Vacuum ultraviolet photochemistry of the conformers of the ethyl peroxy radical

Abstract

We study the conformers of the ethyl peroxy radical (C₂H₅O₂), the simplest peroxy radical having more than one conformer, by combining synchrotron radiation vacuum ultraviolet (VUV) photoionization mass spectrometry with theoretical calculations. The ethyl peroxy radical is formed in a microwave discharge flow tube through the reaction of the ethyl radical (C₂H₅) with oxygen molecules, where C₂H₅ is generated via the hydrogen-abstraction reaction of ethane with fluorine atoms. Two kinds of C₂H₅⁺, originating from photoionization of C₂H₅ and from dissociative photoionization of C₂H₅O₂, whose cation is not stable, have been identified and separated in photoionization mass spectra. The photoionization spectrum corresponding to C₂H₅O₂ is obtained and assigned with Franck–Condon calculations. The present findings show that the gauche conformer (G-C₂H₅O₂) of C₂H₅O₂ has favorable Franck–Condon factors in the ionization transitions, whereas the contribution of the trans conformer (T-C₂H₅O₂) to the photoionization spectrum is minor or negligible due to its large geometric changes in the photoionization process. Moreover, the reason for the instability of C₂H₅O₂⁺ and its detailed dissociation mechanisms have been unraveled with the aid of the calculated potential energy curves.

---

1. Introduction

Peroxy radicals (RO₂) are very important reaction intermediates that play crucial roles in the low temperature oxidation of organic compounds in combustion and atmospheric chemistry.^1–3 In the atmosphere, peroxy radicals are formed via the reaction of alkyl radicals (R) with oxygen molecules. Then the formed peroxy radicals can undergo bimolecular reactions with various species, such as NOx (NO and NO₂), HO₂ and other peroxy radicals, self-reaction and unimolecular reaction. These key processes make an important contribution for the production of secondary pollutants such as ozone and secondary organic aerosols (SOA).^2,3 Thus the direct measurement of peroxy radicals is a fundamental issue in atmospheric chemistry.

The spectroscopy and structure of peroxy radicals are of considerable importance for understanding their reactivity, sources and sinks in the atmosphere. In the past decades, benefiting from the constant development of analytical methods such as absorption spectroscopy,^4,5 cavity ring-down spectroscopy (CRDS),^6–9 photoionization mass spectrometry (PIMS)^10,11 and photoelectron photoion coincidence spectroscopy (PEPICO),^12,13 a great deal of advance has been made to unravel their spectroscopy and structures. Among them, PIMS and PEPICO in particular combined with tunable synchrotron radiation as the photoionization light source have attracted attention and are considered as universal and sensitive methods to probe and analyze such elusive species.^14–16 For example, Meloni et al. studied the vacuum ultraviolet (VUV) photoionization of alkylperoxy radicals at the Advanced Light Source by using PIMS and found most of their cations unstable except for the methyl peroxy radical (CH₃O₂).¹⁰ The ionization energy (IE) of CH₃O₂ and the appearance energy (AE) of the CH₃⁺ fragment ion in dissociative photoionization of CH₃O₂ were measured using PEPICO by Voronova et al. at the Swiss Light Source and Tang et al. at SOLEIL, France.^12,13

With the increase in mass, the structure of peroxy radicals becomes complex, having several isomers or conformers. For instance, the ethyl peroxy radical has two conformers, the gauche conformer (G-C₂H₅O₂) with its CCOO dihedral angle of ±60° and the trans conformer (T-C₂H₅O₂) with the CCOO dihedral angle of ±180°, and the propyl peroxy radical has two isomers, 1-C₃H₇O₂ and 2-C₃H₇O₂, each of which also has several conformers.⁶ The knowledge on the isomers or conformers of peroxy radicals is very limited, and identifying them is still a challenge. Recently, using the method of PEPICO at SOLEIL, two isomers of C₃H₇O₂ were successfully identified and assigned in our publication.¹⁷

Here we focus on the ethyl peroxy radical, C₂H₅O₂, and its two conformers, G-C₂H₅O₂ and T-C₂H₅O₂. The ethyl peroxy radical is the simplest peroxy radical having more than one conformer, and has attracted attention.^18–21 Previous theoretical calculations showed that G-C₂H₅O₂ is the global minimum with an energy of ∼80 cm⁻¹ lower than T-C₂H₅O₂.^20,21 Upon photoionization, the nascent C₂H₅O₂⁺ cation is not stable and dissociates to C₂H₅⁺ and O₂ fragments.^10,11 To explain the instability of C₂H₅O₂⁺, Meloni et al. suggested the cation has a hyper-conjugation structure.¹⁰ The AE of the C₂H₅⁺ and O₂ fragments was measured at 10.0 ± 0.1 eV in the photoionization spectrum by these authors.

In this work, VUV photoionization of C₂H₅O₂ has been investigated in detail by using synchrotron-based PIMS complemented by high-level theoretical calculations on the IEs, structures and spectra. After distinguishing two different sources of C₂H₅⁺, originating from the photoionization of the ethyl radical (C₂H₅) and from the dissociative photoionization of C₂H₅O₂, the photoionization spectrum corresponding to C₂H₅O₂ is acquired and has been assigned with Franck–Condon calculations. The present findings will demonstrate that the G-C₂H₅O₂ conformer has favorable Franck–Condon factors in the ionization transition, whereas the contribution of T-C₂H₅O₂ to the photoionization spectrum is minor or negligible due to its large geometric changes in the photoionization process.

2. Methods

The experiments were carried out on the combustion and flame beamline at National Synchrotron Radiation Laboratory (NSRL) in Hefei, China.²² A fast flow tube was employed as the reactor to generate free radicals and initiate their chemical reactions.²³ A home-made photoionization orthogonal acceleration reflectron time-of-flight (TOF) mass spectrometer was installed to probe and analyze reactants, reaction intermediates and products inside the fast flow tube.²⁴ The detailed configurations of the synchrotron beamline, the fast flow tube and the TOF mass spectrometer have already been introduced before and only a brief description is presented here.^22–24

Briefly, on the beamline, synchrotron photons emitted from an undulator were dispersed by a monochromator equipped with two gratings, 200 and 400 l mm⁻¹, covering the photon energy range of 5–21 eV. A gas filter presently filled with Ar gas was adopted to suppress high-order harmonic radiations from the undulator, and the absolute photon energy was calibrated within an accuracy of ±10 meV using the resonant absorption lines of Ar in the gas filter.²⁵ The synchrotron photon flux was measured using a photodiode (SXUV-100, IRD Inc.) and used to normalize ion signals in the photon energy scans.

The fast flow tube mainly including a main tube and a coaxial movable injector was installed inside the source chamber of the TOF mass spectrometer.²³ Fluorine atoms were generated from 1% diluted F₂ gas in helium using a 2.45 GHz microwave discharge generator (GMS-200W, Sairem, France) and introduced into the fast flow tube to initiate chemical reactions. Other reactant gases such as ethane and oxygen, together with helium bath gas, were injected via the arms of the main tube or the injector. The total pressure inside the flow tube was monitored using a capacity gauge and fixed at 2 Torr by using a closed-loop feedback throttle valve.

After passing through one skimmer (1 mm diameter), the gas mixture in the fast flow tube was sampled and entered into the ionization chamber of the TOF mass spectrometer, crossing a synchrotron photon beam at a right angle and then being ionized. A cage-shaped photoionization source connected with an Einzel lens was developed to extract and focus ions, and then to enhance the detection sensitivity of the setup.²⁴ The TOF mass spectrometer with an orthogonal acceleration and reflectron structure was employed to analyze the mass of ions. The total ion flight length was ∼1 m and the mass resolving power of the TOF mass spectrometer was measured to be M/ΔM ∼ 2000 (FWHM, the full width at half maximum).²⁴

To provide further insights into the experimental data, theoretical calculations have been performed to study the adiabatic ionization energies (AIEs), the potential energy curves and the Franck–Condon factors, using the Gaussian 16 package and MOLPRO 2015 software.^26,27 The geometric structures of the two conformers of C₂H₅O₂ and their cations were optimized by using the M06-2X/aug-cc-pVTZ theoretical methods. The AIEs of the two conformers were calculated using the explicitly correlated coupled cluster single-double and perturbative triple excitations approach, (R)CCSD(T)-F12/aug-cc-pVTZ.²⁷ The potential energy curves of the low-lying electronic states of neutral C₂H₅O₂ and its cation along the C–O coordinates have been calculated at the M06-2X/aug-cc-pVTZ level. In order to assign the photoionization spectrum of C₂H₅O₂, the Franck–Condon factors in the ionization transitions have been calculated at the M06-2X/aug-cc-pVTZ level too, using the time-independent adiabatic Hessian Franck–Condon model in the Gaussian program.²⁶

3. Results and discussion

The experiments were performed in sequence and can be divided into two parts. At first, the experiment was carried out without adding oxygen gas into the fast flow tube, and the ethyl radical (C₂H₅) was produced via the hydrogen-abstraction reaction of ethane with fluorine atoms.

F + C₂H₆ → C₂H₅ + HF (1)

Secondly, with the addition of oxygen gas, the nascent ethyl radical reacted with oxygen to generate the products, C₂H₅O₂, ethylene (C₂H₄) and the hydroperoxy radical (HO₂), via the following two reactions.^19,28

C₂H₅ + O₂ → C₂H₅O₂ (2a)

→ C₂H₄ + HO₂ (2b)

The photoionization mass spectrum in the absence of oxygen in the fast flow tube was acquired at the photon energy of hν = 9.6 eV and is presented as a blue line in Fig. 1. An intense peak is observed at m/z = 29 and has been assigned as C₂H₅ produced from the above mentioned hydrogen-abstraction reaction (1). Another mass peak at m/z = 30 is observed in the mass spectrum too, and can be ascribed to the ¹³C₂H₅ isotopic ethyl radical with an intensity of about 4% of the m/z = 29 peak. The IE of ethane locates at 11.52 eV,²⁹ above the present photon energy of 9.6 eV, indicating that ethane does not contribute to the m/z = 30 peak.

Fig. 1 Photoionization TOF mass spectra acquired without (in blue, hν = 9.6 eV) and with (in red, hν = 9.6 eV; in black, hν = 11.0 eV) the addition of oxygen into the fast flow tube.

The photoionization spectrum of C₂H₅ has been measured by scanning the synchrotron photon energy with a step size of 0.05 eV, and is presented as black squares in Fig. 2(a). The overall shape of the photoionization spectrum is similar to the literature results of Ruscic et al.,³⁰ and the difference might be due to the different internal energies of the produced ethyl radicals and the different photo-flux calibrations in the two experiments. Presently the cationic signal appears at hν = 8.33 ± 0.05 eV in the photoionization spectrum, in accordance with the IE of C₂H₅.^29,31

Fig. 2 Photoionization spectra of the m/z = 29 cations acquired without (a) and with (b) the addition of oxygen gas into the flow tube, together with the reference data of C₂H₅ (a, in red) reprinted from Ruscic et al.³⁰ (with the permission of AIP Publishing), the reference data of C₂H₅O₂ (b, in red) adapted from Meloni et al.¹⁰ (Copyright 2021 American Chemical Society), and the calculated results of C₂H₅O₂ (b, in blue).

After adding oxygen gas into the flow tube, the intensity of the m/z = 29 mass peak declines, nearly approaching the baseline of the mass spectrum, as shown using a red line in Fig. 1. In addition, almost no other peaks can be observed in the mass spectrum. Previous results have shown that the reaction of C₂H₅ with oxygen is fast with a rate constant of 2 × 10⁻¹² cm³ molecule⁻¹ s⁻¹ at 298 K,³² and can generate the products of C₂H₅O₂, C₂H₄ and HO₂. The C₂H₅O₂⁺ cation is not stable and the AE of the C₂H₅⁺ and O₂ fragments was measured at 10.0 ± 0.1 eV,¹⁰ above the present photon energy of 9.6 eV. For the other products, C₂H₄ and HO₂, their IEs locate at 10.5138 and 11.359 eV,²⁹ respectively, and thus they are not observed in the mass spectrum.

The synchrotron photon energy was increased to hν = 11.0 eV, and the photoionization mass spectrum was acquired and is presented as a black line in Fig. 1. Similar to the previous results,^10,11 still no signal could be observed at m/z = 61 (C₂H₅O₂⁺) in the mass spectrum. But, presently the m/z = 29 mass peak appears again and this reappearance should have a different source as that of the above C₂H₅ obtained in the absence of O₂. In addition, the product of C₂H₄ (m/z = 28) from the reaction of C₂H₅ with oxygen has been observed in the mass spectrum. Some other peaks are observed in the mass spectrum too and can be ascribed to the secondary reactions occurring in the flow tube.³³ For example, the m/z = 30 mass peak is ascribed to the secondary reaction product of HCHO whose IE is measured at 10.83 ± 0.05 eV in the photoionization spectrum (Fig. S1, ESI†), and the m/z = 62 mass peak can be assigned as C₂H₅OOH, the product of the reaction of C₂H₅O₂ with HO₂.

To explain the reappearance of the m/z = 29 peak in the mass spectrum, its photoionization spectrum was acquired (with the addition of oxygen) and is presented as black squares in Fig. 2(b), together with the literature data of Meloni et al.¹⁰ Unlike the case in Fig. 2(a), presently the m/z = 29 cation does not appear until at hν = 10.08 ± 0.05 eV in the photoionization spectrum, agreeing well with the literature, AE(C₂H₅⁺) = 10.0 ± 0.1 eV in dissociative photoionization of C₂H₅O₂ measured by Meloni et al.,¹⁰ and thus should be ascribed to the contribution of C₂H₅O₂.

The neutral C₂H₅O₂ at the X²A′′ ground electronic state has an open-shell electronic configuration of (11a′)²(12a′)²(3a′′)²(13a′)²(4a′′)¹.³⁴ Theoretical calculations show that removing an electron from the 13a′ orbital or from the 4a′′ orbital through photoionization can lead to the X³A′′ ground electronic state or the a¹A′ first excited electronic state of C₂H₅O₂⁺, respectively. The geometric structures of the two conformers of C₂H₅O₂, G-C₂H₅O₂ and T-C₂H₅O₂, in the X²A′′ ground state are optimized at the M06-2X/aug-cc-pVTZ level of theory and presented in Fig. 3, together with the optimized structures of their individual cations in the X³A′′ and a¹A′ electronic states after checking their individual wavefunction stablity.²⁶ It is shown that G-C₂H₅O₂ takes a C₁ point-group symmetry and T-C₂H₅O₂ has a Cs symmetry. For consistency, in the following text we do not change and will still use the symbols X²A′′, X³A′′ and a¹A′ (instead of X²A, X³A and a¹A) to describe the electronic states of G-C₂H₅O₂ and its cation. We can see that upon photoionization the structure of the two conformers, i.e. their C–O bond length or the CCOO dihedral angle, changes somewhat. This will be discussed in detail in the following paragraphs.

Fig. 3 The optimized structures of the gauche and the trans conformers of C₂H₅O₂ and their cations. The bond lengths are in Å.

The potential energy curves of the low-lying electronic states of the neutral ethyl peroxy radical (the minimal conformer G-C₂H₅O₂) and its cation have been calculated at the M06-2X/aug-cc-pVTZ level of theory, and are presented in Fig. 4.²⁶ Concretely, their structures have been re-optimized at each C–O coordinate and then their energies have been computed. It is shown that the neutral X²A′′ ground state is bound with a C–O equilibrium bond length of 1.446 Å. The cationic X³A′′ ground state of C₂H₅O₂⁺ is calculated to be a quasi-bound state with a very shallow well and a substantially elongated C–O bond at 2.270 Å, which is out of the Franck–Condon region reached by ionization transitions, as shown using a green rectangle in Fig. 4. The singlet a¹A′ first excited state of C₂H₅O₂⁺ is a bound state with a C–O bond length at 1.687 Å, close to that of the neutral X²A′′ ground state, and should be majorly populated in the photoionization process.

Fig. 4 Potential energy curves of the ethyl peroxy radical (G-C₂H₅O₂) and its cation along the C–O coordinates. The Franck–Condon (F–C) region is displayed using a green rectangle.

Theoretical calculations in particular using energy and charge distribution analyses show that the cationic X³A′′ ground state adiabatically correlates to the C₂H₅⁺(X¹A₁) + O₂(X³Σ_g⁻) dissociation limit, and the a¹A′ first excited state of C₂H₅O₂⁺ correlates to the C₂H₅⁺(X¹A₁) + O₂(a¹Δ_g) dissociation limit, not to the C₂H₅(X²A′) + O₂⁺(X²Π_g) dissociation limit.¹⁰ In addition, as shown in Fig. 4, the potential energy curve of the a¹A′ excited state crosses that of the X³A′′ ground state at ∼1.3 Å and their mutual interaction, e.g. via the spin–orbit coupling, can cause the singlet a¹A′ electronic state to pre-dissociate into C₂H₅⁺ and O₂ fragments along the potential energy curve of the triplet X³A′′ ground state. This pre-dissociation should be a major reason why the C₂H₅O₂⁺ cation is not observed and why only the C₂H₅⁺ fragment ion is observed in the mass spectra.^10,11 In the schematic potential energy diagram of Meloni et al.,¹⁰ the a³A′′ repulsive state, which also adiabatically correlates to the C₂H₅⁺(X¹A₁) + O₂(X³Σ_g⁻) dissociation limit, was proposed to cross with the a¹A′ excited state at a larger C–O coordinate and this pre-dissociation might also cause the C₂H₅O₂⁺ cation to be unstable. Presently the vertical ionization energy (VIE) of the second triplet electronic state, A³A′′, of G-C₂H₅O₂⁺ is calculated to be at 12.24 eV at the TD-M062X/aug-cc-pVTZ level of theory,²⁶ as listed in Table 1.

Table 1 The ionization energy (IE) of C₂H₅O₂ and the appearance energy (AE) of the C₂H₅⁺ and O₂ fragments in the dissociative photoionization of C₂H₅O₂

	State	Species	IE/AE (eV)	Method	Reference
a The adiabatic ionization energy. b The vertical ionization energy. c The appearance energy.
Calc.	X^3A′′	C2H5O2^+	9.58^a	QCISD(T)/∞	10
		G-C2H5O2^+	9.60^a/10.40^b	CCSD(T)-F12/aug-cc-pVTZ	This work
		T-C2H5O2^+	9.59^a/10.39^b	CCSD(T)-F12/aug-cc-pVTZ	This work
	a^1A′	C2H5O2^+	10.01^a	QCISD(T)/∞	10
		G-C2H5O2^+	10.13^a/10.42^b	CCSD(T)-F12/aug-cc-pVTZ	This work
		T-C2H5O2^+	10.13^a/10.47^b	CCSD(T)-F12/aug-cc-pVTZ	This work
	b^1A′′	G-C2H5O2^+	11.80^b	TD-M062X/aug-cc-pVTZ	This work
		T-C2H5O2^+	11.89^b	TD-M062X/aug-cc-pVTZ	This work
	A^3A′′	G-C2H5O2^+	12.24^b	TD-M062X/aug-cc-pVTZ	This work
		T-C2H5O2^+	12.18^b	TD-M062X/aug-cc-pVTZ	This work
Exp.		C2H5^+	10.0 ± 0.1^c	PIMS	10
		C2H5^+	10.08 ± 0.05^c	PIMS	This work

---

The AIEs and the VIEs of the two conformers of C₂H₅O₂, corresponding to their individual cationic X³A′′ and a¹A′ electronic states, have been calculated at the (R)CCSD(T)-F12/aug-cc-pVTZ level of theory and are listed in Table 1 too, together with the literature data.¹⁰ The calculated AIEs of the two conformers are very close, as the difference between their two cationic states is within 0.01 eV. In addition, the experimental AE of the C₂H₅⁺ and O₂ fragments is close to the calculated AIE of the a¹A′ electronic state of C₂H₅O₂⁺, in good accordance with the above results of the potential energy curves.

In order to assign the photoionization spectrum in Fig. 2(b), the Franck–Condon factors for the photoionization of C₂H₅O₂ have also been calculated at the M062X/aug-cc-pVTZ level.²⁶ Theoretical calculations show that the G-C₂H₅O₂ conformer has favorable Franck–Condon factors in the photoionization transition, from the neutral X²A′′ ground electronic state to the cationic a¹A′ first excited state, whereas the Franck–Condon factors of the T-C₂H₅O₂ conformer are calculated to be too low to be reliable. The Franck–Condon factors for the transitions to the cationic X³A′′ ground state can be neglected, and its large VIE–AIE difference (0.8 eV), as listed in Table 1, is also an indicator for the negligible ionization cross section. The simulated photoelectron spectrum (PES) of G-C₂H₅O₂ is generated from the Franck–Condon factors by convolving the stick spectrum with a Gaussian function (HWHM = 200 cm⁻¹, half width at half-maximum), shown in Fig. S2 (ESI†). The simulated photoionization spectrum is then obtained by the integration of the PES and presented as a blue dotted line in Fig. 2(b). The simulated photoionization spectrum has been energetically shifted (∼90 meV) to match the experimental data. We can see that the simulated photoionization spectrum satisfactorily agrees with the experimental data, although some differences can be observed at higher energy. One possible reason for the discrepancy is that, besides the a¹A′ electronic state, the excitations of other higher excited states of G-C₂H₅O₂⁺ such as the b¹A′′ and A³A′′ states listed in Table 1 might also have contributed to the photoionization spectrum at higher energy. Another possible reason is that the experimental photoionization spectrum is directly obtained from the signal of the C₂H₅⁺ fragment ion and thus its dissociation mechanisms might also have influences on the ion intensity, all of which need to be theoretically investigated in detail in the future.

The above favorable Franck–Condon transition can be understood from the optimized geometric change in the photoionization processes too. Previous theoretical predictions showed that the G-C₂H₅O₂ conformer is the minimum with an energy of ∼80 cm⁻¹ lower than T-C₂H₅O₂.^18–21 So under the present experimental conditions (∼298 K), both the G-C₂H₅O₂ and T-C₂H₅O₂ conformers should have already been prepared and populated in the near effusive gas beam sampling from the fast flow tube at 2 Torr, and then both of them should have contributed to the photoionization spectrum. But, as shown in Fig. 3, for the T-C₂H₅O₂ conformer, its C–O bond length of 1.446 Å at the neutral X²A′′ ground state should be increased upon ionization to 2.261 Å at the cationic X³A′′ ground state, or its CCOO dihedral angle of ±180° at the X²A′′ ground state should be changed to ±60° at the cationic a¹A′ excited state. Our theoretical calculations on the torsional potential energy curves (not presented here) also show that the vertical transition of the T-C₂H₅O₂ conformer from the neutral X²A′′ ground state is to the top of the torsional barrier of the cationic a¹A′ excited state, whereas the vertical transition of the G-C₂H₅O₂ conformer is close to the minima of the a¹A′ state. The calculated VIE difference (0.05 eV) of the a¹A′ excited state of the two conformers, as listed in Table 1, can somewhat reflect the height of the torsional barrier,³⁴ which also needs to be studied in detail in the future. These large geometric changes will result in poor Franck–Condon factors and thus will not favor their ionization transitions.

4. Conclusions

In conclusion, we have studied the VUV photoionization of C₂H₅O₂ by using a home-made photoionization TOF mass spectrometer at Hefei synchrotron complemented with theoretical calculations. The ethyl peroxy radical was produced in a microwave discharge fast flow tube via the reaction of C₂H₅ with oxygen molecules, whereby C₂H₅ was generated via the hydrogen-abstraction of ethane with fluorine atoms. The present experiments and theoretical calculations show that the C₂H₅O₂⁺ cation is not stable and dissociates to C₂H₅⁺ and O₂ fragments in the photoionization process. Thus, two kinds of C₂H₅⁺, one from the photoionization of C₂H₅ and one from the dissociative photoionization of C₂H₅O₂, have been clearly identified and confirmed in the photoionization mass spectra. The calculated potential energy curves show that the C₂H₅O₂⁺ cation is firstly prepared at the a¹A′ first excited electronic state and then pre-dissociates to the C₂H₅⁺ and O₂ fragments via the spin–orbital coupling of the a¹A′ state with the X³A′′ ground state, which is a quasi-bound state with a very shallow well and adiabatically correlates to the C₂H₅⁺ and O₂ dissociation limit.

The photoionization spectrum corresponding to C₂H₅O₂ is obtained and assigned with the aid of the Franck–Condon calculations. Of the two possible conformers for C₂H₅O₂, our present findings show that the G-C₂H₅O₂ conformer has favorable Franck–Condon factors, from the neutral X²A′′ ground state to the cationic a¹A′ first excited electronic state, whereas the contribution of the T-C₂H₅O₂ conformer to the photoionization spectrum is minor or negligible due to its large geometric changes in the photoionization process. The present work provides a prototype using synchrotron-based photoionization mass spectrometry to probe the spectroscopy and structure of peroxy radicals plus their isomers or conformers involved in complex reactions. Some other complex peroxy radicals are being investigated and will be reported in forthcoming papers.

Author contributions

Conception and design: X. T.; acquisition of data: Z. W., X. T., C. W., C. Z., J. Y., and X. G.; theoretical calculation: X. L., B. L., and W. Z.; analysis and interpretation of data: Z. W., X. L., and X. T.; writing—original draft preparation: X. T.; writing—review and editing: all authors; all authors have read and agreed to the published version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (No. 21773249, 91961123, 42120104007, U1832184), the International Partnership Program of Chinese Academy of Sciences (No. 116134KYSB20170048) and the Key Program of Research and Development of Hefei Science Center, CAS (No. 2020HSCKPRD001). C. F. is grateful to the Chinese Academy of Sciences President's International Fellowship Initiative (No. 2018VMA0055). The authors are grateful to Dr Jean-Christophe Loison at Univ. Bordeaux, for helpful discussions on the Franck–Condon factors. The authors also would like to thank the NSRL staff for running the synchrotron facility and providing beamtime.

References

1. C. K. Westbrook, Chemical kinetics of hydrocarbon ignition in practical combustion systems, Proc. Combust. Inst., 2000, 28, 1563–1577.
2. P. S. Monks, Gas-phase radical chemistry in the troposphere, Chem. Soc. Rev., 2005, 34, 376–395.
3. J. J. Orlando and G. S. Tyndall, Laboratory studies of organic peroxy radical chemistry: an overview with emphasis on recent issues of atmospheric significance, Chem. Soc. Rev., 2012, 41, 6294–6317.
4. G. Chettur and A. Snelson, Alkylperoxy and alkyl radicals. 4. Matrix IR spectra and UV photolysis of ethylperoxy and ethyl radicals, J. Phys. Chem., 1987, 91, 3483–3488.
5. G. S. Tyndall, R. A. Cox, C. Granier, R. Lesclaux and G. K. Moortgat, et al., Atmospheric chemistry of small organic peroxy radicals, J. Geophys. Res.: Atmos., 2001, 106, 12157–12182.
6. E. N. Sharp, P. Rupper and T. A. Miller, The structure and spectra of organic peroxy radicals, Phys. Chem. Chem. Phys., 2008, 10, 3955–3981.
7. D. B. Atkinson and J. W. Hudgens, Chemical kinetic studies using ultraviolet cavity ring-down spectroscopic detection: self-reaction of ethyl and ethylperoxy radicals and the reaction O₂ + C₂H₅ → C₂H₅O₂, J. Phys. Chem. A, 1997, 101, 3901–3909.
8. D. B. Atkinson and J. L. Spillman, Alkyl peroxy radical kinetics measured using near-infrared CW-cavity ring-down spectroscopy, J. Phys. Chem. A, 2002, 106, 8891–8902.
9. C. Zhang, M. Shamas, M. Assali, X. Tang and W. Zhang, et al., Absolute absorption cross section of the à < --X electronic transition of the ethyl peroxy radical and rate constant of its cross reaction with HO2, Photonics, 2021, 8, 296.
10. G. Meloni, P. Zou, S. J. Klippenstein, M. Ahmed and S. R. Leone, et al., Energy-resolved photoionization of alkylperoxy radicals and the stability of their cations, J. Am. Chem. Soc., 2006, 128, 13559–13567.
11. H. Fu, Y. Hu and E. Bernstein, Generation and detection of alkyl peroxy radicals in a supersonic jet expansion, J. Chem. Phys., 2006, 125, 014310.
12. K. Voronova, K. M. Ervin, K. G. Torma, P. Hemberger and A. Bodi, et al., Radical thermometers, thermochemistry, and photoelectron spectra: a PEPICO study of the methyl peroxy radical, J. Phys. Chem. Lett., 2018, 9, 534–539.
13. X. Tang, X. Gu, X. Lin, W. Zhang and G. A. Garcia, et al., Vacuum ultraviolet photodynamics of the methyl peroxy radical studied by double imaging photoelectron photoion coincidences, J. Chem. Phys., 2020, 152, 104301.
14. F. Qi, Combustion chemistry probed by synchrotron VUV photoionization mass spectrometry, Proc. Combust. Inst., 2013, 34, 33–63.
15. A. Bodi, P. Hemberger, D. L. Osborn and B. Sztaray, Mass-resolved isomer-selective chemical analysis with imaging photoelectron photoion coincidence spectroscopy, J. Phys. Chem. Lett., 2013, 4, 2948–2952.
16. X. Tang, X. Lin, G. A. Garcia, J.-C. Loison and C. Fittschen, et al., Threshold photoelectron spectroscopy of the methoxy radical, J. Chem. Phys., 2020, 153, 031101.
17. X. Tang, X. Lin, G. A. Garcia, J.-C. Loison and Z. Gouid, et al., Identifying isomers of peroxy radicals in the gas phase: 1-C₃H₇O₂ vs. 2-C₃H₇O₂, Chem. Commun., 2020, 56, 15525–15528.
18. A. M. Launder, J. M. Turney, J. Agarwal and H. F. Schaefer, Ethylperoxy radical: approaching spectroscopic accuracy via coupled-cluster theory, Phys. Chem. Chem. Phys., 2017, 19, 15715–15723.
19. J. C. Rienstra-Kiracofe, W. D. Allen and H. F. Schaefer, The C₂H₅ + O₂ reaction mechanism: high-level ab initio characterizations, J. Phys. Chem. A, 2000, 104, 9823–9840.
20. G. M. Just, P. Rupper, T. A. Miller and W. L. Meerts, High-resolution cavity ringdown spectroscopy of the jet-cooled ethyl peroxy radical C₂H₅O₂, J. Chem. Phys., 2009, 131, 184303.
21. P. Rupper, E. N. Sharp, G. Tarczay and T. A. Miller, Investigation of ethyl peroxy radical conformers via cavity ringdown spectroscopy of the Ã-X̃ electronic transition, J. Phys. Chem. A, 2007, 111, 832–840.
22. Z. Zhou, X. Du, J. Yang, Y. Wang and C. Li, et al., The vacuum ultraviolet beamline/endstations at NSRL dedicated to combustion research, J. Synchrotron Radiat., 2016, 23, 1035–1045.
23. Z. Wen, X. Tang, C. Wang, C. Fittschen and T. Wang, et al., A vacuum ultraviolet photoionization time-of-flight mass spectrometer with high sensitivity for study of gas-phase radical reaction in a flow tube, Int. J. Chem. Kinet., 2019, 51, 178–188.
24. Z. Wen, X. Tang, C. Fittschen, C. Zhang and T. Wang, et al., Online analysis of gas-phase radical reactions using vacuum ultraviolet lamp photoionization and time-of-flight mass spectrometry, Rev. Sci. Instrum., 2020, 91, 043201.
25. NIST Atomic Spectra Database, https://www.nist.gov/pml/atomic-spectra-database, Accessed June 6, 2021.
26. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria and M. A. Robb, et al., Gaussian 09, Revision A02, Gaussian, Inc., Wallingford CT, 2016.
27. H.-J. Werner and P. J. Knowles, MOLPRO version 2015, http://www.molpro.net.
28. J. D. DeSain, S. J. Klippenstein, J. A. Miller and C. A. Taatjes, Measurements, theory, and modeling of OH formation in ethyl+ O₂ and propyl+ O₂ reactions, J. Phys. Chem. A, 2003, 107, 4415–4427.
29. NIST Chemistry WebBook, http://webbook.nist.gov/chemistry/, Accessed June 6, 2021.
30. B. Ruscic, J. Berkowitz, L. Curtiss and J. Pople, The ethyl radical: Photoionization and theoretical studies, J. Chem. Phys., 1989, 91, 114–121.
31. B. Gans, G. A. Garcia, S. Boye-Peronne, J.-C. Loison and S. Douin, et al., Absolute photoionization cross section of the ethyl radical in the range 8-11.5 eV: Synchrotron and vacuum ultraviolet laser measurements, J. Phys. Chem. A, 2011, 115, 5387–5396.
32. R. X. Fernandes, K. Luther, G. Marowsky, M. P. Rissanen and R. Timonen, et al., Experimental and modeling study of the temperature and pressure dependence of the reaction C₂H₅+ O₂ (+ M)→ C₂H₅O₂ (+ M), J. Phys. Chem. A, 2015, 119, 7263–7269.
33. E. P. Clifford, J. T. Farrell, J. D. DeSain and C. A. Taatjes, Infrared frequency-modulation probing of product formation in alkyl + O2 reactions: I. The reaction of C₂H₅ with O₂ between 295 and 698 K, J. Phys. Chem. A, 2000, 104, 11549–11560.
34. G. E. Quelch, M. M. Gallo and H. F. Schaefer III, Aspects of the reaction mechanism of ethane combustion. Conformations of the ethylperoxy radical, J. Am. Chem. Soc., 1992, 114, 8239–8247.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1cp02655k

‡ These authors contributed equally to this work.

---