8018
J. Phys. Chem. A 2010, 114, 8018–8019 TABLE 1: Recent Experimental Determinations for the Adiabatic Electron Affinity of NO2
Comment on “On the Electron Affinity of Nitromethane (CH3NO2)”
method
James N. Bull, Robert G. A. R. Maclagan,* and Peter W. Harland
laser photodetachment (1981) ion-molecule charge transfer equilibration (1986)10 laser photoelectron spectroscopy (1988)11
Department of Chemistry, UniVersity of Canterbury, Christchurch, New Zealand ReceiVed: April 26, 2010; ReVised Manuscript ReceiVed: June 9, 2010 Nitrogen dioxide, NO2, being a stable free radical is a reactive gaseous species that is well-known as an anthropogenic pollutant; it is the major initiator of photochemical smog in the troposphere and is an ozone-depleting agent in the stratosphere.1 Recently, following the successful multireference ab initio determination of an accurate value for the adiabatic electron affinity (AEA) of nitromethane, CH3NO2,2 it has come to our attention that a similar high-level multiconfigurational convergence of theory and experiment has not been reported for NO2. The previous CH3NO2 study utilizing the N-state multiconfigurational quasi-degenerate perturbation theory of second order method, NS-MCQDPT,3 showed that for CH3NO2 electron capture is dominated by the -NO2 group, and the level of theory in that study should also apply to gas-phase NO2. There have been many previous experimental and theoretical attempts at determination of an accurate AEA, with data ranging from ≈2.8 to ≈2 eV (see refs 4-11 and references therein). Calculations have even attempted to rationalize a peroxy-isomer (OON-) to explain early experimental ion-molecule reaction electron detachment threshold values at ≈1.8 eV, which is now accepted as an experimental anomaly.4,5,12 Most experimental studies involved initial anion preparation followed by electron detachment under geometrical collisional relaxation conditions in order to differentiate the vertical electron detachment energy (VDE) from the AEA.13 In photoelectron methods, this procedure requires extrapolation to the lowest discernible spectral peak, which is difficult when photodetachment wavelengths are not optimized for high resolution studies. The most recent experimental determinations are summarized in Table 1 and are in agreement for an experimental value in the range 2.27-2.30 eV. Theoretically, the AEA assuming the Born-Oppenheimer approximation is calculated as:
AEA ) E(anion) - E(neutral) + ZPV(anion) ZPV(neutral) where ZPV are respective zero-point vibrational energies, and each energy is calculated at its respective optimum geometry. For NO2/NO2-, assuming the respective 2A1 and 1A1 ground electronic states, the B3LYP/aug-cc-pVTZ anharmonic ∆ZPV correction was calculated at 0.027 eV. The well-known multiconfigurational character of NO2 means that any accurate calculation should employ multireference type theoretical methods.14 Following the previous study on CH3NO2,2 the AEA of NO2 and VDE of NO2- are computed using the GAMESS-US computational package.15 The CASSCF reference wave func* To whom correspondence
[email protected].
should
be
addressed.
experimental value (eV) 9
E-mail:
2.275 ( 0.025 2.30 ( 0.10 2.273 ( 0.005
TABLE 2: Calculated AEA and Anion VDE for NO2a level of theory
calculated value (eV)
AEA SS-MCQDPT2(13/14,10)/aug-cc-pVTZ 7S-MCQDPT2(13/14,10)/aug-cc-pVTZ 7S-MCQDPT2(17/18,12)/aug-cc-pVTZ 7S-MCQDPT2(13/14,10)/aug-cc-pVQZ MRCISD(Q)(13/14,10)/aug-cc-pVTZ MRCISD(Q)(17/18,12)/aug-cc-pVTZ CCSD(T)/aug-cc-pVTZ experimental11
1.995 2.266 2.288 2.315 1.828 1.779 2.192 2.273 ( 0.005
Anion VDE 7S-MCQDPT2(13/14,10)/aug-cc-pVTZ 7S-MCQDPT2(17/18,12)/aug-cc-pVTZ experimental11
2.879 2.756 2.72 ( 0.01
a All multiconfigurational aug-cc-pVTZ calculations assume the SS-MCQDPT2(13/14,10)/aug-cc-pVTZ optimized geometries, whereas aug-cc-pVQZ calculations assume the SS-MCQDPT2(13/14,10)/aug-ccpVQZ optimized geometries.
tions for NO2 were selected as (13/14, 10) or (17/18, 12), where, for the former, 13/14 denotes the number of electrons in the neutral and anion, and 10 is the number of active orbitals. The first active space omits the 1s orbitals on all atoms and the 2s orbitals on oxygen atoms, as required to avoid symmetry breaking in the wave function.14 The second active space constitutes a full valence active space. Initially, the NO2 and NO2- geometries were optimized at the single-state SSMCQDPT2(13/14,10)/aug-cc-pVTZ level of theory yielding (experimental parameters in parentheses16): the ONO angle at 133.94° (133.86°); and the N-O bond length at 1.207 Å (1.194 Å). For NO2- these were calculated at 115.7° (117.5 ( 2°) and 1.276 Å (1.25 ( 0.02 Å), respectively. Geometrical optimization with the quadruple-ζ quality SS-MCQDPT2(13/14,10)/aug-ccpVQZ level of theory yields the ONO angle at 133.87° and N-O bond length at 1.203 Å, whereas for NO2- these are 116.03° and 1.271 Å, respectively. These calculations indicate close basis set geometrical convergence with the SS-MCQDPT2(13/14,10)/aug-cc-pVTZ level of theory. The AEA is expected to be reasonably different from vertical processes due to the difference in geometry of the neutral and anion. The calculated AEA and VDE are summarized in Table 2. The SS-MCQDPT2 calculations for both active spaces reveal the AEA to be underestimated at ≈2.00 eV. The 7S-MCQDPT2/ aug-cc-pVTZ calculations, consistent with CH3NO2, determine the AEA very closely, and essentially within experimental error for the (13/14, 10) active space, whereas the larger active space overestimates by ∼0.7%. The VDE is slightly overestimated by ∼6% with the (13/14, 10) active space, whereas the larger active space is in good accord. Inclusion of computationally demanding excitation perturbations higher than second-order are expected to give even closer agreement with the experimental value, however they are not currently available with this method. On the basis of trends in our calculations at slightly different
10.1021/jp103773v 2010 American Chemical Society Published on Web 07/14/2010
Comments
J. Phys. Chem. A, Vol. 114, No. 30, 2010 8019 NO2 represents a precise determination, in addition to theoretically supporting the precision of the previously reported AEA of CH3NO2. The suggested experimental data consistent with calculations in this work are summarized in Figure 1. This comment therefore provides further support to our previously published article on CH3NO2,2 from which it could be concluded that this method for calculating electron affinities could be applied to molecules similar to CH3NO2, that is, applied to molecules where the experimental electron affinities are ambiguous or not well-determined and calculations require multiconfigurational type reference wave functions for an accurate description.
Figure 1. Summary of suggested experimental valence electron affinities for NO2.
geometries, we expect that expensive geometrical optimizations with the larger active space would probably give negligible improvements to the calculated values on the milli-electronvolt scale. Single-reference CCSD(T) calculations yield the AEA slightly lower than experiment, as was found for CH3NO2.2 Interestingly, previous calculations utilizing the single-reference G2 procedure calculated the AEA at 2.344 eV, with G2 typically performing reasonably well (typical accuracy better than 0.1 eV) for other isoelectronic and small triatomics.17,18 Similarly, previous equations-of-motion calculations determined the AEA as 2.25 eV and the VDE at 2.66 eV.19 Unfortunately, when the G2 procedure is applied to CH3NO2, the AEA is overestimated by ∼45%. The MRCISD calculations including Davidson corrections appear to require higher-order excitations in order to reach convergence, which is a consistent observation with previous multireference coupled-cluster calculations at 2.15 eV.20 The small NO2 CASSCF dipole moment of ∼0.31 D would not support a dipole-bound anion.21 Several calculations were also performed on the ground state 2 A′ CH3NO2 dipole-bound anion with the EA-EOM-CCSD treatment using both the ACES II package22 and March 2010 release of GAMESS-US.15 All calculations assumed the CCSD(T)/ aug-cc-pVTZ+6sp7d optimized geometry. EA-EOM-CCSD/ aug-cc-pVDZ+6sp7d gave the value of 8.9 meV, which became 8.4 meV without the 7d diffuse functions, and when using the diffuse sp scheme of Gutsev and Bartlett,23 the value of 6.9 meV was obtained. These data indicate convergence of a theoretical value for the dipole-bound anion at ≈9 meV, therefore supporting our earlier tentative claims of a value at approximately 7-8 meV.2 As a further check, when EA-EOMCCSD is applied to CH3CN, which has a dipole moment (3.92 D) almost identical to CH3NO2 (3.94 D), the value of 13.3 meV (13.8 meV at ∆CCSD(T)+ZPV) was obtained in excellent agreement with experiment and other reported calculations.24 The above AEA calculations indicate that the laser photoelectron spectroscopy determination of the AEA for gas-phase
References and Notes (1) Crutzen, P. J.; Arnold, F. Nature 1986, 324, 651. (2) Bull, J. N.; Maclagan, R. G. A. R.; Harland, P. W. J. Phys. Chem. A 2010, 114, 3622. (3) Nakano, H. J. Chem. Phys. 1993, 99, 7983. (4) Richardson, J. H.; Stephenson, L. M.; Brauman, J. I. Chem. Phys. Lett. 1974, 25, 318. (5) Herbst, E.; Patterson, T. A.; Lineberger, W. C. J. Chem. Phys. 1974, 61, 1300. (6) Refaey, K. M. A. Int. J. Mass Spectrom. Ion Phys. 1976, 21, 21. (7) Smith, G. P.; Lee, L. C.; Cosby, P. C. J. Chem. Phys. 1979, 71, 4464. (8) Chen, E. C. M.; Wentworth, W. E. J. Phys. Chem. 1983, 87, 45. (9) Woo, S. B.; Helmy, E. M.; Mauk, P. H.; Paszek, A. P. Phys. ReV. A 1981, 24, 1380. (10) Chowdhury, S.; Heinis, T.; Grimsrud, E. P.; Kebarle, P. J. Phys. Chem. 1986, 90, 2747. (11) Ervin, K. M.; Ho, J.; Lineberger, W. C. J. Phys. Chem. 1988, 92, 5405. (12) Pearson, P. K.; Schaefer III, H. F.; Richardson, J. H.; Stephenson, L. M.; Brauman, J. I. J. Am. Chem. Soc. 1974, 96, 6778. (13) Kebarle, P.; Chowdhury, S. Chem. ReV. 1987, 87, 513. (14) Blahous III, C. P.; Yates, B. F.; Xie, Y.; Schaefer III, H. F. J. Chem. Phys. 1990, 93, 8105. (15) Schmidt, M. E.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. H.; Kosecki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput. Chem. 1993, 14, 1347. (16) Morino, Y.; Tanimoto, M.; Saito, S.; Hirota, E.; Awata, R.; Tanaka, T. J. Mol. Spectrosc. 1983, 98, 331. (17) Curtiss, L. A.; Raghavachari, K.; Trucks, G. W.; Pople, J. A. J. Chem. Phys. 1991, 94, 7221. (18) Yu, D.; Rauk, A.; Armstrong, D. A. J. Phys. Chem. 1992, 96, 6031. (19) Andersen, E.; Simons, J. J. Chem. Phys. 1977, 66, 2427. (20) Kaldor, U. Chem. Phys. Lett. 1990, 170, 17. (21) Jordan, K. D.; Wang, F. Annu. ReV. Phys. Chem. 2003, 54, 367. (22) Stanton, J. F.; Gauss, J.; Watts, J. D.; Nooijen, M.; Oliphant, N.; Perera, S. A.; Szalay, P. G.; Lauderdale, W. J.; Kucharski, S. A.; Gwaltney, S. R.; Beck, S.; Balkova´, A.; Bernholdt, D. E.; Baeck, K. K.; Rozyczko, P.; Sekino, H.; Hober, C.; Bartlett, R. J. ACES II; Quantum Theory Project, University of Florida; Integral packages included are VMOL(Almlo¨f, J.; Taylor, P. R.); VPROPS(Taylor, P.) ABACUS; (Helgaker, T.; Jensen, H. J. Aa.; Jørgensen, P.; Olsen, J.; Taylor, P. R.) (23) Gutsev, G. L.; Barlett, R. J. J. Chem. Phys. 1996, 105, 8785. (24) Skurski, P.; Gutowski, M.; Simons, J. Int. J. Quan. Chem. 2000, 80, 1024.
JP103773V