An NMR study of sequential intermediates and collateral products in the conversion of 1,3,6,8-tetraazatricyclo[4.4.1.13,8]dodecane (TATD) to 1,3,6,8-tetraazatricyclo[4.3.1.13,8]undecane (TATU)

Article abstract of DOI:10.1016/j.tetlet.2008.01.091

In situ ¹H nuclear magnetic resonance spectroscopy was used to investigate the processes that occur during the synthesis of 1,3,6,8-tetraazatricyclo[4.3.1.1^3,8]undecane (TATU). NMR analysis showed a reaction mixture containing more than one compound. The production of these intermediates and collateral products was rationally supported by a careful ¹H NMR monitoring study. We characterized 1,3,5-triazabicyclo[3.2.1]octane (TABO, 4) and 3-(2-aminoethyl)-1,3,5-triazabicyclo[3.2.1]octane (AETABO, 7) by ¹H and ¹³C NMR in D₂O solution inside the NMR sample tube, as an intermediate and collateral product of the reaction, respectively. Further, a reaction of 1,3,6,8-tetraazatricyclo[4.4.1.1^3,8]dodecane (TATD) with ¹⁵N-labeled ammonium chloride was carried out. The ¹⁵N NMR and GC-MS experiments indicated that ¹⁵N was incorporated into TATU, TABO, and urotropine.

Full text of DOI:10.1016/j.tetlet.2008.01.091

Available online at www.sciencedirect.com
Tetrahedron Letters 49 (2008) 2154–2158
An NMR study of sequential intermediates and collateral products
in the conversion of 1,3,6,8-tetraazatricyclo[4.4.1.13,8]dodecane
(TATD) to 1,3,6,8-tetraazatricyclo[4.3.1.13,8]undecane (TATU)
*
´
´
´
Augusto Rivera , Martın E. Nunez, Eliseo Avella, Jaime Rıos-Motta
˜
Departamento de Quı´mica, Universidad Nacional de Colombia, Ciudad Universitaria, Carrera 30 No. 45-03, Bogota´ DC, Colombia
Received 26 June 2007; revised 18 January 2008; accepted 21 January 2008
Available online 30 January 2008
Abstract
1
In situ H nuclear magnetic resonance spectroscopy was used to investigate the processes that occur during the synthesis of 1,3,6,8-
tetraazatricyclo[4.3.1.13,8]undecane (TATU). NMR analysis showed a reaction mixture containing more than one compound. The
production of these intermediates and collateral products was rationally supported by a careful 1H NMR monitoring study. We
characterized 1,3,5-triazabicyclo[3.2.1]octane (TABO, 4) and 3-(2-aminoethyl)-1,3,5-triazabicyclo[3.2.1]octane (AETABO, 7) by 1H
and 13C NMR in D2O solution inside the NMR sample tube, as an intermediate and collateral product of the reaction, respectively.
Further, a reaction of 1,3,6,8-tetraazatricyclo[4.4.1.13,8]dodecane (TATD) with 15N-labeled ammonium chloride was carried out. The
15N NMR and GC–MS experiments indicated that 15N was incorporated into TATU, TABO, and urotropine.
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: 1,3,6,8-Tetraazatricyclo[4.3.1.13,8]undecane; TATU; 1,3,6,8-Tetraazatricyclo[4.4.1.13,8]dodecane; TATD; Kinetic; In situ NMR; Aminal cage
Recently we reported on the one-pot, but multistep,
preparation of the new cyclic aminal 1,3,6,8-tetraazatri-
cyclo[4.3.1.13,8]undecane (TATU, 1) by an unusual ring
contraction-rearrangement of 1,3,6,8-tetraazatricyclo-
[4.4.1.13,8]dodecane (TATD, 2) when reacted with ammo-
nium fluoride in water.1 The present study was aimed at
establishing the structures of compounds produced in the
reaction of aminal (2) with ammonium fluoride in water
and elucidating the possible pathways of their formation.
The reaction can follow several pathways. In particular,
it can be assumed that TATD undergoes recyclization
under the action of NH4F as a result of the ring-chain
transformation through an unstable open-chain way. This
reaction falls under the most general definition of mole-
cular rearrangements as chemical reactions associated with
a change in the bond sequence (molecular skeleton) and
breaking the principle of minimum structural changes.2
NMR experiments have been extensively used by several
authors to probe reactions without disturbing the system
and providing useful structural and kinetic information
about reactions in equilibrium.3 Thus, in our experimental
study of aminal cage reactivity, we decided to revise the
reaction between 2 and ammonium fluoride in D2O to
obtain 1, two aminals amenable to NMR analysis. The
goals of this study were (1) to obtain information about
product composition; (2) to identify the intermediates
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
NH
1
2
3
4
N
N
N
N
N
N
N
*
N
Corresponding author. Tel.: +57
1 3165000x14464; fax: +57 1
N
N
O
3165220.
NH2
5
6
7
E-mail address: ariverau@unal.edu.co (A. Rivera).
0040-4039/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2008.01.091

A. Rivera et al. / Tetrahedron Letters 49 (2008) 2154–2158
2155
involved in the reaction pathway; (3) to gain a better
understanding of the overall reaction mechanism; (4) to
determine some preliminary evidence of the kinetic
reaction.
1
The H NMR spectra of the crude reaction mixtures
(Fig. 1a–c) showed that both the starting aminal TATD
(2) and TATU (1) were present. In addition, a close inspec-
tion also revealed the presence of species with low intensity
signals, which can be attributed to intermediates or colla-
teral products. The downfield signal appearing as a singlet
at d 4.45 was assigned at the aminalic protons of urotro-
pine (3) by comparison with the spectrum of a mixture of
purified samples of 1, 2, and 3 (Fig. 1d), which showed
well-separated signals between 2.25 and 4.50 ppm. This
spectrum had five singlets, two of which were attributable
to ethylenic protons of 1 and 2 (d 3.01 and 3.03, respec-
tively), while others were attributable to aminalic protons
at 3.75 for 2, 4.04 for 1, and 4.45 for 3. The one-proton
doublet at d 3.62 and 4.31 with geminal coupling of
13.5 Hz was attributed to aminalic protons of 1. The split-
ting pattern, coupling constants, and NMR ratios of the
signal protons at d: 4.15 (J = 10.2 Hz), 3.84 (J =
13.1 Hz), 3.36 (J = 12.7), and 3.18 (J = 10.4 Hz)
(Fig. 2a), with an intensity ratio of 1:2:2:1, can indeed be
attributed to one of the secondary products (intermediate
or collateral products). These resonances were clearly
a
Selective TOCSY on 3.84 ppm
Selective COSY on 2.72 ppm
b
c
4.50
4.25
4.00
3.75
3.50
3.25
3.00
2.75
2.50
2.25
ppm (t1)
Fig. 2. TOCSY-selective experiments for 4.
differentiated as two AB-systems with a typical coupling
constant of geminal protons.
Based on the chemical shifts, we suggested the presence
of CH2-aminalic groups in its structure (4). To isolate the
spin systems from the spectrum of reaction mixture, 1D
TOCSY experiments were performed.4 Irradiation of the
signal at 3.84 ppm produced enhancements at 3.36 and
3.18 ppm (Fig. 2b). The appearance of signal at 3.36 ppm
as a doublet of doublets (J = 12.7 and 2.1 Hz) indicated
the presence of long-range NMR coupling constants. The
homonuclear small coupling constant value, which is
4
attributed to a JHH coupling constant between W-posi-
tioned hydrogen atoms in bicyclic compounds,5 suggested
that compound 4 had a bicyclic structure. Aside from these
resonances, the NMR ratio of the signal proton at
2.71 ppm (Fig. 2a) and its typical signal patterns of parti-
cular –CH2CH2– ring moieties indicated that this resonance
can be assigned to an ethylene bridge in a bicyclic system.
In fact, a COSY irradiation6 of this signal resulted in a
signal enhancement at ꢀ2.92 ppm (Fig. 2c) that overlapped
with another major resonance at 2.97 ppm, all of which
vindicated the presence of an ethane-bridge. Again, com-
a
b
c
1
parison of the H NMR spectra of 4 with those of the
known compounds 3-oxa-1,5-diazabicyclo[3.2.1]-octane
(5)5a and 3,30-ethane-1,2-diylbis-1,3,5-triazabicyclo[3.2.1]-
octane (6)5b (Table 1) showed very similar d and J values
suggesting that compound 4 had the bicyclo[3.2.1]octane
nucleus. Based on these data, compound 4 was identified
as 1,3,5-triazabicyclo[3.2.1]-octane (TABO, 4). Besides
urotropine, TATU, and TATD, the GC–MS analysis of
this solution confirms the formation of 4. TABO (4) had
in its EIMS the molecular ion peak at m/z 113 correct
for the formula C5H11N3, and the fragmentation pattern
was consistent with the speculative fragmentation course.
To the best of our knowledge, although its structure was
claimed in a patent,7 its NMR data are unprecedented.
We also focused on the strong peaks at 3.96 and
2.97 ppm, with approximate intensities, 2.61; 7.12
(Fig. 2a). In principle, these signal protons did not show
d
4.25
ppm (t1)
4.00
3.75
3.50
3.25
3.00
2.75
2.50
2.25
Fig. 1. 1H NMR spectra of a reaction mixture of 2 (1.41 M) and NH4F
(4.50 M) in a D2O at (a) t = 0 s, (b) t = 20 min 17 s, and (c) t = 440 min.
(d) 1H NMR spectrum of a mixture of 1 (0.09 M), 2 (0.09 M), and 3
(0.01 M) in D2O.

2156
A. Rivera et al. / Tetrahedron Letters 49 (2008) 2154–2158
Table 1
Comparison of select d and J values, for 4, 7, 5, and 6
Entry
(AB)2
AB
AA0BB0
4a
7a
5b
6c
3.84 (d, J = 12.9); 3.36 (dd, J = 12.9, 2.1)
3.65 (d); 3.51 (d)
4.75 (d, J = 9.6); 4.26 (d, J = 9.6)
3.62 (dd, J = 10.7, 1.5); 3.40
(dd, J = 10.7)
4.15 (d, J = 10.4); 3.18 (d, J = 10.4)
3.97 (d); 3.30 (d)
4.20; 3.30
3.89 (dd, J = 10.5, 1.9); 3.30
(dd, J = 10.5, 1.5)
2.97; 2.71
3.21 (dd); 2.80 (dd)
3.30; 2.38
2.79 (dd, J = 10.6, 4.4); 3.17
(ddd, J = 10.6, 4.4, 1.9)
a
This work.
Ref. 5a.
Ref. 5b.
b
c
proportion of integrals, probably due to that some of the
products gave overlapping signals. In fact, 1D COSY
experiment (Fig. 2c) clearly demonstrates the superposition
of ethylene signals of 4 and 2.97 as discussed below. After
correcting the relative integral of signal at 2.97 ppm for the
experimental amount of 4 to obtain 5.12, we noted that the
magnitude of relative intensities of signals at 3.96 and
2.97 ppm is near 1:2. Thus, based on the chemical shifts
observed and NMR ratios, we proposed the structure of
imidazolidine (8) for this compound. To further support
our peak assignments, we carried out various 2D NMR
spectroscopic techniques, but most of the information for
the confirmation of the peak assignments was obtained
by carrying out HMBC and HMQC experiments. Here
the assignment of the peaks was confirmed by comparing
the rates of consumption of TATD (2) and the correspond-
ing formation of peaks assigned to 4 and 8. Thus, all
assigned peaks showed a good consistency along time
dependences of their intensities. When all peaks of com-
pounds 1, 2, 4, and 8 were assigned, a number of extremely
weak (on the order of 10ꢁ3 relative to those of 1) unas-
signed peaks still remained. Three of these (at 2.28, 2.58,
and 2.79 ppm) were ascribed to ethylenic protons while,
similarly, two weak peaks were ascribed to aminalic pro-
tons (at 3.43 and 3.28 ppm). 1D TOCSY, 1D COSY, and
1D NMR selective irradiation (Fig. 3) allowed us to esta-
blish the presence of 3-(2-aminoethyl)-1,3,5-triazabi-
cyclo[3.2.1]octane (AETABO, 7), whose bicyclic fragment
is structurally similar to 4.
Selective COSY on 2.79 ppm
Selective COSY on 2.58 ppm
Selective TOCSY on 3.08 ppm
Selective TOCSY on 3.43 ppm
4.50
4.25
4.00
3.75
3.50
3.25
3.00
2.75
2.50
2.25
ppm (t1)
Fig. 3. 1H NMR-selective irradiation experiments for 7.
In addition, the NMR spectroscopic data obtained dur-
ing the investigation provide instantaneous concentrations
of detectable species in the reaction mixture. This made it
possible to describe a preliminary study of the reaction
kinetics between TATD and ammonium fluoride to give
TATU. In Figure 4, three different concentration profiles
can be clearly identified: one of these from substrate (2)
which decreases to zero; others of final products (1, 8), were
increased; and finally, one of an intermediate (4), reaching
a maximum concentration after about 20 min.
Initially, a decrease in the concentration of TATD, and
a corresponding increase in the concentration of 4 and
imidazolidine were observed. It is interesting to note that
the consumption of 2 during the first hour is directly pro-
portional to the rate of formation of 4 and 8. Based on
these results, we consider that 2 was consumed to preferen-
Fig. 4. Concentration (mM) versus time (s) for 1, 2, 3, 4, 7, and 8 in the
reaction of TATD with NH4F.

A. Rivera et al. / Tetrahedron Letters 49 (2008) 2154–2158
2157
tially produce 4 and 8. This trend continues until 4 reaches
a maximum concentration, as seen in Figure 4. At this
point, an increase in the concentration of 1 was observed.
Its formation is directly proportional to the rate of
consumption of 4. This observation suggested that not only
does the formation of 1 depend on reaching a sufficient
concentration of 4, but more intriguingly that 4 is the
precursor of the main product 1.
15.0
20.0
25.0
30.0
35.0
Although the pathway for this conversion likely involves
the formation of imine or iminium intermediates, the mech-
anism involving the imine would be less likely because the
ring-chain tautomeric equilibrium in aminals is favored
toward the formation of open species in the presence of tri-
fluoroacetic acid.8
A plausible mechanism involving
ppm (t1)
ammonium intermediates is depicted in Scheme 1. The
essence of this mechanism is an initial nucleophilic attack
by ammonia at the aminal carbon of the protonated aminal
producing an open-ring intermediate (9). This intermediate
could experience a rearrangement process to produce a
highly reactive ammonium intermediate 3-(2-aminoethyl)-
1-(aminomethyl)-3,5-diaza-1-azoniabicyclo-[3.2.1]octane
(10). Once intermediate 10 is formed, it could either be
converted into AETABO (7) (Scheme 1) (by eliminating a
aminomethyl moiety) or alternatively, to suffer an intra-
molecular nucleophilic attack to produce the {[3-(imida-
4.50
4.00
3.50
ppm (t2)
Fig. 5. 1H–15N HMBC contour plot of reaction mixture.
between 2 and NH4F, we believe that the driving force for
its dimerization into 1, under the studied reaction condi-
tions, is provided by two mechanisms: (i) the additional
stability gained in the aminal cage structure of 1 and (ii)
imidazolidine acting as an effective H+ binder. In fact, we
cannot exclude the possibility that the free aminal cage is
a weak base. For instance, it is possible that it may be even
more difficult to insert a proton into the cage, as compared
to removing it.
zolidin-1-ylmethyl)imidazolidin-1-yl]methyl}amine
(11).
This reaction, as seen in Figure 4, occurs faster than amino-
methyl elimination. The decomposition of 11 could afford
the detected compounds 4 and 8, as shown in Scheme 1.
Dimerization of 4 followed by rearrangements and subse-
quent expulsion of imidazolidine would lead to the forma-
tion of TATU (Scheme 1).
Strong support for this proposal is provided by in situ
15N NMR experiments. TATU labeled with 15N was
synthesized from TATD and 15N-enriched ammonium
chloride according to the procedure described above for
the non-labeled material. Upon the addition of 15N-NH4Cl
to 2, we observed three signals at 30.2, 23.2, and 13.6 ppm
at 296 K relative to NH4Cl. We assign the low-field line to
Although proposed intermediate 4 should be the actual
isolated compound derived from the condensation reaction
TATU (1). Two-bond H–15N HMBC correlation of this
1
N
N
signal to aminals protons justified this assignment
(Fig. 5). The other resonances at 23.2 and 13.6 ppm were
assigned to urotropine (3) and TABO (4), respectively.
The GC–MS analysis of this reaction mixture confirms
the formation of TATU-15N2 (M+ 156, 100%). (We are
grateful to a referee and the editor for suggesting these
analyses).
N
N
N
N
NH2
NH2
H
N
N
N
NH2
N
H2N
NH3
2-H+
9
10
N
N
N
N
+
CH2=NH2
NH2
NH2
N
N
H2N
7
In this work, NMR analysis showed a reaction mixture
constituted by more than one aminal product: 4, 7, urotro-
pine, and imidazolidine, which were characterized by H
N
N
HN
NH
H2N
N
N
HN
NH
+
N
N
1
NH2
N
NH2
H2N
10
11
4
8
and 13C NMR in a D2O solution as intermediate or collat-
eral products. Such a conclusion also implies that the
above experiments indicated that the actual subproduct
of this reaction is imidazolidine but not ethylenediamine,
as previously described.1
N
N
N
2
N
N
N
H
HN
N
N
N
HN
N
N
H
N
HN
H
H
12
N
N
N
N
NH
N
HN
NH
HN
N
NH
+
8
Acknowledgments
NH
N
N
NH
N
1-H+
13
´
The authors would like to thank Direccion Nacional de
Scheme 1. Proposed mechanism for forming compound 1.
´
Investigaciones (DINAIN) and Departamento de Quımica,

2158
A. Rivera et al. / Tetrahedron Letters 49 (2008) 2154–2158
3. Yim, C. T.; Zhu, X. X.; Brown, G. R. J. Phys. Chem. B 1999, 103, 597.
4. (a) Davis, D. G.; Bax, A. J. Am. Chem. Soc. 1985, 107, 7197; (b) Bax,
A.; Davis, D. G. J. Magn. Reson. 1985, 65, 355; (c) Stott, K.;
Stonehouse, J.; Keeler, J.; Hwang, T. L.; Shaka, A. J. J. Am. Chem.
Soc. 1995, 117, 4199.
Universidad Nacional de Colombia for their financial
support.
Supplementary data
5. (a) Hocker, J.; Wendisch, D. J. Chem. Res. (S) 1977, 236; (b) Rivera,
´
A.; Rıos-Motta, J. Tetrahedron Lett. 2005, 46, 5001; (c) Constantino,
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.tetlet.
2008.01.091.
M. G.; Lacerda, V., Jr.; da Silva, G. V. J.; Tasic, L.; Rittner, R. J. Mol.
Struct. 2001, 597, 129; (d) Ellinger, Y.; Subra, R.; Berthier, G. J. Am.
Chem. Soc. 1978, 100, 4961.
6. (a) Bauer, C. J.; Freeman, R.; Frenkiel, T.; Keeler, J.; Shaka, A. J. J.
Magn. Reson. 1984, 58, 442; (b) Stonehouse, J.; Adell, P.; Keeler, J.;
Shaka, A. J. J. Am. Chem. Soc. 1994, 116, 6037.
References and notes
7. Massie, S. N., U.S. Patent 3,914,307, 1975.
1. Rivera, A.; Nu´nez, M. E.; Morales-Rios, M. S.; Joseph-Nathan, P.
˜
8. Eldin, S.; Digits, J. A.; Huang, S.-T.; Jencks, W. P. J. Am. Chem. Soc.
1995, 117, 6631.
Tetrahedron Lett. 2004, 45, 7563.
2. Muller, P. Pure Appl. Chem. 1994, 66, 1077.
¨