Should bromoform absorb at wavelengths longer than 300 nm?

Transcription

1 JOURNAL OF CHEMICAL PHYSICS VOLUME 117, NUMBER 13 1 OCTOBER 2002 Should bromoform absorb at wavelengths longer than 300 nm? Kirk A. Peterson a) Department of Chemistry, Washington State University, Pullman, Washington and the William R. Wiley Environmental Molecular Science Laboratory, Pacific Northwest National Laboratory, Richland, Washington Joseph S. Francisco Department of Chemistry and Department of Earth and Atmospheric Sciences, Purdue University, West Lafayette, Indiana Received 6 March 2002; accepted 2 July 2002 A theoretical study of the low-lying singlet and triplet electronic states of CHBr 3 is presented. Calculations of excitation energies and oscillator strengths are presented using excited state coupled cluster response methods, as well as the complete active space self-consistent field method with the full Breit Pauli spin orbit operator. These calculations predict that for CHBr 3 there is only one singlet state, the à 1 A 2, that is accessible by wavelengths longer than 234 nm. It is, however, predicted to be very weakly absorbing, but may strongly overlap at shorter wavelengths with a higher B 1 E state, which itself is predicted to be strongly absorbing. There is one triplet state, the ã 3 A 2, that lies well below the à 1 A 2 state and has a predicted absorption maximum at about 270 nm. The band origin of the ã 3 A 2 is predicted to lie around 297 nm, but this C 3v symmetry minimum is calculated to be a second-order transition state, ultimately leading to dissociation to HCBr 2 Br American Institute of Physics. DOI: / I. INTRODUCTION Stratospheric chemistry of halogenated compounds is important because of their role in catalytic ozone depletion processes. 1,2 The influence of bromine and its chemistry on stratospheric ozone abundances is well understood. 3 5 Although the concentration of bromine is less than that of chlorine, bromine is more effective at catalytically destroying ozone. Garcia and Solomon 6 report that each bromine atom is roughly 100 times more effective at destroying ozone than chlorine. Bromine chemistry is also effective in perturbing ozone abundances, not only in the stratosphere, but also in the marine boundary layer. 7,8 Halons and methyl bromide are well known sources of atmospheric bromine. 9,10 Bromoform, which is primarily biogenic in origin, is another bromine source that contributes active bromine to the atmosphere. In fact, recent studies by Dvortsov et al. 11 have found that bromoform alone in the stratosphere contributes more inorganic bromine than all the combined sources of halons and methyl bromide. Studies of its reaction with tropospheric hydroxyl radicals 12 or by chlorine atoms in the marine boundary layer 13,14 suggest that bromoform should have a long atmospheric residence time. As a consequence, photodissociation may be a major removal mechanism for bromoform in the atmosphere. There have been several photodissociation studies of bromoform These studies have focused on photodissociation at 193 nm. Studies of the atmospheric impact of reactive bromine atoms on the chemistry in the marine boundary layer 7 rely on the photochemistry initiated by absorption of wavelengths longer than 300 nm. Thus, the issue for atmospheric chemical modeling is whether bromoform absorbs at wavelengths longer than 300 nm and whether II. COMPUTATIONAL METHODS For the present study, the correlation consistent basis sets of Dunning and co-workers were used Geometry optimizations of the X 1 A 1 ground state of CHBr 3 were carried out with the MOLPRO suite of programs 23 at the singles and doubles coupled cluster level of theory with perturbative triples, CCSD T, using the aug-cc-pvdz and cc-pvtz basis sets. The vertical excitation energies of the excited states were calculated at the CCSD T /cc-pvtz optimal geometry of the ground state using either the equation-ofmotion CCSD EOM-CCSD method 27 with the ACESII package 28 or the closely-related linear response CCSD LRCCSD method 29 with the DALTON program suite. 30 Both the aug-cc-pvdz basis set and cc-pvtz with diffuse s and p functions from the standard aug-cc-pvtz basis set on Br cc-pvtz sp were used in these calculations. The oscillator strengths reported in this work for the singlet states were obtained at the LRCCSD level of theory. In all cases, only the valence electrons were included in the correlation treata Electronic mail: k.peters@wsu.edu there is photochemistry at such wavelengths. There have been several studies that have measured the ultraviolet UV spectra and absorption cross-sections for bromoform. Some studies report a tail in the absorption profiles that extends to longer wavelengths beyond 300 nm up to 360 nm, while others observe no such tail. 18,19 At the longer wavelengths, the absorption is relatively weak and either impurities or absorption of bromoform on the cell windows could complicate the measurements. In this work, we have used predominantly coupled cluster response methods to study the lowlying excited states of bromoform. Our primary goal is to identify whether there are electronically excited states, either singlets or triplets, that lie at wavelengths longer than 300 nm /2002/117(13)/6103/5/$ American Institute of Physics

2 6104 J. Chem. Phys., Vol. 117, No. 13, 1 October 2002 K. A. Peterson and J. S. Francisco TABLE I. Optimized CCSD T Geometry Å and degrees for CHBr 3. Species Coordinate aug-cc-pvdz cc-pvtz Expt. a CHBr 3 X 1 A 1 ) CH CBr HCBr a Substitution structure from Ref. 35. ment frozen core approximation. It is important to note that in this work, the 3d electrons of Br are defined as core in the present calculations and were not correlated. This is the appropriate choice when using the corelation consistent basis sets for the third row main group elements. In order to assess the effects of scalar relativity on the calculated excitation energies, LRCCSD/aug-cc-pVDZ calculations were also carried out with the inclusion of the Douglas Kroll DK Hamiltonian. 31,32 In these cases, the s and p functions in the standard aug-cc-pvdz basis set were recontracted using the results from atomic DK Hartree Fock calculations. Oscillator strengths for the triplet states were obtained using the full Breit Pauli spin orbit operator with full valence complete active space self-consistent field CASSCF wave functions cf., Ref. 33 and references therein. In these CASSCF response calculations both the core and Br 3s orbitals were constrained to be doubly occupied. While all the excited states considered in this work were dominated by singly excited determinants relative to the ground state and hence CCSD is expected to perform well, additional calculations were also carried out on the singlet states with a method including a perturbational treatment of triple excitations, CCSDR III. RESULTS AND DISCUSSION A. Ground-state structure of bromoform The only experimental determination of the molecular structure of bromoform, which has C 3v symmetry, has come from the microwave spectroscopy study of Williams et al. 35 The geometric parameters were determined from an analysis of rotational transitions for several isotopic variants of CHBr 3. There have been several low-level ab initio calculations of the geometry of bromoform. These studies have employed either modest electron correlation or small basis sets. 14,36 Kambanis et al. 14 performed optimizations at the MP2/3-21 G** level of theory. The C Br bond length obtained at this level was Å, which is 0.62 Å larger than the experimentally determined value. The MP2/6-31G* level results of Paddison and Tschaikow-Roux 36 is an improvement, and their theoretical value is too large by only Å. Our results calculated at the CCSD T /cc-pvtz level of theory, presented in Table I, compare very favorably with the experimental results, with the C Br bond lengths only Å longer than the experimental value. The remaining discrepancy is consistent with the expected errors arising from the frozen core approximation and the neglect of scalar relativity. The C H bond distance is calculated to be Å at the CCSD T /cc-pvtz level of theory. The C H bond distance from the microwave studies, however, is shorter by Å with an experimental uncertainty of 0.01 Å. Thus, the CCSD T /cc-pvtz structural prediction for the C H bond is expected to be the more accurate value. B. Electronic excited states of bromoform Vertical excitation energies of the singlet and triplet states of bromoform calculated with the EOM-CCSD method are shown in Table II. The corresponding oscillator strengths are given in Table III. The four highest occupied orbitals in the ground state of bromoform 17e, 15a 1,18e, and 5a 2 ) correspond to Br lone pair orbitals, while the first 3 virtuals can be characterized as predominately C H * (16a 1 ) and C Br * (17a 1 and 19e. Nearly all of the low-lying excited singlet states considered here arise from single excitations from the Br lone pair orbitals into the C Br * orbitals, with smaller contributions from excitation into the C H *. The F 1 A 2 state arises from a Br lone pair to C H * excitation. TABLE II. Vertical excitation energies in ev for singlet and triplet states of CHBr 3. CCSD CCSDR 3 State aug-cc-pvdz aug-cc-pvdz-dk a cc-pvtz sp b aug-cc-pvdz %t 1 Ã 1 A B 1 E C 1 A D 1 E Ẽ 1 E F 1 A G 1 A ã 3 A b 3 E d 3 E ẽ 3 E f 3 A a Includes scalar relativistic effects using the Douglas-Kroll Hamiltonian. b Diffuse s and p functions on Br only.

3 J. Chem. Phys., Vol. 117, No. 13, 1 October 2002 Bromoform absorption 6105 TABLE III. Oscillator strengths a for singlet and triplet states of CHBr 3. State aug-cc-pvdz aug-cc-pvdz-dk à 1 A B 1 E 2.2e-2 2.4e-2 C 1 A 1 5.4e-5 4.5e-4 D 1 E 8.8e-3 1.0e-2 Ẽ 1 E 8.6e-4 8.8e-4 F 1 A G 1 A 1 2.1e-2 2.4e-2 ã 3 A 2 9.0e-5 b 3 E 3.9e e-3 d 3 E 9.0e-4 a Obtained at the CCSD level of theory for the singlet states and the CASSCF level for the triplet states including the full Breit Pauli spin orbit operator. The CCSD T /cc-pvtz optimized ground state geometry of Table I was used in all cases. In the triplet manifold, the states ã 3 A 2, b 3 E,... arise from single excitations between the same molecular orbitals with spin inversion. The vertical excitation energies of bromoform shown in Table II span an energy range of about 5 7 ev. The lowest electronically excited state is the à 1 A 2, with a CCSD/augcc-pVDZ vertical excitation energy of 5.1 ev. As shown in Table III, however, transitions to this state are dipole forbidden from the ground state. On the other hand, the B 1 E state is calculated at the same level of theory to have a vertical transition at 5.63 ev with a relatively strong oscillator strength. This is followed by the weak C 1 A 1 state at 5.74 ev and the strong D 1 E state at 6.17 ev. The affects of scalar relativity, basis set incompleteness, and more extensive electron correlation on the CCSD/aug-cc-pVDZ predictions for these band maxima are also shown in Table II, i.e., CCSD calculations with the Douglas Kroll Hamiltonian aug-ccpvdz-dk, CCSD calculations with the cc-pvtz sp basis set, and CCSDR 3 results, respectively. Scalar relativity is observed to decrease the calculated vertical excitation energies by only ev, which tends to be opposed by the effects of increasing the basis set size, since this is calculated to increase the excitation energies by ev. If one assumes additivity, the largest effect would be to raise the F 1 A 2 state over its CCSD/aug-cc-pVDZ value by 0.14 ev while the remaining singlet excitation energies would change by less than 0.06 ev. In the case of the triplet states shown in Table II, both scalar relativity and basis set extensions have very little effect on the vertical excitation energies; the former lowers the CCSD/aug-cc-pVDZ values by ev 0.02 for the first 3 states, 0.05 for the last two and increasing the basis set to cc-pvtz sp raises the triplet states by just ev. Hence, these two effects would appear to nearly cancel for the triplets. The last point to consider is the effect of triple excitations in the calculations using the CCSDR 3 method. As shown in Table II, all of these transitions are dominated by single excitations and hence it is expected that CCSD should yield accurate results see, for instance, Refs. 29, 37, and 38, particularly for the triplet states. Including triple excitations TABLE IV. Optimized CCSD T Geometry Å and degrees for the ã 3 A 2 state of CHBr 3. C 3v symmetry was imposed. Coordinate aug-cc-pvdz cc-pvtz CH CBr HCBr perturbatively for the singlets lowers the vertical excitation energies by ev, with the largest decreases found for the Ẽ 1 E and G 1 A 1 states. Not surprisingly, these are the states with the smallest single excitation character, albeit the CCSD %t 1 s shown in Table II are still 93% for these states. Based on these results, it is also expected that if CCSDR 3 calculations could have been carried out on the triplet states, these excitation energies would only have differed by no more than 0.1 ev from their CCSD values since the single excitation character of these states are all greater than 98%. Gillotay et al. 18 measured the UV absorption spectrum for gas-phase bromoform over the temperature range K. This spectrum displays continuous absorption from 172 to 310 nm. The spectrum shows several overlapping bands with three clear band maxima around 180, 210, and 225 nm. From 310 to 240 nm, the absorption can be characterized as a shoulder band, but there appears to be a weak maximum around 250 nm. Initial inspection of the calculated results in Table II might lead to the conclusion that these peaks correspond to transitions from the ground state to the B, C, and D states, respectively, since these transitions have calculated oscillator strengths with a ratio of roughly 50:1:20. These states show a pattern of strong weak strong absorption, which is consistent with the experimental spectral pattern. Based on the above discussion, our predicted vertical excitation energies for these states are 5.54, 5.65, and 6.07 ev, respectively 224, 219, and 204 nm. The Ẽ 1 E state, with a predicted absorption maximum at 177 nm 7.01 ev, could also contribute in this region. Certainly, our result for the D 1 E state is consistent with the 193 nm photodissociation experiments of McGivern et al., 17 where their measured anisotropy parameter was consistent with excitation to a state of E symmetry. The shoulder band system appearing from 310 to 240 nm is possibly due to weakly overlapping bands or absorption into the low-lying triplet states, which borrow intensity via spin orbit coupling with nearby singlets. The three lowest triplet states, the ã 3 A 2, b 3 E, and 1, have band maxima ranging from 268 to 241 nm, respectively. In particular, as shown in Table III the 1 state has significant oscillator strength and could account for much of the long wavelength shoulder in the bromoform spectrum. Absorption to the lowest state, the ã 3 A 2, could account for the weak tail at even longer wavelengths. Such behavior has previously been noted for NOCl 39 and HOBr. 40 C. Band origin of the lowest triplet state In order to address the question of whether there is a long wavelength tail in the bromoform spectrum, a full geometry optimization of the lowest triplet state was carried

4 6106 J. Chem. Phys., Vol. 117, No. 13, 1 October 2002 K. A. Peterson and J. S. Francisco TABLE V. Total energies and adiabatic excitation energies for the X 1 A 1 and ã 3 A 2 states of CHBr 3. Total energies hartrees Relative energies ev State Method aug-cc-pvdz cc-pvtz aug-cc-pvdz cc-pvtz X 1 A 1 CCSD T a CCSD b ã 3 A 2 CCSD T a LRCCSD b a At the CCSD T optimal C 3V geometry for the specified basis set and electronic state. The CCSD T vertical excitation energies at the CCSD T /cc-pvtz ground state geometry are 4.83 ev aug-cc-pvtz and 4.91 ev cc-pvtz. b At the CCSD T /cc-pvtz optimal C 3v geometry for each state. The CCSD/aug-cc-pVDZ vertical excitation energy from Table II is 4.64 ev. out. From these calculations, the cutoff for absorption into the excited-state manifold can be estimated. Table IV gives the optimized geometry for the ã 3 A 2 state at the CCSD T / aug-cc-pvdz and CCSD T /cc-pvtz levels of theory. These CCSD T calculations were based on restricted openshell Hartree Fock ROHF orbitals. 41,42 As discussed previously, this transition arises from excitation of the electron from a nonbonding Br lone pair into an antibonding C Br * orbital. This suggests that the C Br bond in the lowest triplet state should be elongated relative to the ground state C Br bond. It is indeed the case that in the ground state, r C Br is calculated to be Å at the CCSD T /cc-pvtz level of theory, but elongates to Å in the lowest triplet state. From the CCSD relative energy differences in Table V, the band origin for the lowest triplet state is predicted to be at around 4.18 ev or 297 nm. The CCSD T excitation energies obtained from separate calculations on each state were somewhat higher, but based on the relatively large 0.03 T 1 diagnostic see, e.g., Ref. 43 and references therein in the triplet state compared to the singlet 0.01, the linear response CCSD values were judged to be more reliable within 0.01 ev, particularly given its large %t 1 value 98%. From these results it could be inferred that, in principle, there should be little absorption resulting from bromoform at wavelengths much longer than 300 nm. On the other hand, the minimum energy structure calculated for the triplet was obtained under the constraint of C 3v symmetry. A subsequent harmonic frequency calculation at the CCSD T / aug-cc-pvdz level of theory confirmed that this is only a transition state and not a true minimum. In fact the high symmetry structure is a second-order transition state with two sets of doubly degenerate imaginary modes. One set corresponds to a Br C Br bending mode, and as might be expected, the other imaginary mode is along the CBr stretch, which leads to dissociation to CHBr 2 Br. Thus, absorption into this region of the potential surface could lead to a spectrum with a long wavelength tail such as that observed out to 360 nm in the experiments of Moortgat et al. 19 Of course, effects due to hot band absorption could also contribute to this feature. Additional experimental investigations of the spectra would be useful. ACKNOWLEDGMENTS This research was performed in the William R. Wiley Environmental Molecular Sciences Laboratory EMSL at the Pacific Northwest National Laboratory PNNL. Operation of the EMSL is funded by the Office of Biological and Environmental Research in the U.S. Department of Energy DOE. PNNL is operated by Battelle for the U.S. DOE under Contract DE-AC06-76RLO K.A.P. was supported by the Division of Chemical Sciences in the Office of Basic Energy Sciences of the U.S. DOE. 1 S. C. Wofsy, M. B. McElroy, and Y. L. Yung, Geophys. Res. Lett. 2, Y. L. Yung, J. P. Pinto, R. J. Watson, and S. P. Sander, J. Atmos. Sci. 37, M. J. Prather, M. B. McElroy, and S. C. Wofsy, Nature London 312, S. Solomon, Rev. Geophys. 26, M. B. McElroy and R. J. Salawitch, Science 243, R. R. Garcia and S. Solomon, J. Geophys. Res. 99, L. A. Barrie, J. W. Bottenheim, R. C. Schnell, P. J. Crutzen, and R. A. Rasmussen, Nature London 334, S.-M. Fan and D. J. Jacob, Nature London 359, J. H. Butler, Geophys. Res. Lett. 21, J. M. Lobert, J. H. Butler, S. A. Montzke, L. S. Geller, R. C. Myers, and J. W. Elkins, Science 267, V. L. Dvortsov, M. A. Geller, S. Solomon, S. M. Schauffler, E. L. Atlas, and D. R. Blake, Geophys. Res. Lett. 26, W. B. DeMore, J. Phys. Chem. A 100, A. A. Pszenny, W. C. Kneene, D. J. Jacob, S. Fan, J. R. Maben, M. P. Zeturo, M. Springer-Young, and J. N. Galloway, Geophys. Res. Lett. 20, K. G. Kambanis, Y. G. Laizcrau, and P. J. Pagigicnrokopoulous, J. Phys. Chem. A 101, J. P. Simons and A. J. Yarwood, Trans. Faraday Soc. 57, A. J. Marr, T. J. Sears, and P. B. Savies, J. Mol. Spectrosc. 184, W. S. McGivern, O. Sorkhabi, A. G. Suits, A. Derecskei-Kovacs, and S. W. North, J. Phys. Chem. A 104, D. A. Gillatay, B. Jenouvrier, B. Coquert, M. F. Merienne, and P. C. Simon, Planet. Space Sci. 37, G. K. Mootgat, R. Meller, and W. Schneider, in Tropospheric Chemistry of Ozone in the Polar Regions, edited by H. N. Ki and K. H. Becker Springer-Verlag, Berlin, 1993, p T. H. Dunning, Jr., J. Chem. Phys. 90, R. A. Kendall, T. H. Dunning, Jr., and R. J. Harrison, J. Chem. Phys. 96, A. K. Wilson, K. A. Peterson, D. E. Woon, and T. H. Dunning, Jr., J. Chem. Phys. 110, MOLPRO is a package of ab initio programs written by H.-J. Werner and P. J. Knowles with contributions from A. Almlöf, R. D. Amos, A. Bernhardsson et al. 24 K. Ragavachari, G. W. Trucks, J. A. Pople, and M. Head-Gordon, Chem. Phys. Lett. 157, C. Hampel, K. A. Peterson, and H.-J. Werner, Chem. Phys. Lett. 190, M. J. O. Deegan and P. J. Knowles, Chem. Phys. Lett. 227,

5 J. Chem. Phys., Vol. 117, No. 13, 1 October 2002 Bromoform absorption J. F. Stanton and R. J. Bartlett, J. Chem. Phys. 98, ACES II is a program product of the Quantum Theory Project, University of Florida. Authors: J. F. Stanton, J. Gauss, J. D. Watts et al. Integral packages included are VMOL J. Almlöf and P. R. Taylor ; VPROPS P. Taylor ; ABACUS T. Helgaker, H. J. Aa. Jensen, P. Jørgensen, J. Olsen, and P. R. Taylor. 29 O. Christiansen, H. Koch, A. Halkier, P. Jørgensen, T. Helgaker, and A. S. d. Meras, J. Chem. Phys. 105, DALTON, a molecular electronic structure program, Release , written by T. Helgaker, H. J. Aa. Jensen, P. Jørgensen et al. 31 M. Douglas and N. M. Kroll, Ann. Phys. New York 82, G. Jansen and B. A. Hess, Phys. Rev. A 39, J. Olsen and P. Jørgensen, in Modern Electronic Structure Theory, edited by D. R. Yarkony World Scientific, Singapore, 1995, Vol. 2, p O. Christiansen, H. Koch, and P. Jørgensen, J. Chem. Phys. 105, Q. Williams, J. T. Cox, and W. Gordy, J. Chem. Phys. 20, S. J. Paddison and E. Tschuikow-Roux, J. Phys. Chem. A 102, O. Christiansen, H. Koch, P. Jørgensen, and J. Olsen, Chem. Phys. Lett. 256, H. Larsen, K. Hald, J. Olsen, and P. Jørgensen, J. Chem. Phys. 115, Y. Y. Bai, C. X. Qian, L. Iwata, G. A. Segal, and H. Reisler, J. Chem. Phys. 90, M. Lock, R. J. Barnes, and A. Sinha, J. Phys. Chem. A 100, M. Rittby and R. J. Bartlett, J. Phys. Chem. 92, P. J. Knowles, C. Hampel, and H.-J. Werner, J. Chem. Phys. 99, T. J. Lee and G. E. Scuseria, in Quantum Mechanical Electronic Structure Calculations with Chemical Accuracy, edited by S. R. Langhoff Kluwer Academic Publishers, Dordrecht, 1995, p.47.