J . Phys. Chem. 1988, 92, 33-36 the fitting procedure are given in Table VII. Generally, there is reasonable agreement between the calculated values and the qualitative observations of band intensities, which are of course subjective and affected by band type. The largest intensities are found for the Sn-H stretching [ v 2 ( A 1and ) u7(E)],and the next most intense are the SnH, deformation bands [u4(A1)and vlo(E)] followed by the CH, and SnH, rocking modes [v,(E) and v l l ( E ) ] . The corresponding motions are also the strongest bands in the recorded spectra. The very low intensities of the CH3 deformations [ Y ~ ( A and I ) vs(E)]are also supported by our calculations. The only disagreement is found for the Sn-C stretching, where the calculated intensities appear to be widely underestimated relative to the qualitative observation although ap/aS, is important. Concerning the vibrational assignments based on the analysis of the potential energy distribution, our calculations show that except for the SnD, deformation and the SnC stretch modes of the A , symmetry species in SnD3CH3,the concept of characteristic group frequencies can be applied to the SnH3CH3and SnH3CD3 species in A I symmetry. In the case of the E symmetry species it appears that the CH, rocking, SnH, rocking, and SnH, deformation modes show a significant amount of mixing for CH3SnH, and CD3SnH3whereas the CH, rocking and SnH, deformation become relatively pure in CH3SnD3. The theoretical sequence of band positionning agrees with the experimental sequence obtained by Kimmel and Dillard.I7 Even in the relatively restricted spectral range between 700 and 800 cm-' our final calculations confirm the fact that the degenerate SnH, deformation is located between the degenerate CH3 rock and the symmetric SnH, deformation and has a higher intensity than the latter two. Nevertheless, it should be emphasized that this sequence is not exactly reproduced a t S C F and MP2 levels, where the degenerate SnH, deformation is found to be lower than the corresponding A I mode by about 10 cm-I. In all cases the
33
CH3 rock can be predicted close to 800 cm-l, in good agreement with the experimental value ( e 7 9 6 cm-I), comparable to the E , frequency (822 cm-I) found in ethane. The degenerate CH, deformation is a t a higher wavenumber than the symmetric one: this latter can be predicted to lie at 1236 cm-', somewhat lower than the values observed at 1264 cm-' in methylsilaneZ6and the harmonic values found in ethanes2' For these two modes it should be noted that the excessive weakness of the calculated intensities is in perfect agreement with the qualitative observations. The asymmetric and symmetric CH, stretchings are respectively predicted at 3156 and 3062 cm-', which is consistent with the observed values. For both calculations (SCF and MP2), the difference between the two bands which is found to be about 100 cm-' in CH3SnH3is widely increased in CD,SnH3 according to the infrared data. The antisymmetric and the symmetric SnH, stretches form a close group at about 1944 cm-I; their calculated splitting decreases when the level of theory is improved in our calculations. These bands shift in the spectra of SnD3CH3to about 1380 cm-' but remain unchanged in their assignments. In the low reginn of the IR spectra, the Sn-C stretch is found around 520 cm-' while the SnH, rock is expected at 430 cm-I, in good agreement with the experimental values. Furthermore, we find some mixing of the SnH, deformation or CH3 rock with these modes in agreement with the experimentally observed effect of deuterium isotope substitution. The lowest vibrational frequency of the molecule is the infrared-inactive torsional mode. Our MP2 calculations indicate that this stannyl torsion should appear at 118, 109, and 95 cm-' for the three isotopic species. These values are consistent with those estimated at 110, 101, and 88 cm-' by Kimmel and Dillard, based on the study of combination bands in the infrared spectra. Registry No. Methylstannane, 1631-78-3.
Pyramidal Inversion in Silyl Anions James R. Damewood, Jr.,* and Christopher M. Hadad Department of Chemistry, University of Delaware, Newark, Delaware 1971 6 (Received: April 3, 1987)
The barrier to pyramidal inversion in the silyl anion, SiH3- (l),the disilanyl anion, SiH,SiH,- (2), the formylsilyl anion, HCOSiH2- (3), and the borylsilyl anion, BH2SiH2-(4), are obtained by ab initio methods at the 6-31G* level of sophistication. The pyramidal inversion barriers calculated for these compounds are 34.7, 26.6, 23.8, and 0.7 kcalfmol, respectively. A comparison is made between the inversion barriers at silicon in the silyl anions and those at phosphorus in the isoelectronic phosphines.
Introduction Like their isoelectronic analogues from group 15 (Va), the anions of group 14 (IVa) are tricoordinate with a lone pair of electrons and have the potential to undergo the process of pyramidal inversion.' The parent compound in this series, the methyl anion, has received considerable theoretical and experimental attention2 and (if bound) is believed to reside in a very shallow potential energy well with an inversion barrier of only ca. 1-2 (1) For reviews of pyramidal inversion, see: Rauk, A.; Allen, L. C.; Mislow, K. Angew Chem., Int. Ed. Engl. 1970, 9, 400. Lambert, J. B. Top. Stereochem. 1971, 6, 19. (2) For example, see: Dykstra, C. E.; Hereld, M.; Lucchese, R. R.; Schaefer, H. F. 111; Meyer, W.J . Chem. Phys. 1977, 67, 4071 (theoretical). Ellison, G. B.; Engelking, P. C.; Lineberger, W. C. J. Am. Chem. SOC.1978, 100, 2556 (experimental).
kcal/mol. Unlike the methyl anion, relatively little information is available on the higher homologues from this series., The static and dynamic stereochemistry of silyl anions, for example, is particularly poorly characterized even though these species are some of the more common reagents in modern silicon chemistry. We have performed the ab initio calculations reported in this paper in order to provide information on the pyramidal inversion process in silyl anions. These data will allow for comparisons between (3) For example, see: (a) Rauk, A,; Andose, J. D.; Frick, W. G.; Tang, R.; Mislow, K. J . Am. Chem. SOC.1971, 93, 6507 and references therein. (b)
Keil, F.; Ahlrichs, R. Chem. Phys. 1975, 8, 384. (c) Lambert, J. B.; Urdaneta-Ptrez, M. J . Am. Chem. Soc. 1978,100, 157. (d) Eades, R. A,; Dixon, D. A. J . Chem. Phys. 1980, 72, 3309 and references therein. ( e ) Hopkinson, A. C.; Lien, M. H. Tetrahedron 1981,37, 1105. (0 Gordon, M. S.; Boudjouk, P.; Anwari, F. J . Am. Chem. SOC.1983, 105, 4972. (g) Damewood, J. R., Jr. J . Org. Chem. 1986, 51, 5028 and references therein.
0022-_1~654/88/2092-0033$01.50/0 0 1988 American Chemical Society
34 The Journal of Physical Chemistry, Vol. 92, No. 1. 1988 TABLE I: Calculated Absolute Energies' for Silyl Anions 1-4
Damewood and Hadad TABLE II: Calculated Inversion Barriers for Silyl Anions 1-4
basis set compd
1 2a 2b 3 4 4a 4b
configuration C, (pyramidal) Dlh (planar) Ci (pyramidal) C, (pyramidal) C, (planar) C, (pyramidal) C, (planar) C, (pyramidal) CZ" (planar) C, (orth pyramidal) CsL.(orth planar)
3-21G -289.084 50' -289.030694 -577.660 187 -577.662051 -577.616851 -401.185482 -401 .I47 812 -314.202 112 -314.202 105
6-31G* -290.606 105 -290.550 784 -580.703 798 -580.706 144 -580.663 800 -403.344 037 -403.306 168 -315.862040 -315.860990 -3 15.840 226 -31 5.717 433
inversion barrieP compd 1
2 3 4
3-21G 33.gd 28.4 23.6 0.0
Methods All calculations were performed with complete geometry optimization using the program GAUSSIAN82 (G82)9 and the 6-31G* basis set &e., 6-31G*//6-31G*). Preliminary calculations were performed with the 3-21G basis and these geometries were used as input structures for the 6-31G* studies. While we report energies and structures obtained at both levels of sophistication for the purposes of comparison, the 6-31G* calculations are of a higher quality and, in most instances, we will limit our discussions (4) For examples, see: (a) Egan, W.; Tang, R.; Zon, G.;Mislow, K. J. Am. Chem. SOC.1970, 92, 1442; (b) J. Am. Chem. SOC.1971. 93, 6205. (c) Mislow, K. Trans. N.Y.Acad. Sei. 1973, 35, 221. (5) Swalen, J. D.; Ibers, J. A. J . Chem. Phys. 1962, 36, 1914. (6) (a) Lehn, J. M.; Munsch, B. Chem. Commun. 1969, 1327; Mol. Phys. 1972, 91, 6507. See also: Boggs, J. E. J . Chem. Phys. 1971, 66, 5769 and references therein. (b) Dixon, D. A.; Arduengo, A. J., 111; Fukenaga, T. J . Am. Chem. SOC.1986,108,2461, Arduengo, A. J., 111; Dixon, D. A,; Roe, D. C. J . Am. Chem. SOC.1986, 108, 6821. (7) Stackhouse, J.; Baechler, R. D.; Mislow, K. Tetrahedron Letr. 1971, 3437. (8) (a) Damewood, J. R., Jr.; Mislow, K. Momrsh. Chem. 1980,111,213. (b) Xie, Y.;Boggs, J. E.; Khaikin, L. S. J . Mol. Srrucr. (THEOCHEM)1986, 139, 255. (9) Binkley, J. S.;Frisch, M. J.; DeFrees, D. J.; Raghavachari, K.; Whiteside, R. A.; Schlegel, H. B.; Fluder, M. J.; Pople, J. A. A copy of the program may be obtained from Carnegie-Mellon University. (10) Marynick, D. S.; Dixon, D.A. Faraday Discuss. Chem. SOC.1977, 62. 47.
sum of angles at Sib*' 286.3 285.9 29 1.O 336.6
'In kcal/mol. 'Summation of the angles about silicon for the ground-state structure (using the 6-31G* optimized geometry). In degrees. dUsing the C,, energy from the C M U database.
In hartrees/molecule. bObtained from the C M U database.
silyl anions and their analogues from group 15. Pyramidal inversion at group 15 centers has been studied most extensively for amines and phosphines.'q4 Theoretical and experimental studies of these species have revealed that a variety of substituents may lower the inversion barrier relative to that found for ammonia (NH,) or phosphine (PH,). Substituents which typically lower these inversion bamers include silyl (-SiH3), acyl (-COR), boryl (-BH,), and a variety of other f~nctionalities.~ In all cases, substituents that lower the barrier to pyramidal inversion are either more electropositive than hydrogen or serve as acceptors of electron density from the nitrogen or phosphorus lone pair. Despite this similarity in substituent effects, significant differences between pyramidal inversion at nitrogen and phosphorus are evident. Consider, for example, that the experimentally determined barrier to pyramidal inversion in ammonia is only 5.8 kcal/mo15 while for phosphine it is predicted to be considerably higher at ca. 37 kcal/mol.6 As a result, the manifold of inversion barriers for phosphines is substantially higher in energy than that for amines. As a further illustration of the differences between these compounds, while an excellent correlation exists between the barrier to pyramidal inversion and the ground-state pyramidality at nitrogen,' substituents that lower the bamer to pyramidal inversion in phosphines have little to no effect on the ground-state pyramidality.4*s Thus, while the amine pyramid is rather flexible, phosphines are rigidly pyramidal. It has recently been suggested3* that the silyl anions are more similar to phosphines than amines in that they also remain rigidly pyramidal, even in those cases where substituents lower the barrier to pyramidal inversion at silicon. The calculations described in this work will allow for a test of this hypothesis.
6-31G* 34.7 26.6 23.8 0.7
1
2a
2b 2
3
4
Figure 1. Schematic representation of the calculated structures for 1-4. Bonding parameters obtained at the 3-21G level of sophistication are shown with the molecular representations. Top: the silyl anion (1); left, C,, ground state; right, Qhtransition state. Center top: the disilanyl anion (2); left, C, eclipsed (2a); center, C, staggered (2b); right, C, transition state. Center bottom: the formylsilyl anion (3); left, C, ground state; right, C, transition state. Bottom: the borylsilyl anion (4); left, C, ground state; right, C,, transition state.
to these higher quality data. Where indicated, symmetry constraints were employed in the geometry optimizations. Planar geometries about silicon were employed in the calculation of all transition states. All bonding parameters were completely optimized. While mechanisms other than pyramidal inversion have been considered for inversion in phosphines,6b these mechanisms seem to be operative in those cases where the barrier to pyramidal inversion is extremely high. We therefore consider only pyramidal inversion in the present work. Calculations were performed on the silyl anion (SiH3-, l), the disilanyl anion (SiH3SiHF, 2), the formylsilyl anion (HCOSiHF, 3), and the borylsilyl anion (BH2SiHF, 4). For 1, pyramidal (C3J and planar (D3,Jstructures were considered. For 2, C, pyramidal configurations in which the lone pair on silicon either eclipsed one (2a)or staggered two (2b)adjacent Si-H bonds were investigated. In the case of 3, the ground-state pyramidal structure is of C , symmetry. In the C,, planar form of 3, the planes of the SiH2 and H C O groups become coincident and form the molecular symmetry plane. For 4,the pyramidal form is of C, symmetry and the planar transition state is of C,, symmetry. The barrier to rotation about the Si-B bond in 4 was calculated for structures with both pyramidal and planar geometries at silicon. In these transition-state conformations, the H-B-H and H-Si-H planes are orthogonal and the structures are of C,(orthogonal pyramidal, 4a) and C2, (orthogonal planar, 4b) symmetry, respectively. Absolute (3-21G and 6-31G*) and relative energies for the various configurations of 1-4 are reported in Tables I and 11, respectively. Selected bonding parameters are reported in Figures 1 (3-21G) and 2 (6-31G*).
The Journal of Physical Chemistry, Vol. 92, No. 1, 1988 35
Pyramidal Inversion in Silyl Anions H
- Si
1.490
/H
I200
\.
1
2b
20
2
3
1
4
/
~ 2 . 0 9 41.544 ~~ 9 3 . 1 w H H94!
H
40
compares well with results that have been obtained previously for 1 at the ab initio level (1.525 A, 97.6°;3b 1.530 A, 97.0°;3d 1.559 A, 97.9°;3e1.501 A, 93.4°;3f 1.562 A, 95.2°3f). In the D3,,transition state, the Si-H bond lengths shorten to 1.490 8, and the H-Si-H angles are all 120'. For the disilanyl anion (2), the silyl group may adopt two different orientations in the ground state. In one, the lone pair on silicon eclipses an Si-H bond on the SiH, group (2a) and in the other, the lone pair staggers two of these Si-H bonds (2b). We find that 2b is the lowest energy conformation with 2a 1.5 kcal/mol higher in energy at the 6-31G* level. Using 2b as our ground-state configuration, we calculate an inversion barrier for 2 of 26.6 kcal/mol. Thus, as is the case for amines and phosphine~,',~ the silyl group is able to significantly lower the inversion barrier of the silyl anion. At the 6-31G* level, the Si-Si bond length is 2.393 8,in the ground-state, C,, pyramidal structure (2b). This Si-Si bond shortens significantly to 2.264 8,in the C,,planar transition state. The Si-H bonds at the anionic center are ver similar in length to those observed in 1 with values of 1.531 for the pyramidal structure and 1.483 8, and 1.480 8, for the planar structure (Figure 2). The formylsilyl anion (3) contains a *-acceptor attached to the anionic center and, by analogy to amines and phosphine^,'^^^^^ we therefore expect the inversion barrier to be substantially reduced relative to 1. We calculate an inversion barrier of 23.8 kcal/mol for this molecule in complete agreement with these qualitative predictions. The C-Si bond length in the C, ground state is 1.947 8,and the C-O bond length is 1.208 A. In the C,, planar transition state these bond lengths adopt values of 1.834 and 1.221 A, respectively. The shortening of the Si-C bond with a concomitant lengthening of the C-0 bond in the transition state is strongly suggestive of significant conjugative interaction between the silyl anion slnd the carbonyl functionality (see below). The Si-H bond length shortening observed in 1 and 2 is also seen in 3 with the C, and C, structures adopting average values of 1S27 and 1.479 A, respectively. The borylsilyl anion (4) contains a very strong *-acceptor adjacent to the silyl anion center. Previous ab initio calculations on this anion predicted a planar, C, structure as the ground (Le., an inversion barrier of 0.0 kcal/mol). In contrast to these previous prediction^,^^ at both the 3-21G and 6-31G* level, we find that the ground state of this compound is pyramidal and of C, symmetry. At the 3-21G level, the structure is extremely flexible in the vicinity of the pyramidal form and inversion requires energies well below the error limits of the calculations (Le., an inversion barrier of 0.005 kcal/mol). At the 6-31G* level, we calculate an inversion barrier of 0.7 kcal/mol, a value which represents the most substantial reduction of a pyramidal inversion barrier in a silyl anion yet reported. Structural evidence at the 6-3 1G* level supports the hypothesis of significant silyl anionborane overlap in 4. The Si-B bond length, for example, decreases from 1.927 8, in the C,, ground-state structure to 1.878 8, in the C,, transition state. It is interesting to note that only the latter value compares well with the 1.882 8, reported previously3e for the planar structure. In addition, the angles about silicon are significantly expanded relative to those found in 1-3. While the change in Si-H bond lengths observed for 1-3 is still apparent in 4, the change from the ground state (1.508 A) to the transition state (1.495 8,) is significantly smaller. Ground-State Geometry us Pyramidal Inversion Barrier in Silyl Anions. Because of the excellent correlation between the ground-state pyramidality of amines with their pyramidal inversion barriers,' the question naturally arises as to whether such a correlation exists for silyl anions. I n order to test for such a relationship, we have summed the angles about the silyl anion center and used these values as a measure of pyramidality at silicon." The values so obtained at the 6-31G* level of sophistication are reported in Table 11. A plot of the 6-31G*
4b
Figure 2. Schematic representation of the calculated structures for 1-4b.
Bonding parameters obtained at the 6-31G*level of sophisticationare shown with the molecular representations. Top: the silyl anion (1); left, C, ground state; right, D3* transition state. Center top: the disilanyl anion (2); left, C, eclipsed (2a); center, C, staggered (Zb); right, C, transition state. Center bottom: the formylsilyl anion (3); left, C, ground state; right, C, transition state. Bottom: the borylsilyl anion (4); top left, C, ground state; top right, C, transition state; bottom left, C, orthogonal pyramidal transition state (4a); bottom right, C,, orthogonal planar transition state (4b).
Discussion Pyramidal Inversion Barriers and Structures of 1 - 4 . The barrier to pyramidal inversion in the silyl anion (SiH3-, 1) has been calculated by a variety of ab initio and semiempirical methods. Current estimates for the barrier to pyramidal inversion While few calculations in 1 range from 26.0 to 39.6 on silyl anions have been r e p ~ r t e d , ~ previous ~ , ~ . ~ . ab ~ initio studies have, to a limited degree, investigated the effects of substitution on the barrier to pyramidal inversion in l.3eThese calculations considered SiH,X- compounds (X = H, BH2, CH3, NH2, OH, and F) and obtained barriers of 34.3,O.O (planar), 40.4, 42.8, 50.7, and 57.3 kcal/mol, respectively. Since most of these barriers are higher than that found for 1, in the present study we wanted to investigate a variety of substituents that would be expected to lower the barrier to pyramidal inversion at silicon. We have therefore considered the disilanyl anion (SiH3SiH,, 2), the formylsilyl anion (HCOSiH2-, 3), and the borylsilyl anion (BH,SiH,-, 4). For completeness, we have also investigated the parent silyl anion (1). The pyramidal inversion barriers obtained from these calculations are reported in Table 11. Inspection of these data reveals that in all cases (2-4) the barrier to pyramidal inversion at silicon is decreased relative to that calculated for 1, in accord with our qualitative predictions. Selected optimized structural parameters for 1-4 are reported in Figures 1 (3-21G) and 2 (6-31G*). As mentioned previously, while we report 3-21G data (Figure 1) for comparative purposes, the focus of our discussion will be based on the 6-31G* results (Figure 2). For 1 we calculate a 6-31G* inversion barrier of 34.7 kcal/mol. This value is at the higher end of the range of inversion barriers that have been calculated for 13a9b94e and is in very good agreement with the 34.3 kcal/mol barrier calculated by Hopkinson and Lien.3e The 6-3 1G* structure of 1 in the C,, configuration has Si-H bond lengths of 1.539 8, and H-Si-H angles of 95.4'. This structure
(1 1) Pyramidality at silicon can also be measured by the out-of-plane angle, 6.' The values of 6 obtained at the 6-31G*level for 1-4 are 1, 81.9'; 2, 83.1'; 3, 80.6O; 4, 44.0'.
36 The Journal of Physical Chemistry, Vol. 92, No. 1, 1988
280
290
Sum of
300
320
310
330
340
Angles About Silicon (degrees)
Figure 3. Plot of the calculated 6-31G* inversion barrier for 1-4 and the silacyclopentadienyl anion vs the pyramidality at silicon. The pyramidality at silicon is measured by the sum of the bond angles about the silyl anion center. The curve in this plot is obtained by visual inspection. TABLE 111: Pyramidal Inversion BarriersY
compd SiH,-
silyl anions inversion barrier
26.0,b,d 27.3,b3e27.5,bJ 33.8,b1g34.3,b,h34.7,b3g 39.6c,' SiH,SiH2- 26.6b39 HCOSiHT 23.8b*g C4H,SiH16.2bJ BH2SiH2- 0.7b,g
phosphorus compds inversion barrier compd PH3
37.2,b.k36.0b,'
SiH,P(CH,)* HCOPH2 CdH4PCHj BH2PH,
18.7c,' 25.gbJ 16.7"' 8.Ib"
"In kcal/mol. b A b initio. bCNDO/2. dSee ref 3d. ' S e e ref 3b. /See ref 8a. gThis work. hSee ref 3e. ' S e e ref 3a. ]See ref 3g. &See ref 6. 'See ref 10. " S e e ref 14.
pyramidal inversion barriers for 1-4 and the silacyclopentadienyl anion3g vs the pyramidality at silicon is shown in Figure 3. As inspection of this figure shows, there is a clear correlation between the calculated pyramidality at silicon and the inversion barrier.I2 While data in the ca. 295-330' and ca. 5-15 kcal/mol range are absent, making extensive interpretation of the data difficult, the curve one would obtain from visual in~pection'~ of the data is very similar to that which has been observed for amines.' Before this observation can be placed into proper perspective, however, it is important to compare the results obtained for the silyl anions with those previously reported for phosphines. Comparison of Silyl Anions with Isoelectronic Phosphines. Since silyl anions and phosphines are isoelectronic, one might (12) The question of the necessity of diffuse functions for the accurate description of anions has been debated: ref 3f; Spitznagel, G.; Clark, T.; Chandrasekhar, J.; Schleyer, P. v. R. J . Comput. Chem. 1982, 3, 363; Brenstein, R.J.; Scheiner, S . Int. J . Quantum Chem. 1986, 29, 1191. In order to test the effect of diffuse functions on these silyl anions, we performed calculations on the silyl anion (1) and the borylsilyl anion (4) at the 6-31+G* level (full geometric information is contained in the supplementary material). At this level, the pyramidal conformation is the ground state for both 1 and 4, and the calculated inversion barriers are 28.8 and 0.1 kcal/mol, respectively. Thus, these quantitative differences do not alter the qualitative conclusions reported in this work. (1 3) The curve in Figure 3 was estimated based on the analogous curve reported for amines.'
Damewood and Hadad expect that pyramidal inversion barriers would be similar in the two systems. In order to test this possibility, we have assembled data on the calculated inversion barriers for silyl anions and closely corresponding phosphines in Table 111. Inspection of this table shows that the inversion barrier for the parent compounds are somewhat similar, although several calculations have predicted lower barriers for 1 than for phosphine. As shown in this table, the silyl, formyl, and boryl substituents lower the barrier to pyramidal inversion in both phosphines and silyl anions. In addition, m e t h y l p h ~ s p h o l e(C4H4PCH3) ~~ and the silacyclopentadienyl anion3g have very similar calculated inversion barriers (16.7 and 16.2 kcal/mol, respectively), both representing a significant reduction in the barrier relative to the parent compounds. It is interesting to note the relative ordering of pyramidal inversion barriers for both species. For example, the silyl-substituted phosphine has a lower inversion barrier (1 8.7 kcal/mol) than formylphosphine (25.8 kcal/mol), whereas the opposite is true in the case of 2 (26.6 kcal/mol) and 3 (23.8 kcal/mol) (Table 111). This difference in ordering may suggest that interaction with r-acceptor substituents in the silyl anions is more substantial than in the phosphines. This hypothesis is supported by the 8.1 kcal/mol barrier calculated for pho~phinoboranel~ and the 0.7 kcal/mol calculated for 4. In order to investigate the possibility that conjugation with the a-acceptor is greater in the case of silicon vs phosphorus, we calculated the barrier to rotation in 4. In these calculations we considered rotation in the pyramidal ground state (4a)as well as in the planar (4b)form. The barriers so calculated are 13.7 and 90.1 kcal/mol, respectively. Since the corresponding ' ~ can barriers in phosphinobcrane are 6.6 and 40.4 k ~ a l / m o l ,we conclude that conjugation with the a-acceptor center is more significant in the case of silyl anions than for phosphines. This observation offers a possible explanation for the difference in the relative conformational energy ordering of these two systems. Further tests of this hypothesis await experimental efforts. In conclusion, we find that silyl anions, like amines,'show a clear correlation between ground-state pyramidality and pyramidal inversion barriers (Figure 3). As found for the isoelectronic phosphines, the barrier to inversion in 1 is reduced significantly by the silyl, formyl, and boryl substituents. While the boryl substituent is able to lower the barrier to pyramidal inversion in both phosphines and silyl anions, a significant decrease in pyramidality is only observed for the latter. We attribute this observation to the increased overlap between boron and silicon in the borylsilyl anion relative to boron and phosphorus in phosphinoborane (see above). With information on the structure and pyramidal inversion barriers of 1-4 and the silacyclopentadienyl a n i ~ n , ~weg are able to estimate the shape of the pyramidality-inversion barrier curve for silyl anions (Figure 3).
Acknowledgment. Acknowledgment is made to the donors of the Petroleum Research Fund, administered by the American Chemical Society, and the University of Delaware Honors Program for support of this research. Registry No.
1, 15807-96-2; 2, 53199-49-8; 3, 88810-38-2; 4,
78450-48-3.
Supplementary Material Aoailable: Coordinates and bonding parameters for all calculated structures (23 pages). Ordering information is given on any current masthead page. (14) Gropen, 0. J . Mol. Struct. 1977, 36. 11 1
