Acid-catalysed rearrangements between 3,4-dihydro-2H-1,5-benzooxazocine and -benzothiazocine and their macrocyclic (16-membered, 24-membered and 32-membered) oligomers

Article abstract of DOI:10.1039/a704255h

Treatment of 2-(3-azidopropoxy)benzaldehyde with triphenylphosphine in diethyl ether, followed by sequestration of the products with nickel(II) thiocyanate and subsequent liberation from the complex with aqueous ammonia, produces the 16-membered dimer of

Full text of DOI:10.1039/a704255h

View Article Online / Journal Homepage / Table of Contents for this issue
Acid-catalysed rearrangements between 3,4-dihydro-2H-1,5-
benzooxazocine and -benzothiazocine and their macrocyclic
(16-membered, 24-membered and 32-membered) oligomers
David C. R. Hockless,a Leonard F. Lindoy,b Gerhard F. Swiegers*,†,a and
S. Bruce Wilda
a Research School of Chemistry, Australian National University, Canberra,
Australian Capital Territory 0200, Australia
b School of Chemistry, University of Sydney, Sydney, NSW 2006, Australia
Treatment of 2-(3-azidopropoxy)benzaldehyde with triphenylphosphine in diethyl ether, followed by
sequestration of the products with nickel(II) thiocyanate and subsequent liberation from the complex
with aqueous ammonia, produces the 16-membered dimer of 3,4-dihydro-2H-1,5-benzooxazocine.
The dioxadiaza diimine macrocycle 3 undergoes an acid-catalysed rearrangement in solution into
an equilibrium mixture that includes the trioxatriaza triimine (24-membered) 8 and tetraoxatetraaza
tetraimine (32-membered) 9 macrocycles, as well as the diimine. The oligomers in [2H]chloroform
differ in the 1H NMR chemical shifts of the protons on the central carbon atoms of the propane-1,3-diyl
links, thereby allowing an investigation of the factors involved in the rearrangements between the
macrocyclic oligomers. Thus, the rearrangements of the macrocyclic polyimines into equilibrium mixtures
of the various oligomers occurs with t 7–10 min, but the position of the equilibrium in each case is
¹
¯
²
dependent upon the imine equivalent concentration of the solution. 3,4-Dihydro-2H-1,5-benzothiazocine
undergoes a similar rearrangement. The equilibrations have been ascribed to facile intermolecular
transiminations between the oligomers in the presence of traces of acid and to the similarities of the
thermodynamic stabilities of the 16-, 24- and 32-membered oligomers. The crystal and molecular
structures of ammonium nitrate salts of the 16-membered diamine 11 and 32-membered tetraamine 10
derivatives of the corresponding macrocyclic benzooxazocines are reported.
Introduction
The importance of metal-ion size in the template synthesis of
large macrocycles by condensation reactions is well known; an
NH
N
N
O
N
N
O
alteration in the metal employed in the cyclisation can produce
an oligomeric macrocycle of different ring size by a kinetic or
thermodynamic template effect.1 If the macrocycles exist in a
metal-independent equilibrium with higher cyclic oligomers
however, the thermodynamic stabilities and solubility proper-
ties of the various oligomers may play the deciding role.
O
N
HN
2
n
1a (n = 1)
1b (n = 1)
1c (n = 3)
In our recent syntheses of imine macrocycles based on
3
1,2,3,4-tetrahydro-1,5-benzodiazocine,
the
32-membered
octaaza tetraimine 1c was found to exist in a metal-independent
acid-catalysed equilibrium with the 16-membered tetraaza
diimine 1a and 24-membered hexaaza triimine 1b.2 The equi-
librium was dominated by the poor solubility of 1c, which was
the product obtained from mixtures of the oligomers in solu-
tions containing traces of acid. The more soluble lower poly-
imines were isolated by selective sequestration or fractional
crystallisation from equilibrium mixtures quenched with a non-
protic solvent.
X
N
N
X
S
N
S
N
Acid-catalysed, metal-independent equilibria involving
macrocyclic imines have been documented for the parent3 and
several substituted dihydrobenzo-azepines and -azocines,2,4–10
and for certain other macrocyclic polyimines such as TAAB
and TRI.11 The equilibria, however, have not been observed
previously for ring sizes larger than 16 members. For example,
the closely related substituted benzoazepine dimers 4 (X = CH2,
O, S, AsMe)3–8 rearrange into equilibrium monomer (7-
4 (X = CH2, NH, O, S, AsMe)
5
N
N
N
N
N
N
N
membered)
dimer (14-membered) mixtures under slightly
acidic conditions. In each case the monomer is strongly
† Present address: Chemistry Department, University of Wollongong,
Northfields Avenue, Wollongong, NSW 2522, Australia.
TRI
TAAB
J. Chem. Soc., Perkin Trans. 1, 1998
117

View Article Online
favoured in solution, but removal of solvent affords the dimer
as a crystalline solid. For 4 (X = NH)9 a similar rearrangement
occurs in which the dimer is favoured; this was ascribed to the
increased basicity of the amino groups present.10 In contrast,
the rearrangements between the benzodiazocine oligomers 1a–c
do not appear to involve the 8-membered monomer, but only
the higher oligomers.
N
N
O
O
N
O
O
N
N
N
O
O
O
We have now prepared the 16-membered dioxadiaza diimine
dimer 3 of 3,4-dihydro-2H-1,5-benzooxazocine 2, as well as the
16-membered dithiadiaza diimine dimer of 3,4-dihydro-2H-
1,5-benzothiazocine, 5,6 and investigated the acid-catalysed
rearrangements in each case. The results are reported below;
they indicate that acid-catalysed rearrangements involving 16-,
24- and 32-membered macrocycles of this type are common to
3,4-dihydro-2H-1,5-benzoheteroazocines containing NH-, O-
and S-heteroatoms and originate in self-assembly processes in
which rapid kinetics permit the formation of the thermo-
dynamically most favoured products in each case.
N
8
9
NH
O
O
O
HN
O
HN
O
NH
NH
O
NH
Results and discussion
Benzoazocine oligomers
10
11
The 14-membered diimine 6,7,15,16-tetrahydrodibenzo[ f,m]-
[1,8,4,11]dioxadiazacyclotetradecine 4 (X = O) had been pre-
pared previously in our laboratory from 2-(2-azidoethoxy)-
benzaldehyde by the aza-Wittig reaction.8 The same reaction
was employed for the preparation of the 16-membered dioxa-
O
HN
HN
NH
O
diaza
diimine
7,8,17,18-tetrahydro-6H,16H-dibenzo[b,j]-
[1,9,5,13]dioxadiazacyclohexadecine 3 (Scheme 1). Because of
O
CHO
OH
CHO
(i)
O
Br
12
6
resonance at δ 2.11. The mass spectrum of the compound dis-
played a peak at m/z 644.0, which corresponded to the
tetraimine 9. The mass spectrum of 3 under similar conditions
contained a molecular ion peak at m/z 321.1. Reduction of 9
with lithium aluminium hydride led to the corresponding, kinet-
ically inert, tetraoxatetraamine 10. Treatment of a methanolic
solution of 10 with nitric acid led to the isolation of colourless
crystals of 10ؒ4HNO3ؒCH3OH that were suitable for a single
crystal X-ray structural determination. The molecular structure
(ii)
CHO
(iii)–(vi)
3
O
N3
7
Scheme 1 Reagents and conditions: (i) BrCH2CH2CH2Br, (ii) NaN3,
(iii) Ph3P, Et2O, (iv) CHCl3, heat, 24 h, (v) NiCl2, KSCN(aq.), (vi)
NH4OH
of the cation in 10ؒ4HNO3
ؒCH OH is shown in Fig. 1; within
3
the unit cell each 32-membered tetraammonium macrocycle is
associated with six nitrate ions, four of which are shared with
cations in neighbouring cells.
The dioxadiaza diamine 11 was prepared by reduction of 3
under similar conditions. Although the H NMR spectrum of
11 is very similar to that of 10, the fragmentation pattern of 11
in the mass spectrum is significantly different; the spectrum
contains a peak for the molecular ion at m/z 326.2, whereas 10
exhibits a peak at m/z 652.2. An X-ray crystal structure
determination was performed on a crystal of 11ؒ2HNO3. The
molecular structure of one of two similar but independent cat-
ions in 11ؒ2HNO3 is shown in Fig. 2. Each cation lies on a
centre of inversion. The crystal contains discrete assemblies
involving four of the 16-membered diammonium macrocycles
juxtaposed about eight bridging nitrate ions.
the triphenylphosphine oxide by-product in the reaction
mixture and the facile rearrangement of 3 in solution, the
diimine was isolated by selective sequestration with nickel()
thiocyanate and subsequent liberation from the metal complex
with aqueous ammonia.6 The use of aqueous ammonia was
necessary to bring about precipitation of 3 after release of the
metal, thereby eliminating subsequent rearrangements in
solution.
1
In our investigation of the acid-catalysed rearrangement of
3, a concentrated solution of the macrocycle in [2H]chloroform
1
was monitored by H NMR spectroscopy over several hours;
the resonances characteristic of the diimine were gradually
replaced by two sets of new resonances. The most diagnostic of
these resonances were those arising from the protons on the
central carbon atom of the propane-1,3-diyl link, in which min-
imal overlap occurred. Thus, the multiplet at δ 2.35, associated
with 3, was replaced by new multiplets at δ 2.11 and 2.25. As
observed previously for other benzodiazocine rearrangements,2
the new resonances suggested the presence of the triimine 8 and
tetraimine 9. The slow crystallisation of a concentrated chloro-
form solution of 3 afforded colourless crystals that exhibited
Crystal data, information relating to data collection and
refinement details for 11ؒ2HNO3 and 10ؒ4HNO3ؒCH3OH are
given in the Experimental section.‡
‡ Atomic coordinates, thermal parameters and bond lengths and angles
have been deposited with the Cambridge Crystallographic Data Centre
(CCDC). Any requests for this material should be accompanied by a
full literature citation together with the reference number 207/157.
For details of the scheme, see Instructions for Authors, J. Chem. Soc.,
Perkin Trans. 1, available via the RSC Web pages (http//www.rsc.org/
authors).
1
new H NMR resonances in [2H]chloroform characteristic of
one of the rearrangement products, including the diagnostic
118
J. Chem. Soc., Perkin Trans. 1, 1998

View Article Online
Fig. 1 ORTEP view of the cation in 10ؒ4HNO3ؒCH3OH showing
atom-labelling scheme for non-hydrogen atoms. Thermal ellipsoids
enclose 50% probability levels. The cation occupies a centre of inversion
in the crystal.
Fig. 3 Relative equilibrium proportions of oligomers of 2 in CDCl3
solution as a function of imine equivalent concentration. The percent-
ages were measured using the relative integrals of diagnostic 1H NMR
resonances at δ 2.11 (3), 2.25 (8) and 2.35 (9) after equilibration.
reduction of a concentrated, week-old, chloroform solution of
5, followed by an ion chromatogram mass spectrum of the oil
after removal of solvent, nevertheless confirmed the presence of
the dithiadiaza diamine (m/z 358.1), trithiatriaza triamine (m/z
537.2) and tetrathiatetraaza tetraamine (m/z 717.0) oligomers
(mixture 13). Pure 13 (n = 1) (the diamine) exhibited the peak at
Fig. 2 ORTEP view of one of two similar, but independent cations in
the unit cell of 11ؒ2HNO3. Each cation lies on a centre of inversion.
The atom-labelling scheme for non-hydrogen atoms in this particular
cation is shown. Thermal ellipsoids enclose 50% probability levels.
Having established the identities of two of the three species
involved in the equilibrium, viz. 3 and 9, an attempt was made
to isolate the third component, which exhibited a diagnostic 1H
NMR resonance at δ 2.25. The attempted fractional crystallis-
ation from dichloromethane of a sample enriched in the
unknown species was unsuccessful due to rearrangements.
Thus, a concentrated, week-old solution of 3 was reduced with
sodium cyanoborohydride in acetic acid–methanol. This led to
the isolation of a mixture of the kinetically stable oxaaza
polyamines 10, 11 and 12 that were separated as their hydro-
chlorides on cation exchange resin by elution with aqueous
methanolic hydrochloric acid of increasing acidity (1–4 ). The
second band to be eluted from the column was identified as the
triamine hydrochloride 12ؒ3HCl by its mass spectrum, which
contained a peak at m/z 489.4 corresponding to the triamine.
The first and third bands recovered from the column gave crys-
talline solids that had properties identical to those of the separ-
ately prepared 11ؒ2HCl and 10ؒ4HCl, respectively. Treatment
of 12ؒ3HCl with aqueous sodium hydroxide produced free 12 as
a colourless oil. In view of the isolation of 12, the third species
in the equilibrium mixture containing 3 and 9 was identified as
the trioxatriaza triimine 8.
S
HN
S
NH
n
13 (n = 1,2,3)
m/z 358.1 only. Because of poor solubility in water and aqueous
alcohols, the mixture of polyamines 13 could not be separated
by ion exchange chromatography of the corresponding hydro-
chlorides.
Acid-catalysed rearrangements between 3, 8 and 9
₁
The t values for rearrangements of 3 and 9 into equilibrium
₂
mixtures of 3, 8 and 9 in [2H]chloroform, as well as the posi-
tions of the resulting equilibria, were determined as functions
of concentration by H NMR spectroscopy. Imines 3 and 9
rapidly rearranged into equilibrium mixtures in solution. The
spectra were recorded at 2–3 min intervals for the first 30 min
after each solution was prepared, and then at increasing inter-
1
Benzothiazocine oligomers
The dithiadiaza diimine 5 was prepared as described previ-
ously.6 The diimine rearranges over several days in [2H]chloro-
form, as indicated by the appearance of two new sets of reson-
ances in the propane-1,3-diyl central methylene region and a
₁
vals over several hours. By this means the t for 9 was found to
₂
be ca. 10 min, whereas for 3 it was ca. 7 min. Concentrations of
0.2  imine equivalents (mol imine bonds per litre [2H]chloro-
form) were used in both cases to normalise the comparisons.
The position of the equilibrium was found to be markedly
dependent on the concentrations of the solutions employed
(Fig. 3). The compounds equilibrated within several hours, but
solutions were kept for ca. 15 h to ensure complete rearrange-
ment. Fig. 3 shows the proportions of each oligomer (measured
1
single new azamethine peak in the H NMR spectrum. Unlike
the solution for 3 however, significant overlap of the resonances
in the spectrum meant that a detailed analysis was not possible.
Moreover, the poor stability of 5 in chloroform precluded the
isolation of the higher oligomers by fractional crystallisation
or selective sequestration with nickel(). The cyanoborohydride
J. Chem. Soc., Perkin Trans. 1, 1998
119

View Article Online
using the integrals of the diagnostic central propane-1,3-diyl
bridge H NMR multiplets) as a function of the imine equiv-
X-Ray crystallographic analysis of 11ؒ2HNO3
1
Crystal data: C20H28N4O8 (M = 452.46), colourless plate, tri-
clinic, crystal dimensions 0.20 × 0.10 × 0.04 mm, Dc 1.378 g
alent concentration. The diimine 3 was thermodynamically
favoured in dilute solutions, but the tetraimine 9 became more
favoured as the concentration of the solution was increased. At
0.006 , the ratio of the integrals indicated 3:8:9 = 72:23:5.
The percentage due to 3 fell whereas that of 9 rose as the con-
centration of the solution was increased over the range exam-
ined. The proportion of 8 increased to a maximum of 37% at
0.090 , and then decreased slowly as the concentration was
increased to 0.200 . At 0.200  the ratio of 3:8:9 determined
in this way was 44:31:25. Thus, of every 100 imine-containing
molecules in a 0.200  [2H]chloroform solution, 16 were
tetraimine molecules (9), 26 were triimine molecules (8) and 58
were diimine molecules (3). Imine equivalent concentrations
greater than 0.200  could not be investigated because of
broadening and overlap of the diagnostic 1H NMR resonances
employed.
cmϪ3, space group P1 (no. 2), a = 10.082(2) Å, b = 12.007(2) Å,
¯
c = 9.782(1) Å, α = 105.89(1)Њ, β = 106.34(1)Њ, γ = 82.14(1)Њ,
V = 1090.7(3) Å3, Z = 2, F000 = 4480.00. All measurements were
performed on a Rigaku AFC6R diffractometer with graphite
monochromated Cu-Kα radiation (λ = 1.541 78 Å), µ = 9.13
cmϪ1. The data were collected at 23.0 ЊC using the ω–2θ scan
technique to a maximum 2θ value of 120.1Њ. Of the 3445 reflec-
tions collected, 3233 were unique (Rint = 0.054). The data were
corrected for Lorentz and polarisation effects. The structure
was solved by direct methods12 and expanded using Fourier
techniques.13 The non-hydrogen atoms were refined aniso-
tropically. The hydrogen atom coordinates were refined but
their isotropic temperature factors were held fixed with a
common value. The final cycle of full-matrix least-squares
refinements was based on 1607 observed reflections [I > 3.0σ(I)]
and 373 variable parameters. The final R factors were R =
¹
2
Conclusion
²
Σ Fo| Ϫ |Fc /Σ|Fo| = 0.057, Rw = [Σw(|Fo| Ϫ |Fc|) /ΣwFo] = 0.052.
The standard deviation of an observation of unit weight was
2.54. The maximum and minimum peaks on the final differ-
ence Fourier map corresponded to 0.47 and Ϫ0.36 e ÅϪ3
respectively.
The origin of equilibria involving 16-membered diimine, 24-
membered triimine and 32-membered tetraimine oligomers of
3,4-dihydro-2H-1,5-benzooxazocines and -benzothiazocines,
and 1,2,3,4-tetrahydro-1,5-benzodiazocines can be ascribed to
facile acid-catalysed rearrangements between the macrocyclic
imines that allow the establishment of thermodynamic mixtures
of the respective oligomers. The diimine, triimine and tetraimine
oligomers possess similar thermodynamic stabilities so that all
three species are present under weakly acidic conditions. In the
case of the benzooxazocines 3, 8 and 9, entropic factors deter-
mine the relative proportions of each oligomer, with the equi-
librium position shifting in favour of the larger molecules as the
concentration of the solution is increased. The difference in
behaviour between the substituted benzoazepine dimers 4
(X = O, S, NH), in which the 7-membered monomer is always
present, and the benzoazocine oligomers 1a–c, 8 and 9, and
oligomers of 5, in which the 8-membered monomer is not ob-
served, suggests that rearrangements of the latter are governed
by ring strain. The monomeric 7-membered benzoazepines are
thermodynamically more stable than the 14-membered dimers,
whereas the 8-membered benzooxazocines and benzothiazoc-
ines are significantly less stable than the corresponding 16-, 24-
and 32-membered oligomers. These systems therefore represent
a self-assembly process in which complex structures are spon-
taneously assembled from smaller building blocks; in each case
the kinetic instability of the imine bonds in weakly acidic solu-
tions permits the formation of the thermodynamically favoured
higher oligomers. A mechanism for such rearrangements has
been proposed previously.2 The ease with which some of the
oligomers can be isolated also suggests that the rearrangement
is useful for the preparation of hitherto inaccessible 24- and
32-membered macrocyclic polyamines. Future work will be
directed towards the isolation of the higher homochiral oligo-
mers of 1,2,3,4-tetrahydro-1-methyl-5,1-benzoazarsocine, the
8-membered homologue of the 7-membered benzoazarsepines
previously prepared in our laboratory.7
X-Ray crystallographic analysis of 10ؒ4HNO3ؒCH3OH
Crystal data: C41H60N8O17 (M = 936.97), colourless block, tri-
clinic (crystallised from methanol), crystal dimensions
0.18 × 0.08 × 0.16 mm, Dc 1.379 g cmϪ3, space group P1 (no.
¯
2), a = 8.3422(8) Å, b = 10.533(1) Å, c = 13.979(1) Å, α =
94.871(9)Њ, β = 100.733(7)Њ, γ = 108.814(8)Њ, V = 1128.4(2) Å3,
Z = 1, F000 = 498.00. All measurements were performed on a
Rigaku AFC6R diffractometer with graphite monochromated
Cu-Kα radiation (λ = 1.541 78 Å), µ = 9.13 cmϪ1. The data were
collected at 23.0 ЊC using the ω–2θ scan technique to a max-
imum 2θ value of 120.1Њ. Of the 3581 reflections collected, 3370
were unique (Rint = 0.019). The data were corrected for Lorentz
and polarisation effects. A correction for secondary extinction
was applied [coefficient = 1.4(3) × 10Ϫ6]. The structure was
solved by direct methods12 and expanded using Fourier tech-
niques.13 The non-hydrogen atoms were refined anisotropically
except for the minor component of a disordered nitrate ion. A
methanol molecule was included with half-occupancy over a
centre of symmetry. Hydrogen atoms were refined isotropically.
Methanol hydrogen atoms could not be located. The final cycle
of full-matrix least-squares refinements was based on 2414
observed reflections [I > 3.0σ(I)] and 426 variable parameters.
The final R factors were R = Σ Fo| Ϫ |Fc /Σ|Fo| = 0.048, Rw =
¹
2
²
2
[Σw(|Fo| Ϫ |Fc|) /Σw|Fo | ] = 0.044. The standard deviation of an
observation of unit weight was 3.12. The maximum and min-
imum peaks on the final difference Fourier map corresponded
to 0.33 and Ϫ0.25 e ÅϪ3 respectively.
In the crystallographic structure determinations of
10ؒ4HNO3ؒCH3OH and 11ؒ2HNO3 described above, neutral
atom scattering factors were taken from Cromer and Waber.14
Anomalous dispersion effects were included in Fc;15 the values
for ∆fЈ and ∆fЉ were those of Creagh and McAuley.16 The
values for mass attenuation coefficients are those of Creagh and
Hubbell.17 All calculations were performed using the TEXSAN
crystallographic software package from the Molecular Struc-
ture Corporation.18
Experimental
1H and 13C NMR spectra were recorded at 20 ЊC on a Varian
Gemini 300 spectrometer operating at 299.96 MHz. Chemical
shifts are quoted as δ values in [2H]chloroform relative to
internal Me4Si unless otherwise stated; coupling constants J are
given in Hz. Elemental analyses and molecular weight meas-
urements were performed by staff within the Research School
of Chemistry. Melting points were determined on a Reichert
melting point apparatus and are uncorrected. All reactions were
performed under an inert atmosphere. The benzothiazocine
dimer 5 and its diamine 13 (n = 1) were prepared as previously
described.6
2-(3-Bromopropoxy)benzaldehyde 6
This compound was prepared in 50% yield by a similar pro-
cedure to that described for the ethylene analogue8 and was
isolated as a pale yellow oil, bp 122–124 ЊC (0.1 mmHg)
(Found: C, 50.0; H, 4.6. C10H11BrO2 requires C, 49.4; H, 4.6%);
δH(299.95 MHz; CDCl3; Me4Si) 2.38 (2 H, t, J 7, CH2CH2CH2),
3.63 (2 H, t, J 7, OCH2), 4.21 (2 H, t, J 7, CH2Br), 7.02 (2 H, m,
ArH), 7.54 (1 H, m, ArH), 7.62 (1 H, d, J 7, ArH) and 10.48 (1
120
J. Chem. Soc., Perkin Trans. 1, 1998

View Article Online
H, s, CHO); δC(299.95 MHz; CDCl3) 29.82 (CH2CH2CH2),
32.01 (OCH2), 65.85 (CH2Br), 112.54 (ArC), 120.92 (ArC),
128.40 (ArC), 136.05 (ArC), 160.91 (ArC) and 189.38 (CHO);
chloroform–dichloromethane (6 ml) and the solution was
allowed to stand overnight in a sealed vessel. Upon the addition
of diethyl ether (14 ml), 9 separated as colourless crystals. The
product was isolated, washed with a small quantity of cold
diethyl ether, and dried (60 ЊC, 0.01 mmHg) for 1 h. The stoi-
chiometry of the fractional hydrate obtained did not alter with
further drying. Yield: 0.20 g (0.31 mmol), mp 215–220 ЊC
(Found: C, 73.9; H, 6.8; N, 8.5. C40H44N4O4ؒ0.3H2O requires C,
73.9; H, 6.9; N, 8.6%); δH(299.95 MHz; CDCl3; Me4Si) 1.75 (0.6
H, br m, H2O), 2.11 (8 H, m, CH2CH2CH2), 3.80 (8 H, t, J 7,
OCH2 or NCH2CH2), 4.06 (8 H, t, J 7, OCH2 or NCH2CH2),
6.86–6.95 (8 H, m, ArH), 7.33 (4 H, t, J 7, ArH), 7.91 (4 H, d, J
7, ArH) and 8.81 (4 H, s, ArCHN); δC(299.95 MHz; CDCl3)
30.42 (CH2CH2CH2), 58.44 (OCH2 or NCH2CH2), 65.09
(OCH2 or NCH2CH2), 111.84 (ArC), 120.65 (ArC), 127.09
m/z (EI) 244.0 (M ϩ).
᝽
2-(3-Azidopropoxy)benzaldehyde 7
This compound was prepared by the method used for the ethyl-
ene analogue.8 The crude product was filtered through a short
silica gel column with dichloromethane as eluent and the eluate
was evaporated to leave a pale brown oil. Dissolution of the
oil in diethyl ether and cooling of the solution in a dry ice–
methanol bath gave colourless crystals of the pure products
that were separated and stored at Ϫ20 ЊC. Yield: 89% of a col-
ourless oil (at room temperature) (Found: C, 58.5; H, 5.5; N,
20.8. C10H11N3O2 requires C, 58.5; H, 5.4; N, 20.5%); δH(299.95
MHz; CDCl3; Me4Si) 2.14 (2 H, t, J 7, CH2CH2CH2), 3.57 (2 H,
t, J 7, OCH2), 4.19 (2 H, t, J 7, CH2N3), 7.01 (2 H, m, ArH),
7.56 (1 H, t, J 7, ArH), 7.83 (1 H, d, J 7, ArH) and 10.50 (1 H,
s, CHO); δC(299.95 MHz; CDCl3) 28.68 (CH2CH2CH2),
48.13 (OCH2), 65.09 (CH2Br), 112.49 (ArC), 121.01 (ArC),
128.61 (ArC), 135.08 (ArC), 160.96 (ArC) and 189.52 (CHO);
(ArC), 131.96 (ArC) and 157.94 (ArC); m/z (EI) 644.0 (M ϩ); Mr
᝽
(osmometry in CH2Cl2) Calc. 644.8, Found 650.2.
7,8,9,10,17,18,19,20-Octahydro-6H,16H-dibenzo[b,j][1,9,5,13]-
dioxadiazacyclohexadecine dihydrochloride hydrate, 11ؒ2HClؒ
H2O
m/z (EI) 121.0 (M ϩ).
Lithium aluminium hydride (1.06 g, 28 mmol) was suspended in
anhydrous tetrahydrofuran (300 ml). The mixture was cooled
(0 ЊC) and treated with 3 (3.00 g, 9.0 mmol) in small portions
over 1 h, after which it was warmed to room temperature and
stirred for 2 h. The following reagents were added to the reac-
tion mixture sequentially at 0 ЊC: water (1.06 ml), 4  NaOH
(1.06 ml), water (3.18 ml). The mixture was then filtered and the
precipitate of aluminium oxides was washed with tetrahydro-
furan (50 ml). The filtrate was evaporated in vacuo to dryness; a
pale yellow oil remained that could not be crystallised or dis-
tilled without decomposition. Dissolution of the oil in meth-
anol and the dropwise addition of concentrated hydrochloric
acid led to the precipitation of the dihydrochloride as a micro-
crystalline solid. The product was isolated, washed with cold
methanol and dichloromethane, and dried for 1 h (60 ЊC, 0.01
mmHg). Yield: 3.01 g (7.2 mmol) (80%), mp 266–270 ЊC
(Found: C, 57.4; H, 7.5; N, 6.5. C20H26N2O2ؒ2HClؒH2O requires
C, 57.6; 7.2; 6.7%); δH(299.95 MHz; 2:1 CD3OD–D2O; Me4Si)
2.42 (4 H, br m, CH2CH2CH2), 3.45 (4 H, m, OCH2 or
NCH2CH2), 4.30 (4 H, m, OCH2 or NCH2CH2), 4.52 (4 H, s,
Ar-CH2-N), 7.21 (4 H, m, ArH) and 7.61 (4 H, m, ArH);
δC(299.95 MHz; 2:1 CD3OD–D2O) 26.74 (CH2CH2CH2), 45.81
(OCH2 or NCH2CH2), 48.11 (OCH2 or NCH2CH2), 65.67
(ArCH2N), 112.40 (ArC), 122.36 (ArC), 132.70 (ArC), 133.25
᝽
7,8,17,18-Tetrahydro-6H,16H-dibenzo[b,j][1,9,5,13]dioxadi-
azacyclohexadecine 0.6 hydrate, 3ؒ0.6H2O
The azide 7 (3.14 g, 15.3 mmol) was dissolved in diethyl ether
(30 ml) and triphenylphosphine (4.01 g, 15.3 mmol) was added.
The mixture was stirred for 3 h until nitrogen evolution had
ceased. The phosphine oxide was filtered off and washed with
diethyl ether. The filtrate and washings were evaporated to leave
a yellow oil that solidified over 15 h. The crude product was
dissolved in chloroform (250 ml) and the solution was heated
under reflux for 24 h. An aqueous methanol solution (20 ml,
50%) of NiCl2ؒ6H2O (1.95 g) and KSCN (2.00 g) was added
to the chloroform solution and heating was continued for 1 h.
The organic layer was separated and the pale purple complex,
[Ni(SCN)2(3)] (2.30 g, 66%), was filtered off, washed with
chloroform and diethyl ether, and dried. The complex was sus-
pended in aqueous ammonia (350 ml, ca. 20%) and the mixture
was stirred for 4 days in a sealed vessel. During this period the
free diimine separated as a colourless microcrystalline solid. It
was filtered off, washed with a small portion of cold aqueous
ammonia followed by water, and dried for 4 h (60 ЊC, 0.01
mmHg). The stoichiometry of the fractional hydrate obtained
did not alter with further drying. Yield: 1.88 g {87% based on
[Ni(SCN)2(3)]}; mp 155–184 ЊC (Found: C, 72.1; H, 7.0; N,
8.4. C20H22N2O2ؒ0.6H2O requires C, 72.1; H, 7.0; N, 8.4%);
δH(299.95 MHz; CDCl3; Me4Si) 1.80 (1.2 H, br m, H2O), 2.35
(4 H, m, CH2CH2CH2), 3.97 (4 H, m, OCH2 or NCH2CH2),
4.02 (4 H, m, OCH2 or NCH2CH2), 6.98 (4 H, m, ArH), 7.32
(2 H, m, ArH), 7.91 (2 H, d, J 7, ArH) and 8.98 (2 H, s,
NCH); δC(299.95 MHz; CDCl3) 29.14 (CH2CH2CH2), 58.21
(OCH2 or NCH2CH2), 65.56 (OCH2 or NCH2CH2), 114.73
(ArC), 121.56 (ArC), 126.78 (ArC), 131.95 (ArC) and 158.47
(ArC) and 158.15 (ArC); m/z (EI) 326.2 (11 ϩ). The free diamine
᝽
was liberated from the salt by treatment with saturated aqueous
sodium hydroxide. The colourless oil that separated had the
following spectroscopic properties: δH(299.95 MHz; CDCl3;
Me4Si) 1.85 (0.2 H, br m, H2O), 2.03 (4 H, m, CH2CH2CH2),
2.87 (4 H, m, OCH2 or NCH2CH2), 3.86 (4 H, s, NCH2Ar),
3.98 (4 H, m, OCH2 or NCH2CH2), 6.83 (2 H, d, J 7, ArH), 6.91
(2 H, t, J 7, ArH) and 7.26 (4 H, m, ArH); δC(299.95 MHz;
CDCl3) 29.94 (CH2CH2CH2), 46.87 (OCH2 or NCH2CH2),
49.97 (OCH2 or NCH2CH2), 66.00 (ArCH2N), 110.80 (ArC),
120.46 (ArC), 128.58 (ArC), 130.92 (ArC) and 157.13 (ArC);
(ArC); m/z (EI) 321.1 (M ϩ).
᝽
7,8,17,18,27,28-Hexahydro-6H,16H,26H-tribenzo[b,j,r]-
[1,9,17,5,13,21]trioxatriazacyclotetracosine 8
m/z (EI) 326.2 (M ϩ).
᝽
A solution of 3ؒ0.6H2O (0.05 g, 0.15 mmol) in [2H]chloroform
(0.6 ml) was allowed to stand overnight whereupon the 1H
NMR spectrum of the solution included resonances consistent
with 8; δH(299.95 MHz; CDCl3; Me4Si) 2.25 (6 H, m, J 7,
CH2CH2CH2), 3.85 (6 H, t, J 7, OCH2 or NCH2CH2), 4.19 (6 H,
t, J 7, OCH2 or NCH2CH2), 6.92 (6 H, m, ArH), 7.32 (3 H,
m, ArH), 7.95 (3 H, d, J 7, ArH) and 8.82 (3 H, s, ArCHN).
7,8,9,10,17,18,19,20,27,28,29,30-Dodecahydro-6H,16H,26H-
tribenzo[b,j,r][1,9,17,5,13,21]trioxatriazacyclotetracosine
trihydrochloride sesquihydrate, 12ؒ3HClؒ1.5H2O
Crystalline 3 (4.00 g, 12 mmol) was dissolved in chloroform
(40 ml) and the solution was set aside for one week. Sodium
cyanoborohydride (3.43 g, 55.6 mmol) was then added to the
solution with stirring, followed by the immediate addition of
1:1 methanol–acetic acid (250 ml). The mixture was stirred for
1 h and then cooled to 0 ЊC. Saturated aqueous sodium hydrox-
ide (250 ml) was slowly added and the mixture was extracted
with dichloromethane (3 × 150 ml). The organic layer was sep-
arated, dried over MgSO4, and evaporated to dryness, leaving a
7,8,17,18,27,28,37,38-Octahydro-6H,16H,26H,36H-tetrabenzo-
[b,j,r,z][1,9,17,25,5,13,21,29]tetraoxatetraazacyclodotriacontine
0.3 hydrate, 9ؒ0.3H2O
The triimine 3ؒ0.6H2O (1.00 g, 3.0 mmol) was dissolved in 1:1
J. Chem. Soc., Perkin Trans. 1, 1998
121

View Article Online
colourless oil. The oil was dissolved in a methanol–hydrochloric
acid solution (66% by volume methanol, 1  acid) and the solu-
tion was chromatographed on DOWEX 50W-X2 ion exchange
resin by elution with 2:1 methanol–hydrochloric acid of
increasing acidity (1–4  in acid). The triamine trihydrochloride
was eluted as the second band from the column with the eluent
3.2–3.4  in acid. The colourless solid remaining after removal
of the solvent was purified by Soxhlet extraction into methanol.
Yield of 12ؒ3HClؒ1.5H2O: 1.30 g (2.1 mmol) (17% of the ori-
ginal mixture), mp 146–148 ЊC (Found: C, 57.6; H, 7.6; N, 6.6.
C30H39N3O3ؒ3HClؒ1.5H2O requires C, 57.6; H, 7.2; N, 6.7%);
δH(299.95 MHz; 2:1 CD3OD–D2O; Me4Si) 2.50 (6 H, m,
CH2CH2CH2), 3.44 (2 H, br m, NH or OH), 3.62 (6 H, m, OCH2
or NCH2CH2), 4.35 (6 H, m, OCH2 or NCH2CH2), 4.44 (6 H, s,
ArCH2N), 7.22 (6 H, m, ArH) and 7.60 (6 H, m, ArH); δC(299.95
MHz; 2:1 CD3OD–D2O) 26.59 (CH2CH2CH2), 46.32 (OCH2 or
NCH2CH2), 47.45 (OCH2 or NCH2CH2), 66.31 (ArCH2N),
112.90 (ArC), 120.02 (ArC), 122.38 (ArC), 132.58 (ArC) and
H2O), 1.97 (8 H, m, CH2CH2CH2), 2.74 (8 H, t, J 7, OCH2 or
NCH2CH2), 3.76 (8 H, s, ArCH2N), 4.04 (8 H, t, J 7, OCH2 or
NCH2CH2), 6.86 (8 H, m, ArH) and 7.20 (8 H, m, ArH);
δC(299.95 MHz; CDCl3) 29.87 (CH2CH2CH2), 46.28 (OCH2 or
NCH2CH2), 49.60 (OCH2 or NCH2CH2), 66.06 (ArCH2N),
111.01 (ArC), 120.31 (ArC), 128.29 (ArC), 129.75 (ArC) and
156.93 (ArC); m/z (EI) 652.2 (M ϩ); Mr (osmometry in CH2Cl2)
᝽
Calc. 652.9, Found 659.
References
1 (a) S. M. Nelson, Pure Appl. Chem., 1980, 52, 2461 and references
cited therein; (b) M. G. B. Drew, F. S. Esho and S. M. Nelson,
J. Chem. Soc., Dalton Trans., 1983, 1653.
2 P. A. Gugger, D. C. R. Hockless, G. F. Swiegers and S. B. Wild,
Inorg. Chem., 1994, 33, 5671.
3 I. R. Goldman, J. K. Larson, J. R. Tretter and E. G. Andrews, J. Am.
Chem. Soc., 1969, 91, 4941.
4 R. W. Kluiber and G. Sasso, Inorg. Chim. Acta, 1970, 4, 226.
5 J. W. L. Martin, K. P. Wainwright, K. D. V. Weerasuria and
S. B. Wild, Inorg. Chim. Acta, 1985, 99, L5.
157.72 (ArC); m/z (EI) 489.4 (12 ϩ). Addition of saturated
᝽
aqueous sodium hydroxide to the hydrochloride, followed by
extraction of the mixture with dichloromethane, produced,
after separation of the organic phase (which was dried over
MgSO4) and evaporation to dryness, 12ؒ1.5H2O as a colourless
oil that could not be crystallised or distilled (Found: C, 69.6; H,
8.3; N, 8.4. C30H39N3O3ؒ1.5H2O requires C, 69.7; H, 8.2; N,
8.2%); δH(299.95 MHz; CDCl3; Me4Si) 2.00 (11 H, m, NH, H2O
and CH2CH2CH2), 2.80 (6 H, t, J 7, OCH2 or NCH2CH2), 3.81
(6 H, s, ArCH2N), 4.05 (6 H, t, J 7, OCH2 or NCH2CH2), 6.86
(6 H, m, ArH) and 7.24 (6 H, m, ArH); δC(299.95 MHz; CDCl3)
29.62 (CH2CH2CH2), 46.15 (OCH2 or NCH2CH2), 49.39
(OCH2 or NCH2CH2), 66.00 (ArCH2N), 111.01 (ArC), 120.42
6 J. W. L. Martin, G. J. Organ, K. P. Wainwright, K. D. V. Weerasuria,
A. C. Willis and S. B. Wild, Inorg. Chem., 1987, 26, 2963.
7 J. W. L. Martin, F. S. Stephens, K. D. V. Weerasuria and S. B. Wild,
J. Am. Chem. Soc., 1988, 110, 4346.
8 P. A. Duckworth, F. S. Stephens, K. P. Wainwright, K. D. V.
Weerasuria and S. B. Wild, Inorg. Chem., 1989, 28, 4531.
9 J. Bergman and A. Brynolf, Tetrahedron Lett., 1989, 30, 2979.
10 D. C. R. Hockless, A. J. Leong, L. F. Lindoy, G. F. Swiegers and
S. B. Wild, Inorg. Chim. Acta, 1996, 246, 371.
11 (a) G. A. Melson and D. H. Busch, J. Am. Chem. Soc., 1965, 87,
1704; (b) G. A. Melson and D. H. Busch, J. Am. Chem. Soc., 1964,
86, 4834.
12 M. C. Burla, M. Camalli, G. Cascarano, G. Polidori, R. Spanga and
D. Viterbo, J. Appl. Cryst., 1989, 22, 389.
13 P. T. Beurskens, G. Admiraal, G. Beurskens, W. P. Bosman,
S. Garcia-Granda, R. O. Gould, J. M. M. Smits and C. Smykalla,
DIRDIF92, The DIRDIF program system, Technical Report of
the Crystallography Laboratory, University of Nijmegen, The
Netherlands, 1992.
14 D. T. Cromer and J. T. Waber, in International Tables for X-ray
Crystallography, Kynoch Press, Birmingham, 1974, vol. IV, Table
2.2A.
15 J. A. Ibers and W. C. Hamilton, Acta Crystallogr., 1964, 17, 781.
16 D. C. Creagh and W. J. McAuley, in International Tables for
Crystallography, ed. A. J. C. Wilson, Kluwer Academic Publishers,
Boston, 1992, vol. C, Table 4.2.6.8, p. 219.
17 D. C. Creagh and J. H. Hubbell in International Tables for
Crystallography, ed. A. J. C. Wilson, Kluwer Academic Publishers,
Boston, 1992, vol. C, Table 4.2.4.3, p. 200.
(ArC), 128.27 (ArC) and 129.74 (ArC); m/z (EI) 489.5 (M ϩ).
᝽
7,8,9,10,17,18,19,20,27,28,29,30,37,38,39,40-Hexadecahydro-
6H,16H,26H,36H-tetrabenzo[b,j,r,z][1,9,17,25,5,13,21,29]-
tetraoxatetraazacyclodotriacontine 0.2 hydrate, 10ؒ0.2H2O
Lithium aluminium hydride (0.14 g, 3.7 mmol) was suspended
in anhydrous tetrahydrofuran (150 ml) and the mixture was
cooled (0 ЊC). Solid 9 (0.55 g, 0.85 mmol) was added to the
mixture in small portions over 1 h. The mixture was then
warmed to room temperature, heated under reflux for 4 h, and
stirred overnight. The following reagents were then added
sequentially to the reaction mixture at 0 ЊC: water (0.14 ml), 4 
NaOH solution (0.14 ml), water (0.43 ml). The mixture was
filtered and the grey precipitate of aluminium oxides was
washed with tetrahydrofuran. The filtrate and washings were
evaporated to dryness and diethyl ether was added to induce
crystallisation of the colourless product. Yield: 0.45 g (0.68
mmol) (81%); mp 149–150 ЊC (Found: C, 73.0; H, 8.1; N, 8.3.
C40H52N4O4ؒ0.2H2O requires C, 73.2; H, 8.1; N, 8.5%);
δH(299.95 MHz; CDCl3; Me4Si) 1.80 (4.3 H, br m, NH and
18 TEXSAN, Crystal Structure Analysis Package, Molecular Structure
Corporation, 1985 and 1992.
Paper 7/04255H
Received 17th June 1997
Accepted 1st September 1997
122
J. Chem. Soc., Perkin Trans. 1, 1998