Discordant Supramolecular Fibres Reversibly Depolymerised by Temperature and Light

Article abstract of DOI:10.1002/chem.202004115

Synthetic stimuli responsive supramolecular polymers attract increasing interest for their ability to mimic the unique properties of natural assemblies. Here we focus on the well-studied benzene-1,3,5-tricarboxamide (BTA) motif, and substitute it with two

Full text of DOI:10.1002/chem.202004115

Accepted Article
Accepted Article
Title: Discordant supramolecular fibres reversibly depolymerised by
temperature and light
Authors: Marieke Gerth, José A. Berrocal, Davide Bochicchio, Giovanni
M. Pavan, and Ilja Karina Voets
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.202004115
Link to VoR: https://doi.org/10.1002/chem.202004115
01/2020

10.1002/chem.202004115
Chemistry - A European Journal
FULL PAPER
DISCORDANT SUPRAMOLECULAR FIBRES REVERSIBLY
DEPOLYMERISED BY TEMPERATURE AND LIGHT
Marieke Gerth,[a,b,c,d] José A. Berrocal,[e] Davide Bochicchio,[f] Giovanni M. Pavan,[f,g] and Ilja K.
Voets*[a,c,h]
[a]
Dr. M. Gerth, Prof.Dr. I.K. Voets
Self-Organizing Soft Matter group, Faculty of Chemical Engineering and Chemistry
Eindhoven University of Technology
Groene Loper 3, 5612AE Eindhoven, The Netherlands
E-mail: i.voets@tue.nl
[b]
[c]
[d]
[e]
[f]
Dr. M. Gerth
Laboratory of Physical Chemistry, Faculty of Chemical Engineering and Chemistry
Eindhoven University of Technology
Groene Loper 3, 5612AE Eindhoven, The Netherlands
Dr. M. Gerth, Prof.Dr. I.K. Voets
Institute for Complex Molecular Systems
Eindhoven University of Technology
Groene Loper 3, 5612AE Eindhoven, The Netherlands
Dr. M. Gerth
Van’t Hoff Laboratory for Physical and Colloid Chemistry, and Debye Institute for Nanomaterials Science
Utrecht University
Padualaan 8, 3584CH Utrecht, The Netherlands
Dr. J.A. Berrocal
Adolphe Merkle Institute
University of Fribourg
Chemin des Verdiers 4, CH-1700 Fribourg, Switzerland
Dr. D. Bochicchio, Prof.Dr. G.M. Pavan
Department of Innovative Technologies
University of Applied Sciences and Arts of Southern Switzerland
Galleria 2, Via Cantonale 2c, CH-6928 Manno, Switzerland
Prof.Dr. G.M. Pavan
[g]
Department of Applied Science and Technology
Politecnico di Torino
Corso Duca degli Abruzzi 24, 10129 Torino, Italy
Supporting information for this article is given via a link at the end of the document.
Synthetic stimuli responsive supramolecular polymers attract
increasing interest for their ability to mimic the unique properties
of natural assemblies. Here we focus on the well-studied
benzene-1,3,5-tricarboxamide (BTA) motif, and substitute it with
two (S)-3,7-dimethyloctyl groups and an azobenzene
photoswitch. We demonstrate the UV (λ = 365 nm) induced
depolymerisation of the helical hydrogen-bonded polymers in
methylcyclohexane (MCH) through circular dichroism and UV-
vis spectroscopy in dilute solution (15 μM), and NMR and iPAINT
super-resolution microscopy in concentrated solution (300 μM).
Repeated depolymerisation can be reversed without degradation
by irradiating at λ = 455 nm. Molecular dynamics simulations
show that the most energetically favourable configuration for
these polymers in MCH is a left-handed helical network of
hydrogen-bonds between the BTA cores surrounded by two
right-handed helices of azobenzenes, which is experimentally
recovered after depolymerisation. The responsiveness to two
orthogonal triggers across a broad concentration range holds
promise for use in e.g. photo-responsive gelation.
Supramolecular chemistry offers a broad palette of handles to
fine-tune the hierarchical nanostructure and associated properties
of materials through the strong dependence of the configuration
of supramolecular assemblies on external variables and exact
chemical structure of the monomers.[1] By incorporating selected
chemical groups into supramolecular moieties, the responsivity of
supramolecular constructs to specific external cues can thus be
extended by design.[2] Being able to reversibly switch between a
material with and without emerging properties, especially through
local (de)activation, is a continuous quest in material science that
is being tackled from the molecular to the macro-scale.[3]
In contrast to, for example, thermal cues, light can easily be
applied in a patterned fashion to achieve localised responses in a
4]
non-invasive fashion.[3a, Numerous light-responsive systems,
based on supramolecular assembly, have been presented in the
past years, which upon irradiation can, e.g., covalently
polymerise[5], undergo supramolecular polymerisation upon
release of protons from a photo-acid[6], change `secondary
structure' of the backbone[2a], catastrophically depolymerise due
to strain build-up[7], or act as nanoscopic switches for macroscopic
actuation[3c, 8]. Synthetic accessibility of azobenzenes and related
light-responsive molecular switches allows targeted design of
photoswitches that are responsive to light of different
wavelengths.[9]
Introduction
1
This article is protected by copyright. All rights reserved.

10.1002/chem.202004115
Chemistry - A European Journal
FULL PAPER
A supramolecular construct that has been studied extensively,
both experimentally[10] and computationally[11], is the synthetically
accessible benzene-1,3,5-tricarboxamide (BTA) motif. Such
BTAs can form temperature-responsive supramolecular
structures in certain solvents and are versatile supramolecular
building blocks, since they can be designed for self-assembly in
The UV-vis absorption at 200 ≤ λ ≤ 275 nm originates from the
BTA core (Figure 2A).[15] The trans-azobenzene moieties absorb
between λ ≈ 300-400 nm, and the low-intensity UV-vis absorption
peak above λ ≈ 420 nm is attributed to cis-azobenzene. This is a
minority species (Figure S14) in the thermodynamic equilibrium
distribution at 20 °C, but exhibits an unexpectedly strong Cotton
effect in CD (Figure 2A). The CD spectrum of azoBTA in MCH
exhibits a strong negative signal centred around λ = 225 nm,
which is attributed to the preferred left-handed helical assembly
of γ-(S)-BTA (Figure 2A).[10b, 16] The positive CD signal originating
from the azobenzene units indicates that also these are helically
arranged. In the Fourier Transform Infrared (FTIR) spectra, we
observe characteristic vibrational peaks for non-hydrogen-
bonded species (ν = 3450 cm−1 and 1666 cm−1) in CHCl3 – a
solvent wherein azoBTA is molecularly dissolved – which shift to
12]
13]
aqueous environments[11b,
or apolar alkane solvents.[10b,
Here, inspired by previous work on azobenzene-decorated
BTAs[3b, 14], we synthesise a BTA derivative (Figure S1, S5, S9,
S13) comprising an azobenzene as a molecular photoswitch
(azoBTA, Figure 1) for light-induced cycling between an
associated and dissociated state in apolar solvents such as
methylcyclohexane (MCH). We show through a combination of
spectroscopic experiments on dilute solutions (15 μM), and NMR
and super-resolution microscopy on concentrated solutions (300
μM) that the photoisomerisation and (dis)association is reversible
and occurs over a broad range of concentrations. It should be
noted that the above-mentioned azobenzene-decorated BTAs are
C3-symmetrical, which the azoBTA presented here is not. The
desymmetrisation is expected to influence the (light-induced
de-)polymerisation, as well as the structure of the fibres. We
indeed find that the helices formed by the hydrogen bonds are
left-handed, whereas the helices formed by the azobenzene-
bearing moieties are predominantly right-handed, yielding a
discordant helical structure.
lower wavenumbers in MCH (ν = 3238 cm−1 and 1642 cm−1),
17]
indicative for hydrogen-bond formation (Figure 2B).[10b,
We
record the variable temperature (VT)-UV-vis and CD (Figure 2C
and S17) spectra of 15 μM azoBTA in MCH to find the melting
temperature of the azoBTA polymers. The red-shifted UV-vis
absorption shoulder of the trans-azobenzene peak at low
temperature indicates the presence of J-type aggregates, which
are aggregates characterised by a large angular offset.[18] Upon
heating above the melting temperature of the fibres (Tm ≈ 75 °C),
the vanishing CD signal for both the BTA and azobenzene
chromophores indicates that azoBTA is molecularly dissolved
(Figure 2C). Supramolecular assembly into one-dimensional
fibres is confirmed by small-angle X-ray scattering (SAXS). We
obtain for 3 mM azoBTA in MCH a core radius of 13.8 Å and a
Figure 1. Chemical structure of azobenzene functionalised BTA and a cartoon
representation of their discordant superstructure and light-induced
dissociation. In the cartoon the benzene rings of the BTA cores are indicated
as black disks, the hydrogen bonds are depicted in white, and the
azobenzenes are represented by the orange rectangles. Trans-azobenzenes
are flat rectangles, whereas cis-azobenzenes are slightly twisted rectangles.
Irradiation with light of λ = 365 nm (λ365nm) leads to (partial) depolymerisation of
the equilibrated discordant fibres, which is reversible through irradiation at λ =
455 nm or thermal relaxation.
Figure 2. AzoBTA forms H-bond stabilised polymers in methylcyclohexane. A)
A representative UV-vis absorption and CD spectrum of azoBTA at 15 μM in
MCH at 20 °C. The absorption peak originating from the (hydrogen bonded)
BTA is in the range 200 ≤ λ ≤ 275 nm, from trans-azobenzene between λ = 300
- 400 nm, and from cis-azobenzene above ~420 nm. B) FTIR spectroscopy
shows signature peaks at ν = 3238 cm−1 and 1642 cm−1 for hydrogen-bonded
BTAs at 500 μM in MCH, and monomers at 500 μM in CHCl3. C) Evolution of
CD signal at diagnostic wavelengths for BTA (λ = 225 nm), trans-azobenzene
(λ = 365 nm) and cis-azobenzene (λ = 455 nm) assembly with temperature. The
concentration of azoBTA is 15 μM in MCH. The temperature was increased at
10 °C∙hr−1, and the sample was kept isothermal during the measurements of the
spectra. D) The SAXS profile (open circles) of a 3 mM sample in MCH and fit
(line) match a core-shell cylindrical structure.[19]
Results and Discussion
AzoBTA forms H-bond stabilised polymers in methylcyclohexane
First, we investigate spectroscopically whether azoBTA
molecules are able to form helical superstructures when the
azobenzene is present in the side chain of the monomer structure.
2
This article is protected by copyright. All rights reserved.

10.1002/chem.202004115
Chemistry - A European Journal
FULL PAPER
thickness of 19.0 Å using a core-shell cylinder model[31] to
describe the SAXS profile (Figure 2D). The length of the fibres is
too long to resolve.
cores, we observe that the order in the helical arrangement of the
azobenzene units is much more impaired.
From these MD simulations, we can also compare the different
fibre structures in terms of interactions between the monomers in
the assembly as well as with the external solvent. From an energy
point of view, the simulations suggest that the best arrangement
for the azoBTA fibre is a double azobenzene helix discordant with
respect to the inner H-bond architecture (S4: see Figure 3a). In
fact, the S4 arrangement minimises both the total potential energy
of the system (Figure 3c: sum of solute-solute, solute-solvent and
solvent-solvent interactions) and in particular the azoBTA-
azoBTA monomer-monomer interaction energy (see Supporting
Figure S19), while the other 3 configurations are relatively higher
in energy (and quite similar between them). We decomposed the
azoBTA-azoBTA interaction energy in 4 different terms, as shown
in Figure 3c, in order to investigate the major driving force for this
difference. Namely, we decomposed the monomer-monomer
interaction energy in terms of how much of this is due to the
monomer cores, tails, etc. (data and scheme in Figure 3c: how
much the cores interact between themselves, the tails between
themselves, etc. – energy values expressed per-monomer). We
find that, in general, the core-core interaction (Figure 3c: 1,
orange plot) is much more relevant than the azobenzene-
azobenzene interaction (2, purple plot) for the monomer-
monomer interaction energy, which is consistent with the
evidence that during the MD simulations on average the core-core
stacking is preserved to a higher extent compared to that between
the azobenzene units. S2 and S4, the configurations in which the
helicities are opposed, are the ones that minimise the
azobenzene-azobenzene interaction (Figure 3c: 2, purple plot).
However, from the decomposition analysis, it becomes clear that
the better stability of S4 comes mainly from the optimisation of the
interactions between carbon chains (Figure 3c: 4, grey plot),
which is found about 8 kcal mol−1 more favourable than in the
other cases. Interestingly, S4 is also the system in which the
azobenzene-azobenzene displacement and aggregation differs
more compared to the others (Supporting Figure S20: higher g(r),
indicating a stronger compaction of azobenzene moieties). In S4
the azobenzene moieties are more compactly arranged due to the
stronger folding and interaction of the alkyl chains. Altogether, the
MD results indicate that at T = 300 K these azoBTA monomers
prefer to form a 1D stack with opposed helicities for BTAs and
azobenzenes. From an energy point of view, this is thus
suggested to be the most favourable, and thus likely, the structure
which these assemblies adopt in solution under the experimental
conditions.
Molecular dynamics simulations provide an insight into the
internal structure
Intrigued by the fact that the CD spectra seem to suggest the
presence of opposite helicities for the BTA cores and azobenzene
units, we performed all-atom molecular dynamics (MD)
simulations to investigate the internal arrangement of the azoBTA
fibres at atomistic resolution. We built an atomistic model of the
azoBTA monomer and parametrised it in the framework of GAFF
(General Amber Force Field),[20] as previously done for other BTA
21]
variants[11b,
and other azo-containing self-assembling
monomers.[7, 22] Then, we constructed a model fibre composed of
30 initially extended azoBTA monomers arranged in 4 different
ways (generating 4 different starting configurations for the fibres,
see Figure 3a), spanning through the simulation box through
periodic boundary conditions (PBC) to effectively model a portion
of the bulk of infinite supramolecular polymers. Among the four
initial configurations, two have concordant (the same) helicities for
the hydrogen bonds involving the amide groups of the BTA
monomers and the stacking of the azobenzene groups (S1 and
S3), and two have discordant (opposite) helicities between BTA
and azobenzenes (S2 and S4). Two configurations have a single
azobenzene helix (S2 and S3) and two have a double helical
structure (S1 and S4), see Figure 3a. In all cases, we solvated the
initially extended fibre models with the same quantity (4000) of
explicit MCH solvent molecules. These models have then been
relaxed and equilibrated through 400 ns of MD simulations
conducted in NPT conditions (constant N: number of particles, P:
pressure and T: temperature during the run) at the temperature of
300 K and 1 atm of pressure. During this time all systems were
seen to equilibrate in the MD regime. Then, we compared the last
200 ns (equilibrated part) of each MD simulation.
As a common feature in all four cases we could observe that
during the MD simulations the fibres deviate quite a lot from their
initially extended configuration, while at the same time these
substantially preserve the order in terms of stacking of the BTA
cores (see Supporting Figure S18: g(r) between the BTA cores).
The average number of H-bonds per BTA monomer is in every
case around ~2.5. This indicates a relatively high stability and
order of the stacking of the BTA cores, which is preserved in a
better way than in BTA variants designed to self-assemble in
water, where we previously found a value of around ~2.1 (as seen
recently, the maximum of 3 H-bonds per-monomer is not reached
in these systems in solution due to thermal oscillations/vibrations
and deviations from perfection).[11a, 11b, 21-22] Compared to the BTA
3
This article is protected by copyright. All rights reserved.

10.1002/chem.202004115
Chemistry - A European Journal
FULL PAPER
Figure 3. Molecular dynamics simulations. a) The four different initial azoBTA arrangements used in this work: S1 and S3 have concordant helicities between the
amides H-bonding (in red) and azobenzene stacking (in blue), while S2 and S4 have discordant helicities. b) Total potential energy in the systems (sum of the total
solute-solute, solute-solvent and solvent-solvent interactions) taken from the equilibrated part (last 200 ns) of MD – this is minimized in the S4 assembled structure.
c) Decomposition of the azoBTA-azoBTA interaction energy (values per-monomer): core-core (1: orange), azobenzene-azobenzene (2: purple), polar group-polar
group (3: yellow) and alkyl chain-alkyl chain (4: grey).
Reversible photoisomerisation
7.0 ppm for the trans-azoBTA model compound. Upon
photoisomerisation the doublet disappears while the multiplet
shifts downfield to δ ≈ 6.8 ppm, which is indicative for the
presence of the cis isomer. As expected, no signal is visible in the
spectrum for azoBTA in d14-MCH at room temperature before
irradiation ‒ apart from the reference CHCl3 at δ = 7.26 ‒ since
the majority of the azoBTA is assembled (Figure 4C). Signals
appear in d14-MCH at T = 75 °C before irradiation (Figure 4D, δ =
7.97, 7.12 and 7.01 ppm), and at room temperature after
irradiation with UV light (Figure 4E, respectively). Before
irradiation the chemical shifts of the peaks correspond well with
the trans isomer of the model compound, and the appearance of
peaks after UV irradiation indicates that the fibres have (partially)
depolymerised.
With the confirmation that the supramolecular polymers melt at
high temperatures and after establishing the most likely
conformation of the polymers in solution, we study the
photoresponsivity of the motifs. We investigate the addressability
of the azobenzenes with UV light (λ365nm) through 1H NMR
experiments (Figure 4) on a model compound azoBTA-CO2Et
(Figure S21) in deuterated chloroform (CDCl3), and next
investigate the light response of the assembled azoBTA in
deuterated MCH (d14-MCH). A model compound was used in
CDCl3 due to the limited availability of azoBTA; since there is no
self-assembly in chloroform, we are confident it would have the
same isomerisation behaviour. We focus on the aromatic region
(chemical shifts 8.9 < δ < 5.9 ppm), which is where the diagnostic
peaks originating from the hydrogens of the aromatic rings of the
azobenzene appear. A comparison of the 1H NMR spectra in
CDCl3 at room temperature for monomeric azoBTA before (Figure
4A) and after (Figure 4B) irradiation with λ365nm reveals a
characteristic doublet at δ = 7.85 ppm and a multiplet around δ =
4
This article is protected by copyright. All rights reserved.

10.1002/chem.202004115
Chemistry - A European Journal
FULL PAPER
Figure 4. 1H NMR spectra of azoBTA and a model compound azoBTA-CO2Et
under various conditions. 1H NMR spectrum of 300 μM azoBTA-CO2Et A)
before irradiation in CDCl3, and B) irradiated with λ365nm in CDCl3. 1H NMR
spectrum of 300 μM azoBTA C) before irradiation in d14-MCH at room
temperature, D) before irradiation in d14-MCH at T = 75 °C, and E) irradiated
with λ365nm in d14-MCH. Irradiation was performed for 15 mins at 600 mA
Figure 5. (Photo-)isomerisation of non-associated azoBTA. A) UV-vis spectra
of stepwise trans (blue) to cis (yellow) photoisomerisation through irradiation at
λ365nm, and B) the reverse cis (blue) to trans (yellow) is achieved by irradiation
at λ455nm. Irradiation experiments in A and B are performed at 15 μM in CHCl3 at
room temperature with small doses of weak-intensity light corresponding to 1 s
at 10% of the LED intensity at 1000 mA. C) Thermal cis-to-trans relaxation of
300 μM azoBTA-CO2Et in CDCl3 at room temperature as monitored by 1H NMR
spectroscopy. The monomers were isomerised to cis by 15 mins irradiation with
high intensity UV light at 600 mA. D) The photoisomerisation of azoBTA is
reversible for at least 20 cycles as shown by UV-vis absorption values at A365nm
for the trans-azoBTA (circles) and A455nm for cis-azoBTA (squares). The 15 μM
solution in chloroform was irradiated at room temperature for 10 s at 100% of
the LED intensity at 1000 mA.
We now turn to UV-vis spectroscopy to study the
photoisomerisation in more detail and attempt to obtain control
over the fraction of isomerised azoBTA by stepwise irradiation.
Stepwise irradiation (steps of 1 s separated by 59 s intervals for
spectral acquisition and sample handling) facilitates the
investigation of the isomerisation and dissociation states that lie
between the predominantly trans pristine state, and the
predominantly cis state obtained after irradiation at λ = 365 nm. A
typical UV-vis absorption spectrum of pristine, non-associated,
predominantly trans-azoBTA in chloroform is depicted in Figure
5A in blue. It shows a strong π→π* absorption around λ = 365 nm
and a weak n→π* absorption around λ = 420 nm, both originating
from trans-azobenzene. Upon stepwise irradiation with λ365nm the
peak originating from the π→π* transition gradually decreases in
intensity, and a weak absorption centred at λ = 445 nm appears
for the n→π* transition of the cis isomer. Within 29 steps of 1 s
irradiation we reach a characteristic spectrum for predominantly
cis-azobenzene that hardly changes upon continued irradiation.
Subsequent exposure to stepwise irradiation at λ = 455 nm
induces the reverse process: the absorption band at λ = 365 nm
increases in intensity, while simultaneously the n→π* band
undergoes a hypsochromic shift from λ = 455 nm to 420 nm
(Figure 5B). After 34 steps of 1 s irradiation, the absorption
spectrum appears constant with continued irradiation.
recovery from 80% to approximately 100% trans-azobenzene in
Figure 5C. This strongly suggests that it is possible to isomerise
the remaining fraction of cis-isomers (Figure 5C), additionally
considering that trans-azoBTA is the most favourable
conformation (Figure S22, S23).
Having established that the photoisomerisation is fully reversible,
though slow, we evaluate the robustness of the process. To this
end a 15 μM azoBTA solution is alternatingly illuminated with 365
nm and 455 nm light for 10 s (Figure 5D). Gratifyingly, we observe
little to no difference in the maximum A365nm and A455nm between
20 successive cycles as we repeatedly switch between trans- and
cis-azoBTA.
Fibre dissociation through irradiation
Now that we have confirmed that the azobenzenes are light-
responsive in the assembled azoBTA molecules, we investigate
the effect of the photoisomerization on the assembly state in more
detail through UV-vis and CD spectroscopy. We perform stepwise
irradiation experiments in MCH at room temperature and observe
the changes in UV-vis and CD spectra. Upon stepwise irradiation
at λ365nm, we again observe a gradual decrease of the intensity of
the trans-π→π* absorption peak around 365 nm, and the
simultaneous increase in intensity of A455nm (Figure 6A). The
consecutive spectra recorded for λ365nm irradiation in MCH are
spaced closer together than the equivalent spectra recorded for
azoBTA in chloroform (Figure 5A), and a longer total irradiation
time is needed to reach a state where the spectrum barely
changes with continued irradiation. These combined observations
indicate that in MCH a higher energy barrier for isomerisation
Surprisingly, the intensity of the absorption maximum at λ = 365
nm after a more-than-equivalent number of backward irradiation
steps with λ455nm is approximately 20% below its initial value.
Therefore, we investigate the thermal relaxation of the azoBTA-
CO2Et model compound through 1H NMR. We determine the
percentage trans-azobenzenes as a function of relaxation time
from the ratio of peak integrals at characteristic chemical shifts for
the trans and cis isomers – specifically the ratio between peaks at
δ = 7.85 and 8.35. Herewith we determine the half-life of the cis
isomer in CDCl3 to be about 24 hrs. From these kinetics we find
that the thermal recovery of the last 20% or so of cis-azobenzene
still present after 30×1 s irradiation steps at λ455nm will take
approximately 100 hrs at room temperature; equivalent to the
5
This article is protected by copyright. All rights reserved.

10.1002/chem.202004115
Chemistry - A European Journal
FULL PAPER
exists, which we attribute to the strong attractions and close
pacing of the molecules in the fibres. Irradiation at λ455nm recovers
the trans isomer to a large extent. The shape of the obtained UV-
vis absorption spectrum after trans-cis-trans photoisomerisation
has features of both trans-azoBTA in CHCl3 and of the initial
spectrum in MCH (Figure S24), which indicates that the
azobenzenes are less or differently organised than before
irradiation.
We exploit the inclusion of the two γ-(S) stereogenic centres per
azoBTA monomer to monitor the presence of excess right- or left-
handed superstructures by CD spectroscopy. We find that the
trans-to-cis isomerisation upon stepwise irradiation with λ365nm
(Figure 6A) is accompanied by a complete loss of the azobenzene
CD signal, while the Cotton effect of the BTA is merely reduced in
intensity (Figure 6B). This implies that the photoisomerisation to
cis-azoBTA triggers the disappearance of the helical azobenzene
superstructure, while the BTA core of the fibres dissociates but
does not disintegrate entirely.
Azobenzene-substituted BTA monomers in MCH spontaneously
form hydrogen-bonded supramolecular fibres in MCH below T ≈
75 °C, as shown by FTIR, iPAINT, and CD and UV-vis
spectroscopy. The stability of the hydrogen bonded network
between BTA cores is also confirmed by MD simulations. The
azoBTAs form a 1D helical assembly wherein the helicity of the
BTA cores is opposite to the double helix formed by the
azobenzene substituents. Additionally, the azobenzenes remain
addressable with UV irradiation at a wavelength of λ = 365 nm
when the azoBTA is assembled into fibres, and the UV-induced
trans-to-cis isomerisation depolymerises the supramolecular
fibres. The formation of a discordant helical nanoarchitecture
appears robust; it is regenerated after thermal depolymerisation,
after cis-trans photoisomerisation, and after thermal relaxation.
This work demonstrates that a single photoswitch per monomer
suffices to depolymerise a supramolecular BTA construct.
Experimental Section
Synthesis and characterisation of methyl-(E)-4-(4-((4-(2-((tert-
butoxycarbonyl)amino)ethoxy)phenyl)diazenyl)phenoxy)butanoate
(1): (E)-4,4’-(diazene-1,2-diyl)diphenol (2.51 g, 11.67 mmol), tert-butyl (2-
bromoethyl)carbamate (1.72 g, 7.67 mmol), methyl 4-bromobutanoate
(3.17 g, 17.51 mmol), potassium carbonate (8.08 g, 134.7 mmol) and
sodium iodide (187 mg, 1.25 mmol) were suspended in DMF (55 ml) by
sonication. The reaction mixture was stirred at 100 °C for 24 hours under
argon. The mixture was poured into 500 mL water (pH 6), the precipitate
was collected by filtration and washed with acidic water (pH 6). The crude
product was dried in a vacuum oven, and subsequently purified by column
chromatography (SiO2, CH2Cl2:EtOAc from 20% to 40% EtOAc in 15
column volumes) to yield compound 1 (yellow powder, 1.55 g, 49% yield).
1H NMR (400 MHz, CDCl3, 97:3 E/Z mixture, signals of the main isomer)
δ: 7.86 (d, J = 8 Hz, 4H), 6.99 (d, J = 8 Hz, 4H), 5.00 (t, J = 8 Hz, 1H), 4.09
(q, J = 8 Hz, 2H), 3.71 (s, 3H), 3.58 (m, 2H), 2.56 (t, J = 8 Hz, 2H), 2.15 (p,
J = 8 Hz, 2H), 1.46 (s, 9H). 13C NMR (100 MHz, CDCl3) δ: 174.02, 161.09,
147.08, 124.58, 124.55, 114.84, 67.14, 51.95, 30.69, 28.51, 24.72. UV-vis
(CHCl3): ε358.5 = 6.77∙104 cm-1 M-1. FT-IR (ATR) ν (cm-1): 3281, 3060, 3004,
2979, 2969, 2930, 2061, 1985, 1908, 1735, 1719, 1687, 1595, 1542, 1496,
1460, 1439, 1386, 1361, 1329, 1302, 1275, 1249, 1207, 1176, 1144, 1107,
1064, 1048, 1039, 1017, 997, 939, 883, 869, 812, 779, 732, 695, 640, 585,
566, 556, 533, 518, 484, 464. MALDI-ToF-MS (m/z): [M+] calcd for
Figure 6. Depolymerization of azoBTA fibres by UV irradiation. A) UV-vis
absorption and B) CD spectra of the stepwise irradiation of a 15 μM azoBTA
solution in MCH at room temperature with small doses corresponding to 1 s at
10% of the LED intensity at 1000 mA of λ365nm shows the trans (blue) to cis
(yellow) photoisomerisation of the azobenzene in the azoBTA fibres. Super-
resolved iPAINT reconstructions before C) and after D) irradiation of a 300 μM
azoBTA solution with λ365nm (1 s at 50% of the LED intensity at 1000 mA) show
that the large fibrillar structures disappear upon exposure to UV light.
C
24H31N3O6 457.22; found 457.25.
Synthesis and characterisation of methyl 4-(4-((E)-(4-(2-(3,5-bis(((S)-
3,7-
dimethyloctyl)carbamoyl)benzamido)ethoxy)phenyl)diazenyl)pheno
xy)butanoate (2): 3,5-bis(((S)-3,7-dimethyloctyl)carbamoyl)benzoic acid
(400 mg, 0.82 mmol) was synthesised via
a previously reported
protocol,[24] and was suspended in dry DCM (15 ml) under argon. Ghosez’s
reagent (164 mg, 1.23 mmol) was added. The suspension was stirred for
3 hours and eventually became a clear solution. The volatiles were
removed under vacuum. Compound 1 (374 mg, 0.82 mmol) was dissolved
in 5% TFA in chloroform. After 3 hours of stirring, the volatiles were
removed under vacuum. The dried residue was redissolved in chloroform
(20 mL) and triethylamine (331 mg, 3.27 mmol) was added. This mixture
was stirred at RT for 10 minutes and subsequently added dropwise to a
solution of the acyl chloride residue in chloroform (10 mL) at 0 °C. After the
addition, the reaction was allowed to proceed at room temperature for 12
hours under argon. The solvent was removed and the crude product
purified by column chromatography (SiO2, Reveleris 120 g, CHCl3:EtOAc
from 20% to 40% EtOAc in 10 column volumes) to obtain compound 2
(yellow powder, 0.51 g, 75% yield). 1H NMR (400 MHz, CDCl3, 94:6 E/Z
mixture, signals of the main isomer) δ: 8.35 (d, J = 4 Hz, 2H), 8.33 (t, J =
4 Hz, 1H), 7.86 (d, J = 4 Hz, 4H), 7.01-6.96 (m, 5H), 6.43 (t, J = 8 Hz, 2H),
4.22 (t, J = 8 Hz, 2H), 4.08 (t, J = 8 Hz, 2H), 3.93 (q, J = 8 Hz, 2H), 3.70
We employ iPAINT super-resolution microscopy[23] to visualise
the fibres and confirm that the disappearance of the CD signal
upon UV irradiation is not explained by e.g. helix inversion.
Therefore, we study a 300 μM sample before and after a brief
irradiation (1 s) at λ = 365 nm, which should result in partial
depolymerisation. Before irradiation we observe fibres of several
microns long with a high polydispersity in length (Figure 6C and
Figure S25), while we find only sub-micron sized fragments after
1 s of irradiation (Figure 6D and Figure S25). Although the details
of the fragmentation process have not been elucidated, these
results once more suggest that we actively disrupt the fibres by
photoisomerising azobenzenes in the assembled state.
Conclusion
6
This article is protected by copyright. All rights reserved.

10.1002/chem.202004115
Chemistry - A European Journal
FULL PAPER
(s, 3H), 3.51-3.44 (m, 4H), 2.55 (t, J = 8 Hz, 2H), 2.15 (p, J = 8 Hz, 2H),
1.65-1.09 (m, 20H), 0.93 (d, J = 8 Hz, 6H), 0.85 (d, J = 8 Hz, 12H). 13C
NMR (100 MHz, CDCl3) δ: 173.71, 166.15, 165.69, 161.04, 160.38, 147.53,
147.17, 135.56, 134.95, 128.33, 128.17, 124.55, 124.53, 114.83, 114.79,
67.12, 66.87, 53.57, 51.83, 39.92, 39.38, 38.69, 37.26, 36.77, 30.90, 30.63,
28.09, 24.77, 24.72, 22.84, 22.74, 19.63. UV-vis (CHCl3): ε357 = 3.16∙104
cm-1 M-1. FT-IR (ATR) ν (cm-1): 3238, 3071, 2953, 2926, 2870, 1739, 1638,
1599, 1559, 1500, 1467, 1437, 1381, 1366, 1299, 1247, 1172, 1149, 1107,
1056, 1005, 943, 905, 885, 841, 805, 731, 692, 589, 551. MALDI-ToF-MS
(m/z): [M+Na+] calcd for C49H71N5O7841.54 ; found 864.53.
Commercial chemicals were obtained from TCI and/or Sigma Aldrich.
Used as received without further purification. Methylcyclohexane was
dried over mol sieves and by bubbling N2 before use.
CD and UV-vis spectra were recorded simultaneously on a Jasco J815
CD spectrometer, equipped with a PTC-423s/5 Peltier cell holder.
Measuring parameters were chosen appropriately. A sealable quartz cell
was used. The spectra in Figure 2A, C, S17 were recorded in continuous
mode between 550-190 nm at every 2 nm, with a sensitivity of 100 mdeg,
at 100 nm min−1. The integration time was set to 0.25 s. The temperature
was changed at 10 °C hr−1. The spectra in Figure 6A and B were recorded
simultaneously and in continuous mode between 550-190 nm at every 0.5
nm, with a sensitivity of 100 mdeg, at 1000 nm min−1.The response time
was set to 0.125 s, and spectra were recorded at an interval of 1 min. The
spectra in Figure S22 were recorded continuously between 450-350 nm
and 270-200 nm at every 0.5 nm, with a sensitivity of 100 mdeg, at 200 nm
min−1. The response time was set to 0.5 s, and spectra were recorded at
an interval of 1 min. The spectra in Figure S23 were recorded continuously
between 550-200 nm at every 0.5 nm, with a sensitivity of 100 mdeg, at
500 nm min−1. The response time was set to 0.5 s, and spectra were
recorded at an interval of 5 mins for the first 5 hrs, after which manually
another series was started at an interval of 10 mins.
Synthesis and characterisation of 4-(4-((E)-(4-(2-(3,5-bis(((S)-3,7-
dimethyloctyl)carbamoyl)benzamido)ethoxy)phenyl)diazenyl)pheno
xy)butanoic acid (3): Compound 2 (480 mg, 0.58 mmol) was dissolved in
1,2-dichloroethane (10 mL), and trimethylstannanol (6.12 mg, 3.38 mmol)
was added. The mixture was stirred at 70 °C for 24 hours. The solvent was
removed and the crude product dried in a vacuum oven. The dried crude
was loaded on a paper filter and washed copiously with water to remove
Sn by-products. The obtained product (463 mg, 0.57 mmol, 98% yield) was
dried overnight in a vacuum oven. 1H NMR (400 MHz, CDCl3) δ: 8.36 (d, J
= 4 Hz, 2H), 8.34 (t, J = 4 Hz, 1H), 7.85 (d, J = 8 Hz, 4H), 7.02-6.97 (m,
5H), 6.39 (t, J = 8 Hz, 2H), 4.23 (t, J = 8 Hz, 2H), 4.08 (t, J = 8 Hz, 2H),
3.93 (q, J = 8 Hz, 2H), 3.55-3.41 (m, 4H), 2.52 (t, J = 8 Hz, 2H), 2.12 (p, J
= 8 Hz, 2H), 1.57-1.10 (m, 20H), 0.94 (d, J = 8 Hz, 6H), 0.86 (d, J = 8 Hz,
12H). 13C NMR (100 MHz, CDCl3) δ: 178.61, 166.11, 165.64, 161.27,
160.34, 147.57, 147.08, 135.55, 134.94, 128.33, 128.17, 124.53, 124.51,
114.85, 114.83, 67.52, 66.90, 53.57, 39.93, 39.39, 38.69, 37.27, 36.78,
31.30, 30.90, 28.10, 25.47, 24.77, 22.85, 22.74, 19.64. UV-vis (DMSO):
ε364 = 1.58∙104 cm-1 M-1. FT-IR (ATR) ν (cm-1): 3244, 3072, 2954, 2926,
2869, 1641, 1596, 1562, 1500, 1468, 1383, 1297, 1248, 1198, 1148, 1106,
1052, 908, 841, 773, 731, 692, 548. MALDI-ToF-MS (m/z): [M+Na+] calcd
for C47H67N5O7 814.08; found 836.49.
UV-vis spectra (Figure 5) were recorded on a Shimadzu 2700 UV-visible
spectrophotometer. Spectra were recorded between 500-290 nm with 1
nm sampling. The wavelength was continuously changed ‘fast’ during the
recording of a spectrum, and spectra were recorded at an interval of 60s.
The light source was changed at 310 nm. A sealable quartz cell with a path
length of 1 cm was used. The temperature was held constant at 20 °C.
IR spectra were recorded on a Perkin Elmer Spectrum One 1600 FT-IR
spectrometer or a Perkin Elmer Spectrum Two FT-IR spectrometer,
equipped with a Perkin Elmer Universal ATR Sampler Accessory. FT-IR
spectra in liquid were recorded in a sealed cell with CaF2 windows.
Synthesis and characterisation of 3,4,5-tris(dodecyloxy)benzyl4-(4-
((E)-(4-(2-(3,5-bis(((S)-3,7-
dimethyloctyl)carbamoyl)benzamido)ethoxy)phenyl)diazenyl)pheno
xy)butanoate (azoBTA): Compound 3 (160 mg, 0.2 mmol) was dissolved
in dry DCM (15 mL) under argon, then Ghosez reagent (40 mg, 0.30 mmol)
was added. The solution was stirred for 3 hours and the volatiles were
removed under vacuum. The crude material was redissolved in dry DCM
(10 mL) and the solution cooled down to 0 °C (ice bath). Triethylamine (80
μL, 0.59 mmol) was added, followed by a dropwise addition (syringe, 5
minutes) of 3,4,5-tris(dodecyloxy)benzyl alcohol (131 mg, 0.2 mmol)
dissolved in dry DCM (5 mL). The solution was stirred for 10 minutes at
0 °C, and then at room temperature for 12 hours. The solvent was removed
and the crude product purified by column chromatography (SiO2, Reveleris
Silica 80 g, CHCl3:EtOAc from 0% to 40% EtOAc in 15 column volumes)
to obtain compound 4 (yellow powder, 60 mg, 20% yield). 1H NMR
(400MHz, CDCl3, 88:12 E/Z mixture, signals of the main isomer) δ: 8.35 (d,
J = 4 Hz, 2H), 8.33 (t, J = 4 Hz, 1H), 7.85 (d, J = 8 Hz, 4H), 7.15 (t, J = 8
Hz, 1H), 6.98 (dd, J1 = 8 Hz, J2= 12 Hz, 4H), 6.57 (t, J = 8 Hz, 2H), 6.54 (s,
2H), 5.03 (s, 2H), 4.21 (t, J = 8 Hz, 2H), 4.08 (t, J = 8 Hz, 2H), 3.96-3.90
(m, 8H), 3.50-3.43 (m, 4H), 2.60 (t, J = 8 Hz, 2H), 2.16 (p, J = 8 Hz, 2H),
1.81-1.40 (m, 20H), 1.30-1.26 (m, 56H), 1.14 (t, J = 8 Hz, 6H), 0.92 (d, J =
8 Hz, 6H), 0.87 (t, J = 8 Hz, 9H), 0.86 (d, J = 8 Hz, 12H). 13C NMR (100MHz,
CDCl3) δ: 173.09, 166.27, 165.80, 160.98, 160.38, 153.34, 147.48, 147.14,
138.35, 135.54, 134.93, 130.84, 128.35, 128.20, 124.53, 124.51, 114.79,
114.73, 107.12, 73.55, 69.26, 67.09, 66.93, 66.82, 39.90, 39.36, 38.68,
37.25, 36.74, 32.07, 32.06, 30.93, 30.89, 30.47, 29.88, 29.87, 29.83, 29.79,
29.78, 29.75, 29.56, 29.53, 29.50, 28.06, 26.26, 26.24, 24.75, 22.82, 22.72,
19.60, 14.25. UV-vis (CHCl3): ε358.5 = 2.61∙104 cm-1 M-1. FT-IR ν (cm-1):
3239, 3072, 2922, 2853, 1735, 1638, 1596, 1558, 1501, 1467, 1439, 1380,
1332, 1299, 1245, 1149, 1115, 1057, 948, 904, 841, 721, 692, 550.
MALDI-ToF-MS (m/z): [M+Na+] calcd for C90H145N5O101457.17; found
1480.10.
1H NMR and 13C NMR spectra were recorded either on a Varian Mercury
Vx 400 MHz (100 MHz for 13C) or Varian Oxford AS 500 MHz (125 MHz
for 13C) NMR spectrometers. Chemical shifts are given in ppm (δ) values
relative to residual solvent or tetramethylsilane (TMS). Splitting patterns
are labelled as s, singlet; d, doublet; t, triplet; q, quartet; p, pentet; m,
multiplet.
MALDI-ToF-MS (Matrix assisted laser desorption/ionisation time of flight
mass spectroscopy) was performed on a PerSeptive Biosystems Voyager
DE-PRO spectrometer or a Bruker autoflex speed spectrometer using α-
cyano-4-hydroxycinnamic acid (CHCA) and 2-[(2E)-3-(4-tert-butylphenyl)-
2-methylprop-2-enylidene]malononitrile (DCTB) as matrices.
Karl-Fisher titrations were carried out on
a Mettler-Toledo C30
Coulometric KF titrator containing CombiCoulomat Frit KF reagent.
Samples of approximately 0.5 ml were injected directly into the medium,
and averages of at least 2 measurements were taken as the water content.
SAXS (Small-angle X-ray scattering) measurements were performed on a
Saxslab Ganesha vacuum system with a Pilatus 300k solid-state photon-
counting 2D-detector and a high brilliance Microfocus Cu radiation source,
Genix3D, wavelength λ = 1.54184 Å. Measurements were obtained in
transmission mode using a sample-to-detector distance of 80 mm for
WAXS and up to 1400 mm for SAXS. A silver behenate standard was used
to calibrate the q-scale. The samples were prepared in 2 mm quartz glass
capillaries. Fitting of the data with a core-shell cylinder model[19], with a
scattering length density of 3.8 ± 0.02 ∙10−6 Å−2 for the core, 1.2 ± 0.01 02
∙10−6 Å−2 for the shell and 1.1 ± 0.01 02 ∙10−6 Å−2 for the solvent, yields a
radius of 13.8 Å for the core and a shell thickness of 19.0 Å. The cylinder
length L is longer than the attainable resolution of the instrument.
7
This article is protected by copyright. All rights reserved.

10.1002/chem.202004115
Chemistry - A European Journal
FULL PAPER
Super-resolution microscopy. A recently reported protocol[23] for iPAINT
imaging[25] is modified and used to image the adsorbed supramolecular
fibres in DMSO.[26] Briefly, a 300 μM solution of azoBTA in MCH was
flushed into the imaging cell and dried under a gentle stream of nitrogen.
The dried sample was stained by flushing 0.5 % v/v of Cage-552 (10 mM
in DMSO) and 1 % v/v of i-PrOH in MCH into the imaging cell. iPAINT
images were acquired on an inverted N-Storm Nikon microscope equipped
with a λ = 561 nm laser (≈490 mW cm−2), and a λ = 405 nm laser (≈160
mW cm−2). The incident light passes through a quad-band pass dichroic
mirror (95335 Nikon) and is focused on the sample with a Nikon objective
(oil immersion, 100×, NA = 1.49). The sample illumination occurs in a
quasi-Total Internal Reflection Fluorescence (TIRF) geometry. iPAINT
images were taken on 256×256 pixels region of interest at an acquisition
frame rate of 47 fps on an EMCCD Ixon3 Andor camera (pixel size 17×17
μm). Photons are collected over 5 × 103 frames during which 0.5 % of the
UV laser line is illuminating the sample continuously, along with 100 %
power of the 561 nm laser. The UV laser is used to stochastically uncage
a small amount of the caged (dark, non-fluorescent) probes warranting a
spatial separation greater than the diffraction limit of light. UV irradiation
turns the dark probes into an open (bright, fluorescent) conformational
state of which the absorption maximum falls in the visible region of the
spectrum, enabling excitation and bleaching by the 561 nm laser. The
localisation of single molecules is carried out by NIS-element Nikon
software. The super-resolved iPAINT images were corrected for
background localisations using a density-based algorithm. The point-cloud
of localisations is screened for all data points with n neighbours within a
certain area of radius δ. With the iPAINT images being a 2D projection of
3D localisations, the density of single molecules identified along the fibres
is higher than those localised on the coverslip, granting a straightforward
identification of the supramolecular structures.
Dutch Ministry of Education, Culture and Science (Gravity
Program 024.001.035) for financial support. The ICMS Animation
Studio (Eindhoven University of Technology) is acknowledged for
providing the artwork for Figure 1. GMP acknowledges the
funding received by the Swiss National Science Foundation
(SNSF grant number 200021_175735) and by the European
Research Council (ERC) under the European Union’s Horizon
2020 research and innovation programme (grant agreement no.
818776 – DYNAPOL). M.M.R.M. Hendrix is acknowledged for
performing the SAXS measurements, and A. Aloi for performing
the iPAINT imaging.
Keywords: supramolecular polymer • photoswitch • responsive •
photoisomerisation • benzene-1,3,5-tricarboxamide
[1]
a) S. Cantekin, T. F. A. de Greef, A. R. A. Palmans, in
Chemical Society Reviews, Vol. 41, 2012, pp. 6125-6137;
b) S. Yagai, K. Ohta, M. Gushiken, K. Iwai, A. Asano, S.
Seki, Y. Kikkawa, M. Morimoto, A. Kitamura, T. Karatsu, in
Chemistry - A European Journal, Vol. 18, 2012, pp. 2244-
2253.
[2]
a) B. Adhikari, Y. Yamada, M. Yamauchi, K. Wakita, X. Lin,
K. Aratsu, T. Ohba, T. Karatsu, M. J. Hollamby, N. Shimizu,
H. Takagi, R. Haruki, S.-i. Adachi, S. Yagai, in Nature
Communications, Vol. 8, Nature Publishing Group, 2017, p.
15254; b) E. Fuentes, M. Gerth, J. A. Berrocal, C. Matera,
P. Gorostiza, I. K. Voets, S. Pujals, L. Albertazzi, Journal of
the American Chemical Society 2020, 142, 10069-10078;
c) K. K. Kartha, N. K. Allampally, A. T. Politi, D. D. Prabhu,
H. Ouchi, R. Q. Albuquerque, S. Yagai, G. Fernández, in
Chemical Science, Vol. 10, The Royal Society of Chemistry,
2019, pp. 752-760.
Sample preparation. For UV-vis and CD spectroscopy (15 μM), solution-
IR spectroscopy (0.5 mM), SAXS (3 mM), and iPAINT and 1H NMR (0.3
mM) analysis, azoBTA (4) or the model compound azoBTA-CO2Et (5,
Figure S21) was dissolved at target concentration in dried MCH at 75 °C,
or in CHCl3 at room temperature, and allowed to equilibrate at rt for at least
24 hrs before measurements.
[3]
a) R. Klajn, P. J. Wesson, K. J. M. Bishop, B. A. Grzybowski,
in Angewandte Chemie International Edition, Vol. 48,
WILEY‐VCH Verlag, 2009, pp. 7035-7039; b) S. Lee, S.
Oh, J. Lee, Y. Malpani, Y.-S. Jung, B. Kang, J. Y. Lee, K.
Ozasa, T. Isoshima, S. Y. Lee, M. Hara, D. Hashizume, J.-
M. Kim, in Langmuir, Vol. 29, 2013, pp. 5869-5877; c) O. M.
Wani, R. Verpaalen, H. Zeng, A. Priimagi, A. P. H. J.
Schenning, in Advanced Materials, Vol. 31, 2019, pp.
1805985-1805985.
Molecular Dynamics simulations: Modelling and simulation
procedures. The atomistic model of the azoBTA monomer has been
parametrised in the framework of GAFF (General Amber Force Field).[20]
The force field parameters for the BTA core and alkyl groups have been
[4]
S. S. D. Lafleur, L. Shen, E. J. T. W. Kamphuis, S. J. A.
Houben, L. Balzano, J. R. Severn, A. P. H. J. Schenning, C.
W. M. Bastiaansen, in Macromolecular Rapid
Communications, Vol. 40, 2019, pp. 1800811-1800811.
J. Kim, J. Lee, W. Y. Kim, H. Kim, S. Lee, H. C. Lee, Y. S.
Lee, M. Seo, S. Y. Kim, in Nature Communications, Vol. 6,
2015, p. 6959.
C. Maity, W. E. Hendriksen, J. H. Van Esch, R. Eelkema, in
Angewandte Chemie - International Edition, Vol. 54, 2015,
pp. 998-1001.
J. W. Fredy, A. Méndez-Ardoy, S. Kwangmettatam, D.
Bochicchio, B. Matt, M. C. A. Stuart, J. Huskens, N.
Katsonis, G. M. Pavan, T. Kudernac, in Proceedings of the
National Academy of Sciences of the United States of
America, Vol. 114, National Academy of Sciences, 2017, pp.
11850-11855.
taken from previous simulation works on BTA supramolecular
21]
polymers.[11b,
The force field parameters for the additional trans
azobenzene unit in the azoBTA has been taken from other previous
works.[7, 27] The molecular models for the azoBTA stacks (Figure 3a in the
main paper) have been then inserted into simulation boxes filled with
explicit MCH molecules. The simulation boxes were built in such a way to
effectively model a portion of the bulk of infinite supramolecular polymers
via periodic boundary conditions (PBC). The all atom molecular dynamics
(AA-MD) simulations were performed with the GROMACS software (2016
version).[28] After an initial minimization, all AA-MD runs have been
conducted in NPT conditions (constant N: number of particles, P: pressure
and T: temperature) at 300 K (using the v-rescale thermostat[29] with a
coupling constant of 1 ps) and 1 atm (using semi-isotropic pressure scaling
to allow the fibres to fold) with a coupling constant of 1 ps. The Particle
Mesh Ewald (PME)[30] method was used to treat long-range electrostatics.
The LINCS algorithm was used to constrain all bonds involving
hydrogens.[31] For what concerns the analysis of the AA-MD simulation
results, we used GROMACS built-in analysis tools: gmx energy for
energetic analysis and gmx rdf for the calculation of radial distribution
functions.
[5]
[6]
[7]
[8]
[9]
J. Lee, S. Oh, J. Pyo, J.-M. Kim, J. H. Je, in Nanoscale, Vol.
7, 2015, pp. 6457-6461.
a) A. A. Beharry, O. Sadovski, G. A. Woolley, in Journal of
the American Chemical Society, Vol. 133, American
Chemical Society, 2011, pp. 19684-19687; b) D. Bléger, J.
Schwarz, A. M. Brouwer, S. Hecht, in Journal of the
American Chemical Society, Vol. 134, American Chemical
Society, 2012, pp. 20597-20600; c) D. Manna, T.
Udayabhaskararao, H. Zhao, R. Klajn, in Angewandte
Chemie International Edition, Vol. 54, 2015, pp. 12394-
12397.
Acknowledgements
[10]
a) R. P. M. Lafleur, X. Lou, G. M. Pavan, A. R. A. Palmans,
E. W. Meijer, in Chemical Science, Vol. 9, The Royal
Society of Chemistry, 2018, pp. 6199-6209; b) Y. Nakano,
The authors acknowledge the Netherlands Organisation for
Scientific Research (NWO VIDI Grant 723.014.006) and the
8
This article is protected by copyright. All rights reserved.

10.1002/chem.202004115
Chemistry - A European Journal
FULL PAPER
[31]
B. Hess, H. Bekker, H. J. C. Berendsen, J. G. E. M. Fraaije,
in Journal of Computational Chemistry, Vol. 18, John Wiley
& Sons, Ltd, 1997, pp. 1463-1472.
T. Hirose, P. J. M. Stals, E. W. Meijer, A. R. A. Palmans, in
Chemical Science, Vol. 3, The Royal Society of Chemistry,
2012, pp. 148-155.
[11]
a) D. Bochicchio, G. M. Pavan, in ACS Nano, Vol. 11,
American Chemical Society, 2017, pp. 1000-1011; b) M.
Garzoni, M. B. Baker, C. M. A. Leenders, I. K. Voets, L.
Albertazzi, A. R. A. Palmans, E. W. Meijer, G. M. Pavan, in
Journal of the American Chemical Society, Vol. 138,
American Chemical Society, 2016, pp. 13985-13995; c) A.
J. Markvoort, H. M. M. ten Eikelder, P. A. J. Hilbers, T. F. A.
de Greef, E. W. Meijer, in Nature Communications, Vol. 2,
2011, p. 509.
[12]
[13]
a) L. Albertazzi, N. van der Veeken, M. B. Baker, A. R. A.
Palmans, E. W. Meijer, in Chemical Communications, Vol.
51, The Royal Society of Chemistry, 2015, pp. 16166-
16168; b) C. M. A. Leenders, L. Albertazzi, T. Mes, M. M.
E. Koenigs, A. R. A. Palmans, E. W. Meijer, in Chemical
Communications, Vol. 49, The Royal Society of Chemistry,
2013, pp. 1963-1965.
a) M. A. J. Gillissen, T. T. Hoeben, A. J. H. Spiering, J. A. J.
M. Vekemans, A. R. A. Palmans, E. W. Meijer, in Israel
Journal of Chemistry, Vol. 51, 2011, pp. 1118-1127; b) T.
Mes, M. M. J. Smulders, A. R. A. Palmans, E. W. Meijer, in
Macromolecules, Vol. 43, American Chemical Society,
2010, pp. 1981-1991.
[14]
[15]
[16]
[17]
[18]
K. Kreger, P. Wolfer, H. Audorff, L. Kador, N. Stingelin-
Stutzmann, P. Smith, H.-W. Schmidt, in Journal of the
American Chemical Society, Vol. 132, 2010, pp. 509-516.
M. M. J. Smulders, A. P. H. J. Schenning, E. W. Meijer, in
Journal of the American Chemical Society, Vol. 130, 2008,
pp. 606-611.
M. M. J. Smulders, T. Buffeteau, D. Cavagnat, M. Wolffs, A.
P. H. J. Schenning, E. W. Meijer, in Chirality, Vol. 20, 2008,
pp. 1016-1022.
M. Masuda, P. Jonkheijm, R. P. Sijbesma, E. W. Meijer, in
Journal of the American Chemical Society, Vol. 125, 2003,
pp. 15935-15940.
S. Yagai, Y. Goto, X. Lin, T. Karatsu, A. Kitamura, D.
Kuzuhara, H. Yamada, Y. Kikkawa, A. Saeki, S. Seki, in
Angewandte Chemie International Edition, Vol. 51, 2012,
pp. 6643-6647.
[19]
[20]
S. Kline, R. Heenan, 2013.
J. Wang, R. M. Wolf, J. W. Caldwell, P. A. Kollman, D. A.
Case, in Journal of Computational Chemistry, Vol. 25, 2004,
pp. 1157-1174.
[21]
M. B. Baker, L. Albertazzi, I. K. Voets, C. M. A. Leenders,
A. R. A. Palmans, G. M. Pavan, E. W. Meijer, in Nature
Communications, Vol. 6, 2015, p. 6234.
[22]
[23]
D. Bochicchio, S. Kwangmettatam, T. Kudernac, G. M.
Pavan, in ACS Nano, Vol. 13, 2019, pp. 4322-4334.
B. Adelizzi, A. Aloi, N. J. Van Zee, A. R. A. Palmans, E. W.
Meijer, I. K. Voets, in ACS Nano, Vol. 12, American
Chemical Society, 2018, pp. 4431-4439.
[24]
[25]
[26]
M. L. Ślęczkowski, E. W. Meijer, A. R. A. Palmans, in
Macromolecular Rapid Communications, Vol. 38, Wiley-
Blackwell, 2017, p. 1700566.
A. Aloi, N. Vilanova, L. Albertazzi, I. K. Voets, in Nanoscale,
Vol. 8, The Royal Society of Chemistry, 2016, pp. 8712-
8716.
A. Aloi, A. Vargas Jentzsch, N. Vilanova, L. Albertazzi, E.
W. Meijer, I. K. Voets, in Journal of the American Chemical
Society, Vol. 138, American Chemical Society, 2016, pp.
2953-2956.
[27]
[28]
L. Pesce, C. Perego, A. B. Grommet, R. Klajn, G. M. Pavan,
in Journal of the American Chemical Society, 2020, p.
jacs.0c03444.
M. J. Abraham, T. Murtola, R. Schulz, S. Páll, J. C. Smith,
B. Hess, E. Lindahl, in SoftwareX, Vol. 1-2, 2015, pp. 19-
25.
[29]
[30]
G. Bussi, D. Donadio, M. Parrinello, in The Journal of
Chemical Physics, Vol. 126, 2007, p. 014101.
T. Darden, D. York, L. Pedersen, in The Journal of
Chemical Physics, Vol. 98, 1993, pp. 10089-10092.
9
This article is protected by copyright. All rights reserved.

10.1002/chem.202004115
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents
The synthesis and characterisation of a chiral supramolecular motif are reported. The BTA motif self-assembles into helical
structures, wherein the hydrogen-bonded core is of opposite helicity from the surrounding double helix formed by azobenzenes. This
discordant structure is confirmed the most favoured conformation by MD simulations. The supramolecular fibres can be reversibly
(de)polymerised by thermal cues and light irradiation.
10
This article is protected by copyright. All rights reserved.