Download Ionization and Reactions of Hydroperoxides with O2 and F-: A Mass Spectrometry Study and more Papers Chemistry in PDF only on Docsity!
Computational and experimental studies of chemical ionization
mass spectrometric detection techniques for atmospherically
relevant peroxides
Benjamin M. Messer, David E. Stielstra, Christopher D. Cappa, Kurtis W. Scholtens,
Matthew J. Elrod*
a Department of Chemistry, Hope College, Holland, MI 49423, USA Received 13 April 1999; accepted 9 November 1999
Abstract
We report the results of computational and experimental studies concerning the chemical ionization mass spectrometric detection of hydrogen peroxide (HOOH) and methyl hydroperoxide (CH 3 OOH). GAUSSIAN2 (G2) electronic structure calculations are used to predict structures, natural charges of the atoms and energies for the neutral species, as well as for the cation, anion, and the proton and fluoride adduct species. These calculations are used to predict ion–molecule reaction thermodynamics as a guide to the experimental development of chemical ionization mass spectrometric detection methods. Both HOOH and CH 3 OOH are predicted to react exothermically with O 21 and F^2 to yield the cationic and fluoride adduct species, respectively. In addition, CH 3 OOH is predicted to react exothermically with H 3 O^1 to yield the proton adduct species. The feasibility of F^2 chemical ionization mass spectrometric detection of peroxides was experimentally explored through kinetic studies. The fluoride adduct formation reactions for both HOOH and CH 3 OOH were found to proceed at or near collision-limited rates. (Int J Mass Spectrom 197 (2000) 219 –235) © 2000 Elsevier Science B.V.
Keywords: Chemical ionization; Computational; Thermodynamics; Kinetics; Peroxides
1. Introduction
Hydrogen peroxide (HOOH) is the dominant oxi- dant in clouds, fogs, or rain in the atmosphere [1]. It is formed from the reaction of two hydroperoxyl radicals
HO 2 1 HO 2 3 HOOH 1 O 2 (1)
Methyl hydroperoxide (CH 3 OOH) is an important atmospheric species formed by the oxidation of meth- ane in the atmosphere [1]
CH 4 1 OH 3 CH 3 1 H 2 O (2)
CH 3 1 O 2 1 M 3 CH 3 O 2 1 M (3)
CH 3 O 2 1 HO 2 3 CH 3 OOH 1 O 2 (4)
Under high nitrogen oxide (NO x ) conditions, the following reactions lead to ozone production:
CH 3 O 2 1 NO 3 CH 3 O 1 NO 2 (5)
NO 2 1 h n~ l , 380 nm) 3 NO 1 O (6)
- Corresponding author. E-mail: elrod@hope.edu O 1 O 2 1 M 3 O 3 1 M (7) 1387-3806/00/$20.00 © 2000 Elsevier Science B.V. All rights reserved PII S 1 3 8 7 - 3 8 0 6 ( 9 9 ) 0 0 2 6 0 - 2
International Journal of Mass Spectrometry 197 (2000) 219 –
Recently, Wennberg et al. [2] reported the mea- surement of higher HO x levels than predicted in the upper troposphere of the northern hemisphere. Be- cause this result suggests that this region of the atmosphere is not ordinarily dominated by NO x chem- istry, it is therefore more susceptible to anthropogenic NO x emissions than previously thought. This is an important finding since it indicates that increased air traffic in the upper troposphere may lead to a substan- tial increase in ozone levels. It was suggested that uncertainties in the reactions that interconvert HO x and its peroxide reservoirs (such as HOOH and CH 3 OOH) might be leading to the underprediction of HO x levels in the upper troposphere. The study of reactions involving peroxide chemis- try in both the laboratory and the field has been hindered by the analytical detection methods currently available. In laboratory kinetics environments, ultra- violet–visible optical detection methods have tradi- tionally been used, with the difficulty arising from the need to determine temperature-dependent absolute absorption cross sections at wavelengths specific to each peroxide [3]. Because of these problems, there is considerable uncertainty concerning the actual prod- uct distribution of reaction (4), with the possibility of a CH 2 O 1 H2O 1 O 2 channel as the major product pathway. In field detection environments, the most successful analytical method involves the indirect approach of preconcentration of the peroxides, deri- vatization and fluorescence detection. This method is plagued by sampling artifacts, low sensitivity and poor time resolution [4]. Direct gas phase sampling of CH 3 OOH has been tested by using gas chromatogra- phy (GC) separation and traditional electron impact mass spectrometry (EIMS) methods, but extensive fragmentation of the CH 3 OOH^1 ion ruled out the method for direct use in field studies [5]. Therefore, it is clear that the development of a new analytical technique for the measurement of peroxides would aid in both laboratory and field investigations of atmo- spheric peroxide chemistry. EIMS detection methods are widely used in labo- ratory atmospheric chemistry applications because the approach is much more general than the often mole- cule-specific optical techniques. However, EIMS sen-
sitivity and selectivity levels are usually inferior to those of the competing optical techniques. The method of chemical ionization mass spectrometry (CIMS) has found increasing use in atmospheric chemistry applications because of its potential for increased selectivity and sensitivity over traditional EIMS methods (which are largely due to the advan- tages of the ambient ionization conditions typical of CIMS approaches). CIMS has been implemented in laboratory kinetics settings [6,7], as well as in field measurement settings [8 –10]. Although the method is, in principle, totally general, chemical ionization schemes (sufficiently fast ion–molecule reactions) must be developed for each chemical species under study. In addition, the complete sample matrix must be evaluated for potential interference reactions that could hinder the proposed chemical ionization scheme. Although chemical ionization schemes currently exist for the study of many atmospherically relevant species [11], there exists no systematic procedure for the determination of feasible chemical ionization schemes for other species of interest. Currently, chemical ionization schemes are proposed by analogy to similar systems and are empirically tested. We believe that the CIMS method has not been more widely implemented because of this less-than- straightforward developmental aspect of the tech- nique. It is therefore the goal of this article to describe the systematic development of a chemical ionization scheme for use with chemical ionization mass spec- trometry, by using the peroxides HOOH and CH 3 OOH as specific illustrations of the combined computational/experimental approach.
2. Methodology
In this study, we computationally investigate the four ion–molecule reactions listed below in order to establish the various thermodynamic quantities that are required to evaluate potential chemical ionization schemes:
A 1 e^2 3 A^2 (8)
bond orbitals [22]. For radical species, the NBO analysis utilizes separate calculations for both the alpha and beta spin electrons, the characteristics of which are then combined to create an overall elec- tronic structure. These “natural” bond orbitals provide insight into the electronic structural changes which occur during the ionization process, and are used to explain observed differences in structure between the neutral and ionic species. All calculations were per- formed on a Silicon Graphics Indigo2 R4400 work- station.
4. Experimental
4.1. Synthesis of methyl hydroperoxide
Methyl hydroperoxide is not commercially avail- able; however, O’Sullivan et al. have detailed a straightforward synthesis [23] which was modified here to avoid a potentially dangerous purification step. A mixture of water (14 mL), 30% hydrogen peroxide (4.0 mL), and dimethyl sulfate (1.9 mL), under continuous stirring at 0 °C, was treated with drop- wise addition of 40% potassium hydroxide (15 mL). The solution was then slowly heated to 50 °C over a period of 30 min, and the temperature was maintained for another 30 min to drive the reaction to completion. The desired product was collected by slowly bubbling argon gas through the solution, and trapping the escaping vapor in a collection trap containing 25 mL water at 0 °C. In this way, CH 3 OOH was collected and CH 3 OOCH 3 , the other product of the synthesis, was not. The product was positively identified by IR spectroscopy, with characteristic absorbance peaks at 2960 and 1320 cm^21 [24]. The product was stored in aqueous solution at 5 °C.
4.2. Mass spectrometric detection
All mass spectrometric measurements for this study were conducted utilizing the chemical ioniza- tion mass spectrometer apparatus depicted in Fig. 1. A negative ion chemical ionization scheme (F^2 ) was used to detect HOOH and CH 3 OOH. F^2 was pro-
duced in a polonium-210 alpha emitting ionization source by passing a large N 2 flow (10 STP L min^21 ) and 1.0 STP mL min^21 of NF 3 through the ionization source. The commercial ionization source consisted of a hollow cylindrical (69 mm length by 12.7 mm diam) aluminum body coated with 10 mCi of poloni- um-210 on the interior walls. Ions were detected with a quadrupole mass spectrometer housed in a two-stage differentially pumped vacuum chamber. Flow tube gases (neutrals and ions) were drawn into the front chamber through a 0.1 mm aperture, which was held at a potential of 2 210 V. The ions were focused by three lenses constructed from 3.8 cm i.d., 4.8 cm o.d aluminum gaskets. The front chamber was pumped by a 6 in. 2400 L s^21 diffusion pump. The gases entered the rear chamber through a skimmer cone with a 1. mm orifice (held at 2 130 V) that was placed approx- imately 5 cm from the front aperture. The rear chamber was pumped by a 250 L s^21 turbomolecular pump. Once the ions passed through the skimmer cone, they were mass filtered and detected with a quadrupole mass spectrometer.
4.3. Sample standardization, introduction, and mass spectral acquisition
The peroxide samples were standardized by chem- ical titration using the I 32 method described by Klas- sen et al. [25]. The HOOH solutions used were prepared by serially diluting 1 mL of 30% HOOH with distilled water to a final 0.1% HOOH solution.
Fig. 1. Experimental apparatus.
About 5.0 STP mL min^21 N2 was passed through a bubbler containing ;5 mL of the HOOH solution. This gas phase mixture was then injected into the apparatus and added to a large flow to N 2. The gas phase concentration of HOOH was calculated using the Henry’s law coefficient for the 0.1% HOOH solution (as calculated from O’Sullivan et al. [23]), the fractional flow rate of N 2 through the bubbler and the total pressure. Concentrations of about 1 3 1012 molecule cm^23 were used to collect mass spectra. The CH 3 OOH solution was used as prepared and intro- duced to the system (and concentrations were calcu- lated) in a manner analogous to HOOH. It should be noted that water vapor (; 1 3 1014 molecule cm^23 ) was also introduced to the system through the use of the bubbler method.
4.4. Ion–molecule kinetics
In order to estimate the rate constant for the reaction of F^2 with HOOH and CH 3 OOH, fixed distance ion–molecule kinetics studies were per- formed in a manner similar to previous work by Huey et al. [11]. All measurements were performed at 298 K and 100 Torr, and turbulent flow conditions were maintained. Previous work by Seeley et al. [26] has demonstrated that the turbulent flow experimental conditions of the experiments performed here result in fast mixing of reactant gases such that homogeneous mixing is complete within a few cm of the reactant introduction port. The measurements were performed using pseudo first order conditions ([neutral] .. [ion]) and the product ion signal was monitored as a function of neutral concentration. The apparatus was tested on the SF 62 1 NO 2 reaction, and it was found that the measurements resulted in a systematically lower rate constant (by a factor of about 2) than was previously reported in the literature [11]. As discussed by Adams et al. [27], systematic errors in the mea- surement of ion–molecule rate constants can occur when the plasma velocity is not matched to the neutral gas velocity. Therefore, in order to eliminate system- atic errors in the determination of bimolecular rate constants in our apparatus, relative rate measurements were conducted. The well-studied reaction [28] of
F^2 1 Cl 2 was used as the reference for the HOOH and CH 3 OOH 1 F^2 kinetics studies. For most exper- iments, the mean gas velocity was held constant at around 960 cm s^21 over a reaction distance of 7. cm, yielding a reaction time of 7.6 ms.
5. Results and discussion
5.1. Validation calculations
In order to assess the accuracy of the computa- tional ion thermodynamics method used in this study, several benchmark calculations were performed at various levels of theory for O 2 , NO 2 , H 2 O, and HCl. These calculations were chosen because of the exis- tence of accurate experimental ion thermodynamic data for each of the atmospheric species and because each neutral is experimentally known to affect one of the processes (8)–(11). The results of these calcula- tions are compiled in Table 1. It may be seen that all calculated ion thermodynamic values of interest are not in good agreement with the experimental values for calculations at levels of theory lower than the MP4/6-311 1 G( d , p ) level. Furthermore, the relatively high quantitative accuracy desired in this study (0. eV) is not consistently achieved for calculations at a lower level of theory than that of the G2 compound method. For the anions, it is clear that the use of diffuse orbitals (indicated by the “ 1 ” notation in the basis set description) is more important than the use of high level electron correlation, which may be noted by comparing the much better quantitative accuracy of the calculations at the MP2/6-311 1 G( d , p ) level compared to those at the MP4/6-311G( d , p ) level. The exothermicity of the majority of ion–molecule reac- tion schemes is on the order of 1 eV, indicating that the 0.1 eV accuracy achieved by the G2 method is sufficient for evaluating most proposed reactions. As an additional verification of the computation- ally estimated thermodynamic properties used, the standard enthalpy of formation of CH 3 OOH was calculated through the use of the following isodesmic (“equal bond”) reaction, in which the number and
description of the bonding in anion species. In the NBO analysis of the HOOH^2 species, the alpha electron interpretation is best described as a complex of an RO radical with an OH^2 anion, while the beta electron interpretation indicates a bonding character similar to the neutral species (i.e. with an intact O–O bond). The resulting structure then represents a com- promise between the remaining covalent bonding
character in the O–O bond and the orientation depen- dent electrostatic forces between the RO and OH^2 subunits. Indeed, the weakened covalent bonding in the HOOH anion (as described by the beta electron interpretation) is manifested in the very large O–O bond length (2.25 Å) calculated for HOOH^2. In addition, the charge– dipole forces (as described by the alpha electron interpretation) favor hydrogen
Table 2 Ionic standard thermodynamic properties (eV) for peroxides and peroxide-related species HF/6-31G( d ) MP2/6-31G( d ) MP4/6-311G( d , p ) a^ MP2/6-311 1 G( d , p ) a^ MP4/6-311 1 G( d , p ) a^ G2a^ Expt. E.A. (HOOH)
2 0.88 2 0.85 2 0.55 0.43 0.50 0.
I.P. (HOOH)
9.46 9.79 9.96 10.19 10.37 10.70 10.58c
F.A. (HOOH)
2.38 2.88 2.51 1.62 1.61 1.
P.A. (HOOH)
6.97 6.86 6.76 6.84 6.92 6.85 6.99d
P.A. (HOO) 6.37 7.09 7.20 7.12 7.04 6.65 6.84d E.A. (CH 3 OOH)b
2 0.88 2 0.43 2 0.87 2 0.84 2 0.
I.P. (CH 3 OOH)
8.53 6.29 9.32 9.59 9.51 9.
F.A. (CH 3 OOH)
2.33 2.78 2.54 1.57 1.56 1.
P.A. (CH 3 OOH)
7.38 7.22 7.48 7.20 7.31 7.
P.A. (CH 3 OO)
7.31 7.59 7.82 7.68 7.68 7.
a (^) Geometry optimized at the MP2/6-31G( d ) level. b (^) Geometry optimized at the MP2/6-31 11 G( d ) level and all energies calculated with doubly diffuse basis sets for CH 3 OOH (^2). c (^) See [35]. d (^) See [36].
Table 3 MP2/6-31G( d ) geometric parameters of HOOH and relevant ions HOOH HOOH^2 HOOH^1 HOOH 21 HOOHF^2 Bond length (Å) H 1 –O 2 0.976 0.976 1.007 0.999 1. O 2 –O 3 1.468 2.247 1.351 1.463 1. O 3 –H 4 0.976 0.976 1.007 0.991 1. O 2 –H 5 or H 1 –F5 0.999 1. Bond angles (deg) H 1 –O 2 –O3 98.7 56.8 102.3 103.5 92. O 2 –O 3 –H4 98.7 56.8 102.3 99.7 92. H 1 –O 2 –H5 or O 2 –H1–F5 109.4 146. Dihedral angles (deg) H 1 –O 2 –O3–H4 121.2 227.1 180.0 122.9 0. H 5 –O 2 –O3–H4 or O 3 –O2–H1–F5 237.1 0.
bonding between the neutral OH radical and OH^2 and are manifested by the unusually acute /HOO bond angles (56.8°) calculated for HOOH^2. This interpre- tation is in agreement with previous calculations by Hrusak et al., who characterize HOOH^2 as a stable intermediate on the reaction pathway between HOOH 1 e^2 and the more thermodynamically stable O^2 zH 2 O [32]. Such significant structural changes for CH 3 OOH^2 are not observed. Apparently, the addi- tional steric constraints imposed by the methyl sub- stituent prevent the alpha electron interaction from playing a significant role in determining the equilib- rium structure of CH 3 OOH^2.
5.4. Electron affinities
The electron affinity of HOOH was determined to be 0.82 eV at the G2 level. Using the previous lower level calculations of Hrusak et al. for various HOOH^2 species [32] and the experimental bond dissociation energy for HOOH of 50.5 kcal/mol [33], an electron affinity value of 0.88 eV is obtained for HOOH, which is in good agreement with the results of this study. The electron affinity of CH 3 OOH was calcu-
Table 5 MP4/6-311 1 G( d , p ) natural charges of HOOH and relevant ions Atom HOOH HOOH^2 HOOH^1 HOOH 21 HOOHF^2 H 1 0.495 0.434 0.581 0.588 0. O 2 2 0.495 2 0.934 2 0.081 2 0.437 2 0. O 3 2 0.495 2 0.934 2 0.081 2 0.272 2 0. H 4 0.495 0.434 0.581 0.533 0. H 5 or F 5 0.588 2 0.
Table 6 MP4/6-311 1 G( d , p ) natural charges of CH 3 OOH and relevant ions Atom CH 3 OOH CH 3 OOH^2 CH 3 OOH^1 CH 3 OOH 21 CH 3 OOHF^2 C1 2 0.098 2 0.231 2 0.283 2 0.126 2 0. H 2 0.153 0.125 0.272 0.220 0. H 3 0.136 0.113 0.274 0.191 0. H 4 0.146 0.131 0.273 0.191 0. O 5 2 0.324 2 0.651 0.140 2 0.160 2 0. O 6 2 0.471 2 0.820 2 0.255 2 0.472 2 0. H 7 0.458 0.334 0.579 0.578 0. H 8 or F (^8)
0.578 2 0.
Table 4 MP2/6-31G( d ) geometric parameters of CH 3 OOH and relevant ions CH 3 OOH CH 3 OOH^2 CH3OOH^1 CH 3 OOH2^1 CH3OOHF^2 Bond length (Å) C1–H2 1.092 1.097 1.085 1.094 1. C 1 –H3 1.094 1.095 1.089 1.089 1. C 1 –H4 1.093 1.095 1.088 1.089 1. C 1 –O5 1.420 1.428 1.469 1.445 1. O 5 –O6 1.470 1.470 1.272 1.492 1. O 6 –H7 0.977 0.990 0.990 0.998 1. O 6 –H8 or H 7 –F 8 0.998 1. Bond angles (deg) H 2 –C 1 –H 3 110.3 109.5 112.7 110.8 110. H 2 –C 1 –H 4 110.1 110.7 112.8 110.8 109. H 2 –C 1 –O 5 104.2 104.3 102.3 99.7 106. C 1 –O5–O 6 104.5 106.1 111.1 106.6 104. O 5 –O6–H7 98.4 100.6 105.7 102.7 99. O 5 –O6–H8 or O 6 –H7–F8 102.7 173. Dihedral angles (deg) O 6 –O5–C1–H2 177.4 177.0 177.8 180.0 183. O 6 –O5–C1–H3 296.3 294.9 243.3 296.7 303. O 6 –O5–C1–H4 59.0 57.8 58.7 63.2 64. H 7 –O6–O5–C1 124.0 82.7 179.8 123.7 65. H 8 –O6–O5–C1 or F 8 –H7–O 6 –O5 236.3 299.
[35] indicates that the G2 results are quite accurate for HOOH^1. The G2 I.P. for CH 3 OOH was determined to be 9.86 eV. Additionally, the G2 enthalpy of HOOH^1 and CH 3 OOH^1 can be compared to those for HO 2 and CH 3 O2 to yield a calculated proton affinity of HO 2 and CH 3 O 2 , respectively. The calculated P.A. of HO 2 was found to be 6.65 eV, which is also in good agreement with the experimental value of 6.84 eV [36]. The calculated P.A. of CH 3 O2 was found to be 7.42 eV.
5.7. HOOH 2^1 and CH 3 OOH 2^1 : structures and natural charges
The results indicate that the addition of a proton to each species is predicted to occur at the terminal oxygen atom, leading to two chemically identical hydrogen atoms from both the structural and natural charge point of view. Except for a lessened negative natural charge on the terminal oxygen atoms in HOOH 21 and CH 3 OOH 21 , these species are very similar to their neutral parent species.
5.8. Proton affinities
The calculated P.A. of HOOH was found to be 6.85 eV at the G2 level, which is in good agreement with the experimental value of 6.99 eV [36]. The calculated P.A. of CH 3 OOH was found to be 7.25 eV at the G2 level.
5.9. F^2 z HOOH and CH 3 OO^2 z HF: structures and natural charges
The most unusual change in geometry of the various ionization products can be seen in the fluoride adducts of HOOH and CH 3 OOH. In each case, the resulting product cyclized to form either a five or six membered ring (see Fig. 2), respectively. This cy- clization significantly adds to the stability of the fluoride adduct, as calculations which held the perox- ide geometry fixed and optimized only the position of the fluoride anion yielded fluoride affinities 0.2– 0. eV lower than those in which all degrees of freedom are optimized. Because of the chemical symmetry of HOOH, the F^2 zHOOH ion forms a symmetrical
pentagonal structure in which the electron density of the fluoride ion is shared evenly between the two hydrogen atoms (thus the notation F^2 zH 2 O 2 would actually more accurately represent the structure of the adduct). In the CH 3 OOH case, the chemical symmetry is broken, and the resulting fluoride adduct more closely resembles that of an ROO^2 anion complexed with HF. The natural charge analysis and calculated bond lengths for CH 3 OOH are consistent with this interpretation, as the terminal oxygen of the peroxide functional group possesses a stronger negative charge than the central oxygen atom and the H–F bond length is more similar to that of the HF molecule than that in the F^2 zHOOH ion.
5.10. Fluoride affinities
The calculated fluoride affinities of HOOH and CH 3 OOH are 1.65 and 1.59 eV, respectively. In order to further investigate the interesting ring formation observed in the fluoride adducts of HOOH and CH 3 OOH and to determine if fluoride adduct forma- tion might be a general chemical ionization scheme for ROOH species, additional calculations were per- formed on CH 3 CH 2 OOH. In this case, the MP2/6- 31G( d ) level geometry optimization found that a seven membered ring was the minimum energy struc- ture, indicating that ring formation seems to be general trend for the fluoride adducts of small ROOH species. The calculation of G2-level fluoride affinities was not possible with the available computational resources, but a MP4/6-311 1 G( d , p ) level fluoride affinity for CH 3 CH 2 OOH was determined and found to be very close to the value (1.67 eV) calculated with the same level of theory for CH 3 OOH. Therefore, we estimate that the fluoride affinities for CH 3 CH 2 OOH and CH 3 OOH are essentially identical.
5.11. Evaluation of potential CIMS detection schemes
The small electron affinity of HOOH and the negative electron affinity of CH 3 OOH essentially rule out the use of electron transfer as a possible chemical ionization scheme for these species. The E.A. of
HOOH (0.82 eV) is calculated to be lower than that of SF 6 (1.05 eV [37]), thus eliminating SF 62 (which is generally a very versatile reagent ion) as a possible ionizing agent. The calculated E.A. of HOOH is higher than the E.A. of O 2 (0.45 eV [38]), indicating that O 22 could be used as a possible ionizing agent for HOOH. However, we are not aware of any experi- mental observation of the HOOH^2 anion. The ionization potentials of both HOOH (10. eV) and CH 3 OOH (10.41 eV) are less than the experimental ionization potential of oxygen (12. eV) [39], indicating that both species could be de- tected by chemical ionization with O 21 :
HOOH 1 O 21 3 HOOH^1 1 O (^2)
D H 5 2 1.48 eV (15)
CH 3 OOH 1 O 21 3 CH 3 OOH^1 1 O (^2)
D H 5 2 1.65 eV (16)
The calculated proton affinity of HO 2 (6.65 eV) is less than the experimental proton affinity of water (7. eV) [36], thus ruling out the use of the H 3 O^1 reagent ion as a chemical ionization scheme for HO 2. How- ever, the calculated proton affinity of CH 3 O2 (7. eV) is higher than that of water, indicating that CH 3 O could be detected by using chemical ionization with the H 3 O^1 reagent ion:
CH 3 O 2 1 H 3 O^1 3 CH 3 OOH^1 1 H 2 O
D H 5 2 0.26 eV (17)
In fact, the preliminary findings from this work were the basis for the chemical ionization scheme recently employed by our group in our kinetic study of the neutral CH 3 O2 1 NO reaction [40]. As for the RO 2 species, the calculated proton affinity of HOOH (6.85 eV) is less than experimental proton affinity of water (7.16 eV), but the calculated proton affinity for CH 3 OOH (7.25 eV) is greater than that of water, suggesting that CH 3 OOH could be detected using chemical ionization with the H 3 O^1 reagent ion:
CH 3 OOH 1 H 3 O^1 3 CH 3 OOH 21 1 H 2 O
D H 5 2 0.09 eV (18)
The fluoride affinities of all three peroxide species are less than the experimental value of the fluoride affinity of SF 5 (1.657 eV [11]), indicating that a fluoride transfer from SF 62 to the target peroxide species is not exothermic. However, all three species are expected to react exothermically with the fluoride anion, indicating that all three species could be detected by chemical ionization with the F^2 reagent ion:
HOOH 1 F^2 3 F^2 zHOOH
D H 5 2 1.65 eV (19)
CH 3 OOH 1 F^2 3 CH 3 OO^2 zHF
D H 5 2 1.59 eV (20)
CH 3 CH 2 OOH 1 F^2 3 CH 3 CH 2 OO^2 zHF
D H < 2 1.6 eV (21)
Although the calculated F.A. of hydrogen peroxide is potentially greater, within the error of the calcula- tions, than the experimental fluoride affinity of SF 5 , and could possibly be detected via fluoride transfer with SF 62 , the reaction with F^2 is a general detection method for peroxides. Because fluoride adduct forma- tion results in a significant reduction in entropy (in contrast to the other chemical ionization schemes in which the numbers of reactants and products are the same), we have calculated the free energy change in order to be certain that the fluoride adduct reactions are in fact spontaneous. For example, the calculated free energy change for reaction (19) was found to be 2 1.34 eV, indicating that enthalpy effects do indeed dominate the free energy change for these reactions.
5.12. Effect of fluoride hydrate formation on the feasibility of fluoride adduct chemical ionization schemes
The formation of fluoride hydrates [F^2 (H2O) n )] is often unavoidable in the experimental production of
system. Therefore, the ROOH product ions are also present in hydrated form: F^2 zROOHz(H 2 O) n. Figs. 3 and 4 show that the most significant ion signal is contained in the n 5 2 and 3 hydrates. Therefore, since it appears that one of the hydrated water molecules is lost in the reaction between F^2 z(H 2 O) n and ROOH, the dominant reaction pathway seems to be the water exchange reaction described previously:
F^2 z(H 2 O) n 1 ROOH 3 F^2 zROOH(H 2 O) n 21 1 H2O (22) It was a somewhat surprising finding that the additional water (; 1 3 10 14 molecule cm^23 ) added from the ROOH sample introduction had little effect on the observed hydrate distribution ( n 5 4 remained the predominant hydrate). Thermodynamic consider- ations do not serve to explain any unusual stability in the n 5 4 hydrate [42]. In order to investigate this effect, we added a large amount of additional water vapor (; 2 3 1017 molecule cm^23 ) to the system. Under these conditions, the maximum in the hydrate distribution did shift towards larger hydrates ( n 5 6 was the largest observed hydrate), but again the effect was less than expected from thermodynamic consid- erations. We propose two possible explanations for the relative insensitivity of the fluoride hydrate distri- bution to added water vapor. It is possible that the ion sampling potentials on the front aperture and the skimmer cone are great enough to cause some disso- ciation of the larger hydrates, so that the ions that are
transmitted to the quadrupole mass filter are not representative of the ions formed in the reactor. We attempted to investigate this effect, but were not able to obtain ion signals for front aperture potentials of less than 2 130 V (the distribution did not change in the range from 2 130 to 2 210 V). A second expla- nation rests on the details of the kinetics of the fluoride hydrate formation itself. Although thermody- namic measurements do not indicate any special stability for the n 5 4 hydrate, it is possible that there is a kinetic bottleneck in the formation of the n 5 5 hydrate, such that the hydrate distribution is relatively insensitive to water concentration. In any case, this effect is fortuitous because it allows the ion signal to stay relatively concentrated in the n # 4 hydrates and makes the measurement relatively insensitive to changing water vapor concentrations (which may be encountered in laboratory or field applications of the technique).
5.14. Ion–molecule kinetics
In order to address the ultimate sensitivity of this method for both laboratory kinetic and field detection purposes, the rates of the fluoride adduct formation reactions were estimated. The rate of the relevant ion–molecule reaction is one of the factors that directly determine the ultimate sensitivity of the CIMS method since technical limitations usually re- quire relatively short ion–molecule reactions times. Unlike selected ion flow tube (SIFT) techniques, which isolate a single m/z carrier for kinetic studies, our technique does not allow a separate study of the rates of each of the individual fluoride adduct hydrates with Cl2, HOOH and CH 3 OOH. In addition, because of interferences at the ion reactant m/z ratios, the signals resulting from the formation of product ions were followed. Therefore, our kinetic studies repre- sent a weighted average of the relative rates of all processes that lead to a particular product ion. Previ- ous SIFT studies of the Cl 2 1 F^2 z(H 2 O) n reactions have indicated that there is relatively little dependence of the rate constant on hydration levels of n , 4 [28]. Therefore, we must assume that our measurements of the relative rates for reactions involving several
Fig. 4. F^2 (H2O) n chemical ionization mass spectrum for CH 3 OOH.
F^2 z(H 2 O) n species will not be substantially affected by our monitoring procedure. For the Cl 2 , HOOH, and CH 3 OOH kinetic studies, we therefore simply chose to follow the largest product ion signal. The following reactions were studied:
Cl 2 1 F^2 z(H 2 O) n 3 F^2 zCl 2 1 n H 2 O (23)
HOOH 1 F^2 z(H 2 O) n
3 F^2 zHOOH(H 2 O) 2 1 ~ n 2 2! H 2 O (24)
CH 3 OOH 1 F^2 z(H 2 O) n
3 CH 3 OO^2 zHF (H 2 O)2 1 ~ n 2 2! H 2 O (25)
Therefore, the reactions involving Cl 2 , HOOH, and CH 3 OOH were monitored at m/z ratios of 89, 89, and 103, respectively. The relevant rate equation for the bimolecular ion–molecule reactions is given as
d @product ion] dt^5 k [molecule][reagent ion]^ (26)
Because the kinetics measurements were performed under pseudo-first-order conditions ([molecule] .. [ion]), the rate equation simplifies to
d [product ion] dt^5 k^9 [reagent ion]^ (27)
(were k 9 5 k [molecule]), and the integrated rate law in terms of the original bimolecular rate constant k is
[product ion] 5 [reagent ion] (^0)
3 ~ 1 2 e^2 k [molecule] t! 1 C (28)
where C is a constant of integration. For reasons of convenience, we choose to replace [reagent ion] 0 with [product ion]final. These two quantities are related by the proportionality factor A ( A is equal to unity for a reaction which possesses a single reaction channel and has reached completion). If we further relate both concentrations explicitly to the appropriate mass spec- trometer signal, Eq. (28) becomes
[product ion signal] 5 A 3 @product ion signal] (^) final
3 ~ 1 2 e^2 k [molecule] t! 1 C (29)
The ion–molecule bimolecular rate constant is ob- tained by measuring the product ion signal as a function of the molecule concentration (for a fixed reaction time). In order to directly compare kinetic runs performed with different molecule concentra- tions and reaction times, we define the relative time as follows:
t rel 5 [molecule] t (30)
If the signal is also defined on a relative basis (calculated by dividing [product ion signal] t by [prod- uct ion signal] t 5 final; the relative signal thus takes on values from 0 to 1) for each kinetic run, plots of relative signal vs. relative time for experiments with different conditions (different detection sensitivities and different reaction times and molecule concentra- tions) may be directly compared as Eq. (29) simplifies to
[relative product ion signal] 5 A 3 ~ 1 2 e^2 kt rel! 1 C (31) Although the constant (C), which originates from the integration of the rate law, is rigorously zero accord- ing to the boundary conditions, we retain it as a fitting parameter to account for background signal. In summary, the basic experiment and data analy- sis method is as follows: the ROOH ion product signal [i.e. F^2 zROOH (H 2 O) 2 ] is followed as a function of ROOH concentration. Each relative product signal data point is calculated by dividing each absolute product signal data point by the product signal at the highest molecule concentration ([signal F^2 zROOH (H 2 O)2]final). The relative reaction time ( t rel) for each data point is calculated from the absolute reaction time (reaction distance/flow velocity) and the mole- cule concentration for each data point via Eq. (30). The relative product signal as a function of t rel is then fitted via nonlinear least squares techniques to Eq. (31), with A , k , C as adjustable parameters. The process is repeated for the reference reaction (Cl 2 1
molecule present in the chemical system with either the reagent ion or one of the product ions. These reactions have the potential to lower the sensitivity of the method and to create multiple signal carriers for a single m/z ratio. For laboratory experiments, where the number of chemical constituents can be con- trolled, it is relatively straightforward to consider the effect of interference reactions. However, for field experiments, the presence of high concentrations of water vapor and numerous abundant chemical species can make this consideration more difficult. Although the F^2 (H 2 O) n chemical ionization schemes presented here for peroxides have a good probability of being useful in laboratory settings [F^2 (H 2 O) n methods have previously been successfully implemented in labora- tory kinetics studies [40,45], the hydrate formation issue and the reactivity of F^2 (H2O) n with atmospheric acids may complicate the implementation of this approach in field settings. Indeed, our computational prediction that CH 3 OOH should react with H 3 O^1 may be a more promising route for the field detection of this compound, as the H 3 O^1 (H 2 O) n reagent ion is one of the predominant ions formed in ionization sources operating under atmospheric conditions.
6. Conclusions
The equilibrium structures, natural charges and energies for neutral ROOH (R 5 H, CH 3 ) species and their cationic, anionic, proton adduct, and fluoride adduct analogs have been calculated via G2 ab initio methods. The anions are found to be either marginally thermodynamically stable or not stable, and therefore are not expected to be useful as chemical ionization products. However, the ionization potentials of both HOOH and CH 3 OOH were predicted to be less than that of O 2 , suggesting that both species could be detected using chemical ionization mass spectromet- ric methods and an O 21 chemical ionization scheme. The calculated proton affinity of CH 3 OOH suggests that this species could be detected with a H 3 O^1 chemical ionization scheme. The fluoride adduct an- alogs of all three ROOH species were predicted to form exothermically via the reaction ROOH 1 F^2.
Each fluoride adduct species was predicted to form a five-, six-, or seven-member intramolecular ring sys- tem for R 5 H, CH 3 , and CH 3 CH 2 , respectively. Computational studies predict that this cyclization increases the stability of the fluorine adduct by 0.2– 0.3 eV over F^2 zROOH complexes which were fixed at the neutral ROOH geometry. The fluoride adduct reactions involving fluoride hydrate reactants and ROOH were also calculated to be thermodynamically feasible. The feasibility of the fluoride adduct forma- tion chemical ionization method was experimentally verified for HOOH and CH 3 OOH using a chemical ionization mass spectrometer. Unique mass signals were observed at the parent mass and at masses corresponding to hydrated parent ion. The rates of these reactions were found to proceed at or near the collision limited rate, thus making them good candi- dates for sensitive CIMS applications. As demonstrated for the case of ROOH, unique chemical ionization schemes can be developed for atmospherically relevant systems through the com- bined use of computational and experimental studies. Using high level ab initio thermodynamic calcula- tions, the range of possible ionizing agents can be sufficiently narrowed to aid the experimental confir- mation and kinetic study of a proposed chemical ionization mass spectrometric detection method. We hope that the procedure described in this work pro- vides a model for a general development method of CIMS detection schemes, and that the applications developed here aid in the study of peroxides in the laboratory and field environments.
Acknowledgements
The authors thank Dr. Nicole Bennett for her assistance with the synthesis of CH 3 OOH. The au- thors acknowledge support from the Camille and Henry Dreyfus Foundation, the American Chemical Society, Petroleum Research Fund, Research Corpo- ration, the Michigan Space Grant Consortium, and the National Science Foundation (ATM-9874752).
References
[1] J.H. Seinfeld, S.N. Pandis, Atmospheric Chemistry and Phys- ics: From Air Pollution to Climate Change, Wiley, New York,
[2] P.O. Wennberg, T.F. Hanisco, L. Jaegle, D.J. Jacob, E.J. Hintsa, E.J. Lanzendorf, J.G. Anderson, R.-S. Gao, E.R. Keim, S.G. Donnelly, L.A. Del Negro, D.W. Fahey, S.A. McKeen, R.J. Salawitch, C.R. Webster, R.D. May, R.L. Herman, M.H. Proffitt, J.J. Margitan, E.L. Atlas, S.M. Schauf- fler, F. Flocke, C.T. McElroy, T.P. Bui, Science 279 (1998)
[3] W.B. DeMore, S.P. Sander, D.M. Golden, R.F. Hampson, M.J. Kurylo, C.J. Howard, A.R. Ravishankara, C.E. Kolb, M.J. Molina, Chemical Kinetics and Photochemical Data for Use in Stratospheric Modeling, JPL Publication No. 97-4, Jet Propulsion Laboratory, Pasadena, CA, 1997. [4] M. Lee, B. Noone, D. O’Sullivan, B.G. Heikes, J. Atmos. Oceanic Technol. 12 (1995) 1060. [5] J. Polzer, K. Bachman, J. Chromat. A 653 (1993) 283. [6] P.W. Villata, L. Huey, C.J. Howard, J. Phys. Chem. 99 (1995)
[7] J.B. Lipson, M.J. Elrod, T.W. Beiderhase, L.T. Molina, M.J. Molina, J. Chem. Soc. Faraday Trans. 93 (1997) 2665. [8] L.G. Huey, E.J. Dunlea, E.R. Lovejoy, D.R. Hanson, R.B. Norton, F.C. Fehsenfeld, C.J. Howard, J. Geophys. Res. 103 (1998) 3355. [9] F.L. Eisele, H. Berresheim, Anal. Chem. 64 (1992) 283. [10] O. Mohler, F. Arnold, J. Atmos. Chem. 13 (1991) 33. [11] L.G. Huey, D.R. Hanson, C.J. Howard, J. Phys. Chem. 99 (1995) 5001. [12] K.M. Ervin, J. Ho, W.C. Lineberger, J. Phys. Chem. 92 (1988)
[13] A.E.S. Miller, T.M. Miller, A.A. Viggiano, R.A. Morris, J.M. Van Doren, S.T. Arnold, J.F. Paulson, J. Chem. Phys. 102 (1995) 8865. [14] M.T. Bowers, Gas Phase Ion Chemistry, Academic, New York, 1979. [15] M.J. Frisch, G.W. Trucks, H.B. Schlegel, P.M.W. Gill, B.G. Johnson, M.A. Robb, J.R. Cheeseman, T. Keith, G.A. Peters- son, J.A. Montgomery, K. Raghavachari, M.A. Al-Laham, V.G. Zakrzewski, J.V. Ortiz, J.B. Foresman, J. Cioslowski, B.B. Stefanov, A. Nanayakkara, M. Challacombe, C.Y. Peng, P.Y. Ayala, W. Chen, M.W. Wong, J.L. Andres, E.S. Re- plogle, R. Gomperts, R.L. Martin, D.J. Fox, J.S. Binkley, D.J. Defrees, J. Baker, J.P. Stewart, M. Head-Gordon, C. Gonza- lez, J.A. Pople, GAUSSIAN 94 (Revision E.2), Gaussian, Inc., Pittsburgh, PA, 1995. [16] L.A. Curtiss, K. Raghavachari, G.W. Trucks, J.A. Pople, J. Chem. Phys. 94 (1991) 7221. [17] L.A. Curtiss, P.C. Redfern, K. Raghavachari, J.A. Pople, J. Chem. Phys. 109 (1998) 42. [18] B.M. Messer, M.J. Elrod, Chem. Phys. Lett. 301 (1999) 10.
[19] Calculated frequencies for all species are available from the authors upon request. [20] L.A. Curtiss, J.E. Carpenter, K. Raghavachari, J.A. Pople, J. Chem. Phys. 98 (1993) 1293. [21] E.D. Glendening, A.E. Reed, J.E. Carpenter, F. Weinhold, NBO Version 3.1. [22] J.P. Foster, F. Weinhold, J. Am. Chem. Soc. 102 (1980) 7211. [23] D.W. O’Sullivan, M.L. Lee, B.C. Noone, B.G. Heikes, J. Phys. Chem. 100 (1996) 3241. [24] G.L. Vaghjiani, A.R. Ravishankara, J. Phys. Chem. 93 (1989)
[25] N.V. Klassen, D. Machington, H.C.E. McGowan, Anal. Chem. 66 (1994) 2921. [26] J.V. Seeley, J.T. Jayne, M.J. Molina, Int. J. Chem. Kinet. 25 (1993) 571. [27] N.G. Adams, M.J. Church, D. Smith, J. Phys. D. 8 (1975)
[28] J.V. Seeley, R.A. Morris, A.A. Viggiano, J. Phys. Chem. 100 (1996) 15821. [29] J. Koller, M. Hodoscek, B. Plesnicar, J. Am. Chem. Soc. 112 (1990) 2124. [30] R.D. Bach, P.Y. Ayala, H.B. Schlegel, J. Am. Chem. Soc. 118 (1996) 12765. [31] R.H. Hunt, R.A. Leacock, C.W. Peters, K.T. Hecht, J. Chem. Phys. 42 (1965) 1931. [32] J. Hrusak, H. Frierichs, H. Schwarz, H. Razafinjanahary, H. Chermette, J. Phys. Chem. 100 (1996) 100. [33] X. Luo, P.R. Fleming, T.R. Rizzo, J. Chem. Phys. 96 (1992)
[34] J. Hrusak, S. Iwata, J. Chem. Phys. 106 (1997) 4877. [35] F.S. Ashmore, A.R. Burgess, J. Chem. Soc. Faraday Trans. 2 73 (1977) 1247. [36] E.P. Hunter, S.G. Lias, J. Phys. Chem. Ref. Data 27 (1998)
[37] E.C.M. Chen, J.R. Wiley, C.F. Batten, W.E. Wentworth, J. Phys. Chem. 98 (1994) 88. [38] M.J. Travers, D.C. Cowles, G.B. Ellison, Chem. Phys. Lett. 164 (1989) 449. [39] NIST Standard Reference Database 69, November 1998 Release, NIST Chemistry WebBook (http://webbook.nist.gov/ chemistry). [40] K.W. Scholtens, B.M. Messer, C.D. Cappa, M.J. Elrod, J. Phys. Chem. 103 (1999) 4378. [41] S.S. Xantheas, L.X. Dang, J. Phys. Chem. 100 (1996) 3989. [42] R.G. Keesee, A.W. Castleman, J. Phys. Chem. Ref. Data 15 (1986) 1011. [43] E.A. Cohen, H.M. Pickett, J. Mol. Spectrosc. 87 (1981) 582. [44] E. Barbagli, M. Maestro, Chem. Phys. Lett. 24 (1974) 567. [45] J.V. Seeley, R.F. Meads, M.J. Elrod, M.J. Molina, J. Phys. Chem. 100 (1996) 4026. [46] R.G. Tonkyn, J.W. Winniczek, M.G. White, Chem. Phys. Lett. 164 (1989) 137.
