100 C for hexanes), probe temperature 350 C; bath and sheath gas flow rates 300 standard liters (...truncated)