Article pubs.acs.org/ac
Online Characterization of Isomeric/Isobaric Components in the Gas Phase of Mainstream Cigarette Smoke by Tunable Synchrotron Radiation Vacuum Ultraviolet Photoionization Time-of-Flight Mass Spectrometry and Photoionization Efficiency Curve Simulation Yang Pan,*,† Yonghua Hu,‡ Jian Wang,† Lili Ye,† Chengyuan Liu,† and Zhixiang Zhu† †
National Synchrotron Radiation Laboratory, University of Science and Technology of China, Hefei, Anhui 230029, P. R. China Center of Technology, China Tobacco Anhui Industrial Co, Ltd., Hefei, Anhui 230088, P. R. China
‡
S Supporting Information *
ABSTRACT: A newly developed, qualitative and quantitative method based on tunable synchrotron radiation vacuum ultraviolet photoionization time-of-flight mass spectrometry (SR-VUV-PI-TOFMS) and photoionization efficiency (PIE) curve simulation was applied for the online analysis of isomers and isobaric compounds in the gas phase of mainstream cigarette smoke. After blocking the particulate phase components by the Cambridge filter pad, a puff of fresh gas-phase cigarette smoke was immediately introduced into a vacuum ionization chamber through a heated capillary, then was photoionized, and analyzed by a TOF mass spectrometer. The PIE curves for the mass peaks up to m/z = 106 were measured between 8.0 and 10.7 eV. Some components could be directly identified by their discriminated ionization energies (IEs) on the PIE curve. By simulating the PIE curve with the sum of scaled absolute photoionization cross sections (PICSs), complex isomeric/isobaric compounds along with their mole fractions could be obtained when the best-fitting was realized between experimental and simulated PIE curves. A series of reported toxic compounds for quantification, such as 1,3butadiene (m/z = 54), 1,3-cyclopentadiene (m/z = 66), benzene (m/z = 78), xylene (m/z = 106), 2-propenal (m/z = 56), acetone and propanal (m/z = 58), crotonaldehyde (m/z = 70), furan and isoprene (m/z = 68), were all found to have other isomers and/or isobaric compounds with considerable abundances. Some isomers have never been reported previously in cigarette smoke, like C5H6 isomers 1-penten-3-yne, 3-penten-1-yne, and 1-penten-4-yne at m/z = 66. Isomeric/isobaric compounds characterization for the mass peaks and mole fraction calculations were discussed in detail below 10.7 eV, an energy value covering several conventional used VUV light sources.
C
species during smoking cannot be detected, and the real dynamics information of fresh cigarette smoke components cannot be revealed. Recently, photoionization mass spectrometry (PI-MS) has been applied for the online analysis of the chemical components and their highly dynamic processes during pyrolysis and combustion of biomass.2,3,5−7,11−13 Compared to traditional “hard” electron ionization (EI) methods for the gaseous components analysis, photoionization produces little or no fragments, making the identification and interpretation of complex cigarette smoke in real-time possible. By using vacuum ultraviolet (VUV) single photoionization (SPI) and resonanceenhanced multiphoton ionization (REMPI) in combination with online time-of-flight (TOF) mass spectrometer, Zimmermann and co-workers have qualitatively and quantitatively analyzed a large variety of aliphatic and aromatic substances and
igarette smoke is a highly complex and dynamic chemical mixture containing neutral constituents, free radicals, and certain ions, as a result of pyrolysis and combustion reactions of tobacco products.1−9 To standardize the analyses, cigarette smoke is generally divided into the gas phase and the particulate phase. Cambridge filter pads are usually used to trap the particulate particles larger than 0.1 μm, and gas-phase components pass through the pads.10 Up to now, more than 4,800 chemical species have been characterized in cigarette smoke, some of which are highly toxic and/or carcinogenic to human body. Routinely used qualitative and quantitative methods for the smoke components are mostly chromatography or chromatography/mass spectrometry techniques, where trapping, derivatization, and chromatographic separation of smoke collected from cigarette puffs are essentially needed.1−3,6,10 These widely used methods are well-established and reliable but are off-line techniques. Some smoke composition will be altered during tedious pretreatment. Moreover, owing to the complexity and instability character of the smoke, a lot of nascent products including short-lived © 2013 American Chemical Society
Received: September 16, 2013 Accepted: November 19, 2013 Published: November 19, 2013 11993
dx.doi.org/10.1021/ac402955k | Anal. Chem. 2013, 85, 11993−12001
Analytical Chemistry
Article
was measured to be around 1 × 1013 photons s−1, and the energy resolving power (E/ΔE) was about 1500.20 The schematic of the experimental setup is shown in Figure 1. A
have studied their puff-by-puff behaviors in cigarette smoke.2,3,5−7 This research showed great superiority over initial IR-based online methods.14 However, the online PI-MS method still has two problems that affect the qualitative and quantitative accuracy. On one hand, isomers cannot be discriminated from mass spectrum. Even some isobaric species are not distinguishable by low resolution mass spectrometer. On the other hand, the overlap of unresolved isomeric/isobaric species will make the quantification problematic. For instance, the ion signal of m/z = 58 in cigarette smoke produced by 118 nm laser (10.49 eV) had to be attributed to that of acetone, while the contributions from propanal and even ethanedial with higher IE of 10.2 eV were excluded.3 To solve this problem, a fast gas chromatography (GC) was introduced before TOFMS analysis to perform a two-dimensional characterization of cigarette smoke, and some isomers like acetone and propanal were successfully discriminated.5 However, due to the limited separation ability of the online fast GC, many isomers still cannot be distinguished. As we know at least four major species at m/z = 68 in cigarette smoke could be ionized by the electron-beam pumped rare gas excimer light source (EBEL), while only furan and isoprene have been successfully separated and quantified.5 However, the different cigarette types and measurement time scales used in the studies limit the direct comparability. Tunable synchrotron radiation vacuum ultraviolet photoionization time-of-flight mass spectrometry (SR-VUV-PITOFMS) is a powerful approach for products diagnosis in combustion and pyrolysis processes.15−18 In comparison with conventional VUV light sources, synchrotron radiation is more advanced with high tunability. Currently, we developed a SRVUV-PI-TOFMS-based technique for the characterization of species in mainstream cigarette smoke.19 In this paper, we will characterize the isobaric/isomeric compounds in the gaseous mainstream smoke by two methods: 1) “Pre”-qualification: Characterization by comparing the reported ionization energy (IE) values with the IEs rationalized by onsets on the photoionization efficiency (PIE) curves. Some unknown IEs of possible components were obtained by theoretical calculation. 2) PIE curve simulation: Qualification and quantification of more complex isomeric/isobaric species via PIE curve simulation. The mole fractions of major isomeric/ isobaric species in each mass peak were determined below 10.6 eV, an energy covering three conventional VUV light sources, i.e., Krypton lamp (10.6 eV), 118 nm VUV laser (10.49 eV) via a 2-fold frequency tripling of fundamental Nd:YAG laser pulses, and EBEL (9.8 ± 0.04 eV).
Figure 1. Schematic of the smoking source and photoionization mass spectrometric apparatus.
modified glass syringe was used as the smoking source. Two glass tubes (6 mm O.D., 3 mm I.D.) were set at the same side of the glass syringe. One tube was connected to a cigarette holder (with Cambridge filter pad), and another tube was connected to the ionization chamber via a heated fused silica capillary (I.D. 100 μm, 40 cm in length, 200 °C). Cigarettes were lit with a Borgwaldt electric lighter. After pulling the pistol of the syringe and closing the valve beside the holder, a puff of gas-phase cigarette smoke (∼35 mL puff volume, ∼2 s puff duration) in the syringe would be introduced into the ionization chamber (∼1 × 10−4 Pa) by virtue of the vacuum difference. The flow rate of the smoke was estimated to be around 1 mL min−1. The end of the capillary was put in the central area of the ion extraction field of the TOF mass spectrometer. Cigarette smoke beam out from the capillary was perpendicular crossed and ionized by the SR-VUV light. The generated ions were extracted and accelerated into the field-free region and then reflected by a reflector. The ions were detected by two multichannel plates and amplified by a preamplifier (VT120C, ORTEC, USA). An 1 GHz multiscaler (FAST Comtec P7888, Germany) was used to record the TOF mass spectra.21 The PIE curves were obtained by keeping the step size of 4 Å/step and the accumulation time of 12 s for each mass spectrum. For comparison, quantitative analysis for acetone and propanal were also fulfilled with the LC-MS-based standard method.22−24 Total yields of acetone and propanal in a cigarette were measured to be 298.7 and 68.2 ± 5 μg/cigarette. Theoretical Calculations. All the quantum chemistry calculations were carried out using the Gaussian 09 program.25 The experimental IEs of some cigarette smoke components are not available. To fulfill the characterization the IEs of Nmethylacetamide (m/z = 73), 2-propenethiol (m/z = 74), 2methylfuran (m/z = 82), 4-methyl-1-cyclopentene (m/z = 82), 3-hexyne (m/z = 82), 2-methyl-1,3-pentadiene (m/z = 82), and glycerol (m/z = 92) were calculated at the CBS-QB3 level of theory.26 The CBS-QB3 method is a compound chemistry model, providing satisfactory accuracy for relatively larger molecules at a moderate computational cost. Specifically, the
■
EXPERIMENTAL AND THEORETICAL METHODS Materials. A kind of Virginia type cigarettes was purchased from domestic markets of China. These cigarettes are 84 mm in length, and release tar, nicotine, and carbon monoxide yields of 13, 1.3, and 13 mg, respectively, under ISO smoking machine conditions. Before the experiments, the cigarettes and Cambridge filter pads were stored for one week under controlled conditions of 60% relative air humidity and 22 °C. Experimental Procedure. All the experiments were carried out on the mass spectrometric analysis endstation of the National Synchrotron Radiation Laboratory, China. The synchrotron radiation and TOF mass spectrometer setup has been reported elsewhere, so only a brief description is given here.19,20 The beamline used for photoionization covers the photon energy from 7.8 to 24.0 eV. The average photon flux 11994
dx.doi.org/10.1021/ac402955k | Anal. Chem. 2013, 85, 11993−12001
Analytical Chemistry
Article
PIE Curve Simulation. Accurate quantification is based on the reliable qualification. For the complex cigarette smoke, it means that isomers and isobaric compounds for a designated mass peak should be known prior to quantification, and knowledge of their mole fractions is essential. “Pre”qualification through PIE analysis could offer us an easy way for the characterization of some mass peaks with one or more species. However, the numbers of isomers and isobaric compounds increase dramatically with the increase of molecular weight. Overlap of the cross sections of isomers, approximate IEs of each component, slow ascendant trend, and signal fluctuation make the isomers discrimination really difficult only by searching onsets from the PIE curves. Cool et al. reported that the PIE curve of isomers could be simulated by the sum of their absolute photoionization cross sections (PICSs).27 In doing so, possible components with known PICS should be chosen first. Then, by adjusting the arbitrary scale for the signals as required to match the scaled PICS of each isomer according to eqs 1 and 2, composition characterization will be realized and mole fractions of the isomers can be obtained. σ = xaσa + xbσb + ... + xnσn (1)
geometry optimization and frequency calculations were performed at the B3LYP/CBSB7 basis set, followed by a series of single-point energy corrections including CCSD(T), MP4, MP2, and CBS extrapolation calculations. Energies of neutral molecules and corresponding cations were computed, respectively, and then the IEs were determined using the formula IE = E[cation] − E[neutral].
■
RESULTS AND DISCUSSION Photoionization Mass Spectra and Photoionization Efficiency Curves. Figure 2 shows the mass spectra of the
xa + xb + ... + xn = 1
(2)
Here the PICS σ of a mass peak is expressed as a combination of cross sections σa to σn for isomers a, b, ..., n weighted by their respective mole fractions xa, xb, ..., xn.27 That means complicated components in cigarette smoke can be characterized by reconstructing the experimental PIE curve with individual component’s PICS. Figure 3 shows the result of an experiment for the m/z = 40 ions in this work conducted to test our procedure for
Figure 2. Photoionization mass spectra of the gas-phase components in the mainstream cigarette smoke at different photon energies: (a) 9.0 eV; (b) 10.0 eV; (c) 11.0 eV; (d) 12.1 eV.
gaseous components in the mainstream cigarette smoke at four photon energies. Only a few mass peaks with relatively low abundances were found at 9.0 eV, which were attributed to some unsaturated hydrocarbons, oxygenated and nitrogenous species. More compounds with higher IEs were gradually ionized and observed with the increase of photon energies. Due to the “soft” near-threshold photoionization character, nearly all the mass peaks could be attributed to parent ions with little or no fragments. As reported previously, the PIE curve of each mass peak can be obtained by plotting the ion intensity versus corresponding photon energy.19 Generally, the ion signal intensity of a mass peak will rise greatly from the baseline when ionization occurs. A distinguishable onset on the PIE curve could be found, corresponding to the IE of a species. Some cigarette smoke components, such as acetylene at m/z = 26, can be easily “pre”qualified through the clear signal intensity jump. Sometimes, even isomers like allene and propyne could be discriminated from the PIE curve.20 However, this “pre”-qualification method is only applicable for limited mass peaks with relative lower molecular weights.
Figure 3. The PIE curve of m/z = 40 (open circle), scaled PICSs of propyne and allene, and the sum of the scaled PICSs.
determinations of isomers from PIE measurements. To eliminate the influences of flux variations with photon energy, the ion signals have been normalized by the monitored photon flux. In addition, considering many species undergo changes in concentration levels with the lapse of time, the PIE curve has also been corrected by the decay of signals. As can be seen in Figure 3, two thresholds near 9.69 and 10.36 eV were clearly observed on the PIE curve, corresponding to the IEs of allene and propyne. Thus, the PIE curve was tentatively simulated by scaling the absolute PICS of allene (σallene)28 and propyne (σpropyne),27 as depicted in 11995
dx.doi.org/10.1021/ac402955k | Anal. Chem. 2013, 85, 11993−12001
Analytical Chemistry
Article
Table 1. Major Gaseous Compounds in Mainstream Cigarette Smoke Characterized in This Work, Their IEs, and Mole Fractions m/z
formula
species
IE (lit.a)
IE (this work)
mole fraction (±5−10%)
28 30 34 40
C2H4 NO H2S C3H4 C3H4 C3H6 C2H4O C2H6O C2H6O CH4S C4H4 C4H6 C4H6 C4H6 C4H6 C3H4O C4H8 C4H8 C4H8 C3H6O C3H6O C2H2O2 C2H5NO C2H5NO C5H6 C5H6 C5H6 C5H6 C4H4O C5H8 C5H8 C5H8 C5H10 C5H10 C4H6O C4H6O C5H10 C5H10 C4H8O C4H8O C4H8O C3H7NO C3H6O2 C3H6S CS2 C6H6 C6H6 C5H5N C6H8 C6H8 C4H4N2 C5H6O C6H10 C6H10 C6H10 C5H8O C6H12 C5H10O C5H10O C7H8
ethylene nitric oxide hydrogen sulfide propyne allene propylene acetaldehyde ethanol dimethyl ether methyl mercaptan butenyne 1,3-butadiene 1-butyne 1,2-butadiene 2-butyne 2-propenal (acrolein) 2-butene isobutene 1-butene acetone propanal ethanedial acetamide N-methylformamide 1,3-cyclopentadiene 1-penten-3-yne 3-penten-1-yne 1-penten-4-yne furan isoprene cyclopentene 3-methyl-1,2-butadiene 2-methyl-2-butene 2-pentene methyl vinyl ketone crotonaldehyde 2-methyl-1-butene 3-methyl-1-butene tetrahydrofuran 2-butanone 2-methylpropanal N-methylacetamide methyl acetate 2-propenethiol carbon disulfide fulvene benzene pyridine 1,3-cyclohexadiene 1,4-cyclohexadiene pyrazine 2-methylfuran 3-methylfuran 3-hexyne 2-methyl-1,3-pentadiene 3-methyl-2-butenal 2-hexene tetrahydropyran 2-pentanone toluene
10.5 9.26 10.46 10.35 9.69 9.73 10.23 10.48 10.03 9.44 9.58 9.07 10.18 9.23 9.58 10.11 9.1 9.22 9.55 9.7 9.96 10.18 9.69 9.83 8.56 8.97 9.02 9.95 8.88 8.86 9.01 8.9 8.69 9.04 9.65 9.75 9.12 9.52 9.40 9.52 9.71 10.25 10.07 8.36 9.24 9.26 8.25 8.82 9.29 8.38 9.4 ∼8.9 9.25 9.38 8.83
10.5 9.28 10.45 10.31 9.7 9.71 10.2 10.46 10.01 9.41 9.57 9.05 9.24 9.1 9.56 9.69 9.96 9.68 9.82 8.55 ∼8.86 ∼8.86 9.01 8.68 9.12 9.4 9.5 9.7 9.43(9.44b) 10.26 9.02(9.04b) 10.06 8.36 9.23 9.24 8.26 8.83 9.32 8.37 8.59 (8.60b) 9.33 (9.33b) 8.48 (8.49b) 9.4 8.88 9.26 9.39 8.82
100% (
