摘要:

J O U R N A L O F C H E M I S T R Y Materials Feature Article High resolution electron microscopy in materials research Gustaaf Van Tendeloo EMAT , University of Antwerp, Groenenborgerlaan 171, B-2020 Antwerpen, Belgium The historical evolution and the future aspects of high resolution electron microscopy are covered briefly. As for the applications for materials science and inorganic chemistry, we are positive about the future of electron microscopy, particularly when combined with other techniques.The strong interaction between sample and electron beam provides chemical, structural and electronic information through the emission of X-rays, inelastically scattered electrons, light, etc.We try to illustrate the strong points of high resolution electron microscopy with a few examples in diVerent fields of solid state chemistry, including high Tc superconducting materials, colossal magnetoresistant materials and ionic conductors. The tendency of modern technology towards reduced dimensionality increases the importance of electron microscopy as a quality control device.Transmission electron microscopy (TEM)? Introduction The idea of this introduction is not to provide a course in Electron microscopy started with Ruska and Knoll who built electron microscopy; for those interested we refer to specialised their first electron microscope in 1931.The magnification was books such as the Handbook of Microscopy1 or Electron only 17 times, but it showed that an enlarged image could be DiVraction T echniques.2 Our aim is rather to provide the basics obtained using accelerated electrons as information carriers.of the technique and illustrate its possibilities as well as its By 1933 Ruska obtained a resolution of 50 nm, which was an limitations for practical problems. order of magnitude better than the resolution of the optical The working principle of the electron microscope is closely microscope. Its practical use, however, was not evident, since related to that of the optical microscope; it is represented the accelerated electrons burnt any living material, which schematically in Fig. 1. Assume a plane wave of coherently actually was the main purpose for the development. After 1945 accelerated electrons is incident on the sample under investi- materials scientists in particular realised the possible impact gation.Within the sample, such a plane wave is no longer an of electron microscopy and the instrument was further develeigenfunction of the system and generates diVerent Bloch oped and underwent ‘plastic surgery’. By that time the resowaves. If the sample is very thin, the wavefunction f (x,y) at lution of the instrument had decreased to the nanometer range the exit plane of the object is not dependent on the thickness and in 1956 Menter published the first lattice image of copper of the sample; in practice f (x,y) will be the projected (crystal) phthalocyanine; revealed a direct resolution of 1.2 nm and potential. This exit plane will act as a planar source of spherical showed the presence of crystal defects such as dislocations.waves (a Huygens source).Through an electromagnetic lens From that moment on, the idea of visualising individual atoms (the equivalent of the optical lens) the diVracted beams are no longer remained a dream. In the early seventies Hashimoto and co-workers as well as Ottensmeyer et al. demonstrated a resolution of 2 A ° and the direct observation of heavy atoms such as gold or thorium.Now, at the turn of the century, most commercial electron microscopes have a resolution of around 2 A ° and specially designed or high voltage high resolution microscopes reach the 1 A° level. Atomic resolution or structure resolution can therefore be performed on most inorganic materials. The focus is therefore no longer on the ultimate resolution, but on the application of this resolution in materials science.‘Frontiers of Electron Microscopy in Materials Science’ should now become ‘Frontiers in Materials Science by Electron Microscopy’. Electron microscopy has the enormous advantage of combining real space information down to the atomic scale with reciprocal space information in the diVraction pattern. This diVraction information can be obtained from areas of only a few nanometers in size.This makes electron microscopy a local probe, which is particularly strong at revealing the defect structure or the local structure of nano-scale inclusions. With modern technology entering nano-scale or even atomic-scale dimensions, electron microscopy has acquired a new goal: to guide scientists in ‘tailoring’ new materials. This contribution will briefly outline the basics of electron microscopy and of high resolution electron microscopy in particular.We will then try to illustrate the possibilities of the technique for diVerent problems in solid state chemistry or Fig. 1 Schematic representation of the image formation by the objective lens in an ideal transmission electron microscope materials science. J.Mater. Chem., 1998, 8(4), 797–808 797focused onto the back focal plane of this lens. The diVracted Scherzer resolution limit (ca. 1.7 A ° in this case) intensive computer simulations and retrieval methods are required. For amplitudes of the diVerent beams are given by the Fourier transform of the object function, i.e. F(u,v)=F[ f (x,y)]. If the more details on the retrieval of the projected potential using diVerent methods, we refer to the contribution of Van Dyck object is periodic (i.e.perfectly crystalline), the diVraction pattern will consist of sharp spots; any deviation from perfec- in ref. 1. Finally, I would like to note that the term ‘resolution of the tion will result in scattering away from the sharp Bragg reflections. Because the interaction between the accelerated microscope’ is sometimes confusing: the resolution of the microscope can be expressed in terms of the Scherzer resolution electrons and the material is very strong, electron diVraction is very sensitive to minor and local deviations from perfection.limit (1.7 A ° in Fig. 2) or in terms of the instrumental resolution limit (1.1 A ° in Fig. 2). In a second stage of the imaging process, the back focal plane acts as a set of Huygens sources of spherical waves which again interfere in the image plane (Fig. 1). This is an inverse Pros and cons of TEM Fourier transform, which restores the object function: The fact that the interaction of accelerated electrons with y(x,y)={F[F(u,v)]} matter is very strong has enormous advantages: only very small amounts of material are needed to perform the diVraction The intensity in the image plane is then given by |y(x,y)|2 and for this (ideal) microscope we obtain an enlarged image of the experiments and to obtain high quality images.The drawback of this strong interaction, however, is that the scattering is no projected crystal potential. The resolution limit for this ideal microscope which does longer kinematic and that the classical structure retrieval from diVraction evidence only (as in X-ray diVraction) is very not suVer from any lens aberrations, should be determined by only the sample and the wavelength of the accelerated electrons diYcult.Recently, however, with the installation of high quality CCD cameras on modern microscopes and with advanced (Abbe’s principle).In reality, microscope parameters such as spherical aberration, chromatic aberration or focus determine retrieval methods, this problem can partly be solved. Details of this method and its application to solving the structure of the resolution. Mathematically all these factors can be combined in the so-called ‘transfer function’, exp[ix(u,v)], which is nanometer sized precipitates can be found in the work of Zandbergen et al.3 nothing other than ‘the way reality is distorted by the microscope’.The scattered amplitude in the image plane is then Another problem, related to the strong electron–matter interaction present, is the easy absorption of accelerated elec- calculated as trons; one has to work in vacuum and the samples—or at y(x,y)=F{F(u,v)exp[ix(u,v)]} least the interesting parts—have to be very thin; i.e.of the order of a few hundred nanometers for normal microscopy In Fig. 2 we represent the imaginary part of this transfer function for a modern electron microscope at optimum defocus and below 20 nm for high resolution electron microscopy (HREM). Special sample preparation techniques therefore had of the objective lens.Down to distances of 6 nm-1 (i.e. ca. 1.7 A° in real space) the transfer function is largely negative to be developed, ensuring no damage to or alteration of the material under investigation. A widely used technique in and interpretation is relatively straightforward. For distances between 1.7 A ° and 1.1 A ° the structural details will be highly materials chemistry is the so-called ion milling technique, where argon ions with an energy between 1 and 5 keV are distorted by the fluctuating transfer function; below 1.1 A ° no details are visible.This transfer function is very much a bombarded on to a rotating sample at a glancing angle of between 1° and 15°. This technique is particularly useful when function of the focus of the objective lens and of the exact values of the spherical and chromatic aberration constants.interfaces or grain boundaries are of interest. When only the bulk structure is of concern, the sample can just be crushed in Particularly for the correct interpretation of details below the an agate mortar and the powdered flakes then dispersed on a copper grid, covered with a lacy carbon film. This method holds for a large number of inorganic materials, as long as their anisotropy is not too large.Very anisotropic materials, such as graphite or dichalchogenides, can be prepared by cleavage between Scotch tape; the glue is dissolved in benzene or chloroform. As mentioned before, I would like to underline that the interpretation of HREM images at the atomic level is not always straightforward.When the image is taken at the ideal Scherzer defocus, for very thin regions, with a short periodicity along the electron beam direction, with the electron beam perfectly along a low order zone axis, on a perfectly aligned microscope, the heavy atoms tend to be imaged as black dots and interpretation will not cause too many problems down to the Scherzer resolution limit. All other cases, however, will generally require simultaneous computer image calculations for diVerent defocus and thickness values.The eVect of a number of these deviations from perfection can be simulated by the computer. This is particularly important for a slight misalignment of the crystal or of the microscope; even if the average orientation of the crystal is perfect, local deviations tend to occur in the vicinity of crystal imperfections, incoherent precipitates, etc.Most recent electron microscopes no longer have LaB6 type filaments but field emission sources; these provide not only a more intense beam but also a more coherent beam, which Fig. 2 Transfer function (imaginary part) of a high resolution electron results in a transfer function as in Fig. 2, where the information microscope at optimum defocus; the horizontal dashed lines indicate the noise level limit is pushed further beyond the Scherzer resolution limit. 798 J. Mater. Chem., 1998, 8(4), 797–808Correct quantitative interpretations in this region as a function methods (e.g. in determining the exact oxygen content) are less quantitative than some of the classical chemical methods, but of the projected crystal potential can no longer be done on a single image; one needs a series of images at diVerent defocus these determinations are local and are particularly useful for determination of deviations from an average composition or values in order to retrieve the projected crystal potential. This method has now proven its usefulness in unravelling crystal to detect local fluctuations in the composition.Exactly how quantitative the EDX and EELS methods are is hard to define structures of nano-crystalline domains. In combination with the quantitative analysis of the electron diVraction patterns explicitly; it depends on a number of parameters such as the specific elements present in the compound, the size of the this has become a powerful technique in crystallography.More details on the technique can be found in refs. 3 and 4. domain investigated, the availability of a standard with a fixed composition, etc. One should also realise that most published HREM images are taken at room temperature. Electron microscopy is possible between liquid He temperature and about 1000 °C but at Structure of fluorinated YBa2Cu3O6Fx through HREM temperatures diVerent from ambient temperature the eVects of heating or cooling seriously hamper the resolution.The reso- The possibility of synthesising new superconducting materials is limited by the problem of maintaining electroneutrality and lution is then no longer governed by the aberrations of the lens, but by the stability of the specimen stage.Practically this matching between diVerent chemical bonds in their layered structures. This problem can be reduced significantly and, means that at temperatures diVerent from ambient temperature, a point resolution below 2.5–3.0 A ° is extremely hard to obtain. therefore, new superconductors can be prepared if oxygen anions located in the insulating blocks are replaced by fluorine.O2- and F- anions have diVerent charges but possess similar HREM combined with other microscopy techniques atomic radii and may form similar structures. The preparation HREM by itself may be highly informative and produce of Cu-based oxyfluorides, however, is often made diYcult by spectacular images; the strength of electron microscopy is the high stability of simple fluorides or oxyfluorides; the use mainly in its combination with several techniques within the of soft chemistry techniques can therefore be a powerful tool same instrument, on the same specimen area.While HREM in overcoming this problem and preparing new materials. For provides real space information, the electron diVraction pattern the fluorination of the high Tc superconducting compound provides complementary information on the sample. The so YBa2Cu3O7 (YBCO), XeF2 can be used successfully as a called convergent beam electron diVraction (CBED), in particu- fluorinating reagent at temperatures not higher than 300 °C.lar, will provide information on the local lattice parameters The resulting compound YBa2Cu3O6+xFy is also superconand the local symmetry (point group as well as space group).ducting (Tc=94 K). The total electron–sample interaction is a very complex With the aim of obtaining Cu-containing superconductors, phenomenon and elastic scattering—used for imaging—is only a fluorination treatment should keep the CuO2 layers one of the eVects (see Fig. 3). The incoming high energy unaltered. Many researchers tried to substitute oxygen by electron is also at the origin of the production of X-rays, light, fluorine but they obtained either multiphase mixtures or Auger electrons, back scattered electrons, inelastically scattered partially fluorinated compounds with lower Tc values than electrons, etc.All of these electrons, X-rays and photons carry that for pure YBa2Cu3O7. In the case of YBCO an anion information on the sample area irradiated by the incoming deficient starting material is definitely needed since anion electrons.Since the electron beam can be focused down to exchange in oxidised YBCO demands higher temperatures 1 nm or even smaller, local chemical or electronic information leading to the formation of very stable admixtures. Using is readily available.The most developed and practically appli- HREM we determined the crystal structure of the new material cable methods are EDX (energy dispersion of X-rays) and and using EDX we provided proof that fluorine has definitely EELS (electron energy loss spectroscopy). EDX analyses the entered the structure. outcoming X-rays which provide quantitative information on YBa2Cu3O6.95 and YBa2Cu3O6.11 formulations were used the local composition, particularly the heavier elements.EELS for fluorination. All subsequent operations, including the fluoralso provides chemical information—particularly for the lighter ination, were carried out in a glove box in a dried N2 elements—but is also sensitive to the binding of the considered atmosphere. Syntheses were carried out in Ni crucibles placed element (e.g. it will easily diVerentiate the sp2 binding in in a hermetically sealed steel or Ni container.More details on graphite from the sp3 binding in diamond). These chemical the preparations can be found in ref. 5. The fluorination of YBa2Cu3O6.95 did not lead to any significant changes in the structure or to a change in Tc, which led us to conclude that in this case there is no noticeable insertion of fluorine into the structure or a replacement of O atoms by F atoms. The fluorination of YBa2Cu3O6.11, however, revealed a strong dependence of the properties and the (micro)structure on the synthesis conditions and annealing temperature.The highest value of Tc=94 K was found after 50 h at 350 °C. The fluorinated samples exhibited a large volume fraction of superconducting phase, which confirms the bulk nature of the observed superconductivity.The fact that no orthorhombic YBCO phase was observed on ED or HREM images implies that the superconducting transition can only be attributed to the fluorinated phase. EDX measurements inside the electron microscope indicate that fluorine had indeed entered the crystal structure of YBCO.However, the fluorine concentration varied between diVerent grains and even within a single grain. The insertion of fluorine into the YBCO structure may occur at two diVerent positions. The most probable is occupation of positions in the Cu(1) layer. When up to one fluorine atom Fig. 3 Schematic representation of the electron–specimen interaction for high energy electrons (100 kV or more) per unit cell is incorporated, the fluorine may either occupy J.Mater. Chem., 1998, 8(4), 797–808 799one of the two available oxygen positions in the Cu(1) layer and form a square Cu(1) co-ordination polyhedron or it may be distributed statistically on the 1/2,0,0 and 0,1/2,0 sites. In the first case such intercalation should be accompanied by an orthorhombic distortion of the initial tetragonal unit cell similar to the case in the YBa2Cu3O7 structure.A statistical arrangement would maintain tetragonal symmetry. We will call this the F1 phase [Fig. 4(a)]. If the fluorine content is more then one atom per unit cell (F2 phase), the extra atoms will occupy the remaining sites in the Cu(1) plane. The presence of two F atoms in this plane results in the formation of an octahedral co-ordination for the Cu(1) atoms [Fig. 4(c)]. Such an arrangement will lead to a distortion of the octahedra due to a Jahn–Teller eVect and a substantial increase of the Cu(1)–Cu(2) separation. An increase of the Cu(1)–O distance (from ca. 1.90 A ° in reduced YBCO to 2.3–2.5 A ° in the F2 phase) will lead to a c-parameter for the new F2 phase of about 13.0 A ° .Obviously, such fluorination will result in the oxidation of the Cu atoms and keep the superconducting CuO2 layers unaltered. A second possibility for the intercalation of fluorine in the YBCO structure is to insert it into the Y layer. This, however, will automatically suppress superconductivity because the F atoms will create a fluctuation of the charge density and induce hole localisation.The combination of diVraction and high resolution microscopy (HREM) allowed us to determine the approximate structure while EDX unambiguously confirms the presence of F even on a nanometer scale. The [001]* diVraction pattern of Fig. 5(a) shows an intensity and spot distribution typical of Fig. 5 Electron diVraction patterns of the main reciprocal zones for tetragonal symmetry with a=3.86 A ° .The overwhelming the F1 and F2 phases. The [001]* zone is common to both compounds. majority of the crystals show this type of tetragonal symmetry, without the presence of extra reflections or diVuse streaks. While the [001]* patterns are the same for all crystals, two tions, which can all be indexed in a tetragonal cell with a= 3.869 A ° and c#11.66 A ° ; i.e.very closely related to the undoped diVerent patterns are observed along [hk0]*; this is illustrated in the [100] and [110] sections of Fig. 5(b)–(e). Such patterns YBCO. The sharp reflections as well as the absence of streaks reveal a perfectly crystalline structure; they are typical of the can be obtained from neighbouring areas, even within the same crystal, as indicated in Fig. 6. F1 phase. The [100]* zone of Fig. 5(b) on the other hand exhibits Patterns 5(c) and 5(e) exhibit clear and well defined reflec- Fig. 4 Schematic crystal structures of: (a) orthorhombic YBa2Cu3O7, (b) tetragonal YBa2Cu3O6, (c) fully fluorinated YBa2Cu3O6F2 800 J. Mater. Chem., 1998, 8(4), 797–808Fig. 6 HREM image of fluorinated YBCO material, together with the optical diVraction patterns obtained from diVerent areas of the HREM image, corresponding to phases F1 (A), F2 (C) and the interface (B) between F1 and F2 planar defects, inducing streaking of the reflections along c*.DiVraction patterns along the [110]* zone axis [Fig. 5(d)] not only exhibit streaks along the c*-axis but very often also show additional streaks (or rows of diVuse spots) parallel to the c*- direction at positions h+1/2, k+1/2, l.This would imply that the structure determination of the F2 phase as tetragonal with a=3.869 A ° and c#13 A ° is only a first approximation. The real unit cell would be something like aÓ2×aÓ2×c. The exact origin of these extra reflections, however, is not clear at this stage. [100] HREM images are able to elaborate the diVerence between the F1 and the F2 phase.The F1 phase shows a contrast very similar to that for the undoped YBCO (Fig. 8). Direct measurement of the c/a ratio, from the HREM image or by using the Fourier transform of limited parts of the HREM images (see, e.g., Fig. 6A) allow us to deduce the c- Fig. 7 Enlarged ED pattern along [100]* (from Fig. 5d) illustrating the presence of F1 and F2 phases with diVerent c* parameter. Only parameter as 11.7 A ° , which is in good agreement with the reflections with l=3n are marked. corresponding X-ray data. The F2 phase, with an enlarged c-parameter, exists as an intergrowth within the F1 matrix (Fig. 6) or occurs as an pronounced diVuse streaks along the c* axis. Such streaking isolated defect with a limited extension (see below).Optical may arise due to the presence of a large amount of planar diVraction obtained from areas corresponding to the phases (001) defects in the F1 phase. A detailed examination of such F1 and F2 (marked A and C respectively in Fig. 6) and its ED patterns indicated the periodic presence along c* of extra interface (marked B) confirm the ED data.The diVerence in intensity maxima apart from the maxima due to the F1 phase the c-parameter for the F1 and F2 phases is clearly visible in (Fig. 7). These reflections are compatible with a new unit cell the enlarged Fig. 9. Using the F1 phase as an internal standard, with lattice parameters a=3.869 A ° and c#13 A ° . This phase we obtain a c-parameter of 13.0–13.2 A ° for the F2 phase.It is we will call the F2 phase. It is clear, however, that, in contrast to the F1 phase, the F2 phase does contain a large number of clear visually that in the F2 phase, the distance between the J. Mater. Chem., 1998, 8(4), 797–808 801Fig. 8 HREM image along [100] for the F1 phase; no defects are Fig. 9 Enlarged area showing an alternation of F1 and F2. The detected over large areas but EDX measurements indicate the presence corresponding unit cells are indicated; the diVerence in the c-axis of fluorene is obvious.Cu(1)–O layer and the neighbouring BaO layers has increased HREM of materials exhibiting colossal magnetoresistance drastically, pointing towards a fluorine uptake within the Cu(1) plane. Recently colossal magnetoresistance (CMR) has been discovered in diVerent manganese perovskite materials with general In order to determine the stacking sequence for this new structure we performed image simulations for diVerent models formula (Ln1-xAx)MnO3 with Ln=La, Pr, Nd, Sm and A= Ba, Ca, Sr.For some of these materials the ratio between the based on partially and fully fluorinated structures. For the F2 phase we assumed that in the Cu(1) plane all the oxygens resistance in zero magnetic field and the resistance in a 5 T magnetic field can be as high as 1011.The exact origin of this have been replaced by fluorine, leading to a structural formula YBa2Cu3O6F2. The atomic coordinates were deduced accord- eVect is still under debate, but it is clear that two important parameters are driving the magnitude of the magnetoresistance: ing to the larger unit cell, taking into account a change of the Cu(1)–O(apical ) separation due to a Jahn–Teller eVect. The the number of itinerant holes and the length and direction of the Mn–O bond. The first is directly connected to the oxidation Cu(2)–Cu(2) distance was kept unaltered with respect to the undoped YBCO.A structural model for the new YBa2Cu3O6F2 state of the Mn ion (the ratio of Mn4+ to Mn3+); the second parameter is connected to the average size of the interpolated compound is shown in Fig. 4(c). Calculated images for diVerent thicknesses and defocus values, based on the above models cation in the perovskite structure. These parameters have in a number of cases been optimised leading to the reported show a reasonable agreement with the experimental images.The remaining diVerence between the calculated and the magnoresitance values. The average room temperature structure has been deter- experimental images is attributed mainly to the presence of local strains due to the intergrowth of the F1 and the F2 phases. mined by other diVraction techniques as orthorhombic with symmetry Pnma and with lattice parameters a#apÓ2; b#2ap; Electron diVraction, as well as direct images, indicates a high interweaving of the F1 and F2 phase.The F2 phase never c#apÓ2. Electron microscopy shows the presence of a large number of orientation variants (domains related by a rotation exists over large areas but mainly occurs as bands within an F1 matrix (see Fig. 6). These bands are more or less parallel of 90°); most electron diVraction patterns therefore show a superposition of the diVerent orientation variants.HREM to the (001) planes but there is no rigid interface. In less fluorinated material, F2-type intercalation exists as isolated images can be obtained along most of the low index zone axes and clearly allow us to separate these diVerent variants defects within an F1 matrix; an example is shown in Fig. 10. The extra wide separation between the Cu(1)–O layer and the (Fig. 11). The contrast is not only thickness dependent (as expected), but also region dependent. Such changes are often surrounding BaO layers means that the atom planes are no longer straight but become wavy when a ‘pancake’ of fluorine observed in the microscope and generally are attributed to local changes in the crystal orientation or in the height of the is intercalated. In material which has been less fluorinated such defects are actually the only direct sign of fluorination.It sample (i.e. the focus of the sample). In the present case, however, it was verified that these eVects did not play. is clear that such isolated defects act as the nucleus for the fluorination process and the formation of the F2 phase.Moreover, in most parts of the image the 0.77 nm spacing (2ap 802 J. Mater. Chem., 1998, 8(4), 797–808Fig. 10 The intercalation of fluorine into the Cu(1) layer occurring as an isolated defect (indicated by arrows); HREM image along [100] Fig. 11 HREM image of Nd0.5Ca0.2Sr0.3MnO3 across a twin boundary.Note the change in contrast and the pronounced 2ap contrast along the b-axis in the left domain. spacing along the b-axis) is clearly observed (see left lower part they are isolated point-like defects, clearly visible in the [101] images [Fig. 12(a)]. In this image the bright dots can be of Fig. 11). In the calculated images for diVerent focii and diVerent thicknesses, this doubling of the unit cell is never correlated with the positions of the (Pr, Ca, Sr) sites in this case while the lower intensity dots are associated with the Mn observed.Even including slight tilts, small misalignments or inclusion of the influence of upper Laue zones did not produce configuration. The intensity of the (Pr, Ca, Sr) sites is very constant, as can be seen from the horizontal scan in Fig. 12(b). the correct experimental images. Careful inspection of the HREM images, and comparison with the calculated images, The intensity of the Mn dots, however, varies considerably within a limited area [see also the scan in Fig. 12(c)]. Care is revealed that the double periodicity is related to the (MnO)2 layer in particular. As soon as we introduce a monoclinic taken to make sure that these defects are not introduced by the intense electron beam used for HREM observations; i.e.distortion, creating space group P21/c, by slight displacements of the oxygens in the (MnO)2 layer, the calculated images fit the density of defects does not alter with observation time. One could visualise three origins for these defects: (i) vacancies the experimental ones.The calculated diVraction patterns remain in agreement with experiment. The atom shifts involved or atomic substitution on the Mn sites, (ii) oxygen vacancies in the MnO6 octahedra, and (iii) variation of the oxygen environ- are less than 5%. As in most superstructures, based on a high symmetry parent ment around the manganese. Neutron, as well as EDX measurements, do not show any significant deviation of the structure—the perovskite structure in this case—twin boundaries and antiphase boundaries occur regularly.Depending on oxygen or the Mn content; the high density of defects can therefore only be explained by a local modification of the the exact composition and the preparation method, the density of these two-dimensional defects can be varied. Independent manganese environment.Taking into account the MnIII and MnIV mixed valence, the variation of contrast is probably of composition or preparation, however, another type of defect occurs in almost all of the perovskite-based CMR materials: closely related to a local ordering of MnIII and MnIV, so that J. Mater. Chem., 1998, 8(4), 797–808 803Fig. 13 HREM image along [100]p of Pr1.4Ca1.45Ba0.15Mn2O7.The inset shows the calculated image for a defocus of -57 nm and a sample thickness of 5 nm. average unit cell with a#ap; c#1.93 nm. Electron diVraction, however, clearly indicates an orthorhombic distortion with lattice parameters a#b#Ó2ap. The perfect structure, as viewed by HREM along the [100]p direction, is shown in Fig. 13; without too much image calculation (shown as inset), the stacking sequence can be deduced immediately.The structure is built up as a sequence of 10 layers, grouped as two blocks of 5 layers shifted over (1/2,1/2,0), i.e. the structure is inner centred. The stacking within one block is AO–MnO2–AO–AO–MnO2, with A=Pr, Ca, Sr (Ba). Alternatively, the structure can be described as a sequence of perovskite blocks, AMnO3, interconnected every two blocks by a rocksalt structured layer.Electron diVraction detects weak extra reflections pointing to an orthorhombic distortion resulting in an enlarged unit cell; the intensity of these reflections is very variable, even within one crystallite. HREM along [001] or [110]p clearly reflects this increase of the unit cell. Apart from this change in symmetry, all samples contain a large number of crystal defects, making the local structure sometimes deviate strongly from the average structure.Some of these will certainly influence the physical properties of the materials. We will not discuss here the intergrowth errors, which do not occur very often and Fig. 12 (a) HREM image of Pr0.75Sr0.25MnO3 along [101].The heavy which are believed not to be essential for the understanding of cations are imaged as intense bright dots, the manganese positions as the physical properties; neither will we focus on the large smaller bright dots. Intensity variations at the Mn positions are particularly evident. (b) A (horizontal) scan along the Ln bright lines. number of dislocation loops, which create a high stress inside (c) A scan along the Mn less intense lines.the material. Details of these can be found in the paper by LaVez et al.7 The main defect, we believe, related to the magnetoresitant properties of these materials is illustrated in a variation of the Mn–O distances around the manganese Fig. 14. It shows the co-existence of the pseudocubic perovskite would be involved, in agreement with the diVerent sizes of structure with general formula AMnO3, as narrow strips of MnIII and MnIV.Moreover, the geometry of the MnIII octahedra this material within the intergrowth structure (Fig. 14). The will be diVerent from that of the MnIV octahedra, owing to the interface between both phases is strictly (001) and in view of Jahn–Teller eVect. More details on these eVects are to be found the lattice match, no strain field or interface dislocation is in a recent paper by Hervieu et al.6 detected at this interface.Surprisingly, however, the perovskite Furthermore, it is challenging to expand the investigation structure within these nanometer wide strips shows exactly the of the manganites to other structures related to the perovskite same local structure as the peroskite single phase discussed structure but where the overlapping of the Mn–O–Mn occurs above.A doubling of the unit cell along the b-axis is present in a diVerent way, as in the materials described above. It was in the HREM images, indicating the same structural deforindeed found that magnetoresistance occurs in layered perovskmations as in the bulk; i.e.the structure is at least orthorhomite- related structures. In that respect we investigated bically deformed. Also the point-like clusters, discussed above intergrowths with two perovskite layers and one rocksalt layer (Fig. 12) and related to the Mn oxidation state, are present with typical structural formulae Ln2-yA1+yMn2O7. TEM techwithin the nanometer sized thin slabs. niques are excellent for determining the intergrowth structure and to elucidate the local structure of such materials; as we VV–VIV mixed valencies in the ionic conductor Bi4V2O11-x will illustrate, perovskite slabs tend to form between the regular intergrowths, causing doubts about the true magnetoresistance Bi4V2O11 is the parent compound of the so-called BIMEVOX family, exhibiting interesting oxide anion conductivity.All such properties of the intergrowth. Polycrystalline samples of composition Nd2-yCa1+yMn2O7, Bi2 O3 based oxides are related to the Aurivillius series, where Bi2O2 sheets alternate with a perovskite-like layer containing Nd1.4Ca1.6-xSrxMn2O7 and Pr1.4Ca1.6-xBaxMn2O7 were prepared in the usual way. All materials crystallise in a tetragonal oxygen vacancies which are responsible for the electrical 804 J.Mater. Chem., 1998, 8(4), 797–808Fig. 14 HREM image along [100]p of Pr1.4Ca1.45Ba0.15Mn2O7 showing strips of the perovskite-based phase a few nanometers wide, within an intergrowth structure Fig. 15 (a) HREM image of Bi4V2O10.66 along [100]; groups of three vanadium ions are highlighted in the image as black dots between two intense white dots; (b) corresponding electron diVraction pattern.(c) Computer simulated HREM image for a defocus of 45 nm and a crystal thickness of 5 nm. (d) Image of a defective Bi4V2O10.66; locally blocks of four vanadiums are observed; they are indicated by arrows. properties of these materials. Although the basic Aurivillius structure is known, the actual structure is more complex, due to the presence of twinning, the formation of incommensurate modulations and problems of stoichiometry.For an overview Fig. 16 Electron diVraction patterns of Bi4V2O11 along [001] and of the chemical, physical and structural properties, we refer along [100], clearly showing the commensurate modulation along the b-axis the reader to ref. 8 and references therein.J. Mater. Chem., 1998, 8(4), 797–808 805Fig. 18 (a) [10] HREM image of the intergrowth of one unit cell of a four layer unit within the 123 structure observed in a PrBCGaO barrier layer (x=0.7). The inset shows a simulated image for d= 2.4 nm; d=-25 nm based on the proposed model (right hand figure). Fig. 17 (a) HREM overview of a ramp-type JJ with a PrBCGaO barrier of 10 nm, showing that the interfaces at the ramp are well lems which may arise with HREM, related to the use of an defined and free of secondary or amorphous phases.(b) Schematic intense electron beam. Upon heating, the reduction from VV representation of the investigated ramp-type junctions. to VIV is enhanced under the electron beam and the a-phase is no longer stable.In other samples, e.g. SiO2–quartz, ionisation damage is formed very rapidly when studying the material Bi4V2O11 is the upper limit of the solid solution at room temperature; when heating, however, the ionisation 2Bi2O3–xV2O5 within the Bi2O3–V2O5 binary phase diagram; damage anneals out faster than the production of new defects however, pure Bi4V2O11 is practically never obtained since and the material is perfectly stable.In this way we have studied small traces of BiVO4 are systematically observed. Its ‘average’ without too many problems the phase transition at 573 °C structure is orthorhombic (a#0.55; b#0.56, c#1.53 nm) but between the a-phase and the b-phase of quartz. When working three diVerent polymorphs are known as a function of temperawith an intense electron beam, particularly for a field emission ture. The main reason for this complexity is likely to be related source, one has to realise that this may alter the initial material to the ability of vanadium to undergo a VV to VIV reduction (even burn a hole in the sample).when the experimental environment is altered by thermal treatment or oxygen partial pressure. This can lead to the HREM of inorganic materials with reduced dimensionality formation of mixed valence phases, and ultimately to the formation of Bi6V3O16 (Bi4V2O10.66) with VIV/VV=1/2.With the progressive miniaturisation of the present technology, systems of reduced dimensionality (thin films, nano-wires or Electron microscopy oVers the possibility of following the structural changes during in situ heating.At any moment of quantum dots) gain more and more interest. The connection (epitaxially or not) between diVerent layers or between layers, the transformation, the heating can be stopped and the corresponding electron diVraction and corresponding HREM image wires, dots and the substrate are crucial in determining their properties. High resolution electron microscopy is one of the can be recorded.The complex [100] diVraction pattern of the reduced mate- few techniques which provides real space information down to the atomic level for such interfaces. Moreover HREM contrib- rial Bi4V2O10.66 is illustrated in Fig. 15(b). together with the corresponding HREM image [Fig. 15(a)]; a calculated image utes to the monitoring of the production of these man-made materials.We will discuss here briefly the problem of growing based on the available X-ray data for a thickness of 5 nm and a defocus close to the Scherzer defocus is shown as Fig. 15(c). superconducting Josephson junctions (JJ) on a SrTiO3 substrate. Although this is a specific problem, we hope it will The vanadium configuration is highlighted as groups of black dots between two intense white dots.This block of three illustrate some of the particular diYculties with nanometer thick films on a substrate. vanadium atoms along the b-direction corresponds to one VIV and two VV in the structure [see Fig. 15(a)]. The waviness of YBa2Cu3O7 Josephson junctions can be fabricated by sandwiching a thin barrier of non-superconducting material these layers is clearly revealed in the HREM image.This periodicity of one VIV and two VV is sometimes violated, as in between two superconducting electrodes. To be able to interpret correctly the measured properties, one has to perform Fig. 15(d), where blocks of four vanadium atoms are observed, indicating domains where the reduction of the compound is cross-section HREM in order to obtain local structural information from the junction area itself.However, the cross-section lower than the ultimate VIV/VV=1/2. These blocks of four vanadiums [indicated by arrows in Fig. 15(d)] likely result sample preparation of JJ for HREM investigations is not easy, particularly when diVerent layers are deposited. Details on this from a regular arrangement of one VIV for three remaining VV.This defect actually introduces an antiphase boundary specific preparation method by ion beam bombardment can be found in refs. 9 and 10. with R=1/3[010] in the structure. The structure of the basic Bi4V2O11 (the a polymorph) is The preparation of the ramp-type JJ is summarised as follows. The base electrode (DyBCO) and the separating not entirely clarified; the unit cell has been determined and the heavy V ions located, but the exact oxygen configuration was insulating layer (PrBCO) are deposited in situ.The ramps are structured by an Ar ion beam source using a photoresist mask. still unresolved until recently. The structure becomes slightly monoclinic and an extra modulation (6-fold) occurs along the After stripping the photoresist, the ramp surface is cleaned by ion-milling and is re-oxidised in situ.Subsequently, the bar- b-axis; this is evident from the [001] and the [100] diVraction patterns in Fig. 16. HREM of this phase, however, is barely rier layer (PrBCGaO) and the top electrode (DyBCO) are deposited. The Ga content of barrier layer is nominal. A possible; under the intense electron beam for high resolution observation, a gradual transformation occurs due to the VV to schematic drawing of the complete ramp-type JJ is shown in Fig. 17, together with the TEM image observed experimentally. VIV reduction. We included this example deliberately to illustrate the prob- In the PrBCO layer Cu can be substituted by Ga for 806 J. Mater. Chem., 1998, 8(4), 797–808Fig. 19 [001] HREM micrograph of the interface between (a) PrBa2Cu2.3Ga0.7O7-d and ion-etched SrTiO3.Dark contrast Ga-rich areas in the substrate are indicated by arrows. substitution levels 0.01,x,0.9. XRD shows that bulk content is increased the density of this type of intergrowth is also increased. The presence of Ga apparently favours the PrBa2Cu3-xGaxO7-d samples have a structure identical to tetragonal ReBa2Cu3O7-d.No Ga-containing impurity phase presence of the intergrowth. HREM combined with local microanalysis allows us to propose a model for the Ga- is present in PrBCGaO with x=0.7 indicating the full substitution of Cu by Ga. The unit cell parameters are a= containing intergrowth layer. The unit cell for this new structure is reproduced in the right hand part of Fig. 18. There is a 0.39288(6) nm and c=1.1756(5) nm. For the substitution level x=1 two impurity phases were found, namely CuO and a- clear inequivalence in contrast between the CuO planes and the plane in the middle of the intergrowth unit cell. The BaGa2O4. The fact that a Ga-containing phase is present as an impurity phase indicates that x=1 is above the maximum contrast above and below the CuO planes is similar.The contrast strongly resembles a double perovskite with an substitution level for Ga in PrBa2Cu3-xGaxO7-d . Bulk PrBa2Cu3-xGaxO7-d with x0.7 has been carefully investi- inequivalent lattice plane in the middle of the cell. This was interpreted as showing the occurrence of a full Ga plane. gated by EM and SAED. There is no evidence for any modulation or structural change due to the Ga substitution in A model with layer stacking PrBCO–CuO–(Pr,Ba)O– GaO–(Pr,Ba)O–CuO–PrBCO, based on one slab of the Cu(1) chain. Electrical properties of ramp-type JJ with PrBCGaO barriers (PrBa)CuGaOz sandwiched between PrBCO unit cells, was created and image simulations were carried out.Ga and its (x=0.1 and 0.4) have been measured. The eVect of a PrBCO barrier layer with Ga substitution on the junction parameters surrounding oxygens were shifted from the ideal perovskite positions due to the tetrahedral co-ordination of Ga.The is an increase in the resistivity of up to an order of magnitude for the same barrier thickness as compared to junctions with computer simulated images based on this model yield good agreement with the observed HREM image (see the inset in unsubstituted PrBCO as the barrier.However, this trend is not continued for the higher substitution levels x=0.7 and 1. the image of Fig. 18). Within the SrTiO3 substrate, pronounced black patches are On the contrary, no increase in resistivity for x=0.7 was observed as compared to x=0. The origin of this remarkable observed in the uppermost 10 nm of the interface region with the barrier layer (see Fig. 19). This specific contrast exists in fact could only be traced by HREM. A cross-section TEM investigation provides information all parts of the SrTiO3 substrate, even on the inclined rampedge in SrTiO3, which have been exposed to Ar etching to about the local structural of the barrier layer as incorporated in the JJ (Fig. 17). No peculiarities were found at the interfaces create the ramp-edges. Such contrast variations are not observed normally for ReBCO films on SrTiO3 and we can or inside the superconducting layers responsible for the abnormal behaviour of the JJ with substitution levels x=0.7 or 1 eliminate ion-beam thinning sample preparation as a possible origin. Possible other causes could be oxygen depletion, or for the barrier layer.The junctions discussed here have a ramp angle of 21°–28° and a typical barrier layer thickness of interdiVusion either of the photoresist material or of gallium from the barrier layer. The contrast of the black patches is 10–30 nm. All layers grow epitaxially with the c-axis orientation perpendicular to the substrate. All interfaces are sharp only observed for the combination of a Ga-containing layer on an ion-etched SrTiO3 substrate.Therefore, we believe that and flat. The interfaces between barrier and superconducting layers at the ramp-edge are well defined and free of secondary Ga diVusion in the surface region of the ion-etched SrTiO3 substrate is the reason for the local structural distortions. This phases or amorphous layers.This allows us to correlate the observed microstructure of the PrBa2Cu3-xGaxO7-d barrier can be understood if we assume that lattice imperfections, such as oxygen vacancies or point defects, which are created in the layer and the electrical properties of the ramp-type JJ with this barrier material. SrTiO3 substrate by the ion-etching, facilitate the diVusion of gallium in the interface region of the substrate. HREM images show that the barrier layer contains isolated unit cells with a c-parameter diVerent from c123; a single strip We conclude that two distinct mechanisms of Ga segregation in thin barrier layers are operative; the formation of a Ga- is shown in Fig. 18. These unit cells, with a spacing along the c-direction of approximately ci=0.8 nm, diVer from the normal containing intergrowth in the barrier layer itself and diVusion of Ga in the ion-etched substrate. The microstructure of YBCO-123 structure and are not known as common defects in bulk YBCO-123.The intergrowth only occurs in the PrBa2Cu3-xGaxO7-d (x=0.7 or 1) thin barrier layers is clearly diVerent from that in the bulk material. PrBa2Cu3-xGaxO7-d PrBa2Cu2.3Ga0.7O7-d barrier layer, not in the DyBCO bottom or top layer nor in the PrBCO separating layer.If the Ga with 0.1x0.4 can be used as barrier layer in ramp-type J. Mater. Chem., 1998, 8(4), 797–808 807Josephson junctions to enhance reproducibly the normal state on this matter, I believe it will strongly contribute to the future success of transmission electron microscopy.resistivity. However, higher Ga substitutions are not eVective. This example of the Josephson junction on a SrTiO3 substrate illustrates the possibilities of HREM in the study of the I would like to thank my co-workers at the EMAT research centre of the University of Antwerp (K. Verbist, P. LaVez, structural aspects related to the deposition of a thin layer or sequences of thin layers on a given substrate.Other interesting M. Huve�, O. I. Lebedev, D. Van Dyck, J. Van Landuyt and S. Amelinckx), as well as collaborators from outside the phenomena which are beneficially studied by TEM include the change of growth mechanism as a function of temperature, the University (D. Blank, R. V. Shpanchenko, E. V. Antipov, M. Hervieu) for the use of results from common publications.deposition and growth of diamond thin films, the study of semiconductor devices, the growth of (carbon) nanotubes or Financial support is provided by IUAP 4/10. the characterisation of nanometer or sub-nanometer sized particles. References 1 Handbook of Microscopy, ed. S. Amelinckx, D. Van Dyck, Future of HREM in materials science J. Van Landuyt and G. Van Tendeloo, VCH,Weinheim, 1997. Since the discovery of electron microscopy in 1931, the instru- 2 Electron DiVraction T echniques, ed. J. M. Cowley, Oxford mental resolution has decreased steadily, to reach the atomic University Press, 1993. 3 H. W. Zandbergen, S. J. Andersen and J. Janssen, Science, 1997, level in the seventies and the 1 A° level in the nineties. However, 277, 1221. to prove this 1 A ° resolution limit, a strongly scattering sample 4 D. Van Dyck, in Electron Microscopy, ed. S. Amelinckx, D. Van with a minimal thickness (preferentially in the 1 nm range) is Dyck, J. Van Landuyt and G. Van Tendeloo, VCH, Weinheim, required. It is my personal impression that the future of 1997. electron microscopy for materials science is not so much in 5 R. V. Shpanchenko, M. G. Rozova, A. M. Abakumov, improving the ultimate resolution below the mythical 1 A ° but E. I. Ardashnikova, M. L. Kovba, S. N. Putilin, E. V. Antipov, O. I. Lebedev and G. Van Tendeloo, Physica C, 1997, 280, 272. should rather concentrate on: (i) a correct and ‘user friendly’ 6 M. Hervieu, G. Van Tendeloo, V. Caignaert, A. Maignan and interpretation of the high resolution images down to the B. Raveau, Phys. Rev. B, 1996, 53, 14274. instrumental resolution. The feasibility of this has been proven3 7 P.LaV ez, G. Van Tendeloo, R. Seshaddri, M. Hervieu, C. Martin, but ‘everyday use’ has still some way to go; (ii ) a quantitative A. Maignan and B. Raveau, J. Appl. Phys., 1996, 80, 5850. interpretation for the combination of electron diVraction and 8 M. Huve, R-N. Vannier, G. Nowogrocki, G. Mairesse and HREM. With CCD cameras becoming more and more stan- G. Van Tendeloo, J.Mater. Chem., 1996, 6, 1339. 9 K. Verbist, O. I. Lebedev, G. Van Tendeloo, M. A. J. Verhoeven, dard, the evolution of this will be more or less straightforward; A. J. H. M. Rijnders and D. H. A. Blank, Supercond. Sci. T echnol., (iii ) the use of the electron–sample interaction to its full 1996, 9, 978. capacity, i.e. we shall no longer limit ourselves to the trans- 10 K. Verbist, O. I. Lebedev, G. Van Tendeloo, M. A. J. Verhoeven, mitted and the elastically scattered electron beams, but also A. J. H. M. Rijnders, D. H. A. Blank and H. Rogalla, Appl. Phys. use the information available in the emerging X-rays, inelas- L ett., 1997, 70, 1167. tically scattered electrons, secondary electrons, backscattered electrons, light, etc. Although a lot of progress has been made Paper 7/08240A; Received 17th November, 1997 808 J. Mater. Chem., 1998, 8(4), 797–8

ISSN:0959-9428    

DOI:10.1039/a708240a

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 Carbazole photorefractive materials Yadong Zhang,a Tatsuo Wada*a,b and Hiroyuki Sasabea,b aCore Research for Evolutional Science and T echnology (CREST), JST and bFrontier Research Program, T he Institute of Physical and Chemical Research (RIKEN), Hirosawa 2–1, Wako, Saitama 351–0198, Japan Considerable progress has been made in organic photorefractive materials, since the first observation of photorefractive phenomena from organic materials.Within recent years, a large number of organic photorefractive materials, especially amorphous materials, have been developed based on polymeric composites, fully functional polymers and the multifunctional chromophore approach. Among these organic photorefractive materials, some of them containing carbazole components as a charge transporting function have been demonstrated to exhibit high performance photorefractive eVects. The carbazole building blocks with charge transporting functionality or multifunctions play a very important role in photorefraction and have been widely used in the molecular design approach to new organic photorefractive materials. Based on carbazole functional building blocks, amorphous multifunctional chromophores, amorphous monolithic chromophores and amorphous dendrimers have also been developed as new types of organic photorefractive materials.This paper reviews the recent progress of organic photorefractive materials, especially organic photorefractive materials containing carbazole functional components. A change in refractive index (photorefraction) of a material optical nonlinearities, low cost, low relative permittivity (e), structural modification flexibility and ease of fabrication.may have many origins. Most of the mechanisms generally Compared with inorganic photorefractive crystals, the low lead to irreversible photorefraction processes. However, some relative permittivity of organic materials is an important reason are reversible which is very important for potential applifor pursuing the development of organic photorefractive mater- cations.1,2 Reversible photorefraction processes can be due to ials.A useful figure of merit for photorefractive materials can several microscopic mechanisms, such as the space charge field be defined as Q=n3re/e, where n is the refractive index, re the induced photorefractive eVect, photodimerization, photoisoe Vective EO coeYcient, and e the dc relative permittivity.Thus merization, thermo-optic eVects and photoinduced inter- or Q is approximately measured as the ratio of the optical intra-molecular structural changes.3,4 Except for the space nonlinearity to the screening of the internal space charge charge field induced processes, other processes have local distribution caused by polarization of the medium.For inor- mechanisms which lack the nonlocal property of the photorefganic materials, the optical nonlinearity is driven chiefly by ractive eVect arising from the physical motion of the charges the large ionic polarizability. The EO eVects from the space in the materials.Only the space charge field induced reversible charge in inorganic crystals appear to be limited, since any process can cause a phase shift between the refractive index increase in a component of the EO coeYcient of a material is grating and the light intensity pattern. An important consee Vectively counterbalanced by an increase in the corresponding quence of this phase shift is energy transfer between the two e value.Therefore, Q values remain low in most inorganic light beams interfering in a photorefractive medium. This space crystals.14 For organic materials, the nonlinear optical (NLO) charge field induced photorefractive eVect is defined as the response is a molecular property arising from the asymmetric spatial modulation of a material’s refractive index in response distribution of the electronic charges in the molecular ground to an optically induced charge distribution and is a nonlocal and excited state.15 For this reason, in organic materials large physical process.In this paper, the progress of the space charge EO coeYcients are not accompanied by large dc e values; thus field induced photorefractive eVects in organic materials will a potential improvement in the performance of the photorefracbe discussed.tive eVects of organic materials can be achieved by a suitable The space charge field induced photorefractive eVect was and reasonable molecular design. It is well known that the first observed by Ashkin et al.5 in 1966 in inorganic electrogrowth and preparation of single crystals is generally a timeoptic (EO) LiNbO3 crystals.The photorefractive eVect has consuming and diYcult processes. This is an even more now been observed in a large number of inorganic mate- important factor when attempting to engineer the properties rials,1,2,6,7 such as KNbO3, BaTiO3, Bi12SiO20 (BSO), of a single crystal by the modification of the crystal to include B12GeO20 (BGO), GaAs, and InP5Fe.Many diVerent devices desired functionalities for the materials. On the other hand, have been developed for numerous applications, including high the photorefractive properties of amorphous organic materials density optical data storage, optical image processing, dynamic may be improved by both chemical and physical modifications. holography, optical computing and phase conjugated Amorphous organic materials are generally more amenable to mirrors.8–10 processing into device structures with large areas and useful In 1990, the first observation of the photorefractive eVect in geometries by coating and other methods.an organic doped crystal was reported by Sutter et al.11,12 In this review, recent progress in organic photorefractive However, the growth of high quality organic single crystals is materials with carbazole moieties is summarized.An introducdi Ycult. The IBM group reported the first amorphous poly- tion to the necessary functional components for molecular meric photorefractive material based on a guest–host com- design approaches to organic photorefractive materials, posite system in 1991.13 characterization of organic photorefractive materials and Amorphous organic photorefractive materials can oVer identifying the nature of nonlocal photorefractive eVects are also described.many advantages over photorefractive crystals, such as large J. Mater. Chem., 1998, 8(4), 809–828 809for charge generation from any of the other components, such Necessary functional components for approaches to as second-order NLO chromophores, will increase other com- photorefractive design peting processes.22 When we design molecules for photorefractive materials, the absorption coeYcient at the operating To be photorefractive, a material must have photoinduced charge generation, charge transporting and charge trapping wavelength from the charge generation molecules should first be confirmed by a spectroscopic measurement.Any absorption properties, as well as an EO response. When light is incident on a photorefractive material, if the incident light is not due to other functional components should be avoided from the operating laser wavelength for photorefraction. uniform in intensity, photogenerated charges will migrate through the transporting component from the illuminated area to the dark area, where these charges get trapped by trapping Photoconductivity for the establishment of the space charge field centers.The resulting charge redistribution creates the space charge fields in the material. These fields produce measurable Photoconductivity in amorphous organic materials consists of photocharge generation and charge transport processes.The changes in the refractive index through the linear EO eVect in noncentrosymmetric materials. The mechanism behind the charge generation molecules absorbing photons produce electron –hole pairs which, under the influence of a driven electric photorefractive eVect is summarized from the considerable body of prior work on the inorganic photorefractive crystals.16 field, dissociate to produce electrons and holes.23 These free carriers then migrate through the material via a hopping The formative process of the refractive index grating in organic photorefractive materials is similar to that of the inorganic mechanism by the charge transport components.24 Measurement of these properties is an important first step for crystals.14,16,18 According to the requirements and the mechanism of the photorefractive eVects, photorefractive materials the characterization of a potential organic photorefractive material.The photocharge generation quantum yield and the must have two main functions, photoconductivity for the establishment of a space charge field and the linear EO photoconductivity can be measured using a simple photocurrent technique or a xerographic discharge technique.25 eVect for the formation of a refractive index grating.Photoconductivity in organic materials consists of photocharge Charge mobility can be measured by the time-of-flight technique. 26 One detailed study on the photoconductivity of an generation and charge transporting processes. In amorphous organic photorefractive materials, photocharges can be induced organic photorefractive material has been presented.27 by addition of appropriate sensitizers, such as organic dyes; generated charges can be transported through the hole trans- Linear EO eVect for formation of a refractive index grating porting component, such as carbazole and triphenylamine; the defects in the materials can play a trapping role for these A number of diVerent methods have been developed to measure the EO coeYcient in polymer films.The most widely used charges. Second-order NLO chromophores can provide the linear EO eVect when the dipole orientation of chromophores methods for thick polymer films are the reflection technique28 and the Mach–Zehnder interferometric technique.29,30 These is achieved by an applied electric field.Thus the multifunctionality of organic photorefractive materials can be achieved by two methods are also widely used for measuring the EO coeYcients of amorphous organic photorefractive materials. It two molecular design approaches: the guest–host composite approach14,17 and the fully functional polymeric material should be pointed out that the EO coeYcients obtained at low frequencies contain the contribution from the field induced approach.18 Most of the reported amorphous photorefractive materials were based on guest–host composite systems using birefringence.More accurate measurements of the EO coeYcients should be performed at high modulation ac fre- second-order NLO polymers, a charge transporting polymer, or an inert polymer as a host doped with the corresponding quencies at which the birefringence eVect is insignificant and only the EO eVect of chromophores can respond to these ac necessary functional components.14,17 Recently, bifunctional chromophores combining both charge transporting function frequencies.31 and EO function doped in inert polymers have also been reported.14,17 The composite materials approach has the advan- Two-beam coupling measurements tage of easy optimization of the multifunctionality by independently varying the nature and concentration of each In order to unambiguously distinguish between the photorefractive eVect and other types of gratings, two-beam coupling component.However, there are inherent problems in phase separation of these doped systems which limit the concen- measurements must be performed.The phase shift between the refractive index grating and the light interference pattern can trations of active moieties. In order to overcome this problem, fully functional polymers containing all necessary functional be determined by two-beam coupling techniques. This phase shift, or nonlocal nature of the photorefractive eVect, gives rise groups either in the polymeric main-chain or in the side-chain have the evident advantage of long-term stability and mini- to an asymmetric energy transfer between the two writing beams which does not occur in any of the other refractive mized phase separation.18 However, the time-consuming chemical synthesis and diYculty in rational design are constant index change processes.Therefore, the nonlocal character of the photorefractive eVect can be directly confirmed by an challenges. More recently, amorphous multifunctional and monolithic chromophore approaches to photorefractive mate- observation of the phase shift or the asymmetric energy transfer in two-beam coupling experiments. Experimental details of rials have also been developed.19–21 These approaches can result in non-polymeric supporting photorefractive materials.two-beam coupling techniques can be found in the literature.12,32,33 Characterization of organic photorefractive Four-wave mixing measurements materials The four-wave mixing experiments reveal a large amount of Absorption coeYcient for the charge generation information about photorefractive materials.Using four-wave mixing techniques, grating formation dynamics can be studied. In order to observe a photorefractive eVect, a material should have a suitable optical absorption coeYcient at the operating The diVraction eYciency, a key parameter judging from the performance of the photorefractive eVects can also be measured laser wavelength for the generation of the photocharges.In the optimal case, this absorption coeYcient should come from by a four-wave mixing technique. Such measurements have been carried out in detail for many organic photorefractive the contribution of the charge generation molecules, such as C60 and charge-transfer complex. The absorption coeYcient materials.14,17,18 810 J. Mater. Chem., 1998, 8(4), 809–828mined from independent experimental results such as from the Identifying characteristics of nonlocal diVraction eYciency of four-wave mixing and from the two photorefractive eVect beam coupling gain, are in agreement within experimental error.We can say that the nonlocal index grating is a dominant Other local processes as mentioned above can also lead to changes in the refractive index of the materials.How can we mechanism.21 distinguish the space charge field induced photorefractive eVects from other local photorefraction processes? There are Amorphous organic photorefractive materials many ways to distinguish photorefraction from competing mechanisms that lead to changes in the refractive index. Here without carbazole functional components some of the important characteristics for identifying nonlocal Second-order NLO polymers as hosts photorefractive eVects are listed.To date, three types of second-order NLO polymers, cross- Phase shift and energy transfer linkable epoxy polymers,13,32,36,37 copolymers of methyl methacrylate and acrylate38–40 and linear epoxy polymers,41–44 have When the grating is only produced by the migration of the been used as the host matrix for photorefractive composites.charge by diVusion away from the illuminated regions (nonlocal process), the phase of the index grating is shifted by 90° with respect to the light intensity grating. An important consequence of this phase shift is asymmetric energy exchange in two-beam coupling.1,32 Photorefractive eVects involving other processes do not show a 90° phase shift.A critical element in this coupling is the nonzero phase shift between the refractive index grating and the light intensity grating. Techniques for measuring the grating phase are reported.12,32 Hologram erasability The traps that lead to the formation of the space charge fields in photorefractive materials must themselves be photoionizable.If so, the photorefractive diVraction grating can be erased simply by bathing the material in a uniform field of the appropriate frequency. The photorefractive index changes can be written and erased repeatedly over a long time without noticeable change in the chemical structure of the material.34 For photochemical eVects, reversibility generally requires heating or other chemical treatment.Most photochromic gratings tend to be only partially erasable under uniform illumination.4 EO response and photoconductivity Most local eVects, such as photochromic and photochemical eVects, do not depend upon the absence of the linear EO and the photoconductive eVects. For photorefractive materials, the photoconductivity and EO nonlinearity of materials must be measurable at the same wavelengths as used for photorefractive measurements.These two functional properties are basic necessary conditions for the nonlocal photorefractive eVects.14,17 Enhancement by external fields Drift mobilities in amorphous organic photoconductors generally scale with the square root of the applied electric field.35 For amorphous organic materials, the linear EO eVects can be enhanced by the high orientation degree of second-order NLO chromophores at the high applied electric field.Application of a high external field can enhance photorefractive eVects due to the increase of both the EO nonlinearity and the photoconductivity. 14,17 Polarization anisotropy of the grating readout and modulation of refractive index (Dn) N N NO2 O OH O OH N O OH O OH O2N N O O OH N R OH NO2 NO2 O R COOCH3 O N NO2 N N SCH3 O O N bisA-NAT bisA-NAS R = R = n x y R = R = bisA-NPDA NNDN-NAN PMMA-PNA PMMA-MSAB The photorefractive response is dependent upon the polarization of the incident beam.If pure s- or p-polarized light is In this material system, benzaldehyde diphenylhydrazone (DEH) was used as a charge transporting agent and the used in writing the grating, the diVraction eYciency will be diVerent for the s- and p-polarized components of an oblique second-order NLO chromophores played two roles: as the EO function and the charge generation function.The first photoref- reading beam.14 Unlike the two-beam coupling, the four-wave mixing diVraction eYciency is independent of the phase shift ractive eVect was found in this system based on a second-order NLO cross-linkable epoxy polymer doped with DEH as a hole of the index grating.In principle, a local grating induced by many other mechanisms such as photochemistry and photo- transporting agent.13 The photorefractive nature of this material was demonstrated by two-beam coupling.13,36 It was found chromism can also contribute to the four-wave mixing diVraction eYciency.If the modulation of the refractive index deter- that the photorefractive grating was dependent on the applied J. Mater. Chem., 1998, 8(4), 809–828 811electric field: at low electric fields a weak in-phase grating and composites, in which each of three components exhibits one necessary function.46–49 the low EO coeYcient prevent observation of a phase shifted photorefractive grating and at high electric fields a much stronger photorefractive index grating with a phase shift approaching 90° could be observed,32,36 New phenomena, electric field stabilization of photorefractivity and grating revelation which cannot be observed in inorganic photorefractive crystals, was observed from the second-order polymer bisA-NAT doped with 40 wt% of DEH.37 Some photosensitizers, such as borondiketonate and C60, as a charge generation functional component were also used as dopants in second-order NLO polymers. The performance of various sensitizers used to generate mobile charges at suitable long wavelengths were studied.39–41 In order to improved photorefractive eVects, other charge transporting components, such as p-(diethylamino)benzaldehyde N-(1-naphthyl)-N-phenylhydrazone, N-{[(4-diethylamino)phenyl]methylene}-9Hcarbazol- 9-amine and tritolylamine have also been used.36,40 The results indicate low photorefractive performance with the two-beam coupling gains of <2.2 cm-1 and diVraction eYciencies of <0.11% for the photorefractive system using second-order NLO polymers as a functional matrix.Recently linear epoxy polymers with 4,4¾-nitroaminostilbene (bisA-NAS) as a second-order NLO chromophore containing 29 wt% of PMMA composites doped with 30 wt% of DEH (a charge DEH were reported, with the stilbene dye substitutent NAS transporting moiety), 30 wt% of NPP (a second-order NLO also serving as a charge generation function.42 This composite chromophore) and less than 0.1 wt% of a squarylium dye (SQ) material exhibited good photorefractive performance with twoor 0.2 wt% of tetracyanoquinodimethane (TCNQ) (a charge beam coupling gains as large as 50 cm-1 at an applied electric generation moiety) showed lower glass transition temperature field of 70 V mm-1.However, no net gain was obtained due (Tg) due to the high doping level of DEH and NPP.The to the large absorption coeYcient (a >200 cm-1) at the photorefractive nature of these inert composites was confirmed wavelength of the laser beam (l=650 nm) used. by the two-beam coupling. A phase shift of 90° between a refraction grating and a fringe pattern with applied electric Non-carbazole charge transporting polymer as a host fields was obtained.46 A diVraction eYciency of 1.1% and a beam-coupling gain coeYcient of 10 cm-1 were achieved at a Only one non-carbazole charge transporting polymer, poly(4- dc electric field of 34.9 V mm-1.47 Burzynski et al.48 used PC n-butoxyphenylethylsilane) (PBPES), was used for photorefracas a host for a charge transporting agent, TTA (30 wt%), tive composites.45 In this polymer, either 20 wt% of the laser a second-order NLO chromophore, either NPP (20 wt%) or dye coumarin-153 (C-153) or 40 wt% of (E)-b-nitro-(Z)-b- 6-propionyl-2-dimethylaminonaphthalene (PRODAN) methyl-3-fluoro-4-N,N-diethylaminostyrene (FDEAMNST) as (20 wt%), and a charge generation sensitizer, C60 (0.25 wt%).a second-order NLO chromophore, and 0.2 wt% of either Photoconductivity and the EO eVect were observed in both trinitrofluorenone (TNF) or the fullerene C60 as a sensitizer composites along with their electric field dependence.In the were doped. case of composites containing PRODAN as a second-order NLO chromophore, the chromophore showed absorption at short wavelength. This enabled the authors to demonstrate photorefractivity at a wavelength as short as 488 nm. A maximum diVraction eYciency of 3.2% (at 75 V mm-1) and a net two-beam coupling gain coeYcient of 20 cm-1 (at 42.5 V mm-1) have been obtained for a PC–TTA–NPP–C60 composite at a wavelength of 633 nm.More recently, a novel photorefractive polymeric composite using PMMA as a host was reported.49 In this material, 2,6-di-n-propyl-4H-pyran-4-ylidenemalonoitrile (DPDCP), a transparent, highly dipolar chromophore with large anisotropy of the optical polarizability and negligible second-order nonlinearity was chosen as a functional dye.N,N¾-Bis(3-methylphenyl)-N,N¾-bis(phenyl)- benzidine (TPD) was employed as an eYcient charge transport agent.50 The fullerene C60 was used as a sensitizer. The photorefractive measurements were carried out on a doped polymer material consisting of 30 wt% DPDCP, 15 wt% TPD, The photorefractive properties of these guest–host polymeric 55 wt% PMMA, and 0.3 wt% C60.A 100 mm thick film of this composites were characterized by both four-wave mixing and material exhibits a steady-state diVraction eYciency of 25% two-beam coupling. Steady-state diVraction eYciencies as high and net two-beam coupling gain of 50 cm-1 at a bias field of as 10-4 and rise times as short as 39 ms were obtained by 100 V mm-1.It was found that the microscopic mechanism of four-wave mixing. This is one of the fastest responses in organic the photorefractive eVect in this material involves the formation photorefractive polymers. Net two-beam coupling gains were of a refractive index grating through a space-charge field- also observed in this system. modulated Kerr eVect.Inert polymers as hosts Bifunctional chromophores doped in inert polymers Two inert polymers, poly(methyl methacrylate) (PMMA) and Usually multicomponent composite photorefractive systems, in which each component exhibits one functional process polycarbonate (PC) were used as hosts for photorefractive 812 J. Mater. Chem., 1998, 8(4), 809–828mentioned above, are inherently unstable and have a tendency the four-wave mixing diVraction eYciencies and their response towards phase separation which results in loss of the optical times) increase with the increase in the chromophore concenquality. tration.54 The relation between the photorefractive eVects and Among these functional components, the linear EO and trap density also has been studied. It is found that the gain photoconductive moieties play the main roles in photorefrac- coeYcient and diVraction eYciency can be enhanced by adding tive process. However, phase separation limits the concen- small amounts of the trap molecule, 1,4-bis(N,N-dimethyltration of EO and charge transporting functional components. amino)benzene (BDB).57 Simultaneously the phase-shift In order to obtain a high degree of loading for these two between the illumination pattern and the refractive index functional components, one chromophore combining several grating decrease with increasing BDB concentration.The trap functions might be desirable. This multifunctional design molecules have an eVect on the creation of the space charge field.Chromophore concentration dependence of the EO approach might circumvent the problem of phase separation coeYcients, photoconductivity, four-wave mixing diVraction which limits the concentrations of main components. The eYciency and the two-beam coupling gain coeYcients indicates bifunctional molecules combining both second-order NLO and the great potential of this bifunctional design approach.53,56 charge transporting properties51–57 have already been developed based on charge transporting molecule building blocks, such as triphenylamine (TPA)24 and N,N¾-diphenyl-N,N¾-bis(3- methylphenyl)-1,1¾-biphenyl-4,4¾-diamine (TPD).58 Therefore, Glass molecules for photorefractive materials by increasing the concentration of a bifunctional chromophore in a supporting matrix, a higher photorefractive figure-of-merit Recently the IBM group reported on a family of dihydropyri- (large EO coeYcient and photoconductive eVect) and faster dines 1–8.response (higher charge mobility) can be simultaneously obtained. Several examples of using bifunctional chromophores in photorefractive composites have been reported by Stankus et al.,53 Zhang et al.56 and Bolink et al.57 In all studies, C60 was used as a photocharge generation sensitizer to extend the photoconductivity to longer suitable wavelengths. Inert polymers poly(methyl methacrylate) (PMMA) for 4¾-(N,N¾-di-ptolylamino)- b,b-dicyanostyrene (DTADCST), poly(n-butyl methacrylate) (PBMA) for 4-(N,N¾-diphenylamino)-b-nitrostyrene (DPANST) as well as 4-(N,N¾-diphenylamino)-4¾-nitrostilbene (DPANS) and polystyrene (PS) for monodicyanoethene–TPD (MDCETPD) were used as host matrices, respectively.Among these bifunctional chromo- These molecules can be used to formulate organic photorefracphores, chromophore MDCETPD is found to be fully amorph- tive materials in which the chromophore serves as a charge ous, with a Tg of 110 °C. In order to enhance the orientational transporting function and the optical birefringence species, and degree of MDCETPD at room temperature, the plasticizer additionally acts as the amorphous host.19,59 Most of the work dioctyl phthalate (DOP) was added to decrease Tg.57 All reported here was done on methyl N-isobutyl-2,6-dimethylcomposites with a certain component ratio show low Tg.These 4H-pyridin-4-ylidene(cyano)acetate 4 (2BNCM) which exhiblow Tg materials allow the measurements of the photorefractive its a Tg at 25 °C and, at room temperature, persists without eVects to be carried out at room temperature.Results indicate crystallization. When a small amount (<1 wt%) of a photosensitizer, TNF, was added to these glassy monomers, the resulting that the photorefractive eVects (the two-beam coupling gains, J.Mater. Chem., 1998, 8(4), 809–828 813material exhibited outstanding photorefractive eVects such as materials are expected to exhibit a faster response. Thiophene copolymer, poly[3-octylthiophene-co-N-(3-thienyl)-4-amino- high diVraction eYciencies and two-beam coupling gain. 2-nitrophenol] (POMDT) with second-order NLO chromophores used as a bifunctional host for photorefractive materials has been reported.65 To 61 wt% of POMDT was added 30 wt% of diisooctyl phthalate (DOP) as a plasticizer, 7.6 wt% of 3-cinnamoyloxy-4-[4-(N,N-diethylamino)-2-cinnamoyloxyphenylazo] nitrobenzene (CNNB-R) as a chromophore to enhance the linear EO eVect and 1.4 wt% of TNF as a sensitizer. The linear and second-order NLO properties, photoconductivity and photorefractivity for this material were investigated; a relatively large two-beam coupling gain coeYcient of 24.5 cm-1 was obtained.Carbazole photorefractive composite materials It is well known that organic photorefractive materials consist of two main functional components: a charge transporting agent and a linear EO agent. Carbazole polymers, such as poly-n-vinylcarbazole (PVK), are well known as exhibiting good charge transport properties and their photocharge generation eYciency can be sensitized by the formation of a charge transfer complex between the carbazole moiety and acceptor molecules, such as TNF.66–68 It was found that significant increases in the hologram growth rate can be obtained by doping the chromophore with a small amount of inert polymers, such as 10 wt% of PMMA and polysiloxane, or polyvinylcarbazole. 2BNCM–PMMA–TNF (9059.750.3 wt%) was selected for a more detailed study on photorefractivity. High diVraction eYciency of 80% and twobeam coupling gain of 69 cm-1 were obtained from this system at an external field of 40 V mm-1. High density holographic digital data storage has also been studied in this system.Fully functional polymers In the last few years fully functionlized polyurethanes,60 conjugated polymers61 and polyimides62 with relatively high Tg have been developed by Yu et al. for photorefractive applications.18 Photoconductive, EO and photorefractive studies have been done on these polymers. Two-beam coupling gains without applied electric fields were obtained from one high Tg polymer. 62 Recently a new conjugated polymer with an ionic tris(bipyridyl)ruthenium complex as the charge generating species has been synthesized. A very large net optical gain of 200 cm-1 was obtained at a zero external electric field. However, four-wave mixing diVraction eYciency was still low.63 Bifunctional polymers It is well known that conjugated polymers, due to their delocalized p-electron distribution, have much large carrier mobilities.64 Thus, conjugated polymer-based photorefractive 814 J.Mater. Chem., 1998, 8(4), 809–828the poling electric field and a linear dependence was obtained up to 100 V mm-1, with r33=4.2 pm V-1 at 62.5 V mm-1. The photoconductivity sensitivity was measured as 5.4×10-11 (V cm)-1/W cm-2) with 703 nm illumination at 62.5 V mm-1.Photorefractive gratings are written at a wavelength of 703 nm and four-wave mixing diVraction eYciencies as high as 2% have been achieved in films less than 200 mm thick. Asymmetric energy exchange was also observed and a twobeam coupling gain of 7 cm-1 was obtained. The application of this material to information storage and image processing has been explored with demonstrations of holographic image recording and retrieval.76 Two-beam coupling net gain of organic photorefractive materials was first obtained from a PVK-based composite material.77 The photoconducting polymer PVK was doped with the optically nonlinear chromophore 3-fluoro-4-N,N- PVK with charge transporting functionality has been widely diethylamino-b-nitrosyrene (F-DEANST) (33 wt%) and sensit- used as the charge transporting host component for organic ized for charge generation with TNF (1.3 wt%).The fluoro- photorefractive materials.14,17 An initial report of photorefracsubstituted second-order NLO chromophore was used due to tivity in a carbazole system was based on a PVK polymer its absorption at relatively short wavelength. This composite composites in 1992.69 There has since been increasing interest exhibited Tg at about 40 °C, considerably below the Tg of pure in such systems owing to the native charge transporting ability PVK (212 °C), owing to the plasticization by the NLO chromo- of PVK.The high performance of the photorefractive eVects phore and residual solvent. A small amount of TNF was (near 100% diVraction eYciency for the readout of a hologram added, which forms a charge transfer complex with PVK,66,78 as well as more than 200 cm-1 net two-beam coupling gain) to provide long wavelength photosensitization.The PVK–F- have been observed in some composite materials based on DEANST–TNF system exhibits diVraction eYciencies as high PVK charge transporting polymers.70,71 To date, many photoas 1%, at least two orders of magnitude larger than previously refractive composites based on carbazole polymers, multifuncreported values for any organic photorefractive polymer, and tional polymers with carbazole functional components, carbamore importantly, this material is the first organic to show net zole sol–gel systems, amorphous carbazole oligomers and internal two-beam coupling gain (>10 cm-1).Grating growth carbazole dendrimers for photorefractivity have been times of the order of 100 ms were observed. These photorefrac- developed. tive values make PVK–F-DEANST–TNF comparable with some of the inorganic photorefractive materials,79 such as Bi12SiO20 and BaTiO3. After this observation, various diVerent PVK as a charge transporting host second-order NLO chromophores were examined as a functional dopant in a PVK–TNF matrix.80 It was found that Zhang et al.69 reported the first carbazole organic composite system for photorefractive materials.C60 (0.48 wt%) as a doped PVK composite exhibits enhanced performance of photorefractive eVects compared with the previously reported photosensitizing molecule and 4-N,N-diethylamino-b-nitrostyrene (DEANST) (32 wt%) as a second-order NLO photorefractive polymers.This improvement in performance is partly due to the formation of larger space charge fields and chromophore were doped into PVK. Silence et al.39 also used C60 as a sensitizer due to the presence of multiple stable the higher nonlinearity that results from the application of larger poling electric fields.However, the large photoinduced reduced state72 and high triplet yield.73 C60 can form chargetransfer complexes with PVK. For PVK doped with C60, field refractive index changes in these new materials cannot be explained by the simple models based on the EO photorefrac- dependent photocarrier generation followed the Onsager model74 with very large generation eYciency above 106 tive eVect which have worked so well for previously known photorefractive materials.V cm-1. The high solubility of the C60 and the presence of absorption extending throughout the visible led to the In order to explain the large photoinduced refractive changes, a new orientational enhancement mechanism81,82 was evaluation of C60 as a sensitizer for photorefractive materials.The photorefractive properties as a function of C60 concen- applied to doped PVK77,80 and other systems.83 The enhancement relies on the ability of the NLO chromophores to be tration have been studied.39 The EO coeYcient and the photoconductivity of this PVK-based material were deter- aligned not only by the externally applied electric field but also in situ by the space charge field itself during grating mined.Four-wave mixing signal from a 100 mm thick film was measured. It was found that when no electric field was formation. The resulting periodic poling of the sample leads to a modulation of the birefringence of the materials and to a applied, no diVracted signal was detected. When an external field was applied, a diVracted signal was built up.The strong modulation of EO response, the combination of which contributes favorably to the diVraction eYciency fields in the appro- field dependence of the four-wave mixing diVraction eYciency can be explained by the photorefractive eVect. This depen- priate polarization. A nonlocal photorefractive eVect was also observed in a low Tg non-EO PVK-based composite.84 It was dence was due to the enhancement of photoconductivity and linear EO coeYcient by the applied electric fields.A maximum shown that the eVects of orientation birefringence and modulation of EO coeYcient resulting from the periodic local field diVraction eYciency of 2×10-5 was obtained at an external field of 50 V mm-1. induced by the orientation of dipolar chromophores have to be properly taken into account to describe the formation of Thiopyrylium dye (TPY) as a new photosensitizer was also used in PVK–DEANST systems.75,76 The composite con- the refractive index grating.N,N-Diethyl-substituted para-nitroaniline (EPNA) (39 wt%) tained 79 wt% of PVK, 20.8 wt% of DEANST and 0.2 wt% of TPY. Due to the high concentration of DEANST, the as a second-order NLO chromophore was used to dope the well known photoconductor PVK sensitized with 0.1 wt% of composite showed a Tg of 53 °C.The DEANST chromophore can be electrically poled at room temperature in this low Tg TNF.85 For the alignment of the EPNA molecules, which is essential for the activation of the EO eVect, corona poling with composite. This could be confirmed by the observation of EO modulation at room temperature, when a dc electric field was a tungsten needle is used.This poling technique oVers the advantage of easier sample preparation compared with using applied. The EO coeYcient was measured as a function of J. Mater. Chem., 1998, 8(4), 809–828 815a sandwich cell. The two-beam coupling experiments provided Secondly, it increases the solubility of the dye in PVK inhibiting undoubted evidence of photorefractivity for this PVK– crystallization.The resulting composite can reproducibly form EPNA–TNF. A phase shift of 90° was obtained from the stable, optically transparent films with good reorientation modulation of the transmitted power of the two beams when mobility. No separate plasticizing agent is required. The photothe voltage across the sample increased.Asymmetric energy refractive composite consisted of PVK–EHDNPB–TNF in a transfer between the two beams was observed when the corona mass ratio of 4455551. This materials exhibited a 60% device voltage was switched on. The two-beam coupling net gain of steady-state diVraction eYciency and 120 cm-1 two-beam 18 cm-1 and response time were measured on a 65 mm thick coupling gain, well in excess of its absorption coeYcient of film at an electric field of about 10 kV. 3.5 cm-1, at a wavelength of 676 nm. Control of charge trapping has great technological impor- Sandalphon et al.101 reported dual-grating formation through tance: optimization of charge trapping might be the route photorefractivity and photoisomerization in azo-dye-doped toward the longer storage times and high diVraction eYciencies PVK–TNF polymeric composite.TNF as a sensitizer and required for optical data storage applications. Based on the disperse red I (DR1) as a second-order NLO chromophore PVK–EPNA–TNF system, studies of the modification of the were doped in the PVK host. Photorefractivity and photoisotrap density have been carried out by Malliaras.86,87 The merization in polymer composites have been studied.Both types response time and the phase shift of the photorefractive grating of hologram were observed and reversible. The photorefractive as a function of doping with various amounts of DEH have diVraction eYciency is strongly dependent on the external been studied. Measurements indicate that at low concen- applied field, as expected for other photorefractive polymers. trations, DEH acts as a trap due to its lower ionization High diVraction eYciencies and storage times of several minutes potential, while at higher concentrations a new charge trans- have been demonstrated.The photoisomerization diVraction port pathway through hopping between DEH molecules is eYciency was highest in the absence of an applied field and established.This behavior has also been observed in the case showed some reduction as an external field was applied. of PVK doped with TPD.88 Due to this photoconductive Photoisomerization may turn out to be a useful way to enhance behavior, the response time and the phase shift of photorefracthe photorefractive properties of these azo-dye-doped polymeric tive gratings decrease with a small amount of DEH and composites for erasable optical storage applications.increase at relatively higher concentrations of DEH. A small Multiple-grating formation was reported in photorefractive amount of DEH acts as a trap, decreasing the mobility of composite polymers.102 In this material, PVK–TNF was doped holes in PVK. As the response time is proportional to the with two second-order NLO chromophores, (E)-4-N,N-diethylmobility, photorefractive response time follows the same trend.aminocinnamonitrile (DEACST) and DR 1. DR 1 has been When the concentration of DEH exceeds 1 per 100 carbazole proven to show strong photoisomerization grating formation. units, DEH begins to contribute to dominant charge transport It was present in the composite at a very low concentration, and the response time increases. The dependence of the phase so that it contributed negligibly to the formation of photoref- shift on the DEH concentration can be understood according ractive gratings.The composition of the mixture by mass was to the standard theory of photorefractivity.89 in the ratio 10054050.550.5 (PVK–DEACST–DR 1–TNF). The transient behavior of the photorefractive grating in the Similar results to those observed by Sandalphon et al.101 could PVK–EPNA–TNF system doped with various amounts of be obtained; photorefractive gratings, and photoisomerization DEH has also been studied.90,91 The influence of the trap intensity and polarization gratings were observed in this mate- density induced by DEH on the hole drift mobility was directly rial system, essentially a photorefractive polymer ‘seeded’ with measured.Study of the transient behavior of the photorefracan azo-dye to allow the formation and separate optimization tive gratings in polymers is very interesting, because charge of photoinduced gratings. transport in these materials is very diVerent from that in inorganic crystals, displaying a highly dispersive character, and strong temperature and electric field dependence.92 The holographic time-of-flight technique was applied in a photorefrac- PVK as a charge transporting host with the plasticizers tive polymer composite system.93,94 In this composite, PVK as a charge transporting host was doped with 40 wt% of 4- Most organic photorefractive materials reported so far are (hexyloxy)nitrobenzene (HONB) as a second-order NLO host–guest polymeric composites.To break the centrosymmechromophore and 0.1 wt% of TNF as a sensitizer. Two-beam try of the materials and to obtain a macroscopic EO response, coupling experiments on a 100 mm thick film at 633 nm have the second-order NLO chromophores have to be aligned by shown the PVK–HONB–TNF composite to be purely photoan applied electric field (poled).14,17 The eYciency of the refractive, giving rise to asymmetric energy exchange with a poling process is strongly dependent on the orientational gain coeYcient of 3 cm-1 at 70 V mm-1.The measured electric freedom of the second-order NLO chromophores. At tempera- field, temperature and drift length dependencies of the holetures below the Tg, the polymer chains are frozen and the drift mobility in this composite are in agreement with literature orientational mobility of the chromophores is very low.As data for PVK.95 the temperature is raised close to the Tg, the orientational Based on PVK–EPNA and PVK–HONB, much detailed freedom of the chromophores increases, allowing eYcient research work on the space charge field formation96 and design poling.When large amounts of second-order NLO chromo- of functional components such as the role of absorbing NLO phores are doped in PVK, the Tg of PVK (200 °C) is chromophores22,97 has been done also by Malliaras et al. substantially lowered and the chromophores can be oriented Recently they summarized their research work on the mechaneven at room temperature.Recently a large increase in the ism of photorefractivity and unique photorefractivity in PVKphotorefractive performance of a PVK-based polymer com- based composite materials.98 posite was observed after the incorporation of additional An azo derivative 1-(2¾-ethylhexyloxy)-2,5-dimethyl-4-(4¾- plasticizer molecules.70,71 This was caused by an increase in nitrophenylazo)benzene (EHDNPB) has been used to provide the orientational mobility of second-order NLO chromo- an EO response in PVK–TNT polymer composite.98 This type phores, due to the additional lowing of the Tg, resulting in a of azo chromophore was first used by Kippelen et al.100 In higher net alignment and hence a larger EO eVect.Moreover, this case, the chromophore has been modified to incorporate in such systems, the chromophores are reoriented under the a racemic ethylhexyl group.This nonpolar functionality has influence of the space charge field.81 In this way, a number of two important roles. First, it renders the dye a plasticizer and as such activates the orientation enhancement mechanism.81 polymeric composites with excellent performance, which 816 J.Mater. Chem., 1998, 8(4), 809–828approaches or even exceeds that of existing inorganic mate- matrix and a large dipole moment of the NLO chromophore. rials, have been reported.14,17 Two kinds of plasticizers, inert At a field of 140 V mm-1, the EO coeYcient is 37.6 pm V-1, plasticizers and a charge transporting functional plasticizer, which represents a major increase of the EO activity compared have been used.to the related higher Tg composite without plasticizer, PKV–C60–DEANST.69 Both the diVraction eYciency and the Inert plasticizer. Since an electric field is always applied to two-beam coupling gain in low Tg polymeric materials are facilitate photocharge generation, eYcient photorefractive highly dependent on the applied electric field.The electric field eVects in low Tg PVK-based polymeric composites should be plays a crucial and multifunctional role: it forces a noncentroreasonable. The eYcient electric field-induced alignment of the symmetric alignment of second-order NLO chromophores,105 it chromophore can be obtained in these low Tg polymer com- enhances the quantum yield of photocarrier generation,106 and posites.In order to obtain low Tg composite, the material it assists in transport of the photogenerated carriers.107 containing PVK and about 0.56 wt% of C60 was doped with DiVraction eYciency as high as 40% and asymmetric net two- 25 wt% of DEANST and 20 wt% of a chemically inert plas- beamcoupling gain coeYcients in excess of 130 cm-1, surpassing ticizer, dibutyl phthalate (DBP).103 those of known inorganic single crystalline photorefractive media, were obtained from this composite at an applied electric field of 110 V mm-1.More recently, the IBM group reported a high performance photorefractive polymer based on PVK doped with the secondorder NLO chromophore, 4-piperidin-4-ylbenzylidenemalononitrile (PDCST), the liquid plasticizer butyl benzyl phthalate (BBP) and C60 for increased charge generation at longer wavelengths.71 In comparison to the use of the crystal plasticizer, BBP showed dramatic suppression in crystallization.This photorefractive composite, PVK–PDCST– BBP–C60, with a mixing ratio of 49.553551550.5 wt% showed Tg at 28 °C. The electric field dependence of photorefractive eVects was measured.This photorefractive polymer composite with improved material stability exhibited a high two-beam coupling gain coeYcient of 200 cm-1, and a fast response time of 50 ms at 100 V mm-1 and at 1 W cm-2. Overmodulation of the diVraction eYciency and a high sensitivity of about 3 cm3 kJ-1 were also obtained from this photorefractive material. All the photorefractive properties of this material compare with those of the faster inorganic crystals.14 Temperature dependence studies of the photorefractive eVect in PVK–TCP–DEANST–C60 prepared in the mixtures in the mass ratio 60.0353653.7550.22% have been carried out in the temperature range 20–75 °C.108 It was found that as the temperature of the sample increases, an increase in the photorefractive figures of merit, such as diVraction eYciency, two-beam coupling In situ electrical poling experiments on low Tg PVK– gain coeYcients and speed of grating formation and erasure, DEANST–C60 doped with DBP demonstrated that at room are seen.At elevated temperatures, decreased rigidity of the host temperature the SHG signal strongly increases with the DBP matrix leading to increased orientational mobility of the second- concentration.Moreover, the response time of the NLO active order NLO chromophores results in enhanced photorefractive DEANST molecules to the poling field was observed to decrease performance. When the applied electric field was fixed at 90 V into the subsecond regime in the PVK–DBP matrix. The mm-1, the diVraction eYciency increased with temperature. For composite shows a relatively high field-dependent EO coeYcient temperatures above 37 °C, the diVraction eYciency approached and quantum yield of photocarrier generation.The photorefrac- 100%. Also an increase in the two-beamcoupling gain coeYcient tive diVraction eYciencies obtained from this plasticized comfrom 55 cm-1 at 25 °C to 95 cm-1 at 40 °C was obtained. The posite compare well with those of well-known inorganic observed increase in the speed of erasure with temperature is materials and reached 1.5% at a applied electric field of about consistent with the results of measurements of the quantum 40 V mm-1.The four-wave mixing diVraction eYciency was yield of photocarrier generation and the carrier mobility for the found to be strongly electric field dependent.This behavior polymeric composite discussed. At a poling field of 9 V mm-1 originates in the field dependence of EO activity of the material and temperature of 20 °C, the quantum yield of photocarrier and the space charge field formation. The photorefractive nongeneration was of the order of 10-6, and the carriers’ mobility local character of the observed eVect in the sample was finally was found to be 10 cm2 V-1 s-1.When the sample was heated confirmed by performing a two-beam coupling experiment. A to 58 °C, the quantum yield of photocarrier generation increased two-beam coupling gain of 4 cm-1 was also obtained at electric fourfold and the hole mobility increased by more than one field of 40 V mm-1. The composite shows very fast kinetics of order of magnitude. In order to achieve a broad spectral photorefractive grating formation and erasure, occurring in the response, a new second-order NLO chromophore, 4-[N- millisecond time scale.(2-hydroxyethyl)-N-methylaminophenyl]-4¾-(6-hydroxyhexyl- A liquid inert plasticizer, tricresyl phosphate (TCP),104 was sulfonyl)stilbene (APSS) with high transparency over a broad also used in the PVK–C60–DEANST composite system for wavelength was used together with a PVK–TCP–C60 matrix.108 enhancement of photorefractive performance.This composite The composite PVK–TCP–APSS–C60 was prepared with a contains PVK, 36 wt% of TCP, 3.75 wt% of DEANST and mass ratio of 47.7547.654.550.2%. This composite with low T g 0.22 wt% of C60. The Tg of PVK–TCP–DEANST–C60 was shows high photorefractive figures of merit at 488, 514.5 and determined to be lower than 14 °C.Due to the low Tg of the 632.8 nm. DiVraction eYciencies up to 40%, as well as net composite, the chromophore with a large dipole moment could two-beam coupling of 60 cm-1 have been achieved at these be perfectly aligned at room temperature. A large EO coeYcient was obtained due to eYcient plasticization of the host polymeric operating wavelengths.J. Mater. Chem., 1998, 8(4), 809–828 817N-Ethylcarbazole as a charge transporting plasticizer. poling field in the photorefractive material is essentially dc for According to charge transporting functional requirements of four-wave mixing, the large contribution from the poling photorefractive materials, Kippelen et al.first used N-ethylcar- birefringence to the total refractive index modulation in the bazole (ECz) as a charge transporting functional plasticizer.100 DMNPAA doped composite more than compensates for the In comparison to the use of inert materials as plasticizers, ECz smaller contributions arising from the linear EO and the Kerr shows important charge transporting for photorefraction and eVects.This technique should be very useful for the screening allows one to keep charge transporting functional moieties at of many candidate materials in the search for those with a high ratio. We believe that further improvements in stability optimized performance. In traditional photorefractive matewill be achieved when suitable liquid or amorphous functional rials, the refractive index modulation arises solely from the plasticizers are developed.Based on carbazole polymer PVK, space charge field acting on the EO eVect of the materials.1 one composite system containing ECz as a functional plas- Thus, the photorefraction in high T g polymers is related purely ticizer has been reported to exhibit high photorefractive eVects to EO eVects.In contrast, photorefractive polymers with a low by Meerholz et al.70 Much research work on photorefractive T g exhibit an orientational contribution to the refractive index materials followed this report using the same type of second- modulation. The chromophore design approach to orienorder NLO chromophores99,109,110 or the same carbazole tational enhancement of photorefractive has recently become functional plasticizer.103,109,111 an interesting research topic.49,116 ECz plasticized PVK-based photorefractive system doped To date, among organic photorefractive materials, the PVK– with 2,5-dimethyl-4-( p-nitrophenylazo)anisole (DMNPAA) as DMNPAA–ECz–TNF composite is the best photorefractive an EO chromophore was first reported by Kippelen el al.100 polymeric material.However, there are inherent problems of As the dipoles of the second-order NLO molecules are ran- phase separation with such a high density of functional compodomly oriented, there is a priori no overall EO eVect in these nents in the multi-component system. In order to improve materials. Second-order NLO properties of the bulk material long-term stability of this high performance photorefractive can only be induced by poling the chromophores by an composite system, two approaches have been tried.First, a external electric field. Azo-dye-doped PVK composites exhib- racemic ethylhexyl group has been incorporated into the azo ited rather low T g, because the chromophores act as a plas- chromophore.99 Secondly, the eutectic mixture of the two ticizer for the PVK matrix, i.e.with increasing dye isomeric azo second-order NLO chromophores, DMNPAA concentration the glass transition temperature decreases. and 3,5-dimethyl-4-( p-nitrophenylazo)anisole (3,5DMNPAA) However, the concentration of the dye is limited due to could considerably lengthen shelf-life.117 Devices using this crystallization. In order to overcome this problem and further eutectic composite remain clear for over one year under decrease T g at around room temperature, ECz was added as ambient laboratory conditions.The internal performance of a charge transporting plasticizer. The chromophores were then these improved materials was found to be comparable to that aligned by the dc electric field that is applied for photorefractive of the best materials known previously.and photoconductive recording. The charge transfer complex The photorefractive polymeric composite DMNPAA–PVK– between PVK and TNF provides photosensitivity in the visible ECz–TNF shows diVraction eYciencies as high as 86%, twospectrum. Compared with other azo dyes, such as DR 1, the beam coupling gain coeYcients of more than 200 cm-1, refracbest results were obtained for PVK–ECz–TNF doped with tive index modulations up to 7×10-3, good sensitivity, and DMNPAA.112 A suitable second-order NLO chromophore azo reasonably fast response times (about 0.5 s) and is useful as a dye DMNPAA with an absorption band at relatively short recording medium for dynamic holographic interferometry and wavelength was selected.The performance of PVK– pattern recognition.70,118–121 Devices can be operated with DMNPAA–ECz–TNF was improved considerably by inexpensive low-power laser diodes, unlike many other holo- Meerholz el al.70 to 86% diVraction eYciency and a net two- graphic recording materials that require expensive high-power beam coupling gain of more than 200 cm-1.This performance laser systems. The device performance can be easily adjusted by far surpasses that of the other organic photorefractive by means of an external voltage. Information is reversibly materials reported to date.High performance in photorefrac- stored in this polymer. Thus, storage, readout, and erasure can tive eVects was obtained from this material due to the usage be carried out such that one device can be used for real-time of an appropriate second-order NLO chromophore, plasticizer monitoring with no additional intermediate developing steps.and component ratio. PVK–DMNPAA–ECz system doped This photorefractive polymeric material has been successfully with C60 as a photosensitizer was also demonstrated to exhibit demonstrated in eYcient optical image processing applihigh performance photorefractive eVects.113 cations.120 Furthermore, the use of such high performance The birefringence, the linear EO and Kerr properties of low photorefractive polymers in an optical pattern recognition T g PVK–DMNPAA–ECz–TNF photorefractive polymers system for security verification has also been demonstrated have been measured by a simple frequency-dependent ellipsorecently. 123 Good quality interferograms of the mode patterns metric technique.114,115 The birefringence induced by the orienof a vibrating membrane with good fringe contrast were tation of chromophores in this low T g composite polymer obtained by using inexpensive laser sources, such as laser plays a major role in the overall field-induced refractive index diodes or HeNe lasers.120 change. For the first time, Kerr contributions in photorefractive A systematic study of the eVect of plasticization on the polymers were identified.This study clearly shows that moduphotorefractive performance of PVK based polymeric com- lated birefringence is responsible for the high steady-state posites was also reported.124 It is shown that ECz can be used diVraction eYciencies measured in four-wave mixing experias an eYcient plasticizer, leading to a large increase in the ments in PVK–DMNPAA–ECz–TNF due to the orientation gain coeYcient and the diVraction eYciency, which arises of DMNPAA.In order to compare the contribution from the solely due to an improvement in the orientational mobility of birefringence with other second-order NLO chromophores, Fthe NLO chromophores and was not caused by the alteration DEANST was used.114 It was found that the DMNPAA– of the space charge field.Phase separation was observed in PVK–ECz–TNF composite has a much higher orientational samples with a high ECz concentration. birefringence contribution to the total refractive-index modulation at low ac frequencies compared with that of the FDEANST –PVK–ECz–TNF composite. The diVerence in orien- Carbazole polysiloxanes as charge transporting hosts tational birefringence can explain the higher four-wave mixing Zobel et al.110 first reported photorefractive materials based diVraction eYciency observed for the DMNPAA composite over that of the F-DEANST composite.Because the local on carbazole polysiloxane (PSX). 818 J. Mater. Chem., 1998, 8(4), 809–828conductivity have been observed.Holographic four-wave mixing and two-beam coupling experiments proved the photorefractive nature of the sol–gel composite DHD–Cz–TNF. A large number of carbazole-containing polymers have been prepared.125–127 Among these carbazole polymers, PSX shows relatively low T g, and the T g could be controlled in a very wide range from -45 to 51 °C by introducing diVerent spacer lengths.PSX with a T g at 51 °C was selected as a host for photorefractive materials due to the relatively high ratio of charge transporting function. A composite was obtained from a mixture of PSX (56 wt%), the second-order NLO chromophore DMNPAA (43 wt%) and TNF (1 wt%). Due to the low T g of this PSX composite, no plasticizers were necessary in this case.A laser beam of wavelength 650 nm was used for four-wave mixing and two-beam coupling measurements. A large diVraction eYciency of 60% and net optical gain of 220 cm-1 were obtained from this composite. According to the behavior of both the phase and amplitude of the resulting refractive index grating, the photoisomerization and the photorefractive grating are competitive.However, at electric fields larger than about 10 V mm-1, the refractive index change due to the photorefractive grating exceeds that of the photoisomerization grating. It can be shown that with this composite material the orientational eVects of the DMNPAA dominate the photorefractive eVect. The authors pointed out that in order to obtain pure electronic photorefractive eVects, the photoconductive and the second-order NLO properties of this material must be further improved.The same carbazole PSX and electron acceptor molecule TNF have also been used in photorefractive composites by the IBM group.128 In this case, 33 wt% of F-DEANST was used as a second-order NLO chromphore. Photorefractive and optical data storage measurements were carried out on this composite.High photorefractive performance was also demonstrated. PSX composites exhibit excellent optical clarity and low optical scattering characteristics. These optical properties are necessary in high density holographic digital data storage applications. The utility of PSX based photorefractive polymers for storage applications was demonstrated by recording digital data at a density of 0.52 Mbit cm-2 and reading it back The materials show a low diVraction eYciency (0.01%) and a without error up to 5 min after recording.128 small two-beam coupling gain coeYcient (0.3 cm-1). In Chaput’s case (DR 1–Cz–TNF), noncentrosymmetric chromophore orientation and a large stable EO response (15 pm V-1) Carbazole sol–gel photorefractive materials were obtained.A large two-beam coupling gain of 110 cm-1 was obtained at a wavelength of 633 nm. However, comparison Sol–gel processed optically transparent silica based materials with the absorption coeYcient of 450 cm-1 means that a net are a new generation of multifunctional molecular composites gain cannot yet be achieved. Such a two-beam coupling gain used in the design of optical devices.129,130 Sol–gel technology has been obtained without applying external electric fields provides an attractive route to the preparation of rigid amorphduring the two-beam coupling measurements.Although these ous inorganic oxide glass matrices at ambient temperatures, in materials show a low diVraction eYciency and non-net two- which dopants such as inorganic and organic functional moietbeam coupling gain coeYcient, the sol–gel processing of com- ies can be successfully incorporated.A wide variety of unique posite, seems to be a promising approach towards the prep- functional composites can be formed using sol–gel inorganic– aration of photorefractive materials. Sol–gel processing oVers organic processing techniques. In the last few years, studies materials with excellent optical qualities and the potential for have also focused on NLO sol–gel systems.131,132 Two sol–gel retaining the orientation of chromophores during the formation processed composites exhibiting photorefractive eVects have of the sol–gel by applying an electric poling field.recently been reported by Burzynski et al.133 and Chaput et al.134 In both cases, a carbazole derivative (Cz) was used as a charge transporting agent and formed a charge transfer Carbazole multifunctional photorefractive polymers complex with TNF to facilitate photocarrier generation at the visible wavelengths. 4-[N,N-Bis(b-hydroxyethyl)amino]-4¾- Many types of multifunctional functional polymers have been developed for photorefractive materials to suppress further nitrostilbene derivatives (DHD) and DR 1 were used as secondorder NLO chromophores by Burzynski et al.and Chaput phase separation. Among these polymers, polymers containing the carbazole moiety as a charge transporting agent, bifunc- et al., respectively. In Burzynski’s case, EO eVects and photo- J. Mater. Chem., 1998, 8(4), 809–828 819tional chromophores or multifunctional chromophores have absorption changes due to permanent photobleaching are domibeen reported.135–139 Some multifunctional polymers with both nant in the absence of any external electric eVect for this polymer.EO and charge transporting moieties must be doped with a The dynamics of the erase–write behavior of gratings was also suitable sensitizer. However, phase separation cannot observed studied in these polymers. Due to low mobilities, the material owing to the very low concentration of the dopants.has a slow response. However, it is believed that mobilities can The first fully functional photorefractive polymer was devel- be increased and the absorption grating can be avoided by using oped based on carbazole moieties [polymer (a)]135 in 1992. suitable sensitizers and a longer wavelength of a laser.PMMA-like polymer (b), DCVANMA–CzEtMA–EA In this polymer, some of the carbazole groups were tricyanovinyl- (255:3)137 doped with 2.3 wt% of TNF for photorefractive ated. Carbazole substituted with tricyanovinyl group had two materials has also been reported. In this system, carbazole can functions: second-order nonlinearity and photocharge generform charge transfer complexes with TNF.This complex acts ation. Carbazoles with no acceptor groups could oVer a charge as a photocarrier generation function at a wavelength of transporting function. A photoconductivity of 9.8×10-10 633 nm due to its absorption coeYcient at the same wavelength. V-1 cm-1 and EO coeYcient of 6.1 pm V-1 were measured. DiVraction eYciency was of the order of 10-6; the change in Evidence for absorptive and photorefractive gratings has been the refractive index was reversible in this polymer system.It obtained by four-wave mixing experiments and EO measurewas found that the observed photorefraction was mainly due ments.136 The photorefractive grating was studied by investigating the electric-field dependence of the diVraction eYciency.The to the molar refraction change in the carbazole moiety caused 820 J. Mater. Chem., 1998, 8(4), 809–828by photoinduced ionization. This provides evidence that Since the first observation of photorefractive eVects in amorphous organic polymeric composites, many amorphous trapped ion radicals exist in the polymer forming grating, which is necessary for real photorefractivities.organic materials for photorefraction have been developed based on two molecular design approaches: the guest–host Another low T g polymer PENHCOM138 [polymer (c)] based on the PMMA structure doped with 0.2 wt% of TNF has also composite approach and the fully functional approach. In these two approaches, materials always contain multicom- been developed. In this polymer, 4-[N-ethyl-N-(2-hydroxyethyl) amino-4¾-nitrostilbene] was used as a second-order NLO ponents to oVer multifunction properties.Development of bifunctional chromophores is the first approach to try to chromophore and the carbazole moiety was used as a charge transporting agent. In order to obtain a copolymer with a low develop one chromophore with more than one function.53,55 High performance of photorefractive eVects has been obtained Tg, long aliphatic octyl chains were attached to the side chain as a plasticizer. The ratio of the functions X5Y5Z is 17530553.using this design approach. These chromophores provide two main functions, such as EO activity and suYcient charge This material showed an absorption coeYcient of 25 cm-1 at a wavelength of 633 nm.A four-wave mixing diVraction transport properties for photorefractive behavior; they also provided charge trapping capability which allowed the first eYciency of 0.9% and a two-beam coupling gain coeYcient of 7.5 cm-1 have been obtained at an electric field of 100 V mm-1. demonstration of truly long-lived gratings in a photorefractive polymer, quasi-nondestructive readout.57 These materials also More recently, polymers containing a single multifunctional carbazole chromophore [polymer (d)] have been obtained in exhibited improved optical quality due to the reduction in the number of dissimilar constituents. Recently, the IBM group our laboratory.139 In this polymer, carbazole substituted with two acceptor groups exhibited multifunctional properties.The reported that a family of related dihydropyridines can be used to formulate organic photorefractive materials which represent carbazole chromophores lie parallel to each other, in a ‘shoulder-to-shoulder’ arrangement. In this arrangement, align- significant advances in many respects.19 The most novel aspect of these chromophores is that in high concentrations they form ment of dipole moment is more readily achievable by applying an electric field than in the structures where dipole moments stable organic glasses instead of microcrystallites.In this case, there is no need to devote sample volume to a host polymer are pointing along the polymer main chain. The T g was strongly dependent on the length of the alkyl spacer between or to plasticizing constituents because the chromophores form optical quality films even at 100% chromophore concentration. the carbazole chromophores and of the alkyl group at the 9- position of the carbazole ring.The T g could be controlled in These chromophores show three functions: second-order NLO properties, charge transporting properties and glass state for- the range from 35–87 °C.Among these carbazole main-chain polymers, polymers with a relatively low T g enable photorefrac- mation properties. High performance photorefractive eVects have been obtained from these so-called photorefractive glass tive measurements to be made at room temperature. These carbazole main-chain polymers have proved to have both molecules doped with small amounts of a sensitizing agent, such as C60 or TNF.photoconductivity and EO activity.140 The photorefractive properties of the carbazole main-chain polymers were studied Recently, we developed fully amorphous chromophores for photorefractive materials.20,148 These chromophores were syn- by four-wave mixing and two-beam coupling techniques. The two-beam coupling gain of 14 cm-1 was obtained at an applied thesized based on carbazole building blocks.The design approach is shown in Fig. 1. The chromophores combining electric field of 23 V mm-1, with an absorption coeYcient of 8 cm-1. The photorefractive gain at this electric field was photoconductive and EO functions are plasticized by introducing a suitable flexible alkyl chain. The use of diVerent plastic larger than the absorption coeYcient.A net gain of 6 cm-1 was obtained from the carbazole main-chain polymer. A chains can provide us with a chance to obtain amorphous compounds with controllable T g. We believe that our carba- diVraction eYciency of about 1.5% was also obtained with the same electric field. zole-based design approach has more flexibility for chemical modification, and diVerent types of materials can be obtained.Carbazole main-chain polymers with additional functional moieties [polymer (e)] exhibit more eYcient photorefractive Amorphous carbazole conjugated oligomers and amorphous carbazole dendrimers have been developed for photorefractive eVects141 compared with pristine polymer.139 It was found that the polymer with a carbazole moiety as a charge transporting materials in this way.A conjugated carbazole structure was used to produce photorefractive materials because of its excel- functional side group showed a net two-beam coupling gain of about 45 cm-1 and a diVraction eYciency of 2.5% at an lent charge transporting properties and relatively high carrier mobility of the conjugated carbazole polymers and applied electric field of 23 V mm-1. Carbazole main-chain polymers with an additional second-order NLO chromophore oligomers.149 in the side chain exhibited enhanced second harmonic coeYcients. Carbazole photorefractive chromophores Design approach Since the carbazole molecule has a structure isoelectronic with diphenylamine, the introduction of electron-withdrawing groups in the 3- and/or 6-position induces intracharge-transfer and a mesomeric dipole moment.Depending on the electrona Ynity of acceptor groups, polarizabilities of carbazole derivatives can be tuned by appropriate molecular design of the substituent groups.142 Acceptor-introduced carbazoles have been shown to be very promising as second-order NLO chromophores.143,144 Besides the 3- and 6-substitution positions, N-substitution (9-position) allows various chemical modifications: solubilization and amorphism of substituted carbazole by diVerent length of alkyl chain on the 9-pos- Second-order nonlinear optical moiety Charge transporting moiety Photosensitizer moiety Plastic chain ition,145,146 and control of noncentrosymmetric packing in the Fig. 1 Molecular design approach to amorphous non-polymer photorefractive materials crystalline state through hydrogen bonding.147 J.Mater. Chem., 1998, 8(4), 809–828 821It is well known that photorefractive materials are multifunctional. Until now, this multifunctionality has been limited in molecular design approaches to multicomponent systems for photorefraction. One of our design targets is to develop multifunctional chromophores based on carbazole building blocks by chemical modification.We have tried to find one way of designing photorefractive materials with second-order NLO chromophores.150,151 Some successful examples follow. Amorphous carbazole trimers as multifunctonal chromophores In carbazole conjugated trimers (X) with acceptor groups for photorefractive materials, three carbazole rings are linked by ethynyl groups, with the peripheral carbazoles substituted with electron-withdrawing groups which make the materials EO active.The central carbazole can act as a charge transporting Fig. 2 Absorption spectra of trimer doped with diVerent TNF concenfunction. In order to obtain amorphous carbazole trimers with trations: (a) 0, (b) 0.055, (c) 0.06, (d) 0.0625 and (e) 0.125 wt% a low T g, the three long aliphatic groups are introduced to the trimers at the 9-position of each of the carbazole moieties. Fig. 3 Photoconductive sensitivity as a function of the applied electric field for dicyanovinyl-substituted carbazole trimer doped with diVerent TNF concentrations: ($) 0.05, (#) 0.06, (%) 0.125 and (&) 0.5 wt% trimer, the chromophores could be eVectively aligned at room temperature by applying a dc electric field across the sample.The EO coeYcient was measured on the same sample (134 mm) for the photorefractive measurements using a transmission technique.28 In order to avoid the birefringence contributed from orientation of second-order NLO chromophores, more Multifunctional carbazole conjugated trimers accurate measurements of EO coeYcients have been performed at high modulation ac frequencies at which the birefringence A dicyanovinyl substituted trimer with a T g of 29 °C has been eVect is insignificant and only the EO eVect of chromophores used as a model example for studies of photorefractive eVects.21 can respond to the applied ac frequencies.31 Fig. 4 shows the A single carbazole trimer with a dicyanovinyl group does not electric field dependence of the EO coeYcient. These values exhibit any absorption coeYcient at a wavelength of 633 nm.are comparable with those obtained from other photorefractive However, the trimer could be sensitized by addition of a small materials systems.16,99 The photorefractive nature of the trimer amount of TNF. The charge transfer complex formed by the central cabazole moiety of the trimer and TNF shows a controllable absorption coeYcient at 633 nm.Fig. 2 shows the absorption spectra of the trimer and the trimer doped with diVerent concentrations of TNF. The photoconductive sensitivities were measured on a 10 mm thick sample sandwiched between Al and indium-tin oxide (ITO) electrodes.25 It was found that the TNF concentration has a significant influence on photoconductivity, and the maximized photoconductivity sensitivity value was observed in a 0.06 wt% TNF doped trimer (Fig. 3:). The density of charge transfer complex units which act as the charge generation functions increased with the doping concentration of TNF. Uncomplexed TNF can also act as the hole trap. The density of free carbazole units, which are responsible for hole transport decreased simultaneously, and sensitively influenced the magnitude of charge drift mobility.66 Therefore, there is a balance between the charge generation and charge transporting components, which Fig. 4 The electro-optic coeYcient of dicyanovinyl-substituted carbaresults in the observation of an optimized TNF concentration zole trimer as a function of the applied electric field at an ac frequency of 6 KHz of 0.06 wt% in photoconductivity. Due to the low T g of the 822 J.Mater. Chem., 1998, 8(4), 809–828was confirmed by a two-beam coupling experiment. Two important photorefractive eVects were observed. According to the standard model of photorefractivity,89 the phase shift between the illumination pattern and the refractive index grating increased with the applied electric field (Fig. 5:). It was found that the phase shift at low electric fields is non-zero between 0 and 90°. At relatively high electric field, the phase shift approaches 90°. Asymmetric energy transfer between the two beams caused by the non-zero phase shift was also observed from a trimer sample when an electric field was applied. The nonlocal nature of the index grating arises from an applied electric field which makes the sample photoconductive and EO active.14 Fig. 6 shows the two-beam coupling gain as a function of the applied electric field for the trimer doped with diVerent concentrations of TNF. The diVraction eYciency was measured with a four-wave mixing experiment. It was Fig. 7 The diVraction eYciency as a function of the applied electric found that the diVraction eYciency depended strongly on the field for dicyanovinyl-substituted carbazole trimer doped with diVerent TNF concentrations: (&) 0.05, (%) 0.055, (#) 0.06 and ($) 0.0625 wt% applied electric field (Fig. 7) and reaches 18.3% at a field of 30.6 V mm-1 in a 0.06 wt% TNF doping sample. The eVect of optimized TNF concentration was also found in the photorefcoupling gain, independently.The modulation of the refractive ractive two-beam coupling and four-wave mixing experiments. index from both measurements is in good agreement within Unlike the two-beam coupling experiment, the four-wave experimental error (as shown in Fig. 8) This result indicates mixing diVraction eYciency is not dependent only on the phase that the nonlocal index grating is the dominant mechanism of shift of the grating.In principle, local gratings induced by the holographic gratings. many other mechanisms such as photochromism, thermochro- The amorphous trimer as the first multifunctional chromo- mism and photochemical eVects, can also contribute to the phore has been demonstrated to exhibit good photorefractive diVraction eYciency. In order to confirm the diVraction eVects.This trimer approach has several advantages: (1) high eYciency of the trimer arising from pure space charge field concentration of carrier transporting agent and second-order induced eVects, the modulation of the refractive index can be NLO moieties, (2) large carrier mobility arising from the determined from the diVraction eYciency and the two-beam conjugated structure, (3) good film-forming properties without any amorphous supporting matrix, (4) flexibility in optimizing the photorefractivity by adjusting the concentration of TNF, and (5) phase-separation-free due to very low concentration of the dopant.The amorphous carbazole trimer as a monolithic photorefractive chromophore. A novel multifunctional conjugated carbazole trimer X with nitro acceptor groups has been found to be the first monolithic photorefractive material.It was found that this trimer doped with no other functional components showed an eYcient photorefractive eVect.152 In this trimer, carbazole rings were also linked by the ethynyl group and peripheral carbazoles were substituted with nitro groups. The nitrosubstituted carbazole trimer displays a suitable absorption coeYcient of 8.2 cm-1 at a wavelength of 532 nm.This absorption coeYcient can allow observation of the photorefractive properties of this trimer at an operating wavelength of 532 nm. Fig. 5 The phase shift of index grating of dicyanovinyl-substituted This trimer was demonstrated to be both photoconductive carbazole trimer doped with 0.06 wt% TNF as a function of the applied electric field Fig. 8 Magnitude of index modulation of dicyanovinyl-substituted Fig. 6 The photorefractive gain coeYcient as a function of the applied carbazole trimer doped with 0.06 wt% of TNF versus applied electric field: calculated from (#) p-polarized two-beam coupling gain, ($) s- electric field for dicyanovinyl-substituted carbazole trimer doped with diVerent TNF concentrations: (%) 0.055, (#) 0.06, ($) 0.0625 and (&) polarized two-beam coupling gain and (%) p-polarized diVraction eYciency 0.125 wt% J.Mater. Chem., 1998, 8(4), 809–828 823Fig. 9 The SH intensity of carbazole trimer with a nitro group as a Fig. 11 The two-beam coupling gain of a nitro-substituted carbazole function of incidence angle at an applied electric field of 23 V mm-1 conjugated trimer as a function of the applied electric field.The dashed line is the absorption coeYcient at a wavelength of 532 nm. Fig. 10 The intensity of beam 1 (upper trace) monitored as beam 2 (lower trace) is switched on at time t=0 s and oV at t=255 s, and the Fig. 12 The diVraction eYciency of a nitro-substituted carbazole conintensity of beam 2 monitored as beam 1 is turned on and oV.The jugated trimer as a function of the applied electric field applied electric field was 33 V mm-1 and to have second-order NLO activity. The noncentrosymmetric alignment of the chromophores can be achieved by an electric poling field at room temperature due to its low T g of 20 °C and this can be confirmed by a second harmonic generation (SHG) measurement.The SHG experiment was carried out on the same sample for the photorefractive measurements at a fundamental wavelength of 1064 nm in transmission mode. With no electric field applied, the SH intensity could not be observed, as a result of the centrosymmetric random arrangement of the chromophores. After switching on the electric field, repeatable orientation of the chromophores was realized, reaching a stable plateau value within a few seconds.This partial orientation of the chromophores at room temperature came as a result of the low T g of the carbazole trimer. The SH intensity is strongly dependent on the applied electric field. Fig. 9 shows the angular dependence of the SH intensity at a poling electric field of 23 V mm-1.The photoconductive properties were studied on a sample sandwiched between an ITO and a gold coated glass substrate at a wavelength of 532 nm.25 The photocurrent was almost independent of the laser intensity, but strongly depen- Fig. 13 Experimental geometry for optical image reconstruction and dent on the applied electric field. The photoconductive sensi- phase conjugation via four-wave mixing for the correction of distorted images: (a) input image, (b) distorted image and (c) conjugated image.tivity of the carbazole trimer with nitro groups was measured Reproduced by permission from ref. 153. to be 1.2×10-11 cm S W-1 at an external field of 39 V mm-1. No detectable dark conductivity was observed. It was found that the photocurrent increased rapidly with a time constant transfer between the two beams was observed when an electric field was applied.This provided proof that an electronic <0.1 s upon exposure to light as a result of the large carrier mobility of the carbazole trimer. photorefractive eVect is present.14 Fig. 10 shows typical asymmetric behavior for the monolithic carbazole trimer at an The photorefractive properties of the trimer were characterized by a two-beam coupling and a four-wave mixing.In applied electric field of 33 V mm-1. The two-beam coupling gain coeYcient could be estimated from the asymmetric energy the two-beam coupling experiment, an asymmetric energy 824 J. Mater. Chem., 1998, 8(4), 809–828Fig. 14 Chemical structures and dij values of carbazole dendrimers transfer. The two-beam coupling gain increases monotonically sized in our laboratory.153 These carbazole dendrimeric oligomers have film-forming properties and show glass trans- with the applied electric field as shown in Fig. 11. At an applied electric field of 33 V mm-1, a photorefractive gain of 35.0 cm-1 ition behavior. Values of T g could be controlled by the length of spacer, and the number of carbazole rings or acceptor was obtained.The absorption coeYcient for this trimer was 8.2 cm-1, leading to a net two-beam coupling gain coeYcient groups. Amorphous molecular solid films could be prepared without a supporting matrix by spin-coating. These thin films of 26.8 cm-1. The applied electric field plays an important role in enhancing the photorefractive eVects because of improved could be poled above T g to achieve noncentrosymmetric alignment of the molecular dipoles required for an EO alignment of the second-order NLO chromophores and higher photoconductivity at a higher electric field.If an external response. Second-order NLO responses were examined on thin films by SHG. The values of the second-order NLO coeYcients electric field is not applied during writing, no detectable gratings are observed due to the centrosymmetric random (dij) were strongly dependent on the acceptor groups.The chemical structures and dij values are summarized in Fig. 14. alignment of second order NLO chromophores. Four-wave mixing was used to determine the steady-state diVraction Photoconductive properties of this dendrimer system have been examined by means of a xerographic discharge technique.eYciency of the carbazole trimer. Fig. 12 shows the dependence of the diVraction eYciency on the applied electric fields. At an It is clear that these molecular systems have multifunctional properties, i.e. both photoconductiveity and second-order NLO electric field of 33.3 V mm-1, a diVraction eYciency of 13.2% was reached.Optical image reconstruction of distorted images responses. Two-beam coupling experiments on dendrimers with carbazole substituted with dicyanovinyl groups indicated (as shown in Fig. 13) using phase conjugation was demonstrated in this monolithic photorefractive material.153 that the induced index grating is shifted by 90° with respect to the light intensity grating.This phase shift, or nonlocal nature of the photorefractive eVect, gives rise to an asymmetric Carbazole dendrimers. The development of materials with new chemical structures for photorefractivity is an extremely energy transfer between the two writing beams, which does not occur in any of other processes. The two-beam coupling active field.14,17 In order to develop new amorphous molecules with good modification flexibility, dendrimers have been selec- gain of 11.8 cm-1 was obtained for this dendrimer with a zero applied electric field.ted as a molecular design approach to photorefractive materials. Dendrimeric structures have several advantage for design of photorefractive materials: (1) they are amorphous, (2) the Summary and outlook core and diVerent generation can be modified with diVerent functions for meeting the multifunctional requirements of Considerable progress has been made in understanding both the photorefractive origins as well as the molecular design of photorefractive materials, (3) diVerent dendrons with diVerent functions can also meet the multifunctional requirements.Their amorphous organic photorefractive materials.For example, many interesting new phenomena which do not occur in spherical structure is expected to impart unusual properties. Several dendrimers with mono-acceptor substituted carba- crystalline materials have been observed. The material design approach based on a single chromophore has been developed. zoles as the multifunctional chromophores have been synthe- J. Mater.Chem., 1998, 8(4), 809–828 82525 J. S. Schildkraut, Appl. Phys. L ett., 1991, 58, 340. Many organic amorphous materials with low cost and ease of 26 E. Mu� ller-Horsche, D. Haarer and H. Scher, Phys. Rev. B, 1987, fabrication exhibiting better photorefractive performance than 35, 1273. inorganic crystals have been developed. However, before 27 J. C. Scott, L. Th. Pautmeier and W.E. Moerner, J. Opt. Soc. Am. amorphous organic photorefractive materials can be con- B, 1992, 9, 2059. sidered for practical applications, many important issues have 28 C. C. Teng and H. T. Man, Appl. Phys. L ett., 1990, 56, 1734. to be addressed. Development of new material systems with 29 M. Siegelle and R. Hierle, J. Appl. Phys., 1981, 52, 4199. 30 K. D. Singer, M.G. Kuzyk, W. R. Holland, J. E. Sohn, optimized photorefractive properties and fabrication abilities S. J. Lalama, R. B. Comizzoli, H. E. Katz and M. L. Schilling, remains a major challenge for chemical research work. Appl. Phys. L ett., 1988, 53, 1800. Almost all amorphous organic photorefractive materials 31 W. E. Moerner, S. M. Silence, F. Hache and G. C. Bjorklund, reported must be induced to exhibit photorefractive eVects by J.Opt. Soc. Am. B, 1994, 11, 320. a high applied electric field. Therefore, this high electric field 32 C. A.Walsh and W. E. Moerner, J. Opt. Soc. Am. B, 1992, 9, 1642. must limit practical device applications of these materials. New 33 J. P. Huignard and A. Marrakchi, Opt. Commun., 1981, 38, 249. 34 J. Feinberg, D. Heiman, A. R.Tanguay, Jr. and R. W. Hellwarth, high performance materials that do not require an applied J. Appl. Phys., 1980, 51, 1297. electric field or with low driven electric field should be devel- 35 D. H. Dunlap, E. P. Parris and V. M. Kenkre, Phys. Rev. L ett., oped for practical applications. Several successful examples of 1996, 77, 542. photorefractive materials, such as high T g polymers,69,141 36 W.E. Moerner, C. Walsh, J. C. Scott, S. Ducharme, D. M. sol–gel composite materials,134 and amorphous dendrimers,153 Burland, G. C. Bjorklund and R. J. Twieg, Proc. SPIE, 1991, have been reported and demonstrated to display photorefrac- 1560, 278. 37 S. Ducharme, B. Jones, J. M. Takacs and L. Zhang, Opt. L ett., tive eVects with a driven applied electric field. 1993, 18, 152.For low T g photorefractive materials, the high performance 38 S. M. Silence, C. A. Walsh, J. C. Scott, T. J. Matray, of photorefractive eVects reported mainly comes from the R. J. Twieg, F. Hache, G. C. Bjorklund and W. E. Moerner, Opt. contribution of orientation induced birefringence. According L ett., 1992, 17, 1107. to the application requirements, materials with photorefractive 39 S.M. Silence, C. A. Walsh, J. C. Scott and W. E. Moerner, Appl. eVects contributed by EO eVects should be promising candi- Phys. L ett, 1992, 61, 2967. 40 Y. Cui, Y. Zhang, P. N. Prasad, J. S. Schildkraut and dates for practical applications. D. J. Williams, Appl. Phys. L ett., 1992, 61, 2132. Due to the limitation of space, this paper did not include 41 S. M.Silence, F. Hache, M. Donckers, C. A. Walsh, D. M. photorefractive liquid crystals154–156 and only summarizes Burland, G. C. Bjorklund, R. J. Twieg and W. E. Moerner, Proc. amorphous organic photorefractive materials based on mate- SPIE, 1993, 1852, 253. rials chemistry. 42 M. Liphardt, A. Goonesekera, B. E. Jones, S. Ducharme, J. M. Takacs and L. Zhang, Science, 1994, 263, 367. 43 A. Darwish, N. Kukhtarev, R. Copland, R. Sliz, References P. Venkatesewarlu, A. Williams, J. Caulfield, S. Ducharme, J. M. Takacs and L. Zhang, Proc. SPIE, 1996, 2850, 33. 1 Photorefractive Materials and T heir Applications I and II, ed. 44 A. Goonesekera, S. Ducharme, J. M. Takacs and L. Zhang, Proc. P. Gu� lnter and J.-P. Huignard, Springer Verlag, Berlin, 1988. SPIE, 1996, 2850, 41. 2 Introduction to Photorefractive Nonlinear Optics, ed. P. Yeh, 45 S. M. Silence, J. C. Scott, F. Hache, E. J. Ginsburg, P. K. Jenkner, Wiley, New York, 1993. R. D. Miller, R. J. Twieg and W. E. Moerner, J. Opt. Soc. Am. B, 3 L aser-Induced Dynamic Grating, Springer Series in Optical 1993, 10, 2306. Sciences, ed. H. J. Eichler, P. Gu� lnter and D. W. Pohl, Springer, 46 K. Yokoyama, K.Arishima, T. Shimada and K. Sukegawa, Jpn. Berlin, 1986, vol. 50. J. Appl. Phys. 1994, 33, 1029. 4 Photochromism: T echniques of Chemistry, ed. G. H. Brown, Wiley- 47 K. Yokoyama, K. Arishima and K. Sukegawa, Appl. Phys. L ett., Interscience, New York, 1971, vol. I–III. 1994, 65, 132. 5 A. Ashkin, G. D. Boyd, J. M. Dziedzic, R. G. Smith, A. A. 48 R. Burzynski, Y. Zhang, S. Ghosal and M.K. Casstevens, Ballmann and K. Nassau, Appl. Phys. L ett., 1966, 9, 72. Polymeric Materials Science and Engineering, ACS Spring 6 A. M. Glass, A. M. Johnson, D. H. Olson, W. Simpson and Meeting, Anaheim, California, 1995, vol. 72, p. 292 ACS, A. A. Ballman, Appl. Phys. L ett., 1984, 44, 948. Washington DC. 7 J. Feinberg, Physics T oday, 1988, 41, 46. 49 R. Wortmann, C. Poga, R.J. Twieg, C. Geletneky, C. R. Moylan, 8 G. Roosen, J.-P. Huignard and M. Cronin-Golomb, J. Opt. Soc. P. M. Lundquist, R. G. DeVoe, P. M. Cotts, H. Horn, J. E. Rice Am. B, 1990, 7, 2242. and D. M. Burland, J. Chem. Phys., 1996, 105, 10 637. 9 J. Feinberg and B. Fischer, J. Opt. Soc. Am. B, 1992, 9, 1404. 50 S. Heun and P. M. Borsenberger, Chem. Phys., 1995, 200, 245. 10 L. Hesselink, E.Kratzig and K. H. Ringhofer, J. Opt. Soc. Am. B, 51 C. R. Moylan, R. J. Twieg, V. Y. Lee, S. A. Swanson, 1994, 11, 1648. K. M. Betterton and R. D. Miller, J. Am. Chem. Soc., 1993, 115, 11 K. Sutter, J. Hullinger and P. Cu� lnter, Solid State Commun. 1990, 12 599. 74, 867. 52 J. J. Stankus, S. M. Silence, R. J. Twieg, D. M. Burland, 12 K. Sutter and P. Gu� lnter, J. Opt.Soc. Am. B, 1990, 7, 2274. R. D. Miller, J. C. Scott, W. E. Moerner and G. C. Bjorklund, 13 S. Ducharme, J. C. Scott, R. J. Twieg and W. E. Moerner, Phys. Proc. SPIE, 1994, 2285, 204. Rev. L ett., 1991, 66, 1846. 53 S. M. Silence, R. J. Twieg, G. C. Bjorklund and W. E. Moerner, 14 W. E. Moerner and S. M. Silence, Chem. Rev., 1994, 94, 127. Phys. Rev. L ett., 1994, 73, 2047. 15 Nonlinear Optical Properties of Organic Molecules and Crystals, 54 S.M. Silence, J. C. Scott, J. J. Stankus, W. E. Moerner, ed. D. S. Chemla and J. Zyss, Academic Press, Orlando, 1987, C. R. Moylan, G. C. Bjorklund and R. J. Twieg, J. Phys. Chem., vol. 1 and 2. 1995, 99, 4096. 16 G. C. Valley and N. B. Klein, Opt. Eng., 1983, 22, 704. 55 Y. Zhang, S. Ghosal, M. Casstevens and R. Burzynski, Appl.Phys. 17 Y. Zhang, R. Burzynski, S. Ghosal and M. K. Casstevens, Adv. L ett., 1995, 66, 256. Mater., 1996, 8, 111. 56 Y. Zhang, S. Ghosal, M. Casstevens and R. Burzynski, J. Appl. 18 L. Yu, W. K. Chan, Z. Peng and A. Gharavi, Acc. Chem. Res., Phys., 1996, 79, 8920. 1996, 29, 13. 57 H. J. Bolink, C. Arts, V. V. Krasnikov, G. G. Malliaras and 19 P. M. Lundquist, R. Wortmann, C.Geletneky, R. J. Twieg, G. Hadziioannou, Proc. SPIE, 1996, 2850, 69. M. Jurich, V. Y. Lee, C. R. Moylan and D. M. Burland, Science, 58 M. Stolka, J. F. Yanus and D. M. Pai, J. Phys. Chem., 1984, 1996, 274, 1182. 88, 4707. 20 Y. D. Zhang, T.Wada, L.Wang and H. Sasabe, T etrahedron L ett., 59 P. M. Lundquist, R. J. Twieg, R. Wortmann, R. M. Shelby, 1997, 38, 1758. C. Geletneky, M.Jurich and D. M. Burland, Proc. SPIE, 1996, 21 L. Wang, Y. D. Zhang, T. Wada and H. Sasabe, Appl. Phys. L ett., 2850, 78. 1996, 69, 728. 60 L. Yu, W. K. Chan, Z. N. Bao and S. Cao, Macromolecules, 1993, 22 H. J. Bolink, V. V. Krasnikov, G. G. Malliaras and 26, 2216. G. Hadziioannou, Adv.Mater, 1994, 6, 574. 61 L. Yu, Y.M. Chen, W. K. Chan and Z. H. Peng, Appl. Phys. L ett., 23 L.Onsager, Phys. Rev., 1938, 54, 554. 24 G. Pfister, Phys. Rev. B, 1977, 16, 3676. 1994, 64, 2489. 826 J. Mater. Chem., 1998, 8(4), 809–82862 Z. H. Peng, Z. N, Bao, Y. M. Chen and L. Yu, J. Am. Chem. Soc., Polymers, ed. P. N. Prasad and D. J. Williams, Wiley-Interscience, New York, 1991. 1994, 116, 6003. 63 Z. H. Peng, A. R. Gharavi and L. Yu, Appl. Phys. L ett., 1996, 106 A.Twarowski,Phys., 1989, 65, 2833. 107 L. Pautmeier, R. Richert and H. Bassler, Synth. Met., 1990, 37, 69, 4002. 64 M. Gailberger and H. Bassler, Phys. Rev. B, 1991, 44, 8643. 271. 108 B. Swedek, P. N. Prasad, Y. Cui, N. Cheng, J. Zieba, J. Winiarz 65 L. Li, K. G. Chittibabu, Z. Chen, J. I. Chen, S. Marturunkakul, J. Kumar and S. K. Tripathy, Opt. Commun., 1996, 125, 257.and K. S. Kim, Proc. SPIE, 1996, 2850, 89. 109 S. F. Wang, Z. W. Huang, Z. J. Chen, Q. H. Gong, Z. J. Zhang 66 W. D. Gill, J. Appl. Phys., 1972, 42, 5033. 67 H. Hoegl, J. Phys. Chem., 1965, 69, 755. and H. Y. Chen, Chin. Phys. L ett., 1997, 14, 474. 110 O. Zobel, M. Eckl, P. Strohriegl and D. Haarer, Adv. Mater., 68 Poly(N-vinylcarbazole), ed. J. M. Pearson and M. Stolka, Gordon and Breach, New York, 1981. 1995, 7, 911. 111 S. X. Dou, J. S. Zhang, P. X. Ye, H. P. Hong and C. Ye, Acta Phys. 69 Y. Zhang, Y. Cui and P. N. Prasad, Phys. Rev. B, 1992, 46, 9900. 70 K. Meerholz, B. L. Volodin, Sandalphon, K. Kippelen and Sinica, 1995, 4, 663. 112 B. Volodin, K. Meerholz, Sandalphon, B. Kippelen and N. Peyghambarian, Nature, 1994, 371, 497. 71 A. Gunnet-Jepsen, C.L. Thompson, R. J. Twieg and W. E. N. Peyghambarian, Proc. SPIE, 1994, 2144, 72. 113 Z. J. Chen, L. F. Xu, Z. W. Huang, Q. H. Gong, Z. J. Zhang, Moerner, Appl. Phys. L ett., 1997, 70, 1515. 72 P.-M. Allemand, A. Koch, F. Wudl, Y. Rubin, F. Diederich, Y. K. He and H. Y. Chen, Chin. Phys. L ett., in press. 114 Sandalphon, B. Kippelen, K. Meerholz and N. Peyghambarian, M. M. Alvarez, S.J. Anz and R. L. Whetton, J. Am. Chem. Soc., 1991, 113, 1050. Appl. Opt., 1996, 35, 2346. 115 B. Kippelen, Sandalphon, K. Meerholz and N. Peyghambarian, 73 J. W. Arbogast, C. S. Foote and M. Kao, J. Am. Chem. Soc., 1992, 114, 2277. Appl. Phys. L ett., 1996, 68, 1748. 116 B. Kippelen, F. Meyer, N. Peyghambarian and S. R. Marder, 74 Y. Wang, Nature, 1992, 356, 585. 75 Y. Zhang, C.A. Spencer, S. Ghosal, M. Casstevens and J. Am. Chem. Soc., 1997, 119, 4559. 117 K. Meerholz, R. Bittner, C. Brauchle, B. L. Volodin, Sandalphon, R. Burzynski, Appl. Phys. L ett., 1994, 64, 1908. 76 Y. Zhang, C. A. Spencer, S. Ghosal, M. K. Casstevens and B. Kippelen and N. Peyghambarian, Proc. SPIE, 1996, 2850, 100. 118 B. Volodin, K. Meerholz, Sandalphon, B. Kippelen and R.Burzynski, J. Appl. Phys., 1994, 76, 671. 77 M. C. J. M. Donckers, S. M. Silence, C. A. Walsh, F. Hache, N. Peyghambarian, Proc. SPIE, 1995, 2405, 143. 119 B. Kippelen, K. Meerholz, Sandalphon, B. Volodin and D. M. Burland, W. E. Moerner and R. J. Twieg, Opt. L ett., 1993, 18, 1044. N. Peyghambarian, Nonlinear Opt., 1995, 11, 263. 120 B. L. Volodin, Sandalphon, K. Meerholz, B. Kippelen, 78 G. Weiser, J.Appl. Phys., 1972, 43, 5028. 79 P. Yeh, Appl. Opt., 1987, 26, 602. N. V. Kukhtarev and N. Peyghambarian, Opt. Eng., 1995, 34, 2213. 80 S. M. Silence, M. C. J. M. Donckers, C. A. Walsh, D. M. Burland, R. J. Twieg and W. E. Moerner, Appl. Opt., 1994, 33, 2218. 121 K. Meerholz, Angew. Chem., Int. Ed. Engl., 1997, 36, 945. 122 C. Halvorson, B. Kraabel, A. J. Heeger, B.L. Volodin, 81 W. E. Moerner, S. M. Silence, F. Hache and G. C. Bjorklund, J. Opt. Soc. Am. B, 1994, 11, 320. K. Meerholz, Sandalphon and N. Peyghambarian, Opt. L ett., 1995, 20, 76. 82 D. M. Burland, G. C. Bjorklund, W. E. Moerner, S. M. Silence and J. J. Stankus, Pure Appl. Chem., 1995, 67, 33. 123 B. L. Volodin, B. Kippelen, K. Meerholz, B. Javidi and N. Peyghambarian, Nature, 1996, 383, 58. 83 D. M. Burland, R. G. Devoe, C. Geletneky, Y. Jia, V. Y. Lee, P. M. Lundquist, C. R. Moylan, C. Poga, R. J. Twieg and 124 H. J. Bolink, V. V. Krasnikov, G. G. Malliaras and G. Hadziioannou, J. Phys. Chem., 1996, 100, 16 356. R. Wortmann, Pure Appl. Opt., 1996, 5, 513. 84 M. E. Orczyk, J. Zieba and P. N. Prasad, Appl. Phys. L ett., 1995, 125 P. Strohriegl,Makromol.Chem. Rapid Commun., 1986, 7, 771. 126 M. Lux, P. Strohriegl and H. Hocker, Makromol. Chem., 1987, 63, 311. 85 G. G. Malliaras, V. V. Krasnikov, H. J. Bolink and 188, 811. 127 P. Strohriegl,Makromol. Chem., 1993, 194, 363. G. Hadziioannou, Appl. Phys. L ett., 1994, 65, 262. 86 G. G. Malliaras, V. V. Krasnikov, H. J. Bolink and 128 C. Poga, P. M. Lundquist, V. Lee, R. M. Shelby, R. J. Twieg and D. M. Burland, Appl. Phys. L ett., 1996, 69, 1047. G. Hadziioannou, Proc. SPIE, 1995, 2526, 94. 87 G. G. Malliaras, V. V. Krasnikov, H. J. Bolink and 129 Sol–Gel T echnology for T hin Films, Fibers, Preforms, Electronics and Specialty Shapes, ed. L. C. Klein, Noyes, Park Ridge, 1988. G. Hadziioannou, Appl. Phys. L ett., 1995, 66, 1038. 88 D. M. Pai, J. F. Yanus and M. Stolka, J. Phys. Chem., 1994, 130 Sol–Gel Science: T he Physics and Chemistry of Sol–Gel Processing, ed. C. J. Brinker and G. Scherer, Academic Press, 88, 4714. 89 N. V. Kukhtarev, V. B. Markov, M. Soskin and V. L. Vinetskii, Boston, 1990. 131 S. Kalluri, Y. Shi, W. Steier, Z. Yang, C. Xu, B. Wu and Ferroelectrics, 1979, 22, 949. 90 G. G. Malliaras, V. V. Krasnikov, H. J. Bolink and L. R. Dalton, Appl. Phys. L ett., 1994, 65, 2651. 132 J. Chen, S. Marturunkakul, L. Li, R. J. Jeng, J. Kumar and G. Hadziioannou, Proc. SPIE, 1995, 2527, 250. 91 G. G. Malliaras, V. V. Krasnikov, H. J. Bolink and S. K. Tripathy,Macromolecules, 1993, 26, 7379. 133 R. Burzynski, M. K. Casstevens, Y. Zhang and S. Ghosal, Opt. G. Hadziioannou, Appl. Phys. L ett., 1995, 67, 455. 92 H. Scher and E. W. Montroll, Phys. Rev. B, 1975, 12, 2455. Eng., 1996, 35, 443. 134 F. Chaput, D. Riehl, J. P. Boilot, K. Cargnelli, M. Canva, Y. Levy 93 G. G. Malliaras, V. V. Krasnikov, H. J. Bolink and G. Hadziioannou, Phys. Rev. B, 1995, 52, 14 324. and A. Brun, Chem. Mater., 1990, 8, 312. 135 K. Tamura, A. B. Padias, H. K. Hall Jr. and N. Peyghambarian, 94 G. G. Malliaras, H. Angerman, V. V. Krasnikov, G. ten Brinke and G. Hadziioannou, J. Phys. D: Appl. Phys., 1996, 29, 2045. Appl. Phys. L ett., 1992, 60, 1803. 136 B. Kippelen, K. Tamura, N. Peyghambarian, A. B. Padias and 95 A. R. Tahmasbi and J. Hirsch, Solid State Commun., 1980, 34, 75. 96 G. G. Malliaras, V. V. Krasnikov, H. J. Bolink and H. K. Hall, Jr., Phys. Rev., 1993, 48, 10 710. 137 T. Kawakami and N. Sonoda, Appl. Phys. L ett., 1993, 62, 2167. G. Hadziioannou, Proc. SPIE, 1996, 2850, 24. 97 D. Morichere, G. G. Malliaras, V. V. Krasnikov, H. J. Bolink and 138 C. Zhao, C. Park, P. N. Prasad, Y. Zhang, S. Ghosal and R. Burzynski, Chem. Mater., 1995, 7, 1237. G. Hadziioannou, J. Chim. Phys., 1995, 92, 927. 98 G. G. Malliaras, V. V. Krasnikov, H. J. Bolink and 139 Y. D. Zhang, T. Wada, L. Wang, T. Aoyama and H. Sasabe, Chem. Commun., 1996, 2325. G. Hadziioannou, Pure Appl. Opt., 1996, 5, 631. 99 A. M. Cox, R. D. Blackburn, D. P. West, T. A. King, F. A. Wade 140 T. Aoyama, T. Wada, Y. D. Zhang, H. Sasabe and K. Sasaki, Proc. SPIE, 1997, 3147, 103. and D. A. Leigh, Appl. Phys. L ett., 1996, 68, 2801. 100 B. Kippelen, Sandalphon, N. Peyghambarian, S. R. Lyon, 141 Y. D. Zhang, H. Kimura-Suda, T. Wada, L. Wang, T. Aoyama and H. Sasabe, to be published. A. B. Padias and H. K. Hall Jr., Electron. L ett., 1993, 29, 1873. 101 Sandalphon, B. Kippelen, N. Peyghambarian, S. R. Lyon, 142 H. Sasabe, Supramol. Sci., 1996, 3, 91. 143 T. Wada, Y. D. Zhang, M. Yamakado and H. Sasabe, Mol. Cryst. A. B. Padias and H. K. Hall Jr. Opt. L ett., 1994, 19, 68. 102 M. A. Smith, N. R. King, G. R. Mitchell and S. V. O’Leary, Proc. L iq. Cryst., 1993, 227, 85. 144 Y. D. Zhang, L. Wang, T. Wada and H. Sasabe, Macromolecules, SPIE, 1996, 2850, 14. 103 M. E. Orczyk, J. Zieba and P. N. Prasad, J. Phys. Chem., 1994, 1996, 29, 1569. 145 Y. D. Zhang, T. Wada and H. Sasabe, J. Polym. Sci. Part A: 98, 8699. 104 M. E. Orczyk, B. Swedek, J. Zieba and P. N. Prasad, J. Appl. Polym. Chem., 1996, 34, 2289. 146 T. Wada, Y. D. Zhang, L. Wang and H. Sasabe, Nonlinear, Opt., Phys., 1994, 76, 4995. 105 Introduction to Nonlinear Optical EVects in Molecules and 1995, 9, 276. J. Mater. Chem., 1998, 8(4), 809–828 827147 T.Wada, Y. D. Zhang, Y. S. Choi and H. Sasabe, J. Phys. D: Appl. 152 Y. D. Zhang, L. Wang, T.Wada and H. Sasabe, Appl. Phys. L ett., 1997, 70, 2949. Phys., 1993, 26, B221. 148 Y. D. Zhang, T. Wada, L. Wang and H. Sasabe, Chem. Mater., 153 T. Wada, L. Wang, Y. D. Zhang, M. Tian and H. Sasabe, Nonlinear Opt., 1996, 15, 103. 1997, 9, 2798. 149 C. Beginn, J. V. Grazulevicius and P. Strohriegl,Macromol. Chem. 154 I. C. Khoo, H. Li and Y. Liang, Opt. L ett., 1994, 19, 1723. 155 G. P. Wiederrecht, B. A. Yoon and M. R. Wasielewski, Science, Phys., 1994, 195, 2353. 150 D. Williams, Angew. Chem., Int. Ed. Engl., 1984, 23, 690. 1995, 270, 1794. 156 G. P. Wiederrecht, B. A. Yoon, W. A. Svec and M. R.Wasielewski, 151 S. R. Marder, D. N. Beratan and L.-T. Cheng, Science, 1991, 252, 103. J. Am. Chem. Soc., 1997, 119, 3358. Paper 7/05129H; Received 17th July, 1997 828 J. Mater. Chem., 1998, 8(4), 809–828

ISSN:0959-9428    

DOI:10.1039/a705129h

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, structure and properties of a novel trisulfide double-bridged TTF dimer Hideki Fujiwara,* Emiko Arai and Hayao Kobayashi* Institute forMolecular Science,Myodaiji, Okazaki 444, Japan We have synthesized novel trisulfide-bridged tetrathiafulvalene (TTF) dimer molecule in which two TTF moieties are connected to each other by two trisulfide chains.X-Ray crystal structure analysis reveals that the molecule has a unique cyclophane-like U-shape structure. The structure of the ClO4 salt is also described. Research towards new donors for organic conductors based on the tetrathiafulvalene (TTF) framework has been actively pursued since the discovery of the metallic TTF–tetracyanoquinodimethane complex.1 Among them, the synthesis of tetrathiafulvalenophane was first reported in 1980 and extensively developed in the 1990s because of its unusual molecular structure and interesting intramolecular interaction.2,3 In particular several of these compounds have molecular structures like the dimeric structure of the coventional conducting salts such as the k-type bis(ethylenedithio)–TTF salts,4 which is related to the strong electron correlation system and Mott S S S S SCH2CH2CN SCH2CH2CN MeS MeS S S S S MeS MeS S Zn S S S S S S S SMe SMe i, ii iii 1 (TBA)2 3 4 insulating state.However they often have a largely dislocated Scheme 1 Reagents and conditions: i, 28 wt% sodium methoxide in structure in the transverse direction because of the flexibility methanol (2 equiv.), dry CH3CN–MeOH (551, v/v), room temperature, of the polymethylene chains which connect the two TTF 10 min; ii, zinc(II) chloride (0.5 equiv.), tetra-n-butylammonium bromoieties and this situation may weaken the interplanar inter- mide (1 equiv.), iii, SCl2 (1 equiv.), dry CH3CN, 55 °C, 2 h action.With this in mind we focused on a polysulfide chain as a bridging chain because it has a more rigid framework than mide to generate the corresponding zinc dithiolene complex 4, that of the polymethylene chain, and large interplanar interwhich was finally reacted with sulfur dichloride in dry aceto- action between two TTF moieties may be achieved in a dimer.nitrile at 55 °C. The resultant precipitates were filtered and washed with methanol. The crude product was purified by column chromatography (silica gel, carbon disulfide) and recrystallized from chlorobenzene to aVord air-stable black plate-like crystals in 17% yield.† X-Ray crystal structure analysis‡ revealed that the TTF dimer 1 has a cyclophane-like U-shape structure (Fig. 1). Each of the two TTF moieties has S S S S S S S S S S S S S S S S S S MeS MeS S S S S S S MeS MeS S S 2 1 a planar but slightly bent structure with dihedral angles of Furthermore it is expected to show the intermolecular inter- 16.4 and 24.1°.They stack in a ring-over-bond overlap mode action through the sulfur–sulfur contacts between the polysulwith an intradimer interplanar distance of 3.86 A ° . TTF dimers fide chains. The trisulfide double-bridged molecule 2 has been form sheet-like structures in the crystal which resemble the reported as the 1,3-dithiole-2-thione derivative,5 however, no dimerized stacking structure in conducting sheets.The electro- TTF analog has been prepared so far. In this communication chemical properties were investigated by cyclic voltammetry. we report the synthesis and structure of a novel cyclophane- As shown in Fig. 2, the TTF dimer shows two pairs of reversible like trisulfide-bridged TTF dimer 1 in which two TTF moieties one-electron redox waves (+0.57 and 0.67 V vs.Ag/AgCl in are connected to each other by two trisulfide chains. We also PhCN) and one pair of reversible two-electron redox waves describe the structure and property of its dicationic ClO4 salt. Usually the synthesis of TTF derivatives is carried out by † Mp 144–145 °C (decomp.); 1H NMR (270 MHz, carbon disulfide, the phosphite mediated coupling reaction of thione or ketone [2H6]benzene): d 2.27 (12H, s); Calc.for C16H12S18: C, 24.60; H, precursors. However the bis(1,3-dithiole-2-thione) analog 2 1.55%. Found: C, 24.95; H, 1.82%. was reported to react with trimethyl phosphite to yield the ‡ Crystal data for 1: C16H12S18, M=781.35, monoclinic, space group thiophosphonate.5 Therefore we can not obtain the TTF dimer C2/c, a=12.597(8), b=10.62(1), c=44.867(9) A ° , b=91.93(4)°, V= 1 by the cross-coupling reaction.Thus the synthesis of TTF 5996(7) A ° 3, Z=8, m(Mo–Ka)=13.03 cm-1. The total number of independent reflections measured at room temperature was 6335, of which dimer 1 was performed according to Scheme 1. 4,5-Bis(2¾- 1478 were considered to be observed [I>3.00s(I)]. The structure was cyanoethylthio)-4,5-bis(methylthio)-TTF 36 was hydrolyzed refined by full-matrix least squares to R=0.054, Rw=0.046. Full with 28 wt% methanol solution of sodium methoxide in dry crystallographic details, excluding structure factors, have been acetonitrile–methanol at room temperature and successively deposited at the Cambridge Crystallographic Data Centre (CCDC).treated with zinc chloride and tetra-n-butylammonium bro- 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 1145/83. * E-mail: fuji@ims.ac.jp J. Mater. Chem., 1998, 8(4), 829–831 829Fig. 3 Molecular structure of dicationic donor molecule 12+ in the Fig. 1 Molecular structure of 1: (a) top view and (b) side view ClO4 salt: (a) top view and (b) side view Fig. 2 Cyclic voltammogram of 1 under the following conditions: Fig. 4 Crystal structure of 1·(ClO4)2. One ClO4 anion is pentahedrally Bu4NPF6 0.1 mol dm-3 in PhCN, Pt electrode, 20 °C, scan rate disordered. 50 mV s-1, E vs.Ag/AgCl crystal structure analysis.§ Unfortunately the composition of (+1.15 V). The first oxidation potential is slightly high com- § Crystal data for the ClO4 salt: C16H12S18Cl2O8, M=980.25, triclinic, pared to that of BEDT–TTF (0.53 V). The first and second space group P19, a=11.348(2), b=20.044(7), c=8.053(2) A ° , a= oxidation processes are separated by an interval of 0.1 V, 94.77(2), b=99.30(1), c=106.21(2)°, V=1719.9(9) A° 3, Z=2, suggesting the existence of intradimer interactions between m(Mo–Ka)=13.23 cm-1. The total number of independent reflections two TTF moieties.The preparation of cation radical salts was measured at room temperature was 3305, of which 1669 were conperformed in THF by electrochemical oxidation, and the sidered to be observed [I>3.00s(I)].The structure was refined by full-matrix least squares to R=0.042, Rw=0.041. structure of the ClO4 salt obtained was revealed by X-ray 830 J. Mater. Chem., 1998, 8(4), 829–831Chem. Commun., 1991, 843; (c) M. Adam, V. Enkelmann, H.-J. the ClO4 salt is 152 (donor5anion) and the donors are in the Ra�der, J. Ro� hlich and K. Mu� llen, Angew.Chem., Int. Ed. Engl., 1991, dicationic state (i.e. each TTF is present as the radical cation 31, 309; (d) K. Takimiya, Y. Shibata, K. Imamura, A. Kashihara, species). Fig. 3 shows the molecular structure of the dicationic Y. Aso, T. Otsubo and F. Ogura, T etrahedron L ett., 1995, 36, 5045; TTF dimer. The TTF moieties have a planar structure and (e) J. Tanabe, T. Kudo, M. Okamoto, Y.Kawada, G. Ono, A. Izuoka eclipse each other almost perfectly with a comparatively short and T. Sugawara, Chem. L ett., 1995, 579; ( f ) K. Matsuo, K. Takimiya, Y. Aso, T. Otsubo and F. Ogura, Chem. L ett., 1995, interplanar distance of 3.39 A ° compared to that of the neutral 523; (g) K. Takimiya, Y. Aso, F. Ogura and T. Otsubo, Chem. L ett., one (3.86 A ° ). Each pair of adjacent dimers is almost separated 1995, 735; (h) K.Takimiya, Y. Aso and T. Otsubo, Synth.Met., 1997, by the counter anions and the interaction between donors 86, 1891; (i ) K. Takimiya, K. Imamura, Y. Shibata, Y. Aso, F. Ogura exists only through the sulfur atoms of methylthio groups in and T. Otsubo, J. Org. Chem., 1997, 62, 5567; ( j) Y. Yunoki,irection of the molecular long axis (Fig. 4). The electrical K. Takimiya, Y. Aso and T. Ostubo, T etrahedron L ett., 1997, 38, 3017. properties of this ClO4 salt were measured; it appeared to be 4 (a) A. Kobayashi, R. Kato, H. Kobayashi, S. Moriyama, Y. Nishio, an insulator because of the dicationic state of the donors and K. Kajita and W. Sasaki, Chem. L ett., 1987, 459; (b) H. Urayama, its mixed stacking crystal structure.Preparation of other H. Yamochi, G. Saito, K. Nozawa, T. Sugano, M. Kinoshita, S. Sato, derivatives of 1 and other conducting salts is in progress. K. Oshima, A. Kawamoto and J. Tanaka, Chem. L ett., 1988, 55; (c) A. M. Kini, U. Geiser, H. H. Wang, K. D. Carlson, J. M. Williams, W. K. Kwok, K. G. Vandervoort, J. E. Thompson, D. L. Stupuka, D. Jung and M.-H. Whangbo, Inorg.Chem., 1990, 29, 2555. References 5 (a) X. Yang, T. B. Rauchfuss and S. R. Wilson, J. Am. Chem. Soc., 1 (a) J. Ferraris, D. O. Cowan, V. V. Walatka and J. H. Perlstein, 1989, 111, 3465; (b) C. P. Galloway, D. D. Doxsee, D. Fenske, J. Am. Chem. Soc., 1973, 95, 948; (b) M. Narita and C. U. Pittman, T. B. Rauchfuss, S. R. Wilson and X. Yang, Inorg. Chem., 1994, Jr., Synthesis, 1976, 489; (c) A. Krief, T etrahedron, 1986, 42, 1209; (d) 33, 4537. G. Schukat, A. M. Richter and E. Fangha�nel, Sulfur Rep., 1987, 7, 6 (a) N. Svenstrup, K. M. Rasmussen, T. K. Hansen and J. Becher, Synthesis, 1994, 809; (b) L. Binet, J. M. Fabre, C. Montginoul, 155; (e) M. R. Bryce, Chem. Soc. Rev., 1991, 20, 355. K. B. Simonsen and J. Becher, J. Chem. Soc., Perkin T rans. 1, 1996, 2 T. Otsubo, Y. Aso and K. Takimiya, Adv. Mater., 1996, 8, 203. 783. 3 (a) H. A. Staab, J. Ippen, C. Tao-pen, C. Krieger and B. Starker, Angew. Chem., Int. Ed. Engl., 1980, 19, 66; (b) F. Bertho-Thoraval, A. Robert, A. Souizi, K. Boubekeur and P. Batail, J. Chem. Soc., Communication 8/00202I; received 6th January, 1998 J. Mater. Chem., 1998, 8(4), 829–831 8

ISSN:0959-9428    

DOI:10.1039/a800202i

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 nonlinear optical chromophores containing electronexcessive and -deficient heterocyclic bridges. The auxiliary donor– acceptor eVects Ching-Fong Shu* and Yuh-Kai Wang Department of Applied Chemistry, National Chiao T ung University 1001 T a-Hsueh Road, Hsin-Chu, Taiwan, 30035, R.O.C. treated with BunLi at -78 °C, and the halogen–lithium exchange product was quenched with N-formylmorpholine to Push-pull substituted nonlinear optical chromophores with thiazole and thiophene rings and interposed ethylene units as produce 2-formylthiazole 1.6 Reduction of the formyl group of 1 with NaBH4 gave 2-hydroxymethylthiazole 2, which was p-conjugated bridges were synthesized.The eVects of the nature and location of the heterocycles on the energy of the transformed into 2-chloromethylthiazole 3 upon chlorination with carbon tetrachloride–triphenylphosphine.The reaction of charge transfer transition for the chromophores are discussed. chloromethyl derivative 3 with triphenylphosphine yielded the corresponding phosphonium salt 4.7 Wittig reaction of 2- diethylamino-5-formylthiophene8 with the phosphonium salt 4 in ButOK–benzene led to compound 5a.The thiazole por- Dipolar chromophores of the form D-p-A, where D (A) is an tions of compounds 7 were prepared from 2-chlorothiazole,9 electron donor (acceptor) group and p is a conjugated bridge which was allowed to react with BunLi yielding 2-chloro-5- possessing large molecular second-order nonlinear optical lithiothiazole.Adding N,N-diethylformamide to the 2-chloro- (NLO) responses and good thermal stability, are currently of 5-lithiothiazole solution and quenching with water yielded 2- interest because of their applicability to electro-optic devices.1 diethylamino-5-formylthiazole 6.10 Wittig–Horner conden- In such molecules, the donor and acceptor substituents provide sation of diethyl 2-thienylmethylphosphonate with the thiazole the requisite ground-state charge asymmetry, whereas the p- aldehyde derivative 6 produced compound 7a.Compounds 5a conjugation system provides a pathway for the redistribution of electric charges under the influence of electric fields. Benzene rings are widely used in combination with polyenes in pconjugated bridges because of their thermal and oxidative stability, synthetic availability and substituent positional selectivity.However, the barrier resulting from the aromatic delocalization energy of the benzene ring leads to diminution of b values.2 Synthetic studies have demonstrated that replacing the benzene ring of a chromophore p bridge with easily delocalizable five-membered heteroaromatic rings, such as thiazole and thiophene, results in enhanced molecular hyperpolarizability. 3 Recently, computational studies have suggested that heteroaromatic rings play a subtle role in influencing the second-order NLO response properties of donor–acceptor compounds.4 While the aromaticity of heteroaromatics aVects electronic transmissions between donor and acceptor substituents, the electron-excessive or electron-deficient nature of the heterocyclic ring systems also plays a major role in determining the overall electron-donating and -accepting ability of the substituents: electron-excessive heterocycles act as auxiliary donors and electron-deficient heterocycles act as auxiliary acceptors.4b Thus, attaching a strong donor to an electronexcessive heteroaromatic, such as thiophene,5 and a strong electron acceptor to an electron-deficient heteroaromatic, such as thiazole,5 will yield chromophores with significantly enhanced NLO responses.Contrarily, reversing the architecture of the heterocyclic rings in the p-conjugated system will lead to chromophores with diminished NLO responses. In this report, we present novel NLO chromophores with p-conjugating moieties that contain thiazole and thiophene, with an ethylene segment in between, and show that the location of the heterocyclic rings indeed has a pronounced eVect on the energy of charge-transfer (CT) transition of the chromophores, as was predicted in previous theoretical calcu- N S Br N S O N S OH N S Cl i ii iii 1 2 3 N S Cl N S PPh3 4 6 Cl– iv N S N O 7 N S N S R 5 S N N S R vi v vii 5a, 7a R = H 5b, 7b R = CHO 5c, 7c R = C(CN)=C(CN)2 5d, 7d R = CH=C(CN)2 viii ix x lations.4 The synthetic routes used to prepare the NLO Scheme 1 Reagents: i, BunLi, N-formylmorpholine; ii, NaBH4; iii, CCl4, chromophores 5c–d and 7c–d containing thiazole and thio- PPh3; iv, PPh3; v, 2-diethylamino-5-formylthiophene, ButOK; phene are shown in Scheme 1.The thiazole parts of compounds vi, BunLi, N,N-diethylformamide; vii, diethyl 2-thienylmethylphosphonate, ButOK; viii, BunLi, DMF; ix, BunLi, TCNE; x, CH2(CN)2 5 were synthesized starting from 2-bromothiazole, which was J.Mater. Chem., 1998, 8(4), 833–835 833Table 1 Electronic absorptions and thermal stabilities lmax/nm chromophore in dioxane e/104 M-1 cm-1 Td/°Ca 702 5.84 217 590 4.18 237 622 4.97 242 514 2.94 268 640b 245 (240b) 513c aDSC, 10 deg min-1.bRef 13. cRef 3(d). and 7a, which lack electron-withdrawing groups at the ends have slightly blue-shifted CT bands. This result again indicates that replacing a benzene ring on the donor end with an of their conjugating moieties, were lithiated with BunLi and subsequently quenched with tetracyanoethylene and N,N- electron-deficient heterocycle, such as thiazole, decreases the electron-donating ability of the dialkylamine group, and dimethylformamide to give formyl derivatives 5b and 7b, and tricyanovinyl-substituted 5c and 7c.A further Knoevenagel counteracts the eVects of reduced aromaticity. The relative thermal stabilities of the chromophores were reaction of 5b and 7b with malononitrile aVorded the dicyanovinyl- substituted chromophores 5d and 7d.All the compounds also studied using diVerential scanning calorimetry (DSC). The onset temperatures (Td) of the chromophore thermal decompo- studied here were characterized using conventional spectroscopic techniques.11 sition exotherms are shown in Table 1. Although, as expected, the replacement of a benzene ring with a heterocyclic ring does Most donor–acceptor substituted NLO chromophores are characterized by a long-wavelength CT transition that contrib- not lead to substantially lower thermal stability, we note that chromophores with lower charge-transfer transitional energies utes strongly to the second-order molecular hyperpolarizability. 12 All of the chromophores in this study had strong tend to possess lower thermal stabilities.In summary, compounds 5c,d and 7c,d represent the first bands in the visible region of the spectrum. Table 1 lists the positions of the absorption maxima for the chromophores in example of NLO chromophores synthesized in which the conjugating moieties contain both electron-excessive hetero- this study, and for related chromophores.3d,13 Chromophores 5c,d, which have an electron-excessive heterocycle (thiophene) cycles (thiophene) and electron-deficient heterocycles (thiazole). 5 As predicted by theoretical calculations, the CT on their donor ends, and an electron-deficient heteocycle(thiazole) on their acceptor ends, show pronounced bathochromic transitional energies of the chromophores depends not only on the natures but also strongly on the locations of the shifts (ca. 80 nm) in their CT bands as compared with chromophores 7c,d, which have reversed heterocycle substitution heterocyclic rings, which in turn may have significant eVects on the molecular second-order NLO response properties.4 patterns. This result is reasonable since in chromophores 5c,d, the thiophene and thiazole rings act as additional donor/ Experiments to measure the second-order molecular hyperpolarizabilities of these chromophores are in progress.acceptor groups, facilitating CT transitions. In chromophores 7c,d, the thiazole ring reduces the donation ability of the donor substituent and the thiophene ring attenuates the acceptance We are grateful to Professor Alex K.-Y. Jen, Northeastern University, for useful discussions.We also thank the National power of the acceptor substituent, an arrangement that impedes CT transitions. When compared to chromophores 8c,d, Science Council (ROC) (NSC 85-2113-M-009-003) for financial support. chromophores 7c,d have no red-shifted CT bands, and even 834 J. Mater. Chem., 1998, 8(4), 833–8357 A. Dondoni, G. Fantin, M. Fogagnolo, A. Medici and P.Pedrini, References T etrahedron, 1988, 44, 2021. 8 F. A. Mikhailenko and L. I. Shevchuk, Chem. Heterocycl. Compd. 1 (a) P. N. Prasad and D. J. Williams, Introduction to Nonlinear (Engl. T ransl.), 1974, 10, 1151. Optical EVects in Molecules and Polymers, Wiley, New York, 1991; 9 K. Ganapathi and A. Venkataraman, Proc. Indian Acad. Sci., Sect. (b) D. R. Kanis, M. A. Ratner and T.J. Marks, Chem. Rev., 1994, A, 1945, 22, 362. 94, 195; (c) L. R. Dalton, A. W. Harper, R. Ghosn, W. H. Steier, 10 I. Sawhney and J. R. H. Wilson, J. Chem. Soc., Perkin T rans. 1, M. Ziari, H. Fetterman, Y. Shi, R. V. Mustacich, A. K.-Y. Jen and 1990, 329. K. J. Shea, Chem.Mater., 1995, 7, 1060. 11 1H NMR spectra were recorded using CD2Cl2 as solvent and TMS 2 (a) K. D.Singer, J. E. Sohn, L. A. King, H. M. Gordon, H. E. Katz as internal standard. Chemical shifts are in ppm, coupling con- and C. W. Dirk, J. Opt. Soc. Am. B, 1989, 6, 1339; (b) S. R. Marder, stants are in Hz. 5c: d 1.21 (t, 6H, J 7.1), 3.42 (q, 4H, J 7.1), 6.01 (d, D. N. Beratan and L.-T. Cheng, Science, 1991, 252, 103. 1H, J 4.6), 6.46 (d, 1H, J 14.3), 7.24 (d, 1H, J 4.6), 7.85 (d, 1H, J 3 (a) C.W. Dirk, H. E. Katz, M. L. Schilling and L. A. King, Chem. 14.3), 8.34 (s, 1H). 5d: d 1.17 (t, 6H, J 7.1), 3.33 (q, 4H, J 7.1), 5.82 Mater., 1990, 2, 700; (b) L.-T. Cheng, W. Tam, S. R. Marder, (d, 1H, J 4.3), 6.48 (d, 1H, J 14.9), 7.05 (d, 1H, J 4.3), 7.69 (d, 1H, J A. E. Steigman, G. Rikken and C. W. Spangler, J. Phys. Chem., 14.9), 7.71 (s, 1H), 7.99 (s, 1H). 7c: d 1.18 (t, 6H, J 7.0), 3.48 (q, 4H, 1991, 95, 10643; (c) V. P. Rao, A. K.-Y. Jen, K. Y. Wong and J 7.0), 6.47 (d, 1H, J 15.5), 7.05 (d, 1H, J 4.4), 7.32 (d, 1H, J 15.5), K. J. Drost, T etrahedron L ett., 1993, 34, 1747; (d) A. K.-Y. Jen, 7.35 (s, 1H), 7.83 (d, 1H, J 4.4). 7d: d 1.17 (t, 6H, J 7.1), 3.45 (q, 4H, V. P. Rao, K. Y. Wong and K. J. Drost, J. Chem. Soc., Chem. J 7.1), 6.48 (d, 1H, J 15.5), 6.96 (d, 1H, J 4.2), 7.20 (d, 1H, J 15.5), Commun., 1993, 90; (e) S.-S.P. Chou, D.-J. Sun, H.-C. Lin and P.- 7.22 (s, 1H), 7.49 (d, 1H, J 4.2), 7.65 (s, 1H). HRMS (m/z) 5c: Obs. K. Yang, T etrahedron L ett., 1996, 37, 7279. 365.0745, calc. 365.0768; 5d: obs. 340.0810, calc. 340.0816; 7c: obs. 4 (a) P. R. Varanasi, A. K.-Y. Jen, J. Chandrasekhar, 365.0751, calc. 365.0768; 7d: obs. 340.0814, calc. 340.0816. I. N. N. Namboothiri and A. Rathna, J. Am. Chem. Soc., 1996, 118, 12 (a) J. L. Oudar and P. S. Chemla, J. Chem. Phys., 1977, 66, 2664; 12443; (b) I. D. L. Albert, T. J. Marks and M. A. Ratner, J. Am. (b) J. L. Oudar and J. Zyss, Phys. Rev. A., 1982, 26, 2016. Chem. Soc., 1997, 119, 6575. 13 A. K.-Y. Jen, Y. Cai, P. V. Bedworth and R. S. Marder, Adv.Mater., 5 T. L. Gilchrist, Heterocycl Chemistry, Wiley, New York, 1985. 1997, 9, 132. 6 A. Dondoni, G. Fantin, M. Fogagnolo, A. Medici and P. Pedrini, Synthesis, 1987, 998. Communication 7/08689J; Received 2nd December, 1997 J. Mater. Chem., 1998, 8(4), 833–835 835

ISSN:0959-9428    

DOI:10.1039/a708689j

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 A novel cathode Li2CoMn3O8 for lithium ion batteries operating over 5 volts Hiroo Kawai,a Mikito Nagata,b Hisashi Tukamotob and Anthony R. Westa aDepartment of Chemistry, University of Aberdeen, Meston Walk, Aberdeen, UK AB24 3UE bCorporate R&D Center, Japan Storage Battery Company L imited, Nishinisho, Kisshoin, Minami-ku, Kyoto 601, Japan package.Powder XRD patterns for various cation distributions on 8a and 16d sites were generated theoretically, using the The spinel phase, Li2CoMn3O8, can be used as a cathode in atomic positions of the spinel phase Li4Mn5O12, i.e., secondary lithium batteries, Li/LiPF6,propylene Li8a[Li0.33Mn1.67]16dO4,12 with isotropic thermal parameters carbonate/Li2CoMn3O8, with a discharge capacity of ca.of 0.05 for all atomic positions, and compared with the 40 mAh g-1 at 5.0 V increasing to ca. 130 mAh g-1 at 3.8 V. observed pattern. Rietveld refinement was then carried out for This appears to be the first single cell system to operate models having likely cation distributions. The most probable reversibly over 5 volts. model, with minimum R values,13 was found to be Li8a[Co0.5Mn1.5]16dO4 : Li fully occupies tetrahedral 8a sites; Co and Mn are distributed in octahedral 16d sites with a molar ratio of 153.Cation ordering in octahedral 16d sites remains a possibility because, although there is no evidence of Lithium ion rechargeable batteries which employ lithium superstructure reflections in Fig. 1, Co and Mn have very transition metal oxide cathodes and carbon anodes have similar atomic scattering factors.applications ranging in size from portable electronic devices For conductivity measurements, a pellet (8 mm diameter; to zero emission vehicles (ZEV).1 Recent developments include 2–3 mm thickness) was cold-pressed uniaxially at 150 MPa, 4 V cathodes based on LiCoO2, LiNiO2 and LiMn2O4,2 low sintered at 950 °C for a few hours in order to increase the temperature synthesis of new cathodes such as LiMnO23 and mechanical strength, slowly cooled to 600 °C and maintained Li1.5Na0.5MnO2.85I0.124 working below ca. 4 V and improved at 600 °C for 3 days before quenching to room temperature. electrolytes5 which have made it possible to explore the In/Ga paste electrodes were coated on opposite sides of the potential range to 5 V vs.Li/Li+. Tarascon et al.6 observed sintered pellet. Impedance measurements were carried out over two oxidation–reduction peaks near 4.5 V and 4.9 V for the frequency range 30 mHz–1 MHz, using combined LiMn2O4. Cells with high operating voltage have been reported Solartron 1250/1286 and Hewlett-Packard 4192 instrumenwith cathodes based on spinel structure compounds, including tation.The data were analyzed in the complex impedance (Z*), 4.8 V for LiNiVO47 and LiCrXMn2-XO48 and 4.7 V for admittance (Y *) and modulus (M*) formalisms,14 using in- LiNiXMn2-XO4.9,10 For some electrochemists, especially house software. aiming to commercialize high power ZEV, it has been a Impedance data at -23 °C are shown in Fig. 2.Three arcs primary target to discover a cathode which works over 5 V, are seen. The high frequency well resolved arc has an associated linked to high specific capacity and good cycling stability. We capacitance of 4.7×10-12 F, typical of a bulk response. The show here that a complex spinel Li2CoMn3O8 possesses an middle frequency arc, with an associated capacitance of operating voltage over 5 V, with excellent cycle performance 4.5×10-10 F, corresponds to a grain boundary resistance.The and specific discharge capacity, up to the potential limit of low frequency arc has capacitance ca. 2.3×10-7 F, and may 5.3 V, of ca. 40 mAh g-1. To our knowledge, this is the first be indicative of Schottky barrier phenomena15 at the sample– cathode working over 5 V. electrode interface.No spikes attributable to a blocking elec- An oxidizing atmosphere was used in the initial synthesis of trode response were seen in the measured temperature and the spinel phase Li2CoMn3O8.11 In the present study, frequency ranges. The main current carriers thus appear to be Li2CoMn3O8 was obtained more simply by conventional cer- electrons. Bulk conductivity data are shown as a function of amic synthesis in air.Starting materials were Li2CO3, CoO temperature in Arrhenius format in Fig. 3. Activation energy and MnCO3, all reagent grade. A stoichiometric mixture was for conduction is estimated to be ca. 0.31 eV. The conductivity ground intimately, fired in air, initially at 650 °C to drive oV is ca. 6.8×10-4 S cm-1 at 30 °C. High electronic conductivity CO2 , then at 800 °C for 3 days with intermittent regrinding, in oxides is most frequently observed in mixed valency semito complete the reaction, and finally at 600 °C for 3 days, conductors.16 In the case of Li2CoMn3O8, if some of the Mn4+ before quenching to room temperature.Oxygen loss from the are reduced to Mn3+, charge balance is preserved by oxidation as-prepared sample occurred above ca. 650 °C, involving poss- of Co2+ to Co3+. Thus, from a combination of the Rietveld ible structural changes, and thus the above synthesis procedure refinement and impedance results, Li2CoMn3O8 is most with the final annealing at 600 °C was chosen as the optimum likely written as Li8a[Co2+0.5-YCo3+YMn3+YMn4+1.5-Y]16dO4, conditions to obtain a pure phase. Further study on the 0<Y0.5 (exact value not known).phase(s) formed above ca. 650 °C are in progress. Electrochemical measurements were made in a glass test cell The powder X-ray diVraction (XRD) pattern of containing a Li2CoMn3O8 composite as the positive electrode, Li2CoMn3O8 (Stoe Stadi/P diVractometer, Cu-Ka1 radiation) Li metal foil as the negative electrode, Li metal as a reference was indexed in the cubic space group Fd3m with lattice and 1 M LiPF6 dissolved in propylene carbonate as the electroparameter a=8.1317(17) A ° , Fig. 1. High fluorescence back- lyte. The Li2CoMn3O8 composite electrode was prepared by grounds, caused by an interaction between Co/Mn and Cu- blending 5 wt.% acetylene black as an active material and Ka1 radiation, allowed us to determine only a likely cation 5 wt.% mixture of polyvinylidene fluoride and N-methyldistribution by the Rietveld method, using the pattern fitting 2-pyrrolidone as a binder.Cycle tests were carried out galvanostatically at a current density of 0.5 mA cm-2. structure refinement (PFSR) program in the Stoe software J. Mater. Chem., 1998, 8(4), 837–839 837Fig. 1 Powder XRD profile for Li2CoMn3O8 Fig. 3 Conductivity Arrhenius plot for Li2CoMn3O8 Fig. 2 Impedance data for Li2CoMn3O8 at -23 °C The cell was cycled four times between 3.0 V and 5.1 V, 76.5 mAh g-1, which corresponds to a change in X of 1.04. The composition at 5.3 V is then estimated to be before charging to 5.3 V. Fig. 4 shows a potential profile for the first 5.3 V cutoV measurement. Two plateaux are seen in Li0.12CoMn3O8, indicating that at least 6% Li still remains inside the positive electrode.It is thus likely that the plateau both charge and discharge curves and correspond to two-step extraction/reinsertion of lithium from/into the cathode mate- centered on 5.2 V continues over 5.3 V, finally yielding a plausible end composition Co4+Mn4+3O8 . Improvement in rial. On charge, the plateau centered on 4.0 V has the capacity, estimated at the sharp transition point between the electrolyte stability is needed to measure electrochemical properties at higher potentials.Discharge capacity is esti- two plateaux, of 61.8 mAh g-1. On converting the capacity to X in Li2-XCoMn3O8 , X reaches 0.84 at the transition mated roughly to be ca. 60 mAh g-1 at the plateau centered on 5.1 V and ca. 70 mAh g-1 at the plateau centered on 3.9 V, point. The potential scan was limited to 5.3 V owing to possible electrolyte decomposition at higher potentials. Up to Fig. 4. The two plateaux appear to originate in the redox reactions (1) Co2+16d<Co3+16d, (2) Co 3 +16d<Co4+16d, (3) 5.3 V, the plateau centered on 5.2 V possesses the capacity of 838 J. Mater.Chem., 1998, 8(4), 837–839lithium ion batteries.17 LiMnO23 and Li1.5Na0.5MnO2.85I0.124 are possible candidates for substituting for the expensive and toxic LiCoO2, subject to improved cycle performance in the former and working voltage in the latter. Possible applications are, however, limited to portable electronic devices, owing to operating voltages below ca. 4 V. The new Li2CoMn3O8 cathode reported here possesses a very high working voltage, over 5 V for part of the charge–discharge cycle, with excellent cycling stability.Easy preparation procedure reduces the costs for synthesis. Although the specific capacity at the plateau centered on 5.1 V needs to be improved, Li2CoMn3O8 and its possible solid solutions are promising candidates for cathodes in high voltage (>5.0 V) lithium ion cells, which could be used in high power, emission-free vehicles, with improved oxidation–resistant electrolytes.H. K. thanks CVCP for an ORS Award. Fig. 4 Potential profile for Li2CoMn3O8 References 1 See, for example, K. Brandt, Solid State Ionics, 1994, 69, 173. 2 See, for example, R. Koksbang, J. Barker, H. Shi and M. Y. Saidi, Solid State Ionics, 1996, 84, 1. 3 A. R. Armstrong and P. G. Bruce, Nature, 1996, 381, 499. 4 J. Kim and A. Manthiram, Nature, 1997, 390, 265. 5 J.M. Tarascon and D. Guyomard, Solid State Ionics, 1994, 69, 293. 6 J. M. Tarascon, W. R. McKinnon, F. Coowar, T. N. Bowmer, G. Amatucci and D. Guyomard, J. Electrochem. Soc., 1994, 141, 1421. 7 G. T-K. Fey, W. Li and J. R. Dahn, J. Electrochem. Soc., 1994, 141, 2279. 8 C. Sigala, D. Guyomard, A. Verbaere, Y. PiVard and M. Tournoux, Solid State Ionics, 1995, 81, 167. 9 K. Amine, H. Tukamoto, H. Yasuda and Y. Fujita, J. Electrochem. Soc., 1996, 143, 1607. 10 Q. Zhong, A. Bonakdarpour, M. Zhang, Y. Gao and J. R. Dahn, J. Electrochem. Soc., 1997, 144, 205. 11 G. Blasse, J. Inorg. Nucl. Chem., 1964, 26, 1473. 12 T. Takada, H. Hayakawa and E.Akiba, J. Solid State Chem., 1995, Fig. 5 Variation in discharge capacity of the plateau centered on 3.9 V 115, 420. ($) and the plateau centered on 5.1 V (#) upon cycling for 13 H. M. Rietveld, J. Appl. Crystallogr., 1969, 2, 65. Li2CoMn3O8 (& denotes the total discharge capacity; &=$+#) 14 I. M. Hodge, M. D. Ingram and A. R. West, J. Electroanal. Chem., 1976, 74, 125. Co2+16d<Co4+16d and/or (4) Mn3+16d<Mn4+16d . Further 15 See, for example, H. H. Sumathipala, M. A. K. L. Dissanayake and work is needed to elucidate which reactions are responsible A. R. West, J. Electrochem. Soc., 1995, 142, 2138. 16 E. J. W. Verwey, P. B. Braun, E. W. Gorter, F. C. Romeijn and for the electrochemical performance. Fig. 5 shows the vari- J. H. van Santen, Z. Phys. Chem., 1951, 198, 6. ation in discharge capacity with cycle number. Both plateaux 17 T. Nagaura and K. Tozawa, Prog. Batteries Sol. Cells, 1990, 9, 209. centered on 5.1 V and 3.9 V exhibit excellent cycling stability, maintaining their discharge capacity to well over 50 cycles. Communication 8/00604K; Received 22nd January, 1998 To date, LiCoO2 is the only cathode used in commercial J. Mater. Chem., 1998, 8(4), 837–839 839

ISSN:0959-9428    

DOI:10.1039/a800604k

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 New conductive molecular composites: aniline derivatives as guest molecules encapsulated and polymerized within the channels of the host 3D-coordination polymers [(Me3E)3Fe(CN)6]2 where E=Sn or Pb Amany M. A. Ibrahim Department of Chemistry, Faculty of Science, Ain Shams University, Cairo, Egypt New conductive molecular composites have been prepared by chemical oxidation of polyanilines within the expandable wide channels of the 3D-coordination polymers [(Me3E)3Fe(CN)6]2 where E=Sn or Pb, in the presence and absence of HCl.The UV and visible absorption spectra of the protonated and deprotonated molecular composites are compared with those of polyanilines, polyemeraldine and nigrosine. The protonated molecular composites exhibit conductivities higher than the deprotonated ones.Also, polymerization of some aniline derivatives within the channels of the 3D-polymers, in absence of HCl, leads to vast improvement in conductivity owing to self protonation of the polyemeraldine base. The dramatic shift in optical response of the molecular composite containing polyemeraldine base with change in pH could be used as an optical pH sensor over the pH range investigated.Polyanilines can be readily protonated and deprotonated,24 Introduction and yet the number of electrons on the polymer chain remains The interest in composite materials formed from conducting constant. This has led to the proposal of a two-dimensional polymers and conventional polymers is sparked by a desire to surface to adequately describe the state of the polymer.25,26 couple the physical properties of conventional polymers with However, there are still a great deal of unresolved details the conductivity exhibited by conducting polymers.1–3 The concerning the structure of the polymer chain, and the conforattractive characteristic features of 3D-coordination polymers mational and electronic changes.of the type [(Me3E)3M(CN)6]2 suggested the idea that aniline Here, the electronic absorption spectra of polyaniline and and its derivatives could be encapsulated into the expandable its derivatives encapsulated within the channels of the 3Dchannels of these polymers to yield conductive molecular polymers [(Me3E)3M(CN)6]2 (E=Sn or Pb) are studied composites.4 These coordination polymers can be synthesized and compared with those of benzene, aniline, nigrosine and by the reaction of organotin(IV) or organolead(IV) compounds polyemeraldine described quantitatively by CNDO/S3 conwith hexacyano–d-transition metal ions (Fe and Co) to form figuration-interaction (CI) calculations.27,28 unprecedented modes of supramolecular architecture having three-dimensional polymeric networks.5,6 These 3D-polymers have zeolite like structures involving Me3E(NC)2 units with Experimental trigonal bipyramidal configuration and remarkably wide parallel channels with cross-sections of ca. 10×10 A° , whose walls The host 3D-polymers were prepared by dissolving a 351 are internally coated by constituents of the lipophilic Me3E molar ratio of Me3SnCl or Me3PbCl and K3Fe(CN)6 groups.6 in the minimum amount of water under a nitrogen atmos- The choice of polyaniline and its derivatives is due to their phere in the dark to yield [(Me3Sn)3Fe(CN)6]2 I and many interesting properties.Such properties make polyanilines [(Me3Pb)3Fe(CN)6]2 II. The precipitates were filtered oV, suitable for applications such as: materials for modified elec- washed with water then CH2Cl2 and dried under vacuum at trodes,7–10 corrosion inhibitors for semiconductors in photo- room temperature. The purity and identity of these 3Delectrochemical assemblies,11 in microelectronics,12,13 electro- polymers were checked as described elsewhere,4,6 Table 1.chromic materials,14–17 in rechargeable lithium batteries,18–21 Aniline derivatives were doubly distilled under reduced pressin high-resolution light images22 and recently in developing a ure while the solid starting materials were of highly pure grade conductive molecular ‘wire’ of near-molecular dimensions23 purchased from Aldrich or Merck.A cold (5 °C) 1 M HCl where the polyaniline filaments are essentially 3 nm wide solution, containing an excess of ammonium peroxodisulfate because they are formed inside the channels of an aluminowith respect to the desired reaction stoichiometry was added silicate crystal with 3 nm diameter pores.dropwise to a cold, stirred solution of aniline derivatives in Polyanilines represent a class of polymers built up from 1 M HCl to yield black polyaniline 1¾, poly(o-toluidine) 2¾, benzenoid (A) and quinonoid (B) repeat units.The combipoly( m-toluidine) 3¾, dark brown poly( p-toluidine) 4¾, black nation of one (A) and one (B) unit comprises ‘emeraldine’ poly(o-anisidine) 5¾, poly( p-anisidine) 6¾, poly(2-chloroaniline) whereas four (B) moieties constitute ‘nigrosine’; which is also 7¾, poly(3-chloroaniline) 8¾, brown poly(4-chloroaniline) 9¾, and referred to as ‘nigraniline’. However, polyanilines exhibit a greenish brown poly(2,4,6-trimethylaniline) 10¾.The molecular much greater complexity of properties and structure which are composites were prepared by the addition of cold (<5 °C) mainly due to the diVerent preparation conditions. aniline derivatives to the cold freshly prepared 3D-polymer I or II, suspended in a minimum amount of water, in a molar ratio of 351.This mixture was ground well in the absence of direct light for at least 30 min and then left for about 10 days. The molecular composites, 1–10 obtained from polymer I and 11 and 12 obtained from polymer II were washed with ethanol and dried under vacuum. The compositions of these molecular N H NH2 A N NH B J. Mater. Chem., 1998, 8(4), 841–846 841Table 1 Colour and elemental analysis of polymers I, II and their molecular composites elemental analysis found (calc.) (%) no.compound colour C H N Fe I [(Me3Sn)3Fe(CN)6]2 orange 25.61 3.87 11.94 7.94 (25.76) (3.93) (11.87) (7.86) II [(Me3pb)3Fe(CN)6]2 yellowish orange 18.59 2.81 8.67 5.76 (18.26) (3.02) (8.34) (5.68) 1 [(aniline)1.5+I]n black 34.19 4.48 12.46 6.62 (34.10) (4.62) (11.98) (6.42) 2 [(o-toluidine)1.5+I]n black 35.44 4.72 12.15 6.46 (35.30) (4.92) (12.00) (6.18) 3 [(m-toluidine)1.5+I]n black 35.44 4.72 12.15 6.46 (35.40) (4.87) (11.98) (6.25) 4 [(p-toluidine)1.25+I]n dark brown 34.06 4.57 12.13 6.66 (34.20) (4.82) (11.99) (6.51) 5 [(o-anisidine)1.5+I]n black 34.48 4.59 11.89 6.28 (34.31) (4.84) (11.21) (6.30) 6 [(p-anisidine)1.5+I]n black 34.48 4.59 11.89 6.28 (34.28) (4.91) (11.52) (6.18) 7 [(o-chloraniline)+I]n black 30.35 4.00 11.80 6.72 (30.26) (4.12) (11.20) (6.53) 8 [(m-chloraniline)+I]n black 30.35 4.00 11.80 6.72 (30.45) (4.36) (11.23) (6.63) 9 [(p-chloraniline)0.9+I]n brown 29.94 3.99 11.81 6.82 (29.90) (4.34) (11.42) (6.72) 10 [(2,4,6-trimethylaniline)+I]n greenish brown 34.37 4.81 11.69 6.66 (34.25) (4.67) (11.29) (6.63) 11 [(aniline)1.5+II]n black 28.98 3.80 10.56 20.83 (28.71) (3.72) (10.21) (20.62) 12 [(o-anisidine)1.5+II]n black 29.46 3.92 10.10 19.93 (29.31) (4.12) (9.72) (19.72) composites were investigated by elemental analysis, Table 1.The protonated molecular composites, 1H–10H, 11H and 12H, were prepared by the same procedure in the presence of 1 M HCl solution. They have the same colours as the corresponding polyaniline derivatives. The electronic absorption spectra were recorded on Shimadzu 240Wand Perkin-Elmer Model Lambda 3B doublebeam UV-VIS spectrophotometers within the wavelength range 200–1000 nm as Nujol mull matrices.Conductivity measurements of pressed pellets of diameter 0.45–0.5 cm and thickness 0.08–0.1 cm were performed at 298 K using the fourprobe technique and a super Megohmmeter Model RM 170 instrument (Avo Ltd., Dover, England).Results and Discussion The electronic absorption spectra of the new molecular composites 1–12, as Nujol mull matrices reveal mainly five absorption bands at 232–775 nm, Fig. 1 and 2 and Table 2. These bands are due to the electronic transitions of polyanilines and the anionic host 3D-polymers; tris (trimethyltin) (or lead) hexacyanoferrate(II).Thus, it would be of interest to investigate, separately, the spectra of the host 3D-polymers and those of the polyanilines. The spectra of the 3D-coordination polymers I and II in Nujol mull display three absorption bands at 220, 300 and 430 nm for I and at 225, 310 and 420 nm for II, respectively. The low energy band at 430 nm corresponds to p*÷p transitions of the FeIII(CN)6 building blocks. This band shifts to 320 and 325 nm in the spectra of the isostructural Fig. 1 The electronic absorption spectra of I, II and the molecular polymers [(Me3E)4Fe(CN)6]2, where E=Sn or Pb, respectcomposites 1–10 ively, and corresponds to p*÷p transitions of the FeII(CN)6 building blocks. Thus, these 3D-polymers are excellent materials for both accommodating the polyanilines within their within the individual benzenoid systems in polyanilines.28 They are due to the localized p*÷p transitions within the phenyl wide channels as well as being essentially optically transparent in the visible region.On the other hand, the spectra of moieties in the polymers as is commonly found in pendant group polymers.29 They show a slight blue shift relative to polyanilines 1¾–10¾ exhibit five absorption bands at 210–550 nm, Table 3.The first three bands at 210–230, 235–245 those of aniline and its derivatives. The broad band at 400–450 nm resembles the visible band of the in situ optical and 280–300 nm resemble those of aniline and its derivatives and correspond to 1B÷1A, 1La÷1A and 1Lb÷1A transitions absorption spectra of a polyaniline film on a conducting Pt 842 J. Mater.Chem., 1998, 8(4), 841–846band assignment, the first two bands appearing around 240 and 280 nm in the spectra of the molecular composites 1–12 are due to 1La÷1A and 1Lb÷1A transitions within the individual benzenoid systems in polyanilines. The band at 320–340 nm is attributed to p*÷p transitions of the FeII(CN)6 building blocks.The spectrum of 1 shows this as a broad composite band at 380 nm due to p*÷p transitions of both the FeII(CN)6 building blocks and polyemeraldine. The presence of the band at 320–340 nm and the disappearance of the band at 420 nm can be considered as good evidence for the occurrence of a redox reaction between the host polymers and the aniline derivatives acting as guest molecules.The last band at ca. 550 nm corresponds to a molecular exciton bound to locally distorted segments of the oligomer backbone. The presence of this band is a further indication of the redox reaction and polymerization of polyanilines within the channels of the host Fig. 2 The electronic absorption spectra of the molecular composites polymers. 11, 12, 11H and 12H In addition to the above bands, the spectra of the molecular composites 2, 3, 5, 8 and 12 exhibit a broad band at 730–775 nm.This band does not appear in the spectra of the electrode as a function of the applied oxidation potential.30 corresponding polyanilines or those of the other molecular This band could be taken as indicative of polymerization and composites. To throw light on this band, the spectra of the formation of polyemeraldine units rather than nigrosine units molecular composites prepared in the presence of 1 M HCl since it appears in the spectrum of polyemeraldine while it is were measured in Nujol mull matrices, Table 4 and Fig. 2. On not observed in the spectrum of nigrosine.31 This is also protonation, the bands around 410 nm and 730–775 nm in the supported by the presence of the band due to p*÷p transitions spectra of 2, 3, 5, 8 and 12 exhibit high intensity and are red of nigrosine at 335–365 nm in the spectra of 6¾, 9¾ and 10¾.shifted. On the other hand, the spectra of the other protonated Generally, p*÷p transitions are localized on benzenoid rings molecular composites show two absorption bands at of the polyemeraldine units. The last band located at 440–455 nm and 770–820 nm while the band around 550 nm 540–560 nm arises from the charge transfer exciton from the disappears and the high energy bands (lmax<340 nm) remain benzenoid to quinoid segments (pQ÷pB) of the emeraldine at more or less the same positions, Table 4.base similar to bulk polyaniline.32,33 The presence of this band These changes in absorption spectra upon protonation, are presents also an indication of the polymerization of aniline in accordance with the disappearance of the localised quinoid and its derivatives where the lmax for transitions of isolated structure and the formation of a polaron lattice or bipolaron chains should be much lower34 as observed in the spectra of 6¾ and 9¾ (ca.lmax=450 and 440 nm).According to the previous according to the following structures. Table 2 The electronic absorption spectra of the molecular composites 1–12 p*÷p p*÷p CT-exciton bipolaron compound 1La÷1A 1Lb÷1A FeII(CN)6 polyemeraldine pQ÷pB transitions 1 240 280 385 — 555 — 2 250 285 325 410 560 740 3 250 285 330 415 555 770 4 232 283 320 380 550wa — 5 240 283 320 420 538 730 6 240 280 325 390 540 — 7 240 280 320 380 565 8 242 285 330 410 520 775 9 245 285 335 390 480 — 10 245 285 340 410 580 — 11 240 285 310 380 570 — 12 240 280 305 420 560 750 aw=Weak.Table 3 Electronic absorption spectra of polyanilines 1¾–10¾ p*÷p CT-exciton compound 1B÷1A 1La÷1A 1Lb÷1A polyemeraldine pQ÷pB 1¾ 230 245 280 400 550 2¾ 220 235 300 — 550 3¾ 215 235 300 410 550 4¾ 220 241 300 415 540wa 5¾ 215 240 300 400 550 6¾ 210 240 300 360b 450 7¾ 215 240 280 415 550 8¾ 215 240 300 415 560 9¾ 215 236 280 335–365b 440 10¾ 218 245 300 340b 550wa aw=Weak.bNigrosine. J. Mater. Chem., 1998, 8(4), 841–846 843Table 4 The electronic absorption spectra of the protonated molecular composites 1H–12H p*÷p p*÷p polyemeraldine bipolaron polyemeraldine bipolaron compound base transitions compound base transitions 1H 450 790 7H 440 770 2H 450 805 8H 450 810 3H 455 820 9H 440 760 4H 440 790 10H 445 770 5H 450 800 11H 450 795 6H 455 780 12H 445 800 N N H H + N N H H + N N H H + Polaron lattice Bipolaron lattice The presence of these polaron or bipolaron states would produce two levels in the gap with three allowed directed optical transitions for the polaron and two for the bipolaron.32 Since the spectra comprise only two bands in the visible region, the bipolaron is the most favourable structure which may be formed upon protonation.The spectra of the deprotonated molecular composites 2, 3, 5, 8 and 12 show both bands due to an exciton (550 nm) and bipolaron lattice (730–775 nm). This is due to the fact that the monomers of the aniline derivatives are oxidatively polymerized within the channels of the host polymers, which N N H H + N N H H + n [(Me3E)3FeII(CN)6] n [(Me3E)3FeII(CN)6] n N N H H + N N H n [(Me3E)3FeII(CN)6] n [(Me3E)3FeII(CN)6] nH – – – – + are reduced to the isostructural anionic homologue Scheme 1 [(Me3E)3FeII(CN)6-]2, producing protons.These protons are captured by the negatively charged channels of the host polymer and can link some of the unprotonated imine nitrogens of the emeraldine base forming the partially protonated emeraldine.This implies the presence of a dynamic equilibrium proton transfer between the partially protonated emeraldine and the negatively charged channels of the polymer, Scheme 1. The dramatic changes in the absorption spectra of the unprotonated and protonated molecular composites indicate that they can be used as optical pH sensors over a wide pH range.The Nujol mull matrix of the molecular composite 1, containing the emeraldine base form, was soaked in solutions of diVerent pH values up to pH 7 (the host polymer decomposes in alkaline media). It was removed after 5 min, dried and then recorded, Fig. 3. Both bands at 385 and 555 nm undergo gradual shift to longer wavelengths (ca. 450 and 790 nm) as Fig. 3 The electronic absorption spectra of the molecular composite the pH is decreased. The plot of the maximum absorption 1 doped in solutions of diVerent pH values; 1=1.1, 2=2, 3=2.95, 4=4, 5=5.1, 6=6, 7=7 wavelengths as a function of pH gives a straight line within the pH range 1.1–5.1 while above this pH range a negative deviation was observed, Fig. 4. posites (s298=5.3×10-4–0.69 S cm-1). The semiconducting character of these molecular composites was supported by Conductivity of the molecular composites studying the variation of DC-electrical conductivity as a function of temperature for the molecular composites, Fig. 5. In all The molecular composites 1–8, 11 and 12 behave as good semiconductors having conductivities in the range cases, there are positive temperature coeYcients of electrical conductivity indicating semiconducting behaviour.The acti- 6.3×10-2–0.82×10-4 S cm-1 while 9 and 10 behave as weak semiconductors (s298=3.2×10-6 and 5.0×10-7 S cm-1, vation energies of the protonated molecular composites [DE ca. 0.21 (1H), 0.24 (5H) and 0.21, 0.18 eV (11H)] are lower respectively), Table 5.The conductivity increases by three orders of magnitude upon protonation of the molecular com- than the corresponding values of the nonprotonated ones 844 J. Mater. Chem., 1998, 8(4), 841–846Table 5 Electrical conductivities at 298 K of the protonated and deprotonated molecular composites compound s298 K/S cm-1 compound s298 K/S cm-1 compound s298 K/S cm-1 compound s298 K/S cm-1 1 2.2×10-4 7 6.3×10-2 1H 0.65 7H 0.19 2 3.6×10-2 8 2.8×10-2 2H 0.21 8H 0.15 3 5.8×10-2 9 3.2×10-6 3H 0.25 9H 2.8×10-3 4 2.1×10-4 10 5.0×10-7 4H 0.02 10H 5.3×10-4 5 2.1×10-2 11 3.1×10-4 5H 0.29 11H 0.69 6 8.2×10-5 12 4.2×10-2 6H 0.08 12H 0.30 [DE ca. 0.33 (1), 0.31 (5) and 0.34 eV (11)] indicating that protonation facilitates the mobility of charge, and hence the protonated molecular composites would exhibit higher electrical conductivity than the corresponding non-protonated species.Derivatives with ortho and meta substituents exhibit higher conductivities than aniline and the other derivatives owing to the fact that incorporation of electron-donating substituents at the ortho or meta positions decreases the oxidation potential of the monomer35,36 and hence the higher oxidation states are more facile to polymerize and consequently to be protonated by protons produced in the polymerization process.Also, the steric hindrance of the meta substituent or the presence of a substituent at the ortho position reduces sidecoupling, producing more regular chains. In the case of aniline, molecular composites 1 and 11, para coupling is not exclusive and radical cation coupling at the ortho position may lead to low yields of other products37 causing a decrease in the conductivity. However, the protonated molecular composites 1H and 11H exhibit the highest conductivity.On the other hand, the molecular composites containing the para-substituted derivatives, 4, 6 and 9 exhibit low conductivity due to the high possibility of radical cation coupling at the ortho position Fig. 4 The variation of the maximum absorption wavelength of the leading to a less planar conformation and consequently short long wavelength band of the molecular composite 1 as a function of pH conjugated chains.37 For 10, the presence of methyl groups makes the primary oxidation products rather less stable being more prone to polymerization and induces more deformation along the polymer backbone which results in a decrease of the degree of conjugation and hence a decrease in conductivity.Conclusion It has been demonstrated that novel conductive composites of 3D-coordination polymers and a variety of polyanilines can be prepared by chemical polymerization of polyanilines within the channels of the 3D-polymeric network.The improved conductivity of the deprotonated molecular composites is due to the presence of a dynamic equilibrium proton transfer causing partial protonation of emeraldine base and the formation of regular polyaniline chains, due to steric factors, into the channels of the polymeric network which prevent sidecoupling.The protonated molecular composites exhibit a dramatic increase in conductivity due to the formation of bipolaron lattice. The optical response of the molecular composites towards acid–base conditions make their use as optical pH sensors possible. The characteristics of the composites can be varied by changing the aniline derivatives as well as the host matrix. Therefore, the molecular composites can be tailored to meet specific properties. References 1 J.Yano, J. Electrochem. Soc., 1991, 138, 455. 2 X. Bi and Q. Pei, Synth.Met., 1987, 22, 145. 3 M. G. Kanatzidis, L. M. Tonge and T. J. Marks, J. Am. Chem. Soc., 1987, 109, 3797. 4 S. E. H. Etaiw and A. M. A. Ibrahim, J. Organomet. Chem., 1993, 456, 229. Fig. 5 log s vs. 1000/T for the molecular composites 1, 1H, 5, 5H, 11 5 P.Brandt, A. K. Brimah and R. D. Fischer, Angew. Chem., Int. Ed. Engl., 1988, 27, 1521. and 11H J. Mater. Chem., 1998, 8(4), 841–846 8456 U. Behrens, A. K. Brimah, T. M. Soliman, R. D. Fischer, 21 E. M. Ge`nies, P. Hany and C. Santier, J. Appl. Electrochem., 1988, 18, 751. D. C. Apperely, N. A. Davis and R. K. Harris, Orgonometallics, 22 H. Yoneyama, N. Takahashi and S.Kuwabota, J. Chem. Soc., 1992, 11, 1719. Chem. Commun., 1992, 716. 7 A. F. Diaz and J. A. Logan, J. Electroanal. Chem., 1989, 111, 111. 23 C. G. Wu and T. Bein, L ast Week’s Sci., 1994, 264, 1757. 8 N. Oyama, Y. Chnuki, K. Chiba and T. Ohsaka, Chem. L ett., 24 E. J. C. Chiang and A. G. MacDiarmid, Synth.Met., 1986, 13, 193. 1983, 1759. 25 W. R. Salaneck, I. Lundstrom, W.S. Huang and A. G. 9 K. Chiba, T. Ohsaka, Y. Ohnuk and N. Oyama, J. Electroanal. MacDiarmid, Synth.Met., 1987, 21, 121. Chem., 1987, 219, 117. 26 E. M. Ge`nies and E. Viet, Synth. Met., 1987, 20, 97. 10 N. Oyama, T. Ohsaka and M. Nakanishi, J. Macromol. Sci., Rev. 27 W. R. Salaneck, I. Lundstrom, A. G. MacDiarmid and W. S. Macromol. Chem., A, 1987, 24, 375. Huang, Synth.Met., 1986, 13, 294. 11 R. Noufý� , A. J. Nozik, T. White and L. F. Warren, J. Electrochem. 28 N. O. Lipari and C. B. Duke, J. Chem. Phys., 1975, 63, 1768. Soc., 1982, 129, 26. 29 C. B. Duke, Mol. Cryst. L iq. Cryst., 1979, 50, 63. 12 E. P. Lofton, J. W. Thackeray and M. S. Wrighton, J. Phys. Chem., 30 A. P. Monkman, D. Bloor, G. C. Stevens and J. C. H. Stevens, 1986, 90, 6080.J. Phys. D: Appl. Phys., 1987, 20, 1337. 13 S. Chao and M. S. Wrighton, J. Am. Chem. Soc., 1987, 109, 6627. 31 C. B. Duke, E. M. Conwell and A. Poton, Chem. Phys. L ett., 1986, 14 T. Kobayashi, H. Yoneyama and H. Tamura, J. Electroanal. Chem., 131, 82. 1984, 161, 419; 177, 281, 293. 32 A. J. Epstein and A. G. MacDiarmid, in Electronic Properties of 15 E. M. Genie`s, M. Lapkowski, C. Santier and E. Vieil, Synth. Met., Conjugated Polymers, ed. H. Kuzmany, M. Mehring and S. Roth, 1987, 18, 631. Springer, Berlin, 1989. 16 A. Watanabe, K. Mori, Y. Iwasaki, Y. Nakamura and S. Niizuma, 33 A. P. Monkman and P. Adams, Synth.Met., 1991, 40, 87. Macromolecules, 1987, 20, 1793. 34 S. Stafstro�m and J. L. Bredas, Synth.Met., 1989, 29, E219. 17 A. Akhtar, H. A. Weakliem, R. M. Paiste and K. Gaughen, Synth. 35 M. Leclerc, J. Guay and L. H. Dao, J. Electroanal. Chem., 1988, Met., 1988, 26, 203. 251, 21;Macromolecules, 1989, 22, 649. 18 T. Osaka, S. Ogano and K. Naoi, J. Electrochem. Soc., 1988, 135, 36 Y. Wei, W. W. Focke, G. E. Wnek, A. Ray and A. G. MacDiarmid, 539. J. Phys. Chem., 1989, 93, 495. 19 T. Osaka, S. Ogano, K. Naoi and N. Oyama, J. Electrochem. Soc., 37 L. T. Yu, M. S. Borredon, M. Jozefowicz, G. Belorgey and R. Buvet, J. Polym. Soc., 1987, 10, 2931. 1989, 136, 306. 20 T. Osaka, T. Nakajime, K. Naoi and B. Owens, J. Electrochem. Soc., 1990, 137, 2139. Paper 7/08827B; Received 8th December, 1997 846 J. Mater. Chem., 1998, 8(4), 841&ndash

ISSN:0959-9428    

DOI:10.1039/a708827b

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 Vitrigens Part 2.†—Low molecular weight organic systems with high glass transition temperatures I. Alig,a D. Braun,*a R. Langendorf,a H. O. Wirth,a M. Voigtb and J. H.Wendorffb aDeutsches KunststoV-Institut Schloßgartenstraße 6 D-64289 Darmstadt Germany bFachbereich Physikalische Chemie and Centre of Material Science Marburg Germany Low molecular weight organic materials characterized by a strong tendency towards glass formation and by glass transition temperatures around 100 °C are considered. They are constructed either as bulky odd-shaped molecules or as twin molecules where two bulky groups are linked via flexible or semi-flexible central groups. The synthesis of such materials their structure – glass formation relationship and their miscibility with each other and with polymer matrices are described.Low molecular weight organic compounds characterized by a The use of polymer based guest–host systems has several advantages. A major advantage is that thin films with good reduced tendency towards crystallization and by glass transition temperatures well above room temperature are the excep- dimensional and mechanical properties can easily be prepared by solution casting or spin-coating. A further advantage is tion rather than the rule.1,2 The incorporation of small organic units into oligomer or polymer chain backbones or the attach- that the systems may be optimized by choosing the host system from a large pool of commercially available and well ment of such groups to polymer chain backbones on the other hand leads to systems which can be prepared in the majority characterized polymers.Polymer based guest–host systems on the other hand have of cases as amorphous glasses. The glass transition temperature is known to increase in both cases with increasing chain length some serious disadvantages. (i) The miscibility of the chromophores is limited in the and the tendency towards crystallization is reduced either for steric or for kinetic reasons.3 This contribution is concerned majority of cases to a few percent or less. The major reason is that the entropy of mixing in such binary systems is strongly with low molecular weight organic systems which exhibit glass transition temperatures in the range characteristic of polymers reduced due to the presence of the long chain molecules.11 The long chain molecules on the other hand give rise to the high such as poly(methyl methacrylate) or polystyrene and which can be prepared in the fully noncrystalline state.glass transition temperatures required for most applications. It is obvious that low molecular weight matrix systems with Low molecular weight organic materials with high glass transition temperatures may be of interest both from a scientific high glass transition temperatures oVer advantages in terms of miscibility while keeping the glass transition temperature in and application point of view.4 With respect to fundamental science one wants to know the principal factors which control the required range. (ii) Polymer matrices have the tendency to age a process the position of the glass transition temperature. Such factors are among other things the molecular flexibility the presence which is accompanied by changes in the properties of the matrix.Chromophores dispersed in the polymer matrix are of steric hindrance and specific interactions such as the known to undergo relaxation processes which lead to a decay cohesion energy. of their function.5,6 It seems possible that low molecular weight Low molecular weight glass forming systems may find many vitrigens may be less sensitive in this respect. applications. It is well documented in the current literature The concept proposed in this paper is straightforward in its that guest–host systems in which a glassy matrix serves as a first stage it envisages the replacement of the polymer matrix host for functional molecules meet with a broad range of with a matrix composed of small molecules—which we will call possible applications.5–10 NLO-active systems for second harin the following vitrigens—with the specific requirement that monic generation (SHG) or for electro-optical modulation they have a high glass transition temperature of the order of (Pockels eVect) arise for instance if chromophores charac- 100 °C and above.In fact it should even be possible to construct terized by the absence of an inversion center are dissolved in chromophores in such a way that they themselves act as vitrigens. a polymer matrix which is subsequently poled in a strong A second requirement is of course that the guest–host electric field.5,6 Chromophores able to undergo a light-induced systems can be obtained as thin solid films by spin coating trans–cis–trans isomerization cycle have also been dispersed in for instance and that the films possess good mechanical glass forming matrices.7,8 Such guest–host systems can be used properties.The concept introduced above may thus have to for digital optical storage or holographic storage. A final be modified to take this requirement into account. The modifi- example is organic light emitting diodes (LEDs) composed of cation of the concept consists of replacing only a part of the chromophores dispersed in matrices having specific charge polymer component by the low molecular weight vitrigen. A carrier transport properties.9,10 The glassy matrix serves in all suYcient compatibility of chain molecules and low molecular cases to allow film formation and to reduce the tendency of weight glass formers has consequently to exist.The approaches the chromophores to crystallize. The glassy matrix may even taken by us to achieve this goal and the results obtained so provide special functions such as hole conduction as in the far are described here. case of LEDs. In the majority of cases the matrix is currently composed of polymers such as poly(methyl methacrylate) or polycarbonate. Selection of appropriate chemical structures Considerations based on incremental methods12 or on free volume concepts13 suggest that bulky molecular systems and † Part 1 J. Prakt. Chem. in press. J. Mater. Chem. 1998 8(4) 847–851 847 systems with strong attractive interactions should exhibit a tendency towards high glass transition temperatures. The requirement is on the other hand that the molecular system should have a weak tendency towards crystallization despite the strong intermolecular attraction.Molecules with irregular shapes and molecules able to exist in a large number of diVerent conformations in the fluid state at nearly iden- Synthesis tical energy should possess a weakened tendency towards crystallization mainly for kinetic reasons. 1H NMR spectra were obtained on a Bruker WM-300 instru- Based on such considerations we selected the following ment and chemical shifts are given in ppm downfield from model systems. Me4Si. Mass spectra were obtained on a Varian 311A instru- (i) Twin molecules (e.g. BECADHOX) where two identical ment in the field-desorption mode. DiVerential scanning calorbulky groups are linked via a central group (symmetric twins). imetry measurements were carried out on a Du Pont 912 or a The central links can be selected to be flexible or semi-flexible.Perkin-Elmer DSC 7 instrument. Samples of 5–10 mg in solid These types of molecules allow one to easily introduce a broad form were put in aluminium pans and heated at a scan rate of range of further modifications such as a variation of the length 10 or 20 °C min-1 under a nitrogen flow. The melting points flexibility and shape of the central link. were measured at the first heating. Indium metal was used as standard. After melting the samples were rapidly cooled to room temperature. The resulting glasses were heated under the same conditions again to measure the glass transition. Elemental analyses were performed using a Perkin-Elmer 240 elemental analyser. 1,4-Bis(pyren-1-ylmethylidene)aminomethylbenzene (PYDAPX) In a 500 ml three-necked round-bottomed flask equipped with a water separator (Dean–Stark) reflux condenser and a dropping funnel were placed 2.30 g (10 mmol) of pyrene-1-carbaldehyde dissolved in 100 ml of toluene.The reaction mixture was heated to reflux followed by addition of a solution of 0.68 g (5 mmol) of 1,4-bis(aminomethyl)benzene dissolved in 50 ml of toluene over a period of 1 h. Then the solution was refluxed for 3 h. After cooling to room temp. the precipitated crystals were filtered and the crude product was recrystallized from toluene and dried in vacuo. Yield 2.10 g (75%) mp 213 °C glass transition (Tg) 95°C. dH (300 MHz [2H8]THF) 5.04 (s 4H N-CH2-) 7.49 (s 4H Ar-H) 8.00–8.26 (m 14H Ar-H) 8.63–8.66 (d 2H Ar-H) 9.20–9.23 (d 2H Ar-H) 9.49 (s 2H C-H).Calc. for C42H28N2 (M 560.7) C 89.97; H 5.03; N 5.00. Found C 90.05; H 5.04; N 4.91%. MS (m/z) 561 (M+). 1,3-Bis(benzo[b]carbazol-9-ylmethyl )benzene (BCX) In a 100 ml three-necked round-bottomed flask equipped with a reflux condenser and a dropping funnel were placed 4.35 g (20 mmol) of benzo[b]carbazole (Ru� tgerswerke AG) in 30 ml N N N (CH2)5 N H N CH2 N H CH2 N N BECADHOX BECADHC5 PYDAPX BCX of dimethyl sulfoxide. Under nitrogen and with stirring 1.68 g (ii) A second approach was based on bulky odd-shaped (30 mmol) of potassium hydroxide were added. The suspension molecules which allowed the introduction of the functional was stirred about 1 h at 50 °C. Then the reaction mixture was groups i.e. chromophores directly into the glass forming cooled to room temp.and a solution of 1.75 g (10 mmol) of compounds. Fulvenes were considered in particular (e.g. TPCP a,a¾-dichloro-m-xylene dissolved in 20 ml of dimethyl sulfoxide derivatives). Details of the synthesis and the properties will be was added slowly. The solution was warmed to 90 °C for 1 h. given below. Then 200 ml of water were added and the crude product was separated. Then the crude product was dissolved in hot chloroform and dropped in an excess of methanol (1051 v/v). After separating the product was filtered and dried in vacuo. Yield 4.9 g (91%) mp 247 °C Tg 90 °C. Calc. for C40H28N2 (M 536.7) C 89.52; H 5.26; N 5.22. Found C 89.56; H 5.40; N 5.04%. MS (m/z) 536 (M+). 3,4-Dihydrobenzo[a]carbazole derivatives 1,5-Bis[(3,4-dihydrobenzo[a]carbazol-9-yl]pentane (BECADHC5).In a 100 ml three-necked round-bottomed flask equipped with a reflux condenser and a dropping funnel were placed 2.19 g (10 mmol) of 3,4-dihydrobenzo[a]carbazole (Ru� tgerswerke AG) dissolved in 30 ml of 1-methylpyrrolidin- 2-one under a stream of nitrogen. Then the mixture was warmed at 80 °C and 0.70 g (12.5 mmol) of potassium hydrox- 848 J. Mater. Chem. 1998 8(4) 847–851 ide was added; the colour changed from pale to yellow. After 4 h 1.15 g (5 mmol) of 1,5-dibromopentane dissolved in 20 ml of 1-methylpyrrolidin-2-one were added during 30 min. The reaction was stopped after 10 h. The solution was then filtered and poured in an excess of methanol (1051 v/v). The separated white crystals were filtered and dried in vacuo. Yield 1.77 g (70%) mp 149 °C Tg 40 °C. dH (300 MHz CDCl3) 1.37–1.45 (m 2H -CH2-) 1.88–1.98 (m 4H -CH2-) 2.85–2.98 (m 8H -CH2-) 4.32–4.37 (t 4H -CH2-N) 7.09–7.32 (m 12H Ar-H) 7.46–7.49 (d 2H Ar-H) 7.55–7.57 (d 2H Ar-H).Calc. for C37H34N2 (M 506.7) C 87.71; H 6.76; N 5.53. Found C 87.60; H 6.92; N 5.48%. MS (m/z) 506 (M+). 1,2-Bis[(3,4-dihydrobenzo[a]carbazol-9-yl )methyl]benzene (BECADHOX). This was synthesized analogously to BECADHC5 from 2.19 g (10 mmol) of 3,4-dihydrobenzo[a] Fig. 1 X-Ray diagram of PYDAPX as obtained from synthesis carbazole and 0.88 g (5 mmol) of a,a¾-dichloro-o-xylene. Yield 1.76 g (65%) mp 245 °C Tg 112 °C. dH (300 MHz CDCl3) 2.96 (m 8H -CH2-) 5.43 (s 4H CH2-N) 6.99–7.03 (m 2H Ar-H) 7.10–7.22 (m 14H Ar-H) 7.30–7.32 (m 2H Ar-H) 7.59–7.61 (d 2H Ar-H). Calc. for C40H32N2 (M 540.7) C 88.85; H 5.97; N 5.18. Found C 88.81; H 6.06; N 5.13%.Fulvenes 6-(4-Benzyloxyphenyl)-1,2,3,4-tetraphenylfulvene (TPCPBO). In a three-necked round-bottomed flask equipped with a reflux condenser and a dropping funnel were placed 1.85 g (5 mmol) of 1,2,3,4-tetraphenylcyclopenta-1,3-diene (Aldrich) and 3.18 g (15 mmol) of 4-benzyloxybenzaldehyde dissolved in 45 ml of absolute methanol. The reaction mixture was heated to reflux followed by addition of 90 ml of sodium methoxide (4 wt% in methanol) during 30 min. After refluxing for 18 h Fig. 2 DSC diagram of (a) the melting endotherm of PYDAPX and the mixture was cooled to room temperature. The precipitated (b) the melt quenched PYDAPX red crystals were filtered by suction and washed three times with methanol before drying in vacuo at 40 °C for 24 h. Yield 0.73 g (26%) mp 169 °C Tg 73 °C.dH (300 MHz CDCl3) 4.96 (s 2H -CH2-) 6.48–7.61 (m 30H Ar-H and C-H). Calc. for C43H32O (M 564.7) C 91.46; H 5.71. Found C 91.21; H 5.58%. MS (m/z) 564 (M+). 6-(4-Bromophenyl)-1,2,3,4-tetraphenylfulvene (TPCP-BR). This was synthesized analogously to TPCP-BO from 0.93 g (2.5 mmol) of 1,2,3,4-tetraphenylcyclopenta-1,3-diene 1.85 g (10 mmol) of 4-bromobenzaldehyde in 30 ml of absolute methanol and 45 ml (ca. 28 mmol) of sodium methoxide (4 wt% in methanol). Reaction time 6 h yield 1.26 g (94%) mp 187 °C Tg 84 °C. dH (300 MHz CDCl3) 6.76–7.33 (m 25H Ar-H and C-H). Calc. for C36H25Br (M 537.5) C 80.45; H 4.69. Found C 80.05; H 4.51%. MS (m/z) 536/538 (M+). Fig. 3 X-Ray diagram of the melt quenched PYDAPX 1,2,3,4-Tetraphenyl-6-(pyren-1-yl)fulvene (TPCP-PY ). This was synthesized analogously to TPCP-BO using 1.85 g (5 mmol) of 1,2,3,4-tetraphenylcyclopenta-1,3-diene 2.30 g Glass formation of the low molecular weight (10 mmol) of pyrene-1-carbaldehyde in 60 ml absolute meth- systems anol and 90 ml (ca.56 mmol) of sodium methoxide (4 wt% in The tendency towards glass formation was analysed using methanol). Reaction time 10 h yield 1.33 g (46%) mp 282 °C calorimetric and X-ray scattering investigations. The results Tg 118 °C. dH (300 MHz CDCl3) 6.35–8.22 (m 30H Ar-H and obtained will be discussed predominantly using the vitrigen C-H). Calc. for C46H30O (M 582.7) C 94.81; H 5.19. Found PYDAPX as an example. The powder obtained from the C 94.97; H 5.16%. MS (m/z) 582 (M+). synthesis was crystalline as is apparent from the X-ray diagram (Fig. 1).Experimental techniques The diagram is characterized by the absence of any trace of an amorphous halo the occurrence of which would indicate a X-Ray studies partially crystalline state. The reflections are narrow even for higher order reflections. This can be taken as an indication of The structure of the bulk solid state of the vitrigens and of the blends with polymers was analysed at room temperature in rather perfect and large crystals. Based on the location of the reflections we tentatively assign this scattering diagram to the glassy state employing a wide angle X-ray goniometer (Siemens D 5000) in the reflection mode. an orthorhombic unit cell with the cell parameters a=23.6 J. Mater. Chem. 1998 8(4) 847–851 849 Fig. 5 Variation of the glass transition temperatures of binary blends Fig.4 Melting points of binary mixtures of PYDAPX and BCX of PYDAPX and BCX with composition Table 1 Melting temperatures (Tm) glass transition temperatures (Tg) crystallization temperatures (Tc) and decomposition temperatures (Tz) substance Tg/°C Tc/°C Tm/°C Tz/°C PYDAPX 95 — 213 280 BCX 90 — 247 270 BECADHOX 112 208 245 300 BECADHC5 40 108 149 270 TPCP-BO 73 — 169 290 TPCP-BR 84 — 187 225 TPCP-PY 118 191 282 310 b=8.0 and c=3.8 A ° . The crystallographic density amounts to 1.3 g cm-3 which agrees approximately with the pyknometer result (1.35 g cm3). Fig. 2 displays in the the upper curve (1) the melting Fig. 6 Variation of the glass transition temperatures of binary blends endotherm of this compound. The melting peak is narrow and of (+) PYDAPX and Durel and ($) PYDAPX and PMMA with composition it is located at 216 °C; the enthalpy of fusion amounts to 65.1 kJ mol-1.Such a high value of the enthalpy of fusion is in agreement with the pnce of a perfect crystalline state of reported so far are that the compounds under investigation the compound as found by X-ray analysis. can be transferred into the glassy state using non-exotic (i.e. The melt may be quenched into the glass state at moderate moderate cooling) conditions and that the glass transition cooling rates of the order of 1–10 K min-1. The X-ray diagram temperatures approach the temperature range required for obtained from the quenched state (Fig. 3) is characterized by application in guest–host systems (.100 °C). This even holds a broad halo showing two maxima. Such a scattering diagram for chromophores such as the fulvenes studied here.is characteristic of a truly amorphous state the two maxima reflect the packing of the anisometric building blocks of the molecule. Glass formation of mixtures of the low molecular The DSC heating curve of the quenched state reveals (Fig. 2 weight systems lower curve) a stepwise increase of the specific heat which is of the order of 0.18 J g-1 K-1 which is in the range also It can be expected that the tendency towards crystallization and that the melting points of the systems are reduced in typically observed for polymer systems.14 The glass transition temperature occurs at 95 °C and is thus in the temperature binary mixtures. The experimental results agree with such expectations. The mixing of two vitrigens leads to a strong range required for the applications described above.The results described so far for the model system PYDAPX decrease of the melting temperature at intermediate concentrations. This is evident from the results of the calorimetric are characteristic of all the twin systems and the odd-shaped ones. Table 1 displays the melting temperatures (Tm) the glass analysis shown in Fig. 4 for the particular case of mixtures of the compounds PYDAPX and BCX. transition temperatures (Tg) the crystallization temperatures (Tc) and the decomposition temperatures (Tz). It is apparent The diagram evidently shows eutectic behavior as expected for the case of an immiscible crystalline state and a miscible that glass transition temperatures up to 118 °C were obtained. Of particular interest is the case of the fulvenes.These chromo- molten state. The melting temperature can be reduced to about 470 K in the eutectic mixture. We find only one glass transition phores form a glassy state at elevated temperatures so that they may be directly used to produce thin solid functional films. temperature of the melt-quenched material which varies very smoothly with the composition of the melt (Fig. 5). The conclusions which can be drawn from the results Table 2 Characteristic parameters of the polymers used for the blends polymer short name producer Mw Mn r/g cm-3 Tg/°C poly(ethyl acrylate) PEA Ro�hm 17000 1.12 -15 poly(methyl meth-acrylate) PMMA Polysciences Inc. 100000 53000 1.15 110 poly(vinyl acetate) PVAc Polysciences Inc. 500000 1.05 30 polyarylate PAR/Durel Hoechst 54500 27600 1.22 190 850 J.Mater. Chem. 1998 8(4) 847–851 magnitude of the stepwise increase of the specific heat connected with the glass transition of the two components are probably the origin.15 A similar observation was made for mixtures with PMMA (Fig. 6 lower curve). The closeness of the glass transition temperatures of PMMA and of the low molecular weight compound however makes it impossible to conclude that the glassy state is homogeneous. The optical clarity however points in this direction. Finally we consider the case of blends with PVAc (Fig. 7). We observe two well separated glass transition temperatures the two compounds are not miscible. The conclusion is thus that homogeneous mixtures of polymer and low molecular weight components with elevated glass transition temperatures can be achieved in certain cases whereas the entropic driving force towards mixing is not strong enough in other cases.Fig. 7 Variation of the glass transition temperatures of binary blends of PYDAPX and PVAc with composition Conclusions This is not surprising in view of the fact that the two glass transition temperatures are about equal. The diagram shown in Low molecular weight organic materials able to form the glassy state and characterized by glass transition temperatures Fig. 5 does thus not allow conclusions to be drawn on the state of mixing of the two components in the glassy state. Yet the well above room temperature—vitrigens—can be constructed on the basis of twin molecules and bulky irregularly-shaped eutectic behavior described above strongly suggests miscibility in the fluid and thus also in the glassy state.The observations molecules. The building blocks can be selected to be chromophores. Thin films with good optical properties can be prepared that the glassy films are transparent and reveal no inhomogeneity under the microscope point in the same direction. Similar by spin-coating either using mixtures with amorphous polymers or using the pure vitrigens. results were obtained for other binary blends of vitrigens. Glass formation of mixtures of the low molecular References weight systems with amorphous polymers 1 G. Tammann Z. Phys. Chem. 1898 25 441. Information on the miscibility of the low molecular weight 2 D. R. Uhlmann in Amorphous Materials ed. R. W. Douglas and B. Ellis Wiley Interscience London 1970 p. 205. compounds with the amorphous poly(methyl methacrylate) 3 H.A. Stuart Die Physik der Hochpolymeren vol. III Springer (PMMA) Durel (a polyarylate) and poly(vinyl acetate) (PVAc) Verlag Berlin 1955 p. 649. were obtained using calorimetry and X-ray diVraction. 4 W. Wedler D. Demus H. Zaschke K. Mohr W. Scha� fer and Characteristic parameters of the polymers are given in Table 2. W. Weissflog J.Mater. Sci. 1991 1 347. 5 D. J. Williams in Nonlinear Optical Properties of Organic Molecules and Crystals ed. D. S. Chemla and J. Zyss Academic Press New York 1987 vol. 1 p. 405. 6 D. J. Williams Angew. Chem. 1984 96 637. 7 G. Smets Adv. Polym. Sci. 1983 50 17. 8 M. Eich PhD thesis TH Darmstadt 1987. 9 J. A. Osakeni and S. A. Jenekhe,Macromolecules 1994 27 739. O C O C O O C CH3 CH3 n Durel 10 H. Vestweber J. Pommerehne R.Sandner R. F. Mahrt A. Greiner W. Heitz and H. Ba�ssler Synth.Met. 1995 68 263. First both the X-ray analysis and the calorimetric studies 11 P. J.Flory in Principles of Polymer Chemistry Cornell University reveal that the low molecular weight compounds dispersed in Press Ithaca NY 1983. the polymer do not crystallize on cooling. The variation of the 12 D. W. Van Krevelen Properties of Polymers Elsevier Amsterdam glass transition with composition indicates miscibility at all New York 1990. 13 F. Bueche Physical Properties of Polymers Interscience Publishers concentrations in the case of mixtures with Durel (Fig. 6 New York 1962 p. 85. upper curve). 14 B.Wunderlich and J. Grebowicz Adv. Polym. Sci. 1994 60/61 1. The temperature at which a stepwise increase of the specific 15 P. R. Couchman Macromolecules 1978 11 1156; 1980 13 1272. heat is observed in the blends varies continuously with the composition. The nonlinearity of the curve connecting the glass transition temperatures is not unusual. DiVerences in the Paper 7/06955C; Received 10thMarch 1997 J. Mater. Chem. 1998 8(4) 847–8

ISSN:0959-9428    

DOI:10.1039/a706955c

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 Theoretical calculations of sensitivity of deprotection reactions for acrylic polymers for 193 nm lithography Nobuyuki N. Matsuzawa,*† Takeshi Ohfuji,‡ Koichi Kuhara,§ Shigeyasu Mori,¶ Taku Morisawa,|| Masayuki Endo** and Masaru Sasago** Yokohama Research Center, Association of Super-Advanced Electronics T echnologies, 292 Yoshida-cho, T otsuka-ku, Yokohama 244, Japan The reaction energy of deprotection reactions, density of the reaction site, glass transition temperature, gas permeability, density and relative permittivity of photoresists of poly(TCDA5–RMA3–MAA2) and poly(TCDMACOOR4–TCDMACOOH6) with various protection groups were calculated.The most-enhanced exothermicity was calculated for protection groups containing an ethoxyethyl group as compared to the other protection groups: tetrahydropyranyl, tricyclodecanyl and tert-butyl.For the ethoxyethyl protection groups, a good correlation was found between the experimental sensitivity and the calculated values of the relative permittivity and the glass transition temperature of the polymers. This indicates that calculating these properties of polymers can provide a quick way to identify polymers having a high sensitivity for ArF lithography.The wavelength of light used in photolithography is getting that of the KrF laser used in the conventional lithography process. shorter in an attempt to realize smaller and smaller semiconductor devices. In line with this trend, development of photo- Possible factors which govern the dependence of sensitivity are: (1) heat of reaction of the deprotection reaction at the lithography at a wavelength of 193 nm using the ArF excimer laser will be used to make devices that will appear in the ester unit, (2) van der Waals volume of a segment of polymers, (3) density of polymers, (4) permeability of acids generated in beginning of the next century.1,2 For photoresist materials at this wavelength, a wide variety of acrylic polymers with alicyclic polymers, (5) glass transition temperature (Tg) and (6) relative permittivity (e) of polymers.We think that the first factor plays features are now being examined, because the conventional Novolac and/or polyvinylphenol polymers are not suYciently an important role if the sensitivity is dominated by the deprotection reaction itself. The other factors are important if transparent at a wavelength of 193 nm.3–17 Acrylic ester functionality is introduced to the polymer as a protecting group, the sensitivity is dominated by the diVusion of acid molecules.We note that the second factor represents the density of the and this ester unit decomposes to carboxylic acid in the presence of acid photochemically generated, thus exhibiting a reaction sites, i.e.the ester units, so that it should correspond to the distance the acid molecules have to diVuse. lithographic performance. It has been reported that characteristics of these photoresist materials, such as sensitivity (or dose) and dissolution rate, diVer depending on the protection group introduced to the Calculations photoresist polymer.14–17 However, no detailed theoretical/ Heat of reaction for the deprotection reaction at the ether unit molecular orbital studies of factors which control these characof acrylic polymers was calculated by applying the molecular teristics have been reported as far as they apply to ArF orbital theory both at a semiempirical and an ab initio level.lithography.We thus decided to carry out theoretical studies For the former, the MNDO Hamiltonian18 with the PM-3 of these characteristics. This would hopefully enable us to parameterization19,20 as implemented in the program predict these characteristics with reasonable accuracy and in MOPAC21 was applied for the calculations of geometry optimi- a shorter time as compared to performing actual experiments, zations and succeeding energy calculations of molecules.leading to an acceleration of the development of the new For the latter, local (LDFT) and nonlocal (NLDFT) density lithography process. The property that we chose as a starting functional theory22–24 was applied by using the program point for our theoretical studies is the dependence of sensitivity DGAUSS.25–28 The exchange-correlation potential derived by on the choice of protection group in acrylic polymers, because Vosko, Wilk and Nusair (VWN)29 was used for the LDFT for this property, a variety of experimental results are already calculations, whereas for the NLDFT calculations, the Becke available.10–17 It should be noted that the sensitivity of ArF exchange functional30–32 and the Lee–Yang–Parr correlation photoresists is one of the most important properties to be energy functional33 were applied.The geometries were fully improved, because, to reduce the damage to glass materials, optimized by applying analytic gradient methods.34–38 The the intensity of the ArF laser must be reduced as compared to basis set used for the calculations is of a valence doublef+ polarization functions quality, called DZVP,39 having a † Research Center, Sony Corporation, 174 Fujitsuka-cho, Hodogaya- form of (621/41/1) for the carbon and oxygen atoms and of ku, Yokohama 240, Japan.(41) for the hydrogen atom. The fitting basis set for the electron ‡ ULSI Device Development Laboratories, NEC Corporation, 1120 density and the exchange-correlation potential used in the Shimokuzawa, Sagamihara, Kanagawa 229–11, Japan.§ Microelectronics Research Center, Sanyo Electric Co., Ltd., 180 calculations was in the form of [7/3/3] for the carbon and Ohmori, Anpachi-cho, Anpachi-gun, Gifu 503–01, Japan. oxygen atoms and of [4] for the hydrogen atom. The applied ¶ Central Research Laboratory, Sharp Corporation, 2613–1 numerical grid for the integration was the ‘medium’’ grid in Ichinomoto-cho, Tenri, Nara 632 Japan.the program DGAUSS. Criteria for the SCF convergence and || Central Research Laboratory, Hitachi Ltd., 1–280 Higashigeometry optimization were those set by the ‘medium’’ options Koigakubo, Kokubunji, Tokyo 185, Japan. in the program. Thermodynamic correction terms to obtain ** ULSI Process Technology Center, Matsushita Electronics Corporation, 19 Nishikujo-Kasugano, Minami-ku, Kyoto 601, Japan. DGrxn in the gas-phase at 25 °C from DErxn were calculated J.Mater. Chem., 1998, 8(4), 853–858 853Table 1 Experimental sensitivities reported depending on the literature.40 All of the molecular orbital calculations were done using a Cray J916/12–4096 sensitivity (dose)/mJ cm-2 supercomputer.protection Calculations of the van der Waals volume of a segment of group Ref. 14 Ref. 14 Ref. 15 polymers, and the density, permeability, glass transition tempoly( TCDA5–RMA3–MAA2) perature and relative permittivity of the polymers were per- ETE 1.0a 1.4c formed by applying a graph theoretical treatment of molecular MEE 1.4a 1.6c properties,41–43 as implemented in the ‘Synthia’ module in the AEE 0.4a 0.8c program system POLYMER.44 All the calculations of this AdEE 10.0a graph theoretical treatment were performed on a COMTEC AdCEE 4.0a Solid Impact R-10000 work-station.THP 7.5a 3.0c poly(TCDMACOOR4–TCDMACOOH6) ETE 6.8b Results THP 8.2b Systematic experimental investigations on the sensitivity for aPAG: TPS (1 wt%); pre-baking: 80 °C and 60 s; PEB (post-exposure ArF lithography have been performed14–17 for several acrylic baking): 60 °C and 60 s for ETE, AEE, MEE, 70 °C and 60 s for THP, polymers with various protection groups. 100 °C and 60 s for AdEE and AdCEE; solution for development: TMAH, 0.0476 wt% in water. bPAG: SIT (1 wt%); pre-baking: 80 °C and 60 s; PEB: 70 °C and 60 s; solution for development: TMAH, 0.0476 wt% in water.cPAG: TPS (1 wt%); pre-baking: 80 °C and 60 s; PEB: 50 °C and 60 s; TMAH, 0.048 wt% in water. whereas a comparison of the values for diVerent protection groups for a fixed base polymer can be made. As shown in Table 1, two sets of experimental values are available for the terpolymer, although we note that the order of the sensitivity is the same for the two measurements.Semiempirical MO and DFT calculations For our semiempirical MO and DFT calculations, these copolymers are computationally too large for calculations to be done. CH2 CH CH2 C O O C CH2 C C C O R O O O H Me Me CH2 C C O O CH2 C C Me Me O O COOR COOH Poly(TCDMACOOR4-TCDMACOOH6) R = ETE, THP 60 40 Poly(TCDA5-RMA3-MAA2) R = ETE, MEE, AEE, THP, AdEE, AdCEE 20 30 50 CH Me OEt CH Me OC2H4OMe CH Me OC2H4OCOMe CH Me OC2H4O CH Me OC2H4OCO O THP AdCEE AdEE AEE MEE ETE= = = = = = For the terpolymer, poly(tricyclodecanyl acetate-co-methacrylate- co-methacrylic acid) [poly(TCDA5–RMA3–MAA2)], the sensitivity of the polymers with a protection group consisting of ethoxyethyl (ETE), methoxyethoxyethyl (MEE), acetoxyethoxyethyl (AEE), tetrahydropyranyl (THP), adamantyl- Me C Me Me O O R C O O R H C Me Me O O R R = ETE, THP R = ETE, AEE, MEE, AdEE, AdCEE, THP, TCD R = TCD 1 2 3 oxyethoxyethyl (AdEE), and adamantylcarbonyloxyethoxyethyl (AdCEE) groups has been reported with the use of Thus, model compounds 1–3 were chosen for the calculations of heat of reaction (DHrxn, or DGrxn).In order to model deprotection triphenylsulfonium triflate (TPS) as a photoacid generator (PAG).14,15 For the partially protected poly(carbonyl- reactions for poly(TCDA5–RMA3–MAA2), we chose the model molecules 1, where the polymer chain is terminated by methyl tricyclodecanyl methacrylate) [poly(TCDMACOOR4– TCDMACOOH6)], the sensitivity for ETE and THP with the groups. The model molecule 1 was also used for poly(TCDMACOOR4–TCDMACOOH6) for the calculations of use of hydroxysuccinimide tosylate (SIT) as a PAG has been measured.14,15 The experimental sensitivity of these polymers DGrxn (or DHrxn) of the deprotection reactions where the tricyclodecanyl group detaches from the polymer main chain.For the is summarized in Table 1.14,15 Experimental parameters for the poly(TCDA5–RMA3–MAA2) system are not exactly the same possible deprotection reactions of the tricyclodecanyl group detaching from poly(TCDA5–RMA3–MAA2), the model as those for the poly(TCDMACOOR4–TCDMA-COOH6) system (Table 1),14,15 so that a direct comparison between molecule 2 was used.For deprotection reactions for poly(TCDMACOOR4–TCDMACOOH6), where reaction the sensitivity for poly(TCDA5–RMA3–MAA2) and poly(TCDMACOOR4–TCDMACOOH6) cannot be made, occurs at the ester unit adjacent to the ETE or THP group, the 854 J.Mater. Chem., 1998, 8(4), 853–858Table 2 Calculated reaction energy of the hydrolysis reaction [eqn. For the hydrolysis reaction shown in Table 2, the (1)] in kcal mol-1 MNDO/PM-3 calculated order of magnitude of the reaction energy does not agree with that at the DFT levels. Furthermore, MNDO/PM-3 VWN/DZVP BLYP/DZVP the reaction energy for the But group at the MNDO/PM-3 DHrxn DGrxn a DGrxn a level is calculated to be slightly more exothermic than those R R¾ /kcal mol-1 /kcal mol-1 /kcal mol-1 for the groups containing an ethoxyethyl unit, which is contrary But ETE -11.9 -8.8 -8.8 to experimental results showing lower sensitivity for the But TCD ETE -10.4 -10.1 -11.2 group than for the THP group and the groups containing an But MEE -8.2 -7.9 -8.5 ethoxyethyl unit.14–17 The VWN/DZVP values are slightly less But AEE -11.4 -9.8 -10.3 exothermic than the BLYP/DZVP values.The BLYP/DZVP But AdEE -11.2 -7.4 -10.7 values for the model compounds 1 and 3 are essentially the But AdCEE -11.0 -9.9 -10.8 But THP -11.7 -3.7 -4.4 same (see the values for R¾=ETE and THP in Table 2).This TCD THP -9.4 -5.2 -5.7 shows that the reactivity in terms of the hydrolysis reaction But But -12.3 -2.8 -5.4 should be the same for poly(TCDA5–RMA3–MAA2) and But TCD -9.4 0.8 -1.3 poly(TCDMACOOR4–TCDMACOOH6), if the protection Pri TCD -9.3 -1.9 -3.2 groups present are the same. In the case of R¾=TCD, the BLYP/DZVP values for R=But and Pri are, again, essentially aDGrxn in the gas-phase at 25 °C. the same, showing that the reactivity in terms of the detachment Table 3 Calculated reaction energy of eqn.(2) in kcal mol-1 for model of the TCD group in poly(TCDA5–RMA3–MAA2) and polymer 1 poly(TCDMACOOR4–TCDMACOOH6) does not diVer significantly. MNDO/PM-3 VWN/DZVP BLYP/DZVP For this reaction, the BLYP/DZVP values of DGrxn are protecting DHrxn DGrxn a DGrxn a calculated to be exothermic, so the reaction is predicted to group /kcal mol-1 /kcal mol-1 /kcal mol-1 proceed thermally.The DGrxn values of the groups containing ETE 12.8 15.0 -2.0 an ethoxyethyl group (R¾=ETE, MEE, AEE, AdEE and MEE 14.3 15.5 -0.7 AdCEE) are essentially the same with an exothermicity of AEE 13.7 14.3 -3.5 about 8–11 kcal mol-1 (1 cal=4.184 J) at the BLYP/DZVP AdEE 13.7 12.4 -4.9 level.The value for the THP group is calculated to be less AdCEE 14.5 14.8 -4.0 exothermic by 4–6 kcal mol-1 than that for the groups contain- THP 9.6 14.6 -2.4 But 3.5 16.2 -3.5 ing an ethoxyethyl group. The value for the But group is TCD 14.7 21.2 2.1 essentially the same as that for the THP group, and still less exothermic values are calculated for the TCD protection group, aDGrxn in the gas-phase at 25 °C.so that the general trend of the order of the exothermicity of the calculated values becomes TCD<But~THP<groups con- model molecule 3 was used. We note that because most atoms taining an ethoxyethyl group (ETE, MEE, AEE, AdEE and present in the polymers are sp3 hybridized, the diVerence in AdCEE). Experimentally, it is known that the TCD group is electronic structure between the polymers and model compounds hardly detached at all, whereas the detachment of the But is not expected to be significant, especially at the ester unit, so group can be observed, although the sensitivity for the But that the eVect of this modeling on the magnitude of calculated group is lower than that for the THP group and the groups DGrxn (or DHrxn) is expected to be negligible. containing an ethoxyethyl group.14–17 In addition, as shown DGrxn (or DHrxn) values for two reactions of possible relin Table 1, the THP group shows a lower sensitivity than that evance to the deprotection reaction in the photoresists were for the groups containing an ethoxyethyl group.Thus, the calculated. The calculated DGrxn (or DHrxn) values for hydrolycalculated trend in DErxn is, in general, in agreement with the sis at the ester group45 present in the polymers catalyzed by experimental trend in sensitivity, although it does not account an acid [eqn.(1)] are listed in Table 2. for the diVerence in the sensitivity for the groups containing R-COO-R¾+H2O�R-COOH+R¾OH (1) an ethoxyethyl group.DGrxn (or DHrxn) values calculated for the other reaction For DGrxn (or DHrxn) of eqn. (2) (Table 3),50 a similar result [eqn. (2)] are tabulated in Table 3. for the hydrolysis reaction is obtained; the MNDO/PM-3 calculated order of the magnitude of DHrxn does not agree with that at the DFT levels. The reaction energy for the But group at the MNDO/PM-3 level is calculated to be less endothermic than those for the groups containing an ethoxy- R C O O CH CHR¢R¢¢ R¢¢¢ R C O O H C CR¢R¢¢ R¢¢¢ H + (2) ethyl unit, which is, again, contrary to experimental results.14–17 This shows that the semiempirical method does not predict This is a pyrolysis reaction catalyzed by an acid,46,47 often the reaction energy with the necessary accuracy to allow a referred to as the reaction occurring in the photoresist.6,48,49 qualitative discussion.Thus, although the semiempirical MO As shown in eqn. (2), the reaction involves a breaking of a method has an advantage that it is computationally less OMC bond present in the ester group with formation of a expensive than ab initio methods, methods with no empirical double bond in the protection group.parameters such as the DFT and ab initio MO method must We note that when the deprotection group is THP or TCD, bepplied to the prediction of reactivity of deprotection groups eqn. (2) should be read as eqn. (3) or eqn. (4), respectively. in photoresists. The VWN/DZVP values for eqn. (2) are more endothermic by about 15–20 kcal mol-1 than the BLYP/DZVP values, although we note that the order of magnitude of DGrxn at the VWN/DZVP level is not diVerent from that at the BLYP/DZVP level. Values for the groups containing an ethoxyethyl group are again essentially the same.In addition, values for the THP and But groups are essentially the same as those for the groups containing an ethoxyethyl group. The value for the TCD group is calculated to be more endothermic than R C O O O R C O H O O R C O O R C O H O + + (3) (4) J.Mater. Chem., 1998, 8(4), 853–858 855that for the other groups containing an ethoxyethyl group, the diVusion coeYcient,51 which may account for the lack of correlation in our results, suggesting that the sensitivity may which is in agreement with the experimental trend in sensitivity that the TCD group exhibits a lower sensitivity than the other not simply be described by the product of the two values.For the glass transition temperature for groups.14–17 These results show that the calculated results do not account for the diVerence in the sensitivity for the protec- poly(TCDA5–RMA3–MAA2), there is a lowering of sensitivity (an increase in dose) with an increase in the glass transition tion groups calculated except for the TCD group. temperature, as established by the plot between the calculated and experimental results as shown in Fig. 1(d), although points Results of the graph theoretical treatment are somewhat scattered. For poly(TCDMACOOR4– Calculated values of the van der Waals volume, density, glass TCDMACOOH6), the experimental sensitivity for the ETE transition temperature, gas permeability, relative permittivity group is higher than that for the THP group (Table 1), and and density using the graph theoretical treatment are listed in the calculated glass transition temperature is higher for the Table 4.In Fig. 1, we plotted experimental sensitivity against THP group than that for the ETE group. This agrees with the the calculated properties for the values for poly(TCDA5– trend in Fig. 1(d) found for poly(TCDA5–RMA3–MAA2). RMA3–MAA2). For relative permittivity, there is a good correlation between We calculated the van der Waals volume for a segment the calculated value and the experimental sensitivity for composed of five TCDA units, three RMA units and two poly(TCDA5–RMA3–MAA2). It is shown in Fig. 1(e) that if MAA units for poly(TCDA5–RMA3–MAA2), whereas for the relative permittivity becomes larger, the sensitivity becomes poly(TCDMACOOR4–TCDMACOOH6), the calculated higher (or the dose decreases).For poly(TCDMACOOR4– van der Waals volume corresponds to four TCDMACOOR TCDMACOOH6), the relative permittivity for the ETE group units and six TCDMACOOH units. We note that this value is calculated to be larger than that for the THP group, again represents the density of the reaction site, with the larger showing that a larger relative permittivity leads to a higher volume corresponding to a lower density of the reaction site.sensitivity. For the protection groups containing an ethoxyethyl group (ETE, MEE, AEE, AdEE and AdCEE), there is a weak Discussion tendency for the sensitivity to be lowered (or the dose to be increased) with an increase in the van der Waals volume for For the deprotection reaction, the calculated DGrxn seems to poly(TCDA5–RMA3–MAA2) [Fig. 1(a)]. However, we note dominate the experimental sensitivity if the protection group that points for the AEE and AdCEE groups are very scattered. becomes less reactive [DGrxn of eqn. (1)>ca. -6 kcal mol-1, In Fig. 1(b), the plot for the density of the polymers for or DGrxn of eqn. (2)>~0 kcal mol-1]. This suggests that, for poly(TCDA5–RMA3–MAA2) is shown. The density of polythe TCD protection group, the magnitude of the sensitivity is mers was calculated because this value may represent the dominated by the deprotection reaction itself. However, the magnitude of the free-volume of polymers.However, Fig. 1(b) magnitude of the calculated DErxn does not account for the shows no correlation between the experimental sensitivity and diVerence in sensitivity for groups containing an ethoxyethyl the density, with the values for the AEE and AdEE groups group (ETE, MEE, AEE, AdEE and AdCEE groups). For being very scattered. Thus, it can be concluded that this value these groups, the reaction energy is calculated to be essentially is not related to the sensitivity.the same and to be the most exothermic among the groups We calculated the gas permeability for nitrogen gas as a calculated, suggesting that the rate-determining factor for these parameter which would hopefully represent the mobility of compounds is not in the reaction, but is in the diVusion acids in the photoresist and hence dominate the magnitude of processes.This is supported by the presence of the correlation the sensitivity. We also calculated oxygen gas permeability, between the experimental sensitivity and the calculated properand found that the order was similar to that for nitrogen gas. ties of the van der Waals volume, the relative permittivity and As shown in Fig. 1(c), there is, again, no correlation between the glass transition temperature, although we note that the the calculated value and the experimental sensitivity for correlation with the van der Waals volume was the most poly(TCDA5–RMA3–MAA2). It can be concluded that this scattered among the three. quantity does not have a significant bearing on the sensitivity, For the correlation with the relative permittivity, a higher although we note that the mean error for the calculation of value for the polymer of this quantity is expected to cause a this value is reported to be significantly large (~50%; standard decrease in the value of the pH of the acid molecule, leading deviation).41–43 Gas permeability is reported to be proportional to an enhanced quantity of protons (or oxonium ions) formed to the product of the solubility of the gas in the medium and by the dissociation of the acid molecule.Consequently, the size of chemical species to diVuse becomes smaller, leading to an enhanced degree of diVusion. In addition, transport proper- Table 4 Calculated glass transition temperature, gas permeability, relative permittivity and density of the polymers ties of molecules are known to be related to the solubility of the molecules in the medium.51 In a previous study on alkali van der N2 gas metal cation transport across polymer-supported liquid memprotection Waals volume density Tg permeability relative branes,52 it was reported that the experimental Na+ flux is group /cm3 mol-1 /g cm-3 /°C (Dow unit)a permittivity correlated to the relative permittivity of the liquid membrane poly(TCDA5–RMA3–MAA2) solvent. In that study, the Na+ flux increases with an increase ETE 9.8×102 1.134 100 6.1×102 2.72 in the relative permittivity, which is consistent with our results.MEE 10.1×102 1.146 95 6.4×102 2.76 It has been reported that the dissolution rate of unexposed AEE 10.5×102 1.166 96 4.8×102 2.82 poly(TCDA5–RMA3–MAA2) where R=ETE, MEE, AEE, and AdEE 12.2×102 1.157 111 5.7×102 2.61 THP depends on er of the protecting groups, with the more AdCEE 12.6×102 1.180 112 4.2×102 2.67 polar protecting groups exhibiting a higher experimental dis- THP 9.5×102 1.213 119 3.2×102 2.66 But 9.6×102 1.119 108 6.9×102 2.67 solution rate.14 This means that the presence of protecting groups with low polarity leads to a larger inhibition eVect, if poly-(TCDMACOOR4–TCDMACOOH6) we consider the diVusion of the polar solvent TMAH (tetra- ETE 16.6×102 1.178 124 3.5×102 2.80 methylammonium hydroxide) used for the development.14 A THP 16.7×102 1.199 138 3.0×102 2.78 similar situation may hold true for acids diVusing in polymers, leading to a higher sensitivity in more polar media.We note a1 Dow unit=cm3 mil/(day×100 inches2×atm), where 1 mil= 0.001 inches. here that for the dissolution rate, the correlation was obtained 856 J. Mater. Chem., 1998, 8(4), 853–858Fig. 1 Plot of the logarithm of the experimental sensitivity against (a) van der Waals volume, (b) density, (c) nitrogen gas permeability, (d) calculated glass transition temperature and (e) relative permittivity of the polymer for the relative permittivity of the protection groups, whereas THP was diVerent from that of the protection groups containing an ethoxyethyl group.This means that when the density in our study, the correlation for the sensitivity is for the relative permittivity of the polymers. This may indicate that a local of the reaction site decreases, the sensitivity becomes lower.However, the correlation for this case is much more scattered relative permittivity plays an important role in the diVusion of the TMAH solvent, whereas for acids, the average relative than the other cases which are shown in Figs. 1(d) and (e), so that the density of the reaction site may not be the dominant permittivity of the medium becomes important.Another correlation found is that of the experimental sensi- factor in controlling the sensitivity. Thus, although it is not clear which value is the dominant tivity to the calculated glass transition temperature. This correlation is slightly more scattered than that for the relative factor in controlling the sensitivity, it can be concluded that the sensitivity is not dominated by the reaction alone, but is permittivity.It has been reported53,54 that the protonic or ionic transport in polymeric membranes is strongly aVected mainly dominated by the values related to the property of diVusion for the polymers with the protecting groups contain- by the glass transition temperature, with a higher glass transition temperature leading to a higher conductivity.This is ing an ethoxyethyl unit. For the THP and But groups, the calculated DGrxn for eqn. consistent with the correlation we obtained. We further note that the diVusion of molecules larger than a diatomic gas is (1) is less exothermic than that for the groups containing an ethoxyethyl group, which is consistent with the experimental known to be related to the polymer dynamics of segmental mobility55 which could be represented by the glass transition trend of a lower sensitivity of the THP and But groups than that of the groups containing an ethoxyethyl group.On the temperature for a set of polymers having the same or a similar main-chain structure. other hand, for the case of eqn. (2), the calculated DGrxn values for the THP and But groups are essentially the same as those Another possible explanation to account for the diVerence in sensitivity for the groups containing an ethoxyethyl unit for the groups containing an ethoxyethyl group.However, the calculated relative permittivity for the THP and But groups is can be extracted from Fig. 1(a); there is a weak tendency for the sensitivity to be lowered (the dose to be increased) with smaller, and the glass transition temperature is higher than that for the ETE, MEE and AEE groups containing an an increase in the van der Waals volume, if we ignore the point for THP in Fig. 1(a), because the calculated DGrxn of ethoxyethyl group, which is, again, inconsistent with the J. Mater. Chem., 1998, 8(4), 853–858 85714 S. Iwasa, K.Maeda, K. Nakano, T. Ohfuji and E. Hasegawa, experimental trend. Thus, for the THP and But protecting J. Photopolym. Sci. T echnol., 1996, 9, 447. groups, our calculated results do not clarify the dominant 15 T. Ohfuji, K. Maeda, K. Nakano and E. Hasegawa, Proc. SPIE, factor in controlling the sensitivity, suggesting that a balance 1996, 2724, 386. between the reaction and the diVusion process plays an import- 16 R.D. Allen, R. Sooriyakumaran, J. Opitz, G. M. WallraV, ant role for these groups. We further note that the calculated G. Breyta, R. A. DiPietro, D. C. Hofer, R. R. Kunz, U. Okoroanyanwu and C. G. Willson, J. Photopolym. Sci. T echnol., density for the THP and But groups does not account for their 1996, 9, 465. lower sensitivity than the ETE, MEE and AEE groups, because 17 T.Naito, K. Asakawa, N. Shida, T. Ushirogouchi and M. Nakase, the value for the THP and But groups is smaller than that for Jpn. J. Appl. Phys., 1994, 33, 7028. the ETE, MEE and AEE groups. 18 J. J. P. Stewart, J. Comput.-AidedMol. Design, 1990, 4, 1. 19 J. J. P. Stewart, J. Comput. Chem., 1989, 10, 209. 20 J. J. P. Stewart, J. Comput. Chem., 1989, 10, 221.Conclusion 21 J. J. P. Stewart, QCPE Program 455, 1983; version 6.00. 22 R. G. Parr and W. Yang, Density Functional T heory of Atoms and We found a correlation between the experimental sensitivity Molecules, Oxford University Press, New York, 1989. and the calculated values of the glass transition temperature 23 P. Hohenberg and W. Kohn, Phys. Rev. B, 1964, 136, 864. and the relative permittivity for the photoresists of 24 W.Kohn and L. J. Sham, Phys. Rev. A, 1965, 140, 1133. poly(TCDA5–RMA3–MAA2) and poly(TCDMACOOR4– 25 J. Andzelm and E. Wimmer, J. Chem. Phys., 1992, 96, 1280. 26 J. Andzelm, E. Wimmer and D. R. Salahub, in T he Challenge of d TCDMACOOH6) for protecting groups whose DGrxn of the and f Electrons: T heory and Computation, ACS Symp. Ser., No. 394, deprotection reaction (eqn. (2)] is more than about 0 ed. D. R. Salahub and M. C. Zerner, Am. Chem. Soc., Washington kcal mol-1 [or DGrxn of eqn. (1)> ca. -6 kcal mol-1]. If the DC, 1989, p. 228. reaction is more endothermic than this value, the experimental 27 J. Andzelm, in Density Functional Methods in Chemistry, ed. sensitivity is found to be controlled by the magnitude of DGrxn.J. Labanowski and J. Andzelm, Springer, New York, 1991, p. 155. 28 DGAUSS is a density functional program which is part of the These results show that the balance between the two processes UNICHEM software package and is available from Oxford of reaction and diVusion dominates the dependence of the Molecular Group, Beaverton, OR, USA. experimental sensitivity on the choice of the protecting group. 29 S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58, 1200. It is further suggested that calculating the relative permit- 30 A. D. Becke, J. Chem. Phys., 1986, 84, 4524. tivity and glass transition temperature of the polymers, in 31 A. D. Becke, Phys. Rev. A, 1988, 38, 3098. addition to calculating DGrxn of the deprotection reactions, 32 A. D. Becke, J.Chem. Phys., 1988, 88, 2547. 33 C. Lee, R. G. Parr and W. Yang, Phys. Rev. B, 1988, 37, 785. has the potential to provide a quick way to identify polymers 34 H. F. King and A. Komornicki, J. Chem. Phys., 1986, 84, 5645. having a high sensitivity for ArF lithography. 35 H. F. King, R. N. Camp and J. W. McIver, Jr., J. Chem. Phys., 1984, 80, 1171. This work was performed under the management of 36 P.M. W. Gill, B. G. Johnson, J. A. Pople and M. J. Frisch, Chem. Association of Super-Advanced Electronics Technologies Phys. L ett., 1992, 197, 499. 37 B. G. Johnson, P. M. W. Gill and J. A. Pople, J. Chem. Phys., 1993, (ASET) in the Ministry of International Trade and Industry 98, 5612. (MITI) Program of Super-Advanced Electronics Technologies 38 C. W. Murray, N.C. Handy and G. J. Laming, Mol. Phys., 1993, supported by New Energy and Industrial Technology 78, 997. Development Organization (NEDO). The authors would like 39 N. Godbout, D. R. Salahub, J. Andzelm and E. Wimmer, Can. to acknowledge Dr A. Ishitani and Mr H. Setoya for useful J. Chem., 1992, 70, 560. discussions. 40 D. A. McQusrrie, Statistical Mechanics, Harper and Row, New York, 1976. 41 J. Bicerano, Prediction of Polymer Properties, Marcel Dekker, New York, 1993. References 42 L. B. Kier and L. H. Hall, Molecular Connectivity in Chemistry and 1 T. Ogawa, J. Photopolym. Sci. T echnol., 1996, 9, 379. Drug Research, Academic Press, New York, 1976. 2 D. C. Hofer, G. Allen, G. WallraV, H. Ito, G. Breyta, P. Brock, 43 L. B. Kier and L. H. Hall, Molecular Connectivity in Structure– R.DiPietro and W. Conley, J. Photopolym. Sci. T echnol., 1996, Activity Analysis, Wiley, New York, 1986. 9, 387. 44 POLYMER is available from Molecular Simulations Inc., San 3 S. Takechi, Y. Kaimoto, K. Nozaki and N. Abe, J. Photopolym. Sci. Diego, CA. T echnol., 1992, 5, 439. 45 E. Fischer, Ber., 1909, 42, 1224. 4 Y. Kaimoto, K. Nozaki and S. Takechi, Proc.SPIE, 1992, 1692, 66. 46 L. H. Klemm, E. P. Antoniades and C. D. Lind, J. Org. Chem., 5 R. D. Allen, G. M. WallraV, R. A. DiPietro, D. C. Hofer and R. R. 1962, 27, 519. Kunz, J. Photopolym. Sci. T echnol., 1994, 7, 507. 47 W. P. Ratchford, Org. Synth., 1955, 3, 30. 6 K. Nakano, K, Maeda, S. Iwasa, J. Yano, Y, Ogura and 48 M. Rothschild, A. R. Forte, M. W. Horn and R. R. Kunz, IEEE E. Hasegawa, Proc. SPIE, 1994, 2195, 194. J. Selected T opics in Quantum Electronics, 1995, 1, 916. 7 T. Ushirogouchi, N. Kihara, S. Saito, T. Naito, K. Asakawa, 49 R. R. Kuntz, R. D. Allen, W. D. Hinsberg and G. M.WallraV, Proc. T. Tada and M. Nakase, Proc. SPIE, 1994, 2195, 205. SPIE, 1993, 1925, 167. 8 K. Nakano, K. Maeda, S. Iwasa, T. Ohfuji and E. Hasegawa, Proc. 50 For this reaction, DErxn for the model molecule 3 was not calcu- SPIE, 1995, 2438, 322. lated, because there is no significant diVerences in DErxn between 9 R. D. Allen, G. M. WallraV, R. A. DiPietro, D. C. Hofer and R. R. those for the model molecules 1 and 3 as found for the hydrolysis Kunz, Proc. SPIE, 1995, 2438, 474. reaction. 10 C. K. Ober and A. H. Gabor, J. Photopolym. Sci. T echnol., 1996, 51 H. A. Daynes, Proc. R. Soc. L ond., 1920, A97, 286. 52 P. R. Brown, J. L. Hallman, L. W. Whaley, D. H. Desai, M. J. Pugia 9, 1. and R. A. Bartsch, J.Membr. Sci., 1991, 56, 195. 11 U. Schaedeli, E. Tinguely, K. Cherubini, B. Maire, A. J. Blakeney, 53 R. A. Wallace, J. Appl. Phys., 1971, 42, 3121. P. Falcigno and R. R. Kunz, J. Photopolym. Sci. T echnol., 1996, 54 R. E. Barker, Jr. and C. R. Thomas, J. Appl. Phys., 1964, 35, 87. 9, 435. 55 For example, see reviews: (a) J. D. Ferry,Macromolecules, 1991, 24, 12 N. Shida, T. Ushirogouchi, K. Asakawa and M. Nakase, 5237; (b) E. D. von Meerwall, Adv. Polym. Sci., 1983, 54, 1. J. Photopolym. Sci. T echnol., 1996, 9, 457. 13 S. Takechi, M. Takahashi, A. Kotachi, K. Nozaki, E. Yano and I. Hanyu, J. Photopolym. Sci. T echnol., 1996, 9, 475. Paper 7/07225B; Received 6th October, 1997 858 J. Mater. Chem., 1998, 8(4), 853–858

ISSN:0959-9428    

DOI:10.1039/a707225b

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 Investigation into the structures of binary-, tertiary- and quinternarymixtures of n-alkanes and real diesel waxes using high-resolution synchrotron X-ray powder diVraction Steven R. Craig,a† Gerard P. Hastie,b Kevin J. Roberts,b* Andrea R. Gerson,a John N. Sherwooda and Robert D. Tackc aDepartment of Pure and Applied Chemistry, University of Strathclyde, Glasgow, UK G1 1XL bCentre forMolecular & Interface Engineering, Department ofMechanical and Chemical Engineering, Heriot-Watt University, Edinburgh, UK EH14 4AS cParamins, Exxon Chemical L td., Milton Hill, Abingdon, Oxfordshire, UK OX13 6BB High-resolution X-ray powder diVraction using synchrotron radiation has been used to determine the unit-cell parameters of binary (n-C20H42/n-C21H44 and n-C20H42/n-C22H46), tertiary (n-C19H40 to n-C21H44 and n-C20H42 to n-C22H46) and quinternary (with alkanes ranging from n-C18H38 to n-C26H54) n-alkane mixed homologous systems, together with a number of real diesel waxes.The structures have been found to conform to the orthorhombic structure with four molecules per unit cell and space group Fmmm predicted by Luth et al.8 for binary mixtures.Line profile analysis of the powder patterns revealed the existence of incipient end-chain and interchain disorder; a crystal packing model consisting of interchain mixing, end-chain twisting and chain migration is proposed to account for the observed disorder. For a broad spectrum of natural linear-chain molecular systems to C60H122 using high resolution synchrotron X-ray powthe occurrence of polydispersed aggregates is much more der diVraction.In this we found that, in the solid state, in common than the existence of pure compounds: for example, agreement with the broad conclusions of Luth et al.,8 the wax fractions from petroleum, linear polymers and biological n-alkanes (CnH2n+2) in general adopt three structurally lipids.For multicomponent n-alkane solid-state systems, from a distinct groups; triclinic [12<n(even)<26], orthorhombic comprehensive literature review published by Mnyukh,1 rules [n(even)36 and n(odd)] and monoclinic [28<n(even) 36]. were formulated to define conditions favourable for the existence These materials crystallise in the form of thin plates with of stable solid solutions.It was stated that the addition of a regular faces in which the chain direction is more or less solute molecule A to the solvent B should not increase the space perpendicular to the lamella surface. Prior to melting the group symmetry. Furthermore, the molecular volumes of the alkanes are known to exhibit solid–solid phase transitions to two species should nearly be the same.While these rules seemed crystalline ‘rotator’ phases.21 In the low temperature phase to hold for many paraYn solid solutions, there are examples (phase I) the chains adopt the overall trans configuration, after where the existence of an apparently stable solid solution contra- which the material may transform to a rotator phase (phase dicts the constraints placed on unit cell symmetry, in particular.II) in which the chains undergo hindered rotation about the A number of studies have been undertaken to investigate main carbon axis and the molecules are hexagonally packed. the ability of n-alkanes to form polydispersed crystalline aggre- Such a transformation is accompanied by an introduction of gates using calorimetry,2–4 powder X-ray diVraction tech- gauche bonds into an otherwise all trans structure.19 Employing niques,5–12 X-ray and surface tension measurements,13 electron line profile analysis, we also found evidence for a degree of diVraction,4,14–18 FTIR spectroscopy19 and atomic force structural disorder along the crystallographic c-axis for the microscopy.15 Studies of the binary mixtures of alkanes have longer homologues which, we suggested, implied that the shown that in the liquid state all chain lengths are fully process of end-chain bending/folding found in polyethylenes miscible, but in the solid phases the miscibility is strongly and in n-alkanes with a chain length greater than 102 carbon influenced by the carbon number diVerence and odd–even atoms22,23 was, to a lesser extent, incipient in these systems.carbon number eVects. An additional general feature of binary In this paper we present crystallographic high resolution mixtures is that they often assume untilted orthorhombic synchrotronX-ray diVraction data on a number of binary, tertiary structures when the pure components are tilted (likewise for and quinternary mixed homologue wax systems. The single hommulticomponent mixtures14).The studies have also shown ologues used to prepare the mixtures ranged from C18H38 to that, despite the diVerent chain lengths in the lamellae, the C26H54: these were selected to represent an average carbon length lamella surface is still ordered enough to allow for further present within ‘real’ diesel systems. Together with this, we also nucleation of subsequent layers.Models involving either trans- present data on a number of real diesel waxes containing a range lational or conformational disorder along the 001 plane have of average carbon chain lengths. In addition to refined unit cell evolved to try and account for this process: however, the exact parameters a number of benchmarking parameters were extracted mechanism for crystallisation is still not fully understood.from the preparation and X-ray data notably the impurity iso- In a previous paper,20 as a step towards understanding the paraYn content, amorphorous content, as well as correlation of crystallisation process of these multihomologous paraYn the deviation of c lattice parameter from ideality to: systems, we presented our investigations into the structures of some n-alkanes within the homologous series C13H28 $ the average carbon number for the n-alkane distribution $ the variance of the carbon number distribution $ the thermal disorder [as expressed via an isotropic (Wilson † now at Blacksmith (A Division of Champion Blacksmith), Minto Ave., Aberdeen AB1 4JZ, UK factor U2)] J.Mater. Chem., 1998, 8(4), 859–869 859$ the lattice strain to the planes as calculated from line with the same size reciprocal lattice and n is the number of peaks.A successful indexing is defined by the number of peaks profile analysis. indexed from the first 20 observed and requires the diVerence The overall strategy was a combinatorial one where qualitative between the observed and calculated positions to be <0.03°.and quantitative indicative parameters are combined in a self- Following indexing a final refinement of the unit-cell paramconsistent manner producing a solid-state structural model eters were made using the REFCEL program.26a which rationalises the behaviour of wax systems ranging from Previous studies12,28 suggested that the existence of conforsimple binary mixtures to complicated multihomologous real mational disorders, arising from the presence of chains with diesel wax systems.The overall aim of this is to characterise, diVerent lengths, is most evident along the crystallographic cand account for, the nature and stability of the solid-state axis direction. In this study we monitored the degree of structures formed. positional disorder in the (001) plane using line profile analysis of the 00l reflections to determine the lattice strain and the Experimental isotropic thermal parameter using Wilson’s method.29 Sample preparation Calculation of lattice strain.Determination of lattice strain The n-alkanes used in the preparation of the homologous eVects in the broadening of line profiles was carried out using mixtures were purchased from Aldrich: the purities were all a simplified method of integral breadths,30 a modification of stated as being >99%.Samples for X-ray analysis were the Williamson and Hall31 approach. In the strain analysis the prepared by recrystallisation from the melt or twice from n- peak shapes of the 00l reflections were fitted using a pseudododecane followed by removal of solvent by placing the sample Voight32 function with the PKFIT program.In the use of between absorbent paper and applying pressure for a period Voightian profiles this method is based on the assumption that of several days. Real diesel wax samples were supplied by the broadening profile due to strain eVects is Gaussian and Exxon Chemical Ltd. These samples had been crystallised size eVects Lorentzian.Furthermore, since all the particles from diesel and were not further prepared in any way. The under analysis were very large, then it was assumed that any composition of all samples were determined by gas chromato- strain dependent profile broadening resulted from crystalline graphic analysis on an HP 5890 gas chromatograph. molecular disorder and not grain size.An approximation of the lattice strains along the (001) plane Synchrotron X-ray powder diVraction of the multihomologous mixtures and real waxes studied was obtained using the relationship30 in eqn. (3), Powder diVraction patterns were collected using beamlines 8.3 and 2.324 at Daresbury synchrotron radiation source. The 1/(ds)=L-[4 sin h0/l(ds)]2e2L (3) instrument employed on these two beamlines was the identical where L is the volume-weighted average eVective crystallite two-circle diVractometer providing an incremental angular size along [hkl], e is the upper limit for lattice distortions precision of±0.0001° and an overall angle measuring accuracy ( lattice strain) and ds is the experimentally derived integral of 0.004°.The patterns were recorded using h/2h geometry breadth.with 2h ranging from 2 to 95° in steps of 10 millidegrees. The data were collected after a number of experimental sessions with wavelength use varying, session to session, within the Wilson’s method. In order to provide a benchmark parameter 1–2 A ° range. The wavelength and zero point corrections were for statistical disorder, an isotropic temperature factor was refined from the positions of the first five peaks observed for estimated from the powder data using Wilson’s method.29 In a silicon standard and this resulted in a wavelength accurate this, the fall-oV of the intensity of the harmonic series of well to 0.00003 A ° and a zero-point accuracy in 2h of 0.0001°.resolved 001 reflections was examined with respect to 2h.An DiVraction scans were recorded with a rotating sample in isotropic thermal parameter U was calculated using the order to average out the eVects of preferred sample orientation. relationship in eqn. (4), With the exception of diesel waxes 80441/92-18, 80079/87-11 ln Iobs=ln C-4p2U2Q (4) and 80308/90-112, whose measurements were carried out at low temperatures maintained using a variable-temperature where C is a constant.In this work, the calculated thermal cell,25 all measurements were carried out under ambient parameter was used as a benchmark to assess whether postemperatures. itional or conformational disorder is brought about by the size diVerences in chain length mismatches and the number of Data reduction and analysis varying chain lengths.Unit cell indexing. The data were normalised using the Amorphous content estimations. An examination of the PODSUM program.26a The peak shapes and resulting peak amorphous content, from the powder patterns of the systems positions were fitted using a pseudo-Voight function with the studied, was also carried out in order to generate a comparative PKFIT program.26a The unit cell was indexed from the peak model that could identify samples exhibiting poorer crystalline positions using the ITO program27 which is specifically order.The amorphous content is represented by a diVuse halo optimised for indexing powder patterns taken from materials at the bottom of a set of diVraction peaks, corresponding to having low-symmetry crystal structures. This program provides an average d-spacing of 4.5 A ° (see Fig. 1) which roughly a figure of merit (Fm) describing the accuracy of the fit to the corresponds to the inter-chain separation in n-alkanes. The proposed unit-cell lattice parameters. This is given by eqn. (1) method of analysis involved normalising sample diVraction patterns against the (111) diVraction peak using the PLOTEK Fm=. n l=1 (Qobs-Qcalc)/ .n l=1 (Q¾obs-Q¾calc) l=1 l=1 (1) program.26b Once the main diVraction peak had been normalised, the integrated area under all the broad scattering features where Q is the magnitude of reciprocal lattice vector for each which reflect scattering from the amorphous regions were peak as defined by eqn. (2), calculated. This was used as a qualitative benchmark expressed (2 sin h/l)2=1/d2=Q×10-4 (2) as a percentage of amorphous to crystalline fractional intensity for assessing relative amorphous content from sample to where d is the lattice plane spacing and Q¾obs and Q¾calc are, respectively, the observed and calculated values for the system.sample. Variation in (111) peak intensity, sample to sample, was not considered in this analysis. Qobs and Qcalc are the same values but for an arbitrary system 860 J.Mater. Chem., 1998, 8(4), 859–869Fig. 2 The central section of a powder diVraction pattern of a sample containing C20H42(18%)/C22H46(82%) which has been recrystallised twice from C12H26. This illustrates the possible profusion of peaks in powder diVraction patterns of mixed phase samples. Fig. 1 Plot of a typical wide angle scattering diVraction pattern for a the longer homologue in the sample increases.The systematic wax indicating the two diVerent regions: (a) sharp crystalline diVraction absences observed in the powder diVraction patterns for the profile and (b) broad amorphous halo diVraction profile samples analysed were (hkl): h+k, k+l, l+h=2n+1, which is consistent with the orthorhombic structure of space group DiVerence length calculations.The calculation of the diVer- Fmmm with four molecules per unit cell proposed by Luth ence length value is based upon information derived from the et al.8 for the high temperature (b0) phase of binary mixtures. X-ray diVraction data and that taken from gas chromato- In order for an orthorhombic unit cell containing a mixture graphic traces.Firstly, the cell edge calculated from the indexed of even and odd n-alkanes to conform to a Fmmm space group lattice parameters describes the actual cell edge length for a there must be at least two-fold disorder of each alkane molecule sample (average length of the c-axis backbone). Secondly, the so that each molecule then has an internal symmetry of mmm.average carbon number of the n-alkanes can be derived from Luth8 suggested that this disorder is brought about by motion the gas chromatograph by calculating the area under each of induced by heating of the molecules within the lattice structure, the retention peaks for the n-alkanes, and subsequently arriving that is, the disorder is dynamic. at a weighted average value of n-alkanes present.Using eqn. Three phases, each of which is orthorhombic, are exhibited (5) below, a theoretical c-axis value can be calculated using within the structure of the even–even binary mixtures when the newly derived carbon number average, whereby this should the percentage of the longer homologue exceeds 60%; such relate to the expected value. This can only be true if the wax behaviour is not evident in similar mixtures crystallised from structure was held within an ideal packing arrangement (i.e.the melt.8,34 The exact nature of the orthorhombic structures to that of a pure n-alkane). present in each phase diVer: it is suggested that the diVerent Using the end-group gap of 4.5 A° reported by Stokhuyzen,33 phases correspond to microcrystalline segments within the wax the average number of carbon atoms per molecule, n, was mixture induced into the high and low temperature orthorhomcalculated from eqn.(5). bic states. c/2=1.27(n-1)+4.5 (5) Solid solutions of tertiary and quinternary mixtures The diVerence between the observed and theoretical cell edge will indicate the extent of non-ideal packing that has formed X-Ray powder diVraction patterns were obtained for tertiary mixtures of C19H40 with C20H42 and C21H44, and C20H42 with for a particular wax system.C21H44 and C22H46, along with quinternary mixtures with alkanes ranging from C18H38 to C26H54. Gas chromatograhic Results and Discussion analysis indicated a relatively low impurity content within each of the systems studied (see Table 1, column 5).The refined Solid solutions of even–even and even–odd binary mixtures unit cell parameters determined from the data, together with In order to study the solid state structure of n-alkane mixtures the final mole ratios of the systems studied, are given in Table 1. of even–even and odd–even homologues crystallised from In all cases only a single structural phase is formed within dodecane, X-ray powder diVraction patterns were recorded of the solid solutions of the mixed alkanes.The lattice parameters mixtures of C20H42 with C21H44 or C22H46. Gas chromatorecorded for the mixtures were again found to be orthorhombic graphic analysis revealed the relative impurities for the even– structures of space group Fmmm, with unit cells similar to that odd mixtures comprised negligible neighbouring homologues, of the binary mixtures.The similarity between the powder whereas the even–even mixtures were found to contain a diVraction profiles also indicates that the structures are maximum of 9.5% dodecane. The resultant X-ray powder isostructural. diVraction patterns proved to be complicated due to the high density of peaks combined with, in some systems, the presence Crystal packing model for multihomologous wax systems of multiple phases within the crystal structures formed (see Fig. 2).Refined unit cell parameters determined from the data, While the structural characterisation has been clarified whereby all the wax mixtures are found to pack in an orthorhombic together with the final mole ratios of the systems studied, are given in Table 1.manner, there still exists a lack of fundamental knowledge as to the exact nature of the packing in relation to the composi- It is evident that all the binary mixtures crystallise to form orthorhombic lattices. The similarity between the a and b tion. Single odd n-alkanes crystallise to form an orthorhombic structure of space group Pbcm20 and, as has been shown, the parameters for all the systems indicates that the structures are isostructural; the c parameter increases as the percentage of multihomologous mixtures crystallise to form an orthorhombic J.Mater. Chem., 1998, 8(4), 859–869 861Table 1 Unit cell parameters, figure of merit (Fm) and molecular volumes (Vm) for the homologous mixtures of n-alkanes; plus the percentage impurity content present within the mixture (obtained from GC) no.of Cn mole mean (%) components range ratio Cn impurity a/A ° b/A ° c/A ° Fm Vm/A ° 3 2 20,21 69531 20.31 0.0 5.021 7.707 55.910 5 2164 (0.003 0.003 0.01) 64536 20.36 0.0 5.055 7.643 56.142 6 2168 (0.003 0.004 0.03) 47553 20.53 0.0 5.109 7.630 56.462 6 2200 (0.002 0.002 0.01) 40560 20.60 1.0 5.055 7.463 56.405 5 2127 (0.001 0.008 0.001) 26574 20.74 1.24 5.111 7.624 56.826 5 2213 (0.003 0.003 0.016) 15585 20.85 0.0 5.027 7.617 57.079 6 2186 (0.003 0.003 0.018) 9591 20.91 0.0 5.059 7.619 57.056 7 2199 (0.002 0.002 0.013) 5595 20.95 0.0 5.014 7.631 57.093 5 2183 (0.003 0.009 0.015) 1599 20.99 0.0 5.064 7.461 57.074 10 2156 (0.002 0.002 0.012) 2 20,22 67533 20.66 0.0 5.026 7.691 57.022 5 2204 (0.004 0.01 0.03) 64536 20.72 0.0 5.044 7.670 57.474 7 2224 (0.003 0.003 0.03) 40560 21.20 0.26 5.036 7.619 58.522 5 2244 (0.002 0.001 0.01) 28572 21.44 1.22 5.180 7.192 55.733 5 2076 (0.001 0.001 0.002) 5.603 7.622 58.664 5 2505 (0.002 0.001 0.010) 5.771 7.648 59.287 5 2555 (0.002 0.001 0.011) 18582 21.64 8.68 5.783 7.617 59.220 5 2608 (0.001 0.001 0.001) 5.663 7.478 58.207 5 2465 (0.002 0.001 0.011) 5.095 7.182 55.739 5 2040 (0.001 0.001 0.005) 14586 21.72 9.53 5.655 7.938 59.438 5 2668 (0.002 0.002 0.012) 5.604 7.618 58.287 5 2488 (0.004 0.003 0.048) 5.409 7.181 55.737 5 2165 (0.001 0.001 0.001) 3 19,20,21 658559 20.83 1.82 5.083 7.700 55.258 6 2164 (0.003 0.006 0.04) 10566524 20.14 1.26 5.027 7.694 55.534 28 2148 (0.0006 0.0008 0.01) 13544533 20.30 0.34 5.024 7.696 55.985 32 2164 (0.0006 0.0009 0.005) 19521560 20.41 0.0 5.025 7.688 56.307 25 2176 (0.0004 0.0005 0.005) 3 20,21,22 8538554 21.46 8.57 5.152 7.520 59.004 11 2269 (0.001 0.001 0.011) 12567521 21.09 0.0 5.000 7.487 57.917 4 2168 (0.001 0.001 0.003) 13537550 21.37 0.0 5.159 7.635 58.650 4 2310 (0.002 0.003 0.016) 0538562 21.62 0.0 5.145 7.671 59.140 10 2333 (0.002 0.001 0.010) 5 18–22 155512522560 21.35 0.0 5.025 7.656 58.771 8 2308 (0.002 0.004 0.05) 19–23 6512520530532 21.70 1.18 5.021 7.666 59.928 5 2260 (0.002 0.003 0.03) 20–24 3510521531535 22.85 0.35 5.250 7.143 62.952 6 2360 (0.002 0.003 0.017) 21–25 25952132536 23.91 0.58 5.115 7.334 65.825 5 2470 (0.002 0.002 0.014) 22–26 6513523529529 24.62 0.78 4.984 7.481 67.869 5 2530 (0.002 0.002 0.012) structure of space group Fmmm.A comparison of the molecular odd n-alkanes are more densely packed than for any of the homologous mixtures. This point is perhaps obvious if we volumes obtained for the crystalline homologous mixtures with those obtained for the single odd n-alkanes within the same consider that the single alkane structure has little or no disorder (i.e. it is ideally packed) whereas the Fmmm structure average carbon chain length range, that is, C19H40 to C25H52 (see ref. 20), reveals that the unit cells formed by the single does have inherent disordered packing constraints. This 862 J. Mater. Chem., 1998, 8(4), 859–869disorder will be most evident at the lamellae region where the analysis of the packing behaviour of the diVerent mixed systems.existence of chain length mismatches will aVect the lamellae Disorder arising from the chain length mismatches is one of packing arrangement. To determine the extent and nature of the factors thought to aVect the packing behaviour and stability the non-ideal packing along the lamellae (001) plane, a number of the solid solutions formed from the co-crystallisation of the of the mixed homologue systems were further analysed.diVerent alkane homologues. Fig. 3 and Fig. 4(a) indicate that Fig. 3 highlights the proportional relationship that exists as the average carbon chain length increases, and as the between the experimentally observed c-axis cell length for the number of components in the mixed wax systems increases, homologous mixture solid solutions formed and the average the diVerence length decreases.carbon number. The series of points do not correspond to a It has been suggested that the variance (measure of the chain truly proportional relationship, as the graph can be seen to length distributions) will control the stability of the wax curve away as the carbon number average increases. There produced. A plot of variance versus the diVerence length for also exists a degree of scatter and, although this relationship each of the wax systems [Fig. 4(b)] reveals that as the size of can be seen to be linear, a least-squares fit line emphasises the chain length mismatches (variance) increase the diVerence that there still exists a degree of diversity which as yet remains length decreases. While there does exist a large degree of unexplained through any physical or structural rationalisation. scatter in the plot the resulting pattern does perhaps suggest The graph also illustrates the extent of non-ideal packing that the ideal c-cell length is approached as the number of exhibited by these systems; the expected c axis lengths deterdi Verent homologues present within the solid solution mined theoretically using eqn.(5) are greater than the actual increases. A similar pattern is displayed for the Wilson param- values. This diVerence in length is tabulated in Table 2 and is eter (U2) [Fig. 4(c)] and the lattice strain along the (001) plane used as a benchmark to allow for a qualitative comparative [Fig. 4(d)]: as the number of homologues increase in a mixture there is an increase in incipient conformational disorder and lattice strain, and a decrease in the diVerence lengths.These results can be rationalised if we consider the diVerent packing arrangements that would arise at the inter-lamellae region as the size of chain length mismatches increases. The packing behaviour of a molecule in a single alkane (zero variance) crystalline system can be easily characterised as an alignment of the end chain methylene groups into a region or gap in between the head and tail of the stacked molecules.This produces a well defined lamellae region35 as observed in Fig. 5(a). Mixtures of n-alkanes with low variance (small chain length mismatches) will exhibit similar packing behaviour resulting in low disorder at the lamella interface (see Table 2). As the chain length mismatches increase to two or more carbon atoms the chain ends protruding into the interlamellae region will try and enter the thermodynamically more stable crystalline environment via void filling in adjacent lamellae, i.e.interchain mixing [Fig. 5(b)]. In this, the all-trans backbone Fig. 3 Plot of average carbon number (Cn) as a function of the c-axis structure of the alkane chain will be maintained resulting in parameter for ($) n-C20H42/n-C21H44, (&) n-C20H42/n-C22H46, (+) nvery little conformational disorder (reflected in a low Wilson C19H40/n-C20H42/n-C21H44, (,) n-C20H42/n-C21H44/n-C22H46 and (2) parameter); there will, however, be an increase in the disorder quinternary alkane mixtures recrystallised from C12H26.Solid line is of the lamellae region (seen as an increase in the overall the actual c-axis length, while the dashed line represents the expected c-axis length. percentage disorder—Table 2). Table 2 Calculated physical parameters for the homologous mixtures of n-alkanes co-crystallised from dodecane; diVerence length, variance (Vn), thermal parameter (U2), relative strain and relative percentage disorder no of Cn mean diVerence disorder components range Cn length/A ° Vh U2/A ° 2 rel.strain (%) 2 20,21 20.60 0.94 0.24 2.163 1.000 3.24 20.74 0.91 0.19 1.560 0.491 2.45 20.85 0.92 0.13 1.495 1.553 3.03 20.91 0.99 0.08 1.093 1.127 3.03 20.95 1.02 0.05 1.155 0.959 3.54 20.99 1.06 0.01 1.213 1.066 — 2 20,22 21.44 0.88 0.81 1.404 0.138 11.64 21.64 0.87 0.59 1.830 0.724 5.17 21.72 0.86 0.48 1.684 0.855 1.40 3 19,20,21 20.03 0.818 0.15 3.510 0.751 25.47 20.14 0.82 0.32 3.704 1.194 10.0 20.30 0.80 0.47 3.369 1.108 74.49 20.41 0.79 0.62 4.371 1.652 16.41 3 20,21,22 21.46 0.77 0.41 2.287 2.029 3.73 21.09 0.83 0.32 1.571 1.175 3.33 21.37 0.82 0.49 2.268 0.901 4.23 21.62 0.88 0.24 1.600 0.497 2.51 5 18–22 21.35 0.76 0.89 2.920 8.393 16.11 19–23 21.70 0.65 1.45 3.943 3.254 35.16 20–24 22.85 0.61 1.21 2.973 5.502 3.41 21–25 23.91 0.54 1.10 3.164 5.226 1.74 22–26 24.62 0.44 1.44 — — 1.88 J.Mater. Chem., 1998, 8(4), 859–869 863interlamellae region resulting from the presence of chain twists and folds trying to accommodate within the lamella packing would result in an increased interlamellae spacing and thus higher c-cell lengths than for the lower variance systems.The proposed model would be consistent with the results obtained: lamella packing in the low variance systems dominated by interchain mixing and the higher variance systems with chain folding and twisting. It is likely, however, that while each system may be dominated by one process over another, both processes will probably be in operation simultaneously in varying degrees in each system [Fig. 5(d)]. Previous studies employing FTIR spectroscopy on solid solutions of multihomologous mixtures9,19 have also shown the existence of conformational disorders in such systems. Solid solutions of real diesel wax systems In order to determine a crystal packing model for real diesel waxes X-ray powder diVraction patterns were recorded for a number of wax-crystallised diesel fuel extracts of varying carbon number range and variance.Gas chromatographic analysis revealed that the waxes displayed the typical Gaussian distribution of n-alkanes traditionally observed for fuel-crystallised systems, enabling the characterisation of the wax materials (see Fig. 6). In addition to the n-alkanes, a minor series of methyl-branched alkanes is also observed indicating the existence of an iso-paraYn impurity present within the fuel systems.The lattice parameters determined for the wax samples filtered from diesel fuels are given in Table 3. Most of the waxes were indexed using the program ITO. A number of the poorer crystalline materials possessed less than 20 diVraction peaks, which was below the minimum number of peak positions required to generate a lattice parameter, whereby a figure of Fig. 4 Plot of (a) carbon number average, (b) variance, (c) Wilson merit would not be obtained. In these cases an estimate was parameter (U2) and (d) lattice strain along the [00l] plane for ($) nmade as to the correct index for the remaining peaks (using C20H42/n-C21H44, (&) n-C20H42/n-C22H46, (+) n-C19H40/n-C20H42/nan indexed wax as a model comparison) and the unit cell C21H44, (,) n-C20H42/n-C21H44/n-C22H46 and (2) quinternary alkane mixtures recrystallised from C12H26.parameters were refined using REFCEL. Lattice parameters for the wax materials were, like in the multihomologous mixtures, all found to be orthorhombic with four molecules per unit cell.The c-axis parameter (Fig. 7) and molecular volume were both found to increase linearly with average carbon number. A two phase system was also found to appear in the real diesel waxes assumed to relate to regions with diVering packing arrangements, whereby the phase behaviour can be seen to correlate directly with the average carbon number of the wax. The higher melting point waxes (or higher Fig. 5 Schematic representation of interlamellae packing for (a) single alkane exhibiting a well ordered lamellae region, (b) low variance alkane mixture undergoing interchain mixing, (c) high variance alkane mixture exhibiting end-chain twisting and folding disorders and (d) combined interchain mixing and end-chain bending/twisting disorders at the lamella interface As the chain length mismatches increase to three or four carbon atoms diVerence, disorder results at the lamellae interface as alkane chain ends protrude into the interlamellae region.These chain ends, no longer held by the constraints of the adjacent van der Waals forces, would be able to twist and fold back into the crystal lattice resulting in conformational disorder (deviations from the all-trans structure) at the lamellae interface [Fig. 5(c)]. While such conformational disorders are thermodynamically less favourable than the all-trans structure they are necessarily introduced in order to enhance the stability of the solid solution formed. The re-entering chains would also Fig. 6 Carbon number distribution plot for a range of waxes crystalcause an increase in strain along the [001] plane as they push lised from real diesel obtained from a high temperature GC capillary parallel packed chains apart as they strive to re-enter the column; Cn=(2) 23.57, (&) 23.66, (+) 33.92, (x) 23.93, (6) 24.36, ($) 25.05 and (,) 26.80.Average carbon number range: 20.72Cn22.53. stability of the crystalline lattice. The increased disorder in the 864 J.Mater. Chem., 1998, 8(4), 859–869Table 3 Unit cell parameters, figure of merit (Fm), crystal phase (b—low temperature, b0—high temperature) and molecular volumes (Vm) for the real diesel waxes; plus the percentage iso-paraYn content present within the wax system (obtained from GC) mean wax type Cn range Cn iso-paraYn (%) a/A ° b/A ° c/A ° Fm Vm/A ° 3 T /°C phase 80734/84-24 15–27 19.13 — 4.849 8.061 54.734 1.35 2139 22.0 b0 (0.005 0.005 0.065) 80441/92-18 15–29 20.11 1.81 4.995 7.430 55.772 2.62 2070 -7.0 b (0.008 0.006 0.113) 80074/87-11 15–29 20.73 1.20 4.951 7.838 57.709 0.09 2239 22.0 b0 (0.001 0.001 0.005) 80079/87-11 12–25 20.83 1.04 4.997 7.421 57.580 0.14 2135 -7.7 b (0.001 0.001 0.007) 80653/90-14 14–29 21.55 1.62 4.991 7.738 60.085 0.44 2321 22.0 b0 (0.002 0.002 0.021) 80708/89-14 15–30 21.96 2.31 5.020 7.648 61.859 0.22 2375 22.0 b0 (0.001 0.001 0.011) 80002/88-13 16–28 22.06 1.40 5.025 7.629 60.873 0.46 2333 22.0 b0 (0.002 0.002 0.018) 80653/90-10 15–29 22.12 2.40 5.013 7.673 61.688 0.08 2373 22.0 b0 (0.001 0.001 0.004) 80167/88-19 15–31 22.18 1.77 4.981 7.753 61.538 1.05 2377 22.0 b0 (0.004 0.004 0.063) 80447/87-10 14–29 22.25 2.07 5.022 7.642 62.677 0.98 2405 22.0 b0 (0.004 0.004 0.051) TK197-10 12–26 22.27 — 5.030 7.625 62.602 1.06 2401 22.0 b0 (0.004 0.004 0.045) TK197-16 17–29 22.35 0.99 5.011 7.668 62.097 0.90 2386 22.0 b0 (0.003 0.003 0.054) 80308/90-11 17–28 22.54 3.10 5.024 7.639 62.450 1.11 2397 22.0 b0 (0.005 0.005 0.059) 80308/90-112 17–29 22.59 2.96 4.984 7.418 62.366 0.25 2301 -7.0 b (0.001 0.001 0.011) tk186-7a 17–29 22.61 3.15 5.013 7.675 62.465 0.28 2403 22.0 b0 (0.001 0.001 0.011) M 5.027 7.615 62.758 0.41 2402 22.0 b0 S (0.002 0.001 0.017) TK186-7b 15–25 22.71 2.95 Q 5.004 7.482 63.584 1.38 2381 22.0 b0 P (0.005 0.004 0.059) 80312/86-13 15–32 22.82 3.39 4.983 7.579 63.055 2.93 2381 22.0 b0 (0.011 0.010 0.138) 80226/89-16a 18–31 23.36 2.93 5.004 7.484 65.376 0.49 2448 22.0 b (0.002 0.002 0.028) 80226/86-16 14–32 23.46 3.62 5.021 7.624 63.892 0.58 2445 22.0 b0 (0.002 0.002 0.028) 80226/86 14–32 23.46 5.38 5.001 7.482 65.268 0.50 2442 b (0.002 0.002 0.023) M 5.015 7.645 64.531 0.44 2474 22.0 b0 S (0.002 0.001 0.019) 80152/86-16 18–27 23.50 4.65 Q 5.000 7.487 66502 0.33 2489 b P (0.001 0.001 0.011) 80291/87-10 18–29 23.57 2.78 5.001 7482 65.301 0.02 2443 22.0 b (0.001 0.001 0.001) 80226/89-21 16–30 23.66 2.50 5.000 7.482 66.309 2.12 2481 20.0 b (0.010 0.009 0.126) M 5.020 7.639 64.739 0.50 2483 22.0 b0 S (0.002 0.002 0.028) 80619/88-8 18–30 23.67 3.06 Q 5.002 7.491 65.620 0.85 2459 b P (0.004 0.003 0.049) 80312/86-8 15–32 23.67 3.35 5.000 7.842 66.126 0.29 2474 22.0 b (0.001 0.001 0.013) 80619/88-3 18–30 23.69 3.13 5.001 7.490 66.521 0.46 2492 22.0 b (0.002 0.002 0.027) 80226/86-16c 15–31 23.92 2.34 4.999 7.481 66.421 0.42 2484 22.0 b (0.002 0.002 0.025) 80442/92+5 18–32 23.95 3.92 4.999 7.487 66.111 0.44 2474 22.4 b (0.003 0.002 0.028) 80312/86-7 17–35 24.20 2.02 4.994 7.483 69.626 0.36 2602 22.0 b (0.001 0.001 0.018) 80226/89-11 11–31 24.36 1.87 4.993 7.483 67.962 0.41 2539 24.0 b (0.002 0.002 0.027) 89015/91-14 19–32 25.05 1.89 4.995 7.490 69.132 0.29 2586 22.0 b (0.001 0.001 0.015) Para32,720-4 18–32 25.60 2.36 4.989 7.480 75.920 0.52 2833 22.0 b (0.002 0.002 0.032) Shell Wax 20–37 26.80 5.98 4.989 7.471 74.147 0.82 2764 22.0 b (0.003 0.003 0.048) Para32,721-2 17–34 27.4 4.55 4.986 7.476 80.416 0.29 2998 22.0 b (0.001 0.001 0.020) Astor B Wax 21–43 32.06 — 4.980 7.455 90.921 0.18 3375 22.0 b (0.001 0.001 0.024) J.Mater. Chem., 1998, 8(4), 859–869 865Fig. 7 Plot of average carbon number (Cn) as a function of the c-axis parameter for a range of real diesel waxes. Solid line is the actual caxis length, while the dashed line represents the expected c-axis length. average carbon number) maintain their structural integrity at ambient temperatures whereby the lower melting point waxes, which are only partially crystalline, undergo a rotator transition resulting in a phase change.The two phases, identified previously by Luth et al.8 for binary mixtures, were identified as orthorhombic in nature. Crystal packing model for real diesel wax systems The crystallinity of a particular wax sample is known to be dependent upon the iso-paraYn content,36 thermal history,37 size of chain length mismatches and molecular mass.5,6 Comparison of the X-ray powder patterns obtained for a single alkane,20 the homologous mixtures and a diesel wax (Fig. 8) indicates a decrease in crystallinity as the number of compo- Fig. 8 Comparison of powder patterns obtained for (a) n-C20H42, (b) nents in the system increases.The increase in the fall-oV in n-C20H42(26%)/n-C21H44(74%), (c) n-C20H42(13%)/n-C21H44(37%)/nintensity of the (001) reflections as the range of n-alkanes co- C22H46(50%), (d) quinternary mixture (n-C20H42 to n-C24H50) and (e) crystallising increases, also indicates a growth in the magnitude real diesel wax (80002/83-13) [the inset illustrates the (00l) reflections on which line profile analysis was carried out] of the positional disordering along the crystallographic c-axis (carbon chain) as observed in the Wilson plots for the various systems (Fig. 9).As with the multihomologous mixtures, a qualitative comparative approach was taken to determine the nature and extent of the disordered packing for the diVerent wax systems of varying carbon number range and variance.Fig. 7 illustrates the extent the actual c-cell parameter lengths observed for the wax systems deviated from the ideal packing length. While the plot does show a degree of scatter, the diVerence lengths observed for the real wax systems are smaller than those observed for the homologous mixtures. Table 4 tabulates the experimentally derived physical properties for the wax materials.Plots of diVerence length versus mean carbon chain length, variance, disorder parameter (U2) and relative strain for the real wax systems (Fig. 10) reveal, Fig. 9 Wilson plots obtained for (&) n-C20H42(15%)/n-C21H44(85%), unlike for the homologous mixtures, no discernible pattern. (+) n-C20H42(8%)/n-C21H44(38%)/n-C22H46(54%), (,) five component mixture with alkane range n-C19H40 to n-C23H48 and (2) real Most of the waxes crystallised from the fuels do not have diesel wax (Shell ) illustrating the increase in the fall-oV in intensity of dramatically diVerent carbon number averages or variance the (00l) reflections with increasing number of components in the values.When comparing systems with such subtly diVerensolid- solution mixtures tiating physical properties, relative to the defect nature of the wax, it would be unlikely that any distinct trends would be observed.Comparing, however, the derived physical values for cate the existence of gauche conformers present within the solid solution of the real wax systems. It would be diYcult to the waxes with those obtained for the homologous mixtures, it would seem reasonable to suggest that the lamella packing assert this as definitive evidence for the existence of the end chain distortions solely, as the presence of iso-paraYns, ident- model proposed for the homologous mixtures, consisting of interchain mixing and end-chain deformation, could be applied ified from the GCs, would also result in gauche conformers and lead to higher disorder parameters.to the real wax systems. The small diVerence length observed for the waxes could also be indicative of an interlamellae The diVraction data recorded for the real diesel wax systems indicate the presence of the crystalline phase co-existing with packing dominated by the chain folding and twisting process, as suggested for the high variance multihomologous systems. a poorer ordered phase (amorphous type), observed as an amorphous halo at the baseline of the crystalline diVraction FTIR spectroscopic analysis19,38 of the real wax systems indi- 866 J.Mater. Chem., 1998, 8(4), 859–869Table 4 Calculated physical parameters for the real diesel wax systems; peaks. Fig. 11 depicts the amorphous peaks displayed in the diVerence length, variance (Vn), thermal parameter (U2), relative strain background diVraction data for the real diesel waxes.The and relative percentage disorder main characteristic peaks appear at about 20, 40 and 77° (2h) and correspond approximately to d-spacings of 4.5, 2.2 and C mean diVerence rel. rel. (%) 1.2 (±0.5) A ° . These values roughly correspond to interchain range Cn length/A ° Vn U2/A ° 2 strain disorder phase separations of primary, second and third harmonics corre- 15–27 19.13 0.12 3.96 6.420 1.792 100.0 b0 lations respectively in the crystalline n-alkanes fraction thus 15–29 20.11 0.70 5.50 7.508 1.672 4.91 b0 implying that disorder in the adjacent chain packing could 15–29 20.73 0.55 5.24 23.501 3.702 60.00 b0 account for the amorphous nature. 15–25 20.83 0.70 5.82 9.632 5.026 8.85 b While the disorder parameter (U2) and the percentage 15–29 21.55 0.44 4.32 22.909 1.821 75.96 b0 disorder (see Table 4) may be interrelated, the former refers to 15–30 21.96 0.15 4.34 20.715 4.579 56.54 b0 16–28 22.06 0.64 2.91 5.230 3.104 86.93 b0 conformational disorder along a particular axis (carbon chain) 15–29 22.12 0.38 4.84 21.913 3.895 52.19 b0 and denotes the eYciency of packing within the lamella region, 15–31 22.18 0.50 5.46 8.502 1.721 88.23 b0 while the latter is based predominantly on the peaks corre- 14–29 22.25 0.12 4.82 23.418 4.732 54.42 b0 sponding to the adjacent interchain packing.The inclusion of 17–28 22.54 0.50 3.78 19.262 12.395 52.69 b0 solvent, iso-paraYns and conformational distortions in the 17–29 22.59 0.58 3.83 6.591 1.471 5.07 b parallel chain packing and interlamellae packing will undoubt- 17–29 22.61 0.56 5.87 7.080 1.814 69.89 b0 15–25 22.71 K0.55 3.35 4.811 1.832 53.21 b0 edly result in some of the disorder identified in the diVraction L0.22 3.35 6.167 1.344 3.33 b patterns.The diVraction profiles relating to the adjacent chain 15–32 22.82 0.54 5.26 8.559 10.896 79.71 b0 packing do indicate a poorer crystallinity, as many peaks are 18–31 23.36 0.16 4.67 23.117 5.349 9.20 b inadequately resolved and have large peak halfwidths.On 18–27 23.50 0.64 4.57 5.691 4.229 13.75 b comparison, however, of the relative percentage disorder 18–29 23.57 0.40 3.88 17.407 1.694 5.96 b (amorphous content) with the lattice strain and Wilson param- 16–30 23.66 0.10 5.01 22.828 5.877 7.61 b 15–32 23.67 0.18 5.26 19.944 0.866 9.01 b eter for the wax crystals (see Table 4), we find that no corre- 18–30 23.69 0.04 3.90 6.693 3.688 6.67 b lation exists between these parameters.This indicates that the 15–31 23.92 0.31 4.51 20.874 1.219 6.11 b amorphous and crystalline regions are not mutually dependent, 18–32 23.93 0.45 4.39 5.207 13.331 3.73 b and that an increasing amorphous content does not adversely 11–32 24.36 0.15 5.38 5.606 3.919 4.03 b influence the ordered packing within the lattice. 19–32 25.05 0.38 4.22 5.561 5.889 3.27 b Comparing the relative percentage disorder values with the 20–37 26.80 0.15 7.97 9.444 4.756 9.61 b temperature at which the wax phases are held (Table 4), it is evident that a relationship does exist between these.The solid solutions held within the high temperature crystal phase (b0) exhibit a far greater amorphous (relative disorder) content than the lower temperature phase (b). It is known the temperature conditions5 (or thermal history) under which the wax crystals are kept can substantially aVect the packing of the nalkanes, which are known to be very mobile well below the melting point.If the sample temperature is raised at any point, Fig. 11 Comparison of the background amorphous profiles observed in the powder patterns of the real diesel waxes. The plot shows the Fig. 10 Plot of (a) carbon number average, (b) variance, (c) Wilson amorphous profiles obtained for waxes with mean carbon chain lengths (a) 0 18.3, (b) 21.3, (c) 21.5, (d) 21.6, (e) 22.0, ( f ) 22.7, (g) 22.8, (h) 23.2, parameter (U2) and (d) lattice strain along the [00l] plane for the real diesel waxes (I) 24.7 and ( j) 26.9.J. Mater. Chem., 1998, 8(4), 859–869 867this will induce molecular migration within the crystal altering In summary, the waxes crystallising in the higher temperature (b0) form pack with a degree of (rotational ) disorder. This the packing stability within; as found for binary mixtures with chain length diVerences greater than six carbon atoms.13,16 If provides less restriction on molecular migration than the more ordered (b) phase.Thus in the b0 phase segregation between the wax crystal is held above the first main rotator transition temperature, this will induce a favourable ordering rearrange- those chains of similar length (more crystalline fraction) and the residual components ( less ordered components) can be ment in the lamellae, reducing the overall disorder within this region.expected which perhaps explains the higher amorphous content observed in the samples crystallising in the b0 phase. This Based on these results and taking into account the external factors that can change or alter the packing, a broader crystal disorder can either be manifested via the residual molecular component aligned within the grain boundaries or within the packing model (Fig. 12)—based on the simplified models proposed for the homologous mixtures—is tentatively pro- amorphous regions characterised by poorly defined interlaminae regions [e.g. Fig. 12(c)]. posed for the real diesel waxes.The large range of carbon chain lengths present within the wax systems will lead to a large amount of disorder in the lamella region, thus disrupting Research towards improving our understanding of the structhe interlamellae packing and hindering further nucleation tural aspects of wax crystallisation has been supported for a along that plane. The system thus appears to compensate for number of years through research grants from the ESPRC and the chain length mismatches via interchain mixing [to lesser Exxon Chemical Ltd.We gratefully acknowledge Exxon (i) or greater degree (ii)] and end-chain twisting (as discussed Chemical Ltd. for the financial support of a studentship above) so maintaining the integrity of the lamella interface (S.R.C.), ESPRC for providing beam time on the Daresbury [Fig. 12(a)]. Such processes produce kinks (gauche conformers) SRS and for the financial support of a studentship (G.P.H.) along the carbon backbone as the lattice pushes apart to and senior fellowship (K.J.R.). accommodate the twisted back and intermixing chains. This overall process results in an increase in the interchain spacing References and an increased strain which is observed as a broadening of the (hk0) peaks.At higher temperatures (b0 phase) the molecu- 1 Y. V. Mnyukh, J. Phys. Chem. Solids, 1963, 24, 631. lar mobility of the chains allows for chain diVusion and 2 G. Zerbi, R. Piazza and K. Holland-Moritz, Polymer, 1982, 23, migration, as chains of similar chain length come together to 1921. 3 B. Flaherty, J. Appl. Chem. Biotechnol., 1971, 21, 144. form thermodynamically more favourable (less disordered) 4 D. L. Dorset and R. G. Snyder, Macromolecules, 1995, 28(24), lamella interfaces [Fig. 12(b)]. A looser packing arrangement 8412. could also be expected to display kinks within the carbon 5 K. E. Russel, B. K. Hunter and R. D. Heyding, Eur. Polym. J., backbone or exist as a separate phase of poorly ordered 1993, 29, 211.straight chains that are unable to cocrystallise due to the large 6 E. C. Reynhardt, J. Phys. D: Appl. Phys. 1986, 19, 1925. chain length mismatches [Fig. 12(c)]. 7 M. Dirand, Z. Achour, B. Jouti, A. Sabour and J. C. Gachon, Mol. Cryst. L iq. Cryst., 1996, A275, 293. 8 H. Luth, S. C. Nyburg, P. M. Robinson and H. G. Scott, Mol. Cryst.L iq. Cryst., 1972, 27, 337. 9 M. Maroncelli, H. L. Strauss and R. G. Snyder, J. Phys. Chem., 1985, 89, 5260. 10 G. I. Asbach and H. G. Kilian, Polymer, 1991, 32, 3006. 11 D. L. Dorset, J. Hanlon and G. Kavet, Macromolecules, 1989, 22, 2169. 12 I. Denicolo, A. F. Craievich and J. Doucet, J. Chem. Phys., 1984, 12, 80. 13 X. Z. Wu, B. M. Ocko, H. Tang, E. B. Sirota, S. K. Sinha and M.Deutsch, Phys. Rev. L ett., 1995, 75, 1332. 14 D. L. Dorset, Acta Crystallogr., Sect. B, 1995, 51, 1021. 15 D. L. Dorset and B. K. Annis, Macromolecules, 1996, 29, 2969. 16 D. L. Dorset and R. G. Snyder, J. Phys. Chem., 1996, 100, 9848. 17 D. L. Dorset, Macromolecules, 1985, 18, 2158. 18 D. L. Dorset, Macromolecules, 1987, 20, 2782. 19 G. P. Hastie and K. J. Roberts, J.Mater.Sci., 1994, 29, 1915. 20 S. R. Craig, G. P. Hastie, K. J. Roberts and J. N. Sherwood, J.Mater. Chem., 1994, 4, 997. 21 H. E. King Jr., E. B. Sirota, H. Shao and D. M. Singer, J. Phys. D: Appl. Phys., 1993, 26, B133. 22 See A. Keller in; Sir Charles Frank: An 80th Birthday T ribute, ed. R. Chalmers, J. E. Enderby, A. Keller, A. R. Lang and J. W. Steeds, Adam Hilger, Bristol, 1991. 23 G. Ungar, J. Stejny, A. Keller, I. Bidd and M. C. Whiting, Science, 1985, 229, 386. 24 R. J. Cernik, P. K. Murrey, P. Pattison and A. N. Fitch, J Appl. Crystallogr., 1990, 23, 292. 25 S. R. Craig, K. J. Roberts, J. N. Sherwood, K. Sato, M. Iwahashi and R. J. Cernick, J. Cryst. Growth, 1993, 128, 1263. 26 (a) Powder DiVraction Program Library, CLRC Daresbury Fig. 12 The packing groups thought to exist within diesel crystallised Laboratory; (b) PLOTEK, Data Acquisition Group; unpublished, CLRC Daresbury Laboratory, Warrington, WA4 4AD. wax. Region (a) represents the regions of high strain resulting from long n-alkane protrusions, (b) represents the end chain packing of a 27 J.W. Visser, J. Appl. Crystallogr., 1969, 2, 89. 28 J. J. Retief, D. W. Engel and E. G. Boonstra, J. Appl. Crystallogr., wax with both types of high and low lamellae disorder and (c) demonstrates the presence of amorphous regions possessing no lamel- 1985, 18, 156. 29 G. H. Stout and L. H. Jensen, X-ray Structure Determination, lae order with a minimal adjacent chain packing order. The grain boundary also contributes to the amorphous content of the material. Wiley-Interscience, New York, 2nd edn., 1989. 30 S. R. Craig, G. P. Hastie and K. J. Roberts, J. Mater. Sci. L ett., Although drawn in this manner there is no requirement to assume orthogonality of regions (a) and (b) versus (c). Regions (i ) and (ii) refer, 1996, 15, 1193. 31 G. K. Williamson and W. H. Hall, ActaMetallurgica, 1953, 1, 22. respectively, to areas involved in end-chain mixing to a lesser or greater degree. 32 R. A. Young and D. B. Wiles, J. Appl. Crystallogr., 1982, 15, 430. 868 J. Mater. Chem., 1998, 8(4), 859–86933 R. Stokhuyzen, An investigation of the structural problems in 36 R. M. Butler and D. M. MacLeod, Can. J. Cem. Eng., 1961, 39, 53. relation to some synthetic waxes, 1969, MSc thesis, Rhodes 37 J. J. Retief and J. H. le Roux, S. Afr. J. Sci., 1983, 79, 234. University, Grahamstown, South Africa. 38 S. R. Craig, PhD Thesis, Synchrotron Radiation Studies of the 34 A. R. Gerson, PhD Thesis, Structural and Kinetic Aspects of the Structure of n-Alkanes and Homologous Mixtures, University of Crystallisation of n-Alkanes, Homologous Mixtures and Real Strathclyde, Glasgow, 1995 p. 152. Waxes, University of Strathclyde, Glasgow, 1990. 35 G. Strobl, B. Ewen, E. W. Fischer and W. Piesczek, J. Chem. Phys., 1974, 61, 5257. Paper 7/06532I; Received 8th September, 1997 J. Mater. Chem., 1998, 8(4), 859–869 869

ISSN:0959-9428    

DOI:10.1039/a706532i

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 N-Acyl-b-D-glycopyranosylamines containing 1,4-disubstituted cyclohexyl and phenyl rings: mesomorphism and molecular structure relationships David F. Ewing,a Mandy Glew,a John W. Goodby,a Julie A. Haley,a Stephen M. Kelly,*a†‡ Bernd U. Komanschek,b Philippe Letellier,a Graham Mackenzie*a and Georg H. Mehla aT he School of Chemistry, Hull University, Hull, UK HU6 7RX bCCL RC Daresbury, Daresbury, UK A variety of N-acyl-b-D-glycopyranosylamines incorporating aliphatic and/or aromatic groups has been prepared regiospecifically and in good yield in a one step reaction of the required acid chloride with commercially available D-glycosylamines.The dependence of mesophase behaviour on the degree and nature of intermolecular hydrogen bonding, as well as the size and nature of the hydrophilic and hydrophobic parts of the core and the linking groups between the two have been studied.Of the compounds investigated only the N-(4-alkoxybenzoyl)-b-D-glucopyranosylamines and N-(trans-4-pentylcyclohexylacetyl)-b-Dglucopyranosylamine exhibited an observable thermotropic liquid crystal phase. The compounds synthesized were essentially insoluble in water and no lyotropic mesomorphism was observed.It is clear that it is the number of groups on the hydrophilic carbohydrate part of the molecule capable of hydrogen bonding, after a crucial length of the hydrophobic part of the molecule has been attained, that primarily determines thermotropic mesophase behaviour. However, it has also been found that an odd number of units in the central linkage between the hydrophilic and lipophilic parts of the molecule leads to significantly higher transition temperatures than those of compounds incorporating linkages with an even number of units.X-Ray diVraction studies imply that there is no intercalation of the hydrophobic parts of the molecule in the crystalline state. However, some intercalation is thought to occur after melting to form the bilayer structure of the smectic A* phase.Glycolipids with long aliphatic chains have been reported over nature of the linkage between the hydrophilic and hydrophobic parts of liquid crystalline carbohydrates has been studied to a the last 150 years to exhibit unusual melting behaviour or much lesser extent. Therefore, it was decided to synthesise a double melting points.1–4 It is now clear that these carbolimited number of derivatives of readily accessible carbo- hydrates were in fact liquid crystals exhibiting5–14 lamellar hydrates, incorporating the chemically and thermally stable smectic A* (SmA*),8,13,14 discotic or cubic15–17 mesophases.amide linkage at the anomeric position. This would allow Furthermore, glycolipids often possess amphotropic behavinvestigation of the eVect of introducing an additional site for iour,18 since they exhibit liquid crystalline properties both on hydrogen bonding next to the carbohydrate core.Initial studies melting the pure material to generate a thermotropic mesohave shown that this can lead to higher melting and clearing phase and also in the presence of solvents, e.g.with water to points than is observed for related compounds, such as esters produce lyotropic mesophases,19,20 which are also temperature or acetals.44–46 The amide linkage in carbohydrates is of dependent. especial interest since it seems to play an important role in The use of derivatives of naturally occurring monosaccharthe supramolecular structure of diverse, naturally occurring ides as solvents for non-denatured proteins,21,22 antibacterial glycoconjugates.47,48 and antiviral agents,23,25 surfactants,26 artificial blood,27 drug Some X-ray diVraction data are available for several of the delivery systems,28 optically active building blocks for chiral limited number of liquid crystalline carbohydrate derivatives nematic and ferroelectric liquid crystals,29,30 etc.has been reported with an amide linkage, but only for the crystalline reported. Furthermore, the structural and biological function state.49,53 Therefore, it was hoped that X-ray diVraction studies of monosaccharides, oligosaccharides and polysaccharides in on these new compounds in the crystalline and mesomorphic the organization and function of cell membranes is becoming state would yield information on the molecular structure in of increasing interest.31,32 It is becoming more apparent that both of these states.Two related models for the molecular the molecular factors determining the type of thermotropic organisation of the lamellar SmA* phase of liquid crystal and lyotropic mesophases observed for liquid crystalline carbocarbohydrates have been proposed.10–14 One model for the hydrate derivatives are the configuration of the hydrophilic bilayer lamellar structure of the smectic A* phase found for (carbohydrate) part of the molecule, the number and length of many carbohydrate derivatives suggests that the layers are the hydrophobic substituents, such as alkyl chains characterheld together by hydrogen bonding of the carbohydrate moiet- istic of liquid crystalline carbohydrate derivatives, and the ies along the median of the layers. The non-interdigitated degree and strength of hydrogen bonding with neighbouring aliphatic chains form the peripheral regions of the layers.10 An molecules. The factors that are important are becoming more alternative model proposes that the aliphatic chains are inter- apparent from evaluation of the large number of modified calated in the SmA* phase and located at the centre of the monosaccharides and oligosaccharides exhibiting thermotropic layers, whereas the carbohydrate moieties self-assemble in the and lyotropic liquid crystal properties synthesized recently, outer regions of the bilayers.11–14 Although X-ray diVraction especially in the last five years.33–43 studies primarily carried out on such liquid crystalline carbo- However, the dependence of mesomorphic behaviour on the hydrate derivatives in the crystalline state indicate that intercalation of the alkyl chains is indeed present,8,13 it is not necessary for there to be any correlation between the ordering in the †E-mail: s.m.kelly@chem.hull.ac.uk ‡EPSRC advanced fellow.crystalline state and in the mesomorphic state.11–14 J. Mater. Chem., 1998, 8(4), 871–880 871A limited number of previous investigations had shown that temperatures determined by optical microscopy. DiVerential scanning thermograms (scan rate 10 °C min-1) were obtained amphiphilic carbohydrate derivatives incorporating aromatic and aliphatic cores commonly found in non-amphiphilic liquid using a Perkin-Elmer DSC 7 PC system operating on DOS software.The results obtained were standardised with respect crystals could also exhibit mesomorphism at elevated temperatures. 41–44 Some substituted phenyl derivatives were found to to indium (measured onset 156.68 °C, DH 28.47 J g-1, literature value 156.60 °C, DH 28.45 J g-1), nitrotoluene (measured be more water-soluble than the corresponding carbohydrates with an aliphatic carbon chain in place of the aromatic ring.41 onset 51.17 °C, DH 118.49 J g-1, literature value 51.63 °C, DH 122.58 J g-1) and benzil (measured onset 94.42 °C, DH This has potential ramifications for technological applications based on lyotropic solutions of carbohydrate derivatives in 108.52 J g-1, literature value 94.87 °C, DH 92.68 J g-1).Comparison of the transition temperatures determined by water, such as those involving drug delivery, detergents and surfactants, as many thermotropic liquid crystalline sugar optical microscopy and diVerential scanning calorimetry shows some discrepancies of about 1–3 °C. Discrepancies may be due derivatives are only sparingly soluble in water at ambient temperatures.Therefore, it was also decided to extend these to two factors: firstly, the two methods use diVerent instruments which are calibrated in diVerent ways, and secondly, and more studies and synthesize a number of compounds related to the recently reported liquid crystalline, N-acyl-b-D-glycosyl- importantly, the carbohydrates tend to decompose at elevated temperatures at various rates depending on the rate of heating, amines,45,46 incorporating aromatic and aliphatic cores.Combinations of cyclohexyl and phenyl rings with terminal the time spent at an elevated temperature and the nature of the supporting substrate, e.g. the materials decomposed more and lateral substituents found in typical nonamphiphilic liquid crystals are incorporated in order to clarify whether or not quickly in aluminium DSC pans than on glass microscope slides.there is any correlation between the thermotropic mesomorphism of amphiphilic and nonamphiphilic liquid crystals incorporating identical moieties in the molecular core. DSynthesis of N-acyl-b-D-glycopyranosylamines from alicyclic Glucopyranosides and D-mannopyranosides were chosen with and aromatic carboxylic acids the appropriate structural diversity in order to study the relationship between molecular shape, chain length, lateral A direct acylation of commercially available D-glycosylamines using triethylamine as a base and N,N¾-dimethylformamide substituents and degree of aromaticity and liquid crystalline behaviour.Primarily, D-glucose derivatives were prepared in (DMF) as the solvent yielded the desired N-acyl-b-D-glycosylamines (1–17) in good yield and purity. The required acid order to allow comparisons with known systems to be made as most thermal data reported in the literature for liquid chlorides were either commercially available or were prepared from the corresponding acids by reaction with thionyl chloride.crystal transition temperatures refer to glucose derivatives. 11,12,36–43 Additionally, mannose, xylose and ribose deriva- Good selectivity was achieved without the necessity of using thiadiazole reagents.45 The reaction was stereospecific tives were prepared in order to investigate the dependence of the mesophase behaviour of this class of compound on sugar in all cases, giving only the b anomers.The N-(trans-4- pentylcyclohexylacetyl)-b-D-glucopyranosylamine 23 was pre- configuration. pared in a similar fashion utilising trans-4-penylcyclohexylacetyl chloride and D-glucosylamine. Experimental Techniques General procedure PCl5 (1 equiv.) was added in portions to a mixture of the The structures of the intermediate and final products were determined by 1H and 13C NMR spectroscopy (JEOL JNM- carboxylic acid (1.0 g, 1 equiv.) in THF–diethyl ether (20 cm3; 151; v/v) and the resulting mixture stirred until a clear solution GX 270 spectrometer), mass spectrometry (Finnigan-MAT 1020 GC–MS spectrometer) and infrared spectroscopy (Perkin- was obtained (ca. 3 h). The solvent was removed by evaporation under reduced pressure to yield the crude acid chloride, which Elmer 457 grating spectrophotometer). 1H chemical shifts were measured, in [2H6]DMSO, relative to (CH3)4Si and 13C was dissolved in DMF (20 cm3) and the resultant solution added dropwise to a solution of b-D-glucopyranosylamine chemical shifts relative to the solvent (d 39.5).J Values are in Hz. The 1H and 13C NMR data clearly showed the pyranose (1 equiv.) and DMF (20 cm3) at room temperature.The reaction mixture was stirred for 3 h, then the solvent was removed form for the derivatised monosaccharides and also indicated the stereochemistry associated with the anomeric centre. The by evaporation under reduced pressure (azeotroped with toluene). The crude product was taken up in n-butanol and the purity of each compound was determined by thin layer chromatography (TLC), high performance liquid chromatography resultant solution was washed with water (3×20 cm3).The organic phase was evaporated down and the resultant residue (HPLC), elemental analysis (C, H, N) and diVerential scanning calorimetry (DSC). 4×8 cm precoated TLC plates, SiO2 SIL purified by column chromatography on silica gel using chloroform –methanol (551; v/v) as eluent and then recrystallised G/UV254, layer thickness 0.25 mm (Machery-Nagel, Du� ren, Germany) were utilized.from ethanol. The white crystals so obtained were dried in vacuo over P2O5. Column chromatography was carried out using silica gel 60 (230–400 mesh ASTM). Reaction solvents and liquid reagents were purified by distillation or drying shortly before use.N-(trans-4-Pentylcyclohexylcarbonyl )-b-D-glucopyranosyl- Reactions were carried out under dry N2 unless water was amine 1 present as a reagent or a solvent. Mesophase identification and the transition temperatures of the carbohydrates syn- dH (DMSO) 0.86 (m, 5H, alkyl), 1.24 (m, 11H, alkyl), 1.74 (m, 4H, alkyl), 2.08 (t, 2H, J 7.5, alkyl), 2.98–3.21 (m, 4H, H-2, thesised were determined by optical microscopy using either Olympus BH-2 or a Zeiss Universal polarizing light microscope H-3, H-4, H-5), 3.40 (m, 1H, J5,6¾ 5, J6,6¾ 12, H-6¾), 3.61 (dd, 1H, J5,6 6, J6,6¾ 12, H-6), 4.45 (m, 1H, OH), 4.66 (t, 1H, J1,NH 9, in conjunction with a Mettler FP 82 microfurnace and FP 80 Central Processor.Homeotropic sample preparations suitable H-1), 4.76, 4.85, 5.00 (3×d, 3H, 3×OH), 8.19 (d, 1H, JNH,1 NHCO).dC (DMSO) 13.95, 22.12, 25.95, 28.81, 29.15, 31.61, for phase characterisation were prepared by using very clean glass microscope slides (washed with water, acetone, water, 32.08, 36.53, 36.84 (alkyl ), 60.92 (C-6), 70.01, 72,42, 77.58, 78.49 (C-2–C-5), 79.46 (C-1), 176.56 (CONH). Elemental analysis concentrated nitric acid, water and dry acetone).DiVerential scanning calorimetry was used to determine for C18H33NO6; Calc.: C, 60.14; H, 9.25; N, 3.90; Found: C, 60.54, H, 9.39, N, 3.83%. enthalpies of transition and to confirm the phase transition 872 J. Mater. Chem., 1998, 8(4), 871–880N-(trans-4-Pentylcyclohexylcarbonyl )-b-D-mannopyranosyl- m), 0.88 (CH3, 3H, t). dC (DMSO) 165.9 (CO), 161.1–113.6 (aromatic), 80.2 (C-1), 78.6 (C-5), 77.6 (C-3), 72.0 (C-2), 70.0 amine 2 (C-4), 67.6 (OCH2), 61.0 (C-6), 31.2–22.1 (aliphatic), 19.9 dH (DMSO) 7.79 (NH, J 8.15, 1H, d), 4.99–4.69 (4×OH, 4H, (CH3).Elemental analysis for C21H33NO7; Calc.: C, 61.30; m), 4.42 (H-1, J 1.07, 1H, m), 4.33–3.16 (H-2–H-6¾, 6H, m), H, 8.08; N, 3.40; Found: C, 61.01; H, 7.91; N, 3.29%. 2.15 (COCH, 1H, m), 1.71 (CH2, 2H, m), 1.27–1.04 (aliphatic [a]D22 +15.1 (c 0.0182 g cm-3; DMSO); Rf 0.13 chloroform– chain, 13H, m), 0.88 (CH3, 3H, t). dC (DMSO) 174.8 (CO), methanol (551). 78.9 (C-1), 77.3 (C-5), 73.9 (C-3), 70.8 (C-2), 66.7 (C-4), 61.3 (C-6), 43.7–18.5 (aliphatic), 13.9 (CH3). Elemental analysis for N-(4-Nonyloxybenzoyl )-b-D-glucopyranosylamine 8 C18H33NO6; Calc.: C, 60.14; H, 9.25; N, 3.90; Found: C, 59.98; H, 9.17; N, 3.83%.[a]D23 -1.6 (c 0.0146 g cm-3; DMSO); dH (DMSO) 8.66 (NH, J 9.98, 1H, d), 7.89–6.97 (aromatic, Rf 0.46 chloroform–methanol (251). Jortho 8.10, 4H, m), 5.11–4.90 (3×OH, 3H, m), 4.89 (H-1, J 8.16, 1H, m), 4.51 (OH, 1H, t), 4.02 (OCH2, 2H, t), 3.67–3.10 (H- 2–H-6¾, 6H, m), 1.54 (CH2, 2H, m), 1.35 (aliphatic, 12H, m), N-(trans-4-Pentylcyclohexylcarbonyl )-b-D-xylopyranosyl- 0.88 (CH3, 3H, t).dC (DMSO) 166.0 (CO), 161–126.1 (aro- amine 3 matic), 80.3 (C-1), 78.7 (C-5), 77.6 (C-3), 72.0 (C-2), 70.0 dH (DMSO) 8.17 (NH, J 8.45, 1H, d), 5.00–4.82 (3×OH, 3H, (C-4), 67.6 (OCH2), 61.0 (C-6), 31.3–22.1 (aliphatic), 13.9 m), 4.58 (H-1, J 8.31, 1H, m), 3.61–2.99 (H-2–H-5¾, 5H, m), (CH3). Elemental analysis for C22H35NO7; Calc.: C, 62.10; 2.05 (COCH, 1H, m), 1.71 (CH2, 4H, m), 1.30–1.12 (aliphatic, H, 8.29; N, 3.29; Found: C, 61.94; H, 8.07; N, 3.19%. 13H, m), 0.87 (CH3, 3H, t). dC (DMSO) 174.1 (CO), 78.6 [a]D22 +13.9 (c 0.0146 g cm-3; DMSO); Rf 0.15 chloroform– (C-1), 76.0 (C-3), 70.5 (C-2), 68.1 (C-4), 65.7 (C-5), 42.5–18.5 methanol (551). (aliphatic), 12.3 (CH3). Elemental analysis for C17H31NO5; Calc.: C, 61.98; H, 9.49; N, 4.25; Found: C, 61.84; H, 9.21; N-(4-Decycloxybenzoyl )-b-D-glucopyranosylamine 9 N, 4.03% [a]D25 +16.6 (c 0.0151 g cm-3; DMSO); Rf 0.40 dH (DMSO) 8.66 (NH, J 8.64, 1H, d), 7.88–6.98 (aromatic, chloroform–methanol (251).Jort 9.45, 4H, m), 5.01–4.90 (3×OH, 3H, m), 4.90 (H-1, J 8.03, 1H, m), 4.56 (OH, 1H, t), 4.02 (OCH2, 2H, t), 3.65–3.15 N-(4-Pentyloxybenzoyl )-b-D-glucopyranosylamine 4 (H-2–H-6¾, 6H, m), 1.57 (CH2, 2H, m), 1.35 (aliphatic chain, dH (DMSO) 8.74 (NH, J 9.45, 1H, d), 7.84, 7.28 (aromatic, 10H, m), 0.88 (CH3, 3H, t).dC (DMSO) 165.9 (CO), Jortho 8.10, 4H, m), 5.03–4.91 (3×OH, 3H, m), 4.89 (H-1, J 161.10–113.6 (aromatic), 80.2 (C-1), 78.6 (C-5), 77.6 (C-3), 72.0 8.13, 1H, m), 4.51 (OH, 1H, t), 3.68–3.10 (H-2–H-6¾, 6H, m), (C-2), 70.0 (C-4), 67.6 (OCH2), 61.0 (C-6), 31.2–22.1 (aliphatic), 1.57 (CH2, 2H, m), 1.28 (aliphatic, 4H, m), 0.87 (CH3, 3H, t). 13.9 (CH3). Elemental analysis for C23H37NO7; Calc.: C, 62.85; dC (DMSO) 166.5 (CO), 161.2–114.2 (aromatic), 80.2 (C-1), H, 8.48; N, 3.19; Found: C, 62.82; H, 8.60; N, 3.10%. 78.5 (C-5), 77.6 (C-3), 72.0 (C-2), 70.0 (C-4), 67.2 (OCH2), 61.0 (C-6), 34.9, 21.9 (aliphatic), 13.9 (CH3).Elemental analysis for N-[4-(trans-4-Pentylcyclohexyl )benzoyl]-b-D-glucopyranosyl- C18H27NO7; Calc.: C, 58.52; H, 7.37; N, 3.79; Found: C, 58.47; amine 10 H, 7.31; N, 3.77%. [a]D22 +17.2 (c 0.0153 g cm-3; DMSO); dH (DMSO) 8.72 (NH, J 8.38, 1H, d), 7.82–7.31 (aromatic, Rf 0.15 chloroform–methanol (551). Jortho 8.10, 4H, m), 5.01–4.90 (3×OH, 3H, m), 4.89 (H-1, J 7.96, 1H, m), 4.50 (OH, 1H, t), 3.67–3.11 (H-2–H6¾, 6H, m), 3.05 N-(4-Hexyloxybenzoyl )-b-D-glucopyranosylamine 5 (ArCH, 1H, m), 1.81 (CH2, 4H, m), 1.57–1.05 (aliphatic, 13H, dH (DMSO) 0.86 (m, 3H, alkyl), 1.15–1.40 (m, 6H, alkyl), 1.70 m), 0.88 (CH3, 3H, t).dC (DMSO) 166.5 (CO), 150.9–126.4 (m, 2H, alkyl), 3.07–3.38 (m, 4H, H-2, H-3, H-4, H-5), 3.43 (m, (aromatic), 80.2 (C-1), 78.7 (C-5), 77.6 (C-3), 72.1 (C-1), 70.0 1H, J5,6¾ 5, J6,6¾ 12, H-6¾), 3.67 (dd, 1H, J5,6 6, J6,6¾ 12, H-6), (C-4), 61.0 (C-6), 43.7–22.1 (aliphatic), 13.9 (CH3).Elemental 4.00 (t, J 6.5, OCH2Ph), 4.50 (t, J 6, 1H, OH), 4.87, 4.91 (d, analysis for C24H37NO6; Calc.: C, 66.18; H, 8.56; N, 3.22; 2H, 2 OH), 4.94 (t, 1H, J1,NH=J1,2 9, H-1), 5.01 (d, 1H, OH), Found: C, 66.07; H, 8.31; N, 3.16%.[a]D22 +16.1 7.00 and 7.90 (2d, J 8, Ph), 8.65 (d, 1H, JNH,1 9, NHCO). dC (c 0.0201 g cm-3; DMSO); Rf 0.12 chloroform–methanol (551). (DMSO) 13.93, 22.09, 25.18, 28.59, 31.19, 61.03 (C-6), 67.68 (alkyl ), 70.10, 72.08, 77.63, 78.65 (C-2–C-5), 80.29 (C-1), 113.80, N-[4-(trans-4-Pentylcyclohexyl )benzoyl]-b-D-mannopyranosyl- 126.13, 129.52, 161.24, (aromatic), 166.07 (CONH).Elemental amine 11 analysis for C19H29NO7; Calc.: C, 59.52; H, 7.62; N, 3.65; dH (DMSO) 8.06 (NH, J 8.15, 1H, d), 7.78–7.32 (aromatic, Found: C, 59.14; H, 7.76; N, 3.59%. Jortho 8.19, 4H, m), 5.23–4.75 (4×OH, 4H, m), 4.46 (H-1, J 1.10, 1H, t), 3.70–3.16 (H-2–H-6¾, 6H, m), 2.75 (ArCH, 1H, m), 1.78 N-(4-Heptyloxybenzoyl )-b-D-glucopyranosylamine 6 (CH2, 4H, m), 1.51–1.04 (aliphatic, 13H, m), 0.88 (CH3, 3H, dH (DMSO) 0.86 (m, 3H, alkyl), 1.15–1.40 (m, 8H, alkyl), 1.70 t).dC (DMSO) 165.4 (CO), 151.3–126.7 (aromatic), 79.2 (C-1), (m, 2H, alkyl), 3.07–3.38 (m, 4H, H-2, H-3, H-4, H-5), 3.43 (m, 78.0 (C-5), 73.9 (C-3), 70.7 (C-2), 66.8 (C-4), 61.3 (C-6), 1H, J5,6¾ 5, J6,6¾ 12, H-6¾), 3.67 (dd, 1H, J5,6 6, J6,6¾ 12, H-6), 43.7–22.1 (aliphatic), 13.9 (CH3).Elemental analysis for 4.02 (t, J 6.5, OCH2Ph), 4.51 (t, J 6, 1H, OH), 4.90 (m, 4H, C24H37NO6; Calc.: C, 66.18; H, 8.56; N, 3.22; Found: C, 66.29; 3 OH, H-1), 6.95 and 7.87 (2d, J 8, Ph), 8.65 (d, 1H, JNH,1 9, H, 8.27; N, 3.18%. [a]D24 -1.4 (c 0.0173 g cm-3; DMSO); NHCO). dC (DMSO) 13.95, 22.05, 25.45, 28.43, 28.61, 31.31, Rf 0.16 chloroform–methanol (551). 61.03 (C-6), 67.64 (alkyl ), 70.08, 72.06, 77.63, 78.67 (C-2–C-5), 80.27 (C-1), 113.78, 126.11, 129.51, 161.22 (aromatic), 166.04 N-[4-(trans-4-Pentylcyclohexyl )benzoyl]-b-D-xylopyranosyl- (CONH).Elemental analysis for C20H31NO7; Calc.: C, 60.44; amine 12 H, 7.86; N, 3.52; Found: C, 60.50; H, 7.97; N, 3.41%. dH (DMSO) 8.71 (NH, J 8.64, 1H, d), 7.81–7.31 (aromatic, Jortho 8.10, 4H, m), 5.06–4.95 (3×OH, 3H, m), 4.83 (H-1, J 9.16, N-(4-Octyloxybenzoyl )-b-D-glucopyranosylamine 7 1H, t), 3.69–3.14 (H-2–H-5¾, 5H, m), 3.05 (ArCH, 1H, m), 1.80 (CH2, 4H, m), 1.47–1.05 (aliphatic, 13H, m), 0.87 (CH3, 3H, dH (DMSO) 8.66 (NH, J 8.64, 1H, d), 7.88–6.98 (aromatic, Jortho 9.45, 4H, m), 5.01–4.90 (3×OH, 3H, m), 4.90 (H-1, J 8.03, t).dC (DMSO) 166.7 (CO), 151.0–126.5 (aromatic), 81.0 (C-1), 77.6 (C-3), 71.9 (C-2), 69.6 (C-4), 67.5 (C-5), 43.7–22.1 (ali- 1H, m), 4.56 (OH, 1H, t), 4.02 (OCH2, 2H, t), 3.65–3.15 (H-2–H-6¾, 6H, m), 1.57 (CH2, 2H, m), 1.35 (aliphatic, 10H, phatic), 13.9 (CH3).Elemental analysis for C23H35NO5; Calc.: J. Mater. Chem., 1998, 8(4), 871–880 873C, 68.12; H, 8.70; N, 3.45; Found: C, 68.07; H, 8.67; N, 3.39%. Results [a]D21 +5.9 (c 0.0159 g cm-1; DMSO); Rf 0.14 chloroform– Phase characterisation by thermal optical microscopy methanol (551).All the thermotropic mesophases were found to be of the same N-[4-(trans-4-Pentylcyclohexyl )benzoyl]-b-D-ribopyranosyl- type and exhibit the same optical appearance during optical amine 13 microscopy. For the carbohydrates with an enantiotropic mesophase, the crystals melt on heating at a discrete tempera- dH (DMSO) 8.60 (NH, J 8.45, 1H, d), 7.80–7.30 (aromatic, ture (Tm) to form a birefringent, fluid texture with an oily Jortho 8.42, 4H, m), 5.15 (H-1, J 9.25, 1H, m), 4.81–4.64 (3×OH, streak appearance of webbed focal-conic like defects typical of 3H, m), 3.92–3.42 (H-2–H-5¾, 5H, m), 2.50 (ArCH, 1H, m), a SnA* phase, see Fig. 1. Upon further heating the material 1.82 (CH2, 4H, m), 1.46–1.1 (aliphatic, 13H, m), 0.87 (CH3, became optically extinct at the clearing point (TSmA*-I). 3H, t). dC (DMSO) 166.8 (CO), 150.9–126.5 (aromatic), 76.9 Ba�tonnets are observed on cooling from this isotropic liquid, (C-1), 70.9 (C-2), 69.1 (C-3), 67.2 (C-4), 64.2 (C-5), 43.7–22.1 which coalesce quickly in the bulk to form focal-conic domains.(aliphatic), 13.9 (CH3). Elemental analysis for C23H35NO5; As each sample was cooled further, the hydrophilic end of the Calc.: C, 68.12; H, 8.70; N, 3.45; Found: 67.99; H, 8.51; carbohydrate molecules adhered more strongly to the glass N, 3.43%. [a]D23 -2.6 (c 0.0093 g cm-3; DMSO); Rf 0.17 surface via hydrogen bonding. Thus, most of the resultant chloroform–methanol (551).textures became homeotropic and optically extinct. These observations indicate that the phase is optically uniaxial (if N-(4=-(Nonyloxybiphenyl-4-ylcarbonyl )-b-D-mannopyranosylthe mesophase were biaxial then a residual birefringence for amine 14 the sample would be observed). However, focal-conic defects dH (DMSO) 8.21 (NH, J 10.8, 1H, d), 7.93–7.03 (aromatic, 4H, could still be observed around air bubbles and at the edges of m), 5.28–4.73 (4×OH, 4H, m), 4.48 (H-1, J 1.14, 1H, m), 4.02 the sample. This optical behaviour, i.e.simultaneous presence (OCH2, 2H, t), 3.72–3.15 (H-2–H-6¾, 6H, m), 1.73 (CH2, 2H, of both homeotropic and focal-conic textures, indicates that m), 1.30–1.22 (aliphatic, 12H, m), 0.88 (CH3, 3H, t). dC (DMSO) the mesophase is a calamitic smectic A* phase.8,13,14,33 The 165.2 (CO), 158.9–114.9 (aromatic), 79.4 (C-1), 78.4 (C-5), 78.1 notation smectic A* is used to describe the smectic A phase (C-3), 73.9 (C-2), 70.0 (C-4), 67.5 (OCH2), 61.2 (C-6), 31.2–22.1 exhibited by these compounds as they are optically active and, (aliphatic), 13.9 (CH3).Elemental analysis for C28H39NO7; therefore, the A* phases formed by them must have reduced Calc.: C, 67.04; H, 7.84; N, 2.79; Found: C, 66.91; H, 7.61; symmetry.33 The characterisation of these defects classifies the N, 2.71%.[a]D23 -2.8 (c 0.0100 g cm-3; DMSO); Rf 0.23 mesophase as being smectic A* with a layered structure where chloroform–mhanol (551). the long axes of the molecules are on average orthogonal to the layer planes and the in-plane and out-of-plane positional N-(2=-Fluoro-4=-octyloxybiphenyl-4-ylcarbonyl )-b-D-manno- ordering of the molecules is short range, see X-ray studies.pyranosylamine 15 DiVerential scanning calorimetry dH (DMSO) 8.05 (NH, 1H, t), 7.91–7.02 (aromatic, 7H, m), 5.28–4.77 (4×OH, 4H, m), 4.47 (H-1, J 1.16, 1H, t), 4.01 The enthalpy values for the remaining (Tm) and clearing points (OCH2, 2H, t), 3.69–3.15 (H-2–H-6¾, 6H, m), 1.72 (CH2, 2H, (TSmA*-I) of the N-(4-alkoxybenzoyl)-b-D-glucopyranosylm), 1.41–1.5 (aliphatic, 10H, m), 0.87 (CH3, 3H, t).dC (DMSO) amines (4–9) are typical of liquid crystalline carbohydrates 162.9 (CO), 159.4–114.9 (aromatic), 79.4 (C-1), 77.4 (C-5), 74.5 (e.g. 17.1 and 1.34 J g-1, respectively, for compound 9). The (C-3), 73.9 (C-2), 70.6 (C-4), 67.6 (OCH2), 61.2 (C-6), clearing point enthalpies are relatively small in comparison to 31.20–22.1 (aliphatic), 13.9 (CH3).Elemental analysis for the melting enthalpies, and the values measured are of a similar C27H36NO7F; Calc.: C, 64.14; H, 7.18; N, 2.77; Found: C, 64.03; magnitude to those found in conventional liquid crystal systems H, 7.09; N, 2.73%. [a]D22 -3.9 (c 0.0123 g cm-3; DMSO); which exhibit SmA* to isotropic liquid transitions.A typical Rf 0.18 chloroform–methanol (551). heating thermogram for the N-4-decyloxybenzoyl-b-D-glucopyranosylamines 9 is shown in Fig. 2. It is clear that the transitions Tm and TSmA*-I are both first-order transitions. The N-(2-Fluoro-4=-tetradecyloxybiphenyl-4-ylcarbonyl )-b-Ddecomposition generally observed at TSmA*-I is shown especially mannopyranosylamine 16 clearly in Fig. 2. TSmA*-I is often lower on cooling due to dH (DMSO) 8.05 (NH, J 8.19, 1H, d), 7.92–7.03 (aromatic, 7H, m), 5.28–4.78 (4×OH, 4H, m), 4.49 (H-1, J 1.10, 1H, t), 4.01 (OCH2, 2H, t), 3.70–3.16 (H-2–H-6¾, 6H, m), 1.72 (CH2, 2H, m), 1.30–1.24 (aliphatic, 22H, m), 0.87 (CH3, 3H, t). dC (DMSO) 162.0 (CO), 150.3–114.9 (aromatic), 79.4 (C-1), 77.2 (C-5), 74.6 (C-3), 73.0 (C-2), 70.1 (C-4), 67.5 (OCH2), 61.9 (C-6), 31.2–22.2 (aliphatic), 13.9 (CH3).Elemental analysis for C33H48NO7F; Calc.: C, 67.21; H, 8.20; N, 2.38; Found: C, 67.12; H, 8.14; N, 2.36%. N-(2=,3=-Difluoro-4=-octyloxybiphenyl-4-ylcarbonyl )-b-Dmannopyranosylamine 17 dH (DMSO) 8.25 (NH, 1H, d), 7.52–7.05 (aromatic, 6H, m), 5.24–4.82 (4×OH, 4H, m), 4.50 (H-1, J 1.03, 1H, m), 4.02 (OCH2, 2H, t), 3.70–3.15 (H-2–H-6¾, 6H, m), 1.55 (CH2, 2H, m), 1.40–1.37 (aliphatic, 10H, m), 0.88 (CH3, 3H, t).dC (DMSO) 162.3 (CO), 155.8–116.4 (aromatic), 85.8 (C-1), 84.6 (C-5), 73.5 (C-3), 72.1 (C-2), 68.2 (C-4), 64.8 (OCH2), 61.9 (C-6), 32.9–22.9 (aliphatic), 14.4 (CH3). Elemental analysis for C27H35NO7F2; Calc.: C, 61.94; H, 6.74; N, 2.68; Found: C, 61.77; H, 6.59; Fig. 1 Photomicrograph at 185 °C of the oily smectic A* texture N, 2.54%. [a]D23 -2.5 (c 0.0188 g cm-3; DMSO); Rf 0.22 observed on melting N-(4-decyloxybenzoyl)-b-D-glucopyranosylamine 9 (magnification ×160) chloroform–methanol (551). 874 J. Mater. Chem., 1998, 8(4), 871–880for the recording of suYcient data before degradation of the sample sets in.Thus synchrotron radiation was employed, using the experimental set-up of station 8.2 at CCLRC Daresbury, described elsewhere.54–56 The samples were prepared as polycrystalline powders in Lindemann tubes and kept at a controlled temperature allowing for the recording of diVraction data whilst performing a temperature scan of 2 °C min-1 in the temperature interval of 150 to 210 °C or keeping the sample at a chosen temperatures.The selected experimental set-up was limited to the recording of data relating to lattice parameters greater than 17.8 A ° and the use of wet rat-tail collagen as calibration standard led to a systematic error of 3% of the observed d spacings.57 Fig. 2 DiVerential scanning thermogram as a function of temperature for the first heating cycle for N-(4-decyloxybenzoyl)-b-D-glucopyranosylamine 9 (scan rate 10 °C min-1) decomposition.On cooling there is a greater tendency to form glasses rather than to recrystallise. The DSC thermogram N-octadecanoyl-b-D-glucopyranosylamine 28: Cr–SmA*, 170 °C; shown in Fig. 3 for the N-(trans-4-pentylcyclohexyl-acetyl )-b- SmA*–I, 220 °C45,46 D-glucopyranosylamine 23 shows Tm and TSmA*-I close together The experimental results for N-octadecanoyl-b-D-glucopyr- followed by thermal decomposition.The thermograms depicted anosylamine 2845,46 are shown in Fig. 4. Throughout the in Fig. 2 and 3 are those observed for the first heating cycle. crystalline state reflections relating to lattice parameters of d1 Thermal decomposition above TSmA*-I and the formation of 46.6 A ° and d2 37.2 A ° were observable.Whereas the ratio of glasses on cooling often led to non-reproducible thermograms the intensities of d15d2 is approximately 3551 at 150°C, a ratio for the second or third heating cycles. which remains almost constant up to 166 °C, the relative intensity of the reflections relating to d2 increases sharply Lyotropic mesomorphism above 170 °C with rising temperature.Above that temperature Attempts to obtain lyotropic phases for the new carbohydrate an intensity maximum relating to a d spacing d2 of 37.2 A ° derivatives under the conditions described in the Experimental could be detected, relating to the full formation of the SmA* were unsuccessful as they were all essentially insoluble in water. phase. The recorded transition temperature of 170 to 172 °C The presence of one, two or combinations of cyclohexyl and with a biphasic area up to 178 °C is in line with results phenyl rings with various chains and lateral substituents in observed by optical polarizing microscopy and by DSC the glycopyranosylamines 1–17 does not increase the water measurements.Upon reaching the liquid-crystalline state the solubility of these carbohydrates compared to that of the diVraction pattern coalesced into a meridional maximum, corresponding N-acyl-b-D-glycopyranosylamines.45,46 This is indicative of some macroscopic alignment of the material via not purely a question of the high Tm of the glycopyranosyl- interactions with the walls of the capillary.Elevation of the amines 1–17 as the N-(4-alkoxybenzoyl)-b-D-glucopyranosyl- temperature led to a reduction in the maximum, in line with amines 4–9 exhibit similar Tm to those of the acyl substituted a loss of macroscopic ordering.Up to 208 °C the d spacing compounds.45,46 fluctuated between 37.2–35.3 A ° . Above that temperature the X-Ray investigations In order to characterize the solid state of these materials the aromatic N-(4-decyloxybenzoyl)-b-D-glucopyranosylamine 9 and the related, but non-aromatic, N-octadecanoyl-b-Dglucopyranosylamine 2845,46 and N-octadecanoyl-2-amino-2- deoxy-b-D-glucopyranose 2945,46 were investigated using X-ray diVraction. The tendency of carbohydrates to decompose at high temperatures required an experimental set-up allowing Fig. 4 Variation of the d spacings (layer thickness) of N-octadecanoylb- D-glucopyranosylamine 28 with temperature; (%) strong and (&) weak diVraction reflections obtained for the crystalline phase via temperature scans at 2 °C min-1, (#) strong and ($) weak reflections Fig. 3 DiVerential scanning thermogram as a function of temperature for the biphasic region, and (6) strong and (+) weak reflections for the liquid crystalline region, observed after heating from ambient for the first heating cycle for N-(trans-4-pentylcyclohexylacetyl)-b-Dglucopyranosylamine 23 (scan rate 10 °C min-1) temperature J.Mater. Chem., 1998, 8(4), 871–880 875Fig. 5 A model of N-octadecanoyl-b-D-glucopyranosylamine 28 obtained from CERIUS 2.0 material either reached its liquid state or rapid thermal degradation set in, leading to the absence of small angle intensity maxima.The data recorded by performing a temperature scan of 2 °C min-1 are in line with those found by heating samples from ambient temperature and starting the collection of data after thermal equilibrium was indicated by the temperature control unit. The spacing of 46.6 A ° at 150 °C indicates that, compared to the overall length of one molecule 26.4 A ° [obtained via molecular modelling of one molecule in the gas phase at 0 K using the CERIUS 2.0 (MSI) software, see Fig. 5], Fig. 8 Variation of the d spacings (layer thickness) of N-(4- decyloxybenzoyl)-b-D-glucopyranosylamine 9 with temperature; a bilayer structure must be present in the crystalline state. (%) reflections obtained for the crystalline phase via temperature scans The observed values of d2 of 37.2–35.3 A ° in the liquid at 2 °C min-1, (#) reflections for the biphasic region, and (6) crystalline state up to the disappearance of the maximum reflections for the liquid crystalline state, observed after heating from indicate a contraction of the layering with respect to the ambient temperature crystalline modification in the type of an intercalated bipolar structure of the SmA* phase depicted schematically in Fig. 6 and observed earlier for the crystalline phases of diverse liquid- ture dependent phase behaviour was observed. Whereas at 170 °C the material was found to be predominantly crystalline crystalline carbohydrates.11–14,38,41,52,53,58 For N-4-decyloxybenzoyl-b-D-glucopyranosylamine 9, whose with a lattice parameter d1 of 41.8 A ° , a temperature scan from 180 to 220 °C showed at 180 °C reflections relating to d spacings molecular length was determined by molecular modelling, as above, to be 23.4 A ° , see Fig. 7, a fundamentally similar tempera- d1 and d2 of 41.8 and 37.2 A ° and from 182 °C upwards the Fig. 6 A schematic model of the layer structure of N-octadecanoyl-b-D-glucopyranosylamine 28 in the smectic A* phase Fig. 9 A model of N-octadecanoyl-2-amino-2-deoxy-b-D-gluco- Fig. 7 A model of N-(4-decyloxybenzoyl)-b-D-glucopyranosylamine 9 obtained from CERIUS 2.0 pyranose 29 obtained from CERIUS 2.0 876 J. Mater. Chem., 1998, 8(4), 871–880Table 1 Melting points (°C) of the N-trans-(4-pentylcyclohexyl- Table 4 Melting points (°C) of variously substituted N-[4¾-alkoxybiphenyl- 4-ylcarbonyl)-b-D-mannopyranosylamines 14–17 carbonyl)-b-D-glycopyranosylamines 1–3 compound sugar Cr I compound LC Cr I 1 $ 260 (decomp.) $ 14 $ #235 (decomp.) $ 2 $ #230 (decomp.) $ 15 $ #240 (decomp.) $ 3 $ #230 (decomp.) $ 16 $ #230 (decomp.) $ 17 $ #235 (decomp.) $ Table 2 Transition temperatures (°C) for the N-(4-alkoxybenzoyl)-b- D-glucopyranosylamines 4–9 Table 5 Transition temperatures (°C) for the 1-deoxy-1-(4- decyloxyphenyl)-b-D-glucopyranose 18, 4-decyloxybenzyl-b-D-glucopyranose 19 and N-(4-decyloxybenzoyl)-b-D-glucopyranosylamine 9 compound n Cr SmA* I 4 5 $ 198 — $ 5 6 $ 194 $ (178)a $ compound Z Cr SmA* I ref. 6 7 $ 187 $ 199 $ 7 8 $ 191 $ 214 $ 18 — $ 107 $ 173 $ 41 8 9 $ 181 $ 220 $ 19 OCH2 $ 117 $ 166 $ 44 9 10 $ 187 $ 224 $ 9 NCHO $ 187 $ 224 $ aRepresents a monotropic transition temperature.absence of a maximum relating to d1 and a fluctuation of the spacings d2 between 36.5 and 34.1 A ° , see Fig. 8. These values suggest the occurrence of an non-intercalated structure in the Table 3 Melting points (°C) of the N-[4-trans-4-pentylcyclohexyl)- crystalline phase as opposed to an intercalated structure in the benzoyl]-b-D-glycopyranosylamines 10–13 liquid-crystalline state as observed for N-octadecanoyl-b-Dglucopyranosylamine 28.The small reduction of d2 for 28 in the d spacing in the liquid-crystalline state and the reduction and then subsequent increase and broadening of the d spacing observed for 9, which is unlike the behaviour generally observed in thermotropic liquid crystals, requires some consideration.Although any explanation of the behaviour of these compound sugar Cr I partially oriented liquid-crystalline materials close to their isotropisation temperature has to be tentative, the concept of the nature of the liquid-crystalline state in carbohydrates oVers some insight.58 With rising temperatures the increased molecu- 10 $ #230 (decomp.) $ lar motion and disordering can either be accommodated by a decreasing overlap of the alkyl chains, leading to increased d spacings, a feature not observed in the investigated materials, or by an increasing overlap of the carbohydrate groups leading 11 $ #235 (decomp.) $ to a decrease in the d spacing, but also to an expansion of the layers in their planes, required to allow for the overlap of the bulky head groups, and accommodating for the increased 12 $ #230 (decomp.) $ mobility of the flexible alkyl chains. The reduced values of the intensities of the diVraction data and the broadening of the reflections for temperatures immedi- 13 $ #240 (decomp.) $ ately preceding the isotropisation temperature suggest the onset of the breakdown of liquid-crystalline (long-range) order.The data correspond to the formation of species with periodicities between 27.6–45.5 A° . These lengths correspond roughly to J. Mater. Chem., 1998, 8(4), 871–880 877Table 6 Transition temperatures (°C) of the N-trans-4-pentylcyclohexylcarbonyl-b-D-glucopyranosylamine 1, octyl b-D-glucopyranoside 20, trans- 4-butylcyclohexyl b-D-glucopyranoside 21, 1-O-(trans-4-propylcyclohexyl)methyl b-D-glucopyranoside 22 and N-trans-4-pentylcyclohexylacetylb- D-glucopyranosylamine 23 compound structure Cr SmA* I ref. 20 $ 69 $ 110 $ 11 21 $ 137 $ 167 $ 42 22 $ 125 $ 147 $ 42 1 $ 259 $ — $ 23 $ 232 $ 235 $ the length of a single molecule and of a dimer. At higher although a general tendency of Tm to decrease with increasing chain length is apparent. temperatures thermal degradation sets in.The replacement of the alkoxy chain in the N-(4- alkoxybenzoyl)-b-D-glucopyranosylamines 4–9 by a cyclohexyl ring to yield the N-[4-(trans-4-pentylcyclohexyl)benzoyl]-b- D-glucopyranosylamine 10, shown in Table 3, suppresses the liquid crystalline behaviour by increasing Tm. No mesomorphism can be observed before decomposition occurs. This is also the case for the carbohydrates 10–12 where the incorporation N-octadecanoyl-2-amino-2-deoxy-b-D-glucopyranose 29; Cr–I, 197 °C45,46 of the additional 1,4-disubstituted phenylene ring into the esters 1–3 to yield the compounds 10–12 probably increases As expected for N-octadecanoyl-2-amino-2-deoxy-b-D- Tm, so that only decomposition is observed at elevated temperaglucopyranose 2945,46 liquid-crystalline behaviour was not tures.The compound 13 is also non-mesomorphic below the observed. This is consistent with observations by optical decomposition temperature. microscopy and DSC.45,46 However, between 182–194 °C inten- Increasing the size of the hydrophobic part in the liquid sities corresponding to values of 47.7–48.8 A° were detected, crystalline carbohydrate in the mannosylamine 2 and 11 with which compared to the maximum length of 28.1 A ° of the one phenyl ring to yield the mannosylamines 14–17 with two compound, as determined by modelling and shown in Fig. 9, substituted phenyl rings again does not increase the tendency is indicative of a solid state of similar nature as observed for for liquid crystal formation, see Table 4.This implies that the the other two materials. incorporation of more than one aliphatic or aromatic ring into a liquid crystalline carbohydrate gives rise to such a high value for Tm that no liquid crystalline behaviour can be observed Discussion before decomposition takes place. This suggests that a normal The N-(trans-4-pentylcyclohexylcarbonyl)-b-D-glycopyranosyl- alkyl chain or one alicyclic or aromatic ring is suYcient for amines 1–3 shown in Table 1 are not mesomorphic. The values mesophase formation at elevated temperatures.Attempts to recorded in the table are melting points (Tm) or the temperature actually lower Tm and TSmA*-I would be more rewarding for at which significant visible decomposition occurs. The con- practical applications, since solubility in water should increase.figuration of the carbohydrate moiety seems to be much less The thermal data collated in Table 5 show that the presence important than the high degree of hydrogen bonding attribu- of an additional site for hydrogen bonding in the linkage Z, table to the hydroxy groups on the carboyhydrate part of the e.g. an amide linkage in 9, has a much greater influence on Tm molecule and the amino and carbonyl groups of the amide and TSmA*-I than the presence or absence of a linkage, linkage in generating these high Tm values.e.g. compare TSmA*-I for the compounds 18 and 19. The transition temperatures and enthalpies of transition for The transition temperatures for the carbohydrates listed in the N-(4-alkoxybenzoyl)-b-D-glucopyranosylamines 4–9 are Table 6 indicate that alicyclic rings, such as cyclohexane, lead collated in Table 2.As usual liquid crystalline behaviour is to higher Tm and TSmA*-I; for example compare the thermal found after a certain critical chain length (n=6) is reached. data for the compounds 20 and 21. However, the data also TSmA*-I increases as the number of methylene units in the imply that the odd numbers of units in the central linkage give alkoxy chain increases.Tm is almost independent of chain rise to higher TSmA*-I than an even number of units; see the data for compounds 20 and 21 and the compounds 1 and 23. length, as is often the case for liquid crystalline carbohydrates, 878 J. Mater. Chem., 1998, 8(4), 871–880Table 7 Transition temperatures (°C) of the 4-cyanophenyl trans-4-octylcyclohexanoate 24, 4-cyanophenyl trans-4-(trans-4-pentylcyclohexyl)- cyclohexanecarboxylate 25, 4-cyanophenyl trans-4-(trans-4-pentylcyclohexylmethyl )cyclohexanecarboxylate 26 and the 4-cyanophenyl trans-4- cyanophenyl trans-4-[2-(trans-4-pentylcyclohexyl)ethyl]cyclohexanecarboxylate 27 compound structure Cr SmB N I ref. 24 $ 62 — $ 79 $ 59 25 $ 92 — $ 232 $ 60 26 $ 110 — $ (108)a $ 61 27 $ 90 $ (65)a $ 197 $ 61 aRepresents a monotropic temperature.An additional site of hydrogen bonding has a bigger influence References on the transition temperatures than the nature of the linkage, 1 M. Berthelot, Compt. Rend., 1855, 41, 452; J. Prakt. Chem., 1856, compare 22 and 1. This is exactly the opposite situation to 671, 235. that observed for the thermotropic liquid-crystalline behaviour 2 E. Fischer and B.Helferich, L iebigs Ann. Chem., 1911, 383, 68. of non-amphiphilic liquid crystals. This is illustrated clearly 3 A. H. Salway, J. Chem. Soc., 1913, 103, 1022. 4 P. Gaubert, Compt. Rend., 1919, 168, 277. by the thermal data collated in Table 7 for a number of non- 5 V. Vill, H. Kelkenberg and J. Thiem, L iq.Cryst., 1992, 11, 459. amphiphilic liquid crystals incorporating cyclohexane rings 6 C. R. Noller and W. C. Rockwell, J. Am. Chem. Soc., 1938, 60, 2076. (24–27). It is evident that the compounds with an aliphatic 7 J. C. Chabala and T. Y. Shen, Carbohydr. Res., 1978, 67, 55. ring incorporating no linkage (25) or an even number of 8 E. Barral, B. Grant, M. Oxsen, E. T. Samulski, P.C. Moews, carbon atoms (27) exhibit significantly higher clearing points J. R. Knox, R. R. Gaskill and J. L. Haberfeld, Org. Coat. Plast. than that (26) with an odd number of carbon atoms in the Chem., 1979, 40, 67. 9 V. Vill, Mol. Cryst. L iq. Cryst., 1992, 213, 67. linkage. This has been attributed to the linear conformation 10 J. W. Goodby,Mol. Cryst. L iq. Cryst., 1984, 110, 205.of the former and a bent or kinked conformation for the latter. 11 G. A. JeVrey, Acc. Chem. Res., 1986, 19, 168. 12 G. A. JeVrey and L. M. Wingert, L iq. Cryst., 1992, 12, 179. 13 G. A. JeVrey and S. Bhattacharjee, Carbohydr. Res., 1983, 115, 53. Conclusions 14 H. A. van Doren and L. M. Wingert, Mol. Cryst. L iq. Cryst., 1991, 198, 381. The amide linkage in liquid crystalline carbohydrates generally 15 R.G. Zimmermann, G. B. Jameson, R. Weiss and G. Demailly, gives rise to high transition temperatures and to insolubility, Mol. Cryst. L iq. Cryst., L ett., 1985, 1, 183. at least in the pyranose form, in water. This is probably due 16 B. Kohne, W. Praefcke, W. Stephan and P. Nu�rnberg, Z. to a high degree of intermolecular hydrogen bonding also Naturforsch., T eil B, 1985, 40, 981. 17 W. V. DahlhoV, Z. Naturforsch., T eil B, 1987, 42, 661. involving the amide linkage. There is no correlation between 18 B. Pfannemu� ller, W.Welte, E. Chin and J. W. Goodby,Mol. Cryst. the thermotropic liquid crystalline behaviour of non-amphi- L iq. Cryst., 1986, 1, 357. philic and amphiphilic liquid crystals as a function of molecular 19 M. Marcus and P.L. Finn, L iq. Cryst., 1988, 30, 381. structure. This is exemplified by the opposite odd-even eVect 20 Y. J. Chung and G. A. JeVrey, Biochim. Biophys. Acta, 1989, 985, on the mesomorphic behaviour of the central linkage between 300. 21 T. E. Thomson and A. Baron, Biochim. Biophys. Acta, 1975, 382, the hydrophilic and hydrophobic parts of liquid crystalline 276. carbohydrates and that of related non-amphiphilic liquid crys- 22 J.Diesenhofer and H. Michel, Angew. Chem., 1989, 101, 872. tals. The presence of alicyclic rings, such as cyclohexane, or 23 E. Lederer, Chem. Phys. L ipids, 1976, 16, 91. more than one phenyl ring, or combinations of cyclohexane 24 D. Asselineau and J. Asselineau, Prog. Chem. Fats Other L ipids, and phenyl rings leads to such high melting points that no 1978, 16, 59.mesomorphism can be observed before thermal decomposition 25 H. Arita, K. Sugita, A. Nomura, K. Sato and J. Kawanami, Carbohydr. Res., 1978, 62, 143. occurs. A non-intercalated structure is found for the crystalline 26 K. Hill, in Carbohydrates as Raw Materials II, ed. G. Descotes, state of the carbohydrates investigated. The chains appear to VCH Verlag, Weinheim, 1993, p. 163 and references therein. adopt an intercalated structure in the smectic A* phase after 27 J. G. Riess and J. Greiner, in Carbohydrates as Raw Materials II, melting has occurred. ed. G. Descotes, VCH Verlag, Weinheim, 1993, p. 209 and references therein. 28 M. J. Lawrence, Chem. Soc. Rev., 1994, 417. We gratefully acknowledge the EPSRC for support of post- 29 V.Vill, H.-W. Tunger and M. Paul, J.Mater. Chem., 1995, 6, 2283. graduate (M. G.), postdoctoral (P. L. and G. M.) and Advanced 30 V. Vill, H.-W. Tunger and M. von Minden, J. Mater. Chem., 1996, Fellowships (S. M. K.). Mrs B. Worthington (1H NMR), Mr R. 6, 739. Knight (IR), C. Kennedy (CHN) and Mr A. D. Roberts (MS) 31 Structure of Biological Membranes, ed. S. Abrahamson and I.Pascher, 1977, Plenum Press, New York. are thanked for their assistance. J. Mater. Chem., 1998, 8(4), 871–880 87932 L iquid Crystals and Biological Systems, ed. J. J. Wolken and 47 P. Leon-Ruaud, M. Allainmat and D. Plusquellec, T etrahedron L ett., 1991, 1557. G. H. Brown, 1980, Academic Press, New York. 48 H. Kunz, Angew. Chem., Int. Ed. Engl., 1987, 26, 294. 33 J.W. Goodby, J. A. Haley, G. Mackenzie, M. J. Watson, 49 U. Jeschke, C. Vogel, V. Vill and H. Fischer, J.Mater. Chem., 1995, D. Plusquellec and V. Ferrie`res, J.Mater. Chem., 1995, 5, 2209. 5, 2073. 34 J. W. Goodby, J. A. Haley, G. Mackenzie, M. J. Watson, S. M. 50 H. A. van Doren, R. van der Geest, C. F. de Ruijer, R. M. Kellogg Kelly, P. Lettelier, O. Douillet, P. Gode�, G. Goethals, G.Ronco and H. Wynberg, L iq. Cryst., 1990, 8, 109. and V. Villa, L iq. Cryst., 1997, 22, 367. 51 B. Pfannemu� ller and W. Welte, Chem. Phys. L ipids, 1985, 37, 227. 35 J. W. Goodby, J. A. Haley, G. Mackenzie, M. J. Watson, S. M. 52 D. Baeyens-Volant, P. Cuvelier, R. Fornasier, E. Szalai and Kelly, P. Lettelier, O. Douillet, P. Gode�, G. Goethals, G. Ronco, C. David,Mol. Cryst. L iq. Cryst., 1985, 128, 277. B. Harmouch, P. Martin and V. Villa, L iq. Cryst., 1997, 22, 497; 53 D. Baeyens-Volant, R. Fornasier, E. Szalai and C. David, Mol. J. Carbohydr. Chem., 1997, 16, 479. Cryst. L iq. Cryst., 1986, 135, 93. 36 P. Lettelier, D. F. Ewing, J. W. Goodby, J. A. Haley, S. M. Kelly 54 J. A. Boustra, G. S. Gooris, W. Bras and H. Talsma, Chem. Phys. and G. Mackenzie, L iq. Cryst., 1997, 22, 609. L ipids, 1993, 64, 83. 37 H. Prade and R. Miethchydr. L ett., 1994, 1, 19. 55 W. Bras and J. A. Boustra, NIMPR, 1993, A326, 587. 38 H. Prade, R. Miethchen and V. Vill, J. Prakt. Chem., 1995, 337, 427. 56 E. Towns-Andrews, A. Berry, J. G. Bordas, R. Mant, P. K. Murray, 39 P. Stangier, V. Vill, S. Rohde, U. Jeschke and J. Thiem, L iq. Cryst., K. Roberts, I. Summer, J. S. Worgan, R. Lewis and A. Gabriel, 1994, 17, 589. Rev. Sci. Instrum., 1989, 60, 2346. 40 V. Vill, T. Bo� cker, J. Thiem and F. Fischer, L iq. Cryst., 1989, 6, 349. 57 W. Folkhard, W. Geercken, E. Knoerzler, E. Mosler, 41 G. A. JeVrey, Mol. Cryst. L iq. Cryst., 1984, 110, 221. H. Nemetschek-Gansler, T. Nemetschek and M. H. J. Koch, 42 C. Tschierske, A. Lunow and H. Zaschke, L iq. Cryst., 1990, 8, 885. J.Mol. Biol., 1987, 193, 405. 43 D. Joachimi, C. Tschierschke, H. Mu� ller, J. H. WendorV, 58 I. Pascher and S. Sundell, Chem. Phys. L ipids, 1977, 20, 171. L. Schneider and R. Kleppinger, Angew. Chem., Int. Ed. Engl., 1993, 59 H.-J. Deutscher, F. Kuschel, S. Ko� nig, H. Kresse, D. PfeiVer, 32, 1165. A. Wiegeleben, J. Wulf and D. Demus, Z. Chem. (L eipzig), 1977, 44 P. Lettelier, D. F. Ewing, J. W. Goodby, J. A. Haley, S. M. Kelly 17, 64. and G. Mackenzie, L iq. Cryst., 1997, 22, 609. 60 R. Eidenschink, Kontakte, 1977, 1, 15. 45 D. F. Ewing, M. Glew, J. W. Goodby, J. A. Haley, S. M. Kelly, 61 H.-J. Deutscher, R. Frach and R. Krieg, 1986, Proceedings of P. Lettelier and G. Mackenzie, Proceedings of the 26th Freiburger the 16th Freiburger Arbeitstagung Flu� ssigkristalle, Freiburg, Germany. Arbeitstagung Flu� ssigkristalle, Freiburg, Germany, 1997, p. 55. 46 D. F. Ewing, J. W. Goodby, J. A. Haley, S. M. Kelly, P. Lettelier and G. Mackenzie, L iq. Cryst., 1997, 23, 759. Paper 7/07013F; Received 29th September, 1997 880 J. Mater. Chem., 1998, 8(4), 871

ISSN:0959-9428    

DOI:10.1039/a707013f

出版商:RSC    

年代:1998

数据来源: RSC