Liquid crystals represent the fourth state of matter, being intermediate between the solid and liquid states. As such, they are anisotropic fluids and it is this fact which has led to the great interest in, and applications of, these materials. Liquid crystals were first identified in the mid-to-late 18th century, but interest in them really took off in the early 1970s when George Gray, at the University of Hull, discovered the cyanobiphenyl liquid crystals. These materials had all of the correct properties for application in liquid crystal displays and they were commercialised by a British consortium. The LCD industry was born!
Since that time, the interest in both applied and academic topics in liquid crystals has been substantial, witness the biennial International Conference which tends to attract around 800 participants. Our interests in the field are in the broad area of synthesis and how the molecular structure relates to the observable phases formed and their properties, and our 'twist' is that many of the systems we investigate contain a metal atom as an integral part of the structure.
One of the best parts of the subject is the characterisation of the different liquid crystal mesophases for which we use, as our main technique, polarised optical microscopy. As the phases are anisotropic, then they are birefringent and so viewed between crossed polarisers, characteristic interference patterns can be observed. These are rather beautiful and some examples are shown below.
Left Nematic mesophase of an organomanganese complex Right Smectic C phase of a polycatenar mesogen
Columnar mesophase of a polycatenar mesogen.
Liquid crystals may be first classified according to the manner in which the order of the solid state is destroyed to form the different mesophases. Thus, in thermotropic liquid crystals, it is the primary action of heat, driving temperature changes, which causes transitions to happen. Thermotropic liquid crystals can be further classified into high molar mass (i.e. polymers) and low molar mass (i.e. non-polymeric) materials.
It is the low molar mass materials that concern us for most of this website and so it is of interest to know that they may be further and broadly sub-classified as calamitic (rod-like), discotic (disk-like) or polycatenar (see below) according to their shape. Calamitic mesogens show primarily nematic and smectic mesophases, while discotic mesogens show nematic and columnar mesophases.
The other major classification of liquid crystals is the lyotropic family, where the order of the solid state is destroyed by the action of a solvent, typically, but not always, water. Molecules that form lyotropic mesophases are disk-like (forming nematic and columnar phases), polymeric (forming nematic and hexagonal phases), or surfactant (forming cubic, hexagonal and lamellar phases).
The earliest liquid crystal which contained a metal as an integral part of the structure was a diphenylmercury(II) compound reported in 1923 by Daniel Vorländer. [I-1] Not much happened then until the late 1970s with the discovery of the metal dithiolenes by Anne-Marie Giroud and Ulrich Mueller-Westerhoff, [I-2] but in the mid 1980s, the field really started to take off. There are many reasons why one might want to include a metal in a liquid-crystalline structure, quite apart from natural curiosity. For example, metals are stable radicals, their complexes are coloured, they are centres of polarisable electron density and, of course, they are centres of reactivity. Some of the more recent reviews and overviews of the subject are collected below. [I-3]
What follows is a brief description of some of the projects which are either ongoing or which have been completed recently in the group. Note that not everything has a metal in there! We will concentrate first on thermotropic liquid crystals, and then look at lyotropic examples and how they may be applied.
[I-1] See eg DW Bruce, K Heyns and V Vill, Liq Cryst, 1997, 23, 813.
[I-2] A-M Giroud and UT Mueller-Westerhoff, Mol Cryst, Liq Cryst Lett, 1977, 41, 11.
[I-3] B Donnio, D Guillon R Deschenaux and D W Bruce in "Comprehensive Coordination Chemistry II", Eds. J A McCleverty and T J Meyer, Pergamon, Oxford, 2003, Vol. 7, chapter 7.9, pp 357-627; B Donnio and D W Bruce, "Structure and Bonding", Ed. D M P Mingos, Springer-Verlag, 1999, 95, chapter 5, pp 193; S R Collinson and D W Bruce, "Transition Metals in Supramolecular Chemistry", Ed. J-P Sauvage, Wiley, Chichester, 1999, chapter 7, pp 285; L Oriol and J-L. Serrano, Adv Mater, 1995, 7, 348; L Oriol, M Pinol and J-L Serrano, Prog Polym Sci, 1997, 22, 873; J-L Serrano, (ed) "Metallomesogens", VCH, Weinheim 1996; K Binnemans and C Görller-Walrand, Chem Rev, 2002, 102, 2303
Liquid-crystalline metal complexes have been dominated for many years by metals capable of existing in linear and/or planar geometries, notably those of Groups 9-11. Exciting as this is, it excludes a large number of metals as candidates for liquid crystal complexes and so it was of interest to discover how liquid crystal may be constructed in which there was a metal with a higher coordination number. [HCN-1]
Orthometallated Imines: Thus, in 1994/5, we reported some mesomorphic complexes of octahedral Mn(I) and Re(I). [HCN-2] We were able to obtain liquid crystallinity in these systems as we countered the reduction in anisotropy which accompanied the use of a six-coordinate metal, by employing a highly anisotropic ligand. Representative structures are given below.
We have varied this system to investigate certain structure/property relationships [HCN-3] and have looked at, for example, bimetallic systems[HCN-4] and systems containing fluorinated chains. [HCN-5]
The systems with the fluorinated chains are particularly interesting as the ligands (e.g. below) show a thermotropic cubic phase. An interpretation of why this might be is found in reference .
[HCN-1] D W Bruce, Adv Mater, 1994, 6, 699
[HCN-2] D W Bruce and X-H Liu, Chem Commun, 1994, 729; D W Bruce and X-H Liu, Liq Cryst, 1995, 18, 165.
[HCN-3] X-H Liu, M N Abser and D W Bruce, J Organomet Chem, 1998, 551, 271 (Publisher's Erratum: 1999, 577, 150); X-H Liu, I Manners and D W Bruce, J Mater Chem, 1998, 8, 1555; X-H Liu, B Henirich, I Manners, D Guillon and D W Bruce, J Mater Chem, 2000, 10, 637; M-A Guillevic, M J Danks, S K Harries, S R Collinson, A D Pidwell and D W Bruce, Polyhedron, 2000, 19, 249
[HCN-4] M-A Guillevic, M E Light, S J Coles, T Gelbrich, M B. Hursthouse and D W Bruce, J Chem. Soc, Dalton Trans, 2000, 1437.
[HCN-5] M-A Guillevic and D W Bruce, Liq Cryst, 2000, 27, 153; M-A Guillevic, T Gelbrich, M B Hursthouse and D W Bruce, Mol Cryst, Liq Cryst, 2001, 363, 147.
Diazabutadienes: We were able to show that the general approach of using highly anisotropic ligands as a means to realise mesomorphic octahedral complexes was more general and one of the systems we described employed diazabutadienes as ligands to halorhenium(I) centres as shown below.  Here we found that the transition temperatures were strongly dependent on the nature and size of the anionic ligand bound to the Re centre. 
 S Morrone, G Harrison and D W Bruce, Adv Mater, 1995, 7, 665.
 S Morrone, D Guillon and D W Bruce, Inorg Chem, 1996, 35, 7041.
Bipyridines: 5,5'-Disubstituted-2,2'-bipyridines also turned out to be suitable ligands to generate liquid crystals from halorhenium(I) complexes,  but in this case, there was also interesting mesomophism to be found in the ligands themselves.  In particular, we synthesised a series of polycatenar bipyridines  (more on polycatenars later) and established a classic phase diagram for such systems which is shown below. More recently, we have extended this work on a -diimine systems to phenanthrolines.
One of the reasons for trying to understand how high coordination number metal centres may be incorporated into liquid crystals was our desire to move to complexes of lanthanides where the coordination number is typically nine. In collaboration with Galyametdinov and Ovchinnikov in Kazan, Guillon in Strasbourg, Haase in Darmstadt and Binnemans in Leuven, we first studied structure/property relationships in complexes of salicylaldimines [15-22] and established that, for example, the anion had a crucial rôle to play in the mesophase transition temperatures, [17, 18] while the metal could have an important influence on the mesomorphic range.  Similarly, the structure of the ligand was found to be a crucial factor. 
We also looked at salicylaldimine complexes derived from ethylenediamine and tren and found that here, we were able to generate neutral, mesomorphic complexes which did not possess a bound counter-anion.  The structure of one such model lanthanum complex is shown below.
Some simple lyotropic derivatives have also been realised.  Yu G Galyametdinov, G Ivanova, I V Ovchinnikov, A Prosvirin, D Guillon, B Heinrich, D A Dunmur and D W Bruce, Liq Cryst, 1996, 20, 831; I Bikchantaev, Yu G Galyametdinov, O Kharitonova, I Ovchinnikov, D W Bruce, D A Dunmur, D Guillon and B Heinrich, Liq Cryst, 1996, 20, 489.
The structure of the basic alkoxystilbazole ligand which we have used is shown below.
We have worked with this ligand since our interest in liquid crystals began and an overview of these studies, which includes work on optically non-linear materials and on Langmuir-Blodgett systems, has appeared. 
The general structure of these complexes is shown below and comprises two stilbazoles bound in a linear fashion about a silver(I) centre, with an anion to maintain charge neutrality. Our work with these materials has recently been collected into a review article.  Of greatest initial interest were the materials in which the anion was dodecylsulphate (C12H25OSO3), for there we found materials with nematic, SmA, SmC and cubic phases. The symmetry of the cubic phase was determined by both X-ray diffraction  and freeze-fracture electron microscopy. 
We also studied polycatenar derivatives and found that while the mesomorphism was rather similar to that of conventional polycatenar mesogens,  there were some subtle variations due to both the physical presence of the anion and to the intermolecular electrostatic interactions which it generated. A combination of X-ray diffraction and dilatometry allowed us to propose a mechanism for the columnar-to-cubic transition in polycatenar systems. 
Current studies concentrate on the mesomorphism of isomeric ligand derivatives as well as on lyotropic behaviour. 
 D W Bruce, Adv Inorg Chem, 2001, 52, 151.
 D W Bruce, Acc Chem Res, 2000, 33, 831.
 D W Bruce, B Donnio, S A Hudson, A-M Levelut, S Megtert, D Petermann and M Veber, J Phys II France, 1995, 5, 289.
 B Donnio, D W Bruce, H Delacroix and T Gulik-Krzywicki, Liq Cryst, 1997, 23, 147,
 See e.g. H-T Nguyen, C Destrade and J Malthête, Adv Mater, 1997, 9, 375 for an overview of the mesomorphism in polycatenar mesogens.
 B Donnio, D W Bruce, B Heinrich, D Guillon, H Delacroix and T Gulik-Krzywicki, Chem Mater, 1997, 9, 2951.
 A I Smirnova and D W Bruce, Chem Commun, 2002, 176.
Stilbazoles will also bind to Pd(II) and Pt(II) to form trans square planar complexes as shown below:
Here, structure/property relationships have concentrated on polycatenar derivatives  and the observation that when the anionic ligand is changed from Cl to an alkanoate (h1-CnH2n+1CO2), only nematic phases are seen.  There are interesting comparisons to be made between these systems and the analogous silver complexes above in order to understand what drives the mesomorphism. [33, 34] B Donnio and D W Bruce, J Chem Soc, Dalton Trans, 1997, 2745.
Polycatenar liquid crystals are fascinating materials. In general, they are constructed from long, rod-like cores which possess between three and six terminal chains, often but not always arranged symmetrically. Two common (generalised) examples are shown below (a 'tetracatenar' and a 'hexacatenar' system):
The liquid crystal properties of these materials are diverse and, for example, hexacatenar materials will typically show a columnar hexagonal phase and/or a columnar rectangular phase. Discussion of the arrangement of the molecules in these phases may be found in reference  and papers cited therein.
Of special interest are tetracatenar mesogens of the type shown above, for at short chain length they show nematic and/or smectic C phases, while at longer chain lengths, they show columnar phases. Thus, simply by varying the chain length, the mesomorphism can change from that typical of calamitic materials to that characteristic of discotic mesogens. It is the change from calamitic to discotic behaviour which provides a particular fascination in this class of materials - how do the layers suddenly arrange into columns? This issue is addressed in references  and . For example, in the bipyridines shown above, the transition is through the intermediacy of a cubic phase, while in the polycatenar Pd stilbazoles in the preceding section (X = Cl), it is via a new lamellar phase. However, in the case of the polycatenar phenanthrolines shown below, the change is abrupt and without the intermediacy of any other phase. 
This interest in the calamitic/columnar crossover then led to an interest in the factors which influence the formation of the cubic phase. This is, to a large degree, what drove the synthesis of the large range of tetracatenar metal complexes which we have to hand as this enabled us to carry out a systematic structural comparison. The findings are argued in references [33-35].
Further, with collaborators in Strasbourg, Halle and Tours, we have studied some of our hexacatenar palladium stilbazole complexes in detail using X-ray diffraction and dilatometry, and have been able to interpret the data to propose a generalised model for the packing of polycatener species in columnar phases. Two single crystal structures are shown below.
We have also worked with hydrogen-bonded liquid crystals, in particular developing systems where stilbazoles are hydrogen-bonded to phenols.  Other work, some in collaboration with Professor Takashi Kato in Tokyo, is collected in reference .
Using one of these systems, namely stilbazoles hydrogen-bonded to 2,4-dinitrophenol, we were able to investigate the position of the hydrogen-bonded proton as a function of temperature and found that, while a neutral hydrogen bond existed in the solid state, as the temperature increased through the SmA mesophase, it transferred (double-well potential) from the oxygen to the nitrogen to give an ionic hydrogen bond. 
In addition to the hydrogen bond, there exists an analogous non-covalent interaction known as the halogen bond - the analogy is shown in the figure below and much systematic work was carried out by Tony Legon.
Resnati  has used halogen bonding to construct a variety of molecular solids using various pyridines and electron-poor iodobenzenes
We  have now shown that halogen bonding may be used to induce liquid crystal phase formation from non-mesomorphic components by allowing a non-covalent interaction to form between an alkoxystilbazole and iodopentaflurobenzene:
Reproduced with permission from J. Am. Chem. Soc., 2004, 126, 16. Copyright Am. Chem. Soc. 2004. A C Legon, Angew Chem Int Ed, 1999, 38, 2687; A C Legon, Chem Eur J, 1998, 4, 1980.
An abiding interest in the group is the search for the biaxial nematic phase (representation below) in low molar mass thermotropic mesogens, where so far there has not been an unequivocal demonstration that the phase exists. An overview of synthetic approaches to such systems in currently in press . We have looked at the use of 5,15-disubstituted metalloporphyrins as materials containing the features of both rod and disk  and while were successful in generating nematic and smectic materials (showing the overall rod-like nature of the materials),  low-melting materials were difficult to come by. 
Current interests surround the use of Pd(II) complexes as the basis for materials where we seek enhanced lateral correlations  following the simulations by Berardi and Zannoni: 
We are also looking into covalently linked rod-disk compounds which we have shown to possess substantial molecular biaxiality. 
Interestingly, the same rod-disk compound is able to act as a 'shape amphiphile' in that it can homogenise an immiscible mixture formed from the separate rod- and disk-like components. 
This work, carried out in collaboration with Professor Martin Schröder at the University of Nottingham, has looked at the generation of liquid-crystalline materials from functionalised macrocycles. In one aspect of the work, we have studied functionalised thioether macrocycles such as the one shown below. The three rings on each side are the minimum necessary to generate a mesophase in the free macrocycle as the thioether ring acts as a flexible spacer and so effectively 'decouples' the anisotropy of each side.  This 'linked' anisotropy is then restored when the thioether macrocycle is complexed. 
Smaller rings may also be functionalised and we found that both calamitic and polycatenar mesogens could be generated by N-functionalisation of aneNS2 macrocycles which are also illustrated below. 
aneN3 and aneN3 macrocycles may be monofunctionalised starting from an orthoamide precursor, and we found that using an alkoxybenzyl function, the macrocycle containing the residual bridge showed lyotropic mesophases (vide infra) in water. The single crystal structure of the parent aneN3 material is shown below.
While not macrocyclic, we also collaborate with Martin Schröder (and the Strasbourg group) on mesomorphic complexes of divalent metals derived from imine ligands. For example, MII complexes of the 2,6-disubstituted pyridine below shows a range of columnar mesophases. 
When an amphiphilic material is dissolved in a solvent such as water, the above a particular concentration, the critical micelle concentration (cmc), micellar aggregates form driven by entropic effects. Above the cmc, addition of further amphiphile simply increases the concentration of micelles until their concentration becomes so high that the micelles order. These ordered phases of micelles are the so-called lyotropic liquid crystal phases. Below is shown a schematic structure of a simple amphiphile, in which one would expect to find a polar organic headgroup, typically a cationic alkylammonium cation, or perhaps a carboxylate anion, or even a neutral oligo(ethylene oxide). Alongside is shown a schematic for a metallosurfactant in which the headgroup is now a metal complex.
Using this approach, we have looked at surfactants based on, for example, simple ethylenediamine derivatives of Cr(III) and Co(III).  However, , most of our work has centred on amphiphilic derivatives of tris(bipyridine)ruthenium(II) and we looked at four systems, namely 4- and 5- monoalkyl and symmetric 4,4'- and 5,5'-dialkyl derivatives:
We have looked at the basic surfactant properties with Dr James Bowers of this Department of these complexes and have found, for example, that while the 5,5-dialkyl materials do form micelles,  they are very slow in forming surface monolayers, leading to misleading results when cmc is measured by surface tension. 
At higher concentrations, lyotropic phases form from these materials and, as is predicted, monoalkyl surfactants form I1 cubic phases, while dialkyl materials form H1 hexagonal phases.  Mesophases can also be realised using surfactant derivatives of [M(terpy)2]n+ (M = Ru, n = 2; M = Rh, n = 3). 
A recent development of this work has been the use of the Ru surfactants as templates for the formation of mesoporous silicates. We use the so-called true liquid crystal templating approach developed by Attard  and find that after calcination, we generate a hexagonal mesoporous silicate bearing RuO2 particles which are, on reduction, active catalysts for the selective reduction of 1-hexene.  Control of the calcination conditions leads to very high-activity catalysts. 
TEM image from Dr Wuzong Zhou (St Andrews)
 D W Bruce, I R Denby, G J T Tiddy and J M Watkins, J Mater Chem, 1993, 3, 911.
 J Bowers, M J Danks, D W Bruce and R K Heenan, Langmuir, 2003, 19, 292.
 J Bowers, M J Danks, D W Bruce and J R P Webster, Langmuir, 2003, 19, 299.
 D W Bruce, J D Holbrey, A R Tajbakhsh and G J T Tiddy, J Mater Chem, 1993, 3, 905.
 J D Holbrey, D W Bruce and G J T Tiddy, J Chem Soc, Dalton Trans, 1995, 1769.
 G S Attard, J C Glyde and C G Göltner, Nature (London), 1995, 378, 366.
 H B Jervis, M Raimondi, R Raja, T Maschmeyer, J M Seddon and D W Bruce, Chem Commun, 1999, 2031.
 M J Danks, H B Jervis, M Nowotny, W Zhou, T A Maschmeyer and D W Bruce, Catal Lett, 2002, 82, 95
The LiqCryst Database is a remarkable and unique database containing in excess of 70,000 compounds.