摘要:

J O U R N A L O F C H E M I S T R Y Materials Feature Article Structure and reactivity of oxide surfaces: new perspectives from scanning tunnelling microscopy Russell G. Egdell and Frances H. Jones Inorganic Chemistry L aboratory, South Parks Road, Oxford, UK OX1 3QR Scanning tunnelling microscopy (STM) has emerged in the past few years as a uniquely powerful tool for the investigation of oxide surfaces.STM is capable of real-space imaging of periodic structures with atomic resolution and of characterising local atomic arrangements associated with defect sites. The power of the technique is further enhanced by the ability to probe filled and empty density of states profiles at specific atomic sites by monitoring variations in tunnelling currents with applied voltage. The interpretation of STM images is not always unambiguous, a question of central importance being whether oxygen or metal ions appear as maxima in the images. The necessity for adequate sample conductivity imposes some constraints on the applicability of the technique, although developments in instrumentation and in techniques for sample preparation are helping to overcome these limitations. The range and limitations of the technique are illustrated by reference to work on tungsten oxides, titanium dioxide and iron oxides.The review concludes with a discussion of recent developments in the study of molecular adsorbates. published in 1994 presented only two images where anything 1 Introduction approaching atomic resolution had been achieved on an oxide There has been a huge upsurge in interest over the past few surface and concluded that although ‘the results are encouragyears in the surface properties of metal oxides.This has been ing they have not yet been of real value in a surface structure driven in part by the technological importance of these mate- determination on oxides.’ An exception to this statement was rials in areas such as catalysis and gas sensing, where surface the acquisition before 1990 of atomically resolved images from properties are of crucial importance, and in part by a continu- cleavage surfaces of layered oxide superconductors, particularly ing fascination with the diverse range of structural, chemical Bi2Sr2CaCu2O85–7 and Tl2Ba2CaCu2O8.8 However as will be and electronic properties exhibited by oxides.One major aim seen below these are very ‘easy’ surfaces to work with because of academic surface science is to characterise surface structure of the weak interlayer bonding and the inertness of cleavage at an atomic level and to develop an understanding of the surfaces. The application of STM to oxide superconductors ways in which molecules interact with specific surface sites.has already been reviewed9 and in the present contribution we Most technological applications involve polycrystalline mate- focus attention on materials other than the cuprates. Within rial, but the starting point for tackling the objectives outlined the past four years many of the apparent diYculties in studying above usually lies in the study of well defined single crystal these surfaces have been overcome and there is now a rapidly surfaces under ultrahigh vacuum (UHV) conditions.1,2 The growing body of work where STM has provided definitive intrinsic complexity of oxides is reflected in the fact that even information about surface structure.now only a handful of these idealised surface structures have The organisation of this review paper is as follows.We first been solved by diVraction techniques using electrons or X- present a general introduction to the technique of STM. Next rays. Moreover, there is a general consensus that the chemical we address the important question of interpretation of STM reactivity of oxide surfaces is probably dominated in many images, touching on the interplay between structural and cases by defects.DiVraction techniques oVer the prospect of electronic eVects which is of particular importance when rigorous determination of periodic surface structures, but are dealing with oxides. Experimental problems associated with unsuited to investigation of local structure associated with sample conductivity are then assessed, followed by a discussion defect sites.Scanning tunnelling microscopy (STM) allows of the merits of carrying out STM measurements in UHV imaging of both periodic and defect structures with atomic rather than under ambient conditions. The ideas introduced in the earlier sections are illustrated by reference to work on resolution. Moreover, by measuring the variation of tunnelling the tungsten oxides WO310–14 and NaxWO3,15–17 on the rutile current with applied voltage it is possible to obtain a measureform of TiO2,18–36 and on the iron oxides Fe2O337,38 and ment of filled and empty densities of states.STM therefore Fe3O4.39–44 Finally some recent developments in the study of seems to be the ideal tool for tackling many of the outstanding adsorbates on oxide surfaces are highlighted.45–47 Here there problems in oxide surface science.are clear indications that STM is not only a tool for investi- STM made its debut in 1982 by providing a real space gation of static surface structures, but that it is also able to solution to the surface structure of Si(111) (7×7). Extension give a real time view of dynamic processes, such as molecular of the technique to other semiconductor surfaces and to metal decomposition and diVusion.surfaces followed in rapid succession and by about 1990 it had become a routine matter to obtain high quality images on surfaces of this type. Related techniques using the technology of STM also appeared, notably atomic force microscopy 2 The technique of STM (AFM). In the 1990s the field has settled into a period of 2.1 Background maturity with established series of national and international conferences devoted to scanning probe microscopy and with STM relies on quantum mechanical tunnelling. An atomically the appearance of several textbooks.3,4 sharp tip is brought within a few A ° ngstroms of a surface using Against this background, the application of STM to oxide an arrangement of piezoelectric transducers and a voltage is surfaces developed very slowly in the initial years.Henrich applied between the tip and the sample: here we adopt the usual convention of quoting the bias of the sample relative to and Cox’s definitive monograph on oxide surface science1 J. Mater. Chem., 1998, 8(3), 469–484 469the tip. At small separations the wavefunctions for electron states at the surface overlap with the wavefunctions associated with the tip and electrons may tunnel into or out of the surface, depending on the bias.The tunnelling current shows an exponential variation with tip–surface separation because of the exponential decay of the wavefunctions into the vacuum. The tip is moved parallel to the surface in the xy plane whilst the z position (i.e.the distance from the surface) is controlled. In most cases the xy scanning involves a sawtooth variation in x with progressive increments in y. There are two principal modes of image acquisition. In the constant height mode the tunnel current is simply measured as a function of xy coordinates. In the alternative constant current mode a negative feedback is applied to the z piezo to keep the tunnel current constant and the z displacement is monitored as a function of Fig. 1 Two sequentially acquired 25 A ° ×25 A ° images from the (4×1) xy coordinates. The constant height mode allows for more reconstructed surface of SnO2(110), both with 1 nA tunnel current. rapid scanning across the surface but there is a significantly The left hand image is taken at +1.5 V sample bias, the right hand at larger risk of ‘tip crash’ on surfaces containing irregular +1.0 V bias.The same unit cell is highlighted in each image. Note the protrusions or large numbers of steps. Constant current imag- reversal of topographic contrast. A third image taken at +1.5 V bias was very similar to the first, indicating that there had been no tip ing is therefore more usual.The image is built up from a series change. Adapted from ref. 48. of corrugation profiles, i.e. plots of z displacement as a function of x displacement. These are finally transformed into greyscale images which represent maximum outward z displacement as when the tip is deliberately dipped into an oxygen-rich area white and maximum inward displacement as black. These of the surface (allowing transfer of O atoms onto the tip), the images may in turn be enhanced by various filtering techniques, adatoms are imaged as minima (Fig. 2). application of alternative colour schemes and by transformation into ‘false’ 3D views which usually overemphasise the z 2.3 Models of STM corrugation. Most commercial software allows the bias voltage Most qualitative interpretations of STM images are based to be changed between the forward x sweep and the reverse, implicitly on the TersoV–Hamann model.50 Here the tip elec- so that two images of exactly the same area of the sample can tronic wavefunctions are assumed to be described by spheri- be acquired synchronously at diVerent bias.In addition the cally symmetrical s-waves. For small sample bias it may be feedback loop may be disengaged at each pixel point in the shown that the tunnelling current I is given by: image and a current–voltage (I–V ) curve measured at that point.It may be shown that the normalised diVerential con- I3.s |Ys(r0)|2d(Es-Ef) (1) ductance (dI/dV )/(I/V ) is directly proportional to the filled or empty density of states. Thus STM complements photowhere the summation is taken over amplitudes of sample emission spectroscopy (PES) and inverse photoemission wavefunctions s at the centre of the tip whose coordinates are spectroscopy (IPES) in providing atomically resolved elecspecified by r0 and the subscripts refer to the energy of tronic structure information. However, the spatial characteriselectronic states in the sample (s) relative to the Fermi energy ation of electronic states is gained at the expense of information of the tip (f ).By definition the summation corresponds to the about the crystal momentum, which is provided by the E–k local density of sample electronic states at the centre of dispersion curves derived from PES and IPES. curvature of the tip and therefore the constant current images correspond to contours of constant density of sample electronic 2.2 General considerations in the interpretation of STM images states.Taking explicit account of the decay of the sample and of oxides tip wavefunctions into the tunnelling gap we have: A typical STM image of an oxide surface will contain an I3.s |Ys|2 exp-[2k(R+s)]d(Es-Ef) (2) ordered array of greyscale maxima and minima, usually interspersed with steps, dislocations and other defects.It is rarely the case that all ions within the surface unit cell show up as where R is the tip radius and s is the tip to sample separation. The decay constant k is given by: greyscale maxima. A major issue of interpretation is therefore to establish the correspondence between maxima in greyscale k=Ó(8p2mw/h2) (3) images and atomic positions within the unit cell.In general oxygen anions and metal cations may have the same surface where m is the electron mass and w is the local surface barrier. If s is in A ° ngstroms and w is in eV, k can be expressed in a periodicity, so the corrugation profiles do not in themselves solve this problem in all cases and there is therefore scope for simple numerical form: controversy! The problems of interpretation are illustrated by 2k=1.025Ów (4) some experiments we have carried out48 on a (4×1) reconstruction on SnO2(110) (Fig. 1). Here features which appear as For very large sample–tip separations, the surface barrier would be expected to be simply the average of sample and tip depressions in the+1.5 V image transform into bright maxima in the image taken at lower sample bias.This implies that the workfunctions. However, at the separations appropriate to most STM experiments, the slow decay of exchange correlation images cannot be explained solely in terms of topography and consideration of electronic structure is crucial. An additional potentials causes w to be much less than the workfunction. In fact w can be determined experimentally from the relationship: complication is that image contrast may depend on the state of the tip.In the case of the simple adsorbate system of O on w=0.952(d lnI/ds)2 (5) Cu(110) the image contrast may be totally reversed by the deliberate adsorption of an oxygen atom onto the tip.49 Thus One reported value of 1.6 eV for the oxide surfaces Na0.82WO3(001)15 is much less than typical oxide at +0.03 V sample bias and 1 nA tunnel current oxygen adatoms appear as ‘Mexican hat’ maxima when imaged with workfunctions of the order 4–5 eV.There are already many approximations at this stage. Most a tip that has been rigorously cleaned by field emission, but 470 J. Mater. Chem., 1998, 8(3), 469–484Fig. 2 (a) 50A° ×50 A° STM image of Cu(110) exposed to oxygen and then cooled to 4 K.Image taken with a cleaned Pt/Ir tip at 0.03 V sample bias and 1 nA tunnel current. Rows of the O induced (2×1) reconstruction are apparent, along with an isolated O atom that appears as a ‘Mexican hat’ in the corrugation profile shown under the image in the central panel. (b) 65A ° ×65 A ° image now taken with an O terminated tip. The four isolated O atoms now appear as greyscale minima.Reproduced with permission from ref. 49. notably, STM experiments almost invariably use W or Pt/Ir various heights above a TiO2(110) surface in an attempt to understand experimental empty state images from that surface. tips and if these do indeed terminate with a single transition metal atom, the tip wavefunctions will certainly not be a s In a similar spirit, Diebold et al.30 computed contour plots of empty state charge density for TiO2(110) using a full plane wave. For example, model calculations on a cluster providing a realistic model of a W tip revealed a strong 5dz2 resonance wave pseudo-potential method implemented within the local density approximation.These calculations are discussed in near the Fermi level.51 In general predominance of directional tip states of this sort will lead to greatly enhanced surface Section 3.A limitation in this approach is that no explicit account can be taken of the influence of tip structure on corrugation. Notwithstanding these diYculties eqn. (2) can be cast into STM images. A potentially important recent development in interpretation a very simple form15 that allows rudimentary qualitative interpretation of STM images from compound materials such of STM images of oxides has been direct calculation of tunnelling currents using the so-called electron scattering quan- as oxides: tum chemistry (ESQC) approach.Here the tip, surface and gap are treated as an impurity in an otherwise perfect conduc- I(x,y,z)3.i Di exp-(1.025siÓw) (6) tor and both the tip and surface can act as sources and sinks of electrons.The tip is modelled explicitly as a well defined Here the current I for a tip at position (x,y,z) is a summation cluster and impurity atoms may be added to the tip at will. over contributions from atoms i on the surface in the proximity This model has been used with notable success to treat of the tip.The atoms are assumed to be hard spheres separated adsorbed C6H6 molecules on Rh(111)52 and ordered layers of from the surface of the hard sphere tip by distances si and the S on Re(0001).53 In the latter system adsorption of a single S parameter Di gives a measure of the relative contribution of atom onto a single Pt (1-Pt) or 3-Pt terminated tip led to each atom to the total density of electronic states at the energy contrast reversal for S atoms adsorbed at on-top positions.appropriate to the tunnelling conditions which are being used. The model has recently been extended to treat an FeO(111) Eqn. (6) tells us that when performing constant current imaging layer on Pt(111).54 Here it is found that topographic maxima of a flat surface containing atoms of two types, say j and k, a in the images occur over O positions for Pt terminated tips density of states ratio Dj/Dk=0.1 will translate into an apparent but over Fe positions for O terminated tips.topographic height diVerence between the two atoms such that sj-sk=1.77 A ° . The highly ionic nature of oxides means that density of states ratios of this order are not fanciful and 3 Experimental aspects of STM electronic structure eVects will exert a major influence on topographic images. 3.1 Sample conductivity: which oxides can be studied and which A more direct approach to dealing with electronic structure cannot? eVects is to evaluate the square of sample wavefunctions above the surface [the ys2 of eqn. (2)] in an explicit fashion. This The tunnel current flowing in an STM is typically of the order 0.1–20 nA.There is an obvious requirement that the sample approach has been developed by Gulseren et al.31 using a firstprinciples atomic-orbital based scheme that was benchmarked should have suYcient conductivity to allow passage of the tunnel current. Experience with a number of oxides indicates against more rigorous ab initio plane wave pseudo-potential calculations. In particular they calculated charge densities in that a conductivity of the order of 1 V-1 cm-1 is necessary to allow trouble free imaging.conduction band states up to 1.5 eV above the Fermi level at J. Mater. Chem., 1998, 8(3), 469–484 471The conductivity requirement is obviously satisfied by cleavage in UHV, possibly at low temperature to minimise oxygen loss.This type of sample preparation has proved metallic oxides. Those that have been imaged with atomic resolution include a number of high temperature oxide necessary to obtain reliable photoemission spectra of conducting cuprate phases such as YBa2Cu3O7 and La2CuO4 that superconductors such as Bi2Sr2CaCu2O8-x and its substituted derivatives,5–7,55–61 Tl2Ba2Ca1+nCu2+nO8+2n,8,62,63 are less robust than Bi2Sr2CaCu2O8, and low temperature cleavage in microscopes with cryogenic capabilities has allowed Pb2Sr2(Y,Ca)Cu3O864 and YBa2Cu3O7,65–69 ReO3,70 WO271 and sodium tungsten bronzes15–17 NaxWO3 together with true atomic imaging of YBa2Cu3O7.68,69 If the problem of sample conductivity of bulk oxide crystals oxide bronzes such as the blue bronzes M0.3MoO3,72–77 and Fe3O4 whose room temperature conductivity is very high if at room temperature proves intractable, it may be possible to carry out the STM measurements at elevated temperatures not truly metallic.39–43 For materials which are metallic, filled electronic states below the Fermi level can be accessed at where the conductivity is enhanced. There are obvious problems of thermal drift at high temperature, but a new generation positive sample bias and empty states at negative sample bias, so that STM imaging should be possible at either positive or of commercially available STM instruments allows fairly routine measurements at temperatures up to 1500 K.High negative sample bias. At the other extreme, low conductivity rules out the study temperature atomic scale imaging has been achieved on a cleaved NiO(100) surface at 473 K (Fig. 3);90 and on of important wide bandgap main group oxides such as MgO and Al2O3, although even materials of this sort can be studied UO2(111)91 and UO2(110)92 at 573 K, both oxides having insuYcient room temperature conductivity to allow stable at very low resolution if decorated with a thin conducting layer of a metal such as Au.78 So far as we are aware there tunnelling.Another alternative is to completely circumvent the problem have been no studies to date on bulk oxide crystals whose conductivity is dominated by ion transport and which have of conductivity by growing a thin epitaxial oxide layer on a metal substrate. Freund et al.2 have provided a recent definitive low electronic conductivity (e.g.Y or Ca doped ZrO2). Maximal oxidation state transition metal oxides with a d0 review of this area and identify three distinct approaches. The first involves simple oxidation of a metal. For example, configuration typically have bandgaps in excess of 2 eV and are intrinsically poor conductors if truly stoichiometric. NiO(100) can be grown by oxidation of Ni(100),93 and Cr2O3(0001) on Cr(110).94 The oxide layers are often highly However, annealing in UHV or a partial pressure of hydrogen induces bulk oxygen loss.The resulting d1 donor levels usually strained owing to mismatch with the metal. However, atomically resolved STM has been performed on Cr2O3 on Cr(110)95 sit close to the conduction band minimum so that these materials become extrinsic n-type semiconductors.UHV and on NiO(100) grown on Ni(100).93 The second approach involves evaporation of a metal onto a metal substrate, fol- annealed TiO2,18–36 WO3,10–14 SrTiO379–82 and BaTiO383 all acquire suYcient conductivity to be studied by UHV STM. lowed by oxidation of the deposit: the metals need not be the same and therefore there is a possibility of lattice matching The number of filled donor states is usually small in relation to the number of empty conduction band states.Thus it is the target oxide to the metal substrate. Oxide surfaces prepared in this way and subsequently studied by STM include NiO(111) possible to image these surface by tunnelling into empty states at negative sample bias, but filled state imaging is more diYcult on Au(111),97 FeO(111) on Pt (111),98 TiO2(001) and MgO on Mo(100),99 and Al2O3(0001) on Re(0001).98 In the last of unless the bias is suYcient to access filled O 2p levels.An alternative approach to increasing conductivity that has been these studies the oxide layers were used as substrates for deposition of Au, Pd and Ni metal particles in an attempt to used with V2O5 is to dope with a very small concentration of an interstitial alkali metal during crystal growth,84,85 although prepare model supported metal catalysts.98 Thirdly and finally, oxide layers can be grown by oxidation of alloys.One system it should be noted that even the undoped material will sustain tunnelling in air.86 Likewise the intergrowth bronze Rb0.03WO3 of particular note in this context is oxidised NiAl(110), which yields very well ordered Al2O3 films two atomic layers has been imaged.87 Finally for suYciently high levels of reduction the oxygen vacancies in d0 transition metal oxides thick.99,100 The symmetry of the overlayer in relation to the substrate allows the existence of domains whose basis vectors will arrange themselves into ordered crystallographic shear planes giving rise to distinct new Magne`li-type phases.These are rotated by 24° relative to each other. The periodicity within the overlayer is commensurate with the NiAl(110) substrate may be metallic or have very low activation energies for conduction and most will probably be amenable to STM along the (11 : 0) direction but incommensurate along (001).This gives rise to antiphase domain boundaries within the investigation. In some recent elegant experimental work Rohrer and coworkers have followed the evolution of surface structure rotational domains. These complex surfaces have proved amenable to high quality STM imaging, which at low sample biases in the series MoO3, Mo18O52, Mo8O23, Mo4O11, although the parent insulating MoO3 was studied by AFM rather than of the order of 30 meV reveals atomic structure within the surface unit cell.100 The oxide overlayers have been used in STM.88,89 Binary or ternary dn oxides of the later first row transition turn as substrates for deposition of Pd, Pt and Rh.99,101,102 At elements are usually Mott–Hubbard or charge transfer insulators with a localised dn configuration.Comparatively few of these materials (again excluding the special case of oxide superconductors and their parent phases) have been studied by STM, one notable exception being Fe2O3 which under UHV has a propensity to lose oxygen to give Fe3O4 or even FeO-like surfaces.37,38 In a chemical sense the conductivity is again n-type. Marginally conducting p-type oxides pose a greater technical challenge.Mobile holes are introduced by oxidation of the system by introduction of interstitial oxygen or cation vacancies, or by substitution of a countercation A in ternary oxides such as AMO3 by a countercation with charge one less than A. UHV annealing will always tend to induce oxygen loss and thus to decrease the room temperature Fig. 3 Empty state STM image (+0.7 V sample bias, 1.0 nA tunnel conductivity.In fact some of the superconducting cuprate current) of NiO(001) taken at about 473 K. The square array of phases will even lose oxygen at room temperature. The optimal greyscale maxima are believed to correspond to the Ni positions. Reproduced with permission from ref. 90. sample preparation for materials of this sort therefore involves 472 J.Mater. Chem., 1998, 8(3), 469–48490 K Rh nucleates randomly on the Al2O3, but at 300 K there topographic features with similar corrugation.89 It was also demonstrated that it was possible to obtain air STM images is preferred nucleation at the antiphase domain boundaries.102 Whilst this is an interesting observation in itself, it does for Mo8O23 and Mo4O11 which have 3D structures derived from the ReO3 structure rather than from the layer structure highlight the point that non-commensurate oxide overlayers may show subtle diVerences in their surface properties from of MoO3.An air stable (6×2) reconstruction has also been identified on the cubic perovskite SrTiO3.82 Nonetheless it is true bulk oxide surfaces which do not exhibit these substrate induced defects. remarkable that atomically resolved STM images have apparently been obtained in air from YBa2Cu3O7,65,66 given that this material reacts rapidly on a macroscopic scale with both 3.2 UHV versus ambient conditions CO2 and H2O.In this case it impossible to believe that the Model studies of oxide single crystal surfaces using electron image could arise from true tunnelling across an air gap to an based techniques such as X-ray and UV photoemission, low atomically clean surface.energy electron diVraction (LEED) and electron energy loss spectroscopy (EELS) are invariably carried out under UHV 4 STM of tungsten oxides conditions. Vacuum conditions are necessary to prevent interference of gas phase molecules with electrons in their passage 4.1 Introduction to WO3 and NaxWO3 through the electron optics, but the more stringent need for ultrahigh vacuum arises mainly from the necessity to maintain Tungsten trioxide (WO3) is a 5d0 transition metal oxide.The idealised structure is based on a cubic ReO3-like framework surface cleanliness. On the other hand the technique of STM does not in itself impose requirements as to the medium in of corner sharing WO6 octahedra.However, tilting and distortion of the WO6 octahedra results in deviations from the which it is applied. STM studies in both UHV and under ambient conditions therefore become a possibility, provided ideally cubic structure. At room temperature the bulk structure is monoclinic, with lattice parameters that essentially represent that sample cleanliness can be maintained in the latter.Oxides with a layer structure held together by weak van der a 2×2×2 superstructure on an idealised cubic unit cell of 3.7 A ° . However, for the present purposes we can treat WO3 as Waals forces usually present chemically inert cleavage planes; for example, water only weakly physisorbs onto the (001) if it were cubic. The bandgap is about 2.6 eV at 300 K, but the bulk solid state properties are determined by the propensity surface of Bi2Sr2CaCu2O8.Layered materials are therefore particularly favourable prospects for air STM and indeed the of the material to become oxygen deficient WO3-x. For O vacancy concentrations in excess of x=1×10-4 point defects quality of atomically resolved images obtained under ambient conditions from materials including Bi2Sr2CaCu2O8,7,55,58,61 are eliminated by the formation of well defined shear planes consisting of edge sharing WO6 octahedra running along the Tl2Ba2Ca2Cu3O108,9,62,63 and Pb2Sr2(Ca,Y)Cu3O864 are comparable to those obtained under UHV.Rohrer and coworkers85 <1m0> directions. The WO3 structure is based on a sequence of ionic planes have extended atomically resolved ambient STM and AFM to the non-cuprate oxides V2O5 and V6O13.To minimise water with stoichiometries {O}–{WO2}–{O} and formal ionic charges {2-}–{2+}–{2-} (Fig. 4). However, a repeating adsorption it is also possible to carry out the STM experiment in a closed environment glove box, although even if H2O and dipolar sequence such as this normal to a surface gives rise to an infinite surface energy.This precludes termination of WO3 CO2 levels are reduced to the ppm level oxide surfaces will suVer huge exposures to these gases by the normal standards in either a {O} or a {WO2} surface plane. However, if the final WO2 plane is covered in half a monolayer of oxygen, the of UHV experiments. The mechanism of imaging layered materials in air has sequence {O0.5}–{WO2}–{O} has formal ionic charges {1-}–(2+}–{2-}which can be bracketed into the repeating aroused considerable controversy and in some cases at least it is probable that the tip makes direct contact with the sample.quadrupolar sequence ({1-}–{2+}–{1-})–({1-}–{2+}– {1-}) with the 2- charge of the subsurface {O} layers split The apparent topographic contrast arises from ‘sliding’ of a raft of material across the surface with Moire type modulation between two quadrupolar units.This sequence avoids a divergent surface energy. Two possible simple arrangements of half of the conductivity as the top layer falls in and out of registry with the underlying layers.103 However, a recent study by a monolayer of oxygen ions on top of a WO2 layer can be envisaged.A situation in which oxygen ions are present on Smith and Rohrer has demonstrated that both air and UHV STM images of the layered material Mo18O52 reveal similar alternate W ions along the [100] and [010] directions will Fig. 4 The idealised cubic ReO3 structure of WO3, shown in terms of a sequence of charged ionic planes. The termination on the left has a repeating dipole normal to the surface and an infinite surface energy.Removal of half a monolayer of on-top O allows termination with a repeating quadrupolar charge sequence which no longer has an infinite energy. J. Mater. Chem., 1998, 8(3), 469–484 473Fig. 5 The cubic perovskite structure of NaxWO3, showing two possible terminations for the (001) surface .The lattice parameter corresponds to x=0.665. give rise to a (Ó2×Ó2)R45° reconstruction. Alternatively, the presence of oxygen ions on alternate rows of tungsten ions along [100] or [010] directions will give a (2×1) or (1×2) reconstruction. The energies associated with the two alternative Fig. 6 130 A° ×130 A° STM image of WO3(001) acquired at 1 nA tunnel superstructures have been calculated using atomistic modelling current and+1.5 V sample bias.The (Ó2×Ó2)R45° unit cell is shown. techniques to be 1.39 J m-2 and 1.67 J m-2 respectively.11,12 Note the dark defect troughs running through the image. Adapted Thus a (Ó2×Ó2)R45° termination is to be anticipated. from ref. 11. Introduction of sodium atoms into the WO3 lattice results in a series of prototype oxide bronzes.These have the general formula NaxWO3 where 0.0<x<0.9. They are of particular seem to display a quasi-periodic distribution and are coninterest as their structural and electronic properties vary with strained approximately to the [100] or [010] directions, sodium content. For low values of x the sodium tungsten although the troughs meander somewhat in direction and have bronzes retain the monoclinic structure of WO3 and they are ragged edges.semiconducting. However, for x values greater than 0.43 they At an unrelaxed surface the on-top O ions sit about 1.9 A° adopt an essentially cubic perovskite structure, based on a above the W ions of the WO2 plane. Atomistic simulations cubic framework of corner sharing WO6 octahedra, with the suggest that the on-top oxygen ions relax inward and the sodium atoms occupying a fraction of the twelve-coordinate underlying W ions relax outward, shortening the bond length interstitial sites of the host lattice (Fig. 5). Again the structure by 0.5 A ° .12 The ‘bare’ W ions also relax inward by 0.19 A ° so can be described as a series of layers stacked along [001] with that the height diVerence between on-top O and ‘bare’ W is stoichiometries {NaxO}–{WO2}–{NaxO}–{WO2}. The Na 3s reduced to 1.6 A ° .The conduction band states in WO3 are of levels lie about 10 eV above the bottom of theW5d conduction dominantW5d atomic character but there is significant mixing band104 and each Na ion therefore donates one electron into with O 2p states, which becomes stronger the higher the energy the set of W 5d levels of local t2g symmetry.For x values in above the bottom of the conduction band. The bulk band excess of 0.26 the bronzes are metallic with conductivities structure calculations of Bullett105 on monoclinic WO3 give approaching that of copper in the cubic regime. In the cubic an O 2p/W 5d ratio of 0.17 for states 1.5 eV above the bottom metallic regime the occupied part of the conduction band is of the conduction band. Through eqn.(6) and assuming a about 1 eV wide and has a nearly free electron like shape. local barrier height of 2 eV one can estimate that this ratio Photoemission spectroscopy reveals the expected non-vangives a reduction in the apparent height diVerence between W ishing density of states at the Fermi energy. Early surface and O of 1.2 A ° .Thus in this case structural eVects should structural work on (001) surfaces of sodium tungsten bronzes dominate over electronic structure eVects and greyscale identified (2×1) and (3×1) reconstructions, which were attrimaxima correspond to the on-top oxygen. buted to sodium ordering. The periodicity was a function of Consider next the defect troughs. The apparent depth of the sodium content.troughs was generally found to be of the order of 2 A ° which Taken together WO3 and the metallic bronzes NaxWO3 are corresponds roughly to the bulk WMO bond length of 1.9 A ° . of particular interest in that they provide a pair of materials Spatially resolved measurements of dI/dV versus V curves on in which to explore the influence of a non-metal to metal a trough dissected surface are shown in Fig. 7.14 In contrast transition on the scope of STM. to the regular terraces, there is substantial tunnelling from filled states at biases above 1 eV when the tip is in the proximity 4.2 STM of WO3(001)10–14 of a defect trough. For this reason it is believed that defect troughs are associated with areas where the O0.5 on-top oxygen After annealing (001) oriented crystals of WO3 in low (10-5 mbar) partial pressures of oxygen a high degree of is missing to give termination in a bare WO2 plane. Electrical neutrality is maintained if the surface W ions are reduced to surface order is achieved, as gauged by LEED.Stable tunnelling can only be achieved at positive sample bias typically between WV to give ions with a localised 5d1 electron configuration.Localised electron states are observed in photoemission spectra +1.5 V and +2 V, as expected for an n-type semiconducting oxide with the Fermi level pinned close to the conduction from oxygen deficient WO3 surfaces at binding energies above about 1 eV. band minimum by d1 states associated with oxygen vacancies. 10,11 Typical images showed large flat terraced areas To further explore the eVects of oxygen deficiency, WO3(001) (Ó2×Ó2)R45° was deliberately modified by argon ion bom- supporting the expected (Ó2×Ó2)R45° reconstruction (Fig. 6). However, a striking feature of the images is that the terraces barding and then annealing in UHV.11,13,14 Argon ion bombardment of oxide surfaces is known to result in preferential are dissected by defect troughs running across them.These 474 J. Mater. Chem., 1998, 8(3), 469–484new reconstruction emerges with (1×1) periodicity.13,14 This is confined to bright strips apparently lying 2–3 A ° above the underlying terraces (Fig. 9). Similar structures were observed on surfaces that were prepared simply by annealing in vacuo for long periods at 700 °C.The (1×1) periodicity is clearly seen in high resolution empty state images of a vacuum annealed surface (Fig. 10). The arrangement of the atomic rows along [010] and [100] implies that the reconstruction corresponds either to a situation where all of the W ions in the WO2 plane are capped by on-top oxygen ions, or alternatively that all of the on-top oxygens have been removed and the maxima are the W ions in the bare WO2 layer.Since ion bombardment or high temperature annealing results in pro- Fig. 7 DiVerential conductance spectra for a WO3(001) surface displaying a c(2×2) terrace reconstruction dissected by defect troughs. Blue line: on terraces. Red line: on troughs. The tip was stabilised at 1 nA tunnel current at +2.0 V sample bias at each pixel point.The feedback loop was then disengaged to measure the I–V curves. Each curve is derived from 20 points within an image. Adapted from ref. 14. sputtering of oxygen from the surface. The first eVect observed in the STM was the formation of new line defects on terraces which otherwise retain the original periodicity. These new linear defects are narrower and shallower than the original troughs and adhere strictly to the [100] and [010] directions.With more prolonged reduction, the two types of defect structure are observed to merge on regions of the crystal surface producing striking branched patterns and the terraces now exhibit a p(2×2) periodicity (Fig. 8). This reconstruction is believed to correspond to a situation in which half the original on-top oxygen ions have been removed to leave a WO2 layer terminated by just a quarter of a monolayer of oxygen ions.This is necessarily accompanied by reduction of half the surface W cations from WVI to WV. Atomic scale filled state imaging was possible for the p(2×2) reconstruction. The Fig. 9 280 A ° ×280 A ° STM image of WO3(001) acquired at +1.5 V reduced WV ions in the (2×2) structure support localised filled sample bias and 1 nA tunnel current after several prolonged cycles of electronic states close to the Fermi energy, thus enabling filled ion bombardment and annealing.The bright raft running across the state imaging. image supports a (1×1) reconstruction. The darker areas support a p(2×2) reconstruction. Adapted from ref. 14. After further ion bombardment and anneal cycles, a further Fig. 8 Left hand panel: STM images after bombardment of WO3(001) by 500 eV argon ions at 6 mA for 60 min followed by UHV annealing at 700 °C overnight. +2.0 V sample bias, 1 nA tunnel current. Note branched defect structure. Right hand panel is a high resolution scan of typical terrace structure corresponding to image on left. The new p(2×2) unit cell is highlighted.Adapted from ref. 11. J. Mater. Chem., 1998, 8(3), 469–484 475Fig. 10 3D rendering of high resolution image of (1×1) reconstruction on WO3(001): 1 nA, +1.2 V sample bias. Adapted from ref. 13. gressive removal of oxygen ions, it is more likely that the image is that of bare W ions. This is supported by XPS data which show a progressive decrease in the O/W ratio as one moves through (Ó2×Ó2)R45°, p(2×2) and (1×1) reconstructions.Removal of the on-top oxygen ions from the WO2 layer leaves behind electrons, resulting in reduction of the tungsten ions from WVI to WV. Thus, although the ionic layers Fig. 11 (Top) filled state STM 25 A ° ×25 A ° image of 1×1 reconstruc- have stoichiometries {WO2}–{O}–{WO2}–{O}, the formal tion on WO3(001) taken at 1 nA tunnel current and -0.4 V sample charges on the planes will be {1+}–{2-}–{2+}–{2-} which bias.(Bottom) corrugation profiles along the [010] and [100] direccan be grouped into the quadrupolar sequence tions. Adapted from ref. 13. {(1+)–(2-)–(1+)}–{(1+)–(2-)–(1+)}. This avoids the divergent surface energy associated with a repeating dipole at between -2.0 V and +2.0 V.This is to be expected for a the surface. metallic material. Fig. 12 shows a pair of images taken at A further consequence of the reduction of the W ions on +0.6 V and -0.6 V sample bias acquired simultaneously by oxygen loss is the enhanced amenability of the (1×1) structure switching the voltage between forward and reverse travels of to filled states imaging as compared to the larger periodicity the x piezo drive.The periodicity in the square array of reconstructions. Imaging at negative sample biases is still greyscale maxima is characteristic of the (Ó2×Ó2)R45° reconhighly unstable and usually of short duration, but it is nonethestruction. By analogy with WO3 it is probable that this less possible and gives higher quality images than those reconstruction arises from an ordered array of on-top oxygen achieved for the p(2×2) reconstruction.Fig. 11 shows a filled ions in a O0.5 surface plane. Na ions may be confined to states image of the vacuum annealed surface taken at a sample subsurface planes, although it is hard to establish this definibias of -0.4 eV. The average periodicity measured along the tively. Again we consider that we are imaging oxygen ions, [100] and [010] directions matches that measured from empty rather than the W ions on which the density of states at the state images.The maxima in the images again correspond to Fermi energy is mainly localised. Owing to covalency the the positions of the reduced WV cations. The electron density electronic states at the Fermi level again have something of associated with these cations now allows a significant the order of 15% O 2p atomic character and this coupled with tunnelling current to flow at negative sample biases.the fact that the on-top O ions sit more more than 1.5 A ° In summary, WO3 surfaces show a range of oxygen above the WO2 plane allows the O ions to appear as the stoichiometries which give rise to at least three surface reconmaxima.Filled state images are very similar to the empty structions. The evolution between these reconstructions states images in displaying a square array of greyscale maxima. involves well defined defect structures in which oxygen vacan- However, it is additionally possible to discern extra spots in cies demonstrate a marked propensity to aggregate. This the centre of the squares.In fact, these subsidiary maxima are behaviour in many ways parallels the aggregation of bulk also present in the filled states image, but are less well defined. defects to give crystallographic shear planes. This is clarified by comparison of the corrugation profiles taken along the [100] direction from the filled and empty 4.3 STM of NaxWO3(001)15–17 states images.On average, the maxima are 0.05 A ° above the minima in +0.6 V empty state images, but 0.09 A ° above Two groups have studied STM of (001) surfaces of sodium tungsten bronzes. In our own STM experiments16,17 we have the minima in -0.6 V filled state images. For comparison, the dominant maxima are 0.20 A ° above the minima in empty state concentrated on a single sodium content Na0.665WO3 and have characterised two diVerent reconstructions.Annealing at high images and 0.21 A ° above the minima in filled state images. In our model, the subsidiary maxima correspond to bare W ions temperatures in UHV between 650 and 800 °C gives surfaces exhibiting a (Ó2×Ó2)R45° reconstruction, but lower in the WO2 plane. The variation in the height of the maxima with bias voltage can be understood in terms of the density of annealing temperatures yields a (2×1) surface.STM imaging of high temperature annealed sodium tungsten states ratio. The O contribution to the density of states decreases on moving down the conduction band and for a bronzes can be performed over a wide range of sample biases 476 J. Mater. Chem., 1998, 8(3), 469–484allows preparation of a (2×1) reconstruction17 without switch over to the high temperature (Ó2×Ó2)R45° reconstruction described above.As expected, LEED implies the presence of 90° rotated (2×1) and (1×2) domains. Again STM imaging was possible over a wide range of sample biases. STM showed that the nominally ordered (2×1) surface contained many areas where the atomic ordering is in fact far from perfect, although orthogonal (2×1) and (1×2) domains were observed in many images. A high resolution STM image is shown in Fig. 13. The unreconstructed (001) surface would be expected to display a square array of topographic maxima, but the image clearly shows a doubling of the periodicity along the [100] direction, with rows of double maxima running along [010]. Some intensity is apparent between the rows, although it should be noted that such intensity was not always observed.The images obtained in our work show considerably greater anisotropy than those previously obtained by Rohrer and coworkers.15 Line profiles taken along the atomic rows clearly demonstrate the (2×1) periodicity. The average periodicity along [010] is 3.8 A° (the lattice parameter for x=0.665 is 3.839 A° ), and the corrugation height 0.49 A ° .The average periodicity along [100] is 7.7 A ° (twice the lattice parameter), with alternating separations of 2.2 A° and 5.5 A° between adjacent maxima. The weak features between the dimer rows are 0.46 A ° above the minima. The observed lateral relaxation is so extreme that there is no obvious way of correlating the images with the ionic packing of a WO2 surface plane.However, the images are easily understood in terms of an Na0.5O termination. Displacement of the oxygen ions by 0.8 A ° along [100] doubles Fig. 12 (Top) 40 A ° ×40 A ° STM images of (Ó2×Ó2)R45° reconstruction on Na0.665WO3(001) taken at 1 nA tunnel current and -0.6 V sample bias (upper) and +0.6 V sample bias (lower).The [110] direction runs vertically. The (Ó2×Ó2)R45° unit cell is highlighted. Note square array of greyscale maxima with subsidiary maxima in the upper panel. (Bottom) corrugation profiles along [100] which reveal subsidiary maxima in both images. Adapted from ref. 16. perfectly cubic perovskite there can in fact be no mixing at the bulk C point at the bottom of the band.Thus in filled state images where one is tunnelling predominantly from electronic states 0.6 eV below the Fermi energy (and therefore only about 0.4 eV above the conduction band minimum), the bare W atoms make an enhanced contribution with respect to the on-top oxygen ions. Large area STM images of the tungsten bronze (Ó2×Ó2)R45° reconstruction were notably free of anything corresponding to the trough defects of WO3.However, anti- Fig. 13 (Top) high resolution ‘false’ 3D view of Na0.665WO3(001) from phase boundaries were imaged as a displacement of the maxima STM image acquired at 1 nA tunnel current and -0.4 V sample bias by 2.7 A ° along the [110] direction. The displacement occurs showing dimer structure within the (2×1) unit cell. Note also weak abruptly in a single Ó2 unit cell.intensity between the dimer rows. (Bottom) corrugation profiles corresponding to the high resolution image. Adapted from ref. 17. Delicate optimisation of annealing conditions below 650 °C J. Mater. Chem., 1998, 8(3), 469–484 477the surface periodicity in this direction. The weak intensity major diVerence between the two materials is that NaxWO3 surfaces show a much less rich defect chemistry, with no between the dimer rows is then attributed to ordered Na+ ions.This is shown schematically in Fig. 14. The imaging of indication of the cooperatively organised line defects that are characteristic of WO3. In some ways this simply mirrors the Na+ ions is unexpected in terms of electronic structure arguments, because the density of Na states close to the Fermi bulk defect chemistry: the metallic bronzes have no propensity toward accommodating the intriguing shear plane defects that energy is very small.However, other mechanisms for creation of apparent topographic contrast may be operative, such as are found in WO3. This in turn reflects the fact that sodium doping brings about chemical reduction of the WO3 framework variation in the local barrier height.An apparent problem is that the OMO separation within the dimeric surface species and there is a less strong driving force for reduction by oxygen loss. (2.2 A ° ) is significantly less than twice the ionic radius of the O2- ion [2r(O2-)=2.8 A° ], suggesting incipient bonding between the ions. A dimer consisting of two O2- ions would 5 STM of TiO2 have the formal valence electron configuration sg2pu4pg4su2.Close approach of the O2- ions will raise the upper antibond- 5.1 Background ing su orbital. Movement of this level above the Fermi level Titanium dioxide is of enormous industrial importance as a must result in transfer of electron density out of the orbital. white pigment, but there is also interest in application of the In the limit of complete electron transfer, the surface dimer material as a catalyst or catalyst support; in photocatalytic thus approximates to the peroxide species O22- with the devices; and as a gas sensor.Moreover, single crystals of the electron configuration sg2pu4pg4. However, the OMO bond rutile form of TiO2 are perhaps more easily and cheaply length of a true peroxide anion (e.g. 1.49 A ° in Na2O2) is available than for any other conducting metal oxide. For these considerably shorter than the OMO separation measured from and other reasons TiO2 has been the object of more STM the STM image, implying that the transfer of electron density investigations than any other oxide (with the possible exception is incomplete. Consideration of the origin of the (Ó2×Ó2)R45° of layered cuprates).At least six groups have established their reconstruction observed at high temperatures provides insight credentials in this field by obtaining images with atomic or into the driving force behind such electron transfer. In the unit cell resolution.20–22,26,27,30 As is perhaps to be expected, former case, the high surface energy associated with an O controversies have arisen over both the experimental data and surface layer (formal charge 2-) on top of aWO2 layer (formal its interpretation.charge 2+) is reduced by removal of half the surface O2- ions, The tetragonal rutile structure of TiO2 is shown in Fig. 15. generating the reconstruction. Dimerisation of oxygen ions, Viewed perpendicular to the (110) direction the rutile strucaccompanied by transfer of electron density, provides an ture is built up of ionic planes with stoichiometry alternative mechanism for the reduction of surface charge {O}–{Ti2O2}–{O}–{O}–{Ti2O2} etc., carrying formal ionic without loss of oxygen ions.charges {1-}–{2+}–{1-}–{1-}–{2+}. A stable quadrupo- In summary, the sodium tungsten bronzes show the lar sequence of charged layers is therefore obtained if the outer tunnelling characteristics of a metal and both filled and empty {Ti2O2} layer is terminated by an outer {O} layer of oxygen state imaging can be performed.Two diVerent reconstructions ions as shown schematically in Fig. 16. These oxygen ions have been identified, both arising from the need to reduce the complete six-fold coordination of half the surface Ti cations, charge in the outermost ionic layer.The (Ó2×Ó2)R45° reconleaving the remaining Ti cations five-coordinate. Argon ion struction is very similar to that on WO3(001), but the (2×1) bombardment results in selective sputtering of oxygen, along reconstruction has no analogue in the parent material. Another with significant disruption of the surface.However, oxygen stoichiometry can be partially restored by low temperature annealing to give a (2×1) reconstructed surface.107 Prior to STM experiments it was generally presumed that in this reconstruction half the rows of outer bridging oxygen ions were missing to leave four-fold coordinate surface cations as well as the original bare five-fold cations, as shown schematically in Fig. 16. The roughly octahedral crystal field experienced by the Ti cations splits the Ti 3d conduction band into three-fold degenerate t2g and two-fold degenerate eg components. A gap of just over 3 eV separates the empty 3d states from the filled valence band of O 2p states. Oxygen deficiency introduces occupied Ti 3d donor levels just below the conduction band minimum so that like WO3-x, TiO2-x is an n-type semiconductor.Fig. 14 Schematic representation of the Na0.5O terminated Na0.665WO3(001) (2×1) surface. Small spheres are Na+, large spheres are O2-, assigned their conventional ionic radii. The upper panel shows the unrelaxed structure, the lower panel the structure inferred from STM. The oxygen ions are seen to overlap in the latter. Adapted Fig. 15 The tetragonal rutile structure of SnO2 and TiO2 from ref. 17. 478 J. Mater. Chem., 1998, 8(3), 469–484Fig. 17 Empty state image of TiO2(110) (1×1) taken at +1.55 V sample bias. Note bright rows running along [001] direction, with gaps ( labelled C) in the rows. The rows are attributed to on-top O ions and the gaps to O vacancies. Reproduced with permission Fig. 16 Models for the (1×1) and (1×2) reconstructions on from ref. 26. TiO2(110). Small red spheres are Ti, large dark blue spheres are O. Note the rows of on-top O, represented by light blue spheres. 5.2 STM of TiO2(110) Despite early diYculties in obtaining STM images from TiO2(110) with the periodicity expected from the bulk crystal structure,18,19 at least 11 papers have now appeared in the literature containing unit cell resolved STM images from the (1×1) reconstructed surface.20–30 A general feature in all of these images is a series of bright rows running along the [001] direction with the correct separation along [110] for the (1×1) reconstruction.As expected from the electronic structure, imaging of TiO2 surfaces is most easily achieved at positive sample bias under conditions which probe empty electronic states.Since these states are of dominant Ti 3d atomic character, one interpretation of STM images assumes that the five-coordinate cation positions appear as topographic maxima. However, Fig. 18 Empty state image of slightly oxygen deficient TiO2(110) based on molecular orbital calculations on a [Ti7O24]20- (1×1) taken at +1.2 V sample bias and 0.5 nA tunnel current.The arrows point to features believed to be O vacancies which here appear cluster, Fischer et al. have argued that strong TiMO covalency as bright features between the rows of the (1×1) reconstruction. at TiO2 surfaces gives rise to a substantial contribution of O Reproduced with permission from ref. 30. 2p states to conduction band levels.26,28 Thus whilst electronic structure eVects partially oVset the protrusion of the bridging O atoms, they cannot completely dominate topography.In support of this assignment it was also noted that the bright topographic rows contain a large number of vacant sites, which in the ‘O imaged’ model correspond to missing on-top O (Fig. 17). It is harder to account for these vacancies in the ‘Ti imaged’ model.However, Diebold et al.30 found a diVerent sort of characteristic image associated with defects in which additional bright features appear between the bright rows (Fig. 18). Following the same line of argument as Fischer et al. obviously leads to assignment of the bright rows to Ti cation positions. Evaluation of charge density contours as outlined in Section 2.3 supports this conclusion.As shown in Fig. 19 the charge density contours for empty electronic states suggest that the STM tip should image Ti cation positions as greyscale maxima, except under conditions where the tip is very close to the surface (i.e. with very high tunnelling current). James and coworkers31 reached a similar conclusion from evaluation Fig. 19 Contour plots of charge density averaged along the [001] of empty state charge densities at fixed heights.At 4 A ° above direction for empty states within 2 eV of the conduction band minimum the surface, maximum empty state charge density is found for TiO2(110): (a) relaxed (1×1) reconstruction. Ti positions are at above Ti positions, whereas 2 A °above the surface the the edge of the plot; (b) relaxed missing row (1×2) reconstruction.Reproduced with permission from ref. 30. maximum charge density corresponds to O positions. The J. Mater. Chem., 1998, 8(3), 469–484 479theoretical work suggests that it might be possible to observe a reversal of image contrast as the tunnel current or bias is varied, thus varying the tip–sample separation. An additional aid to assignment is provided by adsorption studies.Formate ions are expected to adsorb on surface cation sites. Onishi and coworkers obtained atomically resolved STM images from TiO2(110) that had been exposed to HCOOH.22,23,25 Bright spots were found to decorate the bright rows, with a disordered arrangement of the adsorbate induced features at low coverage and a (2×1) ordering at higher coverage.The obvious interpretation of this data is to assign the spots to individual formate ions which are imaged via the LUMO of the ion. It follows that the bright rows of the underlying reconstruction correspond to Ti positions. Interpretation of STM images from the (1×2) reconstruction has proved to be equally controversial. As discussed above, the simplest model for the (1×2) reconstruction involves missing rows of bridging O atoms, thus leaving rows of fourcoordinate Ti ions between alternate pairs of rows of fivecoordinate Ti.Very high resolution STM images of this surface Fig. 21 (a) A depiction of added Ti2O3 rows on TiO2(110). A single obtained by Murray et al.27 appeared consistent with this O vacancy is included in the upmost O layer to allow viewing of the model (Fig. 20), but revealed a 0.5 A ° relaxation of five-coordiunderlying Ti positions. (b) The added Ti2O3 may be imaged as double nate Ti ions toward the missing oxygen row, as well as a stranded rows in STM (image taken at +1.0 V sample bias, 0.4 nA 0.25 A° inward relaxation. However, based on dynamic obser- tunnel current). Ordering of the added rows is postulated to produce vation on TiO2 surfaces under various thermal and oxygen the (1×2) reconstruction.Reproduced with permission from ref. 29. exposure treatments, Onishi and coworkers22,25,29 proposed an alternative model in which Ti ions diVuse to the surface to to the surface. However, in the added row model O ions will give ‘added rows’ of stoichiometry Ti2O3 (Fig. 21). The cations desorb from the edges of the added rows to give a double are not ‘bare’ in this reconstruction and, in agreement with banded structure with a node normal to the surface. The latter this model, formate ions were found not to adsorb on the conforms with experimental observations.bright rows of the (1×2) reconstruction. The model also explains the observation that the double stranded rows of the 5.3 STM of TiO2(100) (1×2) reconstruction appear to sit 2 A ° above the terraces of adjacent terrace areas of the crystal surface with (1×1) period- The (100) surface of TiO2 has a higher surface energy than icity.Very recent electron stimulated desorption studies TiO2(110). Ion bombardment and annealing inevitably leads support the model of added Ti2O3 rows.107 If the missing row to a (1×3) reconstruction.The initial model for this reconstrucmodel were correct, the desorbing flux of O ions produced by tion again involved missing oxygen rows. This model accounted electron irradiation should maximise in a single band normal for the appearance of bandgap states in photoemission associated with reduced Ti3+ ions. However, a grazing incidence Xray diVraction study suggested a more complex reconstruction involving (110) microfacets.STM provided spectacular confirmation of this model in what was probably the first atomically resolved STM study32 of a 3D oxide (Fig. 22). To reconcile the microfacet structure with photoemission and STM observation it was, however, necessary to assume that oxygen ions were missing from the top of the ridges of the microfacet structure.32,33 The Ti ions adjacent to these oxygen vacancies are reduced from TiIV to TiIII.The reduced Ti ions act as centres for nucleation of C60 molecules which can be imaged directly in STM.35 An extension of the work on TiO2(100) examined a vicinal surface cut 2.6° oV the (100) plane towards [001]. Here it was found that the vicinal oV-cut steps expected along [001] induce up–down steps along [010] in which the size of the (110) microfacets is extended.This eVectively replaces areas of (001) surface at the [001] steps by the lower energy (110) surface.34 6 STM of iron oxides 6.1 Background Iron forms three principal oxides: haematite a-Fe2O3, magnetite Fe3O4 and wustite Fe1-xO. The structure of a-Fe2O3 is based on a hexagonally close packed (hcp) array of oxygen Fig. 20 High resolution empty state STM image of TiO2(110) (1×2) ions within which two thirds of the octahedral holes are taken at 0.5 nA tunnel current and +1.0 V sample bias, together with occupied by FeIII. Fe3 O4 by contrast has a spinel structure a model for the surface derived from the STM image. Large open based on a cubic close packed (ccp) array of oxygen ions.One circles: on-top O. Large shaded circles: in-plane O. Small black half of the octahedral holes and one eighth of the tetrahedral complete circles: Ti4+. Small black incomplete circles at cell centre are holes are occupied by iron. In contrast to ‘normal’ spinels Ti3+.The bright rows are attributed to Ti4+ ions and the dark rows to on-top O. Reproduced with permission from ref. 27. where the three-valent MIII cations exclusively occupy the 480 J. Mater. Chem., 1998, 8(3), 469–484the Fe2O3. The greyscale maxima in the STM images were interpreted as arising from Fe cations that would be tetrahedrally coordinated in the bulk. These occupy one quarter of the three-fold hollow sites in the hcp oxygen ion layer. With slightly diVerent surface preparation procedures, involving a final anneal in oxygen at 1073 K, a more complex LEED pattern is observed with floreting of the primary LEED spots by smaller spots in a hexagonal array.38 Empty state STM images of these surfaces revealed an ordered hexagonal array of two types of mesoscopic islands with a superlattice dimension of about 40 A° (Fig. 23).From a consideration of atomically resolved structure within the islands, it was concluded that the island corresponded to regions of a- Fe2O3(0001) and FeO(111).Somewhat surprsingly it appeared that oxygen ion positions were imaged in the former. Fourier transform of the STM images generated a pattern that displayed the same essential features as the floreted LEED patterns, thus establishing that that this structure is due to the mesoscopic ‘biphase’ ordering. The driving force for biphase ordering was proposed to be the small mismatch between OMO separations in FeO and Fe2O3, which limits the size of FeO(111) islands that can grow on Fe2O3(0001). 6.3 STM of Fe3O4 The most interesting and comprehensive STM experiments on Fig. 22 (a) Microfacet model for the (1×3) reconstruction of Fe3O4 relate to the (111) surface42,43 which has the same TiO2(100).Large spheres are O, small spheres are Ti. The unit cell is hexagonal symmetry as Fe2O3(0001). Earlier work on indicated. (b) 3D view of the (1×3) reconstruction obtained from Fe3O4(100)39,40 and Fe3O4(110)41 has been reviewed constant current empty state STM image taken at +2 V sample bias and 0.3 nA tunnel current.Reproduced with permission from ref. 34. elsewhere.44 Surface cleaning by 500 eV ion bombardment followed by annealing in UHV at 1050 K leads to surfaces displaying sharp octahedral (Oh) sites and the two-valent MII cations occupy hexagonal LEED patterns. Very high quality STM images the tetrahedral (Td) sites, Fe3O4 is an inverse spinel. Half of from these surfaces were obtained at positive sample bias.42 the FeIII cations occupy the Td sites, whilst the Oh sites are Two diVerent types of terrace reconstruction, designated as occupied by both FeII and FeIII.Below the Verwey transition type A and type B were imaged on the same crystal, with step temperature of 120 K, the FeII and FeIII are partially ordered, heights of 3.8 A ° down from A to B and 0.5 A ° from B to A.but they probably become disordered above this temperature. Within the Fe3O4 structure there are two types of Fe containing Fe1-xO has a rocksalt-like structure, again based on a ccp (111) plane (Fig. 24). The first contains only Fe cations that array of oxygen ions. The bulk material has no range of would be Oh coordinated in the bulk. These so-called Feoct2 thermodynamic stability at room temperature and dispropor- ions occupy 3/8 of the three-fold hollow sites of the underlying tionates into Fe and Fe3O4.However, at elevated temperatures close packed (111) O layer. The second type of layer contains the rocksalt phase is entropically stabilised but always with a Fe ions that would be Oh and Td coordinated in the bulk. The significant Fe deficiency. The FeIII ions which are necessarily Feoct1 Oh ions occupy 1/8 of the three-fold hollow sites of the present in Fe1-xO are displaced into Td sites within the ccp underlying (111) O layer.The Td Fe ions are of two sorts. One oxide array and interstitial Td FeIII and Oh cation vacancies organise into so-called Koch clusters, which introduce a structural element reminiscent of Fe3O4.The Fe oxides are thus seen to be all based on close packed oxygen arrays. The OMO separation within the hexagonally packed O layers decreases progressively from 3.04 A° for Fe1-xO to 2.97 A ° for Fe3O4 and 2.90 A ° for Fe2O3. The close structural relationships between the iron oxides allows facile interconversion between the diVerent bulk phases. Moreover, a metastable c form of Fe2O3 with a ccp oxygen ion array may be prepared by careful oxidation of Fe3O4.This has a spinel-like structure, but with an ordered array of vacancies on the Oh iron sublattice 6.2 STM of Fe2O3(0001) In the first UHV STM study of Fe2O3 an (0001) oriented single crystal was studied.37 The cleaning procedure involved cycles of argon ion bombardment and annealing in UHV, followed by a final anneal in 10-6 mbar of O2 at 1000 K.The resulting surface displayed a hexagonal LEED pattern rotated by 30° relative to crystallographic axes of a-Fe2O3. STM images were obtained at positive sample bias and displayed Fig. 23 A 3D rendering of an STM image of Fe2O3(0001) taken at hexagonal symmetry with a 6.0 A ° periodicity. These obser- -2.0 V sample bias and 1 nA tunnel current.The image shows an vation were consistent with a model proposed in earlier LEED arrangement of two types of island arranged in a mesoscopic hexagonal superlattice of dimension 40 A° . Adapted with permission from ref. 38. studies which involved an epitaxial layer of Fe3O4 on top of J. Mater. Chem., 1998, 8(3), 469–484 481temperatures of around 1100 K showed a biphase hexagonal superstructure with two types of triangular mesoscopic island.43 The hexagonal superlattice had a cell length of about 50 A ° .It was shown that the islands involved co-existing Fe3O4(111) and FeO(111) islands. The greater periodicity of this superstructure as compared with that found on Fe2O3(0001) can be explained in terms of the better match between OMO separation in Fe3O4 and FeO as compared with Fe2O3 and FeO. 7 Dynamic studies of adsorbates on oxide surfaces As discussed in Section 5, Onishi and coworkers have shown that formate ions adsorbed on TiO2 surfaces can be imaged Fig. 24 A schematic representation of the structure of Fe3O4, showing the arrangment of the two types of Fe layers relative to the close under negative sample bias as extremely bright greyscale packed O layers. Small spheres are Fe, large spheres are O.Reproduced maxima. At saturation coverage on the (1×1) reconstructed with permission from ref. 42. TiO2(110) surface, the formate ions form an ordered (2×1) layer. Assuming that the formate bonds to the surface Ti ions via oxygen ions, this saturates all the available surface Ti sites set (designated Fetet1) occupies another 1/8 of the three-fold hollow sites: in an unrelaxed bulk truncation these ions would because each formate ion contains two oxygen ions.In an extremely elegant experiment,45,46 a square void in the ordered be beneath the Feoct1 ions. The other set (Fetet2) occupy 1/4 of the positions on-top of underlying O and sit above the Feoct1 overlayer of deuterated formate ions was created by scanning the STM tip across a 200 A ° ×200 A ° area of the saturated ions.From a consideration of the step heights and the structure within the high resolution images, the type A terraces were surface at +4.5 V sample bias. The mechanism for formate desorption in the proximity of the tip was not established, attributed to the Feoct2 layers, with each trimer of Fe cations capped by an oxygen ion (Fig. 25). These ions could be imaged although a field induced mechanism was postulated. At room temperature the formate ions have significant mobility and as greyscale maxima at both positive and negative sample bias. The type B images were then assigned to the second type of over a timescale of the order of 60 minutes, ions are observed to diVuse into the void area (Fig. 26). The surface diVusion of Fe layer. The outer ionic plane was shown to contain only Feoct1 and Fetet1 ions (i.e. ions occupying hollow sites), which formate ions was found to be highly anisotropic, with much greater mobility along the Ti rows in the [001] direction than were imaged as distinguishable features in the STM images only at positive sample bias.across the O ridges that are encountered in the [110] direction. The ability of STM to image simple molecules in a direct Surfaces prepared by annealing at somewhat more elevated way clearly allows surface coverages to be determined simply by counting the number of molecules in a defined area. This idea was exploited to study the kinetics of decomposition of acetate ions on TiO2(110).At saturation coverage, acetate ions form a (2×1) overlayer on TiO2(110) (1×1) and, as with formate, the acetate ions are imaged as greyscale maxima at Fig. 25 STM images of Fe3O4(111) showing the two types of image (referred to as A and B in the text) in the right hand panels of (a) and (b) respectively. Models for the terraces are shown in the left hand panels.In both cases the large grey circles are O atoms. In (a) the small white circles are Feoct2 atoms. These group into trimer pairs each capped with an O atom represented by a large black circle. The Fig. 26 Series of empty state STM images (320 A ° ×320 A ° , +1.0 V, capped trimers define the unit cell as shown in the figure. Fe atoms in the underlying layer are represented by small black circles (Fetet1) and 0.3 nA) of (2×1) deuteroformate (DCOO) saturated TiO2(110) following desorption of formate ions from a central square region using the small white circles (Feoct1).In (b) Feoct1 atoms (labelled e) are drawn as small white circles and Fetet1 atoms ( labelled d) as small dark circles. STM tip. The elapse times at room temperature increase from top left to bottom right panel in the order 0.5 min, 9 min, 15 min (top row), Feoct2 atoms in the underlying Fe layer are drawn as small white circles.Vacancies at Feoct1 (e) positions are designated by black squares. 30 min, 40 min, 45 min (centre row) 51 min, 60 min, 79 min (bottom row). Reproduced with permission from ref. 45. Reproduced with permission from ref. 42. 482 J. Mater.Chem., 1998, 8(3), 469–484positive sample bias (Fig. 27). By subjecting a crystal surface was possible to derive a first order rate constant of (4±1)×10-3 s-1. to a temperature jump from 510 K to 540 K, it was possible to induce decomposition of the adsorbed acetate ions.47 The number of ions appearing in a 100 A ° ×100 A ° scan was counted 8 Concluding remarks in series of frames at an interval of 16.6 s per frame.The decrease in the number of adsorbed atoms per frame followed The application of STM to problems of oxide surface structure the expected first order decay (Fig. 28). From the analysis it and dynamics has advanced very rapidly in the past few years. The technique is valuable in complementing other techniques in the study of regular periodic structures: whilst STM alone cannot provide a complete description of surface structure of the sort that could be obtained from surface X-ray diVraction or LEED it has nonetheless already proved its worth in a number of cases in establishing initial models for more refined subsequent analysis.Of perhaps greater long term importance is the ability of STM to investigate defect structures of various sorts and to characterise surfaces where one or more diVerent terrace structures coexist on the same surface. These capabilities combined with the recent possibility to undertake STM experiments at elevated temperatures bring the technique to a point where it is now able to give a realistic atomic view of the complex sorts of surface process that are undoubtedly involved in many catalytic reactions.References 1 V. E. Henrich and P. A. Cox, T he Surface Science ofMetal Oxides, Cambridge University Press, Cambridge, 1994. 2 H. J. Freund, H. Kuuhlenbeck and V. Staemmler, Rep. Prog. Phys., 1996, 59, 283. 3 C. J. Chen, Introduction to Scanning T unneling Microscopy, Oxford University Press, Oxford, 1993. 4 R. Wiesendanger, Scanning Probe Microscopy and Spectroscopy, Cambridge University Press, Cambridge, 1994. 5 M. D. Kirk, J. Nogami, A. A. Baski, D. B. Mitzi, A. Kapitulnik, T. H. Geballe and C. F. Quate, Science, 1988, 242, 1673. 6 C. K. Shih, R. M. Feenstra, J. R. Kirtley and G. V. Chandrashekhar, Phys. Rev. B, 1989, 40, 2682. 7 M. Tanaka, S. Yamazaki, M. Fujinami, T. Takahashi, H. Katayama-Yoshida, W. Mizutani, K.Kajimura and M. Ono, J. Vac. Sci. T echnol. A, 1990, 1, 475. 8 X. L. Wu, C. M. Lieber, D. S. Ginley and R. J. Baughman, Appl. Phys. L ett., 1989, 55, 2129. 9 Z. Zhang and C. M. Lieber, J. Phys. Chem., 1992, 96, 2030. 10 F. H. Jones, K. Rawlings, J. S. Foord, P. A. Cox, R. G. Egdell, J. B. Pethica and B. M. R. Wanklyn, Phys. Rev. B, 1995, 52, R14392. 11 F. H. Jones, K. Rawlings, J.S. Foord, P. A. Cox, R. G. Egdell, Fig. 27 Bottom panel: model for a (2×1) acetate monolayer on J. B. Pethica, B. M. R. Wanklyn, S. C. Parker and P. M. Oliver, TiO2(110). Top panel: STM image of a (2×1) acetate monolayer on Surf. Sci., 1996, 359, 107. TiO2(110) acquired at +2.0 V sample bias and 0.3 nA tunnel current. 12 P. M. Oliver, S. C. Parker, R. G. Egdell and F. H. Jones, J. Chem.Reproduced with permission from ref. 47. Soc., Faraday T rans., 1996, 92, 2049. 13 F. H. Jones, R. A. Dixon and A. Brown, Surf. Sci., 1996, 369, 343. 14 R. A. Dixon, J. J. Williams, D. Morris, J. Rebane, F. H. Jones and R. G. Egdell, Surf. Sci., in press. 15 G. S. Rohrer, W. Lu, M. L. Norton, M. A. Blake and C. L. Rohrer, J. Solid State Chem., 1994, 109, 359. 16 F. H. Jones, K.Rawlings, S. Parker, J. S. Foord, P. A. Cox, R. G. Egdell and J. B. Pethica, Surf. Sci., 1995, 336, 181. 17 F. H. Jones, K. Rawlings, J. S. Foord, P. A. Cox and R. G. Egdell, J. Chem. Soc., Chem. Commun., 1985, 2419. 18 G. S. Rohrer, V. E. Henrich and D. A. Bonnell, Science, 1990, 250, 1239. 19 G. S. Rohrer, V. E. Henrich and D. A. Bonnell, Surf. Sci., 1992, 278, 146. 20 M.Sander, and T. Engel, Surf. Sci., 1994, 302, L263. 21 D. Novak, E. Garfunkel and T. Gustafsson, Phys. Rev. B, 1994, 50, 5000. 22 H. Onishi and Y. Iwasawa, Surf. Sci., 1994, 313, L783. 23 H. Onishi and Y. Iwasawa, Chem. Phys. L ett., 1994, 226, 111. 24 Y. Yamaguchi, H. Onishi and Y. Iwasawa, J. Chem. Soc., Faraday T rans., 1995, 91, 1661. Fig. 28 Semi-log plot of number of acetate ions imaged by STM in a 25 H.Onishi, K. Fukui and Y. Iwasawa, Bull. Chem. Soc. Jpn., 1995, 68, 2447. 100 A° ×100 A° frame on a TiO2(110) surface that had been saturated with acetate (CH3COO) at room temperature and then subject to a 26 S. Fischer, A. W. Munz, K. D. Schierbaum and W. Gopel, Surf. Sci., 1995, 337, 17. temperature jump from 510 K to 540 K against frame number. The elapse time between frames is 16.6 s.Reproduced with permission 27 P. W. Murray, N. G. Condon and G. Thornton, Phys. Rev. B, 1995, 51, 10989. from ref. 47. J. Mater. Chem., 1998, 8(3), 469–484 48328 S. Fischer, A. W. Munz, K. D. Schierbaum and W. Gopel, J. Vac. 69 I. Maggio-Aprile, Ch. Renner, A. Erb, E. Walker and O. Fischer, Phys. Rev. L ett., 1995, 75, 2754. Sci. T echnol.B, 1996, 14, 961. 70 S. Ikebe, K. Suzuki and H. Nishikawa, Jpn. J. Appl. Phys., 1992, 29 H. Onishi and Y. Iwasawa, Phys. Rev. L ett., 1996, 76, 791. 31, 2221. 30 U. Diebold, J. F. Anderson, K. O. Ng and D. Vanderbilt, Phys. 71 F. H. Jones, R. G. Egdell, A. Brown and F. R. Wondre, Surf. Sci., Rev. L ett., 1996, 77, 1322. 1997, 374, 80. 31 O. Gulseren, R. James and D. W. Bullett, Surf.Sci., 1997, 72 J. Heil, J. Wesner, B. Lommel, W. Assmus and W. Grill, J. Appl. 377–379, 150. Phys., 1989, 65, 5220. 32 P. W. Murray, F. M Leibsle, H. J. Fisher, C. F. J. Flipse, 73 D. Anselmetti, R. Wiesendanger, H. J. Guntherodt and C. A. Muryn and G. Thornton, Phys. Rev. B, 1992, 46, 12877. G. Gruner, Europhys. L ett., 1990, 12, 241. 33 P. W. Murray, F. M. Leibsle, C. A.Muryn, H. J. Fisher, 74 G. Rudd, D. Novak, D. Saulys, R. A. Bartynski, S. Garofalini, C. F. J. Flipse and G. Thornton, Surf. Sci., 1994, 321, 217. K. V. Ramanujachary, M. Greenblatt and E. Garfunkel, J. Vac. 34 P. W. Murray, F. M. Leibsle, C. A. Muryn, H. J. Fisher, Sci. T echnol. B, 1991, 9, 909. C. F. J. Flipse and G. Thornton, Phys. Rev. L ett., 1994, 72, 689. 75 A. Zettl, L. C.Bourne, J. Clarke, M. F. Crommie, M. F. Hundley, 35 P. W. Murray, J. K. Gimzewski, R. R. Schittler and G. Thornton, R. E. Thompson and V. Walter, Synth.Met., 1989, 29, F445. Surf. Sci., 1996, 367, L79. 76 U. Walter, R. E. Thomson, B. Burk, M. F. Crommie, A. Zettl and 36 P. W. Murray, J. Shen, N. G. Condon, S. J. Pang and J. Clarke, Phys. Rev. B, 1992, 45, 11 474. G. Thornton, Surf.Sci., 1997, 380, L455. 77 W. Lu, N. Nevins, M. L. Norton and G. S. Rohrer, Surf. Sci., 37 N. G. Condon, P. W. Murray, F. M. Leibsle, G. Thornton, A. R. 1993, 291, 395. Lennie and D. J. Vaughan, Surf. Sci., 1994, 310, L609. 78 S. C. Langford, M. Zheni, L. C. Jensen and J. T. Dickinson, J. Vac. 38 N. G. Condon, F. M. Leibsle, A. R. Lennie, P. W. Murray, Sci. T echnol. A, 1990, 8, 3470.D. J. Vaughan and G. Thornton, Phys. Rev. L ett., 1995, 75, 1961. 79 T. Matsumoto, H. Tanaka, T. Kawai and S. Kawai, Surf. Sci., 39 R. Wiesendanger, I. V. Shvets, D. Burgler, G. Tarrach, 1992, 278, L153. H. J. Gutherodt, J. M. D. Coey and S. Graser, Science, 1992, 80 Y. Liang and D. A. Bonnell, Surf. Sci., 1993, 285, L510. 255, 583. 81 H. Tanaka, T. Matsumoto, T. Kawai and S. Kawai, Surf.Sci., 40 G. Tarrach, D. Burgler, T. Schaub, R. Wiesendanger and 1994, 318, 29. H. J. Guntherodt, Surf. Sci., 1993, 285, 1. 82 Q. Jiang and J. Zegenhagen, Surf. Sci., 1996, 367, L42. 41 R. Jansen, V. A. M. Brabers and H. van Kempen, Surf. Sci., 1995, 83 H. Bando, T. Shimitzu, Y. Aiura, Y. Harauyama, K. Oka and 328, 237. Y. Nishihara, J. Vac. Sci. T echnol. B, 1996, 14, 1060. 42 A.R. Lennie, N. G. Condon, F. M. Leibsle, P. W. Murray, 84 R. L. Smith, W. Lu and G. S. Rohrer, Surf. Sci., 1995, 322, 293. G. Thornton and D. J. Vaughan, Phys. Rev. B, 1996, 53, 10244. 85 R. L. Smith, G. S. Rohrer, K. S. Lee, D. K. Seo and M. H. 43 N. G. Condon, F. M. Leibsle, T. Parker, A. R. Lennie, Whangbo, Surf. Sci., 1996, 367, 87. D. J. Vaughan and G. Thornton, Phys.Rev. B, 1997, 55, 15885. 86 R. A. Goschke, K. Vey, M. Maier, U. Walter, E. Goering, 44 F. M. Leibsle, P. W. Murray, N. G. Condon and G. Thornton, M. Klemm and S. Horn, Surf. Sci., 1996, 348, 305. J. Phys. D: Appl. Phys., 1997, 30, 741. 87 M. L. Norton, J. G. Mantovani and R. J. Warmack, J. Vac. Sci. 45 H. Onishi and Y. Isawara, L angmuir, 1994, 10, 4414. T echnol. A, 1989, 7, 2898. 46 H. Onishi and Y. Isawara, Surf. Sci., 1996, 357–358, 773. 88 G. S. Rohrer, W. Lu, R. L. Smith and A. Hutchinson, Surf. Sci., 47 H. Onishi, Y. Yamaguchi, K. Fukui and Y. Isawara, J. Phys. 1993, 292, 261. Chem., 1996, 100, 9582. 89 R. L. Smith and G. S. Rohrer, J. Solid State Chem., 1996, 124, 104. 48 F. H. Jones, R. Dixon, J. S. Foord, R. G. Egdell and J. B. Pethica, 90 M. R. Castell, P. L. Wincott, N. G. Condon, C. Muggelberg, Surf. Sci., 1997, 376, 367. G. Thornton, S. L. Dudarev, A. P. Sutton and G. A. D. Briggs, 49 J. Buisset, H.-P. Rust, E. K. Schweizer, L. Cramer and Phys. Rev. B, 1997, 55, 7859. 91 M. R. Castell, C. Muggelberg, G. A. D. Briggs and D. T. Goddard, A. M. Bradshaw, Surf. Sci., 1996, 349, L147. J. Vac. Sci. T echnol. B, 1996, 14, 966. 50 J. TersoV and D. R. Hamann, Phys. Rev. L ett., 1983, 50, 1998. 92 C. Muggelberg, M. R. Castell, G. A. D. Briggs and D. T. Goddard, 51 G. Doyen, E. Kotter, J. P. Vigneron and M. ScheZer, Appl. Phys. Proc. 4th Nordic Conf. Surf. Sci, ed. S. Raaen and J. Bremer, 1997, A, 1990, 51, 281. p. 103. 52 P. Sautet and C. Joachim, Chem. Phys. L ett., 1991, 185, 23. 93 M. Baumer, D. Cappus, H. Kuhlenbeck, H. J. Freund, 53 P. Sautet, J. C. Dunphy, D. F. Ogletree, C. Joachim and G. Wilhelmi, A. Brodde and H. Neddermeyer, Surf. Sci., 1991, M. Salmeron, Surf. Sci., 1994, 315, 127. 253, 116. 54 H. Galloway, P. Sautet and M. Salmeron, Phys. Rev. B, 1996, 94 F. Rohr, M. Baumer, H. J. Freund, J. A. Meijas, V. Staemmler, 54, 11145. S. Muller, L. Hammer and K. Heinz, Surf. Sci., 1997, 372, L291. 55 V. P. S Awana, S. B. Samanta, P. K. Dutta, E. Gmelin and 95 N. M. D. Brown and H. You, Surf. Sci., 1990, 233, 317. A. V. Narlikar, J. Phys.: Condens. Matter, 1991, 3, 8893. 96 C. A. Ventrice, H. Hannemann, A. Brodde and H. Neddermeyer, 56 C. K Shih, R. M. Feenstra and G. V. Chandrashekhar, Phys. Rev. Phys. Rev. B, 1994, 49, 5773. B, 1991, 43, 7913. 97 H. C. Galloway, J. J. Benitez and M. Salmeron, J. Vac. Sci. 57 Y. S. Luo, Y. N. Yang and J. H. Weaver, Phys. Rev. B, 1992, T echnol. A, 1994, 12, 2302. 46, 1114. 98 C. Xu and D. W. Goodman, Chem. Phys. L ett., 1996, 263, 13. 58 Z. Zhang and C. M. Lieber, Phys. Rev. B, 1992, 46, 5845. 99 Th. Bertrams, F. Winkelman, Th. Uttich, H. J. Freund and 59 M. Oda, C. Manabe and M. Ido, Phys. Rev. B, 1996, 53, 2253. H. Neddermeyer, Surf. Sci., 1995, 331–333, 1515. 60 Ch. Renner and O. Fischer, Phys. Rev. B, 1995, 51, 9208. 100 J. Libuda, F. Winkelman, M. Baumer, H. J. Freund, Th. Bertrams, 61 X. L. Wu, Y. L. Wang, Z. Zhang and C.M. Lieber, Phys. Rev. B, H. Neddermeyer and K. Muller, Surf. Sci., 1994, 318, 61. 1991, 43, 8729. 101 M. Baumer, J. Libuda, A. Sandell, H. J. Freund, G. Graw, 62 Z. Zhang, C.-C. Chen and C. M. Lieber, Phys. Rev. B, 1992, Th. Bertrams and H. Neddermeyer, Ber. Bunsen-Ges. Phys. 45, 982. Chem., 1995, 99, 1381. 63 Z. Zhang, C. M. Lieber, D. S. Ginley, R. J. Baughman and 102 S. Stemple, M. Baumer, M. Frank, J. Libuda and H. J. Freund, B. Morosin, J. Vac. Sci. T echnol. B, 1991, 9, 1009. Proc. 4th Nordic Conf. Surf. Sci., ed. S. Raaen and J. Bremer, 1997, 64 P. K. Dutta, S. B. Samanta, A. V. Narlikar, C. Chen, J. W. Hodby p. 112. and B. M. R. Wanklyn, Philos.Mag. A, 1992, 66, 507. 103 J. B. Pethica, Phys. Rev. L ett., 1986, 57, 3235. 65 L. E. C. Van de Leemput, P. J. M. Van Bentum, L. W. M. Schreurs 104 M. Kielwein, K. Saiki, G. Roth, J. Fink, G. Paasch and and H. Van Kempen, Physica C, 1988, 152, 99. R. G. Egdell, Phys. Rev. B, 1995, 51, 10320. 66 L. E. C. Van de Leemput, P. J. M. Van Bentum, H. Van Kempen 105 D. W. Bullett, J. Phys. C, 1983, 16, 2197. and L. W. M. Schreurs, Physica C, 1988, 153–155, 996. 106 M. C.Wu and P. J. Moller, Surf. Sci., 1989, 224, 265. 67 H. L. Edwards, J. T. Markert and A. L. de Lozanne, Phys. Rev. 107 Q. Guo, I. Cocks and E. M. Williams, Phys. Rev. L ett., 1996, 77, 3851. L ett., 1992, 69, 2967. 68 H. L. Edwards, D. J. Derro, A. L. Barr, J. T. Markert and A. L. de Lozanne, Phys. Rev. L ett., 1995, 75, 1387. Paper 7/06191I; Received 26th August, 1997 484 J. Mater. Chem., 1998, 8(3), 469–484

ISSN:0959-9428    

DOI:10.1039/a706191i

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Feature Article Cholesterol-based functional tectons as versatile building-blocks for liquid crystals, organic gels and monolayers Seiji Shinkai*a and Kazutaka Muratab aDepartment of Chemical Science & T echnology, Faculty of Engineering, Kyushu University, Fukuoka 812, Japan bCHEMIRECOGNICS Project, ERATO, Research Development Corporation of Japan, 2432-3 Aikawa-cho, Kurume, Fukuoka 830, Japan This review article introduces the application of the cholesterol moiety, which has been commonly utilised as a versatile building block in liquid crystals, organic gels and monolayers, to molecular recognition.EVort has especially been made to clarify the characteristics of cholesterol-based assembly systems and the phenomena common to these three systems.Cholesterol is a well-known natural product and frequently reactions in the cholesteric chiral medium has met with only mediocre success.5 With our approach, we do not try to control appears as a building block in molecular assemblies. Its versatility stems from its unique structural characteristics not the guest but allow the guest to control the chirality of its environment.found in other compounds. (1) It is commercially available as a cheap natural product; (2) its structure is rigid and possesses Steroidal crown compounds have been synthesised by several groups and their liquid crystalline properties and membrane- multiple chiral carbon atoms which are prerequisites for chiral recognition;1 (3) its derivatives can form several unique aggre- forming properties have also been studied.6,7 To the best of our knowledge at the beginning of this study, few groups had gates such as liquid crystals, organic gels and monolayers; (4) the absolute configuration of C-3 can be easily inverted.attempted to apply this system to molecular recognition. Our first foray into the realm of environmental amplification Factor (3) in particular most clearly characterises the nature of cholesterol as a building block. Cholesteric liquid crystals employs cholesteric crown ethers 1.A mixture of cholesteryl nonanoate and cholesteryl chloride is known as a room are aggregates composed of chirally-oriented planes; organic gels are fibrous (mostly one-dimensional) aggregates which are temperature cholesteric liquid crystal and generates visible reflected light because the pitch length is comparable to the capable of gelatinizing organic media; monolayers are twodimensional aggregates formed at the air–water interface.To wavelength of visible light. A ternary blend of cholesteryl nonanoate, cholesteryl chloride and 1 (ca. 25 mol%) changes the best of our knowledge, cholesterol is the only compound that can provide so many diVerent mesophases.2 Our present its helical pitch length upon addition of alkali metal salts and, therefore, its colour.8 Since the pitch length change is dependent research interest is in molecular recognition combined with supramolecular systems.It thus occurred to us that oriented on the size of added alkali metal cations, one can easily detect the metal cations from the visible colour change.8 The results molecular assemblies based on cholesterol aggregates might be useful in molecular recognition processes diYcult to attain indicate the potential to apply cholesteric liquid crystals as a new metal sensory system.Interestingly, the pitch length for 2 with other monomeric host compounds. For example, the was scarcely aVected by addition of these metal cations.8 This cholesterol–cholesterol interaction may be aVected by guest means that if a spacer is inserted between the metal-binding molecules intercalated between them or by host–guest intersite and the liquid-crystal-forming site, the event occurring at actions with the host moiety appended to the cholesterol the metal-binding site is absorbed by the flexible molecular skeleton.These chemical signals should aVect the orientation segment in the spacer. and/or the pitch length in the cholesterol stack, which can eventually be read out as a change in the spectroscopic properties (i.e. physical signals). In this review article we survey cholesterol-based assemblies (liquid crystals, organic gels and monolayers at the air–water interface) with potential in the design of novel molecular recognition systems.Liquid crystal systems DiVerentiation by colour should be the magnum opus of chiral discrimination. In homogeneous solutions this idea has been realised in chromogenic chiral crown ethers and calixarenes.3 For example, Kaneda et al. employed chiral azophenolic acerands in which chiral point interactions can be directly ‘read’ by a colour change.3a,b Our method is somewhat diVerent and employs a chiral environment as outlined in the introduction.The use of cholesterol as a chiral environment is not a new idea: many people have studied and attempted to control chiral reactions in liquid crystals,4,5 but control of chiral * E-mail seijitcm@mbox.nc.kyushu-u.ac.jp J.Mater. Chem., 1998, 8(3), 485–495 485achieved employing just 1H NMR spectroscopy (CDCl3).11 The molar ratio of the extracted species could be conveniently estimated by the integral intensities of selected proton resonances of the monosaccharide versus either the phenyl or the alkenyl protons of compound 3. In the case of the D-fucose complex the anomeric proton of the monosaccharide at d 5.9 and the alkenyl proton of compound 3 at d 5.4 were employed. In general, all the saccharides form a 152 complex with compound 3.Two main structural classes exist for the extracted As the helical pitch of the mixed cholesteric liquid crystal is monosaccharides; either two five-membered rings are formed chiral, this system may be applicable to chiral molecular (1,2–3,4 complex) or a five- and a six-membered ring are recognition.It was found that alkali (R)- and (S)-mandelates formed (1,2–4,6 and 2,3–4,6 complexes: Scheme 1). can change the helical pitch in an enantioselective manner.9 trans-Six-membered rings are less stable than the The largest reflectance wavelength maximum (lR) diVerence five-membered ring.13 From Fig. 1 it can be seen that this between (R) and (S) (69 nm) was observed for the combination structural analysis falls short of describing the actual situation.of 1b and potassium salts, which could be visually diVerentiated Some systems that could have formed perfectly good five- [green for (R) and blue for (S)].9 On the other hand, such a membered rings opt for the six-membered ring alternative.colour change was not observed for either the combination of Why do these systems prefer the six-membered ring? With 1b and Bu4N+ mandelates or for the combination of 2 and galactose and talose the ring is a cis rather than a trans six- potassium mandelates.9 The more straightforward chiral recogmembered one, and it is known from decalin that the cis fused nition with ternary cholesteric liquid crystals was achieved for six-membered ring system is conformationally flexible whereas optically-active ammonium ions through the RNH3+–benzothe trans system is fixed. This conformational lability may 18-crown-6(1b) interaction.10 In a-amino acid ester derivatives, provide enough advantageous entropic energy to tip the bal- the lR for D-isomers shifted to longer wavelength whereas that ance in favour of the six-membered ring.With allose and for L-isomers shifted to shorter wavelength (except alanine mannose a trans six-membered ring is formed, but this may methyl ester hydrochloride which has the smallest residual not be an inherent preference for the six-membered ring but group). The largest lR diVerence was observed for phenylrather a preference for the 2,3 five-membered ring over the 1,2 alanine methyl ester hydrochloride (+45 nm for D-isomer and five-membered ring.In a 152 complex once the 2,3 five ring -19 nm for L-isomer). has been formed, the only choice remaining is the formation With saccharides and cholesterol boronic acid 3 similar of a trans six-membered ring. doping experiments were carried out and produced similar The induced shifts produced by the extracted monosacchar- colour changes.11 Since the interaction of boronic acid with ides in a cholesteryl nonanoate–cholesteryl chloride composite saccharides produces a cyclic boronate ester, this covalent liquid crystal are given in Table 1.An example for the visible linkage allows us to isolate the complexes. Once the complexes colour change is shown in Fig. 2. Analysis of Table 1 shows are isolated the opportunity exists to examine structure– that saccharides with similar structural features induce shifts activity relationships. Such an opportunity was not available in the same sense (Fig. 1).11 Within the saccharides extracted with alkali metals and ammonium ions. Solvent extraction of three such structure–shift group relationships exist: the first saccharides was carried out at 25 °C using solid–liquid (CDCl3) group includes saccharides with two (cis) five-membered extraction.As much by serendipity as by design 3 was a much boronate ester rings (Group I). When the first ring (following more eYcient saccharide extractor than the lipophilic boronic normal saccharide numbering) is down with respect to the acids previously used;12 complete extraction was observed for saccharide plane and the second ring is up, a red shift is all monosaccharides.Selection within the saccharides similar induced (D-fucose and L-arabinose). Conversely, when the first to that previously observed was seen in the rate of extraction, ring is up and the second is down, a blue shift is induced (D- the less favourable saccharides taking longer to be completely fructose, D-arabinose and L-fucose).The next structure–shift extracted. However, after 48 h all the saccharides investigated sub-group contains saccharides with one cis five-membered were completely extracted.11 Characterisation of the extracted saccharides could be ring and either a trans five- or a trans six-membered ring B HO N O O H HO = R O O HO HO OH OH HO HO HO OH OH OH O O B O O B O O O B O O B O O O B O O O B O O B O O B O HO R R R HO R R OH R R R OH or GENERAL SACCHARIDE FRUCTOSE 1 fucose arabinose xylose lyxose galactose glucose allose mannose fructose 1,2-3,4 complex 1,2-4,6 complex 2,3-4,6 complex 2,3-4,5 complex 1 2 3 4 5 6 1 2 3 4 5 6 Scheme 1 Structures of 251 complexes (3–monosaccharide) 486 J.Mater. Chem., 1998, 8(3), 485–495Fig. 2 Visible colour change induced by a 3 fucose 251 complex behaviour may not be anomalous if a threshold must be overcome before a shift in the pitch occurs. From the above qualitative structural analysis and the result that the 151 complex of 2-deoxy-D-galactose causes no shift, the spatial disposition of two cholesteryl boronic acid moieties is the driving force for the change in the cholesteric pitch.Structural analysis of the complexes utilising a molecular orbital calculation reveals a quantitative relationship between the magnitude and direction of the induced shift and the angle between the phenyl planes of the two boronic acid moieties (Fig. 3). For the correlation depicted in Fig. 3, one apparent anomaly needs to be clarified. Why are blue shift additives ‘eVective’ at all angles, but those compounds inducing a red shift have a threshold value of about 40° (D-fructose)? This can be easily explained by the inherent twist of the support liquid crystal; complexes that add to this inherent twist cause blue shifts, and those that subtract cause red shifts, with D-fructose (40°) acting as the eVective zero or threshold.D-arabinose D-allose D-fucose L-arabinose L-fructose L-fucose D-arabinose D-fructose D-glucose L-glucose L-mannose D-allose D-mannose D-xylose L-lyxose L-xylose D-lyxose The foregoing findings demonstrate that the pitch length in Fig. 1 Cartoon representation of 251 complexes: the central hexagon cholesteric liquid crystals is useful not only for size recognition denotes the pyranose ring of metal cations but also for chiral recognition of natural products.Furthermore, it is noteworthy that the recognition events can be detected by a visible colour change. (Group II). When the first ring (cis five-membered) is down, a red shift is induced (D-glucose, D-allose and D-xylose) and conversely when this ring is up, a blue or no shift is induced Polymer–liquid crystal composite membrane (L-glucose and D-mannose). The final structure–shift group is systems the most anomalous; the saccharide contains both a cis fivemembered ring and a cis six-membered ring (Group III). When Biological membranes are composed of various kind of phospholipids, cholesterols and proteins, and the fundamental the first ring is either up or down, a blue shift is induced (Dgalactose and L-galactose), but D-talose induces no shift.This functions such as permeation and selectivity are closely associated with the gel–liquid crystal phase transition. Therefore, the anomalous behaviour may be a result of the conformational lability of the cis six-membered ring, or as explained below the phase transition would be one of the most essential functions Table 1 Shifts in the reflection light maxima (l nm) caused by added 251 complex: 3–saccharide11 induced shifta relative to compound 3 group 251 complex 3–saccharide mol% 2.4b mol% 1.2c I D-fucose 123±13 L-arabinose 27±8 L-fucose -164±12 -95±6 D-arabinose -82±10 -45±6 D-fructose -37±11 -23±11 II D-glucose 91±10 42±12 D-allose 69±10 19±10 D-xylose 34±15 L-glucose -71±10 -61±6 D-mannose -3±10 5±6 III D-galactose -44±10 -33±11 L-galactose -17±10 -21±7 D-talose 3±11 blank 2-deoxy-D-galactose 6±15 aShifts are the average of five repeats; errors given are the maximum deviation from the mean.bAt 2.4 mol% compound 3 has a base reflectance of 625±10 nm. cAt 1.2 mol% compound 3 has a base reflectance of 650±6 nm.J. Mater. Chem., 1998, 8(3), 485–495 487above 100 °C. It was found, however, that a mixture of 1a and 1b or 1a and 2 results in the cholesteric liquid-crystal phase even at room temperature.17,18 A composite membrane composed of 1a, 2 and Perplene [a random copolymer consisting of oxytetramethyleneoxyterephthaloyl and poly(oxytetramethylene) oxyterephthaloyl units] permeated alkali thiocyanates and the rate of ion permeation was in the order Na+>K+ >Cs+.17 This order is identical with the metal aYnity of benzo-15-crown-5.On the other hand, when a mixture of 1a and 1b was used, the order was changed to K+>Na+> Cs+.18 Obviously, this is due to the contribution of 1b which shows the highest aYnity for K+. The ion permeation selectivity of these membranes was further examined using 1, 2 and 6.19 Compound 6b provided a stable cholesteric mesophase even at room temperature and the composite membrane showed remarkably high Li+ selectivity.A similar result was obtained from a 6a–6b mixture. A 1a–6a mixture still kept the high Li+ selectivity, whereas a 1a–2 mixture showed the selectivity order of Na+>K+>Cs+ >Li+.The results imply that in the mixed systems the crown Fig. 3 Plots of reflectance wavelength (lR) vs. calculated Ph–Ph ether with the smallest ring size acts as a critical ‘gate’ for ion dihedral angle for 251 complexes. Saccharides used: a, D-fructose; b, permeation: that is, 6 allows Li+ to pass through the benzo- L-galactose; c, D-galactose; d, D-mannose; e, L-arabinose; f, D-arabinose; g, D-glucose; h, L-glucose; i, D-allose; j, L-fucose. 13-crown-4 ring but does not allow Na+ to do so. It was proposed that ion permeation in these systems occurs according to an ion-channel mechanism but not according to provided by phospholipid membranes. Kajiyama and co- an ion-carrier mechanism.18,19 The first evidence for this proworkers14 –16 demonstrated that composite membranes in posal is the activation energy for the Arrhenius plots (ca.which the liquid crystalline material is embedded in a polymer 27.3–48.7 kJ mol-1). These values are consistent with the ionmatrix are applicable to membrane mimetic permeation con- channel mechanism which includes an ion translocation protrol, because a distinct change in the thermal molecular motion cess.In fact, very similar values have been obtained for natural occurs at the crystal–liquid crystal phase transition tempera- and artificial channel-forming compounds.20,21 In contrast, the ture. For example, in ternary composite membranes composed ion-carrier mechanism is known to need a higher activation of a polymer (polycarbonate, PC), a liquid crystal (4: crystal– energy (90–120 kJ mol-1).22,23 The second piece of evidence is nematic liquid crystal phase-transition temperature, 305 K) obtained from dilution of the membrane with nonionophoric and crown ethers having a fluorocarbon chain (5) K+ transport cholesterol derivatives. For example, when crown-containing is ‘completely’ suppressed below 305 K whereas K+ is trans- cholesterols were diluted with 7 which provides the cholesteric ported rapidly above 305 K.16 The results imply that ion liquid-crystal phase at room temperature, ion permeation was permeation across these membranes is drastically aVected by strikingly suppressed.If ion permeation occurs according to the fluidity of the membrane medium. the ion-carrier mechanism, the rate should gradually decrease with increasing concentration of 7.The construction of an ion channel was further corroborated through the ion-conducting behaviour.24 In composite films with Li+ or Na+ salts, the activation energy for ion conduction was lower under the cholesteric liquid-crystal conditions than under the corresponding isotropic conditions. The finding indicates that the crown rings ordered in the cholesteric liquid-crystal phase contruct a channel-like array and facilitate specific ion conduction.It was also shown that in the composite film containing a small amount of 8a, significant photoinduced ion-conductivity switching is realized by alternate irradiation with UV and VIS light.24 This phenomenon is attributed to the order– disorder cycle of the liquid crystal state induced by cis–trans photoisomerisation of the azobenzene moiety.The foregoing references reveal that cholesteric liquid crystals are useful not only for the orientation of appended functional groups but also as a medium to selectively permeate guest cations. As mentioned in the previous section, the pitch length of cholesteric liquid crystals changes enantioselectively in response to chirality of intercalated guest molecules.We thus believe that cholesterol-based membranes are eVective for chiral separation of racemic compounds. Gel systems Recently, a new class of cholesterol-based gels, held together O O O O O NCH2CH2CH2(CF2)7CF3 C2H5O CH N C4H9 R O O (CH2) nCO O CH(CH2)7CO CH3(CH2)7CH O 4 5 6a n = 0 6b n = 2 7 O O by weak hydrogen bonding or van der Waals interactions, was reported.25–33 Our interest was sparked, since we have pre- If steroidal crown compounds are appropriately embedded in a polymer matrix and the stacked crown column is formed, viously nurtured cholesterol derivatives bearing crown ether, azobenzene or boronic acid moieties.8–11,17–19,24,34 Our it is expected to act as a sort of ion channel.The cholesteric liquid-crystal phase of steroidal crown ethers usually appears continuing theme is the development of signal-responsive 488 J.Mater. Chem., 1998, 8(3), 485–495N N CH3O COO COO N N CH3O trans-10Me in the gel phase cis-10Me in the sol phase UV VIS Scheme 2 Sol–gel phase transition induced by photoisomerisation of 10Me in butan-1-ol chemistry, employing molecular transducers capable of As mentioned in the preceding section cholesterol-bound saccharides have a dramatic influence on the pitch length in translating host–guest interactions into readable outputs.Gelation has provided us with a new medium in which we can mixed cholesteric liquid crystals.11 Addition of 2 or 3 mol% of the 251 saccharide complex to a composite liquid crystal further explore such interactions.We met with an early success; the sol–gel phase transition temperature (T gel ) can be con- membrane alters the pitch length in a direction relative to the absolute configuration of the complexed saccharide.This trolled by both metal cations and photo-isomerisation of the azobenzene moiety.35,36 For example, 8b can gelatinize a mixed change could be read-out by eye as a colour change in the liquid crystal.11 It occurred to us that these complexes might solvent of methylcyclohexane–benzene (451 v/v).Since alkali metal and ammonium salts are scarcely soluble in the gelation act as gelators of organic solvents. If so, one can obtain a variety of ‘chiral’ gelators by just changing the saccharide solvent, these metal salts should be solubilised through complexation with the crown ring.The T gel in the presence of Li+, source. The gelation test was carried out as follows: the complexes (0–5 mass%) were mixed with solvent in a septum- Na+, K+, Rb+ and NH4+ rose with increasing metal or ammonium concentrations whereas that in the presence of capped test tube and the mixture was heated until the solid was dissolved.37 The solution was cooled to room temperature Cs+ decreased.35,36 Conceivably, the Cs+ ion which tends to form a 152 metal–crown sandwich complex with 18-crown-6 (G in Table 2 denotes that a gel is formed at this stage).When a gel was not formed at room temperature, the solution was eVectively disorders the aggregation structure of 8b. On the other hand, a variety of azobenzene-containing cholesterol cooled in a refrigerator (at -6 °C) for one day (Gc in Table 2 denotes that a gel was formed at this stage).The results are derivatives (9, 10R and 11R) were synthesized and some of them were shown to be useful for photo-control of the sol–gel summarised in Table 2.37 Inspection of Table 2 reveals an interesting general trend: optical pairs of saccharides lie at phase transition.35,36 For example, butan-1-ol is gelatinized in the presence of only 0.1 mass% 10Me.When the gel was diVerent points along a path of increasing molecular interaction; the path from solution to crystallinity. From our photoirradiated by a high-pressure Hg-lamp through a VIScut colour glass filter (330<l<380 nm: the trans-to-cis iso- previous work with composite liquid crystal membranes red shift complexes increase the pitch length and blue shift com- merisation is induced), the gel phase was changed to the sol phase.The gel phase was re-generated when the sol was plexes decrease the pitch length of the liquid crystal membrane. Obviously, factors that work to increase the pitch length act photoirradiated through a UV-cut colour glass filter (l>460 nm: the cis-to-trans isomerisation is induced). Hence, to reduce the molecular interaction; likewise a decrease in the pitch length can be correlated with an increase in molecular the sol–gel phase transition could be controlled reversibly by a light switch (Scheme 2).cohesion. With carbon tetrachloride as solvent the following O O O O O N N OCH2CO O n CO O N N CO O N N CH3(CH2) nO CO O N N N CH3(CH2) n CH3(CH2) n 8a n = 1 8b n = 2 9 10Me 10Et 10Pr 10Bu 10Dec n = 0 n = 1 n = 2 n = 3 n = 9 11Me 11Et 11Pr n = 0 n = 1 n = 2 J.Mater. Chem., 1998, 8(3), 485–495 489Table 2 Results of the gelation tests carried out with monosaccharide complexesa lyxose xylose mannose glucose galactose D- L- D- L- D- L- D- L- D- Lsolvent blueb redb redb blueb blueb redb redb blueb blueb redb hexane SW SW I I I I I I I I benzene G G G G S R S I G G toluene G G G G I G S I G G dichloromethane G G G I I Gc S I G Gc chloroform G G G G Gc G S I S S carbon disulfide G G G SW G G I I S S diethyl ether I S SW I R S I I S S tetrahydrofuran S S S S I S I I S S 1,4-dioxane G S S G S S S S S S ethyl acetate S S R R S S R I S S acetone R S R R R R R R I R methanol R R R I I I I I I I ethanol S S S I R R R I R R a[Gelator]=0.1–5 mass%; Gel formed when cooled to 23 °C (G) or cooled in a refrigerator to -7 °C (Gc); Gel not formed because of crystallisation (R), soluble (S), or insoluble (I); SW=swollen.R means that the gelator is soluble at high temperature and precipitated at 23 °C whereas I means that the gelator is not soluble even at high temperature.bColour induced in a composite liquid crystal membrane. occurred: L-lyxose (gel ), D-lyxose (insoluble); D-xylose (gel ), L-xylose (insoluble); L-mannose (gel ), D-mannose (insoluble); D-glucose (solution) and L-glucose (recrystallises). For the chiral pairs of complexes the former are predicted to have weak cohesion from the liquid crystal work (red shift), and the latter strong intermolecular cohesion (blue shift).This prediction clearly holds under this set of conditions, and from further inspection of Table 2 it can be seen that there exists a general correlation between a blue–red rule in the liquid crystal system and a solubility rule in the gel system.37 With lyxose and xylose in certain solvents both isomers form gels, but the relative stability of the gel is not the same.The stability of the gel can be conveniently ascertained by plotting the sol–gel phase transition temperature versus mol% of the complex. It has become clear that L-xylose and D-lyxose have strong intermolecular interactions (blue shift in the liquid crystal system) whereas D-xylose and L-lyxose possess weak intermolecular interactions (red shift in the liquid crystal system).For further evaluation of the gel phase both scanning electron microscopy (SEM) and CD spectroscopy were performed on the gels formed with D- and L-xylose complexes. The CD spectra of the optical pairs are inverted, implying a diVerent chirality for the gels. The SEM pictures show that open fibrous gels are formed by these complexes; twists are also clearly visible but these structures are too large for individual fibrils and so must be the result of multiple copies intertwining (super helixes).Both the CD and SEM results confirm that the cholesterol moiety is performing a similar Fig. 4 (a) Linear dichroism (LD) spectra, (b) CD spectra and (c) task in the gel and liquid crystal system, holding together a absorption spectra for 10Me (0.2 mass%) in n-butanol: (i) the gel state was observed at 25 °C and (ii ) the solution state was observed chiral helix.As a corollary, this work represents a rare example at 60 °C of monosaccharide gelation. Also, the gelation ability of the complexes is controlled in a predictable manner by the chirality of the monosaccharide. This description of saccharide chirality is general since it is confirmed by the predictions made in the CD spectrum shows a positive exciton coupling which consists of the positive first Cotton eVect and the negative liquid crystal system.37 Through these studies we noticed that the CD spectra in second Cotton eVect, and the wavelength (310 nm) at [h]=0 agrees with the lmax of the absorption maximum obtained in the sol–gel transition system show very strange behaviour.We thus decided to investigate thoroughly the possible the gel phase. In the cholesterol gel systems (8, 10R, 11R, 12R and 13R), the solutions were totally CD-silent in the sol relationship between the gelation properties and the CD spectra.36,38 As shown in Fig. 4, the spectra for 10Me are phase whereas the clear exciton coupling band appeared in the gel phase.These results suggest that in the gel phase quite diVerent above and below T gel: the absorption maximum in the isotropic solution phase (360 nm) shifts to shorter cholesterol moieties aggregate in a specific, chiral direction, which enforces azobenzene chromophores to interact in an wavelength (high energy region) in the gel phase (310 nm).This means that the excited state of the azobenzene chromo- asymmetric manner. The drastic spectral changes are induced by the sol–gel phase transition, which can be readily phore is destabilised by molecular aggregation. Interestingly, the strong CD spectrum appears only in the gel phase. The ‘read-out’ as a change of CD spectra. 490 J. Mater. Chem., 1998, 8(3), 485–495CO O N N CH3(CH2) nO CO O N N N CH3(CH2) n CH3(CH2) n 12Me 12Et 12Pr 12Bu 12Pent 12Dec n = 0 n = 1 n = 2 n = 3 n = 4 n = 9 13Me 13Et 13Pr n = 0 n = 1 n = 2 The spectral features which are the sign of CD spectra or the (R)-chirality.In other words, the gel with (R)-chirality is a metastable phase formed by rapid cooling. shifts of the absorption spectra are influenced by specific structural features, such as the absolute configuration at C-3 of a The results support the view that azobenzene-appended cholesterol-based gelators aggregate in an asymmetric manner.steroidal skeleton. The absolute configuration at C-3 of 8, 10R and 11R is retained as the natural cholesterol b-configuration In particular, the appearance of the exciton coupling band in CD spectroscopy substantiates the fact that these aggregates [(S)-chirality], whereas that of 12R and 13R is the inverted aconfiguration [(R)-chirality].In general, the natural (S)-C-3 possess a helical structure. As expected, from SEM or optical microscopy, the fibrillar aggregates, which include a helical derivatives have a tendency to give the positive exciton CD spectra independent of solvents, and the lmax of the absorption structure, can be observed in the gel [Fig. 6(a)], and the crystals which were grown from the gel phase also include a spectra shifts to the high energy region by the sol–gel phase transition. On the other hand, the inverted (R)-C-3 derivatives clear helical structure [Fig. 6(b)]. Interestingly, the chirality of the CD spectra agreed with the chirality of the helical structure induce no energy shifts in the absorption spectra, but the sign of the CD spectra is dependent on the solvent.The positive observed by microscopy. We now consider that cholesterolbased gels are constructed by one-dimensional helical stacking. exciton coupling was favoured in polar solvents whereas the negative exciton coupling was favoured in nonpolar solvents. Recently, molecular recognition using a complementary hydrogen-bonding network between hosts and guests has Fig. 5 shows energy-minimised structures of 10Me and 12Me (the a epimer of 10Me). 12Me with the inverted configuration become an active area of endeavour. This concept has been at C-3 adopts an L-shaped bent conformation, whereas 10Me with the natural configration at C-3 adopts an extended conformation which is apparently more aggregative than that of 12Me. In the gel system, the natural derivatives showed a much stronger cohesive nature than the inverted derivatives (in melting point, liquid crystallinity, solubility and T gel).36 It is also expected that this structural diVerence would be reflected by the diVerence in the spectral properties.Interestingly, the sign of the CD spectra is sometimes inverted depending upon the cooling speed of gelatinisation. In the 13Me–methylcyclohexane gel, for example, the cooling of the solution (60 °C) in water at 30 °C (slow cooling) gave a negative exciton CD spectrum whereas the cooling in water at 2 °C (rapid cooling) gave a positive exciton CD spectrum. When the gel with (R)-chirality was gradually heated, the sign changed from (R) to (S) at around 27 °C.When the gel with (S)-chirality was gradually cooled to -10 °C, the h values were increased in the negative direction and inversion of the sign did not take place. The results indicate that, basically, the 13Me gel with (S)-chirality is more stable than that with Fig. 6 (a) SEM picture of the dried gel for a 13Me–cyclohexane system: the gel was prepared by slow cooling: the fibrils have a left-handed helical structure whereas the CD showed a negative exciton coupling Fig. 5 Energy-minimised structures of (a) 10Me with (S)-chirality and band in this gel system. (b) Optical microscope picture of the fibrous crystal for 10Me: the crystal was grown from the 10Me–n-octanol gel.(b) 12Me with (R)-chirality J. Mater. Chem., 1998, 8(3), 485–495 491applied to the formation of stable organic gels.39–43 It seemed recognition.45–47 Of particular interest is the potential application to chiral discrimination: chiral guest molecules in the that if the hydrogen-bonding interaction acts cooperatively with the cholesterol–cholesterol interaction, the synergistic subphase interact with chiral amphiphilic compounds forming the monolayer and the resultant ‘diastereomeric complexes’ eVect may create ‘super-gelators’ which can gelatinize organic solvents at very low gelator concentrations.To test this idea change the surface-pressure (p–A) isotherm in an asymmetric manner.48,49 It was recently demonstrated that certain choles- compounds 14 and 15 were synthesized.44 Although 14 could gelatinize several organic solvents by itself, 15 always resulted terol derivatives (such as 1, 2 and 6a) form beautiful monolayers. 50 Judging from the p–A curves, the crown ether moieties, in precipitates from these solvents. In contrast, a 151 14+15 mixture gave reinforced organic gels because of the action of which tend to spread on the water surface and adopt a parallel orientation, stand up gradually with increasing surface pressure their complementary hydrogen-bonding sites.44 It was shown, however, that the IR spectrum of the 14+15 gel is very and are finally packed as a monolayer with a vertical orientation.Interestingly, the p–A isotherms responded specifically complicated (e.g. there are many NH2 and CNO stretching bands).When the mixture was thoroughly dissolved and then to added alkali metal cations because of the metal–crown interaction.50 cooled very slowly, it gave a crystalline precipitate, the IR spectrum of which was more simplified than that of the gelled It occurred to us that this system may be applicable to chiral discrimination of a-amino acid derivatives because the mixture.We now consider, therefore, that the gelled mixture does not involve a neat tape structure (as shown in Fig. 739c) 18-crown-6 moiety in 1b should bind the NH3+ group also at the air–water interface. Addition of D-PheOMe HCl (D-phenyl- but a partially disordered, kinetically stable structure. alanine methyl ester hydrochloride) to a subphase under a monolayer of 1b induced a large expansion of the monolayer; when L-PheOMe HCl was added only a small expansion occurred.Chiral discrimination was also achieved for other aamino acid derivatives (Table 3), but the largest eVect was observed with PheOMe HCl. Why is such chiral discrimination of a-amino acid derivatives, which is diYcult with conventional chiral amphiphilic compounds, readily realised in a cholesterol-based monolayer system? It is certain that the NH3+ moiety is bound to the 18-crown-6 ring.As the a-amino acid residue CH2R is more hydrophobic than the CO2Me group, this group should be trapped in the hydrophobic cholesterol stacks. In this binding mode, the cholesterol skeleton with a wide chiral plane can more advantageously constrain the orientation of a-amino acid derivatives than conventional chiral amphiphiles with simple point chirality.Thus, the a-amino acid derivatives are recognised at two points (NH3+ by the crown ring and CH2R by the cholesterol plane) O (CH2)6 N N N O O (CH2)6 N N NH2 O H O H H2N N H H 14 15 by a monolayer of 1b. Organic gels have been considered to be troublesome systems We previously found that cholesteric liquid crystals containto which neither conventional recrystallisation nor spectroing 1b can asymmetrically recognise a-amino acid derivatives: scopic analysis can be applied. As reviewed in this section, L-isomers stabilise the liquid crystal phase to shorten the pitch however, prerequisites for organic gel formation have been length whereas D-isomers destabilise the liquid crystal phase elucidated and it has been shown that gelators assemble into to elongate the pitch length.10 Although the monolayer phase fibrous aggregates with an oriented structure in organic solis diVerent from the liquid crystal phase, the microscopic vents, reflecting the molecular shape and the intermolecular environments where one amino acid residue is flanked by two interaction mode: chiral gelators such as cholesterol derivatives cholesterol planes should be similar to each other.Important tend to form helical structures and hydrogen-bonding gelators is the fact that in both systems D-isomers disorder the cholestend to aggregate so that they can satisfy complementary terol-based oriented phases more eYciently than L-isomers. donor–acceptor interactions.We believe that gel systems are Presumably, the space formed between two cholesterols fits an under-exploited field which has great potential for molecular the asymmetric shape of L-isomers more complementarily than recognition and switch functions and by which one can create that of D-isomers. a number of new three-dimensional molecular assemblies. The relationship between molecular structure and physical Cholesterol derivatives would play a central role in such gel properties of monolayers of complexes between an amphiphilic, systems because of their versatility.cholesterol-substituted phenylboronic acid 3 and monosaccharides were studied at the air–water interface.51 Phase Monolayer systems transition, compressibility and limiting molecular area of monolayers of 3 in the presence of monosaccharides are Recently, great interest has developed around the application correlated with the calculated structures of the phenylboronic of monolayers formed at the air–water interface to molecular acid–monosaccharide complexes.The monolayer of 3 exhibits Table 3 Limiting area A0 and lift-oV area A1 of 1b on 100 mM amino acid methyl ester hydrochlorides A0/nm2 molecule-1 A1/nm2 molecule-1 amino acid L-isomer D-isomer L-isomer D-isomer Phe 0.72 0.91 1.06 1.44 N N N N N H H H H O N N H O N O H H O N N Ch O N H O H Ch Ch H Ala 0.82 0.79 1.20 1.44 Fig. 7 Tape structure with complementary hydrogen-bonding inter- Trp 0.81 0.84 1.34 1.33 actions between a triaminopyrimidine unit and an isocyanuric acid: Val 0.71 0.77 1.04 1.14 Ch denotes a cholesterol moiety 492 J.Mater. Chem., 1998, 8(3), 485–495Table 4 Water-facing angles of the complexes between 3 and mono- chiral discrimination towards optical isomers of monosaccharsaccharides in the pyranose or furanose form for the MO-optimised ides.51 As explained above for liquid crystals of 3–saccharide structures complexes, most monosaccharides form 251 complexes with the pyranose form in non-aqueous media (see Scheme 1), from calculated calculated which the Ph–Ph angle for the most stable structures was angle of angle of binding pyranose binding furanose theoretically computed.11 In this system the boron atom is sp2- saccharide site complexes site complexes hybridised.In contrast to that in non-aqueous systems, monosaccharides preferably exist as furanoses in boronic acid xylose 1,2–3,4 153.5 1,2–3,5 93.2 complexes in aqueous alkaline bulk solution, where the boron fructose 2,3–4,5 143.3 1,2–4,6 115.3 atom is sp3-hybridised.52 Furthermore, it is uncertain whether galactose 1,2–4,6 134.6 1,2–5,6 132.3 3 forms similar 251 complexes. To solve these complex mannose 2,3–4,6 131.6 2,3–5,6 166.7 arabinose 1,2–3,4 126.9 1,2–3,5 103.5 problems at the air–water interface we first determined the fucose 1,2–3,4 123.3 1,2–3,5 100.9 stoichiometry by using a mass spectrometer. A built-up film glucose 1,2–4,6 95.8 1,2–5,6 145.5 of 80 layers of an amphiphilic phenylboronic acid on a allose 2,3–4,6 95.6 2,3–5,6 119.1 subphase containing D-glucose at pH 11 was prepared by vertical dipping of a steel plate. The mass spectrum of the sample thus collected supported the double-charged 251 complex (Fig. 8).53 We could not solve the furanose vs. pyranose monolayers of complexes with a small water-facing angle a such as the glucose complex require a lower surface pressure problem, however. For the present study, therefore, the following two series of calculations were taken into account for to achieve the phase transition and to rearrange both the structure of the complex and packing in the monolayer com- model compounds of all experimentally studied monosaccharide complexes of 1 in order to determine their water-facing pared with complexes with a larger angle a, e.g.with fructose. Alternatively, we also tried to correlate the transition pressure angle (Fig. 9): (i) sp2-hybridised boron and pyranoses, because these complexes are most stable in non-aqueous solvents and with the angle for the complexes with furanose-type monosaccharides with the boron atom sp3-hybridised,51 because in the air–water interface provides an environment distinct from the aqueous bulk solution and (ii) sp3-hybridised boron atom aqueous bulk solution monosaccharides adopt this form in boric acid complexes.52 However, there is no correlation and furanoses, because these complexes are more stable in aqueous alkaline solution and the pH of the subphase is higher between transition pressure and the water-facing angle for the furanose complexes.The diVerence in correlation suggests that than pKa (pKa ca. 9). The calculations are shown in Table 4.It is seen from Fig. 10 that the complexes are clearly divided the pyranose form is predominant under the experimental conditions. The monolayer must provide an environment into two groups according to their structure. Those causing a red shift in the cholesteric liquid crystal system and destabilis- unlike the bulk aqueous solution. This finding is in agreement with the previously proposed reduced solvating capability of ing it (D-xylose, L-mannose, L-arabinose, D-fucose, D-glucose and D-allose) fit a linear plot in Fig. 10(a), while those causing water molecules near the air–water interface.54 The monolayer formation properties of cholesterol-based a blue shift in the same system and stabilising it (L-xylose, D-fructose, D-galactose, D-mannose, D-arabinose, L-fucose and azobenzene amphiphiles with the natural (S)-C-3 configuration (10Me) and inverted (R)-C-3 configuration (12Me) were exam- L-glucose) fit a linear plot in Fig. 10(b). A remarkable correlation between monolayer property and molecular structure is ined at the air–water interface.55 The close correlation between the gel system and the monolayer system was observed here observed: the smaller the angle a between two intramolecular cholesterol moieties, the lower the transition pressure.The again. The computational study reveals that 10Me adopts an extended conformation whereas 12Me adopts an L-shaped bent conformation.36 10Me gave an expanded phase with A0 ( limiting area)=0.60 nm2 molecule-1 and A1 ( lift-oV area)= 0.64 nm2 molecule-1 whereas 12Me gave a condensed phase with A0=0.49 and A1=0.54 nm2 molecule-1.Examination using reflectance spectroscopy established that 12Me forms a monolayer with a J-aggregation mode (lmax 407 nm) and with an increase in the compressibility it changes to an H-aggregation mode (lmax 336 nm). The finding indicates that the azobenzene moieties interact with each other although the –B O HO O H O H H HO CH2 H H O O B– OH O C16H33 O C16H33 aggregation mode is dependent on the compressibility.This Fig. 8 251 complex with sp3-hybridised borons formed at the air– spectroscopic character is very similar to a spectral change water interface. The complex structure is illustrated assuming the observed for the sol–gel phase transition of 10Me gels.36,38 pyranose form. On the other hand, 12Me forms a monolayer having a lmax (362 nm) comparable with the monomeric absorption maximum, indicating that the electronic interaction among the azobenzene moieties is absent.This trend is again very similar to the sol–gel phase transition of 12Me gels which does not show any significant spectral change.36,38 One can conclude, therefore, that the diVerence in the gel formation between 10Me and 12Me is well reproduced in the monolayer formation.The morphological changes in the monolayers were directly observed by optical microscopy and reasonably correlated with the spectroscopic changes. Octadecyl Rhodamine B (used as a fluorescent probe) is more miscible with 12Me ( less cohesive36) than with 10Me (more cohesive36). This diVerence is reflected by the p–A isotherms: the p–A isotherm for 12Me plus octadecyl Rhodamine B is not so diVerent from that for 12Me only, whereas the collapse pressure for 10Me plus Fig. 9 Relationship between the Ph–Ph angle and the water-facing angle a. octadecyl Rhodamine B is significantly lower than that for J. Mater. Chem., 1998, 8(3), 485–495 493Fig. 10 Plots of surface pressure of monolayers of 3 at phase transition vs.water-facing angle a of the complexes which destabilise or stabilise the cholesteric liquid crystal; $ at 293 K, and , at 303 K 10Me only. The domain structure was not observed immedi- phase transition, spectroscopic properties, etc. and (iii ) a slight change in the cholesterol structure leads to a large change in ately after spreading a benzene solution containing 12Me, but after a few minutes the domain structure appeared.From the aggregation properties, which is characteristic of molecular assembly systems. We have noticed that in these systems, it is reflectance spectroscopy we noticed that the reflectance intensity fluctuates with time above A1. Optical microscopy suggests essential to obtain an insight into the relation between the aggregation mode of the cholesterol moieties and that of the that the microaggregates of 12Me are drifting on the water surface, and the reflectance intensity increases when they come functional groups appended to their 3-OH.The packing mode of cholesterol derivatives in the crystal state has been studied into the vision field of the microscope. The findings support the view that 12Me also forms the microaggregates even at by several groups.56–60 However, the relation between the cholesterol structure and the packing mode is not yet well 0 mNm-1.At p=19.0 mN m-1 ( just below pc), a large, nearly homogeneous monolayer of 12Me was observable. At p>pc, correlated60 and to simply predict the aggregation mode of metastable and partially-disordered phases mentioned in this on the other hand, long stripes appeared in the homogeneous domain.These wrinkles are formed by a compression that is article from the X-ray-determined single-crystal structure is not appropriate. Probably, one has to find some alternative strong enough to collapse the monolayer. On the other hand, 10Me gave a number of bright spots above A1. As demon- approach to this problem. On the other hand, recent studies have shown that crystals derived from steroid derivatives can strated by reflectance spectroscopy, this pattern corresponds to the formation of microaggregates with a J-aggregation provide novel crystal lattices and inclusion compounds.61 This suggests that they are also useful as a part of molecular mode.When they are compressed to a monolayer, one can observe a well-dispersed island structure.This corresponds to tectonics.We believe that cholesterol-based molecular assembly systems can be further extended as novel molecular orientation an expanded monolayer phase with a J-aggregation mode. Further increase in the compression changed the island struc- systems to chiral molecular recognition, molecular switches, controlled electron or energy transfer systems, membrane ture to a network structure with large cracks. Conceivably, the monolayer is collapsed here and a crystal (or liquid crystal) separations, etc.with an H-aggregation mode is produced on the surface. The monolayer system is diVerent from the liquid crystal system and the gel system in the following two ways: it is an References event occurring in a two-dimensional domain and the param- 1 (a) E.B. Kyba, K. Koga, L. R. Sousa, M. G. Siegel and D. J. Cram, eters obtained from the condensed subphase more or less J. Am. Chem. Soc., 1973, 95, 2692; (b) Y. Chao and D. G. Cram, reflect those of the crystal phase. In cholesterol-based amphi- J. Am. Chem. Soc., 1976, 98, 1015; (c) D. J. Cram, R. C.Hegelson, philes, however, the parameters and their perturbation induced L. R. Sousa, J. M. Timko, M. Newcomb, P. Moreau, F. de Jong, G. W. Gokel, D. H. HoVman, L. A. Domier, S. C. Peacock, by guests added into the subphase are surprisingly similar to K. Madan and L. Kaplan, Pure Appl. Chem., 1975, 43, 327; each other. The results imply that the aggregation mode in (d) M. Newcomb, J.L. Toner, R. C. Helgeson and D. J. Cram, these systems is primarily governed by a cholesterol moiety J. Am. Chem. Soc., 1979, 101, 4941; (e) J.-M. Lehn and C. Sirlin, which inherently possesses an aggregative tendency, whereas J. Chem. Soc., Chem. Commun., 1978, 143; ( f ) J.-M. Lehn, Science, appended functional groups play only a secondary role in the 1985, 227, 849. structural orientation. 2 For applications of steroidal compounds in supramolecular chemistry see: (a) A. P. Davis, Chem. Soc. Rev., 1993, 22, 243 (and references cited therein); (b) R. P. Bonar-Law and J. K. M. Sanders, Conclusions J. Am. Chem. Soc., 1995, 117, 259; (c) H.-P. Hsieh, J. G. Muller and C. J. Burrows, J. Am. Chem. Soc., 1994, 116, 12077; (d) Y. A. Cheng, The main purpose of the present feature article is to T.Suenaga and W. C. Still, J. Am. Chem. Soc., 1996, 118, 1813; demonstrate that cholesterol derivatives (or more generally, (e) V. Janout, M. Lanier and S. L. Regen, J. Am. Chem. Soc., 1996, steroids) act as versatile building-blocks commonly useful for 118, 1573; ( f ) U. Maitra and L. J. D’Souza, J. Chem. Soc., Chem. Commun., 1994, 2793; (g) P. Venkatasan, Y.Cheng and D. Kahne, the formation of liquid crystals, organic gels and monolayers. J. Am. Chem. Soc., 1994, 116, 6955; (h) A. P. Davis and J. J. Walsh, Through these studies it has been shown that (i ) cholesterol J. Chem. Soc., Chem. Commun., 1996, 449; (i ) A. P. Davis, can provide chiral and moderately rigid platforms for molecu- S. Menzer, J. J. Walsh and D. J. Williams, Chem.Commun., 1996, lar orientation and molecular recognition, (ii ) although the 453. aggregation modes are restricted by boundary conditions 3 For example, see (a) T. Kaneda, K. Hirase and S. Misumi, J. Am. inherent to each phase, one can raise many correlation relation- Chem. Soc., 1989, 111, 742; (b) K. Yamamoto, K. Isoue, Y. Sakata and T. Kaneda, J. Chem. Soc., Chem. Commun., 1992, 791; ships among diVerent phases, e.g.in guest selectivity, cohesivity, 494 J. Mater. Chem., 1998, 8(3), 485–495(c) Y. Kubo, S. Maeda, S. Tokita and M. Kubo, Nature, 1996, 31 (a) P. Terech, J. Colloid Interface Sci., 1985, 107, 244; (b) R. H. Wade, P. Terech, E. A. Hewat, R. Ramasseul and F. Volino, 382, 522. 4 For chiral recognition in liquid crystal systems see: (a) C. Eskenazi, J.Colloid Interface Sci., 1986, 114, 442. 32 P. Terech, L iquid Crystals, 1991, 9, 59. J. F. Nicoud and H. B. Kagan, J. Org. Chem., 1979, 44, 995; (b) F. D. Saeva, P. E. Sharpe and G. R. Olin, J. Am. Chem. Soc., 33 (a) T. Brotin, R. Utermo� hlen, F. Fages, H. Buas-Laurent and J.-P. Desvergne, J. Chem. Soc., Chem. Commun., 1991, 416; 1975, 97, 204. For regioselectivity control in liquid crystal systems see: (c) J.M. Nerbonne and R. G. Weiss, J. Am. Chem. Soc., 1978, (b) T. Brotin, J.-P. Desvergne, F. Fages, R. Utermo� hlen, R. Bonneau and H. Bouas-Laurent, Photochem. Photobiol., 1992, 100, 2571; (d) T. Nagamatsu, C. Kawano, Y. Orita and T. Kunieda, T etrahedron. L ett., 1978, 28, 3263; (e) W. J. Leigh and D. S. 55, 349. 34 For a comprehensive review see T.D. James, H. Kawabata, Mitchell, J. Am. Chem. Soc., 1988, 110, 1311. 5 For the use of thermotropic liquid crystals as solvents see; R. G. R. Ludwig, K. Murata and S. Shinkai, T etrahedron, 1995, 51, 555. 35 K. Murata, M. Aoki, T. Nishi, A. Ikeda and S. Shinkai, J. Chem. Weiss, T etrahedron, 1988, 44, 3413. 6 G.-X. He, F. Wada, K. Kikukawa, S. Shinkai and T.Matsuda, Soc., Chem. Commun., 1991, 1715. 36 K. Murata, M. Aoki, T. Suzuki, T. Harada, H. Kawabata, J. Org. Chem., 1990, 55, 541, 548. 7 (a) G. W. Gokel, J. C. Hernandez, A. M. Viscariello, K. A. Arnold, T. Komori, F. Ohseto, K. Ueda and S. Shinkai, J. Am. Chem. Soc., 1994, 116, 6664. C. F. Champana, L. Echegoyen, F. R. Fronczek, R. D. Gandour, C. R. Morgan, J. E. Trafton, S. R. Miler, C.Minganti, D. Eiband, 37 T. D. James, K. Murata, T. Harada, K. Ueda and S. Shinkai, Chem. L ett., 1994, 273. R. A. Schultz and M. Tamminen, J. Org. Chem., 1987, 52, 2963; (b) L. E. Echegoyen, L. Portugal, S. R. Miller, J. C. Hernandez, 38 K. Murata, M. Aoki and S. Shinkai, Chem. L ett., 1992, 739. 39 (a) K. Hanabusa, K. Okui, K. Karaki, T. Koyama and H. Shirai, L. Echegoyen and G.W. Gokel, T etrahedron L ett., 1988, 29, 4065; (c) L. E. Echegoyen, J. C. Hernandez, A. E. Kaifer, G. W. Gokel J. Chem. Soc., Chem. Commun., 1992, 1371; (b) K. Hanabusa, M. Yamada, M. Kimura and H. Shirai, Angew. Chem. Int. Ed. and L. Echegoyen, J. Chem Soc., Chem. Commun., 1988, 836. 8 S. Shinkai, T. Nishi, A. Ikeda, T. Matsuda, K. Shimamoto and Engl., 1996, 35, 1949; (c) K.Hanabusa, T. Miki, Y. Taguchi, T. Koyama and H. Shirai, J. Chem. Soc., Chem. Commun., 1993, O. Manabe, J. Chem. Soc., Chem. Commun., 1990, 303. 9 S. Shinkai, T. Nishi and T. Matsuda, Chem. L ett., 1991, 437. 1382. 40 (a) N. A. J. M. Sommerdijk, M. H. L. Lambermon, M. C. Feiters, 10 T. Nishi, A. Ikeda, T. Matsuda and S. Shinkai, J. Chem. Soc., Chem. Commun., 1991, 339. R. J. M.Nolte and B. Zwanenburg, Chem. Commun., 1997, 455; (b) R. J. H. Hafkamp, B. P. A. Kokke, I. M. Danke, H. P. M. 11 T. D. James, T. Harada and S. Shinkai, J. Chem. Soc., Chem. Commun., 1993, 857 and 1176 (corrigendum) Geurts, A. E. Rowan, M. C. Feiters and R. J. M. Nolte, Chem Commun., 1997, 545. 12 (a) S. Shinkai, K. Tsukagoshi, Y. Ishikawa and T. Kunitake, J. Chem. Soc., Chem. Commun., 1991, 1039; (b) K.Tsukagoshi and 41 E. J. de Vries and R. M. Kellogg, J. Chem. Soc., Chem. Commun., 1993, 238. S. Shinkai, J. Org. Chem., 1991, 56, 4089; (c) K. Kondo, Y. Shiomi, M. Saisho, T. Harada and S. Shinkai, T etrahedron, 1992, 48, 8239. 42 M. Takafuji, H. Ihara, C. Hirayama, H. Hachisako and K. Yamada, L iquid Crystals, 1995, 18, 97. 13 (a) G.WulV, B. Heide and G.Helfmeier, J. Am. Chem. Soc., 1986, 108, 1089; (b) G. WulV and H.-G. Poll, Makromol. Chem., 1987, 43 J.-E. Sohna and F. Fages, Chem. Commun., 1997, 327. 44 S. W. Jeong, K. Murata and S. Shinkai, Supramol. Sci., 1996, 3, 83. 188, 741. 14 (a) T. Kajiyama, H. Kikuchi and S. Shinkai, J. Membr. Sci., 1988, 45 (a) K. Kurihara, K. Ohto, Y. Tanaka, Y. Aoyama and T. Kunitake, J. Am. Chem.Soc., 1991, 113, 444; (b) K. Kurihara, K. Ohto, 36, 243; (b) T. Kajiyama, S. Washizu, A. Kumano, I. Terada, M. Takayanagi and S. Shinkai, J. Appl. Polym. Sci., Appl. Polym. Y. Honda and T. Kunitake, J. Am. Chem. Soc., 1991, 113, 5077; (c) Y. Ikeura, K. Kurihara and T. Kunitake, J. Am. Chem. Soc., Symp., 1985, 41, 327. 15 (a) S. Shinkai, S. Nakamura, S. Tachiki, O. Manabe and 1991, 113, 7342. 46 Y.Ishikawa, T. Kunitake, T. Matsuda, T. Otsuka and S. Shinkai, T. Kajiyama, J. Am. Chem. Soc., 1985, 107, 3363; (b) S. Shinkai, S. Nakamura, K. Ohara, S. Tachiki, O. Manabe and T. Kajiyama, J. Chem. Soc., Chem. Commun., 1989, 1937. 47 H. Kawabata and S. Shinkai, Chem. Expr., 1993, 8, 765. Macromolecules, 1987, 20, 21. 16 (a) S. Shinkai, K. Torigoe, O. Manabe and S. Shinkai, J. Chem. 48 (a) E. M. Arnett, N. G. Harvey and P. L. Rose, Acc. Chem. Res., 1989, 22, 131; (b) N. G. Harvey, D. Mirajovsky, P. L. Rose, Soc., Chem. Commun., 1986, 933; (b) J. Am. Chem. Soc., 1987, 109, 4458. R. Verbiar and E. M. Arnett, J. Am. Chem. Soc., 1991, 111, 1115. 49 P. Qian, M. Matsuda and T. Miyashita, J. Am. Chem. Soc., 1993, 17 S. Shinkai, K. Shimamoto, O. Manabe and M. Sisido, Makromol. Chem., Rapid Commun., 1989, 10, 361. 115, 5624. 50 T. Nishi and S. Shinkai, Chem. Expr., 1993, 8, 173. 18 S. Shinkai, G.-X. He, T. Matsuda, K. Shimamoto, N. Nakashima and O. Manabe, J. Polym. Sci., Polym. L ett. Ed., 1989, 27, 209. 51 R. Ludwig, T. Harada, K. Ueda, T. D. James and S. Shinkai, J. Chem. Soc., Perkin T rans. 2, 1994, 697. 19 S. Shinkai, T. Nishi, K. Shimamoto and O. Manabe, Is. J. Chem., 1992, 32, 121. 52 (a) S. Chapelle and J.-F. Verchere, T etrahedron, 1988, 44, 4469; (b) J. C. Norrild and H. Eggert, J. Am. Chem. Soc., 1995, 117, 1479. 20 U. F. Kragten, M. F. M. Roks and J. M. Nolte, J. Chem. Soc., Chem. Commun., 1985, 1275. 53 R. Ludwig, K. Ariga and S. Shinkai, Chem. L ett., 1993, 1413. 54 Y. Honda, K. Kurihara and T. Kunitake, Chem. L ett., 1991, 681. 21 S. B. Hladky and D. A. Haydon, Biochim. Biophys. Acta, 1972, 274, 294. 55 H. Kata, K. Murata, T. Harada and S. Shinkai, L angmuir, 1995, 11, 623. 22 M. Sirai, T. Orikata and M. Tanaka, J. Polym. Sci., Polym. Chem. 56 B. M. Craven and G. T. DeTitta, J. Chem. Soc., Perkin T rans. 2, Ed., 1985, 23, 463. 1976, 814. 23 J. Donis, J. Grandjean, A. Grosjean and P. Lazlo, Biochim. 57 N. G. Guerina and B. M. Craven, J. Chem. Soc., Perkin T rans. 2, Biophys. Acta, 1981, 102, 690 and references cited therein. 1979, 1414. 24 H. Tokuhisa, K. Kimura, M. Yokoyama and S. Shinkai, J. Chem. 58 H.-S. Shieh, L. G. Hoard and C. E. Nordman, Acta Crystallogr. Soc., Faraday T rans., 1995, 91, 1237. Sect. B., 1981, 37, 1538. 25 Y.-C. Lin and R. G. Weiss, Macromolecules, 1987, 20, 414. 59 H.-P. Weber, B. M. Craven and P. Sawzik, Acta. Crystallogr. Sect. 26 Y.-C. Lin, B. Kachar and R. G. Weiss, J. Am. Chem. Soc., 1989, B., 1991, 47, 116. 111, 5542. 60 E. Otsuni, P. Kamaras and R. G. Weiss, Angew. Chem., Int. Ed. 27 Y.-C. Lin and R. G. Weiss, L iquid Crystals, 1989, 4, 367 and Engl., 1996, 35, 1324. references cited therein. 61 For a comprehensive review see M. Miyata and K. Sada, 28 R. Mukkamala and R. G. Weiss, J. Chem. Soc., Chem. Commun., Comprehensive Supramolecular Chemistry, ed. J.-M. Lehn, 1995, 375. Pergamon Press, London, 1996, p. 147. 29 E. Ostuni, P. Kamaras and R. G. Weiss, Angew. Chem., Int. Ed. Engl., 1996, 35, 1234. 30 L. Lu and R. G.Weiss, L angmuir, 1995, 11, 3630. Paper 7/04820C; Received 7th July, 1997 J. Mater. Chem., 1998, 8(3), 485–495 495

ISSN:0959-9428    

DOI:10.1039/a704820c

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Feature Article Conducting polymer image formation with photoinduced electron transfer reaction Norihisa Kobayashi,*a Kenjiro Teshimab and Ryo Hirohashia aDepartment of Image Science, Chiba University, Chiba 263, Japan bGraduate School of Science and T echnology, Chiba University, Chiba 263, Japan Conducting polymers are promising materials which show attractive electric and optical properties with potential importance in advanced technologies.Electrochromism is one of the noteworthy characteristics of conducting polymers and is eVective in forming conducting polymer images and patterns on an electrode. Space selective image formation is also possible without an electrode by means of the combination of photoinduced electron transfer reaction with electrochromism.This has potential as a novel method of surface modification for any geometry in any place and for micropatterning. In this article, some recent aspects of (1) the electrochromism of conducting polymers and (2) the polymerization and electrochromism of conducting polymers induced by photoillumination leading to image formation are considered from the materials viewpoint.Further, possible future applications for these materials and systems are commented upon. Traditional photography plays an important role in our life depending on doping and dedoping. Doping a conducting polymer makes the oscillator strength, associated with the for storing momentary dynamic and transient information, and printing enables us to make large numbers of copies.valence band–conduction band transition, shift into the free carrier of the visible and infrared region. Therefore, if the These technologies have now been extended to digital imaging, 3D holography, lithography for memory devices and related energy gap between the valence and the conduction band is large enough, doping will be accompanied by a dramatic information recording systems.Common to all, is that photoenergy is utilized as a power source. Image formation materials spectral change. This phenomenon is particularly induced by electrochemical stimuli and is called electrochromism. thus play an important role, and high sensitivity and high resolution are required to convert photoenergy eVectively and However, electrochromism (even for inorganic materials) has not penetrated the display market yet despite having better to produce high contrast images.In the conventional photographic process, the image is composed of silver converted color variation, wider spectral range, and larger viewing angle and memory eVects than liquid crystals. The slower response from silver halide by photoreduction and the developing process.The quality of images depends on the characteristics times in electrochromism due to mass (ion) transport eVects limit its application. of the image forming material. Silver in photography, ink in printing and photopolymers in lithography provide excellent Recently, a potential target for the application of electrochromism has been directed toward curtainless smart images, but the resulting image lacks additional functionality.It is a passive image. If an image could be composed of a windows for heat and daylight control, anti-dazzling rear-view mirrors for vehicles, solar panel technology etc. Such kinds of functional material such as a conducting polymer with optical and electrical properties and be formed by photoillumination, electrochromism are also assumed to be applicable to materials for image formation because fast response is not a prerequisite.it would open up a wide range of applications leading to organic molecular systems with an active image whose activity In this feature article, some recent aspects of electrochromism in conducting polymers will be considered first. Then, polymer- might be controlled by external stimuli after image formation (Fig. 1). ization and electrochromism of conducting polymers induced by photoillumination leading to image formation will be It was discovered two decades ago that shiny poly(acetylene) could be prepared with dramatically diVerent conductivity considered from a materials viewpoint. The formation of images with tunable functions (including electrochromism) by depending on doping with oxidizing or reducing agents.1 Numerous other conducting polymers were subsequently illumination will also be discussed from the viewpoint of active images.developed.2 Some were commercially produced and were applied to electronic devices. Particularly, batteries containing conducting polymers as the cathodic material were marketed.Conducting polymers also have great interest because of their optical properties. Changes in the electronic structure resulting from electrochemical doping and dedoping results in diVerent absorption spectra, which could be applied for passive display materials.3 Recently, conducting polymers having certain electronic characteristics have become strong candidates for active displays such as organic electroluminescent devices.4 Attention has been focused increasingly on exploiting this new class of conducting polymers for commercial flat panel displays.Changes in conductivity of eight orders of magnitude (from semiconductor to conductor or vice versa) and dramatic color changes with electrochemical stimuli can be achieved photoillumination photoenergy conversion system polymerization redox reaction conducting polymer image with some functions active image Fig. 1 Pathways to form conducting polymer image with some functions * E-mail: norichan@tcom.tech.chiba-u.ac.jp J. Mater. Chem., 1998, 8(3), 497–506 497without serious damage to the polymer.14 This electrochromic Electrochromism in conducting polymers characteristic seems to be better than that of poly(pyrrole).Electrochromism is one method of image formation and has Poly(thiophene) has other advantages relating to chromic been investigated mainly among inorganic materials. Inorganic properties. Substitution of alkyl groups on the thiophene ring materials show better electrochromic characteristics and gave a polymer soluble in organic solvents on reduction.reliability and durability although the organic electrochromic Solutions containing poly(3-alkylthiophene) showed solvamaterials are gradually improving. Tungsten trioxide (WO3) tochromism as well as electrochromism.15 based electrochromic devices have already been applied to Further, poly(3-alkylthiophene) enables Langmuir–Blodgett sunglasses and camera viewfinders (Nikon F5).However, why films to be prepared when mixed with stearic acid.16 This have such materials not become more familiar in everyday possibility is useful in preparing thin films for chromic devices. life? It might be due to the slower response times, higher cost, In order to prepare pin-hole free flexible thin films and to and so on. Color variation is also a problem in inorganic narrow the absorption band, oligothiophene of well-defined electrochromic materials.structure was substituted on a non-conjugated vinyl polymer In contrast, organic materials have the advantage in that as side chain pendant, and its electrochromic characteristics color variation due to easier molecular design can be produced were studied.17 This polymer underwent a reversible clear color over a wider range.Chemical substitution of p-conjugated change from green to yellow and vice versa in contrast to groups or other functional groups in a substance is one common poly(thiophene) with its electrochromic color change approach to realizing this kind of modification and to adjusting from red to blue. This kind of p-conjugated system pendant properties. This approach is particularly eVective in organic polymer is expected to lead to a new family of organic conducting polymers because the electronic state of their large p-electron systems.p-conjugated polymer backbone system is very sensitive to The electrochromic characteristics of poly(aniline) were chemical or physical modifications. Further, by utilizing con- pointed out at an early stage,18 but work was only intensively ducting polymers, conventional and simple film manufacturing begun since 1984.19 One of the most attractive advantages for process can be applied in device fabrication because polymers poly(aniline) as compared with the above mentioned conare more flexible and processible than low molecular mass ducting polymers is color variation.Poly(aniline) showed materials.On this basis, many kinds of electrochromic con- multiple color changes from pale yellow to green, to blue, to ducting polymers have been reported. These are broadly deep purple at-0.2 to 1.0 V (vs. SCE).19 This electrochromic divided into three categories; (1) electrochromic conducting behavior was unstable on repetition over this potential range, polymers, (2) electrochromic conducting polymers with elec- but was very stable with repetition of more than 106 times if trochromic dopants, (3) electrochromic conducting polymers the potential range was limited to that causing a color change with inorganic electrochromic materials.from pale yellow to green or vice versa. This performance is better than other conducting polymers. Poly(aniline) was commonly prepared via chemical or Electrochromic conducting polymers electrochemical oxidation in acidic solution.Electrochromic characteristics were also evaluated in acidic solution. Problems The electrochromic characteristics of conducting polymers were reported for poly(N-methylpyrrole) by Diaz et al. in 19815 and using acidic aqueous solution have been reported, for example degradation of poly(aniline) during polarization with wide extensive work has been carried out since then.Initially, target conducting polymers were focused on poly(pyrrole), poly(ani- potential ranges in acidic aqueous solution.20 The preparation and characterization of poly(aniline) and its derivatives in non- line), poly(thiophene) and their derivatives. For poly(pyrrole), it was found that oxidation of pyrrole led to a black powder aqueous solution without a proton donor have already been reported,21–24 and similar color changes as that seen in acidic with high conductivity in 1968 by Dall’Olio et al.6 However, the electrochemical characteristics of poly(pyrrole) have been aqueous solution obtained.By preparing poly(aniline) at a lower temperature, its electrochemical characteristics were more intensively studied since the preparation of a conducting film by electropolymerization.7 Generally, poly(pyrrole) was improved.The degradation was also restricted by employing a specific counter anion (dopant). The electrochemical charac- prepared by electropolymerization with a constant current or constant potential method, e.g.for the constant potential teristics and film properties of poly(aniline) depend on the counter anion, similarly to other conducting polymers. By method, pyrrole was electropolymerized by applying 0.8–1.3 V (vs. standard calomel electrode SCE) to obtain a film with employing camphorsulfonic acid, the post-cycling stability in acidic solution was improved.25 Extended color variation was reasonable mechanical stability.8 The film obtained exhibited a single reversible redox couple at around -0.2 V (vs.SCE) attempted by employing chemical substitution of the poly(aniline) backbone. Poly(N-naphthylaniline) has been reported to in its cyclic voltammogram. The film is brown–black in the oxidized state, and turns pale-yellow on going to cathodic show reversible and durable multicolor changes from pale yellow to red to purple to dark blue at 0.5–0.8 V vs.Ag/AgCl potential. Several kinds of pyrrole derivatives such as Nsubstituted pyrroles were polymerized and their electrochromic (Fig. 2).26 The possibility of multicolor electrochromism with colors of blue, green and red (BGR) was also observed by characteristics were studied.9 Further, several anions were investigated as dopants to extend the color variation.However, operating three poly(aniline derivatives) of poly(o-phenylenediamine) (pale yellow U<blue), poly(aniline) (pale yellow almost all of these films showed limited color changes from brown–black or –red in the oxidized state to pale yellow in U<green) and poly(metanilic acid) (pale yellow <red) independently (Fig. 3).27 The response time of the color change is the neutral state, despite the fact that the conductivity of poly(pyrrole derivatives) is sensitive to substituted groups and also an important factor when materials are applied to electrochromic devices. Very short response times were obtained dopants.10 For poly(thiophene derivatives), their electrochromic for poly(2-methoxy- or 2-ethoxy-aniline) films of about 50 nm thickness.The film was mentioned to have adequate contrast characteristics have been intensively studied since 1982–1983.11–13 Poly(thiophene) showed a change from blue and uniformity. This electrochromism required less than 2.5 ms switching time for about 90% conversion in relative trans- in the oxidized state to red in the neutral state.Despite this limited color variation and the lack of a colorless state, mission at 633 nm for both oxidizing and reducing potential changes.28 The film thickness may be one factor for this very poly(thiophene derivatives) are expected to be promising electrochromic conducting polymers. It has been reported that the fast response. Other conducting polymers have also been investigated as response time between each state is several tens of milliseconds and that color changes can be cycled more than 105 times novel electrochromic materials.Poly(isothianaphthene)29 and 498 J. Mater. Chem., 1998, 8(3), 497–506Table 1 Electrochromic properties of ruthenium complex conducting polymer36 oxidation state solution color 2 orange 1 purple 0 blue -1 green–blue -2 brown -3 rust -4 cherry red Electrochromic conducting polymers with electrochromic dopants Conducting polymers were commonly prepared via chemical Fig. 2 Absorption spectra of a poly(N-naphthylaniline) film dipped in 1 M LiClO4–CH3CN at various applied voltages: (a) blue (0.8 V), or electrochemical polymerization as described above. As- (b) purple (0.7 V), (c) red (0.6 V) and (d) yellow (0.0 V).(Reproduced prepared conducting polymers were obtained in the oxidized by permission from ref. 26.) (doped) state unless subjected to further treatment. Therefore, during polymerization, anions coexisting in the polymerization solution were incorporated into a conducting polymer matrix through electrostatic interaction. If one employed a functional anion showing electrochromic properties as the conducting polymer dopant, its electrochromism can be combined or mixed with that of the conducting polymer. This combination potentially produces the problem of interference of each material resulting in spectroelectrochemical behavior more complex than the simple superimposition of these constitutive building blocks.However, this strategy is useful and an easy method of extending color variation and realizing multicolor electrochromism in a device.Anionic phthalocyanine was successfully incorporated into a poly(pyrrole) matrix as dopant, and its electrochromic characteristics were studied.37 By scanning a potential ranging Fig. 3 Absorption spectra of (a) PA, (b) PMA and (c) PPD, films from -1.6 to 1.2 V vs. SCE, each electrochromism was found oxidized at 0.6 V (Ag/AgCl) in 0.2 M LiCl containing 0.01 M HCl.independently and four distinct colors could be obtained (Reproduced by permission from ref. 27.) (Fig. 4). The film was more stable as an electrode material than poly(pyrrole) alone, and had a higher degree of crystalpoly( N-methylisoindole)30 showed unique electrochromic linity. The same research group also examined ferrocene or character.Each monomer unit has an extra benzene ring with porphyrin incorporated poly(pyrrole) to obtain stable and the thiophene or pyrrole unit, and the presence of this extra durable multicolor electrochromism.38 benzene ring allows resonance structures with quinoid and benzenoid groups, which are expected to aVect its electronic properties considerably. Each polymer showed a color change from a colored state to near colorless on oxidation due to absorption in the near infrared but low absorption in the visible region in the doped polymer.This color change is the opposite to that for common p-type conducting polymers. A similar color change from highly opaque to colorless has been reported for poly(dithieno[3,4-b53¾,4¾-d]thiophene).31 This is expected to have application in optical shutter-like liquidcrystal devices.On the other hand, for multicolor purposes, a series of poly(diarylamines),32 a series of poly(pyrrolopyrroles), 33 metal-complexed poly(thiophenes),34 dinuclear CoII polymer complex with pyrrole substituents35 and a ruthenium complex conducting polymer36 have been reported.The ruthenium complex conducting polymer is of particular interest. The polymer underwent vivid color changes corresponding to seven stable oxidation states of the ruthenium complex (Table 1). The color observed in each oxidation state is distinct. This film also showed very rapid electrochromic response times for 90% conversion of several tenths of a millisecond. Further, the stability of the film was retained on repeatedly stepping the potential for a total of 5×105 cycles.Although the color change is due to the redox reaction of the ruthenium complex, incorporation of a metal complex into a Fig. 4 Cyclic voltammogram of a PP–TSCoPC film on an indium– conducting polymer moiety is one way to realize stable multi- tin-oxide electrode in 0.1 M TEAP–acetonitrile.(Reproduced by permission from ref. 37.) color electrochromism. J. Mater. Chem., 1998, 8(3), 497–506 499Other anionic metal complex was also examined as a dopant Prussian Blue is another electrochromic inorganic material. for cationic poly(pyrrole).39 Positively charged poly(pyrrole) It becomes blue in its oxidized state in contrast to WO3. incorporating quaternized pyridine moieties was prepared and Therefore, it has frequently been used as a counter electrode combined with an anionic iron phenanthroline complex.The for WO3.47 Electrochromic Prussian Blue has also been incorresulting composite showed electrochromism with color change porated into poly(aniline).48 Since this color characteristic is from bright red to transparent brown within 100 ms (DOD: similar to poly(aniline), enhanced electrochromism has been 10–90%).This color change was due to the redox reaction of obtained in comparison with poly(aniline) alone. In addition, the incorporated iron complex, with the cationic poly(pyrrole) stability over 4×104 cycles was observed in its composite films. moiety working as an electron wire. An interesting property The combination of a conducting polymer and another of this system is that other electro-optical functions can be material has many advantages, as described above.Since introduced by varying the metal complex. When an anionic conducting polymers alone cannot provide suYcient characterruthenium complex was employed as dopant, electrogenerated istics such as stability and mechanical strength at present, such chemiluminescence was observed in the resulting composite.a combination should help remove the basic problems seen Electrochromism is just like a game of ball (electron) between when an electrochromic cell is assembled. This methodology wall (electrode) and player (organic material). Undoubtedly, is simple because hundreds of electrochromic materials are fatigue (durability) becomes a problem in a player (organic potential candidates, but at the same time, mechanistic commaterial) sooner or later.Since metal complexes may be plexities may be a problem. Such combinations may yet realize classified as inorganic materials rather than organic, their outstanding systems. stability and durability imparts suYcient stability and durability on the organic material.The dopant mentioned above is purely anionic because it is Photopolymerization leading to conducting polymer used for p-type conducting polymers which have superior images stability compared with n-type. Examples of incorporating cationic dopants into p-type conducting polymers have been Electrochromism is one of the most successful ways to obtain also reported. Methylene Blue is a cationic molecule and is conducting polymer images on an electrode.However, on known to show a color change from colorless to blue on closer inspection of the process and mechanism of electrochrooxidation. Methylene Blue has been incorporated into mism, one notices that electron transfer between electrode and poly(pyrrole)40 or poly(aniline).41 The resulting film showed the material is required to form the image, and color change better electrochromism than the conducting polymer alone.In is diYcult without an electrode. If color change including particular, in the case of poly(aniline) doped with Methylene conducting polymer image formation could be realized at any Blue bound to anionic Nafion film, the film had improved place by external stimuli, this would open up a wide range of mechanical strength.Poly(aniline) not only worked as an applications for image recording systems, micro-patterning, electrochromic material, but also provided a conducting memory devices and microelectronics, as well as image network in a Nafion matrix. formation. It is certain that light plays an important role in the Electrochromic conducting polymers with inorganic microelectronics and computer industries from the viewpoint electrochromic materials of high speed information transmission and as an energy source.Therefore, light is a promising candidate as the external Inorganic materials have also been employed as electrochromic stimulus for forming conducting polymer images. On this basis, dopants or particles.It is well known that inorganic electroconducting polymer image formation by illumination has chromic materials have narrow color variation, but excellent received much interest and many studies have been reported. durability, reliability and stability. If a suitable combination One way to form conducting polymer images is photopolymer- between conducting polymer and inorganic electrochromic ization.As-prepared colored conducting polymers can be material is achieved, the resulting system would possess both obtained in an oxidized (doped) state unless subjected to characteristics. further treatment. Accordingly, illumination through a suitable When pyrrole was electropolymerized in suspended solution mask would produce a clear conducting polymer image on a containing WO3 particles, WO3 incorporated poly(pyrrole) substrate via photopolymerization.Such photopolymerizations was obtained.42 Since WO3 shows electrochromism (colorless leading to conducting polymer images are broadly divided to blue in the reduced state) the resulting film showed color into two categories according to their mechanism (Fig. 5); change from blue to pale yellow to black with appropriate (1) photopolymerization leading to conducting polymers with composition (Table 2).photocatalytic systems; (2) photoexcitation of the monomer WO3 incorporated poly(aniline) was also examined to obtain itself leading to conducting polymers. multicolor changes.43 By changing the potential from negative to positive, the color of the composite film changed from blue, which is due to WO3 in the reduced state, to pale yellow to green, due to the poly(aniline) in its oxidized state.The combination of conducting polymer and WO3 has been widely studied and their composite based electrochromic cells have been reported.44–46 Table 2 Dual electrochromisms which will appear under an ideal condition at polypyrrole films containing WO342 conductivity type insulating conducting composition reduced PPy reduced PPy oxidized PPy and HxWO3 and WO3 and WO3 color pale blue pale yellow black Fig. 5 Schematic representation of each photopolymerization leading (HxWO3) (reduced PPy) (oxidized PPy) conducting polymer image: (1) photocatalytic sustem and (2) monomer photoexcitation system. ÷ negative E0WO3 E0PPy positive � 500 J.Mater. Chem., 1998, 8(3), 497–506Photopolymerization leading to conducting polymers with photocatalytic systems The first attempt involved photoelectrochemical deposition of poly(pyrrole) on n-type GaAs49 or n-type Si wafers50 under an applied external bias to prevent photodegradati of the GaAs surface in its application in solar cells. The investigation of photopolymerization of conducting polymers for the purpose of image formation was begun about five years later.In the early stages, inorganic semiconductors such as n-type TiO251 deposit and n-type Si wafer52 were employed to induce a photoelectrochemical polymerization of pyrrole. In these eVorts, an external bias was applied to the semiconductor electrode to perform the polymerization. However, after pointing out some disadvantages of an electrode system, light-localized deposition of poly(thiophene) and poly(pyrrole) on n-type Si wafer was achieved.53 In this system, photogenerated holes in the n-type Si water could oxidize pyrrole or thiophene derivatives initiate polymerization.The remaining electron in n-type Si wafer could reduce Ag ions, which exist in the reaction solution, leads to Ag Fig. 6 (a) Wavelength dependence of photo-sensitized polymerization deposition at the back surface of the Si wafer. Contrast is not of pyrrole by Ru(bpy)32+ and (b) absorption spectrum of Ru(bpy)32+. suYcient due to the dark and shiny Si wafer surface, but a Relative yield of poly(pyrrole) was obtained from the absorbance good light-localized pattern has been reported.The particle change at 800 nm due to poly(pyrrole) formation. (Reproduced by size of TiO2 was also employed to initiate photopolymerization permission from ref. 55.) of pyrrole.54 Illumination at less than 300 nm in a solution containing Ag+, anionic porphyrin, pyrrole and TiO2 gave photopolymerization mechanism was assumed, and PPD poly(pyrrole) on particulate TiO2.This procedure is eVective worked as an initiator for the photopolymerization (Scheme 1). in modifying the TiO2 particle surface. The reduction in the When both ruthenium complex and methyl viologen were size of sensitizing materials may open up a new direction in incorporated in a Flemion matrix, polymerization eYciency molecular modification leading to molecular opto-electronics.was considerably enhanced. A clearly defined image with ca. In photosensitizing systems, including photopolymerization, 2 mm could be easily obtained by visible light illumination photoinduced electron transfer plays an important role. It is with a conventional xenon lamp through a suitable photomask certain that photoinduced electron transfer in inorganic semi- (Fig. 7).59 conductors proceeds with higher eYciency. However, the diminution of the size of features on inorganic materials is limited Photoexcitation of monomer itself leading to conducting by processing techniques and quantum eVects. Photoinduced polymer electron transfer has also been found in functional molecules. Another way to induce photopolymerization is by the direct Since these functional units are smaller than inorganic particles photoexcitation of monomer leading to conducting polymer and are easily modified by chemical substitution, a system without photosensitizer. The advantage of this method is to employing photoinduced electron transfer between functional simplify the process of conducting polymer image formation molecules would be useful for easier fabrication of molecular because photosensitizer as a system component to initiate functional systems.On this basis, photopolymerization leading the polymerization is not necessary. 1,2-Poly(azepine) was to conducting polymers with photoinduced electron transfer obtained by photolysis (photopolymerization) of phenyl azides. between functional molecules has been performed. This photopolymerization was reported to proceed via inter- Ruthenium and cobalt complexes were used as photosensimediates which in turn were formed from the excited singlet tizers and electron acceptors to achieve photopolymerization state of phenyl azide, and was performed in either inert gas or of pyrrole (Fig. 6).55 The conductivity of the photopolymerized solvent. The slightly oxidized polymer obtained was easily poly(pyrrole) was 3×10-4 S cm-1 which is lower than that of oxidized on exposure to I2 or AsF5, resulting in a color change the chemically polymerized one.However, when each complex and an increase in conductivity to 10-2 S cm-1. High-resowas incorporated into a Nafion matrix and was illuminated lution patterns with less than 10 mm width were obtained when through a mask, a pattern of poly(pyrrole) with 10 mm width the irradiation was performed through a photomask (Fig. 8).60 was successfully obtained. A copper complex56 was also used The well-known conducting polymer, poly(pyrrole), was also as photosensitizer to obtain photopolymerized poly(pyrrole). obtained by photopolymerization without photosensitizer. The A characteristic of the above mentioned molecular systems procedure is very simple, only involving the exposure of an is that visible light is employed for illumination. In contrast acidic aqueous solution of pyrrole to sunlight, but it has the to the above, UV (254 nm) light induced photopolymerization of pyrrole was carried out in halogenated solvent with pyrrole and ferrocene or iron–arene salts.57 Photogenerated halogen attacked the growing chain, resulting in the loss of pconjugation of the poly(pyrrole) obtained.Therefore, a lower conductivity (10-5 S cm-1) and lower electrochemical activity were obtained. Poly(aniline) was also photocatalytically obtained in the bilayer system composed of ruthenium complex-incorporated Nafion and viologen-pendant-poly(siloxane) in acidic aqueous solutions.58 It is interesting that photopolymerization does not occur in acidic solution containing only aniline.N-phenyl- NH NH2 NH n NH2 NH NH2 Ru(bpy)3 2+ or p-phenylenediamine (PPD, head-to-tail coupling dimer of Scheme 1 Schematic representation of the photopolymerization of aniline derivatives with ruthenium complex aniline) was necessary to induce photopolymerization. The J.Mater. Chem., 1998, 8(3), 497–506 501from DHITN to CCl4 or oxygen. This photopolymerization seems to include the contribution of photoinduced cationic polymerization. In the latter case, the photopolymerization was initiated by electron transfer from the photoexcited oligothiophene to the electron acceptor. The molecular mass of the polymer obtained was not so high (weight-averaged molecular mass ca. 400–700), but it was aVected by the concentration of the electron acceptor. Poly(aniline) was also obtained by irradiating an Au electrode with a Nd5YAG laser in a solution containing aniline under an applied external bias.63 Visible Ar laser irradiation also produced polymer on the electrode. A detailed polymerization mechanism is not known, but is of interest from the viewpoint of microprocessing.From the viewpoint of technological applications of the resulting conducting polymer, photodeposition of amorphous poly(diacetylene derivative) from monomer solution is interesting. 64 Despite the considerable volume of literature available on solid state polymerization of diacetylene derivatives, solution state photodeposition has never been reported, as far as we know.It was pointed out that the resulting films with thicknesses of 1 mm had optical qualities superior to that grown by standard techniques. The resulting films exhibit good third order non-linear susceptibilities. Photoinduced electrochromism leading to conducting polymer images Photoinduced electron transfer is an attractive reaction to convert photoenergy to electrical energy.In the preceding section, it was mentioned that photoinduced electron transfer plays an important role in oxidizing monomers leading to conducting polymers by photoexcited sensitizer. If such a reaction is also applied to oxidize or reduce the reduced or oxidized state of a conducting polymer respectively, combination of photoinduced electron transfer and electrochromism must enable us to induce electrochromism by photoillumination.Photoinduced electrochromism, called photoelectrochromism, has been investigated and reported. Photoelectrochromism (photooxidation) of poly(N-methyl- 1mm pyrrole) on an n-Si surface has been performed.65 The polymer Fig. 7 Micrographs of photopolymerized poly(aniline) pattern (top) was photoelectrochemically prepared on an Si surface under and photomask (bottom).an applied external bias. Even though the oxidation of the polymer was performed in this way with a light pulse (photoelectrochromism was not therefore performed only by light) the possibility of optical memory was demonstrated. In this work, the polymer was electrochemically or photoelectrochemically prepared on an Si surface.A simple casting method from polymer solution seems viable. Since this work, the photoredox reaction of conducting polymer with photosensitizer has developed in detail and scope. This research is broadly divided into three categories (two are shown in Fig. 9): i.e. (1) photoreduction systems: (2) photooxidation systems and (3) others. Photoreduction for conducting polymer image formation WO3 incorporated poly(aniline) was prepared as a photoelectrochromic film by electropolymerization of aniline in solution containing suspended WO3.By illuminating this film, photoexcited electrons generated in WO3 reduced the oxidized Fig. 8 SEM photograph of 1,2-poly(azepine) as grown pattern-wise state of poly(aniline) to induce the color change, photo- on fused silica with so-called contact illumination. The bars represent electrochromism.The reduced pattern could be oxidized 10 microns. The mask used exhibits equal lines and spacers. by electrochemical reaction via electrode, and the repetition (Reproduced by permission from ref. 60) of photoreduction and electrooxidation could be cycled (Fig. 10).66 TiO2 was also incorporated in poly(aniline) by the disadvantage in that it takes a long time to obtain the poly(pyrrole) pattern (100 nm after 15 h).same method as WO3 incorporation.67–69 It is of interest that the photoreduction of deprotonated poly(aniline) can be 1,3-Dihydroisothianaphthene (DHITN)61 or oligothiophene62 was photopolymerized in acetonitrile solution in the performed even in aqueous neutral solution.Generally, the deprotonated poly(aniline) (quinone diimine form) is electro- presence of CCl4 or p-dinitrobenzene as an electron acceptor. In the former case, photopolymerization was initiated by the inactive despite it being converted to a protonated form. In this system, since methanol added in the reaction solution intermediate which in turn was produced by electron transfer 502 J.Mater. Chem., 1998, 8(3), 497–506Fig. 9 Schematic representation of each photoredox reaction leading conducting polymer image (1) photoreduction system and (2) photooxidation system Fig. 12 Light image formation on poly(aniline)–TiO2 film immersed in 0.5 mol dm-3 phosphate buVer (pH 7) containing 20 wt% methanol by projecting the positive image on the poly(aniline)–TiO2 film with illumination by 500W xenon lamp for 1 min.(Reproduced by permission from ref. 68.) acidic aqueous solution. The reason spreading was not found for the TiO2 system in neutral aqueous solution is explained by the character of the TiO2 surface. Since TiO2 has many surface hydroxy groups which are involved in proton dissociation and association in aqueous solution, photogenerated protons are likely to be rapidly trapped by these surface hydroxy groups and give high acidity only in the illuminated area.This prevented the spreading of the image for the TiO2 system. Further, a quantum eYciency of ca. 10% was obtained Fig. 10 Change in absorption spectra of poly(aniline)–WO3 film for this light image formation determined at 355 nm under low caused by illumination with a 500W xenon lamp under open circuit illumination intensity. By employing a TiO2 system, an excel- (solid line) taken at 10 s intervals in PC+DME (151) containing 20v/o lent picture image was demonstrated (Fig. 12).68 methanol and 1 M LiClO4, and absorption spectrum obtained by succeeding polarizing at 0.7 V vs. SCE for 5 s (broken line). (Reproduced by permission from ref. 66.) Photooxidation for conducting polymer image formation The above mechanism is believed to be due to photoreduction works as a scavenger for holes photogenerated in TiO2, methanol was oxidized by UV irradiation to generate protons. of the oxidative state of poly(aniline) by photoexcited inorganic semiconductor particles. Therefore, the illuminated part was The attachment of these released protons to the deprotonated poly(aniline) occurs to induce the photoreduction even in bleached, and this provided a positive-type images.Image formation by utilizing photooxidation was also examined. The neutral aqueous solution (Fig. 11). This explanation was con- firmed by monitoring the change in mass with a quartz crystal photooxidation system provided a negative-type image because the illuminated part was colored; this might be applicable to microbalance. CdS was also incorporated in poly(aniline) to compare it with the TiO2 system.70 However, spreading of the photographic and optical disk systems. Photooxidation of poly(aniline) with photoinduced electron transfer between a photoreduced image was found for the CdS system.This is due to the escape of the protons released beyond the illumi- ruthenium complex and viologen was reported.71 Since photoinduced electron transfer between molecules was employed, it nated area by diVusion. Such a spreading was also found when the photoreduction in the TiO2 system was carried out in would be useful for easier fabrication of molecular functional NH NH N N NH NH + + A– A– NH NH NH NH + + A– A– NH NH NH NH e– h+ hn CH3OH HCHO + 2H+ H+ +A– e– –A– A– = anion IP IQ PIQ Fig. 11 Schematic illustration of the photoreduction of the deprotonated form in methanol solution (ref. 69) J. Mater. Chem., 1998, 8(3), 497–506 503systems. Ruthenium complex incorporated Nafion was coated on the poly(aniline) deposited on an ITO electrode. This electrode in acidic aqueous solution containing methyl viologen was illuminated and a colored image of poly(aniline) was obtained. In this system, poly(aniline) was oxidized by the oxidation state of the ruthenium complex generated via photoinduced electron transfer between the photoexcited ruthenium complex to methyl viologen.This photooxidation of poly(aniline) was also performed in an all solid state cell by employing a polymer electrolyte composed of Flemion.Its application to optical memory and compact disk systems was indicated.72 Instead of using casting methods to prepare the poly(aniline) film, a negative poly(aniline) image was also formed via photopolymerization in a Nafion film electrostatically incorporating a ruthenium complex and methyl viologen.This image could be erased by immersing the film in a hydrazine solution, and the erased image was reproduced by illumination. Repetition Fig. 14 Time course of absorbance changes at 800 nm induced by between photooxidation and reduction with hydrazine was photooxidation and photoreduction for the single layer film in 1.0 M therefore obtained as shown in Fig. 13.59 HCl solution with 10 vol% MeOH.Further, the study to induce not only photooxidation of poly(aniline) but also its reduction photocatalytically in a film of poly(3-hexylthiophene) containing one of these chemisystem was examined. One easy method is the combination of cals, a colored and insoluble conducting polymer pattern was a ruthenium complex incorporated into a Nafion with TiO2 obtained.74,75 Since the solubility of the pattern decreased system.73 The repetition between photooxidative image forma- upon illumination because the dopant worked as crosslinker tion and photoreductive image erasing was successful for films between polymer chains (Fig. 15), only the illuminated image composed of photopolymerized poly(aniline) and Flemion was obtained on the substrate by washing the illuminated film containing a ruthenium complex and TiO2 as shown in Fig. 14. with appropriate solvent. This is expected to be an alternative Photooxidation and photoreduction were induced by visible and unique technique of lithography. illumination and UV illumination, respectively. Therefore, this It was also reported that the repetition of photochemical poly(aniline) system enabled photolytically tunable conduc- doping and electroreduction could be obtained when poly(3- tivity and electrochromism leading to image formation. This hexylthiophene)–gel electrolyte (containing photoinduced system showed good memory eVects, an advantage character- dopant) bilayer was sandwiched between indium–tin-oxide istic of electrochromism.The work must be regarded as (ITO) electrodes.The electrolyte layer contained low molecular preliminary, however. There are some problems such as slow mass solvent to facilitate mass transport, but this could indicate response and durability to overcome. Further analysis must a possibility for a solid state rewritable optical recording be carried out to fabricate more attractive systems. system. This photochemical doping can be called a kind of negative type photoresist.Similar work was reported for Photochemical doping, photocrosslinking, photodecomposition poly(3-hexylthiophene) with no other dopant.76 A crosslinked and others insoluble polymer pattern was obtained by illumination at 422 nm. The mechanism of the crosslinking reaction was Photochemical doping, photocrosslinking and photodecomexplained by the photooxidation of the alkyl side chain leading position were employed to form conducting polymer images.to crosslinking or Diels–Alder addition of photoexcited singlet An advantage of these systems is easier fabrication of solid state of oxygen to the main chain thienyl residue. The resulting state systems, but tunability and repetition are problems. For polymer pattern could be oxidized by nitrosonium tetra- photochemical doping systems, tetraphenyliodonium tetra- fluoroborate to give a bulk conductivity of ca. 6 Scm-1. fluoroborate and diphenyliodonium hexafluoroarsenate were cis-1,4-Poly(butadiene) was studied to prepare fluorescent used as photoinduced dopants because these are known to conducting patterns with good stability.77 The conductivity of release BF4- and AsF6- upon illumination.By illuminating a trans-1,4-poly(butadiene) could be increased by eight orders of magnitude upon conjugation and self-doping induced by I2 at room temperature, whereas cis-1,4-poly(butadiene) could not be made conducting by reaction with I2 under the same conditions. Since this photoisomerization from cis to trans can Fig. 15 Residual content of poly(3-hexylthiophene) film remaining Fig. 13 Time course of absorbance changes at 860 nm induced by after light irradiation and then washing with chloroform relative to the original. (Reproduced by permission from ref. 75.) photooxidation (—) and chemical reduction (,) for the single layer film 504 J. Mater. Chem., 1998, 8(3), 497–506be achieved by UV irradiation, I2 treatment of the irradiated some tunable functionality such as tunable conductivity unlike conventional metal images.film produces a conducting polymer image. The image shows strong fluorescence emission and has remained stable for 10 Light is an attractive tool to write conducting polymer images on an appropriate canvas. The image can be prepared months in air. Poly(aniline) also has good stability in air and could be a space-selectively, meaning that one can attach functions depending on the conducting polymer for any geometry in any promising candidate for an image formation material.Poly(aniline) containing triphenylsulfonium hexafluoronate place. Such a technique may explore new technologies beyond the mere application to photolithography, photographic sys- (onium salt) was studied as a negative type photoresist.78 Since the onium salt dissociated into the acid and protons by tems, optical recording systems or optical memory devices. A promising example is the use of photodeposited images irradiation with UV light or electron beam, only the irradiated part of poly(aniline) was oxidized via doping of the photo- composed of amorphous poly(diacetylene derivatives)64 which exhibit good third order non linear susceptibilities.It may be generated acid and proton. The resulting oxidized state of poly(aniline) became insoluble in N-methylpyrrolidin-2-one. useful to fabricate photofunctional elements on a microchip for advanced technology. The other example is easier fabri- By employing an electron beam, conducting polymer lines with a high resolution of 0.25 mm were patterned in thick films.cation of photoenergy conversion systems. The conducting polymer photochemically prepared on a photosensitizer works Since electron beams can be focused as small as 10 nm, this system has considerable interest as an alternative method to as a donor for photosensitizers. Therefore, eYcient photoenergy conversion systems composed of conducting polymers as the current lithographic processes.As another unique negative-type photoresist, polymeric films donor–photosensitizer–acceptor can be simultaneously fabricated during photopolymerization, because the conducting containing poly(bipyridine) complexes of ruthenium have been studied.79 Two sequential photochemical ligand substitutions polymer would be prepared so that the photoenergy for photopolymerization is converted most eYciently in the system.occurred in a stepwise manner in the film upon illumination to vary the absorption spectrum and redox potential of the Such a system can also be fabricated on inorganic semiconductors instead of photosensitizing molecules. This indi- film in three stages.Line patterns composed of diVerent stage components with image resolution below 10 mm could be cates a possibility to modify the surface of semiconductor chips with photofunctional conducting polymers. Further, since obtained. Interestingly the potential control of the underlying ITO electrode allows for selective oxidation of one or both many photosensitizers act as phosphors, microelectroluminescent devices may also be prepared in a chip if the system stage components and spatially controlled color change in image films.The polymer shows redox conducting behavior, configuration, relative energy level of each material, electrode contact and other problems are solved. which is strictly speaking diVerent from the above mentioned conducting polymers.However, this gives an interesting direc- Reviewing the character of conducting polymer produced by photochemical reaction from the viewpoint of these appli- tion for the preparation of active images with multi-photo and electric functions. cations, photochemical reactions mentioned in this paper seem to produce suYcient quantities of conducting polymer on a It is well known that positive type (photodecomposition type) photoresists also play an important role in the field of substrate although there is a problem with response time.Thus prepared conducting polymers show similar physicochemical photolithography. Conducting polymer image formation with positive type photoresists was also examined. Copolymers of characteristics such as conductivity and electrochemistry to that obtained by other methods.thiophene and thiophene derivatives and other poly(thiophene) derivatives were illuminated with a high pressure mercury For fabricating a device with such a conducting polymer, connecting an electrode to the photochemically prepared con- lamp to shorten the p-conjugation system.80 This could be confirmed by absorption spectra. The former copolymer in ducting polymer pattern may be a problem on a sub-micron or nanometre scale.However, STM and AFM techniques would particular showed no redox activity after illumination. Poly(3-alkylthiophene) containing FeCl4- was also prepared be useful in achieving this. Further, if complete optical control were to become feasible in a system, it is not necessary to as a positive type photoresist.81 Upon illumination at 366 nm, a drastic change in the conductivity from 6 to 10-5 S cm-1 consider electrodes.Some of the resulting images are fairly robust for device fabrication. For example, the pattern composed was found. This conductivity change was explained by the direct photodegradation of FeCl4- as a dopant or photoreduc- of I2 treated trans-1,4-poly(butadiene) remains stable for about 10 months in air,77 and poly(aniline) also has good stability in tion of FeCl4- via photoinduced electron transfer from a bipolaronic residue.The above two systems enable the forma- air. However, it is necessary to analyze stability (chemical stability, physical aging and so on) under more severe conditions tion of conducting polymer images by distinguishing the illuminated less-conductive pattern from the unilluminated in order to employ materials in practical applications. It is also certain that we have to overcome many barriers conductive pattern. Irrespective of whether they are of positive or negative type, the conductive image can be fabricated more related to the total system to realize these exciting technologies and their applications.Detailed analysis is needed to clarify easily in the above mentioned systems than in commercial photoresists possibly because these enable a decrease in the relative molecular arrangements, each mechanism concerning photoexcitation at a photosensitizer, photoinduced electron number of processing steps to be made. Further improvement on response and resolution will, however, be required to give transfer reaction in the system, the following polymerization or redox reaction and the total electron transfer throughout the application of these systems a firmer foundation. the system.Extensive research on conducting polymers concerning Future application design, preparation method, conduction mechanism and application has been carried out world-wide.Further, photoinduced This article has given an overview of conducting polymer image formation with electrochemical reaction or photo- electron transfer systems have also been studied extensively. Both of these have independently gathered scientific and induced electron transfer reactions. The former is called electrochromism, and its application is directed mainly toward technological interests, and their combination is of great interest. Simple ways to combine these two independent fields large-scale technologies such as smart windows and solar energy related technology.The latter seems to be advantageous are mixing, layering and chemical modification. Since each method has both benefits and disadvantages, each does not for smaller scale technologies. The characteristics of images composed of conducting polymers depend on the features of provide enough results to satisfy the many demands on the system.However, the combination of these two independent the conducting polymer employed. Such materials would give J. Mater. Chem., 1998, 8(3), 497–506 50539 T. Iyoda, M. Aiba, T. Saiki, K. Honda and T. Shimizu, J.Chem. fields is believed to have the potential to generate entirely Soc., Faraday T rans., 1991, 87, 1765. unexpected attractive technologies in the near future. We hope 40 T. Amemiya, K. Itoh and A. Fujishima, Ber. Bunsenges. Phys. this article will be of some help in awakening interest for Chem., 1989, 92, 682. further investigations and applications of conducting polymer 41 S.Kuwabata, K. Mitsui and H. Yoneyama, J. Electroanal. Chem., image formation with photoinduced electron transfer. 1990, 281, 97. 42 H. Yoneyama and Y. Shoji, J. Electrochem. Soc., 1990, 137, 3826. 43 H. Yoneyama, S. Hirao and S. Kuwabata, J. Electrochem. Soc., 1992, 139, 3141. 44 A. M. Rocco, M.-A. DePao, A. Zanelli and M. Mastragostine, References Electrochim. Acta, 1996, 41, 2805. 1 H. Shirakawa, E. J. Louis, A. G. MacDiarmid, C. K. Chiang and 45 T. Ohtsuka, T. Wakabayashi and H. Einaga, Synth. Met., 1996, A. J. Heeger, J. Chem. Soc., Chem. Commun., 1977, 578. 79, 235. 2 Handbook of Conducting Polymers, ed. T. J. Skotheim, Marcel 46 M. Sima and T. Visan, Romanian J. Phys., 1995, 40, 1073. Dekker, New York, 1986. 47 N. Kobayashi, R. Hirohashi, H. Ohno and E. Tsuchida, Solid State 3 (a) A.O. Patil, A. J. Heeger and W. Wudl, Chem. Rev., 1988, 88, Ionics, 1990, 40/41, 491 and references cited therein. 183; (b) P. M. S. Monk, R. J. Mortimer and D. R. Rosseinsky, in 48 N. Leventis and Y. C. Chung, J. Electrochem. Soc., 1990, 137, 3321. Electrochromism: Fundamentals and Applications, VCH,Weinheim, 49 R. Noufi, D. Tench and L. F. Warren, J. Electrochem.Soc., 1980, 1995, p, 9; (c) R. J. Mortimer, Chem. Soc. Rev., 1997, 26, 147. 127, 2310. 4 Abstracts, in International Conference on Electroluminescence of 50 T. Skotheim, I. Lundstro�m and J. Prejza, J. Electrochem. Soc., 1981, 128, 1625. Molecular Materials and Related Phenomena held in Fukuoka, 51 M. Okano, K. Itoh, A. Fujishima and K. Honda, Chem. L ett., Japan, 1997. 1986, 469. 5 A. F. Diaz, J. I. Castillo, J. A. Logan and W. Y. Lee, J. Electroanal. 52 H. Yoneyama and M. Kitayama, Chem. L ett., 1986, 657. Chem., 1981, 129, 115. 53 H. Yoneyama, K. Kawai and S. Kuwabata, J. Electrochem. Soc., 6 A. Dall’Olio, Y. Dascola, V. Varacco and V. Brocchi, C. R. Acad. 1988, 135, 1699. Sci. Ser. C, 1968, 267, 433. 54 T. Shimizu, T. Iyoda and K. Honda, Pure Appl.Chem., 1988, 7 A. F. Diaz and K. K. Kanazawa, J. Chem. Soc., Chem. Commun., 60, 1025. 1979, 854. 55 H. Segawa, T. Shimizu and K. Honda, J. Chem. Soc., Chem. 8 A. F. Diaz and B. Hall, IBM J. Res. Develop., 1983, 27, 342. Commun., 1989, 132. 9 R. Bjorklund, S. Andersson, S. Allenmark and I. Lundstro�m, Mol. 56 J.-M. Kern and J.-P. Sauvage, J. Chem. Soc., Chem. Commun., Cryst. L iq.Cryst., 1985, 121, 263. 1989, 654. 10 M. Satoh, K. Kaneto and K. Yoshino, Synth. Met., 1986, 14, 289. 57 J. F. Rabek, J. Lucki, M. Zuber, B. J. Qu and W. F. Shi, 11 G. Tourillon and F. Garnier, J. Electroanal. Chem., 1982, 135, 173. J.Macromol. Sci., Pure Appl. Chem., 1992, A29, 297. 12 K. Kaneto, K. Yoshino and Y. Inuishi, Jpn. J. Appl. Phys., 1982, 58 K. Teshima, K. Yamada, N.Kobayashi and R. Hirohashi, Chem. 21, L567. Commun., 1996, 829. 13 F. Garnier, G. Tourillon, M. Gazard and J. C. Dubois, 59 K. Teshima, S. Uemura, N. Kobayashi and R. Hirohashi, in J. Electroanal. Chem., 1983, 148, 299. contribution. Note; Flemion is perfluorosulfonic acid polymer 14 K. Kaneto, H. Agawa and K. Yoshino, J. Appl. Phys., 1987, 61, 1197. (ref. 72) with a similar structure to Nafion and is obtainable from 15 S.D. D. V. Rughooputh, A. J. Heeger and F. Wudl, Synth. Met., Asahi Glass Co. Ltd. 1987, 21, 41. 60 E. W. Meijer, S. Nijhuis and F. C. B. M. Vroonhoven, J. Am. Chem. 16 Y. Watanabe, J. Chem. Soc., Chem. Commun., 1989, 123. Soc., 1988, 110, 7209. 17 I. Imae, K. Nawa, Y. Ohsedo, N. Noma and Y. Shirota, 61 M. Kitano, T. Iyoda and T. Shimizu, Polym.J., 1995, 27, 875. Macromolecules, 1997, 30, 380. 62 M. Fujitsuka, T. Sato, H. Segawa and T. Shimizu, Chem. L ett., 18 D. M. Mohilner, R. N. Adams and W. J. Argersinger, J. Am. Chem. 1995, 99. Soc., 1962, 84, 3612. 63 N. Teramae, T. Uchida and T. Ishioka, Kagakukogyo, 1994, 45, 19 T. Kobayashi, H. Yoneyama and H. Tamura, J. Electroanal. Chem., 952. 1984, 161, 419. 64 M. S.Paley, D. O. Frazier, H. Abdeldeyem, S. Armstrong and 20 T. Kobayashi, H. Yoneyama and H. Tamura, J. Electroanal. Chem., S. P. McManus, J. Am. Chem. Soc., 1995, 117, 4775. 1984, 177, 273. 65 O. Ingana�s and I. Lundstro�m, J. Electrochem. Soc., 1984, 131, 1129. 21 N. Kobayashi, K. Yamada and R. Hirohashi, Chem. L ett., 1990, 66 H. Yoneyama, S. Hirao and S. Kuwabata, J. Electrochem. Soc., 1983. 1992, 139, 3141. 22 M. C. Miras, C. Barberro, R. Kotz and O. Haars, J. Electrochem. 67 S. Kuwabata, N. Takahashi, S. Hirao and H. Yoneyama, Chem. Soc., 1991, 138, 335. Mater., 1993, 5, 437. 23 N. Kobayashi, K. Yamada and R. Hirohashi, Electrochim. Acta, 68 H. Yoneyama, N. Takahashi and S. Kuwabata, J. Chem. Soc., 1992, 37, 2101. Chem. Commun, 1992, 716. 24 K. Teshima, K. Yamada, N. Kobayashi and R. Hirohashi, 69 S. Kuwabata, A. Kishimoto and H. Yoneyama, J. Electroanal. J. Electroanal. Chem., 1997, 426, 97 and references cited therein. Chem., 1994, 377, 261. 25 M. C. Bernard, A. H. Le GoV, V. T. Bich and W. Zeng, Synth.Met., 70 H. Yoneyama, M. Tokuda and S. Kuwabata, Electrochim. Acta, 1996, 81, 215. 1994, 39, 1315. 26 J. G. Guay and L. H. Dao, Polym. Commun., 1989, 30, 149. 71 K. Yamada, M. Syakuto, K. Teshima, N. Kobayashi and 27 A. Kitani, J. Yano and K. Sasaki, J. Electroanal. Chem., 1986, R. Hirohashi, J. Soc. Photogr. Sci. T echnol. Jpn., 1994, 57, 445. 209, 227. 72 N. Kobayashi, T. Yano, K. Teshima and R. Hirohashi, Electrochim. 28 P. J. S. Foot and R. Simons. D, Appl. Phys., 1989, 22, 1598. Acta, in the press. 29 H. Yashima, M. Kobayashi, K.-B. Lee, A. J. Heeger and F. Wudl, 73 N. Kobayashi, Y. J. Kim, K. Teshima and R. Hirohashi, J. Electrochem. Soc., 1987, 134, 46. unpublished work. 30 N. M. Hanly, D. Bloor, A. P. Monkman, R. Bonnett and 74 K. Yoshino, M. Ozaki and R. Sugimoto, Jpn. J. Appl. Phys., 1985, J. M. Ribo, Synth.Met., 1993, 60, 195. 24, L373. 31 A. Bolognesi, M. Catellani, S. Destri, R. Zamboni and C. Taliani, 75 K. Yoshino, H. Takahashi and R. Sugimoto, Jpn. J. Appl. Phys., J. Chem. Soc., Chem. Commun., 1988, 246. 1991, 30, L657. 32 J. Guay, R. Payston and L. H. Dao, Macromolecules, 1990, 23, 76 M. S. A. Abdou, Z. W. Xie, A. M. Leung and S. Holdcroft, Synth. 3598 and references cited therein. Met., 1992, 52, 159. 33 H. Hashimoto, N. Oyama, T. Ohsaka, S. Tanaka and T. Miyashi, 77 L. Dai, H. J. Griesser, X. Hong, A. W. H. Mau, T. H. Spurling, J. Electrochem. Soc., 1991, 138, 2003 and references cited therein. Y. Yang and J. W. White, Macromolecules, 1996, 29, 282. 34 J. L. Reddinger and J. R. Reynolds, Macromolecules, 1997, 30, 673. 78 M. Angelpoulos, J. M. Shaw and K.-L. Lee, Polym. Eng. Sci., 1992, 35 X. Ren, S. K. Mandal and P. G. Pickup, J. Electroanal. Chem., 32, 1535. 1995, 389, 115. 79 R. M. Leasure, W. Ou, J. A. Moss, R. W. Linton and T. J. Meyer, 36 C. M. Elliott, S. J. Schmittle, J. G. Redepenning and E. M. Balk, Chem. Mater., 1996, 8, 264. 80 M. Sandberg, S. Tanaka and K. Kaeriyama, Synth. Met., 1993, J.Macromol. Sci.-Chem., 1988, A25, 1215. 60, 171. 37 M. Velazquez, T. A. Skotheim and C. A. Linkous, Synth. Met., 81 M. S. A. Abdou and S. Holdcroft, Chem. Mater., 1994, 6, 962. 1986, 15, 219. 38 M. Velazquez, T. A. Skotheim and C. A. Linkous, Polym. Prepr., 1984, 25, 258. Paper 7/06386E; Received 1st September, 1997 506 J. Mater. Chem., 1998, 8(3), 497–506

ISSN:0959-9428    

DOI:10.1039/a706386e

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Communication Room temperature synthesis of hybrid organic–inorganic nanocomposites containing Eu2+ E. Cordoncillo,a B. Viana,b P. Escribanoa and C. Sanchez*b aDepartamento de Quimica Inorganica y Organica, Universitat Jaume I, Castellon, Spain bL aboratoire de Chimie de la Matie`re Condense`e UMR CNRS 7574, Universite� Pierre etMarie Curie, 4 place Jussieu, 75252, France the SiMH group as an in situ reducing agent which allows the formation of metal–silica nanocomposites.15 For the first time, the room temperature synthesis of new Eu2+ In the present work, the in situ formation of hydrogen doped hybrid materials together with their absorption and provided by the cleavage of the SiMH bonds is used to emission properties are reported.These hybrids containing generate, during the first step of hydrolysis and condensation Eu2+ and a small amount of Eu3+ are obtained through the reactions, europium in the divalent state. This communication hydrolysis and condensation of diethoxymethylsilane, addresses the room temperature synthesis of two new Eu2+ methyltriethoxysilane and zirconium tetrapropoxide hybrid materials. Absorption and emission properties of these precursors in the presence of europium trichloride.These Eu2+ Eu2+ doped hybrid matrices, which can be processed as bulks doped materials exhibit a strong host dependent Eu2+ or coatings, are also presented. luminescence, the intensity of which does not change upon air The first matrix (system A) is obtained through the storage, confirming that Eu2+ ions are eYciently trapped hydrolysis–condensation of diethoxymethylsilane inside these hybrid matrices.[HSi(CH3)(OCH2CH3)2, MDES] and methyltriethoxysilane [(CH3)Si(OCH2CH3)3, MTEOS] precursors in the presence of a europium(III) salt and zirconium tetrapropoxide (ZrP). MDES and MTEOS were first hydrolyzed with neutral water in ethanol. The molar ratio MDES5MTEOS5H2O5ethanol was 1515252.An ethanolic solution of ZrP and europium The Eu2+ ion is particularly unique because its broad band trichloride in the presence of acetylacetone was then added to luminescence 4f65d1A4f7 is strongly host dependent with the siloxane based sol. The final molar Zr5Eu5Si ratio was emission wavelengths extending from the UV to the red range 150.0159.The ZrP precursor plays a double role: it is used as of the electromagnetic spectrum.1 Therefore, the luminescent a Lewis acid catalyst to assist the cleavage of the SiMH properties of Eu2+ doped solids have been intensively studied bonds,21 and it can help to host Eu ions which usually demand during the past three decades. Theses studies have led to use high coordination numbers.of these compounds as phosphors, notably blue-emitting The second matrix (system B) was synthesized as follows. Eu2+5BaMgAl10O17 in lamp and plasma display panels and europium(III ) chloride was first cohydrolyzed with MDES UV-emitting Eu2+5SrB4O7 for medical applications and skin and triethoxysilane [HSi(OCH2CH3)3, TREOS]. The molar tanning. Crystalline or glassy Eu2+ doped materials are usually ratio MDES5TREOS5Eu5H2O5ethanol was 1.450.650.025 processed at relatively high temperatures.1–4 Moreover, the 251.This siloxane based sol containing europium was then synthesis and stabilization of europium in the divalent state added dropwise to a solution of ZrP in propanol. The Zr5Si under mild synthetic conditions are not easy tasks. To the best molar ratio was 2510.of our knowledge no work has been reported on the synthesis For both A and B, the mixing between the zirconium and optical properties of room temperature processed Eu2+ propoxide solution and the siloxane sol leads to the evolution doped matrices. of hydrogen gas which was used as a reducing agent to decrease The mild conditions provided by the sol–gel process allow the valence of europium cations from III to II.The resulting the synthesis of new hybrid organic–inorganic materials.5–10 clear sols (A and B) were magnetically stirred for 30 min in an The formation of the hybrid macromolecular network involves argon atmosphere. In order to obtain transparent monolithic hydrolysis and condensation of organically functionalized gels and coatings a few microns thick, an appropriate amount metal alkoxide precursors such as alkoxysilanes RxSi(OR¾)4-x of the sol (A or B) was poured into a plastic cuvette or (R=methyl, H or any organic function).The nature of the R deposited onto previously cleaned glass sheets. group allows the tailoring of materials which exhibit a wide The IR spectra of the xerogels A and B show the presence range of optical, mechanical, electrochemical and catalytic of a weak broad band located at 3500 cm-1 (nOMH) which properties.5–14 corresponds to some residual hydroxyl groups, and absorption Organic hydrosilanes HSi(OR)3 are precursors which are bands located at about 2200 cm-1 assigned to stretching particularly versatile for the synthesis of new hybrid mate- vibrations of residual SiMH bonds.Several bands are also rials.15–18 It has been recently demonstrated that hybrid observed at 1145–1025 cm-1, due to SiMOMSi bond matrices synthesized from HSi(OEt)3 and HSi(CH3)(OEt)2 vibrations and indicate that the siloxane network is partly precursors are high performance host matrices for spirooxazine homocondensed. Moreover, the presence of an IR band located photochromic dyes.19 These dyes embedded within such at 960 cm-1 assigned to SiMOMZr linkages22 shows that the matrices exhibit photochromic kinetics much faster than those siloxane species are linked to zirconium oxopolymers.This reported for spirooxazine in any other solid matrix.19,20 kind of hybrid network made from hydrolysis and conden- On the other hand, dehydrocondensation of organic hydro- sation of alkoxysilane precursors and zirconium alkoxides can silanes with silanols is one of the common methods for the be described as nanocomposites built from siloxane polymers synthesis of the siloxane linkage.21 This reaction occurs with cross-linked by zirconium oxo species.1,23 A more complete the evolution of hydrogen gas.In this sense, alkoxide precursors characterization of these new hybrid matrices by MAS NMR and SAXS is under investigation. containing SiMH groups demonstrate the possibility of using J.Mater. Chem., 1998, 8(3), 507–509 507The absorption spectra have been recorded at room temperature for sols and xerogels (A and B) on a Cary 5 (Varian) spectrophotometer using as reference undoped sols and xerogels.For a given system (A or B) the spectra are similar in the sol and in the xerogel states. Absorption spectra are constituted of a broad absorption band in the UV range (200–400 nm) attributed to the 4f75d0A4f65d1 (Eu2+) transition. As an example Fig. 1 shows the Eu2+ absorption spectra of coatings processed from hybrid systems A and B. In samples A, small peaks attributed to Eu3+ at 536 nm, 465.9 nm and 394.3 nm are also present in the visible range corresponding to transitions from the fundamental 7F0 level to the excited 5D0,1,2,3 and 5L6 levels.Strong diVerences are observed between the two systems: absorption maxima are located at 365 nm (ca. 27400 cm-1) and 310 nm (ca. 32250 cm-1) respectively for A and B. The areas of the absorption curves are two order of magnitude larger for divalent than for trivalent europium as expected according to the nature of the electronic transition.Fig. 3 Emission spectrum of europium doped hybrid xerogel of system The emission spectra of the two systems under excitation at B (lexc=355 nm) 355 nm show a broad emission corresponding to the intercon- figurational 4f65d1A4f75d0 transition centered at 460 nm (ca. 21800 cm-1) and 430 nm (ca. 23250 cm-1) for the A and B to about 9000 cm-1 for the hybrid films obtained from system B. The shift between the absorption and emission energies of systems respectively and an intraconfigurational 4f–4f Eu3+ emission at longer wavelengths. Several bands are obtained Eu2+ located in an oxygen ligand field has been assigned to a combination of crystal field and nephelauxetic eVects.1 corresponding to the 5D0A0,1,2,3 transitions.The more intense is the allowed dipolar electric 5D0A7F2 transition at Structures containing oxygen atoms in higher coordination number environments (highly coordinated by metal atoms) 610 nm (see Fig. 2). A Stokes shift value of the Eu2+ luminescence around 5600 cm-1 is obtained for the hybrid coatings produce Eu2+ emissions at longer wavelengths, and distortion of the oxygen polyhedra from ideal coordination geometry processed from system A while the Stokes shift value increases results in a larger Stokes shift.Large Stokes shifts have also been associated with a more asymmetric dopant geometry.2,3 The diVerences observed in the optical responses between A and B are probably related to diVerences in the processing of the hybrid materials which lead to a modification of its coordination sphere.For system A, europium is first reacted with zirconium propoxide and then cocondensed with siloxane precursors. As a consequence, in system A europium should be eYciently sequestered in the zirconium–oxo domains. Its coordination polyhedra should be mainly made of EuMOMZr bonds and therefore should be quite homogeneous. For system B, europium is first reacted with the siloxane precursors and then with zirconium propoxide.Thus the sequestering of europium by zirconium is probably less eYcient. The second neighbors of the europium ion should be, not only zirconium atoms, but also silicon atoms (EuMOMZr and EuMOMSi). The Eu2+ coordination sphere in system B should be consequently more distorted.Moreover, because EuMOMSi Fig. 1 Absorption spectra of europium doped hybrid xerogels: systems bonds are made with m2-oxo bridges while EuMOMZr bonds A and B are built with m3-oxo and m4-oxo bridges, for system B the europium coordination polyhedron should contain less oxygen atoms in higher coordination number environments. As a consequence, for system B, the absorption and emission bands are blue shifted and moreover this system also exhibits a larger Stokes shift.Lifetime measurements have been performed at diVerent wavelengths in A and B xerogels. At 610 nm, the 5D0 (Eu3+) emitting level presents an exponential decay profile with a lifetime value around 670 ms while the Eu2+ lifetime measured at the maximum of the broad emission band is estimated to be about 0.5 ms.With our intensified optical multichannel analyzer detector, it is also possible to observe the 5D1 emission around 550 nm. This emitting level presents a short lifetime value of 2 ms as this lifetime is shortened by the non-radiative 5D1A5D0 multiphonon relaxation mechanism. First measurements indicate a larger Eu2+5Eu3+ concentration ratio in xerogel B than in system A.A qualitative approach based on the measured absorption and emission coeYcients of the divalent and trivalent europium species indicate that the Fig. 2 Emission spectrum of europium doped hybrid xerogel of system A (lexc=355 nm) Eu2+5Eu3+ ratios are about 151 and 551 respectively for 508 J. Mater. Chem., 1998, 8(3), 507–50911 Sol–Gel Optics, Processing and Applications, ed.L. C. Klein, samples A and B.24 The higher Eu2+ content of sample B is Kluwer Academic Publishers, Boston, 1993. probably related to the more eYcient reducing medium pro- 12 B. Dunn and J. I. Zink, J.Mater. Chem., 1991, 1, 903. vided by the initial mixture of the europium trichloride with 13 D. Levy and D.Avnir, J. Phys. Chem., 1988, 92, 734. the MDES and TREOS silane precursors. Moreover, the 14 Sol–Gel Optics I, ed. J. D. Mackenzie and D. R. Ulrich, Proc. SPIE, intensity of the Eu2+ luminescence did not change when the 1990, 1328; Sol–Gel Optics II, ed. J. D. Mackenzie, Proc. SPIE, 1992, 1758; Sol–Gel Optics III, ed. J. D. Mackenzie, Proc. SPIE, xerogels (A and B systems) were stored in air for several 1994, 2288; Sol–Gel Optics IV, ed.J. D. Mackenzie, Proc. months, showing that Eu2+ ions are eYciently trapped inside SPIE, 1997, 3136. the hybrid matrix. 15 Advanced Materials and Processes by Sol–Gel T echniques, Proc. 2nd Eur. Conf. on Sol–Gel T echnol., Colmar, 1992, ed. R. Campostrini and S. Dire�, North-Holland, Amsterdam, 1993. E. C. thanks Conselleria de Cultura, Educacio� i Cie`ncia de la 16 G. D.Soraru, D. D’andrea, R. Campostrini and F. Babonneau, Generaritat Valenciana for financial support. The authors J.Mater. Chem., 1995, 5, 1374. would like to acknowledge P. Aschehoug for his help on the 17 M. Pauthe, J. Phalippou, R. Corriu, D. Leclerq and A. Vioux, J. Non Cryst. Solids, 1989, 113, 21. emission measurements. 18 M.Pauthe, J. Phalippou, R. Corriu, D. Leclerq and A. Vioux, J. Non-Cryst. Solids, 1991, 175, 187. 19 B. Schaudel, C. Guermeur, C. Sanchez, K. Keitaro and J. Delaire, References J.Mater. Chem., 1997, 7, 61. 20 Applied Photochromic Polymers Systems, ed. C. B. McArdle, New 1 G. J. Dirksen and G. Blasse, J. Solid State Chem., 1991, 92, 591. York, 1991. 2 A. Diaz and D. A. Keszler, Chem.Mater., 1997, 9, 2071. 21 J. Chrusciel and Z. Lasocki, Pol. J. Chem., 1983, 57, 121. 3 A. Diaz and D. A. Keszler, Mater. Res. Bull., 1996, 31, 147. 22 M. Andrianainarivelo, R. Corriu, D. Leclerq, H. Mutin and 4 H. F. Folkerts and G. Blasse, J.Mater. Chem., 1995, 5, 1547. A. Vioux, J. Mater. Chem., 1995, 5, 719. 5 H. Schmidt and B. Seiferling, Mater. Res. Soc. Symp. Proc., 1986, 23 B. Viana, N. Koslova, P. Aschehoug and C. Sanchez, J. Mater. 73, 739. Chem., 1996, 6, 1665. 6 B. M. Novak, Adv. Mater., 1993, 5, 422. 24 The concentration ratio between Eu2+ and Eu3+ cations has been 7 C. Sanchez and F. Ribot, New J. Chem., 1994, 18, 1007. evaluated through the areas of the absorption bands and taking 8 D. A. Loy and K. J. Shea, Chem. Rev., 1995, 95, 1431. into account the probability of transition which is diVerent for each cation. These probabilities have been estimated from the 9 J. P. Boilot, F. Chaput, T. Gacoin, L. Malier, M. Canva, A. Brun, measured lifetimes. Y. Levy and J. P. Galaup, C. R. Acad. Sci. Ser. IIb, 1996, 322, 27. 10 T ailor-Made Silicon–Oxygen Compounds: From Molecules to Materials, ed. R. Corriu and P. Jutzi, Vieweg, Wiesbaden, 1995. Communication 7/08960K; Received 12th December, 1997 J. Mater. Chem., 1998, 8(3), 507–509

ISSN:0959-9428    

DOI:10.1039/a708960k

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Communication Cyano-bridged cobalt–phthalocyanine dimer: a trapped species by crystallization with a p radical cation Susumu Takano, Toshio Naito and Tamotsu Inabe*† Division of Chemistry, Graduate School of Science, Hokkaido University, Sapporo 060, Japan The molecular structure derived from the X-ray structure analysis‡ is shown in Fig. 1. Since the whole Electrochemical oxidation of the solution containing [(CN) (Pc)CoMCNMCo(Pc)(CN)] unit is crystallographically benzo[c]phenothiazine (B[c]PT) and independent, there are two possible orientations for the dicyanophthalocyaninatocobalt(III ) anion, [Co(Pc)(CN)2]-, bridging CN group.When the structure was refined by fixing gives crystalline solids containing cyano-bridged cobalt– the orientation, the thermal parameters for Cbridge and Nbridge phthalocyanine dimer units.The dimeric species which may were never reasonable; the Beq value for the carbon atom is exist in equilibrium with the monomer is selectively trapped by too small [1.8(2) A ° 2 for CbridgeMCo(1), 2.1(2) A ° 2 for crystallization with the B[c]PT cation radical. CbridgeMCo(46)] and that of the nitrogen atom is too large [3.8(2) and 3.6(2) A ° 2, respectively].The real structure is thus suggested to contain both orientations with equal probability, and the final refinement based on this model gives the improved thermal parameters; Beq of N(44A) is 2.6(2) A ° 2 and that of C(45A) is 2.9(2) A ° 2.6 The slightly bending configuration of the CoMCNMCo backbone is also consistent with other cyano- Phthalocyanine (Pc) is known as an important industrial probridged CoIII complexes7 with simpler ligands (NH3 and CN).duct, and its electronic properties also receive attention from Since the structure is an average of the two orientations, the the viewpoints of photoconductivity and dark conductivity.1 molecule is expected to be symmetrical.Slight deviation from Dicyanophthalocyaninatocobalt(III ) anion, [Co(Pc)(CN)2]-, can be a novel component of multi-dimensional conductors,2 and the neutral radical3 and partially oxidized conductors4 have been obtained by electrochemical oxidation of this anion. Combination of [Co(Pc)(CN)2]- with p-donor components has yielded non-conducting simple salts,5 whereas the salts are expected to be conducting when the composition deviates from 151.During the study of preparing such salts, we have found that benzo[c]phenothiazine, B[c]PT, can form a 151 compound with the dimerized phthalocyanine unit, [(CN)M (Pc)CoMCNMCo(Pc)(CN)] {(m-cyano-C,N)bis[cyano(phthalocyaninato) cobalt(III)]}. This CN-bridged metallophthalocyanine dimer is assumed to exist as a minor species of equilibrium in the acetonitrile solution of [Co(Pc)(CN)2]-, and is selectively trapped and accumulated in the form of a crystal through molecular recognition by the co-existing B[c]PT radical cation.Fig. 1 Perspective view and a part of the atom numbering scheme of the [(CN) (Pc)CoMCNMCo(Pc)(CN)] unit. Selected bond lengths (A ° ) and angles (°): Co(1)MN(44A) 1.986(7), Co(46)MC(45A) 2.070(7), N(44A)MC(45A) 1.130(8), Co(1)MN(44A)MC(45A) 167.5(7), Co(46)MC(45A)MN(44A) 170.2(7).‡ Crystal data: A prism-like single crystal with dimensions of [B[c]PT][(CN) (Pc)CoMCNMCo(Pc)(CN)] was obtained 0.70×0.35×0.05 mm3 was used for X-ray structure analysis. All measurements were made on a Rigaku AFC7R diVractometer with by an electrochemical oxidation technique.K[Co(Pc)(CN)2] graphite-monochromated Mo-Ka radiation (293 K). C83H43N20Co2S, (20 mg) and B[c]PT (8 mg) was dissolved in acetonitrile (ca. M=1470.31, monoclinic, space group P21/a, a=17.680(4), b= 30 ml) and placed in a glass-fritted two-compartment 20.692(4), c=17.599(3) A ° , b=96.31(2)°, V=6399(1) A ° 3, Z=4, and electrochemical cell. A constant current of 1 mA applied m(Mo-Ka)=6.21 cm-1.Of 15 571 reflections measured, 6703 indepenbetween the two platinum electrodes immersed in each dent reflections with I>3.0s(I) were used for the refinement. The final R is 0.072 and Rw is 0.058 (955 variables). Full crystallographic data, compartment for two weeks yielded typically 10 mg of the excluding structure factors, have been deposited at the Cambridge product crystals.Crystallographic Data Centre (CCDC). See Information for Authors, J. Mater. Chem., 1998, Issue 1. Any request to the CCDC for this material should quote the full literature citation and the reference number 1145/74. † E-mail: inabe@rigaku1.science.hokudai.ac.jp J. Mater. Chem., 1998, 8(3), 511–513 511Fig. 3 DiVuse reflectance spectra of [B[c]PT][(CN) (Pc)CoMCNM Co(Pc)(CN)] (a), K[Co(Pc)(CN)2] (b), and B[c]PT Br (c).f (R) is the Kubelka–Munk function which corresponds to the absorption. Background level for [B[c]PT][(CN) (Pc)CoMCNMCo(Pc)(CN)] is shifted to the marker at the right vertical axis. by equilibrium in the solution of [Co(Pc)(CN)2]- is also Fig. 2 Crystal packing of the [B[c]PT]2[(CN) (Pc)CoMCNM supported by the fact that the formation of the cyano-bridged Co(Pc)(CN)]2 unit (a) and two B[c]PT sandwiched by the two [(CN) (Pc)CoMCNMCo(Pc)(CN)] units represented by the van der polymer, [Co(Pc)(CN)]n, was achieved by simply refluxing Waals radii (b); for the latter only the CoMCNMCo fragments Na[Co(Pc)(CN)2] in water.8 In this situation, dissociated are shown NaCN preferentially stays in the liquid phase, so that the equilibrium moves toward promoting the polymerization.the symmetrical geometry is due to the crystallographically The electrical conductivity of [B[c]PT][(CN) (Pc)CoM diVerent environment for each Pc ring. CNMCo(Pc)(CN)] is very poor, <10-8 V-1 cm-1. This is to A part of the crystal packing is shown in Fig. 2(a). Two Pc be expected, since the crystal consists of open-shell cations rings in the [(CN) (Pc)CoMCNMCo(Pc)(CN)] unit are not enclosed by discrete closed-shell anions.Conduction through parallel. Two B[c]PTs are located at the wide opened site, and the overlap between Pc and B[c]PT is not likely owing to the the shape of B[c]PT is just fit for the space formed by the two large diVerence between their ionization potentials. CN-bridged dimeric Pc units [Fig. 2(b)].This may be the On the other hand the cyano-bridged metallophthalocyanine reason why B[c]PT can form a 151 crystal with polymers are known to be conducting without doping with [(CN) (Pc)CoMCNMCo(Pc)(CN)]. Other p-donors always oxidizing agents,8–10 while it is not clear why they are conform a salt with [Co(Pc)(CN)2]-. ducting. Considering the CoMCNMCo distance (5.186 A ° ), it Since the product crystals are obtained at the anode, virtually is possible for only some of the Pc rings in neighboring no reducing process occurs.Therefore, the oxidation state of polymer chains to be interleaved with each other. Though it Co is reasonably assigned as +3, the same as the starting was suggested that p–p overlap still plays some role in charge monomeric anion.If no oxidation of the Pc ring occurs, the transfer,9 the conduction through the MCoM(CONMCoM)n dimeric species can be considered as a monoanion, backbone may not be ruled out. [(CN) (Pc)CoMCNMCo(Pc)(CN)]-. On the other hand, the The optical spectroscopic measurements were carried out first oxidation potential of B[c]PT, (0.62 V vs. Ag/AgCl in using powder samples.Fig. 3 shows the diVuse reflectance acetonitrile) is much lower than that of [Co(Pc)(CN)2]- spectrum of [B[c]PT][(CN) (Pc)CoMCNMCo(Pc)(CN)]. (1.05 V under the same conditions). Therefore, the product is Since the absorption by [B[c]PT]+ in the visible region is assigned as a simple salt, and the electrochemical process is relatively weak, the Q-band of Pc is clearly seen in the related only to B[c]PT.The formation of the dimer is, thus, spectrum. The band shape is practically unchanged compared generated by equilibrium in the acetonitrile solution, as shown with the spectrum of K+[Co(Pc)(CN)2]-, though a slight in Scheme 1. The existence of the dimer (or higher oligomer) blue-shift (ca. 10 nm) can be seen. These data also suggest that the p–p interaction between the Pc rings is negligibly small.The CN stretching mode in the FTIR spectra using a KBr disc specimen was so weak that it is not detectable with a reliable S/N ratio. In conclusion, we have found that [(CN) (Pc)CoMCNM Co(Pc)(CN)]- can selectively be trapped and accumulated in the form of a crystal by co-existence of the B[c]PT radical cation. Since it is extremely diYcult to synthesize this type of dimer with high purity, the crystal growth based on molecular recognition by a gradually produced counter ion is an advantageous method for the isolation of such a chemical species. This work was partly supported by a Grant-In-Aid for Scheme 1 Proposed mechanism for the formation of [B[c]PT][(CN) (Pc)CoMCNMCo(Pc)(CN)] Scientific Research, from the Ministry of Education, Science 512 J.Mater. Chem., 1998, 8(3), 511–5136 Parameters for C(45B) and N(44B) are fixed as the same as those and Culture, Japanese Government and Tokuyama Science of N(44A) and C(45A), respectively. Foundation. 7 B. C. Wang, W. P. Schaefer and R. E. Marsh, Inorg. Chem., 1971, 10, 1492; F. R. Fronczek and W. P. Schaefer, Inorg. Chem., 1974, 13, 727. 8 J. Metz and M. Hanack, J. Am. Chem. Soc., 1983, 105, 828. 9 M. Hanack, R. Polley, S. Knecht and U. Schlick, Inorg. Chem., References 1995, 34, 3621. 1 T. J. Marks, Angew. Chem., 1990, 102, 886; Angew. Chem., Int. Ed. 10 M. Hanack, A. Datz, R. Fay, K. Fischer, U. Keppeler, J. Koch, Engl., 1990, 29, 857, and references therein. J. Metz, M. Mezger, O. Schneider and H.-J. Schulze, in Handbook 2 T. Inabe, Y. Maruyama and T. Mitsuhashi, Synth. Met., 1991, of Conducting Polymers, ed. T. A. Skotheim, Marcel Dekker, New 41–43, 2629. York, 1986; M. Hanack, K. Du� rr, A. Lange, J. Osý�o Barcina, J. Pohmer and E. Witke, Synth. Met., 1995, 71, 2275; M. Hanack 3 K. Morimoto and T. Inabe, J.Mater. Chem., 1995, 5, 1749. and M. Lang, Adv.Mater., 1994, 6, 819. 4 T. Inabe and Y. Maruyama, Bull. Chem. Soc. Jpn., 1990, 63, 2273. 5 H. Hasegawa, S. Takano, N. Miyajima and T. Inabe, Synth. Met., 1997, 86, 1895. Communication 7/07540E; Received 20th October, 1997 J. Mater. Chem., 1998, 8(3), 511–513

ISSN:0959-9428    

DOI:10.1039/a707540e

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Communication Synthesis of phthalocyanine-doped silica mesostructured materials by ferrocenyl surfactant H. S. Zhou,a H. Sasabeb and I. Honmaa aEnergy Division, Electrotechnical L aboratory (ET L) 1-1-4, Umezono, T sukuba, Ibaraki 305, Japan bNanophotonics L aboratory, Frontier Research Program, T he Institute of Physical Research (RIKEN), Hirosawa 2-1, Wako-shi, Saitama, 351-01, Japan Synthesis of MCM-41 (hexagonal structure) has been described by many groups.1–3 We prepared colored silica- Photosensitive copper phthalocyanine-(CuPc) doped silica mesoporous materials (in powder) have been directly functional surfactant mesostructured materials as follows: 1.0 g 11-ferrocenyltrimethylundecylammonium bromide is dissolved synthesized by a self-organizing coassembly process using a functional surfactant, 11-ferrocenyltrimethylundecyl- in 50 ml H2O, followed by stirring for about 30 min to obtain a homogeneous ferrocenyl TMA orange aqueous solution.ammonium (ferrocenyl TMA) molecules. This simple method shows that it is not necessary to dope molecules into Copper phthalocyanine (CuPc; 0.08 g) is added to the above aqueous solution.Then, the suspension is sonicated for mesoporous channels after the pore becomes opened by calcination. X-Ray diVraction patterns (XRD) show the typical 5–10 min, and stirred for 3–5 days at room temperature to attain solubilization equilibrium. Undissolved organic com- hexagonal diVraction pattern of the mesoporous materials. Optical absorption spectra of these powders show the typical pounds in the solution are removed by centrifugation and filtration to give a yellow–green homogeneous ferrocenyl band of the CuPc molecule. TMA/CuPc aqueous solution.Then, 4.0 g 95% tetraethylorthosilicate (TEOS) is added. Finally, 20 g 35% HCl aqueous solution is added to acidify the above solution, followed by Self-assembling organic–inorganic molecules into highly stirring.Then the solution becomes completely colorless ordered nanostructured architectures has attracted increasing because of the precipitation of the colored powders at the attention because these materials provide a rich source for bottom of the beaker, which indicates that all of the ferrocenyl scientific research and technological applications.1–6 For TMA/CuPc reacts with silica species to form the mesostruc- example, the mesomaterials derived from mesostructured comtured powder.A schematic diagram of the two new synthetic posites can be used in molecular sieves, catalysts, and host– processes is described in Fig. 1. In this case, acidic synthesis of guest materials.1,2,7,8 Recently, the production of functional silica mesostructures through a mediated pathway can form a molecules contained in self-assembled mesostructured macharged S+X-I+ interface: /CuPc@Fe…(CH3)3N+/Br-/+ terials has been a interesting branch of advanced materials H2O-Si-O/, while all of the surfactants have photoabsorbing research.9,10 A conducting carbon wire of poly(acrylonitrile) molecules CuPc at the tail ends.The Pc-doped silica/ferrocenyl confined in the ordered hexagonal channel has been successfully synthesized9 by introducing the monomer molecule into the channel and carrying out a radical polymerization reaction therein.In this work, we have investigated the synthesis of photosensitive mesostructured materials for optical device applications where the photoabsorbing dyes are doped into the mesochannels by a direct self-organizing process of surfactants, not by external doping after the calcination of the channels. If the synthetic path is found to dope functional molecules into mesochannels by a self-organized co-assembly process, it will enable controlled design for the production of functional mesostructured materials for applications such as sensors, photoconversion or luminescent materials.We report the synthesis of CuPc-doped silica mesostructured materials by using a functional surfactant directly instead of alkyltrimethylammonium salt. In this case, CuPc molecules are, supposedly, embedded and self-assembled among the functional surfactant’s hydrophobic tails and organized with a periodic array of the lipid micelle structures. We chose 11- ferrocenyltrimethylundecylammonium (ferrocenyl TMA) bromide as the surfactant in this synthesis process because many organic molecules, such as phthalocyanines, are soluble in ferrocenyl TMA aqueous solution11,12 by being contained in the micelle or vesicle of ferrocenyl TMA.The ferrocenyl TMA surfactant has a ferrocenyl ligand at the lipid tail with a distance of eleven carbons to the trimethylammonium cation head group.The ferrocenyl TMA (I0) and its oxidized form Fig. 1 The self-assembled synthesis method of molecular-doped I+ show that the surfactants form redox-active micelles. The mesostructured materials by using ferrocenyl TMA surfactants to isolated molecules have an absorption band at 440 nm which incorporate CuPc molecules inside the surfactant micelle, assembling with silica (SiO2) frameworks results in the orange color of the surfactants.J. Mater. Chem., 1998, 8(3), 515–516 515TMA mesostructured powder can be obtained after washing and drying in air at room temperature or at temperatures below 60 °C. In this new process, it is not necessary to dope molecules into mesoporous channels after the pores become opened by calcination.The hexagonal mesostructures of CuPc-doped silica mesoporous material powders were confirmed by XRD (Fig. 2). In spite of the lower concentration of surfactants, the hexagonal phase was successfully produced and the products were colored (yellow–green) by the surfactant’s ferrocenyl ligands and CuPc molecules. Fig. 2 shows a typical hexagonal diVraction pattern of the mesoporous materials.The two peaks were observed in the low-angle region for [100] and [110]. The interplanar distance d100=36.4 A ° , which is twice as long as the ferrocenyl TMA surfactant molecule, is basically the same as those of silica MCM products where lipid micelles form the mesochannel structure. The hexagonal unit cell length a=2d100/Ó3=42.3 A ° . The XRD shows that the interplanar distance and hexagonal Fig. 3 Absorption spectra of silica/ferrocenyl TMA mesostructured materials and CuPc-doped silica/ferrocenyl TMA mesostructured unit cell length of CuPc-doped silica mesostructured materials materials are the same as those of undoped materials.13 This phenomenon can be explained by part of a CuPc molecule inserting in 660 nm result from the phthalocyanine chromophore. It dis- between the self-assembled ferrocenyl TMA surfactants’ hydroplays the typical absorption band of phthalocyanines, which phobic tails in a periodical manner within a mesostructured is slightly broader than that of the CuPc molecule only in channel.Therefore, it is unlikely that CuPc molecules are organic solvent. This phenomenon results from interaction located in the center of the ferrocenyl TMA micelle, which between the CuPc molecules.Because phthalocyanine aggre- would expand the micelle size very much.11,12 CuPc molecules gation results in a much broader absorption band, an only inserted between ferrocenyl TMA hydrophobic tails would slightly broadened absorption band indicates that the CuPc slightly enlarge the channel size, and may stabilize the silica molecules lie among the functional surfactant’s hydrophobic (SiO2) framework.In the case of silica, because of the amorphtails and are organized with a periodic array of the lipid ous nature of the Si–O networks, a curved interior surface micelle structures, not an aggregated CuPc cluster in the center such as in the hexagonal phase is possible even in Pc-doped of the ferrocenyl TMA micelle.mesostructured materials. This is why ferrocenyl TMA/SiO2 We directly synthesized CuPc-doped silica mesostructured shows a hexagonal phase. materials. The mesostructure has been investigated and proved Fig. 3 shows the absorption spectrum of a undoped ferroby X-ray diVraction. This communication reports a new and cenyl TMA/SiO2 powder,13 and two absorption bands at simple method for the preparation of functional molecular- 440 nm and 640 nm were clearly seen.The absorption at doped silica-functional mesostructured materials. Because 440 nm was ascribed to the reduced state (original state) of the many organic molecules are soluble in surfactant aqueous ferrocenyl ligand and the one at 640 nm was an oxidized state.solution by being contained in the micelle or vesicle of the Because the ferrocenyl dyes are incorporated in the channel surfactant, eVorts are being made to synthesize films of by a self-assembly process, the absorption of the products functional mesostructured composite materials for optical comes from concentrated dyes at the channel center. The device applications.absorption band at 440 nm is a reduced state and is identical to the absorption of isolated Fe-TMA molecules in the solution, References while the one at 640 nm, an oxidized state, arose from reaction with air (oxygen) as well as moisture after the synthesis. As 1 C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and the products are exposed to air at about 100 °C, the relative J.S. Beck, Nature (L ondon), 1992, 359, 710. 2 J. S. Beck, J. C. Vartuli, W. J. Roth, M. E. Leonowicz, C. T. Kresge, absorption intensity ratio of the 640 nm and 440 nm bands K. D. Schmitt, C. T-W. Chu, D. H. Olson, E. W. Sheppard, increases with time. The increase of the absorption intensity S. B. McCulen, J. B. Higgins and J. L. Schlenker, J. Am. Chem. ratio indicates oxidation of the sample by heat aging.Soc., 1992, 114, 10835. Fig. 3 also shows the absorption spectra of Pc-doped silica 3 Q. Huo, D. I. Margolese, U. Ciesla, D. G. Demuth, P. Feng, mesoporous materials including the ferrocenyl TMA surfac- T. E. Gier, P. Sieger, A. Firouzi, B. F. Chmelka, F. Schuth and tant. The absorption bands located at about 730 nm and G. D. Stucky, Chem.Mater., 1994, 6, 1176. 4 P. T. Tanev and T. J. Pinnavaia, Science, 1995, 267, 865. 5 G. S. Attard, J. C. Glyde and C. G. Goltner, Nature (L ondon), 1995, 378, 367. 6 P. T. Tanev, M. Chibwe and T. J. Pinnavaia, Nature (L ondon), 1994, 368, 321. 7 T. Yanagisawa, T. Shimizu, K. Kuroda and C. Kato, Bull. Chem. Soc. Jpn., 1990, 63, 988. 8 Q. Huo, D. I. Margolese, U. Ciesla, P. Feng, T. E. Gier, P. Sieger, R. Leon, P. M. PetroV, F. Schuth and G. D. Stucky, Nature (L ondon), 1994, 368, 317. 9 C.-G. Wu and T. Bein, Science, 1994, 266, 1013. 10 C.-G. Wu and T. Bein, Chem.Mater., 1994, 6, 1109. 11 T. Saji, K. Hoshino and S. Aoyagui, J. Am. Chem. Soc., 1985, 107, 6865. 12 T. Saji, K. Hoshino, Y. Ishii and M. Goto, J. Am. Chem. Soc., 1991, 115, 450. 13 H. S. Zhou, H. Sasabe and I. Honma, to be published. Fig. 2 The X-ray diVraction pattern of the CuPc-doped silica/ferro- Communication 7/07884F; Received 3rd November, 1997 cenyl TMA. The interplanar distances of the hexagonal unit cell d100, d110, d200 are 36.4 A° , 21.0 A° and 18.2 A° respectively. 516 J. Mater. Chem., 1998, 8(3), 515–516

ISSN:0959-9428    

DOI:10.1039/a707884f

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Communication Chemical tailoring of the charging energy in metal cluster arrangements by use of bifunctional spacer molecules Ulrich Simon,*† Renate Flesch, Hartmut Wiggers, Gu�nter Scho�n and Gu�nter Schmid Universita� t GH Essen, Institut fu� r Anorganische Chemie, Universita�tsstr. 5–7, 45117 Essen, Germany with the formation of H2O2. The oxygen-free cluster Pd561phen36 now provides very active surface sites which can be The insertion of bifunctional spacer molecules into a threedimensional arrangement of Pd561phen36O200 clusters leads to coordinated by the NH2 groups of 4,4¾-diamino-1,2-diphenylethane, the spacer molecule.Generally, the reaction described an increase of the charging energy from 0.02 eV to 0.05 eV, which was determined from the temperature dependence of the here is suitable for every bifunctional molecule with terminal NH2 groups.This particular spacer molecule was chosen, since conductivity. (i) it is large enough to lead to suYcient stretching of the packing, and (ii) steric hinderance rules out folding of the molecule as well as the possibility to bind to the same cluster with both termini.The latter could be the case if a,v-diaminoal- Recent progress in chemistry allows the synthesis of well kanes are used. This means that the spacer may be regarded to be quasi-stiV with respect to the purpose discussed here. As defined nanoscaled metal or semiconductor particles with uniform size and high reproducibility.1 With respect to possible the naked cluster surface places are oriented in all directions the spacing procedure leads to an insoluble precipitate with applications in microelectronics we illustrated recently2 that some of these materials have potential uses resulting from their three-dimensional cluster linkage.This network exhibits an increased inter-particle spacing with respect to the closest electronic inter-particle interactions.If metallic particles with a size of a few nanometers are arranged in a small spatial sphere packing of the non-modified cluster, like in a pressed pellet or a layer deposited from solution. distance of ca. 1 nm, they build tunnel junctions with electrical capacitances down to 10-19 F. This allows controlled charge The charging energy, EC, i.e. the energy barrier that has to be overcome to transfer a single electron from an initially transport between the particles by single electron tunneling (SET) up to around room temperature, which has been recog- neutral cluster to a neutral nearest neighboring cluster, is dependent of the inter-particle capacitance C, as follows from nized to be a fundamental concept for ultimate miniaturisation in electronics.3 Inspired by this fascinating idea, many eVorts EC=e2/2C where e is the charge of the electron.It can be determined directly from the temperature dependence of the have been made to tailor one-, two- and three-dimensional metal nanoparticle arrangements and to examine experimen- dc conductivity [s(T )] of the cluster arrangements, which were investigated in this work on pressed pellets by means of a tally4,5 and theoretically6 the electrical properties of these materials.Most recently a single electron transistor has been Keithley 6517 electrometer. Earlier investigations on the electrical properties of these ligand stabilized metal clusters have constructed using alkanedithiol stabilized gold nanoparticles as tunnel junctions.5 shown that even at high temperatures thermally activated electron hops between nearest neighbors, instead of hops of The family of ligand stabilized metal clusters like Au55- (PPh3)12Cl6, Pt309phen*36O30 or Pd561phen36O200 (phen*= variable range, dominate the charge transport, giving a temperature dependence according to the Arrhenius relation lns(T ) batho-phenanthroline, phen=1,10-phenanthroline) are uniform, chemically tailored metallic nanoparticles in the size #EA/kBT , where EA is the activation energy.8 This is confirmed by this work, as we found the simple activated behavior over range 1.4–2.4 nm. They consist of a metal core with a well defined number of atoms, which is surrounded or ‘dressed’ by the temperature range 80–300 K.Fig. 2 shows the Arrhenius plot, in which the slopes of the straight lines correspond to a protecting ligand shell.If these particles are attached to each other the ligand shell acts like a dielectric spacer between the the activation energy of the charge transport, i.e. the charging energy of the particles in the three-dimensional arrangement. metal cores. The electrical capacitance which is built up between them determines the charging energy for inter-particle While the close packing of the cluster material shows an activation energy of 0.02 eV the insertion of the spacer mol- electron transitions and it depends on the size of the metal core as well as on the thickness and chemical nature of the ecules increases the value to 0.05 eV.Correspondingly the capacitance decreased from initially 4.0×10-18 F down to ligand shell.In this work we report the chemical tailoring of the charging 1.6×10-18 F. The specific conductivity follows the same trend, energy by insertion of bifunctional dielectric spacer molecules into a three-dimensional arrangement of Pd561phen36O200 clusters. The basic idea is to stretch the cluster package in comparison to densest sphere packing and by this to increase the inter-particle spacing, which denotes the spatial distance between the surface of neighboring clusters (see Fig. 1). This should lead to an increase of the charging energy, i.e. a decrease of the electrical capacitance between the clusters. The spacing of the Pd561phen36O200 cluster, the synthesis of which is described elsewhere,7 starts with its deoxygenation by hydrogen in a water–pyridine solution at room temperature Fig. 1 Schematic illustration of the insertion of spacer molecules into a three-dimensional arrangement of clusters, which leads to an increase of the spatial distance of the cluster surfaces † E-mail: u.simon@uni-essen.de J. Mater. Chem., 1998, 8(3), 517–518 517backbone, e.g. due to p-conjugation, to determine the spacing definitely by the molecule length.Financial support by the Bundesminister fu� r Bildung, Wissenschaft und Forschung (BMBF) under contracts No. 03N1012A7 and No. 03N1012B0 is gratefully acknowledged. References 1 Clusters and Colloids, ed. G. Schmid, VCH,Weinheim, 1994. 2 G. Scho�n and U. Simon, Colloid Polym. Sci., 1995, 273, 101; 202. 3 K. K. Likharev, IBM J. Res. Dev., 1988, 32, 144; Single Charge Fig. 2 Arrhenius plot of the Pd561 cluster (a) in dense packing (Ea= T unneling and Coulomb Blockade Phenomena in Nanostructures, ed. 0.02 eV) and (b) with inserted spacer molecules (3-D network; Ea= M. H. Devoret and H. Grabert, Nato ASI Series, vol. 294, Plenum 0.05 eV) Press, New York, 1992. 4 R. P. Andres, J. D. Bielefeld, J. I. Henderson, D. B. Janes, V.R. Kolagunta, C. P. Kubiak, W. J. Mahoney and R. G. Osifchin, as might be expected due to the decrease of the volume fraction Science, 1996, 273, 1690; J. P. Spatz, A. Roescher, S. Sheiko, of the metal in the volume of the sample. Thus, the trend G. Krausch and M. Mo� ller, Adv.Mater., 1995, 7, 731; R. L. Whetten, observed is in agreement with the results found by SchiVrin J. T. Khoury, M.M. Alvarez, S. Murthy, I. Vezmar, Z. L. Wang, et al.9 who investigated 2.2 nm and 8.8 nm colloidal gold P. W. Stephens, Ch. L. Cleveland, W. D. Luedtke and U. Landman, Adv.Mater., 1996, 8, 428. nanoparticles with interconnecting alkanedithiols. 5 T. Sato, H. Ahmed, D. Brown and B. F. H. Johnson, J. Appl. Phys., In conclusion, this work provides one example of electrical 1997, 82, 696.capacitance between metal clusters in a three-dimensional 6 M. P. Samanta, W. Tian, S. Datta, J. I. Henderson and C. P. arrangement; by this the charging energy can be chemically Kubiak, Phys. Rev. B, 1996, 53, 7626; V. Gasparian and U. Simon, tailored by use of bifunctional spacer molecules. The decrease Physica B, in press; U. Simon and V. Gasparian, Phys. Status Solidi of EC, which was determined by the tempeB, in press. 7 G. Schmid, M. Harms, J.-O. Malm, J.-O Bovin, J. van Ruitenbeck, of the conductivity, proved that chemical control of the physical H. W. Zandbergen and W. T. Fu, J. Am. Chem. Soc., 1993, 115, 2046. properties of cluster arrangements is possible, which contri- 8 M. P. J. van Staveren, H. B. Brom and L. J. de Jongh, Phys. Rep., butes to an understanding of the structure–property relation- 1991, 208, 1. ship in these nanomaterials. The power of this strategy will be 9 M. Brust, D. Bethell, D. J. SchiVrin and Ch. J. Kiely, Adv. Mater., examined in future works, where we will extend our investi- 1995, 7, 795. gations to even smaller clusters like Au55(PPh3)12Cl6 with a variety of spacer molecules with variable length, and a rigid Communication 7/07544H; Received 20th October, 1997 518 J. Mater. Chem., 1998, 8(3), 517–518

ISSN:0959-9428    

DOI:10.1039/a707544h

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials A comparison of two convergent routes for the preparation of metalloporphyrin-core dendrimers: direct condensation vs. chemical modification† Keith W. Pollak,a Elizabeth M. Sanfordb and Jean M. J. Fre�chet*a aDepartment of Chemistry, University of California, Berkeley, CA 94720–1460 USA bHope College, Holland,Michigan 49422-9000, USA Porphyrin-core dendrimers consisting of benzyl ether dendrons assembled around a porphyrin core have been prepared by two diVerent convergent syntheses.The first involves the direct condensation of convergent dendrons having 3,5-disubstituted benzaldehyde focal points with an equivalent amount of pyrrole. This route which requires very mild conditions is especially useful for the rapid assembly of small dendrimers but suVers from steric limitations as the size of the dendrons increases above the fourth generation.The second route involves the attachment of pre-formed benzyl bromide dendrons to a functionalized porphyrin through a simple Williamson synthesis. This route is also very practical but it requires careful purification of the final product from the partially functionalized porphyrin dendrimers that are also obtained.MALDI mass spectrometry proved to be a very useful tool both for monitoring the formation of the dendritic porphyrins and for their characterization. Metalloporphyrins with specific architecture have been devel- of dendrons with an aldehyde functionality at their focal point, and their subsequent assembly into a porphyrin by Lindsey12 oped to model various biological systems; two of the common natural archetypes are chlorophyll, and heme proteins.The condensation with pyrrole. Until now, this in situ assembly of porphyrin-core dendrimers had not been reported. light-harvesting and electron-transfer properties of chlorophyll have been modelled by synthetic metalloporphyrins with sub- In view of the potential application of site-isolated porphyrin nuclei for electron transfer or other catalytic processes, this stituents which modify the photochemical behavior of the porphyrin.1 Other metalloporphyrins have been synthesized study explores and compares the synthesis of metalloporphyrin- core dendrimers of generations 1–4 via two routes utilizing with either one or both faces of the porphyrin sterically obstructed to model proteins like hemoglobin and myoglobin the convergent growth approach.Route I, in which dendritic aldehyde and pyrrole are combined in a Lindsey porphyrin which bind and transport molecular oxygen.2 But the most common purpose for synthesizing porphyrins with defined synthesis to generate the porphyrin core in situ (Scheme 1) and route II, in which zinc tetrakis(3,5-dihydroxyphenyl)porphy- architecture is to model natural oxidation catalysts like cytochromes.3,4 rin, a porphyrin core with eight reactive phenolic sites,8 is alkylated with a dendritic bromide in a Williamson ether Some of the latest biological models embed a porphyrin at the core of a dendrimer.The synthesis of such a macromolecule synthesis (Scheme 2).could follow either the divergent5 or convergent6 approach to dendrimers, and indeed both routes have been used by diVerent research groups.7–10 For example Diederich and co-workers7 Results and Discussion have used the divergent approach to grow a polyamide den- Direct formation of the porphyrin core from a dendritic drimer from a porphyrin core. Although this method has been aldehyde and pyrrole successful for the preparation of large porphyrin-core dendrimers, the sensitivity of the porphyrin moiety must be con- Benzyl ether dendrons were prepared through the iterative sidered in the growth of the dendrimers, a process that typically bromination and alkylation steps described by Hawker and involves multiple preparative steps and demanding purification Fre� chet.6a,11 To prepare the desired generation of dendritic procedures.Such potential problems could be avoided by aldehyde, dendritic bromide was used to alkylate 3,5- incorporating the porphyrin core into the macromolecule at dihydroxybenzaldehyde (Scheme 3). Classical Lindsey condenthe last step of the synthesis,8 as is common practice in the sation12 of the dendritic aldehyde and an equivalent amount synthesis of dendrimers through the convergent approach.6 of pyrrole at a concentration of 10-2 M in chloroform was A convenient route for the synthesis of porphyrin-core done at room temperature in the presence of a catalytic amount dendrimers involves the attachment of convergent dendrons of trifluoroacetic acid, under very mild conditions that can be to a pre-formed porphyrin core.Aida and co-workers8 have tolerated by various functionalities which might be present in used the convergent approach to prepare polyether dendrons the dendrons. Monitoring by TLC proved to be eVective and then attached the dendrons to a functionalized tetraaryl because the starting materials, desired product, and polypyrrylporphyrin, tetrakis(3,5-dihydroxyphenyl)porphyrin, through a methane by-products exhibit very diVerent Rf values.However, Williamson ether synthesis. Similarly, Moore et al. attached analysis by UV–VIS spectroscopy allowed for a more accurate polyester dendrons, prepared through the convergent evaluation of the progress of the condensation because the approach, to a metalloporphyrin core using dicyclohexycar- desired product absorbs at sharp, characteristic wavelengths: bodiimide (DCC) coupling.10 A less obvious convergent route a strong Soret band at 424 nm and four weak additional bands to porphyrin-core dendrimers might involve the preparation at 510, 550, 590 and 650 nm while the polypyrrylmethane byproducts have a broader absorption (Fig. 1). When the relative intensities of these bands stabilize, the condensation is at † Presented at the Third Internation Conference on Materials equilibrium.As the generation number is increased the reaction Chemistry, MC3, University of Exeter, Exeter, July 21–25 1997. * E-mail: Frechet@cchem.berkeley.edu became more sluggish and reaction times increased from 1.5 h J. Mater. Chem., 1998, 8(3), 519–527 519for generations 1 and 2, to 10 and 36 h for generation 3 and NMR was facilitated by the high degree of symmetry in the macromolecule.Evidence for the free-base porphyrin core of 4, respectively, though the latter required diVerent reaction conditions. For the first three generations, the porphyrinogen 3a–d is provided by the eight b-pyrrole and two free-base protons seen as singlets at 8.9 and -2.9 ppm, respectively.The condensation product was oxidized directly after condensation and in the same reaction flask to the free-base porphyrin-core spectral features of the dendrons, which surround the porphyrin core with generational tiers of benzyl ethers, are also readily dendrimers 3a–c in 25–30% yield by heating it to reflux in the presence of chloranil.Preparation of the fourth generation attributed. The phenyl protons located on the outermost phenyl rings are seen at 7.2–7.5 ppm, and those on the aromatic free-base porphyrin-core dendrimer 3d was attempted using these reagents and conditions, but it failed. In the condensation ring at the meso position of the porphyrin appear as a oneproton triplet at 7.10 ppm and a two-proton doublet at step, formation of porphyrin was observed in the reaction samples both by TLC and UV–VIS spectroscopy, yet after the 7.50 ppm.The benzyl protons appear as a series of singlets around 4.6–5.0 ppm, depending on the generation number. oxidation step, no porphyrin was produced. This was probably due to in situ decomposition of the product as a result of the Not all of the 13C NMR resonances for the dendrons can be assigned because of the many overlapping signals,11 and at elevated reaction temperature required for the oxidation reaction using chloranil.Using a stronger oxidizing agent, 2,3- high generation, the intensities of some of the resonances of the core carbons are too weak to be seen. However when the dichloro-5,6-dicyano-1,4-benzoquinone (DDQ), the oxidation could be performed at room temperature,reby preserving core carbons are detectable, the porphyrin gives rise to three distinct resonances:13 Ca at 145.8 ppm, Cb at 130.6 ppm, and the porphyrinogen and producing 3d in 14% yield.Purification of all of the free-base porphyrin-core dendrimers by column Cmeso at 119.6 ppm. Analysis of the free-base porphyrin-core dendrimers 3a–d by matrix-assisted laser-desorption ionization chromatography proved to be very simple because the desired product elutes much faster than the polypyrrylmethane by- (MALDI) mass spectrometry (Fig. 2) shows that well defined macromolecules exhibiting the expected peaks corresponding products. Analysis of the free-base porphyrin core dendrimers by 1H toM+H+, M+Na+ and/or M+K+ are obtained. 520 J. Mater. Chem., 1998, 8(3), 519–527Scheme 1 Reagents and conditions: i, CF3CO3H; ii, DDQ; iii, Zn(OAc)2 J. Mater. Chem., 1998, 8(3), 519–527 521Scheme 2 Reagents and conditions: i, K2CO3, 18-crown-6 522 J. Mater. Chem., 1998, 8(3), 519–527Fig. 1 (a) UV–VIS spectrum of the crude reaction mixture obtained in the preparation of a generation 3 porphyrin-core dendrimer 3c via Route I.(b) UV–VIS absorption spectrum of pure 3c. phyrin 6 is obtained. Metallation of this porphyrin with zinc acetate produced zinc 5,10,15,20-tetrakis(3,5-dihydroxyphenyl )porphyrin 4, the octafunctional core for the porphyrincore dendrimers. The dendrons used for the subsequent alkylation are readily obtained through the iterative bromination and alkylation steps described by Hawker and Fre� chet.11 Formation of the porphyrin-core dendrimers requires that all eight phenolic pendant groups of the core be alkylated.This Scheme 3 Reagents and conditions: i, K2CO3, 18-crown-6 is best done using a 20% excess of the benzylic bromide dendron while monitoring the progress of the reaction. Thin layer chromatography (TLC) is not suYcient for this purpose The free-base porphyrin-core dendrimers 3a–d were quantibecause some of the partially alkylated cores such as those tatively metallated by dissolving the macromolecule and zinc with six or seven dendrons added to the core exhibit an Rf acetate in methanol–chloroform (151) and heating at reflux value similar to that of the fully alkylated product.In the final overnight.Spectroscopic confirmation of the metallation reacstages of the reaction, monitoring by MALDI mass spection is provided by the UV–VIS absorption spectra since the troscopy is more eVective as the partially alkylated cores are metallated dendrimers 1a–d exhibit a strong Soret band at readily identified as shown in Fig. 3. 430 nm and two additional bands at 560 and 600 nm.14 In Careful control of the reaction temperature is required addition, the 1H NMR spectra of the products no longer during the alkylation as temperature aVects both the rate and exhibit the characteristic signal for the two protons located the site of alkylation.If the temperature is significantly higher inside the porphyrin ring of the free-base starting materials. than 60 °C, the reaction rate increases but significant Calkylation involving the 2 position of the phenyl ring of the Preparation by dendron alkylation of a pre-formed porphyrin porphyrin core is observed as evidenced by the presence of core two singlets of equal intensity at 7.82 and 8.59 ppm characteristic of the two remaining phenyl protons in the 1H NMR As will be seen below, a comparison of this approach first reported by Aida and co-workers8 with the direct Lindsey- spectrum of the product mixture.Below 60 °C, the extent of C-alkylation is insignificant but the reaction times increase type synthesis shows that each route has its own advantages and shortcomings. considerably. A reaction temperature of 60 °C produced a good compromise aVording primarily O-alkylated product in a Specifically, this route requires that a porphyrin core with multiple reactive functionalities be prepared first then coupled reasonable reaction time.As expected in view of increased steric requirements around the core, the reaction time required to the appropriate number of benzyl ether dendrons. For example, Scheme 4 shows the preparation of a core with eight for complete alkylation varies as a function of generation: 1, 3 and 5 d for generation 2, 3 and 4, respectively. Regardless of protected phenolic functionalities by condensation of pyrrole and 3,5-dimethoxybenzaldehyde.The resulting porphyrinogen conditions used the final product must be purified by column chromatography to remove by-products or partially alkylated is then oxidized to the free base 5,10,15,20-tetrakis(3,5-dimethoxyphenyl) porphyrin 5 in 52% yield.Following removal of products. Purification is best achieved by chromatography through silica, eluting with a continuous linear gradient from the methoxy protecting groups by reaction with boron tribromide, the free base 5,10,15,20-tetrakis(3,5-dihydroxyphenyl)por- hexane to CH2Cl2. Isolation of the desired product by column J.Mater. Chem., 1998, 8(3), 519–527 523Fig. 2 MALDI mass spectra of the generation 1–4 free-base porphyrin-core dendrimers prepared via Route I. Sodium and potassium adducts are clearly seen at the right of the main peaks. (a) Generation 1, (b) 2, (c) 3 and (d) 4. Fig. 3 MALDI mass spectrum of the crude product obtained in the preparation of fourth generation porphyrin-core dendrimer 1d via Scheme 4 Reagents and conditions: i, BF3 OEt2; ii, DDQ; iii, BB3; Route II. Partly alkylated dendrons are clearly seen at m/z 10 200 iv, Zn(OAc)2 and 11 800. chromatography is diYcult since the excess dendritic bromide, be recovered. Since high generation dendrons require considersome partially alkylated core, and any by-products (i.e.C- able synthetic eVort, recovering the unreacted starting material alkylated core) behave very similarly to the desired product is often desirable. After evaporating the solvent, the product on a silica gel column. However, with careful chromatography, was isolated as a violet glass. Subsequent trituration with methanol allowed the porphyrin-core dendrimer to be handled pure product is isolated, and the excess dendritic bromide can 524 J.Mater. Chem., 1998, 8(3), 519–527as a powder. Dendrimers 1b, 1c, and 1d were prepared by this inside the porphyrin ring, and specific sites of the porphyrin ring are designated by a, b and meso. route in 71, 68 and 20% yields respectively. General procedure for the preparation of dendritic benzyl bromides11 Conclusions To a mixture of the appropriate dendritic benzylic alcohol This study demonstrates the versatility of the convergent (1.00 equiv.) and carbon tetrabromide (1.25 equiv.) in the synthesis for the preparation of dendritic porphyrins.A clear minimum amount of dry THF was added triphenylphosphine advantage of the two convergent approaches described herein (1.25 equiv.), and the reaction mixture was stirred under is that few steps involving dendrimers are involved; this helps nitrogen for 20 min.The reaction mixture was then poured ensure that impurities or unreacted functionalities do not onto water and extracted with CH2Cl2. The combined organic accumulate as might be the case in a divergent synthesis extracts were dried (MgSO4) and evaporated to dryness. The involving multiple steps for the growth of the dendrimer onto crude product was purified by column chromatography.the porphyrin core. This may be particularly useful in instances Analyses agreed with those published.11 where very accurate structural control is required or where the end-functionalities of the porphyrin ring itself might not General procedure for preparation of porphyrin-core dendrimers survive the reactions used during multistep syntheses. via Williamson ether synthesis While both convergent routes benefit from the architectural control aVorded by the convergent synthesis, each route oVers The second, third and fourth generation porphyrin-core dendriunique synthetic advantages. The direct Lindsey-type conden- mers were synthesized by coupling zinc tetrakis(3,5-dihydroxysation is done in the presence of a catalytic amount of acid, phenyl)porphyrin 4 to the appropriate dendric bromide in a Williamson ether synthesis.Under nitrogen, zinc tetrakis(3,5- while the alkylation route is done under prolonged basic dihydroxyphenyl)porphyrin (1.00 equiv.) and the dendritic conditions. As a result, the choice of a specific route may be bromide (9.60 equiv) were dissolved in acetone.To this solution, determined by any sensitive functionality that may be present K2CO3 (16.0 equiv.) and 18-crown-6 (1.60 equiv) were added, on the dendrons. A clear advantage of the Lindsey-type route and the mixture was stirred and warmed to 60 °C. The solvent is the ease of both monitoring of product formation and was evaporated, and the residual solids were partitioned purification of the desired free-base dendritic porphyrins by between water and CH2Cl2.The layers were separated, and simple column chromatography. Reaction times for the direct the aqueous layer was extracted with CH2Cl2 (3×). The solvent condensation are also consistently shorter than those for route was evaporated, and the residual solids were adsorbed onto II.However, despite the larger number of steps required for silica (20 ml ), and the crude product was purified by chroma- the chemical modification route it aVords higher yields than tography through a 40 ml silica column eluting with a linear the direct Lindsey-type synthesis, especially in the preparation gradient of solvent from pure hexane to pure CH2Cl2.The of low generation porphyrin-core dendrimers. The chemical product fractions were collected, and the solvent was evapor- modification route appears to be more suitable than the ated. The product was dissolved in a minimum amount of Lindsey route for very large dendrimers since it does not CHCl3 then precipitated into methanol. The precipitate was appear to be as sensitive to steric constraints.It is anticipated then dissolved in a minimum amount of diethyl acetate and that similar trends would be observed for syntheses using other precipitated into diethyl ether. The product was obtained as a types of dendrons. dark purple powder. Finally, since the convergent synthesis of the dendritic porphyrin aVords precise control of the dendrimer architecture Zn[G-2]4P 1b.This was prepared as above from zinc it provides the means for incorporating potential prosthetic tetrakis(3,5-dihydroxyphenyl)porphyrin 4 and [G-1]Br; yield: functionalities at precise locations within the assembly. 71%; UV–VIS l/nm (e): 428 (498 000), 560 (20 000), 600 (7000). Through careful design, functionalizing the dendrons with dH 9.18 (s, 8 H, b-H), 7.48 (d, 8 H, Ar-H), 7.33 (m, 80 H, Ph-H), either specific guest-binding sites or rate-enhancing ligands 7.03 (t, 4 H, Ar-H), 6.73 (d, 16 H, Ar-H), 6.56 (t, 8 H, Ar-H), near the catalytic core could create ‘suprabiotic’4 catalysts. 5.06 (s, 16 H, Bn-H), 4.96 (s, 32 H, Bn-H). dC 160.09, 157.71 (Ar C-O); 149.92 (a C); 144.80 (Ar C-porph); 139.15 (Ar C-CH2); 136.63 (Ph C-CH2); 132.10 (b C); 128.46, 127.88, Experimental 127.43 (Ph C-H); 120.67 (meso C); 115.02, 106.43, 101.65, 96.66 (Ar C-H); 70.00 (Ar/Ph-CH2).Mass spectrum (MALDI-TOF): General directions m/z, calc. 3225.09; found 3228.3 (Calc. for C212H172N4O24Zn; Silica for flash chromatography was Merck Silica Gel 60 C 78.95, H 5.38, N 1.74; found C 79.06, H 5.57, N 1.62%). (230–400 mesh). Matrix-assisted laser-desorption ionization time of flight (MALDI-TOF) mass spectroscopy was per- Zn[G-3]4P 1c.This was prepared as above from zinc formed on a Finnigan Lasermat or Perseptive Voyager DE. tetrakis(3,5-dihydroxyphenyl)porphyrin 4 and [G-2]Br; yield: The energy source was a 337 nm nitrogen laser. Samples were 68%; UV–VIS, l/nm (e): 428 (534 000), 518 (6000), 560 (20 000), prepared as 10-4 M solutions in tetrahydrofuran.The matrix 600 (7000). dH 8.96 (s, 8 H, b-H), 7.48 (s, 8 H, Ar-H), 7.17 (m, was a solution consisting of 0.2 M indoleacrylic acid and 160 H, Ph-H), 7.03 (s, 4 H, Ar-H), 6.67 (d, 16 H, Ar-H), 6.53 7×10-4 M sodium dodecyl sulfate in tetrahydrofuran. Four (d, 32 H, Ar-H), 6.43 (t, 8 H, Ar-H), 6.39 (t, 16 H, Ar-H), 5.06 microlitres of the sample and 40 microlitres of matrix were (s, 16 H, Bn-H), 4.83 (s, 32 H, Bn-H), 4.78 (s, 64 H, Bn-H).dC combined and analyzed. All NMR spectra were recorded as 160.08, 160.00, 157.75 (Ar C-O); 149.79 (a C); 139.25, 139.16 solutions in CDCl3 on a Bruker WM 300 (300 MHz) spec- (Ar C-CH2); 136.62 (Ph C-CH2); 128.40, 127.81, 127.40 (Ph trometer with the solvent proton signal as standard: 7.24 ppm C-H); 120.67, 106.57, 106.25, 101.65, 101.52, 96.67 (Ar C-H); for 1H and 77.0 ppm for 13C.The following abbreviations are 69.89 (Ar/Ph-CH2). Mass spectrum (MALDI-TOF): m/z calc. used: Dendritic aldehydes are referred to as [G-n]CHO, where 6621; found 6630 (Calc. for C436H364N4O56Zn; C 79.09, H 5.54, n designates the dendrimer generation. Dendritic free-base N 0.85; found C 78.98, H 5.96, N 0.50%).porphyrins are referred to as H2[G-n]4P, where n, again, designates the dendrimer generation. Ar refers to the aromatic Zn[G-4]4P 1d. This was prepared as above from zinc repeat units within the dendrimer. Ph refers to the phenyl end- tetrakis(3,5-dihydroxyphenyl)porphyrin 4 and [G-3]Br; yield: groups, the surface, of the dendrimer. Bn refers to the benzylic 20%; UV–VIS, l/nm (e): 430 (463 000), 560 (21 000), 600 (7000).dH 9.00 (s, 8 H, b-H), 7.48 (s, 8 H, Ar-H), 7.19 (m, 320 positions within the dendrimer. Porph refers to the two protons J. Mater. Chem., 1998, 8(3), 519–527 525H, Ph-H), 6.67 (s, 4 H, Ar-H), 6.52–6.38 (overlapping reson- filtrate was flash chromatographed through a silica column, eluting with chloroform.The product fractions were collected ances, 168 H, Ar-H), 4.91 (s, 16 H, Bn-H), 4.75–4.69 (overlapand evaporated to dryness. The residual solid was taken up in ping resonances, 224 H, Bn-H). dC 160.03, 159.90, 159.86, minimal chloroform and precipitated into methanol, so the 157.76 (Ar C-O); 149.68 (a C); 139.16, 139.08 (Ar C-CH2); product could be handled as a powder. 136.62 (Ph C-CH2); 132.10 (b C); 128.40, 127.78, 127.45 (Ph C-H); 106.21, 101.42 (Ar C-H); 69.80 (Ar/Ph-CH2). Mass H2[G-1]4P 3a. This was prepared from [G-1]CHO 2a. The spectrum (MALDI-TOF): m/z calc. 13 413; found 13 437 (Calc. aldehyde and pyrrole were allowed to react for 1.5 h, and the for C884H748N4O120Zn; C 79.16, H 5.62, N 0.42; found C 79.24, oxidation was carried out for 2 h; yield: 30%; UV–VIS, l/nm: H 5.79, N 0.59%). 280 (dendrimer units); 424 (porphyrin Soret); 516, 550, 590, 648 (porphyrin Q-region); dH 8.95 (s, 8H, b-H), 7.39–7.51 (m, General procedure for synthesis of dendritic aldehydes 40H, Ph-H), 7.60 (d, 8H, Ar-H), 7.17 (t, 4H, Ar-H), 5.31 (s, A mixture of the appropriate dendritic bromide (2.00 equiv.), 16H, Bn-H), -2.82 (s, 2H, porph-H); dC 143.99 (a C), 127.63 3,5-dihydroxybenzaldehyde (1.00 equiv.), potassium carbonate (b C), 119.68 (meso C), 70.38 (CH2O), 127.68, 128.07, 128.64, (3.00 equiv.), and 18-crown-6 (0.20 equiv.) was refluxed under 136.79 (PhC).Mass spectrum (MALDI): m/z, calc. 1464; nitrogen in dry THF for 48 h. The reaction was allowed to found 1469. cool then evaporated to dryness under reduced pressure. The residue was partitioned between water and methylene chloride, H2[G-2]4P 3b.This was prepared from [G-2]CHO 2b. The and the aqueous layer was extracted with methylene chloride. aldehyde and pyrrole were allowed to react for 1.5 h, and the The combined organic layers were then dried (MgSO4) and oxidation was caried out for 2 h; yield: 26%; UV–VIS l/nm: evaporated to dryness. The product was purified by flash 280 (dendrimer units); 424 (porphyrin Soret); 516, 550, 590, chromatography through a silica column, eluting with methyl- 648 (porphyrin Q-region); dH 8.87 (s, 8H, b-H), 7.21–7.39 (m, ene chloride to give dendritic aldehyde as a colorless glass. 80H, Ph-H), 7.49 (d, 8H, Ar-H), 7.05 (t, 4H, Ar-H), 6.76 (d, 16H, Ar-H), 6.59 (t, 8H, Ar-H), 5.15 (s, 16H, Bn-H), 4.98 (s, [G-1]CHO 2a. 3,5-Dibenzyloxybenzaldehyde was pur- 32H, Bn-H), -2.89 (s, 2H, porph-H); dC 141.45 (a C), 128.51 chased from Aldrich and used directly. (b C), 119.43 (meso C), 70.38, 70.18 (CH2O), 99.02, 161.14, 107.08, 139.14 (ArC), 127.44, 127.90, 128.48, 136.68 (PhC). [G-2]CHO 2b. This was prepared from [G-1]Br; yield: Mass spectrum (MALDI): m/z, calc. 3162; found 3171. 83%; n/cm-1 2873 [Ar(CO)MH st.], 1669 cm-1 [Ar(CNO)H st.]; dH 9.87 (s, 1 H, ArCHO), 7.28–7.41 (m, 20 H, Ph-H), 7.05 H2[G-3]4P 3c.This was prepared from [G-3]CHO 2c. The (d, 2 H, Ar-H), 6.81 (t, 1 H, Ar-H), 6.67 (d, 4 H, Ar-H), 6.59 aldehyde and pyrrole were allowed to react for 10 h, and the (t, 2 H, Ar-H), 5.07 (s, 8 H, Bn-H), 5.03 (s, 4 H, Bn-H); dC oxidation was run for 5 h; yield: 29%; UV–VIS, l/nm: 280 70.4, 70.6 (CH2O); 106.6, 107.7, 137.0, 160.6 (ArC); 102.0, 104.7, (dendrimer units); 424 (porphyrin Soret); 516, 550, 590, 648 136.7, 160.5 (ArC); 128.0, 128.4, 128.9, 139.2 (PhC); 198.6 (porphyrin Q-region); dH 8.91 (s, 8H, b-H), 7.19–7.39 (m, 160H, (ArCHO).Mass spectrum (MALDI): m/z, calc. 743; found 767. Ph-H), 7.51 (d, 8H, Ar-H), 7.08 (t, 4H, Ar-H), 6.72 (d, 16H, Ar-H), 6.60 (d, 32H, Ar-H), 6.56 (t, 8H, Ar-H), 6.49 (t, 16H, [G-3]CHO 2c.This was prepared from [G-2]Br; yield: Ar-H), 5.09 (s, 16H, Bn-H), 4.90 (s, 32H, Bn-H), 4.88 (s, 64H, 49%; n/cm-1 2873 [Ar(CO)MH st.], 1683 cm-1 [Ar(CNO)H Bn-H), -2.91 (s, 2H, porph-H); dC 143.93 (a C), 128.53 (b C), st.]; dH 9.84 (s, 1 H, ArCHO), 7.29–7.42 (m, 40 H, Ph-H), 7.08 119.69 (meso C), 69.89, 69.88, 70.20 (CH2O), 101.51, 101.53, (d, 2 H, Ar-H), 6.84 (t, 1 H, Ar-H), 6.66 (d, 8 H, Ar-H), 6.65 106.59, 106.23, 139.17, 139.14, 157.89, 160.03 (ArC), 127.49, (d, 4 H, Ar-H), 6.55 (t, 4 H, Ar-H), 6.54 (t, 2 H, Ar-H), 5.00 128.06, 128.53, 136.67 (PhC).Mass spectrum (MALDI) m/z, (s, 16 H, Bn-H), 4.99 (s, 4 H, Bn-H), 4.95 (s, 8 H, Bn-H); dC calc. 6557; found 6577. 69.9–70.2 (CH2O); 106.4, 107.5, 136.7, 160.1 (ArC); 101.9, 103.9, 138.8, 160.1 (ArC); 127.7, 128.1, 128.7, 139.0 (PhC); 197.2 H2[G-4]4P 3d.Pyrrole (0.043 ml; 0.626 mmol) and 2d (ArCHO). Mass spectrum (MALDI): m/z, calc. 1592; found (2.059 g; 0.626 mmol) were dissolved in dry CHCl3 and con- 1616. densed in the presence of TFA (0.025 ml; 0.207 mmol). The solution was shielded from ambient light and stirred at room [G-4]CHO 2d.This was prepared from [G-3]Br; yield: temp. for 36 h. Then DDQ (0.177 g; 0.782 mmol) was added, 56%. n/cm-1 2873 [Ar(CO)MH st.], 1694 cm-1 [Ar(CNO)H and the solution was allowed to stir at room temp. for an st.]; dH 9.84 (s, 1H, ArCHO), 7.26–7.41 (m, 80 H, Ph-H), 7.08 additional 2 h. After cooling, the solution was evaporated to (d, 2 H, Ar-H), 6.84 (t, 1 H, Ar-H), 6.64–6.67 (m, 28 H, Ar-H), dryness under reduced pressure and the residue was taken up 6.54–6.56 (m, 14 H, Ar-H), 4.93, 4.96, 5.01 (s, 60 H, Bn-H); dC in a minimum amount of chloroform, then purified by flash 69.8–71.1 (CH2O); 101.6, 106.4, 138.9, 160.1–160.3 (ArC); 127.5, chromatography through a silica column eluting with chloro- 128.0, 128.6, 136.8 (PhC).Mass spectrum (MALDI) m/z, calc.form. The product fractions were collected and evaporated to 3290; found 3313. dryness. The residual solid was taken up in a minimum amount of chloroform and precipitated into methanol, so the product General procedure for the preparation of porphyrin-core could be handled as a powder; yield: 14%; UV–VIS, l/nm: dendrimers via Lindsey synthesis 280 (dendrimer units); 424 (porphyrin Soret); 516, 550, 590, 648 (porphyrin Q-region); dH 8.88 (s, 8 H, b-H), 7.38 (s, br, Ar- A solution of the appropriate dendritic aldehyde 2a–d (1.00 H), 7.19–7.37 (m, Ph-H), 7.02 (s, br, Ar-H), 6.40–6.65 (m, Ar- equiv.), freshly distilled pyrrole (1.00 equiv.), and 1 drop of H), 4.98–4.85 (overlapping resonances, Bn-H), -2.91 (s, 2H, freshly distilled trifluoroacetic acid (TFA) was prepared in dry porph-H); dC 160.03, 159.90, 159.86, 157.76 (ArC-O); 149.68 (a methylene chloride whose volume was calculated so either C); 139.16, 139.08 (ArC-CH2); 136.62 (Ph C-CH2); 132.10 (b starting material had a concentration equal to 10-2 M.The C); 128.40, 127.78, 127.45 (PhC-H); 106.21, 101.42 (ArC-H); reaction was shielded from ambient light and stirred at room 69.80 (Ar/Ph-CH2).Mass spectrum (MALDI-TOF): m/z, calc. temp. for a period of time designated in the following text. 13 350 g mol-1; found 13 376. Then chloranil (8.00 equiv.) was added, and the reaction mixture was warmed to reflux. This was stirred for a period General procedure for the metallation of free-base porphyrinof time designated in the following text.The solution was core dendrimers allowed to cool then evaporated to dryness under reduced pressure. The residue was taken up in minimal chloroform, Introduction of zinc into dendrimers 3a–d was achieved by dissolving the free-base porphyrin-core dendrimer (1.00 equiv.) and insoluble solids (i.e. excess chloranil ) were removed. The 526 J. Mater. Chem., 1998, 8(3), 519–527and Zn(OAc)2 2H2O (1.10 equiv.) in chloroform–methanol saturated NaHCO3, washed once with distilled water, then dried over Na2SO4.The solvent was evaporated, and the (151) (100 ml) and heating the solution at reflux overnight. The solvent was then evaporated, and the residual solids were product was isolated as purple crystals which were dried under vacuum at 40 °C; yield: 100%; UV–VIS l/nm (e): 424 (242 000), partitioned between water and CH2Cl2. The organic layer was separated and dried (Na2SO4).The solvent was evaporated to 516 (14 000), 552 (5000), 592 (4000), 648 (2000); dH ([2H6DMSO, 2.49 ppm) 8.93 (s, 8 H, b-H), 7.05 (s, 8 H, Ar-H), aVord a violet solid. The solid was redissolved in a minimum amount of chloroform then precipitated into methanol, so the 6.64 (s, 4 H, Ar-H), 3.81 (s, br, Ar-OH).dC ([2H6]DMSO, 39.5 ppm) 156.54 (ArC-OH); 144.97 (ArC-porph); 127.50 (b C); compound could be handled as a powder. (Nb the characterization of 1b–d is given earlier in this section.) 117.81 (meso C); 114.15, 102.50 (ArC-H). Mass spectrum (MALDI-TOF): m/z calc. 743; found 753. Zn[G-1]4P 1a. This was prepared from H2[G-1]4P 3a; Financial support of this research by the National Science yield: 100%; UV–VIS l/nm (e) 428 (568 000), 560 (22 000), 600 Foundation (DMR-9641291) and by the MURI program of (7000).dH 8.98 (s, 8 H, b-H), 7.48 (d, 8 H, Ar-H), 7.35 (m, 40 AFOSR is acknowledged with thanks. H, Ph-H), 7.02 (t, 4 H, Ar-H), 5.18 (s, 16 H, Bn-H). dC 157.78 (ArC-O); 149.89 (a C); 144.63 (ArC-porph); 136.75 (PhC-CH2); 131.97 (b C); 128.58, 128.01, 127.63 (Ph C-H); 120.67 (meso References C); 115.06, 96.67 (Ar C-H).Mass spectrum (MALDI-TOF): 1 (a) G. R. Seely, Photochem. Photobiol., 1978, 2, 107; (b) R. Wagner, m/z calc. 1527; found 1538 (Calc. for C100H76N4O8Zn; C 78.65, J. RuYng, B. Breakwell and J. Lindsey, T etrahedron. L ett., 1991, H 5.02, N 3.67; found C 78.60, H 5.05, N 3.54%). 32, (14), 1703; (c) R. Wagner, J. Lindsey, I. Turowska-Tyrk and W. Scheidt, T etrahedron, 1994, 50, 11097. Zinc tetrakis(3,5-dihydroxyphenyl)porphyrin 4. Tetrakis(3,5- 2 J. Collman, J. Brauman, J. Fitzgerald, P. Hampton, Y. Naruta, J. Sparapany and J. Ibers, J. Am. Chem. Soc., 1988, 110, 3477. dihydroxyphenyl)porphyrin 6 (700 mg, 0.942 mmol) and 3 (a) S. Quici, S. Banfi and G. Pozzi, Gazz.Chim. Ital., 1993, 123, 597; Zn(OAc)2 2H2O(228 mg, 1.043 mmol) were dissolved in meth- (b) C. Quintana, R. Assink and J. A. Shelnutt, J. Inorg. Chem., 1989, anol (20 ml ). The solution was heated to reflux for 4 h, then 28, 3421; (c) J. K. M. Sanders, Proc. Ind. Acad. Sci., Chem. Sci., distilled water (60 ml ) was added, and the methanol was 1994, 106 (5), 983; (d) K. M. Faulkner, S.I. Liochev and evaporated under vacuum. The turbid solution was placed in I. Fridovich, J. Biol. Chem., 1994, 269 (38), 23471; (e) T. Katsuki, the refrigerator overnight, and the purple crystalline product Kikan Kagaku Sosetsu, 1993, 19, 67; ( f ) D. R. Benson, R. Valentekovich, S. W. Tam and F. Diederich, Helv. Chim. Acta, was filtered and dried in vacuo at 40 °C; yield: 99%; UV–VIS, 1993, 76 (5), 2034; (g) H.L. Anderson, R. P. Bonar-Law, l/nm (e): 428 (523 000), 560 (20 000), 600 (7000). dH L. G. Mackay, S. Nicholson and J. K. M. Sanders, NATO ASI ([2H6]DMSO, 2.49 ppm) 9.60 (s, 8 H, Ar-OH), 8.86 (s, 8 H, b- Ser., Ser. C, 1992, 371 (Supramol. Chem.), 359; (h) D. Mansuy and H), 7.01 (d, 8 H, Ar-H), 6.62 (t, 4 H, Ar-H). dC ([2H6DMSO, M. Fontecave, Biochem.Biophys. Res. Commun., 1982, 104 (4), 39.5 ppm) 156.21 (ArC-OH); 148.93 (a C); 144.46 (ArC-porph); 1651. 131.33 (b C); 120.23 (meso C); 114.16, 101.72 (ArC-H). Mass 4 (a) F. R. Long, ‘Porphyrin Chemistry Advances’ Ann Arbor Science Publishers Inc., Ann Arbor, 1979a; (b) F. Montanari and spectrum (MALDI-TOF): m/z, calc. 806; found 814. L. Casella, ‘Metalloporphyrins Catalyzed Oxidations’, Kluwer Academic Press, Boston, 1994; (c) R.A. Sheldon, Tetrakis(3,5-dimethoxyphenyl)porphyrin 5. 3,5-dimethoxy- ‘Metalloporphyrins in Catalytic Oxidations’, Dekker, New York, benzaldehyde (1.500 g, 9.026 mmol) and freshly distilled pyr- 1994. role (626 ml, 9.026 mmol) were dissolved in dry chloroform 5 (a) D. A. Tomalia, A. M. Naylor and W. A. Goddard III, Angew.(903 ml) under nitrogen atmosphere. After adding BF3 OEt2 Chem., Int. Ed. Engl., 1990, 29, 138; (b) G. R. Newkome, C. N. Moorefield and G. R. Baker, Aldrichim. Acta, 1992, 25, 31; (c) (364 ml, 2.979 mmol) to the mixture, the solution was shielded B. I. Voit, Acta Polym., 1995, 46, 87; (d) E. M. M. de Brabander- from ambient light and stirred at room temperature for 90 min. van den Berg, A.Nijenhuis, M. Mure, J. Keulen, R. Reintjens, After adding DDQ (1.537 g, 6.770 mmol), the mixture was B. Vandenbooren, B. Bosman, R. de Raat, T. Frijns, S. v.d. Wal, stirred for an additional 90 min, then triethylamine (415 ml, M. Castelijns, J. Put and E. W. Meijer, Macromol. Symp., 1994, 77, 2.979 mmol) was added to neutralize the acid. The solvent was 51; (e) D. A.Tomalia, Adv.Mater., 1994, 7/8, 529. evaporated, and the residual solids were adsorbed onto silica 6 (a) J. M. J. Fre� chet, Y. Jiang, C. J. Hawker and A. E. Philippides, Preprints IUPAC Int. Symp. Functional Polym., Seoul, 1989, p. 19; (20 ml ). The crude product was purified by chromatography (b) C. J. Hawker and J. M. J. Fre� chet, J. Chem. Soc., Chem. through a 40 ml silica column eluting with a linear gradient of Commun., 1990, 1010; (c) J.M. J. Fre� chet, Science, 1994, 263, 1710. solvent from pure hexane to pure CH2Cl2. After collecting the 7 (a) P. J. Dandliker, F. Diederich, M. Gross, C. B. Knobler, product fractions and evaporating the solvent, the product A. Louati and E. M. Sanford, Angew. Chem. Int., Ed. Engl., 1994, was obtained as purple crystals; yield: 52%; UV–VIS, l/nm (e) 33, 1739; (b) J.P. Collman, L. Fu, A. Zingg and F. Diederich, Chem. 424 (280 000), 516 (19 000), 548 (9000), 586 (9000), 648 (5000). Commun., 1997, 193. 8 (a) R. Jin, T. Aida and S. Inoue, J. Chem. Soc., Chem. Commun., dH ([2H6DMSO, 2.49 ppm) 8.95 (s, 8 H, b-H), 7.01 (s, 8 H, Ar- 1993, 1260; (b) D. Jiang, R. Jin and T. Aida, Chem. Commun., 1996, H), 6.65 (s, 4 H, Ar-H), 3.33 (s, 24 H, ArO-CH3), -3.07 (s, 2 1523; (c) Y. Tomoyose, D. Jiang, R. Jin, T. Aida, T. Yamashita, H, N-H); dC ([2H6DMSO, 39.5 ppm) 156.60 (ArC-OCH3); K. Horie, E. Yashima and Y. Okamoto, Macromolecules, 1996, 142.89 (ArC-porph); 131.12 (b C); 119.94 (meso C); 114.19, 29, 5236. 102.28 (ArC-H); 48.64 (Ar-OCH3) (Calc. for C52H46N4O8; C 9 K. W. Pollak, J. W. Leon and J. M. J. Fre� chet, Polym. Mater. Sci. 73.05, H 5.42, N 6.55; found C 73.11, H 5.29, N 6.43%). Eng., 1995, 73, 137; K. W. Pollak, J. W. Leon and J. M. J. Fre� chet, Chem. Mater., in the press. 10 P. Bhyrappa, J. K. Young, J. S. Moore and K. S. Suslick, J. Am. Tetrakis(3,5-dihydroxyphenyl)porphyrin 6. Tetrakis(3,5- Chem. Soc., 1996, 118, 5708. dimethoxyphenyl)porphyrin 5 (740 mg, 0.866 mmol) was dis- 11 C. J. Hawker and J. M. J. Fre� chet, J. Am. Chem. Soc., 1990, 112, solved in dry CH2Cl2 (20 ml ) under nitrogen atmosphere. The 7638. solution was cooled to 0 °C, then BBr3 (7.27 ml, 7.271 mmol, 12 J. Lindsey, I. Scheirman, H. Hsu, P. Kearney and A. Marguerettaz, J. Org. Chem., 1987, 52, 827. 1 M in CH2Cl2) was slowly added to the reaction. After 13 R. J. Abraham, G. E. Hawkes, M. F. Hudson and K. M. Smith, addition, the mixture was allowed to warm to room temp. as J. Chem. Soc., Perkin T rans. 2, 1975, 204. it stirred overnight. Enough methanol was then added to 14 (a) D. Dolphin, ‘T he Porphyrins vol. III’, Academic Press, NY 1978, deactivate any unreacted BBr3, distilled water (30 ml ) was pp. 12–16; (b) J. E. Falk, ‘Porphyrins and Metalloporphyrins’, added, and the mixture was stirred for 2 h. After evaporation Elsevier Publishing Company, NY, 1964, vol. 75–76, pp. 243–246. of the organic solvents, a green powder was filtered from the water. It was dissolved in diethyl ether, washed twice with Paper 7/05410F; Received 28th July, 1997 J. Mater. Chem., 1998, 8(3), 519&n

ISSN:0959-9428    

DOI:10.1039/a705410f

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Tailoring thermotropic cubic mesophases: amphiphilic polyhydroxy derivatives† Konstanze Borisch,a Siegmar Diele,b Petra Go�ring,b Horst Kresseb and Carsten Tschierske*a aInstitut fu� r Organische Chemie, Martin-L uther-Universita� t Halle-W ittenberg, Kurt-Mothes Str. 2, D-06120 Halle, Germany bInstitut fu� r Physikalische Chemie, Martin-L uther-Universita� t Halle-W ittenberg, Mu�hlpforte 1, D-06108 Halle, Germany Novel amphiphilic polyhydroxy compounds [N-(3,4-dialkoxybenzoyl)-1-amino-1-deoxy-D-glucitols (glucamides), N-(3,4- dialkoxybenzoyl)-1-deoxy-1-methylamino-D-glucitols, N-(3,4,5-trialkoxybenzoyl)-1-deoxy-1-methylamino-D-glucitols (Nmethylgucamides), 1-benzoylaminopropane-2,3-diols, 2-benzoylaminopropane-1,3-diols, 2-(3,4,5-tridodecyloxybenzoylamino)-2- (hydroxymethyl)propane-1,3-diol and (3,4,5-tridodecyloxybenzoyl)bis(2,3-dihydroxypropyl)amine] have been synthesized. Their thermotropic liquid crystalline phases were investigated by means of polarizing microscopy, diVerential scanning calorimetry and X-ray diVraction.Depending on the chain length, and the size of the hydrophilic polyhydroxy units, diVerent mesophases have been found: smectic A phases (SA), inverted bicontinuous cubic phases (CubV2, Ia3d), hexagonal columnar phases (ColH2) and micellar cubic mesophases (CubI2, Pn3m or P43n).In strong analogy to lyotropic systems, the type of thermotropic mesophase depends on the ratio between the volume of the lipophilic moiety and the surface area of the hydrophilic moiety at the hydrophilic–lipophilic interface.The crossing from zero interface curvature (SA phase) to the finite negative curvature of the inverted cylindrical aggregates of the columnar mesophase takes place via bicontinuous cubic mesophases. The cylindrical aggregates of the columnar mesophase are stable over a rather broad range of variation of the structural parameter.At a certain degree of the size of the lipophilic moiety in respect to the surface area of the hydrophilic group, however, the transition from the hexagonal columnar to a micellar cubic mesophase takes place. On the basis of proton conductivity measurements and from packing considerations we propose that this cubic lattice is built up by eight closed micelles per unit cell which have a rod-like shape and represent small segments of extended columns.Therefrom we can propose a model for the transformations between these diVerent thermotropic mesophases. Cubic mesophases represent ordered supermolecular arrange- toward the regions with stronger cohesive forces). From a ments which are optically isotropic. They are common in crystallographic point of view one can distinguish between surfactant–solvent and lipid–solvent systems1 and have diVerent cubic lattices: primitive (Pn3m/Q224, Pm3n/Q223), body attracted considerable interest due to their application in the centred (Ia3d/Q230, Im3m/Q229) and face centred (Fm3m/Q225, pharmaceutical industry, their potential use in drug release Fd3m/Q227).The diVerent types of cubic lattices found up to systems and as templates for the preparation of mesoporous now in the diVerent regions of the lyotropic phase sequence silicates and also because of their biological significance.2Well- are collected in the principal phase diagram given in Fig. 1.1,8–10 defined cubic structures can also be found in amphiphilic block It has to be mentioned that additional intermediate mesophases copolymers3 and polyelectrolytes.4 (ribbon-phases, rhombohedral, tetragonal and ortorhombic As shown in Fig. 1 several diVerent cubic phases can be phases11,12 and isotropic hexagonal phases13) can be found observed in lyotropic systems when the concentration of the between the lamellar phase and the normal and inverted surfactant is changed. They can either occur as intermediate columnar phases.phases between lamellar and hexagonal columnar phases In contrast to lyotropic cubic systems relatively few thermo- (bicontinuous cubic phases, V-phases), or between the hexag- tropic compounds with cubic phases are known. They have onal columnar phases and the micellar solutions (discontinuous been found for nitro- and cyano-substituted biphenyl carcubic phases, I-phases).5,6 The first type can be regarded as boxylic acids,14,15 dibenzoylhydrazides,16 strontium soaps,17 interwoven networks of branched columns, the second one has polycatenar compounds,18–21 discotic molecules,22,23 silver a diVerent microstructure.It is assumed to consist of a cubic complexes,24,25 a spherical compound,26 a few calamitic comarrangement of closed spherical or nonspherical micelles.The pounds27,28 and polypeptides.29 In most cases, bicontinuous value of the interface curvature between hydrophilic regions structures with a body centred lattice (Ia3d, Im3m) have been and lipophilic regions was recognized as the key factor found for these cubic phases.18,15 More recently diVerent types determining the morphology of the polymolecular aggregates of thermotropic cubic phases have been observed for amphiforming the mesophase ( layers, columns, closed micelles),7 philic oligoethylene imines,30 diols31,32 and carbohydrate whereas the sign of the interface curvature distinguishes derivatives.33–36 In the case of double chain carbohydrates between normal (type 1, the interface curvature is directed body centred cubic phases (Ia3d) were usually found whereas away from the regions with stronger cohesive interactions) and cubic phases with primitive cubic lattices (Pm3n or P43n) have inverted phase types (type 2, the interface curvature is directed only recently been detected for the thermotropic phases of some triple chain amphiphilic carbohydrates36 and diol compounds32 and also for some dendrimers.37 From packing considerations it was concluded that the cubic phases of † Presented at the Third International Conference on Materials these compounds should represent inverted micellar cubic Chemistry, MC3, University of Exeter, Exeter, 21–25 July, 1997.* E-mail: coqfx@mlucom.urz.uni-halle.de mesophases. J. Mater. Chem., 1998, 8(3), 529–543 529N-methylglucamides, 2,3-dihydroxypropylamides and some related acylated amino alcohols with diVerent numbers of alkyl chains and hydroxy groups.39,40 The results of the investigation of these compounds are summarized in this report.Results and Discussion Syntheses The homologous N-benzoyl-1-amino-1-deoxy-D-glucitol derivatives 1, N-benzoyl-1-deoxy-1-methylamino-D-glucitols 2 and 3 were synthesized according to the procedure given for the hexyl and dodecyl derivatives.36 The other materials were synthesized by aminolysis of substituted benzoyl chlorides with an excess of 1-aminopropane-2,3-diol (compounds 5–7) (Scheme 1), 2-aminopropane-1,3-diol (compounds 8 and 9) and 2-amino-2-(hydroxymethyl)propane-1,3-diol (compound 10) respectively. The tetraol 11 was obtained by aminolysis of 3,4,5-tridodecyloxybenzoyl chloride with diallylamine, followed by OsO4-catalyzed dihydroxylation of the olefinic double bonds.The final compounds were purified by column chromatography and repeated crystallization. Details are given in the Experimental part. Amphiphilic carbohydrate derivatives Optically isotropic mesophases of double chain carbohydrates —bicontinuous cubic phases.At first we investigated the homologues series of the double chain glucamides 1 and Fig. 1 Schematic representation of the major lyotropic liquid crystal- the analogous N-methylglucamides 2 (Tables 1 and 2). In both line phase types occurring in detergen–solvent and lipid–solvent series of compounds optically isotropic mesophases were found systems depending on the solvent concentration (refs. 1,8); abbrevifor one or two compounds with a medium chain length. ations: La=lamellar a-phase, V=bicontinuous cubic mesophase, H= In the case of the glucamides 1 the pentyloxy derivative 1/5 hexagonal columnar phase, I=discontinuous (micellar) cubic phase; and the hexyloxy derivative 1/6 have optically isotropic meso- subscripts: 1=normal phase (type 1, the interface curvature is directed away from the regions r cohesive interactions), 2= phases.The isotropic mesophase of 1/6 has recently been inverted phases (type 2, the interface curvature is directed toward the investigated and it was found that this phase is a cubic regions with stronger cohesive forces). The positions of the known mesophase with a body centred lattice (Ia3d; acub=8.31 nm types of lyotropic cubic phases in the ideal phase diagram are shown at 175 °C).36 The lower homologues (compounds 1/4 and 1/5) together with their space-groups and space-group numbers.Curiously, display smectic A phases (beside an isotropic mesophase in some inverse systems exhibit phase sequences which are not in accord the case of 1/5) and the higher homologue 1/7 with two with this figure.It is unknown why in some systems the sequences H2–V2 or V2–La occur on increasing the surfactant concentration (ref. 1). The ‘rod-like’ versions of the bicontinuous cubic phases are shown. They can also be regarded as bilayers draped on the underlying minimal surfaces (refs. 1,41). Additional to the shown cubic phases a chiral cubic phase with the space group P4332 (Q212) was found in a ternary lipid–protein–water system (ref. 9). Recently, a novel variant of the cubic phases with the space group Im3 has been found in a ternary lyotropic system between a rectangular and a lamellar phase (ref. 10). It seems that the mesophases of the pure carbohydrate- and polyhydroxy-amphiphiles represent thermotropic analogues of the diVerent lyotropic mesophases occurring in detergent– solvent systems.In order to confirm this analogy, we have initiated a systematic study of the influence of slight variations of the lipophilic and the polar regions of amphiphilic polyhydroxy compounds on the appearance of the diVerent types of thermotropic cubic mesophases.‡ For this purpose we have synthesized novel double chain and triple chain glucamides, Scheme 1 Reagents and condition: i, RBr, K2CO3, cyclohexanone; ii, ‡ In a related study the influence of the alkyl chain size on the lyotropic mesophase structure of 1-alkanoyl-1-deoxy-1-methylamino- KOH, H2O, EtOH; iii, H2O, H+; iv, SOCl2; v, DMF, DMAP, 2,3- dihydroxypropylamine D-glucitols has been investigated (ref. 38). 530 J. Mater. Chem., 1998, 8(3), 529–543Table 1 Transition temperatures and associated enthalpy values (lower lines; in italics) of the N-(3,4-dialkoxybenzoyl)-1-amino-1-deoxy-Dglucitols 1a T /°C compound R DH/kJ mol-1 1/4 C4H9 K 174 (SA 156) Iso 66.3 0.7 1/5 C5H11 K 172 (M 162 SA 168) Iso 68.7 0.5 1/6b,c C6H13 Fig. 2 Optical photomicrograph (crossed polarizers) of the isotropic K 172 CubV2 185 (ColH2 185) Iso 73.5 0.7 mesophase of compound 1/6 growing from the hexagonal columnar 1/7 C7H15 mesophase at 184 °C K 169 ColH2 203 Iso 72.9 1.1 1/9 C9H19 K 166 ColH2 239 Iso 94.8 2.6 1/12b C12H25 head groups.They are connected three-by-three and separated K 162 ColH2 254 Iso 91.2 0.9 by a gyroid minimal surface41 within the lipophilic continuum of the alkyl chains (see Fig. 1). aAbbreviations: K=crystalline solid, SA=smectic A phase, ColH2= More detailed investigation of the cubic to liquid phase inverted hexagonal columnar mesophase, CubV2=inverted bicontinutransition has been done.On heating, a direct transition from ous cubic mesophase, M=optically isotropic (probably cubic) mesothe cubic phase to the liquid state occurs at 185 °C. On fast phase of unknown structure, Iso=isotropic liquid; values in parentheses refer to metastable (monotropic) mesophases.In the cooling from the liquid however, a spherulitic texture occurs compound numbering X/Y, X represents the compound type, and Y at the same temperature. This texture is typical for hexagonal the length of the alkyl chain. bRef. 36. cThe texture of a columnar columnar phases and is identical with the optical texture of mesophase is only detected on cooling (see text).the hexagonal columnar mesophases of the glucamides 1/7–1/12. Immediately after this phase transition took place Table 2 Transition temperatures, associated enthalpy values (lower the cubic phase appear as highly viscous isotropic domains lines) and lattice parameter of the smectic phase (dlam) and of the with cornered boundaries (see Fig. 2). These domains grow hexagonal columnar mesophases (ahex) at the measurement temperarapidly and coalesce to a homogeneous area. On re-heating a tures (lower lines in brackets) of the N-(3,4-dialkoxybenzoyl)-1-deoxydirect transition from the highly viscous cubic phase to the 1-methylamino-D-glucitols 2a fluid liquid was found at 185 °C. The spherulitic texture cannot be found on very slow cooling (<2 K min-1).Obviously the formation of the cubic phase can be supercooled and therefore a columnar phase is observed first as a metastable (monotropic) phase. Interestingly the clearing temperature of the columnar mesophase and the transition temperature of the cubic phase to the liquid state seem to be identical (DT<1 K). T /°C dlam/nm ahex/nm The lower homologue 1/5 displays an enantiotropic smectic compound R DH/kJ mol-1 (T /°C) (T /°C) A phase which turns into an isotropic mesophase on cooling below 162 °C.This phase transition can also be supercooled 2/6b C6H13 3.19 — K 93 SA 130 Iso 29.5 0.7 (down to 158 °C); however this SA phase is a thermo- (120) dynamically stable (enantiotropic) phase in the temperature 2/7c C7H15 — — K 96 M 118 (SA 118) Iso 27.5 0.4 range between 162 and 168 °C.Remarkably, the appearance 2/8 C8H17 — 4.27 of the isotropic domains is diVerent from that of the CubV2 K 87 ColH2 155 Iso 18.5 0.1 (80) phase of compound 1/6. They appear as circular domains (the 2/9 C9H19 — 4.46 same texture as shown in Fig. 3). Due to rapid crystallization K 68 ColH2 174 Iso 18.0 0.2 (80) of this compound no X-ray investigation of its isotropic 2/12b C12H25 — 4.88 mesophase can be carried out to confirm a cubic structure.K 54 ColH2 194 Iso 22.5 0.9 (120) No transition to an optically isotropic mesophase was found 2/16 C16H33 — — K 82 ColH2 194 Iso 45.7 1.2 on cooling the SA phase of the dibutoxy derivative 1/4 down to 146 °C. At this temperature rapid crystallization sets in.Also the N-methylglucamide 2/7 has an optically isotropic aAbbreviations as in Table 1. bRef. 36. cThe fan-like texture of a SA phase is only detected on cooling (see text). The value of the transitions mesophase. This mesophase again occurs intermediate between to the isotropic liquid was found by microscopic investigations. In the a smectic A phase (compound 2/6) and a hexagonal columnar DSC heating cycle an endotherm was detected at 125 °C.phase (compound 2/8). Comparing the homologous series of the glucamides 1 with the N-methylglucamides 2 indicates that in the series of the N-methylglucamides 2 the isotropic meso- heptyloxy chains has a hexagonal columnar mesophase. Thus, with respect to the chain length, the cubic phase occurs as an phase occurs at a longer chain length.Therefore, one can conclude that the N-methylglucamide group is slightly larger intermediate phase between the SA phase and the inverted hexagonal columnar phase. In analogy to lyotropic systems it than the glucamide group. However, it has to be considered that the transition temperatures of the glucamides are signifi- can therefore be concluded that this cubic phase should be an inverted bicontinuous cubic phase (V2-phase).It should consist cantly higher than those of the corresponding N-methylglucamides (due to the additional hydrogen bonding of the NMH of two interwoven yet non-connected networks of cylinders containing the hydrogen bonding networks of the carbohydrate group) and therefore their alkyl chains have an increased J.Mater. Chem., 1998, 8(3), 529–543 531phase type also changes. In analogy to lyotropic systems (Fig. 1), an inverted hexagonal columnar phase (ColH2) can be found between the isotropic mesophase of compound 2/7 and the inverted micellar cubic phase (CubI2) of compound 3/12. Thus, the same type of binary phase diagram as for compound 1/6 (forming a CubV2 phase) with 3/12 was found.36 Therefore, we conclude that this isotropic mesophase should represent an inverted bicontinuous network structure (CubV2).However, due to the low number and intensity of scatterings in the SAXRS we were not yet able to determine the type of cubic lattice. Micellar cubic mesophases of triple chain carbohydrate amphiphiles. Now let us turn to triple chain compounds, which have an increased size of the lipophilic moiety and should form strongly curved aggregates.The phase transition temperatures Fig. 3 Optical photomicrograph (crossed polarizers) of the optically of the triple chain N-methylglucamides 3 are summarized isotropic mesophase of compound 2/7 growing from the SA phase in Table 3. at 90 °C In contrast to the analogous double chain compounds 2, in this homologous series the compounds with a short and mobility which gives rise to a larger eVective volume of their medium chain length (3/6 and 3/8) have a hexagonal columnar lipophilic region.mesophase and an optically isotropic mesophase occurs on Also for the N-methylglucamide 2/7 a direct transition from elongation of the alkyl chains, i.e. on increasing the interface the isotropic mesophase to the liquid state can only be found curvature between polar and lipophilic regions.Therefore it in the heating cycles. On cooling from the liquid state a fan- can be concluded that this phase represents a cubic phase like texture with homeotropic regions and oily streaks can be consisting of a cubic arrangement of closed inverted micelles.observed and not a spherulitic texture as found in the case of The nonyloxy derivative 3/9 and the decyloxy derivative the glucamide 1/6. These textural features are typical for a 3/10 display a columnar–cubic dimorphism. In contrast to the smectic A phase. Immediately after the appearance of the SA double chain carbohydrates, the triple chain glucamides have phase small circular and optically isotropic domains occur (see the cubic phases as high temperature mesophases above the Fig. 3), which rapidly grow and coalesce with formation of a hexagonal columnar phase. The X-ray pattern of these cubic uniform optically isotropic area. The viscosity of these circular phases can be indexed on the basis of primitive cubic lattices domains is significantly larger than that of the surrounding (e.g.acub=7.93 nm at 121 °C for compound 3/9). The space SA phase. Thus, it can easily be distinguished from the homeo- group can be either Pm3n or P43n. Interestingly, the cubic tropically aligned regions of the SA phase. On re-heating only lattice parameter amounts to approximately twice the hexaga direct transition to the fluid isotropic phase was found at onal lattice parameter of the hexagonal columnar phase below 118 °C and no SA phase occurs.Obviously, the formation of (e.g. ahex=4.08 nm at 90 °C for 3/9). The same relationship the isotropic mesophase is again kinetically hindered. This between the hexagonal and cubic lattice parameter is found kinetic eVect points to a three dimensional structure of this for other triple chain amphiphiles and will be discussed below.mesophase. This and the high viscosity of the isotropic mesophase N-(3,5-Didodecyloxybenzoyl)-1-amino-1-deoxy-D-glucitol represent typical features of cubic mesophases. 4/12. The 3,5-disubstituted glucamide 4/12 shows exclusively The inverted bicontinuous structure of the isotropic mesoa hexagonal columnar mesophase. phase of compound 1/6 can be confirmed by miscibility experiments.The binary phase diagram of the system 2/7–3/12 is shown in Fig. 4. Because the structural parameter continuously changes in the contact region between these two compounds, the meso- If one compares this compound with the related 3,4-disubstituted double chain compound 1/12 and the triple chain carbohydrates 3/12, a significantly lower mesophase stability is found for 4/12.The absence of a cubic mesophase leads to the assumption that this 3,5-disubstitution pattern gives rise to behaviour more similar to the 3,4-disubstituted compound than to the triple chain compound 3/12. The columnar mesophases of the double and triple chain glucamides. The hexagonal columnar phases of the amphiphilic polyhydroxy compounds consist of extended aggregates of hydrogen bonding networks surrounded by the molten alkyl chains.Comparison of the hexagonal lattice parameter of the double chain compound 2/8 (ahex=4.27 nm at T=80 °C) and the triple chain compound 3/8 (ahex=3.83 nm at T=80 °C) Fig. 4 Binary phase diagram of the system 2/7–3/12 which have both the same chain length and the same molecular 532 J.Mater. Chem., 1998, 8(3), 529–543Table 3 Transition temperatures, associated enthalpy values (lower lines) and lattice parameter of the hexagonal columnar mesophases (ahex) and the cubic mesophases (acub) at the measurement temperatures (lower lines in brackets) of the N-(3,4,5-trialkoxybenzoyl)-1-deoxyl-1-methylamino- D-glucitols 3a T /°C ahex/nm acub/nm compound R DH/kJ mol-1 (T /°C) (T /°C) 3/6b C6H13 3.49 — K 94 ColH2 145 Iso 40.2 1.3 (100) 3/8 C8H17 3.83 — K 75 ColH2 147 Iso 48.8 0.7 (80) 3/9 C9H19 4.08 7.93 K 50 ColH2 102 CubI2 143 Iso 62.1 1.1 0.8 (52) (121) 3/10 C10H21 — — K 59 ColH2 89c CubI2 158 Iso 55.2 0.3 0.6 3/12b C12H25 — 8.55 K 75 CubI2 185 Iso 112.3 1.2 (90) aAbbreviations: CubI2=inverted micellar cubic mesophase, other abbreviations as in Table 1.bRef. 36. cThis phase transition refers to the heating cycle. On cooling the cubic–columnar transition occurs is at 56 °C. length (L=2.55 nm as estimated from CPK models) indicates that the (average) diameter of the columns of the double chain compound is significantly larger then the diameter of the columns of the triple chain compound. Further examples are collected in Table 4. Assuming a density of r=1 g cm-3 the number n of molecules arranged side by side in a single slice of the columns with a thickness (h) of 0.45 nm was estimated according to eq.(1). n=(a2/2)Ó3h(NA/M)r (1) The parameter a is the hexagonal lattice parameter, NA the Avogadro constant and M the molecular mass. The estimated Fig. 5 Cross section of the columnar mesophases of (a) the triple chain values of n amount to about 5 for the triple chain compounds amphiphiles and (b) the double chaing amphiphiles 3/6–3/8.This value corresponds to those values found in columnar phases of other tapered polyhydroxy of the columns and/or that they are arranged randomly compounds.42–45 oriented in the hexagonal lattice, the cross sections of these Obviously, the three alkoxy chains can eVectively surround aggregates should be circular on average, giving rise to the the polar regions of the hydrogen bonding networks and a typical scattering pattern of hexagonal columnar phases.46 nearly circular shape of the columns can be assumed.In contrast, in the columnar mesophases of all investigated Amphiphilic diols, triols and tetraols double chain compounds (1/8–1/12) a much larger number, about eight molecules per slice, was observed.This number is Single chain, double chain and triple chain 1-benzamidoprolarger than found for other wedge shaped polyhydroxy amphi- pane-2,3-diol derivatives. In a next step we decided to investigate philes. Probably the columns of these double chain compounds the influence of the structure of the hydrophilic parts of the are non-circular (see Fig. 5). Taking into account the fact that amphiphilic molecules on their mesophase behaviour. At first the orientation of the deformations can change along the axis we synthesized and investigated single chain, double chain and triple chain 1-benzamidopropane-2,3-diol derivatives (compounds 5–7 in Table 5). These compounds are structurally Table 4 Comparison of the hexagonal lattice parameter (ahex), the molecular length (L ) (as determined from CPK models assuming an related to the glucamides 1–3, but they have only two instead all trans-conformation of the chains) and the number of molecules (n) of five hydroxy groups.In this way the number of attractive arranged in each slice of the columnar mesophases (with a high of hydrogen bonds is decreased and consequently the phase 0.45 nm) of the double chain and triple chain amphiphiles with the transition temperatures are shifted to lower values.same length of the alkyl chains As in the case of related D-glucamides,35 three diVerent types triple chain double chain of mesophases were found for the 1-benzamidopropane-2,3- amphiphile amphiphile diol derivatives depending on the number of alkyl chains.Compound 5/12 with only one dodecyloxy chain forms a ahex/nm ahex/nm smectic A phase, a hexagonal columnar phase (ColH2) is found R L /nm Comp. (T /°C) n Comp. (T /°C) n for the double chain compound 6/12 and a cubic phase was detected for compound 7/12 with three dodecyloxy chains C8H17O 2.55 3/8 3.83 5.0 2/8 4.27 7.7 (80) (80) (Table 5).C9H19O 2.7 3/9 4.08 5.4 2/9 4.46 8.0 The inverted micellar structure of the cubic phase of com- (52) (80) pound 7/12 was at first proven by means of miscibility C6H13O 1.9 7/6 3.13 4.7 6/6 3.48 7.2 experiments. The optical microscopic picture of the contact (75) (84) region between the smectic phase of 5/12 and the cubic phase J. Mater. Chem., 1998, 8(3), 529–543 533Table 5 Transition temperatures, associated enthalpy values (lower lines) and lattice parameter of the smectic phase (d), of the hexagonal columnar mesophases (ahex) and the cubic mesophases (acub) at the measurement temperatures (lower lines in brackets) of the 1-benzoylaminopropane- 2,3-diols 5–7a T /°C d/nm ahex/nm acub/nm compound R1 R2 R3 DH/kJ mol-1 (T/°C) (T/°C) (T/°C) 5/12 OC12H25 H H 4.03 — — K1 80 K2 89 SA 132 Iso 11.6 36.6 0.8 (85) 6/6 OC6H13 OC6H13 H — 3.48 — K 79 ColH2 87 Iso 22.0 0.3 (84) 6/12 OC12H25 OC12H25 H — 4.2 — K 98 ColH2 148 Iso 60.4 1.4 (105) 7/6 OC6H13 OC6H13 OC6H13 — 3.13 — K 49 ColH2 91 Iso 23.6 1.4 (75) 7/7 OC7H15 OC7H15 OC7H15 — 3.16 — K 46 ColH2 92 Iso 25.7 1.5 (45) 7/8 OC8H17 OC8H17 OC8H17 — 3.33 6.89 K 59 ColH2 74b CubI2 85c Iso 38.6 0.6 0.4 (50) (75) 7/9 OC9H19 OC9H19 OC9H19 — — 7.06 K 49 CubI2 104d Iso 24.4 0.3 (65) 7/12 OC12H25 OC12H25 OC12H25 — — 7.94 K1 45 K2 69 CubI2 126 Iso 36.0 11.5 0.7 (45) aAbbreviations as in Tables 1–3.bThis phase transition refers to the cooling cycle. On heating the columnar–cubic phase transition is observed at 79 °C due to the ‘overheating’ of this phase transition. cOn cooling the cubic phase appears at 77 °C.dThis value was found by microscopic investigations. In the DSC–heating cycle an endotherm was detected at 107 °C. of compound 7/12 between crossed polarizers is shown in Fig. 6. In the contact region between the cubic phase of the triple chain diol 7/12 and the smectic phase of the single chain diol 5/12 a hexagonal columnar phase is induced (large spherulites in the centre of Fig. 6). The detailed phase diagram of the system 5/12–7/12, obtained by investigation of binary mixtures, is shown in Fig. 7. The induction of a columnar mesophase in the contact region of these two compounds is an important piece of evidence for the proposed inverse micellar structure of the cubic mesophase of the triple chain diol 7/12.The X-ray diVraction pattern of this cubic phase can be indexed on the basis of a primitive cubic lattice (Pm3n or P43n) with a lattice parameter acub=7.45 nm at T=90 °C. The dependence of the mesophase type on the chain length (see Table 5) was investigated for the triple chain diols 7. The Fig. 7 Binary phase diagram of the system 5/12–7/12 short chain diols 7/6 and 7/7 have columnar mesophases; a columnar/cubic dimorphism was found for the octyloxy derivative 7/8. Fig. 8 shows the transition from the hexagonal columnar phase to the cubic mesophase as it can be seen on heating compound 7/8 between crossed polarizers. The higher homologues 7/9 and 7/12 have exclusively cubic mesophases. The cubic phases of 7/8 and 7/12 have the same kind of primitive cubic lattice (Pm3n or P43n).As an example the diVraction pattern of the cubic phase of compound 7/12 is shown in Fig. 9. The DSC heating and cooling traces of compound 7/8 are Fig. 6 Optical photomicrograph (crossed polarizers) of the ColH2 shown in Fig. 10. It indicates that the crystalline state can be phase developing in the contact region between the micellar cubic strongly supercooled.Even after prolonged storage only partial CubI2 phase of compound 7/12 (optical isotropic region at the righthand side) and the SA phase of compound 5/12 ( left-hand side) at 95 °C crystallization takes place. Therefore any heating cycle after 534 J. Mater. Chem., 1998, 8(3), 529–543Fig. 9 X-Ray pattern of the micellar cubic mesophase of compound 7/12 at 70 °C Fig. 10 DSC heating and cooling trace of 7/8 (10 K min-1) dimensional structure (cubic phase) to the columnar mesophase occur without supercooling eVects. Therefore, the values for this transition were taken from the cooling cycles. All other reported values correspond to the heating cycles. There is another interesting point concerning the transition from the optically isotropic mesophases to the fluid liquid state.This transition can be detected microscopically by the sudden increase of the fluidity. In many cases however (e.g. compound 7/9) the corresponding endotherm in the DSC heating curve is detected at higher temperatures then observed by microscopy. The reason for this behaviour is not clear. However, related phenomena were reported for several other compounds with transitions from three-dimensional ordered mesophases to isotropic liquids47 (cubic phases of long chain Fig. 8 Optical photomicrographs (crossed polarizers) of the CubI2 phase developing on heating the ColH2 phase of compound 7/8.nitro-biphenyl carboxylic acids, TGB-phases and blue phases (a) ColH2 phase at 74 °C; CubI2 phase (black regions) growing into the of chiral liquid crystals).ColH2 phase at (b) 78 and (c) 79°C (the reduced sharpness of this Comparing the homologous series of the triple diols 7 with picture is due to the rapid growing of the cubic domains during the that of the corresponding triple chain N-methylglucamides 3 time of exposition). reveals that the same phase sequence is found in both series of compounds; however the transition from the columnar to the cubic phase occurs in the series of compounds 7 which the first one gives significantly lower melting enthalpies and sometimes also diVerent melting temperatures.All melting already have a chain length of eight instead of nine or ten carbon atoms. This means, that the eVective size of the 1- temperatures and enthalpies given in Tables 1–3, 4, 6 and 8 refer to the first heating cycles.amidopropane-2,3-diol unit at the hydrophilic–lipophilic interface should be slightly smaller than that of the 1-methyl- Due to the three dimensional structure of the cubic phases the formation of these isotropic phases can be kinetically amidoglucitol group. This is also evident from the comparison of the double chain compounds 2/6 and 6/6.The N-methylglu- hindered. On cooling, the transition from the liquid to the cubic phase is often supercooled and on heating the transition camide 2/6 is a smectic liquid crystal whereas the diol 6/6 forms a columnar mesophase. from the columnar to the cubic phase can be overheated. Therefore diVerent values can result for the heating and cooling cycles depending on the heating–cooling rates (see for example 2-Benzamidopropane-1,3-diols.In the next step, two selected examples of 2-benzamidopropane-1,3-diols (compounds 8/6 Fig. 10). If the cubic phase represents the high temperature mesophase, only on cooling does the transition from the three- and 9/12, Table 6) have been synthesized. These compounds J. Mater. Chem., 1998, 8(3), 529–543 535Table 6 Transition temperatures and associated enthalpy values (lower lines) of the 2-benzoylaminopropane-1,3-diol 8/6 and 9/12a T /°C compound R1 R2 R3 DH/kJ mol-1 8/6 OC6H13 OC6H13 H K164 K2 108 [CubV2 50 (ColH2 50)] Iso 4.5 31.6 9/12 OC12H25 OC12H25 OC12H25 K 67 (ColH2 62b) CubI2 104c Iso 39.3 2.1 0.4 aAbbreviations as in Tables 1–3.bThis phase transition refers to the cooling cycle.On heating the columnar–cubic phase transition is observed at 72 °C due to the ‘overheating’ of this phase transition. cFound by microscopic investigations. In DSC heating cycles the phase transition occurs at 108 °C. diVer from the 1-benzamidopropane-2,3-diols 6/6 and 7/12 only in the position of their hydroxy groups to each other. Interestingly, the triple chain 1,3-diol 9/12 displays a columnar/cubic dimorphism, whereas the corresponding 1,2- diol 7/12 with the same number and length of the chains forms a micellar cubic phase exclusively.This means that the 1,3- diol unit represents a significantly larger hydrophilic group than the corresponding 1,2-diol unit. This is in accordance with the diVerent molecular areas found in the densely packed monomolecular layers of 1,2-diols (0.21 nm2 per molecule)48 and 1,3-diols (0.24 nm2 per molecule)49 at the air–water interface.The double chain compound 8/6 has only monotropic liquid crystalline properties. On rapid cooling from the isotropic melt, at first the typical texture of a hexagonal columnar phase appears, which at 50 °C immediately turns into an isotropic phase. On heating, a direct transition into the isotropic liquid takes place and no columnar phase can be detected.This behaviour is analogous to that found for the Fig. 11 Plot of the specific conductivity k of (&) 7/8 and (+) 9/12 double chain glucamide 1/6 and therefore it is very likely that versus the reciprocal absolute temperature 1/T the optically isotropic mesophase of this compound represents a bicontinuous cubic phase (CubV2 phase).Due to the monotropic character and the rapid crystallization, no X-ray investi- imental finding strongly supports the inverted micellar structure of the cubic phases of the triple chain amphiphiles under gations were possible to confirm the proposed cubic structure of this phase. investigation. The next question concerns the shape of these closed micelles.Because the lipophilic residues of the double chain 1,3-diol 8/6 and the related 1,2-diol 6/6 are identical, the occurrence Spherical, oblate and prolate shapes are under discussion for micellar cubic phases of lyotropic systems. If one assumes the of a bicontinuous cubic phase instead of a columnar phase again can be explained by the larger size of the 1,3-diol unit existence of spherical micelles, than one micelle is located in each corner of a simple cubic unit cells (see Fig. 12). Therefore, of 8/6 in comparison to the 1,2-diol unit of compound 6/6. the diameter of these micelles should be equal to the cubic lattice parameter. In this case the diameter of the micelles Investigation of the micellar cubic phases of the triple chain amphiphiles.In order to prove the inverted micellar structure would be significantly larger than twice the length of the single molecules (see Table 7). This contradiction between molecular of the cubic phases of the triple chain amphiphiles, proton conductivity investigations have been carried out. If the cubic size and cell dimensions is often observed in cubic phases. Therefore, Fontell et al.50 suggested an alternative structure phases under discussion are built up by closed inverted micelles, then the hydrogen bonding networks between the head-groups for lyotropic micellar cubic phases located between the micellar solution and the normal hexagonal phase (CubI1) of lyotropic should be located inside the micelles and therefore they should be isolated from each other by the lipophilic regions of the systems.They proposed that the cubic unit cell is built up by eight short rod-like aggregates with an axial ratio of less than alkyl chains. In a bicontinuous cubic phase and in the columnar phase however, the hydrogen bonding network should be 251. One of these rod-shaped micelles is placed in each corner of the unit cell, one in the centre, and two at each surface of extended over large distances.Since these hydrogen bonding networks can act as proton conductors, the conductivity of the cell. The micelles centred at the corners and in the centre should be statistically disordered or freely rotating, whereas non-oriented samples is expected to decrease at the transition from the hexagonal columnar to the inverted micellar cubic those at the surfaces are only rotationally disordered.50 We propose that the thermotropic cubic phases of the triple mesophase.Those compounds with a columnar/cubic dimorphism are therefore especially useful for these investigations. chain compounds 3/9, 3/10, 3/12 and 7/8–7/12 have the same structure. However, they are built up by inverted micelles and Indeed, a strong decrease of the specific conductivity is found in all cases investigated at the transition from the appear in the absence of any solvent.Thus, the cubic lattice should be built up by non-spherical micelles (short rods), each hexagonal columnar phase to the cubic phase. The dependence of the specific conductivity k on the inverse temperature 1/T consisting of approximately 40–50 individual molecules.In order to organize these molecules in each of these closed for compounds 7/8 and 9/12 is shown in Fig. 11. This exper- 536 J. Mater. Chem., 1998, 8(3), 529–543in a face centred lattice (0.74) which allow more eYcient packing. Such arrangements have indeed been found in lyotropic systems for some micellar cubic phases of the type 1;8 however they have never been observed for an inverted system (see Fig. 1). Interestingly, Pm3n lattices have also been found for thermotropic cubic phases of cone-shaped dendrimers. 37 Here eight nearly spherical (spherical and tetrahedral distorted) micelles were found in the cubic lattice. Because these compounds diVer in chemical structure, they could not be compared directly with the compounds described herein.Triple chain benzamides with other hydrophilic groups. The influence of the hydrophilic group on the mesomorphic properties is obvious from the comparison of the tridodecyloxy benzamides 3/12, 7/12, 9/12, 10/12 and 11/12 given in Table 8. All compounds have the same hydrophobic residue and diVer only in the number of hydroxy groups and their position with respect to each other.It clearly shows that increasing the number of hydroxy groups destabilizes the inverted micellar cubic mesophase in Fig. 12 Two possible models of the structure of the thermotropic favour of the hexagonal columnar phase. Hence, the tetraol inverted micellar cubic mesophases CubI2 of the compounds under 11/12 displays exclusively a columnar phase in contrast to the discussion.The model based on rod-like closed micelles refers to a triol 10/12 and the diols 7/12 and 9/12. proposal by Fontell et al. for normal micellar cubic phases in lyotropic systems (ref. 50). (a) Arrangement of the micelles in the primitive cubic The important influence of the position of the hydroxy lattice. The rod-like micelles are indicated a circles, the statistically groups has already been discussed for the diols 7/12 and 9/12.disordered micelles at the corners and in the centre are indicated by Comparing the N-methylglucamide 3/12 with the tetraol 11/12 black dots. (b) View upon a face of the cubic lattice. (c) Schematic reveals that, despite the fact that the number of hydroxy presentation of a rod-like micelle. groups of 3/12 is larger, it has a micellar cubic mesophase whereas the tetrahydroxy compound 11/12 is only columnar.Table 7 Comparison of the lattice parameter of the micellar cubic In the case of 11/12 the connecting position is between the mesophases (acub) and of the hexagonal columnar mesophases (ahex) four hydroxy groups. Here, all hydroxy groups are positioned of the compounds 3/9, 7/8, 9/12 and 10/12, which have a CubI2/ColH2 close to the hydrophilic–lipophilic interface giving rise to a dimorphisma large area of the hydrophilic group at the interface.In the acub/nm ahex/nm open chain carbohydrate derivative 3/12 however, the lipocompound (T /°C) L /nm ncell d1/nm nmic (T /°C) philic residue is attached to a terminal position of the polar group. In this case, not all hydroxy groups are located directly 3/9 7.93 2.7 414 40 52 4.08 at this interface and therefore the hydrophilic group appears (121) (52) to be smaller. This clearly shows that not only the number of 7/8 6.89 2.2 340 3.4 42 3.33 (75) (50) hydroxy groups but also their position in respect to the 10/12 7.99 2.6 395 4.0 49 4.05 lipophilic–hydrophilic interface is an important factor govern- (67) (40) ing the hydrophilic–lipophilic interface curvature and thus the type of mesophase.Therefore, the area at the polar–apolar aAbbreviations: L=molecular length according to CPK models interface which is occupied by the N-methylglucamide group (extended all trans-conformation of the alkyl chains), ncell=number of is even nearly comparable with that of the 1-amidopropane- molecules in each unit cell of the cubic lattice (calculated using the equation n=Vcell(NA/M)r and assuming a density of r=1 g cm-3); 2,3-diol group in compound 6/12.The small diameter of the nmic=number of molecules in each rod-like micelle of the cubic N-methylglucamido head group may be a result of the hindered mesophase. rotation about the MCH(OH)MCH(OH)M bonds which is much more diYcult than that of alkyl or oxyethylene groups, due to the larger steric interactions of OH compared to H.cylindrical micelles they should have an axial ratio d2/d1 in the order of about 1.551.51 The proposed model (Fig. 12) requires that two micelles are Conclusions arranged side by side at each cell surface of the cubic lattice. Therefore the short diameter of these rods (d1) should amount In summary we have investigated the influence of structural variations of amphiphilic benzamides on their thermotropic to not more than one half of the cubic lattice parameter acub.Indeed in all investigated compounds with a ColH2/CubI2 phase behaviour. Lamellar (SA), columnar (ColH2) and two types of cubic mesophases (CubV2 and CubI2) have been dimorphism the hexagonal lattice parameter of the columnar phases amounts to approximately half the lattice parameter of detected.The same diversity of diVerent microstructures as known for lyotropic lipid–water systems,1 for polycatenar the cubic mesophase (see Table 7). Thus, in all these cases the short diameter (d1) of the rod-like micelles in the cubic meso- thermotropic compounds18 and for polyelectrolytes4 can be found in the thermotropic phase sequence of these amphiphilic phase is nearly identical to the diameter of the columns in the hexagonal columnar mesophase below.Therefrom we can polyhydroxy derivatives. In strong analogy to lyotropic systems the type of thermotropic mesophase depends on the surface conclude that these short rod-like micelles should result from the collapse of the extended columns of the ColH2 phase into area of the hydrophilic parts of these molecules at the hydrophilic –lipophilic interface and the size of the lipophilic groups, small segments.Probably the arrangement of eight rod-like micelles in each unit cell of such a cubic lattice represents an which determine the mean interface curvature. The cylindrical aggregates of the columnar mesophase are stable over a rather energetic minimum which could be stabilized, for example by quadrupole interactions.50 If spherically micelles were to form, broad range of variation of the structural parameter.The average diameter of the columns of the double chain com- their packing coeYcient in a simple cubic lattice would amount to only 0.52.Therefore it can be expected that spherical pounds is significantly larger than that of the corresponding triple chain compounds. This could be due to a non-circular micelles should preferably arrange in a body centred (0.68) or J. Mater. Chem., 1998, 8(3), 529–543 537Table 8 Comparison of the thermotropic phase transition temperatures of the 3,4,5-tridodecyloxybenzamides 3/12, 7/12, 9/12, 10/12 and 11/12 with diVerent hydrophilic groupsa and enthalpy values of the phase transitions of compounds 10/12 and 11/12 (lower lines) T/°C DH/kJ mol-1 compound structure K ColH2 Cub12 Iso 3/12b $ 75 — — $ 185 $ 7/12 $ 45/69 — — $ 127 $ 9/12 $ 67 $ (62) $ 104c $ 10/12 $ 63d $ 64 $ 87e $ 65.3 1.3 1.0 11/12 $ 40 $ 142 — — $ 67.8 1.6 aAbbreviations: see Table 1.bRef. 36. cObtained by microscopic investigations. In the DSC heating cycles the phase transition occurs at 108 °C. dOnly partial crystallization was found. eObtained by microscopic investigations, in the DSC heating cycle the phase transition was found at 93 °C. cross section of the cylinders formed by the double chain decrease in the number of alkyl chains gives rise to a noncircular cross section of the columns.Probably there is a strong compounds. On introduction of the third chain, an eYcient packing in circular columns becomes possible. At a certain undulation within the columns of the double chain compounds which increases with decreasing chain length. These columns degree of the interface curvature however, the transition from the hexagonal columnar to a micellar cubic mesophase takes can fuse with formation of an interwoven network structure ordered in a three-dimensional cubic lattice.On further place. This can be found for compounds consisting of three long aliphatic chains and a polar group with a small area at decreasing the chain length, the interface curvature becomes zero by turning into a smectic layer structure. the polar–non-polar interface.On the basis of proton conductivity measurements and from packing considerations we pro- On increasing the alkyl chain length of appropriate triple chain compounds, density fluctuations could occur in the polar pose that this cubic lattice is built up by eight closed micelles which have a rod-like shape and represent small segments of regions of the columns leading to a disruption of the hydrogen bonding networks within the columns with formation of short the extended columns of a columnar mesophase.The transition from the cylinders of the columnar phase to rod-like micelles. Probably the arrangement of eight such rodlike micelles in a primitive cubic lattice represents an energetic the non-curved lamellar phase occurs via bicontinuous cubic mesophases (CubV2).minimum which allows an eYcient packing of the molecules. The analogy of these thermotropic mesophases with the Therefrom we can propose a model for the transition between the diVerent mesophases. If one starts with the inverted inverted phases of lyotropic systems is obvious. The hydrogen bonding networks correspond to the aqueous regions in columnar mesophase of the triple chain compounds, the 538 J.Mater. Chem., 1998, 8(3), 529–543Table 9 Melting temperatures of the 3,4-dialkoxybenzoic acids and detergent–water systems. It can therefore be assumed that all the 3,4,5-trialkoxybenzoic acids and reference values from the literature other phases which are known from lyotropic systems1,8,10 can also be observed in the thermotropic phase sequence of appropriately designed polyhydroxy amphiphiles.Furthermore the mesomorphic organization of these compounds can be largely influenced by addition of protic solvents.36 Because these protic solvents are incorporated into the head group region, signifi- cant changes of the mesophase types occur. Thus the molecules reported represent typical examples of amphotropic compounds, 52 which bridge the gap between the thermotropic mp/°C mp/°C liquid crystalline phases of rod-like, polycatenar and disc-like molecules on the one hand and the lyotropic mesophases of n found reported ref.found reported ref. detergent–water mixtures on the other. 4 146 5 142 Experimental section 6 130 128 19d 39 7 129 127 19d 41 40–41 55 Techniques 8 123 125 19d 53 9 124 124 19d 46 Transition temperatures (given in °C throughout) were meas- 10 127 123 19d 51 ured using a Mettler FP 82 HT hot stage and control unit in 12 120 121 19d 56 58 56 conjunction with a Nikon Optiphot 2 polarizing microscope and these were confirmed using diVerential scanning calorimetry (Perkin-Elmer DSC-7, heating and cooling rate: (Merck, silica gel 60 F 254).Measurements of the optical 10 K min-1).The accuracy of the transition temperatures is rotation were carried out with a Perkin-Elmer Polarimeter 341. about±0.5 K. All phase transitions, except those of cubic and crystalline phases, were completely reversible. If not otherwise Synthesis of the benzamides 1–10 stated the transition enthalpies of enantiotropic phases were obtained from the first heating scan, those of monotropic The appropriately substituted benzoic acid (1.5 mmol) and phases from the second heating scan.The accuracy of the thionyl chloride (10 ml ) were refluxed for 3 h. The excess enthalpy values is about±0.2 kJ mol-1. thionyl chloride was distilled oV and the residue was dissolved X-Ray diVraction patterns were obtained on a Guinier in dry methylene chloride (15 ml ).The amino alcohol diVractometer (Huber) operating with a Cu-Ka1 beam. The (15 mmol) was dissolved in dry dimethylformamide (30 ml ) refraction patterns were recorded with a film camera. under inert conditions and DMAP (10 mg) was added. To this Phase diagrams were established by the penetration experisolution the benzoyl chloride dissolved in methylene chloride ments and by investigation of binary mixtures.These mixtures was added with stirring at 80 °C. The resulting mixture was were investigated by optical polarizing microscopy between heated at this temperature for 4 h and was stirred for additional crossed polarizers. 24 h at room temp. Afterwards the solvent was evaporated in Measurements of the conductivity (AC, f=1 kHz) were vacuo and the residue was purified by crystallization or by carried out on cooling the samples from the isotropic melt in column chromatography. a microcapacitor (A=2 cm2, d=0.02 cm) without orientation.N-(3,4-Dibutoxybenzoyl )-1-amino-1-deoxy-D-glucitol 1/4. Materials Synthesized from 3,4-dibutyloxybenzoic acid (0.4 g, 1.5 mmol) and 1-amino-1-deoxy-D-glucitol (2.7 g, 15 mmol). The residue Ethyl 4-hydroxybenzoate, ethyl 3,4-dihydroxybenzoate (Aldrich), ethyl 3,4,5-trihydroxybenzoate (Fluka), 1-amino- was crystallized from methanol; Yield: 0.37 g (57%), K 174 (SA 156) Iso (found: C, 58.78; H, 7.91; N, 3.39. C21H35O8N requires: ethanol (Merck), 1-aminopropane-2,3-diol (Merck), 2-aminopropane- 1,3-diol (Aldrich), 2-amino-2-hydroxymethylpropane- C, 58.73; H, 8.21; N, 3.26%); [a]D20 3.3 (c 0.3, chloroform); dH (200 MHz; [2H6]DMSO; 27 °C) 0.96 (6H, t, J 6, CH3), 1,3-diol (Serva), 1-deoxy-1-methylamino-D-gluctitol (Aldrich), diallylamine (Aldrich) and the 1-bromoalkanes (Merck) were 1.44–1.49 (4H, m, CH2), 1.69–1.76 (4H, m, CH2), 3.40–3.78 (8H, m, CHOH, CH2OH, CH2N), 4.05 (4H, t, J 4, OCH2), used as obtained.The substituted carboxylic acids were obtained by etherifi- 4.32 (2H, t, J 4, OH), 4.44 (2H, m, OH), 4.87 (1H, d, J 5, OH), 7.0 (1H, d, J 7, H-ar), 7.44 (2H, t, J 2, H-ar), 8.2 (1H, cation of ethyl 3,4-dihydroxybenzoate or ethyl 3,4,5-trihydroxybenzoate with alkyl bromides and potassium carbonate in br, NH); dC (100 MHz; [2H6]DMSO; 27 °C) 166 (CO), 151, 148, 127, 121, 113 (C-ar), 72, 69, 68, 63 (CH2OH, CHOH), 68 cyclohexanone,53 followed by saponification.The resulting alkoxybenzoic acids were purified by repeated crystallization (CH2), 43 (CH2N), 31, 18 (CH2), 14 (CH3); m/z 429 (M+ 5%). from methanol. The melting temperatures of the benzoic acids, together with the reported values from the literature are N-(3,4-Dipentyloxybenzoyl )-1-amino-1-deoxy-D-glucitol 1/5. Synthesized from 3,4-dipentyloxybenzoic acid (0.4 g, 1.5 mmol) collected in Table 9.Confirmation of the structures of the products was obtained and 1-amino-1-deoxy-D-glucitol (2.7 g, 15 mmol). The residue was crystallized from methanol; Yield: 0.34 g (50%), K 172 (M by 1H and 13C NMR spectroscopy (Varian Gemini 200, Varian Unity 400 and Varian Unity 500) and mass spectrometry 162 SA 168) Iso (found: C, 60.11; H, 8.49; N, 3.02. C23H39O8N requires: C, 60.36; H, 8.60; N, 3.06%); [a]D30 13.5 (c=0.2, (Intectra GmbH, AMD 402, electron impact, 70 eV).J Values are given in Hz throughout. chloroform); dH (200 MHz; [2H6]DMSO; 27 °C) 0.88 (6H, t, J 6, CH3), 1.32–1.48 (8H, m, CH2), 1.64–1.74 (4H, m, CH2), Microanalyses were performed using a CHNF-932 (Leco Co.) elemental analyzer.Owing to the hygroscopic properties 3.41–3.78 (8H, m, CHOH, CH2OH, CH2N), 3.97 (4H, t, J 4, OCH2), 4.30 (2H, t, J 4, OH), 4.44 (2H, dd, J 5, OH), 4.84 of some compounds moisture was absorbed during sample preparation and therefore correct combustion analyses were (1H, d, J 5, OH), 7.0 (1H, d, J 4, H-ar), 7.44 (2H, t, J 2, Har), 8.2 (1H, t, J 5, NH); dC (100 MHz; [2H6]DMSO; 27 °C) not obtained for some compounds.In these cases the water content was determined by Karl–Fischer titration54 and the 168 (CO), 153, 149, 128, 122, 114 (C-ar), 73, 71, 70, 65 (CH2OH, CHOH), 68 (CH2), 41 (CH2N), 30, 29, 23 (CH2), 15 (CH3); amount of absorbed water was taken into account. The purity of all compounds was checked by thin layer chromatography m/z 458 (M+1 100%).J. Mater. Chem., 1998, 8(3), 529–543 539N-(3,4-Diheptyloxybenzoyl )-1-amino-1-deoxy-D-glucitol 1/7. chloroform–ethanol, 1053) and crystallized twice from methanol; Yield: 0.25 g (29%), K 68 ColH2 174 Iso (found: C, 64.53; Synthesized from 3,4-diheptyloxybenzoic acid (0.5 g, 1.5 mmol) and 1-amino-1-deoxy-D-glucitol (2.7 g, 15 mmol). Crystallized H, 9.72; N, 2.17. C32H57O8N 0.5H2O requires: C, 64.83; H, 9.86; N, 2.36%); [a]D20 -5.8 (c 1, methanol); dH (500 MHz; twice from methanol; Yield: 0.40 g (52%), K 169 ColH2 203 Iso; (found: C, 62.82; H, 8.94; N, 2.58.C27H47O8N requires: C, [2H6]DMSO; 30 °C) 0.85 (6H, t, J 7, CH3), 1.25–1.43 (24H, m, CH2), 1.65–1.72 (4H, m, CH2), 2.95 (3H, s, NCH3), 63.12; H, 9.23; N, 2.73%); [a]D20 40 (c 0.3, chloroform); dH (500 MHz; [2H6]DMSO; 30 °C) 0.85 (6H, t, J 6, CH3), 3.32–3.56 (8H, m, CHOH, CH2OH, CH2N), 3.92–3.97 (4H, m, OCH2), 4.27–4.44 (4H, br, OH), 4.87 (1H, br, OH), 6.95 1.26–1.45 (16H, m, CH2), 1.68–1.71 (4H, m, CH2), 3.35–3.76 (8H, m, CHOH, CH2OH, CH2N), 3.97 (4H, q, J 5, OCH2), (3H, br, H-ar); dC (126 MHz; [2H6]DMSO; 30 °C) 171 (CO), 149, 148, 129, 120, 113 (C-ar), 71, 70, 68, 63 (CH2OH, CHOH), 4.33 (2H, t, J 6, OH), 4.42 (1H, d, J 5, OH), 4.46 (1H, d, J 5, OH), 4.87 (1H, d, J 5, OH), 6.96 (1H, d, J 4, H-ar), 7.42 (2H, 68 (CH2), 31 (NCH3), 29, 25, 22 (CH2), 14 (CH3); m/z 583 (M+ 3.6%).t, J 2, H-ar), 8.2 (1H, t, J 6, NH); dC (126 MHz; [2H6]DMSO; 30 °C) 166 (CO), 151, 148, 126, 120, 112 (C-ar), 72, 69, 68, 63 (CH2OH, CHOH), 68 (CH2), 31, 29, 28, 25, 22 (CH2), 14 N-(3,4-Dihexadecyloxybenzoyl )-1-deoxy-1-methylamino-Dglucitol 2/16.Synthesized from 3,4-dihexadecyloxybenzoic acid (CH3); m/z 513 (M+ 9.3%). (0.9 g, 1.5 mmol) and 1-deoxy-1-methylamino-D-glucitol (2.9 g, 15 mmol). The residue was crystallized twice from methanol; N-(3,4-Dinonyloxybenzoyl )-1-amino-1-deoxy-D-glucitol 1/9. Yield: 0.60 g (51%), K 82 ColH2 194 Iso (found: C, 68.42; H, Synthesized from 3,4-dinonyloxybenzoic acid (0.6 g, 1.5 mmol) 10.90; N, 1,79.C46H85O8N 1.5H2O requires: C, 68.45; H, 10.99; and 1-amino-1-deoxy-D-glucitol (2.7 g, 15 mmol). The residue N, 1.74%); [a]D30 2.1 (c 0.6, chloroform); dH (400 MHz; was crystallized twice from chloroform–methanol, 1051; Yield: [2H6]DMSO; 40 °C) 0.84 (6H, t, J 6, CH3), 1.22–1.42 (52H, 0.37 g (43%), K 166 ColH2 239 Iso (found: C, 65.10; H, 9.72; m, CH2), 1.67–1.70 (4H, m, CH2), 2.95 (3H, s, NCH3), N, 2.31. C31H55O8N requires: C, 65.35; H, 9.73; N, 2.46%); 3.39–3.55 (8H, m, CHOH, CH2OH, CH2N), 3.92–3.97 (4H, [a]D20 8.9 (c 0.8, chloroform); dH (500 MHz; [2H6] m, OCH2), 4.24–4.41 (4H, br, OH), 4.82 (1H, d, J 5, OH), DMSO; 30 °C) 0.85 (6H, t, J 7, CH3), 1.25–1.42 (24H, m, 6.94 (3H, br, H-ar); dC (126 MHz; [2H6]DMSO; 40 °C) 171 CH2), 1.68–1.70 (4H, m, CH2), 3.36–3.75 (8H, m, CHOH, (CO), 149, 148, 129, 120, 113 (C-ar), 72, 71, 70, 68, 63 (CH2OH, CH2OH, CH2N), 3.98 (4H, q, J 6, OCH2), 4.25 (2H, t, J 6, CHOH), 31 (NCH3), 29, 25, 22, 21 (CH2), 14 (CH3); m/z 779 OH), 4.36 (1H, d, J 5, OH), 4.39 (1H, d, J 5, OH), 4.79 (1H, (M+ 0.6%).d, J 4, OH), 6.95 (1H, d, J 9, H-ar), 7.42 (2H, br, H-ar), 8.16 (1H, s, NH); dC (126 MHz; [2H6]DMSO; 30 °C) 166 (CO), N -(3,4,5 -Trioctyloxybenzoyl ) -1-deoxy-1-methylamino-D- 151, 148, 127, 120, 113 (C-ar), 72, 71, 70, 69, 68, 63 (CH2OH, glucitol 3/8.Synthesized from 3,4,5-trioctyloxybenzoic acid CHOH), 68 (CH2), 29, 25, 22 (CH2), 14 (CH3); m/z 569 (0.8 g, 1.5 mmol) and 1-deoxy-1-methylamino-D-glucitol (2.9 g, (M+ 10%). 15 mmol). Purified by column chromatography (eluent chloroform –ethanol, 1053); Yield: 0.26 g (25%), K 75 ColH2 147 Iso N- (3,4 - Diheptyloxybenzoyl ) - 1 - deoxy - 1 - methylamino -D- (found: C, 66.04; H, 10.11; N, 1.94. C38H69O9N 0.3H2O glucitol 2/7. Synthesized from 3,4-diheptyloxybenzoic acid requires: C, 66.21; H, 10.18; N, 2.03%); [a]D20-0.5 (c 1, (0.5 g, 1.5 mmol) and 1-deoxy-1-methylamino-D-glucitol (2.9 g, methanol); dH (500 MHz; [2H6]DMSO; 30 °C) 0.86 (9H, t, J 15 mmol).Purified by column chromatography (eluent chloro- 6, CH3), 1.25–1.42 (30H, m, CH2), 1.62–1.70 (6H, m, CH2), form–ethanol, 1053) and crystallization from methanol; Yield: 2.94 (3H, s, NCH3), 3.30–3.48 (8H, m, CHOH, CH2OH, 0.21 g (27%), K 96 M 118 (SA 118) Iso (found: C, 62.62; H, CH2N), 3.86 (2H, t, J 6, OCH2), 3.92 (4H, t, J 6, OCH2), 9.34; N, 2.47.C28H49O8N requires: C, 62.89; H, 9.58; N, 2.72%); 4.28–4.45 (4H, br, OH), 4.88 (1H, br, OH), 6.62 (2H, br, H- [a]D20 -3.0 (c 1, methanol); dH (500 MHz; [2H6]DMSO; ar); dC (50 MHz; [2H6]DMSO; 27 °C) 166 (CO), 152, 138, 132, 30 °C) 0.86 (6H, t, J 6, CH3), 1.26–1.39 (16H, m, CH2), 119 (C-ar), 72, 71, 68, 63 (CH2OH, CHOH), 68 (CH2), 51 1.65–1.72 (4H, m, CH2), 2.95 (3H, s, NCH3), 3.30–3.56 (8H, (CH2N), 31 (NCH3), 30, 29, 26, 22 (CH2), 14 (CH3); m/z 683 m, CHOH, CH2OH, CH2N), 3.92–3.97 (4H, m, OCH2), (M+ 15%). 4.27–4.44 (4H, br, OH), 4.86 (1H, br, OH), 6.94 (3H, br, Har); dC (126 MHz; [2H6]DMSO; 30 °C) 170 (CO), 149, 148, N- (3,4,5-Trinonyloxybenzoyl ) -1-deoxy-1-methylamino-D- 129, 120, 112 (C-ar), 72, 68, 63 (CH2OH, CHOH), 68 (CH2), glucitol 3/9.Synthesized from 3,4,5-trinonyloxybenzoic acid 31 (NCH3), 29, 28, 25, 22 (CH2), 14 (CH3); m/z 527 (M+ 1.8%). (0.8 g, 1.5 mmol) and 1-deoxy-1-methylamino-D-glucitol (2.9 g, 15 mmol). Purified by column chromatography (eluent chloro- N- (3,4 - Dioctyloxybenzoyl ) - 1 - deoxy - 1 - methylamino - D - form–methanol, 1051); Yield: 0.16 g (15%), K 50 ColH2 102 glucitol 2/8.Synthesized from 3,4-dioctyloxybenzoic acid (0.6 g, CubI2 143 Iso (found: C, 68.13; H, 10.67; N, 2.06. C41H75O9N 1.5 mmol) and 1-deoxy-1-methylamino-D-glucitol (2.9 g, requires: C, 67.83; H, 10.41; N, 1.93%); [a]D20 -6.1 (c 1, 15 mmol). Purified by column chromatography (eluent chloro- methanol); dH (200 MHz; [2H6]DMSO; 27 °C) 0.87 (9H, t, J form–ethanol, 1053) and crystallized from methanol; Yield: 6, CH3), 1.28–1.71 (42H, m, CH2), 2.96 (3H, s, NCH3), 0.28 g (34%), K 87 ColH2 155 Iso (found: C, 63.69; H, 9.62; N, 3.49–3.52 (8H, m, CHOH, CH2OH, CH2N), 3.87–3.98 (6H, 2.32.C30H53O8N 0.5H2Orequires: C, 63.98; H, 9.66; N, 2.49%); m, OCH2), 4.25–4.55 (4H, br, OH), 4.80–5.05 (1H, br, OH), [a]D20 -5.9 (c 1, methanol); dH (500 MHz; [2H6]DMSO; 6.60–6.80 (2H, br, H-ar); dC (100 MHz, [2H6]DMSO, 27 °C) 30 °C) 0.85 (6H, t, J 6, CH3), 1.25–1.43 (20H, m, CH2), 166 (CO), 152, 138, 132, 103 (C-ar), 72, 71, 68, 63 (CH2OH, 1.67–1.72 (4H, m, CH2), 2.95 (3H, s, NCH3), 3.30–3.55 (8H, CHOH), 31 (N-CH3), 30, 29, 28, 25, 22 (CH2), 14 (CH3); m/z m, CHOH, CH2OH, CH2N), 3.92–3.97 (4H, m, OCH2), 725 (M+ 28.6%). 4.27–4.45 (4H, br, OH), 4.87 (1H, br, OH), 6.95 (3H, br, Har); dC (126 MHz; [2H6]DMSO; 30 °C) 170 (CO), 149, 148, N- (3,4,5 -Tridecyloxybenzoyl ) - 1-deoxy -1-methylamino-D- 129, 120, 113 (C-ar), 71, 68, 63 (CH2OH, CHOH), 68 (CH2), glucitol 3/10. Synthesized from 3,4,5-tridecyloxybenzoic acid 31 (NCH3), 29, 26, 22 (CH2), 14 (CH3); m/z 555 (M+ 2.1%).(0.9 g, 1.5 mmol) and 1-deoxy-1-methylamino-D-glucitol (2.9 g, 15 mmol).The residue was purified by chromatography (eluent chloroform–methanol, 1051) and crystallized twice from ace- N- (3,4 - Dinonyloxybenzoyl ) - 1 - deoxy - 1 - methylamino - D - glucitol 2/9. Synthesized from 3,4-dinonyloxybenzoic acid tone; Yield: 0.19 g (15%), K 59 ColH2 89 CubI2 158 Iso (found: C, 68.60; H, 10.54; N, 1.89. C44H81O9N requires: C, 68.80; H, (0.6 g, 1.5 mmol) and 1-deoxy-1-methylamino-D-glucitol (2.9 g, 15 mmol).Purified by columnn chromatography (eluent 10.63; N, 1.82%); [a]D30 -5.6 (c 1, methanol); dH (200 MHz; 540 J. Mater. Chem., 1998, 8(3), 529–543[2H6]DMSO; 27 °C) 0.87 (9H, t, J 6, CH3), 1.27–1.72 (48H, (CHOH), 69, 68 (OCH2), 64 (CH2OH), 43 (CH2NH), 31, 29, 25, 22 (CH2), 14 (CH3); m/z 563 (M+ 78.3%).m, CH2), 2.96 (3H, s, NCH3), 3.44–3.52 (8H, br, CHOH, CH2OH, CH2N), 3.84–3.98 (6H, m, OCH2), 4.31–4.51 (4H, br, OH), 4.85–5.05 (1H, br, OH), 6.60–6.76 (2H, br, H-ar); dC 1-(3,4,5-Trihexyloxybenzoylamino)propane-2,3-diol 7/6. (100 MHz, [2H6]DMSO, 27 °C) 165 (CO), 152, 138, 132, 104 Synthesized from 3,4,5-trihexyloxybenzoic acid (0.6 g, (C-ar), 72, 71, 68, 63 (CH2OH, CHOH), 31 (N-CH3), 30, 29, 1.5 mmol) and 1-aminopropane-2,3-diol (1.4 g, 15 mmol). 25, 22 (CH2), 14 (CH3); m/z 767 (M+ 11.4%). Purified by column chromatography (eluent: chloroform–ethanol 1053) followed by preparative thin layer chromatography N-(3,5-Didodecyloxybenzoyl )-1-amino -1-deoxy-D-glucitol (Chromatotron, eluent: chloroform). Yield: 0.19 g (26%); K 49 4/12. Synthesized from 3,5-didodecyloxybenzoic acid (0.7 g, ColH2 91 Iso (found: C, 67.50; H, 9.69; N, 2.78.C28H49O6N 1.5 mmol) and 1-amino-1-deoxy-D-glucitol (2.7 g, 15 mmol). requires: C, 67.85; H, 9.96; N, 2.83%); dH (500 MHz, Crystallized twice from methanol; Yield: 0.10 g (10%), K1 45 [2H6]DMSO; 30 °C) 0.88 (9H, t, J 4, CH3), 1.2–1.45 (18H, m, K2 87 ColH2 175 Iso (found: C, 67.44; H, 10.59; N, 2.0. CH2), 1.6–1.75 (6H, m, CH2), 3.3–3.4 (5H, m, CH2NH, CHOH, C37H67O8N 0.2H2O requires: C, 67.59; H, 10.33; N, 2.13%); CH2OH), 3.85 (2H, t, J 6, OCH2), 4.0 (4H, t, J 6, OCH2), [a]D20 2.3 (c 1, methanol); dH (500 MHz; [2H6]DMSO; 30 °C) 4.53 (1H, t, J 6, OH), 4.75 (1H, d, J 5, OH), 7.15 (2H, s, H- 0.84 (6H, t, J 6, CH3), 1.23–1.39 (36H, m, CH2), 1.65–1.69 ar), 8.3 (1H, t, J 6, NH); dC (126 MHz; [2H6]DMSO; 30 °C) (4H, m, CH2), 3.35–3.74 (8H, m, CHOH, CH2OH, CH2N), 166 (CO), 152, 140, 129, 106 (C-ar), 72 (CHOH), 70, 68 3.95 (4H, t, J 6, OCH2), 4.32 (2H, t, J 4, OH), 4.41 (1H, d, J (OCH2), 64 (CH2OH), 43 (CH2NH), 31, 30, 29, 25, 22 (CH2), 5, OH), 4.46 (1H, d, J 5, OH), 4.85 (1H, d, J 5, OH), 6.56 14 (CH3); m/z 495 (M+ 7.4%).(1H, s, H-ar), 6.96 (2H, s, H-ar), 8.28 (1H, t, J 5, NH); dC (50 MHz; [2H6]DMSO; 27 °C) 166 (CO), 159, 136, 106, 104 1-(3,4,5-Triheptyloxybenzoylamino)propane-2,3-diol 7/7.(C-ar), 72, 71, 69, 68, 63 (CH2OH, CHOH), 68 (CH2), 31, 29, Synthesized from 3,4,5-triheptyloxybenzoic acid (0.7 g, 25, 22 (CH2), 14 (CH3); m/z 653 (M+ 5.0%). 1.5 mmol) and 1-aminopropane-2,3-diol (1.4 g, 15 mmol). Purified by column chromatography (eluent: chloroform–etha- 1-(4-Dodecyloxybenzoylamino)propane-2,3-diol 5/12.Synnol 1053) and by preparative thin layer chromatography thesized from 4-dodecyloxybenzoic acid (0.5 g, 1.5 mmol) and (Chromatotron, eluent: chloroform). Yield: 0.22 g (27%); K 46 1-aminopropane-2,3-diol (1.4 g, 15 mmol). The residue was ColH2 92 Iso (found: C, 68.97; H, 10.44; N, 2.57. C31H55O6N purified twice by column chromatography (eluent: chloroform– requires: C, 69.22; H, 10.31; N, 2.61%); dH (500 MHz, ethanol 1053; chloroform–methanol 1051).Yield: 0.38 g (33%); [2H6]DMSOt; 30 °C) 0.86 (9H, t, J 5, CH3), 1.2–1.45 (24H, K1 80 K2 89 SA 132 Iso (found C, 69.74; H, 9.85; N, 3.70. m, CH2), 1.6–1.75 (6H, m, CH2), 3.1–3.3, 3.5–3.7 (5H, m, C22H37O4N requires: C, 69.62; H, 9.83; N, 3.69%); dH CH2NH, CHOH, CH2OH), 3.87 (2H, t, J 6, OCH2), 3.98 (4H, (200 MHz, CDCl3; 27°C) 0.85 (3H, t, J 6, CH3), 1.2–1.5 (18H, t, J 6, OCH2), 4.52 (1H, t, J 6, OH), 4.75 (1H, d, J 5, OH), m, CH2), 2.2–2.4 (2H, m, CH2), 3.1–3.2 (2H, m, OH), 3.6 (5H, 7.1 (1H, s, H-ar), 8.3 (1H, t, J 6, NH); dC (126 MHz; m, CH2NH, CHOH, CH2OH), 3.95 (2H, t, J 6,OCH2), 6.55 [2H6]DMSO; 30 °C) 166 (CO), 152, 140, 129, 106 (C-ar), 72 (1H, t, J 3, NH), 6.9 (2H, d, J 2, H-ar), 7.7 (2H, d, J 2, H-ar); (CHOH), 70, 68 (OCH2), 64 (CH2OH), 43 (CH2NH), 31, 30, dC (126 MHz; [2H6]DMSO; 30 °C) 166 (CO), 161, 129, 126, 29, 28, 25, 22 (CH2), 14 (CH3); m/z 537 (M+ 81.4%). 114 (C-ar), 71 (CHOH), 67 (OCH2), 64 (CH2OH), 43 (CH2NH), 31, 29, 28, 25, 22 (CH2), 14 (CH3). 1-(3,4,5-Trioctyloxybenzoylamino)propane-2,3-diol 7/8. Synthesized from 3,4,5-trioctyloxybenzoic acid (0.8 g, 1-(3,4-Dihexyloxybenzoylamino)propane-2,3-diol 6/6.Syn- 1.5 mmol) and 1-aminopropane-2,3-diol (1.4 g, 15 mmol). thesized from 3,4-dihexyloxybenzoic acid (0.5 g, 1.5 mmol) and Purified twice by column chromatography (eluent: chloroform– 1-aminopropane-2,3-diol (1.4 g, 15 mmol). The residue was ethanol, 1053; chloroform–methanol, 1051).Yield: 0.41g (47%); purified twice by column chromatography (eluent: chloroform– K 59 ColH2 74 CubI2 85 Iso (found: C, 70.26; H, 10.75; N, 2.30. ethanol, 1053; chloroform–methanol, 1051) and crystallized C34H61O6N requires: C, 70.41; H, 10.61; N, 2.42%); dH twice from methanol. Yield: 0.22 g (24%); K 79 ColH2 87 Iso (500 MHz, [2H6]DMSO; 30 °C) 0.85 (9H, t, J 6, CH3), 1.2–1.35 (found C, 65.76; H, 9.40; N, 3.20.C22H37O5N 2H2O requires: (24H, m, CH2), 1.4–1.48 (6H, m, CH2), 1.6–1.65 (2H, m, CH2), C, 65.28; H, 9.24; N, 3.47%); dH (200 MHz, [2H6] 1.7–1.75 (4H, m, CH2), 3.1–3.2, 3.3–3.4, 3.5–3.7 (m, 5H, DMSO; 20 °C) 0.9 (6H, t, J 6, CH3), 1.25–1.5 (12H, m, CH2), CH2NH, CHOH, CH2OH), 4.50 (1H, t, J 6, OH), 4.75 (1H, 1.65–1.75 (4H, m, CH2), 3.5 (5H, m, CH2NH, CHOH, CH2OH, d, J 5, OH), 7.25 (1H, s, H-ar), 8.3 (1H, t, J 6, NH); dC 5H), 3.9 (4H, t, J 3, OCH2), 4.6 (1H, t, OH), 4.8 (1H, d, J 5, (126 MHz; [2H6]DMSO; 30 °C) 166 (CO),152, 140, 129, 106 OH), 6.9 (1H, d, J 9, H-ar), 7.45 (2H, d, J 3, H-ar), 8.3 (1H, (C-ar), 72 (CHOH), 70, 68 (OCH2), 64 (CH2OH), 43 (CH2NH), br, NH); dC (100 MHz; [2H6]DMSO; 27 °C) 166 (CO), 151, 31, 31, 30, 29, 25, 22 (CH2), 14 (CH3); m/z 579 (M+ 100). 148, 127, 121, 113, 112 (C-ar), 71 (CHOH), 69, 68 (OCH2), 63 (CH2OH), 43 (CH2NH), 31, 29, 25, 22 (CH2), 14 (CH3); m/z 395 (M+ 23.5%). 1-(3,4,5-Trinonyloxybenzoylamino)propane-2,3-diol 7/9. Synthesized from 3,4,5-trinonyloxybenzoic acid (0.8 g, 1.5 mmol) and 1-aminopropane-2,3-diol (1.4 g, 15 mmol). 1-(3,4-Didodecyloxybenzoylamino)propane-2,3-diol 6/12.Synthesized from 3,4-didodecyloxybenzoic acid (0.7 g, Crystallized three times from methanol. Yield: 0.33 g (35%); K 49 CubI2 104 Iso (found: C, 70.76; H, 10.61; N, 2.19. 1.5 mmol) and 1-aminopropane-2,3-diol (1.4 g, 15 mmol). Purified by column chromatography (eluent: chloroform–meth- C37H67O6N 0.2 H2O requires: C, 71.04; H, 10.86; N, 2.24%); dH (500 MHz, [2H6]DMSO; 30 °C) 0.84 (9H, t, J 7, CH3), anol, 1051) and crystallized four times from methanol.Yield: 0.44 g (52%); K 98 ColH2 148 Iso (found: C, 72.22; H, 10.97; 1.2–1.35 (30H, m, CH2), 1.4–1.48 (6H, m, CH2), 1.58–1.66 (2H, m, CH2), 1.68–1.74 (4H, m, CH2), 3.12–3.2, 3.3–3.4, N, 2.47. C34H61O5N requires: C, 72.42; H, 10.90; N, 2.48%); dH (200 MHz, [2H6]DMSO; 27 °C) 0.9 (6H, t, J 6, CH3) 3.58–3.64 (5H, m, CH2NH, CHOH, CH2OH), 3.87 (2H, t, J 6, OCH2), 3.95 (4H, t, J 6, OCH2), 4.5 (1H, t, J 6, OH), 4.75 1.2–1.5 (36 H, m, CH2), 1.6–1.8 (4H, m, CH2), 3.5–3.7 (5H, m, CH2NH, CHOH, CH2OH), 4.0 (4H, t, J 6, OCH2), 4.6 (1H, (1H, d, J 5, OH), 7.15 (2H, s, H-ar), 8.3 (1H, t, J 6, NH); dC (126 MHz; [2H6]DMSO; 30 °C) 168 (CO), 152, 140, 129, 106 t, J 5, OH), 4.8 (1H, d, J 5, OH), 7.0 (1H, d, J 9, H-ar), 7.5 (2H, br, H-ar), 8.3 (1H, br, NH); dC (50 MHz; [2H6]DMSO; (C-ar), 72 (CHOH), 71, 69 (OCH2), 64 (CH2OH), 43 (CH2NH), 31, 30, 29.0, 26, 22 (CH2), 14 (CH3); m/z 621 (M+ 100%). 25 °C) 166 (CO), 151, 148, 127, 121, 113, 112 (C-ar), 71 J.Mater. Chem., 1998, 8(3), 529–543 5411-(3,4,5-Tridodecyloxybenzoylamino)propane-2,3-diol 7/12. additional 24 h at room temp.Afterwards the solvent was evaporated in vacuo and the residue was purified by column Synthesized from 3,4,5-tridodecyloxybenzoic acid (1.0 g, 1.5 mmol) and 1-aminopropane-2,3-diol (1.4 g, 15 mmol). chromatography (eluent chloroform–methanol, 1051) to give N,N-diallyl-3,4,5-tridodecyloxybenzamide. Yield: 0.86 g (76%); Purified by chromatography (eluent: chloroform–ethanol, 1053) and crystallized twice from methanol.Yield: 0.82 g mp 32 °C. N,N-Diallyl-3,4,5-tridodecyloxybenzamide (0.75 g, 1 mmol) and N-methylmorpholine N-oxide (0.23 g of a 60% (73%); K1 45 K2 69 CubI2 126 Iso (found: C, 73.98; H, 11.67; N, 1.87. C46H85O6N requires: C, 73.85; H, 11.45; N, 1.87%); solution in water, 2 mmol) were dissolved in acetone (20 ml ). OsO4 (5 ml of a 0.01 M solution in tert-butyl alcohol) was dH (500 MHz, [2H6]DMSO; 40 °C) 0.84 (9H, t, J 6, CH3), 1.2–1.8 (CH2, m, 60H), 3.35–3.65 (5H, m, CH2NH, CHOH, added and the mixture was stirred at room temp.for 24 h. Afterwards Na2SO3 (5 ml of a saturated solution) was added CH2OH), 3.8–4.0 (6H, br, OCH2), 4.45 (1H, t, J 6, OH), 4.68 (1H, d, J 4, OH), 7.18 (2H, s, H-ar), 8.24 (1H, br, NH); dC and the resulting slurry was stirred for 2 h at 25 °C.The mixture was filtered over a silica bed. The residue was washed (126 MHz; [2H6]DMSO; 40 °C) 166 (CO), 152, 140, 135, 129, 106 (C-ar), 72 (CH-OH), 70, 68 (OCH2), 64 (CH2OH), 43 twice with acetone (50 ml ). The organic solutions were combined and the solvent was evaporated. The residue was dis- (CH2NH), 31, 30, 29, 25, 22 (CH2), 14 (CH3); m/z 747 (M+ 100%).solved in diethyl ether, washed with water (25 ml ), 5% H2SO4 (25 ml ), saturated NaHCO3 solution (25 ml ) and brine (25 ml ). The organic layer was dried with Na2SO4 and the solvent was 2-(3,4-Dihexyloxybenzoylamino)propane-1,3-diol 8/8. Syndistilled oV. The residue was crystallized twice from methanol. thesized from 3,4-dihexyloxybenzoic acid (0.5 g, 1.5 mmol) and Yield: 0.31 g (38%); K 40 ColH2 142 Iso (found: C, 71.12; 2-aminopropane-1,3-diol (1.4 g, 15 mmol).Purified by column H, 10.85; N, 1.68. C26H57O9N requires: C, 71.42; H, 11.15; chromatography (eluent: chloroform–ethanol 1053) and by N, 1.70%); dH (200 MHz, [2H6]DMSO; 40 °C) 0.84 (9H, preparative thin layer chromatography (Chromatotron, eluent: J 6, CH3), 1.24–1.69 (60H, m, CH2), 2.47–2.51 (10H, m, chloroform).Yield: 0.35 g (59%); K1 64 K2 108 (CubV2 50) Iso N(CH2CHOHCH2OH)2), 3.85–3.95 (6H, m, OCH2), 4.46 (2H, (found: C, 66.48; H, 9.37; N, 3.37. C22H37O5N requires: C, br, OH),4.88 (2H, br, OH), 6.69 (2H, s, H-ar); dC (50 MHz; 66.79; H, 9.43; N, 3.54%); dH (500 MHz, [2H6]DMSO; 25 °C) [2H6]DMSO; 40 °C) 171 (CO), 152, 132, 106 (C-ar), 69, 68, 51 0.87 (6H, t, J 6, CH3), 1.2–1.35 (8H, m, CH2), 1.4–1.5 (4H, m, [N(CH2CHOHCH2OH)2], 72, 41, 39, 38, 31, 30, 29, 26, 21 CH2), 1.65–1.75 (4H, m, CH2), 3.5 (5H, t, J 6, CHNH, (CH2), 14 (CH3); m/z 821 (M+ 16%).CH2OH), 3.9–4.05 (4H, m, OCH2), 4.6 (2H, t, J 5, OH), 6.98 (1H, d, J 8, H-ar), 7.45 (2H, m, H-ar), 7.74 (1H, d, J 3, NH); This work was supported by the Deutsche dC (50 MHz; [2H6]DMSO; 27 °C) 163 (CO), 151, 148, 127, Forschungsgemeinschaft and the Fonds der Chemischen 121, 103, 102 (C-ar), 69 (OCH2), 61 (CH2OH), 54 (CHNH), Industrie. 31, 29, 25, 22 (CH2), 13 (CH3); m/z 395 (M+ 20.7%). 2-(3,4,5-Tridodecyloxybenzoylamino)propane-1,3-diol 9/12. References Synthesized from 3,4,5-tridodecyloxybenzoic acid (1.0 g, 1 J. M. Seddon and R. H. Templer, in Handbook of Biological 1.5 mmol) and 2-aminopropane-1,3-diol (1.4 g, 15 mmol).Physics, vol. 1, ed. R. Lipowsky and E. Sackmann, Elsevier, Crystallized twice from methanol. Yield: 0.65 g, (58%); K 67 Amsterdam, 1995. (ColH2 62) CubI2 104 Iso (found: C, 73.82; H, 11.64; N, 1.78. 2 G. Lindblom and L. Rilfors, Biochim. Biophys. Acta, 1989, 988, 221. C46H85O6N requires: C, 73.83; H, 11.46; N, 1.87%); dH 3 (a) A.E. Skoulios, in Developments in Block Copolymers, vol. 1, ed. I. Goodman, Applied Science Publishers, London, 1982, p. 81; (200 MHz, [2H6]DMSO; 27 °C) 0.95 (9H, t, J 6, CH3), 1.2–1.8 (b) A. E. Skoulios, in Advances in L iquid Crystals, vol. 1 ed. G. H. (60H, m, CH2), 3.5 (5H, t, J 6, CH2NH, CH2OH), 3.95 (6H, Brown, Academic Press, New York, 1975, p. 169. m, OCH2), 4.6 (2H, t, J 5, OH), 7.2 (1H, s, H-ar), 7.85 (1H, 4 M.Antonietti and C. Go� ltner, Angew. Chem., 1997, 109, 944. br, NH); dC (126 MHz; [2H6]DMSO; 40 °C) 165 (CO), 152, 5 (a) V. Luzzati and A. P Spengt, Nature (L ondon), 1967, 215, 701; 140, 129, 106 (C-ar), 72 (OCH2), 68 (CH2OH), 54 (CHNH), (b) J. M. Seddon and R. H. Templer, Phil. T rans. R. Soc. L ond., 31, 30, 29, 26, 25, 22 (CH2), 14 (CH3); m/z 747 (M+ 100%). 1993, A344, 377. 6 G. J. T. Tiddy, Phys. Rep., 1980, 57, 1. 7 J. N. Israelachvili, D. J. Mitchell and B. W. Ninham, J. Chem. Soc., 2 - (3, 4, 5 - Tridodecyloxybenzoylamino) - 2 - (hydroxymethyl) Faraday T rans. 2, 1976, 72, 1525. propane-1,3-diol 10/12. Synthesized from 3,4,5-tridodecyloxy- 8 A. Gulik, H. Delacroix, G. Kirschner and V. Luzzati, J. Phys. II benzoic acid (1.0 g, 1.5 mmol) and 2-amino-2-(2-hydroxymeth- Fr., 1995, 5, 445.yl )propane-1,3-diol (1.8 g, 15 mmol). Purified by column 9 P. Mariani, V. Luzzati and H. Delacroix, J.Mol. Biol., 1988, 165. 10 Y. Hendrikx and B. Pansu, J. Phys. II Fr., 1996, 6, 33. chromatography (eluent: chloroform–ethanol, 1053) and crys- 11 V. Luzzati, in Biological Membranes, Vol. 1, ed. D. Chapman, tallized twice from methanol.Yield: 0.29 g, (25%); K 63 ColH2 Academic Press, London 1968, pp. 71–123 64 CubI2 87 Iso (found: C, 71.25; H, 10.98; N, 1.68. 12 P. KekicheV and B. Cabane, J. Phys., 1987, 48, 1571. C47H87O7N 0.6H2O requires C, 71.55; H, 11.27; N, 1.78%); dH 13 M. Clerc, J. Phys. II Fr., 1996, 6, 961. (500 MHz, [2H6]DMSO; 40 °C) 0.84 (9H, t, J 7, CH3), 1.2–1.45 14 (a) G.W. Gray, B. Jones and F. Marson, J. Chem. Soc., 1957, 393; (54H, m, CH2), 1.6 (2H, m, CH2), 1.8 (4H, m, CH2), 3.65 (6H, (b) D. Demus, G. Kunicke, J. Neelsen and H. Sackmann, Z. Naturforsch., T eil A, 1968, 23, 84; (c) S. Diele, P. Brandt and d, J 6, C(CH2)3, 3.9 (2H, t, J 6, OCH2), 4.0 (4H, J 6, OCH2), H. Sackmann, Mol. Cryst. L iq. Cryst., 1992, 17, 163; 4.7 (3H, t, J 6, OH), 7.05 (2H, s, H-ar), 7.15 (1H, s, NH); dC (d) S.Kutsumizu, M. Yamada and S. Yano, L iq. Cryst., 1994, (125 MHz; [2H6]DMSO; 40 °C) 167 (CO), 152, 140, 130, 106 16, 1109. (C-ar), 72 (OCH2), 68 (CH2OH), 62 (C(CH3)3), 31, 29, 26, 22 15 (a) G. Etherington, A. J. Leadbetter, X. J. Wang, G. W. Gray and (CH2), 13 (CH3); m/z 777 (M+ 17.9%). T. Tajbakhsh, L iq. Cryst., 1986, 1, 209; (b) S.Yano, Y. Mori and S. Kutsumizu, L iq. Cryst., 1991, 9, 907. 16 (a) H. Schubert, J. Hausschild, D. Demus and S. HoVmann, Z. Synthesis of N,N-bis(3,4-dihydroxypropyl )-3,4,5-tridodecyloxy- Chem., 1978, 18, 256; (b) D. Demus, A. Gloza, H. Hartung, benzamide 11/12 I. Rapthel and A. Wiegeleben, Cryst. Res. T echnol., 1981, 16, 1445. 17 P. A. Spengt and A. E. Skoulios, Acta Crystallogr., 1966, 21, 892.The 3,4,5-tridodecyloxybenzoic acid (1.0 g, 1.5 mmol) and 18 H.-T. Nguyen, C. Destrade and J. Malthete, Adv. Mater., 1997, thionyl chloride (10 ml ) were refluxed for 3 h. The excess 9, 375. thionyl chloride was distilled oV and the residue was dissolved 19 (a) Y. Fang, A. M. Levelut and C. Destrade, L iq. Cryst., 1990, 7, in dry methylene chloride (15 ml ). Diallylamine (1.5 g, 265; (b) A.M. Levelut and Y. Fang, Colloq. Phys., 1990, C7, 299; 15 mmol) and DMAP (10 mg) were added. The resulting (c) J. Malthete, H. T. Nguyen and C. Destrade, L iq. Cryst., 1993, 13, 171; (d) H. T. Nguyen, C. Destrade and J. Malthete, L iq. Cryst., mixture was heated for 4 h at 80 °C and was stirred for an 542 J. Mater. Chem., 1998, 8(3), 529–5431990, 8, 797; (e) C.Destrade, H. T. Nguyen, C. Alstermark, 39 Thermotropic properties of single chain N-alkanoyl-1-deoxy-1- methylamino-D-glucitols have been reported: J. W. Goodby, M. A. G. Lindsten, M. Nilsson and B. Otterholm,Mol. Cryst. L iq. Cryst., 1990, 180B, 265. Marcus, E. Chin, P. L. Finn and B. Pfannemu� ller, L iq. Cryst., 1988, 3, 1569. More recently the amphotropic behaviour of double chain 20 H.T. Nguyen, G. Sigaud, M. F. Achard, F. Hardouin, R. J. Twieg and K. Betterton, L iq. Cryst., 1991, 10, 389. N-acyl-N-alkyl-1-amino-D-glucitols has been described: H. van Doren and K. R. Terpstra, J. Mater. Chem., 1995, 5, 2153. Smectic 21 W. Weissflog, G. Pelzl, I. Letko and S. Diele, Mol. Cryst. L iq. Cryst., 1995, 260, 157. and columnar liquid crystalline salts of glucamine derivatives have been obtained on addition of carboxylic acids: V.Vill, 22 J. Billard, H. Zimmermann, R. Poupko and Z. Luz, J. Phys. Fr., 1989, 50, 539. H. Kelkenberg and J. Thiem, L iq. Cryst., 1992, 11, 459. 40 Columnar mesophases were found for simple benzamides: 23 B. Kohne, K. Praefcke and J. Billard, Z. Naturforsch., T eil B, 1986, 41, 1036. U. Beginn and G.Lattermann, Mol. Cryst. L iq. Cryst., 1994, 241, 215; and for N-(3,4-didecyloxybenzoyl)-2-aminoethanol: 24 (a) D. W. Bruce, D. A. Dunmur, S. A. Hudson, E. Lalinde, P. M. Maitlis, M. P. McDonald, R. Orr, P. Styring, A. S. Cherodian, U. Stebani and G. Lattermann, Macromol. Rep., 1995, A32 (Suppl. 3), 385. R. M. Richardson, J. L. Feijoo and G. Ungar, Mol. Cryst. L iq. Cryst., 1991, 206, 79; (b) D.W. Bruce, S. C. Davis, D. A. Dunmur, 41 S. Anderson, S. T. Hyde, K. Larsson and S. Lidin, Chem. Rev., 1988, 88, 221. S. A. Hudson, P. M. Maitlis and P. Styring,Mol. Cryst. L iq. Cryst., 1992, 215, 1; (c) D. W. Bruce and S. A. Hudson, J. Mater Chem., 42 K. Praefcke, A.-M. Levelut, B. Kohne and A. Eckert, L iq. Cryst, 1989, 6, 267. 1994, 4, 479; (d) D. W. Bruce, B.Donnino, D. Guillon, B. Heinrich and M. Ibn-Elhaj, L iq. Cryst., 1995, 19, 537; (e) D. W. Bruce, 43 K. Praefcke, P. Marquardt, B. Kohne, W. Stephan, A.-M. Levelut and E.Wachtel,Mol. Cryst. L iq. Cryst., 1991, 203, 149. B. Donnino, S. A. Hudson, A. M. Levelut, S. Megtert, D. Petermann and M. Veber, J. Phys. II (Fr.), 1995, 5, 289. 44 V. Percec, J. Heck, D. Tomazos, F. Falkenberg, H. Blackwell and G. J. Ungar, J. Chem. Soc., Perkin T rans 1, 1993, 2799. 25 M. J. Baena, P. Espinet, M. C. Lequerica and A. M. Levelut, J. Am Chem. Soc., 1992, 114, 4182. 45 (a) G. Lattermann and G. Staufer, L iq. Cryst., 1989, 4, 347; (b) M. Ebert, R. Kleppinger, M. Soliman, M. Wolf, J. H. WendorV, 26 C. Soulie, P. Bassoul and J. Simon, J. Chem. Soc., Chem. Commun., 1993, 114. G. Lattermann and G. Staufer, L iq. Cryst., 1990, 7, 553. 46 H. Hagsla� tt, O. So�derman and B. Jo�nsson, L iq. Cryst., 1994, 17, 27 V. Vill, F. Bachmann, J. Thiem, I. F. Pelyvas and P. Pudlo, Mol. Cryst. L iq. Cryst., 1992, 213, 57. 157. 47 J. W. Goodby, D. A. Dunmur and P. J. Collings, L iq. Cryst., 1995, 28 S. Yano, Y. Moriand and S. Kutsumizu, L iq. Cryst., 1991, 9, 907. 29 A. Douy and B. Gallot,Makromol. Chem., 1986, 187, 465. 5, 703 and references cited. 48 D. Joachimi, A. O� hlmann, W. Rettig and C. Tschierske, J. Chem. 30 (a) U. Stebani, G. Lattermann, M. Wittenberg, R. Festag and J. H. WendorV, Adv.Mater., 1994, 6, 572; (b) U. Stebani, G. Lattermann, Soc. Perkin T rans 2, 1994, 2011. 49 W. Rettig, G. Brezesinski, A. Ma�dicke, C. Tschierske, H. Zaschke R. Festag, M. Wittenberg and J. H. WendorV, J. Mater. Chem., 1995, 5, 2247. and F. Kuschel, Mol. Cryst. L iq. Cryst., 1990, 193, 115. 50 (a) K. Fontell, K. K. Fox and E. Hansson, Mol. Cryst. L iq. Cryst. 31 (a) G. Lattermann and G. Staufer, Mol. Cryst. L iq. Cryst., 1990, 191, 199; (b) G. Staufer, M. Schellhorn and G. Lattermann, L iq. L ett., 1985, 1, 9; (b) K. Fontell, Colloid Polym. Sci., 1990, 268, 264. 51 A values of 1.3551 was obtained by fluorescence quenching experi- Cryst., 1995, 18, 519; (c) M. Schellhorn and G. Lattermann, L iq. Cryst., 1994, 17, 529; (d) M. Schellhorn and G. Lattermann, ments of normal cubic mesophases of lyotropic systems: L. B.-A. Johansson and O. So�derman, J. Phys. Chem., 1987, 91, 5275; a Macromol. Chem. Phys., 1995, 196, 211. value of 1.551 was found by the same method in poly(oxyethylene) 32 K. Borisch, S. Diele, P. Go� ring, H. Krrske, surfactant water systems: B. Medhage, M. Almgren and L. Alsins, Angew. Chem., 1996, 109, 2188, Angew. Chem., Int. Ed. Engl., 1996, J. Phys. Chem., 1993, 97, 7753. 36, 2087. 52 C. Tschierske, Prog. Polym. Sci., 1996, 21, 775 and references cited. 33 K. Praefcke, B. Kohne, A. Eckert and J. Hempel, Z. Naturforsch., 53 (a) L. Claisen and O. Eisleb, Ann. Chem., 1913, 401, 29; T eil B, 1990, 45, 1084. (b) D. Coates and G. W. Gray, J. Chem. Soc., Perkin T rans 2, 34 S. Fischer, H. Fischer, S. Diele, G. Pelzl, K. Jankowski, R. R. 1976, 863. Schmidt and V. Vill, L iq. Cryst., 1994, 17, 855. 54 R. Dunkel, M. Hahn, K. Borisch, B. Neumann, H.-H. Ru� ttinger 35 K. Borisch, S. Diele, P. Go� ring and C. Tschierske, Chem. Commun., and C. Tschierske, L iq. Cryst., in the press. 1996, 237. 55 V. Percec, D. Tomazos, J. Heck, H. Blackwell and G. Ungar, 36 K. Borisch, S. Diele, P. Go�ring and C. Tschierske, L iq. Cryst., 1997, J. Chem. Soc., Perkin T rans 2, 1994, 31. 22, 427. 56 H. Meier, E. Praß, G. Zerban and F. Kosteyn, Z. Naturforsch., T eil 37 V. S. K. Balagurusamy, G. Ungar, V. Percec and G. Johansson, B, 1998, 43, 889. J. Am. Chem. Soc., 1997, 119, 1539. 38 C. Hall, G. J. T. Tiddy and B. Pfannemu� ller, L iq. Cryst., 1991, 9, 527. Paper 7/05359B; Received 24th July 1997 J. Mater. Chem., 1998, 8(3), 529&ndash

ISSN:0959-9428    

DOI:10.1039/a705359b

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials Self-assembled monolayers of n-hexanethiol and 6-[2¾,5¾-di(2- thienyl )pyrrol-1¾-yl]hexanethiol on polycrystalline nickel substrates† Z. Mekhalif,a A. Lazarescu,b L. Hevesi,b J-J. Pireauxa and J. Delhalle*c aL aboratoire Interdisciplinaire de Spectroscopie Electronique bL aboratoire de Chimie Organique and cL aboratoire de Chimie T he� orique des Surfaces et des Interfaces, Faculte�s Universitaires Notre- Dame de la Paix 61, rue de Bruxelles, B5000, Namur, Belgium It is shown in this work that n-hexanethiol and 6-[2¾,5¾-di(2-thienyl)pyrrol-1¾-yl]hexanethiol form stable monolayers on electrochemically reduced polycrystalline nickel substrates, but not on the oxidised ones.Both thiols, when directly adsorbed on the nickel surfaces without electrochemical pretreatment, show significant proportions of oxidised sulfur species (sulfinates and sulfonates) in their XPS spectra.In contrast, the electrochemical pretreatment leads to monolayers exhibiting S 2p levels essentially characteristic of sulfur atoms bound to metallic sites and corroborating the good chemical stability and resistance to electrochemical oxidation of such monolayers.but could be removed in their doped state.12 In the two-step approach, films are adherent to the substrate in both the Introduction electrochemically compensated and doped states.2 We suspect that, in the one-step process, TPT-C3SH undergoes ready Conducting polymers such as polythiophene are often synelectropolymerisation and that adhesion results from the bind- thesised by electrochemical oxidative coupling of the heteroing with the metallic substrate of the available alkylthiol cyclic moieties of the monomers. Direct electropolymerisation pendant groups of the electrodeposited polymer in the interfa- at a metal anode is of special interest from the point of view cial region.In the approach where the electrode is chemically of surface coatings with organic films.First, the conductivity modified with phenyl- or pyrrolyl-alkanethiols, the thiol of the growing layer allows an easy control of the film thickness, strongly binds to the metallic substrate and attractive inter- which is in contrast to similar processes where intrinsically actions, possibly including covalent bonds, are established insulating polymers are electrochemically deposited.In that between the phenyl or pyrrolyl groups at the first stages of the respect, the method has many features in common with galvanopolymer film growth, hence the resulting adhesion. techniques, the main diVerence being in the electrode bias. Our main goal is to construct compact and organised Second, possible applications are many and promising: they monolayers on oxidisable metals that can sustain the elec- range from protective layers to electronic devices.1 trodeposition of conductive polymers and yield electroactive In most cases, electrodeposition of conducting polymers films with good morphological, adherence and electronic (aniline, pyrrole, etc.) is carried out on inert electrodes and properties.A diYculty with the otherwise attractive approach results in the precipitation of the growing polymer to form, proposed by Kowalik et al. is the need to graft alkanethiol upon the electrode surface, an insoluble layer, mechanically chains on the polymerisable monomers. This can have adverse attached, but easily peeled oV. To circumvent the problems consequences such as complexity and cost of the synthesis of raised by such weakly bound coatings, a two-step procedure the monomers,13 diminished conductivity12,14 and quality in has been proposed by which self-assembled monolayers (SAM) the morphological and electronic properties of the polymer of selected thiols, e.g.n-alkanethiol, phenyl- and pyrrolyl- films. In the two-step approach, the choice of the monomer is alkanethiols, are formed on the metal substrate and then in principle independent of the nature of the SAM.However, suitable monomers are electrodeposited on the modified electhe electrochemical properties of the modified electrode should trodes.2–5 The approach exploits the aYnity of the thiol group, allow electrodeposition without oxidation of the metal MSH, for noble (Au, Pt)6 as well as several oxidisable (Cu, Fe, substrate and promote strong (ideally covalent) attractive Ni, etc.)2,7–11 metals.For instance, very good adhesion is interactions between the organic monolayer and the elec- observed for the films electrodeposited on the phenyl- and trodeposited polymer. Accordingly, it is reasonable to assume pyrrolyl-alkanethiol monolayers, but not on the alkanethiols.that, the lower the oxidation potential of the end groups at In 1993, Kowalik et al. reported a one-step oxidative polymeristhe monolayer surface, the better the chances to promote ation directly from a solution of 6-[2¾,5¾-di(2-thienyl)pyrrolelectrodeposition of conductive polymers. In that respect, TPT- 1¾-yl]propanethiol (TPT-C3SH) on Pt and Au electrodes to C3SH, with a reported oxidation potential of +0.89 V vs.produce films exhibiting remarkable stability and toughness.12 Ag/Ag+ measured12 by cyclic voltammetry in acetonitrile Common to both approaches is the fact that adherence propersolution using EtN4BF4 (0.1 M) is certainly a good candidate. ties of the electrodeposited films are observed for alkanethiol The corresponding value for pyrrole in the same conditions chains terminated with reactive groups such as phenyl, pyrrolyl would be +0.91 V (adapted from ref. 15). By its structure, and 2,5-di(2¾-thienyl )pyrrol-1-yl (TPT). In the one-step proanalogous to TPT-C6SH shown in Fig. 1, TPT-C3SH also cedure, thin films of TPT-C3SH on platinum could not be provides two possible sites, a and b, for the oxidative coupling detached when in their electrochemically compensated state, to take place. In this contribution we study the formation and properties of n-hexanethiol and TPT-C6SH monolayers (Fig. 1) adsorbed † Presented at the Third International Conference on Materials on polycrystalline nickel substrates, either oxidised (Ni-ox) or Chemistry, MC3, University of Exeter, Exeter, 21–25 July 1997.* E-mail: joseph.delhalle@fundp.ac.de electrochemically reduced (Ni-red). Both n-hexanethiol (C6SH) J. Mater. Chem., 1998, 8(3), 545–551 545CH3 S Ni-ox or Ni-red ( a ) (b ) S Ni-ox or Ni-red S S a b a b N Fig. 1 Schematic description of (a) n-hexanethiol (C6SH) and (b) 6- [2¾,5¾-di(2-thienyl )pyrrol-1¾-yl]hexanethiol (TPT-C6SH) molecules on the polycrystalline surfaces, Ni-ox or Ni-red and TPT-C6SH include the same alkanethiol chain, M(CH2)6MSH; hence the properties of their chemisorbed monolayers are reported comparatively to identify and account for possible influences of the bulky TPT group.Like other common oxidisable metals (Fe, Cu, etc.), Ni is an interesting substrate for technological applications. For instance, Ni can be used as the final coating in separable electrical contacts on which growth on the metal of poorly conducting layers (oxides, contaminants) due to environmental conditions has up to now limited its use to most common S O S O NHMe2Cl 4 3 i ii S O NMe2 5 S O 6 S O iii iv S N S C6H12OH 7 80% + S N S C6H12OAc 8 v S N S C6H12OMs 9 88% vi S N S C6H12SH 2 51% vii–ix Scheme 1 Reagents and conditions: i, (CH2 O)n, Me2 NH HCl, EtOH, applications.We have recently shown that excellent tribological conc.HCl, reflux, 40 h; ii, aq. NH4OH; iii, thiophene-2-carbaldehyde, behaviour can be obtained when electrochemically reduced NaCN, DMF, room temp.; iv, H2NC6H12OH, C6H6, AcOH, reflux, polycrystalline Ni substrates are modified with n-dodecanethiol 50 h; v, EtOH, aq. NaOH, reflux, 4 h; vi, MsCl, DMAP, pyridine, molecules.16 CH2Cl2; vii, thiourea, reflux; viii, NaOH, reflux; ix, HCl TPT-C6SH diVers from TPT-C3SH12 by the presence of three additional methylenes in the alkanethiol pendant chain.yield, dH (CDCl3): 3.4 (s, 4H), 7.14 (m, 2H), 7.64 (m, 2H), 7.81 This slight diVerence in the structural feature has been selected (m, 2H). as a compromise between expected improvedecular organisation in the monolayer, known to be promoted by longer Synthesis of alcohol 7.Diketone 6 (2.5 g, 10 mmol), 6- alkyl chains,17 and the possibility for future electrodeposition aminohexan-1-ol (1.2 g, 10.5 mmol) and acetic acid (12 ml ) of conductive polymers to proceed. Furthermore, there are were refluxed in 25 ml of benzene for 50 h. At this point the clues that electrodeposition of conductive polymer films on reaction had reached only 50% conversion.On cooling the organised monolayers leads to chains with longer conjugation mixture to room temp., the unreacted diketone crystallised out pathways.18 Finally, the delamination rate of coatings is deterand was recovered by filtration. The solution was concentrated mined not by the thickness of the coating, but by the bonds by evaporation, the residue taken up with diethyl ether, washed that prevail at the interface as well as the packing and the successively with aqueous sodium hydrogen carbonate, then molecular organisation in the monolayer,9 hence the use of a with water, dried over MgSO4 and evaporated.The crude C6 alkyl chain in the present study.product (1.91 g) thus obtained contained the desired alcohol 7 and its acetate. For hydrolysis of the latter, the mixture was Experimental dissolved in ethanol (20 ml ) containing 1 g of NaOH and 2 ml of water and refluxed for 4 h. Conventional work-up of this 6-[2¾,5¾-di(2-thienyl )pyrrol-1¾-yl]hexanethiol 2 (TPT-C6SH) reaction mixture followed by column chromatographic purifi- (Scheme 1) cation (SiO2, eluent pentane–diethyl ether, 50550 v/v) led to 1.33 g of pure 7 (80% yield based on consumed 6).Analysis Synthesis of 2-acetylthiophene 3. A solution of tin tetrachloride (5.22 g, 20 mmol) and freshly distilled acetyl chloride for C18H21NOS2, calc. C: 65.21, H: 6.39; found C: 64.24, H: 6.26%. n/cm-1 ( liq. film): 3356, 3102, 3071, 2931, 2857, 1461, (1.57 g, 20 mmol) in 10 ml of dry dichloromethane was added dropwise at room temp over 0.5 h to a dichloromethane 1404, 1052, 843, 767, 730, 697.dH (CDCl3): 1.18 (m, 4H), 1.40 (m, 2H), 1.50 ( bs, 1H), 1.58 (m, 2H), 3.52 (t, J 7 Hz, 2H), 4.17 (10 ml ) solution of thiophene (1.68 g, 20 mmol) under argon atmosphere. The mixture was stirred for an additional 30 min. (m, 2H), 6.33 (s, 2H), 7.05 (m, 2H), 7.32 (m, 2H).dC (CDCl3): 24.9, 26.0, 30.9, 32.3, 44.9, 62.4, 110.7, 125.1, 125.8, 127.2, Water (20 ml ) and diethyl ether (100 ml) were added, the organic phase was washed once more with water and aqueous 128.4, 135. hydrogen carbonate, dried over MgSO4 and evaporated. The essentially pure crude product (2.42 g, 96% yield) was used Synthesis of mesylate 9. The alcohol 7 (1.324 g, 4 mmol), methanesulfonyl (mesyl) chloride (0.58 g, 5 mmol), 4-dimethyl- without purification.n/cm-1 ( liq. film): 3089, 3001, 2922, 1662, 1517, 1414, 1355, 1273, 725. dH (CDCl3) 2.55 (s, 3H), 7.12 (m, aminopyridine (DMAP, 0.61 g, 5 mmol) and pyridine (1.28 g, 16 mmol) were dissolved in 15 ml of dry dichloromethane at 1H), 7.67 (m, 2H). 0 °C and stirred for 6 h at room temp.The mixture was then acidified with 10% HCl, extracted with diethyl ether, and the Synthesis of compounds 4, 5 and 6. These compounds were prepared according to literature procedures:19 4, mp ethereal layer washed with water, dried and evaporated to give 1.77 g of crude product. Purification by column chromatogra- 179–180 °C, 72% yield; 6, oV-white solid, mp 129–131 °C, 71% 546 J.Mater. Chem., 1998, 8(3), 545–551phy (SiO2, eluent pentane–diethyl ether, 50550 v/v) led to 1.43 g (88% yield) of 9. dH (CDCl3) 1.1–1.25 (m 4H), 1.5–1.63 (m, 4H), 2.93 (s, 3H), 4.05–4.20 (m, 4H), 6.32 (s, 2H), 7.05 (m, 4H), 7.31 (m, 2H). Synthesis of thiol 2 (TPT-C6SH). The mesylate 9 (1.43 g, 3.5 mmol) and thiourea (0.38 g, 5 mmol) were refluxed in 10 ml of degassed ethanol under argon for 16 h (disappearance of 9).Concentrated aqueous sodium hydroxide (5 ml) was added and the mixture refluxed for another 10 h. After cooling to room temp., the mixture was acidified with 10% HCl, extracted with diethyl ether and the ethereal layer washed with water, dried and evaporated. Column chromatographic purification of the crude product (SiO2, eluent pentane–diethyl ether, 951 v/v) led to 0.63 g (51% yield) of thiol 2 (TPT-C6SH) as a yellow–green fluorescing viscous liquid which solidified on standing.Analysis for C18H21NS3: calc. C: 62.20, H: 6.09, found: C: 62.09, H: 5.96%. n/cm-1 ( liq. film): 3101, 3069, 2929, 2854, 2565, 1461, 1402, 843, 766, 695. dH (CDCl3) 1.10–1.32 (m, 5H), 1.42–1.60 (m, 4H), 2.42 (q, J 7 Hz, 2H), 4.17 (t, J 7 Hz), 2H), 6.35 (s, 2H), 7.06 (m, 4H), 7.32 (m, 2H).dC (CDCl3) 24.3, 25.6, 27.5, 30.8, 33.5, 44.8, 110.7, 125.2, 127.2, 128.2, 134. Chemical modification of polycrystalline nickel surfaces with C6SH and TPT-C6SH Chemicals. n-Hexanethiol (Acros, 98%, 21527–0050), HClO4 p.a. (Acros, 22. 331. 21), acetone (Aldrich, 99.9%, HPLC grade 27, 072-5), H2O, ultrapure water (18 MV cm), absolute ethanol p.a.(Merck, 1.000983.2500), NaOH Suprapur (Merck, 017 C762966) were used without additional purification. Metal substrates and monolayer preparation. Disk-shaped metal substrates (thickness 6 mm) were cut from polycrystalline Ni rods (Ø=6.35 mm, Aldrich, 99.99%, 26,707-4) and mechanically polished using various grit diamond pastes down to 1 mm. Before use for monolayer adsorption, these substrates were rinsed with copious amounts of acetone.Two types of metallic substrates diVering in their surface chemical state were compared. The first type (Ni-ox) consists of nickel surfaces mirror polished, cleaned as indicated above and used directly for the monolayer adsorption. The second type (Ni-red) concerns nickel surfaces, polished, cleaned and electrochemically Fig. 2 XPS spectra of the (a) Ni 2p and (b) O 1s core levels of a bare reduced prior to thiol adsorption. The electrochemical pretreat- Ni-ox polycrystalline surface ment consists of a 20 min reduction in an aqueous solution of HClO4 (1 M) of the nickel substrates at -0.7 V vs. a saturated Associated with Ni(OH)2 a shake-up structure around 861.4 eV calomel electrode (SCE).also occurs. The Ni 2p3/2 and Ni 2p1/2 lines characteristic of The monolayers were formed by immersion of the nickel metallic nickel are found at 852.7 and 870.0 eV, respectively.20 substrates in 1×10-3 M solutions of C6SH or TPT-C6SH in The shape of the spectrum and the line positions in Fig. 2(a) hexane for 12 h. The samples were then quenched with hexane point to a Ni-ox surface composed of a mixture of metallic and ultrasonically cleaned for 15 min in hexane to remove the and oxidised nickel, Ni(OH)2, atoms.Three features are also excess of thiol molecules physisorbed on the monolayer surnoted in the O 1s line; they arise at 529.4, 531.6 and 532.7 eV. faces. The samples were finally dried under an argon flow and They nicely correspond to values reported for slightly hydrated used immediately for characterisation.Ni(OH)220 where three components are also found; they are assigned to ionic oxygen (529.0 eV), Ni(OH)2 and C with O Monolayer characterisation (530.5 eV), and finally H2O and O with C (531.8 eV), respectively. 9,20 The exposure time to the atmosphere necessary to X-Ray photoelectron spectroscopy (XPS) is mainly used here to control the elemental composition of the monolayer and transfer the Ni-red surfaces from the electrochemical cell to the spectrometer and mount the samples in the analysis identify the oxidation states of the nickel and sulfur atoms located at the top layers.The spectra were recorded with an chamber is too long (~30 min) to maintain the surface nickel atoms in their reduced state, hence the corresponding XPS SSX-100 spectrometer using monochromatised Al-Ka radiation (1486.6 eV).All reported spectra, except the one shown spectra resemble those of Ni-ox substrates and are not shown in this paper. in Fig. 6(a) (9 weeks), were recorded using identical conditions. The samples were analysed at 35° take oV angle, which In the case of the modified nickel surfaces with the thiol molecules, the Ni 2p, O 1s and S 2p core levels were calibrated corresponds to a sampled thickness of approximately 4 nm.Fig. 2(a) and (b) show the XPS spectra of the Ni 2p and O by reference to the recorded C 1s peak characteristic of the alkyl moiety of the compounds and conventionally set at 1s regions of a representative Ni-ox surface. From published literature on nickel hydroxide, b-Ni(OH)2 and Ni(OH)2 284.5 eV, which is a commonly used procedure for organic films.9,21 In the case of the hexanethiol chain, the C 1s line slightly hydrated,20 it is known that the Ni 2p3/2 and Ni 2p1/2 lines are located around 855.8 and 873.4 eV, respectively.mainly consists of carbon atoms linked to hydrogen atoms J. Mater. Chem., 1998, 8(3), 545–551 547and emerges as a well-defined and intense structure.With a calibration performed with respect to that intense structure, the energy at which the S 2p3/2 signal arises for the free thiols, disulfides, thiolates, sulfinates and sulfonates is equal to 163.3, 163.3, 161.8, 165.5 and 168.0 eV, respectively.9 In the case of TPT-C6SH, the sulfur atoms from the two thiophenes contribute to the S 2p region; the S 2p3/2 component for thiophene is located at 164.3 eV.20 The S 2p line is naturally a doublet structure, (S 2p3/2) and (S 2p1/2), where the spacing between the components is equal to 1.13 eV and the theoretical intensity ratio (S 2p3/2)/(S 2p1/2) is equal to 2/1.Deconvolution of the S 2p lines has been carried out assuming a doublet structure and the theoretical intensity ratio 2/1 between the two components.Electrochemical characterisation of nickel with and without monolayers was carried out in a conventional three-electrode cell. A calomel electrode saturated with KCl was used as reference electrode, a platinum grid as counter-electrode and the nickel disks as working electrodes. The cell was connected to a TACUSSEL PJT 24 potentiostat/galvanostat linked to an IMT1 interface.Characterisation was performed in a 0.1 M NaOH aqueous solution (deoxygenated by argon bubbling for 30 min prior to measurements) with a potential scan ranging from -0.3 to +0.6 V at 10 mV s-1. Voltammograms were digitally recorded using the VOLTMASTER program. Results and Discussion This section is organised in two main parts.In the first one we focus our attention on the XPS characteristics of the C6SH and TPT-C6SH monolayers freshly formed on both Ni-ox and Ni-red surfaces. The second part concerns the stability of such monolayers under ambient conditions for periods of 9 weeks and response to cyclic voltammetry in a rather aggressive medium. XPS characteristics of the monolayers Fig. 3 shows the Ni 2p, O 1s and S 2p core level regions of the monolayers of C6SH chemisorbed on Ni-ox and Ni-red substrates recorded 30 min after the dipping step. Significant diVerences are observed between the Ni-ox and Ni-red samples. As in the case of the reference Ni-ox surface [Fig. 2(a)], the Ni 2p spectrum of the modified Ni-ox substrate [Fig. 3(a)] exhibits the spectral features of both oxidised, Ni(OH)2, and metallic nickel atoms with comparable intensities.The situation is totally diVerent in the case of the C6SH monolayer adsorbed on a Ni-red substrate for which only the characteristic features of metallic nickel are detected in the Ni 2p spectrum [Fig. 3(a)]. This means that reduction in the electrochemical cell immediately followed by immersion in the C6SH solution keeps the nickel atoms under the monolayer in their metallic state.This is corroborated by the spectroscopic characteristics of the O 1s [Fig. 3(b)] and S 2p [Fig. 3(c)] lines. In the case of C6SH on Ni-ox, the intense O 1s line exhibits the three peaks already observed for the reference Ni-ox surface [Fig. 2(b)]. The presence of these lines at 529.4, 531.6 and 533.3 eV indicates the presence of Ni(OH)2, H2O and possibly oxygen containing carbon contaminants in amounts comparable to what was observed on the Ni-ox reference substrate [Fig. 2(b)]. This is in contrast to the Ni-red case [Fig. 3(b)] for which a low Fig. 3 XPS spectra of the (a) Ni 2p, (b) O 1s and (c) S 2p core levels intensity O 1s line appears as a single peak centred at of (i) Ni-ox and (ii) Ni-red polycrystalline surfaces modified with n- 532.2 eV, which is more typical of oxygen in the form of hexanethiol (C6SH) hydroxy (MOH).9,20 In the wide scan spectra, not reported here but very similar to those reported in figure 1 of ref. 4 for 2p line is composed of two large features of very low intensity: n-dodecanethiol, the O 1s peak intensity was also substantially one centred around 162.9 eV and the other around 169.3 eV.lower for Ni-red than Ni-ox surfaces. Very instructive is the The first one corresponds to the doublet structure (S 2p3/2) inspection of the S 2p spectra shown in Fig. 3(c) for C6SH at 161.8 and (S 2p1/2) at 163.0 eV as observed in the case of thiols chemisorbed on gold substrates in the form of thiolates. monolayers on Ni-ox and Ni-red.In the case of Ni-ox, the S 548 J. Mater. Chem., 1998, 8(3), 545–551Fig. 5 XPS spectra of the (a) Ni 2p and (b) S 2p core levels of Ni-red polycrystalline surfaces modified with n-hexanethiol (C6SH), (i ) fresh and (ii) after 9 weeks of exposure to atmospheric oxygen able to that of the thiolates and disulfides. A diVerent situation prevails for the monolayer adsorbed on the Ni-red surface.First, the group at 169.3 eV characteristic of the oxidised sulfur species is not detected, which correlates well with the already mentioned low intensity of the O 1s line. The XPS data point to the fact that the electrochemical treatment, as in the case of n-dodecanethiol,4 has reduced the amount of nickel oxides to the point that the oxidised species cannot be detected. The S 2p doublet is also closer to the expected theoretical ratio 2/1; the small deviation is likely due to small amounts of disulfides as suggested by the deconvolution in the figure.A similar analysis can be carried out on the Ni 2p, O 1s and S 2p core level regions of the TPT-C6SH chemisorbed monolayers on Ni-ox and Ni-red substrates shown in Fig. 4. A comparison between Fig. 3 and 4 reveals a close similarity Fig. 4 XPS spectra of the (a) Ni 2p, (b) O 1s and (c) S 2p core levels in the XPS features. The only diVerences to be noted are in of (i ) Ni-ox and (ii ) Ni-red polycrystalline surfaces modified with 6- the shape of the S 2p levels due to the thiophene sulfur [2¾,5¾-di(2-thienyl )pyrrol-1¾-yl]hexanethiol (TPT-C6SH) contribution, on the one hand, and the presence of traces of oxidised species in the TPT-C6SH monolayer on Ni-red The doublet structure is not very apparent because of the low (Fig. 4), on the other hand. intensity and poor quality of the spectrum in that region, XPS analysis of the freshly prepared monolayers of C6SH which also indicates very low chemisorption on the substrate. and TPT-C6SH on Ni-ox and Ni-red point to similar properties The group centred at 169.3 eV suggests that the amount of oxidised sulfur species (sulfinates and sulfonates) is compar- and suggest that the TPT terminal group in TPT-C6SH does J.Mater. Chem., 1998, 8(3), 545–551 549not interfere with a proper chemisorption of the molecule. The the increase in the intensity of the Ni(OH)2 features in the Ni 2p spectra is more pronounced at comparable exposures to low S 2p line intensities of Ni-ox substrates modified with either C6SH or TPT-C6SH indicate very low amounts of the atmosphere for TPT-C6SH [Fig. 6(a)] than C6SH [Fig. 5(a)]. With the S 2p spectra [Fig. 5(b) and 6(b)] it is chemisorbed species present on these substrates. possible to see a qualitative diVerence in the modification of the chemical state of the sulfur atoms bound to the surface.Stability of the monolayer After 9 weeks the amount of oxidised sulfur species (sulfinates Since the modified Ni-ox substrates are of poor quality, in this and sulfonates) increases in both C6SH and TPT-C6SH monopart we assess only the stability of the monolayers on the sole layers. However, a larger proportion of thiolates remains in Ni-red surfaces. The first test is on the barrier properties of the case of C6SH, while they have practically disappeared from the monolayer against atmospheric oxygen permeation.If the S 2p spectrum of the TPT-C6SH monolayer (9 weeks). oxygen reaches the metal surface, oxidation of the metal occurs, This diVerent behaviour suggests that TPT-C6SH molecules which in turn leads to oxidation of the sulfur species bound form less densely packed monolayers than C6SH.It is interesto the substrate. The second test is on the electrochemical ting to note the increase in disulfide species with increasing stability and blocking eYciency of the modified electrode. oxidation. This suggests that oxidation gradually evolves from the thiolates to the highly oxidised species (sulfinates and Resistance to atmospheric oxygen permeation.In Fig. 5 and sulfonates) via the formation of disulfides. At present, however, 6 are shown the Ni 2p and S 2p levels of C6SH and TPTthis is only a conjecture based on the present observations C6SH monolayers, respectively, recorded 30 min (fresh) and 9 which corroborate similar trends noted in the case of the Ni weeks after preparation.The shape of the O 1s line being surfaces.11 essentially the same for both samples after 30 min and 9 weeks, the corresponding spectra have been omitted. Both the Ni 2p Electrochemical stability. Here we report preliminary tests and S 2p spectra indicate that the C6SH monolayers act as on the blocking eYciency of the C6SH and TPT-C6SH monobetter barriers to atmospheric oxygen than TPT-C6SH.First, layers chemisorbed on Ni-red. The results are shown in Fig. 7(a) Fig. 6 XPS spectra of the (a) Ni 2p and (b) S 2p core levels of Ni-red Fig. 7 Cyclic voltammogram of (i) polycrystalline nickel and (ii ) Nired substrates modified with (a) n-hexanethiol (C6SH) and (b) 6-[2¾,5¾- polycrystalline surfaces modified with 6-[2¾,5¾-di(2-thienyl)pyrrol-1¾- yl]hexanethiol (TPT-C6SH) (i ) fresh and (ii) after 9 weeks of exposure di(2-thienyl )pyrrol-1¾-yl]hexanethiol (TPT-C6SH) (0.1 M NaOH, argon bubbling) atmospheric oxygen 550 J.Mater. Chem., 1998, 8(3), 545–551P. C. Lacaze, Surface Interface Anal., 1993, 20, 749, Handbook of and (b), respectively. For comparison, the cyclic voltammogram Conducting Polymers, ed.T. A. Skotheim, Marcel Dekker, New of a clean nickel substrate (solid line) is superimposed on those York, 1986. vol. 1 and 2; ed. W. R. Salaneck, D. T. Clark and of the modified substrates (dotted lines). During the anodic E. J. Saumelsen, Science and Applications of Conducting Polymers, potential sweep, the clean polycrystalline nickel electrode Adam Hilger, Bristol, 1991; Organic Materials for Electronics, ed.shows a peak at 0.355 V assigned to the oxidation of the metal J. L. Bre�das, W. R. Salaneck and G. Wegner, North-Holland, Amsterdam, 1994; Handbook of Organic Conductive Molecules and and the formation of a passive nickel oxide layer. The high Properties, ed. Hari Singh Nalwa, Wiley, Chichester, 1997, current rise starting at +0.52 V is due to oxygen evolution, vol. 2–4. while the cathodic peak at +0.234 V corresponds to the 2 P. Lang, Z. Mekhalif and F. Garnier, J. Chim. Phys., 1992, 89, reduction of the oxide layer. Fig. 7(a) shows the voltammogram 1063; P. Lang, Z. Mekhalif and F. Garnier, V ide, Couches Minces, (dotted line) corresponding to a Ni-red substrate modified with 1993, 368, 1021; P.Lang, Z. Mekhalif and F. Garnier, Adhesion, C6SH. By comparison with the clean nickel surface, the modi- 1994, 402; Z. Mekhalif, P. Lang and F. Garnier, J. Chim. Phys., 1995; 92, 831; Z. Mekhalif, F. Garnier, P. Lang, R. Caudano and fied electrode exhibits a substantial decrease in its electrochemi- J. Delhalle, J. Electrochem. Soc., submitted for publication. cal activity. A similar conclusion holds in the case of a TPT- 3 R.J. Willicut and R. L. McCarley, J. Am. Chem. Soc., 1994, 116, C6SH monolayer chemisorbed on a Ni-red substrate 10824; R. J. Willicut and R. L. McCarley, L angmuir, 1994, 11, 296; [Fig. 7(b)]. This is in line with the previous observations on R. J. Willicut and R. L. McCarley, Adv.Mater., 1995, 7, 759. the stability of the monolayer to ambient atmosphere.It 4 C. N. Sayre and D. M. Collard, L angmuir, 1995, 11, 296; constitutes additional evidence that the freshly formed organic C. N. Sayre and D. M. Collard, L angmuir, 1997, 13, 714. 5 D. B.Wurn, S. T. Brittain and Y.-T. Kim, L angmuir, 1996, 12, 3756. layer is firmly and densely chemisorbed on the surface to the 6 A. T. Hubbard, Chem. Rev., 1988, 88, 633; R.G. Nuzzo, point that water and electrolytic species are prevented from L. H. Dubois and D. Allara, J. Am. Chem. Soc., 1990, 112, 558; reaching the metal thiol interface. C. D. Bain, E. B. Troughton, Y.-T. Tao, J. Evall, G. M. Whitesides The cyclic voltammetry studies which were carried out on and R. G. Nuzzo, J. Am. Chem. Soc., 1989, 111, 121. freshly prepared monolayers do not reveal substantial diVer- 7 P.E. Laibinis, G. M. Whitesides, D. L. Allara, Y. T. Tao, ences in the electrochemical behaviour of C6SH and TPT- A. N. Parikh and R. G. Nuzzo, J. Am. Chem. Soc., 1991, 111, 321; P. E. Laibinis and G. M. Whitesides, J. Am. Chem. Soc., 1992, C6SH as was observed in the XPS spectra of samples exposed 114, 9022. to atmospheric oxygen for 9 weeks. In the future, it would be 8 M.Itoh, H. Nishihara and K. Aramaki, J. Electrochem. Soc., 1994, interesting to carry out detailed electrochemical characteris- 141, 2018; M. Itoh, H. Nishihara and K. Aramaki, J. Electrochem. ations on such layers. Soc., 1995, 142, 1839. 9 M. Stratmann, Adv.Mater., 1990, 2, 191; M. Volmer, M. Stratmann and H. Viefhaus, Surf. Interface Anal., 1990, 16, 278; M.Stratmann, Conclusion W. Fu� rbeth, G. Grundheimer, R. Lo�sch and C. R. Reinartz, Corrosion Mechanisms in T heory and Practice, ed. P. Marcus and This study show that n-hexanethiol, C6SH, and 6-[2¾,5¾-di(2- J. Oudar, Marcel Dekker, New York, 1995. thienyl)pyrrol-1¾-yl]-hexanethiol, TPT-C6SH, do chemisorb on 10 A. D. Vogt, T. Han and T. P. Beebe, L angmuir, 1997, 13, 3397. reduced polycrystalline nickel, a technologically important 11 Z.Mekhalif, J. Riga, J.-J. Pireaux and J. Delhalle, L angmuir, 1997, metal, and form stable self-assembled layers, much like n- 13, 2285. 12 J. Kowalik, L. Tolbert, Y. Ding, L. A. Bottomley, K. Vogt and dodecanethiol.11 P. Kohl, Synth. Metals, 1993, 55–57, 1171. Preliminary results indicate that electropolymerisation of 13 J. P.Ferraris and G. D. Skiles, Polymer, 1977, 28, 179. 2,5-di(2¾-thienyl)-N-methylpyrrole on the TPT-C6SH mono- 14 A. F. Diaz, J. Castillo, K. K. Kanazawa, J. A. Logan, M. Salmon layers on polycrystalline Ni-red substrates leads to strongly and O. Fajardo, J. Electroanal. Chem., 1982, 133, 233. adherent films.22 More systematic studies of the chemisorption 15 A. F. Diaz, K. K. Kanazawa and G. P. Gordini, J. Chem. Soc., parameters (time, concentration, solvents, nature of the metal), Chem. Commun., 1979, 645. 16 Z. Mekhalif, J. Delhalle, J. J. Pireaux, S. Noe�l, F. Houze� and stability of the monolayers and electrodeposition of conductive L. Boyer, J. Coat. Surf. T echnol., in the press. polymers will be necessary to elaborate adherent electroactive 17 C. D. Bain and G. M. Whitesides, J. Am. Chem. Soc., 1989, 111, 321. films on oxidisable metals with good morphological, adherence 18 Z. Mekhalif, P. Lang and F. Garnier, J. Electroanal. Chem., 1995, and electronic properties. 399, 61. 19 H. Wynberg and J. Matselaar, Synth. Commun., 1984, 14, 2221. 20 J. F. Moulder, W. F. Stickle, P. E. Sobol and K. D. Bomben, This work was funded by the Re�gion Wallonne (Project Handbook of X-ray Photoelectron Spectroscopy, Perkin-Elmer No 3119). Corporation, Eden Prairie, MN, 1992, pp. 84–85; A. N. Mansour, Surf. Sci. Spectra, 1996, 3, 211. 21 G. Beamson and D. Briggs, High Resolution XPS of Organic References Polymers. T he Scienta ESCA300 Database Wiley, Chichester, 1992. 22 Z. Mekhalif et al., work in progress. 1 C. A. Ferreira, S. Aeiyach, M. Delamar and P. C. Lacaze, J. Electroanal. Chem., 1990, 352, 284; S. Aeiyach, A. Kone, M. Dieng, J. J. Aaron and P. C. Lacaze, J. Chem. Soc., Chem. Commun., 1991, 822; F. Beck and R. Michael T echnol., 1992, 59, 64; C. A. Ferreira, S. Aeiyach, M. Delamar and Paper 7/07441G; Received 15th October, 1997 J. Mater. Chem., 1998, 8(3), 545–551 551

ISSN:0959-9428    

DOI:10.1039/a707441g

出版商:RSC    

年代:1998

数据来源: RSC