Masakazu
Nakajima
a and
Yasuki
Endo
b
aDepartment of Basic Science, The University of Tokyo, 3-8-1 Komaba, Meguro-ku, Tokyo 153-8902, Japan. E-mail: nakajima@bunshi.c.u-tokyo.ac.jp
bDepartment of Applied Chemistry, National Yang Ming Chiao Tung University, 1001 Ta-Hsueh Rd., Hsinchu 300093, Taiwan. E-mail: endo@nycu.edu.tw
First published on 10th April 2025
Vinylperoxy radicals were produced in a pulsed supersonic jet by discharging a gas mixture of vinyl bromide and molecular oxygen largely diluted in the Ar or Ne buffer gas. Two conformers of the radical, s-trans and s-cis, were detected through their pure rotational transitions via Fourier-transform microwave spectroscopy. Fine and hyperfine components in the observed spectra were fully assigned and analyzed to determine precise molecular constants for each conformer. The Fermi coupling constants determined for the –CH2 protons indicate that non-negligible unpaired spin density is located on the terminal carbon atom, although the radical is generally considered as the oxygen-centered radical. The intensities of the observed spectra are much weaker than expected, probably because most of nascent vinylperoxy radicals formed by the association of the vinyl radical and O2 promptly dissociate either to the vinoxy radical and atomic oxygen, or to formaldehyde and the formyl radical, even under jet-cooled conditions.
As briefly mentioned above, many experimental and theoretical studies on the C2H3 + O2 reaction have been reported because of its importance in hydrocarbon oxidation. However, knowledge on the intermediate, VPR, is quite limited. So far, several spectroscopic studies have been reported for this radical through its vibrational and electronic transitions. UV-visible spectra of the radical have been observed in aqueous solutions25,26 and in the gas-phase.24,27 The observed spectra show a structureless broad absorption in the range of 220–500 nm with a strong feature centered at about 250 nm. In the gas-phase experiments,24,27 the vinyl radical generated by photolysis of ether vinyl bromide (C2H3Br) or methyl vinyl ketone (CH3COC2H3) was converted to the VPR by the association with O2, and the transient absorption was observed. On the other hand, the reported transient absorption was absent in a similar gas-phase experiment with a lower concentration of the vinyl radical.13 The spectral carrier of the reported UV-visible gas-phase spectra24,27 seems to be still open for discussion. An infrared spectrum of the VPR was reported using the matrix-isolation technique by Yang et al.17 They prepared the radical by the C2H3 + O2 reaction in an Ar matrix, and two absorption features at 1140.7 and 875.5 cm−1 were identified to be attributed to the VPR, with the help of density functional calculations and isotope substitution experiments. As far as the author's knowledge, only a preliminary observation has been reported for the near-IR absorption of the gas-phase VPR by the Miller group.28 They have systematically observed near-IR cavity-ringdown absorption spectra for many alkylperoxy radicals prepared by the alkyl radical association with O2 in the gas phase.29 The preliminary observation provided highly congested spectra in which the carrier species is strongly inferred to be the VPR. However, the identification of the spectral carrier is still tentative.28
In this paper, we report Fourier-transform microwave spectra of two conformers (s-trans and s-cis) of the VPR generated by the reaction C2H3 + O2. Because the radical is an open-shell species and has three protons with a nuclear spin of 1/2, its rotational spectrum splits into a number of fine and hyperfine components. We observed microwave spectra showing such congested structures, and the spectral carrier was unambiguously identified to be the VPR by fully analyzing the observed fine and hyperfine components. Precise molecular constants were determined for each of the conformers by least-squares analysis for a data set of the observed line frequencies. At the end of this article, we will make a brief discussion on the product branching of the reaction C2H3 + O2, based on the spectral intensities of the VPR, the vinoxy radical, and formaldehyde, observed in the same discharged jet.
As shown in Fig. 1, the VPR has two stable conformers depending on the direction of the O–O bond relative to the C–C bond.16,17,21,27,32 Both of them are planar with Cs symmetry, and one conformer with the O–O bond located at the trans position is referred to as s-trans-VPR, and the other is s-cis-VPR. Molecular geometry of each conformer was optimized by referring to the electronic energy calculated at the RCCSD(T)-F12a level of theory using the cc-pCVTZ-F12 basis set, with all the core electrons included in the electron correlation as well as the valence electrons. The optimized bond lengths and angles of each conformer are given in Fig. S1 of the ESI.† The optimized geometries show reasonable agreements with those previously reported.15–18,27,32 Zero-point energy correction on the electronic energy at the optimized geometry was made by calculating harmonic vibrational frequencies at the B3LYP/aug-cc-pVTZ level of theory. The zero-vibrational level of the s-trans-VPR was predicted to be lower in energy by 312 cm−1 than that of the s-cis-VPR. Zero-point vibrational corrections on the equilibrium rotational constants corresponding to the optimized geometry of each conformer were made by using the vibration–rotation α constants calculated at the B3LYP/aug-cc-pVTZ level of theory. The theoretical harmonic frequencies and the vibration–rotation constants used for the zero-vibrational correction are listed in Table S1 of the ESI.† The theoretical energy difference between the two conformers and the predicted rotational constants are summarized in Table 1.
![]() | ||
Fig. 1 Two conformers of the vinylperoxy radical. The s-cis-conformer is predicted to be higher in energy than the s-trans-conformer by 312 cm−1 (cf.Table 1). The direction of the theoretically calculated permanent dipole moment is shown for each conformer as a bold gray arrow together with the principal a- and b-axes. |
s-trans | s-cis | |
---|---|---|
a Zero-point energy is estimated from theoretical harmonic frequencies calculated at the B3LYP/aug-cc-pVTZ level of theory. b Zero-vibrational corrections are made using theoretical vibration–rotation constant α calculated at the B3LYP/aug-cc-pVTZ level of theory. | ||
ΔEe/cm−1 | 0 | 284 |
A e/MHz | 51![]() |
21![]() |
B e/MHz | 4964 | 7009 |
C e/MHz | 4525 | 5282 |
ΔE0a/cm−1 | 0 | 312 |
A 0 /MHz | 50![]() |
21![]() |
B 0 /MHz | 4934 | 6946 |
C 0 /MHz | 4497 | 5239 |
|μa|/D | 2.44 | 2.28 |
|μb|/D | 0.65 | 0.60 |
The components of the permanent dipole moment along the principal axes were obtained for each conformer by computing the electronic energies at the same level of theory as the geometry optimization, with an external electric field virtually applied along the principal axes. The results are listed in Table 1. The direction of the dipole moment (see Fig. 1) is almost parallel to the principal a-axis of each conformer. Fine and hyperfine coupling constants of each conformer were also predicted by using the NMR properties keyword of Gaussian 16 program at the B3LYP/aug-cc-pVTZ level of theory.
The minimum energy path of the interconversion between the two conformers is thought to be along the coordinate of the CCOO dihedral angle. Fig. 2 shows the B3LYP/aug-cc-pVTZ potential energy curve calculated by changing the dihedral angle from 0° (s-cis) to 180° (s-trans) with other geometrical parameters fully relaxed. The interconversion barrier is estimated to be roughly 2000 cm−1, which is in good agreement with the previously estimated values.21,28
A Balle–Flygare type Fourier-transform microwave (FTMW) spectrometer34 operated in the frequency region of 4–40 GHz was used for the observation of microwave transitions. In order to reduce the Doppler broadening of spectral lines, the discharged jet expanded into the cavity of the spectrometer coaxially with the microwave propagating axis, although two Doppler components appeared in the spectrum for each MW transition.35 The rotational temperature of molecules in the jet was thought to be less than 3 K. The Earth's magnetic field in the center region of the MW cavity was cancelled out by using three Helmholtz coils to prevent spectral line splitting due to the Zeeman effect.
An FTMW-MW double-resonance technique36 was employed for observing several weak b-type transitions of the VPR. In the double-resonance experiment, a certain microwave transition was monitored with an FTMW spectrometer, and then the frequency of the pump MW radiation, which illuminated the discharged jet after the irradiation of the probe MW pulse for FTMW spectroscopy, was scanned. In this case, the macroscopic polarization induced by the probe MW pulse is “broken” if the pump MW is resonant with a transition of which either the upper or lower level is shared by the monitored transition, so that depletion of the monitored signal is observed as the double-resonance signal.
A reduced form of the spin-rotation Hamiltonian37 for doublet asymmetric top molecules of Cs symmetry, together with the hyperfine interaction Hamiltonian due to the three protons of the VPR, was used to reproduce the energy levels of the radical. A least-squares analysis of the experimental frequency data was carried out with a homemade program. A somewhat detailed description of the analysis program and comparisons with the spfit/spcat program suite,38,39 which is an analysis tool widely used for high-resolution spectroscopy, has been given elsewhere.40 Molecular constants determined for the s-trans-VPR are listed in Table 2 together with the theoretically predicted constants. Twenty-one parameters are determined from the 81 line positions, and all the observed line positions are reproduced with the standard deviation of 3.1 kHz, comparable to the expected experimental accuracy ∼ 3 kHz. The determined molecular constants show reasonable agreements with the theoretical ones, being strong evidence for the identification of the spectral carrier.
s-trans-conformer | s-cis-conformer | |||
---|---|---|---|---|
Experimenta | Theoryb | Experimenta | Theoryb | |
a Values in parentheses are 1σ error and applied to the last digits. b Rotational constants are cited from Table 1. Centrifugal distortion constants and spin coupling constants are predicted using Gaussian 16 software suite,31 by computing vibrational–rotational coupling (VibRot keyword), and NMR shielding tensors and magnetic susceptibilities (NMR keyword), respectively, at the B3LYP/aug-cc-pVTZ level of theory. c Fixed. | ||||
A | 50![]() |
50![]() |
21![]() |
21![]() |
B | 4925.66940(57) | 4934 | 6951.4084(11) | 6946 |
C | 4490.74247(63) | 4497 | 5235.4680(11) | 5239 |
103ΔN | 1.161(21) | 1.10 | 5.03c | 5.03 |
103ΔNK | −7.03(36) | −5.47 | −18.32(60) | −20.0 |
103ΔK | 507c | 507 | 80.9c | 80.9 |
103δN | 0.121c | 0.121 | 1.55c | 1.55 |
103δK | 6.30c | 6.30 | 10.2c | 10.2 |
ε aa | −2970.6129(62) | −2292.3 | −616.078(12) | −494.4 |
ε bb | −88.3943(44) | −76.0 | −374.8321(93) | −285.2 |
ε cc | −1.0690(38) | 2.7 | −0.8914(90) | 3.3 |
(εab + εba)/2 | −163.78(13) | −363.7 | 294.768(80) | 274.3 |
a F (H1) | −10.4117(39) | −11.194 | −10.063(13) | −10.770 |
T aa (H1) | 2.5423(48) | 2.733 | 6.810(21) | 8.429 |
T bb (H1) | −2.002(15) | −2.233 | −6.084(46) | −7.603 |
T ab (H1) | −5.49(34) | −6.674 | −1.472c | −1.472 |
a F (H2) | −8.6212(46) | −9.695 | −10.007(14) | −10.856 |
T aa (H2) | −0.6959(79) | −2.372 | 2.362(17) | 1.136 |
T bb (H2) | 2.459(16) | 4.140 | 1.007(33) | 1.881 |
T ab (H2) | 3.51(37) | 2.971 | 1.22(42) | −0.132 |
a F (H3) | 2.9815(31) | 3.252 | 1.1987(64) | 2.644 |
T aa (H3) | 4.5530(47) | 5.249 | 5.420(11) | 5.983 |
T bb (H3) | 0.818(18) | 0.283 | −1.380(28) | −1.714 |
T ab (H3) | −6.73(33) | −7.319 | −2.712c | −2.712 |
For the s-cis-VPR, four a-type transitions, 101–000, 202–101, 212–111, and 211–110, were observed by FTMW spectroscopy, and one b-type transition 111–000 was observed with the double-resonance technique by monitoring the 101–000 transition. The intensities of the observed FTMW spectra were about one-half to one-third of those of the s-trans-VPR, probably reflecting the energy difference between the two conformers. A total of 48 lines including fine and hyperfine components were observed for the s-cis-VPR (cf. Table S3 of the ESI†), and 18 molecular parameters were determined by a least-squares analysis, as listed in Table 2. The standard deviation of the fit is 6.0 kHz, which is slightly larger than the expected experimental accuracy. A larger error may be introduced in reading line positions from the s-cis VPR spectra, because they were generally weak and observed with poor S/N ratios.
The ground electronic state of the VPR is 2A′′, so that the unpaired electron is sitting in a molecular π-orbital consisting of the out-of-plane p-orbitals of the oxygen and carbon atoms. Following the McConnell relationship,41,42 the Fermi coupling constant aF of the protons of the VPR is predominantly proportional to the unpaired spin density on the carbon atom to which the proton is bonded. Due to the spin-polarization effect, a positive spin density on the carbon atom causes a negative spin density at the proton nucleus, resulting in a negative aF value for the proton. As seen in Table 2, this is true for the –CH2 protons of the VPR. The aF values of all the –CH2 protons of the VPR are almost constant, ca. − 10 MHz, and no conformer dependence is clearly seen. On the other hand, the signs of the aF values of the –CH protons (H3) are positive. This may be caused by a negative unpaired spin density43 on the –CH carbon.
For halogen-substituted methyl radicals, CH2F44 and CH2Cl,45 which are typical carbon-centered π-radicals, each proton has a negative aF value of ca. −60 MHz, which is 6 times as large as those of the –CH2 protons of the VPR. Because ca. 85% unpaired electron density is estimated to be located on the carbon atom of the halogen-substituted methyl radicals,45 about 10% unpaired electron density is expected on the –CH2 carbon of the VPR. The VPR is generally considered as an oxygen-centered radical, described as structure 1 of Fig. 3. However, the experimental aF values infer the non-negligible unpaired spin density on the –CH2 carbon. Lewis structures with the unpaired electron located at the –CH2 carbon can be depicted as structures 2 and 3 of Fig. 3, in which formal charges on the carbon and oxygen atoms are inverted to each other. Taking into account the electronegativities of carbon and oxygen, Lewis structure 2 seems to be the more appropriate formula. The zwitterionic character and the locations of the formal charges of structure 2 may reasonably explain the relatively large permanent dipole moments calculated for VPRs (cf.Table 1) and their directions (cf.Fig. 1). Furthermore, in Lewis structure 2, the –CH carbon has a vacant p-orbital, and the terminal oxygen has three lone pairs. If electrons of one of the three lone pairs are donated to the vacant p-orbital, a C–O bond may be formed between the –CH carbon and the terminal oxygen, resulting in the three-membered ring isomer, ˙CH2-c-C(H)OO (dioxiranyl-methyl), which is postulated as an intermediate connecting the VPR to the HCHO + HCO dissociation channel, as mentioned in the Introduction section. The contribution of Lewis structure 2 may be an advantage for the formation of the three-membered ring intermediate from both of the two conformers.
![]() | ||
Fig. 3 Lewis structures of the vinylperoxy radical. (1) oxygen-centered radical. (2) and (3) carbon-centered radicals. |
The intensity of the FTMW spectrum observed for the VPR was much weaker than that expected from spectral intensity, for example, of CH3OO46 which is produced by a similar production scheme to the VPR, the association reaction of CH3 + O2, followed by collisional relaxation with third-body buffer gas atoms. It seems that collisional stabilization occurs only for a small portion of vibrationally excited nascent VPRs, and most of them might quickly dissociate either to the vinoxy radical + O or to formaldehyde + HCO under the present experimental conditions. For comparison of the spectral intensity of the VPR to those of the dissociation products, FTMW spectra of the vinoxy radical and formaldehyde were observed under the same experimental conditions as in the VPR observation. The observed spectra for the 202–101 transition of the s-trans-VPR, the 101–000 transition of the vinoxy radical, and the 211–212 transition of formaldehyde are shown in Fig. 4 with the same intensity scale. The most intense fine/hyperfine components are present in the spectra of the s-trans-VPR and the vinoxy radical. Note that the intensity of the formaldehyde spectrum cannot be directly compared to those of the s-trans-VPR and the vinoxy radical, because the spectra of the radicals split into many fine and hyperfine components due to the unpaired electron and the three protons. However, it is obvious from Fig. 4 that the concentration of the vinoxy radical is higher than that of the s-trans-VPR in the jet expansion.
Taking into account the spectral intensity factors (thermal population, effect of fine/hyperfine splittings, and so on), relative concentrations of the three species in the jet expansion are roughly estimated from the spectral intensities of Fig. 4 to be 5:
25
:
70 for the VPR, the vinoxy radical, and formaldehyde, respectively, by assuming a rotational temperature of 2.5 K. Detailed descriptions of this estimation are given in the ESI.† If the final product channels of the C2H3 + O2 reaction are only three, yielding the stabilized VPR, the vinoxy radical, or formaldehyde, and all molecules of these three species were produced by this reaction in the present experiment, more than 90% of the association products (vibrationally excited VPRs) promptly proceed to the final product channels, CH2CHO + O and HCHO + HCO. Only 5% of the association products are stabilized by collisions with the buffer gas even under the jet cooled conditions. Interestingly, Matsugi and Miyoshi13 reported that the previously reported near-UV absorption of gas-phase VPR was absent in their experiment, although they conducted a similar experiment to the previous measurements.24,27 They suggested that the previously reported spectra may be attributed to other transient species. There is no contradiction between the absence of the near-UV absorption and our estimation of the very small concentration of the gas-phase VPR. The branching fraction of the vinoxy radical + O channel has been experimentally determined to be ca. 20% by Oguchi et al.12 under low pressure conditions. The experimentally determined fraction is reproduced by the theoretical calculation based on the RRKM theory.13 The same calculation predicts the branching fraction of the formaldehyde + HCO channel to be slightly less than 80%. The present estimation of the relative concentrations of the three reaction products shows good agreement with the previous experimental and theoretical results of the reaction kinetic studies.
Footnote |
† Electronic supplementary information (ESI) available: Theoretical equilibrium geometries, vibration–rotation constants, and frequency lists of observed microwave transitions for the s-trans- and s-cis-vinylperoxy radicals. See DOI: https://doi.org/10.1039/d5cp00649j |
This journal is © the Owner Societies 2025 |