3960
A. Gamba, V.
TABLE I ! : Twist Angles Determined from the Ratios of the Ued'sa
Case 1 Case 2 Case 3
1/11
2/11
3/11
4/11
Av
67 66 Ei 6
70
68 66
65 65 65 li
68 66
67 66 65
64b
*
aAli values are in degrees These values were obtained by inverting the assignment of the two Oositlons
struct, we used three sets of values for the parameters and determined i99p4L for each set. From among the values cited in calculations on heteromolecules,ll we chose three sets which would span the commonly used range of values: case 1,12 hN = 0.84, k C N = 1.0; case 2,13,14 h N = 0.7, h C N =: 1.0; ease 3,1*,15h N = 0.5, kcrj = 1.0. The effect of the value of h N on the twist angle is small; the average t19p2,c value varies by only 2" (Table 11). With hN = 0.5 the order of the ued's a t positions 3 and 4 is inverted; however, if we also invert the assignment of the
Malatesta, 6 .ivlorosi, and M. Simonetta
coupling constants at these positions, the value of is in good agreement with that derived from the other sets of parameters. A similar inversion was noted by Pedersen and Muus14 for the quinoxaline anion radical upon changing the value ofhN from 0.7 to 0.4.
Acknowledgment. The support of the National Science Foundation (Grant No, GP-31414X) and the Robert A. Welch Foundation is gratefully a c ~ n o ~ ~ e d The g e ~ esr . spectrometer was purchased with funds provided by the ~ National Science Foundation (Grant N QGP-2090). (11) For example, C. L. Talcott and R. J. Myers, Mol. Phys., 12, 549 (1967); A. R. Buick, T. J. Kemp, G. T. Neal, and T. J. Stone, J. Chem. SOC. A, 1609 (1967); A . Carrington and J dos SantosViega, Moi. Phys., 5 , 21 (1962); 0 . R. Geske and G. R. Padmanabhan,J. Amer. Chem. SOC.,87, 1651 (1965). (12) B. J. Tabner and J. R. Yandle,J. Chem. SOC.A . 381 (1968) (13) J. C. M. Henning, J. Chem. Phys., 44, 2139 (1966). (14) J, A . Pedersen and L. T. Muus, Moi. Phys , 16, 589 (1969). (15) R. Zahradnik and J. Koutecky, Advan. Hsterocycb Chem., 5, 72 (1965).
lectron Spin Resonance Spectra of the Anion yl- and Diphenylethylenes amba, V. Malatesta, G . Morosi, and M. Simonetta" C.N.R.Center for the Study of the Relationship between Structure and Chemicai Reactivity and The Institute of Physical Chemistry, University of Milan, 20733 Milan, Italy
(Recieveddune 27, 1972)
Pubiication costs assisted by the Physical Chemistry Institute of the University of Milan
A number of previously unknown esr and uv spectra have been observed for aryl- and diarylethylene anion $radicalsgenerated by IG (internal generation) and EG (external generation) electsolytical reductions in dimethyl sulfoxide and acetonitrile, respectively. Semiempirical self-consistent field calculations combined with limited configuration interaction (LCI-SCF) according to Longuet-Higgins and Pople and Roothaan methods, allow a satisfactory interpretation of the electronic spectra of the anion radicals. Also uv spectra of parent neutral molecules have been taken into consideration and discussed. Spin densities, obtained by the aforementioned and McLachlan's methods, have been translated into coupling constants through McConnell's relationship and compared with experimental data. These three methods allow an unequivocal assignment of proton hyperfine splitting (hfs) constants for all considered anion radicals.
Introduction While the esr spectra of ion radicals have been the subject of extensive experimental and theoretical investigationsl only scanty information about the ultraviolet spectra of the same compounds has been collected. Recent literature referenl@esin the field can be found in papers by Zahradnik and Carsky2 and by Shida and I ~ a t a In . ~the present work we present the uv and esr spectra of a series of anion radicals as well as the uv spectra of the parent neutral molecules. An attempt is made to interpret all the experimental data by means of a single MO calculation. The latter has been performed according to different The Journal of Physical Chemistry, Voi. 76, No. 26, 1972
methods in order to confront the merits of the different semiempirical MO theories in predicting both energy and electron distribution in radical anions. The skeleton of p nitrophenylethylene was present in all compounds under examination. The formulas together with twist angles are shown in Figure 1, where the twist angles are taken counterclockwise looking from the ethylenic carbon. (1) See, for example, (a) E. T. Kaiser and i., Kevan, Ed., "Radical Ions," Interscience, New York, N. Y . . 1968; (b) F. Gerson, "High Resolution E.S.R. Spectioscopy," Wiley, New York, N. Y., 1970. (2) R. Zahradntk and P, Carsky, J. Phys. Chem., 74, (a) 1235, (b) 1240, (c) 1249 (1970). (3) T. Shidaand S. Iwata: J. Phys. Chein., 75, 2591 (1971).
4
5
6
TABLE I: Polarographic Dataa
Materials. I-p-ni~~op~~enyl-l-phenylethylene, trans- and cis-2-bffonso-l-p-n~t~opheny~-~-pheny~ethylenes, trans- and cis- 2-ch ioro- P - p nitrop:henyP- 1-phenylethylenes, and transand cis-P-bromo-p-nit.rostyrenes have been kindly supplied by P'. 1,. Beltrame. For the synthesis of the compounds see ref 4 an 5 . The high degree of purity did not require further puriification. Acetonitrile (ACN) was SOL 16/66, It was further purified following given i n ref 6~Dimethyl sulfoxide (DMS) was RS, fi.rth.er purified following the method given in ref 7-9. Isooctane (IS(>)was a Rudipont product for spes:tra'photoHne.ry. Tetraet,hylammonium perchlorate (TEAPj weis a C ~ f l Erba o produict for polarography. P:Feporation. of i h i o n Radicals and Measurements. Anion rad.icais of the considered molecules have been been e l e ~ t r generated. ~ ~ ~ t Tvrio ~ ~ techniques ~ ~ ~ ~ have ~ used. The IG (internd genemtion)691*technique places the m.eueuu.y pool ::atbode within the microwave cavity. This technique h s khe advantage of simplicity and is particdarly suitable for short lifetime radicals. The EG (external generatioril t e c h n i q i ~ e ~places ~ - ~ ~the cell outside the micrwwave cavity. The arrangement necessitates transfer. of the d c t i o n prior' the esr observation, but allows rapid quarrb-itatiiie reduction of material because it permits t;be use of large ekctrtDde area. complete reduction is ,des,irablii: b~ecausemany of the radicals undergo rapid electron exc:hange reaction with their precursor, with line broadeni13.g in the spectrum. Moreover, the Past technique prese:nts the advantage of measurinlg in parallel SSP and UT?spectm of radicals, which gives more confidance in the assignameot of the observed electrornic spectra onitrile have been used. as ques, respectively. In both s have been made in vacutials by using a multipurposes A IMEL Moclel 436 polarograph. Reduction potentials far each 1c:ompourrd .have been evaluated €rom polarogmphic cui:vvei recorlded at room temperature. A saturated c~lomef.e18i:!ctroc3e e) W E I S the external refwence. About 0.1 l'd sal,utiLon of t , e t r a - n - e t h y l a m r n o n i ~perchlorate ~~ (TEA P,! was used i18 sirppiorting electrolyte, and the solu-
Solvent
~
1 2
i.OO 0.96 0.97
3 4 5
0.058 0.050
0.066
I .2a
0.062
1.05
0.070
1.?3 0,060 1.13 0.5633 1.36 0.062 1 .Q5 0.065 3.12 0.060 1 .Q7 0.858 1 .Q7 0.065 E1,4-.E1,2 = 0.056/n ( n is the elec-
0.95 0.061 ? 0.96 0.061 a It appears from Tomes' relation, trons number), that the first wave is reversible and monoelectronic (J. Tomes, Collect. Czech. Chem. Commun.. 9, 12, 81. 150 (1937)).
tion was ea. M . Dissolved oxygen was removed by bubbling purified nitrogen gas through the sample solution. Polarogmphie data measured in DMS and ACN are collected in Table 1. Electron spin resonance spectra were obtained with a Varian 4500-10 A X-band spectrometer with a LOO-kHz magnetic field modulation. Electronic absorption spectra were taken with a Becknsan DK-2A s p e c ~ ~ o p h o t o ~ ~ ~ e ~ ~ ~ . Esr Experimental Data. 1-p-Nitroph.emyl-l-phevzyle~h~lm e . The electrolytic reduction a t proper potential for the first reversible polarographic wave produces a blue-green (4) P. Beltrame and P. L. Beltrame, Gazz. Chim. Ita/., $8,367 (1968). (5) P. Beltrame, P. L. Beltrame, and b. Bellotti, J . Chem. Soc. B, 9'32 f1969I \ . - - - I .
(6) D. H. Geske and A . il. Maki, J. Arneo. Chem. Soc., 82, 2671 (1960); 53, 1852 (1961). ( 7 ) E. R. Talaty and G. A. Russel, J . Amer. Chem. Soc., 87, 4867 11965). (8) k.~F.-'Nelson,B. M. Trost, and D. H. Evans, d. Amsr. Chem. Soc., 83,3034 (1967). (9) I . Bernal arid G. K. Fraenkel. J. Amer. Chem. SOC, 86, 1671 (1964). (10) A. H. Maki and D. H. Geske, J. Chem. P h y s , 30, 1356 (1959); 33, 825 (1960). (11) P. H. Rieger, I. Bernal, W. H. Reinmuth, and G. K, Fraenkel, J . Amer. Chem. Soc., 85, 683 :1963). (12) C. S. Johnson, Jr., and R. Chang, J Chern. Phys., 43, 3183 f 1965) (13) j. k. Bolton arid 6. K. Fraenkel, J. Chem. Phys., 40, 33G7 (1964). (14) (a) C. Qliva, Thesis, University of Milano, 1969; ( 5 ) R. Mariano, Thesis, University of Milano, 1970. The Journal of Physical Chernislry, Vol. 76, No. 26, 1972
A. Gamba, V. Malatesta, G. Morosi, and M . Simonetta
3962
TABLE I I I: Calculated and Experimental Proton Hyperfine Splittings Theoreticala Anion radical
H
positionb Hortho Hmeta
Hi3 Hortho
I-P-NiTROPHENIL-I-PHENTLETHYLENE
Hmeta /a
I
u
H8 Hortho Hmeta
HI3 Hortho Hmeta
HB Hortho Hmeta
Hortho
TABLE II: Spin Denskies for 1-p-Nitrophenyl-1 -phenylethylene
Ha
Ha
A*/
Hortho Hmeta Ha
HB
N Q! 0 2
3' 4' 5' 6' 7' 8'
I\
5
C
0.002:!1 0.09348 0.1 7580 -0.021 - 1 1 - 0.02 1 09 0.14528 0.13599 O,OEI6:!8 0.22173 0.0861 3 0.08598 0.0O?27 -. Q,OQ2:16 -0.00 1Ell O.OOlt;4 0.00138 --Q .OO6CI1
-0.00509 0.13169 0.17356 -0.01409 -0.01453 0.14188 0.13342 0.10878 0.17648 0.08455 0.08431 0.00832 -0.00303 -0.00215 0.00150 0.001 17 - 0.00679
-0.02126 0.1 0093 0.15515 -0.03740 -0.03740 0.12316 0.12316 0.01767 0.21151 0.18302 0.18302 0.00113 - 0.00080 - 0.00080 - 0.00046 - 0.00046 -0.00013
a A represents spin densities calculated according LCI-SCF (Roothaan method,'" €3 spin densities calculated according LCI-SCF (LonguetHiggins and Pople) method,178 and C spin densities calculated according McLachian method.'*
solution in ACN and a grey-green solution in DMS, respectively. Well-resolved esr spectra of 53 lines in ACN and 64 lines in DMS have been recorded and interpreted on the basis of four triplets. On the basis of the spin densities obtained by MO calculations (see Table IT), the four triplets were assigned. to the nitrogen and three groups of two protons of equal coupling constants in ortho and meta The Journal of Physical Chemistry, Vol 76, No. 26, 1972
3.100 -0.837 2.740
C
3.091 3.184 3.163 -0.990 -0.816 -0.979 2.196 1.890 1.814
--1.070
3.180 2.124 3.15 -1.10 0.95 1.020
3.371 -1.131 0.532
3.438 -0.621 0.484
3.443 -1.202 0.533
- 1.000
3.166 -0.918 2.069
3.109 3.092 -0.749 -0.979 2.053 2.195
- 1.074
3.444 -1.203 0.531
- 1.052
3.352
- 1.074 3.089 -1.151 -0.297 2.156
ACN
DMS
3.135 3.220 3.250 2.922 -0.637 - 1.00s -1,150 -1.558 1.618 2.009 1.800 2.758
3.404 -0.555 0.584 3.137
3.139
- 1.444
-1.077
-0.296 1.869
-0.544 2.061
3.193 3.253 3.091 -1.219 - 1.529 -0.999 -0.217 -0.209 -0.398 1.869 2.006 1.383
3.320
3.051
2.141
3.23 -1.05 2.16
3.220 -1.073 1.089 0.945
3.248
3.220
- 1.080 - 1.520 2.260 3.250 -1.030 0 2.100
3.1 19 -1.058 -1.496 2.543
3.12 -1.34 0 2.20
a A represents spin densities calculated according LCI-SCF (Roothaan) method, 17b B spin densities calculated according LCI-SCF (LonguetHiggins and Pople) method,178 and C spin densities calcuiated according McLachlan method.'* Ortho and meta position to the nitro group. Hfs constants were optimized by least-squares fitting.'5
Pi"
2 3 4 5 6 7 8
B
0.616
Hmeta
1
A
Hl3
Figure 2. (a) Esr spectrum in ACN of l-p-nitrophenyl-l-phenylethylene. (b) Simulated spectrum.'s
Experimentai
TABLE IV:Calculated' and Experimental Nitrogen Hyperfine Splittings Exoeriment Anion radical
Theoryb
UMS
ACN
1
7.835 7.671 8.927 7.671 8.928 7.854 7.695
8.80 8.06 8.75 7.90 8.55 8.06 7.85
8.77
2
3 4 5 6
7
=
+
+
9.50 8.96 8.95 7.97 8.13
QNcN 2 Q ~ o PN" ~ ) f Q C N ~ ~f C2 ?Q o ~ ~ p o ' ( S f " 0 ~ 0= ~ lt(99.0 ) f 10.2) G;QNON I- 'F: (35.8 5.9)6 ,OCN' = 0.19b Nitrogen spin density evaluated by McLachlan method.'* a
QNCN
(SN
+2
positions to the nitro group and in the 2 position. The protons of the unsubstituted ring do not contribute to the hyperfine pattern. The spectra recorded in ACN and simulated with optimized hfs constants15 are shown in Figure 2. Optimized hfs constants are shown in Tables 111and IV. trans- and cis-2-Bromo-1-p-nitrophenyl-1-phenylethylene. Electrolytical reduction in DMS produces a lightgreen solution in the case of trans isomer and a dark-green solution in the case of cis isomer. Both spectra are well (15) J. Heinzer, "Least-Squares Fitting of Isotropic Multiline Esr Spectra," Program 197, Quantum Chemlstry Program Exchange (QCPE), Indiana University.
Twist angles
(degrees)
Resonance integrals (eV) _I~--_l_l__~____.___I
1,
Ionization potent als ( I P ) , electron affinities( A p ) . monocentric electron repulsicin integrals ( ' y P p ) , and Slater exponents (6,)
c N1
6i CI
Di-
11.IBd 28.85!je 17.860e 26.3aof 21.43Qg
*,
___._^
--
Q.03d 12.26e
3.87e 15.09f
Ywp(
/-A )
11.13 !6.595
13.99 11.29 11 .925h
6, _1.625
3.125 2.275 2 150 2.271
a L. E. Sutton, Chem. Sac:., Spec. Pub/., No. I i (1958). 6 Reference 21. Reference 28. G . Favini, I. Vandoni, and M. Simonetta, T h e m Chim. Acta, 3, 45 (1965). ". Pilcher a r d H . A. Skinner, J. Inorg. Nucl. Chem.. 24, 937 (1962). r G. Favini, S. Carra, and M , Simonetta, Gasz. Chim. Ita/.. 90,247 (1960), and references therein a (Br2') has been obtained fromexperimental value of ionization energy of the process Br+(szp4,3P) Br'+(s2p3, 4S) -19.2 e'/ (Landolt-Bbrnstei,l, "Atome und lonen," I . Band - I . Tail, p 212 Springer-Verlag, West Berlin, 1950, an0 from the energy difference between tnese anti valance States 81T (s2x2yz, V 2 ) and Br'+ (s2x2y2, VI) the energies of which are 0.49 and 2.72 eV, respectively (5,A . Skinner end t!. 0 . Pritchard. Tfans. Faraday Soc.. 49, 1254 (1953)). Evaluated according to the relationship - / H ~ H=~ ycici ' 6 c i ) . For anion radicais \he parameters have been corrected according to ref 3. -+
(aHi.
C l ' j 2-CHL090-i-P-WITA0PHE~~L-l-PHENYLETHYLE~~ -
,.e
M
Figure 3 . (a) Esr spectrum in ACN of os-2-chloro-l-~-nilroph@nyi-1 -phen~le~i7ylenejb) S mulated spectrum.
resolved and have been interpreted on the basis of three triplets, due t o nitrogen and protons in ortho and meta positions to the nitro group, and a doublet due to the proton in 2 position. Optimized hfs constants are shown in Tables HliI and IY The radicals are moderately stable and after aboul 0.5 hr sipnificaiit changes appear in the spectrum. When A C N is used as the solvent the spectrum changes rapidly with different rates for trans and cis isomers. The radical for the cis abomer is stable enough to record the esr spectrum, which has been completely resolved and interpreted. The optimized hfs constants are reported in T a -
Figure 4. (a) Esr spectrum of frans-/?-broma-p-nitrostyrene in ACN. (b) Simulated spectrum.
bles III and IV. In the case of the trans isomer the spectrum is rapidly evolving to a final spectrum which is identical with that of 1-p-nitrophenyl-1-phenylethykne. The same spectrum can be recorded for the cis isomer 1 hr after the start of electrolysis. This spectral evidence suggests bromine exchange with solvent protons"16 To substantiate (16) Particular care has been devoted to avoid the presence of traces of water in the vacuum electrolytic cell.
The Journal of Physicai Chemistry Val 76 No 26 1972
A.
3964
I_
Gamba, V . Malatesta, G . Morosi, and M. Simonetta
TABLE VI: Energy and Oscillator Strength of Doublet-Doublet Absorption Bands for Anion Radicals
1.0
,
-
i
'2
Figure 5.
Theory
U v spectra of 1 -p-nitrophenyl-1-phenylethylene:
(1)
radical anion (scale on the right, absorbance vs. wavelength); ( 2 ) neutral molecule (scale on the left, log t vs. wavelength); (3) radical after air exposure
2w
I 300
400
em
A m 1r m r '
ExperimentC Bcal,rib
A&, eV
f
A&, eV
1.783 1.819 2.957 2.996 4.072 4.41 1 4.673
0.124 0.069 0.079 0.004 0.204 0.039
1.614
0.177
1.805
0.004 0.000
1.776 2.138 2.833 2.853 4.312 4.494
0.005
1.695 2.204 2.935 3.060
0.005
0.055 0.201 0.003 0.014 0.290 0.039
0.104 0.067 0.003
2.840 3.116 4.161 4.492 4.683
0.077 0.039
4.427
1.603 1.932 2,826 3.483
0.005
(1.823) 2.214 2.95'1
1.749 2.009 2.780 2.912 4.338 4.425
0.005 0.161 0.000
0.005
(1.907)
0.042
2.097
0.005
0.005 0.100 0.059 0.018
0.110
1.610 1.819 2.736 3.477 4.210 4.572 4.748
1,720 1.954 3.293 4.032 4.321 4.369
0.004 0.223 0.032 0.323 0.052 0.093
1.702 1.899 3.269 4.132 4.360 4.484
0.151
Figure 8. Uv spectra of cis-P-bromo-p-nitrostyrene: (1) radical anion (scale on :he right, absorbance vs. wavelength); (2) neutral molecule (scale on t h e left, !og vs. wavelength); (3) radical after air exposure!.
this reaction mechanism and to verify if bromine-proton exchange occurs with configuration retention, experiments with CD36N solvent are in progress. trans- and cis-2.ChEoro-1-p-aitrophenyi-1 -phenylethylene. In both solvents radical anions undergo a fast decay. However, vveH-resolved spectra have been recorded and the bfs constants fitting experimental data are shown in TabTes III and IV. The magnetic nuclei that contribute to hyperfine patterns tire the same as for bromo-substituted compounds. The spectra of cis isomer recorded in ACN and simulated w%thoptimized hfs constants are shown in Figuse 3. trans- and cis-,3-Bromo-p-nitrostyrene. Stable radicals and well-resolved spectra have been obtained in both solvents. In DMS the trans isomer radical anion exhibits a mint green solution and a spectrum of 53 lines. In the case of the cis isomer a violet solution is observed and the spectrum shows 37 lines. In ACN the solution is dark green and indigo for: the trans and cis anion radicals, respectively. The spectra are very similar to those recorded in DMS and bromine elimination has not been observed. The same spectra. have been recorded when CD&N has been used as solvent. The spectrum of the trans isomer has h'een interpreted by means of five hfs constants, that is, three triplets duel! to nitrogen and protons in ortho and The Journal of Physical Gliemistry, Vol. 76, No. 26, 1972
0.024 0.034
1.922 2.119
(1. 8 9 3 ) d (2,025)d
0.000 0.029
0.036
0.000 0,008
0.022 0.045 0.223
(1.968) 2.101 3.178 3.646
0.000
-d on 0.022 WI 8
3.203 3.646
0.290 0.043
1.693 2.109 2.925 2.979 4.332 4,583 4.635
0.004 0.257 0.042
1.878 2.032
0.005
0.004 0.009
0.01 7 0.152
A€, eV
1.745 2.026 2.696 2.857 4.261 4.744
1.765 2.114 2.904 2.913 4.324 4.470
0.200
f
0.01 2 0.040 0.102
0.221 0.036 0.236 0.142
3.237
3.935 1.812 1.999 2.194 4.786
0.050
1.746 0.057 1.718 0.018 1.961 1.882 2.137 0.207 1.751 0.183 2.977 0.035 2.81 7 0.023 2.946 3.819 0.235 4.094 0.253 3.350 4.226 0.148 4.243 0.034 4.132 4.370 0.028 4.534 0.030 4.523 0.201 0.128 4.681 Acaicd derived by the LCI-SCF (Roothaan) method17b Bca]cd derived by the LCI-SCF (Longuet-Higgins and Pople) method.17a Solvent acetonitrile. Solvent dimethyl sulfoxide. Q
meta position to the nitro gxoup and two doublets due to a and P protons, respectively. For the cis isomer the doublet of a proton is not observable in the spectrum. Hfs constants fitting experimental spectra are reported in Table I11 and IV. Esr spectra of the trans isomer measured in ACN and simulated are shown in Figure 4. Uu Experimental Data. Visible and uv spectra for the anion radical and the neutral molecule of l-p-nitrophenyl-lphenylethylene and cis-P-bromo-p-nitrostyrene are shown in Figures 5 and 6.
Figure ‘ 7 ~Regression of theoretical excitation energies of the first, second, third, and fourth bands of neutral molecules observed frequencies (solid line: correla:ion coefficients r = 0.983, n = 24; 1, . (For numbering o i molecules see Figure I)
~~~~~~~~~~~~~
Calculations ~f spin densities and excitation energies were carried oint by means of semiempirical methods based on the crln approximation, due to the large dimenmons o l the considered molecules. Namely, restricted LCISCF methods, 111 the versions given by Pople and Long ~ e t - W i g g i n s ~B~ ”T by ~ Roothaan,17b and the McEachlan method,18 have lJaeri used. ‘The starting H MgS a have been evaluated adopting the ~ a ~ ~ m e t r i ~suggested ~ t i o ~ i by Rieger and F r a e n k e P for carbon, nitrogen, and oxygen, and the parameters reported in ref 20 for bromine and chlorine atoms. Careful erriinlinntion of X-ray data for related compounds2 1-24 a~lPowedthe assumption of idealized geomeIrks fur nrxatml mokcules. The geometries for the ions weye assunied equa’, to those of parent neutral molecules. These geometrics are described in Table ’V. Xn the same table ttie energy parameters adopted through the calculations and the corresponding references are reported. The nonplanarity of the moAecales was introduced by reducing the appropriate resmmce integrals through the reletion
‘Two-center Gcuhan 5 repulsion Integrals were estimated using the Parif er-Parr approximations.25 At distances greater thalr, 2 813 A. two-center integrals were caleulatebd theoretically with f ~ r r n u l a sg;ven by Ro0fhaan.2~In the
case of interatomic distances less than 2.80 ‘4,interpolation formulas have been obtained for each psir of atoms following the criteria indicated by H o f f ~ ~ ~eia nal, 27 In the @I calculations we considered the intersction of the ground state with the singly excited ~ o n f ~ ~ ~ofr ~ ~ ~ o n s types A, 13, and Ca according lo the definition reperted an ref %a.The 41 configurations, namely, four configurations of type A,four of type B, sixteen ol type Co7and sixteen of type ea, represent all the ~ ~ ~ ~arising € ~ ~ u from one-electron transitions between the four highest doubly occupied MO’s, the singly occupied MO, a n d the four lowest anoccupied MO’s. Energy and o s d l a l o s strength matrix elements for ~ Q U blets based ~ o n g u e t - ~ ~ ~ gand g ~ nh sp l e SCF-
e,,
(17) (a) H . C . Longuet-Higgins and J, A . Pople, Proc. Phys. Soc. A , 68? 591 (1955); (b) C. C. J~ Roothaan, Rev. Mod, Phys, 32, 979 (1960). (18) A. D. McLachlan, Moi. Phys., 3, 233 (1960). (19) P. II.Rieger and G. K. Fraenkel, J. Chein. Phyg., 39,609 (19F3). (20) A. Streitwieser, Jr., “Molecular Orbilal Theory for Organic Chemists,” Wiley, New York, N. Y . , 1962, p 135, (21) G . Casaione, C. Mariani, A. Mugnoli, and hl. :Simenetta, Ac:a Crysfailogr., 22, 225 (1967). (22) 6 . Casalone, 6. Mariani, A. Mugnoli, and M. Sirnonelte. Theor. C h i m A c f a , 8, 228 (1967). (23) G . Zasalane, A. Gavezzotti, C . Mariafii, A. iuiugnali. a n d M. Simunetta, Acta Crystaaiioyr,, S e d . 8,2 6 , 1 (197U). (24) G , Casalone and M . Simoneea, J. Chem. SOC.8,1180 (:97’! j . (25) R. Pariser and R. G . Parr, J. Chem. Phys.. 21, 466, 767 (1953;. (26) 6.C . 2 . Roothaan, S. Chern. Phys., 19, :445 (.iSS?) (27) A . lioffmann. A. lnamura, and 0 ,B. Zaiss, 2. Amer. Chem. Soc., 89, 5216 (1967). The Journal of Physical Chemtstiy, Vo!. 75, No 26, 1972
A.
3966
Gamba, V. Malatesta, G . Morosi, and M . Simonetta
1i
r
t
Figure 8. Regression of theoretical excitation energies of the first and second bands of radical anions on observed frequencies (solid line): correlation coefficient f = 0.885, n = 14. Styrenes Neutral
EIB)
Y I 1 0 - 3 (cm-1)
Anion
Y I
;;[[:
’.‘.
U ? I l
-1.5
-1.0
.-’“?,
“A
- 33.
7.1
.A
‘.’. ’.
-0.5
~
mi:].
5.: r 4
0 L
is
8
15 5
16.
16.5
1; Y x,o-3(cm-~)
0.5
1.0
Figure 9. Huckel’s MO diagram for l-p-nitrophenyl-ltrans-6-bromo-p-nitrostyrene. The phenylethylene and general features of the diagram are c\c”on to all studied diarylethylenes and styrenes were taken from Ishitani and Nagakura.28 Matrix elements in Roothaan ‘,s method have been recently published by ZQhradnik and dXrsky..2 Upon rederivation of‘ formulas we found one mnnor misprint. The correct expression for the matrix element ( 2 9 c ( a ,( E k ) ( H I 2 l c c n(~h I ) ) is 2(hhIG(li) - (hklGIil) G,,(mklGllm) + GRi(mhlGlim) withk ts lorr f h. Both SCF-CT procedures have been programmed and the programs checked against published data for some aromatic hydrocarbon ims.2g8,29 Singlet-singlet excitation energies and probabilities for neutral molecules have been evaluated according to stan-
+
-
The Journal of Physical Chemistry, Vol. 76, No. 26, 1972
-
Figure 10. ( A ) Plot (solid line) of the first absorption band of radical anions against the energy difference between the second and first absorption bands of neutral molecules for diarylethyienes (left y axis): r = 0.935, n = 5. (B) Plot ( - - - - ) of the second absorption band of radical anions against the first absorption band of neutral molecules for diarylethylenes (right y axis): r = 0.908, n , = 5. dard PPP method adopting the parametrization used for ions (see Table V). Forty-nine singly excited configurations have been considered for diarylethylenes and full configuration interaction for trans- and cis-nitrostyrenes, respectively. Calculated excitation energies together with experimental results are collected in Table VI for radical ions and in Table VI1 for neutral molecules. Proton hyperfine coupling constants obtained from calculated spin densities are shown in Table 111 together with the values obtained by esr data. Calculated spin densities have been translated into coupling constants by means of the usual (28) A. lshitani and S . Nagakura, Theor. Chim. Acta, 4, 236 (1966). (29) Y . A. Kruglyak and E. V. Mozdor, Theor Chim. Acta, 15, 365 (1369).
61.174
0.21
4.025
0.25
4958 5.559
0.27
6.198
-6.2
4.05 1
0.27
0.34
5.835
5.584
0.64
3.937
4.061
5.509
0.34
0.62
0.37
0.20 0.28
4.460 4.515
4.903 5.3213 5.726 6.075 8.225
(3.791) (4.188) 4.842
-0.10 -0.10 0.49
0.85
(3.850) 4.823
6.198 6,456
Q.76
0.80 1.36
(3.8503 4.927
5.42 13.34
0.134
5.40 1
0.36
6.334
0.015 0.672 0.089
6.536
0.890
4.509 4.800
0.048
4.805
0.152 0.235
5.03'1 5.081 6.026
6.14 6.352 6.276
0.040 0.527 0.001 0.0-19
4.105
0.75
6.198
4.188
0.18
4.216
0.13
4.919
0,35
4.900
0,59
6,267
1.26
4.011
0.49
5.485
0.31
6.198
0.40
6.473
0.89
4.094
0.35
5.586
0.29
6.439
0.50
0.007 0.618 0.037
1.243 0.792
4.413 4.933 5.719 6.849
0.028 0.010 0.010
4.064
0.090
4.160
0.33
)
5.71 2
6,133 6.306
0.525
6.396 6.804
0.450 3.042
6.439
3.849
0,688
4.424
0.830
3.935 4.1 32 4.245
5.365 5.832 6.04.4
0.596
6.600 6.696
0.42
~
0.112
3.809
6.21.1
3.986 (4 '791 (4.881 5.41
0.225
0.333
1.028 0.076
0.37
5.766
0.521
0.621;
a Interaction among ri.9 singly excited configurations has been considered for 4-5 systems. For smaller systems ( 6 and 7) full configuration interactioii has been uerformed.
cConnel relationship. For each method of calculation d~ffereIitQ constants fcir each proton position have been imized through 9 least -squares fitting of experimental constants us. the corresponding spin densities. Nitrogen hyperfine cmpling coxistants have been obtained only by means of McLmHan spin densities. The calculated er with the relationship between spin densiling consitants, the numerical values of the Q constants, and the corresponding experimental values are reported in Table 117. ~~$~~~~~~~~
Uu S,gectra From Tables VI and VI1 it i s clear which assignment of alrsorptiora bands to electronic transitions is suggested from calcukted transition energies and oscillator strengths. Id such assignments are accepted experi-
mental and theoretical transition energies can be correlat . ed. The results are shown in Figures 7 and 8, from which i t is evident that the correlation is satisfactory iv general. A better correlation is shown in the case of neutral molecules, where up to four bands are availabie and the energy spread amounts to about 2.6 eV. In the scope of the Wuckel approximation a relationship between the first two bands of each ary%ethgleiaeor styrene molecule and the corresponding anion 1s present, as shown in Figure 9. Such a relationship is not supvorted by experimental findings on an absolute scale, but in the ease of arylethylenes, where a series of at least five examples is available, a correlation between the pairs of spectra does exist (see Figure 10). As usual the Muckel approximation is meaningful if data concerning a series of similar compounds are considered. The Journal of Physical Chemistry, Voi 76
No 26. 7972
B. Meyer and T. Stroyer-Hansen
3968
Esr Siaectra. Datai reported in Tables I11 and IV are the results of a number of well-resolved esr spectra, which have allowed the unequivocal assignment of all the proton and nitrogen coupling constants for a series of unknown radical anions. The spectra do not show a significant dependence on the solvents used. The three theoretical methods that we have used always agree as to relative magnitude and sign of the calculated this basis we gave a sign also to experiment a1 coupling constants. When the three different methods are compeired, the McLachlan procedure seems to give a general better agreement with experimental data For this reason spin densities obtained by this method have been used to derive nitrogen coupling constants. From Table IV it appears that the order of magnitude of such constants is calculated correctly, but the relative order for the seven anions i s not reproduced. That i s due to the extremely small range of experimental and theoreti cal values.
Conclusions From the discussion of uv and esr spectra it appears that our approximate wave functions allow reasonable predictions of state energies and electron distributions for the considered molecules and anions. This fact is particularly gratifying since all the compounds are considerably distorted from planarity so that the validity of the a-T approximation was not assured in advance. The important influence of the absence of planarity is substantiated by the fact that in diarylethylenes the spin densities in the unsubstituted phenyl ring are practically zero. This also agrees with the fact that calculated coupling constants of the /3 protons in anions 3 and 5, where the twist angle of the substituted phenyl was assumed as too low, about 50%of the experimental value. Acknowledgment. One of us (G. .I thanks Professor C. Moser for hospitality at CECAM laboratories in Brsay; programs have been written during this stay.
. Mieyer* and T. Stroyer-Hansen Department of Chemistry, iiniversify of Washington, Seattie, Washington 98195
(Received May 26, 1972)
In matrices containing S Z a t an M I R ratio of 100 to 500 a t 2WK, uv absorption at 530 nm characteristic for S4 appears, and ir bands at 668, 483, 320, and 270 6m-I are found. Both the uv and ir spectra gain intensity during annealing and irradiation with visible light. Fine structure of the 668-cm absorption, in Kr a t 688, 681, 668, 660, 654, 647, 640, and 636 cm-I, is due to different normal vibrations, and to lattice sites and to a transient species which absorbs a t 647 cm-I. A very weak absorption around 625 nm, apparently related to the 647-cm-l peak, belongs to a transient species, probably S6 chains or $4 rings.
Recently S3 and S d were produced in low-temperature s o h t i ~ n and , ~ theei; uv spectrum was determined.2 Furthermore, it was discovered that S4 forms when S z in a krypton matrix is illuminated with visible light. This led us to restudy the ir absorption around 668 cm-1 which is found in matriccs containing trapped Sz 334 and in pure trapped sulfur vapor.6 The present study followed earlier experimental methods3 except that ir spectra were recorded with a P E 225 instrument, and samples were trapped on a CsI target instead of a sapphire window. We used M I R ratios between 100 and 500 and studied samples in various rare gases of varying thickness. The sulfur vapor was always at 1000°K and 0.1 Torr to optimize S 2 and eliminate all other sulfur For each sample, spectra were recorded four times: (a) immediately after deposit at 20°K; (b) after 1-hr exposure to a tungsten-iodine lamp; (c) after evaporation of the matrix, annealing at 76"K, and quenching to 20°K; and (d) at room temperature. As observed earlier,4 the spectra always i3howed the strong uv system of $ 2 and the ir absorption around 668 cm.-l In earlier ~ o r k , 3 -the { ~ ;elation between the uv spectrum of Sz and the 668-cm-l absorption was described. The ir The Journal of Physical Chemistry, Voi. 75. No. 26, 7972
band was tentatively assigned to matrix-induced absorption of Sz.The present work confirmed that the ir band increases with S2, but it was noted that the intensity of the ir band increases more than linearly with the concentration of S I and that at the onset of diffusion of S Z the ir band intensity increases further. In addition, in experiments with M I R 5 200 the matrix is red and a strong visible absorption a t 530 nm appears during deposition simultaneously with the 668-cm-l bands. This visible absorption is characteristic for S4.2 This suggests that the 668-cm-1 bands belong to S4, rather than S2.3-5 This interpretation is in accord with fluorescence work7 and a recent Raman spectrum,s both of which show a vibrational frequency of about 710 em-I for matrix-isolated Sz. In the present work, new ir bands were observed in adE. Meyer, T. V. Oommen, and D. Jensen, S. Phys. Chem.. 75, 512 (1971), B. Meyer, T. Stroyer-Hansen, and T . V. Oommen, J . Moi. Spectrosc., 42, 335 (1972). B. Meyer, J Chem. Phys.. 37, 1577 (1962) L. Brewer, 6 .D. Brabson, and R. Meyer, J. Chem. Phys.. 42, 1385 (1965). B. Meyer, Helv. Chim. Acta, 43, 1333 (1960). J. Berkowitz and J. 8 . Marquart, J . Chem. Phys., 39, 275 (1963). L. Brewer and G . D. Brabson, J. Chem. Phys.. 44, 3274 (1966).
