J. Phys. Chem. 1992,96, 623-626 The type of matrix clearly influences radical formation. In the more mobile7hmatrices, CF3CC13and CF2C1CFCl2,the Me2S'+ and Me2S4Me2+radicals are observed simultaneously after X irradiation, whereas in the rigid, more crystalline, CFC13 or CH2C12frozen solutions, they are detected separately depending on the concentration. Surprisingly, formation of the C13P~PC13+ homodimer is not observed in the mobile matrices,whereas in spite of ionization potential differences, the Cl3P&SMe2+heterodimer is readily detected in CF2C1CFCl2(but not in CF3CC13!). The exact mechanism controlling these phenomena is not clear. As a consequence of the absence of C13PLPC13+ radical cation in the mobile matrices, the secondary TBP-e C4P' and "X" radicals are also not detected in these frozen solvents. The experiments show that hyperfine couplings are fairly invariant to the type of freon. In CH2C12the coupling for the same radical structure is usually slightly increased. It is noted that by following the reactions in different halocarbon solvents a consistent description
623
of radical formation and reactivity is obtained.
Acknowledgment. This investigztion has been supported by the Netherlands Foundation for Chemical Research (SON) with financial aid from the Netherlands Organization for Scientific Research (NWO). Use of the services and facilities of the CAOS/CAMM Center, under grant numbers SON-1 1-20-700 and STW-NCH-44.0703, is gratefully acknowledged. We thank Mr. Henk Eding for his contributions in the art work. Registry No. C13P,7719-12-2;Me2S, 75-18-3;C1,P radical cation (l+), 136889-80-0; Me,S radical cation (I+), 34480-65-4;C1,P dimer radical cation (l+), 136804-95-0;Me2S dimer radical cation (I+), 76796-56-0; CI,P.Me2S radical cation (l+), 136804-91-6; CllP radical, 20762-59-8;C13P.Me2S-Me2S radical cation (1 +), 136804-92-7;CI,PC13P.H2S-H2S radical cation (1 +), H2Sradical cation (1 +), 136804-93-8; 136804-94-9; H2S radical cation (l+), 77544-69-5; H2S, 7783-06-4;H2S dimer radical cation (l+), 13604-96-1;CIsP, 58765-98-3;CI2P+, 75601-84-2;CI2Pradical, 20762-58-7.
Electronic Structure and Aromaticity of 1,3-Azaphosphole and 1,3-Azarsole TamPs VeszprBmi,*gt Uszl6 NyulPszi,t J6zsef RBffy,t and Joachim Heinicket Department of Inorganic Chemistry, Technical University of Budapest, 1521 Hungary, and Sektion Chemie, Ernst- Moritz- Arndt Universitat, Greifswald, Germany (Received: June 13, 1991)
Three different aspects of the aromaticity have been studied for 1,3-azaphospholesand azarsoles, namely the molecular geometry, the bond separation reactions, and the UV photoelectron spectra. Ab initio molecular orbital theory has been used to calculate the geometrical and electronic structure of the molecules. Comparing the two possible tautomeric forms, the 1H derivatives (type I) have planar structures while the 3H derivatives (type 11) show a puckered structure. The C-P or C-As bond lengths in the type I molecules are definitely shorter and the aromatic stabilization energies calculated from the bond separation reactions are larger than in type I1 molecules. Within each type I or type I1 molecules no characteristic difference in the aromaticity could be found. The band structure of the photoelectron spectra of type I molecules is similar to that of pyrrole, and the respective aromatic bands could be recognized.
The large variety of nitrogen-containing heteroaromatic systems inspired chemists dealing with organophosphorus compounds to synthesize heteroaromatic systems containing heavy atoms which replace the nitrogen. The achievements of the synthetic chemists in the 1980s have not been accompanied by widespread theoretical investigations. The photoelectron spectrum of the unstable parent compound phosphaethene' has been observed, and numerous quantum-chemical calculations have been conducted on this compound.2 Among the potentially aromatic metallocycles, ph~sphabenzene~ and phosphole4have been studied. In our recent work, investigations of 1,3-benzazaphosphole and -benzoxaphosphole have been reported.s The photoelectron spectra of methyl-substituted 1,2,4,3-triazaphosphole isomers have been studied, and on the basis of the change of some close-lying states the failure of Koopmans' theorem has been alleged.6 In a simple description of the electronic structure of nitrogen-containing heteroaromatic compounds two different types of nitrogen atoms are considered. For the first type the simplest example is pyrrole. The nitrogen atom in pyrrole is substituted by hydrogen that is easily removable by bases possessing acidic character to the parent molecule. In terms of M O theory the aromaticity and stability of pyrrole is the consequence of the interaction of the nitrogen lone pair with the cis-butadiene backbone forming a six-electron *-system. For the second type pyridine is the most common example. In this compound nitrogen possesses a basic character. In terms of M O theory this is a consequence of the free lone pair in the plane of the molecule. Technical University of Budapest. * Ernst-Moritz-Arndt Universitat.
The aromatic *-sextet can be formally constructed here from the interaction of a C=N double bond and the cis-butadiene systems. The least complex compound containing both types of nitrogens is imidazole. According to theoretical investigations the two different nitrogen atoms show similar conjugative ability. The partial charges and the linear coefficients in *-orbitals do not show significant differences between the two different types of nitrogen atoms of imidazole, benzimidazole, and ~ e r i m i d i n e . ~ . ~ When replacing nitrogen by a heavier Group V/a element a diminished interaction is expected due to the small overlap between carbon 2p and phosphorus 3p (or arsenic 4p) orbitals. However, in a recent ab initio study: phosphabenzene was found as 88-90% aromatic as pyridine is, while phosphole is a nonplanar compound in contrast to pyrrole indicating that the aromatic stabilization of the planar state is insufficient to overcome the high inversion ~
~~~~
(1)Lacombe, S.;Gonbeau, D.; Cabioch, J.-L.; Pellerin, B.; Denis, J.-M.; Pfister-Guillouzo, G. J. Am. Chem. SOC.1988,110,6964. (2)Gonbeau, D.; Pfister-Guillouzo, G.; Barrans, J. Can. J . Chem. 1983, 61,1371.Bruna, P. J.; Krumbach, V.; Peyerimhoff, S. D. Can. J. Chem. 1985, 63, 1594. Schoeller, W.W.J. Chem. SOC.,Chem. Commun. 1985,334. (3) Batich, C.; Heilbronner, E.; Hornung, V.; Ashe, A. J.; Clarck, D. T.; Cobley, U. T.; Kilcast, D.; Scanlan, I. J. Am. Chem. SOC.1973,95, 928. (4)Schafer, W.; Schweig, A.; Mathley, F. J . Am. Chem. SOC.1976,98, 407. Mathey, F. Chem. Rev. 1988,88, 429. (5) Nyuliszi, L.;Csonka, G.; Rbffy, J.; Veszprbmi, T.; Heinicke, J. J . Organomet. Chem. 1989,373,49,57. (6)Gonbeau, D.; Pfister-Guillouzo, G.; Barrans, J.; Palmer, M. H. Chem. Phys. 1985,95, 243. (7)Nyuliszi, L.; Pasinszki, T.; Rbffy, J.; Veszprbmi, T.; Thiel, W.; Fabian, J. Struct. Chem. 1990,I , 367. (8) Baldridge, K. K.; Gordon, M. S. J . Am. Chem. Soc. 1988,110,4204.
0022-36S4/92/2096-623$03.00/00 1992 American Chemical Society
624
Veszpr6mi et al.
The Journal of Physical Chemistry, Vol. 96, No. 2, 1992
TABLE I: Bond Lengths (A) and Bond Angles (deg)
a
1.299
b
1.369
C
1.379
d e a
1.352 1.385 106.3
B
111.0
Y
107.3
6 t
1.821 (1.783)" 1.333 (1.343) 1.468 (1.438)
1.698
1.837
1.873
2.001
1.361
1.275
1.339
1.264
1.365
1.424
1.377
1.428
1.353 1.772 88.9
1.329 1.814 86.4
1.341 1.938 83.6
1.324 1.966 82.4
113.1
114.3
113.0
113.1
114.8
112.1
116.8
115.4
105.7
112.1
116.6
114.7
118.9
109.7
111.7
110.2
111.8
110.0
89.08 (90.7) 111.1 ( 110.0) 114.1 (114.1)
1.719 (1.733) 1.381 (1.413) 1.384 (1.384)
1.878 (1,850) 1.37 1 (1.390) 1.386 (1.401)
100.5 (101.1) 125.4 (124.4) 123.2 (1 23.7) 122.5 (122.8)
96.3 (97.3) 124.7 (125.1) 124.7 (124.2) 124.7 (124.4)
Average values from the X-ray analysis of l-benzylphosph~le.~~
barrier for tervalent phosphorus. These facts indicate that the two different phosphorus atoms are not equally good building blocks of aromatic systems as happens in the case of nitrogen. The simplest phosphorus compound that can form an aromatic system containing either a P=C subunit or a PH structure is 1,3-azaphosphole, which can exist in two tautomeric forms (I and 11). The comparison of the electronic structure of these two
I I
I1
tautomers gives rise to the investigation of the conjugative ability of the lone pair of the phosphorus atom relatively to the M subunit. Compound I has only recently been synthesized in phosphorus and its arsenic analogue,9pl0while compound I1 has not yet been synthesized. In this work we will compare the molecular structure of the two forms using quantum-chemical calculations and UV photoelectron spectroscopy.
Experimental Methods and Calculations The synthesis of the compounds investigated has already been reported.9,'0 The photoelectron spectrometer has been described elsewhere.'l Spectra were recorded with He I (21.22 eV) and He I1 (40.81 eV) radiation. Spectral calibration was achieved using Me1 and Ar. Under the experimental conditions the resolution was about 45 meV (fwhm) for the Ar zP3/2line. Minimum energy geometries were computed at the SCF level for all of the investigated molecules in their ground electronic states. For testing the calculated minima second derivative matrix were calculated for all of the molecules. As the calculated harmonic frequencies were positive at the geometries obtained, the equilibrium structures are considered as true minima at the potential surface. For the ionization energies we assumed the validity of Koopmans' theorem. Calculations were carried out using the quantum-chemical program CADPAC.]~The basis set used was the split valence 4-31G type augmented by one set of d polarization functions on phosphorus (sd= 0.57).13 For arsenic a basis set (12s,9p,3d) contracted to [4s,3p,ld]I4was used. As a comparison
TABLE Ik BOAS
(A)
Calculated Bond
Lengths for Isolated Single and Double
estimated from bond C-N c-P C-AS
4-31G (P)
covalent radii
H,X-CH,
HX=CH,
1.47 1.87 1.98
1.45 1 1.867 2.022
1.256 1.635 1.801
all of the arsenic compounds were calculated with a 3-21G basis set (including arsenic too). Since no essential difference were found, only the results of the fmt basis set will be presented. All SCF calculations were of the restricted H F type.
Results and Dlscsasion The calculated structural data of the investigated molecules are presented in Table I. Since there is no available geometrical data for the investigated molecules, calculations were performed on phosphabenzene, arsabenzene, and phosphole for the purpose of checking the reliability of the method. Comparing the results of our calculations to the experimental datal5-'' (in parentheses in Table I) and to some of the available results of ab initio quantum-chemical calculations,* they reproduce the known structures within acceptable error. In order to study the double-bond character of the ring atoms, an estimation of the C-X bond length can be made using the sum of the covalent radii'* of C and X atoms. A better comparison of single and double C-X bond lengths can then be made from the results of the calculations of CH3-XH2 (single bond) and CHz=XH (double bond) using the same level of theory. These data are summarized in Table 11. As expected, the C-X bond lengths reflect the partial double bond character in both types of molecules. In type I molecules the bond shortening and the double-bond character are definitely larger than in type I1 (seeTable 111). (The bond shortening refers to the calculated CH,-XHz bond distance and the double bond percent refers to the difference between the C-X bond lengths of CH3-XHz and CHz=XH. In lH-azaphosphole the average of the two P-C bond lengths, 1.735 A, is very near to the value reported for the P-C bond in (14) Pokier, R.; Kari, R.; Csizmadia, I. G . Handbook of Gaussian Basis
~~
~
(9) Heinicke, J. Tetrahedron Lerr. 1986, 27, 5699. (10) Heinicke, J. Z . Anorg. Allg. Chem. 1989, 29, 71. (11) VeszprCmi, T.; Zsombok, Gy. Magy. KZm.Foly. 1986, 92, 39. (12) Amos, J.; Rice, E. CADPAC: The Cambridge Analytic Derivatives Package, issue 4.0, Cambridge, 1987. (13) Wong, M. W.; Gill, P. M.; N o h , R. H.; Radom, L. J . Phys. Chem. 1988, 92, 4815.
Set; Elsevier: Amsterdam, 1985. (15) Wong, T. C.; Bartell, L. S.J . Chem. Phys. 1974,61, 2840. (16) Wong, T. C.; Ashe, A. J., 111; Bartell, L. S. J. Mol. Srruct. 1975, 25, 65
(17) Coggon, P.; Engel, J. F.; McPhail, A. T.; Quin, L. D. J . Am. Chem. SOC.1970, 92, 5779.
(1 8 ) Cotton, F. A.; Wilkinson, G . Aduanced Inorganic Chemistry, 2nd 4.; Wiley: New York, 1966.
The Journal of Physical Chemistry, Vol. 96, No. 2, 1992 625
1,3-Azaphospholeand 1,3-Azarsole TABLE III: Double-Bod Character of tk Investigated Molecules
/
bond shortening W double bond W
6.0 45
TABLE I V Bod-Separation Reactions comwund AEa % compound 46.67
100
I
5.1 41
2.5 20
AEa
W
18.74
40
18.15
39
2.3 18
4.2 38
1.9 18
A
I 1
52.74
113
1
&-
He11
A
A
AE in kilocalories/mole.
phosphabenzene (1.733 A).'$ The calculated average As< bond length in 3H-azarsole, 1.906 A, is somewhat longer than that in arsabenzene (1.850 A).16 In the case of structure I1 both the P-C and the As-C bond distances are considerably longer, which makes them closer to the length of a single P-C (As-C) bond. The CXC bond angle decreases with the increasing size of heteroatom; however, no obvious trend could be found between the type I and I1 molecules. The increased C-X bond length and the decreased CXC bond angle minimize the distortion of the ring. Thus,the structural change in the opposite part of the ring is small, e.g. only 0.01 A is the total change in the C-C bond length from X = N to As in the I molecule. While 1H-substituted compounds are planar, 3H-substituted molecules have a puckered structure similarly to phosphole. In case of 3H-azaphosphole the P atom moves out of the plane of the other four ring atoms with about 5". In case of As derivative this angle is about 3". The P-H (As-H) bond is nearly perpendicular to the plane of the CXC moiety. Earlier experimental and theoretical investigations of the aromaticity of phosphole evoked controversial conclusions.4~'9~20 A recent method used8.21to provide a quantitative measure of the delocalization stabilization,M,in potentially aromatic compounds is the calculation of the stability of the molecule investigated relative to its simplest fractions (bond separation reaction)22or to its larger fractions (homodesmic, superhomodesmic react i ~ n ) . * ~The - ~ bond-separation ~ reactions of the group V fivemembered rings*suggest that the delocalization stabilization for pyrrole is much greater than that for the other members of the series. A comparative study of the superhomodesmic reactions proved that the stabilization exhibited by the bond-separation reactions for the latter compounds is already contained in butadiene before the ring is formed. Only in pyrrole is the delocalization of the butadienyl moiety greater than that of isolated butadienee8 (19) Epiotis, N. D.; Cherry, W . J . Am. Chem. Soc. 1976, 98, 4365. (20) Egan, W.;Tang, R.; Zon, G.; Mislow, K.J . Am. Chem. SOC.1970, 92, 1442. (21) Schlegel, H. B.;Coleman, B.;Jones, M., Jr. J . Am. Chem. Soc. 1978, 100.6499. (22) Hehre, W.J.; Ditchfield, R.; Radom, L.; Pople, J. A. J . Am. Chem. SOC.1970, 92,4196. (23) George, P.; Trachtman, M.; Bock, C. W.; Brett, A. M. Theor. Chim. Acta 1975, 38, 121. (24) Hess, B. A., Jr.; Schaad, L. J. J . Am. Chem. SOC.1983,105,7500.
18
16
14
12
10
8
IPleV
Figure 1. Photoelectron spectra of the investigated molecules.
In order to evaluate the aromatic stability of the two tautomers, the following bond-separation reactions were investigated: