Estimated core electron binding energies and their application in the

Sergio H. D. M. Faria, João Viçozo da SilvaJr., Roberto L. A. Haiduke, Luciano N. Vidal, Pedro A. M. Vazquez, and Roy E. Bruns. The Journal of Physi...
0 downloads 0 Views 513KB Size
J. Phys. Chem. 1986,90, 6790-6793

6790

TABLE VII: Comparison of the contributioao of Various Electronic Configurations for the 32;and 3ZZ;(II) States of Ga, and AI, at Their Eouilibrium Geometries % contributions ’2;

3 q r r )

configurations

Ga2

1u:1 UtP,P, 1 4 1u,2ugiT; 1ujl U,2Ug“j 1ut1Ug2bg7,Ag

66 12.3 1.3

AI2 49 30 1.7

9

3

Rydberg

Ga2 4 76 0.6 2 4

A12 12 67 2 2

states is smaller than the corresponding splitting of the atomic states. Thus, although the lower 311uand 3Zgstates would be split O:, l,, 2,) and (Oi,lg) wy states, this splitting appears into (O;, to be small enough that it would not affect our assignment of experimentally observed bands. The spin-orbit contamination (mixing of different A-s states) in such cases is not substantially

different at various internuclear distances. Consequently, the effect of spin-orbit interaction on spectroscopic properties such as Re, T,, and we of Ga2 should be small.

Conclusion In this investigation we carried out CASSCF/FOCI calculations of 18 low-lying electronic states of Ga2. The observed electronic spectra and the ground state of Ga2 are reassigned on the basis of our calculated results. The ground state of Ga2 is found to be 311uwith a 3Z8 state only 410 cm-’ above the ground state. The absorption bonds in the region of 15 590 cm-I are assigned to A(311g). The absorption bands in the region of 29000 X(311,) 311,(11) transition, while the cm-I are assigned to X(311,) emission bands in the region of 20 000 cm-’ are assigned to 32; transition.

-

-

+

Acknowledgment. This research was supported in part by the National Science Foundation Grant No. C H E 8520556. Registry No. Ga,, 74508-24-0.

Estimated Core Electron Bkrdlrtg Energies and Their Appiicatlon in the Interpretatlon of Valence Ionlzation Potentials William L. Jolly Department of Chemistry, University of California, and the Materials and Molecular Research Division, Lawrence Berkeley Laboratory, Berkeley, California 94720 (Received: July 28, 1986)

Empirical group shifts for estimation of the core electron binding energies of compounds of 14 elements are evaluated. It is shown, for elements of the first row of the periodic table, that simple oneparameter groups shifts are inadequate for estimating core binding energies of highly positively charged atoms with polarizable ligands. In the case of carbon 1s binding energies, a method involving two parameters-an ‘electronegativity” parameter and a ’polarizability” parameter-is devised which significantly improves the accuracy of the predictions. The nitrogen 1s binding energies of the aminoboranes, NH2BH2, N(CHJ2BF2, NH(CH3)BF2,and NH2BF2,are estimated. These data, when combined with the nitrogen p~ ionization potentials, lead to the conclusion that there is an important degree of N-B 7 bonding in all of these molecules.

One-Parameter Group Shifts

It has been shown that shifts in atomic core electron binding energy can be expressed as the sum of empirical “group shifts” corresponding to the groups bonded to the atom in question.’-5 Early evaluations of group shifts for the elements carbon,’ nit r ~ g e n silicon? ,~ p h o s p h o r u ~ , and ~ . ~ arsenic3 involved both solidstate and gas-phase binding energies, some of which were known to only f0.2 or f0.3 eV. Binding energies are now known for gaseous compounds of a wide variety of elements with an accuracy which is usually fO.l eV or better. With these data, it is possible to determine group shifts which not only permit relatively accurate predictions of core binding energies but which permit study of periodic table trends in group shifts. Table I gives the group shifts, calculated from the available data:.’ for 14 groups and 14 ionizing atoms. In general, the binding energy shifts are calculated relative to the hydride; Le., (1) Gelius, U.; Heder, P. F.; Hedman, J.; Lindberg, B. J.; Manne, R.; Nordberg, R.; Nordling, C.; Siegbahn, K.Phys. Scr. 1970, 2, 70. (2) Jolly, W. L. J. Am. Chem. Soc. 1970, 92, 3260.

(3) Lindberg, B. J.; Hedman, J. Chim.Scr. 1975, 7, 155. (4) Gray, R. C.; Carver, J. C.; Hercules, D. M. J . Efectron Spectrosc. Relat. Phenom. 1916, 8, 343. (5) Hedman, J.; Klasson, M.; Lindberg, B. J.; Nordling, J. C. In Electron Spectroscopy,Shirley, D . A., Ed.; North-Holland Amsterdam, 1972; p 681. (6) Jolly, W. L.; Bomben, K. D.; Eyermann, C. J. At. Dora Nucf. Dora Tables 1984, 31, 433. (7) The EB for (CF,),C*OH was taken from Davis, D. W. Ph.D. Thesis, Lawrence Berkeley Laboratory Report LBL-1900, May 1973.

0022-3654/86/2090-6790$01 .50/0

the group shift for hydrogen is taken as zero. Thus the carbon 1s shift for CH3F is given by the relation AJ?B(CH~F) = EB(CH3F) - EB(CH4) The binding energy shift of an atom is assumed to be the sum of the group shifts of the groups bonded to the atom. For example, the silicon 2p shift for CH3SiH12is estimated from data in Table I as follows: AEB = Ssi(CH3) + 2Ssi(I) = -0.34 2 X 0.56 = 0.8 eV

+

The reliability of the method, at least for elements beyond the first row of the periodic table, can be judged from the data of Table 11, which gives the experimental and estimated shifts for 22 silicon compounds, none of which were used to evaluate group shifts. The standard deviation for this set of data is 0.11 eV. Some significant facts can be discerned in the data of Table I. The group shift for the hard (nonpolarizable) fluorine group is fairly constant at approximately 2.7 for the first-row series of atoms from carbon to fluorine. The fluorine group shift for boron (calculated from data for BF3) is markedly lower, however, probably because the fluorine atoms in BF3 are engaged in considerable F B 7-donor bonding.s For the second-row atoms, silicon, phosphorus, and chlorine, the fluorine group shift is much

-

(8) Beach, D. B.; Jolly, W. L. J . Phys. Chem. 1984, 88, 4647.

0 1986 American Chemical Society

Interpretation of Valence Ionization Potentials

The Journal of Physical Chemistry, Vol. 90, No. 26, 1986 6791

TABLE I: Group Shifts in Core Electron Bindii Energy (eV), Based on Hydrogen as the Reference Group

group

B'

Cb

N

0

ionizing atom (ligancy indicated parenthetically) F Si P S C1 Tic

(3) 2.0 1.1 0.7 0.3 -0.1

(4) 2.75 1.5 1.2 0.5 -0.2 -0.5 -0.6 -1.1 1.2 0.6

(3) 2.87

(2) 2.7

(1) 2.5

(4) 1.1 0.72 0.6 0.5 -0.34

(3) 1.6 0.9

(2)

(1) 1.8 0.4

(4) 1.2 0.4 0.15 (0.0)

Ge

Br

Sn

I

(4) 1.15 0.7 0.55 0.28 0.33

(1)

(4)

(1)

F C1 0.0 0.55 0.3 Br 0.0 0.42 -0.3 I -0.7 -1.4 0.25 -0.3 CH3 -0.32 -0.7 -1.3 -0.4 -0.7 -1.2 -1.0 -0.32 -0.9 SiH, -0.7 -0.4 -0.85 -1.2 -0.9 GeH, -1.4 -0.8 -1.7 -1.4 Ge(CH313 -2.0 -3.7 -2.4 -2.3 -2.0 CF3 1.0 1.4 0.6 1.1 0.4 0.44 -0.4 0.2 CCI, 0.1 -0.4 C6HJ 0.3 -1.3 -1.2 -1.2 OCH, 0.3 1.4 0.1 0.4 -0.1 -0.6 N(CH3)2 0.4 -1.4 -3.6 -3.1 -2.3 Mn(CO)5 -1.3 'B 1s binding energy of BH, estimated as 196.8 eV. Binding energy of B2H6 is 196.5 eV. bGroup shifts calculated from data for compounds of the type CH,X. These group shifts are not recommended for estimating carbon core binding energies. (See text.) cBinding energy shifts relative to TiIa. TABLE II: Experimental and Calculated Silicon 2p Binding Energy Shifts (eVh Usinn Parameters of Table I

compound CH,SiH,Br (CH3)2SiHBr (CH,),SiBr CH3SiHBr2 (CH,)2SiBr2 CH3SiBr3 CH3SiH2C1 CHpSiHC12 CH3SiCl3 CH,SiH2F CH3SiHF2 CH,SiH21 (CH,),SiHCI (CH3)2SiC12 (CH3)2SiHF (CH,),SiHI (CH,),SiCl (CH,),SiF (CH,),SiI (CH3),SiOC2H5 (CH,)2Si(OC2HS)2 CH3Si(OC2HS),

AEB

0.4 0.0 -0.3 1.o 0.7 1.5 0.5 1.2 1.9 0.7 1.7 0.1 0.1 0.8 0.3 -0.2 -0.3 -0.1 -0.5 -1 .o -0.6 -0.2

TABLE 111: Effect of Group Polarizability on the Constancy of Group Shifts

AEB(calcd) 0.3 0.0 -0.4 1.o 0.6 1.6 0.4 1.1 1.8 0.8 1.9 0.2 0.0 0.8 0.4 -0.2 -0.3 0.1 -0.5 -0.9 -0.5 0.0

lower (1.1-1.8 eV), probably because of (1) the greater size of the second row atoms and (2) the fact that, for a given change in atomic charge, binding energy shifts are approximately proportional to the reciprocal of atomic size.9 In the case of a polarizable ligand, there is a significant negative contribution to the core binding energy (essentially a-relaxation energy contribution) associated with the ionization-induced polarization of the ligand.lO*ll This relaxation energy would be expected to increase with increasing electronegativity of the core-ionizing atom, and indeed this trend is found in the decreasing methyl group shifts on going from left to right in the periodic table, from boron to fluorine and from silicon to chlorine. The magnitudes of the methyl group shifts for first- and second-row atoms in the same vertical family of the periodic table are remarkably similar, in contrast to the corresponding fluorine group shifts. The approximate constancy in group shift on going from first- to second-row elements is probably due to a fortuitous cancelling of the effects of the decrease in electron withdrawal and the decrease in relaxation energy. Many more group shifts than are given in Table I can be easily calculated from the available binding energy data, which are (9) Siegbahn, K. et al. ESCA. Atomic, Molecular and Solid State Structure Studied by Meum of Electron Spectroscopy: Almqvist and Wiksells: Uppsala, 1967; pp 79-82. (10) Gelius, U.; Siegbahn, K.Discuss. Faraday SOC.1972, 54, 257. (1 1) Jolly, W. L.; Bakke, A. A. J . Am. Chem. Soc. 1976, 98, 6500.

compound CH3F CH2F2 CHF, CF4 CHlCl CC14 CH,Br CBr4

carbon 1s binding energy, re1 to CH4, eV calcd group shift, eV 2.7 5.5 8.3 11.0

2.7 2.75 2.77 2.75

1.5 5.5

1.5 1.38

1.2 3.9

1.2 0.98

tabulated in ref 6. Table I gives only values for groups for which data for at least three different ionizing atoms are available.

Two-Parameter Group Shifts In the case of carbon compounds, there are enough data to show that the errors in binding energies estimated with simple oneparameter group shifts are much greater than the experimental uncertainties. The errors in the estimated values are particularly great when polarizable groups are bonded to carbon atoms with high positive charges. This effect of group polarizability can be clearly seen in Table 111, where the data for the fluorine group (essentially nonpolarizable) can be compared with those for the polarizable chlorine and bromine groups. Whereas the fluorine group shift is constant within f0.05 eV, the chlorine group shift drops 8% on going from CH3C1to CC14, and the bromine group shift drops 20% on going from CH3Br to CBr4. We believe this effect is due to an effective donation of negative charge to the carbon atom caused by the polarization of the group-a polarization proportional to the positive charge on the carbon atom. The carbon binding energies can be reproduced quite well by the following equation: where x and y are empirically determined group parameters and the summations are carried out for all of the groups bonded to the carbon atom. Values of x and y , and the compounds whose binding energies were used in the evaluation of x and y, are given in Table IV. The x parameter may be taken as a kind of electronegativity of the group, and t h e y parameter as a measure of the effective polarizability of the group. Thus, in the above equation, the term - y C x corrects the term x for the induced polarization of the groups. The importance of accounting for the polarizability of the groups and the reliability of the two-parameter method for carbon compounds can be seen from the data in Table V, where U B values estimated by both the one-parameter and two-parameter methods can be compared with experimental values, none of which

6792 The Journal of Physical Chemistry, Vol, 90, No. 26, 1986 TABLE I V x and y Parameters for the Estimation of Carbon Is Binding Energy Shifts compounds used for

group H F CI Br I OH NH2 CH3 SiH, GeH, CF, CCI, OCH, N(CH3)2

=o

4 H 2 4c12

X

Y'

0 2.75 1.542 1.275 0.543 1.53 0.74 0.0 -0.56 -0.67 1.286 0.66 1.56 0.45 3.88 -0.104 0.0 0.42 0.23 -1.226 -1.555

0 0.0 0.0260 0.059 0.0788 (0.02) (0.05) 0.07 (0.1) (0.1) 0.067 (0.09) (0.1) (0.1) 0.0722 0.042 0.073 0.055 0.1 1 0.103 0.164

parameter determn CH2F2; CF4 CH,Cl; CC14 CH,Br; CBr4 CH31; CFJ CHqOH CHiNH2 (CH,),CCl; (CH,),CO; C2H6 CH,SiH, CHiGeH, CH,CF,; C2F6 CHBCCl, (CH3)20 N(CH,), CH2O; COF2 C2H4; CH2CC12 C2CI4; CH2CCl2 CH2CF2; C2F4 CH,GeCI,; CF,GeCl, CF,Ge(CH,),; Ge(CH,)4 CH3Mn(CO)5;CF3Mn(CO)5

=C F2 GeC1, Ge(CH3), Mn(CO)5 'Parenthesized values are estimates.

TABLE V Carbon Is Binding Energy Shifts (eV) Calculated by the One-Parameter and Two-Parameter Methods

calcd AER 1-parameter AEa method"

exptl comwund CClF, CC12F2 CCIjF CBrF, CH2C12 CHCle CH,C*Cl, CFIGeHl (CFI)IC*OH CFoC*FCF2 C*HCICC12

coc12

C*O(OCH,)2 CO(NW2 NHZCHO CH$*02CH3 HCOZH

co2

9.4 8.0 6.6 8.4 3.0 4.2 4.1 6.6 4.1 3.8 1.5 5.9 5.2 3.9 3.6 4.0 4.9 6.8

9.6 8.3 6.9 9.2 3.0 4.5 3.8 7.6 5.1 4.4 1.5 6.6 6.4 5.0 4.3 4.8 5.1 7.2

2-parameter method 9.5 8.1 6.8 9.0 2.9 4.3 3.9 6.8 4.2 3.9 1.4 6.1 5.1 4.4 4.1 4.1 4.9 6.6

"Calculated by using parameters from Table I and &(OH) = 1.5; S c ( d F 2 ) = 0.4; Sc(4C12) = 0; S c ( ' 0 ) 3.6; Sc(NH2) 0.7. were used in evaluating the parameters. The standard deviations for the one- and two-parameter data of this table are 0.44 and 0.25 eV, respectively. We believe that the two-parameter method for estimating binding energy shifts should probably be used for all ionizing atoms from the first row of the periodic table. However, only in the case of carbon are there enough data to allow the method to be applied. There is some evidence that binding energy data for nitrogen and oxygen should be treated in this way. Thus the SN(CH3)values calculated from the data for CH3NH2,(CH3)2NH,and (CH3),N are -0.4, -0.35, and -0.27, respectively. Similarly the So(CH3) values calculated for CH30Hand (CH3)zO are -0.8 and -0.65, respectively. Hence an estimate of the N 1s binding energy of a compound such as (CH3)2NFobtained by using group shifts from Table I would probably be seriously in error. However, in view of the results of Table 11, it seems likely that the one-parameter method, with group shifts from Table I, can be used for elements which are not in the first row of the periodic table, as long as one is not dealing with ionizing atoms with very positive charges or groups which are extraordinarily polarizable. Consider, however, the extreme case of S ~ F , M ~ I ( C Oin) ~which , the silicon

Jolly TABLE VI: Nitrogen Is Binding Energies, Nitrogen p r Ionization Potentials, and Nitrogen LOIPs for Aminoboranes, eV EB(N N p r Np A (net stabilization

molecule NH2BH2 NHzBFz NH(CH3)BFp N(CH3)2BF2

Is)' 405.6 406.5 406.2 405.9

IP

LOIP

of N p r orbital)

11.36b 1 1 .47b3c 10.45' 9.49'

10.0, 10.7, 10.52 10.2,

1.3 0.7 -0.1 -0.8

'Estimated by the methods of this paper. See text. bReference 14. Reference 15. atom is both highly positive and bonded to the highly polarizable MII(CO)~group. The silicon AEB value for this compound, estimated by using the Ssi(Mn(CO),) value calculated from the AEB value of SiC13Mn(CO)5,is 0.5 eV higher than the experimental value. Other Estimation Methods When appropriate binding energy data for directly calculating a group shift are unavailable, it may still be possible to estimate the group shift by periodic-table extrapolation or interpolation, using either the available binding energies6 or the data of Table I. When a substituent is not directly bonded to the ionizing atom, but has another atom connecting it to the ionizing atom, advantage may be taken of the fact that the group shift is usually 15-25% of the value corresponding to direct bonding to the ionizing atom. Sometimes it is possible to estimate binding energies by utilizing an established correlation between binding energies and some other physical constant. For example, the carbon 1s and oxygen 1s binding energies of transition-metal carbonyl complexes show a close linear correlation with the corresponding C-0 stretching force constants.I2 The Interpretation of Valence Ionization Potentials The interpretation of the valence photoelectron spectrum of a molecule can be aided by a knowledge of the core binding energies of the atoms of the m o l e c ~ l e . ' ~Hence the estimation methods of this paper are useful in those relatively frequent situations in which the valence photoelectron spectrum, but not the core photoelectron spectrum, of a molecule is known. Carefully estimated core binding energies are better than none at all. We shall illustrate the use of estimation methods in the interpretation of the valence photoelectron spectra of several aminoboranes,14J5 in which the extent of N-B T bonding is a point of interest. The first ionization potentials of NH2BH2, N(CH&BF2, NH(CH3)BF2, and NH2BFzare given in Table VI. These ionization potentials are clearly identified with the p r orbitals and are mainly on the nitrogen atoms of the molecules. The main question of interest to chemists is: is the N-B u bonding important; Le. are the nitrogen p~ orbitals significantly delocalized onto the otherwise vacant boron PT orbitals? Westwood and WerstiukI4 concluded that u bonding is significant but weak in NHzBH2. Kroto and M c N a u g h t ~ nconcluded '~ that, in the aminodifluoroboranes, there is negligible u bonding and that the observed planarity of the molecules is due to repulsion between the fluorine atoms and the nitrogen pu electron pairs. Unfortunately the core binding energies of these aminoboranes have not been measured. However, the nitrogen core binding energy of diethylaminoborane, (C2Hs)2NBH2,is known, and this value, listed in Table VI, can be used with other data to estimate the nitrogen core binding energies of the other compounds. Two factors would be expected to cause the nitrogen bonding energy of (CH&NBF2 to be higher than that of (C2HS),NBH2. First is the effect of replacing the ethyl groups with methyl groups. Numerous data indicate that this substitution causes an increase (12) Avanzino, S. C.; Bakke, A. A.; Chen, H. W.; Donahue, C. J.; Jolly, W. L.;Lee, T. H.; Ricco, A. J. Inorg. Chem. 1980, 19, 1931. (13) Jolly, W. L. Ace. Chem. Res. 1983, 16, 370. (14) Westwood, N. P. C.; Werstiuk, N. H. J . Am. Chem. Sor. 1986, 108, 891. (15) Kroto, H. W.; McNaughton, D. J . Chem. SOC.,Dalton Trans. 1985, 1767.

J. Phys. Chem. 1986,90, 6793-6800 of approximately 0.3 eV. (For example, the binding energy of (CH3),NH is 0.32 eV higher than that of (C2H5)2NH.6) Second is the effect of replacing the two hydrogens on the boron atom with fluorine atoms. We assume that this causes an increase of 0.9 eV, equal to the increase in the CH3binding energy on going from CH3CH3to CH3CHF2. Summation of these effects yields a net estimated shift of 1.2 eV, corresponding to a nitrogen 1s binding energy of 405.9 eV for (CH3)2NBF2. Similar treatment of the data for the other aminoboranes, together with the SN(CH3) value of -0.32 eV from Table I, leads to the estimated binding energies for NH2BH2,NH(CH3)BF2,and NH2BF2given in Table VI. The nitrogen 1s binding energy and nitrogen p r ionization potential of the planar ammonia molecule are 405.3 and 9.8 eV, re~pective1y.l~The nitrogen p r orbital of planar ammonia is strictly nonbonding. By applying the approximation that shifts in nonbonding valence orbital ionization potentials are eight-tenths of the corresponding core binding energy shifts, we calculate the "localized orbital ionization potentials" (LOIPs) for the nitrogen p-xorbitals of the various aminoboranes given in Table VI. These are the ionization potentials that the nitrogen p?r orbitals would have if they were strictly nonbonding. The last column of Table VI gives the difference between the actual ionization potentials and the LOIP values. Positive differences correspond to net stabilization of the nitrogen p r orbitals and constitute evidence for N-B ?r bonding; negative differences correspond to net destabilization off the nitrogen p r orbitals and constitute evidence for repulsive interactions of those orbitals. It can be seen that the nitrogen p r orbital of NH2BH2is stabilized by 1.3 eV, corresponding to a significant degree of N-B ?r bonding. In NH2BF2 this stabilization is considerably reduced, yet still large enough to constitute evidence for N-B ?r bonding. The reduction in N-B T bonding from NH2BH2 is probably due to a reduction in the ?r acceptor character of the boron atom by the donation of electron

6793

density to the boron atom from the p?r orbitals of the fluorine atoms. When one of the NH, hydrogen atoms is replaced by a methyl group, repulsion between the CH3bonding orbital and the nitrogen p?r orbital further reduces the net stabilization of the nitrogen p r orbital, an effect which leads to an overall destabilization of 0.8 eV with two methyl groups. The fact that the nitrogen p r orbital undergoes overall destabilization in NH(CH3)BF2and N(CH3)2BF2should not be interpreted as evidence for the lack of N-B ?r bonding. The A values in Table VI are the net effects of the repulsive interactions of the methyl groups and the stabilizing, *-acceptor interactions of the BF2 group. Indeed the N-B ?r bonding in NH(CH3)BF2 and N(CH3)2BF2is probably greater than in NH2BF2because the repulsive effect of the methyl group makes the NH(CH3) and N(CH3)2groups better *-donors than the N H 2 group. Hence, contrary to the conclusions of Kroto and McNaughton, who were B(p?r) unwilling to interpret their data in terms of N(p?r) bonding, we conclude that there is significant N-B ?r bonding in NH2BF2and even more in NH(CH3)BF2 and N(CH3)2BF2. Although there are uncertainties of *0.2 or f0.3 eV in the estimated nitrogen 1s binding energies of Table VI, and although uncertainty in the factor of 0.8 used in the LOIP method introduces a further uncertainty of f O . l eV, we believe that the combination of these uncertainties is still small enough that the qualitative conclusions we have drawn from the A values of Table VI are correct.

-

Acknowledgment. This work was supported by the Director, Office of Basic Energy Sciences, Chemical Sciences Division, of the US.Department of Energy under Contract No. DE-AC0376SF00098. Registry No. F, 14762-94-8; C1, 22537-15-1; Br, 10097-32-2; I, 14362-44-8; Mn(CO)5, 15651-51-1; NH2BHz, 14720-35-5; NH2BF2, 50673-31-9; NH(CH,)BF,, 99646-58-9; N(CHj)IBFz, 359-1 8-2.

Two-Photon, Thermal Lensing Spectroscopy of Monosubstituted Benzenes in the

Jane K. Rice* Chemistry Department, University of Southern California, Los Angeles, California 90089-0484

and Roger W. Anderson Chemistry Board of Studies, University of California, Santa Cruz, Santa Cruz, California 95064 (Received: September 17, 1984; In Final Form: June 12, 1986)

-

-

The 'B2,('Lb) 'AIg('A) and 'BIu('La) 'AIg('A) electronic transitions in neat benzene, fluorobenzene, toluene, phenol, and aniline are examined by two-photon thermal lensing spectroscopy. The energies, integrateajntensitie, and polarization ratios (circular/linear) are given for both transitions. The 'B2,,(lLb) 'A18('A) transition shows intensity mainly through vibronic coupling, in particular through the vI4 vibration. In addition, pure electronically allowed intensity is seen in toluene and aniline. The integrated intensities vary by a factor of 2 for the molecules benzene, fluorobenzene, toluene, and phenol. Aniline has an integrated intensity of 9 times that for benzene. The 0 transition energies for phenol and aniline are lowered from those for the other molecules indicating the presence of charge-transfer contributions. However, only in aniline is the integrated intensity enhanced. The lBI,,('La) 'A18('A) intensity strongly increases with the strength of inductive substituents in the order F C OH C NH2 where the latter has an intensity 160 times that of benzene. Again the 0-0 energies for phenol and aniline indicate the presence of charge-transfer contributions. The 'BlU('La) 'A18('A) intensity increases moderately for the hyperconjugative substituent -CH3. These results display the importance of the pseudoparity designation of altemant hydrocarbons in determining two-photon transition strengths. This work is consistent with current theoretical approaches which consider explicitly or implicitly pseudoparity designations. The results also point out the necessity of considering the effect of hyperconjugativesubtituents like methyl in theoretical treatments. The results indicate that when integrated intensities are used to give two-photon intensities charge-transfer states behave like inductive perturbations.

-

Introduction The research in the nature of benzene electronic states has &n prolific, and Ziegler and Hudson' have provided a thorough review (1) Ziegler,

L. D.;Hudson, B. S. Excired Stares 1982, 5, 41.

0022-3654/86/2090-6793$01 .50/0

-

-

of experimental results. Recent experimental advances in twophoton, multiphoton, and resonance Raman spectroscopies have provoked new questions about the mechanism of intensities in unsubstituted and substituted benzene electronic bands. Experimental and theoretical workf3 have demonstrated that one-photon 0 1986 American Chemical Society