C3-Symmetric, amino acid based organogelators and thickeners: a systematic study of structure-property relations

Article abstract of DOI:10.1016/j.tet.2007.02.066

A class of C₃-symmetric amino acid based organogelators and thickeners featuring a rigid core have been developed. Structural variation yielded a number of compounds, the aggregation behaviour and resulting aggregates and gels of which were studied by FTIR spectroscopy, dropping ball measurements, differential scanning calorimetry and transmission electron microscopy. These studies showed that the nature of the core unit, the type of hydrogen-bonding units and the applied amino acids have a strong influence on the interactions, resulting in large differences in aggregation properties, thermal stability and morphology between the various compounds. The results provide a basis for a better understanding of the relation between aggregate/gel properties and molecular structure. The structural variation available for these compounds allows fine-tuning of the gelators with respect to aggregation behaviour and gel properties.

Full text of DOI:10.1016/j.tet.2007.02.066

Tetrahedron 63 (2007) 7285–7301
C3-Symmetric, amino acid based organogelators and thickeners:
a systematic study of structure–property relations
c
a,
Maaike de Loos,a,b Jan H. van Esch,a, Richard M. Kellogg and Ben L. Feringa
*
*
aDepartment of Organic and Molecular Inorganic Chemistry, Stratingh Institute for Chemistry, University of Groningen,
Nijenborgh 4, 9747 AG Groningen, The Netherlands
bBioMaDe Technology Foundation, Nijenborgh 4, 9747 AG Groningen, The Netherlands
cSyncom BV, Kadijk 3, 9747 AT Groningen, The Netherlands
Received 16 October 2006; revised 25 January 2007; accepted 9 February 2007
Available online 20 February 2007
Abstract—A class of C3-symmetric amino acid based organogelators and thickeners featuring a rigid core have been developed. Structural
variation yielded a number of compounds, the aggregation behaviour and resulting aggregates and gels of which were studied by FTIR spec-
troscopy, dropping ball measurements, differential scanning calorimetry and transmission electron microscopy. These studies showed that the
nature of the core unit, the type of hydrogen-bonding units and the applied amino acids have a strong influence on the interactions, resulting in
large differences in aggregation properties, thermal stability and morphology between the various compounds. The results provide a basis for
a better understanding of the relation between aggregate/gel properties and molecular structure. The structural variation available for these
compounds allows fine-tuning of the gelators with respect to aggregation behaviour and gel properties.
Ó 2007 Published by Elsevier Ltd.
1. Introduction
structural variation.6 LMW hydrogelators belonging to this
class of compounds have already been reported.6 Herein we
report the design and synthesis of C3-symmetric organogela-
tors, their aggregation behaviour and the properties of the
aggregates formed, with emphasis on the effect of the inter-
action strength (determined by the various structural
parameters) on the gelation ability.
Low molecular weight gelator compounds (LMWGs) which
gel solvents have attracted considerable attention, owing to
their striking self-assembling properties, their responsive
behaviour and their applicability in, e.g., cosmetics, food
or pharmaceutics.1 The quest for new LMWGs inspired sev-
eral research groups, including those of Shinkai,2 Terech,1a
Weiss,1c Hanabusa,3 Menger4 and our group,1b,5 to perform
systematic structural studies aimed at the development of
criteria for the design of LMWGs. As a result, the field of
LMWGs has progressed from a stage in which new LMWGs
were discovered by chance, to a point at which the rational
design of LMWGs has become feasible.1 However, whereas
the development of new LMWGs has come within reach, the
tuning of properties like gel stability (mechanical or ther-
mal), the scope of gelated solvents and morphology of the
gels are still less readily achieved. Obviously, these proper-
ties depend on the structures of the LMWGs, although the
precise relations between various structural parameters and
the gel properties under investigation are still not well under-
stood. In an effort to gain a better understanding of these
relations, we developed a class of C3-symmetric amino acid
based LMW gelators that allowed a systematic and broad
2. Results and discussion
2.1. Design
A schematic representation of the structure of the C3-sym-
metric compounds is shown in Figure 1a. The C3-symmetric
compounds are built up in sections from the centre towards
the periphery. The first central section is formed by a C3-
symmetric rigid core. The C3-symmetry of this core reduces
the occurrence of polymorphism and allows a better control
over the intermolecular and fiber/solvent interactions. The
rigid geometry of the core restricts the conformational free-
dom of the molecules1b,5c and provides an ideal platform to
study the influence of the intermolecular interactions on the
gelation ability.
The second section consists of hydrogen-bonding units, in
general ureas or amides. The application of different types
of these units enables variation in the interaction strength.
The units have to be connected to the core in an orientation
Keywords: Amino acids; Gel properties; Gels; Self-assembly; Organogela-
tors.
* Corresponding authors. Tel.: +31 50 3634235; fax: +31 50 3634296;
e-mail addresses: j.h.van.esch@rug.nl; b.l.feringa@rug.nl
0040–4020/$ - see front matter Ó 2007 Published by Elsevier Ltd.
doi:10.1016/j.tet.2007.02.066

7286
M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
OAlkyl
XAlkyl
R
U
Rigid C3-symmetric core
R
R
O
O
O
O
Hydrogen bonding units
Peripheral substituents
U
R
U
U
R
U
OAlkyl
XAlkyl
U
U
R
R
R
R
U
O
U
O
a)
b)
U
U
R
R
AlkylO
AlkylX
X = O; NH
R
R
R
R
U
U
R
R
U
U
U
U
R
R
U
U
R
U = urea; amide
U
U
Figure 2. General structures of the different classes of compounds.
R
c)
number of hydrogen bond donors available for self-assem-
bly: tris-urea and tris-amide derivatives with a-amino acid
ester chains. For the cyclohexane core compounds a third
type differing in the number of hydrogen bond donors
and acceptors was envisioned: tris-amide derivatives with
a-amino acid amide chains. All types of compounds have
been prepared in a convergent manner by coupling an amino
acid derivative at its N-terminus to the core molecule pre-
pared beforehand.
Figure 1. (a) Schematic design of the C3-symmetric compounds. (b) Repre-
sentation of the hydrogen-bonded stack formed by cyclohexane-based com-
pounds. (c) Representation of the hydrogen-bonded stack of benzene based
compounds.
that enables simultaneous, self-complementary, intermolec-
ular hydrogen bonding of the three units, allowing one-
dimensional aggregation and hence gelation to occur. An
excellent core and hydrogen-bonding-unit candidate that
fulfils these criteria is the cis,cis-1,3,5-cyclohexanetricar-
boxamide scaffold.7,8 In this structure (Fig. 1b), the amide
groups are orientated perpendicularly to the mean plane of
the molecule and exhibit a parallel alignment with respect
to each other.6–8 This provides the strong uni-axial inter-
molecular interactions affording one-dimensional columnar
self-assembly. A comparable scaffold is the 1,3,5-benzene-
tricarboxyamide core, which has also been reported to have
a hydrogen-bonded columnar type of packing (Fig. 1c).9–12
The capacity of both scaffolds to form one-dimensional
stacks was previously exploited in the preparation of simple
alkyl derived LMW gelators or thickeners.8,10–12 These
studies already indicate that some dependence of the gela-
tion ability of the C3-symmetric 1,3,5-cyclohexane and
1,3,5-benzene based compounds on the strength of the
intermolecular interactions (determined by the core, the
hydrogen-bonding units and the peripheral substituents) is
present. This makes these types of compounds attractive
candidates to study the influence of various structural param-
eters on the gelation ability and as such to acquire a better
understanding of their effect and of the balance between
the various components in the LMWGs.
The effect of the peripheral substituents was investigated
with the a-amino acid esters and amides presented in
Figure 3. Of these compounds, the esters 1–3 were prepared
under Dean–Stark conditions by reaction of the a-amino
acid with 1 equiv of 1-octanol in the presence of 1.1 equiv
of p-toluenesulfonic acid monohydrate,13 whereas the amides
4–6 were prepared according to standard peptide synthetic
methods. All a-amino acid derivatives were obtained in
high yields and used without further purification.
The 1,3,5-benzene tris-urea derivatives were prepared fol-
lowing standard procedures by reaction of an amino acid
ester with 1,3,5-benzenetriisocyanate 8 (Scheme 1).14
Compound 8 was prepared from commercially available
1,3,5-benzenetricarboxylic acid via 1,3,5-benzenetricar-
bonyl trichloride 7. Compound 7 was obtained as a slightly
yellowish oil, which crystallised upon standing at 4 ꢀC and
was stable for months. This allowed synthesis of 7 on a large
scale made it possible for us to keep it in stock. Reaction of
acid chloride 7 with sodium azide afforded 1,3,5-benzenetri-
carbonyl triazide as a white precipitate. The triazide was not
isolated because it easily detonates,15 but was immediately
taken up in toluene and separated from the reaction mixture.
The toluene solution was gradually heated to 100 ꢀC and
stirred till gas evolution ceased, yielding 1,3,5-benzenetri-
isocyanate 8 in situ via a Curtius rearrangement.16 Sub-
sequently, 8 was allowed to react with the a-amino acid
derivatives 1 and 2 to afford the 1,3,5-benzene tris-urea com-
pounds 9 and 10. Owing to their good solubility in organic
solvents, purification of these compounds could be achieved
by column chromatography, yielding 9 and 10 in low yields
as analytically pure, sticky solids.
To realise this goal, a more extensive and thorough structural
variation is required. Therefore, in our design the final pe-
ripheral section (Fig. 1a) is based on a-amino acid building
blocks. These offer easy access to both urea and amide hy-
drogen-bonding units via their N-terminus and afford a large
structural variety, providing the possibility to tune both the
intermolecular and the fibre/solvent interactions. Addition-
ally, derivatisation of the peripheral C-terminus offers fur-
ther opportunities to influence the fibre/solvent interactions
and can determine the solvent compatibility.1b,f To achieve
the gelation of organic solvents, long alkyl chains are com-
monly introduced.1
1: X = O; R = CH2Ph
2: X = O; R = CH2CH(CH3)2
3: X = O; R = H
4: X = NH; R = CH2Ph
5: X = NH; R = CH2CH(CH3)2
6: X = NH; R = H
O
H2N
XC8H17
2.2. Synthesis
R
LMW compounds were prepared based both on a benzene
core and a cyclohexane core (Fig. 2). For both classes of
compounds, two types are envisioned, which differ in the
Figure 3. a-Amino acid esters and amides applied in the synthesis of the
LMWGs.

M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
7287
OC8H17
R
O
O
O
O
O
OCN
NCO
HO
OH
Cl
Cl
N
O
b, c
a
H
H
N
H
N
O
d
'in situ'
N
O
NCO
H
OC8H17
HO
Cl
O
O
O
R
8
7
N
H
H
R
N
H
17C8O
O
9: R = CH2Ph
10: R = CH2CH(CH3)2
H
e
O
H
17C8O
R
N
O
R
OC8H17
O
O
N
H
O
H
N
O
11: R = CH2Ph
12: R = CH2CH(CH3)2
13: R = H
R
H17C8O
Scheme 1. Synthesis of the 1,3,5-benzene tris-urea and tris-amide derivatives. Reagents and conditions: (a) SOCl2, DMF (cat.), D, 3 h, 100% conversion; (b)
NaN3, water/THF, 0 ꢀC, 2 h; (c) toluene, D; (d) 3.3 equiv 1 or 2, toluene, rt, 18 h, 20–40%; (e) 3 equiv 1, 2 or 3, Et3N, CH2Cl2, 50–60%.
The 1,3,5-benzene tris-amide derivatives were synthesised
following standard amide synthesis procedures from the acid
chloride 7 and the a-amino acid derivatives 1–3 (Scheme 1).
After purification by column chromatography compounds
11–13 were obtained as analytically pure, sticky solids.
Compared to the tris-urea compounds 9 and 10, the tris-
amide compounds 11–13 have the advantage that their syn-
thesis is much more facile and higher yields are obtained in
the final coupling step.
available cis,cis-1,3,5-cyclohexane tricarboxylic acid was
allowed to react with diphenyl phosphoryl azide (DPPA) to
give the acyl azide. Subsequent heating of the mixture led
to a in situ Curtius rearrangement and on quenching with
benzyl alcohol the tris-benzyl carbamate 14 was obtained
as a white powder. Treatment of compound 14 with a strong
acid (33% HBr/HOAc) yielded the HBr-salt of cis,cis-1,3,5-
triaminocyclohexane, from which amine 15 was liberated as
a pale yellow solid, using alkaline conditions.
Because of the low yields in which the benzene tris-urea de-
rivatives were obtained, the cis,cis-1,3,5-cyclohexane tris-
urea derivatives were synthesised using a different method
for the preparation of isocyanates. This method involves
the in situ conversion of the amino group of a-amino acid
esters into an isocyanate without epimerisation by using
(BOC)2O and DMAP,17 followed by reaction with cis,cis-
1,3,5-triaminocyclohexane.
Subsequently, the a-amino acid esters 1 and 2 were con-
verted in situ to the isocyanate17 and allowed to react with
0.33 equiv of 15 (Scheme 2). After evaporation of the sol-
vent, the remaining materials were heated to reflux in etha-
nol and after cooling the obtained solid or gel was filtered
off to yield cis,cis-1,3,5-cyclohexane tris-urea derivatives
16 and 17, respectively, as analytically pure white powders.
Unfortunately, the yield of the coupling reaction did not im-
prove when compared to the synthesis of the benzene tris-
urea compounds, although the purification of the products
is much easier.
First, cis,cis-1,3,5-triaminocyclohexane was prepared fol-
lowing the procedure described in Scheme 2. Commercially
OBzl
O
NH
CO2H
NH2
a, b, c
59%
d, e
O
86%
OBzl
HN
N
H
H2N
NH2
CO2H
HO2C
BzlO
O
14
15
OC8H17
R
N
O
O
N
H
O
H
H
N
H
O
N
O
OC8H17
f, g
~40%
H2N
O
R
OC8H17
R
N
H
R
N
O
1: R = CH2Ph
2: R = CH2CH(CH3)2
16: R = CH2Ph
17: R = CH2CH(CH3)2
H
H17C8O
Scheme 2. Synthesis of the cis,cis-1,3,5-cyclohexane tris-urea derivatives. Reagents and conditions: (a) DPPA, Et3N, benzene, rt, 30 min; (b) D, 30 min;
(c) benzyl alcohol, D, 18 h; (d) HBr/HOAc; (e) aq NaOH; (f) (BOC)2O, DMAP, CH2Cl2, rt, 30 min; (g) 0.33 equiv 15, CH2Cl2, 40 ꢀC, 48 h.

7288
M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
The cis,cis-1,3,5-cyclohexane tris-amides derived by reac-
tion with a-amino acid esters or amides were prepared
following standard procedures from the a-amino acid deriv-
atives 1–6 and cis,cis-1,3,5-cyclohexane tricarboxylic acid
(Scheme 3). First commercially available cis,cis-1,3,5-
cyclohexane tricarboxylic acid was allowed to react with
thionyl chloride to yield the corresponding trichloride 18
as a slightly yellowish oil, which crystallised at 4 ꢀC. In con-
trast to the synthesis of 1,3,5-benzenetricarbonyl trichloride
7 a catalytic amount of DMF was not added. Analogously to
compound 7, cyclohexanetricarbonyl trichloride 18 could be
stored for months at 4 ꢀC.
stirred in (hot) ethanol to remove the soluble impurities.
After purification in this fashion, the desired products were
obtained in high yields as analytically pure, white powders.
2.3. Aggregation properties of the benzene core
compounds
The aggregation ability of the benzene core compounds 9–13
was investigated in the usual way by first heating a weighed
amount of the solid in 0.5 or 1.0 mL solvent followed
by cooling to room temperature. The results for several sol-
vents are shown in Table 1; the solvents are ranked in order
of increasing polarity according to their ET(30)-values.
O
O
O
O
HO
OH
Cl
Cl
In contrast to many bis- and tris-urea or -amide com-
pounds3a,b,5 it was found that in most solvents, compounds
9–13 are readily soluble, even without heating. Just occa-
sionally a viscous solution is formed, indicative of the
formation of higher aggregates. Generally, however, these
viscous solutions are formed at higher concentrations and
for none of the solid-solvent combinations is the formation
of a gel observed. The hydrogen-bonding unit appears to
have only a minor influence on the aggregation ability of
these compounds. The tris-urea derivatives 9 and 10 exhibit
more or less the same aggregation capacities as the tris-
amide derivatives 11–13. Almost the same applies for the
influence of the R-group. Only a slight difference is observed
between the benzyl (9 and 11), the iso-butyl (10 and 12) and
the hydrogen (13) compounds, namely somewhat poorer ag-
gregation behaviour for the iso-butyl compounds. This can
be rationalised by the fact that the benzyl groups are able
to form additional intermolecular p–p stacking interactions
to favour aggregation. Furthermore, the iso-butyl group is
sterically more demanding than the benzyl group,18 which
might hinder self-assembly.
a
HO
Cl
O
c
O
18
b
C8H17
N
O
C8H17
H
O
O
H
R
N
O
R
H
N
H
N
R
N
O
R
O
O
N
H
C8H17
OC8H17
O
O
O
N
H
O
H
N
N
H
O
H
O
O
R
R
H17C8
H17C8
22: R = CH2Ph
23: R = CH2CH(CH3)2
24: R = H
19: R = CH2Ph
20: R = CH2CH(CH3)2
21: R = H
Apparently, the C3-symmetric benzene based compounds
exhibit very poor aggregation properties, especially com-
pared to bis- and tris-urea compounds developed in our
group.5 Also compared to trialkyl-1,3,5-benzene tris-urea12
and -tris-amide10 derivatives the aggregation properties have
diminished. Although comparison with the trialkyl-tris-
urea derivatives is difficult due to the limited data available,
these compounds were at least able to gelate a solvent.12 The
trialkyl-tris-amide derivatives in turn were able to thicken a
Scheme 3. Synthesis of the cis,cis-1,3,5-cyclohexane tris-amide a-amino
acid ester and tris-amide a-amino acid amide derivatives. Reagents and con-
ditions: (a) SOCl2, D, 20 h, 100% conversion; (b) 3 equiv 1, 2 or 3, Et3N,
CH2Cl2, 25–88%; (c) 3 equiv 4, 5 or 6, Et3N, CH2Cl2, 60–90%.
Reaction of 18 with the amino acid esters 1–3 afforded the
cis,cis-1,3,5-cyclohexane tris-amides 19–21. These com-
pounds were readily soluble in the reaction medium, thus
the Et3N$HCl salts formed upon reaction could be removed
by extraction of the reaction mixture. Compounds 19 and 20
were purified by column chromatography and obtained as
colourless sticky solids in moderate yields. Compound 21
did not require further purification and was obtained as
a white powder in high yield. The compounds were found
to be pure within the limits of 1H NMR and 13C NMR detec-
tion and further characterised by elemental analyses. The
synthesis of these compounds is simple and straightforward
and the desired products are obtained in only three steps
starting from the commercially available compounds.
Table 1. Aggregation properties of the benzene core compounds 9–13a
Solvent
Tris-urea
Tris-amide
9
10
11
12
13
n-Hexane
vs (20)
gp
vs (30)
s
s
s
s
vs (20)
s
s
s
vs (20)
vs (160)
s
s
s
s
—
s
vs (10)
p
vs (10)
s
s
s
s
s
s
s
s
s
s
s
s
c
p
s
s
s
s
s
s
s
s
n-Hexadecane
Cyclohexane
p-Xylene
Tetraline
n-Butyl acetate
Cyclohexanone
Olive oil
1,2-Dichloroethane
1-Octanol
s
s
vs (5)
s
s
The cis,cis-1,3,5-cyclohexane tris-amide a-amino acid
amide compounds 22–24 were obtained by reaction of the
a-amino acid amides 4–6 with compound 18. In contrast to
the synthesis of compounds 19–21, the products were not
readily soluble in the reaction medium and a viscous mixture
was formed. Therefore, the solvents were evaporated from
the reaction mixture and the remaining materials were
s
a
Abbreviations: vs: viscous solution (digits: minimal concentration at
which thickening of the solvent was observed by eye (mg mLꢁ1)); gp:
gel-like precipitate; s: soluble at room temperature (solubility
>20 mg mLꢁ1); p: precipitate; c: crystals.

M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
7289
broad range of solvents.10 The diminished aggregation
behaviour of compounds 9–12 is presumably due to the
sterically demanding R-groups. The behaviour of the unsub-
stituted compound 13 (R¼H) compared to tridodecyl-1,3,5-
benzene tris-amide10 suggests that the introduced ester
groups also have a contribution.
with results obtained for trialkyl-tricarboxamide deriva-
tives,8,10,11 replacement of the benzene core by the cyclo-
hexane core resulted in a dramatic increase in the gelation
and aggregation abilities of the compounds. All cyclohexane
core compounds form a gel with one or more solvents, with
minimum gelation concentrations as low as 1 mg mLꢁ1. Re-
markably, protic solvents like ethanol were also gelated by
several of the compounds. The increased gelation ability
compared to the benzene core compounds is presumably re-
lated to a more favourable orientation of the hydrogen-bond-
ing units with respect to the mean plane of the core. For the
cyclohexane core, published crystal structures show that
the amide (or urea groups) is orientated perpendicularly to
the core, enabling one-dimensional hydrogen bond forma-
tion (Fig. 4A).6c,7 For the benzene core, however, conjuga-
tion causes the hydrogen-bonding units to have a larger
2.4. Aggregation properties of the cyclohexane core
compounds
In contrast to the benzene core compounds, the cyclohexane
core compounds were sparingly soluble in most of the com-
mon organic solvents. Their aggregation properties were
investigated in the same way as described above for the benz-
ene core compounds and the results are shown in Table 2.
At first glance it can be concluded that, in agreement
Table 2. Gelation and aggregation properties of the cyclohexane core compoundsa
Solvent
Tris-urea
17
Tris-amide
Tris-amide-amide
16
19
20
21
22
23
24
n-Hexane
p
p
p
5
p
10
vs (5)
p
5
vs (20)
i
gp
5
2
10
vvs (10)
vvs (10)
10
20
20
vvs (10)
vvs (5)
10
vvs (5)
vvs (5)
s
vs (20)
s
pg
p
p
i
vs (5)
pg
20
10
1
5
5
p
1
i
p
i
i
i
i
i
i
pg
5
gp
2
P
20
2
n-Hexadecane
Cyclohexane
p-Xylene
2
5
5
i
5
pg
p
10
i
10
20
10
vvs
p
20
5
i
5
10
p
5
p
gp
p
i
Tetraline
n-Butyl acetate
Cyclohexanone
Olive oil
1,2-Dichloroethane
1-Octanol
Ethanol
20
2
5
20
s
5
2
Abbreviations: see Table 1; digits: minimal gelation concentration in mg mLꢁ1; vvs: viscous solution exhibiting almost no flow; pg: partially gelated.
a
Figure 4. (A) Representation of the translational hydrogen-bonded stack formed by cyclohexane tricarboxamides (from crystal structure).7 (B) Representation
of the triple-helical hydrogen-bonded stack formed by a benzene tricarboxamide derivative (from crystal structure).9 (C) Model of the stacks formed by the
methyl ester derivative of 19.

7290
M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
preference to be in the plane of the core9,10 and the units have
to be rotated out of this plane to enable hydrogen bond forma-
tion. Crystal structures of benzene tricarboxamides show that
this is reflected in a partially tilted (Fig. 4B)9 or parallel10
conformation of the amides with respect to the benzene core.
It can be concluded that in general the tris-amide compounds
19–21 exhibit the best gelation and aggregation abilities.
For the leucyl and phenylalanyl derivatives, introduction of
additional hydrogen-bonding donors and/or acceptors by
substituting the amides with urea moieties (16 and 17) or
the esters with amides (22–24) results in a reduced gelation
ability due to a lower solubility. The lowering in solubility
is most likely caused by the stronger intermolecular
interactions. Together with the reduced gelation ability, the
minimal gelation concentrations decreased due to these
stronger interactions. Only for the L-leucyl derivatives did
the introduction of urea or additional amides result in a pos-
itive effect on the gelation ability.
The type and number of hydrogen-bonding units and the
nature of the R-group have a significant but complicated
influence on the gelation and aggregation ability of the com-
pounds. First we will consider the influence of the hydrogen-
bonding units by comparing 16 with 19, 22 (R¼CH2Ph) and
17 with 20 and 23 (R¼CH2CH(CH3)2) and 21 with 24
(R¼H).
For compounds 16, 19 and 22 (R¼CH2Ph) tris-amide 19 ex-
hibited the best gelation and thickening behaviour and
formed clear gels or viscous solutions with all solvents in-
vestigated. The minimal gelation concentrations are not
very low except for n-hexane and the alcohols. Substitution
of the amide groups with urea moieties (yielding 16) in-
creases the number of hydrogen bond donors, and leads to
a diminished solubility of the compound. This resulted in
a decrease in the range of gelated solvents, especially for
the more apolar solvents. Substitution of the ester groups
with amides (22), thereby increasing the number of both
hydrogen bond donor and acceptor groups, also resulted in
a decrease of the scope of gelated solvents. Furthermore,
this compound is insoluble or precipitates in many solvents.
However, for the solvents that were gelated a decrease of the
minimal gelation concentrations was observed compared to
19. Apparently, upon addition of extra amide groups and
thus introduction of additional intermolecular interactions,
gelation becomes more efficient but at the same time the
lower solubility diminishes the scope of solvents gelated.
The influence of the R-group on the gelation ability was as-
sessed by comparing 16 with 17 (urea), 19 with 20 and 21
(amide) and 22 with 23 and 24 (amide–amide). From the po-
sition of the R-groups in the narrow interior of the stacks, as
revealed by a preliminary model of the methyl ester deriva-
tive of 19 (Fig. 4C), it can be expected that the steric require-
ments of the R-groups have a significant influence on the
packing and thus gelation ability. In this model, for instance,
it seems that the benzyl groups force the stacked molecules
to rotate with respect to each other, contrary to the transla-
tional aggregate in Figure 4A.
Both the tris-urea compounds 16 and 17 are able to form a
gel or viscous solution with several of the solvents tested.
Compound 17 (R¼CH2CH(CH3)2) seems to be slightly more
efficient than 16 (R¼CH2Ph), judged from the fact that the
minimal gelation concentrations are lower. Furthermore,
for compound 17 precipitation is more often observed, espe-
cially for the apolar aliphatic solvents. Apparently, for a ge-
lator containing several strong hydrogen bond forming units,
introduction of some steric hindrance (iso-butyl groups) and
thus slightly weakening the intermolecular interactions can
have a positive effect on the gelation ability.
For the L-leucyl derivatives 17, 20 and 23 the tris-amide 20
displayed the poorest gelation ability and formed a gel
only with n-hexadecane. However, a viscous solution was
formed with several of the other solvents, indicating that ag-
gregation was still taking place. Substitution of the amides
by urea groups (17) increased the gelation ability and a gel
was formed with several of the solvents tested. Conversion
of the ester into an amide (23) also increased the gelation
scope of the compound and lowered the minimal gelation
concentrations. However, the compound was insoluble in
half the solvents tested. These differences with the phenyl-
alanyl compounds can most likely be attributed to the loss
of p–p stacking and the steric hindrance caused by the iso-
butyl group,18 resulting in less efficient stacking of the mole-
cules and subsequent lower gelation abilities for compound
20 compared to 19. Introduction of additional hydrogen-
bonding donors and/or acceptors compensates in these cases
for the unfavourable steric effects, resulting in good gelation
abilities for compounds 17 and 23.
For the tris-amides 19–21 the influence of the R-group
is much more pronounced. Tris-amide 19 (R¼CH2Ph) was
the most efficient gelator and formed a gel or highly viscous
solution in all the solvents tested. Introduction of the steri-
cally more demanding iso-butyl group18 (20) led to a de-
crease in the gelation ability of the compound compared to
19, possibly due to steric hindrance. Apparently, the amide
hydrogen bonds are not strong enough to compensate for
these unfavourable effects. Furthermore, for compound 19
p–p stacking between the phenyl groups might contribute
to the intermolecular interactions and subsequently the gela-
tion ability.19 The compound with the least steric hindrance
(21: R¼H) was found to gelate a range of solvents compara-
ble to compound 19. Additionally, the minimal gelation con-
centrations had decreased. Apparently, the absence of steric
hindrance improves the packing in the stacks and thus the
intermolecular interactions, resulting in increased gelation
abilities.
For the glycine derivatives (21 and 24) the tris-amide 21
showed the best gelation ability. With most solvents investi-
gated a gel was obtained or a viscous solution. The minimal
gelation concentrations for the more polar solvents were
low. Substitution of the esters by amides (24) resulted in
a slight decrease in gelation ability, especially for the
apolar, aliphatic solvents in which the compound was not
soluble.
For the tris-amide-amides 22–24, the influence of the R-
group is rather clear. Comparing 23 (R¼CH2CH(CH3)2)
with 22 (R¼CH2Ph) it seems that as for the tris-urea com-
pounds 16 and 17 the iso-butyl group has a positive effect
on the gelation ability of the compound, resulting in an in-
crease in the scope of gelated solvents. For the compound

M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
7291
with the least steric demand (24: R¼H) and thus expected
to have the best packing within the stacks, the gelation abil-
ity is comparable to 23 (R¼CH2CH(CH3)2) with a slightly
changed solvent scope and decreased minimal gelation con-
centrations.
Supplementary data for data and discussion). The results
show that aggregation and gelation of the compounds
studied is accompanied by hydrogen bond formation. The
concentration-dependent measurements and the resulting
association constants translate the observed differences in
aggregation behaviour to the aggregation ability of the sep-
arate components: the core (benzene<cyclohexane), the
hydrogen-bonding units (amide<urea) and the R-groups
(CH2CH(CH3)2<CH2Ph). For compounds 9, 17 and 20,
the estimated lower limit for the association constant is al-
ready much higher than the association constants observed
for the aggregation of cyclohexane bis-urea compounds,20
demonstrating their strong aggregation and gelation ability.
When aggregation (viscous solution or gel) is not visible
even up to high concentrations, hydrogen bond formation
is not observed by FTIR (12). However, if at low concentra-
tions hydrogen bond formation and thus aggregation are
present, this will not always lead to thickened or gelated so-
lutions (compare 9 with 10). This is probably related to the
size of the formed aggregates, i.e., a highly thickened solu-
tion is only observed when the aggregates are large enough.
Preliminary dynamic light scattering measurements indicate
that indeed the aggregates formed in a non-thickened cyclo-
hexane solution of 10 (31.6 mM) are smaller than the aggre-
gates formed in a thickened cyclohexane solution of 9
(31.6 mM).
Thus, whereas for the benzyl compounds the facile packing
together with the additional p–p stacking seem to cause a
diminished solubility and thus gelation ability upon intro-
duction of additional hydrogen bonding interactions, for
the iso-butyl compounds the poorer packing appears to be
compensated by the introduction of extra hydrogen bonding
interactions resulting in increasing gelation abilities. For the
glycine compounds, facile packing can be expected but no
additional p–p stacking is present, thus the intermolecular
interactions are not too strong and also with additional
hydrogen bonding interactions gels are still obtained instead
of precipitates.
Apparently for these compounds, a subtle balance between
hydrogen bonding strength, steric effects and other interac-
tions, like p–p stacking, are present in order to yield gels,
viscous solutions or precipitates. A preliminary model
(Fig. 4C) shows that sterically demanding R-groups indeed
adopt an important position in the stack, most likely causing
deviation from the stacking pattern found for cyclohexane
tricarboxamides (Fig. 4A)7 (presumally twisting of the orig-
inally translational aggregate). However, further studies
are needed to gain more insight in the precise role of the
R-groups.
2.6. Thermotropic properties of the gels studied
by dropping ball measurements21
Dropping ball measurements reveal the temperature at
which the gel loses its mechanical stability, the so-called
melting temperature (also called gel–sol transition).21 Fall-
ing of the steel ball does not necessarily correspond to loss
of (intermolecular) aggregation, but rather to disintegration
of the gel network to a point at which the gel is not able to
support the steel ball anymore. Measurements were per-
formed on gels formed by cyclohexane-based compounds
and the results are shown in Figure 5. Depending on the
scope of gelated solvents, gels of 1-octanol (Fig. 5a) or tetra-
line (Fig. 5b) were studied. The gels were prepared in sealed
vials to avoid solvent evaporation and allowed to stabilise for
4–7 days prior to the measurements.
Comparison of the tris-urea compounds 16 and 17 with
known cyclohexane bis-urea organogelators5 reveals that
in general these cyclohexane bis-urea gelators gel a broader
scope of solvents and have lower minimal gelation concen-
trations. Apparently, the additional urea group present in 16
and 17 did not affect these properties positively. However,
a fair comparison is difficult, as these bis-urea compounds
have substantially different peripheral substituents.
An analogous comparison can be made between the
tris-amide compounds and cyclohexane bis-amide organo-
gelators:3a a smaller solvent scope is observed for the cyclo-
hexane tris-amides, most likely due to the presence of
stereogenic centres with large R-groups. This is consistent
with the fact that the poorest gelation behaviour is displayed
by the compound exhibiting the sterically most demanding
R-group (20: R¼iso-butyl). Compared to the trialkyl-cis,
cis-1,3,5-cyclohexane tricarboxamides reported by Hana-
busa8 the solvent scope had shifted, but, except for the
iso-butyl derivative, the gelation ability had not significantly
decreased for most compounds. The shift in solvent scope
might be related to the ester groups present in the com-
pounds 19–21.
All different gels melted thermoreversibly at temperatures
varying from 50 ꢀC to >155 ꢀC, depending on the concen-
tration, the hydrogen bonding unit and the R-group. For
most of the gels it was observed that the melting temperature
increased with increasing concentration up to a certain level,
a behaviour common for low molecular weight gelators. In
contrast, the tetraline gels of 22 were found to melt almost
independently of concentration. For several of the gels it
was observed that after melting a viscous solution (22 in tet-
raline) or gel-like precipitate (16, 17 and 23 in 1-octanol) re-
mained. For compound 23, the melting of gels could not be
observed for c>17 mM, because the apparatus only allowed
measurements up to 155 ꢀC.
2.5. Infrared spectroscopy
FTIR spectroscopy was used to study further the differences
in aggregation behaviour observed for the C3-symmetric
compounds. Concentration- and temperature-dependent
measurements were performed on cyclohexane solutions
or gels of the tris-ureas 9, 10 and 17 and the tris-amides 12
and 20 and compared to measurements on cast films (see
The influence of the hydrogen-bonding unit can be discussed
by comparing the 1-octanol gels of 16 with 19 (R¼CH2Ph),
21 with 24 (R¼H) and 17 with 23 (R¼CH2CH(CH3)2).
Comparison of the 1-octanol gels of 16 (tris-urea) with the
gels of 19 (tris-amide) showed considerable higher melting

7292
(a)
M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
tetraline gels of the tris-amide-amides 22 with 23. Com-
parison of the tris-urea 16 (R¼CH2Ph) and 17 (R¼
CH2CH(CH3)2) reveals higher melting temperatures and
thus a higher thermal stability for the benzyl derivative 16,
especially at lower concentrations. This can be explained
from the additional p–p-stacking interactions and lower
steric hindrance for R¼CH2Ph, which will enhance the inter-
molecular interactions. These results are comparable to the
aggregation and FTIR results obtained for the benzene core
compounds 9 and 10, in which the iso-butyl group also has
a destabilising effect compared to the benzyl derivative.
Comparison of the tris-amides 19 (R¼CH2Ph) with 21
(R¼H) shows that these compounds melt in a similar tem-
perature range, depending on concentration. Apparently,
removal of the benzyl group does not have a significant in-
fluence on the thermal stability of the gels. Most likely the
loss of the p–p-stacking interactions is compensated by the
diminished steric hindrance, which enables a closer packing
and thus increases the intermolecular interactions.
(b)
Comparison of thetris-amide-amides 23 (R¼CH2CH(CH3)2)
with 24 (R¼H) reveals at low concentrations lower melting
points and thus a lower thermal stability for the iso-butyl de-
rivative 23. This is in accordance with the larger steric hin-
drance of the iso-butyl group, which will hinder a close and
thus strong packing of the molecules. At high concentrations,
however, the melting points and thermal stability of 24 are
lower than those of 23. This is contrary to expectations and
difficult to explain.
Comparison of the tetraline gels of the tris-amide-amides 22
(R¼CH2Ph) with 23 (R¼CH2CH(CH3)2) also reveals ther-
motropic behaviour, that is, in contrast to expectation, since
the benzyl derivative 22 melts at lower temperatures than the
iso-butyl derivative 23. However, after melting it is observed
that the solution of 22 is still thickened and thus the aggre-
gates are not completely dissociated, whereas for 23 thick-
ening is not observed. Most likely, for compound 22 some
structural rearrangement at the transition temperature causes
the gel fibres to dissociate into single strands, which thicken
the solution. A reason for the thermal instability of the gel of
22 compared to 23 is not obvious, but might be related to the
aromatic solvent used.
Figure 5. Thermal behaviour of: (a) 1-octanol gels of 16 (-:-), 17 (-7-),
19 (---), 21 (-)-), 23 (-C-) and 24 (-B-); (b) tetraline gels of 22 (-;-)
and 23 (-6-).
temperatures for the tris-urea 16. For compound 16 the inter-
molecular interactions were so strong that after dropping of
the ball a gel-like precipitate was still present, which re-
mained even at temperatures of 150 ꢀC. Thus, as expected
the introduction of additional hydrogen bond donors and hy-
drogen bonds immediately enhanced the thermal stability of
the gels. This also corresponds with the higher association
constants of the urea compounds (K10) compared to the
amide compounds (K12) as found with FTIR (Supplementary
data). A similar conclusion can be drawn by comparison of
the glycine derivatives 21 (tris-amide) and 24 (tris-amide-
amide); substitution of the ester functionality by amides to
introduce additional hydrogen bonds also resulted in an
increase in melting temperature of 50 ꢀC.
Summarising, the number of hydrogen bonding interactions
has a significant influence on the thermal stability of the gels.
Both the tris-urea and the tris-amide-amides form stronger
gels than the tris-amides. Between the tris-urea and the
tris-amide-amide also a difference is observed, which is re-
lated to the nature of the hydrogen-bonding unit. The influ-
ence of the R-group is less straightforward. Substitution of
the benzyl group with a hydrogen group does not result in
a significant difference. This fact might be explained by
a compensation of the loss of p–p-stacking interactions
with a closer packing due to loss of steric hindrance. How-
ever, comparison of these R-groups with the iso-butyl group
yields peculiar and so far unexplainable results.
Comparison of 17 (tris-urea) with 23 (tris-amide-amide) re-
veals a considerably lower melting temperature for the tris-
urea 17. Thus it seems that, at least for R¼CH2CH(CH3)2,
the use of the urea unit produces thermally less stable gels
than the combination of two amide groups. A straightfor-
ward explanation is not available, however. Perhaps a single
amide hydrogen bond is stronger than a single urea hydrogen
bond (of the two hydrogen bonds an urea group can form).
Comparison of the melting behaviour of the cyclohexane
tris-urea derivatives 16 and 17 with cyclohexane bis-urea
gels5g reveals higher melting points for the tris-urea. Thus,
it can be concluded that the introduction of an additional
The influence of the R-group can be discussed by comparing
the 1-octanol gels of the tris-urea 16 with 17, the tris-amides
19 with 21, the tris-amide-amides 23 with 24 and the

M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
7293
Table 3. Thermotropic properties of 1-octanol gels of 16, 19 and 23–24a
urea group and as a consequence of further intermolecular
hydrogen bonding has a positive influence on the thermal
stability. For the tris-amides 19 and 21 the melting tem-
peratures are comparable with those of cyclohexane bis-
amides.3a Possibly, the presence of stereogenic centres
has a destabilising effect. Unfortunately, no thermotropic
data are available for the trialkyl cis-1,3,5-cyclohexane tri-
carboxamides,8 thus the exact influence of the stereogenic
centres cannot be deduced.
Gel sample
Tm (ꢀC)
T1 (ꢀC)
T2 (ꢀC)
T3 (ꢀC)
16 (c¼31 mM)
19 (c¼30 mM)
23 (c¼66 mM)
24 (c¼41 mM)
125b
73
>154
132
54 (6.4)
72 (19.4)
53 (3.2)
37 (3.6)
158 (8.1)
183 (8.9)
111 (15.8)
52 (5.1)
176 (9.3)
140 (20.9)
a
Tm is the melting temperature of the gels as deduced by the dropping ball
method. T1, T2 and T3 are the temperatures at the maxima of the endo-
therms observed by DSC. The numbers in parentheses are the associated
enthalpies (kJ molꢁ1).
b
A gel-like precipitate remained after melting.
2.7. Thermotropic properties of the gels studied by
differential scanning calorimetry
Differential scanning calorimetry (DSC) was carried out to
gain insight into the phase transitions of the gels. The mea-
surements were performed on 1-octanol gels of 16, 19, 23
and 24, which were prepared inside large volume sample
pans by heating in the DSC apparatus. Subsequently, two
sets of heating and cooling scans were recorded with an in-
terval of 2–3 h. These subsequent scans revealed identical
curves for each compound, indicating that their gelation
was fully thermoreversible. Figure 6 presents the heating
curves recorded for the different gel samples, showing that
large differences between the various gels are present. The
cooling scans are not discussed, since clear phase transitions
were not observed upon cooling. The common exotherm at
T¼65 ꢀC is related to the softener present in the Viton O-
ring of the sample pans. For compound 19 this peak would
coincide with the phase transition of the gel itself, therefore
measurements on the gel of 19 were performed without the
presence of the Viton O-ring.22
nature of the solvent. This agrees with the finding that for
long alkyl bis-urea gelators the gelation process in polar al-
kanol solvents is primarily driven by solvophobic interac-
tions (entropic) instead of hydrogen bonding interactions.5g
Substitution of the amide groups of 19 by ureas to obtain the
tris-urea 16 (R¼CH2Ph), led to a completely different DSC
heating curve (Fig. 6). Instead of one cooperative transition,
three low endothermic transitions are observed at T1¼54 ꢀC,
T2¼158 ꢀC and T3¼183 ꢀC, indicating that several pro-
cesses are taking place (Table 3). None of these can be re-
lated to the melting temperature of 125 ꢀC determined by
dropping ball measurements (vide supra). Furthermore, the
dropping ball measurements showed that after falling of
the ball only the mechanical stability of the gel was lost,
whereas still a gel-like precipitate was present that remained
even at the uppermost temperature (155 ꢀC) of the dropping
ball apparatus. The temperatures T2 and T3 at which two of
the observed transitions occur exceed this limit and thus the
macroscopic events at these temperatures were not observed.
However, most likely one of these transitions includes the
complete dissolution of the compound into the solvent. In
addition, the peak at T3 could be related with decomposition
of the urea groups, although this disagrees with the decom-
position temperature of the solid (Tdecomp.ꢂ230 ꢀC). The en-
dotherm at T1¼54 ꢀC is also present in the gels formed by 23
and 24, suggesting that it is related to a phase transition
associated with a common structural reorganisation. The
enthalpies calculated for the transitions of 16 are small.
However, the total of these enthalpies is larger than the
enthalpy associated with the melting of the 1-octanol gel
of 19. Although a direct comparison cannot be made, this
does support the high aggregation ability and strength of
the urea compounds compared to the amide compounds as
found with FTIR and dropping ball measurements.
The 1-octanol gel of the tris-amide 19 (R¼CH2Ph) displayed
a heating curve with a strong endothermic transition at
T¼72 ꢀC in a rather narrow temperature range, which
indicates that this transition is cooperative (Fig. 6).22 The
dropping ball measurements (vide supra) showed that the
gel is completely transformed into a homogeneous solution
at 73 ꢀC, thus the transition was assigned as the gel–sol tran-
sition of the gel. The enthalpy calculated for the gel–sol tran-
sition was 19.4 kJ molꢁ1 (Table 3). In apolar aprotic solvents
a much higher enthalpy is expected for the breaking up of
three hydrogen bonds,23 thus the low value of the melting en-
thalpy is most likely caused by the more polar and protic
For the tris-amide-amide 23 (R¼CH2CH(CH3)2) three endo-
therms are observed at T1¼53 ꢀC, T2¼111 ꢀC and
T3¼176 ꢀC (Fig. 6; Table 3). As discussed above, the rela-
tively small transition at T1 is also present in the gels of 16
and 24. The two transitions at higher temperatures are very
broad, indicating that the processes related to these transi-
tions are not cooperative. Dropping ball measurements
(vide supra) show that melting of the gel does not take place
below 154 ꢀC, thus the transition at T2 cannot be assigned to
the melting of the gel. Instead most likely some kind of re-
organisation takes place that does not result in a macroscopic
event. The phase transition at T3 is more likely to be due to
melting of the gel into a homogeneous solution, although
other processes cannot be excluded.
Figure 6. DSC heating curves of 1-octanol gels of 16 (31 mM), 19 (30 mM),
23 (66 mM) and 24 (41 mM).

7294
M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
Also for the tris-amide-amide 24 (R¼H) three endotherms
are observed (Fig. 6; Table 3). The origin of the peak
observed at T1¼37 ꢀC is unknown, but might be related to
a conformational rearrangement of the peripheral substitu-
ents. The transition at T2¼52 ꢀC is comparable to transitions
observed for 16 and 23 and is most likely related to a common
structural reorganisation. The third transition at T3¼140 ꢀC
coincides with the melting point of the gel as deduced by
dropping ball measurements (vide supra) and is therefore
most likely related to the gel–sol transition of the gel.
that the fibres are brittle. A clear splitting and fusing of the
fibres to form thegel network could not be seen. Furthermore,
despite the chirality of the molecule, no twisting of the fibres
was observed. In p-xylene similar fibre morphology was ob-
served. The observed fibre morphology might be related to
the fact that 16 precipitates in many of the solvents tested.
In contrast, when the benzyl group is replaced with an iso-
butyl group (17; Fig. 7B) elongated fibres are observed,
which exhibit a clear left-handed twisting with a pitch of
w200 nm. The fibres appear to be more flexible and vary
in diameter, with an observed minimal size of 18 nm. The
twisted fibres join regularly to form thicker fibres of identical
handedness (Fig. 7C). Most likely, joining and splitting of
the fibres are responsible for the formation of a three-dimen-
sional gel network.
The fact that only for the tris-amide 19 a single melting tran-
sition is observed, whereas for the thermally more stable tris-
urea 16 and tris-amide-amides 23 and 24 multiple peaks are
found suggesting that this is related to the increased number
of intermolecular hydrogen bonding interactions. Probably
the stronger hydrogen bonding interactions are still able to
hold the aggregates together after the initial destabilising
events have taken place. However, the presence of different
hydrogen bonding units could also result in different mor-
phologies involving other transitions. The TEM studies
(vide infra), however, do not give a clear and consistent re-
lation between the morphology of the gels and the transitions
observed with DSC, since the compounds that show three
transitions can form short thin fibres (16) as well as elon-
gated sheets (24).
Substitution of the urea groups with amides to obtain the cy-
clohexane tris-amides 19–21, resulted in a completely differ-
ent morphology and a different effect of the R-group (Fig. 8).
In contrast to the morphology of its tris-urea counterpart
(16), the benzyl derivative 19 formed is elongated, thin fibres
with a monodisperse diameter of 13 nm (Fig. 8A). The fibres
exhibited a right-handed helicity with a pitch of 80–100 nm.
The fibres neither split nor fuse, but align into thicker bun-
dles. Interestingly, in other solvents like tetraline and 1,2-di-
chloroethane a similar morphology was observed, which is
uncommon for other LMWGs.5
2.8. Morphology of the gels
To determine the morphology of the obtained gels, several of
them were subjected to transmission electron microscopy
(TEM) studies. For this purpose TEM samples were prepared
of gels, which were dried and Pt-shadowed prior to
investigation. The morphological differences between the
gels of the various compounds will be discussed with respect
to the differences in R-group and hydrogen-bonding unit.
Chirality aspects will not be considered in any detail.
Figure 7 represents micrographs of 1-octanol gels of the
cyclohexane tris-urea derivatives 16 (R¼CH2Ph) and 17
(R¼CH2CH(CH3)2). In both gels, fibres are present, how-
ever, a clear difference in morphology can be observed. For
compound 16 (Fig. 7A) short and thin fibres (diameter of
w13 nm) are present, which cluster into thicker bundles.
The presence of large amounts of small fragments indicates
Whereas the tris-urea iso-butyl derivative 17 formed twisted
fibres, the tris-amide iso-butyl derivative 20 formed elon-
gated sheet-like fibres, for which nearly no twisting was ob-
served. (Fig. 8B). Their diameter of 40–360 nm was larger
than those of the other compounds discussed so far. Occa-
sionally the fibres fuse or split to form the gel network.
The observed difference in morphology might be related to
the different solvent used,24 which results from the fact
that 20 gels only n-hexadecane.
The morphology of 21 (R¼H; Fig. 8C), somewhat resembles
the morphology observed for the iso-butyl derivative 20.
Also elongated sheet-like fibres are observed with a diameter
of 10–300 nm, which fuse and split to connect into a three-
dimensional gel network. A minor part of these fibres appear
Figure 7. Electron micrographs of 1-octanol gels of the cyclohexane tris-urea: (A) 16 (R¼CH2Ph; c¼10 mg mLꢁ1); (B) 17 (R¼CH2CH(CH3)2; c¼5 mg mLꢁ1);
bars represent 500 nm and (C) detail of the gel of 17; bar represents 100 nm.

M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
7295
Figure 8. Electron micrographs of gels of the cyclohexane tris-amide derivatives: (A) 19 in 1-octanol (R¼CH2Ph; c¼10 mg mLꢁ1); (B) 20 in n-hexadecane
(R¼CH2CH(CH3)2; c¼10 mg mLꢁ1) and (C) 21 in 1-octanol (R¼H; c¼10 mg mLꢁ1). Bars represent 500 nm.
to be twisted, equally both left-handed and right-handed.
The pitch is mostly irregular and variable. It seems remark-
able that the achiral compound 21 forms twisted fibres,
whereas for the chiral counterpart 20 nearly no twisting is
observed. However, this could well be related to the different
solvents used.24
it was observed that the fibres were even more sheet-like and
exhibited no twist at all.
Comparing all micrographs, it can be observed that almost
all compounds form fibres with high aspect ratios, which re-
flect the high anisotropy of the interactions. Exceptions are
the benzyl derivatives 16 and 22, which form relatively short
fibres. Possibly this is related to the fact that these com-
pounds experience the strongest interactions due to the com-
bination of strong hydrogen-bonding units (urea groups or
two combined amides, respectively) with p–p-stacking
groups. These strong interactions most likely cause the mol-
ecules to preferably aggregate fast into new fibres, instead of
first diffuse towards existing fibres to co-aggregate with
these to form more elongated fibres. This process would
lead to a larger number of fibres with a lower fibre aspect ra-
tio. In both cases, decreasing the intermolecular interactions
by substitution of the benzyl group with non-aromatic (and
sterically demanding) groups or by reducing the number of
hydrogen bonds resulted in the formation of elongated, flex-
ible fibres.
For the tris-amide-amide derivatives 22 and 24 differences in
morphology are observed, which are comparable to the tris-
amide derivatives 19 and 21 (Fig. 9). The benzyl derivative
22 formed short fibres with a diameter of 15–70 nm
(Fig. 9A). Similar to its tris-amide counterpart 19, the fibres
exhibited a right-handed twist. In contrast to the regular
pitch observed for 19, the fibres of 22 exhibit an often irreg-
ular pitch of about 100–150 nm. Thinner fibres fuse into
thicker (bundles of) fibres, which results in the formation
of junction zones and consequently a three-dimensional
gel network.
Analogously to its tris-amide counterpart 21, the achiral tris-
amide-amide 24 forms sheet-like fibres, which exhibit occa-
sional irregular twisting, both left-handed and right-handed
(Fig. 9B). Probably, the sheet-like morphology of the fibres
from 21 and 24 is related to the absence of molecular chiral-
ity. The diameter of the fibres ranges from 20–250 nm and
compared to 21, the fibres formed by 24 seem to be thinner.
The fibres split and fuse to form the gel network. In tetraline
Comparison of the morphologies observed for these com-
pounds reveals in general a similar diversity as found for
cyclohexane bis-urea5a,e or bis-amide3a organogelators.
The exception is formed by compound 19, which displays
a homogeneous, solvent-independent morphology.

7296
M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
Figure 9. Electron micrographs of gels of cyclohexane tris-amide-amides: (A) 22 in tetraline (R¼CH2Ph; c¼10 mg mLꢁ1) and (B) 24 in 1-octanol (R¼H;
c¼5 mg mLꢁ1). Bars represent 500 nm.
3. Conclusions
cyclohexane bis-urea 1-octanol gels5g the thermal stabilities
have increased, indicating that the additional urea positively
affects the thermal stability.
In conclusion, based on rational considerations a new class
of chiral C3-symmetric a-amino acid based organogelators
and thickeners have been developed. Starting from commer-
cially available building blocks, compounds with a large
structural diversity are easily accessible in three or four
steps. Variation was applied in the central core, the hydrogen
bonding units and the amino acids. These modifications have
a strong influence on the intermolecular interactions, result-
ing in large differences in aggregation properties, thermal
stability and morphology between the various compounds.
This allows fine-tuning of the gelators with respect to aggre-
gation behaviour and gel properties.
The work described here comprises only a limited variation
in the peripheral substituents. Further understanding of the
influence of steric hindrance, p–p stacking etc. could be ob-
tained by using other natural or unnatural amino acids. Also
variation of the stereochemistry of these amino acids offers
interesting possibilities. Additional opportunities can be ex-
pected from the use of other than alkyl esters. Accordingly
we have previously shown that the use of amino acids that
are not esterified at the carboxylic acid terminus or esterified
with hydrophilic substituents, resulted in a shift of the sol-
vent scope from organic solvents to water and aqueous solu-
tions.6
The central core unit has a strong influence on the aggrega-
tion ability and both gelation studies and FTIR spectroscopy
showed that the cyclohexane core was superior to the benz-
ene core. This is consistent with the results obtained for tri-
alkyl-tricarboxamide derivatives.8,10,11 The type and number
of hydrogen-bonding units influenced the number of possi-
ble hydrogen bonds and as such the strength of the intermo-
lecular interactions. Generally, stronger hydrogen bonding
interactions (urea or amide-amide) resulted in stronger
aggregation and higher thermal stabilities. However, this is
balanced by the a-amino acids (R-groups) used, which influ-
enced the magnitude of the steric hindrance and the possible
presence/absence of p–p stacking. It seems that too strong
intermolecular hydrogen bonding interactions (leading to
insolubility, crystallisation or precipitation) can be compen-
sated by the introduction of steric hindrance to yield
effective gelators. Weaker hydrogen bonding interactions
(amides) seem to be strengthened by the presence of p–p-
stacking interactions and less steric hindrance.
4. Experimental section
4.1. General remarks
Melting points (uncorrected) were determined using a Stuart
Scientific SMP1 melting point apparatus; no attempt was
made to measure the melting points of sticky solids. H
1
NMR and 13C NMR spectra were recorded on a Varian Gem-
ini-200 (200 MHz or 50.32 MHz, respectively) or a Varian
VXR-300 (300 MHz or 75.48 MHz, respectively) operating
at ambient temperature, unless stated otherwise. Chemical
shifts are denoted in d units (ppm) relative to the residual sol-
vent peaks (1H NMR: CDCl3, d¼7.26; DMSO-d6, d¼2.49.
13C NMR: CDCl3, d¼76.91; DMSO-d6, d¼39.5) and con-
verted to the TMS scale. The splitting patterns are designated
as follows: s (singlet), br s (broad singlet), d (doublet), br
d (broad doublet), dd (double doublet), t (triplet), dt (double
triplet), br t (broad triplet), q (quartet), m (multiplet) and br p
(broad peak). FTIR spectra were recorded on a Nicolet
Nexus FTIR apparatus, ~n values are listed in cmꢁ1 units.
Elemental analyses and mass spectrometry were performed
in the analytical department of this laboratory.
FTIR spectroscopy showed that for most of the compounds
the tendency to aggregate is so strong that even at very low
concentrations all molecules are fully hydrogen bonded.
This strong aggregation ability is most likely caused by
the favourable orientation and cooperativity of the three
sets of hydrogen-bonding units, enforced by the rigid cores.
The hydrogen bond formation most likely causes the high
thermal stabilities of the gels. Indeed, compared to
Synthesis of the a-amino acid esters and amides 1–6 is de-
scribed in the Supporting data. cis,cis-1,3,5-Cyclohexane

M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
7297
3
tricarboxylic acid was purchased from FLUKA. Diphenyl
phophoryl azide (DPPA) and 1,3,5-benzenetricarboxylic
acid were purchased from Aldrich. Di-tert-butyl dicarbonate
was purchased from Acros.
3H), 7.09 (s, 3H), 6.30 (d, J¼7.7 Hz, 3H), 4.23–4.16 (m,
3H), 4.08–3.98 (m, 6H), 1.68–1.46 (m, 15H), 1.20 (br s,
30H), 0.91–0.80 (m, 27H); 13C NMR (DMSO-d6):
d¼173.30, 154.58, 140.53, 100.19, 64.34, 50.80, 40.86,
31.81, 28.58, 28.54, 28.06, 25.31, 24.34, 22.85, 22.09,
21.62, 13.91; IR (film, NaCl): ~n¼3347 (N–H), 1740
(C]O, ester), 1645 (C]O, amide I), 1558 cmꢁ1 (N–H,
amide II); MS (ES): m/z 1880.5 [2M+NH+4]+, 1417.4
[3M+Na++NH+4]2+, 953.7 [M+Na+]+, 948.8 [M+NH+4]+,
931.9 [M+H+]+.
Caution: the synthetic procedures for compounds 9 and 10
involve the preparation of 1,3,5-benzenetricarbonyl triazide,
a highly explosive compound in its solid state!15
4.2. Syntheses
4.2.1. 1,3,5-Benzenetricarbonyl trichloride (7). 1,3,5-
Benzenetricarboxylic acid (6.0 g, 28.7 mmol) was placed
in a flask together with SOCl2 (12 mL) and a drop of
DMF. The suspension was refluxed for 3 h, affording a clear
solution. The remaining SOCl2 was evaporated in vacuo,
yielding 7 as a pale yellow oil that crystallised upon standing
at 4 ꢀC (7.7 g, 28.7 mmol, 100%). 1H NMR (CDCl3):
d¼9.08 (s, 3H).
4.2.4. 1,3,5-Benz(Am-L-Phe-O-octyl)3 (11). To a cooled
solution of 1 (2.0 g, 7.2 mmol) and triethylamine (0.73 g,
7.2 mmol) in dry CH2Cl2 (50 mL) was added a solution of
1,3,5-benzenetricarbonyl trichloride 7 (0.64 g, 2.4 mmol)
in dry CH2Cl2 (5 mL). The solution was slowly brought to
room temperature and stirred for 20 h. Subsequently
CHCl3 (20 mL) was added and the solution was extracted
successively with dilute HCl (3ꢃ50 mL), water
(2ꢃ50 mL), 10% aqueous sodium carbonate (3ꢃ50 mL),
water (2ꢃ50 mL) and brine. The solution was dried over
MgSO4 and the solvents were evaporated in vacuo, yielding
a sticky solid. Column chromatography (CH2Cl2/MeOH;
100:1) on silica gel yielded 11 (1.2 g, 1.21 mmol, 50%) as
a white sticky solid. 1H NMR (DMSO-d6): d¼9.12 (d,
3J¼7.3 Hz, 3H), 8.39 (s, 3H), 7.26–7.18 (m, 15H), 4.71–
4.2.2. 1,3,5-Benz(U-L-Phe-O-octyl)3 (9). To a cooled (0 ꢀC)
solution of NaN3 (1.75 g, 27 mmol) in H2O (10 mL) was
added a cooled (0 ꢀC) solution of 1,3,5-benzenetricarbonyl
trichloride 7 (1.0 g, 3.77 mmol) in THF (10 mL). The mix-
ture was stirred for 2 h at 0 ꢀC resulting in the formation
of 1,3,5-benzenetricarbonyl triazide as a white precipitate.
1H NMR (200 MHz, CDCl3): d¼8.86 (s, 3H). This precipi-
tate can easily be isolated by filtration, however, because of
the explosive nature of the dry solid this is strongly rejected.
Thus, cold toluene (100 mL) was added to the mixture to
take up the acyl azide. The toluene layer was separated
and washed with water (3ꢃ80 mL) and brine and dried
over MgSO4. Subsequently, the toluene solution was gradu-
ally heated to 100 ꢀC and stirred till gas evolution stopped,
yielding in situ the corresponding triisocyanate 8. The solu-
tion was allowed to cool to room temperature and 1 (3.45 g,
12.44 mmol) in toluene (30 mL) was added. The mixture
was stirred for one night at room temperature, after which
the solvents were evaporated in vacuo yielding a sticky solid.
Column chromatography using an eluent gradient (CH2Cl2/
MeOH; 200:0–200:5) on silica gel yielded 9 slightly con-
taminated. A second column chromatography (CH2Cl2/
MeOH; 200:10) on silica gel yielded 9 as a pure colourless
3
3
4.63 (m, 3H), 4.01 (t, J¼6.4 Hz, 6H), 3.13 (d, J¼8.4 Hz,
3
6H), 1.48 (s, 6H), 1.17 (s, 30H), 0.80 (t, J¼6.6 Hz, 9H);
13C NMR (DMSO-d6): d¼171.55, 165.54, 137.58, 134.16,
129.29, 129.02, 128.26, 126.49, 64.57, 54.62, 36.32,
31.19, 28.58, 28.04, 25.28, 22.07, 13.92; elemental analysis
calcd (%) for C60H81N3O9 (988.3): C 72.92, H 8.26, N 4.25;
found: C 72.83, H 8.33, N 4.25.
4.2.5. 1,3,5-Benz(Am-L-Leu-O-octyl)3 (12). Compound 12
was synthesised following the same procedure as described
for 11, using 2 (2.0 g, 8.2 mmol), triethylamine (0.83 g,
8.2 mmol) and 1,3,5-benzenetricarbonyl trichloride
7
(0.73 g, 2.73 mmol). Column chromatography (CH2Cl2/
MeOH; 100:1) on silica gel yielded 12 (1.48 g, 1.67 mmol,
62%) as a white sticky solid. 1H NMR (DMSO-d6):
3
d¼9.02 (d, J¼7.3 Hz, 3H), 8.49 (s, 3H), 4.54–4.47 (m,
3H), 4.10–4.01 (m, 6H), 1.84–1.51 (m, 15H), 1.23–1.19
1
3
sticky solid (0.76 g, 0.74 mmol, 20%). H NMR (DMSO-
(m, 30H), 0.93–0.88 (m, 18H), 0.83 (t, J¼6.6 Hz, 9H);
d6): d¼8.62 (s, 3H), 7.31–7.16 (m, 15H), 7.08 (s, 3H),
6.27 (d, 3J¼7.7 Hz, 3H), 4.48–4.41 (m, 3H), 3.99 (t,
3J¼6.4 Hz, 6H), 3.04–2.91 (m, 6H), 1.48 (s, 6H), 1.20 (s,
30H), 0.83 (t, 3J¼6.6 Hz, 9H); 13C NMR (DMSO-d6):
d¼172.40, 154.66, 140.65, 136.93, 129.34, 128.55, 126.89,
100.55, 64.76, 54.04, 37.66, 31.41, 28.79, 28.20, 25.50,
22.26, 14.13; IR (film, NaCl): ~n¼3328 (N–H), 1738
(C]O, ester), 1638 (C]O, amide I), 1555 cmꢁ1 (N–H,
amide II); elemental analysis calcd (%) for C60H84N6O9
(1033.4): C 69.74, H 8.19, N 8.13; found: C 69.72, H 8.23,
N 8.12.
13C NMR (DMSO-d6): d¼172.44, 165.81, 134.30, 129.40,
64.43, 51.25, 39.28, 31.15, 28.55, 28.08, 25.30, 24.42,
22.83, 22.05, 21.21, 13.91; IR (film, NaCl): ~n¼3226 (N–
H), 1753 (C]O, ester), 1640 (C]O, amide I), 1555 cmꢁ1
(N–H, amide II); elemental analysis calcd (%) for
C51H87N3O9 (886.3): C 69.12, H 9.89, N 4.74; found: C
69.24, H 9.96, N 4.77.
4.2.6. 1,3,5-Benz(Am-Gly-O-octyl)3 (13). Compound 13
was synthesised following the same procedure as described
for 11, using 3 (2.02 g, 10.8 mmol), triethylamine (1.1 g,
10.8 mmol) and 1,3,5-benzenetricarbonyl trichloride 7
(0.93 g, 3.5 mmol). After extraction, compound 13 was ob-
tained as a yellowish sticky solid (1.77 g, 2.44 mmol, 70%).
1H NMR (CDCl3): d¼8.24 (s, 3H), 8.01 (t, 3J¼5.7 Hz, 3H),
4.23–4.14 (m, 12H), 1.72–1.65 (m, 6H), 1.32–1.28 (m, 30H),
4.2.3. 1,3,5-Benz(U-L-Leu-O-octyl)3 (10). Compound 10
was synthesised following the same procedure as described
for 9, using NaN3 (1.75 g, 27 mmol), 7 (1.0 g, 3.77 mmol)
and 2 (2.75 g, 11.3 mmol). After reaction an orange sticky
solid was obtained and column chromatography (CH2Cl2/
MeOH; 200:5) on silica gel yielded 10 as a pure, sticky solid
(1.25 g, 1.34 mmol, 36%). 1H NMR (DMSO-d6): d¼8.49 (s,
3
0.89 (t, J¼6.6 Hz, 9H); 13C NMR (CDCl3): d¼170.59,
166.44, 134.41, 128.49, 65.78, 41.91, 31.70, 29.15, 29.09,
28.45, 25.82, 22.55, 13.97; elemental analysis calcd (%)

7298
M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
for C39H63N3O9 (724.0): C 65.23, H 8.85, N 5.86; found: C
64.70, H 8.94, N 5.72.
30H), 0.84 (t, 3J¼6.4 Hz, 9H); 13C NMR (DMSO-d6,
80 ꢀC): d¼172.01, 156.17, 136.79, 128.67, 127.73, 126.00,
63.91, 53.61, 45.38, 37.60, 30.70, 28.01, 27.65, 24.83,
21.53, 13.32; MS (ES): m/z 1061.7 [M+Na+]+, 1056.9
[M+NH+4]+, 1039.8 [M+H+]+.
4.2.7. cis,cis-1,3,5-Tris{[(benzyloxy)carbonyl]amino}cy-
clohexane (14). A solution of cis,cis-1,3,5-cyclohexane tri-
carboxylic acid (4.0 g, 18.5 mmol), diphenyl phophoryl
azide (DPPA; 15.4 g, 56 mmol) and triethylamine (8 mL)
in benzene (150 mL) was stirred for 30 min at room temper-
ature to obtain cis,cis-1,3,5-cyclohexane tricarbonyl triazide
in solution. Subsequently the solution was refluxed for
30 min till gas evolution stopped. Benzyl alcohol (6.4 mL)
was added and the solution was refluxed for 18 h. Subse-
quently the mixture was cooled to 4 ꢀC and the product
was filtered off and washed with cold benzene, yielding 14
as a white powder (5.8 g, 10.9 mmol, 59%). 1H NMR
(DMSO-d6): d¼7.32 (s, 15H), 4.98 (s, 6H), 3.05–3.00 (m,
3H), 1.87 (br p, 3H), 1.07 (br p, 3H); 13C NMR (DMSO-
d6): d¼154.23, 136.16, 127.82, 127.33, 126.78, 64.14,
45.62, 39.74.
4.2.10. cis,cis-1,3,5-Chex(U-L-Leu-O-octyl)3 (17). Com-
pound 17 was synthesised following the same procedure as
described for 16, using di-tert-butyl dicarbonate (1.66 g,
7.6 mmol) in dry CH2Cl2 (10 mL), 4-dimethylamino pyri-
dine (88 mg, 0.72 mmol) in dry CH2Cl2 (6 mL), 2 (1.75 g,
7.2 mmol) in dry CH2Cl2 (12 mL) and cis,cis-1,3,5-triami-
nocyclohexane 15 (0.284 g, 2.2 mmol) in CH2Cl2 (10 mL).
After cooling, the solvent was evaporated in vacuo. The
waxy residue was refluxed in ethanol and after cooling the
obtained gel was filtered off and washed with ethanol and di-
ethyl ether, yielding 17 as a white solid (0.94 g, 1.00 mmol,
1
45%). Mp: >215 ꢀC (dec); H NMR (CDCl3+5% TFA):
3
d¼4.46 (br p, 3H), 4.20 (t, J¼6.6 Hz, 6H), 3.68 (br p,
3H), 2.26 (br d, 3J¼11.0 Hz, 3H), 1.68–1.52 (m, 15H),
1.27 (s, 33H), 0.96–0.84 (m, 27H); 13C NMR (CDCl3+5%
TFA): d¼175.84, 158.18, 67.51, 52.73, 46.55, 40.79,
37.41, 31.62, 28.98, 28.93, 28.04, 25.52, 24.69, 22.49,
22.20, 21.30, 13.85; IR (film, NaCl): ~n¼3307 (N–H), 1744
(C]O, ester), 1639 (C]O, amide I), 1562 cmꢁ1 (N–H,
amide II); elemental analysis calcd (%) for C51H96N6O9
(937.4): C 65.35, H 10.32, N 8.97; found: C 65.05, H
10.44, N 9.33.
4.2.8. cis,cis-1,3,5-Triaminocyclohexane (15). Compound
14 (2.75 g, 5.2 mmol) was dissolved in 33% HBr/glacial
acetic acid (18 mL, 24.9 g). The mixture was allowed to
stand for 90 min under occasional stirring. Diethyl ether
(100 mL) was added and the product was filtered off and
washed with diethyl ether, yielding cis,cis-1,3,5-triamino-
cyclohexane$3HBr as a white powder (1.8 g, 4.83 mmol,
1
93%). H NMR (DMSO-d6): d¼8.16 (br p, 9H), 3.25 (br
p, 3H), 2.22 (br p, 3H), 1.42 (br p, 3H); 13C NMR
(DMSO-d6): d¼44.81, 32.80; ES/MS (MeOH) m/z 130.1
(M$3HBr +H+).
4.2.11. cis,cis-1,3,5-Cyclohexanetricarbonyl trichloride
(18). cis,cis-1,3,5-Cyclohexane tricarboxylic acid (3.0 g,
13.9 mmol) was placed in a flask together with SOCl2
(8 mL). The suspension was refluxed for 20 h, resulting in
a clear solution. The remaining SOCl2 was evaporated in va-
cuo, yielding 18 as a pale yellow oil that formed crystals
upon standing at 4 ꢀC (3.77 g, 13.9 mmol, 100%). 1H
NMR (CDCl3): d¼2.95–2.64 (m, 6H), 1.80–1.60 (m, 3H);
13C NMR (CDCl3): d¼174.20, 52.06, 30.17.
Subsequently,
cis,cis-1,3,5-triaminocyclohexane$3HBr
(1.8 g, 4.83 mmol) was dissolved in water (15 mL) and
4 equiv of NaOH (0.8 g, 20.0 mmol) was added. The mix-
ture was stirred for 18 h at room temperature and subse-
quently heated at reflux for 1 h. After cooling to room
temperature, the product was extracted from the aqueous so-
lution by a continuous extraction (48 h) with CH2Cl2 (down-
ward displacement). The CH2Cl2 solution obtained was
filtered and concentrated invacuo, yielding 15 as a yellowish,
crystalline solid (0.58 g, 4.50 mmol, 93%). 1H NMR
4.2.12. cis,cis-1,3,5-Chex(Am-L-Phe-O-octyl)3 (19). To
a cooled solution of 1 (1.0 g, 3.6 mmol) and triethylamine
(0.36 g, 3.6 mmol) in dry CH2Cl2 (20 mL) was added a solu-
tion of cis,cis-1,3,5-cyclohexane tricarbonyl trichloride 18
(0.33 g, 1.2 mmol) in dry CH2Cl2 (5 mL). The solution
was slowly brought to room temperature and stirred for
20 h. Subsequently, the solution was extracted with dilute
HCl (3ꢃ20 mL), water (2ꢃ20 mL), 10% aqueous sodium
carbonate (3ꢃ20 mL), water (2ꢃ20 mL) and brine. The so-
lution was dried over MgSO4 and the solvent was evaporated
in vacuo, yielding a sticky solid. Column chromatography
(CH2Cl2/MeOH; 100:5) on silica gel yielded 19 (0.3 g,
0.3 mmol, 25%) as a white sticky solid. 1H NMR (CDCl3):
d¼7.24–7.01 (m, 15H), 5.85 (d, 3J¼7.7 Hz, 3H), 4.84–
4.79 (m, 3H), 4.09–4.01 (m, 6H), 3.12–3.04 (m, 6H), 2.09
3
(CDCl3): d¼2.90–2.59 (m, 3H), 1.98 (br d, J¼11.2 Hz,
3H), 0.89–0.92 (m, 3H).
4.2.9. cis,cis-1,3,5-Chex(U-L-Phe-O-octyl)3 (16). To a solu-
tion of di-tert-butyl dicarbonate (0.83 g, 3.8 mmol) in dry
CH2Cl2 (10 mL) was added successively a solution of 4-di-
methylamino pyridine (44 mg, 0.36 mmol) in dry CH2Cl2
(3 mL) and a solution of 1 (1.0 g, 3.6 mmol) in dry
CH2Cl2 (6 mL). The mixture was stirred for 30 min at
room temperature till gas evolution stopped and subse-
quently cis,cis-1,3,5-triaminocyclohexane 15 (0.14 g,
1.1 mmol) in CH2Cl2 (5 mL) was added. The obtained turbid
mixture was first stirred at room temperature for 30 min and
then at 40 ꢀC for 48 h. After cooling, the solvent was evap-
orated in vacuo. The residue was refluxed in ethanol and af-
ter cooling the solid was filtered off and washed with ethanol
and diethyl ether, yielding 16 as a white powder (0.5 g,
3
3
(t, J¼12 Hz, 3H), 1.91 (d, J¼12 Hz, 3H), 1.54 (s, 9H),
1.23 (s, 30H), 0.83 (t, J¼6.6 Hz, 9H); 13C NMR (DMSO-
3
d6): d¼174.16, 171.70, 137.30, 128.98, 128.12, 126.44,
64.37, 53.40, 42.09, 36.64, 31.20, 28.53, 27.98, 25.19,
22.05, 13.93; IR (hexane gel, NaCl): ~n¼3286 (N–H), 1742
(C]O, ester), 1650 (C]O, amide I), 1550 cmꢁ1 (N–H,
amide II); elemental analysis calcd (%) for C60H87N3O9
(994.4): C 72.46, H 8.82, N 4.23; found: C 72.47, H 8.98,
N 4.26.
1
0.48 mmol, 43%). Mp: >230 ꢀC (dec); H NMR (DMSO-
d6): d¼7.28–7.13 (m, 15H), 6.03 (br s, 6H), 4.40–4.33 (m,
3H), 3.98–3.94 (m, 6H), 3.35 (br p, 6H), 2.92–2.88 (m,
6H), 1.83 (br d, J¼8.8 Hz, 3H), 1.46 (br p, 6H), 1.22 (s,
3

M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
7299
4.2.13. cis,cis-1,3,5-Chex(Am-L-Leu-O-octyl)3 (20). Com-
pound 20 was synthesised following the same procedure
as described for 19, using 2 (2.0 g, 8.2 mmol), triethylamine
(0.83 g, 8.2 mmol) and cis,cis-1,3,5-cyclohexane tri-
carbonyl trichloride 18 (0.74 g, 2.73 mmol). Column chro-
matography using an eluent gradient (CH2Cl2/MeOH;
200:5–200:10) on silica gel yielded 20 (1.06 g, 1.19 mmol,
44%) as a white sticky solid. 1H NMR (DMSO-d6):
trichloride 18 (0.33 g, 1.2 mmol). The solvent was evapo-
rated in vacuo, yielding a white waxy solid. Stirring in etha-
nol to remove the Et3N$HCl salts followed by filtration
afforded 23 as a white powder (0.62 g, 0.70 mmol, 58%).
Mp: >230 ꢀC (dec); 1H NMR (CDCl3+5% TFA): d¼7.60 (d,
3J¼8.1 Hz, 3H), 7.19 (s, 3H), 4.65–4.45 (m, 3H), 3.35–3.17
3
2
(m, 6H), 2.41 (t, J¼11.3 Hz, 3H), 1.98 (d, JAB¼12.1 Hz,
3H), 1.69–1.51 (m, 18H), 1.26 (s, 30H), 0.93–0.85 (m,
27H); 13C NMR (CDCl3+5% TFA): d¼176.13, 173.31,
52.41, 42.71, 40.58, 40.37, 31.58, 30.25, 28.94, 28.86,
28.38, 26.48, 24.64, 22.47, 21.97, 21.81, 13.86; elemental
analysis calcd (%) for C51H96N6O6 (889.4): C 68.86, H
10.89, N 9.45; found: C 68.86, H 11.15, N 9.40.
3
d¼8.11 (d, J¼7.7 Hz, 3H), 4.26–4.18 (m, 3H), 4.04–3.94
3
(m, 6H), 2.27 (t, J¼12.1 Hz, 3H), 1.69–1.35 (m, 21H),
1.23 (s, 30H), 0.88–0.80 (m, 27H); 13C NMR (DMSO-d6):
d¼174.37, 172.69, 64.27, 50.17, 42.31, 31.19, 28.58,
28.50, 28.04, 25.24, 24.35, 22.73, 22.08, 21.21, 13.97; IR
(solid, NaCl): ~n¼3294 (N–H), 1736 (C]O, ester), 1644
(C]O, amide I), 1539 cmꢁ1 (N–H, amide II); elemental
analysis calcd (%) for C51H93N3O9 (892.3): C 68.65, H
10.51, N 4.71; found: C 68.86, H 10.76, N 4.72.
4.2.17. cis,cis-1,3,5-Chex(Am-Gly-NH-octyl)3 (24). Com-
pound 24 was synthesised following the same procedure as
described for 22, using 6 (1.0 g, 5.37 mmol), triethylamine
(0.54 g, 5.37 mmol) and cis,cis-1,3,5-cyclohexane tricar-
bonyl trichloride 18 (0.49 g, 1.8 mmol). The solvent was
evaporated in vacuo, yielding a white waxy solid. Stirring
in hot ethanol to remove the Et3N$HCl salts followed by fil-
tration afforded 24 as a white powder (0.94 g, 1.3 mmol,
4.2.14. cis,cis-1,3,5-Chex(Am-Gly-O-octyl)3 (21). Com-
pound 21 was synthesised following the same procedure as
described for 19, using 3 (1.53 g, 8.2 mmol), triethylamine
(0.83 g, 8.2 mmol) and cis,cis-1,3,5-cyclohexane tricarbonyl
trichloride 18 (0.74 g, 2.73 mmol). After drying of the
solution over MgSO4 and evaporation of the solvent invacuo,
21 was obtained as an analytically pure white solid (1.67 g,
1
72%). Mp: >210 ꢀC (dec); H NMR (DMSO-d6, 80 ꢀC):
3
d¼7.65 (br t, J¼5.5 Hz, 3H), 7.46 (br s, 3H), 3.64 (d,
3J¼5.5 Hz, 6H), 3.09–3.02 (m, 6H), 2.29 (br t,
1
2
2.4 mmol, 88%). Mp: >210 ꢀC (dec); H NMR (DMSO-
3J¼12.1 Hz, 3H), 1.88 (br d, J¼12.5 Hz, 3H), 1.48–1.19
3
3
3
d6): d¼8.21 (t, 3H, J¼5.7 Hz), 4.00 (t, 6H, J¼6.6 Hz),
(m, 39H), 0.86 (t, J¼6.4 Hz, 9H); 13C NMR (DMSO-d6,
3
3
3.77 (d, 6H, J¼5.8 Hz), 2.27 (br t, 3H, J¼12.3 Hz), 1.77
80 ꢀC): d¼174.06, 168.25, 42.37, 41.84, 38.23, 31.10,
30.77, 28.67, 28.23, 28.11, 25.95, 21.54, 13.33; elemental
analysis calcd (%) for C39H72N6O6 (721.1): C 64.95, H
10.07, N 11.66; found: C 64.82, H 10.21, N 11.54.
3
(br d, 3H, J¼12.8 Hz), 1.53–1.24 (m, 39H), 0.84 (t, 9H,
3J¼6.6 Hz); 13C NMR (DMSO-d6): d¼174.63, 169.94,
64.28, 42.24, 40.56, 31.22, 28.59, 28.09, 25.29, 22.07,
13.95; IR (1-octanol gel, NaCl): ~n¼1730 (C]O, ester),
1644 (C]O, amide I), 1548 cmꢁ1 (N–H, amide II); elemen-
tal analysis calcd (%) for C39H69N3O9 (724.0): C 64.70, H
9.61, N 5.80; found: C 64.66, H 9.67, N 5.64.
4.3. Gelation experiments
In a typical gelation experiment a weighed amount of the
compound under investigation and 0.5 mL or 1.0 mL of
the solvent are placed in a closed 1.5 mL vial (B12 mm).
The vial was heated using a heating gun or a heating block
until the solid had dissolved, unless the solvent started to re-
flux prior to dissolution. The solution was allowed to cool to
room temperature and subsequently examined. Gelation was
considered to have occurred when a homogeneous substance
was obtained that exhibited no gravitational flow. The exis-
tence of a viscous solution was determined by eye.
4.2.15. cis,cis-1,3,5-Chex(Am-L-Phe-NH-octyl)3 (22). To
a cooled solution of 4 (1.0 g, 3.6 mmol) and triethylamine
(0.36 g, 3.6 mmol) in dry CH2Cl2 (50 mL) was added a solu-
tion of cis,cis-1,3,5-cyclohexane tricarbonyl trichloride 18
(0.33 g, 1.2 mmol) in dry CH2Cl2 (5 mL). A viscous, turbid
mixture was formed, which was slowly brought to room tem-
perature and stirred for 20 h. The solvent was evaporated in
vacuo, yielding a white solid. Stirring in ethanol to remove
the Et3N$HCl salts followed by filtration afforded 22 as a
white powder (1.1 g, 1.11 mmol, 92%). Mp: >215 ꢀC (dec);
4.4. Dropping ball measurements21
3
1H NMR (CDCl3+5% TFA): d¼7.68 (d, J¼8.4 Hz, 3H),
7.24–7.05 (m, 15H), 6.59 (s, 3H), 4.73–4.65 (m, 3H),
Gels with a volume of 1.0 mL were prepared as described
above. A stainless steel ball (63 mg; B2.5 mm) was placed
on top of the gel and the vial was closed. A series of these
samples were placed in a heating block that was slowly
heated (5 ꢀC hꢁ1) while observing the positions of the balls
with a video camera and concurrently monitoring the tem-
perature by means of a thermocouple placed in the heating
block. The melting temperature of the gel was taken as the
temperature at which the steel ball reached the base of the
flask. The upper temperature at which the apparatus could
run was limited to 155 ꢀC.
3.17–3.11 (m, 3H), 3.00–2.90 (m, 9H), 2.31 (t, 3J¼11.5 Hz,
2
2
3H), 1.78 (d, JAB¼12.5 Hz, 3H), 1.45 (dt, JAB¼12.5 Hz,
3JAB¼11.5 Hz, 3H), 1.22–1.02 (m, 36H), 0.82 (t,
3J¼6.8 Hz, 9H); 13C NMR (CDCl3+5% TFA): d¼176.54,
172.17, 134.43, 128.92, 128.89, 127.69, 55.73, 42.59,
40.68, 38.20, 31.65, 30.12, 28.93, 28.89, 28.21, 26.39,
22.50, 13.83; IR (tetraline gel, NaCl): ~n¼3285 (N–H),
1640 (C]O, amide I), 1541 cmꢁ1 (N–H, amide II); elemen-
tal analysis calcd (%) for C60H90N6O6 (991.4): C 72.68, H
9.16, N 8.48; found: C 72.20, H 9.25, N 8.37.
4.2.16. cis,cis-1,3,5-Chex(Am-L-Leu-NH-octyl)3 (23).
Compound 23 was synthesised following the same procedure
as described for 22, using 5 (0.98 g, 4.0 mmol), triethylamine
(0.36 g, 3.6 mmol) and cis,cis-1,3,5-cyclohexane tricarbonyl
4.5. Differential scanning calorimetry (DSC)
For DSC measurements a weighted amount of solid together
with a weighted amount of solvent was placed in a large

7300
M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
(g) Gronwald, O.; Shinkai, S. Chem.—Eur. J. 2001, 7, 4328;
(h) Amaike, M.; Kobayashi, H.; Shinkai, S. Chem. Lett. 2001,
620; (i) Jung, J. H.; Shinkai, S.; Shimizu, T. Chem.—Eur. J.
2002, 8, 2684; (j) Kobayashi, H.; Friggeri, A.; Koumoto, K.;
Amaike, M.; Shinkai, S.; Reinhoudt, D. N. Org. Lett. 2002, 4,
1423; (k) John, G.; Jung, J. H.; Masuda, M.; Shimizu, T.
Langmuir 2004, 20, 2060; (l) Asai, M.; Sugiyasu, K.; Fujita,
N.; Shinkai, S. Chem. Lett. 2004, 33, 120.
volume stainless steel DSC pan (for compound 19 a weighted
amount of gel was used and the sealing Viton O-ring was not
placed in the pan) and heated at 140 ꢀC (19) or 175 ꢀC (16, 22
and 24) for 1 h, followed by cooling to 25 ꢀC (cooling rate of
10 ꢀC minꢁ1). After ageing for 5 h, heating and cooling scans
were recorded on a Perkin–Elmer DSC 7 instrument from
25 ꢀC up to 140 ꢀC (19) or 175 ꢀC (16, 22 and 24) at a scan
rate of 5 ꢀC minꢁ1. For each sample two sets of heating
and cooling scans were recorded, with an interval of 2–3 h.
3. (a) Hanabusa, K.; Yamada, M.; Kimura, M.; Shirai, H. Angew.
Chem., Int. Ed. 1996, 35, 1949; (b) Hanabusa, K.; Shimura, K.;
Hirose, K.; Kimura, M.; Shirai, H. Chem. Lett. 1996, 885; (c)
Kato, T.; Kondo, G.; Hanabusa, K. Chem. Lett. 1998, 193; (d)
Kato, T.; Kutsuna, T.; Hanabusa, K.; Ukon, M. Adv. Mater.
1998, 10, 606; (e) Inoue, K.; Ono, Y.; Kanekiyo, Y.;
Hanabusa, K.; Shinkai, S. Chem. Lett. 1999, 429; (f)
Hanabusa, K.; Maesaka, Y.; Suzuki, M.; Kimura, M.; Shirai,
H. Chem. Lett. 2000, 1168; (g) Hanabusa, K.; Nakayama, H.;
Kimura, M.; Shirai, H. Chem. Lett. 2000, 1070; (h) Suzuki,
M.; Yumoto, M.; Kimura, M.; Shirai, H.; Hanabusa, K.
Chem.—Eur. J. 2003, 9, 348; (i) Suzuki, M.; Yumoto, M.;
Kimura, M.; Shirai, H.; Hanabusa, K. Helv. Chim. Acta 2004,
87, 1; (j) Suzuki, M.; Yumoto, M.; Kimura, M.; Shirai, H.;
Hanabusa, K. Tetrahedron Lett. 2004, 45, 2947.
4.6. Transmission electron microscopy
Gels were prepared as described above. Collidon- and car-
bon-coated 400 mesh copper grids were prepared, following
standard procedures. A tiny piece of gel was carefully placed
on a grid, dried at low pressure (<10–5 Torr) and shadowed
with platinum (angle: 30ꢀ–40ꢀ, distance:w15 cm). The sam-
ples were examined in a JEOL 1200 EX transmission elec-
tron microscope operating at 80 kV. First patches of gel
were searched for, to be sure that the observed structures
originate from the gel. Micrographs were taken from the
periphery of the gel patches.
4. (a) Menger, F. M.; Venkatasubban, K. S. J. Org. Chem. 1978,
43, 3413; (b) Menger, F. M.; Yamasaki, Y.; Catlin, K. K.;
Nishimi, T. Angew. Chem., Int. Ed. Engl. 1995, 34, 585; (c)
Menger, F. M.; Caran, K. L. J. Am. Chem. Soc. 2000, 122,
11679.
Acknowledgements
P. Oosting is acknowledged for preparing compounds 14–15.
Financial support from the Netherlands Organization for
Scientific Research (NWO-CW) and BiOMaDe Technology
Foundation is gratefully acknowledged.
5. (a) van Esch, J.; Kellogg, R. M.; Feringa, B. L. Tetrahedron
Lett. 1997, 38, 281; (b) van Esch, J.; De Feyter, S.; Kellogg,
R. M.; De Schryver, F.; Feringa, B. L. Chem.—Eur. J. 1997,
3, 1238; (c) van Esch, J.; Schoonbeek, F.; de Loos, M.;
Kooijman, H.; Spek, A. L.; Kellogg, R. M.; Feringa, B. L.
Chem.—Eur. J. 1999, 5, 937; (d) Schoonbeek, F. S.; van
Esch, J. H.; Wegewijs, B.; Rep, D. B. A.; de Haas, M. P.;
Klapwijk, T. M.; Kellogg, R. M.; Feringa, B. L. Angew.
Chem., Int. Ed. 1999, 38, 1393; (e) Schoonbeek, F. S.; van
Esch, J. H.; Hulst, R.; Kellogg, R. M.; Feringa, B. L.
Chem.—Eur. J. 2000, 6, 2633; (f) de Loos, M.; Ligtenbarg,
A. G. J.; van Esch, J.; Kooijman, H.; Spek, A. L.; Hage, R.;
Kellogg, R. M.; Feringa, B. L. Eur. J. Org. Chem. 2000,
3675; (g) Brinksma, J.; Feringa, B. L.; Kellogg, R. M.;
Vreeker, R.; van Esch, J. Langmuir 2000, 16, 9249; (h) de
Loos, M.; Friggeri, A.; van Esch, J.; Kellogg, R. M.; Feringa,
B. L. Org. Biomol. Chem. 2005, 3, 1631.
6. (a) van Bommel, K. J. C.; van Esch, J. H.; de Loos, M.; Heeres,
A.; Feringa, B. L. Pat. Appl. WO 03/097587, 2002; (b) Heeres,
A.; van der Pol, C.; Stuart, M.; Friggeri, A.; Feringa, B. L.; van
Esch, J. J. Am. Chem. Soc. 2003, 125, 14252; (c) van Bommel,
K. J. C.; van der Pol, C.; Muizebelt, I.; Friggeri, A.; Heeres, A.;
Meetsma, A.; Feringa, B. L.; van Esch, J. Angew. Chem., Int.
Ed. 2004, 43, 1663.
Supplementary data
Supplementary data associated with this article can be found
in the online version, at doi:10.1016/j.tet.2007.02.066.
References and notes
1. (a) Terech, P.; Weiss, R. G. Chem. Rev. 1997, 97, 3133; (b) van
Esch, J.; Schoonbeek, F. S.; de Loos, M.; Veen, E. M.; Kellogg,
R. M.; Feringa, B. L. Supramolecular Science: Where It Is and
Where It Is Going; NATO ASI Series C: Mathematical and
Physical Sciences; Kluwer Academic: Dordrecht, 1999; Vol.
527, p 233; (c) Abdallah, D. J.; Weiss, R. G. Adv. Mater.
2000, 12, 1237; (d) van Esch, J. H.; Feringa, B. L. Angew.
Chem., Int. Ed. 2000, 39, 2263; (e) Estroff, L.; Hamilton,
A. D. Chem. Rev. 2004, 104, 1201; (f) de Loos, M.; Feringa,
B. L.; van Esch, J. H. Eur. J. Org. Chem. 2005, 17, 3615; (g)
Molecular Gels; Terech, P., Weiss, R. G., Eds.; Springer:
New York, NY, 2006.
7. Fan, E.; Yang, J.; Geib, S. J.; Stoner, T. C.; Hopkins, M. D.;
Hamilton, A. D. J. Chem. Soc., Chem. Commun. 1995,
1251.
8. (a) Hanabusa, K.; Kawakami, A.; Kimura, M.; Shirai, H. Chem.
Lett. 1997, 191; (b) Shikata, T.; Ogata, D.; Hanabusa, K. J. Soc.
Rheol. Jpn. 2003, 31, 229.
9. Lightfood, M. P.; Mair, F. S.; Pritchard, R. G.; Warren, J. E.
Chem. Commun. 1999, 1945.
10. (a) Hanabusa, K.; Koto, C.; Kimura, M.; Shirai, H.; Kakehi, A.
Chem. Lett. 1997, 429; (b) Ogata, D.; Shikata, T.; Hanabusa, K.
J. Phys. Chem. 2004, 108, 15503.
2. (a) Amanokura, N.; Yoza, K.; Shinmori, H.; Shinkai, S.;
Reinhoudt, D. N. J. Chem. Soc., Perkin Trans. 2 1998, 2585;
(b) Yoza, K.; Amanokura, N.; Ono, Y.; Akao, T.; Shinmori,
H.; Takeuchi, M.; Shinkai, S.; Reinhoudt, D. N. Chem.—Eur.
J. 1999, 5, 2722; (c) Amanokura, N.; Kanekiyo, Y.; Shinkai,
S.; Reinhoudt, D. N. J. Chem. Soc., Perkin Trans. 2 1999,
1995; (d) Luboradzki, R.; Gronwald, O.; Ikeda, M.; Shinkai,
S.; Reinhoudt, D. N. Tetrahedron 2000, 56, 9595; (e)
Gronwald, O.; Sakurai, K.; Luboradzki, R.; Kimura, T.;
Shinkai, S. Carbohydr. Res. 2001, 331, 307; (f) Gronwald,
O.; Shinkai, S. J. Chem. Soc., Perkin Trans. 2 2001, 1933;

M. de Loos et al. / Tetrahedron 63 (2007) 7285–7301
7301
11. Yasuda, Y.; Iishi, E.; Inada, H.; Shirota, Y. Chem. Lett. 1996,
575.
19. For p–p-stacking effects in related aggregating C3-symmetric
compounds see Refs. 7 and 12.
12. van Gorp, J. J.; Vekemans, J. A. J. M.; Meijer, E. W. J. Am.
Chem. Soc. 2002, 124, 14759.
20. de Loos, M.; van Esch, J.; Kellogg, R. M.; Feringa, B. L.
Angew. Chem., Int. Ed. 2001, 40, 613.
13. Yamada, N.; Okuyama, K.; Serizawa, T.; Kawasaki, M.;
Oshima, S. J. Chem. Soc., Perkin Trans. 2 1996, 2707.
14. March, J. Advanced Organic Chemistry, 4th ed.; Wiley:
New York, NY, 1992; p 903.
15. According to personal experience, see also: Curtius, T. J. Prakt.
Chem. 1916, 91, 88.
21. Takahashi, A.; Sakai, M.; Kato, T. Polym. J. 1980, 12,
335.
22. Due to the absence of the O-ring, the DSC pans are less well
sealed and thus less able to resist the building up of pressure
upon heating the solvent. Because of the resulting slight leak-
age the base of the curve is not straight.
_
16. March, J. Advanced Organic Chemistry, 4th ed.; Wiley:
New York, NY, 1992; pp 1091–1092.
23. Jadz_yn, J.; Stockhausen, M.; Zywucki, B. J. Phys. Chem. 1987,
91, 754.
€
17. Knolker, H.-J.; Braxmeier, T. Synlett 1997, 925.
18. Charton, M. J. Am. Chem. Soc. 1975, 97, 1552.
24. Fibre morphology is often solvent dependent; see e.g., cyclo-
hexane bis-urea gelators (Ref. 5c).