FULL PAPER
“This is the peer reviewed version of the following article: Adv. Mater. Interfaces. 2018, 5 (9), 1701585, which has been published
infinal form at DOI: 10.1002/admi.201701585. This article may be used for non-commercial purposes in accordance with Wiley Terms
and Conditions for Self-Archiving published at http://olabout.wiley.com/WileyCDA/Section/id-820227.html."

FULL PAPER
Enhanced Conductivity for Carbon Nanotube based Materials
through Supramolecular Hierarchical Self-Assembly
Gustavo A. Zelada-Guillén,*[a,b] Martha V. Escárcega-Bobadilla,*[a,c] Marcin Wegrzyn,[d] Enrique
Giménez,[e] Gerhard Maier[b] and Arjan W. Kleij*[c,f]

Abstract: The currently limited comprehension of hierarchical control
over out-of-equilibrium (dynamic) self-assembly processes in
nanoscience and nanotechnology has limited the exploitation of
multicomponent systems in the design of new nanostructured
functional materials. In our contribution, molecular building blocks with
tailored nanoscale anisotropic supramolecular self-assembly
behavior enable the creation of mesoscale percolation networks of
multiwall carbon nanotubes through collinear interconnections at the
microscale. This strategy affords polymeric composites with tunable
properties at the macroscale, where the organization mechanism is
regulated by dynamic self-assembly at 4 hierarchical levels of autoorganization. Such multilevel self-assembly system reduces up to 8fold the nanotube concentration required for percolation and
enhances conductivity up to 6 orders of magnitude against blanks,
thus yielding anisotropically semiconducting and conducting materials.
Our approach is based on casting-from-solution, thus simplifying
preparative steps when compared to state-of-the-art electron carrier
counterparts such as single-walled carbon nanotube-, graphene- or
indium-tin-oxide-based technologies. Finally, promising material
transparency levels can be reached across the visible and nearinfrared regimes for compositions above the percolation threshold,
which provides new opportunities beyond the current spectral
restrictions in commercial transparent conductors.

Introduction
The deployment of nanoscale objects into industrially viable
multicomponent functional materials requires for the

[a]

[b]
[c]

[d]

[e]
[f]

Dr. G. A. Zelada-Guillén, Dr. M. V. Escárcega-Bobadilla
School of Chemistry, National Autonomous University of Mexico
(UNAM), Circuito Escolar s/n, Ciudad Universitaria, Mexico City
04510 (Mexico). E-mail: g.zelada@unam.mx, mesbo@unam.mx
Dr. G. A. Zelada-Guillén, Dr. G. Maier; Polymaterials AG,
Innovapark 20, Kaufbeuren 87600 (Germany)
Dr. M. V. Escárcega-Bobadilla, Prof. A. W. Kleij
Institute of Chemical Research of Catalonia (ICIQ), the Barcelona
Institute of Science and Technology (BIST), Av. Paisos Catalans 16,
43007 – Tarragona (Spain). E-mail: akleij@iciq.es
Dr. M. Wegrzyn, Instituto Tecnológico del Plástico (AIMPLAS), C.
Gustavo Eiffel 4, Paterna 46980 (Spain). Current address: School of
Engineering, Engineering Research Institute, Jordanstown campus,
Shore Road, Newtownabbey, Co. Antrim, BT37 0QB, Ulster (UK)
Dr. E. Giménez, Instituto de Tecnología de Materiales, Universidad
Politécnica de Valencia, Valencia 46022 (Spain)
Prof. A. W. Kleij, Catalan Institute for Research and Advanced
Studies (ICREA), Pg. Lluis Companys 23, Barcelona (Spain)
Supporting information for this article is given via a link at the end of
the document.

simultaneous resolution of complex problems. Those problems
include optimal orientation, ordering and distribution of
nanostructures,[1] their hierarchical control at multiple scales[2]
(e.g. nanoscale, microscale, mesoscale and macroscale) and
cost-effectiveness of the raw materials[3] together with simple and
scalable processing steps. In this sense, the hierarchical selfassembly of molecular, macromolecular or nanostructured
building blocks (BBs) via supramolecular interactions has
emerged as a promising strategy that allows for spontaneous
control of matter at different scales. Hierarchically organized
systems have found many uses including recognition elements in
theranostic tools,[4] stimuli-responsive parts in soft materials,[5]
transducers and mesoporous components in chemical sensors, [6]
hybrid bio-organometallic assemblies in optical switches,[7]
multilayer topologies in energy conversion/storage systems[8] and
patterned conductors[9] in electronic devices.
However, transparent electrically conducting components in
optoelectronic systems still depend on metal oxide technologies
(rather than hierarchical self-assembly designs) where indium-tin
oxide (ITO) has played a leading role for several decades. [10] Most
of the recently explored alternatives in transparent conductors
include low-dimensional metal-oxide- or carbon-based electrically
conducting nanostructures such as nanoparticles, nanowires,
nanorods and nanosheets.[11] Unfortunately, the development of
those alternatives by precisely controlling these structures at the
nanoscale has remained a difficult task so far. [5,9c,12] An interesting
alternative is the use of highly-stable, non-conducting transparent
polymers (e.g. polymethyl methacrylate, PMMA; polyethylene
terephthalate, PET; polystyrene, PS; polycarbonate, PC) as
matrices in multicomponent (composite) material films or as
substrates for randomly deposited 1D and 2D electron carriers
such as carbon nanotubes (CNTs) or graphene. [13]
In the case of CNTs, the matrix may contribute to a high
physical resistance in the film while the CNTs should work as
nanoscale ‘electrical wires’ when they exceed a percolation
threshold (pc), i.e. the limit beyond which an efficient
interconnection between individual electron carriers reaches the
dimensions of the system. Nowadays, CNT-based conducting
systems rely on single-walled carbon nanotubes (SWCNT)
because they are easier to solubilize, offering good conductivity
values ( from 10-2 S to 10-3 S) and allowing for transparency
degrees between 70 % and 90 %. The latter performance has not
been possible when their yet much less expensive and industrially
produced
multi-walled
counterparts
(MWCNT)
are
used,[12a,12d,13b,14] even if hierarchical topologies (i.e. twisted
yarns)[15] are employed.
However, in all the examples using MWCNTs, higher  values
or larger percolation networks were only possible if multiple
nanotube layers or higher loadings thereof were used.

FULL PAPER
Nevertheless, in microscale/mesoscale systems the film
dimensions exceed the length of the nanotubes and the last
option results unpractical; the higher the loadings the higher the
amount of agglomerates which negatively affect percolation at the
macroscale.[16] Conversely, electrical percolation in the bulk also
depends on the nanotube-nanotube connectivity angle average at
the nanoscale, with fewer collinearly interconnections, the lower
the volumetric conductivity (v).[17] Unfortunately, the achievement
of a balance between collinearity in interconnections at the
nanoscale, adequate dispersion at the microscale and
preservation of a continuous conductive network at the
mesoscale, have remained elusive so far. In a recent work,
chemical unzipping of MWCNTs was used to produce nanotubes
bridged by graphene nanoribbons in conductive 2D networks
supported on fluorine-doped tin oxide substrates, where
percolation depended on the unzipping degree. [18] On the contrary,
in solid matrices, percolation among carbon nanostructures
depends on stochastic and entropic factors, which has limited the
development of high-throughput processes.

how translating self-assembly principles at different levels of
organization and designing new materials that enable the
potential of new conducting systems with promising transparency
levels.

Results and Discussion
Mechanism of self-assembly
For the self-assembly process, we employed two tetrakis-Schiff
base metallic complex variants, 1 and 2 (which is a version of 1
with an extended aromatic surface), in systems meeting the
requisites to render the same self-assembly behavior (Figure 2,
see supporting information). These facile to prepare and easily
modulated BBs[20] were used to provide a potential solution for the
issues of CNT debundling and their collinear interlinking in a
polymer matrix through self-organization.

Figure 1. Supramolecular synthon based on a modular tetrakis-Schiff base
metal complex able to work as a self-assembly BB. In the structure: R = R’ =
Aromatic, M = Cu(II), Ni(II) or Zn(II).

Nature offers examples where supramolecular self-assembly
has evolutionary served as the simplest route to overcome such
entropic limitations in out-of-equilibrium (dynamic) systems. This
can translate, for instance, molecular information into adaptive
patterns in information delivery networks such as in the
cytoskeleton in eukaryotic cells and neurons in animal nervous
systems. Recently, we designed a supramolecular synthon that
self-assembles into nanoarchitectures that mimic at the
mesoscale webs of neurons in the brain either on surfaces or in
polymer matrices by simple casting-from-solution.[19] In these
webs, dispersed nanorings (the ‘neurons’) are interlinked by long
slender nanowires (the ‘axons’/’dendrites’), where the BB is
derived from a modular tetrakis-Schiff base complex (Figure 1).
However, in those preliminary studies, the mechanism by which
the system could be potentially used as a self-assembly template
for creating networks of CNTs in the mesoscale through their
relocation from random clusters remained essentially unknown.
More particularly the exact nature of the relationship between the
nanoscale-controlled microscale self-assembly process of the
Schiff base complex and the directed organization of the CNTs on
the macroscale properties such as electrical percolation or optical
transparency in the UV-VIS-NIR region remained an open
question.
In this contribution we describe a detailed mechanistic
manifold helping to increase the fundamental understanding of

Figure 2. BBs used in this work: a) supramolecular synthons 1 and 2, b) network
self-assembly behavior observed by TEM for a dropcast solution prepared with
2 (70 m) in DCM.

Dynamic light scattering (DLS) and transmission electron
microscopy (TEM) studies show that the auto-organization of
nanotubes is mediated by interfacial mass transfer and energy
dissipation during solvent evaporation through hierarchical selfassembly involving four levels of organization.
This multilevel mechanism (Scheme 1) explains the features
observed both in drop-cast dichloromethane (DCM) solutions
without polymer and in polymer films after solvent evaporation. In
a polymer solution, the nanotube bundles are individualized by a
superficially self-assembled multilayer of the BBs (level 1) at the

FULL PAPER
nanoscale via - interactions and van der Waals forces. During
solvent evaporation, the synthon concentration raises to a critical
concentration beyond which they kinetically favor the formation of
microscale patchy vesicles with nanotubes adhered to their
surface by integrative self-assembly (level 2).
During evaporation, the gradual reduction in solvent volume
progressively increases concentration of the components. Such a
minimization in the continuous medium promotes mutual
adhesion of the vesicles until a close-packed arrangement is
reached beyond the microscale. Afterwards, minimization in
surface tension at the vesicle-solvent interface arrives with their
collapse through toroidal intermediaries (circular-shaped
topologies without matter in their geometric center). This collapse
displaces the nanotubes from the vesicle surface to the regions
which should be the precursors for the rings (level 3). At this latter
stage, the aggregates of collapsed vesicles give birth to the
nanowires via interfacial migration of matter that emerges from
the contact points between vesicles. Simultaneously, the same
migration and rearrangement of matter displaces neighboring
CNTs from the vesicle edges to the nanowires and distributes the
nanotubes along the rim of the final rings, favored by the shape
anisotropy of the nanotubes. This process simultaneously occurs
throughout the whole system thus promoting formation of longrange networks at the mesoscale (level 4).

Scheme 1. Mechanism of self-assembly in a polymer solution during solvent
evaporation until dryness. Level 0, BBs in a free form and CNTs in random
clusters. Level 1, CNTs covered by a nanoscale self-assembled multilayer of
BBs in solution (for clarity purposes, only the closest layer to CNT surface is
shown). Level 2, integrated vesicles at the microscale from BBs and CNTs at
their surface. Level 3, a large number of intermediary patchy vesicles are in
contact beyond the microscale and afford collapsed systems as the solvent
evaporates. Level 4, after complete evaporation, an interconnected network of
CNTs in a polymer matrix at the mesoscale.

Dynamic self-assembly in solution
As discussed earlier, the self-assembly of BBs should be related
with the evaporative process. Therefore, the type and size of
aggregates at different concentrations and compositions would
provide useful information to determine the dynamic route
towards auto-organization of matter in solution. To study the
dynamic self-assembly of the systems, DLS was used to assess
the size distribution profiles of aggregates in DCM (Figure 3) for

samples containing (i) the BBs 1 or 2, (ii) the MWCNTs and (iii)
their mixtures at different concentrations.
Loss of larger MWCNT agglomerates (at 20 g/mL) was
observed when 1 or 2 was present at a 70 M concentration at
which long range self-assembled networks can be formed after
evaporation with intermediate concentrations between 1 M and
100 M. Interestingly, only unimodal size distribution profiles were
detected in these mixtures, where the aggregate size values were
in the same order of magnitude as found for single nanotubes.
The trend suggests the predominance of individual nanotubes in
solution and the lack of higher order agglomerates. However, a
fourfold increase in the concentration of the previous mixtures at
a solvent evaporation degree of 75 % showed a threefold increase
in average size of aggregates preserving a unimodal distribution.
Recently, the role of synthon aromaticity of porphyrin
oligomers in their supramolecular interactions with the aromatic
surface of SWCNTs was spectroscopically demonstrated via -
interactions and van der Waals forces.[21] The study was possible
taking the advantage of the small size of the nanotubes used and
their consequently facilitated solubility in a different solvents. In
our systems, both complexes 1 and 2 possess -conjugation
potential across the ligand framework, while the surface of the
MWCNTs are of the same aromatic nature as any of their singlewalled counterparts. Unfortunately, the lack of solubility in DCM
of these relatively large nanotubes represent a limitation to further
spectroscopic monitoring of BB-CNT interactions. In this regard,
DLS analysis is a powerful alternative to the latter, by comparing
the aggregate size distributions in BB solutions before and after
including the nanotubes, in order to determine whether an
association between BBs and CNTs may occur in DCM. When
pure BB solutions are analyzed, it is possible to observe that not
only large self-assemblies are present (small vesicles, based on
further TEM analysis), but also mono-to-oligomolecular entities
smaller than ca. 4 nm occur, which indicates the presence of a
dynamic equilibrium between both aggregation types (i.e.
intermediate sizes might be kinetically unfavored). After including
nanotubes in solution, the smaller entities disappear while the
unimodal size distribution mentioned in the previous paragraph is
the only observed aggregation, thus indicating a displaced
equilibrium towards the association between BBs and CNTs.
Taking into account the -conjugation in both BBs and nanotubes,
a similar interaction phenomena as in Ref. 21 should be expected.
Therefore, the observed trend in size regulation of the aggregates
governed by the composition of the samples suggests that
individualization of nanotubes in solution can be addressed by the
aromatic surfaces in the BBs 1 and 2. This implies that these BBs
might interact with the external nanotube walls via interfacial van
der Waals and - interactions and this process is dynamically
controlled by the evaporation degree. We expect that in future
studies, spectroscopic titrations using easier-to-solubilize
SCWNTs and similar BBs might be helpful to finally confirm the
latter.

FULL PAPER

Figure 3. Dynamic self-assembly of BBs 1 or 2, and MWCNTs in solution. a)
Size of aggregates from 1 and 2 in DCM solutions (70 M); b) size distribution
of nanotube dispersions at 20 g/mL in DCM and their counterparts when 1 or
2 are also included at 70 M, compared with a threefold increase in size by
increasing 4 times the concentration of the pristine mixtures at a solvent
evaporation degree of 75 %.

Hierarchical auto-organization as a function of solvent
evaporation
The hierarchical self-assembly of the components after
evaporation was further studied to evaluate the possibility of
dynamical control in size regulation of the aggregates. The same
solutions used for the DLS studies were analyzed by transmission
electron microscopy (TEM) and compared with samples having
different ratios of MWCNT and complex 2 as a model system for
the self-assembly mechanism at different evaporation stages. At
low MWCNT concentration (20 g/mL), the BB 2 is involved in a
supramolecular wrapping process on the nanotube surface at the
nanoscale, where the BB multilayer thickness () depends on the
solution concentration (Figure 4a). In this process, the nanotubes
are also individualized (Figure 4b) by such nanometer-sized
homogeneous interfacial self-assembled multilayer of 2
(hierarchy level 1).

Figure 4. Hierarchy levels 1 and 2 in self-assembly for selected mixtures and
the effect of increasing concentration of components. a) In self-assembly level
1 for mixtures containing nanotubes (20 g/mL) and 2 (70 M), average doubles
as the concentration of the mixture increases four times: e.g. in the same selfassembly level (b), the self-assembled multilayer of molecule 2 (70 M) covers
the walls of the nanotubes (20 g/mL) to yield their individualization, while (c)
higher concentration of the mixture (fourfold) provides small CNT aggregates
with a thicker multilayer of self-assembled 2 (higher ); d-e) depending on the
concentration of nanotubes, self-assembly level 2 becomes more evident, for
example, integration of vesicles from 2 (70 M) is carried out with individual (d)
or aggregates (e) of nanotubes (20 g/mL and 80 g/mL, respectively). Panels
(b-e) are TEM images of dropcast DCM solutions after evaporation.

A fourfold concentration of a mixture of CNT and BB follows
the same trend yet yielding a small number of self-assembled
MWCNTs, where each nanotube preserves the superficial
multilayer (Figure 4c) with a two-fold increase in its thickness. As
the solvent evaporates, the BBs generate vesicles at the
microscale (hierarchy level 2) through integrative self-assembly,
thus leaving the wrapped nanotubes self-assembled onto the
surface. In those vesicles, the number of nanotubes per vesicle is
apparently proportional to the concentration of CNTs in the
original solution. For example, nanotubes at 20 g/mL including
BB 2 at 70 M produce vesicles occasionally bearing up to three
self-assembled nanotubes (Figure 4d). However, solutions with
nanotubes having a fourfold of that concentration but the same
BB content, generate similar vesicles containing larger numbers
of nanotubes (i.e. five self-assembled nanotubes, Figure 4e).
Those nanotubes self-assembled to the vesicles can also drag
non-self-assembled nanotubes that are mechanically attached to
the former ones. This behavior is similar to the mechanism of the

FULL PAPER
‘hook-and-loop’ fasteners (or Velcro® tapes). The nanotubes
which are self-assembled to the vesicle surface play the role of
the ‘hooks’, whereas the dragged nanotubes act as the ‘loops’.
Hence, in solution, the vesicles perform a triple role involving
recollection, transportation and dispersion of the nanotubes,
which also explains the unimodal distribution profiles observed by
DLS.

Figure 5. Hierarchy levels 3 and 4 in self-assembly for selected mixtures. Level
3 arises from mutual adhesion of vesicles, which are still organized at level 2
(a), while further relocation of matter at hierarchy level 3 proceeds through
toroidal intermediaries (b). At level 4, the network formation addresses a
relocation of the nanotube positions to the assemblies where matter from 2 is
present (c): d) nanorings from 2 contain nanotubes tangentially distributed
across the curvature; e-f) nanowires from 2 promote collinear self-assembly of
nanotubes. Panels are TEM micrographs from dropcast DCM solutions at
selected compositions: 2 at 70 M (a-b), CNTs at 80 g/mL and 2 at 70 M (cf), (d) is an amplification from dashed box in (c).

Regardless the nanotube concentration in the original solution,
solvent elimination simultaneously results into mutual adhesion of
a large number of vesicles (previously organized at hierarchy level
2) at larger areas in the microscale (hierarchy level 3, Figure 5a).
The same evaporative process results into a collapse in vesicle
shape through toroidal intermediaries (e.g. Figure 5b) and a mass
migration process that results in the final network which occurs
either in presence or absence of CNTs (see supporting
information). During the collapsing of the vesicle, the superficial
MWCNTs migrate to the edges to give place to the rim of the rings
(Figure 5c-d). Simultaneously, the matter present at the contact
points between the vesicles generates nanowires (Figure 5e-f)
promoting the reorientation of the nanotubes that remained close
to the nanowire-growing points to yield collinear self-assemblies
assisted by the anisotropy in shape of both nanotubes and
nanowires (hierarchy level 4). The nanowire growth advances as
long as the amount of neighboring material allows for an
equilibrium point between the final size of the rings and the length

of the nanowires. Thus, the relocation of matter evolves until
equilibrium is reached between the availability of displaceable
material and the mechanical strength of the network, which in
optimal conditions may reach mesoscale dimensions. In any case,
in the final network (hierarchy level 4), the nanotubes remain
covered by the originally self-assembled multilayer of the BB
which occurred since earlier stages (e.g. in hierarchy level 1,
Figure S2).
In our previous work,[19] DFT and molecular dynamics,
together with surface-enhanced Raman scattering microscopy,
high-resolution TEM and nanobeam electron diffraction,
demonstrated that this type of self-assembled BBs were
anisotropically oriented along the longitudinal direction of the
nanowires. Those studies also revealed that, on the contrary, BBs
were radially auto-organized to the center of the nanorings and
concentrically arranged to their rim. Nevertheless, stabilization of
these two architectures in the network was possible by four types
of intermolecular orientations upon self-assembly. These
molecule-to-molecule orientations, namely, back-to-back,
stacked, face-to-face and side contact, were derived from the high
aromaticity and wedge shape in the BBs and also stabilized the
formation of vesicles and oligomers in solution. In the present
study, we speculate that those four intermolecular orientations
may coexist at each hierarchy level and are responsible for the
stabilization of the interconnected nanotubes via  interactions at
hierarchy level 4 after evaporation. Moreover, the same oriented
associations would also explain the favored presence of two
particle sizes for BBs in DCM (Fig. 3a), stable oligomers and
nanosized vesicles, where the last ones were also detected by
TEM. In this sense, the variation in  observed by TEM for the
superficial multilayer of BB on the nanotubes at different
concentrations (Fig. 4a-c) and the extinction of mono-tooligomolecular entities in solution observed by DLS for BB+CNT
mixtures (Fig. 3), may be explained by a nucleation mechanism
occurring at the beginning of level 1. In this process, the highly
aromatic surface of the nanotubes should play the role of a
nucleation site for self-assembly with free/oligomerized BBs,
while this preliminary step may facilitate further association of
additional molecules to earlier layers of the same, which in turn,
yield the observed multilayer. Interestingly, the growth observed
for the BB multilayer as a function of concentration is coherent
with the previous hypothesis and we expect that further
spectroscopic studies with single-walled counterparts may also
confirm this route.
Self-assembly
at
the micro
and
mesoscale
in
multicomponent materials.
The self-assembly in multicomponent systems was then
assessed using the previous casting-from-solution approach
applied in the preparation of polymer composite films with a
nominal thickness d of 100 m. This was done to assess
implications of the micro and mesoscale topologies after the selfassembly process on the macroscale properties such as electrical
percolation and optical transparency in further evaluations. The
films were prepared in combinatorial compositions containing one
of the BBs 1 or 2 and nanotubes at different concentrations.
Therefore, first stock DCM solutions used during the casting

FULL PAPER
procedure were prepared by including the selected BBs at the
same concentration as the experimental counterparts in the
absence of polymer.
Bis-phenol-A based polycarbonate (10 % m/v) was introduced
into those stock solutions as a transparent polymer matrix model.
After solvent evaporation, solid samples with a nominal content of
BB equal to 0.065 % wt ( 0.005 % wt) were obtained. A
microstructural analysis of the films (Figure 6) showed that the
BBs simultaneously promoted the collinear orientation at the
nanoscale and interconnection of nanotubes at the microscale
along the nanowires, similarly to Figure 5e-f. These features were
found by TEM regardless of the direction in which the material
was analyzed (e.g. transversal vs. longitudinal internal sections in
Figure 6c and 6d, respectively).
This trend confirms the creation of interconnected nanotube
networks at the microscale throughout the bulk and opens an
opportunity to modulate not only v but also surface conductivity
(s) in the materials. On the contrary, blank samples which only
included nanotubes displayed random CNT bundles in the bulk
(Figure 6a). The trend observed by TEM at the nanoscale and
microscale inside the materials was consistent with the surface
features respectively observed by scanning electron microscopy
(SEM) on the native films without (Figure 6e) or with BB (Figure
6f), thus confirming this behavior at the mesoscale too.

Figure 6. Structural analysis of polymer films at different compositions. Panels
(a-d) are TEM micrographs from internal sections in the PC films. a) Transversal
cut from a blank film with CNTs at 3 % wt; b) transversal section from a film with
2; c) transversal cut view from a sample containing 2 and CNTs 3 % wt; d)
longitudinal cut from a specimen with 2 and 1 % wt nanotubes. Panels (e-f) are
SEM micrographs from the surface of selected films. e) Blank film with
nanotubes at 3 % wt; f) film containing BB 2 and CNTs at 3 % wt.

Macroscale functional properties in films.
The macroscale properties of the multicomponent materials were
determined by variation of film compositions using a wide range
of nanotube concentration (1  10-4 % wt to 5 % wt). Preliminarily,
analysis of specimens at the latter compositions demonstrated
the possibility to modulate the electrical properties. Film samples
were electrically tested to define the trend on v and s as a
function of the nanotube concentration in self-assembled and
non-self-assembled scenarios. Assessment of v (Figure 7a)
demonstrated that the hierarchical self-assembly process induced
by either 1 or 2 at the nano, micro and mesoscale, promoted close
to a two-fold decrease in the nanotube concentration at which
percolation begins in control experiments (pc, 0.079 % wt and
0.096 % wt for 1 and 2 versus 0.169 % wt for the blank,
respectively).

FULL PAPER

Figure 7. Electrical and optical properties of films. Panels (a-b) show the effect
of CNT concentration (from 10-4 % wt to 5 % wt) and BBs 1 or 2 on v (a) and
s (b) for a selected d = 100 m; pc are indicated at intersection of trend lines
and plotted as bar chart insets; insets in dashed squares are magnifications
from the indicated areas; error bars for (a-b) are 1 SD for N = 3. c) Combined
effect from CNT concentration (0.5 % wt vs. 5 % wt) and film thickness (d = 100
m vs. d = 10 m) for selected BB 1 compared to ITO with d = 130 nm and a
standard blank PC sample with d = 100 m; in (c), ITO and film specimen with
d = 10 m are deposited on a d = 50 m PET film.

On the other hand, the same outcome was observed evaluating
s (Figure 7b), where the BBs triggered an 8-fold reduction in the
percolation threshold concentration as opposed to non-selfassembled equivalents (respectively, pc 0.174 % wt and 0.198 %
wt for 2 and 1 versus 1.363 % wt for the blank).
In addition, self-assembled systems prepared above the
percolation threshold concentration raised the v values between
1–to–3 orders of magnitude as compared to blank counterparts,
where the conductivity values reached up to 10-5 S·cm-1 to 10-1
S·cm-1 between 1 % wt and 5 % wt, respectively. In contrast, self-

assembled systems afforded significant improvement in s over
their respective blanks above the percolation threshold. The selfassembly process enhanced s of about 6–7 orders of magnitude
for CNT concentrations between 1 % wt and 3 % wt up to around
10-6 S to 10-5 S for loadings beyond 2 % wt while higher nanotube
contents inhibited this effect.
Such a difference in the trend observed for s compared to v
is evidence that not only the self-assembly mechanism provides
macroscale conductivity but it also means that anisotropic
percolation is reached at the microscale and mesoscale. This
anisotropic behavior might be a result of a higher density of the
number of conductive paths and more extensive lengths along the
surface of the prepared films at the mesoscale in contrast with a
lower density of the number of continuous interconnections from
one face of the sample to the opposite one in the microscale. The
latter is also coherent with both the surface features detected by
SEM and the internal structures observed by TEM from samples
prepared at concentrations beyond the percolation threshold.
Visible and NIR are appealing regions for potential
applications in energy conversion systems such as solar cells. In
this sense, the NIR range is still underexploited in commercial
devices. For comparison purposes, in the form of a thin film (ca.
102 nm), standard transparent conductors such as ITO present an
acceptable balance between electrical conductivity ( ca. 10-2 to
10-1 S) and optical transmittance mostly in the visible range (%T
around 80 %), whereas it is opaque in the NIR region.[11,22]
Therefore, the assessment of the optical transmittance of the
previous films across the entire UV-VIS-NIR range would
represent an interesting starting point to search for potential
substitutes for ITO in the future. Both BBs 1 and 2 showed a
negligible effect on the optical properties in the VIS-NIR as their
signals remained irrelevant in comparison with those from the
polymer matrix (Figure S3). Moreover, the nanotubes did not
contribute with additional absorption peaks between 500 nm and
3200 nm, yielding flat spectral regions, but average transmittance
(%T) was reduced down to ca. 60% for CNT loadings above 0.5 %
wt.
However, modulation of the film thickness d should help to
control transparency according to the Lambert-Beer law, as the
thicker the film the lower the transparency at a selected
wavelength at equivalent compositions. Therefore, nominal d was
evaluated at two levels (100 m and 10 m) whereas the
nanotube content was incorporated at 0.5 % wt and 5 % wt
(Figure 7c) using 1 as a model system. The results showed that
thinner films with a higher nanotube loading afforded higher
transparencies from 400 nm to 2700 nm when compared to
thicker films with a lower nanotube concentration. This confirmed
that it is possible to control transparency in self-assembled
systems by modulating the film thickness. Interestingly, both selfassembled systems operated much more competitively at  >
2100 nm than ITO films with d = 130 nm, while films with CNTs at
5 % wt and d = 10 m showed comparable transparency levels to
ITO in the range 17002100 nm and acceptable transparency
values in the range 4001700 nm. Therefore, the transparency
was measured at a standard  = 550 nm (%T550) for samples with
different nanotube loadings at the two nominal d values for both
BBs (Figure S4).

FULL PAPER
In both d versions, film transparency dropped exponentially as
the nanotube concentration was stepwise increased from 10 -4 %
wt to 5 % wt. The samples with d = 10 m afforded the lowest
impact on the transparency across the whole range, making it
possible to achieve very competitive %T550 values between
70100 % for the whole loading range. On the other hand, the
thicker films were only competitive when the CNT loading is kept
below 1 % wt. Interestingly, the self-assembly process produced
a weak decrease on %T550 for nanotube loading above 0.2 % wt
in comparison with blanks. This effect may be attributed to three
factors. First, compared to the blanks, a reduction in CNT
agglomerates by self-assembly should homogenize the number
of nanotubes per unit area and decrease the void zones that are
exposed to light in these tests. Second, when 1 or 2 are present,
they should promote an increase in the number of interconnected
nanotubes in the material thus establishing tridimensional
networks that at larger nanotube loading imply thicker selfassembled structures (cf., thinner self-assemblies at 1 % wt in
Figure 6d versus thicker interconnections at 3 % wt in Figure 6c).
Third, the combined effect of the previous two factors should
result into smaller relative standard deviation in average %T 550
when measured in the self-assembled films above the percolation
threshold. The three previous suppositions were consistent with
the combined DLS, TEM and SEM observations as well as
with %T550 for the whole nanotube loading range.
Finally, one of the most remarkable contributions from our
study is that a percolation network is reached through hierarchical
self-assembly at the nanoscale, microscale and mesoscale which
enables the regulation of macroscale properties such as electrical
conductivity at very competitive values in semiconducting and
conducting regimes (10-5 S·cm-1 to 10-1 S·cm-1). The preliminary
results on promising optical transparency in the visible range and
superior optical properties in the NIR region using MWCNTs for
the first time represent an appealing starting point for future
developments using scalable multistage auto-organization.

Experimental Section
Synthesis and characterization of chemical compounds: Synthesis of
all compounds and their characterization by elemental analysis, Matrixassisted laser desorption/ionization-Time of flight (MALDI-TOF) and High
Resolution Mass Spectrometry, 1H and 13C Nuclear Magnetic Resonance
(1H NMR and 13C NMR), and 13C{1H} Distortionless Enhancement by
Polarization Transfer (DEPT) are described below. All starting materials
were purchased from commercial sources and used without further
purification. Elemental analyses were performed at the Unidad de Análisis
Elemental of the University of Santiago de Compostela (Spain). All NMR
measurements were carried out on a Bruker-400 MHz spectrometer at
ambient temperature unless stated otherwise, and chemical shifts are
given in parts per million versus tetramethylsylane (TMS). Mass
spectrometric data were obtained from the Research Support Unit of the
ICIQ and MALDI-TOF experiments were carried out with pyrene as matrix.
Compounds (E)-2-(((2-aminophenyl)imino)(phenyl)methyl)phenol,[23] 3,3´diformyl-2,2´-dihydroxy-1,1´-biphenyl,[24]
2-hydroxy-[1,1'-biphenyl]-3carbaldehyde,[25] 3,3´-bis((E)-((2-((E)-((2-hydroxyphenyl)(phenyl)methylene)amino)phe-nyl)imino)methyl)-[1,1´-biphenyl]-2,2´-diol (ligand portion
of 1), and complex 1[26] were prepared using previously reported
methodologies.
(S)-3,3'-bis((E)-((2-((E)-((2-hydroxyphenyl)(phenyl)methylene)amino)
phenyl)imino) methyl)-[1,1'-binaphthalene]-2,2'-diol (ligand portion of
2): (E)-2-(((2-aminophenyl)imino)(phenyl)methyl)phenol (200 mg, 0.7
mmol) in 2 mL of CHCl3 and (S)-2,2'-dihydroxy-[1,1'-binaphthalene]-3,3'dicarbaldehyde (119 mg, 0.35 mmol) in 5 mL of MeOH were refluxed for
24 h. The precipitate formed was then filtered off. Yellow solid, 300 mg,
97%. [α]D25 = −36.37 deg cm3 g−1 dm−1 (c = 0.84 g cm−3, CH2Cl2). 1H NMR
(400 MHz, CDCl3): δ = 6.28 (22, JHH = 7.8, 1.4 Hz, 1H), 6.42 (td, JHH = 7.7,
1.3 Hz, 1H), 6.55 (m, 2H), 6.69 (dd, JHH = 7.9, 1.2 Hz, 1H), 6.76 (m, 1H),
6.84-7.39 (m, 26H), 7.84 (d, JHH = 8 Hz, 2H), 7.95 (s, 2H), 8.51 (s, 2H),
12.36 (s, 2H), 14.06 (s, 2H). 13C NMR (125 MHz, CDCl3): δ = 117.88,
118.00, 119.50, 120.17, 121.44, 121.57, 123.56, 123.64, 124.88, 125.03,
125.67, 126.99, 127.75, 128.00, 128.79, 129.02, 129.04, 132.31, 133.29,
133.38, 133.50, 134.59, 135.03, 135.09, 135.79, 140.54, 154.44, 154.95,
162.61, 163.92. 1H NMR spectrum, and 13C{1H} (DEPT) spectra can be
found in Figure S5 and Figure S6, respectively. High-resolution mass
spectrometry (TOF ES+) m/z 883.3326 ([M+H]) C60H43N4O4 requires
883.3284.

Conclusions
In conclusion, this work has demonstrated that the bioinspired
nanoscale self-assembly of supramolecular BBs derived from
simple and modular Schiff base compounds can be translated to
a microscopic control of hierarchically auto-organized and
collinearly interlinked MWCNTs to yield mesoscale percolation
networks in a polymer matrix. Consequently, those networks
converted the polymer system into a functional multicomponent
material with tunable, anisotropic electrical conductivity
throughout semiconducting and conducting regimes at the
macroscale. The mechanism of auto-organization is driven
through energy-dissipative processes and multilevel assembly
steps. Moreover, the optical performance observed through the
visible and NIR regions opens new opportunities beyond the
current spectral restrictions observed for commercial transparent
conductors, thus paving the way towards a new approach for the
generation of conductive materials under scalable and attractive
conditions.

Tetrakis-Schiff base complex 2: To a solution of ligand (S)-3,3'-bis((E)((2-((E)-((2-hydroxyphenyl)(phenyl)methylene)amino)phenyl)imino)methyl)-[1,1'-binaphthalene]-2,2'-diol (100 mg, 0.13 mmol) in 6 mL of
anhydrous tetrahydrofuran (THF), ZnEt2 (0.29 mL 1 M in hexanes, 0.29
mmol) were slowly added. The reaction was stirred at room temperature
18 h. The red precipitate was filtered off. Red solid, 127 mg, 97%. 1H NMR
(400 MHz, DMSO-d6): δ = 6.16 (m, 1H), 6.44 (m, 3H), 6.77-7.38 (m, 26H),
7.70 (d, JHH = 8.1 Hz, 2H), 7.76 (d, JHH = 7.5 Hz, 2H), 8.22 (s, 2H), 9.26 (s,
2H). 13C NMR (125 MHz, DMSO-d6): δ = 112.15, 117.18, 119.86, 120.58,
123.14, 123.25, 123.53, 123.96, 124.42, 125.08, 125.71, 126.27, 127.35,
128.13, 128.21, 128.33, 128.50, 128.92, 129.00, 132.99, 134.17, 137.03,
137.75, 137.85, 139.65, 140.32, 164.28, 165.75, 172.19, 173.61. 1H NMR
spectrum, and 13C{1H} (DEPT) spectra can be found in Figure S7 and
Figure S8, respectively. Mass spectrometry (MALDI-TOF) m/z 1009.1
([M+H]). Elemental analysis calcd. (%) for C60H38N4O4Zn2·4H2O C 66.62,
H 4.29, N 5.18 found C 66.58, H 4.00, N 5.94.
Transmission electron microscopy: TEM studies were carried out on a
JEOL model JEM-1011 electron microscope with an acceleration voltage
of 100 keV, 50 m C2 aperture, spot size 3 and variable dose rate.
Analyzed liquid samples were prepared by dropcasting 70 M solutions of
the corresponding BB 1 or 2 and/or 20 g/mL of the MWCNT (DCM 99.8 %

FULL PAPER
purity as solvent) on formvar carbon film-covered square mesh copper
grids and dried completely for 4 h. Deviations from the latter concentrations
are mentioned in the text. Solid film samples were embedded in an epoxybased resin, cured for 72 h at 80 °C, cut with a microtome and fixed on a
copper double grid fixation system for analysis.
Scanning electron microscopy: Microscopic analysis was carried out
using a JEOL microscope, model JSM-6400 at ambient temperature, work
distance of 8 mm, 10-4 Torr pressure, beam voltage 10 kV and variable
probe current. Solid samples were fixed on standard sample holders
without any previous treatment and directly analyzed at the microscope.
Dynamic light scattering: DLS analysis was performed using a Zetasizer
Nano ZS (Malvern Instruments Ltd.) with 532 nm laser radiation source.
Analytical procedure was carried out in all the cases with 1 mL of the
corresponding solution on a standard fluorescence glass cuvette with 1 cm
optical path.
Preparative solutions: Polymer films were obtained from preparative
DCM solutions containing 10 % m/v high purity optical storage media
grade bis-phenol-A polycarbonate provided by Bayer, the required BB 1 or
2 (70 M) and/or the multiwalled carbon nanotubes (>95 % purity, 100–
700 nm in length, 15–42 nm in external diameter, purchased at Nanocyl
SA, Sambreville, Belgium). The amount of nanotubes:polymer ratio ranged
between 10-4:99.9999 to 5:95 in mass, where the systems which also
contained one of the BBs alternatively included the last one in a ratio
BB:polymer+nanotubes equal to 0.065:99.935 in mass. The solutions were
prepared using anhydrous dichloromethane (99.8%) purified, filtered and
dried through a high-throughput solvent purification system.
Preparation of films: Films were prepared under high-efficiency
particulate air-filtered laminar flow conditions following two different
casting-from-solution strategies, depending on the nominal d value
required. Specimens with d = 100 m were prepared by casting
preparative solutions at the required nanotube:polymer ratio on a closed
injection spin-coating system. An aliquot of the solution (commonly 1 mL
to 3 mL, depending on the film thickness desired) was deposited on a
circular flat aluminum mold (6 cm diameter) fixed to the rotation system.
The solution was then dispersed by spinning at 60 rpm and dried at 25 °C
and 1 Atm. The films formed were dried at low vacuum (10-100 mBar),
25 °C for 1-2 h and separated from the mold. Calibration of average d was
carried out using an Extramess 2000 inductive digital comparator (Mahr,
Germany) with 0.1 m resolution and confirmed by SEM. Conversely, films
with d = 10 m were prepared by casting preparative solutions on 50 m
PET substrates and deposited over the last one using the doctor blade
technique. An aliquot of the solution (commonly 1-5 mL, depending on the
surface to cover and the film thickness desired) was deposited at the
application region on the substrate and subsequently spread at 2.5 mm/s
to 30 mm/s at 25 °C using a calibrated steel blade with sub-micrometer
resolution adjustable height. Calibration of average d was carried out
adjusting the blade height and confirmed with the inductive digital
comparator. The formed films were dried at low vacuum (10-100 mbar)
and 25 °C for 1-2 h and directly used after preparation. ITO films with d =
130 nm (10-2 S) were provided by Sigma-Aldrich Chemie GmbH
(Steinheim, Germany) supported on a 50 m thickness PET sheet; prior to
use, ITO was subsequently cleaned with ethanol, dried at vacuum (10-100
mbar) and 25 °C for 1-2 h and directly analyzed afterwards.
Electrical characterization of materials: Surface and volumetric
electrical conductivity of solid films at conductivity values below 10-6 S (net
readout) was determined with a Keithley Model 8009 high-resistivity test
fixture with integrated Faraday box electrostatic shielding system coupled
to a Keithley 6487 Picoammeter/Voltage source. Film samples with the
dimension 6 cm  6 cm were directly analyzed within the fixture at applied

voltages of 100 V and 500 V without any previous treatment. v and s of
solid films at net conductivity readout above 10-6 S was determined
following the two- and four-point probe method, using a Keithley High
Performance Digital Multimeter model 2000 under automatic sweeping
mode. Samples with the same abovementioned dimensions were pretreated at the measuring contact points with 4 mm width/15 mm spacing
thin layer strips of Electrodag 1415 silver ink supplied by Agar Scientific.
Optical characterization of films: UV-VIS-NIR measurements were
carried out on a Lamba 1050 PerkinElmer spectrophotometer equipped
with PMT, InGaAs and PbS detectors, double beam optics, double
monochromator and D2 and W light sources. %T550 optical transmission
measurements were recorded on a Shimadzu UV1800 Spectrophotometer.

Acknowledgements
The Spanish Ministerio de Economía y Competitividad (CTQ2014–60419-R and the Severo Ochoa Excellence Accreditation
2014–2018, project SEV-2013–0319), the European Community
Seventh Framework Program FP7-PEOPLE-ITN-2008 (grant
number 238363), the DGAPA-PAPIIT program from National
Autonomous University of Mexico (grant numbers IA205616 and
IA205316), the National Council for Science and Technology from
Mexico (CONACYT, grant number 251533), CERCA/Generalitat
de Catalunya and ICREA are all acknowledged for support.
Keywords: dynamic self-assembly • hierarchical autoorganization • carbon nanotubes • polymer composites •
conducting materials
[1]

[2]

[3]

[4]
[5]
[6]

[7]
[8]
[9]

a) T. P. Bigioni, X.-M. Lin, T. T. Nguyen, E. I. Corwin, T. A. Witten, H. M.
Jaeger, Nat. Mater. 2006, 5, 265-270; b) M. Qian, A.C. Reber, A. Ugrinov,
N. K. Chaki, S. Mandal, H. M. Saavedra, S. N. Khanna, S. Sen, P. S.
Weiss, ACS Nano 2010, 4, 235-240.
a) A. Dong, J. Chen, P. M. Vora, J. M. Kikkawa, C. B. Murray, Nature
2010, 466, 474-477; b) B. G. Sumpter, L. Liang, A. Nicolaï, V. Meunier,
Acc. Chem. Res. 2014, 47, 3395-3405; c) E. Y. S. Oliveira, R. Bode, M.
V. Escárcega-Bobadilla, G. A. Zelada-Guillén, G. Maier, G. New J. Chem.
2016, 40, 4625-4634.
a) X. Liu, Y.-Z. Long, L. Liao, X. Duan, Z. Fan, ACS Nano 2012, 6, 18881900; b) S. E. Mastroianni, T. H. Epps, Langmuir 2013, 29, 3864-3878;
c) A. R. Chandrasekaran, H. Wady, H. K. K. Subramanian, Small 2016,
12, 2689-2700.
K. J. M. Bishop, Angew. Chem. Int. Ed. 2016, 55, 1598-1600.
A. H. Gröschel, A. Walther, T. I. Löbling, F. H. Schacher, H. Schmalz, A.
H. E. Müller, Nature 2013, 503, 247-251.
a) M. Brutschy, M. W. Schneider, M. Mastalerz, S. R. Waldvogel, Adv.
Mater. 2012, 24, 6049-6052; b) M. Giese, L. K. Blusch, M. K. Khan, W.
Y. Hamad, M. J. MacLachlan, Angew. Chem. Int. Ed. 2014, 53, 88808884.
Y. Tian, X. Yan, M. L. Saha, Z. Niu, P. J. Stang, J. Am. Chem. Soc. 2016,
138, 12033-12036.
a) C. F. J. Faul, Acc. Chem. Res. 2014, 47, 3428-3438; b) C. Zhang, P.
Chen, H. Dong, Y. Zhen, M. Liu, W. Hu, Adv. Mater. 2015, 27, 5379-5387.
a) H. Ma, J. Hao, Chem. Soc. Rev. 2011, 40, 5457-5471; b) S. A.
Majetich, T. Wen, R. A. Booth, ACS Nano 2011, 5, 6081-6084; c) S. Ilday,
F. Ö. Ilday, R. Hübner, T. J. Prosa, I. Martin, G. Nogay, I. Kabacelik, Z.
Mics, M. Bonn, D. Turchinovich, H. Toffoli, D. Toffoli, D. Friedrich, B.
Schmidt, K.-H. Heinig, R. Turan, Nano Lett. 2016, 16, 1942-1948.

FULL PAPER
[10]
[11]
[12]

[13]

[14]

[15]

[16]
[17]

P. P. Edwards, A. Porch, M. O. Jones, D. V. Morgan, R. M. Perks, Dalton
Trans. 2004, 2995-3002.
D. Angmo, F. C. Krebs, J. Appl. Polym. Sci. 2013, 129, 1-14.
a) A. Kumar, C. Zhou, ACS Nano 2010, 4, 11-14; b) K. Sakakibara, J. P.
Hill, K. Ariga, Small 2011, 7, 1288-1308; c) K. Ellmer, Nat. Photonics
2012, 6, 809-817; d) C. J. M. Emmott, A. Urbina, J. Nelson, Sol. Energy
Mater. Sol. Cells 2012, 97, 14-21; e) D. Jariwala, V. K. Sangwang, L. J.
Lauhon, T. J. Marks, M. C. Hersam, Chem. Soc. Rev. 2013, 42, 28242860.
a) J. K. Wassei, R. B. Kaner, Mater. Today 2010, 13, 52-59; b) M. F. L.
De Volder, S. H. Tawfick, R. H. Baughman, A. J. Hart, Science 2013, 339,
535-539.
a) S. Roth, H. J. Park, Chem. Soc. Rev. 2010, 39, 2477-2483; b) K. Jiang,
J. Wang, Q. Li, L. Liu, C. Li, S. Fan, Adv. Mater. 2011, 23, 1154-1161; c)
D. S. Hecht, L. Hu, G. Irvin, Adv. Mater. 2011, 23, 1482-1513; d) X. Sun,
T. Chen, H. Peng, Acc. Chem. Res. 2013, 46, 539-549; e) J. Du, S. Pei,
L. Ma, H.-M. Cheng, Adv. Mater. 2014, 26, 1958-1991; f) M. Layani, A.
Kamyshny, S. Magdassi, Nanoscale 2014, 6, 5581-5591; g) J. Li, G. P.
Pandey, Annu. Rev. Phys. Chem. 2015, 66, 331-356;.
S. H. Kim, C. S. Haines, N. Li, K. J. Kim, T. J. Mun, C. Choi, J. Di, Y. J.
Oh, J. P. Oviedo, J. Bykova, S. Fang, N. Jiang, Z. Liu, R. Wang, P. Kumar,
R. Qiao, S. Priya, K. Cho, M. Kim, M. S. Lucas, L. F. Drummy, B.
Maruyama, D. Y. Lee, X. Lepró, E. Gao, D. Albarq, R. Ovalle-Robles, S.
J. Kim, R. H. Baughman, Science 2017, 357, 773-778.
L. Hu, D. S. Hecht, G. Grüner, Nano Lett. 2004, 4, 2513-2517.
a) P. N. Nirmalraj, P. E. Lyons, S. De, J. N. Coleman, J. J. Boland, Nano
Lett. 2009, 9, 3890-3895; b) M. Y. Timmermans, D. Estrada, A. G.

[18]
[19]

[20]

[21]
[22]
[23]
[24]
[25]
[26]

Nasibulin, J. D. Wood, A. Behnan, D. Sun, Y. Ohno, J. W. Lyding, E. P.
Hassanien, E. Kauppinen, Nano Res. 2012, 5, 307-319.
Z. Yang, M. Liu, C. Zhang, W. W. Tjiu, T. Liu, H. Peng, T. Liu, Angew.
Chem. Int. Ed. 2013, 52, 3996-3999.
M. V. Escárcega-Bobadilla, G. A. Zelada-Guillén, S. V. Pyrlin, M.
Wegrzyn, M. M. D. Ramos, E. Giménez, A. Stewart, G. Maier, A. W. Kleij,
Nat. Commun. 2013, 4, 2648.
a) S. Akine, M. Miyashita, T. Nabeshima, J. Am. Chem. Soc. 2017, 139,
4631-4634; b) P. D. Frischmann, S. Guieu, R. Tabeshi, M. J. MacLachlan,
J. Am. Chem. Soc. 2010, 132, 7668-7675; c) S. J. Wezenberg, G.
Salassa, E. C. Escudero-Adán, J. Benet-Buchholz, A. W. Kleij, Angew.
Chem. Int. Ed. 2011, 50, 713-716; d) G. Salassa, M. J. J. Coenen, S. J.
Wezenberg, B. L. M. Hendriksen, S. Speller, J. A. A. W. Elemans, A. W.
Kleij, J. Am. Chem. Soc. 2012, 134, 7186-7192; e) G. Consiglio, I. Pietro
Oliveri, S. Failla,S. Di Bella, Inorg. Chem. 2016, 55, 10320-10328; f) G.
Consiglio, S. Failla, P. Finocchiaro, I. Pietro Oliveri, S. Di Bella, Inorg.
Chem. 2012, 51, 8409-8418; g) M. Mastalerz, H.-J. S. Hauswald, R. Stoll,
Chem. Commun. 2012, 48, 130-132.
J. K. Sprafke, S. D. Stranks, J. H. Warner, R. J. Nicholas, H. L. Anderson,
Angew. Chem. Int. Ed. 2011, 50, 2313-2316.
R. Gordon, MRS Bulletin 2000, 25, 52-57.
A. Kämpfe, E. Kroke, J. Wagler, Eur. J. Inorg. Chem. 2009, 1027-1035.
H.-C. Zhang, W.-S. Huang, L. Pu, J. Org. Chem. 2001, 66, 481-487.
C. Görl, H. G. Alt, J. Organomet. Chem. 2007, 692, 5727-5753.
M. V. Escárcega-Bobadilla, G. Salassa, M. Martínez Belmonte, E. C.
Escudero-Adán A. W. Kleij, Chem.-Eur. J. 2012, 18, 6805-6810.

FULL PAPER
Entry for the Table of Contents

FULL PAPER
Gustavo A. Zelada-Guillén,* Martha V.
Escárcega-Bobadilla,* Marcin Wegrzyn
Enrique Giménez, Gerhard Maier and
Arjan W. Kleij*

Page No. – Page No.

Putting things to order: Nanoscale anisotropic supramolecular interactions
between carbon nanotubes and molecular building blocks based on bimetallic
Zn(salen) complexes allow for hierarchical self-assembly control at the microscale to
produce mesoscale electrical percolation networks, which enhances macroscale
conductivity in polymer composites up to six orders of magnitude.

Enhanced Conductivity for Carbon
Nanotube based Materials through
Supramolecular Hierarchical SelfAssembly