The Structure and ClâO Dissociation Energy of the ClOO Radical

Subscriber access provided by UNIV OF DURHAM

Article

The Structure and Cl-O Dissociation Energy of the ClOO Radical: Finally, The Right Answers for The Right Reason Adam S. Abbott, and Henry F. Schaefer J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.8b00394 • Publication Date (Web): 14 Feb 2018 Downloaded from http://pubs.acs.org on February 14, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry A is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

The Structure and Cl–O Dissociation Energy of the ClOO Radical: Finally, the Right Answers for the Right Reason Adam S. Abbott and Henry F. Schaefer III∗ Center for Computational Quantum Chemistry, The University of Georgia, Athens, Georgia 30602 USA E-mail: [email protected]

1

ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract The chlorine peroxy radical (ClOO) has historically been a highly problematic system for theoretical studies. In particular, the erratic ab initio predictions of the Cl–O bond length reported in the literature thus far exhibit unacceptable errors with respect to the experimental structure. In light of the widespread disagreement observed, we present a careful and systematic investigation of the ClOO geometry towards the basis set and correlation limits of single reference ab initio theory, employing the cc-pVXZ (X = D, T, Q, 5, 6) basis sets extrapolated to the complete basis set limit and coupled cluster theory up through single, double, triple, and perturbative quadruple excitations [CCSDT(Q)]. We demonstrate a considerable sensitivity of the Cl–O bond length to both electron correlation and basis set size. The CCSDT(Q)/CBS structure is found to be re (ClO) = 2.082, re (OO) = 1.208, and θe (ClOO) = 115.4◦ , in remarkable agreement with Endo’s semi-experimentally determined values re (ClO) = 2.084(1), re (OO) = 1.206(2), and θe (ClOO) = 115.4(1)◦ . Moreover, we compute a Cl–O bond dissociation energy of 4.77 kcal mol−1 , which is likewise in excellent agreement with the most recent experimental value of 4.69 ±0.10 kcal mol−1 .

2


Page 2 of 26

Page 3 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction Interest in the chlorine peroxy radical (ClOO) arises from its integral role in processes that drive the destruction of ozone in the upper atmosphere. In particular, ClOO is thought to be an important intermediate in catalytic cycles involving both the ClO self reaction 1–3 and the BrO + ClO cross reaction, 4–7 which together are thought to account for the vast majority of observed stratospheric ozone depletion at the Earth’s poles. 8 These processes evolve according to the following mechanisms:

ClO + ClO + M −−→ ClOOCl + M ClOOCl + hν −−→ Cl + ClOO ClOO + M −−→ Cl + O2 + M 2 [Cl + O3 −−→ ClO + O2 ]

Net:2 O3 −−→ 3 O2

ClO + BrO −−→ ClOO + Br ClOO + M −−→ Cl + O2 + M Cl + O3 −−→ ClO + O2 Br + O3 −−→ BrO + O2

Net:2 O3 −−→ 3 O2

The other dominant product channel of the BrO + ClO reaction leads to the symmetric isomer OClO, which is thought to further produce ClOO through photoisomerization. 9,10 A myriad of studies concerning ClOO has emerged in the wake of its realized atmospheric implications. The transient character and peculiar electronic structure of ClOO have plagued 3



experimentalists and theorists alike in regard to the accurate determination of its properties. ClOO is notably short-lived under experimental conditions, due to the high reactivity of the free radical and the relatively low dissociation energy of the Cl–O bond, which has been decisively determined to be just under 5 kcal mol−1 . 11–16 Nevertheless, numerous experimental studies have been successful in the detection and characterization of this elusive species. ClOO was first suggested as an intermediate in a flash photolysis study of Cl2 and O2 by Porter and Wright. 17 A tentative identification was made by Rochkind and Pimental, who observed an infrared (IR) spectrum of the O–O stretching vibration attributed to ClOO in the photolysis of ClClO in a nitrogen matrix. 18 Later, Arkell and Schwager observed a more complete IR spectrum of ClOO in an argon matrix, and they were the first to assign all three vibrational modes: 1441 cm−1 (O–O stretch), 407 cm−1 (bend), and 373 cm−1 (Cl– O stretch). 19 In a kinetic study using the molecular modulation technique, Johnston and coworkers measured ClOO in the gas phase for the first time, obtaining UV and IR spectra. They assigned a fundamental frequency at 1443 cm−1 to the O–O stretching vibration. 20 In a detailed IR study in various matrices, Nelander and coworkers 21 found that one of the aforementioned vibrational modes near 400 cm−1 determined by Arkell and Schwager 19 was in actuality an overtone of a lower vibrational mode near 200 cm−1 . The observed band at 192.6 cm−1 was assigned to the Cl–O stretch. 21 Shortly thereafter, M¨ uller and Willner confirmed this low energy Cl–O stretching vibration, reporting a band at 201.4 cm−1 obtained by IR spectroscopy in a neon matrix. 22 More recently, Endo and coworkers first reported high resolution gas phase spectroscopy of ClOO utilizing Fourier transform microwave spectroscopy. 23 In addition to the observed r0 structure r0 (ClO) = 2.075 ˚ A, r0 (OO)=1.227 ˚ A, and θ0 =116.4◦ , they also coupled their experimental results with ab initio computations, discussed later. The combined spectroscopic and ab initio results were used to obtain equilibrium geometric parameters re (ClO) = 2.084(1) ˚ A, re (OO)=1.206(2) ˚ A, and θe (ClOO)=115.4(1)◦ . In a separate study using cavity ring-down spectroscopy, Endo and coworkers reported accurate equilibrium constants for the formation of ClOO from Cl and

4


Page 4 of 26

Page 5 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


O2 , as well as a Cl–O bond dissociation energy of 4.69 ± 0.10 kcal mol−1 . 16 The weak bonding and open-shell electronic structure of ClOO results in its notoriously difficult theoretical characterization. A multitude of single reference, 24–26 multireference, 11,16,23,27–29 and density functional theory 26,30–36 (DFT) studies have been performed on ClOO. For all of these studies, the reported geometries fail to agree with the experimental structure given by Endo and coworkers. 23 The observed deviation from experiment in the Cl–O bond length ranges from 0.02 ˚ A to well over 0.1 ˚ A. DFT in particular seems to be an untrustworthy method for this system, and the performance of various functionals appears to be random. Even for functionals that happen to obtain reasonable geometries, one finds that the corresponding harmonic vibrational frequencies and other properties tend to be poorly predicted. As such, DFT approaches are not considered in the present study. We highlight the following computational investigations which made key strides in properly treating ClOO with theory. An important early study by Peterson and Werner used multireference configuration interaction (MRCI) to determine optimized geometries, potential energy surface features, harmonic and fundamental frequencies, dipole moments, and bond dissociation energies for ClOO and OClO. 11 The reported results agreed reasonably well with the sparse experimental data available at the time, and their suggestions and insights guided subsequent experimental work. The computational part of the studies done by Endo and coworkers, discussed earlier, also achieved fairly good results. 16,23 In the latter work, they employed MRCI singles and doubles with the Davidson correction (MRCISD+Q) combined with the aug-cc-pVXZ (X = D, T, Q, 5) basis sets, among other methods. The Cl–O bond length at the MRCISD+Q/aug-cc-pV5Z level of theory was predicted to be 2.116 ˚ A, compared to their experimentally measured r0 Cl–O bond length of 2.075 ˚ A. The harmonic and anharmonic force constants obtained with MRCISD+Q/aug-cc-pV5Z agreed well with those found by experiment, 22,23 supporting the assumption that the computed harmonic and anharmonic forces are appropriate for obtaining

5



an estimated equilibrium geometry from the experimentally determined r0 structure. To date, their suggested equilibrium geometry (re (ClO) = 2.084(1) ˚ A, re (OO)=1.206(2) ˚ A, and θe (ClOO)=115.4(1)◦ ) 23 based on combined experimental and ab initio results remains as the most reliable structure reported in the literature thus far. More recently, Martin and coworkers utilized the Wn family of methods to obtain benchmark-quality thermochemical properties for various halogen oxides. 26 Special attention was paid to quantifying the degree of multireference character in these species. In the case of ClOO, they demonstrated the importance of including nondynamical correlation through analysis of the total atomization energy (TAE) at various levels of theory. For ClOO, the total valence electron correlation contributions to the TAE was found to be far from converged for coupled cluster theory with singles, doubles and perturbative triple excitations [CCSD(T)], but mostly converged through the addition of full triple and perturbative quadruple excitations [CCSDT(Q)]. The combined higher-order correlation contributions from iterative quadruple [CCSDTQ] and quintuple [CCSDTQ5] excitations amend the TAE by less than 0.5 kcal mol−1 . They also featured a DFT performance study of these halogen oxides based on binding energy predictions. Pervasive error was observed across dozens of functionals, and ClOO was noted as particularly problematic, echoing the difficulties observed in other DFT studies. 31 Martin’s reported geometry re (ClO) = 2.032 ˚ A, re (OO)=1.208 ˚ A, and θe (ClOO)=115.4◦ (CCSD(T)/cc-pV(Q+d)Z) is notably different from that determined by Endo. 23 Finally, Varandas and coworkers 29 computed a global ground state potential energy surface (PES) of the ClO2 system using a double many-body expansion technique based on energies at MRCISD+Q extrapolated to the CBS limit. A detailed topological analysis of the surface was presented, including 10 stationary points, with ClOO as the global minimum. However, the reported structure and frequencies for ClOO are in fair agreement with experiment, which is expected from a level of theory appropriate for a global PES. A resolution of past theoretical failures for the structural prediction of ClOO is long

6


Page 6 of 26

Page 7 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


overdue. Herein, we present a rigorous theoretical study of the structure and properties of ClOO with current state-of-the-art ab initio methods. We analyze the sensitivity of the geometry of ClOO with respect to the level of theory utilized. Employing extrapolation schemes, we investigate the convergence of the ClOO geometry up to the practical limit of both correlation in coupled cluster theory and basis set size. Further, we compute the harmonic frequencies and Cl–O bond dissociation energy. The relative performance of single and multireference methods for this moderately multireference system is discussed. Results are compared to experiment and to other ab initio studies.

Computational Methods In all computations, restricted open-shell Hartree-Fock (ROHF) reference wavefunctions were employed to avoid spin contamination in the single determinant description. For correlated computations, the frozen-core approximation was used unless otherwise indicated. Electron correlation effects were computed with coupled cluster (CC) theory up through single, double, triple, and perturbative quadruple excitations [CCSDT(Q)]. 37–40 The correlation-consistent polarized valence (cc-pVXZ, X=D,T,Q,5,6) basis sets of Dunning 41 were employed to facilitate the extrapolation of energies to the complete basis set (CBS) limit. For chlorine, an extra set of tight-d functions (cc-pV(X+d)Z) was utilized due to the known deficiency in the original Dunning basis sets for second row elements. 42 For long, weak bonds such as the Cl–O bond in ClOO, it is typical to use the augmented variants of the Dunning basis sets. 43 In this study such basis sets were not used in every case, since the effect of including augmented basis functions was found to be negligible for larger basis set cardinalities, and therefore such augmented basis sets are not required for values obtained at the CBS limit (see Table 1). All non-CBS CCSD(T) geometries were computed with analytic gradients in CFOUR. 44 The CCSDT/cc-pVDZ, CCSDT/cc-pVTZ, and CCSDT(Q)/cc-pVDZ geometries were opti-

7



mized with analytic gradients in MRCC. 45 The CCSDT/cc-pVQZ and CCSDT(Q)/cc-pVTZ geometries were obtained via optimizations with numerical gradients based on energies computed in CFOUR and MRCC, respectively. Geometry optimizations based on energy extrapolations were likewise computed with numerical gradients. All numerical gradients were found through finite differences of the appropriate energies, with displacements and gradients generated by the optimizer in Psi4. 46 For optimizations based on energy extrapolation schemes, the computationally demanding levels of theory CCSDT/cc-pVXZ (X = Q, 5, 6) and CCSDT(Q)/cc-pVXZ (X = T, Q, 5, 6) were not computed explicitly. Rather, these energies were recursively approximated through basis set additivity, which for the CC level CC(n) and basis set cardinality X is defined by: CC(n)

EX

i h CC(n) CC(n−1) CC(n−1) ≈ EX−1 + EX − EX−1

This procedure is analogous to that used in the focal-point approach of Allen and coworkers. 47–49 For energy extrapolations to the CBS limit, the Hartree-Fock energies were extrapolated according to the three-point scheme of Feller, 50,51 and the correlation energies were extrapolated according to the two-point scheme of Helgaker and coworkers. 52 For extrapolated energies, the CCSD(T) energies were computed with Psi4, 46 and CCSDT and CCSDT(Q) energies were computed with MRCC 45 through the Psi4/MRCC interface. Variant B (Kállay and Gauss) for the perturbative quadruple excitations was used. 39 A more detailed description of the CBS extrapolations and a sample Psi4 input file for a CCSDT(Q)/CBS energy are given in the supplementary material. Harmonic vibrational frequencies with the CCSDT(Q)/CBS method were computed in a similar manner to the geometry optimizations, utilizing a numerical hessian found by finite differences of CCSDT(Q)/CBS energies. For the Cl–O bond dissociation energy, the geometries of ClOO and O2 were optimized with CCSDT(Q)/CBS. Electronic energies of ClOO, Cl, and O2 were computed at the same level of theory. In order to obtain chemical accuracy,

8


Page 8 of 26

Page 9 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


several corrections were added to these electronic energies. First, the harmonic zero point vibrational energies derived from CCSDT(Q)/CBS hessians were included. Second, to account for the energy contribution of core electron excitations, a core correction was computed through the difference in energy between the all-electron CCSD(T)/cc-pCVQZ energy and frozen-core CCSD(T)/cc-pCVQZ energy. Finally, scalar relativistic corrections were computed using second-order direct perturbation theory [DPT(2)] of the Dirac equation 53–56 in CFOUR 44 at the CCSD(T)/cc-pCVQZ level of theory, correlating all electrons.

Results and Discussion A comprehensive analysis of the ClOO geometry for various methods and basis sets has been performed. Table 1 shows the peculiar sensitivity of the Cl–O bond length with respect to level of theory. For CCSD(T), the predicted geometry deviates further and further from the experimental geometry with increasing basis set size. This undesirable basis set dependence vanishes with the inclusion of more correlation. The bond length appears to converge to within reasonable range of the experimental value only with very large basis sets at the CCSDT(Q) level of theory. We note here that this convergence of the bond length at the CCSDT(Q) level would appear even more quickly if augmented variants of the Dunning basis sets were utilized in the computations, as indicated by the trend observed in the CCSD(T) bond lengths. The CCSDT(Q)/CBS result provides remarkable agreement with experiment. In the case of the multireference computations done by Endo and coworkers 23 given in Table 2, we see an improved behavior in basis set dependence compared to CCSD(T). However, even with a full-valence active space and an extrapolation to the CBS limit with large basis sets, the Cl–O bond distance converges to a value 0.025 ˚ A longer than the experimental value. Perhaps a geometry optimized with MRCISD+Q/CBS extrapolation including a 6-ζ basis set and larger active space would agree better with experiment for the Cl–O bond. However, this would still fall short of an accurate depiction due to the inherent limit

9



of dynamic correlation and size-extensivity in the method. The small change in property predictions between the full valence active space (19 electrons in 12 orbitals) and the significantly larger active space which includes 2p electrons on Cl (25 electrons in 15 orbitals) at the MRCISD+Q/aug-cc-pVQZ level of theory suggests that the inclusion of states corresponding to core electron excitations in the reference configurations may slightly improve the results. The simple molecular orbital visualization given in Figure 1 provides an explanation as to why the performance of low-level single and multireference methods vary so drastically in describing the weak nature of the Cl–O bond. In single reference methods (i.e., CC theory), the large contribution of the antibonding electron configuration is neglected. While configurations which cover this antibonding character are incorporated in CC theory through excitations from the ground state, it is apparent from our results that sufficient treatment of these configurations is only accounted for through the addition of full triple and perturbative quadruple excitations. A similar amount of electron correlation can be achieved through increasing the active space size in the CASSCF reference for MRCISD+Q, but computations hitherto reported in the literature consider active spaces with only one or two virtual orbitals. This choice to limit the virtual space for CASSCF configurations is more computationally feasible, but presumably neglects some low-lying configurations that are important yet not accessed through merely single and double excitations in MRCI. In the case of CC theory, excitations to the virtual space are more complete. It seems reasonable to conclude from the results of Table 2 that in this case, a robust flexibility in virtual space configurations is more important than the static correlation provided by a CASSCF reference wavefunction. Referring to the results of Tables 1 and 2, CCSDT(Q) in fact outperforms the MRCI computations done by Endo et al. for quadruple-ζ and quintuple-ζ basis sets, as well as the CBS limit, despite the neglect of augmented functions in our highest level computations, which would presumably lower the error in the CCSDT(Q) geometrical parameters even further. The inclusion of higher-order correlation, relativistic, and core contributions in the geometry

10


Page 10 of 26

Page 11 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


optimization might result in a slight alteration of the Cl–O bond length, but such expensive additions were outside the scope of this work. Table 3 contains a summary of the vibrational frequencies of ClOO obtained by theory and experiment. Overall, we see reasonable agreement between the ab initio computations and experimental observations. The experimentally derived harmonic frequency of the O–O stretch (ω1 ) is in excellent agreement with the corresponding CCSDT(Q)/CBS harmonic frequency. This is not the case for ω2 and ω3 , and the discrepencies could be due to multiple reasons. First, it is sometimes difficult to derive harmonic vibrational frequencies from experimental vibrational results. The experimentally derived ω2 found by M¨ uller and Willner 22 overestimates the corresponding harmonic frequencies found by theoretical methods. This may be explained by the likely perturbation of ν2 =413.7 cm−1 due to a Fermi resonance with the overtone 2ν3 , reported to be 394.29 cm−1 . 22 Such a Fermi resonance would blue-shift the measured value of ν2 , as well as the corresponding estimation of ω2 . Furthermore, the band corresponding to 2ν3 would be red-shifted, resulting in a perturbed degree of anharmonicity in ν3 . Oddly, the theoretical predictions of ω2 and ω3 are very close to the observed fundamentals ν2 and ν3 which, if taken as correct, would imply these modes exhibit minimal anharmonicity. On the other hand, the anharmonic corrections by Peterson and Werner 11 for ν2 and ν3 are each roughly 12 wavenumbers. It is also unclear whether or not matrix isolation is affecting the experimentally derived vibrational frequencies to any significant degree. In the case of the O–O stretching vibration, Johnston and coworkers reported a gas phase measurement of 1443 cm−1 , 20 which is very similar to the values obtained in various matrices. However, the bending and Cl–O stretching vibrations vary by up to 10 wavenumbers depending on the matrix used, and no gas phase measurements for these modes have been reported. A complete gas phase study of the vibrational modes of ClOO is still needed, though such measurements may be incredibly difficult to obtain due to the weak Cl–O bond dissocation energy. Of course, it is critical to examine whether our approach is truly getting the “right answer

11



for the right reason.” First, we are convinced that the convergent trend observed in Table 1 shows that the the CCSDT(Q)/CBS geometry is near the reliable limit of CC theory. The results of the previously mentioned investigation of correlation contributions to the TAE by Martin and coworkers 26 also suggest that higher-order excitations would only minimally augment the relative energies and therefore the equilibrium geometry, which Martin et al. did not predict beyond CCSD(T). Second, the accuracy of other computed properties further supports the validity of this method. In Table 3, the harmonic frequencies obtained by the CCSDT(Q)/CBS level of theory are in good agreement with the literature. The decisively accurate determination of the Cl–O bond dissociation energy given in Table 2 further shows the impressive predictive capacity of our approach. Finally, we note that extrapolation schemes to some variety of CCSDT(Q)/CBS have a successful history in the literature, though only used recently. In a joint theoretical and experimental study of the HO3 radical, Allen, Douberly, and coworkers achieved similarly remarkable agreement with experiment for geometrical parameters and other properties at the CCSDT(Q)/CBS level of theory. 57 Our group has also recently applied this approach to determining the geometries and frequencies of the methylene amidogen radical 58 and phosgene 59 with impressive results. We also note here that the error in the presently computed CCSDT(Q)/CBS bond length and harmonic frequency of ground state diatomic oxygen compared to experiment 60 is merely 0.0007 ˚ A and 0.1 cm−1 , respectively (see supplementary material). However, no comprehensive benchmark study of CCSDT(Q)/CBS for property predictions currently exists. As for the computational expense of our approach, we find the accuracy to cost ratio to be quite promising. In our experience, the energies at CCSDT and CCSDT(Q) required for a full CCSDT(Q)/CBS energy extrapolation require roughly the same amount of time as CCSD(T) with a 5 or 6-ζ basis set for small systems. Therefore, if one is already obtaining CCSD(T) numerical gradients and hessians within reasonable times with large basis sets, we suggest considering the addition of the energies required for a full CCSDT(Q)/CBS extrapolation. This would typically only double or at most triple the computational time,

12


Page 12 of 26

Page 13 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


but presumably lead to a better description of challenging systems.

Conclusions Failed predictions of the structure of ClOO have become an embarrassment for theoretical chemistry. The geometry and select properties of ClOO have been reported in the limit of nonrelativisitic ab initio theory within the Born-Oppenheimer approximation. We report an unusually sporadic sensitivity of the ClOO geometry, in particular the Cl–O bond, to the level of higher-order dynamical correlation and basis set size. Despite indications of multireference character, the properties of ClOO as determined by the CCSDT(Q)/CBS potential energy surface are of the highest accuracy reported thus far. Indeed, it appears that in the case of a moderately multireference system such as ClOO, the inclusion of ample dynamic correlation can supercede the performance of explicit multireference treatments. The use of extrapolation schemes facilitates implementation of this high level of theory within reasonable computational costs. In light of the multifaceted agreement with experiment observed, we conclude that the quantities reported by the CCSDT(Q)/CBS level of theory are reliable, and such a method may be appropriate for other highly challenging systems.

13



Page 14 of 26

Table 1: Theoretical Cl–O bond lengths of ClOO: dependence on both basis set and treatment of electron correlationa

TZ QZ 5Z 6Z cc-pV(D+d)Z cc-pV(T+d)Z cc-pV(Q+d)Z cc-pV(5+d)Z cc-pV(6+d)Z CBS Experimentalb a

b

CCSD(T) cc-pVXZ 2.068 2.037 2.018 2.014 CCSD(T) 2.230 2.062 2.032 2.018 2.014 2.011

CCSD(T) cc-pV(X+d)Z 2.062 2.032 2.018 2.014 CCSDT 2.281 2.076 2.042 [2.032] [2.028] 2.024

CCSD(T) aug-cc-pV(X+d)Z 2.039 2.022 2.016 2.013 CCSDT(Q) 2.469 2.190 [2.112] [2.095] [2.087] 2.082 2.084

Quantities in brackets indicate that the energies for the numerical gradients were obtained through an additivity scheme. See text. Suma, Sumiyoshi, and Endo. 23

14


Page 15 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 2: Optimized geometries and dissociation energies of ClOO at various levels of theory.

MRCISD+Qa MRCISD+Qb MRCISD+Qb MRCISD+Qb CCSD(T)c CCSDT(Q)c Experiment a b c d e

re (ClO) ˚ A aug-cc-pV(Q+d)Z 2.120 aug-cc-pV(Q+d)Z 2.127 aug-cc-pV(5+d)Z 2.116 CBS Limit 2.109

re (OO) ˚ A 1.207 1.207 1.207 1.205

θe (ClOO) deg. 115.6 115.6 115.5 115.5

D0 kcal mol−1 3.73 3.47 3.68 4.26

aug-cc-pV(Q+d)Z 2.022 CBS Limit 2.082

1.209 1.208

115.3 115.4

– 4.77

1.206(2)d

115.4(1)d

4.69 ± 0.10e

2.084(1)d

Active space of 25 electrons in 15 valence orbitals, see reference 23. Active space of 19 electrons in 12 valence orbitals, see reference 23. This research. Suma, Sumiyoshi, and Endo. 23 Endo and coworkers. 16

15



Figure 1: Valence natural orbitals and occupations of the two most important CASSCF configurations, computed here with a full valence active space in Molpro. 61 Orbital visualization displayed using wxMacMolPlt. 62 The ground state configuration (1) has a contribution to the CASSCF wavefunction of C 2 = 0.79, and the antibonding configuration (2) has coefficient C 2 = 0.11. The squares of all other configuration coefficients are less than 0.01.

16


Page 16 of 26

Page 17 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 3: Harmonic (ω) and fundamental (ν) vibrational frequencies of ClOO (cm−1 )

MRCI/3d2f a MRCISD+Q/aug-cc-pV5Z MRCISD+Q/CBS b CCSDT(Q)/CBS c Experiment d,e MRCI/3d2f a Experiment d Experiment f Experiment g Experiment h a b c d e

f g h

b

O–O str ω1 1538.3 1499.1 1498.0 1471.2 1477.8 ν1 1505.6 1438.6 1442.8 1436.3 1443

bend ω2 403.2 412.5 422.0 414.5 432.4 ν2 390.8 413.7 408.3 – –

Cl–O str ω3 193.6 198.0 203.0 193.0 214.8 ν3 181.2 201.4 192.4 202.8 –

See Peterson and Werner. 11 Suma, Sumiyoshi, and Endo. 23 This research. IR spectroscopy in neon matrix. See M¨ uller and Willner. 22 Reported experimental “harmonic” values based on the approximate removal of spectroscopically observed anharmonicity. See M¨ uller and Willner. 22 IR spectroscopy in argon matrix. See Johnsson, Engadahl, and Nelander. 21 IR spectroscopy in oxygen matrix. See Johnsson, Engadahl, and Nelander. 21 IR spectroscopy, gas phase. See Johnston, Morris, and Van den Bogaerde. 20

17



Acknowledgments We gratefully acknowledge the helpful discussions, comments, and suggestions by B. Zhang, K. B. Moore III, P. R. Hoobler, and A. V. Copan. This research was supported by the U.S. National Science Foundation, Grant CHE-1661604.

Supporting Information Available The supporting information contains further details on energy extrapolations and the dissociation energy computation, a complete list of geometries, and a sample Psi4 input file for CCSDT(Q)/CBS energies.

References (1) Hayman, G. D.; Davies, J. M.; Cox, R. A. Kinetics of the Reaction ClO + ClO: Products and its Potential Relevance to Antarctic Ozone. Geophys. Res. Lett. 1986, 13, 1347– 1350. (2) Molina, L. T.; Molina, M. J. Production of Chlorine Oxide (Cl2 O2 ) from the SelfReaction of the Chlorine Oxide (ClO) Radical. J. Phys. Chem. 1987, 91, 433–436. (3) Sander, S. P.; Friedl, R. R.; Yung, Y. L. Rate of Formation of the ClO Dimer in the Polar Stratosphere: Implications for Ozone Loss. Science 1989, 245, 1095–1098. (4) Yung, Y. L.; Pinto, J. P.; Watson, R. T.; Sander, S. P. Atmospheric Bromine and Ozone Perturbations in the Lower Stratosphere. J. Atmospheric Sci. 1980, 37, 339–353. (5) Solomon, S. Progress Towards a Quantitative Understanding of Antarctic Ozone Depletion. Nature 1990, 347, 347–354.

18


Page 18 of 26

Page 19 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(6) McElroy, M. B.; Salawitch, R. J.; Wofsy, S. C.; Logan, J. A. Reductions of Antarctic Ozone Due to Synergistic Interactions of Chlorine and Bromine. Nature 1986, 321, 759–762. (7) Hills, A. J.; Cicerone, R. J.; Calvert, J. G.; Birks, J. W. Kinetics of the BrO + ClO Reaction and Implications for Stratospheric Ozone. Nature 1987, 328, 405–408. (8) Dameris, M., and S. Godin-Beekmann (Lead Authors), S. Alexander, P. Braesicke, M. Chipperfield, A.T.J. de Laat, Y. Orsolini, M. Rex, and M.L. Santee, Update on Polar Ozone: Past, Present, and Future, Chapter 3 in Scientific Assessment of Ozone Depletion: 2014, Global Ozone Research and Monitoring Project Report No. 55, World Meteorological Organization, Geneva, Switzerland, 2014. (9) Gole, J. L. Photochemical Isomerization of Chlorine Dioxide and the Low-Lying Electronic State of the Asymmetric Isomer: Possible Implications for Matrix Isolation Spectroscopy. J. Phys. Chem. 1980, 84, 1333–1340. (10) Vaida, V.; Solomon, S.; Richard, E. C.; R¨ uhl, E.; Jefferson, A. Photoisomerization of OClO: a Possible Mechanism for Polar Ozone Depletion. Nature 1989, 342, 405–408. (11) Peterson, K. A.; Werner, H.-J. Multireference Configuration Interaction Calculations of the Low-Lying Electronic States of ClO2 . J. Chem. Phys. 1992, 96, 8948–8961. (12) Nicovich, J. M.; Kreutter, K. D.; Shackelford, C. J.; Wine, P. H. Thermochemistry and Kinetics of the Cl + O2 Association Reaction. Chem. Phys. Lett. 1991, 179, 367–373. (13) Baer, S.; Hippler, H.; Rahn, R.; Siefke, M.; Seitzinger, N.; Troe, J. Thermodynamic and Kinetic Properties of the Reaction Cl + O2 + M = ClOO + M in the Range 160–300 K and 1–1000 bar. J. Chem. Phys. 1991, 95, 6463–6470. (14) Mauldin, R. L.; Burkholder, J. B.; Ravishankara, A. R. A Photochemical, Thermodynamic, and Kinetic Study of ClOO. J. Phys. Chem. 1992, 96, 2582–2588. 19



(15) Francisco, J. S.; Sander, S. P. Structure and Thermochemistry of ClO2 Radicals. J. Chem. Phys. 1993, 99, 2897–2901. (16) Suma, K.; Sumiyoshi, Y.; Endo, Y.; Enami, S.; Aloisio, S.; Hashimoto, S.; Kawasaki, M.; Nishida, S.; Matsumi, Y. Equilibrium Constants of the Reaction of Cl with O2 in the Formation of ClOO. J. Phys. Chem. A 2004, 108, 8096–8099. (17) Porter, G.; Wright, F. J. Studies of Free Radical Reactivity By the Methods of Flash Photolysis. Disc. Faraday Soc. 1953, 14, 23–34. (18) Rochkind, M. M.; Pimentel, G. C. Photolysis of Matrix-Isolated Dichlorine Monoxide: Infrared Spectra of ClClO and (ClO)2 . J. Chem. Phys. 1967, 46, 4481–4490. (19) Arkell, A.; Schwager, I. Matrix Infrared Study of the Chlorine Superoxide Radical. J. Am. Chem. Soc. 1967, 89, 5999–6006. (20) Johnston, H. S.; Morris, E. D.; den Bogaerde, J. V. Molecular Modulation Kinetic Spectrometry: ClOO and ClO2 Radicals in the Photolysis of Chlorine in Oxygen. J. Am. Chem. Soc. 1969, 91, 7712–7727. (21) Johnsson, K.; Engdahl, A.; Nelander, B. A Matrix Isolation Study of the Chlorine Superoxide Radical. J. Phys. Chem. 1993, 97, 9603–9606. (22) M¨ uller, H. S. P.; Willner, H. Vibrational and Electronic Spectra of Chlorine Dioxide, OClO, and Chlorine Superoxide, ClOO, Isolated in Cryogenic Matrices. J. Phys. Chem. 1993, 97, 10589–10598. (23) Suma, K.; Sumiyoshi, Y.; Endo, Y. Fourier Transform Microwave Spectroscopy and Fourier Transform Microwave–Millimeter Wave Double Resonance Spectroscopy of the ClOO Radical. J. Chem. Phys. 2004, 121, 8351–8359.

20


Page 20 of 26

Page 21 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(24) Grant, D. J.; Garner, E. B.; Matus, M. H.; Nguyen, M. T.; Peterson, K. A.; Francisco, J. S.; Dixon, D. A. Thermodynamic Properties of the XO2 , X2 O, XYO, X2 O2 , and XYO2 (X, Y= Cl, Br, and I) Isomers. J. Phys. Chem. A 2010, 114, 4254–4265. (25) Fernández, B.; Christiansen, O.; Jørgensen, P.; Byberg, J.; Gauss, J.; Ruud, K. Hyperfine and Nuclear Quadrupole Coupling in Chlorine and Fluorine Dioxides. J. Chem. Phys. 1997, 106, 1847–1855. (26) Karton, A.; Parthiban, S.; Martin, J. M. Post-CCSD(T) ab initio Thermochemistry of Halogen Oxides and Related Hydrides XOX, XOOX, HOX, XOn , and HXOn (X = F, Cl), and Evaluation of DFT Methods for These Systems. J. Phys. Chem. A 2009, 113, 4802–4816. (27) Jafri, J. A.; Lengsfield, B. H.; Bauschlicher Jr, C. W.; Phillips, D. H. The Lower Electronic States of ClOO: a Computational Investigation. J. Chem. Phys. 1985, 83, 1693–1701. (28) Grein, F. Theoretical Studies on ClOO: Electronic Spectra, Ionization Potential, and Electron Affinity. Can. J. Phys. 2011, 89, 891–897. (29) Teixeira, O. B. M.; Mota, V. C.; Garcia de la Vega, J. M.; Varandas, A. J. C. SingleSheeted Double Many-Body Expansion Potential Energy Surface for Ground-State ClO2 . J. Phys. Chem. A 2014, 118, 4851–4862. (30) Filatov, M.; Cremer, D. Bonding in the ClOO (2 A”) and BrOO (2 A”) Radical: Nonrelativistic Single-Reference versus Relativistic Multi-Reference Descriptions in Density Functional Theory. Phys. Chem. Chem. Phys. 2003, 5, 2320–2326. (31) Alcami, M.; Mó, O.; Yanez, M.; Cooper, I. The Performance of Density Functional Theory in Challenging Cases: Halogen Oxides. J. Chem. Phys. 2000, 112, 6131–6140.

21



(32) Beltrán, A.; Andrés, J.; Noury, S.; Silvi, B. Structure and Bonding of Chlorine Oxides and Peroxides: ClOx , ClO− x (x= 1–4), and Cl2 Ox (x= 1–8). J. Phys. Chem. A 1999, 103, 3078–3088. (33) Li, Q.; L¨ u, S.; Xu, W.; Xie, Y.; Schaefer, H. F. Molecular Structures and Electron Affinities for the Chlorine Oxides ClOO, ClOOO, and ClO3 (C3v ). J. Phys. Chem. A 2002, 106, 12324–12330. (34) Zhu, R. S.; Lin, M. C. Ab initio Studies of ClOx Reactions. VIII: Isomerization and Decomposition of ClO2 Radicals and Related Bimolecular Processes. J. Chem. Phys. 2003, 119, 2075–2082. (35) Fängström, T.; Edvardsson, D.; Ericsson, M.; Lunell, S.; Enkvist, C. Density Functional Study of Chlorine–Oxygen Compounds Related to the ClO Self-Reaction. Int. J. Quantum Chem. 1998, 66, 203–217. (36) Morris, V. R.; Bhatia, S. C.; Dibble, T. S.; Francisco, J. S. Evaluating the Accuracy of Density Functional Methods for ClOO. J. Chem. Phys. 1996, 104 . (37) Kállay, M.; Surján, P. R. Higher Excitations in Coupled-Cluster Theory. J. Chem. Phys. 2001, 115, 2945–2954. (38) Kállay, M.; Gauss, J. Approximate Treatment of Higher Excitations in Coupled-Cluster Theory. J. Chem. Phys. 2005, 123, 214105. (39) Kállay, M.; Gauss, J. Approximate Treatment of Higher Excitations in Coupled-Cluster Theory. II. Extension to General Single-Determinant Reference Functions and Improved Approaches for the Canonical Hartree–Fock Case. J. Chem. Phys. 2008, 129, 144101. (40) Kállay, M.; Gauss, J. Coupled-Cluster Methods Including Noninterative Corrections for Quadruple Excitations. J. Chem. Phys. 2005, 123, 054101.

22


Page 22 of 26

Page 23 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(41) Dunning, T. H. Gaussian Basis Sets for Use in Correlated Molecular Calculations. I. The Atoms Boron Through Neon and Hydrogen. J. Chem. Phys. 1989, 90, 1007–1023. (42) Dunning, T. H.; Peterson, K. A.; Wilson, A. K. Gaussian Basis Sets for Use in Correlated Molecular Calculations. X. The Atoms Aluminum Through Argon Revisited. J. Chem. Phys. 2001, 114, 9244–9253. (43) Kendall, R. A.; Dunning, T. H.; Harrison, R. J. Electron Affinities of the First-Row Atoms Revisited. Systematic Basis Sets and Wave Functions. J. Chem. Phys. 1992, 96, 6796–6806. (44) CFOUR, a quantum chemical program package written by J.F. Stanton, J. Gauss, L. Cheng, M.E. Harding, D.A. Matthews, and P.G. Szalay with contributions from A.A. Auer, R.J. Bartlett, U. Benedikt, C. Berger, D.E. Bernholdt, Y.J. Bomble, O. Christiansen, F. Engel, R. Faber, M. Heckert, O. Heun, C. Huber, T.-C. Jagau, D. Jonsson, J. Jusélius, K. Klein, W.J. Lauderdale, F. Lipparini, T. Metzroth, L.A. M¨ uck, D.P. ONeill, D.R. Price, E. Prochnow, C. Puzzarini, K. Ruud, F. Schiffmann, W. Schwalbach, C. Simmons, S. Stopkowicz, A. Tajti, J. Vázquez, F. Wang, J.D. Watts and the integral packages MOLECULE (J. Almlöf and P.R. Taylor), PROPS (P.R. Taylor), ABACUS (T. Helgaker, H.J. Aa. Jensen, P. Jørgensen, and J. Olsen), and ECP routines by A. V. Mitin and C. van W¨ ullen. For the current version, see http://www.cfour.de. (45) MRCC, A Quantum Chemical Program Suite Written by M. Kállay, Z. Rolik, J. Csontos, P. Nagy, G. Samu, D. Mester, J. Csóka, I. Ladjánszki, L. Szegedy, B. Ladóczki, K. Petrov, M. Farkas, and B. Hégely. See also Z. Rolik, L. Szegedy, I. Ladjánszki, B. Ladóczki, and M. Kállay, J. Chem. Phys. 139, 094105 (2013), as well as: www.mrcc.hu. (46) Parrish, R. M.; Burns, L. A.; Smith, D. G.; Simmonett, A. C.; DePrince, A. E.; Hohenstein, E. G.; Bozkaya, U.; Sokolov, A. Y.; Di Remigio, R.; Richard, R. M. et al. Psi4 1.1: An Open-Source Electronic Structure Program Emphasizing Automation,

23



Advanced Libraries, and Interoperability. J. Chem. Theory Comput. 2017, 13, 3185– 3197. (47) East, A. L. L.; Allen, W. D. The Heat of Formation of NCO. J. Chem. Phys. 1993, 99, 4638–4650. (48) Császár, A. G.; Allen, W. D.; Schaefer, H. F. In Pursuit of the ab initio Limit for Conformational Energy Prototypes. J. Chem. Phys. 1998, 108, 1007–11599. (49) Schuurman, M. S.; Muir, S. R.; Allen, W. D.; Schaefer, H. F. Toward Subchemical Accuracy in Computational Thermochemistry: Focal Point Analysis of the Heat of Formation of NCO and [H,N,C,O] Isomers. J. Chem. Phys. 2004, 120, 11586–11599. (50) Feller, D. Application of Systematic Sequences of Wave Functions to the Water Dimer. J. Chem. Phys. 1992, 96, 6104–6114. (51) Feller, D. The Use of Systematic Sequences of Wave Functions for Estimating the Complete Basis Set, Full Configuration Interaction Limit in Water. J. Chem. Phys. 1993, 98, 7059–7071. (52) Helgaker, T.; Klopper, W.; Koch, H.; Noga, J. Basis Set Convergence of Correlated Calculations on Water. J. Chem. Phys. 1997, 106, 9639–9646. (53) Kutzelnigg, W. Perturbation Theory of Relativistic Corrections. 1. The Non-relativistic Limit of the Dirac Equation and a Direct Perturbation Expansion. Z. Phys. D 1989, 11, 15–28. (54) Kutzelnigg, W. Perturbation Theory of Relativistic Corrections. Z. Phys. D 1990, 15, 27–50. (55) Klopper, W. Simple Recipe for Implementing Computation of First-Order Relativistic Corrections to Electron Correlation Energies in Framework of Direct Perturbation Theory. J. Comput. Chem. 1997, 18, 20–27. 24


Page 24 of 26

Page 25 of 26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(56) Tarczay, G.; Császár, A. G.; Klopper, W.; Quiney, H. M. Anatomy of Relativistic Energy Corrections in Light Molecular Systems. Mol. Phys. 2001, 99, 1769–1794. (57) Liang, T.; Magers, D. B.; Raston, P. L.; Allen, W. D.; Douberly, G. E. Dipole Moment of the HOOO Radical: Resolution of a Structural Enigma. J. Phys. Chem. Lett. 2013, 4, 3584–3589. (58) Wiens, A. E.; Copan, A. V.; Rossomme, E. C.; Aroeira, G. J. R.; Bernstein, O. M.; Agarwal, J.; Schaefer, H. F. Reinterpreting the Infrared Spectrum of H + HCN: Methylene Amidogen Radical and its Coproducts. J. Chem. Phys. 2018, 148, 014305. (59) Murphy, K. V.; Schaefer, H. F.; Agarwal, J. Phosgene at the Complete Basis Set Limit of CCSDT(Q): Molecular Structure and Rovibrational Analysis. Chem. Phys. Lett. 2017, 683, 12–17. (60) Herzberg, G. Molecular Spectra and Molecular Structure; Read Books Ltd, 2013; Vol. 1. (61) Werner, H.-J.; Knowles, P. J.; Knizia, G.; Manby, F. R.; Sch¨ utz, M.; Celani, P.; Korona, T.; Lindh, R.; Mitrushenkov, A.; Rauhut, G. et al. MOLPRO, version 2010.1, a Package of Ab Initio Programs. 2010; see http://www.molpro.net. (62) Bode, B. M.; Gordon, M. S. MacMolPlt: A Graphical User Interface for GAMESS. J. Mol. Graph. Model. 1998, 16, 133–138.

25



Graphical TOC Entry

26


Page 26 of 26

The Structure and ClâO Dissociation Energy of the ClOO Radical

Recommend Documents

The Structure and ClâO Dissociation Energy of the ClOO Radical