вход по аккаунту


Bond Stretch Isomerism in Bicyclo[1.1.0]tetrasilanes. Contrasts Between Strained Silicon and Carbon Ring Systems

код для вставкиСкачать
group in 14c has been introduced for two reasons: firstly,
for the later conversion of the photolactones 4 and 6 into
the stereoisomeric dienone lactones and, secondly, to
achieve regioselective Wessely acetoxylation. The oxidation of 14c with Pb(OAc),, however, must then be catalyzed by Lewis acid: with BF3.ether in a mixture of methanol and ethyl acetate["I the o-quinol acetates 1 and 3 are
isolated in 85% yield.
To sum up: the synthesis of 8 shows, for the first time,
how useful photolactonization for the synthesis of macrolides really is and, once again, that in a cyclic system the
information of a remote center of chirality can be transmitted to a high degree to a newly developing stereogenic center.
Received: December 22, 1986;
revised: January 26, 1987 [Z 2021 IE]
German version: Angew. Chem. 99 (1987) 363
[ I ] G. Quinkert, N. Heim, J. W. Bats, H. Oschkinat, H. Kessler, Angew.
Chem. 97 (1985) 985; Angew. Chem. I n f . Ed. Engl. 24 (1985) 987.
[2] G. Quinkert, G. Fischer, U.-M. Billhardt, J. Glenneberg, U. Hertz, G.
Durner, E. F. Paulus, J. W. Bats, Angew. Chem. 96 (1984) 430; Angew.
Chem. Inf. Ed. Engl. 23 (1984) 440.
131 Diene ketenes from o-quinol acetates: G. Quinkert, E. Kleiner, B.-J.
Freitag, J. Glenneberg, U.-M. Billhardt, F. Cech, K. R. Schmieder, C.
Schudok, H.-C. Steinmetzer, J. W. Bats, G. Zimmermann, G. Durner, D.
Rehm, E. F. Paulus, Helu. Chim. Acfa 69 (1986) 469.
[4] After the configuration of the lower homologues of 4 and 6 containing
a 16-membered ring had been determined by single crystal X-ray structural analysis IS], it was possible to differentiate 4 and 6 NMR spectroscopically by analogy.
[S] G. Quinkert, U:M. Billhardt, H. Jakob, G. Fischer, J. Glenneberg, P.
Nagler, V. Autze, N. Heim, M. Wacker, T. Schwalbe, Y. Kurth, J. W.
Bats, G. Durner, G. Zimmermann, H. Kessler, Helu. Chim. Acta. in
161 The azido anion is a better leaving group than the alkoxide anion of the
lactone group and a poorer leaving group than the dienolate anion of
the enol acetate group.
[7] After 90 min irradiation with light of the wavelength region >340 nm in
hexane a stationary composition of (E,E)-, (E.3-, (2.E)-and (2.2)-stereoisomers in the ratio 10 : 12 :2 : 1 is established. The (E.9-stereoisomer
5 can be separated quantitatively by flash chromatography. After repeating the irradiation-chromatography cycle three times the total yield
of 5 amounts to 72%.
[S] The reagent was prepared in situ from diisobutylaluminum hydride and
in toluene at 0°C; see S. Iguchi, N. Nakai, M. Hayashi, H. Yamamoto, M. Maruoka, Bull. Chem. SOC.Jpn. 54
(1981) 3033.
191 Besides 7 the C-6 epimer is also formed (ratio 250 :I).
[lo] X-ray structure analysis of roc-7b: a=8.5357(8), b = 18.582(3),
c = 12.715(1)
8=91.605(8)", V=2015.9(7) A'; monoclinic; P2,/c;
2 = 4 ; p= 1.109 g/cm'; Enraf Nonius CAD4 diffractometer; CuKnradiation; quadrant lo 2 6 = 1 lo", 2388 independent reflections with
I > n ( I ) ; structure determination by direct methods. R(F)=0.039
wR(F)=0.032; SDP program system.-Further details of the crystal
structure investigation are available on request from the Fachinformationszentrum Energie, Physik, Mathematik GmbH, D-7514 EggensteinLeopoldshafen 2 (FRG) on quoting the depository number CSD-52 166,
the names of the authors, and the journal citation.
[ 1I] The more or less planar dienone lactone ring moiety subdivides threedimensional space into two half spaces. The polymethylene ring moiety
is in one half space; in the other half space reduction takes place by the
bulky reagent.
1121 W.C. Still was the first to point out that in the case of medium and large
ring systems, merely a methyl group at a center of chirality sufficed to
control a newly formed stereogenic center by favoring a certain conformation; see W. C. Still, 1. Galynker, Tetrahedron 37(1981) 3981.
[13] The compound rac-7b comes from the racemic series which was used for
optimization; see N. Heim, Dissertation, Universitat Frankfurt am Main
(FRG) 1987.
1141 J. P. Yardley, H. Fletcher 111, Synthesis 1975, 244.
[IS] A. Butenandt, E. Hecker, M. Hopp, W. Koch, Justus Liebigs Ann. Chem.
658 (1962) 39.
[16] B. Seuring, D. Seebach, Helu. Chim. Acfa 60 (1977) 1175.
[I71 For oxidation with Pb(OAc),, catalyzed by BF3.ether, see U.-M. Billhardt, Dissertation, Universitat Frankfurt am Main (FRG) 1985.
0 VCH Verlagsgesellschaft mbH. 0-6940 Weinheim. 1987
Bond Stretch Isomerism in BicycloIl.l.Oltetrasilanes.
Contrasts Between Strained Silicon and Carbon
Ring Systems**
By Paul von Rague &hieyer,* Alexander F. Sax,
Josef Kalcher, and RudolfJanoschek
Dedicated to Professor Tilman J. De Boer, Amsterdam
The strains in small ring silicon systems afford interesting contrasts with their carbon counterparts:"] Three-membered silicon rings are more strained and four-membered
silicon rings less strained than their carbocyclic analogs.",*' We now show how these differences may lead to
bond stretch
in bicycle[ I. 1 .O]tetrasilanes. Although many such possibilities have been considered in
carbon ring system^,'^.^^ no example of isomers which differ principally in the length of a central bond has been
demonstrated experimentally.
Masamune et al. have synthesized and determined the
X-ray structure of a highly strained bicyclo[l. 1.O]tetrasilane derivative, l.lSP,bl
Although the ring inversion is extremely facile (barrier ca. 15 kcal/mol), the central bond
length, 2.373(3) A, is not very much longer than that found
1 R=2.6-Et&H,
for the other Si-Si bonds, which average 2.322
In this
respect, bicyclotetrasilane is similar to bicyclobutane 2
(where the two types of CC bonds have the same, nearly
normal lengths),"] rather than to 2,4-disilabicyclo[ 1.1.O]bu-
1 d61
tane 3, where both experiment and calculations find- the
central C-C bond to be exceptionally long (ca. 1.78 A)."]
Since other heteroatom substituted bicyclobutanes have
abnormally long central bonds,['"' we have examined the
nature of Si4H6,the parent compound of 1, calculationallY.
The results are quite surprising: At all levels of theory
employed, two distinctly different minima were found:
411a1 corresponds closely to Masamune's experimental
structure, and 5 , which is lower in energy, to a geometry
[*] Prof. Dr. P. von R. Schleyer
Institut fur Organische Chemie der Universitat Erlangen-Nurnberg
Henkestrasse 42, D-8520 Erlangen (FRG)
Dr. A. F. Sax, Dr. J. Kalcher, Prof. Dr. R. Janoschek
Institut fur Theoretische Chemie der Universitat
Mozartgasse 14, A-8010 Graz (Austria)
This work was supported by the Deutsche Forschungsgemeinschaft and
the Fonds der Chemischen Industrie. We thank Alexander Kos and
Bernd Reindl, Erlangen, for carrying out some of the calculations, and
Prof. Dr. W. W. Schoeller, Bielefeld, for information prior to publication
and an unknown referee for suggestions.
0570-0833/87/0404-0364 $ 02.50/0
Angew. Chem. Inr. Ed. Engl. 26 (1987) No. 4
calculated recently for Si4H6 by Dabisch and Schoeller.[81
These are "bond stretch" isomers; the bridgehead Si-Si
distance in 5 exceeds that in 4 by ca. 0.5 A (Table 1).
1i 7 1
' t'
The Si4H6 structures were optimized at the all electron
level with the 3-21G (split valence) and d-orbital augmented 3-21G* (polarized) basis
as well as with the
pseudopotential multiconfiguration-self consistent field
(MC-SCF) method, using a triply split but unpolarized valence basis set.Ilo."l Of the five configurations employed
for the MC-SCF treatment, only two proved to be of major
importance, namely yb=...(5a,)'
and ya==...(3b,)2;
whereby 5a, is the bonding M O of the Si-Si bridge, and
3b, is the corresponding antibonding MO.
At all three levels two optimized structures, 4 and 5 ,
were found. Both have CZvsymmetry; the main differences
are not only in the Si-Si distances mentioned above, but
also in the interflap angles between the three-membered
ring moieties (ca. 120" vs. 140") and in the Si-Si-H angles
at the bridgehead positions (ca. 140" vs. 90'). The key geometrical parameters of 4 and 5 are compared in Table 1,
along with the corresponding experimental values for 1
When one considers that the latter involves heavy substitution and that polarized basis sets are needed for the best
results with second row molecules,[91 the agreement between the X-ray geometry and the 3-21G* data for 4 is
quite satisfactory.
However, our data indicates structure 5 to be lower in
energy than 4. The difference depends on the theoretical
level, but is 2.6 kcal/mol with the 6-31G* basis set (3-21G*
geometries) and increases to 8.4 kcal/mol with inclusion of
electron correlation via perturbation theory (MP2/63 IG*//3-21G*). The multiconfiguration treatment (MCSCF), which should be better suited for this purpose, gave
similar geometric parameters (Table 1) and an even greater
energy difference, 17 kcal/mol, favoring 5 over 4 . While
part of this further increase may be due to the lack of dorbitals in the MC-SCF basis set, 5 should be considerably
more stable than 4. Furthermore, the energy barrier separating 4 from 5 is calculated to be quite small, l kcal/mol
o r less. Hence, 4 is not likely to be a viable isomer for
Si4H6 itself. The MC-SCF lowering of the relative energy
of 5 occurs due to considerable admixture of Yainto the
total wave function, Y,,( 5 ) = 0.95 yb( 5 ) - 0.30 Ya(5 ) ,
which reflects its diradicaloid character. In contrast, the
wave function of structure 4 is well represented by yb(4)
only, Y,,,,(4) = 0.98 Yb(4)- 0.12Ya(4). Additional consideration of singly and doubly substituted configurations
(CI-SD) does not change these results. Thus, the electronic
structure of 4 is similar to that of bicyclobutane (2).[6b,c1
The central Si,Si3 bond of 4 has considerable n character
(as in 2 ) and Mulliken populations indicate the Si,Si3 interaction to be antibonding. The bond stretch isomer, 5 ,
has greater Si,Si, antibonding character, and can be regarded to be a singlet diradical. The nature of this species
has been discussed by Schoeller and Dabisch.[slAs in SiH:,
the bond angles at the bridgehead positions in 5 are close
to 90". This results in a rather close approach of the
bridgehead hydrogens. Although 5 is lower in energy than
4 for Si4H6, this order will be reversed if bulky substituents, instead of hydrogens, are present at the bridgeheads. We have shown this calculationally by replacing the
two bridgehead H atoms in 4 and 5 by methyl groups (but
otherwise retaining the Si4H6geometries); this reverses the
relative stabilities. The even larger tBu-bridgehead substituents in 1 cannot be accommodated in a structure of type
5 at all; the 4-like geometry of 1, in which the fBu groups
are separated due to the large RSiSi angles (Table l), results as a consequence.
Si4H6 + 5Si2H6
+ nSi,H6
2i-Si4Ht0+ 2Si3H8
The strain energy of bicyclobutane (2) is 66.5 kcal/
mol.1'2,13b1 A somewhat higher value (70.6 kcal/mol, 321G*)['aJis estimated theoretically for 4 by means of
h o m o d e s m ~ t i c [ equation
(a). This reflects the generally
higher strain energy of three-membered silicon over the
Table I . Key geometrical parameters (bond lengths [A] and angles ["I) of bicyclotetrasilane species, X-ray ( 1 ) and theoretical (4 and 5)
Data source
X-ray [a]
H F/3-2 1G
MC-SCF pseudopotential
HF/"STO-4/3 1 G" [c]
MC-SCF pseudopotential
2.322 [b]
146.9 [b]
108.8 [b]
I 10.9
I 1 1.3
[a] See ref [Sb]. Angles to substituents, instead of to H, are involved. [b] Averaged. [c] See ref. [8].
Angew. Chem. Int Ed. Engl. 26 (1987) No. 4
0 VCH Veriagsgeseilschaft mbH. 0-6940 Weinheim. 1987
0570-0833/87/0404-0365 $ 02.50/0
Table 2. Comparison of strain energies [kcal m o l '1
~ of carbon and silicon small ring systems [a].
3-21G (sp basis)
3-21G' (spd basis)
Pseudopotential (sp basis)
Pseudopotential (spd)
Experiment [d]
(CH2)4, ( D 2 J
(Si H A , (DZJ
40.3 [c]
Si4Hbr4 [bl
[a] Unless indicated otherwise, see [I] for the silicon and [I21 for the carbon data. Based on equation (b). [b] This work. Based on equation (a). [c]Other values: 39.6
(6-31G*//6-31G*); 38.2 (MP2/6-3lG*//6-3lG*); seella]. Id] Ref. 1121. Theoretical calculatlons (6-31G* basis set) give nearly identical values; see K. B. Wiberg, J .
Comput. Chem. 5 (1984) 197
13) W. D. Stohrer, R. Hoffmann, J. Am. Chem. SOC.94 (1972) 779, 1661.
corresponding carbon ring systems (Table 2).1'.'4J However, cyclotetrasilane is less strained than c y c l ~ b u t a n e . ~ ' , ~ ~ [41 M. N. Paddon-Row, L. Radom, A. R. Gregory, J . Chem. SOC.Chem.
Commun. 1976, 477. A. R. Gregory, M. N. Paddon-Row, L. Radom, W.Hence, the opening of bicyclobutane (2) to a hypothetical
D. Stohrer, Aust. J . Chem. 30 (1977) 473.
four-membered ring diradical bond stretch isomer would
IS] a) S. Masamune, Y. Kabe, S. Collins, D. J. Williams, R. Jones, J. Am.
Chem. SOC. 107 (1985) 5552; b) R. Jones, D. J. Williams, Y. Kabe, S .
involve less strain relief than the corresponding change
Masamune, Angew. Chem. 98 (1986) 176; Angew. Chem. lnt. Ed. Engl.
from 4 to 5 : roughly 40 kcaI/mol for 2 vs. 52 kcal/mol for
25 (1986) 173; c) Si-Si bond lengths in unstrained organosilicon com4 (this assumes that the four membered ring diradicals
pounds normally afe 2.34-2.35 A, although steric effects can produce
have the same strain energy as cyclo-Si4H, and cycZo-C4H,,
lengthening to 2.7 A: N. Wiberg, H. Schuster, A. Simon, K. Peters, Angew. Chem. 95 (1986) 100; Angew. Chem. Int. Ed. Engl. 25 (1986) 79. For
respectively). A similar treatment has been used by Wiberg
a discussion of Si. . . Si nonbonded distances, see M. J. Michalczyk, M.
et al. to analyze ring opening processes in strained polyJ. Fink, K. J. Haller, R. West, J. Michl, Organome!allics 5 (1986) 531.
cyclic ring systems.['51This difference in ring-opening ther[61 a) P. H. M. Budzelaar, E. Kraka, D. Cremer, P. von R. Schleyer, J. Am.
modynamic driving force and the inherently lower Si-Si
Chem. Soc. 108 (1986) 561, and references cited therein; b) M. D. Newton, J . M. Schulrnan, J. Am. Chem. SOC.94 (1972) 767; c) reviews: F. H.
bond dissociation energy (74 kcal/mol in H,Si-SiH3 vs. 88
Allen, Tetrahedron 38 (1982) 645; S. Hoz, ibid.. in press; and ref. I13bl.
kcal/mol in H3C-CH3)1161are responsible for the prefer[71 a) G. Fritz, S. Wartanessian, E. Matern, W. Honte, H. G. von Schnering,
ence of Si4H6 for the "bond-stretch'' structure ( 5 ) in con2. Anorg. Allg. Chem. 475 (1981) 87; b) the 3-21C'" geometry of 2.4trast to C& which favors a geometry with normal bond
disilabicyclobutane has r(CC) 1.754 A,r(SiC) 1.829 A,and an interflap
angle of 125.8" (P. von R. Schleyer, unpublished calculations).
lengths (2). These two factors also contribute to the
I81 T. Dabisch, W. W. Schoeller, J. Chem. SOC.Chem. Commun. 1986.
smaller inversion barrier observed for the bicyclotetrasi896.
lane ( 1)[5b1
than for bicyclobutane (2).lT7,'*I
191 The Gaussian 82 program was used: see W. J. Hehre, L. Radom, P. von
We conclude that 1 only adopts a structure of type 4
R. Schleyer, J. A. Pople: Ab lnirio Molecular Orbital Theory. Wiley, New
York 1986.
due to the spacial requirements of the bulky bridgehead
[lo] R. Janoschek, A. Sax, E. A. Halevi, Isr. J. Chem. 23 (1983) 58. A. F. Sax,
substituents. If these can be replaced experimentally by
J. Comput. Chem. 6 (1985) 469.
hydrogens o r by a bridging ring, we predict that a geome[l I] H.-J. Werner, W. Meyer, J. Chem. Phys. 74 (1981) 5794; H.-J. Werner,
try corresponding to the "bond-stretch" isomer 5 will reL A . Reinsch, ibid. 76 (1982) 3144.
1121 P. von R. Schleyer, J. E. Williams, K. R. Blanchard, J. Am. Chem. Soc.
sult. Indeed, our preliminary calculations indicate pentasi92 (1970) 2377.
la[l.l.l]propel~aneto have only a single minimum with a
1131 a) P. George, M. Trachtman, C. W. Bock, A. M. Brett, Tetrahedron 32
large (ca. 2.9 A) separation between the bridgehead atoms.
(1976) 317; b) A. Greenberg, J. F. Liebman: Strained Organic Molecules.
Received: July 21, 1986;
revised: January 2, 1987 [ Z 1867 IE]
German version: Angew. Chem. 99 (1987) 374
[I] a) P. von R. Schleyer, NATO AS1 Ser. C189 (1986) 69: b) A. F. Sax,
Chem. Phys. Lett. 127 (1986) 163; c) A. F. Sax, ibid. 129 (1986) 66.
[ 2 ] W. Schoeller, T. Dabisch, J . Chem. SOC.Chem. Commun. 1985, 1706.
The strain energies reported in this paper were not evaluated by means
of homodesmotic equations (see ref. [13a]) and are too low: however,
their subsequent use gives results in accord with those in ref. [I] and
Table I (W. W. Schoeller, private communication).
0 VCH Verlagsgesellschafi mbH, 0-6940 Weinheim. 1987
Academic Press, New York 1978.
[14] D. Cremer, J. Gauss, J. Am. Chem. SOC.108 (1986) 7467.
1151 K. B. Wiberg, F. H. Walker, J. Am. Chem. SOC.104 (1982) 5239.
[I61 R. Walsh, Acc. Chem. Res. 14 (1981) 246.
[I71 a) R. B. Woodward, D. L. Dalryrnple, J . Am. Chem. SOC.91 (1969) 4612:
b j for a GVB-PRDDO calculation, see P. G. Gassman, M. L. Greenlee,
D. A. Dixon, S. Richtsmeier, J. Z. Gougoutas, ibid. 105 (1983) 5865.
I181 The transition structure for inversion of Si4H6 needs a n MC-SCF Ansatz
for the wave function [cf. 17bJ: preliminary calculations indicate a transition structure of lower symmetry than DZh.This is corroborated by the
U H F optimized triplet state, which shows a non-planar silicon skeleton:
P. von R. Schleyer and A. F. Sax, unpublished.
0570-0833/87/0404-0366 $ 02.50/0
Angew. Chem. lnt. Ed. Engl. 26 (1987) No. 4
Без категории
Размер файла
337 Кб
bond, stretch, bicycle, isomerism, contrast, silicon, ring, system, tetrasilanes, carbon, strained
Пожаловаться на содержимое документа