Branched non-covalent complexes between carboxylic acids and two tris(amidines)

Article abstract of DOI:10.1039/a809190k

Carboxylic acids and two tris(amidine) bases formed branched 3:1 complexes with high solubility in chlorinated and aromatic solvents, particularly when aromatic carboxylic acids with suitable solubilising substituents were used. Whereas N,N′-diethylamidine complexes 10 proved to be difficult to isolate, the respective imidazoline complexes 14 were easily purified by crystallisation. Association constants were determined for model bis(imidazoline) complexes to be about 10³ dm³ mol^-1 in the competitive solvent mixture CDCl₃-CD₃OD (97:3).

Full text of DOI:10.1039/a809190k

View Article Online / Journal Homepage / Table of Contents for this issue
Branched non-covalent complexes between carboxylic acids and
two tris(amidines)
Arno Kraft
Institut für Organische Chemie und Makromolekulare Chemie II, Heinrich-Heine-Universität
Düsseldorf, Universitätsstr. 1, D-40225 Düsseldorf, Germany.
E-mail: kraft@iris-oc2.oc2.uni-duesseldorf.de
Received (in Cambridge) 24th November 1998, Accepted 4th February 1999
Carboxylic acids and two tris(amidine) bases formed branched 3:1 complexes with high solubility in chlorinated
and aromatic solvents, particularly when aromatic carboxylic acids with suitable solubilising substituents were used.
Whereas N,NЈ-diethylamidine complexes 10 proved to be difficult to isolate, the respective imidazoline complexes 14
were easily purified by crystallisation. Association constants were determined for model bis(imidazoline) complexes
to be about 103 dm3 molϪ1 in the competitive solvent mixture CDCl3–CD3OD (97:3).
of supramolecular complex is the basis of numerous hydrogen-
Introduction
bonded calamitic liquid crystals.11 Since amidine–carboxylic
Amidines are important basic groups in Medicinal Chemistry
acid complexes are strengthened by two linear hydrogen bonds,
where they often serve as binding sites for carboxylates and
their association constants are much larger, typically about 103
phosphates, for example, in trypsin inhibitors containing
dm3 molϪ1 in DMSO.12 The binding of carboxylic acids to a
arginine mimics,1 in platelet fibrinogen receptor antagonists2 or
trifunctional amidine base in a 3:1 ratio seemed therefore to be
in drugs that interact with the minor groove of DNA.3
A
quite promising for the self-assembly of branched or possibly
even dendritic molecules. A branched structure was supposed
to have a beneficial effect on the solubility of highly polar amid-
ine derivatives in non-polar solvents in which hydrogen bonding
and electrostatic ion-pair interactions should favour strong
complexation. Tris(amidine) 313 was the first choice for this
approach. We were, however, unable to obtain 3 in sufficient
purity and amount. The substituted amidine derivatives (with
three N,NЈ-diethyl substituted amidine or imidazoline substitu-
ents) finally proved not only to be much more readily accessible,
but also the resulting complexes had satisfactory solubility
properties in chlorinated and aromatic solvents. Two examples
will be described in this paper.
polymerisable amidine, N,NЈ-diethyl-4-vinylbenzamidine, has
recently been applied by Wulff et al. in the design of molecu-
larly imprinted polymers with esterase-like catalytic activity.4
Hydrogen-bonding between dendrimers has been investigated
by Zimmerman et al. who reported the strong binding of a
monodendron containing an amidinium group to a dendritic
host with a naphthyridine receptor.5 Heterocyclic amidines play
also an increasing rôle in various antihypertensive drugs that
bind to imidazoline receptors in the central nervous system.6
Examples in supramolecular chemistry are Anslyn’s tris(amino-
dihydroimidazolium) receptor for citrate7 and Hosseini’s
crystal structure studies of hydrogen-bonded one- and two-
dimensional networks that are based on 1,2-bis(tetrahydro-
pyrimidin-2-yl)ethane or 1,2-bis(dihydroimidazol-2-yl)ethane
and various dicarboxylic acids, sulfonates and phosphates.8
Amidines attracted our attention when, during the prepar-
ation of an oxadiazole-containing dendrimer, we erroneously
identified a DBU-derived by-product of a palladium-catalysed
carbonylation as the salt of carboxylic acid 1 and DBU.9 A
purification method for complexes of carboxylic acids and
DBU or other amidines was developed at a much later stage,
with the effect that the false assignment, in fact, initiated our
present research project on amidine complexes.
Results and discussion
Synthesis of tris(amidine) 7, tris(imidazoline) 13 and their
complexes 10 and 14
Tris(amidine) 7 was prepared in four steps from benzene-1,3,5-
tricarboxylic acid (4). Reaction of 4 with oxalyl chloride, fol-
lowed by aqueous ethylamine afforded amide 5 (Scheme 1).
Imidoyl chloride 6 was obtained after treatment with SOCl2,
and could be converted to 7 with anhydrous ethylamine in 31%
overall yield. Amidine 7 was further purified by vacuum distil-
lation. It hydrolysed, however, slowly during prolonged storage
at room temperature.
Imidazoline 13 was synthesised in a more straightforward
way (Scheme 2).14 The heterocyclic amidine derivative was
obtained in one step by solution condensation of 4 with ethyl-
enediamine and ethylenediamine dihydrochloride in boiling
ethylene glycol, following a general procedure for the prepar-
ation of imidazolines.15 The synthesis proceeded smoothly, and,
after basic work-up and gradient sublimation, 13 was isolated
in high purity, with yields ranging from 21 to 43%.
H
H
H
H
CO2H
N
N
H
H
N
H
N
H
O
O
Ar
Ar
H
H
N
N
N
N
N
H
N
H
3 Cl
Ar = 4-ButC6H4
Ar = 3,5-But2C6H3
1
2
3
In contrast to amidines, the association constant between, for
example, acetic acid and a simple amine such as triethylamine is
rather weak, amounting to only 3000 dm3 molϪ1 in chloro-
form.10 Despite the single hydrogen bond, complexes between
pyridines and mesogenic carboxylic acids are strong in the con-
densed phase below the isotropisation temperature. This type
1,3,4-Oxadiazole-containing acids 1, 2 and 18 were readily
obtained by a reported method that used a palladium-catalysed
carbonylation of aryl iodides for the conversion to the corre-
sponding acid (Scheme 3).9
Amidine complexes 10 (see Scheme 1) and 14 (see Scheme 2)
formed easily on dissolution of tri-basic 7 or 13 and 3 equiv. of
J. Chem. Soc., Perkin Trans. 1, 1999, 705–714
705

View Article Online
O
NHEt
O
H
H
H
N
HA
N
N
HA
N
HO2C
CO2H
i, ii
iii
EtHN
N
H
N
H
N
N
H
iii
ii
i, ii
4
HA
H
HA
H
HA
H
HA
CO2H
O
NHEt
N
N
N
N
4
5
3 X–
11
12
X
X
= Cl
=
13
Et
Et
HA
H
N
iv
N
N
EtN
Cl
Cl
CF3
H
N
Et
NEt
iv
B
Et
HA
HA
Et
v
CF3
4
N
N
H
EtN
Cl
Et
6
R
7
v
14
R-CO2H
O
H
O
H
vi
a
b
c
d
e
f
CF3CO2H
N
HA
N
C6H5CO2H
R
O
H
O
H
R
3-CF3C6H4CO2H
4-ButC6H4CO2H
N
H
N
H
Et
Et
HA
H
H
Et
H
Et
N
N
N
HA
N
O
O
HA
H
HA
H
1
H
N
Et
H
O
O
N
N
H
N
2
N
N
Et
Et
R
O
O
R
Et
g
18
HA
HA
Et
HA
HA
H
Et
Et
N
H
O
N
H
O
N
H
N
Scheme 2 Reagents and conditions: i, H2NCH2CH2NH2, H2NCH2-
CH2NH2ؒ2HCl, TsOH, ethylene glycol, reflux, 3 h; ii, HCl; iii, NaOH;
iv, NaB[3,5-(CF3)2C6H3]4 (3 equiv.), CH3CN; v, RCO2H (3 equiv.),
EtOH (CHCl3), reflux.
3 X
Et
8
9
X
= Cl
=
vii
R
10 R-CO2H
15 and amine 20, a literature-known hydrolysis product of
DBU that forms when DBU is treated with aqueous hydroxide
(Scheme 4).17 The resulting amide 21 was found to be identical
to a sample isolated after one of our earlier carbonylation
reactions. How it was formed in the first place and whether the
DBU batch used was contaminated with 20 or decomposed
during the reaction is unclear at present.
CF3
CF3
X
B
a
b
c
CH3CO2H
4-ButC6H4CO2H
4
1
Scheme 1 Reagents and conditions: i, (COCl)2, DMF, toluene, 60 ЊC,
3 h; ii, EtNH2; iii, SOCl2, reflux, 3 h; iv, EtNH2, CH2Cl2, Ϫ10 ЊC, 1 h;
v, HCl; vi, RCO2H (3 equiv.), EtOH, reflux; vii, NaB[3,5-(CF3)2C6H3]4
(3 equiv.), CH3CN.
Complex properties
It should be noted that the basicity of amidines varies consider-
ably. Whereas the pKa of N,NЈ-dimethyl-4-chlorobenzamidine
hydrochloride (11.41 in 50% alcohol)15 is almost as large as that
of benzamidinium salts (11.6 in water)18 or DBU–Hϩ (12.9 in
water),19 the pKa of a heterocyclic amidine derivative, such
as protonated 2-phenylimidazoline, tends to be significantly
smaller (9.64 in 50% alcohol).15 The differences in pKa values of
carboxylic acids and protonated amidines were supposed to be
large enough for proton transfer to occur, thus ensuring that, in
non-polar solvents, amidine–carboxylic acid complexes consist
of close ion pairs formed by negatively charged carboxylates
and amidinium cations. In fact, neutralisation takes place in
aqueous and ethanolic solution on combining, for example, 13
and 3 equiv. of a water-soluble acid.
carboxylic acid in a suitable polar solvent or solvent mixture,
usually EtOH or a combination of a solvent and a less volatile
non-solvent, such as EtOH–CHCl3 or MeOH–MeCN. Crystal-
lisation occurred in all cases upon cooling, if necessary, after
evaporation of some of the solvent.
DBU–Carboxylic acid complexes
A literature survey revealed that strong DBU–carboxylic acid
interactions in solution have been noted before.16 Various
DBU–carboxylic acid complexes 19a–c were obtained analytic-
ally pure after crystallisation of an equimolar mixture of DBU
and the acid component from a suitable non-polar solvent (e.g.
CH3CN). The complexes were quite hygroscopic and decom-
posed on silica gel. Nevertheless, complexes 19b–c were stable
enough to exhibit peaks of about 12% intensity for DBUؒ1 ϩ
Hϩ and DBUؒ2 ϩ Hϩ, respectively, in the mass spectra after
chemical ionisation with NH3.
Comparison of the NMR data made it clear that the previ-
ously isolated by-product during the preparation of an
oxadiazole-containing dendrimer9 could not be the DBU salt
19b of carboxylic acid 1, but had to be a structural isomer
instead. A second review of the analytical data suggested amide
21 as a possible alternative. The covalently linked amide 21 was
therefore prepared independently by coupling of acid chloride
Most complexes derived from amidine 7 were highly soluble
in a variety of polar and non-polar solvents.† All investigated
† The high solubility of such complexes may be used to advantage for
the chromatographic purification of carboxylic acids, such as 1 or 18,
that are notorious for their low solubility in most solvents. Concen-
trated solutions of the crude acids and amidine 7 could be made in
CH2Cl2 or CHCl3, which were then transferred to a silica gel column.
Gradient elution allowed the non-polar by-products to be removed first
before the complex was destroyed with a more polar eluent (in the case
of 1, elution of the acid required CH2Cl2–MeOH, 9:1). Under these
conditions most of the amidine remained adsorbed on the silica, but
any traces could be quantitatively removed by acidic aqueous work-up
of the acid-containing fractions.
706
J. Chem. Soc., Perkin Trans. 1, 1999, 705–714

View Article Online
crystallisation and slow hydrolysis made purification of the
X
amidine complexes cumbersome. A different amidine derivative
was accordingly sought that did not have these drawbacks.
Imidazoline complexes 14 were usually obtained free of sol-
vent impurities. Although tris(imidazoline) 13 scarcely dissolves
in non-acidic polar solvents (<2 mg cmϪ3 in methanol or water),
the solubility of its complexes in chloroform is surprisingly
high, extending to 40–50 mg cmϪ3 for 14e and even to >100 mg
cmϪ3 for 14c,d,f. The insolubility of 14b in neat CHCl3
emphasises that the acid component must have suitable solubil-
ising groups. Complexes 14c–f do not self-associate in chloro-
form or benzene to any extent.‡ It was, however, reported earlier
that amidocarboxylic acids give rise to inter-complex hydrogen
bonding in solution.20
In contrast to DBU salts 19b–c, none of the complexes 10, 14
or 23 were stable enough under various conditions tried for
mass spectrometry (chemical ionisation, fast atom bombard-
ment or matrix-assisted laser desorption/ionisation), allowing
only molecular ions and fragments of the components to be
detected. One reason may be that, when 10a, 14a, 19a or 23
were heated under vacuum, only the DBU salt could be sub-
limed without decomposition. An attempt to determine the
molar mass of complexes 10c or 14f by gel permeation chrom-
atography in CH2Cl2 or THF at 25 ЊC also failed. In fact, both
complexes and their components eluted as extensively broad-
ened peaks long after the low-molar mass standards used. This
was not observed with the corresponding acids on their own
and suggests that 10c and 14f adhered to the column material.
Such an adhesion effect of amidines is not without precedence
and has been observed by others before.21
Support for the formation of 3:1 complexes in non-polar
solution could, nevertheless, be obtained by Job’s method of
continuous variation22 and vapour-pressure osmometry studies.
The insolubility of tetrakis[3,5-bis(trifluoromethyl)phenyl]-
borate salts 9 and 12 (obtained from the corresponding
hydrochlorides by ion exchange) in neat chloroform made it
necessary to add a polar co-solvent. When the stoichiometry of
the complexes formed by tris(amidine) 9§ or tris(imidazoline)
12 and tetramethylammonium benzoate was determined in a
CDCl3–CD3CN (6:1) mixture, the maximum of the Job plot
was observed in both cases at a mole fraction of 0.25, as would
be expected for 3:1 complexes (Fig. 1). Vapour-pressure
osmometry studies in CHCl3 at 34 ЊC (against benzil or poly-
styrene 2000 as standard) also supported 3:1 stoichiometry.
The experimentally determined number average molar mass
values Mn for complexes 10b–c and 14c–f were in good agree-
ment with the calculated molar masses (see Experimental).
O
O
Ar = 4-ButC6H4
Ar
i
Ar
N
X = CO2H
X = COCl
1
N
N
N
15
ii
I
O
O
Ar
N
Ar
NH
HN
O
O
O
HN
NH
N
N
N
O
16
O
O
N
N
Ar = 4-ButC6H4
N
N
Ar
Ar
iii
Ar
O
Ar
N
N
O
X
N
N
O
N
O
N
N
N
N
N
N
N
O
O
Ar = 4-ButC6H4
Ar
Ar
X = I
X = CO2H
17
18
iv
Scheme 3 Reagents and conditions: i, (COCl)2, DMF, toluene, 60–
110 ЊC, 7 h; ii, 5-iodoisophthalic dihydrazide, NMP, 25 ЊC, 15 h; iii,
ClSO3H, 25 ЊC, 18 h; iv, LiOHؒH2O, PdCl2, Ph2P(m-C6H4SO3Na),
NMP, CO, 100 ЊC, 1 d, then HCl.
N
N
i
N
N
H
19
Ar-CO2H
a
b
c
4-ButC6H4CO2H
O
O
2
3
1H NMR Spectra of amidine 7 and its complexes
ii
Ar
1
Broad signals are observed in the high field H and 13C NMR
spectra of 7. As with other substituted amidines, slow proton
exchange between tautomers23 and/or hindered rotation
around the C–NEt bond24 can account for the observation of
two sets of broad singlets for the N-ethyl protons at 500 MHz at
room temperature. The barrier to rotation in, e.g., N,NЈ-
dimethylacetamidine (38–105 kJ molϪ1) is known to be in the
same range as that of amides.25 It should be noted that the
structure of unsubstituted and N-monoalkylated or N,NЈ-
dialkylated amidines differs in the twist angle between the aryl
ring and the amidine group which is reported to be close to 0Њ in
the former, but increases up to 62Њ in substituted amidines.26,27
O
O
iii
H2N
N
N
O
N
H
20
O
O
Ar
Ar
N
N
N
N
Ar = 4-ButC6H4
21
‡ Self-association is observed in hexane or cyclohexane. Solubility in
these highly non-polar solvents requires benzoic acids with long alkoxy
side chains, as is the case for a complex formed by 13 and 3,4,5-
tris(dodecyloxy)benzoic acid: A. Kraft and A. Reichert, unpublished
results.
Scheme 4 Reagents and conditions: i, ArCO2H, MeOH, reflux; ii,
KOH, MeOH–H2O, 25 ЊC, 40 h; iii, 15, NMP, 20 ЊC, 15 h.
complexes 10 were rather difficult to crystallise, probably yet
another consequence of the large number of solubilising ethyl
groups. Their tendency to include solvents or moisture during
§ Since the 1H NMR signals of the tris(amidine) broadened extensively
upon addition of benzoate, the signal of the benzoate proton para to
the carboxylate group was used for the Job plot.
J. Chem. Soc., Perkin Trans. 1, 1999, 705–714
707

View Article Online
are influenced by temperature and NMR frequency, and the
samples have to be heated to 100 ЊC until the NMR signals in
[2H6]DMSO become sharp again. We note that the HA signal of
10 at δH ≈ 8.2 is also considerably broadened. Hindered rotation
around the Ar–Camidine bond therefore becomes the most likely
explanation resulting from congestion of the three substituted
amidine groups at the 1,3,5 positions of the same benzene
core. For comparison, monoamidine complex 23 (Scheme 5)
NEt
NHEt
Fig. 1 Job plot for 12 binding tetramethylammonium benzoate. The
total concentration was maintained at 10Ϫ3 mol dmϪ3 in CDCl3–CD3-
CN (6:1). The mole fraction x is defined as [12]/([12] ϩ [NMe4PhCO2]).
The maximum of the Job plot at a mole fraction x of 0.25 is consistent
with a 3:1 complex stoichiometry in this solvent mixture (for a 2:1
complex, i.e. on partial dissociation, the maximum would have been
expected at x = 0.33).
22
i
Et
N
N
H
H
O
O
But
Et
23
Scheme 5 Reagents and conditions: i, 4-ButC6H4CO2H, EtOH, reflux.
displays only a single set of relatively sharp 1H NMR signals for
its N-ethyl protons at 500 MHz in CDCl3. A similar dynamic
behaviour has been reported for 2,4,6-tris(dialkylamino)-s-
triazines in which the sterically crowded alkyl groups of the
dialkylamino substituents perform correlated rotations.28
1H NMR Spectra of the complexes with imidazoline 13
None of these complications can occur with conformationally
rigid heterocyclic amidines. The 1H NMR chemical shift of the
aromatic HA protons of 13 (δH = 8.23 in CD3OD) is character-
istic for benzene derivatives with three weakly electron-
withdrawing substituents. Protonation with HCl leads to slight
downfield shifts (δH ≈ 8.6 in CD3OD and D2O, δH = 8.93 in
[2H6]DMSO). Similarly, the HA and NH resonances of borate
12 with its non-coordinating counter-anion are found at
δH = 8.48 and 9.1, respectively, even in the less polar CDCl3–
CD3CN (6:1) solvent mixture [Fig. 3(a)]. Protonation of 13
Fig. 2 1H NMR spectrum (500 MHz, 25 ЊC) of 10b in CDCl3. Arrows
indicate the aromatic and N-ethyl signals of the amidine component.
Signals of an amide impurity are marked by an asterisk; the amount of
hydrolysis product increases slowly with time upon standing at room
temperature.
The crystal structure of N,NЈ-dimethylbenzamidine further-
more shows a preference for the (E,Z)-configuration at the C–N
amidine bonds.26
Simple space-filling model studies imply that 7 (and, even
more so, its complexes 10) cannot be planar. The presence of
three substituted amidine groups next to a benzene ring gener-
ates a variety of possible stereoisomers and conformers. Low
temperature 1H NMR spectra (Ϫ40 ЊC, CDCl3, 300 MHz) of 7
accordingly show three major singlets for the aromatic HA pro-
tons at δH = 7.22, 7.26 and 7.43 in the ratio 2:3:5 and an even
more complicated pattern for the N-ethyl triplets and quartets,
consistent with a complex mixture of tautomers and rotamers.
The N-ethyl signals simplify to a quartet and a triplet at
lower 1H NMR frequency (90 MHz), in a protic solvent
(CD3OD) and at higher temperature (100 ЊC in [2H6]DMSO),
indicating fast exchange on the NMR timescale under these
conditions.
1
with carboxylic acids gives 3:1 salts that show comparable H
NMR chemical shifts for the HA signal (δH ≈ 8.6) in polar sol-
vents, such as [2H6]DMSO, CD3OD or D2O. In contrast, when a
complex, e.g. 14, has sufficient solubilising groups that it dis-
solves in non-polar solvents in which dissociation can be con-
sidered to be negligible, its HA signal is found at much lower
field (δH ≈ 10.1 in CDCl3 and δH ≈ 10.0 in C6D6) as illustrated
by Fig. 3(b). Such large downfield shifts of ∆δ ≈ 1.5 for an
aromatic proton signal cannot be explained by simple solvent
effects alone, but are attributed to the binding of each carb-
oxylate to the NH groups of two imidazolines in a bridged
arrangement. As was previously confirmed by the crystal struc-
ture of model complex 14a, this binding geometry results in
carboxylate–O ؒ ؒ ؒ HA contacts (2.26–2.63 Å for the shorter,
2.77–3.40 Å for the longer distance) close to the sum of the van
der Waals radii (2.6 Å).14 Both the field induced by the nearby
dipole of the carboxylate–imidazolinium ion pair and close
contacts then explain the unusual deshielding of the HA reson-
ance. The crystal structure of 14a further demonstrated that the
complex lacks symmetry and that it is non-planar. Owing to the
size of the carboxylate group, two carboxylates have to twist
considerably out of the plane of an almost planar tris(imidazo-
line) core.
On the other hand, the 1H NMR spectrum of the protonated
amidines 8 and, to a lesser extent, 9 show two sets of sharp
signals for the N-ethyl and NH protons, consistent with an
(E,Z)-stereochemistry at the partial C–N double bond.25 A high
rotational barrier is also apparent from the vicinal spin-spin
coupling between amide NH and N–CH2 protons observed
for 8.
Complexation of an amidine with carboxylic acids requires
exclusive (E,E)-stereochemistry at the amidine C–N bond
which should be easily identified by NMR. Nevertheless, the
1
300 and 500 MHz H NMR spectra of complexes 10 show
Determination of association constants
again broad singlets for both the tris(amidine)’s aliphatic and
aromatic protons at room temperature (Fig. 2). The lineshapes
In all cases, complex association and dissociation was fast on
708
J. Chem. Soc., Perkin Trans. 1, 1999, 705–714

View Article Online
Table 1 Summary of binding constants of imidazoline receptorsa
Compound
Solventa
Ka/dm3 molϪ1
∆δ
H followed
26
A
A
A
B
800 100
780 110
510 50
59400 7200
990 230
0.79
0.30
0.87
1.11
1.13
HA
HB
HA
HA
27
29b
29b
A
HA (ref. 14)
a Binding constants were determined by 1H NMR dilution studies at
25 ЊC in either solvent A [CDCl3–CD3OD (97:3)] or B [CDCl3–CD3CN
(6:1)]. b A 1:1 ratio of 29:tetrabutylammonium benzoate was used.
R
R
H
N
i, ii, iii
N
HO2C
CO2H
N
N
H
R = H
24
25
R = But
Fig. 3 1H NMR spectra (500 MHz) of (a) 12 and of (b) 14b in CDCl3–
CD3CN (6:1) showing the difference in the HA chemical shift in the
presence of a non-coordinating counter-anion (12) and benzoate (14b).
iv
But
ii
the NMR timescale. A broad singlet at δH ≈ 12–13 (for 10a–c
and 14c–g in CDCl3 or C6D6) can be assigned to the NH
protons and indicates strong hydrogen bonding. The signal is,
however, too broad to be followed during NMR titrations,
especially at lower concentrations. According to Wulff and
Schönfeld, the association constant of an amidine–carboxylic
acid complex similar to 23 exceeds 106 dm3 molϪ1 in CDCl3 at
25 ЊC.21 It was expected that imidazoline complexes would show
weaker binding since non-linear hydrogen bonds are involved.
The 1H NMR chemical shifts of imidazoline complexes 14c–f
in CDCl3 remain almost unchanged (∆δ < 0.1) over a concen-
tration range of 10Ϫ1 to 10Ϫ5 mol dmϪ3 and, for example, when
a solution of 14f in CDCl2CDCl2 is heated to 120 ЊC; in the
latter case, the HA signal shifted even further downfield.
Although the tris(imidazoline) complexes show some line-
broadening below 10Ϫ3 mol dmϪ3, this may be caused by traces
of residual water. Addition of a polar co-solvent (e.g. CD3OD)
resulted in a gradual upfield shift of the HA signal and was
attributed to the onset of complex dissociation as well as to the
increased solvent polarity.
O
R
O
HB
HB
H
H
N
H
N
H
N
N
N
HA
N
N
HA
N
H
H
H
H
O
2 X
O
28
29
X
X
= Cl
v
CF3
CF3
=
B
But
26
27
R = HC
R = But
4
Scheme 6 Reagents and conditions: i, H2NCH2CH2NH2, H2NCH2-
CH2NH2ؒ2HCl, TsOH, ethylene glycol, reflux, 3 h; ii, HCl; iii, NaOH;
iv, 4-ButC6H4CO2H (2 equiv.), EtOH, reflux; v, NaB[3,5-(CF3)2C6H3]4
(2 equiv.), CH3CN.
Association constants for a 3:1 complex can only be derived
in a few special cases, for example when the association process
is highly cooperative. The binding of a carboxylate between two
imidazolinium groups was therefore studied with bis(imid-
azoline) model compounds 26, 27 and 29, since in these cases
only 1:1 and 2:1 complexes are possible. It was found that the
HA signals of these bis(imidazolines) were also sensitive to the
extent of complexation.
concentration, also for the HB and N-CH2 (but not HA)
signals since these protons are affected by the onset of weak
binding of a second carboxylate. Association constants Ka
for various model systems are listed in Table 1 and were found
to be in good agreement with previous results. The smaller Ka
and reduced chemical shift for complex 27 is attributed to a
slight change in binding geometry accompanied by interference
of the sterically demanding tert-butyl group during complex-
ation.
When protonated bis(imidazoline) 29 was titrated with
tetrabutylammonium benzoate in CDCl3–CD3CN mixtures in
which 29 was sufficiently soluble, it was noticed that under these
conditions 2:1 complexes formed with increasing benzoate:29
ratio. This complication could be avoided by 1H NMR dilution
experiments.29 An association constant Ka in excess of 104 dm3
molϪ1 was derived from the dilution of an equimolar ratio of 29
and tetrabutylammonium benzoate in CDCl3–CD3CN (6:1).
Similar dilution studies in a more competitive solvent mixture,
Towards dendritic structures
Despite the fact that several routes to first-generation 1,3,4-oxa-
diazole-containing dendrimers have already been developed,9,30
all our efforts to prepare the second-generation dendrimer
failed so far. It was therefore of interest whether or not non-
covalent binding, such as between carboxylic acid 19 and a
tris(amidine) base, may be a suitable alternative. Such an
approach had already been used successfully for the complex-
ation of a branched amidocarboxylic acid,20 and should be
equally applicable to other monodendrons.
CDCl3–CD3OD (97:3), gave a Ka of 990 230 dm3 molϪ1
.
Since model 2:1 complexes 26 and 27 were easily accessible
(Scheme 6), both complexes with their defined host–guest ratio
were also subjected to dilution studies. The change in δ(HA)
with varying concentration could be evaluated as simple 1:1
host–guest complex formation. As might be expected, some
deviations from hyperbolic dependence of δ as a function of
the concentration were observed for HC and, with increasing
Complex 14g was obtained after precipitation and centri-
J. Chem. Soc., Perkin Trans. 1, 1999, 705–714
709

View Article Online
complexes¶ and the development of strongly self-associating
systems with columnar superstructures.
Experimental
General
All solvents were distilled prior to use. Melting points: Olympus
BH-2 polarisation microscope with a Linkam THMS600 hot
stage and a TMS91 temperature controller. DSC: Mettler DSC
30 with TC 11/TA 4000 Processor (10 ЊC minϪ1; K: crystalline,
X: unidentified phase transition, I: isotropic liquid). NMR:
Varian VXR 300, Bruker DRX 500, Varian Unity plus (13C: 150
MHz). TMS was used as internal standard in the NMR meas-
urements. The multiplicities of 13C signals were determined by
DEPT experiments. IR: Perkin-Elmer Ratio Recording Infrared
Spectrophotometer 1420, Bruker Vector 22 FT-IR. EI-MS:
Varian MAT 311 A (70 eV). CI-MS: Finnigan INCOS 50.
MALDI-TOF-MS were measured at the University of Münster
with 2,5-dihydroxybenzoic acid as matrix. Elemental analyses:
Pharmaceutical Institute of the Heinrich Heine University,
Düsseldorf. VPO: Knauer vapour-pressure osmometer. The
number-average molar mass Mn was determined for solu-
tions in a concentration range between 10 and 30 mg gϪ1
.
Compounds 1,20 11,20 13,20 14a,14 14d,20 14e,14 20,17 22,4 Ph2P-
(m-C6H4SO3Na)31 and sodium tetrakis[3,5-bis(trifluoromethyl)-
phenyl]borate32 were prepared as described in the literature.
3,5-Bis[5-(3,5-di-tert-butylphenyl)-1,3,4-oxadiazol-2-yl]benzoic
acid 2
Fig. 4 1H NMR spectra (500 MHz) of 14g in (a) CDCl3–CF3CO2D
and in (b) CDCl3 (1.5 mg cmϪ3). The broadened signals in neat CDCl3
are a result of strong self-association. Protonation with trifluoroacetic
acid breaks up all stacking interaction between the oxadiazole-
containing ligands. The large excess of a carboxylic acid also causes the
HA signal to shift upfield. Solvent and water signals are marked by X.
Synthesis as described previously20 for 1 with 3,5-bis[5-(3,5-di-
tert-butylphenyl)-1,3,4-oxadiazol-2-yl]-1-iodobenzene9b (5.73
g, 8.00 mmol), lithium hydroxide hydrate (504 mg, 12.0 mmol),
PdCl2 (81.8 mg, 0.461 mmol), NMP (20 cm3) and carbon mon-
oxide. Chromatography (silica gel, CH2Cl2–MeOH, 9:1) gave 2
(3.00 g, 59%) as a colourless solid, mp 358–360 ЊC (from
MeOH) [Found: C, 73.8; H, 7.3; N, 9.1%; C39H46N4O4 (634.82)
requires C, 73.8; H, 7.3; N, 8.8%]; νmax (KBr)/cmϪ1 3500–2700,
2950, 1715, 1620, 1590, 1540, 1245, 1230, 740, 700; δH(300
MHz, CDCl3) 1.42 (s, CH3), 7.67 (br s, 2 H), 8.03 (br s, 4 H,
C6H3), 9.12 (br s, 2 H), 9.14 (br s, 1 H, C6H3CO2H); δC(125
MHz, CDCl3–[2H6]DMSO, 6:1) 31.3 (CH3), 35.0 [C(CH3)3],
121.3, 126.4, 128.1, 130.5 (arom. CH), 122.8, 125.4, 152.0,
fugation. Although 14g is strictly speaking not a dendrimer,
the complex has at least a comparable shape and degree of
branching. Low solubility (up to 10 mg cmϪ3), a tendency to
1
precipitate, and broad H NMR signals in CDCl3 even at a
concentration as low as 10Ϫ5 mol dmϪ3 were accompanied by
upfield shifts of the aromatic signals and downfield shifts of the
N–CH2 singlet with increasing concentration, thus indicating
strong self-association (stacking) of the extended π-system. All
non-covalent interactions between the components of the com-
plex were then broken up by the addition of trifluoroacetic
acid.20 The purity and correct stoichiometric ratio of the com-
163.0, 166.0 (ipso-C, C᎐O, 2 signals missing); m/z (CI, NH ) 652
᎐
3
(M ϩ NH4ϩ, 12%), 536, 535 (M ϩ Hϩ, 47, 82), 251 (19), 234,
233 (23, 100); Rf(ethyl acetate) 0.13.
1
plex could thus be checked by a H NMR spectrum of the
N,NЈ,NЉ-Triethylbenzene-1,3,5-tricarboxamide 5
complex in CDCl3–CF3CO2D (6:1) which showed the expected
sharp signals of the dissociated complex and the proton-
ated components (Fig. 4). These results indicate that complex-
ation of a monodendron with a carboxylic acid group at the
focal point is indeed possible, but again emphasise that
extended π-systems induce self-association and π-stacking
interactions.
Benzene-1,3,5-tricarbonyl chloride (prepared by the reaction of
4 with oxalyl chloride and a catalytic amount of DMF in tolu-
ene at 60 ЊC) (16.8 g, 63.4 mmol) was added dropwise to ice-
cold aqueous ethylamine (70%, 80 cm3, 1 mol). After stirring for
15 min, the mixture was diluted with water (300 cm3) and conc.
HCl (40 cm3). The colourless precipitate was collected by suc-
tion filtration, washed with water and methanol, and dried
(16.4 g, 89%), mp 292 ЊC [Found: C, 61.8; H, 7.2; N, 14.4;
C15H21N3O3 (291.34) requires C, 61.8, H 7.3; N, 14.4%]; νmax
(KBr)/cmϪ1 3330, 3300, 3250, 3060, 2860, 2820, 1640, 1580,
1540, 1295, 1140, 710, 695; δH(300 MHz, [2H6]DMSO) 1.15 (t,
J 7.1, CH3), 3.32 (dq, J 7.1, 5.4, NH-CH2), 8.38 (s, C6H3), 8.70
(t, J 5.4, NH); δC(75 MHz, [2H6]DMSO) 14.7 (CH3), 34.2
Conclusions
Tris(amidine) bases 7 and 13 easily formed 3:1 salts with carb-
oxylic acids. Suitable solubilising groups ensured that these
salts are soluble in non-polar solvents in which complexation
through hydrogen bonding is strong. Some preparative difficul-
ties arose because of the hydrolytic instability of 7 and the ten-
dency of the corresponding amidine complexes 10 to include
solvents. Imidazoline complexes 14, on the other hand, fulfilled
the criteria of both easy accessibility and purification. Appli-
cations of this complexation principle, especially by using carb-
oxylic acids with specific functions, are currently being pursued
and concentrate on the substitution of carboxylic acids with
acidic heterocycles as well as the design of liquid-crystalline
¶ Several complexes described in this paper showed a liquid-crystalline
mesophase with a “sandy” texture above the melting transition, but
only on first heating. Most samples started to rapidly decompose at the
clearing temperature (ca. 240 ЊC for complexes from simple benzoic
acids 14c–d, and, as a consequence of the more extended aromatic
π-systems, >270 ЊC for 14e–g). These findings demonstrate that
(presumably columnar) liquid-crystalline mesophases are possible by
non-covalent complexation of suitable acids with 13.
710
J. Chem. Soc., Perkin Trans. 1, 1999, 705–714

View Article Online
(CH ), 128.2 (C H ), 135.1 (ipso-C), 165.2 (C᎐O); m/z (EI, 70
cmϪ1 1644, 1612, 1580, 1357, 1280, 1124; δH(500 MHz, CDCl3–
CD3CN, 6:1) 4.12 (s, 12 H, N–CH2), 7.54 (br s, 12 H), 7.67
᎐
2
6
3
eV) 291 (Mϩ, 89%), 262 (57), 247 (100), 220 (29).
[br m, 24 H, (CF3)2C6H3], 8.48 (s, 3 H, HA), 9.15 (br s, NH).
N,NЈ,NЉ-Triethylbenzene-1,3,5-tricarboximidoyl trichloride 6
3,5-Bis(2-{3,5-bis[(4-tert-butylphenyl)-1,3,4-oxadiazol-5-yl]-
phenyl}-1,3,4-oxadiazol-5-yl)iodobenzene 17
A solution of 5 (5.69 g, 19.5 mmol) in SOCl2 (30 cm3) was
heated to reflux for 3 h. Excess SOCl2 was then removed by
distillation. The residual oil was dried in vacuo, extracted with
hexane and filtered. After concentrating the filtrate in vacuo and
drying, the oil was crystallised from hexane to yield 4.43 g
(65%) of 6 as a colourless solid, mp 82 ЊC [Found: C, 51.9; H,
5.4; N, 12.0; C15H18N3Cl3 (346.68) requires C, 52.0; H, 5.2; N,
12.1%]; νmax (KBr)/cmϪ1 2860, 2820, 1650, 1425, 1340, 1160,
995, 895, 670; δH(300 MHz, CDCl3) 1.36 (t, J 7.3, CH3), 3.77 (q,
N–CH2), 8.66 (s, C6H3); δC(75 MHz, CDCl3) 14.6 (CH3), 49.0
A mixture of 1 (2.92 g, 5.59 mmol), oxalyl chloride (1.46 cm3,
16.8 mmol), dry toluene (15 cm3) and DMF (2 drops) was
stirred at 60 ЊC for 2 h, then at 110 ЊC for 5 h until gas evolution
had ceased and all starting material was dissolved. The brown
solution was decanted and concentrated in vacuo to afford 15 as
a light brown residue (2.93 g, 97%), mp 245–248 ЊC (from tolu-
ene) [Found: C, 68.6; H, 5.4; N, 10.1. C31H29ClN4O3 (541.05)
requires C, 68.8; H, 5.4; N, 10.4%]; νmax (KBr)/cmϪ1 2950, 1760,
1610, 1495, 1180, 840, 720; δH(300 MHz, CDCl3) 1.40 (s, CH3),
7.59, 8.11 (AA‘XX’, C6H4), 8.95 (d, J 1.6, 2 H), 9.12 (t, 1 H,
C6H3); δC(75 MHz, CDCl3) 31.1 (CH3), 35.2 [C(CH3)3], 126.2,
127.0, 130.2, 131.1 (arom. CH), 120.3, 126.4, 135.4, 156.1,
(CH ), 131.4 (C H ), 136.5 (ipso-C), 140.1 (C᎐N); m/z (CI,
᎐
2
6
3
NH3) 365, 363 (M ϩ NH4ϩ, 7, 7%), 348, 346 (M ϩ Hϩ, 16, 18),
312, 310 (Mϩ Ϫ Cl, 77, 100), 246 (17).
162.0, 165.6, 167.0 (ipso-C, C᎐O); m/z (EI) 545, 543, 542 (Mϩ,
᎐
1,3,5-Tris(N,NЈ-diethylcarbamimidoyl)benzene 7
23, 22, 62%), 529, 527 (Mϩ Ϫ CH3, 49, 100), 255 (24), 92 (50),
91 (64). A mixture of 15 (2.93 g, 5.42 mmol), 5-iodoisophthalic
dihydrazide9b (867 mg, 2.71 mmol) and NMP (15 cm3) was
stirred at room temperature for 15 h. The clear brown solution
was then added slowly to vigorously stirred water (150 cm3).
The ochre precipitate was collected by suction filtration,
washed with water and dried. After extraction with ethyl acet-
ate and filtration through a short column of silica gel (eluent:
ethyl acetate), the crude product was recrystallised from hot
CHCl3–EtOH to yield 16 as a colourless powder (1.53 g, 43%),
mp 252–253 ЊC; δH(300 MHz, CDCl3–[2H6]DMSO, 1:1) 1.38 (s,
CH3), 7.64, 8.14 (AA‘XX’, C6H4), 8.58 (d, J 1.6, 2 H), 8.66 (t,
1 H C6H3I), 8.96 (approx. s, 6 H, C6H3), 11.04 (s, 2 H, NH),
11.30 (s, 2 H, NH); δC(75 MHz, [2H6]DMSO) 30.7 (CH3), 34.8
[C(CH3)3], 95.0 (C-I), 126.1, 126.7, 128.1, 139.0 (arom. CH, 2
signals missing), 120.3, 125.0, 134.2, 134.5, 155.1, 162.4, 163.4,
A solution of 6 (8.60 g, 24.8 mmol) in CH2Cl2 (20 cm3) was
added dropwise to a solution of dry ethylamine (13.0 cm3, 198
mmol) in CH2Cl2 (50 cm3) at Ϫ20 ЊC. After stirring at Ϫ10 ЊC
for 1 h and then at room temperature overnight, the solution
was concentrated in vacuo and dried. The residue was dissolved
in water (50 cm3). After addition of NaOH (10 g), the mixture
was extracted with ethyl acetate (3 × 40 cm3). The combined
organic extracts were concentrated in vacuo and dried to give an
orange-red oil. Distillation (Kugelrohr, 240–250 ЊC/0.05 mbar)
furnished 7 as a colourless solid (5.18 g, 56%), mp 42–46 ЊC
[Found: C, 67.4; H, 9.7; N, 22.4; C21H36N6 (372.56) requires C,
67.7; H, 9.7; N, 22.6%]; νmax (KBr)/cmϪ1 3314, 2967, 1628, 1528;
δH(300 MHz, CD3OD) 1.12 (br t, J 7.2, CH3), 3.13 (br q,
N–CH2), 7.24 (s, HA); δH(300 MHz, [2H6]DMSO, 100 ЊC) 1.04
(t, J 7.1, CH3), 3.08 (q, N–CH2), 7.07 (s, HA); δC(75 MHz,
CD3OD) 16.0 (br, CH3), 41.0 (br, N–CH2), 128.5 (arom. CH),
137.5 (ipso-C), 161.0 (C᎐N); m/z (CI, NH ) 374, 373 (M ϩ Hϩ,
25, 100%). Hydrochloride 8 was obtained as a light yellow solid
after dissolving 7 in aqueous HCl and freeze-drying, mp 80 ЊC;
δH(500 MHz, [2H6]DMSO) 1.13 (t, J 7.2, CH3), 1.27 (t, J 7.2,
CH3), 3.23 (tt, J 7.2, 5.7, CH2), 3.46 (approx. quintet, J 6.8, CH2),
8.17 (s, C6H3), 9.42 (t, J 5.7, NH), 10.04 (br t, NH). A solution
of 8 (22.1 mg, 0.0442 mmol) and NaB[3,5-(CF3)2C6H3]4 (117
mg, 0.133 mmol) in CH3CN (4 cm3)–water (2 cm3) was concen-
trated in vacuo. The residue was then taken up in CH3CN
(2 cm3), membrane-filtered and concentrated again in vacuo.
Drying at 70 ЊC/10Ϫ4 mbar yielded 9 (84 mg, 64%) as a colour-
less glass (Found: C, 47.4; H, 2.7; N, 3.1; C117H75B3F72N6
requires C, 47.4; H, 2.6; N, 2.8%); νmax (KBr)/cmϪ1 1653, 1357,
1281, 1128; δH(500 MHz, CDCl3–CD3CN, 6:1) 1.15 (br s, 18
H), 1.30 (br s, 18 H, CH3), 3.18 (br s, 12 H), 3.39 (br s, 12 H,
NCH2), 7.55 (br s, 12 H), 7.67 [br m, 24 H, (CF3)2C6H3], 7.73 (s,
3 H, HA), 7.83 (br s, 3 H, NH), 7.95 (br s, 3 H, NH).
163.9, 164.5 (ipso-C, C᎐O); R (ethyl acetate) 0.84. A solution of
᎐
f
᎐
3
16 (1.48 g, 1.11 mmol) in chlorosulfonic acid (5 cm3) was stirred
at 0–5 ЊC for 10 min and then at 30 ЊC for 18 h. The pale yellow-
brown solution was added dropwise to water (100 cm3) under
vigorous stirring. The resulting precipitate was collected by
suction filtration, washed with water and further purified by
chromatography (CH2Cl2–MeOH, 100:3) to give 17 as a col-
ourless powder (952 mg, 66%) that was very soluble in CH2Cl2
or CHCl3, but crystallised from concentrated solutions within
10 min and, once crystallised and aggregated, could only be
redissolved in large amounts of solvent, mp 303–304 ЊC
[Found: C, 64.9; H, 4.8; N, 12.8. C70H61IN12O6 (1293.24)
requires C, 65.0; H, 4.8; N, 13.0%]; νmax (KBr)/cmϪ1 2950, 1610,
1545, 1490, 1265, 1250, 1235, 1110, 840, 795, 780, 720; δH(500
MHz, CDCl3, 3 mg/0.7 cm3) 1.40 (s, CH3), 7.60, 8.15 (AA‘XX’,
C6H4), 8.81 (d, J 1.6, 2 H), 9.04 (t, J 1.6, 1 H, C6H3I), 9.10 (t,
J 1.6, 2 H), 9.12 (d, J 1.6, 4 H, C6H3); δC(150 MHz, CDCl3) 31.1
(CH3), 35.2 [C(CH3)3], 95.0 (C-I), 125.8, 126.2, 127.1, 127.4,
127.7, 138.9 (arom. CH), 120.5, 126.6, 156.0, 162.5, 163.1,
163.7, 165.7 (ipso-C, 2 signals missing); m/z (MALDI-TOF)
1167 (M ϩ Hϩ Ϫ I), 1189 (M ϩ Naϩ Ϫ I), 1294 (M ϩ Hϩ),
1317 (M ϩ Naϩ); Rf(CH2Cl2–MeOH, 100:3) 0.42.
1,3,5-Tris(4,5-dihydro-1H-imidazol-2-yl)benzene 13 and borate
salt 12
For preparation and analytical data, see ref. 20; δH(500 MHz,
CD3OD) 3.78 (s, N–CH2), 8.23 (s, HA). Hydrochloride 11 was
obtained as a light brown solid after freeze-drying a solution of
13 in aqueous HCl; δH(500 MHz, CD3OD) 4.21 (s, N–CH2),
8.63 (s, HA); δH(500 MHz, [2H6]DMSO) 4.10 (s, N–CH2), 8.93
(s, HA), 11.03 (br s, NH). For the preparation of 12 a solution
of 11 (10.3 mg, 0.0263 mmol) and NaB[3,5-(CF3)2C6H3]4
(69.9 mg, 0.0789 mmol) in CH3CN (3.0 cm3)–water (0.5 cm3)
was concentrated in vacuo. The residue was then taken up in
CH3CN (4 cm3), membrane-filtered and concentrated again
in vacuo. Drying at 80 ЊC/10Ϫ5 mbar yielded 12 (72 mg, 95%) as
a colourless glass (Found: C, 45.4; H, 2.1; N, 2.9; C111H57-
B3F72N6ؒ3H2O requires C, 45.5; H, 2.2; N, 2.9%); νmax (KBr)/
3,5-Bis(2-{3,5-bis[(4-tert-butylphenyl)-1,3,4-oxadiazol-5-yl]-
phenyl}-1,3,4-oxadiazol-5-yl)benzoic acid 18
Synthesis as described previously20 for 1 with 17 (1.13 g, 0.872
mmol), lithium hydroxide monohydrate (73.2 mg, 1.74 mmol),
PdCl2 (10.0 mg, 0.0564 mmol), Ph2P(m-C6H4SO3Na) (67.7 mg,
0.169 mmol), NMP (10 cm3) and carbon monoxide. Chroma-
tography (first CHCl3–ethyl acetate, 2:1, then CHCl3–MeOH,
9:1†) yielded a colourless solid (804 mg, 76%), mp 387–388 ЊC
(from CHCl3–MeOH) [Found: C, 70.6; H, 5.0; N, 13.6.
C71H62N12O8 (1211.36) requires C, 70.4; H, 5.2; N, 13.9%]; νmax
(KBr)/cmϪ1 2964, 1731, 1615, 1547, 1494, 1269, 1241, 721;
J. Chem. Soc., Perkin Trans. 1, 1999, 705–714
711

View Article Online
δH(500 MHz, CDCl3–CF3CO2D, 6:1) 1.42 (s, CH3), 7.70, 8.18
(AA‘XX’, C6H4), 9.29 (d, J 1.5, 2 H), 9.31 (t, J 1.6, 2 H), 9.40 (d,
J 1.6, 4 H), 9.51 (t, J 1.5, 2 H, C6H3); δC(125 MHz, CDCl3–
CF3CO2D, 6:1) 31.1 (CH3), 35.6 [C(CH3)3], 127.1, 128.0, 129.4,
131.3, 133.1 (arom. CH, 1 signal missing), 118.2, 124.9, 125.5,
125.9, 132.2, 158.8, 162.7, 164.0, 164.3, 167.0, 169.3 (ipso-C,
was obtained as a light brown solid after freeze-drying a
solution of 25 in aqueous HCl, mp 293–297 ЊC; δH(300 MHz,
D2O) 1.44 (s, CH3), 4.19 (s, CH2), 8.06 (t, J 1.7, 1 H), 8.22 (d,
J 1.7, 2 H, C6H3).
General procedure for the preparation of the complexes
C᎐O); m/z (MALDI-TOF) 1212 (80%, M ϩ Hϩ), 1235 (90,
᎐
Amidine base and carboxylic acid (1, 2 or 3 equiv., depend-
ing on the number of amidine groups) were dissolved in hot
ethanol (40 cm3 mmolϪ1), to which, if necessary (as in the case
of 14e–g), a certain amount of CHCl3 (5–10 cm3) was added
as co-solvent. After filtration of the hot solution and concen-
tration, the crude product was crystallised from the solvent
(mixture) indicated for each complex.
M ϩ Naϩ); Rf(CHCl3–MeOH, 9:1) 0.17.
N-[3-(2-Oxoazepan-1-yl)propyl]-3,5-bis[5-(4-tert-butylphenyl)-
1,3,4-oxadiazol-2-yl]benzamide 21
Amine 20 was prepared from DBU according to a literature
procedure17 and obtained as a colourless oil after chromato-
graphy (CHCl3–MeOH–conc. NH3, 9:1:1, then CHCl3–
MeOH, 4:1); δH(500 MHz, CDCl3) 1.62–1.78 (m, 8 H), 2.50
(br s, 2 H), 2.51–2.54 (m, 2 H), 2.77 (t, J 6.6, 2 H), 3.32–3.34 (m,
2 H), 3.47 (t, J 6.6, 2 H). A solution of 20 (48.6 mg, 0.286
mmol) and 15 (155 mg, 0.286 mmol) in dry NMP (4 cm3) was
stirred at 20 ЊC for 15 h. The solution was added dropwise to
water (40 cm3). The resulting light brown precipitate was
collected by suction filtration, dried and further purified by
chromatography (ethyl acetate) to yield 21 (52.1 mg, 27%) as a
colourless solid with identical analytical data as the by-product
reported in ref. 9.
Complex 10a. Yield: quant., oil; νmax (KBr)/cmϪ1 1651, 1574,
1402, 1343, 1288; δH(300 MHz, [2H6]DMSO, 100 ЊC) 1.07 (t,
J 7.1, N–CH2CH3), 1.84 (s, CH3CO2Ϫ), 3.13 (q, N–CH2), 7.30
(s, HA).
Complex 10b. Yield: 74% (from hexane), mp 104–108 ЊC
(Found: C, 69.3; H, 8.9; N, 8.2; C54H78N6O6 requires C, 71.5; H,
8.7; N, 9.3%); νmax (KBr)/cmϪ1 1653, 1589, 1541, 1385; δH(300
MHz, [2H6]DMSO, 100 ЊC) 1.10 (t, J 7.1, CH3), 1.30 (s,
N–CH2CH3), 3.17 (q, N–CH2), 7.39 and 7.82 (AA‘XX’, C6H4),
7.49 (s, HA); Mn (VPO, CHCl3, 34 ЊC) 1150 g molϪ1 (against
benzil as standard; C54H78N6O6 requires 907.25), 1210 g molϪ1
(against polystyrene 2000 as standard).
1,3-Bis(4,5-dihydro-1H-imidazol-2-yl)benzene 24 and conversion
to salts 28 and 29
Isophthalic acid (3.22 g, 19.4 mmol), ethylenediamine (4.25
cm3, 63.6 mmol), ethylenediamine dihydrochloride (8.46 g, 63.6
mmol), toluene-p-sulfonic acid (296 mg, 1.54 mmol) and ethyl-
ene glycol (15 cm3) were heated to reflux for 3 h. About half of
the ethylene glycol was then slowly removed by distillation. The
residual solution was concentrated to dryness at reduced pres-
sure (100 ЊC/0.1 mbar). The residue was dissolved in water (100
cm3)–conc. HCl (3 cm3). Addition of 50% aqueous NaOH (10
cm3) gave a brown precipitate which was purified by another
reprecipitation. Sublimation (230 ЊC/0.04 mbar) afforded yel-
low crystals (809 mg, 19%), mp 255–256 ЊC (lit.,33 244 ЊC; lit.,34
234–235 ЊC) [Found: C, 67.1; H, 6.7; N, 26.2; C12H14N4 (214.27)
requires C, 67.3; H, 6.6; N, 26.2%]; νmax (KBr)/cmϪ1 3157, 1615,
1568, 1492, 1467, 1267, 980, 700; δH(500 MHz, CD3OD) 3.77 (s,
N–CH2), 7.51 (t, J 7.9, 1 H), 7.90 (dd, J 7.9, 1.7, 2 H), 8.12 (t,
J 1.7, 1 H, C6H4); m/z (EI) 215, 214, 213 (Mϩ, 25, 91, 68%), 186,
185 (Mϩ Ϫ CH2CH2N, 36, 100), 156 (49), 78 (33). Hydro-
chloride 28 was obtained as a colourless solid after freeze-
drying a solution of 24 in aqueous HCl, mp 344–345 ЊC; δH(300
MHz, D2O) 4.17 (s, N–CH2), 7.89 (t, J 8.0, 1 H), 8.16 (dd, J 8.0,
1.7, 2 H), 8.22 (t, J 1.7, 1 H, C6H4); δC(75 MHz, D2O) 47.8
(N–CH2), 126.6 (ipso-C), 130.5, 133.6, 136.5 (arom. CH), 168.4
Complex 10c. Yield: 80% (from EtOH–MeOH–H2O), DSC:
K/184 (∆H 11 J gϪ1)/liquid crystalline/231 (∆H 20 J gϪ1)/I
(Found: C, 69.2; H, 6.4; N, 12.7; C114H126N18O12ؒ2H2O requires
C, 69.3; H, 6.6; N, 12.8%); νmax (KBr)/cmϪ1 3105, 2984, 2927,
1647, 1577, 1287, 1037, 726, 687, 614; δH(300 MHz,
[2H6]DMSO, 100 ЊC) 1.17 (t, J 7.3, N–CH2CH3), 1.36 [s,
C(CH3)3], 3.30 (br q, J 7.3, N–CH2CH3), 7.65 and 8.08
(AA‘XX’, C6H4), 7.86 (s, HA), 8.27 (t, J 1.6, 1 H), 8.75 (d, J 1.6,
2 H, C6H3CO2Ϫ); Mn (VPO, CHCl3, 34 ЊC) 2290 g molϪ1
(against benzil as standard; C114H126N18O12 requires 1940.38),
2410 g molϪ1 (against polystyrene 2000 as standard).
Complex 14a. Yield: quant. (complex 14a was obtained by
concentrating the mixture and drying under vacuum until
excess trifluoroacetic acid was removed); δH(500 MHz, CD3CN)
4.08 (s, N–CH2), 9.31 (s, HA), 11.25 (s, NH).
Complex 14b. Yield: 72% (from EtOH), decomp. >240 ЊC
(Found: C, 65.0; H, 5.6; N, 12.9; C36H36N6O6ؒH2O requires C,
64.9; H, 5.7; N, 12.6%); νmax (KBr)/cmϪ1 1637, 1600, 1575, 1385,
1292, 718, 671; δH(500 MHz, CDCl3–CD3CN, 6:1) 4.15 (s,
N–CH2), 7.39 (m, 6 H), 7.44 (m, 3 H), 8.07 (m, 6 H, C6H5),
10.06 (s, HA).
(C᎐N). Borate salt 29 was prepared in 91% yield from 28 analo-
᎐
gously to 12, mp 200–202 ЊC (decomp.) [Found: C, 46.8; H,
1.8; N, 3.0; C76H40B2F48N4 (1942.72) requires C, 47.0; H, 2.1;
N, 2.9%]; δH(500 MHz, CDCl3–CD3CN, 6:1) 4.11 (s, 9 H,
N–CH2), 7.54 (br s, 9 H), 7.67 [br m, 18 H, (CF3)2C6H3], 7.80 (t,
J 8.2, 1 H), 8.02 (dd, J 8.2, 1.9, 2 H), 8.41 (br s, 1 H, C6H4), 8.9
(br s, NH); νmax (KBr)/cmϪ1 3451, 1630, 1356, 1280, 1127.
Complex 14c. Yield: 38% (from MeOH), DSC: K/172 (∆H 25
J gϪ1)/liquid crystalline/250 (∆H 19 J gϪ1)/I (decomp.) [Found:
C, 55.1; H, 3.6; N, 10.1; C39H33F9N6O6 (852.71) requires C, 54.9;
H, 3.9; N, 9.9%]; νmax (KBr)/cmϪ1 1637, 1615, 1579, 1395, 1326,
1278, 1123, 691; δH(300 MHz, CDCl3) 4.19 (s, N–CH2), 7.52
(br m), 7.69 (br m), 8.26 (br m), 8.35 (br s, C6H4CF3), 10.06
(s, HA), 13.0 (br s, NH); Mn (VPO, CHCl3, 34 ЊC) 790 g molϪ1
(against benzil as standard), 835 g molϪ1 (against polystyrene
2000 as standard).
1,3-Bis(4,5-dihydro-1H-imidazol-2-yl)-5-tert-butylbenzene 25
Preparation analogous to 24 starting from 5-tert-butyliso-
phthalic acid. Sublimation (250 ЊC/0.03 mbar) afforded light
yellow crystals (63%), mp 206–208 ЊC [Found: C, 70.9; H, 8.4;
N, 20.6; C16H22N4 (270.38) requires C, 71.1; H, 8.2; N, 20.7%];
νmax (KBr)/cmϪ1 2959, 2865, 1620, 1574, 1496, 987; δH(500
MHz, CD3OD) 1.38 (s, CH3), 3.78 (s, N–CH2), 7.92 (t, J 1.6,
1 H), 8.01 (d, J 1.6, 2 H, C6H3); δC(125 MHz, CD3OD) 31.6
(CH3), 36.0 [C(CH3)3], 50.4 (N–CH2), 124.4, 128.1 (C6H3),
Complex 14d. Yield: 60% (from EtOH), DSC: K/105 (∆H 144
J gϪ1)/X/230 (∆H 4 J gϪ1)/liquid crystalline/248 (∆H 42 J gϪ1)/I
[Found: C, 70.5; H, 7.7; N, 10.5; C48H60N6O6 (817.04) requires
C, 70.6; H, 7.4; N, 10.3%]; νmax (KBr)/cmϪ1 2963, 1639, 1608,
1591, 1387; δH(500 MHz, C6D6) 1.28 (s, CH3), 3.48 (s, N–CH2),
7.49 and 8.58 (AA‘XX’, C6H4), 9.95 (s, HA), 12.1 (br s,
NH); δC(125 MHz, CDCl3) 31.3 (CH3), 34.8 [C(CH3)3], 45.4
131.2, 153.4, 167.3 (ipso-C, C᎐N); m/z (CI, NH3) 288
᎐
(M ϩ NH4ϩ, 7%), 271 (M ϩ Hϩ, 100). Hydrochloride 25ؒ2HCl
712
J. Chem. Soc., Perkin Trans. 1, 1999, 705–714

View Article Online
(N–CH2), 124.9, 129.3, 134.4 (arom. CH), 125.1, 133.9, 154.1,
Complex 23. Yield: 70% (from hexane), mp 149–150 ЊC
162.5, 173.4 (ipso-C, C᎐N, C᎐O); M (VPO, CHCl , 35 ЊC) 820
[Found: C, 75.3; H, 9.1; N, 7.1; C24H34N2O2 (382.54) requires C,
75.4; H, 9.0; N, 7.3%]; νmax (KBr)/cmϪ1 2967, 1643, 1383; δH(500
MHz, CDCl3) 1.16 (t, J 7.3, CH3), 1.30 (t, J 7.6, CH3), 1.34 [s,
C(CH3)3], 2.75 (q, J 7.6, Ar–CH2), 3.05 (br q, J 7.2, N–CH2),
7.23 and 7.39 (AA‘XX’, amidine-C6H4), 7.39 and 8.03
(AA‘XX’, C6H4CO2Ϫ), 12.97 (s, NH); δH(500 MHz, [2H6]-
DMSO) 1.09 (br s, CH3), 1.21 (t, J 7.6, CH3), 1.28 [s, C(CH3)3],
2.69 (q, J 7.6, Ar–CH2), 3.10 (br s, N–CH2), 7.42 (br s, amidine-
C6H4), 7.33 and 7.80 (AA‘XX’, C6H4CO2Ϫ), 12.9 (br s, NH).
᎐
᎐
n
3
g molϪ1 (against benzil as standard), 880 g molϪ1 (against poly-
styrene 2000 as standard); Mn (VPO, toluene, 50 ЊC) 760 g
molϪ1 (against benzil as standard).
Complex 14e. Yield: 87% (from EtOH–CHCl3); DSC: K/188
(∆H 2 J gϪ1)/X/220 (∆H 5 J gϪ1)/liquid crystalline/274 (∆H 4
J gϪ1)/I.
Complex 14f. Yield: 56% (from EtOH–CHCl3), DSC: 109/
glass transition/310 (∆H 8 J gϪ1)/liquid crystalline/323 (∆H 42 J
g
Complex 26. Yield: 32% (from CH3CN–MeOH), DSC: K/86
(∆H 6 J gϪ1)/X/136 (∆H 9 J gϪ1)/liquid crystalline/147 (∆H 38 J
g
H, 7.3; N, 9.4; C34H42N4O4ؒH2O requires C, 69.4; H, 7.5; N,
9.5%); νmax (KBr)/cmϪ1 3159, 2962, 1614, 1589, 1541, 1384;
δH(500 MHz, CDCl3) 1.33 (s, CH3), 4.01 (s, N–CH2), 6.99 (t,
J 8.0, 1 H), 7.40 and 7.96 (AA‘XX’, C6H4), 8.24 (dd, J 8.0, 1.6,
2 H), 9.30 (br t, 1 H, C6H4).
Ϫ1)/I (decomp.) [Found: C, 72.3; H, 7.3; N, 11.6; C132H156
-
Ϫ1)/liquid crystalline/179 ЊC (∆H 27 J gϪ1)/I (Found: C, 68.8;
N18O12 (2186.82) requires C, 72.5; H, 7.2; N, 11.5%]; νmax (KBr)/
cmϪ1 3060, 2950, 1630, 1575, 1490, 1370, 1115, 1010, 840, 780,
720; δH(300 MHz, CDCl2CDCl2, 12 mg/0.7 cm3, 120 ЊC) 1.37 (s,
CH3), 4.29 (s, N–CH2), 7.55 and 8.07 (AA‘XX’, C6H4), 8.80 (t,
J 1.7, 3 H) and 8.94 (d, J 1.7, 6 H, C6H3CO2Ϫ), 10.18 (s, HA);
δH(500 MHz, C6D6) 1.37 (s, CH3), 3.68 (s, N–CH2), 7.69 (br s,
6 H), 8.26 (br s, 12 H, C6H3), 8.69 (br s, 3 H), 9.48 (br s, 6 H,
C6H3CO2Ϫ ), 10.00 (s, C6H3), 13.3 (br s, 6 H, NH); δC(75 MHz,
CDCl3) 31.4 (CH3), 35.1 [C(CH3)3], 45.7 (CH2), 121.4, 126.3,
126.8, 130.7, 135.0 (arom. CH), 123.0, 125.0, 125.5, 135.0,
Complex 27. Yield: 61% (from CH3CN–MeOH), mp 209–
213 ЊC [Found: C, 72.8; H, 8.2; N, 8.9; C38H50N4O4 (626.84)
requires C, 72.8; H, 8.0; N, 8.9%]; νmax (KBr)/cmϪ1 2963m,
1635m, 1540s, 1385s; δH(500 MHz, CDCl3) 1.04 (s, 9 H, CH3),
1.33 (s, 18 H, CH3), 3.89 (s, N–CH2), 7.39 and 7.95 (AA‘XX’,
C6H4), 8.27 (d, J 1.5, 2 H), 9.03 (br s, 1 H, HA).
139.5, 152.0, 163.2, 163.7, 166.0, 171.2 (ipso-C, C᎐N, C᎐O).
᎐
᎐
Complex 14g. Yield: 65% (from EtOH–CHCl3), DSC: K/377
(∆H 36 J gϪ1)/liquid crystalline/391 (∆H 10 J gϪ1)/I (decomp.)
[Found: C, 69.7; H, 5.1; N, 15.3. C228H204N42O24 (3916.39)
requires C, 69.9; H, 5.3; N, 15.0%]; νmax (KBr)/cmϪ1 2962, 1615,
1541, 1494, 1390, 1268, 720; δH(500 MHz, CDCl3–CF3CO2D,
6:1) 1.42 (s, CH3), 4.25 (s, NCH2), 7.68, 8.18 (AA‘XX’, C6H4),
8.76 (br s, 3 H), 9.28 (d, J 1.4, 2 H), 9.30 (t, J 1.4, 2 H), 9.38 (d,
J 1.4, 4 H), 9.50 (t, J 1.4, 1 H, C6H3).
Calculation of association constants
Non-linear regression analysis [with Kaleidagraph ver. 3.09
(Synergy Software)] was used to derive association constants Ka
1
by the H NMR dilution method.29 In the case of 2:1 com-
plexes, the concentration C (= [bis-imidazoline] = [carboxylate]/
1
2) and the experimental H NMR chemical shifts δ as well as
the chemical shifts of the bis(imidazoline) host (δh) and the
complex (δc) were fitted to the equation.
Complex 19a. Yield: 30% (from toluene–hexane), hygro-
scopic solid that could be sublimed at 100 ЊC/0.02 mbar, mp
145–148 ЊC (Found: C, 72.0; H, 9.4; N, 8.7; C20H30N2O2
requires C, 72.7; H, 9.2; N, 8.5%) νmax (KBr)/cmϪ1 1650, 1590,
1544, 1385; δH(500 MHz, CDCl3) 1.32 (s, CH3), 1.63–1.82 (m,
6 H), 2.02 (quintet, J 5.8, 2 H), 2.95–2.98 (m, 2 H), 3.38–3.44
(m, 4 H), 3.54 (t, J 5.7, 2 H, CH2), 7.36 and 8.02 (AA‘XX’,
C6H3), 14.0 (br s, NH); δC(125 MHz, CDCl3) 19.8, 24.2, 27.1,
29.2, 32.1, 38.1, 48.6, 54.1 (CH2), 31.4 (CH3), 34.7 [C(CH3)3],
124.4, 129.2 (arom. CH), 135.4, 152.7, 166.0, 172.9 (ipso-C,
δc Ϫ δh
2C
1
1
δ = δh ϩ
ؒ
ͩ
3C ϩ
Ϫ
(3C ϩ
)2 Ϫ 8C
.
ͪ
√
Ka
Ka
Acknowledgements
Financial support by the Fonds der Chemischen Industrie, the
Deutsche Forschungsgemeinschaft, and generous support by
Professor Dr G. Wulff is gratefully acknowledged. I would also
like to thank Ms H. Fürtges and Ms S. Johann for their assist-
ance in the preparation of starting materials, Dr U. Matthiessen
for CI-MS and Dr H. Luftmann (University of Münster) for
MALDI-TOF-MS measurements.
C᎐N, C᎐O). Owing to the strong hygroscopicity of 19a, only
᎐
᎐
fragments of the complex were observed in the CI-MS.
Complex 19b. Yield: 42%, mp 258–261 ЊC (decomp.) (from
CH3CN) [Found: C, 71.0; H, 7.0; N, 12.5; C40H46N6O4 (674.84)
requires C, 71.2; H, 6.9; N, 12.5%]; νmax (KBr)/cmϪ1 2961, 1648,
1620, 1584, 1542, 1495, 1380, 1364, 1323, 722; δH(500 MHz,
CDCl3) 1.39 (s, CH3), 1.70–1.89 (m, 6 H), 2.10 (quintet, J 5.7,
2 H), 3.02–3.05 (m, 2 H), 3.48–3.53 (m, 4 H, CH2), 3.61 (t, J 5.7,
2 H, CH2), 7.56 and 8.12 (AA‘XX’, C6H4), 8.98 (t, J 1.7, 1 H),
9.02 (d, J 1.7, 2 H, C6H3), 13.6 (br s, NH); m/z (CI, NH3) 675
(DBUؒ1 ϩ Hϩ, 12%), 541, 540 (1 ϩ NH4ϩ, 17, 56), 524, 523
(1 ϩ Hϩ, 25, 92), 305 (2DBU ϩ Hϩ, 79), 153 (DBU ϩ Hϩ,
100); Rf(ethyl acetate) 0.07 (smearing).
References
1 For some recent examples, see: (a) C.-B. Xue, J. Roderick,
S. Jackson, M. Rafalski, A. Rockwell, S. Mousa, R. E. Olson
and W. F. DeGrado, Bioorg. Med. Chem., 1997, 5, 693; (b)
B. K. Blackburn, A. Lee, M. Baier, B. Kohl, A. G. Olivero,
R. Matomoros, K. D. Robarge and R. S. McDowell, J. Med. Chem.,
1997, 40, 717; (c) D. S. Jackson, S. A. Fraser, L.-M. Ni, C.-M. Kam,
U. Winkler, D. A. Johnson, C. J. Froelich, D. Hudig and J. C.
Powers, J. Med. Chem., 1998, 41, 2289.
2 (a) B. Gabriel, M. T. Stubbs, A. Bergner, J. Hauptmann, W. Bode,
J. Stürzebecher and L. Moroder, J. Med. Chem., 1998, 41, 4240;
(b) K. Okumura, T. Shimazaki, Y. Aoki, H. Yamashita, E. Tanaka,
S. Banba, K. Yazawa, K. Kibayashi and H. Banno, J. Med. Chem.,
1998, 41, 4036.
3 W. D. Wilson, F. A. Tanious, D. Ding, A. Kumar, D. W. Boykin,
P. Colson, C. Houssier and C. Bailly, J. Am. Chem. Soc., 1998, 120,
10310.
4 G. Wulff, T. Groß and R. Schönfeld, Angew. Chem., Int. Ed. Engl.,
1997, 36, 1963.
5 S. C. Zimmerman, Y. Wang, P. Bharathi and J. S. Moore, J. Am.
Chem. Soc., 1998, 120, 2172.
6 R. M. Eglen, A. L. Hudson, D. A. Kendall, D. J. Nutt, N. G.
Morgan, V. G. Wilson and M. P. Dillon, Trends Pharmacol. Sci.,
1998, 19, 381.
Complex 19c. Yield: 43%, mp 266–268 ЊC (decomp.) (from
CH3CN) (Found: C, 71.3; H, 8.0; N, 10.0; C48H62N6O4ؒH2O
requires C, 71.6; H, 8.0; N, 10.4%); νmax (KBr)/cmϪ1 2962, 1649,
1625, 1595, 1542, 1384, 1364, 1251, 1236, 793, 703; δH(300
MHz, CDCl3) 1.41 (s, CH3), 1.64–1.88 (m, 6 H), 2.10 (quintet,
J 5.7, 2 H), 3.02–3.05 (m, 2 H), 3.47–3.53 (m, 4 H, CH2), 3.60 (t,
J 5.7, 2 H, CH2), 7.63 (br s, 2 H), 8.02 (br s, 4 H, arom. H), 9.00
(br s, 1 H), 9.04 (br d, J 1.6, 2 H, arom. H), 13.6 (br s, NH); m/z
(CI, NH3) 787 (DBUؒ2 ϩ Hϩ, 11%), 652 (2 ϩ NH4ϩ, 26), 636,
635 (2 ϩ Hϩ, 22, 100), 305 (6), 153 (DBU ϩ Hϩ, 4); Rf(ethyl
acetate) 0.38 (smearing).
J. Chem. Soc., Perkin Trans. 1, 1999, 705–714
713

View Article Online
7 A. Metzger, V. M. Lynch and E. V. Anslyn, Angew. Chem., Int. Ed.
Engl., 1997, 36, 862.
23 (a) C. L. Perrin, D. A. Schiraldi and G. M. L. Arrhenius, J. Am.
Chem. Soc., 1982, 104, 196; (b) R. W. Taft, E. D. Raczynska, P.-C.
Maria, I. Leito, W. Lewandowski, R. Kurg, J.-F. Gal, M. Decouzon
and F. Anvia, Fresenius Z. Anal. Chem., 1996, 355, 412.
24 J. S. McKennis and P. A. S. Smith, J. Org. Chem., 1972, 37,
4173.
25 G. S. Hammond and R. C. Neuman, J. Phys. Chem., 1963, 67, 1655.
26 B. Tinant, J. Dupont-Fenfau, J.-P. Declercq, J. Podlaha and
O. Exner, Collect. Czech. Chem. Commun., 1989, 54, 3245.
27 J. O. Trent, G. R. Clark, A. Kumar, W. D. Wilson, D. W. Boykin,
J. E. Hall, R. R. Tidwell, B. L. Blagburn and S. Neidle, J. Med.
Chem., 1996, 39, 4554.
28 (a) A. R. Katritzky, D. C. Oniciu, I. Ghiviriga and R. A. Barcock,
J. Chem. Soc., Perkin Trans. 2, 1995, 785; (b) A. R. Katritzky,
I. Ghiviriga, P. J. Steel and D. C. Oniciu, J. Chem. Soc., Perkin Trans.
2, 1996, 443.
29 (a) C. S. Wilcox, in Frontiers in Supramolecular Chemistry and
Photochemistry, eds. H.-J. Schneider, H. Dürr, VCH, Weinheim,
1991, p. 123; (b) R. S. Macomber, J. Chem. Educ., 1992, 69, 375.
30 J. Bettenhausen and P. Strohriegl, Macromol. Rapid Commun., 1996,
17, 623.
8 (a) M. W. Hosseini, R. Ruppert, P. Schaeffer, A. De Cian, N.
Kyritsakas and J. Fischer, J. Chem. Soc., Chem. Commun., 1994,
2135; (b) O. Félix, M. W. Hosseini, A. De Cian and J. Fischer,
Angew. Chem., Int. Ed. Engl., 1997, 36, 102; (c) O. Félix, M. W.
Hosseini, A. De Cian and J. Fischer, Tetrahedron Lett., 1997, 38,
1755; (d) O. Félix, M. W. Hosseini, A. De Cian and J. Fischer,
Tetrahedron Lett., 1997, 38, 1933.
9 (a) A. Kraft, Chem. Commun., 1996, 77; (b) A. Kraft, Liebigs Ann./
Recueil, 1997, 1463.
10 G. M. Barrow and E. A. Yerger, J. Am. Chem. Soc., 1954, 76, 5211.
11 T. Kato, in Handbook of Liquid Crystals, Vol. 2B, Low Molecular
Weight Liquid Crystals II, eds. D. Demus, J. W. Goodby, G. W. Gray,
H. W. Spiess and V. Vill, Wiley-VCH, Weinheim, 1998, p. 969.
12 J. A. Roberts, J. P. Kirby and D. G. Nocera, J. Am. Chem. Soc., 1995,
117, 8051.
13 O. Félix, M. W. Hosseini, A. De Cian and J. Fischer, New J. Chem.,
1997, 21, 285.
14 A. Kraft and R. Fröhlich, Chem. Commun., 1998, 1085.
15 V. B. Piskov, V. P. Kasperovich and L. M. Yakovleva, Chem.
Heterocycl. Compd., 1976, 12, 917.
16 N. Ono, T. Yamada, T. Saito, K. Tanaka and A. Kaji, Bull. Chem.
Soc. Jpn., 1978, 51, 2401.
17 C. Heidelberger, A. Guggisberg, E. Stephanou and M. Hesse, Helv.
Chim. Acta, 1981, 64, 399.
18 A. Albert, R. Goldacre and J. Phillips, J. Chem. Soc., 1948, 2240.
19 I. Hermecz, Adv. Heterocycl. Chem., 1987, 42, 83.
20 A. Kraft and F. Osterod, J. Chem. Soc., Perkin Trans. 1, 1998, 1019.
21 G. Wulff and R. Schönfeld, Adv. Mater., 1998, 10, 957.
22 (a) P. Job, Ann. Chim. 1928, 9, 113; (b) M. T. Blanda, J. H. Horner
and M. Newcomb, J. Org. Chem., 1989, 54, 4626.
31 S. Ahrland, J. Chatt, N. R. Davies and A. A. Williams, J. Chem.
Soc., 1958, 276.
32 H. Nishida, N. Takada, M. Yoshimura, T. Sonoda and H.
Kobayashi, Bull. Chem. Soc. Jpn. 1984, 57, 2600.
33 G. Levesque, J.-C. Gressier and M. Proust, Synthesis, 1981, 963.
34 G. Kränzlein and H. Bastian (IG Farben), D. R. P. 695473, 1937
(Chem. Abstr., 1941, 35, 141b).
Paper 8/09190K
714
J. Chem. Soc., Perkin Trans. 1, 1999, 705–714