Dynamic supramolecular polymers based on benzene-1, 3, 5-tricarboxamides: The influence of amide connectivity on aggregate stability and amplification of chirality

Article abstract of DOI:10.1002/chem.200902635

N-Centred benzene-1, 3, 5-tri- carboxamides (N-BTAs) composed of chiral and achiral alkyl substituents were synthesised and their solid-state behaviour and self-assembly in dilute alkane solutions were investigated. A combination of differential scanning calorimetry (DSC), polarisation optical microscopy (POM) and X-ray diffraction revealed that the chiral N-BTA derivatives with branched 3, 7-dimethyl- octanoyl chains were liquid crystalline and the mesophase was assigned as Col_ho. In contrast, N-BTA derivatives with linear tetradecanoyl or octanoyl chains lacked a mesophase and were obtained as crystalline compounds. Variable-temperature infrared spectrosco- py showed the presence of threefold, intermolecular hydrogen bonding be-tween neighbouring molecules in the mesophase of the chiral N-BTAs. In the crystalline state at room temperature a more complicated packing between the molecules was observed. Ultraviolet and circular dichroism spec- troscopy on dilute solutions of N-BTAs revealed a cooperative self-assembly behaviour of the N-BTA molecules into supramolecular polymers with preferred helicity when chiral alkyl chains were present. Both the sergeants-and- soldiers as well as the majority-rules principles were operative in stacks of N-BTAs. In fact, the self-assembly of N-BTAs resembles closely that of their carbonyl (C=O)-centred counterparts, with the exception that aggregation is weaker and amplification of chirality is less pronounced. The differences in the self-assembly of N- and C=O-BTAs were analysed by density functional theory (DFT) calculations. These reveal a substantially lower interaction energy between the monomeric units in the supramolecular polymers of N- BTAs. The lower interaction energy is due to the higher energy penalty for rotation around the Ph - NH bond compared to the Ph - CO bond and the diminished magnitude of dipole-dipole interactions. Finally, we observed that mixed stacks are formed in dilute solution when mixing N-BTAs and C=O BTAs.

Full text of DOI:10.1002/chem.200902635

DOI: 10.1002/chem.200902635
Dynamic Supramolecular Polymers Based on Benzene-1,3,5-tricarboxACTHNUTRGNEUaGN mides:
The Influence of Amide Connectivity on Aggregate Stability and
Amplification of Chirality
Patrick J. M. Stals,[a] Jeffrey C. Everts,[a] Robin de Bruijn,[a] Ivo A. W. Filot,[b]
Maarten M. J. Smulders,[a] Rafael Martꢀn-Rapffln,[a] Evgeny A. Pidko,[b]
Tom F. A. de Greef,[b] Anja R. A. Palmans,*[a] and E. W. Meijer*[a, b]
810
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 810 – 821

FULL PAPER
Abstract: N-Centred benzene-1,3,5-tri-
carboxamides (N-BTAs) composed of
chiral and achiral alkyl substituents
were synthesised and their solid-state
behaviour and self-assembly in dilute
alkane solutions were investigated. A
combination of differential scanning
calorimetry (DSC), polarisation optical
microscopy (POM) and X-ray diffrac-
tion revealed that the chiral N-BTA
tween neighbouring molecules in the
mesophase of the chiral N-BTAs. In
the crystalline state at room tempera-
ture a more complicated packing be-
tween the molecules was observed. Ul-
traviolet and circular dichroism spec-
troscopy on dilute solutions of N-BTAs
revealed a cooperative self-assembly
behaviour of the N-BTA molecules
into supramolecular polymers with pre-
ferred helicity when chiral alkyl chains
were present. Both the sergeants-and-
soldiers as well as the majority-rules
principles were operative in stacks of
N-BTAs. In fact, the self-assembly of
N-BTAs resembles closely that of their
carbonyl (C=O)-centred counterparts,
with the exception that aggregation is
weaker and amplification of chirality is
less pronounced. The differences in the
self-assembly of N- and C=O-BTAs
were analysed by density functional
theory (DFT) calculations. These
reveal a substantially lower interaction
energy between the monomeric units
in the supramolecular polymers of N-
BTAs. The lower interaction energy is
due to the higher energy penalty for
derivatives with branched 3,7-dimeth
ACHTUNGERTNyNUNG l-
ACHTUNGTRENNUNGoctanoyl chains were liquid crystalline
and the mesophase was assigned as
Colho. In contrast, N-BTA derivatives
with linear tetradecanoyl or octanoyl
chains lacked a mesophase and were
obtained as crystalline compounds. Var-
iable-temperature infrared spectrosco-
py showed the presence of threefold,
intermolecular hydrogen bonding be-
ꢀ
rotation around the Ph NH bond com-
ꢀ
pared to the Ph CO bond and the di-
minished magnitude of dipole–dipole
interactions. Finally, we observed that
mixed stacks are formed in dilute solu-
tion when mixing N-BTAs and C=O
BTAs.
Keywords: chirality · circular di-
chrosim · helical compounds · self-
assembly · synthesis
Introduction
the alkyl side chain.[5] A detailed study of the self-assembly
process with UV and circular dichroism (CD) spectroscopy
revealed a strongly cooperative mechanism in BTA self-as-
sembly.[5b]
Benzene-1,3,5-tricarboxamides (BTAs) exhibit thermotropic
liquid crystalline behaviour,[1] form organogels[2] and lead to
nanostructured materials,[3] depending on the exact nature
of the side chains employed. As a result, a variety of inter-
esting properties and potential applications is readily at
hand and fine tuning of the desired property only requires
small changes in the molecular structure of this class of
simple, and easily accessible organic molecules. For carbonyl
(C=O)-centred BTAs, studies in the crystalline state re-
vealed threefold intermolecular hydrogen bonds between
neighbouring molecules, resulting in the formation of colum-
nar structures.[4] The hydrogen bonds were organised in a
helical array as a result of the out of plane rotation of the
amide bond with respect to the central benzene ring. In
dilute alkane solutions the hydrogen bonds were retained
and columnar helical supramolecular polymers with a pre-
ferred helicity were formed by introducing a chiral centre in
While a significant amount of knowledge has become
available with respect to studying and describing the self-as-
sembly behaviour of alkyl-substituted C=O-centred BTAs,
their structure–property relationship and the origin of the
observed cooperativity has not received a great deal of at-
tention. Predictive tools for the rational design of molecules
that self-assemble into nano-objects of defined size, shape
and stability by the molecular information stored in the
chemical structure is highly important to push self-assembly
one step further and to arrive at self-organisation and (cata-
lytic) function. Therefore, we recently started a comprehen-
sive effort to combine theoretical and experimental work to
ultimately arrive at an in-depth molecular understanding of
the BTA self-assembly into dynamic supramolecular poly-
mers, a system that due to its simplicity serves as an ideal
model to understand the self-assembly of more complex
molecules.
We here report on the effect of a single-point mutation in
the BTA design, namely changing the amide connectivity in
this system, on its self-assembly behaviour. That amide con-
nectivity in BTAs plays an important role was recently
shown in the development of nucleating agents for isotactic
polypropylene (iPP).[6] A systematic study of nitrogen-cen-
tred (N-centred) and C=O-centred BTAs revealed that the
crystallisation temperature of iPP was significantly increased
when small amounts of BTAs were mixed in. The amide
connectivity proved to play a prominent role in the nucleat-
ing and clarifying properties of the BTAs: the best perform-
ing BTA was an N-centred derivative that showed a superior
[a] P. J. M. Stals, J. C. Everts, R. de Bruijn, M. M. J. Smulders,
Dr. R. Martꢁn-Rapffln, Dr. A. R. A. Palmans, Prof. Dr. E. W. Meijer
Laboratory of Macromolecular and Organic Chemistry
Eindhoven University of Technology
PO Box 513, 5600 MB Eindhoven (The Netherlands)
Fax : (+31)402451036
E-mail: E.W.Meijer@tue.nl
a.palmans@tue.nl
[b] I. A. W. Filot, Dr. E. A. Pidko, Dr. T. F. A. de Greef,
Prof. Dr. E. W. Meijer
Institute for Complex Molecular Systems
Eindhoven University of Technology
PO Box 513, 5600 MB Eindhoven (The Netherlands)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200902635.
Chem. Eur. J. 2010, 16, 810 – 821
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
811

A. R. A. Palmans, E. W. Meijer et al.
clarity and almost no haze. In addition, for oligo(phenylene-
vinylene)-substituted BTAs[7] and gallic acid modified
BTAs,[8] which differed in amide connectivity, it was shown
that both in solution and in the solid state N-centred deriva-
tives were less prone to form ordered aggregates compared
to C=O-centred BTAs.
ACHTUNGTRENNUNG
Remarkably, the reasons for the significant differences in
behaviour of BTA derivatives that only differ in amide con-
nectivity have not been addressed. We anticipated that the
amide connectivity in alkyl-substituted BTAs 1 and 2, which
Results and Discussion
Synthesis: The synthetic approach to trialkyl substituted N-
centred BTAs is depicted in Scheme 1. The procedure is an
adaptation from literature.[8] 3,5-Dinitroaniline was first re-
serve as model systems, may also have significant effects on
the self assembly behaviour in dilute alkane solution with
respect to the aggregate stability and the degree to which
amplification of chirality takes place. We here show the syn-
thesis and characterisation of alkyl-substituted N-centred
BTAs 1a–d. Characterisation of these new compounds in
bulk was conducted with IR spectroscopy, polarisation opti-
cal microscopy (POM), differential scanning calorimetry
(DSC) and X-ray diffraction. The properties of dilute
alkane solutions of 1a–d were investigated with UV/Vis and
CD spectroscopy. To compare the results presented here
with results previously found for C=O-centred BTAs 2a–b
(Figure 1), we include those data. The origin of the differ-
Scheme 1. Synthesis of N-centred BTAs 1a–d.
acted with acid chlorides 3a–d (octanoyl chloride, tridecano-
yl chloride and (R)- or (S)-3,7-dimethyloctanoyl chloride, re-
spectively) affording 4a–d. We selected alkyl groups which
differed in alkyl chain length (C13H27 and C7H15), in branch-
ing (3,7-dimethyloctanoyl) and in chirality ((R)-3,7-dimeth-
ACHUTNGRENyNUG lACTHUGNRTENNUoG ctanoyl and (S)-3,7-dimethyloctanoyl) to asses the impact
of side chains on the thermal behaviour and self-assembly
properties. Catalytic reduction of the nitro group with H2/Pd
afforded the diamino derivatives 5a–d. Compounds 5a–d
were reacted with the acid chlorides 3a–d to afford the de-
sired symmetrically substituted N-centred BTAs 1a–d. After
column chromatography, compounds 1a–d were obtained in
1
high purity as evidenced by H and 13C NMR spectroscopy
and Maldi-Tof-MS.
Solid-state properties of N-centred BTAs 1a–d: The effect
of amide connectivity in BTAs was first evaluated in the
solid state by employing DSC and POM; the data obtained
were compared to those of C=O-BTA reference compounds
2a,b. The thermal transitions observed in the second heating
run in DSC are summarised in Table 1.
Figure 1. Texture obtained for compound 1c after slow cooling from the
isotropic state; T=2248C.
812
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 810 – 821

Dynamic Supramolecular Polymers
FULL PAPER
Table 1. Summary of the DSC and IR data of BTAs 1a–d and 2a,b.
intermolecular hydrogen bonding between neighbouring
molecules, while vibrations for the NH stretch at 3306 cmꢀ1
and an amide II vibration at 1530 cmꢀ1are typical for a non-
helical organisation of the hydrogen bonds.[1d]
The IR data measured of virgin samples of N-centred
BTAs 1a–d, that is, after isolation and without any thermal
treatment, are summarised in Table 1. In all cases, the NH
stretch is found at 3270 cmꢀ1 while two C=O stretch vibra-
tions at about 1656 and 1615 cmꢀ1 and an amide II vibration
at approximately 1550 cmꢀ1 are observed. To visualise the
differences between chiral 1d and 2b, Figure 2 compares the
[b]
[b]
[b]
T [8C][a]
(DH
[kJmolꢀ1])[a]
nNH
[cmꢀ1
nC
nCꢀ
N
=
O
G
R
G
]
[cmꢀ1
]
[cmꢀ1
ACHTUNGTERNNUNG ]
1a
1b
K 174.9 (41.2) I
3271
3278
1656/1614
1649/1607
1665/1605[d]
1656/1614
1651[d]
1550
1551
K1 94.6 (23.1) K2 149.5 (38.1) I
3310[d]
3269
1536[d]
1550
1c
Colho 215.1 (15.3) I
3250[d]
3272
1523[d]
1553
1d
Colho 218.0 (15.9) I
1656/1615
3236[e]
3306[f]
3223[f]
1640[e]
1557[e]
1531[f]
1564[f]
2a[c]
2b[c]
K 19 (17) Colho 198 (8) I
K 120 (13) Colho 235 (18) I
1644[f]
1637[f]
[a] Data derived from the second heating run in DSC. [b] Measured at
room temperature for an untreated sample. [c] Data obtained from refer-
ence [1d]. [d] IR data obtained at T=1758C. [e] IR data obtained direct-
ly after heating into the mesophase and fast cooling to RT. [f] IR data ob-
tained at room temperature, for 2a this corresponds to a non-helical ar-
rangement of the hydrogen bonds.
Previous research indicated that starting from a hexyl side
chain, C=O-centred BTAs 2a,b are liquid crystalline (LC)
and that the LC phase is an ordered, columnar hexagonal
phase (Colho) (Table 1).[1d] In contrast to the C=O-centred
BTAs, POM investigations on N-centred BTAs revealed
that unbranched derivatives do not form a mesophase. Com-
pounds 1a and 1b are crystalline solids that melt into an iso-
tropic liquid (I) at 174.9 and 149.58C, respectively. Increas-
ing the bulkiness of the alkyl side chain is sufficient to
induce an LC phase. Branched 1c and 1d exhibit a liquid
crystalline phase from 91 to 2248C as evidenced by POM
(Figure 1) and by the first heating run in DSC (ESI, Fig-
ure S1). Interestingly, the transition at 918C does not reap-
pear in the second heating run (Table 1), but a transition at
around 88C is observed, while the LC-to-I transition is ob-
served at 215–2188C. The isotropisation temperature and
concomitant enthalpy of branched N-centred BTA 1c,d is
similar to that of its C=O-centred counterpart 2b.
Figure 2. Comparison of the IR spectra of 1d and 2b at room tempera-
ture.
room temperature IR spectra of 2b (in which threefold hy-
drogen bonding in a helical packing is present) and 1d.
There is less symmetry in the NH stretch vibration of 1d
and the splitting of the C=O vibration suggests that the hy-
drogen bonds are organised in a non-helical way. In fact, the
IR trace of 1d resembles the IR trace previously observed
for compound 2a at room temperature, where a non-helical
arrangement of the hydrogen bonds was assigned.[1d]
We then performed variable-temperature IR spectra of
compound 1b and 1d in order to investigate differences in
the hydrogen bonds as a function of temperature (see
Table 2 and Figure S2 in the Supporting Information). Com-
pound 1b is a crystalline solid that melts at 149.58C. The IR
traces remain identical when heating the sample until the
melting temperature is reached (Figure S2A in the Support-
ing Information). At 1758C, well above the melting point,
the NH stretch is at 3310 cmꢀ1 and the amide II is at
1536 cmꢀ1 indicating the loss of hydrogen bonding (Table 1).
In contrast, the NH stretch in case of 1d lowers to a value
of 3250 cmꢀ1 at 1758C, while the carbonyl vibration merges
X-ray diffraction measurements on a sample of (S)-1c un-
ambiguously confirmed a hexagonally ordered, columnar
mesophase (Colho) phase. At 1608C two features are detect-
ed in the wide-angle region: first, a sharp reflection that cor-
responds to a distance of 3.5 Š and is characteristic of the
stacking of disc-like molecules; second, a broad diffuse halo,
associated to a distance of about 5.0 Š, that arises from
short-range interactions between the aliphatic chains as is
typically observed in liquid crystalline phases. In the small-
angle region, a set of three reflections at 16.3, 9.3 and 8.0 Š
p
with spacings in the reciprocal ratio 2: 3:1 is consistent
with a hexagonal packing of the columns. From these
maxima we can deduce an intercolumnar distance of 18.8 Š.
More information on the organisation of the hydrogen
bonds in the solid state of N-centred BTAs can be obtained
by analysing the IR data. For C=O-centred BTAs, we estab-
lished that IR spectroscopy is a sensitive technique to evalu-
Table 2. Thermodynamic parameters of 1c and 2b (c=3·10ꢀ5 m in MCH)
obtained by fitting the temperature-dependent data.
ate the nature of the intermolecular hydrogen bonding. Vi-
he [kJmolꢀ1
]
Te [K]
Ka
ꢀ1
ꢀ
brations for the N H stretch at ꢁ3240 cm , the C=O
1c
2b
ꢀ50
ꢀ60
328.5
336.6
2”10ꢀ5
stretch at ꢁ1640 cmꢀ1 and the amide II at ꢁ1560 cmꢀ1 have
been attributed to the presence of threefold, a-helical type
ꢂ10ꢀ5/10ꢀ6
Chem. Eur. J. 2010, 16, 810 – 821
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
813

A. R. A. Palmans, E. W. Meijer et al.
into one peak at 1651 cmꢀ1 (Figure S2B in the Supporting
Information). This observation, together with the fact that a
Colho phase is present between 91 and 2248C in the first
heating run, strongly points to a a-helical type packing of
the hydrogen bonds in the liquid crystalline state of N-cen-
tred BTAs similar to the packing observed in C=O-centred
BTA 2b. After cooling to room temperature, the helical ar-
rangement is retained but over time it slowly changes back
into the non-helical packing pattern.
The data presented above indicate that N-centred BTAs
have a complicated solid-state behaviour that depends
strongly on the nature of the alkyl chain. Presumably, pack-
ing effects and maximisation of the crystal density govern
the hydrogen bonding patterns in the solid state, an effect
that was also observed previously for C=O-BTAs with short
alkyl chains.[1d] However, there are strong indications from
IR in the Colho mesophase that the molecules organise into
columns stabilised by threefold intermolecular hydrogen
bonds. Unfortunately, to date, there is no crystal structure
available for an N-centred derivative to validate this hypoth-
esis. Therefore, we proceeded with studying the N-centred
BTAs in dilute solution.
Self-assembly of N-centred BTAs in methylcylohexane and
methanol: The self-assembly of N-centred BTAs into dy-
namic supramolecular polymers was investigated by UV and
CD spectroscopy in methylcyclohexane (MCH) and in
methanol (c=3·10ꢀ5 m). In methanol, N-centred BTAs are
molecularly dissolved, while MCH favours stacking of the
BTAs. This is evidenced by a strong Cotton effect in MCH
(Figure 3A), while no Cotton effect is present in methanol.
Moreover, the UV spectra in methanol and MCH differ: the
lmax in MCH is at 223 nm, while that in methanol is at
240 nm (Supporting Information, Figure S3). The blue shift
upon aggregation has been observed previously for C=O-
centred BTAs and is attributed to a decrease in conjugation
between the benzene unit and the amide substituent, be-
cause of hydrogen bond formation. Mirror image CD spec-
tra were obtained for the enantiomer pair 1c,d in MCH
with De=ꢀ63.1 Lmolꢀ1 cm for 1c and +62.3 Lmolꢀ1 cm for
1d at l=233 nm (Figure 3B). These results point to aggre-
gation of the different enantiomers in helical supramolecular
polymers of opposite helicity.
Figure 3. A) UV and CD spectra of 1c in MCH, c=3·10ꢀ5 m. B) Mirror
image CD spectra of (S)-1c (black line) and (R)-1d (grey line), c=3·10ꢀ5
in MCH. C) CD cooling curve of 1c in MCH, c=3·10ꢀ5 m, l =233 nm.
Variable-temperature CD measurements were performed
on compound 1c in MCH (c=3·10ꢀ5 m). The temperature
was decreased from 90 to 208C and the CD intensity was
monitored at l =233 nm, the maximum of the Cotton
effect. The CD effect decreases with an increase in tempera-
ture and disappears at 558C. A plot of the CD intensity
measured at 233 nm versus the temperature is given in Fig-
ure 3C. The curve shows two regimes; a nucleation regime,
in which all molecules are molecularly dissolved (indicated
by the absence of a CD effect) and an elongation regime in
which the aggregate rapidly grows. Upon cooling, at a cer-
tain (concentration dependent) temperature a large enough
nucleus is formed allowing the formation of an aggregate.
Fitting the cooling curves to the nucleation-growth model of
van der Schoot affords Te (the elongation temperature) and
he (the enthalpy release upon elongation).[5b,9] Moreover,
Ka—the equilibrium constant between the monomeric and
the active state—can be derived and is a measure for the
degree of cooperativity in the system. The Te , he and Ka
data are given Table 2 and compared to data previously de-
rived for the C=O-centred BTA 2b.[5b] It is clear that the C=
O-centred BTAs are more stable and aggregation results in
a higher enthalpy release as evidenced by a higher Te and he
at identical concentration in MCH. Moreover, the aggrega-
tion of 1c appears to be less cooperative than that of 2b
(Figure S4 in the Supporting Information), but unfortunately
the values for Ka cannot be directly compared as they are
determined at a different Te.
814
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 810 – 821

Dynamic Supramolecular Polymers
FULL PAPER
Amplification of chirality in N-centred BTAs: When mole-
cules of conflicting chiral information (i.e., R and S enantio-
mers) are mixed, the so-called majority rules effect is opera-
tive in C=O-centred BTAs. The majority rules effect states
that the enantiomer present in the majority dictates the heli-
cal sense to which the minority enantiomer adjusts. In addi-
tion, adding small amounts of a chiral BTA to an achiral so-
lution consisting of equal amounts of P and M helical supra-
molecular polymers results in stacks of either P or M helici-
ty. This effect is commonly referred to as the sergeants-and-
soldiers effect. Sergeants-and-soldiers (SaS) and Majority
Rules (MR) experiments in C=O-centred BTAs showed a
pronounced amplification of chirality effect, that is, the ob-
served CD effect is much higher then expected based on the
net enantiomeric excess (for the MR experiment) or the
amount of chiral compound added (for the SaS experiment).
For the N-centred BTAs, chiral 1c, the sergeant, was ti-
trated into a solution of achiral 1b, the soldier, and the CD-
effect was followed as a function of the amount of chiral 1c
added (MCH, c=3”10ꢀ5 m). Since achiral compounds have
no preferred helicity, equal amounts of P and M helices are
present and no CD effect is measured. The result of the ti-
tration is shown in Figure 4A. A similar experiment was
then conducted using enantiomers 1c and 1d mixed in dif-
ferent ratios (Figure 4B). Clearly, in both cases the CD
effect is not linear with respect to the enantiomeric excess
(MR experiment) and the amount of sergeant added (the
SaS experiment). This shows that amplification of chirality
is indeed present in N-centred BTAs.
Theoretical models developed by van Gestel[10] and van
der Schoot[11] allow us to quantify the energies that are in-
volved in chiral amplification processes: the mismatch
energy (EM), that is, the penalty that is paid by the minority
enantiomer to be present in a helix of its non-preferred
sense and the helix reversal penalty (ER), that is, the penalty
that it cost for a helix to change its helical sense. Both ener-
gies are derived from experimental data. This model recent-
ly yielded values for EM and ER in bipyridine- and C=O-
based BTAs.[11,12] The results obtained for the N-centred
BTAs 1 are summarised in Table 3 and compared to those
Table 3. Mismatch energy penalty (EM) and helix reversal energy (ER)
for 1 and 2 in MCH ( c=3·10ꢀ5 m).
1
2
EM [kJmolꢀ1
]
ER [kJmolꢀ1
]
EM [kJmolꢀ1
]
ER [kJmolꢀ1
]
SaS
MR
>0.5[a]
1.3ꢃ0.4
7.3ꢃ2.5
12ꢃ4
>0.5[a]
1.9ꢃ0.2
12.6ꢃ2.0
15ꢃ4
[a] Fitting the SaS data only allows to determine a lower limit for EM.
obtained for C=O-centred BTA 2. Both the mismatch penal-
ty (EM =1.3 kJmolꢀ1) and the helix reversal penalty (ER =
12 kJmolꢀ1) are lower in case of the N-centred BTA com-
pared to the C=O-centred BTA (EM =1.9 kJmolꢀ1 and ER =
15 kJmolꢀ1, respectively), showing that the degree to which
amplification of chirality takes place is lower in case of N-
centred BTAs. The difference may very well be related to
the lower degree of cooperativity observed for the aggrega-
tion of N-centred BTAs. Unfortunately, to date there are
very few examples known in literature where amplification
of chirality is studied systematically in cooperative sys-
tems.[11]
Origin of the differences between N-centred and C=O-cen-
tred BTAs: Most likely, the differences in aggregate stability
can be attributed to differences in hydrogen-bonding
strength between the different BTAs. However, estimations
of the pKaꢂs of the NH in 1 and 2 in DMSO show that these
are rather similar.[13] Therefore, one cannot rationalise, a
priori, which hydrogen bond will be stronger. Because of
the unavailability of a crystal structure of an N-centred BTA
and to get more insight in the self-assembly behaviour of
the N- and C=O-centred BTAs, DFT calculations were per-
formed. In order to reduce the computational costs, the cal-
culations were preformed on oligomers and infinite chains
of BTAs 3 and 4 (Figure 5), in which the long aliphatic side
chains were replaced with by methyl groups.
Plane-wave DFT (PW-DFT) electronic-structure calcula-
tions were performed by using the VASP code, as it is well
suited for studies of large periodic systems such as the hy-
drogen-bonded chains formed by BTAs.[14] All DFT calcula-
tions were performed using the PBE exchange-correlation
Figure 4. A) Results from a SaS experiment using 1a and 1c and B) re-
sults from a MR experiment using 1c and 1d. In all cases the Cotton
effect is probed at 233 nm and c=3·10ꢀ5 m in MCH.
Chem. Eur. J. 2010, 16, 810 – 821
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
815

A. R. A. Palmans, E. W. Meijer et al.
Comparison of the structural parameters of the PW-DFT
optimised geometry of the C=O-centred infinite chain of
BTA 4 with that of the reported crystal structure of Light-
foot et al. shows excellent agreement between the calculated
and experimental structural parameters. For example, the
Ca-Ca-Cc-O dihedral angle (q) in the optimised geometry
(Figure 6) of 4 in an infinite chain is ꢀ40.88, while it is
ꢀ41.58 in the X-ray structure. The difference in hydrogen-
bond length (dNꢀO) between the calculated and experimen-
tally determined geometry is only 0.01 Š, while the differ-
ence centroid-to-centroid distance between the two aromatic
rings is somewhat larger. The small discrepancy between the
calculated and reported geometry is most probably the
result of the larger aliphatic side chain in the crystal struc-
ture compared to the smaller methyl group used in our cal-
culations.
Given the good agreement between the calculated infinite
hydrogen-bonded chain of C=O-centred BTA 4 and the re-
ported X-ray structure, we next investigated the differences
between the optimised geometries of the infinite chains of
C=O-centred BTA 4 and N-centred BTA 3. Whereas the Ca-
Ca-Ca-O dihedral angle in the infinite chain of BTA 4 is
ꢀ40.88, the Ca-Ca-N-H dihedral angle (q; Figure 5) in the in-
finite chain of N-centred BTA 3 is ꢀ38.68. Furthermore, the
hydrogen-bond length (dNꢀO) in the infinite helical chain of
N-centred BTA 3 is somewhat longer (3.01 Š) compared to
that in the infinite helical chain of C=O-centred BTA 4
(2.98 Š). This effect is more pronounced in the computed
Figure 5. N-centred BTA 3 and C=O-centred BTA 4 used in the compu-
tational studies. The torsion angle q for each BTA used to characterise
the PW-DFT geometries is also shown.
functional[15] that has been shown to reproduce hydrogen-
bond lengths and energies with high accuracy.[16] An advan-
ACHTUNGTRENNUNGtage of PW-DFT is that it is inherently free from the basis-
set superposition error (BSSE), because plane-wave basis
sets describe any point in the periodic supercell with the
same quality in contrast to Gaussian orbitals spanning only
a local region.[17]
Geometries of infinite chains of N-centred BTA 3 and C=
O-centred BTA 4 were optimised by using the PW-DFT ap-
proach. Since PW-DFT calculations are of inherently peri-
odic nature, an N- or C=O-centred BTA dimeric unit was
placed in an (a”a”c) orthorhombic supercell. A vacuum
layer (a=20 Š) was added along the two directions perpen-
dicular to the direction of the infinite chain in order to
avoid spurious interactions between the periodic images of
the system. Optimum lattice constants c equal to 7.20 and
7.40 Š for 3 and 4, respectively, were chosen from the mini-
misation of the total energy with respect to this parameter.
Figure 6 shows the optimised geometry of the infinite, hy-
ꢀ
N H···O distances (Figure 5). The longer hydrogen bond
and smaller dihedral angle of the infinite chain of N-cen-
tered BTA 3 is expected to result in a lower cohesive energy
per molecular unit compared to
the infinite chain of C=O cen-
tered BTA 4. Indeed, the elon-
gation enthalpies, a measure for
the cohesive energy per molec-
ular unit, obtained from the
analysis of the temperature-de-
pendent CD data in MCH sug-
gests that the enthalpy release
(he) for addition of a single
monoACTHNUGRTNEUNGmer is higher for stacks of
C=O centered BTAs than
stacks composed of N-centered
BTAs by 10 kJmolꢀ1.
ꢀ
Figure 6. PW-DFT optimised structures and N H···O hydrogen-bond lengths (Š) of infinite chains of C=O-
centred BTA 4 and N-centred BTA 3. Energy values presented correspond to the average interaction energy
(DE) between the monomers in the respective supramolecular ensemble.
drogen-bonded chains of BTAs 3 and 4, while the relevant
structural parameters are reported in Table 4 together with
the structural parameters from the crystal structure of a C=
O-centred BTA reported by Lightfoot et al.[4a]
This is in line with the differences in the computed aver-
age interaction energies in the supramolecular chains of 3
and 4 (Figure 6). The average interaction energy, that is co-
hesive energy, per molecular unit was evaluated as DE=
(Eidꢀ2Emon)/2, in which Eid is the computed electronic energy
of the unit cell containing two BTA monomers periodically
repeated to yield the infinite BTA chain and Emon is the
electronic energy of the monomer computed at the same
level of theory.[18] The computed cohesive energy per mole-
cule in the gas phase equals ꢀ73 kJmolꢀ1 and ꢀ41 kJmolꢀ1
for the infinite chains of C=O-centred 4 and N-centred 3
BTA, respectively. Hence, the order of stability as found
Table 4. Structural parameters of the optimised structures of infinite
chains of BTAs 3 and 4 together with the structural parameters from the
crystal structure reported by Lightfoot et al.[4a]
Structure
q [8]
dOꢀN [Š]
dring [Š]
PW-DFT infinite chain of BTA 3
PW-DFT infinite chain of BTA 4
crystal structure C=O-centred 4a
ꢀ38.6
ꢀ40.8
ꢀ41.5
3.01
2.98
2.97
3.60
3.70
3.62
816
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 810 – 821

Dynamic Supramolecular Polymers
FULL PAPER
using the PW-DFT calculations agrees with the order of sta-
bility as found from the analysis of the CD data.
As already mentioned, the pKaꢂs of the NH in the N- and
C=O-centred monomeric BTAs are very similar, indicating
that the lower cohesive energy per molecular unit for infi-
ACHTUNGTRENNUNGnite stacks of 3 with respect to 4 cannot solely be due to the
stronger hydrogen bonds formed by the BTA monomer in
the latter. Analysis of the hydrogen-bond length, which has
been claimed[19] to be a good measure of hydrogen-bond
strength, in both optimised geometries shows that the hydro-
gen bonds in the C=O-centred infinite chain are somewhat
shorter than in the N-centred infinite chain, although the
difference is only minor (0.03 Š) Therefore, we hypothe-
ACHTUNGTRENNUNGsised that the reduction in the computed cohesive energy
ꢀ
Figure 7. Potential energy profiles for torsion around the Caromatic Ccarbonyl
per molecular unit as found for the infinite stacks of N-cen-
tred BTA 3 relative to 4 is primarily due to the differences
in dipole–dipole interactions between the monomers, parti-
ally as a result of the smaller dihedral angle found in the in-
finite hydrogen bonded chains consisting of BTA 3.
ꢀ
bond of C=O-centred BTA 4 and the Caromatic N bond of N-centred BTA
3 (PBE/6-311+G
(d,p)). The energy profiles are obtained by scanning the
torsional potential surface in increments of 58 for the Ca-Ca-Cc-O dihe-
dral angle of 4 and Ca-Ca-N-H dihedral angle of 3.
To correlate the observed difference in stability between
C=O- and N-centred hydrogen-bonded supramolecular poly-
mers to the structures of their respective monomers, we
arene ortho-positions were occupied by hydrogen atoms and
the nitrogen was substituted with a hydrogen atom and a C-
reas
G
ACHTUNGTREN(UNNG sp3) carbon. They found that the average dihedral angle in
molecular polymers is caused by the higher energetic barrier
for rotation of the Ca-Ca-N-H dihedral angle in the N-cen-
tred BTA monomer compared to rotation of the Ca-Ca-Cc-O
dihedral angle in the C=O-centred BTA monomer. Upon
formation of threefold hydrogen bonds in supramolecular
polymers consisting of C=O- or N-centred BTAs, the amide
group of the monomeric BTAs has to rotate out of the
plane of the benzene ring. In order to explore the conforma-
tional change of the aromatic BTA amides 3 and 4 and the
role of amide connectivity, torsional energy profiles were
all the X-ray structures was qꢁ228.
In contrast, the equilibrium structure of N-centred BTA 3
is characterised by the amide group in the same plane as the
benzene (qꢁ08). Previous DFT studies reported by Doerk-
sen et al.[23] on the related compound acetanilide, show that
also in this structure the amide is in plane with the benzene
ring. The steric repulsion between the arene ortho-protons
and the smaller NH moiety in 3 is weaker as relative to that
involving larger C=O group in 4. In addition, the equilibri-
um conformation of N-centred BTA 4 shows an unusually
ꢀ
ꢀ
calculated for the torsions around all of the Caromatic Ccarbonyl
short O···H (C6) intramolecular distance (2.18 Š), which
bonds of BTA 4 and all of the Caromatic N bonds of BTA 3
matches known values[14] for such hydrogen bonds. Such a
weak hydrogen bond may play a role in the stabilisation of
the planar geometry of N-centred BTA 4. The observed dif-
ferences in the computed torsional potential surface for 3
and 4 are influenced by the conjugation of the amide and
arene p orbitals, steric repulsions with arene ortho-hydrogen
atoms and weak intramolecular hydrogen bonding.
ꢀ
(Figure 5). The equilibrium geometries were obtained by
full geometry optimisation at PBE/6-311G+(d,p) level of
theory by using the Gaussian 03 program.[20] Initially, tor-
sional energy profiles were obtained by scanning the tor-
sional potential surface in 58 increments at PBE/6-311G+
(d,p) level of theory. At each scan point, all internal coordi-
nates except the three torsional angles driving the scan were
optimised. The calculated potential energy curves for 3 and
4 as a function of the dihedral angle q are displayed in
Figure 7. The equilibrium geometry of C=O-centred BTA 4
already shows a pronounced nonplanarity between the aro-
matic ring and the amide group. The amide groups in 4 are
tilted out of plane by ꢁ128. The deviation from planarity
probably results from steric repulsion between the arene
ortho-hydrogen atoms and the oxygen atoms of the carbonyl
groups. DFT and MP2 calculations on the related compound
N-methyl benzamide (NMB) conducted by Pophristic
et al.[21] and Vargas et al.[22] indicate that the equilibrium ge-
ometry of NMB is also characterised by pronounced nonpla-
narity between the aromatic ring and the amide group. Fur-
thermore, Vargas et al.[22] analysed X-ray structures contain-
ing a phenyl ring attached to an amide C=O, in which both
As a result of the differences in the steric repulsion be-
tween the arene ortho-protons and the amide H and O
atoms in BTA 4 and BTA 3, respectively, rotation of the
amide in N-centred BTAs is substantially hindered as com-
pared to that in the C=O-centred BTAs (Figure 7). Indeed,
for a dihedral angle of 408, which corresponds to the value
obtained in infinite chains of BTAs 3 and 4, the N-centred
BTA is 18 kJmolꢀ1 higher in energy. Thus stronger and
more unfavourable deformations of the N-centred BTA
monomers are required upon its addition to a supramolec-
ular polymer consisting of N-centred molecules. This results
in the less favourable supramolecular polymerisation pro-
cess of the N-centred BTAs as compared to the C-centred
isomers. Furthermore, rotation of the amide groups in the
C=O-centred BTA 4 results in a somewhat larger dipole
moment as compared to that in the N-centred BTA mono-
Chem. Eur. J. 2010, 16, 810 – 821
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
817

A. R. A. Palmans, E. W. Meijer et al.
mer 3 at all values of the dihedral angle q (Figure 8). The
larger dipole in the perturbed structures of 4 enhances the
dipole–dipole interactions in the corresponding supramolec-
Figure 8. Evolution of the dipole moment in the N-centred 3 and C=O-
centred 4 BTA monomers upon the increase of the dihedral angle
ꢀ
ꢀ
around the Caromatic Ccarbonyl bond and the Caromatic NH bond, respectively
(PBE/6-311+G(d,p)).
ACHTUNGTRENNUNG
ular structures and hence further stabilises supramolecular
polymers consisting of C=O-centred BTAs as compared to
those formed from N-centred BTAs. One expects that the
nonplanarity of the monomeric BTA units gradually increas-
es at the initial stage of the supramolecular polymer growth
and finally reaches the limiting value computed for the in-
finite BTA chains. The resulting evolution of the macrodi-
pole as a function of stack size influences the cooperativity
in the growth of the resulting supramolecular polymer.[25]
We therefore speculate that a difference between the evolu-
tion of the macrodipole as a function of chain length may
explain the differences in cooperativity observed for the N-
and C=O-centred BTAs.
Mixed stacks of N-centred and C=O-centred BTAs: Both N-
and C=O-centred BTAs form columnar aggregates in dilute
solutions stabilised by means of threefold intermolecular hy-
drogen bonding, and the results discussed above indicate
that the aggregation and degree of amplification of chirality
is weaker in case of N-centred BTAs. An intriguing question
is if the different hydrogen-bond strengths can be utilised in
making alternating supramolecular block copolymers, an ob-
servation previously made in two component organogel sys-
tems in which amide- and urea-based BTAs were found to
form a more stable gel in a mixed system.[26]
To investigate the feasibility of this, we performed mixing
experiments of chiral 1c and achiral 2a and measured the
CD spectra as a function of chiral 1c added. As depicted in
Figure 9A, achiral 2a shows no CD effect, but upon addi-
tion of only 5% of 1c, there is a fast increase of the CD
effect. The shape of the CD spectrum between 0 and 20%
1c added is typical for that of a C=O-centred BTA.[5b] Start-
ing from around 20% 1c, the shape of the CD curve
changes into a CD curve typical of an N-centred BTA with
Figure 9. A) CD spectra of 1c added to achiral 2a at low fractions of 1c
(MCH, c=3·10ꢀ5 m. B) CD spectra of 1c added to achiral 2a at high frac-
tions of 1c (MCH, c=3·10ꢀ5 m). C) CD effect at 223 and 233 nm as a
function of the amount of chiral 1c added to achiral 2a (MCH, c=
3·10ꢀ5 m).
a lmax at 233 nm (Figure 9B, see the Supporting Information
for more spectra). Because of the changes in the shape of
the CD spectra it is difficult to select a wavelength at which
to probe the amplification of chirality. We selected l=223
and 233 nm, at which C=O- and N-centred BTAs have a
maximal Cotton effect, respectively, as the wavelength of
choice and plotted the CD intensity as a function of chiral
1c added. As seen in Figure 9C, there is a clear deviation
from linearity when probing the CD effect at l=223 and
233 nm. Remarkably, at l=223 nm the levelling off of the
CD effect is observed at only 10% of the added chiral com-
818
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 810 – 821

Dynamic Supramolecular Polymers
FULL PAPER
ponent and the De remains more or less constant at
ꢀ30 Lmolꢀ1 cmꢀ1. This value is in close agreement with a
fully homochiral C=O-centred BTA stack for which De
values are typically around ꢀ40 Lmolꢀ1 cmꢀ1. At l=233 nm,
the increase is more gradual and no levelling off in the CD-
effect is observed, which is quite similar to the SaS experi-
ment conducted with chiral 1c and achiral 1b. The reverse
experiment, namely adding achiral C=O-centred 2b to achi-
ral N-centred 1b, was more difficult because of the poor sol-
ubility of 1b in MCH at room temperature (see Figure S5
and S6 in the Supporting Information for additional data),
but a clear deviation from linearity was observed also in this
case. Although this is no definite proof of an alternating
supramolecular block copolymer, it is clear that mixed
stacks are formed.
cess and rationalise the decreased stability of N-BTA aggre-
gates with respect to C=O-BTA ones.
The lower cooperativity in the self-assembly of N-centred
BTAs relative to that of the C=O-centred BTAs may cause
the lower degree of amplification of chirality observed. A
more in-depth theoretical investigation on the origin of the
cooperativity and amplification of chirality in the N- and C=
O-centred BTAs will be subject of future work. We antici-
pate that the results will allow understanding these phenom-
ena at the fundamental level.
Experimental Section
Materials: (R)-(+)-Citronellic acid and (S)-(ꢀ)-citronellic acid were pur-
chased from Aldrich and converted into the corresponding (R)- and (S)-
3,7-dimethyloctanoic acid according to a described procedure.[27] Refer-
ence compounds 2a,b were prepared as described previously.[1d] All sol-
vents were obtained from Biosolve, except for DMSO (Acros) and spec-
trophotometric grade methylcylcohexane (Aldrich). All other chemicals
were obtained from either Acros or Aldrich. Dry dichloromethane was
tapped off a distillation setup which contained molsieves, CHCl3 was
dried over molsieves and triethylamine was stored on KOH pellets. All
other chemicals were used as received.
Conclusion
Several N-centred benzene-1,3,5-tricarboxamides (N-BTAs)
comprising chiral and achiral alkyl substituents were syn-
thesised and their solid-state behaviour was studied using
DSC, POM and X-ray diffraction. This revealed that the
chiral N-BTA derivatives with a branched 3,7-dimethylocta-
noyl chain are liquid crystalline and the mesophase can be
assigned as Colho. In contrast, N-BTA derivatives with octa-
noyl and tetradecanoyl side chains lack a mesophase and
are obtained as crystalline compounds. Variable-tempera-
ture IR spectroscopy confirmed the presence of threefold,
intermolecular hydrogen bonding between neighbouring
molecules in the mesophase. In the crystalline state at room
temperature, on the other hand, a more complicated organi-
sation of the hydrogen bonds between the molecules was
observed.
UV and CD spectroscopy on dilute solutions of N-BTAs
revealed that the self-assembly into supramolecular poly-
mers is cooperative. Similar to the C=O-centred BTAs, N-
BTA molecules form columnar stacks with a preferred helic-
ity when a chiral alkyl chain was used. Both the sergeants-
and-soldiers as well as the majority-rules principles are oper-
ative in stacks of N-BTAs. Interestingly, mixed stacks of N-
and C=O-BTAs are was formed in dilute solution, suggest-
ing that such a system is thermodynamically more stable
than separate stacks of N- and C=O-BTAs.
Methods: UV/Vis and circular dichroism measurements were performed
on a Jasco J-815 spectropolarimeter, for which the sensitivity, time con-
stants and scan rates were chosen appropriately. Corresponding tempera-
ture-dependent measurements were performed with a PFD-425S/15 Pelti-
er-type temperature controller with a temperature range of 263–383 K
and adjustable temperature slope, in all cases a temperature slope of
1 Kminꢀ1 was used. In all other measurements the temperature was set
at 208C. In all experiments the linear dichroism was also measured and
in all cases no linear dichroism was observed. Separate UV/Vis spectra
were obtained from a Perkin–Elmer UV/vis spectrometer Lambda 40 at
208C. Cells with an optical path length of 1 cm were employed and spec-
troscopic grade solvents were employed. Solutions were prepared by
weighing in the necessary amount of compound for a given concentration
and transferring it to a volumetric flask. Then the flask was filled three-
quarters filled with the spectroscopic grade solvent and put in an oscilla-
tion bath at 408C for 45 min, after which the flask was allowed to cool
down and filled up to its meniscus. The De=CD effect/(32980cl) in
which c is the concentration in molLꢀ1 and l is the optical path length in
cm. 1H NMR and 13C NMR measurements were conducted on a Varian
Mercury 200 MHz and/or a Varian Gemini 400 MHz. Proton chemical
shifts are reported in ppm downfield from tetramethylsilane (TMS).
Carbon chemical shifts are reported using the resonance of CDCl3 as in-
ternal standard. Maldi-TOF-MS were acquired using a Perserptive Bio-
system Voyager-DE PRO spectrometer. In all cases, a-cyano-4-hydroxy-
cinnamic acid was employed as the matrix material. IR spectra were re-
corded on a Perkin–Elmer spectrum 1 using a universal ATR. Variable-
temperature IR spectra were recorded on an Excalibur FTS 3000 MX
FT-IR from Biorad. Polarisation optical microscopy measurements were
done using a Jenaval polarisation microscope equipped with a Linkam
THMS 600 heating device, with crossed polarisers. The thermal transi-
tions were determined with DSC by using a Perkin–Elmer Pyris 1 DSC
The self-assembly of N-BTAs resembles closely that of
their C=O-centred counterparts, with the exception that ag-
gregation is weaker and amplification of chirality is less pro-
nounced. DFT calculations revealed that there is a higher
under
a nitrogen atmosphere with heating and cooling rates of
ꢀ
energy penalty for the rotation around the Ph NH bond
10 Kminꢀ1. The Tg was determined with heating and cooling rates of
40 Kminꢀ1. The XRD patterns were obtained with a pinhole camera
(Anton-Paar) operating with a point-focused Ni-filtered CuKa beam. In
this setup, the samples are held in Lindemann glass capillaries (0.9 mm
diameter). The capillary axis is perpendicular to the X-ray beam and the
pattern is collected on a flat photographic film perpendicular to the
beam. Spacings are obtained via Braggꢂs law (nl=2dsinq).
ꢀ
compared to the Ph CO bond in the monomer. Therefore,
more unfavourable deformations of the N-BTA are required
upon its addition to a growing supramolecular polymer. This
effect along with weaker dipole–dipole interactions due to
the lower dipole moment of the deformed monomers and
weaker intermolecular hydrogen bonding in the supramolec-
ular polymers result in less favourable polymerisation pro-
Procedure for the synthesis of acid chlorides 3c,d: (S)-3,7-Dimethylocta-
noic acid (2.76 g, 16.00 mmol) was dissolved in dry CH2Cl2 (30 mL) under
an argon atmosphere and cooled on an ice bath. A solution of oxalyl
Chem. Eur. J. 2010, 16, 810 – 821
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
819

A. R. A. Palmans, E. W. Meijer et al.
chloride (1.65 mL, 19.22 mmol, 1.2 equiv) in dry CH2Cl2 (20 mL) was
added dropwise and the solution was stirred under an argon atmosphere
at room temperature for an additional 3 h. Evaporation of the solvent
and the excess oxalylchloride in vacuo yielded 3c as a yellow oil (2.78 g,
14.6 mmol, 92% yield), which was used as such. 1H NMR (CDCl3): d=
2.88 (dd, 1H; (COOH)CH2), 2.71 (dd, 1H; (COOH)CH2), 2.10–2.08 (m,
Compound 1b was obtained as a white solid. 1H NMR (CDCl3): d=7.65
(s, 3H; Ar-H), 7.17 (s, 3H; N-H), 2.31 (t, 6H; COCH2), 1.69 (q, 6H;
COCH2-CH2), 1.39–1.21 (m, 60H; CH2), 0.88 ppm (t, 9H; CH3);
13C NMR (CDCl3): d=187.0 (C=O), 133.3 (Ar-C-N), 105.4 (Ar-C), 67.4,
59.2, 53.5, 44.4, 37.9, 31.9, 29.6, 29,5, 29.3, 25.6, 22.7, 14.9, 14.1 ppm;
MALDI-TOF-MS: m/z observed: 776.59 Da [M+Na]+; calcd: 753 Da.
Compound 1c was obtained as a sticky white solid. 1H NMR (CDCl3):
d=7.68 (s, 3H; Ar-H), 7.44 (s, 3H; N-H), 2.34–1.99 (m, 6H; COCH2),
1.69 (q, 6H; COCH2-CH2), 1.37–1.12 (m, 18H; CH2), 0.97 (d, 9H; CH3),
0.85 ppm (t, 18H; CH3); 13C NMR (CDCl3): d=171.4 (C=O), 138.9 (Ar-
C-N), 106.1 (Ar-C), 45.7, 39.0, 37.0, 30.9, 27.9, 24.7, 22.7, 22.5, 19.6 ppm;
MALDI-TOF-MS: m/z observed: 608.52 Da [M+Na]+; calcd: 585 Da.
Compound 1d was obtained as a sticky white solid. 1H NMR (CDCl3):
d=7.68 (s, 3H; Ar-H), 7.38 (s, 3H; N-H), 2.34–1.97 (m, 6H; COCH2),
1.52 (q, 6H; COCH2-CH2), 1.39–1.12 (m, 18H; CH2), 0.95 (d, 9H; CH3),
0.85 ppm (t, 18H; CH3); 13C NMR (CDCl3): d=171.4 (C=O), 138.9 (Ar-
C-N), 105.9 (Ar-C), 45.7, 39.0, 37.0, 31.0, 27.9, 24.7, 22.7, 22.5, 19.6 ppm;
MALDI-TOF-MS: m/z observed: 608.44 Da [M+Na]+; calcd: 585 Da.
1H; (CH2)2CHCH3), 1.59–1.51 (m, 1H; CH2CHCAHTUNGTRENNNUG
6H; -CH2-), 1.00 (d, 3H; CHCH3), 0.89 ppm (d, 6H; CHCAHTUNGTRENNUNG
FT-IR: n˜ =2957, 2930, 2870, 1799, 1464, 1384, 985, 967, 929 cmꢀ1. Acid
chloride 3d was obtained in an identical procedure.
General procedure for the synthesis of 3,5-dinitrobenzene-1-monocarbox-
amide (4a–d): 3,5-Dinitroaniline (1 equiv) was dissolved in dry CH2Cl2 at
a concentration of 0.5m and triethylamine (Et3N; 1.5 equiv) was added.
The solution was stirred at 08C under a nitrogen atmosphere. The appro-
priate acid chloride 3a–d (1.2 equiv) was dissolved in dry CH2Cl2 at a
concentration of 0.5m and added slowly to the amine solution. After the
addition was complete, stirring was continued overnight at room temper-
ature. The solvents were removed in vacuo and the crude product was di-
luted to 0.1m with chloroform. The solution was sequentially washed
with 1m HCl, water, 0.5m NaOH aqueous solution, water and brine. The
solution was dried with MgSO4 and the solvent was removed in vacuo.
The resulting crude compound was purified with column chromatography
using silica as the stationary phase. For 4a the eluent was chloroform/
methanol 99:1, for 4b–d the eluent was dimethoxyethane/heptane 1:5.
The yields were between 36 and 50% after column chromatography. 4a:
1H NMR (CDCl3): d=8.81 (t, 2H; Ar-H), 8.72 (t, 1H; Ar-H), 8.16 (brs,
1H; N-H), 2.50 (t, 2H; CO-CH2), 1.66 (q, 2H; COCH2-CH2), 1.30–1.10
Computational details: Quantum chemical calculations on both the peri-
odic chains of BTA or BTA monomers were performed within density
functional theory with the gradient-corrected PBE[15] exchange-correla-
tion functional. Electronic structure calculations on the infinite BTA
chains and related calculations on the monomeric species were carried
out within the plane-wave DFT (PW-DFT) approach using the Vienna
ab initio simulation package (VASP).[14] The projected augumented wave
(PAW) method[28,29] was used to describe the electron-ion interactions
and for valence electrons a plane-wave basis set was employed. During
the geometry optimisation the energy cut-off was set to 400 eV. Total
electronic energies used to estimate binding energies within the supra-
molecular ensemble were refined by performing single-point calculations
on the optimised structures with higher energy cut-off of 600 eV. The
Brillouin zone sampling was restricted to the G point.[30] Full geometry
optimisations were performed for each structure with the fixed cell pa-
rameters using a conjugate gradient algorithm. Convergence was as-
sumed to be reached when the forces on each atom were below
0.02 eVŠꢀ1. One-dimensional periodicity in the BTA chains was mod-
elled by using a supercell approach. The elementary molecular unit of
such chains that is a BTA dimmer was periodically repeated in the direc-
tion of c vector of the orthorhombic unit cell. To minimise possible artifi-
cial interactions between the periodic images in a and b directions, the
respective cell vectors were set to 20 Š to assure a ꢁ10 Š vacuum layer
between the BTA chains. The unit cell size in the direction of the c
vector was chosen based on the scan of the potential energy for the 1D
periodic model with respect to this parameter. The minimum energy
structures were obtained for c=7.2 and 7.4 Š for N-centred 3 and C=O-
centred 4 BTA chains, respectively. The resulting parameters of the unit
cells were a=b=20 Š and c=7.2 and 7.4 Š for N-centred 3 and C=O-
centred 4 BTA chains, respectively. Monomeric BTA species were repre-
sented within the same periodic code and level of theory as the chain
models by surrounding the molecular species with vacuum in a 20”20”
20 Š3 supercell.
1
(m, 8H; CH2), 0.88 ppm (t, 3H; CH3). 4b: H NMR (CDCl3): d=8.80 (t,
2H; Ar-H), 8.74 (t, 1H; Ar-H), 7.66 (brs, 1H; N-H), 2.46 (t, 2H;
COCH2), 1.78 (q, 2H; COCH2-CH2), 1.26–1.10 (m, 20H; CH2), 0.88 ppm
(t, 3H; CH3). 4c,d: 1H NMR (CDCl3): d=8.80 (t, 2H; Ar-H), 8.74 (t,
1H; Ar-H), 7.76 (brs, 1H; N-H), 2.36–2.18 (dd, 2H; COCH2), 2.08–2.06
(m, 1H; CH-(CH3)), 1.56–1.17 (m, 7H; CH-(CH3)2 and CH2), 1.00 (d,
3H; CH-(CH3)), 0.86 ppm (d, 6H; CH-(CH3)2).
General procedure for the synthesis of 3,5-diaminobenzene-1-monocar-
boxamide (5a–d): The appropriate dinitro compound (4a–d) was dis-
solved in ethyl acetate/methanol (1/1 v/v) in a concentration of 0.05m
and degassed with nitrogen. A catalytic amount of palladium on carbon
(10%) was added and the solution was hydrogenated using a Parr appa-
ratus until no more hydrogen was consumed. After that, the solution was
filtered and the residue was washed with methanol. The solvents were re-
moved in vacuo and the product was isolated in nearly quantitative yield
and used the same day for the next step. 5a: 1H NMR (CDCl3 + drop of
[D6]DMSO): d=9.20 (s, 1H; NH), 6.21 (s, 2H; Ar-H), 5.80 (brs, 4H;
NH2), 5.60 (t, 1H; Ar-H), 2.20 (t, 2H; COCH2), 1.65 (q, 2H; COCH2-
CH2), 1.26–1.10 (m, 8H; CH2), 0.90 ppm (t, 3H; CH3). 5b: 1H NMR
(CDCl3): d=7.21 (brs, 1H; N-H), 6.34 (t, 2H; Ar-H), 5.79 (t, 1H; Ar-H),
2.52 (t, 2H; COCH2), 1.65 (t, 2H; COCH2-CH2), 1.26–1.10 (m, 2H;
1
CH2), 0.90 ppm (t, 3H; CH3). 5c,d: H NMR (CDCl3): d=7.21 (brs, 1H;
N-H), 6.34 (t, 2H; Ar-H), 5.79 (t, 1H; Ar-H), 3.40 (brs, 4H; NH2), 2.40–
2.20 (dd, 2H; COCH2), 2.10–2.08 (m, 1H; (CH2)2CH
(m, 7H; CH2CH(CH3)2 and -CH2-), 1.00 (d, 3H; CHCH3), 0.89 ppm (d,
6H; CH(CH3)2).
ACHTUNGTREN(UNNG CH3)), 1.59–1.14
Conformational study of monomeric BTA species was performed by
using Gaussian 03 program package.[20] Full geometry optimisations and
partial geometry optimisations with fixed values of intramolecular dihe-
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
dral angles were performed at the PBE/6–311+GACTHNUTRGNE(NUG d,p) level of theory.
General procedure for the synthesis of N-centred BTAs 1a–d: The same
procedure as described for compounds 4a–d was carried out, with the dif-
ference that the solvent employed was dry chloroform and the appropri-
ate acid chloride was added in 2.5 equivalents. The crude product was pu-
rified with column chromatography using silica as the stationary phase.
For 1a the eluent employed was chloroform/methanol 99:1, for 1b–d,
chloroform/ethylacetate 9:1 with a gradient to 8:2 was used. The yields
after purification were typically around 30–50%.
Compound 1a was obtained as a white solid.1H NMR (CDCl3): d=7.66
(s, 3H; Ar-H), 7.28 (s, 3H; N-H), 2.31 (t, 6H; COCH2), 1.69 (q, 6H;
COCH2-CH2), 1.38–1.25 (m, 24H; CH2), 0.88 (t, 9H; CH3); 13C NMR
(CDCl3): d=171.8 (C=O), 139.0 (Ar-C-N), 105.8 (Ar-C), 37.9, 31.6, 29.2,
29.0, 25.6, 22.6, 14.0 ppm; MALDI-TOF-MS: m/z observed: 524.49 Da
[M+Na]+; calcd: 501 Da.
Acknowledgements
The authors would like to thank Koen Pieterse for his help with initial
calculations and art-work and NOW and NRSC-C for financial support.
The National Computing Facilities Foundation (NCF) is acknowledged
for providing computational resources, with financial support from NOW
(SH-125-08).
820
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 810 – 821

Dynamic Supramolecular Polymers
FULL PAPER
[16] a) J. Ireta, J. Neugebauer, M. Scheffler, J. Phys. Chem. A 2004, 108,
5692; b) J. Ireta, J. Neugebauer, M. Schefller, A. Rojo, M Galvµn, J.
Phys. Chem. B 2003, 107, 1432–1437; c) C. Tuma, ; A. D. Boese,
N. C. Handy, Phys. Chem. Chem. Phys. 1999, 1, 3939–3947; A. D.
Boese, N. C. Handy, Phys. Chem. Chem. Phys. 1999, 1, 3939–3947;
d) R. Kaschner, D. Hohl, J. Phys. Chem. A 1998, 102, 5111–5116;
e) T. van der Wijst, C. Fonseca-Guerra, M. Swart, F. M. Bickelhaupt,
Chem. Phys. Lett. 2006, 426, 415–421; f) D. R. Hamann, Phys. Rev.
B 1997, 55, R10157–R10160; g) J. A. Frey, S. Leutwyler, J. Phys.
Chem. A 2006, 110, 12512–12518; h) Y. Zhao, D. G. Truhlar, J.
Chem. Theory Comput. 2007, 3, 289–300.
[17] R. S. Fellers, D. Barsky, F. Gygi, M. Colvin, Chem. Phys. Lett. 1999,
312, 548–555.
[18] A. Rochefort, E. Bayard, D. Hadji-Messaoud, Adv. Mater. 2007, 19,
1992–1995.
[19] a) S. J. Grabowski, J. Phys. Chem. A 2000, 104, 5551–5557; b) S. J.
Grabowski, Chem. Phys. Lett. 1999, 312, 542–547; c) M. Rozenberg,
A. Loewenschuss, Y. Marcus, Phys. Chem. Chem. Phys. 2000, 2,
2699–2702.
[20] Gaussian 03, Revision B.04, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A Robb, J. R. Cheeseman, V. G. Zakrzewski,
[1] a) Y. Matsunaga, N. Miyajima, Y. Nakayasu, S. Sakai, M. Yonenaga,
Bull. Chem. Soc. Jpn. 1988, 61, 207–210; b) Y. Matsunaga, Y. Na-
kayasu, S. Sakai, M. Yonenaga, Mol. Cryst. Liq. Cryst. 1986, 141,
327–333; c) Y. Kobayashi, Y. Matsunaga, Bull. Chem. Soc. Jpn.
1987, 60, 3515–3518; d) P. J. M. Stals, M. M. J. Smulders, R. Martꢁn-
Rapffln, A. R. A. Palmans, E. W. Meijer, Chem. Eur. J. 2009, 15,
2071–2080.
[2] a) K. Hanabusa, A. Kawamaki, M. Kimura, H. Shirai, Chem. Lett.
1997, 191–192; b) K. Hanabusa, C. Koto, M. Kimura, H. Shirai, A.
Kakehi, Chem. Lett. 1997, 429–430; c) D. Ogata, T. Shikata, K. Ha-
nabusa, J. Phys. Chem. B 2004, 108, 15503–15510; d) A. Sakamoto,
D. Ogata, T. Shikata, K. Hanabusa, Macromolecules 2005, 38, 8983–
8986; e) A. Sakamoto, D. Ogata, T. Shikata, O. Urakawa, K. Hana-
busa, Polymer 2006, 47, 956–960; f) T. Shikata, D. Ogata, K. Hana-
busa, J. Phys. Chem.
B 2004, 108, 508–514; g) T. Shikata, Y.
Kuruma, A. Sakamoto, K. Hanabusa, J. Phys. Chem. B 2008, 112,
16393–16402.
[3] a) J. Roosma, T. Mes, P. Lecl›re, A. R. A. Palmans, E. W. Meijer, J.
Am. Chem. Soc. 2008, 130, 1120–1121; b) C. F. C. FitiØ, I. Tomatsu,
D. Byelov, W. H. de Jeu, R. P. Sijbesma, Chem. Mater. 2008, 20,
2394–2404; c) R. van Hameren, P. Schçn, A. M. van Buul, J. Hoog-
boom, S. V. Lazarenko, J. W. Gerritsen, H. Engelkamp, P. C. M.
Christianen, H. A. Heus, J. C. Maan, T. Rasing, S. Speller, A. E.
Rowan, J. A. A. W. Elemans, R. J. M. Nolte, Science 2006, 314,
1433–1436; d) I. Paraschiv, M. Giesbers, B. van Lagen, F. C. Groze-
ma, R. D. Abellon, L. D. A. Siebbeles, A. T. M. Marcelis, H. Zuilhof,
E. J. R. Sudhçlter, Chem. Mater. 2006, 18, 968–974.
[4] a) M. P. Lightfoot, F. S. Mair, R. G. Pritchard, J. E. Warren, Chem.
Commun. 1999, 1945–1946; b) P. P. Bose, M. G. B. Drew, A. K. Das,
A. Banerjee, Chem. Commun. 2006, 3196–3198.
[5] a) L. Brunsveld, A. P. H. J. Schenning, M. A. C. Broeren, H. M. Jans-
sen, J. A. J. M. Vekemans, E. W. Meijer, Chem. Lett. 2000, 292–293;
b) M. M. J. Smulders, A. P. H. J. Schenning, E. W. Meijer, J. Am.
Chem. Soc. 2008, 130, 606–611.
[6] M. Blomenhofer, S. Ganzleben, D. Hanft, H.-W. Schmidt, M. Kris-
tiansen, P. Smith, K. Stoll, D. Mader, K. Hoffmann, Macromolecules
2005, 38, 3688–3695.
[7] J. van Herrikhuyzen, P. Jonkheijm, A. P. H. J. Schenning, E. W.
Meijer, Org. Biomol. Chem. 2006, 4, 1539–1545.
[8] J. J. van Gorp, J. A. J. M. Vekemans, E. W. Meijer, Mol. Cryst. Liq.
Cryst. 2003, 397, 191–205.
[9] P. Jonkheijm, P. van der Schoot, A. P. H. J. Schenning, E. W. Meijer,
Science 2006, 313, 80–83.
J. A. MontgomACTHNUGRTNEUNGery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich,
J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J.
Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli,
C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q.
Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G. Liu, A.
Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin,
D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara,
C. Gonzalez, M. Challacombe, P. M. W. Gill, B. G. Johnson, W.
Chen, M. W. Wong, J. L. Andres, M. Head-Gordon, E. S. Replogle,
J. A. Pople, Gaussian, Inc., Wallingford CT, 2004.
[21] V. Pophristic, S. Vemparala, I. Ivanov, Z. Liu, M. L. Klein, W. F. De-
Grado, J. Phys. Chem. B 2006, 110, 3517–3526.
[22] R. Vargas, J. Garza, D. Dixon, B. Hay, J. Phys. Chem. A 2001, 105,
774–778.
[23] a) R. J. Doerksen, B. Chen, D. Liu, G. N. Tew, W. F. DeGrado, M. L.
Klein, Chem. Eur. J. 2004, 10, 5008–5016; b) R. J. Doersken, B.
Chen, M. L. Klein, Chem. Phys. Lett. 2003, 380, 150–157.
[24] G. R. Desiraju, Acc. Chem. Res. 1996, 29, 441–449.
[25] a) N. Kobko, J. J. Dannenberg, J. Phys. Chem. A 2003, 107, 10389–
10395; b) T. F. A. de Greef, M. M. J. Smulders, M. Wolffs, A. P. H. J.
Schenning, R. P. Sijbesma, E. W. Meijer, Chem. Rev. 2009, 109,
5687–5754.
[10] J. van Gestel, Macromolecules 2004, 37, 3894–3898.
[11] M. M. J. Smulders, I. A. W. Filot, J. M. A. Leenders, P. van der
Schoot, A. R. A. Palmans, A. P. H. J. Schenning, E. W. Meijer, un-
published results.
[12] J. van Gestel, A. R. A. Palmans, B. Titulaer, J. A. J. M. Vekemans,
E. W. Meijer, J. Am. Chem. Soc. 2005, 127, 5490–5494.
[13] The pKaꢂs of Ph-NHCO-CH3 and Ph-CONH2 measured in DMSO
are 21.45 and 23.35, respectively. See F. G. Bordwell, D. Algrim, J.
Org. Chem. 1976, 41, 2507–2508.
[26] Y. Zhou, M. Xu, T. Yi, S. Xiao, Z. Zhou, F. Li, C. Huang, Langmuir
2007, 23, 202–208.
[27] A. Tanatani, A. Yokoyama, I. Azumaya, Y. Takakura, C. Mitsui, M.
Shiro, M Uchiyama, A. Muranaka, N. Kobayashi, T. Yokozawa, J.
Am. Chem. Soc. 2005, 127, 8553–8561.
[28] P. E. Blçchl, Phys. Rev. B 1994, 50, 17953–17979.
[29] G. Kresse, J. Joubert, Phys. Rev. B 1999, 59, 1758–1775.
[30] H. J. Monkhorst, J. D. Pack, Phys. Rev. B 1976, 13, 5188–5192.
[14] a) G. Kresse, J. Furtmüller, Phys. Rev. B 1996, 54, 11169–11186;
b) G. Kresse, J. Furtmüller, J. Comp. Mater. Sci. 1996, 6, 15–50.
[15] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77,
3865–3868.
Received: September 24, 2009
Published online: December 18, 2009
Chem. Eur. J. 2010, 16, 810 – 821
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
821