Breton and Shugart
Con clu sion s
A homologous series of tricyclic diazetines (6a -c),
differing by the number of methylene groups in the
saturated bridges of the fused carbon bicycles, was
q
synthesized. The ∆H ’s of decomposition for each of the
diazetines were determined experimentally. The ground-
state strain energy of each diazetine was estimated
utilizing computationally obtained ∆H ’s for each of the
f
experimentally investigated diazetines as well as several
q
other diazetines whose ∆H ’s had been previously re-
ported in the literature. The sum of the ground-state
q
F IGURE 1. Calculated transition state structure for the
decomposition of 1a at the UB3LYP/6-31+G** level. Structural
data provided in Table 6.
strain energies and ∆H ’s of decomposition for all of the
diazetines was nearly constant, with an average value
of 59 kcal/mol, suggesting that all of the diazetines
decompose via the same mechanism. Generally, the
higher the ground-state strain energy of the diazetine,
the less the ∆H for decomposition. The agreement of the
experimentally determined ∆H values with values ob-
tained by computational modeling studies supports the
reaction mechanism proposed by Yamabe that the elimi-
nation process occurs by an unsymmetrical, yet concerted,
transition state with strong biradical character.
located by the CASSCF(2,2)/6-31G* method (Table 6).
The most significant difference being that the twist of
the N1-N2 fragment relative to the C3-C4 bond of the
UB3LYP/6-31+G** structure as measured by the N1-
N2-C3-C4 dihedral angle (6.0°) was shallower than that
of the CASSCF(2,2)/6-31G* structure (14.9°). However,
the calculated reaction barrier (30.9 kcal/mol) at the
UB3LYP/6-31+G** level matched that reported by Yam-
abe (31.0 kcal/mol). It appeared, therefore, that the
q
q
Exp er im en ta l Section
UB3LYP/6-31+G** method afforded reasonable results
for the calculation.20
Cycloa d d u ct 5c. In a screw-cap high-pressure tube was
combined 4 (0.105 g, 0.76 mmol), 1,3-cycloheptadiene (5 g, 53
mmol), and chlorobenzene (2 mL). The mixture was swirled
This computational method was next applied toward
locating the transition state for decomposition of diazet-
ines 6a -c, and 7. In each case, a transition state was
successfully located (Figure 2). Each of the transition-
state structures exhibited a single imaginary frequency
2
and briefly deoxygenated with a stream of dry N . The tube
was sealed with a teflon cap and heated at 150 °C for 48 h.
After this time, the solution had darkened and some dark
precipitate was apparent. The mixture was cooled and filtered
into a clean round-bottomed flask, and the solvent was
removed in vacuo. Column chromatography (SiO , 1:2 EtOAc/
2
hexane) afforded 94.3 mg (53% yield) of 5c as a white solid:
whose animated motion was consistent with loss of N
2
from the hydrocarbon fragment. The transition-state
geometries at the diazetine ring systems were remark-
ably similar (Table 5). As with the parent diazetine (i.e.,
1
H NMR δ 6.16 (dd, J ) 4.8, 3.2 Hz, 2H), 4.78 (br s, 2H), 3.01
13
(
br m, 2H), 2.99 (s, 3H), 1.8-1.5 (m, 6H); C NMR δ 162.7,
1
a ), each of the substituted diazetines exhibited a small
12 15 3 2
130.9, 67.4, 33.8, 25.5, 25.1, 21.9. Anal. Calcd for C H N O :
N1-N2-C3-C4 dihedral angle of 4.9-5.5° such that the
diazetine portion of the ring was not planar. Yamabe had
previously established for 1a that the symmetry-lowering
C, 61.79; H, 6.48; N, 18.01. Found: C, 61.70; H, 6.57; N, 17.74.
Dia zetin e 6c. To a solution of 5c (93.4 mg, 0.4 mmol) in
1
.5 mL of 2-propanol was added 0.112 g (2 mmol) of freshly
crushed KOH. The resulting mixture was heated to reflux for
h and then cooled to room temperature. A few small pieces
twist of the N fragment relative to the C-C diazetine
2
2
bond in the transition state results in a lowering of the
energy of the transition state due to dual orbital mixing
effects possible for a transition state exhibiting strong
biradical character. The twisted transition state is of even
of ice were added to the solution, and then concentrated HCl
was added dropwise until a solution of pH ) 2 was obtained.
This solution was heated at 50 °C for 5 min and then cooled.
A solution of 5 M aq NH
neutral solution. An aq solution of 3 M CuCl
4
OH was added dropwise to obtain a
2
(1 mL) was added
lower energy than a C
symmetry-allowed” transition-state structure.
Higher-level single-point energy calculations of the
s
-symmetric Woodward-Hoffmann
5
via pipet to the stirring reaction mixture, and a dark-brownish-
red precipitate formed almost immediately. Finally, sufficient
“
5
M aq NH
a clear-deep-blue solution. This solution was washed with 3
× 15 mL Et O, dried, and concentrated. Column chromatog-
raphy (SiO , 1:1 EtOAc/hexane) afforded 58 mg (98% yield) of
c as a thick colorless liquid: 1H NMR δ 5.95 (dd, J ) 5.0,3.1
Hz, 2H), 4.43 (dd, J ) 2.2, 1.7 Hz, 2H), 2.87 (br m, 2H), 1.9-
4
OH was added to convert the reaction mixture to
transition-state structures and the corresponding ground-
state diazetine structures (minimized at the B3LYP/6-
2
3
1G** level) at the B3LYP/6-311+G(3df,2p) level afforded
transition-state barriers for the decomposition reactions
which were corrected for zero-point energies utilizing the
2
6
(
1
.1 (m, 6H); 13C NMR δ 129.7, 82.7, 33.6, 25.0, 22.2. UV λmax
corresponding zero-point correction from the minimiza-
tion calculation). Calculated barriers for 6a (33.3), 6b
(NdN) ) 342 nm.
Gen er a l P r oced u r e for th e Th er m olysis of Dia zetin es
(36.9), 6c (37.5), and 7 (34.6 kcal/mol) corresponded
6
a -c a n d 7. A microtube was fashioned from a standard
reasonably well with the experimental values (Table 1).
The agreement of these calculated values with the
experimental values suggests that Yamabe’s proposed
mechanism of an unsymmetrical, yet concerted, elimina-
Pasteur pipet by first sealing the dispensing tip with a flame.
Most of the upper portion of the pipet was cut off leaving ∼2
cm of the wider glassware attached. A solution of the diazetine
(∼10 mg) in 20 µL of toluene was carefully dispensed into the
sealed portion of the pipet with a syringe, being especially
careful to avoid contaminating the sides of the tube where the
second seal was to be made. The solution was cooled in dry
ice while a vacuum was applied at the open end of the tube,
and the tube was sealed at the top of the tube just below where
2
tion of N from diazetines is quite reasonable.
(20) The spin-squared values for each of the calculations carried out
using the UB3LYP method are reported in the Supporting Information.
8
648 J . Org. Chem., Vol. 68, No. 22, 2003