J O U R N A L O F C H E M I S T R Y Materials Induction of liquid crystalline phases in linear polyamines by complexation of transition metal ions Hartmut Fischer,b Thomas Plesnivy,a Helmut Ringsdorfa and Markus Seitz*,a,† aInstitut fu� r Organische Chemie, J. J. Becherweg 18–20, Johannes Gutenberg-Universita� t, 55099 Mainz, Germany bT echnische Universiteit Eindhoven, Polymer Chemistry & T echnology 5600MB Eindhoven, Netherlands In this paper, we describe the preparation and characterization of polyamine–copper complexes based on N-alkylated polyethyleneimines which were obtained by polymer analogous reduction of the corresponding N-acylated polymers.Liquid crystalline properties of the originating linear polyamides are lost in the resulting polyamines which show a higher conformational mobility.In analogy to low molecular mass azacrown derivatives, the complexation of transition metal ions, here copper(II), may reinduce liquid crystalline behavior. The structure formation in the liquid crystalline state is discussed based on X-ray investigations. The generation of well-defined macroscopic structures by non- cyclene derivative peracylated with 4-dodecyloxybenzoyl chloride.15 covalent interaction of two or more individual components is In general, the use of suitable side groups—e.g.the 3,4- of great interest in liquid crystalline research, and more generdialkoxybenzoyl unit16—provides the formation of columnar, ally in the so-called field of ‘supramolecular chemistry’.1–3 For especially discotic hexagonal mesophases (Dh) for a large instance, two diVerent molecular subunits that do not show number of substituted macrocyclic oligoamides.This was any mesomorphic properties themselves may interact to form related to their discoidal molecular geometry, which in turn a liquid crystalline ‘supramolecule’. The self-organization of had been associated with the restricted conformational mobility such systems may take place either in solution, in the solid or of the central macrocycle by N-acylation with benzoic or even in the liquid crystalline state by various kinds of molecular cinnamoic acid derivatives.17–19 As a consequence, discotic interactions such as hydrogen bonding, donor–acceptor interazacrown derivatives lose their mesomorphic properties by actions or the complexation of transition metal ions.4–9 The complete reduction to the N-alkylated macrocycle, due to the induction of liquid crystalline phases in polymeric materials more flexible linkage of the side chains.8 However, a confor- by means of low molecular mass ‘dopants’, such as electron mational fixation of the N-alkylated core and thus the reinduc- acceptor molecules or small cations that act as central units tion of a columnar mesophase may be achieved by coordination for a complex subunit is particularly interesting.10–12 of the central core with transition metal ions,8,17,20 as shown Inorganic as well as organic amino compounds, especially in Fig. 1 for tetrasubstituted cyclam derivatives. In a similar multivalent (‘polydentate’) amino ligands such as linear and fashion, even the salts of protonated cyclic oligoamines may branched oligomers of the aminoethylene unit (dien and trien be enabled to form thermotropic mesophases.21 etc.), have played an outstanding role in the development of Considering azamacrocycles as cyclic oligomers of ethy- coordination chemistry.13 In the context of the work presented leneimine, well-known principles relating structure and proper- here, macrocyclic amino ligands (azacrowns)14 are of particular ties of N-substituted azacrowns can be applied to N-substituted interest.Their principal ability to serve as a central unit of polyethyleneimines. In particular, the formation of thermo- thermotropic liquid crystals was first recognized for a hexatropic hexagonal columnar mesophases by N-benzoylated as well as N-cinnamoylated azamacrocycles was also realized for * E-mail: seitm@engineering.ucsb.edu the equivalent linear polymers.22–24 The particular packing † Present address: Department of Chemical Engineering, University of behavior of N-3,4-acylated ethyleneamine fragments results in California, Santa Barbara, CA 93106, USA.a nearly identical thermotropic phase behavior for cyclic and ‡ The structure shown was obtained by first minimizing the steric linear oligomers as well as for linear polymers (Fig. 2).24 energy of the decamer of N-dimethoxybenzoylethyleneimine starting For the linear polyamide 3 (PEI-3,4), based on miscibility both, from a ‘prefolded’ chain or a zig-zag conformation (2/1-helix). studies and entropic considerations it has been proposed that The final replacement of the methoxy substituents by decyloxy chains in the second step of the energy minimization gave rise to only gradual the columnar aggregates would consist of single helically folded changes in the position of the atoms in the polymer main chain. polymer chains surrounded radially by the alkyl chains.22,24 However, this structure which was obtained using a standard software The polymer backbone of such a possible energetically favored package (CSC Chem3D for Macintosh) should not be viewed as the conformation is shown in Fig. 3. One helix turn is formed by result of an extensive modelling calculation. It rather provides a refined four to five repeat units resulting in a column center with a and reasonable model that we found very useful for the discussion of diameter of about 5 A ° adopting a shape similar to the tetraaza- the packing requirements along the polymer main chain.Nevertheless, cycles; the pitch height in this model is 7.3 A ° .‡ it is in good agreement with earlier molecular modeling calculations in which a 14/3-helical conformation has been proposed for polyoxazol- Based on the N-substituted alkyleneamine fragment one ines containing chiral centers in the main chain (cf.Y. S. Oh, may think of a variety of diVerent molecular architectures, T. Yamazaki and M. Goodman, Macromolecules, 1992, 25, 6322). In such as branched or dendrimeric structures.24–26 In this context, this context, a discussion of the induction of a helical main chain the phase separation in a hydrophilic central region surrounded conformation by the sterical requirements of tapered side groups has by hydrophobic alkyl groups has been discussed as the main been presented recently (cf.V. Percec, D. Schlueter, J. C. Ronda, driving force for the hexagonal columnar mesomorphism G. Johansson, G. Ungar and J. P. Zhou, Macromolecules, 1996, 29, 1464). observed in dendrimeric polymers of higher generation.27 J.Mater. Chem., 1998, 8(2), 343–351 343N N N N CO CO R R R CO R CO N N N N CH2 CH2 R R CH2 R CH2 R N N N N CH2 CH2 R R CH2 R CH2 R OC10H21 OC10H21 OC10H21 Cu2+ Cu(NO3)2 (1:1) c 196 i red. R = c 105 i c 18 Dh 180 (decomp.) lc 96 Dh 132 i * c1 39 c2 52 i g 67 i R = Fig. 1 Induction of columnar mesophases by transition metal ion complexation as observed for CuII complexes of N-alkylated cyclam derivatives (cf.refs. 8 and 17) Synthesis The synthesis of N-acylated polyethyleneimine derivatives by polymer analogous conversion of linear polyethyleneimine has been extensively described elsewhere.22,24 Also, the reduction of N-acylated polyethyleneimines to poly(N-alkylethyleneimines) has been investigated already by Saegusa and Kobayashi.29 In general, the use of lithium aluminium hydride can lead to a decrease in the degree of substitution due to a CMN cleavage in the transition state of the reaction.30 This side reaction had been attributed to a stabilization of the transition state by the lithium counterion,31 and could thus be suppressed by using AlH3 which was freshly prepared from LiAlH4 with concentrated sulfuric acid in THF solution.8,31,32 A CMN cleavage resulting in a degradation of the polymer main chain was not observed.During this work, the polyamides 3 (PEI-3,4) and 4 (PEI-4) could be completely reduced to the polyamines 6 (PA-3,4) and 5 (PA-4), respectively, using a fivefold excess of aluminium hydride (Scheme 1). reaction was followed by IR spectroscopy, monitoring the disappearance of the CNO stretching band at ca. 1640 cm-1 as the reaction proceeded. In addition, 1H NMR spectra of the polyamines in chloroform solution indicated their higher conformational mobility; the signals of the aromatic protons in the side group CH2 CH2 N C R O R C N N N C C N C R R O O O O R OC10H21 OC10H21 R C N HN N C C NH C R R O O O O R 22 3 g 61 fh 120 i R = 2 c 99 (fh 93) i 1 c 108 Dh 154 i which could only be detected as single broad peaks for the Fig. 2 Thermal phase behavior of cyclic and linear oligoamides as polyamides were now split into doublets. Similar behavior has well as linear polymers based on the same N-benzoylated ethyleneimine been observed for the corresponding azacrown derivatives.19 unit (in °C, g=glassy; c=crystalline; wh=hexagonal columnar; i= isotropic) Poly(N-4-decyloxybenzylethyleneimine), 5 (PA-4).LiAlH4 in THF (Aldrich) (1 M; 10 ml ) was stirred under a dry stream of It was supposed that further typical mesomorphic features nitrogen at 0 °C. After dropwise addition of concentrated of substituted macrocyclic oligoamides may be transferred to H2SO4 (0.26 ml) the mixture was stirred for an additional hour the analogous linear polymers.24 In an earlier communication, during which time the evolution of hydrogen could be observed, we have already addressed this point and briefly described the and a white precipitate was formed.This freshly prepared induction of liquid crystalline phases in linear polyamines via AlH3 solution was filtered under argon. Then, 400 mg complexation of transition metal ions.28 Continuing this work, (1.32 mmol) of poly(N-4-decyloxybenzoylethyleneimine), 4 we now present a more detailed discussion of this concept (PEI-4, n=85), in dry THF was added slowly, and the resulting in combination with the first structural studies by X-ray reaction mixture was stirred at room temp.overnight. The diVraction. next day, excess of AlH3 was destroyed by the addition of 3 ml THF–H2O (151 v/v), the solvent was distilled oV, and the white residue dissolved in a 151 mixture of 1 M NaOH and Experimental dichloromethane.The organic layer was washed with water, Materials dried with sodium sulfate, and concentrated by evaporating most of the solvent. The white polymer was precipit- The chemicals and solvents (p.a. grade) used for synthesis were ated in cold acetone.Yield: 265 mg (69%); dH (CDCl3, purchased from Merck and were used as obtained without 200 MHz) 0.87 [t, 3H, CH3(CH2)9O], 1.15–1.5 [br, 14H, further purification unless explicitly stated. THF used for the CH3(CH2)7CH2CH2O], 1.68 [t, 2H, CH3(CH2)7CH2CH2O], polymer analogous reduction of linear polyamides was freshly 2.40 [br, 4H, CH2NCH2], 3.35 [br, 2H, NCH2Ar], 3.81 [t, distilled after refluxing over potassium in a nitrogen atmosphere for several hours. 2H, CH3(CH2)8CH2O], 6.70 (d, 2H, Ar H-3,5), 7.02 (d, 2H, Ar 344 J. Mater. Chem., 1998, 8(2), 343–351Fig. 3 Helical arrangement of the N-acylated polyethyleneimine 3 (PEI-3,4) in a possible energetically favored conformation (for matters of simplification only the atoms of the polymer backbone are shown; thus note that most of the apparent free volume in the center of the column is actually filled with hydrogen atoms) CH2 CH2 N C O OC10H21 CH2 CH2 N CH2 OC10H21 R R CH2 CH2 N CH2 OC10H21 R n n i n ii Cu2+/4 PEI-3,4 3 R = OC10H21 PEI-4 4 R = H PA-4 5 R = H PA-3,4 6 R = OC10H21 PCu-4/NO3 7 R = H, X– = NO3 – PCu-4/OAc 8 R = H, X– = CH3COO– PCu-3,4/OAc 9 R = OC10H21, X– = CH3COO– Scheme 1 Reagents and conditions: i, AlH3, THF, 0 °C, 16 h; ii, copper(II) nitrate (7) or copper(II) acetate (8, 9), THF or CH2Cl2–H2O, room temp., 1–3 d H-2,6); nmax (KBr)/cm-1 2956, 2923, 2852, 2826 (CH2), 1617, (C/N=26.82); degree of substitution calculated from C/N ratio >95%]. 1511, 1464 (CNC), 1246 (CMO) [Calc. for C19H31NO: C, 78.84; H, 10.79; N, 4.84 (C/N=16.29).Found: C, 78.11; H, 10.93; N, 5.07% (C/N=15.41); degree of substitution calculated Preparation of polyamine–copper complexes from C/N ratio: ca. 93%]. The binding of Cu2+ and Ni2+ to branched polyethyleneimine, resulting in the first known polyamine–metal complexes, was Poly[N-3,4-bis(decyloxy)benzylethyleneimine], 6 (PA-3,4). The reaction proceeded in the same way as described for 5 described in 196333 and has recently been reviewed.34 In the work presented here, the complexes were prepared as described (PA-4), starting from 1.0 g (2.2 mmol) of poly[N-3,4-bis- (decyloxy)benzoylethyleneimine], 3 (PEI-3,4, n=85), 15 ml earlier for N-benzylated azamacrocycles,8 by adding a solution of the polyamine in THF or dichloromethane to an excess of (15 mmol) of 1 M LiAlH4-solution in THF and 0.25 ml of concentrated H2SO4. Yield: 620 mg (63%); dH (CDCl3, the metal salt (Scheme 1).The latter was either dissolved in water or used as a solid (which for the cases of nitrates and 200 MHz) 0.85 [t, 6H, CH3(CH2)9O], 1.15–1.5 [br, 28H, CH3(CH2)7CH2CH2O], 1.70 [m, 2H, CH3(CH2)7CH2CH2O], acetates, which are partially soluble in THF, yielded a homogeneous reaction mixture).The resulting mixtures were stirred 2.35 ( br, 4H, CH2NCH2), 3.30 (br, 2H, NCH2Ar), 3.80 [m, 4H, CH3(CH2)8CH2O], 6.5–6.65 (m, 2H, Ar H-2,5), 6.7 (br, for one to several days, although in most cases the complexation was indicated by a dark blueish or greenish color of 1H, Ar H-6); nmax (KBr)/cm-1 2935, 2849 (CH2); 1525, 1469 (CNC); 1260 (CMO) [Calc. for C29H51NO2: C, 78.15; H, 11.53; the organic phase almost instantaneously. After removal of the solvent, the resulting residues were dissolved in dichloro- N, 3.14 (C/N=24.89).Found: C, 76.18; H, 11.10; N, 2.84% J. Mater. Chem., 1998, 8(2), 343–351 345methane and washed with water several times, and filtered observed using an Ortholux II POL-BK (Leitz) microscope. For the photographs of the observed textures Olympus SC 35 through a 0.45 mm Teflon filter to remove traces of the metal salt.After distillation, the solid complexes were dissolved in (type 12) and Kodak-Ektachrome 400 ASA film material was used. DiVerential scanning calorimetry (DSC) measurements benzene, filtered again and finally freeze-dried. As an example, the analytical data of 7 (PCu-4/NO3) were performed with a DSC-7 (Perkin-Elmer) microcalorimeter using an Epson-PC and DSC-7 multitasking software (Perkin- obtained by reaction of polyamine 5 (PA-4) with copper(II) nitrate are as follows: blue–green powder; dH (CDCl3, Elmer) for data evaluation.Calibration standards were indium and lead. Prior to the measurements, 3–10 mg of the substances 200 MHz) 0.87 [t, 3H, CH3(CH2)9O], 1.15–1.50 [br, 14H, CH3(CH2)7CH2CH2O], 1.70 [t, 2H, CH3(CH2)7CH2CH2O], were sealed in aluminium pans.Observed peak maxima are given as the first order transition temperatures; the inflection 2.60–3.40 ( br, 4H, CH2NCH2), 3.25–4.10 [br, 4H, NCH2Ar and CH3(CH2)8CH2O], 6.70 (d, 2H, Ar H-3,5), 7.01 (d, 2H, points for the DSC traces are taken as the glass transitions.X-Ray structural investigations were performed with Cu-Ka- Ar H-2,6); nmax (KBr)/cm-1 2953, 2923, 2850 (CH2), 1612, 1514, 1468 (CNC), 1385 (NMO, nitrate counterion), 1305, radiation (l=0.1541 nm). The diVracted radiation was analyzed using a graphite monochromator and a Siemens X-1000 1252 (CMO), 1163, 1030, 818; l (THF)/nm 665 (e= 40 l mol-1 cm-1) [Found: C, 67.44; H, 9.49; N, 6.30% (C/N= flat plate detector or a flat picture camera at diVerent sample– detector distances.The temperature of the samples was con- 10.71]. Assuming nitrate as the only counterion 21.7% copper( II ) ions per repeating unit were calculated, resulting in an trolled with an accuracy of 0.1 °C by a Linkam THM 600 heating stand. average composition of CuL4.7 (L=amino ligand).§ For the complexation of diVerent central ions (Ni2+, Co2+, UO22+) and the use of other counterions (SO42-, Cl-, BF4-) Results and Discussion during this work, the procedure as described above was used.Complexation of transition metal ions by N-alkylated However, it seems that this approach cannot be generalized: polyethyleneimines using copper chloride, copper sulfate or copper tetrafluoroborate which are not soluble in suitable organic solvents, the The complexation of the polyamines 5 (PA-4) and 6 (PA-3,4) reaction had to be performed in heterogeneous media.Also, was followed using diVerent spectroscopic techniques, as will most complexation attempts with other metal cations (regardbe discussed for the single-chain substituted polyamine–copper less of the counterion) failed so far, both in homogeneous as system 7 (PCu-4/NO3).Fig. 4(a) shows the FT-IR spectra of well as in heterogeneous reaction media. Only for the system the complex and of the pure polyamine ligand 5 (PA-4). As in 5 (PA-4)/Ni(NO3)2 was the preparation of a polymer–metal the case of low molecular mass azacrown derivatives,8 the CHcomplex achieved in THF solution.The optimization of the stretching band at n=2830 cm-1 disappeared, which can be experimental conditions to promote the polymer–metal comrelated to the methylene group next to the amino donor in the plex formation has yet to be investigated.¶ polymer main chain. The influence of the complexation was also seen in the 1H NMR spectrum [Fig. 4(b)]. As compared Characterization to the relatively high mobility of the polyamine ligand, which in that case led to narrower peaks as well as to a doublet 1H NMR spectra were recorded with a 200 MHz FT–NMR splitting of the aromatic signals, the bands were significantly Spectrometer AC-200 (Bruker).Infrared (IR) spectra were broadened for the metal complex. This was particularly pro- recorded from KBr pellets of the materials with an FT-IR nounced for the resonances of the amino methylene protons spectrometer 5DXC (Nicolet).UV–VIS absorption spectra which, in addition, are shifted to lower field. For the pure were measured in quartz cuvettes (Hellma) with a Lambda 5- polyamine ligand 5 (PA-4) they were detected at d=2.4 (main spectrometer (Perkin-Elmer) using UVASOL grade solvents chain) and d=3.4 (side chain), whereas for the complex 7 (Merck).For polarizing microscope investigations, thin films (PCu-4/NO3) two broad signals were found at d=2.6–3.4 and of the compounds prepared on microscopic slides were d=3.4–4.0. Finally, the ligand–metal interaction could be observed by § The authors wish to apologize for an error in the elemental analysis UV–VIS spectroscopy (spectra not shown here).A broad data given for complex 7 (PCu-4/NO3) in ref. 28. By mistake, the shoulder (l=310 nm) of the aromatic absorption band (lmax= calculated data for the ligand 5 (PA-4) instead of the experimental 275 nm) as well as a very weak d–d band of the central atom data for the complex were reported there. This error has been corrected at l=665 nm (e=40 l-1 mol-1 cm-1) were detected. Because here.Also note that, although found useful for the determination of the average copper–ligand ratio in the complexes with nitrate coun- an exact determination of the coordination number of the terion, the same calculation for the compounds obtained with cop- central ions could not be achieved in all cases,§ it should be per(II ) acetate did not yield meaningful values.From the analytical mentioned that several remarks on the complexation of transdata of 8 (PCu-4/OAc) an average composition CuL2 was calculated. ition metal cations with branched and linear polyethyleneimine Despite repeated filtration the complex seemingly still contains traces have been given in the literature.33–35 Based on this, the low of copper(II) acetate.This also resulted in two sharp reflections in the molar extinction coeYcient may be explained by the polymeric wide angle region of the X-ray diVraction pattern as well as in a very weak first order transition at 116 °C in the DSC scans corresponding nature of the ligand resulting in a less defined complexation to the melting point of the copper salt. However, for 9 (PCu-3,4/OAc) of the CuII ions, since in addition to the expected fourwhich evidently does not contain inorganic impurities an even higher coordination, a possible five-coordination has to be taken into value for the copper content was evaluated.Obviously, the basic account as well.34,36,37 assumption for the calculation, accounting for only one counterion species, does not hold for the complexes obtained from copper acetate.Characterization of thermal phase behavior ¶ Interestingly, the extraordinary stability of complexes formed by copper and branched polyethyleneimine even in acidic solution has In analogy to the cyclic low molecular mass derivatives,8 the been reported in this context (cf. B. L. Rivas and K. E. Geckeler, Adv. corresponding polyamines 5 (PA-4) and 6 (PA-3,4) obtained Polym.Sci., 1992, 102, 173), which even served for the separation of Cu2+ from other metal ions (cf. K. E. Geckeler, E. Bayer, G. A. by polymer analogous reduction from the polyamides 4 (PEIVorob’eva and B. Y. Spivakov, Anal. Chim. Acta, 1990, 230, 171). As 4) and 3 (PEI-3,4), respectively, are partially crystalline matein general, metal complexes of branched polyethyleneimine are con- rials that do not form liquid crystalline phases before melting siderably more stable than those of the linear polymer (cf.S. Kobayashi, into an isotropic liquid (Fig. 5). As in the case of the cyclic K. Hiroishi, M. Tokunoh and T. Saegusa, Macromolecules, 1987, 20, oligomers, this can be attributed to the higher conformational 1496), the observed problems for most metal ions may be explained flexibility of the N-alkylated polymers as compared to the and thus also point to the stability of the copper–polyethyleneimine coordination compounds.N-acylated compounds. 346 J. Mater. Chem., 1998, 8(2), 343–351The latter model resembles the proposed structure for the observed hexagonal columnar mesophase of the corresponding linear polyamides,22,24 and would be comparable with the helical crystal structures of metal complexes formed, e.g.by oligopyridine ligands.38,39 On the other hand, a lamellar layered structure would be in agreement with smectic mesophases observed for diVerent organometal complexes of CuII.40 Due to the observed decomposition of the complex nitrate at elevated temperatures, the introduction of the acetate ion was investigated.Although the thermal stability of the obtained compound 8 (PCu-4/OAc) was only slightly improved, this approach proved successful in so far as the simultaneous decrease of the clearing point below the decomposition temperature now allowed the heating of the material to the isotropic state. On first heating under the polarizing microscope (by 10 °C min-1), the complex formed an unspecific texture which slowly cleared at ca. 94°C (last anisotropic domains disappeared at ~105 °C). On decreasing the temperature, the sample showed a considerable supercooling of the clearing point to 72 °C. The growth of typical ba�tonets41 which eventually resulted in the formation of a fan-shaped texture (Fig. 7) at that temperature points to the existence of a smectic mesophase.In addition, the sample could be sheared easily even after cooling down to 40 °C, until at room temperature a marked increase of the viscosity of the material was observed. The behavior was reversible in the following heating and cooling cycles. The copper complex 9 (PCu-3,4/OAc) obtained from the double side chain polyamine 6 (PA-3,4) and copper(II) acetate was obtained as a highly viscous and sticky material.Under the polarizing microscope, a Schlieren texture indicating a nematic mesophase was observed [Fig. 8(a)] at room temperature. On cooling from the isotropic liquid phase (T ca. 75°C), typical ‘Maltese crosses’ were formed within an almost completely homotropically oriented sample [Fig. 8(b)]. In DSC measurements, 8 (PCu-4/OAc) showed a glass transition at ca. 10°C and two first order transitions at 67 and 98 °C [in addition, a very weak signal at 116 °C was detected Fig. 4 Complexation of copper(II) ions by N-alkylated polyethylene- which can be related to residual traces of copper(II) acetate]. imines as follow by spectroscopic investigations on 7 (PCu-4/NO3); Whereas the latter corresponds to the microscopically deter- (a) 1H NMR spectrum (200 MHz, CDCl3, room temp.); (b) FT-IR mined transition to the isotropic melt, the nature of the first spectrum (KBr pellet) transition could not be completely clarified.As the material could be easily sheared in the microscopic investigations at 40 °C, the first peak may be related to a transition between Based on the results for the low molecular mass derivatives (Fig. 1) where a columnar mesophase was found for complexes two liquid crystalline mesophases. On cooling however, only one peak at 70 °C was detected next to the glass transition. formed from the macrocyclic ligand bearing a single-chained 4-alkoxybenzoyl substituent and copper nitrate,8 the investi- Considering the higher cooling rate used in the DSC measurements (10 °C min-1), this is in good agreement with the gation of the polymeric complex 7 (PCu-4/NO3) appeared most promising.Indeed, thermotropic liquid crystalline observations under the polarizing microscope, where a supercooling of the clearing point and the growth of a fan-shaped behavior was observed for this complex as described in an earlier communication.28 Under the polarizing microscope, a texture at ca. 72°C were seen. Thus, the anisotropic domains observed on heating above 72 °C could also be related to a rather unspecific, but shearable texture was observed above the determined glass transition at 53 °C. However, decompo- biphasic behavior of the material above that temperature. The curves of 9 (PCu-3,4/OAc) showed a first order melting trans- sition of the material was observed above 160 °C, and the observed texture completely disappeared at 173 °C.A more ition in the same temperature region in which the glass transition of 8 (PCu-4/OAc) was detected. The clearing tem- detailed analysis of the phase behavior by microscopic investigation was thus not possible. perature of 68 °C agreed with the microscopic observations.The transition temperatures of the copper(II ) complexes 7 By X-ray diVraction, the existence of a thermotropic liquid crystalline phase for 7 (PCu-4/NO3) now could be ultimately (PCu-4/NO3), 8 (PCu-4/OAc), and 9 (PCu-3,4/OAc) as determined from DSC are listed in Table 1. proven. However, the exact determination of the phase type still remains unclear. In the diVraction pattern, next to a broad X-Ray investigations on a powder sample of 8 (PCu-4/OAc) cooled from the isotropic phase to room temperature proved halo at ~4.6 A ° which corresponds to the disordered alkyl side chains, only a first order Bragg peak was found which allowed the liquid crystalline character of the material.Next to the first order Bragg reflection with a corresponding lattice param- calculation of the lattice parameter as d001 ~35.5 A ° .For this material, the absence of mixed or higher order reflections still eter d001=34.9 A ° , two very weak signals were detected in the low angle region of the flat camera picture [Fig. 9(b)]. Their prevents the experimental distinction between a lamellar mesophase structure [Fig. 6(b)] and a cylindrical complex structure, corresponding spacings of 17.5 A ° (ca.d001/2) and ca. 11.5 A ° (ca. d001/3) identify them as the (002) and (003) reflections of as resulting for instance from a helical arrangement of the polyamine main chain around the central copper ions a lamellar structure in which the copper ions are complexed by the amino head groups of polyamine bilayers. In the wide [Fig. 6(c)]. J. Mater. Chem., 1998, 8(2), 343–351 347Fig. 5 Thermotropic phase behavior of N-alkylated polyethyleneimines 5 (PA-4) and 6 (PA-3,4) (in °C, taken from second DSC heating scan, scan rate: 10 °C min-1); g=glassy; c=crystalline; i=isotropic. * The sample could be sheared under the polarizing microscope below the detected first order transition, proving the semicrystalline nature of the polymer, so that a glass transition has to be assumed below 40 °C (not unambiguously detected in the DSC scans).Fig. 6 Schematic representation of possible structures of the polyamine–copper complex 7 (PCu-4/NO3) within the thermotropic mesophase; (a) layered structure; (b) columnar structure. angle region [Fig. 9(a)], a broad halo of the alkyl chains with diVractogram of the polyamine–copper complex 9 (PCu- 3,4/OAc) measured at room temperature [Fig. 10(a)] exhibited a maximum at ca. 4.8 A ° was found. In addition, 8 (PCu- 4/OAc) showed two additional sharp reflections at 2h ca. 29° a rather broad, but pronounced Bragg peak in the small angle region at 2h ca. 2.7° from which a spacing of ca. 32.8 A° was and ca. 44° (ca. 3.0 and ca. 2.0 A° , respectively) which can be attributed to residual traces of copper acetate (compare dis- calculated; mixed or higher order reflections were not found.In the flat camera picture shown in Fig. 10(b), next to the cussion of DSC results). Contrary to expectations which were based on the nematic Bragg peak (from which a slightly higher value of ca. 34A ° was determined) and the halo related to the alkyl chains at texture observed in polarizing microscopy, the X-ray 348 J.Mater. Chem., 1998, 8(2), 343–351Fig. 7 Microscopic texture of 8 (PCu-4/OAc) after cooling to 70 °C from the isotropic phase with 1 °C min-1 Fig. 9 X-Ray flat camera picture of 8 (PCu-4/OAc) after cooling from the isotropic phase to the liquid crystalline state: (a) complete picture; (b) low angle region 4.4 A ° (halo 2), an additional weak halo at ca. 7.3 A ° (halo 1) was measured. This is particularly interesting as this distance coincides with the proposed pitch height of a helically wound polymer main chain in the model shown in Fig. 2 for the mesomorphic linear polyamide 3 (PEI-3,4). Based on the X-ray data, a layered structure also seems possible for 9 (PCu-3,4/OAc). However, considering the comparably high width of the Bragg peak and the microscopic textures observed, the results point to a nematic columnar phase.This assumption appears reasonable since in comparison with the complexes 7 (PCu-4/NO3) and 8 (PCu-4/OAc) of the polyamine ligand 5 (PA-4) which bear only one side chain per Fig. 8 Microscopic textures of 9 (PCu-3,4/OAc): (a) room temperature, polymer repeating unit, and which both form layered structures directly after preparation; (b) after cooling to 42 °C from the isotropic [Fig. 11(a)], a higher number of side chains have to be phase with 1 °C min-1 incorporated in the aggregates of the double chain derivative Table 1 Phase transition temperatures of copper(II)-polyamine complexes 7(PCu-4/NO3), 8(PCu-4/OAc), and 9(PCu-3,4/OAc) given in °C (taken from the second DSC heating scan measured with a scan rate of 10 °C min-1, for 8(PCu-4/OAc) the thermal transitions in the first cooling curve are also shown): g, glassy; c, crystalline; SX, unidentified smectic/lamellar structure; SA, smectic A/lamellar; NC, nematic columnar; i, isotropic DSC scan T / °C 7 from 5 (PA-4) and Cu(NO3)2 2nd heating g 53 lc 173 (decomp.) 8 from 5 (PA-4) and Cu(OAc)2 2nd heating g 11 c/SX 67 SA 98 i 1st cooling g 10 SA 70 i 9 from 6 (PA-3,4) and Cu(OAc)2 2nd heating c 12 NC 68 i J.Mater. Chem., 1998, 8(2), 343–351 349be wound helically around the central copper ions. On the other hand, more than one polymer chain may constitute the columnar packing which may even result in a crosslinked structure, in which single polymer chains may participate in more than one column.However, based on the observed low viscosity of the material within the mesophase the latter model appears least likely. On the other hand, the packing of several polymer chains in the second model would aVord a much higher order than a single more or less helically folded chain. In addition, although in each model the additional halo found at 7.3 A ° in the flat camera picture could correspond to an average distance of copper atoms in the center of complex sub-units along such columns, the obvious similarity of this distance with the pitch height in the helical model discussed for N-substituted linear polyethyleneimines (Fig. 3) lets us favor the first structural model. Conclusions By polymer analogous reduction of the benzamide group in linear polyamides, N-alkylated polyethyleneimines were obtained which, due to their high conformational mobility cannot form liquid crystalline mesophases.As in the case of azamacrocycles, a mesophase is formed by the materials after the coordination of the polyamine to CuII. Whereas the discoid molecular structure of cyclic oligomers only allows discotic phases, based on the higher flexibility of the polyamine main chain, the polymeric complexes may also adopt layered structures.In addition, even for a higher side chain density for which a loss of mesomorphic features was observed in the case of macrocyclic complexes, now a presumably nematic columnar Fig. 10 X-Ray diVraction patterns of 9 (PCu-3,4/OAc) at room temmesophase is found.Although detailed information about the perature: (a) goniometer scan; (b) flat camera picture conformation of the polymeric ligands in the liquid crystalline complexes has yet to be elucidated, the general principle of mesophase induction by metal ion complexation could be proven for the materials discussed. The results described again show the analogy between low molecular mass cyclic derivatives and linear polymers composed of the same basic monomer unit and are therefore of general interest in polymer liquid crystal research.In addition, based on their complex nature, numerous interesting possibilities for tailoring the physical properties of mesomorphous polymer systems appear. For instance, by simple variation of structural features such as the metal ion, the side chain density or by the introduction of chiral centers into the polyamine ligand, physical parameters such as thermotropic phase behavior and also the optical properties of the materials could be tuned.References 1 (a) J.-M. Lehn, Angew. Chem., Int. Ed. Engl., 1988, 27, 89; (b) J.-M. Lehn, Angew. Chem., Int. Ed. Engl., 1990, 29, 1304; (c) J.-M.Lehn, Supramolecular Chemistry, VCH Verlagsgesellschaft, Weinheim, 1995. 2 F.Vo� gtle, Supramolecular Chemistry, Wiley, New York, 1991. 3 E. C. Constable, Nature, 1990, 346, 314. 4 M.-J. Brienne, J. Gabard, J.-M. Lehn and I. Stibor, J. Chem. Soc., Chem. Commun., 1989, 1868. Fig. 11 Schematic representation of the possible structures of polyam- 5 (a) T. Kato and J. M.J. Fre� chet, J. Am. Chem. Soc., 1989, 111, 8533; ine–copper(II ) complexes within their thermotropic mesophases: (a) (b) T. Kato and J. M. J. Fre� chet, Macromolecules, 1989, 22, 3818. smectic mesophase of 8 (PCu-4/OAc); (b) nematic columnar mesophase 6 (a) G. Lattermann and G. Staufer, L iq. Cryst., 1989, 4, 347; (b) of 9 (PCu-3,4/OAc) M. Ebert, R. Kleppinger, M. Soliman, M.Wolf, J. H. WendorV, G. Lattermann and G. Staufer, L iq. Cryst., 1990, 7, 553. 9 (PCu-3,4/OAc). This is in agreement with the findings 7 W. Paulus, H. Ringsdorf, S. Diele and G. Pelzl, L iq. Cryst., 1991, for the corresponding N-acylated polyethyleneimines where 9, 807. 8 A. Liebmann, C. Mertesdorf, T. Plesnivy, H. Ringsdorf and lamellar crystals or hexagonal columnar mesophases were J.H. WendorV, Angew. Chem., Int. Ed. Engl., 1991, 30, 1375. observed for mono- or bis-decyloxysubstituted side groups, 9 C.M. Paleos and D. Tsiourvas, Angew. Chem., Int. Ed. Engl., 1995, respectively.24 34, 1696. As has been discussed before for the original mesomorphic 10 H. Ringsdorf, R. Wu� stefeld, E. Zerta, M. Ebert and J. H. WendorV, polyamides,24 the formation of the complex columnar aggre- Angew.Chem., Int. Ed. Engl., 1989, 28, 914. gates may be explained by several diVerent structural models 11 (a) V. Percec, G. Johansson, J. Heck, G. Ungar and S. V. Batty, J. Chem. Soc., Perkin T rans. 1, 1993, 1411; (b) V. Percec, [Fig. 11(b)]. For instance, a single polymer main chain may 350 J. Mater. Chem., 1998, 8(2), 343–351G. Johansson, D. Schlueter, J.C. Ronda and G. Ungar,Macromol. 30 V. M. Micovic and M. L. Mihailovic, J. Org. Chem., 1953, 18, 1190. 31 N. M. Yoon and H. C. Brown, J. Am. Chem. Soc., 1968, 90, 2927. Symp., 1996, 101, 43. 12 J. L. M. van Nunen, R. S. A. Stevens, S. J. Picken and 32 T. Saegusa, S. Kobayashi and M. Ishiguro, Macromolecules, 1974, 7, 958. R. J. M. Nolte, J. Am. Chem. Soc., 1994, 116, 8825. 13 (a) D. A. House, in Comprehensive Coordination Chemistry, ed. 33 H. Thiele and K.-H. Gronau,Makromol. Chem., 1963, 59, 207. 34 (a) A. von Zelewsky, L. Barbosa and C. W. Schla�pfer, Coord. Chem. G. Wilkinson, Pergamon Press, Oxford, 1987, p. 23; (b) F. A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry, Rev., 1993, 123, 229; (b) B. L. Rivas and K. E. Geckeler, Adv. Polym. Sci., 1992, 102, 173.Wiley, New York, 1988. 14 R. D. Hancock and A. E. Martell, Chem. Rev., 1989, 89, 1875. 35 S. Kobayashi, K. Hiroishi, M. Tokunoh and T. Saegusa, Macromolecules, 1987, 20, 1496. 15 J. M. Lehn, J. Malthe�te and A. M. Levelut, J. Chem. Soc., Chem. Commun., 1985, 1794. 36 U. Stebani, G. Lattermann, M. Wittenberg and J. H. WendorV, Angew. Chem., Int. Ed. Engl., 1996, 35, 1858. 16 G. Lattermann, L iq. Cryst., 1989, 6, 619. 17 S. Bauer, T. Plesnivy, H. Ringsdorf and P. Schuhmacher, 37 (a) J. Arago� , A. Bencini, A. Bianchi, E. Garcia-Espan� a, M. Micheloni, P. Paoletti, J. A. Ramirez and P. Paoli, Inorg. Chem., Makromol. Chem., Macromol. Symp., 1992, 64, 19. 18 C. Mertesdorf and H. Ringsdorf, Molecular Engineering, 1992, 2, 1991, 30, 1843; (b) B. J. Hathaway and A.A. G. Tomlinson, Coord. Chem. Rev., 1970, 5, 1. 189. 19 D. Tatarsky, K. Banerjee and W. T. Ford, Chem. Mater., 1990, 38 (a) J. M. Lehn, A. M. Rigault, J. Siegel, J. Harrowfield, B. Chevrier and D. Moras, Proc. Natl. Acad. Sci. USA, 1987, 84, 2565; (b) 2, 138. 20 G. Lattermann, S. Schmidt, R. Kleppinger and J. H. WendorV, J.-M. Lehn and A. Rigault, Angew. Chem., 1988, 100, 1121; (c) U. Koert, M. M. Harding and J.-M. Lehn, Nature, 1990, 346, 339; Adv. Mater., 1992, 4, 30. 21 G. Lattermann, S. Schmidt and B. Gallot, J. Chem. Soc., Chem. (d) E. C. Constable, S. M. Elder, J. Healy, M. D. Ward and D. A. Tocher, J. Am. Chem. Soc., 1990, 112, 4590. Commun., 1992, 1091. 22 H. Fischer, S. S. Ghosh, P. A. Heiney, N. C. Maliszewskyj, 39 (a) A. F. Williams, C. Piguet and G. Bernardinelli, Angew. Chem., Int. Ed. Engl., 1991, 30, 1490; (b) C. Piguet, G. Bernardinelli, T. Plesnivy, H. Ringsdorf and M. Seitz, Angew. Chem., Int. Ed. B. Bocquet, A. Quattropani and A. F. Williams, J. Am. Chem., Soc., Engl., 1995, 34, 795. 1992, 114, 7440; (c) A. Williams, Chem. Eur. J., 1997, 3, 15. 23 U. Stebani and G. Lattermann,Macromol. Rep., 1995, A32, 385. 40 (a) C. Carfagna, U. Caruso, A. Roviello and A. Sirigu, Makromol. 24 M. Seitz, T. Plesnivy, K. Schimossek, M. Edelmann, H. Ringsdorf, Chem., Rapid Commun., 1987, 8, 345; (b) U. Caruso, A. Roviello H. Fischer, H. Uyama and S. Kobayashi, Macromolecules, 1996, and A. Sirigu, Macromolecules, 1991, 24, 2606; (c) K. Hanabusa, 29, 6560. J.-I. Higashi, T. Koyama, H. Shirai, N. Hojo and A. Kurose, 25 U. Stebani and G. Lattermann, Adv. Mater., 1995, 7, 578. Makromol. Chem., 1989, 190, 1; (d) L. Oriol and J. L. Serrano, Adv. 26 U. Stebani, G. Lattermann, M. Wittenberg and J. H. WendorV, Mater., 1995, 7, 348. J.Mater. Chem., 1997, 7, 607. 41 (a) D. Demus and L. Richter, T extures of L iquid Crystals, VEB 27 J. H. Cameron, A. Facher, G. Lattermann and S. Diele, Adv. Deutscher Verlag fu� r GrundstoYndustrie, Leipzig, 1978; (b) Mater., 1997, 9, 398. G. W. Gray and J. W. G. Goodby, Smectic L iquid Crystals: 28 H. Fischer, T. Plesnivy, H. Ringsdorf and M. Seitz, J. Chem. Soc., T extures and Structures, Leonard Hill, Glasgow, 1984. Chem. Commun., 1995, 1615. 29 T. Saegusa, A. Yamada, H. Taoda and S. Kobayashi, Macromolecules, 1978, 11, 435. Paper 7/05592G; Received 1st August, 1997 J. Mater. Chem., 1998, 8(2),