Experiments towards the formation of 1,6-dehydroquadricyclane and density functional calculations on this and related molecules

Article abstract of DOI:10.1002/(SICI)1099-0690(199803)1998:3<493::AID-EJOC493>3.0.CO;2-N

1,6-Dibromoquadricyclane (6) was obtained from norbornadiene (11) by hydroboration, oxidation of the diol 12 to the diketone 14 and its conversion into 2,6-dibromonorbornadiene (20) using tribromodioxaphosphole 16b followed by treatment of the mixture 17/18 with potassium tert-butoxide in DMSO and photocyclization of 20. Reaction of 6 with tBuLi (2 equiv.) led to the formation of 1-bromo-6-lithioquadricyclane 7, the NMR spectra of which were observed up to 0°C. 7 did not lose LiBr to give 4, but could be trapped with H₂O and chlorotrimethylsilane to give 21e (53%) and 21f (64%). Reaction of 6 with fBuLi (> 4 equiv.) gave rise to 1,6-dilithioquadricyclane (21c), whose NMR spectra could also be recorded. 21c was converted into the corresponding 1,6-disubstituted quadricyclanes with D₂O (87%), chlorotrimethylsilane (92%), dimethyl sulfate (55%), methyl chloroformate (45%), iodine monochloride (62%), and p-toluenesulfonyl chloride (48%). - Density functional calculations using the B3LYP/6-31G* level of theory showed that 1,6-dehydroquadricyclane (4) is a local energy minimum in its singlet electronic state. 4 contains a unique structure with 4 condensed cyclopropane units. The parent hydrocarbons 27 and 28, hitherto unknown, are also local energy minima in their singlet electronic states.

Full text of DOI:10.1002/(SICI)1099-0690(199803)1998:3<493::AID-EJOC493>3.0.CO;2-N

FULL PAPER
Experiments Towards the Formation of 1,6-Dehydroquadricyclane and Density
Functional Calculations on This and Related Molecules[1]
Susanne Glück-Walthera, Oliver Jaroschb, and Günter Szeimies*b
Institut für Organische Chemie, Universität Münchena,
Karlstraße 23, D-80333 München, Germany
Institut für Chemie, Humboldt-Universität zu Berlin,b
Hessische Straße 1Ϫ2, D-10115 Berlin, Germany
Received September 22, 1997
Keywords: Strained molecules / Quadricyclanes / Lithiation / Density functional calculations
quadricyclane (21c), whose NMR spectra could also be re-
1,6-Dibromoquadricyclane (6) was obtained from norborna- corded. 21c was converted into the corresponding 1,6-disub-
diene (11) by hydroboration, oxidation of the diol 12 to the
stituted quadricyclanes with D2O (87%), chlorotrimethylsi-
diketone 14 and its conversion into 2,6-dibromonorborna- lane (92%), dimethyl sulfate (55%), methyl chloroformate
diene (20) using tribromodioxaphosphole 16b followed by
treatment of the mixture 17/18 with potassium tert-butoxide
in DMSO and photocyclization of 20. Reaction of 6 with tBuLi
(2 equiv.) led to the formation of 1-bromo-6-lithioquadricyc-
lane 7, the NMR spectra of which were observed up to 0°C.
7 did not lose LiBr to give 4, but could be trapped with H2O
(45%), iodine monochloride (62%), and p-toluenesulfonyl
chloride (48%). Ϫ Density functional calculations using the
B3LYP/6-31G* level of theory showed that 1,6-dehydroqua-
dricyclane (4) is a local energy minimum in its singlet elec-
tronic state. 4 contains a unique structure with 4 condensed
cyclopropane units. The parent hydrocarbons 27 and 28, hi-
and chlorotrimethylsilane to give 21e (53%) and 21f (64%). therto unknown, are also local energy minima in their singlet
Reaction of 6 with tBuLi (> 4 equiv.) gave rise to 1,6-dilithio-
electronic states.
Lithium halide elimination has proved as an efficient
method for the generation of strained olefins[2] and hydro-
carbons[3]. This reaction has furnished evidence for the for-
mation of 1,7-dehydroquadricyclane (1)[4][5], 1,5-dehydro-
quadricyclane (2)[6] and 1,2-dehydroquadricyclane (3)[7]
,
which could be trapped as Diels-Alder adducts. Ab initio
MO calculations at the TCSCF level of theory using the
6-31G* basis set showed that 1, 2, and 3 were local energy
minima on the C7H6 energy hypersurface, as was the non-
olefinic 1,6-dehydroquadricyclane (4)[7]. Taking 1 as refer-
ence model compound, 2, 3, and 4 were calculated to be
less stable by 11.8, 4.0, and 12.7 kcal/mol. As 2 was formed
by LiBr elimination from 5, it seemed justified to synthesize
dibromide 6 and convert it to 7, to find out if 4 is exper-
imentally accessible.
A molecule structurally related to 1,6-dehydroquadri-
cyclane 4 is 2,4-dehydrohomocubane (8), which has been
invoked as an intermediate in the reaction of dihalide 9b
with tert-butyllithium (tBuLi)[8]. 8 was able to add tBuLi
across the strained CϪC single bond to afford 10 after
aqueous workup. A similar behavior could be expected of 4.
A. Synthesis of 1,6-Dibromoquadricyclane (6)
9-borabicyclo[3.3.1]nonane or the borane dimethyl sulfide
The synthetic procedure for the preparation of 6 started complex as reducing agents did not lead to better results.
from norbornadiene (11). Its conversion to 2-exo-5-syn-bi- 1,5-cyclooctanediol in the first and dimethyl sulfoxide in the
cyclo[2.2.1]heptanediol (12) by hydroboration in THF has second case could not be separated from 12 easily.
been reported by Zweifel, Nagase, and Brown[9]. In our
The conversion of 12 to bicyclo[2.2.1]heptane-2,5-dione
hands, the yields of 12 did not exceed 35%. At a yield of (14) in 41% yield by Jones oxidation has been reported by
26%, the mono-alcohol 13 was a major side-product. Using Hawkins, Hsu, and Wood[10]. Although this yield could be
Eur. J. Org. Chem. 1998, 493Ϫ500
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1434Ϫ193X/98/0303Ϫ0493 $ 17.50ϩ.50/0
493

S. Glück-Walther, O. Jarosch, G. Szeimies
FULL PAPER
reached closely (39%), alternative methods for this oxi- the expected trapping product. This result implies that 1,6-
dation were investigated. The procedure of Swern and dilithioquadricyclane (21c) was generated in the course of
Omura[11] using DMSO/oxalyl chloride in CH2Cl2 fur- the reaction. Therefore, 6 was treated with 4.0 equiv. of
nished 14 in 35%, whereas the “oxon” method of Mori- tBuLi followed by addition of D2O, which led to the iso-
moto[12] was not successful.
lation of 1,6-dideuterioquadricyclane (21d) in 87% yield
and a D2 content of > 90%. In a further experiment, 6 was
reacted with only 2.0 equiv. of tBuLi at Ϫ78°C in ether;
after 2.0 hours of reaction time at 0°C, aqueous workup
resulted in the isolation of 1-bromoquadricyclane (21e) in
53% yield. To promote LiBr elimination from 7, which is
obviously present in the reaction mixture, the reaction was
repeated by addition of tetramethylethylenediamine
(TMEDA) in the first and by lithium bromide in the second
modification, but in both cases only 21e could be isolated.
The reported results indicated that 1-bromo-6-lithioqua-
dricyclane (7) is persistent up to 0°C and does not eliminate
LiBr. In the presence of tBuLi, a second lithium bromine
exchange takes place under the formation of 21c. This
could be confirmed by 13C-NMR spectroscopy. For this
purpose dibromide 6 in [D8]THF was treated at Ϫ78°C
with 2.0 equiv. of tBuLi, spectra of the solution were re-
corded at Ϫ80, Ϫ60, Ϫ40, and 0°C. At all temperatures
clear signals were obtained for all C atoms of 7. C-1 and
C-6 appeared as singlets at δ ϭ 34.50 and 11.58. At 25°C,
the spectrum became unstable and many new signals ap-
peared. When instead of 2.0 equiv. 4.0 equiv. of tBuLi were
used, 13C-NMR spectra of 21c in [D8]THF were recorded
at Ϫ80, Ϫ40, and 0°C, again showing intensive stable sig-
nals at these temperatures. The signal for C-1 and C-6 ap-
peared at δ ϭ 16.75; at 25°C, this signal fainted quickly.
Two reaction paths were pursued to convert 14 into 20.
The Shapiro reaction[13] with bis-hydrazone 15 and 4.0
equiv. of n-butyllithium (BuLi) followed by addition of 1,2-
dibromoethane did not furnish any 20. Alternatively, 14 was
reacted with freshly prepared 2,2,2-tribromo-1,3,2-benzo-
dioxaphosphole (16b)[14] (from 16a and Br2) for 20 h in
CH2Cl2 under reflux, which afforded 2,5,5-tribromobi-
cyclo[2.2.1]hept-2-ene (17) in 61% yield and 2,2,5,5-tetra-
bromobicyclo[2.2.1]heptane (18) in 9% yield. In some
experiments, a small amount of 2,2,5-tribromonortricyclane
(19) (yield < 5%) was also observed. 17 and 18 could be
separated by distillation, the traces of 19 accompanying 18
were hard to remove. The mixture of 17 and 18 was effec-
tively converted into dibromide 20 in 67% yield by reaction
with potassium tert-butoxide in DMSO. Photocyclization
of 20 to 6 was accomplished in 81% yield by irradiating a
solution of 20 in ether in the presence of 5% of aceto-
phenone.
C. Synthesis of Some 1,6-Disubstituted Quadricyclanes
The persistency of 7 and 21c below 0°C encouraged the
reaction of these intermediates with some electrophiles. The
reaction of 7 with chlorotrimethylsilane afforded 21f in 64%
yield. When chlorotrimethylsilane, dimethyl sulfate, methyl
chloroformate, cyanogen iodide, or tosyl chloride were al-
lowed to react with 21c, the corresponding 1,6-disubstituted
B. Reactions of 6 with Organolithium Bases
Under the premise that 4 could be generated from 6 and quadricyclanes 23a؊e were formed in reasonable yields.
trapped by tBuLi, dibromide 6 was treated with an excess After distillation, the diester 23c was completely iso-
(5.0 equiv.) of tBuLi at Ϫ78°C for 2 hours, the mixture merized to dimethyl bicyclo[2.2.1]hepta-2,5-diene-2,5-di-
brought to 0°C and hydrolyzed with water. Besides some carboxylate (24c). This compound was reconverted into 23c
polymeric material, only quadricyclane (21a) could be iso- in 78% yield by irradiation in ether in the presence of aceto-
lated. There was no indication for the formation of 21b, phenone. The isolated 1,6-diiodoquadricyclane (23d) con-
494
Eur. J. Org. Chem. 1998, 493Ϫ500

1,6-Dehydroquadricyclane
FULL PAPER
The Becke3LYP results of Table 1 are consistent with our
earlier calculations in which we used the TCSCF theory[7]
.
Within these levels of theory, 1,7-dehydroquadricyclane (1)
is the most stable of the four C7H6 isomers and 4 is the
least stable. All molecules were local minima on the C7H6
energy hypersurface, as indicated by calculated real positive
vibrational frequencies. However, at 0.9 kcal/mol the differ-
ence in energy between 2 and 4 is only marginal at the
TCSCF/6-31G* level of theory, whereas within the B3LYP/
6-31G* methodology the energy difference rises to 5.1 kcal/
mol. This value indicates that formation of 4 by lithium
halide elimination is energetically a more demanding pro-
cess than the one of generating the olefinic dehydroquad-
ricyclanes 1, 2, and 3. If the 9.6 kcal/mol which 4 is higher
in energy than 1 suffices to prevent its formation, is less
clear.
tained 15% 2,5-diiodobicyclo[2.2.1]heptane (24d), which
could be removed by preparative layer chromatography. 24d
was also formed in the reaction of 21c with iodine, with
p-tolylsulfonyl iodide and with iodine monochloride in
yields of 34, 25 and 17%.
Table 1. B3LYP/6-31G*// B3LYP/6-31G* energies for dehydroqua-
dricyclanes 1؊4, 21a, 21h, 8, 9a, and 9c
The access to 1-bromo-6-trimethylsilylquadricyclane
(21f) and 1,6-diiodoquadricyclane (23d) prompted two
further experiments aimed at reaching the goal of generat-
ing 4. Compound 21f was treated with 8.0 equiv. of CsF in
DMSO[15] for 24 h at room temperature, but only unreacted
21f accompanied by some 2-bromo-5-trimethylsilylbicyclo-
[2.2.1]heptane (22f) could be isolated. 23d and 4 equiv. of
tBuLi afforded after reaction in ether at Ϫ78°C, warming
to 0°C and addition of D2O only 2,5-dideuteriobicy-
clo[2.2.1]heptane (22d) besides polymeric material. This re-
sult shows that 21g is formed as an intermediate, which
shares with 7 the inability to eliminate lithium halide.
Our experimental approach to generate 1,6-dehydro-
quadricyclane (4) did not succeed, because the activation
energy for the LiBr elimination from 7 is energetically too
costly. To overcome this problem, metal exchange of Li in
7 against a metal with lower bond energy to carbon like K,
Cs, or Ag, might be the method of choice. This route will
be followed in future activities in this field.
[b]
mole- B3LYP energy
ZPE[a]
Erel
TCSCF[c] Erel
cule
[au]
[au]
[kcal/mol]
[kcal/mol]
1
2
3
4
Ϫ270.1124892
Ϫ270.1047122
Ϫ270.1076397
Ϫ270.0961003
0.102422
0.101815
0.102053
0.101359
0.126320
0.105912
0.137185
0.161870
0.141148
0.0
4.5
2.8
9.6
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
0.0
11.8
4.0
12.7
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
21a Ϫ271.4448293
21h Ϫ737.9381907
8
Ϫ347.5212474
9a Ϫ348.8459265
9c Ϫ815.3380766
[a]
[b]
Scaled by a factor of 0.9804; see ref.[21]
Ϫ
B3LYP energy ϩ
ZPE, rel. to 1. Ϫ See ref.[7]
.
[c]
The TCSCF/6-31G* structures of 1؊3 differ only slightly
from the B3LYP/6-31G* structures. For the “unsaturated
center” of 1 the C1ϪC7 TCSCF distance was 1.3708 A, the
B3LYP distance was 1.3969 A. The corresponding values
for 2 (C1ϪC5) were 1.4292 A and 1.4030 A, and for 3
(C1ϪC2) 1.3656 A and 1.3719 A. A significant difference
˚
˚
˚
˚
˚
˚
was found in the C1ϪC6 bond length of 4. The TCSCF/6-
31G* calculation showed a CϪC distance of 1.9081 A,
D. Ab Initio Calculations on 1,6-Dehydroquadricyclane and
Related Molecules
˚
whereas the B3LYP/6-31G* value amounted only to
In this section, the questions concerning the existence of
1,6-dehydroquadricyclane (4) as a short-lived reaction inter-
mediate is approached by computational chemistry tech-
niques. The Gaussian 94 program package[16] has been used
throughout. Specifically, the density-functional theory
(DFT)[17] with the Becke3LYP functional[18] and the 6-
31G* basis set has been applied preferentially in this investi-
gation. This method has been used to calculate the energy
difference between singlet and triplet carbenes[19] and has
been applied successfully to solve mechanistic problems in
˚
1.7934 A.
1,6-Dehydroquadricyclane (4) is structurally related to
1,7-dehydroquadricyclane (1) by hydrogen shift from C-1 to
C-6. Furthermore, hydrogen shift from C-5 to C-6 would
convert 4 into 1,5-dehydroquadricyclane (2). We have been
able to locate the saddle points of these two reactions at
E ϭ Ϫ270.016137142 au (1 imaginary frequency at 1122.8i
cmϪ1, ZPE ϭ 0.097297 au) and at E ϭ Ϫ270.0262276 au
(1 imaginary frequency at 1284.2i cmϪ1, ZPE ϭ 0.099400)
using the B3LYP/6-31G* level of theory. IRC calculations
on both saddle points assured that they connected 4 with 1
and, respectively, 4 with 2. The transition structure of the 4
different areas of organic chemistry[20]
.
1. Energies and Structures of Dehydroquadricyclanes
˚
Ǟ 1 reaction showed a long C6ϪC7 bond of 2.0676 A. For
In Table 1 the B3LYP/6-31G*//B3LYP/6-31G* energies the 4 Ǟ 2 reaction, the C5ϪC6 bond length of the tran-
˚
are given for dehydroquadricyclanes 1؊4, for quadricyclane sition-state structure was calculated as 1.7479 A. The unex-
(21a), 1-chloro-6-lithioquadricyclane (21h), for homocub- pected structure of the first saddle point prompted us to
ane (9a), and 2-chloro-4-lithiohomocubane (9c).
recalculate the transition structure of the 4 Ǟ 1 reaction,
Eur. J. Org. Chem. 1998, 493Ϫ500
495

S. Glück-Walther, O. Jarosch, G. Szeimies
FULL PAPER
using the MP2/6-31G*//MP2/6-31G* level of theory. The Table 2. B3LYP/6-31G* selected structural data of singlet and tri-
plet 4
result was in accord with the B3LYP/6-31G* calculation:
The C6ϪC7 bond length amounted to 2.0036 A (E ϭ
˚
bond
bond length
angle
bond angle
Ϫ269.1206747 au, one imaginary frequency at 1074.4i
cmϪ1, ZPE ϭ 0.099038). From the above data, the B3LYP/
6-31G* potential energy barrier (including ZPE correction)
separating 4 from 1 is calculated to be 46.4 kcal/mol. At
41.4 kcal/mol, the energy barrier from 4 to 2 is somewhat
lower. The MP2/6-31G* barrier of the 4 Ǟ 1 reaction is
52.3 kcal/mol and compares reasonably well with its B3LYP
counterpart. The range of these barriers leaves no doubt
that under low-temperature conditions the conversion of 4
to 1 and, respectively, to 2 by hydrogen migration will not
be observed.
(singlet)
(triplet)
singlet
[°]
triplet
[°]
˚
˚
[A]
[A]
1Ϫ2
1Ϫ5
1Ϫ6
1Ϫ7
2Ϫ3
2Ϫ7
1.537
1.510
1.793
1.466
1.518
1.533
1.506
1.531
2.097
1.493
1.523
1.565
2Ϫ1Ϫ5
2Ϫ1Ϫ7
5Ϫ1Ϫ6
6Ϫ1Ϫ7
1Ϫ2Ϫ3
1Ϫ2Ϫ7
3Ϫ2Ϫ7
2Ϫ3Ϫ4
1Ϫ5Ϫ4
1Ϫ5Ϫ6
6Ϫ5Ϫ4
111.5
61.3
51.8
54.1
107.4
57.0
114.1
98.1
97.7
74.1
61.6
106.0
62.9
45.3
46.9
111.8
58.1
109.9
98.3
103.0
87.8
59.0
In the introductory section it was assumed that 2,4-de-
hydrohomocubane (8) might be a good model compound
for comparison with 1,6-dehydroquadricyclane (4). How-
ever, two isodesmic reactions (1) and (2) show that this is
not really true. Obviously, 4 has a considerably higher strain
energy, which makes realistic comparisons between 4 and 8
hardly possible. The 15 kcal/mol in exothermicity which is
shown by equations (1) and (2) should mainly originate
from the difference in strain energy of 4 and 8. If only parts
of this energy enter the activation barrier between 21h and
4, the LiCl elimination from 21h might be retarded to a
point where it is not observed under the experimental con-
ditions at 0°C.
strongly elongated. As expected, the C1ϪC6 bond for the
triplet is considerably longer (2.097 A).
˚
The long C1ϪC6 bond of the singlet of 4 rises the ques-
tion if the triplet state could be the ground state of 4. The
triplet energies of 1؊4 have been calculated using the UB3-
LYP/6-31G*//UB3LYP/6-31G* level of theory. The results
are given in Table 3, which also contains the singlet-triplet
energy difference in kcal/mol.
Table 3 shows that the triplet state at the Becke3LYP/6-
31G* level of theory is higher in energy by close to 20 kcal/
mol for 1 and 2, by 23 kcal/mol for 3 and by 10 kcal/mol
for the non-olefinic 1,6-dehydroquadricyclane (4). This is in
accordance with our earlier results[7]
.
Table 3. Becke3LYP/6-31G* energies for dehydroquadricyclanes
1؊4 triplets, and singlet-triplet energy differences
molecule
UB3LYP energy
[au]
ZPE[a]
[au]
∆ETϪS
[kcal/mol]
1
2
3
4
Ϫ270.0819443
Ϫ270.0732200
Ϫ270.0691118
Ϫ270.0795599
0.100954
0.100652
0.100495
0.100674
18.2
19.0
23.2
9.9
[a]
Scaled by a factor of 0.9804; see ref.[21]
.
2. Some Three-Membered Ring-Condensed Structures
At this point it should be noted that condensation of a
three-membered ring to a side-bond of bicyclo[1.1.0]butane
1,6-Dehydroquadricyclane (4), still awaiting experimental has been achieved experimentally within model compounds
birth, has a fascinating structure, which was calculated at 25[22] and 26b[23]. It seemed interesting to investigate theo-
the B3LYP/6-31G* level of theory. Selected structural data retically if the parent of 4, the tetracycle 27 and its stereo-
for the singlet as well as for the triplet are given in Table 2. isomers 28 and 29 were minima on the corresponding en-
Atoms C1, C5, C6, and C7 form a bicyclo[1.1.0]butane ergy hypersurface.
subunit which is nearly planar (dihedral angle 5617 ϭ
25 has been the subject of intensive ab initio MO calcu-
182.4° for the singlet and 179.7° for the triplet). At the side lations by Wiberg[22]. The most interesting feature of these
bonds C1ϪC7 and C5ϪC6 two three-membered rings are calculations on the structure of 25 at the HF/6-31G* and
condensed to the central bicyclo[1.1.0]butane, for the singlet MP2/6-31G* level of theory was the finding that this mol-
in the exo mode, leading to two additional bicyclo[1.1.0]- ecule had C1 symmetry with bond lengths of C1ϪC3 at
˚
˚
butane subunits C1,C2,C7,C6 and C1,C5,C6,C4. The di- 1.446 A and of C1ϪC4 at 1.577 A and a low energy barrier
hedral angles of these bicyclobutanes (1654 and 6172) (0.9 kcal/mol) between an identical molecule with inter-
amount to 108.2° for the singlet and 106.7° for the triplet. changed bond lengths of C1ϪC3 and C1ϪC4. We have ap-
˚
At 1.793 A, the central bond C1ϪC6 of the singlet is proached this problem with the B3LYP/6-31G*//B3LYP/6-
496
Eur. J. Org. Chem. 1998, 493Ϫ500

1,6-Dehydroquadricyclane
FULL PAPER
The strongly elongated bond length C1ϪC4 of 27 and 28
opens the possibility for a triplet ground state. As seen from
the data of Table 4, this is not the case. 25, 27, and 28 have
singlet ground states, with triplet states energetically well
above the singlet states.
Summing up, our calculations have revealed that 1,6-de-
hydroquadricyclane 4 is a molecule well in reach of exper-
imental verification and that the parent hydrocarbons of
4, the tetracycles 27 and 28, are interesting targets for the
synthetic chemist.
We thank the Deutsche Forschungsgemeinschaft and the Fonds
der Chemischen Industrie for financial support of this investigation.
Experimental Section
For analytical instruments and general procedures, see ref.[7]
.
31G* level of theory and found a similar result: 25 is a
minimum on the C5H6 energy hypersurface in a C1 struc- I. Starting Materials
ture. The bond lengths C1ϪC3 and C1ϪC4 were calculated
Bicyclo[2.2.1]heptadiene (11), N,N,NЈ,NЈ-tetramethylethylenedia-
mine (TMEDA), n-butyllithium (BuLi), tert-butyllithium (tBuLi),
potassium tert-butoxide (KOtBu), and chlorotrimethylsilane were
commercial products.
˚
˚
as 1.454 A and 1.519 A, the dihedral angles 2Ϫ1Ϫ3Ϫ4 and
5Ϫ1Ϫ4Ϫ3 amounted to 122.4° and 136.4°. This result was
also obtained using the QCISD/6-31G*//QCISD/6-31G*
formalism. The corresponding structural parameters were
2-Bromo-1,3,2-dioxaphoshole (16a)[14]: Catechol (110.6 g, 1.004
mol) was mixed with PBr3 (285.2 g, 1.05 mol) with stirring in an
ice bath under nitrogen. Water (1.0 ml) was added as a catalyst,
which caused a vigorous reaction with formation of gaseous HBr.
˚
˚
1.446 A and 1.554 A and 118.3° and 137.6°.
The calculations on the tetracycles 27؊29 revealed that
27 and 28 were local minima on the C6H6 energy hypersur-
face. The central bicyclo[1.1.0]butane unit in 27 As soon as the HBr formation slowed down, the mixture was
heated in an oil bath to 120°C for 5 h. Purification of 16a was
carried out by distillation with a 15-cm Vigreux column, which led
to a colorless oil (185.3 g, 84%) of b.p. 99Ϫ102°C/10 Torr (ref.[14]
b.p. 105Ϫ106/13 Torr). 16a was stored under nitrogen at Ϫ15°C
and converted with Br2 into 16b immediately prior to reaction.
(C1ϪC2ϪC4ϪC5) is bent, the dihedral angle 2Ϫ1Ϫ4Ϫ5
was calculated as 161.9°, the bond length C1ϪC4 as 1.680
˚
A; this is considerably shorter than the C1ϪC6 bond in 4
˚
(1.793 A). The central bicyclo[1.1.0]butane unit in 28
(C1ϪC2ϪC4ϪC5) is planar, at a C1ϪC4 bond length of
˚
1.911 A. However, this unusually long bond does not sever-
II. Synthesis of 1,6-Dibromotetracyclo[3.2.0.02,70.4,6]heptane 6
ely destabilize the molecule, which is more stable by 25.5
kcal/mol than 27. No local minimum with a structure re-
lated to 29 was found on the C6H6 energy hypersurface. The
B3LYP/6-31G*//B3LYP/6-31G* results are given in Table 4.
1. 2-exo-5-syn-Bicyclo[2.2.1]heptanediol (12)[9]. Ϫ a. With the
BoraneϪTHF Complex: A solution of borane in THF (225 ml, 1.80
, 0.405 mol) was added dropwise to a solution of 11 (72.9 g, 0.791
mol) in THF (150 ml), cooled in an ice-bath. After stirring for 3 h
at room temp., the colorless, gelatinous mixture was treated with 3
 NaOH (260 ml) and subsequently with H2O2 (30%, 150 ml),
which were slowly added with stirring and cooling in an ice bath.
The solution was then kept under reflux for 1 h and stirred for 16
h at room temperature. After addition of water (1.0 l) and ether
(1.0 l), the ether layer was separated. It contained exo-bicyclo-
[2.2.1]hept-5-ene-2-ol (13) (23.0 g, 26%). The water layer was con-
tinuously extracted with ether for 3 d with a perforator. After re-
moval of the solvent, 12 (35.0 g, 35%) was isolated as colorless
crystals of m.p. 168Ϫ170°C[24] (ref.[10] 181Ϫ183°C).
Table 4. Singlet and triplet B3LYP/6-31G*//B3LYP/6-31G* ener-
gies for 25, 27, and 28 and singlet-triplet energy differences
mole- B3LYP energy
ZPE
singlet
[au]
B3LYP energy
triplet
ZPE
triplet
[au]
∆ETS
[kcal/
mol]
cule
singlet
[au]
[au]
25[a] Ϫ193.9576082
0.089649 Ϫ193.9295710 0.087648
0.094862 Ϫ231.9610781 0.094529
0.095036 Ϫ231.9809648 0.093343
16.3
14.4
28.0
27
28
Ϫ231.9856058
Ϫ232.0260200
[a]
QCISD/6-31G*//QCISD/6-31G* energy for 25: E ϭ Ϫ193.3206890.
b. With 9-Borabicyclo[3.3.1]nonane (9-BBN): A solution of 11
(9.20 g, 99.8 mmol) in THF (100 ml) was added dropwise to a
solution of 9-BBN in THF (400 ml, 0.50 , 200 mmol). After stir-
ring for 2 h at room temperature, oxidative workup was carried out
by addition of ethanol (160 ml), NaOH (6 , 40 ml), and H2O2
(30%, 80 ml). Following the procedure as described in 1.a., the
ether layer contained 13 (1.34 g, 12%), and the water layer a mix-
ture of 12 (1.92 g, 15%) and cyclooctane-1,5-diol (3.10 g), which
could not be separated by distillation or crystallization.
27 and 28 have C2 symmetry. The calculated structures
of 27 and of 28 show some interesting features, which are
collected in Table 5.
Table 5. Selected structural parameters of 27 and 28, calculated by
B3LYP/6-31G*
mole- C1ϪC4 C1ϪC2 C2ϪC3 C2ϪC4 C3ϪC4 Ͻ1Ϫ4Ϫ2Ϫ3 Ͻ2Ϫ1Ϫ4Ϫ5
˚
˚
˚
˚
˚
cule
[A]
[A]
[A]
[A]
[A]
[°]
[°]
c. With the BoraneϪDimethyl Sulfide Complex: The complex
BH3 ·SMe2 (50 ml, 500 mmol), dissolved in hexane (250 ml), was
added within 30 min to a solution of 11 (46.0 g, 499 mmol) which
27
28
1.680
1.911
1.522
1.459
1.524
1.555
1.441
1.506
1.525
1.476
113.9
116.9
161.9
180.0
Eur. J. Org. Chem. 1998, 493Ϫ500
497

S. Glück-Walther, O. Jarosch, G. Szeimies
FULL PAPER
was kept in an ice bath. The white suspension was stirred for 3 h
solution. The combined pentane phases were extracted six times
at room temperature. After oxidation (500 ml of ethanol, 150 ml with water (500 ml) and dried with MgSO4. After removal of the
of 3  NaOH, 168 ml of 30% H2O2), workup II.1.a. afforded 13 solvent, the oily residue was purified by distillation, affording 20
(10.3 g, 19%) and 12 (18.4 g, 29%). 12 was contaminated by di-
(9.37 g, 67%) as a pale yellow liquid of b.p. 30°C/0.001 Torr. Ϫ IR
methyl sulfoxide (DMSO, 15.0 g), which could not be removed by (film): ν˜ ϭ 2980 cmϪ1, 2951, 1335, 1285, 1276. Ϫ 1H NMR (400
distillation or sublimation.
MHz, CDCl3): δ ϭ 2.35 (m, 2 H, 7-H2), 3.51 (m, 2 H, 1-, 4-H),
6.97 (m, 2 H, 3-, 6-H). Ϫ 13C NMR (100 MHz, CDCl3): δ ϭ 59.62
(d), 72.62 (t), 136.53 (s), 137.86 (d). Ϫ MS (70 eV), m/z (%): 171
(100), 169 (10), 146 (89), 90 (67). Ϫ C7H6Br2 (249.9): calcd. C
33.64, H 2.42; found C 33.20, H 2.55.
2. Bicyclo[2.2.1]heptane-2,5-dione (14): Compound 14 was ob-
tained by chromic acid oxidation of diol 12 following closely the
procedure of Hawkins et al.[10], which furnished 14 in 39% yield as
a waxy white solid. The 1H- and 13C-NMR spectra of 14 were
similar to those reported in the literature[10][25]
.
c. Synthesis of 6: A solution of 20 (5.50 g, 22.0 mmol) and aceto-
phenone (132 mg, 1.10 mmol) in ether (100 ml) with a 150-W mer-
cury high-pressure lamp for 24 h in a glass apparatus. After re-
moval of the solvent and distillation of the oily residue (25°C/0.001
Torr), the distillate was further purified by preparative layer chro-
matography (plc) (Merck, 2 mm, neutral Al2O3 60 F254) giving rise
3. 1,6-Dibromotetracyclo[3.2.0.02,7.04,6]heptane (6). Ϫ a. Reac-
tion of 14 with 2,2,2-Tribromo-1,3,2-benzodioxaphosphol (16b): Bro-
mine (21.6 g, 135 mmol) was added dropwise with stirring to a
solution of 16a (32.8 g, 149 mmol) in CH2Cl2 (40 ml) cooled in an
ice bath and stirring was continued for 1 h at room temperature.
The solution was again cooled in an ice bath and diketone 14 (8.12
g, 65.4 mmol) in CH2Cl2 (40 ml) was added and the mixture kept
under reflux for 20 h. The developing gaseous HBr was carried into
a 2  NaOH solution with a slow nitrogen stream. The solution was
diluted with CH2Cl2 (100 ml) and slowly pored into an ice-cooled
2  soda solution. After separation of the organic phase, extraction
of the water layer twice with 150 ml of CH2Cl2 and extraction of
the combined CH2Cl2 phases twice with 100 ml of 2  NaOH the
solution was dried with MgSO4, the solvent removed in vacuo and
the residue analyzed by NMR spectroscopy from which the forma-
tion of a mixture of 2,5,5-tribromobicyclo[2.2.1]hept-2-ene (17) and
2,2,5,5-tetrambromobicyclo[2.2.1]heptane (18) could be deduced.
Separation was achieved by distillation of 17, which was removed
at 40°C/0.001 Torr and condensed as a colorless liquid (13.2 g,
61%), whereas 18 was isolated by sublimation at 80°C/0.001 Torr
and further purified by crystallization from hexane to give white
crystals (2.43 g, 9%) of m.p. 148°C.
to 6 (4.46 g, 81%) as colorless liquid. Ϫ IR (film): ν ϭ 2935 cmϪ1
,
˜
2859, 1759, 1336, 1258, 1201. Ϫ 1H NMR (400 MHz, CDCl3): δ ϭ
1.98 (dt, J ϭ 5.3 Hz, J ϭ 1.3 Hz, 2 H, 2-, 4-H), 2.13 (t, J ϭ 1.3
Hz, 2 H, 3-H2), 2.55 (d, J ϭ 5.3 Hz, 2 H, 5-, 7-H). Ϫ 13C NMR
1
(100 MHz, CDCl3): δ ϭ 30.93 (s), 31.43 [t, J(13C1H) ϭ 135 Hz],
1
1
33.31 [d, J(13C1H) ϭ 172 Hz], 36.70 [d, J(13C1H) ϭ 195 Hz]. Ϫ
MS (70 eV), m/z (%): 252 (18), 250 (33) [Mϩ], 248 (15), 171 (100),
169 (92), 146 (85), 144 (88). Ϫ C7H6Br2 (249.9): calcd. C 33.64, H
2.42; found C 33.40, H 2.60.
4. Reactions of 6 with Organolithium Bases. Ϫ a. 6 and 5.0 equiv.
of tBuLi: A solution of tBuLi in pentane (5.2 ml, 8.4 mmol) was
added dropwise with stirring to a solution of 6 (420 mg, 1.68 mmol)
in ether (20 ml), which was kept at Ϫ78°C. The solution was stirred
for 2 h at Ϫ78°C, warmed to 0°C, and carefully hydrolyzed with
water. The ether phase was dried, the solvent removed and the re-
sidual oil analyzed by NMR spectroscopy which indicated the for-
mation of polymeric material and of quadricyclane (21a).
17: IR (film): ν˜ ϭ 2989 cmϪ1, 1581, 1442, 1404, 1261, 1208, 959,
683. Ϫ 1H NMR (400 MHz, CDCl3): δ ϭ 1.86 (m, 1 H, 7-H),
2.00Ϫ2.12 (m, 2 H, 2-, 6-H2), 2.52 (m, 1 H, 7-H), 2.98 (m, 1 H,
1-H), 3.65 (m, 1 H, 4-H), 6.18 (m, 1 H, 3-H). Ϫ 13C NMR (100
MHz, CDCl3): δ ϭ 46.68 (t), 50.90 (t), 51.58 (d), 62.73 (d), 63.45
(s), 129.58 (s), 134.08 (d). Ϫ MS (70 eV); m/z (%): 335 (7), 333 (20),
331 (21) [Mϩ], 329 (7), 253 (16), 251 (29), 249 (15), 171 (24), 169
(22), 79 (100).
18: IR (film): ν˜ ϭ 1463 cmϪ1, 1432, 1264, 984, 875, 768, 685,
662. Ϫ 1H NMR (400 MHz, CDCl3): δ ϭ 2.39 (m, 2 H, 7-H2),
2.93 (m, 2 H, endo-2-, -6-H), 3.05 (m, 2 H, exo-2-, -6-H), 3.33 (m,
2 H, 1-, 4-H). Ϫ 13C NMR (100 MHz, CDCl3): δ ϭ 35.55 (t), 54.61
(t, 2 C), 58.22 (d, 2 C), 65.58 (s, 2 C). Ϫ MS (70 eV); m/z (%): 335
(28), 333 (95), 331 (100) [Mϩ Ϫ Br], 329 (31), 253 (8), 251 (15), 249
(7), 171 (26), 169 (22). Ϫ C7H8Br4 (411.8): calcd. C 20.42, H 1.96;
found C 20.70, H 2.01.
b. 6 and 4.0 equiv. of tBuLi: A solution of tBuLi in pentane (10.6
ml, 16.9 mmol) was added dropwise with stirring to a solution of
6 (1.06 g, 4.24 mmol) in ether (10 ml), which was kept at Ϫ78°C.
The solution was stirred for 1 h at Ϫ78°C and carefully hydrolyzed
with 2 ml of D2O. Distillative workup afforded [1,6-D2]tetracy-
clo[3.2.0.02,7.04,6]heptane (21d) (349 mg, 87%) as colorless liquid,
b.p. 25 °C (bath)/12 Torr.
In the 1H-NMR spectrum of 21d the ratio of the signal aeries of
the multiplets at δ ϭ 1.35Ϫ1.48 (2-, 4-, 5-, 7-H) and at δ ϭ 2.02
(3-H2) was 4.0:1, indicating a D2 incorporation > 90%. Ϫ MS (70
eV); m/z (%): 94 (24) [Mϩ], 93 (66), 92 (77), 57 (100).
c. 6 and 2.0 equiv. of tBuLi: A solution of tBuLi in pentane (6.85
ml, 11.0 mmol) was added dropwise with stirring to a solution of
6 (1.37 g, 5.48 mmol) in ether (10 ml), which was kept at Ϫ78°C.
The solution was stirred for 2 h at Ϫ78°C, warmed to 0°C, and
carefully hydrolyzed with water. The ether phase was dried, the
solvent removed and the residual oil analyzed by NMR spec-
troscopy which indicated the formation of polymeric material and
of 1-bromotetracyclo[3.2.0.02,7.04,6]heptane (21e). Distillation of
the oily residue afforded 21e as a colorless liquid of b.p. 25°C
(bath)/0.001 Torr, which was further purified by plc leading to
0.496 g (53%) of 21e. Ϫ 1H NMR (400 MHz, CDCl3): δ ϭ 1.54
(m, 1 H, 4-H), 1.77 (m, 1 H, 6-H), 1.85 (m, 1 H, 2-H), 1.95 (m,
1 H, 3-H), 2.05 (m, 1 H, 7-H), 2.09 (m, 1 H, 5-H), 2.21 (m, 1 H,
3-H). Ϫ 13C NMR (100 MHz, CDCl3): δ ϭ 15.03 (d), 24.50 (d),
25.11 (d), 27.16 (d), 31.58 (t), 32.38 (d), 34.28 (s).
This reaction was repeated several times; in some experiments,
the liquid 17 contained a small amount (< 5%) of 2,2,5-tribro-
monortricyclane (19), which could not be separated from 17.
19: 13C NMR (100 MHz, CDCl3): δ ϭ 24.43 (d), 29.97 (s), 33.16
(t), 37.23 (d), 40.91 (t), 49.88 (d), 65.73 (s).
b. 2,5-Dibromobicyclo[2.2.1]hepta-2,5-diene (20): A solution of
KOtBu (22.9 g, 204 mmol) in DMSO (50 ml) was added dropwise
with stirring to a solution of a mixture of 17 (14.5 g, 43.8 mmol)
and 18 (5.00 g, 12.1 mmol) in DMSO (50 ml), cooled in an ice
bath. After 20 h at room temperature, water (50 ml) was added and
the mixture extracted five times with pentane. Separation of the
The same experiment was repeated, but workup was carried out
1
phases was a slow process, even after addition of a saturated NaCl with deuterium oxide instead of water. In the H-NMR spectrum
498
Eur. J. Org. Chem. 1998, 493Ϫ500

1,6-Dehydroquadricyclane
FULL PAPER
(CDCl3) of the isolated [6-D]21e the signal at δ ϭ 1.77 was absent. Ϫ78°C with tBuLi (13.0 ml, 20.8 mmol). The yellow solution was
Apart from minor chemical shift differences due to different con- stirred for 1 h at Ϫ78°C and charged with chlorotrimethylsilane
centrations, the 13C-NMR spectrum of [6-D]21e was identical to
(1.07 g, 9.85 mmol). The mixture was allowed to warm to room
that of 21e with the exception of the signal for C-6 which appeared temperature, stirred for 16 h and hydrolyzed with ice water. The
at δ ϭ 15.26 as a low intensity triplet. Ϫ [6-D]21e: MS (70 eV); organic material was extracted with ether. Distillative workup af-
m/z (%): 173 (12), 171 (12) [Mϩ], 146 (10), 144 (10), 92 (100), 57 forded a colorless oil [b.p. 10°C (bath)/0.01 Torr], which was
(72). Ϫ C7H6D79Br: calcd. 170.9793; found 170.9794 (HRMS).
further purified by plc to give 23a (1.07 g, 92%) as a colorless
liquid. Ϫ IR (film): ν˜ ϭ 3044 cmϪ1, 2954, 2931, 2856, 1261, 1248.
d. 6, 2.0 equiv. of tBuLi and TMEDA: A solution of tBuLi in
pentane (3.7 ml, 5.9 mmol) was added dropwise with stirring to a
solution of 6 (740 mg, 2.96 mmol) in ether (10 ml), which was kept
at Ϫ78°C. The solution was stirred for 1 h at Ϫ78°C, charged at
Ϫ50°C with N,N,NЈ,NЈ-tetramethylethylenediamine (TMEDA, 347
mg, 2.97 mmol), warmed to 0°C and carefully hydrolyzed D2O.
Distillative workup afforded [6-D]21e (215 mg, 43%) of b.p. 10°C
(bath)/0.01 Torr.
1
Ϫ H NMR (400 MHz, CDCl3): δ ϭ Ϫ0.06 (s, 18 H, SiMe3), 1.22
(d, J ϭ 4.0 Hz, 2 H, 5-, 7-H), 1.26 (dt, J ϭ 4.0 Hz, J ϭ 1.4 Hz,
2 H, 2-, 4-H), 2.08 (t, J ϭ 1.4 Hz, 2 H, 3-H2). Ϫ 13C NMR (100
MHz, CDCl3): δ ϭ Ϫ2.24 (q), 12.13 (s), 18.58 (d), 27.11 (d), 33.76
(t). Ϫ MS (70 eV); m/z (%): 226 (2) [Mϩ], 221 (7), 148 (42), 73
(100). Ϫ C13H24Si2 (236.5): calcd. C 66.02, H 10.23; found C 66.16,
H 10.27.
c. 1,6-Dimethyltetracyclo[3.2.0.02,7.04,6]heptane (23b): Com-
pound 6 (910 mg, 3.64 mmol) in ether (10 ml) was mixed at Ϫ78°C
with tBuLi (8.5 ml, 14.4 mmol). The yellow solution was stirred
for 1 h at Ϫ78°C, dimethyl sulfate (933 mg, 7.40 mmol) was added
and stirring continued for 16 h at room temperature. After hydroly-
sis with 2  ammonia, the ether layer was separated and dried with
MgSO4. Distillative workup afforded 23b (240 mg, 55%) as a color-
less liquid, b.p. 45°C (bath)/40 Torr. Ϫ 1H NMR (400 MHz,
CDCl3): δ ϭ 1.21 (s, 6 H, Me), 1.30 (m, 4 H, 2-, 4-, 5-, 7-H), 1.88
(m, 2 H, 3-H2). Ϫ 13C NMR (100 MHz, CDCl3): δ ϭ 15.13 (q),
21.63 (s), 23.24 (d), 30.82 (d), 32.08 (t). Ϫ MS (70 eV); m/z (%):
120 (48) [Mϩ], 105 (57), 91 (21), 80 (100). Ϫ C9H12: calcd.
120.0939; found 120.0934 (HRMS).
The same reaction with 1.0 equiv. of LiBr instead of TMEDA
led also only to [6-D]21e and polymeric material.
In addition to these reactions, 6 was treated with BuLi (3.0 and
6.0 equiv.), and a mixture of BuLi and tBuLi. In all experiments
considerable amounts of polymeric material and 21e were formed.
e. 1-Bromo-6-lithiotetracyclo[3.2.0.02,7.04,6]heptane (7): A solu-
tion of 6 (34.0 mg, 0.136 mmol) in [D8]THF (1.5 ml) was trans-
ferred into a NMR tube, cooled to Ϫ78°C and carefully mixed
with a solution of tBuLi (0.17 ml, 0.27 mmol) in pentane. All vol-
atile parts were removed at Ϫ78°C/0.001 Torr. The residue was
dissolved in [D8]THF and the 13C-NMR spectrum of the solution
recorded at Ϫ80°C, Ϫ60°C, Ϫ40°C, 0°C, and room temperature.
Ϫ
13C NMR (100 MHz, CDCl3): δ ϭ 11.58 (s, C-6), 30.49 (d, C-
4), 30.71 (d, C-7), 31.95 (d, C-5), 32.05 (d, C-2), 33.54 (t, C-3),
34.50 (s, C-1).
d. Dimethyl Bicyclo[2.2.1]hepta-2,5-diene-2,5-dicarboxylate (24c)
and Dimethyl Tetracyclo[3.2.0.02,7.04,6]heptane-1,6-dicarboxylate
(23c): Compound 6 (1.40 g, 5.60 mmol) in ether (10 ml) was mixed
at Ϫ78°C with tBuLi (14.0 ml, 22.4 mmol). The yellow solution
was stirred for 1 h at Ϫ78°C, methyl chloroformate (1.04 g, 11.0
mmol) was added and stirring continued for 16 h at room tempera-
ture. After hydrolysis with 2  ammonia, the ether layer was sepa-
rated and dried with MgSO4. Distillative workup afforded 24c (510
mg, 45%) as a colorless liquid, b.p. 45°C (bath)/0.001 Torr. Ϫ IR
(film): ν˜ ϭ 3010 cmϪ1, 2953, 1747, 1719, 1585, 1397, 1260, 1153,
In the spectrum taken at room temperature, the signals of 7 could
not be recognized. Many new signals had appeared.
f. 1,6-Dilithiotetracyclo[3.2.0.02,7.04,6]heptane (21c): Compound
6 (23.0 mg, 0.092 mmol) and tBuLi (0.368 mmol) were allowed to
react as described under 4.d. The [D8]THF solution of 21c was
again analyzed by 13C-NMR spectroscopy at Ϫ80°C, Ϫ60°C,
Ϫ40°C, 0°C, and room temperature. Ϫ 13C NMR (100 MHz,
CDCl3): δ ϭ 14.87 (d, C-5, -7), 16.75 (s, C-1, -6), 23.08 (d, C-2, -
4), 32.96 (t, C-3). Ϫ In the spectrum taken at room temperature,
the signals of 21c could not be recognized. Many new signals had
appeared.
1
1063, 986, 877, 770. Ϫ H NMR (400 MHz, CDCl3): δ ϭ 2.27 (m,
2 H, 7-H2), 3.68 (s, 6 H, Me), 4.04 (m, 2 H, 1-, 4-H), 7.75 (m, 2 H,
3-, 6-H). Ϫ 13C NMR (100 MHz, CDCl3): δ ϭ 28.48 (q), 51.53 (d),
73.36 (t), 148.32 (s), 155.14 (d), 164.74 (s). Ϫ MS (70 eV); m/z (%):
208 (50) [Mϩ], 193 (51), 177 (75), 149 (91), 57 (100). Ϫ C11H12O4:
calcd. 208.0775; found 208.0735 (HRMS).
5. Reactions of 7 and 21c. Ϫ a. 1-Bromo-6-trimethylsilyltetra-
cyclo[3.2.0.02,7.04,6]heptane (21f): Compound 6 (1.08 g, 4.32 mmol)
in ether (10 ml) was mixed at Ϫ78°C with tBuLi (5.40 ml, 8.64
mmol). The pale yellow solution was stirred for 1 h at Ϫ78°C and
charged with chlorotrimethylsilane (428 mg, 3.94 mmol). The mix-
ture was allowed to warm to room temperature and hydrolyzed
with ice water. The organic material was extracted with ether. Dis-
tillative workup afforded 21f [25°C(bath)/0.01 Torr], which was
further purified by plc to give 21f (610 mg, 64%) as colorless liquid.
Photocyclization of 24c: Compound 24c (100 mg, 0.480 mmol)
was irradiated in 15 ml of ether with a mercury high-pressure lamp
at Ϫ12°C for 24 h in the presence of acetophenone (2.88 mg, 0.0240
mmol). After removal of the volatile products in vacuo, 23c (78.0
1
mg, 78%) was isolated as a colorless oil. Ϫ H NMR (400 MHz,
CDCl3): δ ϭ 2.43 (dt, J ϭ 4.7 Hz, J ϭ 1.1 Hz, 2 H, 2-, 4-H), 2.61
(t, J ϭ 1.1 Hz, 2 H, 3-H2), 2.86 (d, J ϭ 4.7 Hz, 2 H, 5-, 7-H), 3.67
(s, 6 H, Me). Ϫ 13C NMR (100 MHz, CDCl3): δ ϭ 26.88 (s), 31.58
(t), 51.64 (q), 30.17 (d), 34.71 (d), 171.87 (s).
Ϫ IR (film): ν ϭ 2956 cmϪ1, 2897, 2867, 1275, 1249, 1124, 852,
˜
837, 751. Ϫ 1H NMR (400 MHz, CDCl3): δ ϭ Ϫ0.05 (s, 9 H,
SiMe3), 1.45 (m, 1 H, 5-H), 1.77 (m, 1 H, 7-H), 1.89Ϫ1.96 (m, 3 H,
3-H2, 4-H), 2.29 (m, 1 H, 2-H). Ϫ 13C NMR (100 MHz, CDCl3):
δ ϭ Ϫ2.24 (q), 13.76 (s), 25.61 (d), 28.95 (d), 31.37 (d), 31.57 (d),
32.59 (t), 34.25 (s). Ϫ MS (70 eV); m/z (%): 244 (14), 242 (13)
[Mϩ], 146 (95), 144 (98), 83 (35), 73 (100). Ϫ C10H1579BrSi: calcd.
242.0126; found 242.0096 (HRMS).
e. 1,6-Diiodotetracyclo[3.2.0.02,7.04,6]heptane (23d): Compound
6 (1.00 g, 4.00 mmol) in ether (8 ml) was mixed at Ϫ78°C with
tBuLi (10.0 ml, 16.0 mmol). The solution was stirred for 1 h at
Ϫ78°C, iodine cyanide (1.20 g, 7.85 mmol, freshly crystallized from
chloroform) was added, the mixture, which was protected against
daylight, warmed to room temperature., and after stirring for 2 h
hydrolyzed with water. 20 ml of a 2  sodium thiosulfate solution
b. 1,6-Bistrimethylsilyltetracyclo[3.2.0.02,7.04,6]heptane (23a): was added, the ether layer separated, dried with MgSO4 and the
Compound 6 (1.30 g, 5.20 mmol) in ether (8.0 ml) was mixed at
solvent removed in vacuo from a 0°C bath. According to NMR
Eur. J. Org. Chem. 1998, 493Ϫ500
499

S. Glück-Walther, O. Jarosch, G. Szeimies
FULL PAPER
spectroscopy, the residual oil (0.84 g, 62%) was a 85:15 mixture of 6-31G* level of theory were corrected by a scaling factor of
23d and 2,5-diiodobicyclo[2.2.1]hepta-2,5-diene (24d). Further 0.9804[18]
.
purification of 23d was carried out with plc.
[1]
Taken in part from: S. Glück-Walther, Dissertation, Univ.
München, 1993.
23d: 1H NMR (400 MHz, CDCl3): δ ϭ 2.00 (dt, J ϭ 4.9 Hz,
J ϭ 1.7 Hz, 2 H, 2-, 4-H), 2.20 (t, J ϭ 1.7 Hz, 2 H, 3-H2), 2.35 (d,
J ϭ 4.9 Hz, 2 H, 5-, 7-H). Ϫ 13C NMR (100 MHz, CDCl3): δ ϭ
15.20 (s), 32.26 (t), 33.86 (d), 36.88 (d). Ϫ MS (70 eV); m/z (%):
344 (44) [Mϩ], 217 (53), 192 (47), 91 (100), 90 (49) 65 (85). Ϫ
C7H6I2: calcd. 343.8559; found 343.8573 (HRMS).
[2]
[3]
P. M. Warner, Chem. Rev. 1989, 89, 1067Ϫ1093, and refer-
ences therein.
G. Szeimies in Advances in Strain in Organic Chemistry, vol. 2
(Ed.: B. Halton), Jai Press, London, 1992, p. 1Ϫ55.
O. Baumgärtel, G. Szeimies, Chem. Ber. 1983, 116, 2180Ϫ2204.
O. Baumgärtel, J. Harnisch, G. Szeimies, M. Van Meerssche, G.
Germain, J.-P. Declercq, Chem. Ber. 1983, 116, 2205Ϫ2218.
J. Kenndoff, K. Polborn, G. Szeimies, J. Am. Chem. Soc. 1990,
112, 6117Ϫ6118.
[4]
[5]
[6]
[7]
24d: 1H NMR (400 MHz, CDCl3): δ ϭ 2.22 (m, 2 H, 7-H2), 3.55
(t, 2 H, 1-, 4-H), 7.05 (m, 2 H, 3-, 6-H). Ϫ 13C NMR (100 MHz,
CDCl3): δ ϭ 63.77 (d), 72.95 (t), 104.59 (s), 146.84 (d).
J. Podlech, K. Polborn, G. Szeimies, J. Org. Chem. 1993, 58,
4113Ϫ4117.
[8]
[9]
J. Schäfer, G. Szeimies, Tetrahedron Lett. 1990, 2263Ϫ2264.
G. Zweifel, K. Nagase, H. C. Brown, J. Am. Chem. Soc. 1962,
84, 183Ϫ189.
The intended preparation of 23d from 21c was also carried out
with iodine, iodine chloride, and 4-tolylsulfonyl iodide. In all cases
the NMR spectra of the raw material indicated the formation of
2,5-diiodonorbornadiene (24d) in yields of 34%, 17%, and 25%.
[10]
[11]
R. T. Hawkins, R. S. Hsu, S. G. Wood, J. Org. Chem. 1978,
43, 4648Ϫ4650.
K.Omura, D. Swern, Tetrahedron 1978, 34, 1651-1660; see also:
A. J. Mancuso, S. L. Huang, D. Swern, J. Org. Chem. 1978,
43, 2480Ϫ2482.
f. 1,6-Dichlorotetracyclo[3.2.0.02,7.04,6]heptane (23e): Compound
6 (1.03 g, 4.12 mmol) in ether (8 ml) was mixed at Ϫ78°C with
tBuLi (9.7 ml, 10.5 mmol). The solution was stirred for 1 h at
Ϫ78°C, 4-tolylsulfonyl chloride (1.57 g, 8.23 mmol) was added, the
mixture allowed to warm to room temperature., and after stirring
for 20 h hydrolyzed with water. The ether layer was separated, dried
with MgSO4 and the solvent removed in vacuo. The oily residue
was distilled at 25°C (bath)/0.01 Torr, the distillate further purified
by plc to give 23e (320 mg, 48%) as a colorless oil. Ϫ 1H NMR
(400 MHz, CDCl3): δ ϭ 2.02 (dt, J ϭ 5.4 Hz, J ϭ 1.4 Hz, 2 H,
2-, 4-H), 2.12 (t, J ϭ 1.4 Hz, 2 H, 3-H2), 2.54 (d, J ϭ 5.4 Hz, 2 H,
5-, 7-H). Ϫ 13C NMR (100 MHz, CDCl3): δ ϭ 30.95 (t), 32.75 (d),
36.17 (d), 43.55 (s).
[12]
M. Hirano, M. Oose, T. Morimoto, Bull. Chem. Soc. Jpn. 1991,
64, 1046Ϫ1047.
R. H. Shapiro, Org. React. 1976, 23, 405Ϫ507.
H. Gross, U. Karsch, J. Prakt. Chem. 1965, 29, 315Ϫ318.
H.-G. Zoch, G. Szeimies, R. Römer, G. Germain, J.-P. Declercq,
Chem. Ber. 1983, 116, 2285Ϫ2310.
M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B. G.
Johnson, M. A. Robb, J. R. Cheeseman, T. Keith, G. A. Peters-
son, J. A. Montgomery, K. Raghavachari, M. A. Al-Laham, V.
G. Zakrzewski, J. V. Ortiz, J. B. Foresman, J. Cioslowski, B. B.
Stefanov, A. Nanayakkara, M. Challacombe, C. Y. Peng, P. Y.
Ayala, W. Chen, M. W. Wong, J. L. Andres, E. S. Replogle, R.
Gomperts, R. L. Martin, D. J. Fox, J. S. Binkley, D. J. Defrees,
J. Baker, J. P. Stewart, M. Head-Gordon, C. Gonzalez, J. A.
Pople, GAUSSIAN 94 (Revision D.3), Gaussian, Inc., Pitts-
burgh PA, 1995.
T. Ziegler, Chem. Rev. 1991, 91, 651Ϫ667; J. Labanowski, J.
Adzelm (Eds.), Density Functional Methods in Chemistry,
Springer, Berlin, 1991; R. G. Parr, W. Yang, Density-Functional
Theory of Atoms and Molecules, Oxford University Press, New
York, 1989.
M. J. Frisch, A. Frisch, J. B. Foresman, GAUSSIAN 94 UserЈs
Reference, Gaussian Inc., Pittsburgh, 1995, and references ther-
ein.
C. J. Cramer, F. J. Dulles, J. W. Storer, S. E. Worthington, Chem.
Phys. Lett. 1994, 218, 387Ϫ394; H. M. Sulzbach, E. Bolton, D.
Lenoir, P. v. R. Schleyer, H. F. Schaefer III, J. Am. Chem. Soc.
1996, 118, 9908Ϫ9914; P. R. Schreiner, W. L. Karney, P. v. R.
Schleyer, W. T. Borden, T. P. Hamilton, H. F. Schaefer III, J.
Org. Chem. 1996, 61, 7030Ϫ7039.
See for example: E. Goldstein, B. Beno, K. N. Houk, J. Am.
Chem. Soc. 1996, 118, 6036Ϫ6043; J. Cioslowski, Y. Liu, M.
Martinov, P. Piskorz, D. Moncrieff, J. Am. Chem. Soc. 1996,
118, 5261Ϫ5264; J. C. Poutsma, J. J. Nash, J. A. Paulino, R. R.
Squires, J. Am. Chem. Soc. 1997, 119, 4686Ϫ4697.
J. B. Foresman, A. Frisch, Exploring Chemistry with Electronic
Structure Methods, 2nd ed., Gaussian, Inc. Pittsburg, 1996, p.
64.
K. B. Wiberg, N. McMurdie, J. V. McClusky, C. M. Hadad, J.
Am. Chem. Soc. 1993, 115, 10653Ϫ10657.
F. Alber, G. Szeimies, Tetrahedron Lett. 1994, 4093Ϫ4096.
According to Hawkins et al.[10] a small amount of the isomeric
bicyclo[2.2.1]heptane-2,6-diol might be present and responsible
for the melting point depression.
[13]
[14]
[15]
[16]
6. Reactions of 23d and 21f Towards the Generation of 4. Ϫ a. 23d
and tBuLi: Compound 23d (0.860 g, 2.50 mmol) in ether (6 ml)
was mixed at Ϫ78°C with tBuLi in pentane (6.0 ml, 10.2 mmol).
The solution was stirred for 1 h at Ϫ78°C, allowed to warm to
room temperature., and hydrolyzed with D2O. The ether layer was
separated, dried with MgSO4 and the solvent removed in vacuo.
The oily residue was analyzed by NMR spectroscopy which showed
besides polymeric material the presence of 21d.
[17]
[18]
[19]
b. 21f and Cesium Fluoride: A solution of cesium fluoride (2.00
g, 13.2 mmol) in DMSO (20 ml) was added dropwise at room tem-
perature to a solution of 21f (400 mg, 1.64 mmol) in DMSO. After
24 h at room temperature, the reaction mixture was poured on ice/
water, the organic material extracted with ether, and the ether layer
extracted with water. After drying with MgSO4 and removal of the
solvent in vacuo, NMR analysis indicated the presence of unre-
acted 21f accompanied by some 2-bromo-5-trimethylsilylbicyclo-
[2.2.1]heptane (22f).
[20]
[21]
[22]
7. Ab Initio Calculations: Ab initio calculations were performed
with the GAUSSIAN 94 program package[16]. Calculations em-
ployed the 6-31G* basis set. Spin unrestricted wave functions were
used for triplets, spin restricted wave functions for singlets. DFT
computations were carried out using the B3LYP keyword[26]. Geo-
metries were optimized by use of analytic gradients. The nature of
the stationary points was investigated by calculation of the vi-
brational frequencies. Zero-point energies, calculated at the B3LYP/
[23]
[24]
[25]
[26]
H. Ahlbrecht, M. Dietz, C. Schön, V. Baumann, Synthesis
1991, 133Ϫ140.
A. D. Becke, J. Chem. Phys. 1993, 98, 5648Ϫ5652.
[97284]
500
Eur. J. Org. Chem. 1998, 493Ϫ500