摘要:

J O U R N A L O F C H E M I S T R Y Materials Feature Article Heterosupramolecular chemistry Xavier Marguerettaz, Alan Merrins and Donald Fitzmaurice* Department of Chemistry, University College Dublin, Dublin 4, Ireland Received 21st April 1998, Accepted 24th June 1998 The covalent and non-covalent assembly of nanocrystals largely persist and there exists an associated heterosupramolecand molecules in solution to yield heterosupermolecules ular function.possessing well defined heterosupramolecular functions is Scheme 1 shows two examples of heterosupermolecules, described. Also described is the organisation of heterosup- consisting of a covalently and non-covalently assembled conermolecules into assemblies possessing addressable hetero- densed phase (TiO2 nanocrystal electron donor) and molecular supramolecular function.Strategies are considered that (viologen, electron acceptor) component,4,5 prepared in this will permit the covalent and non-covalent assembly and laboratory. organisation of a wide range of condensed phase and The covalent and non-covalent assembly in solution of molecular components. Also considered are the possible heterosupermolecules possessing well defined heterosupramolelonger term benefits of the development of a systematic cular functions is reviewed.Also reviewed is the covalent and chemistry of both condensed phase and molecular non-covalent organisation of heterosupermolecules in solution components, that is, a systematic heterosupramolecular to yield assemblies possessing addressable heterosupramolecu- chemistry.lar function. Strategies are considered that will permit the covalent and non-covalent assembly and organisation of a 1 Introduction wide range of condensed phase and molecular components. Also considered are the possible long term benefits of Conventionally, a supermolecule is distinguished from a large developing a systematic chemistry of condensed phase and molecule as follows:1 first, the molecular components of a molecular components, i.e.a systematic heterosupramolecular supermolecule are non-covalently linked; secondly, the intrinsic chemistry. properties of these molecular components largely persist; and thirdly, the properties of a supermolecule are not a simple superposition of the properties of the constituent molecular 2 Covalent heterosupermolecules components, i.e.there exists a well defined supramolecular function. The recent past has seen the synthesis of eYcient sensitisers for use in regenerative photoelectrochemical cells.6 Examples With the widespread application of supramolecular concepts throughout chemistry, biology and physics, however, has come include ruthenium complexes containing bipyridine ligands derivatised by addition of carboxylic acid groups which are the need for a more inclusive definition.2 Consequently, the term supermolecule is now also applied to covalently linked chemisorbed at the surface of the constituent nanocrystals of a nanostructured TiO2 photoanode.It has been proposed that molecular components provided, as above, the properties of these constituent components largely persist and there exists a ligands containing such groups displace less basic solvent molecules and chelate Ti4+ sites at the surface of a TiO2 supramolecular function.A more inclusive definition still has been adopted in nanocrystal.7 In this context, Moser and coworkers have studied the discussing recent work directed toward the development of a chemistry of covalently and non-covalently assembled con- chemisorption of a series of model compounds at the surface of the constituent nanocrystals of a nanostructured TiO2 film.7e densed phase and molecular components.3 By analogy with a supermolecule, the properties of the constituent condensed They found salicylate is strongly adsorbed at a single Ti4+ site and is oriented normal to the substrate surface.phase and molecular components of a heterosupermolecule Scheme 1 A covalently and non-covalently assembled heterosupermolecule. J. Mater. Chem., 1998, 8(10), 2157–2164 2157The TiO2 nanocrystals of the heterosupramolecular assembly in Scheme 3 are in ohmic contact with each other and with the conducting support.16 At suYciently negative applied potentials, therefore, electrons occupy the available conduction band states of the nanostructured TiO2 film and are transferred to the viologen molecules covalently linked to the surface of the nanocrystals (Scheme 4).4,15 Under open circuit conditions, bandgap excitation of the heterosupramolecular assembly in Scheme III leads to formation of electron –hole pairs in the nanostructured TiO2 film and, in the presence of a suitable hole scavenger, to the photogenerated conduction band electrons being transferred to the viologen molecules covalently linked to the surface of the nanocrystals (Scheme 4).4,15 The associated heterosupramolecular functions are potential and light-induced electron transfer respectively.Since electron transfer may be initiated either by applying a suYciently negative potential or by bandgap excitation, the following question arises: can the eVects of bandgap excitation be modulated potentiostatically? As stated above, under open circuit conditions, bandgap excitation is followed by electron transfer from a TiO2 nanocrystallite to a viologen (ON state).At suYciently positive applied potentials, however, no electron transfer is observed following bandgap excitation (OFF state).This modulation eVect, see Scheme 2 Light-induced electron transfer in covalently assembled Scheme 5, is attributed to the fact that at positive applied heterosupermolecules. potentials the photogenerated conduction band electrons reduce vacant trap states in a nanocrystal rather than a covalently attached viologen. This finding points to two general advantages of Chemisorption of this molecule is also accompanied by development of a visible charge transfer absorption band.8 It was heterosupramolecular assemblies: first, that the function of the constituent heterosupermolecules of an assembly may be concluded that salicylate could be used to covalently assemble a TiO2 nanocrystal and a viologen molecule in solution.modulated if a bulk property of one or more of the condensed phase components may be modulated, and secondly, that if Accordingly, salicylate–viologen molecules were synthesised and chemisorbed at the surface of the TiO2 nanocrystals in a the property of the condensed phase component that is being modulated can be monitored, the modulation state of each stable aqueous or ethanolic colloidal dispersion at pH 3.0.4,9 As salicylate is adsorbed normal to the crystallite surface at a heterosupermolecule may be inferred.Finally, we note the following limitations of the above single Ti4+ site,7e and as dications are not adsorbed at the positively charged surface of the TiO2 nanocrystals at pH 3.0,10 heterosupramolecular assembly: first, it is not an organised heterosupramolecular assembly, as a consequence of which, it may be assumed that the viologen molecule is oriented normal to the nanocrystal surface as shown in Scheme 2.A the constituent heterosupermolecules are not individually addressable; secondly, the constituent heterosupermolecules series of such viologens, with more or less electron withdrawing moieties (MR), have been prepared.9,11 do not act fully independently, i.e.there is electron transfer between viologens adsorbed at the same or adjacent nanocrys- Recently, it has been found that ruthenium based complexes containing bipyridine ligands derivatised by addition of phos- tals;15 and thirdly, because the constituent condensed phase and molecular components are covalently linked, they may phonic acid groups are even more strongly chemisorbed at TiO2.12 Chemisorption is, as above, discussed in terms of the not be self-assembled.displacement of less basic solvent molecules and chelation of surface Ti4+ sites. Accordingly, the phosphonate–viologen molecules shown in Scheme 2 have also been synthesised.9 4 Non-covalent heterosupermolecules Bandgap excitation of these covalently assembled The covalent assembly of a TiO2 nanocrystal and a viologen heterosupermolecules results in electron transfer from the TiO2 molecule has been described in Section 2.An alternative nanocrystal (electron donor) to the viologen (electron approach is to non-covalently assemble these components.5 acceptor), i.e. light-induced vectorial electron transfer.4 As the Toward this end, TiO2 nanocrystals were prepared in the first reduction potential of the viologen may be varied systempresence of the modified stabiliser shown in Scheme 6.5,17 This atically by appropriate choice of a more or less electron stabiliser, whose synthesis has been described in detail else- withdrawing moiety MR, the rate of electron transfer may be where,9,18 incorporates a diamidopyridine moiety which can determined as a function of the associated change in free recognise and selectively bind, by complementary hydrogen energy change.13 Although similar studies have previously bonding, an uracil moiety.19 been reported,14 an unique advantage of the approach outlined It is possible to state that the above stabiliser is physisorbed above is that the separation and relative orientation of the at the surface of a nanocrystal as the resulting dispersion, TiO2 nanocrystal and viologen molecule is known.which otherwise flocculates on the time-scale of seconds, is stable on the time-scale of months. More quantitatively, 1H 3 Covalent heterosupramolecular assemblies NMR studies show that the methylene and methyl groups of this molecule interact with the surface of the TiO2 nanocrystal The covalent assembly of a TiO2 nanocrystal and a viologen at which it is adsorbed.molecule has been described. In a development of this Also shown in Scheme 6 is a viologen molecule incorporating approach salicylate–viologens have been chemisorbed, as an uracil moiety. This is one of a series of compounds of the shown in Scheme 3, at the surface of one of the constituent same general formula which have been prepared.5,9 Having TiO2 nanocrystals of a 4 mm thick nanostructured film supported on conducting glass.4,15 studied this series, it is possible to conclude the following: if 2158 J.Mater. Chem., 1998, 8(10), 2157–2164Scheme 3 A covalent heterosupramolecular assembly.Scheme 4 Potential and light-induced electron transfer in a covalent heterosupramolecular assembly. J. Mater. Chem., 1998, 8(10), 2157–2164 2159Scheme 5 Potential modulation of light-induced electron transfer in a covalent heterosupramolecular assembly. Scheme 6 Light-induced electron transfer in a non-covalently assembled heterosupermolecule. the two alkane chains contain much less than 25 methylene On mixing, these nanocrystals recognise and selectively bind each other as shown in Scheme 7.5 Further, these nanocrystals groups, the viologen molecule is not soluble in chloroform.If, however, the viologen molecule contains much more than self-organise to form an extended array in solution.5 Evidence for ordering of the nanocrystals in these arrays has been about 25 methylene groups, micellisation is observed in chloroform at the concentrations necessary to enable characterisation obtained.of the resulting heterosupermolecule by 1H NMR and IR. On mixing a colloidal TiO2 nanocrystal dispersion prepared 5 Non-covalent heterosupramolecular assemblies in the presence of the stabiliser incorporating a diamidopyridine moiety with a solution of the viologen molecule incorporat- A covalent heterosupramolecular assembly of TiO2 nanocrystals and viologen molecules has been described in ing an uracil moiety, the former recognises and selectively binds the latter as shown in Scheme 6.5 Bandgap excitation of Section 3.In a development of these studies a non-covalent organised assembly of TiO2 nanocrystals and viologen mol- a TiO2 nanocrystal is followed by electron transfer to the noncovalently bound viologen molecule.5 The associated hetero- ecules has been prepared.The phosphonic acid in Scheme 8 was synthesised.9 The supramolecular function is, therefore, light-induced vectorial electron transfer. long alkane chain ensures this molecule is suYciently hydrophobic to be deposited as a close-packed monolayer using Briefly, these studies have been extended to permit the non-covalent assembly of two TiO2 nanocrystals. As before, Langmuir–Blodgett (LB) techniques.The phosphonic acid head group ensures irreversible attachment of the deposited TiO2 nanocrystals were prepared in the presence of the modi- fied stabiliser incorporating a diamidopyridine moiety.5,17 monolayer to the constituent nanocrystals of a TiO2 substrate, 12 see below.Also synthesised was the viologen shown However, TiO2 nanocrystals were also prepared in the presence of the modified stabiliser incorporating an uracil moiety.5,17 in Scheme 8.9 This molecule contains two alkane chains each 2160 J. Mater. Chem., 1998, 8(10), 2157–2164to that nanocrystal, and that electron transfer between adjacent viologens is not significant. Each TiO2 nanocrystal of the organised heterosupramolecular assembly in Scheme 8 is in ohmic contact with the conducting support.16 Upon application of a suYciently negative potential, therefore, electrons occupy the available conduction band states of a nanocrystal and are transferred to the viologen molecule non-covalently linked to the surface of the same nanocrystal, see Scheme 9.4,15 Under open circuit conditions, bandgap excitation of the organised heterosupramolecular assembly in Scheme 8 leads to formation of electron–hole pairs in a nanocrystal and, in the presence of a suitable hole scavenger, to photogenerated conduction band electrons being transferred to the viologen molecules non-covalently linked to the surface of the nanocrystal, see also Scheme 9.4,15 The associated heterosupramolecular functions are potential and light-induced electron transfer, respectively.Since electron transfer may be initiated either by applying a suYciently negative potential or by bandgap excitation, the following question again arises: can the eVects of bandgap excitation be modulated potentiostatically? As stated above, under open circuit conditions bandgap excitation is followed by electron transfer from a TiO2 nanocrystallite to a viologen (ON state).At suYciently positive applied potentials, however, no electron transfer is observed following bandgap excitation (OFF state). This modulation eVect, see Scheme 10, is again attributed to the fact that at positive applied potentials the photogenerated conduction band electrons reduce vacant trap states in a nanocrystal rather than a covalently attached viologen. Finally, we note the following limitations of the covalent heterosupramolecular assembly have been addressed: first, an organised heterosupramolecular assembly has been prepared, as a consequence of which the constituent heterosupermolecules are individually addressable; secondly, the constituent heterosupermolecules act fully independently, i.e.there is no electron transfer between viologens adsorbed at adjacent nanocrystals; 15 and thirdly, because the constituent condensed Scheme 7 Non-covalent assembly of an organised nanocrystal phase and molecular components are non-covalently linked, assembly.they may be self-assembled. equal in length to that of the phosphonic acid, ensuring that this molecule is also suYciently hydrophobic that it may be 6 Current and future studies deposited as a close-packed monolayer using LB techniques. To prepare the heterosupramolecular assembly shown in Current and future work in this area will be directed toward the preparation and synthesis of a wider range of condensed Scheme 8, a close-packed monolayer of TiO2 nanocrystals is first deposited, using LB techniques, on a conducting glass phase and molecular components and their incorporation into heterosupermolecules and organised assemblies of heterosuper- substrate.A close-packed mixed monolayer of the phosphonic acid and the viologen (ratio of 1251) is then deposited, also molecules possessing novel and diverse functions.Outlined below are two examples of work ongoing in the laboratory. using LB techniques, on the nanocrystal monolayer. Detailed characterisation confirms the deposited molecular monolayer In the first example, the stabilisers incorporating a diamidopyridine and an uracil moiety and incorporating at least one has the structure shown.20 On this basis it is possible to assert that there is a single viologen associated with each nanocrystal, thiol group have been synthesised.9 These stabilisers are strongly chemisorbed at the surface of a silver or gold nanocry- that each viologen has a well defined orientation with respect Scheme 8 A non-covalently organised heterosupramolecular assembly.J.Mater. Chem., 1998, 8(10), 2157–2164 2161Scheme 9 Potential and light-induced electron transfer in a non-covalent organised heterosupramolecular assembly. stals prepared in their presence.21 These nanocrystals recognise of stabilisers between Au and Ag nanocrystals is not possible, and it is predicted that the resulting nanocrystal assemblies and selectively bind each other to form a heterosupermolecule consisting of two non-covalently assembled condensed phase will have Au and Ag nanocrystals in alternating lattice sites.22 The second example relates to the non-covalent assembly of components, see Scheme 11.Because the above stabilisers are chemisorbed at the surface of a given nanocrystal, exchange heterosupermolecules in a wide range of polar solvents.Toward Scheme 10 Potential modulation of light-induced electron transfer in a non-covalent organised heterosupramolecular assembly. 2162 J. Mater. Chem., 1998, 8(10), 2157–2164phase and molecular components whose intrinsic properties largely persist but which possess well defined functions. This is because the use of both condensed phase and molecular components oVers three important advantages over the use, as characterised by conventional solid state chemistry, of only condensed phase components or, as characterised by conventional supramolecular chemistry, of only molecular components.First, the use of both condensed phase and molecular components will permit the assembly and organisation of nanometer-scale structures possessing novel functions that could not be achieved if only condensed phase or molecular components are used.In this context, recent advances in the preparation of nanocrystallites of a wide range of materials possessing well defined sizes, surface properties and crystal structures have been important.6,27 Also of importance have been the development of strategies for linking molecules, typically capping groups or sensitiser molecules, to the surface of these nanocrystallites.6,27 Further, conventional supramolecular chemistry continues to provide an ever increasing number of receptor–substrate pairs that can be used to selfassemble and self-organise the condensed phase and molecular components of a heterosupermolecule or heterosupramolecular assembly.28 Secondly, as stated above, progress toward realisation of practical nanometer-scale devices based on organised assemblies of supermolecules has been slow.A reason for this has been the diYculties encountered in identifying substrates capable of modulating supramolecular function and providing information concerning modulation state of the assembly. The incorporation of condensed phase components in nanometerscale devices, however, oVers general advantages in this respect.Specifically, as the function of a heterosupermolecule in an assembly is dependent on the properties of its constituent condensed phase components, modulation of a bulk property of condensed phase components will, of necessity, modulate the function of the heterosupermolecules in the assembly. Scheme 11 Current and future studies in heterosupramolecular chemistry.Importantly, in this respect, modulation of the property of a nanometer-scale condensed phase component is increasingly a realisable goal.27 Further, if the property of the condensed phase component which is being modulated can be monitored, this end, a series of receptor-substrate pairs known to associate then the modulation state of each heterosupermolecule may in polar solvents are being investigated (Scheme 11).The first, be inferred. based on a crown ether–ammonium cation receptor–substrate Thirdly, the use of both condensed phase and molecular pair may be used to self assemble heterosupermolecules in components will permit heterosupermolecules and heterosup- moderately polar solvents.23 Specifically, this recognition–subramolecular assemblies possessing novel architectures to be strate pair will be used to assemble Au and Ag nanocrystals self-assembled and self-organised in solution.That this will be in acetonitrile.24 The second, based on non self-complementary necessary is increasingly apparent. For example, Gra�tzel and DNA oligomers may be used to assemble heterosupermolecules coworkers have described a stable and eYcient regenerative in polar solvents.25 Specifically, this recognition-substrate pair photoelectrochemical cell for the conversion of solar energy will be used of assemble Au and Ag nanocrystals in water.26 to electrical energy.6 Innovative aspects of the Gra�tzel cell include the use of ruthenium complexes as sensitisers whose 7 Why a systematic heterosupramolecular absorption spectra overlap well the solar emission spectrum chemistry? and the use of 10 mm thick nanoporous nanocrystalline semiconductor films with a surface roughness of>1000 as photoan- The self-assembly and self-organisation of complex nanometerodes. Central to the eYcient operation of the Gra�tzel cell is scale structures in solution is an important objective of matethe fact that the sensitiser molecules, as a consequence of their rials chemistry and physics.The importance of this objective being adsorbed directly at the photoanode, are eVectively is a result of the desire to be able to programme the bottomstacked and the probability of an incident visible photon being up assembly in solution of materials with ever more precisely absorbed is close to unity.It is important to note that while defined chemical and physical properties. In the longer term, the above light-harvesting strategy is clearly based on that of it is expected that it will be possible to assemble nanometergreen plants, its practical implementation utilises a heterosup- scale devices in solution and to organise these devices into ramolecular assembly and that implementation of the same addressable arrays.strategy using only condensed phase or molecular components It must be accepted, however, that progress in respect of has not proved possible. the latter has, to date, been very limited for approaches based In the future it is likely that solid state and synthetic solely on the assembly and organisation of supermolecules.It chemistry will evolve into a seamless continuum of activities is in this context that we have sought to develop a systematic which may be usefully discussed in the framework of a chemistry, termed heterosupramolecular chemistry, of covalently and non-covalently assembled and organised condensed systematic heterosupramolecular chemistry.J. Mater. Chem., 1998, 8(10), 2157–2164 2163115, 4927; (b) S. Yan and J. Hupp, J. Phys. Chem., 1996, 100, The work reviewed was supported in part or in full by grants 6867; (c) S. Yan and J. Hupp, J. Phys. Chem., 1997, 101, 1493. from the Commission of the European Union under the 15 (a) X.Marguerettaz, G. Redmond, S. N. Rao and D. Fitzmaurice, Training and Mobility of Researchers (Network) Programme Chem.Eur. J., 1996, 2, 420; (b) unpublished results. and Forbairt (The Irish National Agency for Science and 16 (a) B. O’Regan, M. Gra�tzel and D. Fitzmaurice, Chem. Phys. Technology). The cover art was prepared by Mr. Niall Murphy Lett., 1991, 183, 89; (b) G. Rothenberger, D. Fitzmaurice and M. Gra�tzel, J. Phys. Chem., 1992, 96, 5983. of Cell Media Ltd. 17 N. Kotov, F. Meldrum and J. Fendler, J. Phys. Chem., 1994, 98, 8827. 18 M.-J. Brienne, J. Gabard and J.-M. Lehn, J. Chem. Soc., Chem. References Commun., 1989, 1868. 19 (a) B. Feibush, A. Fiueroa, R. Charles, K. Onan, P. Feibush and 1 (a) J.-M. Lehn, Angew. Chem., Int. Ed. Engl., 1988, 27, 89; B. Karger, J. Am. Chem. Soc., 1986, 108, 3310; (b) B. Feibush, (b) D. J. Cram, Angew. Chem., Int.Ed. E, 27, 1009; M. Saha, K. Onan, B. Karger and R. Giese, J. Am. Chem. Soc., (c) C. J. Pedersen, Angew. Chem., Int. Ed. Engl., 1988, 27, 1021. 1987, 109, 7531; (c) A. Hamilton and D. Van Engen, J. Am. Chem. 2 (a) V. Balzani and F. Scandola, Supramolecular Photochemistry, Soc., 1987, 109, 5035; (d) A. Bisson, F. Carver, C. Hunter and Ellis Horwood, New York, 1991, ch. 3; (b) J. M Lehn, J. Waltho, J. Am. Chem. Soc., 1994, 116, 10292; (e) M. Tsuboi, Supramolecular Chemistry, VCH, New York, 1995, ch. 8. Appl. Spec. Rev., 1969, 3, 45; ( f ) H. Susi and J. Ard, Spectrochim. 3 (a) L. Cusack, S. N. Rao and D. Fitzmaurice, ACS Symp. Ser., Acta, Part A, 1971, 27, 1549; (g) R. Hamlin, R. Lord and A. Rich, 1997, 679, 17; (b) X.Marguerettaz, L. Cusack and D.Fitzmaurice, Science, 1965, 148, 1734; (h) Y. Kyogoku, R. Lord and A. Rich, Nanoparticle Characterisations and Utilisations, ed. J. Fendler, Proc. Natl. Acad. Sci. (USA), 1967, 57, 250; (i) Y. Kyogoku, R VCH–Wiley, New York, 1998, ch. 14. Lord and A. Rich, J. Am. Chem. Soc., 1967, 89, 496; ( j) Y. 4 (a) X.Marguerettaz, R. O’Neill and D. Fitzmaurice, J. Am. Chem. Kyogoku, R. Lord and A.Rich, Biochim. Biophys. Acta, 1969, Soc., 1994, 116, 2628; (b) X. Marguerettaz and D. Fitzmaurice, 179, 10; (k) G. Zundel, W. Lubos and K. Kolkenbeck, Can. J. Am. Chem. Soc., 1994, 116, 5017. J. Chem., 1971, 49, 3795; (l ) L. Bellamy, The Infrared Spectra of 5 (a) L. Cusack, S. N. Rao, J. Wenger and D. Fitzmaurice, Chem. Complex Molecules,Methuen, London, 1954; (m) G. Pimentel and Mater., 1997, 9, 624; (b) L.Cusack, S. N. Rao and D. Fitzmaurice, A. McClellan, The Hydrogen Bond, Freeman, San Francisco, 1956. Chem. Eur. J., 1997, 3, 202; (c) L. Cusack, R. Rizza, A. Gorelov 20 (a) X. Marguerettaz and D. Fitzmaurice, Langmuir, 1998, 13, and D. Fitzmaurice, Angew. Chem., Int. Ed. Engl., 1997, 36, 848; 6769; (b) A. Merrins, X. Marguerettaz, S. N. Rao and D.(d) L. Cusack, X. Marguerettaz, S. N. Rao, J. Wenger and Fitzmaurice, manuscript in preparation. D. Fitzmaurice, Chem. Mater., 1997, 9, 1765. 21 (a) D. LeV, P. Ohara, J. Heath and W. Gelbart, J. Phys. Chem., 6 (a) B. O’Regan and M. Gra�tzel, Nature, 1991, 353, 737; 1995, 99, 7036; (b) A. Badia, W. Gao, S. Singh, L. Demers, (b) M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, L.Cuccia and L. Reven, Langmuir, 1996, 12, 1262; (c) A. Badia, E. Mu� ller, P. Liska, N. Vlachopoulos and M. Gra�tzel, J. Am. S. Singh, L. Demers, L. Cuccia, G. Brown and R. Lennox, Chem. Chem. Soc., 1993, 115, 6382; (c) A. McEvoy and M. Gra�tzel, Sol. Eur. J., 1996, 2, 359; (d) J. Heath, C. Knobler and D. LeV, J. Phys. Energy Mater. Sol. Cells, 1994, 32, 221; (d) A. Hagfeldt and Chem., 1997, 101, 189; (e) S.Johnson, S. Evans, S. Mahon and M. Gra�tzel, Chem. Rev., 1995, 95, 49. A. Ulman, Langmuir, 1997, 13, 51. 7 (a) U. Ko� lle, J. Moser and M. Gra�tzel, Inorg. Chem., 1985, 24, 22 D. Ahern, S. N. Rao and D. Fitzmaurice, J. Phys. Chem., 2253; (b) H. Frei, D. Fitzmaurice and M. Gra�tzel, Langmuir, 1990, submitted. 6, 198; (c) J. Moser, S. Punchihewa, P.Infelta and M. Gra�tzel, 23 (a) P. Glink, C. Schiavo, F. Stoddart and D. Williams, Chem. Langmuir, 1991, 7, 3012; (d) G. Redmond, D. Fitzmaurice and M. Commun., 1996, 1483; (b) P. Ashton, E. Chrystal, P. Glink, Gra�tzel, J. Phys. Chem., 1993, 97, 6951; (e) J. Moser and M S. Menzer, C. Schiavo, N. Spencer, F. Stoddart, P. Tasker, A. White and D. Williams, Chem. Eur. J., 1996, 2, 709; Gra�tzel, Chem. Phys., 1993, 176, 493. (c) P. Ashton, P. Glink, F. Stoddart, P. Tasker, A. White and D. 8 (a) B. Borgias, S. Cooper, Y.-B. Koh and K. Raymond, Inorg. Williams, Chem. Eur. J., 1996, 2, 729. Chem., 1984, 23, 1009; (b) R. Rodrigues, M. Blesa and 24 N. Zacherroni, S. Wenger, L. Cusack, S. N. Rao, J. Preece, J. F. A. Regazonni, J. Colloid Interface Sci., 1995, 177, 122. Stoddart and D. Fitzmaurice, manuscript in preparation. 9 S. N. Rao and D. Fitzmaurice, Helv. Chim. Acta., 1998, 81, 902. 25 (a) B. Mirkin, R. Letsinger, R. Mucic and J. StorhoV, Nature, 10 D. Bahneman, A. Henglein, J. Lilie and L. Spahnel, J. Phys. 1996, 382, 607; (b) P. Alivisatos, K. Johnson, Xiaongang Peng, T. Chem., 1984, 88, 709. Wilson, C. Loweth, M. Bruchez and P. Schultz, Nature, 1996, 11 (a) S. Hu� nig and W. Schenk, Liebigs Ann. Chem., 1979, 727; 382, 609. (b) S. Hu� nig and W. Schenk, Liebigs Ann. Chem., 1979, 1523; 26 S. Quinn and D. Fitzmaurice, manuscript in preparation. (c) S.Hu� nig and H. Berneth, Top. Curr. Chem., 1980, 92, 4. 27 P. Alivisatos, J. Phys. Chem., 1996, 100, 13 226. 12 P. Pe�chy, F. Rotzinger, M. K. Nazeeruddin, O. Kohle, 28 D. Philp and J. F. Stoddart, Angew. Chem., Int. Ed. Engl., 1996, S. M. Zakeeruddin, R. Humphry-Baker and M. Gra�tzel, J. Chem. 35, 1154. Soc., Chem. Commun., 1995, 65. 13 X. Marguerettaz and D. Fitzmaurice, manuscript in preparation. 14 (a) L. Hong, J. Prieskorn and J. Hupp, J. Am. Chem. Soc., 1993, Paper 8/02968G 2164 J. Mater. Che

ISSN:0959-9428    

DOI:10.1039/a802968g

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 a novel liquid crystalline polymer, poly(2,5-didecyloxy- 1,4-phenylenebutadiynylene) Masashi Kijima,*† Ikuo Kinoshita, Takashi Hattori and Hideki Shirakawa Institute of Materials Science, University of Tsukuba, Tsukuba, Ibaraki 305-8573, Japan Received 27th May 1998, Accepted 11th August 1998 Oxidative polycondensation of 2,5-didecyloxy-1,4-diethynylbenzene with a Cu-amine catalyst gave a new soluble and fusible poly(2,5-didecyloxy-1,4-phenylenebutadiynylene) showing intense fluorescence, semiconducting properties upon doping with H2SO4, and thermotropic liquid crystalline properties.Scheme 1. The synthesis of linear carbon chains, carbynes, has been paid much attention as they represent a new carbon allotrope, although the syntheses have often not been successful due to unwanted side-reactions during the extension of the conjugation number and the instability of the product.The insolubility of the products has often led to ambiguous characterizations.1 In order to overcome their lability, stabilization of the carbon chains could be achieved by means of end capping,2–4 incorporation of inter-chain material,5,6 and introduction of a stable unit in the conjugated carbon chain.Poly(arylene-alt-oligoethnylene)s, in which the arylene unit and the oligoethynylene unit are alternatively linked to form a rod-like polymer, belong to the last case. Although poly(aryleneethynylene) s have been noted as optoelectronic materials in recent years,7–9 there have been few reports on the synthesis of polymers having polyyne segments longer than ethynylene, most of which involve metal–polyyne species.10–12 Thus we Fig. 1 Absorption (solid line) and fluorescence (broken line) spectra of polymer 2 in CHCl3. report the synthesis and properties of a new soluble and fusible rod-shaped poly(arylenebutadiynylene) as an example of fluorescent maximum was 475 nm and the UV–VIS absorption materials containing a linear carbon chain moiety.maximum was 424 nm in CHCl3 solution, which is comparable The synthesis of poly(2,5-didecyloxy-1,4-phenylenebutato the corresponding poly(aryleneethynylene) showing a fluo- diynylene) 2 is shown in Scheme 1. The decyloxy substituent rescent maximum at 475 nm and a UV–VIS maximum at at the phenylene unit is incorporated to increase the solubility 419 nm.and fusibility of the polymer, since unsubstituted poly(1,4- Stimulated by reports that similar rod-like poly(arylene- phenylenebutadiynylene) is almost insoluble and infusible.10 ethynylene)s exhibit thermotropic7 or lyotropic9 liquid crystal- 1,4-Didecyloxy-2,5-diethynylbenzene 17 (660 mg, 1.5 mmol in line properties, the thermal properties of the polymer 2 were 8 mL of THF) was added in a THF solution (30 mL) of a investigated by diVerential scanning calorimetry (DSC) and Hay catalyst13 [37 mg of CuCl and 44 mg of N,N,N¾,N¾- thermo-controlled polarizing microscopy.In the DSC curves tetramethylethylenediamine (TMEDA)] and reacted for 24 h in Fig. 2, two sharp exothermic peaks are observed during at room temperature under O2, giving a powdery yellow cooling, suggesting that the former is an isotropic–liquid polymer.The polymer was soluble in hot CHCl3 and THF, crystalline phase transition (DH=-74.6 J g-1)§ and the latter but the solubility was poor at room temperature. The crude is a liquid crystalline–crystalline phase transition (DH= polymer was purified via normal reprecipitation methods using -15.4 J g-1)§.On the other hand, the total enthalpy of the MeOH–HCl and MeOH successively, and dried under vacuum. two overlapping endothermic peaks is about 86 J g-1, approxi- The polymer structure was confirmed by IR (disappearance of mately corresponding to the total exothermic enthalpy an absorption at 3286 cm-1 due to acetylenic nCMH ), Raman (appearance of a new peak at 2200 cm-1 due to nCOC ), NMR and elemental analysis.‡ The number average molecular weight (Mn) of this polymer was 5400 (Mw/Mn=2.4) vs.polystyrene standards via GPC analysis. The conjugation length (n) is estimated to be about 12 from the Mn value. The low degree of polymerization and the high polydispersity are due to the heterogeneous conditions toward the end of the reaction, brought about by the low solubility of the produced polymer in the reaction medium.The polymer 2 shows an intense luminescence in solution (blue–green) even at a very low concentration (Fig. 1) and in the solid state (yellow). Similar fluorescent properties have been reported in the cases of poly(aryleneethynylene)s.7–9 The Fig. 2 DSC thermograms of polymer 2 at a scan rate of 10 °Cmin-1.†E-mail: kijima@ims.tsukuba.ac.jp J. Mater. Chem., 1998, 8(10), 2165–2166 2165400 kg f cm-2 was of the order of 10-8 S cm-1 without I2 doping, but it increased to the order of 10-5 S cm-1 after H2SO4 doping. A pellet sample of unsubstituted poly( pphenylenebutadiynylene) synthesized by the Hay method10 showed similar conductivity of the order of 10-8 S cm-1 and 10-6 S cm-1 after H2SO4 doping, respectively.The higher conductivity of the doped polymer might be due to the alkoxy substituents, which play a role in lowering the ionization potential of the polymer by analogy with the case of poly( pphenylenevinylene). 16 Consequently, it has been shown that the inherently rod-like polymer 2 having the flexible decyloxy side chains exhibits intense fluorescent, semiconducting, and thermotropic liquid crystalline properties.Furthermore, the polymer solidifies as a crystalline material via a liquid crystalline–crystalline phase transition. Introduction of the linear butadiynylene unit in the polymer main chain explains the increase in crystallinity. Further investigations on the synthesis of polymers with highmolecular weight and good processability for thin films and on structural analysis of the polymer in the mesophase and the crystalline phase are currently under way.The authors are grateful to Dr H. Goto for suggesting polarizing microscopic analyses of the phases. Notes and references ‡ Elemental analysis: Found: C, 82.2; H, 10.3. Calc. for (C30H44O2)n: C, 82.5; H, 10.15%.§ Molar enthalpy values are DH=-32.5 kJ mol-1 and -6.7 kJ mol-1, respectively, taking the repeat unit of the polymer as a molar unit. Mean enthalpies per mole of polymer molecules can be estimated by multiplying the molar enthalpies by the average number of conjugation (n=12). 1 Y. P. Kudryavtsev, R. B. Heimann and S. E. Evshukov, J. Mater. Sci., 1996, 31, 5557. Fig. 3 Optical micrographs of polymer 2 observed between crossed 2 R.Eastmond, T. R. Johnson and D. R. M. Walton, Tetrahedron, polarizers at 200×magnification (A) in a liquid crystalline state at 1972, 28, 4601. 105 °C and (B) in a crystalline state at room temperature. 3 R. J. Lagow, J. J. Kampa, H. Wei, S. L. Battle, J. W. Genge, D. A. Laude, C. J. Harper, R. Bau, R. C. Stevens, J. F. Haw and E.Munson, Science, 1995, 267, 362. (90 J g-1). These enthalpies are ascribed to orientation of the 4 T. Bartik, B. Bartik, M. Brady, R. Dembinski and J. A. Gladysz, rod-like polymer molecules. The polymer 2 melts at 125 °C, Angew. Chem., Int. Ed. Eng., 1996, 35, 414. and when it cools to 105 °C, birefringent droplets with poor 5 L. Kavan, J. Hlavaty�, J. Kastner and H. Kuzmany, Carbon, 1995, fluidity appear and a mesomorphic texture with a Maltese 33, 1321. 6 M. Kijima, T. Toyabe and H. Shirakawa, Chem. Commun., 1996, cross is observed [Fig. 3(A)], which is frequently observed for 2273. rigid backbone polymers in a homeotropic nematic state.11,14 7 M. Moroni, J. Le Moigne and S. Luzzati, Macromolecules, 1994, With further cooling of the sample below 75 °C, the mesophase 27, 562.changes to a crystalline phase exhibiting a spherulitic texture 8 K. Tada, M. Onoda, M. Hirohata, T. Kawai and K. Yoshino, Jpn. [Fig. 3(B)]. In spherulites, the molecular axes of the crystals J. Appl. Phys., 1996, 35, L251. are thought to be perpendicular to the radii of the spherulite.15 9 D. Steiger, P. Smith and C. Weder, Macromol. Rapid Coun. Thus the crystalline phase grows outwards in all directions 1997, 18, 643.from a point with a continual side-by-side accretion of the 10 K. Ohmura, M. Kijima and H. Shirakawa, Synth. Met., 1997, rod-like polymer chains. The thermo-controlled X-ray analysis 84, 41. 11 S. Takahashi, E. Murata, M. Kariya, K. Sonogashira and of the sample shows a broad peak around 2h=15–25° (d= N. Hagihara, Macromolecules, 1979, 12, 1016. 6–3.5 A° ) in the liquid crystalline state and a sharp peak at 12 J. Lewis, M. S. Khan, A. K. Kakkar, B. F. G. Johnson, 2h=23.4° (d=3.8 A° ) in the crystalline state. These peaks are T. B. Marder, H. B. Fyfe, F. Wittmann, R. H. Friend and due to the molecular distances, and peaks due to a layer A. E. Dray, J. Organomet. Chem., 1992, 425, 165; R. D. Markwell, distance typical of smectic phases were not be observed. From I.S. Butler, A. K. Kakkar, M. S. Khan, Z. H. Al-Zakwani and DSC, microscopic and XRD analyses, the liquid crystalline J. Lewis, Organometallics, 1996, 15, 2331. phase is concluded to be a columnar nematic phase consisting 13 A. S. Hay, J. Polym. Sci: Part A, 1969, 7, 1625. of rod-like polymer molecules. 14 S. M. Aharoni, Macromolecules, 1979, 12, 94. Thin film samples can be prepared by casting from a CHCl3 15 C. Bunn and T. C. Alock, Trans. Faraday Soc., 1945, 41, 317. 16 I. Murase, T. Ohnishi, T. Noguchi and M. Hirooka, Polym. or a THF solution on a glass plate but they are brittle, which Commun., 1985, 26, 362. is characteristic of low molecular weight rigid polymers. The electrical conductivity of a pellet sample prepared Communication 8/03953D from the as-prepared powdery product under a pressure of 2166 J. Mater. Chem., 1998, 8(10), 2165–2166

ISSN:0959-9428    

DOI:10.1039/a803953d

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 Coating nanosized iron oxide particles on submicrospherical alumina by a sonochemical method Ziyi Zhong, Yanming Zhao, Yuri Koltypin and Aharon Gedanken* Department of Chemistry, Bar-Ilan University, Ramat-Gan, 52900, Israel Recieved 23rd June 1998, Accepted 22nd July 1998 Iron oxide which adhered strongly to an amorphous alum- crystallized alumina, a c-Al2O3 structure was confirmed by its XRD pattern [Fig. 1(d)] compared to ASTM card-10–425.ina core has been prepared by a sonochemical method; the phase transformation of maghemite to haematite was The weakening of c-Al2O3 peaks in the XRD pattern for coated samples is due to the fact that the alumina surface was greatly retarded in the coating. covered by iron oxide.When the sonication was carried out with the submicrospherical crystallized c-alumina and the The coating of metals and metal oxides on substrates and the product heated to 400 °C in argon, c-Fe2O3 (or Fe3O4) was interaction between coated metallic elements and substrates observed by XRD [Fig. 1(e)]. have long been of interest to materials scientists1,2 and catalysis Fig. 2 shows the TEM images of two samples heated to chemists.3,4 Conventional methods such as evaporation, 400 °C in argon. Fig. 2(a) shows the TEM image of iron oxide impregnation, precipitation and sputtering usually yield a coated on amorphous alumina, and Fig. 2(b) that on crys- coated polycrystalline product. As reported by Suslick et al.5 tallized c-alumina. It is observed that the alumina substrate the sonochemical method can be used as a tool to prepare has a particle size ranging from 180 to 300 nm and a regular nanosized amorphous metals.The formation of the amorphous spherical shape. Most of the Fe2O3 particles coated on amorph- metals is due to the extreme conditions, such as the highest ous alumina are resolved, and separated from each other and transient temperature, which exceeds 5000 K at localized hot possess a relatively regular spherical shape.Moreover, they spots, and the ultrafast cooling rate of >1010 K s-1, that can are strongly adhered to the alumina core and have a radius of be obtained in sonication. Recently, sonication has been further 10–20 nm, whereas the Fe2O3 coated on crystallized c-alumina developed as a tool to drive the deposition of iron oxide, nickel is an agglomerate of small particles with a size distribution of and cobalt on the surface of silica particles.6–8 However, for 1–10 nm, resembling a dense cloud.Most of Fe2O3 coated on materials scientists, since silica can form impurity phases with crystallized alumina is not adhered to the substrate. This result many coated magnetic materials,9–11 Al2O3 is a superior subindicates that the deposited materials are strongly influenced strate and a better matrix material.This is the reason why by the reactivity of the alumina surface. developing a general and eYcient coating method for Al2O3 is Fig. 3 shows DSC curves of two as-prepared sonication an important objective. Here, we report the coating of nanosproducts.Fig. 3(a) shows the DSC curve of as-prepared iron ized iron oxide particles on submicrospherical alumina by oxide coated on amorphous alumina, and Fig. 3(b) that on employing sonication. We also report the new finding that the crystallized alumina. The endothermic peak at ca. 150 °C in known transformation of maghemite (c-Fe2O3) to haematite both Fig. 3(a) and (b) is attributed to the desorption of con- (a-Fe2O3) is greatly retarded owing to a strong interaction between the coated iron oxide and the amorphous alumina.Submicrospherical alumina was prepared by hydrolysis of a dilute solution of aluminum sec-butoxide in a mixture of octan- 1-ol, butan-1-ol and acetonitrile.12 The precipitate was washed thoroughly with acetone and dried in vacuum.Crystallized calumina was obtained when the precipitate was heated to 1000 °C. Sonochemical coating of nanosized iron oxide particles on alumina spheres was carried out under 1.5 atm of argon at 0 °C for 3 h (Ti-horn, 20 kHz, 100 W cm-2). In the sonication cell, 400 mg of pre-prepared alumina submicrospheres, 1 ml of iron pentacarbonyl, and 40 ml decalin were mixed together.Prior to sonication, argon gas was bubbled for 1 h, preventing possible oxidation of iron carbonyl from air. The sonication product was then washed thoroughly with hexane. Fig. 1 shows the XRD patterns of as-prepared alumina and its iron oxide-coated products. It clearly indicates that asprepared alumina and its as-prepared sonication product are formed in the amorphous state [Fig. 1(a)]. This amorphous nature was further confirmed by electron diVraction measurements, which showed only a diVuse ring pattern. However, the as-prepared coated sample heated at 400 °C developed the characteristic patterns of c-Fe2O3 or Fe3O4 (magnetite). DiVerentiating between c-Fe2O3 and Fe3O4 by XRD is some- Fig. 1 XRD patterns of (a) the as-prepared material obtained by what diYcult, because their XRD patterns are almost identical sonicating Fe(CO)5 and amorphous alumina(the alumina is denoted (ASTM card 19–629 for Fe3O4 and 24–81 for c-Fe2O3).as A); (b) iron oxide coated on A and heated at 400 °C in argon; (c) However, Mo�ssbauer spectroscopy results show the product iron oxide coated on A and heated to 550 °C; (d) the as-prepared to be a mixture of c-Fe2O3 (18%) and Fe3O4 (72%).material obtained by sonication of Fe(CO)5 and crystallized c-alumina Additionally, we can not completely exclude the possibility (the alumina is denoted as B and marked with an asterisk); (e) iron that c-Fe2O3 was produced from Fe3O4 in the process of oxide coated on B and heated at 400 °C in argon; (f ) iron coated on B and heated to 550 °C. sample handling, since Fe3O4 is very sensitive to air.For J. Mater. Chem., 1998, 8(10), 2167–2168 2167two as-prepared coated samples were heated to 550 °C under argon, and XRD measurements carried out. The XRD patterns are shown in Fig. 1(c) (iron oxide coated on amorphous alumina) and Fig. 1(f ) (iron oxide on crystallized alumina), respectively. It is clearly seen that Fe2O3 coated on amorphous alumina remains as a c-Fe2O3 phase even after heating to 550 °C, while the Fe2O3 on crystallized alumina is converted to a-Fe2O3.The absence of the 450 °C exothermic peak is a result of a strong interaction between the adhered iron oxide and the amorphous alumina, which results in the elevation of the phase transition temperature of c-Fe2O3. At the same time, unadhered Fe2O3, when coated on crystallized alumina, shows a similar phase transformation behaviour to bulk Fe2O3.The mechanism of the c�a phase transformation is known to be dependent on the particle size and the mechanism of this transformation for small particles (<15 nm) is a chain mechanism15 which involves recrystallization of up to 100 particles to form single a-Fe2O3 flakes of ca. 40–70 nm. Obviously, the strong interaction Fig. 2 TEM micrographs: (a) iron oxide coated on A alumina and between coated iron oxide and alumina can hinder this type of heated to 400 °C in argon; (b) iron oxide coated on B alumina and gathering and recrystallization of iron oxide particles on an heated to 400 °C in argon. alumina surface. In addition, we can not rule out the involvement of trace Al3+ ions, which are known to replace Fe3+ ions in the unit cell, in elevating the transformation temperature.17 If this phenomenon is indeed responsible for the higher transformation temperature, it should occur mainly with the amorphous alumina, where the interaction is stronger.A. Gedanken thanks the Ministry of Science and Technology for supporting this research through the grant for infrastructure. Dr.Yuri Koltypin thanks the Ministry of Absorption for his Giladi scholarship. Dr. Ziyi Zhong and Dr. Yanming Zhao thank the Kort Scholarship fund for supporting their postdoctoral fellowships. The authors thank Professor M. Deutsch, Department of Physics, for extending the XRD facility, and Dr. Shifra Hochberg for editorial assistance.References 1 D. Segal, Chal Synthesis of Advanced Ceramic Materials, Cambridge University Press, 1989. 2 H. Yanagida, K. Koumoto, M. Miyayama, The Chemistry of Ceramics, John Wiley & Sons, New York, 1996. 3 A. A. Tsyganenko and P. Mardilovich, J. Chem. Soc., Faraday Trans., 1996, 92, 4843. 4 D. A. Hucul and A. Brener, J. Phys. Chem., 1981, 85, 496. 5 (a) Ultrsound: Its Chemical, Physical and Biological EVects, ed.K. S. Suslick, VCH,Weinheim, 1988; (b) K. S. Suslick, S. B. Choe, A. A. Cichowlas and M. W. GrinstaV, Nature, 1991, 353, 414. 6 S. Ramesh, R. Prozorov and A. Gedanken, Chem. Mater., 1997, Fig. 3 DSC curves of (a) as-prepared sample of iron oxide coated on 9, 2996. A alumina; (b) as-prepared sample of iron oxide coated on B alumina; 7 S. Ramesh, Y.Koltypin, R. Prozorov and A. Gedanken, Chem. (c) iron oxide prepared by the sonication of Fe(CO)5 under air. Mater., 1997, 9, 546. 8 S. Ramesh, Y. Cohen, R. Prozorov, K. V. P. M. Shafi, D. Aurbach and A. Gedanken, J. Phys. Chem. B, submitted. 9 J. L. Dormann, C. Djega-Mariadasson and J. Jove, J. Magn. taminants such as solvent molecules from the surface of the Magn. Mater., 1992, 104–107, 1576.product. The large exothermic peak in the range 200–350 °C 10 C. Djega-Mariadasson, J. L. Dormann, M. Nogues, G. Viller and corresponds to the crystallization of amorphous c-Fe2O3. This S. Sayouri, IEEE Trans. Magn., 1990, 26, 1819. is consistent with our previous results,13,14 as well as with the 11 D. L. Leslie-Pelecky and R. D. Rieke, Chem. Mater., 1996, 8, above XRD results.On the other hand, it is of particular note 1770. that in contrast to Fe2O3 coated on amorphous alumina, which 12 T. Ogihara, H. Nakajima, T. Yanagawa, N. Ogata, K. Yoshida shows an almost featureless curve in the range 400–550 °C, a and N. Matsushita, J. Am. Ceram. Soc., 1991, 74, 2263. comparatively large exothermic peak was observed for Fe2O3 13 X. Cao, Y.Koltypin, R. Prozorov, G. Kataby and A. Gedanken, J. Mater. Chem., 1997, 7, 2447. coated on crystallized c-alumina [Fig. 3(b)]. It is proposed that 14 X. Cao, R. Prozorov, Y. Koltypin, G. Kataby, I. Felner and transformation of c-Fe2O3 to a-Fe2O3 occurs in this tempera- A. Gedanken, J. Mater. Res., 1997, 12, 402. ture region15,16 (The above observation of a mixture of c-Fe2O3 15 R. M. Cornell and U. Schwertmann, The Iron Oxides–Structure, and Fe3O4 in the sample heated at 400 °C in argon does not Properties, Reactions, Occurrence and Uses, VCH, Weinheim, influence our analysis, since, it is known that, for Fe3O4 particles 1996. smaller than 300 nm, the transformation to a-Fe2O3 includes 16 F. del. Monte, M. P. Morales, D. Levy, A. Fernandez, M. Ocana, an initial formation of c-Fe2O315). To confirm this, two investi- A. Roig, E. Molins, K. O. Gray and C. J. Serna, Langmuir, gations were carried out, First, a pure amorphous Fe2O3 sample 1997,13, 3627. was prepared following the method of Cao et al.13 Its DSC 17 P. S. Sidhu, Clays Clay Miner., 1988, 36, 31. curve is shown in Fig. 3(c) from which it is seen that a similar but larger exothermic peak is observed. Secondly, the above Communication 8/04759F 2168 J. Mater. Chem., 1998, 8(10), 2167–2168

ISSN:0959-9428    

DOI:10.1039/a804759f

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J O U R N A L O F C H E M I S T R Y Materials A novel method for grafting polymers on carbon blacks Jarrn-Horng Lin,a Hsiu-Wei Chen,*a Kuo-Tung Wangb and Feng-Hsiung Liaw b aDepartment of Chemistry, National Sun Yat-Sen University, Kaohsiung, Taiwan, Republic of China bResearch and Development Department, China Synthetic Rubber Corporation, Kaohsiung, Taiwan, Republic of China Received 5th May 1998, Accepted 30th June 1998 A simple and eVective impregnation method to graft polymers on carbon blacks without any coupling agent or complicated pretreatment is reported.Dispersibility measurements, acidity measurements, infrared spectra and surface functional group eVects on the grafting eYciency suggest that the acidic surface functional groups on carbon blacks are key factors in controlling the grafting eYciency of the impregnation process.Table 1 Specific surface areas and volatilities of carbon blacks Introduction Carbon black Specific surface areaa/m2 g-1 Volatilityb/wt% Carbon blacks are important materials that have been widely used as fillers in elastomers, plastics, paints, inks, etc. However, FW200 460 20.0 unmodified carbon blacks cannot be used directly for certain N330 83 0.8 applications.The desired surface properties of carbon blacks BP2000 1475 1.8 cannot always be obtained by the traditional manufacturing aMeasured by BET method with nitrogen at 77 K. Data provided by processes. Therefore, the modification of carbon black surfaces the supplier. bMeasured via TGA under helium flow between 25 by chemical and physical methods, such as oxidation,1–6 and 900 °C.plasma treatment,7,8 polymer grafting,9–19 etc., in particular polymer grafting, have been intensively studied. Most carbon blacks produced are used in combination with polymers, especially with elastomers in the rubber industry. ment of carbon blacks was done by heating the carbon blacks The compatibility of carbon blacks in elastomer matrices is in 3.6 M nitric acid solution at 80 °C for 1 h.The nitric acid markedly improved by grafting polymers onto the surface of treated carbon blacks were then washed with deionized water carbon blacks. Grafting polymers on carbon blacks not only in an ultrasonic cleaner for 10 min and dried in an oven at enables carbon blacks to have good dispersibility in solvents 90 °C before use.Polyethylene glycol (PEG), poly(vinyl and polymers, but also enables carbon blacks to acquire useful alcohol ) (PVA), 1-aminopropan-2-ol, monoethanolamine properties, such as cross-linking ability, bioactivity, and (MEA) and diethanolamine (DEA) have been used to test the photosensitivity. grafting eYciency. Several methodologies have been developed to graft Grafting onto carbon black surfaces was obtained via an polymers onto carbon black surfaces.19 Polymers can be impregnation method.For a typical run, 1 g of carbon black grafted onto carbon surfaces via a termination reaction was mixed with the desired amount of polymer dissolved in between growing polymer radicals and functional groups on 2 ml of water or other solvent described in the text.It was the carbon surfaces, and by a polymerization reaction initiated then stirred at room temperature for 5 min, and finally dried from initiating groups that have been previously introduced in an oven at 100 °C overnight. on the carbon surfaces. Polymers may also be grafted onto The thermogravimetric analysis (TGA) was performed on carbon surfaces by reaction between carbon surface groups a Perkin-Elmer TGA-7A instrument. About 10 mg of carbon and functional polymers.Some complicated pretreatments and black sample was loaded into the instrument in a platinum multiple steps with special coupling agents are required for all sample pan. The temperature was ramped from room temperathe methods of grafting polymers onto carbon blacks men- ture to 950 °C with a rate of 60 °Cmin-1 under a helium flow tioned above.(150 ml min-1). In this paper, we report a simple impregnation method that Infrared (IR) spectra were taken with a Digilab FT 175 FTcan eVectively graft polymers onto carbon blacks without a IR spectrometer and were recorded in absorbance mode with coupling agent. Polymers and organic compounds with func- a resolution of 2 cm-1 and 32 scans at room temperature.The tional groups such as hydroxy or amine groups can be grafted 10 mm diameter sample disk was prepared by mixing of 0.01 g onto carbon blacks by this method. The nature of the polymer of the carbon black with KBr powder in a ratio of 15300. grafted carbon blacks was analyzed and characterized by The pH of polymer grafted carbon blacks was measured Fourier-transform infrared spectroscopy (FT-IR), ultraviolet with a standard combination electrode immersed in an aqueous spectroscopy and thermogravimetric analysis (TGA) under a slurry prepared with 1.5 g carbon black in 20 ml of distilled helium atmosphere.water. The slurry was ultrasonically agitated and stirred for 3 min before introduction of the electrode. The percentage of polymer grafted onto carbon black Experimental surfaces was determined from the diVerence in weight of carbon black before and after the impregnating treatment.To Carbon blacks used in this study were FW200 (Degussa Co.), BP2000 (Cabot Co.), and N330 (CSRC Co.). Their physical eliminate the non-grafted polymer and the physisorbed polymer, the reaction products were washed with 100 ml of deion- properties are shown in Table 1.The carbons were prewashed with deionized water using an ultrasonic cleaner for 10 min ized boiled water using an ultrasonic cleaner for 10 min before measurement of the grafting percentage. The percentage of and dried in an oven at 90 °C before use. Nitric acid pretreat- J. Mater. Chem., 1998, 8(10), 2169–2173 2169Fig. 2 Stability of PEG6000 grafted carbon black dispersion in water Fig. 1 The calibration curve plot of 2 wt% PEG6000 grafted FW200 at room temperature: (6) 2% PEG6000 on N330, (#) FW200, (() carbon black concentration in water versus 249 nm ultraviolet light 2% PEG600 on nitric acid treated N330 carbon black and (%) 20% absorbance. PEG6000 on FW200 carbon black. All grafted carbon blacks were washed with boiled water before stability measurements.grafting was calculated via eqn. (1). PEG6000 grafted carbon blacks produced stable colloidal Grafting (%)=100×[Polymer grafted (g)/ dispersion in water. The stability of PEG6000 grafted carbon carbon black charged (g)] (1) black dispersions was compared with those of untreated carbon The dispersibility of polymer grafted carbon black in water blacks in Fig. 2. As shown in Fig. 2, FW200 carbon black was estimated as follows. Water washed polymer grafted readily precipitated after five days. N330 carbon blacks comcarbon black (0.1 g) was dispersed in 100 ml of deionized pletely precipitated within an hour. N330 carbon black (withwater and stirred with a magnetic stirrer for 10 min. Then the out nitric acid pretreatment) coated with PEG6000 also dispersions were allowed to stand at room temperature. After precipitated within an hour, suggesting that PEG6000 cannot a given time, 2 ml of dispersion liquid was removed with a be grafted onto N330 (without nitric acid treatment).However, pipette, and the content of the dispersed carbon black was more than 60% of the PEG6000 grafted FW200 black and the determined.The amount of the dispersed carbon black was PEG6000 grafted N330 black (nitric acid pretreated) remained estimated by measuring the absorbance of carbon blacks in dispersed in water after a month. the ultraviolet (UV) spectrum at a wavelength where the Fig. 3 shows the TGA results of PEG6000 on FW200 carbon carbon black shows a peak absorbance, i.e. 249 nm for FW200. black. A weight loss peak at 406 °C was observed when a pure A calibration curve plot of the concentration of carbon black PEG sample was tested [Fig. 3(a)], suggesting that the pure in water vs. UV absorbance is shown in Fig. 1. The linear PEG volatilized or decomposed at 406 °C in the He flow. relationship between the carbon black concentrations and the Three weak weight loss peaks at 73, 276 and 700 °C were UV absorbance suggests that the UV method is adequate for observed when the pure FW200 carbon black sample was the dispersed carbon black contents measurement.The dispersibility of carbon blacks in water was estimated using eqn. (2). Dispersibility (%)=100× Carbon blacks dispersed after standing Carbon blacks dispersed before standing (2) Results Polymers with functional groups, such as hydroxy or amine moieties, can be grafted onto carbon blacks by the impregnation method.Polyethylene glycol (PEG, a hydrophilic polymer) with average molecular weight of 6000 (PEG6000) grafted on FW200 carbon black was used as a model to demonstrate the grafting eYciency of the impregnation method. The maximum amount of PEG6000 that can be grafted onto FW200 carbon black by the impregnation method is about 78% of the weight of FW200 carbon black. The results of dispersibility measurements, surface functional group eVects on grafting eYciency, TGA and infrared spectra confirmed that the polymer has been successfully grafted onto carbon surfaces through an interaction between the surface functional groups and the polymers.The interaction can be described as a chemisorption process where the surface functional groups on carbon are the active sites. Chemisorption, by definition, is a process where particles stick to the surface Fig. 3 TGA of (a) pure PEG6000, (b) pure FW200 carbon black and by chemical (usually covalent) bonding, and then find sites (c) 20 wt% PEG6000 on FW200 carbon black in He atmosphere.The which maximize their coordination number with the substrate. He flow rate is 150 ml min-1. The heating rate is 2 °Cmin-1 when The sites used for chemisorption within our impregnation the weight loss is higher than 2 wt% min-1, and is 100 °Cmin-1 when method are most likely the acid surface functional groups on the weight loss is lower than 0.3 wt% min-1.(—) Weight loss and (B) first derivative. carbon blacks. These results are shown in Fig. 2–6. 2170 J. Mater. Chem., 1998, 8(10), 2169–2173tested [Fig. 3(b)]. The strong interaction between PEG6000 and FW200 carbon black led to a decrease of the PEG6000 decomposition peak temperature from 406 °C to as low as 210 °C [Fig. 3(c)]. The quantity of functional groups on a carbon black surface is an important factor in controlling the grafting eYciency of the impregnation method.Here we choose N330 and BP2000 carbon blacks to demonstrate the eVect of surface functional groups on the grafting eYciency of the impregnation method. Table 1 shows some physical properties of carbon blacks used in this study. The number of surface functional groups (demonstrated as volatilized contents of carbon blacks in Table 1) on BP2000 and N330 is lower than 2% of the weight of the carbon blacks.A simple nitric acid pretreatment can increase the quantity of oxygen-containing functional groups, especially the carboxylic groups on a carbon black surface.1,6,20 The dispersibility of N330, BP2000 and PEG6000-coated N330 and BP2000 (without nitric acid pretreatment) in water was poor (see Table 2).Good dispersibility in water can be obtained when PEG6000 was impregnated on nitric acid treated N330 and BP2000 carbon blacks.We also used ethylene glycol as a model to study the eVect of functional groups of PEG polymers on the grafting ability. When the two hydroxy groups of ethylene glycol were replaced by methoxy groups, the grafting ability was reduced from 94% (ethylene glycol on FW200 carbon black) to 6% (1,2-dimethoxyethane on FW200).These results suggest that the interactions between polymers and carbon blacks most likely take place through the reaction between the oxygen-containing surface functional groups on Fig. 4 TGA of (a) 2 wt% PEG6000 on N330 carbon black, (b) 2 wt% carbon blacks and functional groups on the polymers.Fig. 4 PEG6000 on nitric acid treated N330 carbon black, (c) 20% PEG6000 on BP2000 carbon black and (d) 30% PEG6000 on nitric acid treated shows the TGA results of PEG6000 on BP2000 and N330 BP2000 carbon black. (B) Weight loss and (—) first derivative. carbon blacks and on nitric acid pretreated BP2000 and N330 carbon blacks. The peak decomposition temperatures of PEG6000 coated on N330 [398 °C, Fig. 4(a)] and on BP2000 835 cm-1. The grafting eVect of PEG6000 on carbon black shifts the CNO and C–O IR bands of FW200 carbon black [383 °C, Fig. 4(c)] are similar to the peak decomposition temperature (406 °C) of pure PEG6000. However, the peak from 1720, 1596 and 1261 cm-1 to 1742, 1576 and 1238 cm-1, respectively, suggesting that the bond strength of the oxygen- decomposition temperatures of PEG6000 shifted down to 358 and 246 °C, respectively, when PEG6000 was grafted onto ated functional groups, such as CNO and C–O, are altered in the grafted polymer.nitric acid pretreated N330 [Fig. 4(b)] and BP2000 [Fig. 4(d)] carbon blacks. The eVect of the grafted polymer on the acidity of the carbon black is shown in Fig. 6. The pH of the carbon black Infrared spectra were used to examine the interaction between polymer and carbon black. Fig. 5 shows infrared dispersion solution increases with increases in the amount of polymer grafted on the carbon black surface until the pH spectra of PEG6000 and PEG6000 grafted FW200 carbon black. Four major bands were observed for FW200 at 3426, value approaches 4.The shape of the pH plot curve is interesting. The variation of the pH curve of carbon black 1720, 1596 and 1261 cm-1. These bands have been assigned to be the O–H (3450 cm-1), CNO (1720, 1596 cm-1) and with respect to the amounts of grafted polymer may correlate with the quantity of various kinds of acidic functional groups C–O (1261 cm-1) vibrations from the hydroxy, carboxylic, carboxylic anhydride and lactone functional groups on the that are available on the carbon black surface to interact with the impregnated polymers.The grafting eYciency of PEG6000 carbon surface.21 Part of the 1596 cm-1 band may also be due to a contribution from the CNC stretch of the aromatic rings polymer on the FW200 black can only be enhanced a little when the pH value of the carbon black dispersion is higher on the carbon surface.22 Fig. 5(c) is the IR spectrum of PEG6000 grafted FW200 carbon black. The IR bands of than 4.5. This result suggests that the grafting reaction between PEG and carbon black occurs largely through the acidic PEG6000 appear as fine structures on the broad IR bands of carbon black located at 2868, 1469, 1281, 1090, 943 and functional groups on the carbon black surfaces.Table 2 Dispersibilities of polymer/monomer grafted carbon blacks in watera Acid-treated Acid-treated Grafting materials FW200 N330 N330 BP2000 BP2000 Poly(vinyl alcohol ) + — + — + 1-Aminopropan-2-ol + (35%) — + — + Ethylene glycol + (94%) — + — + PEG600 + — + — + PEG3000 + — + — + PEG6000 + (78%) — + — + Monoethanolamine + (89%) — + — + Diethanolamine + (52%) — + — + a+=More than 50% dipersibility in water was observed after the carbon black dispersion slurry had stood for 1 month.—=Carbon blacks precipitated within one day.Numbers in parentheses are the maximum amounts of polymers or monomers that can be grafted on to the carbon surfaces. J. Mater. Chem., 1998, 8(10), 2169–2173 2171Second, the positive relationship between the grafting eYciency and the amount of volatile species on carbon supported the premise that the polymer had been grafted on the carbon surface.A chemical bond can be easily formed between an organic acid group and an alcohol group. The higher grafting eYciency of PEG6000 on carbon black surfaces with higher amounts of volatile species suggests that the sites available for the chemisorption reaction on the carbon surfaces are the surface functional groups.The surface functional groups involved in the grafting reaction are most likely the acidic functional groups on carbon blacks. The acidity eVects, infrared spectra and the polymer grafting on nitric acid pretreated carbon blacks support this hypothesis. The acidity of the carbon black dispersion solution decreases with an increase in the quantity of the polymer grafted, suggesting that the acid sites on the carbon black surfaces were occupied by the grafted polymers.PEG6000 can be grafted on nitric acid pretreated N330 black, but not on N330 black. It is well Fig. 5 FT-IR spectrum of (a) pure PEG6000, (b) FW200 carbon known that the quantity of acidic functional groups, especially black, (c) 20 wt% PEG6000 on FW200 carbon black and (d) 20% the carboxylic acid groups, on a carbon black surface increases PEG6000 impregnated FW200 carbon black after a water-wash with nitric acid treatments.6,20 The positive eVect of nitric acid procedure to remove any physically coated PEG6000.treatment on the grafting eYciency confirmed that the interaction between PEG6000 and carbon blacks takes place via the acidic surface functional groups, especially the carboxylic group.The shifting of the CNO and C–O bonds absorption wavelengths in the polymer-grafted black infrared spectra is consistent with this conclusion. The grafting reaction between PEG and carbon blacks takes place largely on the functional groups on surfaces through a chemisorption process.Carbon surface structures, like defects and pore structures, may also contribute to the grafting behavior during the impregnation process. However, these properties are not solely responsible for the success of the impregnation method. This is proved by the zero or low yield of the grafting reaction between PEG and carbon blacks that have a low percentage of surface functional groups (below 1% of the weight of the carbon blacks).Little grafting yield was observed when N330 black was impregnated with PEG6000. The quantity of surface functional groups on N330 carbon black is lower than 0.9 wt% (see Table 1). BP2000 black has Fig. 6 Relationship between pH and PEG6000 grafting level on FW200 carbon black. wide pore size distributions and high pore volumes23 but the grafting eYciency of PEG6000 on BP2000 black is low.Note that defects and pore structures exist on all kinds of carbon Discussion black surfaces. However, the grafting eYciency is low on carbon blacks with few surface functional groups. Several aspects of the dispersion measurements, TGA and the eVect of surface functional groups on the grafting eYciency Although only a small fraction of commercially available blacks have a significant density of surface groups, the impreg- data support the conclusion that the PEG6000 has been successfully grafted onto the carbon surfaces after the impre- nation method is still good for grafting polymers on blacks.The successful grafting of polymer on nitric acid pretreated gnating process.First, FW200 carbon blacks acquired a hydrophilic property after the grafting process. The PEG6000 grafted N330 black suggest that the impregnation method is applicable to carbon blacks with low amounts of surface functional carbon blacks can produce stable colloidal dispersions in water, suggesting that the PEG6000 has been successfully groups after a simple nitric acid treatment.The decomposition temperature of PEG6000 shifts about grafted onto carbon surfaces by the impregnation method; carbon blacks are hydrophobic materials. A stable colloidal 200 °C after the grafting process. The grafting eVect on the thermal stability of PEG6000 on FW200 carbon black can be dispersion of carbon blacks in water could not be acquired if the hydrophilic polymer was just physisorbed on the sur- easily noted by comparing the TGA of carbon black before and after the impregnation treatment, shown in Fig. 3. This face.11,15 In physisorption processes there is only a van der Waals interaction (for example, a dispersion or dipolar inter- result again supports the premise that a chemical bond has been formed between the polymer and the carbon black.action) between the adsorbate and the substrate. The total amounts of physisorbed species is insensitive to the surface Physical adsorption of the polymer on the carbon surface would not induce such a strong influence on the thermal properties of the substrate. The polymer-grafted carbon blacks have been washed with deionized boiled water before the stability of a polymer. An increase in the thermal degradability of a polymer can be beneficial in terms of environmental dispersibility measurement.Most, if not all, of the polymers would be washed out from the carbon black systems if the protection. A lot of research has been devoted to developing degradable polymers in order to reduce the environmental impregnated polymers were simply physisorbed on the carbon black surface.Physisorption can occur on all kinds of surfaces. pollution caused by large amounts of nondegradable polymers. 24 The weakening of the thermal stability of a polymer The low dispersibility of PEG grafted N330 carbon black in water further proved that the colloidal stability of PEG grafted by modifying the polymer with a functional end group has been reported.24 The weakening of the thermal stability of carbon black cannot be obtained by the physical adsorption mechanism alone.chemisorbed species is also commonly observed in adsorption- 2172 J. Mater. Chem., 1998, 8(10), 2169–2173desorption systems. Consider, for example, that acetophenone China for financial support of this study (NSC 86-2113-M- 110-004). is stable in inert atmosphere at temperatures higher than 200 °C.However, acetophenone decomposes at temperatures as low as 50 °C when chemisorbed on transition metal References surfaces.25 1 E. Papirer, J. Dentzer, S. Li and J. B. Donnet, Carbon, 1991, The grafting of PEG onto carbon black surfaces by the 29, 69. impregnating procedure is a general method that can be 2 S. Asai, K. Sakata, M. Sumita, K.Miyasaka and A. Swakari, Bull. extended to monomers and polymers containing hydroxy or Chem. Soc. Jpn, 1991, 12, 1672. amine groups with diVerent average molecular weights. The 3 S. I. Pyun, E. J. Lee, T. Y. Kim, S. J. Lee, Y. G. Ryu and dispersibilities of some polymer-grafted carbon blacks in water C. S. Kim, Carbon, 1994, 32, 155. are shown in Table 2. Poly(vinyl alcohol ), monoethanolamine, 4 J.B. Donnet, Kautsch. Gummi Kunstst., 1994, 47, 628. 5 R. C. Sosa, D. Masy and P. G. Ronxhet, Carbon, 1994, 32, 1369. diethanolamine, triethanolamine and PEG with average mol- 6 Y. Otaka and R. G. Jenkins, Carbon, 1993, 31, 109. ecular weight between 600 and 8000 have been successfully 7 J. B. Donnet, W. D. Wang and A. Vidal, Carbon, 1994, 32, 199. grafted onto FW200 and nitric acid treated carbon blacks by 8 M.Nakahara, K. Ozawa and Y. Sanada, J. Mater. Sci., 1994, the simple impregnation method. The nitric acid treatment 29, 1646. eVect and the functional group eVect of the monomer on the 9 N. Tsubokawa and K. Yanadori, Kobunshi Ronbunshu, 1992, 49, grafting ability suggest that ester or amide formation is also 865. 10 G. M. Chan, Appl.Surf. Sci., 1982, 10, 377. the most likely mechanism for the binding of low molecular 11 N. Tsubokawa and S. Handa, Pure Appl. Chem., 1993, A30, 277. weight –OH/–NHx species on carbon surfaces. 12 N. Tsubokawa, K. Fujiki and T. Sasaki, Kobunshi Ronbunshu, 1993, 50, 235. 13 S. Drappel, J. M. Gauthier and E. Franta, Carbon, 1983, 21, 311. Conclusions 14 K. Fujiki, N. Tsubokawa and Y.Sone, Polym. J., 1990, 228, 661. Polymers and organic compounds with functional groups such 15 N. Tsubokawa, Polym. Bull., 1989, 22, 655. 16 N. Tsubokawa and H. Tsuchida, Pure Appl. Chem., 1992, A29, as hydroxy or amine groups can be grafted onto carbon black 311. surfaces by a simple impregnation method. The quantity of 17 W. D. Wang, A. Vidal, G. Nanse and J. B. Donnet, Kautsch. acidic surface functional groups on the carbon black is an Gummi Kunstst., 1994, 47, 493. important factor in controlling the grafting eYciency of the 18 E. Papirer and D. Y. Wu, Carbon, 1990, 28, 393. impregnation process. Nitric acid pretreatment can improve 19 N. Tsubokawa, Prog. Polym. Sci., 1992, 17, 417. grafting eYciency on carbon blacks with a low density of 20 J. S. Noh and J. A. Schwarz, Carbon, 1990, 28, 675. 21 J. M. O’Reilly and R. A. Mosher, Carbon, 1983, 21, 47. surface functional groups. From acidity measurements, infra- 22 P. E. Fanning and M. A. Vannice, Carbon, 1993, 31, 721. red spectra and nitric acid eVects on the grafting reaction, we 23 M. Kruk, M. Jaroniec and Y. Bereznitski, J. Colloid Interface Sci., conclude that acidic surface functional groups, especially car- 1996, 182, 282. boxylic groups, play a major role in the grafting reaction of 24 Y. Nagasaki, N. Yamazaki and M. Kato, Macromol. Chem., the impregnation process. Rapid Commun., 1996, 17, 123. 25 H. W. Chen and C. S. Chen, J. Phys. Chem., 1995, 99, 10557. H.-W. Chen and J.-H. Lin would like to express their sincere appreciation to the National Science Council of Republic of Paper 8/03359E J. Mater. Chem., 1998, 8(10), 2169–2173 2173

ISSN:0959-9428    

DOI:10.1039/a803359e

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 Vibrational spectra and thin solid films of a bi(propylperylenediimide) S. Rodriguez-Llorente,a R. Aroca*a and J. Duffb aMaterials and Surface Science Group. University of Windsor, Windsor, ON, Canada N9B 3P4 bXerox Research Centre of Canada. 2660 Speakman Drive, Mississauga, ON, Canada L5K 2L1 Received 24th April 1998, Accepted 19th July 1998 The synthesis, the fabrication of thin solid films and the optical spectra of a member of a new series of bis-perylene materials—the bi(propylperylenediimide), N¾,N¾-dipropyl-N,N-bi[perylene-3,4:9,10-bis(dicarboximide)] (Pr2)— are reported.The objective of the present research eVort is to provide the fundamental spectroscopic characterization of the ground electronic state and the methods for the fabrication of submicron thin solid films of these new materials which are being tested for performing elementary electrical functions of electronic devices.Pr2 is a high molecular weight perylene dye and the molecular organization formed in vacuum evaporated nanometric films is one of the key elements determining the electrical and optical properties of thin solid films. The spectroscopic techniques used in the study of fundamental vibrational modes were transmission and reflection–absorption infrared spectroscopy (RAIRS). The inelastic scattering was investigated using out-ofresonance excitation (FT-Raman), resonance Raman scattering (RRS) and surface-enhanced resonance Raman scattering (SERRS) on silver island films.The vibrational assignment of characteristic wavenumbers was assisted with semi-empirical quantum mechanical calculations. son.14 Here, a new member of the bisperylene family is studied Introduction for the first time.Two diimide perylenes are joined by the The tendency to self-alignment observed in vacuum evaporated imide groups with no substituent chain in between. The Pr2 thin solid films of large aromatic molecules such as phthalo- molecule is shown in Fig. 1. cyanines and perylene tetracarboxylic (PTC) derivatives has been known for several decades,1–3 however our understanding Experimental of the mechanism of film formation and growth is still far from complete. It is agreed that molecular thin solid films are Synthesis of N¾,N¾-dipropyl-N,N-bi[perylene-3,4:9,10- assemblies of organic molecules bound by weak van der Waals bis(dicarboximide)] (Pr2) forces.In aromatic molecules, flatness of the molecular plane To a suspension of 4.76 g (0,011 mole) of perylene-3,4,9, and an extensive p system appear to be suYcient conditions 10-tetracarboxylic acid monoanhydride mono(n-propyl ) for self-ordering.4 It is found that molecular solids may exhibit imide15,16 in 200 ml of 1-methylpyrrolidin-2-one, stirred under various crystals forms that arise from diVerent molecular an argon atmosphere, was added 0.16 g (157 ml, 0.0050 mole) stacking arrangements.5 It is also known that the optical and of anhydrous hydrazine. The mixture was stirred at room electrical properties of organic materials are highly dependent temperature for 30 min then was heated to reflux (202 °C).on these crystal forms. Currently, ultrathin films of these After 4.5 h the reaction mixture was cooled with stirring to materials are being intensively investigated to study the optim- 160 °C then was filtered through a 11 cm glass-fiber filter ization of their optical and electronic properties for potential (Whatman 934AH) in a preheated porcelain funnel.The solid use in a variety of technological applications. In fact, high was washed on the funnel with 5×100 ml portions of boiling quality ultrathin films of perylene derivatives have been predimethylformamide. The initial filtrate was a dark brown pared using the Organic Molecular Beam Deposition (OMBD) technique which produces thin films with a good crystalline structure in ultra high vacuum.The objective of this work has been to obtain well-ordered organic films on inorganic singlecrystalline substrate (epitaxial growth) and to investigate the relation between substrate orientation and film structure. Of the various PTC materials, most of the OMBD work has been focused on the growth of 3,4,9,10-perylenetetracarboxylic dianhydride (PTCDA).6–10 Recently, Armstrong et al.11 have grown ultrathin films of PTCDA and N,N¾-di-n-butylperylene- 3,4:9,10-bis(dicarboximide) on (001) alkali halide surfaces.A theoretical study of a perylene derivative on gold (111) has also been reported.12 Our research eVort is to investigate the electronic spectra, vibrational spectra and thin solid films of a new classes of PTC materials.We have previously reported the film fabrication and spectral properties of 1,3-bis(3-chlorobenzylimidoperylenyl )propane (BPTDPr), a molecule where the two PTC groups are attached to the ends of a propane molecule.13 The work was complemented with the study of the eVect of the chain length (propane versus octane) of the bisperylene in the molecular organization and aggregation in Fig. 1 Chemical structure of the bi(propylperylenediimide) (Pr2) in the solid films: the characterization of 1,8-bis(3-chlorobenzyl- the optimum geometry as calculated with AM1 semiempirical quantum calculations. imidoperylenyl )octane (BPTDOc) was reported for compari- J. Mater. Chem., 1998, 8(10), 2175–2179 2175color. The final hot wash was light orange. The solid was sponding vibronic structure.The absorption and emission spectra of Pr2 solutions in chloroform are shown in Fig. 2. washed with 4×25 ml portions of methanol then was dried at 60 °C to give 3.2 g (74%) of fine grey-black solid, mp 428 °C The absorption spectrum shows the characteristic vibronic structure associated with the p–p* transition of the perylene (decomp.) (by DSC); 1H NMR (300 MHz, ppm, 3:1 CDCl3–trifluoroacetic acid-d, internal TMS): 1.03 (t, 3H, moiety, with a 0–0 band at 526 nm followed by band maxima at 490 nm, 457 nm and 436 nm.A concentration dependent CH3), 1.84 (m, 2H, CH2), 4.25 (t, 2H, N-CH2), 8.89 (m, 8H, aryl-H ). Anal. calcd. for C54H30N4O8: C 75.17, H 3.51, N broad band appears at around 642 nm, probably due to aggregation in solution. The extinction coeYcients obtained 6.50; found: C 73.60, H 3.50, N 6.42.from the absorbance vs. concentration measurements in solu- Thin film preparation tion are shown in Table 1. The fluorescence spectra of the same solutions yielded a The ability of bisperylene materials to form thin films by mirror image of the absorption spectrum with maxima at thermal evaporation without decomposition has been demon- 540 nm, 582 nm and a weak band at ca. 635 nm. strated in previous work.13,14 Pr2 films of diVerent thickness The absorption and emission spectra of a 120 nm evaporated were deposited onto a variety of substrates: transparent and film of Pr2 on glass are similar to those of the KBr dispersed pre-cleaned borosilicate slides (Baxter Cat. M6145) for visible pellet and are also shown in Fig. 2. The absorption spectrum absorption and fluorescence work (they were cleaned by of the solid film shows band broadening with maxima close rubbing with absolute ethanol, and subsequent drying under to the monomer absorption at 509 nm and a red shifted a continuous flow of dry nitrogen gas), polished KBr discs for maximum at 557 nm. The emission obtained with 514.5 nm transmission FTIR and 100 nm mass thickness of silver on excitation is typical of an excimer emitter with a maximum at glass for reflection absorption infrared spectroscopy. Silver 688 nm and the same results are obtained for a KBr diluted islands of 6 nm mass thickness were used as substrate for pellet of Pr2 with a maximum at 697 nm.It is now known SERRS experiments. that for thin solid films thicker than 2 monolayers the predomi- Metal deposition was performed in a Balzers vacuum system nant emission is that of excimers.11 The molecular stacking evaporator, equipped with an Edwards E2M2 rotary vacuum and orientational distribution have a significant eVect on the pump which functions as precursor to the Edwards diVusion spectral profile (bandwidth and symmetry) that explain the pump. Silver shots (Aldrich 20,436-6) were thermally evapor- diVerences between the thin solid film and pellet spectra.ated from a cupped tungsten boat, using a Balzers BSV 080 Complementary information on the electronic spectra were glow discharge/evaporation unit. The background pressure obtained through computations carried out using the AM1 was nominally 10-6 Torr, as measured by the Balzers IKR semi-empirical Hamiltonian as implemented in HyperChem 5, 020 cold cathode gauge.On a substrate preheated to 200 °C, with 7 occupied and 7 unoccupied orbitals used in the consilver was deposited at a rate of 0.05 nm s-1 to a total mass figuration interaction. The geometry of the molecule was first thickness of 100 nm. This deposition rate was allowed to fully optimized using AM1 in Gaussian 94, with a convergence stabilize before the shutter was opened. Film thickness and limit of 0.01 kcal mol-1.The outcome was a C1 symmetry deposition rates were monitored using a XTC Inficon quartz molecule with the two macrocycles perpendicular to each other crystal oscillator. The bulk density of silver employed was (see Fig. 1) and parameters (distances and derivatives) close 10.5 g cm-3, the tooling factor 105%, and the Z-ratio 0.529. to standard values. All frontier orbitals for the Pr2 molecule The organic dye was deposited onto the diVerent substrates are p type. The HOMO is w157, with eHOMO =-8.88 eV, in a diVerent evaporation system with the same setup afore- whereas the LUMO is w158, with eLUMO=-2.51 eV, which mentioned, with the exception that the substrate was not gives a band gap of 6.37 eV, similar to that calculated for heated.The perylene pigment was evaporated from tantalum monoperylene derivatives.17,18 In conclusion, the energy miniboats and the deposition rate for Pr2 was about 0.2 nm s-1. The bulk density employed for the Pr2 material was 1 g cm-3, the tooling factor 92%, and the Z-ratio 0.1.The solutions of the dye were prepared using chloroform (spectroscopic grade, Aldrich) as solvent. The fluorescence and Raman spectra were recorded using the 514.5 nm line of the Ar+ laser as an excitation source and both a THR 1000 spectrograph with a liquid nitrogen cooled CCD detector and the Ramanor U1000 double monochromator spectrometer.All spectra were recorded in a back-scattering geometry using a microscope attachment and a ×100 objective. SERRS spectra were also recorded using the 633 nm He–Ne laser line in the Renishaw system 2000. The infrared spectra were recorded with 1 cm-1 resolution using a Bomem DA3 FTIR equipped with a liquid nitrogen cooled MCT detector. The FT-Raman spectra were measured on a Bomem Ramspec 150 spectrophotometer with an Nd5YAG laser emit- Fig. 2 Absorption and emission spectra of Pr2 solutions in CHCl3 and ting at 1064.1 nm and equipped with a InGaAs detector. The of a Pr2 120 nm evaporated film on glass. The excitation wavelength Raman spectra were recorded with a spectral resolution of of the fluorescence spectrum is the Ar+ laser line at 514.5 nm. The y- 4 cm-1.Semi-empirical quantum chemical calculations, for axis has arbitrary units. geometry and frequency optimization, were carried out using AM1 basis sets on Gaussian 94 and HyperChem 5.0. Table 1 Results and discussion Band center/nm CoeYcient/dm3 mol-1 cm-1 Electronic spectra 436 1.11×103 457 2.69×103 The absorption spectra of PTC derivatives have been recently 490 1.17×104 calculated and assigned17,18 The observed visible absorption 526 1.63×104 spectrum consists of one electronic transition with the corre- 2176 J.Mater. Chem., 1998, 8(10), 2175–2179mization gives a non planar system with the two planar PTC and 853 cm-1 in good agreement with the observations. The moieties perpendicular to each other. At the same time, the assignment of the in-plane and out-of-plane modes of the emission spectra clearly indicate that excimer formation takes perylene moiety provides the basis for the discussion of the place producing a broad band fluorescence (Fig. 2). The average molecular orientation of the chromophores in the question remains about the long range organization in thin evaporated film. The assignment of characteristic infrared solid films of the material.bands is presented in Table 2, while Table 3 compiles observed Raman wavenumbers, relative intensities and bandwidths for Vibrational spectra the Pr2 molecule. Notably, the middle infrared and FT-Raman spectra of Pr2 are relatively simple. The large number of modes The vibrational spectra of the Pr2 material is presented in are seen in the far-infrared region, which is a consequence of Fig. 3. The y-axis is in arbitrary units to allow Raman and very similar infrared cross sections for these highly coupled infrared spectra in one graph. The reference spectrum for the deformation vibrations. The observed far-IR wavenumbers assignment of the IR allowed normal modes is that of an are: 86, 137, 187, 219, 388, 436, 466, 494, 515, 540, 575, 591 isotropic dispersion of Pr2 in KBr shown in Fig. 3. Since the and 605 cm-1. The latter three wavenumbers are also seen in KBr pellet spectrum represents a random spatial distribution the middle infrared spectrum with very weak relative intensity. of monomer and aggregates it is also the reference for the The 219 cm-1 and the 540 cm-1 wavenumbers are observed molecular organization in thin solid films. For a molecular in the resonance Raman spectrum.system with 282 normal modes of vibration, it is reasonable The oV resonance FT-Raman spectrum of the bulk in a to limit the discussion of normal mode assignment to a few pellet and the low region of Resonance Raman Scattering characteristic in-plane and out-of-plane vibrations of the (RRS) of a 120 nm evaporated film are presented in Fig. 3.chromophore moiety.19–21 The computation of fundamental Insignificant diVerences are observed between the two spectra. vibrational wavenumbers using the AM1 was successful and Both the resonant and the non-resonant spectra are completely all eigenvalues were positive. The most characteristic in-plane determined by the vibrational modes of the perylene chromo- molecular vibrations of the PTC group of the Pr2 molecule phore.The characteristic vibrational modes of the perylene are the carbonyl stretching modes: the antisymmetric stretching moiety are observed in the FT-Raman at 1300 cm-1, at 1656 cm-1 and the symmetric CNO stretching at 1379 cm-1, 1574 cm-1 and 1590 cm-1. However, only the FT- 1694 cm-1. The symmetric mode is also observed in the FTRaman spectrum shows the symmetric carbonyl stretching Raman spectrum at 1694 cm-1.The dynamic dipoles of the band at 1694 cm-1. The variation in the bandwidth of Raman symmetric and antisymmetric carbonyl stretches are in the bands may be explained by band overlapping due to the fact PTC plane and they are perpendicular to each other, which that the two PTC moieties that form the Pr2 compound have makes them excellent probes for orientational studies.22 The vibrational modes that are accidentally degenerate in calculated infrared intensities predicted three (accidentally wavenumber.degenerate) strong CNO fundamentals, and in fact three CNO bands can be identified in the IR spectrum and listed in Table 2. The strongest in-plane perylene CNC stretching Infrared spectra and molecular organization vibration is observed at 1594 cm-1 as seen in Fig. 3. A weak band at 1579 cm-1 is also a CNC stretch of the perylene ring. Vibrational spectroscopy provides several observables that Both of these modes are clearly seen in the FT-Raman may be used to extract physical information from thin solid spectrum with their relative intensity reversed (see Fig. 3). The films: the resonance band position, the band shape, and the computed spectrum predicts three strong IR bands in this band intensity. It is this wealth of information that makes region at 1590, 1558 and 1450 cm-1. The out-of-plane wagging vibrational spectroscopy a unique optical probe for determivibrations of the perylene ring have a dynamic dipole perpen- nation of the overall structure, chemical, mechanical and dicular to the perylene plane and thereby perpendicular to localized intermolecular interactions in thick and submicron components of the dipole moment derivatives of the carbonyl thin solid films.In particular, the intensity in the vibrational or CNC stretching vibrations. The out-of-plane modes are spectrum of thin solid films can be explained in terms of the observed with characteristic wavenumber and relative intensity polarization of the incoming infrared beam and the direcat 738 cm-1 and 810 cm-1, and they are assigned to the C–H tionality of the dipole moment derivatives.For instance, the wagging vibrations. The calculated IR spectrum predicts two selection rules operating in the specular reflection absorption strong (accidentally degenerate) bands in this region at 753 infrared spectroscopy (RAIRS) have been extensively used to determine the orientation of nanometric organic films thermally deposited on a reflecting metal surface.23 Transmission and reflection spectroscopic techniques can be used concurrently to enable a definitive account of the molecular orientation of perylene molecules.22 The molecular orientation on a film creates a film anisotropy that can be extracted using the change in relative intensity observed in the film’s spectra recorded in the transmission geometry.In Fig. 4, the transmission IR spectrum of a 100 nm Pr2 film deposited onto KBr is compared with the reflectionabsorption IR spectrum of a 100 nm Pr2 film deposited onto a smooth silver film (100 nm).In the transmission geometry, the polarization of the incident radiation always lies along the substrate surface (shown in Fig. 4). For the vacuum evaporated film, with a particular molecular orientation in the direction perpendicular to the substrate, only bands with a change in dipole moment parallel to the substrate surface will be intense.Infrared bands corresponding to a change in dipole moment Fig. 3 Vibrational spectra of Pr2. FT-Raman of the solid and low perpendicular to the substrate surface will not absorb. In the wavenumber region of the RRS spectrum of Pr2. FTIR spectrum of RAIRS case, the reflecting surface polarizes the light perpen- Pr2 dispersed in a KBr pellet and far-IR spectrum in polyethylene windows.The y-axis has arbitrary units. dicular to the surface plane. Therefore, the intensity of J. Mater. Chem., 1998, 8(10), 2175–2179 2177Table 2 Observed IR wavenumbers (cm-1), relative intensities (in parentheses) and full width at half maximum (FWHM) in cm-1 for Pr2 Pellet Film RAIRS cm-1 FWHM cm-1 FWHM cm-1 FWHM Assignment 487 (3) 6 Ring deformations 580, 592, 605 656 (16) 9 Ring deformations 738 (33) 8 742 (5) 8 739 (32) 9 C–H per.wag 754 (4) 12 753 (2) 10 752 (12) C–H per. wag 794 (13) 5 795 (11) 5 795 (7) 6 alkyl bending 810 (41) 5 811 (13) 5 811 (37) 5 C–H per. wag 851 (7) 5 852 (4) 13 851 (6) 9 C–H per. wag 856 (8) 5 966 (2) 17 950 (2) 7 952 (2) C–H bending 1025 (19) 20 1023 (1) 15 C–H bending 1047 (6) 10 1044 (1) 17 C–H bending 1082 (17) 20 1084 (6) 12 1084 (1) C–H bending 1173 (5) 14 1184 (6) 13 1183 (1) C–H bending 1249 (30) 19 1243 (7) 13 1248 (10) 14 Alkyl bending 1270 (4) 17 1271 (33) 20 1272 (4) 14 Alkyl bending 1287 (3) 8 1280 (5) 15 Alkyl bending 1324 (22) 11 1323 (10) 15 Alkyl bending 1351 (16) 12 1345 (45) 18 1346 (19) 14 C–N stretch 1359 (18) 10 1358 (19) 8 1357 (31) 17 C–N stretch 1375 (7) 13 1369 (9) 13 1375 (10) 17 C–N stretch 1404 (31) 7 1402 (21) 6 1404 (23) 7 Ring stretch 1438 (13) 11 1437 (14) 13 1438 (16) 20 CH2 scissors 1579 (24) 17 1578 (32) 15 1580 (10) CNC Ring stretch 1594 (100) 8 1594 (82) 9 1596 (83) 8 CNC Ring stretch 1644 (19) 17 1643 (12) CNO stretch 1656 (77) 17 1656 (33) 16 1660 (71) 19 CNO stretch 1694 (64) 18 1694 (100) 21 1696 (100) 20 CNO stretch Table 3 Observed Raman wavenumbers (cm-1), relative intensities (in parentheses) and full width at half maximum (FWHM) in cm-1 for Pr2 FT-Raman RRS 514.5 nm SERRS 514.5 nm SERRS 633 nm cm-1 FWHM cm-1 FWHM cm-1 FWHM cm-1 FWHM Assignment 544 (7) 9 542 (20) 11 551 (18) 19 551 (3) 20 Per. deformation 562 (25) 11 573 (1) 576 (3) 8 Alkyl deformation 1049 (3) 13 1064 (13) 13 C–H bend 1074 (4) 12 1074 (6) 1075 (6) 16 1076 (3) 20 C–H bend 1098 (5) 11 C–H bend 1300 (100) 12 1300 (86) 17 1295 (84) 19 1294 (100) 24 C–H bend+ring str. 1319 (7) 24 1312 (18) 32 1308 (20) CH2 bend 1379 (71) 13 1377 (100) 24 1376 (100) 20 1382 (65) 19 Per. ring str. 1455 (17) 10 1454 (10) 11 1452 (16) 20 1452 (12) 16 Per ring str. 1574 (98) 9 1571 (77) 11 1571 (31) 6 1574 (84) 18 CNC stretch 1590 (65) 14 1587 (57) 21 1598 (10) 30 1601 (6) 30 CNC stretch 1612 (6) 16 1610 (10) 21 CNC stretch 1694 (16) 10 1696 (4) 23 CNO stretch vibrational modes with a dynamic dipole perpendicular to the surface are enhanced and those modes parallel to the surface are suppressed.23 No substantial diVerences were observed in the relative intensities of the out-of plane modes observed in the RAIRS spectrum and the spectrum of the KBr reference (Fig. 3 and 4), allowing us to conclude that there is no preferential face-on long range organization in the direction perpendicular to the plane of the substrate. There is, however an increase in the relative intensity of the symmetric CNO stretching band (1696 cm-1) with a dipole moment derivative component along the long molecular axis. The latter may be interpreted as a slight tilted (head-on) long range organization in the film.The thin solid film evaporated onto the KBr surface ( lower trace in Fig. 4) produced a spectrum that represents a minor deviation from the reference. It can be seen that the out-of-plane modes at 739 cm-1 and 810 cm-1 are observed with a decreased relative intensity when compared with the symmetric CNO stretch (1694) in the transmission Fig. 4 Infrared transmission and reflection–absorption spectra of a spectrum. The latter may be interpreted as indication of some 100 nm Pr2 film deposited onto a KBr disc and a 100 nm Ag smooth film, respectively. face-on molecular orientation, a diVerent alignment from what 2178 J. Mater. Chem., 1998, 8(10), 2175–2179molecule are not very diVerent from those of the monoperylene tetracarboxylic derivatives.Conclusion The member of a new class of perylene derivatives (bisperylenes), N¾,N¾-dipropyl-N,N-bi[perylene-3,4:9,10-bis(dicarboximide)] has been characterised by optical spectroscopy. It has been demonstrated that thin solid films of this material can be easily fabricated by vacuum thermal evaporation. The energy minimization gives a non planar system with the two planar PTC moieties perpendicular to each other.The molecular structure with perpendicular planes does not favor molecular stacking and correspondingly the infrared spectra of films Fig. 5 Surface enhanced resonant Raman spectrum of a 12 nm Pr2 show evidence of marginal molecular organization in the film on 6 nm Ag film with a plasmon absorption at 510 nm.Upper perpendicular direction of thin solid films formed on silver or trace is the SERRS spectrum with 514.5 nm excitation wavelength KBr. This behavior is in contrast with the common tendency and the lower trace is the SERRS spectrum with He–Ne 633 nm to stacking and molecular alignment observed in perylene excitation wavelength. The y-axis units are arbitrary. tetracarboxylic derivatives.However, the formation of excimers in solid films was clearly established by the presence of was extracted from the film on smooth metal surface. However, the characteristic broad band in the fluorescence spectrum. the degree of this long range organization is far less than what was observed for 1,8-bis(3-chlorobenzylimidoperylenyl )octane NSERC of Canada is gratefully acknowledged for financial (BPTDOc) where an octyl group is placed between the two support.PTC moieties. The lack of a strong molecular alignment in Pr2 can only be explained by the findings of the geometry optimization indicating that the two planar PTC groups are References perpendicular to each other. Such molecular geometry seems 1 P. E. Burrows, Y. Zhang, E. I.Haskal and S. R. Forrest, Appl. also to be predominant in the solid state. Phys. Lett., 1992, 61, 2417. 2 A. L. Moser, Phthalocyanine Research and Applications, CRC Surface-enhanced resonant Raman scattering Press, Boca Raton, 1990. 3 Phthalocyanines Properties and Applications, ed. C. C. LeznoV and In Fig. 5, the surface-enhanced resonance Raman scattering A. B. P. Lever, VCH Publishers, Inc., New York, 1989.(SERRS) spectra of a 12 nm film of Pr2 deposited onto a 4 C. Taliani and L. M. Blinov, Adv. Mater., 1996, 8, 353. 6 nm silver island film and excited with both the Ar+ 514.5 nm 5 C. A. Jennings, R. Aroca, G. J. Kovacs and C. Hsaio, J. Raman Spectrosc., 1996, 27, 867. and the He–Ne 633 nm laser lines, are shown. The SERRS 6 A. Hoshino, S. Isoda, H. Kurata and T.Kobayashi, J. Appl. spectra are simply the amplified version of the RRS spectrum. Phys., 1994, 76, 4113. Both spectra are entirely due to the PTC moiety. Notably, the 7 C. Ludwig, B. Gompf, W. Glatz, J. Petersen, W. Eisenmenger, carbonyl stretches seen in the FT-Raman are not observed in M. Mobus, U. Zimmermann and N. Karl, Z. Phys. B, 1992, 86, the RRS or SERRS spectra of Pr2. The SERRS spectra as 397.well as the FT-Raman spectrum are very simple for such a 8 C. Ludwig, B. Gompf, W. Glatz, J. Petersen and W. Eisenmenger, Z. Phys. B, 1994, 93, 365. large molecular system and contain less than 10 characteristic 9 T. Schmitz-Hubsch, T. Fritz, F. Sellam, R. Staub and K.Leo, fundamentals. One important diVerence between the SERRS Phys. Rev. B, 1997, 55, 7972.spectra excited at 514.5 nm and 633 nm is the relative intensity 10 C. M. Fisher, M. Burghard, S. Roth and K. v. Klitzing, Europhys. of the combinations and overtones. From Fig. 2, it can be Lett., 1994, 28, 129. seen that the 514.5 nm laser line is in full resonance with the 11 D. Schlettwein, A. Back, B. Scilling, T. Fritz and absorption spectrum of the nanometric film and it contains N.R. Armstrong, Chem. Mater., 1998, 10, 601. 12 D. Lamoen, P. Ballone and M. Piranello, Phys. Rev. B, 1966, combinations and overtones with a strong relative intensity. 54, 5097. The spectrum obtained with the 633 nm excitation is a pre- 13 S. Rodriguez-Llorente, R. Aroca, J. DuV and J. A. DeSaja, Thin resonance Raman spectrum, and a low relative intensity of Solid Films, 1998, 317, 6037.overtones and combination bands is consistent with it. 14 S. Rodrý�guez-Llorente, R. Aroca and J . DuV, J. Mater. Chem., Appreciable relative intensity of combinations and overtones 1998, 8, 629. is only achieved in the RRS spectrum and they are not 15 H. Troster, Dyes Pigments, 1983, 4, 171. 16 E. Spietscha and H. Troster, US Patent 4 709 129 (November, observed in the FT-Raman spectrum. The relative intensity 1987). diVerences between the Red-SERRS (He–Ne excitation) and 17 R. Mercadante, M. Trsic, J. DuV and R. Aroca, J. Mol. Struct. the Green-SERRS (Ar+ excitation) are minimal (with the (Theochem), 1997, 394, 215. exception of the combinations and overtones) and the observed 18 M. Adachi, Y. Murata and S. Nakamura, J. Phys. Chem., 1995, wavenumbers are the same in RRS and SERRS. It is then 99, 14 240. concluded that the Pr2 is physically adsorbed on the silver 19 R. Aroca, A. K. Maiti and Y. Nagao, J. Raman Spectrosc., 1993, 24, 227. islands, a fact of considerable importance for its possible 20 A. K. Maiti, R. Aroca and Y. Nagao, J. Raman Spectrosc., 1993, practical analytical applications. It is also evident from the 24, 351. two spectra shown in Fig. 5 that the bands observed in the 21 R. Aroca and R. E. Clavijo, Spectrochim. Acta, 1991, 47A, 271. 2900–3100 cm-1 region correspond to overtones and combi- 22 E. Johnson and R. Aroca, Appl. Spectrosc., 1995, 49, 472. nations, while the C–H stretching vibrations are not seen in 23 M. K. Debe, Prog. Surf. Sci., 1987, 24, 1-282. the spectrum. In summary, the spontaneous Raman scattering (FT-Raman), the RRS and the SERRS spectra of the Pr2 Paper 8/03180K J. Mater. Chem., 1998, 8(10), 2175–2179 21

ISSN:0959-9428    

DOI:10.1039/a803180k

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 Synthesis and characterization of an organic–inorganic hybrid compound: [WO3 (2,2¾-bipy)] (2,2¾-bipy=2,2¾-bipyridine) Jen Twu,a* Ting-Hua Fang,a Chun-Fu Hsu,a Yang-Yann Yu,a Gwang-Jung Wang,a Chih-Wei Tang,a Kuei-Hsien Chenb and Kwang-Hwa Liic aDepartment of Chemistry, Chinese Culture University, Yang Ming Shan, Taipei, Taiwan, ROC bInstitute of Atomic and Molecular Science, and cInstitute of Chemistry, Academia Sinica, Taipei, Taiwan, ROC Received 1st June 1998, Accepted 22nd July 1998 The hydrothermal synthesis and characterization of an organic–inorganic hybrid compound [WO3(2,2¾-bipy)] (2,2¾- bipy=2,2¾-bipyridine) is presented.This compound was characterized by various techniques including Raman and IR spectroscopy, XRD and thermogravimetric analysis.Structural analysis of the compound by Raman and IR data shows that it has a novel tungstate structure with 2,2¾-bipy coordinated to tungsten. The X-ray powder diVraction data of the compound are indexed to a monoclinic cell with a=14.313(4), b=9.603(2), c=7.288(2) A° , b=102.07(2)°. The structure of [WO3(2,2¾-bipy)] adopts a chain-like framework consisting of corner-shared distorted octahedra (WO4N2) with 2,2¾-bipy acting as a bidentate ligand whose rings fan out from the chain.Upon heating, [WO3(2,2¾-bipy)] retains its structural integrity until 300 °C, followed by a gradual loss of 2,2¾-bipy which leads to the formation of WO3 at ca. 475 °C. The thermal decomposition is observed to proceed through a transitional intermediate, [W4O12(2,2¾-bipy)], the structure of which is proposed and discussed.Introduction Experimental Preparation of [WO3(2,2¾-bipy)] Recently, hydrothermal synthesis of novel metal oxides, using structure-directing agents, has attracted considerable interest All the samples used in this study were obtained from commerowing to their potential applications in catalysis and material cial suppliers and used without further purification.H2WO4 science as well as their rich structural chemistry.1–3 For (Strem), 2,2¾-bipy (Fluka) and distilled water were first mixed, example, layered vanadium and molybdenum salts are formed in a molar ratio of 1525700, in a 23 ml Teflon-lined acid using surfactant as structure-directing templates, under digestion bomb (Hun Kung, Taiwan) and then heated at hydrothermal conditions.4 By contrast, tungstate salts with 160 °C for 96 h under autogeneous pressure.The resultant Keggin-ion structure, e.g. [(C12H25NMe3)6(H2W12O40)] and white solid (yield 80% based on tungsten) was filtered, washed [(C16H36NMe3)6(H2W12O40)], are formed with surfactant ions. with water and then air-dried for 12 h before subjecting it to Attempts to remove the surfactants from these tungstate further analytical studies.salts by thermal treatment result in formation of tungsten Product characterization oxides.3,4 Among the published studies so far, the role of aromatic Powder X-ray diVraction data were recorded at room temperaamines appears to be relatively unexplored, and only a few ture on a Siemens D5000 h–h diVractometer using Bragg– studies involving molybdenum oxide or vanadium oxide, lead- Brentano geometry with back-monochromatized Cu-Ka radiing to compounds [MoO3(2,2¾-bipy)], [Mo2O6(2,2¾-bipy)], and ation.The diVraction pattern was scanned in steps of 0.02° [Mo3O9(2,2¾-bipy)3];5 [VO(VO3)6(VO(2,2¾-bipy)2)2]6 and (2h) over the 2h range 5–60° and a counting time of 5 s per [V3O7(1,10-phen)]7 (2,2¾-bipy=2,2¾-bipyridine, 1,10-phen= step.The peak positions in the powder pattern were determined 1,10-phenanthroline) have been reported. No report with by means of the peak search routine in the PC software tungsten oxide has appeared in the literature. Furthermore, package DIFFAC-AT. The Raman spectrometer used in this details concerning structural evolution upon removal of the study was a Renishaw system 2000 micro-Raman spectrometer aromatic amines from these vanadium and molybdenum com- equipped with a 25 mW He–Ne laser, operating at 632 nm, acting as the excitation source.A 5 mm exit slit oVered a pounds are, as yet, unclarified.5–7 resolution of better than 1 cm-1. In-situ Raman studies were Raman spectroscopy has been used not only for the characcarried out at a heating rate of 3 °Cmin-1 by use of a Linkam terization of a wide range of novel materials,9–14 but also to THM 600 heating cell up to the designated temperatures in delineate molecular speciation of tungstate ions in aqueous air.FTIR spectroscopy was performed on a Perkin Elmer solution and on surfaces of supported WO3 catalysts.12–21 Paragon 1000 IR spectrometer with a resolution of 4 cm-1, Moreover, 2,2¾-bipy is also known to display distinct Raman using KBr pellets.Thermogravimetry was conducted on a TA features exhibited by various geometrical configurations.22–27 Instrument 2950 TG–DTA thermal analyzer using a heating Thus, Raman spectroscopy is capable of discerning molecular rate of 3 °Cmin-1 in air. Chemical analysis for C, H and N details of the synthetic, structural and mechanistic chemistry was performed using an Heraeus elemental analyzer (Found: involving tungstates and 2,2¾-bipy.C, 30.66; H, 2.02; N, 6.99. Calc. for C10H8N2O3W: C, 30.95; Here, we report on the synthesis and characterization of a H, 2.08; N, 7.22%). novel organic–inorganic hybrid compound, [WO3(2,2¾-bipy)], as well as its structural evolution upon removal of organic Results and discussions template upon heating.This study is part of a research program aimed at the synthesis of novel organic–inorganic Fig. 1 shows the Raman spectra of solid 2,2¾-bipy, H2WO4 and [WO3(2,2¾-bipy)] in the range 400–1700 cm-1 at room hybrid compounds via molecular architecture.28,29 J. Mater. Chem., 1998, 8(10), 2181–2184 2181possibility of 2,2¾-bipy being adsorbed on the surface of WO3 can be excluded due to the complete absence of Raman features from WO3,18,19 despite the fact that H2WO4 transforms into WO3 at around 160 °C.31 Thus, Fig. 1(c) and 2 suggest the formation of a novel organic–inorganic hybrid compound which has the formula [WO3(2,2-bipy)] as determined by elemental analysis.In order to explore the possibility of synthesis of compounds with diVerent stoichiometries, a variety of synthetic variations were attempted. Despite adjusting the relative ratio of H2WO4 and 2,2¾-bipy, varying the reaction duration and temperature (in the range 140–180 °C) or substituting H2WO4 by WO3, only [WO3(2,2¾- bipy)] was formed under hydrothermal conditions, as identified by Raman spectroscopy. By contrast, [MoO3(2,2¾-bipy)] and [Mo3O9(2,2¾-bipy)2] could be synthesized by reacting MoO3 with 2,2¾-bipy whereas the synthesis of [Mo2O6(2,2¾-bipy)2] required the participation of cations.5 The indexing of the X-ray diVraction powder pattern of [WO3(2,2¾-bipy)] was performed by using the program TREOR30 which indicated a monoclinic system.The refined Fig. 1 Raman spectra of (a) solid 2,2¾-bipy, (b) H2WO4 and (c) cell parameters were: a=14.313(4), b=9.603(2), c= [WO3(2,2¾-bipy)]. 7.288(2) A° , b=102.07(2) ° and V=979.5 A° 3. The corresponding figures of merit are M20=16 and F20=31(0.0112, 58). The cell parameters are close to those of [MoO3(2,2¾-bipy)]5 and temperature. The Raman spectrum of 2,2¾-bipy [Fig. 1(a)] the indexed powder pattern is given in Table 1.The observed exhibits strong bands at 992, 1570 and 1588 cm-1, medium pattern of [WO3(2,2¾-bipy)] is in excellent agreement with the bands at 1234, 1299, 1445 and 1480 cm-1, and weak bands at calculated powder pattern derived from the cell parameters 437, 610, 762, 811, 1042, 1090, 1143, 1215, 1289 and and the atomic coordinates of [MoO3(2,2¾-bipy)],5 indicating 1306 cm-1.22–24 The Raman spectrum of H2WO4 [Fig. 1(b)] that the bulk product is monophasic and the two compounds exhibits a broad band centered at 637 cm -1 and a sharp band are isostructural. Thus, a coordination environment with two at 942 cm-1.25 The product, [WO3(2,2¾-bipy)] [Fig. 1(c)], terminal WNO bonds, two bridging WMO bonds and two shows strong bands at 930, 1022, 1316 and 1595 cm-1, together WMN bonds, namely WO4N2, is ascribed to tungsten atom.with medium and weak bands at 630, 650, 767, 870, 883, 1055, Each octahedron establishes corner-sharing connections with 1160, 1266 and 1290 cm-1. The distinct Raman features of two adjacent octahedra, through two bridging WMO bonds 2,2¾-bipy, appearing in Fig. 1(a) and (c) (except for bands at (WMOMW), to form an infinite chain-like framework with 870, 883 and 930 cm-1), are known to be associated with a 2,2¾-bipy fanning out along the chain.transition from the trans form (C2h) for the solid compound In addition to XRD results, information concerning the to the cis form (C2v) in coordinated 2,2¾-bipy.24,26,27 The three coordination of tungsten and framework structure of Raman bands not associated with 2,2¾-bipy, WO3 and H2WO4 (at 870, 883 and 930 cm-1), can be ascribed to WMO vibrations of the product.Fig. 2 shows the IR spectrum of Table 1 Powder X-ray data of [WO3(2,2¾-bipy)] [WO3(2,2¾-bipy)] in the range 500–1500 cm-1 at room temperature. Characteristic bands of H2WO4, a weak band at h k l Iobs (%) dobs/A° dcalc a/A° 940 cm-1 and a strong, broad band at 668 cm-1,25 are not observed.In addition to bands due to 2,2¾-bipy, strong bands 1 1 0 100.00 7.905 7.919 2 0 0 24.09 6.993 6.998 at 934, 886, 617 cm-1 and a weak band at 860 cm-1, are 0 2 0 13.19 4.797 4.802 observed and can be ascribed to WMO vibrations of the 3 1 0 23.15 4.197 4.197 product. Comparison of Fig. 1(c) and Fig. 2 with Raman 2 2 0 14.76 3.958 3.959 and/or IR features of H2W12O406- and WO42--related com- -2 2 1 1.51 3.647 3.649 pounds as well as a wide range of tungstates, in the range for 0 0 2 12.56 3.560 3.563 characteristic WMO vibrations (600–1100 cm-1), indicates 4 0 0 6.60 3.496 3.499 -1 1 2 16.85 3.401 3.404 that a novel tungstate structure is formed.15–21 Moreover, the 3 1 1 1.06 3.352 3.351 2 2 1 1.60 3.299 3.299 1 3 0 6.82 3.118 3.120 -3 1 2 5.69 3.009 3.010 2 0 2 3.63 2.936 2.937 -1 3 1 1.00 2.908 2.909 0 2 2 4.73 2.861 2.862 4 2 0 4.23 2.826 2.828 5 1 0 6.31 2.689 2.688 3 3 0 5.08 2.640 2.640 -3 3 1 1.74 2.576 2.576 3 1 2 3.62 2.494 2.495 -4 2 2 2.93 2.423 2.424 0 4 0 1.83 2.401 2.401 -5 1 2 2.54 2.390 2.389 6 0 0 2.50 2.334 2.333 1 3 2 2.22 2.230 2.230 2 4 0 2.34 2.272 2.271 3 3 2 2.32 2.252 2.252 aMonoclinic, a=14.313(4), b=9.603(2), c=7.288(2) A° , b= 102.07(2)°, l=1.5406 A° .Fig. 2 IR spectrum of [WO3(2,2¾-bipy)]. 2182 J. Mater. Chem., 1998, 8(10), 2181–2184[WO3(2,2¾-bipy)] can also be obtained through analysis of vibrational features reported for a wide range of tungstates in the literature.19–21 Geometries such as an isolated octahedron (Ba2WO6, R: 818 cm -1 and IR: 620 cm-1), a corner-shared octahedron (WO3, R: 808 cm-1 and IR: 870 cm-1) and an isolated tetrahedron (Na2WO4, R: 930 cm-1 and IR: 840 cm-1) can be ruled out.19,20 A distorted octahedral geometry with terminal WNO bonds, as found in CdWO4 (R: 896 cm-1 and IR: 884 cm-1) and H2WO4 (R: 948 cm-1 and IR: 948 cm-1), appears more likely.20,27 Raman bands at 930 and 883 cm-1 as well as IR bands at 934 and 886 cm-1 can be assigned to symmetric and asymmetric WNO stretchings19 –21 while the Raman band at 870 cm-1 and IR band at 860 cm-1 can be assigned to WMO stretching,19–21 whereas the broad IR band at 617 cm-1 is assigned to WMOMW stretching.21,25 In addition, three medium bands at 236, 260 and 360 cm-1, and two weak bands at 300 and 320 cm-1 are also observed in the range 200–400 cm-1.However, owing to their overlap with WMN stretching,26,27 WMO bending, WMOMW deformation,19–21 plus C–N and C–C torsion modes,28 detailed assignment of these medium and weak bands cannot be made. Fig. 4 In situ Raman spectra of [WO3(2,2¾-bipy)] at (a) 300 °C, (b) Fig. 3 shows the results of TG–DTG of [WO3(2,2-bipy)], 350 °C, (c) 400 °C and (d) 475 °C, at a heating rate of 3 °Cmin-1 conducted at a heating rate of 3 °Cmin-1 in air.A sharp under ambient conditions. weight loss was observed between 300 °C and 450 °C with a plateau at 450 °C, corresponding to 60.5% of the original weight. This 39.5% weight loss corresponds to a stoichiometry tion of 2,2¾-bipy or a strong fluorescence background is observed below 300 °C when 2,2¾-bipy is adsorbed on surfaces of [WO3(2,2¾-bipy)0.97], assuming loss of bipy and formation of WO3 in good agreement with the formula determined by of metal oxide supports.32 It can be concluded that the stability of [WO3(2,2¾-bipy)] is dependent on that of 2,2¾-bipy (bp= chemical analysis.Additionally, a DTG inflection around 410 °C corresponding to 69.3% of the initial weight (Fig. 3), 273 °C) as well as the strength of bonding between the tungsten atom and 2,2¾-bipy. Any attempt to isolate [W4O12(2,2¾-bipy)] suggests the formation of an intermediate with stoichiometry [W4O12(2,2¾-bipy)] (cf. exptl. ratio W5bipy of 450.96) and in by careful thermal treament resulted only in the restoration of [WO3(2,2¾-bipy)], formation of WO3 and unspecified amorph- situ Raman studies were conducted to shed more light on this intermediate.Fig. 4 shows the in situ Raman spectra of ous species which were identified by thermogravimetric analysis, XRD and Raman spectroscopy (not presented here). The [WO3(2,2¾-bipy)] heated to diVerent temperatures, at a rate of 3 °Cmin-1 under ambient conditions. At 300 °C, Fig. 4(a) inability to isolate products other than [WO3(2,2¾-bipy)], either by thermal treatment or hydrothermal synthesis (vide supra), shows a spectrum almost identical to that in Fig. 1(c). At 350 °C [Fig. 4(b)], in addition to bands from [WO3(2,2¾- contrasts with the MoO3–2,2¾-bipy system for which stable [Mo2O6(2,2¾-bipy)] and [Mo3O9(2,2¾-bipy)2] have been hydro- bipy)], two new bands emerge at 960 and 990 cm-1. At 400 °C [Fig. 4(c)] these bands become prominent at the expense of thermally synthesized.5 These two compounds exhibit chainlike frameworks similar to [MoO3(2,2¾-bipy)] in which the the band at 930 cm-1. At 475 °C [Fig. 4(d)] only two broad bands, centered at 700 and 800 cm-1, can be observed indicat- coordination of the molybdenum atoms varies, viz. 151 and 251, corner-shared octahedra (MoO4N2) and tetrahedra ing complete transformation into WO3.18 The results of the in situ Raman study are consistent with the thermogravimetric (MoO4), respectively.5 Since the intensity ratios of the 960 cm-1 band vs.the 990 cm-1 band at 350 °C [Fig. 4(b)] results (Fig. 3) and provide a more detailed molecular picture. For example, the observation of an identical spectrum at and 400 °C [Fig. 4(c)] remain essentially unchanged they can be ascribed to one species rather than a mixture of unspecified 300 °C to that at room temperature provides unambiguous evidence for the structural integrity of [WO3(2,2¾-bipy)] at this intermediates, with a coincident stoichiometry of [W4O12(2,2¾- bipy)], the relative amount of which present would be tempera- temperature. In contrast, either a significant degree of desorpture dependent.Structural analyses on a wide range of tungstates indicate that the position of the Raman band with the highest wavenumber reflects, in general, the highest bond order in the sample.14–21 For samples containing both tetrahedral and octahedral tungstates such as Na2W2O7, assignments of characteristic bands for each geometry can be made.33 Based on the structures of [WO3(2,2¾-bipy)], [Mo2O6(2,2¾-bipy)] and [Mo3O9(2,2¾-bipy)2], the structure of [W4O12(2,2¾-bipy)]can be tentatively proposed as follows.It most likely exhibits a chain-like framework as does [WO3(2,2¾-bipy)] consisting of octahedra (WO4N2) and tetrahedra (WO4) in a 153 ratio. The band at 960 cm-1 can be assigned to a stretching vibration of WNO in octahedral geometry (WO4N2); and the band at 990 cm-1 can be assigned to a stretching vibration of WNO in tetrahedral geometry (WO4).The shift from 930 to 960 cm-1 for the WNO stretching associated with octahedral geometry (WO4N2) can be explained in terms of the weakening of WMN bonding caused by heating above 350 °C, whereas a Fig. 3 Thermogravimetric analysis (TG and DTG) of [WO3(2,2¾- bipy)] at a heating rate of 3 °Cmin-1 under ambient conditions. further shift from 960 to 990 cm-1 is a consequence of the J.Mater. Chem., 1998, 8(10), 2181–2184 21839 P. K. Dutta, P. K. Gallagher and J. Twu, Chem. Mater., 1993, complete absence of 2,2¾-bipy in tetrahedral geometry (WO4). 5, 1739. However, the detailed structure of [W4O12(2,2¾-bipy)] can not, 10 P. K.Dutta, P. K. Gallagher and J. Twu, Chem. Mater., 1992, as yet, be specified and more research is in progress. 4, 847. 11 J. Twu, P. K. Dutta and C. T. Kresge, J. Phys. Chem., 1991, 95, 5267. Conclusion 12 J. Twu and P. K. Dutta, J. Catal., 1990, 124, 503. 13 J. Twu and P. K. Dutta, Chem.Mater., 1992, 4, 398. Synthetic and X-ray structural details, vibrational and thermal 14 J.M. Stencel, Raman Spectroscopy For Catalysts, Van Nostrand properties of a novel organic–inorganic hybrid compound, Reinhold, New York, 1990. 15 J. Fuchs, R. Palm and H. Hartl, Angew. Chem., Int. Ed. Engl., [WO3(2,2¾-bipy)], are presented. The compound adopts a 1996, 35, 2651. chain-like framework consisting of corner-shared distorted 16 W. P. GriYth and T. D. Wickins, J. Chem.Soc. A, 1966, 1087. octahedra (WO4N2) with 2,2¾-bipy acting as a bidentate ligand 17 W. P. GriYth and T. D. Wickins, J. Chem. Soc. A, 1969, 1066. whose rings fan out from the chain. Upon thermal treatment, 18 F. D. Hardeastle and I. E. Wachs, J. Raman Spectrosc., 1995, [WO3(2,2¾-bipy)] retains its structural integrity below 300 °C, 26, 397. 19 J. A. Horsely, I. E. Waachs, J. M.Brown, G. H. Via and and a transitional intermediate [W4O12(2,2¾-bipy)] is identified F. D. Hardcastle, J. Phys. Chem., 1987, 91, 4014. around 410 °C before final conversion to WO3 around 475 °C. 20 M. Daturi, G. Busca, M. M. Borel, A. Leclaire and P. Piaggio, J. Phys. Chem. B, 1997, 101, 4358. 21 J. Hanuza, M. Maczka and J. H. van der Maas, J. Solid State This research was supported by the National Science Chem., 1995, 117, 177.Foundation of the ROC under contract no. NSC88–2113- 22 S. A. Bagshaw and R. P. Cooney, J. Mater. Chem., 1994, 4, 557. M-034–004. 23 J. A. Bartlett and R. P. Cooney, Spectrochim. Acta., Part A, 1987, 43, 1543. 24 D. A. Barker, L. A. Summers and R. P. Cooney, J. Mol. Struc., References 1987, 159, 249. 25 M. F. Daniel. B. Desbat, J.C. Lassegues, B. Gerand and 1 C. T. Kresge, M. E. Leonwicz, W. J. Roth, J. C. Vartuli and M. Friglarz, J. Solid State Chem., 1987, 67, 235. J. C. Beck, Nature, 1992, 359, 710 26 J. S. Struki and J. L. Walter, Spectrochim. Acta., Part A, 1971, 2 Q. S. Huo, D. I. Margolese, U. Ciesla, P. Y. Feng, T. E. Gier, 27, 223. P. Sieger, R. Leon, P. M. PetroV, F. Schuth and G. D. Stucky, 27 P. K. Dutta and J. A. Incavo, J. Phys. Chem., 1987, 91, 4443. Nature, 1994, 368, 317. 28 J. Twu, C. J. Chuang, K. I. Chang, C. H. Yang and K. H. Chen, 3 A. Stein, M. Fendorf, T. P. Jarvie, K. T. Muller, A. J. Benesi and Appl. Catal. B: Environ., 1997, 12, 309. T. E. Mallouk, Chem. Mater., 1995, 7, 304. 29 J. Twu, C. F. Shih, T. H. Guo and K. H. Chen, J. Mater. Chem., 4 G. G. Janauer, A. Dobley, J. Guo, P. Zavalij and 1997, 7, 2273. 30 P. E. Werner, L. Erikson and M. Westdhal, J. Appl. Crystallogr., M. S. Whittingham, Chem.Mater., 1996, 8, 2096. 1985, 18, 367 5 P. J. Zapf, R. C. Haushalter and J. Zubieta, Chem. Mater., 1997, 31 G. Ramis, C. Cristina, A. S. Elmi and P. Villa, J. Mol. Catal., 9, 2019. 1990, 61, 319. 6 G. H. Huan, J. W. Johnson, A. J. Johnson and J. S. J. Merola, J. 32 N. S. P. Bhuvanesh, S. Uma, G. N. Subbanna and Solid State Chem., 1991, 91, 385. J. Gopalakrishnan, J. Mater. Chem., 1995, 5, 927. 7 C. Y. Duan, Y. P. Tian, Z. L. Liu, X. Z. You and X. Y. Huang, 33 V. H. J. Becher, Z. Anorg. Allg. Chem., 1981, 474, 63. Inorg. Chem., 1995, 34, 1. 8 M. T. Pope, Heteropoly and Isopoly Oxometallates, Springer, Berlin, 1983. Paper 8/04068K 2184 J. Mater. Chem., 1998, 8(10), 2181–2184

ISSN:0959-9428    

DOI:10.1039/a804068k

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 The role of a silane coupling agent in the synthesis of hybrid polypyrrole–silica gel conducting particles Christian Perruchot,a Mohamed M. Chehimi,*a Delphine Mordenti,b Michel Brianda and Michel Delamara aInstitut de Topologie et de Dynamique des Syste`mes (ITODYS), Universite� Paris 7-Denis Diderot, CNRS (UPRESA 7086), 1 rue Guy de la Brosse, 75005 Paris, France.E-mail: chehimi@paris7.jussieu.fr bLaboratoire de Re�activite� de Surface, Universite� Pierre et Marie Curie, CNRS (URA 1106), tour 54, 2e`me e� tage, 4 Place Jussieu, 75252 Paris Cedex 05, France Received 22nd April 1998, Accepted 7th July 1998 The preparation of new hybrid conducting polymer–silica gel particles is described. The silica gel acts as a high surface area substrate (431 m2 g-1) for the in situ chemical synthesis of conducting polypyrrole in aqueous solution in order to obtain hybrid polypyrrole–silica particles.The role of a common silane coupling agent (i.e. aminopropyltriethoxysilane, APS) in the pretreatment of silica gel prior to polymerization and preparation of polypyrrole–APS–silica particles is also investigated.It was found by TGA that the polypyrrole mass loading is higher in polypyrrole–APS–silica than in polypyrrole–silica particles. XPS results indicated that APS-treated silica leads to polypyrrole-rich surface particles not found with the untreated silica. Consequently, the polypyrrole–APS–silica pellets were three orders of magnitude more conductive than those of polypyrrole–silica.The surface area of the polypyrrole–silica (422 m2 g-1), as measured by BET, matched that of the untreated silica whilst that of the polypyrrole–APS–silica (162–184 m2 g-1) is significantly lower.The combination of XPS, TGA, BET and conductivity measurements suggest that pyrrole is predominantly polymerized in the pores of the untreated silica gel whilst the APS pretreatment of silica leads to the formation of a thin overlayer of polypyrrole at the surface of the silica gel in addition to a higher conducting polymer loading in the gel pores.ticles generally exhibit good long term conductivity and chemi- Introduction cal stability. Inherent conducting polymers (ICP) have attracted a great Of relevance to the present work, various forms of hybrid deal of interest owing to their remarkable physical and chemi- inorganic/organic conducting polymer composites were precal properties, such as redox,1,2 acid–base,3–5 ion exchange pared using a metal oxide as a supporting substrate.24–28 For properties,6 and chemical sensing,7–11 in addition to their high example, Maeda and Armes28 have described the synthesis of conductivity.12 Polypyrrole (PPy) is one of the most studied polypyrrole in the presence of ultrafine silica particles in aqueous conducting polymers due to the ease of its electrochemical or media.The ultrafine silica sol acts as a high surface area chemical synthesis in high yield via oxidative polymerisation colloidal substrate for the precipitating polypyrrole leading to at room temperature in various common solvents, including unusual raspberry-shaped polypyrrole–silica nanocomposites water.Furthemore, polypyrrole has fairly good environmental which exhibit long term colloidal stability in water.28 Although stability with regard to air and water.2 However, bulk polypyr- they have a deep black color as bulk polypyrrole, the surface role is infusible, intractable and insoluble in common solvents of these nanocomposites was shown to be silica rich by means which seriously limits its processability.Polypyrrole is also of XPS29 and has an isoelectric point (IEP) at pH 2, matching known to be partly crosslinked13 and suVers poor mechanical that of pure silica.30 Armes and coworkers30 have thus grafted properties. For these reasons, polypyrrole can not, for example, aminopropyltriethoxysilane (i.e.APS) on the surface silanol be solvent cast to produce homogeneous films. To overcome groups of these nanocomposites yielding amino-functionalized these limitations, the preparation of conducting polypyrrole- polypyrrole–silica nanocomposites (APS–PPy–silica) with an based polymer blends,14–17 sterically stabilised colloids17–23 IEP at pH 7.This surface modification was of biological and composite materials24–28 has received increasing interest importance as Saoudi et al.31 found a strong DNA (negatively for it can be an alternative towards more processability. The charged) adsorption onto APS–PPy–silica at neutral pH benefit of such composite materials is the synergistic whereas the untreated PPy–silica nanocomposite had poor combination of the properties of both components.bioadsorptivity towards DNA. The preparation of latexes and sterically stabilised colloidal Wallace and coworkers described the synthesis of both particles of conducting polymers (especially PPy and polyani- PPy-and PANI-modified silica gel particles and their chromatoline, PANI) is well documented.14–23 In 1987, Yassar et al.14 graphic properties were examined by HPLC.32,33 It was shown reported that chemically synthesized polypyrrole could be that these particles behave as a typical reverse stationnary deposited in situ onto spherical polystyrene (PS) latex in phase.However, since these conducting polymer-modified aqueous solution to yield monodisperse PPy–PS particles. silica gels were lacking surface characterization, one can not Armes and coworkers18–20 synthesized sub-micrometer colloid fully interpret the interaction of the analytes with the surface particles of PPy or PANI using various commercial polymers of the stationary phase which governs solid–liquid chromatogor tailor-made copolymers as polymeric stabiliser which raphy.Given the publications of Armes and coworkers,28,29 it becomes either physically adsorbed or chemically grafted onto is interesting to prepare and characterize the surface of such the surface of the precipitating conducting polymer particles, polypyrrole–silica gel particles and check whether or not they are silica or polypyrrole rich.In the case where polypyrrole- producing an interpenetrating polymer network.16 These par- J.Mater. Chem., 1998, 8(10), 2185–2193 2185modified silica gel particles are silica rich, the retention data Synthesis of polypyrrole powders (PPyTS). Pyrrole (1.00 ml, 14.4 mmol) was added via syringe to 100 ml of a stirred of polyaromatic hydrocarbons (PAHs) and other solutes would reflect mixed retention mechanisms due to both the silica gel aqueous solution containing FeCl3·6H2O (9.74 g, 36.0 mmol) and sodium p-toluenesulfonate (7.07 g, 36.0 mmol) at room and the conducting polymer moiety.Indeed, mixed retention mechanisms can occur in all forms of partitioning chromatog- temperature.40 The oxidant-to-pyrrole molar ratio was 2.5, close to the optimal value recommended by Armes.41 The raphy because of the possibility of adsorption on the underlying support.34,35 reaction solution was stirred for 24 h and the resulting black precipitate was vacuum-filtered and washed with copious Faverolle et al.36,37 have demonstated that polypyrrole could be also deposited onto E-glass fibres (14 mm diameter, 300 amounts of de-ionized water until the washings were clear.The powder was then dried in a desiccator overnight and fibres per lock) by oxidative polymerisation. Thereby, in order to increase conducting polypyrrole adhesion to E-glass fibres sieved to 180 mm diameter before being analysed.in a multicomponent system, Faverolle et al.36 have thus pretreated the supporting glass fibres by APS and by a pyrrole- Synthesis of aminopropyltriethoxysilane (APS)-grafted silica functionalized silane coupling agent prior to pyrrole polymeriz- gel particles. 2 ml of APS coupling agent were first hydrolysed ation. Silanes are known to play an important role in modern in a 200 ml water–ethanol (1/9 v/v)39 solution for 6 h and then reinforced plastics enhancing the adhesion between substrate 2 g of bare silre added to the solution and stirred and matrix resin.38,39 It was clearly demonstrated that the overnight.This solution was then Bu� chner-filtered, rinsed with silane coupling agent was eVective in increasing the polypyrro- 50 ml of water–ethanol (1/9 v/v) solution to remove the excess le–glass adhesion on the one hand, and the adhesion between of physically adsorbed silane coupling agent and then dried the as-prepared conducting E-glass fibre and the insulating overnight in a desiccator.polymeric matrix, on the other, this being more marked for the latter. Moreover, scanning electron microscopy (SEM) Synthesis of polypyrrole-silica particles (PPyTS–silica). The examination of the surface morphology of coated E-glass synthesis of this material was made in two steps. Purified fibres shows that the pretreatment by these silane coupling pyrrole was coated on the bare silica gel particles from pentane agents leads to a more regular and homogeneous conducting before proceeding to the oxidative polymerisation of the polypyrrole overlayer than the untreated one.former. To do so, 1 ml of pyrrole was added to a suspension Here, we describe the use of aminopropyltriethoxysilane of 2 g of silica particles in 50 ml of pentane and the mixture (APS) in the preparation of polypyrrole-coated silica gel parwas stirred in a fumehood until free flowing pyrrole-coated ticles (PPy–APS–silica) with the aim of obtaining a polypyrrolesilica particles were obtained.These particles were then added rich hybrid particles surface. This procedure will be compared to 100 ml of an aqueous solution of FeCl3·6H2O (9.74 g, to the method published by Wallace and coworkers32,33 in the 36.0 mmol) and sodium p-toluenesulfonate (7.07 g, preparation of polypyrrole-coated silica gel particles 36.0 mmol) at room temperature.This solution was stirred (PPy–silica) without the use of a silane coupling agent to modify for 24 h and the resulting black polypyrrole–silica particles the silica particles prior to pyrrole polymerization.Our approach were vacuum-filtered and washed with copious amounts of de- diVers from that of Armes and coworkers30 as these authors ionized water until the washings were clear. The powder was used APS to functionalize the polypyrrole–silica nanocomthen dried in a desiccator overnight. The mass of the dried posites, but is rather comparable to that of Faverolle et al.36,37 end-product was ca. 2.5 g. for the preparation of polypyrrole-coated E-glass fibres to give It is important to note that this procedure is more eVective elaborate novel conducting composites. in producing hybrid polypyrrole–silica particles than the direct We have synthesized polypyrrole in the presence of untreated method consisting of oxidizing pyrrole in the presence of silica and APS-treated silica gel particles using either chloride or gel.Indeed, the latter procedure produced both bulk polypyr- p-toluenesulfonate anion dopants. role powder and polypyrrole–silica particles. In addition, The bulk and surface physico-chemical properties of polypyrrole was also coated on the walls of the reaction vessel PPy–APS–silica particles and the reference PPy–silica, leading thus to a much lower yield of hybrid material.We will silica–APS, untreated silica gel and bulk polypyrrole powders thus not report the hybrid materials synthesized so. were determined by elemental analysis, thermogravimetric analysis (TGA), X-ray photoelectron spectroscopy (XPS), BET and the four probe resistivity measurements. In particu- Synthesis of polypyrrole-coated, APS-grafted silica particles lar, the comparison between bulk and surface chemical com- (PPyTS–APS–silica).The APS coupling agent was first positions will be emphasised. grafted onto the silica particles as described above before pyrrole coating and polymerization. Experimental Thermogravimetric analysis (TGA) Materials Thermogravimetric analyses (TGA) were performed using a The silica gel particles (diameter in the range 60–125 mm, Setaram 92-12 TGA–DGA analyser.Each sample was heated Vporous=0.75 cm3 g-1) was provided by Merck and used as from room temperature to 800 °C (1073 K) with a scan rate of received. Aminopropyltriethoxysilane (APS, Acros) was 5 °Cmin-1 and then held isothermally for 3 h. The combustion degassed by N2 before silica pretreatment.De-ionized water was carried out under an air flow rate of 1 ml min-1. and ethanol (Prolabo, 95%) were used as solvent. Pyrrole (Acros) was purified by passing through a column of Pellet conductivity measurements activated basic alumina prior to polymerization. Iron chloride (FeCl3·6H2O) and sodium p-toluenesulfonate Compressed pellets of 16 mm diameter were made with 60 mg (CH3C6H4SO3Na) were obtained from Aldrich and were of each sample.The room temperature conductivity was employed without further purification. measured with a conventional four line probe resistivity measurements apparatus placed on the flat surface of the Synthesis of hybrid and reference materials pellets under test. The current (I/mA) passed through the two outer electrodes and the floating potential (V /mV) was meas- The synthesis of chloride-doped polypyrrole (PPyCl) bulk powder and hybrid materials was carried out as described ured across the inner pair.For a thin pellet of hybrid material or PPy bulk powder, the conductivity of the sample is given below except in the absence of any sodium p-toluenesulfonate. 2186 J.Mater. Chem., 1998, 8(10), 2185–2193by eqn. (1) Elemental analysis Elemental analysis is important to determine the N/Si ratio as s/S cm-1= ln(2)I V pd (1) a function of the APS pretreatment of the host silica gel. It also permits control of the doping level of the conducting where I is the current intensity (mA), V the measured potential polymer moiety in the hybrid materials under test.Table 1 (mV), and d the thickness of the pellet. reports elemental analysis of hybrid polypyrrole–silica materials and the reference silica–APS in wt.%. We have also added Specific surface area measurements (BET) the global N/Si molar ratio. However, in the case of PPy–APS–silica particles, since APS is a nitrogen- and a Specific surface areas (As) were measured with a Quantasorb silicon-containing chemical, these elements contribute to the Jr sorptometer using the BET specific surface area procedure.42 bulk composition and thus aVect the NPPy/Sisilica molar ratio.The samples (20–40 mg) were cooled to liquid nitrogen (77 K) The contribution of nitrogen and silicon from APS to the temperature under a flow of N2–He (30570) and were then total nitrogen and silicon contents of PPy–APS–silica will thus heated to room temperature. The amount of desorbed nitrogen be determined in order to compare the polypyrrole loading in was measured by a thermoconductivity detector and allowed PPy–APS–silica to those of the corresponding PPy–silica.to determine As for each sample. In order to convert weight fractions into molar fraction, we divided the weight fraction of each element by its atomic X-Ray photoelectron spectroscopy (XPS) weight.From Table 1 one can determine the C/N molar ratio XPS signals were recorded using a VG Scientific ESCALAB for PPyTS–silica and PPyCl–silica. The values of these ratios MKI system operated in the constant analyzer energy mode. are 6.17 and 4.19, respectively. The latter is consistent with An Al-Ka X-ray source was used at a power of 200 W polypyrrole backbone whereas the former is higher due to the (20 mA×10 kV) and the pass energy was set at 20 eV.The contribution of the p-toluenesulfonate dopant. As far as the pressure in the analysis chamber was ca. 5×10-8 mbar. Digital doping level is concerned, for chloride dopant, the Cl/N molar acquisition was achieved with a Cybernetix system and the ratio is 22.2%, and for p-toluenesulfonate dopant, the S/N data collected with a personal computer.A home-made data molar ratio is 25.3% and yields approximately 1.8 carbon processing software allowed smooing, linear or Shirley back- atom of dopant for each nitrogen atom. It follows that the ground removal, static charge referencing, peak fitting and C/N ratio in the PPyTS backbone is approximately 4.3.It quantification. Charge referencing was determined by setting should be noted that PPyTS is also doped by chloride anions. the C 1s from the adventitious carbon contamination due to The detection of iron suggests the insertion of FeCl4- anions CMC/CMH component at 285.0 eV in the case of silica and as demonstrated by XPS below.The doping level of hybrid APS. As far as polypyrrole-based materials are concerned, we PPy–silica particles is in good agreement with those obtained preferred to calibrate the spectra by setting the main N 1s for bulk polypyrrole powders. component ( largely due to PPy) at 399.7 eV.43 For PPy–APS–silica particles, both APS and polypyrrole The surface compositions (in atom%) of the various samples contribute to the nitrogen content.However, if the N/Si molar were determined by considering the integrated peak areas of ratio for PPy–APS–silica can be rationalized in a first C 1s (1), N 1s (1.6), O 1s (3.1), Si 2p (1.2), S 2p (2.4), Cl 2p approximation, by eqn. (3) (2.8) and Fe 2p3/2 (7.8) and their respective experimental (N/Si)PPy–APS–silica=(N/Si)silica–APS+(N/Si)PPy–silica (3) sensitivity factors shown in parentheses.The sensitivity factors were determined using a large set of organic and inorganic it follows that the nitrogen content of PPy–APS–silica due to compounds of well defined stoichiometries. the polypyrrole moiety (ca. 4.9% for p-toluenesulfonate and Values of %A, the fractional concentration of a particular 11.0% for chloride anion dopant) is slightly higher than that element A is computed using eqn.(2) in PPy–silica, an indication of a higher mass loading of polypyrrole (for either anion dopants) when silica is pretreated %A= IA/sA S(In/sn) ×100% (2) by APS prior to pyrrole polymerisation. This will be demonstrated further by TGA. where In and sn are the integrated peak areas and the sensitivity factors, respectively.Thermogravimetric analysis (TGA) Fig. 1 shows the thermograms of PPyTS–silica, Results and discussion PPyTS–APS–silica and the reference materials silica and silica–APS. The thermograms have roughly similar shapes at The results reported in this work concern both the bulk and surface physico-chemical properties of the hybrid polypyrrole– high temperature showing a plateau value of weight fraction vs.temperature. The final mass loss for each specimen and silica particles and their reference materials, silica, silica–APS and polypyrrole bulk powders. Elemental analysis and TGA corresponding to the plateau regions of the thermograms are reported in Table 2. concern the bulk compositions whereas XPS and BET are relevant to the surface characteristics.We shall show how The mass loss of bare silica gel is fairly low and known to be due to removal of adsorbed water and adventitious hydro- conductivity measurements can be linked to the surface compositions of the hybrid materials under test. carbon contamination, the latter being usually detected by Table 1 Elemental analysis (wt.%) of the hybrid materials PPyX–APS–silica, PPyX–silica and the reference silica–APS Materials C H N O Si S Cl Fe N/Sia PPyTS–silica 4.02 1.47 0.76 55.46 37.15 0.44 0.20 0.50 4.1 PPyCl–silica 4.85 0.99 1.35 52.65 38.7 — 0.76 0.70 7.0 silica–APS 7.44 2.18 2.36 50.27 37.75 — — — 12.5 PPyTS–APS–silica 11.38 2.49 2.58 49.8 29.7 1.28 2.42 0.35 17.4 PPyCl–APS–silica 11.79 2.37 3.74 44.17 31.95 — 5.38 0.60 23.5 aMolar ratio in %.J. Mater. Chem., 1998, 8(10), 2185–2193 2187XPS Wide scans. Fig. 2(a)–(e) depicts XPS survey scans of silica, PPyTS–silica, silica–APS, PPyTS–APS–silica and bulk PPyTS, respectively. Fig. 2(b)–(d) are dominated by silica as shown by the three intense Si 2p, Si 2s and O 1s peaks. The wide scan shown in Fig. 2(b) for PPyTS–silica is similar to those reported by Maeda et al.29 for polypyrrole-silica nanocomposites and reflects a silica-rich material surface.Indeed, despite the deep black color of this material, the surface exhibits a C 1s peak which is relatively less intense than the Si 2p from the host silica, and a tiny N 1s feature due to polypyrrole. For silica–APS [Fig. 2(c)], the N 1s peak from the pendent aminopropyl group is substantially more intense than that of Fig. 1 Thermogravimetric analysis of bare silica gel (a), polypyrrole– PPyTS–silica [Fig. 2(b)], in agreement with the elemental silica particles (b), APS-treated silica gel particles (c) and analyses and the thermograms depicted in Fig. 1. polypyrrole–APS–silica particles (d). When silica–APS is used as a host substrate for pyrrole polymerisation, the as-prepared material PPyTS–APS–silica Table 2 TGA determination of the total mass loss of hybrid [Fig. 2(d)] depicts a significant increase in both the relative PPyX–APS–silica and PPyX–silica particles and the reference intensity of the C 1s and N 1s peaks from the polypyrrole materials silica, silica–APS and bulk PPy powders moiety. Moreover, the S 2p and Cl 2p peaks due to the ptoluenesulfonate and the co-dopant chloride, respectively, TGA mass loading (%) become distinguishible from the noisy background despite their relative low concentrations. Dopant Material TS Cl The sharp increase in the N 1s/Si 2p intensity ratio on going from PPyTS–silica or silica–APS to PPyTS–APS–silica clearly silica 5.76 5.76 demonstrates that the use of this silane coupling agent is PPyX–silica 11.3 8.35 eVective in producing a hybrid polypyrrole–silica material with silica–APS 14.1 13.4 a much more polypyrrole rich surface.Quantitative evidence PPyX–APS–silica 22.6 23.0 for this is given below. PPyX 90 —a aNot determined, presumed to be comparable to that of PPyTS. The High resolution scans. C 1s. Fig. 3(a)–(c) shows the high mass loss of PPyTS is not 100% owing to the inorganic nature of the resolution C 1s signal for silica–APS, PPyTS–silica and residue due to the oxidizing agent and the dopants and therefore containing most probably sulfonate, chloride and iron (detected PPyTS–APS–silica, respectively. For silica–APS [Fig. 3(a)], by XPS). the C 1s peak was fitted to three components centred at 283.9, 285 and 285.9 eV, respectively, with a relative intensity of ca. 15151 which can be attributed to the three diVerent carbon XPS. At high temperature, dehydration by condensation of atoms in the APS molecule. The three components are due to adjacent surface silanol groups to siloxane may occur. The SiCH2C, CCH2C and CH2NH2 from lowest to highest binding thermogram of PPyTS–silica shows a higher mass loss than energy.The C 1s feature depicted in Fig. 3(a) for silica–APS that obtained with silica which indicates that polypyrrole compares closely with that usually reported in the literature polymerized onto the bare silica gel. The weight loss of for e.g. an E-glass slide surface coated by APS.44–46 silica–APS is significantly higher than that of bare silica and For PPyTS–silica [Fig. 3(b)], although only a weak signal is mainly due to the removal of the APS moiety and is also is recorded, the C 1s signal has a shape that is frequently higher than that of PPyTS–silica, an indication that grafting reported for bulk polypyrrole.47–50 This complex signal was APS onto the bare silica is more massive than the sorption of fitted to five components due to the b carbon from polypyrrole PPyTS (and also PPyCl ).Finally, the hybrid PPyTS– and carbon atoms from the p-toluenesulfonate dopant APS–silica exhibits the lowest plateau value of weight fraction (285 eV), a carbon (286.6 eV) and higher binding energy and thus the highest mass loss. features at 288.5 and 290.1 eV which correspond to ‘disorder The determination of the mass loss due to polypyrrole in type’ carbon and terminal NCNO group.48 The high Eb PPyTS–silica and PPyTS–APS–silica is indicated (Fig. 1) by component centred at 292.2 eV is attributed to a pAp* shakedouble- headed arrows. It follows that the weight loss due to up transition.50 Whilst it is distinctly visible on the C 1s region the conducting polymer is more important for PPy–APS–silica of bulk PPyTS, it is rather obscured by the tailing of the signal than for PPy–silica. This direct evidence for a higher mass in the hybrid materials.loading of polypyrrole in the case of pretreated silica confirms For PPyTS–APS–silica [Fig. 3(c)], while both APS and thus the hypothesis based on elemental analysis. The silane polypyrrole contribute to the C 1s peak, the shape of this coupling agent is thus eVective in promoting a higher mass signal is roughly similar to that of bulk polypyrrole.For this loading of polypyrrole, and this is true for both p-toluenesul- reason we have peak-fitted the C 1s signal from fonate and chloride dopants, however to a lower extent for PPyTS–APS–silica in the same manner as for PPyTS–silica. the latter. This slight increase in polypyrrole mass loading is The change in the relative intensity of the five components probably due to favourable specific interactions of Lewis acid– due to APS is visible on the high binding energy side around base type between the basic amino group and the acidic NMH 287 eV as indicated by an arrow.bonds and the positively charged polypyrrole backbone, keep- Similar decomposition is also obtained for the C 1s signal ing in mind that APS is an adhesion promoter.38,39 One can with chloride anion dopant for both the bulk polypyrrole also consider the surface tension of polypyrrole and treated (PPyCl ) powder and the hybrid polypyrrole–silica particles silica to discuss the increase in the bulk and surface proportion (PPyCl–silica and PPyCl–APS–silica).of polypyrrole.However, since polypyrrole is a high surface It is notable that the C 1s peak fitting of both PPyX–silica energy material (>100 mJ m-2),5 the lower surface tension and PPyX–APS–silica are skewed and in addition they display of silica–APS certainly does not govern this increase in a significant increase of the background intensity (due to electron energy loss) that coincides with the trailing edge of polypyrrole mass loading. 2188 J. Mater. Chem., 1998, 8(10), 2185–2193Fig. 2 XPS survey scan of bare silica gel (a), polypyrrole–silica particles (b), silica-APS particles (c), polypyrrole–APS–silica particles (d) and bulk polypyrrole powder (e). the peak. This is characteristic of polpyrrole and thus clearly on the silica gel pretreatment. Fig. 5(a)–(c) shows the Si 2s–S 2p region for silica, PPyTS–silica and contrasts with the C 1s peak of silica–APS. PPyTS–APS–silica, respectively.The relative intensity of S 2p is high for PPyTS–APS–silica whereas for PPyTS–silica it is N 1s. Fig. 4(a)–(c) shows the N 1s signal for silica–APS, PPyTS–silica and PPyTS–APS–silica, respectively. In the case merely a shoulder on the background on the higher binding energy side of the Si 2s signal from silica.Since the XPS depth of silica–APS [Fig. 4(a)], the N 1s peak was fitted with two components, the major centred at 399.1 eV corresponds to the analysis is in the 5–10 nm range, the increasing S 2p signal intensity clearly demonstrates that the APS pretreatment leads free amine group from APS, whereas the minor component centred at 400.6 eV is due to its protonated form or to an to hybrid PPyTS–APS–silica particles with a higher polypyrrole-rich surface than the PPyTS–silica particles.This MSiOH,H2NM hydrogen bond. Several papers have appeared on such positively charged nitrogen from APS44,45,51 will be quantitatively shown below. and thus we shall not discuss it here. The N 1s peak from PPyTS–silica [Fig. 4(b)] is more Surface composition complex and fitted with four components due to the imine defects (398.2 eV), free NMH from pyrrole repeat units The apparent surface composition of polypyrrole–APS–silica, polypyrrole–silica and the reference materials are reported (399.7 eV) and two positively charged nitrogens centred at 401.7 and 403.8 eV, respectively, according to Kang et al.52 in Table 3.For bare silica gel particle, the O/Si ratio is slightly above Its shape is similar to the N 1s structure usually reported in the literature for bulk polypyrrole. 2 perhaps due to adsorbed water as shown by TGA, and a weak carbon contamination is also detected. For the PPyTS–APS–silica [Fig. 4(c)], the shape of the N 1s peak is similar to that of PPyTS–silica, however with a slightly In the case of hybrid polypyrrole–silica particles, the surface composition is mainly dominated by silica.Indeed, %O is more intense shoulder at high binding energy due to the APS contribution. Therefore, the APS does not aVect the structure 50–60% and %Si above 25% which is quite similar to bare silica gel, whereas %C is significantly higher than that of the of the polypyrrole backbone in PPy–APS–silica particles.contamination of silica but still in the 15–20% range, far below the carbon content of bulk PPys. The nitrogen content is Si 2s–S 2p. For p-toluenesulfonate-doped polypyrrole hybrid materials, the Si 2s region is followed by a S 2p signal from below 5% and even lower than that of APS-treated silica. Since both polypyrrole and APS contain nitrogen atoms, the p-toluenesulfonate dopant whose relative intensity depends J.Mater. Chem., 1998, 8(10), 2185–2193 2189Fig. 4 Characteristic N 1s core line signal of APS-treated silica gel Fig. 3 Characteristic C 1s core line signal of APS-treated silica gel particles (a), polypyrrole–silica particles (b) and polypyrrole– particles (a), polypyrrole–silica particles (b) and polypyrrole– APS–silica particles (c).APS–silica particles (c). of silica–PPyCl and slightly higher than that of bulk PPyCl, neither can be used as a specific elemental marker to distinguish possibly due to iron chloride contamination.49 between the conducting polymer and the silane coupling agent. It is worth noting the increase in the surface content of both However, there is a dramatic change in the carbon and p-toluenesulfonate and chloride anion dopants as a result of nitrogen contents at the surface of the hybrid materials in the the APS pretreatment of the silica gel prior to pyrrole case of polypyrrole–APS–silica particles compared to polypyr- polymerization. role–silica and silica–APS.Indeed, %C levels oV at ca. 50% Fig. 6 compares the bulk and surface N/Si atomic ratio as and the nitrogen is three to four times larger than for determined by elemental analysis and XPS, respectively, for polypyrrole–silica particles.PPyTS–silica, PPyTS–APS–silica and the reference silica–APS. It is notable that whether or not the silica is pretreated by The surface N/Si ratio is at best twice as high as that of the APS, the chloride anion dopant leads to higher carbon and bulk hybrid polypyrrole–silica whereas it is five times larger nitrogen surface contents whilst the TGA analysis indicated a for the surface than the bulk of PPyTS-APS-silica.Therefore, higher PPyTS mass loading by comparison to PPyCl. This Fig. 6 definitely shows that APS is eVective in coating the may be due to diVering surface morphologies of these con- surface of silica with polypyrrole, a situation which contrasts ducting polymers in/on the untreated and APS-treated silica with untreated silica gel particles.particles. The doping levels for PPyTS–silica and PPyCl–silica are Physicochemical properties of hybrid polypyrrole–silica and the similar to those of the corresponding bulk polypyrrole pow- reference materials ders.After APS pretreatment the S/N ratio is lower than that of PPyTS powder due to the APS contribution to N content. Physicochemical properties such as PPy surface relative proportion, surface static charge, specific surface areas and In contrast, the Cl/N for silica–APS–PPyCl is higher than that 2190 J. Mater. Chem., 1998, 8(10), 2185–2193Fig. 6 Bulk and surface N/Si atomic ratio as determined by elemental analysis and XPS for polypyrrole–silica particles, APS-treated silica particles and polypyrrole–APS–silica particles.by XPS analysis [eqn. (4)] %PPy= NPPy Sisilica+SiAPS+NPPy ×100 (4) where NPPy is the contribution of polypyrrole to the total nitrogen content, Sisilica and SiAPS are the silicon contributions arising from silica and APS, respectively, of the hybrid materials.However, since only silica and APS contribute to the total silicon content [eqn. (5)] Sitotal=Sisilica+SiAPS (5) In order to determine the contribution of polypyrrole, it is important to assess the contribution of APS to the total nitrogen content. The value of NPPy can be obtained using eqn. (6) NPPy=Ntotal-NAPS (6) where NAPS is estimated from the Si/N atomic ratio obtained for the pretreated silica–APS assuming that this ratio remains unchanged following polypyrrole sorption in and onto silica–APS.Therefore, the contribution of polypyrrole to the total nitrogen content is given by eqn. (7) Fig. 5 Si 2s–S 2p region for bare silica gel (a), PPyTS–silica particles NPPy=Ntot-(Sitot/r) (7) (b) and PPyTS–APS–silica particles (c).where r= Sisilica–APS Nsilica–APS (8) Table 3 Apparent surface compositions (atom%) of the hybrid PPyX–silica and PPyX–APS–silica materials, and the reference specimens silica, silica–APS and bulk PPyX powders as determined this is easily obtained from Table 3. by XPS Combining eqn. (5) and (7) in eqn. (4), one can determine the relative proportion of polypyrrole at the surface of the Material C N O Si S Cl D/N (%) hybrid materials: silica 4.8 67.9 27.3 PPyTS–silica 15.9 2.7 55.9 25.0 0.5 18.5 %PPy= Ntot-(Sitot/r) Sitot+Ntot-(Sitot/r) ×100 PPyCl–silica 18.3 3.5 53.4 23.6 0.8 22.9 silica–APS 18.0 4.6 52.0 25.4 The surface relative proportion of polypyrrole decreases in the PPyTS–APS–silica 48.7 9.8 27.2 11.1 1.5 15.3 order bulk PPy powder>PPy–APS–silica>PPy–silica.PPyCl–APS–silica 51.9 12.2 22.4 10 3.3 27.0 It is interesting that for an increase in polypyrrole mass PPyTS 75.9 12.6 9.1 2.2 17.5 PPyCl 73 18.6 3 5.2 28.0 loading by a factor of 1.5 as determined by TGA when using APS pretreatment, XPS indicates a four times higher apparent relative proportion of polypyrrole at the surface of the hybrid PPy–APS–silica compared to PPy–silica particles.This result confirms that the APS pretreatment is eVective in increasing electrical conductivity are reported in Table 4 for the various hybrid polypyrrole–silica particles and the reference materials the polypyrrole coating at the surface of the APS-treated silica gel particles for both anion dopants. silica, silica–APS and PPy bulk powders.The relative proportion of PPy at the surface of the various Another interesting result concerns the surface static charge built up by the samples during XPS analysis. The surface static hybrid PPy–silica and PPy–APS–silica particles was derived from the apparent surface composition (Table 3) as determined charge is reported in Table 4 for various hybrid particles and J. Mater. Chem., 1998, 8(10), 2185–2193 2191Table 4 Relative polypyrrole surface proportion, surface static charge, compressed pellets conductivity measurements and surface area measurements of hybrid polypyrrole–silica particles and reference materials XPS %PPy Static charge/eV s/S cm-1 BET/m2 g-1 Dopant Dopant Dopant Dopant Material TS Cl TS Cl TS Cl TS Cl silica 8.9 insulator 431 PPyX–silica 9.7 12.9 4.8 2.2 %10-5 422 421 silica–APS 6.8 insulator 222 234 PPyX–APS–silica 41.3 51.0 0.5 0.1 2.68×10-2 3.71×10-2 162 185 PPyX 100 1.6 23.02 7.13 15–25 the referencematerials.Silica is an electrically insulatingmaterial It is also interesting to compare the diVerences in the specific and this is reflected in the highest surface static charge. For the surface areas (DAS) of the hybrid end products and their hybrid materials, whilst PPyX–silica particles charge up, the corresponding host material: hybrid PPyX–APS–silica particles bear a quasi-neutral surface static charge, within the experimental error of binding energy AS(PPy–APS–silica)-AS(silica–APS)=50–60 m2 g-1 determination (±0.1–0.2 eV).This is a further evidence for a &AS(PPy–silica)-AS(silica)=9–10 m2 g-1 higher conducting polypyrrole overlayer at the outermost surface of the hybrid materials when APS-treated silica particles is Since PPy–silica and PPy–APS–silica have comparable used as host materials for pyrrole polymerisation. polypyrrole loadings but a much higher surface polypyrrole The XPS results indicating a polypyrrole-rich surface of the content of the latter, it is possible that the greater DAS for PPy–APS–silica particles and a silica-rich surface for the APS–containing materials is due to the higher surface pro- PPy–silica particles led us to investigate on the surface area portion of polypyrrole, an essentially non-porous material of silica induced by APS and/or polypyrrole. Table 4 reports having a specific surface area of 15–25 m2 g-1,54,55 one order the specific surface areas (AS) of the hybrid particles and of magnitude lower than those of the former materials.reference materials. The BET measurements show that the A further evidence for a higher polypyrrole proportion at the specific surface area of PPy–silica particles matched that of surface of the hybrid materials induced by the APS pretreatment the untreated silica gel, in contrast, the APS treatment of silica of silica gel is shown by the conductivity of compressed pellets yields a sharp decrease in the surface area of the latter.This of the various materials. These measurements indicate that eVect is exacerbated by polypyrrole loading. PPy–APS–silica is at least three to four orders of magnitude From the mass loading of polypyrrole as determined by more conductive than PPy–silica particles (the conductivity of TGA for PPy–silica, one can estimate the volume occupied by which is below the detection limit of our apparatus) although polypyrrole using a mean density of 1.5.31 We have found a these materials have comparable polypyrrole mass loadings.In volume of 3.7 and 1.7 cm3 for PPyTS and PPyCl, respectively, the case of PPy–APS–silica, the particles stick to each other via corresponding to the loadings of these polymers in 100 g of the surface conducting polymer coatings allowing the flow of PPy–silica particles.These volumes are much lower than the the electrical current via the polypyrrole overlayers. In contrast, porous volume corresponding to the remaining 90 g of bare for PPy–silica particles, polypyrrole is predominantly in the silica gel (determined from the thermograms).Given a surface pores therefore hindering an electrical conductivity via area of polypyrrole–silica matching that of silica, and an the insulating silica-rich surface of the hybrid particles. apparent low relative proportion of PPy at the surface of In this regard, it is worth noting that the p-toluenesulfonate polypyrrole–silica, it is clear that pyrrole is essentially anion dopant leads to more conductive PPy–APS–silica par- polymerized in the pores of the untreated silica gel.ticles than the chloride anion dopant, a trend which parallels For PPy–APS–silica particles, the APS-treated silica was used that obtained for bulk polypyrrole powders.This is also in as a host for pyrrole polymerization. Table 4 indicates a surface agreement with the results reported in the literature for both area of silica–APS twice lower than that of silica. Such a electrochemically prepared films or chemically synthesized decrease can only be obtained if the pores of the silica gel were bulk powders.2 either prefilled or blocked with APS.Added to this, APS can The electrical conductivity of the hybrid particles can be be grafted on the outermost layers of silica. This, however, linked to the surface static charge as determined by XPS. seems negligible as the surface nitrogen contents of silica–APS Thereby, the increase of the electrical conductivity and the and PPy–silica are similar. Combination of TGA and XPS decrease of the surface static charge is an indication that results suggests a sorption of APS into the pores of the silica adsorption of APS onto the silica gel favours the formation gel.However, for comparable mass loadings of APS and of a continuous conducting polypyrrole overlayer at the surface polypyrrole in silica–APS and PPy–silica particles respectively, of APS-treated silica particles in addition to the sorption of amuch higher surface area was observed for the latter.Therefore PPy in the pores partially filled by the APS. These results are APS would be sorbed in the pores but near the outermost in good agreement with those obtained by Faverolle et al.36,37 surface of silica. This situation is likely to be responsible for a in the case of PPy deposited onto APS-treated E-glass fibres.polypyrrole loading in the pore volume of silica still available The conductivity of these polypyrrole-modified fibres was after the APS treatment and at the surface sites modified by higher in the case of the coupling agent pretreatment which APS. This explains why, for comparable polypyrrole loadings, yielded homogeneous conductive polypyrrole overlayers. the surface of PPy–APS–silica has a much higher relative Given the results obtained in this multitechnique study, one proportion of polypyrrole than that of PPy–silica.can view the hybrid materials as illustrated in Fig. 7. Without For nanoporous silica gel particles pretreated by cany APS, PPy is essentially loaded in the silica gel pores methacryloxypropyltrimethoxysilane, Luo et al.53 have clearly whereas the APS treatment yields a substantially higher shown that the silica gel pretreatment entails a significant polypyrrole loading in the porous volume left available and decrease of both porosity and pore radii in addition to specific surface area compared to the untreated silica gel particles. at the outermost surface of the APS treated silica gel. 2192 J. Mater. Chem., 1998, 8(10), 2185–21939 P. R. Teasdale and G. G. Wallace, Analyst, 1993, 118, 329. 10 M. Josowicz, Analyst, 1995, 120, 1019. 11 S. B. Adeloju and G. G. Wallace, Analyst, 1996, 121, 699. 12 Conjugated Polymers and Related Materials, ed. W. R. Salaneck, I. Lundstro�m and B. Ramby, Oxford University Press, London, 1993. 13 H. Naarman, Synth.Met., 1991, 41–43, 1. 14 A. Yassar, J. Roncali and F. Garnier, Polym. Commun., 1987, 28, 103. 15 F. Epron, F. Henry and O. Sagnes, Makromol. Chem., Makromol. Symp., 1990, 35–36, 527. 16 S. Y. Luk, W. Lineton, M. Keane, C. DeArmitt and S. P. Armes, J. Chem. Soc., Faraday Trans., 1995, 91, 905. 17 D. H. Napper, Polymer Stabilization of Colloidal Dispersions, Academic Press, London, 1983. 18 S. P. Armes and M. Aldissi, Polymer, 1990, 31, 569. 19 C. DeArmitt and S. P. Armes, Langmuir, 1993, 9, 652. Fig. 7 Schematic illustration of the silica gel pore filling by poly- 20 P. M. Beadle, L. Rowan, J. Mykytiuk, N. C. Billinghan and pyrrole in polypyrrole–silica (a) and polypyrrole–APS–silica hybrid S. P. Armes, Polymer, 1993, 34, 1561. particles (b). 21 M. R. Simmons, P.A. Chaloner and S. P. Armes, Langmuir, 1995, 11, 222. Conclusion 22 S. P. Armes, M. Aldissi, S. Agnew and S. Gottesfeld, Langmuir, 1990, 6, 1745. The bulk and surface properties of hybrid polypyrrole–silica 23 C. DeArmitt and S. P. Armes, J. Colloid Interface Sci., 1992, particles have been investigated. In addition to its role of 150,134. 24 R. Partch, S. G. Gangolly, E. Matijevic, W.Cai and S. Arajs, promotor adhesion, it was clearly emphasised that the pretreat- J. Colloid Interface Sci., 1991, 144, 27. ment of the silica gel by a silane coupling agent, aminopro- 25 S. P. Armes, S. Gottesfeld, J. G. Beery, F. Garzon and pyltriethoxysilane (APS) prior to pyrrole polymerisation was S. F Agnew, Polymer, 1991, 32, 2325. eVective in producing higher polypyrrole mass loading but 26 S.Maeda and S. P. Armes, Chem.Mater., 1995, 7, 171. above all by increasing the proportion of the conducting 27 A. Bhattacharya, K. M. Ganguly, A. De and S. Sarkar, Mater. polymer at the surface of the hybrid materials. Indeed, the Res. Bull., 1996, 31, 527. 28 S. Maeda and S. P. Armes, J. Mater. Chem., 1994, 4, 935. APS pretreatment leads to a significant decrease of the silica 29 S.Maeda, M. Gill, S. P. Armes and I. W. Fletcher, Langmuir, gel surface area which favors the pyrrole polymerisation in 1995, 11, 1899. the pore volume of silica still available and at the surface 30 M. D. Butterworth, R. Corradi, J. Johal, S. F. Lascelles, S. Maeda whereas onto untreated silica gel the polymerisation mainly and S. P. Armes, J. Colloid Interface Sci., 1995, 174, 510.occurred into the pores of silica particles. The increase in the 31 B. Saoudi, N. Jammul, M. L. Abel, M. M. Chehimi and G. Dodin, Synth. Met., 1997, 87, 97. polypyrrole content at the surface of the hybrid particles could 32 H. Chriswanto, H. Ge and G. G.Wallace, Chromatographia, 1993, be due to favourable specific interactions of Lewis acid–base 37, 423.type between the basic amino group and the acidic NMH 33 H. Chriswanto and G. G. Wallace, Chromatographia, 1996, 42, bonds and the positively charged polypyrrole backbone. 191. Due to a polypyrrole-rich surface, PPy–APS–silica hybrid 34 J. R. Conder, J. High Resol. Chrom. Chrom. Commun., 1982, 5, 341. materials exhibit a significant increase of electrical conductivity 35 Physicochemical Measurement by Gas Chromatography, ed.by comparison to the APS-free hybrid particles. This was also J. R. Conder and C. L. Young, Wiley, Chichester, 1979. reflected in the XPS analysis by a substantial decrease of the 36 F. Faverolle, O. Le Bars, A. J. Attias and B. Bloch, J. Chim. Phys., surface static charge of the hybrid particles with the APS 1995, 92, 943.pretreatment. 37 F. Faverolle, O. Le Bars, A. J. Attias and B. Bloch, Organic This multitechnique study showed that the APS-treated Coatings, ed. P. C. Lacaze,Woodbury, New York, 1996, p. 267. 38 Silanes, Surfaces and Interfaces, ed. D. E. Leyden, Gordon and silica gel is eVective in processing PPy–APS–silica hybrid Breach, New York, 1985. materials which exhibit interesting surface properties such as 39 Silanes and Other Coupling Agents, ed.K. L. Mittal, VSP, Utrecht, polypyrrole-rich surface, fairly good electrical conductivity The Netherlands, 1992. and high specific surface area. 40 S. Rapi, V. Bocchi and G. P. Gardini, Synth. Met., 1988, 24, 217. 41 S. P. Armes, Synth. Met., 1987, 20, 365. 42 S. Brunauer, P. H. Emmett and E. Teller, J. Am. Chem. Soc., 1938, We would like to thank Prof. F. Fievet and G. Cheguillaume 60, 309. (Laboratoire de Chimie des Mate�riaux Divise�s et Catalyse, 43 E. T. Kang, K. G. Neoh, Y. K. Ong, K. L. Tan and B. T. G. Tan, Universite� Paris VII) for the TGA analysis and their helpful Synth. Met., 1990, 39, 69. discussions. The French Ministry of Education and Research 44 D. Wang and F. R. Jones, J. Mater. Sci., 1993, 28, 2481. is gratefully acknowledged for the financial support provided 45 I. Georges, P. Viel, C. Bureau, J. Suski and G. Leayon, Surf. Interface Anal., 1996, 24, 774. as a PhD studentship for one of us (C.P.). 46 W. J. Van Ooij and A. Sabata in ref. 36, p. 323. 47 Practical Surface Analysis, Auger and X-Ray Photoelectron Spectroscopy, ed. D. Briggs and M. P. Seah, John Wiley, References Chichester, 2nd edn., 1990, volandbook of Conducting Polymers, ed. T. A. Skotheim, Marcel 48 E. T. Kang, K. G. Neoh, Y. K. Ong, K. L. Tan and B. T. G. Tan, Dekker, New York, 1986, vol. 1 and 2. Macromolecules, 1991, 24, 2822. 2 J. Rodriguez, H. J. Grande and T. F. Otero, in Handbook of 49 C. Perruchot, M. M. Chehimi, M. Delamar, S. F. Lascelles and Organic Conducting Molecules and Polymers, H. S. Nalwa, John S. P. Armes, Langmuir, 1996, 12, 3245. 50 M. L. Abel and M. M. Chehimi, Synth. Met., 1994, 66, 225. Wiley & Sons Ltd, 1997, vol. 2, p. 415. 51 D. Kowalczyk, S. Slomkowski, M. M. Chehimi and M. Delamar, 3 M. M. Chehimi, M. L. Abel, E. Pigois-Landureau and Int. J. Adhes. Adhes., 1996, 16, 227. M. Delamar, Synth. Met., 1993, 60, 183. 52 E. T. Kang, K. G. Neoh and K. L. Tan, Adv. Polym. Sci., 1993, 4 M. M. Chehimi and E. Pigois-Landureau, J. Mater. Chem., 1994, 106, 135. 4, 741. 53 J. Luo, R. Seghi and J. Lannutti, Mater. Sci. Eng., 1994, C5, 15. 5 M. M. Chehimi, S. Lascelles and S. P. Armes, Chromatographia, 54 T. H. Chao and J. March, J. Polym. Sci. Polym. Chem., 1988, 1995, 41, 671. 26, 743. 6 H. Ge and G. G. Wallace, React. Polym., 1992, 18, 133. 55 S. Maeda and S. P. Armes, Synth. Met., 1995, 73, 151. 7 A. Talaie, Polymer, 1997, 35, 1145. 8 C. N. Aquino-Binag, N. Kumar, R. N. Lamb and P. J. Pigram, Chem. Mater., 1996, 8, 2579. Paper 8/03019G J. Mater. Chem., 1998, 8(10), 2185–2193 2193

ISSN:0959-9428    

DOI:10.1039/a803019g

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 Highly-conducting neutral copper complexes substituted with two tetrathiafulvalenyldithiolato groups Kazumasa Ueda,a Makoto Goto,b Masaki Iwamatsu,a Toyonari Sugimoto,*a Satoshi Endo,a Naoki Toyota,a Koji Yamamotob and Hideo Fujitac aResearch Institute for Advanced Science and Technology, Osaka Prefecture University, Sakai, Osaka 599-8570, Japan bDepartment of Chemistry, Faculty of Integrated Arts and Sciences, Osaka Prefecture University, Sakai, Osaka 599-8531, Japan cDepartment of Natural Environment, Faculty of Integrated Studies, Kyoto University, Yoshida, Kyoto 606-01, Japan Received 18th May 1998, Accepted 1st July 1998 The neutral bis(dialkylthiotetrathiafulvalenyldithiolato)copper complexes, Cu(R2C6S8)2 (R=Me 1a, Et 1b, R2= MCH2CH2M1c) were synthesized by the reaction of bis(tetramethylammonium or tetraethylammonium) bis(dialkylthiotetrathiafulvalenyldithiolato)zincate complexes with CuCl2 in DMF, followed by oxidation with iodine.The room-temperature electrical conductivities were 7.8×10-2, 5.0×10-4 and 3.7 S cm-1 for compressed pellets of 1a, 1b and 1c, respectively. Magnetic susceptibility measurements showed that the remaining spin on the copper atoms is very small for these complexes as a result of large intramolecular electron transfer from CuII atom to the tetrathiafulvalenyl radical cation ligands.Much eVort continues to be directed toward the synthesis of new organic complexes with localized d electrons on the metal atom and conducting p electrons on the ligand.1 So far, oxidatively doped copper(II)2–8 and cobalt(II)9 metallophthalocyanines are well-known typical p–d cooperative systems.These systems form a one-dimensional array of paramagnetic local moments in the conducting framework of the organic phthalocyanine ligands. Furthermore, the copper salt of 2,5- dimethyl-N,N¾-dicyano-p-benzoquinonediimine10 and the charge-transfer complexes of several tetrathiafulvalene derivatives with associated magnetic metal ions11–14 provide two- or three-dimensional p–d cooperative systems.Our interest concerns the synthesis of a new type of organic complexes with magnetic metal ions incorporated through covalent bonds into a two- or three-dimensional p conducting network, since a strong p–d interaction can be expected in their molecular crystals and unique electrical conducting and magnetic properties might result.Here, we report the synthesis of neutral copper complexes substituted with bis(dimethylthio-, diethylthio- and ethylenedithio-tetrathiafulvalenyldithiolato) ligands (1a, 1b and 1c), and their electrical conducting and magnetic S S S S S Zn S S S S S S S RS RS SR SR S S S S SAr SAr RS RS 2- Ar = p-Acetoxybenzyl 1a, 1b, 1c 2R'4N+ R, R = -CH2CH2-, R' = Et R, R = -CH2CH2- 2a 2b 2c 3a 3b 3c i,ii iii,iv R = Me R = Et R, R' = Me R = Me, R' = Et properties.Scheme 1 Reagents and conditions: i NaOMe, MeOH, room temperature, 30 min; ii R¾4NBr (R¾=Me, Et), ZnCl2, room temperature, 2 h; iii CuCl2, DMF, -40 °C, 12 h; iv I2, DMF, -40 °C, 12 h. was stirred for 30 min, followed by the addition of Me4NBr or Et4NBr (0.60 mmol) and ZnCl2 (0.30 mmol) and further stirring for 2 h.The resultant precipitate, bis(tetramethylammonium) or bis(tetraethylammonium) bis(dialkylthiotetrathiafulvalenyldithiolato) zincate 317 was filtered oV, washed RS RS S S S S S Cu S S S S S S S SR SR 1a 1b 1c R = Me R = Et R,R = -CH2CH2- with MeOH and finally dried in vacuo.To CuCl2 (0.12 mmol) in DMF (10 ml ) was added 3 (0.10 mmol) in DMF (10 ml ) at -40 °C under argon, and stirring was continued overnight Experimental at room temperature. After removing the solvent a green solid was collected, washed with MeOH and dried in vacuo. Reaction Synthesis of neutral of this green solid with an excess of iodine in DMF (10 ml ) bis(dialkylthiotetrathiafulvalenyldithiolato)copper complexes 1 was carried out at -40 °C under argon, followed by solvent removal, washing with MeOH and drying in vacuo to give 1.The synthetic procedure is shown in Scheme 1. Sodium (9.0 mmol) was added to bis(p-acetoxybenzylthio)bis(dialkyl- Bis(dimethylthiotetrathiafulvalenyldithiolato)copper 1a:18 45% yield; green solid (mp>300 °C); Anal. Calc. for thio)tetrathiafulvalene 215,16 (0.30 mmol) in MeOH (15 ml ) at room temperature under argon, and the reaction mixture C16H12S16Cu : C, 24.61; H, 1.55.Found: C, 24.48; H, 1.78%. J. Mater. Chem., 1998, 8(10), 2195–2198 2195Table 1 Electrical conductivities for 1a, 1b and 1c at room temperature Bis(diethylthiotetrathiafulvalenyldithiolato)copper 1b: 96% yield; dark green powder [mp #150 °C (decomp.)]; Anal.Calc. srt/S cm-1 Ea/eV for C20H20S16Cu5C, 28.70; H, 2.41. Found: C, 28.77; H, 2.65%. Bis(ethylenedithiotetrathiafulvalenyldithiolato)copper 1a 7.8×10-2 1c: 96% yield; dark green powder (mp>300 °C); Anal. Calc. 1b 5.0×10-4 1c 3.7 0.05 for C16H8S16Cu5C, 24.74; H, 1.04. Found: C, 24.48; H, 1.28%. Electrical conductivities the CuIII state. From this result it is probable that the geometry Electrical conductivities were measured on compressed pellets around the copper is near planar for all the complexes. by using a four-probe method at various temperatures between 5 and 300 K; electrode contacts were made using gold paste.Electrical conducting properties of 1a, 1b and 1c Table 1 shows the electrical conductivities (s) at room Magnetic susceptibilities temperature (srt) and activation energies (Ea) for compressed pellets of the three copper complexes.The srt values of 1a, 1b Magnetic susceptibilities (xobs) were measured in the temperaand 1c are 7.8×10-2, 5.0×10-4 and 3.7 S cm-1, respectively, ture range 5–300 K at an applied field of 1 kOe using a SQUID which are comparable to those of neutral bis(tetrathiafulvalen- magnetometer (MPMS XL, Quantum Design).The paramagyldithiolato) nickel complexes (4) (10-4–10-1 S cm-1)26 and netic susceptibility was obtained by subtracting the diamagbis( benzene-1,2-dithiolato)aurate (1×10-3 S cm-1).27 It is netic contributions, calculated using Pascal’s constants, from notable that the srt value of 1c is remarkably high. The xobs. temperature dependence of the s value was investigated in the temperature range 5–300 K, and semiconducting behavior with EPR spectra a gradual decrease in s with lowering temperature was The EPR spectra were recorded by using a JEOL 1X observed.The s vs. T plot obeyed the relation s= spectrometer with Mn2+/MgO used as calibrant. A exp(Ea/kT ), with Ea equal to 0.05 eV from the slope of the ln s vs.T-1 plot, as shown in Fig. 1. This value is very low, suggesting the possibility of metal-like behavior in a s vs. T MO calculation method plot for, as yet, unavailable single crystals. The srt values are MO calculations were performed with the ZINDO program,19 strongly dependent on the kinds of substituents on the tetrathiwhich allows for the treatment of transition-metal complexes afulavalenyl rings.Thus, the smaller the bulkiness of the and the inclusion of extensive configuration interaction. alkylthio group, the higher is the srt value. A similar tendency was also observed in 4;26 srt values were 10-1, 10-1 and 10-4 S cm-1 on compressed pellets for the methylthio- (4a), Results and Discussion ethylthio- (4b) and n-butylthio-substituted derivatives (4c), MO calculations The molecular structures of 1a, 1b and 1c are unknown since no single crystals suitable for the X-ray structure analysis could be obtained.Furthermore, there are no reports on the crystal structures of any neutral bis(dithiolato)copper complexes. Nevertheless, a number of mono- and bis-anionic salts of bis(dithiolato)copper complexes have been crystallograph- RS RS S S S S S Ni S S S S S S S SR SR 4a 4b 4c R = Me R = Et R = Bun ically characterized and the geometry around the copper atom shown to be variable.The tetra-n-butylammonium salt of respectively. A much larger srt value might be obtained for the ethylenedithio-substituted derivative, unfortunately not bis(maleonitriledithiolato)cuprate(III)20 and the bis(tetra-nbutylammonium) salt21,22 of the corresponding cuprate(II) synthesized in the study, on the basis of the present results.These results suggest that the alkylthio substituent exerts an complex have almost planar geometry around the copper atom. On the other hand, a relatively large distortion from important influence on the stacking mode between the tetrathiafulvalenyl moieties in 1 and 4, which is responsible for the planarity (41.1–57.3°) was seen for the bis(tetramethylammonium) 23 and bis(Methylene Blue) salts24 of bis(maleonitrile- magnitude of electrical conductivity.Information on the stacking modes can be only obtained from the crystal structure of dithiolato)cuprate(II) and for the bis(N-ethylpyridinium) salt of the bis(4,5-dimercapto-1,3-dithiole-2-thionato)cuprate(II) 4b.The geometry around the nickel atom is square-planar and the eight sulfur atoms of one diethylthio-substituted tetrathia- complex.25 ZINDO MO calculations of a parent system of 1, a neutral bis(tetrathiafulvalenyldithiolato)copper complex (1¾) fulvalenyldithiolato ligand are essentially coplanar. Owing to the ethylthio groups projecting out of the molecular plane, the were performed on two extreme cases in which the geometries around the copper atom are planar and distorted by 57.3° molecules can not completely overlap with each other, but are stacked with a interplanar distance of ca. 3.4 A° in a slightly from planarity.The formal charges on the copper atom obtained by Mulliken population analyses were 1.552 and slipped manner. Accordingly, the intermolecular nickel–nickel distance is large (5.17 A° ).Such a partial stacking mode 1.385 for the planar and 57.3°-distorted geometries, respectively. For comparison, similar calculations were also performed between the tetrathiafulvalenyl moieties is considered as one of the reasons for the only moderate electrical conductivity. on the planar bis(maleonitriledithiolato)-copper(II) and -copper( III) complexes, and the formal charges on the copper atom Considering this fact in combination with the higher electrical conductivity for 1c, it is expected that much more tight were 1.442 and 1.539 respectively. Obviously, the copper atom in 1¾ with planar geometry is close to the CuIII state, and with stacking occurs between the tetrathiafulvalenyl moieties as a result of sterically less hindered ethylenedithio groups.On the increasing distortion from planar geometry the copper atom gradually approaches the CuII state. This implies that distortion other hand, 1a and 1b might adopt a similar stacking to that of 4b. However, this conclusion must await X-ray structure from planar geometry decreases the electron transfer from CuII to the TTF radical cation ligands because of a decreased analysis on single crystals of 1a, 1b and 1c, which are now being attempted to be prepared.In addition, the amount of d–p interaction. As is mentioned later, the spin on the copper atoms is almost zero for 1a–1c, where the copper atom is in positive charge residing on each tetrathiafulvalenyl moiety is 2196 J. Mater.Chem., 1998, 8(10), 2195–2198Fig. 2 Temperature dependence of paramagnetic susceptibilities (xp) obtained for 1a (×), 1b (#) and 1c (6). The solid lines are reproduced using the C, h and xp values listed in Table 2. Fig. 1 Temperature dependence of electrical conductivity (s) in the temperature range 5–300 K: (a) s vs. T and (b) ln s vs. T-1 plots. responsible for the comparatively high electrical conductivities. Thus, if copper is in a divalent oxidation state (CuII ) (Scheme 2), a 1+ charge must be placed on each tetrathiafulvalenyl moiety in order to maintain charge neutrality for the complexes.In this case it is diYcult to produce a high electrical Fig. 3 Temperature dependence of xpT for 1a (×), 1b (#) and 1c conductivity since a mixed-valence state can not be achieved (6).The solid lines are reproduced using the C, h and xp values listed in the tetrathiafulvalenyl stacks. In the other extreme, trivalent in Table 2. copper (CuIII ) induces a 0.5+charge into each of the tetrathiafulvalenyl moieties, thereby leading to a mixed-valence state ducting p electrons and the localized d spins coexist and for the tetrathiafulvalenyl stacks.Of course, intermediate interact with each other. By fitting C, h and xp values in the oxidation states of copper also enable production of a high above equation, the observed xp vs. T behavior was well electrical conductivity. As discussed below, the copper atom reproduced for each complex and values are summarized in is virtually in the trivalent oxidation state leading to the Table 2.The C values are 1.0×10-3, 1.0×10-2 and 1.2×10-2 production of a high electrical conductivity for all the emu K mol-1 for 1a, 1b and 1c, respectively, which correspond complexes. to 0.24, 2.7 and 3.2% of that calculated by the equation of C=NmB2g2S(S+1)/3kB, where N is Avogadro’s number, mB the Bohr magneton, kB is the Boltzmann constant, S the spin (TTF0)Cu(III)–(TTF•+) (TTF•+)Cu(III)–(TTF0) (TTF+•)Cu(II)2–(TTF•+) magnetic quantum number, and g an average g factor of the Scheme 2.CuII spin, respectively. As can be readily seen from the very small C values, spin on copper is scarcely present, and the copper atom is essentially in the CuIII state. However, a very Magnetic properties of 1a, 1b and 1c small amount of spin still remains on the copper atom, and The temperature dependence of the paramagnetic susceptibility the interaction between the neighboring spins is weakly-antifer- (xp) was investigated in the temperature range 5–300 K for romagnetic, as shown by the small value and negative sign in 1a, 1b and 1c and results are shown in Fig. 2. Fig. 3 depicts h. The xp values are 9.1×10-4, 1.9×10-4 and 3.6×10-4 the temperature dependence of xpT where a linear relationship emu mol-1 for 1a, 1b and 1c, respectively, and comparable to is observed in the higher temperature region, suggesting the those (1.9×10-4 and 4.3×10-4 emu mol-1) for Cu(pc)I8 and presence of a temperature-independent susceptibility (denoted Co(pc)AsF6,9 respectively.xp). An almost temperature-independent susceptibility in the low temperature region of the Bonner–Fisher curve is obtained EPR spectra as a result of strong antiferromagnetically interacting spins of tetrathiafulvalenyl radical cations in the present complexes.All of the three copper compounds showed only one broad xCu(T ) (=xp-xp) is the contribution from the localized spin signal in their solid EPR spectra. The g values were 2.0075 on the copper atom, and obeys the Curie–Weiss law.for 1a, 2.0089 for 1b and 2.0110 for 1c. By contrast, for both Accordingly, the observed temperature dependence of xp can be expressed by eqn. (1) Table 2 The C, h and xp values for 1a, 1b and 1c xp(T )=xCu(T )+xp=C/(T-h)+xp (1) C/emu K mol-1 h/K xp/emu mol-1 where C is the Curie constant and h the Weiss temperature. 1a 1.0×10-3 ~0 9.1×10-4 Similar xp vs.T behavior has previously been seen in (phthalo- 1b 1.0×10-2 -1.2 1.9×10-4 cyaninato)copper iodide [Cu(pc)I ]8 and (phthalocyaninato)- 1c 1.2×10-2 -4.2 3.6×10-4 cobalt hexafluoroarsenate [Co(pc)AsF6],9 where the con- J. Mater. Chem., 1998, 8(10), 2195–2198 2197planar and non-planar bis(dithiolato)copper(II) complexes Prefecture University) for discussion on the MO calculations. Financial support for this work was provided by the Nagase already characterized, two diVerent signals were observed at gd#2.09 and g)#2.02 in the solid state.A most probable Foundation for Science and Technology and by a Grant-in- Aid for Scientific Research from the Ministry of Education, cause for the present behaviour is the exchange of the CuII spin with other spins, the TTF radical cation in this case, Science and Culture, Japan.giving rise to coalescence of two diVerent signals due to the CuII spin state. Nevertheless, the magnetic measurement results above showed that the spin on the copper atoms is very small (0.24 for 1a, 2.7 for 1b and 3.2% for 1c) because of almost References complete electron transfer from CuII to the TTF moieties.When strong exchange occurs between the CuII spin and the 1 P. Day, Philos. Trans. R. Soc. London, Sect. A, 1985, 314, 145. 2 C. J. Schramm, R. P. Scaringe, D. R. Stojacobic, B. M. HoVman, TTF radical cation, the observed g value can be determined J. A. Ibers and T. J. Marks, J. Am. Chem. Soc., 1980, 102, 6702. by eqn. (2) 3 J. Martinsen, R. L. Greene, S. M. Palmer and B.M. HoVman, g=xCugCu/xtotal+xpgp/xtotal (2) J. Am. Chem. Soc., 1983, 105, 677. 4 J. Martinsen, S. M. Palmer, J. Tanaka, R. L. Greene and xtotal=xCu+xp B. M. HoVman, Phys. Rev. B, 1984, 30, 6269. where gCu and gp, and xCu and xp are the g values and local 5 P. J. Toscano and T. J. Marks, J. Am. Chem. Soc., 1986, 108, 437. 6 J. Martinsen, J.-L. Stanton, R. L. Greene, J.Tanaka, magnetic susceptibilities contributed from the CuII spin and B. M. HoVman and J. A. Ibers, J. Am. Chem. Soc., 1985, 107, the TTF radical cation, respectively, and xtotal the total mag- 6951. netic susceptibility. However, since in the present case the spin 7 M. Y. Ogawa, J. Martinsen, S. M. Palmer, J. L. Santon, on the copper atom is very small, only a very small increase R.L. Greene, B. M. HoVman and J. A. Ibers, J. Am. Chem. Soc., from the g value (2.0072) of the radical cation salt of tetram- 1987, 109, 1115. ethylthio-substituted TTF is expected. Of course, the degree 8 M. Y. Ogawa, B. M. HoVman, S. Lee, M. Yudkowsky and W. P. Halperin, Phys. Rev. Lett., 1986, 57, 1177. of shifting (Dg) is strongly dependent on the amount of spin 9 K. Yakushi, H.Yamakado, T. Ida and A. Ugawa, Solid State on the copper atom. The Dg values are 0.0003, 0.0017 and Commun., 1991, 78, 919. 0.0038 for 1a, 1b and 1c, respectively, which are consistent 10 A. Aumuller, P. Erk, G. Klebe, S. Hunig, J. U. von Schutz and with an increase in the spin on the copper atom along H.-P. Werner, Angew. Chem., 1986, 98, 759; Angew. Chem. Int. this series. Ed.Engl., 1986, 25, 740. 11 P. Day, M. Kurmoo, T. Mallah, I. R. Marsden, M. L. Allan, R. H. Friend, F. L. Pratt, W. Hayes, D. Chasseau, G. Bravic and Conclusions L. Ducasse, J. Am. Chem. Soc., 1992, 114, 10722. 12 T. Enoki, J. Yamaura, N. Sugiyasu, K. Suzuki and G. Saito, Mol. Our initial interest in 1 lies in it containing both p radical Cryst. Liq. Cryst., 1993, 233, 325. cations (TTF ligands) and a d spin on the CuII moieties, which 13 M.Kurmoo, A. W. Graham, P. Day, S. J. Coles, are located separately from each other. A two- or three- M. B. Hursthouse, J. M. Caulfield, J. Singleton, L. Ducasse and dimensional array of 1 could provide a unique organic solid P. Guionneau, J. Am. Chem. Soc., 1995, 117, 12209. 14 A. Kobayashi, T. Udagawa, H. Tomita, T. Naito and system involving both conducting p electrons and localized d H.Kobayashi, Chem. Lett., 1993, 2179. spins in interaction with each other, and novel electrical 15 C. Gemmell, J. D. Kilburn, H. Ueck and A. E. Underhill, conducting and magnetic properties might result. However, in Tetrahedron Lett., 1992, 33, 3923. contrast to this expectation a large degree of intramolecular 16 Y.Misaki, H. Nishikawa, K. Kawakami, S. Koyanagi, T. Yamabe electron transfer from CuII to the TTF radical cation ligands and M. Shiro, Chem. Lett., 1992, 2321. occurs, giving rise to almost no spin on the resultant CuIII 17 K. Ueda, M. Yamanoha, T. Sugimoto, H. Fujita, A. Ugawa, K. Yakushi and K. Kano, Chem. Lett., 1997, 461. atom. On the other hand, the one electron transferred from 18 K.Ueda, M. Goto, T. Sugimoto, S. Endo, N. Toyota, the CuII atom can be equally accepted by the two TTF radical M. Yamamoto and H. Fujita, Synth. Met., 1997, 85, 1679. cation ligands. Consequently, a +0.5 radical cation resides on 19 J. Ridley and M. Zerner, Theor. Chim. Acta, 1973, 32, 111. each TTF ligand, so that a mixed valence state can be achieved, 20 J. D. Forrester, A.Zalkin and D. H. Templeton, Inorg. Chem., if the TTF ligands are eVectively stacked. In fact, the present 1964, 3, 1507. neutral copper complexes exhibited comparatively high electri- 21 A. M. Maki, N. Edelstein, A. Davison and R. H. Holm, J. Am. Chem. Soc., 1964, 86, 4580. cal conductivities without any doping. Several organic metal 22 K. W. Plumlee, B. M. HoVman and J. A. Ibers, J. Chem. Phys., complex-based intrinsic conductors are known so far. Their 1975, 63, 1975. single crystal srt values lie in the range 10-4–10-1 S cm-1. In 23 C. Mahadevan and M. Seshasayee, J. Crystallogr. Spectrosc. Res., comparison the higher srt value of 1c measured on a com- 1984, 14, 215. pressed pellet is of significance. 24 D. Snaathorst, H. M. Doesburg, J. A. A. J. Perenboom and Other magnetic metal ions than CuII, e.g. CrIII (S=3/2), C. P. Keijzers, Inorg. Chem., 1981, 20, 2526. 25 G. Matsubayashi, K. Takahashi and T. Tanaka, J. Chem. Soc., MnII (S=5/2 or 1/2), CoII (S=1/2) and FeIII (S=5/2 or 1/2), Dalton Trans., 1988, 967. can also be used in the preparation of the similar neutral 26 N. L. Narvor, N. Robertson, T. Weyland, J. D. Kilburn, complexes to 1. From these new complexes diVerent electrical A. E. Underhill, M. Webster, N. Seventrup and J. Becher, Chem. conducting and magnetic properties from those of 1 might be Commun., 1996, 1364. observed. We are presently preparating the neutral cobalt and 27 N. C. Sciødt, T. Bjørnholm, K. Bechgaard, J. J. Neumeir, iron complexes. C. Allgeier, C. S. Jacobsen and N. Torup, Phys. Rev. B, 1996, 53, 1773. The authors are grateful to Prof. Kazuo Kitaura (Department Paper 8/03718C of Chemistry, Faculty of Integrated Arts and Sciences, Osaka 2198 J. Mater. Chem., 1998, 8(10), 2195–2198

ISSN:0959-9428    

DOI:10.1039/a803718c

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 Towards elucidating microscopic structural changes in Li-ion conductors Li1+yTi2-yAly[PO4]3 and Li1+yTi2-yAly[PO4]3-x[MO4]x (M=V and Nb): X-ray and 27Al and 31P NMR studies Shan Wong,a Peter J. Newman,b A. S. Best,b K. M. Nairn,a D. R. MacFarlaneb and Maria Forsyth*a aDepartment of Materials Engineering and bDepartment of Chemistry, Monash University, Wellington Road, Clayton, Victoria 3168, Australia.E-mail: maria.forsyth@eng.monash.edu.au Received 14th April 1998, Accepted 17th July 1998 A combination of X-ray powder diVraction (XRD) and nuclear magnetic resonance (NMR) studies has demonstrated that attempted substitutions of Al, V and Nb into the framework of LiTi2(PO4)3 yield several impurity phases in addition to direct substitutions of Al into Ti and V, Nb into P sites.Direct substitutions were confirmed by changes in the unit cell dimensions as indicated by the peak shifts observed in the X-ray diVractographs and by analyses of the 27Al and 31P magic angle spinning (MAS) spectra. A major impurity phase was identified as AlPO4 (found in at least two polymorphs) and the amount present increases with increasing Al additions.The formation of AlPO4 appeared to be enhanced by further V but suppressed by Nb substitution. These results suggest that the presence of AlPO4, together with the non-stoichiometric modified LTP, may be the cause for the observed densification of this material upon sintering and the increased ionic conductivity. S cm-1), although the grain boundary conductivities were Introduction found to be highly dependent on processing conditions.5 The Lithium ion conducting LiTi2[PO4]3 (LTP) is a microporous improved conductivity was attributed by Aono et al.to the ceramic based on a NASICON framework structure of [PO4]3- densification of the sintered pellet and to a decrease in actitetrahedra and [TiO6]8- octahedra in corner sharing arrange- vation energy for the grain boundary ion transport.The ments.1,2 A representation of the NASICON structure is activation energy for the bulk conductivity remained unaVecshown in Fig. 1. Mobile Li+ cations occupy interstitial pos- ted by the Al3+ substitution. The question remains as to the itions to achieve electroneutrality. Hence, these materials are origin for the higher degree of densification in the modified potentially useful in lithium battery applications.Li ion material. The exact nature of site substitution, and its eVect mobility and thereby the Li ionic conductivity are influenced on the framework structure, has also yet to be elucidated. by sizes of the polyhedral cations and ionicity of the Li,O In an earlier solid state nuclear magnetic resonance (NMR) bond.study, tetrahedral and octahedral Al signals were observed in By substitution of trivalent cations for Ti4+ cations in the LATP, when only an octahedral signal for the Al in the octahedral sites of LiTi2[PO4]3 (LTP), Aono et al. reported octahedral (Ti) sites was expected.6 Here, a systematic concenmuch improved Li ionic conductivity in the modified cer- tration-dependent study on the modified ceramics, namely amics.3,4 Further pentavalent cation substitution into the Li1+yAlyTi2-y(PO4)3 (LATP) and Li1+yAlyTi2-y(PO4)3-xtetrahedral phosphorus sites (e.g.using V, Nb, Ta) can also (MO4)x (M=V5+, Nb5 +) (V-LATP and Nb-LATP), is prebe carried out. Among the modified LTP ceramics, Al substi- sented. The aim is to investigate the origin of the two alututed LTP at the nominal composition Li1.3Al0.3Ti1.7[PO4]3 minium signals, to clarify the solid state reactions, and to (LATP) was reported by Aono et al.to have the optimum Li understand the exact nature of site substitution, as part of a ionic conductivity (bulk 3×10-3, grain boundary 9×10-4, continuing eVort in determining the eVect of framework modiand total 7×10-4 S cm-1).Similar bulk conductivities for fication on Li ion mobility. X-Ray powder diVraction and LATP have been reproduced in our laboratory (#2×10-3 high resolution solid state NMR spectroscopy are used. Experimental Appropriate ratios of (NH4)2HPO4, TiO2, Li2CO3, Al2O3 for LATP (and V2O5 or Nb2O5 for V/Nb-LATP) were mixed and the mixture was placed in a graphite crucible and heated to and held at 300 °C until ammonia, carbon dioxide, and water evolution ceased.The coarse solid was ground and passed through a 250 micrometer sieve before being heated at 400 °C in a platinum crucible for 1 h, followed by 950 °C for 2 h. The reaction was carried out under a continuous flow of Ar(25%)–O2 to prevent reduction of the transition metal cations. The reaction product was ball milled in ethanol, and the resultant particle size was estimated to be <2 mm by SEM Fig. 1 A representation of the NASICON structure (adopted from (JEOL JSM 840-A). X-Ray powder diVractographs (Rigaku ref. 3), showing the unit cell stacking along the c-axis ( left) and its Geiger-flex) were acquired to confirm reaction products. In projection onto the basal plane (right).Lithium cations occupy addition to peaks corresponding to LTP, low intensity peaks interstitial positions surrounded by six phosphate tetrahedra and sandwiched by two titanium octahedra. were observed which suggest the presence of residual amounts J. Mater. Chem., 1998, 8(10), 2199–2203 2199of impurities (estimated to be <5%). The significance of these impurity phases will be discussed.NMR experiments were carried out in a 7.05 T superconducting magnet (Oxford Instruments) equipped with a Bruker CXP 300 spectrometer operated by a Tecmag interface (Minimacspect). Room temperature magic angle spinning (MAS) spectra were acquired using a MAS probe (Bruker HP-WB, 4 mm zirconia rotor and Kel-F endcap) at the maximum spinning speed of 14 kHz.Typical radio frequency pulses of 2 ms for 31P and 0.6 ms for 27Al (1/3 of ‘p/2’ pulse) were used. All 27Al spectra were accumulated over 64 transients with a 40 s recycle delay. 31P spectra were averaged over four scans with a 40 s delay between scans. 27Al and 31P chemical shifts were externally referenced to [Al(H2O)6]3+ and 85% H3PO4, respectively. Results and discussion Fig. 3 X-Ray powder diVraction patterns for (a) LTP, (b) LATP ( yAl=0.3), (c) V-LATP ( yAl=0.3 and xV=0.1) and (d) Nb-LATP Results ( yAl=0.3 and xNb=0.1) plotted for comparison. X-Ray powder diVraction. Fig. 2 shows X-ray powder diVraction patterns taken for LATP samples of varying Al Table 1 Unit cell parameters for LTP and LTP-based ceramics concentration. Major reflections were assigned according to the previously reported powder diVraction for LTP.7 The main Nominal composition a/A° c/A° b/° V /A° 3 reaction product was confirmed to be rhombohedral of space LiTi2(PO4)3 a 8.5129(8) 20.878(4) 120b 1310.3 group R39c (analogous to NASICON). The diVraction patterns Li1.3Al0.3Ti41.7(PO4)3 b 8.50 20.82 120b 1303 for LTP, LATP, V-LATP, and Nb-LATP ( yAl=0.3, xV,Nb= Li1.3Al0.3Ti1.7(PO4)2.9(VO4)0.1 8.5044(6) 20.827(5) 120b 1304.5 0.1) are shown in Fig. 3 for comparison. Unit cell dimensions Li1.3Al0.3Ti1.7(PO4)2.9(NbO4)0.1 8.5061(6) 20.836(4) 120b 1305.6 derived from structural analyses of the diVraction patterns for aRef. 7. bRef. 3. V0.1-LATP and Nb0.1-LATP are given in Table 1, together with results reported for LTP7 and for LATP.3 Cell dimensions for all the modified ceramics are noticeably the intensities of the TiP2O7 peaks were much reduced.AlPO4 smaller than LTP, with LATP ( yAl=0.3) being the smallest. (tridymite) and rutile, but not TiP2O7, were present in the The contraction of the LATP unit cell can be explained to a V0.1-LATP systems. Intensities of the tridymite peaks increased first order approximation by partial substitution of smaller with increasing Al added to both LATP and V-LATP ceramics Al3+ for Ti4+ ions (0.535 vs. 0.605 A° ionic radius). The slight (Fig. 2). The sharp reflections of the impurity peaks suggest expansion of the unit cell upon V and Nb substitution results crystalline phases are present, and the flat baselines indicate from much larger V5+ and Nb5+ ions occupying the P5+ sites the absence of substantial amorphous materials.The diVrac- (0.355 and 0.48 A° vs. 0.17 A° , respectively). An attempt to tion pattern for Nb-LATP [Fig. 3(d)] does not show the synthesize single-phased Ta-LATP failed, possibly due to the prominent presence of AlPO4. The significance of AlPO4 will Ta5+ cation being much larger. be discussed in the context of 27AlNMRspectroscopy. Authors Low intensity peaks in the regions 2h 17–19, 21–24 and of previous X-ray studies of LTP and modified LTP did 26–29° do not belong to the NASICON structure.In LATP remark on impurity phases present, but the extra reflections with yAl<0.3, two minor phases were identified to be TiO2 were not documented.7,9–11 (rutile) and TiP2O7 (peaks marked by R and T, respectively). At yAl=0.3, new peaks at 2h=21.7 and 35.8° were assigned 27Al MAS spectroscopy.In a previous NMR study of LATP to AlPO4 of the tridymite phase (peaks marked by A)8 and ( yAl=0.3), we observed 27Al resonances at d -15 and 40 which were assigned to aluminium in octahedral (AlO) and tetrahedral (AlT) sites, respectively.6 Here, 27Al MAS NMR studies were carried out to determine the onset of the AlT signal.Fig. 4 shows the 27Al spectra as a function of Al concentration in LATP, with a plot of the AlT/AlO ratios in the inset. At yAl<0.2, only AlO signals at d-15 were observed. Since no Al-containing impurities were found in the X-ray diVractographs, we assigned this signal to Al occupying the octahedral Ti sites in LTP, and thus confirm direct Al substitution into the LTP framework.At yAl=0.2, the d 40 AlT signal appeared, and the AlT onset is therefore estimated to be between y=0.1 and 0.2. The same concentration-dependent study was carried out for V-LATP and results are shown in Fig. 5. Tetrahedral Al signal was detected at yAl=0.1, and the AlT/AlO ratios were comparable to those found for LATP. An additional broad resonance around d 10 was observed in V-LATP ceramics.Fig. 2 X-Ray powder diVraction patterns for LATP as a function of In contrast to LATP and V-LATP, Nb-substituted LATP (Nbincreasing Al addition: (a) 0.0, (b) 0.05, (c) 0.1, (d) 0.2 and (e) 0.3. LATP) showed a much reduced AlT/AlO ratio (1520 at yAl= All samples show major reflections corresponding to those observed 0.3) (Fig. 6). for LTP (identified by sticks at the base).Small intensities in the regions of 2h=17–19, 21–24 and 26–29° cannot be assigned to LTP 31P MAS spectroscopy. The eVect of Al substitution on the and belong to residual phases present (R=TiO2-rutile, T=TiP2O7, and A=AlPO4-tridymite). local environment of phosphorus was monitored by 31P NMR 2200 J. Mater. Chem., 1998, 8(10), 2199–2203Fig. 4 27Al MAS spectra of LATP show the concentration dependence Fig. 6 27Al MAS spectra for V-LATP, LATP at yAl=0.3 show nearly of the octahedral (d -15) and tetrahedral (d 40) Al signals. The identical AlT/AlO ratios of 1/4, while the Nb-LATP spectrum shows resonance at d 40 appears at yAl=0.1. The ratio of AlT/AlO as a that the tetrahedral signal is reduced by Nb substitution. function of Al concentration is plotted in the inset, showing a monotonic increase.Fig. 7 31P MAS spectra for LATP show gradual and asymmetric broadening of the 31P resonance with Al concentration. The asymmetric lineshapes are characteristic of a distribution of sites arising from a distribution of MMOMP bond angles. Fig. 5 27Al MAS spectra for V-LATP show similar concentration dependence observed for LATP.In the presence of vanadium, the onset concentration for the tetrahedral aluminium signal is lowered Discussion to 0.1. Similar AlT/AlO ratios are found for the concentration range studied. X-Ray powder diVraction. To assign residual reflections in the X-ray diVraction patterns of the Al-modified LTP ceramics requires consideration of the eVect of Al substitution. Two hypotheses are considered.First, if Al substitution (at low spectroscopy. Fig. 7 shows the 31P MAS spectra for LTP, and LATP at various Al concentrations. A shift of d -28 is concentration level ) is distributed randomly throughout a crystal (a solid solution), the unit cell dimensions would be characteristic of orthophosphates. The LTP spectrum was narrow, while resonances for LATP showed gradual and modified by the substitution and shifts in the LTP reflections would be observed.Second, if Al incorporation occurs prefe- asymmetric broadening with increasing Al concentration. (Fine spectral shape was observed for the 31P resonance of rentially on the surface of a crystallite (e.g. as a grain boundary phase) or in a crystal domain (e.g. as trapped impurity), then LTP, the origin of which is not entirely understood).Changes in the phosphorus environments in V-LATP and the structure and symmetry of these Al-containing phases would diVer from LTP. In this case, additional reflections Nb-LATP systems were similarly monitored. 31P spectra for V-LATP at a fixed xV=0.1 and varying Al concentration would be detected. For the LATP, V-LATP and Nb-LATP ceramics, peak shifts were measured which confirmed changes (Fig. 8) showed similar asymmetric broadening observed in the LATP spectra. At yAl0.1, however, an additional well in the unit cell dimensions as a result of direct polyhedral cation substitutions. The detection of additional peaks indi- resolved 31P resonance at d -25 was observed, the intensity of which increased with vanadium concentration (Fig. 9). We cated that secondary phases are present, some of which have been identified. assign this minor resonance to phosphorus in the vicinity of V-substituted phosphorus sites.12 A minor 31P resonance was The presence of secondary phases requires that the main reaction product (LATP, V-LATP and Nb-LATP) be non- also observed in Nb-LATP at d -10 (Fig. 10). J. Mater. Chem., 1998, 8(10), 2199–2203 2201Fig. 10 The 31P spectra for Nb-LATP, V-LATP, LATP (at yAl=0.3) Fig. 8 31P MAS spectra of V0.1-LATP showing similar asymmetric and LTP are plotted for comparison. A weak resonance at d -10 is broadening of the phosphorus resonance with increasing substitution. detected and assigned to those 31P nearing a Nb atom (see the Note that at low Al concentration, a minor resonance at d -25 is expanded insert).observed and is assigned to phosphorus atoms nearest to V atoms. At high Al concentration, this resonance is obscured by the overlap between the major and minor peaks. SiO2). Transitions between polymorphs occur at, berlinite CA 700 °C tridymite CA 1100 °C cristobalite The three polymorphs can be distinguished by X-ray powder diVraction14 and by 27AlNMR spectroscopy.15 27Al resonances for tridymite and cristobalite were found at d 39 and, owing to much reduced quadrupolar coupling constants, their central transitions were relatively unaVected by second order quadrupolar interactions.15 On this basis, the d 40 tetrahedral Al signals observed in LATP and V-LATP at yAl0.1 (Fig. 4 and 5) are assigned to the cristobalite and/or tridymite phases of AlPO4.As the tridymite to cristobalite transformation occurs above the reaction temperature of 950 °C, we therefore assign the AlT signal to tridymite phase of AlPO4. This is in agreement with the assignment made based on the positions of the residual X-ray peaks (2h at 21.7 and 35.8°), although 27Al NMR spectroscopy proves to be more sensitive than X-ray powder diVraction. In contrast to tridymite and cristobalite, the berlinite 27Al MAS resonance should be centered at d 10 and show a stronger second order quadrupolar broadening of its central transition.15 The broad and low intensity resonance at d 10 Fig. 9 31P MAS spectra for V-LATP ( yAl=0.05) demonstrate that observed in LATP and V-LATP at high Al concentration the intensity of the resonance at d -25 increases with increasing V (Fig. 4 and 5) may possibly be attributable to the Al in concentration. This trend is also observed for V-LATP at yAl=0.1. berlinite. The 31P chemical shifts found for the three polymorphs of AlPO4 fall into the shift range of interest (d ca. -28).15 Hence, stoichiometric. Both the non-stoichiometry of the modified we are unable to distinguish the Al31PO4 signal from that of LTP and the presence of impurity phases may aVect pellet the modified LTP.Owing to low concentrations of these densification and the ionic conductivities of these materials. phases, however, their contributions to the total phosphorus Among the impurity phases found (rutile, TiP2O7 and AlPO4), signal are considered small. 31P NMR spectroscopy is sensitive AlPO4 is most interesting as it contains the Al3+ cations to the changes in the PMOMM bond angle, which may aVect intended for the Ti sites on the LTP framework (whereas both either or both the magnitude and the orientation of the 31P TiO2 and TiP2O7 are found in the reaction for LTP). Although magnetic shielding (i.e. the chemical shift anisotropy tensor).we are currently unable to pinpoint the location of AlPO4, Under magic angle spinning, these changes are reduced to previous reports have reported its presence in the intergranular diVerent isotropic chemical shifts. The asymmetric broadening regions of titanium phosphate ceramics.13 of the 31P resonance in LATP in Fig. 7, therefore, reflects slight changes in the LTP framework upon substitution.This 27Al and 31P NMR. The presence of AlPO4 at higher Al structural change is illustrated by considering a ‘chain’ of concentration (and possibly other yet to be identified Alalternating oxide bonds, containing impurities phases) undoubtedly complicates the interpretation of the tetrahedral 27Al NMR signal at d 40. The M(TiMOMPMO,Li,OMPMOMTiMOMPMOM)n structure of AlPO4 is formed of corner sharing [PO4]3- and in which an Al3+ is substituted for the Ti4+ as in, [AlO4]5- tetrahedra.Three polymorphs exist: berlinite, tridymite and cristobalite (the latter two are isomorphous with M(TiMOMPMO,Li,OMPMOMAlMOMPMOM)n 2202 J. Mater. Chem., 1998, 8(10), 2199–2203Table 2 Electronegativities (EN) and ion radii (A° ) fillers to reduce porosity in a pellet.More importantly, these grain boundary phases (in this work postulated to be tridymite) Li+ Ti4+ Al3+ P5+ V5+ Nb5+ O2- may provide low energetic pathways for charge transport through the intergranular regions. This interpretation is con- EN 0.98 1.54 1.61 2.19 1.63 1.6 3.44 sistent with the observed reduction in the activation energy IV: CRa 0.73 0.56 0.53 0.31 0.495 0.62 1.24 IRb 0.59 0.42 0.39 0.17 0.355 0.48 1.38 for grain boundary conductivity observed by Aono et al.VI: CR 0.90 0.745 0.675 0.52 0.68 0.78 1.26 in LATP.3,4 IR 0.861 0.605 0.535 0.38 0.54 0.64 1.40 %c 78 59 57 32 56 57 — Conclusion aTetrahedral (IV) and octahedral (VI ) coordination. CR=covalent radius. bIR=ionic radius. c% ionic character of oxide bond= Two 27Al MAS signals have been observed in the families of 1-exp(-0.25(xA-xB)2), where (xA-xB) is the diVerence in modified LTP ceramics, and they have been assigned to Al in electronegativity. octahedral (d-15) and tetrahedral (d 40) sites.The octahedral signal was assigned to aluminium in the octahedral sites of the LTP framework, and thus confirmed the incorporation of Owing to the size diVerence between the two cations, the Al into LTP.The signal at d 40 was assigned to Al in the AlMOMP bond angle deviates from that of TiMOMP, leading tridymite phase of AlPO4, the presence of which was also to a change in the 31P isotropic shift under the condition of detected by X-ray powder diVraction. 27Al NMR showed magic angle spinning. Increasing Al incorporation gradually increasing aluminium incorporation into LTP as well as changes the distribution of MMOMP angles to result in an increasing presence of AlPO4 (tridymite) with increasing Al increasingly asymmetric lineshape for phosphorus.In contrast, added. Tridymite was much more prominently present in when a V or Nb atom is placed in a P site (M = V or Nb), LATP and V-LATP than in Nb-LATP. The 31PMAS lineshape changes were interpreted as increasing LTP structural distor- M(TiMOMPMO,Li,OMMMOMTiMOMPMOM)n tion resulting from Al occupying the Ti sites.The eVects of V the nearest P is in the second coordination sphere and the or Nb substitutions were confirmed by the observation of PMOMM bond angle change is therefore reduced. Based on minor resonances separated from the main resonance.The this analysis, the asymmetric broadening in the phosphorus presence of AlPO4, and other impurity phases, suggests that spectra reflects mostly structural changes due to Al the modified LTP ceramics are non-stoichiometric and contain substitution. vacancies. In combination with changes in the unit cell param- The eVects of adding V and Nb to LTP are observed as eters derived from analyses of X-ray powder diVraction patseparate lower intensity 31P resonances in Fig. 9 and 10. The terns, we propose that the enhanced bulk lithium ion shifts of the minor 31P peaks may reflect changes in the conductivity (observed by Aono et al. and repeated in our electron density near the substituted site. Table 2 shows that laboratory) can be correlated to both the incorporation of Ti, Al, V, and Nb have similar electronegativity (EN) aluminium and the presence of vacancies in the LTP framevalues16,17 which are lower than that of P so that substitution work, while the improved grain boundary conductivity can be of any of these atoms into a P site can alter the local electron correlated to the presence of grain boundary phase(s) possibly density for the nearby phosphorus.the tridymite phase of AlPO4. A discussion on structure and mobility. As the modified LTP References ceramics are prepared from stoichiometric mixtures of oxides, the presence of impurity phases strongly suggests that the 1 J. B. Goodenough, H. Y.-P. Hong and J. A. Kafalas, Mater. Res. Bull., 1976, 11, 203. nominal compositions do not represent the actual ceramic 2 M.A. Subramanian, R. Subramanian and A. Clearfield, Solid compositions. Assuming Li+ ions occupy interstitial positions, State Ionics, 1986, 18/19, 562. the LTP structure can be represented by,12 3 H. Aono, E. Sugimoto, Y. Sadaoka, N. Imanaka and G.- y. Adachi, J. Electrochem. Soc., 1989, 136, 590. 1 Li+·2 [TiO6/2]2-·3.0 [PO4/2]+ 4 H. Aono, E. Sugimoto, Y. Sadaoka, N.Imanaka and G.- For LATP of nominal composition Li1.3Al0.3Ti1.7[PO4]3, an y. Adachi, Chem. Lett., 1990, 1990, 1825. 5 A. B. Best, K. M. Nairn, S. Wong, P. J. Newman, NMR-derived AlT/AlO ratio of ca. 0.25 implies that at most D. R. MacFarlane and M. Forsyth, Proceedings of the 18th 80% of the total Al added is incorporated into the LTP Australasia Ceramics Conference, Sept., 1998, Melbourne, structure.If AlPO4 were the only impurity phase present, then Australia. the modified ceramic composition could be approximated by 6 K. M. Nairn, M. Forsyth, M. Greville, D. R. MacFarlane and M. E. Smith, Solid State Ionics, 1996, 86–88, 1397. 1.3 Li+·0.24 [AlO3]3-·1.7 [TiO3]2-·2.94 [PO2]+ 7 Nat. Bur. Stand. U.S. Monogr., 1984, 25 (21), 79. 8 Nat. Bur. Stand. U.S. Circ., 1960, 539, 10, 4.where vacancies are present and the material can be considered 9 D. Tran Qui and E. Prince, Z. Kristallogr., 1988, 183, 293. as Li rich. In reality, small quantities of rutile and TiP2O7 are 10 D. Tran Qui, S. Hamdoune, J. L. Soubeyroux and E. Prince, also present, rendering the true stoichiometry of the modified J. Solid State Chem., 1988, 72, 309. ceramic diYcult to establish. 11 D. Tran Qui and S. Hamdoune, Acta Crystallogr., Sect. C, 1988, 44, 1360. A hypothesis. Based on the 27Al MAS and XRD results 12 K. C. Sobha and K. J. Rao, J. Solid State Chem., 1996, 121, 197. 13 Y. Nan, W. E. Lee and P. F. James, J. Am. Ceram. Soc., 1992, presented in this study, the enhanced ionic conductivity found 75, 1641. for the modified LTP ceramics can be correlated to increasing 14 V. O. W. Flo� rke, Z. Kristallogr., 1967, 125, 134. Al added, both in terms of direct substitution and the presence 15 D. Mu� ller, E. Jahn and G. Ladwig, Chem. Phys. Lett., 1984, of tridymite. Structural and compositional changes resulting 109, 332. from framework modification lead to two implications, vacanc- 16 R. D. Shannon, Acta Crystallogr., Sect. A, 1976, 32, 751. ies and grain boundary phases, both of which may influence 17 L. Pauling The Nature of the Chemical Bond, Cornell University Press, IthaNY, 3rd edn., 1960. lithium ion dynamics in the modified LTP framework. Vacancies are well known to promote diVusion in a crystal and thereby aVect bulk ionic conductivity. Impurity phases Paper 8/02752H present at the inter-crystalline regions are important space J. Mater. Chem., 1998, 8(10), 2199–2203 2203

ISSN:0959-9428    

DOI:10.1039/a802752h

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 LaB6 crystals from fused salt electrolysis M. Kamaludeen,*a I. Selvaraj,a A. Visuvasama and R. Jayavelb aCentral Electrochemical Research Institute, Karaikudi - 630 006, Tamil Nadu, India bCrystal Growth Centre, Anna University, Chennai 600 025, India Received 17th April 1998, Accepted 2nd July 1998 Purple crystals of LaB6 with high melting point (2500 °C) have been elecrotrodeposited from an oxyfluoride melt consisting of La2O3–B2O3–LiF–Li2O under an N2 atmosphere.The growth of LaB6 crystals has been observed under a controlled electrodeposition process. The cell set-up employed for the electrodeposition consists of a graphite crucible acting both as the cell container and anode and a centrally placed Mo rod acting as cathode. A stainless steel retort was employed to hold the graphite crucible and fix the electrodes under an N2 atmosphere.The pre-treatment applied to the electrolyte composition before electrolysis has been described. Characterization of the crystalline product by chemical analysis, XRD studies and other physical measurements is reported. Table 1 Methods of synthesizing LaB6 Introduction Molten salt electrolysis LaB6 is a cubic deep purple metallic compound and finds i La2O3+CaCl2+CaB4O77,8 application as an electron emitter in electron microscopes and ii LaCl3+Li2B4O7+LiCl9 as a promising cathode material because of its low work iii La2O3+B2O3+LiF+Li2O5 function (2.6 eV), high current (29A cm2 K-2) and voltage iv La2O3+Li2B4O7+Li3AlF610 capability.1–5 Borides of rare earth metals like Sm, Ce, Y, Yb, v Mixture of oxide, borate and fluoride melts11–13 Solid phase reaction Dy, Ho etc., exhibit a wide variety of interesting magnetic, La powder+amorphous B powder mixed, pelletized and fired at electric and transport properties.6 1375 °C in H2 atmosphere for 1 h then at 1800 °C in H2–Ar in a The borides are not easy to prepare in pure form and the graphite crucible4,14 purification steps are also often diYcult and precise stochiome- Arc melting try is also often hard to achieve.Usually the borides are Arc melting pressed B powder with three-fold excess La powder to give LaB4 then heating LaB4 and B in vacuo at 1600 °C for prepared by high temperature reaction of the constituent 15 min15 elements in powder or pelletized form around 1800 °C initially Flux growth to form LaB4 and then to give LaB6 under controlled Solidification of Al flux containing La and B16,17 conditions.Vapour phase crystallization18 However, these methods are not only expensive but the Floating zone products are also found to be contaminated with the carbides Mixture of respective oxides with B and heated in an induction oven in a tantalum crucible in vacuo, below 10-3 mmHg at 1700 °C for from the crucible. 1 h2,19 Crystals of LaB6 with high melting point (2500 °C) can be conveniently synthesized at lower temperature (850 °C) from molten salt electrolysis. In molten salt electrolysis the driving force for crystallization of LaB6 is provided by the potential (15 mass%)–B2O3 (48 mass%)–LiF (18 mass%)–Li2O (19 gradient as compared with the temperature gradient in the mass%).flux growth. A potential controlled experiment has been The B2O3 may act as a fluxing agent in addition to taking attempted and the results are reported here. part in the cell reaction while Li2O may break down B2O3 DiVerent methods of preparation of LaB6 are summarized aiding B electrodeposition. The viscosity of the melt was in Table 1.7–17 reduced by the addition of LiF.The constituents of the melt in the form of powder were mixed as per the predetermined mass ratio and pressed in the form of pellets of 30 mm diameter, and 20 mm thickness in a Experimental steel die at a load of 5 ton cm-2. The loss of electrolyte by With the aim of synthesizing cubic crystals of LaB6 the volatilization during the melting process was found to be electrolysis was carried out in a high density graphite crucible appreciably reduced by this technique.(Porosity 16%, supplied by M/s. Graphite India Ltd.) with an The graphite crucible containing the electrolyte pellets was inner diameter of 50 mm and depth 80 mm, which served placed in a stainless steel vessel and subjected to predrying at simultaneously as the anode.The cathode was a 10 mm 500 °C for 5 h under partial vacuum by means of an oil diameter Mo rod threaded to a stainless steel rod and was vacuum pump. The electrolyte was then melted slowly under centrally positioned in the melt maintained at 850 °C. a continuous flow of N2 (after passing through hot copper The crucible was placed at the lower end of a cylindrical turnings followed by molecular sieves) which flushed the stainless steel vessel, as shown in Fig. 1, having a vacuum tight system. The melt was equilibrated at 850 °C for ca. 1 h top flange sealing with a water cooling arrangement. The preceding electrolysis. vessel was surrounded by a wire wound furnace with tempera- After the Mo cathode was centrally positioned and immersed ture control.The flange contained provision for fixing elec- into the melt, experiments were carried out at diVerent potentrodes, inlet, outlet for circulating N2 gas and a thermocouple. tials, with a cathode current density of ca. 500–600 mA cm-2 The electrode could be raised or lowered through a special to compare the nature and composition of the deposits at Teflon Swagelok seal arrangement at the flange.varying potentials and results are given in Table 2. Electrochemical studies5 of a similar bath composition under The melt used for the electrodeposition consisted of La2O3 J. Mater. Chem., 1998, 8(10), 2205–2207 2205Fig. 2 SEM micrograph of LaB6 crystals. mixture at 950 °C in a Pt crucible. The boron content was determined by titrating the sample solution, to which mannitol was added, with a standard NaOH solution.Lanthanum was estimated by precipitating it as oxalate with saturated oxalic acid solution followed by dissolution of the precipitate in warm dilute H2SO4 and titrating the solution with standard KMnO4. The La and B contents in the synthesized samples Fig. 1 Electrolytic cell for growth of LaB6 crystals. are given in Table 2.Table 2 Potential range, nature and composition of the deposit XRD analysis Potential Compositiona (%) A powdered sample of LaB6 was subjected to XRD analysis Experiment range in code volts La B Nature (using Cu-Ka radiation) and the XRD pattern is shown in Fig. 3. The lattice constant of LaB6 was determined from the 1.90–1.93 64.84 25.63 clusters of deep GAP 23 XRD data (Table 3) and the value a=4.156 A° was found to (68.17) (31.83) and light match the reported value.20 The absence of any additional purple crystals GAP 30 2.10–2.40 62.28 26.67 clusters of deep (68.17) (31.83) purple cubic crystals aRequired values in parentheses.a ramped applied voltage indicated a decomposition potential of 1.85 V. Experiments carried out at potentials of 1.90–1.93 V were found to yield products with higher La content as compared with the product from electrolysis at higher potentials, which may be attributed to recombination.Results and discussion La and B are simultaneously reduced at the cathode to form Fig. 3 XRD pattern of LaB6. LaB6 while oxides of carbon are evolved at the anode. After electrolysis the Mo cathode covered with the deposit was raised above the melt level and allowed to cool in an N2 Table 3 XRD data corresponding to cubic LaB6 (Cu–Ka radiation) atmosphere before being taken out of the cell.The as grown deposit was scraped oV onto a glass plate and electrolyte 2h/° dobs/A° dstd/A° I/I0 hkl adhering to the boride was leached with a warm 5% HCl solution followed by a 2% NaOH solution and then washed 21.486 4.130 4.149 82 100 30.457 2.928 2.941 100 110 with distilled water.The current eYciency of the process was 37.565 2.390 2.398 53 111 found to be in the range 85–90%. 43.684 2.070 2.079 40 200 49.061 1.854 1.862 41 210 Nature of the deposit 54.160 1.691 1.966 29 211 63.381 1.466 1.471 15 230 The deposits were found to consist of clusters of cubic deep 67.791 1.381 1.385 27 300 purple crystals which ranged from 25 to 250 mm as shown by 71.882 1.312 1.314 20 310 SEM (Fig. 2). 75.970 1.251 1.253 13 311 80.059 1.196 1.199 08 222 84.037 1.151 1.153 09 320 Chemical analysis Crystal data for LaB6: M=180.20, cubic, a=4.145 A° (4.153 A° 20), A known quantity of powdered sample of LaB6 was dissolved density=4.75 g cm-3. in dilute HCl after fusing with an AR Na2CO3–NaNO3 2206 J.Mater. Chem., 1998, 8(10), 2205–2207peaks in the XRD pattern and the formation of the product Athinarayanasamy and Sri. A. Mani for providing XRD characterization. in the form of cubic crystals indicated its high purity. Density References The density of the crystals was determined pycnometrically21 1 H. Ahmad and A. N. Broers, J.Appl. Phys., 1972, 43, 2185. using xylene as the liquid medium and the measured value of 2 B. J. Curtis and H. GraVenberger, Mater. Res. Bull., 1966, 1, 27. 3 I. Batko, M. Batkova, K. Flachbart, V. B. Flippor, Yu. B. 4.75 g cm-3 is in good agreement with the reported value.2 Paderno, N. Yu. Schicevalova and Th.Wagner, J. Alloys Compds., 1995, 217, L1. Mechanism of boride deposition 4 J.M.LaVerty, J.Appl. Phys., 1951, 22, 299. 5 I. V. Zubeck, R. S. Feigelson, R. A. Huggins and Petit, J. Cryst. Various mechanisms have been proposed for boride deposition. Growth, 1976, 34, 85. According to Meerson and Smirnov22 this occurs by simul- 6 H. K. Blomberk, M. J. Merisalo, M. M. Kosukora and taneous discharge of refractory metal and boron, e.g. Ti4+ V. N. Gurin, J.Alloys Compds., 1995, 217, 123. and B3+ form from a borate–fluoride bath in a primary 7 T. Kuroda, Researcher of the Electrochemical Lab., Tokyo, Japan, process. Antony and Welch23 suggested electrodeposition of 1957, vol. 561, p. 63. 8 K. Uchide and M. Shiota, Surf. Technol., 1978, 7, 299. boron followed by reaction with a refractory compound e.g. 9 A. Wold, Air Force Mater.Lab. Tech. Rept. AFML TR 239, ZrB2, whereas Andrieux24 suggested reduction to metal of the 1967, p. 5. refractory metal oxide and B2O3 is accomplished by alkali or 10 K. Uchide, Surf. Technol., 1978, 7, 137. alkaline earth metals which first deposit over the cathode from 11 L. Andrieux, Ann. Chim., 1929, 12, 423. the respective metal halide flux. For example, Ca formed from 12 J. L.Andrieux and D. Baebetti, C.R. Acad. Sci. (Paris), 1932, calcium salts chemically react with B2O3 to give CaB6. 194, 1573. 13 J. L. Andrieux, Ann. Chim (Paris), 1929, 10, 423. For the formation mechanism of LaB6, it is possible to 14 J. R. Rea and E. Kostiner, J. Cryst. Growth, 1971, 11, 110. postulate that electrolytically reduced La reacts either with 15 L. W. Johnson and A.H. Daane, J. Phys. Chem., 1961, 65, 909. electrolytically reduced B or with B2O3, although it is not 16 T. Aita, U. Kawabe and Y. Honda, J. Appl. Phys., 1974, 13, 251. clear, as yet, which mechanism operates. 17 M. M. Korsukova and V. N. Gurin, Mendeleev Chem. J., 1981, 26, 114. 18 T. Niemyski and E. Kierzek-Pecold, J. Cryst. Growth, 1968, 3/4, Conclusion 162. 19 T. Niemyski, I.Procka, J. Jun and J. Paderno, J. Less Common Purple coloured cubic crystals of LaB6 could be favourably Met., 1968, 15, 97. synthesized by electrodeposition employing an oxyfluoride 20 Von Stackeburg and M. Neumann, Z. Phys., Chem. B, 1932, 19, melt consisting of La2O3–B2O3–LiF–Li2O maintained at 314. 850 °C under an N2 atmosphere in a graphite container which 21 J. Johnson and L. H. Adams, J. Am. Chem. Soc., 1912, 34, 563. 22 G. A. Meerson and M. P. Smirnov, Khim. Redk. Elem. Acad. simultaneously acts as an anode with a centrally positioned Nauk. SSSR Inst. Neorg Khim., 1995, 2, 130 (Chem. Abstr., 1956, Mo rod as cathode operating at 1.90–1.93 V with a cathode 50, 3122). current density of 500 mA cm-2. 23 K. E. Antony and B. J.Welch, Aust. J. Chem., 1969, 22, 1993. 24 J. L. Andrieux, Chim. Ind., 1932, 411 (Chem. Abstr., 1932, 26, The authors are very grateful to the Department of Science 3442). and Technology (DST), New Delhi, for financial support for this work. Our thanks are also due to Sri. K. Paper 8/02895H J. Mater. Chem., 1998, 8(10), 2205–2207 2207

ISSN:0959-9428    

DOI:10.1039/a802895h

出版商:RSC    

年代:1998

数据来源: RSC