1H-1,3-Diazepines, 5H-1,3-diazepines, 1,3-diazepinones, and 2,4-diazabicyclo[3.2.0]heptenes

Article abstract of DOI:10.1039/b317099c

Tetrazolo[1,5-a]pyridines/2-azidopyridines 1 undergo photochemical nitrogen elimination and ring expansion to 1,3-diazacyclohepta-1,2,4,6-tetraenes 3, which react with alcohols to afford 2-alkoxy-1H/-13-diazepines 4 (5), with secondary amines to 2-dialkyl

Full text of DOI:10.1039/b317099c

View Article Online / Journal Homepage / Table of Contents for this issue
1H-1,3-Diazepines, 5H-1,3-diazepines, 1,3-diazepinones, and
2,4-diazabicyclo[3.2.0]heptenes†,‡
Ales Reisinger,a Rainer Koch,b Paul V. Bernhardta and Curt Wentrup*a
a Department of Chemistry, School of Molecular and Microbial Sciences,
The University of Queensland, Brisbane, Qld 4072, Australia. E-mail: wentrup@uq.edu.au
b Institut fürReine und Angewandte Chemie, Carl von Ossietetzky Universität,
D-26111 Oldenburg, Germany
Received 6th January 2004, Accepted 4th March 2004
First published as an Advance Article on the web 18th March 2004
Tetrazolo[1,5-a]pyridines/2-azidopyridines 1 undergo photochemical nitrogen elimination and ring expansion to
1,3-diazacyclohepta-1,2,4,6-tetraenes 3, which react with alcohols to afford 2-alkoxy-1H-1,3-diazepines 4 (5), with
secondary amines to 2-dialkylamino-5H-1,3-diazepines 16, sometimes via isolable 2-dialkylamino-1H-1,3-diazepines
15, and with water to 1,3-diazepin-2-ones 19. The latter are also obtained by elimination of isobutene or propene
from 2-tert-butoxy- or 2-isopropoxy-1H-1,3-diazepines 4 or 5. 1,3-Diazepin-2-one 22B and 1,3-diazepin-4-one 24
were obtained from hydrolysis of the corresponding 4-chlorodiazepines. Diazepinones 19 undergo photochemical
ring closure to diazabicycloheptenones 25 in high yields. The 2-alkoxy-1H-1,3-diazepines 4 and 5 interconvert by
rapid proton exchange between positions N1 and N3. The free energies of activation for the proton exchange were
measured by the Forsén–Hoffman method as ∆G‡298 = 16.2 0.6 kcal molϪ1 as an average for 4a–c in CD2Cl2,
acetone-d6, and methanol-d4, and 14.1 0.6 kcal molϪ1 for 4c in acetone/D2O. The structures of 2-methoxy-5,6-
bis(trifluoromethyl)-1H-1,3-diazepine 4k, 1,2-dihydro-4-diethylamino-5H-1,3-diazepin-2-one 22bB, and diazabicyclo-
heptanone 26 were determined by X-ray crystallography. The former represents the first reported X-ray crystal
structure of any monocyclic N-unsubstituted 1H-azepine.
Introduction
1,4-Diazepines are well known for their many pharmaceutical
properties. In contrast, 1,3-diazepines are relatively little
known.3 Some 1,3-diazepin-2-ones and other cyclic ureas have
received considerable attention recently as potential anti-AIDS
drugs.4 In previous communications we have described the
photolysis of variously substituted tetrazolo[1,5-a]pyridines
1T/2-azidopyridines 1A as a convenient method of synthesis
of 1,3-diazepines.1,2 The reaction proceeds via ring expansion
of the first-formed 2-pyridylnitrenes 2 to 1,3-diazacyclohepta-
1,2,4,6-tetraenes 3, which in several cases have been charac-
terized by matrix-isolation IR spectroscopy (Scheme 1).2,5 Here
we report full details of the syntheses of the title compounds
by nucleophilic trapping of 3 as well as the first X-ray crystal
structure of an N-unsubstituted 1H-azepine.
Scheme 1
crystalline in the solid state. Unsymmetrical diazepines of this
kind can exist in two NH tautomeric forms, 4 and 5, of which
the isomers with the substituent the farthest away from the
Results and discussion
1
1H-1,3-Diazepines
NH site usually dominates. For each of the isomeric tetrazole/
azide pairs 6 and 7, 8 and 9, and 10 and 11, the same cyclic
carbodiimide intermediate 3 is formed.5a Consequently, nucleo-
philic trapping affords the same diazepines 4 (5) for each pair.
The compounds and yields are listed in Table 1. The chemical
shifts and coupling patterns in the proton NMR spectra clearly
identify these compounds as the 1H, not the 4H or 5H, isomers.
The proton attached to the carbon atom next to the NH func-
tion couples with the NH proton. Decoupling of the NH
proton by double irradiation or by addition of a drop of D2O
causes a corresponding reduced multiplicity of the CH signal.
Photolysis of tetrazoles/azides (1T/1A) in 1,4-dioxan solution
in the presence of alcohols afforded 2-alkoxy-1,3-diazepines 4
(5) as indicated in Scheme 2 and Table 1. Most of these
diazepines are distillable, yellow to orange compounds, often
† Preliminary reports on parts of this work have been published in
ref 1. This paper is Diazepines, Part 3. For Part 2, see ref 2.
‡ Electronic supplementary information (ESI) available: NMR spectra
of 4a, 4c, 19v
20, 19a
25a, and 19g
25g (Fig. S1–S6),
Arrhenius and Eyring plots for H-exchange in 4b,c (Fig. S7), views of
the X-ray crystal structures of compounds 26 and 22bB (Fig. S8–S9),
bond lengths and angles for compounds 4k, 22bB and 26, preparative
procedures and characterization data for all compounds not described
in the Experimental section, and computational data (Cartesian co-
1
The H NMR spectra were fully assigned in several cases by
means of homonuclear decoupling experiments. The 13C NMR
spectra were fully assigned in several cases by means of 2-D
HMBC experiments (e.g. for 4b), or by means of short- and
long-range C–F couplings in the compounds containing CF3
groups.
ordinates, absolute energies, IR, H and 13C NMR) for 20a–c, 21a–b,
1
22aA–F and 22aB dimer. See http://www.rsc.org/suppdata/ob/b3/
b317099c/
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1227

View Article Online
Table 1 1H-1,3-Diazepines 4(5) from reaction of tetrazoles/azides 1T/1A with alcohols ROH
R1
R2
R3
R4
OR
Yield (%)
a
b
c
d
e
f
g
h
i
j
k
l
m
n
o
p
q
r
H
H
H
H
CF3
CF3
CF3
H
H
H
H
H
H
H
H
H
H
CF3
CF3
CF3
CF3
CF3
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
CF3
CF3
CF3
H
H
Cl
H
OMe
OEt
OiPr
OtBu
OMe
OEt
OtBu
OMe
OEt
OtBu
OMe
OEt
OMe
OEt
OiPr
OtBu
OMe
OEt
OtBu
OMe
OEt
OtBu
OtBu
OEt
53a, b
72a, b
53a
c
92a from 7ad, e, 47a from 6T
89a from 7ae
94a from 7ae
72a from 8Td, 69a from 9Td, e
H
H
74a from 8T
c
H
H
CF3
CF3
CF3
CF3
CF3
CF3
H
H
H
H
H
80f, g
60a
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CH3
CH3
H
95a from 11A, 67a from 10T
93a from 11A, 64a from 10T
95a from 11A
89a
93a
90a
67a
60h
s
t
u
v
x
y
z
52a
c
H
Cl
CF3
CF3
c
i
CF3
CF3
CH3
Cl
Cl
c
OtBu
a Isolated yield, after distillation. b Crude yield >95% by NMR and/or GC c Not isolable due to elimination of isobutene, affording 19 (see Table 4).
d Two precursors, see Scheme 2. e The major isomer 4 is listed; this is in equilibrium with the minor isomer 5. f Isolated by preparative TLC. g X-ray
crystal structure, see Fig. 5. h Yield estimated by 1H NMR. i Not isolated; see Ref. 2.
if one proton in 4b is irradiated, e.g. 4-H, the signal for 7-H
disappears as well (Fig. 1b). If 5-H is irradiated, the signal
for 6-H disappears too (Fig. 1e–f ). Therefore, there is fast
and degenerate interconversion of 4b and 5b, presumably by
exchange of H between 1-N and 3-N. However, the rate of
exchange was not experimentally accessible using conventional
coalescence technique.6 Line broadening occurred in the 1H
NMR spectrum on heating the solution of 4a to 130 ЊC, but
the coalescence temperature was not reached (Fig. S1 in the
ESI material‡). Therefore, we used the Forsén–Hoffman double
resonance saturation transfer method7 to determine the free
energies of activation for the H-exchange interconverting 4
and 5.
The Forsén–Hoffman method is limited to cases where the
rate of exchange approximately equals the longitudinal spin–
lattice relaxation time T 1. Nuclear Overhauser effects in proton
NMR can limit the accuracy of the method when the two
exchanging nuclei are in the same molecule. Observing a 13C
1
nucleus instead of H overcomes this problem. Saturation of
a site, e.g. νb, by irradiation with a strong radiofrequency field
causes perturbation of its spin distribution, which is transferred
to the site νa which is being observed. The lifetime τa of this site
depends on the relaxation time T 1 and the rate constant for the
exchange, ka. Therefore, T 1 must be determined by the standard
inversion–recovery technique using the pulse sequence (T D–π–
τ–π/2)n. τa and ka are then obtained from the equations
Scheme 2
τa = T 1 × {Mza(∞)/Mza(0) Ϫ (Mza(∞)}
(1)
In some cases, individual signals for 4 and 5 can be observed
in the NMR spectra, e.g. for 4h/5h and 4i/5i (see experimental
data in the ESI material for details‡), but in most cases only one
set of signals is observed. In the case of 4h/5h the isomer ratio
and
ka = 1/τa
(2)
1
is 2 : 1. Irradiation of one of the H NMR signals of 4h, for
example, not only allows identification of the proton with
which it couples; it also causes the complete disappearance
of the corresponding peak of 5h. Therefore, the two isomers
exist in a fast equilibrium. Compounds substituted only on 2-C
exhibit individual signals for 4-H, 5-H, 6-H and 7-H, as seen for
where Mza(∞) is the measured intensity of magnetization at site
νa, and Mza(0) is the measured intensity of magnetization in the
absence of irradiation. Knowing ka, activation parameters can
then be obtained from the Arrhenius and Eyring equations.6
Fig. 2 shows an example of a saturation–transfer experiment
using the 2-isopropoxy-1,3-diazepine 4c in CD2Cl2 solution.
1
example in the H NMR spectrum of 4b in Fig. 1a. However,
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1228

View Article Online
Fig. 1 Decoupling experiments on 4b. 1H homonuclear decoupling with selective irradiation at each resonance as indicated by arrows.
Fig. 2 Saturation–transfer experiment: stacked 13C NMR spectra (100 MHz) of 2-propoxy-1H-1,3-diazepine 4c obtained as a function of temper-
atures in CD2Cl2 solution. Carbon C-7 was irradiated and carbon C-4 observed. 32 Scans were collected in each run.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1229

View Article Online
Table 2 Activation parameters for hydrogen exchange between N1 and N3 in 2-alkoxy-1H-1,3-diazepines 4a
∆G‡298/kcal
Compd
Solvent
Ea/kcal molϪ1
A/sϪ1
∆H‡/kcal molϪ1
∆S‡/cal molϪ1 KϪ1
molϪ1
4aa
4ba
4ca
4ca
4ca
4cb
CD2Cl2
Acetone-d6
CD2Cl2
Acetone-d6
CD3OD
Acetone-d6/D2O
7.9 0.3
7.9 1.1
9.2 0.2
11.4 0.1
12.3 0.1
—
3.2 × 106
2.3 × 107
2.8 × 107
4.1 × 109
8.7 × 109
—
7.3
8.3
8.6
10.8
11.7
—
Ϫ30.8
Ϫ26.9
Ϫ24.5
Ϫ16.5
Ϫ15.0
—
16.5 0.3
16.3
1
15.9 0.2
15.8 0.1
16.2 0.1
14.1 0.6
a Forsén-Hoffman method. b Coalescence experiment, Tc = 308 0.5 K.
Fig. 3 Inversion–recovery experiment on 4c. 13C NMR at 125 MHz in CD2Cl2. The times indicated in each spectrum are the delay times in s after
which the 90Њ sampling pulse was applied after the inversion. The derived spin–lattice relaxation times T 1 are C2, 20.37; C4, 5.51; C7, 5.29; C5, 4.76;
C6, 4.84; OiPr(CH), 7.59; OiPr(CH3), 4.04 s.
The corresponding inversion–recovery experiment with deter-
mination of T 1 is shown in Fig. 3. Analogous experiments were
carried out with 4c in acetone-d6 and methanol-d4, with 4a
in CD2Cl2 and with 4b in acetone-d6. In all cases, excellent
Arrhenius and Eyring plots were obtained,8 and the resulting
Fig. 4 1,3-H-exchange in hydrogen bonded 1,3-diazepine molecules.
activation parameters are given in Table 2. The average free
energy of activation, ∆G‡ = 16.2 0.6 kcal molϪ1, for the
298
three diazepines 4a–c reveals that there is hardly any structural
or solvent effect in the three solvents used (average ∆H‡ = 9.3
2 kcal molϪ1; ∆S‡ = 22.7 8 cal KϪ1 molϪ1). The most likely
mechanism is intermolecular H-transfer between neighbouring
diazepine molecules (Fig. 4). Only by using acetone/D2O as
a solvent was a decrease in the activation free energy to 14.1
0.6 kcal molϪ1 observed (Table 2). The deuterium isotope effect
can be expected to cause a slightly decreased rate of exchange
(cf. data for CD3OD in Table 2), but the net rate increase
observed with D2O indicates intermolecular H(D)-exchange
between diazepine and (deuterated) water molecules, whereas in
the other solvents the exchange is between diazepine molecules
(Fig. 4). The exchange rate was fast enough for 4c in acetone/
D2O to be measured by the conventional coalescence technique
(Fig. S2 in the ESI material‡). The coalescence temperature
for H4/H7 was 308
0.5 K, from which the free energy of
activation was calculated.
The structures of the 1,3-diazepines are boat-like. The
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1230

View Article Online
a
Table 3 1H- and 5H-1,3-Diazepines 15 and 16 from reaction of tetrazoles/azides 1T/1A with amines HNR2
R1
R2
R3
R4
NR2
Yield 1H 15 (%)
Yield 5H 16 (%)
a
b
e
H
H
CF3
CH3
CH3
CH3
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
NEt2
—
—
64b
—
—
—
—
61b
N(iPr)2
N(iPr)2
NMe2
NEt2
N(iPr)2
N(iPr)2
67b, c
20b, d
t
67 from 17, 63 from 18b, c
64 from 17, 68 from 18b, e
u
ua
v
46 from 17, 39 from 18b, e
f
a Other 5H-1,3-diazepines have been reported.2 b Isolated yields after distillation or chromatography. c Crude yield >95% by 1H NMR. d This
compound exists in the 7-trifluoromethyl-5H-form. e Two precursors; see Scheme 4. f Rearranges to 2-(diisopropylamino)pyrrole-3-carbonitrile;2
cf. Scheme 11.
structure of the 1H-1,3-diazepine 4k was determined by single
crystal X-ray structure (Fig. 5).§ This is the first reported
crystal structure of any monocyclic N-unsubstituted azepine.
A few polycyclic 1H-1,3-diazepines9 and 1-benzoyl-1,3-diaze-
pines,1b,10 including 12e,1b have been reported, and a 5H-1,3-
diazepine2 was described recently. The bond lengths and angles
are as expected; the C2–N3 (1.275(3) Å) and C2–N1 (1.382(3)
Å) bonds are consistent with a proton residing at N1, and this
was confirmed by its identification from difference maps during
refinement. The C–C single (C5–C6 1.487(4) Å) and double
bonds (C4–C5 1.325(4) Å and C6–C7 1.328(4) Å) within the
seven-membered ring are also clearly defined.
Scheme 3
Fig. 5 ORTEP view of 1H-1,3-diazepine 4k (30% ellipsoids).
The 1H-diazepines react with acid chlorides to afford 2-
benzoyl derivatives 12b,e,h,i,m and the 2-acetyl derivative 13b.
Catalytic hydrogenation of this compound affords the tetra-
hydro-derivative 14. The amidine structure being very stable,
this compound could not be reduced further (Scheme 3).
2. 3H-1,3-Diazepines
Analogous photolysis of the tetrazoles/azides 1 in the presence
of secondary amines presumably also affords the 1H-1,3-
diazepine as the initial product, which was isolable for the first
time in the case of 15e. This compound was fully characterised
but isomerised to the 5H isomer 16e on heating, best by gas
chromatography at 130 ЊC, and partially on flash vacuum
thermolysis at 600 ЊC. In all other cases the 5H isomers 16 were
obtained directly in the photolysis reactions. The isolated com-
pounds are listed in Scheme 4 and Table 3. In agreement with
the experimental observations, theoretical calculations showed
that the 2-alkoxy-1,3-diazepines are most stable in the 1H-form,
Scheme 4
but 2-dialkylamino-1,3-diazepines are usually more stable in
the 5H-form.1a We have not found any limitation in the kind
of substituent that can be used. The unsubstituted 2-dialkyl-
amino-5H-1,3-diazepines 16a–c are reported here, as well as tri-
fluoromethyl- and methyl-substituted analogs. Several chloro-
and alkoxy-substituted 2-dialkylamino-5H-1,3-diazepines have
been described previously,2 and cyano-substituted analogs will
be reported.11
The 1H- and 5H-isomers 15 and 16 are related by 1,5-sigma-
tropic shifts of hydrogen. Substituted derivatives can in
principle exist in two isomeric forms, 15Ј and 16Ј (Scheme 4).
Unlike the alkoxy-substituted 1H-derivatives 4(5) and the single
§ CCDC reference numbers: 4k: 225886; 22bB: 226705 26: 225887.
See http://www.rsc.org/suppdata/ob/b3/b317099c/ for crystallographic
data in cif or other electronic format.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1231

View Article Online
Table 4 Diazepinones 19 and diazabicycloheptenones 25
R1
R2
R3
R4
ROH
Yield 19 (%)
Yield 25 (%)c
a
a
g
j
H
H
H
H
H
CF3
H
H
H
H
H
H
H
H
H
CF3
H
H
H
H
H
H
H
H
H
CF3
H
Cl
Cl
H
Cl
H
H2Oa
76
48
98
—
98
98
98
98
98
—
—
98
98
98
tBuOHb
tBuOH
tBuOHb
tBuOH
tBuOH
H2Oa
CF3
H
CF3
CF3
CH3
H
98d
61
p
s
98d
100e
81
80
75
73
53
67
t
v
tBuOHb
H2Oa
v
H
x
z
CF3
CF3
H
H
H
CH3
Cl
CF3
H
tBuOHb
tBuOHb
H2Oa
aa
a By photolysis of 1T in the presence of water. b By photolysis of 1T in the presence of tBuOH. c By photolysis of 19 for 7–8 h. d By heating 4g or 4p at
100–120 ЊC for 30 min. e By heating 4s at 70–80 ЊC for 15 min.
isolated amino-substituted 1H-derivative 15e, the amino-
substituted 5H-derivatives 16 tend to exist preferably with a sub-
stituent at C4 next to the methylene group at C5 (e.g. 16t,u,ua
(Scheme 4)). 16e is an exception (Table 3). Isomeric tetrazolo-
pyridines 1T afford the same carbodiimides 3 and the same
diazepine product. Thus, tetrazoles 17 and 18 both afforded the
same diazepines 16 (Scheme 4 and Table 3).
Moreover, the diazepinones 19 are conveniently obtained in
high yields by photolysis of the appropriate tetrazoles/azides 1
in the presence of water (Scheme 5 and Table 4). Except for the
4-chloro derivative 19v described below, they are stable, crystal-
line compounds, and they have all been fully characterized.
Although prepared by trapping of 3v in dioxan/water solu-
tion, once isolated, the 4-chlorodiazepinone 19v is sensitive
to moisture. Its reaction with H2O in DMSO-d6 solution was
investigated in situ by 13C NMR spectroscopy as shown in
Fig. S6 in the ESI material.‡ Compound 19v gradually dis-
appeared to be replaced by a new compound, which features
two new quaternary carbons at 170 and 153 ppm, two further
sp2methine carbons at 125 and 106 ppm, and a CH2 group at
34 ppm (verified in a DEPT-135 experiment). The corre-
The barriers to hindered rotations of the dialkylamino
groups about the C–N single bonds in several 2-dialkylamino-
5H-1,3-diazepine derivatives have been measured by NMR
methods (∆G‡ = 15–16 kcal molϪ1).8 The free energies of
298
activation for ring inversion of the 5H-1,3-diazepines (9.5–12
kcal molϪ1) have also been measured8 and will be the subject
of a forthcoming publication.
1
sponding H NMR signals were a doublet of doublets at 6.0
(7-H), a quartet typical of 5H-1,3-diazepines at 5.1 (6-H), and
a doublet at 2.95 ppm (CH2). The 13C NMR spectrum is in
excellent agreement (sum of deviations 15.8 ppm) with the cal-
culated spectrum of the diazepinedione 20a (Scheme 6). NMR
calculations were carried out at the B3LYP/6–31ϩG**//MP2/
6–31G* level, which was previously established as the most
accurate for this type of molecules.12 The other hypothetically
possible tautomers 20b and 20c have calculated energies 16.4
and 19.1 kcal molϪ1 higher than 20a, respectively. Energies were
calculated at the MP2/6–31G* level of theory. Since HCl is
formed in the reaction, the compound could also exist as the
salt 21 (Scheme 6), or in equilibrium with this salt. Of the two
tautomers 21a and 21b, the former is calculated to be 8.5 kcal
molϪ1 lower in energy. The calculated 13C NMR spectra for
these tautomers are in poorer agreement with the experimental
spectrum than that of 20a (sum of deviations 29.7 and 50.7 kcal
molϪ1, respectively). The computational data are tabulated in
the ESI material (Tables S4–S10).‡
3. 1,3-Diazepinones
We found that all the 2-tert-butoxy-1,3-diazepines 4 reported in
Table 1 as well as some of the 2-isopropoxy derivatives (4c,o)
were thermally unstable and eliminated isobutene or propene,
respectively, to furnish 1,3-diazepin-2-ones 19 (Scheme 5 and
Table 4). In several cases the 2-tert-butoxy-1,3-diazepines were
not isolable because they underwent isobutene elimination
already during the initial photochemical preparation from 3.
When the isolable tert-butoxy derivative 4s was heated above
its melting point (40–41 ЊC), the melt suddenly solidified at
ca. 80 ЊC to give white, crystalline diazepinone 19s. When the
1
thermal reaction of 4p was monitored by H NMR spectro-
scopy (DMSO-d6 solution) at 87 ЊC, isobutene was readily
observed at 1.67 and 4.64 ppm along with peaks due to 19s. The
reaction was complete in 45 min. The analogous elimination of
propene from 4o in DMSO-d6 solution had a half-life of ca. 4 h
at 130 ЊC.
Scheme 6
In order to obtain rigorous information on the nature of the
substitution products derived from 19v, the reaction with
amines was investigated. This could in principle give rise to any
Scheme 5
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1232

View Article Online
Table 5 Calculated relative energies (kcal molϪ1; 0 K) of structures
20a–c, 21a–b and tautomers 22aA–F at various levels of theory with
inclusion of scaled ZPVE (cf. Schemes 6 and 7)
NMR could be due to either NH or OH. Strong hydrogen
bonding would make the experimental IR spectra of tautomers
A or B more similar and lower the C᎐O stretching frequency in
᎐
B. A B3LYP/6–31G* calculation of a H-bonded dimer of 22aB
moved the carbonyl group vibration closer to the experimental
frequency (calculated: 1676; experimental: 1621 cmϪ1 with a
weak shoulder at 1650 cmϪ1; for the full calculated IR data
see Table S5 in the ESI material‡).
HF/6–31G*
B3LYP/6–31G*
MP2/6–31G*a
20a
20b
20c
0.0
18.3
20.1
0.0
16.2
18.7
The problem was resolved by recording the IR spectra in Ar
matrices at 20 K, where discrete molecules, devoid of inter-
molecular H-bonding, are observed. The IR spectrum of 22a
showed no band due to OH, but bands at 3444 (NH), 1690
21a
21b
0.0
Ϫ15.3
0.0
8.2
22aA
22aB
22aC
22aD
22aE
22aF
0.0
Ϫ11.6
11.3
Ϫ8.8
10.6
0.0
Ϫ10.3
12.0
Ϫ5.7
8.7
0.0
Ϫ9.8
13.1
Ϫ6.6
9.3
(C᎐O), 1654, and 1612 cmϪ1, and the appearance was in very
᎐
good agreement with the calculated spectrum for 22aB in the
gas phase (B3LYP/6–31G*). Therefore, the discrete molecules
exist in the most stable tautomeric form, B.
An X-ray crystal structure determination of 22b confirmed
that the molecule crystallizes as a centrosymmetric H-bonded
dimer of 22bB (Fig. 6 and Fig. S8). The H atom was located on
N1 and refined without any constraints. There is a strong and
short hydrogen bond (N1–H1 ؒ ؒ ؒ O2Ј 2.04(2) Å, 174(2)Њ, O2Ј
symmetry code 1 Ϫ x, 2 Ϫ y, 1 Ϫ z), and a correspondingly long
Ϫ9.4
Ϫ6.2
Ϫ5.8
a the ZPVE from the HF/6–31G* calculation was used.
of the six tautomers of 22 A–F (Scheme 7). Ab initio calcu-
lations (HF/6–31G*, MP2/6–31G* and B3LYP/6–31G*) con-
firmed chemical intuition that B is the most stable tautomer,
followed by D, F, A, E and C in the order of increasing energy
(Table 5 and Table S8). The compounds 22 obtained show a
CH2 group in the proton and carbon NMR spectra. Tautomers
C, E, and F are in discord with the 1H NMR spectra. Tautomer
C᎐O double bond (1.245(2) Å). Also of note is the trigonal
᎐
planar coordination geometry of the exocyclic amino group
N4 (dihedral angles N3–C4–N4–C8 1.3(3)Њ and C5–C4–N4–C1
1.2(3)Њ); the C4–N4 bond length (1.340(2) Å) is intermediate
between single and double bonds, and the N4–C4–N3–C2–O2
moiety is almost planar. Thus, the compound has some degree
of zwitterionic character with a partial positive charge on
N4 and a partial negative charge on O2 (see structure G in
Scheme 7).
B is in excellent accord with the calculated 13C and H NMR
1
data (for full details see the ESI material, Tables S6–S7‡). How-
ever, the urea-type structures B, D, and F have calculated C᎐O
᎐
stretching vibrations at 1717, 1725 and 1740 cmϪ1, respectively
(for 22a in the gas phase at the B3LYP/6–31G* level; for full
details see Table S5). In contrast, the highest experimental
wavenumbers that come into consideration for a carbonyl
group are in the low-to-middle 1600 cmϪ1 range. Therefore, at
first sight, the 2-hydroxy tautomer A appeared to fit the IR
spectroscopic data better than tautomer B.
Fig. 6 ORTEP view of 1,3-diazepin-2-one 22bB (30% ellipsoids).
The chlorine atom in the 2-ethoxydiazepine 5y/4y is also
prone to substitution.2 This compound is generated by photo-
lysis of the azidopyridine 1Ay in the presence of ethanol
(Scheme 8). The facile dark reaction with diethylamine to
furnish 23 in 81% yield has been described.2 Compound 5y/4y
reacts with water to give a 57% yield of the 1,3-diazepin-4-one
24, which features a carbonyl group at 162 ppm (1705 cmϪ1
)
and an NH function at 7.8 ppm (broad) in the 1H NMR
(Scheme 8).
4
Diazabicycloheptenes
The cyclic ureas 19 all undergo electrocyclic photocyclization,
formally disrotatorily, to afford 2,4-diazabicyclo[3.2.0]hepten-
3-ones 25 in virtually quantitative yield (Scheme 9 and Table 4).
Several NMR spectra showing this transformation are repro-
duced in the ESI material (Fig. S4–S6‡). This typically requires
12 h of photolysis, i.e. the photocyclization is much slower than
Scheme 7
However, the IR spectra in KBr clearly show highly hydrogen
bonded NH or OH groups absorbing between 3600 and 2800
cmϪ1, and a weak and broad signal near 8.5 ppm in the H
1
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1233

View Article Online
Scheme 8
Scheme 10
Scheme 9
the synthesis of the diazepinones 19, thereby allowing the latter
to be prepared. In the case of the photolysis of 6,7-bis(trifluoro-
methyl)tetrazolo[1,5-a]pyridine 1Tk in dioxane/water for 2 h
and 40 min, the diazabicycloheptanone 26 was unexpectedly
obtained in 48% yield. The mechanism of this presumed photo-
reduction reaction had not been investigated, but the structure
of 26 was unequivocally established by X-ray crystallography,‡
thereby lending further support to the structural characteriz-
ation of the bicycloheptenones 25. The crystal structure and
essential data for 26 are given in the ESI material (Fig. S9). We
have observed an unexpected photoreduction/photocyclization
in another case, viz. the formation of 3-ethoxy-2H-2,4-diaza-
bicyclohept-3-ene on photolysis of tetrazole 1Tk in the presence
of ethanol, but this reaction was not investigated further.¶
Prolonged photolysis of some of the tetrazoles/azides 1T/1A
in the presence of amines without isolation of the 2-dialkyl-
amino-1,3-diazepines 15 or 16 also caused photocyclization
to 3-substituted 2,4-diazabicycloheptadienes 27, which were
isolated in modest yields in several cases (Scheme 10). This
presumably takes place via the 1H tautomers 15 (isolable in the
case of 15e, vide supra) and it suggests that there is a dynamic
equilibrium between 16 and 15 (Scheme 11).
Scheme 11
imidazole moiety. The 1H and 13C NMR data rule out the alter-
native 4π electrocyclizations in 15 or 16 leading to 1,6- or 2,7-
diazabicyclo[3.2.0]heptadienes 28–30, which would have been
analogous to the photocyclizations reported for 2-dialkyl-
amino-3H-azepines13a and 2,5-bis-tert-butoxy-3H-azepine.13b
Nevertheless, we know from previous work that photocycliz-
ations of the type 15 29 or 16 30 do take place when appro-
priate substituents are present to drive the reaction forward to
the 3-cyanopyrroles 31 or 32.2 In the present case, it is likely that
a photoequilibrium exists between all the possible electro-
cyclization paths, in which the imidazole derivative 27 becomes
the dominant product (Scheme 11).
The vinylic and aliphatic cyclobutene protons give rise to
1
only one signal each in the H NMR spectra of 27a,b, thereby
implying rapid H-exchange between the N-atoms of the
¶ Other unexpected reactions have been observed in this work, e.g. the
acquisition of an extra carbon atom with formation of 2-formylamino-
3,5-dichloropyridine from 6,8-dichlorotetrazolo[1,5-a]pyridine in
a
dark reaction with diethylamine or dipropylamine. After stirring the
tetrazole with an amine for several days in cyclohexane or in dioxane
the crude residue showed the presence of the starting material and
about 20% (by GCMS) of 3,5-dichloropyridine-2-N᎐CHNEt2 and -2-
Conclusion
᎐
N᎐CHNPr2, respectively. These two products were not isolated, as they
᎐
Photolysis of tetrazolo[1,5-a]pyridines/2-azidopyridines 1T/1A
gives rise to nitrogen elimination and formation of 1,3-diaza-
decomposed on silica gel to give 2-formylamino-3,5-dichloropyridine,
whose structure was verified by X-ray crystallography.
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1234

View Article Online
cyclohepta-1,2,4,6-tetraenes 3. In the presence of alcohols, 3
is trapped to afford 2-alkoxy-1H-1,3-diazepines 4(5) in good
to excellent yields. The 1H-1,3-diazepines 4(5) undergo rapid
hydrogen exchange between positions 1 and 3, with free
diazabicyclo[3.2.0]heptenes are reported below. Melting points
are uncorrected.
Exchange kinetics
energies of activation ∆G‡ of the order of 16.2 0.6 kcal
298
molϪ1 with little or no solvent effect; most likely, this process
takes place by intermolecular H-exchange between neigh-
bouring diazepine molecules (Fig. 4). A lower free energy of
The variable temperature saturation transfer experiments
using the Forsén–Hoffman double resonance method7 were
performed on a Bruker AMX400 spectrometer operating at
400.136 MHz for 1H and 100.614 MHz for 13C. The probe was
temperature calibrated using methanol as standard. The
reported temperatures are estimated to be accurate to 0.5 ЊC.
All solutions were prepared by dissolving 50–60 mg in 1 ml of
the appropriate deuterated solvent and degassing by three
freeze–pump–thaw cycles. The single frequency irradiation,
which was used to irradiate the selected exchanging carbon
peak, was set to the frequency of that carbon resonance.
Irradiation at this frequency was switched on for 40 seconds
(equal to approximately 5 × T 1), and with the minimal delay of
24 microseconds a 70Њ pulse was applied followed by collection
of 64K data points. The irradiation power was selected such
that the irradiated resonance was fully saturated but was not
high enough to irradiate the magnetized resonance. The 1H
Waltz decoupler was switched on all the time. The solution was
allowed to equilibrate for at least 10–15 minutes before 32 scans
were collected at each temperature. The probe tuning and
shimming was adjusted for each temperature. The intensity
of the magnetized resonance was measured directly as the
absolute value. The rates of exchange k were measured over
the temperature range 312–296 K.
activation of 14.1
0.6 kcal molϪ1 for 4c in acetone/D2O
indicates H-exchange between diazepine and water molecules in
this case. X-ray crystallography of 4k reveals a twist-boat type
structure of this molecule.
Trapping of the 1,3-diazacyclohepta-1,2,4,6-tetraenes 3 with
amines affords 2-dialkylamino-5H-1,3-diazepines 16, but the
corresponding 2-diisopropylamino-1H-1,3-diazepines 15e was
isolated in one case and underwent partial isomerization to 16e
on heating. Ab initio calculations confirmed that the 2-alkoxy-
1,3-diazepines are more stable in the 1H form, whereas the
2-dialkylamino-1,3-diazepines are usually more stable in the
5H-form.1a
Trapping of the 1,3-diazacyclohepta-1,2,4,6-tetraenes 3 with
water affords the novel 1,3-diazepin-2-ones 19 in high yields.
The same compounds are also obtained by very easy thermal
elimination of isobutene from the 2-tert-butoxy-1H-1,3-
diazepines 4(5) or less readily from the 2-isopropoxy analogues.
Reaction of 19v with water affords the 4-dialkylamino-1,3-
diazepin-2-ones 22B, which exist as hydrogen bonded dimers in
solution and in the solid state.
Prolonged photolysis of the 1,3-diazepin-2-ones 19 causes
electrocyclization to 2,4-diazabicyclo[3.2.0]hept-6-en-2-ones
25 in virtually quantitative yields. In some cases unexpected
photoreduction to diazabicycloheptane derivatives such as 26
takes place.
In some cases, prolonged photolysis of tetrazoles/azides 1T/
1A in the presence of amines, i.e. conditions expected to lead to
diazepines 15 and/or 16, also causes cyclization to 3-dialkyl-
amino-2H-2,4-diazabicyclohepta-3,6-dienes 27.
The 13C NMR spectra of 4c in CD2Cl2 at various temper-
atures are shown in Fig. 2. When the C-7 carbon was irradiated
at 293 K, the C-4 carbon was fully saturated, and hence no
signal was observed for this site. Lowering the temperature
slowed down the exchange, and already at 283 K a small signal
for C-4 was observed. Gradually lowering the temperature to
223 K resulted in complete recovery of the C-4 signal. Hence
the rate of exchange at this temperature must equal zero on the
NMR time scale. The exchange rate for the process was then
calculated from eqns. (1) and (2) above. The relaxation times T 1
for C-4 and all other carbon atoms were obtained by standard
inversion–recovery experiments and are given in the caption
of Fig. 3. The intensity of the magnetized resonance C-4 in
the absence of irradiation, MzC4(0), and in the presence of
irradiation, MzC4(8), were measured, and the spin lifetime
τ(C-4) was calculated for each temperature following eqn. (1).
The exchange rate k at each temperature was then calculated
from eqn. (2). The data are presented in Table 6. Measurements
of T 1 and k were performed in a similar manner for 4c in
acetone-d6 (Table 6), for 4b in acetone-d6 and for 4a in CD2Cl2.
Relaxation times of all carbons were slightly longer in acetone-
d6 (for 4c 24.2, 5.6, 5.7, 5.1, 5.1, 8.4 and 4.1 s for C-2, C-4, C-7,
C-5, C-6, OiPr(CH) and OiPr(CH3), respectively).
Computational methods
Standard ab initio molecular orbital calculations14 were carried
out using the Gaussian 98 system of programs.15 Geometry
optimisations were performed with the polarised split-valence
6-31G* basis set at the Hartree–Fock (HF), second-order
Møller-Plesset pertubation (MP2) and the hybrid density
functional theory B3LYP levels of theory. The frozen-core
approximation was employed for all correlated calculations.
Harmonic frequencies were calculated at the HF/6–31G* and
the B3LYP/6–31G* levels in order to confirm the stationary
points as minima and to evaluate zero-point vibrational
energies (ZVPEs). The directly calculated ZVPEs were scaled
by 0.9135 (HF/6-31G*)16 and 0.9806 (B3LYP/6-31G*)17 to
account for their overestimation at this level of theory. 13C and
1H NMR chemical shifts were calculated at the B3LYP/6-
31ϩG**//MP2/6-31G* level using TMS as a reference.12 The
calculated harmonic frequencies are required to be scaled to
account for the neglect of anharmonicity effects in the theoreti-
cal treatment. For the B3LYP/6–31G* IR spectra reported
herein, a factor of 0.9614 has been used.17
In the case of 4c in acetone-d6/D2O, 2 drops of D2O was
added to the acetone-d6 solution that had been used previously
for kinetic measurements. This resulted in broadening of all
carbon signals except that of C-2, and therefore saturation by
irradiation was not possible. Increasing the temperature to
308 K caused further broadening of the protonated ring carbon
signals and their complete disappearance in the spectral base-
line. Cooling to 263 K slowed the rate of exchange so that the
peaks became just visible. Further cooling to 243 K resulted
in full recovery of all carbon signals, now indicating slow
exchange (Fig. S2). This experiment indicated that the coales-
cence temperature should be ca. 308 K. Measurements of the
1H NMR spectra as a function of temperature confirmed the
coalescence temperature as 308 0.5 K, from which ∆G‡ for
N,N–D exchange was calculated (Table 2).6
Experimental
The unfiltered light from a 1000 W high pressure Hg/Xe lamp
was used for all irradiations. Tetrazolo[1,5-a]pyridines/2-azido-
pyridines were prepared as previously described,2,5a except 2-
azido-3,6-bis(trifluoromethyl)pyridine 1Aq, which is described
in the ESI material.‡ Preparative details and characterization
data for most of the compounds prepared are given in the ESI
material.‡ Only the general experimental procedure and data for
representative and new types of diazepines, diazepinones and
In the case of 4c in methanol-d4 the initial 1H NMR spectrum
at 301 K showed very broad signals for all protons. Unlike the
previous experiment in acetone-d6/D2O, warming to 330 K did
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1235

View Article Online
Table 6 Exchange rates k/sϪ1 and spin state lifetimes τ/s for C-4 or C-7 at temperatures T /K 0.5 K for 4c in various solvents
CD2Cl2
Acetone-d6
Methanol-d4
T
k
τ(C4)
0.716
0.962
0.866
1.361
4.089
T
k
τ(C7)
0.366
0.858
1.813
2.684
4.380
7.050
12.25
22.84
45.95
T
k
τ(C4)
0.348
0.541
0.712
1.098
1.971
3.262
4.726
9.034
21.53
40.82
283
275
268
260
252
242
233
223
213
1.396
1.040
1.154
0.735
0.245
0.116
0.079
0.023
0.0078
273
263
253
248
243
238
233
228
223
2.73
1.16
0.55
0.37
0.23
0.14
0.082
0.044
0.022
285
280
275
270
263
258
253
247
240
234
2.877
1.849
1.404
0.911
0.507
0.306
0.212
0.111
0.0464
0.0245
8.589
12.809
44.209
128.72
not result in further broadening but marginal sharpening of the
signal. In contrast, upon renewed cooling to 298 K sharpening
to fully resolved and coupled peaks occurred. This was an
irreversible phenomenon caused by initial fast NH,N exchange,
followed by H–D exchange (accelerated and completed on
warming to 330 K), and a slower ND,N exchange in the
resulting N-deuterated diazepine. H-7 had become a doublet
due to coupling with H-6 and uncoupling from the now
deuterated NH function. A Forsén–Hoffmann saturation
transfer experiment could now be carried out on this solution,
where each of the four ring carbon atoms could be irradiated
while its partner was being observed. On irradiation of C-4 at
139.9 ppm above 290 K the C-7 resonance at 132.6 ppm was
fully saturated, indicating fast exchange. Gradual cooling
caused an increased intensity of the C-7 peak till 243 K (slow
exchange). T 1, τ and k for the ND,N exchange were obtained as
described above (Table 6). Except for C-2, the T 1 values were
shorter in methanol-d4 than either acetone-d6 or CD2Cl2 (28.5,
4.3, 4.1, 4.2, 4.4, 6.0 and 3.3 s for C-2, C-4, C-7, C-5, C-6,
OiPr(CH) and OiPr(CH3), respectively).
diazepin-2-ones the starting azido/tetrazolo[1,5-a]pyridine
(ca. 1 mmol) was dissolved in a mixture of absolute dioxane
(120 ml) and the appropriate alcohol or water (20–30 ml) in a
quartz vessel. The mixture was purged with high purity dry
nitrogen for about 1 h. The degassed solution was irradiated
with the high pressure Hg/Xe lamp while stirring the mixture in
an ice bath. Reaction times were usually 1–2 h. Dialkylamino-
diazepines were prepared analogously. When dimethylamine
was needed, the gas was passed directly into the reaction mix-
ture, or a saturated solution of dimethylamine (large excess) in
dioxane was used. After the end of the reaction (monitored
by nitrogen evolution and thin layer chromatography, usually
1–2 h) the volume was reduced under vacuum, and the resulting
oily residue was purified by chromatography on deactivated
aluminium oxide (90, neutral). The aluminium oxide was
deactivated by aqueous methanol (20%) and then dried in the
air overnight at room temperature. Compounds were purified
by Kugelrohr distillation, column or preparative thin layer
chromatography.
Arrhenius and Eyring parameters were derived in the usual
manner6 by plotting ln k vs. 1/T and ln (k/T) vs. 1/T , respec-
tively (Fig. S7 in the ESI material). The r2 values for the
Arrhenius and Eyring plots were 0.980 (4c in CD2Cl2), 0.997 (4c
in acetone-d6), 0.990 (4c in methanol-d4), 0.980 (4a in CD2Cl2),
and 0.990 (4b in acetone-d6). All resulting activation parameters
are collected in Table 2.
5,6-Bis(trifluoromethyl)-2-methoxy-1H-1,3-diazepine 4k
Purified by preparative TLC (silica gel 100, CH2Cl2–MeOH 97 :
3). Crystallized from petroleum ether, yellow–orange crystals,
mp 120–121 ЊC. The crystals were suitable for X-ray data col-
1
lection (see below). Yield: 80%. H NMR (CDCl3, 200 MHz)
δ 7.12 (q, 1 H, 4-H, JH,F = 1.80 Hz), 6.42 (dq, 1 H, 7-H, J1,7
=
6.34 Hz, JH,F = 1.40 Hz), 5.15 (br, 1 H, N-H ), 3.81 (s, 3 H,
OCH3); 13C NMR (CDCl3, 50 MHz) δ 157.0 (2-C), 146.5 (q,
4-C, JC,F = 7.8 Hz), 139.5 (q, 7-C, JC,F = 7.6 Hz), 122.5 (q, CF3,
JC,F = 270.5 Hz), 122.1 (q, CF3, JC,F = 270.5 Hz), 115 (q, 5-C,
JC,F = 31.2 Hz), 113.2 (q, 6-C, JC,F = 32.4 Hz), 56.8 (OCH3);
MS (EI) m/z 260 (Mϩ, 100%), 241 (30), 218 (39), 203 (85), 184
(92), 177 (34), 156 (11), 153 (25), 148 (26), 121 (24), 120 (22),
98 (62), 76 (23), 75 (22), 69 (91), 58 (98), 42 (17). HRMS, calcd.
for 12C8H6N2F6O: 260.0367; found: 260.0390. Anal. calcd. for
C8H6N2F6O: C, 36.94; H, 2.33; N, 10.77%. Found: C, 36.96; H,
2.26; N, 10.73%.
Crystallography
Cell constants were determined by least-squares fits to the
setting parameters of 25 independent reflections measured on
an Enraf-Nonius CAD4 four-circle diffractometer employing
graphite-monochromated Mo Kα radiation (0.71073 Å) and
operating in the ω–2θ scan mode. Data reduction was per-
formed with the WinGX package.18 Structures were solved by
direct methods with SHELXS and refined by full-matrix
least-squares analysis with SHELXL-97.19 All non-H atoms
were refined with anisotropic thermal parameters. Hydrogens
attached to N-atoms were located from difference maps then
restrained in these positions using a riding model. All other H-
atoms were included at calculated positions and again allowed
to ride on their parent C-atom. Drawings were produced with
ORTEP.20 Crystallographic data in CIF format have been
deposited with the Cambridge Crystallographic Data Centre.§
Essential structural parameters are listed in the ESI material.‡
Crystal data. C8H6F6N2O, M = 260.15, monoclinic, space
group P21/c, a = 9.642(7), b = 10.277(7), c = 10.081(7) Å, β =
91.89(3)Њ, U = 998(1) Å3, Z = 4, Dc = 1.731 g cmϪ3, µ = 1.91
cmϪ1, 1854 reflections measured, 1746 unique (Rint = 0.0213),
R1 = 0.0494 (for 1331 observed data, I > 2σ), wR2 = 0.1637 (all
data). Crystallographic data in cif format have been deposited
with the Cambridge Crystallographic Data Centre.§ Essential
structural parameters are listed in the ESI material.‡
General procedures for the synthesis of 1,3-diazepines
2-Diisopropylamino-4-trifluoromethyl-1H-1,3-diazepine 15e
Purified by Kugelrohr distillation (∼50 ЊC, 10Ϫ4 mbar); red oil.
The azides/tetrazoles (100–150 mg) were photolysed in dry,
distilled, N2-purged 1,3-dioxane (“dioxane”) solutions using a
quartz vessel. Alcohols were dried using magnesium metal,
amines were refluxed over and distilled from potassium hydrox-
ide, and dioxane was distilled from Na metal immediately prior
to use.
1
Yield: 64%. H NMR (benzene-d6, 400 MHz) δ 5.90 (d, 1 H,
5-H, J5,6 = 5.7 Hz), 5.61 (dd, 1 H, 7-H, J6,7 = 7.2, J1,7 = 4.6 Hz),
5.40 (m, 1 H, 6-H), 4.30 (br s, 1 H, N-H), 3.91 (sept, 2 H, NiPr),
1.13 (d, 12 H, NiPr), 13C NMR (benzene-d6, 100 MHz) δ 157.9
(2-C), 137.1 (q, 4-C, JC,F = 32.4 Hz), 134.6 (7-C), 123.1 (CF3,
For the synthesis of 2-alkoxy-5H-1,3-diazepines and 1,3-
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1236

View Article Online
JC,F = 273 Hz), 118.2 (5-C), 110.4 (6-C), 44.9 (NiPr), 20.2
(NiPr); IR (neat film) 3398 br, 2954 m, 2946 w, 2913 w, 1611 s,
1553 vs, 1525 s, 1456 m, 1364 m, 1328 vs, 1295 vs, 1254 s, 1232
vs, 1151 m, 1137 m, 1100 m, 1052 m, 1035 w, 1016 w, 985 m,
932 w, 902 w, 893 w, 862 w, 796 w, 741.0 m, 634 w cmϪ1; MS (EI)
m/z 261 (Mϩ, 100%), 247 (8), 246 (29), 232 (14), 202 (33), 201
(12), 189 (6), 150 (49), 148 (13), 105 (21), 104 (8), 87 (36),
NMR (100 MHz, acetone-d6) δ 159.9 (2-C), 158.4 (4-C), 130.7
(7-C), 102.9 (6-C), 44.1 and 43.5 (NCH2CH3)2, 29.3 (5-C), 14.7
and 12.5 (NCH2CH3)2; IR (KBr) 3600–2800 vbr. (νmax 3425 and
3189), 2977 w, 2939 w, 1650 w sh, 1629 s, 1568 vs, 1464 m,
1454 m, 1437 m, 1396 m, 1382 w, 1366 w, 1322 s, 1270 m, 1245
m, 1204 w, 1188 w, 1147 w, 1127 w, 1097 w, 1078 w, 1068 m,
1017 w, 929 w, 823 m, 807 w, 778 w, 760 w, 715 m cmϪ1; MS (EI)
m/z (Mϩ, 100%), 154 (8), 153 (9), 152 (13), 139 (6), 138 (7), 125
(10), 110 (14), 109 (54), 97 (17), 90 (12), 83 (15), 82 (23), 72 (36),
70 (31), 69 (17), 65 (59), 64 ( 22), 63 (17), 58 (25) 56 (17), 55 (14),
54 (18), 45 (20), 44 (17), 32 (17), 30 (18). HRMS, calcd. for
12C9H15N3O: 181.120963; found: 181.121575. Anal. calcd. for
C9H15N3O: C, 59.64; H, 8.34; N, 23.19%. Found: C, 59.98; H,
8.63; N, 22.82%.
78 (13), 77 (22), 56 (6), 54 (5), 44 (7), 42 (11). HRMS, calcd. for
12
C H18F3N3: 261.14528; found: 261.14544. Anal. calcd. for
12
C12H18F3N3: C, 55.16; H, 6.94; N, 16.08%. Found: C, 55.44; H,
5.66; N, 16.61%.
2-Diethylamino-5H-1,3-diazepine 16a
Purified by distillation (Kugelrohr, 50–70 ЊC/0.5 mmHg). Pale
1
red oil. Yield: 61%. H NMR (acetone-d6, 301 K, 400 MHz)
¯
Crystal data. C9H15N3O, M 181.24, triclinic, space group P1,
δ 6.88 (d, 1 H, 4-H, J4,5 = 5.00 Hz), 6.64 (d, 1 H, 7-H, J6,7 = 6.46),
4.53 (q, 1 H, 6-H, J6,7 = 6.46, J5,6 = 6.47, J4,6 = 0.88 Hz), 3.41 (q,
4 H, N(CH2CH3)2, J = 7.04), 2.24 br (2 H, 5-H), 1.01 (t, 6 H,
NCH2CH3, J = 7.04 Hz); the assignments were confirmed by
homonuclear decoupling experiments; 13C NMR (100 MHz,
acetone-d6) δ 159.0 (2-C), 147.7 (4-C), 142.1 (7-C), 95.7 (6-C),
43.3 and 41.2 (N(CH2CH3)2), 32.8 (5-C), 13.7 and 13.6 (N-
(CH2CH3)2); IR (KBr) 3010 w, 2972 m, 2931 m, 2871 w, 1622 vs,
1599 s, 1588 s, 1575 s, 1557 m, 1520 vs, 1516 vs, 1460 m, 1446 m,
1422 m, 1376 m, 1356 s, 1318 m, 1286 m, 1254 m, 1228 w, 1209
w, 1161 m, 1096 w, 1080 m, 1015 w, 916 w, 883 w, 847 w, 783 w,
753 w, 722 w, 701 w cmϪ1; MS (EI) m/z 165 (Mϩ, 11%), 136 (5),
110 (3), 109 (10), 108 (7), 95 (9), 94 (13), 93 (17), 92 (5), 83 (7),
82 (8), 81 (43), 80 (7), 72 (35), 71 (5), 69 (10), 68 (28), 67 (100),
66 (11), 56 (17), 55 (12), 54 (12), 53 (10), 44 (11), 42 (10), 41 (12),
39 (18). HRMS, calcd. for 12C9H15N3: 165.1265; found:
165.1266. Anal. calcd. for C9H15N3: C, 65.42; H, 9.15; N,
25.43%. Found: C, 65.02; H, 9.19; N, 25.39%.
a = 7.607(2), b = 8.040(2), c = 8.369(1) Å, α = 96.79(2), β =
98.36(3), γ = 104.70(2)Њ, U = 483.4(2) Å3, Z = 2, Dc = 1.245 g cmϪ3
,
=
µ = 0.84 cmϪ1, 1839 reflections measured, 1697 unique (Rint
0.0242), R1 = 0.0434 (for 1181 observed data, I > 2σ), wR2 =
0.1244 (all data). Crystallographic data in cif format have been
deposited with the Cambridge Crystallographic Data Centre.§
Essential structural parameters are listed in the ESI material.‡
6,7-Bis(trifluoromethyl)-2,4-diazabicyclo[3.2.0]heptan-3-one 26
This compound was obtained unexpectedly when 250 mg
of 6,7-bis(trifluoromethyl)tetrazolo[1,5-a] pyridine 1Tk was
photolysed in dioxane/H2O (120/30 ml) solution for 2 h and
40 min. Solvent was removed under vacuum, and the residue
was chromatographed on silica gel 100 (70–230 mesh ASTM,
CHCl3/EtOH 85 : 15) to afford 120 mg (48%) of crystalline
solid, mp 80–81 ЊC. 1H NMR (400 MHz, acetone-d6) δ 7.15 (br
s, 1 H, N-H ), 7.05 (br s, 1 H, N-H ), both protons exchange
with D2O, 4.45 (m, 1 H, 1 or 5-H, J1,5 = 8.2 Hz), 4.27 (m, 1 H,
5 or 1-H, J1,5 = 8.2 Hz), 3.58 (m, 1 H, 6 or 7-H), 3.34 (m, 1 H,
7 or 6-H). Upon addition of D2O to simplify the spectrum,
a homonuclear decoupling experiment showed that the proton
at 4.45 couples with the proton at 3.58 ppm, and the proton at
4.27 couples with the proton at 3.34 ppm. Even under such
conditions (NH decoupled), all signals were observed as com-
plex multiplets due to H–F coupling. 13C NMR (100 MHz,
acetone-d6), 163.5 (3-C), 125.3 (q, CF3, JC,F = 275 Hz), 124.9 (q,
CF3, JC,F = 275 Hz), 50.4 (q, 1- or 5-C, JC,F = 3 Hz), 48.7 (q, 5- or
1-C, JC,F = 4 Hz), 46.2 (q, 6- or 7-C, JC,F = 26 Hz), 39.8 (q, 7- or
6-C, JC,F = 27 Hz), DEPT-135 indicated protonated carbons
at 50.4, 48.7, 46.2 and 39.8 ppm. IR (KBr) 3265 br (N-H ),
1710 vs, 1661 vs, 1455 m, 1413 m, 1361 s, 1319 s, 1300 w, 1265 m,
1249 vs, 1235 m, 1211 vs, 1202 s, 1166 s, 1135 s, 1124 m, 1094 vs,
953 w, 928 w, 845 w, 798 w, 738 w cmϪ1; MS (EI) m/z 249 (Mϩ ϩ
1, < 1%), 229 (1), 111 (6), 95 (4), 91 (3), 84 (100), 69 (10), 56
(20), 28 (17), MS (CI) m/z 249 (Mϩ ϩ 1, 100%), 85 (3), 84 (36);
HRMS calcd. for 12C7H7N2F6O: 249.045707; found: 249.
046065. Anal. Calcd for C7H6N2F6O: C, 33.88; H, 2.44; N,
11.31; found: C, 33.88; H, 2.49; N, 11.31.
2-Diisopropylamino-7-trifluoromethyl-5H-1,3-diazepine 16e
Prepared and purified by preparative GC of 15e at 130 ЊC. Red
1
oil, yield: 20%. Decomposes on silica gel and Al2O3. H NMR
(400 MHz, CDCl3) δ 6.27 (t, 1 H, 4-H, J4,5 = 5.2 Hz), 4.90 (t,
1 H, 6-H, J5,6 = 6.7 Hz), 3.95 (br, 2 H, iPr2), 1.65 (br, 2 H, 5-H),
1.11 (br, 12 H, iPr); the assignments were confirmed by homo-
nuclear decoupling experiments; 13C NMR (100 MHz, CDCl3)
δ 158.0 (2-C), 157.6 (4-C), 147.7 (4-C), 141.7 (q, 7-C, JC,F = 31.6
Hz), 123.0 (q, CF3, JC,F = 273 Hz), 95.9 (q, 6-C, JC,F = 3.4 Hz),
46.0 (iPr2), 31.5 (5-C), 20.6 (iPr); IR (neat) 2974 m, 2956 w,
2935 w, 1622 m, 1595 vs, 1542 w, 1520 w, 1511 w, 1487 w, 1472
w, 1398 w, 1370 m, 1322 w, 1301 s, 1206 m, 1164 s, 1118 s, 1071
w, 1021 w, 1008 w, 956 w, 839 w, 757 w, 697 w, 672 w cmϪ1. MS
(EI) m/z 261 (Mϩ, 49%), 242 (11), 218 (100), 203 (12), 201 (8),
175 (64), 161 (14), 150 (10), 149 (38), 148 (6), 126 (33), 117 (10),
107 (12), 83 (11), 69 (7), 58 (12), 54 (3), 43 (25), 41 (37). HRMS,
12
calcd. for C12H18F3N3: 261.14528; found: 261.14551. Anal.
calcd. for C12H18F3N3: C, 55.16; H, 6.94; N, 16.08%. Found:
C, 54.91; H, 6.58; N, 15.75%.
1,2-Dihydro-4-diethylamino-5H-1,3-diazepin-2-one 22bB
Crystal data. C7H6F6N2O, M = 248.14, monoclinic, space
group P21/c, a = 13.243(3), b = 5.6585(5), c = 13.304(3) Å, β =
116.61(1)Њ, U = 891.3(3) Å3, Z = 4, Dc = 1.849 g cmϪ3, µ = 2.09
cmϪ1, 1668 reflections measured, 1561 unique (Rint = 0.0396),
R1 = 0.0423 (for 928 observed data, I > 2σ), wR2 = 0.1223 (all
data). Crystallographic data in cif format have been deposited
at the Cambridge Crystallographic Data Centre.§ The crystal
structure and essential bond lengths and angles are reported in
the ESI material (Fig. S7).‡
This compound was prepared in a similar manner as described
for 22aB using diethylamine. The product was crystallized from
an ether–dichloromethane (10%) mixture placed into a closed
vessel containing hexane. After 12 h, one large crystal (30–40
mg) separated from the solution; a fraction of this was used for
X-ray crystallography. More hexane was added into the mother
liquor to afford more product as small white cubes. Yield:
61%, mp 118–119 ЊC. 1H NMR (400 MHz, acetone-d6) δ 8.3 (br
s, 1 H), 6.22 (d, 1 H, 7-H, J6,7 = 7.2 Hz), 4.98 (q, 1 H, 6-H, J6,7
7.2 Hz, J5,6 = 7.2 Hz), 3.44 (q, 4 H, N(CH2CH3)2, JCH3,CH2
7.2 Hz), 3.01 (d, 2 H, 5-H, J5,6 = 7.2 Hz), 1.19 and 1.07 (br, 6
H, N(CH2CH3)2, JCH3,CH2 = 7.2 Hz); all proton signals were
identified from homonuclear decoupling experiments; 13C
=
=
3-Diisopropylamino-5-trifluoromethyl-2H-2,4-diazabicyclo-
[3.2.0]hepta-3,6-diene 27e
A solution of 2-azido-6-trifluoromethylpyridine 7A (100 mg;
0.47 mmol) in 100 ml of dioxan and 5 ml of diisopropylamine
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1237

View Article Online
1996, 118, 4009; (b) C. Wentrup and H.-W. Winter, J. Am. Chem.
was photolysed for 12 h. Evaporation of the solvent gave a
yellow oil, purified by distillation at 45–50 ЊC/0.5 mbar. 1H
NMR (acetone-d6) δ 6.38 (1 H), 6.32 (1 H), 4.48 (1 H), 3.90
(septet, 2 H, iPr), 1.23 (d, 12 H, iPr); 13C NMR (acetone-d6)
Soc., 1980, 102, 6159.
6 F. A. Bovey, Nuclear magnetic Resonance Spectroscopy, Academic
Press, New York, 2nd edn.; 1988; J. Sandström, Dynamic NMR
Spectroscopy, Academic Press, New York, 1982.
δ 164 (C᎐N), 142.3 (C᎐C), 141.0 (C᎐C), 126.1 (q, CF ), 62.0 (m,
᎐
᎐
᎐
3
7 S. Forsén and R. A. Hoffman, J. Chem. Phys., 1963, 39, 2892.
8 A. Reisinger, PhD Thesis, The University of Queensland, 2001.
9 J. C. Teulade, A. Gueiffier, J. P. Chapat, G. Grassy, A. Carpy,
A. H’Naifi, B. Perly and J. Couquelet, Chem. Pharm. Bull., 1989,
37, 2293; P. Molina, A. Arques, A. Alias, M. C. Foces-Foces and
A. L. Llamas-Saiz, J. Chem. Soc., Chem. Commun., 1992, 424.
10 J. A. Moore, W. J. Freeman, R. C. Gearhart and H. B. Yokelson,
J. Org. Chem., 1978, 43, 787.
C1 and C5), 47.5 (CH(CH3)2), 21.5 (CH3); MS m/z 261 (Mϩ,
35%). Anal. Calcd for C12H18N3F3 C, 55.14; H, 6.95; N, 16.09;
found C, 55.37; H, 7.16; N, 16.08.
Acknowledgements
11 C. J. Addicott, PhD Thesis, The University of Queensland, 2002.
12 R. Koch, B. Wiedel and C. Wentrup, J. Chem. Soc., Perkin Trans. 2,
1997, 1851.
13 (a) R. A. Odum and B. Schmall, J. Chem. Res. (S), 1997, 276;
R. A. Odum and B. Schmall, J. Chem. Res. (M), 1997, 1850; (b)
K. Satake, S. Takami, Y. Tawada and M. Kimura, Chem. Commun.,
2001, 1382.
This work was supported by the Australian Research Council.
We thank Ms Julia Stuthe for assistance with some of the
preparations, Dr Lisa George for the Ar matrix IR spectra, Dr
Colin H. L. Kennard and Mr Karl Byriel for assistance with
collecting the X-ray data, and Ms Lynette Lambert for advice
and assistance with the NMR experiments.
14 W. J. Hehre, L. Radom, P. v. R. Schleyer and J. A. Pople, Ab Initio
Molecular Orbital Theory, Wiley, New York, 1986.
15 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery,
R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli,
C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala,
Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz,
B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi,
R. Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe,
P. M. W. Gill, B. G. Johnson, W. Chen, M. W. Wong, J. L. Andres,
M. Head-Gordon, E. S. Replogle and J. A. Pople, Gaussian, Inc.,
Pittsburgh PA, 1998.
16 J. A. Pople, A. P. Scott, M. W. Wong and L. Radom, Isr. J. Chem.,
1993, 33, 345.
17 A. P. Scott and L. Radom, J. Phys. Chem., 1996, 100, 16502.
18 L. J. Farrugia, J. Appl. Crystallogr., 1999, 32, 837.
19 G. M. Sheldrick, SHELX97 - Programs for Crystal Structure
Analysis (Release 97-2), Institut für Anorganische Chemie der
Universität, Tamstrasse 4, D-3400, Göttingen, Germany, 1998.
20 L. J. Farrugia, J. Appl. Crystallogr., 1997, 30, 565.
References
1 (a) A. Reisinger, R. Koch and C. Wentrup, J. Chem. Soc., Perkin
Trans. 1, 1998, 2247; (b) A. Reisinger and C. Wentrup, Chem.
Commun., 1996, 813.
2 A. Reisinger, P. V. Bernhardt and C. Wentrup, Org. Biomol. Chem.,
2004, 2, 222.
3 D. J. Le Count, Comprehensive Heterocyclic Chemistry II,
A. R. Katritzky. C. W. Rees and E. F. V. Scriven, eds, vol 9, p. 139.
Elsevier, New York, 1996.
4 P. S. Zurer, Chem. Eng. News, May 5, 1997; P. Y. S. Lam,
P. K. Jadhav, C. J. Eyermann, C. N. Hodge, Y. Ru, L. C. Bacheler,
J. L. Meek, M. C. Otto, M. R. Rayner, Y. N. Wong, C.-H. Chang,
P. C. Weber, D. A. Jackson, T. R. Sharpe and S. Erickson-Viitanen,
Science, 1994, 263, 380; C. N. Hodge, P. Y. S. Lam, C. J. Eyerman,
P. K. Jadhav, Y. Ru, C. H. Fernandez, G. V. De Lucca, C.-H. Chang,
R. F. Kaltenbach, E. R. Holler, F. Woerner, W. F. Daneker,
G. Emmett, J. C. calabrese and P. E. Aldrich, J. Am. Chem. Soc.,
1998, 120, 4570; F. Qian, J. E. McCusker, Y. Zhang, A. D. Main,
M. Chlebowski, K. Kokka and L. McElwee-White, J. Org. Chem.,
2002, 67, 4086 and references therein.
5 (a) R. A. Evans, M. W. Wong and C. Wentrup, J. Am. Chem. Soc.,
O r g . B i o m o l . C h e m . , 2 0 0 4 , 2, 1 2 2 7 – 1 2 3 8
1238