FEATURE ARTICLE

www.rsc.org/materials | Journal of Materials Chemistry

Self-organization of nanostructured functional dendrimers{

Joaqun Barbera,a Bertrand Donnio,b Lionel Gehringer,b Daniel Guillon,*b Mercedes Marcos,a Ana Omenata and Jose Luis Serrano*a
Received 17th February 2005, Accepted 23rd May 2005 First published as an Advance Article on the web 16th June 2005 DOI: 10.1039/b502464a In this article, we review some of our recent work on the development and study of the properties of self-organizing supermolecular liquid crystalline dendritic materials. Prior to this purpose, the state of the art in the field of molecular design, structure and properties of liquid-crystalline dendrimers (LCDs) will be briefly reviewed, and illustrated by a selection of pioneering examples. The first type of LC dendrimer takes into account the location of their functional elements (mesogens) in the periphery of the macromolecules and its topology of attachment, and the nature of the mesogen. We will show how functionality can be in-built into such materials so that self-organising functional systems can be created. Side-chain liquid-crystalline dendrimers exhibiting lamellar, columnar and nematic phases are shown. The mechanism of mesophase formation (mesogen interaction versus microphase separation) is also discussed in relation to the molecular structure (the nature of the mesogen and of the dendritic matrix, dendritic generation number). The second type of LC dendrimer introduced presents a different topology in that the mesogens are now inserted within the dendritic matrix and in the periphery. Owing to their particular constitution, it will be shown that the mesophases of the lamellar and columnar types possess unusual morphologies.

I. Introduction
Two different approaches are usually considered in the elaboration of nanomaterials.1 The first one is the top-down approach which has been developed these last decades and which consists of pushing the limit of the lithography techniques

Qumica Organica. Instituto de Ciencia de Materiales de Aragon. Universidad de Zaragoza C.S.I.C., 50009 Zaragoza, Spain. E-mail: joseluis@posta.unizar.es b Institut de Physique et Chimie des Materiaux de Strasbourg. Groupe des Materiaux Organiques. 23, rue du Loess, F-67037 Strasbourg, France. E-mail: daniel.guillon@ipcms.u-strasbg.fr { This work was funded as part of the EU Research Training Network LCDD, Supermolecular Liquid Crystal Dendrimers.

down to the nanometer scale. However, there are intrinsic limitations to going beyond 100 nm with such techniques. The second, more recent one is the bottom-up approach which consists of building up nanostructures and assembling them from individual atoms, molecules or macromolecules.2 With this in mind, dendrimers have been considered, during the last decade, as promising materials for the fabrication and assembly of nanostructures. More specifically, dendrimers functionalised with mesogenic groups are of great interest for the generation of self-organized nanostructures a la carte. ` Since their discovery in the late 1970s,3 and the adjustments of perfectly controlled iterative synthetic processes, the chemistry of dendrimers has led to the most impressive developments and rapidly expanding areas of current

Joaqun Barbera

Dr Joaqun Barbera received his PhD in 1986 under the supervision of Profs. E. Melendez and J. L. Serrano. After working with Dr A. M. Levelut in the Physics Laboratory of Solids (CNRS-University Paris Sud), he obtained a permanent position in the University of Zaragoza, where he teaches Organic Chemistry. His scientific activities have been carried out in the Liquid Crystal Group, and his current research is mainly devoted to the structural studies of liquid crystals.

Dr Bertrand Donnio was born in 1967 (Bretagne) and graduated in chemistry from the University of Rennes in 1991.
This journal is The Royal Society of Chemistry 2005

He obtained his PhD from Sheffield University (UK) in 1996 under the supervision of Prof. D. W. Bruce. After two postdoctoral fellowships (in Neuchatel, Switzerland, Prof. R. Deschenaux, and in Freiburg, Germany, Prof. H. Finkelmann), he obtained a permanent position as Charge de Recherche (CNRS) at the IPCMS-GMO. He is interested in the elaboration of liquid crystalline materials (design and synthesis of dendrimers, Bertrand Donnio elastomers, metallomesogens, conjugated organic mesogens), and in the study of the supramolecular organisation within LC mesophases (small angle X-ray diffraction, dilatometry).
J. Mater. Chem., 2005, 15, 40934105 | 4093

chemistry.4 Dendrons and dendrimers represent an original class of aesthetic, practically monodisperse supermolecules possessing a regular and controlled branched architecture. This enthusiasm is primary caused by the intrinsic and unique molecular features of the dendrimers4 and the possibility of generating numerous and perfectly characterized original chemical structures, offering new synthetic concepts and challenges for chemists5 as well as raising several interesting theoretical problems.6 This field of research has been further boosted by the appreciation of their uses as potentially interesting candidates in widespread applications since dendrimers are able to concentrate an important number of functional groups in a small volume.7 Interests may range when suitably functionalized from biology, as drug or gene delivery devices due to their rough resemblance and similar dimensions to some living components,8 to molecular functional materials with dedicated physical properties.9 Several excellent comprehensive review articles have emphasized the rich chemistry of dendrimers including synthetic

Dr Lionel Gehringer graduated in chemistry from the University of Strasbourg in 2000. He completed a M.Sc. on the synthesis of organic active NLO chromophores in the IPCMS-GMO under the supervision of Prof. J. F. Nicoud. He then undertook a PhD thesis on the synthesis and the study of the physicochemical properties of mainchain liquid crystalline dendrimers under the supervision of Drs Daniel Guillon Lionel Gehringer and Bertrand Donnio. He is now working as a postdoctoral fellow in the University of Mainz (Germany) under the supervision of Prof. H. Frey on the modification of hyperbranched polymers.

methodologies,5 synthesis of chiral systems,10 metal-11 and heteroatom-based12 dendrimers. Some aspects of the supramolecular chemistry and self-assembling ability of elaborated dendrimers into precise and pre-determined fashions have also been discussed.13 As mentioned above, the most important features of these macromolecules are related to their unique structure, which combines a highly branched, perfectly controlled topology with a geometrical rate of growth as the generation number increases and a narrow molecular weight distribution. This is the result of sophisticated genealogical directed syntheses (convergent or divergent directed growth),6 consisting of controlled iterative methods involving the repetition of successive and specific elementary steps. An important consequence of such sequential constructions is the total degree of control of the ultimate macromolecular architecture that is the molecular size and shape, which can be modulated by the generation growth, the multiplicity of the branches and the connectivity of the focal core. Dendrimers may represent an interesting alternative towards the development of materials in which molecular level information is transferred from the initiator core to the periphery (or vice versa) with the expectation of complementary and synergic phenomena, i.e. induction of new properties, and/or cooperative effects, i.e. amplification of the existing properties. It was thus of interest to functionalize such supermolecules with mesogenic elements in order to obtain a new class of liquid crystalline materials.14,15 Indeed, one interesting issue is the possibility to discover new types of mesophases with original morphologies as for liquid crystals with nonconventional molecular architectures.16 Molecular engineering of liquid crystals is also an important issue for controlling the self-organizing process and the ultimate mesomorphic properties. Since dendrimers are in principle monodisperse, the measurements of the physical properties are thus completely reproducible. Finally, an attractive approach in the field of materials science would be to use the ability of such polyfunctional and discrete dendritic structures to self-assemble into mesophases and their subsequent processing

Dr Daniel Guillon received his PhD degree in Chemical Physics in 1976 from the Louis Pasteur University of Strasbourg (France). He is director of research at CNRS, presently deputy director of the Institute of Physics and Chemistry of Materials of Strasbourg (IPCMS) since 2002 and head of the organic materials group of the institute since 1989. His main interests include the description at the molecular level of the Daniel Guillon mesomorphic structures exhibited by low molecular weight and polymer materials, the study of self-organisation processes in supramolecular chemistry, the bottom-up approach of organic and hybrid nanomaterials, and the structure of thin organic films. Sine 2004, he has
4094 | J. Mater. Chem., 2005, 15, 40934105

also been Vice-President of the International Liquid Crystal Society. Dr Mercedes Marcos is currently Investigador Cient fico of Instituto de Ciencia de Materiales de Aragon (CSIC-Universidad de Zaragoza). She obtained her PhD from the University of Zaragoza in 1983, under the guidance of Prof. E. Melendez and Prof. J. L. Serrano. Her research interests are focused Mercedes Marcos on liquid crystals (ferroelectric liquid crystals, low and high molecular weight paramagnetic metallomesogens) and currently dendrimer and pre-megamer liquid crystals. She has published over 90 scientific articles and is co-author of a book on metallomesogens.
This journal is The Royal Society of Chemistry 2005

consequence of their statistical mode of construction.22 Two approaches have also been successfully applied to generate mesomorphism in these systems. One is the random polymerization of mesogenic monomers containing at least two different functional groups A and B (ABm-type monomers with m 2 and A ? B) to yield main-chain systems,23,24 whereas the other one consists of the grafting of mesogenic groups at the periphery of pre-existing hyperbranched polymers, i.e. side-chain systems.25

Fig. 1 Schematic 2D representation of an end-group dendrimer of the second generation with a 4-fold core connectivity (NC 5 4) and a ternary branch multiplicity (NB 5 3). The mesogen can be attached terminally or laterally to yield (a) end-on and (b) side-on LCDs.

II. Control of molecular conformation in side-chain LC dendrimers

The term side-chain liquid crystal dendrimers is assigned, in analogy with the polymer terminology, to those materials which are built by the functionalization of the periphery of a pre-formed dendrimer (poly(amidoamine) PAMAM, poly(propyleneimine) PPI, carbosilane, poly(siloxane), carbosilazane, etc.) with units that promote the formation of supramolecular organisations giving rise to liquid crystal mesophases15,2628 (Fig. 1). For instance, the introduction of rod-like or disk-like units at the periphery of the original dendrimer leads to dendritic architectures that display liquid crystalline properties. These LC dendrimers combine in one single macromolecule two main components with opposite tendencies: on one hand, the central isotropic dendritic architecture, and on the other hand, the structural anisotropic (pro)mesogenic units. The branches of the dendritic core tend to be isotropically distributed in space because of entropic forces, resulting in a pseudo-spherical morphology (star-burst shape in dendrimer terminology). The (pro)mesogenic units at the periphery, in contrast, are subjected to strong anisotropic interactions. In general, enthalpy gains over entropy in these systems, with the result of a mesophase formation. In addition, the incompatible chemical natures of the dendrimer core and the (pro)mesogenic units forces the microsegregation of these two constituent parts at the molecular level, which also favours the formation of mesophases. In summary, the mesomorphic properties of these LC dendrimers (phase type, transition temperatures and thermodynamic stability) are determined by

into multicomponent nanosize objects in which each constituent element could be tailored independently. The most important class of liquid-crystalline dendrimers (LCDs) is represented by the side-chain liquid crystalline dendrimers as schematically represented in Fig. 1. The overall structure of such side-group LCDs (this terminology is used by analogy to side-chain liquid crystal polymers) consists of a flexible branched network emanating from a single multivalent initiator core and mesogenic or pro-mesogenic units attached at the termini of the branches. The control of the molecular conformation in such side-chain LC dendrimers will be considered thoroughly in section II of this paper. This field has now well expanded and one can distinguish several additional sub-classes of LCDs including main-chain systems (see section III), shape-persistent LCDs,17 supramolecular dendromesogens,18 fullerene-containing liquid crystals,19 metallo-mesogenic-dendrimers20 and the so-called mesomorphic polypedes.21 To avoid confusion in the following, let us note that the dendritic systems so far mentioned are distinct from hyperbranched polymers, often referred to as dendrimers, a misleading term, since they are characterized by randomly branched structures with a high degree of branching (without any cross-linking) and broad molecular weight distributions, a

Ana Omenat

Dr Ana Omenat obtained her PhD degree from the University of Zaragoza, and after working with Prof. Ghedini in the University of C a l a b r i a a n d in P h i l i p s Research Laboratories in Eindhoven, she joined the Liquid Crystal Group at the University of Zaragoza. Her research activities have always been related to liquid crystals: low and high molecular weight organic and metallorganic materials.

Prof. Jose Luis Serrano received his PhD in Chemistry in 1980 from the University of Zaragoza (Spain). In 1996,
This journal is The Royal Society of Chemistry 2005

he obtained the position of Full Professor of Organic Chemistry at the University of Zaragoza. Since 1985, he has belonged to the Institute of Materials Science of Arago n (ICMA), and now he is Vice-Director of the Institute of Nanoscience of Arago n (INA). During the last few years, Prof. Serranos research has been devoted to the use of liquid crystals as a tool in supramolecular chemistry, dendrimers and Jose Luis Serrano polymers in order to obtain functional materials with relevant optical and electrooptical properties.
J. Mater. Chem., 2005, 15, 40934105 | 4095

the enthalpy/entropy balance, the degree of chemical incompatibility of the constituent parts, their size and the structure of the (pro)mesogenic unit itself. Pioneering reports Among this category of LC dendrimers, Meijer et al.27 have described two series of poly(propyleneimine) dendrimers functionalised with pentyloxy and decyloxy cyanobiphenyl mesogens at the periphery (Fig. 2). X-Ray diffraction experiments have shown that all the mesophases observed are smectic A in nature. One striking feature is the almost constant value of the layer spacing whatever the generation number for a given spacer. The fact that the layer spacing does not vary as a function of the number of peripheral mesogenic units implies that the dendritic core should be in a pronounced distorted conformation, which seems to indicate an extension with increasing generation number occurring only in two dimensions, i.e. in a plane parallel to the smectic layers. Lattermann et al.28 have also considered poly(propyleneimine) dendrimers of different generations but substituted with 3,4bis(decyloxy)benzoyl groups (Fig. 2). Firstly, these compounds exhibit also mesomorphic behaviour with a melting temperature decreasing with increasing molar mass. Secondly, the observed mesophases have a columnar nature. Carbosilane LC dendrimers Shibaev et al.,15 and to a lesser extent Frey et al.,29 have reported the thermotropic behaviour of carbosilane LC

Fig. 2 Poly(propyleneimine) dendrimers with two types of terminal mesogenic groups.

dendrimers. They showed that for the lower generations there is no significant influence of the dendritic architecture. Structural studies on two series of dendrimers, containing either terminal cyanobiphenyl and methoxyphenyl benzoate groups, show the existence of disordered smectic phases (smectic A and C) up to the fourth generation despite the globular shape of the dendritic core,30 with layers of mesogenic groups alternating with layers formed by the carbosilane dendritic cores and aliphatic spacers. It is interesting to note, here also, that the layer spacing is slightly dependent on generation number but for the high generations only, in the range between 40 and 50 A. The fifth generation of the carbosilane LC dendrimers, containing 128 terminal cyano groups, forms not only a lamellar phase but also a supramolecular columnar nanostructure.31 A smectic A phase exists at low temperature, for which the molecular organisation is the same as for the previous generations with a layer spacing of 53 A despite the presence of the large number of peripheral mesogenic groups. At higher temperature, this lamellar mesophase transforms into a rectangular columnar phase, and finally into a hexagonal columnar phase. The explanation of this behaviour given by the authors is based on a change of the molecular shape. With the increase of the temperature, the carbosilane LC dendrimers become less elongated and their shape more circular, favouring the formation of ellipsoidal columns. Each column would consist of the stacking of ellipsoidal molecules, elongated in a plane perpendicular to the columnar axis. Further increase of the temperature leads to a more symmetrical shape of the LC dendrimer becoming circular and inducing the formation of a hexagonal phase. The columns would result from the stacking of the circular molecules flattened in a direction parallel to the columnar axis. In both rectangular and hexagonal phases, the surface of the columns is covered by the mesogenic groups while their inner part consists of soft dendritic cores. A similar trend has been observed with carbosilane dendrimers terminated with mesogenic groups based on anisic acid derivatives.32 More recently, Shibaev and coworkers have also considered photosensitive LC dendrimers. The interest in such compounds is due to the new opportunities provided by such systems in the production of various optical devices and the preparation of materials suitable for optical data storage. Moreover, the development of photoactive dendrimers capable of forming liquid-crystalline phases is of particular interest, because a low viscosity is expected to induce a quick response of the dendrimers to the action of an external field such as, for example, light irradiation. Thus, fast response photosensitive materials can be generated. The first example concerns a liquid crystalline carbosilane dendrimer of the first generation with azobenzene terminal groups, exhibiting a smectic A mesophase such as described above for other terminal mesogenic groups.33 In this case, the azobenzene terminal moiety serves the dual function, i.e. on one hand, its rigid anisometric shape ensures the development of a mesomorphic state, and, on the other hand, the presence of azo-chromophores ensures that the LC dendrimer is sensitive to light irradiation. It was shown that the EZ photopolymerization of the azobenzene groups proceeds both in solution and in film under UV irradiation. This process is reversible photochemically and thermally.
This journal is The Royal Society of Chemistry 2005

4096 | J. Mater. Chem., 2005, 15, 40934105

Fig. 3 (a) Schematic representation of the molecular model for dendrimers with one-terminal-chain mesogenic units. (b) Model for the smectic A supramolecular organisation.

PAMAM and PPI LC dendrimers In the last years, a systematic study of the properties of sidechain LC dendrimers has been undertaken on poly(amidoamine) (PAMAM), poly(propyleneimine) (PPI) dendimers. In all cases, it was found that the enthalpic gain of the mesogenic units arranged as in a classical liquid crystalline mesophase dominates over the entropic tendency of the dendrimer core to adopt a globular isotropic conformation. The flexibility of the dendritic core (PAMAM, PPI) allows the macromolecule to adopt a microphase separated molecular conformation which gives rise to various types of liquid crystalline supramolecular organisations. The chemical structure of the (pro)mesogenic units (shape, number of terminal alkoxy chains) determines the type of mesophase formed (nematic, lamellar, columnar). The structural study of the mesophases by X-ray diffraction techniques has led to the proposal of idealised models for both the dendrimer molecules and the supramolecular arrangements in the different mesophases. LC dendrimers obtained by functionalization of commercial amino-terminated PAMAM and PPI with mesogenic units with one terminal alkoxy chain, namely 4-alkoxybenzoyloxysalicylaldehyde, display nematic and smectic mesophases.34 In this case, the molecular model proposed consists of a cylinder in which the dendrimer core occupies the central slab and the mesogenic units are arranged parallel to each other, extending up and down from the molecule centre (Fig. 3a). This model explains the mesogenic behaviour of these PAMAM and PPI derivatives since the dendrimeric supermolecules can be considered as large rods that would be ordered parallel to each other promoting the supramolecular order typical of smectic mesophases (Fig. 3b). This model also shows that the dendritic core deforms strongly with increasing generation number, since the layer
This journal is The Royal Society of Chemistry 2005

spacing stays around 45 nm whatever the molecular weight (Fig. 4). This deformation takes place in two dimensions in a plane parallel to the smectic layers. Systematic calculations performed from X-ray diffraction and volume data indicate that the diameter of the dendrimer cylinder increases from 1.3 nm up to 6 nm when going from the lowest generation up to the fourth one containing 64 peripheral mesogenic units. Dendrimers incorporating mesogenic units with two terminal alkoxy chains (3,4-dialkoxybenzoyloxysalicylaldehyde) exhibit a hexagonal columnar mesophase.35 These dendrimers cannot be arranged in a molecular cylindrical model like that proposed for the smectic phases, since the cross-sectional area of the terminal chains is larger than the area occupied by the mesogenic units. In this case, the dendrimers tend to adopt a different and more stable conformation, so that the mesogenic units can be accommodated optimally. It consists of a disk-like radial arrangement that allows the filling of the space in all the three distinct

Fig. 4 Variation of the size of the elementary dendrimer cylinder as a function of generation number (Gn). d is the layer spacing, i.e. the height of the cylinder.

J. Mater. Chem., 2005, 15, 40934105 | 4097

Fig. 5 (a) Schematic representation of the molecular model for dendrimers with two-terminal-chain mesogenic units. (b) Model for the hexagonal columnar supramolecular organisation.

regions: central core, rigid part of the mesogenic units and terminal chains. Therefore, the dendritic molecules fill up thick disks or disk equivalents: for the low generations, two or three molecules are needed to form a complete disk, and from the third generation onwards, just one molecule constitutes a disk, whose thickness depends on the generation number (Fig. 5a). The arrangement of these disks within supramolecular organisations gives rise to cylindrical columns and therefore to the hexagonal columnar mesophase (Fig. 5b). This proposed model was further justified by MD simulation which reproduces the paving of the hexagonal lattice of the dendrimers in a flattened wedge conformation. The result of the calculation evidenced a good filling of the available volume. An enhancement of the micro segregation over the entire simulation experiment time was also observed, contributing to the stabilization of the structure. Furthermore, the compensation of the molecular areas at the dendritic/mesogens interfaces implies the tilt of the peripheral rigid segments with respect to the radial directions. It is interesting to note that the diameter of the columns does not vary significantly with the generation number. Indeed it varies irregularly between 5.3 nm and 6.7 nm, despite the fact that the molecular weight of one single dendrimer varies from 2660 up to 48500 Daltons when going from the lowest generation up to the fourth generation. In other words, this clearly indicates that the dendritic core deforms strongly in one main direction corresponding to that of the columnar axis, whereas the mesogenic units are arranged radially (Fig. 5a) to ensure efficient lateral interactions within and between columnar slices. Similarly to the case of lamellar phases described above, systematic calculations performed from X-ray diffraction and volume data indicate the highest generation dendrimer (containing 64 peripheral mesogenic groups) can fill the equivalent of a disk 1.9 nm thick, whereas the third generation dendrimer containing 16 peripheral
4098 | J. Mater. Chem., 2005, 15, 40934105

mesogenic groups fills a disk 0.5 nm thick. It has been found that, whatever the generation, there is on average 16 mesogenic units over 0.5 nm along the columnar axis. To conclude this section, let us emphasize that both lamellar and columnar mesophases can be obtained with dendrimers, even with those of high molecular weight, the stability of the corresponding phases being ensured by lateral interactions between the (pro)mesogenic units and by a significant deformation of the conformation of the dendritic core. The difference in the symmetry of the mesophase is only related to a small difference in the molecular design of the peripheral mesogenic groups. Those with only one terminal end-chain produce lamellar mesophases, whereas those with two terminal end-chains lead to columnar mesophases. Let us point out also that the dimensions (a few nanometers) of the elementary dendrimer cylinder can be tuned according to the generation number, but the fundamental dimensions of the mesophase structure (layer spacing for the lamellar phases and intercolumnar distance for the columnar phases) do not depend on the size of the dendrimer itself. In the field of liquid crystals, it is well known that many of the materials displaying a columnar mesophase are constituted by disk-like (discotic) molecules. To the best of our knowledge, very little attention has been paid to LC dendrimers containing discotic mesogenic units, namely based on hexasubstituted derivatives of triphenylene. These compounds are also of special interest because of their photoconductive properties. A series of PPI-based dendrimers that incorporate discotic triphenylene mesogenic units at their periphery has been prepared and their properties investigated.36 All the dendrimers except that of the first generation show a hexagonal columnar mesophase in a wide temperature range. The XRD characterisation of this mesophase reveals that its parameters are practically independent of the dendritic generation considered and indicate the existence of a hexagonal sublattice
This journal is The Royal Society of Chemistry 2005

Fig. 6

Schematic representation of the model for the hexagonal columnar mesophase in LC dendrimers with discotic mesogenic units.

in the hexagonal columnar mesophase. This hexagonal sublattice would be generated by the segregated columns containing either the stacked triphenylenes or the dendrimer (Fig. 6). As shown in Fig. 6, each dendrimer column is surrounded by six columns of triphenylene moieties. In this model, the flexible dendritic part adopts a cylindrical arrangement with the four generations having a similar diameter but their height increasing. Once more, the interactions between the mesogenic units determine the supramolecular structure and the dendritic central core adopts the elongated conformation necessary to allow this molecular arrangement. Side-chain LC dendrimers exhibiting the nematic mesophase, which is the most disordered liquid crystalline phase, can also be achieved by the appropriate choice of the mesogenic unit(s) attached to the dendrimer central core. For example, side-chain PPI-derived dendrimers containing mesogenic units with short terminal alkoxy chains (ethoxy, butoxy and pentoxy) or mesogenic units attached laterally, have been prepared in a similar approach to that employed in side-chain liquid crystal polymers.37 In such cases, the side-by-side molecular arrangement typical of lamellar phases is disfavoured, thus promoting a nematic order (Fig. 7) and indeed side-on dendrimers, and some end-on dendrimers bearing mesogens with short aliphatic chains, display a nematic mesophase. This result indicates that these nanophase-structured dendrimers lead to macromolecules with LC properties similar to those of comparable low mass compounds. Tuning the mesophase with codendrimers As described above, the presence of one or two terminal alkoxy chains in the chemical architecture of the mesogenic units induces a drastic modification in the thermotropic behaviour of the dendrimer (from lamellar to columnar structures). The introduction of both types of mesogenic units
This journal is The Royal Society of Chemistry 2005

(containing one and two terminal chains) within the same dendritic structure, i.e. the preparation of random codendrimers, should open a larger possibility of tuning the symmetry of the mesophase exhibited by the material. A series of codendrimers was synthesised, in which both types of mesogenic units were introduced in various proportions38 (Fig. 8). The study of the phase diagram obtained for these codendrimers reveals that, for intermediate compositions, two other mesophases appear between those (smectic A and hexagonal columnar, respectively) of the homodendrimers. Codendrimers with a small content of the two-terminal-chain comonomer exhibit a smectic C phase below the smectic A phase. The appearance of the tilted smectic mesophase is a consequence of the increase in the total number of terminal chains. The tilt of the molecule affords a larger area of the ideal cylinder base, which allows the accommodation of all the terminal chains (Fig. 8a). Larger contents of the twoterminal-chain comonomer in the codendrimers favour the occurrence of a rectangular columnar mesophase. As the number of terminal chains increases, their accommodation within a cylindrical structure becomes unlikely. This elementary dendritic cylinder is deformed into some kind of parallelepiped structure, which in turn promotes the existence of a rectangular columnar mesophase (Fig. 8b). Thus, depending upon the respective concentration of the two types of monomers, orthogonal and/or tilted lamellar mesophases, or else rectangular and/or hexagonal mesophases can be obtained. Tuning the dimensions of the nano-dendritic object with bulky peripheral groups In order to further investigate the influence of the structure of the mesogenic units on the mesomorphic behaviour of the dendrimers, bulky units have been introduced at the periphery
J. Mater. Chem., 2005, 15, 40934105 | 4099

Fig. 7 General structure of nematic dendrimers with end-on and side-on mesogenic units.

of the fourth and fifth generations of amino-terminated PAMAM and PPI (Fig. 9), with the aim of achieving cubic mesophases, since the increased volume around the dendritic core could force the dendrimer to adopt a globular conformation.39 However, all the dendrimers prepared show only a hexagonal columnar mesophase, even those bearing the largest mesogenic unit with nine terminal chains. In contrast to the case of the bulky dendrons reported by Percec et al. which can self-organize into micellar cubic mesophases,40 the coating of the present PAMAM and PPI dendrimers is not sufficient to

lead to a globular shape of the dendrimer able to generate such cubic mesophases. A model of organization of these dendritic supermolecules, deduced from X-ray diffraction results and theoretical calculations made therefrom, has been proposed. This molecular model implies a cylinder, whose inner part is occupied by the dendritic core and the bulky mesogenic units spreading around it, as shown in Fig. 10. The dendritic core adopts an extended conformation which is possible due to the great flexibility (conformational freedom) of the PAMAM and PPI skeletons. In this way, the molecular

Fig. 8 (a) Schematic representation of the molecular model for codendrimers with one- and two-terminal-chain mesogenic units (SmC phase). (b) Molecular model for codendrimers with one- and two-terminal-chain mesogenic units (rectangular columnar phase).

4100 | J. Mater. Chem., 2005, 15, 40934105

This journal is The Royal Society of Chemistry 2005

Fig. 9 Chemical structure of LC dendrimers with bulky mesogenic units.

model of thick disks proposed for the LC dendrimers with two-terminal-chain mesogenic units is transformed into a model consisting of a long cylinder, which is the result of the axial elongation of a disk caused by the extended conformation adopted by the dendrimer core in order to accommodate the bulky units around it. X-Ray diffraction investigations show that the diameter of the cylinder is about the same, between 5.3 and 5.8 nm, whatever the generation and the bulky peripheral group (Fig. 11). In this context, it is interesting to compare these values to those of the expected diameters of the dendrimers in their most extended conformation, between 10 and 12.3 nm. More detailed calculations indicate that the internal diameter of the cylinder corresponding to the dendritic core decreases when the bulkiness of the peripheral group increases, for example going from 2.5 down to 1.8 nm in the case of the fifth generation dendrimers, in agreement with an

elongation of the dendritic core along the columnar axis. Let us emphasize also that such molecular engineering involving bulky peripheral groups is another way to tune the dimensions of the nano-dendritic objects.

III. New morphologies in main-chain LC dendrimers

Regarding the dendrimers considered in this section, the branching points are no longer single atoms (C, N, Si) but consist of rigid and anisotropic molecular moieties instead, these units being linked together through long and flexible alkyl spacers (Fig. 12b). The mesogenic groups are now present

Fig. 10 Molecular model for the LC dendrimers with bulky mesogenic units.

Fig. 11 Variation of the size of the elementary dendrimer cylinder with bulky mesogenic groups. L3, L6 and L9 are for the bulky groups containing 3, 6 and 9 terminal aliphatic chains respectively. 24 and 25 are for the dendrimers of the fourth and fifth generations respectively. h is the height of the cylinder and int. is the internal diameter corresponding to the dendritic part only.

This journal is The Royal Society of Chemistry 2005

J. Mater. Chem., 2005, 15, 40934105 | 4101

Fig. 12 Schematic 2D representation of side-chain (a) and mainchain (b) liquid crystalline dendrimers.

at every level of the dendritic hierarchy, and the dendrimers are forced to adopt constrained and regular structures. As such, the anisometric branches do not radiate isotropically as in side-chain dendrimers, but, in contrast, favour preferentially an anisotropic order by a gain in the enthalpy of the system in order to produce the most stable structure. These main-chain dendrimers may represent an interesting alternative for the development of original molecular materials having a new architecture (Fig. 12b). As far as we are aware, the most closely related compounds which have been reported are the so-called willow-like dendrimers synthesized by Percec et al. some years ago.24,41 Shibaev et al. also foresaw the elaboration of such compounds, but have not yet reported on their synthesis and physical properties.42 Octopus LC dendrimers As already mentioned, conventional end-group dendrimers, where the core matrix is either PAMAM, PPI or carbosilane,

are usually built up using a divergent approachthat is constructed outwards from a central core, and the mesogenic moieties are attached to the periphery at the final stage of the synthesis. For the presently described dendrimers, this synthetic approach appeared unsatisfactory as a number of additional steps were preliminarily required to synthesize the mesogenic moiety at each generation, before moving up to the next generation, with the consequent diminishing overall reaction yields hampered by reactivity and purification problems. Therefore, in order to improve the reactivity of the connecting sites and facilitate the purification of the dendritic materials, a modular synthesis was elaborated for the preparation of the branches. Each constituent part was prepared separately and later assembled selectively together. The anisotropic units selected were a tolane-based or a stilbene-based moiety, due to both their thermal stability and chemical versatility. Such a poor mesogenic segment, despite its anisotropy and rigidity, was also chosen in order to test whether mesomorphism could be induced solely by the dendrimerization process. The dendritic branches, functionalized by a focal acid group, were then coupled to a small tetrapodand core unit, bearing four amino groups, to yield the final dendrimer (Fig. 13).43 The monodispersity and the analytical characterization of the compounds were achieved by SEC, MALDI-TOF and various spectroscopic techniques. Homolithic (X 5 Y) and heterolithic (X ? Y) dendrimers were synthesized and all of them were liquid crystals. Mesomorphism was thus induced by the precise assembling of these non-mesogenic units within the dendritic frame. Moreover, the mesophase stability was found to greatly depend on the localization of the various units within the heterolithic systems. The dendrimers bearing only one aliphatic end-chain at the extremity of the outer tolane unit exhibit smectic behaviour.

Fig. 13 The different octopus dendrimers.

4102 | J. Mater. Chem., 2005, 15, 40934105

This journal is The Royal Society of Chemistry 2005

The X-ray diffraction patterns are characterized by the presence of 34 orders of reflection corresponding to a well defined lamellar stacking. On the basis of these X-ray patterns, the high temperature phase has been assigned as a disordered smectic A phase, whereas the other, low-temperature, phase is due to an extra in-layer order, such as a hexatic smectic B phase, the layer spacing being rather large, in the range 1012 nm. This confirms the prolate conformation of the dendrimers in both smectic phases with the peripheral anisotropic units being almost perpendicular to the layer normal direction. In this case, due to geometric constraints, the mesogenic groups of the first row of the arborescence, located in the smectic sub-layer, ought to be tilted to match the molecular area of roughly two mesogenic groups, in order to occupy a larger apparent surface area (Fig. 14). Of course, the rigid units contained in these aromatic slabs are tilted with respect to the layer normal, but the tilt is not necessarily correlated, i.e. short-range tilt order. In other words, the relative disordered distribution of the rigid parts yields an apparent zero tilt angle, and then a uniaxial smectic sub-layer. This model is consistent with the molecular conformation deduced from MD calculations, and also explains the difference between the periodicity and the molecular length. The SmB-to-SmA transformation corresponds to the loss of the hexatic order consequent on the lateral disorganization of the anisotropic units of the outer slabs.44 Therefore, the morphology of the smectic phases generated by such multiblock molecules is quite unique in that it possesses a two-level molecular organization, each being dependent on the other one. It consists of an internal sublayer made of tilted rigid segments with no correlation of the tilt, flanked by outer slabs inside which the mesogenic groups are arranged perpendicular to the layer (Fig. 14). Molecular modeling supports this view of strongly segregated multilayer structures, with interfaces between the various molecular parts. Obviously, these interfaces are not so well defined due to thermal fluctuations. Nevertheless, let us point out that due to

Fig. 14 Model for the octopus conformation and organization within the smectic phases.

this peculiar structural feature, such layered mesophases cannot exactly be described as purely SmA or SmB phases. As for the other set of dendrimers bearing two or three aliphatic chains at the extremity of the outer tolane or stilbene part, they all exhibit a columnar mesophase with a hexagonal symmetry. The formation of columnar mesophases in nondiscotic systems, and particularly with polycatenar mesogens,45 is a consequence of the mismatch between the surface areas of the aromatic cores and the cross-section of the aliphatic chains, resulting in the curvature of all the interfaces, as has been discussed with side-chain LCDs bearing polycatenar end-groups. In the present case, in order to compensate the discrepancy between the cross-sections of both the anisometric segments and the chains, one can also imagine the former to be tilted and distributed in a splay fashion, with respect to the columnar axis, also resulting in the curvature of the interfaces.46 Indeed, the parameters of the hexagonal lattices obtained experimentally, a 5 910 nm, correspond fairly well to the diameter of the dendrimers in flattened conformation, ranging between 10 and 11 nm as estimated by MD simulation. It is therefore highly probable that the octopus preferably adopts an oblate shape within the columns that is a flattened or a wedge-like conformation with the anisotropic blocks lying more or less in the 2D hexagonal lattice plane, rather than a prolate conformation (cylindrical) as in the smectic systems. The formation of the mesophase results from the selfassembling process of octopus molecules necessarily adopting pre-defined shapes as for the dendromesogens described by Percec et al.,18,47 where columnar structures are generated from the self-assembling of the most stable molecular conformations having either a wedge-like or half-disc shape. The overall molecular conformations of the dendrimers in the mesophase are driven by the steric congestion of the terminal aliphatic chains and depend on the segregation between the different constituent blocks. In the present case, one or two molecular conformations likely predominate to satisfy the geometrical requirements. The formation of supramolecular disks or columns results from the molecular association of these two types of dendritic conformations, as shown in Fig. 15. One possibility consists of assembling together two or three dendrimers with solely the wedge-like conformation and another one of associating three octopuses with both wedgelike and flattened conformations in a 2 : 1 mixture. Clearly, one has to bear in mind that these various arrangements must co-exist in the columns since it is not possible to favour one over the other. Then, the resultant columns further selforganize into a 2D hexagonal lattice. Considering the diblock, alternated chemical nature of these octopus dendrimers, an onion morphology for the columns is most probable.44 In these dendrimers, the morphology of the mesophase is also determined by the number of alkyl chains grafted on the peripheral mesogenic group. Indeed, the change in the number of terminal chains per end group (one or two) modifies the relationships between the hard parts and the soft parts, and consequently the molecules will adopt either a parallel (prolate) or a flat (oblate) conformation. The formation of the smectic lamellar phases is the result of the parallel disposition of the mesogenic groups on both sides of the focal
J. Mater. Chem., 2005, 15, 40934105 | 4103

This journal is The Royal Society of Chemistry 2005

Fig. 15 Schematic representation of the two possible molecular conformations of the dendrimers, and their self-assembling and selforganization processes into the columns of the Colh phase

tetravalent core, the dendrimer adopting the shape of a giant elongated multipode, and then organizing into layers. In contrast, the grafting of additional terminal chains at the periphery prevents such a parallel disposition of the promesogenic groups, which are forced to be radially arranged around the central moiety: the dendrimer can adopt the shape of a flat-tapered object which will self-arrange into supramolecular columns. Although the expected mesomorphism cross-over from lamellar to columnar structures was observed with increasing the number of terminal chains,33,34 the two-phase behavior or the induction of intermediate mesophases (such as bicontinuous cubic or rectangular columnar phase37 for instance) is nevertheless absent within a single dendrimer.

towards a columnar organization with an onion internal morphology. As for the smectic-like phases, the dendrimers adopt a parallel conformation, and due to the particular nature of the molecules, the segments of the first generation, which are located in the smectic sub-layer, are tilted leading to a two-level molecular organization. In addition, and very importantly, the individual dendritic nano-objects forming the elementary bricks of the high-level self-organization, can be tuned in shape and size through suitable molecular design, for example by using mesogenic groups of different types and by building codendrimers or by using bulky promesogenic groups. Thus, the structural concept of these LC dendrimers proves to be innovative and versatile as it offers many new opportunities in the design of a wide range of multicomponent systems with specific properties for potential novel applications. Indeed, it appears that large flexibility and freedom are allowed in the choice of the elementary anisotropic bricks of these LCDs, without the suppression of the mesomorphic properties. Additionally, such bricks can be independently interchanged, and the stability of the mesophases accordingly modulated, proving the good sensitivity of the dendritic scaffold with the nature and the mutual arrangement of the mesogenic segments (spatial location), and the intimate relationships with the mesomorphic properties. The high sensitivity of such dendrimers with the surrounding environment (properties versus molecular structure) could be, in principle, beneficial to access certain kinds of molecular sensors, i.e. to use such supermolecules as tools to test how properties in general may be altered or modulated upon delicate external stimulation. Systems with photoactive moieties can also be envisaged in view of optoelectronic applications.

IV. Conclusion
In conclusion, it is shown that the induction of mesomorphic properties14,16 can be extended to various types of dendrimers of high molecular weight, even with non-mesomorphic building blocks (dendrimerisation promotes LC phases). This is achieved through a precise chemical design, combined with the role of different types of intermolecular interactions. The interactions between the polarizable promesogenic units combined with the nanophase segregation contribute to the stabilization of the aromatic sublayers (in the smectic phases) and of the columnar cores (in the columnar phases). In the case of side-chain LC dendrimers, the presence of one chain per mesogenic unit favours their parallel arrangement; hence the lamellar morphology is promoted. The increased aliphatic chain density due to the introduction of two chains per mesogenic unit, imposes a curved interface with the promesogenic units arranged radially, hence inducing the columnar mesomorphism. The same is true in the case of main-chain LC dendrimers. The high density of aliphatic chains likely imposes also curved interfaces at all hierarchical levels of the dendrimers, i.e. between the promesogenic units (internal and peripheral), the aliphatic spacers and the terminal chains respectively, forcing the molecules to adopt a wedgelike conformation, thus promoting their self-assembling
4104 | J. Mater. Chem., 2005, 15, 40934105

Acknowledgements
This work has been supported by the CICYT of Spain and the FEDER funds (EU) under the projects MAT2002-04118-C0201 and MAT2003-07806-C01, by the European Union under the RTN Project Supermolecular Liquid Crystal Dendrimers (LCDD) (HPRN-CT2000-00016), and by the Diputacion General de Aragon. B.D. and D.G. would like to thank C. Bourgogne and B. Heinrich for technical help and useful discussions.

References
1 G. Cao, Nanostructures and Nanomaterials, Imperial College Press, London, 2004. 2 P. Yong, The Chemistry of Nanostructured Materials, World Scientific, Singapore, 2003. 3 D. A. Tomalia and J. M. Frechet, J. Polym. Sci. A, Polym. Chem., 2002, 40, 2719. 4 (a) D. A. Tomalia and H. Dupont Durst, Top. Curr. Chem., 1993, 165, 193; (b) N. Ardouin and D. Astruc, Bull. Soc. Chem. Fr., 1995, 132, 875; (c) G. R. Newkome, C. N. Moorefield and F. Vogtle, in Dendrimers and Dendrons: Concepts, Synthesis and Applications, Wiley-VCH, Weinheim, 2001. 5 (a) S. M. Grayson and J. M. Frechet, Chem. Rev., 2001, 101, 3819; (b) C. J. Hawker, Adv. Polym. Sci., 1999, 147, 113. 6 D. A. Tomalia, A. M. Naylor and W. A. Goddard, III, Angew. Chem., Int. Ed. Engl., 1990, 29, 138.

This journal is The Royal Society of Chemistry 2005

7 (a) J. Issberner, R. Moors and F. Vogtle, Angew. Chem., Int. Ed. Engl., 1994, 33, 2413; (b) G. M. Dykes, J. Chem. Technol. Biotechnol., 2001, 76, 903. 8 (a) D. Astruc, C. R. Acad. Sci. Paris Ser. II, 1996, 322, 757; (b) K. Uhrich, Trends Polym. Sci., 1997, 5, 388; (c) D. K. Smith and F. Diederich, Chem. Eur. J., 1998, 4, 1353; (d) S. Hecht and J. M. J. Frechet, Angew. Chem., Int. Ed., 2001, 40, 74; (e) U. Boas and P. M. Heegard, Chem. Soc. Rev., 2004, 33, 49. 9 (a) H. F. Chow, T. K. K. Mong, M. F. Nongrum and C. W. Wan, Tetrahedron, 1998, 54, 8543; (b) M. Fischer and F. Vogtle, Angew. Chem., Int. Ed., 1999, 38, 884; (c) K. Inoue, Prog. Polym. Sci., 2000, 25, 453. 10 (a) D. Seebach, P. B. Rheiner, G. Greiveldinger, T. Butz and H. Sellner, Top. Curr. Chem., 1998, 197, 125; (b) B. Romagnoli and W. Hayes, J. Mater. Chem., 2002, 12, 767. 11 (a) E. C. Constable, Chem. Commun., 1997, 1073; (b) C. Gorman, Adv. Mater., 1998, 10, 295; (c) G. R. Newkome, E. He and C. N. Moorefield, Chem. Rev., 1999, 99, 1689. 12 (a) J. P. Majoral and A. M. Caminade, Chem. Rev., 1999, 99, 845; (b) H. Frey, C. Lach and K. Lorenz, Adv. Mater., 1998, 10, 279; (c) H. Frey and C. Schlenk, Top. Curr. Chem., 2000, 210, 69. 13 (a) F. Zeng and S. C. Zimmerman, Chem. Rev., 1997, 97, 1681; (b) V. V. Narayanan and G. R. Newkome, Top. Curr. Chem., 1998, 197, 19; (c) E. C. Constable and C. E. Housecroft, Chimia, 1998, 52, 533; (d) D. K. Smith and F. Diederich, Top. Curr. Chem., 2000, 210, 183. 14 (a) J. W. Goodby, G. H. Mehl, I. M. Saez, R. P. Tuffin, G. Mackenzie, R. Auzely-Velty, T. Benvegnu and D. Plusquellec, Chem. Commun., 1998, 2057; (b) J. W. Goodby, Curr. Opin. Solid State Mater. Sci., 1999, 4, 361; (c) I. M. Saez and J. W. Goodby, J. Mater. Chem., 2005, 15, 26. 15 (a) S. A. Ponomarenko, N. I. Boiko and V. P. Shibaev, Polym. Sci. Ser. C, 2001, 43, 1 and references therein; (b) D. Guillon and R. Deschenaux, Curr. Opin. Solid State Mater. Sci., 2002, 6, 515. 16 (a) D. Demus, Liq. Cryst., 1989, 5, 75; (b) C. Tschierske, Prog. Polym. Sci., 1996, 21, 775; (c) G. Pelzl, S. Diele and W. Weissflog, Adv. Mater., 1999, 11, 707; (d) C. Tschierske, Annu. Rep. Prog. Chem., Sect. C, 2001, 97, 191; (e) C. Tschierske, J. Mater. Chem., 2001, 11, 2647; (f) X. Cheng, M. Prehm, M. K. Das, J. Kain, U. Baumeister, S. Diele, L. Dag, A. Blume and C. Tschierske, J. Am. Chem. Soc., 2003, 125, 10977. 17 H. Meier, M. Lehmann and U. Kolb, Chem. Eur. J., 2000, 6, 2462. 18 (a) V. S. K. Balagurusamy, G. Ungar, V. Percec and G. Johansson, J. Am. Chem. Soc., 1997, 119, 1539; (b) S. D. Hudson, H. T. Jung, V. Percec, W. D. Cho, G. Johansson, G. Ungar and V. S. K. Balagurusamy, Science, 1997, 278, 449; (c) V. Percec, W. D. Cho and G. Ungar, J. Am. Chem. Soc., 2000, 122, 10273. 19 (a) B. Dardel, R. Deschenaux, M. Even and E. Serrano, Macromolecules, 1999, 32, 5193; (b) T. Chuard, B. Dardel, R. Deschenaux and M. Even, Carbon, 2000, 38, 1573; (c) B. Dardel, D. Guillon and B. Heinrich, J. Mater. Chem., 2001, 11, 2814. 20 (a) U. Stebani, G. Lattermann, M. Wittenberg and J. H. Wendorff, Angew. Chem., Int. Ed. Engl., 1996, 35, 1858; (b) J. Barbera, M. Marcos, A. Omenat, J. L. Serrano, J. I. Martnez and P. J. Alonso, Liq. Cryst., 2000, 27, 255; (c) R. Deschenaux, E. Serrano and A. M. Levelut, Chem. Commun., 1997, 1577. 21 (a) G. H. Mehl and I. M. Saez, Appl. Organomet. Chem., 1999, 13, 261; (b) I. M. Saez and J. W. Goodby, Liq. Cryst., 1999, 26, 1101; (c) I. M. Saez and J. W. Goodby, Chem. Eur. J., 2003, 9, 4869; (d) S. Campidelli, C. Eng, I. M. Saez, J. W. Goodby and R. Deschenaux, Chem. Commun., 2003, 1520; (e) I. M. Saez and J. W. Goodby, J. Mater. Chem., 2003, 13, 2727. 22 (a) P. J. Flory, in Principles of Polymer Chemistry, Cornell University Press, Ithaca, NY, 1953; (b) Y. H. Kim, J. Polym. Sci. A; Polym. Chem., 1998, 36, 1685; (c) H. Frey and D. Holter, Acta Polym., 1999, 50, 67; (d) A. Hult, M. Johansson and E. Malmstrom, Adv. Polym. Sci., 1999, 143, 1.

23 (a) V. Percec and M. Kawasumi, Macromolecules, 1992, 25, 3843; (b) S. Bauer, H. Fischer and H. Ringsdorf, Angew. Chem., Int. Ed. Engl., 1993, 32, 1589. 24 (a) V. Percec, P. Chu and M. Kawasumi, Macromolecules, 1994, 27, 4441; (b) S. W. Hanh, S. Y. K. Yun, J. I. Jin and O. H. Han, Macromolecules, 1998, 31, 6417. 25 A. Sunder, M. F. Quincy, R. Mulhaupt and H. Frey, Angew. Chem., Int. Ed., 1999, 38, 2928. 26 K. Lorenz, D. Holter, B. Stuhn, R. Mullhaupt and H. Frey, Adv. Mater., 1996, 8, 414. 27 M. W. P. L. Baars, S. H. M. So ntjes, H. M. Fischer, H. W. I. Peerlings and E. W. Meijer, Chem. Eur. J., 1998, 4, 2456. 28 J. H. Cameron, A. Facher, G. Lattermann and S. Diele, Adv. Mater., 1997, 9, 398. 29 K. Lorenz, D. Holter, B. Stuhn, R. Mulhaupt and H. Frey, Adv. Mater., 1996, 8, 414. 30 R. M. Richardson, I. J. Whitehouse, S. A. Ponomarenko, N. I. Boiko and V. P. Shibaev, Mol. Cryst. Liq. Cryst., 1999, 330, 167. 31 R. M. Richardson, S. A. Ponomarenko, N. I. Boiko and V. P. Shibaev, Liq. Cryst., 1999, 26, 101. 32 (a) S. A. Ponomarenko, E. V. Agina, N. I. Boiko, E. A. Rebrov, A. M. Muzafarov, R. M. Richardson and V. P. Shibaev, Mol. Cryst. Liq. Cryst., 2001, 364, 93; (b) E. V. Agina, S. A. Ponomarenko, N. I. Boiko, E. A. Rebrov, A. M. Muzafarov and V. P. Shibaev, Polym. Sci. Ser. A, 2001, 43, 1000. 33 A. Y. Bobrovsky, A. A. Pakhomov, X. M. Zhu, N. I. Boiko and V. P. Shibaev, Polym. Sci. Ser. A, 2001, 43, 431. 34 J. Barbera, M. Marcos and J. L. Serrano, Chem. Eur. J., 1999, 5, 1834. 35 (a) M. Marcos, R. Gimenez, J. L. Serrano, B. Donnio, B. Heinrich and D. Guillon, Chem. Eur. J., 2001, 7, 1006; (b) B. Donnio, J. Barbera, R. Gimenez, D. Guillon, M. Marcos and J. L. Serrano, Macromolecules, 2002, 35, 370. 36 M. D. McKenna, J. Barbera, M. Marcos and J. L. Serrano, J. Am. Chem. Soc., 2005, 127, 619. 37 L. Pastor, J. Barbera, M. McKenna, M. Marcos, R. Martn Rapu n, J. L. Serrano, G. R. Luckhurst and A. Mainal, Macromolecules, 2004, 37, 9386. 38 J.-M. Rueff, J. Barbera, B. Donnio, D. Guillon, M. Marcos and J. L. Serrano, Macromolecules, 2003, 36, 8368. 39 J.-M. Rueff, J. Barbera, M. Marcos, A. Omenat, R. Martn Rapun, B. Donnio, D. Guillon and J. L. Serrano, submitted. 40 V. Percec, C. M. Mitchell, W.-D. Choo, S. Uchida, M. Glodde, G. Ungar, X. Zeng, Y. Liu, V. S. K. Balagurusamy and P. A. Heiney, J. Am. Chem. Soc., 2004, 126, 6078. 41 (a) V. Percec, P. Chu, G. Ungar and J. Zhou, J. Am. Chem. Soc., 1995, 117, 11441; (b) J. L. Li, K. A. Crandall, P. Chu, V. Percec, R. G. Petschek and C. Rosenblatt;, Macromolecules, 1996, 29, 7813. 42 S. A. Ponomarenko, E. A. Rebrov, N. I. Boiko, A. M. Muzafarov and V. P. Shibaev, Polym. Sci. Ser. A, 1998, 40, 763. 43 L. Gehringer, D. Guillon and B. Donnio, Macromolecules, 2003, 36, 5593. 44 L. Gehringer, C. Bourgogne, D. Guillon and B. Donnio, J. Am. Chem. Soc., 2004, 126, 3856. 45 (a) J. Malthete, H. T. Nguyen and C. Destrade, Liq. Cryst., 1993, 13, 171; (b) H. T. Nguyen, C. Destrade and J. Malthete, Adv. Mater., 1997, 9, 375. 46 (a) D. Fazio, C. Mongin, B. Donnio, Y. Galerne, D. Guillon and D. W. Bruce, J. Mater. Chem., 2001, 11, 2852; (b) A. I. Smirnova, D. Fazio, E. F. Iglesias, C. G. Hall, D. Guillon, B. Donnio and D. W. Bruce, Mol. Cryst. Liq. Cryst., 2003, 396, 227. 47 (a) V. Percec, D. Schlueter, G. Ungar, S. Z. D. Cheng and A. Zhang, Macromolecules, 1998, 31, 1745; (b) G. Ungar, Y. Liu, X. Zeng, V. Percec and W. D. Cho, Science, 2003, 299, 1208.

This journal is The Royal Society of Chemistry 2005

J. Mater. Chem., 2005, 15, 40934105 | 4105