Coarse-grained simulation studies of a liquid crystal dendrimer: towards computational predictions of nanoscale structure through microphase separation Zak E. Hughes, a Mark R. Wilson* a and Lorna M. Stimson b Received 3rd August 2005, Accepted 23rd September 2005 First published as an Advance Article on the web 17th October 2005 DOI: 10.1039/b511082c Coarse-grained simulations are described in which the behaviour of a system of model liquid crystalline dendrimer molecules is studied in both liquid and smectic-A liquid crystalline phases. The model system is based on a third generation carbosilane dendrimer, which is functionalised at the surface by short polymeric chains terminated in mesogenic units. The design of the coarse- grained model is based on initial Monte Carlo studies of a single carbosilane molecule at an atomistic level, which yield structural data. The coarse-grained dendrimer is represented in terms of a combination of spherical sites representing the dendrimer core and polymer chains, and spherocylinders representing the mesogenic groups. A strong coupling is seen between internal molecular structure and molecular environment, with individual dendrimer molecules undergoing a remarkable transition from spherical to rod-shaped at the isotropic–smectic phase transition. The driving force for mesophase formation is provided by nanoscale microphase separation of mesogens and the dendrimer core. 1 Introduction Liquid crystal dendrimers (LCDrs) are unusual materials. The underlying branched structure of a dendrimer should lead to molecular architectures with spherical symmetry. This would appear to rule out the formation of liquid crystalline phases, which normally require the presence of anisotropy in mole- cular shape and in molecular interactions. However, synthetic chemists have produced a number of interesting architectures, which effectively combine a dendritic structure with liquid crystallinity. For example, mesogenic groups may be incorpo- rated into the dendrimer between branching points, leading to the dendrimer molecules themselves becoming anisotropic and able to form calamitic nematic and smectic thermotropic phases. 1 Alternatively, mesogenic groups can be bonded to the ‘‘surface’’ of a dendrimer, either directly or by using short ‘‘decoupling’’ chains as shown by Richardson and co-workers. 2–4 Here, conformational changes allow the dendrimer to rearrange the distribution of its terminal mesogenic groups providing a mechanism for formation of liquid crystal phases. Wilson and co-workers 5 have recently simulated a model for a third generation carbosilane liquid crystalline dendrimer in solution in isotropic, nematic and smectic-A solvents. The results show an interesting coupling between dendrimer structure and molecular environment. In the isotropic solvent, the dendrimer itself forms an isotropic sphere with the meso- genic groups randomly arranged on the periphery. In the nematic solvent, the dendrimer rearranges its molecular structure, so that the mesogenic groups can lie, on average, parallel to the nematic director with an average order para- meter approaching that of the solvent. In the smectic solvent, the dendrimer rearranges further, so that mesogens can lie commensurate with the smectic layers. These simulations were stimulated by recent synthetic and X-ray diffraction studies of carbosilane LCDrs. 2,3 Here it seems that mesophase formation can be influenced by the generation number. In the lower generations, the formation of rods (as seen in the simulations discussed above) leads to smectic behaviour characterised in the X-ray measurements by a smectic layer spacing. In generation 5, the dendrimers can also form a columnar phase. Here it is likely that conforma- tional changes can lead to the dendrimer rearranging structure in a second way to form discs, which can stack to form columns, which in turn pack to form columnar phases. The purpose of the current paper is to design a coarse- grained (CG) model to study carbosilane LCDrs in the bulk and to look at the structure of the phases formed. This is a difficult task. Unlike simpler macromolecules, such as linear homopolymers (which can be coarse-grained by designing a suitable model for a single monomer unit) a CG model of the dendrimer must account for a range of different interaction sites. Moreover, it must be able to represent a complicated molecular shape, which can change structure through con- formational rearrangement, yet be sufficiently ‘‘cheap’’ to simulate a relatively large number of molecules over long simulation times. The design of the simulation model is described in section 2. Simulation results from long simulation runs are described in section 3, and we point to some con- clusions in section 4 and mention the possible extension of this type of model to other problems in soft matter chemistry a Department of Chemistry, University of Durham, South Road, Durham, DH1 3LE, UK. E-mail: mark.wilson@durham.ac.uk b Biophysics and Statistical Mechanics Group, Laboratory of Computational Engineering, Helsinki University of Technology, P.O. Box 9203, 02015-HUT, Finland. E-mail: lorna@lce.hut.fi PAPER www.rsc.org/softmatter | Soft Matter 436 | Soft Matter, 2005, 1, 436–443 This journal is ß The Royal Society of Chemistry 2005